Fibroblast growth factor 18 stimulates the proliferation of hepatic stellate cells, thereby inducing liver fibrosis

Tsuchiya, Yuichi; Seki, Takao; Kobayashi, Kenta; Komazawa-Sakon, Sachiko; Shichino, Shigeyuki; Nishina, Takashi; Fukuhara, Kyoko; Ikejima, Kenichi; Nagai, Hidenari; Igarashi, Yoshinori; Ueha, Satoshi; Oikawa, Akira; Tsurusaki, Shinya; Yamazaki, Soh; Nishiyama, Chiharu; Mikami, Tetuo; Yagita, Hideo; Okumura, Ko; Kido, Taketomo; Miyajima, Atsushi; Matsushima, Kouji; Imasaka, Mai; Araki, Kimi; Imamura, Toru; Ohmuraya, Masaki; Tanaka, Minoru; Nakano, Hiroyasu

doi:10.1038/s41467-023-42058-z

Download PDF

Article
Open access
Published: 09 October 2023

Fibroblast growth factor 18 stimulates the proliferation of hepatic stellate cells, thereby inducing liver fibrosis

Nature Communications volume 14, Article number: 6304 (2023) Cite this article

6258 Accesses
2 Citations
16 Altmetric
Metrics details

Subjects

Abstract

Liver fibrosis results from chronic liver injury triggered by factors such as viral infection, excess alcohol intake, and lipid accumulation. However, the mechanisms underlying liver fibrosis are not fully understood. Here, we demonstrate that the expression of fibroblast growth factor 18 (Fgf18) is elevated in mouse livers following the induction of chronic liver fibrosis models. Deletion of Fgf18 in hepatocytes attenuates liver fibrosis; conversely, overexpression of Fgf18 promotes liver fibrosis. Single-cell RNA sequencing reveals that overexpression of Fgf18 in hepatocytes results in an increase in the number of Lrat⁺ hepatic stellate cells (HSCs), thereby inducing fibrosis. Mechanistically, FGF18 stimulates the proliferation of HSCs by inducing the expression of Ccnd1. Moreover, the expression of FGF18 is correlated with the expression of profibrotic genes, such as COL1A1 and ACTA2, in human liver biopsy samples. Thus, FGF18 promotes liver fibrosis and could serve as a therapeutic target to treat liver fibrosis.

Upregulation of cancer-associated gene expression in activated fibroblasts in a mouse model of non-alcoholic steatohepatitis

Article Open access 20 December 2019

Single-cell and bulk transcriptomics of the liver reveals potential targets of NASH with fibrosis

Article Open access 29 September 2021

Single-cell transcriptome and cell type-specific molecular pathways of human non-alcoholic steatohepatitis

Article Open access 05 August 2022

Introduction

Liver fibrosis results from chronic liver injury triggered by many factors, including viral infection, excess alcohol intake, lipid accumulation, and autoimmune diseases^1,2. Liver fibrosis ultimately develops into liver failure and cancer; hence, liver fibrosis is tightly associated with cause-specific mortality in several liver diseases^3,4. In addition to liver fibrosis caused by chronic viral hepatitis, nonalcoholic steatohepatitis (NASH), which is now renamed metabolic dysfunction-associated steatohepatitis (MASH)⁵, is one of the leading causes of liver fibrosis. Simple steatosis and MASH belong to the same disease category characterized by metabolic dysfunction-associated steatotic liver disease (MASLD)⁵. Approximately 10% of patients with MASLD are ultimately diagnosed with MASH, which is characterized by hepatocyte steatosis, inflammation, and hepatocyte cell death^6,7. Although MASH is presumably caused by cell death due to lipid-mediated toxicity or oxidative stress, the detailed mechanisms are not fully understood^8,9.

Recent advancements in single-cell RNA sequencing (scRNA-seq) technology have revealed that cell‒cell interactions play a crucial role in the development of liver fibrosis^10,11,12. Among the various cells involved in liver fibrosis, hepatic stellate cells (HSCs) are central players in promoting liver fibrosis^13,14. Under physiological conditions, quiescent HSCs mainly localize in the space of Disse and contain large amounts of vitamin A, thereby contributing to vitamin A homeostasis. Quiescent HSCs have retinol‐containing lipid droplets and express the characteristic marker genes, lecithin retinol acyltransferase (Lrat) and desmin (Des)¹⁰. In response to liver injury, HSCs become activated and activated HSCs may lose the expression of Lrat but evolve into α-SMA-expressing myofibroblasts that highly express profibrotic genes^13,14. Various cytokines have been shown to activate HSCs; for example, Hedgehog, connective tissue growth factor (CTGF), and transforming growth factor beta (TGFβ) activate HSCs¹⁵. However, it remains unclear which growth factors control the numbers of quiescent and activated HSCs and myofibroblasts during the development of fibrosis.

Fibroblast growth factors (FGFs) comprise 22 structurally related proteins and regulate various biological and cellular responses, including organogenesis, tissue remodeling, angiogenesis, and regulation of metabolism^16,17. The extracellular FGF family comprises the FGF1, FGF4, FGF7, FGF8, and FGF19 subfamilies, and FGFs transmit signals through cognate receptors, including FGF receptors 1 to 4. The FGF1, FGF4, FGF7, and FGF8 subfamilies exert their activities in a paracrine or autocrine manner, whereas the FGF19 subfamily acts as a hormone^16,17. Of note, members of the FGF family play crucial roles in regulating liver fibrosis¹⁸. For example, FGF1 and FGF2 suppress the activation of HSCs^19,20, thereby attenuating liver fibrosis. However, another study reported that deletion of Fgf1 and Fgf2 attenuates CCl₄-induced liver fibrosis²¹, suggesting that FGF1 and FGF2 exhibit a profibrotic role under certain conditions. While hepatic fibrosis is attenuated in Fgf15-deficient mice in murine MASH and CCl₄-induced models²², administration of an FGF21 analog attenuates hepatic inflammation and fibrosis in a murine MASH model²³. Together, these results suggest that members of the FGF family attenuate or exacerbate liver fibrosis depending on the experimental conditions.

Cellular FLICE-inhibitory protein (cFLIP), which is encoded by Cflar, is a caspase 8 homolog but does not have protease activity; it binds to caspase 8 and suppresses apoptosis²⁴. Germline deletion of Cflar results in embryonic lethality due to enhanced apoptosis and necroptosis^25,26, but the in vivo functions of Cflar have been extensively investigated by deleting Cflar in specific tissues²⁷. We and others have reported that hepatocyte-specific Cflar-deficient (Cflar^LKO) mice have increased susceptibility to death ligand-induced cell death^28,29. Using Cflar^LKO mice, we identified that histone H3 is released from apoptotic hepatocytes and induces endothelial cell injury³⁰.

In the present study, we utilized Cflar^LKO mice to identify factors that promote liver fibrosis. We identified FGF18 as a critical factor promoting liver fibrosis in murine chronic liver injury models. scRNA-seq revealed that overexpression of FGF18 increases the number of Lrat⁺ HSCs. In vitro experiments showed that FGF18 induces the proliferation of HSCs by inducing the expression of Ccnd1 but suppresses TGFβ-induced profibrotic gene expression, thereby providing a microenvironment for the proliferation of HSCs. Moreover, the expression of FGF18 is correlated with the expression of profibrotic genes, such as COL1A1 and ACTA2, in human liver biopsy samples. Our findings suggest that FGF18 promotes liver fibrosis and could serve as a therapeutic target to treat liver fibrosis.

Results

Cflar ^LKO mice develop mild liver fibrosis

Since liver fibrosis is tightly associated with apoptosis of hepatocytes³¹, we crossed Cflar^flox mice with albumin promoter-driven Cre recombinase transgenic (Alb-Cre) mice to generate Cflar^LKO mice that were utilized to identify factors promoting liver fibrosis. As we previously reported²⁹, the expression of cFLIP at the mRNA and protein levels diminished, but was not entirely abrogated, in hepatocytes of Cflar^LKO mice (Supplementary Fig. 1a, b). Serum alanine aminotransferase (ALT) concentrations were slightly increased in Cflar^LKO mice compared to Cflar^FF mice (Supplementary Fig. 1c), suggesting that mild liver injury occurred in Cflar^LKO mice under homeostatic conditions. Immunohistochemistry (IHC) revealed that the numbers of cleaved caspase 3 (CC3)⁺ and TUNEL⁺ cells were slightly increased in the livers of Cflar^LKO mice (Supplementary Fig. 1d–f). Notably, Sirius Red⁺ areas, the hydroxyproline content, and the expression of Col1a2 and Col3a1, which are hallmarks of fibrosis, were increased in the livers of Cflar^LKO mice compared to Cflar^FF mice (Supplementary Fig. 1d, g–i). Moreover, CK19⁺ areas and expression of Krt19, which encodes CK19, were increased in the livers of Cflar^LKO mice (Supplementary Fig. 1d, j, k), suggesting that the ductular reaction was enhanced. These results indicate that Cflar^LKO mice spontaneously developed mild liver fibrosis and ductular reaction.

Liver fibrosis is exacerbated in Cflar ^LKO mice fed the CDE diet

To identify factors that promote liver fibrosis, we treated or fed Cflar^LKO mice various agents or diets. Among them, we chose the choline-deficient ethionine-supplemented (CDE) diet because our preliminary experiments revealed that the CDE diet resulted in drastic exacerbation of liver fibrosis in Cflar^LKO mice (Fig. 1). We first determined the experimental conditions for the CDE diet, such as the sex and ages of mice, to efficiently induce liver fibrosis using wild-type mice. We found that 8-week-old female mice fed the CDE diet for 4 weeks were suitable for the liver fibrosis model (Supplementary Fig. 2). Then, we compared the phenotypes of 8-week-old female Cflar^FF and Cflar^LKO mice fed the normal or CDE diet for 4 weeks. ALT concentrations were significantly elevated in Cflar^LKO mice fed the CDE diet compared to Cflar^FF mice fed the CDE diet or Cflar^LKO mice fed the normal diet (Fig. 1a). The body weights (BWs) of Cflar^FF and Cflar^LKO mice fed the normal diet gradually increased at 12 weeks after birth (left panel, Fig. 1b). However, the BW of Cflar^LKO mice was progressively decreased during the CDE diet (right panel, Fig. 1b). In contrast, BW recovered to pre-diet levels before the CDE diet in Cflar^FF mice (right panel, Fig. 1b). As expected, many hepatocytes from Cflar^FF and Cflar^LKO mice fed the CDE diet were filled with cytoplasmic fat droplets (Fig. 1c). Consistent with an increase in ALT levels in Cflar^LKO mice after 4 weeks of the CDE diet, the numbers of both CC3⁺ and TUNEL⁺ cells were increased in the livers of Cflar^LKO mice (Fig. 1d, e). Moreover, Sirius Red⁺ and desmin⁺ areas, the hydroxyproline content, and the expression of Col1a2 and Col3a1 were increased in the livers of Cflar^LKO mice compared to Cflar^FF mice fed the CDE diet for 4 weeks (Fig. 1c, f–i). CK19⁺ areas and Krt19 expression were increased in the livers of Cflar^LKO mice (Fig. 1c, j, k). Given that desmin is a universal marker for HSCs¹⁰, these results indicate that Cflar^LKO mice rapidly developed severe liver fibrosis along with an increase in the number of HSCs after being fed the CDE diet.

**Fig. 1: CDE diet-induced liver injury is exacerbated in *Cflar*^LKO mice.**

Fibroblast growth factor 18 is elevated in the livers of Cflar ^LKO mice

To identify the gene(s) involved in the exacerbation of liver fibrosis in Cflar^LKO mice, we compared the gene expression profiles of Cflar^FF and Cflar^LKO mice fed the CDE diet for 4 weeks by whole liver RNA sequencing (RNA-seq) (Supplementary Data 1). Gene Ontology (GO) enrichment analysis revealed that genes related to cell adhesion, extracellular matrix organization, and cell migration were enriched in the livers of Cflar^LKO mice compared to Cflar^FF mice (Fig. 2a, b). As the desmin⁺ and CK19⁺ areas were increased in the livers of Cflar^LKO mice (Fig. 1c, g, j), we focused on growth factor-related genes. After investigating the mRNA expression kinetics in the livers of 8-week-old female mice before and after CDE diet feeding for 4 weeks (Fig. 2c), we focused on Fgf18, Tnfsf13b, and Wnt5a. We tested whether overexpression of these genes induced the proliferation of desmin⁺ and CK19⁺ cells in the livers of wild-type mice using the hydrodynamic tail vein injection (HTVi) method³². HTVi is an in vivo transfer method that yields high gene expression levels in the liver. We first confirmed that infused genes were overexpressed in the livers using qPCR and IHC for EGFP as a positive control (Supplementary Fig. 3a–c). Unexpectedly, the Ki67⁺ areas and, to a lesser extent, desmin⁺ areas, but not the CK19⁺ areas, were increased in the livers injected with Fgf18 compared to those injected with Egfp, Tnfsf13b, or Wnt5a 7 days after injection (Fig. 2d, e). Of note, most Ki67⁺ cells did not express CK19, suggesting that these cells were mostly noncholangiocytes (Supplementary Fig. 3d, e). Thus, we focused on FGF18 for subsequent experiments.

To determine the cellular source of FGF18, we examined the livers by RNAscope, since reliable antibodies against FGF18 for IHC are unavailable as far as we explored. We visualized Fgf18 mRNA along with Hnf4a mRNA (a hallmark of hepatocytes) or Lrat mRNA (a hallmark of HSCs) by RNAscope. Fgf18 mRNA⁺ puncta were very closely located to Hnf4a⁺ puncta or Lrat⁺ puncta (Fig. 2f), suggesting that hepatocytes and HSCs expressed Fgf18. Of note, Fgf18 mRNA⁺ puncta areas were increased in the livers of Cflar^LKO mice compared to those of Cflar^FF mice fed the CDE diet for 4 weeks (Fig. 2f, g). Although Fgf18 mRNA⁺ puncta were detected in comparable numbers in Hnf4a⁺ hepatocytes in Cflar^FF and Cflar^LKO mice, Fgf18 mRNA⁺ puncta in nonparenchymal liver cells (NPCs) (presumed Lrat⁺ HSCs) were increased in Cflar^LKO livers compared to Cflar^FF livers (Fig. 2g). Therefore, both HSCs and hepatocytes expressed Fgf18 under these experimental conditions.

Intriguingly, the expression of Fgf18 was also elevated in choline-deficient, L-amino acid-defined, high-fat diet (CDAHFD)-induced murine MASH or 3,5-diethoxycarbonyl-1,4-dihydrocollidine–supplemented (DDC) diet-induced liver fibrosis models in wild-type mice^33,34 (Supplementary Fig. 4a–h). Reanalysis of GSE99010³⁵ revealed that the expression of Fgf18 was correlated with the expression of profibrotic genes, such as Col1a1 and Col3a1, in the livers of mice fed the normal diet or the Western diet combined with CCl₄ injections for 24 weeks but not 12 weeks (Supplementary Fig. 4i, Supplementary Table 1). Of note, the expression of Fgf18 was elevated at relatively late stages compared to that of other profibrotic genes, such as Acta2, Col1a2, Col3a1, and Tgfb1-3 (Supplementary Fig. 4j). These results suggest that increased expression of Fgf18 is observed in multiple liver fibrosis models, especially in the advanced stages of liver fibrosis.

Hepatocyte-specific deletion of Fgf18 attenuates fibrosis in Cflar ^LKO mice

To determine whether FGF18 attenuates or exacerbates liver fibrosis in Cflar^LKO mice, we generated hepatocyte-specific Fgf18-deficient mice (Fgf18^LKO mice) by crossing albumin-Cre mice with Fgf18^FF mice³⁶. Although germline deletion of Fgf18 results in perinatal death due to a defect in alveolar dilatation³⁷, Fgf18^LKO mice were born at the expected Mendelian ratio and grew to adulthood. As expected, Fgf18 expression was decreased in hepatocytes in Fgf18^LKO mice (Supplementary Fig. 5a). The expression of the Cflar gene was slightly elevated in hepatocytes and NPCs in Fgf18^LKO mice (Supplementary Fig. 5a, b), but the underlying mechanism is currently unknown. Fgf18^LKO mice fed the normal diet did not exhibit any liver abnormalities (Supplementary Fig. 5c, d), and compared to Fgf18^FF mice, Fgf18^LKO mice fed the CDE diet did not show further exacerbation or attenuation of liver fibrosis (Supplementary Fig. 5e to m).

We generated mice lacking both Cflar and Fgf18 in hepatocytes (Cflar^LKO;Fgf18^LKO mice). With the normal diet, the expression of Fgf18 was diminished in hepatocytes but elevated in NPCs from Cflar^LKO;Fgf18^LKO mice compared to Cflar^FF;Fgf18^FF mice (Fig. 3a). We also found that Fgf18 mRNA expression was elevated in NPCs from Cflar^LKO mice compared to NPCs from Cflar^FF mice fed the normal diet (Fig. 3b). However, the expression of Cflar at the mRNA and protein levels was diminished in hepatocytes, but not NPCs, from Cflar^LKO;Fgf18^LKO mice (Fig. 3a, c). These results suggest that the expression of Fgf18 in NPCs, but not hepatocytes, was elevated in the absence of Cflar in mouse hepatocytes, even under normal diet conditions.

**Fig. 3: Hepatocyte-specific deletion of *Fgf18* attenuates fibrosis in the livers of *Cflar*^LKO mice fed the CDE diet.**

We then compared the extent of liver injury and fibrosis among Fgf18^LKO, Cflar^LKO, and Cflar^LKO;Fgf18^LKO mice following 4 weeks of CDE diet feeding. We utilized the results of Fgf18^FF and Fgf18^LKO mice fed the CDE diet in Supplementary Fig. 5. Serum concentrations of ALT and aspartate aminotransferase (AST) were elevated in Cflar^LKO mice compared to Fgf18^LKO mice but comparable to those in Cflar^LKO;Fgf18^LKO mice. In contrast, serum concentrations of alkaline phosphatase (ALP) did not differ among Fgf18^LKO, Cflar^LKO, and Cflar^LKO;Fgf18^LKO mice (Fig. 3d). Fgf18 expression was elevated in the livers of Cflar^LKO mice compared to Fgf18^LKO mice, but its expression was reduced in the livers of Cflar^LKO;Fgf18^LKO mice (Fig. 3e), confirming that hepatocytes in the livers of Cflar^LKO mice expressed Fgf18 to some extent. Intriguingly, the Sirius Red⁺ areas and hydroxyproline content, but not the desmin⁺ areas, were elevated in the livers of Cflar^LKO mice compared to Fgf18^LKO mice, but their expression was reduced in the livers of Cflar^LKO;Fgf18^LKO mice (Fig. 3f, g–i). Accordingly, the expression of Col3a1 and Acta2 was elevated in the livers of Cflar^LKO mice compared to Fgf18^LKO mice but attenuated in the livers of Cflar^LKO;Fgf18^LKO mice (Fig. 3j). These results suggest that FGF18 derived from hepatocytes contributes, at least in part, to liver fibrosis. The partial attenuating effect of Fgf18 deletion on liver fibrosis may be caused by residual expression of Fgf18 in HSCs or compensation by other cytokines. In contrast, CK19⁺ cells and Krt19 expression were elevated in the livers of Cflar^LKO mice compared to Fgf18^LKO mice, but their expression was not decreased in the livers of Cflar^LKO;Fgf18^LKO mice (Fig. 3f, k, l).

Overexpression of Fgf18 in hepatocytes promotes liver fibrosis

To determine whether the expression of FGF18 alone induces liver fibrosis in vivo, we generated transgenic mice expressing Fgf18 in hepatocytes. We integrated Fgf18 cDNA into the Rosa 26 locus under the control of the CAG promoter interrupted by a cassette containing loxP-STOP-loxP (LSL) sites (Fig. 4a), enabling us to express Fgf18 in a Cre recombinase-dependent manner³⁸. We crossed Rosa26-CAG-LSL-Fgf18 Tg mice with albumin-Cre Tg mice, generating mice that expressed Fgf18 in hepatocytes. Hereafter, we refer to Alb-Cre;Rosa26-CAG-LSL-Fgf18 Tg mice as Fgf18 Tg mice. Fgf18 Tg mice appeared normal until adulthood based on their BW and appearance. However, Fgf18 Tg mice spontaneously developed liver hypertrophy; their liver weights were approximately 1.5-fold greater than those of non-Tg mice (Fig. 4b–d). Fgf18 mRNA expression was elevated in hepatocytes from Fgf18 Tg mice compared to those from non-Tg mice (Fig. 4e). We also detected the protein expression of FGF18 in hepatocytes, but not NPCs, from Fgf18 Tg mice by Western blotting (Fig. 4f). Of note, due to glycosylation, endogenous mFGF18 in hepatocytes migrated slower than recombinant mFGF18 in E. coli. Moreover, our in-house enzyme-linked immunosorbent assay (ELISA) for human FGF18 detected mouse FGF18 in the culture supernatants of hepatocytes from Fgf18 Tg mice³⁹ (Fig. 4g). In sharp contrast, we could not detect protein expression of FGF18 in either hepatocytes or NPCs from Cflar^LKO, Fgf18^LKO, or Cflar^LKO;Fgf18^LKO mice due to the lower expression levels of Fgf18 (Supplementary Fig. 6a–c).

Serum ALT, AST, and ALP concentrations did not differ between non-Tg and Fgf18 Tg mice (Fig. 4h). H&E-stained liver tissues and IHC revealed that the numbers of mononuclear cells surrounding the large vessels, including Ly6G⁺ neutrophils and CD68⁺ monocytes, were increased in the livers of Fgf18 Tg mice (Fig. 4i, Supplementary Fig. 7a–d). These cells appeared positive for Ki67 and were considered proliferating (Supplementary Fig. 7e, f). Flow cytometry revealed that the percentages of B cells, monocytes (CD11b⁺Ly6G^-), and neutrophils (CD11b⁺Ly6G⁺) were increased in the livers of Fgf18 Tg mice compared to those of non-Tg mice (Supplementary Fig. 7g–i). In several portal areas, the number of small vessels was increased (Fig. 4i), suggesting that aberrant angiogenesis occurred in the livers of Fgf18 Tg mice.

Many cytokines were elevated in the livers of Cflar^LKO and Fgf18 Tg mice compared to the respective control mice (Supplementary Fig. 8a). However, deletion of Fgf18 only marginally attenuated the expression of several inflammatory cytokines (Supplementary Fig. 8b), suggesting that FGF18 in hepatocytes did not play a dominant role in the induction of these cytokines under Cflar-deficient conditions but could upregulate them when it was overexpressed.

Intriguingly, Sirius Red⁺ areas and the hydroxyproline content were elevated in the livers of Fgf18 Tg mice compared to non-Tg mice (Fig. 4j–l). CK19⁺ areas and Krt19 expression were increased in the livers of Fgf18 Tg mice (Fig. 4m–o). These results suggest that overexpression of Fgf18 alone results in liver fibrosis and ductular reaction.

FGF18 increases the number of CD31^-CD34⁺ stromal cells

We performed RNA-seq of the whole liver to elucidate the mechanisms underlying FGF18-induced liver fibrosis (Supplementary Data 2). GO enrichment analysis revealed that genes associated with cell adhesion, angiogenesis, collagen fibril organization, and growth factors were enriched in the livers of Fgf18 Tg mice compared to non-Tg mice (Fig. 5a, b). We extracted genes elevated more than 2-fold in the livers of two murine models: Cflar^LKO mice fed the CDE diet and Fgf18 Tg mice. A total of 600 genes were upregulated in both models (Fig. 5c; Supplementary Data 3). GO enrichment analysis revealed that these overlapping genes were categorized in cell adhesion, extracellular matrix organization, and collagen fibril organization (Fig. 5d), suggesting that expression of Fgf18 alone upregulates, at least in part, signature genes involved in CDE-induced liver fibrosis.

As the expression of Cd34 was elevated in the whole liver RNA-seq of both Cflar^LKO mice fed the CDE diet and Fgf18 Tg mice (Figs. 2b, 5b), we focused on Cd34. CD34 is a marker of hematopoietic progenitor cells but is also expressed in vascular endothelial cells, mesenchymal stromal cells, and activated HSCs⁴⁰. We used flow cytometry to analyze and characterize CD34⁺ cells. CD34⁺ cells were composed of CD31⁺CD34⁺ cells and CD31⁻CD34⁺ cells, and the percentage of the latter cells was increased in the livers of Fgf18 Tg mice (Fig. 5e, f). Most CD31^-CD34⁺ cells expressed PDGFRα and Sca1, with minor populations expressing Thy1 or podoplanin (Fig. 5e–g). Importantly, sorted CD31^-CD34⁺ cells expressed Lrat, Hgf, Ngfr, Acta2, and Col1a2 at higher levels than sorted CD31⁺ cells (Supplementary Fig. 9a, Fig. 5h), indicating that CD31^-CD34⁺ cells indeed contained HSCs. RNAscope and IHC further confirmed that Lrat-expressing and desmin⁺ HSCs were increased in the livers of Fgf18 Tg mice compared to non-Tg mice (Fig. 5i–l).

IHC revealed that the numbers of CD34⁺ cells coexpressing the p75 nerve growth factor receptor (p75NTR) encoded by Ngfr, a marker of HSCs⁴¹, and vimentin were increased in the livers of Fgf18 Tg mice (Supplementary Fig. 9b, c). The expression of Cd34 and the numbers of CD34⁺ cells were increased in the livers of Cflar^LKO mice fed the CDE diet and wild-type mice fed the CDAHFD diet (Supplementary Fig. 9d–i), suggesting that the expansion of CD34⁺ cells is not specific to Fgf18 Tg mice but a more generalized phenomenon in liver fibrosis.

Characterization of CD31^-CD34⁺ stromal cells and cell‒cell communication by scRNA-seq

To further characterize CD31^-CD34⁺ cells, we isolated lineage marker-negative (CD31^- CD45.2^- CD146^- Epcam^- Ter119^-) NPCs from the livers of non-Tg and Fgf18 Tg mice (Supplementary Fig. 10a) and characterized these cells by scRNA-seq. We identified 22 clusters in NPCs, hepatocytes, and myeloid cells isolated from non-Tg and Fgf18 Tg mice (Fig. 6a, Supplementary Fig. 10b, Supplementary Table 2). Lrat⁺ HSC clusters comprised clusters 1, 3, and 8 (Fig. 6a), whereas Acta2⁺ myofibroblast and Thy1⁺ portal fibroblast clusters included clusters 16 and 12, respectively (Fig. 6a). Cell composition analysis of each cluster from non-Tg and Fgf18 Tg mice revealed that the percentages of HSCs were increased in Fgf18 Tg mice (Fig. 6b). Violin plots from pooled data from non-Tg and Fgf18 Tg mice showed that several genes, including Adamtsl2, Des, Fgfr2, Hgf, and Ngfr, were specifically expressed in HSCs. In contrast, other genes, including Tgfbr3, Tgfbi, and Tgfb1i1, were ubiquitously expressed in both HSCs and fibroblasts (Fig. 6c, Supplementary Data 4).

**Fig. 6: Characterization of CD31^-CD34⁺ stromal cells and cell‒cell communication analyzed by scRNA-seq.**

Previous studies reported signature genes of quiescent (qHSCs) and activated HSCs (aHSCs) by scRNA-seq^10,11. Notably, all Lrat⁺ HSC clusters expressed signature genes of qHSCs, including Ank3, Colec11, Ecm1, Fcna, Gucy1b1, Reln, and Vipr1, and signature genes of aHSCs, including Col1a2, Col3a1, Lgals1, Thbs1, and Vim (Supplementary Fig. 10c, Supplementary Data 4). In contrast, Acta2⁺ myofibroblast and Thy1⁺ portal fibroblast clusters expressed signature genes of aHSCs, including Col1a2, Col3a1, Dpt, Igf1, Lgals1, S100a6, and Vim (Supplementary Fig. 10c, Supplementary Data 4). These results suggest that FGF18 mainly expanded cell populations expressing qHSC marker genes.

To integrate the pathways triggered by FGF18 with the phenotypes of the livers in Fgf18 Tg mice, we analyzed them using the cell‒cell communication tool CellChat⁴². Of note, Fgfr2 was expressed in HSCs, whereas Fgfr1 was expressed in fibroblasts and minor populations of HSCs (Fig. 6c). Although read-through transcripts of Fgf18 were expressed in various tissues from Fgf18 Tg mice, FGF18 signals converged in HSCs and fibroblasts through Fgfr2 and Fgfr1, respectively (Fig. 6d, e). Similarly, the TGFβ1 signal merged in HSCs and portal fibroblasts (Fig. 6d, e). Moreover, several chemokines mainly produced by HSCs, including Ccl19, Cxcl12, and Cx3cl1, may induce signals in neutrophils, monocytes, dendritic cells, and B cells (Fig. 6d, e).

As our scRNA-seq did not contain liver sinusoidal endothelial cell (LSEC) and cholangiocyte, we integrated previously published scRNA-seq datasets from unstimulated liver NPCs with our scRNA-seq datasets¹¹. The combined UMAP contained LSEC and cholangiocyte (Supplementary Fig. 11a). CellChat revealed that FGF18 may stimulate cholangiocyte through Fgfr1-3 (Supplementary Fig. 11b, c) and that Vegfa from HSC and cholangiocyte may induce signals in LSEC and HSC (Supplementary Fig. 11b, c). Together, these data may explain the mechanisms underlying the proliferation of CK19⁺ cholangiocytes and angiogenesis.

Regarding liver hypertrophy, FGF18 did not activate hepatocytes (Supplementary Fig. 12a, b). The expression of Hgf was elevated in the livers of Fgf18 Tg mice, and Lrat⁺ HSCs highly expressed Hgf (Figs. 5b, 6c). Given that FGF18 did not upregulate the expression of Hgf in HSCs (Supplementary Fig. 12c), FGF18 may induce the proliferation of hepatocytes by HGF released from expanded HSCs that highly expressed Hgf. HGF induced ERK phosphorylation and Jun, Ccnd1, and Myc upregulation (Supplementary Fig. 12a, d). Therefore, these results suggest that FGF18 stimulates the proliferation of HSCs that provide an HGF-enriched microenvironment to stimulate hepatocyte proliferation, thereby causing liver hypertrophy.

FGF18 stimulates the proliferation of HSCs but suppresses TGFβ-induced profibrotic gene expression

To further investigate the mechanisms by which FGF18 promotes liver fibrosis, we isolated primary HSCs by the Nycodenz density gradient centrifugation method⁴³. Based on the determination of UV-positive or desmin-positive cells, the percentages of HSCs were approximately 80 to 90% (Supplementary Fig. 13a–d). We stimulated HSCs with FGF18 or TGFβ1 as a positive control to induce the upregulation of profibrotic genes⁴⁴. FGF18 did not upregulate the expression of Col1a2, Col3a1, or Acta2, but TGFβ1 did (Fig. 7a). Likewise, based on the anti-α-SMA staining, TGFβ1 induced strong morphological changes, such as loss of a star-shaped appearance and elongated and spindle shaped morphology. In sharp contrast, FGF18-stimulated HSCs retained a star-like appearance, which was observed in untreated HSCs (Supplementary Fig. 14). Surprisingly, FGF18 stimulated the proliferation of HSCs and upregulated the expression of Ccnd1, encoding cyclin D1, which is involved in cell cycle progression⁴⁵ (Fig. 7b, c). Moreover, MEK and Akt inhibitors completely and moderately blocked FGF18-induced Ccnd1 expression (Fig. 7d), respectively, suggesting that the MEK/ERK pathway mainly contributes to FGF18-induced Ccnd1 expression. Since Fgfr2 was selectively expressed in HSCs (Fig. 6c) and its expression was higher than that of Fgfr3 or Fgfr4 (Fig. 7e), FGF18 may transmit signals through FGFR2.

**Fig. 7: FGF18 stimulates the proliferation of HSCs but inhibits TGFβ-induced upregulation of profibrotic genes.**

A previous study reported that bile acids increase the expression of another FGF protein family member, FGF15/19⁴⁶. The serum concentrations of various bile acids were elevated in the sera of Cflar^LKO mice fed the CDE diet (Supplementary Fig. 15a). Nevertheless, bile acids did not upregulate Fgf18 expression in primary hepatocytes (Supplementary Fig. 15b). In contrast, we found that TGFβ1 induced the expression of Fgf18 in primary hepatocytes and HSCs (Fig. 7f, g), whereas FGF18 suppressed TGFβ1-induced profibrotic gene expression as well as the expression of Tgfb2 and Tgfb3 in HSCs (Fig. 7h, i). Thus, the expression levels of Fgf18 and Tgfbs and their target genes were intimately regulated by each other. As expected, FGF18 did not upregulate the expression of profibrotic genes in hepatocytes, whereas TGFβ did activate them, although the induction levels were very low (Fig. 7j, k). Given that the basal expression levels of profibrotic genes in HSCs were higher than those in hepatocytes (Fig. 7a vs. 7j, 7k), the expansion of HSCs by FGF18, even at the relatively quiescent stages, may contribute to liver fibrosis.

Expression of FGF18 is correlated with the expression of COL1A1 and ACTA2 in human liver biopsy samples

To determine whether our data obtained from murine experiments are relevant to human liver fibrosis, we examined the expression of FGF18 in liver biopsy samples to diagnose various human liver diseases (Supplementary Table 3). Although the expression levels of FGF18 in the livers varied among diseases, the expression of FGF18 was correlated with that of COL1A1 and ACTA2 (Fig. 8a). These results suggest that FGF18 contributes to the development of liver fibrosis in humans as well as mice.

We propose the following model of how FGF18 promotes liver fibrosis. In response to various injuries, TGFβ is released from different types of cells, such as macrophages engulfing apoptotic hepatocytes. Then, TGFβ induces the production of FGF18 in HSCs and hepatocytes. FGF18 then stimulates the proliferation of HSCs. Proliferating HSCs further respond to stimuli derived from immune cells, such as scar-associated macrophages, and produce collagens and extracellular matrix, culminating in the development of liver fibrosis (Fig. 8b).

Discussion

In the present study, we found that the expression of Fgf18 was elevated in various murine liver fibrosis models. Deletion of Fgf18 in hepatocytes attenuated CDE-induced fibrosis in Cflar^LKO mice; conversely, overexpression of Fgf18 promoted liver fibrosis by increasing the number of HSCs. Mechanistically, FGF18 stimulated the proliferation of HSCs. Furthermore, the expression of FGF18 was correlated with the expression of COL1A1 and ACTA2 in human liver biopsy samples. Thus, FGF18 is a critical growth factor that promotes liver fibrosis and may be a therapeutic target for treating liver fibrosis.

Few apoptotic hepatocytes were detected in the livers of Cflar^LKO mice under homeostatic conditions. Cflar-deficient hepatocytes are susceptible to death ligand-induced apoptosis^28,30, suggesting that death ligands expressed under homeostatic conditions trigger hepatocyte apoptosis. Notably, Cflar^LKO mice developed mild fibrosis, with an increase in CK19⁺ cells in the liver. Although apoptosis promotes liver fibrosis under various conditions³¹, slight and persistent apoptosis of hepatocytes may be sufficient to induce mild fibrosis. Since macrophages release TGFβ upon engulfing apoptotic cells⁴⁷, the TGFβ released from macrophages stimulates the upregulation of Fgf18 expression in HSCs. Indeed, the expression of Fgf18 was higher in isolated NPCs (presumed HSCs) from Cflar^LKO mice, even when fed the normal diet, compared to those from Cflar^FF mice. This suggests that the slightly increased levels of Fgf18 in NPCs (presumed HSCs) may contribute to, at least in part, fibrosis in Cflar^LKO mice under normal dietary conditions. Moreover, preexisting mild fibrosis and increased susceptibility to death ligand-induced apoptosis of hepatocytes may drastically exacerbate the development of CDE-induced liver fibrosis in Cflar^LKO mice.

Although previous studies have demonstrated an intimate relationship between FGF18 and tumors^48,49,50, the link between FGF18 and fibrosis remains unknown. Given that HSCs play a central role in liver fibrosis², we can surmise that FGF18 activates HSCs, thereby promoting liver fibrosis. Although hematopoietic progenitor cells and CD31⁺ endothelial cells express CD34 under normal conditions, HSCs express CD34 in CCl₄-induced liver injury⁵¹. The scRNA-seq analysis revealed that CD31^- CD34⁺ cells are composed of Lrat⁺ HSCs, Acta2⁺ myofibroblasts, and Thy1⁺ fibroblasts, suggesting that CD34 is a common surface marker of HSCs and fibroblasts, at least in the liver. While Lrat⁺ HSCs mostly expressed Fgfr2, Acta2⁺ and Thy1⁺ fibroblasts expressed Fgfr1, some minor populations of Lrat⁺ HSCs expressed Fgfr1. Thus, FGF18 likely induces signals in HSCs through FGFR2 in vitro and in vivo. Regarding the contribution of FGF18 in hepatocytes, deletion of Fgf18 in hepatocytes moderately attenuated CDE-induced liver fibrosis. However, the effect was not drastic, suggesting that FGF18 produced by HSCs may compensate for the lack of FGF18 in hepatocytes. Hence, it would be interesting to test whether liver fibrosis is attenuated in mice lacking Fgf18 in HSCs by crossing Fgf18^flox/flox mice with Lrat-Cre Tg mice.

The scRNA-seq analysis revealed that the numbers of Lrat⁺ HSCs were increased in Fgf18 Tg mice and that Lrat⁺ HSCs mostly expressed quiescent HSC marker genes. Lrat⁺ HSCs expressed several activated HSC markers, including Col1a2, Col3a1, Igf1, and Lgals1, although their expression levels were relatively low compared to those of Acta2⁺ or Thy1⁺ fibroblasts. Moreover, Lrat⁺ HSCs did not express Timp1 and Tnc, which were elevated in HSCs from CCl₄-treated livers^10,42. Thus, FGF18 mainly induces the expansion of weakly activated or quiescent Lrat⁺ HSCs and does not fully promote terminal differentiation toward Acta2⁺ myofibroblasts.

In addition to fibrosis, the transgenic expression of Fgf18 in hepatocytes resulted in liver hypertrophy, consistent with a previous study⁵². The expression of Fgf18 is elevated in various human cancers, and its expression correlates with poor prognosis^49,50. FGF18 released from tumor cells induces their proliferation in an autocrine or paracrine manner⁴⁸. Under the experimental conditions of the present study, FGF18 did not activate primary hepatocytes. The expression of Hgf, which is a critical cytokine responsible for hepatocyte proliferation, was elevated in the livers of both Cflar^LKO mice fed the CDE diet and Fgf18 Tg mice, but FGF18 did not induce the upregulation of Hgf in HSCs. Given that the expression levels of Hgf were relatively high in HSCs under homeostatic conditions, FGF18 induced the proliferation of hepatocytes, not through the upregulation of Hgf but by stimulating the proliferation of HSCs. Moreover, FGF18 is involved in the proliferation and tube formation of endothelial cells⁴⁸. Consistent with previous studies, aberrant angiogenesis occurred in the livers of Fgf18 Tg mice. This finding was partly because Vegfa from cholangiocytes and HSCs may stimulate LSECs.

Notably, TGFβ upregulates the expression of profibrotic genes in HSCs but does not trigger the proliferation of HSCs. Activation of HSCs, rather than their strong expansion, seems sufficient to induce mild liver fibrosis, as observed in wild-type mice fed a CDE diet for 4 weeks, or mice fed the normal diet or the Western diet treated with CCl₄ for 12 weeks. In contrast, as livers progress to advanced fibrosis, it is reasonable to speculate that the expansion of HSCs may become necessary. When the strength of the stimuli exceeds a certain threshold, the expression of Fgf18 is strongly induced. This situation was observed in the livers of Cflar^LKO mice fed the CDE diet for 4 weeks, mice on the normal diet or the Western diet treated with CCl₄ for 24 weeks. Assuming that expansion of HSCs requires some space, it is reasonable to speculate that FGF18 suppresses the production of extracellular matrix to make a space for HSC proliferation. Moreover, it is worth noting that the expression levels of profibrotic genes in HSCs were relatively higher than those in hepatocytes, even if its expression levels were declined in the presence of FGF18. This finding indicates that the increased HSCs by FGF18 stimulation play an important role in the accumulation of collagens and extracellular matrix. Moreover, previous studies reported that scar-associated macrophages differentiate from circulating monocytes, are recruited to the fibrotic niche, and interact with PDGFRα⁺ collagen-producing mesenchymal cells, resulting in the activation of several profibrogenic pathways^53,54. Indeed, FGF18 induced the expression of inflammatory cytokines and chemokines, thereby recruiting immune cells, including scar-associated macrophages. Thus, proliferating HSCs induced by FGF18 further respond to profibrotic stimuli derived from scar-associated macrophages, such as platelet-derived growth factor (PDGF) and IL-1β⁵³, ultimately contributing to the development of liver fibrosis.

When focusing on a role of FGF18 in suppressing the expression of profibrotic genes, it appears that FGF18 may mitigate liver fibrosis under specific conditions, such as the transient expression of Fgf18 as noted by Tong et al.⁵⁵. However, in the long term, FGF18 seems to stimulate HSC proliferation, and subsequently, the proliferating HSCs become responsive to subsequent profibrotic stimuli as described above, ultimately leading to liver fibrosis. From a therapeutic perspective, it would be intriguing to investigate the effect of shRNA-mediated deletion of Fgf18 in adult mice with liver fibrosis. Further study will be required to address this issue.

We and others reported that FGF18 suppressed the expression of Tgfb2 and Tgfb3 and TGFβ-induced expression of profibrotic genes⁵⁵. In contrast, TGFβ induced upregulation of Fgf18. A previous study showed that TGFβ inhibits the phosphorylation of YAP, thereby preventing degradation of YAP and subsequently leading to fibrosis⁵⁵. Notably, FGF18 counteracts TGFβ-mediated YAP dephosphorylation, facilitating the degradation of phosphorylated YAP and preventing liver fibrosis⁵⁵. However, the specific mechanisms underlying the FGF18-induced phosphorylation of YAP were not elucidated in that study. In contrast, we reported that FGF18 induced the upregulation of Ccnd1, which was completely abolished in the presence of a MEK inhibitor. This finding suggests that the MEK/ERK pathway is crucial in this process. It remains unclear whether the MEK/ERK pathway and the YAP/TAZ pathway finally converge on the induction of proliferation-related or profibrotic genes. RNA-seq analysis using HSCs treated with TGFβ, FGF18, and a combination of the two could serve as a valuable tool to unravel the crosstalk between TGFβ and FGF18.

Methods

All experiments were performed according to the guidelines approved by the Institutional Animal experiments Committee of Faculty of Medicine, Toho University (21-409 and 21-412) and Faculty of Medicine and Graduate School of Medicine, Juntendo University (250071). Human study was approved by the ethics committee of Faculty of Medicine and Graduate School of Medicine, Juntendo University (E22-0085-H01).

Reagents

Human FGF18 (100-28, PeproTech), murine HGF (BL771601, BioLegend), insulin-transferrin-selenium (41400045, Gibco), cholic acid (CA) (C0324, Tokyo Chemical Industry), deoxycholic acid (DCA) (10712-12, Nacalai Tesque), glycocholic acid (GCA) (G2878, Sigma), taurocholic acid (TCA) (T4009, Sigma), human TGFβ1 (100-21, Peprotech), U0126 (CAS 109511-58-2, Calbiochem), and LY294002 (440206, Calbiochem) were purchased from the indicated sources. The following antibodies were used in this study and obtained from the indicated sources: anti-AKT (4691, Cell Signaling, 1:1000), anti-phospho-AKT (4060, Cell Signaling, 1:1000), anti-α-SMA (Ab124964, Abcam, 1:500), PE anti-mouse B220 (50-0452-U100, Tonbo Biosciences, 1:200), anti-APC micro-Beads (130-090-855, Miltenyi, 1:5), anti-PE micro-Beads UltraPure (130-105-639, Miltenyi, 1:5), anti-β-actin (sc-47778, Santa Cruz, 1:5000), anti-cleaved caspase 3 (9664, Cell Signaling, 1:1000), FITC anti-mouse CD3ε (35-0031-U100, Tonbo Biosciences, 1:200), PE anti-mouse CD11b (50-0112-U100, Tonbo Biosciences, 1:500), anti-mouse CD16/CD32 (Bio X Cell, BE0307, 1:200), APC anti-mouse CD31 (17-0311-82, Invitrogen, 1:100), FITC anti-mouse CD31 (102405, BioLegend, 1:200), PE anti-mouse CD31 (102407, BioLegend, 1:100), anti-CD34 (Ab81289, Abcam, 1:1000), APC anti-mouse CD34 (128611, BioLegend, 1:200), APC anti-mouse CD45.2 (558702, BD, 1:100), BV421 anti-mouse CD45.2 (109832, BioLegend, 1:200), PE anti-mouse CD45.2 (109808, BioLegend, 1:100), anti-CD68 (97778, Cell Signaling, 1:200), APC anti-mouse CD146 (134712, BioLegend, 1:100), anti-CK19 (in house, 1:200; MABT913, Merck, 1:1000), anti-desmin (AB32362, Abcam, 1:1000), APC anti-mouse Epcam (563478, BD biosciences, 1:100), anti-ERK (4695, Cell Signaling, 1:1000), anti-phospho-ERK (4370, Cell Signaling, 1:1000), FITC anti-mouse F4/80 (35-4801-U100, Tonbo Biosciences, 1:500), anti-FGF18 (120525, Wuhan Huamei Biotech, 1:1000; 19S-SE5 and 12G7-9 were made in-house), anti-cFLIP (Dave-2, Adipogen, 1:1000), anti-Ki67 (Ab16667, Abcam, 1:200), anti-Ly-6G (87048, Cell Signaling, 1:200), APC anti-mouse Ly-6G (127614, BioLegend, 1:500), anti-p75NTR (Ngfr/Tnfrsf16) (AF1157, R&D, 1:2000), PE anti-mouse PDGFRα (562776, BD Biosciences, 1:200), PE anti-mouse podoplanin (127407, BioLegend, 1:200), PE anti-mouse Sca1 (561076, BD Biosciences, 1:200), APC anti-mouse Ter119 (20-5921-U100, Tonbo Biosciences, 1:100), PE anti-mouse Thy1.2 (105307, BioLegend, 1:200), Alexa 594-conjugated anti-vimentin (677804, BioLegend, 1:400), eBioscience™ Fixable Viability Dye eFluor™ 506 (65-0866-14, eBioscience, 1:5000), and 7-AAD (420404, BioLegend, 1:100). Anti-MHC class I sampletag 1 (1:50), anti-MHC class I sampletag 2 (1:50), anti-MHC class I sampletag 3 (1:50), anti-MHC class I sampletag 4 (1:50), anti-MHC class I sampletag 5 (1:50), and anti-MHC class I sampletag 6 (1:50) were purchased from BD Biosciences. Alexa 647-conjugated donkey anti-rabbit IgG (A31573, 1:500), Alexa 488-conjugated donkey anti-rabbit IgG (A21206, 1:500), and Alexa 488-conjugated donkey anti-goat IgG (A11055, 1:500) antibodies were purchased from Invitrogen. HRP-conjugated donkey anti-rabbit IgG (NA934, 1:5000) and HRP-conjugated sheep anti-mouse IgG (NA931, 1:5000) antibodies were purchased from GE Healthcare. Peroxidase AffiniPure donkey anti-rat IgG (712-035-153, 1:5000) antibody was purchased from Jackson ImmunoResearch. Biotinylated goat anti-rabbit IgG (E0432, 1:200) and streptavidin-HRP (P0397, 1:300) were purchased from DAKO. Streptavidin poly-HRP80 conjugate (65R-S119) was purchased from Stereospecific Detection Technologies GmBH.

Mice

Cflar^FF mice were provided by Y-W. He and have been described previously⁵⁶. Exon 1 of the Cflar gene was flanked by two loxP sites and introduced into ES cells, resulting in the generation of Cflar^FF mice. Fgf18^FF mice (RBRC05691) were obtained from RIKEN Bioresource and were described previously³⁶. Exon 3 of the Fgf18 gene was flanked by two loxP sites and introduced into ES cells, resulting in the generation of Fgf18^FF mice. Albumin-Cre recombinase transgenic (Alb-Cre) mice (003574) and C57BL/6 mice were purchased from Jackson Laboratory and CLEA-Japan, respectively. To generate hepatocyte-specific Cflar-deficient mice, we crossed Cflar^FF mice with Alb-Cre mice, generating Cflar^LKO mice as described previously²⁹.

To generate hepatocyte-specific Fgf18-deficient mice, we crossed Fgf18^FF mice with Alb-Cre mice, generating Fgf18^LKO mice. To generate hepatocyte-specific Cflar and Fgf18 double-deficient mice, we crossed Cflar^FF;Alb-Cre mice with Fgf18^FF mice, resulting in the generation of Cflar^FF;Fgf18^FF;Alb-Cre (Cflar^LKO;Fgf18^LKO) mice.

Mice were housed in 23 ± 2°C, a humidity of 55% ± 5%, and a 12 h dark/light cycle. Only female mice were used for the CDE diet experiments, because our preliminary analysis showed that liver fibrosis were more severe in female mice than in male mice. Only male mice were used for the DDC diet experiments because male mice generally give more reproducible results due to the lack of estrous cycle. Feeding experiments began with 8-week-old mice, lasting either for a duration of 4 weeks (for the CDE and DDC diets) or for 12 weeks (for the CDAHFD diet). Mice fed the normal diet were analyzed at 12 weeks. To collect sera and tissue samples, mice were first anesthetized by isoflurane inhalation and then euthanized by cervical dislocation. To isolate liver cells, mice were anesthetized by intraperitoneal injection of a mixture of midazolam, medetomidine, and butorphanol prior to liver perfusion. Both male and female mice were used for other experiments. Non-Tg and Fgf18Tg mice were fed the normal diet and analyzed at young (6-8 weeks) and adult (18-24 weeks) ages.

All experiments were performed according to the guidelines approved by the Institutional Animal Experiments Committee of Faculty of Medicine, Toho University School of Medicine (21-409 and 21-412) and Faculty of Medicine and Graduate School of Medicine, Juntendo University (250071).

Generation of Rosa 26-LSL-Fgf18 Tg mice

To generate mice expressing Fgf18 in a tissue-specific manner, Fgf18 cDNA was inserted into the CTV vector (CAG promoter-loxP-STOP-loxP-Fgf18 cDNA-polyA) (Addgene #15912)⁵⁷. To induce homologous recombination, we used the CRISPR/Cas9 system; two sets of sgRNA oligos (5’- caccgTGGGCGGGAGTCTTCTGGGC-3’ and 5’- aaacGCCCAGAAGACTCCCGCCCAc-3’ for pX335-Rosa26-3, and 5’- caccGACTGGAGTTGCAGATCACG -3’ and 5’- aaacCGTGATCTGCAACTCCAGTC -3’ for pX335-Rosa26-4) were cloned into BbsI-digested pX335-U6-Chimeric_BB-CBh-hSpCas9n (D10A) plasmid (Addgene #42335). These vectors were electroporated into a feeder-free KTPU8 ES cell line derived from the TT2 ES cell line using a previously described method⁵⁸. Briefly, KTPU8 cells were electroporated using a Gene Pulser set (Bio-Rad). G418 was introduced into the medium after 48 h, and cells were subsequently cultured for an additional 7 days. After G418 selection, several ES clones in which a single transgene was integrated into the Rosa26 locus were obtained. Chimeric mice were produced by aggregation of ES cells with eight-cell embryos of ICR mice. Chimeras were crossed with C57BL/6N mice, resulting in three lines of Fgf18 Tg mice. CAG promoter-driven Fgf18 expression was interrupted by the insertion of the neomycin gene flanked by the loxP sites in the reverse orientation. Thus, Cre recombinase-dependent deletion of the neomycin gene results in tissue-specific expression of Fgf18. To express Fgf18 in hepatocytes, we crossed Rosa 26-LSL-Fgf18 Tg mice with albumin-Cre Tg mice, resulting in the generation of mice with hepatocyte-specific expression of Fgf18 in mice, which were referred to as Fgf18 Tg mice.

Induction of hepatitis

For the CDE diet, eight-week-old female mice were housed individually and fed the choline-deficient diet (02960034, MP Biomedicals) and drinking water containing 0.15% DL-ethionine (E5139, Sigma) for 4 weeks. For the DDC diet, 8-week-old male mice were fed the diet containing 0.1% 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) (CLEA Japan) for 4 weeks. For CDAHFD, eight-week-old mice were fed the CDAHFD diet (A06071302, Research Diets) for 12 weeks.

Measurement of serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP) concentrations

Serum ALT, AST, and ALP concentrations were determined by Oriental Yeast Co. Ltd.

Histological, immunohistochemical, and immunofluorescence analyses

The livers were fixed in 10% formalin or 4% paraformaldehyde and embedded in paraffin blocks. Paraffin-embedded liver sections were used for H&E staining and immunohistochemical and immunofluorescence analyses. Paraffin-embedded liver sections were stained with H&E, anti-CC3, anti-Ki67, anti-CK19, anti-desmin, anti-Ly-6G, anti-CD68, and anti-CD34 antibodies, Sirius Red, or the TUNEL method. The TUNEL method was performed using an In Situ Cell Death Detection Kit (Sigma‒Aldrich) according to the manufacturer’s instructions. Pictures were obtained by an All-in-One microscope (BZ-X700, KEYENCE). The numbers of CC3⁺ and TUNEL⁺ cells were counted manually. The Sirius Red⁺, CD34⁺, desmin⁺, CD68⁺, Ki67⁺, and CK19⁺ areas were automatically calculated by hybrid cell count (BZ-X700, KEYENCE).

For immunofluorescence analysis, tissue sections were preincubated with the MaxBlock^TM Autofluorescence Reducing Kit (MaxVision Biosciences) according to the manufacturer’s instructions. Then, paraffin-embedded sections were treated with Instant Citrate Buffer Solution (RM-102C, LSI Medicine) to retrieve the antigen. Blocking was performed with Blocking One Histo (06349-64, Nacalai Tesque) and 5% donkey serum. Liver tissue sections were stained with the indicated antibodies, followed by visualization with Alexa-conjugated secondary antibodies. Confocal microscopy was performed on an LSM 880 (Zeiss). Images were processed and analyzed using ZEN software (Zeiss).

Nile red staining

Frozen liver sections were first fixed with 10% formaldehyde neutral buffer solution for 1 h at room temperature, followed by staining with 1 μg/mL Nile Red (144-08811, WAKO) in PBS for 15 min at room temperature. After three washes with PBS, samples were mounted with Vectashield mounting medium with DAPI (H-1200, VECTOR).

Sirius red staining

Sirius Red/Direct Red 80 (D4132, Tokyo Chemical Industry) was dissolved at 1 g/L in a saturated aqueous picric acid solution. Dewaxed and hydrated paraffin sections were stained in the solution for 1 h and then washed twice with 0.5% acetate in water.

Quantitative polymerase chain reaction assays

Total RNA was purified from mouse livers and sorted cells using Sepasol-RNA I SuperG (09379-55, Nacalai Tesque) and RNeasy Micro Kit (74004, QIAGEN), respectively. Then, cDNAs were synthesized with the Revertra Ace qPCR RT Kit (FSQ-101, Toyobo). Quantitative polymerase chain reaction (qPCR) analysis was performed with the 7500 Real-Time PCR detection system with the SYBR green method for the target genes and an endogenous control, murine Hprt, or human GAPDH, with 7500 SDS software (Applied Biosystems). The primers used in this study are included in Supplementary Table 4.

Measurement of the hydroxyproline content

The concentrations of the hydroxyproline were determined by a Hydroxyproline Assay Kit according to the manufacturer’s instructions (STA-675, Cell BIOLABS). Frozen liver tissues (50-100 mg) were acid-hydrolyzed with 10 volumes (w/v) of 6 N HCl at 95 °C for 24 h. After centrifugation, the supernatant was used for the colorimetric assay.

Hydrodynamic tail vein injection (HTVi)

To express Egfp for the negative control, murine Fgf18, Tnfsf13b, and Wnt5a in hepatocytes, we amplified these cDNAs by RT‒PCR and subcloned them into the pLIVE vector (MIR5420, Mirus Bio). The pLIVE vector consists of the albumin promoter and alpha-fetoprotein enhancer II, thereby expressing the gene specifically in hepatocytes at high levels for a relatively long period in vivo. HTVi was performed according to the standard procedure³². Briefly, 20 μg of pLIVE-Egfp, Fgf18, Tnfsf13b, or Wnt5a was diluted in 2 mL of TransIT-EE Hydrodynamic Delivery solution (MIR5340, Mirus Bio) and injected into 8-week-old mice through the tail vein. The expression of these genes in the liver was verified by qPCR, and liver sections were analyzed by immunohistochemistry with anti-CK19, anti-desmin, anti-GFP, or anti-Ki67 antibodies one week after injection.

RNAscope

Mouse livers were first fixed with buffered 10% formaldehyde neutral buffer at room temperature for 16–32 h and embedded in paraffin. Dewaxed and hydrated liver sections were processed using the RNAscope Multiplex Fluorescent Reagent Kit v2 (323100, Advanced Cell Diagnostics) and TSA Plus Cyanine 3 and Cyanine 5 kits (NEL744001KT and NEL745001KT, respectively, Akoya Biosciences) according to the manufacturer’s instructions. RNAscope target probes for mouse Fgf18 (495421), Lrat (460641-C2), and Hnf4a (497651-C3) were purchased from Advanced Cell Diagnostics. Fluorescence images were obtained by either BZ-X700 (Keyence) or LSM880 (Zeiss), and fluorescent areas were calculated by Hybrid Cell Count (Keyence) or Adobe Photoshop software. To identify cells expressing Fgf18 mRNA, merged images were used to determine whether Fgf18 mRNA puncta were located close to Lrat mRNA puncta (HSC origin) or Hnf4a mRNA puncta (hepatocyte origin). The number of Fgf18 mRNA puncta in HSCs and hepatocytes was manually counted and statistically analyzed.

Isolation of primary hepatocytes and nonparenchymal liver cells

To purify nonparenchymal liver cells, we performed a modified two-step collagenase perfusion method as described previously⁵⁹. Briefly, the livers from 6- to 8-week-old wild-type, non-Tg, and Fgf18 Tg mice were perfused with liver perfusion buffer (LPB) 1 (136 mM NaCl, 5.4 mM KCl, 0.5 mM EGTA, 0.5 mM NaH₂PO₃ 2H₂O, 0.42 mM Na₂HPO₃, 10 mM HEPES pH 7.5, 5 mM glucose and 4.2 mM NaHCO₃) at a flow rate of 3 ml/min for 5 min. Then, the livers were perfused with LPB2 (136 mM NaCl, 5.4 mM KCl, 5 mM CaCl₂, 0.5 mM NaH₂PO₃ 2H₂O, 0.42 mM Na₂HPO₃, 10 mM HEPES pH 7.5, 5 mM glucose and 4.2 mM NaHCO₃) containing 0.5 mg/ml collagenase type IV (C5138, Sigma‒Aldrich) and 0.06 mg/ml DNase I (DN25, Sigma-Aldrich) at a flow rate of 3 ml/min for 8 min. The digested livers were transferred to a glass dish, and fibrous connective tissues were removed from the livers with a pair of tweezers. For Fgf18 Tg mice, the perfusion of livers with LPB2 alone was insufficient to isolate nonparenchymal cells from fibrotic livers. Thus, we further incubated the livers with LPB2 after perfusion using a stirrer bar at 37 °C for 5 min. Cells were dispersed with tweezers and passed through a 70 μm cell strainer. The supernatant was transferred to a new tube after centrifugation at 50 × g for 2 min. The pellet was resuspended in DMEM and centrifuged at 50 × g for 2 min; then, the pellet was used as hepatocytes. The new supernatant was combined with the former supernatant and then centrifuged at 100 × g for 2 min several times until no cell pellet was visible. The final supernatant was centrifuged at 660 × g for 5 min, and the pellets were resuspended in RBC lysis solution (0.17 mM NH₄Cl, 0.01 mM EDTA, 0.1 M Tris, pH 7.3) and then washed twice in DMEM containing 10% FCS. Cells were then subjected to flow cytometry or cell sorting. For Western blotting, cells were lysed in lysis buffer as described below.

Hepatocytes were pelleted after centrifugation at 50 × g for 2 min, resuspended in DMEM containing 10% FCS, and plated on collagen I-coated 6-well microplates (4810-010, Iwaki Glass). The next day, the medium was changed to remove unattached dead hepatocytes. Cells were used for stimulation or Western blotting by directly adding lysis buffer as described below.

Flow cytometry and cell sorting

Nonparenchymal liver cells were prepared as described above. Single-cell suspensions were incubated with anti-mouse CD16/CD32 antibody and then stained with the indicated antibodies. Cells were analyzed on an LSRFortessa X-20 (BD Biosciences) or sorted by a FACSAria Fusion (BD Biosciences). Data were processed with CellQuest software or FlowJo (BD Biosciences).

Western blotting

Cells were lysed with RIPA buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 25 mM β-glycerophosphate, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 1 μg/mL aprotinin, 1 μg/mL leupeptin, and 1 μg/mL pepstatin) on ice for 20 min. After centrifugation, the cell lysates and cultured supernatants were subjected to SDS‒PAGE and then transferred onto polyvinylidene difluoride membranes (IPVH 00010, Millipore). The membranes were analyzed by immunoblotting with the indicated antibodies and developed with Super Signal West Dura Extended Duration Substrate (34076, Thermo Scientific). The signals were analyzed with Amersham Imager 600 (GE Healthcare Life Sciences).

To prepare a positive control of mouse FGF18, we generated recombinant mouse FGF18 in E. coli as GST-fusion protein as described previously³⁹. Briefly, E. coli BL21(DE3) was transformed with pGEX-6P-mFGF18. The transformed bacteria were incubated in 2 ml of LB medium containing 50 μg/ml ampicillin at 37°C overnight, followed by inoculation in 250 ml of LB medium, and the culture was continued at 37°C. When the bacteria reached 1.0 OD600, IPTG was added to a final concentration of 0.1 mM and incubated at 20 °C for 8 h to induce protein expression. Then, the bacteria were harvested and resuspended in NETN (20 mM Tris pH 8.0, 100 mM NaCl, 0.5% NP-40, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 μg/ml pepstatin), followed by sonication by a Polytron (KINEMATICA). After centrifugation, the supernatants were incubated with 125 μl of glutathione–Sepharose 4B beads (GE Healthcare, 17-0756-01) (1:1 slurry) for 1 h. Then, the beads were washed with NETN 4 times, followed by incubation with PreScission protease (GE Healthcare, 27-0843-01) to release mFGF18 at 4°C overnight. The concentrations of purified proteins were determined by Bradford analysis (Bio-Rad 5000006). The purity of these proteins was evaluated by staining with Coomassie blue.

Cell culture

Isolated hepatocytes were stimulated with FGF18 (100 ng/mL), HGF (10 ng/mL), or insulin (10 ng/mL) for the indicated times. The phosphorylation of the indicated proteins was analyzed by Western blotting. For bile acid stimulation of primary hepatocytes, the medium was first changed to serum-free DMEM. Three hours later, primary hepatocytes were stimulated with 100 μM CA, GCA, TCA, and DCA for 2 h or with 1 ng/mL TGFβ1 for 24 h. The expression of the indicated genes was determined by qPCR.

Isolation of HSCs from wild-type mice and in vitro stimulation

To isolate HSCs, we used wild-type female C57BL/6 mice over 24 weeks old with the Nycodenz density gradient centrifugation method⁴³. Vitamin A-containing granules are increased in HSCs with aging, enabling us to obtain sufficient numbers of HSCs for qPCR and cell proliferation. Briefly, following the perfusion of livers with 0.3 mg/mL collagenase for 8 min as described above, the livers were perfused with LPB2 containing 0.6 mg/mL pronase E (KA-002, KNF) and 0.06 mg/ml DNase I at a flow rate of 3 ml/min for 8 min. The perfused livers were transferred to a glass dish and dispersed with the help of tweezers in LPB2 containing 0.6 mg/ml of both pronase and collagenase. The cell suspension was transferred to a beaker and incubated at 37 °C for 30 min with careful stirring. Cells were passed through a 70 μm cell strainer. After centrifugation at 600 × g for 10 min, the pellet was resuspended in GBSS/B (136 mM NaCl, 5 mM KCl, 1 mM MgCl₂ 6H₂O, 0.28 mM MgSO₄ 7H₂O, 0.22 mM KH₂PO₄, 0.42 mM Na₂HPO₄ 12H₂O, 5.5 mM glucose, 2.7 mM NaHCO₃, 1.5 mM CaCl₂ 2H₂O) containing 0.06 mg/ml DNase I and centrifuged at 600 × g for 10 min two times. The pellets were resuspended in GBSS/B containing 8.8% Nycodenz (18003, SEW), and GBSS/B was gently added to the cell suspension. After centrifugation at 1500 × g for 22 min without braking, the HSCs were found as a condensed white band in the interphase of both solutions. The HSC fractions were collected with a Pasteur pipette and centrifuged at 600 × g for 10 min. We usually obtained 1.0 × 10⁶ cells from two mice.

To evaluate purity of HSCs isolated from the livers, cells were plated and cultured on poly-l-Lysine-coated 12 mm micro coverslips (C012001, MATSUNAMI) in 24-well plates for 24 h. After removing nonadherent cells, cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and stained with anti-desmin antibody, followed by visualization of Alexa 647-conjugated donkey anti-rabbit antibody. UV-positive or desmin-positive cells were counted manually and expressed as the percentages of total cells. Approximately 80-90% of cells were UV-positive or desmin-positive HSCs by a confocal microscopy (LSM880) (Supplementary Fig. 13).

To test whether FGF18 and TGFβ induce morphological changes of HSCs, HSCs were plated as described above. HSCs were starved with 0.2% FBS containing DMEM for 24 h, and then left untreated or simulated with FGF18 (100 ng/mL) or TGFβ1 (1 ng/mL) for 24 h. Cells were stained following the same procedure as described above, with the exception that anti-α-SMA antibody was used as the primary antibody. Then, cells were visualized with Alexa 647-conjugated donkey anti-rabbit antibody. Pictures were obtained by an All-in-One microscope (BZ-X700, KEYENCE).

HSCs were plated onto 48-well plates (1 × 10⁵ cells), starved with DMEM containing 0.2% FBS for 24 h and then stimulated with TGFβ1 (1 ng/mL) or FGF18 (100 ng/mL) in the absence or presence of U0126 (10 μM) or LY294002 (10 μM) for 24 h. The expression of the indicated genes was determined by qPCR.

For the proliferation assay, HSCs were plated onto 96-well plates (1 × 10⁴ cells) and then stimulated with TGFβ1 (1 ng/mL) or FGF18 (100 ng/mL) in DMEM containing 0.2% FBS for the indicated times. Cell viability was determined by the WST (water-soluble 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)[2H] tetrazolium monosodium salt-1) assay using a cell counting kit (343-07623, Dojindo).

Bulk RNA-sequencing of the whole livers

Transcriptome libraries were constructed using total RNA samples from mouse livers. Briefly, polyA RNA was isolated using Dynabeads M-270 Streptavidin (DB65305, Thermo Fisher Scientific) conjugated with biotin-labeled oligo(dT) primer and reverse transcribed using SuperScript II (18064022, Thermo Fisher Scientific). For amplification of the total cDNA, the beads containing cDNA were first subjected to PCR using KAPA Hifi HS Ready Mix according to the manufacturer’s instructions (KK2601, KAPA Biosystems). The first PCR products were purified using an Agencourt AMPure XP kit (A63880, Beckman-Coulter) and were used as templates for the second PCR using KAPA Hifi HS Ready Mix. The second PCR products were purified using the Agencourt AMPure XP kit, and the transcriptome library was subjected to fragmentation/end-repair/polyA-tailing/ligation using the NEBNext Ultra II FS DNA Library Prep Kit for Illumina (E7805, New England Biolabs) according to the manufacturer’s instructions. Reaction products were purified using a double-size selection of the Agencourt AMPure XP kit and amplified by PCR using NEBNext Ultra II Q5 (New England Biolabs) and unique dual indexing primers. Reaction products were purified twice using a double-size selection of the Agencourt AMPure XP kit. Final transcriptome libraries, whose lengths were approximately 300 bp, were pooled and sequenced on an Illumina NovaSeq 6000 S4 flow cell (Illumina). Sequences were mapped to the mouse reference genome GRCm39 using Bowtie2 software, and TCC-GUI (https://github.com/swsoyee/TCC-GUI) was used to normalize sequence count data and conduct differential gene expression analysis. DAVID v6.8 (https://david.ncifcrf.gov/summary.jsp) was used for the functional annotations of differentially expressed genes. Heatmaps were generated using R3.6.1 with Genefilter and Gplots libraries. RNA-seq data were deposited in NCBI as GEO accession number GSE188273.

Enzyme-linked immunosorbent assay (ELISA)

We developed an in-house ELISA for hFGF18 by generating monoclonal antibodies against hFGF18. The generation and characterization of the in-house hFGF18 ELISA system were described previously³⁹. Briefly, we used one rat anti-hFGF18 antibody (clone 12G7-9) and one rabbit anti-hFGF18 antibody (clone 19S-SE5) to develop the in-house ELISA. The plate was coated with clone 12G7-9 (30 µg/mL) in 0.5% BSA-PBS at 4 °C overnight and blocked with 1% BSA-PBS-0.05% Tween (PBS-T) for 1 h. After washing with PBS-T three times, samples were added to the plate and incubated at RT for 1 h. After washing with PBS-T three times, the plate was incubated with biotinylated 19S-SE5 antibody (1.0 μg/mL) at RT for 1 h. After washing with PBS-T three times, the plate was incubated with streptavidin poly-HRP80 conjugate (0.2 μg/mL) at RT for 1 h and developed with TMB substrates. The lower detection limits of hFGF18 and mFGF18 were approximately 10 pg/mL and 500 pg/mL, respectively.

Single-cell RNA sequencing

We isolated nonparenchymal cells from the livers of 6-week-old non-Tg and Fgf18 Tg mice. To enrich HSCs and fibroblasts, we depleted lineage marker-positive cells using APC-conjugated antibodies that react with the following lineage markers (CD31, CD45.2, CD146, EpCAM, and Ter119), followed by anti-APC microbeads and LS columns (130-042-401, Miltenyi). Following depletion, approximately 5 × 10⁴ cells were stained with different DNA barcode-conjugated MHC class 1 antibodies (BD^TM Mouse Immune Single-Cell Multiplexing Kit [MHC H2 Class I], 626545, BD Biosciences) to identify each mouse. Then, lineage-negative cells were sorted by a BD FACSAria^TM III system (BD Biosciences). A mixture of 2.4 × 10⁴ cells from 3 non-Tg and 3 Fgf18 Tg mice (approximately 4 × 10³ cells from each mouse) was subjected to scRNA-seq using the BD Rhapsody^TM Single-Cell Analysis System (BD Biosciences), and the resulting cDNA was amplified by the TAS-Seq protocol as previously described⁶⁰. Briefly, on-bead cDNA was polyC tailed with stochastic termination conditions by terminal transferase, deoxycytidine, and spiked-in dideoxycytidine. Then, second-strand synthesis and whole-transcriptome amplification were performed by PCR. The size distribution of cDNA and BD Sampletag libraries was analyzed by a MultiNA system (MCE-202, Shimazu). The resulting cDNA library was processed to generate the sequencing library using the NEBNext UltraII FS library prep kit for Illumina (New England Biolabs), and sequencing adapters were added to associated BD Sampletag libraries by PCR. Sequencing was performed by an Illumina NovaSeq 6000 sequencer (Illumina, San Diego, CA, USA) and NovaSeq 6000 S4 Reagent Kit v1.5 (200 cycles). The pooled library concentration was adjusted to 2.0 nM, and 12% PhiX control library v3 (Illumina) was spiked into the library. The sequencing configurations were as follows: read1 67 base-pair [bp], read2 151 bp, index1 8 bp, and index2 8 bp. Adapter trimming, quality filtering, and mapping to the cell barcode and GRCm38-101 reference transcriptome or BD Sampletag reference of fastq files were performed using a pipeline (https://github.com/s-shichino1989/TASSeq) as described previously⁶⁰. Briefly, after the trimming of adapter sequences, cell barcode reads were annotated by Python script provided by BD Biosciences with minor modification. cDNA reads were mapped to reference RNA sequences (build GRCm38 release-101) using bowtie2-2.4.2. Demultiplexing of scRNA-seq data by BD sample tag was performed using a pipeline (https://github.com/s-shichino1989/TASSeq) as described previously⁶⁰. Briefly, cell barcode information of each read was added to the bowtie2-mapped BAM files by the python script and pysam 0.15.4 (https://github.com/pysam-developers/pysam), and read counts of each gene in each cell barcode were counted using mawk. The resulting gene expression count matrix was processed for downstream single-cell analyses (integration of six datasets, UMAP dimension reduction, cell cluster identification, and conserved marker identification) using Seurat version 4.1.1 in R version 4.1.3. Briefly, six scRNA-seq datasets were integrated with the merge functions of Seurat. Cells that contained more than 25% mitochondrial transcripts were filtered out. PCA was performed against 6885 highly-variable genes identified by the FindVariableFeatures (selection.method = mvp, mean.cutoff = c(0.1, Inf), dispersion.cutoff = c(0.5, Inf)) function in Seurat. A total of 1:56 PCs were selected by Jackstraw analysis and used for clustering analysis. We set the resolution to 1.5. Moreover, a cell expressing two different lineage marker genes was considered a doublet cell and removed from the datasets. Next, marker genes in each cell cluster were defined by the FindAllMarkers function in Seurat (test method = Wilcox, minimum expression in each cluster ≥ 10%). In the subclustering analysis of clusters 1, 3, 8, 12, and 16, we set the resolution to 0.4.

We used CellChat to infer intercellular communication from scRNA-seq data from non-Tg and Fgf18 Tg mice. We then downsampled cells from Fgf18 Tg mice to those from non-Tg mice by the SubsetData function of Seurat v4.1.1. We set the resolution to 1.4. The communication probability between cell subsets was analyzed by CellChat (https://github.com/sqjin/CellChat).

We extracted the datasets of LSEC and cholangiocyte clusters from normal mouse liver scRNA-seq data¹¹ by the SubsetData function of Seurat. Then we integrated these datasets with those of non-Tg and Fgf18 Tg mice. A total of 1:65 PCs were selected by Jackstraw analysis and used for clustering analysis. We set the resolution to 1.0.

Bile acid analysis

Twenty microliters of the sera from mice of the indicated genotype fed formal diet or CDE diet for 4 weeks were used to analyze bile acids. One hundred eighty microliters of acetonitrile were added to the sera with 2 µl of 1 mM 2,4-dichlorophenoxyacetic acids as an internal standard. After vortexing for 3 min, the mixtures were centrifuged at 20,400 × g for 3 min. The supernatants were loaded on an ultrafiltration membrane (Nanosep 3 K omega, PALL Corp.) and centrifuged (14,000 × g, 30 min). The filtrate was dried and resuspended in 20 µl of MeOH. The LC‒MS conditions were as follows: HPLC, Agilent 1260 infinity series (Agilent Technologies, Inc.), column; Ascentis Express C18 (4.6 ×150 mm, 2.7 µm) (Supelco, Inc.); solvent, (A) 0.1% formic acid aq., (B) 0.1% formic acid in acetonitrile; gradient; 30–75% (B) for 0–30 min, 95% (B) for 30–35 min; flow rate, 0.6 ml/min; column temperature, 40 °C; injection, 10 µl; MS, Agilent 6120 Quadrupole (Agilent Technologies, Inc.); mode, ESI negative, SIM; capillary voltage, 4 kV; dry gas rate, 13.0 l/min (N2); gas temperature, 350 °C; mass range, 150–600 m/z.

Human liver biopsy samples

Human liver biopsy samples were obtained from patients admitted to Juntendo University Hospital to diagnose liver diseases. Half of the biopsy samples were used for histological analysis to determine the diagnosis. RNA was extracted from the remaining liver biopsy samples using an RNeasy Micro Kit (74004, QIAGEN) and analyzed by qPCR. The patients’ diagnosis, age, sex, and concentrations of liver enzymes are listed in Supplementary Table 3. Written informed consent was obtained from the patients before liver biopsy. This study was approved by the ethics committee of Faculty of Medicine and Graduate School of Medicine, Juntendo University (approval number E22-0085-H01).

Statistical analysis

Statistical analysis was performed by two-tailed unpaired Student’s t test, two-way ANOVA with Sidak’s multiple comparison test, two-way ANOVA with Tukey’s multiple comparison test, one-way ANOVA with Dunnett’s multiple comparison test, one-way ANOVA with Tukey’s multiple comparison test, two-sided nonparametric Wilcoxon rank sum test, or Pearson correlation coefficient using GraphPad Prism 9. GO enrichment analyses were performed using DAVID 6.8, and p values were calculated based on one-sided Fisher’s exact test. P-value < 0.05 was considered significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All biological materials, including the Rosa26-LSL-Fgf18 Tg mice used in this study, are available from the corresponding authors upon request. Obtaining Cflar^FF and Fgf18^FF mice requires a material transfer agreement (MTA) with the organization described in the manuscript. The bulk RNA-seq and scRNA-seq datasets were deposited to NCBI under accession numbers GSE188273 and GSE205871, respectively. The authors declare that the data supporting the findings of this study are available within the paper and its supplementary data and supplementary information files. Source data are provided with this paper.

References

Kisseleva, T. & Brenner, D. Molecular and cellular mechanisms of liver fibrosis and its regression. Nat. Rev. Gastroenterol. Hepatol. 18, 151–166 (2021).
PubMed Google Scholar
Schwabe, R. F., Tabas, I. & Pajvani, U. B. Mechanisms of fibrosis development in nonalcoholic steatohepatitis. Gastroenterology 158, 1913–1928 (2020).
CAS PubMed Google Scholar
Angulo, P. et al. Liver fibrosis, but no other histologic features, is associated with long-term outcomes of patients with nonalcoholic fatty liver disease. Gastroenterology 149, 389–397.e310 (2015).
PubMed Google Scholar
Vilar-Gomez, E. et al. Fibrosis severity as a determinant of cause-specific mortality in patients with advanced nonalcoholic fatty liver disease: a multi-national cohort study. Gastroenterology 155, 443–457.e417 (2018).
PubMed Google Scholar
Rinella, M. E. et al. A multi-society Delphi consensus statement on new fatty liver disease nomenclature. J. Hepatol. S0168–8278, 00418–X. https://doi.org/10.1016/j.jhep.2023.06.003 (2023).
Google Scholar
Younossi, Z. et al. Global burden of NAFLD and NASH: trends, predictions, risk factors and prevention. Nat. Rev. Gastroenterol. Hepatol. 15, 11–20 (2018).
PubMed Google Scholar
Loomba, R., Friedman, S. L. & Shulman, G. I. Mechanisms and disease consequences of nonalcoholic fatty liver disease. Cell 184, 2537–2564 (2021).
CAS PubMed Google Scholar
Parthasarathy, G., Revelo, X. & Malhi, H. Pathogenesis of nonalcoholic steatohepatitis: an overview. Hepatol. Commun. 4, 478–492 (2020).
PubMed PubMed Central Google Scholar
Schwabe, R. F. & Luedde, T. Apoptosis and necroptosis in the liver: a matter of life and death. Nat. Rev. Gastroenterol. Hepatol. 15, 738–752 (2018).
PubMed PubMed Central Google Scholar
Rosenthal, S. B. et al. Heterogeneity of HSCs in a Mouse Model of NASH. Hepatology 74, 667–685 (2021).
CAS PubMed Google Scholar
Yang, W. et al. Single-cell transcriptomic analysis reveals a hepatic stellate cell-activation roadmap and myofibroblast origin during liver fibrosis in mice. Hepatology 74, 2774–2790 (2021).
CAS PubMed Google Scholar
Xiong, X. et al. Landscape of intercellular crosstalk in healthy and NASH liver revealed by single-cell secretome gene analysis. Mol. Cell 75, 644–660 e645 (2019).
CAS PubMed PubMed Central Google Scholar
Karin, D., Koyama, Y., Brenner, D. & Kisseleva, T. The characteristics of activated portal fibroblasts/myofibroblasts in liver fibrosis. Differentiation 92, 84–92 (2016).
CAS PubMed PubMed Central Google Scholar
Wang, S. et al. An autocrine signaling circuit in hepatic stellate cells underlies advanced fibrosis in nonalcoholic steatohepatitis. Sci. Transl. Med. 15, eadd3949 (2023).
CAS PubMed Google Scholar
Du, K. et al. Hedgehog-YAP signaling pathway regulates glutaminolysis to control activation of hepatic stellate cells. Gastroenterology 154, 1465–1479.e1413 (2018).
CAS PubMed Google Scholar
Xie, Y. et al. FGF/FGFR signaling in health and disease. Signal Transduct. Target Ther. 5, 181 (2020).
CAS PubMed PubMed Central Google Scholar
Ornitz, D. M. & Itoh, N. The fibroblast growth factor signaling pathway. Wiley Interdiscip. Rev.: Dev. Biol. 4, 215–266 (2015).
CAS PubMed Google Scholar
Seitz, T. & Hellerbrand, C. Role of fibroblast growth factor signalling in hepatic fibrosis. Liver Int. 41, 1201–1215 (2021).
PubMed Google Scholar
Peterova, E., Podmolikova, L., Rezacova, M. & Mrkvicova, A. Fibroblast growth factor-1 suppresses TGF-beta-mediated myofibroblastic differentiation of rat hepatic stellate cells. Acta Med. (Hradec Kralove) 59, 124–132 (2016).
CAS Google Scholar
Sato-Matsubara, M. et al. Fibroblast growth factor 2 (FGF2) regulates cytoglobin expression and activation of human hepatic stellate cells via JNK signaling. J. Biol. Chem. 292, 18961–18972 (2017).
PubMed PubMed Central Google Scholar
Yu, C. et al. Role of fibroblast growth factor type 1 and 2 in carbon tetrachloride-induced hepatic injury and fibrogenesis. Am. J. Pathol. 163, 1653–1662 (2003).
CAS PubMed PubMed Central Google Scholar
Schumacher, J. D. et al. Direct and indirect effects of fibroblast growth factor (FGF) 15 and FGF19 on liver fibrosis development. Hepatology 71, 670–685 (2020).
CAS PubMed Google Scholar
Xu, P. et al. Fibroblast growth factor 21 attenuates hepatic fibrogenesis through TGF-beta/smad2/3 and NF-kappaB signaling pathways. Toxicol. Appl. Pharm. 290, 43–53 (2016).
CAS Google Scholar
Budd, R. C., Yeh, W. C. & Tschopp, J. cFLIP regulation of lymphocyte activation and development. Nat. Rev. Immunol. 6, 196–204 (2006).
CAS PubMed Google Scholar
Yeh, W. C. et al. Requirement for Casper (c-FLIP) in regulation of death receptor-induced apoptosis and embryonic development. Immunity 12, 633–642 (2000).
CAS PubMed Google Scholar
Dillon, C. P. et al. Survival function of the FADD-CASPASE-8-cFLIP(L) complex. Cell Rep. 1, 401–407 (2012).
CAS PubMed PubMed Central Google Scholar
Tsuchiya, Y., Nakabayashi, O. & Nakano, H. FLIP the switch: regulation of apoptosis and necroptosis by cFLIP. Int J. Mol. Sci. 16, 30321–30341 (2015).
CAS PubMed PubMed Central Google Scholar
Schattenberg, J. M. et al. Ablation of c-FLIP in hepatocytes enhances death-receptor mediated apoptosis and toxic liver injury in vivo. J. Hepatol. 55, 1272–1280 (2011).
CAS PubMed Google Scholar
Piao, X. et al. c-FLIP maintains tissue homeostasis by preventing apoptosis and programmed necrosis. Sci. Signal 5, ra93 (2012).
PubMed PubMed Central Google Scholar
Piao, X. et al. Depletion of myeloid cells exacerbates hepatitis and induces an aberrant increase in histone H3 in mouse serum. Hepatology 65, 237–252 (2017).
CAS PubMed Google Scholar
Luedde, T., Kaplowitz, N. & Schwabe, R. F. Cell death and cell death responses in liver disease: mechanisms and clinical relevance. Gastroenterology 147, 765–783.e764 (2014).
CAS PubMed Google Scholar
Matsuda, M. et al. Oncostatin M causes liver fibrosis by regulating cooperation between hepatic stellate cells and macrophages in mice. Hepatology 67, 296–312 (2018).
CAS PubMed Google Scholar
Matsumoto, M. et al. An improved mouse model that rapidly develops fibrosis in non-alcoholic steatohepatitis. Int J. Exp. Pathol. 94, 93–103 (2013).
CAS PubMed PubMed Central Google Scholar
Preisegger, K. H. et al. Atypical ductular proliferation and its inhibition by transforming growth factor beta1 in the 3,5-diethoxycarbonyl-1,4-dihydrocollidine mouse model for chronic alcoholic liver disease. Lab Invest. 79, 103–109 (1999).
CAS PubMed Google Scholar
Tsuchida, T. et al. A simple diet- and chemical-induced murine NASH model with rapid progression of steatohepatitis, fibrosis and liver cancer. J. Hepatol. 69, 385–395 (2018).
PubMed PubMed Central Google Scholar
Kimura-Ueki, M. et al. Hair cycle resting phase is regulated by cyclic epithelial FGF18 signaling. J. Invest. Dermatol. 132, 1338–1345 (2012).
CAS PubMed Google Scholar
Ohbayashi, N. et al. FGF18 is required for normal cell proliferation and differentiation during osteogenesis and chondrogenesis. Genes Dev. 16, 870–879 (2002).
CAS PubMed PubMed Central Google Scholar
Soriano, P. Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat. Genet. 21, 70–71 (1999).
CAS PubMed Google Scholar
Tsuchiya, Y. et al. A high-sensitivity ELISA for detection of human FGF18 in culture supernatants from tumor cell lines. Biochem. Biophys. Res Commun. 675, 71–77 (2023).
CAS PubMed Google Scholar
Lin, G., Finger, E. & Gutierrez-Ramos, J. C. Expression of CD34 in endothelial cells, hematopoietic progenitors and nervous cells in fetal and adult mouse tissues. Eur. J. Immunol. 25, 1508–1516 (1995).
CAS PubMed Google Scholar
Koui, Y. et al. Development of human iPSC-derived quiescent hepatic stellate cell-like cells for drug discovery and in vitro disease modeling. Stem Cell Rep. 16, 3050–3063 (2021).
CAS Google Scholar
Jin, S. et al. Inference and analysis of cell-cell communication using CellChat. Nat. Commun. 12, 1088 (2021).
ADS CAS PubMed PubMed Central Google Scholar
Weiskirchen, S., Tag, C. G., Sauer-Lehnen, S. & Tacke, F. Weiskirchen R. isolation and culture of primary murine hepatic stellate cells. Methods Mol. Biol. 1627, 165–191 (2017).
CAS PubMed Google Scholar
Dewidar, B., Meyer, C., Dooley, S. & Meindl-Beinker, A. N. TGF-beta in hepatic stellate cell activation and liver fibrogenesis-updated 2019. Cells 8, 1419 (2019).
CAS PubMed PubMed Central Google Scholar
Mullany, L. K. et al. Distinct proliferative and transcriptional effects of the D-type cyclins in vivo. Cell Cycle 7, 2215–2224 (2008).
CAS PubMed Google Scholar
Holt, J. A. et al. Definition of a novel growth factor-dependent signal cascade for the suppression of bile acid biosynthesis. Genes Dev. 17, 1581–1591 (2003).
CAS PubMed PubMed Central Google Scholar
Fadok, V. A. et al. Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-beta, PGE2, and PAF. J. Clin. Invest. 101, 890–898 (1998).
CAS PubMed PubMed Central Google Scholar
Gauglhofer, C. et al. Up-regulation of the fibroblast growth factor 8 subfamily in human hepatocellular carcinoma for cell survival and neoangiogenesis. Hepatology 53, 854–864 (2011).
CAS PubMed Google Scholar
Chen, T. et al. Fibroblast growth factor 18 promotes proliferation and migration of H460 cells via the ERK and p38 signaling pathways. Oncol. Rep. 37, 1235–1242 (2017).
CAS PubMed Google Scholar
Zhang, J. et al. FGF18, a prominent player in FGF signaling, promotes gastric tumorigenesis through autocrine manner and is negatively regulated by miR-590-5p. Oncogene 38, 33–46 (2019).
CAS PubMed Google Scholar
Terkelsen, M. K. et al. Transcriptional dynamics of hepatic sinusoid-associated cells after liver injury. Hepatology 72, 2119–2133 (2020).
CAS PubMed Google Scholar
Hu, M. C. et al. FGF-18, a novel member of the fibroblast growth factor family, stimulates hepatic and intestinal proliferation. Mol. Cell Biol. 18, 6063–6074 (1998).
CAS PubMed PubMed Central Google Scholar
Ramachandran, P. et al. Resolving the fibrotic niche of human liver cirrhosis at single-cell level. Nature 575, 512–518 (2019).
ADS CAS PubMed PubMed Central Google Scholar
Fabre, T. et al. Identification of a broadly fibrogenic macrophage subset induced by type 3 inflammation. Sci. Immunol. 8, eadd8945 (2023).
CAS PubMed Google Scholar
Tong, G. et al. Fibroblast growth factor 18 attenuates liver fibrosis and HSCs activation via the SMO-LATS1-YAP pathway. Pharm. Res. 178, 106139 (2022).
CAS Google Scholar
Zhang, N. & He, Y. W. An essential role for c-FLIP in the efficient development of mature T lymphocytes. J. Exp. Med. 202, 395–404 (2005).
ADS CAS PubMed PubMed Central Google Scholar
Thai, T. H. et al. Regulation of the germinal center response by microRNA-155. Science 316, 604–608 (2007).
ADS CAS PubMed Google Scholar
Nakahara, M. et al. Gene-trap mutagenesis using Mol/MSM-1 embryonic stem cells from MSM/Ms mice. Mamm. Genome 24, 228–239 (2013).
CAS PubMed Google Scholar
Okabe, M. et al. Potential hepatic stem cells reside in EpCAM+ cells of normal and injured mouse liver. Development 136, 1951–1960 (2009).
CAS PubMed Google Scholar
Shichino, S. et al. TAS-Seq is a robust and sensitive amplification method for bead-based scRNA-seq. Commun. Biol. 5, 602 (2022).
CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

The authors thank H. Oshima and M. Oshima for technical advice, Y.-W. He for providing Cflar^FF mice, and M. Tanaka, C.J. Okumura, and T. Maruyama for generating anti-human FGF18 antibodies. This work was supported in part by Grants-in-Aid for Scientific Research (B) (20H03475 and 23H02707 to Hiroyasu N), Scientific Research (C) (17K08994 and 20K11589 to YT), and Transformative Research Areas (B) (22H05064 to SS) from Japan Society for the Promotion of Science (JSPS); the Japan Agency for Medical Research and Development through AMED-CREST (22gm1210002 to MT and Hiroyasu N) and AMED-PRIME (21gm6210025 to SS) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan; Toho University Grant for Research Initiative Program (TUGRIP) (to Hiroyasu N); the Science Research Promotion Fund from the Promotion and Mutual Aid Corporation for Private Schools of Japan (to Hiroyasu N and SY); and the Princess Takamatsu Cancer Research Fund (to Hiroyasu N).

Author information

These authors contributed equally: Yuichi Tsuchiya, Takao Seki.

Authors and Affiliations

Department of Biochemistry, Faculty of Medicine, Toho University, 5-21-16 Omori-Nishi, Ota-ku, Tokyo, 143-8540, Japan
Yuichi Tsuchiya, Takao Seki, Kenta Kobayashi, Sachiko Komazawa-Sakon, Takashi Nishina, Soh Yamazaki & Hiroyasu Nakano
Department of Biochemistry, Faculty of Pharmaceutical Sciences, Toho University, 2-2-1 Miyama, Funabashi-shi, Chiba, 274-8510, Japan
Yuichi Tsuchiya
Department of Biological Science and Technology, Faculty of Advanced Engineering, Tokyo University of Science, 6-3-1 Niijuku, Katsushika-ku, Tokyo, 125-8585, Japan
Kenta Kobayashi & Chiharu Nishiyama
Division of Molecular Regulation of Inflammatory and Immune Diseases, Research Institute for Biomedical Sciences, Tokyo University of Science, 2669 Yamazaki, Noda-shi, Chiba, 278-0022, Japan
Shigeyuki Shichino, Satoshi Ueha & Kouji Matsushima
Department of Gastroenterology, Faculty of Medicine and Graduate School of Medicine, Juntendo University, 2-1-1 Hongo, Bunkyo-Ku, Tokyo, 113-8421, Japan
Kyoko Fukuhara & Kenichi Ikejima
Department of Gastroenterology, Toho University Omori Medical Center, 6-11-1 Omori-Nishi, Ota-ku, Tokyo, 143-8541, Japan
Hidenari Nagai & Yoshinori Igarashi
Laboratory of Quality Analysis and Assessment, Graduate School of Agriculture, Kyoto University, Gokasyo, Uji-shi, Kyoto, 611-0011, Japan
Akira Oikawa
Department of Regenerative Medicine, Research Institute, National Center for Global Health and Medicine, 1-21-1 Toyama, Shinjuku-ku, Tokyo, 162-8655, Tokyo, Japan
Shinya Tsurusaki & Minoru Tanaka
Laboratory of Stem Cell Regulation, Institute for Quantitative Biosciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo, 113-0032, Tokyo, Japan
Shinya Tsurusaki & Minoru Tanaka
Department of Pathology, Faculty of Medicine, Toho University, 5-21-16 Omori-Nishi, Ota-ku, Tokyo, 143-8540, Japan
Tetuo Mikami
Department of Immunology, Faculty of Medicine and Graduate School of Medicine, Juntendo University, 2-1-1 Hongo, Bunkyo-Ku, Tokyo, 113-8421, Japan
Hideo Yagita
Atopy Research Center, Faculty of Medicine and Graduate School of Medicine, Juntendo University, 2-1-1 Hongo, Bunkyo-ku, Tokyo, 113-8421, Japan
Ko Okumura
Laboratory of Cell Growth and Differentiation, Institute for Quantitative Biosciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo, 113-0032, Japan
Taketomo Kido & Atsushi Miyajima
Department of Genetics, Hyogo Medical University, 1-1 Mukogawa-cho, Nishinomiya-shi, Hyogo, 663-8501, Japan
Mai Imasaka & Masaki Ohmuraya
Center for Animal Resources and Development, Kumamoto University, 2-2-1 Honjo, Chuo-ku, Kumamoto, 860-0811, Japan
Kimi Araki
Center for Metabolic Regulation of Healthy Aging, Kumamoto University, 1-1-1 Honjo, Chuo-ku, Kumamoto, 860-8556, Japan
Kimi Araki
Hoshi University School of Pharmacy and Pharmaceutical Sciences, 2-4-41 Ebara, Shinagawa-ku, Tokyo, 142-8501, Japan
Toru Imamura
National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Umezono, Tsukuba-shi, Ibaraki, 305-8560, Japan
Toru Imamura

Authors

Yuichi Tsuchiya
View author publications
You can also search for this author in PubMed Google Scholar
Takao Seki
View author publications
You can also search for this author in PubMed Google Scholar
Kenta Kobayashi
View author publications
You can also search for this author in PubMed Google Scholar
Sachiko Komazawa-Sakon
View author publications
You can also search for this author in PubMed Google Scholar
Shigeyuki Shichino
View author publications
You can also search for this author in PubMed Google Scholar
Takashi Nishina
View author publications
You can also search for this author in PubMed Google Scholar
Kyoko Fukuhara
View author publications
You can also search for this author in PubMed Google Scholar
Kenichi Ikejima
View author publications
You can also search for this author in PubMed Google Scholar
Hidenari Nagai
View author publications
You can also search for this author in PubMed Google Scholar
Yoshinori Igarashi
View author publications
You can also search for this author in PubMed Google Scholar
Satoshi Ueha
View author publications
You can also search for this author in PubMed Google Scholar
Akira Oikawa
View author publications
You can also search for this author in PubMed Google Scholar
Shinya Tsurusaki
View author publications
You can also search for this author in PubMed Google Scholar
Soh Yamazaki
View author publications
You can also search for this author in PubMed Google Scholar
Chiharu Nishiyama
View author publications
You can also search for this author in PubMed Google Scholar
Tetuo Mikami
View author publications
You can also search for this author in PubMed Google Scholar
Hideo Yagita
View author publications
You can also search for this author in PubMed Google Scholar
Ko Okumura
View author publications
You can also search for this author in PubMed Google Scholar
Taketomo Kido
View author publications
You can also search for this author in PubMed Google Scholar
Atsushi Miyajima
View author publications
You can also search for this author in PubMed Google Scholar
Kouji Matsushima
View author publications
You can also search for this author in PubMed Google Scholar
Mai Imasaka
View author publications
You can also search for this author in PubMed Google Scholar
Kimi Araki
View author publications
You can also search for this author in PubMed Google Scholar
Toru Imamura
View author publications
You can also search for this author in PubMed Google Scholar
Masaki Ohmuraya
View author publications
You can also search for this author in PubMed Google Scholar
Minoru Tanaka
View author publications
You can also search for this author in PubMed Google Scholar
Hiroyasu Nakano
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Y.T., M.T., and Hiroyasu N. designed the study and interpreted the results. Y.T., T.S., K.K., S.K-S., T.N., and S.T. performed and analyzed most experiments. T.S., S.S., S.U., and K.M. analyzed the scRNA-seq results. A.O. performed bile acid analysis. H.Y., K.O., M.I., K.A., T.I., and M.O. provided critical reagents. Y.T. constructed the plasmids. K.F. and K.I. collected the patients' characteristics and supplied human liver biopsy samples. Hidenari N., Y.I., S.Y., C.N., T.M., T.K., and A.M. supervised the experiments. Y.T. and Hiroyasu N. wrote the manuscript with constructive input from all authors.

Corresponding author

Correspondence to Hiroyasu Nakano.

Ethics declarations

Competing interests

Y.T. and Hiroyasu N. are inventors on the patent application that includes the ELISA to detect human and murine FGF18 and the generation of transgenic mice that spontaneously developed liver fibrosis used in this study. The other authors declare no conflicts of interest.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Tsuchiya, Y., Seki, T., Kobayashi, K. et al. Fibroblast growth factor 18 stimulates the proliferation of hepatic stellate cells, thereby inducing liver fibrosis. Nat Commun 14, 6304 (2023). https://doi.org/10.1038/s41467-023-42058-z

Download citation

Received: 29 June 2022
Accepted: 28 September 2023
Published: 09 October 2023
DOI: https://doi.org/10.1038/s41467-023-42058-z

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.