IntroductionPersistent hepatic injury, regardless of etiology, usually leads to liver fibrosis, a disease state characterized by the accumulation of extracellular matrix (ECM).1 Liver fibrosis can progress to cirrhosis and hepatocellular carcinoma, the leading causes of death worldwide.2 Importantly, in patients with metabolic dysfunction-associated steatotic liver disease (MASLD), the most prevalent chronic liver disease, fibrosis is the critical determinant of morbidity and mortality.3,4,5,6,7 Although resmetirom and semaglutide have recently been approved for the treatment of metabolic dysfunction-associated steatohepatitis (MASH) with fibrosis,8,9,10,11 effective therapeutic strategies that directly and robustly target liver fibrosis and cirrhosis remain limited, indicating that a substantial unmet medical need persists.Hepatic stellate cells (HSCs) are the primary source of ECM production during liver fibrosis.12,13,14,15 In the healthy liver, HSCs remain in quiescent states, in which they mainly store vitamin A lipid droplets. Upon liver injury, these cells transdifferentiate into myofibroblast-like cells and exhibit different activated phenotypes as ECM-producing cells.12,13,16 Thus, targeting and modulating HSC activation is considered a promising strategy for managing liver fibrosis.17 Since the process of HSC activation and the resulting pathophenotypes are governed by complex signaling pathways,12,16 elucidating the detailed mechanisms that control HSC fate is essential for the development of effective antifibrotic therapies.Yes-associated protein (YAP) is a master effector downstream of the Hippo signaling pathway that plays pivotal roles in the regulation of organ size and tissue homeostasis.18,19,20 Consequently, dysregulation of this pathway contributes to the pathogenesis of various diseases, including multiple cancers.18,21,22 Notably, YAP has emerged as a key target in organ fibrogenesis and HSC activation.23,24,25,26 YAP functions as a transcriptional coactivator of the transcriptionally enhanced associate domain (TEAD) family, regulating the expression of downstream target genes involved in fibrogenesis.18,22,27 Nonetheless, little is known about the upstream regulators of YAP signaling in the context of hepatic fibrogenesis. In particular, the mechanisms through which YAP is activated, in contrast to the well-known inhibitory phosphorylation by large tumor suppressor kinase (LATS),18,28 remain poorly understood.The p21-activated kinase (PAK) family comprises serine/threonine kinases and is classified into group I (PAK1–3) and group II (PAK4–6) on the basis of distinct structural and functional characteristics.29 Compared with group I PAKs, group II members, including PAK4, have been relatively less well studied. Owing to its elevated expression in various cancer cells and its role in cytoskeleton remodeling and cell growth, most previous studies have focused on the oncogenic functions of PAK4.29 However, emerging evidence suggests that PAK4 also plays important roles in diverse pathophysiologic situations in noncancerous cells. Our groups recently reported that PAK4 is crucial for the regulation of lipid metabolism30,31 and skeletal muscle functions.32,33 In addition, we found that hepatocyte PAK4 contributes to ischemia‒reperfusion injury in the liver.34 Moreover, while the involvement of PAK1 in integrin signaling in HSCs has been suggested,35 the potential roles of PAK4 and its downstream substrate targets in HSC activation remain largely unexplored.In this study, we found that PAK4 is upregulated in activated HSCs in both human and murine fibrotic livers and revealed a formerly unknown profibrogenic PAK4–YAP signaling axis in HSCs: PAK4 directly phosphorylates YAP at T428, a previously unrecognized site, leading to YAP activation and subsequent HSC activation that drives liver fibrogenesis. To uncover this, we employed multiple liver fibrosis animal models in combination with HSC-specific Pak4 deletion or pharmacological inhibition of PAK4. Complementary approaches–including omics analyses, mutagenesis, and molecular docking–were utilized to explore the underlying mechanism. Finally, the clinical relevance of this signaling axis was validated through analysis of liver tissue samples from patients with liver fibrosis using a newly developed antibody against T428-phosphorylated YAP.ResultsPAK4 is upregulated in activated HSCs in human and murine liver fibrosisTo explore the involvement of PAK4 in HSC activation and liver fibrosis, we initially examined PAK4 expression levels in various murine models of liver fibrosis. In liver tissues from mice chronically treated with carbon tetrachloride (CCl4), a well-established model of liver fibrosis, the protein expression of PAK4 markedly increased, along with that of the HSC activation marker alpha-smooth muscle actin (α-SMA) (Fig. 1a). Given that PAK4 undergoes autophosphorylation upon activation,29,36 we further confirmed that phosphorylation at S474 paralleled the increase in PAK4 expression (Fig. 1a). In contrast, the expression of other PAK family members was not significantly altered by CCl4 treatment, although PAK2 expression was reduced (Supplementary Fig. 1a, b).Fig. 1Full size imageUpregulation of PAK4 in activated HSCs in patients and mice with liver fibrosis. a Western blot analysis of PAK4 and p-PAK4 (S474) in liver tissues from mice treated with CCl4 as depicted in the experimental scheme (Vehicle, n = 3; CCl4, n = 4). Relative band intensities of p-PAK4 and total PAK4 were quantified and normalized to those of HSP90. b–e Coimmunofluorescence (Co-IF) staining of PAK4 and α-SMA in liver tissue from mice injected with CCl4 (n = 3 per group) (b), mice subjected to bile duct ligation (BDL) (n = 4 per group) (c), mice fed a choline-deficient, L-amino acid-defined, high-fat diet (CDAHFD) (n = 3 per group) (d), and human liver samples from individuals with normal, fibrosis or cirrhosis (Normal, n = 8; Fibrosis, n = 11; Cirrhosis, n = 8) (e). The number of PAK4- and α-SMA-double-positive cells was normalized to the total number of DAPI-stained cells. Scale bars: 20 μm. f Western blot analysis and immunofluorescence staining of PAK4 in primary mouse HSCs under quiescent and activated conditions. Scale bars: 50 μm. g qPCR analysis of Pak4 in quiescent (Q) and activated (A) HSCs (n = 4 per group). h Conserved binding motifs for HIF1α or YY1 within the human and murine PAK4 promoters. i ChIP assays validating direct HIF1α or YY1 binding to PAK4 promoter regions in LX-2 cells (n = 3 per group). Western blot analysis of PAK4 expression in LX-2 cells transfected with si-HIF1α (j) or si-YY1 (k) (n = 3 per group). An asterisk indicates the HIF-1α band. The data are presented as the mean ± SEM. Statistical significance was assessed using unpaired two-tailed Student’s t tests (for a–d, g and i–k) or one-way ANOVA followed by Tukey’s multiple comparisons test (for e)PAK4 expression in liver tissues was analyzed in depth using immunofluorescence (IF) staining, which revealed that the colocalization of PAK4 expression with α-SMA-positive cells was significantly increased in fibrotic livers (Fig. 1b). These findings were also verified in other liver fibrosis models, including bile duct ligation (BDL) (Fig. 1c) and choline-deficient, L-amino acid-defined, high-fat diet (CDAHFD) feeding (Fig. 1d). To gain further insight into clinical relevance, we examined PAK4 expression in patients with liver fibrosis. Consistent with the findings in the animal models, compared with that in normal, PAK4 expression in activated HSCs in fibrosis and cirrhosis patients gradually increased according to disease severity (Fig. 1e). Since disease etiology may influence fibrosis pathobiology, we further characterized the patient cohort (Supplementary Table 1) and performed subgroup analyses. PAK4 expression remained significantly different across fibrosis stages in hepatitis B virus (HBV)-associated patients, which represents a major etiology in our cohort (Supplementary Fig. 1c). Among cirrhotic patients, PAK4 expression did not differ between patients with different HBV-related and nonviral etiologies (Supplementary Fig. 1d), suggesting that our findings may be broadly applicable to liver fibrosis across different etiologies.Moreover, in vitro cell models with primary mouse HSCs accurately validated the higher expression levels of PAK4 and p-PAK4 in activated HSCs than in quiescent HSCs (Fig. 1f and Supplementary Fig. 1e). To understand how PAK4 expression is upregulated in activated HSCs, we next assessed Pak4 mRNA levels and found them to be consistently elevated during HSC activation (Fig. 1g), suggesting the transcriptional regulation of PAK4. Using the Evolutionary Conserved Regions (ECR) browser, the transcription factors hypoxia-inducible factor 1 alpha (HIF-1α) and yin yang 1 (YY1) were predicted to bind to the conserved proximal promoter regions of the PAK4 gene in both humans and mice (Fig. 1h). Chromatin immunoprecipitation (ChIP) assays confirmed the binding of these two regulators to the PAK4 promoter (Fig. 1i). Consistent with these findings, silencing either HIF-1α or YY1 expression significantly reduced PAK4 expression in LX-2 cells, a human HSC line (Fig. 1j, k). Collectively, these findings indicate that PAK4 expression is transcriptionally enhanced in activated HSCs, implying a role for PAK4 in the progression of hepatic fibrogenesis.PAK4 deficiency suppresses HSC activation in vitroGiven that PAK4 expression is induced in activated HSCs, we hypothesized that PAK4 deficiency may inhibit HSC activation. To test this hypothesis, we employed two different genetic approaches to silence PAK4 during the course of HSC activation. First, primary HSCs were isolated from the livers of Pak4f/f mice and cultured in vitro for 3 days, after which they were infected with adenovirus expressing Cre recombinase during an additional cultivation period of 3 days (Fig. 2a). PAK4 expression was efficiently silenced in these HSCs, leading to a significant reduction in the expression of α-SMA and type I collagen (COL1A1), a hallmark of liver fibrosis and the most abundant collagen type (Fig. 2b). The suppression of fibrogenesis in PAK4-deficient HSCs was further confirmed by immunostaining for α-SMA (Fig. 2c) and qPCR analysis of the expression levels of multiple fibrogenic genes (Fig. 2d). We next evaluated the effect of PAK4 knockdown using siRNA in HSCs derived from wild-type (WT) mice (Fig. 2e) and observed similar results (Fig. 2f–h). Furthermore, PAK4 silencing exerted an antifibrotic effect on LX-2 cells (Supplementary Fig. 2a), supporting the relevance of our findings in a human HSC model.Fig. 2Full size imageLoss of PAK4 inhibits the fibrogenic activity of HSCs. a, e Schematic diagrams illustrating the experimental design for PAK4 silencing in primary mouse HSCs. b, f Western blot analysis of α-SMA and COL1A1 expression in PAK4-deficient HSCs (n = 3 per group). c, g Immunofluorescence staining of α-SMA in PAK4-deficient HSCs. Scale bars: 50 μm. d, h qPCR analysis of fibrogenic genes in PAK4-deficient HSCs (d, n = 3 per group; h, n = 4 per group). Gene Ontology (GO) (i) and gene set enrichment analysis (GSEA) (j) plots based on RNA-seq data from PAK4 knockdown HSCs (n = 4 per group). k Heatmap showing significantly decreased expression of collagen formation-related genes in PAK4 knockdown HSCs (n = 4 per group). The data are presented as the mean ± SEM. Statistical significance was assessed using unpaired two-tailed Student’s t testsIn addition to their fibrotic properties, activated HSCs exhibit multiple pathological characteristics, including proliferation, migration, and contractility.12,13 Our series of experiments demonstrated that PAK4 knockdown suppressed all of these features, as evidenced by reduced HSC proliferation and migration and attenuated collagen gel contraction (Supplementary Fig. 2b–d).To explore the biological impact of PAK4 on HSC activation in depth, we performed comprehensive RNA sequencing (RNA-seq) on PAK4-silenced HSCs. Consistent with the observed changes in HSC pathophenotypes, gene ontology (GO) analysis revealed marked alterations in genes associated with ECM organization, cell migration, and proliferation upon PAK4 inhibition (Fig. 2i). Additionally, gene set enrichment analysis (GSEA) demonstrated significant downregulation of collagen formation-related genes in PAK4-deficient HSCs (Fig. 2j, k), supporting a functional role for PAK4 in HSC activation and fibrogenesis.Deletion of PAK4 in HSCs attenuates liver fibrosis in vivoTo investigate whether the loss of PAK4 in HSCs decreases liver fibrosis in vivo, we used an established model to delete a target gene in activated HSCs37,38,39,40 and generated Pak4f/f;Pdgfrb-CreERT2 mice. We injected CCl4 into Pak4f/f littermate (LT) and Pak4f/f;Pdgfrb-CreERT2 mice for 2 weeks to induce HSC activation, followed by tamoxifen administration for 5 consecutive days, with CCl4 treatment maintained for an additional 4 weeks (Fig. 3a). The genotypes of the mice and effective HSC-specific deletion of PAK4 by the inducible Cre-mediated knockout (KO) system were verified (Supplementary Fig. 3a–c). Sirius red and Masson’s trichrome staining revealed that the accumulation of collagen significantly decreased in the livers of HSC-specific Pak4-KO (Pak4ΔHSC) mice compared with those of LT controls (Fig. 3b, c), and this decrease was accompanied by the attenuation of the content of hydroxyproline, a key component of collagen (Fig. 3d). Consistently, reduced HSC activation was observed, as evidenced by decreased α-SMA expression (Fig. 3b, e). A secondary antibody-only control experiment confirmed the specificity of α-SMA staining (Supplementary Fig. 3d). The transcript levels of α-SMA (Acta2) and genes encoding ECM proteins such as collagens (Col1a1, Col1a2 and lumican (Lum)) were also decreased (Fig. 3f). Notably, the antifibrotic effects of PAK4 deletion on HSCs were also observed in the BDL model (Fig. 3g–l).Fig. 3Full size imageThe HSC-specific silencing of PAK4 diminishes liver fibrosis in mice. Schematic diagrams illustrating the experimental design for HSC-specific Pak4 KO (Pak4ΔHSC) mice subjected to two different liver fibrosis models: CCl4 treatment (a) or BDL (g). b Histological analyses including Sirius red, Masson’s trichrome, H&E, and α-SMA immunohistochemical staining of liver tissues from CCl4-treated mice. Scale bars: 20 μm. Quantification of the Sirius red–positive area (c), hydroxyproline content (d), α-SMA–positive area (e), and results of the qPCR analysis of fibrogenic genes (f) (LT-Vehicle, n = 6; LT-CCl4, n = 8; Pak4ΔHSC-Vehicle, n = 4; Pak4ΔHSC-CCl4, n = 6). h Histological analyses of Sirius red, Masson’s trichrome, H&E, and α-SMA IHC staining of liver tissues from BDL-operated mice. Scale bars: 20 μm. Quantification of the Sirius red–positive area (i), hydroxyproline content (j), α-SMA–positive area (k), and results of the qPCR analysis of fibrogenic genes (l) (LT-Sham, n = 3; LT-BDL, n = 6; Pak4ΔHSC-Sham, n = 3; Pak4ΔHSC-BDL, n = 6). The data are presented as the mean ± SEM. Statistical significance was assessed using two-way ANOVA followed by Tukey’s multiple comparisons testTo corroborate these findings, we employed an additional approach: Pak4f/f mice were subjected to lentiviral delivery of Cre recombinase under the control of the Col1a1 promoter (Supplementary Fig. 3e). The HSC-specific targeting of this viral system was validated using isolated HSCs and hepatocytes (Supplementary Fig. 3f, g). When PAK4 was successfully deleted in HSCs (Supplementary Fig. 3h), the pathophenotypes of liver fibrosis, as demonstrated by histological and gene expression changes, markedly improved (Supplementary Fig. 3i–m).On the basis of our in vitro findings concerning the role of PAK4 in HSC proliferation, we next evaluated this effect in in vivo models. Similarly, the number of proliferating HSCs in the liver, identified as proliferating cell nuclear antigen (PCNA)- and α-SMA-double-positive cells, was significantly reduced by PAK4 deficiency (Supplementary Fig. 4a–c), suggesting that the attenuation of fibrogenesis observed in vivo is partly attributable to decreased HSC expansion. In contrast, coimmunostaining for cleaved caspase-3, a marker of apoptosis, and α-SMA indicated that HSC apoptosis was not altered by HSC-specific PAK4 deletion in mice (Supplementary Fig. 4d–f); thus, apoptosis was excluded as a potential confounding factor.The activities of serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were not significantly altered in either of the HSC-specific Pak4 KO mouse models, suggesting the primary impact of PAK4 modulation on HSCs (Supplementary Fig. 4g–i). Overall, these findings clearly demonstrate that the silencing of PAK4 in HSCs effectively diminishes the development of hepatic fibrosis.PAK4 stabilizes YAP in activated HSCsTo elucidate the molecular mechanisms by which PAK4 regulates the phenotypes of HSCs, we further analyzed our RNA-seq data using KEGG pathway analysis and found that the Hippo signaling pathway was significantly altered in Pak4-deficient HSCs (Fig. 4a). We subsequently identified YAP as a key molecule closely linked to genes involved in the Hippo signaling pathway (Fig. 4b), prompting us to focus on the potential role of PAK4 in regulating YAP.Fig. 4Full size imagePAK4 inhibition promotes YAP degradation via ubiquitination. KEGG pathway analysis of RNA-seq data (a) and GeneMANIA analysis showing a close association of YAP with the identified Hippo signaling pathway (b). c Western blot analysis of YAP in quiescent and activated HSCs. d YAP-responsive luciferase assays in HEK293T cells overexpressing WT-PAK4 or kinase-dead mutant PAK4S474A together with YAP (n = 3 per group). e qPCR analysis of YAP target genes in LX-2 cells overexpressing the constructs shown in panel d (n = 4 per group). Western blot analysis of YAP (f) and qPCR analysis of Yap and Ctgf (g) expression in PAK4 knockdown HSCs (n = 5 per group). h Cycloheximide (CHX) chase assays were performed to determine YAP protein stability (n = 4 per group). i Immunoprecipitation (IP) assays monitoring the ubiquitination status of YAP under PAK4 overexpression conditions. j Western blot analysis of YAP in LX-2 cells transfected with siRNAs against the indicated E3 ligases in combination with PAK4 knockdown (n = 3 per group). k Schematic illustration of the proposed role of PAK4-mediated YAP phosphorylation in its degradation. The data are presented as the mean ± SEM. Statistical significance was assessed using one-way ANOVA (for d, e), two-way ANOVA (for j) followed by Tukey’s multiple comparisons test, or unpaired two-tailed Student’s t tests (for h, g)We first evaluated the expression of YAP in HSCs and found that its levels increased during HSC activation (Fig. 4c), which is in line with the induction of PAK4 expression (Fig. 1f). To determine whether PAK4 regulates YAP activity, we assessed the transcriptional activity of YAP in the context of PAK4 expression. Overexpression of PAK4 significantly increased YAP-responsive luciferase activity (Fig. 4d) and increased the transcription levels of YAP target genes, including connective tissue growth factor (CTGF) and cysteine-rich angiogenic inducer 61 (CYR61, CCN1), in LX-2 cells (Fig. 4e), indicating PAK4-mediated transactivation of YAP.Considering the concomitant induction of PAK4 and YAP expression, we next investigated the role of PAK4 in regulating YAP expression. Knockdown of PAK4 expression significantly reduced YAP protein levels in primary mouse HSCs (Fig. 4f), which was strengthened by decreased expression of the YAP target gene Ctgf (Fig. 4g). Interestingly, PAK4 silencing did not affect Yap mRNA levels (Fig. 4g), suggesting the possibility of posttranscriptional regulation. Indeed, a cycloheximide (CHX) chase assay demonstrated that forced expression of PAK4 increased YAP protein stability (Fig. 4h). Consistently, PAK4 overexpression decreased YAP ubiquitination in the presence of the proteasome inhibitor MG-132 (Fig. 4i), indicating that PAK4 inhibits the ubiquitin‒proteasome-mediated degradation of YAP. Notably, a kinase-dead mutant of PAK4 (PAK4S474A) abolished all of these effects of PAK4 on the transactivation and stabilization of YAP (Fig. 4d, e, h, i), suggesting that the kinase activity of PAK4 is required for YAP activation.To elucidate the detailed mechanism underlying PAK4-induced stabilization of YAP, we further characterized possible E3 ubiquitin ligases involved in YAP degradation. Previous studies have suggested that YAP is targeted for degradation by beta-transducin repeat-containing protein (β-TRCP) and F-Box and WD repeat domain-containing 7 (FBXW7).28,41,42 Our additional experiments using siRNAs against β-TRCP or FBXW7 revealed that PAK4 silencing-induced YAP destabilization was reversed by knockdown of β-TRCP, but not FBXW7, in LX-2 cells (Fig. 4j). These findings support the concept that PAK4-dependent phosphorylation of YAP promotes its stabilization by blocking β-TRCP-mediated proteasomal degradation (Fig. 4k).PAK4 directly phosphorylates YAP at T428 to promote its activationGiven the importance of PAK4 kinase activity in the activation of YAP, we considered the possibility that PAK4 directly phosphorylates YAP. To test our hypothesis, we first examined the interaction between these two proteins and found that PAK4 binds to YAP, as evidenced by the results of coimmunoprecipitation (Co-IP) and proximity ligation assays (Fig. 5a, b). Since PAK4 did not bind to TAZ, a paralog of YAP (Fig. 5a), we focused on YAP and excluded TAZ as a potential substrate of PAK4.Fig. 5Full size imageYAP is phosphorylated by PAK4 at T428 for its activation. Coimmunoprecipitation (Co-IP) (a) and proximity ligation assays (PLA) (b) monitoring the direct interaction between PAK4 and YAP. Scale bars: 50 μm. c In vitro kinase assays demonstrating direct phosphorylation of YAP by PAK4. d IP assays detecting serine and threonine phosphorylation of YAP under PAK4 overexpression conditions. e Computational model of the interaction geometry between PAK4 and YAP. f Western blot analysis of T428-phosphorylated YAP in HEK293T cells overexpressing PAK4 and YAP (n = 3 per group). g Western blot analysis of T428-phosphorylated YAP in PAK4-knockdown LX-2 cells (n = 3 per group). h IP assays evaluating threonine and serine phosphorylation in WT-YAP and YAPT428A mutants. i IP assays showing the ubiquitination status of WT-YAP and the YAPT428A mutant. j CHX chase assays comparing the protein stability of WT-YAP and the YAPT428A mutant (n = 3 per group). k Immunofluorescence staining of YAP to determine its subcellular localization in LX-2 cells transfected with WT-YAP (1 μg) or the YAPT428A mutant (2.5 μg). Scale bars: 20 μm. l YAP-responsive luciferase assays to estimate YAP transcriptional activity in HEK293T cells overexpressing WT-YAP (1 μg) and YAP mutants (YAPT428A, 2.5 μg; YAPT428E, 0.5 μg) in the presence of the PAK4 construct (n = 3 per group). The data are presented as the mean ± SEM. Statistical significance was assessed using one-way ANOVA (for f, l) followed by Tukey’s multiple comparisons test or unpaired two-tailed Student’s t tests (for g, j)In vitro cell-free kinase assays using recombinant human PAK4 clearly demonstrated the direct phosphorylation of YAP, along with the autophosphorylation of PAK4 (Fig. 5c). Additional IP experiments revealed that overexpression of WT-PAK4, but not the PAK4S474A mutant, markedly increased the phosphorylation of YAP at threonine residues but not at serine residues (Fig. 5d). Next, we used in silico prediction tools, including PhosphoNET and PhosphoSitePlus, to identify potential phosphorylation sites of YAP by PAK4 and identified five putative sites. Among them, competition assays using oligopeptides containing each of these residues revealed that PAK4 specifically phosphorylates YAP at T428 (Supplementary Fig. 5a), a site that is highly conserved across species (Supplementary Fig. 5b). Notably, the phosphorylation of YAP at T428 was further supported by site-specific detection using liquid chromatography‒tandem mass spectrometry (LC‒MS/MS) analysis (Supplementary Fig. 5c).To better evaluate the structural feasibility of the newly identified phosphorylation site, we employed a multistep computational modeling strategy. A peptide segment encompassing YAP-T428 (423MDTGDTINQST433) was first docked to the catalytic domain of PAK4, and the resulting complex was then subjected to short-term molecular dynamics simulations with backbone restraints to relieve steric clashes and stabilize the interface. This integrated workflow consistently yielded complexes in which the T428-containing region of YAP was oriented toward the catalytic pocket of PAK4, with the threonine hydroxyl group positioned within the hydrogen bond distance to the catalytic residue (Fig. 5e). These findings provide a structural rationale that complements the biochemical, mutagenesis, and LC–MS/MS data, reinforcing that T428 is a direct site of PAK4 phosphorylation.Since we discovered that T428 phosphorylation of YAP is a previously unrecognized site, we generated a phospho-specific antibody against p-T428-YAP to validate this modification. In a cell-free system, recombinant human PAK4 induced YAP T428 phosphorylation, and the signal was abolished by peptide competition (Supplementary Fig. 5d), supporting antibody specificity. The enforced expression of PAK4 significantly increased the level of T428-phosphorylated YAP (Fig. 5f), whereas the silencing of PAK4 reduced this protein in LX-2 cells (Fig. 5g), thereby confirming that YAP is phosphorylated at T428 in HSCs. Residual T428 phosphorylation of YAP remained detectable in PAK4-knockdown LX-2 cells, which may reflect incomplete PAK4 silencing under these experimental conditions and/or compensatory phosphorylation by other as-yet-unidentified kinases.To further validate the functional significance of YAP phosphorylation at T428, we generated phosphoinactive (YAPT428A) and phosphomimetic (YAPT428E) mutants. The T428A mutation selectively abolished PAK4-induced threonine, but not serine, phosphorylation of YAP (Fig. 5h). As a consequence, compared with WT-YAP, the inactive YAPT428A mutant exhibited rapid protein degradation with increased ubiquitination in the presence of PAK4 (Fig. 5i, j). Consistent with these findings, the YAPT428E mimetic mutation had the opposite effect (Supplementary Fig. 5e, f), supporting the role of T428 phosphorylation in the regulation of YAP protein stability.Because YAP functions as a transcriptional activator, we next assessed the effects of T428 phosphorylation on its nuclear localization and transcriptional activity. Given the role of PAK4 in regulating YAP stability, the overexpression of WT-YAP, YAPT428A, and YAPT428E mutants at equal plasmid amounts in the presence of PAK4 resulted in different YAP expression levels (Supplementary Fig. 5g). Therefore, we adjusted the amounts of expression vectors to achieve comparable YAP protein levels (Supplementary Fig. 5g), thereby excluding secondary effects arising from differences in protein abundance. Under these conditions, compared with WT-YAP, the YAPT428A mutant exhibited increased cytoplasmic retention (Fig. 5k). Moreover, compared with WT-YAP, the phospho-inactive YAPT428A mutant displayed significantly reduced transcriptional activity, whereas the phospho-mimetic YAPT428E mutant had the opposite effect (Fig. 5l). Collectively, these results support the notion that PAK4-mediated phosphorylation of YAP at T428 stabilizes the protein abundance of YAP, promotes its nuclear accumulation, and enhances its transcriptional activity, thereby contributing to coordinated YAP activation.On the basis of the known role of LATS1 as an upstream kinase involved in the inhibition of YAP signaling,18,28 we investigated whether LATS1 participates in the PAK4-mediated regulation of YAP. The levels of both T1079-phosphorylated active and total LATS1 were unchanged by PAK4 knockdown in LX-2 cells (Supplementary Fig. 6a). In addition, the inhibitory phosphorylation of YAP at S127 induced by LATS1 was decreased in PAK4-deficient LX-2 cells (Supplementary Fig. 6a), presumably due to reduced YAP expression under these conditions. These results further support that the effect of PAK4 on YAP phosphorylation and subsequent stabilization is direct and not indirectly mediated through LATS1 regulation. Moreover, the destabilizing effect of PAK4 silencing on YAP expression remained evident even under LATS1 knockdown conditions (Supplementary Fig. 6b), also supporting a LATS1-independent mechanism. Notably, the YAPT428E mutant (i.e., mimicking phosphorylation by PAK4) was significantly resistant to LATS1-stimulated degradation (Supplementary Fig. 6c, d), indicating that PAK4 counteracts this process.Overexpression of the phospho-inactive YapT413A mutant in HSCs attenuates liver fibrosisTo investigate the functional consequences of YAP phosphorylation by PAK4 in hepatic fibrogenesis, we administered lentiviral vectors carrying Col1a1 promoter-driven overexpression constructs of either Flag-tagged WT-Yap or the phospho-inactive YapT413A mutant (i.e., the mouse Yap T413 site orthologous to the human YAP T428 residue) into mice. During the course of CCl4 treatment, the mice received intravenous injections of the lentiviruses after the first 2 weeks, and phenotypic analyses were conducted following an additional 4 weeks of treatment (Fig. 6a). The successful overexpression of YAP in HSCs was confirmed by the colocalization of α-SMA and Flag (Supplementary Fig. 7a). In addition, Yap mRNA levels were comparably elevated in both the WT-Yap and the YapT413A mutant groups (Supplementary Fig. 7b), whereas Yap protein levels were lower in the YapT413A group than in the WT-Yap group (Supplementary Fig. 7a, c), reflecting a phosphorylation-dependent protein stabilizing effect.Fig. 6Full size imageOverexpression of the YapT413A mutant in HSCs attenuates liver fibrosis. a Schematic diagram illustrating the experimental design for HSC-specific overexpression of Flag-Yap or the Flag-YapT413A mutant in mice with CCl4-induced liver fibrosis. b Histological analyses of liver tissues stained with Sirius red, Masson’s trichrome, H&E, and α-SMA by IHC. Scale bars: 20 μm. Quantification of the Sirius red–positive area (c) and α-SMA–positive area (d) (LV-Con+Veh, n = 8; LV-Con+CCl4, n = 12; LV-Yap+CCl4, n = 11; LV-YapT413A + CCl4, n = 11). qPCR analyses of fibrogenic genes (e) and YAP target genes (f) (LV-Con+Veh, n = 7; LV-Con+CCl4, n = 11; LV-Yap+CCl4, n = 10; LV-YapT413A + CCl4, n = 11). Co-IF staining of YAP and α-SMA (g) or PAK4 (h) in human liver samples from individuals with normal, fibrosis or cirrhosis (Normal, n = 8; Fibrosis, n = 11; Cirrhosis, n = 8). Scale bars: 20 μm. i Positive correlation between the populations of PAK4-expressing and YAP-expressing HSCs on the basis of the quantification shown in panels (g) and (h). j Western blot analysis of p-YAP (T428) in human liver samples from individuals with normal, fibrosis or cirrhosis (Normal, n = 8; Fibrosis, n = 11; Cirrhosis, n = 8). The data are presented as the mean ± SEM. Statistical significance was assessed using one-way ANOVA followed by Tukey’s multiple comparisons testCollagen fiber accumulation in the liver was significantly enhanced by overexpression of WT-Yap in HSCs (Fig. 6b, c). In contrast, overexpression of the inactive YapT413A mutant did not induce such fibrotic responses but instead further decreased collagen deposition compared with that in the CCl4 alone treatment group (Fig. 6b–d), indicating the inhibitory effects of the phospho-inactive Yap mutant on the development of liver fibrosis. The effects of WT-Yap and the YapT413A mutant on fibrogenic and YAP target gene expression were also validated by qPCR (Fig. 6e, f). Similar to the results from the HSC-specific Pak4-KO mouse models, serum ALT and AST activities did not significantly change (Supplementary Fig. 7d). These findings provide compelling evidence for the critical causal role of PAK4-induced YAP phosphorylation in HSC activation and liver fibrogenesis.To clarify the clinical relevance of the PAK4–YAP signaling axis in pathogenesis, we examined its expression in liver tissues from patients with fibrosis. Double IF staining for YAP and α-SMA revealed that YAP expression in activated HSCs progressively increased in patients with fibrosis and cirrhosis (Fig. 6g). Moreover, HSCs with high YAP expression substantially overlapped with PAK4-positive population, indicating a significant positive correlation between PAK4 and YAP expression (Fig. 6h, i). Notably, compared with those in normal tissue, the expression levels of T428-phosphorylated YAP in liver tissue gradually increased in patients with fibrosis and cirrhosis (Fig. 6j). Overall, these findings underscore the importance of PAK4-mediated T428 phosphorylation of YAP in HSC activation and the development of liver fibrosis in patients.Pharmacological administration of a PAK4 inhibitor improves liver fibrosisTo evaluate the therapeutic potential of PAK4 inhibition in liver fibrosis, we employed ND201651, an orally available small-molecule PAK4 inhibitor.30,31 Treatment of primary mouse HSCs with ND201651 significantly reduced the protein levels of YAP and the expression of its downstream transcriptional targets, including Ctgf and Ccn1 (Fig. 7a, b), confirming that PAK4-dependent regulation of YAP occurred. Consequently, ND201651 treatment effectively suppressed HSC activation, as evidenced by decreased α-SMA expression and reduced mRNA levels of Col1a1 and Col1a2 (Fig. 7a, c, d and Supplementary Fig. 8a), without affecting cell viability (Supplementary Fig. 8b).Fig. 7Full size imageThe small-molecule PAK4 inhibitor improves liver fibrosis. a–d In vitro therapeutic effects of the PAK4 inhibitor ND201651 (ND). Western blot analysis of YAP and α-SMA expression (a), qPCR analysis of YAP target genes (b) and fibrogenic genes (d) (n = 3 per group), and IF staining of α-SMA expression (c) in ND-treated primary mouse HSCs. e Schematic diagram illustrating the experimental design for the pharmacological inhibition of PAK4 in mice with CCl4-induced liver fibrosis (Vehicle, n = 6; CCl4, n = 7; CCl4,+ND 10 mg/kg, n = 8; CCl4,+ND 30 mg/kg, n = 8). f–i In vivo therapeutic effects of ND. Western blot analysis of total and phosphorylated forms of YAP and PAK4 (f), qPCR analysis of YAP target genes (g) and fibrogenic genes (k), histological analyses of liver tissues (h, i), and hydroxyproline measurement (j) in ND-treated mice. The data are presented as the mean ± SEM. Statistical significance was assessed using one-way ANOVA followed by Tukey’s multiple comparisons testIn accordance with the results of the in vitro HSC experiments, the efficacy of the PAK4 inhibitor in controlling YAP expression and HSC activation was also validated in vivo. Mice were orally treated with ND201651 during the last 4 weeks of a 6-week CCl4 injection period (Fig. 7e). ND201651 treatment (10 or 30 mg/kg) almost completely inhibited CCl4-induced expression of YAP and its target genes, in parallel with the suppression of PAK4 activity and YAP T428 phosphorylation (Fig. 7f, g). Subcellular fractionation analysis confirmed that ND201651 treatment reduced both nuclear and cytosolic levels of p-PAK4 and YAP (Supplementary Fig. 8c). Consistent with these findings, histological examination, hydroxyproline measurement, and qPCR analysis revealed that ND201651 clearly reduced collagen accumulation and the expression of fibrogenic genes in the liver (Fig. 7h–k). Serum ALT and AST activities decreased in the 30 mg/kg treatment group (Supplementary Fig. 8d). Recovery from CCl4-induced body weight loss by ND201651 treatment further supports its beneficial effects and excludes the possibility of systemic toxicity (Supplementary Fig. 8e). Taken together, these results demonstrate that the PAK4 inhibitor has therapeutic efficacy in suppressing hepatic fibrogenesis through the inhibition of YAP-mediated HSC activation (Fig. 8).Fig. 8Full size imageProposed mechanism of PAK4-mediated T428 phosphorylation of YAP in hepatic fibrogenesis. In activated HSCs, upregulated PAK4 expression leads to the phosphorylation of YAP at T428, leading to YAP activation and subsequent HSC activation and liver fibrogenesis. Therefore, therapeutic strategies targeting PAK4 may represent potential novel treatment optionsDiscussionGiven that both clinical and preclinical studies suggest the reversibility of fibrotic processes, progression or regression,1 elucidating their dynamic pathogenesis and identifying novel therapeutic targets are fundamental. In the present study, we discovered that PAK4 acts as a novel kinase for the activating phosphorylation of YAP and that this newly characterized PAK4-YAP signaling axis plays a central role in HSC activation and liver fibrosis progression.Beyond its limited studies in cancer biology, our findings open new avenues for understanding the pathophysiological role of PAK4 and provide compelling evidence for its novel regulatory functions in HSC activation. On the basis of the upregulation of PAK4 in activated HSCs and fibrotic livers, we revealed the antifibrotic therapeutic effects of PAK4 inhibition. The loss of PAK4 suppressed multiple hallmarks of HSC activation, including ECM production, proliferation, migration, and contractility, as further supported by global gene expression profiling and in vivo validation. Moreover, HSC-specific silencing of Pak4 successfully dampened the development of liver fibrosis induced by both hepatotoxic insult and bile acid overload, highlighting its functional relevance in diverse injury contexts. Pharmacological inhibition of PAK4 also effectively reduced HSC activation and liver fibrosis. Collectively, these findings establish PAK4 as a potential master regulator controlling HSC fate toward activation, highlighting it as a promising therapeutic target for liver fibrosis.Although group I and group II PAK family members share overall protein similarity, they differ substantially in domain composition, regulatory mechanisms, and substrate specificity, resulting in context-dependent redundant or distinct biological functions. A previous study demonstrated that the group I member PAK1 mediates integrin β1 downstream signaling in HSCs and that pharmacological inhibition of PAK1 using IPA3 exerts antifibrotic effects.35 Building on this established framework, our study identifies a critical role for the group II member PAK4 in HSC activation through direct regulation of YAP, thereby revealing a nonredundant PAK signaling axis that expands the current paradigm of PAK family function in hepatic fibrogenesis.To understand the underlying mechanisms of the profibrogenic role of PAK4, we conducted an unbiased transcriptomic analysis and newly identified YAP as a critical downstream effector of PAK4 in HSCs. Importantly, our findings demonstrate that PAK4 directly phosphorylates human YAP at T428 (corresponding to T413 in mouse Yap), a previously uncharacterized and evolutionarily conserved site. This phosphorylation event was rigorously validated through peptide competition assays, mass spectrometry, molecular docking, and mutagenesis, thereby providing convergent evidence for its mechanistic significance. Functionally, PAK4-mediated phosphorylation of YAP at T428 markedly increases YAP protein stability by reducing its ubiquitination, thereby increasing overall protein abundance as a primary mechanism. Moreover, T428 phosphorylation promotes YAP nuclear localization and transcriptional activity. Together, these coordinated effects of PAK4-mediated T428 phosphorylation of YAP drive fibrogenic gene expression and HSC activation. The causal relevance of this modification was further demonstrated by HSC-specific overexpression of the phospho-inactive YapT413A mutant, which abrogated the ability of YAP to drive HSC activation and hepatic fibrogenesis. Notably, we also confirmed the increase in T428-phosphorylated YAP in liver tissues from fibrosis patients, establishing the clinical significance of this regulatory axis.The relationship between PAK4 and YAP has also been suggested by previous work showing that PAK4 inhibition reduces YAP mRNA levels in glioblastoma cells, where this signaling contributes to integrin αvβ3-mediated Glut3 expression.43 In contrast, our study demonstrated that PAK4 promotes YAP protein stabilization through phosphorylation without affecting YAP transcription in HSCs, highlighting a mechanistic distinction from observations in glioblastoma and other cellular contexts.Although YAP and its paralog TAZ share substantial overall sequence similarity (~60%), they differ in key structural features, including WW domain composition and phosphodegron organization, which can influence context-dependent interactions with regulatory proteins.44,45 In this study, we revealed that PAK4 specifically binds to YAP but not to TAZ. Consistent with the role of intrinsically disordered regions in mediating dynamic protein‒protein interactions, our docking model suggests that PAK4 phosphorylates YAP at T428 within a flexible segment of the transactivation domain. This selectivity is further supported by a recent study showing that nuclear-translocated glucokinase specifically binds to and phosphorylates TAZ but not YAP.46 Collectively, these findings indicate that despite their structural similarity, YAP and TAZ engage distinct binding partners and play nonredundant functional roles in different biological contexts.While the inhibitory regulation of YAP by the canonical Hippo pathway component LATS1 has been well recognized, the potential involvement of other upstream activating regulators has remained largely unexplored. This question is particularly important considering the central role of YAP signaling and the diverse disease contexts in which it is involved. A key finding of our study is the activating phosphorylation of YAP by PAK4, which is independent of LATS1. Specifically, LATS1 phosphorylates YAP at S127, leading to its cytoplasmic sequestration through 14-3-3 binding and subsequent protein degradation.18,28 In contrast, we found that PAK4 neither alters LATS1 expression nor its activity in HSCs, indicating that PAK4 directly affects YAP phosphorylation. Importantly, PAK4 was able to stabilize YAP even under LATS1-deficient conditions, and T428 phosphorylation prevented LATS1-mediated degradation. These findings are notable in light of the role of LATS1 in promoting YAP ubiquitination through β-TRCP.28,41 It is therefore reasonable that we observed that PAK4-dependent stabilization of YAP was associated with β-TRCP while simultaneously competing with the inhibitory action of LATS1. Because β-TRCP also mediates the degradation of β-catenin,47 a known factor of HSC activation,48 our data further suggest that this E3 ligase integrates multiple profibrogenic signals, positioning PAK4-YAP interactions within a broader regulatory framework of hepatic fibrogenesis. Collectively, our results underscore the direct and independent actions of PAK4 in promoting YAP stabilization and counteracting LATS1. Additionally, they highlight a fine-tuned regulatory balance between the activation and inhibition of YAP that can be exploited therapeutically. By revealing a mechanism through which PAK4 selectively promotes YAP activity while bypassing LATS1 activity, our findings provide new conceptual insight into the complexity of Hippo signaling and suggest opportunities for targeted intervention.Our series of studies revealed the dynamic regulation of PAK4 expression under diverse pathophysiological conditions. In the present study, we newly uncovered that PAK4 is transcriptionally upregulated by two transcription factors, HIF-1α and YY1, during HSC activation. HIF-1α promotes HSC activation by regulating glycolytic metabolism and by promoting interleukin-6 secretion.49,50 YY1 has also been implicated in the fibrogenic properties of HSCs.51 It has been suggested that hypoxia and Hedgehog signaling contribute to HIF-1α induction49,52 and that transforming growth factor-β1 increases YY1 expression in HSCs,53 although the upstream mechanisms responsible for activating these pathways require further investigation. In addition, the key downstream effectors of these transcription factors remain largely undefined. Our findings expand our knowledge about the critical roles of HIF-1α and YY1 in the control of HSC fate and hepatic fibrogenesis. Since we previously showed that PAK4 levels in hepatocytes and adipocytes are posttranscriptionally modulated by ubiquitin-mediated proteasomal degradation according to nutrient status,30,31 it is highly likely that the regulation of PAK4 expression is context dependent and varies across different pathophysiological conditions.The activation of HSCs is influenced by neighboring cells, including hepatocytes.54,55 Injured hepatocytes exhibit altered secretory profiles that facilitate profibrogenic HSC activation.55 In particular, oxidative damage to hepatocytes plays a crucial role in triggering HSC activation.55 In our previous studies, we demonstrated that inhibition of PAK4 in hepatocytes protects against oxidative injury by activating nuclear factor erythroid 2-related factor 2 and alleviating lipotoxicity by enhancing peroxisome proliferator-activated receptor alpha.31,34 Thus, the antifibrotic effects of pharmacological inhibition of PAK4 may be attributed to its combined effects on both HSCs and hepatocytes, offering additional therapeutic advantages.There has been growing interest in developing therapeutic agents that modulate Hippo signaling. Considerable efforts have focused on disrupting the YAP-TEAD interaction or inhibiting TEAD palmitoylation.18 However, despite these advances, targeting the Hippo pathway remains challenging for clinical translation, with concerns including TEAD inhibition-associated renal toxicity and the risk that complete inhibition of YAP and TAZ could compromise essential physiological functions.56,57 These limitations underscore the need for selective, context-dependent regulation. In this regard, our study identifies a unique PAK4-YAP axis as a promising alternative target. We demonstrate that PAK4 selectively targets YAP without binding to TAZ, providing a novel means of modulating YAP signaling. This specificity may minimize the risk of adverse effects, highlighting the therapeutic relevance of PAK4 inhibition as a safer upstream regulatory strategy. Notably, our recent findings that PAK4 inhibition confers renal protection58 further mitigate safety concerns, reinforcing its translational potential.The therapeutic value of targeting PAK4 is broadened by its favorable metabolic profile. Given that the pathogenesis of MASH and its progression to end-stage liver diseases with poor outcomes are closely linked to cardiometabolic complications,59 it is also noteworthy that PAK4 inhibition also has antiobesity and antidiabetic effects.30,32,60 Such pleiotropic benefits could expand the clinical impact of PAK4-targeted therapy. Furthermore, since both PAK4 and YAP are elevated in hepatocellular carcinoma,61,62 targeting this axis may not only suppress fibrosis progression but also contribute to preventing tumorigenesis, positioning PAK4 as a multifunctional therapeutic target across diverse liver pathologies.Several important avenues arise from our study that merit further exploration. YAP is subject to diverse posttranslational modifications, including acetylation and lactylation,28 and is sensitive to environmental cues such as mechanical stress.18,22 Whether T428 phosphorylation by PAK4 integrates with or counterbalances these regulatory layers remains an intriguing question for future studies. While HSCs are best known for their pathological fibrogenic activity, their emerging homeostatic functions13,38 suggest the need to comprehensively characterize the PAK4-YAP axis in the context of both physiological and pathological conditions. Moreover, despite ongoing efforts to develop PAK4 inhibitors for clinical application, limitations, such as poor bioavailability and potential off-target effects, still exist.63 Therefore, the development of new scaffold small molecules targeting PAK4 and approaches other innovative modalities, including targeted protein degradation, is highly important given the pathological induction of PAK4.In summary, we identify PAK4 as a critical regulator that is upregulated during transcriptional reprogramming in fibrotic HSCs, where it drives HSC activation and liver fibrogenesis through T428 phosphorylation-mediated stabilization and transactivation of YAP. These findings offer novel perspectives for developing therapeutic interventions targeting HSC activation in liver fibrosis and provide broader clinical implications for other YAP-related pathological conditions.Materials and methodsMouse modelsAnimal experiments were approved by the Institutional Animal Care and Use Committee of Jeonbuk National University (No. NON2023-099) and conducted in accordance with institutional guidelines. Wild-type (WT) C57BL/6N and C57BL/6J male mice were provided by Samtako Bio Korea (Osan, Korea) and the Jackson Laboratory (Bar Harbor, ME, USA), respectively. Pak4-floxed mice (B6.129S2(FVB)-Pak4tm2.1Amin/J, Strain No. 015828) and Pdgfr-β-CreERT2 mice (B6. Cg-Tg(Pdgfrb-cre/ERT2)6096Rha/J, Strain No. 029684) were obtained from Jackson Laboratory. Hepatic stellate cell (HSC)-specific Pak4 knockout (KO) (Pak4ΔHSC) mice were generated by crossing Pak4flox/flox mice with Pdgfr-β-CreERT2 mice. Pak4flox/flox littermates (LT) served as controls. Genotypes were confirmed via PCR using specific primers (Supplementary Table 1). The mice were housed under controlled barrier conditions (22 ± 2 °C, 50 ± 5% humidity, 12-h light/dark cycle) with ad libitum access to food and water. The animals were randomly assigned to groups according to their genotype. To activate Cre-ERT2 and induce HSC-specific deletion of the Pak4 gene, tamoxifen (#T5648, Sigma‒Aldrich, Merck KGaA, Darmstadt, Germany) was administered intraperitoneally at 75 mg/kg per day for 5 consecutive days.Liver fibrosis was induced in eight-week-old male mice by intraperitoneal injection of vehicle (corn oil) or carbon tetrachloride (CCl4, 0.6 mL/kg) twice weekly for six weeks. In a surgical model, eight-week-old male mice underwent either a sham operation or bile duct ligation (BDL) surgery as previously described;64 the mice were subsequently sacrificed two weeks after surgery. For dietary induction, six-week-old C57BL/6N male mice were fed either a normal chow diet (NCD) or a choline-deficient, L-amino acid-defined, high-fat diet containing 60 kcal% fat and 0.1% methionine without added choline (CDAHFD) (#A06071302, Research Diets, New Brunswick, NJ, USA) for 12 weeks.Human tissue samplesThis human study was approved by the Institutional Review Board of Jeonbuk National University Hospital (IRB No. JUH 2023-11-024-001) and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants. Human liver tissue samples were collected from patients who underwent liver resection, liver transplantation, or liver biopsy at Jeonbuk National University Hospital between November 2017 and May 2023. The samples were classified according to the degree of hepatic fibrosis (normal liver, fibrosis, and cirrhosis) on the basis of histologic METAVIR staging as reported in the original pathology reports. Underlying etiologies predominantly included chronic hepatitis B virus infection, with additional cases of hepatitis C virus infection, metabolic dysfunction-associated steatotic liver disease, and alcohol-associated liver disease. All cirrhotic patients had compensated cirrhosis. Baseline complete blood count and liver chemistry parameters were included (Supplementary Table 1). Biopsies were either immediately frozen in liquid nitrogen and stored at −80 °C or fixed in formalin and embedded in paraffin. These specimens were analyzed for histopathological and immunohistochemical features, as well as protein expression across fibrosis stages.Cell cultureLX-2 cells (an immortalized human HSC line) were kindly provided by Dr. S.L. Friedman (Icahn School of Medicine at Mount Sinai, NY, USA). Primary mouse HSCs were isolated as previously described.65 Briefly, HSCs were obtained by collagenase type IV perfusion followed by Percoll gradient centrifugation. The cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a humidified atmosphere containing 5% CO2.Generation of p-T428-YAP antibodiesA custom polyclonal antibody specific for YAP phosphorylated at threonine 428 (p-T428-YAP) was produced by GW Vitek (Seoul, Korea). A synthetic phosphopeptide encompassing the T428 residue of human YAP (sequence: EMDTGD[pT]INQSTL) was synthesized and conjugated to keyhole limpet hemocyanin (KLH). The rabbits were immunized according to the company’s standard immunization protocol, which included primary and booster injections, followed by the collection of antisera. Antibody titers were monitored by ELISA against phosphorylated and nonphosphorylated peptides. For purification, the serum samples were first processed by protein A/G chromatography, followed by antigen-specific affinity purification using the phosphorylated peptide. The purified antibody was validated by ELISA and peptide competition assays to confirm its specificity for the phosphorylation of YAP at T428.Lentiviral delivery of the YapT413 mutantTo overexpress Flag-tagged WT-Yap and the phospho-inactive YapT413A mutant (the mouse Yap T413 site corresponding to the human YAP T428 residue; YapT413A: ACC → GCC) in HSCs, lentiviruses (LVs) were generated by Vector Builder (Guangzhou, China). Each LV carried either WT-Yap or the YapT413A mutant under the control of the Col1a1 promoter. After 2 weeks of CCl4 administration, the mice received retro-orbital intravenous injections of control LV, LV-Yap, and LV-YapT413A (1 ×108 TU/mouse), after which the CCl4 treatment was continued for an additional 4 weeks.StatisticsData are expressed as the mean ± standard error of the mean (SEM). The n values shown as individual dots in the figures represent biological replicates. All the results were obtained from at least three independent samples or experiments. Differences between two groups were assessed using a two-tailed unpaired Student’s t test. For comparisons among multiple groups, one-way or two-way ANOVA was applied, followed by Tukey’s post hoc test. Statistical significance was defined as P