Hepatic LGALS4 Alleviates Diet‐Induced Steatosis by Promoting AMPK‐Dependent Fatty Acid Oxidation

ABSTRACT

Aims

Metabolic dysfunction‐associated steatotic liver disease (MASLD) is characterised by impaired hepatic lipid handling, and despite recent therapeutic advances, additional mechanistically distinct therapeutic strategies remain needed. We sought to identify endogenous hepatic regulators induced by energetic challenge that could be leveraged to counteract pathological lipid accumulation.

Materials and Methods

Integrative transcriptomic analyses were performed across multiple physiological energetic challenges, including cold exposure, β3‐adrenergic activation, fasting and exercise, to identify conserved hepatic energy‐responsive genes. The functional role of Galectin‐4 (LGALS4) was investigated using hepatocyte‐specific gain‐ and loss‐of‐function approaches in high‐fat diet (HFD)‐fed mice and primary hepatocytes. Mitochondrial function was assessed via Seahorse metabolic flux analysis.

Results

LGALS4 emerged as a conserved hepatic factor induced by energetic challenge but suppressed during chronic nutrient excess and in human MASLD liver samples. Hepatocyte‐specific overexpression of Lgals4 markedly attenuated HFD‐induced hepatic steatosis and reduced hepatic triglyceride accumulation, with limited effects on systemic glucose tolerance or insulin sensitivity. Conversely, Lgals4 knockdown exacerbated lipid accumulation and impaired fatty acid oxidation (FAO) gene expression in hepatocytes. Mechanistically, LGALS4 increased AMPK and ACC phosphorylation and upregulated the PGC1α–PPARα fatty acid oxidation programme, leading to enhanced mitochondrial oxidative capacity and reduced lipid accumulation in hepatocytes. These effects were largely abolished by AMPK inhibition, supporting an AMPK‐dependent mechanism.

Conclusions

LGALS4 is an energy‐responsive hepatic regulator that protects against steatosis by promoting AMPK‐dependent fatty acid oxidation. These findings identify LGALS4 as a potential metabolically selective target for MASLD treatment.

1. Introduction

The liver serves as a central hub for maintaining systemic energy homeostasis, dynamically balancing glucose and lipid metabolism to meet fluctuating physiological demands [1, 2, 3]. This metabolic flexibility acts as a critical pivot between health and disease states [4, 5, 6]. Under conditions of chronic nutrient excess, this delicate equilibrium is disrupted, leading to the ectopic accumulation of lipids in the liver and driving the development of metabolic dysfunction‐associated steatotic liver disease (MASLD) [7, 8, 9]. MASLD, one of the most prevalent chronic liver diseases globally, encompasses a histological spectrum ranging from simple steatosis to metabolic dysfunction‐associated steatohepatitis (MASH), which is characterised by hepatocellular ballooning, lobular inflammation and progressive fibrosis [10, 11, 12, 13, 14, 15, 16]. Recent therapeutic advances, including the approval of resmetirom and the clinical emergence of GLP‐1 receptor agonists such as semaglutide, have significantly advanced the treatment landscape of MASLD. However, disease heterogeneity and incomplete long‐term disease control continue to highlight the need to identify endogenous pathways that maintain hepatic metabolic homeostasis and may serve as novel therapeutic targets.

Conversely, enhancing hepatic energy expenditure, particularly fatty acid oxidation (FAO) and mitochondrial respiration, can restore metabolic balance and represents a key therapeutic strategy to counteract energy surplus and lipid overload [17, 18, 19]. Under energetic challenges, such as fasting, cold exposure and physical exercise, the liver exhibits metabolic flexibility, rapidly reprogramming towards lipid mobilisation and FAO [20, 21, 22, 23, 24, 25]. However, whether these diverse stimuli converge on a conserved hepatic transcriptional programme, and which effectors orchestrate this response, remains unclear. By tracing the conserved hepatic transcriptional landscapes responsive to diverse energy demands, we aim to uncover key regulators that bridge the gap between physiological adaptation and therapeutic potential. Identifying such factors is essential for developing novel strategies to restore energy balance and mitigate the progression of metabolic diseases.

AMP‐activated protein kinase (AMPK) is a central mediator of hepatic responses to energetic challenge and functions as a key cellular energy sensor. Phosphorylation of AMPK at Thr172 promotes FAO through inhibitory phosphorylation of acetyl‐CoA carboxylase (ACC) while suppressing lipogenesis to restore metabolic balance [26, 27, 28, 29, 30]. Importantly, AMPK activates PGC1α, which coactivates PPARα to induce mitochondrial and peroxisomal β‐oxidation genes, including Acox1 and Cpt1a [31, 32, 33]. Given the central role of AMPK in hepatic FAO, identifying endogenous regulators that engage this pathway may deepen our understanding of hepatic metabolic adaptation and uncover potential targets for MASLD.

Here, we sought to define conserved transcriptional programmes underlying hepatic adaptation to energetic challenges and identify regulators of metabolic flexibility. Integrative transcriptomic analyses across diverse perturbations revealed a core energy‐responsive network enriched for lipid metabolism, within which Galectin‐4 (LGALS4) emerged as a previously unrecognised regulator. Hepatic LGALS4 is induced by energy demand but suppressed during metabolic overload, functioning as a modulator of lipid homeostasis. Mechanistically, LGALS4 activates the AMPK–PGC1α axis to enhance FAO, thereby alleviating hepatic steatosis without broadly affecting glucose homeostasis. Collectively, our findings establish LGALS4 as a novel mediator of the hepatic adaptive response and uncover a regulatory axis with therapeutic potential for MASLD.

2. Research Design and Methods

2.1. Animal Studies

Male C57BL/6J mice were maintained under a 12 h light/dark cycle at 22°C ± 0.5°C in a specific pathogen‐free (SPF) environment with ad libitum access to food and water. Mice were randomly assigned to experimental groups based on their body weight. For thermogenic challenges, 11‐week‐old mice were acclimated at 30°C for 1 week before being maintained at 30°C or exposed to 4°C for 6 h. Chemical interventions involved a single intraperitoneal injection of CL316243 (1 μg/g), T3 (0.75 μg/g) or vehicle, with tissues harvested 6 h post‐injection. Nutritional status was manipulated via fasting (up to 48 h) or a 24‐h fast/24‐h refeeding cycle. For the exercise model, mice underwent an 8‐week treadmill programme (6 days/week), reaching a sustained speed of 28 m/min. To induce MASLD, 6–8 week‐old mice were fed a high‐fat diet (HFD, 60% kcal fat) for 9 or 31 weeks. In the diet‐switch (HFD‐DW) model, mice were transitioned back to standard chow.

Hepatocyte‐specific Lgals4 overexpression was achieved via tail vein injection of AAV8‐TBG‐Lgals4‐flag or AAV8‐TBG‐GFP (1 × 10¹¹ viral particles/mouse). Two weeks post‐injection, mice were challenged with an HFD for 16 weeks to induce MASLD. Mice were euthanised by pentobarbital sodium followed by cervical dislocation; sample sizes are indicated in the figure legends. All procedures for animal experiments were approved by the Fudan University Shanghai Medical College Animal Care and Use Committee and followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

2.2. Human Studies

Human liver tissues were obtained from patients with biopsy‐confirmed MASLD and from histologically normal controls undergoing hepatic surgery for non‐metabolic indications at Zhongshan Hospital, Fudan University. Clinical and demographic characteristics of all subjects are summarised in Table S5. The study was approved by the Institutional Review Board of Zhongshan Hospital and conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent prior to inclusion.

2.3. Cell Culture and Treatment

AML12 and HepG2 cells (ATCC) were cultured in DMEM (Gibco, USA) supplemented with 10% FBS and 1% penicillin–streptomycin. To mimic MASLD in vitro, cells were challenged with 0.5 mM free fatty acids (FFAs; oleic:palmitic acid = 2:1) for 48 h. Plasmids were transfected using Lipofectamine 3000 (Invitrogen). For siRNA knockdown, cells were transfected with siRNA using Lipofectamine RNAi MAX (Thermo Fisher) for 48 h. For AMPK inhibition, cells were pretreated with 10 μM Compound C for 1 h. Primary hepatocytes were isolated from 8 to 12‐week‐old mice using a two‐step collagenase perfusion technique and seeded onto rat tail collagen‐coated plates. For primary hepatocyte experiments, freshly isolated cells were used within 24 h of isolation to preserve their metabolic characteristics. Detailed information on key reagents is listed in Table S2.

2.4. RNA‐Sequencing and Bioinformatics

Total RNA from mouse livers was sequenced on the Illumina NovaSeq 6000 platform. Clean reads were aligned to the GRCm39 genome using HISAT2, and DEGs were identified using DESeq2 (|log₂FC| ≥ 1, adjusted p < 0.05). Public datasets (GSE118787, GSE299228, GSE280324) were obtained from the GEO database. Functional enrichment (GO/KEGG) and GSEA were performed using DAVID and R (v4.4.3) with clusterProfiler and fgsea packages.

2.5. Metabolic and Biochemical Analyses

For GTT and ITT, mice were fasted for 16 h or 4 h, respectively. Investigators performing the tests, blood glucose measurements and data analyses were blinded to group allocation. Glucose (0.5 g/kg) and insulin (1 U/kg) doses were optimised for the HFD model to ensure blood glucose levels remained within the glucometer's detection range during GTT and to adequately challenge the model's inherent insulin resistance during ITT. Blood glucose levels were measured from tail bleeding using a glucometer at 0, 15, 30, 60 and 120 min post‐administration. Plasma and hepatic triglyceride (TG) and total cholesterol (TC) levels, as well as plasma ALT/AST activities, were measured using commercial enzymatic kits (Applygen, China) following the manufacturers' instructions. Detailed information on key reagents, including chemicals and commercial kits, is listed in Table S2.

2.6. Histopathology and Intracellular Imaging

Liver sections were subjected to H&E staining (paraffin‐embedded) or Oil Red O staining (OCT‐embedded frozen sections) to evaluate morphology and lipid accumulation. Histological assessment was performed in a blinded manner according to the NAFLD Activity Score (NAS) established by the NASH Clinical Research Network. Steatosis (0–3), lobular inflammation (0–3) and hepatocyte ballooning (0–2) were scored independently, and the composite NAS was calculated as the sum of these individual scores.

For morphometric quantification of hepatic steatosis, three non‐consecutive frozen liver sections (100 μm apart) from the left lateral lobe were analysed per animal. Five random non‐overlapping fields per section were captured at 20× magnification, excluding large vessels and tissue edges. Oil Red O‐positive areas were quantified by automated thresholding in ImageJ under identical settings by an investigator blinded to group allocation. Hepatic lipid accumulation was expressed as the percentage of Oil Red O‐positive area relative to the total parenchymal area.

Cells were fixed with 4% paraformaldehyde for 15 min at room temperature, washed with PBS and incubated with BODIPY 493/503 (Beyotime) for 15 min to label neutral lipids. Nuclei were counterstained with DAPI. For immunofluorescence, primary hepatocytes were incubated with LGALS4 primary antibodies and Cy3‐conjugated secondary antibodies. Fluorescence images were acquired using a fluorescence microscope (Olympus, Japan).

2.7. Seahorse Analysis

Mitochondrial oxygen consumption rate (OCR) was measured using a Seahorse XFe96 Extracellular Flux Analyser (Agilent). AML12 cells were incubated in Seahorse XF assay medium supplemented with 10 mM glucose, 1 mM pyruvate and 2 mM glutamine (pH 7.4). OCR was measured following sequential injection of oligomycin (1.5 μM), FCCP (1.0 μM) and rotenone/antimycin A (0.5 μM each). Values were normalised to protein content. Basal respiration, ATP production and maximal respiration were calculated according to the manufacturer's instructions.

2.8. RNA Isolation and Real‐Time Quantitative PCR

Total RNA was extracted using RNAiso Plus (Takara) and reverse‐transcribed with the PrimeScript RT reagent kit (Takara). Quantitative PCR was performed using SYBR Green Mix (Beyotime). The thermal profile included an initial denaturation at 95°C (2 min), followed by 40 cycles of 95°C (15 s), 60°C (30 s) and 72°C (30 s). Relative mRNA levels were calculated using the 2^−ΔΔCT method and normalised to Actin. Primer sequences are provided in Table S1.

2.9. Western Blotting

For Western blotting, total proteins were extracted with RIPA buffer (Beyotime) supplemented with protease and phosphatase inhibitors. Equal amounts of protein were separated by SDS‐PAGE, transferred to PVDF membranes (Millipore) and blocked with 5% skim milk. Membranes were incubated with primary antibodies overnight at 4°C, followed by HRP‐conjugated secondary antibodies. Signals were detected using a chemiluminescence system (Tannon) and quantified via ImageJ. Antibody details are listed in Table S2.

2.10. Statistical Analysis

Data are expressed as mean ± SD for experiments involving primary biological variability (including animal studies, primary hepatocyte experiments and human cohort analyses) and as mean ± SEM for cell line‐based assays to reflect experimental precision. Statistical evaluations were performed using GraphPad Prism 9.0. Normality of data distributions was assessed using the Shapiro–Wilk test, and homogeneity of variance was evaluated using the F‐test for two‐group comparisons or the Brown–Forsythe test for multiple‐group comparisons, where appropriate. For datasets with small sample sizes, distributional assessment was additionally supported by visual inspection of individual data points. Comparisons between two groups were analysed by two‐tailed unpaired Student's t‐tests, while multiple‐group comparisons were conducted using one‐way or two‐way ANOVA followed by appropriate post hoc tests. Sample sizes were determined based on prior experience and power analysis. Mice were randomly assigned to experimental groups. Significant differences from the above tests are indicated in the figures as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Non‐significant differences are indicated in the figures as ns.

3. Results

3.1. Integrated Transcriptomics Identifies Lgals4 as a Regulator of Hepatic Metabolic Adaptation

The liver serves as a central metabolic hub that maintains systemic energy homeostasis through dynamic regulation of glucose and lipid metabolism [4, 34, 35]. To define the fundamental molecular signatures that coordinate hepatic adaptation to energetic stress, we performed an integrated transcriptomic analysis across four distinct energetic challenges: cold exposure (Cold), β3‐adrenergic activation using CL316243 (CL), fasting (Fast; GSE118787) and calorie restriction (CR; GSE299228) (Figure 1A). Differential expression analysis revealed widespread transcriptional remodelling in the liver across all conditions (Figure 1B), indicating a robust and conserved hepatic network in response to diverse energy deficits.

FIGURE 1.

Integrated transcriptomic profiling identifies Lgals4 as a conserved hepatic factor responsive to energetic challenges. (A) Schematic overview of the integrative screening strategy. Hepatic transcriptomic profiles were analysed from four independent energy‐challenge models: Cold exposure (Cold, 4°C), pharmacological thermogenic activation (CL316243), fasting (GSE118787) and calorie restriction (GSE299228). (B) Volcano plots illustrating differentially expressed genes (DEGs) in the liver under the indicated energetic stimuli. Red and blue dots represent significantly upregulated and downregulated genes, respectively (threshold: |log₂FC| > 1, p adj. < 0.05). (C–E) Gene set enrichment analysis (GSEA) plots showing significant enrichment of fatty acid metabolic process and fatty acid beta‐oxidation in liver transcriptomes following cold exposure (C), CL316243 treatment (D) and fasting (E). Normalised enrichment scores (NES) and p values are indicated. (F and G) Venn diagram (F) and heatmap (G) identifying 21 core genes consistently upregulated across all four energetic challenges. (H) Gene Ontology (GO) enrichment analysis of the 21 core energy‐responsive genes, ranked by −log₁₀ (p‐value).

Pathway enrichment and gene set enrichment analysis (GSEA) demonstrated that while cold and CL316243 treatments preferentially targeted fatty acid metabolism and PPAR signalling, nutritional interventions (Fast and CR) primarily modulated lipid mobilisation and β‐oxidation. Despite these stimulus‐specific nuances, all four conditions converged on a coordinated reprogramming of hepatic lipid catabolism (Figures 1C–E and S1–S4). Furthermore, expression profiling showed that while impacts on de novo lipogenesis (DNL) were context‐dependent, these stimuli shared a convergent signature in promoting FAO, characterised by marked Pgc1a induction (Figure S5A–C).

To identify the core effectors underlying these adaptive responses, we intersected differentially expressed genes (DEGs) across all conditions, revealing a core module of 21 genes uniformly upregulated under all energy stimuli (Figure 1F). This signature was further validated by heatmap analysis, which confirmed the high concordance of this 21‐gene signature, reinforcing its role as a fundamental transcriptional response to metabolic demand (Figures 1G and S6A). Functional annotation of these core genes revealed strong enrichment in lipid metabolic processes, including triglyceride regulation and fatty acid biosynthesis, as well as the PPAR signalling pathway (Figures 1H and S6B,C). Gene‐concept network analysis demonstrated that while many of these candidates, including Apoa4, Plin5 and Crat, are well‐established regulators in lipid trafficking and oxidation [36, 37, 38], Lgals4, a member of the galectin family, emerged as a prominent but previously uncharacterised factor in hepatic metabolism (Figure S6D).

Beyond thermal and nutritional cues, physical exercise represents a profound physiological challenge that confers systemic metabolic benefits [39, 40, 41]. To assess whether this programme extends to exercise, we analysed a transcriptomic dataset from exercise‐trained mice (GSE280324) and found that 13 of the 21 core genes, especially Lgals4, were significantly induced, further reinforcing its role as a master regulator of energy balance (Figure S6E). Notably, cross‐tissue expression profiling in humans (GSE7905) confirmed that while LGALS4 is highly enriched in the gastrointestinal tract, it maintains substantial expression levels in the liver (Figure S6F), supporting its potential involvement in hepatic adaptive responses. Together, these data uncover a conserved hepatic transcriptional programme responsive to energy challenges and identify Lgals4 as a previously unrecognised candidate regulator of metabolic adaptation.

3.2. Hepatic Lgals4 Is Dynamically Regulated by Energetic and Metabolic Stress

To validate the bioinformatic findings, we systematically examined Lgals4 expression across multiple physiological and pathological metabolic contexts. Hepatic Lgals4 mRNA levels were robustly induced by acute energy‐demanding stimuli, including cold exposure and β3‐adrenergic activation (CL316243) (Figure 2A,B). During fasting, Lgals4 expression exhibited a time‐dependent increase, peaking at 24–48 h and was rapidly suppressed upon refeeding (Figures 2C and S7A). These transcriptional shifts were mirrored at the protein level, as evidenced by the marked elevation of LGALS4 under these conditions (Figure 2D–F). Notably, neither fasting nor cold exposure significantly altered Lgals4 expression in brown adipose tissue (BAT) or quadriceps muscle (Figure S7B,C), highlighting the liver as a primary metabolically responsive site of LGALS4 induction during energetic challenges. We next assessed Lgals4 regulation under sustained physiological energy demand using a chronic aerobic exercise model. Eight weeks of treadmill training significantly upregulated hepatic LGALS4 (Figure S7D,E). Moreover, systemic administration of triiodothyronine (T3), a well‐established driver of hepatic energy expenditure, robustly induced hepatic Lgals4 expression (Figure S7F), confirming its responsiveness to physiological and endocrine signals that orchestrate energy expenditure.

FIGURE 2.

Dynamic regulation of hepatic Lgals4 by physiological energetic cues and metabolic stress. (A and B) qPCR analysis of hepatic Lgals4 mRNA levels in mice subjected to acute cold exposure (4°C, 6 h) (n = 6) (A) or pharmacological β3‐adrenergic activation via CL316243 injection (n = 8) (B). (C) Hepatic Lgals4 mRNA expression was measured in fed, 24‐h fasted and 24‐h refed mice (n = 6). (D–F) Representative Western blots and quantification of hepatic LGALS4 protein levels under cold exposure, CL316243 treatment (D and E) and the fast/refed cycle (F). Tubulin or HSP90 served as loading controls (n = 3–4). (G–J) Hepatic Lgals4 mRNA (G and H) and protein (I and J) levels in mice fed a high‐fat diet (HFD) for 9 or 31 weeks (n = 4–6) and in a high fat diet–diet switch (HFD‐DW) model (n = 5–12). (K) Relative mRNA expression of Lgals4 in primary hepatocytes treated with oleic acid and palmitic acid (OA + PA) (n = 3). (L) Relative mRNA levels of LGALS4, PGC1α, FASN and SREBP1C in liver biopsies from non‐MASLD controls (n = 5) and MASLD patients (n = 11). Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by Student's t‐test or one‐way ANOVA with Tukey's correction for multiple comparisons.

Given its sensitivity to physiological energy deficit, we next investigated whether Lgals4 is dysregulated under conditions of metabolic overload. In contrast, hepatic Lgals4 expression was markedly downregulated in mice following high‐fat diet (HFD) feeding, with progressive suppression observed from 9 to 31 weeks during disease progression. These time points represent early and advanced stages of HFD‐induced MASLD, as reflected by increased hepatic triglyceride content, lipogenic gene expression and hydroxyproline levels (a marker of collagen deposition) (Figure S7G–I). To assess whether this suppression is reversible, we subjected HFD‐fed mice to dietary reversal by switching to chow (HFD–DW). Notably, restoration of metabolic homeostasis robustly rescued Lgals4 mRNA and protein expression (Figure 2G–J), indicating that its regulation remains highly plastic and sensitive to metabolic status. This inhibitory effect of lipid overload was further confirmed in vitro, where treatment with oleic and palmitic acids (OA + PA) significantly reduced Lgals4 expression in primary hepatocytes (Figure 2K). Clinical validation in liver biopsies revealed that LGALS4 expression is significantly downregulated in MASLD patients compared to healthy controls, coinciding with diminished PGC1α and reciprocal induction of FASN and SREBP1C (Figure 2L), reinforcing its clinical significance in human hepatic lipid homeostasis. Collectively, these findings identify Lgals4 as a bidirectional metabolic sensor that is induced by energy deficit but suppressed during chronic nutrient excess and metabolic dysfunction, highlighting its potential role in maintaining hepatic lipid metabolic flexibility and homeostasis.

3.3. Hepatocyte Lgals4 Overexpression Alleviates Diet‐Induced Hepatic Steatosis

Given its induction by energetic stress and inverse association with MASLD, we next investigated the in vivo role of Lgals4 using an adeno‐associated virus (AAV8) system driven by hepatocyte‐specific thyroxine‐binding globulin (TBG) promoter overexpression system in adult C57BL/6J mice. Two weeks post‐injection of AAV‐TBG‐Lgals4 or control AAV‐TBG‐GFP, mice were fed a high‐fat diet (HFD) for 16 weeks to induce MASLD (Figure 3A), with efficient hepatic overexpression confirmed at mRNA and protein levels (Figure 3B). Hepatocyte‐specific Lgals4 overexpression led to a significant reduction in body weight gain and random blood glucose levels in HFD‐fed mice, whereas fasting glucose levels remained largely comparable to controls (Figure 3C–D). However, glucose tolerance and insulin sensitivity, as assessed by GTT and ITT, were not significantly altered (Figure S8A,B). Analysis of circulating lipid profiles revealed a trend towards reduced plasma triglyceride (TG) levels in the Lgals4‐overexpressing group, while total cholesterol (TC) levels remained unchanged (Figure 3E–F). Plasma ALT and AST levels were comparable between the AAV‐TBG‐Lgals4 and control groups (Figure S8C,D).

FIGURE 3.

Hepatocyte‐specific Lgals4 overexpression alleviates HFD‐induced hepatic steatosis. (A) Schematic representation of the experimental design. Adult C57BL/6J mice were injected with AAV‐TBG‐GFP (n = 5) or AAV‐TBG‐Lgals4 (n = 6) and subsequently fed a high‐fat diet (HFD) for 16 weeks. (B) Validation of Lgals4 overexpression in the liver by qPCR and Western blot analysis. HSP90 served as a loading control. (C and D) Body weight curve (C) and random/fasting blood glucose levels (D) of AAV‐GFP and AAV‐Lgals4 mice during 16 weeks of HFD feeding. (E and F) Plasma triglyceride (TG) (E) and total cholesterol (TC) (F) levels at the end of the 16‐week HFD challenge. (G and H) Quantitative analysis of absolute liver weight (G) and tissue‐to‐body weight ratios (H) for liver, epididymal white adipose tissue (eWAT), brown adipose tissue (BAT) and quadriceps (Quads). (I) Representative gross morphological images of livers from the indicated groups. (J) Representative H&E (top) and Oil Red O (bottom) staining of liver sections. Scale bars, 200 μm. (K) Histological scoring of liver sections, including steatosis grade, hepatocyte ballooning, inflammation and Non‐alcoholic Fatty Liver Disease Activity Score (NAS) (n = 5–6). (L) Quantitative analysis of Oil Red O‐positive area (%) in liver sections from the indicated groups (n = 5–6). (M and N) Biochemical quantification of hepatic TG (M) and TC (N) content. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by Student's t‐test.

Since progressive hepatic steatosis is the hallmark of MASLD that drives systemic metabolic collapse, we next examined the effect of Lgals4 on hepatic lipid accumulation. Notably, Lgals4 overexpression significantly decreased both absolute liver weight and the liver‐to‐body weight ratio, without affecting weights of epididymal white adipose tissue (eWAT), brown adipose tissue (BAT) and quadriceps muscle (Figures 3G,H and S8E). Consistent with these findings, gross morphological and histological analyses (H&E and Oil Red O) demonstrated a marked attenuation of hepatic steatosis, characterised by substantial reductions in lipid droplet size and total lipid area (Figure 3I–L). These improvements were corroborated by biochemical quantification, as hepatic triglyceride levels were significantly decreased in the Lgals4‐overexpressing group, while total cholesterol levels remained unchanged (Figure 3M,N). To confirm these effects were liver‐autonomous rather than secondary to systemic changes, we utilised a short‐term (10‐day) HFD model. In this setting, Lgals4 overexpression remained highly liver‐specific without detectable induction in extrahepatic tissues (Figure S8F). Notably, Lgals4 significantly alleviated steatosis, while body weight gain and food intake remained comparable between groups (Figure S8G,H). Furthermore, no significant changes were observed in hepatic Fgf21 expression, BAT thermogenic gene expression or adipose tissue morphology (Figure S8I–K). These findings suggest that Lgals4 improves hepatic steatosis without major remodelling of extrahepatic metabolic tissues.

3.4. Lgals4 Promotes Hepatic FAO by Activating PGC1α–PPARα

Building on its physiological activation by energetic stress and its anti‐steatotic effects, we next investigated how Lgals4 modulates hepatic lipid metabolism. In HFD‐fed mice, hepatocyte‐specific Lgals4 overexpression did not significantly alter key de novo lipogenesis genes (Fasn, Acc1, Srebp1c.) or major lipid uptake transporters (Cd36, Fabp1), although a modest reduction in Fatp5 was observed (Figure 4A,C).

FIGURE 4.

Lgals4 activates the hepatic PGC1α–PPARα programme to drive fatty acid oxidation. (A–C) qPCR analysis of genes involved in lipogenesis (A), fatty acid oxidation (FAO) (B) and lipid uptake (C) in the livers of HFD‐fed AAV‐GFP (n = 5) and AAV‐Lgals4 mice (n = 6). (D and E) Representative Western blots (D) and quantification (E) of key proteins in lipogenesis (FASN, SREBP1c) and fatty acid oxidation (PGC1α, PPARα) pathways. HSP90 served as a loading control (n = 4). Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by Student's t‐test or one‐way ANOVA with Tukey's correction for multiple comparisons.

However, Lgals4 overexpression markedly activated the FAO programme. Specifically, mRNA levels of the key regulators Pgc1a and Ppara, along with their downstream target Acox1, were significantly upregulated in Lgals4‐overexpressing livers (Figure 4B). Consistently, protein analyses confirmed increased levels of PGC1α and PPARα in the liver, whereas lipogenic proteins such as FASN and SREBP1C remained comparable between groups (Figure 4D,E). Genes involved in glycolysis, gluconeogenesis, glycogen metabolism, inflammation and fibrosis were largely unaffected (Figure S9A,B), indicating that Lgals4 exerts a relatively selective role in promoting hepatic lipid oxidation, thereby reducing hepatic lipid accumulation and promoting metabolic homeostasis.

3.5. Lgals4 Promotes Mitochondrial FAO and Reduces Lipid Accumulation In Vitro

To confirm the cell‐autonomous role of Lgals4, we challenged AML12 cells and HepG2 cells with 0.5 mM free fatty acids (FFAs; OA:PA = 2:1) to mimic MASLD in vitro. Efficient overexpression of Lgals4 in AML12 cells was validated by qPCR (Figure 5A). Functionally, Lgals4 overexpression significantly attenuated FFA‐induced intracellular triglyceride accumulation, lipid droplet size and abundance, as evidenced by BODIPY staining (Figure 5B,C).

FIGURE 5.

Lgals4 mitigates lipid accumulation in hepatocytes by selectively activating the FAO programme in vitro. (A) qPCR analysis of Lgals4 mRNA levels in AML12 cells transfected with empty vector (EV) or Lgals4 expression plasmids, under basal or FFA‐challenged (0.5 mM, 48 h) conditions (n = 3). (B and C) Intracellular triglyceride (TG) content (B) and representative images of BODIPY 493/503 staining (green) (C) in AML12 cells. Nuclei were counterstained with DAPI (blue). Scale bar, 50 μm. (D and E) Oxygen consumption rate (OCR) profiles (D) and quantified basal respiration, ATP production and maximal respiration (E) in FFA‐challenged AML12 cells measured by Seahorse XF analysis (n = 6). (F–H) qPCR analysis of genes involved in de novo lipogenesis (DNL) (F), fatty acid oxidation (FAO; G) and lipid uptake (H) in AML12 cells under the indicated conditions (n = 3). (I and J) Representative Western blots (I) and quantification (J) of key metabolic proteins in AML12 cells (n = 3). Tubulin served as a loading control. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by Student's t‐test or one‐way ANOVA with Tukey's correction for multiple comparisons.

To directly assess mitochondrial function supporting its potential involvement, we performed Seahorse‐based oxygen consumption rate (OCR) analysis. Lgals4 overexpression increased maximal respiratory capacity, with a trend towards elevated basal respiration, while ATP‐linked respiration remained largely unchanged (Figure 5D,E). These results indicate that Lgals4 enhances mitochondrial oxidative capacity under lipotoxic conditions.

Consistent with in vivo findings, Lgals4 did not significantly impact major DNL genes (Fasn, Acc1 and Srebp1c) or lipid uptake genes (Cd36 and Fatp5) (Figure 5F,H). In contrast, Lgals4 robustly activated the FAO programme, upregulating the expression of key regulators and enzymes governing mitochondrial and peroxisomal β‐oxidation, including Pgc1a, Ppara and Acox1, particularly under FFA challenge (Figure 5G). These transcriptional changes were corroborated at the protein level, where increased expression of PGC1α and PPARα was observed, while lipogenic proteins remained unchanged (Figure 5I,J). Similar effects were observed in HepG2 cells, where Lgals4 consistently attenuated FFA‐induced lipid accumulation (Oil Red O staining) and upregulated FAO‐related genes without affecting lipogenesis or lipid uptake (Figure S10A–E). Together, these data demonstrate that Lgals4 enhances mitochondrial oxidative capacity and selectively promotes FAO, thereby alleviating hepatocellular lipid accumulation and restoring hepatic lipid homeostasis under lipotoxic conditions.

3.6. Lgals4 Promotes Hepatic FAO in an AMPK‐Dependent Manner

Given that AMPK is a central regulator of cellular energy homeostasis and drives FAO via PGC1α and PPARα [32]. We therefore examined whether AMPK signalling contributes to Lgals4‐induced metabolic reprogramming. Immunofluorescence showed cytoplasmic localisation of LGALS4 in primary hepatocytes (Figure S10F). Notably, AAV‐mediated Lgals4 overexpression significantly increased the phosphorylation of AMPK at Thr172 and its downstream target ACC (Ser79) in the liver (Figure 6A,B). Consistent with our in vivo findings, Lgals4 overexpression in AML12 cells robustly enhanced the AMPK–ACC signalling cascade under FFA challenge (Figure 6C,D). These results suggest that Lgals4 triggers a metabolic shift towards FAO by activating the AMPK regulatory signalling.

FIGURE 6.

Lgals4 promotes hepatic FAO in an AMPK‐dependent manner. (A and B) Representative Western blots (A) and quantification (B) of phosphorylated AMPK (Thr172), total AMPK, phosphorylated ACC (Ser79) and total ACC in the livers of HFD‐fed AAV‐GFP and AAV‐Lgals4 mice (n = 4). (C and D) Representative Western blots (C) and quantification (D) of phosphorylated and total AMPK (Thr172) and ACC (Ser79) in FFA‐challenged AML12 cells (n = 3). HSP90 served as a loading control. (E) qPCR analysis confirming Lgals4 mRNA levels in AML12 cells after indicated treatments (n = 4). (F–H) qPCR analysis of genes involved in de novo lipogenesis (DNL) (F), fatty acid oxidation (FAO) (G) and lipid uptake (H) in AML12 cells treated with or without the AMPK inhibitor Compound C (10 μM) (n = 4). (I and J) Representative Western blots (I) and quantification (J) of P‐AMPK (Thr172), total AMPK and PGC1α protein levels in AML12 cells (n = 4). Data are presented as mean ± SD for (A and B) and mean ± SEM for the remaining panels. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by Student's t‐test or one‐way ANOVA with Tukey's correction for multiple comparisons.

To determine whether AMPK activation is required for Lgals4‐mediated metabolic reprogramming, we treated Lgals4‐overexpressing AML12 cells with the AMPK‐specific inhibitor, Compound C. Crucially, the induction of FAO‐related genes, including Pgc1a, Acox1 and Ppara, by Lgals4 was almost entirely abolished upon Compound C treatment, whereas the expression of genes involved in DNL and lipid uptake remained largely unaffected by either Lgals4 or AMPK inhibition (Figure 6E–H). Furthermore, immunoblotting analysis confirmed that the Lgals4‐induced increase in PGC1α protein levels was significantly suppressed upon AMPK inhibition (Figure 6I,J). Collectively, these findings demonstrate that Lgals4 drives a metabolic reprogramming towards enhanced FAO in an AMPK‐dependent manner, establishing the AMPK–PGC1α axis as a critical mediator of its metabolic effects.

3.7. Lgals4 Sustains AMPK–PGC1α Signalling and FAO Through Modulation of Cellular Energy Status

To further establish the physiological necessity of LGALS4 in hepatic lipid metabolism, we performed loss‐of‐function (LOF) experiments using siRNA‐mediated knockdown in AML12 cells and mouse primary hepatocytes (MPHs). Efficient silencing of Lgals4 was confirmed using three independent siRNAs in AML12 cells, with si‐RNA2 showing the highest efficiency and being selected for subsequent studies (Figure 7A,E,F,I).

FIGURE 7.

Endogenous Lgals4 sustains AMPK–PGC1α signalling and FAO through modulation of cellular energy status. (A) qPCR analysis of Lgals4 mRNA levels in AML12 cells transfected with three independent siRNAs targeting Lgals4 (si‐RNA1, si‐RNA2 and si‐RNA3) or negative control siRNA (si‐NC) (n = 3). (B and C) Intracellular triglyceride (TG) content (B) and representative images of BODIPY 493/503 staining (green) (C) in AML12 cells transfected with si‐NC or si‐Lgals4 (si‐RNA2) under basal or FFA‐challenged conditions. Nuclei were counterstained with DAPI (blue). Scale bar, 50 μm. (D) qPCR analysis of fatty acid oxidation (FAO)‐related genes in AML12 cells transfected with si‐NC or si‐Lgals4 under basal (left) or FFA‐challenged (right) conditions (n = 3). (E) Representative Western blots showing LGALS4, phosphorylated AMPK (Thr172), total AMPK and PGC1α protein levels in AML12 cells following Lgals4 knockdown. Tubulin served as a loading control. (F–H) qPCR analysis of Lgals4 mRNA levels (F), intracellular TG content (G) and expression of FAO‐related genes (H) in mouse primary hepatocytes (MPHs) transfected with si‐NC or si‐Lgals4 (si‐RNA2) (n = 3–5). (I) Representative Western blots showing LGALS4, phosphorylated AMPK (Thr172), total AMPK and PGC1α protein levels in MPHs transfected with si‐NC or si‐Lgals4. HSP90 served as a loading control. (J and K) Intracellular ATP levels and ADP/ATP ratios in AML12 cells following Lgals4 overexpression (J) or knockdown (K) (n = 4–6). Data are presented as mean ± SD for (F–H) and mean ± SEM for the remaining panels. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by Student's t‐test or one‐way ANOVA with Tukey's correction for multiple comparisons.

Functionally, Lgals4 knockdown modestly increased TG accumulation under basal conditions but markedly exacerbated lipid deposition under FFA challenge, as evidenced by quantitative TG measurements and BODIPY staining showing increased lipid droplet size and abundance (Figure 7B,C). This effect was consistently observed in primary hepatocytes, where Lgals4 silencing significantly elevated intracellular TG content (Figure 7G).

Consistent with our gain‐of‐function findings, the metabolic defect caused by Lgals4 knockdown was primarily attributed to an impaired FAO programme rather than changes in lipogenesis or uptake. In both AML12 cells and MPHs, si‐Lgals4 significantly reduced the expression of key FAO‐related genes, including Pgc1a, Ppara and Acox1, with more pronounced suppression under lipotoxic conditions (Figure 7D,H). In contrast, genes involved in de novo lipogenesis and fatty acid uptake remained largely unchanged following Lgals4 knockdown in both AML12 cells and MPHs (Figure S11A–D). Furthermore, Lgals4 knockdown attenuated AMPK phosphorylation (Thr172) and reduced the protein levels of PGC1α in both AML12 cells and MPHs (Figure 7E,I). Together, these findings demonstrate that endogenous Lgals4 is required for maintaining AMPK–PGC1α signalling, mitochondrial FAO and lipid homeostasis especially under lipotoxic stress.

To investigate the mechanism underlying LGALS4‐mediated AMPK activation, we examined whether LGALS4 directly interacts with components of the canonical AMPK activation machinery. Co‐immunoprecipitation assays in HEK293T cells revealed no detectable interaction between LGALS4 and AMPKα2 (Figure S11E). Similarly, semi‐endogenous immunoprecipitation assays in AML12 cells failed to detect an interaction between LGALS4 and the upstream AMPK kinase LKB1 (Figure S11F), suggesting that LGALS4 is unlikely to activate AMPK through direct physical association.

We therefore explored whether LGALS4 regulates AMPK activation by modulating cellular energy homeostasis. Notably, LGALS4 overexpression elevated the intracellular ADP/ATP ratio, accompanied by a modest decrease in absolute ATP levels (Figure 7J). Conversely, Lgals4 knockdown reduced the ADP/ATP ratio and slightly increased ATP levels (Figure 7K). These findings suggest that LGALS4 promotes a metabolically active state associated with elevated cellular energy demand, thereby providing a plausible mechanistic basis for AMPK activation. Together, these data support a model in which LGALS4 activates the AMPK–PGC1α axis through modulation of intracellular energy status rather than direct interaction with canonical AMPK signalling components.

4. Discussion

The liver acts as a central metabolic rheostat, maintaining systemic homeostasis by dynamically balancing energy storage and expenditure [42]. While the transition from physiological adaptation to pathological energy surplus drives the global rise of MASLD [11, 43], the endogenous mechanisms that orchestrate this metabolic switch remain poorly defined. Here, we identify LGALS4 as a previously unrecognised metabolic rheostat that senses energetic challenges and directs hepatic lipid flux towards oxidation. Integrated analysis of hepatic transcription under multifaceted energetic challenges shows that LGALS4 is uniformly induced by physiological energy deficits and serves as a critical brake against hepatic lipid accumulation. Our data indicate that LGALS4 functionally engages the AMPK–ACC cascade and the downstream PGC1α–PPARα programme, promoting hepatic FAO. These findings position the LGALS4–AMPK–PGC1α axis as a responsive regulatory network linking energetic stress to hepatic lipid metabolism. However, the precise biochemical mechanism by which LGALS4 influences cellular energy homeostasis and AMPK activation remains to be fully defined.

LGALS4 belongs to the galectin family of β‐galactoside‐binding proteins, which translate glycan‐encoded information into diverse cellular processes, including apoptosis, proliferation, migration, immune regulation and tissue homeostasis [44]. Although several galectins have increasingly been implicated in intracellular signalling and metabolic regulation [45, 46, 47], LGALS4 itself has been studied mainly in the gastrointestinal tract and in tumour biology [48, 49]. In hepatocellular carcinoma, galectin‐4 expression is associated with less aggressive disease and improved survival [50, 51]. Our findings therefore expand the functional repertoire of LGALS4 by identifying it as a regulator of hepatic metabolic adaptation and lipid handling, thereby linking a galectin family member to the control of FAO and steatosis.

A central conceptual advance of this study is the characterisation of a conserved hepatic transcriptional programme that governs the adaptive response to diverse energy‐demanding states. Despite the heterogeneity of stimuli such as cold exposure, fasting and exercise, these stimuli converge on a coordinated reprogramming of lipid metabolic pathways. Our findings position LGALS4 as a core, stimulus‐responsive effector at the intersection of these adaptive pathways, serving as an integral component of the hepatic machinery that prioritises lipid mobilisation and oxidation under metabolic demand. By identifying this conserved axis, we provide a conceptual framework for harnessing endogenous adaptive programmes to counteract the chronic energy surplus that defines MASLD.

Beyond its physiological induction, LGALS4 is reciprocally and dynamically regulated by the systemic energy state: it is robustly upregulated to meet energetic deficits but suppressed during chronic nutrient surfeit. Importantly, this suppression is functionally reversible upon metabolic restoration, highlighting the high regulatory plasticity of LGALS4. Crucially, the clinical relevance of our findings is underscored by the marked downregulation of LGALS4 in liver biopsies from MASLD patients, mirroring the metabolic derangements observed in HFD‐fed mice and suggesting a conserved pathological role. Hepatocyte‐specific restoration of LGALS4 significantly alleviates hepatic steatosis in HFD‐fed mice independently of changes in body weight or systemic glucose disposal (Figure 3). This supports a model in which LGALS4 acts as a metabolic rheostat, sensing systemic energetic cues to effectively drive a shift in hepatic bioenergetics towards enhanced FAO. Moreover, Seahorse‐based metabolic flux analysis revealed that this LGALS4‐driven reprogramming translates into increased mitochondrial respiratory capacity and efficient substrate clearance, thereby protecting hepatocytes from the lipotoxic stress induced by energy oversupply.

The molecular mechanism by which LGALS4 regulates lipid metabolism is centred on AMPK activation. In contrast to systemic metabolic modulators such as metformin, which has shown limited effects on liver histological outcomes in MASLD, LGALS4 exerts its effects through a distinct liver‐autonomous mechanism. Our study demonstrates that LGALS4 induces AMPK phosphorylation (Thr172), leading to ACC phosphorylation and activation of the PGC1α–PPARα axis to drive mitochondrial and peroxisomal FAO. AMPK inhibition largely abolishes LGALS4‐induced FAO gene expression, confirming its essential role in this process. To gain insight into the relationship between LGALS4 and AMPK activation, we examined whether LGALS4 physically associates with AMPKα2 or its upstream kinase LKB1. Co‐immunoprecipitation assays failed to detect stable interactions between LGALS4 and either protein under our experimental conditions. We therefore evaluated cellular energy status and observed that LGALS4 overexpression was accompanied by an increased ADP/ATP ratio, whereas LGALS4 knockdown produced the opposite effect. These alterations coincided with corresponding changes in AMPK activity, suggesting a potential link between LGALS4 expression, cellular energy status and AMPK signalling. However, the present data do not establish a direct causal mechanism by which LGALS4 regulates intracellular energy homeostasis. Thus, while our findings are consistent with a model in which LGALS4 is associated with bioenergetic remodelling that accompanies AMPK activation, the precise biochemical events connecting LGALS4 to cellular energy sensing and AMPK regulation remain to be determined.

Another noteworthy aspect of LGALS4 function is its remarkable metabolic selectivity. Unlike broad‐acting AMPK activators or PPARα agonists, which can perturb glucose metabolism or trigger systemic inflammation [52, 53, 54], LGALS4 primarily targets hepatic lipid content without significantly altering insulin sensitivity (Figure S8). Our findings confirm a liver‐autonomous role for LGALS4, as its anti‐steatotic effects in short‐term HFD models occurred independently of changes in food intake, adipose remodelling, thermogenesis, systemic endocrine signalling or its extrahepatic expression. This precision allows LGALS4 to act as a metabolic switch, directing lipid flux towards catabolism while preserving broader hepatic homeostasis. By minimising systemic off‐target effects, LGALS4 emerges as a promising candidate for the precision treatment of MASLD.

In conclusion, our findings establish LGALS4 as a pivotal hepatic regulator that senses energetic stress and selectively redirects lipid flux from storage towards oxidation. By activating the AMPK–PGC1α–PPARα axis, LGALS4 recruits physiological adaptive mechanisms to clear pathological energy surplus, protecting hepatocytes from lipotoxicity without disrupting systemic homeostasis. These features position the LGALS4–AMPK axis as a promising, metabolically selective target to reprogramme energy balance and counteract chronic nutrient oversupply in MASLD.

Author Contributions

Die Hu: writing – review and editing, writing – original draft, visualisation, resources, methodology, investigation, formal analysis, data curation, conceptualisation. Jingchi Li, Hailong Cui, Xiaofu Zhang, Sihang Fang and Wen Wen: data curation, investigation, methodology, resources. Xiaoying Li: writing – review and editing, supervision, project administration, funding acquisition. Qiuyu Wang: writing – review and editing, supervision, project administration, funding acquisition. All authors reviewed and edited the manuscript and approved the final version.

Funding

This work was supported by China Postdoctoral Science Foundation (2023M730660) and National Natural Science Foundation of China (82270919, 82401004, 82470844).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Transcriptomic profiling of liver under cold exposure.

Figure S2: Transcriptomic analysis of liver following CL316243 treatment.

Figure S3: Hepatic transcriptomic landscape in response to fasting (GSE118787).

Figure S4: Transcriptomic analysis of liver under calorie restriction (GSE299228).

Figure S5: Convergent activation of hepatic fatty acid oxidation across diverse energetic challenges.

Figure S6: Functional annotation of a conserved hepatic energy‐responsive programme identifying LGALS4.

Figure S7: Dynamic regulation of hepatic Lgals4 in response to physiological energy demands.

Figure S8: Assessment of systemic metabolic parameters, tissue weights and extrahepatic characterisation following Lgals4 overexpression.

Figure S9: Assessment of systemic glucose metabolism and hepatic inflammatory profile following Lgals4 overexpression.

Figure S10: Lgals4 alleviates lipid accumulation and induces the FAO programme in HepG2 cells.

Figure S11: Lgals4 regulates lipid homeostasis independently of de novo lipogenesis or fatty acid uptake and activates AMPK without direct physical interaction.

Table S1: Primer sequences for RT‐qPCR.

Table S2: Key drugs, reagents and antibodies used in this study.

Table S3: siRNA oligonucleotide sequences.

Table S4: Differential expression analysis of the 21‐gene signature across various energetic challenges.

Table S5: Clinical characteristics of human liver samples.

Hu D., Li J., Cui H., et al., “Hepatic LGALS4 Alleviates Diet‐Induced Steatosis by Promoting AMPK‐Dependent Fatty Acid Oxidation,” Diabetes, Obesity and Metabolism 28, no. 9 (2026): 7875–7891, 10.1111/dom.70989.

Data Availability Statement

The RNA‐Seq data was uploaded in the SRA database under accession code (PRJNA1442354: Transcriptomic profiling of mouse liver under cold exposure and CL316243 Treatment). All other data are available in the Methods section and Supporting Information or from the corresponding authors upon request.