MEN1/menin deficiency suppresses hepatocellular carcinogenesis via disrupting mitophagy-mediated mitochondrial homeostasis

ABSTRACT

Hepatocellular carcinoma (HCC) is a highly lethal liver cancer with complex pathogenesis intertwined with metabolic and mitochondrial dysfunction. MEN1/menin is a protein with context-dependent functions in liver diseases. While MEN1 has been linked to HCC progression and mitochondrial homeostasis, its precise regulatory mechanism in these processes remains incompletely understood. Here, we report that MEN1 localizes to the outer mitochondrial membrane (OMM) in HCC cells, which is mediated by its N-terminal mitochondrial targeting sequence and the TOMM20 translocase complex. In a genetically engineered DEN- and CCl₄-induced HCC mouse model, hepatocyte-specific men1 deficiency significantly suppressed tumorigenesis, a phenotype associated with impaired mitochondrial homeostasis. Mechanistically, MEN1 deficiency disrupted mitochondrial function by manifesting as promoted mitochondrial fission, impaired oxidative phosphorylation, reduced ATP levels, and elevated reactive oxygen species during energy stress. Critically, MEN1 loss inhibited mitophagy via downregulating the PINK1-PRKN/Parkin pathway, which impaired clearance of dysfunctional mitochondria and promotes their cytotoxic accumulation. Moreover, MEN1 expression was upregulated in human HCC tissues, correlated with poor clinical outcomes and was positively associated with autophagy signatures. Notably, pharmacological activation of mitophagy reversed the tumor-suppressive effects of MEN1 deficiency in vitro and in vivo. These findings identified a noncanonical role of mitochondrial MEN1 in driving HCC progression via regulating mitophagy homeostasis, and highlight the MEN1-mitophagy axis as a potential therapeutic target for HCC.

Abbreviations: Alb-Cre: albumin promoter-driven recombinase Cre; Baf A1: bafilomycin A₁; CCCP: carbonyl cyanide m-chlorophenyl hydrazone; CCl₄: carbon tetrachloride; Co-IP: co-immunoprecipitation; CQ: chloroquine; DEN: diethylnitrosamine; DNM1L: dynamin 1 like; DQ-BSA: self-quenched BODIPY-conjugated bovine serum albumin; Gal: galactose; GOT1/AST: glutamic-oxaloacetic transaminase 1; GPT/ALT: glutamic–pyruvic transaminase; HCC: hepatocellular carcinoma; WT: wild type; HMKO: hepatocyte-specific men1 knockout; IF: immunofluorescence; IHC: immunohistochemistry; IMM: inner mitochondrial membrane; KEGG: Kyoto Encyclopedia of Genes and Genomes; MEFs: mouse embryonic fibroblasts; MEN1-FL: full-length MEN1; MFF: mitochondrial fission factor; MFN1: mitofusin 1; MTS: mitochondrial targeting sequence; OCR: oxygen consumption rate; OMM: outer mitochondrial membrane; OXPHOS: oxidative phosphorylation; PINK1: PTEN induced kinase1; PRKAA1: protein kinase AMP-activated catalytic subunit alpha 1; PRKN: parkin RBR E3 ubiquitin protein ligase; qPCR: RNA extraction and quantitative polymerase chain reaction; RNA-seq: RNA-sequencing; ROS: reactive oxygen species; shMEN1: small hairpin RNA-mediated MEN1 knockdown; TCGA: The Cancer Genome Atlas; TEM: transmission electron microscopy; TOMM20: translocase of outer mitochondrial membrane 20.

Introduction

Liver cancer is the third leading cause of cancer-related mortality worldwide, with approximately 80% being hepatocellular carcinoma (HCC) [1]. Despite advances in diagnosis and treatment, prognosis remains poor owing to the complex, multifactorial pathogenesis of liver cancer [2]. Given that the liver functions as a primary detoxification organ and central metabolic hub, targeting the coordinated interplay between metabolic and detoxification pathways represents a promising therapeutic strategy for HCC.

The MEN1 gene (Men1 in mice) encodes a nuclear protein whose inactivation causes multiple endocrine neoplasia type 1 (MEN1) [3]. MEN1 regulates multiple cellular processes through interactions with functionally distinct proteins, and its functional mutation contributes to the initiation and progression of various human diseases [4–7]. Traditionally characterized as a scaffold protein that modulates gene expression by interacting with chromatin modifiers and transcription factors [8,9], MEN1 exhibits striking stage-dependent duality in liver diseases. Nuclear MEN1 acts as a hepatic protector, inhibiting hepatic steatosis through the SIRT1-CD36-mediated lipid metabolism in aging mice [10]. Paradoxically, during HCC tumorigenesis, MEN1 functions as an oncoprotein through the MLL-mediated trimethylation of histone H3 lysine 4 (H3K4me3)-dependent YAP1 epigenetic activation [11,12]. This functional switch from “hepatoprotection” to “carcinogenesis promotion” suggests tight associations with microenvironmental changes across disease stages, including simple fatty liver, cirrhosis, and HCC. However, whether this functional heterogeneity stems from dynamic shifts in MEN1’s subcellular localization remains unclear.

Emerging evidence shows MEN1 has noncanonical functions beyond its known nuclear roles, with distinct activities linked to specific subcellular compartments. For instance, Ren et al. reported that luminal breast cancer cell lines predominately express cytoplasmic MEN1 [13], and such expression associates with poorer prognosis in sporadic pulmonary carcinoids and neuroendocrine-featured breast carcinomas [13,14]. Functionally, cytoplasmic MEN1 suppresses pancreatic neuroendocrine tumor progression by facilitating monounsaturated fatty acid metabolism and inhibiting MTOR (mechanistic target of rapamycin kinase) activation [15]. At the cell membrane, membrane-associated MEN1 more potently represses RAS signaling through regulation of SOS (SOS Ras/Rac guanine nucleotide exchange factor) [16,17]. These observations raise the possibility that MEN1’s subcellular localization mediates interactions with key energy metabolic organelles like mitochondria, a link reinforced by recent studies showing MEN1 exerts opposing pro- and anti-tumor effects in immunodeficient versus immunocompetent mouse models via the CGAS-STING1 pathway [18]. Notably, CGAS, a sensor of mitochondrial DNA released from damaged mitochondria into the cytosol [19], provides a molecular bridge between MEN1’s subcellular distribution and tumorigenesis through this signaling axis. Such CGAS-mediated sensing of mitochondrial DNA highlights functional crosstalk between MEN1’s localization and mitochondrial integrity, prompting exploration of MEN1’s role in mitochondrial biology.

While MEN1 has been linked to energy metabolism regulation [20,21], its function in mitochondrial homeostasis, especially in the context of liver cancer, has remained undefined. Here, we uncover an unrecognized noncanonical role of MEN1 in the mitochondria of HCC cells and show that MEN1 localizes to the outer mitochondrial membrane (OMM) to promote cell proliferation. In chemically induced HCC mouse models, we show that hepatocyte-specific men1 knockout suppresses tumor progression. Mechanistically, MEN1 deficiency triggers mitochondrial dysfunction, accelerates stress-induced mitochondrial fission, and impairs mitophagy by disrupting the PINK1 (PTEN induced kinase1)-PRKN (parkin RBR E3 ubiquitin protein ligase) pathway. Our study revealed a previously unappreciated mechanism by which MEN1 deletion perturbs mitophagic homeostasis, causing the accumulation of cytotoxic damaged mitochondria that suppress HCC progression. These findings underscored that MEN1’s subcellular localization is critical for regulating crosstalk between mitochondrial homeostasis and autophagy, thereby sustaining tumor growth and offering a potential therapeutic target for HCC.

Results

MEN1 is localized in the outer mitochondrial membrane of liver cancer cells

Quantitative proteomic analysis for MEN1 interactors co-immunoprecipitated (co-IP) from Hep3B whole-cell lysates found that 63 out of 878 binding partners were mitochondrial proteins (Figure S1A), which are mainly involved in mitochondrial translation processes (Figure S1B), suggesting that MEN1 may localize to the mitochondrion. To examine the mitochondrial association of MEN1, we performed subcellular fractionation of nuclear, cytosolic, and mitochondrial fractions from liver cancer cells. Immunoblotting showed that MEN1 was present in the mitochondria of Hep3B and Huh7 cells (Figure 1(A), S1C). Moreover, we isolated mitochondrial fractions of mouse embryonic fibroblasts (MEFs) from men1 knockout (men1^Δ/Δ) and wild type (Men1^{flox/flox [f/f]}) littermate mice and revealed that endogenous MEN1 was detected in the mitochondrial fraction from only the Men1^f/f MEFs (Figure 1(B), S1C). Immunofluorescence (IF) localization analysis using a MEN1 antibody and MitoTracker Red (a mitochondrial fluorescence probe) further confirmed that MEN1 was localized to mitochondria in Men1^f/f MEFs (Figure 1(C)). Similarly, mitochondrial localization of MEN1 was also observed in HCC cells overexpressing flag-tagged MEN1 (oeMEN1) (Figure S1D-F).

Figure 1.

MEN1 is localized in the outer mitochondrial membrane of liver cancer cells. (A) Immunoblotting of MEN1 protein in whole-cell lysate (wcl), nuclei (Nuc), cytosol (cyto) and mitochondria (Mito) fractions from Hep3B and Huh7 cells. (B) Immunoblotting of MEN1 protein in cytosolic and mitochondrial fractions in Men1^f/f and men1^Δ/Δ MEFs. (C) Representative immunofluorescent staining images and fluorescence colocalization analysis of MEN1 (green) and MitoTracker red (red) in Men1^f/f and men1^Δ/Δ MEFs. Nuclei were stained with DAPI (blue); scale bar: 5 µm. (D) Schematic diagram of full-length MEN1 (MEN1-FL) and its four truncated mutants. (E) Immunoblotting of His-tagged proteins in HEK-293T cells transfected with plasmids encoding MEN1-FL or its truncated mutants for 48 h. (F) Representative immunofluorescent staining images of His-tagged proteins (green) and MitoTracker red (red) in Hep3B cells transfected with MEN1-FL or its truncated mutants for 48 h. Scale bar: 20 µm. (G) Quantitative analysis of mitochondrial (mito-MEN1) and nuclear MEN1 (nuc-MEN1) fluorescence intensity from (F) (n = 15 fields). (H) Representative immunofluorescent staining images of His-tagged proteins (red) and TOMM20 (green) in HEK-293T cells transfected with plasmids encoding MEN1-FL, NTD-MTS WT and NTD-MTS^L3G, NTD-MTS^R14G, NTD-MTS^K4,K8G, NTD-MTS[4 G] mutants for 48 h. Scale bar: 20 µm. (I) Bar chart depicts the fluorescence colocalization analysis of His-tagged and TOMM20 proteins from (H) (n = 15 fields). (J) Immunoblotting of indicated proteins in mitochondrial fractions from Hep3B and Huh7 cells, followed by treatment with proteinase K (Pro.K), 1% Triton X-100 (tx-100) alone or their combination for 30 min. (K) Immunoblotting of indicated proteins in mitochondrial fractions from Hep3B and Huh7 cells, followed by treatment with proteinase K at different concentrations. (L) Mitochondrial fractions purified from Hep3B cells were immunoprecipitated with MEN1 antibody followed by immunoblotting analysis of MEN1 and TOMM20 proteins. (M and N) Immunoblotting and quantitative analysis of TOMM20, TOMM70 and MEN1 proteins in nuclei and mitochondrial fractions from Hep3B cells transfected with siTOMM20 or siTOMM70 for 48 h (n = 3 biological replicates). TUBA was used as a cytosolic loading control, and LMNB as a nuclear marker. Statistical analyses were performed using one-way ANOVA (G and I) and two-tailed unpaired Student’s t test (M and N); *p < 0.05, **p < 0.01, ***p < 0.001, and ns, not significant.

Protein sequence analysis using the MitoProt database revealed that MEN1 harbors a putative mitochondrial targeting sequence (MTS) with a cleavage site, spanning amino acid residues 1 to 15 at the N terminus with a probability score of 0.163 (Figure S1G). This sequence is highly conserved across species and enriched in hydrophobic amino acids and positively charged hydrophilic residues (Figure S1G), features that facilitate mitochondrial localization by forming an amphipathic helix. To verify the MTS functionality, we generated four 6×His-tagged truncated constructs of MEN1, each lacking specific domains based on the protein’s domain organization (Figure 1(D,E)). IF staining showed that mitochondrial distribution of the ΔNTD variant that lacks the predicted MTS was significantly reduced compared to the full-length MEN1 (MEN1-FL), whereas no significant differences in the mitochondrial MEN1 (mito-MEN1) levels were observed among MEN1-FL, ΔThumb, ΔPalm, and ΔFingers (Figure 1(F,G)). Moreover, the nuclear localization (nuc-MEN1) of ΔFingers (lacking two nuclear localization signals) was markedly lower than that of MEN1-FL or the other three truncations (Figure 1(F,G)). To further dissect the critical residues within the MTS that mediate MEN1’s mitochondrial localization, we created His-tagged constructs of the NTD harboring point mutations within the MTS and performed IF staining with anti-His antibody and TOMM20 (translocase of outer mitochondrial membrane 20) (Figure 1H). Pearson’s correlation analysis revealed that the mitochondrial colocalization efficiency of two mutants-the double mutant NTD-MTS^K4G,K8G and the quadruple mutant L3G R14G K4G K8G (NTD-MTS[4 G])-was significantly reduced relative to MEN1-FL, NTD-MTS wild type (WT), and single mutants NTD-MTS^L3G and NTD-MTS^R14G (Figure 1(I)). In contrast, NTD-MTS^L3G and NTD-MTS^R14G showed only a slight reduction in mitochondrial localization compared to MEN1-FL or NTD-MTS WT, with no significant differences observed (Figure 1(I)). These findings indicate that the positively charged hydrophilic residues in the NTD-MTS domain are crucial for MTS-mediated mitochondrial targeting of MEN1.

To determine the sub-mitochondrial localization of MEN1, isolated mitochondria from liver cancer cells were treated with Triton X-100 or subjected to osmotic stress combined with proteinase K digestion (Figure S1H). In intact mitochondria, MEN1 was sensitive to proteinase K digestion, a pattern comparable to that of OMM proteins such as MFN1 (mitofusin 1) and TOMM20 (but not VDAC1 [voltage dependent anion channel 1]), whereas inner mitochondrial membrane (IMM) proteins (OPA1 [OPA1 mitochondrial dynamin like GTPase] or COX4I1/COX4 [cytochrome c oxidase subunit 4I1]) were protected from proteinase K digestion (Figure 1(J,K), Figure S1I,J). When both OMM and IMM were broken by chemical dissolution or physical stress, MEN1 exhibited significantly higher sensitivity to proteinase K digestion, consistent with the pattern observed for the OMM proteins (Figure 1(J,K), Figure S1I,J), suggesting that MEN1 is localized to the OMM. Co-IP analysis of mitochondrial fractions demonstrated an association between MEN1 and TOMM20, but not with TOMM70 (data not shown), which was further validated by co-IP with anti-TOMM20 antibody (Figure 1(L), S1K). Glutathione S-transferase (GST) affinity-isolation assays subsequently confirmed that the interaction between MEN1 and TOMM20 is indirect (Figure S1L).

As a core component of the OMM translocase complex, TOMM20 binds to the N terminal MTS of proteins to mediate nuclear-encoded protein transport into the mitochondria [22]. Small interfering RNA (siRNA)-mediated TOMM20 knockdown (siTOMM20) in Hep3B cells prominently reduced the protein abundance of mitochondrial MEN1 (Figure 1(M)), while no reduction was observed in siTOMM70-treated cells (Figure 1(N)). Collectively, these findings demonstrate that MEN1 is translocated to the OMM and that TOMM20 plays a critical role in this translocation process.

MEN1 deficiency impaired mitochondrial function in liver cancer cells

Since mitochondria are the energy factories and major metabolic hubs, we hypothesized that mitochondrial MEN1 regulates cellular energy metabolism. To understand the metabolic processes controlled by MEN1 during HCC progression, hepatocyte-specific men1 knockout mice (HMKO) were generated by crossing Men1^f/f mice with Alb (albumin) promoter-driven recombinase Cre (Alb-Cre) mice. Men1^f/f (WT, without Cre) and HMKO mice underwent an 8-month injection with diethylnitrosamine (DEN) and carbon tetrachloride (CCl₄) to induce HCC models, followed by untargeted metabolomics analysis of liver tumors (Figure S2A-C). The metabolic profiles of liver tumors from WT and HMKO mice were significantly distinct, as demonstrated by hierarchical clustering (Figure 2(A)) and principal component analysis (Figure S2D). The largest category of differential metabolites was amino acids (18.3%), followed by organic acids (15.3%), carbohydrates (11.3%), and lipids (11.3%) (Figure 2(B)). A total of 1,265 ions assigned to known metabolites, of which 300 metabolites exhibited significant abundance changes in HMKO tumors compared with WT tumors, including 157 downregulated and 143 upregulated metabolites (Figure S2E,). Metabolic pathway enrichment analysis further revealed that energy metabolism-related pathways-including glycolysis (a major adenosine triphosphate [ATP]-generating pathway), glycine-threonine metabolism, and the pentose phosphate pathway-were dramatically enriched in HMKO tumors (Figure 2(C)), indicating that MEN1 deficiency disrupts key energy metabolic pathways and causes mitochondrial dysfunction.

Figure 2.

MEN1 deficiency impaired mitochondrial function in liver cancer cells. (A) Heatmap showing metabolic profiles of liver tumor tissues from 8-month-old WT and HMKO mice after EDN and CCl₄ induction. (B) Pie chart illustrating the proportion of differential metabolites in liver tumor tissues from WT and HMKO mice. (C) Metabolic pathway enrichment analysis (MPEA) of differential metabolites in liver tumors from WT and HMKO mice. (D) Immunoblotting for MEN1 protein in shNC and shMEN1 Hep3B or Huh7 cells. (E) Representative transmission electron microscopy images showing mitochondrial morphology in shNC and shMEN1 Huh7 cells. Red arrows, mitochondrial cristae; scale bar: 500 nm. (F) Bar graphs depict the mitochondrial number (n = 3 biological replicates), cristae numbers per mitochondrion (n = 30 fields) and mitochondrial circularity in shNC and shMEN1 Huh7 cells from (E). (G) Representative flow cytometry histograms and quantification of ANXA5 -FITC staining in shNC and shMEN1 Hep3B cells with or without galactose (4.5 g/L) treatment for 36 h (n = 3 biological replicates). (H) Cell viability of shNC and shMEN1 Hep3B cells in galactose medium (4.5 g/L) was examined and normalized to that under normal glucose condition (n = 3 assays). (I) Analysis of mitochondrial OCR in shNC and shMEN1 Hep3B cells with or without galactose (4.5 g/L) for 36 h before OCR detection. Dotted lines indicate the time points of adding oligomycin (O), FCCP (F) and rotenone plus antimycin (R+A). (J and K) Mitochondrial membrane potential (∆Ψm) detected with JC-1 probe by flow cytometric analysis (J) in shNC and shMEN1 Hep3B cells with or without galactose (4.5 g/L) treatment for 36 h and quantification (K) (n = 3 biological replicates). Normal ∆Ψm is indicated by JC-1 aggregates while depolarized membrane potential is indicated by JC-1 monomers. (L) ATP levels in shNC and shMEN1 Hep3B cells with or without low-glucose (1.5 g/L) treatment at indicated time points (n = 3 biological replicates). (M) ROS levels detected by DCFH-DA probe using flow cytometry in shNC and shMEN1 Hep3B cells with or without galactose (4.5 g/L) treatment for 36 h and quantification of DCFH-DA (n = 3 biological replicates). (N) Heatmap showing transcriptional downregulation of genes involved in electron transport, fatty oxidation and mitochondrial biogenesis in liver tumor tissues from WT and HMKO mice (n = 3 biological replicates). (O) qPCR analysis of mRNA expression of indicated differentially expressed genes in HMKO tumors versus WT tumors (n = 7 mice per genotype). Statistical analyses were performed using two-tailed unpaired Student’s t test; *p < 0.05, **p < 0.01, ***p < 0.001, and ns, not significant.

Transmission electron microscopy (TEM) showed that small hairpin RNA-mediated MEN1 knockdown (shMEN1) in Hep3B cells (Figure 2(D)) induced a marked increase in mitochondrial number, reduction in cristae number, and decrease in mitochondrial circularity (Figure 2(E,F)). These structural abnormalities are well-documented hallmarks of impaired mitochondrial oxidative phosphorylation (OXPHOS) function. To confirm functional energy defects, galactose (Gal) was used to enforce mitochondrial ATP production [23]: when glucose is replaced with galactose, cells cannot rely on glycolysis and must use mitochondria for OXPHOS. Flow cytometry analysis of ANXA5/annexin V (a cell death marker) showed minimal cell death in shNC and shMEN1 Hep3B cells grown under normal glucose conditions (Gal 0 h) (Figure 2(G)). Although galactose induced toxicity in both liver cancer cell lines, MEN1-deficient cells exhibited strikingly higher mortality than shNC Hep3B cells (Figure 2(G)). Similar results were observed in MEN1-knockdown Huh7 cells under low glucose conditions (Figure 2(D), S2F). Cell functional assays further confirmed that MEN1 deficiency decreased cell viability and proliferation under galactose exposure or glucose limitation, whereas MEN1 overexpression promoted cellular growth (Figure 2(H), Figure S2G,H). These data suggest that MEN1-deficient liver cancer cells are more sensitive to energy stress, consistent with mitochondrial function impairment.

We next measured mitochondrial respiration using Seahorse analysis to assess oxygen consumption rate (OCR), a direct readout of OXPHOS activity. As shown in Figure 2(I), when cells were grown in normal glucose medium, a moderate OCR difference was detected between shNC and shMEN1 Hep3B cells, suggesting that mitochondrial energy imbalance caused by MEN1 deficiency can be partially compensated by glycolysis activation when the energetic demand is low under basal conditions [24]. However, when cells were suffered from stress of galactose treatment, both basal and maximal OCRs were significantly decreased in MEN1-deficient Hep3B cells compared with those in shNC cells (Figure 2(I)), consistent with the reduced cell survival rate in the MEN1-knockdown Hep3B cells (Figure 2(G)). A similar reduction in basal and maximal OCRs was observed in shMEN1 Huh7 cells exposed to low-glucose stress (Figure S2I). Again, the mitochondrial functional defect induced by MEN1 deficiency mirrored that observed in Hep3B cells, that is, the OCR discrepancy was most pronounced under energetic stress conditions (Figure S2I). In further support of the role of MEN1 in maintaining mitochondrial function under energy stress, MEN1 knockdown liver cancer cells exhibited significantly decreased mitochondrial membrane potential (∆Ψm, measured by JC-1 uptake via flow cytometry) and intracellular ATP levels under both galactose and low-glucose conditions (Figure 2(J-L)). Treatment with carbonyl cyanide m-chlorophenyl hydrazone (CCCP) markedly diminished the JC-1 fluorescence signals (Figure S2J,K), confirming mitochondrial insufficiency. Additionally, MEN1 knockdown increased basal reaction oxygen species (ROS) levels and accelerated ROS generation triggered by galactose exposure or Rosup (a ROS inducer), as measured by flow cytometry using the fluorescent probe DCFH-DA (Figure 2(M), S2L).

To validate MEN1’s regulatory role in mitochondrial function and energetic homeostasis at the molecular level, we analyzed the gene expression profiles in liver tumors from WT and HMKO mice. RNA-sequencing (RNA-seq) analysis identified 891 differentially expressed genes (DEGs) with aberrant gene expression patterns (false discovery rate [FDR]<0.05) (Figure S2M), which were highly enriched in mitochondria-mediated energy metabolism processes (e.g., carboxylic acid metabolism, oxidation-reduction, long-chain fatty acid metabolism) (Figure S2N). A heatmap displayed broad transcriptional downregulation of genes involved in electron transport (mt-Co2, mt-Co3, mt-Nd3 and mt-Nd4), fatty oxidation (Ppara, Cpt1a, Cpt1b and Cpt1c), and mitochondrial biogenesis (Ppargc1a and Nfe2l1) in HMKO tumors (Figure 2(N)). Quantitative polymerase chain reaction (qPCR) analysis further validated decreased mRNA expression of these genes in HMKO tumors (Figure 2(O)), consistent with compromised mitochondrial function in the absence of MEN1.

Together, these data demonstrate that mitochondrial MEN1 is critical for maintaining mitochondrial homeostasis, and its loss disrupts energy metabolism, thereby suppressing HCC progression.

MEN1 deficiency accelerates stress-induced mitochondrial fission of liver cancer cells

Mitochondria homeostasis is maintained via dynamic fusion and fission; dysregulation of this balance impairs mitochondrial function and contributes to cancer progression [25]. Based on the observed mitochondrial dysfunction in MEN1-deficient liver cancer cells, we hypothesize that MEN1 may regulate mitochondrial dynamics under energy stress. We first characterized mitochondrial morphology in liver tissues from WT and HMKO mice with DEN- and CCl₄-induced HCC. TEM revealed that mitochondria within tumors of WT and HMKO mice showed marked swelling, severe cristae and matrix disruption, and increased mitochondrial area and perimeter (Figure 3(A-C)). The mitochondrial number in WT tumors was dramatically higher than that in nontumor livers, whereas no further increase was observed in HMKO tumors relative to nontumor counterparts (Figure 3(D)). Notably, whether in nontumor or tumorous livers, HMKO mice showed a significantly more mitochondrial number than WT mice whereas their mitochondrial area and perimeter were mildly smaller (Figure 3(A-D)). This increase in mitochondrial number driven by men1 deletion occurred independently of mitochondrial mass changes indicated by similar VDAC1 expression levels in nontumor or tumorous livers (Figure S3A). Consistently, HMKO mice showed increased mitochondrial-endoplasmic reticulum (ER) contacts (mitochondria-associated membranes, MAMs), which are contact sites regulating mitochondrial fission (Figure 3(E,F)) [26], suggesting exacerbated mitochondrial fission and potential fragmentation.

Figure 3.

MEN1 deficiency accelerates stress-induced mitochondrial fission of liver cancer cells. (A-D) Representative TEM images showing mitochondria in normal and tumor regions of the liver tissues from WT and HMKO mice (A). Quantitative analysis of mitochondrial area (B) (n = 20 per group), perimeter (C) (n = 15 per group), and number per cell (D) (n = 12 fields). Scale bar: 500 nm; m, mitochondria; red arrows, damaged mitochondria. (E and F) Representative TEM images of normal and tumor regions of liver tissues from WT and HMKO mice (F); quantification for percentage of mitochondria-ER contacts is shown (E) (n = 12 fields per group). Scale bar: 100 nm; red arrows indicate the contact sites. (G) Representative immunofluorescent staining images of MtioTracker red showing mitochondrial morphology in shNC and shMEN1 Hep3B cells with or without galactose (4.5 g/L) treatment for 36 h. Scale bars: 20 µm or 5 µm. (H and I) Quantitative analysis of mitochondrial number (H) and mean branch length of mitochondria per cell (I) from (G) (n = 45 cells per condition). (J) Immunoblotting of indicated proteins in whole-cell lysates from shNC and shMEN1 Hep3B and Huh7 cells with or without galactose treatment for 36 h. (K) immunoblotting of indicated proteins in whole-cell lysates from vector and oeMEN1 Hep3B cells with or without galactose treatment for 36 h. (L) Immunoblotting of indicated proteins in mitochondrial (Mito) and cytosolic (cyto) fractions from shNC and shMEN1 Hep3B cells with or without galactose treatment for 36 h. (M) Quantitative analysis of DNM1L, p-DNM1L (Ser616), MFF, and p-MFF (Ser172) expression in mitochondria relative to cytosol derived from (L) (n = 3 biological replicates). (N) Representative immunofluorescent staining images of DNM1L (green) and COX4I1 (red) in shNC and shMEN1 Hep3B cells with or without galactose treatment for 36 h. Nuclei were stained with DAPI (blue); scale bar: 40 µm. (O) Bar chart depicts the percentage of DNM1L colocalized with COX4I1 in mitochondria (N) (n = 10 fields per condition). (P) Whole-cell lysates from shNC and shMEN1 Hep3B cells were immunoprecipitated with anti-MFF or anti-IgG antibodies followed by immunoblotting analysis of the indicated proteins. (Q) Immunoblotting of indicated proteins in cytosolic and mitochondrial fractions from primary Men1^f/f and men1^−/− hepatocytes with or without low-glucose (1.5 g/L) treatment for 36 h. Statistical analyses were performed using one-way ANOVA (B-E) and two-tailed unpaired Student’s t test (H, I, M and N); *p < 0.05, **p < 0.01, ***p < 0.001, and ns, not significant.

Using the mitochondrial-targeting fluorescent probe MitoTracker Red, we found that MEN1-knockdown liver cancer cells had increased numbers of individual mitochondria and decreased mean branch length, with these effects further enhanced by galactose treatment or glucose starvation (Figure 3(G-I), Figure S3B,C), indicating that MEN1 inactivation promotes mitochondrial fission. Immunoblotting showed that MEN1 deletion markedly upregulated the expression of fission-related proteins (phosphorylated DNM1L/DRP1 [dynamin 1 like; p-Ser616], phosphorylated MFF [mitochondrial fission factor; p-Ser172], MIEF1/MID51, FIS1) but barely affected fusion-associated proteins (OPA1, MFN1) (Figure 3(J)). MEN1-deficient HCC cells exhibited enhanced galactose-induced fission protein expression (Figure 3(J)), with similar increases observed in Hep3B and Huh7 cells when cells were grown in medium with glucose starvation (Figure S3D,E). Conversely, MEN1 overexpression attenuated the protein levels of p-MFF (Ser172), p-DNM1L (Ser616), and MIEF1 in galactose-treated Hep3B cells and glucose-limited Huh7 cells (Figure 3(K), S3F), confirming that loss of MEN1 drives mitochondrial fission.

It is well-established that MFF phosphorylation by activated PRKAA1/AMPKɑ1 (protein kinase AMP-activated catalytic subunit alpha 1) triggers DNM1L translocation from the cytosol to the OMM for assembly into ring structure, thereby facilitating mitochondrial fission [27]. We found that MEN1 knockdown significantly enhanced PRKAA1 phosphorylation at the Thr172 (p-Thr172) under galactose treatment (Figure 3(J)). Consistent with this, MEN1 deficiency augmented the mitochondrial localization and phosphorylation levels of DNM1L and MFF in isolated mitochondrial fractions (Figure 3(L,M), Figure S3G,H). IF costaining with anti-DNM1L antibody and COX4I1 showed that MEN1 deletion significantly increased DNM1L fluorescence intensity and the percentage of DNM1L colocalized with mitochondria when Hep3B cells were treated with galactose (Figure 3(N,O), Figure S3I). Co-IP analysis of mitochondrial fractions further revealed that MEN1 knockdown enhanced the interaction between DNM1L and MFF in Hep3B cells (Figure 3(P)).

To confirm the specificity of these findings, primary hepatocytes (Men1^f/f and men1^−/−) derived from WT and HMKO mice were isolated (Figure S3J,K) and replicated these experiments under glucose starvation conditions. Consistent with observations in MEN1-deficient HCC cells, men1^−/− hepatocytes showed profound alterations in mitochondrial morphology (Figure S3L,M); genetic ablation of Men1 triggered extensive mitochondrial fragmentation, manifesting as increased mitochondrial number and reduced mean branch length (Figure S3N,O). Accordingly, men1 knockout facilitated the mitochondrial localization and phosphorylation of DNM1L and MFF, accompanied by enhanced DNM1L fluorescence intensity and mitochondrial colocalization (Figure 3(Q), Figure S3P-S). Collectively, these findings demonstrate that MEN1 deficiency accelerates energy stress-induced mitochondrial fission via activating the PRKAA1-MFF-DNM1L pathway. This fission dysregulation exacerbates mitochondrial dysfunction and contributes to the accumulation of damaged mitochondria, key events that suppress HCC progression.

MEN1 deficiency suppresses autophagy and disrupts lysosomal degradation in liver cancer cells

Autophagy is crucial for delivering misfolded proteins and damaged organelles to lysosomes for degradation [28], and its dysregulation is closely associated with liver cancer progression and therapy resistance [29–31]. To explore the potential role of MEN1 in autophagy regulation, we ranked 60,660 genes from HCC samples in the The Cancer Genome Atlas (TCGA) dataset by their relative MEN1 expression levels in the MEN1-high group (186 cases) compared with MEN1-low group (185 cases) for Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. Results showed that MEN1-high samples were significantly enriched within the gene sets associated with lysosomal signatures and autophagy-related signaling pathways (Figure S4A). Gene Set Enrichment Analysis (GSEA) affirmed that the autophagy signaling was stronger in MEN1-high expression samples (Figure 4(A)), suggesting that MEN1 might regulate autophagy in HCC.

Figure 4.

MEN1 deficiency suppresses autophagy and disrupts lysosomal degradation in liver cancer cells. (A) GSEA of autophagy signaling gene sets in MEN1-high versus MEN1-low patients from the TCGA HCC cohort. (B and C) Immunoblotting of indicated proteins in vector and oeMEN1 Hep3B cells with or without galactose (4.5 g/L) treatment for 36 h (B). Quantitative analysis of LC3-II:I expression (C) from (B) (n = 3 biological replicates). (D and E) Immunoblotting of indicated proteins in shNC and shMEN1 Hep3B cells with or without CCCP (10 μM) treatment for 24 h (E). Quantitative analysis of LC3-II:I expression (D) from (E) (n = 3 biological replicates). (F and H) representative TEM images of shNC and shMEN1 Hep3B cells with or without galactose (4.5 g/L) treatment for 36 h (F). Scale bar: 500 nm; red arrows indicate autophagosomes. Quantification of autophagosome number (H) from (F) (n = 4 fields). (G and I) Representative immunofluorescent staining images of LC3 in shNC and shMEN1 Hep3B or Huh7 cells with or without galactose (4.5 g/L) treatment for 36 h (G). Scale bar: 20 μm. Quantification of LC3 puncta number (I) from (G) (n = 6 fields per condition). (J) Representative immunofluorescent staining images of mRFP-LC3B (red) and GFP-LC3B (green) in shNC and shMEN1 Hep3B cells transfected with mRFP-GFP-tagged LC3B, followed by treatment with 10 μM chloroquine (CQ) for an additional 24 h. Scale bar: 10 μm. (K) quantification of the ratio of autolysosomes (red puncta) to autophagosomes (yellow puncta) from (J) (n = 15 fields per condition). (L) Representative fluorescent staining images of DQ-BSA (red) and DAPI (blue) in shNC and shMEN1 Hep3B cells with glucose-free medium (starvation), starvation with 10 μM CQ, 10% FBS glucose medium and galactose medium. Scale bar: 40 μm. (M) quantitative analysis of mean fluorescence intensity of DQ-BSA from (L) (n = 6 fields per condition). Statistical analyses were performed using two-tailed unpaired Student’s t test (C, D, H, I, and M) and one-way ANOVA (K); *p < 0.05, **p < 0.01, and ***p < 0.001.

Hierarchical clustering and qPCR analysis revealed that autophagy-related genes were markedly downregulated in the liver tumors of HMKO mice compared to those in WT mice (Figure S4B,C). Immunoblotting revealed that the autophagy marker cytosolic MAP1LC3B/LC3B (microtubule associated protein 1 light chain 3 beta) underwent conversion to LC3B-II in Hep3B and Huh7 cells overexpressing MEN1 under galactose treatment (Figure 4(B,C), Figure S4D,E). Conversely, MEN1 knockdown attenuated CCCP-induced transformation of LC3-I to lipidated LC3-II along with a reduction in endogenous LC3 puncta (Figure 4(D,E), Figure S4D,F). Moreover, the expression of autophagy-related proteins (ATG5, SQSTM1/p62, BECN1 [beclin 1]) was decreased in HCC cells with MEN1 knockdown but increased in MEN1-overexpressing cells (Figure 4(B,E), Figure S4D). Consistently, significant reduction of autophagic vesicles and LC3-II puncta was observed in galactose-treated Hep3B cells with MEN1 knockdown by TEM imaging (Figure 4(F,H)), corroborating the reduced endogenous LC3 puncta levels. IF staining further indicated that MEN1 knockdown dramatically suppressed the generation of endogenous LC3 puncta in HCC cells regardless of galactose treatment (Figure 4(G,I)). These results confirm that MEN1 positively regulates autophagy in liver cancer cells.

MEN1 deficiency appears to impact autophagy initiation, as evidenced by decreased expression of BECN1, a key regulator of autophagy initiation (Figure 4(E), S4D). To determine whether disrupted autophagic flux results from defective autophagosome-lysosome fusion or reduced autolysosomal degradation capacity, a tandem mRFP-GFP-LC3B vector was used to evaluate autophagic turnover. As shown in Figure 4(J,K), Figure S4G,H, the number of LC3B puncta was significantly decreased in MEN1 knockdown Hep3B and Huh7 cells, with a large proportion of these puncta exhibiting GFP⁻ RFP⁺ signaling (autolysosomes) rather than GFP⁺ RFP⁺ signaling (autophagosomes). Conversely, lysosomal degradation was blocked by chloroquine (CQ) or bafilomycin A₁ (Baf A1), MEN1-knockdown cells showed a substantial proportion of LC3 puncta exhibiting GFP⁺ RFP⁺ signaling instead of GFP⁻ RFP⁺ signaling (Figure 4(J,K), Figure S4G,H). Further results revealed that MEN1 overexpression in Hep3B cells significantly augmented the number of both autolysosomes and autophagosomes, and the change patterns in GFP⁺ RFP⁺ and GFP⁻ RFP⁺ signaling caused by CQ treatment resembled those observed in MEN1-deficient cells (Figure S4I,J). These data indicate that MEN1 inactivation impairs autophagosome formation in HCC cells.

We speculated that accumulating fragmented mitochondria in MEN1-deficient HCC cells may result from an impaired autophagic process characterized by defective degradation capacity. To verify this, we employed a self-quenched fluorescent probe DQ-BSA (BODIPY-conjugated bovine serum albumin) that emits red light when degraded in lysosomes to test lysosomal degradation capacity. Hep3B cells under serum starvation (positive control) showed strong DQ-BSA fluorescence, while CQ treatment inhibited starvation-induced DQ-BSA degradation (Figure 4(L,M)). Notably, we observed reduced DQ-BSA fluorescence intensity due to impaired capacity for lysosomal degradation in MEN1 knockdown cells (Figure 4(L,M)). In line with these observations, treatment with galactose or CCCP effectively triggered autophagy of HCC cells as confirmed by immunoblotting (Figure 4(B,E), Figure S4D). These treatments induced a pronounced increase in DQ-BSA fluorescence intensity in Hep3B and Huh7 cells, and that effect was obviously attenuated by MEN1 knockdown (Figure 4(L,M), Figure S4K,L). Altogether, these findings demonstrate that MEN1 deficiency suppresses autophagy initiation and impairs autolysosomal degradation capacity in HCC cells.

MEN1 deficiency inhibits mitophagy through the regulation of PINK1-PRKN pathway in liver cancer cells

Mitophagy is a selective form of autophagy that eliminates dysfunctional mitochondria to maintain mitochondrial homeostasis [25]. Given the established mitochondrial dysfunction and autophagy impairment in the context of MEN1 deficiency, we investigated whether MEN1 deletion impacts mitophagy in HCC cells. To assess mitochondrial engulfment by autophagosomes, we examined the colocalization of autophagosomes and mitochondria via IF costaining assays. Results showed that MEN1 knockdown significantly decreased GFP-LC3 fluorescence intensity and reduced its colocalization with MitoTracker Red in galactose-treated Hep3B cells (Figure 5(A), S5A), suggesting mitophagy inhibition. To verify the positive regulation of MEN1 on mitophagy, we transfected Huh7 cells with Mito-Keima, a mitochondrially targeted pH sensitive fluorescent reporter of autophagy that exhibits green fluorescence at neutral pH and red fluorescence in the acidic environment of autophagosomes. After treatment with CCCP (a classic mitophagy inducer), MEN1-deficient cells showed significantly fewer acidic Mito-Keima puncta representing mitophagy than shNC Huh7 cells (Figure 5(B,C)). Quantification of Mito-Keima red:green ratio via confocal microscopy further confirmed that CCCP-induced increases in this ratio were suppressed by MEN1 knockdown (Figure 5(D)). Similarly, MEN1-deficient cells displayed significantly lower mitophagy levels regardless of CCCP treatment, as shown by flow cytometer analysis (Figure 5(E-G)). To determine whether accumulation of fragmented mitochondria occurred in MEN1-deleted HCC cells, Huh7 cells were costained with Lyso fluorescence dye and MitoTracker Red probe to concurrently label lysosomes and mitochondria, respectively. Notably, large fragmented mitochondria were more prominently observed in MEN1 knockdown cells, whereas a tubular mitochondria network was clearly distinct in shNC control cells (Figure S5B). After treatment with CCCP, the proportion of fragmented mitochondria sequestered within autolysosome was significantly lower in MEN1-knockdown cells compared to shNC cells (Figure S5C,D). These results confirm that MEN1 deficiency leads to the accumulation of damaged mitochondria in HCC cells, predominantly attributed to mitophagy suppression.

Figure 5.

MEN1 deficiency inhibits mitophagy through the regulation of PINK1-PRKN pathway in liver cancer cells. (A) Representative immunofluorescent staining images of GFP-LC3 (green) and MitoTracker red (red) in shNC and shMEN1 Hep3B cells with or without galactose (4.5 g/L) treatment for 36 h. Scale bar: 5 μm. (B) Representative immunofluorescent staining images of Mito-Keima in shNC and shMEN1 Huh7 cells with or without CCCP (10 μM) treatment for 24 h. Scale bar: 20 μm. (C) Quantification of acidic Mito-Keima puncta number from (B) (n = 20 cells). (D) Bar chart depicts the Mito-Keima red:green ratio representing mitophagy from (B) (n = 7 fields). (E) measuring mitophagy with mt-Keima by flow cytometry in shNC and shMEN1 Huh7 cells with or without CCCP (10 μM) treatment for 24 h (n = 3 biological replicates). (F and G) quantification of the mitophagy⁺ gate (F) and the ratio of fluorescence emitting at 611 nm following excitation at 561 nm vs 488 nm (G) from (E). (H) Immunoblotting of indicated proteins in whole-cell lysates (wcl) from shNC and shMEN1 Hep3B or Huh7 cells with or without galactose (4.5 g/L) treatment for 36 h. (I and J) immunoblotting of indicated proteins in cytosolic (cyto) and mitochondrial (Mito) fractions from shNC and shMEN1 Hep3B cells (I) and shNC and shMEN1 Huh7 cells (J) treated with CCCP (10 μM) for 24 h. (K and L) Apoptosis detection by flow cytometry using ANXA5-FITC staining (K) and quantitative analysis (L) of shNC and shMEN1 Huh7 cells with or without CQ (10 μM) treatment for 24 h (n = 3 biological replicates). (M and N) Representative EdU staining images (N) and quantitative analysis (M) for EdU-positive rate in shNC and shMEN1 Huh7 cells with or without 10 μM CQ treatment for 24 h (n = 3 biological replicates). Nuclei were stained with Hoechst (blue); scale bar: 50 µm. Statistical analyses were performed using two-tailed unpaired Student’s t test; *p < 0.05, **p < 0.01 and ***p < 0.001.

Recent studies have identified the PINK1-PRKN pathway as a central signaling cascade mediating mitophagy in mammalian cells [32]. Our results showed that MEN1 knockdown significantly attenuated the expression of COX4I1, PINK1, and PRKN, as well as PINK1 and PRKN phosphorylation in liver cancer cells following galactose treatment, but did not significantly affect the expression of MFN1 or TOMM20 (Figure 5(H)). Since PINK1-PRKN translocate to mitochondria to regulate mitophagy [33], their mitochondrial translocation in MEN1-deficient HCC cells was examined via subcellular fractionation. Immunoblotting revealed that MEN1 knockdown reduced the enrichment and phosphorylation of PINK1 and PRKN in the mitochondrial fraction of CCCP-treated HCC cells, while the enrichment levels of MFN1 and TOMM20 remained unchanged (Figure 5(I,J), Figure S5E,F). Consistently, MEN1 overexpression enhanced galactose-induced mitochondrial accumulation of COX4I1, as well as the enrichment and phosphorylation levels of PINK1 and PRKN (Figure S5G-I). Furthermore, MEN1 knockdown considerably inhibited cytosolic LC3-1 to lipidated LC3-II conversion in the mitochondrial fraction of CCCP-treated HCC cells (Figure 5(I,J), Figure S5E,F), whereas oeMEN1 promoted robust LC3-II lipidation (Figure S5G-I), confirming that MEN1 facilitates mitophagosome formation. These findings were corroborated by decreased mitochondrial colocalization of PINK1 with COX4I1 and PRKN with MitoTracker Red in CCCP-treated Huh7 cells with MEN1 knockdown (Figure S5J-L). Additionally, emerging evidence indicates that FUNDC1 (regulated by phosphorylation by ULK1 on Ser17) and BNIP3L/NIX function as mitophagy receptors to recruit LC3/GABARAP for mitochondrial clearance [34]. Our results showed that MEN1 deletion did not affect BNIP3L and FUNDC1 expression (Figure 5(H)), excluding their contribution to the impaired mitophagy phenotype.

Damaged mitochondria generate excessive ROS that trigger mitochondrial-mediated apoptosis; cancer cells clear dysfunctional mitochondria by enhancing mitophagy to promote survival [35]. GSEA analysis showed that the mitophagy signature was significantly enriched in HCC patients with high MEN1 expression (Figure S5M). Importantly, liver tumors in HMKO mice exhibited a remarkable enrichment in apoptosis and oxidative stress signals compared with WT mice (Figure S5N), suggesting that MEN1 deficiency impairs mitophagy-mediated clearance of damaged mitochondria. Consistent with these findings, when autophagy or mitophagy was inhibited with CQ treatment, MEN1 knockdown in Huh7 cells potently induced apoptosis and suppressed proliferation (Figure 5(K-N)), as assessed by flow cytometry and EdU staining assays. Notably, the apoptotic inhibitor Z-VAD-FMK (ZVF) exerted a more potent inhibitory effect on apoptosis in MEN1-knockdown cells than in shNC Huh7 cells, while the ferroptosis inhibitor Ferrostatin-1 had no such effect (Figure S5O,P), confirming apoptosis dependence. Collectively, these findings demonstrate that MEN1 deficiency suppresses mitophagy by downregulating the PINK1-PRKN pathway, leading to the damaged mitochondrial accumulation, ROS overproduction, and apoptosis induction-thereby inhibiting HCC progression.

MEN1-autophagy signaling is upregulated in HCC tissues and associated with poor prognosis

Based on the above findings, we analyzed MEN1 expression profiles in HCC and normal liver tissues using The International Cancer Genome Consortium (ICGC) and TCGA databases. Comparative analysis showed that MEN1 mRNA levels were significantly upregulated in HCC samples (Figure 6(A)) and positively correlated with pathological grades (Figure 6(B)). Similar results were obtained in three independent HCC datasets from the Gene Expression Omnibus (GEO) (Figure S6A), another cancer-related database. Compared with normal liver, adjacent nontumor, or cirrhosis tissues, HCC tumors showed an obvious elevation in MEN1 mRNA levels (Figure 6(A), S6A). Kaplan-Meier analysis of overall survival (OS), disease-specific survival (DSS), and progression-free survival (PFS) in the TCGA cohort exhibited an inverse correlation between MEN1 expression and survival probability (Figure 6(C), S6B). We then examined protein expression of MEN1 in a published dataset containing protein quantification of 8,943 genes between 165 pairs of normal liver and HCC samples, and demonstrated that MEN1 protein levels were significantly elevated in tumor samples relative to their paracancerous counterparts (Figure 6(D)).

Figure 6.

MEN1-autophagy signaling is upregulated in HCC tissues and associated with poor prognosis. (A) Expression level of the MEN1 gene in normal liver tissues and HCC samples from the ICGC database. (B) Expression level of the MEN1 gene in normal liver tissues and HCC tissues of different pathological grades from the TCGA database. (C) Kaplan−Meier overall survival (OS) in HCC patients stratified by high versus low MEN1 expression. (D) Expression levels of MEN1 protein in normal liver tissues and HCC samples from a published dataset in the fanjia database. (E) qPCR analysis of MEN1 mRNA expression in nontumor liver tissues (N) and HCC tumor (T) tissues (n = 85 pairs). (F) representative IHC staining images of MEN1 in HCC tissues of different differentiation grades (G1-well, G2-moderate, and G3-poor). Scale bars: 2 mm or 50 μm. (G) IHC scores for MEN1 expression calculated by the intensity and percentage of stained cells (n = 8, 68, and 9 for G1, G2, and G3 cohorts). (H) representative immunofluorescent staining images and fluorescence colocalization analysis of MEN1 and TOMM20 in nontumor liver tissues and HCC samples. Scale bar: 20 μm. (I) quantitative analysis of merged fluorescence intensity of MEN1 and TOMM20 from (H) (n = 35 fields). (J and K) Immunoblotting (J) and quantitative analysis (K) of MEN1 protein in whole-cell lysate (wcl), nuclei (Nuc), and mitochondria (Mito) fractions from nontumor liver tissues and HCC tumor tissues (n = 3 biological replicates). (L) Schematic diagram of the HCC model in C57BL/6 mice established by DEN and CCl₄ injection. (M) Representative liver images of normal and HCC model mice. Red box indicates tumorous regions and blue box indicates nontumor liver tissues in the same HCC mice. (N) immunoblotting of indicated proteins in normal liver tissues and liver tumors from control (vehicle) and HCC model mice. (O) Immunoblotting of MEN1 protein in whole-cell lysate and mitochondrial fractions from liver tumors (T) and adjacent nontumor tissues (A) of the same HCC mice. (P) Representative IHC staining images of PINK1 in nontumor liver (N) tissues and tumor (T) tissues from clinical HCC samples. Scale bar: 2 mm or 50 μm. Statistical analyses were performed using two-tailed unpaired Student’s t test (A, D, E, I, and K), one-way ANOVA (B and G) and log-rank (mantel-cox) test (C); *p < 0.05, **p < 0.01, and ***p < 0.001.

To validate these statistical analyses, we compared the mRNA and protein expression levels of MEN1 between 85 pairs of HCC tumors and adjacent nontumor tissues using qPCR and immunohistochemistry (IHC) staining. Results showed that MEN1 was overexpressed in 70.6% of tumors at the mRNA level and 58.8% at the protein level (Figure 6(E), Figure S6C,D), indicating that most HCC patients possess upregulated MEN1 expression. Correlation analysis between MEN1 protein expression and HCC malignancy (based on IHC staining data) showed positive correlations with differentiation grade, tumor size, tumor stage, and tumor number, but not with hepatitis B surface antigen, vascular invasion, or other clinicopathological characteristics (Figure 6(F,G), Table 1). Additionally, MEN1 levels were closely associated with serum levels of AFP (alpha fetoprotein), a well-established HCC biomarker (Table 1). We then investigated the intracellular distribution of MEN1 in clinical HCC tissues using immunoblotting and IF costaining, and observed that endogenous MEN1 was markedly enriched in mitochondria and nucleus fractions in HCC tissues relative to nontumor tissues (Figure 6(H-K), Figure S6E,F). To confirm these clinical findings, a DEN- and CCl₄-induced HCC model was established in C57BL/6 mice (Figure 6(L,M)). qPCR and immunoblotting showed significantly elevated MEN1 mRNA and MEN1 protein levels in liver tumors versus adjacent nontumor tissues (Figure 6(N), S6G). Subcellular fractionation analysis further revealed that MEN1 mitochondrial localization was notably greater in liver tumors than in adjacent noncancerous tissues from the same animals (Figure 6(O)).

Table 1.

Correlation between MEN1 and clinicopathologic characteristics in HCC patients.

Characteristics	MEN1 expression levels
	All cases	Low	High	p value
Total cases	85	35	50
Gender				0.797
Male	65	26	39
Female	20	9	11
Age (years)				0.267
≤50	37	18	19
>50	48	17	31
Hepatitis B surface antigen				0.169
Negative	28	15	13
Positive	57	20	37
Liver cirrhosis				0.497
No	32	15	17
Yes	53	20	33
Tumor encapsulation				0.802
No	22	10	12
Complete	63	25	38
Vascular invasion				0.643
No	58	25	33
Yes	27	10	17
Tumor number				0.065
Single	73	27	46
Multiple	12	8	4
Tumor size (cm)				0.046*
<5	37	20	17
≥5	48	15	33
Differentiation				vs II-III 0.052
Well	9	6	3
Moderate	68	26	42
Poor	8	3	5
Tumor stage				vs II-III 0.0015*
I	50	28	22
II	27	4	23
III	8	3	5
AFP (ng/ml)				0.027*
<20	36	20	16
≥20	49	15	34

AFP, alpha fetoprotein; *p < 0.05, p values from Fisher’s exact test.

The relationships between MEN1 and autophagy or mitophagy-related gene expression in liver cancer samples were further explored. Spearman’s rank correlation analysis of the TCGA cohort revealed a significant positive correlation between MEN1 expression and autophagy-related genes (e.g., ATG5, ATG3, ATG7, ULK1, BECN1, and SQSTM1) (Figure S6H). Consistent with pathway involvement, ATG5 and BECN1 transcript levels were significantly increased in HCC samples relative to normal liver tissues, and their high expression was associated with poor prognosis (Figure S6I,J). In the HCC mouse model, liver tumors displayed elevated autophagy levels (LC3-I to LC3-II conversion) and activated PINK1 signaling pathway, as indicated by increased PRKN expression (Figure 6(N)), with IHC showing higher PINK1 expression in HCC tumor areas than in adjacent nontumor tissues (Figure 6(P), S6K). These findings suggest that aberrant activation of mitochondrial MEN1-mediated mitophagy contributes to HCC malignancy and progression.

Men1 deficiency suppresses HCC tumorigenesis via disrupting the mitophagy homeostasis

Mitophagy regulates mitochondrial number and maintains quality control by selectively eliminating damaged mitochondria; aberrant mitophagy promotes cell death or survival by disturbing mitochondrial homeostasis [30,31,36]. We used T0467, a novel mitophagy inducer designed to alleviate PINK1-PRKN pathway suppression and enhance PINK1-mediated PRKN recruitment to mitochondria [37], to verify whether restoring mitophagy reverses the tumor-suppressive effects of MEN1 deficiency (Figure S7A,B). T0467 treatment reversed MEN1 knockdown-induced proliferation inhibition in HCC cells under galactose culture conditions (Figure 7(A), S7C) and alleviated MEN1 deficiency-mediated inhibition of clonogenic survival in Hep3B and Huh7 cells (Figure S7D,E). Furthermore, MEN1 deletion increased basal apoptosis levels of HCC cells (Figure 7(B,C)), while MEN1 overexpression mildly diminished apoptotic activity (Figure S7F,G). Importantly, T0467 effectively reversed Baf A1-induced mitochondrial apoptosis in MEN1-deficient Hep3B cells, restoring apoptosis levels to those observed in Baf A1-treated shNC cells (Figure 7(B,C)). Like MEN1 downregulation, siPINK1 (Figure S7H) dramatically elevated the apoptosis rate in Huh7 cells, and this effect was completely counteracted by MEN1 overexpression (Figure S7F,G). These in vitro results confirm that MEN1 downregulation induces cytotoxicity in HCC cells via mitophagy suppression.

Figure 7.

Men1 deficiency suppresses HCC tumorigenesis via disrupting the mitophagy homeostasis. (A) Cell viability of shNC and shMEN1 Hep3B cells with or without T0467 (10 μΜ) treatment for 48 h in galactose medium (4.5 g/L) (n = 6 assays). (B and C) apoptosis detection by flow cytometry using ANXA5-FITC staining (B) and quantitative analysis (C) in shNC and shMEN1 Hep3B cells treated with T0467 (10 μΜ), Baf A1 (10 μΜ) alone or their combination for 48 h (n = 3 biological replicates). (D) Schematic diagram illustrating the effect of T0467 on tumor growth in Men1-deficient HCC model with DEN and CCl₄. (E-J) Liver images (E) and quantitative analysis of liver weight (F), tumor foci number (G), tumor burden (H), maximum tumor diameter (I), and serum GPT and GOT1 levels (J) in WT and HMKO HCC model mice injected with T0467 (n = 8 mice per group) or vehicle (n = 7 mice per group). (K) Representative images of H&E and IHC staining for MKI67, PINK1, PRKN, and cleaved-CASP3 of liver tumor tissue sections from WT and HMKO HCC mice injected with T0467 or vehicle. Scale bars: 2 mm or 50 μm. (L) ATP levels in liver tumors from WT and HMKO HCC mice injected with T0467 or vehicle. (M) Representative immunofluorescent staining images of TOMM20 (red) and PRKN (green) in liver tumor tissue sections from WT and HMKO HCC mice treated with T0467 or vehicle. Nuclei were stained with DAPI (blue); scale bar: 20 μm. (N and O) Quantification of PRKN fluorescence intensity (N) and fluorescence colocalization analysis of PRKN and TOMM20 (O) from (M). (P) Representative TEM images of liver tumor tissues from WT and HMKO HCC mice treated with or T0467 or vehicle. Scale bar: 500 nm. (Q) Quantitative analysis of mitophagosome numbers per 25 μm² in liver tumor tissues from (P). Statistical analyses were performed using one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, and ns, not significant.

To further evaluate the role of MEN1-regulated mitophagy in HCC tumorigenesis, WT and HMKO mice at 5 months (early HCC initiation stage) after DEN and CCl₄ induction were treated with T0467 (Figure 7(D)). In both vehicle and T0467-treated groups, tumorigenesis in HMKO mice was markedly alleviated relative to WT mice (Figure 7(E)). Liver weight (but not body weight) and tumor foci number were obviously decreased in HMKO mice compared with WT mice (Figure 7(F,G), Figure S7I). H&E staining also displayed a significant reduction in tumor burden and maximum tumor diameter in HMKO mice (Figure 7(H,I,K)). Consistently, serum GPT/ALT (glutamic – pyruvic transaminase) and GOT1/AST (glutamic-oxaloacetic transaminase 1) levels were significantly lower in HMKO mice (Figure 7(J)), and IHC results showed that men1 knockout substantially reduced MKI67/Ki67-positive proliferative areas (Figure 7(K), S7J). These data further demonstrate that MEN1 acts as an oncogenic driver in HCC and that hepatocyte-derived MEN1 promotes tumor development and progression.

Notably, T0467 treatment accelerated tumor progression in both genotypes compared with vehicle controls and its protumorigenic effect was more pronounced in WT mice than in HMKO mice (Figure 7(E), Table 2); T0467-treated WT mice showed a 2.45-fold increase in tumor foci number, whereas T0467-treated HMKO mice exhibited a 2.24-fold increase in the same parameter (Figure 7(G)). Similar trends were observed in tumor burden (1.90-fold in WT mice vs. 1.63-fold in HMKO mice), liver weight (1.74-fold in WT mice vs. 1.55-fold in HMKO mice), and maximum tumor diameter (2.81-fold in WT mice vs. 2.07-fold in HMKO mice) (Figure 7(F,I)). Importantly, T0467 had no significant impacts on body weight or serum GPT and GOT1 levels (Figure 7(J), S7I), suggesting minimal systemic toxicity. This was further validated by H&E staining of lung, spleen, heart, and kidney tissues from T0467-treated (3 months) WT and HMKO mice without DEN or CCl₄ induction, which showed no significant histological damage compared to vehicle controls (Figure S7K).

Table 2.

The efficacy of T0467 in primary liver cancers.

Primary liver tumors		Treatment condition		Relative to Vehicle treatment (set to 1)
		Vehicle	T0467	Ratio of T0467 toVehicle
Tumor foci number	WT	15.57 ± 5.42	22.50 ± 4.56	2.45
	HMKO	4.57 ± 3.99	10.25 ± 5.56	2.24
Tumor burden (%)	WT	21.35 ± 4.13	40.52 ± 8.06	1.9
	HMKO	10.02 ± 3.50	16.35 ± 3.93	1.63
Liver weight (g)	WT	3.31 ± 0.93	5.77 ± 1.92	1.74
	HMKO	1.88 ± 0.21	2.92 ± 1.05	1.55
Maximum tumor diameter (mm)	WT	8.14 ± 2.09	22.88 ± 3.10	2.81
	HMKO	3.14 ± 2.17	6.50 ± 3.87	2.07
Mitophagosomes number/25 μm2	WT	1.83 ± 1.16	6.83 ± 1.17	3.73
	HMKO	0.83 ± 0.75	1.50 ± 1.05	1.81

HMKO tumors displayed a marked alteration in mitophagy and apoptosis levels, as evidenced by the diminished PINK1 and PRKN, and the increased cleaved-CASP3/caspase 3 positive areas relative to WT tumors (Figure 7(K), S7J). Men1 deficiency significantly impaired ATP production in the liver tumors of HMKO mice (Figure 7(L)). T0467 treatment reactivated the mitophagy and inhibited apoptosis signaling pathway in HCC tumors, as demonstrated by the restored expression of PRKN (but not PINK1) and cleaved-CASP3, alongside modestly recovered ATP levels (Figure 7(K,L), Figure S7J). IF staining further revealed that Men1 deficiency attenuated the fluorescence intensity of PRKN and PINK1, as well as their mitochondrial recruitment, in HMKO tumors versus WT tumors (Figure 7(M-O), S7L-N). T0467 significantly facilitated the PRKN expression and mitochondrial localization but had no effect on PINK1 in both genotypes (Figure 7(M-O), S7L-N). TEM observations showed that mitophagosome formation was blocked in HMKO tumors compared to WT tumors, and T0467 effectively rescued this defect: T0467 increased mitophagosomes by 3.73-fold in WT tumors and 1.81-fold in HMKO tumors (Figure 7(P,Q)). Taken together, these findings confirm that MEN1 deficiency suppresses HCC tumorigenesis at least in part by disrupting the PINK1-PRKN-mediated mitophagy pathway.

Discussion

Although MEN1 is primarily a nuclear protein that interacts with multiple nuclear factors [9], it is also detectable in cell membrane, cytoplasm, and other subcellular fractions [15,16,38,39]. However, little is known about how MEN1’s subcellular localization relates to its biological functions. Several studies have revealed that non-canonical functions of MEN1 depend on its targeting to subcellular compartments, with localization in membrane vesicles and cytoplasmic compartment observed under different contexts. In pancreatic β cells, MEN1 localized in the cytoplasm in response to glucose or insulin to promote β-cell proliferation [40]. Consistently, MEN1 was found to reside within insulin containing secretory vesicles, where it presumably regulates secretory granule formation and cellular trafficking [39]. At the plasma membrane, MEN1 modulates membrane-associated signaling cascades, as exemplified by its role in facilitating the translocation of AKT1 from the cytoplasm to the plasma membrane [41]. In hepatocytes, MEN1 has been shown to interact with cellular energy sensors (SIRT1, PPARG/PPARɣ, and FOXO1) in the cytoplasm to maintain hepatic metabolic homeostasis by facilitating glucose uptake and restraining fatty acid uptake [20]. In the present study, we detected a substantial pool of MEN1 in the mitochondria of HCC cells and tissues. MEN1 deletion leads to mitochondrial dysfunction, accompanied by impaired lysosomal degradation capacity and reduced autophagic activity. Under energy stress conditions, knockdown of MEN1 disrupts the PINK1-PRKN dependent mitophagy pathway, thereby promoting the accumulation of dysfunctional mitochondria, ROS generation, and cell death in HCC cells. Hepatocyte-specific Men1 deficiency suppresses liver tumorigenesis and progression in DEN- and CCl₄-induced HCC mouse models, and this inhibitory effect is attenuated-with accelerated tumor growth-upon blockage of the PINK1-PRKN mediated mitophagy pathway. These diverse studies show that the subcellular localization of MEN1 dictates its biological functions and thus control cell fate in HCC.

Mitochondrial quality and quantity are governed by the dynamic balance between fusion and fission, orchestrated by dynamin-related GTPases (MFN1, MFN2, OPA1) and cytosolic DNM1L with its OMM receptors, respectively [42]. Dysregulation of this equilibrium impairs mitochondrial functions including respiration, perturbing cellular fate decisions [43-45]. Experimental studies in both human and murine models have demonstrated that disrupted mitochondrial homeostasis is a hallmark of cancer, and accumulating evidence links HCC progression to mitochondrial dysfunction, highlighting mitochondrial function as a potential anticancer target via pharmacological or genetic interventions [29–31]. Here, we found that MEN1 localizes to the OMM via a TOMM20-dependent translocation manner. Loss of MEN1 causes mitochondrial dysfunction characterized by reduced mitochondrial membrane potential, ATP synthesis, and oxygen consumption rate, which strongly correlates with suppressed HCC cell proliferation and tumor growth. Naturally, this prompts inquiry into the functional implications of MEN1’s mitochondrial localization. Mitochondria, often termed cellular powerhouses for their ability to efficiently produce ATP to sustain normal cellular physiology, are central to such roles [46]. Our data demonstrate that mitochondrial MEN1 is critical for maintaining mitochondrial homeostasis. However, no significant alternations in mitochondrial fusion proteins were observed, suggesting MEN1 specifically regulates mitochondrial fission. Consistent with this, MEN1 downregulation induces extensive mitochondrial fragmentation and elevates energy stress-induced mitochondrial apoptosis in HCC cells. Under glucose deficiency, PRKAA1 activation signaling pathway facilitates the clearance of dysfunctional mitochondria by phosphorylating MFF, regulating DNM1L and activating ULK, an upstream kinase in autophagy that triggers mitophagy [22,47]. Concordantly, MEN1 deletion enhances PRKAA1 and MFF phosphorylation, promotes mitochondrial recruitment and translocation of DNM1L and MFF, which in turn accelerates mitochondrial fission.

Mitophagy preserves cellular homeostasis by precluding superfluous ROS production from dysfunctional mitochondria [48]. Mounting evidence indicates that defective or excessive mitophagy accelerates the pathogenesis of various disorders, including metabolic diseases, diabetes, Parkinson disease, and cancer [49,50]. Mitochondria-targeted PINK1 and PRKN are well recognized for their synergistic role in mediating mitophagy of defective mitochondria [48]. In this study, we show that MEN1 knockdown in HCC cells significantly reduces PINK1 and PRKN expression and blocks their mitochondrial translocation. Moreover, MEN1 deficiency impairs the autophagic process characterized by disrupted mitophagosome formation and reduced lysosomal degradation capacity; this impairment drives the accumulation of cytotoxic damaged mitochondria, excessive ROS generation, and mitochondrial apoptosis. These findings align with our previous observations in lung cancer, where MEN1 maintains lysosomal homeostasis and autophagic flux through epigenetic regulation of SQSTM1 and MAP1LC3B expression independent of the PRKAA1-MTORC1 pathway [51].

In this study, we identify MEN1 as a predictor of recurrence and overall survival in HCC patients, with elevated autophagy/mitophagy signaling positively associated with poor prognosis, this suggests aberrant activation of the MEN1-autophagy and mitophagy axis in high-energy-demand proliferating HCC cells. Mechanistically, we establish that MEN1 expression or mitochondrial localization links to PINK1-PRKN signaling to mediate mitophagy during HCC progression: MEN1 deletion downregulates PINK1-PRKN, impairs mitophagy, and triggers accumulation of damaged mitochondria-alongside reduced ATP production and elevated ROS-directly compromising HCC cell proliferation and tumor growth. This tumor-suppressive effect of mitophagy inhibition is mitigated by T0467 (a PINK1-PRKN activator), confirming the axis’s functional relevance. Our findings challenge the conventional view of autophagy as a context-dependent “double-edged sword” in tumorigenesis [50,52]. While autophagy can suppress initiation or fuel progression, MEN1-dependent mitophagy here acts as a critical adaptive process sustaining malignant HCC cells. This highlights that autophagy’s role is governed by upstream regulators like MEN1, advancing understanding of MEN1’s functional plasticity. Importantly, it provides a rationale for targeting the MEN1-mitophagy axis in HCC treatment, supported by precedents such as revumenib, an FDA-approved MEN1-MLL inhibitor that exhibits superior efficacy in relapsed/refractory acute leukemia harboring KMT2A translocations [53,54].

Nevertheless, our mechanistic validations were performed in vitro, and in vivo verification using pink1 knockout mice was not included in this study. Future investigations will employ liver-specific pink1 knockout mice to further confirm our findings in a more physiological environment.

Materials and methods

Cell culture

Hep3B (HB-8064) and HEK-293T (CRL-11268) were obtained from the American Type Culture Collection/ATCC. Huh7 (CRL-10741) was purchased from the Japanese Collection of Research Bioresources. Cells were routinely maintained in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific, C11995500BT) supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, A5256701), 1% penicillin (10,000 units/ml), and 1% streptomycin (100 μg/ml). All cell lines were cultured in a humidified cell incubator at 37°C with 5% CO₂. For experimental conditions requiring replacement of glucose with galactose (Macklin, C16853494) in the culture medium, MEFs, Huh7, or Hep3B cells were washed with phosphate-buffered saline (PBS; Beyotime, C0221A) to remove residual glucose and then cultured in glucose-free IMDM (Thermo Fisher Scientific, 11,966,025) containing 10% dialyzed FBS and 4.5 g/L galactose. To establish a low-glucose culture environment, these cells were rinsed three times with PBS and maintained in glucose-free DMEM supplemented with 10% dialyzed FBS and 1.5 g/L glucose. All cell lines were tested as free for mycoplasma contamination.

Human HCC specimens

HCC specimens and paired adjacent nontumor liver tissues were obtained from surgical resection at the Department of Hepatobiliary Surgery, the Affiliated Hospital of Guizhou Medical University. All studies in clinical specimens were performed in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Affiliated Hospital of Guizhou Medical University (Ethics license, 2,023,086). The collection and use of clinical samples were approved and conducted in compliance with data protection regulations to ensure patient confidentiality. The diagnosis of HCC was confirmed by pathologists and informed consents were obtained from all patients. Patients who had received chemotherapy or interventional treatments prior to surgical resection were excluded. Following surgical resection of primary tumors, a section of each specimen was immediately fixed with 4% (w:v) paraformaldehyde (Beyotime, P0099) solution for 48 h before paraffin embedding, and the remaining tissues were stored at −80°C. A total of 85 HCC samples (65 males and 20 females) aged 26–80 years, and other clinicopathological characteristics are summarized in.

Establishment of hepatocyte-specific men1 knockout mice and carcinogenesis induced by chemical carcinogens

Men1 floxed mice were generated by GemPharmatech Co., Ltd. Briefly, single-guide RNA (sgRNA), Cas9 mRNA and a donor vector targeting exon 2 and 5 were microinjected into mouse zygotes to create a Men1-flox allele with two LoxP sites flanking exons 2 and 5. The vector was transfected into mouse embryonic stem (ES) cells via electroporation. Following standard protocols, correctly targeted ES cell clones confirmed by Southern blotting were microinjected into blastocysts of albino C57BL/6 mice. F1 mice produced by crossing C57BL/6 mice with chimeric mice were subjected to Southern hybridization with 3′ and 5′ external probes to confirm germline transmission of the Men1 floxed allele. Genotyping primers used in this study are listed in Table S6.

To generate hepatocyte-specific men1 knockout mice, Men1^f/f mice were crossed with Alb-Cre transgenic mice, which express Cre recombinase under the control of the hepatocyte-specific albumin promoter to restrict its activity exclusively to hepatocytes. Offspring carrying both the Men1^f/f allele and the Alb-Cre transgene were identified as HMKO mice, where Cre-mediated recombination selectively deletes the floxed Men1 exons in hepatocytes. Littermates lacking the Alb-Cre transgene served as WT controls. All animals were housed under standard conditions at Guizhou Medical University under standard conditions with food and water ad libitum and maintained on a 12 h light – dark cycle.

The DEN- and CCl₄-induced HCC mouse model was established as previously described [55], mice aged 14 days received a single intraperitoneal (i.p.) injection of DEN (Macklin, C14391668) at 25 mg/kg, followed by 8 injections of CCl₄ (Macklin, C805325) every two weeks (1.2 ml/kg, diluted 1:4 in olive oil, i.p.), starting 4 weeks after DEN administration. For HCC induction in WT and HMKO mice, either T0467 (5 mg/kg, biweekly; MedChemExpress, 859,518-94–8) or vehicle control was administered via i.p. injection for 5–8 months after DEN initiation. Mice were sacrificed at indicated time points following DEN injection. All studies in mice subjects were approved by the Institutional Animal Care and Use Committee of Guizhou Medical University (Ethics license, 2,303,047). The animal experiments were in strict accordance with good animal practice as defined by the Laboratory Animal Center of Guizhou Medical University.

Mitochondrial respiration assay

Oxygen consumption rate was measured using a Seahorse XFp Analyzer (Agilent Technologies, 103,015-100) in liver cancer cells. Briefly, HCC cell lines were seeded in XF96 cell culture plates at a density of 20,000 cells per well and allowed to adhere overnight in complete medium at 37°C with 5% CO₂. Subsequently, cells were washed twice with PBS and maintained in either complete medium or galactose medium for 36 h. Prior to the assay, cells were washed and incubated in XF basal medium (Agilent, 103,575-100) supplemented with 2 mM glutamine (Solarbio, G0200)-with 1 mM sodium pyruvate (Roche, 105,477) and 4.5 g/L glucose added for the control group, while omitted for the galactose group. The prepared test solutions were preincubated in a CO₂-free incubator at 37°C for later use. Cells were then equilibrated in a CO₂-free 37°C incubator for 1 h to stabilize prior to OCR measurement. Mitochondrial OCR was determined using the Seahorse XF96 Analyzer. The OCR was monitored over time both before and after sequential injections of mitochondrial modulations at the following final concentrations: oligomycin (1 μM), FCCP (1 μM), and a combination of rotenone (0.5 μM) and antimycin A (0.5 μM).

Transmission electron microscope observations

For mitochondrial ultrastructural analysis, HCC cells or fresh liver tissues from WT and HMKO mice were fixed in 2.5% glutaraldehyde solution (Macklin, G849973) at 4°C for 12 h. After washing with PBS, the cells or tissues were refixed in 1% osmium tetroxide (EMCN, GP18456). They were then dehydrated in graded series of acetone (XiLong Scientific, 67–64–1), infiltrated with acetone-Epon-812 resin mixtures, and embedded in Epon-812 resin (Sigma-Aldrich, 25,038–04–4). Ultrathin sections were counterstained with uranyl acetate (EMCN, GS02624) and lead citrate (EMCN, GS10701-1). Images were captured using a JEM-1400FLASH transmission electron microscope (JEOL, Japan) and recorded on copper grids.

Detection of mitophagy in vitro

To visualize mitophagy using a mitophagy-specific probe, shNC and shMEN1 Huh7 cells were transfected with pMito-Keima (MBL, AM V0251), followed by treatment with 10 μM CCCP (MedChemExpress, 555–60–2) for 24 h. Keima fluorescence under neutral pH was observed at an excitation wavelength of 430 ± 20 nm and emission wavelength of 624 ± 20 nm, while fluorescence under acidic conditions was observed at 562 ± 20 nm (excitation) and 624 ± 20 nm (emission). Mito-Keima thus serves as an effective tool for evaluating mitophagy flux. To visualize and quantify mitophagy, we employed two complementary approaches: confocal microscopy (Olympus, Japan) and flow cytometry (ACEA NovoCyte). Acidic Mito-Keima puncta and fluorescence ratios were quantified using ImageJ and FlowJo software.

IHC staining

IHC staining procedure was performed following standard protocol. Paraffin-embedded samples were dewaxed and rehydrated sequentially. Antigen retrieval was performed by heating sections in 0.01 mM sodium citrate buffer ([MedChemExpress, HY-B1610N], pH 8.0) using a pressure cooker. To block endogenous peroxidase activity, sections were treated with 1% hydrogen peroxide solution (Macklin, H650970) for 20 min. After blocking nonspecific binding with 5% bovine serum albumin (BSA; Solarbio, A8010) for 1 h at room temperature, sections were incubated overnight at 4°C with primary antibodies diluted in PBS containing 5% BSA. Anti-MEN1 (Bethyl Laboratories, A300-105A; 1:1000), MKI67 (Cell Signaling Technology, 9129; 1:500), PINK1 (Proteintech, 23,274–1-AP; 1:200), PRKN (Cell Signaling Technology, 4211s; 1:200), and cleaved-CASP3 (Proteintech, 25,128–1-AP; 1:100). The following day, sections were treated with biotinylated secondary antibody (ZSGB-BIO, PV-9000) at 37°C for 30 min, developed using DAB chromogenic solution (ZSGB-BIO, ZLI-9018), and counterstained with hematoxylin to visualize. Quantification of IHC staining was carried out by calculating the ratio of the positive staining area to the total area using Image-Pro Plus software. Staining intensity was scored as 0 (negative), 1 (weak), 2 (moderate) and 3 (strong). The positive area score was defined based on the stained area as 1 (≤25%), 2 (26%–50%), 3 (51%–75%) and 4 (≥76%). IHC score was calculated by multiplying the intensity score by the positive area score. The results for MKI67, PRKN, cleaved-CASP3 were assessed based on the percentage of positive-staining cells.

Proteinase K protection assay

Freshly purified mitochondria were divided into two equal aliquots. One aliquot was resuspended in either osmotic buffer (25 mM sucrose [Solarbio, S8271], 10 mM MOPS-KOH [Solarbio, M8150], pH 7.2) or 1% Triton X-100 (Sigma-Alrich, T8787) to disrupt the outer mitochondrial membrane, while another aliquot was resuspended in regular mitochondrial buffer and treated with or without proteinase K (Beyotime, ST533) for 30 min. All samples were treated with proteinase K (at concentrations ranging from 0.1 to 10 mg/ml) for 0.5 h, 1 h, or 2 h on ice to digest surface-exposed proteins. The reaction was terminated by adding 2 mM PMSF (Solarbio, P0100) and samples were subsequently analyzed by immunoblotting.

GPT and GOT1 analyses

For serum biochemical analysis, blood samples were collected from anesthetized WT and HMKO mice and centrifuged at 5,000 × g for 5 min to isolate serum. The activities of GOT1 and GPT in the serum were measured using commercial Kit (Solarbio, BC1565 or BC1555) according to the manufacturer’s protocol. Briefly, a standard curve was generated by diluting the 20 mol/L sodium pyruvate solution with distilled water. After mixing the serum samples with the respective reagents, the absorbance was measured at 505 nm with an Epoch Ultra-Micro Microplate Spectrophotometer (BioTek, Epoch). GPT or GOT1 activity (U/L) =12 × X × F, (X represents the concentration and F denotes the dilution factor).

qPCR assay

RNA extraction, cDNA synthesis, and qPCR were performed as described previously. Total RNA was isolated from cultured cells and liver tissues from mice or human using the RNAiso Plus reagent (Takara, 9109). For qPCR analysis, 1 μg of total RNA was reverse-transcribed into cDNA using the HiScript® II 1st Strand cDNA Synthesis Kit (Yeasen, 11141ES60). Subsequent PCR amplification was carried out with TB Green® Premix Ex Taq^TM II (Takara, RR820A) on a CFX Connect Real-Time System (Bio-Rad). Relative mRNA expression levels were calculated by 2^−ΔΔCt method, with GAPDH or ACTB/β-actin serving as the internal reference control. Reverse transcription (RT-PCR) was performed with the following program: 95°C for 2 min; 95°C for 30 s, 65°C for 30 s, and 72°C for 1 min, 30 cycles; 72°C for 5 min. PCR products were run in 2% agarose gels stained with 0.1‰ ethidium bromide dye (Macklin, C14141868). Bands were visualized with a GleUV system (Baygene Biotech). Primer sequences used in this study are shown in Table S6.

Immunoblotting

For SDS-PAGE analysis, both whole-cell lysates and mitochondrial fractions were collected and lysed on ice in RIPA buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl [Solarbio, T8230], 1% Triton X-100, 0.1% SDS [Solarbio, S8010], 1% NP-40 [Beyotime, ST2045], and 1% sodium deoxycholate [Beyotime, ST2048]) supplemented with protease inhibitor (1 mM PMSF). After centrifugation (22,000 × g, 10 min, 4°C), supernatants were retained, and protein concentrations were quantified using a BCA assay (Solarbio, PC0020) before SDS-PAGE analysis. Separated proteins were transferred to the polyvinylidene difluoride membranes (Thermo Fisher Scientific, 88518). After blocking with 5% nonfat milk TBST (Servicebio, G0004) for 1 h at room temperature, membranes were washed with TBST and incubated with primary antibodies against target proteins at 4°C overnight. Antibodies against TOMM20 (66777–1-Ig), DNM1L (12957–1-AP), PRKN (14060–1-AP), COX4I1 (66110–1-Ig), PINK1 (23274–1-AP), FIS1 (10956–1-AP), TOMM70 (14528–1-AP), MFF (17090–1-AP), ATG5 (10181–2-AP), BECN1 (11306–1-AP), MFN1 (13798–1-AP), VDAC1 (10866–1-AP), MIEF1 (20164–1-AP), FUNDC1 (28519–1-AP), cleaved-CASP3 (25128–1-AP), GAPDH (10,494–1-AP), IgG (B900620), ACTB (66009–1-Ig), TUBA (11224–1-AP) were purchased from Proteintech. Antibodies against p-DNM1L (4494), PRKN (4211), LC3A/B (12741), LMNB (17416), MKI67 (9129), SQSTM1 (23214), p-PRKAA1 (2535), His (12698), BNIP3 (44060) were purchased from Cell Signaling Technology. Antibodies against MEN1 was purchased from Bethyl Laboratories (A300-105A). Membranes were then washed with TBST buffer and incubated with horseradish peroxidase-conjugated secondary antibodies (Proteintech, 10494–1-AP; B900620) for 1 h at room temperature. Following additional washes, immunoreactive signals were visualized using an enhanced chemiluminescence detection system (Tanon, 5200).

EdU incorporation and detection

A Click™ EdU-488 Cell Proliferation Detection Kit for Imaging (Beyotime, C0071S) was used to assay cell proliferation through 5-ethynyl-2^’-deoxyuridine (EdU; Beyotime, C0071S-1) incorporation. Briefly, cells were seeded into 96-well plates (1.0 × 10³ cells per well) and treated with DMSO (Solarbio, ID9011) or CQ (MedChemExpress, 1954–5–7) for 24 h. Cells were labeled with 10 μM EdU at 37°C for another 24 h. Cells were then fixed with 4% paraformaldehyde in PBS and the Click™ reaction was performed according to the manufacturer’s instructions. Nuclei were stained with Hoechst 33,342 (Beyotime, C1027) for 15 min at room temperature in the dark. Plates were imaged with ImageXpress Micro-4 (Molecular Devices) or Nikon Ar1 (Nikon) microscope. Image analysis was performed MetaMorph software. Data from nine different fields in each condition were quantified per experiment. The EdU entire population nuclear intensity and the percentage of cells incorporating EdU were measured. EdU intensity was also detected only for those cells that incorporated EdU.

Immunofluorescence and other cell staining

For cellular IF staining, cells were first seeded on coverslips in 6-well plates. After attachment, cells were washed three times with PBS and fixed with 4% paraformaldehyde solution for 15 min at room temperature. Fixed cells were then permeabilized with 0.1% Triton X-100 in PBS for 10 min and blocked with blocking buffer (5% BSA, and 0.1% Tween-20 [Beyotime, ST825]) for 1 h at room temperature. Cells were incubated with primary antibody overnight at 4°C, followed by incubation with fluorescently conjugated secondary antibodies (Thermo Fisher Scientific, A11005; A11008) for 60 min at 37°C. Nuclei were stained with 4,’6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich, D9542]) for 30 min. For DQ-BSA lysozyme activity analysis, the indicated cells were preincubated with DQ-BSA Red (Thermo Fisher Scientific, D-12051) at final concentration of 10 μg/ml for 1 h. Cells were then treated galactose, CCCP, or vehicle for an additional 24 h. After fixation, cells were counterstained with DAPI as described above. The fluorescence intensity of DQ-BSA was quantified using ImageJ software from images acquired with confocal microscopy. For mitochondrial morphology analysis, cells (1 × 10⁵) were seeded into confocal dishes (NEST, 801,001) and cultured for 24 h. Cells were then incubated with MitoTracker Red CMXRos (Cell Signaling Technology, 9082) for 45 min at 37°C followed by three washes with PBS. Fluorescent images were captured randomly from each group using confocal microscopy and mitochondrial footprint was analyzed using the Mitochondrial Network Analysis (MiNA) toolset of Fiji software [56]. The colocalization of flag-MEN1 and TOMM20 was assessed using the colocalization tool in ImageJ software. Mitochondrial membrane potential was measured with JC-1 (Beyotime, C2006). Live-cell images were captured using a confocal microscopy and quantification for ∆Ψm assay was performed using Fiji software by calculating the fluorescence intensity ratio of JC-1 aggregates to JC-1 monomers.

For IF staining of liver tissues, paraffin-embedded sections were deparaffinized in xylene, followed by antigen retrieval via boiling in 0.1 M citrate buffer (pH 6.0) in a water bath for 15 min. After blocking with 5% BSA for 1 h at room temperature, sections were incubated overnight at 4°C with primary antibodies against PINK1 (Proteintech, 23,274–1-AP; 1:200), PRKN (Cell Signaling Technology, 4211; 1:200). Sections were then washed three times in PBS (5 min per wash) and incubated with Alexa Fluor 594 (Thermo Fisher Scientific, A11005; 1:200) or Alexa Fluor 488 (Thermo Fisher Scientific, A11008; 1:200) secondary antibody for 1 h at 37°C. Following three additional PBS washes, sections were stained with DAPI for 20 min at room temperature in the dark, then washed three times more with PBS. Finally, sections were mounted with anti-fluorescence quenching medium and coverslipped. Images were acquired using a laser scanning confocal microscopy. Integrated mean fluorescence intensity was calculated as the product of area and integrated density. Relative protein expression levels were determined based on fluorescence intensity as described above and normalized to DAPI intensity. Confocal images were merged and analyzed using ImageJ software, with the Colocalization Finder plugin employed to compute the Pearson’s correlation coefficient (Rr), which ranges from −1.0 to 1.0. Negative values denote a negative correlation while positive values indicate a positive correlation, and the strength of the correlation increases with the absolute value of Rr [57].

ATP content assay

ATP levels were measured using a luciferase-based luminescence assay with ATP detection Kit (Beyotime, S0026). Hep3B was washed twice with PBS, cells were lysed on ice using ATP detection lysis buffer and centrifuged at 12,000 × g for 5 min at 4°C. For fresh mouse liver tissues, 150 µl of ATP lysis solution were added per 20 mg of tissue, following by homogenization with a glass homogenizer (Sigma-Aldrich, D8938). The homogenate was centrifuged at 12,000 × g for 5 min at 4°C and the supernatant was collected. The fluorescence intensity of ATP in the supernatant was measured with a chemiluminotmeter, and protein concentration was measured via BCA assay (Solarbio, PC0020). ATP concentration was calculated using a standard curve from the luminescence assay and normalized to the corresponding protein concentration.

Liver cancer patient survival analysis

Clinical parameters of liver hepatocellular carcinoma cohorts were retrieved from TCGA database (https://tcga-data.nci.nih.gov/tcga/). RNA-Seq data from 240 HCC and 202 normal liver tissues were obtained from ICGC database (https://dcc.icgc.org/) for analysis of MEN1 mRNA expression. To evaluate the correlation between MEN1 expression levels and overall survival, disease-specific survival, and progression-free survival, patients in the TCGA cohort were stratified into high- and low-MEN1 expression groups based on the median MEN1 expression level, followed by Kaplan – Meier survival analysis. The GSE76427, GSE39791, and GSE63898 datasets were downloaded from GEO (http://www.ncbi.nlm.nih.gov/geo) to analyze the association of MEN1 expression with normal liver tissues, adjacent nontumor tissues, cirrhosis tissues, and HCC tumors. The relationship between MEN1 expression and autophagy-related genes in the TCGA liver cancer cohort was analyzed using an online tool (http://gepia2.cancer-pku.cn/).

Flow cytometric analysis

Mitochondrial membrane potential of liver cancer cells with or without 10 μM CCCP treatment was measured using the JC-1 assay Kit. For apoptosis analysis of HCC cells cultured in galactose or low glucose medium, staining was performed according to the protocol of ANXA5 Apoptosis Detection Kit (Keygenbio, KGA1101-100). Briefly, 2 × 10⁵ cells from each group were incubated with 500 μl of staining solution for 15 min at room temperature in the dark, followed by flow cytometric analysis. Additionally, reactive oxygen species levels in Hep3B cells with or without galactose treatment were detected using the Reactive Oxygen Species Assay Kit (Beyotime, S0033S) according to the manufacturer’s instructions, and stained cells were analyzed by flow cytometry.

RNA-seq and data analysis

Eight months after DEN and CCl₄ induction, total RNA was extracted from liver cancer tissues of WT and HMKO mice using TRizol reagent (Yeasen, 10606ES60). RNA quality and concentration were assessed using a Nanodrop^TM OneCspectrophotometer (Thermo Fisher Scientific). For samples with high-quality RNA, cDNA libraries were constructed using the KC-DigitalTM RNA-seq Library Preparation Kit (HealthSeq, DR08502) according to the manufacturer’s protocol. The kit eliminates duplication bias during sequencing by labeling pre-amplified cDNAs with an 8-base random unique molecular identifier. Library products of approximately 150 nt in size were enriched, quantified, and sequenced on an Illumina Novaseq 6000 platform at the Kangce Technology Co., Ltd (Wuhan, China). Sequencing depth averaged ~30 million 150-nucleotide paired-end reads per sample.

For RNA-seq data analysis, raw sequencing data were quality-filtered using FastQC (version 0.11.5), low-quality reads were discarded and reads containing adaptor sequences were trimmed to generate clean reads. These clean reads were further processed with in-house scripts to eliminate duplication bias introduced during library preparation and sequencing. Clean reads were aligned to the reference genome using HISAT2 software (version 2.0.1-beta). valid mapped reads were counted per gene using featureCounts software (version v1.6.0) based on the genome annotation file, and expression levels were normalized as FPKM (Fragments per kilobase of exon per Million mapped reads). Differential expression analysis was performed using R package. Genes with a FDR < 0.05 and absolute log₂ fold change (log₂FC) >1 were defined as significantly DEGs. Draw scatter plots were generated for all DEGs, with red indicating upregulated genes and blue indicating downregulated genes, and volcano plots were also constructed. Hierarchical clustering analysis and heatmap visualization of DEGs were performed using the “pheatmap” package in R software (version 1.0.12; https://github.com/raivokolde/pheatmap). Gene Ontology (GO) and KEGG enrichment analyses were conducted using KOBAS software (version 2.1.1), with a significance cutoff of adjusted p value < 0.05. GSEA was performed using GSEA software (version 4.0.3; https://github.com/GSEA-MSigDB/gsea-desktop). Normalized clean reads from RNA-seq data were analyzed against gene signatures from the MSigDB database.

Metabolomic analysis

Liver tumor tissues (20 mg) from WT and HMKO mice after DEN and CCl₄ induction were weighed, and mixed with 400 μl of 7:3 methanol:water solution. The mixture was shaken at 1,500 × g for 5 min, placed on ice for 15 min, and centrifuged at 12,000 × g for 10 min. A 300 μl aliquot of the supernatant was collected and stored at −20°C for 30 min, then centrifuged again at 12,000 × g for 3 min. A 200 μl aliquot of the resulting supernatant was transferred for liquid chromatography-mass spectrometry (LC-MS) analysis. All samples were analyzed using two LC-MS methods. Positive ion mode: separation was performed on a Waters ACQUITY Premier HSS T3 Column (1.8 µm, 2.1 mm × 100 mm) with solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile) using the following gradient: 5 to 20% mobile phase B over 2 min, increased to 60% mobile phase B over the next 3 min, increased to 99% mobile phase B over 1 min (held for 1.5 min), returned to 5% mobile phase B over 0.1 min (held for 2.4 min). Analytical conditions: column temperature 40°C; flow rate 0.4 ml/min; injection volume 4 μl. Negative ion mode: the elution gradient was identical to that used for positive ion mode. MS analysis was carried out using electrospray ionization in positive and negative ion modes using full-scan detection over an m/z range of 75–1000 at a resolution of 35,000. Putative metabolites were identified by matching accurate m/z values of detected peaks with an in-house database integrated from HMDB, KEGG, CHEBI, and other public databases. Significant metabolites were determined using Student’s t test and visualized using the MetaboAnalyst platform.

Statistical analysis

Statistical analyses were performed using GraphPad Prism software 10.0. No statistical methods were used to predetermine sample sizes. All experiments were conducted with at least three technical replicates and three independent biological replicates. Numerical data are presented as mean ± SD. Differences between two groups were analyzed using a two-tailed unpaired Student’s t test while differences among three or four groups were determined using one-way analysis of variance (ANOVA). Survival curves were generated by the Kaplan-Meier method and compared with the log-rank (Mantel-Cox) test. The sample size (n) refers to the number of biological replicates per experiment. A p-value < 0.05 was considered statistically significant.

Funding Statement

This work was supported by grants from National Natural Science Foundation of China (82573918 and 32160144); Natural Science Foundation of Guizhou Provincial Education Department (Young Science and Technology Top-notch Talent Project) ([2024]327); Guizhou Provincial Key Technology R&D Program Project ([2023]191); Guizhou Provincial Basic Research Program (Natural Science) (ZD[2025]061 and ZD[2025]015); Guizhou Provincial Natural Science Foundation ([2021]4029 and [2022]4017); Excellent Young Talents Plan of Guizhou Medical University ([2022]106); Guizhou Provincial Health Commission Joint Medical Research Fund for High-Quality Development (2024GZYXKYJJXM0050); Natural Science Foundation of Guizhou Provincial Education Department (Young Science and Technology Talent Project) [[2024]327].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data supporting the finding of this study are available from the corresponding author upon reasonable request. Raw RNA-seq data have been deposited in the National Center for Biotechnology Information (NCBI) under accession number SUB15438132 for WT and HMKO mice. All other relevant data are included in the article or uploaded as supplementary information.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2026.2677182