Abstract
BACKGROUND: MicroRNA (miRNA) and mRNA levels in matching specimens were used to identify miRNA-mRNA interactions. We aimed to integrate transcriptome, immunophenotype, methylation, mutation, and survival data analyses to examine the profiles of miRNAs and target mRNAs and their associations with breast cancer (BC) diagnosis. METHODS: Based on the Gene Expression Omnibus (GEO) database and The Cancer Genome Atlas (TCGA), differentially expressed miRNAs and targeted mRNAs were screened from experimentally verified miRNA-target interaction databases using Pearson's correlation analysis. We used real-time quantitative reverse transcription polymerase chain reaction to verify BC and benign disease samples, and logistic regression analysis was used to establish a diagnostic model based on miRNAs and target mRNAs. Receiver operating characteristic curve analysis was performed to test the ability to recognize the miRNA-mRNA pairs. Next, we investigated the complex interactions between miRNA-mRNA regulatory pairs and phenotypic hallmarks. RESULTS: We identified 27 and 359 dysregulated miRNAs and mRNAs, respectively, based on the GEO and TCGA databases. Using Pearson's correlation analysis, 10 negative miRNA-mRNA regulatory pairs were identified after screening both databases, and the related miRNA and target mRNA levels were assessed in 40 BC tissues and 40 benign breast disease tissues. Two key regulatory pairs (miR-205-5p/High mobility group box 3 (HMGB3) and miR-96-5p/Forkhead Box O1 (FOXO1)) were selected to establish the diagnostic model. They also had utility in survival and clinical analyses. CONCLUSIONS: A diagnostic model including two miRNAs and their respective target mRNAs was established to distinguish between BC and benign breast diseases. These markers play essential roles in BC pathogenesis.