ABSTRACT
INTRODUCTION: Complex critical syndromes like sepsis and COVID-19 may be composed of underlying subclasses, or 'endotypes,' which may respond differently to treatment. We previously reported the discovery and validation of a 33-mRNA host response classifier which defined three sepsis endotypes across 1,300 patients with bacterial sepsis at hospital or ICU admission. Here, we aimed to test whether our 33-mRNA bacterial sepsis endotypes classifier recapitulates the same clinical and immunological endotypes in COVID-19. METHODS: In this prospective, single-center observational cohort study, we recruited adult patients with RT-PCRconfirmed COVID-19 within 24 hours of admission to an Athens, Greece hospital. RNA was extracted from whole blood collected in PAXgene RNA tubes, and then profiled on the NanoString nCounter® platform to quantify the 33 mRNAs. The endotypes classifier then assigned one of three endotypes (Inflammopathic, Adaptive, or Coagulopathic) to each patient. We tested endotype status against other clinical parameters including lab values, severity scores, and outcomes. RESULTS: We enrolled 71 patients with COVID-19, of which 33 went on to severe respiratory failure (SRF), of which 6 (8%) died. Patients were assigned as Inflammopathic (34%), Adaptive (39%), or Coagulopathic (27%);Adaptive patients had lower rates of SRF and no mortalities. Coagulopathic and Inflammopathic endotypes had 12% and 16% mortality rates. The Coagulopathic group was significantly associated with D-dimers, and the Inflammopathic group showed high clinical severity and highest C-reactive protein and IL-6 levels. CONCLUSIONS: Our predefined 33-mRNA endotypes classifier recapitulated immune phenotypes in viral sepsis (COVID-19) despite its prior training and validation only in bacterial sepsis. Further work should focus on continued validation of the endotypes and their interaction with immunomodulatory therapy. If confirmed with future studies, the 33-mRNA classifer could be used as a companiondiagnostic test to guide a precision-medicine-based intervention.
ABSTRACT
Background: COVID-19 is a pandemic caused by the SARS-CoV-2 virus that shares and differs in clinical characteristics of known viral infections. Methods: We obtained RNAseq profiles of 62 prospectively enrolled COVID-19 patients and 24 healthy controls (HC). We collected 23 independent studies profiling 1,855 blood samples from patients covering six viruses (influenza, RSV, HRV, Ebola, Dengue and SARS-CoV-1). We studied host whole-blood transcriptomic responses in COVID-19 compared to non-COVID-19 viral infections to understand similarities and differences in host response. Gene signature threshold was absolute effect size ≥1, FDR ≤ 0.05%. Results: Differential gene expression of COVID-19 vs HC are highly correlated with non-COVID-19 vs HC (r=0.74, p< 0.001). We discovered two gene signatures: COVID-19 vs HC (2002 genes) (COVIDsig) and non-COVID-19 vs HC (635 genes) (nonCOVIDsig). Pathway analysis of over-expressed signature genes in COVIDsig or nonCOVIDsig identified similar pathways including neutrophil activation, innate immune response, immune response to viral infection and cytokine production. Conversely, for under-expressed genes, pathways indicated repression of lymphocyte differentiation and activation (Fig1). Intersecting the two gene signatures found two genes significantly oppositely regulated (ACO1, ATL3). We derived a third gene signature using COCONUT to compare COVID-19 to non-COVID-19 viral infections (416 genes) (Fig2). Pathway analysis did not result in significant enrichment, suggesting identification of novel biology (Fig1). Statistical deconvolution of bulk transcriptomic data found M1 macrophages, plasmacytoid dendritic cells, CD14+ monocytes, CD4+ T cells and total B cells changed in the same direction across COVID-19 and non-COVID-19 infections. Cell types that increased in COVID-19 relative to non-COVID-19 were CD56bright NK cells, M2 macrophages and total NK cells. Those that decreased in non- COVID-19 relative to COVID-19 were CD56dim NK cells & memory B cells and eosinophils (Fig3). Conclusion: The concordant and discordant responses mapped here provide a window to explore the pathophysiology of COVID-19 vs other viral infections and show clear differences in signaling pathways and cellularity as part of the host response to SARS-CoV-2.