VDAC1 as prognostic marker and therapeutic target in lung adenocarcinoma: a study integrating bioinformatics and experimental validation

Abstract

Background

Research on the expression and molecular mechanisms of voltage-dependent anion channels (VDACs) in lung adenocarcinoma (LUAD) remains limited.

Materials and methods

Multiple datasets were utilized to analyze VDACs expression in LUAD and investigate the clinical significance of VDACs-associated genes and signaling pathways. The database analysis results were further validated through cellular and animal experiments.

Results

The expression levels of VDAC1 and VDAC2 increase with advancing tumor stage. Subsequent survival analysis revealed that elevated mRNA expression of VDAC1 (HR = 1.6, log-rank P = 0.0015), VDAC2 (HR = 1.5, log-rank P = 0.0088), and VDAC3 (HR = 1.7, log-rank P = 0.0026) was significantly associated with shorter overall survival in LUAD patients. The VDACs-based prognostic signature holds significant value for risk stratification, with VDAC1 demonstrating the poorest prognostic impact. We demonstrated that both in vitro and in vivo experiments consistently showed that the combination of trametinib with VBIT-12 markedly suppresses tumor growth.

Conclusion

This study integrates bioinformatic insights with experimental validation to clarify the clinical significance and therapeutic potential of VDACs in LUAD, providing a valuable foundation for prognosis assessment and targeted therapy development.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12672-025-03955-5.

Introduction

Worldwide cancer statistics consistently identify lung cancer as the leading cause of cancer-related mortality [1], with a global five-year survival rate remaining below 20% [2]. Non-small cell lung cancer (NSCLC) accounts for approximately 85% of all cases, with lung adenocarcinoma (LUAD) being the most prevalent histological subtype [3, 4]. Although treatment modalities for LUAD have expanded to include targeted therapies and immunotherapy alongside conventional approaches, many patients are diagnosed at advanced stages due to the absence of early specific symptoms, leading to unfavorable outcomes [5]. Therefore, the development of biomarkers with high specificity and sensitivity as prognostic indicators is crucial to improve clinical outcomes and promote personalized treatment strategies.

As the predominant protein in the mitochondrial outer membrane, the voltage-dependent anion channel (VDAC) serves as a crucial gateway governing metabolite exchange and energy homeostasis between mitochondria and cytosol [6]. In normal cellular metabolism, VDAC facilitates the transfer of ATP produced in the mitochondria to the cytoplasm to support cellular activities, while transporting ADP from the cytoplasm back into the mitochondria for ATP synthesis [7]. Beyond its metabolic role, VDAC is a key regulator of apoptosis, interacting with proteins such as Bcl-2 family members and hexokinase to modulate mitochondrial-mediated cell death [8].

VDAC has been increasingly implicated in the pathogenesis of diverse diseases, spanning neurodegenerative, cardiovascular, metabolic, and oncological conditions [6]. In mammals, VDAC exists as three isoforms: VDAC1, 2, and 3. VDAC1 and 2 are the most abundantly expressed, whereas VDAC3 shows the lowest expression in most mammalian cells, including cancer cells [9]. VDAC1 expression is significantly elevated in mesothelioma patients, with expression levels increasing at later disease stages, correlating with poor prognosis. Conversely, decreased VDAC1 expression drives metabolic rewiring, resulting in the inhibition of tumor growth and modulation of the tumor microenvironment (TME) [10]. In glioma cells, knockdown of VDAC1 reduces the levels of metabolism-related enzymes and cancer stem cell markers. Ultimately, VDAC1 depletion suppresses tumorigenesis by inhibiting metabolic activity and reversing malignant phenotypes, thereby effectively blocking oncogenic signaling cascades [11–13].

In colorectal cancer, elevated peripheral blood cholesterol promotes CD8⁺ T cell exhaustion and VDAC1 expression, facilitating immune evasion [14]. Similarly, elevated VDAC1 expression in breast cancer correlates with adverse clinical outcomes and modulates immune cell infiltration patterns [15]. The pleiotropic effects of VDAC1 inhibition in bladder cancer spanned tumor-intrinsic mechanisms (growth and stemness suppression) and microenvironmental reprogramming (vascular, immune, and stromal compartment modulation), suggesting comprehensive therapeutic potential [16].

Targeting VDAC2 in melanoma demonstrates remarkable enhancement of both tumor suppression and immunotherapy effectiveness. VDAC2 blockade coordinates CD8⁺ T cell and interferon-γ-dependent adaptive immunity with tumor cell-intrinsic innate immune responses, positioning VDAC2 as a bifunctional therapeutic target to combat immune escape mechanisms [17]. Moreover, silencing VDAC2 expression increases the sensitivity of melanoma cells to erastin [18]. Compared to wild-type cells, erastin shows enhanced antitumor potency against HRAS-, KRAS- or BRAF-mutated cancers via VDAC2-targeted mechanisms [19]. Elevated VDAC2 expression has also been reported in most hepatocellular carcinomas, where Celastrol directly targets VDAC2 to induce mitochondria-dependent cell death and effectively inhibit tumor cell proliferation [20, 21]. In contrast, VDAC2 knockout reduces the cytotoxic efficacy of anti-cancer drugs against colon cancer cells and diminishes their tumor-suppressive activity, indicating that VDAC2 is required to promote drug-induced tumor cell apoptosis [22]. However, the function of VDAC2 appears context-dependent, as its expression is downregulated in bladder cancer, where it acts as a growth suppressor [23]. Furthermore, VDAC2 has been implicated in increasing the sensitivity of renal cell carcinoma cells to erastin [24].

Research on VDAC3 remains relatively limited compared to that on other isoforms. Current evidence, however, suggests its potential as a therapeutic target in breast cancer, where dankastatin B exerts anti-tumor effects via direct interaction with VDAC3 [25]. Elevated VDAC3 expression has also been observed in colorectal adenocarcinoma (COAD), with patients exhibiting low VDAC3 expression showing significantly shorter overall survival,conversely, higher VDAC3 expression correlated with later pathological stages. This suggests that VDAC3 may function as an independent prognostic marker for COAD [26]. Additionally, VDAC3 knockout induces resistance to erastin, indicating that VDAC3 contributes to the anti-tumor effects of erastin [3, 19].

In summary, the three VDAC isoforms generally promote tumor progression and are are regarded as potential therapeutic targets in most studies. However, some findings suggest that VDAC2 and VDAC3 may also exhibit tumor-suppressive effects under certain conditions. Notably, no systematic analysis of all three VDAC isoforms has been reported in LUAD. To address this gap, we conducted a comprehensive analysis of public databases to delineate the expression profiles, correlation with disease stages, and prognostic value of VDAC1/2/3 in LUAD. Furthermore, cellular and animal studies validated VDAC1 inhibition as a viable therapeutic strategy. Importantly, combining a VDAC1 inhibitor with trametinib synergistically enhanced the anti-tumor efficacy of trametinib, revealing a promising combination therapy for LUAD.

Materials and methods

GEPIA 2

GEPIA 2 (http://gepia2.cancer-pku.cn/) is a versatile online platform providing customizable features, including tumor/normal differential expression profiling, analysis across cancer types or pathological stages, survival prognosis evaluation, homologous gene identification, correlation studies, and dimensionality reduction. We analyzed the differential expression of VDAC1, VDAC2, and VDAC3 across 31 cancers and corresponding paracancerous tissues, as well as in lung adenocarcinoma (LUAD), considering cancer stage (Log2FC Cutoff:1;, p-value Cutoff: 0.01) overall survival (Cutoff-High(50%); Cutoff-Low(50%)), and disease-free survival (Cutoff-High(50%); Cutoff-Low(50%)). The mRNA expression data for VDAC1, VDAC2, and VDAC3 were individually analyzed. Their expression patterns were visualized using the "Boxplots" module, correlations with pathological stages were assessed using the "Stageplots" module, and prognostic significance was evaluated through "Survival Analysis". VDAC1, VDAC2, and VDAC3 were entered with the "Multiple Genes Comparison" option selected to compare their mRNA expression levels across different cancer types.

Assistant for clinical bioinformatics

The Assistant for Clinical Bioinformatics platform (https://www.aclbi.com/static/index.html#/) includes sample data from 33 tumor types in the TCGA database, seven pediatric/hematologic tumors in the TARGET database, tumor and non-tumor samples in the GEO database, cell line data from the CCLE database, and information on 24 tumor types from the ICGC database. This platform was used to generate Sankey plots for VDACs, develop prognostic models, perform landscape mutation analysis of VDAC1, and conduct expression analysis across different cellular subpopulations. Sankey Diagram and Mutation Analysis: Using the TCGA-LUAD cohort, a Sankey diagram was plotted (choose Single Sample and Sankey Diagram) to relate VDAC1/2/3 expression (distribution variables) to clinical features (status, pTNM stage, gender, age). The VDAC1 mutation landscape was analyzed (choose Complex Analysis and Genetic Mutation Landscape). Data normalized as log2(TPM + 1); p < 0.05 deemed significant.

Prognostic analysis was performed using the Prognosis, Gene and Survival module for the LUAD (Non-Small Cell Lung Cancer) cohort. VDAC1, VDAC2, and VDAC3 were entered as candidate genes. The analysis was configured with the grouping method set to "Medium", survival times at 1, 3, and 5 years, and "OS" (Overall Survival) as the prognostic parameter. This generated the LASSO (Least Absolute Shrinkage and Selection Operator) coefficient profiles for the VDAC family, employing tenfold cross-validation for feature selection. Subsequently, a multivariate Cox regression analysis was applied to construct the prognostic model. The optimal model was identified through an iterative process using the step function. For Kaplan–Meier curves, statistical significance was assessed with the Log-rank test (reporting p-values), and hazard ratios (HR) with 95% confidence intervals (CI) were derived from univariate Cox regression. Gene expression data were normalized using log2(TPM + 1), and a p-value < 0.05 was considered statistically significant.

The Assistant for Clinical Bioinformatics downloaded the single-cell data in.h5 format and its annotation results from TISCH. Use the R software MAESTRO and Seurat to process and analyze the single-cell data. VDAC1 in the NSCLC was visualized using the "Single-Gene" module of the Assistant for Clinical Bioinformatics platform. This was achieved by selecting the "Lung" and "NSCLC" datasets, querying for "VDAC1", which generated a t-SNE plot and a bar chart depicting its expression across different cell types [27].

LinkedOmics

LinkedOmics (https://www.linkedomics.org/login.php) is an open-access platform hosting multi-omics datasets from 32 TCGA cancer types and 10 CPTAC cancer cohorts. The LinkInterpreter tool performs functional enrichment analyses including Gene Ontology (GO) terms, pathway annotations, network clusters, and additional functional classifications. LinkCompare enables data visualization via dynamic Venn diagrams, correlation plots, interactive heatmaps, and cross-study meta-analysis, facilitating both individual cancer type and pan-cancer multi-omics investigations.This platform was used to generate heatmaps illustrating positive and negative correlations with VDAC1, as well as to perform GO and KEGG pathway analyses. For the LUAD cohort, RNA-seq data from the Hiseq platform were analyzed to identify genes correlated with VDAC1 expression using a Pearson correlation test. The results were visualized in a heatmap displaying genes positively and negatively correlated with VDAC1. Subsequently, functional enrichment analysis was performed on the correlated gene sets using GO-BP and KEGG pathways.

The human protein atlas

The Human Protein Atlas (HPA, https://www.proteinatlas.org/) systematically maps the spatial distribution of all proteins encoded by the human genome across human cells, tissues, and organs. Utilizing antibody-based immunohistochemistry (IHC) and immunofluorescence (IF) technologies, it conducts large-scale protein localization analyses on tissue microarrays (TMA) and cell lines. The database encompasses protein expression data for over 20,000 human genes, covering 44 major tissue types. Immunohistochemical images for VDAC1/2/3 in normal lung tissue and LUAD samples were retrieved from the Human Protein Atlas (HPA) database. This was accomplished by searching for each gene, selecting the "Tissue" section for lung tissue, and the "Cancer" section for LUAD.

Xiantao academic

Xiantao Academic (https://www.xiantaozi.com/) is primarily an online platform that leverages public databases—such as GEO, TCGA, and GTEx—to support a range of bioinformatic functions. These include dataset management, differential expression analysis, enrichment analysis, immune infiltration analysis, clinical significance evaluation, and machine learning. The diagnostic potential of the VDAC family was evaluated by generating a ROC curve. This analysis was performed using the "Clinical Significance" module, specifically the "Diagnostic Category" and "Diagnostic ROC" functions, applied to the "Lung Cancer (TCGA, LUAD)" dataset.

Cell culture, xenograft mouse model, and drug treatment

The A549 human lung adenocarcinoma cell line was acquired from the Chinese Academy of Sciences Cell Bank (Shanghai, China). Cell maintenance was performed in DMEM medium containing 10% FBS, 100 U/mL penicillin, and 100 U/mL streptomycin, incubated at 37 °C with 5% CO₂. Trametinib and VBIT-12 (purchased from MedChemExpress, Shanghai) were prepared as DMSO stock solutions and serially diluted to working concentrations prior to experimentation.

All animal studies were approved by the Ethics Committee of Shanghai Pulmonary Hospital. Five-week-old male nude mice (BALB/c) were purchased from Shanghai Model Organisms Company (Shanghai, China) and anesthetized by intraperitoneal injection of 1.25% Tribromoethanol (from Nanjing Aibei Biotechnology Co., Ltd, China) at 0.2 ml/10 g. A549 cells (5 × 10⁶ cells/0.1 mL/mouse) were subcutaneously injected into the mice to establish xenograft models. Tumors were allowed to grow to approximately 100 mm³ over 14 days, after which the mice were randomly divided into 4 groups (n = 5 per group): saline (control), VBIT-12 (10 mg/kg), trametinib (0.2 mg/kg), and trametinib (0.2 mg/kg) + VBIT-12 (10 mg/kg).

The dosing regimens for the mouse model were determined as follows. For VBIT-12, no previous application in tumor models was found; therefore, its dosage was referenced from a murine model of ulcerative colitis, where a single dose of 20 mg/kg was administered [28]. Since our study required alternate-day dosing, the concentration was halved to 10 mg/kg. For trametinib, studies in A549 mouse models reported similar inhibitory efficacy within the range of 0.1–0.5 mg/kg [29]. Therefor,e a lower dose of 0.2 mg/kg was selected.

Treatments were administered every other day until mice were euthanized with cervical dislocation (28 days), and tumors were harvested and weighed. From the first day of treatment, the longest (a) and shortest (b) tumor diameters were measured every two days using digital calipers, and tumor volume was calculated using the formula: V = (a × b²)/2. All tumor-injected mice developed measurable tumors by day 14 post-injection, with no mortality observed prior to the terminal dissection timepoint.

Flow cytometry

Cell apoptosis was assessed using FITC-Annexin V/PI double staining (Elabscience Kit). After treatment, cells were collected (300 × g, 5 min), PBS-washed, and counted. For staining, 1–5 × 10⁵ cells were resuspended in 500 μL binding buffer, then incubated with 5 μL each of Annexin V-FITC and PI (50 μg/mL) for 15–20 min (dark, RT). For identifying dead cells, Fixable Viability Stain 700 (BD Biosciences) was used. Ki-67 staining (PE Mouse Anti-Ki-67, BD Biosciences) was performed following cell membrane permeabilizationp for 40 min. Samples were processed using a BD LSRFortessa flow cytometer, and data was analyzed with FlowJo (Tree Star, Inc.).

Statistical analysis

All statistical analyses were conducted using GraphPad Prism 9 (GraphPad Software, San Diego, USA). Number of 3 times experiment repeated. Data significance was assessed using an unpaired t-test, with significance levels indicated as *p < 0.05; ** p < 0.01; *** p < 0.001; and **** p < 0.0001.

Results

Transcriptional expression of VDACs in cancers

In 31 cancer types, VDAC1, VDAC2, and VDAC3 were found to be significantly overexpressed in lymphoid neoplasm diffuse large B-cell lymphoma (DLBC) and thymoma (THYM) tumors, with little or low expression in adjacent normal tissues. Conversely, in acute myeloid leukemia (LAML), VDAC1, VDAC2, and VDAC3 were expressed at lower levels in tumor tissues but showed higher expression in paracancerous tissues. In lung adenocarcinoma (LUAD), a slightly higher expression of VDAC1 was observed in tumor tissues compared to paracancerous tissues, whereas VDAC2 and VDAC3 exhibited no significant differences (Fig. 1).

Fig. 1.

VDACs mRNA expression levels in diferent types of cancers. The comparison of VDAC expression in different types of cancers and normal tissues from the GEPIA2 database

Expression of VDAC1 and VDAC2 correlates with tumor progression

Comparative analysis revealed no significant differences in VDAC1/2/3 expression levels between tumor tissues and matched adjacent normal tissues (Fig. 2A–C). Immunohistochemical analysis revealed that VDAC1, VDAC2, and VDAC3 were absent in lung normal tissues but exhibited varying degrees of positive staining in LUAD (Supplement Fig. 2). However, VDAC1 and VDAC2 expression increased with tumor stage, while VDAC3 expression showed no clear association with staging (Fig. 2D–F). In LUAD patients with high VDAC1 and VDAC2 expression, the numbers of deceased and surviving patients were approximately equal, whereas in patients with low expression levels of these genes, the number of survivors exceeded the number of deaths. Most patients in stage I exhibited low VDAC1 and VDAC2 expression, with the proportion of high expressers increasing progressively in stages II, III, and IV. In stage I, the majority of patients were female, while stages II, III, and IV showed an equal distribution between males and females. Patients over 60 years of age dominated both genders (Fig. 2G–H).

Fig. 2.

Expression of VDACs in tumors, relationship with clinical staging and distribution characteristics of LUAD from the GEPIA2 and Assistant for Clinical Bioinformatics dataset. A–C The expression of VDAC (1,2,3) in normal and cancer samples basing on LUAD from GEPIA2. D–F The expression of VDAC (1,2,3) in different stages of LUAD patients from GEPIA2. G–I Distributional characteristics of VDAC (1,2,3) in LUAD cancer patients from Assistant for Clinical Bioinformatics

For VDAC3, the survival pattern differed slightly; patients with high VDAC3 expression had an almost equal ratio of survival to death, whereas those with low expression predominantly survived. In stage I, the proportions of high and low VDAC3 expression were similar. In stage II, low expressers slightly predominated, and in stages III and IV, the proportions of high and low expressers were nearly equivalent. Female predominance was noted in stage I, but stages II to IV exhibited a balanced male-to-female ratio. Again, the majority of patients in all stages were over 60 years old (Fig. 2I).

Prognostic value of VDACs in LUAD

Higher VDAC1 and VDAC2 expression levels were associated with reduced overall survival but showed no significant impact on disease-free survival (Fig. 3A–D). In contrast, elevated VDAC3 expression predicted poorer outcomes for both overall and disease-free survival (Fig. 3E–F). These findings suggest that VDAC could serve as prognostic markers in LUAD.

Fig. 3.

High VDACs expression is associated with poor OS in LUAD patients. The survival of VDAC1 (A-B), VDAC2 (C-D), VDAC3 (E–F) expression (high and low) in GEPIA2 dataset

Prognostic signature of VDACs is valuable for risk stratification

VDAC expression profiles were analyzed using least absolute shrinkage and selection operator (LASSO)-penalized Cox regression for prognostic modeling.The regression coefficients are visualized in Fig. 4A, with VDAC1 emerging as the gene with the highest score. The optimal model included all three genes (Fig. 4B). Samples with higher risk scores demonstrated shorter survival times and higher VDAC expression levels. Kaplan–Meier survival curves showed that the low-risk group had a significantly higher survival probability (p = 0.000463) in this validation cohort. The diagnostic ROC curve (AUC) values are shown in Fig. 4C, with a maximum AUC of 0.893, indicating high predictive accuracy. Overall, the VDAC-based prognostic signature proved useful for risk stratification.

Fig. 4.

LASSO coefficient profiles of VDACs. A–B Cvfit and lambda curves showing the least absolute shrinkage and selection operator (LASSO) regression was performed with the minimum criteria. C The curve of risk score, kaplan–Meier survival analysis of the three-gene signature, heatmap of the expression profles of the three prognostic genes in low and high risk group, and diagnosis depends on ROC analysis the of the three-gene signature in LUAD from Assistant for Clinical Bioinformatics and Xiantao Academic

Somatic nucleotide variation profiling in LUAD

A horizontal histogram displayed the genes with the highest mutation frequencies in patients with LUAD, including TP53 (48%), TTN (45%), MUC16 (40%), CSMD3 (37%), and RYR2 (35%) (Fig. 5A, B). Missense mutations were the most prevalent mutation type, with single nucleotide polymorphisms (SNPs) far exceeding insertions (INS) or deletions (DEL) (Fig. 5B). The predominant base substitution was C > A (Fig. 5B). The distribution of mutations per sample is shown in Fig. 5B, where each color in the boxplot represents a specific mutation type. As observed in many cancer genomes, especially solid tumors, localized hypermutations were a prominent feature.

Fig. 5.

A Landscape of mutation profles in LUAD samples from Assistant for Clinical Bioinformatics. A Mutation information of each gene in each sample was shown in the waterfall plot, where diferent colors with specifc annotations at the bottom meant the various mutation types. The barplot above the legend exhibited the number of mutation burden. B Cohort summary plot displaying distribution of variants according to variant classifcation, type and SNV class. Bottom part (from left to right) indicates mutation load for each sample, variant classifcation type. A stacked barplot shows top ten mutated genes

Enrichment analysis of VDAC1 co-expressed genes in LUAD

To explore the potential mechanisms and biological roles of VDAC1 in LUAD, co-expression analysis was performed using LinkedOmics. The top 50 positively and negatively correlated genes are presented as heatmaps (Fig. 6A, B). GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses indicated upregulation in RNA localization, tRNA metabolism, ribosome, and proteasome processes, while pathways such as bone development and heat morphogenesis were downregulated (Fig. 6C, D).

Fig. 6.

Analysis of VDAC1 and its co-expression genes in LUAD from LinkedOmics. A, B Top 50 positively and negatively corrected genes with VDAC1. C, D GO analysis and KEGG pathway enrichment analysis of VDAC1 and its corrected

Distribution of VDAC1 in different cell types of NSCLC

Single-cell RNA sequencing data comprising 9,055 T-cell and monocyte profiles from 14 NSCLC patients revealed major cell types including CD4Tconv, CD8T, CD8Tex, Mono/Marco, Tprolif, and Treg (Fig. 7A). VDAC1 was most highly expressed in Tprolif cells, followed by Treg cells (Fig. 7B, C). Both cell types are known to suppress anti-tumor immunity. Multiple cancer model studies have demonstrated the abundant presence of Tprolif and Treg cells in the TME [30, 31]. Consistent with our results, the tumor-promoting factor NUSAP1 is also predominantly expressed in Tprolif cells in LUAD [32].

Fig. 7.

VDAC1 expression levels in different cell subpopulations of non-small cell lung cancer patients from Assistant for Clinical Bioinformatics. A The t-SNE plot of single-cell clustering. B The t-SNE plot of the expression distribution of VDAC1 in different cells, where different colors represent expression abundance. C The bar chart of the expression abundance of VDAC1 in different cells

VBIT-12 enhances the therapeutic efficacy of trametinib

Our study established a correlation between elevated VDAC1 expression and advanced tumor stage coupled with poor prognosis in lung adenocarcinoma (LUAD). To evaluate the therapeutic potential of VDAC1 inhibition, A549 cells were treated with the VDAC1 inhibitor VBIT-12 and the MEK inhibitor trametinib. Individually, both agents effectively induced apoptosis and suppressed proliferation of A549 cells in vitro (Fig. 8A–C; Supplement Fig. 1, 4). Notably, the combination therapy demonstrated superior anti-tumor efficacy compared to either agent alone (Fig. 8B, C; Supplement Fig. 1, 4). In xenograft mouse models, while VBIT-12 (TGI:33.79%) monotherapy showed minimal efficacy, trametinib (TGI:70.38%) alone significantly suppressed tumor growth. Critically, the combination therapy (TGI:94.69%) demonstrated superior efficacy, surpassing either single agent (Fig. 8D–F; Supplement Fig. 3A). This enhanced anti-tumor activity was achieved without significant body weight loss, supporting a favorable safety profile for the combination (Supplement Fig. 3B). Collectively, these findings indicate that the combination of trametinib and VBIT-12 exerts potent anti-tumor effects in vitro and in vivo, positioning it as a highly promising therapeutic strategy for LUAD.

Fig. 8.

The effectiveness of the VDAC1 inhibitor in combination with trametinib in A549 cells and xenograft models. A Apoptotic effects of different concentrations of VBIT-12 on A549 cells. B, C Apoptosis and proliferation effects of VBIT-12 (10 µM) combined with trametinib (0.02 µM) on A549 cells. D Nude mice (n = 5) with xenograft tumors by injecting subcutaneously A549 cells (5 × 10⁶cells/0.1 mL) were treated with saline, VBIT-12 (10 mg/kg), trametinib (0.2 mg/kg), trametinib (0.2 mg/kg) + VBIT-12 (10 mg/kg) respectively, over a period of 14 days. E, F The last tumor volume (E) and mean tumor weight (F) (n = 5). Significant difference in the unpaired t-test with *p < 0.05 and **p < 0.01, ***p < 0.001,****p < 0.0001, ns = not significant

Discussion

This study provides a systematic analysis of the expression patterns and prognostic significance of the VDAC family in LUAD, with a dedicated focus on the clinical and therapeutic relevance of VDAC1. Our findings establish VDAC1 as an independent prognostic factor associated with adverse outcomes in LUAD. Furthermore, the combination of the VDAC1 inhibitor VBIT-12 and the MEK inhibitor trametinib demonstrated marked synergistic antitumor activity in both in vitro and in vivo models. These results not only confirm VDAC1 as a promising therapeutic target in LUAD but also introduce a novel combination strategy centered on VDAC1 inhibition, highlighting its potential as a translatable approach to address targeted therapy resistance in this malignancy.

VDAC isoforms are widely expressed in various tissues of the human body but with varying degrees of specificity, with VDAC1 and VDAC2 being more abundantly expressed than VDAC3. The three isoforms of VDAC differ in their roles: VDAC1 has a general role, VDAC2 is involved in apoptosis, while VDAC3 participates in sex determination and development [33]. VDAC plays a crucial role in tumorigenesis and progression by driving metabolic reprogramming in cancer cells. Specifically, VDAC1 is frequently overexpressed in malignancies, enhancing glucose and glutamine uptake to fuel rapid proliferation [34, 35]. This isoform-specific upregulation is evident in prostate cancer, where VDAC1 mRNA levels are markedly elevated compared to normal tissue, while VDAC2 and VDAC3 expression remains largely unchanged [36].

Breast malignancies are characterized by significant VDAC1 upregulation, a marker of poor prognosis. Functional studies confirm that VDAC1 promotes breast cancer progression by enhancing proliferation, migration, and invasion through Wnt pathway activation. Notably, the local anesthetic lidocaine has been shown to suppress breast cancer activity and induce apoptosis via VDAC1 downregulation [37, 38]. Similarly, elevated VDAC1 expression in endometrial carcinoma correlates with aggressive clinicopathological features such as high tumor grade, deep myometrial invasion, and lymph node metastasis [39]. Beyond these cancers, VDAC1 knockdown consistently inhibits proliferation in a wide range of malignancies, includinglung, prostate, colon, glioblastoma, liver, and pancreatic cancers [40].

Our analysis also demonstrated that VDAC1 is associated with lung adenocarcinoma stage and poor prognosis, with higher VDAC1 expression linked to later stages and shorter overall survival. Furthermore, VDAC1 expression was positively correlated with genes such as HSPA4 and PPP2CA and negatively correlated with genes such as ARHGEF17 and DAPK2. In the prognostic model, VDAC1 appeared to be the most malignant and was correlated with immune cell infiltration. Single-gene analysis showed that VDAC1 was most highly expressed in Tprolif cells.

In hepatocellular carcinoma cells, knockout of VDAC1, VDAC2, and VDAC3 all resulted in inhibited cell proliferation. Analysis of the TCGA database identified VDAC2 as a prognostic indicator associated with poor outcomes [18, 41, 42]. Simultaneous silencing of VDAC1 and VDAC2 suppressed HT-29 colon cancer cell proliferation and induced G1 phase cell cycle blockade. The combined knockdown additionally attenuated mTORC1 pathway [43].

VDAC2 exhibits context-dependent roles across various cancers. It is highly expressed in thyroid and bladder cancers, where it is associated with poor prognosis and modulates sensitivity to therapeutics like sorafenib, highlighting its potential as a prognostic biomarker and therapeutic target [44–48]. Conversely, VDAC2 demonstrates tumor-suppressive functions in other malignancies. In glioma, higher VDAC2 expression in non-stem tumor cells (NSTCs) inhibits dedifferentiation,its ablation promotes the acquisition of glioma stem cell-like properties and correlates with higher tumor grade and poorer patient survival [49]. Similarly, in ovarian cancer, VDAC2 overexpression attenuates, while its knockout promotes, tumor progression [50].

Our results demonstrate that high VDAC2 expression is associated with advanced stage and poor overall survival in LUAD. This observation aligns with functional studies reporting that VDAC2 is upregulated in lung cancer and that its knockdown inhibits cancer cell proliferation and invasion while sensitizing cells to Artesunate-induced apoptosis [51–53]. Collectively, these findings position VDAC2 as a promoter of LUAD progression.

Research on VDAC3 in cancer remains limited. While it is the least expressed VDAC isoform and plays fundamental roles in development, recent studies suggest it may influence cancer outcomes. In esophageal squamous carcinoma, VDAC3 expression correlates with poor prognosis, and its knockdown increases cisplatin sensitivity [54]. Furthermore, higher VDAC3 levels are generally associated with advanced tumor stages and reduced overall and disease-free survival, indicating a potential prognostic significance.

In vitro experiments demonstrated that the VDAC1 inhibitor VBIT-12 promoted apoptosis in lung adenocarcinoma cells. Among the known VDAC1 inhibitors, 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS) enhances glycolysis, thereby improving cellular resilience and increasing the sensitivity of cancer cells to radiation. This suggests that VDAC1 inhibitors may enhance tumor radiosensitivity [55]. The VDAC-targeting agent erastin was found to stimulate AMPK activity while suppressing mTOR signaling and cell proliferation in human umbilical vein endothelial cells. These findings identify VDAC1 as a previously unrecognized upstream modulator of the mTOR pathway in endothelial cells, highlighting its therapeutic potential for anti-angiogenic treatment strategies [56]. Furthermore, VDAC1 inhibitors not only mitigate the toxic effects of cisplatin but also serve as potential therapeutic agents to increase cisplatin sensitivity, offering a meaningful combination therapy to enhance chemotherapy efficacy [57]. Our results corroborate that VBIT-12, a specific VDAC inhibitor, promotes apoptosis and suppresses proliferation in tumor cells in vitro. Moreover, the combination of VBIT-12 with trametinib demonstrates superior efficacy in both in vitro and in vivo models.

Conclusion

In conclusion, our results indicate that VDAC1 serves as an independent prognostic biomarker for lung adenocarcinoma. Moreover, both ex vivo and in vivo models demonstrated that combining trametinib with a VDAC1 inhibitor holds promise as a novel therapeutic strategy to enhance the survival outcomes of patients with lung adenocarcinoma.

Author contributions

L.C and DP.Z designed the study and planned the experimental design, L.C performed the experiments and organized the article writing. DP.Z and L.C check and correctlanguage expression. All authors contributed to writing or editing the manuscript.

Funding

The work was funded by Shanghai Hospital Development Center (Grant No. SHDC22024308).

Data availability

The data that support the findings of this study are available from from GEPIA 2 (http://gepia2.cancer-pku.cn/), The Assistant for Clinical Bioinformatics platform (https://www.aclbi.com/static/index.html#/ LinkedOmics(https://www.linkedomics.org/login.php), The Human Protein Atlas (HPA, https://www.proteinatlas.org/) and Xiantao Academic (https://www.xiantaozi.com/). Other data are available from from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted with the approval of the Ethics Committee of Shanghai Pulmonary Hospital, School of Medicine, Tongji University (K25-577). The maximum tumor size approved by the Ethics Committee did not exceed 1500 mm³. All tumors in this animal study were confirmed to have a maximum volume not exceeding 1500 mm³. All animal experiments were conducted in accordance with the ARRIVE guidelines 2.0.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.