Longitudinal proteomic profiling of high-risk patients with COVID-19
Pre-print now available: Markers of severity and predictors of fatal disease identified in high-risk patients with COVID-19
End-stage kidney disease (ESKD) patients are at high risk of severe COVID-19, and in this study, longitudinal protein profiling with five different Olink® Target 96 panels was used to measure 436 circulating proteins in 55 hospitalized/non-hospitalized ESKD patients with COVID-19, as well as 51 non-infected ESKD controls. Results were then verified in an independent cohort of 46 patients.
One unique aspect of this study was the dense serial sampling of subjects in both the inpatient and outpatient setting, enabling the monitoring of intra-individual changes in circulating proteins as patients moved from early disease to clinical deterioration. This allowed not only the identification of potential biomarkers characteristic for different disease stages, but also the examination of temporal patterns of protein expression that were associated with severe versus non-severe COVID-19. Some of the main conclusions were:
221 proteins differentially expressed in ESKD patients with COVID-19 (with consistent results in the independent cohort)
203 proteins associated with clinical severity scores
– included IL-6 and multiple proteins involved in monocyte recruitment, neutrophil activation and epithelial injury
Machine learning analysis identified predictors of current or future severity, as well as predictors of death
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
Overall, these 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 also provide a valuable resource for identifying drug targets in COVID-19. Exemplifying the spirit of collaboration and openness that has become ever more apparent during the course of the pandemic, the authors of this study have posted the manuscript directly on medRxiv and made the raw protein data freely available to the wider research community (see below).
Citation and data access
Gisby J, Clarke C, Medjeral-Thomas N, et al., Longitudinal proteomic profiling of high-risk patients with COVID-19 reveals markers of severity and predictors of fatal disease (2020) medRxiv DOI: 10.1101/2020.11.05.20223289
What the authors say
Corresponding author on the manuscript, Dr. James Peters from Imperial College, had the following to say about the study: “A unique aspect of this study was the ability to perform serial proteomic profiling with Olinkimmunoassays in both the inpatient and outpatient setting in a high-risk cohort of patients. This longitudinal aspect of the study enabled us to capture intra-individual changes in circulating proteins as patients move from early disease to clinical deterioration. As a result, we were able to identify protein predictors of fatal disease. In addition, our data shed light on the pathobiology of severe COVID-19, revealing a signature of monocyte recruitment, neutrophil activation, integrin-mediated cell adhesion abnormalities, and epithelial/endothelial injury. We believe that our results indicate that epithelial/endothelial breakdown is a key feature of severe COVID-19 and explains the blood clots seen in these patients. We hope that the data will lead to the identification of new therapies for severe COVID-19”
About the author and his work
Dr. James Peters is a UKRI Innovation Fellow at Health Data Research UK and Clinical Reader in Rheumatology, and moved to Imperial College as a Group Leader in 2019.
The research from Dr. Peters’ team lies at the interface of molecular epidemiology, computational biology and clinical medicine. The focus is on understanding the molecular and cellular basis of immune-mediated inflammatory diseases using a combination of genomics, transcriptomics and proteomics. A wide range of techniques are employed, including longitudinal molecular profiling, polygenic risk scores, and Mendelian randomisation, leveraging both in-house datasets and publically available resources (e.g. GWAS summary statistics, UK Biobank).