Frontiers in Genetics · Published 2026-07-12 · DOI 10.3389/fgene.2026.1831053
Ya Yang, Tingyuan Fan, Xiaye Miao, Chuanli Ren
BackgroundTumor-associated inflammation is an important contributor to cancer progression and therapeutic resistance. However, the molecular regulators linking inflammatory signaling with tumor biology in lung adenocarcinoma (LUAD) remain incompletely defined.MethodsWe implemented an integrative multi-cohort and machine learning framework to identify inflammation-associated candidate regulators in LUAD. Immune-inflammatory genes derived from hallmark pathways were intersected with differentially expressed and prognosis-associated genes in TCGA-LUAD. Transcriptomic and clinical data from TCGA and 13 independent GEO cohorts were integrated for model development and validation. Multi-algorithm prioritization was applied to derive a consensus prognostic signature and nominate core candidates. Multi-omics analyses were conducted to characterize genomic alterations and immune associations. Functional relevance was examined using loss-of-function experiments in lung cancer cell lines.ResultsWe identified 52 inflammation-associated candidate genes and derived a 10-gene core signature with moderate prognostic performance across independent cohorts, while cohort and endpoint heterogeneity remained evident. Among these genes, PANX1 emerged as a focused candidate for further characterization based on an integrative assessment of model contribution, recurrence across analyses, and biological plausibility. Multi-omics analyses linked PANX1 expression to copy-number alterations, mutational context, and inferred immune microenvironment features. PANX1 was significantly upregulated in lung cancer cells, and its knockdown inhibited cell proliferation, reduced extracellular ATP release, and suppressed pro-inflammatory cytokine expression, including IL6, TNFA, and IL1B.ConclusionBy integrating large-scale multi-cohort analysis, machine learning prioritization, and experimental validation, this study identifies PANX1 as a potential inflammation-associated regulator in LUAD. These findings support a previously underrecognized association between PANX1 and inflammatory tumor biology and suggest its potential value as a biomarker candidate, while further mechanistic and translational validation is required.
Abstract from DOAJ. Public domain (CC0 1.0).
Read the article at the publisher →
Yang, Y., Fan, T., Miao, X., et al. (2026). Machine learning prioritization identifies PANX1 as an inflammation-associated candidate regulator in lung adenocarcinoma. Frontiers in Genetics. https://doi.org/10.3389/fgene.2026.1831053