Epithelial MLCK deficiency alleviates alcohol-associated liver disease via dendritic–Th17 cell axis Authors: Xianda Wang, Juan Wang, Lei Xiong, Sainan Tang, Shuang Wu, Junhao Zhao, Zihan Wang, Jiayi Zhu, Yaqun Li, Jialu Zhuang, Junjie Wu, Jinruo Zhang, Ziqi Zhou, Jianshang Huang, Haiyuan Shen, Miaomiao Wu, Lina Qi, Wei He, Ruolei Hu, Jerrold R Turner, Yongping Huang, Hua Wang, Li Zuo Published: 2026 ABSTRACT AbstractBackground & AimsIntestinal dendritic cells (DCs) have a key role in immune regulation, especially during injury. The role of intestinal epithelial paracellular permeability in DC antigen presentation and its regulation is unclear although it is known that myosin light chain kinase (MLCK) regulates barrier permeability. Thus, we examined the role of MLCK in DC function and immune responses in alcohol-associated liver disease (ALD).MethodsWe used systemic and intestinal epithelial-specific Mylk-knockout models (n = 6 mice per group) to investigate the MLCK-mediated leak pathway in ALD. The effects of constitutively active MLCK in intestinal epithelial cells were also assessed (n = 6). Mechanistic investigations focused on LPA-LPAR2-Ca2+-dependent MLCK activation, while also evaluating tight junction integrity, DC antigen uptake and presentation by flow cytometry, and pathogenic Th17 cell responses (n = 3).ResultsThe MLCK-mediated leak pathway was upregulated in ALD (n = 6, p <0.05). Both systemic and intestinal epithelial-specific Mylk knockout attenuated ALD progression, as evidenced by reduced liver injury markers (ALT decreased by 50%, n = 6, p <0.05), whereas constitutively active MLCK accelerated disease (n = 6, p <0.05). LPA-LPAR2 signaling activated MLCK via Ca2+/CaM-dependent pathways, disrupted tight junctions, and enhanced DC antigen uptake and antigen-presenting capacity (by 1.4-fold, n = 3, p <0.05). This promoted pathogenic Th17 cell responses and GM-CSF production (by twofold, n = 6, p <0.05) associated with liver inflammation.ConclusionsThis study highlights the role of MLCK-regulated intestinal permeability in DC function and Th17 cell differentiation in ALD. Thus, targeting this pathway could provide new therapeutic strategies for ALD.Impact and implicationsThis study reveals the crucial role of MLCK-regulated intestinal paracellular permeability in ALD, expanding our understanding of the gut–liver interaction. The findings provide new insights into immune regulation and barrier function, offering potential therapeutic implications for the treatment of ALD. By highlighting the role of MLCK-mediated barrier function in DC function and Th17 cell differentiation, this research opens new avenues for clinical application, particularly in developing targeted therapies for ALD. For researchers, these findings offer a fresh perspective on the interplay between the gut microbiota and immune responses, advancing the field of immunology and gastrointestinal diseases.Keywords: ALD, MLCK, Intestinal barrier, DCs, Th17 cells, Gut–liver axis Introduction Alcohol-associated liver disease (ALD) represents a major global health burden.[1], [2], [3] In 2019, ALD was responsible for ∼11 million life-years lost worldwide, with a sharp increase in mortality. Epidemiological studies indicate that the prevalence of ALD is ∼3.5% in the general population, rising to 26.0% in hazardous drinkers and reaching 55.1% among individuals with alcohol use disorders.4 Its severe form, alcohol-associated hepatitis (AH), has a very poor prognosis, with a 30-day mortality rate of 30–50%.5 Current clinical treatments are limited, underscoring the urgent need to explore new therapeutic targets.[6], [7], [8] The gut–liver axis is considered a central driver in the initiation and progression of ALD.9,10 Recent research has revealed that intestinal barrier dysfunction is closely associated with various liver diseases, including alcoholic fatty liver disease, non-alcoholic fatty liver disease (NAFLD), and cirrhosis.11,12 Significantly increased intestinal permeability has been observed in both patients with NAFLD and animal models.13 Similarly, in ALD, ethanol (EtOH) disrupts the intestinal barrier structure and immune homeostasis, leading to the translocation of microbial products and pathogens to the liver, thereby exacerbating liver injury.14,15 Consequently, targeting the intestinal barrier function is recognized as a potential strategy for intervening in the progression of liver disease.16 Tight junctions are crucial structures that regulate paracellular permeability,17,18 thereby maintaining intestinal epithelial integrity and preventing luminal antigens from translocating into the lamina propria.19 However, the precise mechanisms by which epithelial tight junctions modulate the gut–liver axis remain incompletely understood. The myosin light chain kinase (MLCK) family, which comprises serine/threonine-specific protein kinases,20 has a crucial role in various cellular processes. Previous studies demonstrated that calmodulin (CaM) is a key Ca2+ sensor that regulates MLCK activity through Ca2+-dependent binding to its calmodulin-binding regulatory domain. Upon Ca2+ binding to the four EF-hand Ca2+-binding sites of CaM, conformational changes promote the interaction of Ca2+/CaM with the calmodulin-binding domain of MLCK, thereby activating the kinase.21,22 Activated MLCK subsequently phosphorylates the myosin light chain,23 inducing actomyosin contraction that reorganizes tight junction proteins, such as occludin and zonula occludens-1 (ZO-1), ultimately leading to increased epithelial permeability.24 Thus, MLCK serves as a key regulator of intestinal barrier function. Nevertheless, whether the MLCK-mediated leakage pathway contributes to liver disease and the underlying mechanisms have not been fully elucidated. In this study, we demonstrate that gut epithelial MLCK deletion restores alcohol-induced intestinal barrier dysfunction, reduces the antigen-presenting capacity of dendritic cells (DCs), attenuates DC–Th17 cell interactions, and diminishes portal vein-mediated transfer of granulocyte-macrophage colony-stimulating factor (GM-CSF) and IL-17A to the liver, ultimately alleviating ethanol-induced liver injury. Thus, we reveal a novel MLCK-mediated pathway that influences gut–liver crosstalk by regulating DC function, thereby not only deepening the understanding of ALD immunopathology, but also identifying a potential new therapeutic target for this disease. Materials and methods Mice Mylk knockout (MylkKO), Mylkflox/flox (Mylkf/f), and Villin-CA-Mylk (tMylk) Tg (MylkCA) mice were provided by Jerrold R. Turner (Brigham and Women's Hospital, Boston, MA, USA). Villin-Cre, Alb-Cre, and C57BL/6J mice (7–8-week-old) were obtained from GemPharmatech (Nanjing, China).[25], [26], [27] Intestinal epithelial-specific knockout mice (MylkΔIEC) and hepatocyte-specific (MylkΔHep) knockout mice were generated by crossing Mylkflox/flox mice with the respective Cre lines. All mice had a C57BL/6J background, with Cre-negative littermates as controls. All animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Anhui Medical University (approval no. LLSC20252903). For the ALD model, 8–10-week-old male mice were fed a Lieber–DeCarli ethanol liquid diet for 10 days, followed by a single ethanol gavage at a dose of ∼5 g/kg body weight (volume of 20 μl/g body weight using 31.5% [v/v] ethanol).28 Control mice received an isocaloric diet and a maltose-dextrin gavage. Mice were euthanized 9 h post gavage and serum and liver tissues were collected for analysis. Male mice were used to minimize hormonal variability and to ensure experimental consistency across genetic and mechanistic analyses. Additional methods Detailed methods for the following are provided in the supplemental data: intestinal organoid polarity reversal; evaluation of intestinal organoid barrier integrity; confocal immunofluorescence imaging of suspended organoids; single-cell RNA sequencing and bioinformatics analysis; untargeted LC-MS-based fecal metabolomic analysis; serum biomarker detection using Luminex; ELISAs; calpain activity assays; intestinal permeability assays; quantification of serum alanine aminotransferase (ALT)/aspartate aminotransferase (AST), and hepatic triglycerides (TGs); histological scoring of intestinal injury; histopathology and immunohistochemistry (IHC); immunofluorescence (IF) staining; flow cytometry; in vivo antigen uptake by intestinal DCs; measurement of intracellular calcium concentration; Caco-2 cell transfection, Western blot analysis (WB); and real-time quantitative PCR (RT-qPCR). Statistical analysis Data are expressed as the mean ± SEM. Normality was assessed for all the datasets. Between-group comparisons were performed using the unpaired Student's t test (normal distribution) or Mann-Whitney U test (non-normal distribution). Multiple group comparisons were performed using one-way ANOVA with Welch's correction for unequal variances (normal distribution) or the Kruskal-Wallis test (non-normal distribution). Statistical significance was defined as p <0.05. All analyses were performed using the GraphPad Prism (v10.1.1) software (GraphPad Software LLC, Boston, MA, USA). Results MLCK is upregulated in intestine epithelium in ALD To investigate the intestinal barrier function in ALD, we established a Gao-binge mouse model (Fig. S1A–E). Concurrently, alcohol directly disrupted the integrity of the intestinal mucosa, leading to a ‘leaky gut’ phenotype (Fig. 1A). Alcohol caused the most severe duodenal damage (Fig. 1B). Previous results showed that MLCK is a key regulator of the leak pathway.29 Therefore, we measured MLCK expression in the duodenum of ALD mice.30 MLC phosphorylation and MLCK expression were significantly upregulated in the duodenum following alcohol exposure compared with the control group (Fig. 1C–E; Fig. S1F). Moreover, consistent upregulation of phosphorylated myosin light chain (pMLC) and MLCK was also observed in the jejunum, ileum, and colon of alcohol-fed mice (Fig. 1D,E). Notably, MLCK positively correlated with intestinal permeability (Fig. 1F), suggesting that alcohol disrupts intestinal tight junctions and barrier integrity via MLCK activation. Importantly, MLCK expression was positively correlated with serum ALT and AST levels as well as hepatic TG content (Fig. 1F). Collectively, these data indicated that elevated intestinal epithelial MLCK expression is associated with ALD. Upregulation of MLCK in intestinal epithelium contributes to ALD pathogenesis. (A) Serum fluorescein levels indicating intestinal permeability from pair-fed and EtOH-fed mice. (B) Representative H&E images of duodenum, jejunum, ileum, and colon tissues from pair-fed and EtOH-fed mice. (C) WB of pMLC in duodenum from pair-fed and EtOH-fed mice. (D) IF of pMLC and (E) MLCK in duodenum, jejunum, ileum, and colon tissues from pair-fed and EtOH-fed mice. (F) Correlation analysis of Mylk mRNA expression in duodenal tissues with intestinal permeability, serum ALT, serum AST, and liver TG levels from EtOH-fed mice. Data are presented as mean ± SEM; n = 6; analyzed with two-tailed unpaired Student’s t test (A,C–E) and Pearson correlation coefficient (F). Scale bars: 100 μm (B), 20 μm (D,E). ALD, alcohol-associated liver disease; ALT, alanine aminotransferase; AST, aspartate aminotransferase; EtOH, ethanol; IF, immunofluorescence; MLCK, myosin light chain kinase; pMLC, phosphorylated myosin light chain; TG, triglyceride; WB, Western blot. Mylk knockout alleviates EtOH-induced liver injury and gut barrier dysfunction To investigate the potential role of MLCK in ALD progression, systemic MylkKO mice were subjected to alcohol feeding (Fig. S2A). Knockout efficiency was validated by qPCR and IF (Fig. S2B,C). After alcohol feeding, MLCK knockout significantly alleviated alcohol-induced liver injury and ameliorated hepatic steatosis and lipid metabolic dysregulation (Fig. 2A; Fig. S2D-E). In addition, Mylk knockout markedly reduced intrahepatic neutrophil infiltration (Fig. 2A) and expression of inflammatory cytokine-related genes (including Tnfa, Il1b, and Il6) (Fig. 2B). Mylk KO attenuates EtOH-induced liver injury and gut barrier dysfunction. (A) Representative images, and quantification, of H&E, Oil Red O, and MPO staining of liver tissues. (B) RT-qPCR analysis of Tnfa, Il1b, and Il6 in liver and (C) serum fluorescein levels indicating intestinal permeability from mice with or without Mylk knocked down (MylkKOvs. MylkWT, respectively) fed an EtOH diet. (D) Representative images of H&E of duodenum from MylkKO or MylkWT mice fed the pair and EtOH diets. (E) IF of ZO-1, occludin, F4/80, and MPO and (F) RT-qPCR analysis of Tnfa, F4/80, and Ly6g in duodenum from MylkKO or MylkWT mice fed an EtOH diet. Data are presented as mean ± SEM; n = 6; analyzed with two-tailed unpaired Student’s t test (A–C) or one-way ANOVA with post hoc Tukey’s HSD tests (D–F). Scale bars: 100 μm (D), 50 μm (A,E). EtOH, ethanol; IF, immunofluorescence; KO, knockout; MPO, myeloperoxidase; Mylk, myosin light chain kinase; RT-qPCR, real-time quantitative PCR; WT, wild-type; ZO-1 zonula occludens-1. We further examined the effect of Mylk knockout on the intestinal barrier function. Using a fluorescent tracer gavage assay, we found that alcohol feeding significantly increased intestinal permeability, which was attenuated by Mylk knockout (Fig. 2C). H&E staining further confirmed that Mylk knockout ameliorated duodenal tissue damage and improved histology score (Fig. 2D). Moreover, Mylk knockout rescued alcohol-induced downregulation of the tight junction gene Tjp1 (Fig. S2F) and effectively increased the protein expression of ZO-1 and occludin in the duodenum (Fig. 2E). Interestingly, alcohol feeding led to elevated expression of F4/80 and myeloperoxidase (MPO) in the duodenum, which was mitigated by Mylk knockout (Fig. 2E,F). These data further suggest that Mylk contributes to ALD progression by impairing the intestinal barrier function, and its deletion alleviates duodenal injury and inflammation, thereby attenuating alcohol-induced liver injury. Mylk deficiency in IECs but not in hepatocytes alleviates EtOH-induced liver injury of mice To avoid potential artificial effects associated with whole-body knockout, we used MylkΔIEC and established a Gao-binge model in these mice (Mylkfl/fl; Villin-cre) (Fig. S3A–D). Consistent with the observations in whole-body knockout mice, MylkΔIEC mice exhibited alleviated liver injury and steatosis after EtOH feeding compared with Mylkf/f controls. This was evidenced by reduced plasma ALT and AST levels, lower hepatic TG content, and improved histopathology in both H&E-stained and Oil Red O-stained liver sections (Fig. 3A,B; Fig. S3E). Furthermore, mRNA expression of hepatic inflammatory genes (Tnfa, Il1b, and Il6) was significantly downregulated in MylkΔIEC mice (Fig. 3C), along with diminished MPO-positive cell infiltration in the liver (Fig. 3B). To determine whether MLCK influenced alcohol metabolism, we measured key alcohol-metabolizing enzymes. The results showed comparable expression levels of cytochrome P450 2E1 (Cyp2E1), alcohol dehydrogenase (Adh)-1, and Adh2 in EtOH-fed MylkΔIEC and Mylkf/f mice (Fig. S3F). Mylk deficiency in IECs attenuates EtOH-induced liver injury in mice. (A) Serum ALT, AST, and hepatic TG levels. (B) Representative images, and quantification, of H&E, Oil Red O, and MPO staining of liver tissues. (C) RT-qPCR analysis of Tnfa, Il1b, and Il6 in liver. (D) Representative images of H&E and IF staining of ZO-1, occludin, F4/80, and MPO in duodenum and histology score. (E) RT-qPCR analysis of Tnfa, F4/80, Ly6g, and Tjp1 in duodenum. (F) Serum fluorescein levels indicating intestinal permeability. Data are presented as mean ± SEM; n = 6; analyzed with two-tailed unpaired Student’s t test. Scale bars: 100 μm (D), 50 μm (B,D). ALT, alanine aminotransferase; AST, aspartate aminotransferase; EtOH, ethanol; IEC, intestinal epithelial cell; IF, immunofluorescence; MPO, myeloperoxidase; Mylk, myosin light chain kinase; RT-qPCR, real-time quantitative PCR; TG, triglyceride; ZO-1, zonula occludens-1. Notably, alcohol-induced villous injury was markedly attenuated in MylkΔIEC mice, which displayed a well-preserved intestinal villus architecture (Fig. 3D). Assessment of tight junction proteins revealed that the loss of ZO-1 and occludin in the intestinal epithelium was significantly reduced in MylkΔIEC mice compared with Mylkf/f controls, indicating enhanced barrier function (Fig. 3D; Fig. S3G). Concurrently, MylkΔIEC also suppressed EtOH-induced upregulation of F4/80 and MPO in the duodenum (Fig. 3D) and reduced the expression of inflammatory genes (Tnfa, F4/80, and Ly6g) (Fig. 3E). Furthermore, compared with Mylkf/f mice, MylkΔIEC mice exhibited upregulated Tjp1 expression and decreased permeability, which alleviated damage to the intestinal barrier (Fig. 3E,F). Notably, hepatocyte-specific Mylk knockout (Mylkfl/fl; Alb-cre) failed to ameliorate AH pathology (Fig. S4A–D). Taken together, these findings indicate that deleting Mylk from intestinal epithelial cells protects against alcohol-induced increase in intestinal permeability. Constitutively active Mylk in IECs exacerbates ALD To further validate the pathogenic role of intestinal MLCK in alcohol-related injury, we systematically evaluated intestinal barrier function and liver pathological changes in intestinal epithelial-specific constitutively activated mice (MylkCA) (Fig. S5A-D). In contrast to the aforementioned knockout experiments, intestinal epithelial-specific activation of Mylk significantly exacerbated alcohol-induced intestinal barrier dysfunction, as evidenced by the marked increase in intestinal permeability (Fig. 4A). At the molecular level, the expression of inflammatory genes (Tnfa, F4/80, and Ly6g) in duodenal tissues was significantly upregulated, whereas that of Tjp1, encoding the tight junction protein, was downregulated (Fig. 4B), suggesting that sustained activation of Mylk aggravated intestinal inflammation and disrupted epithelial barrier integrity. Histological analysis further confirmed that the MylkCA alcohol-fed group exhibited more severe inflammatory cell infiltration and structural mucosal damage in the duodenum (Fig. 4C). Consistent results from IF and WB analyses showed significantly reduced expression of the tight junction proteins ZO-1 and occludin in the intestines of this group, along with markedly increased expression of the macrophage marker F4/80 and the neutrophil marker MPO (Fig. 4C and Fig. S5E). Constitutively active Mylk in IECs exacerbates ALD. (A) Serum fluorescein levels indicating intestinal permeability. (B) RT-qPCR analysis of Tnfa, F4/80, Ly6g, and Tjp1 in duodenum. (C) Representative images of H&E and IF staining of ZO-1, occludin, F4/80, and MPO in duodenum and histology score. (D) Serum ALT, AST, and hepatic TG levels, body-weight changes and liver-to-body weight ratios. (E) Representative images, and quantification, of H&E, Oil Red O, and MPO staining of liver tissues (n = 6). (F) RT-qPCR analysis of Tnfa, Il1b, and Il6 in liver. Data are presented as mean ± SEM; n = 6; analyzed with two-tailed unpaired Student’s t test. Scale bars: 100 μm (C), 50 μm (C,E). ALD, alcohol-associated liver disease; ALT, alanine aminotransferase; AST, aspartate aminotransferase; IEC, intestinal epithelial cell; IF, immunofluorescence; MPO, myeloperoxidase; Mylk, myosin light chain kinase; RT-qPCR, real-time quantitative PCR; TG, triglyceride; ZO-1, zonula occludens-1. MylkCA alcohol-fed mice exhibited more severe liver injury, with significantly elevated serum ALT, AST, and hepatic TG levels (Fig. 4D). An increase in the severity of hepatic steatosis was observed in the MylkCA group compared with the control group (Fig. 4E). Furthermore, alcohol exposure increased overall body weight and liver-to-body weight ratio (Fig. 4D), confirming that the intestinal epithelial-specific activation of MLCK exacerbates systemic metabolic disturbances and liver pathology. Consistent with these findings, neutrophil infiltration was markedly aggravated, and the mRNA expression of liver inflammation-related genes (Tnfa, Il1b, and Il-6) was also significantly upregulated (Fig. 4F). Taken together, these results lend support to the opposite perspective, validating the crucial role of MLCK in the development of ALD. LPA-LPAR2-Ca2+/CaM signaling mediates MLCK activation and intestinal barrier disruption To elucidate the upstream mechanisms by which EtOH disrupts the intestinal barrier function, we performed an untargeted metabolomic analysis of fecal samples. Principal component analysis (PCA) clearly distinguished between EtOH-fed and control mice (Fig. 5A), indicating significant remodeling of the gut metabolic landscape upon ethanol exposure. Metabolite set enrichment analysis (MSEA) revealed that glycerophospholipid metabolism, a pathway strongly associated with intestinal inflammation, was among the most significantly altered metabolic pathways (Fig. S6A), which were strongly associated with intestinal inflammation. Interestingly, we observed a marked increase in the level of lysophosphatidic acid (LPA) in the EtOH-treated group (Fig. 5B,C). Given that LPA is a known regulator of cytoskeletal reorganization and calcium mobilization,31 we hypothesized that EtOH-induced LPA production serves as a crucial upstream signal that triggers MLCK-mediated barrier disruption. LPA–LPAR2–Ca2+/CaM signaling mediates MLCK activation and intestinal barrier disruption. (A) PCA plot of untargeted metabolomics from fecal samples of each group of mice. (B) Heat maps of differential metabolin expression. (C) Response intensity of LPA in pair-fed and EtOH-fed mice. (D) RT-qPCR analysis of LPAR2 in Caco-2 cells. (E) Flow cytometric measurement of intracellular Ca2+ using Fluo-4 AM in LPA-treated and LPAR2-silenced Caco-2 cells. (F) Calpain activity in Caco-2 cells after LPA stimulation and LPAR2 silencing. (G) Schematic of the apical-out intestinal organoid model for investigating luminal–epithelial interactions. (H) Representative images of organoids exhibiting apical-out polarity following suspension culture. (I) Standardized FITC-dextran permeability assay for quantitative assessment of paracellular barrier function. (J) Quantitative results demonstrating LPA-induced barrier dysfunction in organoids. (K–N) Immunofluorescence staining of ZO-1 (vehicle n = 10; LPA n = 5), occludin (vehicle n = 10; LPA n = 7), MLCK (vehicle n = 8; LPA n = 7), and pMLC (vehicle n = 6; LPA n = 11) in organoids after LPA treatment. Data are presented as mean ± SEM; n = 3 (F,J), 5 (C), or 6 (D,E); analyzed with two-tailed unpaired Student’s t test (C,D,J–N) or one-way ANOVA with post hoc Tukey’s HSD tests. Scale bars: 20 μm (J,K–N). CaM, calmodulin; EtOH, ethanol; LPA, lysophosphatidic acid; LPAR, lysophosphatidic acid receptor; MLCK, myosin light chain kinase; PCA, principal component analysis; pMLC, phosphorylated myosin light chain; RT-qPCR, real-time quantitative PCR; ZO-1, zonula occludens-1. LPA signals primarily through G protein-coupled receptors (LPARs).32 To identify which LPARs mediate this effect, we examined the major receptors reportedly expressed in intestinal epithelial cells (LPAR1–3). RT-qPCR analysis showed that LPA selectively upregulated LPAR2 expression (Fig. 5D), whereas LPAR1 and LPAR3 showed no significant changes (Fig. S6B). To determine the functional requirement of LPAR2, we performed short interfering (si)RNA-mediated knockdown, and silencing efficiency was confirmed by RT-qPCR (Fig. S6C). Flow cytometric analysis with Fluo-4 AM showed that LPA stimulation induced a marked increase in intracellular Ca2+ levels, which was substantially attenuated after LPAR2 knockdown (Fig. 5E). Accordingly, the Ca2+-bound form of CaM was enhanced by LPA and suppressed by LPAR2 knockdown (Fig. 5F). These results indicate that LPA activates Ca2+/CaM signaling through LPAR2. To test whether LPA is sufficient to disrupt barrier function, we used apical-out murine intestinal organoids (Fig. 5G), a physiologically relevant model that recapitulates the luminal–epithelial interface for mechanistic studies. As early as 1 day after being placed in suspension culture, organoids everted to exhibit apical polarity (Fig. 5H). Meanwhile, we established a FITC-dextran permeability assay to quantitatively assess paracellular flux dynamics (Fig. 5I). LPA treatment significantly increased FITC-dextran penetration, indicating impaired barrier function (Fig. 5J). Immunofluorescence staining revealed that LPA-treated organoids exhibited significant downregulation of ZO-1 and occludin compared with untreated controls (Fig. 5K,L), with concomitant upregulation of MLCK expression and enhanced MLC phosphorylation (Fig. 5M,N). Together, these findings support a mechanistic pathway in which alcohol-induced LPA accumulation acts through LPAR2-dependent Ca2+/CaM signaling, leading to MLCK activation and tight junction disruption, thereby driving intestinal barrier dysfunction. scRNA-seq reveals crosstalk between epithelial cells and dendritic cells To further investigate the mechanism by which Mylk knockout protects against ALD, we performed single-cell RNA sequencing of duodenal tissues from EtOH-fed MylkWT and MylkKO mice. UMAP dimensionality reduction identified 12 distinct cell clusters (Fig. 6A; Fig. S7A). Feature plot analysis showed that Mylk was expressed across most intestinal cell populations, predominantly in epithelial cells, and its expression increased overall after alcohol fed (Fig. 6B). Although the proportions of major intestinal cell subsets remained largely unchanged (Fig. S7B), the most pronounced shifts in differential abundance were observed within the myeloid cluster (Fig. 6C). To further dissect these alterations, we performed pathway enrichment analysis on each myeloid subpopulation. This identified functional dysregulation in DCs (Fig. S7C,D), characterized by downregulation of Rap1 signaling and FcγR-mediated phagocytosis pathways essential for antigen cross-presentation and DC–T cell adhesion (Fig. 6D; Fig. S7E,F).[33], [34], [35], [36], [37] Specifically, the Rap1 signaling and FcγR-mediated phagocytosis pathways were significantly downregulated upon Mylk knockout (Fig. 6D). Single cell RNA sequencing delineates an active signaling network between epithelial cells and DCs. (A) UMAP plot of duodenal single-cell transcriptomes from EtOH-MylkWT and EtOH-MylkKO mice showing major cell clusters. (B) Feature plot showing Mylk expression across intestinal cell populations. (C) Summary of differentially expressed genes across duodenal cell lineages. (D) KEGG enrichment analysis of downregulated pathways in DC cluster. (E) Flow cytometry of CD11c+MHC-II+ DCs from MylkWT and MylkKO mice fed the pair or EtOH diet. (F) Flow cytometry of CD11c+ MHC II+ DCs from EtOH-MylkΔIEC and EtOH-MylkCA mice. (G) Serum ALT and AST levels. (H) Hepatic TG levels. (I) Representative images, and quantification, of H&E, Oil Red O, and MPO staining of liver tissues (n = 6). Scale bar: 50 μm. Data are presented as mean ± SEM; n = 3 (E,F) or n = 6 (G,H); analyzed with two-tailed unpaired Student’s t test (F) or one-way ANOVA with post hoc Tukey’s HSD tests (E,G–I). ALT, alanine aminotransferase; AST, aspartate aminotransferase; DC, dendritic cell; EtOH, ethanol; KO, knockout; MPO, myeloperoxidase; Mylk, myosin light chain kinase; MylkΔIEC, intestinal epithelial cell-specific knockout; MylkCA, intestinal epithelial-specific constitutively activated; TG, triglyceride; WT, wild-type. To further characterize the DC phenotype, we conducted flow cytometry analysis of mouse duodenal tissues (Fig. S7G) and found that alcohol stimulation markedly increased the expression of MHC-II (Fig. 6E), which is crucial for antigen presentation. Importantly, intestine-specific knockout of Mylk reduced MHC-II expression (Fig. 6F), whereas intestine-specific overexpression of Mylk enhanced the antigen-presenting ability of DCs (Fig. 6F). These results suggest that Mylk deletion attenuates alcohol-induced DC activation. To determine whether DCs are required for the hepatotoxicity induced by epithelial MLCK, we performed DC depletion using anti-CD11c monoclonal antibodies in wild-type (WT) and MylkCA mice (Fig. S8A). Depletion efficiency was confirmed by flow cytometry, showing >80% reduction of CD11c+MHC-II+ cells in intestinal lamina propria (Fig. S8B). In MylkWT mice, DC depletion significantly ameliorated alcohol-induced liver injury, as evidenced by decreased serum ALT and AST levels (Fig. 6G), and reduced hepatic steatosis and injury (Fig. 6H,I). Strikingly, DC depletion also reduced the liver injury observed in MylkCA mice (Fig. 6G–I). Together, these findings indicate that the pathogenic effects of epithelial MLCK hyperactivation are mediated largely through DCs, placing DCs as essential downstream effectors of MLCK-driven epithelial-immune crosstalk. MLCK aggravates ALD by orchestrating DC-mediated Th17 cell activation To further demonstrate that the disruption of the barrier function enhances the antigen capture by DCs, we conducted an in vivo antigen-tracking experiment. By orally administering fluorescently labeled OVA antigen (OVA-AF647), we directly evaluated the antigen uptake ability of DCs in vivo. Flow cytometry analysis showed that the proportion of OVA+ DCs in the intestinal DCs of EtOH-fed mice was significantly higher than that of the control group (Fig. 7A). MLCK disruption impairs the proinflammatory DC–Th17 cell axis in the intestine, thereby ameliorating ALD. (A) Flow cytometric analysis of OVA–Alexa Fluor 647 uptake by intestinal DCs in pair-fed and EtOH-fed mice. (B) Interaction strength between DCs and CD4+ T cells in EtOH-MylkWT and EtOH-MylkKO mice. (C) Cell–cell communication analysis of ligand–receptor co-expression between DC and Th17 cells in EtOH-MylkWT and EtOH-MylkKO mice. (D) RT-qPCR of Plcb2 and Mavs mRNA expression in duodenum. (E–J) ELISA of GM-CSF and IL-17A in duodenum, serum, and liver from (E–G) EtOH-MylkΔIEC mice compared with EtOH-Mylkf/f mice and (H–J) EtOH-MylkWT and EtOH-MylkCA mice. (K) Serum levels of GM-CSF and IL-17A in samples from patients with HCC and ALD (HCC n = 16; ALD n = 54). Data are presented as mean ± SEM; n = 6 (D–J) analyzed with two-tailed unpaired Student’s t test (A,D–K). ALD, alcohol-associated liver disease; DC, dendritic cell; EtOH, ethanol; GM-CSF, granulocyte-macrophage colony-stimulating factor; HCC, hepatocellular carcinoma; KO, knockout; MLCK, myosin light chain kinase; MylkCA, intestinal epithelial-specific constitutively activated; Mylkf/f, Mylkflox/flox; MylkΔIEC, intestinal epithelial cell-specific knockout; RT-qPCR real-time quantitative PCR; WT, wild-type. To further investigate how enhanced antigen presentation by DCs affects intestinal immunity and the development of ALD, we performed a bioinformatics analysis on DCs. Cell communication analysis revealed that alcohol stimulation strengthened the interaction between DCs and Th17 cells, whereas Mylk knockout significantly decreased this interaction (Fig. 7B; Fig. S9A,B). In addition, Mylk deficiency attenuated the interaction between major histocompatibility complex class II (H2) molecules on DCs and CD4 molecules on Th17 cells (Fig. 7C). KEGG pathway enrichment analysis revealed that the downregulated genes in Th17 cells were significantly enriched in multiple immune- and inflammation-related pathways (Fig. 7D; Fig. S9C), which are closely associated with Th17 cell function,[38], [39], [40], [41], [42] consistent with in vivo findings (Fig. 2, Fig. 3). Previous studies reported that DCs can drive the conversion of Th17 cells into pathogenic Th17 cells,[43], [44], [45], [46] characterized by the production of large amounts of cytokines, such as GM-CSF and IL-17A, thereby promoting disease progression.47,48 To directly validate Th17 cell functional polarization in vivo, we performed flow cytometry analysis of intestinal CD4+ T cells and found that alcohol exposure increased the proportion of pathogenic Th17 cells (IL-17A+GM-CSF+), while reducing nonpathogenic Th17 cells (IL-17A+IL-10+), and that this shift was reversed by MLCK disruption (Fig. S9D–F). We then used qPCR to validate the expression levels of Csf2 (encoding GM-CSF) and Il17a in liver and duodenal tissues from each group of mice (Fig. S9G,H). We also measured GM-CSF and IL-17A levels in different tissues using ELISA. Intestine-specific knockout of Mylk reduced the expression of GM-CSF and IL-17A in duodenal tissues (Fig. 7E), portal vein blood (Fig. 7F), and liver tissues (Fig. 7G) compared with the alcohol-fed group, whereas constitutive activation of Mylk enhanced their expression (Fig. 7H–J). Notably, GM-CSF and IL-17A were also upregulated in samples from patients with ALD (Fig. 7K). These findings suggest that GM-CSF and IL-17A produced in the intestine reach the liver via the portal circulation, contributing to alcohol-induced liver injury. In summary, our data demonstrate that Mylk knockout attenuates DC–Th17 cell interaction and reduces pathogenic Th17 cell polarization. Together with our findings that MLCK disruption enhances intestinal barrier integrity (Fig. 2, Fig. 3, Fig. 4), these results support a model in which epithelial MLCK drives ALD by facilitating DC-mediated Th17 cell activation downstream of tight junction dysfunction. Discussion In this study, we identified intestinal epithelial MLCK as a central regulator that integrates barrier integrity and immune activation in ALD. Using systemic and epithelial-specific genetic models, we showed that alcohol exposure markedly upregulated MLCK expression in the duodenal epithelium, leading to tight junction disruption, increased intestinal permeability, and aggravated hepatic injury. Mylk deficiency effectively prevented these pathological changes, whereas the intestinal epithelial-specific constitutive activation of Mylk exacerbated steatosis, inflammation, and neutrophil infiltration. These findings provide in vivo evidence that epithelial MLCK activation is both necessary and sufficient to drive gut barrier failure and liver injury in ALD. Although Chen et al. identified MLCK as a mediator linking intestinal inflammation to ALD, their work relied on a whole-body Mylk-knockout model.49 The present study advances this concept by specifically manipulating MLCK expression in IECs, providing direct evidence that epithelial MLCK activation alone is sufficient to drive alcohol-induced gut barrier disruption and liver injury. We demonstrated that alcohol exposure upregulated MLCK expression in the intestinal epithelium, particularly in the duodenum, which correlated with increased intestinal permeability and liver injury markers. Importantly, using an apical-out intestinal organoid model, we observed that MLCK translocated to the tight junction region upon LPA stimulation, a localization pattern precisely required for its barrier-regulating function.50 Under basal conditions, although MLCK is expressed, it remains largely cytoplasmic and does not affect permeability. Alcohol and its metabolites, such as acetaldehyde, disrupt mucosal integrity by impairing tight junctions, leading to a ‘leaky gut’.51 In ALD, the tight junction proteins occludin and ZO-1 are markedly reduced, suggesting that barrier dysfunction represents one of the earliest pathological events. Using systemic and IEC-specific Mylk-knockout models, we confirmed that MLCK deficiency protected against alcohol-induced gut barrier dysfunction and liver injury, whereas the intestinal epithelial-specific constitutive activation of Mylk exacerbated steatosis, inflammation, and neutrophil infiltration. These findings firmly established intestinal MLCK as a key regulator of ALD pathogenesis. Our findings identified LPA as a key metabolic mediator linking alcohol-induced metabolic remodeling to intestinal barrier dysfunction. Yu et al. demonstrated that dysregulated LPA signaling compromises intestinal epithelial integrity by disrupting tight junction organization and promoting barrier leakage.52 Building upon this framework, our study extends the role of LPA to alcohol-associated contexts, revealing that LPA directly activated the MLCK/MLC pathway in intestinal epithelial cells. Mechanistically, we further demonstrated that this effect was mediated through the LPA-LPAR2-Ca2+/CaM signaling cascade, because LPA stimulation induced rapid intracellular Ca2+ elevation, whereas LPAR2 knockdown markedly attenuated Ca2+ responses and downstream MLCK activation. This activation induces cytoskeletal contraction and ZO-1-associated tight junction disassembly, ultimately resulting in increased paracellular permeability. These observations not only provide a mechanistic bridge connecting alcohol-induced metabolic changes with epithelial barrier breakdown but also suggest that LPA–MLCK signaling represents a convergent axis through which diverse metabolic insults, including alcohol exposure, impair gut–liver homeostasis. A key conceptual advancement of this work lies in defining how epithelial barrier disruption reshapes intestinal innate immunity. Through single-cell transcriptomics and functional validation, we revealed that MLCK-dependent tight junction disruption enhanced DC access to luminal antigens. To directly test this, we performed in vivo antigen tracking using OVA-AF647, which confirmed increased antigen uptake by DCs in alcohol-fed mice. Furthermore, we demonstrated that the enhanced DC–Th17 cell interaction induced by alcohol crucially depended on MLCK activity. A more rigorous analysis of Th17 heterogeneity revealed that alcohol feeding preferentially expanded pathogenic (IL-17A+GM-CSF+) Th17 cells, while reducing the nonpathogenic (IL-17A+IL-10+) subset, both in an MLCK-dependent manner. Mylk knockout markedly attenuated this crosstalk and reduced portal transport of GM-CSF and IL-17A to the liver, thereby mitigating hepatic inflammation. These findings place MLCK at the intersection of epithelial integrity and adaptive immune polarization, illustrating how local barrier failure translates into systemic inflammatory responses via DC-mediated T cell instruction. However, several questions remain. For instance, although we demonstrated that Th17-derived cytokines can directly translocate to the liver via the portal circulation, we also observed that alcohol feeding upregulated IL-17 and GM-CSF transcription within the liver itself. This suggests an additional, indirect pathway: gut-derived microbial products that penetrate the compromised barrier might activate hepatic innate immune cells to locally upregulate Th17-associated cytokines.50 Zeng et al. demonstrated that alcohol induces the migration of Candida albicans-specific Th17 cells from the intestine to the liver, where they exacerbate ALD via the IL-17–Kupffer cell axis.53 In the setting of MLCK deletion, improved barrier function reduces translocation of such microbial ligands, thereby attenuating hepatic cytokine gene expression. Second, whether other barrier-disrupting metabolites or bacterial components similarly affect DC function through MLCK warrants further investigation. The potential of MLCK inhibition as a therapeutic approach for human ALD also merits clinical evaluation. In conclusion, our findings established MLCK as a central regulator of gut–liver communication that coordinates epithelial barrier integrity with DC-mediated immunity. By controlling antigen access and DC activation, MLCK shapes the intestinal immune landscape and determines the severity of alcohol-induced liver injury. Importantly, we identified LPA as a crucial upstream signal that activates MLCK and regulates intestinal permeability in response to alcohol. Our data further define a complete LPA–LPAR2–Ca2+–MLCK signaling axis and demonstrate its functional consequences through enhanced DC antigen uptake, increased T cell priming capacity, and skewing toward pathogenic Th17 cell responses. Thus, targeting the epithelial–immune axis, including LPA–MLCK signaling, could offer novel therapeutic opportunities for ALD and other barrier-related inflammatory disorders. Abbreviations ADH, alcohol dehydrogenase; AH, alcoholic hepatitis; ALD, alcohol-associated liver disease; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CaM, calmodulin; CYP2E1, cytochrome P450 2E1; DC, dendritic cell; EtOH, ethanol; GM-CSF, granulocyte-macrophage colony-stimulating factor; ICH, immunohistochemistry; IEC, intestinal epithelial cell; HCC, hepatocellular carcinoma; IF, immunofluorescence; KO, knockout; LPA, lysophosphatidic acid; LPAR, lysophosphatidic acid receptor; MLCK, myosin light chain kinase; MPO, myeloperoxidase; MSEA, metabolite set enrichment analysis; MylkCA, intestinal epithelial-specific constitutively activated; Mylkf/f, Mylkflox/flox; MylkKO Mylk knockout; MylkΔHep, hepatocyte-specific knockout; MylkΔIEC, intestinal epithelial cell-specific knockout; NAFLD, non-alcoholic fatty liver disease; PCA, principal component analysis; pMLC, phosphorylated myosin light chain; RT-qPCR, real-time quantitative PCR; siRNA, short interfering RNATG, triglycerides; tMylk, Villin-CA-Mylk; ZO-1, zonula occludens-1. Authors contributions JW, LX, and XW Designed and performed the experiments, analyzed the data, and drafted the manuscript. LX, ST, SW, YL, and JZ conducted experiments. JZ, JZ, JW, JZ, ZZ, JH, LQ, WH, and RH assisted with data analysis. YH provided single-cell sequencing bioinformatics analysis. JT provided the mice. HS and MW provided the ideas and revised the manuscript. LZ and HW supervised the project and revised the manuscript. All authors interpreted the data, reviewed, and revised the drafts of the manuscript, and approved the final manuscript for submission. Data availability statement Mouse single-cell RNA sequencing data have been uploaded to Zenodo (https://doi.org/10.5281/zenodo.18212354). Other datasets used and/or analyzed during the study are available from the corresponding author upon reasonable request. This study did not report the original code. Financial support This work was supported by the National Natural Science Foundation of China (grant nos. U24A20674, 82225008, U21A20375, 82270589), the Ministry of Science and Technology of China (MOST) National Key R&D Program (grant nos. 2023YFA1800800 and 2024ZD0530600), Shenzhen Medical Research Funds (grant no. B2302007), Natural Science Foundation of Anhui (grant nos. 2022AH010047, 2023AH020034), Scientific Research Project of Anhui Education Department (grant no. 2025AHGXZK50101), the Anhui Postdoctoral Scientific Research Program Foundation (grant no. 2025A1009), Grant from Anhui Medical University (grant no. 2023xkjT008), Ph.D., and Master Program of Anhui Medical University (grant nos. YJS20230111, YJS20240042, YJS20240062 and YJS20240084). Conflicts of interest The authors declare no conflicts of interest. Please refer to the accompanying ICMJE disclosure forms for further details. Acknowledgements We thank Jerrold R. Turner (Harvard Medical School) for kindly providing the MLCK knockout (MLCKKO), Mylkflox/flox (Mylkf/f), and Villin-CA-MLCK (tMLCK) Tg (MLCKCA) mice. Footnotes Author names in bold designate shared co-first authorship Supplementary data to this article can be found online at https://doi.org/10.1016/j.jhepr.2026.101849. Contributor Information Yongping Huang, Email: huangyp1226@whu.edu.cn. Hua Wang, Email: wanghua@ahmu.edu.cn. Li Zuo, Email: zuoli@ahmu.edu.cn. Supplementary data The following are the Supplementary data to this article: Multimedia component 1 Multimedia component 2 Multimedia component 3 Multimedia component 4 References Associated Data Supplementary Materials Multimedia component 1 Multimedia component 2 Multimedia component 3 Multimedia component 4 Data Availability Statement Mouse single-cell RNA sequencing data have been uploaded to Zenodo (https://doi.org/10.5281/zenodo.18212354). Other datasets used and/or analyzed during the study are available from the corresponding author upon reasonable request. This study did not report the original code.