key: cord-0292613-gfzg6976 authors: Leimanis-Laurens, M. L.; Gil, D.; Kampfschulte, A.; Krohn, C.; Prentice, E.; Sanfilippo, D.; Prokop, J. W.; Lydic, T.; Rajasekaran, S. title: The Feasibility of Studying Metabolites in PICU Multi-Organ Dysfunction Syndrome Patients Over an 8-day Course Using An Untargeted Approach date: 2020-12-07 journal: nan DOI: 10.1101/2020.12.04.20244053 sha: acfa3b846282200a5692afc5d7d5f2e28e00bbb9 doc_id: 292613 cord_uid: gfzg6976 Metabolites are generated from critical biological functions and metabolism. This pediatric study reviewed plasma metabolites in patients suffering from multi-organ dysfunction syndrome (MODS) in the pediatric intensive care unit (PICU) using an untargeted metabolomics approach. Patients meeting criteria for MODS were screened for eligibility and consented (n=24), and blood samples were collected at baseline, 72 hours, and 8 days; control patients (n=4), were presenting for routine sedation in an outpatient setting. A sub-set of MODS patients (n=8) required additional support with veno-atrial extracorporeal membrane oxygenation (VA-ECMO) therapy. Metabolites from thawed blood plasma were determined from ion pairing reversed-phase LC-MS analysis. Chromatographic peak alignment, identification, relative quantitation, statistical and bioinformatics evaluation were performed using MAVEN and MetaboAnalyst 4.0. Metabolite analysis revealed 115 peaks per sample. From the PLS-DA with VIP scores above 2.0, 7 dynamic metabolites emerged over the 3 time points: tauro-chenodeoxycholic acid (TCDCA), hexose, p-hydroxybenzoate, hydroxyphenylacetic acid (HPLA), 2_3-dihydroxybenzoic acid, 2-keto-isovalerate, and deoxyribose phosphate. After Bonferonni adjustment for repeated measures hexose and p-hydroxybenzoate were significant at one time point, or more. Kendalls tau-b test was used for internal validation of creatinine. Metabolites may be benign or significant in describing a patients pathophysiology and require operator interpretation. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 7, 2020. ; https://doi.org/10.1101/2020.12.04.20244053 doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. 2 of 10 of metabolic pathways, are growing in appeal medically over the last decade for their potential in 47 disease characterization, drug discovery and precision medicine [8, 9] . We have previously described the current cohort of patients for patient whole blood 49 transcriptomics [10, 11] , and plasma lipidome [12] . This has revealed a complex biology in a 50 heterogenous patient population with a non-uniform patient response to treatments over an 8-day 51 course (stabilization and recovery phases) of illness during a PICU admission. Complimentary to 52 these previously reported analytic modalities from whole blood [10] [11] [12] , the aim of this current report 53 was two-fold: 1) to characterize total blood plasma metabolites (polar, charged) using an untargeted 54 approach, 2) to determine change in metabolites over an 8-day PICU course. There is a gap in our 55 understanding of the complex interaction between pediatric critical illness, specifically multi-organ 56 dysfunction syndrome (MODS) [13] (affecting twenty percent of PICU admissions [14] , resulting in 57 ten times the mortality rate [15] ), and their respective blood metabolites. After IRB approval, a short-term longitudinal design was adopted at Helen DeVos Children's 61 Hospital (2016-062-SH/HDVCH). Samples were collected under the protocol and study design [10-62 12] in a quaternary-care, urban, pediatric hospital in Western, Michigan. In brief, patients who were 63 identified as having MODS were enrolled, 24 in total, with an additional 4 sedation-control patients. These 24 patients were then further classified as needing veno-arterial extracorporeal membrane 65 oxygenation (VA-ECMO) as a therapeutic modality (n=8) according to Extracorporeal Life Support 66 Organization (ELSO) criteria [16] . Blood samples from the patients were obtained and placed into 67 EDTA-filled tubes, plasma was processed and stored at -80 o C for later use. All samples had 68 undergone one freeze-thaw before processing and analysis. Plasma samples (~50 microliters) were subjected to biphasic extraction using 73 chloroform/methanol/water as described previously [17] to remove nonpolar matrix interferences 74 and recover polar metabolites in the aqueous extraction phase. Stable isotope labeled (D 4 )-succinate 75 was added to plasma during extraction for use in estimation of metabolite recovery and for relative 76 quantitation across experimental groups. Samples were filtered through 0.2 micron syringe filters 77 (Fisher Scientific) and reconstituted in 100 microliters of 50 % methanol for use in ion pairing 78 reversed-phase LC-MS analysis. Targeted polar metabolite identification utilized a Thermo Scientific model TSQ Vantage triple 80 quadrupole mass spectrometer operating in negative ion mode. The mass spectrometer was 81 coupled to a Shimadzu Prominence HPLC with thermostated column oven and autosampler. Ten Detection parameters for each precursor/product ion pair of interest have been optimized based 88 using commercially available standards. LC-MS data analysis of chromatographic peak alignment, compound identification, relative 93 quantitation, and statistical evaluation across experimental groups will be performed using MAVEN 94 software [19] . Only relative quantitation of analytes against a selected internal standard was 95 performed for comparison of values across experimental treatment groups. "Absolute" quantitation 96 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in Figure 1 ). Metabolic profiles for sedation-controls were compared to MODS or ECMO patients, quantified 100 as percent of total. Metabolites with >30% of cases with zero values were excluded from further 101 analysis; consequently 66 metabolites were analyzed over the three time points (Figure 1 ). Using Analytical flow chart is presented in Figure 1 . Percent total of metabolites and change over time All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 7, 2020. ; https://doi.org/10.1101/2020.12.04.20244053 doi: medRxiv preprint (also a 6-carbon sugar-data not shown), and this is closely monitored at the PICU bedside, given blood 170 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 7, 2020. Furthermore, it was of interest to determine whether any of those metabolites identified by PLS- DA with high VIP scores where statistically significant over time, as this may provide additional 194 understanding and potential biomarker identification of this cohort of untargeted metabolites. When comparing to sedation-controls and correcting for Bonferonni adjustment for repeated 196 measures (P-value <0.008), both hexose and p-hydroxybenzoate were significant at, at least one time 197 point (Table 1) . All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 7, 2020. ; https://doi.org/10.1101/2020.12.04.20244053 doi: medRxiv preprint All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 7, 2020. ; https://doi.org/10.1101/2020.12.04.20244053 doi: medRxiv preprint 7 of 10 acceptable to use in this case. Spearman's Rho measures the rank correlation (how the ranks of the 229 x and y values align), and Kendall's Tau measures the percent of concordant pairs, which is also 230 based on ranks but considered to be a more robust measure. Creatinine was found to be high at 231 baseline, which correlated with clinical creatinine values. and is of ongoing concern given reports of higher mortality and multi-organ injury [33] . HPLA, a 268 phenylcarboxylic acid, has been speculated to be a marker of sepsis in adult cardiac surgery, however This indicates that the metabolic profile may illustrate a different landscape on patient recovery, and 276 metabolites may be organ specific. Limitations of the work include a low sample volume, capturing high-abundance metabolites, 278 and sample integrity may have been compromised by a previous freeze-thaw cycle. A second cohort 279 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in Dietary Intake Influences Metabolites in Infants: A Scoping Review Proteomic 315 and Metabolomic Characterization of COVID-19 Patient Sera Virus-induced genetics revealed by multidimensional precision medicine 338 transcriptional workflow applicable to COVID-19 Gene expression signatures identify paediatric patients with multiple organ 342 dysfunction who require advanced life support in the intensive care unit Pediatric Multi-Organ Dysfunction Syndrome: Analysis by an Untargeted Shotgun 346 Lipidomic Approach Reveals Low-abundance Plasma Phospholipids and Dynamic Recovery Over Center Observational Study. medRxiv 2020 Epidemiology of sepsis and multiple organ 350 dysfunction syndrome in children Day 1 multiple organ 352 dysfunction syndrome is associated with poor functional outcome and mortality in the pediatric 353 intensive care unit Data Analysis, I.; Pediatric Acute Lung, I.; Sepsis Investigators, N. Outcomes of Day 1 Multiple Organ 356 Dysfunction Syndrome in the PICU Physiology of Extracorporeal Life Support (ECLS) Extracorporeal Life Support for Adults Global analysis of retina lipids by complementary precursor ion and 361 neutral loss mode tandem mass spectrometry LC-MS data processing with MAVEN: a metabolomic 368 analysis and visualization engine Using MetaboAnalyst 4.0 for Comprehensive and Integrative 371 The pls Package: Principal Component and Partial Least Squares Regression 373 in R Classication and Regression Training Large-Scale Human Metabolomics Studies: A Strategy for Data (Pre-) Processing Validation Box and whisker plots for local climate datasets interpretation and creation using Excel 379 Hyperglycemia in critically ill children MetaboAnalyst 4.0: 384 towards more transparent and integrative metabolomics analysis Metabolomics 387 analysis reveals large effects of gut microflora on mammalian blood metabolites Bioconversion of toluene to p-hydroxybenzoate . via the construction and 390 characterization of a recombinant Pseudomonas putida Bioproduction of p-hydroxybenzoate from 393 renewable feedstock by solvent-tolerant Pseudomonas putida S12 The Kolbe-Schmitt Reaction Microbial synthesis of p-hydroxybenzoic acid from glucose Acid Synthesis in Mycobacterium tuberculosis Association of Blood Glucose Control and Outcomes in Patients with COVID-19 and Pre-existing Type 404 2 Diabetes Are phenylcarboxylic acids really markers in severe 406 sepsis? Critical Care All rights reserved. No reuse allowed without permission.perpetuity.preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint inThe copyright holder for this this version posted December 7, 2020. All rights reserved. No reuse allowed without permission.perpetuity.preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint inThe copyright holder for this this version posted December 7, 2020. ; https://doi.org/10. 1101