Abstract 3971: Genome-wide tissue-based microRNA signature in healthy women predicting breast cancer risk

• Published in: Cancer research (Chicago, Ill.) Ill.), 2015-08, Vol.75 (15_Supplement), p.3971-3971
• Main Authors: Taslim, Cenny, Weng, Daniel Y., Brasky, Theodore M., Dumitrescu, Ramona G., Huang, Kun, Kallakury, Bhaskar V. s., Krishnan, Shiva, Llanos, Adana A., Marian, Catalin, Schneider, Sallie S., Spear, Scott L., Troester, Melissa A., Freudenheim, Jo L., Geyer, Susan, Shields, Peter G.
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2015-3971

Introduction: Small non-coding microRNAs (miRNAs) play important roles in both normal breast development and breast carcinogenesis. The goal of this study is to identify miRNAs in normal breast tissues which are related to breast cancer risk. Materials and Methods: We used a high-throughput digital counting of miRNAs without amplification (Nanostring®) to examine miRNA expression in 161 reduction mammoplasty (RM) tissues from two independent studies. A multivariate model was used to identify miRNAs associated with breast cancer risk (based upon Gail risk scores) in a training study (n = 90) then the model was validated in a replication study (n = 71). Risk-related microRNAs were then evaluated in serum for associations with real breast cancer cases using publically available prospective cohort (Sister Study, n = 410). Results: We identified a 41-miRNA signature in healthy women distinguishing high risk from low risk women with a prediction accuracy of 82% (95% CI = 80% to 87%) in the training study. Predictive accuracy was 69% (95% CI = 65% to 73%) in the replication study. 34 of 41 serum miRNAs that mapped to public data predicted women who developed breast cancer within 18 months after blood draw from those who remained cancer free with accuracy of 59% (95% CI = 57% to 61%). We have also shown that these accuracies were significantly higher than random chance (P < 0.0001). IPA canonical pathway analysis revealed that the risk-related microRNAs targets were significantly enriched for HER-2 signaling in breast cancer, and estrogen-dependent breast cancer signaling, and other important cancer pathways such as molecular mechanisms of cancer, PI3K/AKT signaling, PTEN signaling, and TGF-beta signaling. Conclusion: Our results indicate that miRNA profiling from breast tissue of healthy patients may identify clinically useful predictors of breast cancer risk and these miRNAs may also work as non-invasive biomarker for early breast cancer prediction. Citation Format: Cenny Taslim, Daniel Y. Weng, Theodore M. Brasky, Ramona G. Dumitrescu, Kun Huang, Bhaskar V. s. Kallakury, Shiva Krishnan, Adana A. Llanos, Catalin Marian, Sallie S. Schneider, Scott L. Spear, Melissa A. Troester, Jo L. Freudenheim, Susan Geyer, Peter G. Shields. Genome-wide tissue-based microRNA signature in healthy women predicting breast cancer risk. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelp