key: cord-297209-84gs67bn authors: Livanos, A. E.; Jha, D.; Cossarini, F.; Gonzalez-Reiche, A. S.; Tokuyama, M.; Aydillo, T.; Parigi, T. L.; Ramos, I.; Dunleavy, K.; Lee, B.; Dixon, R.; Chen, S. T.; Martinez-Delgado, G.; Nagula, S.; Ko, H. M.; Reidy, J.; Naymagon, S.; Grinspan, A.; Ahmad, J.; Tankelevich, M.; Gordon, R.; Sharma, K.; Britton, G. J.; Chen-Liaw, A.; Spindler, M. P.; Plitt, T.; Wang, P.; Cerutti, A.; Faith, J. J.; Colombel, J.-F.; Kenigsberg, E.; Argmann, C.; Merad, M.; Gnjatic, S.; Harpaz, N.; Danese, S.; Rahman, A.; Kumta, N. A.; Aghemo, A.; Petralia, F.; van Bakel, H.; Garcia-Sastre, A.; Mehandru, S. title: Gastrointestinal involvement attenuates COVID-19 severity and mortality date: 2020-09-09 journal: medRxiv : the preprint server for health sciences DOI: 10.1101/2020.09.07.20187666 sha: doc_id: 297209 cord_uid: 84gs67bn Given that gastrointestinal (GI) symptoms are a prominent extrapulmonary manifestation of coronavirus disease 2019 (COVID-19), we investigated the impact of GI infection on disease pathogenesis in three large cohorts of patients in the United States and Europe. Unexpectedly, we observed that GI involvement was associated with a significant reduction in disease severity and mortality, with an accompanying reduction in key inflammatory proteins including IL-6, CXCL8, IL-17A and CCL28 in circulation. In a fourth cohort of COVID-19 patients in which GI biopsies were obtained, we identified severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) within small intestinal enterocytes for the first time in vivo but failed to obtain culturable virus. High dimensional analyses of GI tissues confirmed low levels of cellular inflammation in the GI lamina propria and an active downregulation of key inflammatory genes including IFNG, CXCL8, CXCL2 and IL1B among others. These data draw attention to organ-level heterogeneity in disease pathogenesis and highlight the role of the GI tract in attenuating SARS-CoV-2-associated inflammation with related mortality benefit. hospital admission, age > 18 years and having a multiplexed cytokine panel performed during 160 their admission (Supplementary Fig. 1) . The basic demographics and clinical characteristics are 161 summarized in Table 1 We analyzed diarrhea, nausea and vomiting at the time of hospital admission. Two 171 hundred and ninety-nine patients (47%) reported any GI symptoms (nausea, vomiting and/or 172 diarrhea) with diarrhea being the most common (245 patients, 39%), followed by nausea (157 173 patients, 25%), and then vomiting (82 patients, 13%) ( Table 3) . Patients with GI symptoms were 174 significantly younger (average age 61 years) than those without GI symptoms (average age 67 175 years) (p<0.0001) ( Table 1) . The distribution of race, ethnicity and co-morbid illnesses including 176 obesity, HTN, DM and inflammatory bowel diseases (IBD) were comparable between those with 177 GI symptoms and those without GI symptoms ( Table 1) In an effort to determine the association between GI symptoms and COVID-19 severity, we initially 182 performed univariate analysis comparing the distribution of COVID-19 severities in patients with 183 and without GI symptoms. Patients presenting with GI symptoms had less severe disease than 184 patients without GI symptoms (p<0.001 Chi-square test). Importantly, mortality was significantly 185 lower in COVID-19 patients with GI symptoms (15.7%) than those without GI symptoms (31.0%; 186 p<0.0001 Fisher's exact test) (Table 2, Fig. 1a) . Furthermore, each individual GI symptom 187 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint (nausea, vomiting and diarrhea) was associated with less severe disease (p<0.02 Fisher's exact 188 test) and lower mortality (p<0.001 Fisher's exact test) (Fig. 1a) . These findings were further 189 emphasized by Kaplan-Meier estimates of survival over short-term follow-up of 25 days (p<0.001 190 log-rank test) (Fig. 1b, Supplementary Fig. 2a) . Consistent with prior reports 35,40,41 older age and 191 higher disease severity were associated with higher mortality in the discovery cohort (Table 4) , providing validity to our findings. Next, we decided to account for multiple comorbidities in determining the impact of GI 194 symptoms on COVID-19 outcomes. Herein, we adjusted for age, body mass index (BMI), gender, 195 race, diabetes, HTN, chronic lung disease and heart disease in a multivariate model to determine 196 the impact of any GI symptoms and each of the individual GI symptoms (nausea, vomiting or 197 diarrhea) on disease outcomes. In the multivariable models, any GI symptoms, diarrhea, nausea, 198 and vomiting, were inversely associated with disease severity and mortality, while age and BMI 199 were positively associated with these outcomes. African-American race was inversely associated 200 with disease severity but not mortality (Fig. 1c, Supplementary Fig. 2b , Supplementary Table 201 2). Patients who presented with GI symptoms had 50% reduced odds of having severe disease 202 (odds ratio of 0.56) and death from COVID-19 (odds ratio 0.54), compared to the patients who 203 presented without GI symptoms (Fig. 1d, Supplementary Table 2 ). An external validation cohort further confirms decreased mortality in COVID-19 patients 206 with GI symptoms 207 Next, we sought an external validation cohort, distinct from the MSH discovery cohort, and studied 208 well-characterized patients (n=287) from Milan, Italy, to determine the impact of GI symptoms on 209 COVID-19 associated outcomes. In this cohort, GI symptoms on admission were characterized 210 as presence (n=80, 27.9%) or absence of diarrhea on admission ( Table 5) . Consistent with the 211 discovery cohort, patients with diarrhea on admission were significantly younger (60.6 ± 13.9 vs 212 65.5 ± 13.3 for patients without diarrhea, p = 0.0056) and had significantly lower mortality (10.0% 213 in patients with diarrhea vs 23.7% in patients without diarrhea, p=0.008). Additionally, patients 214 with diarrhea had lower composite outcome of mortality or ICU admission compared to those 215 without diarrhea (20% vs 40%, p=0.0014) ( Table 5) . Distinct from the discovery cohort, the 216 proportion of male patients was lower in those patients with diarrhea (57.5%) compared to those 217 without diarrhea on admission (72%) (p=0.0238) ( Table 5) . Next, the association between 218 diarrhea and mortality was evaluated by multivariate logistic regression adjusting for age, gender, 219 BMI, diabetes, chronic heart and lung disease and other confounders. Even after adjusting for 220 these covariates, the presence of diarrhea on admission was found to be significantly inversely 221 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint associated with mortality with a median odds ratio of 0.33 over 1000 bootstrap iterations (Fig. 222 1e). Presence of GI symptoms can be used to predict reduced disease severity and mortality 224 in patients with COVID-19 in a second validation cohort 225 After observing significantly reduced mortality in COVID-19 patients with GI symptoms in our 226 discovery and external validation cohort, we developed a predictive model based on the discovery 227 cohort and applied it to a distinct internal validation cohort comprising of 242 well-characterized 228 patients with COVID-19, admitted between April 16, 2020 and April 30, 2020 to MSH. The 229 inclusion of 'any GI symptoms' to a model consisting of age and BMI at baseline, improved the 230 ability to predict severity and mortality with a median area under the curve (AUC) of 0.59 (age + 231 BMI) vs. 0.64 (age + BMI + any GI symptoms) for disease severity and 0.70 (age + BMI) vs. 0.73 232 (age + BMI + any GI symptoms) for mortality (Fig. 1f, Supplementary Table 3 ). In addition, the 233 effect of GI symptoms, age and BMI on the AUC was evaluated by excluding each variable one 234 at a time from the model and calculating the consequent reduction in AUC. The exclusion of GI 235 symptoms resulted in a significant reduction in AUC with a median value of 0.054 for disease 236 severity and 0.03 for mortality. Notably, the effect of GI symptoms on the AUC was more dramatic 237 than that of age (AUC reduction of 0.054 versus 0.025) for disease severity (Fig. 1f To gain mechanistic insights into significantly reduced COVID-19 severity and mortality across 243 these three large cohorts, we began by analyzing systemic biomarkers, comparing patients with 244 and without GI symptoms admitted to MSH (discovery and internal validation cohorts). Initially, 245 we examined, a set of 4 cytokines, IL-6, IL-8, TNF-α, and IL-1β, measured on admission in all the 246 patients as part of routine clinical care. IL-6, and IL-8 levels, known to be associated with poor 247 survival 35 were found to be significantly reduced in circulation of patients with GI symptoms (FDR 248 10%) (Supplementary Fig. 3, Supplementary Table 4 ). Next, to facilitate high dimensional analyses of potential immunological differences 250 between patients with and without GI symptoms, we performed a validated, multiplexed proteomic 251 assay (O-link), simultaneously quantifying 92 protein analytes in 238 patients (from among the 252 discovery and internal validation cohorts; GI symptoms (n=104), no GI symptoms (n=134)) where 253 serum samples were available for analyses. Unsupervised consensus clustering of these 92 254 analytes revealed six groups of analytes with similar expression patterns across all COVID-19 255 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint patients (Fig. 2a, Supplementary Table 5) . Notably, analytes in clusters 5 and 6 displayed less 256 correlation in patients with GI symptoms compared to those without GI symptoms (Fig. 2a, 257 Supplementary Fig. 4) . Next, we interrogated biological pathways over-represented in each 258 cluster of soluble analytes. We found that the "KEGG Jak/Stat Signaling Pathway" was 259 significantly enriched in Cluster 5; while the "Hallmark Inflammatory Response" pathway was 260 significantly enriched in Cluster 4 (Fisher's exact test 10% FDR). These pathways were 261 downregulated in patients displaying diarrhea symptoms based on pathway level signatures 262 (p<0.05 from t-test) (Fig. 2b) ; suggesting a reduced inflammatory response in patients affected 263 by GI symptoms. In addition, we found that clusters 1, 2, 3, 5 and 6 were significantly 264 downregulated in patients with GI symptoms compared to those without (FDR 15%) (Fig. 2c) . This downregulation seemed to be driven mostly by diarrhea since the same clusters 1, 2, 3, 5 266 and 6 were significantly downregulated with FDR correction at 10% in patients who presented 267 with diarrhea compared to those without symptoms. The lack of signal registered for nausea and 268 vomiting might be due to reduced statistical power given by the smaller number of samples 269 displaying vomiting (n=29) or nausea (n=54) symptoms. 275 symptoms compared to those without (t-test FDR 10%) (Fig. 2d,e, Supplementary Table 6 ). When looking at each individual GI symptom, diarrhea had the most significantly differential 277 analytes. As a caveat, this difference between diarrhea, nausea and vomiting might be due to 278 reduced statistical power given that fewer patients presented with nausea and vomiting than those 279 who presented with diarrhea. Consistent with GI symptoms as a group, IL-7 was significantly 280 increased, in addition, MCP-2 was significantly increased in patients presenting with diarrhea. We also quantified total anti-spike protein IgA, IgG and IgM antibodies and compared 282 patients with and without GI symptoms and patients with and without diarrhea. There were no 283 significant differences between the groups (Supplementary Fig. 5) Thus, overall, GI symptoms are associated with significantly reduced levels of key 285 inflammatory cytokines like IL-6, IL-8, IL-17 and CCL28 that are known to be associated with poor 286 COVID-19 outcomes. The GI tract was endoscopically uninflamed in all patients 289 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint Next, we sought to obtain GI tissue-level mechanistic insights regarding disease pathogenesis. To this end, we enrolled 18 cases with COVID-19 and 10 SARS-CoV-2 uninfected controls who 291 underwent upper GI endoscopy (SARS-CoV-2 infected n=16, uninfected n=8), colonoscopy 292 (SARS-CoV-2 infected n=1, uninfected n=1) or both upper endoscopy and colonoscopy (SARS- CoV-2 infected n=1, uninfected n=1) ( swab after the procedure, it was considered to be 0 days from last PCR positive) (Fig. 3a) . COVID-19 symptoms on presentation and treatment regimens were diverse as detailed in The GI mucosa was endoscopically uninflamed in all subjects regardless of the severity brush border in both COVID-19 cases and controls ( Fig. 4a-h) . Additionally, we detected SARS-323 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint CoV-2 nucleocapsid protein in small intestinal enterocytes of COVID-19 patients ( Fig. 4i,n, Supplementary Fig. 7) , but not controls (Fig. 4m,r, Supplementary Fig. 8 337 nm viral particles in the enterocytes of the duodenum and ileum (Fig. 4s-u) . Pleomorphic, 338 spherical structures, morphologically consistent with viral particles were observed in conjunction 339 with small vesicles within enterocytes along the basolateral surface of the enterocytes (Fig. 4s) 340 and blebbing off of the enterocyte apex (Fig. 4u) . Particles with distinct, stalk-like projections 341 (corona) were seen within the enterocyte cytoplasm (Fig. 4t) . Overall, of the 13 patient samples in none of the 12 duodenal biopsies. Altogether these data suggest that SARS-CoV-2 was present 356 at low copy numbers in the GI tract of these early-convalescent COVID-19 cases. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint Next, we performed mass cytometry (CyTOF) based immunophenotypic analyses on the GI 360 tissues of 13 and peripheral blood of 10 COVID-19 cases and 10 controls ( Supplementary Fig. 9 ). GI tissues were divided into lamina propria 362 (LP) and epithelial compartment (EC) fractions and analyzed separately. Immune populations 363 were clustered on the basis of cell-specific markers for both the LP and EC (Fig. 5a,c,g) . While 364 the overall distribution of canonical immune cell subsets in the GI LP were comparable between 365 COVID-19 and control patients (Fig. 5a,b) , few immune populations showed differences as 366 detailed below. Additionally, no clear differences in the LP could be discerned between patients 367 with asymptomatic/mild/moderate disease and those with severe disease (Fig. 5b In the LP, among myeloid cells, CD206 + CD1c + "inflammatory" cDC2 (conventional DCs) 45 370 were reduced in COVID-19 cases compared to controls (0.4-fold decrease, p=0.01). Additionally, plasmacytoid DCs (pDCs) were reduced in COVID-19 cases (0.5 fold decrease, p=0.07) ( Fig. 372 5d,e), analogous to changes described in the peripheral blood of COVID-19 patients 36 . Among 373 other LP populations, effector (PD-1 + CD38 + ) CD4 + and CD8 + T cells were significantly increased 374 in COVID-19 cases compared to the controls (Fig. 5f) , while CD8 + CD103 + T cells (tissue resident 375 memory subset) trended higher in COVID-19 cases compared to controls (1.7-fold increase, 376 p=0.06) (Supplementary Fig. 10a ). Neutrophils (3-fold), eosinophils (1.9-fold), CD4 -CD8 -T cells (3.5-fold) and CD4 + CD103 + T cells (1.8-fold) were increased in COVID-19 cases, but these 378 differences were not statistically significant possibly due to the sample size. We also observed a 379 trend of non-significantly decreased regulatory CD4 + T (TREG) cells (0.6-fold decrease) and 380 increased IgM + plasma cells (2.3-fold increase) in the LP of infected cases vs controls 381 ( Supplementary Fig. 10a ). The distribution of naïve and memory CD4 + and CD8 + T cells was 382 altered in COVID-19 patients, with a reduction of naïve CD4 + T cells and EMRA (effector memory 383 re-expressing RA) CD8 + T cells in LP of COVID-19 patients (Supplementary Fig. 10b ), but this 384 difference did not reach statistical significance. Similar to the LP, the EC showed a reduction of CD206 + cDC2 in COVID-19 cases 386 compared to controls (0.4-fold decrease, p=0.05), while the CD4 -CD8subset of IELs was 387 significantly increased (1.6-fold-increase, p=0.03) (Fig. 5h) . CD8 + T cells, the dominant IEL 388 population, showed an increase (2.6-fold) in COVID-19 cases compared to controls but the 389 difference did not reach statistical significance (p=0.4) ( Supplementary Table 12a ), likely owing 390 to inter-patient variability, also observed by light microscopy. A subset of CD8 + IELs, CD8 + CD69 + 391 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint T cells, showed a non-significant increase in COVID-19 cases vs controls (3.9-fold, p=0.2), while 392 plasma cells were comparable between the cases and controls ( Supplementary Fig. 11 a,b,c) . In the peripheral blood, cell type assignments were carried out using specific cell surface 394 markers (Supplementary Fig. 12a) . We observed that effector (PD-1 + CD38 + ) T cells (for both 395 CD4 + and CD8 + T lymphocytes) were significantly increased in PBMCs of SARS-CoV-2 infected 396 individuals (Fig. 5i) . CD14 -CD16 + inflammatory monocytes trended lower in COVID-19 cases 397 compared to controls (0.4-fold decrease, p=0.09). In contrast, CD14 + CD16 -"classical" monocytes 398 were comparable in COVID-19 cases and controls (p=0.3) (Supplementary Fig. 12b ). Additionally, a non-significant increase in IgG + plasma cells (7.2-fold, p=0.13) and a non- The majority of DEGs were detected in the LP (1061, false discovery rate [FDR] £ 0.05), compared 419 to 12 DEGs in the EC that largely overlapped with the LP (Fig. 6a) . Both LP and EC showed (Fig. 6b) . Downregulation of pathways linked 425 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. We considered the possibility that the observed expression changes could imply IFI44, IFI6 and OAS3 was increased (Fig. 6e) . Next, using gene set enrichment analysis (GSEA), we rank ordered the EC DEGs 457 according to effect size (logFC * -logPvalue) and tested for enrichment in the reported SARS- CoV-2 infected gene signatures 29 (Supplementary Fig. 15a) . The genes upregulated in EC 459 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint showed a significant enrichment in genes upregulated in the SARS-CoV-2 infected intestinal 460 organoid gene datasets. We then carried out Hallmark pathway enrichment analyses on this 461 ranked EC gene list and found that the top two processes associated with genes upregulated in 462 EC were interferon alpha response (normalized enrichment score (NES) 1.91, FDR<0.005) and 463 interferon gamma response (NES=1.8, FDR=0.005) (Supplementary Fig. 15b) . This enrichment 464 is indicative of the host antiviral response against SARS-CoV-2 in human intestines of COVID-19 465 patients. We then evaluated cytokines and chemokines in intestinal samples projecting our RNA-467 seq dataset with published data from the human bronchial epithelial cells infected with SARS- CoV-2 30 . Remarkably, we found that the many of the inflammatory cytokines and chemokines 469 such as IL-1b, IFNg, CCL24 and CXCL8 were downregulated in COVID-19 patients (Fig. 6e) . The to SARS-CoV-2, which in this case is associated with a survival advantage. GI manifestations, reported in nearly half of patients within our discovery cohort were 490 notably higher than some of the early reports 3,7 but similar to more recent studies 8 , likely 491 attributable to greater clinical awareness of GI symptoms and therefore more documentation of 492 such symptoms. A recently published study of 278 patients from a Columbia University Hospital 493 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. The fact that GI symptoms remain significantly associated with better COVID-19 outcomes 504 after adjusting for multiple covariates, known to be associated with reduced survival, in 3 505 independent cohorts totaling 1163 patients, across two different countries enhances the 506 robustness and validity of our findings. Furthermore, our model to predict COVID-19 severity and 507 mortality was enhanced by the inclusion of GI symptoms suggesting that intestinal parameters 508 should be considered in initial assessments and severity stratification of COVID-19 patients. On examination of circulating proteins, including analyses of cytokines and chemokines 525 known to be associated with COVID-19 severity 30,31,35 , we found that multiple 526 cytokines/chemokines were downregulated in patients with GI symptoms compared to patients 527 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. Transcriptome analyses further demonstrated the downregulation of a number of 556 important pro-inflammatory gene products, including IFNG, IL1B, IKBKB and STAT3B, which 557 contribute to TH17 cell differentiation and IBD pathogenesis. Calprotectin, a heterodimer 558 encompassing calgranulin A and calgranulin B, which are encoded by S100A8 and S100A9, 559 respectively, was recently identified as a biomarker of severe COVID-19 disease and suggested 560 to represent a trigger of cytokine release syndrome 81 . We observed no induction of S100A8 or 561 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. the "pro-inflammatory changes" were not observed in the intestines. A separate study by our 593 group that analyzed stool samples from patients going through the acute phase of COVID-19, 594 found little evidence of an active gut inflammatory response. In particular, such stool samples 595 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint strikingly lacked any significant increase of IL-1b, IL-6, TNF-a and IL-10, despite the detection of 596 viral genomes (MEDRXIV/2020/183947). In summary, we have observed an unexpected but significant reduction in COVID-19 598 severity and mortality when patients demonstrate GI symptoms like diarrhea, nausea or vomiting. These data suggest a previously unappreciated tissue-specific response to SARS-CoV-2 and 600 provide the rationale for more studies aimed at determining the mechanisms underpinning the A total of 634 subjects were included in the discovery cohort (Supplementary Figure 1) . In addition to demographic information (including race and ethnicity and primary language), 616 clinical characteristics, laboratory data and outcomes data was extracted from the medical charts. Co-variates that were studied included: history of smoking, BMI (obesity defined as BMI >30) and 618 comorbid conditions including, hypertension, diabetes, chronic lung disease (including asthma 619 and COPD), heart disease (including coronary artery disease, atrial fibrillation and heart failure), 620 chronic kidney disease, cancer, HIV, and inflammatory bowel disease (IBD). (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint ventilation), or Mechanical ventilation AND pressor medications OR creatinine clearance <30 OR 630 new renal replacement therapy OR ALT > 5x upper limit of normal. GI symptoms were defined as more than one episode of either diarrhea, nausea, and/or 632 vomiting at the time of admission. If only one episode of either diarrhea, nausea, and/or vomiting 633 was specifically documented, patients were not considered to have GI symptoms. Additionally, 634 we did not consider GI symptoms that developed during the course of hospitalization, as they 635 could reflect nosocomial or treatment-related effects. Disease severity (as described above) and mortality were considered as outcomes The demographic characteristics of these patients and controls are provided in Fig. 3a, Table 6 , 662 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint patients and 0 otherwise. Significant association based on 95% confidence interval (CI) are 697 reported in Fig. 1c and Supplementary Table. 2. CI of odds ratio were computed based on 1000 698 bootstrap iterations. At each bootstrap iteration, patients were sampled with replacements and 699 logistic regressions were estimated considering as outcome severity and mortality. Then, 95% CI 700 of coefficients and odds ratio were estimated across bootstrap iterations (Fig. 1d) . (Fig. 1e) . For this analysis, we considered 233 patients with clinical data including age, BMI, and GI 715 symptoms. In order to evaluate the predictive performance of each model, bootstrapping was 716 performed. Specifically, at each bootstrap iteration, we randomly sampled patients in the 717 discovery cohort with replacement and estimated a logistic regression to model each outcome as 718 function of a particular GI symptom, age and BMI. In this analysis, only age and BMI were adjusted 719 for since they were the only variables significantly associated with both outcomes across different 720 GI symptoms models in the discovery cohort (Fig. 1c) . Then, the estimated model was utilized to 728 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. The ELLA platform is a method for rapid cytokine measurement using microfluidics ELISA assays. The assay measured TNF-α, IL-6, IL-8, and IL-1β, previously validated by the Mount Sinai Human Consensus clustering was performed using the R packages ConsensusClusterPlus 91 based on 761 z-score normalized data. Specifically, markers were partitioned into six clusters using the Kaplan 762 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint considered in order to derive cluster z-score signatures via package GSVA 92 . Based on these 764 signatures, the association between different clusters and GI symptoms were derived via logistic 765 regression with outcome corresponding to each GI symptom. Fig. 2c shows the signed FDR (-766 log10 scale). P-values were adjusted via Benjamini-Hochberg 90 . Defining associations between GI symptoms and Olink protein markers 769 Associations between GI symptoms and Olink proteomic data were derived using unpaired t-test 770 comparing the symptomatic and asymptomatic groups. P-values were adjusted via Benjamini- Hochberg 90 . Only associations passing a 10% FDR were reported as significant (Fig. 2d) . (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint room temperature for 10 min. Cryovials were immediately transferred to -80 until the sample was 797 acquired for mass cytometry as detailed below. (which was not certified by peer review) 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 September 9, 2020. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint potential cytopathic effect (CPE). Cell culture supernatants were also collected and assessed for presence of infective 866 particles by plaque assay. Briefly, ten-fold serial dilutions were performed in infection media for (which was not certified by peer review) 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 September 9, 2020. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint Table 6 : Clinical characteristics of patients who underwent endoscopic GI biopsies * Age is provided as a range to obscure identifying information related to individuals **Patient 10 was excluded from analysis because SARS-CoV-2 nasopharyngeal PCR testing was repeatedly negative and her COVID-19 antibodies were also negative. Her inclusion initially was based on an atypical pulmonary infiltrate on chest X-ray and a viral syndrome in two of her family members. . Cluster assignment derived based on consensus clustering is reported on the top of the heatmap. b, Boxplot of "Hallmark Inflammatory Response" and "KEGG JAK/STAT Signaling pathway" z-scores stratified by GI symptoms. P-values from unpaired t-tests are reported. "Hallmark Immune Response" and "Hallmark JAK/STAT Signaling" pathways were found significantly enriched at 10% FDR in Cluster 4 and Cluster 5, respectively. c, Association between proteomic clusters and GI symptoms. This association was derived by comparing clusters signatures between asymptomatic and symptomatic groups via t-test. Associations significant at 10% (dark blue) and 15% (light blue) FDR are reported. d, Analytes associated with GI symptoms at 10% FDR based on unpaired t-test. The intensity of the color is proportional to the -log10 p-value. Negative associations are displayed in blue; while positive associations in red. On the right side of the heatmap, the cluster assignment for each marker is reported. c, Boxplot of select differentially expressed markers stratified by GI symptoms. P-values from unpaired t-tests are reported. patients (b,d) and controls (a,c) with ACE2 (green), EPCAM(red) and DAPI (blue) including isotype (e,g) and no primary (f,h) controls. i-r, Immunofluorescent staining of duodenal (i-m) and ileal (n-q) biopsies from COVID-19 patients (i-l, n-q) and controls (m,r) with SARS-CoV-2 nucleocapsid (green), EPCAM (red) and DAPI (blue) including isotype (k,p) and no primary (j,q) controls. s-u, Electron microscopy of duodenal biopsies (s, t) and an ileal biopsy (u) from COVID-19 patients showing viral particles (arrows). Scale bars; 100 µm (a-r), 0.5 µm (s,u), 0.1 µm (t). i. Severe Asymp/mild/mod Control All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. A Novel Coronavirus from Patients with Pneumonia in China A novel coronavirus outbreak of global 1028 health concern Clinical features of patients infected with 2019 novel coronavirus in 1030 Wuhan Clinical Characteristics of Covid-19 in New York City Epidemiological and clinical features of asymptomatic patients with SARS-1034 CoV-2 infection Symptoms of Coronavirus: Centers for Disease 1036 Control and Prevention Epidemiological, clinical and virological characteristics of 74 cases of 1039 coronavirus-infected disease 2019 (COVID-19) with gastrointestinal symptoms Clinical Characteristics of COVID-19 Patients With Digestive Symptoms in 1042 China: A Descriptive, Cross-Sectional, Multicenter Study Extrapulmonary manifestations of COVID-19 Review article: gastrointestinal features in COVID-19 1047 and the possibility of faecal transmission Gastrointestinal Manifestations of SARS-CoV-2 Infection and Virus 1049 Load in Fecal Samples From a Hong Kong Cohort: Systematic Review and Meta-analysis Quantitative mRNA expression profiling 1052 of ACE 2, a novel homologue of angiotensin converting enzyme Intestinal inflammation modulates the expression of ACE2 and 1055 TMPRSS2 and potentially overlaps with the pathogenesis of SARS-CoV-2 related disease Epidemiologic Features and Clinical Course of Patients Infected with 1058 SARS-CoV-2 in Singapore Evidence for Gastrointestinal Infection of SARS-CoV-2 First Case of 2019 Novel Coronavirus in the United States Prolonged presence of SARS-CoV-2 viral RNA in faecal samples An overview of their replication and 1066 pathogenesis Enteric involvement of severe acute respiratory syndrome-associated 1068 coronavirus infection Detection of SARS coronavirus in patients with suspected SARS Epidemiological, demographic, and clinical characteristics of 47 cases of 1072 Middle East respiratory syndrome coronavirus disease from Saudi Arabia: a descriptive 1073 study Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-1075 PCR Aberrant gut microbiota alters host metabolome and impacts renal 1077 failure in humans and rodents AGA Institute Rapid Review of the GI and Liver Manifestations of COVID-1079 19, Meta-Analysis of International Data, and Recommendations for the Consultative 1080 Management of Patients with COVID-19 Respiratory disease in rhesus macaques inoculated with SARS-CoV-1082 2 Proteomics of SARS-CoV-2-infected host cells reveals therapy targets TMPRSS2 and TMPRSS4 promote SARS-CoV-2 infection of human small 1086 intestinal enterocytes Infection of bat and human intestinal organoids by SARS-CoV-2 SARS-CoV-2 productively infects human gut enterocytes Imbalanced Host Response to SARS-CoV-2 Drives Development of 1092 COVID-19 Longitudinal analyses reveal immunological misfiring in severe COVID-19 Analysis of inflammatory parameters and disease severity for 88 1096 hospitalized COVID-19 patients in Wuhan, China Dynamic Interleukin-6 Level Changes as a Prognostic Indicator in Patients 1098 With COVID-19 The role of cytokine profile and lymphocyte subsets in the severity of 1100 coronavirus disease 2019 (COVID-19): A systematic review and meta-analysis An inflammatory cytokine signature helps predict COVID-19 1103 severity and death. medRxiv A dynamic COVID-19 immune signature includes associations with 1105 poor prognosis Deep immune profiling of COVID-19 patients reveals distinct 1107 immunotypes with therapeutic implications Longitudinal characteristics of lymphocyte responses and cytokine profiles 1109 in the peripheral blood of SARS-CoV-2 infected patients Systems biological assessment of immunity to mild versus 1112 severe COVID-19 infection in humans Presenting Characteristics, Comorbidities, and Outcomes Among 1114 5700 Patients Hospitalized With COVID-19 in the New York City Area Following TRAIL's path in the immune 1119 system TRAIL-Mediated Suppression of T 1121 Cell Receptor Signaling Inhibits T Cell Activation and Harnessing the biology of IL-7 for therapeutic 1124 application Single-Cell Analysis of Crohn's Disease Lesions Identifies a Pathogenic 1126 Cellular Module Associated with Resistance to Anti-TNF Therapy Intra-and Inter-cellular Rewiring of the Human Colon during 1129 A novel role for constitutively expressed epithelial-derived 1131 chemokines as antibacterial peptides in the intestinal mucosa Tissue distribution of ACE2 protein, the functional receptor for SARS 1134 coronavirus. A first step in understanding SARS pathogenesis Expression of the SARS-CoV-2 cell receptor gene 1137 ACE2 in a wide variety of human tissues Gastrointestinal Symptoms and Coronavirus Disease 2019: A Case-1139 Control Study From the United States COVID-19 in pediatric patients: a case series from the Bronx, 1141 NY Multisystem Inflammatory Syndrome in Children (MIS-C) Associated with COVID. AJR 1144 COVID-19 Digestive System Involvement and Clinical Outcomes in a 1146 Large Academic Hospital in Gastrointestinal symptoms of 95 cases with SARS-CoV-2 infection COVID-19: consider cytokine storm syndromes and 1150 immunosuppression High-Dose Intravenous Immunoglobulin as a Therapeutic Option for 1152 Deteriorating Patients With Coronavirus Disease Interleukin-1 blockade with high-dose anakinra in patients with COVID-1155 19, acute respiratory distress syndrome, and hyperinflammation: a retrospective cohort 1156 study The origin, transmission and clinical therapies on coronavirus disease 1158 2019 (COVID-19) outbreak -an update on the status Pulmonary Vascular Endothelialitis, Thrombosis, and Angiogenesis 1160 in Covid-19 Histopathologic Changes and SARS-CoV-2 Immunostaining in the Lung of 1162 a Patient With COVID-19 Kidney disease is associated with in-hospital death of patients with 1164 COVID-19 Acute kidney injury in critically ill patients with COVID-19 Endothelial cell infection and endotheliitis in COVID-19 Single-cell landscape of bronchoalveolar immune cells in patients with 1170 COVID-19 Histopathology and ultrastructural findings of fatal COVID-19 1172 infections in Washington State: a case series Structure and signalling in the IL-17 receptor family Targeting IL-17 and TH17 cells in chronic inflammation TH17 responses in cytokine storm of COVID-19: An emerging target 1178 of JAK2 inhibitor Fedratinib Cell host response to infection with novel human coronavirus EMC 1180 predicts potential antivirals and important differences with SARS coronavirus Distinct immune response in two MERS-CoV-infected patients: can we go 1183 from bench to bedside? 1186 72. Pan, J., et al. A novel chemokine ligand for CCR10 and CCR3 expressed by epithelial cells 1187 in mucosal tissues Differential effects of interleukin-15 1189 (IL-15) and IL-2 on human neutrophils: modulation of phagocytosis, cytoskeleton 1190 rearrangement, gene expression, and apoptosis by IL-15 Airway epithelial IL-15 transforms monocytes into dendritic cells IL-15 is expressed by dendritic cells 1194 in response to type I IFN, double-stranded RNA, or lipopolysaccharide and promotes 1195 dendritic cell activation IL-10: A Multifunctional Cytokine in Viral 1197 Infections Profiling serum cytokines in COVID-19 patients reveals IL-6 and IL-10 are 1199 disease severity predictors Reduction and Functional Exhaustion of T Cells in Patients With 1201 Coronavirus Disease 2019 (COVID-19) Association of Interleukin 7 Immunotherapy With Lymphocyte Counts 1203 Among Patients With Severe Coronavirus Disease 2019 (COVID-19) TRAIL and its receptors in the colonic epithelium: a putative role in the 1206 defense of viral infections Neutrophil calprotectin identifies severe pulmonary disease in COVID-19. 1208 medRxiv Comparative replication and immune activation profiles of SARS-CoV-2 1210 and SARS-CoV in human lungs: an ex vivo study with implications for the pathogenesis 1211 of COVID-19 Host-Viral Infection Maps Reveal Signatures of Severe COVID-19 Patients Elevated serum levels of S100A8/A9 and HMGB1 at hospital admission 1215 are correlated with inferior clinical outcomes in COVID-19 patients The mucus and mucins of the goblet cells and enterocytes provide 1218 the first defense line of the gastrointestinal tract and interact with the immune system Proinflammatory cytokines in the pathogenesis of inflammatory 1221 bowel diseases The immunology of mucosal models of 1223 inflammation Intestinal 1225 mucosal tolerance and impact of gut microbiota to mucosal tolerance Mapping Systemic Inflammation and Antibody Responses in 1228 Controlling the False Discovery Rate: A Practical and 1230 Powerful Approach to Multiple Testing ConsensusClusterPlus: a class discovery tool with 1233 confidence assessments and item tracking GSVA: gene set variation analysis for 1235 microarray and RNA-seq data A Streamlined CyTOF Workflow To Facilitate Standardized Multi-Site 1237 Immune Profiling of COVID-19 Patients. medRxiv Integrated Transcriptome and Network Analysis Reveals 1239 Spatiotemporal Dynamics of Calvarial Suturogenesis edgeR: a Bioconductor package for 1241 differential expression analysis of digital gene expression data RSEM: accurate transcript quantification from RNA-Seq data with or 1244 without a reference genome A scaling normalization method for differential expression 1246 analysis of RNA-seq data Profiler--a web-based toolset 1248 for functional profiling of gene lists from large-scale experiments The histograms shows the decrease in the expression of CD206 and CD123 in dendritic cell populations of COVID-19 patients (red) compared to the controls (blue). e, Unsupervised analyses of immune cells: The bar plots show the relative frequencies of different myeloid and lymphocytic immune populations namely CD206+ cDC2 and plasmacytoid DCs in lamina propria of controls (blue) and COVID-19 patients (red).. f, Supervised analyses: The bar graphs depict alterations in PD-1+ CD38+ (exhausted) CD4+ and CD8+ T cell subsets in lamina propria of controls (blue) and COVID-19 patients (red). g, Uniform manifold approximation and projection (UMAP) intestinal biopsies. h, Unsupervised analyses of immune cells: The bar plots show the relative frequencies of CD206+ cDC2 dendritic cells and CD4-CD8-T cells in the intraepithelial compartment of controls (blue) and COVID-19 patients (red). i, Supervised analyses on PBMCs indicating immune cell changes in blood : The bar graphs depict alterations in PD-1+ No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity instrument on a NovaSeq S1 Flowcell, with an average yield of 39 million PE reads/sample. (which was not certified by peer review) 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 September 9, 2020. All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint Biopsy specimens for electron microscopy were placed in 3% buffered glutaraldehyde. Following (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. (which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint 1259 1260 All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 September 9, 2020. . https://doi.org/10.1101/2020.09.07.20187666 doi: medRxiv preprint