Osteoblasts or bone marrow stromal cells are required as supporting cells for the in vitro differentiation of osteoclasts from their progenitor cells. Soluble receptor activator of nuclear factor-kappaB ligand (RANKL) in the presence of macrophage colony-stimulating factor (M-CSF) is capable of replacing the supporting cells in promoting osteoclastogenesis. In the present study, using Balb/c-derived cultures, osteoclast formation in both systems-osteoblast/bone-marrow cell co-cultures and in RANKL-induced osteoclastogenesis-was inhibited by antibody to tumor necrosis factor-alpha (TNF-alpha), and was enhanced by the addition of this cytokine. TNF-alpha itself promoted osteoclastogenesis in the presence of M-CSF. However, even at high concentrations of TNF-alpha the efficiency of this activity was much lower than the osteoclastogenic activity of RANKL. RANKL increased the level of TNF-alpha mRNA and induced TNF-alpha release from osteoclast progenitors. Furthermore, antibody to p55 TNF-alpha receptors (TNF receptors-1) (but not to p75 TNF-alpha receptors (TNF receptors-2) inhibited effectively RANKL- (and TNF-alpha() induced osteoclastogenesis. Anti-TNF receptors-1 antibody failed to inhibit osteoclastogenesis in C57BL/6-derived cultures. Taken together, our data support the hypothesis that in Balb/c, but not in C57BL/6 (strains known to differ in inflammatory responses and cytokine modulation), TNF-alpha is an autocrine factor in osteoclasts, promoting their differentiation, and mediates, at least in part, RANKL's induction of osteoclastogenesis.

The calcium-regulating hormone, 1,25(OH)(2)D(3), induces tumor necrosis factor-alpha (TNF-alpha) synthesis and release from bone marrow macrophages (BMMs). To investigate the mechanism of this regulation, we have examined the effects of 1,25(OH)(2)D(3) on the cytokine message. 1,25(OH)(2)D(3) increased TNF-alpha mRNA abundance in a dose- and time-dependent manner. The combined treatment of BMMs with LPS and 1,25(OH)(2)D(3) resulted in a synergistic increase of TNF-alpha. The steroid also increased the expression of CD14 (LPS receptor). Vitamin D receptors (VDRs) mediate 1,25(OH)(2)D(3) genomic effects by forming homodimers or heterodimers with retinoic acid receptors (RARs) or retinoic X receptors (RXRs). The RXR ligand, 9-cis retinoic acid (9cRA), reduced TNF-alpha mRNA abundance in BMMs, but increased CD14 mRNA levels. 1,25(OH)(2)D(3) or LPS did not affect TNF-alpha transcript stability. 9cRA, however, caused TNF-alpha mRNA destabilization. Next, we searched for potential vitamin D response elements (VDREs) in the promoter region (1.2 kb) of the TNF-alpha gene, and identified six such sequences. Using electrophoresis mobility shift assay (EMSA) we identified one of those sequences (-1008 to -994) as a likely candidate to be a VDRE (tnfVDRE). The binding of tnfVDRE to BMM-derived nuclear extract was increased following cell treatment with 1,25(OH)(2)D(3). No induction was observed with 9cRA treatment, but the retinoid enhanced the activity of 1,25(OH)(2)D(3) when added together. Previously characterized VDREs (mouse osteopontin and rat osteocalcin) competed effectively with tnfVDRE, demonstrating the nature of the TNF-alpha-derived sequence as a VDRE. We observed super-shift and block-shift of the complex in the presence of either anti-VDR or anti-RXR antibodies. Our data suggest that 1,25(OH)(2)D(3) increases TNF-alpha transcript abundance in BMMs via a transcriptional mechanism; 9cRA decreases TNF-alpha mRNA by destabilizing the transcript, and possibly also by forming transcriptionally inactive complex with 1,25(OH)(2)D(3) on the tnfVDRE. The receptor complex interacting with tnfVDRE found in the promoter of the cytokine gene is probably composed of VDR-RXR heterodimer.