Thyroid hormone receptor does not heterodimerize with the vitamin D receptor but represses vitamin D receptor-mediated transactivation.

The 9,000 Mr calcium-binding protein calbindin-D9k (CaBP9k) is markedly induced by 1,25-dihydroxyvitamin D3 [1,25-(OH)2D3] in mammalian intestine. However, although a vitamin D response element (VDRE) has been reported in the promoter of the rat CaBP9k gene (at -490/-472), the CaBP9k promoter is weakly transactivated by 1,25-(OH)2D3. Previous studies indicated that when MCF-7 cells are transfected with the rat CaBP9k VDRE ligated to the thymidine kinase promoter and treated with both 1,25-(OH)2D3 and T3 there is an enhancement of the response observed with 1,25-(OH)2D3 alone, suggesting direct cross-talk between thyroid hormone and the vitamin D endocrine system and activation via the formation of vitamin D receptor (VDR)-thyroid hormone receptor (TR) heterodimers. To determine whether the weak response of the rat CaBP9k natural promoter to 1,25-(OH)2D3 could be enhanced by T3, CaBP9k promoter/reporter chloramphenicol acetyltransferase constructs were transfected in MCF-7 cells, and the cells were treated with the two hormones alone or in combination. No induction with T3 alone and no enhancement of reporter activity in the presence of both hormones was observed. To determine whether a lack of effect by T3 was specific for the CaBP9k promoter and to further examine the possibility of cross-talk between the TR- and VDR-signaling pathways, the 1,25-(OH)2D3-responsive rat 24 hydroxylase [24(OH)ase] promoter and the rat osteocalcin VDRE (-457/-430), both fused to reporter genes were similarly examined in MCF-7 cells. Again, no enhancement of the response to 1,25-(OH)2D3 was observed in the presence of T3. In addition, a similar lack of response to T3 but responsiveness to 1,25-(OH)2D3 was observed when UMR106-01 osteosarcoma cells [which, like MCF-7 cells, express VDR, TR, and the retinoid X receptor (RXR) endogenously] were transfected with a 1,25-(OH)2D3 responsive mouse osteopontin promoter reporter. In vitro DNA binding assays were carried out using purified human VDR, human RXRalpha, and chick T3Ralpha and 24(OH)ase, osteocalcin, osteopontin, and CaBP9k VDRE oligonucleotide probes. No VDR-TR heterodimer binding on any of these VDREs was observed, although, as expected, there was binding by the VDR-RXR complex and strong TR-RXR binding to a consensus thyroid hormone response element. Simultaneous gel retardation assays using similar and lower concentrations of TR with RXR showed strong binding of TR-RXR on a 32P-labeled thyroid response element. Studies using the yeast two-hybrid system also did not provide evidence for the formation of a VDR-TR protein-protein interaction. In addition, in vivo data showed that transfection of TR, in fact, repressed VDR-mediated transcription and that the repression could be reversed by the addition of RXR. Thus, in vitro and in vivo experiments do not support ligand-sensitive transactivation mediated by VDR-TR heterodimer formation but rather suggest that TR expression can repress 1,25-(OH)2D3-induced transcription predominantly by sequestering RXR.