Intracellular IL-24 ameliorates lipid metabolic disorders in metabolic dysfunction-associated steatohepatitis by restoring the autophagy-lysosome pathway

Gene expression data were obtained from three datasets (GSE48452 [19], GSE63067 [20], and GSE89632 [21]) in the GEO database [22], encompassing 154 liver tissue samples (82 NAFLD patients and 72 healthy controls) for autophagy-related gene (ARG) expression analysis. Gene symbols were matched to corresponding probes, and expression values for genes with multiple probes were averaged. Subsequent preprocessing involved log2 transformation, followed by batch effect removal using the "removeBatchEffect" function and data normalization via the "normalizeBetweenArrays" function, all within the R limma package. The integrity of the integrated datasets was confirmed by principal component analysis (PCA). A comprehensive set of 246 ARGs was compiled from the Human Autophagy Database (HADb) and a detailed literature review. Differential expression analysis of ARGs (DE-ARGs) between the NAFLD and control groups was performed using the limma package [23], with statistical significance set at |logFC| >0.2 and a Benjamini-Hochberg adjusted P < 0.05. Visualization of the results was achieved through heatmaps, volcano plots, and boxplots generated using the "heatmap" and "ggplot2" packages in R. To optimize feature selection, we utilized least absolute shrinkage and selection operator (LASSO) regression, a regularization technique that enhances predictive accuracy by selecting variables through penalized optimization. The "glmnet" R package [24] was employed for a 10-fold cross-validation procedure to tune the penalty parameter and identify the most critical features. For the analysis of public datasets, the term NAFLD is retained due to insufficient metabolic criteria in the original sample annotations.

Male C57BL/6J mice (6–8 weeks old, weight range 18–20 g) were sourced from HFK Bioscience (Beijing, China) for the purposes of this study. The animals were housed in a specific pathogen-free facility under a controlled 12-h light/dark cycle with unrestricted access to standard laboratory diet and water. Following random assignment to experimental groups, the study concluded with euthanasia of all mice under isoflurane anesthesia, collection of blood via cardiac puncture, and subsequent harvesting of liver tissues for further investigation.

Following a 7-day acclimation period during which all mice were fed standard chow, the experimental group was placed on a high-fat high-fructose diet (HFFD) for 18 weeks to induce MASLD. The HFFD consisted of 60% of calories from fat (XTHF60, Jiangsu Xietong Medicine Bioengineering Co., Ltd., Jiangsu, China) and 20% (w/v) fructose in the drinking water (D809612, Macklin, Shanghai, China). The control group was maintained on a normal chow diet (NCD) providing 10% of calories from fat (XTCON50J, same manufacturer) for the same duration [25].

The full-length IL-24 (NM_053095) cDNA was PCR-amplified and inserted into the pHBAAV-CMV-MCS-3flag AAV vector. The resulting IL-24 overexpression (oe) plasmid was then packaged into adeno-associated virus serotype 8 (AAV8) to generate the AAV8-IL-24 vector (AAV8-m-IL-24-3xflag-Null). Both the AAV8-IL-24 and the control AAV8-Null vectors were obtained from HanBio (Shanghai, China). At weeks 6 and 12 of the dietary intervention, mice were administered intravenously via the tail vein with 1 × 10¹¹ viral genomes of AAV8-IL-24 or AAV8-Null in 100 µL of PBS.

MASLD patients (n = 300) were recruited from the Department of Traditional and Western Medical Hepatology, Third Hospital of Hebei Medical University from January 2022 to December 2023. Inclusion criteria: (1) Diagnosis confirmed by abdominal ultrasound (independently assessed by two senior sonographers) and FibroScan (Controlled Attenuation Parameter, CAP >240 dB/m), consistent with the Guidelines for the prevention and treatment of metabolic dysfunction-associated (non-alcoholic) fatty liver disease (Version 2024) [26]; (2) Complete hepatic biochemical profiles available, with liver biopsy performed when necessary for histological assessment; (3) Signed informed consent. Exclusion criteria included: viral hepatitis (HBsAg/HCV antibody positivity), alcoholic liver disease, hepatic neoplasms, systemic inflammatory diseases, immunosuppressive therapy, and hepatic decompensation. 150 healthy controls were enrolled from routine health check-ups.

Methodological details (including primary mouse hepatocytes isolation and cell culture, construction and transfection of lentiviral vectors, cell transfection, construction of the cellular model, histopathological examination, qPCR, immunohistochemical staining, immunofluorescence staining, lysotracker red and bodipy 493/503 staining, autophagic flux analysis, flow cytometry, electron microscopy, western blot, transcriptomic analysis, metabolomics analysis, blood chemistry and cytokine quantification) have been archived in the. Supplementary Methods

Quantitative data were presented as mean ± standard deviation (SD). unless stated otherwise, and subjected to statistical analysis using a two-tailed unpaired Student's t test, one-way or two-way ANOVA with Bonferroni's post hoc test, or Spearman's rank correlation coefficient. All analyses were performed using GraphPad Prism version 9.0 software, and a p-value less than 0.05 (P < 0.05) was considered statistically significant.

Based on transcriptomic data from human liver tissue samples in the GEO database, PCA illustrated tighter clustering of samples within each group following normalization (Fig. 1A), indicating effective batch effect correction and strong intra-group consistency, validating sample processing reliability. Differential expression analysis revealed 26 ARGs associated with NAFLD, including IL-24 (Fig. 1B-C). Hepatic IL-24 mRNA levels were downregulated in NAFLD patients compared to controls (P < 0.0001; Fig. 1D), which was further supported by qPCR analysis in the HFFD-induced mouse model of NAFLD (Fig. 1E). LASSO regression analysis (Fig. 1F) selected "lambda.1se" as the optimal penalty parameter, with IL-24 remaining a significant feature with a non-zero coefficient (β = −0.277), suggesting an inverse relationship with NAFLD progression. Correlation analyses (Fig. 1G) demonstrated a significant positive correlation between IL-24 expression and the autophagy marker MAP1LC3B mRNA (R = 0.39, P < 0.001), and a significant negative correlation with mTOR expression (R = − 0.55, P < 0.001), implying a potential role for IL-24 in promoting autophagy through mTOR inhibition.

In the HFFD-induced mouse model of MASH, the observed bioinformatics data were consistent with in vivo findings: HFFD-fed mice showed reduced hepatic IL-24 expression along with increased protein levels of the autophagy receptors p62 and the autophagosome marker LC3-II (Fig. 1H; quantification in Fig. S1A), aligning with an inhibited autophagy-lysosome pathway [7]. Analysis of serum samples from MASLD patients recruited at our hospital revealed significantly lower serum IL-24 concentrations in MASLD patients (median [IQR]: 191.25 [171.97–212.61]) compared to healthy controls (220.545 [205.08–245.83]; P < 0.0001; Fig. 1I). Moreover, a further significant decrease in serum IL-24 (116.02 [95.48–135.75]) was observed in MASLD patients with concomitant transaminitis compared to those with MASLD alone (P < 0.0001). Correlation analysis (Fig. 1J) revealed that IL-24 expression levels demonstrated significant negative correlations with CAP values (r = −0.61, P < 0.0001), liver stiffness measurement (LSM, r = −0.39, P < 0.0001), AST (r = −0.65, P < 0.0001), and ALT (r = −0.75, P < 0.0001), indicating a negative association between IL-24 and hepatic steatosis severity, fibrosis progression, and liver injury in MASLD patients. Baseline clinical characteristics are detailed in Table S1.

In the 18-week HFFD-induced C57BL/6 mouse model, IL-24 treatment was initiated at week 6 of the diet (Fig. 2A). Remarkably, IL-24-treated mice exhibited a significant divergence in body weight trajectory compared to the HFFD group from week 7 onward (Fig. 2C). By the conclusion of the study at week 18, IL-24 administration resulted in a significant reduction in both body weight and liver-to-body weight ratio relative to HFFD-fed mice (Fig. 2C, D). Intraperitoneal glucose tolerance test (IPGTT) and insulin tolerance test (ITT) results indicated that IL-24 effectively ameliorated HFFD-induced glucose intolerance and insulin resistance, as evidenced by statistically significant differences in the area under the curve (AUC) between the treatment groups (Fig. 2E-F; quantification in Fig. S1B).

Serum biochemical analysis revealed that IL-24 not only significantly reduced the elevated ALT and AST levels in HFFD-fed mice (Fig. 2G) but also concurrently downregulated the expression of pro-inflammatory cytokines TNF-α, IL-6, and IL-1β in serum (Fig. S1D). H&E staining of liver sections showed that HFFD feeding resulted in a NAS value of 6–7, which was significantly reduced by IL-24 intervention, as evidenced by decreased liver fatty degeneration, inflammatory infiltration, and hepatocyte ballooning (Fig. 2H, I). Further immunohistochemical staining demonstrated that IL-24 intervention markedly decreased positive signals of macrophage marker F4/80 and neutrophil marker Ly-6G in the liver, indicating suppressed infiltration of both macrophages and neutrophils (Fig. S1E). Additionally, Masson's trichrome and Sirius red staining demonstrated that IL-24 treatment attenuated perisinusoidal fibrosis (Fig. 2J), which was supported by reduced protein levels of α-SMA and COL1α1 following IL-24 intervention (Fig. S1F).

Confirmation of successful in vivo IL-24 overexpression was achieved by Western blot analysis of the Flag-tag, which also demonstrated significantly lower endogenous IL-24 expression in liver and serum of HFFD-fed mice compared to NCD controls (Fig. 3A-B). Immunohistochemical analysis (Fig. 3C) showed that IL-24 intervention significantly reduced p62, p-TFEB expression, p-mTOR/mTOR ratio, while increasing nuclear TFEB accumulation compared to the HFFD group, supporting the hypothesis that IL-24 enhances autophagy through modulation of mTOR-related pathways (IHC images and quantification of mTOR and TFEB are provided in Fig. S1G). Further Western blot analysis revealed that HFFD feeding resulted in increased expression of p-mTOR and p-TFEB, along with reduced p-AMPK and nuclear TFEB levels. These changes were accompanied by decreased LAMP2 expression and accumulation of LC3-II and p62, findings that are consistent with impaired autophagic-lysosomal function. IL-24 treatment alleviated these alterations, as evidenced by increased p-AMPK levels and decreased p-mTOR and p-TFEB expression, suggesting its potential role in regulating autophagy through the AMPK/mTOR/TFEB axis, a mechanism that was validated in our subsequent in vitro experiments. Furthermore, the simultaneous increase in LAMP2 and decrease in LC3-II and p62 are indicative of a potential restoration of autophagy-lysosome function (Fig. 3D; quantification in Fig. S2A).

TEM demonstrated substantial accumulation of lipid droplets and a scarcity of autophagosomes in hepatocytes from HFFD-fed mice. In contrast, liver sections from IL-24-treated mice showed a marked decrease in lipid droplets and an increase in autolysosomes, which contained lipid inclusions and were often surrounded by membranous structures, indicating the presence of lipophagy (Fig. 3E). These findings suggest a potential mechanism whereby IL-24 was found to exert protective effects against MASH progression by promoting autophagy, which was later demonstrated to reduce intrahepatic lipids through enhanced lipolysis and lipid oxidation.

To corroborate the proposed mechanism of IL-24 at the cellular level, we performed experiments using AML12 and primary hepatocytes (PHs). Immunofluorescence analysis revealed a significant reduction in IL-24 expression upon palmitic acid (PA) stimulation in both AML12 cells and PHs compared to the control, whereas overexpression of IL-24 substantially reduced PA-induced expression of p62 and p-mTOR (Fig. 4A-D). Strikingly, IL-24-treated cells displayed significantly increased nuclear localization of TFEB (Fig. 4B, D; quantification in Fig. S2D for AML12 cells), suggesting enhanced nuclear translocation following TFEB dephosphorylation, potentially representing a cellular attempt to compensate for PA-mediated autophagy inhibition. Western blot analysis demonstrated that PA stimulation induced dose-dependent increases in LC3-II and p62 expression in PHs with escalating concentrations (Fig. S2E). Additionally, PA treatment led to reduced p-AMPK levels, enhanced p-mTOR and p-TFEB, diminished nuclear TFEB protein, and downregulated LAMP2 expression (Fig. 4E; quantification in Fig. S2F). These findings collectively indicate that PA treatment leads to molecular alterations consistent with impaired autophagic-lysosomal function, including the accumulation of autophagy substrates and inhibition of key regulatory kinases. Importantly, IL-24 intervention attenuated these PA-induced alterations.

To further dissect the signaling pathway, we evaluated the role of AMPK using compound C (CC), an AMPK inhibitor, in PHs. Compared to the PA group, cells exposed to both PA and CC exhibited a more pronounced inhibition of p-AMPK and a further increase in the expression of p-mTOR and p-TFEB, leading to reduced nuclear TFEB translocation and a more severe impairment of autophagic flux, evidenced by elevated LC3-II and p62 levels. Critically, IL-24 treatment was able to alleviate these effects even under PA + CC conditions (Fig. 4F; quantification in Fig. S3A; p-AMPK immunofluorescence in Fig. S3B). Collectively, these findings suggest that IL-24 may regulate autophagic-lysosomal function through the AMPK/mTOR/TFEB axis.

To investigate the mechanism of IL-24 in regulating autophagic flux, we employed an mCherry-EGFP-LC3 dual-fluorescence reporter system in both AML12 and PHs. PA-stimulated cells exhibited a significant reduction in red puncta (autolysosomes) and an increase in yellow-green puncta (autophagosomes) compared to the control groups, indicating a blockage in autophagosome maturation. IL-24 intervention increased the intensity of the mCherry signal, suggesting a restoration of autolysosome formation. Co-administration of rapamycin (RAPA), an established autophagy inducer, further enhanced the red fluorescence, demonstrating a synergistic effect with IL-24 on autophagic flux. Following the addition of the autolysosome inhibitor chloroquine (CQ) to the PA + IL-24 treatment group, the number of autolysosomes decreased but remained significantly higher than in the PA-only group, as evidenced by a relative increase in red fluorescent puncta (Fig. 5A-B). Population-based quantification using flow cytometry (to avoid sampling bias from limited imaging fields [27]) confirmed that IL-24 overexpression significantly increased the percentage of mCherry-positive cells, further supporting the conclusion that IL-24 restores autophagic-lysosomal function (Fig. 5C, Fig. S3C shows control and oeIL-24 groups).

Consistent with the aforementioned results, Western blot analyses in PHs demonstrated that RAPA treatment following PA stimulation decreased p-p70S6K (indicating mTOR inhibition), reduced levels of p-TFEB, p62, and LC3-II, and increased LAMP2 expression and nuclear TFEB accumulation, collectively indicating restored autophagy-lysosome function. IL-24 intervention further enhanced these effects, indicating a potential synergistic action in promoting autophagosome-lysosome fusion. Remarkably, IL-24 significantly reduced p62 and LC3-II levels while increasing LAMP2 expression even under CQ treatment, providing functional evidence that IL-24 promotes autophagic flux by enhancing lysosomal pathway (Fig. 5D-E; quantification in Fig. S3D-E).

To assess the functionality of the late stages of the autophagy-lysosome pathway, we employed LysoTracker staining in both AML12 cells and PHs to visualize acidified autolysosomes and lysosomes. Upon co-staining with BODIPY 493/503, a marker for lipid droplets, PA-treated cells showed a significant decrease in LysoTracker red fluorescence, indicative of impaired lysosomal acidification. IL-24 intervention significantly restored the red fluorescence intensity while simultaneously reducing the abundance of BODIPY-labeled lipid droplets (Fig. 6A-B). Colocalization analysis further demonstrated a substantial reduction or complete disappearance of lipid droplet fluorescence in regions enriched with LysoTracker signal. This spatial pattern suggests a potential link between the restored lysosomal activity and the degradation of lipids. Collectively, these findings support the conclusion that the autophagy-lysosome pathway is a contributor to the IL-24-induced reduction of intracellular lipids, a finding further validated by multi-omics analyses.

Transcriptomic profiling of murine liver tissue provided further evidence for IL-24's regulatory function in autophagy. Heatmap analysis revealed that IL-24 treatment led to the downregulation of genes associated with mTOR signaling and the upregulation of genes involved in autophagy (Fig. 6C). Reactome pathway enrichment analysis identified significant clustering of differentially expressed genes within key pathways, including "Macroautophagy" and "mTOR signaling" (Q-value < 0.05) (Fig. 6D). These expression patterns were consistent with earlier bioinformatic predictions derived from human samples (Fig. 1). Integrated analysis of transcriptomic data and prior experimental results supports a model in which IL-24 activates AMPK to suppress mTOR activity, promotes TFEB nuclear translocation (Figs. 3 and 4), thereby enhances autophagic flux while reducing lipid droplet accumulation.

To identify the functional receptor complex type of IL-24 in hepatocytes (IL20R1/IL20R2 vs. IL22R1/IL20R2), we first examined the activation of downstream STAT1 and STAT3. In both animal and cellular models, IL-24 significantly increased the p-STAT3/STAT3 ratio but did not markedly alter p-STAT1/STAT1 levels, preliminarily suggesting that IL-22R1/IL-20R2 may be its primary functional receptor (Fig. 7A-B, Fig. S4A). To validate this hypothesis, we knocked down IL20R1 and IL22R1 separately in AML12 cells (Table S2, Fig. S4B). Western blot results demonstrated (Fig. 7C-D, Fig. S4C) that compared to the PA group, IL20R1 knockdown increased p-STAT3 and p-AMPK expression while reducing LC3-II and p62 levels, effects similar to those observed in the PA + oeIL24 group. Conversely, IL22R1 knockdown reversed the protective effects of IL-24, with protein expression levels resembling those of the PA group. Notably, neither knockdown significantly affected STAT1 activation, further supporting that IL-24 primarily functions through IL22R1/IL20R2 rather than IL20R1/IL20R2. Phenotypically, Oil Red O staining (Fig. 7E-F) and biochemical assays of cell supernatants (Fig. 7G) were consistent with these molecular changes: under IL-24 intervention, IL22R1 knockdown exacerbated PA-induced lipid accumulation and cellular damage (elevated AST and ALT), whereas IL20R1 knockdown partially mimicked the protective effects of IL-24, alleviating lipid accumulation and hepatocyte injury. These results indicate that IL-24 regulates autophagy and mitigates hepatocyte steatosis and damage through the IL-22R1/IL-20R2 receptor complex.

The regulatory effects of IL-24 on lipid metabolism in MASH were investigated through integrated transcriptomic and metabolomic analyses. PCA revealed significant changes in gene expression (Fig. 8A) and metabolic profiles (Fig. 8B) following IL-24 treatment in HFFD-fed mice. Differential gene expression analysis demonstrated the downregulation of genes involved in lipid synthesis (FASN, ACACA) and the upregulation of genes associated with lipid catabolism (CPT1A, ACADL, HADHA) (qPCR-validated P < 0.05, Fig. 8C-E). GSEA further confirmed the significant enrichment of differentially expressed genes in fatty acid degradation pathways (NES = 2.098, Adjusted P = 0.005, Fig. 8F). Metabolomic profiling revealed a significant decrease in a range of carbohydrate-related metabolites (such as Adenosine-5'-Diphosphoglucose, UDP-glucose, D-Glucose 6-Phosphate, Glucose) upon IL-24 intervention (Fig. 8G), suggesting a decrease in glucose metabolic intermediates. Furthermore, we observed increased levels of short-, medium-, and long-chain acylcarnitines (CACs) (Fig. 8H, Fig. S4D). This finding, when integrated with the upregulation of fatty acid oxidation pathways in the transcriptomic data, is consistent with increased flux through β-oxidation in the liver. The concurrent occurrence of these glycolipid metabolic changes suggests that IL-24 intervention may promote metabolic reprogramming, a shift in metabolic priority from carbohydrate metabolism toward lipid metabolism, ultimately manifesting as reduced hepatic lipid droplets (ORO staining, Fig. 8I; quantification in Fig. S4E) and improved serum metabolic parameters (glucose, triglycerides, cholesterol, uric acid) (Fig. 8J, Fig. S4F). KEGG pathway enrichment analysis showed a significant clustering of differential metabolites within autophagy-related pathways (Fig. 8K). Further analysis of key metabolites (e.g., GDP, PE) indicated that their alteration patterns suggest potential enhancement of autophagic processes. In summary, these results demonstrate that IL-24 reprograms cellular metabolism toward catabolic processes, including autophagy, to maintain energy homeostasis.