MicroRNA in combination with HER2-targeting drugs reduces breast cancer cell viability in vitro

HER2-positive (HER2+) breast cancer patients that do not respond to targeted treatment have a poor prognosis. The effects of targeted treatment on endogenous microRNA (miRNA) expression levels are unclear. We report that responsive HER2+ breast cancer cell lines had a higher number of miRNAs with altered expression after treatment with trastuzumab and lapatinib compared to poorly responsive cell lines. To evaluate whether miRNAs can sensitize HER2+ cells to treatment, we performed a high-throughput screen of 1626 miRNA mimics and inhibitors in combination with trastuzumab and lapatinib in HER2+ breast cancer cells. We identified eight miRNA mimics sensitizing cells to targeted treatment, miR-101-5p, mir-518a-5p, miR-19b-2-5p, miR-1237-3p, miR-29a-3p, miR-29c-3p, miR-106a-5p, and miR-744-3p. A higher expression of miR-101-5p predicted better prognosis in patients with HER2+ breast cancer (OS: p=0.0392; BCSS: p=0.0125), supporting the tumor-suppressing role of this miRNA. In conclusion, we have identified miRNAs that sensitize HER2+ breast cancer cells to targeted therapy. This indicates the potential of combining targeted drugs with miRNAs to improve current treatments for HER2+ breast cancers. Four HER2 positive breast cancer cell lines (SKBR3, BT-474, KPL4 and SUM190PT) were treated with trastuzumab (10 µg/mL), lapatinib (100 nM) or trastuzumab plus lapatinib for 24 hours. miRNA expression levels were measured in untreated and treated cells using the one-color microarray SurePrint Human miRNA Microarray with design ID 031181, release 16.0, 8 x 60K (Agilent Technologies, Santa Clara, CA, USA) according to the protocol supplied by the manufacturer (miRNA Microarray System v2.3). Scanning was performed on Agilent Scanner G2565B. Samples were processed using Feature Extraction (FE) version 10.7.3.1 (Agilent Technologies). Quality was assessed by the quality control parameters in FE. The data were log2-transformed and for each sample, considering only expressed miRNAs, the data were median centered. All non-expressed miRNAs across samples were set to a common minimum value.