Inhibiting sorting nexin 10 promotes mucosal healing through SREBP2-mediated stemness restoration of intestinal stem cells

Abstract

Intestinal stem cell (ISC) is a promising therapeutic target for inflammatory bowel disease. Cholesterol availability is critical for ISC stemness. Low plasma cholesterol is a typical feature of Crohn's disease (CD); however, its impact on mucosal healing remains unclear. Here, we identified an essential role of sorting nexin 10 (SNX10) in maintaining the stemness of ISCs. SNX10 expression in intestinal tissues positively correlates with the severity of human CD and mouse colitis. Conditional SNX10 knockout in intestinal epithelial cells or ISCs promotes intestinal mucosal repair by maintaining the ISC population associated with increased intracellular cholesterol synthesis. Disassociation of ERLIN2 with SCAP by SNX10 deletion enhances the activation of SREBP2, resulting in increased cholesterol biosynthesis. DC-SX029, a small-molecule inhibitor of SNX10, was used to verify the druggable potential of SNX10 for the treatment of patients with CD. Our study provides a strategy for mucosal healing through SREBP2-mediated stemness restoration of ISCs.

Fig. 1.. SNX10 expression in intestinal tissues positively correlates with the severity of IBD.

(A) SNX10 mRNA expression in health controls (n = 26) and active patients with CD (n = 66) of the GSE112366 dataset. (B) The correlation between SNX10 mRNA expression in biopsy samples and ileum simple endoscopic score for CD (ses-cd) from GSE112366 was analyzed by Spearman (n = 253). (C) SNX10 mRNA expression in inflamed and noninvolved areas of different intestinal segments from the GSE100833 dataset. (D) Patients with CD were divided into the high SNX10 expression group and the low SNX10 expression group according to the median of SNX10 mRNA level in the biopsy samples (GSE100833, n = 1717). Gene set enrichment analysis (GSEA) of indicated gene sets in CD patients with SNX10 low expression versus SNX10 high expression was conducted. In one analysis, a nominal P value was shown. (E) Immunofluorescent staining of SNX10 in normal or typical lesion areas of the colon from patients with CD. All scale bar, 50 μm. (F) Immunofluorescent staining of SNX10 in the areas of varying degrees of inflammation of colon tissues from patients with CD. White arrows indicate SNX10-positive epithelial cells. Scale bar, 100 μm. Data are represented as means ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001. DAPI, 4′,6-diamidino-2-phenylindole.

Fig. 2.. Intestinal epithelium SNX10 deletion alleviated DSS-induced colitis and mucosal injury.

(A) Immunofluorescent staining of SNX10 in colon sections of DSS-treated mice. Scale bar, 50 μm. (B) Relative mRNA levels of Snx10 in colonic epithelial cells (EPCAM⁺CD45⁻) isolated from DSS-treated mice (n = 6). (C) Survival curve of Snx10^fl/fl (WT) and Snx10^fl/fl;Vil-Cre⁺ (Snx10^ΔIEC) mice (n = 16) subjected to colitis induction with 3% DSS for 6 days, followed by regular drinking water. (D) Body weight change of mice treated with 3% DSS for 6 days, followed by regular drinking water for 3 days (n = 10). (E) The colon length of WT and Snx10^ΔIEC mice treated with DSS on day 9. (F) Relative mRNA levels of proinflammatory cytokines in the colon tissue measured by reverse transcription quantitative polymerase chain reaction (RT-qPCR) (n = 7 to 10). (G) Representative hematoxylin and eosin (H&E) staining images of WT and Snx10^ΔIEC mice on day 9. Scale bar, 100 μm. (H) Body weight change of mice (n = 6) treated with 1% (w/v) TNBS presensitization solution for 6 days, followed by 2.5% (w/v) TNBS intrarectal (ir) delivery (at day 0). (I) The colon length of WT and Snx10^ΔIEC mice induced with TNBS on day 4. (J) Representative H&E staining images of WT and Snx10^ΔIEC mice treated with TNBS. Scale bars, 100 μm. (K) Relative mRNA levels of proinflammatory cytokines in the colon tissue measured by RT-qPCR (n = 6). Data are represented as means ± SD. ns, not significant *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 3.. SNX10 deficiency enhanced the stemness of CBC stem cells and promoted the differentiation of secretory-lineage cells.

(A) Representative immunohistochemistry (IHC) images of Ki67 in colon sections and positive cell count in each crypt. (B) Immunofluorescent staining of AXIN2 in colon sections. (C) Immunofluorescent staining of Ki67 and SNX10 in colon tissues from patients with CD. White arrows: SNX10-positive epithelial cells. (D) ISC signature gene expression in colonic epithelial cells isolated from colitis mice. (E) Small intestinal (SI) organoids cultured from WT or Snx10^ΔIEC mice. (F) Average surface area and (G) structure quantification (based on the number of buds) of SI organoids (n = 5, 30 organoids per mouse). (H) Immunofluorescent staining of Ki67 or AXIN2 in SI organoids on day 7. (I) Representative images of colon organoids. (J) Average surface area and (K) quantification of colon organoids (n = 6, 30 organoids per mouse). (L) Immunofluorescent staining of Ki67 or AXIN2 in colon organoids on day 7. (M and N) Immunofluorescent staining of differentiation markers (Alpi or CAII: absorptive cells, Lyz: Paneth cells, Muc2: goblet cells, ChgA: enteroendocrine cells, COX1: tuft cells) in organoids on day 5. (O) Representative IHC images of β-catenin in the colon sections. (P) Immunofluorescent staining of β-catenin and SNX10 in the areas with different degrees of inflammation of colon tissues from patients with CD. Scale bar, 100 μm. (Q) Immunoblots for active β-catenin in colonic epithelial cells isolated from colitis mice. Data are represented as means ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 4.. Lgr5-ISC−specific SNX10 deficiency retained the ISC pool for accelerating the epithelial repair.

(A) GSEA of indicated gene sets in CD patients with SNX10 low expression versus SNX10 high expression from GSE100833 (n = 1717). (B) Correlation analysis by Spearman between SNX10 and ISC marker gene expression in CD patients with moderate or severe endoscopic activity from GSE112366 (n = 56). (C) Immunofluorescent staining of SNX10 and LGR5-EGFP (enhanced green fluorescent protein) in colon sections from colitis mice. (D) Immunofluorescent staining of OLFM4 and SNX10 in inflamed colon tissues from patients with CD. (E) Body weight change of WT and Snx10^ΔISC mice (n = 6, tamoxifen-induced) subjected to colitis induction with 3% DSS for 6 days, followed by regular drinking water. (F) Colon length of WT and Snx10^ΔISC mice on day 9. (G) Histology of the recovery region in the colon from WT or Snx10^ΔISC mice on day 9. (H) Quantification of crypts with 0, 1, 2, or 3+ EGFP⁺ cells in recovery region (~50 crypts per mouse, six mice per group). (I) Histology of the inflamed region in the colon from WT or Snx10 ^ΔISC DSS-induced mice (day 6). Scale bar, 100 μm. (J) Images of SI organoids and (K) colon organoids on day 5 from WT or Snx10^ΔISC mice (left). The average surface area of 30 organoids per mouse (n = 5) was measured (right). Data are represented as means ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 5.. The increased cholesterol biosynthesis by SNX10 deletion contributed to intestinal epithelial repair.

(A) GSEA of gene sets enriched in colonic epithelial cells from Snx10^ΔIEC mice (versus WT mice) with DSS colitis (n = 3). (B) Correlation analysis (by Spearman) matrix between ISC signature genes and cholesterol biosynthesis genes in the inflamed area from patients with CD of GSE100833 (n = 369). (C) Immunofluorescent staining of 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR) and SNX10 in inflamed areas of colon tissues from patients with CD. Scale bar, 100 μm. (D) Free cholesterol content in colonic epithelial cells isolated from colitis mice (n = 6). (E) Immunofluorescent staining of HMGCR in colon sections. (F) SI and (G) colon organoids were treated with 500 μM methyl-β-cyclodextrin (MβCD) for 16 hours, and then replaced with normal medium for 16 hours. Free cholesterol content at different times was measured. (H) Representative images of organoids treated with or without lovastatin (5 μM) on day 6. (I) Free cholesterol content of organoids (n = 5). (J) Relative mRNA levels of cholesterol biosynthesis gene in organoids (n = 3). (K) Immunofluorescent staining of Dishevelled2 (DVL2) in colon sections from DSS-treated mice. Scale bar, 50 μm. (L) Immunofluorescent staining of DVL2 and LRP6 in HCT116 cells that were treated with serum-free medium containing 500 μM MβCD for 16 hours, and then replaced with serum-free medium for 8 hours, followed by additional Wnt3a (200 ng/ml) for another 2 hours. Data are represented as means ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001. NES, normalized enrichment score.

Fig. 6.. Disassociation of ERLIN2 with SCAP induced by SNX10 deletion enhanced SREBP2 activation contributing to the increased cholesterol biosynthesis.

(A) Representative immunoblots for precursors (P) and N-terminal cleavage fragments (N) of SREBP1 and SREBP2 in SI or colon organoids with or without lovastatin from WT or Snx10^ΔIEC mice. (B) Immunofluorescent staining of SREBP2 in WT or SNX10 KO HCT116 cells with or without cholesterol (10 μg/ml) treatment for 12 hours in low-glucose Dulbecco’s modified Eagle’s medium (DMEM). (C) WT or SNX10 KO HCT116 cells 36 hours after transfected with CellLight Golgi-GFP were cultured with low-glucose DMEM in the presence or absence of cholesterol (10 μg/ml) for 12 hours, followed by immunofluorescent staining of SREBP2. (D) Lysates of HCT116 cells transfected with empty vector (EV)-flag or SNX10-flag plasmids were prepared and subjected to immunoprecipitation with anti-FLAG M2 agarose beads, followed by a WB with the indicated antibodies. (E) Lysates of WT or KO HCT116 cells without or with cholesterol treatment for 12 hours were subjected to immunoprecipitation with anti-ERLIN2 antibody, followed by a WB with the indicated antibodies. (F and G) Immunofluorescent staining of SNX10 and SCAP or ERLIN2 in HCT116 cells without or with cholesterol treatment for 12 hours. IgG, immunoglobulin G.

Fig. 7.. SNX10 PPI DC-SX029 promoted epithelial repair by enhancing the SREBP2-mediated stemness of ISCs.

(A) Representative immunoblots for SREBP2 in SI or colon organoids with or without DC-SX029 (50 μM) treatment for 7 days. (B) Free cholesterol content in SI or colon organoids with or without DC-SX029 treatment for 7 days. (C and D) Relative mRNA levels of ISC signature genes and cholesterol biosynthesis genes in SI or colon organoids measured by RT-qPCR. (E) SI organoids were cultured in ENR medium without or with DC-SX029 (50 μM) for 7 days, followed by immunostaining with the antibodies against Ki67 and Axin2. (F) Colon organoids were cultured in WENR medium without or with DC-SX029 (50 μM) for 7 days, followed by immunostaining with the antibodies against Ki67 and Axin2. (G) Representative IHC images of Ki-67 and β-catenin in colon sections from each group on day 9. Scale bars, 50 μm. Mice (n = 10) were treated with 3% DSS for 6 days, followed by regular drinking water for 3 days. Mesalazine (50 mg/kg per day) or DC-SX029 (2 mg/kg per day) was given intragastrically during the whole process. (H) Free cholesterol content in colonic epithelial cells isolated from the indicated groups (n = 6). (I) Relative mRNA levels of ISC signature genes in colon epithelial cells isolated from the indicated groups measured by RT-qPCR (n = 10). (J) Immunofluorescent images of colon sections from Lgr5-EGFP mice in each group. Scale bar, 50 μm. Data are represented as means ± SD. Data of Fig. 6 (B to D) were compared by two-sided unpaired t test. Data of Fig. 6H were compared by one-way analysis of variance (ANOVA) followed by the Bonferroni post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001.