D-type cyclins (cyclins D1, D2, and D3) promote G1-S progression and are aberrantly expressed in cancer. We reported previously that all-trans-retinoic acid chemo-prevented carcinogenic transformation of human bronchial epithelial (HBE) cells through proteasomal degradation of cyclin D1. Retinoic acid is shown here to activate distinct mechanisms to regulate different D-type cyclins in HBE cells. Retinoic acid increased cyclin D2, decreased cyclin D3 and had no effect on cyclin D1 mRNA expression. Retinoic acid decreased cyclin D1 and cyclin D3 protein expression. Repression of cyclin D3 protein preceded that of cyclin D3 mRNA. Proteasomal inhibition prevented the early cyclin D3 degradation by retinoic acid. Threonine 286 (T286) mutation of cyclin D1 stabilized cyclin D1, but a homologous mutation of cyclin D3 affecting threonine 283 did not affect cyclin D3 stability, despite retinoic acid treatment. Lithium chloride and SB216763, both glycogen synthase kinase 3 (GSK3) inhibitors, inhibited retinoic acid repression of cyclin D1, but not cyclin D3 proteins. Notably, phospho-T286 cyclin D1 expression was inhibited by lithium chloride, implicating GSK3 in these effects. Expression of cyclin D1 and cyclin D3 was deregulated in retinoic acid-resistant HBE cells, directly implicating these species in retinoic acid response. D-type cyclins were independently targeted using small interfering RNAs. Repression of each D-type cyclin suppressed HBE growth. Repression of all D-type cyclins cooperatively suppressed HBE growth. Thus, retinoic acid repressed cyclin D1 and cyclin D3 through distinct mechanisms. GSK3 plays a key role in retinoid regulation of cyclin D1. Taken together, these findings highlight these cyclins as molecular pharmacologic targets for cancer chemoprevention.