# A 12-gene immune signature predicts prognosis and identifies KRT6B as a therapeutic target in lung adenocarcinoma

## TL;DR

A 12-gene immune signature predicts survival in lung adenocarcinoma patients and identifies KRT6B as a potential treatment target.

## Contribution

A novel 12-gene immune signature and the identification of KRT6B as a therapeutic target in lung adenocarcinoma.

## Key findings

- The 12-gene signature predicted survival in LUAD patients across multiple datasets.
- KRT6B was identified as a central hub gene and validated as overexpressed in LUAD tumor cells.
- The model's accuracy for 1-, 2-, and 3-year survival was confirmed with AUC values of 0.624–0.788.

## Abstract

Lung adenocarcinoma (LUAD) exhibits high mortality and heterogeneity. While immune-related signatures show prognostic potential, robust models validated through both computational screening and experimental methods are lacking.

Transcriptomic and clinical data from The Cancer Genome Atlas (TCGA) database and three Gene Expression Omnibus (GEO) cohorts (GSE3141, GSE30219, and GSE50081) were analyzed. A 12-gene immune-related prognostic signature was constructed using LASSO Cox regression. The model was subsequently validated using three independent external cohorts. Its prognostic performance was comprehensively assessed using time-dependent receiver operating characteristic (ROC) curves. Functional enrichment analyses (GO, KEGG, and GSEA), tumor microenvironment (TME) profiling (via CIBERSORT and ESTIMATE algorithms), and drug sensitivity analyses were conducted. Protein-protein interaction (PPI) network analysis identified KRT6B as a central hub gene. KRT6B expression and its functional role were further validated through tissue microarray immunohistochemistry (IHC), as well as in vitro and in vivo experiments.

We developed a prognostic model for LUAD based on 12 immune-related genes and derived a risk score via LASSO regression. High-risk patients exhibited significantly worse overall survival compared to low-risk patients in both the training set (TCGA) and the three independent validation cohorts (GSE3141, GSE30219, and GSE50081) (all P < 0.05). Time-dependent ROC analysis confirmed the model’s predictive accuracy for 1-, 2-, and 3-year survival (AUC: 0.624–0.788). A nomogram incorporating the risk score and key clinical indicators further enhanced prognostic performance (AUC: 0.753 to 0.763). PPI network analysis pinpointed KRT6B as a core hub gene within the signature. Subsequent experimental validation confirmed the overexpression of KRT6B in LUAD tumor cells and demonstrated its tumor-promoting functions both in vitro and in vivo.

We established and validated an immune-related gene signature for prognostic prediction and identified KRT6B as a promising prognostic biomarker and potential therapeutic target in LUAD.