Longitudinal proteomic profiling of dialysis patients with COVID-19 reveals markers of severity and predictors of death

Abstract

End-stage kidney disease (ESKD) patients are at high risk of severe COVID-19. We performed dense serial blood sampling in hospitalised and non-hospitalised ESKD patients with COVID-19 (n=256 samples from 55 patients) and used Olink immunoassays to measure 436 circulating proteins. Comparison to 51 non-infected ESKD patients revealed 221 proteins differentially expressed in COVID-19, of which 69.7% replicated in an independent cohort of 46 COVID-19 patients. 203 proteins were associated with clinical severity scores, including IL6, markers of monocyte recruitment (e.g. CCL2, CCL7), neutrophil activation (e.g proteinase-3) and epithelial injury (e.g. KRT19). Random Forests machine learning identified predictors of current or future severity such as KRT19, PARP1, PADI2, CCL7, and IL1RL1 (ST2). Survival analysis with joint models revealed 69 predictors of death including IL22RA1, CCL28, and the neutrophil-derived chemotaxin AZU1 (Azurocidin). Finally, longitudinal modelling with linear mixed models uncovered 32 proteins that display different temporal profiles in severe versus non-severe disease, including integrins and adhesion molecules. Our findings point to aberrant innate immune activation and leucocyte-endothelial interactions as central to the pathology of severe COVID-19. The data from this unique cohort of high-risk individuals provide a valuable resource for identifying drug targets in COVID-19.

Mendelian randomisation identifies alternative splicing of the FAS death receptor as a mediator of severe COVID-19

“…These data suggest that genetic propensity to higher soluble FAS levels influences COVID-19 severity, but not susceptibility. Observational data from the analysis of Gisby et al 11 revealed a similar pattern. Plasma FAS was not significantly differentially abundant in the comparison of all COVID-19 cases versus uninfected controls (Benjamini-Hochberg adjusted P value 0.280), but it was highly significantly associated with COVID-19 severity grading within-cases (Benjamini-Hochberg adjusted P 0.019) ( Fig.…”

“…We first identified a list of proteins associated with COVID-19 clinical severity grading by examining two studies that performed broad proteomic profiling of COVID-19 patient plasma samples using the Olink proteomics platform 10 , 11 . We took the lists of severity-associated proteins from each study (Benjamini-Hochberg adjusted P <0.05) and intersected these to provide a high-confidence list of 157 severity-associated proteins ( Fig.…”

“…We identified plasma proteins that were significantly associated (5% FDR) with COVID-19 severity in the two studies that used Olink proteomics platform (Filbin et al 10 and Gisby et al 11 ). To define a set of robust COVID-19 biomarkers, we took the intersect of the lists of significant associations from these two studies.…”

“…We used data on COVID-19 patients and sex, age and ethnicity matched non-infected controls from the study by Gisby et al 11 . The study design involved serial plasma sampling from the COVID-19 patients.…”

“…A wide range of therapies directed at specific elements of the inflammatory response have been developed for autoimmune and inflammatory diseases 8 , 9 , and present potential repurposing opportunities for treatment of COVID-19. Profiling of plasma proteins in COVID-19 patients has revealed a signature of innate immune cell activation (including upregulation of IL-6, monocyte chemokines and neutrophil proteins) and epithelial/endothelial injury in severe disease 10 , 11 . A limitation of such observational studies, however, is that they cannot distinguish causal mediators of immunopathology from secondary downstream consequences of inflammation and/or tissue injury.…”