Functional and mechanistic inactivation of immune hubs orchestrates tumor development and progression in lung adenocarcinoma Authors: Abdullah Alghamdi, Mohammed Alissa Published: 2026 ABSTRACT AbstractBackgroundLung adenocarcinoma (LUAD) is the most common subtype of non-small cell lung cancer (NSCLC) and remains a leading cause of cancer-related deaths worldwide. Despite advances in targeted therapies and immune checkpoint inhibitors, survival outcomes remain poor, underscoring the need for novel immune-related biomarkers to improve diagnosis, prognosis, and therapeutic strategies.MethodologyWe retrieved 1,284 immune-related genes from the ImmPort database and constructed a protein–protein interaction (PPI) network using STRING, with hub genes identified via Cytoscape. Expression validation was performed using RT-qPCR with TaqMan assays in LUAD and normal lung cell lines, and extended validation was carried out through GSCA (TCGA datasets), HPA, OncoDB, cBioPortal, KM Plotter, TISIDB, and TargetScan. Functional assays including proliferation, colony formation, wound-healing, and xenografts were performed in LUAD cell lines (A549, H1299) following overexpression of hub genes.ResultsFour hub genes, IFNG, TNF, CD4, and CD8A, were identified as central immune regulators. All were significantly downregulated in LUAD tissues and cell lines, with ROC analysis confirming diagnostic potential. TCGA and HPA datasets validated reduced expression at both mRNA and protein levels, while promoter methylation and CNV alterations were linked to transcriptional repression. Kaplan–Meier analysis revealed that low expression predicted poor overall survival. TargetScan identified four candidate miRNAs (miR-130b-3p, miR-326, miR-4319, and miR-4644), validated as upregulated in LUAD, with ROC curves showing strong diagnostic accuracy. Immune subtype and infiltration analysis confirmed strong associations of hub genes with T-cell–mediated immunity and immune checkpoint molecules, while drug sensitivity analysis linked their expression to responses to chemotherapy and targeted agents. Functional assays demonstrated that hub gene overexpression suppressed proliferation, migration, and invasion in vitro and inhibited tumor growth in xenograft models.ConclusionThis integrative study establishes IFNG, TNF, CD4, and CD8A as novel immune-related biomarkers in LUAD, regulated through genetic, epigenetic, and post-transcriptional mechanisms. Their diagnostic, prognostic, and therapeutic significance highlights their potential as targets for precision oncology in LUAD.Supplementary InformationThe online version contains supplementary material available at 10.1007/s12672-026-05157-z.Keywords: LUAD, Immune-related hub genes, Bioinformatics, miRNA regulation, Functional validation Introduction Lung cancer remains one of the most common malignancies worldwide and is the leading cause of cancer-related deaths, accounting for nearly 1.8 million deaths annually [1, 2]. Among the histological subtypes, lung adenocarcinoma (LUAD) represents the most frequent and aggressive form [3, 4]. Despite advances in surgery, radiotherapy, chemotherapy, and targeted agents, the overall five-year survival rate for LUAD patients remains dismal, particularly in advanced stages [5]. The heterogeneous molecular landscape of LUAD complicates clinical management and highlights the urgent need for novel diagnostic, prognostic, and therapeutic biomarkers [6]. A better understanding of the molecular mechanisms driving LUAD progression is essential for improving patient outcomes. In recent years, the tumor immune microenvironment (TIME) has emerged as a central player in the development and progression of LUAD [7, 8]. Tumor cells evade immune surveillance through complex interactions with immune and stromal components, which not only facilitate tumor proliferation but also promote metastasis and therapeutic resistance [9]. With the clinical success of immune checkpoint inhibitors such as PD-1/PD-L1 and CTLA-4 blockers, immunotherapy has revolutionized LUAD treatment [5, 10]. However, not all patients benefit from these therapies, and primary or acquired resistance remains a major clinical obstacle. Consequently, identifying immune-related molecular signatures and hub genes that regulate tumor-immune interactions has become a focus of cancer research. Immune-related hub genes represent highly connected nodes in gene interaction networks, often serving as key regulators of tumor biology [11, 12]. Such genes may influence immune cell infiltration, cytokine signaling, and antigen presentation, thereby shaping the tumor immune landscape [11, 12]. Previous studies have highlighted the significance of immune-associated genes such as CD274 (PD-L1), CTLA4, and LAG3 in LUAD progression and therapy response [13, 14]. However, a systematic exploration of novel immune-related hub genes with potential diagnostic and therapeutic implications in LUAD remains limited. Our study aimed to fill this gap by identifying and characterizing immune-related hub genes in LUAD through an integrative approach. To ensure translational relevance, we validated the role of the hub genes in LUAD cell lines and Xenograft model using functional assays, thereby complementing the bioinformatics predictions with in vitro and in vivo evidence. Through this combined in silico, in vitro, and in vivo strategy, our study identified immune-related hub genes that are not only differentially expressed in LUAD but also correlated with patient prognosis, immune infiltration, and functional pathways. These findings hold promise for improving LUAD diagnostics, prognostication, and therapeutic targeting. Importantly, the hub genes identified in our analysis may serve as potential biomarkers for predicting immunotherapy response, guiding patient stratification, and informing personalized treatment strategies. In summary, LUAD continues to impose a significant health burden, and existing therapeutic options remain inadequate for many patients. Immunotherapy has provided new hope, but its clinical benefits are not universal. Our study contributes to this field by systematically identifying immune-related hub genes with potential clinical and biological importance in LUAD. By integrating bioinformatics analyses with laboratory validation, we offer a robust framework for discovering novel biomarkers and therapeutic targets. The identification of these genes not only deepens our understanding of LUAD pathogenesis but also lays the groundwork for future translational and clinical studies aimed at improving patient care. Methodology Immune-related genes retrieval ImmPort (https://www.immport.org/) is supported by the National Institutes of Health (NIH) to promote data sharing with the public. To download immune-related genes, we navigated to the ImmPort database homepage, clicked on the “Resources” button, selected “Gene Lists” from the “Resources” page, and then clicked on “Gene Summary.” PPI network, hub genes identification, and pathway enrichment The STRING database is a tool for identifying interacting genes and constructing PPI networks based on known and predicted interactions [15]. Using STRING, we constructed the PPI network for immune-related genes with a confidence score of ≥ 0.4. The PPI network was then visualized using Cytoscape software (version 3.7.2). Additionally, we utilized the CytoHubba plug-in in Cytoscape to calculate gene degrees using the “degree” method, identifying the top 10 hub genes for further analysis and verification. Moreover, DAVID [16] database was used to explore the hub genes-associated signaling pathways. Cell culture Six human LUAD cell lines (A549, H1299, H1975, H1650, H460, and H358) and six normal lung epithelial or fibroblast cell lines (BEAS-2B, NHBEC, WI-38, MRC-5, HEL299, and HFL1) were obtained from ATCC, Manassas, VA, USA. All cell lines were authenticated by short tandem repeat (STR) profiling and tested negative for mycoplasma contamination prior to experimentation. Cells were cultured in RPMI-1640 or DMEM (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific) and 1% penicillin-streptomycin (Thermo Fisher Scientific). Cultures were maintained at 37 °C in a humidified incubator with 5% CO₂. RNA extraction and quantitative real-time PCR (RT-qPCR) Total RNA was extracted from LUAD and normal lung cell lines using the TRIzol Reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol, and RNA concentration and purity were determined with a NanoDrop spectrophotometer. One microgram of RNA was reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific), and quantitative real-time PCR was carried out with TaqMan Gene Expression Assays (Applied Biosystems, Thermo Fisher Scientific) on a real-time PCR system. The following TaqMan probe IDs were used: IFNG (Hs00989291_m1), TNF (Hs00174128_m1), CD4 (Hs01058407_m1), CD8A (Hs00233520_m1), and GAPDH (Hs02786624_g1) as an internal control. Relative gene expression levels were calculated using the 2^−ΔΔCt method, with GAPDH used as the endogenous reference gene. GAPDH is widely used as a housekeeping control in lung cancer gene expression studies due to its stable expression across multiple cell types. However, recent studies suggest that the geometric mean of multiple reference genes may further improve normalization accuracy in cancer transcriptomic analyses [17, 18]. In the present study, GAPDH-based normalization was complemented by independent validation using TCGA transcriptomic datasets and HPA protein expression data, which demonstrated consistent expression patterns for the hub genes. Extended validation of hub gene expression The mRNA expression levels of hub genes in LUAD were validated using the Gene Set Cancer Analysis (GSCA) database, which integrates TCGA transcriptomic datasets [19]. Differential expression analysis was performed to compare LUAD tissues with normal controls, and expression patterns across pathological stages were also examined. Protein expression of hub genes was assessed using immunohistochemistry data from the Human Protein Atlas (HPA) database [20]. Because HPA provides pre-annotated images rather than raw staining intensity data, the analysis was performed qualitatively without formal semi-quantitative scoring [21]. To explore functional relevance, Gene Set Enrichment Analysis (GSEA) was carried out through GSCA, using default parameters. Promoter methylation analysis Promoter methylation levels of the hub genes IFNG, TNF, CD4, and CD8A were analyzed using the OncoDB platform [22], which integrates DNA methylation profiles from TCGA datasets. Differential methylation between LUAD and normal tissues was assessed under default parameters. OncoDB provides gene-level methylation data from TCGA, which summarizes the methylation levels at key genomic positions, such as the promoter region and gene body. These methylation levels are derived from multiple CpG probes within the promoter regions, but detailed probe-level analysis (e.g., individual CpG site methylation) and CpG context (exonic vs. intronic) [23] were not assessed. To examine the relationship between promoter methylation and gene expression, correlation analysis was performed through the GSCA database [19] using Spearman’s rank correlation, and significance was determined with false discovery rate (FDR)–adjusted p-values. Genomic alteration analysis Genomic alterations of the hub genes IFNG, TNF, CD4, and CD8A in LUAD were analyzed using the cBioPortal for Cancer Genomics platform, which integrates TCGA data [24]. The types and frequencies of genetic alterations were identified under default parameters. Protein domain mapping of variants was performed within cBioPortal to visualize mutation locations relative to functional domains. Copy number variations (CNVs) were also assessed. To evaluate the clinical significance of these alterations, overall survival (OS), progression-free survival (PFS), and disease-specific survival (DSS) were analyzed using cBioPortal’s survival module with Kaplan–Meier estimates. Prognostic analysis, miRNA prediction, and validation The prognostic significance of IFNG, TNF, CD4, and CD8A was evaluated using the KM Plotter database [25], which integrates TCGA survival data to generate Kaplan–Meier survival curves and hazard ratios under default parameters. Potential upstream regulatory miRNAs targeting the 3′UTRs of hub genes were predicted using the TargetScan database [26], based on conserved seed sequence binding. Candidate miRNAs were experimentally validated in six LUAD and six normal lung cell lines using TaqMan MicroRNA Assays (Applied Biosystems, Thermo Fisher Scientific). The following probe IDs were used: hsa-miR-130b-3p (Assay ID: 000456), hsa-miR-326 (Assay ID: 000542), hsa-miR-4319 (Assay ID: 465345_mat), hsa-miR-4644 (Assay ID: 464312_mat), and U6 snRNA (Assay ID: 001973) as the endogenous control. Relative expression levels were calculated with the 2^−ΔΔCt method. Dual-luciferase reporter assay Dual-luciferase assays were performed using the Thermo Fisher Scientific dual-luciferase reporter system according to the manufacturer’s instructions. Briefly, the predicted miRNA-binding fragments of the 3′UTR of IFNG, TNF, CD4, and CD8A were amplified and cloned downstream of the firefly luciferase gene in a luciferase reporter vector (WT). Corresponding mutant (MUT) reporters were generated by site-directed mutagenesis of the seed-matching sites. Human LUAD cell lines (A549, H1299, H1975, H1650, H460, and H358) were seeded in 24-well plates and co-transfected with WT or MUT reporter plasmids together with the indicated miRNA mimics (miR-4319, miR-130b-3p, miR-4644, or miR-326) or negative control mimic using Lipofectamine™ 3000 (Thermo Fisher Scientific). A Renilla luciferase plasmid was co-transfected as an internal control to normalize transfection efficiency. After 24–48 h, cells were lysed using Passive Lysis Buffer, and firefly and Renilla luciferase activities were measured sequentially on a luminometer following addition of the kit substrates. Results are presented as relative luciferase activity (firefly/Renilla) and expressed as fold change compared with the corresponding control group. Immune subtype, immune modulator, infiltration, and drug sensitivity analysis The association between hub gene expression and LUAD immune subtypes was analyzed using the TISIDB database [27], which integrates molecular and immunological data across TCGA cohorts. Correlations between hub gene expression and immune checkpoint molecules were also assessed through TISIDB. Immune infiltration patterns were evaluated using the GSCA database [19], which estimates the relative abundance of immune cell subsets based on gene expression profiles. The relationship between hub gene expression and drug sensitivity was further analyzed using GSCA by correlating expression levels with the half-maximal inhibitory concentration (IC50) of a panel of anticancer drugs, including chemotherapeutics and targeted agents. Inducing overexpression of hub genes using expression vectors in LUAD cells LUAD cell lines A549 and H1299 (ATCC, Manassas, VA, USA) were transfected with overexpression (OE) plasmids encoding IFNG, TNF, CD4, or CD8A, together with corresponding empty vector controls, using Lipofectamine™ 3000 (Thermo Fisher Scientific) according to the manufacturer’s protocol. To control for potential transfection reagent effects, parallel empty vector–transfected cells were used as controls in all experiments. Forty-eight hours post-transfection, expression efficiency was confirmed at both the mRNA level by RT-qPCR (TaqMan Gene Expression Assays, Applied Biosystems) and at the protein level by western blotting. Functional assays were performed using identical transfection conditions across groups to ensure comparable exposure to the transfection reagent. A previously reported Western blot protocol [28, 29] was employed in our study. Cell proliferation assay Cell viability was assessed using the CellTiter-Blue Cell Viability Assay (Thermo Fisher Scientific). Transfected and control cells were seeded in 96-well plates and incubated for 24–72 h. Fluorescence intensity was measured on a microplate reader, and proliferation rates were calculated relative to controls. Colony formation assay For clonogenic assays, transfected cells were seeded at low density (approximately 500–1000 cells per well) in 6-well plates and cultured for 10–14 days under standard conditions to allow colony formation. Colonies were fixed with methanol and stained with 0.1% crystal violet. Only colonies containing ≥ 50 cells were counted manually. Colony numbers were quantified relative to the corresponding empty vector control group, ensuring identical plating conditions across all experimental groups. This assay reflects the long-term proliferative and survival capacity of individual cells following transfection. Wound-healing assay Cell migration was evaluated using a wound-healing assay. Briefly, cells were seeded in 6-well plates until reaching near confluence, after which a linear scratch was created using a sterile pipette tip. Detached cells were removed with PBS, and images of wound closure were captured at 0 and 24 h. Migration rates were quantified as the percentage of wound closure relative to the initial wound width. This assay measures overall collective cell migration over the observation period but does not capture single-cell tracking dynamics or directional persistence parameters. Xenograft model For in vivo validation, female BALB/c nude mice (6–8 weeks old, 18–22 g) were obtained from the institutional animal facility and maintained under standard pathogen-free conditions (22 ± 2 °C, 12 h light/dark cycle, 50–60% humidity) with free access to food and water. A549 cells transfected with either control or hub gene–overexpressing vectors were suspended in sterile PBS (1 × 10⁶ cells/100 µL) and injected subcutaneously into the right flank under brief isoflurane anesthesia (2% in oxygen). Tumor growth was monitored twice weekly using digital calipers, and volume was calculated as (length × width²)/2. After four weeks, or earlier if humane endpoints were reached, mice were deeply anesthetized with ketamine (80 mg/kg) and xylazine (10 mg/kg) via intraperitoneal injection, followed by cervical dislocation to ensure a painless and rapid death. Tumors were excised, weighed, and processed for histological and molecular analyses. All efforts were made to minimize animal discomfort and to adhere to ethical principles of reduction, refinement, and replacement. Statistical analysis All experiments were performed in triplicate unless otherwise stated. Data are presented as mean ± standard deviation (SD). Statistical comparisons between two groups were performed using an unpaired two-tailed Student’s t-test, while comparisons among multiple groups were analyzed by one-way ANOVA followed by Tukey’s post hoc test. For bioinformatics analyses, database-integrated statistical tools were used under default parameters, and false discovery rate (FDR)–adjusted p-values were applied where applicable. Receiver operating characteristic (ROC) curve analysis was used to evaluate diagnostic performance, and Kaplan–Meier survival curves were compared using the log-rank test. Correlation analyses were performed using Spearman’s rank correlation. P*-value < 0.05, P**-value < 0.01, and P***-value < 0.001 were considered statistically significant. Results Retrieval of immune-related genes and hub genes identification A total of 1,284 immune-related genes were initially retrieved from the ImmPort database and used to construct a protein–protein interaction (PPI) network through the STRING database (Fig. 1A). The resulting network was imported into Cytoscape software, where hub genes were identified using the degree method. This analysis highlighted four genes (IFNG, TNF, CD8A, and CD4) as the top hub genes with the highest connectivity scores (Fig. 1B). Subsequent KEGG pathway enrichment analysis of the hub genes revealed their involvement in several immune-related pathways, including the “T cell receptor signaling pathway, PD-L1 expression and PD-1 checkpoint pathway in cancer, epithelial cell signaling in Helicobacter pylori infection, antigen processing and presentation, and rheumatoid arthritis.” Additional pathways identified were the “IL-17 signaling pathway, hematopoietic cell lineage, Chagas disease, amoebiasis, and Yersinia infection (Fig. 1C).” Other significantly enriched pathways included the “phospholipase D signaling pathway, JAK-STAT signaling pathway, human cytomegalovirus infection, cytokine–cytokine receptor interaction, and shigellosis (Fig. 1C).” To validate these findings, we assessed the expression levels of the hub genes in six LUAD cell lines compared with six normal lung cell lines using RT-qPCR. The results demonstrated significant downregulation of IFNG, TNF, CD4, and CD8A in LUAD cells relative to normal controls (p-value < 0.05; Fig. 1D). Furthermore, ROC curve analysis based on the RT-qPCR data showed that all four hub genes achieved high AUC values, indicating strong diagnostic potential (Fig. 1E). Retrieval of immune-related genes and hub gene identification in LUAD. A Construction of a PPI network for 1,284 immune-related genes retrieved from the ImmPort database using STRING. B Identification of top hub genes (IFNG, TNF, CD4, CD8A) using the degree method in Cytoscape. C Enriched KEGG pathways. D RT-qPCR validation of hub gene expression in six LUAD and six normal lung cell lines. E ROC curve analysis of RT-qPCR data showing diagnostic accuracy of hub genes. P-value < 0.05 Validation of hub gene expression in LUAD using TCGA datasets The mRNA expression of the hub genes was validated using TCGA-LUAD datasets through the GSCA platform. As shown in Fig. 2A, all four genes were significantly (p-value < 0.05) downregulated in LUAD tissues compared with normal controls. To explore the relationship between hub gene expression and disease progression, we analyzed their expression across LUAD pathological stages. Heatmap and trend analyses (Fig. 2B-C-D) revealed that CD4, CD8A, and IFNG expression declined progressively with advancing tumor stage, whereas TNF expression showed minimal change. We next examined protein expression levels using immunohistochemistry data from the HPA. Consistent with the mRNA data, CD4, CD8A, TNF, and IFNG were readily detected in normal lung tissues, with moderate staining intensity, but appeared reduced in LUAD tissues compared with normal lung tissues based on qualitative inspection of HPA immunohistochemistry images (Fig. 2E). To investigate the biological functions of the identified hub genes, GSEA was conducted using the GSCA database. The enrichment plots (Fig. 2F) demonstrated significant enrichment of gene sets in LUAD development. Validation of hub gene expression in LUAD using TCGA and HPA datasets. A mRNA expression of hub genes in LUAD and normal tissues from TCGA using GSCA. B, C Heatmap and trend plots showing hub gene expression across LUAD pathological stages. D Differential expression analysis of hub genes by stage. E Protein expression validation of hub genes in normal lung and LUAD tissues using the HPA database. F GSEA plots showing functional enrichment of hub gene-associated pathways in LUAD. P-value < 0.05 Promoter methylation analysis of hub genes in LUAD We next examined whether epigenetic alterations contribute to the dysregulation of the hub genes in LUAD. Using the OncoDB platform, promoter methylation profiles of IFNG, TNF, CD4, and CD8A were analyzed (Fig. 3A). IFNG and TNF genes exhibited significantly (p-value < 0.05) higher promoter methylation levels, while CD4 and CD8A exhibited significantly (p-value < 0.05) lower promoter methylation levels in LUAD tissues compared with normal controls (Fig. 3A). To assess the impact of methylation on transcriptional regulation, correlation analyses were performed using the GSCA database (Fig. 3B). A consistent negative association was identified between promoter methylation and gene expression for each hub gene. The Spearman correlation coefficients were − 0.27 for IFNG, − 0.34 for TNF, − 0.68 for CD4, and − 0.66 for CD8A, all statistically significant (Fig. 3B; FDR < 0.001). Promoter methylation analysis of hub genes in LUAD. A Promoter methylation profiles of IFNG, TNF, CD4, and CD8A in LUAD and normal tissues using OncoDB. B Correlation between promoter methylation and mRNA expression of hub genes in LUAD from GSCA. P-value < 0.05 Genomic alterations of hub genes in LUAD Genomic alterations of the hub genes were explored using the TCGA-LUAD cohort in cBioPortal. Among the analyzed cases, a subset of approximately 20 samples exhibited detectable alterations in at least one hub gene. CD4 showed the highest mutation frequency (~ 10%), followed by IFNG (~ 6%), CD8A (~ 4%), and TNF (~ 2%) (Fig. 4A-B). Variant classification analysis revealed that missense mutations were the most common type, followed by splice-site and nonsense mutations (Fig. 4C). The single nucleotide variant (SNV) spectrum indicated that C > T substitutions accounted for the majority of observed alterations. The median number of variants per sample was one. Protein domain mapping (Fig. 4D) showed that mutations in these hub genes were distributed across different functional domains, potentially affecting their protein activity. CNV analysis demonstrated that hub genes were subject to both heterozygous and homozygous amplifications, as well as deletions (Fig. 4E). To determine the prognostic implications of CNVs, OS, PFS, and DSS were analyzed in LUAD patients (Fig. 4F). Kaplan–Meier survival analysis suggested a trend toward poorer clinical outcomes for patients harboring CNV alterations (amplifications or deletions) compared with wild-type groups; however, these differences did not reach statistical significance (Fig. 4F). Genomic alterations of hub genes in LUAD, A Oncoplot showing mutation profiles of hub genes across LUAD samples. B Variant classification of mutations in hub genes. C Heatmap showing SNV mutation frequencies. D Protein domain mapping of hub gene mutations. E Copy number variation (CNV) analysis of hub genes. F Kaplan–Meier curves showing OS, PFS, and DSS stratified by CNV status of hub genes. P-value < 0.05 Prognostic significance, miRNA regulation, experimental validation, and diagnostic value of hub genes in LUAD Kaplan–Meier survival analysis using the KM Plotter database revealed that lower expression levels of IFNG, TNF, CD4, and CD8A were significantly (p-value < 0.05) associated with poor overall survival in LUAD patients (Fig. 5A). These results indicate that reduced expression of the identified hub genes is linked to poor patient prognosis. To explore potential upstream regulators of the hub genes, we performed miRNA–mRNA interaction analysis using the TargetScan database. Different miRNAs were predicted to target the 3′ UTRs of IFNG, TNF, CD4, and CD8A (Fig. 5B). Notably, hsa-miR-130b-3p, hsa-miR-326, hsa-miR-4319, and hsa-miR-4644 were identified as potential regulators with strong binding predictions (Fig. 5B). To validate the predicted miRNA–mRNA interactions, we assessed the expression of the candidate miRNAs in six LUAD cell lines and six normal control cell lines using RT-qPCR. As shown in Fig. 5C, all four miRNAs (miR-130b-3p, miR-326, miR-4319, and miR-4644) were significantly upregulated in LUAD cells compared with normal controls (Fig. 5C). This inverse relationship supports their potential role in downregulating IFNG, TNF, CD4, and CD8A in LUAD. The diagnostic accuracy of these miRNAs was further evaluated using ROC curve analysis (Fig. 5D). The AUC values were 0.80 for miR-130b-3p, 0.80 for miR-326, 0.76 for miR-4319, and 0.70 for miR-4644 (Fig. 5D), indicating moderate to high discriminatory power in distinguishing LUAD from normal samples. To experimentally validate the predicted miRNA–mRNA interactions, dual-luciferase reporter assays were performed in LUAD cell lines. Reporter constructs containing the wild-type (WT) 3′UTRs of IFNG, TNF, CD4, and CD8A were co-transfected with their corresponding miRNA mimics (miR-4319, miR-130b-3p, miR-4644, and miR-326, respectively). As shown in Fig. 5E, co-transfection of the miRNA mimics significantly reduced the luciferase activity of the WT reporter constructs compared with controls, whereas mutation of the predicted binding sites (MUT) largely abolished this inhibitory effect across multiple LUAD cell lines (A549, H1299, H1650, H1975, H358, and H460). These results confirm that miR-4319, miR-130b-3p, miR-4644, and miR-326 directly target the 3′UTRs of IFNG, TNF, CD4, and CD8A, respectively, providing functional evidence that these miRNAs may contribute to the suppression of these hub genes in LUAD. Prognostic relevance of hub genes, miRNA regulatory prediction, expression validation, diagnostic performance, and luciferase confirmation of miRNA–mRNA interactions in LUAD. A Kaplan–Meier survival curves for hub gene expression in LUAD from KM Plotter. B Predicted miRNA–mRNA binding interactions for hub genes using TargetScan. C RT-qPCR validation of candidate miRNA expression in six LUAD and six normal lung cell lines. D ROC curve analysis of candidate miRNAs showing diagnostic performance. E Dual-luciferase reporter assays validating the direct interactions between the candidate miRNAs and their predicted target genes. Co-transfection of miRNA mimics (miR-4319, miR-130b-3p, miR-4644, and miR-326) significantly reduced luciferase activity of the wild-type (WT) 3′UTR reporter constructs of IFNG, TNF, CD4, and CD8A, respectively, while mutation of the predicted binding sites (MUT) attenuated this effect across LUAD cell lines (A549, H1299, H1650, H1975, H358, and H460). Data are presented as mean ± SD from three independent experiments. P-value < 0.05 Association of hub genes with immune subtypes, immune modulators, immune infiltration, and drug sensitivity in LUAD Using the TISIDB database, we assessed the expression of hubs IFNG, TNF, CD4, and CD8A across six immune subtypes of LUAD (Fig. 6A). All hub genes displayed significant expression variability among subtypes (Kruskal–Wallis test, p < 0.001). Notably, expression levels were higher in immune-active subtypes compared with immune-suppressed ones, suggesting that these genes are closely linked to the immune phenotype of LUAD (Fig. 6A). We next explored the relationship between hub gene expression and immune inhibitor molecules using TISIDB (Fig. 6B). A broad range of inhibitory immune checkpoint molecules, including PDCD1 (PD-1), CD274 (PD-L1), CTLA4, LAG3, and TIGIT, showed significant positive correlations with IFNG, TNF, CD4, and CD8A expression (Fig. 6B). Immune infiltration analysis via the GSCA database revealed strong positive correlations between hub gene expression and multiple immune cell populations. CD4 and CD8A expression correlated most significantly with CD8 + T cells, NK cells, iTreg cells, and macrophages, while IFNG and TNF also showed strong associations with cytotoxic lymphocytes and Th1 cell subsets (Fig. 6C). Finally, drug response analysis was performed using GSCA to determine the relationship between hub gene expression and sensitivity to anticancer agents. Several clinically relevant compounds, including 5-fluorouracil, gefitinib, and paclitaxel, showed significant correlations between predicted IC50 values and hub gene expression in GSCA pharmacogenomic datasets (Fig. 6D). Decreased expression of IFNG, CD4, and CD8A was associated with increased sensitivity to immune-modulating or cytotoxic drugs, suggesting that these genes may influence therapeutic responses in LUAD. Association of hub genes with immune subtypes, immune checkpoints, infiltration, and drug sensitivity. A Expression of hub genes across six LUAD immune subtypes using TISIDB. B Correlation between hub gene expression and immune inhibitory molecules in LUAD. C Association between hub gene expression and immune cell infiltration from GSCA. D Correlation between hub gene expression and drug sensitivity to anticancer agents. P-value < 0.05 Functional characterization of hub genes in LUAD cell lines and Xenograft model To experimentally validate the biological roles of the hub genes, overexpression (OE) vectors for IFNG, TNF, CD4, and CD8A were transfected into LUAD cell lines (A549 and H1299). RT-qPCR and western blot confirmed successful upregulation of each target gene at both mRNA and protein levels compared with control cells (Fig. 7A–B, 8A–B, 9A–B, 10A–B, and Supplementary data Fig. 1). Functional assays demonstrated that higher expression of IFNG, TNF, CD4, and CD8A significantly suppressed LUAD cell proliferation, as shown by reduced cell viability (Figs. 7C, 8C, 9C and 10C). Consistently, colony formation assays revealed marked decreases in the number of colonies in overexpressing cells relative to controls (Figs. 7D-E, 8D-E, 9D-E and 10D-E), indicating impaired long-term proliferative capacity. Wound-healing assays further confirmed reduced migratory ability, with OE cells showing markedly slower closure of wound gaps at 24 h compared with control groups (Figs. 7F–G, 8F–G, 9F and 10F–G). These findings suggest that IFNG, TNF, CD4, and CD8A restrain LUAD cell motility and metastatic potential. To evaluate the tumor-suppressive role of each hub gene independently in vivo, A549 cells stably overexpressing IFNG, TNF, CD4, or CD8A were subcutaneously injected into nude mice, and tumor growth was monitored longitudinally. Each gene was analyzed as a separate experimental group and compared with its corresponding vector control. Consistent with the in vitro findings, xenografts derived from IFNG-overexpressing cells exhibited significantly reduced tumor size and volume compared with control tumors (Fig. 10H–I). Similarly, TNF overexpression markedly suppressed tumor expansion, resulting in smaller tumor mass and volume (Fig. 10H-I). Independent overexpression of CD4 also led to a significant reduction in tumor size and volume. Likewise, tumors formed from CD8A-overexpressing cells demonstrated significantly attenuated growth compared with controls (Fig. 10H-I). Furthermore, the proposed mechanistic model (Fig. 10J) suggests that downregulation of IFNG, TNF, CD4, and CD8A in LUAD may facilitate immune escape by impairing antigen presentation, reducing pro-apotic signaling, and enhancing immunosuppressive pathways within the tumor microenvironment. Conversely, enforced expression of these genes partially counteracts these protumorigenic processes, thereby restraining tumor progression in vivo. Functional assays of IFNG and TNF in A549 LUAD cells. A RT-qPCR confirmation of IFNG and TNF overexpression. B Western blot validation of IFNG and TNF protein levels. C Cell proliferation analysis following overexpression. D, E Colony formation assays of control and OE groups. F, G Wound-healing assays demonstrating reduced migration after overexpression. P***-value < 0.001 Functional assays of IFNG and TNF in H1299 LUAD cells. A RT-qPCR confirmation of IFNG and TNF overexpression. B Western blot validation of IFNG and TNF protein levels. C Cell proliferation analysis following overexpression. D–E Colony formation assays of control and OE groups. F–G Wound-healing assays demonstrating reduced migration after overexpression. P***-value < 0.001 Functional assays of CD4 and CD8A in A549 LUAD cells. A RT-qPCR confirmation of CD4 and CD8A overexpression. B Western blot validation of CD4 and CD8A protein levels. C Cell proliferation analysis following overexpression. D, E Colony formation assays of control and OE groups. F Wound-healing assays demonstrating reduced migration after overexpression. P***-value < 0.001 Functional assays of CD4 and CD8A in H1299 LUAD cells and xenograft validation. A RT-qPCR confirmation of CD4 and CD8A overexpression. B Western blot validation of CD4 and CD8A protein levels. C Cell proliferation analysis following overexpression. D–E Colony formation assays of control and OE groups. F–G Wound-healing assays demonstrating reduced migration after overexpression. H Representative images of xenograft tumors in the control and hub gene overexpression groups (OE tumors are smaller). I Tumor growth curves showing suppressed growth in xenografts derived from hub gene–overexpressing A549 cells compared with controls. J Schematic illustration of proposed mechanism by which downregulation of hub genes promotes immune escape and LUAD progression. P***-value < 0.001 Discussion LUAD is the most prevalent subtype of NSCLC and remains a leading cause of cancer-related mortality worldwide [30]. Despite advances in early detection, targeted therapies, and immune checkpoint inhibitors, the five-year survival rate for LUAD patients remains disappointingly low, especially in advanced stages [31]. A major reason for this poor prognosis is the complex molecular and immunological heterogeneity of LUAD, which contributes to immune evasion, drug resistance, and variable responses to immunotherapy [32, 33]. While PD-1/PD-L1 inhibitors have provided durable responses in subsets of patients, many either fail to respond or develop resistance during treatment [34, 35]. Therefore, there is a critical need to identify novel immune-related biomarkers that not only improve our understanding of LUAD biology but also serve as prognostic indicators and potential therapeutic targets. In this study, we systematically identified and validated four immune-related hub genes (IFNG, TNF, CD4, and CD8A) through an integrative approach combining large-scale bioinformatics analyses with in vitro and in vivo validation. Our findings reveal that these hub genes are consistently dysregulated in LUAD. Consistent with earlier reports, we observed marked downregulation of IFNG, TNF, CD4, and CD8A in other cancers [36, 37]. IFNG is a critical cytokine for antitumor immunity, with reduced levels linked to poor prognosis in lung cancer [38]. Similarly, TNF has dual roles in tumor biology but its loss has been associated with immune suppression and tumor progression [39]. CD4 and CD8A, representing T helper and cytotoxic T-cell markers respectively, corroborating earlier findings of impaired T-cell immunity in LUAD [40, 41]. Importantly, survival analysis confirmed that reduced expression of all four hub genes correlated with poor overall survival, extending prior evidence by demonstrating that these genes are not only differentially expressed but also clinically relevant prognostic indicators. Our promoter methylation analysis revealed that IFNG and TNF were hypermethylated, while CD4 and CD8A showed hypomethylation in LUAD. Prior studies demonstrated IFNG promoter hypermethylation as a mechanism of immune evasion in several cancers [42, 43], while TNF methylation has been less frequently studied. The observed hypomethylation of CD4 and CD8A contrasts with reports of generalized hypermethylation-driven immune silencing [44, 45], suggesting that additional regulatory mechanisms (e.g., miRNA suppression) may explain their reduced expression. Interestingly, although CD4 and CD8A showed relatively lower promoter methylation levels in LUAD compared with normal tissues, correlation analysis within the LUAD cohort still revealed a significant negative relationship between methylation and gene expression. This observation is not uncommon in large cancer cohorts, where group-level methylation differences and within-cohort methylation–expression correlations may reflect distinct regulatory contexts. In such cases, hypomethylation relative to normal tissue does not necessarily eliminate the inhibitory effect of residual methylation on transcription. Similar observations have been reported in recent multi-omics studies demonstrating that partial correlations integrating additional regulatory layers such as chromatin accessibility and non-coding RNA activity can resolve apparent methylation–expression paradoxes [46]. Therefore, the observed relationship for CD8A likely reflects the complex interplay of epigenetic and post-transcriptional regulatory mechanisms, rather than contradicting the role of methylation in modulating gene expression in LUAD. In the present study, the mutation frequencies of the identified hub genes were relatively low, with CD4 showing the highest alteration rate (~ 10%). This observation is consistent with previous studies reporting that immune regulatory genes in LUAD often exhibit low mutation frequencies but strong functional relevance in immune signaling pathways. Large-scale genomic studies have demonstrated that comprehensive mutation landscape analysis across the full TCGA cohort can provide deeper insights into environmental and immune-related drivers of lung cancer [47]. Therefore, although mutation events in the hub genes are relatively infrequent, their biological significance may arise primarily through transcriptional, epigenetic, and immune regulatory mechanisms rather than high mutation burden. Although CNV alterations in the hub genes showed a trend toward poorer survival outcomes, the differences did not reach statistical significance in the present analysis. This may reflect limited statistical power or the heterogeneous molecular landscape of LUAD. Recent studies have suggested that subtype-stratified analyses can improve the detection of CNV-associated prognostic effects, particularly for genes involved in oncogenic signaling pathways [48]. Therefore, future studies incorporating LUAD molecular subtype stratification may provide a more refined understanding of the clinical impact of these genomic alterations. Kaplan–Meier survival analysis indicated that lower expression of IFNG, TNF, CD4, and CD8A was associated with poorer overall survival in LUAD. It should be noted that the KM Plotter platform primarily performs univariate survival stratification, and the analysis does not account for potential clinical confounders such as patient age, tumor stage, or smoking status. Recent studies have emphasized that multivariable Cox regression models incorporating clinical covariates are important for establishing the independent prognostic value of immune-related biomarkers [49, 50]. Therefore, although our findings suggest a potential prognostic role for these hub genes, further validation using multivariate survival modeling in independent LUAD cohorts would be required to confirm their independent prognostic significance. In the present study, candidate miRNAs targeting the hub genes were initially predicted using the TargetScan database. Although TargetScan provides multiple prioritization metrics such as context + + scores and conservation filters, our study further prioritized miRNAs based on experimental expression validation and functional reporter assays. Importantly, dual-luciferase reporter assays confirmed that miR-4319, miR-130b-3p, miR-4644, and miR-326 directly target the 3′UTRs of IFNG, TNF, CD4, and CD8A, respectively. Recent studies have emphasized that computational predictions alone may not fully capture true regulatory interactions and that experimental 3′UTR validation remains essential for confirming miRNA–mRNA targeting relationships [51]. Therefore, the combination of bioinformatic prediction, expression validation, and luciferase assays in our study strengthens the evidence supporting these miRNA–hub gene regulatory interactions in LUAD. We further demonstrated that hub gene expression was strongly associated with immune subtypes, checkpoint molecules, and immune infiltration. This aligns with reports that high IFNG and CD8A levels predict enhanced immune infiltration and response to immunotherapy [52, 53]. Interestingly, we found that hub gene expression correlated with sensitivity to chemotherapeutics such as gefitinib and paclitaxel. Previous studies have described immune gene expression as predictors of immunotherapy outcomes [54, 55], but their role in modulating chemotherapy sensitivity is less explored, marking this as a novel aspect of our study. These associations were derived from pharmacogenomic correlations implemented in the GSCA platform using drug response data from large cancer cell line panels such as GDSC. However, these analyses do not incorporate LUAD subtype stratification or pharmacogenomic covariates such as EGFR mutation status, which are known to strongly influence response to EGFR inhibitors. Therefore, the observed associations should be interpreted as hypothesis-generating correlations rather than direct predictors of therapeutic response. Future studies integrating LUAD-specific pharmacogenomic datasets and molecular subtype information may further clarify the clinical relevance of these relationships. Our in vitro assays revealed that overexpression of IFNG, TNF, CD4, and CD8A suppressed LUAD cell proliferation, colony formation, migration, and invasion, consistent with their tumor-suppressive functions. In vivo xenograft experiments confirmed that overexpression of these genes reduce tumor growth. While previous studies have shown IFNG and TNF to inhibit tumorigenesis via JAK–STAT and NF-κB pathways [56], experimental validation of CD4 and CD8A in LUAD cell lines has been limited. Our findings uniquely demonstrate that restoring CD4 and CD8A expression directly suppresses tumor aggressiveness, highlighting a novel link between T-cell–associated genes and intrinsic tumor cell behavior. Although our study provides compelling evidence for the roles of IFNG, TNF, CD4, and CD8A in LUAD, a few limitations should be acknowledged. First, while we validated findings using both in silico and in vitro approaches, our experimental work was restricted to a limited number of LUAD cell lines and xenograft models, which may not fully capture the heterogeneity observed in patient tumors. Second, the use of public databases such as TCGA, GEO, and GSCA introduces potential biases related to sample composition, clinical annotation, and data processing pipelines. Third, although we identified candidate miRNAs and methylation changes regulating hub gene expression, mechanistic assays directly confirming these regulatory interactions were not performed. Fourth, drug sensitivity analysis was based on computational predictions, and functional validation in clinical or preclinical therapeutic models is needed. Another limitation of this study relates to the hub gene identification strategy. The PPI network was constructed using a STRING confidence score ≥ 0.4 and hub genes were selected based on degree centrality in CytoHubba. Recent studies suggest that incorporating network permutation testing and multiple centrality metrics (e.g., betweenness, closeness, and MCC) may further enhance hub gene robustness in biological networks [57, 58]. Although the degree method is widely applied in cancer bioinformatics studies and the identified genes were subsequently validated through multi-omics datasets and functional experiments, future studies may incorporate these additional network validation strategies to further strengthen hub gene prioritization. One more methodological consideration relates to potential batch effects in TCGA-based transcriptomic analyses. Although GSCA provides normalized TCGA datasets, cross-cohort batch effects may influence stage-specific expression patterns. Recent studies have suggested that harmonization methods such as ComBat correction can improve reliability of progression analyses in large transcriptomic datasets [59]. Future studies incorporating explicit batch-effect correction may further refine stage-related expression trends of the identified hub genes. Another limitation relates to the immune infiltration analysis performed using TISIDB and GSCA databases. Although these platforms provide immune cell estimation based on established gene expression signatures, the present study did not incorporate multiple immune deconvolution algorithms such as CIBERSORTx, EPIC, TIMER2, or MCP-counter, nor did it perform tumor purity adjustment. Recent studies suggest that combining multiple immune inference methods may improve robustness of immune landscape analyses. Future work incorporating multi-algorithm immune deconvolution and purity correction may further refine the immune associations of the identified hub genes in LUAD. Finally, our study was not supported by prospective patient cohort validation, which would strengthen the clinical applicability of the identified biomarkers. Future work should address these limitations through larger-scale, multi-omics integration and clinical validation studies. Conclusion In conclusion, our study highlights the central role of IFNG, TNF, CD4, and CD8A in LUAD progression and immune regulation. By integrating computational predictions with laboratory validation, we provide compelling evidence that these genes act as tumor suppressors and immune regulators. Their downregulation, mediated by promoter methylation, genomic alterations, and miRNA regulation, promotes immune evasion and tumor progression. Our findings expand current knowledge by identifying novel regulatory mechanisms and functional roles for these hub genes and suggest that restoring their activity may represent a promising therapeutic strategy for LUAD. Supplementary Information Supplementary Material 1. Acknowledgements The authors extend their appreciation to Prince Sattam bin Abdulaziz University for funding this research work through the project number (PSAU/2024/03/31564). Author contributions Abdullah Alghamdi and Mohammed Alissa conceived and designed the study. Abdullah Alghamdi performed the bioinformatics analyses, carried out the experiments, and prepared the figures. Mohammed Alissa supervised the project, validated the findings, and provided critical revisions to the manuscript. Both authors contributed to data interpretation, manuscript drafting, and approved the final version of the paper. Funding None. Data availability The datasets analyzed during the current study are publicly available in the Gene Expression Omnibus (GEO) database under accession numbers GSE13850, GSE16088, and GSE12865. Additional datasets were obtained from public repositories including The Cancer Genome Atlas (TCGA), cBioPortal, Human Protein Atlas (HPA), TISIDB, TargetScan, OncoDB, and GSCA databases. All data generated or analyzed during this study are included in this published article and its supplementary information files. Declarations Ethics approval and consent to participate The use of the laboratory animals was approved by the Standing Committee of Bioethics Research (SCBR) at Prince Sattam bin Abdulaziz University, Al-Kharj, Saudi Arabia (Approval no. SCBR-579/2025). The study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki and the ARRIVE guidelines. All animal procedures for the xenograft experiments were performed in compliance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, USA), and efforts were made to minimize animal suffering. Nude mice were maintained under pathogen-free conditions with controlled temperature, humidity, and a 12-hour light–dark cycle, with ad libitum access to food and water. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement The datasets analyzed during the current study are publicly available in the Gene Expression Omnibus (GEO) database under accession numbers GSE13850, GSE16088, and GSE12865. Additional datasets were obtained from public repositories including The Cancer Genome Atlas (TCGA), cBioPortal, Human Protein Atlas (HPA), TISIDB, TargetScan, OncoDB, and GSCA databases. All data generated or analyzed during this study are included in this published article and its supplementary information files.