Identification of an eight-lncRNA prognostic model for breast cancer using WGCNA network analysis and a Cox-proportional hazards model based on L1-penalized estimation

Abstract

An ever-increasing number of long noncoding (lnc) RNAs has been identified in breast cancer. The present study aimed to establish an lncRNA signature for predicting survival in breast cancer. RNA expression profiling was performed using microarray gene expression data from the National Center for Biotechnology Information Gene Expression Omnibus, followed by the identification of breast cancer‑related preserved modules using weighted gene co‑expression network (WGCNA) network analysis. From the lncRNAs identified in these preserved modules, prognostic lncRNAs were selected using univariate Cox regression analysis in combination with the L1‑penalized (LASSO) Cox‑proportional Hazards (Cox‑PH) model. A risk score based on these prognostic lncRNAs was calculated and used for risk stratification. Differentially expressed RNAs (DERs) in breast cancer were identified using MetaDE. Gene Set Enrichment Analysis pathway enrichment analysis was conducted for these prognostic lncRNAs and the DERs related to the lncRNAs in the preserved modules. A total of five preserved modules comprising 73 lncRNAs were mined. An eight‑lncRNA signature (IGHA1, IGHGP, IGKV2‑28, IGLL3P, IGLV3‑10, AZGP1P1, LINC00472 and SLC16A6P1) was identified using the LASSO Cox‑PH model. Risk score based on these eight lncRNAs could classify breast cancer patients into two groups with significantly different survival times. The eight‑lncRNA signature was validated using three independent cohorts. These prognostic lncRNAs were significantly associated with the cell adhesion molecules pathway, JAK‑signal transducer and activator of transcription 5A pathway, and erbb pathway and are potentially involved in regulating angiotensin II receptor type 1, neuropeptide Y receptor Y1, KISS1 receptor, and C‑C motif chemokine ligand 5. The developed eight-lncRNA signature may have clinical implications for predicting prognosis in breast cancer. Overall, this study provided possible molecular targets for the development of novel therapies against breast cancer.