# A metabolic-radioimmune signature predicts therapy response and immune reprogramming in non-small cell lung cancer

## TL;DR

This study identifies a metabolic signature in non-small cell lung cancer that predicts treatment response and immune changes, offering new strategies to improve radiotherapy outcomes.

## Contribution

A novel metabolic-radioimmune signature is developed to predict therapy response and identify radiosensitizing agents in NSCLC.

## Key findings

- Radiotherapy induces metabolic heterogeneity in the tumor microenvironment, with distinct changes in tumor cells and immune cells.
- A seven-gene RRMG signature robustly predicts poor survival and identifies immunosuppressive tumor environments.
- Ouabain and two novel compounds are identified as potential radiosensitizers based on experimental validation.

## Abstract

This study systematically investigates radiotherapy-induced metabolic remodeling across the TME, encompassing tumor cells, immune cells, and tumor-draining lymph nodes (TDLNs), and establishes a prognostic signature based on radioresistance-related metabolic genes (RRMGs) to optimize therapeutic stratification and radiosensitizer discovery.

Bulk transcriptomic datasets of NSCLC tumor cells and tumor-draining TDLNs were systematically integrated, along with single-cell RNA-seq data from tumor tissues, to reconstruct metabolic flux maps using the single-cell Flux Estimation Analysis (scFEA) algorithm. WGCNA and Cox regression modeling of TCGA radiotherapy cohort were used to identify core RRMGs. A prognostic nomogram was developed using risk scores derived from these genes, while CIBERSORT and TIDE algorithms were used to evaluated TIME features and immunotherapy responses. Candidate radiosensitizing agents were predicted via the oncoPredict platform and validated by molecular docking, qRT-PCR and western blotting in radioresistant NSCLC cells.

Radiotherapy induced profound metabolic heterogeneity across the NSCLC TIME: Tumor cells and draining TDLNs exhibited suppressed tricarboxylic acid (TCA) cycle activity and N-glycan biosynthesis, while immune cells displayed upregulated serine metabolism alongside divergent shifts in lymphoid subsets. Seven RRMGs were identified as key prognostic determinants, including PGD, IDH2, G6PD, ALDH3A1, UPP1, XYLT2, AACS. The RRMG-based risk model robustly predicted poor overall survival (HR = 4.726, 95% CI: 2.154-10.371; P<0.001), with high predictive accuracy (AUC for 1-, 3-, and 5-year: 0.752, 0.778, and 0.879). High-risk patients demonstrated an immunosuppressive TIME marked by elevated tumor-promoting immune cell infiltration and TIDE scores. The model’s generalizability was verified in an independent radioimmunotherapy cohort (AUC: 0.618). Experimental validation revealed significant upregulation of high-risk RRMGs in radioresistant NSCLC cells. Ouabain and two novel compounds (BRD-K28456706, BRD-K42260513) were nominated as promising radiosensitizers.

Radiotherapy-induced metabolic reprogramming in TIME drives resistance of NSCLC. The RRMG signature predicts radioimmunotherapy outcomes for patient stratification. Identifying ouabain and novel compounds highlights targeting metabolic vulnerabilities as a translatable strategy to overcome resistance.