Fig 2: Increased FGF19 in hepatocellular carcinoma patients is associated with disturbances of the hydrosodic balance . A. Serum FGF19 distribution in a cohort of 173 patients with advanced HCC (FGF19 measured by ELISA). Median value and values at transitions between low, intermediate and high groups are indicated. B. Natremia (left panel) and MELD-Na score (right panel) in patients of the cohort according to the plasma FGF19 concentration C. Creatinin clearance in patients of the cohort according to the plasma FGF19 concentration. Mann Whitney test significance is indicated. ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001, ∗∗∗∗p ≤ 0.0001. n. s: not significant.

Fig 3: GSEA Pathways enriched in FGF19/Aldafermin + MYC driven tumors.(A) Upset plot depicting the number of unique and shared GSEA pathways enriched in each sample groups analyzed by RNAseq. FGF19 + MYC and Aldafermin + MYC samples were pooled for this analysis. (B) GSEA analysis of the 18 pathways specifically enriched in FGF19/Aldafermin + MYC vs sgTrp53 + MYC tumors. Kolmogorov–Smirnov statistical test significance is indicated. Data information: Threshold of significance for GSEA: p < 0.01. Complementary data are provided in Dataset EV1.

Fig 4: FGF19 and FGF15 cooperate with MYC to trigger liver carcinogenesis.(A) Experimental strategy to test FGF19/15 cooperation with other oncogenic events. Female C57BL/6J mice are subjected to hydrodynamic gene transfer (HGT) of a combination of oncogenes and the Sleeping-Beauty transposase 100 (SB100). (B) Representative livers and hematoxylin & eosin & Saffron (HES) stained sections following HGT with empty vector (RFP) or FGF19 plasmids, as indicated. (C) Kaplan–Meier curve of tumor-free survival for distinct oncogenic combinations, as indicated. Statistical significance was determined using Mantel-Cox (Log-rank) test; p value is indicated. (D) Representative livers and HES stained sections following hydrodynamic injection with RFP + MYC, FGF19 + MYC and FGF15 + MYC. (E) Liver/body weight ratio and tumor number per liver following HGT with the indicated combinations of oncogenes. (F) RT-qPCR expression analysis of transfected FGF 19/15 and HCC markers Glypican 3 and Alpha fetoprotein in RFP parenchyma (NT) and FGF19 + MYC and FGF15 + MYC tumors (T). Data information: Scale-bars: 1 cm (macroscopic liver), 100 µm (HES staining). Data are represented as mean ± SD. Number of mice per group and tumor incidence indicated are indicated in the figure. The data are representative of at least two independent experiments and all replicates shown are biological replicates. All tumoral samples come from independent mice. Statistical significance was evaluated using Mann–Whitney test performed on liver/body weight (E) or mRNA levels (F). Significant p values are indicated: ns (non significant) p > 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. RFP red fluorescent protein, T tumor, NT non-tumor parenchyma. Source data are available online for this figure.

Fig 5: Non-cell-autonomous activation of hepatocellular STAT3 by FGF19.(a) in vivo study design. 11–12-week old db/db mice received a single intraperitoneal injection of 1 mg kg−1 FGF19 or vehicle (saline), and livers were harvested 2 h post dose (n=5 per group). (b) Immunoblot analysis of pSTAT3Y705 in liver lysates of db/db mice treated with recombinant FGF19 protein. Anti-total-STAT3 and anti-β-actin serve as loading control. pERK and total ERK levels were also determined. (c) Primary hepatocytes were isolated by collagenase digestion followed by low-speed centrifugation and plating onto collagen-coated plates. (d) Lack of pSTAT3Y705 activation in primary mouse hepatocytes by FGF19. Cell lysates were prepared at the indicated time points following FGF19 stimulation and analysed for phosphorylation of the various proteins. Mouse IL-6 (mIL-6) was included as a positive control. (e) Lack of pSTAT3Y705 activation by FGF19 in primary human hepatocytes. Human IL-6 (hIL-6) was included as a positive control.