Pan-cancer atlas of ITGA6 unveils its multifaceted oncogenic roles and therapeutic potential in uveal melanoma

Abstract

Background

Integrin alpha 6 (ITGA6) has been implicated in tumorigenesis and progression, but its roles across cancer types and its functional significance in uveal melanoma (UVM) remain incompletely characterized.

Methods

We performed an integrated pan-cancer analysis of ITGA6 expression, diagnostic performance, prognostic relevance, immune associations, pathway activity, and drug sensitivity. In addition, we evaluated its functional role in UVM using in vitro assays, including proliferation, migration, invasion, and colony formation experiments following ITGA6 knockdown.

Results

ITGA6 was upregulated in multiple tumor types, including lung squamous cell carcinoma and hepatocellular carcinoma. Diagnostic analysis indicated strong discriminatory value in several cancers, with AUC values > 0.95 in pheochromocytoma/paraganglioma and cholangiocarcinoma, and AUC values of 0.85–0.95 in pancreatic adenocarcinoma, hepatocellular carcinoma, kidney chromophobe, and head and neck squamous cell carcinoma. Survival analyses revealed that the prognostic association of ITGA6 was cancer-type dependent: higher ITGA6 expression correlated with better outcomes in skin cutaneous melanoma and uterine corpus endometrial carcinoma, but with worse outcomes in lung adenocarcinoma, lower-grade glioma, and UVM. In UVM, elevated ITGA6 expression was associated with poorer overall survival, progression-free interval, and disease-specific survival, and was positively correlated with immune infiltration and malignant functional states. ITGA6 expression showed positive correlations with all 14 functional state gene sets in UVM, with particularly strong correlations for angiogenesis, proliferation, and stemness. Pathway analyses further suggested that high ITGA6 expression was associated with epithelial-mesenchymal transition and TNF-α/NF-κB signaling. Drug sensitivity analyses identified several compounds that may be associated with ITGA6-related transcriptional signatures. Functional experiments showed that ITGA6 knockdown significantly inhibited proliferation, migration, invasion, and colony formation in UVM cells.

Conclusions

These findings suggest that ITGA6 may serve as a potential diagnostic and prognostic biomarker in selected cancers and support a tumor-promoting role for ITGA6 in UVM.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-026-16287-6.

Introduction

Cancer is a highly heterogeneous group of diseases and remains one of the leading causes of death worldwide. According to the latest cancer statistics in 2025, cancer remains the first or second leading cause of death in most countries worldwide [1, 2]. Globally, cancer mortality rates have declined, driven by improved prevention of risk factors, expanded early detection programs, and advancements in therapeutic interventions [3]. Early diagnosis significantly reduces mortality risks for breast, cervical, colorectal, lung, and prostate cancers [4]. With the development of bioinformatics tools and public resources such as The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) [5], researchers can now perform pan-cancer analyses to investigate cancer-related genes more systematically [6, 7]. These platforms also enable systematic evaluation of the associations between individual genes and clinical prognosis, tumor progression, and immune infiltration [8].

Integrins are a family of widely expressed cell adhesion molecules that mediate interactions between cells and extracellular matrix ligands, thereby regulating biological processes such as thrombosis, inflammation, and cancer progression [9, 10]. Structurally, integrins are heterodimeric glycoproteins composed of α and β subunits [11–13]. Integrin α6 (ITGA6), an important member of the integrin family, has been reported to be dysregulated in multiple malignancies and is associated with tumor cell proliferation, metastasis, and therapeutic resistance [14, 15]. For instance, previous studies have demonstrated that ITGA6 expression correlates with intravesical recurrence risk in bladder cancer and drives tumor cell proliferation and invasion [16]; it also modulates platinum-based chemotherapy responses in epithelial ovarian cancer, where ITGA6 inhibition enhances drug sensitivity [17]. Furthermore, ITGA6 suppression has been shown to inhibit proliferation and metastasis in hepatocellular carcinoma, bladder cancer, and laryngeal squamous cell carcinoma [18–20].

Pan-cancer analyses provide a useful framework for comparing gene functions across tumor types and identifying shared as well as context-specific molecular features [21]. Although accumulating evidence has implicated ITGA6 in oncogenesis and prognosis, comprehensive pan-cancer investigations of ITGA6 remain limited [22]. In this study, we comprehensively analyzed ITGA6 expression patterns and their associations with pan-cancer diagnostics, prognostic stratification, tumor immune microenvironment features, activation of cancer-related pathways, and drug sensitivity, identifying a significant correlation between ITGA6 and uveal melanoma (UVM). Building on these findings, we further elucidated the mechanistic role of ITGA6 in UVM pathogenesis and validated its functional impact through experimental models.

Materials and methods

Data acquisition and processing

Pan-cancer ITGA6 expression profiles and clinical features were collected from TCGA (https://cancergenome.nih.gov/) and Genotype-Tissue Expression (GTEx, https://gtexportal.org/) via the UCSC Xena Browser [23]. ITGA6 mutation frequency in TCGA cohorts was calculated and analyzed using cBioPortal (https://www.cbioportal.org/). TCGA samples underwent comprehensive pan-cancer analysis via SangerBox 3.0 (http://sangerbox.com/) [24]. ITGA6 expression correlations with the tumor immune microenvironment were explored using multi-algorithm immune infiltration analysis in TIMER2.0 (http://timer.cistrome.org/). Human normal tissue and cancer cell line expression profiles were obtained from the Human Protein Atlas (HPA, https://www.proteinatlas.org/). Chemotherapy data from GDSC (https://www.cancerrxgene.org/), CTRP (http://portals.broadinstitute.org/ctrp/), and PRISM were aggregated to assess drug-ITGA6 expression interactions. Cancer immune cycle data were sourced from the Tumor Immunophenotype (TIP, https://biocc.hrbmu.edu.cn/TIP/) database. Transcriptomic data were obtained from the TCGA Pan-Cancer Atlas project, which includes molecular profiles of 33 cancer types. We used the uniformly processed and batch-corrected RNA-seq dataset (EBPlusPlusAdjustPANCAN_IlluminaHiSeq_RNASeqV2.geneExp.tsv) to minimize platform-related biases and ensure comparability across samples. Clinical and survival data were retrieved from the TCGA-CDR resource. After intersecting expression and clinical datasets, a total of 11,060 samples (10,323 tumor and 737 normal samples) were included for downstream analyses. Given that this study aimed to perform systematic pan-cancer analysis rather than model construction, no random split into training and validation cohorts was performed. For specific analyses, sample sizes varied after quality control. Outliers were removed in differential expression analysis based on z-score filtering, and samples with missing or zero survival time were excluded from survival analyses.

To prepare the gene expression matrix for differential expression analysis, a standard preprocessing workflow was applied. First, low-expressed or unreliable genes were filtered out, retaining only those with expression values greater than 0.1 in more than half of the samples, in order to reduce background noise and improve the robustness of downstream analyses. Since the merged TCGA and GTEx datasets originated from different experimental platforms, significant non-biological batch effects were present. To correct for these technical variations across sequencing platforms and centers, the “ComBat” algorithm was applied to adjust the merged expression matrix. The original expression data, obtained from the Xena platform, had already been uniformly normalized in the form of log2(TPM + 1) or log2(FPKM-UQ + 1). These normalized values were used directly in all subsequent analyses.

Expression and genetic variation analysis

ITGA6 mRNA expression differences between tumor and normal tissues were compared using Wilcoxon rank-sum tests, followed by paired-sample validation. Protein-level expression was externally verified using GEO and Clinical Proteomic Tumor Analysis Consortium (CPTAC) datasets. Organ-specific ITGA6 expression was visualized using the gganatogram R package. ITGA6 mutation frequency, variant types, copy number alterations (CNAs), and genetic alteration traits were extracted from cBioPortal (http://www.cbioportal.org). Immunohistochemistry (IHC) staining images were obtained from the HPA database. The diagnostic potential of ITGA6 across cancers was evaluated by calculating area under the curve (AUC) values using the pROC R package.

Survival analysis and clinical outcomes

Survival data were obtained from the TCGA database. The correlation between ITGA6 expression and clinical outcomes—including overall survival (OS), disease-specific survival (DSS), progression-free interval (PFI), and disease-free interval (DFI)—was assessed using the “survival” and “survminer” R packages. Kaplan–Meier analysis and univariable Cox regression were performed to evaluate whether ITGA6 functioned as a risk or protective factor across different cancer types. Results from the Cox regression analyses were visualized using the “forestplot” R package. To account for multiple testing across the genome or transcriptome and control the false discovery rate, all p-values derived from log-rank tests were adjusted using the Benjamini–Hochberg (BH) procedure. Thus, all reported significant p-values have been corrected for multiple comparisons.

To determine whether ITGA6 expression is an independent prognostic factor, univariate and multivariate Cox proportional hazards regression analyses were performed in the TCGA-UVM cohort. Variables with potential clinical relevance, including ITGA6 expression, pathologic stage, age, gender, and BMI, were included in the analysis. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. Variables with P < 0.1 in univariate analysis or with clinical significance were further included in the multivariate model.

Pathway and mechanistic analysis

To investigate the functional mechanisms of ITGA6, tumor specimens were stratified into high- and low-expression groups based on the upper and lower 30th percentiles of ITGA6 expression, which allows clear separation of expression groups while maintaining adequate sample size for statistical analysis. Gene set activity was compared between these groups using Gene Set Enrichment Analysis (GSEA), focusing on 50 hallmark gene sets and 83 metabolic pathway gene sets from the Molecular Signatures Database (MSigDB), including Hallmark, KEGG metabolic, immune-related, and aging-related collections.

Differential expression analysis was conducted with the limma package to obtain log2 fold changes (log2FC) for all genes, which were subsequently ranked by log2FC. Enrichment analysis for Hallmark and KEGG metabolic gene sets was performed using the GSEA function in clusterProfiler, which calculated enrichment scores (ES) and assessed significance with multiple hypothesis correction. Results were visualized using bubble plots.

Additionally, 14 functional state gene sets related to tumor behaviors such as apoptosis, invasion, and stemness were selected from the CancerSEA database. Single-sample gene set scores were computed for all specimens using the GSVA R package with z-score normalization, followed by data scaling to derive final gene set activity scores. Functional associations were evaluated by calculating Pearson correlation coefficients between ITGA6 expression and each gene set score when data followed approximately normal distributions. In cases where data did not meet normality assumptions or involved ranked variables, Spearman correlation coefficients were applied, and significance p-values were adjusted using the Benjamini-Hochberg (BH) method.

Tumor microenvironment exploration

Pan-cancer stemness was comprehensively analyzed using Spearman correlation. Immune infiltration scores and tumor purity were evaluated using seven algorithms (CIBERSORT, CIBERSORT-ABS, EPIC, MCPCOUNTER, QUANTISEQ, TIMER, XCELL). Associations between ITGA6 expression and immune-related genes (Major Histocompatibility Complex [MHC], chemokine receptors, immunestimulators, immunosuppressors) were validated [25].

Screening for ITGA6-targeting compounds

Our analysis was performed using the CMap build 02 dataset (Broad Institute, first-generation CMap), which contains gene expression profiles for 1,288 compounds. We obtained the CMAP_gene_signatures RData file. Following established analytical protocols, we utilized the Connectivity Map (CMap) database to identify small molecule drugs capable of counteracting the biological effects of ITGA6 dysregulation across various cancers. Using the XSum (eXtreme Sum) optimal feature matching method, we compared ITGA6-associated gene signatures with CMap compound profiles to generate similarity scores for all compounds. Lower scores indicate a higher likelihood that a compound can inhibit the oncogenic effects mediated by ITGA6. Of note, the compounds identified are candidate molecules for hypothesis generation; not all are FDA-approved drugs, and no comparison between CMap versions was performed.

Cell culture methods

The 92–1 and MUM2B cell lines were purchased from the Cell Resource Center of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Beijing, China), with STR profiling certification. Cells were cultured in RPMI 1640 medium (Vivacell, USA) supplemented with 10% fetal bovine serum (FBS) at 37 °C under 5% CO₂.

Western blotting

Proteins were extracted and quantified from treated cells. Equal amounts of total protein were separated by electrophoresis using the Mini-PROTEAN Tetra Cell system (Bio-Rad Laboratories Inc. Gels) and transferred to PVDF membranes. Membranes were blocked with Protein-Free Rapid Blocking Buffer (PS108P, Epizyme, China) at room temperature for 30 min. Primary antibodies were incubated with membranes overnight at 4 °C. Secondary antibodies were applied for 1 h at room temperature. Protein bands were visualized using BeyoECL Plus reagent (P0018S, Beyotime, China).

siRNA-mediated knockdown of ITGA6

To silence ITGA6 expression, two specific small interfering RNAs (siRNAs) targeting the coding region of human ITGA6 were designed and synthesized. The sense strand sequences (5′ → 3′) were as follows:si-ITGA6-1: GGAUCAAGAUCAUCCUGAATT. si-ITGA6-2: CCACCAAGAGCUCAACAAUTT. A non‑targeting scrambled siRNA was used as the negative control. Cells were transfected with 50 nM siRNA using Lipofectamine™ RNAiMAX transfection reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol.

Cell counting Kit-8 (CCK-8) assay

Cells were seeded in 96-well culture plates at a density of 2 × 10³ cells per well (five technical replicates per group). Following 24, 48, and 72 h of incubation under standard culture conditions, 10 μL of CCK-8 assay reagent (Biosharp, China) was added to each well. After 2 h of incubation at 37 °C, absorbance was measured at 450 nm using a microplate reader (BioTek Cytation 5, USA). The CCK-8 assay measures cellular metabolic activity as a proxy for relative cell viability; it does not directly distinguish between live and dead cells.

Colony formation assay

Cells were plated in 6-well plates at a density of 500 cells per well and cultured for 14 days. Colonies were fixed with 4% paraformaldehyde for 15 min, stained with 0.5% crystal violet (Biosharp, China) for 30 min, and air-dried at room temperature. Visible colonies (defined as clusters containing ≥ 50 cells) were counted manually under bright-field microscopy.

In vitromigration and invasion assays

Cell metastatic potential was assessed using BioCoat™ Transwell chambers (Corning, USA). For invasion assays, chambers were pre-coated with Matrigel (BD Biosciences, USA), whereas uncoated chambers were used for migration assays. Briefly, 5 × 10⁴ cells suspended in serum-free RPMI-1640 medium were seeded into the upper chamber. The lower chamber contained complete medium with 10% FBS as a chemoattractant. After 24-h incubation at 37 °C, cells that successfully migrated/invaded to the lower membrane surface were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet, and quantified by counting five randomly selected microscopic fields (200 × magnification) per membrane using ImageJ software (NIH, USA).

Statistical analysis

Statistical analyses were performed using SPSS 26.0 (SPSS, Chicago, IL) and GraphPad Prism 8.0 (GraphPad Software, LLC, SanDigeo, USA). All experiments were independently repeated ≥ 3 times, with data presented as mean ± Standard Deviation (SD). Two-tailed Student's t-tests were used to determine statistical significance between groups. Bioinformatics analyses and data visualization were conducted using R. Statistical significance was set at P < 0.05. Pearson or Spearman correlation analyses were selected based on data distribution and analysis purpose.

Result

ITGA6 expression analysis in human tissues

Analyses of unpaired and paired samples from TCGA revealed significantly elevated ITGA6 mRNA expression in multiple cancers, including cholangiocarcinoma (CHOL), kidney chromophobe (KICH), hepatocellular carcinoma (LIHC), and lung squamous cell carcinoma (LUSC) (Fig. 1A, B). These findings were validated in the CPTAC database, confirming elevated ITGA6 protein levels in LUSC and LIHC (Fig. 1D). Figure 1C illustrates the organ-specific distribution of ITGA6 expression in tumor and normal tissues using standardized median Z-scores derived from TPM values. As this panel is based on Z-score transformation rather than raw expression values, it reflects the relative expression pattern of ITGA6 across organs rather than absolute expression rankings. In normal tissues, ITGA6 showed relatively higher standardized expression in some organs, whereas lower relative expression was observed in others. In tumor tissues, the relative expression pattern also varied across organs, with positive Z-scores indicating expression above the overall mean and negative Z-scores indicating expression below the mean. IHC analysis identified distinct cytoplasmic immunoreactivity of ITGA6 in colorectal, skin, head and neck cancers (Fig. 1E).

Fig. 1.

Pan-cancer expression profile of ITGA6. A Comparison of ITGA6 mRNA levels between tumor and paired adjacent normal tissues in TCGA. B ITGA6 mRNA expression in tumor vs. normal tissues across cancer types. C Organ-specific distribution of ITGA6 expression in tumor and normal tissues visualized using median Z-scores derived from TPM values. GTEx normal tissue TPM data and TCGA tumor TPM data were standardized using Z-score transformation to enable cross-organ comparison of relative expression patterns. Positive Z-scores indicate expression above the overall mean, whereas negative Z-scores indicate expression below the mean. D Protein-level validation of ITGA6 expression in CPTAC. E Proportion of patients with high, medium, low, or no ITGA6 expression across cancer types, based on HPA clinical pathological data

The mutational landscape of ITGA6

To characterize ITGA6 mutations in cancer, we conducted comprehensive analyses of its pan-cancer mutational profile. The pan-cancer analysis of ITGA6 mutations revealed that missense mutations are the predominant type of single-nucleotide variants (SNVs) (Fig. 2A). Through comparative mutation rate analysis, uterine corpus endometrial carcinoma (UCEC) was identified as the cancer type with the highest frequency of ITGA6 mutations (Fig. 2B). Conversely, ITGA6 alterations were absent (0% mutation frequency) or occurred at extremely low frequencies in several other cancer types, including pheochromocytoma/paraganglioma (PCPG), kidney chromophobe (KICH), and thyroid carcinoma (THCA). Figure 2C illustrates the differential methylation levels of ITGA6 across genetic loci in a variety of cancers. The mutation rates of ITGA6 varied significantly among different cancer types, with missense mutations, deep deletions, and amplifications being the most common alterations. UCEC and melanoma exhibited the highest frequencies of ITGA6 mutations (Fig. 2D). The pan-cancer distribution of ITGA6 mutation hotspots is shown in Fig. 2E.

Fig.2.

ITGA6 mutational landscape. A The waterfall plot illustrates the distribution of ITGA6 mutations and the classification of single nucleotide variation (SNV) types across diverse cancer types. Tumor mutation burden is displayed in the top bar plot, while the right bar plot summarizes the frequencies of different SNV types. Cancer types and mutation classifications are indicated at the bottom and left of the figure, respectively. B Mutational landscape of ITGA6 and classical oncogenic signaling pathways across cancers. The horizontal axis denotes different tumor types, while the vertical axis represents ITGA6 and multiple canonical oncogenic signaling pathways. Color intensity reflects the proportion of patients carrying mutations within each tumor cohort, with darker red indicating a higher number of mutated samples and white representing the absence of mutations or a mutation frequency of zero. Numerical labels within color blocks indicate the count of samples harboring mutations in the corresponding gene for each cancer type. A label of "0" signifies that no mutations were detected in the gene coding region, and absence of a number indicates no mutations were identified in any region of the gene. C Bubble plot showing differential methylation levels of ITGA6 loci in pan-cancer analysis. X-axis: cancer types; Y-axis: methylation levels; bubble size: methylation magnitude; color code: red/blue indicates hyper-/hypo-methylation in tumors vs. normal tissues. D Mutation frequencies and types of ITGA6 alterations in top-ranked cancers. E Spatial distribution of ITGA6 mutation hotspots

Correlation of ITGA6 with pan-cancer clinical features

The association between ITGA6 expression and clinical features was systematically investigated across multiple cancer types. ITGA6 expression showed cancer type-specific associations with clinicopathological characteristics. In kidney renal clear cell carcinoma (KIRC) and the pan-kidney cohort (KIPAN), ITGA6 expression tended to be lower in more advanced T stages (Fig. 3A). Significant differences in ITGA6 expression were also observed across N stages in bladder urothelial carcinoma (BLCA) and KIPAN; however, these differences did not necessarily indicate a uniform stepwise trend (Fig. 3B). In KIRC and KIPAN, M1-stage tumors exhibited significantly lower ITGA6 expression than non-metastatic groups, suggesting an association between ITGA6 expression and metastatic status in these cancers (Fig. 3C). Similarly, ITGA6 expression tended to be lower in more advanced pathological stages in KIRC and KIPAN (Fig. 3D). Higher tumor grades were associated with lower ITGA6 expression in stomach adenocarcinoma (STAD), esophageal carcinoma (STES), KIPAN, and KIRC (Fig. 3F). In addition, females with KIPAN showed significantly higher ITGA6 expression than males, although the underlying biological basis remains unclear and requires further investigation (Fig. 3E). ITGA6 expression also exhibited age-dependent patterns, showing negative correlations in diffuse large B-cell lymphoma (DLBC), cervical squamous cell carcinoma (CESC), and KIPAN, but positive correlations in thymoma (THYM), uterine carcinosarcoma (UCS), and lung adenocarcinoma (LUAD) (Fig. 3G). Overall, these findings suggest that the relationship between ITGA6 expression and clinicopathological stage is context-dependent and varies across cancer types, rather than representing a universal stepwise marker of tumor progression.

Fig. 3.

Correlation between ITGA6 Expression and Clinical Characteristics Across Multiple Cancer Types. A Correlation between ITGA6 expression and T stage in pan-cancer. B Correlation between ITGA6 expression and N stage in pan-cancer. C Correlation between ITGA6 expression and M stage in pan-cancer. D Correlation between ITGA6 expression and tumor stage in pan-cancer. E Correlation between gender and ITGA6 expression in pan-cancer. F Correlation between ITGA6 expression and tumor grade in pan-cancer. G Correlation between age and ITGA6 expression

The diagnostic value of ITGA6

We systematically evaluated the diagnostic performance of ITGA6 across multiple cancer types using both the TCGA dataset and the combined TCGA-GTEx (Genome Tissue Expression) dataset (Fig. 4A). The results showed that the AUC values varied substantially across different cancers. In some cancer types, AUC values were close to 0.5, indicating limited discriminative ability, whereas in others, ITGA6 exhibited relatively higher diagnostic performance.

Fig. 4.

Correlation between ITGA6 expression and pan-cancer diagnosis. A Diagnostic efficiency of ITGA6 expression in different cancers visualized as AUC values. Blue and red curves represent results from TCGA-GTEx and TCGA-only datasets, respectively; (B) ROC curves for selected cancer types (KICH, PCPG, and HNSC) that showed relatively higher AUC values in panel A

Notably, ITGA6 demonstrated AUC values greater than 0.95 in paraganglioma (PCPG) and cholangiocarcinoma (CHOL), indicating strong diagnostic potential in these cancers. Additionally, in pancreatic adenocarcinoma (PAAD), liver hepatocellular carcinoma (LIHC), kidney chromophobe (KICH), and head and neck squamous cell carcinoma (HNSC), AUC values ranged between 0.85 and 0.95, suggesting relatively high diagnostic performance. Figure 4B presents representative ROC curves for selected cancer types (KICH, PCPG, and HNSC) that exhibited comparatively higher AUC values in Fig. 4A. Overall, these findings indicate that the diagnostic value of ITGA6 is cancer type–dependent and may have potential as a biomarker in specific cancers rather than uniformly across all tumor types.

Prognostic relevance of ITGA6 expression in pan-cancer

To evaluate the clinical relevance of ITGA6, prognostic analyses were performed across multiple malignancies. ITGA6 expression showed pan-cancer prognostic significance, functioning as an independent risk or protective factor depending on cancer type. ITGA6 acted as a protective factor (HR < 1) in KIRC, skin cutaneous melanoma (SKCM), and UCEC, but as a risk factor (HR > 1) in lower-grade glioma (LGG), LUAD, pancreatic PAAD, and UVM (Fig. 5A). KM survival curves revealed significant associations between ITGA6 expression and OS: high ITGA6 correlated with prolonged OS in SKCM and UCEC, whereas low ITGA6 predicted better outcomes in LUAD, LGG, HNSC, and PAAD (Fig. 5B). These findings suggest that the prognostic significance of ITGA6 is heterogeneous across cancer types, likely reflecting differences in tumor biology and context-dependent roles driven by tumor heterogeneity.

Fig. 5.

Prognostic landscape of ITGA6 in pan-cancer. A The associations between ITGA6 expression and clinical outcomes, including overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI), and progression-free interval (PFI), were assessed using univariate Cox regression and Kaplan–Meier analysis. Color coding was used to denote results: red (risk factor, indicating high expression associated with worse survival), green (protective factor), white (non-significant), and gray (missing data). Only statistically significant results (p < 0.05) from both analytical methods are displayed. B Survival curves of OS according to the expression level of ITGA6 in patients with different types of cancer

Enrichment analysis of ITGA6 across pan-cancer types

To gain insights into the functional patterns of ITGA6 across various cancer types, we integrated characteristic gene expressions to reflect the activity of the ITGA6 pathway. Based on ITGA6 expression, we selected the top 30% and bottom 30% samples with the highest and lowest expressions, respectively, for each tumor type, and conducted differential gene analysis using the limma package. Subsequently, we performed pan-cancer GSEA to reveal that ITGA6 expression may activate oncogenic pathways such as epithelial-mesenchymal transition (EMT), TNFαSignaling via NFκB, and UV response downregulated, while potentially inhibiting oxidative phosphorylation (Fig. 6A). We applied the z-score parameter from Gene Set Variation Analysis (GSVA) to 14 functional status gene sets (angiogenesis, apoptosis, cell cycle, differentiation, DNA damage, DNA repair, EMT, hypoxia, inflammation, invasion, metastasis, proliferation, quiescence, and stemness), obtained composite z-scores, and further calculated the Pearson correlation between ITGA6 expression and each gene set score. The results indicated a positive correlation between ITGA6 expression and angiogenesis, as well as metastasis (Fig. 6B).

Fig. 6.

A Differences in the enrichment of ITGA6 were examined across 50 Hallmark gene sets and 83 metabolic pathway gene sets. Samples were stratified into high‑ and low‑expression groups based on the upper and lower 30% of ITGA6 expression levels, respectively. Differential expression analysis was performed using the limma package, followed by Gene Set Enrichment Analysis (GSEA) implemented in the clusterProfiler package. For each gene set, the enrichment score (ES) was computed, and statistical significance was assessed with multiple‑hypothesis correction. Results were visualized using bubble plots, where the normalized enrichment score (NES) indicates the direction of enrichment: a negative NES represents significant enrichment in the ITGA6 low‑expression group, while a positive NES corresponds to enrichment in the ITGA6 high‑expression group. Color intensity reflects the absolute value of ES, and bubble size denotes the significance level after adjustment, with larger bubbles indicating smaller corrected p‑values. B The relationship between ITGA6 expression and 14 malignant features of cancer is presented. The vertical axis indicates the normalized score for each functional state derived from combined z-score values, while the horizontal axis represents the z-score of ITGA6 expression. Pearson correlation coefficient (R) values are provided alongside each scatterplot

To further examine whether the associations observed at the pan-cancer level are consistent across individual tumor types, we performed cancer-specific analyses in eight representative malignancies (SKCM, UCEC, LUAD, LGG, HNSC, PAAD, KIRC, and UVM). The correlations between ITGA6 expression and 14 functional state gene sets were calculated using GSVA-derived scores. As shown in Supplementary Figure S1 A-H, the correlation patterns between ITGA6 expression and functional states varied across cancer types. While positive associations with certain functional processes, such as angiogenesis, EMT, invasion, and metastasis, were observed in multiple cancers, the strength and direction of these correlations were not consistent across all tumor types. In addition, variability was also observed in other functional categories, including proliferation, differentiation, and hypoxia. Overall, these findings indicate that the relationship between ITGA6 expression and tumor-related functional states is heterogeneous across cancers, suggesting context-dependent biological roles.

The relationship between ITGA6 and functional proteins as well as cellular pathways

To delve deeper into the role of ITGA6 in cancer pathogenesis, we systematically analyzed the interactions between ITGA6 and functional proteins in the TCPA database. Our findings revealed that in UVM, ITGA6 exhibited significant positive correlations with CKIT, ACC-pS79, PKCALPHA-pS657, ATM, and PKCALPHA, while demonstrating significant negative correlations with AKT, BECLIN, RB, CASPASE8, and CYCLIND1 (Fig. 7A). Figure 7B displays the four functional proteins with the most significant positive correlations with ITGA6 in UVM. Subsequently, we evaluated the activity scores of 10 cancer-related pathways summarized from previous studies, including TSC/mTOR, RTK, RAS-MAPK, PI3K-AKT, hormone ER, hormone AR, EMT, DNA damage response, cell cycle, and apoptosis. Based on the median expression level of ITGA6, patients were stratified into high-expression and low-expression groups, and differences in pathway activity scores between these groups were compared. Red indicates that the pathway activity was significantly higher in the ITGA6 high-expression group, while green signifies significantly lower activity. Across multiple cancer types, ITGA6 demonstrated a significant correlation with the activation status of RTK, and it was also closely associated with the inhibition of the DNA damage response (Fig. 7C). Notably, the results reflect differences in inferred pathway activity associated with ITGA6 expression, rather than direct protein–protein interactions.

Fig. 7.

A The top five functionally related proteins to ITGA6 across various cancers in the TCPA database. Significant positive correlations are represented in red, significant negative correlations in blue, and non-significant correlations in white. The darker the color, the greater the absolute value of the correlation. B The most significantly related functional proteins to ITGA6 in UVM from the TCPA database. C The relationships between ITGA6 and 10 cancer-related pathways across multiple cancer types. Patients were stratified into high-expression and low-expression groups based on ITGA6 expression levels, and differences in pathway activity scores between these groups were calculated using the wilcox.test function. Red indicates higher pathway activity in the ITGA6 high-expression group, green signifies lower activity, and white represents no significance. These results represent inferred pathway activity changes rather than direct protein–protein interactions

Correlation between ITGA6 and the immune microenvironment in cancer

We systematically analyzed the correlation between ITGA6 and 150 immune regulatory factors, including chemokines, chemokine receptors, MHC molecules, immunosuppressants, and immunostimulants, across multiple cancer types. The results indicated that in UVM, high expression of ITGA6 was significantly positively correlated with the expression of chemokines, chemokine receptors, immunosuppressants, immunostimulants, and MHC molecules. In adrenocortical carcinoma (ACC), high ITGA6 expression was similarly significantly positively correlated with the expression of immunostimulant molecules. Conversely, in KIRC, high ITGA6 expression was significantly negatively correlated with the expression of almost all immune regulatory factors. These findings suggest that ITGA6 may play a significant role in the immune microenvironment of UVM, ACC and KIRC (Fig. 8A). Subsequently, we employed multiple algorithms to evaluate the correlation between ITGA6 mRNA expression and immune cell infiltration. The results revealed that in UVM, ITGA6 expression was significantly positively correlated with the infiltration of CD8 + T cells and macrophages. In TCGA database, ITGA6 expression was also significantly positively correlated with the infiltration of B cells and CD8 + T cells. Notably, across all cancer types, ITGA6 expression was significantly positively correlated with the infiltration of endothelial cells (Fig. 8B). To provide a quantitative assessment of these relationships, the Pearson correlation coefficients between ITGA6 expression and immune modulators across cancer types are summarized in Supplementary Table S1.

Fig. 8.

Relationship between ITGA6 Expression and Immune Infiltration. A Correlation between ITGA6 and 150 immune regulatory factors across multiple cancer types; (B) Correlation between ITGA6 expression and cancer immune infiltration assessed using seven algorithms

The impact of ITGA6 on chemotherapy response and development of targeted therapeutics

To investigate the relationship between ITGA6 and cancer chemotherapy responsiveness, we conducted a drug sensitivity analysis using the PRISM, CTRP, GDSC1, and GDSC2 databases. The results demonstrated a significant correlation between ITGA6 expression and the sensitivity to numerous drugs (Fig. 9A).

Fig. 9.

Analysis of Chemotherapy Resistance. Four distinct databases were utilized to investigate the correlation between ITGA6 expression and drug sensitivity. A PRISM, GDSC2, GDSC1, and CTRP. B Potential compounds targeting ITGA6 across multiple cancers were identified using Connectivity Map (CMap) analysis. A heatmap displays the similarity score of each compound in every tumor type, with lower scores (indicated by darker blue shading) suggesting greater potential to inhibit ITGA6-mediated oncogenic effects. C In parallel, candidate small molecules and drugs capable of counteracting the biological effects of dysregulated ITGA6 expression are presented for four selected tumor types. Each point in the scatter plot represents a compound, plotted against its similarity score on the y‑axis; lower-scoring compounds are considered more likely to suppress ITGA6-driven pro‑carcinogenic activity

Compounds that significantly modulate ITGA6 activity could facilitate the discovery of novel therapeutic agents for cancer treatment. Utilizing data from the Cmap database, we employed the optimal feature matching method XSum to compare gene-related features with those in the cMap dataset, yielding similarity scores for 1,288 compounds. Compounds with lower scores are more likely to inhibit the expression of this gene. The results indicated that arachidonyltrifluoromethane, stock in 35,874, and X4.5dianilinophthalimide could inhibit ITGA6 activity in most cancers (Fig. 9B). Figure 9C illustrates specific potential targeted drugs in STAD, UVM, UCEC, and TGCT. X4.5dianilinophthalimide, MS.275, and W.13 significantly inhibited ITGA6 activity in STAD, UVM, UCEC, and TGCT, respectively.

The relationship between ITGA6 and UVM

In this study, we observed a association between ITGA6 expression and UVM, with higher ITGA6 expression being associated with poorer prognosis. To delve deeper into the potential link between ITGA6 expression and UVM, we conducted a more detailed analysis. Firstly, we thoroughly analyzed the correlation between ITGA6 expression and the prognosis of UVM. The results indicated that high expression of ITGA6 was a poor prognostic factor for OS, PFI, and DSS in UVM patients (Fig. 10A). To further evaluate whether ITGA6 serves as an independent prognostic factor in UVM, univariate and multivariate Cox regression analyses were performed incorporating available clinical variables. As shown in Table 1, univariate analysis indicated that high ITGA6 expression was significantly associated with worse overall survival (HR = 3.325, 95% CI: 1.310–8.443, P = 0.011). Importantly, multivariate analysis demonstrated that ITGA6 remained an independent prognostic factor after adjustment for clinical covariates (HR = 3.628, 95% CI: 1.364–9.655, P = 0.010). In addition, advanced pathologic stage and older age were also significantly associated with poor prognosis.

Fig. 10.

A The relationship between ITGA6 expression and OS, PFI, and DSS in UVM. B The relationship between ITGA6 expression and the immune microenvironment in UVM. C Correlations between ITGA6 expression and functional state gene sets in UVM. Different functional states showed varying correlation strengths, although all were positively associated with ITGA6 expression

Table 1.

Univariate and multivariate Cox regression analysis of overall survival in TCGA-UVM cohort

Characteristics	HR(95% CI) Univariate analysis	P value Univariate analysis	HR(95% CI) Multivariate analysis	P value Multivariate analysis
ITGA6
Low	Reference		Reference
High	3.325 (1.310—8.443)	0.011	3.628 (1.364—9.655)	0.010
Pathologic stage
Stage II	Reference		Reference
Stage III	1.180 (0.474—2.940)	0.722	0.969 (0.381—2.468)	0.948
Stage IV	69.945 (6.778—721.852)	< 0.001	50.870 (4.656—555.855)	0.001
Gender
Female	Reference
Male	1.542 (0.651—3.652)	0.325
Age
< = 60	Reference		Reference
> 60	2.123 (0.914—4.933)	0.080	2.678 (1.090—6.579)	0.032
BMI
< = 30	Reference
> 30	1.920 (0.705—5.234)	0.202

Additionally, we explored the interaction between ITGA6 expression and the immune microenvironment in UVM. The results showed that high expression of ITGA6 was significantly correlated with the immune infiltration of B cells, CD8 + T cells, and macrophages (Fig. 10B). Subsequently, we evaluated the correlation between ITGA6 expression and functional status gene sets in UVM. The results revealed that ITGA6 exhibited a significant positive correlation with all 14 functional status gene sets (R > 0.45, p < 0.001), particularly in terms of angiogenesis, proliferation, and stemness, with correlation coefficients as high as R > 0.7 (Fig. 10C). This variability may also reflect functional heterogeneity within the tumor. Overall, these findings indicate that ITGA6 expression is positively associated with malignant phenotypes in UVM and provide a rationale for subsequent experimental validation.

The effect of ITGA6 on UVM cells

Currently, there is a lack of experimental evidence regarding the specific effects of ITGA6 on UVM cells. Therefore, we successfully constructed an ITGA6 knockdown UVM cell model and verified the knockdown efficiency through RT-qPCR and Western blot experiments (Fig. 11A, B). CCK-8 assay assay results showed that the viability of UVM cells was significantly reduced after ITGA6 knockdown (Fig. 11C). Transwell experiments further confirmed that the migration and invasion abilities of UVM cells were significantly weakened after ITGA6 knockdown (Fig. 11D, E). Colony formation assay results also supported our findings, demonstrating that the colony-forming ability of UVM cells was significantly decreased after ITGA6 knockdown (Fig. 11F).

Fig. 11.

A, B Verification of ITGA6 knockdown efficiency by RT-qPCR and Western blot. C Metabolic assay of tumor cells after ITGA6 knockdown. D, E Transwell experiments detecting changes in migration and invasion abilities of tumor cells after ITGA6 knockdown. F Colony formation assay detecting the number of colonies formed by tumor cells after ITGA6 knockdown. **p < 0.01; ***p < 0.001

Discussion

ITGA6, a member of the integrin family, has been implicated in the onset, progression, and treatment of various tumors, as evidenced by numerous studies [26, 27]. This research focuses on exploring the expression patterns, mutational status, clinical feature correlations, diagnostic value, prognostic significance, functional enrichment, cellular pathway associations, immune microenvironment correlations, and responses to chemotherapy of ITGA6 across multiple cancers, with particular emphasis on its potential link with UVM.

Study data indicate that ITGA6 expression is significantly upregulated in various cancers. Notably, in cancers such as LUSC and LIHC, the upregulation of ITGA6 expression is particularly prominent, suggesting a pivotal role in cancer initiation and progression. These findings align with previous studies demonstrating ITGA6 overexpression in multiple cancer types [28, 20, 29].

Our in-depth analysis of the mutational distribution and types of ITGA6 across pan-cancer reveals that missense mutations are the primary single-nucleotide polymorphism type. Notably, ITGA6 exhibits the highest mutation rate in UCEC, potentially linked to the pathogenesis of UCEC. Meanwhile, the mutation rate of ITGA6 varies across cancer types, with the most common mutation types including mutation, deep deletion, and amplification, which may lead to functional alterations in ITGA6 and consequently affect cancer developmen [30–32]. It should be noted that mutation profiles and expression patterns represent distinct molecular layers, and their relationships with clinical features may not be directly comparable. Specifically, linking specific SNV types (e.g., missense mutations) to expression–gender or expression–age correlations is not biologically valid, as these operate on different analytical levels and no known mechanism supports such a linkage. Furthermore, all ITGA6 mutations analyzed in this study are tumor-associated somatic alterations derived from TCGA/cBioPortal datasets, not inherited germline mutations. Therefore, these alterations should not be interpreted as hereditary traits or as changes transmitted via germ cells.

Regarding the relationship between ITGA6 and clinical characteristics across pan-cancer, our results show that ITGA6 expression is significantly correlated with clinical features such as cancer stage and grade. As tumor malignancy increases, ITGA6 expression levels gradually decrease, potentially associated with ITGA6 dysfunction in early cancer stages. Previous studies have confirmed that high ITGA6 expression promotes proliferation, migration, and invasion in malignancies such as bladder cancer and metastatic breast cancer [33, 16, 34–37]. Furthermore, we found that ITGA6 expression levels correlate with patient age and gender, suggesting that ITGA6 expression may be influenced by multiple factors [38–40].

In terms of diagnostic value, ITGA6 demonstrates high diagnostic accuracy in multiple cancers such as pheochromocytoma and PCPG and CHOL, indicating its potential as an effective diagnostic biomarker for these cancers [41]. Regarding prognostic significance, ITGA6 expression levels are significantly associated with the prognosis of various cancers, serving as a risk or protective factor. The divergent prognostic patterns observed in Fig. 5B – where high ITGA6 expression is associated with better outcomes in SKCM and UCEC but with worse outcomes in LUAD, LGG, and UVM – likely reflect tumor heterogeneity and context-dependent biological roles of ITGA6. However, further studies are required to validate their clinical utility and to explore whether targeting ITGA6 could provide therapeutic benefits [42–44].

The onset of malignant tumors is closely related to the abnormal activation of multiple signal transduction pathways. Through GSEA, we found that ITGA6 is significantly positively correlated with oncogenic pathways such as EMT and TNFα signaling via NFκB, while negatively correlated with the oxidative phosphorylation pathway, further revealing the potential mechanisms of ITGA6 in cancer development. Furthermore, ITGA6 expression promotes angiogenesis and metastasis, potentially due to its ability to form heterodimers with ITGB4 (CD104) or ITGB1 (CD29), generating α6β4 integrin or α6β1 integrin, respectively [45, 46, 27]. These integrins are cellular adhesion molecules crucial for regulating signaling pathways involved in tumor development, metastasis, and angiogenesis [13, 15, 47].

We also found significant associations between ITGA6 and various functional proteins, particularly in UVM, where ITGA6 is highly positively correlated with CKIT and ACC-pS79, and highly negatively correlated with AKT and BECLIN. These results provide important clues for further understanding the function of ITGA6 in cancer. Additionally, through Pearson correlation analysis, we revealed the correlations between ITGA6 and multiple immune genes across pan-cancer. The results show significant correlations between ITGA6 and various immune regulatory factors. Notably, in UVM, multiple immunosuppressants, immunostimulants, and MHC molecules are significantly upregulated in the high ITGA6 expression cohort, suggesting that ITGA6 plays a crucial role in the tumor immune microenvironment [48–51]. However, these findings are based on correlation analyses and do not imply causality; further experimental validation is required to elucidate the underlying mechanisms.

The interaction between ITGA6 and the tumor immune microenvironment may be related to drug resistance in cancer chemotherapy. Previous studies have shown that ITGA6 is closely associated with drug resistance in laryngeal squamous cell carcinoma and epithelial ovarian cancer [17, 18]. Therefore, understanding the drug sensitivity of ITGA6 can enhance the efficacy of cancer chemotherapy. Our study of ITGA6 drug sensitivity found significant correlations between ITGA6 expression and multiple drugs. In particular, compounds such as arachidonyltrifluoromethane, stock in 35,874, and X4.5.dianilinophthalimide can inhibit ITGA6 activity in multiple cancers, providing new avenues for identifying novel cancer therapeutic drugs.

Previous studies have reported that ITGA6 is involved in tumor progression and metastasis in several cancers [52]. Our findings are consistent with these reports and further extend the role of ITGA6 to a pan-cancer context, while highlighting its potential significance in UVM. Through this research, we found a significant correlation between ITGA6 and UVM, an area that has received little attention in previous studies. Therefore, we further analyzed the association between ITGA6 and UVM. In UVM, high ITGA6 expression is a risk factor for OS, PFI, and DSS, and it exhibits significant correlations with the immune microenvironment and functional status gene sets. These results suggest that ITGA6 may play a key role in the onset and progression of UVM. Our experimental data confirm that knocking down ITGA6 inhibits proliferation, migration, and invasion. This fills a gap in integrin family research on UVM, providing new perspectives and clues for studying the integrin family in tumor biology and offering new insights into the metastatic mechanisms of this rare tumor [53–55]. The functions of ITGA6 in UVM cells suggest that targeting ITGA6 may represent a potential therapeutic strategy. However, given the complexity of integrin-related signaling networks, future studies should explore whether feedback regulatory mechanisms or compensatory pathway activation may influence therapeutic efficacy, and whether combination or adaptive strategies could enhance treatment outcomes. By inhibiting ITGA6 expression or function, the proliferation, migration, and invasion capabilities of UVM cells can be significantly reduced, thereby inhibiting tumor growth and metastasis [20, 37].

Several limitations should be acknowledged. First, drug sensitivity predictions were derived from computational tools (CMap, GDSC) and lack experimental validation. Second, the lack of independent external cohorts for certain analyses (e.g., UVM) limits generalizability. Future studies should include prospective clinical validation and in vivo models to confirm the therapeutic potential of targeting ITGA6. We acknowledge that the CCK-8 assay used in this study reflects metabolic activity rather than directly quantifying live/dead cell proportions. Future studies should incorporate direct cell counting, apoptosis detection (e.g., Annexin V/PI staining), or other viability discrimination assays to further elucidate the mechanisms by which ITGA6 knockdown reduces UVM cell viability.

In summary, this study comprehensively analyzes data from multiple databases to delve into the expression characteristics, mutational status of ITGA6, and its correlations with clinical features, prognosis, immune microenvironment, and chemotherapy responses across multiple cancers. These findings offer new insights and bases for further understanding the functions of ITGA6 in cancer and its applications in clinical diagnosis and treatment. Future research can further explore the specific mechanisms of ITGA6 in cancer and the development and application of targeted therapeutic drugs against ITGA6.

Abbreviations

ACC

Adrenocortical carcinoma

AKT

Protein kinase B

AR

Androgen receptor

AUC

Area under the curve

BH

Benjamini-Hochberg

BLCA

Bladder urothelial carcinoma

CCK-8

Cell Counting Kit-8

CESC

Cervical squamous cell carcinoma

CHOL

Cholangiocarcinoma

CI

Confidence interval

CMap

Connectivity Map

CNA

Copy number alteration

CPTAC

Clinical Proteomic Tumor Analysis Consortium

DFI

Disease-free interval

DLBC

Diffuse large B-cell lymphoma

DSS

Disease-specific survival

ECL

Enhanced chemiluminescence

EMT

Epithelial-mesenchymal transition

ER

Estrogen receptor

FBS

Fetal bovine serum

GEO

Gene Expression Omnibus

GSEA

Gene Set Enrichment Analysis

GSVA

Gene Set Variation Analysis

GTEx

Genotype-Tissue Expression

HNSC

Head and neck squamous cell carcinoma

HPA

Human Protein Atlas

HR

Hazard ratio

IHC

Immunohistochemistry

ITGA6

Integrin alpha 6

KEGG

Kyoto Encyclopedia of Genes and Genomes

KICH

Kidney chromophobe

KIPAN

Pan-kidney cohort

KIRC

Kidney renal clear cell carcinoma

LGG

Lower-grade glioma

LIHC

Liver hepatocellular carcinoma

LUAD

Lung adenocarcinoma

LUSC

Lung squamous cell carcinoma

MAPK

Mitogen-activated protein kinase

MHC

Major histocompatibility complex

MSigDB

Molecular Signatures Database

NF-κB

Nuclear factor kappa B

OS

Overall survival

PAAD

Pancreatic adenocarcinoma

PCPG

Pheochromocytoma/paraganglioma

PCR

Polymerase chain reaction

PFI

Progression-free interval

PI3K

Phosphoinositide 3-kinase

PVDF

Polyvinylidene difluoride

RAS

Rat sarcoma virus

RTK

Receptor tyrosine kinase

RT-qPCR

Reverse transcription quantitative PCR

SD

Standard deviation

siRNA

Small interfering RNA

SKCM

Skin cutaneous melanoma

SNV

Single-nucleotide variant

STAD

Stomach adenocarcinoma

STES

Esophageal carcinoma

STR

Short tandem repeat

TCGA

The Cancer Genome Atlas

TCPA

The Cancer Proteome Atlas

TGCT

Testicular germ cell tumor

THYM

Thymoma

TNF

Tumor necrosis factor

UCEC

Uterine corpus endometrial carcinoma

UCS

Uterine carcinosarcoma

UVM

Uveal melanoma

XSum

eXtreme Sum

Authors’ contributions

YL, CS and MD analyzed the data, and wrote the manuscript; JL and FZ conceived the project and revised the manuscript; while YG, TL and WZ edited the manuscript; MD, CS and YL researched the data. All authors read and approved the final manuscript.

Funding

This work was supported by National Natural Science Foundation Project (No.82460215); Yinchuan Science and Technology Plan Project (No. 2024SF006); Ningxia Natural Science Foundation Project (No.2024AAC03515); Ningxia Medical University School-level Scientific Research Project (No. XY2024058).

Data availability

The study incorporated data from publicly accessible websites, as detailed in the “Methods” section. Additional information is available upon contacting the corresponding authors.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.