Epithelial Regnase-1 inhibits colorectal tumor growth by regulating IL-17 signaling via degradation of NFKBIZ mRNA

Abstract

Regnase-1 is a ribonuclease that regulates inflammation in immune cells by degrading cytokine mRNA. Regnase-1 was identified as one of the frequently mutated genes in the inflamed colorectal epithelium of patients with ulcerative colitis; however, its significance in intestinal epithelial cells during the tumorigenic process remains unknown. Therefore, we developed an Apc^Min/+ mouse model lacking Regnase-1 in intestinal epithelia. Regnase-1 deletion significantly enhanced colon tumor growth accompanied by elevated levels of extracellular signal-regulated kinase (ERK) phosphorylation in tumor tissues. Transcriptome analysis of the tumor tissues revealed that Nfkbiz, a mediator of the interleukin (IL)-17 signaling pathway, was the primary degradative target of Regnase-1 in enterocytes and that Regnase-1 deficiency enhanced IL-17 signaling. The treatment with antibiotics or IL-17-neutralizing antibody canceled the proliferative effect of colon tumors due to Regnase-1 deletion, suggesting the protective role of Regnase-1 against colon tumor growth was dependent on IL-17 signaling triggered by gut microbes. Analysis of the Nfkbiz knockout mouse model demonstrated that the tumor-suppressive effect of Regnase-1 depended on Nfkbiz expression. Remarkably, oral treatment of dimethyl fumarate, a potential inhibitor of Regnase-1 protein inactivation, suppressed tumor growth, downregulated Nfkbiz, and suppressed ERK activation. Furthermore, TCGA data analysis revealed that low Regnase-1 expression in colorectal cancer tissue was related to poor prognosis. Therefore, Regnase-1 represses colon tumor growth by regulating IL-17 signaling via Nfkbiz mRNA degradation. Regnase-1 could be a potential therapeutic target in colon tumors.

Keywords: IL-17; Regnase-1; colorectal cancer.

Fig. 1.

Phenotypes of Apc^Min/+ mice with intestine-specific Regnase-1 deletion (Reg1^KO-Min) compared to control Apc^Min/+ mice (Reg1^WT-Min). (A) Regnase-1 mRNA levels in intestinal or splenic tissues from 120-d-old Reg1^WT-Min and Reg1^KO-Min mice (N = 4). (B) Body weight of 120-d-old Reg1^WT-Min and Reg1^KO-Min mice (N = 14). (C) Representative intestines at 120 d old. (D) Tumor number and area in the proximal half, distal half of small intestines, and colons of 120-d-old Reg1^WT-Min and Reg1^KO-Min mice (N = 14). (E) Representative hematoxylin and eosin (HE) staining of colon tumors at 120 d old. (F) Representative IHC staining of pERK in colon tumor tissues. Quantification of the pERK-positive area (nine tumors from four Reg1^WT-Min mice and 20 tumors from four Reg1^KO-Min mice). Data are presented as mean ± SE.

Fig. 2.

Transcriptome analysis of colon tumor tissues from Reg1^WT-Min and Reg1^KO-Min mice. (A) Volcano plot presenting the result of RNA-seq performed in colon tumor tissues of Reg1^WT-Min and Reg1^KO-Min mice (N = 5). Differentially expressed genes are shown in red. Genes previously reported as a degradative target of Regnase-1, except Nfkbiz, are shown in blue. (B) Enriched gene sets detected by GSEA of RNA-seq results using Mouse MSigDB (M2: CP, Reactome). (C) Estimated level of intratumoral immune cell infiltration calculated using mMCP counter. (D) Quantification of immune cells per sight field in colon tumor tissues by IHC of Ly6G (13 tumors from five Reg1^WT-Min mice and 25 tumors from six Reg1^KO-Min mice) and CD4 (18 tumors from nine Reg1^WT-Min mice and 42 tumors from ten Reg1^KO-Min mice). (E) Representative IHC of Ly6G and CD4 in colon tumor tissues. Arrowheads indicate positively stained cells. (F) Enrichment plot for the gene set REACTOME_CHEMOKINE_RECEPTORS_BIND_CHEMOKINES. Black arrows indicate Cxcl1, Cxcl2, and Cxcl10. Data are presented as mean ± SE.

Fig. 3.

Upregulation of Nfkbiz in Reg1^KO-Min intestinal tissues. (A) IL-17 signaling pathway schema. (B) Representative IHC of IκBζ in colon tumor tissues. Quantification of the IκBζ-positive area (nine tumors from four Reg1^WT-Min mice and 17 tumors from four Reg1^KO-Min mice). (C) Enrichment plot for the gene set of Nfkbiz-downstream genes obtained from RNA-seq performed in Fig. 2. (D) mRNA expression levels of Il17a, Il17f, Il17ra, Il17rc, Il23a, and Rorc in colon tumors from 120-d-old Reg1^WT-Min and Reg1^KO-Min mice (N = 7). (E) Representative IHC of RORγ in colon tumor tissues. Arrowheads indicate positively stained cells. Quantification of Th17 cells per sight field in colon tumor tissues (seven tumors from four Reg1^WT-Min mice and 20 tumors from four Reg1^KO-Min mice) by IHC of RORγ. (F) mRNA levels of Nfkbiz and Il6 in nontumor intestinal tissues from 120-d-old Reg1^WT-Min and Reg1^KO-Min mice (N = 8). Data are presented as mean ± SE.

Fig. 4.

Phenotypes of Reg1^WT-Min mice and Reg1^KO-Min mice under antibiotics treatment or intestinal-specific Nfkbiz deletion. (A) Quantification by qPCR of bacteria in feces from Apc^Min/+ mice with or without oral antibiotics (Abx) treatment for 9 wk (N = 20 for control and N = 8 for Abx-treated). (B) Representative colon from mice with or without oral Abx treatment since weaning until 120 d old. (C) Tumor area in colons of mice without Abx treatment (N = 14 for Reg1^WT-Min and Reg1^KO-Min each) and mice with oral Abx treatment (N = 3 for Reg1^WT-Min and Reg1^KO-Min each) since weaning until 120 d old. (D) Representative colon from mice intraperitoneally injected weekly with isotype control (IgG) or anti-IL17A antibody from 30 d old until 120 d old. (E) Tumor area in colons of mice treated with IgG (N = 4 for both Reg1^WT-Min and Reg1^KO-Min) and those treated with anti-IL17A antibody (N = 4 for both Reg1^WT-Min and Reg1^KO-Min) from 30 d old until 120 d old. (F) Representative colon from mice with or without Nfkbiz^KO at 120 d old. (G) Tumor area in colons of mice without Nfkbiz^KO (N = 14 for Reg1^WT-Min and Reg1^KO-Min each) and mice with Nfkbiz^KO (N = 6 for Reg1^WT-Min and N = 7 for Reg1^KO-Min) at 120 d old. Data are presented as mean ± SE.

Fig. 5.

Effect of DMF on human colon cancer cell line HT29. (A) Immunoblotting using lysates of HT29 cells after IL-17A (50 ng/mL) treatment with or without DMF pretreatment. (B) Immunoblotting using lysates of HT29 cells after 3 h of IL-17A treatment with or without DMF pretreatment. (C) mRNA levels of Nfkbiz, Regnase-1, and Lcn2 in HT29 cells 3 h after IL-17A (50 ng/mL) treatment with DMF pretreatment. Data are presented as mean ± SE.

Fig. 6.

Effect of oral DMF treatment on mice. (A) Immunoblotting using lysates of nontumor cecum tissues from Reg1^WT mice treated with vehicle or DMF for 5 wk. (B) mRNA levels of Nfkbiz and Regnase-1 in nontumor cecum tissues from Reg1^WT and Reg1^KO mice (N = 4) treated with vehicle or DMF for 5 wk. (C) Representative colon from mice treated with vehicle or DMF since weaning age until 120 d old. (D) Tumor area in colons of 120-d-old mice treated with vehicle of DMF (N = 4 for Reg1^WT-Min and Reg1^KO-Min each). (E) Quantification of neutrophils and CD8⁺ T cells per sight field in colon tumor tissues of mice treated with or without DMF by IHC of Ly6G (20, 14, 42, 15 tumors from each group, respectively) and CD8 (6, 12, 25, 15 tumors from each group, respectively), respectively. Data are presented as mean ± SE.

Fig. 7.

Expression of Regnase-1 in human colorectal cancer cases. (A) Representative IHC of Regnase-1 and IκBζ (NFKBIZ) proteins in colorectal tumor tissues. T and NT represent tumor and nontumor tissues, respectively. (B) Quantification of the Regnase-1-positive area in tumor and nontumor tissues. (C) Correlation between Regnase-1-positive area and IκBζ-positive area rate in tumor tissues. (D) Representative image showing spots with depleted Regnase-1 expression and elevated NFKBIZ expression. (E) Overall survival and (F) disease-free survival of colorectal cancer cases with IL17A/IL17F expression higher than the average of each from the TCGA Firehose Legacy dataset.