Computational analysis of TMPRSS2 expression in normal and SARS-CoV-2-infected human tissues

Abstract

The transmembrane serine protease 2 (TMPRSS2) is a key molecule for SARS-CoV-2 invading human host cells. To provide insights into SARS-CoV-2 infection of various human tissues and understand the potential mechanism of SARS-CoV-2 infection, we investigated TMPRSS2 expression in various normal human tissues and SARS-CoV-2-infected human tissues. Using publicly available datasets, we performed computational analyses of TMPRSS2 expression levels in 30 normal human tissues, and compared them between males and females and between younger (ages ≤ 49 years) and older (ages > 49 years) populations in these tissues. We found that TMPRSS2 expression levels were the highest in the prostate, stomach, pancreas, lungs, small intestine, and salivary gland. The TMPRSS2 protein had relatively high expression levels in the parathyroid gland, stomach, small intestine, pancreas, kidneys, seminal vesicle, epididymis, and prostate. However, TMPRSS2 expression levels were not significantly different between females and males or between younger and older populations in these tissues. The pathways enriched in TMPRSS2-upregulated pan-tissue were mainly involved in immune, metabolism, cell growth and proliferation, stromal signatures, and cancer and other diseases. Many cytokine genes displayed positive expression correlations with TMPRSS2 in pan-tissue, including TNF-α, IL-1, IL-2, IL-4, IL-7, IL-8, IL-12, IL-18, IFN-α, MCP-1, G-CSF, and IP-10. We further analyzed TMPRSS2 expression levels in nasopharyngeal swabs from SARS-CoV-2-infected patients. TMPRSS2 expression levels showed no significant difference between males and females or between younger and older patients. However, they were significantly lower in SARS-CoV-2-infected than in healthy individuals and patients with other viral acute respiratory illnesses. Interestingly, TMPRSS2 expression levels were positively correlated with the enrichment levels of four immune signatures (B cells, CD8⁺ T cells, NK cells, and interferon response) in SARS-CoV-2-infected patients but likely to be negatively correlated with them in the normal lung tissue. Our data may provide insights into the mechanism of SARS-CoV-2 infection.

Keywords: Gene co-expression network; Gene expression profiles; Immune signatures; Pathway and gene ontology; SARS-CoV-2; Transmembrane serine protease 2.

Fig. 1

TMPRSS2 expression in various human tissues. (A) Comparison of TMPRSS2 expression levels across 30 human tissues in GTEx [13]. Comparison of TMPRSS2 expression levels between males and females (B) and between older (ages > 49 years) and younger (ages ≤ 49 years) (C) in individual human tissues in GTEx. Two-tailed Student's t-test was used in (B) and (C). The adjusted P value was calculated by the Benjamini and Hochberg method [17]. TMPRSS2: Transmembrane serine protease 2; GTEx: Genotype-Tissue Expression.

Fig. 2

Association between TMPRSS2 expression and immune signatures in various human tissues. Correlation between TMPRSS2 expression levels and immune signature enrichment levels in various human tissues in males and females (A) and in older (ages > 49 years) and younger (ages ≤ 49 years) populations (B). Pearson correlation coefficients (r) and P values are shown in (A, B). *P < 0.05, **P < 0.01, and ***P < 0.001. NK: natural killer.

Fig. 3

Pathways and gene ontology associated with TMPRSS2 expression in pan-tissue. (A) The KEGG pathways enriched in TMPRSS2-high pan-tissue which were identified by GSEA [15]. (B) 12 gene modules and their representative gene ontology differentially enriched between TMPRSS2-high and TMPRSS2-low pan-tissue, which were identified by WGCNA [16]. The P values are shown in parenthesis.

Fig. 4

Genes having significant expression correlation with TMPRSS2 in pan-tissue. (A) Four KEGG pathways associated with 86 genes having strong expression correlations with TMPRSS2 (Pearson correlation coefficient (r) > 0.7). (B) 23 genes having the strongest positive expression correlations with TMPRSS2 (r > 0.8). (C) Positive expression correlation between TMPRSS2 and 14 genes encoding inflammatory cytokines. Pearson correlation coefficients (r) and P values are shown. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 5

TMPRSS2 expression in SARS-CoV-2-infected human tissues. (A) Comparisons of TMPRSS2 expression levels between males and females and between older (ages > 49 years) and younger (ages ≤ 49 years) in SARS-CoV-2-infected human tissues. (B) Comparisons of TMPRSS2 expression levels between SARS-CoV-2-infected and normal and other-viruses-infected human tissues. Two-tailed Student's t-test P values are indicated in (A, B). FC: fold change. (C) Pearson correlation between TMPRSS2 expression levels and immune signature enrichment levels in SARS-CoV-2-infected human tissues. *P < 0.05, **P < 0.01, and ***P < 0.001. (D) Expression correlation between TMPRSS2 and androgen receptor gene AR in SARS-CoV-2-infected human tissues and in normal human pan-tissue. Pearson correlation coefficients (r) and P values are shown.