Amino Acid Metabolism-Related lncRNA Signature Predicts the Prognosis of Breast Cancer

Abstract

Background and Purpose: Breast cancer (BRCA) is the most frequent female malignancy and is potentially life threatening. The amino acid metabolism (AAM) has been shown to be strongly associated with the development and progression of human malignancies. In turn, long noncoding RNAs (lncRNAs) exert an important influence on the regulation of metabolism. Therefore, we attempted to build an AAM-related lncRNA prognostic model for BRCA and illustrate its immune characteristics and molecular mechanism. Experimental Design: The RNA-seq data for BRCA from the TCGA-BRCA datasets were stochastically split into training and validation cohorts at a 3:1 ratio, to construct and validate the model, respectively. The amino acid metabolism-related genes were obtained from the Molecular Signature Database. A univariate Cox analysis, least absolute shrinkage and selection operator (LASSO) regression, and a multivariate Cox analysis were applied to create a predictive risk signature. Subsequently, the immune and molecular characteristics and the benefits of chemotherapeutic drugs in the high-risk and low-risk subgroups were examined. Results: The prognostic model was developed based on the lncRNA group including LIPE-AS1, AC124067.4, LINC01655, AP005131.3, AC015802.3, USP30-AS1, SNHG26, and AL589765.4. Low-risk patients had a more favorable overall survival than did high-risk patients, in accordance with the results obtained for the validation cohort and the complete TCGA cohort. The elaborate results illustrated that a low-risk index was correlated with DNA-repair-associated pathways; a low TP53 and PIK3CA mutation rate; high infiltration of CD4⁺ T cells, CD8⁺ T cells, and M1 macrophages; active immunity; and less-aggressive phenotypes. In contrast, a high-risk index was correlated with cancer and metastasis-related pathways; a high PIK3CA and TP53 mutation rate; high infiltration of M0 macrophages, fibroblasts, and M2 macrophages; inhibition of the immune response; and more invasive phenotypes. Conclusion: In conclusion, we attempted to shed light on the importance of AAM-associated lncRNAs in BRCA. The prognostic model built here might be acknowledged as an indispensable reference for predicting the outcome of patients with BRCA and help identify immune and molecular characteristics.

Keywords: amino acid metabolism; breast cancer; immunity; long non-coding RNA; prognostic model; prognostic signature.

FIGURE 1

A flow chart of the data analysis and process.

FIGURE 2

GO and KEGG analysis of amino acid metabolism-associated DEGs. (A) GO and (B) KEGG.

FIGURE 3

Amino acid metabolism-associated lncRNA signature based on training sets. (A) Univariate cox analysis (B) Kaplan–Meier curves, (C) multi-index ROC analysis, (D) risk score, and (E) time-dependent ROC analysis. Univariate and multivariate Cox analyses of the expression of AAM-related lncRNAs. (F) Univariate and (G) multivariate analyses.

FIGURE 4

Construction of the mRNA–lncRNA regulatory network (A). Heatmap of the clinicopathological manifestations and AAM-related lncRNA prognostic signature (B).

FIGURE 5

Nomogram of both prognostic AAM-associated lncRNAs and clinical–pathological factors (A). Calibration plot for the nomogram (B). Stratification analysis of the risk score in BRCA. (C,D) Age (age >65 and age ≤65 years). (E,F) Tumor stage (I–II or III–IV). (G,H) Tumor T stage (T1–2 or T3–4). (I,J) Tumor M stage (M0 or M1). (K,L) Tumor N stage (N0 or N1–3).

FIGURE 6

Gene enrichment analysis for AAM-related lncRNAs based on TCGA in the high (A) and low (B) BRCA risk groups. Correlation between the TMB and the two risk subsets (C). Association between the TMB and risk score (D). Prominently mutated genes in the patients with BRCA in the different risk subgroups. The mutated genes (rows, top 20) are ranked according to mutation rate; samples (columns) are arranged to emphasize the mutual exclusivity among mutations. The right panels depicts the mutation percentage, and the top panel indicates the overall number of mutations. The color coding indicates the mutation type (E).

FIGURE 7

Evaluation of the TME and levels of lymphocyte infiltration in the two groups. (A) Associations between the risk score and the immune and stromal scores. (B) Associations between the risk score and immune cell types.

FIGURE 8

Immune cell infiltration levels and corresponding function determined by ssGSEA (A). Expression of m6A-related genes in both groups (B). Expression of immune checkpoint-related genes in both groups (C).

FIGURE 9

Heatmap and table showing the distribution of the BRCA PAM50 molecular subtypes (basal, luminal A, luminal B, HER2⁺, and TNBC) in the risk subgroups (A). Relationships between the risk score and chemotherapeutic sensitivity (B–G).