635 Background: Soy isoflavone consumption may reduce breast cancer risk, and should cause measureable changes in gene expression in the breast. We conducted a Phase IIB trial of soy intervention (1:1 randomization to G- 2535or placebo) and studied effects on the expression of a 28- gene panel, and other biomarkers. Methods: Breast sampling was performed using random fine needle aspiration; cDNA was prepared from 100 ng of purified RNA and amplified. We selected 30 genes: 14 targets of genistein, 9 related to ERα, 5 associated with epithelial atypia, and 2 housekeeping (HSK) genes (GAPDH and HPRT1). Taqman low density arrays (TLDA) were used to perform quantitative PCR reactions. The mean Ct values of triplicates were normalized against HSK and month-6 minus baseline differences were t-tested. ANOVA was used for global comparisons between groups. Results: RNA analyses could be performed on 89 of 98 women (9 had insufficient RNA). The soy group showed increased expression of 18/28 genes post-intervention; Benjamini-Hochberg adjustment reduced this number to 14 genes. These included 7 related to genistein effect (BCL-2, CDKN1A, CDKN2A, DDIT3, FAS, PARP-1, and TP53), 5 related to estrogen response (ESR1, FOXA1, MYB, PGR, SCUBE2) and 2 related to breast atypia (AR and Wnt5B). The control group showed increased expression of 8 genes, but after adjustment, one (IGF-1) remained significant. On comparing the change in gene expression between the soy and control groups, there were significant increases in 4 genes: FAS, FGFR3, FOXA1, MYB, which were non-significant after adjustment. The mean change across all 28 genes in the soy group from baseline to 6-months was 1.56-fold, whereas for control subjects it was 1.25-fold (p=0.02). At 6-months, gene expression in the soy group was 1.66-fold that of HSK, compared to 1.31 for the control subjects (p=0.0001). Conclusions: We saw significant increases in soy-induced genes in the breast following a 6-month intervention suggest an estrogenic effect on the breast. However, spontaneous variation in gene expression resulted in statistically non-significant change between groups. The reproducibility of expression of specific genes needs to be considered in design of future studies.
Read full abstract