Hypoxia-anoikis-related genes in LUAD: machine learning and RNA sequencing analysis of immune infiltration and therapy response.

Abstract

Hypoxia plays a crucial role in the pathogenesis of various cancers, especially lung adenocarcinoma (LUAD), by altering cancer metabolism to promote escape mechanisms. Anoikis, a specialized form of programmed cell death, is evaded by LUAD cells during tumor progression and metastasis through upregulation of anti-apoptotic proteins. Investigating the impact of hypoxia-anoikis-related genes on prognosis and therapy prediction in LUAD is essential. Gene expression and clinical data from 489 LUAD patients and 49 normal tissues in The Cancer Genome Atlas (TCGA) dataset were used as the training set, while GSE72094, GSE31210, and GSE30219 datasets were used for validation. Weighted Gene Co-Expression Network Analysis (WGCNA) identified genes associated with hypoxia and anoikis. Machine learning models were evaluated using the C-index. Kaplan-Meier survival analysis, immune cell infiltration, tumor mutational burden (TMB), and sensitivity to therapy were assessed based on risk scores. A total of 21 hypoxia-anoikis-related prognostic genes were identified. The Random Survival Forest (RSF) model had the highest C-index. High-risk patients had significantly lower survival rates. Immune analysis showed higher immune infiltration in the low-risk group, with lower immune escape potential in these patients. Risk scores were correlated with sensitivity to targeted therapy and chemotherapy. MCF2 was identified as a key prognostic gene, and its knockdown inhibited LUAD cell proliferation and metastasis. These 21 genes offer insights into LUAD prognosis and therapy response, guiding personalized treatment strategies for LUAD patients.