Identification and Validation of Three-Gene Signature in Lung Squamous Cell Carcinoma by Integrated Transcriptome and Methylation Analysis

Guanghua Li; Libo Wu; Jiaxing Yu; Siyang Zhai; Hailong Deng; Qiushi Wang

doi:10.1155/2022/9688040

Identification and Validation of Three-Gene Signature in Lung Squamous Cell Carcinoma by Integrated Transcriptome and Methylation Analysis

J Oncol. 2022 Sep 23:2022:9688040. doi: 10.1155/2022/9688040. eCollection 2022.

Authors

Guanghua Li¹, Libo Wu², Jiaxing Yu¹, Siyang Zhai¹, Hailong Deng³, Qiushi Wang¹

Affiliations

¹ Department of Thoracic Surgery, The Second Affiliated Hospital of Harbin Medical University, Harbin 150000, China.
² Department of Thoracic Surgery, The Second Affiliated Hospital of Hainan Medical College, Haikou 570100, China.
³ Department of Thoracic Surgery, Hailun People's Hospital, Hailun 152300, China.

Abstract

Since DNA methylation (DNAm) is associated with the carcinogenesis of various cancers, this study aimed to explore potential DNAm prognostic signatures of lung squamous cell carcinoma (LUSC). First, transcriptomic and methylation profiles of LUSC were obtained from The Cancer Genome Atlas database (TCGA). DNAm-related genes were screened by integrating DNAm and transcriptome profiles via MethylMix package. Subsequently, a prognostic signature was conducted with the least absolute shrinkage and selector operation (LASSO) Cox analysis. This signature combined with the clinicopathological parameters was then utilized to construct a prognostic nomogram via the rms package. A signature based on three DNAm-related genes claudin 1 (CLDN1), ATP-binding cassette subfamily C member 5 (ABCC5), and cystatin A (CSTA) that were hypomethylated and upregulated in LUSC was constructed. Univariate and multivariate Cox regression analysis suggested that this signature, combined with age and TNM.N stage, was significantly correlated with survival rate. Time-dependent receiver operating characteristics and calibration curves suggested the nomogram constructed with age and TNM.N stage variables could accurately evaluate the 3- and 5-year outcome of LUSC. Finally, the average mRNA and protein expression levels of CLDN1, ABCC5, and CSTA in LUSC were verified to be significantly higher than those in paracancerous tissues. Moreover, silencing CLDN1, ABCC5, and CSTA expressions could significantly reduce the carcinogenesis of the A549 cell line. The DNAm-driven prognostic signature consists of CLDN1, ABCC5, and CSTA incorporated with age and TNM. N stage could facilitate the prediction outcome of LUSC.