A Novel Risk Score Model of Lactate Metabolism for Predicting over Survival and Immune Signature in Lung
Zhou Jiang1, Yongzhong Luo1, Lemeng Zhang1
1Thoracic Medicine Department, Hunan Cancer Hospital/The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Tongzipo Rd 283#, Yuelu District, Changsha 410013, China.
Background:
The role of lactate acid in tumor progression was well proved. Recently, it was found that lactate acid accumulation induced an immunosuppressive microenvironment. However, these results were based on a single gene and it was unclear that lactate acid genes were associated with immunotherapy and able to predict overall survival.
Methods:
Genes and survival data were acquired from TCGA, GEO and GENECARDS. PCA and TSNE were used to distinguish sample types according to lactate metabolism-associated gene expression. A Wilcox-test examined the expression differences between normal and tumor samples. The distribution in chromatin and mutant levels were displayed by Circo and MAfTools. The lactate metabolism-associated gene were divided into categories by consistent clustering and visualized by Cytoscape. Immune cell infiltration was evaluated by CIBERSORT and LM22 matrix. Enrichment analysis was performed by GSVA. We used the ConsensusClusterPlus package for consistent cluster analysis. A prognostic model was constructed by Univariate Cox regression and Lasso regression analysis. Clinical specimens were detected their expression of genes in model by IHC.
Results:
Most lactate metabolism-associated gene were significantly differently expressed between normal and tumor samples. There was a strong correlation between the expression of lactate metabolism-associated gene and the abundance of immune cells. We divided them into two clusters (lactate.cluster A,B) with significantly different survival. The two clusters showed a difference in signal, immune cells, immune signatures, chemokines, and clinical features. We identified 162 differential genes from the two clusters, by which the samples were divided into three categories (gene.cluster A,B,C). They also showed a difference in OS and immune infiltration. Finally, a risk score model that was composed of six genes was constructed. There was significant difference in the survival between the high and low risk groups. ROC curves of 1, 3, 5, and 10 years verified the model had good predictive efficiency. Gene expression were correlated with ORR and PFS in patients who received anti-PD-1/L1.
Conclusion:
The lactate metabolism-associated genes in LUAD were significantly associated with OS and immune signatures. The risk scoring model that was constructed by us was able to well identify and predict OS and were related with anti-PD-1/L1 therapy outcome.
