INHIBITION OF 1,25-DIHYDROXYVITAMIN D-3 STIMULATED OSTEOCALCIN GENE-TRANSCRIPTION BY TUMOR-NECROSIS-FACTOR-ALPHA - STRUCTURAL DETERMINANTS WITHIN THE VITAMIN-D RESPONSE ELEMENT

Control of osteoblast function requires the coordinate activity of sys temic and local regulatory factors. We have investigated the mechanism of interaction between the secosteroid 1,25-dihydroxyvitamin D-3 [1,2 5-(OH)(2)D-3] and the cytokine tumor necrosis factor-alpha (TNF-alpha) by measuring their effects on two 1,25-(OH)(2)D-3 responsive matrix p rotein genes, osteocalcin (OC) and osteopontin (OF). Our previous stud ies revealed that an inhibitory effect of TNF-alpha on 1,25-(OH)(2)D-3 -stimulated OC gene transcription is conferred by the same 25 base pai r region of 5'-flanking DNA that confers a response to vitamin D (VDRE ). Gel mobility shift studies of [P-32]VDRE binding to ROS 17/2.8 cell nuclear extract revealed that TNF-alpha inhibits 1,25-(OH)(2)D-3 stim ulated formation of specific retinoid X receptor/vitamin D receptor (R XR/VDR)-DNA complexes in vitro. To determine if TNF-alpha was inhibiti ng nuclear protein-VDRE binding by modulation of VDR availability, we measured intranuclear VDR in cells treated with 1,25-(OH)(2)D-3 (10(-6 ) M), TNF-alpha (100 ng/ml), or both, by western blot. 1,25-(OH)(2)D-3 caused upregulation of the nuclear VDR Treatment with TNF-alpha inhib ited the 1,25-(OH)(2)D-3-stimulated up-regulation of VDR nuclear prote in content. However, down-regulation of VDR was unlikely to be the mec hanism of TNF-alpha action because TNF-alpha had no effect on 1,25-(OH )(2)D-3 stimulation of steady state OP messenger RNA or transcription of an OP-VDRE-chloramphenicol acetyl transferase reporter construct. T hese results suggest that decreased VDR alone does not explain the mec hanism of TNF-alpha action. VDRE structural requirements for TNF-alpha action were characterized by comparing binding of mutant and hybrid f orms of mouse (m)OP-, rat (r)OC-, and human (h)OC-VDRE probes to nucle ar protein from cells treated with 1,25-(OH)(2)D-3 and/or TNF-alpha. T hese homologous vitamin D response elements differ in that an AP-1 seq uence is included in the rOC-VDRE and hOC-VDRE but not in the OP-VDRE. Gel mobility shift analysis revealed that TNF-alpha inhibited 1,25-(O H)(2)D-3 stimulation of nuclear protein binding to rOC-VDRE and hOC-VD RE to 59% and 69% of control, respectively, but had no effect on 1,25- (OH)(2)D-3 stimulation of nuclear protein binding to OP-VDRE. The effe ct of TNF-alpha could not be conferred in a mutant OP-VDRE in which th e rOC-VDRE AP-1 sequence was inserted. Similarly, the effect of TNF-al pha on the hOC-VDRE could not be prevented by randomization of the hOC -VDRE AP-1 sequence, suggesting that the structural specificity of OC- VDRE response to TNF-alpha was exclusive of the activator protein-1 se quence. Two hybrid VDREs were synthesized in which the [5']/[3'] half- sites were derived from [rOC]/[mOP] or [mOP]/[rOC]. When these hybrid VDRE were used as probes, the inhibitory effect of TNF-alpha on 1,25-( OH)(2)D-3 stimulated nuclear protein binding was only observed when th e rOC-VDRE 5' half site GGGTGA was used. This sequence is common to th e rOC and hOC-VDRE and therefore, confers the response to TNF-alpha in both of these hormone response elements. These results: 1) confirm th at TNF-alpha inhibits transcription of the rOC gene by decreasing 1,25 -(OH)(2)D-3 stimulated RXR/VDR binding to the VDRE, 2) suggest that mo dulation of VDR availability, per se, does not explain the inhibitory effect of TNF-alpha on 1,25-(OH)(2)D-3-stimulated gene transcription, 3) show that the inhibitory effect of TNF-alpha on RXR/VDR binding to the VDRE is exclusive of the AP-1 sequence, 4) identifies the 5' half- site GGGTGA as necessary to confer the effect of TNF-alpha.