Cutting Edge: Hypoxia-Induced Ubc9 Promoter Hypermethylation Regulates IL-17 Expression in Ulcerative Colitis Free

Inflammatory bowel diseases are the chronic relapsing inflammatory disease of the gastrointestinal tract that includes ulcerative colitis (UC) and Crohn disease. Emerging evidence suggests a crucial role of IL-17 in gastrointestinal inflammation that is produced by Th17 cells, γδT cells, and innate lymphoid cells (ILCs) (1). The orphan nuclear receptor ROR-γt is the major transcription factor that binds to ROR response elements within the IL-17a promoter to induce IL-17a transcription (2). However, the molecular mechanisms by which the expression of ROR-γt is regulated to prevent unrestricted inflammation remain unclear.

Posttranslational modification by small ubiquitin-like modifier (SUMO) proteins catalyzes a cascade of biochemical reactions that mediates E1, E2, and E3 enzymes. Ubc9 is the E2 enzyme that is used by the SUMO pathway as a conjugation enzyme to transfer SUMO to the substrate proteins (3–5). We have shown earlier that Ubc9 inhibits IL-17 expression in the colonic mucosa via the SUMOylation of ROR-γt (6).

Hypoxia is a key feature of inflammatory conditions of the intestine via the transcription factor hypoxia-inducible factor (HIF), which regulates key target genes involved in inflammation (7). In this study, we report that HIF-1α binds to CpG island within the Ubc9 promoter is hypermethylated in the colonic mucosa of UC patients, resulting in reduced Ubc9 expression, causing elevated IL-17a expression.

C57BL/6 mice and Rag1^−/− mice were purchased from The Jackson Laboratory. All experiments were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of Baylor Research Institute and University of Texas Southwestern Medical Center.

Abs used in these studies were anti-SUMO1 (1:500, no. FL-101; Santa Cruz Biotechnology), anti–ROR-γt (1:800; BD Biosciences), anti-Ubc9 (1:1000; Cell Signaling Technology), anti–HIF-1α (1:200; BD Biosciences), mouse IgG (Santa Cruz Biotechnology), PE-CD45Rb (BioLegend), APC-CD4 (BioLegend), FITC-CD25 (Tonbo Biosciences), Clean-Blot IP Detection Reagent (HRP) (1:50; Pierce Biotechnology), DMEM (Life Technologies), FBS (Life Technologies), QIAamp DNA Mini Kit (QIAGEN), EpiTect Bisulfite Kit (QIAGEN), EpiTaq HS polymerase (TaKaRa), and Chromatin Immunoprecipitation (ChIP) Assay Kit (MilliporeSigma).

Colon tissue samples were collected from UC patients who underwent a colectomy at the Baylor University Medical Center. The study was approved by the Institutional Review Board of the Baylor Scott & White Research Institute.

The CpG island of Ubc9 gene was predicted by University of California, Santa Cruz, genome browser and by the MethPrimer software using an island size of 150 nt, at least 50% guanine-cytosine percentage, and observed/expected CpG ratio of 0.6. MethPrimer is used to design the primers for methylation-specific PCR and unmethylation-specific PCR amplification of the target gene.

Genomic DNA was isolated from the samples using the QIAamp DNA Mini Kit (QIAGEN) according to manufacturer’s protocol. Two micrograms of genomic DNA was modified with sodium bisulfite using EpiTect Bisulfite Kit (QIAGEN). Fragments were amplified using EpiTaq HS (TaKaRa). The primer sequences for the genes shown in Supplemental Table 1.

Jurkat T cells were transfected with HIF-1α and Ubc9 promoter plasmid using Lonza Cell Line Nucleofector Kit. Cells were kept in hypoxia incubator chamber (2% O₂). After 24 h of transfection, cells were stimulated with PMA (50 ng/ml) and ionomycin (1 μg/ml). Lysates were prepared, and luminescence was measured.

CD4⁺ T cells were isolated from the spleen of wild-type mice using CD4⁺ T cell isolation kit (Miltenyi Biotec). The CD4⁺CD25⁻CD45RB^hi cells were FACS sorted and transduced with LentiCRISPR–HIF-1α viral particles. Transduced cells were cultured in Th17 conditions (8) and then injected i.p. (3 × 10⁵ cells per mouse) into Rag1^−/− mice and monitored for disease severity up to 8 wk.

The data were analyzed with GraphPad Prism 8 software to determine statistical significance using the paired Student t test. The data are expressed as mean ± SD. A p value <0.05 was considered significant: *p < 0.05, **p < 0.01, and ***p < 0.001.

To investigate the potential dysregulation of Ubc9–ROR-γt pathway (6) in UC patients, we tested Ubc9 expression by immunoblotting the surgically resected colon tissue samples from UC patients. As shown in Fig. 1A, we observed reduced Ubc9 level in UC patients compared with control samples. Consistent with a previously published report (9), we also observed reduced level of Ubc9 transcripts in UC patient’s samples (Supplemental Fig. 1). Next, we analyzed the level of IL-17a expression in these samples, and as shown in Fig. 1B, we found a strong inverse correlation (Pearson coefficient, p = −0.76) between Ubc9 mRNA expression and IL-17a level.

Next, we investigated if reduced Ubc9 expression resulted in reduced SUMOylation of ROR-γt in UC patients. We immunoprecipitated ROR-γt and analyzed SUMOylation by immunoblotting using anti-SUMO1 Ab. We observed a significant decrease in the SUMOylation of ROR-γt in UC samples (Fig. 1C, 1D). These data suggest that reduced Ubc9 expression leads to decreased SUMOylation of ROR-γt in the inflamed colonic mucosa of UC patients.

Epigenetic modifications regulate gene expression without affecting genomic sequences. DNA methylation is the most well-established form of epigenetic regulation mediated by DNA methyltransferases (10). To investigate the potential mechanism for reduced Ubc9 expression in UC patients, we analyzed Ubc9 promoter within a 2.4-kb region upstream from first start codon. We observed a CpG island within the exon 1 and promoter of Ubc9 gene (guanine-cytosine content = 63.3%; ratio of observed CpG versus expected CpG = 1.11 and CpG count = 167) (Fig. 2A). We hypothesized that hypermethylation of this CpG island could be the cause for reduced transcription of Ubc9 in UC patients. We performed methylation-specific PCR and unmethylation-specific PCR. Interestingly, we observed DNA methylation in most of the UC patients, whereas no DNA methylation was observed in most of the controls (Fig. 2B).

Hypoxic conditions promote DNA methylation in inflammatory bowel disease and other inflammatory diseases (11). Therefore, we hypothesized that hypoxia-induced transcription factors might promote DNA methylation of the CpG island. In support of this hypothesis, we found a consensus hypoxia-response element (HRE) in human Ubc9 and mouse Ubc9 promoter (Supplemental Fig. 2A, 2B). Because HIF-1α is the major factor that is induced during hypoxic conditions (7), we performed ChIP assays using colonic lamina propria lymphocytes (cLPLs) isolated from mice and kept under hypoxic conditions (2% O₂). As shown in Fig. 3A, anti–HIF-1α Ab, but not the control IgG, precipitated Ubc9 promoter DNA, suggesting that HIF-1α binds to HREs within the CpG island.

To investigate if HIF-1α binding to Ubc9 promoter modulates Ubc9 expression, next, we performed a Ubc9 promoter–driven luciferase assays. We transfected cells with HIF-1α expressing construct along with pGL4–human Ubc9 promoter construct. Cells were subjected to hypoxia (2% O₂) for 24 h and lysed, and luciferase activity was measured. As shown in Fig. 3B, luciferase activity of Ubc9 promoter was decreased when wild-type HIF-1α was expressed with pGL4–human Ubc9 promoter construct.

To confirm that binding of HIF-1α to Ubc9 promoter inhibits Ubc9 expression, we knocked down HIF-1α expression in mouse cLPLs by using LentiCRISPR viral particles specific to HIF-1α. The efficiency of knockdown was confirmed by Western blotting (Fig. 4A). We then stimulated these cLPLs with PMA and ionomycin and subjected them to hypoxia (2% O₂). As shown in Fig. 4B, 4C, knocking down HIF-1α substantially rescued the cells from reduced Ubc9 expression. Further, lentiviral mediated depletion of HIF-1α substantially reduced IL-17a and HIF-1α target genes such as PTGS1, PTGS2, and CXCR4 expression (Fig. 4D, 4E). Also, hypoxia induced DNA methylation in in vitro–generated Th17 cells, but not in the cells in which HIF-1α was knocked down (Fig. 4F). These data collectively suggest that HIF-1α negatively regulates Ubc9 expression and modulates IL-17a expression. Previous studies have shown that hypoxia and HIF-1α induce promoter DNA methylation and regulate gene expression (12–14). Several mechanisms have been proposed including inhibition of the activity of ten-eleven translocation (TET) enzymes (15) and by the induction of DNA methyltransferases (DNMTs) (16). It is likely that a similar mechanism is involved in HIF-1α–mediated regulation of Ubc9 promoter.

As has been shown earlier, the HIF-1α enhances the development of Th17 cells. Moreover, the murine T cells with HIF-1α deficiency are resistant to induction of Th17-dependent experimental autoimmune encephalitis (17). Therefore, to investigate the physiological impact of HIF-1α–mediated regulation of Ubc9 and IL-17–mediated inflammation, we used the Th17 cell adoptive transfer colitis model (18, 19). We sorted CD4⁺CD25⁻CD45RB^hi cells and knocked down HIF-1α using LentiCRISPR viral particles. The cells were differentiated under Th17-polarizing conditions for 6–7 d, and 3 × 10⁵ cells were adoptively transferred into Rag1^−/− mice (8). Rag1^−/− mice that received Th17 cells with reduced HIF-1α showed reduced loss of body weight, lower fecal occult blood and diarrhea scores, and lower weight-to-length ratio of the colon compared with mice that received control T cells (Fig. 5A–E). These mice exhibited lower IL-17a mRNA levels in the colonic mucosa, and colonic cultures showed lower IL-17a secretion than control T cells (Fig. 5F, 5G). These mice showed increased Ubc9 mRNA levels compared with mice receiving control T cells (Fig. 5H); correspondingly, histologic analysis showed lower diseases scores in the Rag1^−/− mice receiving Th17 cells with reduced HIF-1α (Fig. 5I, 5J). Further, the expression of IL-17–induced genes such as MMP1, MMP2, MMP3, MMP9, and CXCL1 was substantially reduced in Rag1^−/− mice that received Th17 cells with reduced HIF-1α (Fig. 5K–O). Together, these findings suggest that HIF-1α negatively regulates Ubc9 expression and modulates inflammation in the colonic mucosa. Earlier studies have shown that conditional deletion of HIF-1α in epithelial cells, dendritic cells, and regulatory T cells leads to increase in the intestinal permeability and more severe colitis (20–22). Knockdown of HIF-1α in myeloid cells ameliorates dextran sulfate sodium–induced colitis, which was associated with reduced monocyte infiltration and attenuated IL-17 levels (23). These studies along with our findings further highlight important cell-specific proinflammatory roles of HIF-1α that should be taken into account when designing strategies to treat gut inflammation by targeting HIF-1α.

The authors have no financial conflicts of interest.