C/EBP homologous protein-induced loss of intestinal epithelial stemness contributes to bile duct ligation-induced cholestatic liver injury in mice

Abstract

Impaired intestinal barrier function promotes the progression of various liver diseases, including cholestatic liver diseases. The close association of primary sclerosing cholangitis (PSC) with inflammatory bowel disease highlights the importance of the gut-liver axis. It has been reported that bile duct ligation (BDL)-induced liver fibrosis is significantly reduced in C/EBP homologous protein knockout (CHOP^-/- ) mice. However, the underlying mechanisms remain unclear. In the current study, we demonstrate that BDL induces striking and acute hepatic endoplasmic reticulum (ER) stress responses after 1 day, which return to normal after 3 days. No significant hepatocyte apoptosis is detected 7-14 days following BDL. However, the inflammatory response is significantly increased after 7 days, which is similar to what we found in human PSC liver samples. BDL-induced loss of stemness in intestinal stem cells (ISCs), disruption of intestinal barrier function, bacterial translocation, activation of hepatic inflammation, M2 macrophage polarization and liver fibrosis are significantly reduced in CHOP^-/- mice. In addition, intestinal organoids derived from CHOP^-/- mice contain more and longer crypt structures than those from wild-type (WT) mice, which is consistent with the upregulation of stem cell markers (leucine-rich repeat-containing G-protein-coupled receptor 5, olfactomedin 4, and SRY [sex determining region Y]-box 9) and in vivo findings that CHOP^-/- mice have longer villi and crypts as compared to WT mice. Similarly, mRNA levels of CD14, interleukin-1β, tumor necrosis factor-alpha, and monocyte chemotactic protein-1 are increased and stem cell proliferation is suppressed in the duodenum of patients with cirrhosis.

Conclusion: Activation of ER stress and subsequent loss of stemness of ISCs plays a critical role in BDL-induced systemic inflammation and cholestatic liver injury. Modulation of the ER stress response represents a potential therapeutic strategy for cholestatic liver diseases as well as other inflammatory diseases. (Hepatology 2018;67:1441-1457).

Fig. 1. CHOP deficiency protected BDL-induced liver injury

WT and CHOP^−/− mice were subjected to BDL or sham operation for 14 days. (A) Total bile acid (TBA) levels, ALP, AST and ALT activity levels in serum were determined. (B) Representative images of H&E, Masson’s Trichrome staining of liver paraffin sections. Scale bar 100μm. (C) Metavir fibrosis score based on H&E and Masson’s Trichrome images. (D) The relative mRNA levels of α-SMA, Collagen-1, LOXL2 and Fibronectin in the liver were measured by qRT-PCR and normalized using HPRT1 as an internal control. (E) Hepatic TBA levels. (F) Hepatic hydroxyproline levels. Results are represented as Mean ± S.E. (n=6–9). Statistical significance relative to the WT sham group, *p<0.05, **p<0.01, ***p<0.001; relative to the WT BDL group, #p<0.05, ##p<0.01, ###p<0.001.

Fig. 2. CHOP deficiency attenuated BDL-induced hepatic inflammatory responses

(A) Representative images of IHC against CD11b, F4/80, merged images with counterstaining of DAPI, and magnified views of portal area from liver sections of 14-day BDL or sham WT and CHOP^−/− mice. Scale bar 50 μm. (B–F) The relative mRNA levels of target genes in the liver of 14-day BDL or sham WT and CHOP^−/− mice were measured by qRT-PCR and normalized using HPRT1. Results are represented as Mean ± S.E. (n=6–9). Statistical significance relative to the WT sham group, *p<0.05, **p<0.01, ***p<0.001; relative to the WT BDL group, #p<0.05, ##p<0.01.

Fig. 3. The effect of CHOP deficiency on BDL-induced hepatocyte apoptosis and innate immunity

(A) The relative mRNA levels of CHOP, ATF4 and GRP78 in the livers of WT mice 1, 3, 5, 7 and 14 days after BDL. (B) Representative immunoblot images of cleaved-caspase-3 in liver samples of 14-day BDL or sham WT and CHOP^−/− mice. The relative protein levels of cleaved-caspase-3 were analyzed using β–Actin as a loading control. (C) WT and CHOP^−/− mice were subjected to BDL or sham operation for 14 days. Representative images of immunofluorescent staining against cleaved-caspase-3 and merged images with DAPI counterstaining are shown. The number of cleaved-caspase-3 positive cells was counted. Scale bar 50 μm. (D) The relative mRNA levels of TLR2, TLR4, CD14, IL-18 and IL-1β, in GRP78 in the liver of WT mice 1, 3, 5, 7 and 14 days after BDL. (E) The relative mRNA levels of TLR2, TLR4, CD14, IL-18 and IL-1β in the livers of normal human and PSC patient. (F) The relative mRNA levels of TLR2, TLR4, CD14, IL-18 and IL-1β in the liver of WT and CHOP^−/− mice with 14-day BDL. Results are represented as mean ± S.E. from each group (n=5–9 for mice samples, n=10 for human samples). Statistical significance relative to the WT sham group or BDL 0 day group, *p<0.05, **p<0.01, ***p<0.001; relative to the WT BDL group, #p<0.05, ##p<0.01, ###p<0.001.

Fig. 4. CHOP^−/− mice were protected from BDL-induced disruption of intestinal barrier function and bacterial translocation

WT and CHOP^−/− mice were subjected to BDL or sham operation for 14 days. (A) FITC-Dextran levels in the serum. Colony-forming units (CFUs) of bacteria from (B) MLNs and (C) blood. (D) Representative images of H&E staining of small intestine sections and magnified views. Yellow arrows indicate infiltration of inflammatory cells into intestinal epithelium. Red arrows indicate dilated lymph vessel. Scale bar 200 μm. (E) The relative mRNA levels of IL-1β, TNF-α and MCP-1 in the small intestine were measured by qRT-PCR and normalized using HPRT1. Results are represented as Mean ± S.E. from each group (n=5–8). Statistical significance relative to the WT sham group, *p<0.05, **p<0.01, ***p<0.001; relative to the WT BDL group, #p<0.05, ##p<0.01.

Fig. 5. CHOP deficiency attenuated BDL-induced loss of tight junctions and apoptosis in small intestine

WT and CHOP^−/− mice were subjected to BDL or sham operation for 14 days. (A) The relative mRNA levels of OCLN, ZO-1 and E-Cad in the intestine, normalized using HPRT1. (B) The protein levels of OCLN, E-Cad and β-Actin in the small intestine villi were determined by Western blot analysis. Representative images are shown. (C) Representative images of immunofluorescent staining of small intestine sections against E-Cadherin, OCLN, and merged images with counterstaining of DAPI are shown. Scale bar 50 μm. (D) TUNEL assay of small intestine sections. The number of TUNEL positive cells in intestinal epithelium is shown. Scale bar 50 μm. (E-F) The relative mRNA levels of CHOP, XBP-1s, ATF4 and GRP78 in intestine were measured by qRT-PCR and normalized using HPRT1. Results are represented as Mean ± S.E. from each group (n=5–8). Statistical significance relative to the WT sham group, *p<0.05, ***p<0.001; relative to the WT BDL group, #p<0.05.

Fig. 6. CHOP^−/− mice were protected from BDL-induced suppression of ISCs

(A) WT and CHOP^−/− mice were subjected to BDL or sham operation for 14 days. The length of intestinal villi and intestinal crypts was measured based on the H&E staining images. (B–C) WT and CHOP^−/− mice were subjected to BDL or sham operation for 3, 5, 7 or 14 days. The relative mRNA levels of Lgr5, Olfm4, c-Myc and Ki67 in isolated intestinal crypts were measured by qRT-PCR and normalized using HPRT1. (D) IHC staining against c-Myc, Ki67 in the small intestine, and magnified Ki67 staining in the intestinal crypt base from WT and CHOP^−/− mice with 14-day BDL. Scale bar 50 μm. (E) The representative images of BrdU staining in the small intestine from WT and CHOP^−/− mice subjected to BDL for 3, 7, and 14 days. Scale bar 200 μm. (F) The relative mRNA levels of CHOP, ATF4 and GRP78 in the small intestine from WT and CHOP^−/− mice 3, 7, and 14 days after BDL were measured by qRT-PCR and normalized using HPRT1. Results are represented as Mean ± S.E. from each group (n=6–10). Statistical significance relative to the WT sham group or BDL 0-day group, *p<0.05, **p<0.01; relative to the WT BDL group, #p<0.05, ##p<0.01.

Fig. 7. CHOP deficiency prevented LPS- and TG-induced suppression of intestinal organoid growth

(A) Representative images of WT and CHOP^−/− intestinal organoids cultured for 2, 4, 6, 8 and 10 days. (B) WT and CHOP^−/− intestinal organoids were cultured for 4 days, then treated with LPS (500 ng/ml) or TG (200 nM) for 6 days. Representative images at 2, 4, 6 days are shown. (C–E) The relative mRNA levels of Lgr5, Ki67 and CHOP in crypt structures of intestinal organoids after 6-day treatment were measured by qRT-PCR and normalized using HPRT1. Results are represented as Mean ± S.E. from each group (n=3). Statistical significance relative to the WT organoids vehicle control group, *p<0.05, **p<0.01; relative to the WT organoids counterparts, #p<0.05, ##p<0.01, ###p<0.001.

Fig. 8

Schematic diagram of the proposed mechanism underlying the protective effects of CHOP deficiency on BDL-induced intestine dysfunction and cholestatic liver injury.