key: cord-0694970-9vu8orf0
authors: Sheridan, Steven D.; Thanos, Jessica M.; De Guzman, Rose M.; McCrea, Liam T.; Horng, Joy E.; Fu, Ting; Sellgren, Carl M.; Perlis, Roy H.; Edlow, Andrea G.
title: Umbilical cord blood-derived microglia-like cells to model COVID-19 exposure
date: 2021-03-19
journal: Transl Psychiatry
DOI: 10.1038/s41398-021-01287-w
sha: 2cb92c62a32a811e2dbcded8517f93c7519a4c82
doc_id: 694970
cord_uid: 9vu8orf0

Microglia, the resident brain immune cells, play a critical role in normal brain development, and are impacted by the intrauterine environment, including maternal immune activation and inflammatory exposures. The COVID-19 pandemic presents a potential developmental immune challenge to the fetal brain, in the setting of maternal SARS-CoV-2 infection with its attendant potential for cytokine production and, in severe cases, cytokine storming. There is currently no biomarker or model for in utero microglial priming and function that might aid in identifying the neonates and children most vulnerable to neurodevelopmental morbidity, as microglia remain inaccessible in fetal life and after birth. This study aimed to generate patient-derived microglial-like cell models unique to each neonate from reprogrammed umbilical cord blood mononuclear cells, adapting and extending a novel methodology previously validated for adult peripheral blood mononuclear cells. We demonstrate that umbilical cord blood mononuclear cells can be used to create microglial-like cell models morphologically and functionally similar to microglia observed in vivo. We illustrate the application of this approach by generating microglia from cells exposed and unexposed to maternal SARS-CoV-2 infection. Our ability to create personalized neonatal models of fetal brain immune programming enables non-invasive insights into fetal brain development and potential childhood neurodevelopmental vulnerabilities for a range of maternal exposures, including COVID-19.

Maternal immune activation can result from exposures ranging from metabolic conditions to stress and infection, with potential in utero consequences to the developing fetus [1] [2] [3] [4] [5] [6] . In particular, epidemiologic studies strongly suggest that maternal viral and bacterial infections in pregnancy may be associated with adverse neurodevelopmental outcomes in offspring, particularly autism spectrum disorders, schizophrenia, and cerebral palsy, but potentially including mood and anxiety disorders as well [1] [2] [3] [7] [8] [9] . For instance, individuals who were fetuses during the 1957 influenza pandemic had a significantly increased risk for being hospitalized for schizophrenia as an adult 10 . The magnitude of these effects varies, but their consistency is difficult to ignore. Although mechanisms underlying neurodevelopmental morbidity in offspring remain unclear, microglial priming toward a pro-inflammatory phenotype with consequent altered synaptic pruning has been suggested as a candidate mechanism [11] [12] [13] [14] [15] [16] [17] .

Microglia, brain-resident tissue macrophages, play a key role in normal neurodevelopment by modulating synaptic pruning, neurogenesis, phagocytosis of apoptotic cells, and regulation of synaptic plasticity [18] [19] [20] [21] . Fetal yolk sac-derived macrophages are the progenitors for the permanent pool of brain microglia throughout an individual's lifetime [22] [23] [24] [25] . As such, inappropriate fetal microglial priming ("trained immunity" 26 ) due to in utero immune activation may have lifelong neurodevelopmental consequences. The central role of mononuclear cells, including macrophages, in COVID-19 pathogenesis 27 suggests that the potential risk to exposed fetal microglia requires investigation.

We have previously developed and validated adult patient-specific models of microglia-mediated pruning by reprogramming induced microglial cells from human peripheral blood mononuclear cells (PBMCs), and assaying them with isolated synapses (synaptosomes) derived from neural cultures differentiated from induced pluripotent stem cells (iPSCs) 28, 29 . We demonstrated robust evidence of abnormalities in microglia and synaptosomes from individuals with schizophrenia, shown to be complement-dependent through a C3 receptor neutralizing antibody and rescued in a dose-responsive fashion with a small molecule probe 28 . Other groups have similarly applied in vitro synaptic pruning assays to provide insight into the pathogenesis of autism spectrum disorder and neurodegenerative disorders [30] [31] [32] .

In this study, we investigated whether our validated reprogramming methods for adult PBMCs could be successfully adapted and applied to umbilical cord bloodderived mononuclear cells (CB-MNCs) from both SARS-CoV-2 infected and uninfected pregnancies to create personalized models of fetal brain microglia. Such models could have a wide range of application in investigating effects of in utero exposure on neurodevelopment. To illustrate this application, we demonstrate successful induction of microglia-like cells (CB-iMGs) from CB-MNCs from both infected and uninfected pregnancies.

This study was approved by the Partners Institutional Review Board. Informed consent was obtained from all participants.

Umbilical cord blood was collected at the time of delivery into EDTA tubes for plasma and CB-MNC isolation. After spinning at 1000g for 10 min to separate plasma, samples were processed for CB-MNC isolation using a Ficoll density gradient 33 . Briefly, blood was transferred into a 50 mL conical tube and then diluted to 1:1 ratio with Hanks' balanced salt solution without calcium or magnesium (HBSS minus). This diluted blood was then gently layered on top of Ficoll at 2:1 ratio (two volumes of blood diluted with HBSS minus to one volume Ficoll). The conical tube was then centrifuged at 1000g for 30 min at room temperature with brake inactivated to allow layering of cellular components. The cloudy ring below the plasma and above the Ficoll (i.e. the CB-MNC layer) was collected and placed in a new 15 mL conical tube, with HBSS minus added to bring the volume to 15 mL. This tube was then centrifuged at 330g for 10 min with high brake. The supernatant was removed and the CB-MNC pellet was washed with HBSS minus and resuspended in 10 mL HBSS minus for counting. Cells were counted on a hemocytometer in a 1:10 dilution of trypan blue. Cells were frozen in freezing medium consisting of RPMI 1640 Medium with 1% penicillin-streptomycin, L-glutamine, 1% sodium pyruvate, 1% non-essential amino acids, 20% fetal bovine serum (FBS), and 10% DMSO at 5-10 million cells/ vial, placed in a chilled Mr. Frosty, then into −80°C. The following day, CB-MNC cryovials were transferred to liquid nitrogen for long-term storage. Isolated cryopreserved adult peripheral blood mononuclear cells were obtained from a single healthy control donor by Vitrologic (https://vitrologic.com) cat# MNC-300.

Derivation of induced microglia-like cells from CB-MNCs by direct cytokine reprogramming iMGs were derived using previously described methods 28, 29 while CB-iMGs were derived from CB-MNCs with modifications as noted. Briefly, cryopreserved PBMCs or CB-MNCs were rapidly thawed at 37°C, diluted into media consisting of RPMI 1640 with 10% FBS and 1% penicillin/ streptomycin. The cell suspension was centrifuged at 300g for 5 min at room temperature, with the brake off. After aspirating the supernatant, the cell pellet was resuspended in media, counted, and plated on Geltrex-coated 24-well plates at 1 × 10 6 cells per 0.5 mL per well. After cells were incubated at 37°C for 24 h, the media was completely replaced with RPMI 1640 with GlutaMAX, 1% penicillin-streptomycin, 100 ng/mL of human recombinant IL-34 (Peprotech), and 10 ng/mL of GM-CSF (Peprotech). Media was replaced on day 13 after the initial cytokine reprogramming, and realtime live cell imaging or immunocytochemistry was performed on day 14.

iPSC generation and neural differentiation for synaptosome isolation iPSCs were reprogrammed from fibroblasts and used to derive neural progenitor cells, which were differentiated into neural cultures, as previously described 28, 29 . In brief, adult human fibroblasts were reprogrammed to iPSCs using non-integrating synthetic RNA pluripotency factors, expanded and cryopreserved by Cellular Reprogramming, Inc. (https://www.cellular-reprogramming.com). To initiate neural progenitor induction, iPSCs were cultured feeder-free in E8 medium (Gibco) on Geltrex-coated sixwell plates and passaged using 50 mM EDTA and trituration with ROCK inhibitor (10 mM Thiazovivin; Stemgent). iPSCs were further purified using magneticactivated cell sorting with Tra-1-60 microbeads (Miltenyi Biotec) on LS columns as described by vendor. Neural progenitor cells (NPCs) were derived from these iPSCs using neurobasal medium (Thermo Fisher Scientific) with 1× Neural Induction Supplement (Thermo Fisher Scientific), expanded using a neural expansion medium, and purified by double sorting using MACS against CD271 and CD133. NPCs were immunostained for markers, including Nestin, SOX1, SOX2, and Pax-6. Validated NPCs were seeded for neural differentiation on Geltrexcoated T1000 five-layer cell culture flasks (Millipore Sigma # PFHYS1008) and grown in neuronal differentiation medium (Neurobasal media (Gibco # 21103049) supplemented with 1× each (N2 supplement (Stemcell Technologies SCT # 7156), B27 supplement without Vitamin A (Gibco # 12587010), non-essential amino acids (NEAA Gibco # 11140050), penn/strep), 1 μM ascorbic acid, 10 ng/mL BDNF and GDNF (Peprotech), and 1 μg/ mL mouse laminin (Sigma # L2020) for 8 weeks.

Synaptosome isolation by sucrose gradient was adapted for iPSC-derived differentiated neural cultures from previously described protocols [34] [35] [36] . First, media was aspirated from flasks and cells were washed or scraped with 1× gradient buffer (ice-cold 0.32 M sucrose, 600 mg/L Tris, 1 mM NaH 3 CO 3 , 1 mM EDTA, pH 7.4 with added HALT protease inhibitor-Thermo Fisher # 78442), homogenized using a dounce homogenizer and centrifuged at 700g for 10 min at 4°C. The pellet was resuspended in 10 mL of 1× gradient buffer (transferred to a 30 mL and centrifuged at 15,000g for 15 min at 4°C. The second pellet was resuspended in 1× gradient buffer and slowly added on top of a sucrose gradient with 1× gradient buffer containing 1.2 M (bottom) and 0.85 M (middle) sucrose layers. The gradient and cell mixture was centrifuged at 26,500 r.p.m. (~80,000g) for 2 h at 4°C, with the brake set to "slow" so as not to disrupt the final bands. The mixtures were handled carefully and bands were inspected to confirm successful fractionation. The synaptosome band (in between 0.85 and 1.2 M sucrose) was removed with a 5-mL syringe and 19Gx1 ½″ needle and centrifuged at 20,000g for 20 min at 4°C. The final pellet was resuspended in an appropriate volume of 1× gradient buffer with 1 mg/mL bovine serum albumin (BSA) with protease and phosphatase inhibitors, aliquoted and slowly frozen at −80°C. Protein concentration was measured by BCA, and synaptosomes were further analyzed by transmission electron microscopy for heterogeneity of size and morphology, and enrichment of presynaptic (synapsin, SNAP-25) and post-synaptic (PSD-95) markers was determined by western blot analysis as previously described 28, 29 .

Real-time live cell imaging of CB-iMGs was performed as previously described 28, 29 . Briefly, cells were imaged on the IncuCyte ZOOM live imaging system (Essen Biosciences) while incubated at 37°C with 5% CO 2 . Synaptosomes were sonicated and labeled with pHrodo Red SE (Thermo Fisher Scientific) and added to CB-iMGs at 15 µg total protein per well in 24-well plates. Phase contrast and red fluorescence channel images were taken at a resolution of 0.61 µm per pixel every 45 min for a total of 315 min. Images were exported as 16-bit grayscale files and analyzed using CellProfiler 37 

We adapted previously reported methods 28, 29 for generating iMGs from adult-derived PBMCs to reprogram umbilical CB-MNCs from neonates of SARS-CoV-2 negative (n = 4) and SARS-CoV-2 positive (n = 2) mothers, delivered between 7 June 2020 and 6 July 2020. After 2 weeks of cytokine exposure, analogous to PBMCderived iMGs, CB-iMGs displayed typical ramified microglial morphology (Fig. 1a, i) and stained positive for canonical microglial markers IBA1, CX3CR1, PU.1, and P2RY12 (Fig. 1a, ii-v) , demonstrating that iMGs can be generated from umbilical cord blood-derived CB-MNCs. These cells exhibit morphologic and features and markers of cell identity comparable to those we have demonstrated for adult blood-derived PBMCs (Fig. 1b, i-v) . Quantitation of immunopositive cells for microglial markers (Fig. 1c) and upregulation of canonical microglial genes (Fig. 1d) indicate comparable efficiency of iMG transdifferentiation by PBMCs and CB-MNCs.

We next characterized CB-iMG function in a model of synaptic pruning using highly purified isolated nerve terminals (synaptosomes), allowing quantitation of synaptic engulfment in vitro with greater signal-to-noise than intact neural cultures 28, 29 . Figure 2a illustrates the real-time imaging-based phagocytosis assay workflow with CB-iMG cultures upon addition of iPSC-derived neural culture purified synaptosomes labeled with pHrodo Red SE, a pH-sensitive dye that fluoresces upon localization to lysosomes post-engulfment. Engulfment of synaptosomes can be robustly quantified in real-time live imaging (Fig. 2b, c) as well as by endpoint confocal microscopy of fixed immunostained cells 28, 29 .

To characterize kinetics of phagocytosis over time, pHrodo-labeled synaptosomes were added to the CB-iMG cultures and engulfment was visualized every 45 min for approximately 5 h using real-time live fluorescence imaging (Fig. 3a) . Phagocytotic index was determined at each time point by measuring pHrodo area per cell (Fig. 3b) using CellProfiler as described in "Methods" and Supplementary Information. CB-iMGs demonstrate robust phagocytosis of synaptosomes, with a time course of synaptosome engulfment rising over time qualitatively similar to that observed with adult PBMCderived iMGs 28, 29 .

To illustrate the application of these models to study maternal exposures, we generated CB-iMGs from umbilical cord blood of neonates from mothers who tested positive for SARS-CoV-2 (n = 2). As observed above for CB-iMGs derived from SARS-CoV-2 unexposed pregnancies, these cells also display ramified and amoeboid morphology by phase contrast imaging (Fig. 4a) , express microglial-specific markers IBA1, CX3CR1, PU.1, and P2RY12 by immunostaining (Fig. 4b, i-iv) , and engulf pHrodo-labeled synaptosomes in our synaptic pruning assay as visualized by real-time live imaging (Fig. 5a-c) and quantification (Fig. 5d) . These results demonstrate the capacity to create phenotypically characteristic and functionally active CB-iMGs in a quantitative model of synaptic pruning to provide for further investigations into potential effects of maternal exposures such as SARS-CoV-2.

Our results demonstrate the successful creation of neonatal patient-specific models of microglia-mediated synaptic pruning via cellular reprogramming of neonatal cord blood mononuclear cells. These models should facilitate novel insights into fetal brain development in the setting of maternal exposures, including but not limited to SARS-CoV-2 infection. Both SARS-CoV-2exposed and -unexposed umbilical cord blood-derived microglia-like cells express canonical microglial markers IBA1, CX3CR1, PU.1, and P2RY12, and demonstrate a range of morphologies with varying degrees of ramification, potentially reflecting a range of activation states that can be perturbed in experimental systems. Importantly, the induced microglia phagocytose synaptosomes demonstrating that they can recapitulate this key function of microglia in the developing brain. This work suggests the potential for umbilical blood mononuclear cells to serve as a non-invasive, personalized biomarker of fetal brain microglial priming. To our knowledge, microglia have not previously been modeled from umbilical cord blood or used to predict neurodevelopmental vulnerability at a time when there is a window for intervention. These models provide the potential for quantifiable endpoints that can be used to assess microglial programming in the setting of various maternal exposures, and later can be used to test the efficacy of potential therapies to ameliorate in utero priming of microglia toward a proinflammatory phenotype 38 .

Previous studies have demonstrated the impact of various maternal exposures, including maternal stress, metabolic disorders, air pollution, and infections, on fetal brain development [2] [3] [4] 7 . Such exposures may lead to offspring neurodevelopmental morbidity via maternal immune activation [39] [40] [41] [42] [43] [44] [45] [46] , which results in aberrant microglial programming in the developing brain [12] [13] [14] [15] 18 . In turn, maternal immune activation models have pointed to aberrant differentiation of fetal microglia and dysregulation of cytokine networks as key mechanisms underlying abnormal fetal brain development, with microglia primed toward a pro-inflammatory phenotype and altered synaptic pruning implicated in offspring morbidity [11] [12] [13] [14] [15] . Given the extent of synapse formation and pruning that occurs in fetal and neonatal life, developmental microglial function represents a critical target for investigation.

We have previously demonstrated feasibility for the concept of using other monocyte populations to model microglial behavior using another monocyte type, fetal placental macrophages or Hofbauer cells, as a potential biologic surrogate for fetal microglial function in preclinical models of maternal obesity 16 , and have shown that maternal obesity-associated inflammation primes both fetal brain microglia and resident placental macrophages toward a highly correlated pro-inflammatory phenotype 16 . Personalized assays using more readily available cord blood mononuclear cells, as presented here, can be used to assess both baseline microglial function, and behavior in response to "second hit" inflammatory stimuli, which will be helpful in informing risk assessments for an individual fetus.

The COVID-19 pandemic, with its associated maternal immune activation and pro-inflammatory cytokinemediated physiology 47, 48 , may pose risk to the developing fetal brain. While current data suggest that vertical transmission of SARS-CoV-2 is relatively rare [49] [50] [51] , the profound immune activation observed in a subset of infected individuals suggests that, even if the virus itself does not cross placenta, the developing fetal brain may be impacted by maternal inflammation and altered cytokine expression during key developmental windows [52] [53] [54] . In this work, we suggest a model system that may be applied to investigate risk associated with the COVID-19 pandemic. Extending our work with adult peripheral blood mononuclear cells to cord blood-derived microglial Fig. 4 Phenotypic characterization of CB-iMGs derived from CB-MNCs exposed to maternal SARS-CoV-2 infection. a Phase contrast ramified morphology of patient-derived CB-iMGs. Scale bar: 30 μm. b Immunostained images of exposed CB-iMG cells with nuclei (Hoechst) and indicated microglial markers (i) IBA1, (ii) CX3CR1, (iii) P2RY12, and (iv) PU.1. Scale bar: 25 μm. 5 Functional characterization of synaptosome engulfment by CB-iMGs derived from CB-MNCs exposed to maternal SARS-CoV-2 infection. a Representative live real-time images in phase contrast/red fluorescence overlay mode showing cellular uptake and b red fluorescence channel alone of pHrodo (red)-labeled synaptosomes uptake at the end of the phagocytosis assay (315 min). Scale bar: 60 μm (boxes show the magnified view of engulfing CB-iMG). c Representative pHrodo (red)-labeled synaptosome engulfment in exposed CB-iMGs during live real-time imaging used for quantification. Scale bar: 20 μm. d Quantification of labeled synaptosome uptake by exposed CB-iMGs cells during live imaging. The phagocytic index represents the mean pHrodo+ area per iMG cell over N = 8 fields per well × 3 wells per line and N = 2 separate SARS-CoV-2exposed CB-IMG line derivations. Error bars represent SEM. models will allow for rapid, scalable models to investigate risk, yielding non-invasive, personalized assays of the impact of SARS-CoV-2 on fetal brain microglial priming and synaptic pruning function. This approach can complement more traditional approaches that will require large longitudinal cohort studies and may require years or even decades (in the case of schizophrenia, for example) to fully capture risk. The ability to detect priming of fetal brain microglia toward a pro-inflammatory phenotype extends beyond SARS-CoV-2 to include numerous other maternal infections in pregnancy, as well as the myriad maternal exposures that have been suggested to impact fetal microglial development.

In sum, we demonstrate the potential for umbilical cord blood mononuclear cells to serve as a non-invasive, personalized model of fetal brain microglial priming. These models provide the potential for quantifiable endpoints that can be used to assess microglial programming in the setting of various maternal exposures. We determined that CB-iMGs can recapitulate microglial characteristics and function in vitro, providing key insight into cells from the neonatal brain that are otherwise inaccessible at birth and throughout childhood. We illustrated their application to investigate effects of maternal infection, including SARS-CoV-2, on the developing brain 28 . Beyond characterizing any consequent abnormalities, the scalability of this approach may enable investigation of targeted therapeutic strategies to rescue such dysfunction.

Influence of infection during pregnancy on fetal development

Long-term risk of neuropsychiatric disease after exposure to infection in utero

The fetal origins of mental illness

Beyond infection-maternal immune activation by environmental factors, microglial development, and relevance for autism spectrum disorders

Inflammatory mechanisms in obesity

The impact of maternal obesity during pregnancy on offspring immunity

Infections and brain development

Contributions of maternal and fetal antiviral immunity in congenital disease

Maternal infection during pregnancy and autism spectrum disorders

Adult schizophrenia following prenatal exposure to an influenza epidemic

A novel model for prenatal brain damage. II. Long-term deficits in hippocampal cell number and hippocampaldependent behavior following neonatal GABAA receptor activation

Apoptosis and microglial activation in influenza encephalopathy

Controversies and prospects about microglia in maternal immune activationmodels for neurodevelopmental disorders

Prenatal infection leads to ASD-like behavior and altered synaptic pruning in the mouse offspring

Maternal immune activation-induced PPARgamma-dependent dysfunction of microglia associated with neurogenic impairment and aberrant postnatal behaviors in offspring

Placental macrophages, a window into fetal microglial function in maternal obesity

A starring role for microglia in brain sex differences

Microglia function during brain development, new insights from animal models

Synaptic pruning by microglia is necessary for normal brain development

Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner

Microglia shape adult hippocampal neurogenesis through apoptosis-coupled phagocytosis

Fate mapping analysis reveals that adult microglia derive from primitive macrophages

Tissue-resident macrophages originate from yolksac-derived erythro-myeloid progenitors

Origin of microglia: current concepts and past controversies

Development and homeostasis of "resident" myeloid cells: the case of the microglia

Microglial priming as trained immunity in the brain

Pathological inflammation in patients with COVID-19: a key role for monocytes and macrophages

Increased synapse elimination by microglia in schizophrenia patient-derived models of synaptic pruning

Patient-specific models of microglia-mediated engulfment of synapses and neural progenitors

CUB and Sushi Multiple Domains 1 (CSMD1) opposes the complement cascade in neural tissues

Progranulin deficiency promotes circuit-specific synaptic pruning by microglia via complement activation

Cytoplasmic-predominant Pten increases microglial activation and synaptic pruning in a murine model with autism-like phenotype

Isolation of lymphocytes, granulocytes and macrophages

The isolation of nerve endings from brain: an electron-microscopic study of cell fragments derived by homogenization and centrifugation

Method and validation of synaptosomal preparation for isolation of synaptic membrane proteins from rat brain

Comparison of simple sucrose and percoll based methodologies for synaptosome enrichment

CellProfiler 3.0: next-generation image processing for biology

SARS-CoV-2 detection in different respiratory sites: a systematic review and meta-analysis

Brain changes in a maternal immune activation model of neurodevelopmental brain disorders

Maternal immune activation and autism spectrum disorder: from rodents to nonhuman and human primates

Maternal immune activation by Poly I:C as a preclinical model for neurodevelopmental disorders: a focus on autism and schizophrenia

Maternal immune activation alters nonspatial information processing in the hippocampus of the adult offspring

Maternal immune activation yields offspring displaying mouse versions of the three core symptoms of autism

Maternal immune activation and neuropsychiatric illness: a translational research perspective

Maternal immune activation and related factors in the risk of offspring psychiatric disorders. Front

The risk for behavioural deficits is determined by the maternal immune response to prenatal immune challenge in a neurodevelopmental model

Longitudinal characteristics of lymphocyte responses and cytokine profiles in the peripheral blood of SARS-CoV-2 infected patients

COVID-19: consider cytokine storm syndromes and immunosuppression

Vertical transmission of coronavirus disease 2019: a systematic review and meta-analysis

Infant outcomes following maternal infection with SARS-CoV-2: first report from the PRIORITY Study

Transplacental transmission of SARS-CoV-2 infection

Cytokine induction in fetal rat brains and brain injury in neonatal rats after maternal lipopolysaccharide administration

The classical complement cascade mediates CNS synapse elimination

Prenatal exposure to maternal infection alters cytokine expression in the placenta, amniotic fluid, and fetal brain

We are grateful to Dana Cvrk, CNM; Muriel Schwinn, NP; Lydia Shook, MD; Adeline Boatin, MD, MPH; Robin Azevedo, RN; Laurel Gardner, RN; Suzanne Stanton, RN; Natalie Croul, BA; Nicola Young, BA; and Samantha Devane, BS for their assistance with recruitment and sample collection; to all members of the MGH Obstetric-Pediatric COVID-19 Biorepository Processing Team for their assistance with sample processing and storage, to Lael Yonker, MD and Alessio Fasano, MD for their partnership on the Obstetric-Pediatric biorepository; to Jon Li, MD and Xu Yu, MD for critical infrastructural and regulatory support; and to Anjali Kaimal, MD,MAS, and Jeffrey Ecker, MD, for assistance with study infrastructure and departmental support. Most importantly, we thank the participants for being part of the study. This work was supported by R01-MH120227 (to R.H.P.) and R01HD100022 and 3R01HD100022-02S2 (to A.G.E.).