To obtain a comprehensive understanding of the impact of SARS-CoV-2 infection on patients with severe COVID-19 with at least 3 negative virus tests, we analysed the transcriptional profiles of whole blood cells via RNA-Seq analysis. Five healthy volunteers and 14 patients with severe COVID-19 were enrolled in this study (the characteristics of the patients are listed in Table 1). The patients were enrolled in 2 batches at two different time points (April 2020 and May 2020 separately). The first batch (including 5 patients) and the second batch (including 9 patients) both came from the East Campus of Renmin Hospital of Wuhan University. The common feature of these two batches of patients was at least three consecutive negative SARS-CoV-2 virus tests.
Compared with the healthy donors, the severe COVID-19 patients exhibited 33788 upregulated genes, 1007 downregulated genes, and 20347 genes that remained unchanged. A volcano map (Supplementary Fig. 1) was used to visually display the differentially expressed gene (DEG) distribution between the patients and the healthy donors. We used Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses to investigate gene enrichment. The most significantly enriched GO terms were "detection of stimulus involved in sensory perception" and "sensory perception of smell" (Fig. 1A). The most significantly enriched KEGG terms were "olfactory transduction", "neuroactive ligand-receptor interaction", and "taste transduction" (Fig. 1B). Most of the genes described in those terms overlapped and were classified as genes controlling sensory functions, mainly smell and taste. The expression levels of these genes varied among the different batches of patients. In the first batch of patients, only one patient showed abnormally increased expression of the smell- (Fig. 2A) and taste-related (Fig. 2B) genes; however, in the second batch, more than half of the patients showed obviously elevated expression of these genes. The results indicate that patients infected with SARS-CoV-2 at different periods have significantly different symptoms regarding the genes controlling smell and taste functions.
Changes in the expression of genes related to “cytokine storm” and inflammatory responses
Sepsis is defined as life-threatening organ dysfunction caused by a dysregulated host response to infection, including bacterial, fungal, parasitic, or viral infections (The Third International Consensus Definitions for Sepsis and Septic Shock, Sepsis-3) [13]. Since SARS-CoV-2 is an infectious viral pathogen, it is reasonable to consider that severe COVID-19 is a subtype of sepsis [14, 15]. The “cytokine storm” and inflammatory-related responses are the main focusses of sepsis studies. However, our RNA-Seq results showed that in the patients with severe COVID-19, there were large individual differences in the expression of inflammatory factors (Fig. 3A), chemokines (Fig. 3B), and adhesive molecules (Fig. 3C). Patient A5 from batch 1 and patients B1, B2, B3, B4, and B8 from batch 2 showed significant upregulation of these genes, whereas no obvious changes were found in patients A1, A2, A3, and A4 from batch 1 and patients B5, B6, B7, and B9 from batch 2 (Fig. 3A-C). In addition, the expression levels of cytokines, chemokines, and adhesive molecules showed no significant difference between the survival and death groups, which is in line with recent COVID-19 studies [16, 17] (Supplementary Fig. 2–4). According to the results, in those patients with severe COVID-19 but negative viral tests, there were large individual differences in inflammatory responses, and there was no obvious relationship between the inflammatory responses and severity and mortality.
The adaptive immune function of all the patients was severely impaired
Compared with the number of upregulated genes (33,788), only 1,007 genes were downregulated. Among them, many genes were related to immune functions. According to the KEGG enrichment analysis results, there were 10 significantly changed KEGG pathways (p-adj < 0.05), and 4 of them were related to immune responses, namely, “primary immunodeficiency”, “T cell receptor signalling pathway”, “haematopoietic cell lineage”, and “Th1 and Th2 cell differentiation”. Seven genes were shared by these 4 categories: CD3D, CD3E, CD3G, CD4, CD8A, CD8B, and CD40LG. Compared with those in the healthy volunteers, the expression levels of these genes in the patients were dramatically decreased (Fig. 4A). In addition, genes related to B cell function, including CD5, CD19, CD20, CD21, CD22, CD23, CD79a, and CD79b, were also significantly downregulated (Fig. 4B).
The major histocompatibility complex (MHC) is a collection of genes that code for MHC molecules found on the surface of all nucleated cells [18, 19]. In humans, MHC genes are referred to as human leukocyte antigen (HLA) genes. MHC molecules play an important role in the antigen presentation process, which is a key step in the activation of the adaptive immune response. As shown in Fig. 4C, the expression of multiple HLA genes, in addition to the most reported HLA-DR gene, was decreased to a very low level.
Antigen-induced signals transferred from antigen-presenting cells by MHC molecules to the T cell receptor (TCR) alone are insufficient to activate T cells. Costimulatory and coinhibitory receptors play a pivotal role in T cell activation or inhibition, as they determine the functional outcome of TCR signalling [20, 21]. The expression of four classic costimulatory molecules, namely, CD27, CD28, CD40LG, and TNFRSR25, was reduced to a very low level (Fig. 4D). However, the expression of other costimulatory molecules (Supplementary Fig. 5A) as well as coinhibitory molecules [22], such as CTLA-4, PD-1, and PD-L1 (Supplementary Fig. 5B), did not show a significant increasing or decreasing trend.
Together, these results strongly suggest that the adaptive immune response is still severely and universally impaired in patients with severe COVID-19 who have negative viral tests and that this immunosuppression is not due to the upregulation of classic immune checkpoint molecules such as PD-1.