FGF4 protects the liver from immune-mediated injury by activating CaMKK β-PINK1 signal pathway to inhibit hepatocellular apoptosis

Abstract

Immune-mediated liver injury (ILI) is a condition where an aberrant immune response due to various triggers causes the destruction of hepatocytes. Fibroblast growth factor 4 (FGF4) was recently identified as a hepatoprotective cytokine; however, its role in ILI remains unclear. In patients with autoimmune hepatitis (type of ILI) and mouse models of concanavalin A (ConA)- or S-100-induced ILI, we observed a biphasic pattern in hepatic FGF4 expression, characterized by an initial increase followed by a return to basal levels. Hepatic FGF4 deficiency activated the mitochondria-associated intrinsic apoptotic pathway, aggravating hepatocellular apoptosis. This led to intrahepatic immune hyper-reactivity, inflammation accentuation, and subsequent liver injury in both ILI models. Conversely, administration of recombinant FGF4 reduced hepatocellular apoptosis and rectified immune imbalance, thereby mitigating liver damage. The beneficial effects of FGF4 were mediated by hepatocellular FGF receptor 4, which activated the Ca²⁺/calmodulin-dependent protein kinasekinase 2 (CaMKKβ) and its downstream phosphatase and tensin homologue-induced putative kinase 1 (PINK1)-dependent B-cell lymphoma 2-like protein 1-isoform L (Bcl-X_L) signalling axis in the mitochondria. Hence, FGF4 serves as an early response factor and plays a protective role against ILI, suggesting a therapeutic potential of FGF4 and its analogue for treating clinical immune disorder-related liver injuries.

Keywords: Ca2+/Calmodulin dependent protein kinase 2; Fibroblast growth factor 4; Fibroblast growth factor receptor 4; Immune liver injury; Mitochondria-associated apoptosis; PTEN-induced putative kinase 1.

Graphical abstract

Figure 1

Intrahepatic FGF4 levels are associated with ILI pathogenesis. (A) Representative H&E staining images of human liver sections from patients with AIH and hepatic hemangioma. The areas infiltrated by lymphocytes are indicated by black triangles. Scale bar = 100 μm. (B) Serum levels of ALT and AST in the indicated groups. (C, D) Analyses of the mRNA levels of Fgf4 in human liver sections from the indicated groups using RNA scope. (E) Representative IHC images and quantification of FGF4 immunostaining in human liver sections from the indicated groups. Scale bar = 50 μm. (F) Representative H&E-stained images and statistical analysis (%) of focal hepatocyte death in liver tissues of mice treated with ConA for 0, 6, 12, 18, 24, 48, and 72 h. The area of dead hepatocytes is circled by broken white trace lines. Lymphocytes are indicated by black arrows. Scale bar = 100 μm. (G) Serum levels of ALT and AST in the indicated mice. (H, I) Analysis of the mRNA levels of Fgf4 in liver tissues from the indicated mice using RNA scope. Scale bar = 50 μm. (J) Relative mRNA levels of Fgf4 in liver tissues from the indicated mice as measured using qRT-PCR. (K, L) Western blot analysis of the expression of FGF4 in liver tissues from the indicated mice. Representative H&E-stained images (left) and HAI analysis (right) of liver sections from mice treated with S-100 for 0, 28, and 42 days. The areas infiltrated by lymphocytes are indicated by black triangles. Scale bar = 100 μm. Serum levels of ALT and AST in mice. Relative mRNA levels of Fgf4 in liver tissues from the indicated mice measured by qRT-PCR. Western blot analysis of FGF4 in liver tissues from the indicated mice. (A–E) Control group: n = 10, mild group: n = 14, severe group: n = 21; (F, G, I, J, L‒P) n = 6 per group. Data are the mean ± SEM. (B) Ordinary two-way analysis of variance (ANOVA) followed by Tukey's post hoc test; (D, E, M, O, P) Ordinary one-way ANOVA followed by Tukey's post hoc test. (N) Ordinary two-way ANOVA followed by Sidak post hoc test. n.s., not significant.

Figure 2

Hepatocyte-specific deletion of Fgf4 exacerbates ConA- and S-100-induced liver damage. (A) Representative H&E-stained images (left) and statistical analysis (%) (right) of focal hepatocyte death in liver tissues from saline- or ConA-treated Fgf4^fl/fl and Fgf4^−/− mice. Areas of dead hepatocytes are circled with white dotted trace lines, and lymphocytes are indicated by black arrows. Scale bar = 100 μm. (B, C) Serum levels of ALT and AST in the indicated groups. (D) Serum levels of pro-inflammatory cytokines in the indicated groups. (E) Changes in the infiltration rate of CD3⁺CD4⁺ T-cells in the livers of the indicated groups. (F) Representative H&E-stained images (left) and HAI analysis (right) of liver tissues from saline- or S-100-treated Fgf4^fl/fl and Fgf4^−/− mice. Lymphocytes are indicated by black arrows. Scale bar = 100 μm. (G) Serum levels of ALT and AST in the indicated groups. (H) Serum levels of pro-inflammatory cytokines in the indicated groups. (I) Serum levels of IgG in the indicated groups. (J) Changes in the infiltration rate of CD3⁺CD4⁺ T-cells in the livers of the indicated groups. (A‒D, F–I) n = 6 per group; (E, J) n = 3 per group. Data are the mean ± SEM. (A–E) Ordinary two-way ANOVA followed by Tukey's post hoc test. (F–J) Ordinary one-way ANOVA followed by Tukey's post hoc test. n.s., not significant.

Figure 3

Administration of recombinant FGF4 ameliorates ConA- and S-100-induced liver damage. (A) Schematic of the rFGF4 treatment regimen for mice with ConA-induced ILI. (B) Representative H&E staining images (left) and statistical analysis (%) (right) of focal hepatocyte death in liver tissues of ConA-induced ILI groups following treatments with different concentrations (L: 1.0 mg/kg, M: 1.5 mg/kg, H: 2.0 mg/kg body weight) of rFGF4. Areas of dead hepatocytes are circled using white dotted trace lines, and lymphocytes are indicated by black arrows. Scale bar = 100 μm. (C) Serum levels of ALT and AST in the indicated groups. (D) Serum levels of pro-inflammatory cytokines in the indicated groups. (E) Changes in the infiltration rate of CD3⁺CD4⁺ T-cells in the livers of the indicated groups. (F) Schematic representation of the rFGF4 treatment regimen in mice with S-100-induced ILI. (G) Representative H&E-stained images (left) and HAI analysis (right) of liver tissues from S-100-treated mice after rFGF4 administration. Lymphocytes are indicated by black arrows. Scale bar = 100 μm. (H) Serum levels of ALT and AST in the indicated groups. (I) Serum levels of IgG in the indicated groups. (G) Serum levels of pro-inflammatory cytokines in the indicated groups. (K) Changes in the infiltration rate of CD3⁺CD4⁺ T-cells in the livers of the indicated groups. (B‒D, G‒J) n = 6 per group; (E, K) n = 3 per group. Data are the mean ± SEM. (B) Ordinary one-way ANOVA followed by Dunnett's post hoc test. (C, H) Ordinary two-way ANOVA followed by Sidak post hoc test. (D, E, G, I–K) Two-tailed unpaired Student's t-test. n.s., not significant.

Figure 4

FGF4 targets hepatic FGFR4 to protect against ILI. (A) Relative mRNA levels of Fgfr1‒4 in the sorted hepatocytes, macrophages, T-cells, and NK cells from the saline or ConA-treated mice measured by qRT-PCR. (B) Co-localisation of FGFR4 expression (red) in hepatocytes with albumin (green) by immunofluorescence staining. Scale bar = 20 μm. (C) Schematic of experiment with hepatic knock-down of Fgfr4 (Fgfr4-LKD) in mice. (D) Relative mRNA levels of Fgfr4 in hepatocytes of Fgfr4-LKD and control mice measured by qRT-PCR. (E) Western blot analysis of FGFR4 in hepatocytes of Fgfr4-LKD and control mice. (F) Representative H&E-stained images and statistical analysis (%) of focal hepatocyte death in liver tissues from ConA- and rFGF4-treated Fgfr4-Ctrl and Fgfr4-LKD mice. Areas of dead hepatocytes are circled with white broken trace lines, and lymphocytes are indicated by black arrows. Scale bar = 100 μm. (G) Serum levels of ALT and AST in the indicated groups. (H) Serum levels of pro-inflammatory cytokines in the indicated groups. (I) Representative images and quantification of F4/80 (green) immunofluorescence staining with DAPI (blue) counterstaining in liver sections from the indicated groups. Scale bar = 50 μm. (J) Changes in the infiltration rates of CD3⁺CD4⁺ T-cells in the livers of the indicated groups. (A, J) n = 3 per group; (I) n = 4 per group; (D‒H) n = 6 per group. Data are the mean ± SEM. (A) Ordinary one-way ANOVA followed by Sidak post hoc test. (D, E) Two-tailed unpaired Student’s t-test. (F‒J) Ordinary two-way ANOVA followed by Tukey’s post hoc test. n.s., not significant.

Figure 5

FGF4 inhibits intrinsic apoptosis caused by ConA treatment in mouse hepatocytes. (A) Normalised enrichment scores of intrinsic apoptosis-related pathways in liver tissues from ConA-treated Fgf4^fl/fl and Fgf4^−/− mice based on RNA-seq data. n = 3 per group. (B) Normalised enrichment scores of intrinsic apoptosis-related pathways in liver tissues of ConA- and rFGF4-treated mice based on RNA-seq data. (C) Heatmap of representative differentially expressed genes related to intrinsic apoptosis in liver tissues of ConA-treated Fgf4^fl/fl and Fgf4^−/− mice based on RNA-seq data. (D) Heatmap of representative and differentially expressed genes related to intrinsic apoptosis in liver tissues from ConA- and rFGF4-treated mice based on RNA-seq data. (E, F) Western blot analysis of proteins related to the intrinsic apoptotic pathway upon hepatic FGF4 deficiency or rFGF4 treatment in total liver lysates from the indicated groups. (G–H) Western blot analysis of proteins related to the intrinsic apoptotic pathway upon hepatic FGF4 deficiency or rFGF4 treatment in hepatocytes isolated from the indicated groups. (A–D) n = 3 per group; (E–H) n = 6 per group. Data are the mean ± SEM. (E, G) Ordinary two-way ANOVA followed by Tukey's post hoc test. (F, H) Two-tailed unpaired Student's t-test. n.s., not significant.

Figure 6

Hepatic FGF4 inhibits apoptosis-evoked macrophage and T-cell infiltration under ILI conditions. (A, J) Changes in the frequency of Annexin V⁺ apoptotic hepatocytes isolated from the livers of the indicated groups. (B, K) Relative mRNA levels of Cx3cl1 in total liver lysates from the indicated groups. (C, L) Representative FCM of Ly6G⁻CD45⁺F4/80⁺CD11b⁺ macrophage infiltration in livers of the indicated groups. (D, M) Changes in the infiltration of Ly6G⁻CD45⁺F4/80⁺CD11b⁺ macrophages in livers of the indicated groups. (E, N) Recruitment of macrophages to apoptotic hepatocytes as indicated by albumin (blue), c-Casp3 (green), and F4/80 (blue) immunofluorescent staining in liver tissues from the indicated groups. Scale bar = 20 μm. (F, O) Changes in the intrahepatic infiltration of the Ly6G⁻F4/80⁺CD11b⁺MHCII⁺CD206^‒ M1-type and Ly6G⁻F4/80⁺CD11b⁺MHCII^‒CD206⁺ M2-type macrophages of the indicated groups. (G, P) Relative mRNA levels of Cx3cl1 in livers of the indicated groups. (H, Q) Relative mRNA levels of Hmgb1 in liver lysates of the indicated groups. (I, R) Changes in the infiltration of TCRβ⁺CD3⁺CD4⁺ T-cells in livers of the indicated groups. (A, D, I, J, M, R) n = 3 per group; (F, O) n = 3–4 per group; (B, G, H, K, P, Q) n = 6 per group. Data are the mean ± SEM. (A, G, J, P) Ordinary two-way ANOVA followed by Tukey's post hoc test; (B, D, H, I, K, M, Q, R) Two-tailed unpaired Student's t-test. n.s., not significant.

Figure 7

FGF4 regulates mitochondrial mPTP opening and Cyt C efflux through the CaMKKβ/PINK1/Bcl-X_L pathway. (A, B) Heatmap representation of differentially expressed genes involved in the CaMKKβ pathway upon hepatic FGF4 deficiency or rFGF4 treatment in liver tissues from the indicated groups based on RNA-seq data. (C, D) Western blot analysis of proteins involved in the CaMKKβ pathway in hepatocytes isolated from the indicated groups. (E, F) Western blot analysis of proteins downstream of the CaMKKβ pathway in mitochondria isolated from hepatocytes of the indicated groups. (G) Reciprocal co-immunoprecipitations of CaMKKβ and PINK1 in HEK293 cells. (H, I) Changes in the Calcein AM⁺ hepatocyte population isolated from the indicated groups reflecting mPTP opening. (J, K) Change in the flux of Cyt C in hepatocellular cytoplasm and mitochondria isolated from the indicated groups. (L) Western blot analysis of CaMKKβ phosphorylation levels in hepatocytes isolated from the indicated groups. (M) Western blotting analysis of proteins related to the PINK1/BCL-X_L pathway in hepatocellular mitochondria isolated from livers of the indicated groups. (A, B, H, I) n = 3 per group; (C–F, J‒M) n = 6 per group. Data are the mean ± SEM. (C–F, H, I) Two-tailed unpaired Student's t-test. (J, K) Ordinary two-way ANOVA followed by Sidak post hoc test. (L, M) Ordinary two-way ANOVA followed by Tukey's post hoc test. n.s., not significant.

Figure 8

Activation of the CaMKKβ/PINK1 pathway by FGF4 underlies its protective effect against ILI. (A, F) Representative H&E-stained images and statistical analysis (%) of focal hepatocyte death in liver tissues from the indicated groups. Areas of dead hepatocytes are circled with white dotted trace lines, and lymphocytes are indicated by black arrows. Scale bar = 100 μm. (B, G) Serum levels of ALT (upper) and AST (lower) in the indicated groups. (C, H) Serum levels of pro-inflammatory cytokines in the indicated groups. (D, I) Representative images and quantification of F4/80 (green) immunofluorescent stains with DAPI (blue) counterstains in liver sections from the indicated groups. Scale bar = 50 μm. (E, J) Changes in the infiltration of CD3⁺CD4⁺ T-cells in livers of the indicated groups. (K) Graphic model depicting the role of FGF4 in protecting the liver from ConA-induced ILI pathologies by activating the CaMKKβ-PINK1-Bcl-X_L signalling axis. (A‒C, F–H) n = 6 per group; (D, I) n = 4 per group; (E, J) n = 3 per group. Data are the mean ± SEM. (A–J) Ordinary two-way ANOVA followed by Tukey's post hoc test. n.s., not significant.