key: cord-1053069-yd8p8wpk authors: Gustafson, D.; Ngai, M.; Wu, R.; Hou, H.; Schoffel, A.; Erice, C.; Mandla, S.; Billia, F.; Wilson, M. D.; Radisic, M.; Fan, E.; Trahtemberg, U.; Baker, A.; McIntosh, C.; Fan, C.-P. S.; dos Santos, C. C.; Kain, K. C.; Hanneman, K.; Thavendiranathan, P.; Fish, J. E.; Howe, K. L. title: Cardiovascular Signatures of COVID-19 Predict Mortality and Identify Barrier Stabilizing Therapies date: 2022-02-09 journal: nan DOI: 10.1101/2022.02.08.22270636 sha: 2308c47916ca6282c10b17a86db2d0a2da2a597b doc_id: 1053069 cord_uid: yd8p8wpk Background: Endothelial cell (EC) activation, endotheliitis, vascular permeability, and thrombosis have been observed in patients with severe COVID-19, indicating that the vasculature is affected during the acute stages of SARS-CoV-2 infection. It remains unknown whether circulating vascular markers are sufficient to predict clinical outcomes, are unique to COVID-19, and if vascular permeability can be therapeutically targeted. Methods: Evaluating the prevalence of circulating inflammatory, cardiac and EC activation markers, and the development of a microRNA atlas in 241 patients with suspected SARS-CoV-2 infection, allowed their prognostic value to be assessed by a Random Forest model machine learning approach. Subsequent ex vivo experiments assessed EC permeability responses to patient plasma and were used to uncover modulated gene regulatory networks from which rational therapeutic design was inferred. Findings: Multiple inflammatory and EC activation biomarkers were associated with mortality in COVID-19 patients and in severity-matched SARS-CoV-2-negative patients, while dysregulation of specific microRNAs at presentation was specific for poor COVID-19-related outcomes and revealed disease-relevant pathways. Integrating the datasets using a machine learning approach further enhanced clinical risk prediction for in-hospital mortality. Exposure of ECs to COVID-19 patient plasma resulted in severity-specific gene expression responses and EC barrier dysfunction which was ameliorated using angiopoietin-1 mimetic or recombinant Slit2-N. Interpretation: Integration of multi-omics data identified microRNA and vascular biomarkers prognostic of in-hospital mortality in COVID-19 patients and revealed that vascular stabilizing therapies should be explored as a treatment for endothelial dysfunction in COVID-19, and other severe diseases where endothelial dysfunction has a central role in pathogenesis. plasma and were used to uncover modulated gene regulatory networks from which rational 13 therapeutic design was inferred. 14 15 Findings: Multiple inflammatory and EC activation biomarkers were associated with mortality in 16 COVID-19 patients and in severity-matched SARS-CoV-2-negative patients, while dysregulation 17 of specific microRNAs at presentation was specific for poor COVID-19-related outcomes and 18 revealed disease-relevant pathways. Integrating the datasets using a machine learning approach 19 further enhanced clinical risk prediction for in-hospital mortality. Exposure of ECs to COVID-19 20 patient plasma resulted in severity-specific gene expression responses and EC barrier dysfunction 21 which was ameliorated using angiopoietin-1 mimetic or recombinant Slit2-N. 22 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint Interpretation: Integration of multi-omics data identified microRNA and vascular biomarkers 1 prognostic of in-hospital mortality in COVID-19 patients and revealed that vascular stabilizing 2 therapies should be explored as a treatment for endothelial dysfunction in COVID-19, and other 3 severe diseases where endothelial dysfunction has a central role in pathogenesis. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint Evidence before this study 2 While diagnostic testing has allowed for the rapid identification of COVID-19 cases, the lack of 3 post-diagnosis risk assessment metrics, especially among the highest-risk subgroups, thereby 4 undermined the cascade and allocation of care. To date, the integration of clinical data with broad 5 omics technologies has opened up new avenues for efficiently delineating complex patient 6 phenotypes and their associations with clinical outcomes, with circulating profiles of plasma 7 microRNAs (miRNA), in particular, having been shown to be tightly associated with disease, and 8 capable of providing not only detailed prognostic information but also mechanistic insight. 9 10 Added value of this study 11 Markers of endothelial dysfunction at presentation, while indicative of poor outcomes in COVID-12 19 -positive patients, likely reflect systemic vascular dysfunction in critically ill patients and are 13 not specific to SARS-CoV-2 infection. More so, the generation of a plasma microRNA atlas 14 uncovers COVID-19-specific prognostic markers and multiple disease-specific pathways of 15 interest, including endothelial barrier dysfunction. Furthermore, synthesis of electronic health 16 record data with clinically relevant multi-omic datasets using a machine learning approach 17 provides substantially better metrics by which mortality can be estimated in patients with severe 18 COVID-19. Finally, targeted stabilization of the endothelial barrier with Q-Peptide and Slit2- N 19 are novel therapeutic avenues that should be explored in COVID-19 patients. 20 21 Implications of all the available evidence 22 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint Together, our work provides biological insight into the role of the endothelium in SARS-CoV-2 1 infection, the importance of miRNA as disease-and pathway-specific biomarkers, and the exciting 2 possibility that endothelial barrier stabilizing treatments might hold promise in COVID- 19 . is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint Vascular endothelial-cadherin . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. have high rates of morbidity and mortality.(3) Substantially higher rates of morbidity and mortality 7 have been observed as variants of concern become more predominant.(4) While diagnostic testing 8 has allowed for the rapid identification of COVID-19 cases, the lack of post-diagnosis risk 9 assessment metrics, especially among the highest-risk subgroups, have undermined the cascade 10 and allocation of care. 11 12 It has been proposed that the use of existing cardiovascular and respiratory parameters could serve 13 as a metric of risk prediction.(5, 6) However, case-fatality rates of those with comorbidities remain 14 particularly high (e.g., preexisting CVD at ~10.5%) with cardiorespiratory the aforementioned 15 markers having limited utility.(7) In this regard, while standard metrics including measures of 16 cardiac damage (e.g., troponin values above the 99 th percentile), the extent of inflammatory 17 activation (e.g., C-reactive protein expression), and cardiovascular imaging have elucidated the 18 spectrum of COVID-19 complications, they have only modestly elucidated the risk of adverse in-19 hospital outcomes and often provide limited insight into disease mechanism.(8-10) Albeit 20 encouraging, current COVID-19 therapeutics focus on stemming aberrant immune responses and 21 controlling viral reproduction (e.g., tocilizumab and remdesevir, respectively), which may neglect 22 key elements of the host response contributing to outcomes.(11) From this perspective, clinical 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. data and autopsy studies revealing endotheliitis and thrombosis have raised the possibility that 1 endothelial dysfunction, particularly fluctuations in vascular integrity and coagulative capacity, 2 could be a driver of clinical outcomes. (12) Thrombosis, fluid extravasation, and microangiopathy 3 observed in the small vessels and capillaries of the lungs directly support the notion that an intense 4 vascular reaction takes place in those with severe disease.(13) 5 6 Early data suggested that viral tropism towards angiotensin-converting enzyme 2 and acetylated 7 sialic acid residues, which are highly expressed by vascular endothelial cells (EC), could instigate 8 cardiovascular dysfunction.(14-19) However, recent in vitro evidence has alternatively suggested 9 that SARS-CoV-2 may have limited infectious potential and replicative ability in ECs.(20) In this 10 respect, mechanisms secondary to direct infection, such as cascading immunological activation 11 may instead be the driving factor behind the observed endothelial dysfunction; particularly among 12 the populations with coexisting conditions where EC dysfunction is already evident.(21, 22) In 13 fact, recent single cohort analysis of select endothelium-related biomarkers such as 14 thrombomodulin(23), von Willebrand Factor(24), angiopoietin-2 (Ang-2)(25), and soluble 15 triggering receptor expressed on myeloid cells-1 (sTREM-1)(26) have shown utility in 16 prognostication, being associated with both disease severity and in-hospital mortality. While 17 markers of endothelial function may aid in prognostication, it seems unlikely that a simple 18 combination of markers can provide insight significant enough to adjudicate the level of care a 19 patient will need, nor is it apparent that these markers associate specifically with COVID-19 20 pathology. To date, the integration of clinical data with broad omics technologies has opened up 21 new avenues for efficiently delineating complex patient phenotypes and their associations with 22 clinical outcomes.(27, 28) Circulating profiles of plasma microRNAs (miRNA), in particular, have 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint 1 and Online Figure I) between May 2020 to December 2020 (prior to vaccine availability). 1 Infection status of admitted patients was confirmed by at least two SARS-CoV-2 polymerase-chain 2 reaction tests. Patients with SARS-CoV-2 but noninfectious etiologies were not enrolled (e.g., 3 blunt force trauma). Medical history, physical examination, clinical laboratory values, and acute 4 illness scores (Acute Physiologic Assessment and Chronic Health Evaluation II and Sequential 5 Organ Failure Assessment) were recorded upon admission (day 0 or 1; t0-1), two to three days later 6 (t2-3), and up to five days (t4-5), along with the synchronous collection of blood samples (Online 7 Figure I ). As a result of sampling logistics, t0 or t1 (days) were grouped as the earliest timepoint 8 available for admitted patients. Similarly, collapsed sampling timepoints at t2-3 and t4-5 allowed 9 evaluation of patient trajectories. Since the analysis was conducted retrospectively, clinical care 10 was dictated by individual care providers with the primary outcome being mortality. 11 The cohort of 241 patients was categorized into three groups that reflected conventional concepts 13 of COVID-19 severity (National Institutes of Health, 'Clinical Spectrum of SARS-CoV-2 14 Infection'), as well as two analogous symptom/severity matched control groups.(31) The resulting 15 five groups were: SARS-CoV-2 negative patients presenting to outpatient clinics with symptoms 16 consistent with a respiratory tract illness (n=30, "mild negative"); SARS-CoV-2 positive patients 17 presenting to outpatient clinics with symptoms consistent with a respiratory tract illness (n=27, 18 "mild COVID-19"); admitted SARS-CoV-2 positive patients requiring supplemental oxygenation 19 (n=39, "moderate COVID-19"); SARS-CoV-2 positive patients requiring high-level in-patient 20 ICU care (n=76, "severe COVID-19"); and SARS-CoV-2 negative patients who exhibited 21 symptoms of a severe respiratory disease requiring high-level in-patient ICU care (n=69, "severe 22 negative"). Patients testing negative by nasopharyngeal SARS-CoV-2 polymerase chain reaction 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint had immunological history (i.e., antibody reactivity) ascertained through spike antigen cross-1 reactivity using a Federal Drug Administration approved enzyme-linked immunosorbent sandwich 2 assay (Online Figure II) . 3 4 Data Visualization and Statistical Analysis: All data generated and analyzed which support the 5 findings of this study are included in this article. Associated supplementary information files are 6 available on a publicly accessible archive (see below). Descriptive Analysis -Clinical 7 characteristics were characterized using summary statistics. Continuous variables were described 8 using median and inter-quartile range (IQR), and dichotomous or polytomous variables were 9 described using frequencies. Between-group differences were evaluated using Wilcoxon rank-sum 10 tests for continuous variables and Fisher's exact tests for dichotomous/polytomous variables. 11 Correlation between continuous variables were quantified using Spearman rank correlation. 12 Descriptive outcome analysis -The Kaplan-Meier survival method was applied to assess the in-13 hospital death, and the between-group differences in the freedom from death were evaluated using 14 log-rank tests. The length of hospitalization/ICU was characterized using competing risk models 15 in terms of cumulative incidence rate function. Univariable Cox proportional hazard regression 16 were applied to assess and quantify the association of the baseline clinical characteristics with in-17 hospital/ICU death. The associations of continuous variables were modeled using natural cubic 18 splines. Biomarker Analysis -Comparisons between two independent groups were made using t-19 tests for normally distributed continuous variables or Wilcoxon rank-sum tests non-normally 20 distributed continuous variables. When more than two groups were compared, either a one-way 21 ANOVA with a Tukey or Bonferroni post-hoc test (where appropriate) for multiple testing 22 correction, Kruskal-Wallis one-way analysis of variance with Dunn's multiple comparison 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint correction. Two-way ANOVA was used to estimate how the mean quantitative variable changes 1 according to time and group differences in leak experiments. Where appropriate, Benjamini-2 Hochberg false discovery rate (FDR) was utilized with adjusted P values (or Q value where stated) 3 of <0.05 being considered statistically significant and indicated in the graphs as reported by the 4 analysis software with significance thresholds of P<0.05, P<0.01, P<0.001, and P<0.0001 5 indicated as *, **, ***, **** respectively. MiRNA pathway analysis was conducted using 6 BioCarta/KEGG/Reactome databases and tested for enrichment by a hypergeometric test with 7 adjustment for multiple comparisons using the Benjamini-Hochberg FDR, with P ≤0.05 considered 8 to be statistically enriched in a gene set of interest. (32) (33) (34) ) Although many hypotheses were tested 9 throughout the manuscript, no experiment-wide multiple test correction was applied. Unless into the study by a lawfully entitled substitute decision-maker on behalf of a participant when 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. and 159 (66.0%) patients having more than one co-existing condition. In this respect, there were 20 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint calcium channel blockers (P=0.0338), and diuretics (P=0.0088) between groups (Table 1) . 1 Although a small proportion of patients had documented smoking status, there were 40 (total 138; 2 29.0%) current or former smokers, with 33 (total 234; 14.1%) patients having chronic obstructive 3 pulmonary disease. 33 died, with Kaplan-Meier survival curves indicating that patients with severe disease higher 16 mortality risk than did patients with less severe phenotypes (P<0.001, Figure 1A ). Furthermore, 17 the only baseline clinical metrics associated with mortality within admitted patients were the 18 history of coronary artery disease (Log-rank, P=0.011) and age at hospital admission (Log-rank, 19 P=0.046, Online Table IV is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. matched SARS-CoV-2-negative controls, patients with severe COVID-19 had on average longer 1 ICU stays (mean 13 [IQR, 7-35] days versus mean 8 [IQR, 3-15] days, P<0.0001), were more 2 likely to display acute respiratory distress syndrome (ARDS, P=0.0195), and had overall worse 3 oxygenation, requiring higher FiO2 (P<0.0001) as well as having lower PaO2/FiO2 ratios 4 (P=0.0003, Online Table V) . Interestingly, although many cardiovascular metrics were unchanged, 5 admitted COVID-19 patients had an increase in the number of noted arrhythmic events (P=0.007) 6 and a higher number of secondary cardiovascular events (P=0.001, Online Table V is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. standalone assays measuring myeloperoxidase (MPO) and high-sensitivity cardiac troponin (hs-1 cTnI). Correlation analysis revealed that Ang-2, sE-Selectin, and sICAM had moderate 2 correlations with numerous other inflammatory markers, with correlation coefficients ranging 3 from -0.89 (IL-6:MPO) to 0.89 (sICAM-1:Ang-2; Online Figures IV-IX). Myeloperoxidase was 4 the only biomarker without a significant correlation. Among these markers, two were significantly 5 different between mild SARS-CoV-2 positive patients and SARS-CoV-2 negative patients with 6 mild illness at t0-1 (elevated Ang-2, P=0.0200; elevated MPO, P=0.0439, Figure 2A severity within the SARS-CoV-2 positive cohort but failed to demonstrate significance between 10 the critically ill patient groups that did or did not have COVID-19. 11 12 Amongst all patients admitted with COVID-19, univariable analysis revealed that only Ang-2 was 13 associated with mortality (P=0.015; Online Table IV) , suggesting higher concentrations are 14 associated with higher mortality, while both Ang-2 (P=0.020) and sVCAM-1 (P=0.012) were 15 associated with mortality when looking specifically at the severe COVID-19 patients ( Figure 2C , 16 D, Online Table VI) . Sub-stratifying these markers by severity, there were significantly higher t0-17 1 concentrations of Ang-2, IL-6, and MPO in non-survivors with COVID-19 when compared to 18 the severe SARS-CoV-2 negative patients (Online Figure XI) . Over time, only IL-6 and MPO 19 remained significantly different between severe COVID-19 and severe negative patients (Online 20 Figure XII and XIII). Taken together, while markers of inflammation/endothelial dysfunction were 21 observed at early timepoints and associated with severity, the majority were not specific to either 22 COVID-19 status or mortality. 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 1 Mortality: While select inflammatory/EC activation markers were informative for ICU mortality, 3 they lacked the ability to distinguish between COVID-19 and non-COVID-19 pathology. We 4 postulated that assessment of the circulating miRNA transcriptome may provide further precision 5 with respect to patient subgroups, since miRNA profiling (in contrast to circulating protein 6 markers) has been shown to effectively differentiate complex disease etiologies.(40, 41) Using 7 whole transcriptome miRNA sequencing (2,083 mature miRNAs) we screened plasma obtained 8 from the differing groups of disease severity to identify meaningful differences in miRNA 9 composition (n=30, negative mild; n=14, mild COVID-19; n=15, moderate COVID-19; n=36, 10 severe COVID-19; n=33, negative severe). Comparative analyses indicated that there were 11 substantially higher numbers of differentially expressed miRNA as disease severity progressed 12 is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. Learning: Given the high dimensionality of the dataset generated, we sought to examine the utility 2 of models developed using machine learning to predict in-hospital mortality of COVID-19 patients 3 at admission, based on common clinical data, protein expression data, and miRNA expression data. 4 We performed 250 experiments using repeated randomized stratified sub-sampling cross-5 validation into disjoint sets of 80% training and 20% testing, to train a set of Random is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint 3), we next sought to identify candidate miRNA markers that significantly contribute to mortality 1 risk. We considered miRNAs within the top 50% of abundance and cross-examined the differential 2 expression between the both survivors and non-survivors of COVID-19 as well as SARS-CoV-2 3 severe negatives. More so, we took into consideration existing biological relevance, thereby 4 specifically analyzing miR-1 which is associated with myocardial injury(43, 44), miR-199a-3p 5 which has been shown to be cardioprotective(45), miR-181a-5p which has been shown to restrict 6 vascular inflammation(46), along with members of the miR-30 family which are enriched in ECs 7 and capable of modulating inflammation(47-49), as well miR-339-3p and miR-6080 which were 8 among the highest differentially expressed. Univariable hazard ratios and log-rank P-values were 9 generated to determine the relationship of the miRNA expression measured in plasma with 10 mortality. When ranking by significance of independent association with mortality, miRNAs are 11 among the highest ranking factors comparing to other clinical metrics (Online Table IV is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. underscored the impact of endothelial barrier disruption on disease outcomes.(52) The data 1 presented so far suggested that dysfunction of the endothelium may also accompany COVID-19 2 and may be associated with poor outcomes. To further examine this hypothesis, we utilized two 3 validated permeability platforms using an endothelial monolayer model: continuous monitoring of 4 transendothelial electrical resistance ( Figure 5A ) using the xCelligence platform, and a transwell 5 system consisting of a confluent monolayer of pooled human umbilical vein ECs on a permeable 6 membrane. In a timeframe that would exclude viral replication(20, 53), cells were treated with 7 20% (v/v) plasma from the t0-1 samples from across the disease severity spectrum of COVID 8 positive and negative patients. Moderate and severe COVID-19 patient plasma induced significant 9 endothelial barrier dysfunction, while the mild and mild negative patient plasma did not induce 10 significant EC leak in the xCelligence assay ( Figure 5B) . Similarly, using a validation cohort, the 11 integrity of the monolayer was gauged through leakage of a large dextran tracer across the EC 12 barrier in both acute (i.e., one-hour) and longer-term (i.e., six-hour) treatment. This revealed 13 barrier disruption in response to moderate and severe COVID-19 patient plasma, but not in 14 response to plasma from severe negative patients (Online Figure XIX ). Of the mediators tested, 15 EC permeability correlated with levels of Ang-2, hs-cTnI, ET-1, IL-6, IL-8, sTREM-1, and (Online 16 Figure XX) . These experiments provide in vitro evidence that barrier dysfunction can be 17 independent of direct SARS-CoV-2 infection. We reasoned that elucidating the unique pathways 18 through which the more severe COVID-19 phenotypes exert their barrier disruptive effects may 19 reveal potential therapeutic approaches to maintain EC barrier. 20 21 Dysfunction: To gain a better understanding of how the endothelium is modulated by plasma 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint components, we investigated the gene-level changes using RNA-sequencing of ECs after six hours 1 of co-incubation with patient-derived plasma. On the transcriptome level (16,285 total quantified 2 genes, having ≥10 reads in at least five samples), biological replicates within groups were tightly 3 correlated (Pearson r=0.975-0.988, n=4-5), suggesting robust intra-group clustering even among a 4 heterogeneous patient group, similar to previous studies.(27) Principal component analysis of the 5 normalized transcriptome showed segregation between experimental groups, with the severe group 6 being clearly distinct from the mild and moderate groups ( Figure 5C ). Pair-wise differential 7 expression analysis with the mild SARS-CoV-2 negative cohort as the control revealed 393, 49, 8 and 246 genes are differentially expressed (FDR < 0.05, log2 fold change >±0.58) in the mild, 9 moderate, and severe COVID-19 cohorts, respectively ( Figure 5D ). Gene set enrichment analysis 10 revealed that co-incubation with either severe or moderate COVID-19 patient plasma altered the 11 expression of endothelial genes related to acute inflammatory response, angiogenic programs, or 12 histone deacetylase activity, whereas administration of plasma from mild COVID-19 patients 13 altered the expression of genes involved in priming of antiviral responses ( Figure 5E and Online 14 Data Files VI, VII, and VIII). Similarly, pathway enrichment analysis using gProfiler with 15 significant differentially expressed genes highlighted a predominance of KEGG pathways relating 16 to interferon in the mild group, while in contrast, pathways in ECs exposed to moderate and severe 17 COVID-19 plasma related more broadly to cell motility, developmental processes, cell stress 18 responses, cell structure reorganization, and actin mobilization ( Figure 5F ). Collectively, these 19 results suggested endothelial structural changes could be occurring that might be amenable to 20 treatment with barrier stabilizing agents. 21 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint Plasma Exposure in vitro: We next examined two principal structural contributors to the EC 2 barrier, vascular endothelial-cadherin (VE-Cadherin) and Claudin-5. VE-cadherin is an essential 3 adherens junction protein that regulates cell-cell junctional stability, and Claudin-5 is a tight 4 junction protein that regulates size-dependent paracellular permeability pathways.(54) 5 Immunostaining confirmed significant disruptions to both VE-cadherin and Claudin-5 expression 6 as well as junctional localization, particularly in ECs exposed to the moderate and severe COVID-7 19 plasma ( Figure 6A ). To gauge the potential clinical importance of these changes in relation to 8 vascular leak, we next investigated whether targeted therapeutics that are known to stabilize the 9 vascular barrier or suppress EC activation can prevent COVID-19 plasma-induced permeability in 10 vitro. The following drugs that have been reported to reduce endothelial dysfunction were utilized: 11 Q-peptide, a synthetic integrin-binding motif of angiopoietin-1(55); Slit2-N, a recombinant 12 member of the Slit family of secreted extracellular matrix glycoproteins that stabilizes adherens 13 junctions(56); Nangibotide, a TREM-1 inhibitor(57); and dexamethasone, a potent synthetic 14 adrenal corticosteroid currently used in COVID-19 treatment(58). Co-treatment with either Q-15 peptide or Slit2-N prevented the disruption of the EC barrier, as examined through junctional 16 protein expression ( Figure 6B , C, left) and permeability as examined by xCelligence readout 17 ( Figure 6B, C, right) . While other agents could maintain junctional protein expression (e.g., 18 Nangibotide, dexamethasone) in specific settings, no other agents were universally able to 19 maintain barrier protein expression in the face of moderate or severe disease plasma exposure and 20 prevent the physiologic barrier disruption measured by electrical resistance. 21 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint DISCUSSION: Recent literature on SARS-CoV-2 pathogenesis has suggested that the induction 1 of substantial acute respiratory distress phenotypes is driven by a mismatched inflammatory 2 response together with broad vascular dysfunction.(59, 60) While several detailed reports 3 implementing multi-omic approaches have provided insight into the immune cell phenotypes 4 involved in these processes(61-63), risk stratifying immune markers specific to COVID-19 have 5 not been fully elucidated. Herein, we provide a comprehensive, multi-omics-based description of 6 the molecular antecedents to COVID-19 mortality, yielding new insights pertaining to the 7 vasculature. Our study further delineates the gradient of vascular dysfunction observed in patients 8 across the spectrum of COVID-19 severity, particularly among those with severe illness. While 9 our findings are consistent with smaller cohort studies examining single markers of cardiovascular 10 dysfunction(26, 64), our report is the first to use proper, disease-negative controls, which allowed 11 us to ask etiological questions. Our analysis revealed that although markers of cardiovascular 12 dysfunction (such as Ang-2 and sVCAM-1, Figure 2A severe SARS-CoV-2 negative controls using samples gathered on admission. Through this study, 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint we provide further empirical evidence for the value of data from novel biomarkers over metrics 1 traditionally utilized in healthcare systems (i.e., clinical demographics and laboratory data). 2 Notably, we show that the data from electronic medical records fails to adequately capture the risk 3 of in-hospital mortality. Using a Random Decision Forest machine learning model on our multi-4 omic biologically relevant datasets we were able to develop an exploratory prognostic risk 5 prediction model that incorporates markers that are COVID-19 and vascular specific. Importantly, 6 there are several platforms that allow the interrogation of candidate miRNAs within clinically 7 relevant timeframes.(67, 68) While still investigational, given the extent of pre-clinical and clinical 8 research on miRNAs(40), reasonable to expect miRNA datasets will soon be amenable for clinical 9 implementation. As an example of the importance of a multi-omic approach that includes clinically 10 relevant disease negative controls, miR-1 (which has been intensively studied as a cardiac enriched 11 miRNA and associated with numerous cardiac etiologies(41, 43, 44, 69)) failed to show an 12 association with in-hospital mortality in our study. If the promising miRs-30b/c/e, -6080, -181a-13 5p, -199a-3p, and -339 identified here are validated in larger populations they would represent new 14 biomarkers that could be utilized for rapid in-hospital risk assessment. As miRNAs represent 15 functional biomarkers, having active roles in gene regulation, they also present an important 16 opportunity to understand the pathophysiological relevance of endothelial-based processes 17 affected by SARS-CoV-2 infection. Several lines of evidence support the hypothesis that endothelial function is a major determinant 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. With the known roles of Ang-1 and Ang-2 as ligands of the endothelial cell-specific integrins, and 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint their antagonistic relationship as modulators of endothelial survival, the barrier maintenance seen 1 through Q-peptide could be due to antagonism of higher Ang-2 levels in more severe COVID-2 19.(78, 79) Additionally, measurement of the levels of endogenous sSlit2 in SARS-CoV-2 patient 3 plasma revealed a significant upregulation only in patients with severe disease suggesting a 4 possible compensatory mechanism and furthering the biological relevance of Slit2-N treatment 5 (Online Figure XXI) . Therefore, as a proof-of-concept, these data support the hypothesis that Taken together, we have provided the first highly salient comparison of a diverse group of 10 cardiovascular markers between COVID-19 positive and negative patients, highlighting that well-11 known markers of inflammation, cardiac damage, and endothelial dysfunction are not specific to 12 COVID-19 pathology. Notably, with a targeted miRNA transcriptomic approach, we were able to 13 discern specific markers that showed better discrimination between these two groups. Using 14 machine learning, incorporation of protein and miRNA markers improved the prediction of in-15 hospital mortality over baseline clinical variables. While exploratory, this is a clinically feasible 16 approach that has the advantage of using pathophysiologically relevant SARS-CoV-2's markers, 17 as opposed to the surrogate markers used in most other published risk stratification models.(10, 18 26) Finally, our data provide several lines of evidence supporting the notion that endothelial barrier 19 function is affected in a SARS-CoV-2 specific manner, that is distinct from the pathways involved 20 in critically ill patients with non-COVID-19 severe respiratory illnesses. Our data reinforce the 21 idea that barrier dysfunction is likely independent of direct viral infection and instead secondary 22 to yet undiscovered mediators in the plasma. We further pursued our observations using assays of 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint EC function using an in vitro model, which served as a platform for rational therapeutic choice. 1 Here, we show that EC barrier is reduced by the addition of COVID-19 patient plasma in a disease 2 severity-dependent manner and that this can be prevented by stabilization with synthetic Ang-1 3 (Q-peptide) and Slit2-N. Together, our work provides biological insight into the role of the 4 endothelium in SARS-CoV-2 infection, the importance of miRNA as disease-and pathway-5 specific biomarkers, and the exciting possibility that endothelial barrier stabilizing treatments 6 might hold promise in COVID-19. Moreover, we provide insights into the use of this approach to 7 find therapeutic options that might prove useful in other critical illnesses and emerging infectious 8 diseases where endothelial permeability is central to disease pathophysiology. The in vitro system used in this manuscript represents a reductionist approach to disease modelling, 18 and as such, vascular cell co-culture or examination of specific endothelial beds (i.e., coronary or 19 pulmonary) may better inform us about microenvironment changes in COVID-19. Putative 20 therapies could be further assessed in co-culture models (e.g., organ on a chip) that better mimic 21 tissue complexity and will require testing in animal models. Nevertheless, our findings provide 22 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint Network, and the Toronto Academic Vascular Surgeons. We would also like to gratefully 1 acknowledge the participants who graciously contributed to the study, teams of research 2 coordinators, as well as the clinicians, nurses, and biobank members who assisted with the 3 collection of samples that were utilized in this study. Parts of figures (including the entire graphical 4 abstract) was created with BioRender.com. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint 1 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint 1 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) In-Hospital Mortality for Severe COVID-19 Patients. Assessment of datasets using repeated 2 randomized stratified sub-sampling cross-validation (Random Forest machine learning) for (a) 3 clinical data, (b) clinical data and protein expression metrics, and (c) clinical data and miRNA 4 atlas expression metrics. Univariable log hazard ratios of candidate microRNAs (d) hsa-miR-6080, 5 (e) hsa-miR-30e-5p, (f) hsa-miR-30c-5p, and (g) hsa-miR-30b-5p in relation to mortality. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint 1 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. respectively, based on a log fold-change of >±0.58. (e) Gene set enrichment plot of the top-ranked 20 gene set, TNF Targets Up (FDR = 7.05x10 -9 , NES = 3.13), WT1 Targets Up (FDR = 1.45x10 -8 , 21 NES = 2.66), and interferon responsive genes (FDR = 4.52x10 -9 , NES = 3.24) using all genes 22 ranked by their magnitude of association with each respective disease severity group (the 23 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 9, 2022. ; https://doi.org/10.1101/2022.02.08.22270636 doi: medRxiv preprint PRIMARY TABLES: Expression Is Required for SARS-CoV-2 To Infect Primary Human Endothelial Cells 2 and Induce Inflammatory and Procoagulative Responses SARS-CoV-2 infection of human iPSC-derived cardiac cells reflects cytopathic features 5 in hearts of patients with COVID-19 Broken barriers: a new take on 8 sepsis pathogenesis Inhibition of 10 SARS-CoV-2 Infections in Engineered Human Tissues Using Clinical-Grade Soluble 11 Human ACE2 VE-cadherin and claudin-5: it takes two to tango Angiopoietin-1 15 promotes cardiac and skeletal myocyte survival through integrins Slit2-Robo4 18 signalling promotes vascular stability by blocking Arf6 activity Triggering receptor 21 expressed on myeloid cells-1 inhibitor targeted to endothelium decreases cell activation. 22 Frontiers in immunology Changes in cytoskeletal and tight junctional proteins correlate with decreased 25 permeability induced by dexamethasone in cultured rat brain endothelial cells. 26 Neuroscience letters Facing COVID-19 in the ICU: vascular 28 dysfunction, thrombosis, and dysregulated inflammation Vascular 31 endothelial damage in the pathogenesis of organ injury in severe COVID-19 The 34 global phosphorylation landscape of SARS-CoV-2 infection Single-cell multi-omics analysis of the immune response in COVID-19 COVID-19 immune features revealed 40 by a large-scale single-cell transcriptome atlas Clinical and pathological investigation of 42 patients with severe COVID-19 The Prognostic Value of Plasma MicroRNA-155 and 44 MicroRNA-146a Level in Severe Sepsis and Sepsis-Induced Acute Lung Injury Patients Accuracy of circulating 1 microRNAs in diagnosis of sepsis: a systematic review and meta-analysis Point-of-4 care genetic testing for personalisation of antiplatelet treatment (RAPID GENE): a 5 prospective, randomised, proof-of-concept trial MicroRNA 7 amplification and detection technologies: opportunities and challenges for point of care 8 diagnostics MicroRNAs miR-1, miR-133a, miR-133b 10 and miR-208 are dysregulated in human myocardial infarction Immune consequences 13 of endothelial cells' activation and dysfunction during sepsis Sepsis and endothelial permeability Vascular endothelial growth 18 factor may contribute to increased vascular permeability in acute respiratory distress 19 syndrome Mitigates SARS-CoV-2-Induced Inflammatory Lung Vascular Leakage and Mortality in 22 Humanized K18-hACE-2 Mice Impaired type I 25 interferon activity and inflammatory responses in severe COVID-19 patients BET 28 inhibition blocks inflammation-induced cardiac dysfunction and SARS-CoV-2 infection. 29 Cell Interleukin-6 promotes a sustained loss 31 of endothelial barrier function via Janus kinase-mediated STAT3 phosphorylation and de 32 novo protein synthesis Influenza Virus-34 cytokine-protease cycle in the pathogenesis of vascular hyperpermeability in severe 35 influenza. The Journal of infectious diseases Endothelial Cell Survival Through the Phosphatidylinositol 3′-Kinase/Akt 38 Signal Transduction Pathway Angiopoietin-1 40 and its receptor Tie-2 participate in the regulation of capillary-like tubule formation and 41 survival of endothelial cells Robo4-dependent Slit signaling to survive the cytokine storm in sepsis and influenza. 44 Science translational medicine Intermedin protects against sepsis 1 by concurrently re-establishing the endothelial barrier and alleviating inflammatory 2 responses Dysregulation 4 of angiopoietin-1 plays a mechanistic role in the pathogenesis of cerebral malaria A double-blind, 7 placebo-controlled, randomised, multicentre, proof-of-concept and dose-finding phase II 8 clinical trial to investigate the safety, tolerability and efficacy of adrecizumab in patients 9 with septic shock and elevated adrenomedullin concentration (AdrenOSS-2) Claudin-5, P=8.0×10 -3 . Mild vs moderate: Claudin-5, P=2.9×10 -2 . Mild vs severe: Claudin-5 -e) Left; Representative confocal microscopy images depicting a combinatorial 12 screen of COVID-19 plasma and matched controls treated with either (b) Q-peptide, (c) Slit2-N, 13 (d) Nangibotide, and (e) Dexamethasone on pHUVECs reveals modulated maintenance of VE-14 cadherin (red) and Claudin-5 (green) expression; n=3-4 per group Imaging P values were 16 determined by one-way ANOVA with Dunnett's multiple comparisons test (all compared to 17 control). Nangibotide; Moderate vs negative: VE-cadherin *P=1 Severe vs negative: VE-cadherin *P=2.4×10 -2 , Claudin-5 *P=3 Quantification of VE-cadherin and Claudin-5 for each respective group Distribution -no. (%) Male Sex, no./total no. (%)