Introduction and Objective: Diets rich in flavonoids are associated with reduced risk of prostate cancer (PCa). Despite the large number of different flavonoids (over 4000) only a few have been studied in PCa to date. Based on a previous screening of 30 flavonoids in our lab, we identified two novel flavonoids- 2,2’-dihydroxychalcone (2,2’-DHC) and fisetin with potent anti-proliferative properties in PCa cell lines. Our objective in this study was to determine molecular mechanisms of action of 2,2’-DHC and fisetin by examining global gene expression effects in human PCa cell lines using cDNA microarray. Methods: LNCaP and PC3 human prostate cancer cells were exposed to 15μM 2,2’-DHC, 25μM fisetin, or vehicle control (0.1% DMSO) for 6, 12 and 24h. cDNA from each sample (including DMSO control) was labeled using cyanine5 (cy5) and hybridized onto a custom made 2-color microarray slide spotted with 22,000 genes. cDNA obtained from universal human RNA sample was labeled with cyanine 3 (cy3) and hybridized onto all slides. The gene expression level of all samples from the scanned slides was normalized using the LOWESS method. 2-fold changed genes were identified and ANOVA was used to determine statistically significant gene expression alterations between groups. K-means cluster analysis and biological data mining using the GO database was performed. Quantitative RT-PCR and Western blotting for 10 genes was performed to validate the results of the microarray. Results: A total of 736 significantly altered genes were identified in the analysis. Classification of the genes by gene ontology (GO) showed that key genes involved in cell cycle and mitosis were down-regulated by both flavonoids at all time points. Among this group were the G2 checkpoint proteins cyclin B1, cdc2 and cdc25A, while the cyclin dependent kinase inhibitor p21 was consistently up-regulated. Furthermore, a number of spindle checkpoint genes showed reduced expression, including bub1 and plk1. Changes in known mitogenic pathways included reduction in levels of hepatocyte growth factor and its receptor, and reduction of IGF1 receptor. Finally a number of novel genes such as TWIST1 and sestrin were highly differentially regulated. K-means clustering by expression pattern demonstrated that the majority of cell cycle genes were being modulated by flavonoids in a similar fashion, as the majority of these genes clustered together. The results of quantitative RT-PCR and Western blotting performed for 10 genes were in direct agreement with the microarray data. Conclusions: Flavonoid treatment results in alterations in multiple cellular pathways that are important in prostate carcinogenesis. Modulation of such pathways may explain many of the beneficial anti-cancer properties associated with flavonoids.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]