Identification of VDR-responsive gene signatures in breast cancer cells

Objectives: Defining transcriptional profiles which predict cancer cell anti-proliferative responsiveness towards 1,25dihydroxyvitamin D-3 [1 alpha,25(OH)(2)D-3] is required to improve and tailor the chemotherapeutic application of this secosteroid hormone to individual cancer patients. Methods: We undertook a transcriptomic approach with Affymetrix human U133 GeneChips to determine responsive and resistant gene signatures in MCF-7 breast cancer cells and 1 alpha,25(OH)(2)D-3-resistant MCF-7(Res) cells, respectively. Principal component and hierarchical clustering analyses demonstrated that the patterns of responsiveness between the 2 cell types differed clearly and were used to generate heat maps. Differentially regulated gene targets were validated with Q-RT-PCR and the biological impact upon proliferation measured. Results: In untreated MCF-7 Res cells, 163 genes were up-regulated and 274 down-regulated ( with a log 2 ratio of > 0.5) compared to the MCF-7 controls. Using the same gene expression threshold, 1 alpha,25(OH)(2)D-3 treatment (100 nM, 6 h) of MCF-7 cells up-regulated 91 genes and down-regulated 5, whereas in MCF-7 Res, despite their resistance to the anti-proliferative effects, 156 genes were modulated with 91 being down-regulated. Strikingly, CYP24 was the only induced gene that was common to the genetic profiles of the 2 sets of 1 alpha,25(OH)(2)D-3-treated cells. Heat map analyses defined 2 sub-clusters of genes: (1) basal expression patterns associated with insensitivity towards 1 alpha,25(OH)(2)D-3 and (2) regulated expression patterns associated with 1 alpha,25(OH)(2)D-3 sensitivity. This latter cluster contained BAX, GADD45 alpha, IGFBP-3, EGFR, MAPK4 and TGF-beta(2). Time course studies confirmed the 1 alpha,25(OH)(2)D-3 regulation of TGF-beta(2) in MCF-7 and non-tumourigenic MCF-12A cells but not in MCF-7 Res cells. Co-treatment of MCF-7(Res) cells with exogenous TGF-beta(2) plus 1 alpha,25(OH)(2)D-3 enhanced anti-proliferative and vitamin D receptor transcriptional effects. Conclusions: Basal and regulated gene patterns can be used to predict and monitor the cellular response towards vitamin D-3 compounds and may possibly be applied as a further diagnostic tool.