Commitment to the Osteoblast Lineage Is Not Required for RANKL Gene Expression

Differentiation of bone-resorbing osteoclasts from hematopoietic precursors depends upon expression of the cytokine receptor activator of NFκB ligand (RANKL) by fibroblastic stromal cells, which some evidence suggests are of the osteoblast lineage. We have shown previously that hormonal-responsiveness of the murine RANKL gene is mediated in part by a distal enhancer that binds Runx2, a transcription factor required for commitment to the osteoblast lineage, supporting the idea that osteoclast-supporting stromal cells may be osteoblasts or their progenitors. However, in this study we demonstrate that parathyroid hormone (PTH) stimulation of RANKL in mice is not affected by a significant reduction in the number of osteoblasts. Consistent with this, neither Runx2, nor Cbfb, a binding partner essential for Runx activity, are required for basal or PTH-stimulated RANKL expression in fibroblastic stromal cell models. Nonetheless, RANKL responsiveness to PTH was elevated in cultured calvaria cells expressing high levels of osterix, another transcription factor required for osteoblast differentiation, and this was associated with elevated PTH receptor expression. The responsiveness of RANKL to 1,25-dihydroxyvitamin D₃ was not elevated in the osterix-expressing cells. Together, these results suggest that commitment to the osteoblast lineage is not a requirement for RANKL gene transcription in fibroblastic stromal cells but may enhance responsiveness of this gene to specific hormones via control of their receptors.In the adult skeleton, bone is constantly renewed via the coordinated activity of osteoclasts that resorb bone and osteoblasts that form bone. These cells function within an anatomically distinct structure known as the basic multicellular unit (BMU), in which the osteoclasts are located in the lead and are followed by osteoblasts (1). Because of this organization, bone formation occurs only at sites of prior bone resorption and the recruitment of osteoblasts to such sites is known as coupling. While the mechanisms that underlie coupling are unknown, two different explanations have been proposed. According to the first, release of factors, such as TGFβ, from the bone matrix as a consequence of osteoclast activity recruits osteoblast progenitors and promotes their differentiation (2). Thus, osteoblastogenesis in this “serial” pathway of coupling is a consequence of osteoclastogenesis (3). However, in view of the fact that osteoclastogenesis depends upon support from stromal cells that may be of the osteoblastic lineage, the existence of a parallel pathway has been proposed (4). According to the parallel pathway model, osteoclast and osteoblast differentiation occur simultaneously due in part to the requirement of osteoblast lineage cells for osteoclast differentiation.The idea that osteoblast lineage cells are required for osteoclast differentiation originated from studies showing that osteoblasts or osteoblast-like cells, not osteoclast precursors, are targets of hormones that stimulate bone resorption (5–7). It has since been demonstrated that these hormones stimulate osteoclast differentiation by acting directly on stromal cells to stimulate expression of receptor activator of NFκB ligand (RANKL),³ suppress expression of the RANKL antagonist osteoprotegerin (OPG), or both (8, 9). However, whether the stromal cells that are the targets of these hormones are in fact osteoblasts or osteoblast precursors is unclear. One reason for this uncertainty is that the calvaria and bone marrow stromal cell cultures commonly used to study osteoblastic cells contain many cell types, including fibroblastic cells that may not be of the osteoblastic lineage (10). In addition, analysis of RANKL expression during osteoblast differentiation in vitro has produced conflicting results with differentiation both promoting (11) and inhibiting (12) RANKL expression. Attempts to identify RANKL-expressing cells in remodeling bone using histologic methods have also produced inconsistent results (8, 13–15). More importantly, conditional ablation of osteoblasts in transgenic mice did not alter osteoclast numbers or bone resorption (16), and many mouse models with increased osteoblast number do not exhibit increased osteoclast number (17–20). Therefore, it remains unclear whether matrix-synthesizing osteoblasts or their precursors are required for the support of osteoclast differentiation. This uncertainty is reflected by the use of deliberately vague terms, such as “stromal” or “stromal/osteoblastic,” when referring to osteoclast support cells.Whatever their lineage may be, the cells that support osteoclast formation do so by expressing RANKL, which is indispensable for osteoclast formation in vivo (21). To gain insight into the biology of osteoclast support cells, we have sought to understand the mechanisms that control the cell type-specific expression of the murine RANKL gene. We identified a transcriptional enhancer that mediates hormonal control of RANKL in stromal cells (22, 23). This enhancer, designated the RANKL distal control region (DCR), is located 76-kb upstream from the transcription start site. Gel shift and chromatin immunoprecipitation (ChIP) assays revealed that the DCR contains a binding site for runt related transcription factor 2 (Runx2), a transcription factor that is essential for osteoblast differentiation (24, 25). Furthermore, deletion of the Runx2 binding site blunted the hormonal responsiveness conferred by the DCR (22). These results suggested that Runx2 may be a factor linking osteoclast formation to osteoblast formation via cell type-specific control of RANKL expression.Consistent with the idea that Runx2 is required for RANKL expression in stromal cells, Runx2-deficient mice exhibit a reduced number of osteoclasts (24). Moreover, calvaria cells from Runx2-deficient mice were less capable of supporting osteoclast formation in vitro (26). In contrast, expression of a dominant negative Runx2 protein in a stromal cell line did not alter basal or stimulated RANKL expression (27). In addition, cell lines derived from Runx2-deficient mice or cell lines in which Runx2 was suppressed by RNA interference still expressed RANKL in response to signaling pathways activated by hormones that stimulate bone resorption (28, 29). It has also been recently proposed that Runx2 may exert an inhibitory effect on basal RANKL expression by condensing chromatin and thus reducing transcription (29). Conflicting results were also obtained in two similar transgenic mouse models overexpressing Runx2 in osteoblasts, which showed either an increase in RANKL expression and osteoclast number (30) or no change in these measurements (31). Thus, as with the identity of the cells that produce RANKL, the role of Runx2 in RANKL expression is unclear.In the present study we investigated the relationship of osteoblasts and their precursors to the cells that support osteoclast differentiation via expression of RANKL. We found that both mature osteoblasts and Runx family proteins are dispensable for RANKL expression. In addition, enrichment of cells committed to the osteoblast lineage was not associated with an increased ability to express RANKL, although it was associated with increased responsiveness to PTH.