Folliculin depletion results in liver cell damage and cholangiocarcinoma through MiT/TFE activation

Abstract

Mutations in the tumor suppressor gene Folliculin (FLCN) are responsible for Birt-Hogg-Dube' (BHD) syndrome, a rare inherited condition that predisposes affected individuals to skin tumors, pulmonary cysts, and kidney tumors. FLCN regulates key cellular pathways, including TFEB, TFE3, and mTORC1, which are critical for maintaining cell homeostasis. Loss of FLCN leads to both hyperactivation of mTORC1 and constitutive activation of TFEB and TFE3, contributing to tumorigenesis. While previous studies showed that Flcn liver-specific conditional knockout (Flcn^LiKO) mice are protected from developing liver fibrosis and damage upon high-fat diet exposure, the potential role of FLCN loss in liver carcinogenesis remained unexplored. Here, we demonstrate that hepatic loss of FLCN in mice results in cancer associated with inflammation and fibrosis with features of cholangiocarcinoma (CCA). This phenotype emerges in mice over 90-week-old, with a male predominance. Moreover, Flcn^LiKO mice are more prone to develop diethylnitrosamine (DEN)- or 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC)- induced liver tumors with heterogenous histological features. Notably, depletion of TFE3, but not TFEB, in the liver of Flcn^LiKO mice fully rescues the cancer phenotype and normalized mTORC1 signaling, highlighting TFE3 as the primary driver of liver cancer and mTORC1 hyperactivity in the absence of FLCN.

Fig. 1. Characterization of liver-specific Flcn knockout male mice.

A Flcn expression levels after CRE recombination in 12-week-old Flcn^LiKO and control mice (n = 3 CTRL and n = 4 Flcn^LiKO). Serum ALT (B) and AST (C) levels in Flcn^LiKO (n = 6 at 4 weeks and n = 10 at 12 weeks) and control mice (n = 5 at 4 weeks and n = 7 at 12 weeks) at the indicated time points. D H&E staining of liver section from mice of the indicated genotypes at 4 and 12 weeks of age, showing clear cells and hepatocyte ballooning. Red arrows indicate immune infiltrates. E, F PAS and Sirius red staining (E), along with immunostaining for the indicated markers (F) in liver sections from 12-week-old Flcn^LiKO and control male mice, with relative quantification (n = 3 per group). G Volcano plot of differentially expressed genes (DEGs) between 12-week-old male Flcn^LiKO and control mice. H Gene Ontology (GO) analysis, including KEGG and Biological Process terms, highlighting significantly upregulated (red) and downregulated (blue) pathways in Flcn^LiKO mice compared to control (FDR < 0.05). All the data refer to male mice. Each dot represents an individual mouse. Data are mean ± standard error. Statistical analysis: Student t-test *p-value < 0.05, **p-value < 0.01.

Fig. 2. Sexual dimorphism in TFE3 and TFEB activation in Flcn-depleted livers.

A Immunoblot analysis of TFE3, TFEB, and mTORC1 target P-4EBP1 in total liver lysate from male and female Flcn^LiKO and control mice at 12 weeks of age, with relative quantifications (n = 3-4 per group). Nuclear fraction analysis with quantifications (n = 3-4 per group) (B) and immunostaining with relative quantifications (n = 3 per group) (C), showing TFE3 and TFEB nuclear localization in male and female Flcn^LiKO and control mice. Each dot represents an individual mouse. Blue bars indicate male, and purple bars indicate female mice. Data are mean ± standard error. Statistical analysis: One-way ANOVA: *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001 ; ****p-value < 0.0001.

Fig. 3. Spontaneous development of liver tumors in Flcn^LiKO mice.

A–D. Gross liver appearance and tumor area quantification (n = 6 Flcn^LiKO males and n = 5 Flcn^LiKO females) (A), liver-to-body weight (LW/BW) ratio (B), serum ALT (C) and AST (D) levels in 90-week-old male and female Flcn^LiKO mice and controls (n = 6 CTRL and n = 8 Flcn^LiKO). Tumor area is represented as percentage of the total liver area. The arrow in (A) indicates liver tumor. E, F. Histological characterization of liver pathology in 90-week-old Flcn^LiKO mice assessed by H&E, PAS, and Sirius red staining, as well as immunostainings for the indicated markers with relative quantifications (n = 3 per group). G Volcano plot of differentially expressed genes (DEGs) between 90-week-old male Flcn^LiKO and control mice. H Gene Ontology (GO) analysis, including KEGG and Biological Process terms, highlighting significantly upregulated (red) and downregulated (blue) pathways in Flcn^LiKO mice compared to control (FDR < 0.05). Each dot represents an individual mouse. Data are mean ± standard error. Statistical analysis: One-way ANOVA: *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001; ****p-value < 0.0001.

Fig. 4. DEN-induced tumorigenesis due to Flcn depletion.

A Schematic representation of the experimental plan. B Gross liver morphology and tumor area quantification in Flcn^LiKO and control mice injected with DEN at P21 and collected 8 months post-injection. Tumor area is represented as a percentage relative to controls (n = 3 CTRL, n = 14 male Flcn^LiKO mice, n = 3 female Flcn^LiKO mice). C, D H&E staining (C) and immunostainings for the indicated markers with relative quantifications (n = 3 per group) (D) in liver sections from Flcn^LiKO and control mice injected with DEN at P21 and collected 8 months post-injection. H&E staining reveals normal histological architecture in control mice (a,b), whereas Flcn^LiKO mice exhibit various histological subtypes of hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA), including clear cells (c), lobular and trabecular formation (d), biliary reaction (e) cyst formation (f) and immune infiltrates (d,g,h). Red arrows indicate immune infiltrates. Each dot represents an individual mouse. Data are mean ± standard error. Statistical analysis: One-way ANOVA: *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001; ****p-value < 0.0001.

Fig. 5. Impaired cell homeostasis due to FLCN loss during chronic liver damage and recovery.

A Schematic representation of the experimental plan. B Change in body weight of Flcn^LiKO and control mice during the injury protocol (n = 5 per group). C Liver-to-body weight (LW/BW) ratio of Flcn^LiKO and control mice at the indicated time points during liver injury and recovery (n = 4-5 per group). D, E Gross liver appearance (D) and serum ALT and AST levels (E) in Flcn^LiKO and control mice 2 weeks after removal of DDC-containing food (n = 5 per group). The blue arrow indicates a liver cyst. F–H H&E staining (F) and immunostaining for the indicated markers (G, H) in liver sections from Flcn^LiKO and control mice at the specified time points, with relative quantification (n = 3 per group). All the data refer to male mice. Each dot represents an individual mouse. Data are mean ± standard error. Statistical analysis: Two-way ANOVA or Student t-test: *p-value < 0.05; ****p-value < 0.0001.

Fig. 6. TFE3 is the principal contributor to liver pathology in Flcn^LiKO mice.

A–E Liver-to-body weight (LW/BW) ratio (A) and serum ALT (B) and AST (C) levels in control, Flcn^LiKO, FBKO, F3KO, and TKO mice 8 months after DEN injection (n = 12 CTRL and Flcn^LiKO, n = 6 FBKO, F3KO, TKO). Gross liver morphology (D) and histological analysis with relative quantifications (n = 3 per group) (E), showing partial rescue upon TFEB depletion and complete rescue in TFE3- or TFEB/TFE3-depleted livers. F Gene expression analysis of the indicated genes in liver samples from mice of the indicated genotypes (n = 4 per group). G Luciferase assay of Sox9 promoter activity demonstrating transactivation upon TFEB and TFE3 overexpression (n = 3 per group). All the data refer to male mice. Each dot represents an individual mouse. Data are mean ± standard error. Statistical analysis: One-way ANOVA: *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001; ****p-value < 0.0001.