Identification and validation of a copper homeostasis-related gene signature for the predicting prognosis of breast cancer patients via integrated bioinformatics analysis

TL;DR

This study identifies a gene signature related to copper homeostasis that predicts breast cancer patient outcomes and suggests targeting copper balance as a treatment strategy.

Contribution

A novel gene signature based on copper homeostasis-related lncRNAs is developed to predict breast cancer prognosis.

Findings

A gene signature with four copper homeostasis-related lncRNAs predicts poorer outcomes in high-risk breast cancer patients.

The model's prognostic value reached an AUC of 0.795 after ten years.

Higher expression of the four lncRNAs was observed in tumor tissues compared to normal tissues.

Abstract

The prognostic value of copper homeostasis-related genes in breast cancer (BC) remains largely unexplored. We analyzed copper homeostasis-related gene profiles within The Cancer Genome Atlas Program breast cancer cohorts and performed correlation analysis to explore the relationship between copper homeostasis-related mRNAs (chrmRNA) and lncRNAs. Based on these results, we developed a gene signature-based risk assessment model to predict BC patient outcomes using Cox regression analysis and a nomogram, which was further validated in a cohort of 72 BC patients. Using the gene set enrichment analysis, we identified 139 chrmRNAs and 16 core mRNAs via the Protein–Protein Interaction network. Additionally, our copper homeostasis-related lncRNAs (chrlncRNAs) (PINK1.AS, OIP5.AS1, HID.AS1, and MAPT.AS1) were evaluated as gene signatures of the predictive model. Kaplan–Meier survival analysis revealed that patients with a high-risk gene signature had significantly poorer clinical outcomes. Receiver operating characteristic curves showed that the prognostic value of the chrlncRNAs model reached 0.795 after ten years. Principal component analysis demonstrated the capability of the model to distinguish between low- and high-risk BC patients based on the gene signature. Using the pRRophetic package, we screened out 24 anticancer drugs that exhibited a significant relationship with the predictive model. Notably, we observed higher expression levels of the four chrlncRNAs in tumor tissues than in the adjacent normal tissues. The correlation between our model and the clinical characteristics of patients with BC highlights the potential of chrlncRNAs for predicting tumor progression. This novel gene signature not only predicts the prognosis of patients with BC but also suggests that targeting copper homeostasis may be a viable treatment strategy.

Figures