Hyperglycemia-triggered ATF6-CHOP pathway aggravates acute inflammatory liver injury by β-catenin signaling

Abstract

Although hyperglycemia has been documented as an unfavorable element that can further induce liver ischemia-reperfusion injury (IRI), the related molecular mechanisms remain to be clearly elaborated. This study investigated the effective manner of endoplasmic reticulum (ER) stress signaling in hyperglycemia-exacerbated liver IRI. Here we demonstrated that in the liver tissues and Kupffer cells (KCs) of DM patients and STZ-induced hyperglycemic mice, the ER stress-ATF6-CHOP signaling pathway is activated. TLR4-mediated pro-inflammatory activation was greatly attenuated by the addition of 4-phenylbutyrate (PBA), one common ER stress inhibitor. The liver IRI in hyperglycemic mice was also significantly reduced after PBA treatment. In addition, deficiency of CHOP (CHOP^-/-) obviously alleviates the hepatic IRI, and pro-inflammatory effects deteriorated by hyperglycemia. In hyperglycemic mice, β-catenin expression was suppressed while the ATF6-CHOP signal was activated. In the liver tissues of PBA-treated or CHOP^-/- hyperglycemic mice, the expression of β-catenin was restored. Furthermore, CHOP deficiency can induce protection against hyperglycemia-related liver IRI, which was disrupted by the knockdown of β-catenin will cause this protection to disappear. High glucose (HG) treatment stimulated ATF6-CHOP signaling, reduced cellular β-catenin accumulation, and promoted the TLR4-related inflammation of BMDMs. But the above effects were partially rescued in BMDMs with CHOP deficiency or by PBA treatment. In BMDMs cultured in HG conditions, the anti-inflammatory functions of CHOP^-/- were destroyed by the knockdown of β-catenin. Finally, chimeric mice carrying WT or CHOP^-/- BMDMs by bone marrow transplantation were adopted to verify the above conclusion. The current study suggested that hyperglycemia could trigger ER stress-ATF6-CHOP axis, inhibit β-catenin activation, accelerate inflammation, and deteriorate liver IRI, thus providing the treatment potential for management of sterile liver inflammation in DM patients.

Fig. 1. ATF6-CHOP signaling pathway was induced in liver tissues from DM patients and STZ mice.

A Quantification analysis of mRNA level of ER stress-related pathway molecules (IRE1, XBP1, PERK, ATF4, ATF6, and CHOP) in human livers (n = 15). *P < 0.05, **P < 0.01. B Protein level detection of cATF6 and CHOP in human liver. C CHOP in clinical samples was detected by IF staining. CHOP-positive cells from different groups were counted and presented in the right panel. **P < 0.01. D IRE1, XBP1, PERK, ATF4, ATF6, and CHOP gene mRNA levels in mice undergoing sham operation were quantified using qRT-PCR (ratios of target gene/HPRT). (n = 4~6). **P < 0.01. E The protein expression of cATF6 and CHOP in different groups of mice undergoing sham operation were examined by western blot. F Dual-immunofluorescence staining of CD68+ macrophages (red) and CHOP (green) in mouse liver sections. Histograms presented in the right panel were scored semi-quantitatively by averaging the number of positively stained cells per field (×200). Data are based on three independent experiments with similar conclusion. *P < 0.05, **P < 0.01.

Fig. 2. ATF6-CHOP pathway was required for hyperglycemia-exacerbated liver IRI.

The ER stress antagonist PBA was administered to DM mice before the start of liver ischemia. Untreated diabetic mice were used as controls. A cATF6 and CHOP protein expression in control mice, diabetic mice, and PBA-treated diabetic mice was measured by western blot analysis. Liver IRI was measured at 0, 6, and 24 h by. B The quantification results of sALT and sAST. C H&E staining of ischemic liver tissue (n = 4~6). Suzuki’s histological score was used to evaluate liver damage, and results were presented as a histogram in the right. D TUNEL staining of representative ischemic liver lobes (n = 4~6). TUNEL+ cells were measured by recording the positive cell numbers per area. E Western blot detection of Bcl-xL, Bcl-2, Cleaved caspase-3, Bax, and GAPDH. Data are representative of three experiments. *P < 0.05, **P < 0.01.

Fig. 3. ATF6-CHOP pathway was essential for the hyperglycemia-promoted pro-inflammatory response in liver IRI.

A TNF-α, IL-6, and IL-10 mRNA levels were examined by qRT-PCR and B the serum cytokine levels were measured by ELISA. C, D The infiltration of neutrophil and macrophage were analyzed by IHC (×200). CD68+ and Ly6G+ cells were quantitated by counting positive cells and the quantified results are shown as the right panels. E Protein expression detection of p-IKBα, p-NF-κB p65, p-IKKα/β, and GAPDH. Data are representative of three independent assays with consistent results. *P < 0.05, **P < 0.01.

Fig. 4. CHOP deficiency alleviated hyperglycemia-exacerbated liver injury and TLR4-mediated innate immune responses after IR.

A, B Liver function was evaluated by the value of sALT and sAST. C Histopathologic analysis of livers at 6 h after reperfusion, and the severity of liver IRI was judged according to Suzuki’s histological grading. D Protein expression of Bcl-2 and Cleaved caspase-3 in CHOP^−/− and WT diabetic mouse ischemic livers was detected using western blot. E Apoptosis was evaluated in CHOP^−/− and WT diabetic mouse ischemic livers by immunohistological staining with anti-cleaved caspase-3 (×200), and representative TUNEL-stained ischemic liver lobes (magnification ×400). TUNEL+ cells were recorded based on positive cell numbers/area. Cytokines level of liver and serum were examined through qRT-PCR (F) and ELISA (G), respectively. H CD68 and Ly6G staining of ischemic liver lobes (×200). Data are representative of three independent assays with similar results. *P < 0.05, **P < 0.01.

Fig. 5. Hyperglycemia-triggered ATF6-CHOP pathway depressed β-catenin signaling in liver tissues, and β-catenin was required for the protective effect of CHOP deficiency in hyperglycemia-aggravated liver IRI.

A Western blots of β-catenin and GAPDH in sham and ischemic livers, at 6 h post-reperfusion after 90 min ischemia. B CHOP and β-catenin protein expression in CHOP^−/− and WT diabetic mouse ischemic livers were detected using western blot. CHOP^−/− diabetic mice were delivered with Alexa Fluor 488-labeled siRNA through tail vein injection at 4 h before ischemia. C IF staining was conducted to exhibit CD68-positive macrophages (red) and control siRNA (green) labeled by Alexa Fluor 488 and in ischemic liver lobes. Nuclei were blue after staining with DAPI. D Western blot analysis of β-catenin in KCs isolated from CHOP^−/− diabetic mouse administered with NS siRNA or β-catenin siRNA. E Histological analysis of ischemic liver tissue. Suzuki’s histological grading was the common standard for judging liver IRI severity. F The concentration of sALT and sAST in CHOP^−/− DM mice treated with β-catenin siRNA or NS siRNA after 90 min of liver ischemia and 6 h of reperfusion. G, H The mRNA abundance and protein levels of TNF-α, IL-10, and IL-6 in mouse were determined using qRT-PCR and ELISA, respectively. I IHC analysis of CD68+ macrophages and Ly6G+ neutrophils (magnification ×200) and TUNEL staining (magnification ×400) in ischemic livers. *P < 0.05, **P < 0.01. Data are representative of three independent assays.

Fig. 6. ATF6-CHOP pathway was triggered by HG and promoted pro-inflammatory responses in macrophages.

A ER stress pathway-related molecules (IRE1, XBP1, PERK, ATF4, ATF6, and CHOP) in BMDMs cultured with LG and HG medium. B cATF6 and CHOP protein expression levels measured by western blot. PBA was added into BMDMs cultured in HG DMEM and were stimulated by LPS for another 24 h; untreated BMDMs differentiated under HG conditions were used as controls. C Protein levels of TNF-α, IL-6, and IL-10 in culture supernatants measured by ELISA. D PBA was added into BMDMs cultured in HG DMEM and were stimulated by LPS for another 24 h. Western blot detection of β-catenin, p-Akt, p-NF-κB p65, and β-actin. *P < 0.05, **P < 0.01. All the results are representative of at least three independent experiments.

Fig. 7. β-Catenin negative regulation by CHOP was critical for HG-mediated inflammatory responses in macrophages.

A, C CHOP in differentiated BMDMs cultured in HG conditions was inhibited by transfection of siRNA, and endogenous CHOP was validated by qRT-PCR and western blot. **P < 0.01. B, C β-Catenin levels were increased after CHOP downregulation in hyperglycemia-stressed BMDMs, as shown by qRT-PCR and western blot. *P < 0.05, **P < 0.01. D TNF-α, IL-6, and IL-10 concentration in culture supernatants were measured by ELISA at 0, 2, 6, 12, and 24 h. E Western blot detection of β-catenin expression in different groups of BMDMs with LPS (1 µg/ml) stimulation. F Quantification analysis of mRNAs of cytokines TNF-α, IL-6, and IL-10 (6 h). G Cytokine in supernatants was evaluated by ELISA (24 h). *P < 0.05. H Protein expression of p-Akt, p-NF-κB p65, and β-actin (6 h) was analyzed using western blotting. *P < 0.05, **P < 0.01. Data stand for three independent experiments with similar results.

Fig. 8. Administration of HG-cultured macrophages exacerbated hepatic IRI by ATF6–CHOP axis-mediated β-catenin.

A Flow chart of administration of HG-cultured macrophages into mice. B, C Liver function was evaluated by sALT and sAST level (n = 6). *P < 0.05, **P < 0.01. D Inflammatory cytokine levels in serum in different groups were determined by ELISA. E Histopathologic analysis of livers harvested 6 h after 90 min of liver ischemia, and severity elevation of liver injury was judged according to Suzuki’s histological grading. *P < 0.05, **P < 0.01.