Integrating zinc homeostasis network and immune landscape: a five-gene prognostic framework for precision oncology in lung adenocarcinoma

Background: Lung adenocarcinoma (LUAD) exhibits marked heterogeneity in clinical outcomes and therapeutic responses, underscoring the imperative for reliable prognostic biomarkers. Dysregulation of zinc homeostasis is an emerging hallmark of cancer, contributing to tumor progression through multifaceted mechanisms including exacerbated oxidative stress and sustained oncogenic signaling. This study aimed to develop and validate a novel prognostic signature based on zinc homeostasis network-related genes for stratifying LUAD patients into distinct risk groups to predict clinical outcomes and inform therapeutic strategies.

Methods: Transcriptomic and clinical profiles of LUAD cases from the TCGA database were integrated to screen for differentially expressed genes (DEGs) involved in zinc homeostasis network. A prognostic risk model was constructed via univariate, multivariate, and LASSO regression analyses, and externally validated using GEO datasets. Model performance was evaluated using a nomogram, time-dependent ROC curves, and decision curve analysis. To characterize immune microenvironment heterogeneity across risk subgroups, we applied seven deconvolution algorithms, ssGSEA and single-cell profiling. Spearman correlation analysis and Wilcoxon rank-sum tests were used for investigating the associations between risk stratification and immunomodulatory markers, tumor mutational burden (TMB), as well as the predicted responsiveness to conventional chemotherapeutic and targeted therapies. In vitro experiments validated the expression levels of key candidate genes and confirmed their biological functions.

Results: Differential expression analysis identified 124 zinc homeostasis network-related DEGs, of which 16 showed significant prognostic relevance. A five-gene risk model stratified patients into distinct prognostic groups, with high-risk cases showing markedly reduced overall survival. Risk score correlated positively with advanced clinical stages. Multivariate Cox regression confirmed the model as independently predictive of LUAD prognosis. A nomogram integrating risk score and clinical features was constructed for predicting 1-, 3-, and 5-year survival. Immune profiling showed that low-risk cases had hot tumor phenotypes, elevated immune scores and infiltration by immune cells, while those at high risk showed raised levels of immunotherapy resistance markers and increased TMB. Drug sensitivity analysis indicated differential responses to chemotherapeutic and targeted agents across risk groups. Independent knockdown of SLC16A3 or overexpression of EGR2 significantly suppressed malignant behaviors in LUAD cells. Additionally, SLC16A3 downregulation reduced cisplatin sensitivity in LUAD cells.

Conclusion: Our study highlights the clinical translational potential of a five-gene zinc homeostasis network signature in predicting prognosis and guiding personalized therapeutic strategies for LUAD.

1 Introduction

Among all types of lung cancer, lung adenocarcinoma (LUAD) is a prevalent and aggressive non-small cell lung cancer subtype, accounting for nearly 40% of cases and exhibiting marked intertumoral heterogeneity (1). Despite advancements in targeted therapies and immune checkpoint inhibitors, over 60% of patients develop therapeutic resistance within 12 months (2–4). This high rate of treatment failure underscores the critical unmet need for novel prognostic frameworks to both predict clinical outcomes and guide personalized therapeutic decisions.

Zinc ions function as essential intracellular second messengers (5), with tightly regulated homeostasis required for key physiological processes, including cellular growth, development, and immune regulation (6–8). Increasing evidence indicates that zinc homeostasis dysregulation is a hallmark of tumor progression (9, 10), wherein aberrant zinc flux drives tumor progression by activating oncogenic pathways such as MAPK and Akt to promote oncogenesis (11–13), while simultaneously inducing epithelial-mesenchymal transition (EMT) and enhancing matrix metalloproteinase-driven extracellular matrix degradation to facilitate invasiveness and metastasis (9, 14). Proteins involved in zinc transport serve as critical regulators of intracellular zinc homeostasis, and dysregulation of their expression has been linked to multiple malignant phenotypes. For instance, Yang et al. demonstrated that ZIP4 overexpression accelerates pancreatic tumor growth and cachexia in mice by stimulating RAB27B-mediated extracellular vesicle production (15). Similarly, Franklin et al. reported that reduced expression of hZIP1 contributes to early prostate cancer progression (16). Although zinc homeostasis network dysregulation promotes malignant progression in various cancers and specific transporters like SLC39A4 are implicated in lung cancer pathogenesis (17, 18), their collective impact on LUAD immune microenvironment remodeling remains uncharacterized. The potential of multi-gene signatures for constructing prognostic models and guiding clinical therapeutic regimens remains unexplored. To address these gaps, we propose to develop an integrated prognostic model based on zinc homeostasis network to provide insights into novel therapeutic approaches for LUAD.

In this study, zinc homeostasis network-associated hub genes were identified and analyzed in LUAD, and their potential roles in tumor progression were elucidated through functional enrichment analysis. A model incorporating five signature genes was constructed and validated for predicting prognosis employing multivariate Cox regression and machine learning algorithms. We further assessed the model’s association with the tumor immune landscape and its potential in predicting responses to chemotherapy and targeted agents. Taken together, these findings offer a novel framework for personalized therapeutic decision-making and prognostic marker selection in LUAD, shedding new light on the role of zinc homeostasis network in disease progression.

2 Materials and methods

2.1 Data acquisition and analysis of zinc homeostasis network-related differentially expressed genes

RNA-seq datasets in TPM format, along with matched clinical data for 541 LUAD tumors and 59 adjacent normal tissues were retrieved from TCGA. Wilcoxon rank-sum tests were utilized to identify differentially expressed genes (DEGs), with thresholds set at false discovery rate (FDR) < 0.05 and |log₂ fold change| > 1.0, using a volcano plot for visualization. Zinc homeostasis network-related gene sets were sourced from GeneCards (https://www.genecards.org/) using the search term “zinc transport”, as these transporters represent the primary regulators of intracellular zinc homeostasis, and 124 zinc homeostasis network-related DEGs were determined via Venn diagram intersection. Protein-protein interaction (PPI) networks were constructed using STRING (http://string-db.org) and visualized with Cytoscape. To define a highly interconnected gene cluster suitable for functional enrichment analysis, the top 60 hub genes were selected using the Maximal Clique Centrality (MCC) algorithm within the CytoHubba plugin for Cytoscape.

2.2 Prognostic risk model construction and validation in LUAD

LUAD cases with overall survival <30 days were excluded to minimize immortal time bias. Sixteen genes linked to prognosis were identified in the univariate analysis (p < 0.05). Subsequent integration of LASSO regression and multivariate Cox analysis refined the selection to five genes demonstrating significant prognostic value. These genes were used to establish a zinc homeostasis network-related gene signature for risk stratification. The risk score was calculated as follows:

$$\text{Risk\hspace{0.17em}Score}=\text{Σ}(\text{expression\hspace{0.17em}level\hspace{0.17em}of\hspace{0.17em}each\hspace{0.17em}gene}\times \text{corresponding\hspace{0.17em}regression\hspace{0.17em}coefficient}).$$

Cases were assigned to high-risk (HR) and low-risk (LR) groups based on the median risk score. Kaplan-Meier (K-M) survival curve and time-dependent Receiver Operating Characteristic (ROC) curve analyses were conducted using the “survival” and “timeROC” R packages. Dimensionality reduction and visualization were performed using Principal Component Analysis (PCA) and t-distributed Stochastic Neighbor Embedding (t-SNE) with the “limma,” “Rtsne,” and “ggplot2” R packages. Model performance was further validated in three independent GEO cohorts: GSE26939, GSE72094, and GSE13213, along with corresponding clinicopathological features and survival outcomes derived from the GEO database.

2.3 Assessment of immune microenvironment

Lymphocyte infiltration in HR and LR groups was profiled using the “IOBR” R package (19), with group differences visualized via heatmap. Seven established deconvolution algorithms-including XCELL, TIMER, QUANTISEQ, MCP-COUNTER, EPIC, CIBERSORT-ABS, and CIBERSORT-were employed to comprehensively evaluate correlations of risk score with immune cell subpopulations in the tumor microenvironment (TME). Immune function across groups was systematically analyzed using the “ssGSEA” algorithm. TME characteristic scores, including immune score, stroma score, and ESTIMATE score, were determined with the “estimate,” “reshape2,” and “ggpubr” R packages. For immunotherapy response prediction, intergroup comparisons of TIDE Score, CAF Score, MDSC Score, and Exclusion Score were facilitated by the R packages “ggplot2,” “ggpubr,” and “tidyverse”.

2.4 Human tissues and cells

The LUAD tissue and adjacent lung tissue samples used in this study were all obtained from five primary patients who had signed informed consent forms, with samples collected from the 900th Hospital of the Joint Logistic Support Force of China. According to the Declaration of Helsinki, this study has been approved by the Ethics Committee of the 900th Hospital of the Joint Logistic Support Force of China (Approval No.: 2025-047). Human LUAD cell lines A549 and H1975 were purchased from the Cell Bank of the Shanghai Institute of Biological Sciences (Shanghai, China). All cells were cultured in a humidified incubator at 37°C with 5% CO₂. A549 cells were cultured in DMEM/F-12 medium containing 10% fetal bovine serum (FBS), and H1975 cells were cultured in RPMI 1640 medium containing 10% FBS.

2.5 Cell transfection

Cell transfection experiments were performed using lipid nanoparticle-based reagents (GA-RNA Transfection Kit and GA-DNA Transfection Kit, GeneAdv Co. Ltd, Suzhou, China). SLC16A3 expression was knocked down using two different types of siRNA, with the following sequences:

si-SLC16A3#1: GGAGCAUCAUCCAGGUCUATT,

si-SLC16A3#2: CCGUCAGUGUCUUCUUCAATT.

Cells transfected with nonspecific siRNA (si-con) served as the negative control. For EGR2 overexpression, the pCMV-C-3xFLAG-h-EGR2 plasmid and its corresponding empty vector control were used. Transfection efficiency was assessed by Western blot 48 hours after transfection.

2.6 Cellular function assays

For CCK-8 viability assays, cells were seeded in 96-well plates at 1,000 cells/well. After 72 hours incubation, cell viability was assessed by measuring the optical density at 450 nm using a microplate reader following incubation with CCK-8 reagent (Meilunbio, Dalian, China) for 1.5 hours at 37°C. For EdU proliferation assays, cells were plated in 24-well plates at 2×10⁴ cells/well and cultured for 72 hours. Cells were fixed with 4% paraformaldehyde for 30 min and permeabilized with Triton X-100 for 15 min before staining according to the EdU kit protocol (Beyotime, Shanghai, China), and the percentage of EdU positive nuclei was determined from three randomly selected fields under a fluorescence microscope. Cell migration was evaluated using “Transwell” chambers with 8-μm pores. Briefly, cells resuspended in serum-free medium were seeded into the upper chamber at 3.5×10⁴ cells/chamber, while the lower chamber contained 0.6 mL of 20% serum-supplemented medium. After 24 hours incubation, migrated cells on the lower membrane surface were fixed with 4% paraformaldehyde for 30 min, stained with 2.5% crystal violet for 20 min, and counted in three random fields per chamber under an inverted microscope. These functional assay protocols have been validated in previous studies (20, 21).

2.7Cisplatin sensitivity and apoptosis assays

The GSE213102, GSE222187, and GSE108214 datasets were downloaded from the GEO database, and SLC16A3 expression levels were extracted and compared between cisplatin-sensitive and cisplatin-resistant LUAD cell lines. For the in vitro drug sensitivity assay, A549 cells were transfected with si-con or si-SLC16A3#1. At 48 hours post-transfection, cells were seeded into 96-well plates at equal densities, and then treated with increasing concentrations of cisplatin (Meilunbio, Dalian) for an additional 24 hours. Cell viability was measured using the CCK-8 assay, and the IC50 value for cisplatin was calculated from the dose-response curve. To evaluate cisplatin-induced apoptosis, A549 and H1975 cells were transfected with si-con or SLC16A3 siRNA. For the TUNEL assay, cells were treated with 10 μM cisplatin for 24 hours at 48 hours after transfection, followed by staining using the TUNEL Apoptosis Detection Kit (Beyotime, Shanghai). Images were captured under a fluorescence microscope, and the percentage of TUNEL-positive cells was quantified. For Annexin V/PI flow cytometry, cells were similarly treated with 10 μM cisplatin for 24 hours at 48 hours post-transfection. Cells were collected, washed with PBS, and stained using the Annexin V-FITC/PI Apoptosis Detection Kit (Meilunbio, Dalian). After incubation at room temperature in the dark for 10–15 minutes, samples were analyzed by flow cytometry, and the proportions of apoptotic cells were determined using FlowJo software.

2.8 Statistical analysis

All bioinformatics analyses were conducted using the R (version 4.3.3). Functional assays included three independent biological replicates, with data expressed as mean ± SD. Statistical analysis used GraphPad Prism 7.0, with group differences assessed by one-way ANOVA or Student’s t-test. Statistical significance was defined as p < 0.05.

3 Results

3.1 Identification of zinc homeostasis network-related DEGs between normal and LUAD tissues

The overall workflow of this study is illustrated in Figure 1. Comparison of the transcriptomic data of LUAD and normal tissue pairs identified 2,749 DEGs, of which 1,868 were upregulated and 881 were downregulated. These are illustrated in the volcano plot (Figure 2A). Intersection analysis with zinc homeostasis network-related gene sets indicated 124 zinc homeostasis network-associated DEGs (Figure 2B). To pinpoint core regulators, a PPI network was generated, from which 60 hub genes were identified using topological analysis (Figure 2C). These hub genes included transcriptional regulators (e.g., PPARG, EGR2), metabolism-related genes (e.g., SLC16A3, LDHA), and transmembrane signaling molecules (e.g., CFTR, CAV1). The expression levels of hub genes in tumor versus normal tissues were visualized through heatmap analysis (Figure 2D).

Figure 1

Figure 1. Flowchart of the study.

Figure 2

Figure 2. Identification of zinc homeostasis network-related DEGs between normal and LUAD tissues. (A) Volcano plot showing the distribution patterns of DEGs between LUAD and normal tissues. (B) Venn analysis illustrating the intersection of DEGs with zinc homeostasis network-related gene sets. (C) PPI network of 60 hub genes. (D) Heatmap depicting the differential expression profiles of 60 hub genes between tumor and adjacent normal tissues.

3.2 GO and KEGG

To evaluate functional roles and signaling pathways associated with zinc homeostasis network-related DEGs, GO and KEGG enrichment analyses were performed on the 60 hub genes, incorporating both standard and log₂FC-weighted methods. GO enrichment covered BP, CC, and MF, revealing significant associations with key physiological functions such as hypoxia response, glucose homeostasis, lipid localization, among others (Supplementary Figures S1A, B). KEGG pathway analysis identified significant enrichment in oncogenic pathways, especially the PPAR and HIF-1 signaling pathways (Supplementary Figure S1C), suggesting potential molecular mechanisms by which zinc homeostasis network-related genes contribute to LUAD initiation and progression. These results were further supported by log₂FC-weighting (Supplementary Figure S1D).

3.3 Construction of zinc homeostasis network-related prognostic risk model

Univariate Cox analysis of 60 hub genes identified 16 genes significantly associated with LUAD prognosis (Figure 3A), including key regulatory genes such as PPARG, CAV1, and LDHA. LASSO regression (Figures 3B–D) followed by multivariate Cox analysis (Figure 3E) yielded a final prognostic model consisting of five risk genes: CAV1, KLF4, LDHA, EGR2, and SLC16A3. Patients were stratified into HR and LR groups based on the median risk score. The risk factor analysis demonstrated a significant inverse correlation between risk score and clinical outcomes in the TCGA-LUAD cohort (Figures 3F, G). K-M survival analysis revealed significantly elevated mortality in the high-risk group (Figure 3H), with HR = 2.29 (95% CI: 1.69-3.09). The predictive ability of the model at 1-, 3-, and 5-year intervals was supported by time-dependent ROC curves, and the C-index of the five-gene signature was 0.6766 in the TCGA cohort (Figure 3I). Moreover, the AUC values of our integrated model surpassed those of any individual risk gene (Supplementary Figures S2A–E). PCA and t-SNE analyses confirmed separation of the groups (Figures 3J, K). A sankey diagram of clinicopathological correlations indicated a prevalence of early-stage (I-II) tumors in LR cases (Figure 3L). Subgroup analysis showed marked positive relationships between risk score and higher pathological, T, and N stages (Figures 3M–O). Furthermore, HR cases had markedly higher proportions of advanced-stage tumors compared to LR patients (Figures 3P–R). A chord diagram illustrates a functional cooperative network among the five risk genes during LUAD progression (Figure 3S), suggesting their biological interactions may contribute to disease advancement.

Figure 3

Figure 3. Construction of zinc homeostasis network-related prognostic risk model. (A) Univariate Cox regression analysis to identify zinc homeostasis network-related prognostic gene. LASSO regression analysis (B-D) and multivariate Cox regression analysis (E). (F) Risk score distribution plot based on the median risk score. (G) Bar plot comparing mortality rates between HR and LR groups. (H) Kaplan-Meier survival curves (HR = 2.29, 95% CI: 1.69-3.09, log-rank p < 0.001). (I) ROC curves to evaluate predictive efficacy for 1-, 3-, and 5-year survival. PCA analysis (J) and t-SNE analysis (K). (L) Sankey diagram illustrating the relationship among risk groups, pathological stages, and survival status. Differential analysis of risk score across clinicopathological subgroups: Pathological stage (M), T stage (N), N stage (O). Composition of clinicopathologic subgroups in high- and LR groups: Pathological stage (P), T stage (Q), N stage (R). (S) Interaction network of risk genes in LUAD. (*p < 0.05, **p < 0.01, ***p < 0.001).

3.4 Construction of a prognostic nomogram model

According to Univariate Cox analysis, a significant positive correlation was observed between the risk score and unfavorable prognosis (Figure 4A). This association remained statistically significant after adjusting for potential confounders in multivariate analysis, supporting the role of the risk score as an independent prognostic indicator (Figure 4B). A nomogram integrating the risk score with clinicopathological features showed improved prognostic performance, with a C-index of 0.713 (Figure 4C). Calibration curves indicated excellent agreement between predicted and observed survival rates at 1, 3, and 5 years (Figure 4D). Time-dependent ROC analyses (Figures 4E–G) and decision curve analyses (Figures 4H–J) demonstrated that the nomogram outperformed individual clinical parameters in clinical net benefit across 1- to 5-year prediction periods, underscoring its potential utility for prognostic evaluation.

Figure 4

Figure 4. Construction of prognostic nomogram model. Univariate (A) and multivariate (B) Cox analyses assessing the prognostic significance of the risk score. Nomogram-based survival prediction for LUAD patients using the risk score (C). Calibration curve associated with the nomogram (D). ROC curves (E–G) and DCA (H–J) comparing the clinical benefit of the nomograms, risk score, and established clinical parameters at 1-, 3-, and 5-year follow-up intervals. (***p < 0.001).

3.5 External validation of the prognostic risk model

In the GSE26939 (Figure 5A), GSE72094 (Figure 5B), and GSE13213 (Figure 5C) datasets, the risk factor distribution, K-M survival curves, and ROC analyses consistently aligned with findings from the TCGA-LUAD cohort, demonstrating that patients in the HR group displayed significantly poorer prognoses. The PCA further supported the reliability of the risk stratification, confirming clear separation between high- and LR groups. In addition, the C-index values of the five-gene signature in the GSE26939, GSE72094, and GSE13213 cohorts were 0.6379, 0.6513, and 0.6441, respectively, indicating a relatively stable discriminative ability across independent validation sets.

Figure 5

Figure 5. External validation of the prognostic risk model. External validation of the prognostic model: (A) GSE26939, (B) GSE72094, and (C) GSE13213.

3.6 The potential molecular mechanism of the prognostic risk model

This study further delineated the biological functions and signaling pathways distinguishing high- and LR groups. GO analysis revealed that DEGs were primarily enriched in BP related to mitotic regulation and cell adhesion modulation. At the CC level, DEGs were concentrated in cell-substrate junction and focal adhesion. MF analysis further highlighted growth factor binding and MHC class II protein complex interactions as key features (Supplementary Figure S3A). KEGG pathway heatmap based on GSVA revealed preferential activation of pathways involving ATPase inhibitor, steroid hydroxylase, and related pathways in the LR group, whereas the HR group displayed enrichment in glucose catabolic and glycolytic process, as well as cadherin binding pathways (Supplementary Figure S3B). GSEA identified activation of the glycolysis and gluconeogenesis in the HR group, indicative of metabolic reprogramming (Supplementary Figure S3C), while the LR group showed selective activation of Th1/Th2 cell differentiation pathways (Supplementary Figure S3D). These integrative analyses provide mechanistic insights into the biological divergence between risk subtypes and offer a conceptual basis for identifying actionable therapeutic targets in LUAD.

3.7 Assessment of TME and immunotherapy

Heatmap visualization integrated with immune cell infiltration data derived from multiple algorithms revealed a higher abundance of effector immune cells, including B lymphocytes and various T cell subsets, within the TME of LR LUAD patients compared to their HR counterparts (Figures 6A-C). Immune function analysis using ssGSEA further indicated increased activity of pathways associated with the immune systems in the LR group, particularly in antigen-presenting dendritic cells (aDCs), Type II interferon response, T helper cells, neutrophils, and immature dendritic cells (iDCs) (Figure 6D). Furthermore, both ssGSEA and CIBERSORT analyses revealed strong associations between the identified risk genes’ expression levels and distinct immune signatures (Figures 6E, F). The LR group also showed significantly elevated stromal score, immune score, and estimate score (Figure 6G), consistent with a “hot tumor” phenotype and elevated immunotherapy responsiveness. The HR group showed increased expression of biomarkers associated with immune evasion and therapy resistance, including elevated TIDE, CAF, MDSC, and Exclusion score (Figures 6H-K), suggestive of inherent resistance to immune checkpoint blockade. These findings support the utility of the risk model in delineating immune landscape heterogeneity and predicting immunotherapeutic response in LUAD.

Figure 6

Figure 6. Assessment of TME and immunotherapy. (A) Overview of immune cell infiltration analysis across various algorithms. (B, C) Heatmap depicting differential infiltration levels of B lymphocytes and T lymphocytes between HR and LR groups. (D) Differential activation of immune functions between the groups. (E, F) Correlation analysis between independent risk genes and immune cell infiltration. (G) TME scores: stromal score, immune score, and estimate score. (H) TIDE score. (I) CAF score. (J) MDSC score. (K) Exclusion score. (*p < 0.05, **p < 0.01, ***p < 0.001).

3.8 Single-cell dissection of the immune microenvironment

Based on the previous immune infiltration analysis suggesting a heightened immunosuppressive status in the HR group, we performed scRNA-seq on 9 LUAD samples to further investigate the underlying mechanisms. UMAP dimensionality reduction identified and annotated the major cell types within the tumor microenvironment (Figure 7A), and illustrated their compositional proportions and abundance (Figure 7B). Heatmap analysis demonstrated unique marker gene expression profiles for each cell cluster (Figure 7C). The spatial expression patterns of canonical markers, including CD3D (T cells), KRT7 (epithelial cells), and CD68 (macrophages), among others, on UMAP plots further validated the reliability of cell annotations (Figure 7D). To investigate the applicability of our risk model at the single-cell level, epithelial cells were further stratified into HR and LR groups according to their risk scores (Figures 7E, F). Re-clustering of T/NK cells identified functional subsets such as CD4⁺ T cells, Tregs, and NK cells (Figure 7G). Cell communication analysis indicated that HR epithelial cells exhibited stronger interaction networks with Tregs compared to their low-risk counterparts (Figures 7H, I). Furthermore, macrophage subpopulation analysis revealed close interactions between HR epithelial cells and M2-type macrophages (Figures 7J-L). Collectively, these single-cell findings suggest that HR tumor cells may actively recruit and interact with both Tregs and M2 macrophages, thereby contributing to the establishment of an immunosuppressive microenvironment.

Figure 7

Figure 7. Single-cell dissection of the immune microenvironment. (A) UMAP projection showing major cell types in the TME. (B) Compositional proportions and abundance of identified cell types. (C) Heatmap of cell type-specific marker gene expression patterns. (D) UMAP visualization of canonical marker genes validating cell annotations. (E, F) Stratification of epithelial cells into HR and LR groups based on risk scores. (G) Re-clustering of T/NK cells identifies functional immune subsets. (H) Cell-cell communication networks between epithelial subgroups and T/NK cells. (I) Key ligand-receptor pairs mediating Treg interactions with high- and low-risk epithelial cells. (J) Re-clustering of macrophages identifies functional immune subsets. (K) Cell-cell communication networks between epithelial subgroups and macrophages. (L) Key ligand-receptor pairs mediating macrophages interactions with high- and low-risk epithelial cells.

3.9 Correlation analysis of risk score with immunomodulators and chemokines

Immunomodulators and chemokines serve as pivotal drivers of tumorigenesis and cancer progression by mediating immune evasion, remodeling the TME, and promoting metastatic dissemination. As these molecules represent critical therapeutic targets, this study systematically examined the relationship between risk stratification and these immune regulatory factors. Correlation analyses between the risk score and selected immunomodulators and chemokines were visualized via heatmap (Figure 8A), revealing significant associations. The risk score correlated with several immune checkpoints (e.g., CD274, CD96, CTLA4, CSF1R), immune activators (e.g., CD27, CD28, CD276, CD40LG), chemokines (e.g., CCL17, CCL19, CCL7, CXCL14), and chemokine receptors (e.g., CCR2, CCR4, CCR7). Further analysis showed positive correlations between the risk score and CXCL5, CCL7, CXCL8, PVR, and CD276, while negative correlations were observed with CXCL14, CXCL16, CD40LG, CD28, TNFRSF14, and CD48 (Figures 8B-L). These associations highlight the immunological distinctions between risk groups and offer a theoretical foundation for developing risk-adapted immunotherapeutic approaches in LUAD.

Figure 8

Figure 8. Correlation analysis of risk score with immunomodulators and chemokines. (A) Heatmap showing the correlation between risk score and immunomodulators and chemokines. (B–L) Scatter plots with regression lines illustrating associations between risk score and the expression levels of CXCL14, CXCL5, CXCL16, CCL7, CXCL8, CD40LG, PVR, CD276, CD28, TNFRSF14, and CD48.

3.10 Chemotherapeutic agent sensitivity profiling

Given the immune microenvironment characteristics of HR patients, particularly their pronounced resistance to immune checkpoint inhibitors, we therefore conducted drug sensitivity analyses to compare the half-maximal IC50 values of various chemotherapeutic agents across different risk subgroups. The objective was to assess differences in chemoresistance and guide treatment strategies for HR individuals. The results indicated that the HR group displayed significantly greater sensitivity to several agents, including Epothilone B, Cisplatin, Docetaxel, Doxorubicin, Etoposide, Gemcitabine, and Vinorelbine (Figures 9A-G). In comparison, Methotrexate was identified as a potentially effective chemotherapeutic option for LR patients (Figure 9H).

Figure 9

Figure 9. Chemotherapeutic agent sensitivity profiling. Sensitivity responses to individual chemotherapeutic agents are shown: (A) Epothilone B, (B) Cisplatin, (C) Docetaxel, (D) Doxorubicin, (E) Etoposide, (F) Gemcitabine, (G) Vinorelbine, and (H) Methotrexate.

3.11 Risk stratification guides targeted therapy strategies for LUAD

Mutational analysis revealed that HR patients showed elevated tumor mutational burden, with driver mutations primarily enriched in genes involved in the regulation of genomic stability, including TP53, TTN, MUC16, CSMD3, and RYR2 (Supplementary Figure S4A). Drug sensitivity profiling showed that LR patients were more responsive to the mTOR inhibitor rapamycin (Supplementary Figure S4B) and the EGFR inhibitor erlotinib (Supplementary Figure S4C). HR patients exhibited greater sensitivity to a range of agents, including multi-kinase inhibitors (sorafenib, sunitinib), BCR-ABL/PDGFR inhibitors (imatinib), ALK/MET inhibitors (TAE684, CH5424802), and epigenetic regulators (cyclopamine, FTI-277), suggesting that their aggressive disease phenotype may be linked to dysregulated kinase signaling and epigenetic alterations (Supplementary Figures S4D-J). Other compounds, such as VX-680, midostaurin, PF-562271, PHA-665752, and MG-132, also showed increased efficacy in this group (Supplementary Figures S4K-O).

3.12 In vitro functional validation of SLC16A3 and EGR2 in lung adenocarcinoma

Given that this model shows significant association with lung adenocarcinoma prognosis, the independent risk genes may critically regulate malignant tumor phenotypes. As the functional roles of the two representative signature genes SLC16A3 and EGR2 in lung adenocarcinoma progression remain incompletely elucidated, we selected them for expression and functional validation using a series of in vitro assays. Both qRT-PCR (Figures 10A, B) and Western blot (Figures 10C, D) analyses revealed significant upregulation of SLC16A3 and downregulation of EGR2 in tumor tissues compared to adjacent normal tissues, which was corroborated by immunohistochemistry (Figures 10E, F). To investigate their biological functions, we separately performed SLC16A3 knockdown and EGR2 overexpression in lung adenocarcinoma cells, with transfection efficiency confirmed by Western blotting (Figures 10G, H). CCK-8 (Figures 10I, J) and EdU (Figures 10K, L) assays demonstrated that SLC16A3 knockdown significantly inhibited proliferation in A549 and H1975 cells versus controls. EGR2 overexpression similarly suppressed proliferation (Figures 10M-P). Furthermore, “Transwell” migration assays showed that SLC16A3 knockdown reduced cellular migration (Figures 10Q, R), while EGR2 overexpression likewise inhibited migration (Figures 10S, T). These results demonstrate that SLC16A3 acts as an oncogene by promoting proliferation and migration, while EGR2 functions as a tumor suppressor. Both modulating these genes attenuated these malignant behaviors in vitro, providing functional validation that aligns with and strengthens the reliability of our prior bioinformatic analyses.

Figure 10

Figure 10. In Vitro Functional Validation of SLC16A3 and EGR2 in Lung Adenocarcinoma​. qRT-PCR and Western blotting were applied for testing the expression levels of SLC16A3 (A, C) and EGR2 (B, D) in lung adenocarcinoma tissues and adjacent normal tissues. Representative immunohistochemical staining images of SLC16A3 (E) and EGR2 (F) in LUAD tissues and adjacent normal tissues. The intervention efficiency of SLC16A3 knockdown (si-SLC16A3#1, si-SLC16A3#2) and EGR2 overexpression (EGR2-OE) was validated by Western blotting (G, H). CCK-8 assay measured cell proliferation after SLC16A3 depletion (I, J) or EGR2 overexpression (M, N). EdU assay evaluated the proliferative capacity of A549 and H1975 cells after SLC16A3 knockdown (K, L) or EGR2 overexpression (O, P). Cell migration were measured by “Transwell” assays (Q-T). (*p < 0.05, **p < 0.01, ***p < 0.001).

3.13 Downregulation of SLC16A3 reduces cisplatin sensitivity of LUAD cells

Prior drug sensitivity analysis indicated that the HR group may be more sensitive to cisplatin (Figure 9B). Given that cisplatin is a first-line chemotherapeutic agent for lung adenocarcinoma, we further investigated the relationship between the risk gene SLC16A3 and cisplatin sensitivity. Analysis of the GSE213102 and GSE222187 datasets revealed that SLC16A3 expression was significantly lower in multiple cisplatin-resistant lung adenocarcinoma cell lines compared to their corresponding parental cells (Figures 11A-D). Similarly, in the GSE108214 dataset, SLC16A3 expression was also higher in cisplatin-sensitive lung adenocarcinoma cell lines (Figure 11E). At the cellular level, knockdown of SLC16A3 in A549 cells resulted in a markedly increased IC50 for cisplatin, suggesting reduced cisplatin sensitivity (Figure 11F). Furthermore, in A549 and H1975 cells with SLC16A3 knockdown and treated with 10 μM cisplatin for 24 hours, both TUNEL staining and flow cytometric analysis demonstrated that SLC16A3 downregulation significantly attenuated cisplatin-induced apoptosis compared to the control group (Figures 11G–J).

Figure 11

Figure 11. Downregulation of SLC16A3 reduces cisplatin sensitivity of LUAD cells. SLC16A3 expression in cisplatin-sensitive versus resistant LUAD cell lines was assessed using GSE213102 (A–C), GSE222187 (D), and GSE108214 (E) datasets. (F) CCK-8 assay for cell viability and IC50 calculation in control and SLC16A3-knockdown A549 cells after 24-hour treatment with a cisplatin concentration gradient. Evaluation of apoptosis by TUNEL staining (G, H) and Annexin V/PI flow cytometry (I, J). (*p < 0.05, **p < 0.01, ***p < 0.001).

4 Discussion

The incidence of LUAD continues to rise, contributing to a growing global patient burden (22). Despite progress in prevention, early detection and treatment, the overall prognosis remains poor (23). Consequently, further exploration of novel and robust prognostic biomarkers is critically needed to improve patients’ overall survival. Zinc homeostasis network-related genes, as key regulators of zinc homeostasis, contribute to intracellular zinc imbalance through aberrant expression and are implicated in the malignant progression of lung cancer (17, 24). While most research has focused on the effects of individual dysregulated genes in tumor progression, more comprehensive models that consider multiple risk genes provide a more robust approach for investigating the collective roles of zinc homeostasis network-related genes in LUAD pathogenesis. Thus, identifying key molecular biomarkers related to zinc homeostasis network and understanding their involvement in disease progression is essential.

In this study, we systematically identified 60 zinc homeostasis network-related DEGs in LUAD, which form significant interaction networks and are predominantly enriched in oncogenic pathways such as PPAR and HIF-1 signaling. This implies that dysregulation of the zinc homeostasis network contributes to LUAD progression through metabolic reprogramming and hypoxia adaptation, consistent with known roles of zinc in energy metabolism and hypoxia (25–28). Through an integrated approach incorporating LASSO regression, univariate Cox analyses, and multivariate Cox analyses, a five-gene prognostic signature was established. Risk stratification demonstrated that HR patients had significantly reduced overall survival and higher mortality rates. The risk score was strongly correlated with advanced T stage, N stage, and pathological stage. The five-gene risk model demonstrated consistent discriminative ability and prognostic patterns across the TCGA-LUAD cohort and multiple external validation cohorts, supporting its stable and reproducible performance. Based on these findings, we propose this model as an auxiliary tool for prognostic evaluation and risk stratification. In clinical practice, quantifying the expression of these five genes in surgical or biopsy specimens generates an individualized risk score that estimates prognostic risk. Moreover, integrating this risk score with clinicopathological features into a nomogram improved the overall discriminative performance and clinical net benefit compared to using clinical variables alone. Our results indicate that the five-gene risk score complements clinical assessment by enabling finer prognostic stratification and providing a valuable reference for personalized treatment decisions, rather than serving as a stand-alone classification tool. GSEA further revealed that HR tumors exhibit features of metabolic reprogramming, including activated glycolysis and HIF-1 signaling, which are established drivers of malignancy and immune evasion in LUAD (29–31).

Significant evidence has shown an intimate relationship between zinc homeostasis network and tumor immunology (32, 33). Using multiple algorithms, our immune profiling revealed significant correlations between risk score and the infiltration of various immune cell populations, including B and T cells, along with macrophages. The LR group exhibited an immunologically active TME, characterized by elevated infiltration of APCs, such as aDCs and iDCs, together with a higher presence of tumor-infiltrating lymphocytes. In contrast, the HR group displayed distinct immunosuppressive traits, including elevated TIDE scores predicting resistance to immune checkpoint inhibitors, and significantly reduced lymphocyte infiltration. Our study further employed single-cell analysis to reveal more active intercellular communication networks between HR tumor cells and immunosuppressive cells, including Tregs and M2-type macrophages. This suggests that HR tumor cells may actively recruit and interact with these immunosuppressive populations. Extensive evidence indicates that the cellular composition and functional state of the TME significantly influence the efficacy of immunotherapy (34, 35). Specifically, Tregs can impair the anti-tumor function of effector T cells through direct contact or the secretion of inhibitory cytokines such as IL-10 and TGF-β (36), while M2-type macrophages promote angiogenesis and suppress T-cell activity by releasing mediators like VEGF and arginase (37). Recent studies have further confirmed that Tregs and M2-type macrophages together constitute key immunosuppressive hubs within the TME, which collectively attenuate the effector functions of infiltrating CD8⁺ T cells and other antitumor immune cells, thereby inhibiting tumor cell killing and limiting the efficacy of immune-based therapies (38–40). Therefore, the synergistic interaction between HR tumor cells and immunosuppressive cells is likely an important mechanism underlying enhanced immune escape and increased disease aggressiveness in HR patients. These insights highlight the importance of targeting immunosuppressive pathways in the development of novel therapeutic strategies for LUAD patients with high-risk profiles. Further analysis of immunoregulatory molecules revealed that the risk score was positively associated with the expression levels of CXCL5, CCL7, CXCL8, PVR, and CD276. Previous reports have shown that these molecules play critical roles in creating an immunosuppressive microenvironment through immune cell recruitment regulation and direct suppression of effector immune cell functions (30, 41–44). In comparison, immunostimulatory molecules such as CXCL14, CXCL16, CD40LG, CD28, and TNFRSF14 displayed negative correlations with risk score. These molecules promote anti-tumor immunity by recruiting effector cells such as NK cells and CD8+ T cells, activating antigen-presenting cells, and directly inhibiting tumor progression (45–50). Together, these findings offer novel insights into the mechanisms through which zinc homeostasis network-related DEGs regulate the immune microenvironment in LUAD.

Based on drug sensitivity profiling, we propose a risk-stratification-guided therapeutic approach. LR patients may benefit from targeted therapies including mTOR inhibitors (e.g., rapamycin) and EGFR inhibitors (e.g., erlotinib), which display greater efficacy in this subgroup. In contrast, HR patients demonstrate increased sensitivity to multi-target kinase inhibitors (e.g., sorafenib, sunitinib) and epigenetic regulators (e.g., cyclopamine). The HR group also shows increased responsiveness to conventional chemotherapeutic agents such as cisplatin and gemcitabine, likely due to compromised DNA repair mechanisms associated with genomic instability (51).

A body of evidence has indicated that the five genes in our model are all involved in the zinc homeostasis network (52–56). In LUAD, they function as a central hub within this network, and their synergistic dysregulation drives disease progression by impacting metabolic reprogramming and immune microenvironment remodeling. Among the five risk genes identified, the functions of LDHA, KLF4, and CAV1 in the onset and progression of LUAD have been relatively well studied (57–59). In contrast, the roles of SLC16A3 and EGR2 in LUAD are relatively limited. Available evidence indicates that SLC16A3, as a member of the monocarboxylate transporter family, promotes immune evasion and metastasis in various tumors by mediating lactate efflux to maintain tumor microenvironment acidification (60, 61). EGR2 is a transcription factor that can restrict tumor growth by regulating tumor suppressor pathways like PTEN (62), but its precise function in LUAD is unknown. To validate their biological functions, we performed systematic in vitro assays. We found that SLC16A3 was significantly upregulated and EGR2 was downregulated in LUAD tissues at both mRNA and protein levels. Further functional experiments revealed that knocking down SLC16A3 or overexpressing EGR2 could significantly inhibit the proliferation and migration of LUAD cells. These results confirm that SLC16A3 acts as an oncogene and EGR2 as a tumor suppressor in LUAD, aligning with their identification as key risk genes in our bioinformatic analysis. Moreover, we further discovered that the risk gene SLC16A3 may be involved in regulating the response of LUAD cells to cisplatin. Analysis of public datasets revealed that SLC16A3 expression is higher in cisplatin-sensitive lung adenocarcinoma cells. Moreover, knocking down SLC16A3 in lung adenocarcinoma cells attenuated cisplatin-induced apoptosis. These findings align with the trend observed in our drug sensitivity analysis, where the high-risk group appeared to derive greater benefit from cisplatin treatment, suggesting the potential utility of this five-gene risk model in guiding adjuvant treatment strategy selection.

This study has several strengths. We conducted a comprehensive investigation of zinc homeostasis-associated genes in LUAD and developed a five-gene risk model. The model’s risk score was associated with aggressive clinicopathological features and exhibited consistent prognostic performance across multiple cohorts. Furthermore, because the model includes only five genes, it can be easily combined with clinical characteristics in a nomogram to improve overall predictive performance and clinical net benefit. The signature genes are implicated in oncogenic pathways and immune regulation, suggesting their potential utility in predicting immunotherapy response and guiding risk-based precision therapy. Finally, experimental validation of SLC16A3 and EGR2 underscores the biological relevance of our model. Nevertheless, certain limitations should be acknowledged. The overall predictive accuracy of the model, as indicated by the AUC and C-index, was moderate. This may stem from the inherent biological and clinical heterogeneity of LUAD, the retrospective nature of the public cohorts, and potential sample selection biases. Thus, validation in larger, prospective, multicenter cohorts is still required. Furthermore, functional validation was performed exclusively in vitro and focused on two key genes, while CAV1, KLF4, and LDHA were not functionally examined in this study. Elucidating the functions of these genes is a priority for future research to fully uncover the model’s link to tumor progression. Additionally, while our analyses suggest that the signature influences the tumor immune microenvironment via regulatory T cells and M2 macrophages, the precise molecular mechanisms require further investigation.

5 Conclusion

A prognostic risk model for LUAD was established based on zinc homeostasis network-related genes, demonstrating significant correlations with clinical stage and with patient survival. HR patients show poorer prognoses and features of an immunosuppressive microenvironment, whereas LR patients display a “hot tumor” phenotype associated with greater immunotherapy responsiveness. Furthermore, risk stratification guides potential therapeutic strategies, revealing differential sensitivities to chemotherapeutic and targeted agents. This model bridges the zinc homeostasis network with actionable precision therapy and may have important implications for personalized treatment and risk assessment in patients, with potential for clinical translation.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

Ethics statement

The studies involving humans were approved by the Ethics Committee of the 900th Hospital of the Joint Logistic Support Force. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

FC: Writing – review & editing, Investigation, Writing – original draft, Formal analysis, Visualization, Conceptualization, Data curation, Methodology. JX: Writing – review & editing, Methodology, Formal analysis, Data curation, Investigation, Visualization, Writing – original draft, Conceptualization. WZ: Writing – review & editing, Software, Resources, Data curation, Validation. XZ: Writing – review & editing, Data curation, Validation, Resources. YC: Resources, Project administration, Writing – review & editing, Supervision. NZ: Writing – review & editing, Formal analysis. CL: Formal Analysis, Writing – review & editing. ZY: Conceptualization, Funding acquisition, Writing – review & editing, Supervision.

Funding

The author(s) declared financial support was received for this work and/or its publication. This research was supported by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0529400 & 2024ZD0529406), the Fujian Medical University: Set sail Fund Project (2022QH1333) and the External Cooperation of Science and Technology Program of Fujian Province (202210034).

Acknowledgments

We extend our gratitude to all the authors involved in this study.

Conflict of interest

The authors declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu.2025.1691179/full#supplementary-material

Supplementary Figure 1 | GO (A, B) and KEGG pathway (C) enrichment analysis. (D) LogFC-weighted enrichment validation.

Supplementary Figure 2 | Time-dependent ROC curves of (A) CAV1, (B) EGR2, (C) KLF4, (D) LDHA, and (E) SLC16A3 for predicting 1-, 3-, and 5-year overall survival.

Supplementary Figure 3 | (A) GO analysis. (B) GSVA and GSEA (C, D) between HR and LR groups.

Supplementary Figure 4 | (A) Mutation site analysis in high- and LR patient groups. Comparative assessment of therapeutic response to targeted agents between the two patient cohorts: (B) Rapamycin, (C) Erlotinib, (D) Sorafenib, (E) Sunitinib, (F) Imatinib, (G) TAE684, (H) CH5424802, (I) Cyclopamine, (J) FTI-277, (K) VX-680, (L) Midostaurin, (M) PF-562271, (N) PHA-665752, (O) MG-132.

References

1. Denisenko TV, Budkevich IN, and Zhivotovsky B. Cell death-based treatment of lung adenocarcinoma. Cell Death Dis. (2018) 9:117. doi: 10.1038/s41419-017-0063-y

PubMed Abstract | Crossref Full Text | Google Scholar

2. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. (2021) 71:209–49. doi: 10.3322/caac.21660

PubMed Abstract | Crossref Full Text | Google Scholar

3. Siegel RL, Miller KD, Fuchs HE, and Jemal A. Cancer statistics, 2021. CA Cancer J Clin. (2021) 71:7–33. doi: 10.3322/caac.21654

PubMed Abstract | Crossref Full Text | Google Scholar

4. Lin JJ, Cardarella S, Lydon CA, Dahlberg SE, Jackman DM, Jänne PA, et al. Five-year survival in EGFR-mutant metastatic lung adenocarcinoma treated with EGFR-TKIs. J Thorac Oncol. (2016) 11:556–65. doi: 10.1016/j.jtho.2015.12.103

PubMed Abstract | Crossref Full Text | Google Scholar

5. Yamasaki S, Sakata-Sogawa K, Hasegawa A, Suzuki T, Kabu K, Sato E, et al. Zinc is a novel intracellular second messenger. J Cell Biol. (2007) 177:637–45. doi: 10.1083/jcb.200702081

PubMed Abstract | Crossref Full Text | Google Scholar

6. Prasad AS. Discovery of human zinc deficiency: its impact on human health and disease. Adv Nutr. (2013) 4:176–90. doi: 10.3945/an.112.003210

PubMed Abstract | Crossref Full Text | Google Scholar

7. Roohani N, Hurrell R, Kelishadi R, and Schulin R. Zinc and its importance for human health: An integrative review. J Res Med Sci. (2013) 18:144–57.

PubMed Abstract | Google Scholar

8. Chasapis CT, Loutsidou AC, Spiliopoulou CA, and Stefanidou ME. Zinc and human health: an update. Arch Toxicol. (2012) 86:521–34. doi: 10.1007/s00204-011-0775-1

PubMed Abstract | Crossref Full Text | Google Scholar

9. Bafaro E, Liu Y, Xu Y, and Dempski RE. The emerging role of zinc transporters in cellular homeostasis and cancer. Signal Transduct Target Ther. (2017) 2:17029–. doi: 10.1038/sigtrans.2017.29

PubMed Abstract | Crossref Full Text | Google Scholar

10. Gumulec J, Masarik M, Adam V, Eckschlager T, Provaznik I, and Kizek R. Serum and tissue zinc in epithelial Malignancies: a meta-analysis. PloS One. (2014) 9:e99790. doi: 10.1371/journal.pone.0099790

PubMed Abstract | Crossref Full Text | Google Scholar

11. Kim EK and Choi EJ. Pathological roles of MAPK signaling pathways in human diseases. Biochim Biophys Acta. (2010) 1802:396–405. doi: 10.1016/j.bbadis.2009.12.009

PubMed Abstract | Crossref Full Text | Google Scholar

12. Bismar TA, Yoshimoto M, Vollmer RT, Duan Q, Firszt M, Corcos J, et al. PTEN genomic deletion is an early event associated with ERG gene rearrangements in prostate cancer. BJU Int. (2011) 107:477–85. doi: 10.1111/j.1464-410X.2010.09470.x

PubMed Abstract | Crossref Full Text | Google Scholar

13. Krohn A, Freudenthaler F, Harasimowicz S, Kluth M, Fuchs S, Burkhardt L, et al. Heterogeneity and chronology of PTEN deletion and ERG fusion in prostate cancer. Mod Pathol. (2014) 27:1612–20. doi: 10.1038/modpathol.2014.70

PubMed Abstract | Crossref Full Text | Google Scholar

14. Hojyo S, Fukada T, Shimoda S, Ohashi W, Bin BH, Koseki H, et al. The zinc transporter SLC39A14/ZIP14 controls G-protein coupled receptor-mediated signaling required for systemic growth. PloS One. (2011) 6:e18059. doi: 10.1371/journal.pone.0018059

PubMed Abstract | Crossref Full Text | Google Scholar

15. Yang J, Zhang Z, Zhang Y, Ni X, Zhang G, Cui X, et al. ZIP4 promotes muscle wasting and cachexia in mice with orthotopic pancreatic tumors by stimulating RAB27B-regulated release of extracellular vesicles from cancer cells. Gastroenterology. (2019) 156:722–34.e6. doi: 10.1053/j.gastro.2018.10.026

PubMed Abstract | Crossref Full Text | Google Scholar

16. Franklin RB, Feng P, Milon B, Desouki MM, Singh KK, Kajdacsy-Balla A, et al. hZIP1 zinc uptake transporter down regulation and zinc depletion in prostate cancer. Mol Cancer. (2005) 4:32. doi: 10.1186/1476-4598-4-32

PubMed Abstract | Crossref Full Text | Google Scholar

17. Wu DM, Liu T, Deng SH, Han R, and Xu Y. SLC39A4 expression is associated with enhanced cell migration, cisplatin resistance, and poor survival in non-small cell lung cancer. Sci Rep. (2017) 7:7211. doi: 10.1038/s41598-017-07830-4

PubMed Abstract | Crossref Full Text | Google Scholar

18. Liu Z, Hu Z, Cai X, and Liu S. SLC39A5 promotes lung adenocarcinoma cell proliferation by activating PI3K/AKT signaling. Pathol Res Pract. (2021) 224:153541. doi: 10.1016/j.prp.2021.153541

PubMed Abstract | Crossref Full Text | Google Scholar

19. Zeng D, Ye Z, Shen R, Yu G, Wu J, Xiong Y, et al. IOBR: multi-omics immuno-oncology biological research to decode tumor microenvironment and signatures. Front Immunol. (2021) 12:687975. doi: 10.3389/fimmu.2021.687975

PubMed Abstract | Crossref Full Text | Google Scholar

20. Zhang C, Yu R, Li S, Yuan M, Hu T, Liu J, et al. KRAS mutation increases histone H3 lysine 9 lactylation (H3K9la) to promote colorectal cancer progression by facilitating cholesterol transporter GRAMD1A expression. Cell Death Differ. (2025) 32:2225-38. doi: 10.1038/s41418-025-01533-4

PubMed Abstract | Crossref Full Text | Google Scholar

21. Liu J, Su Y, Zhang C, Dong H, Yu R, Yang X, et al. NCOA3 impairs the efficacy of anti-PD-L1 therapy via HSP90α/EZH2/CXCL9 axis in colon cancer. Int Immunopharmacol. (2025) 155:114579. doi: 10.1016/j.intimp.2025.114579

PubMed Abstract | Crossref Full Text | Google Scholar

22. Zhang L, Wang S, and Wang L. Comprehensive analysis identifies YKT6 as a potential prognostic and diagnostic biomarker in lung adenocarcinoma. BMC Cancer. (2024) 24:1235. doi: 10.1186/s12885-024-12975-3

PubMed Abstract | Crossref Full Text | Google Scholar

23. Ziora P, Skiba H, Kiczmer P, Zaboklicka N, Wypyszyńska J, Stachura M, et al. Ten-year observational study of patients with lung adenocarcinoma: clinical outcomes, prognostic factors, and five-year survival rates. J Clin Med. (2025) 14:2552. doi: 10.3390/jcm14082552

PubMed Abstract | Crossref Full Text | Google Scholar

24. Ni C, Lou X, Yao X, Wang L, Wan J, Duan X, et al. ZIP1(+) fibroblasts protect lung cancer against chemotherapy via connexin-43 mediated intercellular Zn(2+) transfer. Nat Commun. (2022) 13:5919. doi: 10.1038/s41467-022-33521-4

PubMed Abstract | Crossref Full Text | Google Scholar

25. Luo J, Shang Y, Zhao N, Lu X, Wang Z, Li X, et al. Hypoxia-responsive micelles deprive cofactor of stearoyl-CoA desaturase-1 and sensitize ferroptotic ovarian cancer therapy. Biomaterials. (2025) 314:122820. doi: 10.1016/j.biomaterials.2024.122820

PubMed Abstract | Crossref Full Text | Google Scholar

26. Morand J, Briançon-Marjollet A, Lemarie E, Gonthier B, Arnaud J, Korichneva I, et al. Zinc deficiency promotes endothelin secretion and endothelial cell migration through nuclear hypoxia-inducible factor-1 translocation. Am J Physiol Cell Physiol. (2019) 317:C270–c6. doi: 10.1152/ajpcell.00460.2018

PubMed Abstract | Crossref Full Text | Google Scholar

27. Costello LC and Franklin RB. The clinical relevance of the metabolism of prostate cancer; zinc and tumor suppression: connecting the dots. Mol Cancer. (2006) 5:17. doi: 10.1186/1476-4598-5-17

PubMed Abstract | Crossref Full Text | Google Scholar

28. Zhao L, Oliver E, Maratou K, Atanur SS, Dubois OD, Cotroneo E, et al. The zinc transporter ZIP12 regulates the pulmonary vascular response to chronic hypoxia. Nature. (2015) 524:356–60. doi: 10.1038/nature14620

PubMed Abstract | Crossref Full Text | Google Scholar

29. Ge W, Wang Y, Quan M, Mao T, Bischof EY, Xu H, et al. Activation of the PI3K/AKT signaling pathway by ARNTL2 enhances cellular glycolysis and sensitizes pancreatic adenocarcinoma to erlotinib. Mol Cancer. (2024) 23:48. doi: 10.1186/s12943-024-01965-5

PubMed Abstract | Crossref Full Text | Google Scholar

30. Lin C, He H, Liu H, Li R, Chen Y, Qi Y, et al. Tumour-associated macrophages-derived CXCL8 determines immune evasion through autonomous PD-L1 expression in gastric cancer. Gut. (2019) 68:1764–73. doi: 10.1136/gutjnl-2018-316324

PubMed Abstract | Crossref Full Text | Google Scholar

31. Kim S, Jang JY, Koh J, Kwon D, Kim YA, Paeng JC, et al. Programmed cell death ligand-1-mediated enhancement of hexokinase 2 expression is inversely related to T-cell effector gene expression in non-small-cell lung cancer. J Exp Clin Cancer Res. (2019) 38:462. doi: 10.1186/s13046-019-1407-5

PubMed Abstract | Crossref Full Text | Google Scholar

32. John E, Laskow TC, Buchser WJ, Pitt BR, Basse PH, Butterfield LH, et al. Zinc in innate and adaptive tumor immunity. J Transl Med. (2010) 8:118. doi: 10.1186/1479-5876-8-118

PubMed Abstract | Crossref Full Text | Google Scholar

33. Gao H, Zhao L, Wang H, Xie E, Wang X, Wu Q, et al. Metal transporter Slc39a10 regulates susceptibility to inflammatory stimuli by controlling macrophage survival. Proc Natl Acad Sci U S A. (2017) 114:12940–5. doi: 10.1073/pnas.1708018114

PubMed Abstract | Crossref Full Text | Google Scholar

34. Lu Y, Liu Y, Zuo X, Li G, Wang J, Liu J, et al. CXCL12(+) tumor-associated endothelial cells promote immune resistance in hepatocellular carcinoma. J Hepatol. (2025) 82:634–48. doi: 10.1016/j.jhep.2024.09.044

PubMed Abstract | Crossref Full Text | Google Scholar

35. Ke H, Li P, Li Z, Zeng X, Zhang C, Luo S, et al. Immune profiling of the macroenvironment in colorectal cancer unveils systemic dysfunction and plasticity of immune cells. Clin Transl Med. (2025) 15:e70175. doi: 10.1002/ctm2.70175

PubMed Abstract | Crossref Full Text | Google Scholar

36. Shiri AM, Zhang T, Bedke T, Zazara DE, Zhao L, Lücke J, et al. IL-10 dampens antitumor immunity and promotes liver metastasis via PD-L1 induction. J Hepatol. (2024) 80:634–44. doi: 10.1016/j.jhep.2023.12.015

PubMed Abstract | Crossref Full Text | Google Scholar

37. Lu Y, Han G, Zhang Y, Zhang L, Li Z, Wang Q, et al. M2 macrophage-secreted exosomes promote metastasis and increase vascular permeability in hepatocellular carcinoma. Cell Commun Signal. (2023) 21:299. doi: 10.1186/s12964-022-00872-w

PubMed Abstract | Crossref Full Text | Google Scholar

38. Su W, Han HH, Wang Y, Zhang B, Zhou B, Cheng Y, et al. The polycomb repressor complex 1 drives double-negative prostate cancer metastasis by coordinating stemness and immune suppression. Cancer Cell. (2019) 36:139–55.e10. doi: 10.1016/j.ccell.2019.06.009

PubMed Abstract | Crossref Full Text | Google Scholar

39. Jiang J, Chen H, Zhao C, Li T, Zhang C, Ma L, et al. PRTN3 promotes IL33/Treg-mediated tumor immunosuppression by enhancing the M2 polarization of tumor-associated macrophages in lung adenocarcinoma. Cancer Lett. (2025) 616:217584. doi: 10.1016/j.canlet.2025.217584

PubMed Abstract | Crossref Full Text | Google Scholar

40. Thommen DS and Schumacher TN. T cell dysfunction in cancer. Cancer Cell. (2018) 33:547–62. doi: 10.1016/j.ccell.2018.03.012

PubMed Abstract | Crossref Full Text | Google Scholar

41. Zhang W, Wang H, Sun M, Deng X, Wu X, Ma Y, et al. CXCL5/CXCR2 axis in tumor microenvironment as potential diagnostic biomarker and therapeutic target. Cancer Commun (Lond). (2020) 40:69–80. doi: 10.1002/cac2.12010

PubMed Abstract | Crossref Full Text | Google Scholar

42. Takacs GP, Kreiger CJ, Luo D, Tian G, Garcia JS, Deleyrolle LP, et al. Glioma-derived CCL2 and CCL7 mediate migration of immune suppressive CCR2(+)/CX3CR1(+) M-MDSCs into the tumor microenvironment in a redundant manner. Front Immunol. (2022) 13:993444. doi: 10.3389/fimmu.2022.993444

PubMed Abstract | Crossref Full Text | Google Scholar

43. Kučan Brlić P, Lenac Roviš T, Cinamon G, Tsukerman P, Mandelboim O, and Jonjić S. Targeting PVR (CD155) and its receptors in anti-tumor therapy. Cell Mol Immunol. (2019) 16:40–52. doi: 10.1038/s41423-018-0168-y

PubMed Abstract | Crossref Full Text | Google Scholar

44. Liu HJ, Du H, Khabibullin D, Zarei M, Wei K, Freeman GJ, et al. mTORC1 upregulates B7-H3/CD276 to inhibit antitumor T cells and drive tumor immune evasion. Nat Commun. (2023) 14:1214. doi: 10.1038/s41467-023-36881-7

PubMed Abstract | Crossref Full Text | Google Scholar

45. Westrich JA, Vermeer DW, Silva A, Bonney S, Berger JN, Cicchini L, et al. CXCL14 suppresses human papillomavirus-associated head and neck cancer through antigen-specific CD8(+) T-cell responses by upregulating MHC-I expression. Oncogene. (2019) 38:7166–80. doi: 10.1038/s41388-019-0911-6

PubMed Abstract | Crossref Full Text | Google Scholar

46. Parikh A, Shin J, Faquin W, Lin DT, Tirosh I, Sunwoo JB, et al. Malignant cell-specific CXCL14 promotes tumor lymphocyte infiltration in oral cavity squamous cell carcinoma. J Immunother Cancer. (2020) 8:e001048. doi: 10.1136/jitc-2020-001048

PubMed Abstract | Crossref Full Text | Google Scholar

47. Jung HY, Lee DK, Lee M, Choi SH, Park JD, Ko ES, et al. ELK3-CXCL16 axis determines natural killer cell cytotoxicity via the chemotactic activity of CXCL16 in triple negative breast cancer. Oncoimmunology. (2023) 12:2190671. doi: 10.1080/2162402X.2023.2190671

PubMed Abstract | Crossref Full Text | Google Scholar

48. Tang T, Cheng X, Truong B, Sun L, Yang X, and Wang H. Molecular basis and therapeutic implications of CD40/CD40L immune checkpoint. Pharmacol Ther. (2021) 219:107709. doi: 10.1016/j.pharmthera.2020.107709

PubMed Abstract | Crossref Full Text | Google Scholar

49. Holling GA, Chavel CA, Sharda AP, Lieberman MM, James CM, Lightman SM, et al. CD8+ T cell metabolic flexibility elicited by CD28-ARS2 axis-driven alternative splicing of PKM supports antitumor immunity. Cell Mol Immunol. (2024) 21:260–74. doi: 10.1038/s41423-024-01124-2

PubMed Abstract | Crossref Full Text | Google Scholar

50. Zhu YD and Lu MY. Increased expression of TNFRSF14 indicates good prognosis and inhibits bladder cancer proliferation by promoting apoptosis. Mol Med Rep. (2018) 18:3403–10. doi: 10.3892/mmr.2018.9306

PubMed Abstract | Crossref Full Text | Google Scholar

51. Dasari S and Tchounwou PB. Cisplatin in cancer therapy: molecular mechanisms of action. Eur J Pharmacol. (2014) 740:364–78. doi: 10.1016/j.ejphar.2014.07.025

PubMed Abstract | Crossref Full Text | Google Scholar

52. Joseph LJ, Le Beau MM, Jamieson GA Jr., Acharya S, Shows TB, Rowley JD, et al. Molecular cloning, sequencing, and mapping of EGR2, a human early growth response gene encoding a protein with “zinc-binding finger” structure. Proc Natl Acad Sci U S A. (1988) 85:7164–8. doi: 10.1073/pnas.85.19.7164

PubMed Abstract | Crossref Full Text | Google Scholar

53. Pramchu-em C, Meksawan K, and Chanvorachote P. Zinc sensitizes lung cancer cells to anoikis through down-regulation of akt and caveolin-1. Nutr Cancer. (2016) 68:312–9. doi: 10.1080/01635581.2016.1142582

PubMed Abstract | Crossref Full Text | Google Scholar

54. Ikegami N, Szegedi I, Kiss C, and Petrás M. Regulatory role of zinc in acute promyelocytic leukemia: cellular and molecular aspects with therapeutic implications. Int J Mol Sci. (2025) 26:9685. doi: 10.3390/ijms26199685

PubMed Abstract | Crossref Full Text | Google Scholar

55. Michaelis V, Kasper S, Naperkowski L, Pusse J, Thiel A, Ebert F, et al. The impact of zinc on manganese bioavailability and cytotoxicity in hepG2 cells. Mol Nutr Food Res. (2023) 67:e2200283. doi: 10.1002/mnfr.202200283

PubMed Abstract | Crossref Full Text | Google Scholar

56. Sasaki S, Kobayashi M, Futagi Y, Ogura J, Yamaguchi H, and Iseki K. Involvement of histidine residue his382 in pH regulation of MCT4 activity. PloS One. (2014) 10:e0122738. doi: 10.1371/journal.pone.0122738

PubMed Abstract | Crossref Full Text | Google Scholar

57. Chen Y, Yan H, Yan L, Wang X, Che X, Hou K, et al. Hypoxia-induced ALDH3A1 promotes the proliferation of non-small-cell lung cancer by regulating energy metabolism reprogramming. Cell Death Dis. (2023) 14:617. doi: 10.1038/s41419-023-06142-y

PubMed Abstract | Crossref Full Text | Google Scholar

58. Zhou H, Guan Q, Hou X, Liu L, Zhou L, Li W, et al. Epithelial-mesenchymal reprogramming by KLF4-regulated Rictor expression contributes to metastasis of non-small cell lung cancer cells. Int J Biol Sci. (2022) 18:4869–83. doi: 10.7150/ijbs.73548

PubMed Abstract | Crossref Full Text | Google Scholar

59. Lin Q, Zong S, Wang Y, Zhou Y, Wang K, Shi F, et al. Breast cancer-derived CAV1 promotes lung metastasis by regulating integrin α6β4 and the recruitment and polarization of tumor-associated neutrophils. Int J Biol Sci. (2024) 20:5695–714. doi: 10.7150/ijbs.94153

PubMed Abstract | Crossref Full Text | Google Scholar

60. Yu T, Liu Z, Tao Q, Xu X, Li X, Li Y, et al. Targeting tumor-intrinsic SLC16A3 to enhance anti-PD-1 efficacy via tumor immune microenvironment reprogramming. Cancer Lett. (2024) 589:216824. doi: 10.1016/j.canlet.2024.216824

PubMed Abstract | Crossref Full Text | Google Scholar

61. Qian Y, Galan-Cobo A, Guijarro I, Dang M, Molkentine D, Poteete A, et al. MCT4-dependent lactate secretion suppresses antitumor immunity in LKB1-deficient lung adenocarcinoma. Cancer Cell. (2023) 41:1363–80.e7. doi: 10.1016/j.ccell.2023.05.015

PubMed Abstract | Crossref Full Text | Google Scholar

62. Unoki M and Nakamura Y. Growth-suppressive effects of BPOZ and EGR2, two genes involved in the PTEN signaling pathway. Oncogene. (2001) 20:4457–65. doi: 10.1038/sj.onc.1204608

PubMed Abstract | Crossref Full Text | Google Scholar