Presentation Abstract

Abstract Number: 5049
Presentation Title: Serum miRNAs as an early marker for breast cancer
Presentation Time: Wednesday, Apr 04, 2012, 8:00 AM -12:00 PM
Location: McCormick Place West (Hall F), Poster Section 7
Poster Section: 7
Poster Board Number: 25
Author Block: Ashley C. Godfrey, Zongli Xu, Clarice R. Weinberg, Paul A. Wade, Lisa A. Deroo, Dale P. Sandler, Jack A. Taylor. NIEHS, RTP, NC
Abstract Body: MicroRNAs (miRNAs) are small, non-coding, single-stranded RNAs between 18-22 nucleotides that regulate gene expression by affecting mRNA translation and degradation. It has been proposed that miRNAs could be master regulators of gene expression by regulating entire gene networks. Expression of miRNAs is altered in tumor compared to normal tissue and there is growing evidence that miRNA profiles in serum may differ in cancer cases as compared to healthy controls. However, these studies use samples collected after diagnosis so treatment and/or disease effects cannot be excluded. We are using serum samples from the Sister Study, a prospective cohort study of 50,000 women who had never had breast cancer at the time of enrollment, but were at increased risk because they had a sister previously diagnosed with breast cancer. We are exploring whether serum miRNA profiles can be used as prediagnostic markers of breast cancer. In an initial pilot analysis comparing 5 women diagnosed with breast cancer 1-2 months after their blood draw to 5 controls, there were 25 miRNAs significantly different at the 0.05 level (paired T test). Of those 25 miRNAs, 7 have been previously associated with breast tumor tissue in published studies showing differential expression in tumor tissue or in breast cancer cell lines. We have now completed a larger follow up study examining 205 women diagnosed with breast cancer within 18 months of their blood draw compared to 205 controls frequency matched to cases by age, 2 month interval of blood draw and race. We are using Affymetrix miRNA microarrays to profile miRNA expression in serum samples, and have optimized this platform for use with less than .5mLs of serum. After normalization and background correction, we found 439 miRNAs (of 1105 human miRNAs on the chip) expressed above background in at least 50 samples. When the average expression of controls was compared to cases using logistic regression, we found 16 miRNAs significantly different at the .05 level. For cases, the mean time between blood draw and cancer diagnosis was 10 months. We hypothesize that, if miRNA is a prediagnostic marker, cases with a longer time between blood draw and diagnosis would have a more control-like miRNA profile. To examine this, we are conducting a more detailed analysis among cases with a short time between blood draw and diagnosis. The next steps of our analysis will be to examine how far in advance of cancer diagnosis can we detect case-control differences in miRNA expression profiles.