Dysregulated miRNA Expression and Androgen Receptor Loss in Racially Distinct Triple-Negative Breast Cancer

Abstract

Androgen receptor (AR)-negative triple-negative breast cancer (TNBC), often termed quadruple-negative breast cancer (QNBC), disproportionately impacts women of African descent, leading to poorer overall survival (OS). MiRNAs regulate the expression of gene drivers involved in critical signaling pathways in TNBC, such as the AR gene, and their expression varies across races and breast cancer subtypes. This study investigates whether differentially expressed miRNAs influence AR transcription, potentially contributing to the observed disparities between African American (AA) and European American (EA) QNBC patients. Race-annotated TNBC samples (n = 129) were analyzed for AR expression status and revealed the prevalence of QNBC in AA patients compared to EA (76.6% vs. 57.7%) and a significant association of AR loss with poor survival among AAs. The Cancer Genome Atlas (TCGA) RNA-seq data showed that AAs with TNBC (n = 32) had lower AR mRNA levels than EAs (n = 67). Among TCGA patients in the AR-low group, AAs had significantly poorer OS than EAs. In our cohort, 46 miRNAs exhibited differential expression between AAs and EAs with QNBC. Ten of these miRNAs (miR-1185-5p, miR-1305, miR-3161, miR-3690, miR-494-3p, miR-509-3-5p, miR-619-3p, miR-628-3p, miR-873-5p, and miR-877-5p) were predicted to target the AR gene/signaling. The loss of AR expression is linked to poorer prognoses in AA women. The understanding of the specific miRNAs involved and their regulatory mechanisms on AR expression could provide valuable insights into why AA women are more prone to QNBC.

Keywords: androgen receptor; miRNA; prognosis; quadruple negative; race; triple-negative breast cancer.

Figure 1

AR immunostaining and overall survival analysis in TNBC samples of EA and AA patients. (A) Micrographs representing AR staining in tumor tissues from EAs (80%) and AAs (6%), AR (brown) and nuclei (blue), Insets: 20× objective; (B) Kaplan–Meier plots of overall survival in TNBC (n = 40) and QNBC (n = 89) patients; (C) cases that were EA and AA TNBC and QNBC by IHC.

Figure 2

DNA methylation of AR gene of AA and EA patients of TCGA database. (A) Methylation status of AR promoter region in normal and QNBC (AA vs. EA). Beta value ranges from 0 to 1 (no to complete methylation); significance based on unpaired t-test; (B) oncoprint showing AR mutations and CNAs in AR-low samples (n = 278).

Figure 3

Supervised hierarchical cluster (SHC) of global miRNA expression profiling of QNBC cases of EA and AA patients. Heatmap shows 46 DE miRNAs between AA QNBC (n = 33, blue bars) and EA QNBC (n = 9, green bars) from discovery cohort. Shown below clinical data of patients [age (>50 years, 50 years), tumor size (>3 cm, ≤3 cm, and grade (2, 3), and survival status (alive, deceased)].

Figure 4

MiRNA expression profiling analysis of QNBC cases of EA and AA patients. (A) Partial supervised hierarchical cluster (SHC) of global miRNA profiling showing DE miRNAs between AA QNBC (n = 10, blue bars) and EA QNBC (n = 9, green bars) (validation cohort). (B) Six miRNAs DE between AA QNBC and EA QNBC with expression directions common in discovery and validation cohorts.

Figure 5

ROC analysis and corresponding p-values of six miRNAs (common to both discovery and validation cohorts of patients). Black line: test real classifier value; Red line: random classifier value (AUC = 0.5).

Figure 6

Network of six miRNAs and experimentally validated targets of AR gene (color circles). Solid lines: protein–protein interaction; dashed lines: miRNA–mRNA interaction (Cytoscape 3.9.1).