Identification of Transcription Factors Regulating SARS-CoV-2 Entry Genes in the Intestine.

Gastrointestinal symptoms of coronavirus disease 2019 (COVID-19), including diarrhea, nausea, and vomiting, are more common than previously thought. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) likely causes these symptoms by infecting the epithelial cells lining the gastrointestinal tract, with angiotensin I converting enzyme 2 (ACE2) functioning as the viral receptor and transmembrane serine protease 2 (TMPRSS2) functioning in viral spike protein priming. Du et al recently reported that ACE2 not only is expressed in lung alveolar type II (AT2) cells but also highly expressed in absorptive enterocytes. However, the regulatory mechanisms and transcription factors driving expression of Ace2 and Tmprss2 remains unclear. Using epigenomic approaches and mouse genetic models, we identify 4 key transcriptional regulators (caudal-type homeobox 2 (CDX2), hepatocyte nuclear factor 4 (HNF4), Smad family member 4 (SMAD4), or GATA binding proteins) that bind to the loci of these genes, alter chromatin looping, shape epigenetic modifications, and, ultimately, show a dramatic impact on Ace2 and Tmprss2 gene expression upon transcription factor knockout.

We began by investigating the expression of COVID-19–related host genes throughout the body. Chromatin accessibility and RNA transcript levels show a tissue-specific expression pattern for Ace2 and Tmprss2, with greatest expression observed in intestine, kidney, and lung tissues (Figure 1A and B and Supplementary Figure 1). Ace2 and Tmprss2 are expressed more robustly in isolated intestinal epithelium compared with the remaining subepithelium (Figure 1C). Temporally, transcript levels of Ace2 and Tmprss2 increase during embryonic development (Figure 1D and F). In the adult tissue, Ace2 transcripts are increased in the villus compared with the crypt (Figure 1E and F). We next focused on the epithelial cell populations expressing the SARS-CoV-2–related genes. single cell RNA sequencing analysis defined cell populations within crypt epithelium that express markers of stem, progenitor, and differentiating epithelial cells (Supplementary Figure 2A). Cells expressing canonic goblet, Paneth, tuft, enteroendocrine, or enterocyte lineage markers were each identified, as expected (Supplementary Figure 2B–F, respectively). Other transcripts are expressed more broadly throughout the epithelium (Supplementary Figure 2G). We found Ace2 expression to be enriched in cells co-expressing mature enterocyte markers, whereas Tmprss2 was expressed more broadly throughout the epithelium (Supplementary Figure 2H).

Figure 1

(A) DNase sequencing (DNase-seq) (GSE51336) enrichment at gene loci of Ace2 and Tmprss2 among different tissues. (B) RNA sequencing (RNA-seq) (GSE36025) tracks of Ace2 and Tmprss2 in different tissues. (C) Microarray data of mesenchymal and epithelial cells of the developing intestine (GSE6383). Data are collected from microarray probes showing representative expression fold change. (D) RNA sequencing of intestinal epithelium (GSE115541) shows that Ace2 and Tmprss2 increasingly are expressed across developmental time, from embryonic day (E)12.5 to adult. (E) In the adult tissue, Ace2 is expressed more robustly in the villus vs the crypt cells (GSE53545, GSE70766, and GSE102171). Data are presented as means ± SEM. ∗∗∗P < .001. (F) integrative genomics viewer tracks of RNA sequencing. ADVil, adult villi; FPKM, fragments per kilobase of transcript per million mapped reads.

We next examined how intestinal transcription factor regulatory networks impact the expression of genes important for SARS-CoV-2 infection. CDX2 is required for specification of the intestine during embryonic development, and in adult life is required for intestinal maturation and proper enterocyte function. HNF4 factors are required for maturation of the embryonic intestine, and work in conjunction with SMAD4 to promote expression of adult enterocyte genes. GATA family transcription factors are important for intestinal regionalization. Chromatin immunoprecipitation sequencing of these transcription factors shows multiple binding regions at the loci of genes involved in CoV-2 infection (Figure 2A). Notably, the binding regions occur at locations of accessible chromatin (indicated by assay for transposase-accessible chromatin (ATAC)-seq in pink in Figure 2A), consistent with transcriptional enhancer functions. Active chromatin modifications were observed at these loci, and, importantly, chromatin modifications at the Ace2 locus showed dependence on knockout of CDX or HNF factors in the epithelium using the Villin-Cre driver (indicated by asterisks in Figure 2B). Aside from Ace2, we also observed increased levels of these active chromatin modifications at the Tmprss2 locus upon knockout of HNF4 or CDX transcription factors (Figure 2B). Altered chromatin structure at the Ace2 and Tmprss2 loci upon CDX and HNF4 loss would predict a corresponding change in transcript levels in these knockout models. Indeed, Ace2 was down-regulated significantly upon loss of CDX2, HNF4 factors, SMAD4, or GATA factors (Figure 2C–F). Although Ace2 levels were diminished in these knockout models, Tmprss2 showed an increase in transcript levels (Figure 2C–F). Finally, consistent with their dynamic expression changes in response to HNF4 knockout, we also observed dynamic chromatin looping events at these loci, as measured by Hi-C chromatin immunoprecipitation (HiChIP). Although the Ace2 locus showed fewer chromatin looping events upon HNF4 knockout, the Tmprss2 locus showed increased contacts between presumed enhancers and the Tmprss2 transcriptional start site (Figure 2G).

Figure 2

HNF4, CDX2, SMAD4, and GATA transcription factors are key regulators of (A) Chromatin accessibility (ATAC-seq) and transcription factor binding (chromatin immunoprecipitation sequencing [ChIP-seq]) at COVID-19–related gene loci. (B) Histone post-translational modifications associated with enhancers and promoters are reduced upon knockout of either CDX or HNF4 at the Ace2 locus (indicated by asterisks), and, conversely, are unchanged or increase at the Tmprss2 locus. (C–F) Transcriptome analysis to measure corresponding gene expression changes of COVID-19–related host genes upon knockout of these transcription factors (∗P < .05, ∗∗P < .01, and ∗∗∗P < .001). (G) H3K4me3-HiChIP assays to measure 3-dimensional enhancer–promoter looping show multiple contacts between COVID-related gene promoters and nearby regulatory elements. Loops are visualized by Sushi, and shown with q ≤ 0.0001 and counts ≥ 8 (2 combined biological replicates per condition). ATAC-seq, assay for transposase-accessible chromatin with high throughput sequencing; FPKM, fragments per kilobase of transcript per million mapped reads; H3K4me3-HiChIP, Hi-C chromatin immunoprecipitation of H3K4me3; MNase, micrococcal nuclease; WT, wild-type.

This study provides an overview of how the genes known to facilitate SARS-CoV-2 infection are spatially and temporally expressed and transcriptionally regulated. Transcription factors function in complex and collaborative networks to promote proper cell function. It is interesting that 4 key intestinal transcription factors, CDX2, HNF4 factors, SMAD4, or GATA factors, all activate Ace2 in the intestine. In addition, these transcription factors also might work as suppressors for Tmprss2, because Tmprss2 is increased upon loss of these key intestinal transcription factors. These findings could help understand variable susceptibility to COVID-19 within the population owing to variable utilization of these transcription factors or their binding sites, which might cause variable expression of Ace2 or Tmprss2. These regulatory mechanisms ultimately could lead to potential avenues for altering host gene expression to reduce the susceptibility or severity of SARS-CoV-2 infection.