Here, we describe an integrated bioinformatics, functional analysis, and translational pathology approach to identify novel miRNAs involved in breast cancer progression. Coinertia analysis (CIA) was used to combine a database of predicted miRNA target sites and gene expression data. Using two independent breast cancer cohorts, CIA was combined with correspondence analysis and between group analysis to produce a ranked list of miRNAs associated with disease progression. Ectopic expression studies were carried out in MCF7 cells and miRNA expression evaluated in two additional cohorts of patients with breast cancer by in situ hybridization on tissue microarrays. CIA identified miR-187 as a key miRNA associated with poor outcome in breast cancer. Ectopic expression of miR-187 in breast cancer cells resulted in a more aggressive phenotype. In a test cohort (n = 117), high expression of miR-187 was associated with a trend toward reduced breast cancer-specific survival (BCSS; P = 0.058), and a significant association with reduced BCSS in lymph node-positive patients (P = 0.036). In a validation cohort (n = 470), high miR-187 was significantly associated with reduced BCSS in the entire cohort (P = 0.021) and in lymph node-positive patients (P = 0.012). Multivariate Cox regression analysis revealed that miR-187 is an independent prognostic factor in both cohorts [cohort 1: HR, 7.37; 95% confidence interval (CI), 2.05-26.51; P = 0.002; cohort 2: HR, 2.80; 95% CI, 1.52-5.16; P = 0.001] and in lymph node-positive patients in both cohorts (cohort 1: HR, 13.74; 95% CI, 2.62-72.03; P = 0.002; cohort 2: HR, 2.77; 95% CI, 1.32-5.81; P = 0.007). miR-187 expression in breast cancer leads to a more aggressive, invasive phenotype and acts as an independent predictor of outcome.
University College Dublin ->