Development and Validation of a Novel Stemness-Index-Related Long Noncoding RNA Signature for Breast Cancer Based on Weighted Gene Co-Expression Network Analysis

Abstract

Background: Breast cancer (BC) is a major leading cause of woman deaths worldwide. Increasing evidence has revealed that stemness features are related to the prognosis and progression of tumors. Nevertheless, the roles of stemness-index-related long noncoding RNAs (lncRNAs) in BC remain unclear. Methods: Differentially expressed stemness-index-related lncRNAs between BC and normal samples in The Cancer Genome Atlas database were screened based on weighted gene co-expression network analysis and differential analysis. Univariate Cox and least absolute shrinkage and selection operator regression analyses were performed to identify prognostic lncRNAs and construct a stemness-index-related lncRNA signature. Time-dependent receiver operating characteristic curves were plotted to evaluate the predictive capability of the stemness-index-related lncRNA signature. Moreover, correlation analysis and functional enrichment analyses were conducted to investigate the stemness-index-related lncRNA signature-related biological function. Finally, a quantitative real-time polymerase chain reaction was used to detect the expression levels of lncRNAs. Results: A total of 73 differentially expressed stemness-index-related lncRNAs were identified. Next, FAM83H-AS1, HID1-AS1, HOXB-AS1, RP11-1070N10.3, RP11-1100L3.8, and RP11-696F12.1 were used to construct a stemness-index-related lncRNA signature, and receiver operating characteristic curves indicated that stemness-index-related lncRNA signature could predict the prognosis of BC well. Moreover, functional enrichment analysis suggested that differentially expressed genes between the high-risk group and low-risk group were mainly involved in immune-related biological processes and pathways. Furthermore, functional enrichment analysis of lncRNA-related protein-coding genes revealed that FAM83H-AS1, HID1-AS1, HOXB-AS1, RP11-1070N10.3, RP11-1100L3.8, and RP11-696F12.1 were associated with neuroactive ligand-receptor interaction, AMPK signaling pathway, PPAR signaling pathway, and cGMP-PKG signaling pathway. Finally, quantitative real-time polymerase chain reaction revealed that FAM83H-AS1, HID1-AS1, RP11-1100L3.8, and RP11-696F12.1 might be used as the potential diagnostic biomarkers of BC. Conclusion: The stemness-index-related lncRNA signature based on FAM83H-AS1, HID1-AS1, HOXB-AS1, RP11-1070N10.3, RP11-1100L3.8, and RP11-696F12.1 could be used as an independent predictor for the survival of BC, and FAM83H-AS1, HID1-AS1, RP11-1100L3.8, and RP11-696F12.1 might be used as the diagnostic markers of BC.

Keywords: WGCNA; breast cancer; cancer stem cells; prognosis; stemness-index-related lncRNAs.

FIGURE 1

Identification of stemness-index-related module and lncRNAs based on WGCNA. Samples clustering analysis to remove outliers (A), determination of soft threshold and inspection of scale-free network (B), 11 modules were identified and presented in different colors by setting MEDissThres as 0.2 and minModuleSize as 30 (C), 11 modules were grouped into four clusters by correlation analyses (D), purple module has lowest correlation with other modules (E), and magenta module was most significantly negatively correlated with mRNAsi (p < .05 and correlation coefficient = −0.62) (F).

FIGURE 2

Construction and validation of a stemness-index-related lncRNA signature associated with survival of BC patients. Results of univariate Cox regression analysis (A) and LASSO Cox regression analysis (B), Kaplan–Meier survival analysis between high-risk and low-risk groups in TCGA database (C) and GSE20585 dataset (D), ROC curve evaluated efficiency of stemness-index-related lncRNA signature for predicting 1-, 3-, and 5-year OS in TCGA database (E) and GSE20585 dataset (F), and lncRNAs expression profiles, risk scores distribution, and patients' survival status in TCGA database (G) and GSE20585 dataset (H).

FIGURE 3

Kaplan–Meier survival stratifcation analyses in TCGA database based on stemness-index-related lncRNA signature. Age > = 60 years (A), age < 60 years (B), female (C), M0 (D), N0 (E), N1–N3 (F), stages i–ii (G), stages iii-iv (H), T1–T2 (I), T3–T4 (J).

FIGURE 4

Stemness-index-related lncRNA signature was an independent prognostic factor in BC. Univariate Cox regression analysis (A) and multivariate Cox regression analysis (B) to identify independent prognostic factors from stemness-index-related lncRNAs and other clinicopathological characteristics in TCGA database.

FIGURE 5

Functional annotation of stemness-index-related lncRNA signature. GO-Biological processes (A), GO-Cellular component (B), GO-Molecular function (C), and KEGG pathway enrichment analysis (D). Immune cell infiltration between high- and low-risk groups (E). GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

FIGURE 6

Potential regulatory mechanisms of lncRNAs in stemness-index-related lncRNA signature. Venn diagram of protein-coding gene associated with all of six lncRNAs (A), interaction network of five most relevant protein-coding genes and each lncRNA (B), and Sankey diagram showed five most relevant protein-coding genes of each lncRNA (C). KEGG pathway enrichment analysis of each lncRNA-related protein-coding gene. FAM83H-AS1 (D), HID1-AS1 (E), HOXB-AS1 (F), RP11-1070N10.3 (G), RP11-1100L3.8 (H), and RP11-696F12.1 (I). KEGG, Kyoto Encyclopedia of Genes and Genomes.

FIGURE 7

Investigation of diagnostic value of lncRNAs in stemness-index-related lncRNA signature. Expression levels of these lncRNAs in stemness-index-related lncRNA signature in TCGA database (A), and ROC curves to evaluate their capability in distinguishing BC and normal samples in TCGA database (B). Validation of expression of lncRNAs in stemness-index-related lncRNA signature by quantitative real-time polymerase chain reaction (C). Results were shown as mean ± SD. *p < .05 **p < .01 vs. MCF-10A.