Genes Involved in Osteoclastogenesis

Abstract

Osteoclast formation is a complex process requiring the temporal activation of a yet unknown number of transcription factors. Osteoclast differentiation is dependent on two cytokines: macrophage colony stimulating factor (M-CSF) and receptor activator of NF-kB ligand (RANKL). These agents induce gene expression changes during the differentiation process, presumably by inducing transcription factors. A search for genes that are regulated in the developing osteoclast was performed using both differential display and gene arrays. Differential display revealed a novel member of the krüppel-associated box (KRAB) containing transcription repressor, KROCS, which was shown to be down regulated during osteoclast formation. The potential targets for this gene remain unknown, as do the targets of the majority of the other members of the krüppel associated box (KRAB) containing family of transcription repressors. In addition to KROCS, three other genes were also identified including ADO21, PRO1859 and an endogenous retrovirus like gene and the regulation of vitamin D up-regulated protein (VDUP) was also confirmed. Array analysis identified a number of other transcription factors regulated during osteoclast formation including the up-regulated NFATc1, GABP?, FBP, EGR1 and the repressed RelB and KOX31, a KRAB containing transcription repressor. The array also identified calmodulin 1 a member of the NFAT activation pathways as up-regulated by RANKL. The expression of NFATc1 to 4 in human osteoclasts was investigated showing NFATc1 to be the most expressed NFATc in osteoclasts. The transcription variants of NFATc1 were tested for expression differences showing that the mRNAs encoding the protein isoforms B and C were most expressed. The involvement of calmodulin, calcineurin and NFATc1 involvement in osteoclast formation was further studied by the use of inhibitors. BAPTA-AM is an intracellular chelator of calcium that prevents changes in calcium concentration. Phenoxybenzamine irreversibly binds calmodulin in the presence of calcium, inhibiting the action of calmodulin. Cyclosporin A (CsA) is an inhibitor of calcineurin. Use of both BAPTA-AM and phenoxybenzamine resulted in inhibition of osteoclast formation, decreasing the percentage of multinucleated cells from 54% in control cultures to 7.9% and 7.1% respectively. Both BAPTA-AM and phenoxybenzamine treated cells showed a marked reduction in TRAP activity with only 14.5% and 16.8% respectively staining positive for TRAP. This represents an approximate 60% reduction in TRAP positive cells compared to control osteoclasts. Both BAPTA-AM treated and phenoxybenzamine treated cells were negative for bone resorption. Addition of increasing doses of cyclosporin A (CsA) to M-CSF and RANKL treated cells resulted in the inhibition of multinucleated osteoclast formation. At 1000ng/mL CsA the formation of TRAP positive cells with more than one nucleus had reduced to less than 5% from 54% without the presence of CsA. Cells treated with 1000ng/mL CsA were unable to resorb bone, however the percentage of cells that were TRAP positive remained unchanged with CsA treatment. No significant decrease in expression of cathepsin K or TRAP transcripts were observed by real-time quantitative PCR (Q-PCR) in cells treated with CsA. Although all three agents inhibited the formation of multinuclear giant cells, both BAPTA-AM and phenoxybenzamine resulted in TRAP negative cells, whereas CsA resulted in TRAP positive cells. These results implicate the intracellular calcium increase caused by RANKL and calmodulin activation as a regulator of TRAP but place the calcineurin activation of NFATc1 downstream of TRAP induction. The regulation of a series of other genes was tested to determine if some RANKL mediated regulation of osteoclast genes were 'sensitive to CsA while others were 'resistant'. Of 28 genes tested, 13 were significantly affected by CsA and were considered 'sensitive' while the RANKL mediated regulation of 15 genes was unaffected by CsA and these were considered 'resistant'. This is strong evidence for two pathways of gene activation in osteoclasts, a CsA 'sensitive' pathway involving calcineurin, NFAT and possibly other transcription factors and a CsA 'resistant' pathway of gene activation, not dependent on calcineurin. Surprisingly, the RANKL mediated induction of NFATc1 was not inhibited by CsA, suggesting that NFATc1 induction is dependent on the resistant pathway. The identity of the second pathway (or pathways) is yet to be established, however the data indicate that this pathway mediate the RANKL sensitive regulation of at least one half of genes in human osteoclasts. The corollary if that only one half of osteoclast genes are dependent on calcineurin and presumably NFATc1 activation. There was no unifying principle that separated the CsA resistant from sensitive pathways of RANKL regulation. Cell surface markers, chemokines and transcription factors were among those affected by CsA. Even classical osteoclast markers fell neatly into two categories. The RANKL mediated induction of calcitonin receptor (CalcR) was inhibited by more than 100 fold in the presence of CsA implicating NFAT/calcineurin in the regulation of CalcR expression in osteoclasts. In contrast, the RANKL mediated induction of TRAP or cathepsin K, two prominent osteoclast markers, was totally unaffected by CsA. The expression of a series of chemokines and receptors was investigated. MCP-1 and RANTES were RANKL induced, and this induction was sensitive to CsA. The CC chemokines MCP-1 and RANTES were down regulated by around 10 fold in the presence of CsA. In contrast the RANKL mediated induction of MCP-1 receptor was resistant to CsA. The existence of chemokine and receptor in the same cell provides for a RANKL inducible autocrine loop, suggesting that MCP-1 should act directly on osteoclasts. The fact that the RANKL induction of the MCP-1 receptor, CCR2B, is unaffected by CsA suggests that exogenous MCP-1 should still signal in CsA treated osteoclasts. Addition of either MCP-1 or RANTES to CsA treated cultures resulted in a recovery of 70-80% of the multinuclear TRAP positive phenotype. The MCP-1 and RANTES induced multinuclear cell could not overcome the CsA induced inhibition of bone resorption. Surprisingly, MCP-1 and RANTES induced multinucleation in the absence of RANKL (M-CSF and chemokine treated cells) resulting in 50% of the normal multinucleation present in cells treated with RANKL. The data suggest that chemokines produced by osteoclasts are involved in promoting a multinuclear phenotype. When inhibited by CsA, osteoclasts fail to produce both MCP-1 and RANTES, although their respective receptors are present. This failure to produce MCP-1 and RANTES prevents the formation of an autocrine loop. When provided with MCP-1 or RANTES the CsA inhibited osteoclasts are subsequently able to pass through to the stage of a multinucleated giant cell. Similarly, in the absence of RANKL, chemokines promote the formation of TRAP positive osteoclast-like giant cells visually indistinguishable from osteoclasts. However, the multinuclear cells formed by chemokines in the absence of RANKL were also incapable of bone resorption. In order to determine if chemokines were capable of stimulating bone resorption, after osteoclasts had formed, pre-differentiated mature osteoclasts were plated onto bone and treated with a range of cytokines. The results showed that bone resorption occurred only in cultures that were exposed continuously to RANKL. These data indicate that chemokine induction by RANKL is required for multinucleation but that RANKL is required for bone resorption. The functional testing of genes detected by array analysis proved crucial involvement of both the NFAT pathway and CC chemokines in osteoclast formation knd function. Other genes identified such as GABP and FBP, are likely to be key factors in the development of a functional osteoclast. Future works investigating human osteoclast formation should take into strong consideration the genes identified in this thesis as targets for further functional studies.