IBMS BoneKEy | Commentary
Osteocytes communicate with osteoclast lineage cells via RANKL
Koichi Matsuo
DOI:10.1038/bonekey.2012.36
Commentary on: Nakashima T, Hayashi M, Fukunaga T, Kurata K, Oh-Hora M, Feng JQ, Bonewald LF, Kodama T, Wutz A, Wagner EF, Penninger JM, Takayanagi H. Evidence for osteocyte regulation of bone homeostasis through RANKL expression. Nat Med. 2011; 17 (10):1231-1234 and Xiong J, Onal M, Jilka RL, Weinstein RS, Manolagas SC, O'Brien CA. Matrix-embedded cells control osteoclast formation. Nat Med. 2011; 17 (10):1235-1241.
Using conditional knockout mice, two groups recently reported phenotypes of mice with osteocyte-specific deletion of the receptor activator of nuclear factor-κB ligand (RANKL), the cytokine essential for osteoclast differentiation. Both groups, Nakashima et al. and Xiong et al., found that mice lacking RANKL (encoded by Tnfsf11) in osteocytes develop postnatal osteopetrosis, indicating that matrix-embedded osteocytes induce RANKL-dependent osteoclast formation in bone remodeling. These are indeed paradigm-shifting studies.
Previously, many of us believed that osteoblasts or stromal cells are the major source of RANKL throughout life, although many other cell types including T lymphocytes and hypertrophic chondrocytes express RANKL. The prevailing idea that osteoblasts direct osteoclastogenesis has been reinforced by widely used co-culture methods in which osteoclast formation from hematopoietic precursor cells is supported by 'calvarial osteoblasts', which in hindsight most likely contain osteocytes. Nakashima et al. observed high RANKL expression in both primary cells purified by alternate collagenase-EDTA extractions and dentin matrix protein 1 (Dmp1)-positive osteocytes purified from mouse calvarial and cortical bone by cell sorting. They demonstrated that purified osteocytes more potently support osteoclastogenesis than do osteoblasts and bone marrow stromal cells in co-culture. Nakashima et al. proposed that osteocytes derived from osteoblasts might provide RANKL to osteoclast precursor cells. This laboratory had been interested in whether T lymphocytes are a major RANKL source under pathological conditions, such as in joints of rheumatoid arthritis patients. LoxP sites were introduced flanking Tnfsf11 exons 3 and 4, and 'floxed' mice were bred with mice expressing Cre recombinase driven by the osteocyte-specific Dmp1 promoter or the T-cell-specific Lck promoter. Although the T-cell-specific knockout mice showed no phenotype under physiological conditions, osteocyte-specific knockout mice developed osteopetrosis postnatally.
Independently, O'Brien and co-workers generated an allele carrying floxed exons 3 and 4 of Tnfsf11, and crossed resultant mice with various 'deleter mice', such as Prx1-Cre (for excision in early limb bud mesenchyme and in a subset of craniofacial mesenchyme, but not in the axial skeleton), Osterix 1 (Osx1)-Cre (early osteoblasts and late-proliferating and hypertrophic chondrocytes), Osteocalcin (Ocn)-Cre (mature osteoblasts and hypertrophic chondrocytes), Dmp1-Cre (osteocytes) and collagen X (ColX)-Cre (hypertrophic chondrocytes). Cre-recombinase-mediated deletion is irreversible; once a gene is deleted in the osteoblasts, osteocytes derived from these osteoblasts carry deleted alleles. Following careful analysis of phenotypes of various conditional knockout mice, both Nakashima et al. and Xiong et al. found that osteocyte-specific RANKL deletion resulted in postnatal osteopetrosis, suggesting that osteocytes are a major source of RANKL during bone remodeling. Xiong et al. demonstrated that Dmp1-Cre-mediated deletion occurs not only in osteocytes, but also in osteoblasts. However, they also showed that RANKL deletion mediated by doxycycline withdrawal over a period of 2 months in osteoprogenitors of adult Osx1-Cre mice (via the Tet-off system) did not alter bone remodeling, suggesting that osteoprogenitor RANKL does not contribute to bone remodeling. By contrast, in bone modeling characterized by bone growth associated with changes in shape during development, resorption of the primary spongiosa is dependent on RANKL produced by mature osteoblasts or hypertrophic chondrocytes, as illustrated by unresorbed calcified cartilage seen in conditional RANKL-knockout mice carrying either Osx1-Cre, Ocn-Cre or ColX-Cre. Thus, for the first time, osteoblast/hypertrophic cartilage-derived RANKL has been linked to osteoclastogenesis during 'bone modeling', whereas osteocyte-derived RANKL is associated with 'bone remodeling' in maintenance of bone.
What is the significance of RANKL production by osteocytes in bone remodeling, particularly in coupling of bone resorption and formation? Osteocytes communicate with osteoblasts and negatively regulate osteoblast differentiation by secretion of sclerostin (encoded by Sost) through dendrites extended into canaliculi. Sclerostin is a potent inhibitor of osteoblast differentiation, which binds to LRP5/6 receptors and antagonizes Wnt signaling. As it is now clear that osteocytes communicate with osteoclast lineage cells through RANKL, both osteoclasts and osteoblasts are likely controlled by osteocytes during bone remodeling (Figure 1). On the other hand, osteoclast-derived hypothetical 'coupling factors' facilitate transition from bone resorption to bone formation. How is this three-part relationship regulated to maintain bone integrity? It is plausible that bone resorption in the absence of bone formation occurs when osteocytes produce both RANKL and sclerostin at high levels. When coupling occurs, osteocytes likely stop producing both to allow the transition from bone resorption to formation. Further studies will be required to determine whether osteoclast-derived coupling factors, which by definition activate osteoblastic bone formation, regulate this switching in osteocytes.
Numerous interesting questions for future study emerge from these highly stimulating papers.
How is RANKL expression in osteocytes induced by unloading? Xiong et al. demonstrated that RANKL mRNA levels and osteoclast number were elevated in control but not conditional Dmp1-Cre mice, based on 3 weeks of tail suspension experiments. If osteocytes sense mechanical loading in the form of fluid-flow shear stress, how does loss of fluid-flow shear stress result in elevated RANKL expression? As sclerostin stimulates RANKL expression in osteocytes, and unloading increases osteocyte expression of sclerostin, unloading may induce RANKL expression through sclerostin. Is there concerted regulation of RANKL expression in multiple interconnected osteocytes by dendrites? Conversely, do multiple dendrites from a single osteocyte deliver different levels of RANKL depending upon orientation or geometry of the lacuna-canalicular system? Technology is currently being developed to analyze the lacuna-canalicular system three-dimensionally by X-ray imaging, which can be combined with immunogold staining of RANKL.Do parathyroid hormone, 1,25-dihydroxyvitamin D3, bisphosphonate or other molecules regulate osteocytic RANKL expression in vivo? Do osteocytes express RANK for RANKL to function in an autocrine manner? If so, what is the function of this autocrine loop?There may be spatial, temporal and species variations in osteocyte RANKL production. How do osteocytes differ between cortical and cancellous bone in terms of RANKL expression? Is RANKL not expressed in osteocytes in developing animals? Do new and old osteocytes in adult animals express similar amounts of RANKL? A simple hypothesis is that old osteocytes function to induce osteoclastogenesis to replace old bone. If so, how do aged osteocytes produce more RANKL? Does RANKL expression by apoptotic osteocytes regulate osteoclast precursor differentiation? Do osteocytes ever dedifferentiate into osteoblasts ex vivo? How is RANKL production regulated in 'acellular' (lacking osteocytes) bone during bone remodeling, as occurs in some teleost fish such as medaka?Which among two membrane-bound forms of RANKL and a secreted form are produced by osteocytes? How does RANKL reach the bone surface? At least in osteoblastic cells, RANKL is localized to secretory lysosomes and stimulation with RANK-Fc-conjugated beads promotes its translocation to the cell surface mediated by Rab 27 small GTPases.
The regulatory system depicted in Figure 1 is further complicated by osteoprotegerin (OPG), a soluble decoy receptor for RANKL that blocks osteoclast formation by inhibiting binding of RANKL to RANK. As OPG is produced by osteocytes as a target of Wnt signaling, it is reasonable to assume that when sclerostin (and RANKL) production is reduced, OPG expression is elevated. Osteocytes are a source of constant surprise to bone biologists. Previously, we were astonished when we learned that osteocytes are hormone-producing cells that communicate with kidneys at a distance by secreting the hormone fibroblast growth factor 23, which regulates phosphate metabolism. What will be the next unanticipated finding in bone metabolism? Major bone resorption by osteocytes under certain conditions? Morphogenesis controlled by osteocytes? Osteocyte-targeted therapy? Whatever happens, osteocytes' waving arms serve to attract many bone scientists.
This work is licensed under a
Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.