Osteocytes: Regulating the Mineral Reserves?

Abstract

Osteocytes are increasingly recognized as multifunctional cells that help orchestrate bone and mineral metabolism—and are currently one of the hottest topics in the field. In addition to their putative role as detectors/transducers of mechanical strain in bone, osteocytes are a major source of sclerostin, a key inhibitor of Wnt signaling and bone formation, and appear to be important regulators of osteoclast function via their production of NF-κB ligand (RANKL) and osteoprotegerin. It has also long been believed that osteocytes could remove or even replace mineralized bone around their lacunae and canaliculi. Thanks to a number of studies and reviews in recent years,1-9 this once-controversial concept of “osteocytic osteolysis” is now very much back on the agenda. In this issue of the Journal of Bone and Mineral Research, Kogawa and colleagues10 present evidence that sclerostin could act to stimulate the dissolution of bone mineral by osteocytes by increasing their expression of some of the same resorptive machinery used by osteoclasts. The authors show that treatment of cultured MLO-Y4 and primary human osteocyte-like cells with recombinant human sclerostin increased the expression of mRNA for a number of resorption markers, most notably carbonic anhydrase 2 (CA2). The authors found that sclerostin caused both intracellular and extracellular acidification in cultures of the osteocyte-like cells, albeit in a non–dose-dependent manner. The classic CA inhibitor, acetazolamide decreased the sclerostin-stimulated extracellular acidification in MLO-Y4 cells, whereas antisense knockdown of CA2 led to an extracellular alkalinization, with elimination of the effect of acetazolamide. Sclerostin additionally caused a slight stimulation of calcium release from mineralized layers of primary osteocyte-like cells, an effect that was also blocked by acetazolamide. Stimulation of mRNA for CA2, tartrate-resistant acid phosphatase (TRAP), chloride channel 7 (CLCN7), and cathepsin K by sclerostin in the MLO-Y4 cell line was inhibited by small interfering RNA (siRNA) constructs for Lrp4, Lrp5, or Lrp6, suggesting the possible involvement of Wnt signaling. Last, sclerostin treatment increased the plan area of osteocyte lacunae by about 25% in cultured samples of trabecular bone taken from patients undergoing hip surgery, an effect that was again abrogated by acetazolamide. The concept of osteocytic osteolysis was first mooted in 19th century, although is perhaps now most closely associated with Bélanger.11 He suggested that either parathyroid hormone (PTH) or a low-calcium diet can induce osteocytes to enlarge their lacunae. In its broadest sense, the term osteocytic osteolysis implies dissolution of not only the mineral component of bone adjacent to lacunae and canaliculi but of the collagenous matrix, too. These ideas provoked vigorous rebuttals12, 13 but began to gain traction again following reports that osteocyte lacunar area was markedly increased in rats continuously infused with PTH for 4 weeks1 and in mice infused with prednisolone for 3 weeks.2 More modest increases in lacunar size were also observed in mice subjected to ovariectomy.2 More recently, some of the coauthors of the paper by Kogawa and colleagues10 have presented a comprehensive study of the effects of lactation on osteocyte lacunae in mice.7 Lactation was shown to be associated with increases in lacunar area of about 20–40% in tibias and lumbar vertebrae; these increases were reversed with 7 days of cessation of lactation. The effects were mimicked by PTH-related protein (PTHrP) and abrogated by targeted deletion of PTH receptor 1. Increased expression of mRNAs for osteoclast-specific markers (including cathepsin K, which cleaves collagen) was also noted in “osteocyte-enriched” bones of lactating animals. These reversible changes in osteocyte lacunar size have been referred to as “remodeling,”7, 9 although this process clearly differs from bone remodeling by osteoclasts and osteoblasts, which involves obvious turnover of substantial quantities of both mineral and organic matrix. Despite suggestions that removal and replacement of collagenous matrix around osteocytes might occur in extreme pathophysiological situations, such as in the ultra-high turnover bone of egg-laying hens,3 there seems to be little wider evidence for such a process. Removal of mineral alone might be sufficient to account for the enlargement of osteocyte lacunae observed by light or scanning electron microscopy if the resulting thin layer of demineralized collagenous matrix shrinks back against the lacunar walls during dehydration for specimen preparation. In vivo, such a layer of demineralized (hydrated) matrix adjacent to the osteocyte could presumably serve as a scaffold to facilitate subsequent rapid mineral reaccretion. One potential histological approach to testing whether mineral alone or both mineral and matrix are removed during osteocytic osteolysis might be to study high-quality frozen bone sections (or undecalcified resin-embedded sections) stained with toluidine blue, which differentiates mineralized matrix and osteoid.14 The enlargement of osteocyte lacunae in sclerostin-treated bone samples described by Kogawa and colleagues10 is consistent with size changes reported for stimulation by lactation, PTH/PTHrP and prednisolone.1, 2, 7 Although these changes (about 20–40% in terms of cross-sectional area) are apparently modest, they are potentially important because osteocytes comprise at least 90% of the cells in bone, and each osteocyte may throw out up to 50 canalicular processes.6 The surface area of the osteocyte lacunocanalicular system and its contact with mineralized bone matrix is estimated to be several orders of magnitude greater than the area of the macroscopic bone surface.4 Thus, osteocytes are uniquely placed to regulate bone mineral. As Bonewald4 points out, removal of only a few angstroms of mineral per osteocyte might be expected to have significant effects on circulating, systemic levels of calcium and phosphate. Following on from the initial report of CA2 expression by osteocytes,7 the work of Kogawa and colleagues10 highlights the potential role played by carbonic anhydrase in the regulation of mineral solubility by osteocytes. This enzyme, which, as the authors acknowledge, is widely expressed by cells, catalyses the formation of carbonic acid from carbon dioxide and water, and is thought be the key source of the protons used by osteoclasts to dissolve bone mineral. It should be noted, though, that the evidence presented for stimulation by sclerostin of CA2 protein (as opposed to mRNA) in the osteocyte-like cells was equivocal. Whether proton excretion due to the basal metabolic activity of osteocytes (ie, not specifically involving CA2) could also play a role in regulating pericellular mineralization remains to be determined. In addition to protons, a number of other inorganic factors produced by osteocytes such as CO2, bicarbonate, and pyrophosphate (PPi) could also directly influence mineralization. PPi, a potent, ubiquitous physicochemical inhibitor of mineralization that functions as the body's “natural water softener,”15 may be of particular interest in this regard.6 Much of the PPi in the extracellular compartment is likely to be generated at the cell surface by the action of ecto-nucleotide pyrophosphatase/phosphodiesterases, which liberate PPi from ATP.15, 16 The osteocyte-like MLO-Y4 cell line, in common with many cell types, has been shown to release ATP to the extracellular environment.17 As Atkins and Findlay6 note in their excellent recent review, there are numerous protein factors made by osteocytes that are likely to play roles in modulating pericellular mineralization. These include members of the small integrin-binding ligand, N-linked glycoprotein (SIBLING) family such as matrix extracellular phosphoglycoprotein (MEPE), osteopontin (OPN), the enzyme PHEX (phosphate-regulating gene with homologies to endopeptidases on the X-chromosome), dentine matrix protein 1 (DMP1), and bone sialoprotein. Several SIBLINGs, including MEPE and OPN contain an acidic serine and aspartate-rich peptide sequence (ASARM) that, when cleaved from the parent protein and phosphorylated, binds hydroxyapatite crystals, inhibiting further mineral deposition. An earlier report from Atkins and colleagues5 indicated that sclerostin could also inhibit mineralization in primary human osteoblast cultures through a MEPE-ASARM–dependent mechanism. Kagawa and colleagues10 emphasize that the full spectrum of sclerostin's mode of action on bone is still being unraveled. Previous work from the same group was suggestive of sclerostin effects that did not involve Wnt signaling.5, 18 However, the siRNA knockdown results presented in the present paper are consistent with roles for LRP4, LRP5, and LRP6 in mediating the stimulatory effect of sclerostin on expression of mRNAs for CA2, cathepsin K, and TRAP, at least in the MLO-Y4 cell line. The authors propose that osteocytes, which appear to be a major source of sclerostin, are themselves important targets for it. This idea is based not only on the present work but on another recent study, showing that sclerostin increased the ratio of RANKL:osteoprotegerin (OPG) mRNA expressed by cultured osteocyte-like cells and stimulated osteoclastogenesis in cocultures.18 Understanding the regulation of this type of autocrine/paracrine feedback may well pose some challenges. It should be emphasized that not all of the results presented by Kagawa and colleagues10 are supportive of a significant role for sclerostin in regulating mineralization around osteocytes. The authors also mention briefly that measurements in 14-week-old mice overexpressing human SOST showed no change in osteocyte lacunar area in trabecular bone and only a small increase in cortical bone (despite presumably showing a low bone mass phenotype). They speculate that this might be because of compensatory responses to chronic stimulation by SOST.10. Part of the excitement of these recent studies on osteocytic osteolysis is in the plethora of further questions they raise. Do osteotropic agents other than sclerostin and PTH/PTHrP (eg, Dickkopf-related protein 1 [DKK1], gonadal steroids, glucocorticoids) also regulate expression of the cellular machinery for mobilizing mineral in osteocytes? Does osteocytic mineral dissolution (or the reverse) occur in pathophysiological settings other than lactation? Do animals (or human patients) treated with carbonic anhydrase inhibitors such as acetazolamide show reduced osteocyte lacunar size? Do osteocytes in vivo also use pyrophosphate to reduce pericellular mineralization? Do bisphosphonates, some of which penetrate into the osteocytic lacunar system,19 affect osteocytic mineral dissolution? What are the relative contributions in vivo of osteocyte and osteoclast-mediated mineral release following treatment with sclerostin, PTH/PTHrP, and glucocorticoids, as well as in settings such as ovariectomy and lactation? In view of the very high surface area of interaction of osteocytes with mineralized matrix, is it possible that osteocytic mineral dissolution could offer a more rapid response than osteoclastic bone resorption? What happens to mineral homeostasis in mice subjected to targeted osteocyte ablation using the dentin matrix acidic phosphoprotein 1 (DMP-1)/diphtheria toxin method?20 Studying the biology of osteocytes directly is hard, owing to their location in bone and their multifunctional nature. The most clearly measurable activity of living osteocytes in situ still appears to be their ability to control the mineralization (and thus size) of their lacunae and canaliculi. Two-dimensional cultures of immortalized, osteocyte-like cell lines or primary cells have already provided many useful pointers, but as Atkins and Findlay6 point out, we ultimately need to refer to large animals or humans and use three-dimensional in vitro models to further our understanding of these fascinating cells. The author has had a research grant from Amgen and acted as a consultant for Amgen and UCB Celltech.