Transcriptome Analyses Identify a Metabolic Gene Signature Indicative of Antitumor Immunosuppression of EGFR Wild Type Lung Cancers With Low PD-L1 Expression

Min Wang; Jie Zhu; Fang Zhao; Jiani Xiao

doi:10.3389/fonc.2021.643503

Transcriptome Analyses Identify a Metabolic Gene Signature Indicative of Antitumor Immunosuppression of EGFR Wild Type Lung Cancers With Low PD-L1 Expression

Front Oncol. 2021 Sep 14:11:643503. doi: 10.3389/fonc.2021.643503. eCollection 2021.

Authors

Min Wang¹, Jie Zhu², Fang Zhao³, Jiani Xiao⁴

Affiliations

¹ Department of Respiratory and Geriatrics, Chongqing Public Health Medical Center, Chongqing, China.
² Department of Respiratory and Critical Care Medicine, West China Hospital, Sichuan University, Chengdu, China.
³ Department of Intensive Care Unit, The People's Hospital of Tongliang District, Chongqing, China.
⁴ Department of Pediatric Dentistry, West China Hospital of Stomatology, Sichuan University, Chengdu, China.

Abstract

Purpose: With the development and application of targeted therapies like tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs), non-small cell lung cancer (NSCLC) patients have achieved remarkable survival benefits in recent years. However, epidermal growth factor receptor (EGFR) wild-type and low expression of programmed death-ligand 1 (PD-L1) NSCLCs remain unmanageable. Few treatments for these patients exist, and more side effects with combination therapies have been observed. We intended to generate a metabolic gene signature that could successfully identify high-risk patients and reveal its underlying molecular immunology characteristics.

Methods: By identifying the bottom 50% PD-L1 expression level as PD-L1 low expression and removing EGFR mutant samples, a total of 640 lung adenocarcinoma (LUAD) and lung squamous carcinoma (LUSC) tumor samples and 93 adjacent non-tumor samples were finally extracted from The Cancer Genome Atlas (TCGA). We identified differentially expressed metabolic genes (DEMGs) by R package limma and the prognostic genes by Univariate Cox proportional hazards regression analyses. The intersect genes between DEMGs and prognostic genes were put into the least absolute shrinkage and selection operator (LASSO) penalty Cox regression analysis. The metabolic gene signature contained 18 metabolic genes generated and successfully stratified LUAD and LUSC patients into the high-risk and low-risk groups, which was also validated by the Gene Expression Omnibus (GEO) database. Its accuracy was proved by the time-dependent Receiver Operating Characteristic (ROC) curve, Principal Components Analysis (PCA), and nomogram. Furthermore, the Single-sample Gene Set Enrichment Analysis (ssGSEA) and diverse acknowledged methods include XCELL, TIMER, QUANTISEQ, MCPcounter, EPIC, CIBERSORT-ABS, and CIBERSORT revealed its underlying antitumor immunosuppressive status. Besides, its relationship with somatic copy number alterations (SCNAs) and tumor mutational burden (TMB) was also discussed.

Results: It is noteworthy that metabolism reprogramming is associated with the survival of the double-negative LUAD and LUSC patients. The SCNAs and TMB of critical metabolic genes can inhibit the antitumor immune process, which might be a promising therapeutic target.

Keywords: GEO; The Cancer Genome Atlas Program (TCGA); epidermal growth factor receptor (EGFR) wild-type; low expression of programmed death-ligand 1 (PD-L1); lung adenocarcinoma; lung squamous carcinoma.