A five-gene expression signature of centromeric proteins with prognostic value in lung adenocarcinoma

Centromere proteins (CENPs) form a large protein family. Sixteen proteins in this family are positioned at the centromere throughout the cell cycle. The overexpression of CENPs is common in many cancers and predicts a poor prognosis. However, a comprehensive analysis of CENPs expression has not been conducted, and their clinical significance in lung adenocarcinoma (LUAD) is unclear.

We investigated the expression differences of the CENP family in LUAD using The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) cohorts. Kaplan-Meier curve survival analysis was performed to assess their independent prognostic values. We then tested 5 clinical LUAD specimens by quantitative real time polymerase chain reaction (qRT-PCR). The risk model was constructed with least absolute shrinkage and selection operator (LASSO). Cox regression analyses were carried out to determine independent prognostic indicators. Weighted gene coexpression network analysis (WGCNA) was employed to define the coexpression networks.

The messenger RNA (mRNA) expression of 15 differential CENP proteins was higher in LUAD than in normal lung tissues. Among them, 10 CENP proteins had significant prognostic value. The risk model comprising CENPF, CENPU, CENPM, CENPH, and CENPW showed a significant correlation [hazard ratio (HR) 1.75, 95% confidence interval (CI): 1.3-2.35; P=2e-04]. However, the prognostic accuracy was not strong [1-year survival: area under curve (AUC) 0.63; 3-year survival: AUC 0.62; 5-year survival: AUC 0.6]. The qRT-PCR results showed that the 5 CENPs were upregulated in LUAD tissues compared to in normal lung tissues. A total of 441 hub genes coexpressed with the 5 CENPs were identified.

CENPF, CENPU, CENPM, CENPH, and CENPW have prognostic values and may be potential targets for LUAD treatment.