Identification of CAV1 and CDH5 as potential diagnostic and prognostic biomarkers in lung adenocarcinoma

Introduction

Lung adenocarcinoma (LUAD) is the most prevalent subtype of lung cancer and according to statistics reported in 2024, LUAD continues to rank among the leading malignancies in both incidence and mortality worldwide. In the United States, LUAD accounted for ~45% of lung cancers in 2021, with an age-adjusted incidence of ~22 per 100,000 population (1). For example, recent advances include low-dose computed tomography screening that reduces lung-cancer mortality, and precision therapies such as EGFR-targeted osimertinib (with or without chemotherapy), RET inhibitors (selpercatinib), KRAS p.G12C inhibitors (sotorasib), and immune checkpoint blockade (pembrolizumab-based chemoimmunotherapy). Despite notable progress in diagnostic and therapeutic strategies, such as low-dose CT screening, EGFR-targeted therapy and immune checkpoint inhibitors, the 5-year survival rate for LUAD remains poor at ~15-20% (2,3). With the rapid development of high-throughput sequencing and bioinformatics, researchers can now systematically identify key molecular targets related to tumor initiation and progression at the genomic level (4-6); however, precise therapeutic targets for LUAD are still lacking.

In addition to genetic and transcriptomic characterization, advances in nanotechnology and biomolecular engineering have introduced new opportunities for improving LUAD diagnosis and treatment. For example, macrophage-mediated sulfate-based nanomedicine has demonstrated enhanced drug delivery efficacy within the tumor microenvironment of lung cancer (7), whilst nano-assisted radiotherapy strategies have shown potential in increasing treatment precision and efficacy in non-small cell lung cancer (8). Furthermore, self-assembled DNA-based biosensors have enabled rapid and portable detection of circulating tumor cells in lung cancer patient blood samples (n=46), achieving 100% specificity and 86.5% sensitivity, supporting early diagnosis and disease monitoring in lung cancer (9). These technological innovations underscore the need to explore molecular-level biomarkers that could complement or integrate with such novel platforms to improve clinical outcomes in LUAD.

The present study employed integrated bioinformatics tools and machine learning algorithms and conducted in silico validation using independent GEO (GSE31210 discovery; GSE68465 validation) and TCGA-LUAD cohorts, as well as molecular validation by reverse transcription-quantitative (RT-q) PCR on paired LUAD and adjacent normal tissues (n=50) to identify and validate potential therapeutic targets for LUAD, with the aim to providing a foundation for future targeted treatment strategies.

Materials and methods

Identification of differentially expressed genes (DEGs)

A total of two LUAD-related datasets, GSE31210 and GSE68465(10), were retrieved from the GEO database (11), in which the control samples comprised non-tumorous normal lung tissues rather than patient-matched adjacent tissues (GSE31210: 20 normal, 226 LUAD; GSE68465: 19 normal, 443 LUAD); these datasets were chosen for their large sample sizes, availability of normal controls, and rich clinical annotation (including overall survival), making them widely used benchmark cohorts for discovery (GSE31210) and external validation (GSE68465). DEGs were identified using the ‘limma’ R package (v3.60.6, Bioconductor), implemented in R v4.4.1 (www.r-project.org/foundation/) (12), with screening criteria set as |log2fold change|>1 and P<0.05. Volcano plots were generated using the ‘EnhancedVolcano’ package for visualization (13).

Weighted gene co-expression network analysis (WGCNA)

WGCNA was performed on the GSE31210 dataset using the ‘WGCNA’ R package (14), because this dataset contains both tumor and normal samples with comprehensive clinical annotation and a sufficiently large sample size, making it suitable for constructing reliable co-expression networks to identify key gene modules associated with LUAD. A soft-thresholding power β=6 was chosen based on the scale-free topology criterion (scale-free R²>0.85). The minimum module size was set to 30 to ensure module robustness. The resulting sample dendrogram and trait heatmap are presented in Fig. S1. Modules with a Pearson correlation coefficient of r>0.92 and P<0.05 were considered statistically significant, with correlations evaluated between module eigengenes and clinical traits (such as tumor vs. normal status, stage), measured from the sample annotation of the GSE31210 dataset.

Gene set enrichment analysis (GSEA)

GSEA was performed using GSEA software (v4.3.2) (15) on the entire set of DEGs, whereas Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed using the ‘clusterProfiler’ R package (v4.12.0; www.r-project.org/foundation/) (16,17) on the same DEG set. GO analysis included biological process, cellular components and molecular functions. KEGG analysis was used to identify significantly enriched signaling pathways. P<0.05 and q<0.05 were considered to indicate a statistically significant difference.

Construction and analysis of the protein-protein interaction (PPI) network

Identified DEGs were submitted to the STRING database to construct a PPI network (18), with a confidence score threshold of >0.9 (high confidence). Interaction data were imported into Cytoscape (v3.10.0; https://cytoscape.org/) for visualization. Network topology was analyzed using two centrality algorithms: Degree Centrality and Betweenness Centrality (19).

Machine learning analysis

A total of two machine learning algorithms were applied: Random Forest (RF) and Support Vector Machine-Recursive Feature Elimination (SVM-RFE) (20,21). RF analysis was performed using the ‘randomForest’ package (v4.7-1.2) in R v4.4.1 (www.r-project.org/foundation/), and feature importance was ranked accordingly according to the default feature importance measures implemented in the ‘randomForest’ R package (Mean Decrease Accuracy and Mean Decrease Gini, as recommended by the package developers). SVM-RFE analysis was performed using the ‘e1071’ package (v1.7-14) in R v4.4.1 (www.r-project.org/foundation/) to further select optimal features based on gene importance. A linear kernel was applied with the default cost parameter (C=1), and feature ranking was performed using 10-fold cross-validation. The gene subset corresponding to the lowest cross-validated classification error was selected as the optimal feature set. P<0.05 was considered to indicate a statistically significant difference.

RT-qPCR

RT-qPCR (22) was used to validate the expression levels of caveolin-1 (CAV1) and cadherin 5 (CDH5) in tumor and adjacent normal tissues from 50 patients with LUAD, with adjacent tissues collected at least 5 mm away from the tumor margin. All tissue specimens were collected during surgical resection at Hebei Provincial Hospital of Traditional Chinese Medicine (Shijiazhuang, China) between May 2021 and June 2023, immediately snap-frozen in liquid nitrogen, and stored at -80˚C until RNA extraction. Among the patients, 28 were male and 22 were female. Inclusion criteria were histopathologically confirmed LUAD and availability of paired tumor and adjacent normal tissues; exclusion criteria included prior chemotherapy, radiotherapy, or other malignancies. The median age was 63 years (range, 45-78 years). The study protocol was approved by the Ethics Committee of Hebei Provincial Hospital of Traditional Chinese Medicine (Shijiazhuang, China; approval no. HBZY2023-KY-076-01) and followed the ethical principles of The Declaration of Helsinki. Written informed consent was obtained from all participants.

Total RNA was extracted using TRIzol® reagent (Thermo Fisher Scientific, Inc.), and RNA purity and concentration were assessed using a NanoDrop 2000 spectrophotometer. High-quality RNA was reverse transcribed into cDNA, and qPCR was performed using SYBR Green PCR Master Mix on a QuantStudio 6 Real-Time PCR System (Thermo Fisher Scientific, Inc.). Relative gene expression was calculated using the 2-ΔΔCq method (22), with GAPDH used as the internal control. Primers were designed and validated using Primer-BLAST. The following primer sequences were used: GAPDH forward, 5'-AATGGACAACTGGTCGTGGAC-3' and reverse, 5'-CCCTCCAGGGGATCTGTTTG-3'; CAV1 forward, 5'-GCAGAACCAGAAGGGACACAG-3' and reverse, 5'-CCAAAGAGGGCAGACAGCAAGC-3'; and CDH5 forward, 5'-AGGTGCTAACCCTGCCCAAC-3' and reverse, 5'-CGGAAGACCTTGCCCACATA-3'.

Receiver operating characteristic (ROC) curve analysis

ROC curve analysis was performed using the ‘pROC’ package (v1.18.5) in R v4.4.1 (www.r-project.org/foundation/) (23). An area under the curve (AUC) of >0.9 was considered to indicate notable diagnostic performance. P<0.05 was considered to indicate a statistically significant difference.

Single-cell RNA sequencing (scRNA-seq) analysis

scRNA-seq data related to LUAD were obtained from the Tumor Immune Single Cell Hub (TISCH) database (http://tisch.comp-genomics.org) (24). The expression levels of key genes across different cell types were visualized using heatmaps and feature plots, which were generated directly from the TISCH database portal based on preprocessed single-cell RNA-seq data provided by the database.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 9.0 (Dotmatics). Differences between tumor and normal tissues in The Cancer Genome Atlas (TCGA) and GEO datasets were assessed using unpaired Student's t-test, while differences between matched tumor and adjacent normal tissues in the clinical validation cohort and paired comparisons within the TCGA cohort (Fig. 3D) were assessed using paired Student's t-test. For comparisons among multiple subgroups, one-way ANOVA followed by Tukey's post hoc test was applied. Kaplan-Meier survival analysis was performed using the Kaplan-Meier Plotter database (http://kmplot.com/analysis/) based on TCGA-LUAD data; patients were stratified into high- and low-expression groups using the median expression value as the cut-off, and statistical significance was evaluated using the log-rank test. P<0.05 was considered to indicate a statistically significant difference.

Figure 3 Expression and prognostic analysis of key genes. Expression of CAV1 and CDH5 in the (A) GSE31210, (B) GSE68465 and (C) TCGA datasets. (D) Paired expression analysis of CAV1 and CDH5 in TCGA samples. (E) RT-qPCR validation in tissues from patients with LUAD (n=50). (F) Kaplan-Meier survival curves for CAV1 and CDH5, comparing patients with high expression and low expression (grouped according to the median expression level). P-values were calculated using the log-rank test. (G) Diagnostic nomogram based on CAV1 and CDH5. (H) Calibration curve evaluating nomogram accuracy. (I) Receiver operating characteristic curve analysis in the GSE31210, GSE68465 and TCGA datasets. ***P<0.001 and ****P<0.0001. CAV1, caveolin-1; CDH5, cadherin 5; TCGA, The Cancer Genome Atlas; RT-qPCR, reverse transcription-quantitative PCR; LUAD, lung adenocarcinoma; AUC, area under the curve; TRP, true positive rate; FRP, false positive rate.

Results

Identification of overlapping genes between DEGs and key WGCNA modules

Based on the GSE31210 dataset, a total of 9,292 DEGs were identified, including 5,399 upregulated and 3,893 downregulated genes in LUAD tumor tissues compared with controls, as this dataset contains both tumor and non-tumor lung tissues with comprehensive clinical annotation, making it suitable for DEG identification. A volcano plot was generated to visualize these DEGs (Fig. 1A). To further identify key genes, a WGCNA was performed using the same dataset, resulting in the identification of 25 co-expression modules (Fig. 1B).

Figure 1 Integrated analysis of DEGs and key WGCNA modules. (A) Volcano plot of DEGs from the GSE31210 dataset (x-axis: log2 fold change; y-axis: -log10 adjusted P-value). (B) Gene clustering dendrogram. ‘Dynamic Tree Cut’ indicates modules initially defined by hierarchical clustering, and ‘Merge Dynamic’ represents merging of highly similar modules. (C) Heatmap of module-trait correlations. (D) Scatter plot of gene significance vs. module membership in the brown module. (E) Venn diagram of overlapping genes between DEGs and the key brown module. (F) Gene Ontology enrichment analysis of overlapping genes (x-axis: gene counts; y-axis: enrichment score with adjusted P-values). (G) Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis of overlapping genes (x-axis: gene counts; y-axis: enrichment score with adjusted P-values). Gene counts and the corresponding adjusted P-value levels are shown for each enriched term or pathway. **P<0.01 and ***P<0.001. DEG, differentially expressed gene; WGCNA, weighted gene co-expression network analysis; LUAD, lung adenocarcinoma; MF, molecular function; CC, cellular component; BP, biological process; sig, significant.

Correlation analysis between these modules and LUAD phenotypes revealed that 17 modules were significantly associated with LUAD, among which the brown module showed the strongest correlation (Fig. 1C). This suggested that this module may serve a critical role in the pathogenesis and progression of LUAD. Further scatter plot analysis of module membership and gene significance demonstrated a strong positive correlation within the brown module (Pearson's r=0.92, P=1x10-1,000), indicating its relevance as a key module (Fig. 1D).

Subsequently, a Venn diagram was used to identify overlapping genes between DEGs and those in the key module identified by WGCNA, resulting in 1,059 common genes (Fig. 1E). GO and KEGG enrichment analyses were then performed on these overlapping genes. GO analysis revealed enrichment in signaling-related molecular functions such as ‘signaling receptor binding’, ‘kinase activity’ and ‘protein-containing complex binding’, cellular components such as ‘vesicle’, ‘endomembrane system’ and ‘extracellular region part’, as well as biological processes such as ‘regulation of cell communication’, ‘cellular developmental process’ and ‘negative regulation of cellular process’ (Fig. 1F). KEGG analysis revealed significant enrichment in ‘TGF-β signaling pathway’, ‘TNF signaling pathway’, ‘PI3K-Akt signaling pathway’ and ‘pathways in cancer’, suggesting that these genes may serve regulatory roles in LUAD initiation and progression (Fig. 1G).

Identification of key LUAD genes based on PPI network and machine learning

A PPI network was constructed using the 1,059 overlapping genes (Fig. 2A). Topological analysis in Cytoscape using the Degree Centrality and Betweenness Centrality algorithms identified the top 20 hub genes (Fig. 2B and C, respectively). The intersection of these two analyses yielded 13 shared hub genes (Fig. 2D).

Figure 2 Screening of lung adenocarcinoma key genes based on PPI network and machine learning. (A) PPI network of overlapping genes. PPI subnetworks ranked by (B) Degree and (C) Betweenness. (D) Venn diagram of overlapping hub genes from both algorithms. Top-ranked genes from the (E) RF and (F) SVM-RFE algorithms. (G) Venn diagram identifying shared genes from the RF and SVM-RFE algorithms. GSEA results for (H) CAV1 and (I) CDH5. Feature importance and GSEA significance were determined based on model-specific thresholds. PPI, protein-protein interaction; RF, Random Forest; SVM-RFE, Support Vector Machine-Recursive Feature Elimination; GSEA, gene set enrichment analysis; CAV1, caveolin-1; CDH5, cadherin 5; ES, enrichment score; NP, normalized P-value; NS, normalized score.

Subsequently, two machine learning approaches, RF and SVM-RFE, were employed to further refine key gene selection. The RF algorithm identified the top five genes based on feature importance scores (Fig. 2E), whereas SVM-RFE selected another top five genes through iterative optimization (Fig. 2F). A Venn diagram analysis of both methods identified two overlapping candidate genes: CAV1 and CDH5 (Fig. 2G).

Furthermore, GSEA indicated that both CAV1 and CDH5 were significantly enriched in signaling pathways such as MAPK, Wnt and TGF-β, suggesting their involvement in LUAD development, progression and modulation of the tumor immune microenvironment (Fig. 2H and I).

Expression analysis and prognostic model construction of key genes

To explore the roles of CAV1 and CDH5 in LUAD, their expression levels were analyzed in the GSE31210 dataset. Both genes were significantly downregulated in LUAD tissues compared with those in normal tissues (Fig. 3A). Moreover, consistent downregulation trends were observed in the GSE68465 and TCGA datasets for CAV1 (Fig. 3B and C). Although CDH5 expression was upregulated in the LUAD tissues in the GSE68465 dataset, it was downregulated in the TCGA cohort. To further investigate whether tumor stage contributes to this inter-dataset discrepancy, a subgroup analysis was performed using the GSE68465 dataset, stratifying patients by N stage (N0, N1 and N2). As shown in Fig. S2, CDH5 expression was significantly higher in N0-stage samples compared with that in N1 (P<0.001) and a significant difference was also observed between N1 and N2 (P<0.05). This stage-associated decline supports a biological trend of progressive CDH5 downregulation during LUAD advancement.

Pairwise analyses in the TCGA dataset demonstrated significantly lower expression of CAV1 and CDH5 in LUAD tissues compared with available adjacent normal lung tissues, eliminating potential interference from inter-individual biological variation (Fig. 3D). Similarly, RT-qPCR results from 50 LUAD samples demonstrated significantly decreased expression levels of CAV1 and CDH5 in tumor tissues compared with those in matched adjacent normal tissues (P<0.0001; Fig. 3E). Notably, although the difference in CDH5 expression between tumor and control tissues was relatively small in magnitude, it remained statistically significant. This expanded validation supports the reliability and reproducibility of the bioinformatics-driven identification of these genes.

Furthermore, Kaplan-Meier survival analysis revealed that high expression of CAV1 and CDH5 was significantly associated with improved overall survival, suggesting that CAV1 and CDH5 have potential tumor suppressor functions (Fig. 3F). A diagnostic nomogram model based on the expression of CAV1 and CDH5 was then constructed (Fig. 3G), with calibration curves indicating strong agreement between predicted and actual outcomes, as the calibration curves closely overlapped with the ideal 45˚ reference line (Fig. 3H). In addition, ROC curve analysis demonstrated notable diagnostic performance for CAV1 (AUC=0.979) and CDH5 (AUC=0.969) in GSE31210, and this was validated in both GSE68465 and TCGA datasets (Fig. 3I), supporting the potential of CAV1 and CDH5 as diagnostic biomarkers for LUAD.

Expression features of key genes in single-cell transcriptomics

To precisely characterize the cell-type-specific expression and potential functions of CAV1 and CDH5, three LUAD-related scRNA-seq datasets (EMTAB6149, GSE127465 and GSE131907) from the TISCH database were analyzed: EMTAB6149 includes ~20,000 cells in total from the primary LUAD tissues of 4 patients; GSE127465 comprises ~14,000 cells in total from surgically resected LUAD tumors and adjacent tissues from 7 patients; and GSE131907 includes ~208,506 cells in total derived from 11 patients with LUAD. These datasets represent a diverse range of LUAD microenvironments and provide robust resolution of immune and stromal cell populations. Both genes were revealed to be highly expressed in endothelial cells (Fig. 4A), suggesting potential roles in tumor-associated angiogenesis, vascular barrier maintenance or microenvironment regulation. To visualize their spatial expression across immune cell subpopulations, t-distributed stochastic neighbor embedding plots were generated for CAV1 and CDH5, which showed that both genes were predominantly expressed in endothelial cell clusters, with lower expression observed in other immune cell subtypes (Fig. 4B). Additionally, violin plots were used to quantify expression levels across major cell types, revealing marked heterogeneity in gene expression patterns, with CAV1 and CDH5 showing higher expression in endothelial cells but minimal expression in most immune cell subsets, indicating substantial variability across cell populations (Fig. 4C). These findings not only corroborate the high expression observed in heatmaps, particularly in endothelial cells, but also highlight the complexity of gene regulation in the tumor microenvironment, providing a theoretical basis for further functional studies.

Figure 4 Single-cell expression features of CAV1 and CDH5 in lung adenocarcinoma. (A) Heatmaps of gene expression across immune cell types for CAV1 (left) and CDH5 (right). (B) t-distributed stochastic neighbor embedding visualization of spatial gene expression patterns. (C) Violin plots of expression heterogeneity across immune subpopulations. Data were retrieved from the Tumor Immune Single Cell Hub database. Cell type annotations were based on original dataset classifications. TPM, transcripts per million. Cell type abbreviations: Tconv, conventional CD4+ T cells; Treg, regulatory T cells; T Tex, exhausted CD8+ T cells; DC, dendritic cells; Mono/Macro, monocytes/macrophages; NK, natural killer cells. CAV1, caveolin-1; CDH5, cadherin 5.

Discussion

LUAD exhibits notable heterogeneity in terms of its molecular characteristics, cellular composition, and clinical behavior, posing ongoing challenges for both diagnosis and targeted therapy (25). The present study systematically identified key genes participating in LUAD, namely CAV1 and CDH5, through integrated differential expression analysis across multiple GEO and TCGA datasets, WGCNA, PPI network topology screening and dual machine learning algorithms. The expression profiles and potential functions of these genes in LUAD were then further explored.

CAV1 is a scaffolding protein involved in membrane signal integration, cytoskeletal remodeling and tumor-suppressive signal transduction, and has been implicated in LUAD proliferation, migration and mechanisms of drug resistance (26). CDH5 (coding for VE-cadherin) is an endothelial cell-specific adhesion molecule that maintains vascular integrity and participates in tumor angiogenesis and immune evasion (27). In the context of LUAD progression, the downregulation of CAV1 and CDH5 may actively contribute to tumor aggressiveness through multiple mechanisms. CAV1, a structural component of caveolae, is known to regulate membrane-associated signaling and actin cytoskeleton dynamics (28). Its loss has been associated with the activation of the MAPK/ERK and Wnt/β-catenin pathways, both of which are implicated in promoting epithelial-mesenchymal transition (EMT), increased motility and the metastatic capacity of LUAD cells (29). CAV1 downregulation also facilitates invadopodia formation, enabling extracellular matrix degradation and local invasion (30). Furthermore, loss of CDH5 expression has been associated with enhanced invasiveness in several cancers, including breast, gastric and colorectal cancers, as demonstrated in a recent pan-cancer analysis (31). In the present study, survival analysis demonstrated that high expression of both CAV1 and CDH5 was significantly associated with improved overall survival, suggesting the tumor-suppressive roles of these genes in LUAD.

CDH5/VE-cadherin serves a critical role in maintaining endothelial junctional integrity (32), and reduced CDH5 expression in LUAD may disrupt the vascular endothelial barrier, thereby increasing vascular permeability and promoting tumor cell intravasation into the bloodstream. This mechanism supports distant metastasis and has been associated with enhanced extravasation and colonization at secondary sites (33). Furthermore, loss of CDH5 expression may impair immune surveillance by altering leukocyte transmigration, thereby contributing to the formation of an immune-privileged tumor niche, characterized by reduced immune cell infiltration and evasion from host immune responses (34). These effects highlight the functional importance of CAV1 and CDH5 beyond their diagnostic value, further supporting their role as potential therapeutic targets in LUAD. Additionally, a diagnostic nomogram model constructed based on these two genes exhibited strong predictive accuracy. ROC analysis across multiple independent datasets also demonstrated robust diagnostic potential of CAV1 and CDH5, supporting their reliability as candidate molecular biomarkers for LUAD. While CDH5 expression was found to be upregulated in the GSE68265 dataset, this trend was not replicated in other independent datasets. To explore whether clinical staging may account for this discrepancy, a subgroup analysis stratified by N stage was performed using the GSE68465 cohort. The results demonstrated a significant stepwise decline in CDH5 expression from N0 to N2 stages, supporting a stage-dependent downregulation pattern. This finding aligns with the trends observed in the TCGA dataset and the RT-qPCR validation cohort conducted in the present study. Taken together, these data suggest that CDH5 downregulation is a consistent feature in LUAD progression, and that isolated inter-dataset variability may be attributable to sample composition, stromal content or platform-specific differences. These observations underscore the importance of integrating stratified analysis in biomarker evaluation to mitigate confounding effects and enhance biological interpretability.

In terms of underlying mechanisms, GSEA revealed that CAV1 and CDH5 were significantly enriched in key signaling pathways, including MAPK, Wnt and TGF-β. These pathways are known to be involved in cell proliferation, migration, metastasis and immune regulation in LUAD (35-37). CAV1 generally acts as an inhibitor of the MAPK/ERK and Wnt/β-catenin pathways, while loss of CDH5 function disrupts endothelial junctions and facilitates pro-tumorigenic signaling, thereby highlighting their tumor-suppressive roles. Notably, the activation of MAPK and Wnt signaling has been reported to promote tumor cell growth and maintain cancer stemness (38), whereas the TGF-β pathway in LUAD has shown stage-dependent effects, exhibiting tumor-suppressive roles in early stages but promoting EMT and metastasis in later stages (39).

Further analysis of single-cell transcriptomic data revealed that CAV1 and CDH5 are predominantly expressed in endothelial cells, suggesting the hypothesis that they may be involved in maintaining endothelial function (40), regulating tumor-associated angiogenesis and contributing to the formation of the immune barrier. A recent study has reported that dysfunction of tumor endothelial cells is associated with abnormal vascularization, immune escape and distant metastasis, which is consistent with the findings of the present study (41). Although these processes were not directly investigated in the present study, such associations can be hypothesized based on the enriched terms and pathways identified in the bioinformatics analysis, which are consistent with the findings of the previous report. Whilst this endothelial-specific expression pattern suggests a possible involvement in angiogenic regulation, this inference is based on gene localization and prior biological evidence rather than direct functional association. In the current study, the proportion of endothelial subclusters expressing CAV1 or CDH5 were not quantified, and associations with angiogenesis-related signatures or histological vessel density were not assessed. Therefore, future studies integrating these analyses will be necessary to confirm the functional roles of these genes in LUAD-associated angiogenesis.

From a clinical standpoint, the incorporation of CAV1 and CDH5 into diagnostic workflows may hold promising translational potential. Given their consistent differential expression between LUAD and normal tissues, both genes could be evaluated using immunohistochemistry on formalin-fixed paraffin-embedded lung biopsy specimens, allowing pathologists to assess protein levels directly in tumor samples. Alternatively, emerging liquid biopsy techniques, such as detection of circulating tumor RNA or protein in plasma, could potentially enable the non-invasive, dynamic monitoring of these biomarkers during disease progression or treatment response. Although CAV1 and CDH5 are cytoskeleton-associated proteins and are unlikely to be secreted directly, their dysregulation may still be detected indirectly through circulating tumor RNA or extracellular vesicles. A study demonstrated the feasibility of liquid biopsy for monitoring molecular alterations in patients with LUAD (42). Although further validation is necessary, particularly at the protein and clinical implementation level, the findings of the present study provide a rationale for the future development of CAV1 and CDH5 as clinically useful biomarkers in LUAD diagnosis and patient management.

Furthermore, whilst the nomogram and ROC analyses in the present study demonstrated notable diagnostic performance for CAV1 and CDH5, these results were derived from datasets with limited clinical covariates and without multivariate survival adjustment. In particular, the Kaplan-Meier survival curves were based on univariate analysis and did not account for potential confounders, such as tumor stage, patient age or treatment modalities. Similarly, although multiple datasets were used for validation, the high AUC values may partially reflect dataset-specific characteristics or overfitting. Therefore, future studies incorporating multivariate Cox regression analysis, larger prospective cohorts and bootstrap or external validation frameworks will be necessary to confirm the independent clinical utility of these biomarkers.

In addition, whilst the expression levels of CAV1 and CDH5 were validated at the mRNA level in an expanded clinical cohort, there was a lack of protein-level validation and histological confirmation of cell-type specificity within LUAD tissues. Future studies involving immunohistochemical staining and spatial transcriptomic profiling will be required to fully elucidate the functional roles and cellular localization of CAV1 and CDH5.

In summary, CAV1 and CDH5 appear to have notable biological roles in LUAD and may serve as promising targets for diagnosis and therapy. A limitation of the present study is that it focused on bioinformatics integration and clinical validation and did not include in vitro or in vivo functional experiments. However, the consistent expression patterns of CAV1 and CDH5 across multiple datasets and RT-qPCR analysis support their tumor-suppressive potential. Mechanistic validation, such as gene knockdown or overexpression studies in LUAD models, should be performed in future research to further elucidate the functional roles and therapeutic implications of these genes.

In conclusion, the present study systematically identified CAV1 and CDH5 as potential key tumor-suppressor genes in LUAD. Both genes were significantly downregulated in tumor tissues and associated with a favorable patient prognosis. The diagnostic model based on their expression demonstrated a notable predictive performance. Moreover, single-cell transcriptomic analysis highlighted the specific expression of CAV1 and CDH5 in endothelial cells, suggesting their potential involvement in the regulation of the tumor microenvironment. These findings provide a theoretical basis and novel molecular targets for the diagnosis and targeted treatment of LUAD.

Supplementary Material

Sample clustering dendrogram and phenotype heatmap for the GSE31210 dataset. Hierarchical clustering was performed using average linkage and Euclidean distance. The lower color bar represents clinical phenotype classification (purple, LUAD; blue, normal tissue). The dendrogram and heatmap confirm consistency between sample grouping and phenotype labels and indicate no obvious outliers. This visualization supports the reliability of subsequent module-trait correlation analyses. LUAD, lung adenocarcinoma.

CDH5 mRNA expression in patients with lung adenocarcinoma from the GSE68465 dataset, stratified by lymph node metastasis status (N stage). CDH5 expression was significantly higher in N0-stage samples compared with that in N1 (P<0.001) and a significant difference was also observed between N1 and N2 (P<0.05). *P<0.05 and ***P<0.001. CDH5, cadherin 5.

Acknowledgements

Not applicable.

Funding

Funding: The present study was funded by the Research Project of the Hebei Provincial Administration of Traditional Chinese Medicine (grant nos. 2020103 and 2022065).

Availability of data and materials

The datasets generated in the present study may be requested from the corresponding author.

Authors' contributions

LZ conceived and designed the study, developed the methodology and drafted the manuscript. YW contributed to software implementation, validation and project administration. YuL performed formal analysis, participated in data acquisition and interpretation and contributed to securing funding. YaL conducted experimental investigation and provided essential resources (clinical samples and reagents). LW curated the data, performed statistical analysis and visualization. CW contributed to the study conception and manuscript writing, reviewing and editing. LZ and CW confirm the authenticity of all the raw data. All authors read and approved the final manuscript.

Ethics approval and consent to participate

The study was approved by the Ethics Committee of Hebei Provincial Hospital of Traditional Chinese Medicine (approval no. HBZY2023-KY-076-01). Written informed consent was obtained from all participants.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References