TNFα-Antagonist Use and Mucosal Inflammation Are Associated with Increased Intestinal Expression of SARS-CoV-2 Host Protease TMPRSS2 in Patients with Inflammatory Bowel Disease.

Dear Editors:

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) viral transfection is dependent on angiotensin-converting enzyme 2 (ACE2) and transmembrane serine protease 2 (TMPRSS2), and increased intestinal receptor expression might support viral replication.1, 2, 3 Therefore, we read with great interest the work by Krzysztof et al, reporting that anatomic location, intestinal inflammation, and age are key determinants of intestinal expression of ACE2 and TMPRSS2 in patients with inflammatory bowel disease (IBD). However, several additional questions remain as the authors themselves also describe. It is important to understand the effects of immunomodulating drugs on ACE2 and TMPRSS2 expression, and the cell-type–specific context in which these genes are expressed. In addition, it is important to understand the genetic determinants of intestinal ACE2 and TMPRSS2 expression. To address these remaining questions, we queried host genetic data and RNA sequencing data of intestinal biopsies from patients with IBD. We independently confirm that anatomic location, intestinal inflammation, and age are determinants of intestinal ACE2 and TMPRSS2 expression. Furthermore, we demonstrate that immunomodulating drugs are associated with increased intestinal TMPRSS2 expression, and highlight the cell-type–specific context of ACE2 and TMPRSS2 expression.

We analyzed bulk RNA sequencing data of 92 ileal and 199 colonic snap frozen mucosal biopsies from 168 patients with IBD, as a part of the 1000IBD project. Genotypes were obtained using both imputed Global Screening Array data and whole exome sequencing. We analyzed clinical data regarding the use of immunomodulating drugs, age, sex, diagnosis (Crohn’s disease, ulcerative colitis, or unclassified IBD) and body mass index (BMI). Multivariate linear mixed regression analyses were performed to assess the effects of these clinical factors on the intestinal gene expression levels of ACE2 and TMPRSS2 (R v.3.6.0). In addition, we assessed the effects of host genetic variation on gene expression (cis-expression quantitative trait loci). All of these clinical factors, as well as sequencing batch, were included as covariates. We performed deconvolution analyses on bulk RNA sequencing data to assess cell-type–specific expression of ACE2 and TMPRSS2. We validated this approach using single-cell RNA sequencing data from an independent set of 18 colonic biopsies from 11 patients with ulcerative colitis (unpublished data; 1000IBD cohort; 2020).

First, we replicated the effects of anatomic location, intestinal inflammation, and age on intestinal expression of ACE2 and TMPRSS2. The expression of ACE2 was higher, and TMPRSS2 was lower in ileum compared to colon (P < 2.2 × 10–16). Moreover, the expression of ACE2 was lower and the expression of TMPRSS2 was higher in inflamed ileum compared with a noninflamed ileum, independent of medication use, age, sex, diagnosis, and BMI (P = 4.4 × 10–6 and P = 8.6 × 10–8, respectively). Ileal TMPRSS2 expression was associated with increasing age (P = .04). Furthermore, we found increased TMPRSS2 expression in intestinal (ileal and colonic) biopsies of male patients (P = .02).

Second, we studied the effects of immunomodulating drugs on ACE2 and TMPRSS2 expression in intestinal biopsies. We found increased TMPRSS2 expression in ileal biopsies of patients using tumor necrosis factor alpha antagonists (P = 8.8 × 10–6), independent of intestinal inflammation, age, sex, diagnosis, and BMI. Because aminosalicylates were only used in the context of ulcerative colitis, we assessed the influence of aminosalicylates in colonic tissue only, and observed an increased TMPRSS2 expression (P = .02). The use of thiopurines or steroids was not associated with differential expression of ACE2 or TMPRSS2. Furthermore, host genetic variation was not associated with differential expression of ACE2 or TMPRSS2.

Third, we interrogated the cell-type–specific contexts of intestinal ACE2 and TMPRSS2 expression. Using bulk RNA sequencing data, we quantified cell-type proportions and observed that an enrichment of epithelial cells is associated with increased expression of ACE2 and TMPRSS2 in both ileum and colon (all P < .004). Using single-cell RNA sequencing data from colonic biopsies, we observed that TMPRSS2 is mainly expressed by absorptive enterocytes, and that intestinal inflammation is associated with increased TMPRSS2 expression within absorptive enterocytes (P = 3.2 × 10–22). ACE2 was also primarily expressed by absorptive enterocytes, but expression was not affected by intestinal inflammation.

In conclusion, ACE2 and TMPRSS2 are key proteins for cellular entry of SARS-CoV-2, and are highly expressed in the intestinal mucosa. Next to intestinal inflammation, age, sex, and anatomic location, the use of tumor necrosis factor alpha antagonists and aminosalicylates influence intestinal expression of TMPRSS2 as well. We demonstrate that intestinal inflammation is associated with an increased expression of TMPRSS2 in absorptive enterocytes, suggesting that the increased expression is not merely an effect of change in cellular composition during inflammation. Altered intestinal expression could render patients with IBD particularly susceptible to COVID-19 and absorptive enterocytes could provide targets for interventional studies. Indeed, clinical studies are needed to monitor the impact of COVID-19 on patients with IBD.