Smad4 expression in hepatocytes is upregulated during NAFLD progression
To elucidate the relationship between Smad4 expression and NASH, the expression of Smad4 in healthy liver and NASH tissues from patients was assessed by tissue microarray using immunohistochemistry. We observed that Smad4 expression was significantly upregulated in NASH tissues compared to that in healthy controls (Fig. 1A, B). Moreover, we analyzed the publicly available Gene Expression Omnibus dataset GSE164760 and compared Smad4 expression between healthy and NASH tissues. The analysis revealed a significant increase in Smad4 mRNA levels in the NASH group compared to healthy controls (Fig. 1C).
To corroborate these findings, we used a short-term HFD mouse model of NAFLD (Fig. 1D). Immunohistochemical staining demonstrated a significant upregulation of Smad4 expression in the hepatocytes of fatty liver tissues (Fig. 1E). Double immunofluorescence staining confirmed that Smad4 was expressed in most hepatocytes (Fig. 1F). Consistent with this, Western blot analysis indicated a significant increase in Smad4 protein levels in fatty liver tissues of HFD-treated mice (Fig. 1G).
Collectively, these results suggest that Smad4 is activated in hepatocytes during NAFLD, indicating a potentially crucial role for Smad4 in NAFLD pathogenesis.
Hepatocyte-specific deletion of Smad4 attenuates high-fat diet-induced non-alcoholic fatty liver disease.
To further investigate the role of Smad4 in hepatocytes during NAFLD, we used a conditional Smad4 deletion approach in murine hepatocytes, as previously described 26. We generated hepatocyte-specific Smad4 knockout mice (Albumin-cre; Smad4flox/flox, AlbSmad4−/−) by crossing mice carrying the Loxp-flanked Smad4 allele with Albumin-cre mice. The AlbSmad4−/− mice were born at the expected Mendelian ratio, viable, and fertile. We used Smad4fl/fl littermates as controls. Smad4 deletion in primary hepatocytes of AlbSmad4−/− mice was confirmed using double immunofluorescence staining (Fig. 2A).
To delineate the role of Smad4 in NAFLD, we established a HFD-induced NAFLD model using AlbSmad4−/− mice and their control littermates. In response to HFD feeding, AlbSmad4−/− mice demonstrated a less pronounced increase in both body and liver weight compared with control littermates (Fig. 2B, C). After 3 months of HFD, AlbSmad4−/− mice displayed lower serum alanine aminotransferase (ALT), triglyceride (TG), and non-esterified fatty acid (NEFA) levels, while serum aspartate aminotransferase (AST) and total cholesterol (TC) levels were similar between AlbSmad4−/− and control mice (Fig. 2D). Furthermore, hepatic TG and NEFA levels were notably reduced in AlbSmad4−/− mice compared with Smad4fl/fl mice (Fig. 2E). AlbSmad4−/− mice also exhibited impaired glucose tolerance compared with Smad4fl/fl mice (Fig. 2F). The absence of Smad4 in hepatocytes in AlbSmad4−/− mice was further validated using double immunofluorescence staining and Western blot (Fig. 2G, H). Deletion of Smad4 in hepatocytes resulted in decreased fat accumulation, as evidenced by hematoxylin and eosin and Oil Red O staining. No significant differences were observed between AlbSmad4−/− and Smad4fl/fl mice who were fed a normal diet (Fig. 2G). Collectively, these data suggest that hepatocyte-specific Smad4 deficiency attenuates the development of HFD-induced NAFLD.
Smad4 deficiency in hepatocytes attenuated liver inflammation and CXCL1 secretion
To investigate whether Smad4 modulates liver inflammatory cell infiltration and hepatocyte proliferation, we stained liver tissues of AlbSmad4−/− and Smad4fl/fl mice via immunofluorescence. The infiltration of F4/80+ macrophages and CD11b+ monocytes was diminished in AlbSmad4−/− mice compared with that in Smad4fl/fl mice when fed a HFD. When fed the normal diet, immune cell infiltration was similar between AlbSmad4−/− and control mice (Fig. 3A). During NAFLD progression, activated chemokines evoke multiple cellular and tissue responses, including hepatocyte proliferation, activation, necrosis, angiogenesis, and immune cell recruitment27, 28. Notably, CXCL1 is a key gene involved in NAFLD progression. Therefore, we examined CXCL1 expression in the livers of AlbSmad4−/− and Smad4fl/fl mice through double immunofluorescence staining. CXCL1 levels were significantly lower in the livers of AlbSmad4−/− mice than those in Smad4fl/fl mice (Fig. 3B). This decrease in CXCL1 levels was further confirmed using quantitative reverse transcription polymerase chain reaction (RT-qPCR) in hepatocytes of NAFLD of AlbSmad4−/− mice (Fig. 3C).
To further clarify the function of Smad4 in hepatocytes, we knocked down Smad4 in AML12 cells using siRNA and verified Smad4 protein levels using Western blot (Fig. 3D). The cells were then treated with palmitic acid (PA) for 24 h to simulate an in vitro NAFLD environment. RT-qPCR analysis revealed that Smad4 deficiency mitigated the PA-induced CXCL1 expression in hepatocytes (Fig. 3E). The protein expression level of CXCL1 was further analyzed using enzyme-linked immunosorbent assay (ELISA) with AML12-conditioned medium (CM). Smad4 deficiency partly reduced the secretion of CXCL1 in PA-induced AML12 CM (Fig. 3F).
In addition, AML12 cells were transfected with a lentiviral vector to knock down Smad4, and the expression levels of Smad4 protein were measured using Western blot (Fig. 3G). In line with previous results, both the expression and secretion of CXCL1 in PA-stimulated Smad4-knockdown AML12 cells were significantly reduced compared to those in control cells (Figs. 3H, I). Taken together, these results suggest that hepatocyte-specific deletion of Smad4 lessens liver inflammation and CXCL1 secretion.
Hepatocyte Smad4 promotes CXCL1 secretion via the ASK1-P38-JNK signaling pathway
Previous studies have highlighted the involvement of the JNK and p38 MAPK cascades in the regulation of CXCL1 secretion29. Accordingly, we examined the levels of total and phosphorylated proteins involved in ASK1, P38, and JNK signaling. We found that the expression of phosphorylated ASK1 (p-ASK1), p38 (p-p38), and JNK (p-JNK) was diminished in the liver tissue of HFD-treated AlbSmad4−/− mice compared with Smad4fl/fl mice (Fig. 4A).
In subsequent experiments, we used si-Smad4 and sh-Smad4 to knock down Smad4 in AML12 cells, which were then exposed to PA. The Western blot analysis showed that the ASK1, P38, and JNK signaling pathways were activated in hepatocytes in response to PA administration. However, this activation was remarkably suppressed by Smad4 knockdown, suggesting that Smad4 knockdown considerably inhibited PA-induced activation of the ASK1–P38–JNK pathway in vitro (Figs. 4B, C).
To further elucidate the signaling pathways involved in the induction of CXCL1 secretion by Smad4, we employed the JNK inhibitor (SP600125) and the p38 MAPK inhibitor (SB203580). Both inhibitors suppressed PA-induced CXCL1 secretion and mRNA expression, as evidenced using RT–qPCR and ELISA analysis (Figs. 4D, E). In conclusion, these results suggest that Smad4 facilitates CXCL1 secretion via the ASK1-P38-JNK signaling pathways during NAFLD progression.
CXCL1 promotes fatty acid synthesis in hepatocytes by binding to CXCR2
To further investigate the role of Smad4 in hepatic lipid deposition, RT-qPCR was used to determine the expression levels of genes involved in fatty acid synthesis and consumption. We found that Smad4 deficiency considerably inhibited the expression of genes critical for fatty acid synthesis (ACC1, FASN, and SCD1) and fatty acid binding protein 1 (FABP1). However, we observed no substantial differences in the expression of genes related to fatty acid uptake (FATP1) and fatty acid β-oxidation (CPT1a and ACOX1) (Fig. 5A).
To corroborate these findings, we isolated primary hepatocytes from AlbSmad4−/− and Smad4fl/fl mice and then exposed them to PA. Following Smad4 knockout, the expression of genes responsible for fatty acid synthesis (ACC1, FASN, and PPAR-γ) was substantially reduced in primary hepatocytes (Fig. 5B). We further observed decreased lipid deposition in primary hepatocytes of AlbSmad4−/− mice compared with that in Smad4fl/fl mice following PA treatment, as evidenced by Oil Red O staining (Fig. 5C). This conclusion was confirmed using si-Smad4 AML12 and sh-Smad4 AML12 cells (Figs. 5D-G). Taken together, these findings suggest that hepatocyte Smad4 promotes NAFLD development primarily through fatty acid synthesis.
CXCL1 binds to specific receptor CXCR2. We then examined whether CXCL1-induced fatty acid synthesis occured in hepatocytes. We assessed the expression levels of CXCR2 in hepatocytes using Western blot (Fig. 5H). To further elucidate the molecular mechanisms underlying the effects of CXCL1 on fatty acid synthesis, we cultured AML12 cells with CXCL1 recombinant protein and analyzed the expressions of genes related to fatty acid synthesis using RT-qPCR. We observed that ACC1, FASN, and SCD1 were substantially upregulated by CXCL1. However, when CXCR2 activation was inhibited by the CXCR2 inhibitor, SB225002, in AML12 cells, CXCL1-induced ACC1, FASN, and SCD1 expression was reduced (Fig. 5I). These results suggest that CXCL1 induces fatty acid synthesis in hepatocytes via CXCR2.
CXCL1 promotes macrophage M1 polarization
Aberrant lipid-mediated hepatic inflammatory-immune dysfunction and chronic low-grade inflammation play important roles in NAFLD pathogenesis. Macrophage polarization is an important mechanism that regulates inflammatory responses30. Therefore, we assessed the quantities of CD86+ (M1 marker) and CD206+ (M2 marker) macrophages in NAFLD tissues. Our findings revealed a substantial decrease in M1 macrophages in AlbSmad4−/− mice compared with Smad4fl/fl mice, while M2-like macrophages remained comparable (Figs. 6A, B). We further evaluated the expression levels of genes associated with M1-like (IL-6, MCP1, and TNF-α) and M2-like (Arg1, IL-10, and YM1) phenotypes using RT-qPCR and found similar results (Figs. 6C, D). Given that the absence of hepatocyte Smad4 resulted in diminished CXCL1 secretion, we postulated that hepatocyte Smad4 might facilitate macrophage M1 polarization via CXCL1. We confirmed the expression of CXCR2 in macrophages using western blot (Fig. 6E). We cultured RAW264.7 cells with LPS and IFN-γ for 24 h to induce M1 polarization. It was found that LPS and IFN-γ activated the expression of CD86 in macrophages in comparison with the control group, CXCL1 recombinant proteins further upregulated CD86. This expression was further augmented by the addition of CXCL1 recombinant proteins, as evidenced by immunofluorescence staining. However, the treatment of RAW264.7 cells with the CXCR2 inhibitor SB225002 eliminated CXCL1-induced expression of CD86 (Fig. 6F). Following M1 polarization induction, the expression levels of iNOS, MCP1, and TNF-α were significantly increased in RAW264.7 cells compared with those in the control group, and the addition of exogenous CXCL1 recombinant protein further amplified M1-related gene expression (Fig. 6G).
For M2 polarization, RAW264.7 cells were cultured with IL-4 and IL-13 for 48 h. This treatment activated the expression of Arg1, IL-10, and YM1 in macrophages compared with the control group. No significant differences were observed after the administration of CXCL1 and SB225002 (Fig. 6H). These findings suggest that CXCL1 promotes the M1-type polarization of macrophages via CXCR2.