Glucose metabolism transcriptome clustering identifies subsets of resectable lung adenocarcinoma with different prognoses

Abstract

Objectives

Reprogramming of energy metabolism is a well-established hallmark of cancer, with aerobic glycolysis classically considered a prominent feature. We investigate the heterogeneity in glucose metabolism pathways within resectable primary lung adenocarcinoma and its clinical significance.

Methods

Using The Cancer Genome Atlas data, RNA expressions were extracted from 489 primary lung adenocarcinoma samples. Prognostic influence of glycolytic, aerobic, and mitochondrial markers (monocarboxylate transporter [MCT]4, MCT1, and translocase of outer mitochondrial membrane 20, respectively) was assessed using Kaplan-Meier analysis. Clustering of 35 genes involved in glucose metabolism was performed using the k-means method. The clusters were then analyzed for associations with demographic, clinical, and pathologic variables. Overall survival was assessed using the Kaplan-Meier estimator. Multivariate analysis was performed to assess the independent prognostic value of cluster membership.

Results

Classical statistical approach showed that higher expression of MCT4 was associated with a significantly worse prognosis. Increased expression of translocase of outer mitochondrial membrane 20 was associated with a nonsignificant trend toward better prognosis, and increased expression of MCT1 was associated with a better outcome. Clustering identified 3 major metabolic phenotypes, dominantly hypometabolic, dominantly oxidative, and dominantly mixed oxidative/glycolytic with significantly different pathologic stage distribution and prognosis; mixed oxidative/glycolytic was associated with worse survival. Cluster membership was independently associated with survival.

Conclusions

This study demonstrates the existence of distinct glucose metabolism clusters in resectable lung adenocarcinoma, providing valuable prognostic information. The findings highlight the potential relevance of considering metabolic profiles when designing strategies for reprogramming energy metabolism. Further studies are warranted to validate these findings in different cancer types and populations.