An immune-resistance dual-axis framework prioritizes MAL2 for functional investigation in lung cancer

Background Immune checkpoint blockade has reshaped lung cancer treatment, yet durable benefit remains restricted to a subset of patients. Biomarkers incompletely capture interactions between immune activation and tumor-intrinsic resistance. Integrating immune and epithelial expression programs may help identify response-associated states and candidates for functional investigation. Methods Public bulk transcriptomes were analyzed with format-specific preprocessing, empirical Bayes differential expression, standardized cross-cohort evidence, and gene-set enrichment. Five discovery-informed modules formed an immune-resistance score. Conditional leave-one-out and leave-cohort-out analyses refitted normalization within training data and compared the fixed score with a nested ridge-logistic model. Survival, composition-adjustment, and prioritization sensitivity analyses were performed. Single-cell counts underwent quality control, unsupervised clustering, and patient-level pseudobulk analysis. MAL2 was investigated in A549 and H1299 cells using knockdown, growth, motility, apoptosis, HLA-ABC, cytokine, and PBMC co-culture assays. Results No individual gene met FDR <0.05 in any of the three discovery contrasts. Favorable-response genes were enriched for adaptive immune processes, whereas the unfavorable list showed weaker DNA-replication enrichment. Conditional leave-cohort-out AUCs for the dual-axis score were 0.964, 0.859, and 0.773; performance did not consistently exceed the immune axis. In GSE190266, outcome analysis yielded a cohort-relative AUC of 0.617 with incomplete module coverage. Higher scores were associated with longer PFS in two evaluated cohorts. Quality control retained 91,762 single cells. MAL2 remained epithelial-enriched and ranked third among 58 evaluable candidates, but its patient-level epithelial response contrast and standalone PFS associations were not significant. Knockdown reduced growth and motility-related readouts, increased apoptosis and HLA-ABC expression, and altered cytokine and co-culture activation readouts. Conclusion The immune-resistance framework links cohort-level response patterns to epithelial candidates for functional investigation. MAL2 depletion alters growth and immune-related phenotypes in lung cancer cells, supporting further investigation of its biological role.