Candidate diagnostic miRNAs that can detect cancer in prostate biopsy.

BACKGROUND: While histopathological evaluation remains the gold standard for diagnosis of prostate cancer (PCa), sampling errors remain a frequent problem; therefore, use of tissue biomarkers that can distinguish between benign and malignant prostate disease is a potentially beneficial diagnostic strategy.

METHODS: Deep sequencing of the miRNA transcriptome of 14 benign prostatic hyperplasia (BPH) and 60 cancerous and non-cancerous prostate samples extracted from 34 cancer-bearing prostates removed by prostatectomy was performed; of the latter 60 samples, 16, 21, and 23 samples contained <10%, >30%, and no dysplastic cells, respectively. The predictive value of selected miRNAs was then tested by quantitative reverse-transcribed PCR (qRT-PCR), using two separate chemistries, Exiqon and Taqman, to evaluate the tissue samples obtained by prostatectomy. Validation experiments were also performed for a subset of miRNAs by qRT-PCR of 87 prostate core biopsies.

RESULTS: We identified 123 miRNAs significantly dysregulated in PCa (adjusted P-values <0.05); 110 and 13 miRNAs were dysregulated only in cancerous samples and non-cancerous samples extracted from cancer-bearing prostates, respectively, while 31 were dysregulated regardless of the dysplastic cell content of the studied specimens. The clinical utility of eight selected miRNAs was analyzed using the same sample set with two qRT-PCR chemistries. Measurable qRT-PCR signals were obtained for seven and six miRNAs using the Exiqon and Taqman chemistries, respectively, and expression levels of six and four of these miRNAs differed significantly between BPH and PCa samples, regardless of dysplastic cell content. Validation experiments on core biopsies using qRT-PCR confirmed differential expression between BPH and PCa of four miRNAs (miR-187-3p, miR-183-5p, miR-32-5p, and miR-141-5p) using the Exiqon and one miRNA (miR-187-3p) with the Taqman chemistry.

CONCLUSIONS: Our sequencing analyses identified several candidate diagnostic miRNAs and confirmed some which have previously been reported as diagnostic in prostate malignancy. The results of this study suggest also that some of selected miRNAs can differentiate between non-malignant and malignant prostates even when neoplastic cells are missing from the studied specimen.