Evaluating the Immune Response in Treatment-Naive Hospitalised Patients With Influenza and COVID-19

Abstract

The worldwide COVID-19 pandemic has claimed millions of lives and has had a profound effect on global life. Understanding the body's immune response to SARS-CoV-2 infection is crucial in improving patient management and prognosis. In this study we compared influenza and SARS-CoV-2 infected patient cohorts to identify distinct blood transcript abundances and cellular composition to better understand the natural immune response associated with COVID-19, compared to another viral infection being influenza, and identify a prognostic signature of COVID-19 patient outcome. Clinical characteristics and peripheral blood were acquired upon hospital admission from two well characterised cohorts, a cohort of 88 patients infected with influenza and a cohort of 80 patients infected with SARS-CoV-2 during the first wave of the pandemic and prior to availability of COVID-19 treatments and vaccines. Gene transcript abundances, enriched pathways and cellular composition were compared between cohorts using RNA-seq. A genetic signature between COVID-19 survivors and non-survivors was assessed as a prognostic predictor of COVID-19 outcome. Contrasting immune responses were detected with an innate response elevated in influenza and an adaptive response elevated in COVID-19. Additionally ribosomal, mitochondrial oxidative stress and interferon signalling pathways differentiated the cohorts. An adaptive immune response was associated with COVID-19 survival, while an inflammatory response predicted death. A prognostic transcript signature, associated with circulating immunoglobulins, nucleosome assembly, cytokine production and T cell activation, was able to stratify COVID-19 patients likely to survive or die. This study provides a unique insight into the immune responses of treatment naïve patients with influenza or COVID-19. The comparison of immune response between COVID-19 survivors and non-survivors enables prognostication of COVID-19 patients and may suggest potential therapeutic strategies to improve survival.

Keywords: COVID-19; adaptive; blood; immune response; influenza; innate; survival; transcriptome.

Figure 1

Top 12 clusters identified with BioLayout. (A) Enrichment of gene clusters in blood of patients with influenza (annotated in red) and COVID-19 (annotated in blue). Increased abundances of gene transcripts in influenza patients are involved with an innate immune response, while in COVID-19 clusters are involved with an adaptive immune response, blood coagulation and neutrophil degranulation. (B) After TMM normalisation a significant difference in gene clusters between patients with influenza or COVID-19 was detected. The abundance of gene transcripts involved with an innate immune response and plasmacytoid dendritic cell were observed to be higher in influenza patients. In contrast, the abundance of gene transcripts involved with an adaptive immune response and neutrophil degranulation was higher in COVID-19 patients.

Figure 2

Differences in immune response indicated by predicted cell types in patients with COVID-19, who either survived or died, and patients with influenza. (A) M0 macrophages, resting natural killer (NK) cells, plasma cells, cytotoxic CD8+ T cells and regulatory T cells were found to be significantly higher in COVID-19 patients. In influenza patients a significantly higher proportion of activated dendritic cells was detected. (B) A statistically significant higher count of neutrophils in COVID-19 patients who died after 30 days indicating the presence of an elevated innate immune response. While an adaptive immune response was detected in COVID-19 survivors as can be seen by the statistically significant higher count of naïve B cells, and CD4+ and CD8+ T cells.

Figure 3

Adaptive immune response associated with COVID-19 and a positive patient outcome. Volcano plots (A) between patients with COVID-19 or influenza and (B) between COVID-19 survivors and non-survivors, threshold criteria used FDR < 0.05 and log2 fold change < -1 or >1, transcript which met criteria were used for enrichment analysis with ToppGene. (C) Enrichment analysis of the transcripts with an increased abundance in patients with COVID-19 identified an increased adaptive immune response which was also detected in (D) patients with COVID-19 who were still alive 30 days after hospital admission. (E) Increased innate immune response in patients who died of COVID-19 after 30 days of hospital admission. Percentage in annotation is the ratio of the input query genes overlapping with the genes in the pathway annotation.

Figure 4

Receiver Operating Characteristic (ROC) curves showing prediction accuracy COVID-19 survivors and non-survivors. (A) Genes identified with EdgeR and gene co-expression analysis and used for subsequent modelling. (B) ROC curves according to the three models used [Boosted Logistic Regression (LogitBoost), Bayesian Generalised Linear (Bayesglm) and RandomForest (rf)]. (C) In total three different models were used [RandomForest (rf), Boosted Logistic Regression (LogitBoost) and Bayesian Generalised Linear (Bayesglm)]. The 47 genes identified with gene co-expression and differential gene expression analysis were used as input. The highest sensitivity obtained was 75% and for specificity 93%.