key: cord-0025145-ujj973hd
authors: nan
title: Proceedings of the World Molecular Imaging Congress 2021, October 5-8, 2021: General Abstracts
date: 2022-01-04
journal: Mol Imaging Biol
DOI: 10.1007/s11307-021-01693-y
sha: bb89f899a1d35c68d74a9c09548ab3977cd25241
doc_id: 25145
cord_uid: ujj973hd

nan

. 2D fluorescence imaging of biodistribution and tumor targeting of anti-HER2 Nb (2RS15dNOTAG) randomly labeled with 3 distinct Cy5-derivatives in mice bearing subcutaneous HER2+ SKOV3 tumors on right hind limb evaluated over 24h after injection.

Although TDM1 accumulation in CAV1-low PDXs was similar in control and lovastatin cohorts (Fig1C), TDM1-tumor accumulation was 1.8-fold and 1.4-fold higher in CAV1-high PDX #1 (CAV1, IHC 3+) and PDX #2 (CAV1, IHC 2+), respectively. Lovastatin enhances antibody efficacy and improved mice survival in a CAV1-high PDX (TP53 and KRAS mutant with HER2 IHC 3+) obtained from a patient who did not respond to trastuzumab therapy (Fig1E) . ADCC experiments indicate that lovastatin enhances antibody efficacy, which depends on cytokinemediated NK expansion and antibody's Fc domain (Suppl. Fig1c) . The preclinical findings were further validated in the retrospective analyses showing that statin users treated with trastuzumab had higher survival than non-statin users (p = 0.02; Suppl. Fig1d). Conclusions: This study identifies how tumoral CAV1 expression and statin use relate to GC response in HER2-targeted imaging and therapeutic approaches. In addition to show the preclinical synergy of CAV1 depletion with anti-HER2 antibodies, we demonstrate that statin users had better clinical responses to antibody-based therapies in HER2+ heterogeneous GC. [11C] ROCK201 showed fast brain kinetics and good specific binding signals in rats and NHP, which allowed reliable estimation of kinetic and binding parameters using 1TC. Further validation studies are ongoing to confirm the correlation between the PET signal and ROCK protein levels in different brain regions. [11C] ROCK201; (b) Whole brain TACs of [11C] ROCK201 in Sprague-Dawley under baseline and self-blocking (ROCK201, 3.33 mg/kg, i.v.) conditions; (c) Summed SUV PET images of [11C] ROCK201 in the brain of a rhesus monkey from 15 to 30 min imaging window under baseline and blocking (ROCK201,0.05 mg/kg, i.v.) conditions; (d) TACs from selected regions in the rhesus monkey brain under baseline and self-blocking (ROCK201, 0.05 mg/kg, i.v.) conditions. Abstract Body : Introduction: The use of noninvasive ultrasound (US) for in vivo tumor tissue characterization is an exciting prospect for anticancer treatment response monitoring. A new technology termed H-scan US imaging has been developed and links differences in the raw backscattered US signals to various-sized tissue structures. The purpose of this research was to develop a 3-dimensional (3-D) H-scan US imaging system and method for tissue characterization in volume space using a preclinical animal model of breast cancer. Further, the proposed H-scan US imaging technique has been validated using matched histological data. Statement of Contribution/Methods: Preliminary studies were conducted using female nude athymic mice (N = 15, Charles River Laboratories) implanted in the mammary fat pad with 1 million breast cancer cells ATCC) . Once tumors reached about 1 cm in size, animals were sorted into three groups so that mean tumor size in each was comparable (N = 5 per group). Animals were then US imaged at baseline and before receiving intraperitoneal injections, namely: (1) 0.3 mg sterile saline (control), (2) 0.2 mg of agnostic TRA-8 monoclonal antibody to human death receptor 5 (DR5) + 0.1 mg sterile saline, and (3) 0.1 mg TRA-8 + 0.2 mg paclitaxel. Image data was acquired using a programmable US scanner (Vantage 256, Verasonics Inc) equipped with a volumetric imaging transducer (4DL7, Vermon) at baseline and again every 24 h for 3 d. To generate the H-scan US images, a set of Gaussian-weighted Hermite filters were convolved with the radiofrequency (RF) data to measure the relative strength of the received signals. Finally, the lower frequency backscattered signals were assigned to a red (R) channel and the higher frequency components to a blue (B) channel. The unfiltered original RF signal was assigned to the green (G) channel to complete the RGB colormap and 3-D H-scan US image display. After US imaging on day 3, animals were humanely euthanized and tumors excised for histological processing using immunofluorescent staining for DAPI. Result/Discussion: The in vivo results show that 3-D H-scan US imaging is considerably more sensitive to tumor changes after neoadjuvant treatment as compared to traditional B-scan US. While there was no difference at baseline (p = 0.52), repeat H-scan US results from treated tumors exhibited percentage decreases in percentage of image intensity (-19.8 ± 8.2 %, and -62.2 ± 11 .7% for the TRA-8 and TRA-8 + paclitaxel treated animals at day 3, respectively; p 0.5, p Conclusions: 3-D H-scan US imaging is a promising technique that allows visualization of the heterogenous tumor tissue microenvironment and allows in vivo detection of an early treatment response within days of chemotherapy administration.

References: 1. Yousaf T, Dervenoulas G, Politis M. Advances in MRI Methodology. In: International Review of Neurobiology. Vol 141. Academic Press Inc.; 2018:31-76. doi:10.1016/bs.irn.2018 .08.008 2. Errante Y, Cirimele V, Mallio CA, Di Lazzaro V, Zobel BB, Quattrocchi CC. Progressive increase of T1 signal intensity of the dentate nucleus on unenhanced magnetic resonance images is associated with cumulative doses of intravenously administered gadodiamide in patients with normal renal function, suggesting dechelation. Invest Radiol. 2014; 49 (10) Abstract Body : Introduction: Bladder cancer (Bca) is the fourth most common type of cancer in men in the United States and is associated with increasingly concerning morbidity and mortality. The prevalence of bladder cancer increases with age, making early detection of Bca extremely important as disease progression is rapid. Common imaging modalities for assessing the extent of Bca include computed tomography (CT), ultrasonography (US), magnetic resonance imaging (MRI), all of which show evidence of enhancing diagnostic accuracy in patients [1] . However, these techniques do not account for phenotypic alterations caused by genomic changes during tumor development, and such changes would only be detected on a morphological level. Immunopositron emission tomography (immunoPET) has emerged as a pivotal imaging modality to assess Bca due to the robust specificity and sensitivity of the technique [2] . This study examines the performance of the radioimmunoconjugate [89Zr]Zr-DFO-Panitumumab and its ability to detect EGFR expression in bladder tumors, a potent oncogene commonly altered in Bca [3] . Methods and Results: Panitumumab was conjugated with the acyclic bifunctional chelator DFO-Bz-SCN and then radiolabeled with the positronemitting radionuclide zirconium-89. The radioimmunoconjugate prepared had a radiochemical yield of 90.36 %, a radiochemical purity of 99.18 %, and specific activity of 25 MBq/nmol.

[89Zr]Zr-DFO-Panitumumab displayed specific binding to EGFR-expressing UMUC3 cells with a Bmax value of 5.9 x 104 EGFRs/cell in vitro. ImmunoPET/CT in vivo images show localization of the antibody in both subcutaneous and orthotopic UMUC3 xenografts (Fig. 1) . Imaging results in a mouse bearing subcutaneous EGFR+ UMUC3 Bca cells at 72 h postinjection of [89Zr]Zr-DFO-Panitumumab demonstrated 0.98 ± 0.20 and 15.05 ± 2.40 % ID/g (percentage of injected dose per gram) in the non-tumor bladder and subcutaneous tumor, respectively. In the orthotopic UMUC3 xenografts, imaging results at 48 h post-injection of [89Zr]Zr-DFO-Panitumumab showed higher antibody-tumor uptake in the bladder-tumor when compared with non-tumor cohorts, whose accumulation correlated with respective tumor volume (r = 0.99, P = 0.0068, Spearman's correlation). These observations were also confirmed with ex vivo biodistribution analysis. Conclusions: These preclinical data show the potential of EGFRtargeted immunoPET to image and, therefore, diagnose BCa with altered protein levels of EGFR. 

Country: United States antibody-MMAE conjugate (PSMA-ADC) which was withdrawn from clinical trial due to toxicity [10, 11] . In in vitro studies, PSMA-1-VcMMAE was 48-fold more potent in killing PSMA-positive PC3pip cells than killing PSMA-negative PC3flu cells, while free MMAE showed no selectivity in killing PC3pip and PC3flu cells. In in vivo studies, PSMA-1-VcMMAE significantly inhibited tumor growth leading to prolonged animal survival in different animal models, including metastatic prostate cancer models. The maximum tolerated doses (MTD) of MMAE, PSMA-ADC and PSMA-1-VcMMAE using 20% body weight loss as the reference were 700nmol/kg, 640nmol/kg and 7640nmol/kg, respectively. According, the therapeutic indexes of MMAE, PSMA-ADC and PSMA-1-VcMMAE were 1, 12.8 and 20 . It is speculated that replacement of antibody by small molecule ligand changes the pharmacokinetics of the conjugates, resulting in overall improved therapeutic index of PSMA-1-VcMMAE. In summary, the small molecule-drug conjugate reported here can be easily synthesized, are more cost efficient and have a better therapeutic index than antibody-drug conjugates. Abstract Body : Background The coronavirus pandemic, has resulted in more than 153 million confirmed cases and over 3 million deaths [1] , along with morbidities. Infection of the SARS-CoV-2 virus produces a wide variability of outcomes, ranging from asymptotic to loss of life. Predicting the occurrence and severity of COVID-19 is an important area of research. COVID-19 disease is primarily a respiratory one. Computed tomography (CT) imaging of the lungs has elucidated aspects of the disease including peripheral ground-glass opacities, crazypaving pattern and consolidation that relate to severity. Polymerase chain reaction (PCR) testing has become ubiquitous for detection of disease, but CT imaging can demonstrate abnormalities in the lungs prior to positive PCR results [2] . In this study, a combination of CT imaging radiomic features and clinical pathology features were selected as input to machine learning based (ML-based) predictive models to classify SARS-CoV-2 exposure versus mock exposure in nonhuman primates (NHPs). Methods Across three experiments, 12 crab-eating macaques (Macaca fasciularis) were exposed to SARS-CoV-2 and another 12 were instilled with vehicle only (mock). The macaques were observed for development of clinical signs of infection. Scans by CT (Gemini PET/CT, Philips Healthcare) were performed prior to exposure and on Day 2, Day 4, Day 6 and Day 8. Blood sampling, performed on the same days as CT scanning, provide concomitant serum chemistry, hematology and immunology laboratory values. CT and clinical pathology data from Day 2, Day 4 and Day 6 were selected for building machine learning models. The lung field was automatically segmented from the CT scan using a deep learningbased method that was initially developed for liver and modified for lungs [3] . Using PyRadiomics within 3D slicer, 110 radiomic features were extracted from the lung [4] . 61 clinical pathology features were collected. Due to the limited number within this study's sample set (72; 24 NHPs with three timepoints each), the minimum redundancy, maximum relevance (MRMR) [5] feature selection was employed to reduce the features to the 10 most relevant and least redundant for each category (radiomics, clinical pathology). Multiple ML algorithms were trained ( Figure 2 ) using these selected features to determine which had the best predictive power. Eighty percent of the samples were used for training while the remainder was reserved for testing. During training, six-fold cross validation was performed, and the following measurements of predictive power were estimated: accuracy, sensitivity, and specificity. Results While disease severity was mild, showing no cage-size changes in behavior, Day 2, Day 4 and Day 6 demonstrated the largest abnormalities on CT. Figure 1 shows an axial CT section of the lung with the deep learning-based lung segmentation in blue and lesion segmentation in cyan [3] . Figure 2 shows the 10 radiomic and clinical pathology features selected by the MRMR algorithm and the overall accuracy, sensitivity and specificity of the 5 machine learning algorithms evaluated. Random forest using radiomic selected features had an accuracy of 92.86%, sensitivity of 100% and specificity of 85.71%. Conclusions The SARS-CoV-2 virus exposure elicited mild disease in crab-eating macaques, making prediction analysis difficult. The selected radiomic features outperformed the selected clinical pathology features in predicting which samples were from SARS-CoV-2-exposed versus mock exposed NHPs. The random forest ML model outperformed the four other algorithms evaluated. Future efforts will focus on combining radiomic and clinical pathology features into a single predictive model as well as exploring additional feature selection techniques and ML models, such as least absolute shrinkage and selection operator (LASSO). Furthermore, we hope to use these observations to interrogate the predictive relationships between imaging features and disease heterogeneity and outcome, including histopathology, as more severe disease models become available. Table illustrates list of top 10 selected features using mRMR technique from radiomic and clinical pathology data sets. Accuracy measurements for 5 machine learning algorithms (linear discriminant analysis (lda), k-nearest neighbor (knn), random forest (rf), support vector machine (svm) and classification and regression trees (cart)) are shown on the right. Random forest (rf) classifier had the best performance. MIP1 alpha is macrophage inflammatory protein; IL is interleukin; IFN is interferon. 

Elena Rama, RWTH Aachen University, erama@ukaachen.de Category: Bioengineering, Synthetic Biology and Basic Biology Abstract Body : The longitudinal monitoring of the remodelling and the pathological dysfunctions of implanted vascular prosthesis remains challenging. To overcome this issue, novel approaches toward the incorporation of MR contrast agents into the textile scaffold of vascular prostheses have already been explored [1, 2] . Nonetheless, the transition from late-invitro bioreactor maturation to early-in-vivo implantation of vascular implants remains inadequately investigated. Therefore, this study aimed to further develop in vitro non-invasive imaging methods to monitor the passage between these two critical phases. For this purpose, we developed a biohybrid tissue-engineered vascular graft (TEVG) consisting of biodegradable poly(lactic-co-glycolic acid) (PLGA) fibers labeled with ultrasmall superparamagnetic iron oxide nanoparticles (USPIO), non-degradable polyvinylidene fluoride (PVDF) scaffold, and fibrin gel containing endothelial and smooth muscle cells. The degradation of the USPIO-labeled PLGA fibers was longitudinally monitored via MRI. Moreover, the deposition of newly synthesised extracellular matrix (ECM) components, as replacement of the degraded PLGA fibers, was assessed in TEVGs by molecular MR (mMR) using elastin-and collagen type I-targeted Gdcontaining MR molecular probes (Gd-DTPA-ESMA and EP-3533) in comparison to untargeted Gd-DTPA (Magnevist). The tissue response was additionally investigated in terms of endothelial coverage by molecular US (mUS) using RGD-targeted poly(butyl cyanoacrylate) microbubbles (RGD-MB) targeting αvβ3 integrin as a marker of inflammatory reaction and thrombus formation. RGD-MB binding was longitudinally evaluated in TEVGs and compared to RADcontrol MB used as control. A competition study was also performed by injecting free RGD before the injection of RGD-MB. The longitudinal monitoring of the degradation of USPIOlabeled PLGA fibers was performed by embedding the fiber sheets in fibrin gel and incubating them at 37°C for 8 weeks and at 70°C for 3 weeks. Furthermore, their degradation was evaluated in hybrid conditions. First, USPIO-labeled PLGA fibers were monitored in combination with the PVDF non-degradable scaffold for 3 weeks and, subsequently, combined with cellular components for 7 days (A). Quantitative R2 analysis reflects the correlation between the degradation of the USPIO-labeled PLGA fibers and the decrease of their relaxation rate values over time due to the release of the USPIO. Interestingly, the degradation rate of the PLGA fibers accelerated in combination with the PVDF scaffold and, subsequently, became even faster in combination with cells (B-C). mMR monitoring of the deposition of newly synthesised ECM components showed no Gd-DTPA-ESMA binding in TEVGs that was, furthermore, confirmed through immunofluorescence microscopy (D,J) . A 3-fold +/-higher binding of the Gd-DTPA-ESMA probe was instead detected in human umbilical arteries (HUA) used as positive control. In contrast, a strong signal, given by the binding of EP-3533 to collagen type I, was observed in TEVGs and, subsequently, validated by histology(E-F,J). αvβ3 integrin expression was longitudinally monitored via mUS after 0, 3, and 14 days of bioreactor maturation of the TEVGs (G). Of great relevance, no difference between RGD-MB and RAD-control MB binding was observed showing that the shear stress used during the bioreactor conditioning of the TEVGs did not cause distress to the endothelial cells (H). Therefore, TNF-α was used for pre-activating the endothelium and to mimic an inflammatory state. In this respect, RGD-MBs showed a 70-fold +/-higher binding expressed as differential targeted enhancement (d.T.E.) (I). Further histological analyses confirmed the expression of αvβ3 integrin only after TNF-α preconditioning (J). Herein, we present in vitro and non-invasive imaging approaches to monitor the late-in-vitro maturation phase as well as the early-in-vivo implantation phase of biohybrid tissue-engineered vascular grafts. Moreover, our findings demonstrate the replacement of the PLGA fibers by deposition of collagen type I and the detectability of endothelial inflammation via RGD-MB targeting. To conclude, the presented imaging methods may be highly valuable to foster the clinical translation of TEVGs.

References: [1] (Magnevist) . E-F: No Gd-DTPA-ESMA binding and, therefore, no elastin is observed in TEVG. In contrast, strong EP-3533 binding to collagen type I is observed in TEVG (E) and HUA (F). G: TEVG molecular US investigation comparing RGD-MB to RAD-control MB binding in untreated and TNF-α treated TEVGs. Images of the competition study are also shown. H: Longitudinal US monitoring of untreated TEVG shows no difference between RGD-MB and RAD-control MB binding, and no changes in d.T.E. are observed in RGD-MB binding after free-RGD pre-injection. I: 70-fold higher binding for RGD-MBs is observed after TNF-α treatment of TEVGs. J: Immunofluorescence representative images of elastin, collagen type I, and αvβ3 integrin expression in TEVGs and HUA. (*p < 0 .05, **p < 0 .001 and ***p < 0 .0001) (Stars indicate the TEVGs lumen).

First Name: Elena Last Name: Rama Email: erama@ukaachen.de

Country: Germany ID: GA71 Pulsed Focal ultrasound as a non-invasive method to deliver exosomes in the brain/stroke Ahmet Alptekin, Augusta University, aalptekin@augusta.edu

Abstract Body : Exosomes, a component of extracellular vesicles, are shown to carry important small RNAs, mRNAs, protein, and bioactive lipid from parent cells and are found in most biological fluids. Investigators have demonstrated the importance of mesenchymal stem cells derived exosomes in repairing stroke lesions. However, exosomes from endothelial progenitor cells have not been tested in any stroke model, nor has there been an evaluation of whether these exosomes target/home to areas of pathology. Targeted delivery of intravenous administered exosomes has been a great challenge, and a targeted delivery system is lacking to deliver naïve (unmodified) exosomes from endothelial progenitor cells to the site of interest. Pulsed focused ultrasound is being used for therapeutic and experimental purposes. There has not been any report showing the use of low-intensity pulsed focused ultrasound to deliver exosomes to the site of interest in stroke models. In this proof of principle study, we have shown different parameters of pulsed focused ultrasound to deliver exosomes in the intact and stroke brain with or without intravenous administration of nanobubbles. The study results showed that administration of nanobubbles is detrimental to the brain structures (micro bleeding and white matter destruction) at peak negative pressure of >0.25 megapascal , despite enhanced delivery of intravenous administered exosomes. However, without nanobubbles, pulsed focused ultrasound enhances the delivery of exosomes in the stroke area without altering the brain structures. 

Ramanpreet Sembhi, The University of Western Ontario, rsembhi@uwo.ca Category: Neuroscience Abstract Body : INTRODUCTION: Inhaled hyperpolarized (HP) 129Xe magnetic resonance imaging (MRI) is a non-invasive imaging method, which is currently used to measure lung structure and function.1,2 This MRI approach provides a way to obtain simultaneous ventilation/perfusion (V/P) lung measurements of functional gas exchange within the lungs due to high natural solubility of xenon in lung tissue compared to other imaging gases. Due to various physical properties of the 129Xe isotope, it can serve as a new probe for brain blood flow, grey and white matter mapping3,4 and functional measurements. These measurements are possible due to distinct and large range of chemical shifts (~200ppm) of 129Xe when residing within lung tissue, brain tissue and red blood cells compared to the gas phase. 129Xe-based imaging aims to improve sensitivity where traditional MRI may fall short, and by reaching beyond the resolution limitations of positron emission tomography (PET).

[15O]-water PET is the gold standard imaging method for determining cerebral perfusion. 5, 6 In this study, simultaneous 129Xe-based MRI and [15O]-water PET images were collected and compared. METHODS: For the initial validation phase in phantoms, in a 60mL plastic syringe, 30mL of hyperpolarized 129Xe gas was dissolved in (30mL) of [15O]-water solution. After dissolving, all leftover xenon gas was removed from the syringe. Hyperpolarized 129Xe gas was obtained from a turn-key, spin-exchange polarizer system (Polarean 9820 129Xe polarizer). 129Xe dissolved phase images were acquired in a 3T PET/MRI (Siemens Biograph mMR) scanner with RF coil tuned to 34.09 MHz. A fast Gradient Recalled Echo sequence was utilized using the following parameters: Matrix Size=104x48; Slice thickness=250mm; TE/TR=2.04/20ms; BW=660Hz/pixel; Flip angle= 110 (a Constant Flip Angle approach was used); and FOV= 150x150mm2.

[15O]-water PET data (half-life time of 2min; Matrix size=344x344; 10x60sec frames and OSEM with 3 iterations) were acquired simultaneously with 129Xe MRI for 600sec using the integrated PET system in a 3T PET/MRI. RESULTS: Figure 1 A, B , C, D respectively shows two consecutive 2D axial 129Xe MRI images and two (2D and 3D) [15O]-water PET images acquired simultaneously. 129Xe/PET images indicate that the diameter of the phantom from both PET and MRI images are similar clearly indicating the feasibility of the simultaneous hyperpolarized 129Xe MRI and [15O]-water PET measurements. 129Xe images demonstrate a sufficient SNR level suggesting that 3D 129Xe imaging is possible. DISCUSSION AND CONCLUSIONS: As per knowledge, this is the first ever demonstration of the simultaneous 129Xe MRI and [15O]-water PET scans. This demonstration will enable the next step, namely, in-vivo double tracer brain perfusion imaging. Important validation work for 129Xe-based brain perfusion techniques, directly and simultaneously with [15O]-water PET will be performed using a small animal model. The comparison results show similarity between both imaging modalities and tracers, moving towards the next step in validating the xenon imaging technique as a potential for brain perfusion measurement. The ability to use 129Xe as a non-radioactive tracer, providing similar and complimentary information as [15O]-water PET may be a much more cost-effective alternative to PET for imaging stroke, brain cancer and other brain diseases and will significantly increase the number of the 129Xe MRI clinical applications.

References: 1 Kaushik, S. S. et al. Single-breath clinical imaging of hyperpolarized (129)Xe in the airspaces, barrier, and red blood cells using an interleaved 3D radial 1-point Dixon acquisition. Magn Reson Med 75, 1434 -1443 , doi:10.1002 /mrm.25675 (2016 . 2 Kaushik, S. S. et al. Probing the regional distribution of pulmonary gas exchange through single-breath gas-and dissolved-phase 129Xe MR imaging. J Appl Physiol (1985 Physiol ( ) 115, 850-860, doi:10.1152 Physiol ( /japplphysiol.00092.2013 Physiol ( (2013 . 3 Kershaw, J. et al. Confirming the existence of five peaks in 129Xe rat head spectra. Magn Reson Med 57, 791-797, doi:10.1002 Med 57, 791-797, doi:10. /mrm.21186 (2007 . 4 Wakai, A., Nakamura, K., Kershaw, J. & Kanno, I. In vivo MR spectroscopy of hyperpolarized Xe-129 in rat brain. International Congress Series 1265 Series , 139-143, doi:10.1016 Series /j.ics.2004 Series .04.063 (2004 Abstract Body : Objective: Genetically encoded optical sensors and advancements in microscopy have revolutionized the scientific toolbox used for answering complex biological questions especially in terms of studying brain activity. Most optical sensors currently used are based on fluorescence and are very useful for single cell imaging in superficial brain regions. However, there remains a need to develop new tools for reporting neuronal activity in vivo within deeper structures without the need for hardware to be implanted within the brain such as lenses or fibers. We expect this can be achieved by replacing the fluorescent elements of the existing biosensors with bioluminescent producing elements ( Figure 1A ). This will bypass the need of external light sources to illuminate the sensor and overcome limitations such as limited light penetration depth, light diffraction and tissue heating that are all associated with the excitation light used for fluorescent imaging. We are developing neurotransmitter and voltage indicators based on bioluminescent light emission that have robust changes in light output in response to neurotransmitter presentation or membrane depolarization that do not photobleach (figure 1B). The neurotransmitter indicators will allow neuroscientists to monitor activity associated with a specific neurotransmitter as it relates to behavior in a variety of psychiatric disorders. We are building on recent advancements in two fields: genetically encoded optical sensors for real-time sensing changes in neurotransmitter levels; and genetic evolution of new enzymes that produce biological luminescence at unprecedented intensities. Methods and Results: We have bioengineered the first bioluminescent neurotransmitter indicators using a multistep screening approach. First rational design using a variety of split luciferases and linker variants to fuse the luciferases to a sensing protein. Then an automated workflow to screen for improved variants in mammalian cells. The variants were cloned using Gibson assembly with primers that incorporate randomized amino acids in the linker regions between the two NanoLuc halves and the glutamate binding protein Glt1. Then used automation for high throughput mini preps and transfection of mammalian cells. Currently, we have produced the first bioluminescent neurotransmitter indicators by adapting sensing domains used in fluorescent indicators and incorporating bioluminescent producing luciferases. As a result, we created a BioLuminescent Indicator of the Neurotransmitter Glutamate (BLING). Using our protein evolution approach in mammalian cells, we were able to improve our BLINGs response by 2.5x from screening 1200 variants ( Figure 1D ). Our improved glutamate indicator demonstrated a 255% +/-50 increase in response to neurotransmitter presentation when expressed in mammalian cells (HEK cells), can successfully report changes in extracellular glutamate when used in a plate reader and outperforms the fluorescent sensors iGluSnFr and GCamP6 in plate reader assays ( Figure 1C-E) . Finally, to test the feasibility of imaging BLING we used real-time microscopy ( Figure 1F and 1G) which demonstrated up to a 310% increase in light emission detectable at the single cell level. Conclusion: We expect these indicators to perform well for molecular imaging neuronal activity in deep brain structures with imaging equipment that is readily accessible to researchers. We are also continuing with our protein engineering strategy to improve BLING, other neurotransmitter indicators, to develop improved voltage indicators and spectral variants. These indicators will be able to immediately benefit ongoing research efforts to study the mechanisms that give rise to a wide array of psychiatric disorders and provide researchers with significantly improved approaches to study neuronal activity at the level of the cell, network, and behaving animals longitudinally. Furthermore, since these reporters produce their own light, they can be used as activators for light sensitive proteins to carry out a variety of downstream functions within a cell.

Image/Figure Caption: A. Schematic of the BioLuminescent Indicator of the Neurotransmitter Glutamate (BLING) with a split luciferase (blue) and glutamate sensing domain (tan) displayed on the cell surface. B. Comparison of bioluminescent vs fluorescent imaging in superficial and deep brain regions. Bioluminescence light is able to be detected because no excitation light is needed and there is nearly no background noise or autofluorescence. C. Trace of BLINGs response to glutamate when expressed in HEK cells using a plate reader. D. High throughput screening of BLING linker variants produced a new optimized BLING. E. Optimized BLING sensor has 2.5x the response amplitude compared to the parental BLING and outperforms existing fluorescent indicators of glutamate iGluSnFr response and the response of GCamp6m to calcium. F. Example trace of a single cell response to 1000uM glutamate addition at 120 seconds. G. Bioluminescent microscopy comparing background luminescence to luminescence with 1000uM glutamate. Abstract Body : Objective: The creation of a novel synthetic circuit based on a "switch" that is controlled in a non-invasive manner can become a valuable tool for the fields of synthetic biology and molecular imaging. Many circuit designs rely on external light or the injection of a chemical to stimulate the circuit. These methods can be limited in their stimulation of systems due to light having poor penetration in tissue and chemicals lacking spatial and temporal control due to variability in diffusion and cellular uptake. A circuit activated by magnetic fields can overcome these challenges as fields are able to penetrate deep into tissue and create a homogenous stimulation to cells. The "switch" for this novel circuit design is based on the Electromagnetic Perceptive Gene (EPG; GenBank: MH590650.1) which was discovered in Kryptopterus vitreolus (glass catfish). When EPG is expressed in mammalian cells, electromagnetic fields (EMF) can elicit a measurable response in the cell 1. Here we introduce a novel method of controlling enzymes used in molecular imaging simply by applying EMF. Methods: All cloned constructs were produced using the EPG gene and Gibson Assembly creating fusion proteins attached to the C and N terminal of EPG. For bacterial expression we used pET101 plasmid and BL21 E. Coli and expression was induced using Magic Media (Thermo Fisher). For Mammalian cells expression we used pcDNA 3.1 plasmid and HeLa and HEK293T cells. FRET readings were done using VICTOR Nivo (Perkin Elmer), NanoLuc imaging was performed using IVIS (Perkin Elmer), and APEX2 was read using Cytation 5 (BioTek). Results: Utilizing Förester Resonance Energy Transfer (FRET) with NanoLuc and mVenus fused to either terminus of EPG, we observed in HeLa cells a change of 2-4% in N=15 wells the DF/F (change in ratio of mVenus to Nanoluc after stimulus) following constant static magnet stimulation in contrast to controls without stimulation in a single copy EPG (Persuasive data Figure P1A ) and a 10-15% in N=15 wells an increase in a DF/F when two copies of EPG were cloned in tandem (double EPG; Persuasive data Figure P1B ). An EPG split Nanoluc2 fusion protein was cloned using Gibson Assembly with the EPG fused inside of the split NanoLuc fragments. This construct was expressed in BL21 cells and measured using the IVIS with EMF induction from an electromagnet (~35 mTesla). Reads were taken every 10 seconds with an open filter for 10 minutes with stimulation period of constant EMF for 2 minutes after 2 minutes of initial reads. In this EPG split NanoLuc construct we observed a 40% increase in photon flux with cell lysates (Fig 1A) and 70% increase with whole cell Fig 1B during stimulation with EMF and decreased after EMF stimulation was removed. The effect of this protein has also shown promising results when fused to the split peroxidase APEX23. HEK 293FT cells that were stimulated for 30 minutes with a static magnet show an increase in enzyme activity over the non-stimulated control (Persuasive data Figure P2 ). Conclusion: We were able to demonstrate through FRET and a split protein approaches, the potential to control enzymes by using EPG as a "biological hinge" that allows bringing together two proteins or two parts of a split protein by remote EMF induction. This is the first step toward engineering a new platform ID: GA76 Imaging of olfactory nerve using a selective Nav1.7 sodium channel inhibitor In the diagnosis of smell disorders Dauren Adilbay, Memorial Sloan-Kettering Cancer Center, dr.dauren@gmail.com Category: Neuroscience Abstract Body : Background: Sense of smell (olfaction) provides a range of information that enriches our lives and, in some cases, protects us from harm or saves our lives. Disorders of smell can be observed in a wide range of severity from minor parosmia's to total loss of smell or anosmia. Anosmia/hyposmia affects 13.3 million people in the US alone according to the recent U.S. National Health and Nutrition Examination Survey (NHANES) [1] . Hundreds of thousands more people with persistent smell issues will be added to this number due to the COVID pandemic [2] . Currently, smell disorders are assessed using self-or physician-administered smell tests that are mostly subjective and unable to describe the level of dysfunction. Nav1.7 sodium channels are abundantly expressed in the olfactory nerve, both in the epithelium and the olfactory bulb. Patients with loss-of-function in SCN9A, the gene encoding Nav1.7, experience anosmia in addition to congenital insensitivity to pain [3] . Hsp1a is a recently discovered peptide that binds to Nav1.7 with high potency and selectivity [4] . We examined whether a fluorescenttagged version of Hsp1a could be used to visualize normal and damaged mouse olfactory nerves. Methods: We first assessed the possibility of visualizing the olfactory nerve in normal mice using Hsp1a-IR800, a tracer in which Hsp1a is covalently conjugated to the fluorescent dye IR-800. Athymic nude mice were intravenously injected with Hsp1a-IR800 (1 nmol, 100μL PBS), phosphate-buffered saline (PBS), or a combination of Hsp1a-IR800 and unmodified Hsp1a. To assess the difference in radiance efficiency in the normal and damaged olfactory nerve mouse model, we conducted a second experiment in which mice were injected with methimazole to model olfactory damage [5] . All mice were restrained using intraperitoneal ketamine/xylazine cocktail injection and in-vivo epifluorescence images obtained using an IVIS Spectrum (PerkinElmer) using a filter set (excitation 745 nm, emission 810 nm). Animals were sacrificed and olfactory nerve/bulb, muscle, heart, spleen, kidney, and liver were removed and imaged ex vivo using an IVIS imaging system. Autofluorescence was removed through spectral unmixing. Semiquantitative analysis of the Hsp1a-IR800 signal was conducted by measuring the average radiant efficiency (in units of [p/s/cm2/sr]/[μW/cm2]) in regions of interest (ROIs) that were placed on the region of olfactory epithelium/bulb. Results: Immunohistochemistry performed using a Nav1.7 antibody confirmed that mouse olfactory epithelium sensory neurons, axon bands, and olfactory bulb abundantly express Nav1.7 ( Figure 1 ). In epifluorescence in-vivo images in mice receiving imaging agent, the area of olfactory epithelium/bulb were clearly visible. The radiant efficiency was significantly less in both mice injected with PBS and in mice injected with the unmodified peptide in combination with imaging agent (blocking, Fig. 1 ). We observed a 150-fold increase in radiant efficiency compared to mice injected with PBS and a 61fold increase compared to blocking mice. Mice with damaged olfactory nerve had a visibly diminished radiance. The olfactory nerve area of methimazole-treated mice had a statistically significant 7-fold decrease in radiant efficiency compared to normal mice. Ex vivo images showed a similar statistically significant difference between mice groups in the olfactory epithelium/bulb region. The heart and kidneys are the only internal organs with higher fluorescence signals compared to control and blocked mice. Conclusion: We demonstrated that it is possible to image the mouse olfactory epithelium/bulb using Hsp1a-IR800, a fluorescent peptide that targets Nav1.7. Radiance efficiency was reduced in mice with a damaged olfactory epithelium/bulb suggesting that Hsp1a tracers hold promise in the diagnosis of smell disorders. Hsp1a-IR800/Hsp1a formulation, respectively. Images were taken 30 min after tail vein injection. (b) Fluorescence intensity quantification of panel a. (c) Epifluorescence images of animals. Control -normal mice injected with Hsp1a-IR800; Experimental -animals treated with methimazole 50 mg/kg on day 0, day 3 and injected with Hsp1a-IR800 on day 8; Block -Hsp1a-IR800/Hsp1a formulation respectively. Images were taken 30 min after tail vein injection (d) Fluorescence intensity quantification of panel c. (e) Histological confirmation of NaV1.7 expression in olfactory bulb and epithelium. The first row is the overview of the olfactory epithelium IHC slide, epithelium with 20x magnification, and IgG control. The second row is the overview slide with the cut at the level of the olfactory bulb, the slide with 20x magnification, and the IgG control. (B) . Western blots confirmed the presence of OATP1B3 at 90 kDa and NIS at 120 kDa in MRI/PET cells (C) . Functional tests with MRI showed over a 30% increase in average R1 of MRI/PET cells than in naive cells (p < 0 .05) (D) . Significantly higher (p < 0 .001) 

showing uptake of 18F-TFB uptake in NIS expressing tissues as well as our MRI/PET tumour (F). A T1-weighted MRI image showing contrast enhancement in the MRI/PET tumour and no contrast enhahncement in the naive tumour (G). Abstract Body : Purpose: Bladder cancer is a prevalent disease with high morbidity and mortality. Intravesicular therapy is a modality of treatment that injects therapeutic agents directly into the bladder to minimize systemic effects. Recent targeted agents have shown potential against bladder cancer. Our goal is to develop a preclinical modality that can be used to evaluate real-time distribution of potential intravesicular therapies. We hypothesize that 3-dimensional optical/CT of optically labeled therapies can be used as a preclinical tool to determine systemic exposure of locoregionally delivered agents. To validate real-time 3D optical/CT for distribution studies, we used Cy7 labeled albumin as a therapeutic surrogate to validate this technique for future use with optically labeled therapeutics. Material and Methods: Mouse albumin (ALB) (Sigma, MO) was labeled with Cy7 according to manufacture instruction (Lumiprobe, MD). MB49 (ATCC, VA) mouse urothelial carcinoma cells stably expressing the fly luciferase were implanted orthotopically into the bladder of C57BL/6 mice (B6). MB49 tumor growth was detected by bioluminescence imaging (3D optical CT scanner, MILabs, Netherlands) and mice were injected intravesicularly with Cy7-labeled ALB (Cy7ALB). 3D optical scans with microCT for anatomical reference were done 24h, 3 and 5 days after Cy7ALB injection. To confirm Cy7 tracer uptake by MB49 cells some animals were dissected 24h, 3 and 5 days after tracer injection, bladders were sectioned and imaged using RVL-100G Echo Revolve microscope (Echo, San Diego, CA). Results: We found that 80-90% of the mice injected orthotopically with MB49 murine tumor cells developed bladder cancers, as confirmed by bioluminescent imaging as early as 7 days after implantation. We infused Cy7ALB intravesicularly for 2 hours and imaged real-time distribution with 3D optical/CT imaging and found expected high concentration of NIR activity in the bladder but not other tissues, which was confirmed postnecropsy. Interestingly, we found trace fluorescent activity in the region of the bladder cancer after extended imaging if we did not wash out the Cy7ALB after treatment, possibly due to macropinocytosis. Conclusion: Our study demonstrated that 3D optical/CT is a promising way to evaluate real-time distribution of NIR-labeled intravesicular proteins. Furthermore, using optically labeled probes allows evaluation of tumor accumulation versus urothelial uptake, which is important when developing targeted therapeutics. Thus, Cy7-labeled therapeutic candidates in conjunction with 3D optical/CT can play a valuable role in preclinical screening of candidate therapies being considered for intravesicular delivery. Abstract Body : Introduction: In vivo imaging and tracking of cells can be used to noninvasively improve the accuracy, efficacy and safety of novel immune-modulatory treatments and cell therapies. Zirconium-89 has a long half-life (t1/2 = 78.4 h) and has therefore been applied for direct cell labeling and imaging of cells in vivo using positron emission tomography (PET), in particular with the tracer [89Zr]Zr-oxine [1] - [3] . However, the water insolubility of [89Zr]Zr-oxine and challenging synthesis method required may limit its implementation as a routine clinical imaging tool. This work has evaluated a novel cell labeling approach employing the desferrioxamine (DFO) chelator, commonly used to label antibodies for immuno-PET, attached to the cell penetrating peptide MPG ([89Zr] Zr-DFO-MPG△NLS) which is a fusion peptide with a nuclear localization sequence (△NLS), in direct comparison to [89Zr]Zr-oxine ( Fig.1) . Methods: DFO-MPG△NLS peptide was produced via solid phase peptide synthesis.

[89Zr]Zr-oxine was synthesized according to a published procedure using 0. 9-2.4 MBq of [89Zr]Zr(oxalate)4 [4] . DFO and DFO-MPG△NLS were labeled by adding 1.4-2.2 MBq neutralized [89Zr]Zr(oxalate)4, that was also used as a control for cell labeling, to 10-20 μg DFO and DFO-MPG△NLS, respectively. All 89Zr-labeled compounds were incubated at room temperature (RT).

[89Zr]Zr-DFO-MPG△NLS was purified via a preconditioned Sep-Pak Plus light C18 cartridge. Complete chelation of zirconium-89 for all compounds was observed using instant thin layer chromatography (iTLC) and in addition high performance liquid chromatography (HPLC) in the case of [89Zr]Zr-DFO-MPG△NLS. Jurkat T-cells, human peripheral blood mononuclear cells (PBMCs), isolated human T-cells and sheep mesenchymal stem cells (MSCs) at 5×106 cells/mL were labeled with 94-902 kBq [89Zr]Zr-oxine, 60-842 kBq [89Zr]Zr-DFO-MPG△NLS, 151-669 kBq [89Zr]Zr(oxalate) 4, Zr-DFO in a 1.5 mL Eppendorf vial in 500 μL phosphate buffered saline (PBS) by a 10-30 min incubation at RT. The cell associated activity and uptake were determined after 2 washes with PBS. Cell pellets were resuspended in the appropriate complete cell medium and cellular activity retention was examined at different time points post-labeling alongside viability by trypan blue exclusion. Results: All 89Zr-labeled compounds were synthesized at a radiochemical yield of >90%. The recovery of [89Zr]Zr-DFO-MPG△NLS after C18 cartridge purification was >90 % (n = 6). Both [89Zr]Zr-oxalate and [89Zr]Zr-DFO showed no significant uptake into Jurkat T-cells ([89Zr] Zr- Zr-oxine were 37.9 ± 6.1 % and 40.7 ± 13.0 %, respectively (Fig. 2  A) .

[89Zr]Zr-DFO-MPG△NLS exhibited fast cellular uptake kinetics in Jurkat T-cells cells independent of the incubation time as could be shown and has been described previously for [89Zr]Zr-oxine [2] , [5] (Fig. 2 B) .

[89Zr]Zr-DFO-MPG△NLS showed significantly higher retention in PBMCs and sheep MSCs at 24 h and a non-significant difference in cellular retention in T-cells and Jurakt T-cells compared to [89Zr]Zr-oxine ( Fig. 3 A) ; there was no difference in the cellular retention within Jurkat T-cells over 4 days, compared to [89Zr] Zr-oxine ( Fig. 3 B) . No significant loss of viability was observed throughout all experiments as indicated by trypan blue exclusion. Conclusion: This work is the first to show that 89Zr-labeled DFO can be utilized for intracellular cell labeling via conjugation to the cell penetrating peptide MPG and that similar or superior performance, compared to the state-of-the-art tracer [89Zr]Zr-oxine could be achieved.

[89Zr]Zr-DFO-MPG△NLS offers significant advantages over [89Zr]Zr(oxinate)4: excellent aqueous solubility, high specific activity of [89Zr]Zr-DFO, the additional purification step, and the potential to transfer the synthesis to an automated synthesis module. Future work will investigate the in vivo cell tracking over several days. Abstract Body : Introduction Presented metal-free MR contrast agent is based on phosphoruscontaining polymer with phosphorothioate group (pTMPC), which is extremely rare in living organisms [1] . This feature ensure 31P MR signal chemical shift from the biological phosphoruscontaining compounds, presented here by a polymer with phosphoester group (pMPC) and rat mesenchymal stem cells (rMSC). MR properties of the polymer were investigated using 4.7 T MR scanner with 1H/31P custom-made dual coils and 1.5 T relaxometer. Methods Presented probes pTMPC and pMPC were synthesized by controlled radical polymerization technique of the corresponding zwitterionic monomer. MR imaging and spectroscopy of probes were performed on 4.7 T scanner using dual 1H/31P radiofrequency surface coil intended for small laboratory animals. 1H-MR imaging (RARE sequence, TR/TE=2500/12 ms, FOV=10 cm) was applied for localization of phantom (1 mL, cP = 100 mmol L−1). T1 (single pulse sequence, TR=200-4000 ms) and T2 (CPMG spin lock sequence, TE=2-1200 ms) phosphorus relaxation times were measured. For 31P-MRI chemical shift imaging was used (CSI, TR=500 ms, scan time ST=15 min-3 h, field of view FOV=3.6 cm, resolution 2.25×2.25×3.6 mm). The same parameters were used to compare the signal from pTMPC (200 μL, cP = 100 mmol L−1) and rMS cells (200 μL, 1.25×108 mL−1) using more sensitive 1H/31P solenoid coil. 1H/31P-MRI overlapping and phosphorus MRI processing were obtained using ImageJ software. Signal-tonoise ratio (SNR) was calculated for MRI and MRS data. 1H relaxation times were obtained using 1.5 T relaxometer (cP = 10-100 mmol L−1). AlamarBlue Assay was used for cytotoxicity testing of the polymer (cP = 10-0.01 mg mL−1) using primary human fibroblasts (HF) and rMSC. Results and Discussion Results confirm that pTMPC do not influence 1H-MR T1/T2relaxation times (2510/2067.2 ms). 31P-MR T1/T2 relaxation times (2018.3/119.9 ms) were found to be adequate for further MR experiments. Between the polymers with phosphorothioate and phosphoester group, 31P-MR spectroscopy showed a chemical shift of 56.07 ppm, which enable to acquire 31P images from both phantoms during one CSI measurement by frequency selection (Figure 1 -I). SNR calculated from 31P spectra results in for pTMPC probe and 12.5-163.7 for pMPC. Imaging analyses results in SNR of 6.3-13.6 (ST=15 min-3 h) and 3.1-5.4, respectively. Measurement of the pTMPC and rMSC solution resulted in two clearly distinguishable signals (Figure 1 -II). The difference in chemical shift between these the two peaks (Δδ=54.5 ppm) was similar to what we observed for phantoms. Imaging of both cells and polymers was possible to perform within short acquisition time (ST=2 min). Cytotoxicity testing confirms that pTMPC is not influencing cells viability. High phosphorus concentration and SNR within a short scan time is beneficial for further application in various in vivo animal models. Conclusion Preliminary results of a novel phosphorus-containing pTMPC contrast agent based on 31P-MR chemical shift proved high sensitivity of the probe. Large chemical shift from biological 31P signal, resulting in clearly separated two 31P signals, was observed in spectroscopy and imaging. This outcome, together with great biocompatibility, indicate that pTMPC could serve as an efficient phosphorus in vivo contrast agent. Acknowledgements The study was supported by the Charles University Grant Agency GA UK (project number 358119); Ministry of Health of the Czech Republic (project number NU20-08-00095); Institute for Clinical and Experimental Medicine IKEM, IN00023001.

References: [1] Abstract Body : Breast cancer is a highly heterogenous disease, both phenotypically and genetically. Of critical importance, a correlation between intra-tumoral heterogeneity, drug resistance and negative clinical outcome has been established. Previously, quantity or subcellular location of protein biomarkers have been compared to identify tumor tissue types. To-date, mass cytometry and single-cell sequencing studies have been used to assess spatial intra-tumor heterogeneity by identifying a range of phenotypic and genetic alterations across different regions of a single tumor. Recently, cell morphology has been indicated as a direct readout of the functional phenotypic state of an individual cancer cell. Here, the hypothesis was proposed that the spatial context of organelles within cancer cells, specifically their subcellular location and inter-organelle relationships (topology), can be used to inform breast cancer cell classification. Since numerous correlations between biologic behaviors and pathologic findings have been wellestablished, organelle topological heterogeneity reveals the long-term adaptation of organelles and cytoskeletal networks to match breast cancer cell type status. Thus, Organelle Topologybased Cell Classification Pipeline (OTCCP) was introduced as a novel approach that quantifies, for the first time, the topological features of subcellular organelles, removing the bias of visual interpretation, to classify different breast cancer cell lines using a machine learning method. This study aims to investigate the heterogeneity of organelle topology and morphology in breast cancer cells to increase our basic understanding of breast cancer biology on a subcellular level. OTCCP is a novel approach that quantifies, for the first time, the feature of organelles in breast cancer cells, removing the bias of visual interpretation to classify different cell lines based on organelle morphology and topology. This method was tested on three different organelle datasets: mitochondria, early endosomes and recycling endosomes in a panel of human breast cancer cells, including T47D (estrogen receptor-positive), MDAMB231, MDAMB436 and MDAMB468 (triple negative) and AU565 (HER2 positive), as well as non-cancerous mammary epithelial MCF10A cells. High resolution Airyscan microscopy was used to collect z-stacks across cells labeled with fluorescently labeled-transferrin uptake (ERC) and immunostained with anti-Tom20 (Mito) and anti-EEA1 (EEC) to label the recycling and early endosomal and mitochondria networks, respectively. Subsequently, 3D surface rendering of organelle objects was performed using IMARIs image analysis software. A morphometric evaluation of mitochondrial and endosomal compartments resulted in 34 topology and morphology parameters. Application of Random Forest machine learning based classification to 18 of these 34 parameters generated the highest accuracy in breast cancer cell classification using MATLAB software. We systematically evaluated how different parameter combinations affected the machine learning-based cancer cell classification and discovered that topology parameters were crucial to achieve the highest classification accuracy. Based on organelle spatial distribution and their interaction with neighbor organelles (topology) a classification accuracy over 95% was achieved to distinguish between a variety of human breast cancer cell lines of differing subtype and aggressiveness. Topology parameters generated the highest accuracy (97.1%) in breast cancer classification, indicating that organelle topology is a novel, highly precise classifier to differentiate cell lines of differing subtype and aggressiveness. These findings lay the groundwork for using quantitative topological organelle features as an effective method to analyze and classify breast cancer cell phenotypes.

References: Wang, L., Ward, J., Bouyea, M., and Barroso, M., 2020 Abstract Body : Background and Motivation: High throughput capabilities can be useful in preclinical imaging to promote larger sample sizes while limiting experimental durations and costs. High throughput imaging facilitates the efficient use of experimental radiotracers, which tend to be available in small yet expensive batches and often use radionuclides with short halflife. Although multi animal imaging is a frequent practice, animal welfare and consistent imaging conditions (e.g. stable body temperature in [18F] FDG imaging, Ref. [1] ) can be overlooked where multi-animal imaging platforms lack the ability to monitor physiological signals, stabilize temperature and adjust the concentration of anesthetic gas for individual subjects. A system for high throughput total body dynamic imaging in mice is presented here. It comprises a PET/CT system and a new 3-mouse cradle, enabling monitoring of physiological signals and dedicated anesthetic gas supply. This set up facilitates stable imaging conditions for consistent PET imaging, even with prolonged acquisition times. Methods: A new animal cradle with capacity for 3 mice was used. The cradle enables individualized monitoring of body temperature and respiration in each subject (figure 1). The cradle also includes animal warming using warm air, anaesthesia gas exhaust and the option to separately adjust anaesthesia for each animal. Two animals are located on the lower deck and the third is positioned above these on a removable tray for easier access and 2-mouse imaging. This animal cradle features an interface for quick connection of all lines on multiple imaging platforms (PET/CT, PET/SPECT/CT and PET/MR instruments). The animal cradle was tested with a PET/CT system with 150 mm FOV for total body imaging in a single cradle position. The CT performance of the system was assessed in Ref. [2] and the PET NEMA characterization can be found here [3] . The cradle and scanner combination was tested in a multi mouse [18F] FDG dynamic PET experiment using 3 healthy mice. Each mouse was injected via the tail vein with 10 MBq of [18F] FDG at the start of acquisition. 1% isoflurane was used for maintenance of anaesthesia and the scan lasted for 2 hrs. The LM file was rebinned after the acquisition using this scheme: 1x20 s, 8x5 s, 8x10 s, 8x60 s, 8x300 s and 7x600 s. The image was reconstructed with MLEM using 12 iterations and a 0.5 mm voxel size. PMOD was used for organ segmentation and statistical analysis. Results: The resulting images show the typical murine biodistribution of [18F] FDG. despite the prolonged acquisition, no signs of motion or brown adipose tissue uptake were apparent (figure 2). This indicated stable animal conditions and proper temperature control. Regions of interest were drawn in the vena cava to derive a surrogate of the blood tracer concentration as well as in the myocardium. The chart provided in the Persuasive Data section shows the time activity curves for each animal in each region of interest. Conclusion: A system comprising the Si78 PET/CT and a 3-mouse high throughput imaging cradle was tested in a dynamic [18F] FDG PET experiment. Physiological monitoring and controlled temperature in all positions of the 3-mouse bed allowed stable conditions to be maintained for as long as 2 h during the measurement. The system will facilitate high throughput imaging while maintaining gold-standard imaging conditions. This will particularly benefit core facilities, pharma, and centers with high workloads. Figure 2 . Dynamic frames at different time points of the [18F] FDG acquisition. Initial frame shows injection moment Abstract Body : Objectives: Microglia, the resident innate immune effector cells of the central nervous system (CNS), carry out a wide range of functions from passively surveilling the brain to contributing to the post-injury healing process, and play a critical role as first responders to brain injury. Despite the contribution of prolonged microglial activation and aberrant function to deleterious neuroinflammation seen in chronic neurological diseases such as Alzheimer's Disease (AD) and neuropathic pain, there are no readily available methods for specifically tracking these cells and their function in vivo. Positron Emission Tomography (PET) is a highly sensitive functional imaging modality well-suited to non-invasive, longitudinal visualization of these molecular processes. A recent gene profiling study identified P2Y12R as an important gene strongly associated with homeostatic microglia1. P2Y12R protein expression in the CNS is exclusively on microglia at constant levels in healthy adults, but quickly and significantly declines with the onset of a pro-inflammatory state (e.g, surrounding dense, mature Ab plaques in AD)2,4,5. While P2Y12R is an extremely promising functional PET biomarker of microglial function, the only published tracer is not suitable for in vivo use4. Here, we seek to develop the first P2Y12R PET radioligand amenable to in vivo imaging. Methods: We identified P2Y12R antagonist AZD1283 as a promising CNS tracer candidate based on its high affinity (KD = 11nM) and favorable lipophilicity (LogD=2.8)3. [11C] AZD1283 was labeling using the radioactive precursor [11C] NH4CN, which was obtained from cyclotron produced [11C] CO2 using the automatic module GE ProCab (Fig. 1A) . [11C] NH4CN was delivered to a TRACERlab FX-FN module and trapped in a solution of 0.5M aqueous KOH (2mL) and K2.2.2 (3mg in 1mL tetrahydrofuran) in a glass reactor. A solution of 1mg synthetic precursor and 2mg palladium tetrakis-triphenylphosphine in 500mL THF was added to the reactor, and the resulting mixture heated and stirred at 90oC for 5 minutes. This crude reaction mixture was purified via semipreparative HPLC to afford pure [11C] AZD1283. Final product was reformulated via pluslite C18 cartridge to yield [11C] AZD1283 in saline/EtOH, and analyzed via analytical HPLC at 0 and 90 minutes after reformulation to confirm both tracer identity and stability in formulation. Stability of [11C] AZD1283 was assessed in human plasma via HPLC at 0, 5, 15, 30, 60, and 90 minutes. In vitro incubation with mouse and human liver microsomes and a Caco-2 monolayer assay were performed to assess metabolic stability and passive permeability, respectively. Finally, key physiochemical parameters (e.g. cLogP, pKa, total polar surface area) were calculated and a published Multi-Parameter Optimization (MPO) scoring tool used to predict potential for CNS uptake6. Results: AZD1283 was radiolabeled for the first time with carbon-11 to afford [11C] AZD1283 in a yield of 9.17 mCi with high chemical and radiochemicaly purity (>99%) and molar activity of 395 mCi/mmol(average of n=3). In vitro stability of [11C] AZD1283 was found to be high with >99% intact tracer present after 90 minutes incubation in formulation and human serum as demonstrated by HPLC chromatograms (Fig. 1B -D) . Half-life in mouse and human liver microsomes was found to be 37 and >160 minutes, respectively, and Caco-2 assay predicts passive CNS uptake with a small amount of active efflux. AZD1283 received an MPO score of 4.2, indicating over-all alignment with known CNS penetrant druglike molecules, and high potential for in vivo CNS uptake. Conclusions: Herein we describe the successful synthesis of [11C] AZD1283, and demonstrate its high stability in both formulation and human plasma. Our new P2Y12R tracer can be reliably synthesized in ample yield and purity for in vivo imaging. Future PET imaging and metabolism studies in mice and non-human primates will further elucidate the in vivo behavior and translational potential of this very promising compound.

References: 1. Butovsky et al. Nat. Neurosci, 2014 . 17:131-143. 2. Haynes et al. Nat Neurosci. 2006 , 19(12): 1512 -9. 3. Bach et al. J. Med. Chem, 2013 , 56(17): 7015-7024 4. Villa et al. Theranostics, 2018 : 5400-18. 5. Walker et al. Int J Mol Sci. 2020 , 21:687. 6. Wager et al. ACS Chem. Neurosci. 2010 . [11C] cyanation of synthetic bromoprecursor was carried out in THF at 90 oC for 5 minutes to yield [11C] AZD1283 (A). Identity and radiochemical purity of [11C] AZD(1B, green) was confirmed via co-injection with cold reference standard(1b, blue) onto analytical HPLC (B). [11C] AZD1283 is stable in formulation at 90 minutes post end of synthesis(C). [11C] AZD1283 was incubated in human plasma, with aliquots quenched and analyzed at 0 (1D,red), 5 (1D,blue), 15 (1D,green), 30 (1D,pink), and 90 (1D, brown) minutes to confirm stability in human plasma for 90 minutes (D) . Abstract Body : Objectives: Predicting central nervous system (CNS) uptake and binding specificity of positron emission tomography (PET) tracers prior to in vivo testing remains a significant challenge. For this reason, developing and translating a single CNS PET tracer for clinical use is an extremely resource intensive endeavor, often requiring synthesis and evaluation of numerous candidate molecules without a reliable way to predict likelihood of success in advance. While in silico methods exist to augment the therapeutic drug development process by efficiently screening candidates for potential pitfalls early in development, there is a paucity of such tools tailored to PET tracer design and translation1. In this work, we address this unmet need by generating a unique three-pronged in silico approach to rapidly predict the potential value of candidate CNS PET tracers for clinical use by 1) building a large database of successful and unsuccessful CNS PET tracers following an in-depth meta-analysis of published work, 2) leveraging this database to develop complementary in silico methods to determine multiparameter scoring functions for quantifying likelihood of tracer success and 3) utilizing machine learning (ML) approaches to differentiate successful and unsuccessful tracers. Methods: Following an extensive literature analysis, we collated a diverse library of 75 successful CNS tracers (defined as approved for clinical research) and 65 unsuccessful tracer candidates (most commonly due to insufficient blood-brain barrier permeability or prohibitively high non-specific binding (NSB)). Eight key physiochemical parameters were computed for all tracers: clogP, clogD, molecular weight, topological polar surface area (tPSA), pKa, number of hydrogen bond donors (HBD) of the neutral species, HBD at physiological pH, and net charge at pH 7.4. Tracers were categorized as 'successful' or 'unsuccessful', with the latter further subdivided (Fig 1a) . Based on published methods, physiochemical data was used to create a multi-parametric scoring function for quantifying a tracer's likelihood of success in vivo, with the aim of distinguishing between successful and unsuccessful tracers2. In parallel, a random subset of the physiochemical and classification data (70%) was used to train several ML models that were then evaluated by classifying the remaining tracers (30%). Results: Evaluation of 8 physicochemical parameters showed distinct trends, indicating significant potential for in silico differentiation of successful vs. non-CNS penetrant tracers (Fig 1b) . These distinctions were effectively captured by our novel scoring function, which distinguished successful from non-permeable molecules with 96.7% specificity, 49.3% sensitivity, and positive (PPV) and negative (NPV) predictive values of 97.4 and 43.3, respectively. In particular, the PPV reflects the function's high potential utility for identifying nonpermeable molecules early in development. Conversely, overall physicochemical property profiles of high NSB tracers were strikingly similar to successful tracers, discouraging use of an analogous function for distinguishing these two populations of tracers (Fig 1c) . In comparison, application of a Support Vector Machine learning model showed a 16% increased probability of correct classification for both positive and non-permeable categories (79% vs. 63%), despite access to only a portion of the data (Fig. 1d-e) . Conclusion: Analysis of physicochemical properties of 140 successful and unsuccessful CNS PET radiotracers revealed key differences that we believe significantly impact CNS uptake in vivo. Our proof-of-concept method based on both a manually constructed quantitative function and ML models show promise for ab initio prediction of CNS tracer viability prior to resource intensive in vivo imaging studies. Future work will include expansion of the tracer library in addition to optimization and comparison of our scoring function with ML-derived models. Integration of these complimentary approaches will enable design of a high throughput, easy to use computational tool for CNS PET tracer design, ultimately facilitating translation of high impact tracers for clinical use in a time-and cost-effective manner.

References: 1. Zhang et al. J. Med. Chem. 2013 , 56, 4568-4579. 2. Wager et al. ACS Chem. Neurosci. 2010 Image/ Figure: https://s3.amazonaws.com/amz.xcdsystem.com/1E8A4501-BB41-A02A-1D3B9882B798BFAA_abstract_File10351/GA85_ImageFigure_0513083312.pdf Image/Figure Caption: Figure 1 . The radiotracer library we used to generate this data, depicted in (A), is comprised of successful (n = 75) and unsuccessful (n = 65) radiotracers, with the latter further subdivided into tracers without CNS uptake(n=30), tracers with prohibitively high NSB (n = 25), and tracers that failed for other reasons (n = 10). Head-to-head comparison of the average physiochemical property profile of successful CNS tracers vs. non-permeable PET tracers (B), and tracers with high NSB (C). A Support Vector Machine learning model identified distinct attributes between successful vs. non-permeable negatives allowing for high probability of correct classification on a test set when compared with manual assignment (79% correct assignment) (D) . Just as manual scoring failed to distinguish between positives and high nonspecific binders, the same model also found little distinction and classified all tracers as positive (E). Abstract Body : Background: Melanoma is the most aggressive skin cancer with the highest associated mortality [1] . The 5-year survival rates range from 98.3% (stage I) to 16% (stage IV) emphasizing the importance of diagnosis at an early stage [2] . Non-invasive morphological imaging approaches such as dermoscopy and high-resolution, morphological reflectance confocal microscopy demonstrate high sensitivity but moderate and variable specificity [3] . Noninvasive gene assays based on PRAME (PReferentially expressed Antigen in MElanoma) expression show similarly high sensitivity but moderate specificity [4] ; neither of these approaches have shown potential in stratifying lesions based on malignant potential. Noninvasive molecular imaging using targeted fluorophores such as PARPi-FL (targets PARP1, overexpressed in cancers [5] ), can enable functional evaluation of PARP1 expression to stratify lesions which are malignant and/or have a higher propensity for malignant transformation [6] , thereby improving diagnostic specificity for melanoma detection and prognosis. As a first step towards PARPi-FL imaging in vivo, we investigated PARP1 expression as a biomarker to differentiate melanoma, atypical melanocytic proliferation and benign nevi. Methods: We investigated PARP1 and PRAME expression using immunohistochemistry (IHC) in a spectrum of 132 melanocytic lesions (73-melanoma, 51-benign nevi, 8-atypical intraepidermal melanocytic proliferation) in 5 μm-thin formalin-fixed paraffin-embedded serial sections. IHC was performed on an automated Leica-Bond stainer platform. Both IHC and H&E slides were digitized. Automated quantification for PARP1 area positivity and intensity was performed using Aperio Positive Pixel Algorithm (Leica Biosystems, Germany) on whole slides annotated for tumor, benign and epidermis (used as control) by a pathologist using histology and PRAME expression as gold standard, The quantification was performed using fixed optimized parameters for hue value, hue width and color saturation threshold. Strong positive pixels were selected for analysis which was performed for total area positivity and normalized intensity. Descriptive statistics was performed for three groups-atypical intraepidermal melanocytic proliferation (AIMP), benign nevi and melanoma. ROC analysis was performed only on nevi versus melanoma. All statistics were performed on GraphPad Prism version 8.0. Results: PARP1 expression was found in all benign and melanoma tissues at different levels. Area positivity (%) was highest in melanoma (31±13), followed by AIMP (25±17) and benign nevi (13±16). Oneway ANOVA demonstrated statistically significant differences between benign and melanoma groups (p Conclusions: Our results on a large dataset demonstrate statistically higher PARP1 expression in melanoma as compared to benign melanocytic lesions. Therefore, PARPi-FL will potentially serve as a robust biomarker for identification of malignant and potentially malignant lesions, which will be tested subsequently in excised tissues. A multimodal reflectance and fluorescence confocal microscopy approach will also be tested to further improve the diagnostic accuracy. Given the IND status of PARPi-FL for use in humans, the next set of results on PARPi-FL staining will establish the basis for subsequent in vivo clinical imaging studies for molecular melanoma diagnosis in patients. Acknowledgements: Supported by Melanoma Research Foundation (Dermatology Fellow Award) and National Institutes of Health grants P30 CA008748-50 and R01EB020029. 

Image/Figure Caption: Higher PARP1-expression was found in melanoma and AIMP as compared to benign nevi suggesting PARP1 expression estimation using PARPi-FL can help identify malignant and potentially malignant lesions A. IHC quantification of area positivity and normalized intensity was higher for melanoma (p These results suggest the importance of PARP1 expression and potentially PARPi-FL staining in improving diagnosis of melanoma in vivo in patients. year overall survival rate of 10%. Treatment options with the greatest overall survival include chemotherapeutics such as gemcitabine or combination regimens such as FOLFIRINOX. Treating pancreatic cancers with gemcitabine or other cytostatic drugs has been known to induce senescence, a type of cell arrest, and to make the cancer refractory to additional therapy. Senescent tissue secretes numerous inflammatory markers, inducing increases in vascular access as part of the senescence associated secretory phenotype (SASP), and is considered tumorigenic. Here, we used clinically available antibodies such as Bevacizumab (VEGF-A), Crizanlizumab (P-selectin), Siltuximab (IL-6), and ADAM17 to target some of these secreted markers in order to identify senescent tissue. Current methods for determining senescence are limited to lysosomal activity of beta-galactosidase, which is not unique to senescent tissue and utilizes a small molecule agent that is widely distributed and rapidly cleared, lowering specificity. These SASP-targeting antibodies were radiolabeled using the established 89Zr-Desferoxamine isotope-chelator, with purities greater than 99%, to produce precision ImmunoPET versions of these clinically tested antibodies. We were able to induce SASP with daily administration of trametinib and palbociclib in both orthotopic and subcutaneous models of pancreatic cancer with the MiaPaCa2 cell line. By imaging four different components of SASP in mice undergoing senescence therapy, a greater visualization of chemotherapy induced senescence can achieved with immunoPET. In Figure 1 , variation in tumor uptake was seen, showing not only differences in antibody tumor penetrance, but also injected dose per gram for the four antibodies. By quantifying tumor uptake of each SASP factor, the presence of senescent and resistant pancreatic cancer can be detected earlier, providing time for next-line therapy, and possibly a new therapeutic stratagem. Figure 1 : ImmunoPET of MiaPaCa2 subcutaneous tumors under TP therapy targeting markers of SASP. Tumor heterogeneity between markers identify SASP factor location and intensity (optimization needed).

Renogram curve pattern of bilateral PUJO patient. Obstructive pattern is observed in the right kidney curve, while non-functioning left kidney showed a minimal radioactivity uptake (flat curve). Abstract Body : In the era of personalized medicine in oncology, progress regarding the modulation of tumor microenvironment (TME) and its follow-up by molecular imaging is of crucial importance because of its central role in the development of cancer. Among immunosuppressive cells of the TME, M2-like macrophages have been associated with cancer aggressiveness, therapy resistance and poor prognosis. The aim of this study was to assess if M2like macrophages (i) could be tracked in vivo into the TME with SPECT imaging and (ii) could be modulated by the inhibition of Gp96, an endoplasmic reticulum chaperone involved in inflammatory processes that we have previously shown to be expressed at the cell membrane of human primary M2 macrophages (Chaumonnot et al, 2021) . Tissues from mouse triple negative breast cancer (TNBC, 4T1 cell line) and colon cancer (CC, CT26 cell line) were analyzed by immunohistochemistry (IHC) to detect the presence of Gp96 and CD206 (a marker of M2 macrophages) into the TME. Specific CD206 in vivo imaging on 4T1-and CT26-tumor bearing mice receiving or not a specific inhibitor of Gp96 (PU-WS13) was performed with 99mTc-Tilmanocept SPECT (i.v injection, 15MBq/mouse). Images were performed at 1h, 4h and 24h post-injection. Ex vivo gamma counting of tumors was performed after the last imaging. Radioactivity content measured with gamma counting or on images was expressed as percentage of the injected dose per gram of tissue (% ID/g). For 4T1 tumors, tumor growth and collagen content were also assessed. IHC experiments demonstrated an overexpression of Gp96 in tumor cells as well as the presence of M2-like macrophages expressing both CD206 and Gp96 in 4T1 tumors while CT26 tumors only showed an upregulation of Gp96. These results are confirmed with in vivo SPECT imaging. 99mTc-Tilmanocept tumor uptake was significantly higher in 4T1compared to CT26-tumor bearing mice (2.53 %ID/g and 1.08 %ID/g respectively, p=0.009). Based on these results, the 4T1 model was chosen to study the impact of Gp96 inhibition. Interestingly, PU-WS13 induced a significant decrease in 99mTc-Tilmanocept tumor uptake compared to untreated mice (1.12 %ID/g and 0.78 %ID/g respectively, p=0.0011) as well as a lower number of CD206+ M2-like macrophages compared to untreated mice. These results correlated with a reduced tumor growth and collagen content in 4T1 tumors. 99mTc-Tilmanocept SPECT imaging might represent an innovative non-invasive strategy to quantify CD206+ tumor-associated macrophages as a biomarker relevant for prognosis, therapeutic prediction and/or monitoring of solid tumors. The potential effects of PU-WS13 on the modulation of M2-like macrophages are currently investigated. Abstract Body : Introduction: There has been a lack of real-time intravital quantitative research on cardiovascular adaptive changes of astronauts in weightlessness environment. In this study, the dynamic 18F-FDG PET/CT was used to explore the myocardial metabolic adaptive changes in hindlimb unloading (HU) model of weightlessness. Methods: Twelve healthy Balb/c mice aged 8 weeks were randomly divided into the control group (Ctrl, n = 6) and the HU group (HU, n = 6). The preparation of the simulated weightless model was mainly based on the classic tailsuspension method proposed and improved by Morey-Holton and Chen Jie et al. After 1/2/3/4 weeks of modeling, dynamic 18F-FDG PET/CT myocardial metabolic imaging was performed for 60 min including 31 frames (10×3s; 3×10s; 4×60s; 6×150s; 8 ×300s). The dynamic raw data was analyzed by Carimas software to obtain polar bull 's-eye map of myocardial metabolism (The American Heart Association 17-segment method). With the time activity curve (TAC) of left ventricular as the input function, the original TAC of 17 myocardial segments was fitted by Patlak model to obtain 18F-FDG net uptake rate constant Ki of each segment. Unpaired t test was used to compare Ki of each segment between the Ctrl group and the HU group, and p Results: After 1 week of HU modeling, the Ki of the seg 5 (0.01108 ± 0.00530 vs 0.01750 ± 0.00395, p and seg 6 (0.00985 ± 0.00323 vs 0.01564 ± 0.00221, p ± 0.00515 vs 0.02831 ± 0.00706, p ± 0.00390 vs 0.02230 ± 0.00766, p Conclusions: After short-term simulated weightlessness (suspension 1-2 weeks), Ki of the partial myocardial segments (seg 5, 6, 11, 16) supplied by left circumflex artery (LCX) was significantly reduced, indicating the decrease of glucose metabolism in the myocardial segment supplied by LCX was an important adaptive change. After mid-term simulated weightlessness (suspension 3-4 weeks), Ki of the most myocardial segments was significantly reduced, suggesting the heart was in a low metabolic state. The result of this study for guiding the astronauts to adapt to the weightlessness was that seg 5 and 6 were the myocardial regions with the most significant changes in glucose metabolism during the early weightless simulation, while glucose metabolism of the whole heart reduced during the later weightless adaptation. Figure: 

Pim Van Den Hoven, Leiden University Medical Center, p.van_den_hoven@lumc.nl

Abstract Body : Introduction: In assessing the severity of lower extremity arterial disease (LEAD), physicians often rely on clinical judgement, thereby supported by conventional measurements of macrovascular blood flow including the ankle-brachial index or toe pressure measurement. However, these techniques provide no information about regional tissue perfusion and have shown to be of limited value in patients with chronic limb-threatening ischemia (CLTI) (1) . Near-infrared (NIR) fluorescence imaging using indocyanine green (ICG) has been used extensively in perfusion studies and has shown valuable results in prediction of tissue viability in patients with LEAD. However, with regards to diagnosing LEAD, inconsistency is seen between different Fontaine stages and it is unclear whether advanced stages of LEAD alter the in-and outflow of ICG (2) . Furthermore, there is limited information about the perfusion patterns of ICG NIR fluorescence imaging in control patients. In the search for providing valid and reliable cutoff values for tissue perfusion, it is an essential step to define these perfusion patterns of ICG NIR fluorescence imaging in controls too. Methods: This prospective cohort study was registered at the Dutch Trial Register with number NL7531 and included patients with CLTI and control patients. Patients were excluded based on contra-indications for ICG. The Quest Spectrum Platform® was used to perform ICG NIR fluorescence imaging. Upon intravenous administration of ICG, the NIR fluorescence intensity in both feet was recorded for 10 minutes. The NIR fluorescence videos were analyzed using the Quest Research Framework®. The dorsum of the foot was selected as region of interest and subsequently a time-intensity curve was created. An example of the measured NIR fluorescence intensity with the corresponding time-intensity curve is shown in Figure 1 . Fourteen parameters were extracted from the time-intensity curves. Results: Successful ICG NIR fluorescence imaging was performed in 35 patients with a total of 60 limbs included for analyses. Nineteen patients presented with LEAD of whom 28 limbs were classified as CLTI. The control group consisted of 16 patients with a total of 32 limbs. The mean maximum intensity was significantly lower in the control group (37.9 vs 25.8, p=0 .000). Furthermore, the time to maximum intensity (i.e. Tmax) was reached earlier in the CLTI group (90.5 vs 143.3, p=0.002). When looking more closely at inflow parameters, the maximum slope, the normalized maximum slope and the ingress rate were all significantly higher in the CLTI group (2.0 vs 0.6, p=0.000; 4.2 vs 2.4, p=0.000; 1.0 vs 0.2, p=0.000). For outflow parameters, a significant difference was seen for the maximum slope egress, which was higher in the control group (0.5 vs 0.2, p=0.005). No significant difference was observed for the normalized maximum slope egress (1.0 vs 0.8, p=0.733) . Normalization of the time-intensity curves displayed a more narrow distribution for the time-intensity curves in both groups. Conclusion: This study demonstrates the different perfusion patterns as seen on ICG NIR fluorescence imaging between patients with CLTI and control patients. Interestingly, most inflow parameters observed in patients with CLTI were superior to the control group. No significant differences were observed regarding outflow. Possible explanations for the increased inflow include damage to the regulatory mechanisms of the microcirculation, arterial stiffness and transcapillary leakage. In the search for providing cut-off values for adequate perfusion, more research is needed on in-and outflow patterns for control patients and various stages of LEAD. Abstract Body : Introduction: Intraoperative tumour identification (e.g., tumour extension, margins or presence of metastases) via receptor specific targeting is one of the ultimate promises of fluorescence-guided surgery. The translation of receptor-specific fluorescence tracers into clinical trials forms a critical component in maturing this treatment concept. This step is often hindered by differences in physiology and dosing in the preclinical and clinical setting, as well as differences between the imaging equipment used in both situations. To close the gap between preclinical studies and first-in-human trials the potential of topical tracer application on surgical specimens was evaluated, in order to assess both tracer and (clinical grade) camera performance. Methods: Ex vivo tracer performance was assessed via incubation of freshly excised penile squamous cell carcinoma (pSCC; N=10) and oral squamous cell carcinoma (OSCC; N=10) specimens in a solution containing the c-Met targeting tracer EMI-137 (500 nM; (1)). Three prostate specimens were incubated with the PSMA-targeting tracer EuK-(SO3)Cy5-mas3 (500 nM; (2)). Both tracers contained Cy5 as fluorescent label. All specimens were analysed using a Cy5-compatible prototype clinical grade laparoscopic camera system. In-house developed image processing software allowed video-rate/real-time tumour identification and assessment of the tumour-to-background ratio (TBR). Fluorescence imaging results were related to standard pathological tumour evaluation and immunohistochemistry. The in vivo potential of c-MET targeting, and intraoperative tumour visualization was confirmed following intravenous administration of EMI-137 in five pSCC patients using the same camera system as used for ex vivo measurements. Results: After incubation with EMI-137 or EuK-(SO3)Cy5-mas3 91.3% of tumours could be fluorescently illuminated and detected (9/10 OSCC, 10/10 pSCC and 2/3 prostate tumours). Immunohistochemistry revealed overlap between the fluorescence staining and c-Met or PSMA receptor expression in all illuminated specimens. Non-visualization (2/23, 8 .7%) could be linked to lesions that resided deeper below the resection surface (OSCC; limited tissue penetration) or highly diffuse, low grade tumour with very low density of PSMA positive cells (limited local signal). Tumour margin assessment was improved using video-rate representation of the TBR (median TBR: 2.5 +/-0.2; range 1. 9-4.2) . This visualization technique also allowed recognition of heterogeneity in receptor expression within the lesion. With regard to EMI-137, studies on surgical specimens successfully translated to in vivo fluorescence-guided surgery, confirming the potential of the topical evaluation steps. Conclusion: Ex vivo evaluation on fresh tumour specimens helps 1) evaluate the ability of fluorescent tracers to target tumour specific receptors in human tissue and 2) highlights the possible limitations for in vivo use (e.g., tissue penetration, low/diffuse receptor expression). Initial data obtained with EMI-137 indicate that successful ex vivo staining could be predictive for successful in vivo use. Abstract Body : Introduction: In recent years imaging of the fibroblast activate protein has emerged as a high potential extension of the tracer arsenal within nuclear medicine for visualization of tumor tissue. Based on the success stories of FAPI-PET one may argue that the same target may also help advance the field of image-guided surgery. Therefore, a family of fluorescent FAPI-tracers was created based on the known FAP targeting vector Gly-Pro (GP; (1)) and its variant ERGTGP. As a first step towards assessment of their suitability as FAPtargeting tracer in fluorescence-based image-guided surgery in vitro performance was assessed, wherein focus was placed on their sub-cellular accumulation. Methods: Four different GPfluorescent FAPI tracers were synthesized by functionalizing GP or ERGTGP with QAmine(SO3)Cy5, resulting in QAmine-(SO3)Cy5-GP, QAmine-(SO3)Cy5-ERGETGP, and QAmine-(SO3)Cy5-CPGTEGRE. MCF7 (known FAP positive human breast cancer cells; (1)) cells and three control cell lines (MDAMB231 (human breast carcinoma), RT4D12 (rat schwannoma), GeB3 (dog epithelial)) were incubated with 1uM of tracer, either alone or in competition with the FAP inhibitor Ac-Gly-BoroPro. Incubation with non-functionalized dye variants and blocking with a mitochondrial tracer (1nM) were performed as control. Fluorescence confocal microscopy was used to assess the cellular distribution of the tracer signal (Cy5). Nuclear, lysosomal and mitochondrial counter stains were used as reference for intracellular localization. Immunohistochemical staining with an anti-FAP antibody was applied to confirm the presence of FAP within the different cell cultures. Results: QAmine-(SO3)Cy5-GP, QAmine-(SO3)Cy5-ERGETGP and QAmine-(SO3)Cy5-CPGTEGRE showed intracellular tracer uptake in all cell lines, with a staining pattern similar to literature reports (1) . Nonfunctionalized dyes did not show cellular uptake. Neither incubation time (1 or 4 hours) nor temperature (4 or 37 degrees) influenced the staining of QAmine-(SO3)Cy5-GP, QAmine-(SO3)Cy5-ERGETGP or QAmine-(SO3)Cy5-CPGTEGRE. Interestingly, blocking experiments with the FAP inhibitor Ac-Gly-BoroPro (1000-fold excess) revealed a 62% decrease in fluorescence in RT4D12 cells, while uptake in MCF-7 and Geb3 cells was unaffected. Blocking with a mitochondrial stain affected uptake of the FAPI tracers in MCF-7 (60% signal decrease) but not Geb3 cells while lysosomal and mitochondrial counterstains were completely diminished. In contradiction to fluorescence imaging, FAP-specific immunohistochemistry showed intracellular presence of FAP in MCF-7, but not in RT4D12, MDAMB231 and GEB3 cells. Conclusion: The fluorescent FAPI-tracers were able to stain a wide range of cells, but the observed staining patterns did not seem to correlate with FAP IHC. At this point this discrepancy is unexplained, but the bulky nature of the fluorescent dyes could play a role. Abstract Body : Introduction: Patients with chronic limb-threatening ischemia have a risk of undergoing a major amputation within 1 year of nearly 30% with a substantial risk of reamputation since wound healing is often impaired (1) . Quantitative assessment of regional tissue viability following amputation surgery can identify patients at risk for impaired wound healing. For the assessment of regional perfusion, near-infrared (NIR) fluorescence imaging using Indocyanine Green (ICG) has already shown promising results in reconstructive surgery (2) . Prospective studies on the predictive value of ICG NIR fluorescence imaging in healing of amputation wounds have yet to be performed. Methods: This prospective cohort study included adult patients undergoing lower extremity amputation surgery due to peripheral artery disease or diabetes mellitus. ICG NIR fluorescence imaging was performed within 5 days following amputation surgery using the Quest Spectrum Platform®. Following intravenous administration of ICG, the NIR fluorescence intensity of the amputation wound was recorded for 10 minutes. The NIR fluorescence intensity videos were analyzed and if a fluorescence deficit was observed, this region was marked as "low fluorescence". All other regions were marked as "normal fluorescence". Results: Successful ICG NIR fluorescence imaging was performed in 10 patients undergoing a total of 15 amputations. No "low fluorescence" regions were observed in 11 out of 15 amputation wounds. In 10 out of these 11 amputations, no wound healing problems occurred during follow-up. Regions with "low fluorescence" were observed in 4 amputation wounds. Impaired wound healing corresponding to these regions was observed in all wounds and a reamputation was necessary in 3 out of 4. An example of a patient where a "low fluorescence" region was observed is depicted in Figure 1 . When observing time-related parameters, regions with low fluorescence had a significantly longer time to maximum intensity (113 seconds versus 32 seconds, p=0.003) and a significantly lesser decline in outflow after five minutes (80.3% versus 57.0%, p=0.003). Conclusion: ICG NIR fluorescence imaging was able to predict postoperative skin necrosis in all four cases. Quantitative assessment of regional perfusion remains challenging due to influencing factors on the NIR fluorescence intensity signal, including camera angle, camera distance and ICG dosage. This was also observed in this study, contributing to a large variety in fluorescence intensity parameters among patients. To provide surgeons with reliable NIR fluorescence cut-off values for prediction of wound healing, prospective studies on the intra-operative use of this technique are required. The potential prediction of wound healing using ICG NIR fluorescence imaging will have a huge impact on patient mortality, morbidity as well as the burden of amputation surgery on health care. (6) Abstract Body : Introduction: Current diagnostic modalities for patients with peripheral artery disease (PAD) mainly focus on the macrovascular level and are unable to quantify regional tissue perfusion (1) . For quantitative assessment of tissue perfusion, near-infrared (NIR) fluorescence imaging using indocyanine green (ICG) seems promising. In this prospective cohort study, ICG NIR fluorescence imaging was performed pre-and post-revascularization to assess changes in foot perfusion. Methods: ICG NIR fluorescence imaging was performed in patients with PAD pre-and postintervention. After intravenous bolus injection of 0.1mg/kg ICG, the camera registered the NIR fluorescence intensity over time on the dorsum of the feet for 15 minutes using the Quest Spectrum Platform®. Time-intensity curves were plotted for 3 regions of interest (ROIs): 1. The dorsum of the foot, 2. The forefoot and 3. The hallux. Time-intensity curves were normalized for maximum fluorescence intensity. Extracted parameters were the maximum slope, area under the curve (AUC) for the ingress and the AUC for the egress. The non-treated contralateral leg was used as a control group. Results: Successful revascularization was performed in 32 patients. There was a significant increase for the maximum slope and AUC egress in all three ROIs. The most significant difference was seen for the maximum slope in ROI 3 (3.7 %/s to 6.6 %/s, p Conclusions: This study shows the potential of ICG NIR fluorescence imaging in assessing changes in foot perfusion following successful revascularization. The perfusion of the non-treated leg does not appear to be influenced by a revascularization procedure on the contralateral side. Future studies should incorporate normalized parameters and focus on the value of ICG NIR fluorescence imaging in predicting favorable outcome following revascularization. Abstract Body : PURPOSE The clinical importance of poly (ADP-ribose) polymerase (PARP) as a tumor biomarker warrants the development of novel PARP imaging probes with improved pharmacological properties. PARP inhibitors serve as lead structures for the development of PARP-targeted optical and nuclear imaging probes. The most prominent among such imaging agents are PARPi-FL, an optical probe; its radiolabeled analogue [18F] PARPi; and [18F] FTT, the first clinically applied PARP radiotracer. We present the preclinical in vivo evaluation of 18F-labeled talazoparib, the as-of-yet most potent PARP inhibitor, as second-generation PARP radiotracer. METHODS Enantiomerically pure [18F] talazoparib (>99% enantiomeric excess, >99% radiochemical purity, 45-176 GBq/µmol molar activity; n = 4) was used for this study. Binding of [18F] talazoparib to PARP1 was demonstrated by uptake assays in the HR-deficient cell line HCC1937 and compared to the inactive enantiomer [18F] LT-674. In a pilot in vivo study, HCC1937-xenograft bearing female NOD.CB17-Prkdcscid/J mice (n = 5) were subjected to 1 hour dynamic PET imaging and a 10 minute static PET scan 2 hours after injection of [18F] talazoparib, followed by an anatomical MR scan. Biodistribution was assessed from dynamic PET data as well as ex vivo by gamma-counting of selected organs in order to further investigate the pharmacological properties of the molecule. PARP1 immunofluorescence microscopy confirmed PARP1 expression in the xenograft tissue. RESULTS [18F] Talazoparib was stable in both mouse and human serum over 240 minutes. The radiotracer demonstrated excellent binding to HCC1937 cells in vitro that was blockable by >99% with excesses of both talazoparib and olaparib. As expected, the inactive enantiomer [18F] LT-674 showed only low uptake in HCC1937 cells. Talazoparib was found to block substantially more of the radiotracer signal than olaparib, a phenomenon that was also observed in a cell-based competition assay. In vivo, [18F] talazoparib exhibited balanced clearance by both hepatobiliary and renal excretion pathways and fast blood clearance with a calculated blood half-life of 3.3 minutes. Most radiotracer signal was observed in the excretion-related abdominal tissues (liver, kidney, spleen, and intestine). Moderate bone and muscle uptake was observed, in line with reported specific uptake of previous PARP radiotracers. [1] The HCC1937 xenografts showed [18F] talazoparib signal that intensified over time, resulting in a mean tumor-to-muscle ratio (TMR) of 1.82 ± 0.44 3 hours post injection. The 5.6-fold higher (10.23 ± 1.51) tumor-to-blood ratio underlines the rapid clearance of the tracer, combined with specific muscle uptake. Our data approve feasibility of [18F] talazoparib-PET in mice and warrants exploration of [18F] talazoparib imaging in other cell and animal models. CONCLUSION [18F] Talazoparib presented with rapid and balanced excretion, relatively high bone and muscle uptake, and a TMR that is comparable to [18F] FTT in the same xenograft model (1.9 after 1 hour uptake) [2] . The distinct mode of action compared to first-generation PARP inhibitors suggests that [18F] talazoparib may allow us to study PARP enzymes from a different angle beyond PARP1. [3] To shed light on this hypothesis will be within the scope of future studies.

References: [1] 

Abstract Body : Abstract-Background: One of the complications in a deep inferior epigastric artery perforator (DIEP) flap or profunda artery perforator (PAP) flap is the occurrence of fat necrosis, which occurs in 12-45% of all reconstructions. [1, 2] Intra-operative use of near-infrared fluorescence (NIR) imaging with indocyanine green (ICG) is a technique that has proven its potential to predict fat necrosis, since the use of ICG NIR fluorescence imaging reduces its incidence by more than halve. [3] [4] [5] [6] [7] For the interpretation of the observed NIR fluorescence intensity, the surgeon relies on qualitative assessment of the signal. The aim of this study is to provide a first step in quantification of the NIR fluorescence intensity by gaining insight in the observed perfusion patterns in DIEP and PAP flaps. Method: In this pilot study, NIR fluorescence with ICG was performed in patients undergoing DIEP -or PAP flap reconstruction. After arterial and venous anastomosis, 7.5mg ICG was administered intravenously as a bolus. To prevent autofluorescence, the operation room was cleared of ambient light and gauzes were placed around the flap to limit interfering fluorescence from the thoracic skin. The NIR fluorescence intensity over time was measured with the Quest Spectrum platform®. Postoperatively, four regions were selected by a plastic surgeon based on the NIR fluorescence images: perforator, well-perfused, questionably perfused and poorly perfused. The location of the perforators was selected based on an intra-operatively placed marker. Time-intensity curves were analyzed within these four regions with the Quest Spectrum Framework®. For the quantification, four parameters were extracted from the time-intensity curves: Tmax (seconds), ingress (arbitrary units/seconds), T max slope (seconds) and the max slope (arbitrary units/seconds). Results: During the study period, 14 DIEP flaps and 3 PAP flaps were performed in 14 patients. Thirteen perforators, 17 well-perfused, 8 questionably perfused and 3 poorly perfused regions were identified. The well-perfused regions matched the perforator curves and showed a fast inflow. The questionably perfused regions were more similar to the poorly perfused regions and demonstrated a slow inflow. Outflow was only seen in the perforator and well-perfused regions. There was a significant difference in three of the four measured parameters between the four groups: Tmax: p Conclusion: In the search for the quantitative assessment of tissue perfusion using ICG NIR fluorescence imaging, this study demonstrated well-perfused and perforator regions show a significantly faster inflow. Moreover, within these regions, the outflow phase was reached in all cases, in contrary to the questionably and poorly perfused regions. These results are a first step towards the quantification of tissue perfusion in free flap surgery. To more accurately predict fat necrosis, future studies on ICG NIR fluorescence imaging during free flap surgery, should incorporate quantitative assessment of the measured NIR fluorescence intensity. Abstract Body : Introduction: Cancer surgery is applied in the treatment of solid tumors and aims for the total removal of tumor lesions and complete remission of the patient. Yet, intraoperative identification of tumor margins remains difficult, frequently resulting in positive resection margins and a necessity for adjuvant therapies. Intra-operative guidance using fluorescence molecular imaging has been proven to improve the identification of tumor margins and thus support complete resections. However, insufficient tissue penetration of the fluorescent signal hinders the identification of deep or occult tumor lesions. The combination of fluorescence imaging with intra-operative gamma-probing provides an ideal balance of the deep signal penetration of gamma rays for tumor localization and a fluorescent signal for precise tumor resection. [1] , [2] Preferentially, both modalities are combined in a site-specifically labeled bimodal single-label tracer. In this study we propose a convenient way to label targeting moieties, such as nanobodies, with 99mTc[Tc] and IRDye800CW. 99mTc[Tc] is a gammaemitting radioisotope which can be used pre-and intra-operatively for tumor detection using SPECT/CT or gamma-probing respectively. The near infra-red dye IRDye800CW shows high in vivo signals enabling accurate identification of tumor margins. Methods: Specific nanobodies were engineered to carry a carboxy-terminal hexahistidine-and cysteine-tag separated by a 14 amino acid linker. First, nanobodies were site-specifically labeled with IRDye800CW-maleimide on the cysteine-tag and subsequently with 99mTc[Tc](CO)3 on the hexahistidine-tag.( Figure 1A ) Quality controls were conducted for each step via size-exclusion chromatography and hydrophobic interaction chromatography. (Photo)stability was assessed in function of time and added activity via absorbance and fluorescence measurements. The influence of the bimodal label on in vivobiodistribution and tumor targeting was evaluated. Tumor-bearing mice were imaged using SPECT/CT and fluorescence imaging one hour after intravenous injection of 2 nmol of labeled compound. Ex vivo analysis on organs of interest was performed using fluorescence imaging and gamma-counting. Results: The site-specific labeling of the nanobodies with IRDye800CW and subsequently 99mTc[Tc](CO)3 was successful. However, the labeling strategy was shown to be influenced by the amount of activity used in the second labeling step. Labeling of 25 µg nanobody-IRDye800CW with up to 74 MBq 99mTc[Tc](CO)3 resulted in good preservation of the fluorescent signal. Higher activities led to a decrease in or the loss of the fluorescent signal. Up-scalability of the labeling was proven to be possible by labeling 200 µg nanobody-IRDye800CW with 220 MBq 99mTc[Tc](CO)3 for the in vivo experiments without any consequences for the fluorescent signal. The IRDye800CW/99mTc [Tc] tracers showed an in vivo and ex vivo biodistribution profile with rapid blood clearance, renal clearance, no unexpected background signals, and specific tumor uptake for the targeted nanobody as compared to a non-targeting control nanobody. (Figure 1B -D) Thus, the bimodal label shows no signs of influencing the in vivo biodistribution of the tracer negatively. Conclusion: The bimodal labeling was successfully optimized and could be implemented as a standard bimodal labeling strategy for other nanobodies. Furthermore, the in vivo biodistribution of the tracer is not negatively influenced by the addition of the bimodal label. Therefore, this new bimodal fluorescent/radioactive nanobody-based tracer has the potential to be translated and implemented into clinical practice.

References: [1] -targeting has shown to be a valuable approach for the detection and removal of prostate cancer (PCa) recurrences using radioguidance [1] . Especially within the field of Urology, the rapid adoption of robot-assisted surgery has generated the need for robot-compliant imageguided surgery techniques. To this end, we recently translated a miniaturized DROP-IN gamma probe which has promised to connect radioguidance with robotic surgery [2] . Building on these developments, we set out to evaluate the feasibility and short-term outcomes of robot-assisted PSMA radioguided surgery (RGS) in PCa patients using a dedicated radiotracer and DROP-IN gamma probe. Methods: 30 patients were included in this study with either primary or recurrent prostate cancer, divided over three different high-volume centers in PCa surgery (i.e., Amsterdam -The Netherlands; Hamburg -Germany; Sydney -Australia). Patients were selected based on a positive PSMA positron emission tomography (PET)/CT, where patients with a biochemical recurrent cancer (PSA ≥ 0.2 ng/ml) had either nodal or local recurrent PCa (a maximum of 2 recurrences). Roughly 21 hours before surgery, all participants received an intravenous injection of 99mTc-PSMA I&S (~550 MBq), which was followed by a SPECT/CT, serving as a quality control for successful tracer distribution, where correlation with diagnostic PET/CT was recorded. Robotic PSMA-targeted surgery was realized using the DROP-IN gamma probe. The primary endpoint was feasibility to identify and resect PCa (including recurrences or remnants) in this setting. Secondary endpoints included the comparison between radioactive observations of removed tissue samples and final histopathology, the reduction in prostatespecific antigen (PSA) and the obtainment of a complete biochemical response (i.e., PSA Results: In these PSMA-targeted procedures, the preoperative SPECT/CT images indicated tracer accumulation in lesions as well as background signal in clearance organs. During surgery, interference signals were observed from background tissues (i.e., blood, urine, bladder, and bowel). Due to these background signals, the maneuverability of the DROP-IN seemed to be instrumental for successful target definition in vivo. In both the primary and recurrent setting, nodal and local metastases could be retrieved. Interestingly, in some cases of the primary tumor setting, tumor remnants were still found in the prostate resection bed after an incomplete tumor resection, despite urine contamination. Ex vivo scanning of the specimens excised with a gamma probe, helped to immediately confirm a successful lesion retrieval in all cases. At pathology, nodal metastasis had a median size of 8.7 mm, IQR 3.8-15.0. Postoperatively, no high-grade complications (Clavien-Dindo ≥ grade 3) were observed. Conclusion: PSMA-targeted surgery has proven useful for the intraoperative detection and resection of local and nodal PSMA-avid PCa (recurrences). Based on these initial successes, the multicenter study is being expanded to include more patients and centers. [1-13C] lactate production in metastatic prostate cancer (PCa)2 and established the relationship between Gleason grade and increasing [1-13C] lactate3. However, the specific biological mechanisms behind [1-13C] lactate production in human PCa are still not well characterized. Objective. This study used HP 13C-MRI in patients with low-, intermediate-, and high-risk PCa to identify molecular drivers of [1-13C] lactate production in prostate tumors of varying aggressiveness. Methods. Patients with biopsy-proven tumors visible on proton MRI (1H-MRI) who were also scheduled for radical prostatectomy, underwent 3T HP 13C-MRI following intravenous injection of hyperpolarized [1-13C] pyruvate. Tumor-derived signal-tonoise ratios (SNR) for pyruvate, lactate, and total carbon were calculated in addition to the [1-13C] pyruvate-to- [1-13C] lactate conversion rate (kPL). Whole-mount surgical sections were stained with hematoxylin and eosin, with tumor grade and percent Gleason pattern 4 (%GP4) evaluated by an experienced genitourinary pathologist. The same sections were used for the immunohistochemical analysis of monocarboxylate transporters (MCT) 1 and 4 membrane expression. We also used RNAscope to quantify mRNA expression of lactate dehydrogenase (LDH) subunits A and B as well as pyruvate dehydrogenase E1 subunit alpha 1 (PDHA1). A random forest classifier was used to quantify the overall, tumor-, and stroma-derived expression of these metabolic biomarkers. Results. The study included 10 patients with 15 lesions, of which 2, 11, and 2 harbored Grade Group 1, 2, and 3 disease, respectively. Lactate SNR showed a strong positive correlation with %GP4 (rs = 0.65, P = 0.03), tumor-stroma MCT4 ratio (rs = 0.90, P = 0.002), and LDHA/LDHB ratio (rs = 0.80, P = 0.005). Total carbon SNR demonstrated a strong positive correlation with LDHA/PDHA1 ratio (rs = 0.89, P 10%), the latter demonstrated a significantly higher median tumor-stroma MCT4 ratio (0.37 vs 0.05, P = 0.01), LDHA/LDHB ratio (2.60 vs. 0.31, P = 0.006), and LDHA/PDHA1 ratio (2.20 vs 0.31, P = 0.006), respectively. When the analysis was repeated in single malignant glands harboring clear Gleason pattern 3, 4, and 5 disease, the described ratios were significantly higher in GP5 glands compared to GP3 and GP4 disease (P-value range, Conclusions. We demonstrated differential [1-13C] lactate production in intermediate-risk PCa which contained varying percentages of Gleason pattern 4 disease, confirming that the metabolic signature derived from HP 13C-MRI can provide further stratification of disease aggressiveness. This was associated with increased tumor lactate efflux via MCT4 that was likely driven by a switch from oxidative to glycolytic metabolism as evidenced by the concomitant increase in both the LDHA/LDHB and LDHA/PDHA1 ratios. A single-gland assessment of the described ratios also indicated the presence of two distinct metabolic phenotypes: oxidative and lactate-consuming (Gleason patterns 3 and 4) compared to glycolytic and lactate-producing (Gleason pattern 5) states. These findings provide the potential mechanistic evidence for the increasing [1-13C] lactate production in severe disease, which underpins further clinical translation of HP 13C-MRI as a non-invasive risk-stratification tool to differentiate indolent from aggressive PCa. Abstract Body : Background: With increasing demand for gallium-68 (68Ga) labeled radiopharmaceuticals, alternative sources of 68Ga to the low activity output 68Ge/68Gagenerators are strongly desirable. By using liquid or especially solid targets mounted to medical cyclotrons, it is possible to produce larger amounts of 68GaCl3 via the 68Zn(p,n)68Ga reaction (1, 2) . Purification of the 68Ga from the enriched 68Zn target material and other metal ion impurities is a critical step as these metals compete with 68Ga in the complexation with different chelators, that negatively effects the radiolabeling yields. Predominantly, the metal ion of concern is obviously the dominant Zn2+ present after irradiation, but also the trivalent Fe3+, which, for example, has a higher stability constant (log KML) than Ga3+ for the commonly used chelator DOTA. It was hypothesized that by decreasing the amount of Fe3+ in the cyclotronproduced 68GaCl3 eluate, the 68Ga-radiopharmaceutical yield and its' apparent molar activity (AMA) will consequently increase. In this work, an improved purification methodology is presented, generating low level of Fe3+ in the 68GaCl3 eluate, that can be utilized for high-yield radiolabeling of clinically relevant DOTA-based tracers, such as DOTA-TOC and FAPI-46. By titrations the potential improvement for radiolabelings based on NOTA-and HBED-based tracers was also investigated. Materials and methods: Enriched zinc-68 (68Zn) foils (approx. 120 mg 68Zn, Isoflex USA) were irradiated in a cyclotron with 25 µA, 12.6 MeV protons in a solid target cyclotron system (GE PETtrace 800 and Comecer PTS). The produced 68Ga was purified in an automated purification module (Taddeo PRF, Comecer) using a 2 x UTEVA resin (Triskem) method, as previously described (2) . Modifications of the dilution and wash solutions before purification were performed to decrease the levels of Fe3+ in the final 68GaCl3 eluate. The 68GaCl3 eluate was analysed for metal content by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS), and by titrations (evaluated with radio-iTLC) with three commonly used chelators (DOTA, NOTA, and HBED). The resulting 68GaCl3 eluate was utilized for GMP-radiolabeling of the DOTA-based tracer molecules DOTA-TOC and FAPI-46 using an automated synthesis module (Eckert & Ziegler). Full quality controls (QC) were performed and as an evaluation of the modification's impact, the amount radiolabeled activity, yields, and AMA were determined. Results: Compared to the earlier cyclotron produced 68GaCl3 eluates, the new modifications yielded a 5-fold decrease of Fe impurity levels, and 68Ga titrations with the chelators DOTA, NOTA, and HBED showed an increased AMA of >2-, >3-, and >15-fold, respectively. Notable; compared to earlier investigated generator eluted 68GaCl3, the AMA for DOTA was >40-fold higher (2) . GMP-compliant, radiochemically pure, and stable products of [68Ga]Ga-DOTA-TOC (>5 h stability tested) and [68Ga]Ga-FAPI-46 (>4 h stability tested) were achieved. The final product activity at end of synthesis (EOS) of ~6 GBq/batch of [68Ga]Ga-DOTA-TOC or [68Ga]Ga-FAPI-46 after 68 min bombardment was significant higher, compared to the corresponding batches of ~0.6 GBq/batch, obtained from the generator produced 68GaCl3 eluate (see Table 1 ). Conclusions: In this study, a new improved purification approach was demonstrated to decrease the levels of competing metal ions (i.e., Fe3+) in the cyclotron solidtarget produced 68GaCl3 eluate, which enabled GMP-compliant high-yield DOTA-peptide synthesis of clinically relevant tracers. To our knowledge, this is the first data showing the highest yields obtained for DOTA-based tracers with cyclotron-produced 68Ga. Titrations indicate that radiolabelings of NOTA-or HBED-based tracers may give even better yields using this cyclotron-produced 68GaCl3 eluate. Abstract Body : Introduction: In penile cancer, sentinel lymph node biopsy is a routine procedure for nodal staging. Most commonly, radiocolloids are used to provide various forms of preoperative imaging and intraoperative guidance in this procedure. This has been further extended with intraoperative fluorescence imaging by introduction of hybrid radio-and fluorescence-tracers, such as ICG-99mTc-nanocolloid. In parallel, an alternative magneticparticle based approaches has been introduced that uses superparamagnetic iron-oxide nanoparticles (SPIONs). In this magnetic approach, intraoperative localization is currently limited to detection with a handheld-magnetometer-detection probe only. In an attempt to improve intraoperative localization in the magnetic approach, this study investigates a novel 3D magnetic particle imaging (MPI) and navigation modality (freehand MPI) in procedures that use a combination of magnetic and indocyanine-green (ICG) fluorescence guidance. Methods: The freehand MPI modality was built using a surgical navigation device, optical tracking system, magnetometer probe and dedicated reconstruction software. The combined use of magnetic and fluorescence guidance was facilitated by administering a hybrid cocktail of both SPION and ICG tracers. The use of both the 3D freehand MPI modality and the tracer mixture were investigated in phantoms, in human skin explants and during porcine surgery. Results: The phantom and human skin studies showed that the current freehand MPI modality had a depth penetration up to 1.5 cm, a resolving power of at least 7 mm and a sensitivity of 2.2*10-2 mg/mL SPIONsa. Evaluation during porcine surgery showed that freehand MPI could improve intraoperative target identification by allowing for an augmented reality visualization of the localized tracer uptake directly on the surgical field. Furthermore, the sites of SPION accumulation found in the fhMPI scans could be used as target for virtual reality navigation and depth estimation prior to incision. This boosted surgical confidence, as opposed to using magnetic tracing only. Furthermore, in analogy to the hybrid approach used in radioguided surgery, fluorescence imaging provided a real-time visual confirmation of the sentinel lymph nodes during the actual excision. Conclusion: 3D freehand MPI is feasible in vivo, supporting a better identification of SPION accumulation and providing subsequent surgical navigation. Furthermore, the additional use of ICG provides the option for direct real-time confirmation of the localized lymph nodes. Abstract Body : Within the past decade, the incidence of tick-born encephalitis virus (TBEV) infection has risen significantly and medically speaking, it has become the most important arthropod-born virus in Europe. TBEV is a neurotropic flavivirus that potentially causes lethal encephalitis while among survivors, long-term neurological sequelae are often observed (1) . At present, our knowledge regarding viral tropism and its distribution along specific anatomical brain regions remains limited. In this study, we aimed to visualize flavivirus infection within the mouse brain using Optical Projection Tomography (OPT) (2) and investigated the possibility to map viral spread to specific brain regions by co-registering OPT to anatomical Magnetic Resonance Images (MRI). Immunocompromised mice were intracranially inoculated with Langat virus, a BSL-2 flavivirus closely related to TBEV and at endpoint, brains were harvested and processed for whole-organ OPT. After immunofluorescent labelling of nonstructural protein 5 (NS5), viral and autofluorescent OPT images were acquired and reconstructed into DICOM format. Afterwards, volumetric T1-weighted (MDEFT, TR/TE/TI 3000/3/950 ms, 40 mm3 voxel size) MR images were acquired. Voxel-to-voxel affine transformation matrices were calculated using the autofluorescent OPT images and applied to those displaying viral NS5 signal. Finally, fusion images were created, and virus-containing brain regions were determined according to the Turone mouse brain atlas (3) . Affine registration successfully resulted in fusion images delineating viral signal in a detailed anatomical context provided by high-resolution MRI and revealed infection in gray matter (GM), white matter (WM) and the entire ventricular system. For GM specifically, we observed viral signal in entorhinal cortex, dorsal endopiriform nucleus, piriform area, primary visual field and primary somatosensory cortex (upper limb and barrel field). As for WM, the olfactory limb of the anterior commissure, anterior forceps of corpus callosum and supra-callosal WM contained viral signal. Interestingly, the choroid plexus of the lateral ventricle, subventricular zone and olfactory limb of the anterior commissure, all part of the rostral migratory stream, suggest the specific infection of the ventriculo-olfactory neurogenic system in this mouse model. In this study we demonstrated, for the first time, the successful creation of OPT-MRI fusion images and their value for anatomical brain mapping of flavivirus infection. In the future, the proposed method may significantly contribute to unraveling the tropism of TBEV and provide novel insights regarding its neurotropic properties. Abstract Body : Purpose: Inflammatory bowel diseases (IBD) are progressive disorders of the gastrointestinal tract that hinder patient quality of life if untreated. Systemic treatments have yielded mixed results in mouse models and the clinic, where a lack of selective delivery has yielded low drug concentrations in the colon. Thus, it is fruitful to identify potential drug delivery targets through targeted molecular imaging. Extracellular matrix (ECM) proteins are upregulated in a variety of inflammatory diseases due to increased tissue remodeling and have served as attractive delivery targets 1-4. Tenascin C, in particular, is a glycoprotein with multiple functional isoforms within the ECM. While various tenascin C targeting proteins have been specific in several oncology models5, it remains largely unexplored in IBD. Methods: In this work, colon immunohistochemistry from mice provided with dextran sulfate sodium (DSS) in drinking water revealed expression of the alternatively spliced D domain of tenascin C (TNC D). Novel proteins with sub-micromolar binding affinity to TNC D were generated. Using amine chemistry, we labeled these proteins with IRDye 800CW, a commonly used near infrared (NIR) fluorophore. After confirming minimal changes in binding affinity after bioconjugation via in vitro binding assays, the reagents were dosed in murine DSS colitis. Results: We report evidence of in vivo specificity by dosing the imaging antibody with and without a high molar dose of unlabeled antibody as shown through macroscopic organ-level imaging with a NIR scanner as well as a custom NIR endoscope. Our imaging data revealed TNC D expression in punctate lesions ranging from the distal to proximal ends of the colon. Subsequent ex vivo biodistribution via total organ digest helped quantify the absolute uptake in the diseased colons between blocked and unblocked groups and revealed significant differences in colon uptake and undetectable specific uptake in non-disease organs. Following ex vivo imaging, cellular level specificity was confirmed via IHC and microscopy. Conclusion: It is beneficial to be mindful of delivery applications in addition to imaging. Thus, it's fruitful to identify ECMs with high absolute local expression unique to disease tissue early on in preclinical research. The murine blocking experiments estimate absolute expression of TNC D at the site of lesions is likely greater than 100 nM. This estimate is similar to cell surface molecular imaging targets present in autoimmune diseases such as type 1 diabetes and various cancers with high reported imaging target to background ratio (TBR). Abstract Body : Over the last 40 years, gadolinium (Gd)-based contrast agents have served as the gold-standard for contrast enhancement in magnetic resonance imaging (MRI). Unfortunately, Gd chelates retain high false-positive rates and can lead to harmful side effects, such as nephrogenic systemic fibrosis and Gd-deposition in the brain. Although there has been substantial research using superparamagnetic iron oxide nanoparticles (SPIONs) as contrast agents due to their robust dark T2* MRI signal, these particles suffer from some limitations. SPIONs produce a constant dark signal that is always "on" and can be easily confounded by the negative contrast produced by natural iron stores. Here, we propose poly(lactic-co-glycolic acid) (PLGA) Nano-, Encapsulated, Manganese Oxide (NEMO) particles due to their ability to convert from an "off" to "on" state in low-pH environments, such as tumor microenvironments, rendering the particles pH-activatable. Three types of manganese oxide cores (MnO, Mn2O3, and Mn3O4) were either synthesized through thermal decomposition of manganese(II) acetylacetonate or purchased from Company A or Company B. The composition of the nanocrystals was confirmed using X-Ray diffraction (XRD), which showed impurities in cores from Company A that corresponded to calcium, sodium, and zinc. Transmission electron microscopy (TEM) results showed that the true nanocrystalline sizes directly contradict the reported diameters with up to a 30nm and 60nm difference than reported for Company A and Company B, respectively. Company A's nanocrystals presented with a consistent, non-uniform distribution with some rod-like figures, while Company B's nanocrystals presented with an unknown substance creating a persistent film across all TEM images. The diameters for in-house nanocrystals were ~35nm and ~11nm for MnO and Mn3O4, respectively, with uniform distributions. All manganese oxide cores were then encapsulated within PLGA to form NEMO particles using a single emulsion technique. To produce bright signal on MRI, MnO must degrade to produce Mn2+ ions at low pH. To assess Mn2+ release and resulting MRI signal via controlled release, NEMO particles were incubated at pH 7.4, 6.5, and 5 to mimic pH in blood, tumor extracellular space, and intracellular endosomes, respectively. Supernatants were analyzed by inductively coupled plasma -optical emission spectrometry (ICP-OES) for released Mn content and 1T MRI for T1 signal enhancement. After encapsulation of the first of three replicates, controlled release showed greater than 100% release of Mn2+ at 24h in pH 5 buffer for Company A Mn2O3, Company B MnO, and Company B Mn3O4, which was attributed to impurities and poor encapsulation efficiency. According to scanning electron microscopy (SEM), micron-sized particles were synthesized when encapsulating cores from both Company A and Company B. In-house NEMO particles showed a uniform nano-sized distribution with ~94% Mn2+ release for MnO and ~41% Mn2+ release for Mn3O4 after 24h at pH 5. In general, the T1 MRI signal of particle supernatants increased with decreasing pH as expected due to greater release of Mn2+ under acidic conditions. Specifically, In-house MnO and Company A Mn3O4 presented with the maximum T1 MRI signal. Contrarily, both samples of Company B Mn2O3 presented with no discernable signal change with varying pH indicating that these particles were very poor contrast agents. Based on our preliminary assessment, it appears that in-house synthesis of smaller batches of manganese oxide nanocrystals is currently preferable to commercialized sources to ensure sample purity, a tight size distribution, and maximum T1 MRI signal. Ultimately, determining a precise and reliable way to synthesize commercially available manganese oxide nanocrystals at large scales will allow for reproducibility of NEMO particles bringing them one step closer to being used in a clinical setting. The goal of this work is to develop a PSMA-targeted Near InfraRed Fluorescent probe to offer surgeon a valuable intraoperative tool for allowing a complete tumor resection with the possibility of killing the residual cancer cells by PDT. Methods PSMA-617 binding motif was synthetized as reported5 and conjugated to IRDye700DX-NHS. The PSMA affinity of the dye was examined in human PCa cells with different PSMA expression (LNCaP, PC3, PC3-PIP, and PC3-FLU) by FACS, whereas cellular uptake and dye localization (endolysosomal pathway) was studied by fluorescence confocal microscopy. IRDye700-PSMA-617 photodynamic properties were evaluated in PC3-PIP and PC3-FLU cells after irradiation with a laser at 635 nm (100 J/cm2). MTT assay was performed 1h post irradiation. Finally, PC3-PIP tumor bearing mice were injected with two doses (1 and 5 nmol) of IRDye700-PSMA-617 and imaged at 1h, 4h and 24h post injection. After imaging mice were sacrificed and organs harvested to detect the fluorescence signal. Results and discussion The conjugation of IRDye700DX with PSMA-617 did not affect the exc/em wavelengths of the fluorophore (Persuasive data). As expected, the affinity of IRDye700-PSMA-617 towards PCa cells followed the order of their PSMA expression, i.e. PC3-PIP>LnCap >PC3, PC3-FLU (Fig1.A). Confocal microscopy images confirmed these results (Fig.1C) , and showed a co-distribution of the probe and endolisosomal markers in the perinuclear region (Persuasive data). PDT experiments showed a concentrationdependent decrease of cell viability (from 75% at 10 nM to 12% at 500 nM), whereas controls did not show any cytotoxicity (Fig.1B) . NIRF imaging on mice bearing PC3-PIP tumors showed a significant tumor uptake after the injection of both 1 or 5 nmol (Fig.1D) at any time points. The maximum tumor-to-muscle ratio (ca. 60) was observed, for both doses (Persuasive data), 24h post-injection. Importantly, urine, healthy prostate, and bladder were not fluorescent at 24h post injection (Fig.1E ). Conclusion A novel PSMA-targeted NIRF dye with dual imaging-PDT capabilities was synthetized. The probe displayed cellular uptake and photo-induced cytotoxicity proportional to the PSMA expression. NIRF imaging, performed on a mouse model with high PSMA expression, displayed an excellent and specific tumor uptake with a significantly high tumor-to-background ratio at 24 h post injection, when healthy prostate, kidneys and bladded were not fluorescent. Abstract Body : Introduction: Decision-making and dexterity are considered the hallmark features for improving surgical performance. With the advances in (robot-assisted) minimallyinvasive surgery and image-guided surgery, improving the surgical performance features is becoming increasingly more important. Recently, DROP-IN gamma probes were introduced to facilitate radioguidance in robotic surgery. This study investigates if robotic radioguidance could be further improved with a novel tethered Click-On gamma probe design, which further integrates the detection modality with the robotic instruments themselves. This is done using a thorough investigation of decision-making and dexterity in both a phantom setting and during porcine surgery. Methods: Using computer-assisted design, 3D print prototyping and precision machining, we created a miniaturized gamma probe that can be integrated with a robotic instrument using a clicked concept. To allow for a quantified investigation on the surgical performance, fiducial markers were integrated in the Click-On gamma probe housing and a custom computer-vision algorithm was built to track the gamma probe movements via the robotic laparoscope, allowing for a reconstruction of the probe movement trajectory. Using a dexterity phantom, the tasks performed with the Click-On prototype were compared to those performed with the parental DROP-IN probe using the completion time of the specific-tasks and the quantified features of the extracted movements. To investigate the impact of radioguidance on surgical decision-making, we also performed a blinded study during porcine surgery, wherein surgeons had to localize a hidden radioactive 57Co-source using either palpation or Click-On radioguidance. Results: A fully functional Click-On gamma probe was fabricated, which allowed for a successfully integration with the robotic forceps instrument, while preserving the grasping function and rotational freedom of the instrument. With a diameter procedural time, and a more focused movement.. Conclusion: The integration of gamma probe designs within the surgical instruments themselves provides a next step forwards in the optimization of minimal-invasive surgery using radioguidance. The value of this approach was highlighted by its impact observed on surgical decision-making and dexterity. Abstract Body : Introduction Perfluorocarbon(PFC)-based compounds have been largely used to label immune system cells and track the inflammatory process by Magnetic Resonance Imaging (MRI).1 However, information is still lacking about the impact of these agents on the monocyte/macrophage polarization towards one of their cell subsets and on the best way to image them. In the present work, a PFC-based nanoemulsion (PFCE-NE) was developed to monitor the course of inflammation in a model of spinal cord injury (SCI), a pathology in which the understanding of immunological events is of utmost importance to select the optimal therapeutic strategies.2 Moreover, as in the literature there is not a standard protocol for fluorinated system-based imaging (single or multiple administrations, 19F MRI or PRESS, etc…), the various protocols adopted have been compared in terms of signal quantification, amount of contrast agent delivered to the diseased region, information gained. Finally, the influence of PFCE-NE over macrophage polarization was investigated both in vitro and in vivo. Methods The PFCE-NE was prepared by sonication of a mixture of Perfluoro-15-crown-5 ether, the fluorescent phospholipid Rhodamine-DOPE, and 10 % w/v Kolliphor P188. The effects of PFCE-NE over macrophage polarization were studied in bone marrow-derived macrophages extracted from C57BL/6J mice. Experiments were carried out on M0, M1, or M2 macrophages, incubated with PFCE-NE (40 mM in fluorine) for 1, 6, or 24 h. Then, the PFCE-NE was injected in SCI mice (20 mmol F/Kg) and different cell tracking protocols were tested (single and multiple PFCE-NE administrations, 19F-MRI or MRS detection, and the option of pre-saturating Kupffer cells through a pre-injection of phospholipid-based vesicles) (Figure 1 ). At the end of the MRI study, spinal cords were excised and the impact of PFCE-NE over macrophage polarization was investigated ex-vivo. Results/Discussion M0, M1, and M2 cells avidly took up the nanoemulsion (mean diameter 135 nm), even if with different rates. A switch towards M2 (anti-inflammatory) phenotype was observed, especially for M0 cells. The administration of PFCE-NE to SCI mice allowed to monitor by 19F MRI the ongoing inflammation (Figure 2 ), individuating a possible therapeutic window for stem cell transplantation, around day 8 to 11 post-injury. Pre-saturation of Kupffer cells with liposomes allowed for significantly higher contrast agent accumulation at the lesion site (1038.4 ± 160.6 nmol 19F with saturation, 463.3 ± 45.5 nmol w/o saturation, p Conclusions In conclusion, this work demonstrates that PFCE-NEs can be used as probes to non-invasively track the monocytes/macrophages infiltration in a SCI model by 19F-MRI. Moreover, the influence exerted by PFCE-NE on macrophage phenotypes, with a prevailing switching towards the M2 profile, suggests a possible theranostic function, with tremendous potential future applications. Nevertheless, extensive behavioral and toxicity tests are still required. Acknowledgement The authors acknowledge the Italian Ministry of Research for FOE contribution to the Euro-BioImaging MultiModal Molecular Imaging Italian Node (www.mmmi.unito.it), the Open Lab of Advanced Microscopy (OLMA) at the Molecular Biotechnology Center (MBC) for support.

References: [1] Wu, L, Liu, F, Liu, S, Xu, X, Liu, Z, Sun, X. Perfluorocarbons-Based 19F Magnetic Resonance Imaging in Biomedicine. Int J Nanomedicine. 2020;15:7377-7395. Abstract Body : Background: AbbVie is evaluating the repulsive guidance molecule A (RGMa) inhibitor monoclonal antibody Elezanumab for the treatment of spinal cord injury. There is preclinical evidence suggesting a correlation of motor function recovery after spinal cord injury with regeneration of serotonin axons. The presynaptic serotonin transporter (SERT) has been shown to be modulated after injury and in response to treatment, and the synaptic vesicle glycoprotein 2A (SV2A), expressed on the membrane of synaptic vesicles and present in spinal cord grey matter, may both serve as biomarkers for the evaluation of neuroregeneration in the spinal cord. Here we report the evaluation of two high affinity SERT radioligands and one SV2A radioligand for imaging uninjured spinal cord in nonhuman primates (NHPs). Materials and methods: Precursors and standards for fluorine-18 labeled (S)-2'-(3-fluoropropoxy)methyl)-6-nitroquipazine ((S)- [18F] 1, hSERT Ki: 0.46 ± 0.08 nM)1 and (R)- [18F] 1 (hSERT Ki: 0.12 ± 0.01 nM)1 were prepared by a 5-step synthesis. Radiolabeling of 1 was established by direct fluorination using the tosylate-precursor. The SV2A radioligand [18F] SynVesT-1 was prepared by Cu-mediated radiofluorination of the trimethyltin precursor.2 The distribution and density of SERT and SV2A in the NHP and human brain and spinal cord were evaluated by autoradiography and tissue homogenate binding assays using radioligands (S)- [18F] 1, [3H]UCB-J and [3H]SynVesT-1. A total of eight PET measurements were performed in six cynomolgus macaques. Baseline and pretreatment PET measurements with the SERT inhibitor escitalopram (2 mg/kg) were performed in a FOCUS-220 PET system for (S)- [18F] 1 with the brain or spinal cord in the field of view. Binding of (R)- [18F] 1 in brain and spinal cord and [18F] SynVesT-1 in spinal cord (n=2) was also evaluated. Blood samples were collected for evaluation of metabolism of the radioligand and computation of the arterial input function. PET images were evaluated using MRI and CT-based regions of interest for the brain and spinal cord, respectively. Results: Both [18F]1 enantiomers and [18F] SynVesT-1 could be prepared with high chemical yield and molar activity. The SV2A density in the NHP and human spinal cord was 5-10 fold lower than in brain (cortical grey matter). Metabolism of [18F] 1 was moderate to slow, with ~40% remaining at 60 min post-injection. (S)- [18F] 1 binding was highest in the cervical part of the spinal cord. Binding of (S)- [18F] 1 was reliably quantified in brain, and specific binding was reduced by ~70% following administration of escitalopram. Similar results were observed in cervical spinal cord with escitalopram pretreatment. Baseline PET measurements with (R)- [18F] 1 and [18F] SynVesT-1 demonstrated binding in the cervical spinal cord. The binding kinetics of (R)- [18F] 1 in the brain and spinal cord were slow and not suitable for quantification using compartmental modeling. Spinal cord binding of [18F] SynVesT-1 was reversible and could be quantified using a two-tissue compartment model. Conclusion: In vitro and pre-clinical imaging studies indicate that quantitative molecular imaging of SERT and SV2A, in the spinal cord, may serve as a potential translational biomarker for the evaluation of neuronal regeneration.

References: 1 Gerdes J., Zheng M.-Q., Braden M., Holden D., Ropchan J., Lin S., Carson R., Huang Y. Novel F-18 labeled 6-nitroquipazine derivatives as high affinity tracers for the serotonin transporter. 2012 Journal of Nuclear Medicine, 53:S1, 291. 2 Li S., Cai Z., Wu X., Holden D., Pracitto R., Kapinos M., Gao H., Labaree D., Nabulsi N., Carson R.E., Huang Y. Synthesis and in vivo evaluation of a novel PET radiotracer for imaging of synaptic vesicle glycoprotein 2A (SV2A) in nonhuman primates. 2019 ACS Chem Neurosci, 10:3, 1544-1554. Abstract Body : Background: Intraplaque hemorrhage (IPH) is a pivotal driver of atherosclerotic plaque progression, instability and rupture [1, 2] and is an imaging predictor of stroke and myocardial infarction [3, 4] . Advanced imaging techniques were required to accurately detect IPH for risk stratification of patients and early intervention of unstable plaques [5] . We aimed to use magnetic particle imaging (MPI) combined with computed tomographic angiography (CTA) for precise detection of IPH in vivo. Methods: Twenty-two human carotid atherosclerotic samples were collected from the patients undergoing carotid endarterectomy (CEA) and imaged using the MPI/CT scanner. Histological examinations were used to analyze the source of MPI signals in human tissues. Six-week-old ApoE-/-mice were used to establish the tandem stenosis (TS) mouse model, which provided comparative arteries of unstable plaques with IPH, stable plaques and normal arterial wall in the same mouse [6, 7] . TS ApoE-/-mice were performed with MPI and CTA imaging. 7.0 T T1-weighted fat-suppression MRI and Prussian blue staining were performed to confirm the presence of IPH. Results: Human CEA samples exhibited strong endogenous MPI signals ex vivo. Histological examinations found that MPI signals were related to the presence of IPH and particularly hemosiderin, the degradation product of IPH. In vivo MPI/CTA images of TS ApoE-/-mice revealed that endogenous MPI signals were detected in the unstable plaques rather than stable plaques and normal arterial wall. Quantitative analysis indicated that MPI signals were significantly higher in unstable plaques (7.29 ± 4 .48) compared with stable plaques (2.63 ± 1.12) and normal arterial wall (2.09 ± 0.79) (P 1.34 ± 0.08) than stable plaques (1.00 ± 0.06) and normal arterial wall (0.96 ± 0.02) (P Conclusions: MPI/CTA is a novel technology that allows in vivo detection of atherosclerotic plaques with IPH and holds great promise for identification of unstable plaques in humans. Abstract Body : Introduction: Aberrant cell surface glycosylation has been described as one of the key hallmarks of cancer, with hypersialylation the most commonly observed alteration.1 Changes in the glycome of cells can be monitored by incorporating unnatural sugars bearing bioorthogonal chemical reporters. Monitoring hypersialylation could provide an insight into tumour progression, proliferation and potentially be used for monitoring treatment response.2, 3 We describe here improved in vitro imaging of hypersialylation using a novel triacetylated cyclopropene mannosamine (Ac3ManNCyoc) derivative. Methods: Mannosamine was functionalised with a carbamate linked cyclopropene handle. The monosaccharide was synthesised with different degrees of acetylation (Ac1 to Ac4), which we expected to show different diffusion rates across the cell membrane due to varying hydrophobicity as well as variable aqueous solubility. Ac4ManNCyoc has limited aqueous solubility as a result of its high clogP (2.20) . A tetrazine-PEG11-DyLight649 probe was synthesised and used to label the methyl cyclopropene moiety. The extent of incorporation of the sugar analogues was assessed in three cell lines (COLO205, MDA-MB-231, PANC-1) by flow cytometry. Results: Varying the levels of sugar acetylation resulted in significant differences in measured fluorescence intensity, indicating that there was an optimum level of acetylation that balanced sugar solubility and membrane penetration. In COLO205 cells, this optimum was found with tri-and diacetylated sugar derivatives (SBR 16.6 and 17.4 respectively). The hydrophobicity of Ac4ManNCyoc results in poor aqueous solubility limiting the substrate availability. Ac3ManNCyoc and Ac2ManNCyoc strike a balance, whereas the hydrophilicity of Ac1ManNCyoc results in poor cellular uptake. The improved labelling with Ac3ManNCyoc and Ac2ManNCyoc was also observed in MDA-MB-231 and PANC-1 cell lines (Fig 1B-C) . Ac3ManNCyoc yielded the highest signal-to-background ratio in both (97.6 and 52.8 respectively) with the increased labelling demonstrating the increased hypersialylation in these tumour lines. Ac3ManNCyoc Labelled PANC-1 cells were visualised using fluorescent microscopy ( Fig 1D-E) . Conclusions: Ac3ManNCyoc is a novel candidate for probing tumour hypersialylation by metabolic labelling and subsequent in vivo imaging in murine models. The triacetylated sugar represents the optimal compromise between sugar solubility and membrane penetration. and PANC-1 (C) cells were incubated in the presence or absence of the appropriate ManCyoc sugar 125μM (Ac1, Ac2, Ac3 or Ac4) for 24 hours. They were subsequently treated with the dye solution (5 μM Tz-Dylight647 + 50 nm Sytox green cell death stain) for 1 hour. Statistical analysis was performed using an unpaired t test with Welch correction (**** P≤0.0001,*** P≤0.001, ** P≤0.01, * P≤0.05). Data are median ± SEM (n = 3, per group). Fluorescence microscopy of PANC-1 cell glycans. Cells were cultured with (D) or without (E) Ac3ManNCyoc (125 μM) for 24 h and then incubated with a 5 μM solution of Tz-AF647, Red: AF647, Blue: DAPI nuclear stain, scale bar 50 μm.

Last Name: Parle Email: drp45@cam.ac.uk

Carolyn Bayer, Tulane University, carolynb@tulane.edu

Abstract Body : New-onset hypertension during pregnancy, termed preeclampsia, is a leading cause of maternal mortality and a probable risk factor predisposing both the mother and offspring to cardiovascular disease. While inadequate growth and remodeling of the placental vasculature, beginning at the earliest stages of pregnancy, is the likely cause of preeclampsia, little is understood about the in vivo, longitudinal progression of placental development. The establishment of methods for longitudinal, in vivo molecular imaging of markers of placental angiogenesis and vascular remodeling, therefore, was the objective of this work. Using an established in vivo rat model of preeclampsia and molecularly targeted microbubbles (MBs), we acquired longitudinal contrast-enhanced ultrasound (CEUS) images of placental development. Timed-pregnant Sprague Dawley rats were surgically modified on gestational day (GD) 14 using the reduced uterine perfusion pressure technique (RUPP) to induce preeclampsia [1] . Vevo Target-Ready MicroMarkers (FUJIFILM, VisualSonics) were used for CEUS imaging. MBs were conjugated with biotinylated arginine-glycine-aspartate (RGD), a peptide that binds to ανβ3, a marker of angiogenesis. A full video loop of nonlinear contrast images was acquired before MB administration to obtain a baseline measurement of the nonlinear background signal from tissue. Then, a 200 μL bolus of MBs was administered followed by a 150 μL saline flush while continuously recording nonlinear CEUS images. At 10-minutes after MB injection, a destructive US burst was applied and an additional two minutes of nonlinear CEUS images were acquired. Thirty minutes later, a second bolus of MBs conjugated with biotinylated L-arginineglycine-glutamic acid-serine (RGES), a nontargeted isotype control of RGD, was administered and the same imaging sequence was implemented. All injections were administered via tail vein catheter on GD14 and through the indwelling jugular vein catheter on GD16 and 18. We used a bicompartmental model to evaluate ανβ3-targeted microbubble binding, placental microvasculature architecture, and placental perfusion in both a normal pregnant and a rat model of preeclampsia. A linearized time intensity curve was extracted from each pixel of the images of each CEUS sequence and fitted to the bicompartmental model using nonlinear iterative leastsquares curve fitting [2] . The parameters of molecular expression and perfusion were confirmed with histology. Using the developed imaging methods and models, we have demonstrated in vivo, longitudinal imaging of the reduced expression of the angiogenic marker as well as reduced microvasculature development and perfusion in the preeclamptic animals in comparison to the controls. The imaging methods and the bicompartmental model allow for the assessment of the complex development of the placenta, which has previously not been demonstrated longitudinally and in vivo. The developed imaging and analysis methods will enable the discovery and validation of novel therapeutics for preeclampsia.

References: [1] B. T. Alexander et al., "Reduced uterine perfusion pressure during pregnancy in the rat is associated with increases in arterial pressure and changes in renal nitric oxide," Hypertension, vol. 37, no. 4, pp. 1191 -1195 , Apr 2001 S. Turco, I. Tardy, P. Frinking, H. Wijkstra, and M. Mischi, "Quantitative ultrasound molecular imaging by modeling the binding kinetics of targeted contrast agent," Physics in Medicine Image/ Figure Abstract Body : Phase-changing nanodroplets (PCNDs) are nanosized constructs with perfluorocarbon liquid cores that can be remotely vaporized to form highly echogenic gaseous microbubbles. After vaporization, the presence of the microbubbles within the tissue allows for ultrasound images of the agents to be obtained. The PCND nanometric size results in better biological stability and longer circulation times than conventional microbubble ultrasound contrast agents. Beyond their role as ultrasound contrast agent precursors, the PCNDs can be loaded with fluorophores, to allow for techniques such as near-infrared (NIR) fluorescence imaging to be employed to follow their biodistribution. These probes are particularly useful in studying the dynamics of nanoparticle circulation, accumulation and on-target release1, but also to overcome the limits of fluorophore photostability. A standing challenge in the fabrication of nanoparticle-encapsulated dyes, however, is the aggregation-induced quenching that is associated with large dye loadings. This decreases the quantum yield and severely limits the brightness of dye-loaded nanoparticles2. In this study, we employ a lipophilic carbocyanine DiOC18(7) (DiR) NIR dye to produce NIR-sensitive fluorescent PCNDs. Apart from DiR, the PCNDs are loaded with Epolight 3072, a NIR-absorbing dye with a maximum absorption near 1064 nm. This chromophore absorbs pulsed laser light, resulting in localized heating and pressure waves that cause the liquid perfluorocarbon core to vaporize. In the inactivated (liquid) PCND phase, the loaded DiR dye is in a self-quenched state, as a large number of fluorophore molecules are confined to a small space. When activated, the PCNDs undergo a liquid-to-gas phase transition and increase in diameter approximately 5-fold, allowing for the dye molecules to disperse and un-quench and resulting in increased fluorescence intensity. This phenomenon was observed by varying the amount of loaded DiR and studying the fluorescence intensity change in polyacrylamide phantoms containing the PCNDs. An example of such a phantom is presented in Fig.1 , where panel A shows the 800-nm channel and the white light channel of the phantom activated with an "FMI" stencil, and panel B is an ultrasound frame of a portion of the same phantom. The ratio between the fluorescence signal intensity in activated regions and fluorescence signal intensity in the non-activated, background region was used as a measure of fluorescence enhancement. A fluorescence enhancement of up to 1.5x was observed in the cases of DiR-lipid molar ratios between 1 x 10-3 and 6 x 10-3. Batches with dye-lipid ratios lower than 1 x 10-3 did not show a self-quenching effect, while batches where the lipid-dye ratio exceeded 6 x 10-3 showed a lower enhancement due to the unquenching being only partially achieved. Several nanodroplets concentrations in phantoms are studied, to elucidate whether inter-particle interactions play a role in the quenching process. Overall, these probes are promising in achieving an increased tumor-to-background ratios (TBR) in NIR fluorescence techniques. Nontargeted agents generally suffer from high background signals, lowering their TBR. Even for targeted agents, that offer molecular insights into the tumor microenvironment, unbound agent in the blood can hamper imaging2. Combining the possibility of on-demand fluorescence enhancement like shown herein with molecular targeting could provide new avenues for diagnosis or monitoring of cancer progression. Abstract Body : Pancreatic cancer (PC) is an extremely lethal disease, with an overall survival rate of 10% [1] . Surgical resection remains the only potentially curative option, but resections are complicated by infiltrative disease, proximity of critical vasculature, peritumoral inflammation, and dense stroma [2, 3] . Furthermore, translating preoperative images to the intraoperative setting poses additional challenges for tumor detection, and can result in undetected lesions. Thus, PC has high rates of incomplete resections, and disease recurrence [4] . Fluorescenceguided surgery (FGS) has emerged as a method to improve intraoperative detection of cancer and ultimately improve surgical outcomes. Initial clinical trials have demonstrated feasibility of FGS for PC, but there are limited targeted probes under investigation for FGS in PC. Because of the heterogeneity of PC, it is essential to investigate other targets for FGS, either in combination with existing probes, or independently [5] . Mucin16 (MUC16) is a glycoprotein overexpressed in 60-80% of PCs, yet this biomarker has not been investigated for FGS of this disease. Our previous work demonstrated initial feasibility for targeting MUC16 for FGS in subcutaneous and orthotopic mouse models with a murine antibody conjugate, muAR9.6-IRDye800 [6] . Herein, we investigate the translational potential of the humanized variant of this antibody, huAR9.6-IRDye800, for fluorescence-guided surgery of PC. The objective of this project is to develop preclinical evidence to support the clinical translation of huAR9.6-IRDye800, and to investigate the role of antigen expression and tumor microenvironment on accumulation and contrast. Our hypothesis is that huAR9.6-IRDye800 is a non-toxic fluorescent probe that improves the intraoperative detection of PC with FGS in MUC16-expressing tumors. AR9.6 was humanized via complementarity determining region grafting. Humanization did not impact binding of the antibody to MUC16 ( Figure 1A ). HuAR9.6 was conjugated to IRDye800 via amide coupling synthesis at dye to protein ratios of 0.3, 1, 2, and 4 ( Figure 1B ). An ELISA confirmed that the conjugation ratio did not impact antibody affinity for MUC16 ( Figure 1C ). Western blotting confirmed binding of the humanized conjugate to human PC cell lines ( Figure 1D ). While in vitro binding was not impacted by conjugation ratio, in vivo studies in healthy CD-1 mice, shown in Figures 2A and 2B , demonstrated that increased dye to protein ratios resulted in increased liver signal at 48 hours after injection, consistent with other studies in the literature [7] . Dynamic contrast enhancement observed after injection of 50 µg of huAR9.6-IRDye800 in MUC16-expressing subcutaneous xenografts showed that tumor to background ratios (TBRs) continued to increase up to 9 days after injection (peak TBR 6.95 ±0.392, N=4), as shown in Figure 3A and C. Necropsies conducted at 9 days post-injection showed 5.92 times greater signal in the tumor as compared to the liver ( Figure 3B ), suggesting the potential to differentiate tumor from key background organs in PC, and detect metastatic lesions. Results of the dynamic contrast enhancement were consistent with previous dynamic imaging conducted with muAR9.6-IRDye800 [6] . HuAR9.6-IRDye800 localization within tumors was observed with fluorescence microscopy ( Figure 5 ). A 28-day safety and biodistribution study (not pictured) of both high (80µg) and low (20µg) dose huAR9.6-IRDye800 compared to saline control showed no indication of long-term toxicity from the conjugate, evidenced by no significant changes in animal weight over time, and no significant differences in blood chemistry. Future work will assess the impact of antigen expression on contrast. To that end, patient-derived xenograft models were screened for expression of MUC16, and three models were identified that express low, moderate, and high levels of MUC16 ( Figure 4 ). Our results demonstrate that huAR9.6-IRDye800 binds to MUC16, preferentially accumulates and identifies tumors with high tumor to background ratios, and does not exhibit long-term toxicity. Abstract Body : Introduction: Small cell lung cancer (SCLC) is a notoriously aggressive disease, presenting with a 5-year survival rate of only 7%1. Low-dose computed tomography, the gold-standard in lung cancer screening, is unable to detect early SCLC when interventions are most effective2. Thus, the identification of biomarkers for SCLC and the subsequent creation of targeted therapeutics remain unmet clinical needs. CD133 is a transmembrane protein with minimal expression in normal tissue but overexpression on the surface of a variety of tumors, including SCLC, making it a promising target for imaging and therapy3. Here, we report the development of a novel CD133-targeting radioimmunoconjugate ¾ [89Zr]Zr-DFO-PROM1 ¾ and its in vivo evaluation in a murine model of SCLC via PET imaging and ex vivo biodistribution studies. Methods: The CD133-targeting mAb PROM1 was modified with desferrioxamine (DFO) using p-SCN-Bn-DFO, and the resultant DFO-PROM1 immunoconjugate was subsequently labeled with [89Zr]Zr4+ under basic conditions (pH 7.4) and purified via size exclusion chromatography. The immunoreactivity of [89Zr]Zr-DFO-PROM1 was determined via an in vitro assay with CD133-expressing NCI-H82 human small cell lung cancer cells. To interrogate the performance of the radioimmunoconjugate in vivo, either [89Zr]Zr-DFO-PROM1 (90 µCi, 36 ug) or an isotype-control radioimmunoconjugate ([89Zr]Zr-DFO-hIgG1, 90 µCi, 36 ug) were administered via tail vein to 5-7 week old BALB/c female mice bearing subcutaneous NCI-H82 tumors. PET images were collected 24, 48, 72, 96, and 120 h post-injection, and the mice were sacrificed for ex vivo biodistribution analysis after the final imaging timepoint. Results: [89Zr]Zr-DFO-PROM1 was synthesized in >99% radiochemical yield and >99% radiochemical purity. The radioimmunoconjugate was found to have a post-purification specific activity of 11.1 GBq/µmol and an immunoreactive fraction of >70% with CD133-positive NCI-H82 SCLC cells. PET imaging and biodistribution experiments revealed the in vivo specificity of [89Zr]Zr-DFO-PROM1 for its target. Selective uptake of the radioimmunoconjugate was observed in CD133-positive NCI-H82 xenografts as early as 24 h post-injection. Over the course of 5 days, the activity concentration in the tumor grew to a maximum of 22.4 %ID/g ( ± 6.45 %ID/g), while the uptake in healthy compartments such as the blood, liver, and bone decreased or remained at relatively lower levels. Critically, only low levels of accretion of [89Zr]Zr-DFO-hIgG1 were observed in the tumor, reinforcing the specificity of the CD133-targeted radioimmunoconjugate. Conclusion: We have synthesized and characterized a novel CD133-targeting radioimmunoconjugate ¾ [89Zr]Zr-DFO-PROM1 ¾ and clearly demonstrated its efficacy for the non-invasive delineation of antigen-expressing tumor tissue in a murine model of SCLC. We are currently working to leverage the PROM1 platform for theranostic PET imaging as well as radioimmunotherapy. Zr-DFO-hIgG1 (90 µCi, 36 µg) to athymic nude mice bearing subcutaneous NCI-H82 SCLC xenografts (n = 4 for each antibody). Images on the left are coronal slices, whereas the last image on the right is a maximum intensity projection (MIP) collected at 120 h p.i. (B) Ex vivo biodistribution results. Following the final PET imaging timepoint at 120 h, the mice were sacrificed via CO2 asphyxiation. Selected organs were collected in pre-weighed tubes, and the activity in each tissue was measured using a gamma counter and subsequently normalized to the mass of each tissue. Statistical significance was determined using 2way ANOVA analysis with GraphPad Prism 8 (****p < 0 .0001). Abstract Body : Hypoxia, the diminished availability of molecular oxygen in bodily tissues is the hallmark of many solid tumors which often results in chemoresistance in different types of cancers. Over the last decade a variety of bioresponsive off-on molecular probes have been designed to specifically monitor changes in nitroreductase (NTR) and azoreductase enzymes that are overexpressed under hypoxic conditions. Although several probes have been shown to respond to these reductase enzymes, there is still a great need for fluorescent near infrared (NIR), water soluble probes with favorable pharmaceutical properties that enable clinical translation.

Here we report a novel off -on water soluble fluorogenic probe that indicates cell hypoxia. The probe is a heptamethine cyanine dye that is selectively triggered by NTR to undergo selfimmolative cleavage (1,6-elimination) to produce a large enhancement in NIR fluorescence signal at 780 nm. Additionally, the probe enables microscopic imaging of cell hypoxia with "turn on" fluorescence with no cytotoxicity effects. Michaelis Menten kinetic measurements, NTR blocking studies and colocalization studies are also reported. The probe has excellent photophysical properties and a chemical structure which can be easily manipulated to produce other NIR fluorescent probes that monitor changes in other reductase enzymes elevated under hypoxic conditions. Currently, ~75% of BCa surgeries in the USA are breast conserving surgery (BCS). Positive surgical margins, cancer tissue at the surface of a resected lumpectomy specimen, are found in 20-40% of all BCS resections. Among patients treated by BCS and radiation therapy, positive surgical margins are associated with a 2-fold increase in the risk of local recurrence when compared with negative margins1. Pathological examination is a current gold standard to confirm positive margins. However, microscopic evaluation of tissue samples only assesses about 1/10 of 1% of the entire volume of the removed specimen leading to margin undersampling and local recurrence of patients with pathologically clean margins, i.e. false negative margins. Innovation: Development of a simple real-time imaging technique to rapidly and precisely identify excisional BCa margins of the surgical wound would significantly improve the quality of care and reduce costs associated with incomplete excisions and consequent multiple visits. Exploiting increased protease expression at the edge of BCa, we propose in vivo topical administration of an imaging probe to identify BCa that may remain in the surgical cavity after standard-of-care resection. Methodology: Based on years of preclinical and clinical studies2-4 we have formulated a small molecule, fluorescently-quenched-protease-substrate probe, 6QC-ICG5, as a tool for imaging the surgical cavity after resection of tumor tissues. Different formulations of thermosensitive sol-gel Poloxamer-407, DMSO, and 6QC-ICG were tested to identify the best characteristics for topical application of the imaging probe. Results: Thermosensitive gel mixtures were utilized to allow easy fluid application and solidification once applied to prevent "dripping". Initially, to assess the viscosity and time-to-solidification of gel mixtures, more than forty formulations with different concentration of the gel and DMSO were tested. Results suggested that poloxamer-407/DMSO mixes (with ~20% concentration of each) possessed the best characteristics for "no drip" topical applications of 6QC-ICG. The formulations were tested using mouse animal models including: 1) orthotopic human MDA-MB-468 BCa xenografts in nude mice; or 2) human BCa xenografts in NSG PDX mice. We found that these mixes were good solvents for 6QC-ICG (>100-μM) with a desirable thermoreversible sol-gel transition and a weak mechanical strength that was important for distribution of 6QC-ICG on the surface of tissue. They demonstrated excellent ability to detect cancer tissue after 10min exposure that led to high signal-to-noise ratio (4.5) without pooling on the surface (Spectrum, Perkin-Elmer or Curadel-FLARE, Curadel, imaging cameras, 800-nm filter sets). H&E histology and fluorescence imaging (Odyssey, Li-Cor) of the tumor and surrounding normal tissue samples confirmed that Poloxamer-407/DMSO/6QC-ICG mix penetrated into the 120-μm in 10-min. To assess an efficacy of gel-carrier/prob formulation to detect cancerous tissue in vivo, half of tumor xenograft along with surrounding tissue was excised under anesthesia followed by 10-min topical application of the gel-carrier/probe mix. ICG fluorescent "hot-spots" in animals were inked to indicate the regions of fluorescence during histology. Correlation between the inked areas, i.e. fluorescence, in H&E slides, and the areas of pathologically confirmed cancer was determined. This analysis revealed a per-spot sensitivity=1.0 [95% CI 0.059-1.000] and specificity=0.933 [95% CI 0.765-0.988] for 24 spots in 12 nude mice that corroborated with a sensitivity=1.0 [95% CI 0.791-1.000] and specificity=0.941 [95% CI 0.692-0.997] for 36 spots in 10 PDX mice. Conclusion: Utilization of Poloxamer-407/DMSO/6QC-ICG mix topically to detect BCa in vivo in the surgical wound has the potential to reduce re-excisions as well as the false negative rate from pathology undersampling, with a consequent savings in healthcare costs and enhancement in patient quality of life. 

Abstract Body : Introduction Optogenetic functional magnetic resonance imaging (ofMRI) combines a precise neuronal stimulation technique, with fMRI as indirect readout of neuronal activation1. This enables a cell-type specific mapping of the whole brain dynamic response to the activation or inhibition of neuronal circuits. Despite the advantage of using fMRI to map functional changes of whole brain activity, molecular changes after optogenetic activation have mainly been measured using invasive techniques, such as fast-scan cyclic voltammetry or genetic fluorescence sensors. Although both techniques have been shown to have a high temporal resolution and specificity, they are restricted to a small, preselected region in the brain. In contrast, positron emission tomography (PET) offers the great potential to assess metabolic and molecular changes on whole brain level in vivo. While static [18F] FDG-PET provides no information on metabolic changes induced by a stimulation during the acquisition, functional PET (fPET) enables the detection of metabolic changes with high temporal resolution on single subject level via a continuous [18F] FDG infusion2. In this study, a simultaneous [18F] FDG-fPET/BOLD-fMRI protocol was established in rats subjected to optogenetic stimulation of the nigrostriatal pathway. Material&Methods Male rats (n=4, 329±24 g), were stereotactically injected with 2 µL of an adeno-associated virus vector overexpressing channelrhodopsin-2 into the right substantia nigra compacta (rSNc). Five to six weeks post-surgery, a [11C] methylphenidate PET scan was performed to control for neuronal cell death induced by protein overexpression. Ten weeks post-surgery, an optical fiber (diameter: 200 µm) connected to a 473 nm laser was implanted through a drilled hole into the rSNc. A constant infusion of αchloralose (20 mg/kg/h) and pancuronium bromide (1 mg/kg/h) was used as anesthesia. [18F] FDG-fPET/BOLD-fMRI scans were performed on a 7T small-animal MRI equipped with a PET insert. [18F] FDG (98±29 MBq) was infused using a bolus plus constant infusion protocol and fMRI data were simultaneously acquired using an EPI-BOLD sequence (TR: 2 s, TE: 18 ms). 20 Hz light stimulation protocols were used for optimized readouts. fMRI data were preprocessed and analyzed using SPM12. For PET data analysis, time activity curves (TACs) were generated and the difference between right (rSTR) and left striatum (lSTR) over time was calculated. To exclude temperature induced artifacts from optical stimulations, two non-injected control rats were stimulated and scanned using the protocol described above. Results [11C] Methylphenidate TACs and binding potential values were indistinguishable between the rSTR and lSTR, confirming that protein overexpression did not induce neuronal death. Optogenetic stimulation of the nigrostriatal pathways resulted in BOLD signal changes of 3.6±1.0% in the rSTR, while no changes were detected on the contralateral side. BOLD signal changes highly correlated to the stimulation paradigm and maximum t-values of 7.5±0.4 were observed in the dorsal rSTR on single animal level. [18F] FDG uptake in the rSTR gradually increased over time and the uptake pattern followed the stimulation paradigm. In the rSTR, an increased tracer uptake of 12.0±4.0% compared to the contralateral side was calculated in the last 5-minute time-interval of the PET acquisition. In contrast to the fMRI activation, metabolic activation was more pronounced in the medial part of the rSTR. Light stimulated, non-injected rats did not show BOLD signal changes or right-left differences in [18F] FDG uptake in the STR. Conclusion Here, we present for the first time simultaneous, dynamic [18F] FDG-fPET/BOLD-fMRI data during optogenetic stimulation in rats. We established a protocol to detect brain activation induced by light stimulation of nigrostriatal neurons projecting to the dorso-medial part of the STR. Our data show a clear spatial and temporal mismatch between activity patterns observed with BOLD-fMRI and [18F] FDG-fPET, which may be related to the complementary readouts of both. Further experiments are ongoing to increase the animal number and determine functional connectivity changes on group level. Abstract Body : Cerenkov luminescence (CL) is produced by radioisotopes predominantly when beta (β) particles travel faster than light in a given medium i.e. tissue. [1] The emitted CL is continuous, blue weighted, and proportional to the reciprocal wavelength. [2] CL has been characterized across the visible spectrum of light and into the near infrared region. The wide availability of highly sensitive visible light detectors such as IVIS based systems has allowed CL imaging to be established as a cheap alternative to PET, especially in the preclinical field. [3] However, the scattering and absorption of visible light is a major limitation of CL. Light above 650nm and into the NIR spectrum is absorbed an order of magnitude less than wavelengths in the 400 -600 nm regime. [4] SWIR wavelengths furthermore undergo minimal to no absorption or scattering and detection of CL in this spectrum would have significant benefits. [5] It should be further noted that the human eye does not detect light past ~700nm and LED bulbs are optimized for visible light emission with no emission in the SWIR spectrum. CL has previously been detected in the SWIR regime from medical linear accelerators (LINACs) where particles are accelerated to a magnitude higher energy than radioisotopes. [6] Proportionally they produce an order of magnitude more-light than that of radioisotope CL. [1] In this work we show for the first time that SWIR CL detection can also be performed with radioisotopes. SWIR CL is detected using an InGaAs based detector from five medical radioisotopes. The current setup's sensitivity (0.05 µCi/µl) is characterized along with and the SWIR CL radioisotope spectrum. Furthermore, the pre-clinical applicability of SWIR CL is shown using xenograft models. This work has implications for the rapid adoption of CL for both clinical screening and resection. The advantages of SWIR CL imaging for radioisotopes would include no need for ambient light removal along with reduced scattering, absorption and increased penetration depths in comparison to visible based CL. the SWIR setup of isotope imaging. Imaging was carried out in a dark enclosure with the radioactive source located within a lead ingot (pig) for shielding. The camera was fitted with a 16mm ƒ/1.4 SWIR lens and located ~20 cm from the source with a field of view of ~15x15 cm. Camera control and recordings were managed from the camera software on a laptop. Abstract Body : Introduction. Breast cancer is the most commonly diagnosed cancer for women and clear margins in breast-conserving surgery (BCS) are essential for preventing recurrence1,2. About 20% to 40% of BCS procedures result in margins, which are either positive or suspected of having malignant cells at the margins of the resection region3. The aim of this study is to develop an alternative diagnostic method, based on Nuclear Magnetic Relaxation (NMR) relaxometry and Fast field cycling (FFC) technique, able to identify the presence of tumour tissue in in small tissue samples cut from the surgical specimens, at the margins of tumour resection, before the histopathological analysis. Water proton's T1 relaxation, measured at very low magnetic fields, can be exploited to support the surgeon in real time margin assessment during BCS. Methods. Forty-one patients (aged between 40 and 90 years, mean age 65±15 years) undergoing quadrantectomy or mastectomy for breast cancer were enrolled in the study. From each surgical specimen between two and three samples (weight between 16-114 mg and with a diameter between 2.5-8.9 mm) were obtained, leading to a total of 104 freshly excised breast tissue samples. The individual tissue samples were analyzed by relaxometric methods and then transferred to the pathology laboratory for standard histopathological evaluation (H&E staining). The acquisition of proton longitudinal relaxation time (T1) as a function of the magnetic field strengths (the so-called Nuclear Magnetic Resonance Dispersion, NMRD, profile) was performed on the Stelar SpinMaster Fast Field Cycling (FFC)-NMR relaxometer (Stelar S.n.c., Mede (PV), Italy), endowed with a microcoil of 10 mm diameter. The NMRD profiles (7 points) were acquired at 10°C, in the range of low magnetic field strength from 0.01 to 1 MHz proton Larmor frequencies (corresponding to B0 = 0.24 mT-24 mT). The overall acquisition time was 20'34''. Results/Discussion. The histological analysis allowed to classify the investigated specimens as follows: 40 healthy (H), 21 tumours (T), and 43 mix (M, containing both tumor and healthy tissue), respectively. Both H and T tissues relaxation rates increase when the magnetic field strengths decrease but the relative values and slopes of the two curves are significantly different. This finding appears to be associated with the different water content and water mobility characteristics of the tissues. Tumour tissue has a protein/fat/water ratio that is highly altered in respect to healthy breast tissue in which adipocytes are predominant. As previously reported 4, 5, lipid protons relaxation rates are significatively less dependent than water/protein protons from the magnetic field strength in the range 0.02-10 MHz. Accordingly, we identified two relaxometric criteria, namely i) the slope of the line joining the R1 values measuread at 0.02 and 1 MHz and ii) the sum of the R1 values measured at 0.39 and 1 MHz, that reflect this peculiar behavior and can allow to assess the presence of tumour cells in a breast tissue specimen (sensitivity of 92% and specificity of 85%) (Fig.1 ). Conclusion. The relaxometric method proposed is low cost, fast and it does not rely on the expertise of a pathologist or cytologist as it can be highly authomatised. Its application can allow intraoperative surgical breast specimen margin evaluation by providing the surgeon real time information about the presence of residual tumour tissue. This can lead to a more precise surgical resection of breast cancer in BCS, reducing the need for re-excision. Acknowledgments. This work was performed in the frame of the COST Action AC15209 (EURELAX). This project has received funding from the European Union Horizon 2020 research and innovation program under grant agreement No 668119 (project "IDentIFY") and from AIRC under IG 2019, ID 23267 project (PI Geninatti Crich Simonetta). Abstract Body : Given that 1 in 8 women develop breast cancer throughout their lifetime, the disease is a substantial public health burden. Patient morbidity and mortality occur primarily due to metastatic spread but are compounded by thrombosis or the formation of blood clots. Recently, neutrophils activated by the inflammatory tumor microenvironment have been observed to release web-like DNA fibers known as neutrophil extracellular traps (NETs). NETs have been linked to several negative downstream effects including thrombosis, damage to vascular endothelial cells, and capture of circulating tumor cells to promote metastasis. Due to the pathogenic role of NETs in breast cancer, it is critical to ensure that contrast agents used in diagnostic imaging do not elicit this response. Breast MRI is often used as a supplement to mammography for high-risk women, as it detects breast cancers that may be missed by mammography and provides superior soft-tissue contrast; however, the standard MRI contrast agents used in the clinic (e.g. gadolinium chelates) cause high false-positive rates, as high as 25%. Thus, alternative contrast agents to improve diagnostic accuracy are being developed, including metal oxide nanoparticles such as iron oxide (Fe3O4) and manganese oxide (MnO). Prior studies have demonstrated that certain nanoparticle formulations may promote NETosis, which could encourage cancer progression. This study aimed to evaluate the effect of altered polymer coating of MnO nanoparticles on NET formation. MnO nanocrystals were synthesized using thermal decomposition of Mn (II) acetylacetonate in dibenzyl ether and oleylamine at 280°C. Nanocrystal size and chemistry were evaluated with TEM, XRD, and FTIR. Hydrophobic MnO nanocrystals were encapsulated with biodegradable poly(lactic-co-glycolic acid) (PLGA) and increasing amounts of polyethylene glycol (PEG) including 0%, 2.5%, 5%, and 10%. Hydrodynamic particle size was evaluated by DLS and confirmed to be ~180 nm. Neutrophils were isolated from the bone marrow of healthy female BALB/c mice femurs via density gradient. Neutrophils were stained with CellTracker Deep Red, plated, and stimulated for 3 hours with either media only (negative control), phorbol-12-myristate-13-acetate (PMA) (positive control), unencapsulated MnO nanoparticles, or PLGA MnO nanoparticles with 0-10% attached PEG. DNA was stained with Hoechst and plates were imaged with an inverted confocal fluorescence microscope to evaluate the total NET formation. Nikon General Analysis software was used to quantify the number of neutrophils and NETs per field of view. The greatest NET formation resulted from PMA stimulated neutrophils with an average of 35.83 NETs per 100 neutrophils. Unstimulated neutrophils caused the second-highest amount of NET formation at 19.28 NETs per 100 neutrophils. Bare MnO nanocrystals and PLGA MnO nanoparticles without PEG attached resulted in similar NETosis to unstimulated controls, whereas particles coated with a low PEG percentage provoked fewer NETosis than did stimulated neutrophils; 2.5% PEG PLGA nanoparticles had the smallest response of 11.06 NETs per 100 neutrophils. These results indicated that polymer-coated MnO nanoparticles do not enhance NETosis, supporting their safety as future breast cancer MRI contrast agents. Future studies will elucidate the mechanism responsible for the observed reduced NETotic response with nanoparticle PEG coating, including assessing nanoparticle stability, particle uptake into neutrophils, neutrophil activation, and reactive oxygen species as a result of nanoparticle exposure.

Last Name: Panchal Email: dp00016@mix.wvu.edu

Qing Liang, Actinium Pharmaceuticals, qliang@actiniumpharma.com

Abstract Body : Objectives: This work is investigates the feasibility of utilizing machine learning (ML) to predict the optimal level of I-131 injected activity which could allow reducing the number of time points required for dosimetry calculation for the SIERRA phase III clinical trial. Background: Personalized dosimetry in targeted radiopharmaceutical therapy provides a methodology to estimate the dose for each patient. Dosimetry can guide the amount of administered activity to patients maximizing the radiation dose to the tumor without causing toxicity in healthy organs. Currently, the methodology of the Committee on Medical Internal Radiation Dose (MIRD) to estimate the radiation dose involves imaging at multiple timepoints over a period of several days, which is not always feasible or practical. The SIERRA trial uses an iodine-131 labelled antibody (Apamistamab, Iomab-B) to deliver high doses of radiation to leukemia cells of older patients (55 years and older) with relapsed or refractory acute myeloid leukemia. This is done as a conditioning regimen for bone marrow transplant. Machine learning has become a powerful tool for predictive modelling within healthcare and nuclear medicine imaging. The availability of demographic, anatomic structural, and lab test information within the context of SIERRA opens the possibility to improve dosimetric prediction using machine learning. Methods: 74 patients who received the Iomab-B treatment were retrospectively analyzed in this work. The therapeutic infused activity was personalized based on an initial dosimetric dose with a low activity (7-20 mCi) Iomab-B infusion followed by 3 gamma camera imaging acquisitions performed immediately after infusion is complete, at 24 h, and at 72-96 h post infusion. The biodistribution information at the different imaging timepoints was used to calculate the dose to the liver following the MIRD formalism. 212 features (organ uptake and volume from planar and CT images respectively, demographic information, blood, and marrow tests) were evaluated. Features with greater than 20% of entries missing were discarded. Other remaining missing values were replaced with the median. Interquartile range of features was scaled using the RobustScaler preprocessing method using Python's scikit-learn library. A novel feature selection pipeline was designed and implemented using Python's mlxtend library (SequentialFeatureSelector method). The LASSO linear regression method (α=0.3) was used to train and test the model (70/30 train/test split). The model was used in combination with a single timepoint image (first image) to predict the injected activity required to achieve a 24 Gy maximum dose to the liver. Results: Liver mass, initial liver uptake, serum aspartate aminotransferase value (pre and post dosimetric infusion), body surface area (BSA), serum lactase dehydrogenase, lymphocyte count, and neutrophil count were robust features observed to account for variations in prescribed activity. The machine learning model had r2 = 0.71 and RMSE=179.8 mCi (mean prescribed activity=716.4 mCi). Error of less than 30% was reported in 63.6% of patients, which represents the fraction of patients within the uncertainty threshold of optimized gamma-camera imaging. SIERRA-specific constraints (maximum manufactured activity is 1030 mCi) led to further reduced model error within the clinical setting (RMSE=167.4 mCi). Our ML model outperformed conventional single-variable methods, such as those based on body or liver mass (RMSE=337.2 mCi or 278.3 mCi, respectively). Furthermore, the current method appears to significantly improve upon the predictive capability of liver dosimetry using the single initial time point (RMSE=331.8 mCi). Conclusion: Our results suggest that with carefully selected features, it is feasible to train a ML model that reduces the number of images to predict a personalized I-131 injection for the SIERRA trial. 

Rüdiger Exner, University of Bath, R.M.Exner@bath.ac.uk

Abstract Body : In recent years, the synthesis and modification of functionalized indocyanine dyes incorporating cypate frameworks, denoted Cy5 and Cy7 dyes, has received a lot of academic and commercial interest. Fluorescent probes based on these are of particular interest in in vivo imaging, for instance in fluorescence guided surgery, because their photophysical properties allow for imaging in the near-infrared (NIR) window, ranging from ca. 650 to 1350 nm [1] and more recently for photoacoustic imaging investigations [REF] . In the far-red and NIR spectral range, overall attenuation in biological tissues reaches a local minimum, allowing for higher penetration depths of the emitted light. Most of the Cy5 and Cy7 dyes used in ongoing clinical studies bear sulfonates as peripheral substituents as these modifications are typically associated with higher tumor-to-background ratio in vivo, higher brightness due to reduced aggregation to protein surfaces and faster excretion of corresponding optical imaging probes from healthy tissues. [2] [3] [4] [5] Hereby we report on our investigations into the synthesis and spectroscopic characterization of a new family of sulfonated cyanine dyes incorporating carboxylic acids as design elements of choice for a straightforward conjugation to multiple functional molecules. These yielded access a variety of novel compounds equipped with multiple targeting vectors including moieties suitable for conjugation to biomolecules and/or ligands or prosthetic groups for radiolabeling with metal isotopes of relevance for nuclear medicine applications, such as 89Zr ions, which were fully characterized. Furthermore, the in vitro characteristics of the as-synthesized novel compounds were assessed in vitro in PC3, DU145 or LNCaP cell lines cultured under normoxic or hypoxic conditions, and the vector-driven cellular uptake was rationalized. We envisage that the general design principles, as well as new insights into the chemistry of these dyes, will allow for the practical, facile, and rapid realization of highly modular building blocks of interest in a range of imaging applications including in multimodality regimes.

Models by MRI Janet Morrow, University at Buffalo, SUNY, jmorrow@buffalo.edu

Abstract Body : Transition metal-based MRI contrast agents are of interest as alternatives to gadolinium-based contrast agents. High spin Fe(III) complexes have shown promising proton r1 relaxivity and enhanced contrast in mice MRI studies.1-3 A rational approach to increase the relaxivity of Fe(III) MRI probes is to connect multiple iron centers together in a rigid framework by using aromatic linkers. Such an approach produces self-assembled tetrahedral cages with sixcoordinate Fe(III) centers bound to three catechols each. Such cages have been used for chemical catalysis4-5 but their solution chemistry is not well-studied. 6 Our investigations confirm that that rigidly linked self-assembled Fe(III) centers are robust towards dissociation and produce strong T1 proton relaxation.7

The self-assembled structures produce four kinetically inert iron catechol centers. Tetrahedral Fe(III) cages (Fe4L6) remain intact in solution when challenged with phosphate, carbonate or rat serum and show inertness towards acid, unlike tris-catechol complexes that dissociate under similar conditions. Proton relaxivity at neutral pH in PBS buffer, 37 °C at 4.7 T is 6.5 mM-1s-1 per molecule. In the presence of human serum albumin (HSA), relaxivity increases to 14 mM-1s-1, consistent with strong binding to serum albumin. Variable temperature 17O NMR studies are consistent with the absence of rapidly exchanging inner-sphere water ligands. The efficient proton relaxation is attributed to second-sphere water interactions at the Fe(III) centers that are tightly connected in a slowly tumbling assembly. Characterization of the iron cage by EPR spectroscopy and NMR spectroscopy experiments of Ga(III) analogs provide additional support for the structure of the cage in solution. MRI studies in mice show that Fe4L6 produces enhanced contrast of the vasculature with a half-life of approximately 4 hours for clearance. The iron cage accumulates over a period of several minutes to several hours into different murine tumor models including subcutaneous tumors from murine colon cancer and from ovarian cancer cell lines. Contrast enhancement of the ovarian tumor upon injection with 0.025 mmol/kg iron cage is even greater than that with injection of covalently linked HSA-Gd(III) complexes. This accumulation is attributed in part to the enhanced permeability retention effect of the Fe4L6-bound HSA and to the active uptake of serum albumin into tumor tissue by transcytosis through the vasculature and enhanced uptake at the cellular level. These hypotheses are being addressed by tuning the binding affinity of the cage to serum albumin. Res. 1999, 32, 975-982. 5 

A single-sided magnetic particle imaging scanner for in vivo human imaging Alexey Tonyushkin, Oakland University, tonyushkin@oakland.edu

Abstract Body : Magnetic Particle Imaging (MPI) is an emerging molecular imaging modality that visualizes the distribution of magnetic nanoparticles (MNPs) with an extraordinary sensitivity providing high spatial and temporal resolutions [1] . MPI could address clinical needs for safe diagnostic and therapeutic applications such as cancer detection, cell tracking, and angiography. MNPs may be functionalized to serve as specific tumor markers, which would allow the MPI device to be used for in vivo screening of cancer [2] . In MPI, the continuous excitation of the MNP is provided by the external ac magnetic field with rf frequencies. The nonlinear magnetization response to the excitation is detected by a receive coil. The spatial encoding of the signal is produced by selective saturation of the MNPs with the superimposed magnetic field gradient. Scanning the field-free region (FFR) across the imaging volume senses the local concentration of MNPs that allows a tomographic reconstruction of the tracer distribution inside the volume. The two different FFR are field-free point (FFP) and field-free line (FFL). Scanners with the FFP geometries have a more robust hardware consisting of a smaller number of coils to generate the required field configuration; however, the FFL geometry provides signals that are a factor of ten higher and benefits from a more robust and time saving backprojection image-reconstruction technique [3] . Thus, the FFL scanners may be beneficial to the future clinical applications. To date, a few small-bore MPI systems have been developed, including the commercial preclinical systems; however, human-size MPI scanner has yet to be built. The major challenges are the requirements of high power and safety concern due to highfield gradient being scanned across large cylindrical bore. These cylindrical geometry constraints prevent scaling up the MPI hardware to accommodate full-size human subject. Single-sided geometry, on the other hand, has all the hardware located on one side providing an unrestricted imaging volume, which would allow the MPI device to be used locally for in vivo screening of pathologies and in interventional procedures [4] . Previous state-of-the-art single-sided MPI scanner had an FFP geometry [5] . The major drawback of any single-sided devices is inherently non-homogeneous magnetic fields and nonlinear gradient, which diminishes the performance of the scanner. Our single-sided prototype of the MPI scanner utilizes a unilateral geometry with an FFL that offers higher signal and more robust image reconstruction technique than the corresponding single-sided scanner with the FFP geometry. In our device, a strong excitation magnetic field is generated by a single elongated transmission coil and the MNP response is simultaneously detected by a surface gradiometer receive coil. Here, progress toward the development of a single-sided MPI scanner for in vivo human imaging is presented. In particular, the simulation's results of the 2D image reconstruction imply an achievable spatial resolution of 2 mm. The experimental results from the sensitivity studies show the highest to date MNP detection limit in a single-sided scanner of up to 100 ng of iron that is a factor of 50 better sensitivity than the estimated limit of small tumor detection.

References: [1] B. Gleich and J. Weizenecker, "Tomographic imaging using the nonlinear response of magnetic particles," Nature, vol. 435, no. 7046, pp. 1214 -1217 , Jun. 2005 H. Arami, et al., "Tomographic magnetic particle imaging of cancer targeted nanoparticles," Nanosc., vol. 9, no. 47, pp. 18723-18730, Dec. 2017. [3] J. Weizenecker, B. Gleich, and J. Borgert, "Magnetic particle imaging using a field free line," J. Phys. D, Appl. Phys., vol. 41, no. 10, Art. no. 105009, May 2008. [4] C. Kaethner, et al., "Asymmetric scanner design for interventional scenarios in magnetic particle imaging," IEEE Trans. Magn., vol. 51, no. 2, Art. no. 6501904, Feb. 2015. [5] K. Grafe, et al., "2D images recorded with a single-sided magnetic particle imaging scanner," IEEE Trans. Med. Imag., vol. 35, no. 4, pp. 1056 -1065 , Apr. 2016 Abstract Body : Very late antigen-4 (VLA-4) is a trans-membrane integrin protein that is highly expressed in aggressive forms of metastatic melanoma. This integrin plays a key role in tumor growth, angiogenesis and metastasis by facilitating cancer cell adhesion and migration. LLP2A, a small-molecule high affinity, peptidomimetic, itself does not inhibit cancer cell proliferation and survival, and therefore is an ideal candidate for imaging and delivery of therapeutic payloads. We previously demonstrated the utility of 177Lu-DOTA-LLP2A as a therapeutic agent against metastatic melanoma in mice in combination with immune checkpoint inhibitors, although tumor growth was only inhibited with no cures. We hypothesize that failure to diminish tumor burden after 177Lu-DOTA-LLP2A therapy was due to the rapid renal clearance of the dose from the blood and tumor thereby decreasing efficacy. To improve the pharmacokinetic profile, we designed and synthesized novel analogs of DOTAGA-LLP2A with a 4-(piodophenyl)butyric acid (pIBA) albumin binding moiety. We demonstrate the feasibility of this albumin binding strategy by comparing in vitro cell binding assays and in vivo biodistribution performance of 177Lu-DOTAGA-PEG4-LLP2A (1) to the albumin binding 177Lu-DOTAGA-pIBA-LLP2A (2) . In vitrocell binding assay results for 1 and 2 showed Kd of 0.4 and 1.75 nM and Bmax of 200 and 315 fmol/mg respectively. Preliminary in vivo biodistribution data for both tracers exhibited specific uptake in tumor, spleen, thymus and bone due to endogenous organ expression of VLA-4. Tumor uptake for 1 was highest at 1 h (~15 %ID/g) and 2 at 24 h (~27 %ID/g). Clearance from the tumor for 1 occurs at 24 h ( < 5 %ID/g) while 2 is retained for greater than 96 h (~10 %ID/g). Compound 2 also exhibited longer blood circulation time compared to 1. Collectively, the in vitro and in vivo data suggest that modifying LLP2A with an albumin binder (pIBA) offers similar VLA-4 receptor binding qualities as the non-albumin binding molecule and provides a remarkable improvement in renal clearance and tumor retention. Efficacy studies are underway. Non-small cell lung cancer (NSCLC) accounts for 85% of all diagnosed lung cancers and is treated primarily by surgical resection (1) . In order to preserve lung function, surgeons attempt to remove the minimum amount of lung tissue required, but intraoperative identification of NSCLC is difficult. Surgeons utilize tissue palpation and visual inspection to identify tumor margins leading to disease recurrence in approximately 40% of patients (2) . More specific methods for delineating tumor margins intraoperatively need to be developed to improve the rates of complete resections (3). Fluorescence-guided surgery (FGS) is an emerging technology for aiding surgeons in the specific detection of tumors in real-time using fluorophores targeted to overexpressed tumor markers (4) . Fluorophore development for in vivo imaging has largely focused on the synthesis and characterization of near-infrared (NIR) I fluorophores due to longer wavelength emitters exhibiting increased tissue penetration and decreased light scattering, but longer wavelength fluorophores have further increased signal-to-noise ratios compared to emitters in the visible spectrum due to lower tissue autofluorescence in the NIR-II window (5). This project focuses on the development of NIR-I/II fluorophores targeted to cytosolic phospholipase A2 (cPLA2) and choline kinase α (ChoKα), two enzymes overexpressed in NSCLC (6, 7) . DDAO-arachidonate is a cPLA2-activatable fluorophore. In the presence of cPLA2, DDAO is released from arachidonate and has a measurable fluorescence with maximum emission at 660 nm and extending into the NIR-I window. For this project, freshly resected canine (n=1) and human (n=5) lung tumor and normal lung tissues were obtained from surgeries for operable lung cancers and painted ex vivo with 20 nmol of DDAO-arachidonate. The canine tissue for this project contained baseline fluorescence from ChoKα-targeted JAS239 (700 and 800 nm emission -LI-COR Pearl), and two human tissues contained baseline fluorescence from a folate receptor-targeted fluorophore (800 nm emission -LI-COR Pearl). Tissues were incubated with DDAO-arachidonate, and the fluorescence determined using the LI-COR Pearl (700 nm emission) or PerkinElmer IVIS Spectrum (680 nm emission). Fluorescence in tumor tissue was then compared to normal to determine tumor-to-normal ratios (TNR). Tumor was considered distinguishable from normal tissue if the TNR was greater than 2:1. TNRs for human tissues ranged from just over 1.5:1 to over 5:1 with three of the tumors distinguishable from the normal tissues. TNR for the canine tissue increased from 1.48:1 to 2.17:1 following treatment with DDAO-arachidonate. These data suggest that tumor heterogeneity may be overcome in FGS by employing fluorophores targeted to different enzymes commonly overexpressed in lung cancers with differentiable emissions. To improve the TNR for intraoperative NSCLC imaging, we synthesized a novel family of NIR-II fluorophores targeted to ChoKα by choline mimetic groups. These fluorophores are constitutively active with Stokes shifts over 40 nm and maximum emissions between 980 nm and 1020 nm. The fluorescence tails of these probes extend well past 1200 nm. In this project, we have shown that DDAO-arachidonate is a promising fluorophore for lung cancer imaging in human and canine patients, and we have developed prototype ChoKα-targeted probes with emissions in the NIR-II window making them strong candidates for in vivo lung cancer imaging. and normal lung tissue from the same patient were sliced into four sections each. All tissue slices were thoroughly rinsed in DPBS to remove blood from the tissue surface. Slices were then placed in 1 mL each of phenolred free DMEM in a 24-well clear plate. Slices were then treated with 50 μL of 400 μM DDAO-A (+ DDAO-AA) or 50 μL DPBS (-DDAO-AA). Slices were incubated at 37°C in 5% CO2 for 3 hr. Tissue was rinsed with DPBS twice prior to imaging using the IVIS Spectrum with 640 nm excitation and measuring 680 nm emission. Fluorescence was quantified using Living Image for two trials of the experiment, one using the neuroendocrine lung tumor, and one using a NSCLC, also donated by the Singhal lab. Fluorescence in the tumor tissue was significantly higher than in the normal tissue with an associated tumor-to-background ratio of approximately 4.9:1. *p-value < 0 .05 (unpaired t-test) Abstract Body : Introduction Chemical exchange saturation transfer (CEST) MRI is sensitive to many of the biomolecules in the central nervous system (CNS) altered in multiple sclerosis (MS) that cannot be detected using conventional MRI. Changes in CEST signal generated from two complementary methods, continuous wave (CW) and on resonance variable delay multiple pulse (onVDMP), were recently shown to occur during the progression of the mouse model experimental autoimmune encephalomyelitis (EAE)1,2. Here we investigated the cells and metabolites responsible for these CEST MRI signal changes in MS-associated lymph nodes using matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging, correlating the molecular information from MALDI imaging with CEST signal and conventional EAE metrics. Methods EAE: The model was induced in C57Bl/6 mice by injection of pertussis toxin (250 ng, i.p.) and MOG35-55 (300 μg, s.c.) in IFA supplemented with tuberculin (4 mg/ml) on days 0 and 2. Mice were assessed for model progression using a standard 0-5 disability scoring rubric. MRI: Mice were imaged 13-14 days after model induction (DPI) using a horizontal bore Biospec 11.7T scanner with a 2x2 phased array coil and 72-mm volume resonator at 0.14 mm resolution. For onVDMP CEST MRI, we used 32 pulse-exchange modules with 8 delays (mixing times from 1.14 to 100 ms; B1 of 46.8 μT). For CW CEST MRI, we used a saturation time of 3 s with a B1 of 1 μT. Maps of magnetization transfer asymmetry ratio (MTR) were investigated at frequencies 0.4-6.0±0.2 ppm upon B0 correction using the WASSR method. MALDI imaging: Metabolite data was obtained with Bruker Rapiflex MALDI TOF/TOF instrument at 0.1 mm resolution. Tissue was extracted 13-14 days post induction, flash frozen, cryo-sectioned (10 µm thick), and placed on indium tin oxide (ITO)-coated slides that were sprayed with 1,5-diaminonaphthalene (1,5-DAN, 10 mg/mL). Metabolite identification was confirmed using MS/MS fragmentation spectral analysis. Histology: Adjacent sections were stained with hematoxylin and eosin (H&E) for anatomical referencing of MALDI images. Flow cytometry: Cellular staining was quantified using a MACSQuant 10 flow cytometer and data was analyzed in FlowJo software using a standard gating scheme as reported before3. Results Changes in CW and onVDMP CEST MRI signal were observed in the CNS-draining superficial cervical lymph nodes (cLNs). These changes corresponded with active stages of model progression, wherein the paralysis is increasing (13) (14) . We compared CEST MRI signals to cellular composition in these lymph nodes, and found significant correlation to CD11b+/CD86+ antigen-presenting cells and CD4+/IL-17+ T-cells, both recognized progressors of this model. We then compared CEST MRI signals to metabolite composition in these lymph nodes. Of metabolites confirmed thus far, 5 correlated significantly to CW and onVDMP CEST MRI signal: alanine, lactate, leucine, malate, and phenylalanine. H&E staining revealed regional, i.e. cortex and medulla, differences in metabolite presence, e.g. leucine, within these lymph nodes ( Figure 1 ). Conclusions We linked changes in the CEST MRI signal of MS-associated lymph nodes to cellular and metabolic alterations using flow cytometry and MALDI imaging, a label-free approach for screening a wide-array of biomolecules directly from tissue sections. (13) (14) . Visualization of alanine, lactate, leucine, malate and phenylalanine maps in two representative cLNs each of (top row) control and (middle row) EAE mice. Bottom row: Average (N=5 each) relative signal intensity of these metabolites at their characteristic mass-to-charge ratio (m/z). Optical imaging is shown to highlight the location of the cLNs. H&E staining is shown to highlight intra-nodal regions (medulla and cortex). Arrows highlight a prominent example of these regional differences. Abstract Body : Purpose: Radioligand therapy (RLT) is a promising treatment option in metastatic castration resistant prostate cancer (mCRPC), although much of the focus has been on bone seeking radionuclides and PSMA-directed RLT. Herein, we evaluated if CUB domain containing protein 1 (CDCP1) can be exploited to treat mCRPC with RLT, particularly for subsets like small cell neuroendocrine prostate cancer (SCNC) that would not be expected to respond to bone seeking radionuclides or PSMA directed RLT. Experimental Design: CDCP1 mRNA levels were evaluated in the RNA-seq data from 119 recent mCRPC biopsies. Protein expression was assessed in twelve SCNC and adenocarcinoma patient derived xenografts. Saturation binding assays were performed with 4A06, a recombinant human antibody that targets the CDCP1 ectodomain. The feasibility of imaging and treating mCRPC in vivo was tested with 89Zr-4A06 and 177Lu-4A06. Results: CDCP1 mRNA expression was observed in over 90% of mCRPC biopsies, including SCNC and in adenocarcinoma with low FOLH1 (PSMA) levels. A modest anticorrelation was observed between CDCP1 and PTEN. Overall survival was not significantly different based on CDCP1 mRNA levels, regardless of PTEN status. Full length and/or cleaved CDCP1 was expressed in ten of twelve PDX samples . Bmax values of ~22,000 and ~6,200 fmol/mg were calculated for two human prostate cancer cell lines. Five prostate cancer models were readily detected in vivo with 89Zr-4A06. 177Lu-4A06 significantly suppressed the growth of DU145 tumors compared to control. Conclusions: The antitumor data and the overexpression of CDCP1 reported herein provide the first evidence promoting CDCP1 directed RLT as a treatment strategy for mCRPC. Combined with low expression in normal human tissues, these data provide a compelling scientific rationale for testing CDCP1 directed RLT clinically in mCRPC patients alone or in combination with other systemic therapies. Abstract Body : Cancer immunotherapies have had transformative success and durable clinical responses in patients with intractable malignancies. Immune checkpoint inhibitors (ICIs) have been effective in treating refractory solid tumors like non-small cell lung cancer and melanoma. Still, these treatments can have significant toxicities and there are many patients whose cancers do not respond. Thus, there is a clear need for translational tools that allow for early assessment of a favorable response.1-3 Preclinical evidence shows that immunologically active tumors, containing high levels of CD3+/CD8+ T cell infiltrates, correlate with favorable responses to immunotherapy.2,3 As a result, efforts have been made to image and monitor immune activity, particularly within the tumor microenvironment (TME). These efforts include the direct imaging of global T cell activation markers (eg. Granzyme B and ICOS), as indicators of T cell-mediated immune response.4,5 Still, none of these approaches have yet developed into widespread clinical application and each has a different temporal expression pattern relative to the cell type and immunogen. Our approach uses CD69, the canonical early-activation marker expressed in a variety of activated immune cells including cytotoxic T cells and NK cells, and applies this approach as a biomarker of response to cancer immunotherapy. Here, we show CD69-targeted PET imaging can characterize immune cell activation in the TME in response to checkpoint blockade therapy. The PET imaging probe was produced by conjugating an anti-mouse CD69 monoclonal antibody (H1.2F3) to deferoxamine (DFO) and labeling with Zr-89 ([89Zr]-DFO-H1.2F3). To test this probe in vitro, an uptake study was conducted comparing PMA/Ionomycinstimulated primary mouse T cells and an untreated control.

[89Zr]-DFO-H1.2F3 detected a 15fold increase in CD69 expression between stimulated and unstimulated T cells. To model an in vivo system, the probe was tested using an immunocompetent immune checkpoint inhibitor preclinical model.6 Adult, female, Balb/c mice bearing CT26 syngeneic tumors were injected with [89Zr]-DFO-H1.2F3 on day 15 post tumor inoculation, and were then imaged or sacrificed for organ harvest on day 18 post tumor inoculation, i.e. 72 hours post radiotracer injection. The in vivo PET/CT imaging study showed increased uptake in tumors from responder mice (18.0 ± 2.4 %ID/g, n = 8), relative to tumors from nonresponders and untreated controls (11.8 ± 0.4 %ID/g, n = 2; and 13.3 ± 0.6 %ID/g, n = 5, respectively). In a separate experiment, ex vivo biodistribution validation showed compelling increased tumor uptake in responders (43.1 ± 12.8 %ID/g; n = 6) relative to that of nonresponder and untreated control cohorts (10.3 ± 1.2 %ID/g; n = 4, and 12.6 ± 1.8 %ID/g; n = 5, respectively). The spleens of responder mice also had higher uptake (34.03 ± 2.66 %ID/g) than those of nonresponders and untreated controls (24.21 ± 0.52 %ID/g and 22.84 ± 2.80 %ID/g, respectively). Autoradiography corroborated biodistribution analyses showing increased uptake (%ID/mm2) for responders relative to nonresponders and untreated control mice. Using chromogenic detection methods, immunohistochemistry staining detected increased expression of CD69, OX40, Granzyme B, ICOS, and CD3 expression in tumor sections of responder mice relative to tumor sections from nonresponders and untreated controls. This CD69 PET imaging approach detected CD69 expression with sufficient sensitivity to quantify immune cell activation in a syngeneic mouse immunotherapy model, in which increased tumor uptake correlated with response.

Whee-Soo Kim, California Institute of Technology, wheesoo@caltech.edu Category: New Chemistry, Materials & Probes Abstract Body : A major goal of molecular imaging is to enable the visualization of specific molecular targets and tissue properties such as stiffness. Among molecular imaging modalities, ultrasound has significant advantages such as a high spatiotemporal resolution, low cost and portability. Recently, gas-filled proteins called gas vesicles (GV) were introduced as a new type of ultrasound contrast agent, which are thermodynamically stable and engineerable at the genetic level, allowing their multiplexed detection, functionalization and use as reporter genes [1, 2] . However, in some circumstances it is challenging to differentiate the ultrasound scattering produced by GVs from tissue background. To enhance the signal specificity of ultrasound, previous work has introduced magneto-motive ultrasound (MMUS) imaging [3, 4] , in which magnetic nanoparticles are moved by the applied magnetic field, resulting in unique time-locked ultrasound contrast. However, to produce this contrast, a high concentration of nanoparticles had to be used due to the poor echogenicity of magnetic nanomaterials. In this study, we combine the unique properties of GVs and magnetic nanoparticles to greatly enhance the sensitivity of MMUS and thereby enhance the contrast to noise ratio (CNR) of molecular imaging. Magnetic nanoparticle labeled GVs (MGVs) were constructed using click chemistry, resulting in approximately 186 magnetic nanoparticles attached to each GVs, and the hydrodynamic size was about 800 nm. After constructing the MGVs, MMUS imaging was conducted in hydrogel phantomes, demonstrating a strong signal uniquely produced by MGVs rather than either GVs or MNPs alone. Interestingly, we found that the MMUS signal was stronger in lower stiffness gels than at higher stiffness, which indicated that MGVs could be used as stiffness sensor. To further determine the effects of stiffness, several different concentrations of agarose and Matrigel were used. Lowering the stiffness in both materials led to stronger ultrasound signal. This result indicated that higher surrounding stiffness lowers the MGVs' movement by the magnetic force. We hypothesized that this stiffness dependence could be exploited to visualize biologically driven changes in tissue mechanics. To test this concept, we imaged an organoid model in which stiffness increases when fibrosis occurs. We found that MMUS signal decreased in the fibrosis model while in control organoids the signal was maintained. These results indicate that MGVs are a contrast agent with uniquely advantageous properties for MMUS imaging, which could be used as a tool for diagnostic and therapeutic applications. for patients with papillary thyroid cancer (PTC) remains controversial due to its variable riskbenefit ratio. Molecular fluorescence-guided imaging (MFGI) and quantitative spectroscopy can differentiate tumor from normal tissue after intravenous injection of a near-infrared fluorescent (NIRF) tracer and may help to improve detection of true negative nodal compartments in patients with PTC. However, target antigens for NIRF-tracers in PTC are unknown and no study previously reported about the feasibility MFGI and quantitative spectroscopy in PTC management. Objectives: 1) To identify a PTC-specific antigen targeted by a clinical available NIRF-tracer; 2) associate protein overexpression of the selected antigen with PTC tumor characteristics and locoregional recurrence rates 3) assess in vitro selective binding of the selected NIRF-tracer and 3) evaluate safety and efficacy of the selected NIRF-tracer for the first in-human detection of PTC nodal metastases (NM) using MFGI and quantitative spectroscopy. Methods: The highest-ranked overexpressed gene targeted by a clinically available NIRF-tracer1 was selected by class comparison between PTC and normal thyroid tissue using Functional Genomic mRNA profiling2 of publicly available microarray expression data. Target antigen protein expression was quantified using the H-score calculated from a tissue microarray with primary PTC (n=741) and normal thyroid tissue (n=108) that was immunohistochemically stained with 8189S, a c-Met targeting antibody. An H-score ≥ 150 was considered as positive. Staining intensities were correlated with tumor characteristics and tenyear locoregional recurrence free survival (LRFS). Fluorescence microscopy of a c-Met positive PTC (TPC1) and MET negative cell line (T47D) were performed to assess selective binding of the c-Met targeting NIRF-tracer EMI-137. A multicenter phase 1 dose-escalation study (NCT03470259) of patients receiving intravenous EMI-137 prior to surgery was conducted to evaluate 1) safety; 2) optimal dosage for the ex vivo detection of PTC NM using the IVIS Spectrum and IVIS Lumina II; 3) feasibility ex-vivo PTC NM detection with multidiameter single-fiber reflectance and single-fiber fluorescence (MDSFR/SFF) spectroscopy and 4) in vivo selective binding using fluorescence microscopy and immunohistochemical staining with 8189S of PTC NM and normal lymph nodes (NLN). A p-value Results: MET (ranked 11 of 1,702 overexpressed genes) was selected as the most promising antigen and is targeted by the NIRF-tracer EMI-137. A higher c-Met expression was observed in primary PTC compared to normal thyroid tissue (200.0 (IQR 150.0-258.0) versus 68.0 (IQR 20.0-100.0); p Patients with a positive c-Met expression (n=364) Abstract Body : Objectives: In vivo chemistry driven by bioorthogonal reactions has attracted considerable attention in the past decade, which has led to an exponential increase in pretargeted positron emission tomography (PET) studies. In turn, tetrazine based in vivo chemistry is a promising direction in pretargeted PET due to its instantaneous reaction rates and versatility.1 However, the success of pretargeted PET, in this case, relies on tetrazine radiolabeling, in which [18F] fluoroglycosylation has demonstrated great potential. It was previously reported that 5- [18F] [18F] FDG-Tz and [18F] FDR-Tz were performed in healthy mice (female Balb/cOlaHsd, 9-14 weeks and male Balb/cOlaHsd, 11-12 weeks, respectively). Anesthetized animals (0.4 L/min O2, induction 4-5%, maintenance 1.5-2.5% isoflurane) were sacrificed at 15, 30, and 60 min, respectively, after intravenous tracer administration. The number of animals per time point varied from 2 to 4 (15 min n=3/3, 30 min n=4/4, 60 min n=2/3, [18F] FDG-Tz / [18F] FDR-Tz). Glucose transporter 1 (GLUT1) binding affinities of FDG-Tz and FDR-Tz were determined as described previously.4 Results: [18F] FDR-Tz showed higher uptake in blood, heart, spleen, lungs, and skeletal muscle than [18F] FDG-Tz at 15 min post-injection. In the blood and heart, [18F] FDR-Tz exhibited almost 2 times higher radioactivity concentration than [18F] FDG-Tz, the average %ID/g in blood being 5.9 vs 2.9 and in the heart being 2.9 vs 1.6. These differences are statistically significant (P Half-maximal inhibitory concentration IC50 for FDG-Tz was 324.4 µM and for FDR-Tz was 22.4 µM. Higher GLUT1 affinity for FDR-Tz correlates with higher observed binding of radiolabeled tracer in organs with high GLUT1 expression. Conclusions: We can conclude, that from the two investigated 18F-labeled glycosylated tetrazines, the [18F] FDR conjugated candidate exhibited higher binding in the blood and heart, which can be explained by its higher affinity to GLUT1 transporter. 

Mustafa Utkur, Bilkent University, mustafa.utkur@bilkent.edu.tr

Abstract Body : Introduction: Magnetic Particle Imaging (MPI) is a tracer-based molecular imaging modality with applications such as angiography [1] , cancer imaging [2] , and catheter steering for cardiovascular interventions [3] . The superparamagnetic iron oxide nanoparticle (SPION) tracers suffer from relaxation effects that deteriorate their MPI signal response. On the other hand, by exploiting these relaxation effects, MPI's capability can be extended to detect functional parameters such as viscosity [4] , [5] or temperature [6] , [7] . Exiting techniques for temperature measurements during magnetic fluid hyperthermia (MFH) rely on invasive discrete temperature probes [8] . MPI, on the other hand, is noninvasive, compatible with MFH, and can further enable focused application of MFH in vivo [9] . A hybrid MPI-MFH configuration has the potential to enable cancer imaging, while substantially increasing the efficacy of thermal therapy via real-time temperature mapping. To map the temperature in vivo, however, the viscosity levels must be also taken into account due to the confounding effects of the two parameters on the MPI signal [10] . This work demonstrates proof-of-concept experiments of simultaneous estimation of temperature and viscosity in MPI. Experimental results for temperatures between 25-45°C and viscosities between 0.9-3.6 mPa· s show that our novel relaxation-based MPI technique can enable the simultaneous estimation of these two parameters. Materials and Methods: A total of 20 samples of a commercially available SPION (Perimag, Micromod GmbH) were prepared using water/glycerol mixtures and divided into four groups. For each group, a fixed viscosity level could be achieved at 5 different temperatures. An in-house magnetic particle spectrometer setup was used to acquire SPION responses under 60 different drive (i.e., excitation) field amplitude-frequency pairs ranging between 5-25 mT and 1-7 kHz for all 20 samples. The time-domain SPION responses were analyzed using our novel technique to estimate the effective relaxation time constant at each drive field amplitude-frequency pair. The estimated relaxation time constants were normalized by the period of the drive field at each frequency. This enables a fair comparison of the relaxation effect on the time-domain SPION responses at different frequencies. Results: The percentage changes in the normalized time constants with respect to 1°C change in temperature and 1 mPa· s change in viscosity are shown in Fig.1 . These plots demonstrate the sensitivity of MPI for temperature and viscosity estimation at different drive field settings. The sensitivity for temperature estimations in Fig.1 .A show a relatively mild change with respective to frequency and amplitude, with higher sensitivities seen at higher field amplitudes. The sensitivity for viscosity estimations in Fig.1 .B, on the other hand, show a dramatic change with respective to frequency, with the highest sensitivity observed at relatively low frequencies of around 1-2 kHz. At around 3-4 kHz, the sensitivity for viscosity estimations rapidly falls down and flattens out, making it possible to directly estimate the temperature at those frequencies, preferably using high field amplitudes. Therefore, by performing two measurements at two different drive field parameters (e.g., at a fixed field amplitude of 20 mT and at two different frequencies of 1 kHz and 4 kHz), both the viscosity and the temperature information of the environment can be simultaneously estimated. Conclusions: The benefits of simultaneously mapping temperature and viscosity make MPI a very promising modality for cancer imaging and treatment monitoring in combination with therapeutic hyperthermia. The novel relaxation-based MPI technique presented in this work utilizes the timedomain SPION signal and can be directly extended to estimate temperature and viscosity for multi-dimensional imaging in MPI. drive field frequencies ranging between 1-7 kHz with 1 kHz increments, and field amplitudes ranging between 5-25 mT with 2.5 mT increments. Each subplot in (A) corresponds to a distinct viscosity levels (left to right: 0.9 mPa· s, 1.8 mPa· s, 2.7 mPa· s, and 3.6 mPa· s), and each subplot in (B) corresponds to a distinct temperature (left to right: 25°C, 30°C, 35°C, 40°C, and 45°C). The normalized time constants with respect to viscosity show the highest slope (i.e., highest sensitivity) at lower frequencies of around 1-2 kHz, and falling down to almost zero beyond 3-4 kHz. According to the temperature plots, on the other hand, the highest sensitivity is achieved at lower frequencies and higher field amplitudes. As a result, due to the different trends in these sensitivity plots, the temperature and viscosity levels can be simultaneously estimated by performing two measurements at different drive field parameters. Abstract Body : Background: Neuroregenerative cell therapies of the central nervous system are hampered by high initial cell death and a host anti-graft immune response due to which >90% of transplanted cells can be lost within 1 week of transplantation1,2. We hypothesized that scaffolding cells in composite hyaluronic acid (HA)-based hydrogels of tunable mechanical strengths can significantly improve transplanted cell survival as evidenced by serial in vivo bioluminescence imaging (BLI). Concurrently we aim to quantitatively and directly assess hydrogel degradation post transplantation via chemical exchange saturation transfer (CEST) MRI by virtue of abundant high amide protons of the gelatin component of the hydrogels without necessitating any exogeneous labels. Methods: Allogeneic (mouse) GRPs (mGRPs) carrying proteolipid protein-green fluorescent protein (PLP-GFP) and transduced with a lentiviral vector carrying firefly luciferase (pLenti4-CMV-Luc) genes were derived from fetal forebrain tissue of (E13) transgenic mice. Composite hydrogels were formulated by mixing thiolated HA (HA-S), thiolated gelatin (Gel-S) and poly (ethylene glycol) diacrylate (PEGDA) crosslinker in a ratio of 2:2:1 at various concentrations and their rheological properties were tested. Based on optimal hydrogel stiffness, injectability and high cell viability, a 10 mg/ml formulation was chosen and 2x105 mGRPs were transplanted intracerebrally with/without hydrogel scaffolding into immunocompetent BALB/c mice brain striatum (Fig.1a) . MRI was performed using a 11.7T horizontal bore Bruker Avance scanner equipped with a 15 mm birdcage transmit/receive coil and a rapid acquisition and relaxation enhancement (RARE) pulse sequence. Imaging parameters were: slice thickness=1.5 mm, averages= 2, echo time=19 ms, repetition time= 1500 ms, matrix dimensions=128x128, field of view (FOV)=17x17 mm2 and RARE factor=16. For CEST imaging, a saturation power of B1 = 3.6 µT was used with -5.5 to 5.5 ppm offset range. MTRasym values were derived from fixed regions of interest (ROI) using MATLAB. Results: The survival of scaffolded vs. naked mGRPs transplanted in brain striatum was serially assessed by BLI up until 32 days post injection (Fig.1b) . The BLI signal gradually decreased over 32 days in both groups, but the decrease was significantly higher in the saline group postday 7 (p Conclusion: We demonstrated that hydrogel morphology and biodegradation can be concurrently monitored in vivo with T2-w MRI and CEST MRI respectively, with the hydrogel acting as a barrier for infiltrating host immune cells. In parallel, serial in vivo BLI revealed a 4.5fold enhancement in the survival of transplanted Luc+ allogeneic mGRPs in immunocompetent mice brain. Abstract Body : Delivery of radiation at ultrahigh dose rates has become known as FLASH radiotherapy, and this technique has recently been shown to preferentially spare normal tissues from radiation damage, while preserving as compared to tumor tissue. However, the underlying mechanisms of this phenomenon remain unclear at this time, but perhaps the most widely proposed hypothesis about the biological origin is that the effect stems from substantial oxygen depletion during the FLASH radiation pulse. This oxygen depletion would reduce the radiochemical damage during irradiation becuase the presence of oxygen is well known to fix and enhance radiation damage. This is particularly important in normal tissues, as they are highly oxygenated, and so the reduction in normal tissue damage is thought to be from this depletion process. The radiation responses tumor tissue is largely unchanged in FLASH, presumably due to the fact that the tissue is already nearly hypoxic. Testing of this hypothesis was teh goal of this work, by direct measurement of tissue oxygen in vivo during and after FLASH irradiation. Oxygen measurements were performed in vitro and in vivo using the phosphorescence quenching method and a water-soluble molecular probe Oxyphor 2P. Lifetime fitting of the phosphorescnece emission provides a direct prediction of the in situ oxygen level. The changes in oxygen per unit dose (g-values) were quantified in response to irradiation by 10 MeV electron beam with both ultra-high dose rates of 300 Gy/sec or with conventional radiation therapy rates of 0.1 Gy/sec. In vitro experiments with 5% BSA acqueous solutions resulted in gvalues for oxygen consumption of 0.19-0.21 mmHg/Gy (0.34-0.37 mM/Gy) for conventional irradiation and 0.16-0.17 mmHg/Gy (0.28-0.30 mM/Gy) for UHDR irradiation. In vivo, the total decrease in oxygen after a single fraction of 20 Gy FLASH irradiation was 2.3±0.3 mmHg in normal tissue and 1.0±0.2 mmHg in tumor tissue (p-value These molecular imaging observations suggest that oxygen depletion to radio-biologically relevant levels of hypoxia is unlikely to occur under FLASH irradiation, despite this being the leading hypothesis for the protective effect of FLASH. FLASH irradiation appears to induce less total oxygen consumption than conventional irradiation in vitro, which may be related to the FLASH sparing effect. This latter discovery was not expected but could agree with radical-mediated oxygen depletion effects. The difference in oxygen depletion between FLASH and conventional irradiation was not quantified in vivo, because measurements of oxygen depletion under conventional irradiation are hampered by the continuous resupply of oxygen from the blood. Abstract Body : Recently our lab introduced a sterically shielded heptamethine cyanine dye 775c (previously called s775z) that has no propensity for aggregation and highly photostable and showed that cyclic-RGD (Arg-Gly-Asp) bioconjugation using standard amide bond chemistry (775cRGD) selectively target integrin receptors for tumor accumulation in a A549 lung adenocarcinoma tumor-bearing mouse model.1As the tumor-targeting is mediated by the interaction of the targeting ligand and the macromolecular target, it is critical to assure minimum interference from cyanine conjugation for favorable binding affinity. In this study, the conjugation scope of 775c to cyclic-RGD has been extended using different linker chemistries to demonstrate effective non-invasive bioconjugation and compare targeted cancer imaging performance. Reactive duos 775mRGD and 775aRGD were developed using maleimide-thiol coupling and strain promoted azide-alkyne cycloaddition chemistries respectively. Spectroscopy data obtained in phosphate-buffered saline (pH 7.4) for 775cRGD, 775mRGD and 775aRGD at room temperature show similar photophysical properties and no aggregation. Non-specific protein binding analysis using a series of dye/bovine serum albumin associations found a Ka value of 2.7 x 104M-1 for 775c which increased to 4.0 -7.0 x 104M-1 upon cyclic RGD conjugation. This is about 20-40-fold lower serum albumin association compared to FDAapproved heptamethine cyanine analog, Indocyanine Green. Cell cytotoxicity studies using four different cancer cell lines (A549 lung adenocarcinoma, U87 human glioblastoma, HT1080 human fibro carcinoma, and CHO-K1 Chinese hamster ovary) showed no significant descend in cell viability for each peptide conjugate at low micromolar concentrations common for biological imaging studies. For cellular uptake studies, A549 lung adenocarcinoma cells which were previously reported for αvβ5-type integrin overexpression was exploited.2 A comparative set of fluorescence microscopy studies incubated separate A549 cell samples with each peptide conjugate of 775c for one hour and observed similar cellular uptake. In each case, targeted probe uptake was successfully blocked by pre-incubating the cells with free cyclic RGD peptide that indicates integrin selective binding and endocytosis. For in-vivo studies, an innovative pairedagent imaging strategy was followed where a binary mixture of targeted peptide and passively accumulating control fluorophore was injected into a single A549 subcutaneous tumor-bearing mouse model. A retro-orbital injection of each targeted peptide conjugate and the control (pentamethine cyanine dye 650c without cyclic RGD peptide, 10nmol of each probe) was given to nude mice (N=5) bearing a tumor in the right rear flank and imaged periodically over three hours. All three peptide conjugates showed high tumor-to-background ratios and slower washouts compared to the transient uptake of the control fluorophore over three hours suggesting integrin selective tumor accumulation. Ex-vivo tumor spheroid analysis revealed a heterogenic binding potential for each peptide conjugate and a significantly higher tumor accumulation of 775cRGD than remaining duo given to its inheritably strong amide bond nature. Subsequent biodistribution analysis of excised major mouse organs strongly indicated low background tissue retention and promoted renal clearance of the peptide conjugates at three hours indicating effective excretion of the fluorophores. These results strongly suggest effective cyclic RGD bioconjugation of 775c using three highly utilized linker chemistries, superior tumor targeting for diagnosis, and potential applications for cancer fluorescence-guided surgery. Abstract Body : Alzheimer's disease (AD) is a neurodegenerative disorder that closely associated with high concentrations of reactive oxygen species (ROS) in the brain. However, a powerful tool to monitor cerebral ROS concentration is not at hand. Previously, we have reported CRANAD-61, a fluorescence probe for detecting cerebral ROS at micro-and macro levels in brains of AD mice. However, the short emission wavelength (570 nm) of CRANAD-61 after ROS reaction hampers it's in vivo applications due to limited penetration depth and background interference. In the present study, we designed, synthesized and characterized an oxalatecurcumin-based probe CRANAD-122 that enables noninvasive dual wavelength near-infrared fluorescence (NIRF) imaging in vivo. CRANAD-122 showed an emission wavelength at 740 nm and 620 nm before and after reacting with ROS, which enables in vivo ratiometric quantification of cerebral ROS from the near-infrared dual-imaging channels. Real-time in vivo imaging studies with CRANAD-122 showed with higher signal ratio of short wavelength to long wavelength of transgenic mice compared with wild-type mice within 180 min, indicating fast brain uptake and excellent ROS-responsive capability of CRANAD-122. In summary, CRANAD-122 is a fast responsive, BBB-permeable, dual-wavelength imaging probe that could provide a simple, fast and effective tool to noninvasively monitor ROS levels in AD brains. Figure 1 . In vivo two-photon imaging with CRANAD-123 to identify active plaques and CAAs. a-c) Representative images of plaques and CAAs labelled by CRANAD-123 from red channel (a) and green channel (b) and merged image (c). Arrowhead indicates inactive plaques, in which red signals are dominated, while the yellow spots are active plaques, in which red and green signals both exist. d) quantification of plaques and CAAs through the ratio of intensities from red/green channels. It is clear that two distinct groups of plaques, and lower group is the active plaque group. While CAAs has no apparent separation. cg) Representative images of CAAs labelled by CRANAD-123 from red channel (e) and green channel (f) and merged image (g). Arrowhead indicates inactive CAAss, in which red signals are dominated. Arrow indicates active CAAs, in which the merged red/green into yellow color can be seen. h) quantification of CAAs through the ratio of intensities from red/green channels. It is clear that two distinct groups of CAAs, and lower ratio group is the active CAA group. Abstract Body : Oral administration is unequivocally the most accessible mode of drug delivery, requiring minimal expertise and invasiveness. However, this route is unable to accommodate the delivery of most biological therapeutics such as antibodies and peptides. In addition, it is challenging to target oral delivery to specific segments of the gastrointestinal (GI) tract, for example to treat inflammation or cancerous lesions. Existing approaches to enable this via the protection, imaging and targeting of the payload following intake are lacking in translational value as they compromise on the convenient formulation and dosing of ingestible drugs. Here, we will present a platform that leverages the simple addition of protein nanoparticles called gas vesicles (GVs) to equip oral drug delivery vehicles with in vivo tracking and release capabilities to safely transport biomolecular cargo. This is possible because GVs function simultaneously as in vivo ultrasound contrast agents1-3 and steric blockers of molecular diffusion that can be ultrasonically "erased" to trigger a rapid outflow diffusion of the payload from within the material (Fig. 1) . We demonstrate the utility of this approach in releasing protein therapeutic payloads in the GI tract of rodent models of experimental colitis. This platform has the potential to modernize existing oral drug delivery systems by enabling the convenient transportation of potent small molecules, biologics or gene therapy treatments to specific anatomical locations in complex tissues such as the gastrointestinal (GI) tract. Abstract Body : There is an emerging need to accurately identify both infectious and noninfectious bacteria noninvasively in the body to achieve multiple biomedical objectives. This can be challenging given the current lack of clinically available probes that can reliably distinguish bacteria from mammalian cells. Thus, a comprehensive understanding of bacterial physiology and genetics is required to develop probes for targeted imaging. Since bacteria are evolutionarily and phylogenetically distinct from mammalian cells, fundamental differences in metabolism and cellular structures can be leveraged to develop bacteria-specific imaging agents. Metal transport is a distinct pathway in bacteria that can be exploited to develop specific probes. For instance, Escherichia coli UTI89 utilizes the metallophore, yersiniabactin (YbT) to sequester Cu (II) from the extracellular environment and transport it inside specifically through the outer membrane protein receptor, FyuA (ferric yersiniabactin uptake A)1. We hypothesized that 64Culabeled YbT can facilitate targeted positron emission tomography (PET) imaging of FyuAexpressing bacteria in vivo. To optimize the radiolabeling process, we labeled YbT with four different transition radiometals -55Co, 64Cu, 68Ga, and 89Zr -in various physicochemical conditions. High performance liquid chromatography analysis revealed highest complexation of YbT with 64Cu. The presence of a single well-defined radio-peak with >95% purity at pH 7 and 37°C ( Figure 1A ) motivated us to choose these conditions for our probe preparation. For in vitro stability assessment, we incubated the probe in mouse serum for 0, 0.5, 1, 2, and 4 hrs, and confirmed the presence of at least 60% of intact 64Cu-YbT 4 hrs after the initial incubation ( Figure 1B ). When we administered 64Cu-YbT intravenously (i.v.) in mouse models harboring bacteria in their hind limb muscles, we observed distinctly higher PET signals from muscles infected with FyuA-expressing bacteria: E. coli UTI89, E. coli Nissle wild-type (WT) and Klebsiella pneumoniae compared to those infected with negative control bacteria: Staphylococcus aureus and FyuA-knockout mutant E. coli Nissle (ΔfyuA) (Figures 1, C -E) . We then proceeded to investigate whether 64Cu-YbT can be used to monitor antibiotic treatment outcomes. We administered 64Cu-YbT i.v. in mice following intramuscular injection of E. coli UTI89 and K. pneumoniae. Bioluminescence imaging of luciferase-expressing E. coli UTI89 indicated reduction in bacterial burden, while PET imaging revealed a corresponding reduction in signal intensity in mice that received two oral doses of ciprofloxacin, compared to untreated mice (Figures 1, F and G) . Finally, we aimed to track FyuA-expressing bacteria in diverse, yet clinically relevant in vivo environments. We used 64Cu-YbT to selectively locate K. pneumoniae in the lungs of mice that received intranasal administration of bacteria ( Figure 1H ). P. aeruginosa, which can also infect the lungs as an opportunistic pathogen but does not encode the FyuA transporter, served as the negative control for the study ( Figure 1I ). Furthermore, 64Cu-YbT allowed us to specifically identify probiotic E. coli Nissle, which is a popular bacterial template for various bioengineering applications, in two distinct hypoxic microenvironments -subcutaneous 4T1 tumor and ischemic heart (Figures 1, J-M) . For both models, we first performed i.v. administration of E. coli Nissle or PBS and rested the mice for three days for the bacteria to localize and proliferate in the respective hypoxic cores. Subsequent administration of 64Cu-YbT i.v. yielded significant PET signals only in mouse models that received E. coli Nissle. In conclusion, we have validated that 64Cu-YbT, a metal-metallophore complex, can be prepared with high purity and stability under physiological conditions within an hour. We have demonstrated how this single PET probe can potentially be used to identify specific bacterial populations, monitor antibiotic treatment outcomes, and track pathogenic and probiotic bacteria in diverse niches in vivo. Abstract Body : Background. Indocyanine green (ICG) fluorescence imaging has been used to detect many types of tumors during surgery. However, there are few studies on thymic masses, and the dose and time of ICG injection have not been optimized. We aimed to evaluate the optimal ICG injection dose and timing for detecting thymic masses during surgery. Method. Forty-nine consecutive patients diagnosed with thymic masses on preoperative computed tomography (CT) and scheduled to undergo thymic cystectomy or thymectomy were included. The patients were administered 1, 2, or 5 mg/kg of ICG at different times. Thymic masses were observed during and after surgery using a near-infrared fluorescence imaging system, and the fluorescence signal tumor-to-normal ratio (TNR) was analyzed. Results. Among the 49 patients, 14 patients with thymic cysts showed negative fluorescence signals, 33 patients with thymoma or thymic carcinoma showed positive fluorescence signals, and 2 patients showed insufficient fluorescence signals. The diagnosis of thymic masses based on chest CT was correct in 32 (65%) of 49 cases. However, the differential diagnosis of thymic masses based on NIR signals was correct in 47 of 49 cases (94%); these included 14 cases of thymic cysts (100 %) and 33 cases of thymomas or thymic carcinomas (94%). In addition, the TNR was not affected by the time or dose of ICG injection, histological type, stage, and tumor size. Conclusions. Low-dose intravenous injection of ICG at flexible times can detect thymic masses. In addition, thymic cysts could be distinguished from thymomas or thymic carcinomas during surgery. Hypoxia, one of the key microenvironment factors, is a common feature of many solid tumors, which is caused by the rapid proliferation of cancer cells. Hypoxia is closely related to tumor growth and metastasis, and is believed to be associated with resistance to chemotherapy and radiotherapy. Therefore, hypoxia has become an important target in cancer imaging. Visualization of the hypoxic tissue region and quantification of the level of hypoxia will be helpful to detect cancer at early stage and to monitor the therapeutic responses. Early hypoxia imaging probes are mostly single-emission molecules, whose fluorescence intensity is greatly affected by the local probe concentration and heterogeneous delivery efficiency. Current studies focus on ratiometric probes that have hypoxia sensitive and non-sensitive fluorescence emission peaks, and the ratio of the two can predict hypoxia quantitatively. However, most of these probes lack the ability to actively target cancer cells and require intratumoral injection, which greatly limits their application in small lesions and metastases. The goal of this study was to combine hypoxia sensitive fluorescent dye, nanocarrier, and tumor targeting molecule together to improve the efficiency and specificity of hypoxia imaging in breast cancer. A near-infrared nitroreductase (NTR) activated fluorescent dye Cy7-1 and a "always-on" fluorophore FITC were conjugated to polyamidoamine (PAMAM) dendrimers (PAMAM-FITC-Cy7-1) to allow for quantitative imaging. The conjugates were further modified with low molecular weight hyaluronic acid (LWHA) to form LWHA@PAMAM-FITC-Cy7-1 nanoparticles (NPs). LWHA was used to target CD44 that is overexpressed on the surface of breast cancer cells. The proposed NPs were negatively charged with size ~80nm. The NPs are stable in physiological environment and selectivevly release Cy7-1 in acidic environment, presumably due to the conformational change of PAMAM. In vitro result shows that ratio between Cy7-1 fluorescence and FITC fluorescence (ICy7-1/ IFITC) presents good linear relationship with NTR level and is unaffected by the NPs concentrations. 4T1 breast cancer cells were used to investigate the cellular uptake of the probe. Confocal imaging results show that LWHA@PAMAM-FITC-Cy7-1 NPs were internalized into cancer cells through CD44 mediated endocytosis, and accumulated in cytoplasm after entering the cells. To test the performance of LWHA@PAMAM-FITC-Cy7-1NPs in living cells, CoCl2 was used to induce hypoxia in 4T1 cell culture. Confocal imaging results show that ICy7-1 were significantly higher in hypoxia cells, while IFITC remained similar between hypoxia and normoxia cells. Ratio images ICy7-1/ IFITC were generated and allow for quantitative evaluation of NTR level. The ability of LWHA@PAMAM-FITC-Cy7-1 in hypoxia imaging in vivo was investigated in subcutaneous tumor model. When administered intravenously, fluorescence of Cy7-1 was observed in tumor region with very low background. Fluorescence of FITC was helpful for monitoring the dynamic change of NPs concentration within the tumor region. Unlike ICy7-1 and IFITC that fluctuated with time, ratio of ICy7-1/ IFITC remained stable for ~10 hours, which provides sufficient imaging window and quantitative assessment of tissue hypoxia. Overall, this study successfully developed a hypoxia imaging technique for breast cancer, which could serve as an in vivo imaging tool for assessing tumor microenvironment. This imaging technique could account for variability of NPs delivery and uptake, providing subjective evaluation of tumor hypoxia. Further studies are underway to investigate its ability to detect lymph node metastases and other micro-metastases lesions. Abstract Body : Objective: Monitoring and predicting patient-specific response to immunotherapy in triple negative breast cancer (TNBC) as there are variable clinical response rates. Using advanced positron emission tomography (PET) imaging to probe the tumor microenvironment for T-cell activation, we aim to predict early response to immunotherapy for in mouse models of TNBC tumors. Methods: Female Balb/c or C57/Bl6 mice bearing 4T1 (cohort 1) or E0771 (cohort 2) mammary carcinoma cell tumors, respectively, were administered paclitaxel (PTX; 10 mg/kg), immunotherapy [IMT, anti-PD1 (200 µg/mouse as a high dose or 66 µg/mouse as a low dose) + anti-CTLA4 (100 µg/mouse as a high dose or 33 µg/mouse as a low dose), PTX plus IMT, or vehicle (saline) intraperitoneally. Two mouse models were used as Balb/c mice are genetically biased to Th2 immune response and C57/Bl6 mice are biased to Th1 response. Treatments were given every 3 days and granzyme B specific (GZP) PET imaging (Tcell activation) was performed. Tumors were extracted for histological validation at endpoint and stained with etc, etc, etc. Standard uptake value (SUV) was quantified and heterogenity of SUV through histogram analysis were performed. Statistical analyses were performed with unpaired Mann-Whitney and t-tests. Results: In cohort 1, non-responders to treatment had a significantly higher tumor volume compared to responders starting on day 6 (p Discussion: Preliminary results reveal that chemotherapeutic treatment results in T cell activation when compared to treatment with immunotherapy alone. Non-invasive PET imaging of the tumor microenvironment can provide data on T cell activation and hypoxic response predicting response to combination immunotherapy and chemotherapy. Utilizing clinically available advanced imaging strategies to understand biologically distinct features of the TNBC tumor microenvironment can aid in personalizing anti-cancer therapies. Abstract Body : Introduction: Glioblastoma is a highly aggressive primary brain tumour with a poor long term survival outlook1. Currently, care for this type of cancer involves chemotherapy, radiation, and surgical intervention; however, survival is typically less than a year after diagnosis1, with a two year survival rate of 26-33%2. The use of novel immunotherapies, particularly in combination with the chemotherapy temozolomide (TMZ), shows promise and represents one of the first treatment advances in more than 10 years1,2.

Tumour associated macrophages (TAMs) are implicated in tumour growth and poor prognostic outcome3, and preferentially take up systemically-injected iron oxide contrast agents, making them an ideal target for imaging. Additionally, gadolinium (Gd) contrast agents extravasate into brain tissue when the blood brain barrier (BBB) is damaged. In this work, a novel Magnetic Resonance Fingerprinting (MRF) sequence4 capable of simultaneous T1, T2 and T2* relaxation measurements was used to make concentration maps of iron-labeled TAMs and a Gd agent. We used MRF and traditional T2-weighted MRI anatomical scans to assess survival and tumour immunodynamics in response to two immunotherapy combinations in a mouse orthotopic glioblastoma model. Methods: Fifteen C57BL/6 mice (N=5/group) were intracranially implanted with 5x104 glioblastoma cells. Mice were 1) untreated, 2) treated with anti-PD-1 (200μg/mouse/treatment), or 3) treated with a combination of anti-PD-1 and TMZ (25μg/kg/mouse/treatment). All mice received MRI brain scans twice weekly (pre and post contrast) until the tumours reached endpoint for a maximum of 6 weeks. Both weekly scans included anatomicals and MRF. Mice received an IV tail vein injection of 100μL SPIO Rhodamine B (Biopal) 24 hours before the post-contrast scan. During the post-contrast scan, mice received injections of MultiHance (gadobenate dimeglumine, Bracco) prior to MRF imaging. Upon termination, flow cytometry (FC) was performed on the brains and spleens to compare macrophage populations to the MRF data and investigate cell populations in the brain. Results: As expected, treated mice survived the longest. Initial data obtained from analysis of a MRF scan subset indicates variable macrophage uptake between mice (Figure 1 ). There was no immediately observable link between overall uptake and treatment group, with one mouse in group 2 demonstrating a moderate infiltration of cells, some mice from groups 1 and 3 showing high levels throughout the brain, and others having no discernible macrophages. The MRF image for the mouse from group 2 corresponds with the FC data, with 13.3% of the identified brain cells being macrophages. This is in contrast to the overall mean of 3.0%. The group 1 mouse also showed high levels of iron in the brain, with 9% of cells identified as macrophages. One mouse in group 3 had no visible macrophages with MRF and very few in the FC analysis. We are still analysing maps, particularly Gd maps. Further analysis is required to determine if a correlation exists between imaging and FC data. Early analysis indicates differences in the location of macrophages within the brain tissue at different time points, with later points having higher levels of putative TAMs, and treated mice with tumours having higher levels of macrophage and lymphocyte infiltration in tumours than untreated mice. There are deposits of iron in the ventricles of the brain at later time points, suggesting the early spread of tumour cells (assisted by TAMs) via the cerebral vasculature. Conclusions: To our knowledge, this is the first time MRF has been used in vivo to quantitatively image Gd and SPIO simultaneously. We are able to evaluate macrophage distributions and their longitudinal response to tumour growth using MRF. Further analysis is needed to determine the correlation between imaging and FC data and evaluate BBB damage indicated by Gd maps. Figure 1 -MRF maps of iron-labeled macrophages and FC analysis of macrophages in the brain. Macrophage migration into the brain varied significantly between mice and groups (full analysis still ongoing). In the MRF maps overlaid on the FSE anatomical in these sample mice, tumours are circled in green and colour scales indicate cellular density (cells/mm3). The group 3 mouse with a tumour had far more visible macrophages in the tumour compared to other groups, whereas the group 3 mouse with no tumour had almost no macrophages visible on the MRI or in the flow cytometry. The mouse from group 2 had macrophages visible throughout the brain, whereas the mouse from group 1 had some macrophages visible in the tumour and some in the rest of the brain (not all shown on the single slice). Abstract Body : Contrast-enhanced transcranial functional ultrasound imaging is a powerful technique for non-invasive visualization of neural activity. However, microbubbles and other commonly used ultrasound contrast agents are highly unstable in the bloodstream, necessitating repeated administration during a brief imaging session. Here, we describe nanoscale ultrasound reporters which circulate in mice for at least 2h, enabling prolonged imaging sessions with a single injection. These particles are based on gas vesicles (GVs), a unique class of air-filled protein nanostructure biosynthesized in certain aquatic microbes as a means to regulate buoyancy. GVs comprise a 2-nm thick protein shell encapsulating air which is in equilibrium with the surrounding media. Despite this inherent stability, intravenously injected GVs are rapidly sequestered by liver macrophages. We show that by coating the GV surface with a zwitterionic hydrogel, particle recognition and uptake pathways are disrupted, greatly enhancing circulation time. Acrylate groups are first biochemically attached to primary amines on the GV shell via N-hydroxysuccinimide chemistry. Next, these acrylate groups are used as anchors for radical polymerization with zwitterionic monomers and cross-linkers, forming a thin hydrogel on the GV surface. This process increases the hydrodynamic diameter of the particles from 235 nm to 348 nm, while reducing the zeta potential from -51 mV to -6 mV. The in vitro phagocytosis kinetics of these particles, as measured by flow cytometry, are significantly slower than that of unmodified particles. Upon intravenous administration, the coated GVs circulate with a half-life of approximately 1h, and share the same biodistribution as unmodified GVs. We believe such long-circulating ultrasound contrast agents to be of great practical utility in functional neuroimaging, and may be adapted for molecular detection and drug delivery applications as well. (Figure 1a ) that rapidly establishes a chemical equilibrium with HP 13CO2 in tissues.3 The detectable signals can be used to image tissue pH using the Henderson-Hasselbalch equation pH = pKa + logSBic/SCO2, where S hyperpolarized 13C NMR signal and pKa is known in vivo.2 Solid tumors commonly developed an acidic interstitial microenvironment with pH 6.5 -7.2, in contrast to pH 7.4 in normal tissues, owing to heterogeneous perfusion, high metabolic activity, and rapid cell proliferation. 4 We demonstrated this pH imaging method can distinguish between low-and high-grade prostate cancer in the TRAMP model of prostate cancer according to the level of extracellular acidification (Figure 1b and 1c) . These data suggest that HP 13C MRI could be used to discriminate between indolent and aggressive PCa in humans, motivating clinical translation of the method. Therefore, we have developed optimized methods for polarization and breakdown of 13C-GLC to form HP 13C bicarbonate in preparation for clinical translation. First, we used NMR to identify the optimal conditions for 13C-GLC hydrolysis through systematically varying temperature, amount of base, reaction time, and amount of acid for neutralization ( Figure 2 ). We then investigated the reaction kinetics to determine the reaction rate constants under these different conditions ( Figure 3) . These data showed GLC breaks down through an exponential kinetic profile of hydrolysis reaction with a time constant of 11 s catalyzed by 2.0 eq of NaOH at 60 °C (Figure 3b,3c) . Subsequently, we polarized a mixture of 13C-GLC and GE trityl radical at 0.8 K with the human clinical GE 5T SPINlab polarizer for ~5 hours, where the frozen HP sample is dissolved immediately by 130°C hot water. We replaced H2O with D2O for dissolution to avoid the dipolar-dipolar interaction raised from the intermediate species of 13C-GLC breakdown, which caused a profound loss of polarization during the breakdown and neutralization process (Figure 4 ). We found that the percentage of polarization was significantly increased to 40% in D2O in comparison to 1% in H2O ( Figure 5 , Table 1 ). We established a standard operating procedure, including cooling down, neutralization, removal of radical, and sterilization of the dissolution, to ensure the safety for patient injection ( Figure 6 ). We obtained a human injectable HP 13C-bicarbonate solution with a high polarization ~ 20% at 90 sec of dissolution and long T1 -128 sec at 1.4 T ( Table 2) .

Using a human 3T clinical MRI scanner, we demonstrated the excellent accuracy and precision of this pH imaging method in in vitro phantom studies (Figure 7) . We currently are characterizing it in vivo mouse models and investigating the toxicity in rats. Taken together, these data suggest that the high polarization and purity of the resulting hyperpolarized 13Cbicarbonate will enable the clinical translation of this method. Abstract Body : Introduction: Endometriosis is a gynecological disorder that commonly occurs in 10-15% of women of reproductive age and up to 30% of women suffering from infertility. This disease is defined by the presence of endometrial-like tissue (lesions) outside the uterus in the ovaries and peritoneum. It is accompanied by various symptoms such as cyclic menstrual pain, chronic pelvic pain, dyspareunia, and infertility which all negatively affect patient's quality of life [1] . Retrograde menstruation is thought to contribute to its etiology [2] though hormonal, inflammatory milieu, or immune dysfunction may cause the lesion deposition [3] [4] [5] . Since no symptom is pathognomonic for endometriosis, it is hard to determine whether pain or infertility are the results of this disorder. Even today the final diagnosis for the disease is only established at laparoscopy, which is a painful procedure for which anesthesia is required and multiple side effects (bleeding, infection, etc) are common. Most importantly, the stage of endometriosis on laparoscopy correlates poorly with the extent and severity of pain [1] . Therefore, there is an unmet clinical need for developing novel non-invasive methods for early/reliable detection of endometriosis. Magnetic resonance imaging (MRI) is a technique for diagnostic imaging before treatment planning [6] . However, with MRI different lesions like superficial or deep infiltrating lesions can be easily overlooked because of lack of adequate resolution [7] . The goal of this study was to investigate whether iron oxide nanoparticles can accumulate in endometriotic lesions in a mouse model of endometriosis. We expect to observe nanoparticle accumulation in the lesions largely due to the enhanced permeability and retention effect similar to that in tumors. We believe that this method can help to reduce the number of overlooked lesions by MRI. In future, these nanoparticles can also be used as a potential platform for the delivery of therapeutics to lesions. Methods: Dextran-coated iron oxide nanoparticles (NP, 32.6±4 nm, zeta potential: 27.5 mv) were synthesized with a co-precipitation method. NPs were labeled with a fluorescent dye (Cy5.5) to yield a multimodal MRI/optical imaging probe (Cy5.5-NP). Here, in vivo Pgrcre/+Rosa26mTmG/+ endometriosis mouse model was utilized. Endometriosis was induced by removing one uterine horn, mincing it in sterile PBS, and injecting back minced uterine tissues into the peritoneal cavity. This unique Pgrcre/+Rosa26mTmG/+ mouse model of induced endometriosis provides an opportunity to visualize the progesterone receptor (Pgr)positive cells which express green fluorescent protein GFP (mG). In contrast, the Pgr-negative cells express red fluorescent mTomato (mT), allowing accurate localization and visualization of the endometriotic lesions under fluorescent imaging. Cy5.5-NP (10 mg Fe/kg) was injected intravenously in mice with endometriotic lesions. The animals were imaged with fluorescence imaging (IVIS) pre-injection and 24 hours post-injection followed by ex vivo imaging. Results: Ex vivo images showed fluorescent signal from Cy5.5-NP which was co-localized with the GFP signal from the lesions. NP accumulation was not detected in the muscle tissue, which was used as control (Fig. 1) . We realize that GFP-expressing model is not ideal for in vivo imaging due to the weak signal and abundance of tissue autofluorescence in the green channel. In the future, we will utilize a recently developed model where endometriotic lesions express luciferase and can be detected by a significantly more sensitive bioluminescence imaging (BLI). Conclusion: In this work, we employed a dual-modal, MR/NIRF imaging probe to visualize endometriotic lesions in a mouse model of endometriosis by passive targeting. Cy5.5-NP accumulated and were visible with ex vivo fluorescence imaging 24 hours following administration. Future work is focused on performing MRI studies and exploring different ligands for a targeted delivery approach, which can enhance the specific accumulation of NPs in lesions as a potential theranostic probe. communication between immune and cancer cells for the first time. This system demonstrated antigen-dependent activation of reporter genes in immune cells when engaging CD19 on cancer cells, with minimal off-target activity. Continuing our work in animals, we plan to image this immune-cancer interaction via MRI, a clinically relevant modality, as well as co-engineer cells with a CD19-targeted CAR. We have also begun to broaden the antigens we can target with the highly modular SynNotch system, so that this system can be applied across various cancer cell types. Further development of this system will provide a broadly applicable tool for monitoring cell-based immunotherapies and beyond. We posit these tools will ultimately improve our ability to develop better cell therapies and to understand response/non-response and side effects in individual patients. Abstract Body : Introduction: Daratumumab (DARA) is an FDA approved high-affinity monoclonal antibody targeting CD38 that has shown promising therapeutic efficacy in double refractory multiple myeloma (MM) patients [1] . Despite the well-established clinical efficacy of DARA, not all heavily pretreated patients respond to single-agent DARA, and the majority of patients who initially respond eventually progress [2] . This may be due to upregulation of pathways that inhibit DARA-mediated tumor killing [3] . Antibody-drug conjugates (ADCs) combine the highly targeted tumor antigen recognition of antibodies with the cell killing properties of chemotherapy for effective internalization and processing of the drug. In this study, we evaluated the antitumor efficacy of DARA conjugated to the maytansine derivative, emtansine (DM1), linked via a non-cleavable bifunctional linker. The ADC was labelled with the near-infrared (NIR) fluorophore IRDye800CW (DARA-DM1-IR) to evaluate the biodistribution and pharmacokinetics of the ADC. We hypothesize: 1) the conjugation of DM1 will enhance tumor-killing efficacy of the native DARA and 2) the conjugation of IRDye800CW will not affect the biologic activity or tumor targeting ability of the ADC. Materials and Methods: Daratumumab (Janssen Biotech) was generously provided by the Siteman Cancer Center pharmacy. IRDye800CW (Li Cor) was conjugated following the manufacturer's protocol. Cell viability was evaluated with native and fluorescently labelled DARA (DARA-IR) and DARA ADC conjugates, respectively, in a high (MM.1S) and low (U266) CD38-expressing myeloma cell line. Live cell microscopy on MM.1S cells was performed to evaluate binding and internalization kinetics of DARA-IR and DARA-DM1-IR. In vivo fluorescence imaging was performed on fox chase severe combined immunodeficient beige mice bearing MM1.S subcutaneous tumors (n = 3-5/group) transfected with green fluorescence protein and luciferase. 100µg of DARA-IR and DARA-DM1-IR were administered intravenously via tail-vein and daily fluorescence imaging was performed using the IVIS Spectrum CT (IRDye800CW Ex./Em.: 745 nm/820 nm) (Perkin Elmer). Mean fluorescence intensity (MFI), defined as total counts/region of interest, was determined using ImageJ. Tumor-to-background ratio (TBR) was quantified using tumor MFI divided by the MFI of contralateral tissue. Animals were then euthanized and organs were resected and imaged on the IVIS Spectrum CT. Tissue-to-muscle ratio was quantified by using tissue MFI divided by the MFI of the muscle. Tumors from both mouse cohorts were flash frozen, sectioned and stained with Hoechst 33342 (Thermo Fisher). Fluorescent microscopy was performed using an Olympus BX51 epi-fluorescent microscope. Results and Discussion: After conjugation and purification, spectrophotometric analysis showed a degree of labelling of ~1.2 with a 3:1 molar dye to antibody ratio. CD38-dependent myeloma tumor cell killing was observed following incubation of DARA-DM1 (MM.1S IC50: 0.68±0.07 µg/mL; U266 IC50: 2.82±0.14 µg/mL) and conjugation of IRDye800CW to DARA-DM1 did not significantly affect its therapeutic efficacy (MM.1S IC50: 0.55±0.07 µg/mL; U266 IC50: 4.58±0.21 µg/mL).

MM.1S cells within 3 hours post-incubation. Tumor uptake of the fluorescently-labelled DARA conjugates increased with peak TBR achieved 9 days post-contrast (DARA-IR: 4.00±0.73; DARA-DM1-IR: 3.29±0.36) (Fig. 1A,B) . Clearance of DARA-IR and DARA-DM1-IR was primarily observed through the liver with no significant difference observed in tumoral uptake (Fig. 1C) . Histology of tumor sections following sacrifice demonstrated binding of both conjugates primarily on the tumor cell surface (Fig. 1D) . Conclusions: Here, we demonstrated the conjugation of DM1 improved the therapeutic window of DARA. Fluorescently labelling the ADC with IRDye800CW did not affect the biologic activity of the native antibody and allowed for evaluation of the DARA conjugates on a whole-body and cellular level in our mouse models. Future studies will focus on further in vitro and in vivo characterization of other promising anti-CD38 immunotherapies and ADCs in immunocompetent myeloma mouse models with differential CD38 expression. Abstract Body : Noninvasive reporting of neuronal activity would be invaluable for revealing how specific brain circuits control animal behaviors. Luminescent calcium biosensors based on luciferase enzymes should enable noninvasive activity reporting from deep locations, as luciferases operate without requiring excitation light and auto-bioluminescence is minimized compared to fluorescence in mammals. To date, a seiries of NanoLuc-based luminescent calcium biosensors have been developed [1] [2] [3] . However, the calcium responsivity of these reporters is limited, and NanoLuc substrates that efficiently cross the blood-brain barrier have not yet been identified. Here, we present our efforts to improve CaMBI responsiveness and in vivo brightness in the brain through protein and substrate engineering. We have identified additional sites in NanoLuc where insertion of calcium-binding domains confer calcium sensitivity to light production, and are optimizing the resulting constructs through linker and core mutagenesis. We have also identified a new NanoLuc substrate that produces superior bioluminescence from the brain. Together, these results will advance the future use of luminescent biosensors for calcium and other analytes in the brain. Abstract Body : Background: Exposure to young blood or young blood plasma (YBP) has been increasingly documented to be beneficial for alleviating age-related cognitive decline1-5, however, the neurofunctional change after treatment with YBP has not been well characterized in vivo. Positron emission tomography (PET) is an advanced molecular imaging modality and can noninvasively visualize molecular and functional changes involved in biological processes6. PET imaging with radiolabeled glucose analogue 2-[18F]-fluoro-2-deoxy-D-glucose (18F-FDG) has been extensively used to characterize the neural activity by measuring cerebral glucose metabolism7. In vivo PET imaging requires strict restriction of body movement and generally, is conducted under the anesthetized state by using anesthetic agents in animal studies. However, anesthetics are known to reduce neural activity and cerebral metabolism, thereby interfering with the interpretations of in vivo cerebral metabolic imaging data8. In addition, isoflurane which is an extensively used anesthetic during animal PET imaging has been reported to block synaptic plasticity in the hippocampus of mouse9. Objective: The purpose of the present study is to investigate the in vivo neurofunctional changes after YBP treatment in aged mice under conscious state, as well as to explore the mechanisms underlying the effects of YBP treatment on cognitive function in aged mice. Methods: Natural aging mice (aged 56 weeks) were systemically administered with YBP (extracted from 8-week mice) or PBS for 4 weeks via tail vein. One week prior to the PET imaging, the aged mice were trained twice to adapt to the PET imaging under conscious state, according to the previous method8. In vivo 18F-FDG PET imaging was performed in aged mice under conscious state after systemic treatment with YBP and the image data were analyzed using the SPM software. Novel object-location recognition test, novel partner recognition test, and spontaneous Y-maze test were performed to test the short-term recognition memory, long-term recognition memory, and spatial working memory of the aged mice, respectively. Western blot and immunofluorescent staining were conducted to assess the expression levels of brain-derived neurotrophic factor (BDNF), insulin-like growth factor-1 (IGF-1), IGF-1 receptor (IGF-1R), synapsin 1 (SYN1), postsynaptic density-95 (PSD95), or the alterations of AKT and ERK signaling pathways in brain tissues and in vitro cells. Results: YBP treatment exhibited positive effects on the short-and long-term recognition memory, but not on the spatial working memory of the aged mice. Compared with the PBS control group, YBP treatment group demonstrated higher glucose metabolism in the right hippocampus and the bilateral somatosensory cortex, and lower higher glucose metabolism in the left cerebrum and the right bed nucleus of the stria terminalis (both P < 0 .05, uncorrected). Ex vivo studies demonstrated that YBP treatment increased the activation of AKT signaling pathway (P < 0 .05 in both in the hippocampus and somatosensory cortex) and decreased the activation of the EKT signaling pathway (P < 0 .05, and P < 0 .01 in the hippocampus and somatosensory cortex, respectively). Moreover, YBP treatment significantly increased the expressions of BDNF, IGF-1, IGF-1R, and SYN1 in the hippocampus and somatosensory cortex (P < 0 .01, P < 0 .05, P < 0 .05, and P < 0 .05, respectively). This was further supported by the in vitro cell studies which demonstrated that young but not aged blood plasma upregulated the expressions of BDNF, IGF-1, IGF-1R, and SYN1 (P < 0 .01, P < 0 .01, P < 0 .05, and P < 0 .05, respectively). Conclusion: Our results indicate that YBP treatment triggers positive effects on short-and long-term recognition memory in aged mice, which is likely linked to the increased neuronal activity, upregulated cognition-related protein expression, and alterations of key neuronal signaling pathways in the hippocampus and somatosensory cortex. 18F-FDG PET imaging provides a new avenue for exploring the mechanisms underlying YBP treatment. Effects of YBP treatment on rCGM regional cerebral glucose metabolism in aged mice. YBP treatment group (n = 7) showed higher rCGM in the bilateral somatosensory cortices (B) and right hippocampus (C), but lower in the right bed nucleus of the stria terminalis (D, upper) and left cerebellum (D, lower) than PBS control group (n = 8) (P 2-[18F]-Fluoro-2-deoxy-D-glucose; PET, positron emission tomography. rCGM, regional cerebral glucose metabolism; YBP, young blood plasma; A, anterior; P, posterior; L, left; R, right; x, left (-) to right (+) direction; y, dorsal to ventral direction; z, anterior (+) to posterior (-) direction; z = 0 indicates the position of bregma according to the Mouse Brain Atlas. (E) Schematic illustrating the paradigm of the shortterm novel object-location recognition test. (G) YBP treatment group (n=7) demonstrate higher novel object-location preference ratio compared with PBS control group (n=8). (F) Schematic illustrating the paradigm of long-term novel partner recognition test. (G) YBP treatment group (n=7) demonstrate higher novel partner preference ratio compared with PBS control group (n=8). Data are presented as the mean ± SEM. *P < 0 .05; n.s., no significance; YBP, young blood plasma. [1, 2] . The initial efforts to develop 111In-D6-B483, a HD with covalently linked DOTA attached to the surface hydroxyl groups, indicated the potential for targeting orthotopic GBM in a mouse model but required further optimization for clinical use [3] . Current approaches for imaging of GBM rely on magnetic resonance imaging (MRI) to diagnose and grade the glioma status. 111In-D6-B483 has the potential to quantify the degree of TAM involvement that may correlate to the severity of the glioma. In addition, D6-B483 may be used for radiotherapy using 90Y in place of 111In. HDs have been observed to be retained in TAMs for up to one month after a single administration providing a local reservoir of radiation with limited systemic exposure (systemic clearance within 2 days). Methods: The synthesis of D6-B483 was achieved in 2 reaction steps. During the first synthetic step, partial propargylation of the dendrimer was achieved using sodium hydride and propargyl bromide. In the second step, azide-terminated DOTA was attached to propargyl dendrimer to yield D6-B483. For the synthesis of Cy5 labeled dendrimer, 1-2 propargyl functional groups on dendrimer were reacted with Cy5 azide to achieve Cy5-D6-B483. Cy5 labelled D6-B483 with either 2-3 or 8-10 DOTA (10 mg/kg) was administered IV to mice. Mice (3/timepoint) were sacrificed at 15 min, 4, 24, 48, and 96 h postdose. Amount of Cy5-D6-B483 was measured in kidney and liver after tissue homogenization and extraction [1] . D6-B483 was mixed with 111InCl3 heated to 85oC to yield a labeling efficiency of 73%. Using centrifuge membrane system (3kDa cut off), the product was buffer exchanged to remove free 111In and formulated for injection in PBS. The final product had a radiochemical purity of 96%. Twenty female mice were implanted with 106 GL-261-luc2 cells by stereotactic intracranial (IC) surgery. Brain tumor size and location was measured by bioluminescence (BLI) to confirm tumors were between 15 to 60 mm3 prior to dosing. A separate group of 8 mice were implanted subcutaneous (SC) with 106 GL-261-luc2 cells and dosed once tumors were between 125 to 350 mm3 (caliper measurements). All mice received a single IV dose of 111In-D6-B483 (~230 µCi, 45 µg) and SPECT/CT images were obtained at 3-6, 24, 48, 72 and 96 h post-dose. Within 24 h after terminal sacrifice, gamma counting of tissues was performed. Results: HDs have previously been demonstrated in animals and humans to be renally secreted with minimal liver uptake [4] . D6-B483 with 2-3 DOTA had less exposure to the liver than D6-B483 with 8-10 DOTA with comparable kidney exposure. Stability of the 111In-D6-B483 in mouse serum out to 4 days was greater than 90%. Bioluminescence imaging was used to enroll 12 IC tumor and 4 SC tumor mice into the study. SPECT/CT imaging showed relatively high uptake of 111In-D6-B483 in the IC and SC tumors (approximately 15%ID/g and 8% ID/g respectively) compared with approximately 1%ID/g in contralateral brain. 111In-D6-B483 persisted in the tumors for at least 96 h consistent with the previous study [3] . Conclusion: 111In-D6-B483 is a targeted and highly selective SPECT tracer for non-invasive imaging of brain tumors enabling precise quantitation of HD uptake and corresponding TAM involvement. 111In-D6-B483 is currently being developed for a Phase 1 study in GBM and metastatic brain cancer patients. (1). The potential benefit of using this contrast mechanism was demonstrated in a variety of applications, including cancer detection and grading (2), stroke characterization, muscle metabolism imaging, and kidney disease monitoring (3). However, the qualitative nature of the typical CEST analysis metric, and the long acquisition times required, constitute a barrier for clinical translation. Moreover, the unique biophysical characteristics of each metabolite/compound of interest, mandate a careful and exhaustive design of the CEST-MRI acquisition protocols, which is further challenged by the large number of imaging parameters involved (4). Objective: The goal of this work was to develop an AI-based automatic method for designing rapid CEST-MRI acquisition protocols and simultaneously providing the associated quantitative reconstruction means. Methods: The CEST-MRI acquisition protocol was represented as a computational graph, allowing its optimization using auto-differentiation. The saturation block was realized using the analytical solution of the Bloch-McConnell equations (5) and the spin dynamics were calculated during excitation and relaxation, using the Bloch equations with a discrete-time state-space model (6) . Finally, the resulting MR signals were mapped into CEST quantitative parameters using a fully connected deep reconstruction network. The resulting method, termed here as AutoCEST, was used for inventing quantitative bio-marker imaging protocols, each suitable for a particular clinically relevant compound/metabolite. Specifically, rapid acquisition protocols were automatically generated for amide, amine, aliphatic relayed nuclear Overhauser enhancement (rNOE), phosphocreatine, and Iohexol. The resulting protocols were initially validated in-vitro at 9.4T, using a variety of imaging phantoms assembled from the above-mentioned compounds, and also compared to an unoptimized CEST MR-fingerprinting protocol (7) of similar length. Next, AutoCEST was used for inventing two in-vivo oriented acquisition protocols, aimed for obtaining quantitative biomarker maps of protein and lipid (via the rNOE compound) and semisolid macromolecule concentrations (via the MT pool). A healthy mouse and a tumorbearing mouse were then imaged at 7T and 9.4T using the automatically designed acquisition schedules. Results: The acquisition times required for the protocols generated by AutoCEST were 35s (Iohexol, amine, amide, MT), 49s (rNOE), and 71s (phosphocreatine). The quantitative reconstruction time for a 64x64 image slice was 15 ms, using a laptop equipped with 8 CPU cores. The output phantom compound concentration maps were in good agreement with the known solute concentration (RMSE for Iohexol and phosphocreatine: 2.63 mM, Pearson's correlation r>0.97 between amide, amine, and rNOE with BSA concentrations). The resulting exchange rates were in agreement with known literature values and pH. The comparative MR-fingerprinting protocol demonstrated a poor quantification ability, highlighting the importance of automatic optimization. The reconstructed in-vivo semisolid lipid/macromolecule MT volume fractions were lower in the cortex compared to a white-matter region of interest (ROI), as expected. The in-vivo rNOE volume fraction was significantly lower at the tumor compared to the contralateral ROI (two-tailed t-test, p Conclusions: AutoCEST can automatically generate rapid CEST/MT acquisition protocols that can be rapidly reconstructed into quantitative parameter maps, potentially serving as a clinically useful molecular imaging technique. Abstract Body : Background & Rationale: Nearly 1 in 5 high school students and 1 in 20 middle school students reported using e-cigarettes (ECs) in the United States in 2020. It is well known that smoking can cause cardiovascular disease and thrombosis; however, few studies have investigated the adverse effects of EC use on the cardiovascular system, especially in the lungs. The public perception of ECs as a "safe" alternative to smoking and high rates of use, particularly amongst youth, demand that EC safety be rigorously studied in order to inform appropriate regulations. EC use has been linked to hypercoagulation and a persistent proinflammatory state. Neutrophils and platelets are known to undergo activation and proinflammatory phenotypic changes when exposed to EC vapor, and their aggregation promotes vascular occlusion and further inflammation. Thus, we hypothesize that platelet and neutrophil recruitment will be increased in mice exposed to EC vapor, and that higher rates of aggregate formation will be observed in the pulmonary vasculature of these subjects. Methods: To investigate lung vascular pathology following EC use, we have applied an innovative spinning disk confocal microscopy approach to visualize real-time interactions of blood cells in the lungs of live mice. Identifying neutrophil-platelet aggregates in the lung following EC vapor exposure would suggest that EC users may be at higher risk of atherosclerotic and thrombotic complications that could cause respiratory disease with chronic use. Mice (n=2) were subjected to 3 consecutive days of EC vapor inhalation using a Juul device with Virginia Classic Tobacco e-liquid (5% nicotine) for 3 hours each day for each topographical profile. 3 profiles were investigated: Profile 1-1 puff/min imaged 6 hours later; Profile 2-1 puff/2min imaged 1 day later; and Profile 3-1 puff/min imaged 1 day later. Mice subjected to saline injection via tail vein constituted the control group. Mice were anesthetized to facilitate microsurgery enabling fluorescence microscopy of the living lung 24 hours after completion of the inhalation course. Neutrophils and platelets were labeled with fluorescently tagged antibodies to compare overall presence and aggregation within the lung microvasculature ( Fig. 1) . At least 15 fields of view (FOVs) were collected for each mouse and quantified using Nikon General Analysis. Results & Conclusions: Preliminary data suggest significant increases in platelet (Profile 2 only; 2.22-fold), neutrophil (up to 2.35-fold), and neutrophil-platelet aggregates (Profile 1 and 2; up to 1.88-fold) presence in mice exposed to EC vapor compared to untreated mice. Collectively, this illustrates that EC vapor induces a pro-inflammatory state in the pulmonary vasculature in as little as one day after short term exposure. These findings will prove useful in future studies seeking to model the consequences of long-term EC use, which will ultimately enhance both clinician monitoring of patients at risk of such complications and public awareness of the potential consequences of EC use. Future work will investigate EC effects on platelet clotting pathways using platelet aggregometry and will make use of in vitro assays to explore constitutive phenotypical changes to platelets and neutrophil isolated from mice exposed to EC vapor. 

Le Xue, Zhejiang University, 11918441@zju.edu.cn

Abstract Body : Background: Frontal lobe epilepsy (FLE) is considered to be the second most common type of focal epilepsies of childhood. Focal cortical dysplasias (FCD), highly associated with medication resistant epilepsy, are the most common cause of focal epilepsy in children. In recent years, focal epilepsy is increasingly considered as a disorder of abnormal brain network organization and function that affects large-scale brain networks beyond the epileptic zone. However, to the best of our knowledge, no study has investigated the metabolic brain networks of FLE patients. Objective: The goal of this study was to investigate topological alterations of metabolic brain networks based on 18F-FDG PET in pediatric FLE patients with and without FCD using graph theoretical analysis. Methods: Thirty-three pediatric FLE patients were retrospectively enrolled and divided into the FCD group (n = 15) and the non-FCD group (n = 18). Twenty-five age-matched controls were also enrolled. PET brain images were acquired on a PET/CT scanner, using 3D whole-head acquisition. First, all PET images were preprocessed using SPM12, and the intensity of images was further globally normalized. Afterwards, the mean intensity values of 90 regions were extracted using an AAL90-based atlas for each subject. Metabolic connectivity matrix (90×90) for each group was constructed by calculating Pearson's correlation coefficient of the mean intensity values between pair of regions across subjects. From the connectivity matrix, graph theoretical measures were calculated to investigate topological alterations of metabolic brain networks. Global network measures included small-worldness, global efficiency, local efficiency, clustering coefficient and modularity, while regional network measures included nodal degree, nodal efficiency and betweenness centrality. Betweenness centrality was calculated to identify hubs of the networks. Non-parametric permutation test was performed to compare the statistical differences of the network measures between groups. Results: Compared with controls, both FCD and non-FCD groups showed reduced connections across the whole networks, but the FCD group was much more pronounced. Regarding global network measures, both the networks in patients and controls showed a small-world property. Relative to the non-FCD group, the FCD group showed significantly decreased average degree, global efficiency, local efficiency, clustering coefficient and modularity than controls (p < 0.05). In terms of modular structures, there were two modules in controls, less than three modules in the FCD and non-FCD groups. Regarding regional network measures, compared with controls, the FCD group showed a large number of regions with decreased nodal degree mainly in the frontal and temporal lobes, and a large number of regions with decreased nodal efficiency and increased betweenness centrality mainly in the frontal lobe and subcortical structures. In contrast, the non-FCD group showed a few regions with decreased nodal degree mainly in the frontal lobe, and a few regions with decreased nodal efficiency and increased betweenness centrality mainly in the frontal lobe. Besides, hubs with high betweenness centrality gradually decreased from controls (n = 18) to non-FCD (n = 14) group and finally to FCD (n = 11) group. Hubs in controls were mostly cerebral association regions and distributed over all four lobes. In the non-FCD group, hubs were also distributed over four lobes, but hubs in the frontal lobe were altered. In contrast, in the FCD group, the pattern of distribution of hubs was visibly disrupted, as the hubs were mainly concentrated in the frontal and temporal lobes. Conclusions: Pediatric FLE patients with FCD showed marked topological alterations in metabolic brain networks with network reconfiguration. These findings provide first evidence of altered topological organization of metabolic brain networks in pediatric FLE patients, with disruption of the coordination of cortical and subcortical structures. Abstract Body : Background. Myeloperoxidase (MPO) is a highly oxidative and proinflammatory enzyme that can generate reactive oxygen/nitrogen species and cause tissue damage if overexpressed. MPO is implicated in multiple inflammatory diseases, and its activity can be noninvasively detected by MPO-Gd, an activatable MRI probe. However, the linear chelate used in MPO-Gd raised safety concerns and its relatively small increase in relaxivity after activation limited its translational potential. In this study, we aimed to develop a more stable and efficient MRI probe to detect MPO activity. Methods. We designed and synthesized a highly efficient myeloperoxidase activatable MRI probe (heMAMP) that contains the macrocyclic DOTA as chelating moiety and two MPO-activatable 5-hydroxyindole moieties linked through a rigid amide bond (Fig. 1) . We validated the specificity and efficacy of heMAMP both in vitro and in vivo in rodent models of subcutaneous inflammation with wildtype and MPO-KO mice and unstable carotid atherosclerosis (tandem stenosis) with and without MPO inhibition (AZM198). We compared heMAMP with MPO-Gd and the conventional MRI agent DOTA-Gd. We also performed a molecular docking study to explore the mechanism underlying heMAMP's increased efficiency over MPO-Gd. Results. heMAMP demonstrated markedly higher Gd stability and greater relaxivity change after activation by MPO (Fig. 2) . In a mouse model of subcutaneous inflammation, heMAMP demonstrated a 2-3-fold increased contrast-to-noise ratio (CNR) compared to MPO-Gd, and 4-10 times higher CNR compared to conventional DOTA-Gd. When the dose of heMAMP was reduced from 0.3 mmol/kg to 0.1 mmol/kg, the signal increase was comparable to that of MPO-Gd at 0.3 mmol/kg while little CNR increase was observed in MPO-KO mice, confirming the in vivo specificity of heMAMP (Fig. 3) . The increased efficacy of heMAMP was further confirmed in vivo in a model of unstable atherosclerotic plaque where heMAMP showed a comparable signal change and responsiveness to MPO inhibition at 3-fold lower dosage compared to MPO-Gd. Molecular docking experiments revealed that a hydrogen bond formed intramolecularly between the two amide bonds which increases the rigidity of heMAMP and improved binding affinity of heMAMP to the active site of MPO compared to MPO-Gd. Conclusion: heMAMP is a potential translational candidate with superior stability and efficacy compared to MPO-Gd and other activatable analogues to detect MPO activity and inflammation. Figure 1 . Chemical structures of heMAMP and MPO-Gd. heMAMP contains the macrocyclic DOTA as chelating moiety which confers it more thermodynamically stable and two MPO-activatable 5-hydroxyindole moieties linked through a rigid amide bond. Figure 2 . The dynamic R1 ratios of heMAMP and MPO-Gd after activation by MPO. heMAMP demonstrated more efficient activation with >3 times higher R1 ratio compared with MPO-Gd when incubated with MPO, GOX and glucose over 180 min at 40 ℃ (n = 3, * p = 0.033, Kolmogorov-Smirnov test). Figure 3 . MR imaging from pre-contrast and at ~60 min for heMAMP, DOTA-Gd, MPO-Gd in wild type mice and heMAMP in MPO-KO mice in a mouse model of subcutaneous inflammation. heMAMP demonstrated a 2-3-fold increased contrast-tonoise ratio (CNR) compared to MPO-Gd, and 4-10 times higher CNR compared to conventional DOTA-Gd. Little CNR increase was observed in MPO-KO mice, confirming the in vivo specificity of heMAMP (n = 3 for each group).

Last Name: Wang Email: Wang.Cuihua@mgh.harvard.edu

Hunter Snoderly, West Virginia University, htsnoderly@mix.wvu.edu

Abstract Body : Background & Rationale: Breast cancer (BC) is the most diagnosed cancer amongst women, and metastasis drives the majority of deaths in these patients. The lung is a common metastatic site in all BC subtypes; thus, understanding how the pulmonary microenvironment fosters metastasis is of utmost importance. Crosstalk between neutrophils, platelets, neutrophil extracellular traps (NETs), and BC-associated extracellular vesicles (EVs) promotes the formation of the pre-metastatic niche. However, the exact cellular and molecular mechanisms mediating these interactions are not well understood. NETs form when neutrophils extrude their nuclei as web-like structures decorated with pro-inflammatory granular proteins. In cancer, NETs act as a scaffold allowing circulating tumor cells to escape the shear stresses of the vasculature and extravasate. NETs also enhance thrombosis, which itself facilitates metastasis. Increased platelet activation may also lead to enhanced NET formation as a positive feedback loop forms in which NETs activate platelets which then promote further neutrophil activation. A recent study found tumor EVs to be able to generate NET release; however, it was limited by examining EV-induced NETosis in an in vitro setting solely between neutrophils and cell-culture derived EVs. We hypothesize that platelets will be driven towards a prothrombotic state in the presence of tumor EVs, and that platelets and neutrophils will act synergistically to generate higher rates of NET formation through interaction with tumor EVs. Preliminary data show for the first time that these BC-associated EVs not only modulate NET formation but do so in a tumor progression-specific manner. BC-associated EVs isolated from tumorous mice with early metastases provoked the most NET formation. To investigate whether BC-associated EVs also modulate platelet phenotype, and if they can also indirectly modulate neutrophil behavior via platelet-dependent mechanisms, we will coculture isolated neutrophils and platelet rich plasma with or without BC-associated EVs. Methods: In this study, EVs were isolated via sequential centrifugation from the plasma of mice bearing orthotopic 4T1 murine mammary carcinomas 1, 2, 3, or 4 weeks post-tumor establishment. Neutrophils with or without platelets will be isolated from healthy mice and stimulated with BC-associated EVs ex vivo. EVs isolated from healthy mice will serve as controls. The potential of BC-associated plasma EVs to induce NETosis and/or platelet activation will be compared to 4T1 EVs derived from conditioned cell culture medium. The stimulatory effects of BC-associated EVs on platelets and neutrophils will be compared to our preliminary data in which neutrophils alone were stimulated with BC-associated EVs (Media alone or classical NET stimuli will serve as negative and positive controls, respectively. Confocal fluorescent microscopy will evaluate overall levels of NETosis and platelet/neutrophil association. Observed NETosis will be confirmed via ELISA assay detecting citrullinated histone 3 (a highly specific marker of NETosis) or via assays detecting DNA. Coagulation pathways by which EVs promote platelet activation will be revealed via platelet aggregometry. Finally, EV cargo and size characteristics will be determined using mass spectrometry and nanotracking analysis, respectively. Results & Conclusions: Preliminary results indicate that plasma EVs isolated from week 3 of tumor progression enhance NET formation significantly more than unstimulated neutrophils (media alone) or healthy plasma EV-stimulated neutrophils. All other tumor plasma EVs had higher rates of NET formation than unstimulated neutrophils or those stimulated with healthy EVs; however, this was not significant. We expect the addition of platelets will further enhance NETosis in the presence of BC-associated EVs in general. This research will enable the development of new treatments targeting the mechanisms by which tumors communicate with the rest of the body.

First Name: Hunter Last Name: Snoderly Email: htsnoderly@mix.wvu.edu

Hossein Najafiaghdam, University of California Berkeley, hossein_najafi@berkeley.edu

Abstract Body : Introduction: Identification of tumor cell depth with simplistic 2D imagers has tremendous applications for intraoperative imaging and in vivo imaging. For example, in vivo monitoring of the dynamics of immune response within tumors is of key therapeutic importance in cancer therapy. While novel fluorescence imaging instruments and advanced optical techniques such as light-sheet microscopy have been effective tools for extracting 3D information from ex vivo specimens, they are not practical for in vivo imaging. On the other hand, smaller compact devices such as silicon-based electronic microchip imagers [1, 2] , are able to achieve virtually seamless integration with surgical settings while still being minimallyinvasive. However, due to their smaller form-factor, they exhibit a poorer performance compared to their cumbersome counterparts. While there are some tradeoffs that allow improving the performance of these small sensors, many are significantly costly and some are often not surgically compatible or practical. The tradeoff between miniaturization and image quality, as well as the ability to derive depth information from 2D images, can be partially overcome with computational methods that compensate for non-idealities such as the impact of point-spreadfunction using various optimization techniques such as deconvolution. However, these techniques are, at their core, only linear methods and are blind and indifferent to the content of the images beyond the raw data. Machine learning and neural networks can incorporate and embed knowledge of the cells and perform content-aware optimization and thus have the potential to outperform conventional optimization algorithms. In this work, we present a module that leverages the processing power of neural networks to extract spatial information and resolve the 3D distribution of cells in 3D specimens, providing location maps of cell clusters at any given depth. We also introduce a second micro-imager to the system on the other side of the specimen (fixed to be 1 mm thick) and quantitatively demonstrate how adding an additional micro-imager for a more uniform coverage of the space contributes to the accuracy of the module. Methods: The module is a 3-layer convolutional neural network that was trained on a dataset of randomly generated cell images. The dataset was generated using perlin noise, thresholding and resizing to replicate real-life images obtained with the micro-imager [3] . Upon training, the network is then able to extract the spatial position of every cell cluster in the 3D specimen from the single or pair of input images (depending on the number of micro-imagers being utilized). We evaluate the module in both conditions and report the improvement in prediction accuracy with the additional sensor. The output cell cluster maps are provided in the form of layers, one for each discretized layer of the 3D specimen, where every pixel in the layer is classified as either "cell" (1) or "no cell" (0). Results: Using only one sensor, the module is able to achieve a detection accuracy of 70% in the nearest layer, decreasing to 60% in the farthest layer from the sensor. When using two sensors, the module achieved an 88% accuracy in the nearest layers which decreased to only 82% when moving to the middle layers (farthest from either sensor). A detection accuracy beyond 80% is practically high enough to make clinical assessments on changes in cell distribution. Conclusion: Here we presented a module that enables extraction of 3D information from 2D images obtained with miniaturized micro-imagers that can be seamlessly integrated with surgical instruments. The module performs content-aware extraction of cell cluster locations from 2D images and reconstructs the spatial distribution with minimal amount of data required, while achieving a detection accuracy of at least 82%, enough for clinical purposes. Abstract Body : Introduction: Fluorescence microscopy is critical in obtaining cellular-level information from targets inside the body. One key application is visualizing the tumor response to therapeutic procedures to help assess effectiveness of the therapy during the course of the oncological treatments. Real-time monitoring of multicellular-level changes provides key information regarding disease progression and immune response, which is of great importance in immunotherapy, a game-changing therapeutic unleashing the immune system to attack cancer [1] . While immunotherapy results in durable responses for patients who respond to the treatment, detailed cellular-level information from the immune response and tumor progression is essential to quickly pivot and personalize the therapy. Due to the low resolution (centimeter-scale) of the current oncological imaging modalities, the curative window for cancer is lost before a critical change in the tumor size is detected [2] . Moreover, invasive methods such as biopsies are impractical on a repeated basis. While state-of-the-art chip-scale image sensors provide high resolution microscopy, they can't be utilized in an implantable fashion or they require wiring and hence not practical for chronic real-time in vivo monitoring [3] . To address this problem, we propose a prototype for an implantable fluorescence microscope-on-chip detecting small clusters of fluorescence-labeled cells (via systemic injection of a fluorophore-labeled antibody) in realtime and wirelessly transmitting the data back to an external interrogator through an ultrasonic link. Miniaturization of the system is achieved by: (1) a fiber optic plate coupled with integrated angle selective gratings to eliminate bulky optical filters [4] , (2) a single mm-sized piezoceramic for wireless power and data transfer obviating the need for external wiring and batteries and (3) a mm-sized laser diode replacing optical fibers to provide illumination inside the patient's body. Methods: The fluorescence signal from a cell labeled with fluorophores is proportional to the optical properties of the fluorophores and the excitation power [5] . To obtain a detectable signal from close to 200 cells (~200 fW) with typical optical properties of the fluorophore, Cyanine5.5-NHS Dye (Cy5.5), a mm-sized laser diode with a 30 mW/cm2 excitation power at 635 nm is chosen and is controlled by the imager. 8 distinct illumination intervals ranging from 16 ms to 128 ms can be hardcoded into the implant. To supply the nominal electrical power (36.3 mW) of the excitation source from the ultrasound waves, the implant operates with a 0.5% duty-cycled illumination. The implant takes 20 s to power-up and store enough energy for a subsequent imaging cycle with the optical source turned on. Once a single image is captured, the data is quantized and streamed out for all of the 1440 pixels (36x40 pixel array). The implant will be tested for wireless backscattering using the piezoceramic transceiver in future experiments. We evaluate the resolution of the system with a USAF resolution test target. Results: To verify the performance of the device, a fluorescent Cyanine5.5-NHS dye, distributed over a resolution test target with a fine spatial structure, is imaged and the corresponding image data (11.5 kbits/frame) is steamed out. Interrogated with an AC signal, the imager captures the selected structures with

Yu Fu, Zhejiang University, 12031066@zju.edu.cn

Abstract Body : Background: 18F-FDG, as the most widely used PET imaging tracer, enables the visualization of detailed metabolic processes in human bodies due to its ability to capture both the biochemical and physiological processes. In practice, the PET imaging needs a full dose radioactive tracer to obtain satisfactory diagnostic accuracy, which raises concerns about the potential over dose radiation exposure especially for those pediatric patients. Hence, researchers have tried to reduce the tracer dose as much as possible for the PET scans to mitigate this negative effect. The quality of low-dose PET (L-PET) images, containing more noises and artifacts, is inevitably lower than that of full-dose PET (F-PET) images, which hardly meet the requirements to conduct successful diagnosis. In order to tackle these problems, deep learning models (DLMs) have been developed to synthesis high-quality F-PET images using low dose ones, but how the inner frameworks and image radiation doses affect the performances of DLMs have yet to be fully investigated. Methods: Totally 32 pediatric epilepsy subjects were included. All 3D brain PET images were acquired on a whole-body hybrid PET/MR system. All of the L-PET images were reconstructed by list-mode, which included doses ranged from 5% to 40%. A novel DLM, which consists of a spatial deformable aggregation module (SDAM) and an enhanced generative adversarial network (E-GAN), was proposed in this study to synthesize high-quality F-PET images from the corresponding L-PET images. This DLM can take full advantage of spatial information among the consecutive L-PET slices and preserve various semantic details, encouraging less blurring and more realistic synthesized results. For a better comparison, we also reproduced three recent state-of-the-art image synthesis models, include Pixel2Pixel, CycleGAN and pGAN. Objective: The aim of this study was to quantitatively and qualitatively show if our model has a better synthesis performance on commonly used metrics (peak signal-to-noise ratio (PSNR), structural similarity index measurement (SSIM) and difference maps) over these image synthesis models (Pixel2Pixel, CycleGAN and pGAN) . Results: Qualitatively, the synthesized F-PET images showed marked improvement in image quality compared with the L-PET images and visually resemble the ground truth image (true F-PET). From 5% dose to 40% dose, the PSNR and SSIM of synthesized F-PET images were gradually increased from 38.5 to 41.9 and 0.987 to 0.994, respectively. Leveraging the powerful SDAM and E-GAN, all the synthesized F-PET images derived from only 5% dose displayed delicate and precise textures. No matter for PSNR or SSIM, it's easy to note that our model significantly outperformed than other three models in all levels of low doses. When considering the whole dose range of 5% to 40%, the Pixel2Pixel, which with the PNSR of 28.6 to 31.4 and SSIM of 0.893 to 0.939, performed worst in all the four models. CycleGAN had the comparable performances with Pixel2Pixel on PSNR at the dose range of 5% to 15% (from 28.4 to 30.8) but showed stronger ability in the dose range of 20% to 40% in both PSNR (from 32.0 to 35.1) and SSIM (from 0.949 to 0.971). The pGAN, with the PSNR of 36.2 to 39.1 and SSIM of 0.979 to 0.989, was the model with best performance other than our model. Besides, synthesized F-PET images based on our model also showed the lowest mean differences relative to the true F-PET images. Conclusions: The proposed DLM can boost the synthesis performance of F-PET from corresponding L-PET images. Our model outperformed three recent state-of-the-art DLMs (Pixel2Pixel, CycleGAN and pGAN) with the highest PSNR and SSIM, and the lowest values in pixel-scale difference maps. Due to its high accuracy, the proposed DLM may show excellent prospects in future clinical applications.

Image/Figure Caption: Figure 1 . A: Our proposed DLM, which consists of a spatial deformable aggregation module (SDAM) and an enhanced generative adversarial network (E-GAN). These 2-D FDG PET slices, along with their corresponding adjacent L-PET slices, are first feed to the (a) Offset Prediction Network to generate deformable offset fields. These offset fields are then used as the kernels of the (b) Spatial Deformable Convolution to generate the fused feature maps, leading to the enhanced target L-PET slice. Based on the (c) Process of Discrimination and Generation, the final synthetic F-PET images are generated. B: The ground truth F-PET image and the corresponding synthesized F-PET images using Pixel2Pixel, CycleGAN, pGAN and our model based on 5% dose L-PET images. As shown in the enlarged drawing, it's easy to find that synthesized F-PET image of our model is the closest to the ground truth F-PET image and it retains more semantic details (metabolic information) than other models. C: The pixel-scale difference maps of these synthesized F-PET images relative to corresponding ground truth F-PET images across the dose range of 5% to 40%. Relative to other three synthesis models, our synthesized F-PET images have the smallest difference maps. D: The mean SSIM and PSNR of synthesized F-PET images based on our model. The SSIM and PSNR of synthesized F-PET images were gradually increased from 0.987 to 0.994 and from 38.5 to 41.9, respectively. 

Shensheng Zhao, Electrical and Computer Engineering, University of Illinois, Urbana-Champaign, sz35@illinois.edu

Abstract Body : Ultrasound localization microscopy (ULM) breaks the acoustic diffraction limit, allowing super-resolved imaging with a deep penetration1. It relies on locating the fluctuated microbubbles in blood vessels over time to 'paint' the image. Once combined with photoacoustic imaging (PAI), it expands the imaging capability to map not only the fine micro vasculatures and blood flow from ULM, but also the expression of biomarkers from PAI. However, the long data acquisition time restricts ULM in clinical applications. Although much effort has been made to accelerate ULM, such as compressed sensing-based ULM2 and deep learning-based ULM3, all these prior methods cannot allow data acquisition shorter than one second even without considering processing time, which insufficient for real-time clinical applications. In this research, we aim to develop an accelerated ULM (termed A-ULM) using deep neural networks (termed A-net) towards real-time applications beyond one second. With A-ULM, we aim to demonstrate a real-time dual-mode photoacoustic/ULM for in vivo mouse imaging. Our approach starts with A-net, which is trained to reconstruct subwavelength ultrasound images from power doppler images and sparse localization images. Then, randomized singular value-based clutter filtering4 and cross correlation-based localization method5 are integrated with A-net to enable real-time ULM imaging. Finally, simultaneous A-ULM and PAI are implemented on a dual ultrasound/photoacoustic imaging system. We first evaluated the performance of A-net in simulated blood vessels using a series of sparse localization data reconstructed from different numbers of frames. The simulation results reveal that A-net can indeed extract the detailed blood vessel information from power doppler images and sparse localization images. It can reconstruct super-resolved vascular images with high quality (i.e., the similarity index with the real super-resolved image is higher than 0.9) by using only 5% of the total localization points. In other words, it can reduce the data acquisition time by 95%. For example, it will take 500 milliseconds to acquire a ULM image compared to 10 seconds without A-net. To demonstrate the in vivo imaging capability, we constructed a dual-modal imaging system that accommodates the data acquisition, image processing, and co-registration of both A-ULM and PAI. Our imaging system allows a frame rate up to 0.67 Hz to display a reconstructed ULM image from A-net when imaging a mouse kidney. During this process, only 0.5 seconds of data acquisition and 1.5 seconds of data processing are needed for A-net to reconstruct the ULM image. In contrast, the conventional ULM using the localization method requires 4.5 seconds in data acquisition and around 5 minutes for processing to build the same kidney micro-vasculature map. This drastic reduction in both acquisition and data processing provided by A-ULM and fast pre-ULM data processing allows the implementation of towards real-time super-resolution imaging in vivo. Figure Caption: Fig 1: The implementation of towards real-time ULM/PA imaging using random singular value-based clutter filtering (Doppler processing) , cross correlationbased localization and A-net. (a) The demonstration of data acquisition and data processing. Each frame includes data acquisition for ultrasound (Doppler and ULM) and photoacoustic in hardware. Simultaneously, software experiences the data transfer of last frame, beamforming, doppler and localization processing, the inference of A-net and the image display. In in vivo mice kidney experiment, the data acquisition time for each frame is 0.5s, and the processing time is 1.5 s, indicating the frame rate is determined by computation time due to the synchronization of the hardware and software, which is around 0.67 Hz. (b) A-ULM vascular image from a mouse kidney. A-ULM can be obtained from the input of the power doppler image and the sparse ULM. Fusion of A-ULM and the ground truth shows A-ULM (red) has a high similarity with ULM (green) . (c) the improvement of image quality with increasing data acquisition time. Abstract Body : Because of the advantages in specificity, sensitivity, high throughput, and costeffectiveness, fluorescence imaging has been regarded as a powerful tool for biology and medicine. Combining with endoscopic technology, it can solve the problem of insufficient penetration depth of fluorescence imaging, and has been successfully applied to the clinical diagnosis of cancer, tissue hyperplasia and inflammation. However, most existing fluorescence imaging endoscopes can only obtain intensity information of specific wavelengths through filters, which leads to a high false positive and missed diagnosis rate in clinical applications. An effective solution for this problem is to collect complete spectral information in addition to the intensity information in the field of view. Therefore, we designed a fluorescence spectrum endoscope, which uses a beamsplitter to separate the collected signal light into the camera and the spectrometer to obtain complete spectrum information while performing fluorescence imaging. For endogenous fluorescence, we constructed mice tumor model, used autofluorescence imaging to quickly screen suspicious areas, and used autofluorescence spectrum to achieve qualitative analysis of lesions. In addition, for the exogenous fluorescence generated by the probe, we detected the spectrum shift due to changes in concentration, volume, and pH through the endoscope. The spectrum shift is short ( With the in vitro experiments, our endoscope detected a good mathematical mapping relationship between the spectrum shift and the variables(concentration, volume, and pH). In general, our fluorescence spectrum endoscope can simultaneously acquire the fluorescence images and continuous spectrum with high spectral resolution, and is expected to achieve qualitative and quantitative analysis in clinical diagnosis. Abstract Body : Background: Imaging technologies enabling whole-brain observation of specific neural signals represent a "holy grail" of molecular imaging for neuroscience. While the most widely used neuroimaging approaches with fluorescent genetically encoded calcium indicators (GECIs) have enabled major neuroscience discoveries1, they fall far short of providing simultaneous whole-brain imaging of neural signals due to the physical limitation of light scattering2. Alternatively, ultrasound has unique advantages as a modality for neural imaging due to its ability to penetrate much deeper than light (several cm) while providing relatively high spatial and temporal resolution (3,4, a unique class of air-filled protein nanostructures derived from buoyant microbes, have been developed as reporter genes for bacteria and mammalian cells5,6, and most recently as the first acoustic biosensors for enzymatic activity7. Building upon the design of the first GV-based biosensors, we engineered GVs as the first ultrasonic genetically encoded calcium indicators (UGECIs), enabling calcium imaging in mammalian cells with ultrasound in vitro, and potentially brain-wide calcium imaging in the future. Innovation: The key innovation of this abstract is the development of the first genetically encoded calcium indicators for ultrasound. These sensors are reversible and sensitive in physiological conditions. In addition, this work describes the first user of ultrasound to dynamically sense molecular signals inside mammalian cells. Method: Our basic design of UGECIs relies on the fact that GVs with a softer protein shell undergo larger "buckling" deformations in response to ultrasound pulses8-10 ((persuasive data, Fig. S1a) , resulting in the production of non-linear ultrasound signals, which can be selectively imaged with conventional ultrasound devices8,11. An alpha-helical protein called GvpC, bound to the GV shell surface, acts as a mechanical stiffener, controlling the extent of GV buckling and non-linear contrast9,11(persuasive data, Fig. S1b ). The GvpC was modified to incorporate a peptide domain recognized by a calcium-dependent binding protein (persuasive data, Fig. S1c ). Our hypothesis is that upon Ca2+-dependent binding between these two molecular elements, the engineered GvpC wound undergo an allosteric conformational change resulting in reduced GV shell stiffness, increased buckling, and the appearance of non-linear signal (persuasive data, Fig. S1d ). UGECIs were characterized for their dynamic range, reversibility, sensitivity and kinetics in purified form, and also expressed in HEK293T cells, where the intracellular calcium sensing was tested. Results: UGECIs, in the purified format, showed significant enhancement in non-linear contrast in response to Ca2+ in 37˚C, and the contrast was reversed by addition of calcium chelator EGTA (persuasive data, Fig. S1e-g) . This construct showed a half-maximal contrast at sub-micromolar concentration of Ca2+ in 37˚C in vitro (persuasive data, Fig. S1f ). Furthermore, UGECIs were expressed transiently in HEK293T cells and these cells showed calciumdependent ultrasound contrast. Control cells expressing UGECIs with disabled calcium binding domains did not respond to Ca2++ (persuasive data, Fig. S1h ). Impact: These results provided the first demonstration of genetically encoded calcium indicators for ultrasound, making it possible to conduct ultrasound calcium imaging on mammalian cells. Abstract Body : Background Repetitive mild traumatic brain injury (rTBI) is being steadily recognized to increase the risk of neurodegenerative diseases. However, the mechanisms underlying the adverse effects of rTBI are largely unknown. Preclinical studies in mice have shown alterations in prefrontal cortex-dependent functions and hippocampal-dependent learning impairments following rTBI1-3. In the majority of cases, structural damage following rTBI is not seen on T2-weighted MRI1, hampering diagnosis and clinical care. There is thus a need for more sensitive noninvasive techniques to detect rTBI-induced brain changes thereby potentially predicting long-term cognitive outcomes. Hyperpolarized 13C magnetic resonance spectroscopic imaging (HP 13C MRSI) is a recent imaging method that enables detection of metabolic changes in moderate TBI models4,5. Objective/Hypothesis We investigated whether 13C MRSI of HP [1-13C] pyruvate and [13C]urea could detect metabolic impairment in vivo in a rTBI model. Methods Animals and mild rTBI induction: Eight weeks old male C57BL/6J mice were assigned to a Sham (n=10) or rTBI (n=10) group. rTBI was induced as previously described using the Closed-Head Impact Model of Engineered Rotational Acceleration (CHIMERA) device1,6. Timeline is shown in Suppl.Figure1A. Behavior: Risk-taking behavior was evaluated using the Elevated Plus Maze (EPM). MRI: A 14.1T MR scanner was used to acquire T2-w MRI and HP 2D 13C CSI (parameters shown in Suppl. Table. 1). [1-13C] pyruvate and [13C]urea were copolarized for ~1h. 13C data were acquired 16 seconds post-intravenous injection. MR analysis: Brain volumes were calculated using T2-w data. The area-under-the-curve of HP [1-13C] pyruvate, [1-13C] lactate and [13C]urea, and lactate/pyruvate ratios were calculated for cortex, subcortex and blood. Spectrophotometric assays: Pyruvate dehydrogenase (PDH) activity was measured in cortex and subcortex. Statistical analyses: Statistical significance was evaluated using unpaired Student t-test. *p 7. Syndromic plots were performed using the R syndRomics package8 and PC naming using loadings >|0.45|. Results Mild rTBI increased risk-taking behavior, reflected by longer time spent in open arms and center of the EPM (Figure1.A, p=0.0102). However, no differences in brain volumes were found between Sham and rTBI (Figure1.B). In cortical voxels, no changes in HP [1-13C] lactate production was observed. In contrast, HP [1-13C] pyruvate and [1-13C] urea were increased in rTBI (Figure1.C, p=0.0101; p=0.006, respectively). The subsequent HP [1-13C] lactate/pyruvate ratio was lower in rTBI (p=0.001), as visualized on the corresponding heatmaps. PDH, the enzyme that converts pyruvate into acetyl-coA, was decreased in prefrontal cortex (p=0.0044) and cortex (p=0.0375) of rTBI (Figure1.D). Given the multidimensional nature of risk-taking behaviors, MRI, and enzymatic assays, we examined the relationships between the measured variables using linear PCA to identify sets of covariance that self-organized into functionally related groups based on their cross-correlations, which yielded PC1 (33.8% variance, Figure1.E). PC1 suggests the predictive nature of MRI for rTBI-induced chronic changes: decreases in lactate/pyruvate ratio correlates with increased risk-taking behavior. In subcortical voxels, HP metabolites were increased in rTBI, but the lactate/pyruvate ratio and PDH activity remained unchanged (Suppl.Figure1.B-C). Despite lack of group differences in subcortex, MRI's potential for predicting rTBI-induced changes were observed in this region (Figure1.E-yellow highlights). HP [13C]urea was increased in rTBI's blood and relationships with brain, enzymatic, and behavioral changes were also revealed in PC1. Conclusions rTBI induced long-lasting behavioral deficits invisible on T2w MRI. HP 13C MRSI detected cortical metabolic alterations characterized by decreased HP lactate/pyruvate ratios and PDH activity. Interestingly, HP pyruvate and urea were higher in rTBI, suggesting vascular, and/or blood-brain-barrier alterations9-11. Furthermore, PCA demonstrated the potential of HP 13C MRSI in predicting long-term rTBI-induced behavioral outcomes. Altogether, our results demonstrate that HP 13C MRSI can detect metabolic impairment following injury, and thus offer new possibilities for the detection of rTBIinduced damages in the clinical setting. Representative manual brain masking of T2-weighted MRI data (yellow: cortex, pink: hippocampus; green: ventricles; blue: subcortex) and corresponding volume sizes for Sham and rTBI. (C) T2 image and overlaid grid used for HP 13C MRSI analyses, red voxels indicate cortical areas and their corresponding HP 13C spectra for Sham and rTBI. Quantitative analyses of HP 13C (B) lactate, (C) pyruvate, (D) urea and (E) lactate/pyruvate for cortical areas. Representative HP metabolites heatmaps from a Sham and a rTBI mouse showing higher HP [1-13C] pyruvate and HP 13C urea levels in the entire brain of rTBI compared to Sham. Heatmaps of HP [1-13C] lactate/pyruvate ratio show lower values in cortical areas in rTBI. (D) PDH activity was decreased in cortical areas, including the prefrontal cortex and cortex, of rTBI compared to Sham. (E) Multidimensional analysis of risk-taking behaviors, MRI, and enzymatic alterations due to rTBI were examined using linear principal component (PC) analysis. PC1 accounts for 33.8% of the total variance and denotes relationships between risk-taking behaviors, MRI (brain and blood), and post-mortem enzymatic changes. Positive and negative correlations between the variables listed in PC1 are indicated by the arrows and corresponding numbers. As an example, decreases in lactate/pyruvate ratio correlates with increased risk-taking behavior. On a larger scale, we also observe that several HP 13C MRSI variables are predictive of behavioral outcomes (not just a single variable). (ICG) has been promoted to visualize of the tumor and its resection margins in clinical practice intraoperatively. Objective: Since there are no studies evaluating the ability of intravenous ICG to detect MLN, we aim to use preclinical and clinical studies to explore the feasibility of intravenous ICG to detect MLN and tumor. Patients: A total of 15 patients (13 with lung cancer and 2 with esophageal cancer) were enrolled this study. The patients were 7 males and 8 females, and the average age was 66 ± 10 years (49-84 years). In lung cancer, 11 cases were adenocarcinoma and 2 cases were squamous cell carcinoma, and the tumor diameter was 2.2 ± 1.0 cm (range 0.9-3.9 cm). According to the 8th edition of the TNM staging system, 5 cases were stage IA, 3 cases were stage IB, and 4 cases were stage IIB. In esophageal cancer, all patients were squamous cell carcinoma with stage IIIB, and the tumor diameters was 10 cm and 3.7 cm, respectively. Methods: To evaluate the ability of ICG to detect MLNs, we established four C57BL/6 mouse footpad tumor models using Lewis lung carcinomagreen fluorescent protein (LLC-GFP) cell line. ICG (5 mg/kg) was injected into the tail vein, and the distribution in tumor and lymph nodes (LNs) was detected by NIR fluorescence imaging system at 12 hours. In clinical studies, patients were administered with ICG (2 mg/kg) intravenously 12 hours before surgery, and the tumor and LNs were detected with the NIR fluorescence imaging system. The fluorescence signal intensities of the target region were measured using the fluorescent signal of tumor-to-normal ratio (TNR). Results: We evaluated the ability of ICG to detect MLN by comparing the TNR of MLN and normal lymph nodes (NLN). The TNR of tumor and MLNs was significantly higher than normal NLN in mice (p=0.0015 and p=0.038, respectively). In the fluorescence microscopy image, ICG is distributed in the metastatic site with GFP signal. In patients, a total of 268 LNs, ICG fluorescence signals in LNs have investigated 52 LNs which were selected based on preoperative CT and PET/CT LNs with suspected metastasis, as well as hilar and lobe-specific LNs. Within those 52 LNs, there were 33 LNs with no fluorescent signal confirmed no metastasis. 10 of the rest 19 LNs had metastases, 5 of which were not detected by CT or PET/CT. 9 of 19 LNs showed false positives which were consistent with the results of CT and PET/CT. The TNR of primary cancer and MLN was significantly higher than NLN (p=0.0001 hand p=0.0002, respectively), and there was no significant difference between PC and MLN. Conclusion: Intravenous injection of ICG can not only successfully detect tumors, but also MLN. This research can provide accurate information for tumors and MLN by fluorescence signals during surgery to achieve minimal resection. These findings offer the potential to prevent unnecessary resection by accurately identification from NLN and precisely resect the MLNs without missing them.

Image/ Figure Caption: Figure 1 . Detection of metastatic lymph node by NIR fluorescence imaging system with ICG, and the ex vivo study. Figure 2 . CT, PET, NIR fluorescence, and pathological images for each patient group. Abstract Body : Lipoproteins are natural nanostructures with essential roles in cell function, metabolism, and disease1,2. High-density lipoproteins (HDL), predominantly composed of phospholipids and apolipoprotein A1 (apoA-I), play a crucial role in regulating the immune system. For example, HDL mediates cholesterol efflux from macrophages3,4. Previous work has shown that radiolabeled reconstructed HDL can be used to noninvasively study macrophage-rich inflammatory lesions, such as tumors and atherosclerotic plaques5-7. Here, a radiolabeled apoA-I fragment was developed to monitor macrophage burden by PET imaging. The terminal cysteine of this apoA-I mimetic was functionalized with the chelator deferoxamine (DFO) and subsequently radiolabeled with 89Zr (89Zr-apoA-I mimetic, Fig 1A) . This compound's avidity for HDL was assessed in vitro by size exclusion chromatography (SEC) (Fig 1B) . Pharmacokinetics and biodistribution of the 89Zr-apoA-I mimetic were evaluated by PET imaging and ex vivo gamma counting in wild-type mice (Figs 1C-D) . Results revealed a half-life of 2.8 min (Fig 1C) and high uptake in liver, kidney, spleen, and bone marrow (Figs 1D and E) . We then evaluated 89Zr-apoA-I mimetic's properties in two different mouse models of inflammatory diseases: myocardial infarction (MI), and melanoma. (Figs 1F-M) . PET imaging 24 hours after 89Zr-apoA-I mimetic injection showed heterogeneous accumulation in the tumor (Fig 1F) . The mean maximum standardized uptake value (SUVmax) of the tumor was 2.57 ± 0.66, while the reference (muscle) had a SUVmax of 0.61 ± 0.04 (Fig 1G) . Imaging results were validated by ex vivo gamma counting (Fig 1H) . PET imaging at 24 hours on MI mice revealed an infarct accumulation of 89Zr-apoA-I mimetic, SUVmax of 2.66 ± 0.42 compared to 0.54 ± 0.04 in the reference (Figs 1I and J) . This specific infarct uptake was confirmed by gamma counting compared to the remote myocardium and the non-infarcted control mice (Fig 1K) . To investigate cellular avidity of the apoA-I mimetic, a non-radioactive apoA-I mimetic (90Zr-apoA-I mimetic) was injected to MI and melanoma mice and cytometry time of flight (CyTOF) was then performed on the tumor and infarcted myocardium, respectively. (Figs 1L and M) . We employed an extensive multicolor CyTOF panel that combined myeloid cell markers, including CD11b, Ly6G, and F4/80, as well as neutrophil marker Ly6G. The resulting high-dimensional dataset was visualized using the viSNE algorithm and indicated a Zr signal colocalized with F4/80 and CD11b positive cells, suggesting macrophage uptake. Combined, our results demonstrate 89Zr-apoA-I mimetic's applicability as an HDL-avid PET tracer that allows noninvasively monitoring macrophage burden in inflammatory disease. Abstract Body : Introduction: Preclinical PET imaging has become an important tool to noninvasively study the onset and progression of human diseases in small animals. Specifically, 18F-FDG is the most commonly used radiotracer to monitor tumor progression and transformation, and serves as a benchmark for evaluation of other imaging agents in mice. While clear, standardized protocols such as RECIST or EANM guidelines for tumor imaging have been created for clinical practice, protocols for preclinical imaging tend to vary widely, resulting in poor translatability of preclinical data (1) . It is well recognized that dietary state, ambient temperature and mode of anesthesia will affect glucose metabolism in the background organs of mice, hence affecting the reproducibility across subjects (2) . This study investigates the impact of the standardized imaging protocols on 18F-FDG biodistribution in healthy and tumor-bearing mice. In particular, the effect of fasting status on 18F-FDG tumor uptake is being explored. Methods: Six healthy and three MDA-MB-213-bearing mice (same strain, age and sex) underwent dynamic 18F-FDG PET at 40 min post-injection, before and after 16 h fasting. Mice were kept warmed and under isoflurane before 18F-FDG administration until the end of PET scans. Blood glucose was measured before and after PET. Images of 50 and 60 min postinjection were compared. Results: Using a standardized animal handling procedure throughout the PET scans i.e. warmed and asleep uptake under isoflurane), no distinct uptake differences were seen in the background organs (liver, lung and muscle) before and after 16 h fasting in healthy mice. When applying the same imaging protocol on tumor xenografts, the tumor demonstrated reasonably good uptake (2.46 ± 0.36 and 2.71 ± 0.42 %ID/g, respectively) at 60 min post-injection before and after 16 h fasting, with good tumor-to-muscle ratio observed regardless of dietary status (2.96 and 3.33, respectively) ( Figure 1 ). No significant differences in tumor uptake at 50 and 60 min post-injection after fasting, although a lower uptake on muscle was seen at later timepoint (0.92 ± 0.10 and 0.81 ± 0.21 %ID/g, respectively), giving rise to a slightly higher the tumor-to-muscle ratio (2.77 and 3.33, respectively) . Glucose levels were found to be varied widely across each mice although not significantly lower after fasting. Conclusions: A standardized PET imaging protocol is important to ensure reproducibility of tumor uptake and background organs across small animals. Abstract Body : Introduction. The prevalence of non-alcoholic fatty liver diseases (NAFLD) is increasing worldwide and is estimated to be ~25%. NAFLD include a large spectrum of liver injuries from simple steatosis to non-alcoholic steatohepatitis (NASH) which affects around onethird of those patients. Whereas simple steatosis has a good prognosis, NASH is associated with worse long-term outcome, with progressive fibrosis as well as higher risk of cardiovascular disease. Liver biopsy remains the only diagnostic tool that can distinguish steatosis from NASH and grade disease severity [1] . So, NASH is categorized histologically, by the presence of steatosis, inflammation with hepatocytes injury (eg. ballooning), with or without fibrosis. Hepatic steatosis can be easily diagnosed noninvasively using ultrasound imaging, and the advanced progression of NAFLD can be evaluated using ultrasound elastography by quantifying liver fibrosis. However, despite the fact that advanced fibrosis imaging in NASH has been shown to be a relevant prognostic marker of liver failure and HCC, current imaging tools, are inadequate for early detection of steatohepatitis [1] . In particular, liver inflammation which is an early event in the process cannot be currently evaluated noninvasively; new imaging tools are then greatly needed. Vascular cell adhesion molecule-1 (VCAM-1) contributes to in situ recruitment of inflammatory cells in response to cytokines. Moreover, VCAM-1 was recently deemed to be a top upregulated gene and a key pathophysiological player in NASH [2] . In this study, we evaluated the expression of VCAM-1 in the liver of several NAFLD mouse models and the potential of SPECT imaging, using an anti-VCAM-1 nanobody -99mTc-cAbVCAM1-5 - [3] , to diagnose preclinical liver inflammation. Methods. First, C57BL6/J mice fed a choline deficient (MCD) or a standard diet (STD) were imaged with 99mTc-cAbVCAM1-5 (V-MCD) or with an irrelevant control nanobody (C-MCD) as a proof-of-concept. Then, models with metabolic NAFLD were employed: wild type fed a normal (Ctl) or high fat diet (WH), FOZ fed a high fat diet (FH) and C57BL6/J fed a choline-deficient (CDH) or -supplemented (CSH) high fat diet. NAFLD activity score (NAS) was obtained from liver histology analysis. The hepatic expression of VCAM-1 and proinflammatory markers was measured using RT-qPCR and VCAM-1 protein content using ELISA. 99mTc-cAbVCAM1-5 liver uptake was expressed as a concentration (SUV) or as a total uptake (%ID: % of injected dose). Results. In MCD mice, 99mTc-cAbVCAM1-5 specifically accumulated in V-MCD liver in comparison to C-MCD and STD (SUV: 0.48±0.11 vs 0.28±0.04 and 0.26±0.03, respectively, p WH and CSH were characterized by moderate steatosis score (1.2±0.7 and 0.8±0.4) and mild inflammation score (1.0±0.6 and 0.7±0.5), while FH and CDH exhibited severe steatosis (score of 3) and inflammation (score 2.7±0.5 and 3.0±0.0). VCAM-1 expression was higher in mice with NAFLD than in Ctl (WH: 1.40±0.70, FH: 7.61±2.28, CSH: 2.74±0.38 and CDH: 8.29±4.52 fold increase over Ctl) and correlated with the expression of MCP1 (r=0.77, p 99mTc-cAbVCAM1-5 total liver uptake was 3-fold and 6-fold higher in CDH and FH livers respectively than in controls, strongly correlated with histological inflammation score (r=0.86, p steatohepatitis from simple steatosis with high performance (AUROC=0.89, p=0.035). Conclusions. 99mTc-cAbVCAM1-5 detects liver inflammation and identifies steatohepatitis in vivo noninvasively in NAFLD mouse models. Given that 99mTc-cAbVCAM1-5 is already in phase 1 clinical trial (NCT04483167) for atherosclerotic lesion imaging, these data have the potential for rapid translation and pave the way toward using 99mTc-cAbVCAM1-5 as a novel way to evaluate liver inflammation in the ever growing NAFLD population, a current unmet medical need.

References: [1] Castera L, et al. Gastroenterology. 2019, 156(5) :1264-1281.e4. [2] Furuta K, et al. J Clin Invest. 2021, 131(6) . [3] Broisat A, et al. Circ Res. 2012; 110 (7) Abstract Body : Background: Even though a combination of multiple reporter genes is a feasible approach for indirect cell labeling, therapeutic cells are often engineered using constructs that do not allow insertion of multiple reporter genes for imaging purposes. In this study we offer a solution to this unmet need by developing a small reporter gene that allows multimodality imaging. Our reporter gene system, derived from the 'NanoBiT' split-system [1, 2] , enables bioluminescence (BLI) with radionuclide imaging using PET or SPECT. The LgBiT subunit, one of the split pieces of NanoLuc luciferase, was expressed and anchored in the membrane of cancer cells. Upon administration of small part of the 'NanoBiT' split-system, the HiBiT peptide tag (11 AA) [3] , full NanoLuc luciferase will be reconstituted, upon colocalization with LgBiT subunit. Experimental design: We constructed a chimeric transmembrane reporter gene (TM-LgBiT) by fusion of LgBiT to a C-terminal transmembrane anchoring domain of platelet-derived growth factor receptor (PDGFR) and HA tag for membrane expression on cancer cells. Furthermore, we conjugated the HiBiT peptide to DOTA chelator via a 6-aminohexanoic acid (Ahx) linker and radiolabeled it for single-photon emission-computed tomography (SPECT), yielding [111In] In-DOTA-6-Ahx-HiBiT. In vitro characterization of reporter probes was followed by in vivo evaluation. We determined the characteristics of our reporter system using bioluminescence and radioactivity as detection techniques in mice bearing PC-3 tumors. Results: We established a reliable procedure for labelling HiBiT peptide with indium-111, using the [111In] In-DOTA-6-Ahx-HiBiT tracer we synthesized. Specific uptake of DOTA-6-Ahx-HiBiT was detected using bioluminescence imaging in vitro in cells, and in vivo from PC-3 tumors expressing TM-LgBiT reporter. After a SPECT/CT scan, we found specific [111In] In-DOTA-6-Ahx-HiBiT tracer uptake in mice implanted with target PC-3 positive cells (TM-LgBiT), which we confirmed using ex vivo biodistribution studies. Additionally, and importantly, uptake by tumours expressing TM-LgBiT was shown to be specific by blocking the interaction with excess of unlabeled peptide. Conclusion: We have generated and validated a novel multi-modality gene reporter system that has the potential to be applied in many fields of life sciences and we will further investigate its specificity for therapeutic cells tracking. Tc-tetrofosmin being used.2,3 While these imaging agents delivery in tissues is perfusion-dependent, as cationic species, their retention within mitochondria is determined by ΔΨm-dependent trapping according to the Nernst equation.4 If the retention of such lipophilic cations could be corrected for perfusion, either by pharmacokinetic modelling or parallel injection of non-mitochondrially dependent perfusion imaging agents, these radiotracers could be repurposed to provide a non-invasive imaging index on mitochondrial dysfunction in evolving diseases such as cancer and cardiovascular diseases, as we have recently proposed.5 Positron Emission Tomography (PET) offers significant advantages over SPECT in terms of spatial resolution, sensitivity, and primarily, the capacity for dynamic imaging to allow for such pharmacokinetic perfusion correction modelling. Previous work in our group focussed on the development of fluorine-18 and gallium-68 lipophilic cations provided the foundation for this promising approach.6-9 Aims: (i) To develop a series of PET-compatible ligands for mitochondrial dysfunction, with differing alkyl-functionalisation motifs to provide opportunities for tuneable lipophilicity, as a means of optimising their pharmacokinetic profiles. (ii) Assess their capacity to report on ΔΨm by assessing their pharmacokinetic profiles in isolated perfused rat hearts. Results: We have synthesised and radiolabelled a series NODAGA-based radiotracers, with triarylphosphonium-functionalisation, with gallium-68 to form lipophilic cations. The tracers radiolabel with over 97 % RCY in significantly mild conditions, and exhibiting log D values of between -2.43 and -0.37. The candidate tracer exhibited less than 3 % retention in healthy hearts, which was not significantly diminished by mitochondrial depolarisation with CCCP. This work suggests that while this approach is promising, the lipophilicity of this class of tracers must be increased in order for them to exhibit more favourable cardiac uptake and retention. Southworth, Sci. Rep., 2019 , 9, 216. 6 A. Haslop, L. Wells, A. Gee, C. Plisson and N. Long, Mol. Pharm., 2014 , 11, 3818-3822. 7 S. P. McCluskey, A. Haslop, C. Coello, R. N. Gunn, E. W. Tate, R. Southworth, C. Plisson, N. J. Long and L. A. Wells, J. Nucl. Med., 2019 , 60, 1750 -1756 . 8 A. J. Smith, P. J. Gawne, M. T. Ma, P. J. Blower, R. Southworth and N. J. Long, Dalt. Trans., 2018 A. J. Smith, B. E. Osborne, G. P. Keeling, P. J. Blower, R. Southworth and N. J. Long, Dalt. Trans., 2020 , 49, 1097 -1106 

Rama Jayasundar, Department of NMR, All India Institute of Medical Sciences, New Delhi, India, ramajayasundar@gmail.com

Abstract Body : Introduction Recent times have seen an alarming increase in the incidence of stroke1, a major cause of disability worldwide. With increasing interest in Complementary and Alternative Medicine (CAM), the current study has evaluated the effect of a CAM modality, namely ayurveda, in ischemic stroke patients using MRI and clinical assessment scores. Materials and methods The study was carried out with approval from the Institute Ethics Committee. Twelve first-ever stroke patients (18-60 years) were enrolled after obtaining written informed consent. The patients were assessed using MRI and clinical assessment scores such as The National Institutes of Health Stroke Scale (NIHSS), Modified Rankin Score (mRS) and Barthel Index at pre-and 6 months post-ayurveda intervention. 3D-T1 images were obtained at 3T using the following parameters: echo time 4.8 ms, repetition time 10 ms, slice thickness 1mm, voxel size-0.75x0.75x0.75 mm3, acquisition matrix -320x318x240, flip angle-80. FLAIR images were also obtained using the following parameters: voxel size-1x1x1mm3, acquisition matrix-22x19x380, TR/TE(ms)-4800/302.Standard FreeSurfer (v6.0) software was used for segmentation and measurement of cortical thickness and volume. Results Figure 1 (persuasive data) shows a representative MR segmentation data obtained before treatment (Fig. 1a) and 6 months post-ayurveda intervention (Fig. 1b) and the difference image (Fig. 1c) . The latter shows the areas (outlined by circle)which have registered an increase in cortical thickness, indicating a positive response to the treatment. Figure 2 shows the corresponding clinical scores (NIHSS, mRS and Barthel Index) between pre-and post-intervention, reflecting the clinical improvement in the patient. Conclusion With increasing interest in the use of CAM2, there is a need for rigorous scientific evaluation of the same. This study has evaluated using MRI, the structural changes effected by ayurveda in ischemic stroke patients. The preliminary results have shown an increase in cortical grey matter thickness effected by ayurveda intervention in ischemic stroke patients. Although further in-depth analysis and studies are underway, the study has shown that MRI has a definite role to play in the scientific evaluation and validation of the effect of CAM. 

Benjamin Vidal, Theranexus, benjamin.vidal@theranexus.com

Abstract Body : Neuroimaging has been widely used to map physiopathological underpinnings in attention deficit hyperactivity disorder (ADHD). In particular, functional magnetic resonance imaging studies have suggested disruption of the resting-state functional connectivity in ADHD patients1. Besides, animal models remain a valuable tool to study the pathological mechanisms involved in ADHD and guide towards the development of new treatments2. Here we investigated the resting-state functional connectivity of the spontaneously hypertensive rat (SHR) as compared to the similar genetic background control strain (Wistar-Kyoto), using functional ultrasound imaging (fUS). This new technique, enabling fast measurement of cerebral blood volume changes (CBV), has been shown as promising to study brain functional connectivity in rodents with high spatiotemporal resolution and sensitivity3. One day after a thinned-skull surgery, SHR (n=8) and Wistar-Kyoto (n=7) rats were anesthetized using isoflurane and kept in a stereotaxic frame. Body temperature was maintained using a heated blanket and cardiac and breathing rates were continuously monitored during the imaging session. Doppler images were acquired every 2.8 seconds for 20 minutes with a small animal fUS system (Iconeus, France) over two brain slices (bregma +2.2 and -4.8mm), using a motorized displacement of the ultrasonic probe, to follow spontaneous CBV changes in the prefrontal cortex and the visual areas. After spatial coregistration and frequency filtering, functional long-range connectivity was estimated between pairs of regions from an anatomical atlas. The local connectivity was also investigated using regional homogeneity analysis, with temporal coherence measurement between each pixel and its neighbors4. As compared with the control strain, SHR rats displayed altered long-range functional connectivity between spatially distinct regions, with increases between the primary and the secondary cingulate cortex, and between the primary cingulate and the mediolateral secondary visual cortex. Conversely, the connectivity between the pretectal area and the brainstem was decreased. Additionally, SHR also displayed altered local connectivity, with significant increases of regional homogeneity in the secondary motor cortex and the mediomedial secondary visual cortex, along with significant decreases in the secondary cingulate cortex, the superior colliculus and the pretectal area. As a conclusion, we show that SHR rats display various functional connectivity alterations in different scales, with signs of hyperconnectivity among different cortical regions and especially the visual cortex, and signs of hypoconnectivity in subcortical regions of the visual system such as the pretectal area and the superior colliculus. These findings are consistent with previous clinical studies that showed similar alterations of functional connectivity in ADHD1 and demonstrate the potential of fUS for investigating animal models of CNS disorders for both local and long-range connectivity. of interest as defined by the SIGMA rat brain atlas. The bottom left corner shows the mean connectivity matrix for the Wistar-Kyoto strain; the top right corner shows the mean connectivity matrix for the SHR strain. For statistical comparisons, a Fisher Z-transformation was applied to the correlation coefficients before two-way ANOVA followed by multiple comparisons with a false discovery rate of 0.1 (*q-value < 0.1). On the right: regional homogeneity maps, showing the local functional connectivity (mean for each group). The maps were obtained by estimating the temporal coherence between each pixel and its neighbors. For normalization purposes, each individual map was converted into Z-score. Abstract Body : Genetically engineered bacteria have the potential to outperform traditional diagnostics and therapeutics due to their ability to sense and respond to unique signals, access hard-to-reach areas of the body, and deliver therapeutics directly to these areas [1] [2] . However, current tools to non-invasively monitor their activity and spatial distribution in vivo are limited. Optical imaging methods that use fluorescent and luminescent reporter genes suffer from the poor penetration depth of light in tissue [3] , while nuclear imaging methods rely on radioactive tracers that have relatively low spatial resolution and require a complex synthesis pipeline [4] [5] . Ultrasound imaging has the potential to address these limitations because it is a widely available technique that does not use ionizing radiation and has tissue penetration depth of several centimeters and spatial resolution below 100 microns [6] . Recently, the first genetically encoded ultrasound contrast agents based on gas vesicles, which are air-filled protein nanostructures that are naturally used by aquatic microbes to regulate their buoyancy, were developed to link ultrasound to gene expression in bacteria [7] . However, these first-generation acoustic reporter genes (ARGs) expressed too poorly under in vivo conditions to enable ultrasound imaging of bacteria in therapeutically relevant contexts. Here, we present a new and improved ARG construct that results in high levels of robust gas vesicle expression in the probiotic bacterium E. coli Nissle, and enables ultrasound imaging of these cells with high sensitivity using nondestructive pulse sequences. This second-generation ARG construct -bARGSer -uses genes derived from a different bacterial species and was incorporated into optimized gene circuits for gene expression monitoring. We demonstrate that with bARGSer, we can visualize the spatial distribution of engineered E. coli Nissle after they home to and colonize tumors upon systemic administration [8] . We also demonstrate that the engineered E. coli Nissle can be imaged with ultrasound when colonizing the gastrointestinal (GI) tract of mice. By enabling monitoring of the precise spatial location of engineered probiotic bacteria inside the body, this technology could greatly improve the development and eventual clinical use of this emerging class of microbial cell-based theranostics. (2012) . [6] Maresca, D., Lakshmanan, A., Abedi, A., et al. Biomolecular Ultrasound and Sonogenetics. Annu Rev Chem Biomol Eng 9, 229-252(2018) . [7] Bourdeau, R., Lee-Gosselin, A., Lakshmanan, A. et al. Acoustic reporter genes for noninvasive imaging of microorganisms in mammalian hosts. Nature 553, 86-90 (2018) . [8] Among the ideal characteristics that make natural exosomes a promised nanoplatforms is the heterogeneity in their composition, which allow us to chemically modify their surface. This is known as "engineering of exosomes" and the aim is to improve the uptake of the nanovesicles in the targeted tissue, such as a tumor.2-6 The objective of this study is the enrichment of these natural nanoparticles with folic acid by the covalent functionalization of their surface and the in vitro and in vivo evaluation as diagnostic tool in oncology by optical imaging. Methods: Milk exosomes were isolated by ultracentrifugation and conjugated by two synthesis steps; an initial biological functionalization and further fluorescent labeling. For the initial bioconjugation, folic acid (0.1 mg) was activated in presence of EDC (0.06 mg) and sulfo-NHS (0.2 mg) at RT for 35 min and then, mixed with exosomes (30 μg) over night, at 4ºC. Further optical labeling of the folate-exosomes was performed in presence of Sulfo-Cyanine5 NHS ester (SCy5) in mild conditions (pH 8.5, 4ºC, 90 min) . The purification of fluorescent Fol-Exo-SCy5 exosomes was performed using by Exo-spinTM columns. Biofunctionalized exosomes were characterized by TEM, NTA, Z-potential (ZP), MALDI-TOF MS and Fluorospectrometer. In vitro assays were performed in murine macrophages (RAW 264.7) and human cervical cancer (HeLa) cells by flow cytometry and confocal microscopy. Blocking assays were performed in presence of enriched medium folic acid (1 mg/mL). In vivo and ex vivo validation of nanoprobes by optical imaging was performed in nude HeLa xenograft tumor model (2x106 HeLa cells, 250 µL matrigel:DMEM). Imaging studies performed at 1h, 3h, 6h and 24h after i.v injection of the nanoconjugate (30 μg, 200μL PBS). Histological evaluation of tumor tissue was performed by immunohistochemistry (IHC). Results: TEM confirmed that Fol-Exo-SCy5 conserved their morphology, NTA revealed a slight increase of their size (149.8+/-1nm) and ZP a surface modification (-8.2+/-1mV) . MS spectrum confirmed the incorporation of folic acid and SCy5 into the nanoparticles and fluorescence was confirmed by fluorimetry. In vitro studies revealed a fast uptake of the probe in RAW 264.7 while HeLa requires longer times for the probe uptake. Both cell lines confirmed the uptake reduction of the nanoprobe in presence of free folic acid (Fig 1) . In vivo optical imaging in a HeLa xenographic model showed signals of Fol-Exo-SCy5 in tumor tissue even at short timepoints,1h. Ex vivo biodistribution at 24h revealed main accumulation in liver, tumor, lungs, uterus and kidneys. The tumor cells populations with overexpression of folate receptors (macrophages and tumor cells) were characterized by histology/IHC and correlated with the Fol-Exo-SCy5 signal (Fig 2) . Conclusion: We have optimized a covalent method for bioengineering of natural exosomes with folic acid. The full characterization of the nanoparticles confirmed the successful bioconjugation in the surface of the nanovesicles without modifying their morphology or size. In vitro results showed a selective uptake of Fol-Exo-SCy5 via folate receptors and preliminary in vivo and ex vivo results confirmed the ability of the nanoparticles to reach tumor tissue. Acknowledgement: This work was partially supported by Comunidad de Madrid (Y2018/NMT-4949 (NanoLiver -CM and S2017/BMD-3867 RENIM-CM, co-financed by European Structural and Investment Funds), A. Santos-Coquillat is grateful for financial support from Ministerio de Ciencia e Innovación, Instituto de Salud Carlos III Sara Borrell Fellowship grant CD19/00136. Thrombosis and its complications are responsible for 30% of deaths annually1. The limitations of current methods for diagnosing and treating thrombosis highlight the need to improve diagnosis and treatment strategies2. In this study, we designed and successfully synthesized a novel theranostic nanomaterial based on an iron core and an outer layer of silver with strong absorbance at the near-infrared wavelength. Elemental silver was grown on the surface of small iron oxide nanoparticles. The silver-coated iron oxide nanoparticles (AgIONPs) were stabilized with polyethylene glycol (PEG). For targeting purposes, the nanoparticles were conjugated with RGD, fibrin-binding peptide (FibPeps) and a single chain antibody (ScFv) targeting activated platelets. In vitro human clot binding assays showed that from the three binding ligands tested, ScFv has more specific binding to the clots. The nanoparticles exhibited good r2 relaxivity (30.1 mM-1s-1) at 9.4T and demonstrated its ability to darken T2-weighted signal of in vitro thrombus. In vitro and ex vivo PAI studies showed that AgIONPs enhanced the signal of in vitro thrombus in photoacoustic imaging, especially for excitation wavelengths in the near-infrared region. Laser-enhanced thrombolysis in vitro with human blood clots indicated that an enhanced effect of thrombolysis was observed with the nanoparticles in a dose-dependent manner. Furthermore, hemocompatibility assay on red blood cells and Presto Blue Viability assay on CHO cells suggested good biocompatibility of the nanosystem. Altogether, these results suggest that the nanosystem could potentially act as a multi-imaging theranostic nanomaterial for diagnosis and treatment of thrombosis. Abstract Body : Introduction: The weak thermal polarization of nuclear spins limits the sensitivity of MRI, even for MR-sensitive nuclei as fluorine-19. This leads to the inability to map very low concentrations of targets using 19F-MRI, which thus raises the need to further enhance this platform's capabilities. Here, we employ the principles of CEST-MRI in 19F-MRI to obtain a 900-fold signal amplification of a biocompatible fluorinated agent, which can be presented in a "multicolor" fashion1. Capitalizing on the dynamic interactions in host-guest supramolecular assemblies in an approach termed GEST (guest exchange saturation transfer), we demonstrate that an inhalable fluorinated anesthetic can be used as a single 19F-probe for the concurrent detection of micromolar levels of two targets, with potential in-vivo translatability. Method: 19F-GEST NMR data were acquired on 9.4 T NMR with a pre-saturation pulse B1 (2.5μT/ 3s) applied prior to the 90o RF pulse with its frequency swept from Δω= +6.1ppm to Δω=-6.1ppm. 19F-GEST MRI of phantoms was performed using a CEST-RAREst sequence with the following parameters: B1 = 2.5μT/ 3000 ms, TR/TE=10000/2.93ms, RARE factor=18, 10mm slice, FOV=3.2×3.2cm2, matrix size=32×32, NA=120. 19F-GEST spectra were obtained after B0 correction and 19F-GEST maps were calculated using MTRasym analysis. In vivo experiments were performed on a horizontal 15.2 T scanner followed bilaterally injections of either H1 or H2 into the mouse brain striatum. Then a Low-Flow anesthesia system was used to deliver fluroxene to the mouse brain while placed inside the scanner through inhalation. Localized 19F-GEST experiments were performed on a 2.5×2.5×2.5 mm3 voxel using Image-Selected In-vivo Spectroscopy (ISIS) protocol with a pre-saturation pulse B1=2.5μT/ 2000ms in a total scan time of 16min. Results: Studying 24 different pairs composed of fluorinated anesthetic and macrocycle hosts the fluorinated anesthetic fluroxene (light blue sphere, Fig. 1a) showed the ability to generate a significant 19F-GEST contrast in the presence of two different molecular hosts (H1 and H2, Fig. 1a) with two opposite ∆ω values, either downfield or upfield relative to the chemical shift of unbound fluroxene. While the effect of H2-fluroxene was at ∆ω=-2.2 ppm (green, Fig. 1b ) that of H1-fluroxene was obtained at ∆ω= +1.6 ppm (magenta, Fig. 1b) . Then, 19F-GEST experiments conducted with a fixed concentration of fluroxene and reduced concentrations of either H1 or H2 demonstrating the ability to detect extremely low concentrations of both H1 (500 nM) and H2 (1 μM), respectively. 19F-GEST MRI experiments were then performed and presented as pseudo-multicolor imaging maps showing the ability to spatially map multiplexed systems even at the same ROI (Fig. 1c) . Using a fluorinated anesthetic as an inhalable 19F-agent (i.e., fluroxene), the potential use of the proposed platform for in-vivo studies was demonstrated. For that purpose, solutions of either of the examined hosts were injected intracranially to the two hemispheres of a live mouse (Fig. 1d) , after which the mouse was placed in a 15.2 T MRI scanner and anesthetized with fluroxene (schematics in Fig. 1d , bottom). After confirming successful delivery of the guest to the brain (Fig. 1e) , in-vivo localized 19F-GEST spectroscopy (Fig. 1f ) was performed to show the in-vivo applicability of 19F-GEST. Conclusion: We reported here on the design, development, characterization and implementation of an approach that combines the principles of supramolecular chemistry with CEST and 19F-MRI for multiplexed MRI that can detect micromolar concentrations of targets. Based on the CEST principles, which allow signal amplification, and relying on 19F-MRI, low concentrations of multiple targets could be detected. By using fluorinated anesthetic, which is delivered at a mM concentration to a rodent brain the applicability of in vivo GEST-MRS was demonstrated2 and should be further developed. (green) . The 19F-GEST spectra represent the subtraction of the 19F-NMR spectra obtained when the saturation pulse was applied "on-resonance" from the 19F-NMR spectra generated when the saturation pulse was applied "off-resonance" as noted with arrows. Low-dose spiral CT can detect small and early stage LuCa lesions, but cannot be used widely as an annual LuCa screening tool. Metabolomics detects global metabolite variations under physiology and pathology. We hypothesize that metabolomic profiles measured from blood may reveal LuCa at early stages as a screening tool to triage suspicious patients to CT tests. To test this hypothesis, we designed our innovative study of using paired blood sera samples obtained from LuCa patients both prior to and at the time of their LuCa diagnosis to establish LuCa screening metabolomic profiles. Samples. Sera from non-small cell LuCa (NSCLC) patients and their age, gender and smoking habit matched healthy controls were grouped according to the design of training-testing-validation cohorts in this study. The training cohort included 25 NSCLC sera from patients at the time of diagnosis (TOD) and controls (Ctrl); the testing cohort consisted 25 sera collected 0.5 to 5 yrs prior to diagnosis (PTD) from the 25 NSCLC patients in the training cohort; and the validation cohort recruited sera collected less than 2 yrs PTD from additional 54 NSCLC patients and controls. MR Spectroscopy. High resolution magic angle spinning (HRMAS) MRS analysis of serum samples are performed at 4°C by a 600MHz Bruker spectrometer at 3,600Hz spinning rate. Spectra are analyzed with a MatLab-based curve fitting program. Data Analysis. 57 spectral regions were selected based on the training and testing cohorts. Following selections of these regions, all data analytical procedures, including principal component and canonical analyses, were performed on the training cohort and followed by the testing and validation cohorts. Analyses of native sera HRMAS MRS of the training cohort selected 57 spectral regions for statistical analyses, and PCA and canonical analysis were conducted to differentiate TOD from Ctrl groups, with the testing and validation cohorts passively followed the calculations to produce scores within each case. Using the mean plus one standard error (M+SE) as the threshold, calculated from the canonical score differences between TOD and PTD for each case, i.e. the difference of the two scores for each patient, a Kaplan-Meier survival analysis indicates better survival rate from their time of NSCLC diagnoses for patients with score differences higher than the threshold (p=0.031). Furthermore, for Stage I and IIA patients, survivals can be predicted (p=0.027) from their PTD samples, if their score values are higher than the M+SE threshold. Since neither testing nor validation cohorts were involved in the determination of values of the canonical score, by collectively examining all Stage I and IIA in both cohorts, the resulting Kaplan-Meier survival predicting capability by the threshold (established by the testing cohort) was enhanced significantly (p=0.0044). For 5-year survival, our data showed Sen=0.63, Spe=1.00, PPV=1.00, NPV=0.21, ACC=0.67, and F1=0.78. Thus, serum metabolomics may be a screening candidate for early detections with the potential to further probe into all the related metabolic pathways for better understanding of disease. Abstract Body : Purpose Pancreatic ductal adenocarcinoma (PDAC) is currently the fourth highest cause of cancer-related death in the world, which showed the lowest 5-year relative survival rate of 9% [1] . Early and precision detection is particularly important for PDAC patients [2] . Therefore,accurate imaging for aiding diagnosis and subsequent treatment are indispensable for improving the therapeutic efficacy and patients' survival rate [3] . The aim of this study was to perform multimodality imaging as fluorescence molecular imaging (FMI), magnetic particle imaging (MPI) and magnetic resonance imaging (MRI) for the precision detection of both orthotopic and subcutaneous PDAC xenografts. Methods Plectin-1 targeted superparamagnetic iron oxide nanoparticles (SPIONs) conjugated with IRDye800CW (PTP-Fe3O4-IRDye800CW) were synthesized for targeted imaging probe. Subcutaneous and orthotopic PDAC mouse models were established. A multi-modality imaging method, including FMI, MPI and MRI, was conducted for quantitative evaluation of tumor. Histology and immunohistochemical analysis were used as references for validation. Results We successfully developed the PTP-Fe3O4-IRDye800CW nanoparticles with outstanding targeting specificity of PDAC xenografts. The probe showed good imaging performance for FMI, MPI and MRI in vitro. In both subcutaneous and orthotopic PDAC xenograft models, the PTP-Fe3O4-IRDye800CW nanoparticles showed better targeting specificity, uniform distribution, distinct periphery of tumors and longer retention effects for 7 days compared with CON-Fe3O4-IRDye800CW nanoparticles. In the orthotopic model, multi-modality imaging had the best observations at 48 h post-injection. For FMI imaging, PTP-Fe3O4-IRDye800CW group showed significantly higher normalized FMI intensities than CON-Fe3O4-IRDye800CW group (64.49% ± 7.03% vs. 41.99% ± 0.12%, P MPI imaging, PTP-Fe3O4-IRDye800CW group showed significantly higher normalized MPI signals than CON-Fe3O4-IRDye800CW group (85.72% ± 1.54% vs. 74.41% ± 1.91%, P PTP-Fe3O4-IRDye800CW nanoparticles distributed throughout the tumor and drew distinct periphery of tumors instead of confined to the injection site as CON-Fe3O4-IRDye800CW nanoparticles. Histology and immunohistochemical observations were in accordance with the multi-modality imaging. Conclusion Integration of the individual imaging advantages of FMI, MPI and MRI could allow more comprehensive and precision detection of PDAC with high sensitivity and resolution for better diagnosis at anatomical and molecular levels, which may potentially benefit patients with PDAC for early and sensitive diagnosis. Moreover, the utilization of novel MPI is more suitable for imaging tumors in deep organ due to its special imaging properties. The nanoparticles also possess intraoperative imaging and hyperthermia properties with potential for the guidance of subsequent precision therapy. 

Rama Jayasundar, Department of NMR, All India Institute of Medical Sciences, New Delhi, India, ramajayasundar@gmail.com

Abstract Body : Introduction Stroke is a leading cause of disability worldwide and affects the economy and work productivity of countries.1 Physiotherapy is the conventional rehabilitation technique. The present study evaluates the effect of physiotherapy in post-stroke rehabilitation by Magnetic Resonance imaging (MRI). Materials and methods Eleven first-ever stroke patients (18-60 years) with ischemia in left hemisphere were enrolled for the study after obtaining approval from Institute Ethics Committee and written informed consents. Qualified physiotherapists gave standard physiotherapy to all patients, one hour every day for six consecutive months. National Institutes of Health Stroke Scale was determined at 0, three and six months post-intervention. Diffusion Tensor Imaging (DTI) was carried out at 3T (Ingenia, Philips MR scanner), using the following parameters: repetition time 12210 ms, echo time 83 ms, acquisition matrix 112 X 110, flip angle 900, number of slices 66, gradient direction 32. DTI images were acquired and processed with FSL 5.0.9. Color coding of tractography pathways was based on a standard red-green-blue (RGB) code that was applied to the vector in each brain area (red for right to left, blue for dorsal to ventral, and green for anterior to posterior). Comparisons between groups (pre-and post-physiotherapy) for longitudinal changes were carried out. Results Percentage improvement in NIHSS score was observed for all the patients after three (71.5%) and six (80%) months of physiotherapy. Representative images of processed DTI data shown in Figure-1 shows the physiotherapy induced mean changes in fractional anisotropy of corticospinal tract for left and right hemisphere separately. The fractional anisotropy of left corticospinal tract decreased between on 3 and 6 months, whereas increased in the right corticospinal tract. This shows that the contralateral region has taken the activity and resulted in increase in CST tract volume. Conclusion Physiotherapy plays a major role in post stroke rehabilitation. Although there are reports on physiotherapy induced changes in physiological functions,2 to the best of our knowledge there are no studies to assess the white matter changes effected by physiotherapy in post-stroke recovery. The current study has evaluated for the first time physiotherapy induced corticospinal tract changes in post-stroke recovery using DTI. The results have shown that DTI can play a crucial role in understanding the effects of physiotherapy in post-stroke recovery. Abstract Body : Introduction Gaseous microbubbles (MB) are an established clinical ultrasound (US) contrast agent. They consist of a heavy gas core stabilized by a lipid shell, which can be easily functionalized with adhesion ligands to allow for targeted molecular imaging1-3. MB can also provide MRI contrast, which originates primarily from the susceptibility difference between their gas core and the liquid environment in which they reside4. However, the susceptibility contrast produced by bare MB is too weak, requiring the use of MB doses in excess of what is typically used for clinical US4 or the doping of the lipid shell with superparamagnetic iron oxide nanoparticles for MR detection6. HyperCEST is an indirect MR detection scheme capable of revealing signals from very dilute molecules whose concentration fails below the typical MR sensitivity. HyperCEST has been used for molecular MRI applications both in vitro5,7-10 and in vivo11. The scope of this work was to assess whether bare MB could be detected by using hyperCEST at the fM in-blood concentrations used in clinical US studies to enable their use as a dual-modality molecular imaging agent. Methods First, HyperCEST contrast was simulated using quantitative hyperCEST (qHyperCEST)18 with the Full HyperCEST (FHC) solution19. Simulations were performed for MB with different diameters and concentrations, and for different saturation pulse strengths. In vitro hyperCEST studies were performed on a high resolution 500 MHz NMR spectrometer. Small-sized (0.9 ± 0.4 μm) and medium-sized (2 ± 1 μm) perfluorocarbon MB were prepared in-house as previously described20. The size distributions and concentrations were determined via single particle optical sizing. Large-sized (3.2 ± 0.7 μm), commercially-available MB were purchased directly from Advanced Microbubbles Inc (Newark, CA, USA). Each sample was diluted to reflect standard in blood concentrations used for contrast US (10 μL/kg at 1010 MB per mL for humans17). Before NMR signal acquisition, MB samples were equilibrated to 37 °C to replicate in vivo conditions. Fresh HP 129Xe was periodically bubbled directly into the sample, followed by a continuous-wave presaturation pulse (B1), a 90° excitation pulse, and signal acquisition. The resulting Z-spectra were fit according to the FHC solution19. Results The results showed that MB produced a strong hyperCEST contrast in agreement with the simulation results. Specifically, we found that hyperCEST contrast increases with gas-volume concentration as well as with B1 strength. In addition, an optimal MB size exists that provides an exchange rate that is neither too fast (where the occupation time of 129Xe inside the MB is too short to be appreciably saturated) nor too slow (where the amount of saturated 129Xe exchanging into solution is insufficient on the timescale of an experiment). Exceptional gas turnover rates provided substantial contrast even at low B1 strengths. Efficacy diminishes once the MB diameter exceeds 2 μm. No hyperCEST effect was observed in the absence of MB. Most importantly, MB produced hyperCEST contrast at concentrations as low as 10s of fM, which is at the clinical in-blood concentration used in humans. Conclusions MB are efficient and are capable of accommodating millions of 129Xe atoms at a time. Therefore, they are detectable at concentrations much lower than carriers that maintain a 1:1 stoichiometry. Compared to nanoemulsions21 and gas vesicles7,22, MB are kinetically and magnetically superior; they can be detected at a much lower concentrations, while their narrower gas-phase peak characteristic of their spherical geometry leads to more efficient saturation, lessening concerns about SAR limitations for in vivo translation. As such, hyperCEST detection of gas MB is expected to enable the use of MB as a dual-modality contrast agent for combined US and MRI molecular imaging applications.

Effect of on-resonance saturation on 129Xe NMR spectrum. (A) Sketch of the experimental NMR setup (not on scale). A suspension of MB was placed in an NMR tube. Fresh HP 129Xe gas was bubbled directly into the suspension through a long narrow PEEK tube, connected to an HP-compatible ventilator on one end, and immersed in the NMR tube on the other end. 129Xe exchanges into and out of the MB according to kAB and kBA, respectively. (B) 129Xe NMR spectra from a suspension of MB, with on-resonance (-194.6 ppm) and offresonance RF saturation. Peaks at 0 ppm and -192 ppm correspond to 129Xe dissolved in the solvent, and residual 129Xe gas inside of the PEEK tubing, respectively. At the low MB concentrations used here, the gas-phase signal originating from 129Xe inside of the MB core, and located 2.6 ppm upfield from the gas-phase peak inside of the PEEK tubing, cannot be observed directly. But, when B1 is on-resonance the MB, a significant attenuation of the dissolved-phase signal is observed. (C) NMR 129Xe spectra of a solvent sample without MB, acquired with on-and off-resonance RF saturation. Peaks at 0 ppm and -192 ppm are still visible. However, no solvent signal reduction is observed after RF saturation at -194. 6 Abstract Body : Introduction: Photoacoustic mesoscopy is an emerging modality that can noninvasively visualise vascular networks at high resolution, yielding insight into their role in tumour development and therapy response1. Quantification of these vascular networks remains limited, as it requires accurate vessel segmentation, which is challenging to validate. Using a combination of in silico, in vitro and in vivo photoacoustic images, we assessed the accuracy and performance of two vessel segmentation methods, with and without vesselness filtering, to establish how these methods influence tumour vessel segmentation. Methods: To generate in silico ground-truth (GT) vascular networks, we utilised Lindenmayer systems2 (referred to as Lnets, n=30). We simulated photoacoustic imaging of the L-nets using SIMPA3 (v0.1.1) at 532nm and the k-Wave MATLAB toolbox (v1.3, MATLAB v2020b), assuming a speed of sound of 1500 ms-1. Red-coloured string agar phantoms provided in vitro GT (n=7). Strings were embedded at three different depths from the agar surface (0.5, 1 and 2mm). For in vivo analysis, breast patient-derived xenografts (PDXs) were implanted (n=14). Photoacoustic mesoscopy imaging of string phantoms and PDXs was conducted at 532nm. Excised PDXs underwent ex vivo immunohistochemistry to stain for CD31 (normalised area measures vessel density). We tested a moment preserving thresholding method (included in Fiji v2.1.0, referred to as Auto-Thresholding (AT)) and a learning-based segmentation method using random forest classifiers4 (ilastik v1.3.325, referred to as Random Forest (RF)). Between pre-processing and segmentation, feature enhancement was tested as a variable using Sato filtering (α=0.25) to produce vesselness filtered images (VF). Therefore, 4 segmentation pipelines were tested on all datasets: AT, RF, AT+VF and RF+VF. Results: Using L-nets as in silico GT, we found learning-based RF segmentation methods provide the most accurate segmentation, with lower mean squared errors (MSE) compared to the GT for segmented blood volume (MSE: AT=0.046±0.0005, AT+VF=0.058±0.0008, RF=0.023±0.0002, RF+VF=0.016±0.0002). We observed that RF methods are able to segment vessels distant from the simulated illumination source, where the signal-to-noise ratio (SNR) decreases, whereas AT methods fail to do so. Supporting this, in vitro in string phantoms, with strings placed at different depths, RF methods were able to cope with the depth-dependent decrease in SNR. AT methods failed to accurately segment the full extent of the strings at depth. Visual inspection of the in vivo tumour networks subjected to our processing pipelines suggests that VF increases vessel diameters in vivo, which quantitatively resulted in increased normalised blood volume. All methods resulted in segmented blood volumes that correlated with vessel density measured ex vivo, except AT (Pearson's correlation coefficient r with p-value: AT r=0.22 p=0.45; AT+VF r=0.64 p=0.01; RF r=0.61 p=0.02; RF+VF r=0.55 p=0.04). Discussion: In silico and in vitro GT datasets both highlighted that a learningbased RF methods can better segment vessels at depth, where SNR decreases because of the decrease in light fluence. The performance of RF methods is explained by the inclusion of features other than just intensity, as with AT methods, but also vessel edges and texture, resulting in more accurate segmentations. VF visually increases vessel diameters and ultimately blood volume in vivo, which can be associated with acoustic reverberations observed surrounding vessels in vivo, spreading the apparent extent of vessels and ultimately increased blood volume. Correlations with vessel density measured ex vivo, support our in vivo measured blood volumes in all methods except AT. Conclusion: Using a combination of in silico, in vitro and in vivo photoacoustic mesoscopy datasets, we have compared and validated the performance of two vessel segmentation methods, with or without the application of vesselness filtering, for use in photoacoustic data segmentation and quantification, finding that learning-based methods yield the best performance. Abstract Body : Study objectives and hypothesis: The goal of this study is to evaluate whether the reactive oxygen species (ROS) activated magnetic resonance imaging (MRI) probe Fe-PyC3A can be used to non-invasively image an increase in intrahepatic neutrophil content in a mouse model of acetaminophen overdose. Fe-PyC3A is a low molecular weight iron complex that is oxidized from a low-relaxivity Fe2+ (r1 = 0.18 mM-1s-1, 4.7T, RT) to a high-relaxivity (r1 = 2.4 mM-1s-1, 4.7T, RT) complex by ROS.1 In normal liver ROS levels are low, but the metabolic and inflammatory changes (e.g. infiltration of activated neutrophils) associated with drug induced liver injury results in increased ROS levels and an aberrant oxidizing liver microenvironment.2-4 We hypothesize that ROS imaging using Fe-PyC3A represents a biomarker to detect, quantify, and monitor changes in liver oxidative stress associated with acute liver inflammation. Innovation and Significance. Liver toxicity is the most frequent form of drug related adverse event and is responsible for an estimated 1/5th of post-approval drug withdrawals.5-7 Liver toxicity is also a major reason for discontinuation of drug development programs.7 Drug hepatotoxicity is often linked to an increase in intrahepatic ROS levels related to altered hepatocellular metabolism or acute liver inflammation. 8 The results of this preliminary study demonstrate how Fe-PyC3A generates liver signal increase that correlates with pathologic neutrophil infiltration following acute drug related liver injury. Our preliminary results suggest that liver ROS imaging with Fe-PyC3A represents a potentially powerful biomarker to noninvasively diagnose and monitor inflammation and/ or oxidative stress related to drug hepatotoxicity. Experiment design. We used Fe-PyC3A to image liver inflammation following a hepatotoxic dose of acetaminophen (formulated at 15 mg/mL and administered i.p.). Male C57BL/6 mice (5 weeks of age) were imaged 16 -22 hours after receiving an intraperitoneal dose of 300 mg/kg acetaminophen (N = 9), 450 mg/kg acetaminophen (N = 20), or saline vehicle (N = 12). Each mouse was euthanized after imaging and the right and left liver lobes were harvested for analysis of inflammatory infiltrate by flow cytometry. A total of 41 mice were used in this experiment. The mice were imaged with a 4.7 T small-bore scanner (Bruker Biospec) with a custom-built volume coil. 2D T1-weighted coronal gradient echo images of the abdomen were acquired prior to and 2 minutes after injection of 0.2 mmol/kg Fe-PyC3A via the tail vein. The (post-pre) injection change in liver vs. muscle contrast-to-noise ratio (DCNR) was quantified separately in the left and right lobes of the liver. After imaging, the left and right lobes of the liver were harvested and assayed for neutrophil content by flow cytometry. Flow cytometry data was reported as percentage of live leukocytes. A portion of the right and left lobes of N=11 mice receiving a 450 mg/kg dose of acetaminophen were also assayed for myeloperoxidase positive (MPO+) cell content by immunohistochemistry (IHC). Comparisons in DCNR and intrahepatic leukocyte levels between groups of mice were made using one-way ANOVA followed by Tukey's post-hoc test for multiple comparisons. We tested for correlations between DCNR and percentage intrahepatic neutrophils using the entire set of 82 liver samples analyzed by calculating the Pearson's product-moment correlation coefficient. Results. Figure 1A compares coronal T1-weighted images of mice 18 hours after i.p. injection of either saline or 450 mg/kg acetaminophen acquired prior to and 2 min after Fe-PyC3A injection. DCNR correlated positively and significantly with intrahepatic neutrophil content determined by flow cytometry (P Conclusions. Our data indicate that Fe-PyC3A represents a potentially useful technology for precision imaging of ROS-generating inflammatory cells in the liver. Abstract Body : Introduction: CD19 proved to be an excellent therapeutic target in B cell malignancies. With CD19-directed chimeric antigen receptor (CAR) T cells, bispecific T cell engagers (BiTE), and Fc-optimized monoclonal antibodies (mAbs), three auspicious treatment approaches have been developed for relapsed/refractory B cell lymphoma; however, all suffering from limited response rates and/or considerable toxicities [1, 2] . In clinical routine, therapy decisions are solely based on histological CD19 staining at initial diagnosis, disregarding tumor heterogeneity and altered temporal CD19 expression levels during treatment. In this work, we report on the development of radiolabeled αCD19-mAbs for positron emission tomography (CD19-ImmunoPET) aiming to visualize CD19 expression on B cell lymphomas non-invasively, thus facilitating both personalized patient stratification and therapy guidance. Methods: The antihuman CD19 mAb 4G7 (αhCD19), the GMP-grade Fc optimized anti-human CD19 mAb 4G7SDIE (αhCD19-Fcopt), and the anti-mouse CD19 mAb 1D3 (αmCD19) were conjugated by the metal chelator NODAGA for radiolabeling with copper-64 (64Cu). The radiolabeling efficiency was assessed by HPLC and TLC. Antigen excess cell binding and blocking studies were performed to determine the immunoreactive fraction, specific binding, and internalization of the mAbs. Human Daudi lymphoma cells or CD19KO Daudi cells were injected subcutaneously in CD1 nude mice. The syngeneic A20 lymphoma cells were engrafted in BALB/c mice for imaging studies of the αmCD19. Mice receiving either unrelated 64Cu-control mAb, blocking of CD19 binding sites or a B cell depleting anti-CD20 mAb (syngeneic mouse model) served as control groups. ImmunoPET was performed 6 h, 24 h, and 48 h after 64Cu-αCD19 injection. For ex vivo cross-validation, the organ uptake was measured by γ-counting after the last imaging time point. Most importantly, 64Cu-αhCD19-Fcopt was administered to four B cell lymphoma patients based on compassionate use to evaluate the eligibility for CD19directed therapy. Results: Constantly high radiolabeling efficiency of >90% was achieved for all three 64Cu-αCD19. We revealed immunoreactive fractions of 62% (64Cu-αhCD19), 72% (64Cu-αhCD19-Fcopt), and 67% (64Cu-αmCD19), respectively. The three radioimmunoconjugates bound specifically to CD19 expressing cells and an internalization capacity of 56% after 24 h was assessed for 64Cu-αhCD19-Fcopt. ImmunoPET confirmed in vivo targeting of 64Cu-αhCD19-Fcopt to CD19-positive lymphoma lesions with a significantly higher accumulation in Daudi tumors (6.2 %ID/cc) compared to CD19KO tumors (4.4 %ID/cc, p 50% higher in the target group compared to the control groups. Ex vivo biodistribution analyses confirmed the obtained in vivo results. Furthermore, neither the binding to physiological B cells in the syngeneic A20 lymphoma model nor the enhanced Fc-γ mediated spleen uptake of the 64Cu-αhCD19-Fcopt affected distinct detection of CD19-positive lymphoma sites by CD19-ImmunoPET. Thus, the GMP-grade αhCD19-Fcopt, which was already clinically applied to treat minimal residual disease of childhood leukemia [3] , was radiolabeled for investigations in humans. In four Non-Hodgkin-Lymphoma patients with histologically confirmed CD19 expression 64Cu-αhCD19-Fcopt was able to detect cervical, mediastinal, and abdominal CD19positive lymphoma sites as well as bone marrow infiltration. Conclusion: CD19-ImmunoPET enables the non-invasive in vivo identification of CD19-positive lymphoma sites in different lymphoma mouse models and patients. These results emphasize the potential application of 64Cu-CD19-mAbs as a novel imaging tool to guide emerging CD19-targeting therapies. 

Michel Modo, University of Pittsburgh, mmm154@pitt.edu

Abstract Body : Bioscaffolds consisting of extracellular matrix (ECM) based hydrogel have the potential to promote tissue restoration after a stroke. Infiltration of immune cells, especially macrophages, is a pivotal event to drive hydrogel biodegradation that leads to the invasion of neural cells. However, the spatio-temporal dynamics of the infiltration of immune cells into peri-infarct tissue and the ECM hydrogel remain poorly understood. The tagging of immune cells using perfluorcarbon (PFC) nanoemulsions afford their in vivo visualization using 19F magnetic resonance imaging (MRI). Rats with middle cerebral artery occlusion, a model of ischemic stroke, were implanted with ECM hydrogel 14 days after infarction using MRI-guidance to define the site and volume of injection. One day prior to the ECM implantation, animals were injected with PFC through the tail vein and baseline 19F MR image was acquired to verify the injection and potential passive influx to the stroke-damage brain. No PFC influx to the brain was evident at this time point. Longitudinal "time lapse" imaging over 24 hours indicated that no immune cells invaded the brain for 6 hours after ECM implantation, although a major infiltration in the scalp incision wound was evident within 1 hour after surgery. Infiltration of immune cells was first evident in the peri-infarct area at 9 hours post-implantation with subsequent invasion into the ECM hydrogel. By 24 hours a robust invasion of immune cells into the ECM hydrogel was evident, as verified by immunohistochemistry in post-mortem tissue. Histologic analyses revealed that almost all invading macrophages were labeled with PFC indicating that these were peripheral macrophages rather than brain-derived microglia. ECM hydrogel implantation therefore induces a briefly delayed invasion of macrophages into the stroke-damaged brain through the peri-infarct area that is governed by peripheral macrophage effecting the initial biodegradation events that lead to brain tissue restoration.

Image/ Figure Caption: Immune cells, such as macrophages, were tagged systemically by injecting PFCs through the tail vein. This afforded their in vivo detection using 19F MRI and localization to the peri-infarct area during the infiltration process as well as their invasion into the ECM bioscaffold. A 19F MRI and histology overlay provided confirmation of immune cells corresponding to the PFC signal on 19F MRI. Type I collagen is a major component of fibrotic tissue present in the extracellular space at a concentration above 10 µM, thus making it an acceptable target for MRI. Here we investigated a novel MRI collagen targeting contrast agent as a potential intestinal fibrosis biomarker in a mouse model of IBD. Materials and methods. To induce colitis, 7-week-old male C57BL/6 mice received 2.0% dextran sulfate sodium (DSS) in the drinking water for 5 days followed by 16 days of normal water. To evaluate disease progression, this 3-week cycle was repeated 2, 3 or 4 times (Fig. 1A) . Age-matched mice receiving normal drinking water were used as controls. MR imaging was conducted after 2, 3 and 4 cycles using EP-3533, a type I collagen-targeting MR probe. To minimize bowel peristalsis during scanning, mice received an anti-peristaltic agent (N-Butylscopolamine, 5 mg/kg) via i.p. injection. MR imaging was performed on a 4.7T (Bruker Biospec) scanner. Axial T1-weighted 2D FLASH images, covering the descending colon and the small intestine, were acquired prior to and 15, 30, 40 and 50 minutes after i.v. injection of 5 µmol/kg EP-3533. To demonstrate specific binding to type I collagen, another group of DSS mice were imaged at 12 weeks using EP-3612, an isomeric non-binding control probe. These same mice were imaged again 24h later using EP-3533. Mice were euthanized after each scanning session and the colons collected for histology. Additionally, a tissue section from a mouse that received EP-3533 was used for elemental mapping of Gd by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) to evaluate the probe distribution within the bowel wall. For MR image analysis, signal intensity (SI) of the bowel wall was measured pre and post probe injection and the % increase calculated (%SI); the slope obtained from %SI vs time post injection was used as an indicator of probe washout rate. Results. As demonstrated by histology, the 3 DSS-treated groups developed intestinal fibrosis with an increased collagen deposition compared to the control groups (Fig. 1B) . After EP-3533 administration to DSS injured mice, the initial %SI in the bowel wall was lower than in control mice, but the signal intensity persisted over the course of the study while in control mice the %SI decreased rapidly (Fig. 1C) . The washout rate was highly significant for distinguishing fibrosis from normal intestine. The washout rate of EP-3533 in DSS-treated mice was similar regardless of treatment cycles with DSS, and this correlates with histological analysis where no major difference in disease severity was observed after 2, 3 or 4 cycles of DSS (Fig 1D) . EP-3533 showed significantly higher signal enhancement in DSS injured mice compared to non-binding isomer EP-3612 demonstrating specificity for collagen (Fig. 1E, 1F) . Elemental imaging of a colon section from a mouse that received EP-3533 clearly indicates that the probe distribution correlates with the collagen layer (Fig 1G) . Conclusion. The type I collagen targeting probe EP-3533 is specific for intestinal fibrosis in the DSS mouse model of colitis. Our results lend evidence to the use of EP-3533 as a potentially translatable biomarker of intestinal fibrosis. Figure 1 A) Protocol for colitis induction using 2.0% dextran sulfate sodium water. B) Representative images of colon tissue stained with trichrome for control mice and mice with colitis exposed to 4 cycles of DSS/normal water showing deposition of collagen (blue stain). Yellow arrows point to the collagen layer. C) Time -signal intensity curves for DSS-injured mice (4 cycles) compared to age-matched control mice after injection of 5 µmol/kg collagen targeting MR probe EP-3533. D) Slope of %SI vs time curves showing significant differences in probe washout between the control and DSS-injured mice after 2, 3 and 4 cycles; no significant difference in slopes was observed between the 3 DSS groups. E) Axial T1weighted MR images of a DSS-treated mouse prior to and 50 minutes post injection of the isomeric non-binding control probe EP-3612 (top), and the same mouse imaged 24h later with EP-3533 (bottom). Signal enhancement of the bowel wall was observed only after EP-3533 administration. Arrows point to the bowel wall. F) Change in signal intensity in the bowel wall of DSS-treated mice after injection of EP-3533 or EP-3612. The significant difference indicates the specificity of EP-3533 for detecting intestinal fibrosis. G) EP-3533 distribution in a colon section obtained from a mouse with colitis analyzed by LA-ICP-MS (detection of gadolinium m/z 157). An adjacent section was stained with trichrome. The probe distribution correlates with the collagen layer stained in blue with Masson's trichrome stain, validating the specific detection of intestinal fibrosis. Abstract Body : Background DDX24 is an important nucleolar protein, which was recently identified to be an important oncogen, and also associated to vascular malformations disease named Multi-organ Venous and Lymphatic Defect Syndrome (MOVLD). The purpose of this research is to develop a highly potent peptide probe that could be used for DDX24-related disease detection via binding to nucleolar DDX24 protein. Materials and Methods Screening of binding peptide candidates to DDX24 from a random bacterial surface display library coupled with magnetic cell sorting (MACS) and fluorescence-activated cell sorting (FACS) respectively ( Figure 1A ). Then the binding potency of selected peptides were docked with homology modeling of DDX24 to further identify specific peptide probes. To validate the chosen peptide probe, the peptides were labeled with FITC and tested the binding ability for DDX24 in cell line stably expressed DDX24, via immunofluorescence (IF) and flow cytometry (FCM) experiments in vitro. Moreover, the tumor-bearing mice models were established and tested with CY7-labeled optimal peptide via intratumoral injection to evaluated the specificity and affinity targeting DDX24 in vivo. Results Several potent and selective peptides targeting DDX24 have been successfully screened via bacteria-displayed peptide method. The ATP-binding domain of DDX24 was docked with peptides using homology modeling by Schrodinger software, revealing a peptide2/DDX24 interaction with geometrical and energy match (-10.785 kcal/mol) ( Figure  1B) . IF, FCM and cell saturation binding fluorescence experiments were performed to identify that the peptide 2 was the most highly selective peptide ( Figure 1C ). NIRF imaging revealed that CY7-labeled peptide 2 showed a higher uptake in DDX24-positive tumor than DDX24-negative tumor in live animals ( Figure 1D , E). Conclusions Peptide 2, bound to DDX24 specifically and sensitively. It may be a potential candidate for molecular imaging detection and targeting therapy of DDX24-related diseases. Abstract Body : Introduction: Vascular phenotypes in breast cancer are diverse and influence tumour biology and therapy response1. Photoacoustic imaging (PAI) is an emerging modality that allows non-invasive visualisation of tumour vascular features through the endogenous contrast of deoxy-and oxy-haemoglobin (Hb and HbO2)2. We hypothesised that PAI could provide a non-invasive tool to monitor vascular phenotypes in breast patient-derived xenografts (PDXs). Methods: Two luminal B (lumB) PDXs (AB580, STG143) and two basal PDXs (STG139, STG321) were implanted in NSG mice (nAB580(LumB) = 17, nSTG143(LumB) = 6; nSTG139(Basal) = 26, nSTG321(Basal) = 12). In vivo PAI was performed at the mesoscopy (RSOM) and macroscopy (MSOT) levels. RSOM was conducted weekly once tumours were palpable to yield longitudinal data on tumour blood volume (BVRSOM). MSOT was conducted when tumours reached ~0.4cm. Total haemoglobin (THbMSOT = Hb+HbO2) and blood oxygen saturation images (SO2MSOT = HbO2/THbMSOT) were captured3. Once excised, tumours underwent immunohistochemistry (IHC) to stain CD31 (endothelial cells), CAIX (hypoxia), VEGF (angiogenic factor) and ASMA (%ASMA coverage of CD31+ vessels, indicating vessel maturity). Total RNA was extracted from flash frozen tumour tissue. Libraries for RNA-seq were prepared with Illumina TruSeq and sequenced. Transcript levels were quantified with Salmon and further analysis was conducted in R. Scores for expression of a breast hypoxia gene signature4 were generated using GSVA5. Results: Longitudinal data analyses showed that BVRSOM increased linearly with tumour volume in all models and at a significantly higher rate in basal PDXs compared to lumB PDXs (p=0.03). Supporting this, THbMSOT was significantly higher in basal PDXs (2958±195.8a.u. vs. 1391±254.8a.u., p Vessel maturity was higher in the basal STG321 model (%ASMA coverage:STG321 36.75±1.27% vs. STG139 20.08±1.15%, AB580 28.76±1.35%, STG143 25.64±0.75%, p Discussion: We have examined breast PDXs from basal and lumB subtypes, finding markedly different vascular phenotypes between subtypes as well as between individual models. Our results suggest that while both lumB PDXs (AB580 and STG143) contain relatively few blood vessels, they are mature and perfuse the tumour, leading to high SO2MSOT and low hypoxia. Despite low CAIX expression, GSVA scores of hypoxic gene signature expression were relatively high in AB580 and the hypoxic pathways involved need further investigation. Both basal PDXs (STG139 and STG321) have high blood vessel content however, differences in vessel maturity appear to govern the perfusion, oxygenation and ultimately lead to discordance in their overall levels of hypoxia. This was supported by GSVA scores, which were higher in STG139 than STG321, giving a molecular basis to the phenotypic variation observed non-invasively using PAI. Conclusion: PAI is sensitive to different vascular phenotypes in breast PDXs across scales, providing a non-invasive visualisation of these phenotypes observed using IHC and reflecting underlying gene expression.

Importantly, our results suggest that some vascular features are model dependent, whereas others are subtype dependent potentiating further exploration in additional mouse models and in future, in patients. Abstract Body : Background: The gut-brain axis has been recognized to play a critical role in health and disease. Gut microbiota dysbiosis during aging affects multiple systems including digestive system, circulated system and central nervous system. Fecal microbiota transplants have been applied to treat intestinal disease. However, whether fecal transplants from young donors could improve cognition decline induced by aging is still unclear. In this study, we investigated the effect of young fecal microbiota transplants in cognition of natural aged mice. Methods: Young fecal microbiota was collected from young donors (3 months, n = 10). Aged mice (14 months, n = 40) were randomly assigned into two groups, 14M control (14M WT, n = 20) and 14M Trans (n = 20). All the group received antibiotic treatment for two weeks, followed by young fecal microbiota transplantation twice a week for two weeks in 14M Trans group. After this two-week period, 14M Trans group received microbiota transplantation once a week for five months. Open field assay and passive avoidance assay were used to evaluate the motor function and memory of aged mice, respectively. 18F-fluorodeoxyglucose (18F-FDG) scanning was performed to estimate the glucose metabolism of brain. Western blot, reverse transcriptionpolymerase chain reaction (RT-qPCR), and immunofluorescence were performed to depict the changes of neuroinflammation and microglia activation in the brain. Results: Open field assay showed that aged mice received young fecal microbiota transplants had no difference in motor function compared with 14M WT group (A-D). However, 14M Trans group performed better in passive avoidance assay, which indicated that the cognition of 14M Trans group was improved (E). 18F-FDG scanning showed that the glucose uptake in hippocampus (p Conclusion: Our results showed that young fecal microbiota transplants could attenuate the cognition decline in natural aged mice. Glucose metabolism was improved in natural aged mice, especially in hippocampus and striatum. TNF-α and activated microglia were decreased after young fecal microbiota transplantation. These results suggested that young fecal microbiota transplantation might improve cognition of natural aged mice via affecting neuroinflammation, microglia activation and glucose metabolism. Abstract Body : Background and purpose Nasopharyngeal carcinoma (NPC) is most likely to occur with metastatic lymph nodes (mLN) in head and neck tumors owing to the extensive extravascular lymphatic capillary network under the mucosa1. The presence of mLN is a significant predictor of distant metastasis, especially when extranodal extension (ENE), described as tumor cells perforating the capsule into perinodal tissue, is diagnosed2,3. Yet significant advances in magnetic resonance imaging and positron emission tomography/computed tomography have increased the diagnosed accuracy of mLN, there has insufficient sensitivity and specificity to detect occult mLN with axial diameters smaller than 10 mm, and inadequate spatial resolution to confirm ENE for current imaging methods4,5. The purpose in our study is to evaluate the potential of Carbonic Anhydrases IX targeted probe (CAIX-800) for in vivo detection of occult mLN with higher sensitivity and specificity, and identification of ENE with sufficient spatial resolution using ultrasound guided spectroscopic photoacoustic imaging. Method Cell surface marker CAIX was validated for protein expression by western blot and confocal microscopy in seven human NPC cells. A peptide probe CAIX-800, containing IRDye 800CW and a moiety with high affinity to CAIX was synthetized based on our previous study6. The mLN models were generated by footpad injection of two human NPC cell lines (S18 and 5-8F) with high metastatic potential, respectively (n=10 per group, nude mice). Each mouse was received intravenous injection of 15 μM CAIX-800 and imaged with a closedfield near-infrared fluorescence imager to detect the occult mLN after 3 weeks of footpad injection. The feasibility of ENE identification was further in vivo imaged by a preclinical Vevo 2100 LAZR Imaging system. Harvested LNs were histologically examined ex vivo by the pathologist to confirm their metastatic status. Results Analysis of western blot revealed high-tomoderate levels of CAIX expression in seven NPC cell lines. Confocal microscopy demonstrated the satisfying cell targeting ability of CAIX-800 in vitro. In vivo optical imaging showed a significantly higher mean fluorescence intensity (MFI) in metastatic LNs vs. benign LNs, 3.646 ± 1.609 vs. 1.292 ± 0.737, respectively (mean ± standard deviation; Mann-Whitney U-test, p High area under the curve (AUC) were obtained for MFI (AUC = 0.9225, 95% confidence interval (CI), 0.8449-1.000) in receiver operating characteristic curve (ROC). The optimal threshold value for MFI was 1.960 at which a 80% sensitivity and a 80% specificity was reached (likelihood ratio 4.0). Notably, CAIX-800 clearly visualized most mLN(16/20) with excellent image quality, even if the axial diameter is as small as 1.6 mm( Figure.1B) . Photoacoustic signal in the LNs corresponding to accumulated CAIX-800 were spectrally unmixed using a linear least square algorithm. Metastatic LNs had a third-fold increase in photoacoustic signal compared to benign LNs at 8 hour time point (Fig. 1E , p Conclusion Our findings offer an effective and accurate tool to non-invasively identify occult mLN and extranodal extension, which have a strong potential for the diagnosis of early stage of metastatic NPC . Abstract Body : Objective: Studies involving the binding of proteins and rare earth elements (REE) are attractive to many researchers due to the versatility of proteins and unique attributes of rare earths. Among REEs, gadolinium is the most commonly used element for molecular imaging in the clinic with more than 30 million injections a year1. Several protein-based MRI contrast agents have been reported with attributes such as enhanced relaxivity, stability, and targeting capabilities2. Synthesis of these contrast agents can be simply achieved via bioreactors, allowing mass production at an affordable price while eliminating several steps and byproducts from chemical reactions. It is ideal to assess the presence of free ions left in the solution after initiating a binding to avoid reaction with other molecules in-vivo, and an excess amount of protein/chelate may be added to ensure safety. The answer to the question of how much more should be added above the theoretical ratio could differ with unique binding kinetics and adding too much would decrease performance. A current method of evaluating the existence of free rare earths involves the use of Arsenazo III, which is a calcium sensing dye that reacts with REEs as well as uranium, thorium, and zirconium. Here, we sought out to create a biosynthetically producible recombinant protein that directly indicates binding with high specificity for REEs, which can be utilized downstream as an MRI and optical imaging agent upon binding gadolinium. We have developed such a probe and have termed this new construct GLamouR (Green Lanmodulin-based Reporter, Figure 1 ). Methods: Protein was expressed in E. coli and isolated with HIS-tag purification and size exclusion. Western blot against the V5 tag (fused to the C terminus of the protein) was used to cross-confirm purification. Samples were dialyzed with TRIS buffer to remove residual imidazole in the solution. Fluorescence was acquired every 11 seconds, and REEs were added after the initial two readings. MRI measurements were performed under a field strength of 7T with samples in 1.5 mL centrifuge tubes. T1 relaxation times were acquired in six 1mm slices with 12 TR experiments ranging from 800ms to 17500ms, averaged 3 times. T2 relaxation times were acquired in the same geometry, with 75 echo images from 7.5ms to 562.5ms averaged 7 times. Results: Emission at 510nm (Figure 1 of Persuasive Data, n=5 wells with 10 averages, confirmed with separately purified GLamouR from 3 independent colonies) were observed to change +170.3% (±17.2) for Gd3+ and +169.2% (±21) for La3+, when compared to the first read. Negative controls were -9.5% for Ca2+ (±4.1) and -14.5% (±0.4) for TRIS buffer. Decrease in the controls are thought to be attributed to dilution upon adding REEs. MRI data ( Figure 2 of Persuasive Data, confirmed for 2 independent batches) show the T1 relaxation time was shortened by approximately 50% from 3316 (± 9) ms for TRIS buffer to 1698 (±16) ms for TRIS solution containing the GLamouR-Gd complex. T2 relaxation time was shortened from 288.1 (±2.7) ms to 224.5 (±1.3) ms, resulting in a decrease of about 22%. Conclusion: Biosynthesis of contrast agents such as the GLamouR have the potential to be utilized as a contrast agent for MRI and optical imaging as well as a probe for residual gadolinium reported through an increase of fluorescence. This system can eliminate several steps needed for quality control and synthesis, dramatically simplifying the process. Scaling up production while lowering costs via biosynthesis is being used in the pharmaceutical industry for therapeutics but has yet to be explored with molecular imaging probes. Abstract Body : Introduction Cardiovascular diseases such as stroke, coronary and peripheral artery diseases are the leading cause for morbidities and mortalities in the western world. Life threatening clinical manifestations are based on the formation of an arterial plaque as a consequence of atherosclerosis. This process is initiated by endothelial lesions, followed by immune cell infiltration and intracellular accumulation of lipid molecules finally forming a core of extracellular lipids called atherosclerotic plaques. Herein, vascular smooth muscle cells (VSMCs), which are characterized by their enormous plasticity, are present at all stages and thus play a dominant role in disease progression but also in protection of the pathogenesis (1). Consequently, non-invasive in vivo visualization and tracking of VSMCs in atherosclerosis might represent a novel powerful tool to monitor atherosclerotic plaque formation and to uncover the dedicated role of VSMCs in atherosclerosis. Methods First, we generated experimental mice which express the PET reporter gene sr39tk exclusively in VSMCs (sr39tk/VSMCs) after Cremediated recombination enabling their non-invasive in vivo visualization. Furthermore, to enable visualization of the temporal dynamics of atherosclerotic plaque formation, sr39tk/VSMCs mice were crossed with apolipoprotein-deficient (ApoE-/-) mice which develop atherosclerosis. Thus, the radiotracer [18F] FHBG exclusively accumulates in sr39tk expressing VSMCs after sr39tkmediated phosphorylation enabling non-invasive in vivo detection of VSMCs by PET/MRI. In our experimental setup MRI was used for anatomical co-registration. In addition we performed biodistribution analysis, autoradiography, histopathology (H&E) and immunohistochemistry for ex vivo cross validation of the in vivo [18F] FHBG. Results Non-invasive in vivo [18F] FHBG PET/MRI of transgenic sr39tk/VSMCs mice revealed a significantly increased [18F] FHBG uptake in sr39tk expressing VSMCs in the aorta and its branched vessels but not in wild-type control mice without sr39tk expressing VSMCs (PET signal equals background signal). In line with our expectations we determined an even more profound [18F] FHBG uptake in atherosclerotic plaques of atherosclerosis developing sr39tk/VSMCs x ApoE-/-mice. In these mice [18F] FHBG PET/MRI allowed the precise visualization of arteries as well as the site of aortic atherosclerotic plaque formation. Quantitative analysis of the increased [18F] FHBG uptake within the aortic atherosclerotic plaques of sr39tk/VSMCs x ApoE-/-mice revealed significantly elevated [18F] FHBG uptake values (1.075 %ID/ml ± 0.188 %ID/ml SEM; n = 5 ) when compared to the unaffected aortic arch of sr39tk/VSMCs mice (0.6112 %ID/ml ± 0.058 %ID/ml SEM; n = 5). Ex vivo [18F] FHBG biodistribution and autoradiography analysis verified our in vivo [18F] Abstract Body : The Chemokine CXCL12 /stromal cell-derived factor 1 alpha and its receptor CXCR4 play important roles in various biological and pathological processes, including growth, migration, adhesion, invasion and infiltration of multiple cancer and organ specific metastasis.

There is a pressing unmet need to develop non-invasive molecular imaging, with the additional quantification capability to monitor the CXCL12/CXCR4 axis at different stages of liver metastasis promoted by the high-fat diet at the cell, tissue, and body levels. Here we report the recent development of our pioneered novel class of protein-based MRI contrast agents focusing on humanized Gd3+ based CXCR4 and CXCL12 targeted MRI contrast agents (hProCA32.CXCR4 and hProCA32.CXCL12). Both hProCA32.CXCR4 and hProCA32.CXCL12 exhibits high dual relaxivities that are 5-10 folds of Eovist and Multihance at both 1.4T and 7.0T. They also exhibit much stronger serum stability and metal selectivity for Gd over physiological metal ions than clinically approved agents. Importantly CXCR4 targeted contrast agent has a strong binding capability to expressed CXCR4 on cancer cells (Kd: 0.4 ± 0.05 µM) while not perturbing downstream signaling. This important feature enables us to apply ProCA32.CXCR4 as an excellent companion diagnostic agent for monitoring treatment effects and drug development. We have demonstrated that ProCA32.CXCR4 has enabled early detection and stage of liver metastasis of cancers (including uveal melanoma, ovarian cancer) with muchimproved sensitivity and specificity compared with non-targeted agents and Eovist. The CXCL12 targeted MRI contrast agent (hProCA32.CXCL12) is able to provide the first delineation of the micro-vessel (sinusoid space) structure of the liver using pMRI. We discovered that high fat largely alters liver sinusoid structure with a microenvironment prometastasis. Large production and application of these developed contrast agents will have a major impact on tumor diagnosis and treatment stratification. 

Kiel Neumann, University of Virginia, kdn2e@virginia.edu

Abstract Body : Since 1997, a multidimensional approach has been recommended to diagnose and manage athletes suspected of having a sport-related concussion (SRC).1, 2 The recommended assessment protocol consists of clinic-based measures of balance, neurocognitive function, and concussion-related symptomology which has been demonstrated to have a sensitivity of up to 100% within 24 hours of a diagnosed SRC in collegiate athletes.3, 4 Despite the clinical utility of this multifaceted approach, a growing body of evidence suggests that physiological recovery from a SRC may persist well beyond apparent clinical recovery based on this accepted multi-faceted clinical evaluation. Microglia function to serve and protect the brain, and are wholly dedicated to maintaining homeostasis. Microglia phagocytose apoptotic neurons and secrete inflammatory factors to attract other immune cells to the site of injury. The resolution of this series of events is a tightly regulated process that occurs when injury has been resolved.5,6 However, if left unresolved, such as the case of chronic or pathologic inflammation, neuronal damage and loss of brain function can occur and impede recovery. More recent humanbased studies have investigated central inflammation using positron emission tomography (PET) and single-photon emission computerized tomography (SPECT) following a SRC7-9 targeting the translocator protein (TSPO), which is an 18 kDa transmembrane protein overexpressed in mitochondria of microglia as a result of neuronal insult. For example, Coughlin et al studied [11C] DPA-713 in active and retired NFL athletes and demonstrated increased distribution volume (Vt), as measured by PET, in the hippocampus, amygdala, supramarginal gyrus, and temporal pole compared to healthy age-matched controls. These studies present evidence that chronic neuroinflammation may be a consequence of individuals engaged in career contact sport. The purpose of this pilot study was to determine the presence of neuroinflammation beyond clinical recovery in active, collegiate athletes following a diagnosed SRC using [18F] DPA-714, a radiopharmaceutical specific for TSPO10. We hypothesized that collegiate athletes who were diagnosed with a SRC, but have been permitted to return to play, would show significantly higher Vt compared to healthy age-matched controls. We enrolled 10 athletes and 10 healthy age-matched controls (5 male and 5 female in each category). All subjects completed a standard clinical battery of concussion tests to achieve "return-to-play" status. Before returning to sport, subjects consented to T1-weighted MP-RAGE MRI and a 90 minute dynamic PET scan. Total distribution volume was derived and corrected for metabolites by blood sampling. Regional CNS comparisons between groups showed >75% differences in Vt across 11/12 brain regions. These data suggest neuroinflammation persists in collegiate athletes who have sustained a concussion, despite being cleared to return-to-play using the current clinical gold standard workup. Further studies confirming these findings in an expanded cohort are ongoing. imaging has been designed to assess the biomarker expression in a pre-operative setting, by 68Ga (or 44Sc)-PET and, in case of positive response, exploiting the NIRF moiety to aid surgeons intra-operatively to discriminate between tumor and healthy tissues during tumor resection. The synthetic pathway (2 steps) relied on an innovative synthesis approach based on the native chemical ligation. The first step consisted in the condensation between chelator, preactivated as thioester, and peptide, functionalized with an N-terminal cysteine, in buffer at pH around 7. Afterwards, a solution containing S-Cy5.5, pre-activated as maleimide, was added to the reaction mixture (step 2). Final purification was performed by semi-preparative reverse phase liquid chromatography, to obtain AAZTA-C4-CO-Cys(Sulfo Cy 5.5)-AE105 (1) with 99% of purity and an overall yield of 27%. Metal complexation with natural gallium isotope was easily carried out in 10 minutes using equimolar amounts of 1 and gallium chloride at room temperature in acetate buffer at pH 3.8. The uPAR affinity of Ga-AAZTA-C4-CO-Cys(Sulfo Cy 5.5)-AE105 (2) was evaluated by flow cytometry, showing a highest affinity for the U-87 MG glioblastoma cell line with respect to the control MCF-7. The behavior of the control cell line MCF-7 with probe 2 depicted lower fluorescence values, thus confirming the difference in the urokinase receptor expression of the two cell populations. The demonstration that the binding of probe 2 is driven by the targeting vector is provided by the observation that S-Cy 5.5 dye alone showed no affinity for U-87 MG glioblastoma cell line, with values below 5% in all concentrations range. A further confirmation of uPAR targeting was gained by performing blocking experiments with a 100 fold excess concentration of the AE105 peptide that highlighted a significant decrease in the number of cells that bind probe 2. Overall, these results underline that probe 2 recognize uPAR on the surface of U-87 MG glioblastoma cell line with a dosedependent behavior in the tested concentration range. A preliminary preclinical validation of the dual probe 2, aimed at assessing its ability to be taken up by uPAR expressing tumors, was carried out on mice grafted with U-87 MG cells. The animals were intravenously administered with 5 nmol of probe 2. NIRF images were acquired from 30 min up to 24 h post administration. The images provided a clear indication about the predominant renal elimination of the probe.

The tumor-to-background ratio, using muscles as reference, was comprised between 2 and 3 in the entire time window investigated, with a maximum tumor signal detected 2h post injection. After the last in vivo imaging acquisition, mice were sacrificed and ex vivo imaging of harvested organs showed that 24h post-injection probe 2 accumulated mainly in kidneys and tumors, and slightly less in liver and intestine. The almost null brain uptake is a relevant important finding if considering the use of this probe in surgical treatments of glioblastoma. In summary, a new potential dual PET/NIRF peptide-based imaging probe for the targeting of urokinase receptor, has been efficiently synthesized. The in vitro results confirmed the good affinity and specificity of the probe (complexed with "cold" Ga(III)), whose cell binding was proportional to the expression levels of uPAR. The preliminary in vivo evaluation of probe 2 was performed by NIRF imaging on a mouse model of glioblastoma (U87-MG) showing a good tumor uptake up to 24 h post injection, with a maximum tumor fluorescence signal detected 2h post injection. Fig 1A. The complex remained largely intact in serum for an hour, but integrity dropped to 61 % after 24 h and 44 % after 48 h (Fig 1B) . A PSMA bioconjugate of pypa was synthesised using a lipophilic linker, based on PSMA-617 ( Fig 1C) . [3] [201Tl]Tl(III)-pypa-PSMA showed excretion through the kidneys and bladder with little heart uptake observed, suggesting adequate in vivo stability for delivery to tumours. In contrast, Identifying novel targets for PDAC treatment is a major need for effective treatment strategies since PDAC exhibit limited response to chemo-and radiation therapy. Because pancreatic cancer cells are 'glutamine avid' [1] , targeting glutamine metabolic pathways can provide novel options for treatment. The glutamine transporter, SLC1A5, is being actively investigated as a pharmacological target in cancer [2] . Here we have engineered human pancreatic cells expressing shRNA to downregulate the glutamine transporter SLC1A5. We performed high-resolution 1H magnetic resonance spectroscopy (MRS) to understand the metabolic reprograming that occurs with SLC1A5 downregulation. Patient derived pancreatic cancer cells (Pa04C) were genetically engineered to express shRNA against SLC1A5 (Pa04C_SLC1A5) using lentiviral transduction. Downregulation of SLC1A5 was verified with q-RT-PCR and immunoblotting. Pa04C cells genetically engineered to express an empty vector (Pa04C_EV) using lentiviral transduction were used as controls. Tumors were derived by injecting 2x106 EV or Pa04C_SLC1A5 cells in the flank of 4 to 6-week-old male SCID mice. Tumor growth was monitored, and tumors excised and snap frozen at the end of the experiment. Dual phase extraction was performed on cells (EV n= 4, Pa04C_SLC1A5 n= 4) and tumors (EV n= 7, Pa04C_SLC1A5 n= 9), as previously described [3] . High-resolution 1H MR spectra were obtained from the aqueous phase extracts on a 750 MHz spectrometer. All data acquisition, processing and quantification was performed with TOPSPIN 3.5 software. We verified significant downregulation of SLC1A5 in Pa04C_SLC1A5 cells and tumors compared to EV cells and tumors. Downregulation of SLC1A5 also resulted in a significantly growth delay of Pa04C_SLC1A5 tumors compared to EV tumors. Metabolic heat maps generated from quantitative 1H MRS, of cell and tumor aqueous extracts, identified significant downregulation of the glutamine/glutamate ratio in cells with SLC1A5 downregulated (Figure 1 A) . Other than the significant decrease of glutamine/glutamate in the Pa04C_SLC1A5 cells, no other significant metabolic differences were detected in these cells. In tumors, however, SLC1A5 downregulation resulted in a significant decrease of branched chain amino acids (BCAA), valine, alanine, glutamate, acetate, aspartate, glutathione, choline, phosphocholine, glycine, tyrosine, phenylalanine and histidine (Figure 1 B) . SLC1A5 downregulation clearly resulted in significant tumor growth delay, supporting its importance as a therapeutic target. Our data expand the understanding of the diverse metabolic reprogramming that occurs following SLC1A5 downregulation in tumors in vivo that may lead to the development of additional metabolic targets. These data also identify the role of the tumor microenvironment in the metabolic reprogramming identified in tumors. Supported by NIH R01CA193365 and R35CA209960. D-Met showed clearance from both hepatobiliary (slow) and urinary (rapid) pathways. Early background uptake was observed in the liver, pancreas and kidneys with minimal background uptake in other organs. Importantly, minimal uptake was observed in the GI tract or lungs, despite potential concerns regarding the uptake within the human microbiomes. Additionally, minimal background uptake was observed in the spine, CNS or musculoskeletal system. In 2 

Mengtong Duan, Caltech, mduan@caltech.edu Category: Bioengineering, Synthetic Biology and Basic Biology Abstract Body : Ultrasound is an established technology for noninvasive deep-tissue imaging that overcomes the low penetration depth of optical methods. However, until recently it was impossible to couple ultrasound contrast to gene expression or cellular function. Gas vesicles (GVs) were recently introduced as the first genetically encoded ultrasound contrast agents expressed from acoustic reporter genes (ARGs) in bacteria [1] and mammalian cells [2] . However, the first-generation mammalian acoustic reporter genes (mARGs), based on 9 essential GV genes from a soil bacterium B. megaterium, suffer from a number of limitations: linking individual GV genes into compact mammalian polycistronic expression cassettes severely attenuates GV expression and ultrasound contrast compared to the cotransfection of individual genes, both transient and stably transduced mARGs require coexpression of three separate polycistronic cassettes for GV production, and the acoustic properties of B. megaterium GVs only allow imaging using a destructive ultrasound modality, limiting their use to endpoint measurements. Here we present a new generation of mARGs that overcome the abovementioned limitations. We demonstrate that mARGs2.0 provide strong expression from only two polycistronic expression cassettes with optimized cassette ratio, nonlinear acoustic properties that allow for nondestructive continuous imaging using amplitude modulation pulse sequences [3, 4] , and the ability to modulate acoustic properties based on gene composition. We extensively characterize mARGs2.0 in vitro and demonstrate their utility in vivo by imaging gene expression in an orthotopic cancer model. For this purpose, we stably integrated these reporter genes into the genomes of human breast cancer cells and imaged them nondestructively in vivo after tumor formation in mouse mammary fat pads and chemically induced expression. Second, other regions of signal void lead to low specificity. Magnetic particle imaging (MPI) cell tracking has the potential to overcome these challenges. MPI detects the response of SPIONs to an applied magnetic field. The MPI signal appears as a hot spot with no background, providing high specificity. MPI signal is proportional to iron content and directly quantifiable, providing measures of iron mass and cell number. Sensitivity and resolution in MPI depend heavily on the magnetic properties of the SPION. In this study, we evaluate a SPION called Synomag-D™, a nanoflower-shaped multi-core particle, for the detection and quantification of metastatic cancer cells by MPI. To the best of our knowledge, Synomag-D has not yet been used to label breast cancer cells for imaging with MPI. Objective: To utilize Synomag-D for quantitative MPI of breast cancer cells disseminated in the mouse brain and compare to our previous work where breast cancer cells were labeled with micron sized iron particles (MPIO) and imaged by MRI. Methods: Human brain metastatic breast cancer cells (231BR) were labeled with Synomag-D (MicroMod GmbH). Synomag-D has a core size of 25-30 nm (individual cores are 8-9 nm) which is considered ideal for MPI based on relaxometry and simulations6. Labeling was validated with a Perl's Prussian Blue (PPB) stain for iron and MPI of cell pellets. MPI relaxometry was acquired for triplicate samples of 1uL of Synomag-D, 10uL of MPIO, and 5.0x105 231BR cells labeled with Synomag-D or MPIO. NOD/SCID/ILIIrg−/− mice (n=6) were injected with 2.5x105 cells intracardially (IC). MPI was performed on a MOMENTUM scanner (Magnetic Insight Inc., Almeda, CA, USA) on Day 0 and 7 (n=6) using the 3D high sensitivity isotropic (multichannel) scan mode. MPI data was compared to MRI from a prior study using mice injected IC with 2.5x105 MPIO-labeled 231BR cells. Results: Synomag-D labeled cells with an efficiency of 98.0±0.40% (Fig. A) and 98% viability. Mean iron/cell was 5 pg. There was a linear relationship between iron content and MPI signal (R2 = 0.96, p On Day 0, MPI signal was detected in 3/6 mouse brains (Fig. D) ; the mean iron content was 0.08 ug, which is ~16,000 cells in the mouse brain. MPI signal was also visible in the lung/liver region since cells injected IC are distributed throughout the body. Strong MPI signal may impede detection of low signal in nearby regions; and may result in no signal being detected in 3/6 mice on Day 0. On Day 7 lung/liver signal was diminished and brain signal was detected in 5/6 mice. The signal decreased from Day 0; mean iron content was 0.012ug (~2,400 cells) (Fig. F) . This is expected, as cancer cells die and clear by Day 7 in this model1. In Day 0 MRI, MPIO-labeled cells appear as discrete signal voids (Fig. E) . Quantification of these images is challenging; voids are counted or the percentage of black pixels calculated, however, this does not allow for determination of cell number. Conclusions: This is the first study to demonstrate that brain metastatic breast cancer cells can be labeled with Synomag-D and quantified in vivo with MPI. Employing MPI for experimental cancer cell tracking allows for the detection and quantification of the arrest, clearance, and retention of cancer cells in vivo. Abstract Body : Optical technologies are increasingly important in modern biological research, but identifying sources and targets of optical illumination is challenging in living tissue. Here we introduce liposomal nanoparticle reporters (LisNRs), a novel nanoprobe design that addresses this limitation permitting noninvasive volumetric mapping of illumination profiles in vivo using magnetic resonance imaging (MRI). The LisNR mechanism leverages reversible light-dependent changes in the water permeability of a liposomal membrane encapsulating a large reservoir of paramagnetic contrast agent. This system permits T1 contrast from a large number of encapsulated contrast agents to be modulated simultaneously by a relatively small number of photoisomerizable molecules, enabling amplified light detection. We implement this general MRI contrast mechanism for light detection through the incorporation AzoPC, an azobenzenecontaining phospholipid that promotes high bilayer permeability in its ultraviolet light-favored cis conformation but low permeability in its blue light-favored trans state. We screen a number of lipid compositions to determine optimal performance for in vitro light sensing in MRI. The resulting light LisNR displays maximal T1 relaxivity changes up to 200% with sensitivity to total incident photon counts of approximately 1016. We apply the nanosensor in vivo to measure three-dimensional illumination profiles characteristic of widely used optogenetics and photometry applications in the rat brain. Light exposure as short as 2 minutes, with fiber tip power of 0.3 mW corresponding to < 1 017 total incident photons, generated mean signal changes of approximately 10% and all observed responses were completely reversible over multiple light cycles. Our results reveal previously undescribed spatial profiles of light propagation arising in the presence of physiological tissue microstructure. This work introduces a general basis for mapping light in opaque environments, as well as a potent and adaptable sensor architecture for MRI. Abstract Body : Background: Mounting evidence indicates that innate immune dysfunction, involving chronic activation of macrophages and microglia, plays a critical role in the onset and progression of many neurological diseases. While PET imaging provides a potential noninvasive method to visualize and quantify innate immune activation in the brain and periphery, most available PET tracers are not specific for macrophages/microglia. We identified OP-801, a new synthetic PAMAM hydroxyl dendrimer that crosses the blood-brain barrier in the presence of inflammation and is selectively (>95%) taken up by reactive macrophages/microglia, as a candidate agent to radiolabel. Here we aim to 1) develop the first hydroxyl dendrimer PET tracer targeting reactive macrophages/microglia and 2) evaluate its ability to detect inflammation in the brain and periphery of mice following lipopolysaccharide (LPS)-induced sepsis. Methods: OP-801 was radiolabeled using a 2-step azide fluorination and copper-catalyzed click reaction.Formulated tracer (150-250 μCi, 0.04 ng/μCi in saline and 10% ethanol) was injected into female C57BL/6 mice by tail vein, 24 hours after intraperitoneal administration of LPS (10 mg/kg, n=14) or saline (n=6), and 60 minute dynamic PET/CT images were acquired. Brain PET images were quantified by co-registering PET and CT images followed by fitting a standard brain atlas within the skull of each mouse. Symptoms were assessed using the murine sepsis score (MSS) [1] and three LPS mice with negligible symptoms excluded from subsequent analysis. After imaging, all mice underwent cardiac puncture followed by perfusion with saline. Organs were dissected, weighed, and placed into a gamma counter to measure radioactivity. Tracer distribution within the brain was further evaluated using high resolution autoradiography. Brain slices used for autoradiography were subsequently stained with cresyl violet to enable quantification of 18F-OP-801 uptake in specific brain regions; one saline and one LPS animal were excluded from biodistribution analysis due to poor perfusion. Ex vivo plasma stability of 18F-OP-801 was evaluated at 60-and 150-minutes post-injection (n=4/group). Results: 18F-OP-801 was afforded in >4.7±2.66% yield (n=4 syntheses, decaycorrected) with >95% radiochemical purity. Brain PET/CT images (summed 50-60 minutes) revealed linearly increasing 18F-OP-801 uptake, corelating with MSS [1] (R=0.82, p=0.001). Quantitation of whole brain time activity curves demonstrated significantly elevated signal in LPS-versus saline-injected mice (p 18F-OP-801. Brain atlas analysis of 50-60 min summed PET images showed that uptake differed significantly (p x vivo biodistribution data confirmed PET findings for LPS (n=6) vs. saline (n=5) mice, showing significant differences in organs expected to have elevated reactive macrophages, including liver (2.7±2.20 saline vs. 27.0±30.11 LPS %ID/g, p=0.009), lung (2.9±4.34 vs. 26.2±26.47 %ID/g, p=0.030) , and spleen (2.0±1.60 vs. 21.4±18.24 %ID/g, p=0.004). Importantly, ex vivo plasma stability of 18F-OP-801 was >95%, on average, after 150 minutes. Conclusion: 18F-OP-801 is a new PET tracer with extraordinary potential for highly sensitive and specific detection of activated macrophages/microglia in the whole body and brain. Based on our promising proof-of-concept data, we plan to evaluate the utility of 18F-OP-801 for imaging innate immune activation in mouse models of Alzheimer's disease and multiple sclerosis. In parallel, we are working toward clinical translation of 18F-OP-801 to use for patient stratification and monitoring responses to novel immunomodulatory therapies.

References: Shrum, B., et al. BMC Res Notes. 2014, 7(233) . 

Oluwatosin Ibhagui, Georgia State Uuniversity, oodubade1@student.gsu.edu Category: Systemic Diseases (Kidney, Liver and Pancreas) Abstract Body : Acute and chronic lung diseases including COVID-19, idiopathic pulmonary fibrosis (IPF), and chronic obstructive pulmonary disease (COPD) characterized by diffuse, progressive remodeling of the lung parenchyma with increased extracellular matrix (ECM) deposition, including collagen within the lung with the worst prognosis is a major health crisis affecting over half a billion patients globally. Early and noninvasive detection and staging of these lung diseases are pressing unmet medical needs limiting the potential for treatment monitoring and possible reversal of disease progression. Here we used magnetic resonance imaging with a new human collagen-targeted protein MRI contrast agent (hProCA32.collagen) to progressively detect and quantify fibrosis using both idiopathic pulmonary fibrosis and COPD animal models. hProCA32.collagen exhibits very strong collagen I binding affinity and is able to specifically bind to collagen 1 overexpressed in lung fibrosis of human patient. hProCA32.collagen possesses high relaxivities per particle (r1 and r2) at both 1.4 and 7.0 T enabling early detection by multiple imaging techniques of IPF using our established progressive bleomycin-induced IPF mice model closely mimicking key human morphological features of IPF including reticular opacities, traction bronchiectasis, and honeycombing-like cystic clustering. Both conventional (RARE and T2 weighted MRI) and ultra-short 3D (3D-UTE) MRI pulse sequences were able to detect overexpressed collagen in lung fibrosis with strong sensitivity and specificity. SNR values obtained by molecular MRI correlates well with the amount of histological lung collagen and hydroxyproline content in fibrotic mice. Furthermore, the enhancement pattern of hProCA32.collagen closely resemble the histological morphology of collagen overexpression in mice. The observed MR image enhancement was progressive with 60 % and 139 % SNR increase in early-stage IPF mice (n=4) and late-stage IPF mice (n=6) respectively. Signal retention up to 80.5% was also observed in the more fibrotic late-stage mice compared to healthy mice indicative of selective collagen binding of hProCA32.collagen. This observation was absent in the MR image obtained after injection of clinical contrast agent Gadovist® and our developed non collagen targeted contrast agent hProCA32. This enhancement specificity was further demonstrated in a second mouse model of chronic obstructive pulmonary disease (COPD) with nicotine-induced bronchial collagen accumulation with lung MRI enhancement in the bronchi structure of mice in contrast to alveolar distribution observed in IPF mice model. The low dose attribute of hProCA32.collagen helps mitigates metal toxicity and is expected to overcome the major clinical barriers in early diagnosis, noninvasive detection and staging of lung diseases with have strong translational potential in facilitating effective treatment to halt further chronic lung disease progression. Healthy mice before, 1 h, 3 h and 24 h after injection of hProCA32.collagen; (B) mice treated with bleomycin for 3 weeks (early stage) pre, 1h, 3h and 24 h after injection of hProCA32.collagen with heterogeneous enhancement of collagenous honeycombing at the sub plueral and paraseptal areas of the lung (black arrows) (C) 6 weeks BM treatment (late stage) pre, 1h, 3h and 24 h after injection of hProCA32.collagen with heterogeneous enhancement of collagenous honeycombing at the sub plueral and paraseptal areas of the lung (black arrows) with traction bronchiectasis (arrow heads) (D) 6 weeks BM treatment (late stage) pre, 1h, 3h and 24 h after injection of non-targeted hProCA32 with homogenous and rapid enhancement of collagenous honeycombing at the sub plueral and paraseptal areas of the lung at the 1 h time point (black arrows), at the 3 h time point, a noticeable wash out of hProCA32.collagen was observed. (E) 6 weeks BM treatment (late stage) pre, 30 mins, 3h and 24 h after injection of Gadovist® with no significant enhancement in the lung F. Lung SNR plot of healthy, early stage and late stage IPF mice before and after injection of the MRI contrast agents. H, I Scatter-plots of lung collagen area and modified Ashcroft score against AUC_0-24 respectively showed a linear correlation in BM-induced IPF model. I., Histogram pixel analysis of normal mice before and at peak enhancement (1 h) after injection of hProCA32.collagen, J. 6 weeks BM-treated mice before and at peak enhancement (1 h) after injection of gadovist, K. 3 weeks BM-treated mice before and at peak enhancement (3 h) after injection of hProCA32.collagen with heterogeneous lung SNR distribution compared to the pre time point, the heterogeneity was also observed for 6 weeks treated mice injected with hProCA32.collagen (L). [18F] NaF in a rat model of VC was performed to test this hypothesis. A model of VC was used in which rats were fed a diet containing warfarin and vitamin K1 along with subcutaneous administration of vitamin D3 to induce severe VC.3 Anaesthetised rats were injected with [68Ga]Ga-THP-Pam and scanned using preclinical PET/CT 60-120 min postinjection. The rats were imaged using the same protocol with [18F] NaF the following day. As a control study, animals fed a healthy diet were imaged using the same procedure. Organs were harvested and their radioactivity was measured for ex vivo biodistribution. Organs of interest were fixed in formalin and embedded in paraffin. Paraffin-embedded organs were scanned using μCT. Imaging ( Figure 1A ) showed high uptake of [68Ga]Ga-THP-Pam and [18F] NaF (3.44 ± 0.69 and 0.91 ± 0.24 %ID respectively, p = 0.002, Figure 1B ) in a region of tissue around the stomach, with severe calcification as identified by CT. Additionally, [68Ga]Ga-THP-Pam demonstrated increased uptake in the VC group vs. the healthy group across several major organs, most notably in the kidneys (2.21 ± 0.76 vs. 0.25 ± 0.13 %ID, p = 0.002, Figure 1B ). Ex vivo biodistribution data confirmed the increased uptake of [68Ga]Ga-THP-Pam observed in the imaging data. The presence of calcification in the kidneys, stomach and other organs of interest was confirmed by microCT-based 3D X-ray histology (XRH) and conventional histology ( Figure  1C ). To visualise small areas of interest such as the aorta, a prototype post-reconstruction method to improve the spatial resolution of preclinical PET with gallium-68 was used. Analysis of the mineral composition of the calcifications is ongoing. These results demonstrate that [68Ga]Ga-THP-Pam may offer improved detection of VC in comparison to [18F] NaF, including microcalcifications undetectable by preclinical CT. Abstract Body : Introduction: Therapeutic mesenchymal stem cells (MSC) have great potential for regenerative medicine. After administration, many MSC die due to a hostile, proinflammatory environment and unfortunately, cases of graft failure are only detected after several months based on lack of tissue repair1. Early detection and intervention for failed MSC therapies could be made possible with sensitive and quantitative imaging. Magnetic particle imaging (MPI) with superparamagnetic iron oxide nanoparticles and fluorine-19 (19F) magnetic resonance imaging (MRI) with perfluorocarbon nanoemulsions both produce images which allow for quantification of cells2. MPI is an emerging instrument for cell tracking which is expected to have superior cellular sensitivity compared to 19F MRI, however, this has not been carefully compared. The objective of this study is to assess the cellular sensitivity of MPI and 19F MRI for detection of MSC. Methods: Cell Preparation: MSC were labeled by overnight coincubation with 55µg Fe/mL ferucarbotran and transfection agents3 or 2.5mg/mL perfluoropolyether (PFPE). Cell pellets were produced by serial dilution from 1.024x106 to 1000 cells (n=9 per cell number). Ferucarbotran-labeled MSCs (1x105) or PFPE-labeled MSCs (2x106) were injected to immunocompromised mice by subcutaneous or intraperitoneal administration (n=4 total). Imaging protocol: Ferucarbotran-labeled cells (pellets and in vivo) were imaged with a preclinical Momentum MPI scanner using weak (3T/m) gradients in 2D (1.5 min) and 3D (22.8 min) . 19F MRI of PFPE-labeled MSC were acquired on a 3T clinical MRI using a 4.3cm dual-tuned surface coil and a 3D balanced steady state free precession sequence. Cell pellets were imaged in 1.5 minutes (same as MPI) and 4.5 minutes, and in vivo 19F MRI was 35 minutes. MPI and 19F MRI in vitro cell detection limits were defined as the minimum number of MSC detected with signal-to-noise ratio (SNR) >5. Quantification of MPI and 19F signal was calculated as mean signal intensity times the area of the region of interest. Validation: Ferucarbotran labeling was validated with Perl's Prussian blue stain and by measuring iron mass per cell by MPI2. PFPE labeling was validated by microscopy and by measuring 19F atoms per cell by NMR4. Results: Cell number was correlated with MPI signal (R2 = 0.9988) and 19F signal (R2 = 0.9865). For MPI, MSC labeled with 19.09±2.50pg Fe/cell and as few as 4000 MSC could be reliably detected in 2D (SNR=5.5, 8/9 replicates). Improved sesitivity was seen in 3D; 2000 MSC were detected (SNR=6. 6, 9/9 replicates) and as few as 1000 MSC could be visualized (SNR 1.55 x1011 19F/cell and the range of 256-1024 x103 MSC were detectable in short scans (SNR=7.12, 1.5 min/pellet). Fewer cells (128x103) were detected with longer imaging times (4.5 min/pellet, SNR=5.9, 2/9 replicates). In vivo MPI showed a reduction in MPI signal measured from cells injected subcutaneously (53%) and intraperitoneal (49%) compared to signal from a pellet of 1x105 MSC. With 19F MRI, the same amount of signal was detected from 2x106 cells in pellets (1.1x1018 19F spins) and after subcutaneous (9.3x1018 19F spins) and intraperitoneal (1.1x1018 19F spins) injections. Conclusions: While impossible to make a direct comparison between these cellular imaging modalities, we detected fewer MSC with MPI than with 19F MRI, using the same scan time. There is no doubt that cellular sensitivity will continue to improve with cellular uptake strategies and optimized versions of cell labels. We also found differences in sensitivity in vivo compared to in vitro. The dispersion of cells in vivo can result in some cells to fall below the intravoxel detection limits. Additional factors in vivo that influence cell detection include motion, effects of Brownian relaxation (MPI), and the coil filling factor (19F MRI). To advance imaging of cellular therapeutics, it is essential to continue to understand and improve cellular sensitivity.

References: [1] Nejadnik, H et al. Wires Nanomed Nanobi 11 (2019) . [2] Sehl, OC et al. Tomography 6 (2020) . [3] Thu, M.S et al. Nat Med 18 (2012) . [4] Fink, C. Sci Rep 8:590 (2018) . Abstract Body : Introduction Recent advances in CRISPR/Cas9 technology have enabled its use as a powerful tool for in-vivo gene editing. Previously, we have validated this technique in-vivo by assessing vesicular monoamine transporter 2 (VMAT2) knockdown (KD) effects using [11C] DTBZ and [11C] raclopride PET, behavioral experiments and ex-vivo analyses [1] . However, the effects of targeted gene editing using CRISPR/Cas9 on whole-brain function have not been studied to date. Here, we targeted the Slc18a2 gene, encoding the VMAT2, using Cas9 from Staphylococcus aureus (SaCas9) in rats. We performed simultaneous PET/fMRI scans to elucidate the effects of gene editing on both local molecular and network-level functional connectivity (FC) changes. Methods Baseline [11C] raclopride PET/fMRI scans were acquired in adult rats (n=23) under medetomidine (0.05mg/kg bolus+0.1 mg/kg/h infusion) and 0.5% isoflurane in air over 60 min using a 7T small-animal MRI with a PET insert. The tracer was applied as a bolus and fMRI was acquired using an EPI-BOLD sequence (TR=2500ms, TE=18ms). AAV-SaCas9 and AAV-sgRNA-Slc18a2 were next injected into the right substantia nigra pars compacta (SNc) and DPBS into the left SNc. The rats were rescanned 8-14 weeks after using the same PET/fMRI protocol and additional PET scans were performed using [11C] DTBZ and [18F] FMZ. Following standard preprocessing, the PET data were assessed for molecular changes and fMRI was used to evaluate effects in default-mode (DMN) and sensorimotor (SMN) network FC. Results and Discussion Based on the [11C] DTBZ scans we split the rats into low knockdown (KD 20%, n=10) groups ( Figure 1A) . On a molecular level, we found correlations between VMAT2 KD and D2R expressions (R2 = 0.52, p Conclusion Our study is the first to assess CRISPR/Cas9 gene editing as an in-vivo method for the investigation of not only local, but also network-level brain function. While we detected functional changes at both low and high KD, we could robustly separate the two cohorts, thus recommending the used multimodal study design for future research of both early and advanced disease phenotypes. Therefore, we demonstrated the tremendous potential the combination of CRISPR/Cas9 and PET/fMRI imaging has to bridge the gap in understanding the link between targeted gene alterations and whole-brain functional changes. Abstract Body : Co(II) shift agents loaded into liposomes produce lipoCEST effects comparable to that of liposomes containing Tm(III) and Dy(III) shift agents. [1, 2] These paramagnetic complexes shift the intraliposomal water proton resonance, thereby producing lipoCEST (liposomal CEST) agents due to the exchange of intra and extra liposomal water molecules through the bilayer membrane. Shift agents were encapsulated in the intraliposomal interior or as amphiphilic complexes in the liposomal bilayer. In order to prepare improved lipoCEST agents, new Co(II) complexes that are highly soluble with an overall neutral or cationic charge complexes have been employed. These agents are Co(II) octahedral coordination complexes of the 14-membered tetraaza cyclam macrocycle with bisbenzyl and bisamide/bis-N-glycinamide pendants appended to the cyclam amines as shown in figure 1. Neutral shift agents have lower tonicity than charged analogues imparting higher osmotic shrinkage of the shift agent loaded liposomes when placed in aqueous solutions of similar osmolarity. This yielded a lipoCEST peak 5 ppm away from bulk water at a relatively low saturation power of 6 uT in isotonic 300 mOsm solution for a neutral Co(II) complex loaded into 100 nm liposomes. Being saturation power dependent, the lipoCEST peak was found to shift further with an increase in saturation power. Asymmetry of Z-spectra was plotted to calculate the position of LipoCEST peak against bulk water peak. Moreover, larger shrinking and a corresponding shift in the position of the lipoCEST peak were observed with hypertonic solutions due to an increased bulk magnetic susceptibility (BMS) effect from the more shrunken liposomes. Integration of a freeze-thaw step during the liposome formulation after the extrusion enhanced the core loading efficiency of the liposomes and so the concentration of the intraliposomal shift agent was higher than the liposomes made without freeze-thawing. The encapsulation efficiency of the shift agent inside the liposomes increased from 4.0% to 37.8% with 5 cycles of freezing in liquid nitrogen and thawing at 55 °C. The lipoCEST peak from one such formulation was shifted 9 ppm away from bulk water at 6 uT saturation power under similar isotonic condition. The position of the lipoCEST peak further shifted from 9 ppm to 12 ppm with an increase in saturation power (15.4 uT). Additionally, two amphiphilic Co(II) complexes containing charged or neutral Co(II) complexes are being investigated with the expectation of producing larger shifts and a greater magnitude of the CEST peak. These two contrast agents both have two stearyl long chain pendants alkylated to the macrocycle amines. These complexes will be integrated into the liposomal membrane for liposomes core loaded with hydrophilic shift agents. Both relaxivity and lipoCEST data from these liposomes are employed to image phantoms or animals with these MRI probes.

Julia Stahl, Werner Siemens Imaging Center, julia.carolin.stahl@med.uni-tuebingen.de Category: Immunology: Inflammation & Infection Abstract Body : Introduction: Non-invasive in vivo tracking of T cells enables to follow their migration and homing dynamics in cancer and autoimmune diseases. We aimed to establish simultaneous 89Zr-DFO-CD8-minibody and 19F-Perfluorocarbon (PFC) PET/MR imaging in T cell driven acute and chronic contact hypersensitivity reaction (CHSR). CHSRs are orchestrated by CD8+ and CD4+ T cells and are accompanied by immigration and activation of phagocytic cells. The aim of our study was to distinguish the differential impact of CD8+ T cells and phagocytes in acute and chronic CHSRs by 89Zr-DFO-CD8-minibody/19F-PFC PET/MR imaging of NF-κB.p52-/-mice, which are unable to develop chronic CHSR, and in wild-type mice. Methods: Acute and chronic CHSR was induced by one (acute CHSR) or five (chronic CHSR) repetitive trinitrochlorobenzene (TNCB) challenges of the right ear of TNCB-sensitized C57BL/6J and C57BL/6-NF-κB.p52-/-mice. The left ear served as healthy control. For in vivo PET/MR investigations we used a 89Zr-DFO-murine CD8-specific minibody (IAB42, ImaginAb, USA) and 19F-PFC to track CD8+ T cells and phagocytes namely neutrophils, macrophages, and monocytes non-invasively in vivo with simultaneous PET/MR. Imaging studies were performed before, 12h and 24h after the 1st and 5th TNCB-ear-challenge (n = 5). For ex vivo cross-validation we conducted 89Zr-DFO-CD8-biodistribution, immunohistochemistry (IHC) and flow-cytometry analysis (FACS). Results/Discussion: Our in vivo PET/MRI studies in wild-type mice revealed a strongly, up to 14-fold increased 89Zr-DFOmurine-CD8-minibody uptake in ears with chronic CHSR (2.73 %ID/cc, 24h) and no increase in ears with acute CHSR (0.19 %ID/cc, 24h) when compared to the healthy left ears (acute CHSR: 0.19 %ID/cc, 24h; chronic CHSR: 0.37 %ID/cc, 24h). In the draining cervical lymph nodes (dLNs) of mice with chronic CHSR the 89Zr-DFO-murine-CD8-minibody uptake (8.20 %ID/cc, 24h) was 3-fold enhanced compared to mice with acute CHSR (2.61 %ID/cc, 24h). In inflamed ears of NF-κB.p52-/-mice with chronic CHSR we determined an approximately 50% reduced 89Zr-DFO-CD8-minibody uptake (1.47 %ID/cc, 24h) when compared to wild-type mice (2.73 %ID/cc) and only a moderate reduction in the dLNs of NF-κB.p52-/-mice (5.98 %ID/cc, 24h) when compared to wild-type mice. Focusing on phagocyte tracking, we identified a strongly enhanced 19F-PFC-uptake exclusively in inflamed ears of wild-type mice with chronic CHSR (7.51% PFC uptake normalized to standard, 24h) but not in ears with acute CHSR (3.59%). Interestingly, we identified hardly any differences in the 19F-PFC-uptake in inflamed ears of NF-κB.p52-/-mice (7.11%) when compared to wild-type mice (7.51%). In the dLNs of NF-κB.p52-/-mice (11.67%) with chronic CHSR we detected an even higher PFC uptake when compared to wild-type mice (8.48%). Most importantly, extensive ex vivo biodistribution and FACS analysis as well as IHC confirmed our in vivo PET/MRI data. Conclusions: Highly innovative simultaneous non-invasive in vivo 89Zr-DFO-CD8-minibody-and 19F PFC PET/MR imaging enables differential tracking of CD8+ T cells and phagocytes in secondary lymphatic organs and the sites of acute and chronic CHSR. Thus, this imaging approach is applicable to gain more profound knowledge about the dynamics of cellular processes during disease progression and facilitates dual monitoring of novel anti-inflammatory treatment approaches. Acknowledgement: This work was supported by the Horizon2020 Programme under grant agreement n° 675417 (PET3D) and the EU/EFPIA/Innovative Medicines Initiative 2 Joint Undertaking under grant agreement n° 831514 (Immune-Image). The ovarian cancer is one of the gynaecological tumor with the highest mortality rate1. The aim of this study is to improve the therapeutic effect of the chemotherapy based on theranostic PEGylated liposomes co-loaded with doxorubicin (Doxo) and Gadoteridol using low intensity pulsed ultrasound (LIUS) on ovarian cancer xenograft mouse model. MRI is a technique of election to monitor the tumor progression and visualize the Gd-based agent that can report on the US-triggered drug release and predict the long-term therapeutic outcome. Sonosensitive theranostic liposomes (Gd-Doxo-Lipo) were prepared by using the film hydration method (lipid composition: DPPC/DSPC/ Chol /DSPE-PEG2000, molar ratio 10/5/4/1) and loaded with doxorubicin 1 mg/mL and Gadoteridol 300 mM.2 Female athymic nude mice Foxn1-/ Foxn1-were purchased by Envigo (Milano, Italy) and were inoculated with human ovarian cancer cells (A2780). Tumor volume were evaluated by T2w-MR images acquired at 7 T. Mice were divided in three groups i) not treated group ii) group treated with Gd-Doxo-Lipo, and iii) group treated with Gd-Doxo-Lipo and US local application, T1w MR images were acquired before and after the treatment. The US treatment was performed using US waves generated by a piezoelectric transducer (1 MHz, intensity < 5 W/cm2) purchased by Precision Acoustics (Dorset, UK). The amount of Doxo and Gadoteridol in the tumour was assessed spectrofluorimetrically and by ICP-MS, respectively. After the individuation of the best the US setup conditions for the maximal release of the drug and the Gd-based agent from liposomes, the most appropriate US parameters to obtain the internalization in the tumour cell membrane were optimized in vitro and in vivo. Gadoteridol and Doxo showed a similar ability to be released from the liposomes and be internalized by sonoporation. The tumour US stimulation resulted in a T1-contrast enhancement in the tumour and in the kidney calix and bladder, thus indicating the effective intratumor release of gadoteridol and drug from the liposomes. The tumour uptake of the Gd-agent and the drug was confirmed by confocal images and quantification (Fig. 1) . Notably, the group of mice treated with Gd-Lipo-Doxo and US-stimulated showed an arrest in the tumor progression, and in some cases the lesion completely remitted. Interestingly, the tumour T1 contrast enhancement measured just after the US treatment showed a positive correlation with the therapeutic outcome after three weeks. Low intensity US stimulation seems to be a very appealing method to improve the chemotherapy efficacy of a liposomal formulation of doxorubicin in a mouse model of ovarian carcinoma. The three components of the theranostic agents (liposomes, Doxo, and Gadoteridol) are already used in clinic, making this approach suitable for clinical translation. Moreover, MRI confirms to be an excellent imaging technique to monitor and predict the therapeutic effect. , 0.8 mg of precursor in 100 mL ethanol was added, followed by 100 mL 1N HCl and 100 mL of 3% H2O2. After 15 mins at room temperature reaction was quenched with 100 mL NaHSO3. The reaction mixture was extracted with dichloromethane and purified by chromatography. Human AD post-mortem brain slices (10 μm thick cut using Leica 1850 cryotome) consisting of anterior cingulate (AC) and corpus callosum CC as well as temporal cortex were used for in vitro binding studies. Brain slices were incubated with [125I]IAZA in 50% ethanol PBS buffer pH 7.4 (60 mL;0.02 μCi/mL) at 25 oC for 1.25 hr. Nonspecific binding was measured in the presence of 5-OH-BTA-0. The slices were then washed with cold PBS buffer, 50% ethanolic PBS buffer, 90% ethanolic PBS buffer, PBS buffer and cold water. Brain sections were dried, exposed overnight on phosphor film, and placed on the phosphor films and analyzed using Optiquant software and extent of binding of [125I]IAZA was measured in DLU/mm2. Results: The single step radiosynthesis of [125I]IAZA was very efficient. Product was purified on preparative thin layer chromatography, extracted and the final product taken up in absolute ethanol. Radiochemical yields (>50%) were obtained with a radiochemical purity of >95%. Specific activity based on [125I]iodide was >15 Ci/mmol. For Ab plaques, slices from subjects were positively immunostained with anti-Ab Biolegend 803015 (Biolegend, CA, USA) which is reactive to amino acid residue 1-16 of bamyloid. Selective binding of [125I]IAZA was observed in grey matter regions shown to contain the Ab plaques. The ratio of anterior cingulate to corpus callosum was >20. Similarly, temporal cortex exhibited significant binding of [125I]IAZA. Very little white matter binding was seen when 90% alcohol was used in the washing of the slices. Washing with PBS buffer or even with 50% alcohol was insufficient to remove the nonspecific white matter binding. Incubation with 5-OH-BTA-0 did not reduce Abstract Body : Cancer is the second-leading cause of death in the United States, next to cardiovascular disease. Glioblastoma multiforme (GBM) is the most common and lethal primary central nervous system (CNS) tumor. Despite increased attention, GBM treatment regimens prolong median survival rate by only several months, and long-term patient prognosis remains poor. While the majority of research has gravitated toward profiling/targeting tumor-specific molecules, altered metabolism in tumors is implicated heavily in oncologic survival and progression. Healthy maintain strong ion gradients across the cell membrane. Sodium (Na+) and chloride (Cl-) are high (~150 mM) in blood and extracellular spaces, and low (~15 mM) in intracellular milieu; potassium (K+) concentrations are reversed. These balances are responsible for critical functions like nerve signaling and osmoregulation. Na+ levels in blood (Na+b), extracellular (Na+e), and intracellular (Na+i) spaces specifically engender weak transendothelial (∆Na+end = Na+b -Na+e) and strong transmembrane (∆Na+mem = Na+e -Na+i) gradients. These are involved, respectively, in the integrity of the blood-brain barrier (BBB) and maintenance of the cell membrane potential (Vm). Cancer cells, however, exhibit altered metabolism resulting in inefficient energetics and disturbed gradients, particularly ∆Na+end and ∆Na+mem. Current methods to determine specific compartmental Na+ concentrations are undesirable, requiring biopsies, insertion of voltammetric microelectrodes, and other invasive techniques which preclude human translation. Nuclear magnetic resonance (NMR) has been shown to non-invasively detect the 100%-abundant isotope sodium-23 (23Na), but cannot resolve the degenerate signals from different physiological compartments. Traditional 23Na magnetic resonance imaging (MRI) attempted to isolate specific Na+compartments by exploiting relaxation differences, but these methods lacked complete compartmental specificity. 23Na magnetic resonance spectroscopic imaging (MRSI), following administration of a paramagnetic lanthanide-based contrast agent consisting of a lanthanide(III) ion (Ln3+) complexed with a macrocyclic chelate, was shown to induce compartmental chemical shift differences and resolve the 23Na-NMR signal into separate peaks which could be quantified individually with greater specificity. This 23Na-MRSI technique was applied in vivo in several rat GBM models (RG2, U87, and U251) using the thulium(III) (Tm3+) salt of the chelate 1,4,7,10tetraazacyclododecane-1,4,7,10-tetrakis(methylenephosphonate) (DOTP8-), or TmDOTP5-. The sensitivity of 23Na chemical shift to TmDOTP5-concentration was found to be ~2.1 ppm/mM in vitro, which manifested in three separated peaks in vivo upon intravenous infusion. The shifting was more pronounced in tumor tissue, where a compromised BBB permitted greater TmDOTP5-extravasation, compared to healthy tissue, but was evident throughout the whole brain at various depths. Integration of the separated peaks permitted non-invasive quantification of spatial heterogeneity in compartmental Na+ levels. Na+b concentrations were found to be elevated within all three tumors relative to healthy tissue, and significantly so (p This 3D 23Na-MRSI method is the first to compartmentalize Na+ levels within the tumor microenvironment with minimal invasiveness, and explore spatial heterogeneity in ∆Na+end and ∆Na+mem. It has potential to be used to monitor chemotherapeutic and anti-angiogenic drugs, and also to be used with other agents based on gadolinium (Gd3+) which are already approved for clinical use in GBM patients. TmDOTP5− administration into a rat brain bearing a U251 tumor. 1H-MRI of an axial slice displaying the anatomical tumor boundary (white outline). The 23Na-MRSI is overlaid on top of the 1H-MRI. Candidate voxels (b) inside and (c) outside the tumor are indicated (yellow boxes). Before TmDOTP5− delivery, a single 23Na peak was observed at 0 ppm, both inside and outside the tumor (black spectra). Following TmDOTP5−delivery, compartmental peak separation was achieved to varying extents throughout the brain (blue spectra). (b) Within the tumor, this separation was most pronounced due to a compromised blood-brain barrier (BBB), which permits substantial accumulation of TmDOTP5− in the extracellular space. (c) Outside of the tumor (bottom spectra), such a high degree of extravasation would not be possible, but some shifting is still observed. The TmDOTP5− distribution in the brain warrants labeling the most shifted peak as blood sodium (Na+b ), which occurred consistently around 2 ppm. The unshifted peak, which has no access to TmDOTP5−, is intracellular sodium (Na+i ). The intermediate peak, therefore, is extracellular sodium (Na+e ), which is shifted more inside the tumor than outside in healthy tissue. (d) Spatial distributions of compartmentalized 23Na signals (Na+b, Na+e, Na+i) as well as transendothelial (ΔNa+end) and transmembrane (ΔNa+mem) gradients in an RG2 tumor. The high-resolution 1H-MRI data are shown in the left four columns, whereas the lower resolution 23Na-MRSI data are shown in the next five columns on the right. The left column shows the tumor location (white outline) on the anatomical 1H-MRI (left), whereas the next two columns show the T2 maps (range shown: 0-100 ms) before and after TmDOTP5− injection, and the subsequent column depicts the ∆R2 map (i.e., difference between 1/T2 maps before and after, range shown: 0 -30 s-1), which is proportional to [TmDOTP5-] in healthy and tumor tissues. Since the integral of each 23Na peak represents the [Na+], the respective three columns show the integral maps of Na+b, Na+e, and Na+i from left to right (i.e., ∫Na+b, ∫Na+e, ∫Na+i). The last two columns on the right show ΔNa+end = ∫Na+b-∫Na+e and ΔNa+mem = ∫Na+e-∫Na+i. The ∫Na+b map reveals low values in healthy tissue compared to tumor tissue, and within the tumor boundary a high degree of heterogeneity. The ∫Na+e map reveals low values in tumor and normal tissues, but within the tumor boundary a small degree of heterogeneity is visible while ventricular voxels show very high values. The ∫Na+i map reveals low values ubiquitously except some ventricular voxels. The ΔNa+end map reveals dramatically high values within the tumor only. The ΔNa+end was driven primarily by an increase of ∫Na+b inside the tumor and which was more pronounced in superficial regions of the brain compared to deeper slices. The ΔNa+mem map shows low values in tumor tissue compared to normal tissue, although ventricular voxels show very high values.

The ΔNa+mem is driven primarily by decreased ∫Na+e and thus shows similar level of heterogeneity as the ∫Na+e map. All maps use the same color scale and are relative. Abstract Body : Introduction: Early detection of heart disease is critical in its prognosis, treatment and management. The current paradigm of anatomic imaging and measurement of circulating biomarkers detect heart disease after significant structural and functional changes have occurred. Therefore, there is a need for the use of in vivo molecular imaging techniques that can target a biomarker localized to cardiac tissue that is altered before functional derangements are evident. The hormone ghrelin and its receptor, the growth hormone secretagogue receptor (GHSR) are expressed in the myocardium, and are associated with cardioprotection through promotion of contractility and growth, and inhibition of inflammation and apoptosis. We have previously shown that the dynamics of ghrelin and GHSR are abnormally up-regulated in human heart failure and heart disease, thereby indicating that GHSR is a clinically relevant target for molecular imaging of heart disease. We therefore used a quantitative PET-MRI approach in a preclinical canine model of myocardial infarction to determine the in vivo specificity and sensitivity of a radiolabelled GHSR ligand, [18F] LCE470. Methods: Myocardial infarction (MI) was surgically induced in 11 month-old female hounds (n=4) by 2h ligation of the left anterior descending coronary artery. PET-MR imaging was conducted 3 days prior to surgery, and at multiple time points over a period of 1.5 years. All hounds were first administered 7 MBq/kg body weight of 13N-NH3 i.v. for measurement of cardiac perfusion, with a dynamic list-mode PET acquisition and simultaneous MR cine for heart function, and T1 and 3D T1w for anatomical images. 50 min later, 7 MBq/kg [18F] LCE470 was injected for determination of GHSR using the same PET-MR acquisition parameters. Compartmental modelling of both tracers was done through Carimas (turkupetcentre.fi/carimas/) on segmented polar maps delineating 3 regions of interest: infarct, left circumflex (LCX) and remote myocardium. Net uptake rate (Ki) for 13N-NH3 uptake and distribution volumes (VD) for [18F] LCE470 were calculated for each of these 3 regions for all time points. At the end of the study, animals were euthanized and endomyocardial tissue samples from each region were obtained for analysis of GHSR by confocal microscopy using a fluorescent GHSR ligand, Cy5-cyclo-ghrelin (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) . Fluorescence intensities were determined using the Percentile algorithm in Image J FIJI. Results: The VD for [18F] LCE470 decreased significantly (p Conclusion and Significance: We have characterized the specificity and sensitivity of [18F] LCE470 in detecting myocardial GHSR in a preclinical large animal model. Using this unique PET tracer, we show that regional GHSR distribution is altered after MI independently of changes in perfusion. Therefore, molecular PET imaging of myocardial GHSR may indicate subclinical changes that precede some functional changes associated with heart disease. Abstract Body : Prostate-Specific Membrane Antigen (PSMA) overexpression is a characteristic of prostate cancer and is associated with cancer progression, metastasis, and poor prognosis in patients. Interestingly, increased expression levels of PSMA are also found on the vasculature of a variety solid tumors. Using cutting-edge microfluidic organ on a chip technology and patient derived cell line, we have established an innovative approach to investigate PSMA-dependent mechanisms promoting tumor vascularization. Vascularized tumor on a microfluidic chip allows us to mimic a tumor with its microenvironment in vitro. Cells are migrating, coordinating and organizing themselves into three-dimensional structures similar to those found in a human body. We are able to design an extracellular microenvironment, monitor real time changes in vessels microarchitecture and interactions of tumor and endothelial cells. Our preliminary data confirmed that endothelial cells can be activated to sprout and further driven by secreted cytokines from tumor cells. Endothelial cells with elevated PSMA expression levels are more likely to form vessel-like structures in comparison to endothelial cells lacking PSMA. We can alter some of these interactions using angiogenic inhibitors that successfully impact endothelial cell activation. We hope to gain more understanding of this complex system that reliably bridges the gap between in vitro and in vivo models and allows precise studies of cellular and molecular mechanisms driving PSMA-dependent tumorigenesis. [1] . Nearly 91% of prostate cancers first detected in men are estimated to be clinically localized, qualifying these patients as potential candidates for radical prostatectomy [2] . Clinical radiation therapy is a noninvasive means of killing cancer cells and effectively reducing tumor burden. This method of treatment is prescribed for more than 50% of cancer patients [3] . Although radiation therapy is highly effective for the majority of cancer patients, the nonspecificity of irradiation can result in toxicity for surrounding tissues. Radiosensitizers have been developed to amplify the effects of radiation within tumor cells. Nanoparticles with highatomic-number, such as gold nanoparticles, have been proposed as next-generation radiosensitizers, due to their excellent absorptivity over soft tissue and the generation of secondary electrons, producing biological damage [4] [5] . Adjuvant radiation has been shown to mitigate disease progression, however, there is a big proportion of patients with lethal prostate cancer that do poorly even with post-operation radiotherapy, and they may receive systemic therapy, including androgen deprivation therapy (ADT) and chemotherapy [6] [7] . In randomized clinical trials, concurrent chemotherapy-radiotherapy has demonstrated improved local tumor control and overall survival [8] . A principal rationale for using concurrent chemotherapy with radiotherapy is the ability of chemotherapy drugs to radiosensitize. The objective of this study is to develop theranostic gold nanoclusters for prostate cancer targeted radiotherapy and chemotherapy to improve the therapy efficacy and survival rate. CY-PSMA-1 ligands were synthesized and conjugated with monomethyl auristatin E (MMAE), and then reacted with HAuCl4 via an one-step reaction [9] to produce PSMA-targeted AuNCs-MMAE conjugates (Fig.  1a) . PSMA-targeted AuNCs without MMAE were synthesized as a control. In vitro binding affinity, cell targeting, MMAE release, cytotoxicity and radiotherapy were tested using PC3pip cells (PSMA+) and PC3flu cells (PSMA-). In vivo tumor targeting, biodistribution and clearance of AuNCs were monitored by an IVIS Spectrum Imaging System and ICP-OES. Tumor damage was confirmed by H&E and immunochemistry staining, and radiotherapy and chemotherapy were evaluated by monitoring tumor growth over 24 days. AuNCs-MMAE conjugates with size around 2.5 nm showed strong fluorescence with ex/em at 490nm/670nm, and in vitro cellular uptake demonstrated significantly higher AuNPs uptake in PC3pip cells than in PC3flu cells. MMAE was cleaved from AuNCs by cathepsin and released intracellularly, leading to cell apoptosis, but could be inhibited by the addition of cathepsin inhibitor. Synergetic radiotherapy by MMAE and AuNCs was confirmed with a colony formation assay, showing the best therapy outcome for the AuNCs-MMAE conjugates. In vivo fluorescence results demonstrated higher AuNCs accumulation in PC3pip tumors than in PC3flu tumors, and fast urinal clearance of AuNCs. In vivo therapeutic studies showed that a single injection of AuNCs-MMAE conjugates significantly inhibited PC3pip tumor growth compared to small molecular PSMA-MMAE, and a exposure levels. 18F-pretomanid PET/CT shows high penetration is achieved in infected lungs, with decreased penetration into lung lesions and cavitary walls being attributed to reduced vascular supply. We also observed high penetration of 18F-pretomanid into infected brains. Since the penetration of 76Br-bedaquiline and 18F-linezolid into the brain parenchyma is low, we propose new pretomanid-containing regimens should be tested for the treatment of TBM. The heterogeneous profile of 18F-pretomanid tissue penetration highlights the value added by radiolabeled antimicrobials in the optimization of drug regimens for TB. ROIs of unaffected lung). D. AUC (0-60 min) tissue/plasma rations of 18F-pretomanid distribution in brain lesions and areas of unaffected brain in a mouse model of tuberculous meningitis (n=4 mice, n=7 ROIs of brain lesions, n=7 ROIs of unaffected brain). Representative MIP and brain axial slide of 18F-pretomanid in a mouse with tuberculous meningitis. E. High resolution autoradiography of 18F-pretomanid in Mycobacterium tuberculosis-infected rabbit brain shows decreased penetration in brain lesions, identifiable by hematoxylin & eosin staining. Abstract Body : Macrocyclic complexes of Fe(III) typically have greater kinetic inertness towards dissociation than linear chelates, making them a good choice for designing new contrast agents. Phosphonate pendants offer a way to introduce negatively charged groups to the coordination sphere of MRI contrast agents and, when used in Gd(III) and Mn(II) contrast agents, have been shown to increase proton relaxivity of the agents.1, 2 Using multiple phosphonate pendants on a single macrocyclic-based ligand allows for additional negative charges in the coordination sphere, to produce an anionic complex. High spin iron(III) has a moderately long electronic relaxation time and can be used as an alternative to Gd(III) for the development of MRI contrast agents.3, 4 Using these principles, a series of Fe(III) complexes have been synthesized using 1, 4, 7-triazacyclononane (TACN), containing one, two, or three phosphonate pendants resulting in closed coordination complexes with a varying degree of anionic charge. For the non-phosphonates in this series, hydroxypropyl donor groups have been used. The three phosphonate pendant TACN complex (Fe-NOTP) shows similar r1 relaxivity values (0.98 mM-1 s-1, 1.4 T) compared to the hydroxypropyl derivative (1.0 mM-1 s-1, 1.4 T) and both show slightly higher r1 relaxivity values compared to the tris-carboxylate TACN complex (Fe-NOTA, 0.63 mM-1 s-1, 1.4 T). However, Fe-NOTP showed a 58% increase in relaxivity in the presence of HSA, whereas the hydroxypropyl derivative5 and Fe-NOTA show a 23% and Fe-NOTP showed higher kinetic inertness in the presence of acid (100 mM HCl) than the hydroxypropyl derivative, with no dissociation of Fe-NOTP observed after three days at 37 °C. The phosphonate and hydroxypropyl derivatives both show high kinetic inertness towards dissociation in the presence of HEPES buffer and biologically relevant anions (carbonate and phosphate). The two also showed similar reduction potentials (near -330 mV vs. NHE), suggesting both phosphonates and hydroxypropyl donor groups stabilize iron (III) well. Mouse MR imaging studies with Fe-NOTP show kidney and vena cava T1 contrast enhancement similar to DOTAREM when twice as much Fe-NOTP is injected (100 μmol/kg vs. 200 μmol/kg). However, Fe-NOTP shows slightly prolonged kidney elimination compared to DOTAREM. The hydroxypropyl derivative has been reported to show kidney and vena cava enhancement similar to that of DOTAREM at equal injection concentrations (50 μmol/kg), however, higher doses of the hydroxypropyl derivative were not reported due to lower solubility.5 This data suggested that phosphonate pendants can increase kinetic inertness of the iron complexes in acid, HSA binding, and aqueous solubility. However, the hydroxypropyl donor groups show faster clearance in vivo and slightly higher relaxivity values. Future studies with the mixed phosphonate-hydroxypropyl donor pendants could produce a complex that has increased solubility and inertness as well as in vivo enhancement comparable or better than DOTAREM at equal injection concentrations. in vehicle treated mice, demonstrating the potential for imaging liver fibrogenesis. Conclusion MnL1 is a novel macrocyclic Mn(II) MR probe with a pendant hydrazine carboxylate moiety that enables rapid reaction with aldehydes, a 270% turn on in relaxivity at 1.4T, low nonspecific liver enhancement in healthy mice, but markedly higher liver signal enhancement in mice with ongoing liver fibrosis, and shows potential for noninvasive imaging of liver fibrosis. Abstract Body : Despite recent successes in immune checkpoint blockade therapy of several types of human cancers, pancreatic cancer has a poor response to immunotherapy1-3. Increasing evidence shows that a dense stromal barrier in pancreatic cancer blocks drug delivery and intratumoral distribution4,5. Stromal physical and immunosuppressive biological barriers further limit the number and function of infiltrating effector T cells3,6. It is clear that the presence of a pro-immune tumor microenvironment is required for a good therapeutic response to immune checkpoint therapy6-9. Pancreatic cancer is an immunologically cold tumor that has a low level of neoantigens, lacks infiltration of immune effector cells, and contains massive immunosuppressive cells1, 3,5. To improve therapeutic responses to immune checkpoint therapy in pancreatic cancer, an MRI capable and ultra-small magnetic iron oxide nanoparticle(IONP)based immune checkpoint PD-L1 inhibitor (Nano-iPD-L1) has been developed by conjugating high-affinity PD-L1 binding and blocking peptides (PD1Y) to the polymer coating of 5 nm core size IONPs. Systemic delivery of near-infrared dye-labeled Nano-iPD-L1 led to the selective accumulation of the IONPs in pancreatic tumors of the Kras-driven transgenic mice for optical and MR imaging of intratumoral delivery of the nanoparticle PD-L1 inhibitor. Importantly, IONP mediated delivery of PD-L1 blocking peptides increased intratumoral infiltration of antigen-presenting dendritic cells and CD8+ cytotoxic T cells. To further improve the delivery of Nano-iPD-L1 and activation of tumor-specific immune responses, a receptor-targeted and stroma breaking theranostic IONP has been developed by conjugating a recombinant ligand containing the amino-terminal fragment (ATF) of uPA and the catalytic domain of matrix metalloproteinase-14 (ATFmmp14-IONP), which confers the ability to target uPAR expressing tumor and stromal cells and promotes nanoparticle/drugs to migrate through the stromal cell and extracellular matrix barriers to reach tumor cells. Results showed that co-delivery of Nano-iPD-L1 with ATFmmp14-IONP could overcome stromal barriers and enhance intratumoral delivery and penetration of both IONPs in mouse pancreatic tumors. Furthermore, the combination therapy using ATFmmp14-IONP encapsulated with a chemotherapy drug, Doxorubicin (DOX), and Nano-iPD-L1 significantly inhibited tumor growth in the KPC mouse pancreatic cancer model. However, inhibition of PD-L1 alone using Nano-iPD-L1did not show an anti-tumor effect in this tumor model. Histological analysis showed that targeted co-delivery of these theranostic IONPs into tumors further increased infiltration of dendritic cells and effector T cells and decreased immunosuppressive cells, converting an immune "cold" pancreatic tumor into a "hot" tumor. Flow cytometry analysis revealed a higher level of activated CD8+ cytotoxic T cells in the tumor-infiltrating lymphocytes obtained from the tumors of the mice that received Nano-iPD-L1, ATFmmp14-IONP, or the combination treatment than those in tumors from the no treatment control. Immunofluorescence analysis of tumor tissues revealed that conventional DOX treatment markedly increased the levels of expression of immune checkpoint proteins, PD-L1 and Siglec 1510, which are strong inhibitors of cytotoxic CD8+ T cells. However, treatment with ATFmmp14-IONP/DOX or the combination of ATFmmp14-IONP/DOX and Nano-iPD-L1 inhibited DOX-induced upregulation of PD-L1 and Siglec 15 in tumors. Results of this study suggest that improved drug delivery in tumors by co-administrations of a stroma penetrating ATFmmp14-IONP/DOX and a Nano-iPD-L1 leads to a strong therapeutic response through improved intratumoral delivery and stromal penetration, enhanced direct tumor cell killing and tumor antigen presentation, activating tumor immune responses, and blocking PD-L1 function to enhance the cytotoxicity of CD8+ T cells. Therefore, these targeted theranostic and immunotherapy IONPs and the combination therapy have the potential for further translational development of a new image-guided and targeted cancer immunotherapy for advanced pancreatic cancer.

Claire Rabut, California Institut of Technology, crabut@caltech.edu

Abstract Body : Motivation Ultrasound is playing an emerging role in molecular imaging thanks to new nanoscale contrast agents and reporter genes. Acoustic methods for the selective detection of these agents are needed to maximize their impact in biology. Existing ultrasound pulse sequences use the nonlinearity in contrast agents' response to distinguish them from linear tissue scattering [1, 2, 3] . However, such pulse sequences typically scan the sample using focused transmissions, resulting in a limited frame rate and restricted field of view. Meanwhile, existing wide-field scanning techniques based on plane wave transmissions suffer from limited sensitivity or nonlinear artifacts [4, 5, 6] . To overcome these limitations, we introduce ultrafast amplitudemodulated imaging (uAM): a nonlinear modality combining amplitude-modulated pulses, multiplane wave transmissions and selective coherent compounding.

Methods Bursts of N successive tilted plane waves are repeated three times with modulated amplitudes: +1/2, +1/2 and +1 (Fig.A) . After reception, the subtraction of the two half-amplitude bursts from the full amplitude burst allows the elimination of linear signal and the capture of specifically nonlinear responses. Modulated multiplane waves (MPW) bursts are then repeated N times, and for each repetition the polarities of the successive plane waves are given by the column of the Hadamard matrix of order N (Fig.B) . In conventional MPW imaging, the contribution of all the Hadamard RFi data is summed to produce a coherent recombination of compounded plane wave signals [7] . We however found that the first Hadamard summation systematically leads to a pulse-inversion like summation of polarized pulses and therefore carries inappropriately time-delayed echoes in the recombined signals (Fig.C) . We therefore introduced selective coherent compounding of all Hadamard RF data except the first one H1. The resulting image, for which harmonic residues at inappropriate time have been eliminated, is expected to contain (N-1) times the signal obtained with a single plane wave.

We demonstrate the performance of uAM by imaging gas vesicles (GVs), an emerging class of genetically encodable biomolecular contrast agents.

In vitro results We first evaluated in vitro the performance of uAM in comparison to slower goldstandard amplitude modulation sequence parabolic-AM (pAM) [2] and xAM [3] (Fig.D.E) in a tissue-mimicking phantom containing two rows of 2mm diameter wells filled with GVs. uAM achieves contrast imaging sensitivity (Fig.E) comparable to pAM or xAM (250kHz imaging framerate) but allows the recording of a significantly larger and deeper field of view (Fig.D) and at a much faster imaging framerate of 3.2 kHz. In vivo results These exceptional spatio-temporal performances enable simultaneous nonlinear and linear image formation and allow concurrent monitoring of ultrafast phenomena, such as acoustic reporter gene dynamic and blood volume variations. We therefore demonstrated in vivo the performance of uAM technique by imaging the circulation and phagolysosomal absorption of GVs. We monitored in real-time both the blood flow dynamic and the phagolysosomal function of a mouse liver after intravenous injection of GVs (Fig.F) . Leveraging the ability of uAM processing to extract simultaneously both linear and nonlinear information from the same pulse sequence, we processed two complementary sets of data ( Fig.G) : by applying a clutter filter to the linearly-processed uAM images, we obtained Power Doppler (PD) images of the liver, tracking the vascular enhancement due to circulating GVs (Fig.H) ; while the nonlinear-processed images from the same plane tracked the contrast of GVs only. AM images therefore express the population of circulating GVs, but also GVs taken up by liver tissue (Fig.I) . This in vivo application highlights the capacity of uAM to provide simultaneous access to nonlinear contrast and ultrafast phenomena such as blood flow.

Conclusion uAM's exceptional combination of speed, sensitivity and spatial coverage will give this pulse sequence a bright future in contrast ultrasound. (4 angles) transmission in an agarose phantom with a point source inclusion of GVs. C. After Hadamard summing, individual plane waves are retrieved with amplitude N, except for H1, which contains additional nonlinear echoes. Selective compounding omitting H1 enables the reconstruction of a nonlinear image sans artifact. D. In vitro evaluation of uAM-8angles performance in comparison with parabolic-AM and xAM [2, 3] (2mm diameter GVs-inclusions are embedded in a tissuemimicking phantom): we immediately observe that uAM offers the widest and deepest field of view compared to pAM and xAM. E. Contrast-to-Tissue Ratio (CTR) and Contrast-to-Artifact Ratio (CAR) for each pulse sequence. CTR=upper inclusions/upper background. CAR=upper inclusions/artifact region underneath. Err.b.± SEM (N=3) F. Schematic of the in-vivo experiment: purified GVs were intravenously injected in an anesthetized mouse, and the liver was continuously monitored with uAM. G. uAM imaging of simultaneous nonlinear and blood flow imaging to visualize mouse liver function. H. After injection, the vascular signal quickly reaches a maximum within 60 s, then decreases back to a baseline. I. Meanwhile, the AM signal progressively increases to a maximum around 600 s after the injection, corresponding to the largest concentration of intact GVs in the liver. Macrophages then degrade the GVs, resulting in a gradual decrease towards baseline. Scale bars: 1 mm. Abstract Body : The opportunity to quantitatively study specific cellular phenomenon of a complete human pancreas with preserved spatial 3D context would have extensive significance for pre-clinical and clinical research. For studies of human pathophysiology and underlying anatomy, a method that could selectively label full human organs would be highly desirable. A variety of mesoscopic imaging approaches are currently available for deep tissue imaging of biological samples, including optical projection tomography (OPT)1 and light sheet fluorescence microscopy (LSFM)2-4. Whereas optical 3D imaging techniques have powerfully progressed, they have remained limited due to insufficient antibody labelling within large tissue volumes. In research fields incorporating cell-loss or cell exhaustion studies, like the fate of beta cells in the pancreas in settings of diabetes, a better understanding of disease pathogenesis for the recovery of beta-cell mass is lacking. In this study, we describe a ex vivo concept that enables 3D imaging of the full human pancreas with micrometre resolution. In analogy to a 3D jig saw puzzle, entire human pancreata were sliced in a 3D printed matrix for high-throughput tissue-preparations of cm3-sized tissue cuboids, immunolabelled (against insulin), optically cleared (in BABB), 3D imaged and the resultant image data sets were computationally stitched back in 3D space. The method enabled quantification of antibody labelled cells with preserved organ context. Analysed pancreatic tissues from healthy and type 2 diabetic (T2D) donors encompassing over 200.000 islets of Langerhans provided information about normal and abnormal islet compositions, including the localization of regions with densely packed islets and, in T2D, islets with internal haemorrhaging and loss of sphericity of larger islets. The study provides 3D evidence that human islet mass may exhibit significant heterogeneities in its distribution, and it confirms stereological5 indications of the appearance of islet clusters in the periphery of the organ, be it from non-diabetic or T2D donors. Taking advantage of the autofluorescent properties of tissues as a source of contrast to outline cellular compositions or processes6,7, more specific of red blood cells, thousands of hyperaemic islets within or in colocalization with the antibody labelled regions were identified. However, the impact of these lesions on human islet physiology and whether they are specifically formed in connection with T2D remains to be elucidated. In summary, the presented method provides a mean to characterize previously unrecognized features in larger tissue volumes of the human pancreas with micrometre resolution and maintained spatial context. Further, it may be directly translated to assessments of other markers and organ systems.

immunohistochemistry of inflamed ears and joints confirmed our in vivo PET/MRI imaging data. Conclusion: Simultaneous tracking of the temporal dynamics of the accumulation and homing sites of CD4+ T cell and phagocytes by non-invasive in vivo PET/MRI might help to improve the evaluation of novel immunotherapeutic compounds as well as the monitoring of IMIDs. Thus, this dual 89Zr-CD4-DFO-minibody and 19F-PFC PET/MRI-based imaging approach might help to elucidate the pathogenesis experimental autoimmune diseases such rheumatoid arthritis or psoriasis. Acknowledgment: This work was supported by the Horizon2020 Programme under grant agreement n° 675417 (PET3D) and the EU/EFPIA/Innovative Medicines Initiative 2 Joint Undertaking under grant agreement n° 831514 (Immune-Image). Abstract Body : INTRODUCTION The development of further techniques in synthetic biology is dependent on reporter genes to generate a quantifiable response on whether the base circuit is set up correctly. Reporter genes for Magnetic Resonance Imaging (MRI) would allow genes to be measured in a non-invasive and safe manner, paving the way for translational use in human subjects. One of the best mechanisms for generating such a reporter gene is Chemical Exchange Saturation Transfer (CEST), which generates contrast without addition of outside metals and allows for the contrast mechanism to be controlled. Previous CEST reporter genes [1, 2] can be further optimized via protein evolution but have proven difficult to improve using typical protein optimization engineering methods. We hypothesized that these limitations could be overcome via application of machine learning. INNOVATION We developed the Protein Optimization Evolving Tool (POET). POET uses Genetic Programing, a method of machine learning, to identify motifs within peptides that allows them to produce MRI contrast. From these motifs POET then produces a list of peptides that are predicted to have high contrast. We then ordered the predicted peptides and evaluated them via MRI to determine how much contrast they generate. This in turn becomes new training data for POET to use to predict a new generation of peptides. This method is not only a new way of designing peptides, but due to the nature of the program, a far wider search space of peptides is analyzed, leading to the exploration of novel mechanisms to optimize the production of CEST contrast. METHODS Twelve amino acid long peptides were synthesized by Genscript (NJ) and were tested on a preclinical 7T Bruker scanner in a phantom designed specifically for this purpose. The scans used a modified RARE sequence with a TR of 10000 ms and a TE of 4.74 ms. Saturation was performed with 4.7 µT pulse for 4000 ms. In these experiments POET was used to optimize for contrast at 3.6 ppm, but could be used to optimize for any given frequency. RESULTS & DISCUSSION This process has been used over the course of eight generations and has produced peptides with MRI contrast up to four times higher than previously Poly-L-Lysine (PLL; Figure, top panel, Red line) which is the gold standard for peptide based contrast at 3.6ppm [3] . Further, this mechanism has shown the ability to discover new peptides that have challenged the current understanding about what chemical properties allow MRI contrast to be generated. Although the training data, and prior research was focused on peptides that were both positively charged and basic, POET had led to the discovery of CEST agents that are neutral in charge and pI, but produce more contrast than PLL (Figure, bottom panel) . CONCLUSION This process allows for a wider search space of peptides to be analyzed than via directed evolution and for developing proteins that challenge current theories about how CEST contrast works. This will aid future research into peptides used to generate CEST contrast as well as developing short peptides to fill a variety of other functions. This study demonstrates how protein engineering using machine learning can be leveraged to develop new molecular imaging probes ID: GA296 DeepNI: A cloud computing system for deep learning-based neural image analysis Zi-Min Weng, University of Alabama at Birmingham, zimin99@uab.edu

Abstract Body : Overview: There is a growing number of deep learning-based methods for the processing, classification, identification, and quantification of neural images. As more and more open-source deep learning models are being shared in the open domain, making such models easily accessible and useable for researchers of all levels of computer skills has been an important task as we aim to promote the usage of such advanced computational methods. We present a software platform, DeepNI, that facilitates an easy and fast deployment of deep learning-based neural image processing and analysis tools. Objective: DeepNI, where NI stands for neuroimaging, is an open-source software based on cloud-computing developed with MATLAB with the following objectives: 1) To establish a cloud computing system equipped with graphical processing units (GPUs) to implement deep learning computation of classification, segmentation, and detection tasks on neural imaging. 2) To build up a MATLABbased user-friendly graphic user interface (GUI) for easy and intuitive operations for image processing. 3) To deploy a reliable data transmission and communication between cloud computing server and multiple clients. Method: DeepNI [1] is implemented with MATLAB as the client-end and server-end programs. For the client-end programs, a GUI is developed for data import, job submission, job status checking and data retrieval upon completion of the deep learning-based image processing tasks. A view is also included for results verification and evaluation. For the server-end programs, we have implemented the job handling through the MATLAB Job Manager toolbox [2] as the communication component in DeepNI to transmit jobs and results between the server and clients through the Transmission Control Protocol (TCP) and to manage multiple jobs from clients for the computing processes. The server is established on the Google cloud platform with virtual instances that are equipped with GPU resources. Result: DeepNI has been implemented and currently shared on GitHub for free academic use. Multiple deep learning models have been integrated into DeepNI as more new models being added into our system. More advanced functions, such as a web-based user interface, will continue to be added into DeepNI in order to further stimulate the testing and evaluation of deep learning models for neural image processing and analysis.

Hyperpolarized MR imaging of ascorbate-mediated oxidative stress in cancer Nathaniel Kim, Memorial Sloan Kettering Cancer Center, kimn1@mskcc.org

Abstract Body : Introduction Tumor progression and survival is driven by genetic mutations and environmental conditions. Recent work has revealed that redox metabolism is deregulated in pancreatic cancer and may provide new opportunities for diagnosis and targeted treatment. Specifically, a therapeutic strategy for targeting KRAS mutant cancers using high dose ascorbate was introduced to disrupt redox homeostasis.1 While targeted therapeutic strategies have been introduced, methods to monitor the metabolic reprogramming of redox homeostasis are critically needed. We sought to identify hyperpolarized DHA (HP DHA)2 as an imaging agent for monitoring oxidative stress in patient derived xenograft models (PDXs) of cancer. Methods Ascorbate therapy of pancreatic cancer PDXs -KRAS-driven (PC106) and BRCA-driven (HyMad1) PDX models, created from patient samples and verified using targeted IMPACT sequencing,3,4 were subcutaneously xenografted in the right flank of NSG mice. 2 weeks after implantation, mice were randomly divided into two groups. One group was treated with freshly prepared vitamin C in 250 μL of PBS (4 g/kg) twice a day via IP injection (PC106, n = 7; HyMad1, n = 7). Control group mice were treated with PBS using the same twice a day dose (PC106, n = 7; HyMad1 n = 7). T2 images of these mice were acquired weekly to quantify tumor size and evaluate changes in tumor growth over time. HP DHA imaging of pancreatic cancer PDXs -Ascorbate and vehicle-treated mice of both PDX models were injected with 250 μL of 30 mM HP DHA in D2O to investigate the redox status of the implanted tumor over time. 13C HP MRS was performed using an axial 2D CSI acquisition on a 5T small animal MRI spectrometer to measure DHA/ascorbate metabolism. DHA/ascorbate ratios were quantified by taking the area under the curve and compared between ascorbate-treated and vehicle mice to determine changes in tumor redox upon ascorbate therapy. Metabolomics of solid tumors -Tumor samples of ascorbate and vehicle-treated mice of both PDX models were extracted after 1 week and 5 weeks treatment. Tumor samples were analyzed by untargeted metabolomics to quantify differences in key redox and metabolic pathways between treatment and vehicle groups for both PDX models. Results Ascorbate treatment was shown to slow tumor growth in the PC106 model after 3 weeks (36.5% ± 9.3%, p=.0035) while the effect on tumor growth was significant in the HyMad1 PDX model after 7 days (74.0% ± 12.3%, p=.0001, Figure 1A) . Interestingly, the observed HP DHA/ascorbate ratio in vivo for PC106 PDXs after 1 week of ascorbate treatment did not change (8.5% ± 21.7%, p=.7105), mirroring the lack of significant response observed in tumor growth after 1 week ( Figure 1B) . The observed HP DHA/ascorbate ratio in vivo in HyMad1 PDXs after 1 week of ascorbate treatment was found to decrease by 28.0% ± 12.3% (p=.0593), reflecting the increase in oxidative stress upon ascorbate treatment ( Figure 1C ). Initial metabolomic results for the KRAS-driven PDX model show significant changes in the redox machinery of the tumor after 1 week of ascorbate treatment. NADH levels in tumors treated with ascorbate increased by an average 90% ± 29.8% (p=.0240) while NADP levels increased by an average 100% ± 39.4% (p=.0353) versus tumors treated with vehicle.

Caitlin Tressler, The Johns Hopkins University School of Medicine, ctressl3@jhmi.edu Category: Oncology Abstract Body : Introduction: Gemcitabine is a first line chemotherapy agent used to treat pancreatic cancer and metastatic disease [1] . While gemcitabine metabolism has been well studied by LC-MS and pharmacokinetics, little has been done to determine the spatial distribution of drug delivery and metabolism [1] [2] [3] . This is particularly important as pancreatic tumors are prone to high rates of reoccurrence, which may result from non-uniform distribution and response to treatment across the tumor. Furthermore, the metabolic response to gemcitabine treatment in tumor tissue is unknown. Methods: Athymic nude mice were inoculated in their flanks with Panc1 tumors which were grown to 500 cm3. Prior to euthanasia, animals were injected with 200 mg/kg gemcitabine via tail vein at 0.5, 1.0, and 2.0 hours. Animals were euthanized with CO2, and tumors, pancreases, livers, and kidneys were collected for analysis. Tissue mimetic model was prepared from homogenized untreated mouse liver. All tissues were cryo-sectioned at 10-micron thickness onto pre-chilled indium tin oxide slides. MALDI imaging was performed on a Bruker RapifleX MALDI TOF/TOF instrument. Tissues were imaged at 50micron pixel size in reflectron positive mode with 200 laser shots per pixel. Mass spectral range was m/z 150 to 2,000. Data was normalized to total ion count and analyzed in FlexImaging and SCiLS Lab. Results: Gemcitabine was detected in tumors and quantified using the tissue mimetic model. Concentrations of gemcitabine peaked at 0.5 hours post-injection at around 230 µg per mg tissue in the tumor tissue. In addition, gemcitabine triphosphate was detected in treated tissues. Tumor tissue images were imported into SCiLS lab for unsupervised analysis of the lipid region (m/z 400 to 1, 200) . Segmentation analysis revealed significant differences between the 1 and 2 hour-time points. These differences in segmentation were also observed in the pancreas but not in liver and kidney. Discussion: Our data demonstrate that gemcitabine was metabolized to the active drug metabolite gemcitabine triphosphate in tumor and pancreas and peaked at 1 hour of post-injection. MALDI images of [Gem+H+] at m/z 264 revealed that gemcitabine and gemcitabine triphosphate were not evenly distributed throughout the tumor. Profound overall metabolite and lipid profile changes were observed in tumor and pancreas as shown by segmentation analysis, likely resulting from a metabolic response to gemcitabine treatment. Conclusions: Gemcitabine is delivered and metabolized unevenly across Panc1 flank tumor models. Furthermore, gemcitabine has a profound effect on metabolites and lipids within tumor tissue. We are currently completing measurements of additional biological repeats, histopathology stains (H&E, trichrome) and identifying metabolite and lipid markers of response to gemcitabine treatment. Abstract Body : Non-specific uptake and retention of molecular targeted agents and heterogeneous tissue optical properties diminish the ability to differentiate between tumor and normal tissues using molecular targeted fluorescent agents. Paired-agent imaging (PAI) can increase the diagnostic ability to detect tumor tissue by mitigating these non-specific effects and providing true molecular contrast by co-administration of an untargeted control imaging agent with a targeted agent. This study evaluates the suitability of available clinically translatable untargeted agents for the translation of PAI in fluorescence guided surgery using an affibodybased targeted imaging agent (ABY-029). Three untargeted agents that fluoresce near 700 nm and exhibit good clinical safety profiles (methylene blue, IRDye 700DX and IRDye 680LT) were tested in combination with the clinically tested IRDye 800CW labeled anti-epidermal growth factor receptor (EGFR) Affibody molecule, ABY-029 (eIND 122681). Properties of the untargeted agent important for human use and integrity of PAI were tested: 1) plasma protein binding; 2) fluorescence signal linearity in in vitro whole blood dilution; 3) in vivo pharmacokinetic matching to targeted agent in negative control tissue; and 4) in vivo diagnostic accuracy of PAI vs single agent imaging (SAI) of ABY-029 alone in orthotopic oral head and neck squamous cell carcinomas. Results: Overall, IRDye 680LT outperformed IRDye 700DX and methylene blue with the highest signal linearity (R2 = 0.9998 ± 0.0002, 0.9995 ± 0.0004, 0.91 ± 0.02, respectively), the highest fluorescence yield in whole blood at 1 μM (104.42 ± 0.05, 103.68 ± 0.09, 101.9 ± 0.2, respectively), and the most closely matched ABY-029 pharmacokinetics in EGFR-negative tissues (Binding Potential error percentage = 0.31% ± 0.37%, 10.25% ± 1.30%, and 8.10% ± 5.37%, respectively). The diagnostic ability of PAI with ABY-029 and IRDye 680LT outperformed conventional SAI with an area-under-the-receiveroperating-characteristic curve (AUC) value of 0.964 vs. 0.854, and 0.978 vs. 0.925 in the Odyssey scanning system and Pearl wide field imaging system, respectively. Conclusion: PAI is a highly promising methodology for increasing detection of tumors in fluorescence guided surgery.

Although not yet clinically approved, IRDye 680LT demonstrates promise as an untargeted agent when paired with ABY-029. The clinical translation of PAI to maximize tumor excision, while minimizing normal tissue removal could improve both patient survival and life quality.

First Name: Cheng Image/Figure Caption: At the top, structure of 111In-DTPA-P60. The radiolabeled peptide was administered to healthy mice subcutaneously (blue), intravenously (red) and intraperitoneally (green) . At the bottom, representative SPECT/CT images at 0 h, showing activity located at the injection site (IS), and its retaintion over time (graph in dashed box), bladder (Bl) and in the intraperitoneal space (IP). 

Abstract Body : The arbitrary waveform relaxometer (AWR) is an indispensable platform for developing magnetic nanoparticle tracers for magnetic particle imaging (MPI) applications. By characterizing the point spread function (PSF) of a specific magnetic nanoparticle, the AWR can evaluate the performance (e.g. SNR and spatial resolution) of the magnetic nanoparticle for MPI [1] . The design of wideband (DC-400 kHz) and arbitrary waveforms enables rapid optimization of pulse sequence design. Currently, the AWR serves as a complementary tool for MPI as well as a sensor for many applications [2] [3] [4] . One of the biggest challenges in arbitrary waveform excitation is direct feedthrough interference, which is usually six orders of magnitude larger than the signal from magnetic nanoparticles. In MPI and conventional magnetic particle spectrometers (MPS), a high-pass filter and tuned filter are applied to filter out the single tone interference, which cannot be done for arbitrary waveform excitation. The AWR uses a very fine shimming system, which employs a cantilever and duplicated gradiometer receive coil. This mechanism currently achieves over six orders of magnitude of direct feedthrough attenuation [1] . A challenge in mechanical coil shimming is that it is highly sensitive to spatial variations, which often occur as the shimming process is done before each scan to ensure the highest possible feedthrough rejection. This abstract will show results for a design of a solenoidal gradiometric coil that is less sensitive to spatial variations and still provides adequate direct feedthrough rejection. The main approaches used are that varying the winding pitch at the ends of the transmit (TX) solenoid coil reduces the spatial variation of the magnetic field produced when driven by a current; and varying the winding pitch at the ends of the receive (RX) solenoid coil reduces the spatial variation of the total flux through the loops of the coil. Preliminary results show a factor of 2 reduction in spatial variation with just one varied pitch extra turn on the ends of the TX and RX coils. Abstract Body : Introduction: Current assessment of neuropathic pain relies on self-reported pain and the physiological/behavioral response to evoking stimuli, which do not provide information of changes at the cellular or molecular level.1 PET imaging is a noninvasive molecular imaging technique that can quantify biochemical processes underlying pain mechanisms. Increased expression of the α2δ-1 subunit of voltage-dependent calcium channels has been observed in the spinal cord and nerves in many different animal models of neuropathic pain2. Additionally, gabapentinoids, which bind to the α2δ-1 protein, reverse the increase in α2δ-1 expression and reduce the neuropathic pain phenotype in animal models.2 Therefore, the expression level of α2δ-1 appears to be a robust biomarker of neuropathic pain. Gabapentin, a highly selective α2δ-1 ligand with nanomolar affinity, is a first line treatment for neuropathic pain3. Based on this knowledge, we hypothesized that radiolabeled gabapentin could serve as a new PET tracer and offer objective measures of neuropathic pain. Two novel fluorinated derivatives of gabapentin (cis-and trans-4-fluorogabapentin) and their 18F-labeled versions have been developed recently.4 Here, we describe the evaluation of these novel radiotracers in healthy non-human primates, healthy rats, and rats post spinal nerve ligation (SNL), a model of neuropathic pain. Methods: Radiotracers trans- [18F] 4-fluorogabapentin ( [18F] tGBP4F) and cis- [18F] 4-fluorogabapentin ([18F]cGBP4F) were synthesized according to recently developed method.4 SNL rats were acquired from Charles River ~7 days post-surgery. The specific binding of the tracer was confirmed using in vitro autoradiography on rat spinal cord sections. The binding affinity (IC50) was measured by competitive radioligand binding to spinal cord sections using quantitative autoradiography. Rats were euthanized at 75 min after tracer injection and the critical tissues were dissected and counted using a gamma-counter. Following gamma-counting, the spinal nerves were imaged using ex vivo autoradiography. Dynamic PET imaging and arterial blood sampling were performed on two rhesus monkeys at baseline and after preinjection of gabapentin (5 mg/kg). Blood samples were analyzed by gamma counting and HPLC to determine the radioactivity concentrations in whole blood and plasma as well as radiometabolites. Time activity curves were extracted from brain regions of interest and kinetic analysis was applied to estimate volume of distribution and receptor occupancy. Results: Autoradiography on rat spinal cord sections showed binding of both radiotracers to the dorsal horn region consistent with the reported expression of α2δ-1 protein by immunohistochemistry5. [18F] tGBP4F showed slightly higher binding affinity than gabapentin whereas [18F] cGBP4F showed lower binding affinity (IC50: tGBP4F = 23 ± 11 nM, n = 4; GBP = 35 ± 12 nM, n = 4; cGBP4F > 500 nM, n = 2). Analysis of monkey blood showed that both radiotracers undergo minimal metabolism in vivo (>95% parent fraction 90 min post injection) and that the tracer has low plasma protein binding (free fraction >90%). Monkey dynamic PET imaging showed modest brain penetration and differences between the tracers (brain SUV120-180min: [18F] Abstract Body : Nerve damage plagues surgical outcomes, significantly affecting post-surgical quality of life. Nerve damage occurs in ~8% of all surgeries and intraoperative nerve injuries affect 25 million patients annually worldwide, incurring undue pain, loss of function, and high costs to the healthcare system. Intraoperative nerve detection is completed using anatomical knowledge and conventional white light visualization when possible. However, nerves can be difficult or impossible to identify by white light visualization and neuroanatomy is often varied between patients. Fluorescence-guided surgery (FGS) offers a potential means for enhanced intraoperative nerve identification and preservation. To date, a variety of nerve specific fluorophores have been tested in preclinical models, however a clinically approved nervespecific contrast agent does not yet exist. Synthesis and functional validation of a nerve-specific library has resulted in development of the first-in-class near infrared (NIR) nerve-specific small molecule fluorophores for translation to first-in-human clinical studies using existing clinical infrastructure of FGS systems. Further synthetic tuning of the most promising fluorophores has generated improved brightness, solubility, and safety. Increased brightness levels have resulted in 2-3x increase in nerve signal-to-background ratio (SBR). Improved solubility has enable water-soluble nerve-specific fluorophores with improved toxicity. Single-dose toxicity testing including blood chemistry and 14-day monitoring has facilitated definition of the no observable adverse effect level (NOAEL), informing plans for upcoming investigational new drug (IND) enabling studies. Pharmacodynamics studies have been completed to identify the optimal imaging time point and dose, where nerve contrast is generated within 15 minutes and remains present for up to four hours following systemic administration. Clinically relevant formulation strategies have been explored to enable solubilization of these promising fluorophores for systemic administration when necessary. Micelle based formulations have been identified that provide equivalent performance to the previously utilized co-solvent formulation while possessing clinically relevant safety and stability profiles. Additional work is underway to identify the biomolecular target of these structure-inherent targeting small molecule fluorophores to enable a mechanistic understanding for nerve-specificity. Clinically relevant, minimally invasive surgical procedures guided by NIR nerve-specific fluorescence have been completed in large animal models using the da Vinci surgical robot, demonstrating efficacy in surgically relevant models using existing clinical FGS system intrastructure. Given the interplay between recent synthetic tuning efforts and safety and pharmacology testing, we anticipate translation to the clinic within the next two years. We believe that the continued development and use of these fluorophores in humans will improve nerve identification during surgery and reduce nerve injury, improving outcomes across all surgical specialties. 

Myelin content is decreased in the brain of Down syndrome animal model compared to its littermate wild type during aging process: an in vivo PET 11C-PIB imaging study Larissa Estessi de Souza, University of Sao Paulo Medical School, larissa.estessi@gmail.com

Abstract Body : Introduction: There is evidence that with normal aging, defects in myelin occur, which are an early pathological mark that occurs in oligodendrocytes before other brain cells show signs of degeneration (1) . Molecular neuroimaging allows in vivo and noninvasive studies of cellular and molecular processes, and therefore, is an important tool to understand the cells changing over time. Positron Emission Tomography (PET) imaging using Pittsburgh Compound B labelled with carbon-11 (11C-PIB) is traditionally used for β-amyloid plaque quantification, however, in the last decade has also been used for myelin quantification (2) (3) (4) . Objective: The objective of this study was to evaluate the myelin content changes over the life of an animal model of Down syndrome (Ts65Dn) and its wild type, by 11C-PIB PET imaging. This in vivo evaluation has not been previously performed in the literature. Methods: Transgenic mice (TS65Dn) and its littermate wild type (WT), were evaluated with 11C-PIB PET imaging in 3 different ages: 2, 5 and 14-month-old (WT=6 and Ts65Dn=6 animals). 11C-PIB (±18 MBq) was injected intravenously and 30 min after tracer injection a static PET image of 20 min was acquired in a PET scanner for small animals (β-cube, Molecubes) with the animal anesthetized with isoflurane 2-3% in oxygen. Images were quantified by PMOD™ software using T1-MRI as brain template and the data was expressed in SUV (Standardized Uptake Value) mean±standard deviation in different brain regions: striatum, cortex, hippocampus, thalamus, cerebellum, brainstem and midbrain. Comparisons were made by 2-way ANOVA (SPSS). Statistical differences were considered significant when P≤0.05. Results: 11C-PIB uptake profile is similar in the transgenic and WT animals, showing increased uptake at 5 months of age, compared to 2 months, indicating ongoing myelination process. At 14 months, the 11C-PIB uptake is significantly lower in the Ts65Dn compared to the WT animals indicating that the Down syndrome animals are more affected by myelin loss caused by aging process. The 11C-PIB uptake (SUV) was, WT 2-month-old group: striatum 0.042±0.007, cortex 0.049±0.006, hippocampus 0.043±0.008, thalamus 0.041±0.007, cerebellum 0.044±0.007, brainstem 0.052±0.009 and midbrain 0.051±0.012; WT 5-month-old (compared to 2-months-old): 0.095±0.018 (p=0.001), 0.107±0.0220(p= 0.001), 0.104±0.027 (p=0.001), 0.095±0.026 (p=0.001), 0.109±0.024 (p=0.001), 0.118±0.035 (p=0.002) and 0.117±0.051 (p=0.001) in the respective brain regions; WT 14-month-old (compared to 2-months-old): 0.086±0.031 (p=0.005), 0.105±0.034 (p= 0.001), 0.093±0.032 (p=0.002), 0.089±0.033 (p=0.002), 0.100±0.022 (p=0.001), 0.105±0.037 (p=0.001) and 0.085±0.038 (p=0.004) in the respective brain regions. For the Ts65Dn 2-month-old, the 11C-PIB SUV was: striatum 0.038±0.013, cortex 0.037±0.012, hippocampus 0.033±0.013, thalamus 0.034±0.013, cerebellum 0.038±0.015, brainstem 0.044±0.015 and midbrain 0.039±0.015; Ts65Dn 5-month-age was (compared to 2-months-old)

Kofi Deh, Memorial Sloan Kettering Cancer Center, kofi.deh@gmail.com

Abstract Body : INTRODUCTION: Despite an unprecedented increase in spatio-temporal resolution of 13C metabolic imaging afforded by dissolution dynamic nuclear polarization (dDNP), the currently achievable spatial resolution is unsatisfactory. Even the application of compressed sensing and variable-resolution imaging (1,2) may not be sufficient to obtain satisfactory spatial resolution for metabolic products produced in low concentrations or for probes with low polarization rates (e.g. [1-13C] dehydoxyascorbate, [2-13C] pyruvate, [6-13C] arginine, [1-13C] glutamine) (3) . Researchers have proposed super-resolution (SR) of the acquired 13C image using a corresponding anatomic proton image as a reference (4, 5) . However, this method depends on accurate registration and similarity of contrast between the proton and carbon images as well as accurate segmentation of the proton image, which may not be reasonable with complicated anatomy. This abstract shows that it is possible to generate high resolution 13C metabolic maps by transferring high spatial frequencies from a single highresolution image of a thermal 13C phantom to a low-resolution 13C metabolite map, thus potentially realizing high spatial resolution HP MRI from existing data. METHOD: SR of 13C images via high-frequency transfer was demonstrated using multi-echo EPI data of phantoms and live mice on a preclinical Bruker Biospec 3T with a 1H/13C 42 mm diameter birdcage coil. The thermal phantom for constructing the SR database consisted of syringes and spheres containing 13C-Urea. Hyperpolarized 13C images of a mouse injected with 13C-pyruvate were acquired and separated into metabolite maps using a least-squares algorithm (IDEAL) (6) . Theory: A MATLAB implementation of the example-based super-resolution algorithm (7) was used for this experiment. It extracts and stores overlapping patches from the feature maps of an input high resolution image and its blurred, downsampled version. By ensuring that the candidate patch for adding high-frequency content to the low-resolution image being upsampled matches not only the low-resolution patch being replaced but also the overlapping pixels, an image with a consistent appearance is generated. Fig. 1a describes the extension of this approach to hyperpolarized 13C data.RESULTS: Fig. 1b presents a scheme for generating high-resolution 13C maps by separating chemical species present in multi-echo EPI data with IDEAL and then upsampling the individual maps using SR. Figure 1c shows both the proton and the carbon EPI images of syringes and balls filled with 13C-labeled chemicals used for generation of the SR database at varying target resolution. The SR database is created from the first echo magnitude image of the carbon EPI acquisition and used to upsampled the 13C pyruvate map extracted with IDEAL from a multi-echo EPI imaging data of a mouse injected with 13C-pyruvate. Improved detail can be seen in the SR image compared to bicubic interpolation, which is the best performing standard approach (Fig. 1d) . Extending this further, a noisy 13C-alanine map is upsampled with bicubic interpolation as well as with SR, and then overlaid on an anatomic image (Fig. 1e ). The latter method shows the best visual improvement in visualization of 13Calanine distribution in the heart and kidneys, including spatial resolution of the heart chambers ID: GA338 Dual imaging of the peripheral nerves with chlorin tracers JUNIOR GONZALES, Memorial Sloan Kettering Cancer Center, gonzalj4@mskcc.org Category: New Chemistry, Materials & Probes Abstract Body : Background: Although chemotherapy-induced peripheral neuropathy (CIPN) represents a global, growing issue, affecting millions world-wide, no methods exist to quantitatively measure this dysfunction. 85% of patients in the form of intense burning, tingling, cramps, numbness and acute pain [1] . CIPN is the most common neurotoxic side-effect of antineoplastic agents such as cis-platin and paclitaxel, and it often has long lasting impact on a patient's quality of life [2] . To address this urgent medical challenge, we turned our awareness to an isolated venom peptide Hsp1a. Hsp1a was isolated from Homoeomma Spec. Peru (a tarantula from the rainforest of Peru), and it is highly selective for sodium channel, NaV1.7 [3] . In this work, we have synthesized a red fluorescent agent capable to as well chelate copper-64 that simultaneously features a novel NaV1.7-targeted peptide, Hsp1a. Our multi-modal imaging agent is a conjugate which is fluorescent and PET active. Combining these two imaging methods allowed us to determine selectivity of our agent not just on a whole-body level, but also at subcellular resolution. We also used human tissue (a human vagus nerve) to corroborate that our Hsp1a imaging agent is selective after topical administration. Herein, we report the first radiolabeled Hsp1a peptide, for the selective targeting of sodium channel NaV1.7, a previously validated marker of peripheral neuropathy. Methods: We synthesized ChL-Hsp1a by conjugating a chlorin tracer [4] to a Hsp1a peptide. ChL-Hsp1a was used to synthesize [64Cu]Cu-ChL-Hsp1a. Mice were intravenously injected with ChL-Hsp1a, PBS and block formulation or with [64Cu]Cu-ChL-Hsp1a and their sciatic nerves were exposed 30 min and then resected. For the fluorescence imaging, animals were intravenously injected with ChLHsp1a (10 μM, 1 nmol in 100 μL PBS) or a block formulation of ChLHsp1a (ChL-Hsp1a, 10 μM, 1 nmol, and Hsp1a, (204 μM, 21 nmol, in 100 μL of PBS), n = 6, or PBS (n = 6). Similarly, for the bioluminescence, animals were injected with [64Cu]Cu-ChL-Hsp1a (3.5−4.2 MBq) dissolved in 100 μL PBS and a block formulation [64Cu]Cu-ChL-Hsp1a , and Hsp1a (204 μM, 21 nmol in 200 μL of PBS, n = 6). Using a fluorescence stereoscope, we studied the fluorescence signal and accumulation in mouse sciatic and brachial plexus nerves 30 min after intravenous injection of ChL-Hsp1a (1 nmol, 10 µM of ChL-Hsp1a in 100 µL of PBS) or PBS (n=3/group). Results: We successfully labeled Hsp1a with a chlorin tracer to afford ChL-Hsp1a (25%). [64Cu]Cu-ChL-Hsp1a was obtained in >95% and in 40-50% radiochemical yield (ndc). We found that injections with ChL-Hsp1a (1 nmol, 10 µM in 100 µL of PBS) resulted in high fluorescence coming from the sciatic nerves with elevated 93-fold compared to sciatic nerves from mice injected with block. In a similar pattern, animals injected with [64Cu]Cu-ChL-Hsp1a (3.5−4.2 MBq) resulted in high bioluminescence signal coming from the sciatic nerves with elevated 48-fold compared to sciatic nerves from mice injected with block. We also found that mouse sciatic nerves were clearly visible with the Lumar fluorescence stereomicroscope (n=3/group). This data suggests that it is possible to delineate peripheral nerves with high contrast and that the same agent after radiolabeling can be used in Cerenkov bioluminescence.

Conclusions: ChL-Hsp1a and [64Cu]Cu-ChL-Hsp1a peptides accumulated and highlighted peripheral nerves in mice in vivo. The statistically significant fluorescence/Cerenkov signal observed in peripheral nerves in mice, as a result of using fluorescent/PET active Hsp1a peptides, is clear evidence of peripheral nerve localization. This fluorescent/PET approach could be especially beneficial to stratify patients, as treatments often continue to rely on subjective assessments of these symptoms by physicians after patients' self-reporting. ChL-Hsp1a and [64Cu]Cu-ChL-Hsp1a peptides are promising tools that could potentially help physicians to rapidly identify peripheral neuropathy in patients.

Copper-dopped responsive nanomaterials with improved Pt-based chemotherapy and chemodynamic therapy for cancer Yuanyuan Ma, Southeast University, mayuanyuan.sun@gmail.com

Abstract Body : Cis-diamminedichloridoplatinum(II), known as cisplatin, and subsequent expansion of platinum-based chemotherapy drug family have been used against a wide range of tumors and achieved remarkable success for cancer therapy [1] [2] . However, cisplatin shows its restricted clinical use because of the highly toxic nature, which will cause indiscriminate damage to both cancer cells and normal cells, further inducing severe side effects like nephrotoxicity and hepatotoxicity [3] . Moreover, rapid development of resistance is another major concern of cisplatin. Among all the effecting factors, overproduced glutathione(GSH) is an important one to induce resistance during the cancer therapy [4] . Thus, it is important to develop a novel method to enhance Pt-based chemotherapeutic effect and improve drug resistance, especially by depleting GSH. In this study, hollow-structured mesoporous organosilica nanoparticle(HMON) was used as the cisplatin carrier, owing to the large internal cavity and increased loading capacity [5] . Cubased metal-organic framework(Cu-MOF) was used to encapsulate HMON to form a kind of hybrid nanomaterials(Hcis@Cu-MOF). MOF was responsive to the acid conditions and could be easily biodegraded [6] , which enables the copper ions and loading cisplatin to release in a regulated manner. Moreover, copper ions within the framework of the nanometerials were capable of depleting intracullar GSH, leading to the enhanced cispaltin therapeutic effect. Additionally, copper ion was one kind of Fenton-like agent, which was able to react with intracellular hydrogen peroxide and subsequent produce highly toxic hydroxyl radicals(· OH), resulting in cell apoptosis, which was the chemodynamic therapy [7] . Thus, this synthesized nanomaterials could responsive to the mild acid conditions and release copper and cisplatin precisely. Additionally, the copper ions in the nanomaterials could enhance cisplatin therapeutic effect and combine chemodyamic therapy with chemotherapy via GSH consumption and · OH production, achieving the improved cancer-killing efficacy. Figure 1a -b showed the uniform morphology of HMON and the synthesized Hcis@Cu-MOF monitored by transmission electron microscope. The HMON showed 40nm, while after encapsulation of Cu-MOFs, a thin layer of MOF structure could be observed evidently. In addition, the size of the nanomaterials would show great enhanced permeability and retention(EPR) effect. Copper ions and cisplatin could be released quickly at pH 5.5 (Figure 1c-d) . While at pH 7.4, copper ions and cisplatin could only release less than 30%, indicating the great pH responsiveness of Hcis@Cu-MOF. Figure 1e -f displayed the ability of GSH depletion and · OH production, which were measured by 5,5′-Dithiobis(2-nitrobenzoic acid) and methylene blue(MB) degradation, showing the concentration/time-dependent manner. Cytotoxic effect was mesured by 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide as shown in Figure 1g . Compared to free cisplatin, Hcis@Cu-MOF showed higher cytotoxicity. Moreover, intracellular GSH level and · OH production were also measured as shown in Figure 1h -i. Cells treated by Hcis@Cu-MOF showed a significant decrease of GSH, while the level of · OH increased a lot. The biodistribution of Cy-7 labeled Hcis@Cu-MOF in mice were displayed in Figure 1j . The results showed the nanomaterials possessed great EPR effect and could reach the tumor in 4 hours after intravenous administration. The in vivo results were shown in Figure 1k , showing that the tumor could be efficiently inhibited compared with free cisplatin. Moreover, mice treated by Hcis@Cu-MOF had a higher survival rate as shown in Figure 1l . To conclude, the synthesized Hcis@Cu-MOF could deliver cisplatin efficiently to the tumor site. In addition, the nanomaterals was pHsensitive and could release cisplatin and copper ions within the cancer cells. Copper ions dopped in the nanomaterials could deplete intracellular GSH and improve cisplatin efficiency. Meanwhile, copper ions were able to react with the abundunt hydrogen peroxide, producing , leading to cell death and great chemodynamic therapeutic effect. Furthermore, Hcis@Cu-MOF could inhibit tumor growth in mice, indicating superior biomedical applications for cancer therapy in the future.

Evan Unger, MVT Pharma LLC, e.unger@mvtpharma.com

Abstract Body : Cardiovascular disease is the leading cause of death and disability worldwide. Efforts to remove occlusive thrombi via thromboembolectomy or by mechanical disruption and/or biochemical dissolution have demonstrated efficacy; however, these therapeutic options are time consuming and are accompanied by substantial risk of hemorrhagic complications. Enhance safety and efficacy of thrombus removal have high potential clinical impact. Sonothrombolysis (SL) refers to the use of ultrasound (US) to disrupt thrombi. However, there is a trade-off between time/efficiency and damaging healthy tissue. Ultrasound mediated microbubble (MB) cavitation, which creates shear stress and induces biological response such as nitrite oxide release, can accelerate the disruption of the thrombi while keeping delivered energy low which can decrease the risk of damaging healthy tissue. Moreover, nanodroplets (NDs) can be better for SL than MBs due to their capacity to permeate better to the thrombi. In this study, we developed a safe and stable lipid-based MB and ND formulation. The formulation offers a targeted ND and MB with the capability to recognize fibrin through a fibrin binding peptide (FBP) ligand conjugated to the lipid surface of the MBs. We show in vitro that the targeted MB and ND bind to fibrin clots. We show in a rat hindlimb model of microvascular occlusion (MVO) that ultrasound therapy with NDs are able to disrupt the clots upon US mediated cavitation and relieve MVO (Fig. 1) . These results are proof of concepts for the usage of ND to support further preclinical development and IND-enabling studies of these ND for MVO therapeutic indication. Figure 1 : Treatment of microcirculation obstruction (MVO) with US targeted MB cavitation using phase change contrast agent. During MVO stage, the blood flow to the hindlimb was substantially reduced as indicated by contrast specific ultrasound imaging. Post ultrasound treatment with infusion of NDs, the blood flow to the hindlimb was restored.

Full Name of Abstract's 1st Author : Evan C Unger Abstract Body : Introduction: Sphingosine-1-phosphate (S1P) is a bioactive lipid mediator involved in many biological actions, including vascular homeostasis and immunosurveillance. S1PR1 plays the most important role in immunosurveillance. We speculated that S1P/S1PR1 axis would sense the homeostasis alternations under microgravity during the spaceflight. In this study, the molecular imaging quantification through PET/CT was used to explore the S1PR1 expression changes in hindlimb unloading (HU) model of simulated weightlessness. Methods: Eleven healthy Balb/c mice aged 8 weeks were randomly divided into the control group (CTR, n = 5) and the HU group (HU, n = 6). The preparation of the simulated weightless model was mainly based on the classic tail-suspension method proposed and improved by Morey-Holton and Chen Jie et al.1 After 4 weeks of modeling, dynamic PET/CT imaging was performed for 30 min including 25 frames (10×3s; 3×10s; 4×60s; 6×150s; 2 ×300s). The labeling and synthesis of radiopharmaceuticals (18F-10a) were based on previous published literature. 2 The dynamic raw data was analyzed by Carimas software. With the time activity curve (TAC) of left ventricular as the input function, the original TAC of vital organs (including brain, lung, liver, spleen, renal, muscle, bone, brown adipose tissue) was fitted by Logan REF modeling to obtain corresponding distribution volume rate (DVR) and binding potential (BP).3 Area under the curve (AUC) of TACs, DVR and BP were compared between the CTR and HU group, and p Results: For the above indicators, there was no significant differences for most organs (including brain, lung, liver, spleen, renal, muscle, bone). However, we found that the AUC of the BAT TACs (2563.97 ± 594. 34 vs 1714.76 ± 536.50, p ) in the HU group was significantly higher than that of the CTR group. Simultaneously, the DVR (2.67 ± 0.30 vs 1.56± 0.32, p ± 0.30 vs 0.56 ± 0.32, p were significantly higher than that of the CTR group ( Figure 1 ). Conclusions: After mid-term simulated weightlessness (hindlimb unloading 4 weeks), the specific association of S1PR1 in the BAT was significantly increased, indicating that hindlimb unloading may induced increased S1PR1 expression in BAT. Although the molecular mechanism behind this phenomenon remains to be explored, S1P/S1PR1 axis warrants further investigation regarding BAT metabolism under microgravity.

Peter Clark, UCLA, pclark@mednet.ucla.edu

Abstract Body : PET imaging with the glucose analogue 2-deoxy-2 [18F] fluoroglucose ([18F] FDG) is foundational for the diagnosis and staging of many types of cancer including lung cancer. The biochemical pathway that enables [18F] FDG to concentrate in cancer cells is well described: [18F] FDG is transported across the cell membrane through the GLUT glucose transporters and phosphorylated by the hexokinase enzymes to [18F] FDG-6-phosphate. However, our understanding of the signaling pathways that activate the GLUT transporters and the hexokinase enzymes and that drive [18F] FDG accumulation in cancer cells is limited to a few major pathways in a limited number of cancer types. The full scope of pathways that regulate [18F] FDG accumulation in cancer cells remains largely unexplored. Squamous cell lung cancer (SqCLC) constitutes 30% of non-small cell lung cancer. SqCLC cells consume elevated levels of glucose compared to lung adenocarcinoma and surrounding lung tissue. However, few pathways have been identified that activate glucose consumption and drive [18F] FDG accumulation in SqCLC. The goal of this work is to discover new pathways that drive glucose consumption and [18F] FDG accumulation in SqCLC. To identify new pathways that drive glucose consumption and [18F] FDG accumulation in SqCLC, we used a high-throughput glucose consumption assay. In this assay, cells are treated with 2-deoxyglucose and 2deoxyglucose-6-phosphate levels are measured using a coupled enzyme assay that quantitatively converts the 2-deoxyglucose-6-phosphate to a light output. We measured glucose consumption in three SqCLC cell lines -H520, SK-MES-1, and H596 -treated with 3555 bioactive small molecules. We further validated the results of the high-throughput screen and studied the mechanism by which one of the compounds identified regulates glucose consumption. We identified and validated >50 compounds that blocked glucose consumption in at least one of the SqCLC cell lines. We focused on validated compounds that were not previously known to affect glucose consumption in SqCLC, that significantly decreased glucose consumption by >70%, that blocked glucose consumption in more than one cell line, that had a glucose consumption IC50 M, and that had been in humans. Based on these criteria, we chose to further study the kinase inhibitor Pacritinib. Pacritinib strongly blocks glucose consumption with an IC50 of 2 µM in H520, SK-MES-1, and H596 cells in culture and blocks [18F] FDG accumulation in these cell lines in vivo while having no effect on [18F] FDG accumulation in other highly glycolytic tissue including the heart and brain. Pacritinib blocks mRNA and protein levels of Hexokinase 1 and 2, leading to a decrease in hexokinase activity. The effect of Pacritinib on glucose consumption was rescued by Hexokinase 1 overexpression. These data suggest that Pacritinib blocks glucose consumption by inhibiting Hexokinase 1 expression. Pacritinib targets various kinases, including FLT3, JAK2, and TYK2 with low nanomolar IC50 values. Additional small molecule inhibitors selective for JAK2 and TYK2 but not FLT3 had no effect on glucose consumption in the SqCLC cell lines. However, additional and selective FLT3 small molecule inhibitors blocked glucose consumption, and FLT3 overexpression rescued the effect of Pacritinib on glucose consumption in SqCLC cell lines. These results suggest that Pacritinib blocks glucose consumption through the inhibition of FLT3. Our results demonstrate that FLT3 is a major driver of [18F] FDG accumulation in SqCLC by increasing Hexokinase 1 expression. The implications of this study are that [18F] FDG could potentially function as a pharmacodynamic biomarker of FLT3 inhibitor efficacy and that the degree of [18F] FDG accumulation could potentially serve as a measure of FLT3 pathway activity in SqCLC. Abstract Body : Being able to noninvasively control any neural circuit with a combination of spatial, temporal, cellular, and molecular pathway precision could allow for more precise treatment of neuropsychiatric disorders with fewer side effects. Recently, a paradigm was developed that allows such neuromodulation to be performed noninvasively by combining focused ultrasound, viral vectors, and chemogenetics. In this approach called Acoustically Targeted Chemogenetics1, or ATAC, focused ultrasound (FUS) is used to briefly open tight junctions in the blood-brain barrier (FUS-BBBO) in spatially specific brain region(s)2, allowing localized delivery of adeno-associated viral vectors (AAVs) encoding chemogenetic receptors3-5 to neurons and enabling their subsequent excitation or inhibition with bio-orthogonal, orally bioavailable drugs. FUS targets both deep and cortical brain regions with millimeter precision6; AAVs allow for the selection of a neuronal population using cell-specific promoters7,8; and chemogenetic receptors3-5enable temporally-specific neuromodulation. Relative to what is currently being studied and tested, ATAC is the only neuromodulation technique that achieves temporal, spatial, cell-type, and molecular specificity noninvasively. However, its spatial precision is limited by the precision of FUS itself. While FUS allows for millimeter precision, many brain regions, such as a single layer of cortex or specific hypothalamic nuclei, are significantly smaller than the focus of FUS both in mice and larger animals. Such precision cannot be feasibly achieved in vivo with current FUS devices. To solve this limitation, a novel strategy is proposed that decouples the precision of FUS from noninvasive chemogenetic neuromodulation. Established tools for gene editing are combined with FUS delivery to achieve transduction precision comparable to an invasive intracranial injection. To achieve this, two independent sessions of FUS-BBBO are used, and a spatial overlap of expression between the fields expressing the gene-editing target and the gene-editing enzyme is enabled. The results of this intersectional ATAC strategy, or iATAC, show that expressing both of these components is safe in mice after a systemic delivery and show ~10-fold improvement in targeting precision.

Renal clearable theranostic nanoplatforms for gastrointestinal stromal tumors

The Society for Fetal Urology consensus statement on the evaluation and management of antenatal hydronephrosis

Comparison of the Effectiveness of the Renogram, the Serial Renal Scan and the Diuretic Half Time according to the Renal Function for Interpreting a Diuretic DTPA Scan following

A DROP-IN Gamma Probe for Robot-assisted Radioguided Surgery of Lymph Nodes During Radical Prostatectomy

Glutamate carboxypeptidase II in diagnosis and treatment of neurologic disorders and prostate cancer

Prostate-Specific Membrane Antigen-Guided Surgery

Linker Modification Strategies To Control the Prostate-Specific Membrane Antigen (PSMA)-Targeting and Pharmacokinetic Properties of DOTA-Conjugated PSMA Inhibitors

Reduced lung-cancer mortality with low-dose computed tomographic screening

Highly efficient activatable MRI probe to sense myeloperoxidase activity

ICASSP 2019 -2019 IEEE International Conference on Acoustics, Speech and Signal Processing

NanoLuc Complementation Reporter Optimized for Accurate Measurement of Protein Interactions in Cells

Engineered Luciferase Reporter from a Deep Sea Shrimp Utilizing a Novel

CRISPR-Mediated Tagging of Endogenous Proteins with a Luminescent Peptide

IRBM 2019

Full Prescribing Information for DEFINITY®

The sensitivity and specificity of clinical measures of sport concussion: three tests are better than one

Imaging of Glial Cell Activation and White Matter Integrity in Brains of Active and Recently Retired National Football League Players

Molecular imaging of neuroinflammation in patients after mild traumatic brain injury: a longitudinal (123) I-CLINDE single photon emission computed tomography study

as New PET Tracers for TSPO: A Comparison with

PK11195 in a Rat Model of Herpes Encephalitis

Targeted auger electron-emitter therapy: Radiochemical approaches for thallium-201 radiopharmaceuticals

Fluorine-19 Cellular MRI Detection of In Vivo Dendritic Cell Migration and Subsequent Induction of Tumor Antigen-Specific Immunotherapeutic Response

Quantification and characterization of virus-like particles by size-exclusion chromatography and nanoparticle tracking analysis

Development and Characterization of Novel LipoCEST Agents Based on Thermosensitive Liposomes

Image/Figure Caption: Figure 1 -Co(II) shift agents

Sonosensitive theranostic liposomes for preclinical in vivo MRI-guided visualization of doxorubicin release stimulated by pulsed low intensity non-focused ultrasound

Phosphorylcholine: a novel second messenger essential for mitogenic activity of growth factors

Overexpression of Choline Kinase Is a Frequent Feature in Human Tumor-derived Cell Lines and in Lung, Prostate, and Colorectal Human Cancers

Near Infrared Fluorescent Imaging of Choline Kinase Alpha Expression and Inhibition in Breast Tumors

Metal Ion Sensing Using Ion Chemical Exchange Saturation Transfer 19F Magnetic Resonance Imaging

Elucidating the Ordering in Self-Assembled Glycocalyx Mimicking Supramolecular Copolymers in Water

Acoustic reporter genes for noninvasive imaging of microorganisms in mammalian hosts

Deploying immunotherapy in pancreatic cancer: defining mechanisms of response and resistance

Desmoplasia and chemoresistance in pancreatic cancer

Resistance mechanisms to immune-checkpoint blockade in cancer: tumor-intrinsic andextrinsic factors

Facilitating T cell infiltration in tumor microenvironment overcomes resistance to PD-L1 blockade

Impacting Pancreatic Cancer Therapy in Heterotypic in Vitro Organoids and in Vivo Tumors with Specificity-Tuned, NIR-Activable Photoimmunonanoconjugates: Towards Conquering Desmoplasia?

Optical projection tomography as a tool for 3D microscopy 668 and gene expression studies

Ultramicroscopy: three-dimensional visualization of neuronal networks in the whole mouse brain

Ionescu-Tirgoviste, C. et al. A 3D map of the islet routes throughout the healthy human pancreas

Topologically selective islet vulnerability and self-sustained downregulation of markers for β-cell maturity in streptozotocin-induced diabetes

Image/Figure Caption: Figure 1. a) Chemical structures of Gd complexes studied in this work

40, 24 h, 37°C) in the presence or absence of BSA and BSA-Ald; d) Hydrolysis of BSA-Ald bound Gd-CHyd and Gd-1,4 monitored by longitude relaxation at 60 MHz (PBS

Artificial reporter gene providing MRI contrast based on proton exchange

Sensitive NMR detection of cationic-polymer-based gene delivery systems using saturation transfer via proton exchange

Clinical value of FDG-PET/CT in bacteremia of unknown origin with catalase-negative gram-positive cocci or Staphylococcus aureus

Molecular imaging of bacterial infections: Overcoming the barriers to clinical translation

Evaluation of Macroaggregated Albumin Particle Sizes for Use in Pulmonary Shunt Patient Studies

A peptide inhibitor of FOXP3 impairs regulatory T cell activity and improves vaccine efficacy in mice

Image/Figure Caption: Fig. 1a : Representative images from microPET/CT of healthy mice injected with each formulation of dendrimer. Fig. 1b: Representative images from microPET/CT of orthotopic U87 glioblastoma mice

Full Name of Abstract's 1st Author : Béatrice LOUIS First Name: Béatrice Last Name: Louis Email: beatrice.louis@etu.univ-amu.fr Organization: AIX-MARSEILLE UNIVERSITE Country: France References: 1. Chapman TW, Hill RA

Faria Dde P, et al

Multiple Sclerosis and Related Disorders

An Effective Immuno-PET Imaging Method to Monitor CD8-Dependent Responses to Immunotherapy. Cancer research

Fluorescence imaging of peripheral nerves by a Nav1.7-targeted inhibitor cystine knot peptide

Recent advances of cocktail chemotherapy by combination drug delivery systems

Self-Assembled Copper-Amino Acid Nanoparticles for in Situ Glutathione "AND" H2O2 Sequentially Triggered Chemodynamic Therapy

Image/Figure Caption: Figure 1a-b Representative TEM images of HMON and Hcis@Cu-MOF

The safety of deFinity and Optison for ultrasound image enhancement: a retrospective analysis of 78,383 administered contrast doses

Paramagnetic fluorinated nanoemulsions for sensitive cellular fluorine-19 magnetic resonance imaging

Mortality and causes of death in people diagnosed with HIV in the era of highly active antiretroviral therapy compared with the general population: an analysis of a national observational cohort. The Lancet Public Health

Mazor T et al. Clonal expansion and epigenetic reprogramming following deletion or amplification of mutant IDH1

2-Hydroxyglutarate produced by neomorphic IDH mutations suppresses homologous recombination and induces PARP inhibitor sensitivity

Image/Figure Caption: A) After administration of 2H-labeled substrates (for example

Diels-Alder reaction for tumor pretargeting: in vivo chemistry can boost tumor radiation dose compared with directly labeled antibody

Upper panel, primary tumors: After validation of tumor formation with bioluminescent imaging (BLI), mice with MDA-MB-231 breast cancer were systemically injected with MNP-Cy7-TCO, followed by Tet-Cy5.5 24 hrs later. Accumulation of the particle construct and the small molecule trigger were apparent 24 hrs post injection, respectively. Lower panel, metastatic tumors: Validation of metastases was again performed with BLI

Dissolution of thrombotic arterial occlusion by high intensity, low frequency ultrasound and dodecafluoropentane emulsion: an in vitro and in vivo study

Interactions between ultrasound stimulated microbubbles and fibrin clots

Quantification of myocardial blood flow with ultrasound-induced destruction of microbubbles administered as a constant venous infusion

2 Luo, Z. et al. Syntheses and in vitro biological evaluation of S1PR1 ligands and PET studies of four F-18 labeled radiotracers in the brain of nonhuman primates

18F]-C-SNAT4: an improved caspase-3-sensitive nanoaggregation PET tracer for imaging of tumor responses to chemo-and immunotherapies

Bioorthogonal cyclization-mediated in situ self-assembly of small-molecule probes for imaging caspase activity in vivo

Positron emission tomography imaging of drug-induced tumor apoptosis with a caspase-triggered nanoaggregation probe

Design of a novel class of proteinbased magnetic resonance imaging contrast agents for the molecular imaging of cancer biomarkers

TB-IgG-IR700 (n=4) and TB-Gr1-IR700 (n=3) with (B) averaged mean fluorescence intensity of (left to right) the spleen, kidney, lungs, small intestine, liver, stomach, heart, blood and tumor. (C) Millimolar concentrations normalized to BALB/c mouse spleen weight from 4T1-luc tumor-bearing TB-Ctrl (n=3), Gr1-IR700 (n=3) and Gr1-IR700+PIT (n=3) groups illustrated respectively for (upper row: left to right) valine, isoleucine, lactate, acetate, glutamate, succinate, glutamine, glutathione and

Full Name of Abstract's 1st Author : James D Barnett First Name: James Last Name: Barnett Email: jbarne55@jhmi

At 48h p.i, the blood, SW1222 tumor, non-tumored liver and kidney uptakes

At 6h p.i., the corresponding tumor-to-organ activity ratios (T:NT) for blood, non-tumored liver preliminary absorbed doses to tumor, blood, non-tumored liver, and kidney were respectively 175 mGy/MBq, 6 mGy/MBq (TI= 31

Alpha radioimmunotherapy using 225 Ac-proteus-DOTA for solid tumors -safety at curative doses

Image/Figure Caption: Coronal multiplanar reconstruction (MPR) microCT (A) and PET (B) images, and maximum intensity projection (MIP) administration (Fenestra LC

Table (D) is a summary of the estimated Lu-177 dosimetry for the tumor and abdominal organs, and corresponding therapeutic index. (E) and (F) are histograms and tables showing the tumor and abdominal organs %ID/g uptake (in vivo (E) and ex vivo (F)), and corresponding tumor-to-organs ratios. First Name: Alexandre Last Name: Le Roux Email: lerouxa@mskcc.org Organization: Memorial Sloan Kettering Cancer Center Country: United States References

Noninvasive MR imaging-guided focal opening of the blood-brain barrier in rabbits

Evolving the lock to fit the key to create a family of G protein-coupled receptors potently activated by an inert ligand

DREADDs for Neuroscientists

Unparalleled control of neural activity using orthogonal pharmacogenetics

A Randomized Trial of Focused Ultrasound Thalamotomy for Essential Tremor

References: References: 1

Image/Figure Caption: Figure 1

Distribution of axon diameters in cortical white matter: an electron-microscopic study on three human brains and a macaque

Nabraj Sapkota et al., Characterization of spinal cord white matter by suppressing signal from hindered space

Sex differences in the human corpus callosum microstructure: a combined T2 myelinwater and diffusion tensor magnetic resonance imaging study

The permeability coefficient of water in the cell membrane and the diffusion coefficient in the cell interior

Image/Figure Caption: Figure 1. Results (mean±SD) for (a)FA, (b)MD, (c)RD and (d)AD changes in ECS volume with different noise level: blue -noise free

On the Mechanism of Membrane Damage by Complement: The Effect of Length and Unsaturation of the Acyl Chains in Liposomal Bilayers and the Effect of Cholesterol Concentration in Sheep Erythrocyte and Liposomal Membranes

Image/Figure Caption: Porphyrin-lipid microbubbles enable longitudinal quantitative biodistribution measurements via PET imaging, hyperspectral fluorescence, and gamma counting. A) A series of microbubbles with varying lipid chain lengths (C16, C18) and shell

Cancer statistics, 2020. CA Cancer J Clin

The Miss Rate for Colorectal Adenoma Determined by Quality-Adjusted, Back-to-Back Colonoscopies

ICES Report: New Findings about the Risks and Limitations of Colonoscopy Used in the Early Detection of Colorectal Cancer

Hyperpolarized MRI with Silicon Micro and Nanoparticles: Principles and Applications

3. Maeda, A. et al. In Vivo Imaging Reveals Significant Tumor Vascular Dysfunction and Increased Tumor Hypoxia-Inducible Factor-1α Expression Induced by High Single-Dose Irradiation in a Pancreatic Tumor Model

Preclinical In vivo Imaging for Fat Tissue Identification, Quantification, and Functional Characterization

Image/Figure Caption: MRI biomarkers for NALFD

Vegfr2-targeted three-dimensional ultrasound imaging can predict responses to antiangiogenic therapy in preclinical models of colon cancer

Pharmacokinetic Modeling of Targeted Ultrasound Contrast Agents for Quantitative Assessment of Anti-Angiogenic Therapy: a Longitudinal Case-Control Study in Colon Cancer

US imaging of tumor angiogenesis with microbubbles targeted to vascular endothelial growth factor receptor type 2 in mice

Spatial Characterization of Tumor Perfusion Properties from 3D DCE-US Perfusion Maps are Early Predictors of Cancer Treatment Response

Hyperpolarized 13C-pyruvate MRI detects real-time metabolic flux in prostate cancer metastases to bone and liver: a clinical feasibility study

A metabolite-specific 3D stack-of-spiral bSSFP sequence for improved lactate imaging in hyperpolarized [1-13C]pyruvate studies on a 3T clinical scanner

Coil combination methods for multi-channel hyperpolarized 13 C imaging data from human studies

Image/Figure Caption: Figure 1. (A) Representative HP 13C

Tomographic imaging using the nonlinear response of magnetic particles

The X-space formulation of the magnetic particle imaging process: 1-D signal, resolution, bandwidth, SNR, SAR, and magnetostimulation

Novel Scanning Strategies in x-Space Magnetic Particle Imaging for Improved Imaging Performance and Theranostic Applications

A High-Throughput, Arbitrary-Waveform, MPI Spectrometer and Relaxometer for Comprehensive Magnetic Particle Optimization and Characterization

Quantitative magnetic particle imaging monitors the transplantation, biodistribution, and clearance of stem cells in vivo

The X-space formulation of the magnetic particle imaging process: 1-D signal, resolution, bandwidth, SNR, SAR, and magnetostimulation

Magnetostimulation limits in magnetic particle imaging

Dynamics of chain formation and decay for super-resolution Magnetic Particle Imaging

Signal encoding in magnetic particle imaging: properties of the system function

Sigma-1 receptor agonists as therapeutic drugs for cognitive impairment in neuropsychiatric diseases

Chiral resolution of serial potent and selective σ1 ligands and biological evaluation of (-)-[18F]TZ3108 in rodent and the nonhuman primate brain

Image/Figure Caption: Figure 1. Quantification of brain expression σ1R via

Representative imagine with PET/MRI for Tg (Left) and Wt ( Right)mouse. (Upper: Coronal slice; Middle: Transaxial slice

The BP value of vital brain regions was obtained after fitting by 2TCM model. WB

Immune suppressive landscape in the human esophageal squamous cell carcinoma microenvironment

Nicotine increases microvascular blood flow and flow velocity in three groups of brain areas

C) and (D) were statistical analyses of SUVmax and Ki of sMLN and BLN with SUVmaxs

Pharmacological characterization of crotamine effects on mice hind limb paralysis employing both ex vivo and in vivo assays: Insights into the involvement of voltage-gated ion channels in the crotamine action on skeletal muscles

Analysis of nanoparticle delivery to tumors. Nat Rev Mater 1, 16014, 2016. 4. Attaluri, A. et al. Nanoparticle distribution and temperature elevations in prostatic tumours in mice during magnetic nanoparticle hyperthermia

Reoperation rates after breast conserving surgery for breast cancer among women in England: retrospective study of hospital episode statistics

Photodynamic Therapy Is an Effective Adjuvant Therapy for Image-Guided Surgery in Prostate Cancer

The role of reexcision for positive margins in optimizing local disease control after breastconserving surgery for cancer

A comparison of ink-directed and traditional whole-cavity re-excision for breast lumpectomy specimens with positive margins

Intraoperative margin assessment and re-excision rate in breast conserving surgery

Relationship between initial margin status for invasive breast cancer and residual carcinoma after re-excision

Predictors of re-excision among women undergoing breast-conserving surgery for cancer

Imaging and photodynamic therapy: mechanisms, monitoring, and optimization

Variability in reexcision following breast conservation surgery

Obtaining adequate surgical margins in breast-conserving therapy for patients with early-stage breast cancer: current modalities and future directions

Image/Figure Caption: Fig1.: PSMA as a target for breast cancer (A) Expression of PSMA

Name: Aditi Last Name: Shirke Email: aas151@case

Acknowledgements: We acknowledge Center for Biomedical Neuroscience (CBN) 2020-21, UT Health San Antonio, Texas for generously providing funding for this research. We also thank Glenn Fulbright for isotope production and Katie Strychalski for technical support. 

Cindy Rodriguez, CUNY Hunter College, cindy.rodriguez54@myhunter.cuny.edu

Abstract Body : Introduction: The synthesis of radioimmunoconjugates via the random attachment of bifunctional chelators to lysine residues yields heterogeneous products that can display suboptimal in vitro and in vivo behavior. In light of this, several site-specific and siteselective approaches to bioconjugation have been developed, including strategies based on unnatural amino acids, chemoenzymatic transformations, the manipulation of the heavy chain glycans, and the reduction and modification of interchain disulfides. [1] Yet each of these approaches comes with attendant drawbacks, for example the expense and complexity of incorporation unnatural amino acids into mAb. In this work, we describe the attachment of the chelator desferrioxamine (DFO) to the HER2-binding antibody pertuzumab via a novel, lysinetargeted approach to site-specific bioconjugation as well as its labeling with [89Zr]Zr4+ to create [89Zr]Zr-SSKDFO-pertuzumab. Subsequently, we compared the performance of [89Zr]Zr-SSKDFO-pertuzumab to traditionally synthesized [89Zr]Zr-DFO-pertuzumab via surface plasmon resonance (SPR) analysis, bead-based binding assays, and PET imaging experiments. Methods: Randomly modified DFO-pertuzumab was synthesized via the sequential modification of the antibody with an NHS-bearing, linear azide and a dibenzocyclooctyne-bearing variant of DFO (DBCO-DFO). In contrast, SSKDFO-pertuzumab was synthesized via the site-selective acyl-transfer of a branched bis-azide moiety onto the light chain of the antibody followed by the addition of DBCO-DFO. Both immunoconjugates were labeled with [89Zr]Zr4+ using standard protocols and purified via size exclusion chromatography to yield [89Zr]Zr-DFO-pertuzumab and [89Zr]Zr-SSKDFO-pertuzumab. The immunoreactivity of the radioimmunoconjugates was determined using Ni-NTA beads coated with purified HER2, and the kinetic binding parameters for each construct were determined via SPR. Finally, both [89Zr]Zr-SSKDFO-pertuzumab and [89Zr]Zr-DFO-pertuzumab were administered to mice bearing subcutaneous HER2-positive BT474 human breast cancer xenografts, and PET images were collected 6, 24, 48, 72, 96, 120 , and 144 h after administration. Results: The site-specific bioconjugation of SSKDFOpertuzumab was validated by ESI mass spectrometric analysis.[89Zr]Zr-SSKDFO-pertuzumab and [89Zr]Zr-DFO-pertuzumab were synthesized in >99% radiochemical yield, >99% radiochemical purity, and specific activities of ~2 mCi/mg. The immunoreactivities of both radioimmunoconjugates were >90%, and the binding constants of SSKDFO-pertuzumab and DFO-pertuzumab were found via SPR to be 0.8 x 10-9 M and 1.65 x 10-9 M, respectively. Finally ¾ and most importantly ¾ both [89Zr]Zr-SSKDFO-pertuzumab and [89Zr]Zr-DFOpertuzumab produced high contrast PET images in mice bearing BT474 xenografts. In each case, the tumor could clear be visualized as early as 24 h post-injection, and the accretion of the radioimmunoconjugates continued over the course of the experiment reaching maxima of ~75% ID/g at 120 h. Conclusion: Our lysine-mediated site-specific bioconjugation technology produced a radioimmunoconjugate, [89Zr]Zr-SSKDFO-pertuzumab, that is better defined and better characterized than a traditionally synthesized analog and exhibits excellent in vitro and in vivo behavior. We are currently working to validate [89Zr]Zr-SSKDFO-pertuzumab in other murine models of HER2-positive cancer. Abstract Body : Objectives: The Hepatitis E virus-like nanoparticles (HEVNPs) are small and safe protein capsids (≈ 30 d.nm). All the physical characteristics of the HEV are retained by the HEVNPs while the genome is lacking. The self-assembling ability of the HEVNPs enables them to act as cargo vehicles1,2. LXY30 is a peptide targeting to α3β1 integrin. Previously, LXY30 has been successfully used for targeting HEVNPs in breast cancer2. In order to evaluate LXY30-HEVNP for targeted drug delivery and theranostics of colorectal cancer, the NPs were first radiolabeled with 68Ga and their stability was evaluated. Subsequently, the uptake of [68Ga]Ga-DOTA-HEVNP-LXY30 in HCT 116 tumor cells was determined, and their ex vivo biodistribution was investigated in mice bearing human colorectal xenograft tumors. Methods: Conjugation of DOTA chelator to the LXY30-HEVNP lysine amino acids: DOTA-NHS ester (21.3 mg, 33.66 µmol) was dissolved in 0.01 M phosphate buffer (PB) pH 7.4 and added to LXY30-HEVNPs dispersed in 0.01 M PB pH 7.48 (1.05 mg/ml, 28.05 nmol). The final reaction volume was 1430 µl (adjusted pH to 7.3). The reaction was run overnight at 4oC, followed by a purification with two consecutive PD-10 columns. Gallium-68 radiolabeling: DOTA-HEVNP-LXY30 was dispersed in metal-free 0.25 M ammonium acetate buffer pH 7 (800 µl) and 68GaCl3 in 0.1 M HCl was added (80-140 MBq in 1 ml). The reaction mixture (pH 4.5) was incubated for 30 min at 60oC after which the radiolabeled HEVNPs were purified with a PD-10 column. Stability assay: The purified targeted NPs, [68Ga]Ga-DOTA-HEVNP-LXY30, were incubated in 0.01 M PBS, 50% human plasma, FeCl3 (10 mg/ml), or CO2-independent cell culture medium (Gibco) at 37oC for 1 h. Cell uptake: The HCT 116 (ATCC® CCL247™) cells (2.2 million cells/ml/well) were incubated at 37oC for 3 h with either the targeted 68Ga-HEVNP-LXY30, the non-targeted 68Ga-HEVNP, or 68GaCl3 (0.2 MBq/well). Ex vivo biodistribution: The evaluation of [68Ga]Ga-DOTA-HEVNP-LXY30 was performed in HCT 116 colon tumor-bearing female nude NMRI mice (n = 13, 7-8 weeks, 17-20 g) . [68Ga]Ga-DOTA-HEVNP-LXY30 were injected into the tail vein in 200 µl of PB pH . The mice were euthanized at 15, 30 and 60 min after the administration, and the organs were harvested, weighed and their radioactivity was counted on a gamma-counter. Results: The RCY of [68Ga]Ga-DOTA-HEVNP-LXY30 was 67.87 ± 3.31% (n = 3), the Am was 104 GBq/µmol and the RCP was 98.43 ± 0.38%, n = 3 (radio-TLC).

[68Ga]Ga-DOTA-HEVNP-LXY30 remained intact (≥ 90%) after 1 h of incubation under all the tested conditions (Fig.1A) . The in vitro cellular uptake of the targeted HEVNPs in HCT 116 colorectal carcinoma cells was confirmed by the cell study which revealed 13.27 ± 0.68% (n = 3) internalized NPs in 2 h and 21.04 ± 0.73% (n = 3) in 3 h (Fig.1B) . The internalization of the targeted HEVNPs was double the one of the non-targeted HEVNPs in all time points. Evaluation of the ex vivo biodistribution of [68Ga]Ga-DOTA-HEVNP-LXY30 (n = 4) revealed that only the 0.83 ± 0.39%ID/g of [68Ga]Ga-DOTA-HEVNP-LXY30 accumulated in the subcutaneous HCT 116 xenografts in vivo. At the last time point (1 h) , some [68Ga]Ga-DOTA-HEVNP-LXY30 were still circulating in the blood (1.83 ± 0.55%ID/g), and they were distributed mainly in the liver (39.82 ± 12.97%ID/g). Conclusions: LXY30-HEVNPs were successfully radiolabeled with 68Ga for imaging of the drug delivery carrier distribution. The labeled NPs have high in vitro stability and sufficient in vitro cell internalization capability in HCT 116 cells. Nevertheless, the tumor uptake of [68Ga]Ga-DOTA-HEVNP-LXY30 in vivo was low most likely due to the rapid sequestration of the HEVNPs to the liver. Acknowledgements: Academy of Finland (decision numbers: 298481, 320102 and 318422). than that induced by TEMPO. In addition to the limited stability of TEMPO in vivo (t1/2 = 10-20 min) through nitroxy reduction [7] , which is not an issue for N-OH-based moieties, the requirement for direct conjugation of N-OH Pip to the CEST beacon may provide higher fidelity reporting of a broad range of biomolecular activities in living subjects. Conclusion: The uncaging of radical character of N-OH Pip in the presence of a magnetic field has been applied to develop activity-based contrast agents for diaCEST MRI. The ability of N-OH Pip to broadly suppress signals across a range of diaCEST classes may enable multiplexed molecular MRI of biomolecule interactions, signaling or metabolic pathways.References: [1] Kolanowski JL, et al. Chem Soc Rev 2018, 47:195-208 . [2] James ML, Gambhir SS. Physiol Rev 2012, 92:897-965 . [3] Lux J, Sherry AD. Curr Opin Chem Biol 2018, 45:121-130 . [4] Choi JS et al., Nat Materials 2017, 16:537-542 . [5] Ratnakar JS, et al. JACS 2013, 135:14904-14907 . [6] Shindo M et al. Yakugaku Zasshi 1971, 91:590-593 ; Davidson D et al. J Org Chem 1936, 1:305-314. [7] Paletta J et al. Org Lett 2012, 14:5322-5325 . Abstract Body : Introduction Large pools of labile Zn2+ can be found in specific tissues and mapping Zn2+ distribution is crucial to understand its cellular function and its role in disease development1-3. Here, we show the design of MRI-sensor for Zn2+ that possess reversible ionbinding properties with a rapid dissociation rate allowing the detection of a wide range of biologically relevant concentrations of the cation. By combining CEST and 19F-MRI, in an approach termed 19F-iCEST,4-7 the designed 19F-probe was found to be applicable for a wide range of biological-relevant Zn2+ concentrations and was used for in-vivo mapping of Zn2+ pools in the brain of live mice with high specificity. Method 19F-sensors NIS-1, NIS-2 were synthesized using reductive amination and NIS-3 by using the N-alkylation procedure. 19F-NMR and 19F-iCEST (B1 = 2.3 µT/2 s) experiments were performed on 9.4 T NMR at 37 oC and the obtained data were fitted using the Bloch-McConnell equations8-10 to quantify the the ion-chelate dissociation rate, kex (Fig. 1a) . Phantom was prepared with 7 mM NIS-3 and 100 µM of cation's as noted in (Fig. 1b) and 19F-iCEST was performed on 9.4 T MRI with B1 = 2.3 µT/2 s. In-vivo 19F-iCEST was performed on 15.2 T scanner by continuous delivery of NIS-3 to the location of interest (hippocampal region CA3 or the thalamus (TH) via implanted cannula. 19F-CEST-RAREst sequence was used to acquire the in-vivo data with: TR/TE=3000/10.93 ms, RARE factor=16, FOV=22x22 mm2 (32x32 matrix). B1 = 2μT/1.5s was applied "on-resonance", Δω=+3.2 ppm (NIS-3:Zn2+ bound peak) or "off-resonance", Δω=-3.2 ppm and the obtained data were used to generate the Zn2+map, similar experiments were performed on Δω=18 ppm as another control (Fig, S1g) . Results The fluorinated sensors NIS-1, NIS-2 and NIS-3 were synthesized, purified and characterized. Then, their Zn2+ binding properties, 19F-NMR characteristic, 19F-iCEST performance and binding conformation were studied (Fig. 1a , S1a). Although all examined chelates showed well-resolved downfield peak of NIS-bound Zn2+ in the 19F-NMR spectrum (Δω>3 ppm), NIS-3 showed the highest 19F-iCEST effect. This large iCEST effect is attributed to the fast exchange rate between Zn2+-bound and free NIS-3 (kex = 845 s-1) and thus allow the detection of very low concentrations of Zn2+ (500 nM, Fig. S1f ) . It's specificity as compared to other cations was studied and clearly shows the ability to use NIS-3 to map Zn2+ with no background signal from other competitive cations (Fig. 1b , S1d-e). After validation that NIS-3 does not possess any toxicity effect on cells (Fig, S1b) , in-vivo 19F-iCEST MRI experiments were performed to obtain Zn2+ map in the CA3 region of the hippocampus of live mice. To this end, NIS-3 was delivered to either the CA3 (Zn2+-rich ROI) or to the thalamus (TH, Zn2+-poor ROI) via cannula following by in-vivo 19F-iCEST experiments. Clearly, a significant (29±5%, N=7/group, ***p-value was observed only in Zn2+-rich ROI (CA3, Fig. 1c d) and only when the saturation pulse was applied at (Δω=+3. 2 ppm) showing the specificity of the designed approach to map labile Zn2+ in-vivo. Conclusion We showed here an approach for in-vivo mapping of labile Zn2+ with the use of fluorinated-chelate (NIS-3) that reversibly binds the cation at very low affinity to obtain a relatively fast dissociation rate (kex = 845 s-1) of bound-Zn2+. These properties are not just ideal for the cation mapping but also maintain its basal levels in the tissue. This proposed 19F-iCEST chelate can and should be further applied to map labile Zn2+ in other tissues (Pancreas and prostate) and the iCEST principle outlined in this work should be further extended to rationalize the design of new 19F-iCEST probes to map other metal ions with biological importance. Abstract Body : Preclinical small animal imaging involving agents rated biosafety level-2 (BSL-2) or greater is important for studying infectious diseases and particularly timely due to the SARS-CoV-2 global pandemic. We are investigating the pharmacokinetics of SARS-CoV-2 and subsequent gene transduction in an hACE2 mouse model1. This project requires imaging of mice after administration of spike protein-coated pseudotyped lentivirus in a preclinical PET/CT scanner. Several BSL-3 compliant imaging chambers are commercially available; however, these enclosure systems are expensive, developed for the specific PET/CT scanner, and can accommodate at most two mice simultaneously. Here, we constructed developed a generalizable enclosure system adaptable to many preclinical PET/CT, PET and CT systems for in vivo multianimal imaging of biohazardous agents. Materials and system assembly. Materials were primarily purchased from McMASTER-CARR and include polycarbonate cylinder (4" OD, 8585K55, Figure 1A ), clear polycarbonate disks (8508K78), snap-in pipe and tubing grommets (63595K54), barrel-style DC connector receptacles (8320N128) and push-in plugs (8320N111), plastic barbed tube fittings (5463K709), plastic barbed tubing fittings for thru-wall finishing (5463K81), Wye connectors (2974K391), tubing, plastic quick-turn tube coupling sockets (51525K212) and plugs (51525K122). The enclosure system was designed to easily snap a polycarbonate disk ( Figure 1B ) to the clear polycarbonate cylinder ( Figure 1A ) and connected to a HEPA filter. For the clear polycarbonate cylinder, one side of an impact-resistant polycarbonate clear cylinder was closed with a polycarbonate disc. For the snap-in disk part ( Figure 1B) , a second clear polycarbonate disk was glued to the snap-in pipe and tubing grommet, and airtight fittings with airflow channels. DC connectors ( Figure 1C ) were inserted in the middle of the polycarbonate disk to connect to heating strip plugs on the 3D-printed fourmice imaging bed (Sofie Biosciences). Wires and tubings were connected and sealed ( Figure  1D ). The total cost for one enclosure system was approximately $300. Demonstration of enclosure system for PET/CT imaging with BSL2 procedures. All experiments were performed under protocols approved by the Administrative Panel on Biosafety (APB) and Administrative Panel on Laboratory Animal Care (APLAC) at Stanford University. All work was conducted under a biosafety cabinet except when in transit to and imaging on the PET/CT scanner. Four hACE2-expressing mice were transferred from their cage to an induction chamber. Anesthesia, vacuum exhaust and heating lines were connected between the four-mice bed and the polycarbonate disk. The anesthesia gas entry line and outlet of the vacuum line were fitted with HEPA filters. Mice were anesthetized with 1.5% isoflurane in oxygen and placed on the bed, and the bed was slid into the polycarbonate cylinder. The polycarbonate disk was snapped into the polycarbonate cylinder to seal the system. The surface of the enclosure system was sanitized with 70% ethanol gauze or spray prior to transfer out of the biosafety cabinet and onto the PET/CT scanner sled ( Figure 1D ). The unit with mice remained on the PET/CT for 90 min at room temperature and then returned to the biosafety cabinet to return mice to their cages. Mice respiration and temperatures were monitored visually and using a thermal gun, respectively. Results and conclusion. All hACE2 mice remained stable throughout the handling and 90min scanner procedures. Respiration was within normal ranges while under anesthesia within the enclosure system and their body temperatures remained stable. The sealed enclosure and HEPA filtration system ensured no potential release of infectious particles. The enclosure system was economical compared to the commercially available units and could be easily adapted to most PET/CT systems. Abstract Body : Choline Kinase α (ChoKα) is an enzyme overexpressed in 60% of human lung tumors. It is associated with an aggressive phenotype, high histological tumor grade, and poor clinical outcome1-3. ChoKα catalyzes the phosphorylation of choline to generate phosphocholine, a precursor of the cell membrane phospholipid, phosphatidylcholine (PtdCho)4. A novel near-infrared (NIR) fluorescent ChoKα inhibitor, JAS239, has been developed in our lab (excitation: 745 nm, emission: 775 nm)5. JAS239 binds to the active site of ChoKα, competitively inhibiting phosphocholine production, and thus can be used as a targeted NIR imaging agent for tumor resection guidance in ChoKα-overexpressing tumors. Previously, we have utilized JAS239 to image ChoKα-overexpressing breast5,6and lung tumor xeno-and allografts in mice, as well as lung metastases. Here, we translate JAS239 into the veterinary clinic for intraoperative and back-bench imaging in client-owned dogs with spontaneous lung adenocarcinomas. An initial toxicology study was performed using five experimental dogs that received a single dose of JAS239 at five times the imaging dose of 1 mg/kg (5 mg/kg total). The dogs were observed for overt signs of toxicity, including lethargy, vomiting, and anorexia. Blood was collected at baseline, 0, 6, 12, 24, 48 hours, and 5 days post injection for pharmacokinetics, complete blood count-brief (CBC-brief) and blood chemistry. The dogs were adopted out at the end of the study. To date, three canine patients with operable lung adenocarcinomas have been recruited into an owner informed consent clinical trial. Patients received either 0.5 or 1 mg/kg JAS239 24 hours prior to intraoperative imaging. The VisionSense Imaging System was used to image the lung tumors in situand ex vivo. After tumor resection, all sides of the tumor were imaged, sections of tumor and normal tissue were excised for back table imaging, and the fluorescent borders were marked with sutures for pathology. Cut pieces of the tumors were imaged on the LI-COR Pearl and the IVIS Lumina and submitted to the School of Veterinary Medicine Comparative Pathology Core (CPC) for H&E and ChoKαimmunohistochemistry (IHC) (results pending). The experimental dogs showed no overt signs of toxicity and CBC-brief and blood chemistry results were within normal limits. Pharmacokinetics showed rapid probe elimination from the serum, with 80% eliminated by 6 hours, further indicating low toxicity. Intraoperative tumor imaging showed clear margin delineation with higher fluorescence in the tumor than in the normal lung. Initial pathology results found no malignancies at the fluorescent border indicating clear margins. The purpose of these preliminary clinical cases is to establish the optimal dosing and timing for imaging JAS239 fluorescence in canine lung tumors. The clinical trial will recruit up to 30 patient with the goal of determining the efficacy of JAS239 for detecting ChoKα levels in tumor margins, identifying residual tumor lesions, lymph node involvement, and detection of micrometastases from primary lung or other cancers. The ultimate goal of this study is the translation of ChoKα sensors for fluorescence guidance during human lung tumor resections.

Madhusudhana Gargesha, BioInVision, madhu.gargesha@bioinvision.com Category: Immunology: Inflammation & Infection Abstract Body : Aim: Developing a functional cure for HIV is a major goal of the HIV/AIDS research field. Recently, it has been found that natural killer (NK) cells are effective at killing HIV-infected cells. To validate this in a preclinical setting, we assessed the whole-mouse 3D biodistribution of natural killer (NK) cells to gain a better understanding of NK cell homing to potential latent HIV reservoirs. For accurate whole-body assessment of cell trafficking, we used the CryoVizTM cryo-imaging system which acquires 3D microscopic anatomical color and molecular fluorescence volumes with single-cell sensitivity spanning a whole mouse by serially sectioning and imaging the block-face at ultra-high resolution. Methods: Approximately 10M Qtracker-625 (red) labeled human NK cells were injected into NCr nu/nu mice (n=2). In one of the mice, we injected cells intravenously (IV) while in the other, NK cells were administered intraperitoneally (IP). 48 hours after injection, both mice were euthanized, embedded in a medium, and flash frozen in liquid nitrogen for imaging. Mice were imaged with an in-plane (xy) resolution of 10.2µm and a z-section thickness of 40µm. We used an eGFP long pass (>510nm) fluorescence filter which was optimal for detection of Qtracker-625 (red) NK cells. The imaging system created 2D color anatomical and fluorescence block-face images from the whole mouse samples and automatically processed them to obtain 3D volumes. We semi-automatically segmented in 3D the following organs of interest -lung, liver, and spleen -from the color anatomical volumetric data. Using the machine-learning-based algorithm of the system, we automatically detected NK cells from the fluorescence volumetric data. After a careful manual review and editing of automated cell detection results, we obtained refined cell count estimates globally and locally within the segmented organs. Results: With the IV injection ( Figure 1A) , we observed immune cell homing to the lungs, liver, spleen, lymph nodes, bone marrow and the gastrointestinal tract mucosa. Some cells were also found to be retained in the tail (injection site). With the IP injection, we observed cell clumping at the site of injection (in and around abdominal region) and local cell retention with no evidence of migration ( Figure 1B ). Conclusions: This study has uniquely revealed whole-body trafficking of fluorescently labeled NK cells, which will be of significant value in strategies to develop a functional cure for HIV, a major goal of the HIV/AIDS research field. Accurate whole-body assessment of NK cell trafficking and 3D biodistribution analyses were obtained using CryoVizTM cryo-imaging. In future studies, we will employ this technique to assess the efficacy of strategies to engineer NK cells to improve their homing to sites harboring latent HIV, the primary obstacle to achieving a functional cure for HIV. In (A-B), a 3D volume rendering of true-color anatomy is shown fused with detected NK cells rendered in yellow and segmented organs rendered in pseudo-colors (lung-RED, liver-GREEN, spleen-BLUE). With IV injection (A), we observed NK cell homing mainly to the lungs, liver, spleen, femoral and vertebral bone marrow, with some cells retained in the tail (injection site, magenta arrow in A). With IP injection (B), we observed cell clumping at the injection site in and around the abdominal region (blue arrow in B). We found the IV administration method to be more efficacious than the IP route in terms of immune cell trafficking.

Last Name: Gargesha Email: madhu.gargesha@bioinvision.com

Evangelia Sereti, Lund University, evangelia.sereti@med.lu.se Category: Oncology Abstract Body : Introduction and Objectives: Despite substantial progress in prostate cancer therapeutic landscape, metastatic castration-resistant prostate cancer (mCRPC) remains a lethal disease with poor clinical outcome. DNA damage repair (DDR) defects are present in 20-25% of mCRPC patients, with BRCA1/2 mutations representing the most frequent events [1] . Recent clinical trials have shown that mCRPC patients harboring DDR defects responded well to poly (ADP-ribose) polymerase inhibitors (PARPi). The PARPi olaparib was found to significantly improve survival in patients with BRCA1/2 mutated mCRPC. Although robust data exist for PARPi as monotherapy in selected patients, there is a clinical need for combinational therapeutic approaches to improve the efficacy of PARPi and overcome or prevent PARPi resistance [2] . NOV202 is a novel vascular disrupting agent (VDA) with strong anti-proliferative effects in a large panel of cancer cells. This study aimed to evaluate if NOV202 could improve the efficacy of olaparib in prostate cancer cells with BRCA1/2 mutations in vivo. Materials and Methods: Prostate cancer-based xenografts were developed from human DU145 (BRCA1/2 mutated) cell line. Luciferase-transfected DU145 cells were inoculated subcutaneously in male NMRI nude mice. Animals were treated with olaparib (100 mg/kg), NOV202 (30 mg/kg), or their combination for 21 days, followed by either 1 or 2 weeks of no treatment. Tumor growth and response to therapy were monitored by non-invasive bioluminescence optical imaging (BLI) and tumor size was also measured by caliper. During the experiment, a 25-parameter blood analysis, including hemoglobin concentration, red and white blood cells, and neutrophil granulocytes was performed weekly. Results: Olaparib and NOV202, as single drug treatments, induced tumor growth inhibition in BRCA1/2 mutated xenografts. Compared to monotherapy, the combination of olaparib and NOV202 exerted significant synergistic effect. The combinational effect continued for 2 weeks post-treatment period. Changes in the mean values of hematological parameters were not statistically significant in combinational versus monotherapy groups. Conclusions: Single agent treatment with the VDA agent NOV202 and interestingly also with olaparib reduced tumor growth in prostate cancer xenografts with BRCA1/2 mutations in vivo. The anticancer efficacy of olaparib was significantly increased when combined with NOV202 in BRCA1/2 mutated xenografts. A clinical trial of this combination is planned. Abstract Body : Introduction: Small molecular imaging probes for MRI applications are an important set of sensors that were developed for mapping a variety of analytes ranging from neurotransmitters to metal ion's. Nevertheless, their fast clearance from tissues frequently limits their detectability but also prevents the performance of longitudinal studies. Prolonged infusion strategies were found to be a solution for rapidly cleared responsive probes in in-vivo studues1, 2 but other strategies for extending the tissue retention time of designed probes are still of need for desired applications. Here, we show the design of an MRI responsive probe for Zn2+, which undergoes spontaneous self-assembly into a supramolecular hydrogel under physiological conditions. The fluorinated probe shown was developed to detect Zn2+ based on the combination of CEST and 19F-MRI (19F-iCEST)3-5 and found to be detectable in-vivo even 24 hours after its injection allowing longitudinal in-vivo MRI studies. Method: A fluorinated-chelate that is based on a dipicolylamine motif, which specifically binds Zn2+, was synthesized followed by its coupling to a BTA (benzene-1,3,5-tricarboxamide) backbone to obtain the tripod NIS-4. 19F-NMR and19F-iCEST (B1 = 2.3 µT/2 s) were performed on 9.4 T NMR at 25 oC and 37 oC. Hydrogel formation was evaluated by dynamic light scattering (DLS), diffusion NMR studies, and photography at 25 oC and 37 oC. In-vivo experiments were performed on a 9.4 T MRI scanner after intracranial injection of 2 µL of NIS-4 (0.25M in DMSO) to the CA3 region of the hippocampus. In-vivo 19F-NMR experiments were performed to study the sensor clearance rate from the live animal brain. For 19F-MRI experiment, 19F-RAREst sequence was used with the following parameters: TR/TE=4000/10.8 ms, RARE factor=16, FOV=22X22 mm2 (32x32 matrix). Results: The Zn2+ responsive probe NIS-4 ( Fig. 1a) with its 19F-dipicolylamine motifs as Zn2+ binding moieties was synthesized and characterized based on the hypothesis that tripods of this type tend to self-assemble to large architectures at elevated temperatures6. The effect of the addition of Zn2+ to an aqueous solution of NIS-4 resulted in an additional peak of the Zn2+-NIS-4 complex in 19F-NMR spectrum (Δω=+5 ppm, Fig. 1b) . At 37 oC an additional peak of the formed self-assembly of NIS-4 was obtained upfield. A characteristic 19F-iCEST spectrum showed a high CEST effect (Fig. 1e ) from NIS-4 in presence of labile Zn2+ at 37 oC revealing its potentiality to map the ion with MRI. The transformation of NIS-4 to a hydrogel at 37 oC was observed by a naked-eye (Fig. 1c , inset) and confirmed by both diffusion NMR (Fig. 1c) and DLS measurements (Fig. 1d) . Intracranially injection of NIS-4 to a mouse brain was followed by in-vivo 19F-MRS and 19F-MRI. The long retention time (>20 hours) of the injected probe was clearly evident by longitudinal in-vivo 19F-MRS (Fig. 1f) . In-vivo 19F-MRI show the spatial distribution of NIS-4 21 hours after its injection (Fig. 1g) confirming its prolonged retention time in the delivered region. Conclusion: We showed the design of an MRI-responsive probe for Zn2+ sensing that is self-assembled to a hydrogel upon its injection to a live animal. We have demonstrated that the formed supramolecular hydrogel is capable to detect Zn2+ when implementing the CEST principles in 19F-MRI (an approach termed 19F-iCEST). Moreover, we showed that upon its delivery, the 19F-MRI signal is preserved for more than 20 h after its injection. While the proposed approach was demonstrated here for a Zn2+ responsive probe that was designed for 19F-iCEST it can be generalized for the design of additional MRI responsive probes for applications that require long tissue-retention times for longitudinal in-vivo studies.

to 117% at a resolution of 0.78 mm (Fig. 1e) . CONCLUSION: Multi-shot, multi-echo EPI with IDEAL enables high resolution imaging of low SNR 13C chemicals; however, there is loss of quantitation accuracy with increasing spatial resolution. Clinical applications of this technique will help to better visualize small lesions that may occur in cancer metastasis. Acknowledgements: Starr Cancer Consortium (SCC-I12-0031), Thompson Family Foundation and NIH R01 CA252037. (2):978-986.

Image/Figure Caption: a)Non-localized spectroscopy of phantom used in this study for simulating 4 ppm shift resulting from reduction of [1-13C] DHA to [1-13C] Vitamin C in-vivo. Chemical diagram reproduced with permission from Keshari et al. PNAS 2011. b)Number of signal averages (or equivalently) condition number simulation for determining optimal ΔTE for IDEAL separation of DHA and Vitamin C. Shorter ΔTE with high NSA result in better performance of IDEAL. c) T1-weighted of phantom setup showing syringes filled with DHA and Vitamin C separated by a 15 mL falcon tube for loading the coil. An EPSI image was acquired for verifying the location and quantity of the chemicals by spectroscopy. The magnitude of the EPI image acquired at 1.563 mm resolution and the IDEAL decomposition into Vitamin C and DHA maps are shown. d) The IDEAL decomposition of EPI images acquired at 3 spatial resolutions with and without denoising are shown. Downsampling is performed using bilinear interpolation, after which the results of IDEAL are resampled back to the acquired spatial resolution. f) The quantitative accuracy of IDEAL is assessed by measurement of the mean intensity of a region of interest (ROI) drawn on each chemical species map . The spectroscopic quantitation is determined by peak integration . Abstract Body : A major outstanding challenge in the fields of biological research, synthetic biology, and cell-based medicine is a lack of methods to noninvasively image the functions of natural and engineered cells inside opaque organisms. A promising imaging modality for this purpose is ultrasound (US), with its penetration depth of several cm and spatial resolution of tens of µm. Recently, the first genetically encoded US contrast agents-gas vesicles (GVs)-were shown to link cellular function to an US readout via heterologous expression in both commensal bacteria1 and mammalian cells2. GVs are air-filled protein nanostructures derived from buoyant photosynthetic microbes, in which they serve as cellular flotation devices. GVs are encoded by operons of 8-14 genes, and most of their molecular makeup comprises the structural protein GvpA. The air inside GVs allows them to scatter US. GVs can be imaged nondestructively with high specificity by taking advantage of their nonlinear US scattering properties.3 Just as the discovery of the first fluorescent proteins was followed by the engineering and evolution of their properties, we are working to engineer the properties of GVs as acoustic reporters. In this work, we sought to 1) increase the amount of nonlinear signal produced by bacterially-expressed GV gene clusters under mammalian physiological conditions, 2) identify GV clusters capable of producing nonlinear GVs when expressed in mammalian cells with small genetic footprints to facilitate gene delivery, and 3) engineer GV types compatible with nondestructive multiplexed imaging. We accomplished these goals using two complementary approaches: homolog screening and directed evolution. Critical for both approaches was the development of a highthroughput assay for various US phenotypes of interest in bacterially-expressed GVs. To enable this screening, we developed the first "acoustic plate reader," which enabled the automated screening of >1000 bacterially-expressed GV gene cluster variants in a matter of hours. We used this system to screen GV gene cluster homologs from 11 unique species, and thousands of mutants of the primary GV structural protein GvpA/GvpB. In our homolog screen, we tested candidate gas vesicle gene clusters from several diverse species of bacteria and archaea. We identified two GV clusters with improved properties over existing ones: one with greatly enhanced nonlinear contrast and stability in bacteria,4 and another that enables nonlinear detection of mammalian cells and has a more compact genetic footprint than GV clusters previously used for mammalian expression. The former GV gene cluster allowed us to noninvasively image the spatial distribution of tumor colonization by engineered tumor-homing bacteria in situ in living mice, and the latter enabled noninvasive nonlinear imaging of tumor cells themselves. In our directed evolution screen, we generated scanning site saturation libraries for two homologs of GvpA and screened them in E. coli using our custom-built robotic US plate scanner. Custom imaging pulse sequences were used to assess the acoustic phenotypes of each bacterial patch, including total backscattering, nonlinear scattering, and collapse pressure. Using this technique, we identified mutants of GvpA with >150x higher acoustic signal than their parents, enabling both their specific detection in bacteria and nondestructive multiplexed imaging of two GV types. Here, we report the discovery of novel GV gene clusters that enable nonlinear imaging of bacterial and mammalian cells in situ, as well as clusters that enable multiplexed GV imaging. These results suggest that high-throughput protein engineering could play as large a role in engineering acoustic biomolecules as it has in the development of their fluorescent counterparts.

DeepNI Viewer GUI of deep learning human brain segmentation result. It has been demonstrated that paired-agent imaging (PAI) is both highly sensitive and specific in quantifying cell-surface receptor concentrations3. Here, we extend the same principles to cellpenetrating agents for quantification of intracellular drug binding -intracellular PAI (iPAI) -in a chick chorioallantoic membrane (CAM) xenograft assay.Fertilized chicken eggs were incubated over a 14-day period, during which time eggshells were windowed and pieces of ~1-mm3 A431 human HNSCC xenografts (grown subcutaneously in athymic nude mice) were implanted on the surface of the CAMs. For imaging, a 100 µL fluorescent cocktail (intracellular targeted agent: 6.8 µM Erl-SiTMR; intracellular control agent: 6.8 µM Erl-TMR; extracellular targeted agent: 3.4 µM ABY-029 -IRDye 800CW conjugated to anti-EGFR Affibody molecule; extracellular control agent: 0.3 µM IRDye 680LT conjugated to Imaging Agent Control Affibody molecule) was intravenously injected. Multispectral images (550-700 nm and 650-850 nm in 5 nm increments) were collected using the Maestro fluorescence imaging system in 10minute intervals over two hours. Post-acquisition, motion correction and a spectral unmixing routine were employed to isolate the relative contributions of each of the PAI and iPAI agents, and a ratiometric calculation of binding potential (BP) -proportional to EGFR concentrationwas performed.Results demonstrated that all four imaging agents could be spectrally unmixed using a linear least squares fitting approach. Figure 1 (a) shows a representative fluorescence spectrum from a tumor region collected 1 h post-injection along with the spectral fit and individual unmixed fluorophore signals. Performing this on a pixel-by-pixel basis produced images of each agent's tissue distribution, and permitted calculation of BP maps, i.e., intracellular and extracellular protein receptor availability [Fig. 1(c) ]. Interestingly, iPAI and PAI findings had opposite signal intensities in the tumor regions that resulted in an unexpected negative intracellular BP. This may be attributed to necrosis [ Fig. 1 (b) ] in the bulk of the tumor (where most of the iPAI and PAI BP inconsistencies were), whose effects were amplified since the intracellular agents emit in the dominant blood-absorbing region. Therefore, it is likely that autofluorescence was the signal shown in the TMR channel of the tumors because of the lack of blood flow, while surrounding signal was being suppressed by blood. Extracellular BP results on the other hand were more consistent with what was expected where BP was highest in the tumor regions. Further investigation into egg hemodynamics and accounting for blood absorption are currently being investigated (supplementary), but these initial findings support the potential for the CAM model to provide measures of therapeutic agent delivery and ultimately response.References: 1. C. R. Leemans et al., Nat Rev Cancer 11, 9-22, 2011 2. V. Vigneswara and A. Kong, Ann Oncol, 29(4) , 794-796, 2018 3. K. M. Tichauer et al., Mol Imaging Biol, 14 (5) 

John Kelly, University of Western Ontario, jkell5@uwo.ca Category: Bioengineering, Synthetic Biology and Basic Biology Abstract Body : Introduction: Chimeric antigen receptor T (CAR-T) cells have shown remarkable efficacy in treating blood-based cancers1. However, current challenges of these transformative therapies include low effectiveness in many patients, cell accumulation in normal tissues leading to serious side effects, and insufficient migration into and treatment of solid tumours. Translationally relevant imaging technologies that allow one to monitor the fate of CAR-T cells could enable prediction of patient response and side effects, as well as better evaluation of new CAR-T designs. We recently developed our second-generation CRISPR system for editing cancer cells with multimodality imaging reporter genes at the AAVS1 safe harbor2. However, this system was too inefficient for editing primary cells aimed for clinical use. Our objective in this study was to develop a highly efficient CRISPR system for editing human T cells with multiple transgenes at loci known to improve both safety and efficacy, whilst also making them "visible" with clinically relevant modalities. First, as described previously3, to improve CAR-T cell potency and delay exhaustion, we edited T cells at the TRAC locus to knock-in a CAR and simultaneously knock-out the endogenous T cell receptor (TCR). We then incorporated human MRI, PET or preclinical bioluminescent reporter genes into our CRISPR system for molecular-genetic imaging of edited CAR-T cells. Methods: T cells were nucleofected with ribonucleoprotein complexes (RNPs) that cut genomic DNA at the TRAC or AAVS1 loci. Adeno-associated virus (AAV) particles containing CD19CAR and/or the organic anion transporting polypeptide 1B3 (OATP1B3) MRI reporter gene, sodium iodide symporter (NIS) PET reporter gene or Akaluc bioluminescent reporter gene were used as DNA donors. A truncated human LNGFR gene was also included for magnetic bead sorting of edited cells. Editing efficiency was determined by flow cytometry and integration by PCR analysis. CD19+ human B cell lymphoma (NALM6) cells expressing firefly luciferase (FLuc) were engineered for CAR-T targeted kill assays. Uptake of the PET tracer [18F] tetrafluoroborate was measured with a gamma counter. Uptake of the MRI contrast agent Gd-EOB-DTPA was determined using a 3T clinical MRI scanner. Results: Initial experiments showed >85% GFP+/TCR-editing efficiency when targeting CD19CAR-GFP to the TRAC locus ( Figure 1A , B). T cells edited at the AAVS1 locus showed pre-sort efficiencies of ~21% for OATP1B3-LNGFR and ~50% for NIS-LNGFR ( Figure 2A, B) , which increased to 80-95% purity post-magnetic LNGFR sorting. Junction PCR analysis confirmed that the CAR and reporter genes were inserted at the correct loci. NIS-and OATP1B3-edited T cells showed significant uptake of their respective tracer ( [18F] tetrafluoroborate) and contrast agent (Gd-EOB-DTPA) (Figure, 2C, D) . Multiplexed editing with both CD19CAR and reporter genes together produced TCR-/LNGFR+ cell populations, which effectively killed NALM6 cells ( Figure 3 ). These data indicate that both the CAR and reporter gene components were functional in single and double-edited T cells. To improve efficiency, all-in-one AAV vectors containing CD19CAR-OATP1B3, -NIS or -Akaluc were constructed and targeted to the TRAC locus ( Figure 4A ). Correct integration and bioluminescence signal for CD19CAR-Akaluc edited cells was confirmed ( Figure 4B ). In addition, only TRAC RNP and all-in-one AAV edited T cells killed NALM6 cells, whereas those without the CAR did not ( Figure 4C ). These data indicate that both the CAR and reporter genes are functional using an all-in-one approach. Conclusion: Our work demonstrates the first CRISPR-Cas system for highly efficient editing of cancer-killing CAR-T cells with clinically relevant human reporter genes. Ongoing efforts are extending these studies to MRI and PET of edited CAR-T cells in preclinical mouse cancer models. These advanced CRISPR tools should have broad utility for co-editing primary cells with therapeutic genes and reporter genes to make trackable therapeutic cells with improved efficacy and safety profiles. 

Exploring sensitivity and resolution for cell tracking with magnetic particle imaging: the effects of cell proliferation and intracellular nanoparticle relaxation Maryam Berih, Western University, mberih@uwo.ca

Abstract Body : Introduction: Magnetic resonance imaging (MRI) cell tracking can be used to monitor the fate of cells labeled with superparamagnetic iron oxide (SPIO) nanoparticles. One limitation of this method is the inability to track proliferative cells long-term due to dilution of the SPIO amongst cell progeny1-3. This is particularly concerning for tracking rapidly dividing cells (cancer). Magnetic particle imaging (MPI) is an emerging modality which directly detects SPIOs. In vivo we expect to be able to track proliferation of SPIO labeled cells with MPI as long as the cells and progeny remain in the same field of view. To begin to study this, our first aim is to quantify the dilution of SPIOs in breast cancer cells in vitro using MPI. MPI signal is generated from a combination of Néel (internal rotation of magnetization) and Brownian (physical rotation of nanoparticle) relaxation. Brownian relaxation of SPIO is influenced by the nanoparticle's surroundings and we hypothesize this may have implications for detecting partially immobilized intracellular SPIOs. A second aim is to determine how MPI signal and resolution change when SPIOs are intracellular (live cells) compared to free SPIOs (lysed cells). Methods (Aim 1): 4T1 murine breast cancer cells were labeled in vitro by co-incubation with 55 g Fe/mL ferucarbotran, with heparin and protamine sulfate4. Triplicate samples of 1x106 ferucarbotran-labeled 4T1 cells were collected after 24 hours (day 0), and the remaining cells were returned to culture. After 22 hours (day 1), 1x106 ferucarbotran-labeled cells were collected again, and this was repeated at 41 hours (day 2). MPI of cells was acquired on a Momentum MPI in 2D using dual-channel 5.7T/m gradients and excitation of 20mT (X-channel) and 26mT (Z-channel). Each day, a cytospin of 4T1 cells was performed and cells were stained for iron with Perl's Prussian blue (PPB). Methods (Aim 2): Mesenchymal stem cells (MSC) were labeled using the same protocol. Three samples of 2.5x105 ferucarbotran-labeled MSC were collected. MP relaxometry using rf amplitude 20mT and bandwidth 160mT was conducted for 3 samples of ferucarbotran and ferucarbotran-labeled MSC. MSC were lysed with 70L RIPA buffer then sonicated. Relaxometry was repeated for lysed cells. Sensitivity was evaluated by normalizing curves by sample iron mass and resolution was evaluated by measuring the fullwidth half maximum (FWHM). Results (Aim 1): MPI signal was reduced and PPB stain revealed less iron content in proliferating cells over two days (Fig A) . The iron content in 4T1 cells (measured by MPI) decreased by 68% from day 0 to day 1 (p < 0 .01) and by 93% on day 2 (p < 0 .0001) (Fig B) . (Aim 2): MP relaxometry revealed a 44% reduction in sensitivity (p < 0 .01) and increased FWHM by a factor of 2 (p < 0 .001) as ferucarbotran was incorporated into cells. Following sonication of cells, sensitivity was significantly recovered (p < 0 .01, Fig C, D) . Conclusion: A reduction in MPI signal was measured from SPIO-labeled 4T1 cells following proliferation in vitro. Our measurements of intracellular iron are in close agreement with a theoretical reduction of 66% (day 1) and 87% (day 2). Future work will examine this in vivo. As ferucarbotran was incorporated into MSC, a reduction in MPI sensitivity and resolution were measured; this is likely related to increased Brownian relaxation times of intracellular SPIO5-7. With cell lysis, the MPI signal is recovered. This indicates a loss in MPI sensitivity is induced by cellular uptake of ferucarbotran. For a standard 5.7T/m gradient, the reduction in resolution is equivalent to 1.6mm (free iron) compared to 3.2mm (intracellular iron). MPI cell tracking is in its infancy and these studies contribute important knowledge towards monitoring proliferative cells in vivo and optimizing resolution and sensitivity for cell detection.References: 1. Makela et al. Sci. Rep (7) . . 2. Arbab AS et al. Radiology. 2003 . 3. Foster PJ et al. Neoplasia (10). 2008 . 4. Economopoulos V et al. Transl. Oncol. (6). 2013 . 2015 6. Saritas EU et al. J Magn Reson (229). 2013 7. Suzuka H et al. J Nanosci Nanotechnol (19) . 2019. 8. Arami et al. Med Phys (40) . 2013. Figure: https://s3.amazonaws.com/amz.xcdsystem.com/1E8A4501-BB41-A02A-1D3B9882B798BFAA_abstract_File10351/GA299_ImageFigure_0513035741.jpeg Image/Figure Caption: Figure 1 : (A) MPI signal for 1 x 106 ferucarbotran-labeled 4T1 cells is reduced after collection from culture over 3 consecutive days. PPB stain exhibits sufficient labeling of cells with ferucarbotran labeling (day 0), and reduced iron content as cells proliferate in culture (day 1 and 2). (B) Iron mass per cell, measured by MPI, was significantly reduced over time. (C) Relaxometry curves for ferucarbotran and 2.5 x 105 live and lysed ferucarbotranlabeled MSC. (D) The peak MPI signal (sensitivity) was significantly higher for free ferucarbotran and ferucarbotran released from lysed cells, compared to ferucarbotran-labeled cells. The FWHM was increased in live and lysed ferucarbotran-labeled cells, compared to free ferucarbotran. * p < 0.05, ** p < 0.01, *** p < 0.001. [18F] FRho6G-DEG is an 18F-labeled analogue of rhodamine 6G that is taken up by mitochondria-rich cells such as cardiomyocytes and as such is being developed as a myocardial perfusion PET imaging agent. [1] In swine, [18F] FRho6G-DEG shows a high first-pass extraction fraction (84%) that is independent of flow. However, human PET studies using 10% EtOH/saline as the vehicle resulted in tracer accumulation in the IV apparatus and vein, along with suboptimal uptake into heart vs. liver. These phenomena might be related to tracer solubility and/or the formation of aggregates. If so, an alternative drug formulation might positively modulate [18F] FRho6G-DEG biodistribution. With this in mind, we chose to assess propylene glycol (PG) as a clinically approved alternative to EtOH in the final formulation. We also investigated the use of (2-hydroxypropyl)-β-cyclodextrin (HPβCD) as an excipient, because rhodamine 6G can form an inclusion complex with HPβCD. [2] Methods: [18F] FRho6G-DEG was obtained from its tosylated precursor via a standard nucleophilic aliphatic 18F-fluorination reaction. [3] The tracer was HPLC-purified, extracted via tC18 solidphase extraction cartridge, and eluted with EtOH. Fractions of [18F] FRho6G-DEG solution were concentrated under Ar, then resolubilized as reported in Table 1 . The extent of aggregation was estimated by filtration of [18F] FRho6G-DEG in various matrices through 0.2 µm microcentrifuge filters (14,000 rcf, 5 min), and protein binding was assessed in the same matrices by ultrafiltration (30,000 MW cutoff, Table 1 ). [4] Non-specific membrane binding (NSB) was estimated using formulations incubated in PBS and these factors were used to correct % free fractions in serum. Sprague-Dawley rats were injected with [18F] FRho6G-DEG in Matrices B-E and µPET images were acquired over 60 min. Results: The use of 30% HPβCD as excipient and/or 40% PG as solvent caused a marked increase in % activity passing through 0.2 µm filters, indicating enhanced solubility. A small increase in % free fraction of [18F] FRho6G-DEG in serum was observed using solvent matrices containing 30% HPβCD (Matrices C & E) vs. those without HPβCD (Matrices B & D) . Dynamic PET images (5-min intervals) of [18F] FRho6G-DEG in rats suggest that when using 10% EtOH/saline as the solvent, the presence of 30% HPβCD has little effect on the ratio of activity concentration (Ac) in the left ventricle (LV) vs. liver at later time points (3.0 vs. 3.1 at 55-60 min). A heart:liver ratio of 2.7 was observed when using 40% PG/saline alone. However, when 30% HPβCD was used in conjunction with 40% PG/saline, the observed ratio increased to 5.1. Matrix Formulation % Activity through filtera % free fraction in serum±SDb Ac(LV)/Ac(liver)c 0-5→55-60 min A PBS (pH 7.4) [18F] FRho6G-DEG in 10% EtOH in saline was improved by the addition of 30% HPβCD, but greater improvement was observed with the use of 40% PG in saline as solvent, either with or without HPβCD. All three of these alternative formulations also decreased protein binding. As observed in µPET scans, the use of 30% HPβCD in 40% PG/saline results in higher heart:liver ratios from 0-60 min compared with 10% EtOH/saline, 10% EtOH/saline with 30% HPβCD, or 40% PG/saline only. The relationship between solvent matrices, complexing excipients, and the biodistribution of radiopharmaceutical candidates bearing the rhodamine scaffold remains incompletely understood, but additional investigations of this system are ongoing.

Conclusion We demonstrated HP DHA as an imaging agent for probing oxidative stress in PDX models of cancer. HP DHA was then used to characterize the biochemical response of these PDX models to ascorbate therapy. Changes in DHA/ascorbate metabolism were observed in these tumor models after prolonged treatment, demonstrating a new method for assessing ascorbate therapy in pancreatic cancer. Acknowledgements This work was supported by the Starr Cancer Consortium (SCC-I12-0031), Stand Up 2 Cancer, Thompson Family Foundation and NIH R01 CA252037. Tumor growth curves for patient derived xenograft models of KRAS and BRCA driven cancer. 2 weeks after tumor implantation, mice were randomly divided into two groups. One group was treated with freshly prepared vitamin C in 400 mL of PBS (4 g/kg) twice a day via IP injection (KRAS, n = 7; BRCA, n = 7). Control group mice were treated with PBS using the same twice a day dose (KRAS, n = 7; BRCA, n = 7). Tumor growth change for the KRAS PDX tumors is significant after 21 days of ascorbate treatment (36.5% ± 9.3%, p=.0035) and significant for the BRCA PDX tumors after 7 days pf ascorbate treatment (74.0% ± 12.3%, p=.0001 Abstract Body : Cell therapy has tremendous potential to transform the treatment of many diseases such as cancer and autoimmune disorders. Despite recent advances in synthetic genetic control systems, there is a current lack of technological ability to secrete therapeutic agents from engineered cells with precise spatiotemporal control. The ability to induce a localized modulation would prevent the development of potentially lethal side effects associated with systemic release of therapeutic proteins, such as autoimmune disorders. Here, we are developing temperature-sensitive recombinase circuits to remotely control immune cells deep within the body using high-precision non-invasive techniques such as focused ultrasound hyperthermia. A permanent gene expression transition can be achieved by coupling a short pulse of thermal activation with recombinase expression. We have verified the activity of these circuits in a panel of immune cells demonstrating its broad applicability. Remarkably, these circuits display extremely low baseline activity in the absence of the thermal stimulus, and high induction after stimulation. This tool was used to control CD19 bispecific T-cell engager (BiTE) production and release by macrophages in vitro. Engineered macrophages were shown to traffic into CD19+ Raji tumors in an immunocompromised mouse model. Trafficked macrophages harboring the temperature-sensitive recombinase reporter circuit can be activated in vivo using focused ultrasound hyperthermia (FUS). In addition to enhancing cell therapeutics, this tool also has the potential to address fundamental scientific questions, such as cell migration and multicellular development. Introduction Alzheimer's disease (AD) is a progressive neurodegenerative disease that is characterized by memory loss and cognitive dysfunction. Oxidative stress has been highlighted as a key player in the neurodegeneration associated with this disease. In this study, we measured elevated levels of reactive oxygen species (ROS) in a transgenic mouse model of AD using [18F] ROStrace, a radiotracer that measures ROS. A series of PET imaging studies were conducted in wild type and APPSWE/PS1dE9 female and male mice at 5, 10 and 16 months of age, and the results were compared with histological measures of Aβ burden. Methods PET studies using [18F] ROStrace were conducted using a total of 91 mice. Radiotracer (~ 0.25 mCi) was injected through the tail vein and dynamic images were obtained for 60 min and analyzed using Pmod. The cortex, thalamus, cerebellum, hypothalamus, brain stem, central gray, striatum, hippocampus, amygdala, and midbrain from the Mirrione atlas were included for volume of interest (VOI) quantification. To confirm specific uptake regions of [18F] ROStrace, ex vivo autoradiography was performed and compared with the PET data. Antibody staining to measure Aβ plaque density was conducted in the same mice after PET to assess the correlation between [18F] ROStrace and Aβ burden. Results An increase in brain uptake of [18F] ROStrace was observed from 5 months to 16 months, with the APP/PS1 mice showing higher brain uptake than the wild type mice. We observed higher [18F] ROStrace uptake in cortex and hippocampus by SPM analysis. This was confirmed by two-way ANOVA test, we demonstrated a significant difference in cortex (p Conclusion PET studies of [18F] ROStrace were conducted to measure superoxide levels in a murine model of AD. We observed that [18F] ROStrace brain uptake in APP/PS1 mice was higher than in wild type mice, and specific uptake corresponded to regions with a high Aβ burden. We also observed higher uptake of [18F] ROStrace in female AD mice than male AD mice. Based on these results, we conclude that [18F] ROStrace is a useful radiotracer for imaging the increase in levels of oxidative stress thought to occur in AD. Abstract Body : Introduction. Liver fibrosis can occur in most chronic liver diseases, and can lead to cirrhosis, liver failure, primary liver cancer, and/or organ failure.1 Fibrogenesis is accompanied by upregulation of lysyl oxidase enzymes which catalyze oxidation of lysine eamino groups on extracellular matrix proteins to form the aldehyde containing amino acid allysine which then undergoes cross-linking. MRI probes functionalized with a hydrazine moiety can bind to allysine in vivo to detect fibrogenesis, 2-4 but these are all gadolinium-based and there is growing concern about the long term safety of Gd-based probes.5 Here, we describe the design and synthesis of novel manganese-based MRI probes with high signal amplification for imaging liver fibrogenesis. Results and discussions: The design of hydrazine equipped manganese probes for imaging liver fibrogenesis must meet several requirements: 1) chemically stable Mn2+ complex that does not release free Mn2+ or undergo redox chemistry; 2) limited hepatobiliary elimination to minimize liver background signal; 3) contain coordinated water coligand to promote high relaxivity; 4) achieve higher relaxivity when bound to fibrotic tissue to increase signal at site of fibrosis. To meet these requirements, we designed the probes Mn-2Hyd, Mn-2CHyd and Mn-1CHyd based on the stable Mn-1,4-DO2A chelate (Fig 1a) . 6 In model reactions with butyraldehyde, Mn-2CHyd showed the highest on-rate and conversion yield (Fig  1b) . Temperature dependence of the H217O tranverse relaxivity showed that the inner-sphere number of water molecule decreases from 0.90 in Mn-1,4-DO2A to 0.44 in Mn-2CHyd (Fig 2a) , resulting in a lower relaxivity of Mn-2CHyd in PBS (1.6 mM-1s-1). However, Mn-2CHyd exhibits an almost 5-fold turn-on relaxivity when bound to allysine modified BSA protein (7.7 mM-1s-1, Fig 1c) , indicating high affinity towards allysine residues and increased relaxivity upon binding. Mn-2CHyd is more inert to Mn2+ release than Mn-1,4-DO2A (Fig 2b) . Unlike Mn-1CHyd which showed high Mn liver levels in healthy mice at 60 mins post-injection, Mn-2CHyd showed little accumulation in the healthy liver (Fig 2c) . This combination of stability, turn-on relaxivity, and low uptake in healthy liver indicate that Mn-2CHyd is quite promising as a novel probe for liver fibrogenesis detection. Mn-2CHyd was tested in mice treated with CCl4 or olive oil vehicle for 12 weeks to induce liver fibrosis. Ex vivo analyses of liver hydroxyproline and Sirius red staining indicated consistent fibrosis in the CCl4 group (Fig 2e-g) . Gd-DOTA was used as a comparison to demonstrate the specificity of Mn-2CHyd towards fibrosis. At 45 mins post-injection, there is no liver enhancement with Gd-DOTA in CCl4 mice, but strong signal enhancement and high liver-to-muscle contrast to noise ratio (ΔCNR) in mice injected with Mn-2CHyd (Fig 1d-e) . The vehicle group imaged with Mn-2CHyd showed significantly lower liver enhancement (Fig 2d) . These data indicated that Mn-2CHyd shows high sensitivity and is specific for in vivo detection of liver fibrogenesis. Conclusion. A series of hydrazine equipped manganese MRI probes were designed and synthesized. Mn-2CHyd shows the highest affinity and turn-on relaxivity in allysine containing protein solution, and low liver uptake in normal mice. The ability of this probe to detect liver fibrogenesis was demonstrated in vivo using a CCl4 mouse model, where it shows significantly enhanced liver signal in CCl4 injured mice than in control mice. This study opens the door to study manganese-based hydrazine equipped probes for imaging fibrogenesis.

References: 1. Bataller, R., & Brenner, D. A. J. Clin. Invest. 2005, 115, 209-218; 2. Chen, H. H., Waghorn, P. A., Wei, L., et al. JCI insight, 2017, 2, e91506; 3. Waghorn, P. A., Jones, C. M. et al. Angew. Chem. Int. Ed. 2017, 56, 9825-9828; 4 . Akam, E. A., Abston, E., Rotile, N. J., Slattery, H. R., Zhou, I. Y., Lanuti, M., & Caravan, P. Chem. Sci. 2020, 11, 224-231; 5. Grobner, T. Nephrol. Dial. Transplant. 2006 , 21, 1104 −1108 6. Garda, Z., Forgács, A., Do, Q. N., Kálmán, F. K., et al. J. Inorg. Biochem. 2016, 163, 206-213 . Tumor burden and progression were monitored by weekly bioluminescence imaging (BLI) in mice treated with bortezomib. Mice with established disseminated disease were randomly grouped into untreated and treated cohorts, along with an additional cohort of age and strain matched healthy control cohort. The treated group received bortezomib (1mg/kg) intraperitoneally, twice a week. Small PET imaging was performed with 18FDG and 64Cu-LLP2A at baseline (prior to start of treatment) and then sequentially (one day apart) every week for upto six weeks. Whole body PET/CT images were acquired on the small animal INVEON PET/CT scanner and images were coregistered using Inveon Research Workplace (IRW) software. Regions of interest (ROI) were drawn and selected from PET images using CT anatomical guidelines and the activity associated with the ROI was measured with IRW software. When feasible, tissues were extracted for ex vivo flow cytometry. Results: Flow cytometry studies confirmed high expression of CD49d (α4 subunit of VLA-4) in U266-luc myeloma cells (>99%) compared to MM.1S-luc cells with ~52% cells showing CD49d expression. Bortezomib therapy response was monitored by longitudinal small imaging PET and BLI in MM.1S-luc and U266-luc disseminated tumor bearing mice. BLI data in mice from the untreated cohort showed 4-fold increased total body luciferase activity as compared to the mice from the treated cohort at week 3 in both the mouse models. In MM.1S-luc myeloma mouse model, compared to the treated cohort of mice (where treatment was started at week 0), 64Cu-LLP2A localized with a significantly higher SUVmean in bone marrow rich skeletal sites such as spine (0.58 versus 0.31) and femur (0.72 versus 0.39) at week 4. Similarly, in U266-luc tumor bearing untreated mice, the percent increase in the 64Cu-LLP2A uptake in spine was 4 times higher than the treated mice at week 4.64Cu-LLP2A PET/CT imaging at 4 h post injection of the radiotracer showed significant differences in the uptake among the treated and untreated cohort demonstrating better monitoring of therapy response with this tracer as compared to 18FDG (1 h post injection) in both the myeloma tumor models. Conclusion: We observed qualitative and quantitative differences in the uptake of standard-of-care 18F-FDG and VLA4 targeted 64Cu-LLP2A tracers in human models of MM. Based on the in vitro and in vivo data, we posit that the accuracy of MM detection and treatment monitoring can be significantly improved by the complementary use of the metabolic and plasma cell receptor targeted tracers. Abstract Body : Objective: Neuroendocrine Prostate Cancer (NEPC) is an aggressive subtype of prostate cancer (PC) that is resistant to current standard of care treatments including androgen receptor signaling inhibition (ARSI) [1] . The patients have aggressive disease with visceral metastasis and have short survival time (7-10 months) [2] . Genomic and proteomic characterization of biopsy samples of NEPC lesions indicates loss of androgen receptor (AR) signaling [3] References: 1. Watson, P.A., V.K. Arora, and C.L. Sawyers, Emerging mechanisms of resistance to androgen receptor inhibitors in prostate cancer. Nat Rev Cancer, 2015. 15(12): p. 701-11. 2. Apostolidis, L., Nientiedt, C., Winkler, E.C., et al., Clinical characteristics, treatment outcomes and potential novel therapeutic options for patients with neuroendocrine carcinoma of the prostate. Oncotarget, 2018. 10: p. 17-29. 3 . Puca, L., Gavyert, K., Sailer, V., et al., Delta-like protein 3 expression and therapeutic targeting in neuroendocrine prostate cancer. Sci Transl Med, 2019. 11(484) . Abstract Body : Fluorescence-guided surgery has emerged as an important tool in a number of different disciplines such as cancer resection [1] , plastics and reconstructive surgery [2] and gastric surgery [3] . However, it has not appreciably impacted the field of orthopaedic trauma. One of the dominant challenges in orthopaedic trauma surgery is the management and treatment of infection, which can result in repeated procedures, prolonged morbidity and huge economic cost. To properly address infection, both infected and devitalized (i.e., hypoperfused) bone and soft tissue must be cleared in a thorough debridement. However, no current approach exists to precisely identify these tissues. Our team has established intraoperative methods to quantify perfusion [4] [5] [6] and are now developing infection-sensitive contrast. As a corollary to the Dartmouth-Hitchcock initiated multicenter clinical study funded by NIH/NIAMS, our group is advancing fluorescence-guided debridement (FGD), the use of two contrast agents-one sensitive to perfusion, the other sensitive to infection/inflammation-to guide removal of infected and devitalized bone and tissue. Approval was obtained from the IRB to administer indocyanine green (ICG) and/or 5-aminoluvulinic acid (5-ALA) to patients undergoing redebridement and re-fixation under suspicion of osteomyelitis, and acquire a time-series of fluorescent images using the SPY Elite (Stryker Corp.). Hardware from patients suspected of having implant-associated infection was also removed and analyzed for bacteria. Depth-resolved bacterial biofilms on the surface of tibiotalocalcaneal fusion nails were imaged with 3 µm axial resolution using Ganymede optical coherence tomography (OCT) system (Thorlabs, USA). Functional OCT speckle variance technique applied in 25ms increments over four seconds for each depth slice was used to detect individual bacteria through their temporal motility rates. Nails, screws and rods were soaked in a solution of 10 mM of 5-ALA for 2 hours, and then imaged using a wide-field fluorescence microscope under 405 nm excitation. In a wide spectrum of gram-positive and gram-negative bacteria, 5-ALA induces porphyrin synthesis (e.g., protoporphyrin IX, but others as well), which are photosensitizers and emit red light (635 nm) when excited with blue (405 nm). A 630±30 nm bandpass filter was mounted to a sCMOS camera (PCO, Germany) and imaged at a 1 Hz framerate. Figure 1A -H shows the results of Patient 1 who sustained a Gustilo Type IIIA distal tibia fracture that was complicated by infection. The patient was administered ICG just prior to re-debridement, re-fixation and Masquelet bone grafting. The signal void in on the right side of the field-of-view is suggestive of devitalization. The ingress slope (wash-in rate) shows reduced values and the time-to-peak at the peripheral regions of the compromised region (ROI2) show extremely slow and disrupted kinetics. Two weeks following this surgery, persistent nonunion was confirmed and the patient underwent transtibial amputation. Figure 1I -O shows fluorescence and OCT analysis of a tibiotalocalcaneal fusion nail removed from Patient 2, who sustained a pilon fracture leading to MRSA-positive osteomyelitis. The nail was half-submerged in a 10 mM solution of 5-ALA, and imaged. Figure 1I shows the 635 nm PpIX intensity as a function of position, and J-L show the fluorescence intensity maps corresponding to three positions indicated. Fig. 1M and N show the functional and structural respectively, and Fig 1O shows the ST at a depth of 18. Hardware will undergo quantitative microbiology and NGS 16s rRNA microbiome analysis to determine the presence of viable and dead microbes. Ongoing clinical studies will investigate association between in vivo ICG and 5-ALA images, ex vivo analysis of hardware and microbiome diversity to determine performance of fluorescence for detecting devitalized and/or infectious tissue. Abstract Body : Aim: To assess the feasibility and value of performing dynamic perfusion PET imaging for the evaluation of esophageal tumors over the course of radiation therapy. Methods: Fifteen patients with esophageal tumors were imaged prior to radiation therapy, at the midway point of therapy, and post-therapy, an average of 12 weeks after baseline imaging. Dynamic perfusion imaging was acquired over a single bed position at the site of the primary tumor for 15 minutes immediately following injection of 185 MBq 18F-FDG. Imaging was performed on a digital photon counting system (Philips Vereos). Patients spent the remaining uptake period in the injection suite followed by whole body imaging at 75 minutes post-injection as standard imaging. Listmode data from the dynamic acquisition were reconstructed with 1 minute per frame. Regions of interest were placed over the primary tumor, aorta/blood pool, paraspinal muscle, and a portion of healthy liver. Results: Excellent image quality and quantifiably robust data were produced for all acquisitions. Physiologic motion could readily be corrected for, however very little motion was observed. Differences in perfusion and dynamic uptake patterns in the individual tumors were identifiable. From baseline to post-treatment, patients with Grade 1-2 (well and moderately differentiated) tumors had greater decreases in the initial rate of FDG uptake than did those with Grade 3 (poorly differentiated) tumors, on average an 86% decrease in the slope of the dynamic uptake curve for Grades 1-2 versus only a 56% decrease in slope for Grade 3 tumors. Patients who responded well to radiation therapy also had tumors which tended to show a greater initial perfusion peak on the uptake curve than did those which did not respond well to therapy. Conclusion: We have demonstrated that dynamic perfusion FDG PET/CT is robustly achievable. Quantification of tumor features, including initial tumor perfusion and uptake rates, is robust as well as the assessment of changes in perfusion and uptake in response to therapy. This PET-based methodology can be applied to assess changes in the rates of tumor perfusion and radiotracer uptake over the course of therapy, adding additional insight beyond delayed, static measures of SUV. Abstract Body : Introduction: Pulmonary fibrosis results in thickening of the lung interstitium, abolition of alveolar spaces and eventual respiratory failure.1 It is accompanied by upregulation of lysyl oxidase and Lox-like enzymes which catalyze oxidation of lysine e-amino groups on extracellular matrix proteins (chiefly collagens) to form the aldehyde containing amino acid allysine which then undergoes cross-linking reactions with other proteins to stabilize the matrix. Hydrazine derivatized gadolinium probes can bind to allysine in vivo to stage and quantify fibrogenesis.2-4 Here, we designed a new probe, Gd-1,4, with two hydrazine moieties where dual binding to allysine may enhance on-rate and slow the off-rate relative to the previous reported probe Gd-CHyd,4 and assessed whether Gd-1,4 can be used to monitor treatment response in a model of pulmonary fibrosis with a promising natural product EGCG being developed clinically.5 Methods: Cis-1,4-Gd(CHyd)2 (Gd-1,4) is a Gd-DOTA derivative with two hydrazine arms synthesized by amide coupling of two tBu-piperazin-1-ylcarbamate groups to to 1,4,7,10-tetraazacyclododecane-1,4-bis(tBu-acetate)-4,10-bis(glutaric acid 1-tBu ester), followed by acid deprotection and gadolinium complexation. C57Bl/6 adult male mice at 8 weeks of age received a single intratracheal dose of bleomycin, 1 U/kg body weight (50 μL total volume) as previously described.6 A pair-wise imaging study of Gd-CHyd and Gd-1,4 was carried out in mice at day 14 post bleomycin injury on a 4.7 T Bruker Biospec MRI with a dose of 100 μmol/kg of each probe. For the treatment study, mice were imaged first at day 10 post bleomycin injury, then treated daily with oral gavage of either epigallocatechin gallate (EGCG, 100 mg/kg) or PBS (vehicle) for 11 days. At day 21 post bleomycin injury, the animals were imaged again. MRI consisted of T1-weighted UTE (TR/TE/flip angle = 4 ms/0.011ms/15°) images acquired prior to and up to 30 minutes following intravenous administration of Gd-1,4 (100 μmol/kg). Following the second imaging session, animals were euthanized (60 minutes post injection), and lung tissues were removed for further analysis. Results and discussions: We first compared the dual hydrazine probe Gd-1,4 to Gd-CHyd in a head-to head comparison in vivo. We imaged mice 14 days after bleomycin injury first with Gd-CHyd and then 4 h later with Gd-1,4. This pairwise study showed that the signal generated by Gd-1,4 in fibrotic lung is 3-fold higher (P Conclusion. A novel allysine targeting dual binder MR probe Gd-1,4 was developed and it shows higher sensitivity towards pulmonary fibrogenesis detection than the state of the art. Importantly, Gd-1,4 can noninvasively and sensitively monitor treatment response with an antifibrotic drug currently in clinical development for IPF treatment.

Sheng-Chieh Chiu, The University of Alabama at Birmingham, leon1215@uab.edu

Abstract Body : Overview: A novel framework that is designed as extensions for open-source image software is proposed that facilitates deep learning-based image processing and analysis over cloud computing servers, and to demonstrate the feasibility we present our extension for brain parcellation with deep learning in the widely used 3D Slicer [1] . With the rapid growth of deep learning technologies and their application over medical imaging, more and more openly shared models have been made available and possess a strong potential for advanced image processing. However, for inexperienced users, setting up and testing such open-source models remains to be a challenging task. To address these challenges, we propose to develop computational tools that require minimal effort for users to set up while maintaining the same computational performances. Our framework is based on the open-source image viewer, 3D Slicer, by implementing its extension that utilizes the MatlabBridge Extension [2] in 3D Slicer, the MATLAB Job Manager [3] , and deep learning models configured and executed on the Google Cloud. Objective: We aim to create a versatile and user-friendly software environment for users of all experience levels to perform deep learning-based image processing by eliminating the barrier of the demanding hardware and software requirements. Methodology: The proposed project was divided into three components of implementation. 1) The cloud server environment establishment: The server was established in the Google cloud platform with GPU support. Docker images were used within the server for executing the deep learning models to avoid software conflicts and inconsistencies. 2) Deep learning model implementation: The deep learning model, FastSurfer [4] , was used as a proof of concept as the deep learning-based brain segmentation model in the cloud server. We package FastSurfer and its required libraries into a single Docker image for easy deployment and isolated environment setups. 3) Bridging between 3D Slicer and the cloud server: With the aim to transmit data to the server, MATLAB Job Manager was deployed under MatlabBridge Extension. This job manager packages the data and commands into jobs, submits jobs from the local computer, distributes and executes them on the cloud, and returns the completed results back to the client. The user will only be required to install 3D Slicer and our extension without any specific settings for FastSurfer or other libraries. Result: We have successfully implemented the framework as a 3D Slicer extension. Brain parcellation was greatly accelerated by the 3D Slicer FastSurfer extension while the user does not need local GPU resources. By demonstrating the feasibility and performance improvements with deep learning-based image processing methods through our work, such computational frameworks can be generalized to other deep learning-based image processing and analysis models to stimulate novel and advanced imaging research in the future. Abstract Body : Due to its high versatility and availability, albumin has always been an attractive molecule for biomedical and pharmaceutical applications. A wide range of albuminbased systems, such as drug conjugates and nanoparticles, have been reported over the years as diagnostic imaging agents as well as to improve a drugs' pharmacokinetic, safety and efficacy profile [1] [2] [3] . Besides their long-standing use as a lung perfusion imaging agent, significantly less attention has been given to macroaggregated albumin (MAA) microparticles. One of the hurdles holding back the development of MAA is how the microparticles are prepared, i.e., by heating a bulk solution of albumin, which results in an undesirably large size distribution and batch-tobatch variability [4, 5] . The aim of our study was to develop a method to prepare MAA microparticles in a more defined, tunable and reproducible manner. To do this, a microfluidic chip was prepared using thiol-ene chemistry that had an integrated flow focusing droplet generator and curing network [ Figure 1A ]. Aqueous droplets containing albumin were formed within a continuous oil phase, which then travelled through a 1.2 m long channel heated to 85°C to denature and aggregate the protein into stable MAA microparticles. By carefully adjusting the flow rates into the chip, microparticles between 70 to 300 µm with a polydispersity index of Overall, this study shows that microfluidics is a viable alternative to prepare MAA in a more controlled fashion. Although it was not possible to obtain microparticles smaller than 70 µm with our current set up, changes in the geometry of the flow focusing region or refinement of the continuous oil phase may solve this problem. Current studies are also undergoing to evaluate how drugs can be incorporated within the MAA microparticle matrix for passive lung-targeting delivery.

~0.3, [18F]cGBP4F = ~0.6). Logan and Lassen graphical analyses indicated 60.1 ± 6.3% specific displacement of [18F] cGBP4F by gabapentin at 5.0 mg/kg. In rats, the brain penetration was consistent with that in monkeys. Gamma counting and ex vivo autoradiography studies in SNL rats showed significant increase in tracer accumulation in ligated nerves compared with nonligated nerves (tGBP4F: 205 ± 51%, n = 5; cGBP4F: 166 ± 22 %, n = 4). Conclusion: α2δ-1 PET radiotracers trans-4-fluorogabapentin ( [18F] tGBP4F) and cis-4-fluorogabapentin ( [18F] cGBP4F) were synthesized and evaluated in vitro, in vivo, and ex vivo. [18F] tGBP4F showed >20-fold higher in vitro binding affinity towards α2σ-1 and [18F]cGBP4F showed higher CNS permeability. Both tracers showed significant increase of tracer accumulation at the ligated nerves, which indicate their potential application in PET imaging of neuropathic pain.References: [1] PAIN 135, 7-10 (2008) ; [2] Neuroscience 338, 183-206 (2016) . [3] Nat. Rev. Neurosci. 13, 542-555 (2012) ; [4] J. Nucl. Med. 61 (supplement 1), 68 (2020); [5] Neuroscience 155, 510-521 (2008).

Last Name: Zhou Email: yzhou40@mgh.harvard.edu

Optimizing and functionalizing a radiolabeled dendrimer nanosystem for PET imaging of glioblastoma Béatrice Louis, AIX-MARSEILLE UNIVERSITE, beatrice.louis@etu.univ-amu.fr

Abstract Body : Nanometric size promotes the penetration of chemicals into solid tumors via the enhanced permeability and retention effect (EPR effect) that is not constantly found. Late diagnosis and variability of drug penetration in tumors are both regularly associated with poor prognosis. To these concerns, supramolecular radiolabelable dendrimers self-assembling into "dendrimersomes" were designed to assess EPR effect heterogeneity (see References). Aiming at lowering healthy organ exposure and increasing tumor uptake, dendrimers underwent chemical modifications for optimization. To improve targeting efficiency and specificity, functionalization of the dendrimers with a specific RGD moiety was considered to target the αvβ3 integrin, widely overexpressed in many tumors including glioblastoma. This study aimed at 1) evaluating the impact of chemical modifications on dendrimer biodistribution to select the best formulation and 2) evaluating the impact of the functionalization on the in vitro binding specificity of dendrimersomes and on their in vivo biodistribution and tumor targeting. Methods: Five different dendrimer formulations (A,B,C,D,E) were radiolabeled with gallium-68. Healthy Swiss mice (n=6 per formulation) were injected intravenously and dynamically imaged using microPET/CT. Activities were quantified at 5, 10, 15, 20, 30, 45, 60, 75, 90 , and 120min post-injection in the organs of interest. Six mice per formulation were injected intravenously with gallium-68 dendrimersome, 20μL-blood samples were withdrawn at PET time points, and quantified using a calibrated gamma counter. Mice were euthanized at 120min and organs were harvested and counted using gamma-counter. Quantifications were expressed as %ID/g. Non-compartmental analysis (NCA) and compartmental analysis (CA) were achieved with PKanalix and Monolix softwares respectively. Different ratios of the optimal radiolabelled dendrimer / RGD moiety (6:1, 3:1, 1:0, 1:3, 1:6, 1:9) were evaluated by cell-based receptor binding assay on U87 human glioblastoma cells (n=3). MicroPET/CT studies were performed in athymic nude mice (n=4) bearing orthotopic U87 glioblastoma xenografts with the best-functionalized dendrimersome identified in vitro, compared to the non-functionalized dendrimersome. Results: Radiochemical purity was superior to 95% and stable at least up to 4h in 37°C serum. Amongst the 5 formulations, microPET results with formulation E showed significantly lower retention in the liver and increased urinary elimination compared to the others with significantly higher activity in the bladder at 2h (Fig.1a , Table 1 ). These results were confirmed by gamma-counting ( Table  2) . Dendrimersome E also displayed the lowest systemic exposure with an AUC of 0.038 %ID/mm3.h, the highest clearance (2.656 mL/h), and a short blood terminal half-life (2.21 h), which could be ascribed to a more efficient renal elimination as suggested by bladder data ( Table  3 ). The volume of distribution was higher than for other dendrimers (8.002mL versus 4.031-5.954mL), suggesting a wider distribution in the organism, in line with the higher concentrations observed at the first time points in the main organs. AUCs in the main organs were contrarily lower for dendrimersome E, due to a short half-life. The semi-mechanistic model confirmed other dendrimersomes were retained longer in the liver, with lower transfer rates from the liver to blood. The 6:1, 3:1, and 1:3 ratios showed a significantly improved dendrimersome binding on U87 cells compared to the 1:0 control ratio ( Table 4 ). The 1:3 ratio was selected for in vivo studies. A significant increase of microPET/CT quantifications in tumors was observed with the 1:3 ratio functionalized dendrimersome compared with the non-functionalized dendrimersome (brain-to-cerebellum activity ratios respectively: 1.80± 0.65 and 1.47±0.62; *P =0.0382, Fig.1b) . Conclusion: Formulation E appeared as the optimal dendrimer with an optimized biodistribution profile. Functionalisation of radiolabelled dendrimersomes provides a new perspective for adding molecular targeting to passive EPR targeting, illustrating the high versatility of this type of nanosystem for molecular imaging and offering novel outlooks for theranostics applications. 0.091±0.020 (p=0.001), 0.094±0.015 (p=0.001), 0.096±0.020 (p=0.001), 0.082±0.019 (p=0.001), 0.094±0.010 (p=0.001), 0.103±0.020 (p=0.001) and 0.094±0.023 (p=0.004) in the respective brain regions; Ts65Dn 14-month-old (compared to 2-months-old) 0.073±0.023 (p=0.015), 0.085±0.029 (p= 0.004), 0.076±0.020 (p=0.006), 0.068±0.020 (p=0.013), 0.078±0.024 (p=0.002), 0.083±0.020 (p=0.013) and 0.063±0.020 (p=0.066) in the respective brain regions. At 14-months-old there was uptake difference between WT and Ts65Dn in hippocampus (p=0.031), thalamus (p=0.030), brainstem (p=0.015) and midbrain (p=0.010). Conclusion: 11C-PIB PET imaging showed similar myelination process in the brain of WT and Ts65Dn animals and showed that in older animals (14 months) the myelin content in the Down syndrome animal model is significantly lower than in its littermate wild type indicating that these animals are more affected by the aging process regarding myelin content.which is expected at higher spatial resolutions. Discussion:It has been shown that detail in low resolution 13C metabolite maps can be enhanced using transfer of high spatial frequencies from the scan of a thermal phantom, resulting in better visualization of anatomy such as the heart chambers and kidneys. Unlike the use of deep learning techniques, which require the use of large amounts of data, this approach only requires a single image. In future work, we will perform a quantitative comparison of superresolution algorithms for high-frequency transfer and apply these approaches to reconstruct higher spatial resolution human metabolic images which can not be readily acquired. Acknowledgements: NIH R01CA237466 and R01CA252037. imaging. b) Demonstration of multi-echo EPI imaging of a phantom of 13C-Urea and 13C-Acetate at a resolution of 1.6 mm with a separation of the image into separate maps by IDEAL. An SR database created from an 0.8 mm image of a 13C-Alanine phantom is used to upsample the 1.6 mm maps. The structural similarity index (SSIM) between the natively acquired high resolution image and the bicubic interpolated image is 0.2,, but increases to 0.5 for the SR image. c) Gradient-echo multi-echo maps of various containers with 13-C labeled chemicals for creating a SR database for in-vivo SR upsampling. d) An in-vivo metabolic map of 13C-pyruvate generated using IDEAL of multi-echo EPI images of a mouse injected with 13C-pyruvate. The images are upsampled 4-fold using bicubic interpolation and SR. White arrows show improved anatomic detail. e) An in-vivo metabolic map of 13C-alanine acquired at 3 mm, and upsampled 4-fold using bicubic and SR interpolation. White arrows show improved anatomic detail. Abstract Body : Introduction: Nanobiologic therapy can effectively induce trained immunity to inhibit tumor growth by overcoming the immunosuppressive tumor microenvironment. Trained immunity is a functional state of innate immune cells that is metabolically and epigenetically regulated, and can synergize with current clinically relevant checkpoint blockade therapies [1, 2] . While the simultaneous targeting of trained immunity and immune checkpoints is very effective in several tumor mouse models, the exact changes in tumor immune microenvironment remain to be elucidated. The current study explores a combination of multi-probe PET imaging and single cell sequencing techniques to acquire a full picture of immune reprogramming as a result of the aforementioned combination immunotherapy. Methods and results: Trained immunity-inducing nanobiologics (TI-nanobiologics) were produced by microfluidics from phospholipids, cholesterol, apolipoprotein A-1 (apoA1) and a muramyl tripeptide functionalized phospholipid. TI-nanobiologics have a mean hydrodynamic diameter of 20 nm ± 2 nm (Ð = 0.3) and a discoidal morphology as shown by cryogenic transmission electron microscopy ( Figure 1A ). To determine their in vivo characteristics, nanobiologics were radiolabeled with zirconium-89 (89Zr) to allow studying their biodistribution in both mice and non-human primates using PET/CT and PET/MRI respectively. Upon intravenous administration in mice, PET/CT revealed nanobiologic accumulation in the bone marrow and spleen, with a relatively low liver uptake. Dynamic PET/MRI imaging of two non-human primates showed fast bone marrow and spleen accumulation, as well as liver uptake. At 48 hours post injection, uptake was predominantly found in the bone marrow and spleen to relatively favorable levels as compared to the liver ( Figure 1B ). To locally investigate cell association, live mice were injected with fluorescently labeled TI-nanobiologic as well as green fluorescent dextran and were subjected to intravital microscopy (IVM). The TI-nanobiologic association with myeloid cells was observed throughout the calvarium bone marrow ( Figure 1C ). Myelopoiesis, the production of myeloid cells from hematopoietic stem and progenitor cells (HSPCs), is an integral part of trained immunity resulting in systemic changes in the immune landscape [3] . This process is accompanied with metabolic changes in the bone marrow which was visualized and quantified by injecting C57BL/6 mice with 18F-fluorodeoxyglucose (18F-FDG) at day five after the first TInanobiologic injection followed by PET imaging. Significantly higher 18F-FDG uptake in the bone marrow of TI-nanobiologic treated animals was observed; the mean maximum standardized uptake value (SUVmax) in this treatment group was 1.45 ± 0.31 as compared to a SUVmax of 1.02 ± 0.04 for the control group (P = 0.017) ( Figure 1D ). Resulting changes in myeloid cell accumulation were investigated using PET imaging following intravenously injected 89Zr labeled CD11b-specific nanobodies [4] . PET/CT revealed a significantly higher SUVmax in the spleen one day post-treatment (SUVmax TI-nanobiologic: 2.9 ± 0.4; PBS: 1.8 ± 0.3, P = 0.0079). Bone marrow image analysis showed a trend towards increased CD11b+ cells at day eight (SUVmax MTP10-HDL: 1.6 ± 0.2; PBS: 1.5 ± 0.2, p = 0.1508) ( Figure 1E ). Therapeutic studies in B16F10 bearing mice showed that continuous treatment with TI-nanobiologic in combination with checkpoint inhibitors resulted in a more than eightfold reduction in mean tumor volume, as compared to PBS and checkpoint inhibitors alone, at day nine (p Conclusion: This study demonstrates the value of integrating non-invasive PET imaging with deep profiling to optimize cancer immunotherapy strategies. Specifically in this study, it was found that simultaneously mobilizing the two arms of our immune system i.e. innate and adaptive immunity results in very effective anti-tumor response. (middle) Dynamic PET/MRI scans of a non-human primate 1, 30 and 60 minutes after injection of 89Zr labeled TI-nanobiologic. Fast bone marrow and spleen accumulation, as well as liver uptake can be observed. (right) PET/MRI scan of a non-human primate injected with 89Zr labeled nanobiologics displays a favorably high bone marrow and spleen accumulation relative to the liver. 1C. Intravital microscopy image of a live mouse calvarium eight hours post administration of fluorescently labeled TI-nanobiologic. Clear uptake can be seen throughout the bone marrow. FITC-dextran was injected intravenously to display the vasculature. 1D. 18F-FDG-PET of C57BL/6 mice. 18F-FDG was intravenously injected one hour before PET/CT imaging. A higher SUVmax of the bone marrow was observed in mice injected with TInanobiologic, indicating increased metabolic activity. 1E. C57BL/6 mice inoculated with 1x105 B16F10 cells treated with either TI-nanobiologic or PBS received an intravenous injection of 89Zr labeled CD11b specific nanobody at day seven. There is a higher SUVmax in the bone marrow and the spleen indicating higher amounts of CD11b-expressing cells present. 1F. Tumor growth curves of mice treated with PBS and TI-nanobiologic (dotted lines), checkpoint inhibitor immunotherapy alone (yellow boxes) versus in combination with TI-nanobiologic. Combining anti-PD-1 + anti-CTLA-4 has no significant effect on tumor growth rate and tumor size. Adding TI-nanobiologic dramatically decreases the tumor growth rate, an effect that is even more pronounced after the TI-nanobiologic regimen rises from three to six injections. 

Jagathesh Chandra Bose Rajendran, Masonic Medical research institute, rbose@mmri.edu

Abstract Body : Triple-negative breast cancer (TNBC) is the most aggressive subtype of breast cancer, accounting for the majority of breast cancer-related death. Owing to the lack of specific therapeutic targets, the development of targeted anticancer therapy has been extremely challenging. To overcome this problem, we investigate whether uPAR targeted heterologous miRNA delivery of uPA nano cocktails contain engineered extracellular vesicles (uPA-eEV and Sc-uPA-eEV) coated polymeric nanocarrier (PNCs) separately loaded with antimiR-21 or an antimiR-10b elicits synergistic antitumor responses. To monitor the tumor-specific accumulation of nano cocktails, we conjugated NH2 reactive indocyanine green (ICG) to the 4T1-eEV-PNCs, and then monitored the cocktail accumulation by dual molecular imaging modalities (NIR fluorescence and photoacoustic imaging). The uPA-eEVs coating on PNCs offers a means of enhancing natural tumor targeting affinities and thereby enhancing the antitumor activity of antimiRs -nano cocktails. The systemic administration of the uPA-eEV-PNCs nano cocktail showed a robust tumor tropism in nude and syngeneic mice models ( Figure. 1), which significantly enhanced the combinational antitumor effect of antimiR-21 and antimiR-10b that leads to complete tumor regression and extension of progression-free survival of syngenic 4T1 tumor-bearing mice. Besides, the uPA-eEV-PNCs-antimiRs nano cocktail with low dose doxorubicin displays a synergistic antitumor effect as evidenced by the inhibition of tumor growth, reduction of lung metastasis, and extension of the survival of 4T1 tumor-bearing mice. Our targeted heterologous miRNA therapy can be easily translated into the clinics by using eEVs isolated from autologous tumor cells that have the flexibility for ex vivo expansion and genetic engineering. Figure. a) Schematic illustration for the preparation strategy of 4T1engineered extracellular vesicles that display uPAR targeting uPA or scrambled uPA peptidefunctionalized PLGA nanocarriers separately loaded with AmiR-21 or AmiR 10b and the transmission electron micrograph shows the PLGA nanocarrier and 4T1-eEVs coated PLGA NPs. Negative-stain electron microscopy was performed to visualize the eEVs on the PLGA NCs. Scale bars, 100 nm. In vivo biodistribution and uPAR targeted 4T1 tumor accumulation in syngeneic subcutaneous Nude mice (nu/nu) tumor model. b) IVIS fluorescence imaging shows the whole-body biodistribution and 4T1 -tumor-specific accumulation of uPA and Sc-uPA nano cocktails formulations ICG labeled Sc-uPA and uPA nano cocktail formulations (150 μL) was administered via tail vein on days 0,6 and 12 and imaged on Days 2,7 and 15 using a Lago (Spectral Imaging system). c) Day 16 ( Photoacoustic imaging) for ICG fluorescence and photoacoustic signals. d) Ex-vivo fluorescence imaging shows the uPAR mediated 4T1 tumorspecific accumulation of uPA nano cocktails formulations. The uPA and Sc-uPA nano cocktails formulations injected nude animals were sacrificed on day 17, and the organs (liver, spleen, kidney, heart, lungs, and brain) were collected for ex vivo analysis. e) The biodistribution and tumor-specific accumulation of AmiR-21 and AmiR-10b in uPA and Sc-uPA nano cocktails formulations injected nude mice organs. The Taqman real-time qRT-PCR was used for the organ-specific biodistribution based on the quantification of AmiRs. In vivo biodistribution and uPAR targeted 4T1 tumor accumulation in the syngeneic BALB/CJ animals bearing the 4T1 subcutaneous tumor model. f) IVIS fluorescence imaging shows the whole-body biodistribution and 4T1 -tumor-specific accumulation of uPA and Sc-uPA nano cocktails formulations ICG labeled Sc-uPA and uPA nano cocktail formulations (150 μL) was administered via tail vein on days 0,6 and 12 and imaged on Days 2,7 and 15 using a Lago (Spectral Imaging system) for ICG fluorescence. g) Ex-vivo fluorescence imaging shows the uPAR mediated 4T1 tumor-specific accumulation of uPA nano cocktails formulations in BALB/CJ animals bearing syngeneic 4T1 subcutaneous tumor model. The uPA and Sc-uPA nano cocktails formulations injected nude animals were sacrificed on day 17, and the organs (liver, spleen, kidney, heart, lungs, and brain) were collected for ex vivo analysis. h) The biodistribution and tumor-specific accumulation of AmiR-21 and AmiR-10b delivered using uPA and Sc-uPA nano cocktails formulations injected in BALB/CJ animals bearing 4T1 subcutaneous tumor model. The Taqman real-time qRT-PCR was used for the organ-specific biodistribution based on the quantification of AmiRs.

First Name: Jagathesh Chandra Bose Abstract Body : An iron oxide nanoparticle, Feraheme (ferumoxytol, FH) is currently in clinical use as an intravenous iron supplement, as well as an off-label T2 contrast agent for MR imaging. In cancer patients, FH is known to passively target tumor-associated macrophages; furthermore, a variety of small molecule pharmaceuticals can be non-covalently captured and retained within the exterior carbohydrate shell of FH. FH therefore has promise as a clinical nano-drug platform to target delivery to tumor-associated macrophages and other tumor-infiltrating immune cells, while simultaneously enabling MR monitoring of nanoparticle uptake efficiency and response of the tumor to therapy. We have shown that FH can stably capture a clinically-approved immunostimulatory drug, allowing the nanoparticle-drug complex ("FH-drug") to function as an immunostimulatory agent and polarize macrophages toward an anti-tumorigenic phenotype in cancer. The B16 mouse model of melanoma produces rapidly expanding tumors that are poorly responsive to immunotherapeutic antibodies. Here, we show that FH-drug improves the therapeutic response of B16 to anti-CD40 agonist monoclonal antibodies. Mice with subcutaneous B16 xenografts receiving anti-CD40 in combination with FH-drug exhibited slower tumor growth, increased tumor inhibition, lengthened tumor latency time, and improved survival compared to mice receiving only the antibody therapy. This regimen also was more therapeutically efficacious than anti-CD40 combination therapy with an equivalent dose of free drug. FH-drug, composed of FDA-approved materials, is therefore promising as adjunctive agent to anti-CD40 therapy, and perhaps as an adjunct to other immunotherapeutic regimens in the clinic. The theranostic nature of this particle could allow simultaneous treatment and monitoring of therapeutic response, as well as noninvasive estimation of nanoparticle uptake by the tumor. The AUC was significantly reduced in the combination therapy group compared to anti-CD40 alone (p<0.001). D) Survival of mice after tumor challenge. Mice receiving combination therapy exhibited the longest survival duration compared to all other groups. Abstract Body : For many cancer types, immunotherapy is an emerging first-line treatment, with the potential to reach complete regression in patient subsets. The main effector cells mediating tumor eradication are cytotoxic T lymphocytes (CTL), which kill cancer cells in a cell-contact and tumor-antigen-specific manner. Adoptive transfer of ex vivo activated tumorspecific CTL (ACT) is, thus, a promising strategy to increase anti-tumor immunity but its capacity to control tumor growth is often insufficient. External application of heat in the feverrange (38 -40 °C) has been shown to activate and support immune effector functions in tumors. However, the rational design of ACT and hyperthermia combinations is currently hampered by an incomplete understanding how both therapies synergize and which remaining immunosuppressive mechanisms limit their combined efficacy. By using live-cell microscopy and optical reporters, we monitored structural damage induced by OVA-specific CTL to the cellular and nuclear membranes and DNA double-strand breaks in B16F10/OVA melanoma cells. CTL-mediated damage was predominantly sub-lethal and followed by rapid recovery of the tumor cell. Treatment with 39.5 °C, applied either continuously or for 1 -3 h on 2 consecutive days, significantly increased tumor apoptosis rates and correlated with stabilized CTL-tumor cell contacts and impaired recovery of melanoma cells from CTL-mediated damage. To evaluate the efficacy of ACT and whole-body hyperthermia (WBH) in the context of the melanoma microenvironment, we used intravital multiphoton microscopy combined with an imaging window for longitudinal monitoring of CTL effector function. Following intradermal tumor injection and adoptive CTL transfer, we applied WBH of 39.5 °C for 1 or 2 h and repeated the treatment every other day for one week. Directly following the first treatment, imaging revealed an immediate block of tumor cell proliferation and increased apoptosis rates. Timelapse microscopy showed enhanced CTL killing activity while CTL-tumor cell interaction dynamics remained unchanged, ranging from stable, long-lasting to highly dynamic contacts. The combination of ACT and WBH further induced the infiltration of phagocytic cells which was absent in tumors of mice treated with either therapy alone. Macrophage infiltration partially resolved or stabilized despite repeated WBH treatments. Subregional analysis of tumor cell viability further revealed a resistance of tissue-invading tumor cells to WBH monotherapy which was overcome by combined ACT/WBH treatment. Thus, kinetic imaging and intravital microscopy were successfully applied to deepen the mechanistic understanding of immune cell function during fever-range WHB which forms the basis for improved, rationale design of combination therapies.

Image/Figure Caption: Fever-range thermal therapy to support ACT efficacy. A) Longitudinal monitoring of ACT using imaging windows and intravital multiphoton microscopy (iMPM). Immune cells for transfer were isolated from fluorescent donor mice, in vitro activated and adoptively transferred by i.v. injection in tumor-bearing mice. Using an imaging window, fluorescently labeled tumors and immune cell infiltration were monitored over time (up to 3 weeks). Using iMPM allowed to monitor CTL effector function in tumor subregions at subcellular resolution. B) Repeated 1 h treatments of mice with whole-body hyperthermia at 39.5 °C significantly delayed B16F10/OVA tumor progression during the treatment period. C) Tumor mitosis and apoptosis rates, determined by dynamic iMPM, revealed inhibition of mitotic activity and enhanced tumor apoptosis rates immediately following treatment and (D) enhanced density of macrophages. Error bars, SD. Abstract Body : Introduction: Chimeric antigen receptor T (CAR-T) cell therapy is a relatively new cancer treatment in which a patient's own T cells are isolated and engineered to express a CAR. CAR engineering redirects the T cells to bind to a specific cancer antigen and induce cancer cell death. CAR-T cell therapies have shown promising results in patients with hematological malignancies with up to 70% of patients responding to treatment1. Despite this success, CAR-T cell therapies can cause severe off-target toxicities and not all patients respond to this treatment (up to 30%). Currently, CAR-T cell monitoring is limited to blood tests that are invasive and one-dimensional as they only provide information on the total number of circulating CAR-T cells and lack any information about biodistribution, particularly at tumour sites2. An imaging tool for tracking CAR-T cells could provide important patient-specific data on CAR-T cell fate to inform on potential success or failure of treatment as well as off-target toxicities. Fluorine-19 magnetic resonance imaging (19F MRI) allows for the detection of 19F perfluorocarbon (19F PFC) labeled cells non-invasively and in real time to provide information on cell location(s), cell number, and tumour homing ability. Our goal is to track the fate of 19F PFC-labeled CAR-T cells in leukemia bearing mice using 3 Tesla 19F MRI. Methods: T cells were engineered to co-express a CD19 targeting CAR and GFP, expanded, and then labeled with 5 mg/ml PFC overnight. CD19+ B cell leukemia cells (NALM6) were engineered to express the firefly luciferase (FLuc) reporter gene to enable detection of viable cancer cells with bioluminescence imaging (BLI). NSG mice (n=17) received subcutaneous injections of 1 million NALM6-FLuc cells in their left hind flank. Twenty-one days later, each mouse received an intratumoural injection of 10 million PFC labeled CAR-T cells (n=6), unlabeled CAR-T cells (n=3), PFC labeled T cells (n=5), or saline (n=3). 19F MRI was performed on a 3T clinical scanner days 1, 3, and 7 post PFC labeled cell injection. BLI was performed on an IVIS Lumina system days -1, 5, 10, and 14 from injection day in all treatment groups. Results: Flow cytometry revealed that engineered T cell populations were approximately 68.6% GFP/CAR positive. 19F signal was successfully detected in the tumours of all mice that received PFC labeled cells using 19F MRI on days 1, 3, and 7 post-injection. In vivo BLI data revealed that PFC labeled CAR-T cells show no significant difference in cytotoxicity compared to unlabeled CAR-T cells. Mice treated with labeled or unlabeled CAR-T cells have significantly lower tumour burden by day 14 compared to mice treated with labeled T cells (p Conclusions: We have shown for the first time that 19F MRI is able to detect PFC labeled CAR-T cells up to 7 days post injection using a 3T clinical scanner. Importantly, PFC labeling did not significantly affect CAR-T cell cytotoxicity in vivo. These imaging tools may have broad applications for tracking emerging CAR-T cell therapies in preclinical models and may be useful for the evaluation of CAR-T cells in patients. Abstract Body : Introduction: Focused ultrasound (FUS) in combination with microbubbles has been shown to open the blood-brain barrier (BBB) increasing drug delivery, being effective for small-molecule drugs1. Recently, RaSP FUS has been used to safely deliver fluorescently labelled-dextran (3 kDa) and AuNPs (25 nm)2,3 to the brain. PEGylated liposomal nanomedicines (ca. 100 nm) are clinically available drug delivery systems, but suffer from the inability to cross the BBB. In this study, we examine the ability of RaSP FUS to deliver liposomes safely and efficiently in the brain of mice in vivo, compared to conventional long ultrasound pulse sequences4. Methods: DiD-PEGylated liposomes (DPLs) were synthesized by co-incubation of DiD dye and PEGylated stealth liposomes for 2 hrs/37oC. The DPLs were purified using size exclusion chromatography and characterized using fluorescence/DLS. 36 female wild-type mice underwent FUS after injection of Sonovue® microbubbles and DPLs. The left hippocampus of mice was exposed to either: (i) low-energy short pulses (1 MHz, 5 cycles, MPapk-neg = 0.4 and 0.6 MPa, n = 18 mice) of US emitted at a rapid rate (1.25 kHz) in bursts (0.5 Hz) or (ii) standard long pulses (10 ms, 0.5 Hz, MPapk-neg = 0.4 and 0.6 MPa, n = 18 mice). Mice were sacrificed at 0 hr (n=16) or 2 hr (n=16) after US treatment. The delivered liposome dose was quantified using normalized optical density (NOD) measurements and tissue damage assessed by H&E staining. Staining of neurons, microglia and astrocytes was performed to examine the liposome uptake in these cells. Results: DPLs were synthesized with a dye concentration of 2.1±0.5 μg/mL with no modification to size, zeta potential and PDIs from the original PEG liposomes. DPLs were stable in human serum up to 48 h. The long pulses with 0.6 MPa showed uptake of DPLs with fluorescence signal concentrated around the blood vessels. The delivery with RaSP sequences showed more diffused delivery from vessels. ROIs around the left (US) and right (control) hippocampus were selected. The NOD analysis showed DPL delivery in brains with long pulse and RaSP sonication at 0.6 MPa, but no delivery at RaSP sonication at 0.4 MPa. H&E staining of the brains for RaSP FUS at 0.6 MPa showed no damage while damage was observed for long pulses FUS at 0.4 MPa and 0.6 MPa. Cellular uptake of DPLs was observed only in microglia and neuronal cells. No astrocyte uptake was observed for both RaSP and long pulses. Conclusion: RaSP at 0.6 MPa showed localised delivery of fluorescent liposomes DPLs (~100 nm) across the BBB with a more diffused and safer profile compared to the long pulse ultrasound sequence. The safer delivery can be attributed to the lower energy transfer in RaSP compared to the long pulse sequence. This study could allow the targeted delivery of drugs via liposomes to neuronal cells and desired regions of the brain. Abstract Body : OBJECTIVES: Therapeutic selection and monitoring in relapsing-remitting multiple sclerosis (RR-MS) is impeded by the absence of sensitive biomarkers to track the presence, infiltration and/or activation of specific immune cells in the central nervous system (CNS). Although many FDA-approved immunomodulatory therapies are available for RR-MS1, response rates are highly heterogenous and determining an efficacious treatment strategy for individual patients remains challenging. Therefore, the ability to non-invasively assess therapeutic response in RR-MS patients in real-time is a critical unmet need to improve disease management and patient outcomes. Positron emission tomography (PET) imaging of the triggering receptor expressed on myeloid cells 1 (TREM1) is a sensitive approach to detect the infiltration of pathogenic peripheral myeloid cells (e.g., macrophages, monocytes, neutrophils) into the CNS in the context of neurological injury2. Moreover, TREM1-PET has been shown to identify active disease and distinguish it from periods of remission in a mouse model of RR-MS. Here, we aim to investigate the ability of [64Cu]TREM1-mAb imaging to monitor therapeutic response to prophylactic fingolimod (FTY720) treatment in the RR experimental autoimmune encephalomyelitis (RR-EAE) mouse model of MS. FTY720 is an immunotherapy widely used to treat RR-MS by preventing the activation and infiltration of lymphocytes into the CNS, and has also been shown to modulate myeloid cell responses3. We hypothesize that TREM1-PET will permit detection of early therapeutic response in FTY270-treated animals, as evidenced by reduced CNS [64Cu]TREM1-mAb signal. METHODS: Animals: RR-EAE was induced in female SJL mice using PLP139-151 emulsified in immune adjuvant. RR-EAE mice were randomly split into untreated and prophylactic FTY270 treatment (3mg/kg daily for 12 days) groups. Mice were scored daily for disease severity. Naïve (non-EAE) SJL mice were also used for these studies. TREM1-PET: PET/CT imaging was performed 18-20 h post-injection of a 64Cu-radiolabeled anti-TREM1 monoclonal antibody ([64Cu]TREM1-mAb: 82-99 μCi, >99% radiochemical purity). Ex vivo: Following PET, mice were perfused with saline and CNS tissues were analyzed via high-resolution autoradiography and gamma counting. Brain slices used for autoradiography were subsequently stained with cresyl violet to enable quantification of [64Cu]TREM1-mAb binding in specific brain regions. RESULTS: Prophylactic FTY270 treatment prevented the development of disease in RR-EAE induced mice, whereas untreated RR-EAE mice exhibited disease progression with significantly higher EAE scores seen from day 11 (Suppl. Fig1A). TREM1-PET images revealed high signal in the spinal cords of untreated RR-EAE mice (i.e., mice experiencing active EAE disease), which was markedly reduced in RR-EAE mice treated with FTY270 (Fig. 1A) . Notably, RR-EAE FTY270-treated mice displayed similar CNS [64Cu]TREM1-mAb signal as naïve mice. PET data are supported by ex vivo autoradiography of CNS tissues following perfusion to remove unbound intravascular tracer. High-resolution autoradiography of spinal cords revealed significantly reduced [64Cu]TREM1-mAb signal in the lumbar and thoracic regions of RR-EAE FTY270-treated and naïve mice compared to untreated RR-EAE mice (p Suppl. Fig.1B-C) . These results were further corroborated by gamma counting of CNS tissues (Suppl. Fig.1D ). CONCLUSION:

[64Cu]TREM1-mAb imaging successfully identified therapeutic response in prophylactic FTY270-treated RR-EAE mice and distinguished it from an active disease state in untreated RR-EAE mice. TREM1-PET imaging has high potential for evaluating treatment efficacy in RR-MS, as well as the development and screening of novel immunotherapies targeting the CNS. Studies investigating the ability to detect responses to FTY270 in symptomatic RR-EAE mice are currently underway. Abstract Body : Introduction: The blood-brain barrier (BBB) is a major physical barrier for drug delivery. Sonopermeation, which refers to the combination of ultrasound (US) and microbubbles (MB), has been used to permeabilize the BBB and improve drug delivery to the brain. In order to enhance sonopermeation, MB design parameters such as size, shell composition and active targeting have been investigated [1] . In this study, we investigated the role of MB shape, a new design parameter, to potentiate sonopermeation-mediated drug delivery across BBB. Methods: Spherical MB were synthesized by anionic polymerization of n-butyl cyanoacrylate [2] . Rodshaped MB were synthesized by stretching spherical MB unidirectionally above their glass transition temperature [3] . Rod-shaped and spherical MB were studied to systematically compare their margination profiles, blood half-life and the magnitude of BBB permeation upon US treatment. Results: Cryo-SEM and confocal microscopy confirmed successful formation of spherical and rod -shaped MB. Upon flowing these MB in a microfluidic chip, we observed that rods showed stronger margination both in the absence and presence of blood. Real-time US monitoring in mice demonstrated that rods have a significantly higher blood half-life time than spheres. Likely as a direct result of these phenomena, rod-shaped MB in combination with focused US almost doubled the degree of BBB permeabilization and drug delivery as compared to spheres. Finally, via functionalizing the MB surface with brain endothelium-binding anti-CD71 antibodies, we demonstrate that active targeting of rod-shaped MB enhances BBB permeabilization and drug delivery even more, by up to a fourfold. Conclusion: This is the first study to exemplify the importance of MB shape in margination, circulation half-life and ultrasound-mediated drug delivery. These findings indicate that tailoring MB shape is a promising strategy for potentiating US-mediated opening of BBB to improve brain disorder treatment. Collectively, this highlights the need to develop new host-directed therapeutics to target neuroinflammation, a key determinant of pathophysiology (Jain et al 2018) . N-acetyl-cysteine (NAC) has both antioxidant and anti-inflammatory properties and can cross the blood brain barrier (BBB) when conjugated to dendrimer nanodevices (D-NAC) . Importantly, D-NAC has been shown to improve neurobehavioral outcomes in other models of neuroinflammation (Kannan 2012). We used a previously established rabbit model of TB meningitis mimicking disease in children, and performed 18F-FDG, 18F-py-albumin and 124I-DPA-713 (ligand for TSPO, which is upregulated on activated microglia/macrophages) positron emission tomography (PET) imaging and neurobehavioral scoring to noninvasively monitor response to D-NAC treatment over time. Objective: To noninvasively monitor response to host-directed treatment using PET imaging and neurobehavioral scoring in a pediatric model of TB meningitis. Methods: Young rabbits at post-natal day 5 were injected with live Mycobacterium tuberculosis H37Rv (TB-infected, n=13 rabbits) or PBS (uninfected control, n=6 rabbits) into the brain parenchyma. Intravenous treatment with D-NAC or PBS in TB-infected rabbits began 3 weeks post-infection, was blinded and continued weekly for 2 weeks. The same rabbits were imaged at day 0 and week 2 post-treatment with 18F-FDG (metabolic activity), 18F-py-albumin (vascular permeability) and 124I-DPA-713 (microglial activation) and compared to uninfected control. Regions of interest of brain lesion and unaffected brain were calculated and data was represented as ratios over time and fold change. Neurobehavioral studies and bacillary burden (CFU) were monitored during treatment. Results: We show that 18F-FDG, and 124I-DPA-713 signal correlated with brain lesions in TB-infected rabbits. Higher 18F-py-albumin PET signal was also observed at lesion sites suggesting increased BBB permeability. Brain lesion/unaffected ratios were significantly elevated in TB-infected rabbits compared to uninfected controls. Although TBinfected rabbits had similar brain bacillary burden (P = 0.471), the brain lesion/unaffected tissue ratios of all radiotracers decreased in TB-infected rabbits treated with D-NAC versus those treated with saline (controls). TB-infected rabbits treated with saline had significantly increased 124I-DPA-713 brain lesion/unaffected tissue ratios compared to uninfected rabbits (P = 0.003). This increase was not observed in TB-infected rabbits treated with D-NAC, indicating higher intracerebral inflammation in saline-treated TB-infected rabbits. Additionally, TB-infected rabbits treated with saline had significantly worse neurobehavioral scores compared to uninfected rabbits (P Conclusions: D-NAC treatment decreases intracerebral inflammation and improves neurobehavioral scores in a rabbit model of TB meningitis. Multi-tracer PET imaging can be used to noninvasively monitor intracerebral responses to host-directed therapy. These data suggest that D-NAC is a promising host-directed therapy for TB meningitis that warrants further investigation in combinatorial studies with antimicrobials. 18F-py-albumin (upper panel) and correlates with (D) the lesion on gross pathology. All images correspond to the same rabbit. Lower panel quantifies the longitudinal changes in brain lesion/unaffected ratio of (A) 18F-FDG, (B) 124I-DPA-713, and (C) 18F-py-albumin of all TB-infected rabbits pre-treatment (day 0) and post-treatment (week 2) with D-NAC (blue, n=2 rabbits, n=5 ROIs) or saline (red, n=2 rabbits, n=6 ROIs). Data are represented as fold change of the brain lesion/unaffected ratio. PET signal of all radiotracers reduces more in TB-infected rabbits treated with D-NAC compared to those treated with saline. E. 124I-DPA-713 signal significantly increases in saline-treated rabbits (red, n=2 rabbits, n=6 ROIs) compared to uninfected rabbits (grey, n=2 rabbits, n=2 ROIs) which is not observed in D-NAC-treated rabbits (blue, n=2 rabbits, n=5 ROIs). (1). However, some patients do not respond and about 30% of patients respond initially but develop resistance. Mechanisms of resistance include poor penetration of T-cells in the tumor, and the immunesuppressive function of tumor-associated macrophages (TAMs) (2) . Thus, there is a clinical need to monitor response to immunotherapy co-targeting these immune cells to spare potential nonresponders from toxicity. SPECT is a non-invasive imaging modality that can image two radioisotopes concurrently. We aim to use dual-isotope SPECT imaging to track these immune cells through 99mTc-αCD8-cDb for imaging of cytotoxic T-cells and 111In-αCD68-IgG for imaging of TAMs, respectively, in syngeneic animal models of melanoma. Methods. 99mTc-αCD8-cDb was prepared via site specific conjugation of HYNIC (3) to the αCD8-cDb followed by radiolabeling with 99mTc. 111In-αCD68-IgG was prepared via conjugation of CHX-A''-DTPA to the αCD68-IgG followed by radiolabeling with 111In. Stability and binding affinity (Kd) of the SPECT agents were characterized. The optimal imaging timepoint and in vivo specificity of 111In-αCD68-IgG were determined via biodistribution studies of 0.37 MBq of the 111In-αCD68-IgG and compared with its isotype control in non-tumor bearing mice at 1, 5 and 24h p.i. (n = 3). YUMMER1.7 allografts characterized by high CD8+ T-cells, and high TAM (2) were used for imaging studies. 30 MBq of 99mTc-αCD8-cDb were injected in YUMMER1.7 allografts (n = 4) and single-isotope SPECT/CT was performed at 5h p.i. Imaging was repeated at 1, 10, 14 and 19 days post inoculation of tumor cells to monitor CD8+ T cell infiltration during tumor growth. Separately, YUMMER1.7 allografts (n = 5) were injected at 10 days post inoculation of tumor cells with 30 MBq of 111In-αCD68-IgG and single-isotope SPECT/CT was performed at 24h p.i. Dual-isotope SPECT imaging of a phantom composed of both 99mTc and 111In solutions was performed to determine feasibility of dual-isotope imaging studies. Results. 99mTc-αCD8-cDb was produced with a 80 μCi/μg specific activity, 100% radiochemical yield, and purity. The 99mTc-αCD8-cDb was determined to be stable in PBS and mouse serum with >74% monomer up to 5h post synthesis. 111In-αCD68-IgG was produced with a 10 μCi/μg specific activity, a 100% radiochemical yield, and 80% purity. 111In-αCD68-IgG was determined to be stable in PBS and mouse serum with >70% monomer up to 24h post synthesis. Affinity (Kd) of 111In-αCD68-IgG was found to be 35.5 nM. The 5-h imaging timepoint for 99mTc-aCD8-cDb was chosen based on previous studies (3) . The optimal SPECT imaging time point for 111In-αCD68-IgG was determined via biodistribution studies and was found to be 24h p.i using the macrophage-rich spleen as in vivo standard. The highest spleen uptake of 40 ± 36 %ID/g was observed at 24h p.i. In contrast, the spleen uptake of the isotype control 111In-IgG2a was 4.5 ± 1.4 %ID/g. SPECT imaging quantification showed the greatest uptake of 99mTc-aCD8-cDb in the tumor relative to heart at 14 days (2.1 ± 0.9) of tumor growth but decreased at 19 days (0.8 ± 0.3), when the tumor volume decreased. This tumor model is dependent on Tcells, where tumor growth increases when CD4+ and CD8+ T-cells are depleted (2) . SPECT imaging of 111In-αCD68-IgG showed a tumor:heart ratio of 4.9 ± 2.2. Dual-isotope SPECT imaging of the phantom showed distinct energies between 99mTc and 111In with 140 keV for 99mTc, and 170-and 245-keV for 111In. Dual-isotope SPECT imaging of 111In-aCD68-IgG and 99mTc-αCD8-cDb in YUMMER1.7 are currently ongoing. Conclusions. SPECT Imaging of 111In-aCD68-IgG or 99mTc-αCD8-cDb showed specificity for CD8+ T cells and CD68+ TAMs in YUMMER1.7 allografts. Future studies will consist in applying dual-isotope SPECT in the context of combination therapy. Abstract Body : Introduction. Fluorocarbon (FC) is both hydrophobic and lipophobic and has high biological inertness and gas-dissolving ability. In the clinic, FC colloids are routinely used as ultrasound contrast agents and have been widely studied for oxygen delivery, with a proven safety profile and well-characterized pharmacokinetics.1,2 FC-in-water nanoemulsion (NE) is actively being evaluated in clinical trials for in vivo cell detection by fluorine- 19 (19F) MRI.3While the inertness of FC contributes to their low toxicity, it makes chemical functionalization challenging. The high electronegativity of fluorine imparts very low cohesive energy density and Lewis basicity to FC, making dissolution of other molecules difficult. Functionalization is further complicated by colloidal instability. We have designed new fluorous molecules, in particular chelators4,5 and fluorophores,6,7 that overcome solubility barriers to reside in the FC core, providing a platform for the development of multimodal imaging agents. This strategy minimizes the interaction of encapsulated molecules with biological systems and promotes agent stability in vivo. We have demonstrated the use of lipid FC NE formulations for visualization of macrophage infiltration in mouse models using multimodal imaging, including positron emission tomography (PET), 19F MRI and cryo-fluorescence tomography (CFT) (Figure 1 ). Materials & Methods. A fluorous octadentate hydroxamic acid chelator was synthesized and formulated into FC NE to capture 89Zr in NE for PET imaging.4 A fluorescent FC NE was also prepared by dissolving a fluorous Cy5 dye (lem = 672 nm) in FC.7 NE bearing the 89Zr chelator or Cy5 was formulated with lipid surfactants using a high-pressure homogenization process. The FC NE encapsulating 89Zr (3.7 MBq) or Cy5 was then intravenously injected into mice (8-10 weeks, female CD1 mice , Envigo, Indianapolis, IN) bearing acute inflammation at the left paw, induced by subcutaneous administration of λcarrageenan (2% in saline, 50 μL). Images were acquired 24 h post NE injection on multiple imaging platforms, including PET/CT (Inveon, Siemens), 19F/1H MRI (11.7 T BioSpec, Bruker), fluorescence (IVIS Spectrum, PerkinElmer) and CFT (Xerra, Emit Imaging). Serial blood samples were collected to study the circulation time of NE in vivo. Histology was used to confirm the internalization of FC NE by macrophages. All animal experiments followed protocols that were approved by University of California San Diego's Institutional Animal Care and Use Committee. Results. Lipid FC NE containing fluorous 89Zr chelator or Cy5 dye in FC oil was formulated with a droplet diameter of ~150 nm and a polydispersity NE harboring fluorous chelator was readily labeled with 89Zr, giving a radiolabeling yield of ~60% after purification. Pharmacokinetic studies showed that the NE has a blood circulation half-life of 14.5 h. In PET/CT imaging, prominent signals are observed in the inflamed left paw, but not in the contralateral paw; signals are also observed in liver and spleen (Fig. 1b) . Similarly, in 19F MRI, a hotspot is observed in the left paw (Fig. 1c) , as is also observed superficially with in vivo fluorescence imaging using the fluorous Cy5 NE (Fig. 1d) . The same mouse, subjected to three-dimensional CFT, displays bright fluorescence signals in the left paw, liver, spleen and lymph nodes (Fig. 1e) , where Fig. 1f shows the fluorescent signal overlaid on a white light image for anatomical reference. CFT is effective at yielding a high-resolution, 'ground truth' probe biodistribution at experimental endpoints. Conclusion. Synthesis of novel fluorous molecules, such as chelators and dyes, enable the functionalization of FC NE for multimodal imaging. A high degree of consistency is observed among PET, MRI and CFT in the visualization of local macrophage infiltration and agent biodistribution. Overall, these materials and methods are useful for imaging a broad range of preclinical inflammatory diseases.

Neil Robertson, Michigan State University, robe1104@msu.edu

Abstract Body : Imaging is a powerful and rapidly growing field of biomedical research, proven indispensable for early detection, diagnosis and treatment of various cancer types and tracking cells of interest in vivo. Preclinical development of novel imaging agents and modalities helps to drive the way diseases are treated and diagnosed, and link interdisciplinary approaches. Individual imaging techniques have certain advantages in imaging biological objects, but at the same time each suffers from drawbacks that other modalities may overcome. As such, the formulation of a single universal contrast agent useful in multiple modalities was explored to help bridge specific gaps in information lost with any single modality and aid in enhancing collaboration between various imaging fields. Using a single agent removes the issue of coadministered agents having differing distribution and breakdown kinetics and ensures the ability to image the region of interest with all desired modalities. Furthermore, a universal agent would allow for the manual registration of different modalities from a single administration and work with or without artificial intelligence (A.I.) systems. Therefore, the goal of the study was to develop a contrast agent that combines magnetic resonance imaging (MRI), magnetic particle imaging (MPI), computed tomography (CT), fluorescence (FL), photoacoustic (PA), ultrasound (US) and positron emission tomography (PET)/single-photon emission computed tomography (SPECT) into a single formulation. Development of this agent, termed Omniparticle, was explored using a modified co-precipitation synthesis technique followed by conjugation of small molecule moieties to provide imaging potential in all modalities. The iron core, wrapped in a modified dextran polymer for stability and functionalization, provides the MR and MP imaging potential. The functionalized fluorophore component provides FL, PA and US imaging capabilities. Attachment of chelators, such as DOTA for PET or hynic for SPECT, provides the respective PET or SPECT imaging properties. Finally, attachment of thyroxine (T4) provides CT imaging potential and completes the MRI/MPI/FL/PA/US/CT and PET/SPECT contrast agent. The synthesized agent has the following parameters: core size 15-20 nm, overall size 25-35 nm, final zeta potential +10-15 mV, 100-150 functional amine groups per particle, and moieties include 7-10 Cy5.5/particle, 7-10 DOTA/particle and 5-10 Thyroxine (T4)/particle. Phantoms were imaged in each of the modalities to probe the presence of the individual functional moieties as well as gather information to adjust the ratio of loading on the contrast agent. Fig. 1 shows potent signal generation with each modality upon addition of corresponding moieties. To probe the in vivo functionality of our contrast agent, an initial pilot study was performed across multiple imaging modalities. Exploring a model of breast cancer, we imaged the ductal tree of naïve mice after an intraductal injection of the agent (40 µg Fe/duct), prior to moving into tumor bearing animals. Immediately post injection, the animal was imaged with a microCT, MPI and with the IVIS fluorescence imaging. As seen in Fig. 2 , imaging of the intraductal injection was possible with each of the modalities, with the CT capable of resolving the ductal tree network (Fig. 2a) . After rendering the slices we obtained a 3D view of the entire network (Fig. 2b) .Fluorescence imaging of Cy5.5 (Fig. 2c) as well as the simple 2D (Fig. 2e) , and the 3D MPI scan co-registered with the mouse's skeleton, showed spatial localization of the injected agent (Fig.  2d) . Ex-vivo histology showed accumulation of the agent in the ductal epithelial cells as expected. In the future we plan to utilize this agent to image tumor models along with tracking labeled cells (not shown here). We believe that by providing a universal tool to the biomedical imaging field we can corroborate registered signals and potentially enhance the information acquired in the clinic. Abstract Body : Background: Liver disease is a leading cause of morbidity and mortality in people with human immunodeficiency virus (HIV) (PWH), even in those optimally-treated with antiretroviral therapy (ART) [1] . HIV itself can exert damaging effects on hepatocytes, liverresident macrophages (Kupffer cells), and hepatic stellate cells. Immune cell depletion and other viral pathogenesis mechanisms can also contribute to liver dysfunction, inflammation, and fibrosis [2] [3] [4] [5] [6] [7] . The extent of HIV-associated liver damage and recovery remains unclear. PET/CT is a valuable, non-invasive approach to elucidate liver dysfunction in PWH and establish a connection between imaging findings and hepatic changes that could provide additional guidance in HIV disease management. In this study, we used [18F] -fluorodeoxyglucose (FDG) PET/CT imaging to longitudinally evaluate hepatic structural and functional changes in ART-naïve PWH with advanced disease until two years after treatment initiation. Clinical, laboratory, and immune markers were also measured to identify other factors associated with liver injury and recovery. Methods: Forty-eight viremic PWH (33 males, 15 females), with CD4 counts Results: At baseline, PWH were younger than control participants with median ages of 36.25 and 52.55, respectively (p Conclusions: ART-naïve, late-presenting PWH demonstrated low baseline liver FDG uptake that increased two years after treatment but remained lower than healthy participants. These changes correlated with markers of immune recovery and decreased systemic inflammation following treatment. Our findings suggest a multifactorial model for liver injury in HIV that likely reflects CD4 depletion, monocyte/Kupffer cell activation, and hepatocyte dysfunction. While long-term ART seems to reverse many aspects of hepatic abnormalities, residual liver injury identified non-invasively through FDG PET/CT imaging may exist, at least within two years from treatment initiation, especially in PWH who present with low nadir CD4 counts.Kupffer Cells in Driving Hepatic Inflammation and Fibrosis in HIV Infection. Front Immunol, 2020. 11: p. 1086. Abstract Body : Introduction: Cell tracking is a rapidly advancing field, providing critical insight on transplanted cell fate preclinically.1 This is essential for improving our understanding of biological mechanisms and developing cellular therapies.1 Prior to tracking, cells must be labelled either directly through endocytosis, or indirectly through the introduction of a reporter gene.1 To give complementary information, and limit the associated weaknesses of each method, this project will combine direct and indirect labelling for the tracking of cancer cells. Direct labelling is a highly sensitive method. However, the tracer is diluted upon cell division and may be expelled by cells where it is taken up by scavenger cells.1 Due to this, imaging based on direct labelling is limited to short term tracking.1 The novel Magnetic Particle Imaging (MPI) excels at short term tracking of cells directly loaded with iron nanoparticles due to its high sensitivity, positive signal contrast, and ability to directly quantify iron-loaded cells.2 Indirect labelling is less sensitive, however, as the reporter gene is passed equally to daughter cells during cell division, and is quickly denatured outside of living cells.1 Indirect labelling allows for longterm tracking of both cell viability and proliferation.1 Akaluc/Akalumine based Bioluminescence Imaging (BLI) has shown the ability to detect a single cell at depth in an animal model, due to the low biological attenuation of near-infrared light, released by the oxidation of Akalumine by Akaluc.3 BLI will therefore be used to image indirectly labelled cells in this project. This project will develop the first dual MPI/BLI system for tracking cancer cells, using the BLI reporter Akaluc, providing a more complete picture of cell fate in preclinical models. Methods: A mouse breast cancer cell line (4T1BR5) was transduced to express Akaluc, and the fluorescence reporter TdTomato (TdT). The expression level and stability of this construct was analyzed with flow cytometry over 10 passages. BLI of varying cell numbers was done to determine signal linearity. Cells were loaded with a commercial iron nanoparticle, Synomag-D. Loading efficiency was calculated through iron staining. Viability was assessed for loaded and unloaded cell populations through Zombie Violet staining and analysis with flow cytometry. Additionally, BLI signal intensity was compared for these cell populations following administration of Akalumine (n=9). The resolution and sensitivity of Synomag-D was assessed in comparison to Vivotrax (the industry standard) using in vitro MPI relaxometry (n=3). Calibration lines were created by imaging a range of Synomag-D concentrations, plotting signal versus concentration. Results: The TdT-Akaluc construct was successfully transduced and the percentage of cells (79%) stably expressing TdT did not change over 10 passages. Staining showed clear and efficient Synomag-D uptake, (98% of cells labelled, n=9). Viability and BLI expression were not significantly different for loaded and unloaded cell populations(n=9). BLI signal linearly and positively correlated with cell number (p=0.2). The sensitivity of Synomag-D was roughly 4x that of Vivotrax, with comparable resolutions (p Discussion: Here we developed the first dual MPI/BLI system for cell tracking with the highly sensitive BLI reporter Akaluc. The combined use of MPI and BLI will give complementary information, leading to a more complete understanding of cell fate in the body. The stable expression of Akaluc, offers a highly sensitive marker for cell viability which is not affected by iron loading. As MPI and BLI signal are both linear, quantification of cells should be feasible in a mouse model, and is the focus of our current work Abstract Body : Purpose: Immune checkpoint inhibitors (ICIs) have had great success but remain ineffective in over 50% of patients due to several patient-specific response variabilities. 1 The immunosuppressive effects of the endothelial 'barrier' (also termed EC anergy) have been related to EC surface expression of immunosuppressive markers (i.e. PD-L1) that deter effector T cells; identifying patients with such EC markers could help select patients for ICIs, and spare non-responders. Here, we present initial developments and validation of the first molecular ultrasound microbubble (MB) probes to image immunosuppressive markers on ECs. Method: MBs were fabricated by microfluidic procedure. Briefly, fixed-ratio of lipids i.e. DPPC, DPPA, and PEG5000 PE (Avanti Polar Lipids Inc.) (mole ratio 10:80:8) were hydrated with 2 mL buffer solution with propylene glycol, glycerol. The lipid suspension was then sonicated for 5 minutes and then homogenized by purging through LV1-microfluidic system (Microfluidics, Westwood, MA) operating at 30,000 psi. The homogenate was then supplemented with Pluronic® L-64 surfactant (0.3 mg/mL) and sealed in Wheaton vials pre-filled with perfluoropropane gas (6 mg/mL) and finally activated in a Vialmix (Lantheus Medical Imaging) to produce suspension of MB. The resulting MB were characterized by Accusizer for size distribution and concentration. Tissue-mimicking gelatin-agar based pipe phantoms were prepared as per procedure outlined by Shiying W et al.2 The inner lining of one of the conduit was surface functionalized with PDL-1 protein by EDC/NHS chemistry. PDL-1 targeted and non-targeted bubbles were infused into functionalized phantoms and imaged by linear array transducer MS250 (13-24 MHz) operated by Visualsonics (Vevo 2100) system. The presence of anti-PDL1 antibody on the lipid shells of microbubble and its binding affinity to 4T1 mouse endothelial cells was assessed in parallel plate inverted flow chamber system under steady state flow conditions outlined in earlier work.3 In vitro evaluation of anti-PDL1 and IFN-γ markers influence on T cell and Treg migration towards tumor cells was investigated with the dual chamber setup outlined in Fig 1g. Results: Conjugation of PDL-1 to PEG5000 PE was optimized and confirmed by detection of additional band in Anti-PEG immunoblot assay at 130 kDa (Fig 1a) . FACS analysis of Alexa-647-anti PDL-1 engineered microbubbles led to prominent shift in fluorescent MBs as compared to ratisotype conjugated MB which established the presence of anti-PDL1 antibody on lipid shell of MB (Fig 1b) . Evaluation of concentration and size distribution of MB revealed presence of 4-5 x 108 microbubbles/mL with average size of 1.1 µm (±0.39 µm) which was consistent for both PDL-1 targeted and control MB (Fig 1c) . Assessment of MB stability over time indicated that ~35% of MBs were stable after 4 hours of activation (Fig 1d) . The PDL-1 specific binding affinity of MB was evaluated in PDL1-antigen functionalized tissue mimicking phantom, and revealed consistently higher level of acoustic signal (mean echo power) in PDL-1 antigen functionalized channel as compared to unfunctionalized channel which confirmed the ligand specific affinity and binding ability of anti-PD-L1 MB (Figure 1e /f). In order to identify the implication of PDL-1 molecular imaging as potential screening agent for responders to ICI immunotherapy, we examined the role of immunomodulators (anti-PDL1 and IFN-γ) in altering the expression of CD47 and PDL-1 markers in CT-26 cells in the presence of murine bone marrow derived immune cells. Simultaneous treatment with anti-PDL-1 antibody and IFN-γ led to decline in CD47 and PDL-1 expression levels in CT26, and decline in viability of CT-26 seeded in the lower chamber (Figure 1g 

Yu-Jin Lee, Stanford University, yujinlee@stanford.edu

Abstract Body : INTRODUCTION The current standard of care preoperative imaging to stage head and neck squamous cell carcinoma (HNSCC) is performed using 18F-FDG-PET/CT, which nonspecifically highlights regions with significant metabolic activity. This lack of specificity for cancer presents challenges when incidental findings are identified and require additional invasive biopsy procedures. To address this, a highly specific PET radiotracer (89Zr-radiolabeled panitumumab) was tested for use in combination with conventional 18F-FDG-PET/CT diagnostic imaging of HNSCC in humans. Panitumumab is a monoclonal antibody targeting the epidermal growth factor receptor that is highly expressed in HNSCC. The current study investigated the safety and feasibility of this agent for detection of primary HNSCC and metastatic lesions. MATERIALS AND METHODS We enrolled 14 patients with HNSCC scheduled for surgical resection and cervical lymph node (LN) dissection in a single center study. Patients underwent 18F-FDG-PET/CT as standard of care preoperative imaging prior to study enrollment. 89Zr-panitumumab (1mCi) was administered intravenously 1-5 days prior to surgery, and 89Zr-PET/CT scans were performed prior to surgery. Both 18F-FDG-PET/CT and 89Zr-panitumumab-PET/CT scans were serially reviewed by a nuclear medicine physician. Surgical specimens were processed according to routine standard of care, assessed by a pathologist who was blinded to the imaging results, and used as gold standard for detection of local-regional HNSCC. Diagnostic statistics were evaluated using receiver operating characteristic curves, and logistic regression was performed to evaluate the combined ability of 18F-FDG-PET/CT and 89Zr-panitumumab-PET/CT to detect HNSCC. Likelihood ratio was used to compare the predictive model of 18F-FDG-PET/CT alone and the combined model of 18F-FDG-PET/CT and 89Zr-panitumumab-PET/CT scans to detect HNSCC. RESULTS A total of fourteen patients were enrolled and completed the study. Four patients (28.5%) had areas of high metabolic activity outside the head and neck region with SUVmax (maximum standardized uptake value; quantified PET signal) greater than 2.0 on 18F-FDG-PET/CT that were not detected on 89Zr-panitumumab-PET/CT. All four patients with incidental findings underwent further workup and were found to have no evidence of cancer on biopsy or followup. 89Zr-panitumumab-PET/CT imaging demonstrated uptake of 89Zr-panitumumab in the tumor lesions as early as 22 hours after study drug infusion (primary tumor SUVmax 4.0 ±2.6; mean ± standard deviation). A total of 47 lesions (primary tumor, LNs, incidental findings) with SUVmax signal (9.9 ± 6.1; mean ± SD) were visualized on 18F-FDG-PET/CT and 35 lesions on 89Zr-PET/CT (3.1 ± 2.5; mean ± SD). 89Zr-panitumumab-PET/CT had higher specificity (88.9%) and area under the curve (AUC; 0.89) compared to that of 18F-FDG-PET/CT (74.1%, 0.79, respectively) in detecting HNSCC using pathologic assessment as gold standard (p > 0.05; DeLong's test). The combined ability of 18F-FDG-PET/CT and 89Zr-panitumumab-PET/CT to detect HNSCC was significantly improved with higher specificity (96.3%) and AUC (0.91) (p DISCUSSION/CONCLUSION 89Zr-panitumumab-PET/CT imaging was safe and identified metastatic lymph nodes, with the added ability of distinguishing HNSCC from incidental findings visualized on 18F-FDG-PET/CT. Larger studies with appropriate power are needed to evaluate the potential of 89Zr-panitumumab-PET/CT to accurately stage HNSCC in a minimally invasive manner.

First Name: Yu-Jin Last Name: Lee Email: yujinlee@stanford.edu

Valerie Porter, University of California Davis, vporter@ucdavis.edu

Abstract Body : Introduction: Voxel-wise image segmentation is one of the most common tasks performed in preclinical brain MRI analysis pipelines [1] . Atlas co-registration protocols are useful; however, they are suboptimal in rodent models of acute organophosphate (OP) intoxication because of large deformations in brain regions over time post-exposure [2, 3] . Manual segmentation is time consuming and may be subjective. Therefore, efficient and robust image segmentation tools are needed. The U-net convolutional neural network (CNN) [4] has been utilized for automated segmentation of rodent brain images [5] , however previous studies have not employed these methods in the context of acute OP intoxication. We tested the hypothesis that the U-net CNN is capable of providing automated segmentation of rodent brain MRI scans in a longitudinal study of brain pathology in a rat model of acute OP intoxication. Methods: T2-weighted MRI brain scans (n=33, 59 slices/scan, acquired on a Bruker BioSpec 7T scanner, acquisition resolution 125x125x500 µm) of adult Sprague Dawley rats subjected to acute OP intoxication and follow-up treatments were manually segmented by an experienced human observer to delineate 8 brain regions, namely by hemisphere: hippocampus, lateral and medial thalamus, piriform cortex, and entorhinal cortex. Image data consisted of scans of rats from five treatment groups at three time points (3, 7 , and 28 days post-exposure). Treatment groups included diisopropylfluorophosphate (DFP), DFP plus midazolam (MDZ), DFP plus allopregnanolone (ALO), DFP plus midazolam and allopregnanolone (DUO), and vehicle controls (VEH). The T2-weighted scans provide a range of both acute and chronic pathologies [6] . The training and test datasets consisted of 28 scans (DZP=5, MDZ=9, ALO=6, DUO=5, VEH=3, all five treatment groups) and 5 scans (DFP=1, MDZ=3, ALLO=1), respectively. The image data (size=280x200x59) were preprocessed by center-cropping the images and labels (size=200x200x59) to capture the entire brain and down sampling the in-plane resolution (128x128x59) to decrease the runtime. The U-net CNN used a learning rate of 0.002 with a batch size of 10 and 100 epochs. The image data was augmented by shifting each slice randomly between (-10,10) pixels in the x and y directions and randomly rotating the image between (-10,10) degrees. The augmentation also included the brightness levels of the image slices were increased or decreased by ± 20% of the original value, and Gaussian noise was added to each image slice. Dice coefficients of volumetric overlap were calculated between U-net CNNgenerated labels and manually segmented labels. The results are reported as median (range). Results and Conclusion: The loss function did not change significantly beyond 30 epochs indicating that the network was able to quickly find robust segmentation features despite the heterogeneous presentation of lesions ( Figure 1 ). The model accuracy however improved going from 20 epochs to 100 epochs from 84% to 93% (Figure 1 ), which suggests that additional epochs in future studies may be beneficial. The Dice coefficients were in the moderate range for most regions, being the highest for the right thalamus, 0.62 (0.39-0.68) ( Table 1) . Segmentation accuracy was marginal for the right piriform cortex, 0.27 (0.20-0.28). The training took 15 minutes, while the generation of an individual segmentation took approximately 1 minute. We conclude that the U-net CNN-based segmentation approach was fully automated and efficient; however, its accuracy was in the low to moderate range. Future research will include examining the impact of adding more MRI scans of OP intoxicated rats with varying pathologies to the training data set, MRI bias correction, and skull stripping to improve the accuracy of the segmentation method. Supported by the National Institutes of Health via grants R21AG064599 and U54NS079202. [4] B) The model accuracy and loss graphs over epochs. C) Images from two different test scans that were segmented manually and with the U-net CNN, where the first row is the anatomical image, second row is the manual segmentation of the anatomical image, and the last row is the U-net segmentation. Blue/Purple are the hippocampus, Yellow/Light Green are the lateral thalami, Dark Green is the medial thalamus, Red/Orange are the entorhinal cortexes, and Teal/Light Blue are the piriform cortexes. 

Kelly von Beck, Vanderbilt University School of Medicine, kelly.n.vonbeck@vanderbilt.edu

Abstract Body : The patient is a 4 year-old male with a history prematurity (31-weeks), developmental delay, chronic lung disease, sleep apnea, gastroparesis, and dysphagia treated with two rounds of esophageal Botox injections of the lower esophageal sphincter (LES). He presented to the ED with acute abdominal pain three hours after an esophagogastroduodenoscopy (EGD) with dilation of the LES. He was ill-appearing on presentation with respiratory distress including audible grunting and diffuse subcutaneous emphysema. He underwent a chest x-ray which revealed pneumomediastinum and an esophagram that confirmed a distal esophageal perforation with contrast extravasation into the left pleural space. He was admitted to the PICU for resuscitation. He became febrile, tachycardic, and hypotensive overnight and was therefore taken to the OR due to concern for mediastinitis and septic shock. Bilateral chest tubes were placed under fluoroscopic guidance to drain inflammatory pleural fluid. An EGD was then performed which showed a large distal esophageal perforation. On-table esophagram revealed a linear 2cm posterolateral perforation along the distal 1/3rd of the esophagus with communication to the mediastinum. All contrast was then suctioned out, and the mediastinum was irrigated using a guidewire and kumpe catheter. The distal end of the perforation was located 2cm above the gastroesophageal (GE) junction, which was used as the landing zone for placement of an esophageal stent. External radiopaque markers were used to delineate the GE junction, as well as the distal and proximal ends of the perforation. A Boston scientific 18mm diameter x 10cm length covered self-expanding metal stent (SEMS) was chosen based on the gathered measurements. A kumpe catheter was advanced through the mouth into the stomach, which was then exchanged for a 0.035-inch Amplatz wire for delivery of the stent. The stent was advanced to 1 cm above the GE junction and slowly deployed, with the proximal end appropriately below the pharynx. Positioning was confirmed endoscopically and on x-ray. A catheter and guidewire were then used to place a distal jejunal feeding tube through the stent for enteral nutrition. On postoperative day (POD) 6, the patient failed extubation due to compression of the trachea by the esophageal stent. He remained intubated until the stent was removed on POD 13 with good healing of the esophagus and no evidence of perforation. The patient required prolonged intubation post-procedure given the small size of the child and the available SEMS options. Although frequently cited as a rare complication of esophageal balloon dilation, the risk of esophageal perforation has been shown to be as high as 12.5-14% in pediatric patients [1, 2] . While the application of esophageal stenting has been well established in adults, investigation on the use of SEMS in children has been limited to studies with small sample sizes. One retrospective case series reports the use of SEMS designed for trachea-bronchial use in the treatment of esophageal atresia, fistula, and corrosive ingestion in patients from 4 years to 16 months old [3] . The authors report the safety and efficacy of this approach as a secondary treatment for benign esophageal strictures, citing stent migration as the only complication [3] . Similarly, Manfredi et al demonstrate a 100% technical success rate of placement and retrieval of esophageal stents and an 80% closure success rate in dilation-induced perforations in their cohort of pediatric patients [4] . Surgical repair of esophageal perforations is associated with significant perioperative risks. This case demonstrates that SEMS are a safe and effective alternative treatment in children with acute esophageal perforation. This treatment modality warrants further investigation in order to define criteria for placement in the pediatric population and increase experience with the procedure among pediatric surgeons and pediatric interventional radiologists. (D) confluent with the mediastinum (E). Distal and proximal ends of the perforation were marked externally with radiopaque markers (F), and the stent was placed with the distal end lying between the perforation and GE junction (G) . Chest x-ray shows placement of the stent (H) and resolution of the perforation after stent removal 13 days later (I).

Last Name: von Beck Email: kelly.n.vonbeck@vanderbilt.edu

Celine Taglang, University of California San Francisco, celine.taglang@ucsf.edu

Abstract Body : Introduction The Warburg effect, which is characterized by elevated glucose uptake and flux to lactate, is a metabolic hallmark of cancer1. Recent studies have identified deuterium (2H)-magnetic resonance spectroscopy (MRS) using [6,6'-2H]-glucose as a novel method of imaging the Warburg effect in high-grade primary glioblastomas (GBMs)2. However, its utility for imaging low-grade gliomas and for assessing treatment response has not been tested. The goal of this study was to determine whether [6,6'-2H]-glucose can be used for imaging tumor burden and treatment response in mutant isocitrate dehydrogenase (IDHmut) lowgrade gliomas in vivo. Methods Orthotopic tumor implantation and treatment: We examined mice bearing orthotopic tumors of the BT257 model, which was derived from a patient carrying an IDHmut low-grade astrocytoma3. Eight 6-to-8-week-old female SCID mice were intracranially injected with ~3×105 BT257 cells. Once tumors reached a volume of ~20-25mm3 (day zero, D0), mice were randomized into treatment (100 mg/kg niraparib oral daily) or control (saline) groups. Anatomic MR imaging: MR studies were performed using a 14.1T vertical system (Agilent) equipped with a single-channel volume 1H coil. A spin echo multi-slice sequence was used for axial T2-weighted MRI (TE=20ms, TR=1200ms, field of view=30×30 mm2, matrix=256×256, slice thickness=1mm, number of averages=4). Tumor volume was evaluated as the sum of manually contoured tumor areas in each slice multiplied by slice thickness. 2H-MRS studies: Animals were positioned prone under a 16mm 2H surface coil (DOTY Scientific). Following injection of a bolus of [6,6'-2H] -glucose (1g/kg) via a tail-vein catheter over 25-35s, non-localized 2H-MR spectra were acquired over 66 minutes with a pulseacquire sequence (TR=500ms, averages=500, complex points=512, flip angle=64, spectral width=2kHz, temporal resolution = 4min 10s). 2H spectra were referenced to semi-heavy water (HDO; 4.75 ppm, MestReNova). The area under the curve (AUC) for [3,3'-2H]-lactate was normalized to [6,6'-2H]-glucose and HDO for the sum of all spectra. For spatial localization, a 2D chemical shift imaging (CSI) sequence with TE/TR=1.35/250ms, FOV=30x30mm2, 7x7, 128 points, 2.5kHz spectral width, NA=20 was used and data analyzed using in-house Matlab codes. Statistical analysis: All results are expressed as mean±standard deviation. Statistical significance was assessed using an unpaired two-tailed Student's t-test with p Results 2H-MRS monitors tumor burden in IDHmut gliomas in vivo: In line with higher glucose uptake and lactate production in BT257 neurospheres relative to normal human astrocytes, [6,6'-2H] -glucose metabolism to lactate was observed in BT257 tumor-bearing mice. In contrast, tumor-free mice showed production of glutamate/glutamine, but not lactate. Importantly, 2D CSI confirmed spatial localization of lactate to the tumor vs. contralateral normal brain in mice bearing orthotopic BT257 tumors. 2H-MRS monitors early response to therapy in IDHmut gliomas in vivo: Poly-(adenosine 5′-diphosphate-ribose) polymerase inhibitors (PARPi) inhibit IDHmut glioma growth and are in clinical trials for IDHmut glioma patients4,5. Treatment of mice bearing orthotopic BT257 tumors with niraparib induced tumor shrinkage as assessed by T2weighted MRI, an effect that was apparent by D18 and resulted in significantly enhanced survival (Fig.4C) . Importantly, [6,6'-2H] -glucose flux to lactate was reduced in BT257 tumorbearing mice treated with niraparib relative to controls. Furthermore, this reduction was observed at early time-points (D4-5 and D8-9) when no difference in tumor volume could be detected by MRI pointing to the potential ability of [6,6'-2H] -glucose to assess pseudoprogression. Conclusions We show that IDHmut gliomas display the Warburg effect, which can be leveraged for non-invasive, non-radioactive 2H-MRS-based imaging of tumor burden in vivo. Importantly, [6,6'-2H] -glucose flux to lactate can be used for non-invasive visualization of IDHmut glioma response to therapy at early time-points when volumetric alterations cannot be detected by MRI, suggesting that [6,6'-2H]-glucose has the potential to assess pseudoprogression, which is a major challenge in glioma imaging6.

Neil Robertson, Michigan State University, robe1104@msu.edu

Abstract Body : Objectives: Validation of nanoparticle delivery to the tumor of interest is critical for the successful treatment of cancer. Nanoparticle-based technologies, including theranostic nanomedicine, are fast developing areas of biomedical research offering unprecedented potential for cancer treatment. The ability to image and follow nanoparticle accumulation in vivo, as well as observe clearance time, greatly enhances our ability to treat these tumors. An in vivo multimodal optical/MRI imaging probe was designed with a bioorthogonal chemistry activation and validation technique in mind. The nanoparticle utilizes the highly specific nature of the "click to release" reaction to label the particles taken up by the tumors with a secondary fluorophore, highlighting the location of the particle constructs within the body, validating delivery and activating therapeutic payloads. Bioorthogonal chemistry has previously been explored for potent pretargeting [1, 2] and toxicity studies in vivo [2] [3] [4] , as well as clinical studies [5] . In this study we investigated in vivo function of a nanoparticle construct in combination with bioorthogonal chemistry to noninvasively image nanoparticle delivery to both primary and metastatic tumors. We hypothesized that uptake of the nanoparticle conjugated to trans-cyclooctene (TCO) would occur in the tumors after systemic injection, followed by accumulation of the non-toxic secondary labeling molecule tetrazine. This accumulation would only be possible due to reaction between trans-cyclooctene (TCO) and tetrazine and therefore, should occur only in the tissue with significant nanoparticle uptake. Methods: A dual modality (magnetic resonance and near infrared optical imaging) nanoparticle (MNP) construct labeled with Cy7 and a bioorthogonal TCO moiety (MNP-Cy7-TCO), as well as a fluorophore labeled small molecule probe, tetrazine labeled with Cy5.5, (Tet-Cy5.5) were synthesized. The construct consisted of a dextran-coated superparamagnetic iron-oxide nanoparticle (MNP), modified with amine termini and conjugated with the near-infrared fluorescent dye, Cy7, and the bioorthogonal trans-cyclooctene moiety. The small molecule probe synthesized was a dual amine tetrazine, which was conjugated with the near-infrared fluorescent dye, Cy5.5. The synthetic strategy employed for the particle was a co-precipitation technique as previously reported [6, 7] . For the small molecule tetrazine, a protected tetrazine intermediate was reacted with hydrazine to produce a dual amine probe easily conjugated. The synthesized construct, MNP-Cy7-TCO, was tested in a MDA-MB-231 mouse model of breast cancer. Animals with primary tumors and lymph node metastases were intravenously injected with the nanoparticles followed by intravenous injection of tetrazine 24 hrs later. Controls included tumor bearing mice injected with a non-clickable MNP construct, MNP-Cy7, followed by an injection with tetrazine 24 hrs later and mice who only received the tetrazine injection. Noninvasive imaging was performed with an IVIS Spectrum and verification of the nanoparticle presence was done with histology. Results: Characterization of the particle construct and the probe showed a particle size of roughly 30 nm with a zeta potential of +20-22 mV. The particles were also found to have approximately 7-10 Cy7 molecules per particle. The tetrazine product was characterized with FT-IR spectroscopy, which showed the presence of amine functional groups only post reaction and purification. In vivo studies demonstrated that the same nanoparticle construct was capable of imaging both primary and metastatic breast tumors in mice (Fig. 1) . Fluorescence microscopy showed excellent co-localization of Cy7 and Cy5.5 signals, corroborating our imaging data (Fig.  2) . Conclusions: The multimodal aspect of the nanoparticle construct is valuable for noninvasive imaging, while the biologically inert bioorthogonal reaction provides both an in vivo "tagging event" for validation and an operator-controlled approach to construct activation in future studies. Overall, the nontoxic nature of the explored system further validates future use of the chemistry coupled with therapeutic payloads.International Society of Cerebral Blood Flow and Metabolism 16, 834-840, doi: 10.1097/00004647-199609000-00008 (1996) . 

Jinghang Xie, Stanford University, jinghang@stanford.edu

Abstract Body : Immunotherapy is revolutionizing the management of many cancers. The lack of specific, non-invasive diagnostics monitoring tumor early response, however, remains an unmet clinical need in advancing precision immunotherapy. Granzyme B (gzm B) is a serine proteinase uniquely executing the cytotoxic T cell mediated cancer cell killing and tumor eradication and represents an ideal target for molecular imaging of patient early response to immunotherapy. Based on our highly translatable target-enabled in-situ ligand aggregation (TESLA) probe template, we designed and synthesized a novel gzm B activatable self-assembly small molecule probe G-SNAT for real-time imaging of gzm B activity in vivo. A series of studies in vitro confirmed the gzm B-specific probe activation. Cyanine 5 (Cy5) fluorophore conjugated probe (G-SNAT-Cy5) was highly retained in both activated cytotoxic T cells and CAR (chimeric antigen receptor) T-cell engaged cancer cells in correlation to gzm B expression and activity. In lymphoid tumor (A20) implanted BALB/c Rag2−/−γc−/− mice treated with CD19-targeted firefly luciferase expressing CAR T cells, bioluminescence imaging (D-luciferin) tracked the persistence of injected CAR T cells and G-SNAT-Cy5 successfully reported the cytotoxic response of tumors. To assess the clinical value of G-SNAT for imaging checkpoint blockade therapy and study associated immune cell trafficking, we constructed a colorectal cancer model with transgenic BALB/c mice expressing firefly luciferase in splenic cells and treated with a combined regimen of anti-PD-1 and anti-CTLA-4. Longitudinal bioluminescence imaging with D-luciferin and a gzm B specific luciferin substrate (GBLI2) developed by our group revealed the dynamics of immune cell infiltration and triggered cytotoxicity in responders. Further G-SNAT-Cy5 fluorescence imaging predicted the therapeutic outcomes before the difference of tumor volume was evident. These functional imaging studies also revealed tumors in nontreated and treated nonresponders an "immune desert" lacking immune cells and cytotoxic responses. These results support further development of G-SNAT for PET imaging of immune response to checkpoint blockade and CAR T-cell therapy in patients and highlight the role of multimodality imaging in understanding the fundamentals of immunotherapy.

Kaustav BERA, Case Western Reserve University, kxb413@case.edu

Abstract Body : Introduction -Artificial Intelligence (AI) enabled quantitative radiomic analysis1 which involves high-throughput extraction of sub-visual features from radiographic images has shown great potential in oncology. In non-small cell lung cancer (NSCLC), these radiomic features on CT scans have shown value in diagnosis, prognosis as well as predicting response to different therapies including chemotherapy, radiation as well as most recently immunotherapy2. These radiomic features have so far been derived from inside and immediately adjacent to the tumor. Meanwhile, recent studies have shown that estimation of fat from CT scans can prognosticate outcome in multiple different cancers3, including treatment prediction in NSCLC4. In this novel study, we have build an integrated model which combines traditional radiomic features from inside the tumor as well as radiomic features from different fat regions on multi-slice CT scans from pre-treatment PET/CT images in order to predict response to immunecheckpoint inhibition therapy ICI (nivolumab) in metastatic NSCLC. Methods-A retrospective database review at a single hospital between 2015-2020 after fulfilling inclusion criteria of availability of pre-treatment PET/CT as well as complete response information led to N=38 patients with Stage IV NSCLC treated with nivolumab. ICI response was defined as per RECIST 1.1 into responders(stable disease (SD), partial response (PR), complete response (CR)), and non-responders and non-responders (progressive disease (PD)). A total of 55 nodules were annotated on these scans by an expert radiologist. In addition to the tumor region, the regions associated with fat regions including total body fat, subcutaneous fat, visceral fat, intramuscular fat, and muscle volume region were also delineated using a semi-quantiative approach using inhouse software along with radiologist input. From all of these annotated regions, radiomic textural features were extracted. These features included features from Gabor, Laws, Haralick, and Collage feature families2. The top performing features were then selected within 3-fold cross-validation settings with an embedded classifier for each fold. The mRMR feature selection algorithm5 was used in combination with the Random Forest classifier. Model performance was estimated using AUC, sensitivity, specificity and accuracy of predicting response Results-The three top features during the feature selection method included one from the intratumoral from Laws textural family and two from the visceral fat region from Gabor textural feature family. These texture features represent the degree of heterogeneity among the tumors with the feature values being higher in non-responders showing higher degree of heterogeneity as compared to responders. AUC of 0.72 +/-0.06 was achieved with the integrated tumor+fat radiomics model with the accuracy of 0.72 +/-0.05 and sensitivities and specificity of 0.72 +/-0.14 and 0.71 +/-0.14, respectively in predicting response. Using only the intra-tumoral radiomics model , the AUC of the model was 0.62 +/-0.07 with an accuracy of 0.64 +/-0.05. The addition of Radiomics features from the fat regions thus improved the overall performance of the model.Conclusion-Radiomics analysis of semi-automatically derived visceral fat regions from multislice CT images of pre-treatment PET/CT in NSCLC patients treated with ICI has been shown preliminarily to accurately predict response to these agents. When integrated with traditional radiomic analysis, this can provide a robust method for response prediction. This proof-ofconcept study will be further validated in larger multi-site data sets and finally deployed in a prospective clinical trial before being ready to be used clinically. Abstract Body : While many cellular phenomena in mammalian organisms are studied in cell culture, these in vitro conditions often fail to reconstruct the complexity that cells experience in their native environment. However, studying mammalian cellular processes in a live intact organism requires noninvasive imaging tools that can penetrate deep inside the body. Commonly used light-based imaging techniques, while powerful in vitro, suffer from poor penetration into intact tissue. In contrast, ultrasound is a widely used biomedical technology that enables imaging of cells deep inside the body with high spatiotemporal resolution. Recent work has established gas vesicles (GVs), microbially derived gas-filled protein nanostructures, as genetically encodable acoustic reporter genes1, 2. However, the use of these proteins in mammalian cells has been limited by the need for tedious clonal expansions of sorted cell lines. To facilitate the use of acoustic reporter genes in a broader range of biological contexts such as primary cells and endogenous tissues, there is a need for an efficient, accessible, and a versatile platform to deliver these genes to cells both in vitro and in vivo. Here, we developed a modular viral delivery platform that efficiently integrates acoustic reporter genes into mammalian cells and tissues. This platform is based on viral vector architectures that successfully co-express more than 10 genes to produce strong ultrasound contrast. To demonstrate the modularity of this system, we expressed GVs with different promoters across several cell types, including neuronal and immune cells. We demonstrate the in vivo application of this technology by injecting our engineered viral vectors deep inside the brain and imaging the nonlinear acoustic contrast from in situ transduced cells transracially in live animals. Our viral GV delivery platform serves as a critical resource to greatly expand the in vivo imaging applications of acoustic reporter genes. Abstract Body : Cancer of the breast, with an estimated 279,100 new breast cancer cases for both men and women, is the most diagnosed cancer for the year 2020 in the United States. Despite advances in standard chemo and radiation therapies for breast cancer, patient morbidity and mortality from metastasis remains the most challenging to address. Current treatment paradigms are not yielding effective control or elimination of this lethal aspect of cancer. Critical to the understanding of metastasis are pathways that supply mechanical cues to cancer cells to migrate. Cancer cells invade and migrate through a complex network of collagen fibers and other extracellular matrix (ECM) components such as fibronectin, laminin and hyaluronan on their metastatic journey. In previous studies cyclooxygenase-2 (COX-2) downregulation resulted in a significant decrease of collagen 1 (Col1 fibers), numbers of cancer associated fibroblasts (CAFs), and metastasis, whereas COX-2 overexpression resulted in a significant increase of Col1 fibers, CAFs, and metastasis [1] . COX-2 catalyzes the production of prostaglandin E2 (PGE2) as a cellular response to inflammation via prostanoid receptors. COX-2 is overexpressed in 40% of human breast cancers and has a pleiotropic and multifaceted role in carcinogenesis. Upregulation of COX-2/PGE2 is associated with poor prognosis in many cancers, including breast cancer. Our purpose in this study was to determine if COX-2 driven changes in the ECM, including changes in Col1 fibers and their patterns, altered ECM stiffness. Altered ECM stiffness can provide mechanical cues to cancer cells through mechanotransduction pathways to modify migration and invasion. Establishment of pre-clinical triple negative SUM-149 human breast cancer cells that stably overexpress COX-2 (SUM-149-COX-2) or with an empty vector (SUM-149-EV) was achieved through cloning and construction of a lentivirus vector expressing the COX-2 gene. These cells were inoculated in the mammary fat pad of severe combined immunodeficient (SCID) mice. Nine weeks post-inoculation, excised xenograft tumors were measured using iNano nanoindenter (Nanomechanics Inc.) to measure tissue elastic modulus (i.e. stiffness). In parallel we evaluated molecular changes in ECM proteases such as MMP1, UPA and UPAR that actively play a dynamic role in the formation and degradation of ECM using western blot analysis. Indentation results were processed into resulting photomaps and heatmaps of indented regions and tissue stiffness, respectively (Fig. 1A) . Representative photomap and heatmap of tissue stiffness measured in a SUM-149COX-2 tumor are shown in Figure 1A . Values for multiple indentation points measured per individual tumor and data summarized from elastic modulus-probed tumor tissues of SUM149EV Control (n=8) vs. SUM-149-COX-2 (n=9) are displayed in Figure 1B . Resulting indentation data identified, for the first time, a trend of reduction in tumor stiffness with COX-2 overexpression. COX-2 overexpression also altered the protein expressions of UPAR and MMP1 (Fig. 1C) , which may be associated with the alterations of tumor tissue stiffness. These insights expand our understanding of the molecular interplay between inflammation and ECM remodeling during tumor development that may lead to development of imaging probes and therapeutics to target mechanoresponsive proteins. Abstract Body : Introduction, hypothesis, and objectives: Adipose tissue engineering, often including hydrogels as cell scaffolds, is emerging as a promising alternative to autologous fat transfer, in particular for breast and facial reconstructive surgical procedures.1,2 However, in vivo implantation of voluminous reconstructed adipose tissue (AT) substitutes requires the development of imaging procedures addressing indicators such as volume retention and remodelling.3 In fact, excessive tissue remodeling in vivo is one of the main problematics occuring in AT grafts over time, and the development of in vivo imaging methodologies enabling their follow-up post-implantation appears timely. Reconstructed AT grafts are not always easy to delineate in MRI, and addition of MRI-visible additives into the hydrogels constituting the cell scaffolds, could facilitate delineation and segmentation of the implants.3 In this study, we postulated that the addition of ultra-small iron oxide nanoparticles (USPIONs) in alginate-based cell scaffolds could enhance their MRI-visibility, thus allowing their volume follow-up over a period of several weeks. The impact of hydrogel formulation on the differentiation potential of adipose-derived stromal cells (ASCs) is also demonstrated. Materials and methods: USPIONs (5 nm diam.) were synthesized by thermal decomposition and pegylated by phosphine-oxide PEG. The particles were characterized for size (DLS, TEM), molecular signatures (FTIR) and relaxometric properties (1H T1 and T2 at different fields). USPIONs were mixed with alginate formulations and jellified using a 3D-printing approach which allowed to build functional channels in the hydrogels (perspective of long-term implantation in vivo : oxygenation, vascular growth). Hydrogels with and without USPIONs were s.c. implanted in mice flanks (balb/c female mice, n = 6 each group), and MR-imaged at t = 0, 3, 10, and 14 days post-injection (T1-w. spin-echo; echo time/repetition time: 13.5/704.2 ms). Blood samples were analyzed for signs of inflammation (CBA cytokine test : IL-1, IL-10,IL-12p70; MCP-1; IFNgamma; TNF). Finally, ASCs were incubated with the hydrogel, then induced into adipocytes in order to evaluate the impact of the formulation in the differentiation potential of ASCs, which is a key aspect in the preparation of AT grafts.4 Results: PEGylated USPIONs were successfully synthesized with a core size Conclusion: USPIO-alginate hydrogels can be implanted as cell scaffolds in vivo to provide a persistent visibility in MRI. The gel remains on the site of implantation after 2 weeks, with MRI-visibility preserved. The hydrogel has limited impact on ASCs differentiation, which confirms the potential of the technology for the development of reconstructed AT substitutes and their visualization over time in vivo by MRI (for volume retention quantification). Acknowledgements: Mr P. L. for his contribution to USPIO synthesis. Abstract Body : Photodynamic therapy (PDT) utilizes light and photosensitizer to generate cytotoxic species such as singlet oxygen. The effectiveness of PDT, especially the type II pathways, depends on the availability of oxygen in the vicinity. Many tumors lack oxygen, making them resistant to PDT and other therapies. Hence there is need to enhance the oxygen content of tumors while delivering sufficient photosensitizer dose. Here we employed a perfluorocarbon nanodroplets as an oxygen carrier along with the photosensitizer Benzoporphyrin Derivative (BPD) and a photoacoustic imaging contrast agent Indocyanine Green (ICG) to enhance the imaging signal. The PA signal provides information about the oxygen saturation and drug concentration accumulated in the tumor. In summary, this new nanodroplet platform has the capability to significantly enhance PDT while providing patientspecific treatment options.

Last Name: Mallidi Email: Srivalleesha.Mallidi@tufts.edu

Dongjun Li, Georgia State University , a377682953@gmail.com

Abstract Body : The extracellular matrix protein collagen is a major biomarker for multiple diseases including fibrosis, metastasis, aortic aneurysms, cardiovascular diseases, and fibrotic stroma of cancer. Collagen's expression types, level and spatial pattern are largely dependent on locations, types of diseases and disease progression state. It is essential to develop a non-invasive imaging methodology to allow us to longitudinally map the dynamic changes of collagen expression during disease progression and upon treatment including radiation. Our group designed and developed a human protein-based contrast agent with a collagen-targeting moiety (hProCA32.collagen) based on key determinates for metal binding sites in proteins and relaxivity. In this study, we report our recent breakthrough in optimization, characterization, formulation, and production of a designed novel human protein-based contrast agent with strong collagen-targeting specificity (hProCA32.collagen) pioneered by our team. We have shown that hProCA32.collagen exhibits 6.7-fold and 13.7-fold higher binding affinity for collagen type I than type III and IV respectively. The expression and purification yield have been largely improved by fermentation and large-scale optimization with 2-fold increase in solubility and significantly improved shelf life (up to 12 months) with stable protein conformation and little to no aggregation. Meanwhile, hProCA32.collagen possesses high relaxivities per particle at both 1.4 and 7.0 T and high collagen binding affinity (Kd: 1.0 µM), which enables the robust detection of early-stage liver fibrosis via dual contrast modes. Furthermore, hProCA32.collagen has high Gd3+ binding affinity (2.0 ± 0.25 × 10-22 M) and up to 15-fold higher metal selectivity with r1 and r2 relaxivity of 27 mM-1s-1 and 41 mM-1s-1 per Gd3+. Our in vivo imaging studies and histological analysis revealed that the newly formulated hProCA32.collagen is able to detect tumors less than 2 cm in Hepatocellular carcinoma (HCC) mouse model as well as early-stage liver fibrosis with high sensitivity and specificity using magnetic resonance imaging (MRI). The newly formulated hProCA32.collagen is expected to have broad preclinical and clinical applications for various diseases. 1D3B9882B798BFAA_abstract_File10351/GA359_ImageFigure_0513092121.pdf Image/ Figure To accomplish targeted near-infrared photoimmunotherapy (NIR-PIT) 3, we synthesized an IR700-labeled Gr1 antibodyphotosensitizer conjugate to target Gr1+ cells. In this proof of principle study, we have previously demonstrated the specificity of the Gr1-IR700 conjugate to detect and eliminate splenic MDSC in culture. Here, we validate the effectiveness of Gr1-IR700-mediated PIT in vivo. Methods: Gr1-IR700 and IgG-IR700 synthesis was achieved through the attachment of an NHS-activated NIR phthalocyanine dye, IR700, to the free amine residues on the RB6-8C5 monoclonal antibody and rat IgG2b,κ isotype control. Previously, Gr1-IR700 specificity for spleen-derived Gr1+ cells from tumor-bearing mice was validated using CCK-8 viability assays. In this current work, we inoculated 4T1-luc mammary carcinoma cells in the mammary fat pad of 4-6 week old female BALB/c mice to induce the generation and expansion of MDSC. Intraperitoneal (i.p.) injection of IR700-conjugated antibodies was administered once tumors reached approximately 750 mm3. Biodistribution studies were carried out using this orthotropic tumor model and non-tumor-bearing (NTB) mice to determine the retention of Gr1-IR700 and IgG-IR700 at 24 hours. For NIR-PIT experiments, tumor-bearing (TB) mice were i.p. injected with 100 μg of Gr1-IR700 followed by NIR-PIT of the spleen at 200 J/cm2 48 hours post injection. Metabolite changes in PIT-treated spleens were identified through high resolution 1H magnetic resonance spectroscopy (MRS) of dual-phase extracts using a Bruker Avance III 750 MHz (17.6T) MR spectrometer. A 5 mm inverse triple-resonance (TXI) probe was used for data acquisition, and Topspin 3.5 software was used for data processing, analysis and quantification. In vivo Gr1-IR700-mediated MDSC cell death will be determined using CCK-8 viability assays and fluorescence-activated cell sorting (FACS) analysis of spleens in TB-Ctrl, Gr1-IR700 and Gr1-IR700+PIT-treated groups. Results and Discussion: The Gr1-IR700 and IgG-IR700 biodistribution data in Figure 1A and 1B illustrate a striking increase in mean fluorescence intensity (MFI) of the spleen and lungs of TB mice injected with Gr1-IR700 compared to Gr1-IR700-injected NTB mice and IgG-IR700-injected TB mice, indicating myeloid cell recruitment and binding of Gr1-IR700 in TB mice. Metabolite changes between the TB-Ctrl, Gr1-IR700 and Gr1-IR700+PIT groups were determined. The millimolar concentrations normalized to BALB/c mouse spleen weight presented in Figure 1C from 4T1-luc tumor-bearing TB-Ctrl (n=3), Gr1-IR700 (n=3) and Gr1-IR700+PIT (n=3) groups demonstrate significant changes in valine, isoleucine, lactate, acetate, glutamate, succinate, glutamine, glutathione, aspartate, choline, creatine, P-choline, GPC, total choline, taurine and myoinositol with Gr1-IR700+PIT. The robust change in glutamine is of particular interest because of the role of glutamine in T cell metabolism. These findings, taken together with previous in vitro data, confirm the ability of Gr1-IR700 to eliminate MDSC, illustrating its potential application for MDSC theranostics in the TME.

Alexandre Le Roux, Memorial Sloan Kettering Cancer Center, lerouxa@mskcc.org

Abstract Body : Metastatic cancer is the leading cause of death for patients diagnosed with colorectal cancer (CRC), with the liver being the most common site of metastasis. The tumor microenvironment influences the growth, response to treatment and tumor immune responses in vivo, and subcutaneous tumors do not accurately represent the natural tumor microenvironment. Therefore, we used ultrasound-guided percutaneous inoculation to develop a pre-clinical orthotopic model of CRC liver metastasis. This orthotopic liver xenograft model was used to evaluate an anti-tumor/anti-DOTA bispecific antibody (bsAb) pretargeted radioimmunotherapy (PRIT) platform and radiohaptens developed for imaging and radiotherapy (lutetium-177). We used zirconium-89 (Zr-89) as an imaging surrogate due to its long physical half-life and imaging resolution, and effective half-life similar to the therapeutic radiohapten. A novel radiohapten precursor DO3A-proteus, mirroring the "gold standard" β-radiohapten [177Lu]LuDOTA-Bn was radiolabeled with [89Zr]ZrCl4, for theranostic imaging of the liver xenograft model. DO3Aproteus consists of a commercial 2,2',2''- (10-(17-amino-2-oxo-6,9,12,15-tetraoxa-3azaheptadecyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl) triacetic acid (NH2-PEG4-DO3A) and non-radioactive lutetium 2-(4-isothiocyanatobenzyl)-1,4,7,10-tetraazacyclododecanetetraacetic acid (p-SCN-Bn-DOTA· Lu3+ complex) prepared from p-SCN-Bn-DOTA (Macrocyclics, Inc.) and 175LuCl3· 6 H2O. For preparation of [89Zr]Zr-DO3A-proteus, DO3Aproteus (30 nmol) was mixed with [89Zr]ZrCl4 (UW-Madison Cyclotron Lab, 74 MBq/2 mCi) in 0.5M HEPES (pH 7.5). After 60 min at 95°C, the reaction was purified with a Strata-X SPE cartridge to obtain [89Zr]Zr-DO3A-proteus (RCY: 33%, AM= 0.93 MBq(25 µCi)/nmol, RCP by radio-RP-HPLC: >99%). GPA33-expressing human colorectal cancer SW1222 cells (40 μl bolus of 2 million cells suspended in PBS:Matrigel (1:1 v/v)) were inoculated percutaneously within the liver of athymic nude mice (n=10, female, 6-8 week-old) by freehand using the ultrasound monitor for reference (40 MHz transducer, Vevo 2100 high-resolution imaging system, Visualsonics). Adequate tumor growth was achieved in 5 weeks, and the mice were injected i.v. with bsAb huA33-C825 (250 µg, 1.19 nmol), followed 24 h later with i.v. injection of dendrimer-clearing agent (10% (w/w), 25 µg, 2.76 nmol), and long-lasting microCT contrast agent (10 mL/kg, Fenestra LC, Medilumine). Finally, [89Zr]Zr-DO3A-proteus was injected i.v. 4h after the clearing agent (0.37 MBq (10 µCi), 0.4 nmol). Imaging (Inveon PET/CT) was performed at ~4, 24h and 48h post-injection (p.i.). Ex vivo biodistribution assay of radioactivity was performed at 6h (n= 4 mice) and 48h (n= 6 mice) post-injection. 5 non tumor bearing athymic nude mice were also used as control and received the same doses of bsAb, clearing agent, microCT contrast agent and [89Zr]Zr-DO3A-proteus.[89Zr]Zr-DO3A-proteus showed rapid renal clearance, with minimal organ retention and demetallation. Following ex vivo biodistribution assay of radioactivity, the blood, SW1222 tumor, non-tumored liver and kidney uptakes at 6h p.i. were respectively 0.24 ± 0.07 %ID/g, 4.0 ± 1.2 %ID/g, 0.21 ± 0.03 %ID/g, and Dittgen, T. et al. Lentivirus-based genetic manipulations of cortical neurons and their optical and electrophysiological monitoring in vivo. Proc Natl Acad Sci U S A 101, 18206-18211 (2004) . [8] Kugler, S., Kilic, E. & Bahr, M. Human synapsin 1 gene promoter confers highly neuronspecific long-term transgene expression from an adenoviral vector in the adult rat brain depending on the transduced area. Gene Ther 10, 337-347 (2003) .

Image/ Figure Caption: Ultrasound-based gene delivery is limited in spatial precision because the FUS beam is too long and spans millimeters in length, leading to nonspecific viral transduction and gene expression. Moreover, when targeting near the base of the skull, there is reflection of the FUS beam, causing pressure build-up and undesired damage in the form of hemorrhage. To overcome these limitations, two separate gene delivery sessions can be performed with the gene-editing target delivered first followed by erasure of gene expression in undesired areas of the brain with the delivery of the gene-editing enzyme. By using this advance in precision targeting, desired gene expression will only be restricted to the intended site and allow for targeting difficult-to-reach areas of the brain such as a single layer of cortex. Abstract Body : Background: Besides brain microglia, the translocator protein (TSPO) is expressed by macrophages and other peripheral immune cells, making it a potentially suitable biomarker for tracking reactive changes in peripheral organs in infectious/inflammatory diseases (e.g. lungs, liver, spleen, brain, and bone marrow), including sepsis. The bone marrow is especially important as it plays a central role during inflammatory/infectious events with secondary increased hematopoietic stem cell (HSC) differentiation (hematopoiesis). Still, only a handful of studies have used PET with TSPO specific ligands to evaluate peripheral organ inflammation in sepsis. In our study, using a rat model of LPS-induced inflammation (sepsis model), we performed whole body PET imaging with a TSPO tracer, [18F] DPA-714, to determine changes in TSPO expression in both the brain and lungs, liver, bone marrow, and spleen. In order to assess the physiologic relevance of TSPO upregulation, we correlated our imaging results with several serum and brain biomarkers of inflammation. Methods: Thirteen male Fisher rats (weight 0.32 ± 0.01 kg, age 3.96 ± 0.17 months) underwent dynamic PET scans with [18F] DPA-714 on an Inveon microPET/CT scanner. Scans were collected at baseline and 4 hours after intravenous LPS injection (5mg/Kg). Reconstructed PET images were analyzed using PMOD 3.8 to generate time activity curves (TACs) for brain, lungs, liver, bone marrow, and spleen. Mean Standardized Uptake Value (SUVmean) were calculated using emission data at equilibrium (26-40 mins). We measured serum cytokines and, following euthanasia and perfusion, assessed organ TSPO gene expression (qPCR) levels. Multiplex Immunofluorescence (IF) staining was done on the lung, spleen, and brain tissues for TSPO, CD68 (monocytes/macrophages), and Iba1 (microglia in brain only). In the brain, we measured microglial ramifications' length and cell body area. Paired t-tests were done to evaluate the differences in [18F] DPA-714 uptake between baseline and post LPS status. Repeated measures correlations between SUV values and serum cytokine levels were performed using the rmcorr package in R. Results: Mean TAC (n=13) for whole-brain, lungs, liver, and bone marrow showed increase in [18F] DPA-714 binding post-LPS compared to baseline. Post-LPS SUVmean were significantly higher for brain (p =0.006), lungs (p=0.02), liver (p Conclusion: Using noninvasive molecular imaging with [18F] DPA-714 PET in a rat model of sepsis, further validated by qPCR and IF microscopy, we demonstrated organ-level brain, liver and lung inflammatory changes as well as spleen immune cell efflux and activated hematopoiesis in the bone marrow. [18F] DPA-714 PET can be helpful in better understanding of the pathophysiology of systemic inflammation in sepsis and its organ-level effects. Abstract Body : Introduction: Epithelioid sarcoma (ES) is a rare yet high-grade and highly aggressive soft tissue neoplasm first identified by Enzinger in 1970.1 ES has high reported recurrence and metastasis rates (around 70% and 50%, respectively).2 Unresectable locally advanced and metastatic ES are usually chemoresistant and generally fatal.3 Here, we extended a previous near-infrared photoimmunotherapy (NIR-PIT) study,4 to develop antibody photosensitizer conjugates based on two monoclonal antibodies (mAbs) binding to CD44 or EGFR, both overexpressed on ES tumor cell surface. We conducted cellular studies and established animal models to evaluate the specificity and efficacy of the two mAb conjugates in binding, detecting and eliminating ES. Method: CD44, EGFR mAb or its IgG isotype control was conjugated with a NIR phthalocyanine dye, IR700, to form CD44-IR700, EGFR-IR700 and IgG-IR700, respectively. A bilateral ES tumor model was established by inoculating 1×106 VAESBJ cells in 0.05 ml of Hank's balanced salt solution on either flank of athymic Balb/c (nu/nu) female mouse. Subsequently, 100 μg of CD44-IR700 or EGFR-IR700 or IgG-IR700 was injected intravenously (i.v.), and NIR fluorescence images of IR700 in mice were obtained over a 24-h period (n = 4 per group). For PIT, tumor-bearing mice (n = 4 or 5 per group) received two i.v. injections of 100 μg of CD44-IR700 or EGFR-IR700 at a one-week interval, and tumors were exposed to NIR light exposure at 200 J/cm2 at 24h p.i. Tumor diameters were measured over 3 weeks. IgG-IR700 and PBS-injected mice were used as controls. Results and Discussion: CD44 and EGFR overexpression in VA-ES-BJ cells and human ES sections were validated by flow cytometry and immunohistochemistry (Fig. S1 ). Both CD44-IR700 and EGFR-IR700 demonstrated concentration and exposure-dose dependence of cellular PIT and target-specific cell death of VA-ES-BJ cells. Compared to EGFR-IR700, CD44-IR700 exhibited a higher potency with cellular PIT and had a much lower IC50 value due to the higher expression level of CD44 on VA-ES-BJ cell surface (Fig. S2 ). CD44-IR700 and EGFR-IR700 both exhibited preferential accumulation in VA-ES-BJ tumors and showed approximately two to three-fold higher retention compared to IgG-IR700. EGFR-IR700 showed a higher tumor-to-normal ratio and less retention in organs than CD44-IR700 (Fig. S3) . EGFR-IR700 injection combined with NIR light exposure resulted in significant tumor growth delay in VAESBJ tumors compared to the IgG-IR700 or PBS group or EGFR-IR700 without PIT group; CD44-IR700-PIT also resulted in significant tumor growth delay compared to the PBS group (Fig. 1) . Reduction of CD44 and EGFR levels were observed in the CD44-IR700 or EGFR-IR700-treated tumors as evidenced by flow cytometry and western blotting studies (Fig. S4) . Our data collectively demonstrate the specificity and efficacy of CD44 or EGFR-based PIT as novel therapeutic approaches that can be integrated into an intraoperative setting for achieving clear tumor margins in localized ES or applied as a standalone therapy for inoperable locally advanced and recurrent ES.

Jehyeong Yeon, Yonsei University, yy77887@yonsei.ac.kr

Abstract Body : Introduction The glymphatic system (GS) is assumed to be a waste clearance system of the cerebrospinal fluid (CSF) through the perivascular and interstitial space of the brain [1] . Natural sleep is associated with about 60% increase in the interstitial space, resulting in a conspicuous increase in convective exchanges of CSF with interstitial fluid (ISF) [2] . To date, no studies have proposed a method of observing the influx of CSF into the interstitial space from non-surface to deep water using magnetic resonance diffusion tensor images along cellular space volume. In this study, we hypothesized that the extracellular space (ECS) volume fraction involved interstitial space expands from 30% to 50% during sleep, filling the dilated space with incoming CSF. The diffusion scalars of the corpus callosum (CC) were affected by sleep [3] . Therefore, using Monte Carlo simulation (MCS), we simulated diffusion tensor signals according to an increase in ECS volume ratio in the voxels of the corpus callosum and obtained a diffusion scalars: frcational anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD) and axial diffusivity (AD). The purpose of this work was to detect the most sensitive diffusion scalars from the ECS volume changes using MCS. Methods For MCS, the voxel size of genu of CC was set to 20μm×20μm×20μm and the number of Axons was 420. Axons are acquired by fitting a gamma distribution to the histograms of axon radii as follows [4] . We integrated a threedimensional diffusion motion by updating the position of each water molecule in every Δt during a diffusion time Tdiff. Details of the simulation parameters used in our study are shown in Table  1 [5] [6] [7] [8] [9] [10] . Table 1 . The MCS parameters Parameter Value Meaning θ 0.122 Scale parameter in radius gamma distribution k 3.5 Shape parameter in radius gamma distribution mxOD 2.6µm Maximum of outer diameter N 10000 Total water molecules Δt 1µm Update time Tdiff 120 ms Diffusion time ID/OD 0.63 (g-ratio) Inner diameter to outer diameter ratio Din 2.5×10-3mm2/s Intracellular space diffusivity Dout 3.5×10-3mm2/s Extracellular space diffusivity b 1000s/mm2 b-value G 80mT/m Gradient strength P 5 .5µm/s Permeability coefficient of water molecules The diffusion tensor signal was obtained by increasing gradually to about 5% of the initial ECS volume fraction from 30% to 50%. By adding noise to the obtained diffusion tensor signal, the diffusion scalars according to ECS volume fractions were obtained when SNR = 10, 20, 30, and infinite. MCS was performed using software developed by Sapkota et al [6] . Results The FA, MD, RD and AD ranges of the human corpus callosum were reported by Liu et el [8] . Our results are depicted in Figure 1 . All diffusion scalars for the initial ECS volume fraction representing the non-sleep state were in the normal ranges mentioned above. As the ECS volume fraction increased, MD and RD increased and FA decreased, and AD did not change significantly. Regardless of the SNR, FA and MD can detect changes in ECS volume ratio of 10% or more, whereas RD can detect changes in ECS volume ratio of 5% or more. RD is very sensitive in changes of ECS volume because it can be distinguished even under 5% regardless of SNR. Discussion Decreased FA with increasing ECS volume may be caused by the disruption of myelin and awakening was associated with widespread increases in white matter FA, which were mainly driven by RD reductions [11] . The result can be interpreted as a change in FA and RA because ECS volume was increased by shrinking myelin through MCS. FA and MD were not significantly different for ECS volume difference less than 10%. RD was most sensitive compared to other scalars, and ECS volume changes were discernable from more than 5%. Therefore, RD can be used as an appropriate index to monitor sleep states.

Alex Dhaliwal, University of Toronto, alex.dhaliwal@mail.utoronto.ca

Abstract Body : INTRODUCTION Microbubbles are ultrasound contrast agents used for diagnostic imaging. They derive their contrast from the expansion and contraction of a flexible shell around a gaseous core that occurs upon stimulation with acoustic waves. This interaction can also be manipulated to enhance local drug delivery using focused ultrasound (FUS), whereby interactions between microbubbles and their confining vasculature create openings at sites of ultrasound irradiation. This transient vasculature disruption allows for enhanced extravasation of drugs co-injected alongside the microbubbles or embedded within the microbubbles themselves 1-3. While microbubble-mediated FUS has been utilized to improve chemotherapy delivery to solid tumors 4, little is known about the impact of microbubble composition on important drug delivery parameters such as pharmacokinetics, biodistribution, and degree of FUS-enabled delivery enhancement to tumors. A thorough understanding of the interplay of these parameters is essential towards the intentional and optimal design of microbubble delivery systems. OBJECTIVE To quantitatively elucidate the effect of microbubble shell charge (+/-anionic phosphatidic acid lipids) and lipid chain length (C16, C18) on longitudinal organ biodistribution in the presence and absence of focused ultrasound in an orthotopic breast cancer model. METHODS A series of Definity®-analogue microbubbles was designed with varying lipid chain lengths (C16, C18) and the presence or absence of a negatively charged shell (+/-phosphatidic acid). These compositions were modified through the inclusion of porphyrin-lipid 5, a multifunctional imaging moiety capable of facilitating quantitative microbubble shell tracking in vivo through radionuclide chelation and fluorescence imaging. Microbubbles were coordinated to the radioisotope 64Cu using a simple, one-pot chelation strategy, after which they were injected into 4T1 tumor-bearing Balb/c mice and exposed to focused ultrasound (1 MHz, 300 kPa, 1% duty cycle, 2 minute total sonication time). Longitudinal biodistribution was assessed via PET imaging over 48 hours, while end-point quantification was assessed at 3.5 and 48 hours via gamma counting and hyperspectral fluorescence imaging. Degree of agent extravasation was assessed by comparing to mice perfused prior to dissection using both the porphyrinmicrobubble platform and an Evans Blue standard. RESULTS Biodistribution was dominated by organs of the reticuloendothelial system, sequestering primarily within the liver, followed closely by the spleen. Uptake followed similar shell dependencies as seen with other lipid drug delivery platforms such as liposomes 6-8, but the degree of organ specificity was magnified -specifically, splenic uptake was greater for longer chain lengths, whereas liver uptake was greater for shorter chain lengths and negatively charged bubbles. Gamma counting of blood samples revealed that shell circulation follows a two-phase exponential decay with slow halflives on the order of 4-10 hours depending upon shell composition. This was corroborated by biodistribution in perfusion-dominated organs (heart, lungs, kidneys), which followed the trends set by the microbubble shell clearance kinetics. Importantly, focused ultrasound-mediated tumor delivery varied by shell composition, with an increase in shell extravasation only apparent for microbubbles constructed with C18 chain length lipids without phosphatidic acid. These differences were limited to microbubble shell extravasation and not to bulk changes in tumor vascular permeability as assessed by co-delivery of Evans Blue. IMPACT The trends and relationships uncovered in this robust exploration of microbubble shell biodistribution will better inform the design of microbubble-based delivery systems. Specifically, the use of charged lipids and tailoring of lipid chain length can alter organ selectivity of shell deposition and tumor uptake, providing a framework for those using microbubble-enabled FUS to enhance drug delivery to different tumor sites. Importantly, these quantitative longitudinal measurements revealed pharmacokinetic trends which were previously obscured in studies utilizing shorter study timeframes 9-11 or end-point metrics alone, highlighting the importance of thorough pharmacokinetic evaluation of microbubble platforms. utilize PET data for which 3D contours for each individual organ were constructed and analyzed. C) Liver displays increased signal at early timepoints with dependence on both chain length and charge, showing highest signal for C16PA microbubbles. D) Spleen displays progressive and sustained increased signal for C18 microbubbles with or without PA compared to C16 microbubbles. E) In 4T1 orthotopic breast cancer tumors, C18nPA microbubbles displayed the greatest overall accumulation and degree of improvement following focused ultrasound treatment. F) Relative percentage change in tumor signal comparing FUS-treated and FUSuntreated tumours, illustrating that lipid composition has a substantial influence on the efficacy of a microbubble-based delivery system. (1). Because most colorectal lesions originate from the epithelial cells lining the luminal surface of the colon, biomarkers expressed on the polyp or lesion surface present an opportunity for early, real-time, in vivo detection of pre-cancerous lesions. Preliminary studies from our group have demonstrated the feasibility of using antibodyfunctionalized hyperpolarized (HP) silicon micro-and nano-particles to detect tumors expressing MUC1, a mucin glycoprotein aberrantly expressed in CRC (4) . Although such surfacefunctionalized microparticles (MPs) have great potential value in identifying early polyps/lesions, the complex biological signatures of these lesions merit the concurrent use of multiple biomarkers, to increase specificity in detection. We envision that microparticles with a "universal" surface functionalization by the IgG-binding Protein L would enable modular incorporation of multiple antibodies, against multiple targets. In the present work, we demonstrate the utility of this method with our existing MUC1+ cell systems and tissues. We hypothesize that CRC-targeted microparticles will provide ultrasensitive, non-invasive molecular imaging agents for detection of the majority, if not all, of the CRC lesions. Methods: CRC cell lines (HT29-MTX, SW480, HCT116, and CACO2), and paraffin-embedded sections of primary polyp tissues representing the most common types of colorectal polyps (hyperplastic, sessile serrated, tubular, and villous/tubulovillous), were stained with various antibodies including EGFR, EpCAM, and MUC1 to determine the smallest subset of biomarkers that would collectively identify the most common subtypes of colorectal polyps. Micron-sized iron oxide magnetic beads, pre-coated with protein L, were used in this study as proof-of-concept particles for in vitro and ex vivo binding experiments. MUC1+ (HT29-MTX and SW480) and MUC1-(CACO2) cell lines were incubated with microparticles (MPs) conjugated to MUC1 antibody and a fluorescent dye. Cell lines were imaged with a confocal microscope after an hour of incubation and a wash step. In a separate strategy, the colon tissue of human (hu) MUC1-expressing, Apcmin/+ mice with colorectal tumors, and Apcmin/+ mice with colorectal tumors that do not express huMUC1 were removed, cut longitudinally, and fixed with 4% (v/v) PFA for 12 hours. The tissue then was incubated with MUC1 targeting particles and imaged with a confocal microscope. In all experiments, normal mouse IgG antibody served as control. Results: The primary polyp tissue staining showed differential expression of the cancer epitope of MUC1 across different polyp subtypes. In cell line binding assay, MUC1 antibody-conjugated MPs bound specifically to MUC1+ HT29-MTX and SW480 cells. There was no binding or minimal binding of MUC1 functionalized MPs to MUC1-CACO2 cell lines (Fig 1.) . In fresh, fixed tissue, MUC1 functionalized MPs bound to huMUC1 colorectal tumors with minimal binding of MPs to normal adjacent tissue in both targeting and control groups. Conclusion: Protein L coated MPs can be functionalized with antibodies to target CRC surface markers in vitro. Ongoing work seeks to functionalize MPs with additional CRC targeting antibodies to specifically recognize and target heterogeneous lesion surfaces to achieve early, real-time, and potentially non-invasive, detection of pre-cancerous lesions. Acknowledgments: This study was supported by the UTHealth Innovation for Cancer Prevention Research Training Program Pre-doctoral Fellowship (Cancer Prevention and Research Institute of Texas grant # RP160015); Cancer Prevention and Research Institute of Texas (CPRIT; RP180164); US National Cancer Institute (NCI; U01 CA214263 and R21 CA185536), Institutional Research Grants and a Startup grant from MD Anderson Cancer Center.

Exploring the effects of SBRT on the pancreatic tumor microenvironment using intravital fluorescence microscopy Timothy Samuel, University Health Network, tim.samuel@mail.utoronto.ca

Abstract Body : Background: Despite decades of improvements in cytotoxic therapy, the current standard of care for locally advanced pancreatic cancer (LAPC) provides, on average, only a few months of survival benefit.1 Recently, Stereotactic Body Radiation Therapy (SBRT), a technique that accurately delivers high doses of radiation to tumors, has emerged as a promising therapy to improve local control of LAPC, however, its effects on the tumor microenvironment are still poorly understood.2 Our group and others have shown that high-dose radiotherapy can cause significant damage to tumor blood vessels.3,4 Thus, we hypothesize that this vascular damage may further lead to tumor hypoxia, which can potentially generate a more aggressive tumor phenotype. Exploring the effects of SBRT on the pancreatic tumor microenvironment (including hypoxia) can help guide the development of more effective treatments. Methods: NRG mice were orthotopically inoculated with a human pancreatic cancer cell line (BxPC3) engineered to constitutively express red fluorescent protein (RFP), as well as green fluorescent protein (GFP) under hypoxic conditions (driven by a 5x Hypoxia Response Element promoter). After five weeks of tumor growth, a glass imaging window was surgically implanted over the abdomen to provide optical access to the tumor. Mice were split into two groups (n=5/group), with one receiving SBRT (5x8Gy, given in daily fractions) using a smallanimal irradiator (X-Rad 225Cx), and the other group being left untreated. Intravital fluorescence microscopy (Zeiss LSM710) was performed before and (1, 4, 7 , and 14 days) after treatment to quantify changes in the pancreatic tumor microenvironment, including blood vessels and fibrillar collagen (using APC-CD31 and second harmonic generation, respectively). Immunohistochemical staining was also performed on tissues to validate in vivo findings. Results: Contrary to our hypothesis, a decrease in tumor hypoxia was observed following SBRT as compared to the control group. These results were validated immunohistochemically with CA9 staining. Reoxygenation of the tumor was observed with a decrease in both tumor growth and cell density in the SBRT group. No significant differences in tumor vascularization were detected between groups. Conclusions & Future Directions: Using intravital fluorescence microscopy, we were able to demonstrate a decrease in tumor hypoxia in pancreatic tumor xenografts following 5x8Gy SBRT. This reoxygenation of the tumor was likely due to reduced demand for oxygen as a result of significant cell death and growth arrest from treatment. Understanding how this reoxygenation phenomenon occurs in a dose-dependent manner will help optimize dose/fractionation schemes for clinical SBRT. Abstract Body : Introduction Non-alcoholic fatty liver disease (NAFLD) affects ~25% of the global population, and can progress to non-alcoholic steatohepatitis (NASH), cirrhosis and hepatocellular carcinoma. Diagnostic imaging of NAFLD through ultrasound and elastography is routinely performed clinically, but diagnosis is subjective and occurs late in pathogenesis. Invasive liver biopsy remains the gold standard for assessing NAFLD/NASH. Quantitative, noninvasive biomarkers that are sensitive to early changes in liver pathology are needed for diagnosis and patient management, with particular interest in the early inflammatory response. Perfluorocarbon (PFC) nanoemulsion, administered systemically is taken up by phagocytic cells of the reticuloendothelial system including resident liver Kupffer cells, and peripheral macrophages that traffic to inflammatory sites. 19F MRI can then be used to detect PFCs in vivo. We assess the impact of high fat diet (HFD) on 19F PFC uptake in the liver using MRI, compared with proton MRI based proton density fat fraction (PDFF) (1) and T1 relaxation time (2) . Methods NALFD was induced in 8-9-week-old female C57BL/6 mice HFD (40% fat, 2% cholesterol 20% fructose). Control animals received standard (Std) diet. MRI occurred at 6, 9 and 12 weeks post-diet start. A PFC imaging agent (VS-1000H, Celsense) was administered to mice two days before first imaging. Imaging was performed using an 11.7 T MRI system (Bruker) with a 1H/19F dual-tuned birdcage coil. A rapid relaxation enhancement (RARE) protocol was used to acquire 19F, and 1H anatomical images for co-registration and T1 maps. PDFF was quantified using a multi-gradient-echo (MGE) sequence. Analysis including fitting of MGE in/out of phase data (1) was accomplished in VivoQuant (Invicro) . Subsets of mice were sampled at each imaging time point for H&E and immunohistochemistry (IHC) staining. Results HFD mice show significant increases in liver volume compared to Std. diet mice (Fig. 1a) . Moreover, the total amount of liver PFC remains significantly higher in HFD mice compared to controls at all time points (80-157%, Fig. 1b) . When normalizing total liver PFC to tissue volume (Fig. 1c) , HFD mice show a significant reduction in PFC density compared to Std diet mice at 9 weeks (41%) and 12 weeks (57%). The liver PDFF of Std diet mice treated with vehicle ranges from 5-8% (Fig. 1d) . HFD mice show significantly higher PDFF at all time points (~36%). T1 relaxation times of Std diet mice treated with vehicle is ~2.35 s (Fig.1e) . HFD mice show significantly reduced T1 values compared to Std diet at all time points (56-59%). H&E and IHC confirmed significantly higher steatosis and inflammation in HFD versus Std diet mice (data not shown). Discussion and Conclusions Overall, the MRI-based in vivo biomarkers studied detected the impact of HFD on liver with statistical significance; the total 19F-uptake increased, the 19F-density decreased (due to increased liver volume), PDFF increased and 1H T1 decreased. These findings are consistent with significant increases in steatosis as detected by PDFF, T1 values, and histology relative to normal mice. We speculate that the increase in total 19F-uptake with HFD is due to influx of inflammatory macrophages. However, the degree to which Kupffer cell activation occurs, which may also affect PFC uptake, or change in density during disease course is unknown. Future experiments involving flow cytometry of liver tissues could resolve this question by separating Kupffer and peripheral cell populations that are PFC positive. Moreover, chronic administration of HFD resulted in marked disease in the liver of all animals prior to the start of imaging. Initiation of imaging earlier in the disease time course could enable detection of the onset of disease and provide a biomarker for protective effects of therapeutics prior to significant liver injury.

Last Name: Lister Email: drlister@health.ucsd.edu

Helene Chesnais, Penn Medicine, chesnais@sas.upenn.edu

Abstract Body : Introduction Type II diabetes mellitus has been known to be a significant risk factor for the development of cardiovascular disease (CVD). Amongst diabetics, CVD is the most prevalent cause for morbidity and mortality around the world. Studies have suggested that calcification is a key pathological feature of patients with diabetes and CVD. One such study found that patients with diabetic nephropathy had significantly greater prevalence of coronary artery calcification than normoalbuminuria diabetic patients. 18F-sodium fluoride (18F-NaF) positron emission tomography and computed tomography (PET/CT) is an imaging technique that has been used to investigate calcification activity in the human body. 18F-NaF PET/CT is also approved for the monitoring of bone metastases in prostate cancer. Research on the association between diabetes and prostate cancer suggests that patients with diabetes have a lower risk of developing prostate cancer. However, these studies have not investigated the effect of diabetes on aortic calcification in prostate cancer. Methods Our study included 101 males (ages 53-91) with prostate cancer who had undergone 18F-NaF PET/CT. The resulting mean (standard deviation) BMI was 29.1 (4.7) kg/m2. Mean tracer dose and time to scan were 7.8 (2.5) mCi minutes. Patients who had diabetes type II were identified from physician notes or lab values. An imaging processor (Fiji PET/CT Viewer) was used to manually segment each axial slice of the aorta from the top of the aortic arch to the end of the common iliac arteries, Figure 1 . The CT segmentation was overlaid onto the PET images to calculate the mean and maximum Standard Uptake Values (SUVmean and SUVmax) of the aorta. Aortic SUVmean and SUVmax were compared with age and BMI. Aortic SUVmean and SUVmax were then compared between patients with and without type II diabetes. In order to correct for the effect of BMI or dose on uptake, a multiple regression analysis was performed to test for the effect of BMI, dose, and diabetes on SUVmean and SUVmax. Results We identified 21 cases of type II diabetes within the 101 patients. 71 patients did not have a history of type II diagnosis, and 9 had insufficient information in their charts to determine their diabetic status. Diabetic patients had higher aortic SUVmean (p = 0.015) and SUVmax (p = 0.049) than non-diabetic patients, Figure 2 . SUVmean and SUVmax were correlated with BMI (r = 0.24, p = 0.0199 for mean; r = 0.25, p= 0.0129 for max) but not age. A diagnosis of type II diabetes was still correlated with higher SUVmean (p = 0.026) after correcting for BMI and tracer dose, whereas dose (p=0.006) and BMI (p=0.016) were stronger predictors of SUVmax than a diagnosis of diabetes. Conclusion Our results demonstrate a correlation between type II diabetes and increased calcification of the aorta as measured by higher 18F-NaF PET/CT aortic SUVmean in prostate cancer patients. This suggests that, while patients with diabetes have a lower risk of developing prostate cancer, patients who have both conditions may have a poorer cardiovascular profile than patients with prostate cancer and no diabetes. Future studies should elucidate the potential of 18F-NaF PET/CT to assess atherosclerosis in high-risk prostate cancer patients with diabetes. Abstract Body : Current methods of monitoring islet cell transplants for Type 1 diabetes (T1D) using MRI are limited by its low sensitivity and inability to differentiate signals between live and dead cells post-transplantation. Simultaneous PET/MRI has greater sensitivity and ability to visualize cell metabolism. However, this new dual modality tool currently faces two major challenges in its use for monitoring cell transplantation. Primarily, PET signal decay and spatiotemporal change in radioactivity hinder quantification of the transplant region and prediction of the transplanted cell number. Furthermore, selection bias from different radiologists renders human error in region of interest (ROI) selection. This calls for the development of artificial intelligence (AI) algorithms for the automated analysis of PET/MRI of cell transplantation. Here, we combined K-means++ for ROI segmentation with a Convolutional Neural Network (CNN) to predict radioactivity and cell number in the transplant ROI. To train the algorithm with in vitro data, 3D-printed phantoms were generated in varying shapes. BetaTC6 cell lines were sequentially labeled with VivoTrax (460ul of 5.5ug/ul stock solution per 25 cm² flask, 24 h @ 37 °C) and per 25 cm² flask, 1 h @ 37 °C) prior to imaging and dispersed in 1% agarose gel in the phantoms in various cell numbers. Radioactivity (uCi) measurements of the phantoms were recorded using a dose calibrator after labeling and before scanning. Static PET and T2Map MRI were acquired for all phantoms. Different k-values were tested for segmentation of in vitro data using the elbow method and raw PET volumes with radioactivity values were used to train the CNN on ground truth. The output from K-means++ on coregistered images used the segmented PET ROI as a mask to segment the MRI region. After scanning, inductively coupled plasma optical emission spectroscopy (ICP-OES) was used to determine actual iron content in the ROI of the phantoms. Iron staining was performed to confirm the presence of nanoparticles in the cells. Upon training the CNN to a minimum loss value and confirmation of dose prediction within 1μCi, the algorithm was programmed to estimate cell number by back-calculating to the initial timepoint labeling via F-18 half-life equation. In vivo data was then generated by transplanting the labeled betaTC6 cells under the left kidney capsule of mice (n=5). Image data and radioactivity values were acquired for all mice. The algorithm's performance was then tested on the in vivo data using intraclass correlation coefficient (ICC). Results from the elbow method showed that a k-value of 3 has the least sum-square errors (SSE) and was elected for PET image segmentation throughout the rest of the study (Fig. 1 A) . Dose calibrator measurements of cell radioactivity after labeling indicated a linear correlation between cell number and dose response (Fig. 1 B) . Initial training on ground truth indicates decreasing loss value (MSSE)in the activity range of 5 uCi to 60 uCi ( Fig. 1 C) . In vivo PET/MRI data indicates transplant location of cells with the overlay showing both MRI and PET signal from the cells under the left kidney capsule (Fig. 1 D) . Testing of the algorithm indicates an ICC score with a high degree of agreement with the dose calibrator value and actual transplanted cell number. Here, we observed the ability of combined machine learning and deep learning to segment a PET/MRI ROI and predict radioactivity and cell number in the image. This provides a novel tool for monitoring islet and other cell transplantation therapies through PET/MRI. It also unlocks a dynamic approach to automated, standardized segmentation and quantification of radioactivity of a ROI in simultaneous PET/MRI for radiolabeled nanodrug applications. Abstract Body : Introduction Despite the use of aggressive surgical, radiation and chemotherapies, glioblastomas remain highly lethal, incurable brain tumours with a mere 6.8% fiveyear survival rate.1 This poor prognosis is rooted in the diffuse infiltration of tumour cells beyond surgical margins, leading to incomplete tumour resection, inevitable tumour regrowth and patient death.2 There is consequently a dire clinical need for glioblastoma therapies that target tumour infiltrates; a need that can be addressed by porphryin theranostics. Porphyrins are multimodal organic probes whose preferential accumulation in tumour versus healthy tissue can be leveraged alongside their red/NIR fluorescence to 1) maximize tumour resection via fluorescence-guided surgery (FGS)3, 4 and 2) impart local, healthy tissue-sparing glioblastoma tumouricidal effects through photodynamic therapy (PDT). Despite these advantages, the clinical utility of porphyrins for glioblastoma treatment is limited by their poor transport across the intact blood-brain barrier (BBB) in infiltrative glioblastoma. Supramolecular porphyrin drug-delivery platforms, particularly nanoparticles and microbubbles, may address this limitation. Activelytargeted porphyrin nanoparticles (A-pNP) comprise surface ligands that target receptors at the BBB and tumour cells to facilitate glioblastoma-specific porphyrin delivery.5 This targeted delivery may be enhanced by microbubble-enabled MRI-guided focused ultrasound (MB-MRgFUS), an emerging platform that transiently and safely opens the BBB via ultrasoundtriggered microbubble oscillation within target vasculature. This focal BBB disruption has been shown to enhance drug delivery to glioblastoma tumours using conventional co-delivery of lipidmicrobubbles (MB) and drug agents, or via all-in-one drug-loaded microbubbles.6, 7 Collectively, A-pNPs, co-delivery MB-MRgFUS and all-in-one MB-MRgFUS strategies are thus promising avenues to explore for trans-BBB glioblastoma porphyrin delivery. Objective To compare the trans-BBB, glioblastoma-specific porphyrin delivery efficacy of three nano and micro-MRgFUS supramolecular platforms: 1) A-pNPs, 2) A-pNPs co-delivered with MB-MRgFUS, and 3) all-in-one porphyrin-loaded microbubbles (pMB-MRgFUS). Methodology A-pNPs, pMBs and porphyrin-devoid MBs were fabricated and characterized using literature protocols. 5, 8 MRgFUS was administered to healthy or 9L dual-tumour-bearing rats using an established feedback-controlled system9 (see Figure below for details). Results Similarly effective MRgFUS BBB disruption was achieved by pMBs and MBs in both healthy and tumour tissue as evaluated by gadolinium-T1w MRI. Despite yielding similar bulk vasculature effects, the supramolecular platforms had distinct porphyrin delivery efficacies. In healthy rats, A-pNPs were able to facilitate similar levels of trans-BBB porphyrin delivery as when combined with MB-MRgFUS. This delivery was approximately six and two-fold higher than that via pMBs in unsonicated and sonicated tissue respectively. The A-pNPs also outperformed the pMBs when combined with MB-MRgFUS in delivering porphyrin to infiltrative 9L glioblastoma tumours, whereby the conventional A-pNP/MB-MRgFUS co-delivery strategy yielded an 8-fold higher porphyrin tumour concentration and substantially higher fluorescence contrast at tumour sites compared to the all-in-one pMB/MRgFUS strategy. The porphyrin delivery efficacy of A-pNP/MB-MRgFUS could be amplified by increasing the A-pNP dose ten-fold while maintaining microbubble dose; dosing that was unfeasible for pMBs. Overall, co-delivery of A-pNP and MB-MRgFUS exhibited effective trans-BBB, glioblastoma-specific porphyrin delivery capabilities superior to A-pNP on its own or all-in-one pMB-MRgFUS. Impact This study overcomes the BBB as a primary barrier in delivering porphyrins to infiltrative glioblastoma via A-pNP/MB-MRgFUS, allowing the multitude of advantages of porphyrin-FGS and PDT to become clinically viable. Furthermore, given the increasing focus on using drug-loaded microbubbles within the field of MRgFUS, this study also has implications beyond porphyrin delivery. To the best of our knowledge this is the first in vivo orthotopic glioblastoma study to compare the trans-BBB, tumour-specific drug-delivery efficacy of drug-loaded microbubble-MRgFUS to conventional co-delivery MRgFUS. It highlights the advantages of the co-delivery strategy and demonstrates the limitation of drug dosing still faced by all-in-one microbubbles. This knowledge will invaluably inform more optimal design of microbubbles and MRgFUS for CNS applications.

References: 1. Q. T. Ostrom, G. Cioffi, H. Gittleman, N. Patil, K. Waite, C. Kruchko and J. S. Barnholtz-Sloan, Neuro Oncol., 2019 , 21, v1-v100. 2. A. Giese, R. Bjerkvig, M. E. Berens and M. Westphal, J. Clin. Oncol., 2003 , 21, 1624 -1636 . 3. H. Stepp, T. Beck, T. Pongratz, T. Meinel, F. W. Kreth, J. C. Tonn and W. Stummer, J. Environ. Pathol. Toxicol. Oncol., 2007 , 26, 157-164. 4. W. Stummer, U. Pichlmeier, T. Meinel, O. D. Wiestler, F. Zanella and H.-J. Reulen, The Lancet Oncology, 2006 , 7, 392-401. 5. M. A. Rajora, L. Ding, M. Valic, W. Jiang, M. Overchuk, J. Chen and G. Zheng, Chem. Sci., 2017 Zheng, Nat Nanotechnol, 2015, 10, 325-332. 9. N. McDannold, Y. Zhang, J. G. Supko, C. Power, T. Sun, C. Peng, N. Vykhodtseva, A. J. Golby and D. A. Reardon, Theranostics, 2019, 9, 6284-6299 .

Image/ Figure Caption: The three porphyrin supramolecular platforms studied (a) included actively-targeted porphyrin nanoparticles (A-pNP), porphyrin microbubbles (pMB) and codelivery of A-pNP and porphyrin-devoid microbubbles (MB). Porphyrin was incorporated into A-pNP and pMB as a lipid-conjugate within the particle shells. A-pNP/MB-enabled MRI-guided focused ultrasound (MB-MRgFUS) and pMB-MRgFUS displayed similar levels of BBB disruption (b) in healthy Wistar rats (2 sonication sites, indicated by yellow arrow) and 9L dual tumour-bearing Wistar rats (1 sonication site) as evident from (i) representative T1w MRI images post Gd contrast administration and (ii) associated quantification of Gd contrast enhancement relative to unsonicated regions. However, the strategies yielded differing degrees of porphyrin delivery (c) across the BBB to healthy brain parenchyma and 9L tumours as determined by tissue homogenate spectrofluorometry (i) and ex vivo Maestro fluorescence imaging (ii; yellow and pink circles represent unsonicated and sonicated tumours respectively). MRgFUS was conducted using a feedback controlled acoustic cavitation setup to ensure sonication was matched by cavitation thresholds (0.56 MHz, 10 ms burst length, 1000 ms period, 120 s duration, 0.5 emission target) after the administration of either saline (n=3) A-pNP + MB (n=4) or pMB (n=4) (0.35 mg/kg porphyrin, 3x109 MB/kg). The two microbubbles were size matched via centrifugation size isolation. Animals were euthanized via transcardial perfusion to remove intravascular porphyrin, and transvascular porphyrin delivery was assessed with ex vivo hyperspectral fluorescence imaging (Maestro, red filter, 1000 ms exposure) and quantified by spectrofluorometry of tissue homogenates. Abstract Body : Introduction 3D molecular contrast-enhanced ultrasound (3D mCEUS) provides a powerful quantitative assessment for the functional monitoring of tumor response to oncologic therapy.1-4 Parameters commonly used employ the differential targeted enhancement (dTE), requiring a high mechanical index (MI) ultrasound pulse.2 This limits the feasibility of clinical translation of the technique given evidence of bioeffects in tissues exposed to high-MI ultrasound.5 In this study, quantitative image features (texture and histogram-based features) are extracted from static 3D parametric maps of 3D mCEUS cines with and without contrast destruction. By comparing statistical properties of extracted image features and performance predicting response to anti-angiogenic therapy, we promote furthering clinical translation of this technique in cancer treatment monitoring applications by obviating the need for contrast destruction. Methods Mice bearing LS174T tumors were randomized to serial treatment (n=8; Bevacizumab 10 mg/kg days 0, 1, 3, 7) or control (n=8; saline). A separate cohort with LS174T tumors (n=18) was used to assess repeatability of measurements through repeated acquisition. 3D mCEUS cine images with vascular endothelial growth factor receptor 2 targeted contrast were obtained on days 0, 1, 3, and 10 using Philips EPIQ 7 with an X6-1 transducer. 3D parametric maps were generated from 4D cines using intensity projection methods and other parameters obtained through a sliding window method6 then stratified by breadth of sliding window and requirement of destruction pulse. 91 image features extracted from each parametric map for each mouse, and each scan day were selected for repeatability (intraclass correlation coefficient; ICC≥0.8) and detection of response to therapy (Wilcoxon signed-rank test; p Results/Discussion 54 of 91 image features were repeatable in at least one parametric map as assessed in the separate repeatbility cohort. The average intensity projection (AIP) map had the most repeatable features at 47. Logistic regression and linear determinant analysis showed highest performance (AUC 1.0, AUC 1.0 respectively) predicting response to therapy with only features extracted from before contrast destruction. After recursive feature elimination, only AIP features remained and several demonstrated strong linear correlations (Spearman ρ≥0.8) with dTE parameters. Several of these AIP image features also showed separation of measured values between treated and control animals at scan day 1. Generating dTE maps involves taking the difference of AIP maps and a post-destruction baseline intensity projection to calculate the bound fraction of molecular contrast. While parameters extracted from dTE maps are the common standard in the literature, unintended bioeffects of high-MI pulses5 and 5+ minute acquisition times limit feasibility of clinical translation. Conclusion Utilizing image features attuned to spatial heterogeneities characteristic of tumor vasculature, we show promising performance in early prediction of response to oncologic therapy without contrast destruction. Abstract Body : Purpose: Hyperpolarized (HP) 13C MRI is a novel metabolic imaging technique enabling real-time in vivo measurements of enzyme-catalyzed pyruvate metabolism being developed to address unmet clinical needs for prostate, brain, cardiac, and kidney imaging.1-4 Of primary interest are the conversions of pyruvate to lactate via lactate dehydrogenase (LDH), which is upregulated in cancers as well as animal models of inflammation and fatty liver disease,5 and of pyruvate to alanine via alanine aminotransferase (ALT). HP 13C MRI may provide valuable metabolic biomarkers for staging and assessing treatment response in these diseases. Human clinical research studies are underway at twelve sites around the world. However, whole-abdomen imaging with HP 13C maintains a challenge, mainly due to B0 and B1 inhomogeneities, respiratory motion, and the need for broad spatial coverage. In addition, there is little baseline data about HP 13C metabolite levels in healthy people. To investigate baseline energetics and the feasibility of whole-abdomen coverage, we studied HP 13C-pyruvate metabolism in multiple organs of healthy volunteers. Methods and Materials: After IRB approval, five healthy subjects underwent HP 13C MRI on a clinical 3 T scanner (MR750, GE Healthcare, Milwaukee, WI) with a volumetric transmitter and an 8channel receiver array (Clinical MR Solutions, Brookfield, WI) to increase spatial coverage. T1and T2-weighted images were acquired for anatomic registration. A volumetric multi-echo Dixon scan (IDEAL IQ) was used generate B0 field maps and correct off-centered metabolite frequencies due to in-plane B0 inhomogeneities. Localized shimming was used to correct for B0 inhomogeneities through-plane. After injection of 0.43 mL/kg of 250 mM HP [1-13C] pyruvate, signals from [1-13C] pyruvate and its downstream metabolites, [1-13C] lactate and [1-13C] alanine, were acquired using a multi-slice spectrally-selective echo planar imaging acquisition2 with flip angles of 30° for pyruvate, 60° for lactate and alanine. Total scan time was 1 minute with a 3 s temporal resolution, 2×2×2 cm3 voxels, and 7-14 slices extending from the diaphragm to the kidneys. Data were phase corrected, Fourier transformed, pre-whitened and coil-combined using pyruvate to estimate the coil weights.6 Metabolite signals were summed over the entire time course and ROIs were selected within the liver, both kidneys, the pancreas, and the spleen. Results and Discussion: Figure 1 (A) shows area-under-the-curve (AUC) metabolite maps (summed over the entire time course) overlaid on a T2-weighted anatomical reference from a representative healthy volunteer data set. Figure 1(B) shows the average metabolite signals computed from the ROIs of each abdominal organ (liver, kidneys, pancreas, and spleen) averaged across the five subjects with standard error bars. Absolute metabolite signals (pyruvate, lactate, and alanine) from the kidneys and spleen were 3-6 times higher than those from liver and pancreas, likely reflecting differences in overall perfusion. Figure  1 (C) shows the average metabolite ratios (lactate/pyruvate and alanine/pyruvate) for each organ averaged across the five subjects. Although absolute metabolite signals were lower, lactate/pyruvate and alanine/pyruvate ratios were both higher in the liver than in other abdominal organs (L/P: 0.38 vs. 0.17-0.21, A/P: 0.25 vs. 0.02-0.10), indicating higher production of these metabolites. Right and left kidney metabolism were symmetric as expected. The pancreas showed higher levels of alanine production relative to other organs, possibly due to high levels of protein synthesis. Conclusions: We successfully acquired HP 13C metabolic images of multiple organs throughout the abdomen with excellent spatial and temporal resolution. This broad spatial coverage will be valuable in developing novel applications of HP 13C MRI. Establishing normal values for metabolite ratios and metabolic conversion rates will serve as an important baseline for imaging cancer as well as metabolic liver diseases. Acknowledgements: This work was supported by NIH grants NIDDK 5R01DK115987 and P41 EB013598. Abstract Body : Introduction: Magnetic Particle Imaging (MPI) is an emerging tracer-based modality with promising applications in molecular imaging, in vivo cell tracking, and clinical angiography. Because its unique physics takes advantage of the electronic magnetization of superparamagnetic iron oxide particles (SPIOs), MPI has an excellent safety profile, superb linear "positive contrast", ideal tissue penetration, and a sensitivity of 200 cells [1, 2, 3] . However, the spatial resolution of MPI is currently limited to roughly 1 mm in high gradient (up to 7 T/m) preclinical scanners. Improved MPI resolution at human scale is essential for the clinical translation of MPI. Here we investigate a novel type of MPI tracer, which we call superferromagnetic nanoparticles (SFMIOs), that have experimentally yielded both a 10x improvement in resolution and a 100x improvement in SNR [4] . We hypothesize that when unchained SPIOs are in solution, we will observe signals comparable to typical SPIO magnetic responses. However, when the FWHM and SNR dramatically improve in the signal we observe, it should be indicative of SFMIO behavior: that is, interacting SPIOs being drawn to one another and increasing the magnetic field that each individual SPIO is exposed to. We test the hypothesis that "chaining" of strongly interacting SPIOs is essential for the dramatic improvement seen in our measurements by applying pre-polarization pulses to gauge their influence on the FWHM and SNR of our signal. An initial scan showing 100% SPIO behavior had a FWHM of 5.3 mT. The final scan with dominating SFMIO behavior had a FWHM of 1 mT. Materials and Methods: Single-core magnetite nanoparticles with an oleic acid coating were synthesized via thermal decomposition at UF and at UCB [5] . The estimated concentration of said particles is 12.6 mg/mL of Fe. The magnetization response of SFMIOs in toluene as a function of prepolarization time was measured using the UCB arbitrary wave relaxometer (AWR) [6] . Spatial resolution was quantified via the Houston Criteria by calculating the full-width-at-half-maximum (FWHM) of the PSF. The pulse sequence used was made of 10 ms pre-polarizing magnetic fields (DC, 12 mT amplitude). The particles were then excited by a linear change in amplitude (from 12 mT to -12 mT) that occurred over 50 us (20 kHz). A chain of these pre-polarizing pulses followed by 20kHz excitation was created for a total scan time of 600 ms. Results: By applying pre-polarizing (DC) magnetic pulses before excitation, we successfully measured the "chain formation time", the time point when individual SPIOs form chains, thus causing SFMIO behavior. We also show that the total received signal we measure can linearly be decomposed into two independent portions -one from the decaying population of unchained SPIOs, and the other from the growing population of chained SFMIOs. Conclusion: These observations strongly support the hypothesis that chaining of SPIOs is responsible for dramatic improvement in spatial resolution and SNR responsible for Super-Resolution MPI. individual SPIOs state to chains responsible for SFMIO behavior can be seen as a function of the length of a pre-polarizing pulse. (d,e) Received signal can be decomposed into two orthogonal signals that comprise the SFMIO and SPIO components. d) The SPIO component of our received signal drops as a function of pre-polarization time. The first scan shows 100% SPIO response with a FWHM of 5.3 mT.. e) As the pre-polarization time increases, the SPIO contribution drops and the SFMIO contribution increases with a final FWHM of 1 mT. f) An individual measured PSF and our reconstructed version from the two decomposed signals with a pre-polarization time of 30 ms. Good agreement between the two is seen. Note the small, sharper peak that occurs around ~17 us within the excitation scan -that is the beginning of SFMIO behavior within the sample. (a) 1.1 mg/mL Vivotrax, (b) Abstract Body : Magnetic Particle Imaging (MPI) is an emerging positive-contrast imaging modality that directly images superparamagnetic iron oxide (SPIO) tracers in the body [1] . Preclinical MPI has demonstrated micromolar sensitivity for tracking of stem and white blood cells, as well as in imaging of cancer, lung perfusion and abdominal bleeding [2] [3] [4] [5] . MPI resolution is determined by the ratio of the SPIO saturation strength and the gradient field strength [1, 6, 13] . MPI already has excellent sensitivity and ideal contrast, linearity and penetration depth. The only significant weakness is poor spatial resolution, now about 1.5 mm in a mouse [7, 8] . All MPI studies to date assume non-interacting SPIOs. Recently, Tay et al. experimented with interacting SPIOs to dramatically improve MPI resolution [9] . They showed experimentally that high local concentrations of SPIO tracers showed order-of-magnitude improvements in both SNR and resolution [9] . Several experiments support the hypothesis that the observed super-resolution behavior is due to the formation of SPIO particle chains [9, 10, 11] . We propose that these superresolution chains of SPIOs exhibit "superferromagnetism,'' since superparamagnetism does not exhibit remanence or coercivity [12] . Objective: We performed computational modeling of finite-length linear chains of interacting single-domain superparamagnetic particles. We show that a linear chain of ≥ 2 SPIOs couples via a positive-feedback mechanism and demonstrates magnetic hysteresis. The saturation remanence and coercivity of the multi-particle system depend on its physical parameters such as chain length, particle size, separation and magnetic susceptibility. We show that the image-blurring effects of coercivity dispersion from varying chain lengths can be mitigated, facilitating simple yet robust synthesis. Methods: We consider identical and independent SPIO chains, allowing ensemble averaging across the nth SPIO in each chain. Each nodal SPIO feels an identical applied field plus the local dipole field from neighboring SPIOs. Ensemble averaging across chains permits the use of a Langevin saturator model for each node in the chain. This formulation generates a system of transcendental equations that describe positive feedback, which is consistent with the observed hysteretic behavior. The system is iteratively solved until an energetically stable, self-consistent solution is obtained. The external applied field is stepped, akin to a DC magnetometry study, to obtain the complete magnetization hysteresis curve. Relevant system parameters can be easily altered to accomodate particles of varying physical and magnetic properties. Results: We observe hysteresis in the magnetization curve for chains of ≥ 2 particles. Marked improvements in SNR (~70-fold) and resolution (~30-fold) are evident from the derived MPI point spread function (PSF). Parametric sweeps indicate that the coercive threshold does not vary with chain length above a certain size, allowing for generous practical tolerances in superferromagnetic iron oxide (SFMIO) chain synthesis. The simulation shows that dimers also display hysteresis, paving the way for dimeric tracers with superior sensitivity and resolution that should have negligible magnetomotive drift, which could be a safety concern for longer chains. Simulated data shows the MPI harmonics necessary for MPI reconstruction [13] . We expect some resolution and SNR losses from particle relaxation. We can also incorporate coercivity dispersion, which will manifest as further PSF blurring. We find that coercivity is a monotonic function of SPIO shell thickness, vanishing for very thick shells as expected. Conclusion: We successfully simulated an ensemble of short-chains of SPIOs using a computational model of magnetic dipole-coupled Langevin saturators. The model accurately predicts hysteresis, and also predicts the experimentally measured 10-fold improvements in SNR and spatial resolution. These findings will be crucial for practical SFMIO synthesis and FDA safety. Indeed, the primary value of the simulation is to estimate uniformity constraints on chain length and shell coating thickness for SFMIO behavior.

width half-max (FWHM) resolution when moving from a single SPIO tracer to an SFMIO dimer; (D) Parametric sweeps indicate that chains larger than 7 particles show negligible coercive dispersion, mitigating the problem of resolution loss; (E) Coercivity as a function of center-tocenter particle separation for 40nm particles shows that as particle separation within the chain increases, the ensemble behavior reverts to the Langevin response of non-interacting SPIOs; (F) Frequency spectrum of signal response of an SPIO (left) to a unit-amplitude RF excitation at 1 kHz shows odd harmonics associated with nonlinearity, used to reconstruct the MPI signal. The response of the SFMIO dimer (right) to an identical excitation shows stronger nonlinear harmonics, demonstrating the improved SNR and confirming that MPI reconstruction is viable. (1) is a promising option for tumoral treatment, relying on the selective delivery of boron atoms to cancer cells, followed by the irradiation with a neutron beam of the diseased organ. The innovation of this study lies on the development and test of a nanosized theranostic agent, able to maximize the selective uptake of boron atoms in tumor cells and, at the same time, to quantify (by MRI) the in vivo boron biodistribution that is crucial to determine the optimal neutron irradiation time and to calculate the delivered radiation dose. Methods: In this study we used the theranostic compound AT101(2) (10 B enriched ligand-C-[N-(DOTAMA-C6)carbamoylmethyl]C′-palmitamidomethyl-ocarborane), a compound containing a Gd atom for imaging purposes, and 10 Boron atoms organized in a carborane structure for BNCT. AT101 was internalized into PLGA (Poly,(lacticco-glycolic acid) nanoparticles: these NPs, approved for clinical use, are coated with DSPE-PEG-2000 and produced with and w/o 50% of a PLGA conjugated with a polyhistidine chain(n=15). The NPs were dialyzed to remove the unbound AT101 and treated to remove unreacted compounds. These PLGA were incubated with mesothelioma (AB-22) and healthy mesothelium cell lines (MET-5α) for 24 h in a petri dish. The cells were collected and sonicated. A portion of each batch was mineralized to analyze Gd concentration with ICP-MS, the remaining portion is used to quantify proteins. Results: Preliminary studies on PLGA-poly-his nanoparticles were made using the compound Curcumin. After these preliminary steps, investigating the nanoparticle stability and compound retaining, PLGA NPs containing AT101 were characterized using Dynamic Light Scattering, showing an average size of 132±13 nm, with a Polydispersion index never superior to 0.15, suggesting the good quality and reproducibility of the preparations. Their reproducible Z-potential was also analysed. We observed no significant release of AT-101 from the NPs when incubated for 72 h with Hepes buffer at 37°C and different pH, suggesting good stability. Two different NPs were created, both containing AT-101 but with or w/o PLGA conjugated with poly-histidine, a feature that is, in theory, able to swell and alter the charge of the nanoparticle at different pH. They were incubated with AB-22 and MET-5α cell lines. We observed a significant increase in the nanoparticle uptake in the AB-22 cell line, compared with MET-5α, and interestingly, in AB-22 cells, the NP containing Poly-histidine has a double internalization rate compared to the one without the Poly-His feature. Moreover, both the cells lines showed no significant toxicity derived from the incubation with the nanoparticles. The evaluation of the corresponding contrast enhancement in the T1 weighted MR image is in progress. In vivo studies for BNCT mediated by the theranostic agent AT-101 conjugated with PLGA are also in progress. Conclusion: Our results show how stable PLGA nanoparticles with polyhistidine chains can be reliably produced, and how they can retain the theranostic agent AT-101. We showed how our nanoparticles can favourably enter tumoral cells, compared to healthy mesothelium. This therapy can potentially affect only tumour cells with a lethal dose of radiation, even in case of spreading and infiltrative cases. These promising insights on the possibility to selectively direct a theranostic dual agent directly into the tumoral cell, with the possibility to analyze the drug uptake in a certain organ using MRI, may be an important, versatile and new starting point for the future of BNCT technology and cancer theranostics. (BNCT) is a radiation science emerging as a hopeful tool in treating cancer, by selectively concentrating boron compounds in tumor cells and then irradiating them with an epithermal neutron beam. BNCT relies on the nuclear reaction that occurs when Boron-10, a stable isotope, is irradiated with low-energy thermal neutrons to yield α particles(Helium-4) and recoiling lithium-7 nuclei. In our study, we exploit a dual imaging/therapy theranostic agent, AT-101, containing Boron and Gadolinium, encapsulated in PLGA nanoparticles, to perform Magnetic Resonance Imaging of the organ of interest and selectively damage tumoral cells. (Image made with Biorender.com) Full Name of Abstract's 1st Author : Jacopo Sforzi Abstract Body : Glioblastoma (GBM) is a grade IV brain tumor and a deadly malignancy. Currently, surgery, radiotherapy, and chemotherapy are the standard line of treatment. Despite all improvements in GBM treatment, the survival rate of patients remains low, with a median of 14 months.1 One of the difficulties in the treatment of GBM patients is the delivery of chemotherapeutic drugs across the blood-brain barrier (BBB).2 As a result, the efficacy of GBM treatment is poor while acquiring resistance to chemotherapeutic drugs over time. Novel delivery solutions that improve drug dosage in GBM or supplementation therapy that sensitizes GBM to low dose chemotherapy can increase the survival rate of patients. Our goal was to find a combinatorial treatment option that improves the effectiveness of two chemotherapeutic drugs (temozolomide (TMZ) and doxorubicin (DOX)) in GBM. To this end, we established a methodology to select a panel of microRNAs (miRNA) that target genes with important therapeutic roles in GBM. MicroRNAs, the regulators of gene expressions, have been considered novel targets to reprogram GBM cells for effectively responding to chemotherapy. Since several mechanisms are involved in tumor cell survival and response to therapy, targeting multiple miRNAs regulating various cellular hallmarks of GBM was critical for achieving successful therapy. We defined four important therapeutic pathways to target: cell proliferation and apoptotic signaling, invasion and metastasis, cytokine signaling, and stemness. The abundance of the miRNAs and their target genes motivated us to use available online databases, including Gene Expression Omnibus (GEO) and The Cancer Genomic Atlas (TCGA), for a meta-analysis approach. We introduced a comprehensive method of therapeutic miRNA selection for sensitizing cancer cells to chemotherapy, and the assessment of these miRNAs in vitro in GBM cell lines and in vivo in mouse models by molecular imaging. The rationally identified panels were evaluated in GBM cells by delivering synthetic microRNAs using PLGA-PEG nanoparticles prior to treatment with TMZ and DOX drugs. A panel of five miRNAs (miR-218, miR-490, miR-138, miR-139, and antisense-miR-21) was identified as the most promising combination (Fig. 1a) . STAT3, ZEB1, SMAD7, CDK6, and TGIF2 were identified among the most important genes that are targeted by these miRNAs, and they cover four different therapeutic signaling pathways identified earlier (Fig. 1b) . Cell viability, apoptosis, and the expression of downstream genes were evaluated in cell culture using MTT, FACS, and western blot assays. Results of U87-MG cells treated with different combinations of miRNAs revealed that the panel of selected miRNAs significantly increased the apoptotic population compared to control conditions (~3 -fold higher). After optimizing the miRNA-mediated therapeutic evaluation in GBM cell lines of different phenotypes (U87-MG, T98g, and LN308), the cells were pretreated with the panel of miRNAs 24 hours before being treated with different concentrations of TMZ or DOX. TMZ is a pro-drug that became active upon addition to the medium with a neutral pH. It had higher efficiency when it was given in two consecutive doses to the cells that were pre-treated with the panel of miRNAs loaded in PLGA-PEG-NPs. DOX showed a dose-dependent inhibition of cell viability upon pre-treatment with the microRNAs. Our results supported the hypothesis that we can achieve significant apoptosis induction with lower doses of drugs (Fig. 1c) . Further evaluation in tumor xenografts of U87-MG GBM (expressing Fluc) in animals with the pre-treatment of miRNAs loaded PLGA-PEG-NPs by intravenous injections, and intraperitoneal doses of TMZ, resulted in a significant reduction in tumor volume by the group that received combinatorial treatment of miRNAs and TMZ after four cycles of treatment when compared to tumor growth in the group that received control nanoparticles (CNP) and TMZ, and miRNAs only treatment, p < 0 .05, (Fig. 1d) . (b) Target genes of each miRNAs in GBM were identified from Gene Expression Omnibus and The Cancer Genome Atlas and assigned to a specific therapeutic pathway. Genes are color-coded based on the therapeutic pathways: yellow is the proliferation and apoptotic signaling, red is invasion and metastasis, blue is cytokine signaling, and green is stemness including EMT. (c) FACS analysis showed a significant increase in the apoptosis percentages (red) after treatment of U87-MG cells with miRNAs than control nanoparticles (CNP). DOX is more effective in increasing apoptosis than TMZ, but it has difficulty in passing the blood-brain barrier. (d) Bioluminescence images of animals treated with control nanoparticles (CNP) and TMZ, miRNAs and saline or the combination of miRNAs and TMZ. Significant tumor shrinkage was observed in the group that was treated with the combination of miRNAs and TMZ after four cycles of therapy. Abstract Body : Background: Alzheimer's disease (AD) is a neurodegenerative disorder characterized by a gradual decline in neurocognitive function. The ability of sigma-1 receptor (σ1R) to regulate cognitive impairment, especially its effect of delaying memory loss, has attracted increasing attention by more and more researchers [1] . Whether PET targeting σ1R can apply for the early diagnosis of AD remains to be explored. We aimed to evaluate a novel fluorine-18 labelled (+)-TZ3108 that specifically targets the σ1R to quantify the expression level of σ1R in the brain of mouse model of AD. Methods: Male APP/PS1 doubletransgenic mice, expressed mutant human presenilin (Deltae9) and mouse amyloid precursor protein (APPswe) was assigned into transgenic group (Tg group, 5 month old, n = 3) for animal model of AD. Male age-matched C57BL/6 wild type mice was assigned into control group (Wt group, 5 month old, n =3). The fluorine-18 labelled (+)-TZ3108 was synthesized according previous published references [2, 3] . Brain microPET imaging of (+)- [18F] TZ3108 was performed on a nanoPET scanner (Mediso Inc., Hungary). Anesthesia was maintained at 0.75-2.0% during the whole imaging session and core temperature was maintained at 37ºC. A 10 min CT scan was conducted to confirm positioning of the animal, followed by a 5 min transmission scan for attenuation correction. Thereafter, a 30 min dynamic (10×3s; 3×10s; 4×60s; 6×150s; 2 ×300s) emission scan was performed after i.v. injection of (+)- [18F] TZ3108 in 10% ethanol saline solution. PET emission data were corrected by related attenuation and reconstructed in manufactured software. The brain MRI image of each mouse was collected through 9.4T microMRI scanner (BioSpec 94/20USR, Bruker Inc.), and MRI images were coregistered to the PET/CT images in Carimas software manually. All volume of interest (VOI) were drawn on the MRI image using Carimas (Finland PET Center). Three dimensional VOIs were transformed to the PET space and then overlaid on all reconstructed PET images to obtain TACs. Activity measures were standardized to the mice body weight, and injected dose of radioactivity yielded SUV. With the time activity curve (TAC) of left ventricular as the input function, the original TAC of vital brain regions was fitted by 2TCM model to obtain binding potential (BP). Results: The quantification of (+)- [18F] TZ3108 PET/MRI revealed that the expression of σ1R in Tg group (BPTgWB = 2.41±0.10, BPTgCB = 0.95±0.10, BPTgFC = 1.04±0.10 BPTgHC = 0.92±0.10, BPTgST = 1.36±0.10, BPTgBS = 0.67±0.10, BPTgTHA = 0.63±0.10) was significantly lower than the Wt group (BPWtWB = 3.56±0.10, BPWtCB = 1.30±0.10, BPWtFC = 3.74, BPWtHC = 3.76±0.10, BPWtST = 1.95, BPWtBS = 1.98±0.10, BPWtTHA = 0.79±0.10) for whole brain (WB), cerebellum (CB), frontal cortex (FC), hippocampus (HC), striatum (ST), brain stem (BS), thalamus (THA) respectively. (Figure 1 ) Conclusion: Associated to the cognitive impairment, and effect of delaying memory loss, the expression of σ1R in the brain of 5-month old AD transgenic mice was significantly reduced compared to the Wt controls from this preliminary study indicated that imaging σ1R via (+)- [18F] TZ3108 may be a potential biomarker for early diagnosis for AD. Abstract Body : Inflammation plays an essential role in the progression of many chronic diseases like hepatopathy, atherosclerosis and rheumatoid arthritis. Cytokines such as tumour necrosis factor-α (TNF-α), interleukin-6 (IL-6), cyclooxygenase-2 (COX-2) and inducible nitric oxide synthase (iNOS) play important roles in the regulation of the inflammation. In this study, modular two-in-one nano-cocktails were synthesised to provide treatment of inflammatory diseases and also enable tracking of their delivery to the disease sites. Chit-IOCO loaded with treatment module (cerium oxide nanoparticles) and imaging module (iron oxide nanoparticles) were synthesised by electrostatic self-assembly. Its MRI capability, anti-inflammatory and antifibrosis ability were investigated. Results demonstrated that Chit-IOCO significantly reduced the expression of ROS, TNF-α and COX-2 in the LPS-stimulated macrophages. Cytotoxicity studies showed that the nano-cocktails inhibited the proliferation of macrophages. Additionally, Chit-IOCO is a great MRI contrast agent in macrophages. It was possible to track the delivery of Chit-IOCO to the inflamed livers of CCl4-treated C57BL/6 mice, demonstrated by a shortened T2* relaxation time of the livers after injecting Chit-IOCO into mice. In vivo anti-inflammatory and blood tests demonstrated that Chit-IOCO reduced inflammation-related proteins (TNF-a, iNOS and Cox-2) and bilirubin in CCl4 treated C57BL/6. Histology images indicated that the nanococktails reduced fibrosis of CCl4-treated mouse liver and did not affect other organs of the mice. This is the first reported data on the anti-inflammation and anti-fibrosis efficacy of Chit-IOCO in C57BL/6 mouse liver inflammation model. Overall, Chit-IOCO nano-cocktails have shown great potential in MR imaging/detecting and treating/therapeutic capabilities for inflammatory diseases. Potentially, these nano-cocktails can be modified with antibodies or binding peptides, allowing these nanoparticles to target the inflammatory biomarkers for the diagnosis of diseases. [1] . Most ESCC imaging diagnoses can be performed by computed tomography and ultrasound but it is difficult to detect suspicious minor lesions. However accurate diagnosis of disease progression and lesion location is of great significance for early diagnosis and the determination of neoadjuvant therapy. PET/CT becomes more and more popular in assessing tumors and lymph nodes, being a quantitative parameter the net uptake (influx) rate constant (Ki) may be more sensitive and specific than the max standardized uptake value (SUVmax) in different situations [2, 3] . Methods: Thirty-two patients enrolled in this study were divided into two groups as N0 stage and non-N0 stage, who were examined by 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) dynamically. According to the CT imaging, primary tumors (PTs) and regional lymph nodes were collected to analysis. Respectively the SUVmax and Ki were statistic analyzed to identify primary tumors between N0 and non-N0 stages, as well as to distinguish suspicious malignant lymph nodes ( Representative baseline and TAK-418-blocked co-registered PET/MRI scans are shown in Figure 1 , along with the TAK-418 dose-occupancy curve in the cerebellum. From the repopulation studies, there were no detectible levels of TAK-418 in plasma at either 24 or 48 h. By 24 h occupancy in cerebellum was measured at 30% which dropped to 6% by 48 h. This represents a significant repopulation of the LSD1 enzyme, as a dose of 0.3 mg/kg TAK-418 occupied 91% of the available enzyme at 3 h. From the dosimetry studies it was observed that the elimination of [18F] MNI-1054 takes place via both hepatobiliary and urinary pathways. With the estimated whole-body effective doses scaled from NHP to human 0.0272 mSv/MBq and 0.0267 mSv/MBq for female and male respectively. Conclusions [18F] MNI-1054 showed good brain penetration and heterogenous distribution, which correlated with known LSD1 expression levels. It was found to bind irreversibly to LSD1, and determined to be a suitable PET tracer for determining the central LSD1 occupancy in NHP. There was also significant re-population of the LSD1 enzyme observed by 24 h, which by 48 h had almost returned to pre-block levels. The dosimetry profile of [18F] MNI-1054 is comparable to other 18F small molecule PET tracers, and the next step in its evaluation will be to conduct PET studies in humans. Abstract Body : Radiotherapy constitutes a treatment mainstay for patients with head-and-neck squamous cell carcinomas (HNSCC) and is often combined with concomitant chemotherapy for locally advanced cancers. However, the radiation resistance of individual tumors has been shown to strongly depend on the presence and dynamics of tumor-associated hypoxia. Tumor hypoxia can be longitudinally monitored by 18F-Fluoromisonidazole (18F-FMISO)-PET/CT. The aim of our study was to evaluate radiomics changes of 18F-FMISO-hypoxia imaging during chemoradiotherapy (CRT) as predictors for treatment outcome in head-and-neck squamous cell carcinoma (HNSCC) patients. HNSCC patients undergoing CRT were prospectively recruited. All obtained three FMISO-PET/CT scans at weeks 0, 2 and 5 during CRT (W0/W2/W5). Hypoxic subvolumes (HSV) within the tumor were segmented by a threshold of 1.4 times background uptake. Patients were classified based on variations during CRT of the HSV size (increasing, stable and decreasing) by the percentage relative volume deviation (∆rV ).

Variations in the location of HSV (geographically-static and geographically-dynamic) were also quantified by a new classification parameter (CP) accounting for the spatial overlapping. Additionally, 130 radiomic features (RF) were extracted from HSV at W0, and their variations during CRT were quantified by relative deviations (∆RF). The prediction of treatment outcome based on variation of FMISO-hypoxia quantification parameters was evaluated. To maximize generalizability in the prediction models, positive findings in outcome prediction were considered relevant if (i) they were statistically significant after Benjamin-Hochberg correction for multiple testing, (ii) their predictive value was observed for both time-points during CRT (W0-W2 and W0-W5) and (iii) the outcome prediction was confirmed in both cohorts (initial cohort of 25 patients and validation cohort of 15 patients). HV decreased in 64% of patients between W0 and W2 and in 80% between W0 and W5. Although an overall decrease of hypoxia for patients was observed in the sample, ∆rV did not shown predictive value for treatment outcome. CP distinguished earlier disease progression (geographically-dynamic) from later disease progression (geographically-static) in both time-points and confirmed for both cohorts. The texture feature low-grey-level-zone-emphasis predicted local recurrence with AUCW2 = 0.82 and AUCW5 = 0.81 in initial cohort and AUCW2 = 0.79 and AUCW5 = 0.80 in validation cohort. Radiomics analyses of FMISO-derived hypoxia dynamics and individual texture features were able to predict patient outcome during CRT for HNSCC. Abstract Body : Brown adipose tissue (BAT) plays an important role in thermogenesis and thermoregulation. BAT metabolism is activated by sympathetic stimuli mediated by beta3adrenergic receptors, triggered mainly by exposure to cold. Crotamine is a polypeptide isolated from the venom of the rattlesnake, that besides its antitumoral effect, has a recently described role on the differentiation and metabolic activation of BAT in rodents. Positron emission tomography with 18F-fluorodeoxyglucose (PET-FDG) allows the evaluation of BAT glycolytic activity. OBJECTIVE: to evaluate the effects on BAT metabolism of crotamine administration in rodents at basal conditions and after cold stimulus activation using PET-FDG. METHODS: PET-FDG studies were performed in a small -animal PET scanner (Triumph trimodality-Gamma Medica Ideas, california, USA), 45 min after intravenous administration of 18F-FDG in a group of 8 mice submitted to a cold stymulus at baseline, 10, and 21 days after daily oral administration of 10 μg of a synthetic crotamine (Group 1). PET-FDG ws performed after the same intervals in a control group also receiving crotamine but without cold stymulus (Group 2). Volumes of interest (VOI) were drawn over the interscapular fat on PET-FDG images, and metabolic activity was measured by SUVmax and SUVmean values in the VOI. RESULTS: Group 1 showed a significant reduction in uptake in BAT after 10 days of administration of crotamine when compared to the baseline study, both by analysis of medium SUV and SUVmax. There was no statistically significant difference in uptake after 21 days of treatment with crotamine compared to the baseline study, when the same parameters were analyzed (SUVmean: baseline 3.5+-0.5; T10 2;1+-0.7; T21 3.2+-0.8)(SUVmax baseline 5.2+-1.1;T10 3.2+-1.2; T21 5.3+-1.3). Group 2 (control) showed a significant increase in 18F-FDG uptake in BAT after 10 days of administration of crotamine when compared to the baseline study, both by analysis of SUVmean and SUVmax. There was no statistically significant difference in uptake after 21 days of treatment with crotamine compared to the baseline study when the same parameters were analyzed (SUVmean: baseline 1.2+-0.3; T10 1,6+-0.3; T21 1.2+-0.2)(SUVmax baseline 1.8+-0.5;T10 2.3+-0.4; T21 1.9+-0.5). PET-FDG with cold stimulus shows greater glycolytic activity than studies without stimulus at all intervals studied (baseline, 10 and 21 days with crotamine). CONCLUSION: This is the first study that has evaluated the effects of crotamine in BAT glycolitic metabolism in vivo. Crotamine induced a transient increase in the glycolytic activity of BAT and a transient reduction in the response to the cold stimulus after 10 days, with no significant change compared to baseline after 21 days in both groups.

Prostate-specific membrane antigen (PSMA)-targeted photodynamic therapy enhances the delivery of PSMA-targeted magnetic nanoparticles to PSMAexpressing prostate tumors Ethel Ngen, Johns Hopkins University School of Medicine, engen1@jhmi.edu

Abstract Body : Introduction: Magnetic nanoparticle-induced hyperthermia (MNP-IH) is a less morbid focal therapy that is currently being explored to treat localized, intermediate, and highrisk prostate cancers (PCs).1 MNPs (≥100 nm), have been shown to possess both thermal and magnetic properties ideal for magnetic resonance image (MRI)-guided MNP-IH.2 However, the delivery of both receptor-targeted and non-targeted nanoparticles to tumors, after intravenous administration is still a challenge, especially nanoparticles ≥ 100 nm.3 Current preclinical and clinical practices to deliver high doses of magnetic nanoparticles (MNPs) to tumors for MNP-IH, involve direct intra-tumoral injections.4,5 However, this is an invasive procedure that can cause treatment-related morbidity, and negatively impact patient quality of life after treatment.6 Additionally, given the multi-focal nature of PC lesions, direct tumor injections could miss remote or nearby foci, which could lead to disease recurrence after therapy. Consequently, there is a need for nanoparticle delivery systems capable of increasing tumorspecific accumulation of MNPs, after intravenous administration.7-9 Hypothesis: Enhanced vascular permeability plays an essential role in the delivery of nanoparticles to tumors, through the enhanced permeability and retention (EPR) effect. However, the EPR effect varies greatly within tumors and also between patients.3 The poor clinical performance of some targeted and non-targeted nanoparticles has been partially attributed to this variable EPR effect.3 Consequently, tumor re-engineering approaches designed to specifically enhance the tumor EPR effect and subsequently increase nanoparticle delivery, are currently being explored.10, 11 Prostate-specific membrane antigen (PSMA) is overexpressed on the epithelium of aggressive prostate cancers (PCs) and has been targeted for imaging and therapy. 8, 9 Objective: Here, we evaluated the feasibility of using low dose PSMA-targeted photodynamic therapy (PDT), as a tumor re-engineering approach, to enhance the EPR effect and subsequently increase the delivery of PSMA-targeted MNPs to human PSMA(+) PC tumor xenografts, in a preclinical mouse model. Method: Human PSMA(+) PC3 PIP tumor-bearing NSG mice (Group 1) were intravenously administered 3.3 nmol of a low-molecular-weight PSMA-targeted photosensitizer and treated with fluorescence image-guided PDT, 4 h after. Briefly, the mice were irradiated with NIR light from a 690 ± 20 nm light-emitting diode, at a total fluence of 200 J/cm2.12 The mice were then intravenously administered 50 mg/kg of a PSMA-targeted MNP, immediately after PDT and monitored with both fluorescence imaging and T2-weighted MRI 18 h, 42 h, and 66 h after that.8,9 While the PSMA-targeted photosensitizer was imaged using its 700 nm fluorescence signal; the PSMA-targeted MNP was imaged using its 800 nm fluorescence signal. Additionally, T2-weighted MRI was used to image the PSMA-targeted MNP core. Human PSMA(+) PC3 PIP tumor-bearing NSG mice, not pretreated with PDT and only intravenously administered the PSMA-targeted MNP, were used as primary negative controls (Group 2). Additional control groups were also used to investigate the specificity of the PSMA-targeted photosensitizer and the EPR enhancement effect. Results: An 8-fold increase in the delivery of the PSMA-targeted MNPs was detected using T2-W MRI in the PDT-pretreated PSMA(+) PC3 PIP tumors of the Group 1 mice, 42 h after PDT; compared to the untreated PSMA(+) PC3 PIP tumors of the control Group 2 mice. This finding was validated with fluorescence imaging which showed similar results. Additionally, T2-W MRIs revealed enhanced peripheral intra-tumoral delivery of the PSMA-targeted MNPs. That finding was in keeping with two-photon microscopy, which revealed higher vascular densities at the tumor periphery. Prostate Specific Membrane Antigen (PSMA) is known to be overexpressed in prostate cancer and has also be found to be significantly overexpressed on neovasculature of many solid human tumors, including breast tumor neovasculature and breast cancer cells2. Previously, we combined a potent Photo Dynamic Therapy (PDT), Pc413, and a highly selective PSMA inhibitor, PSMA-1, invented in our lab, to develop a theranostic agent that can selectively identify and destroy prostate cancer (PCa) both in vitro and in vivo to improve fluorescence image guided surgery (FIGS) for prostate cancer3. We recently demonstrated that PDT can be used as an effective adjuvant therapy after FIGS of prostate cancer resulting in reduced PSMs, lower tumor recurrence and extended animal survival following surgery. In this application we will utilize PSMA-Pc413, a theranostic agent, to determine if FIGS followed by PDT can improve the approach to breast cancer management. To confirm PSMA expression in breast cancer cells, adult, female BALB/c mice inoculated with 5x103 cells were allowed to grow until tumors reached ~100mm3 in volume. They then received 0.5mg/kg of PSMA-1-Pc413 through tail vein injection. Fluorescence images were taken at 24-hour post injection -previously determined peak time of PSMA-1-Pc413 -and the tumor was subsequently excised for histological analysis.To validate PSMA expression in the 4T1 model, orthotopic tumor tissue underwent immunofluorescence staining. Tumor tissues were co-incubated with PSMA-1-Alexa488 (synthesized in the lab) and goat anti-rabbit CD31-specific antibody overnight at 4oC. After washing, tissues were incubated with goat anti-rabbit AlexaFluor594 and analyzed11. We have demonstrated an orthotopic syngeneic mouse model of breast cancer where PSMA is overexpressed on the neovasculature of 4T1 breast cancers in BALB/c mice. Strong PSMA expression was observed in tumor tissues, which was blocked in the presence of excess amount of PSMA-1 ligand (Fig.1A) . PSMA expression was found to be co-localized with CD31, indicating PSMA's expression in tumor vasculature. When PSMA-1-Pc413 was given via IV to female BALB-C mice bearing orthotopic 4T1 tumor, selective accumulation in the tumor was observed (Fig.1B) . These data support the feasibility to use PSMA-1-Pc413 to highlight tumor for FIGS and PDT of breast cancer. This work will lay the foundation for development of this agent towards clinical BCS and has the potential to dramatically impact current surgical procedures. In the next stage, we will then utilize this optimized model to investigate the efficacy of FIGS followed by PDT. After tumor implantation into breast fat pads, we will inject the PSMA targeted Pc413 agent and perform FIGS followed by PDT therapy to the resection cavity. Mice will be assessed for survival, local recurrence, metastatic spread to the lungs and elicited immunological response, which, in the future, may allow combination therapy with immune modulators. These studies will result in the use PSMA-1-Pc413 as a theranostic molecular imaging probe provide image guidance for breast tumor resection and ablative therapy to cancer tissues missed during surgery.

Country: United States ID: GA391 Coupled plasmonic gold nanoparticle-coated nanobubbles -multimodal imaging agent for atherosclerosis Sourabh Mehta, IITB Monash research academy, sourabh.mehta@monash.edu

Abstract Body : Cardiovascular diseases (CVDs) contribute to 31% of global deaths, with atherosclerosis-related CVDs being the major contributor1. In high-risk CVDs, like atherosclerosis, real-time information about disease status plays a crucial role in designing treatment strategies. For the detection of atherosclerosis imaging technologies such as ultrasound, computer tomography (CT), and magnetic resonance imaging are approved diagnostic modalities. However, each has its own limitations; thus, multimodal imaging could be a potential solution2. We designed and developed gold nanoparticle (AuNP)-coated nanobubbles (AuNB) by assembling and growing gold nanoparticles on the lipidic nanobubble (NB) template. The synthesized AuNBs exhibit coupled plasmonic resonance at a wavelength of 630 nm, whereas free nanoparticles show plasmonic resonance at a wavelength of 520 nm. Further, the transmission electron microscopy images show AuNP coating on the NB surface to form spherical AuNB of 300-400 nm size, and the X-ray photoelectron spectroscopic (XPS) analysis confirms the presence of AuNP on the NB surface. Furthermore, we observed that the AuNPs did not leak from the AuNBs surface when incubated at 37 oC at 300 RPM for 72 hours. Next, we studied the echogenicity of AuNBs in various conditions such as depth, pH, and dilutions; the results indicated the AuNBs show contrast enhancement in tissue-mimicking phantom in vitro. Also, the AuNBs show contrast enhancement in CT imaging in vitro; thus, AuNBs can be used as a multimodal imaging agent (US/CT). We functionalized AuNB with Biotin for making AuNBs molecular imaging ready. Further, to test the proof of concept, we added Streptavidin-FITC. The absorbance measured at 488 nm indicated the AuNBs successful conjugation of AUNB-Biotin-Streptavidin-FITC. For molecular imaging of atherosclerosis, we tagged scFvanti-GPIIb/IIIa, which binds specifically to activated GP IIb/IIIa on activated platelets at the thrombus site or the atherothrombotic site3, 4. Flow cytometric analysis indicates the successful detection of activated platelets. Hence, the AuNBs can be used as a potential contrast agent for multimodal molecular imaging of thrombosis and atherosclerosis. -125) . Animals were observed daily for signs of paralysis and disease progression was scored on a 1-5 scale. On day 7 post-induction, mice received 250 μg of anti-CD19 mAb, MDX-1342, or control mAb intraperitoneally (n=9/group). On day 14, control and treated mice (n=4/group) were intravenously injected with 3.44±0.37 MBq [64Cu]hCD19-mAb and imaged using PET/CT 18-20h later; these animals were perfused post-scan and organs dissected to determine percent injected dose per gram (%ID/g). CNS tissues were further analyzed using ex vivo autoradiography. PET signal was quantified in the whole brain, cerebellum, pons and medulla using a brain atlas, and normalized to ex vivo blood signal. Finally, on day 16, CNS tissues and spleen were harvested for flow cytometry (n=5/group) to quantify B cells. Results:

[64Cu]hCD19-PET imaging enabled highly sensitive in vivo detection of B cell burden in different tissue compartments of EAE mice after MDX-1342 treatment. Although the disease severity of treated and control groups did not statistically differ (Fig. S1a) , CD19-PET results helped provide insights into why this therapy might not have worked. Notably, PET images showed a marked decrease in [64Cu]hCD19-mAb binding in the spleen (Fig 1a) , but a strikingly incomplete reduction in bone marrow and brain signal in the medulla, pons and cerebellum (Figs. 1b&S1b). Peripheral PET findings were mirrored by ex vivo gamma counting, showing significant signal reduction in spleen (97%, p 0.1 for all). Importantly, flow cytometry verified the dramatic depletion of CD19+, CD45R/B220+, and CD20+ B cell populations in the spleen of treated mice (p Conclusion: Taken together, our CD19-specific PET tracer accurately detected the incomplete depletion of B cells in the CNS and peripheral tissues of mice treated with an anti-B cell therapy. Accordingly, incomplete depletion of B cells may be the reason this particular therapy was not successful in treating these EAE mice. The ability to track CD19+ B cells in the whole body and CNS, both preclinically and clinically, will help provide a more complete picture of the in vivo B cell load during MS pathogenesis, and has the potential to greatly enhance understanding, diagnosis, and treatment of this disease. Such whole-body molecular information is critical to optimize the dose and timing of immunomodulatory therapies and ultimately improve clinical outcomes for MS patients.References: 1. Chen, D. et al. J. Immunol. 196, 1541 -1549 . hCD19-mAb in the whole brain in the mice given the anti-hCD19 mAb when compared to mice given a control mAb. PET analysis normalized to ex vivo blood signal. (C) Ex vivo gamma counting shows significantly lower signal in both the spleen and bone marrow in treated mice compared to control mice which mirrors the PET images. ***p-value ≤ 0.001, **p-value ≤ 0.01, **p-value ≤ 0.05, nsp-value > 0.05.

First Name: Samantha

Email: levines@stanford.edu