THE ANATOMICAL RECORD 224:317-324 (1989) Molecular Mechanisms of Bone Resorption by the Osteoclast ROLAND BARON Yale University School of Medicine, Departments of Orthopedics, Cell Biology, and Medicine, New Hauen, Connecticut 06510 ABSTRACT The osteoclast is a multinucleated cell that is actively engaged in the synthesis of lysosomal enzymes, their vectorial transport toward the apical membrane, and the secretion of these enzymes a t its apical pole. These secreted enzymes are targeted to the apical ruffled-border membrane by mechanisms that involve cation-independent mannose-6-phosphate receptors. These receptors bind to an enzyme-linked mannose-6-phosphate recognition marker in the Golgi complex, and the enzyme-ligand-receptor complex, carried within small coated transport vesicles, dissociates upon reaching the low pH established in the bone-resorbing compartment by the osteoclast. The apical bone-resorbing compartment is sealed off by the attachment of the osteoclast to the calcified matrix and is actively acidified by the osteoclast. The plasma membrane of the cell is divided into distinct domains. The apical membrane at the ruff led-border shares common antigenic determinants with lysosomal and endosomal membranes, including a 100 kD protein and proton pumps that may be involved in the acidification of the extracellular resorbing compartment. The basolateral membrane is highly enriched in sodium pumps. Finally, the cytoplasm of the osteoclast is highly enriched in carbonic anhydrase, and bicarbonate-chloride exchange appears to regulate the intracellular pH of this cell. These observations are consistent with a scheme in which, in the low pH environment of the bone-resorbing lacuna produced by the osteoclast, the mineral phase dissolves, exposing the organic matrix to the action of the secreted enzymes. The activity of these enzymes is in turn presumably favored by the acidic milieu. All constituents of the matrix, whether mineral or organic, then would be reduced to their elemental forms (ions and amino acids) extracellularly. No phagocytic events would be required for the complete degradation of the bone matrix. According to these concepts, therefore, membrane iontransport mechanisms become the most important molecular aspect of bone resorption. Resorption of the calcified extracellular bone matrix by osteoclasts constitutes an essential step in the sequential process of bone remodeling (Frost, 1964; Baron et al., 1984b). In carrying out this resorptive function, osteoclasts are thereby involved in normal bone turnover and growth (Kimmel and Jee, 1980; Tran van et al., 1982a,b), in spontaneous or induced tooth movement (Baron, 1973; Vignery and Baron, 19801,in bone-fracture healing (Frost, 19641, as well as in a number of pathological resorptive processes such as periodontal disease (Saffar and Baron, 1977; Saffar and Makris, 19801, local and systemic resorptive effects of cancers (Mundy and Spiro, 1981), Paget’s disease and osteoporosis (Avioli and Krane, 1977). Our laboratory has been studying the cellular and molecular biology of the osteoclast and its function in bone resorption. The goal of this brief review is to summarize our findings and concepts, based mostly on observations made in our own laboratory, but also on relevant contributions of other investigators. After describing the morphological characteristics of this cell, its synthetic and secretory functions are dis0 1989 ALAN R. LISS, INC. cussed. Both morphological and functional approaches have led to the conclusion that the osteoclast is a highly polarized cell. In this context, therefore, the issue of targeting secretory products to the ruffledborder apical domain, rather than to the basolateral domain of the cell membrane, is discussed. Characterization of the apical compartment then is described in terms of both its content and its limiting membrane, i.e., the ruff led-border membrane. These studies also demonstrate the existence of a distinct basolateral domain and establish the presence and functional importance of some ionic transport systems a t each pole of the cell. The potential role of these systems then are discussed inasmuch as they are directly related to the function of the osteoclast, i.e., bone resorption. Received September 16, 1988; accepted December 8, 1988. 318 R. BARON Fig. 1. Osteoclasts along the bone surface in their resorbing lacunae. Two osteoclasts, each with three nuclei in this plane of section, are present within a resorbing site formed by two Howship’s lacunae; at the light microscopic level, one can distinguish the basophilic cytoplasm, the sealing zones (arrows) limiting the highly vacuolated ruffled-border area (between arrows). Undecalcified section, Goldner stain x 1,250. Modified from Baron and Vignery 1981, by copyright permission of Pergamon Press. MORPHOLOGICAL CHARACTERISTICS OF THE OSTEOCLAST The osteoclast is a giant multinucleated cell formed by the asynchronous fusion of mononuclear hematopoietic precursors, most likely within the mononuclearphagocyte lineage (Jee and Nolan, 1963; Fishman and Hay, 1962; Burger et al., 1982; KOand Bernard, 1981). This cell is usually found singly or in low numbers at one given time and site, characteristically at the interface between soft and calcified tissues in an area where the bone matrix is fully mineralized (Fig. 1). Although this interface includes the periosteum, most of the remodeling activity occurs along the endosteum, i.e., at the interface between the calcified bone matrix and the bone marrow. Osteoclasts and osteoblast-related cells most probably are derived from hematopoietic and stroma1 stem cells, respectively, both of which are present in bone marrow (Owen, 1978). At the light microscopic (LM) level, the osteoclast is characterized by its size (50-100 km in average), its multinucleation (usually two to ten nuclei), and its presence within a resorptive (Howship’s)lacuna along the edge of the calcified matrix-bone marrow interface. In good preparations, the osteoclast is found in very close apposition to the calcified matrix; the area of the cell closest to the matrix is characterized by densely stained patches at the periphery of the cell and a lightly stained, highly vacuolated, and striated center area corresponding to the ruffled-border (Fig. 1,and see below). The cytoplasm is usually strongly basophilic, granular, and foamy, with vacuoles of varying sizes located mostly between the nuclei and the ruffled-border area. The nuclei are characteristically heterogeneous in size, shape, and basophilia, a possible reflection of the asynchronous fusion of mononuclear precursors (Miller et al., 1977). Ultrastructural analysis (Holtrop and King, 1977) of the os- teoclast (Fig. 2) confirms and extends the LM observations. First, the morphological polarity of the cell is evident, with a marked contrast existing between the side facing the calcified matrix and the side of the cell facing the bone marrow. Toward the bone surface, the peripheral zone of the osteoclast is very closely apposed to the extracellular matrix. This so-called “sealing zone” (Schenk et al., 1967) is characterized by a very narrow space (0.2 to 0.5 nm) between the plasma membrane of the cell and the calcified matrix and by the presence of an organelle-free area in the adjacent cytoplasm, the “clear-zone” (Fig. 21, which is characteristically enriched in contractile proteins (King and Holtrop, 1975; Marchisio et al., 1984). Toward the center of the cell-matrix interface, the plasma membrane of the osteoclast manifests progressively deeper infoldings, which reach a high degree of geometric complexity. Numerous folds and ampullar spaces are present which, due to the random obliquity of the sections, often appear as “vacuoles” in the cell’s cytoplasm. This membrane has been suggested to be coated on its cytoplasmic side in a characteristic manner (Kallio et al., 19711, but the exact nature of the coating material has not yet been determined. In all species tested so far (Fig. 2), one of the most consistent morphological features of the cytoplasmic organization of the osteoclast was the perinuclear distribution of multiple Golgi complexes, which entirely surround each of the nuclei of the cell (Baron et al., 1985b). We also recently found that all of these Golgi stacks are functionally oriented with the trans-side away from the nuclei (Baron et al., 1988a). In addition, the osteoclast is characteristically rich in mitochondriae, free polysomes, and coated transport vesicles in the Golgi areas, as well as in multiple vacuolar structures of heterogeneous size and shape, concentrated mostly between the nuclei and the ruffled-border area. As mentioned earlier, many of these vacuoles are pockets of extracellular space cut obliquely through the folds of the plasma membrane. The osteoclast is not, as often thought, poor in rough endoplasmic reticulum (rER): this organelle is quite well developed, but it remains limited in extent relative to the large size of the cell. The rER is usually concentrated in the basal portion of the cytoplasm of the osteoclast and in the perinuclear envelopes (Fig. 2). Distribution of Acid Phosphatase and Other Lysosomal Enzymes It has been known for some time that one of the main characteristics of the osteoclast is its enrichment in acid phosphatase. This has been well demonstrated by LM and electron microscopic (EM) studies (Fig. 2) and by enzyme cytochemistry (Doty and Schofield, 1972; Gothlin and Ericsson, 1971; Lucht, 1971; Baron et al., 1985b, 1986b, 1988a). An isoenzyme of acid phosphatase that is resistant to incubation in the presence of tartrate (tartrate-resistant acid phsophatase: TRAP) seems to be the form in which the osteoclast is particularly (Minkin, 19821, if not specifically (Anderson and Toverud, 19811, enriched. It is most important to emphasize that ultrastructural studies have shown that this high concentration of lysosomal enzymes in the osteoclast is not due to a high concentration of phago- MOLECULAR MECHANISMS O F BONE RESORPTION 319 Fig. 2. General ultrastructural view of an osteoclast after localization of arylsulfatase. This micrograph shows an osteoclast attached to the bone surface a t the sealing zones (SZ, arrows), which delimit the ruffled border area (between arrows) and its complex membrane infoldings; large extracellular recesses (double arrows) are found deep into the cell's cytoplasm and are close to the nuclei (n) and their surrounding Golgi complexes; abundant mitochondria are found in the osteoclast. Arylsulfatase is localized in all components of the biosynthetic pathway, including the endoplasmic reticulum (ER),the numerous perinuclear Golgi complexes, and their associated vesicles, which are often close to the infoldings of the ruffled-border membrane (double arrows). Bar 1 km. x 9,000. Reproduced from Baron et al., 1985, by copyright permission of the Rockefeller University Press. cytic structures such as secondary lysosomes. Instead, these enzymes are found, for the most part, in elements of the exocytic pathway (Baron et al., 1985b, 1986b, 1988a). Consequently, this enrichment in lysosomal enzymes does not reflect a high phagocytic activity, but rather a high biosynthetic activity. Using a variety of techniques to localize multiple enzymes, the presence of these enzymes was demonstrated in the rER, in the Golgi complex, and in numerous transport vesicles of the osteoclast. Thus, localization of arylsulfatase and beta-glycerophosphatase by enzyme cytochemistry and localization of beta-glucuronidase and cathepsin C by immunocytochemistry both have shown the abundant concentration of these enzymes in the lumen of the rER cisternae, including the perinuclear envelopes, in the cisternae of the Golgi complexes and in very numerous small (50 to 75 nm), coated vesicles in the Golgi complex areas and throughout the cytoplasm. These latter are particularly abundant between the nuclei and the deep portions of the ruffled-border membrane folds (Baron et al., 198513, 1988a). In addition, some typical secondary lysosomal vacuoles were found, which are filled also with enzymes but are uncoated, are larger, and have a heterogeneous content. These are, however, relatively few and concentrated mostly in the basal portion of the cell, facing the bone marrow compartment. Evidence for Polarized Secretion of Newly Synthesized Lysosomal Enzymes Although strongly suggested by many studies (Vaes, 1968; Lucht, 1971; Doty and Schofield, 1972; Gothlin 320 R. BARON and Ericsson, 1971), direct evidence for the secretion of lysosomal enzymes by the osteoclast into the sealed-off extracellular bone-resorbing compartment was, until recently, still lacking. Localization of these enzymes by enzyme cytochemistry has never provided convincing results, in part because of artefacts of salt precipitation that occur in the bone-resorbing lacuna (due most probably to the high concentration of calcium and phosphate in this specific microenvironment). It is also possible that an excess of endogenous substrate-to-enzyme ratio in this compartment masks hydrolysis of exogenously added substrate and consequently prevents reliable localization. Therefore, an attempt was made to use immunocytochemical methods to localize extracellular enzymes; recently one such enzyme, cathepsin C (Mainferme et al., 1985), was found a t high and detectable concentrations in the subosteoclastic boneresorbing compartment (Baron et al., 1988a). In contrast, enzyme cytochemistry provided a better method for the localization of lysosomal enzymes in intracellular organelles. Both methods agree, however, in showing high concentrations of these enzymes in the ER, all cisternae of the Golgi complexes, and in most of the small (50-75 nm), coated transport vesicles found in the Golgi regions and in the cytoplasm (between the perinuclear Golgi and the numerous folds and vacuoles of the ruff led-border complex). The overall conclusions from these studies are that the osteoclast is actively engaged in the synthesis of lysosomal enzymes that proceed through the Golgi and are vectorially transported from the trans-Golgi region to the ruff led-border apical membrane in coated transport vesicles. These transport vesicles then fuse with the plasma membrane, exclusively at the ruffled border, and release their content into the bone-resorbing compartment (Baron et al., 1988a). An attempt then was made to determine how these enzymes are sorted from other proteins and targeted to the apical membrane (Farquhar, 1985). cellular organelle. Consequently, the possibility was investigated that the osteoclast could lack such a ligand-receptor system, or dispose of it differently along the exocytic pathway to ensure the proper targeting of the enzymes to its apical membrane. It was found that, indeed, the osteoclast expresses high concentrations of the cation-independent mannose-6 phosphate receptor (Baron et al., 1988a,b). The receptors were found to be entirely codistributed with four different lysosomal enzymes along the exocytic pathway, from their site of synthesis in the rER, through the Golgi apparatus and the transport vesicles, to their delivery site at the ruffled-border membrane, where they were present only at sites of recent budding of transport vesicles. These results were interpreted to indicate that newly synthesized lysosomal enzymes in the osteoclast bear the mannose-6 phosphate recognition marker and are targeted to the apical ruffled-border membrane via their binding to the cation-independent mannose-6 phosphate receptor in the Golgi apparatus. The secretory enzymes are transported, bound to their receptors, in small, coated vesicles from the trans-face of the Golgi apparatus to the apical domain. These transport vesicles fuse with the ruffled border membrane and release the enzymes upon reaching the boneresorbing compartment (Baron et al., 1985b, 1988a), which is actively acidified by the osteoclast. Evidence for a Low pH in the Bone-Resorbing Compartment and for Active Acidification Process by the Osteoclast A number of previous studies suggested the possibility that the osteoclast might secrete acids during bone resorption (Neuman et al., 1960; Vaes, 1968; Gay and Mueller, 1974). These were, however, mostly indirect observations, and the fact that collagenases, which have a neutral pH optimum (Vaes, 1980), were considered to be the only enzymes capable of digesting collagen raised doubts as to whether or not the bone matrix could be degraded in an acidic environment (Blair et Role of the Mannose-6-PhosphateReceptor al., 1986). In most cells, lysosomal enzymes are sorted out of the To address this issue, we developed a culture system main flow of secretory and membrane proteins a t the in which, without isolation of the cells, active funclevel of the Golgi apparatus (Farquhar, 1985; Kornfeld, tional osteoclasts could be directly observed in situ 1986; Griffiths and Simons, 1986; von Figura and Ha- while resorbing a flat surface of bone (Baron et al., silik, 1986) and specifically targeted to late endosomes 1985b). When such flat fragments of bone were incualong the endocytic pathway (Brown et al., 1986). The bated in the presence of the weak base acridine orange, newly synthesized enzymes are recognized by a specific large acidic compartments could be observed by epiflureceptor, the mannose-6-phosphate receptor (Creek orescence in close association with the osteoclasts. Beand Sly, 1984), which is highly concentrated in mem- cause such compartments disappeared upon isolation of branes of the Golgi compartment (Brown and Farqu- the cells from their substratum, it was suggested that har, 1984, 19871, and binds the mannose-6-phosphate the acidic compartment was indeed extracellular, i.e., recognition marker, which is added to the lysosomal in the subosteoclastic bone resorbing compartment enzymes in the cis-Golgi elements. This mechanism en- (Baron et al., 1985b). More recently, this finding was sures that the enzymes are sorted out of the flow of confirmed using an EM probe for low pH (Anderson et other proteins synthesized by the cells. The enzymes al., 1984) and peroxidase-conjugated second antibodies. remain bound to this mannose-6-phosphate receptor This probe accumulated in the subosteoclastic compartuntil they reach an acidified compartment where the ment and between the fingerlike folds of the ruffledlow pH induces their dissociation from the receptor. In border membrane, thereby demonstrating unequivothis manner, enzymes are released in the proper com- cally that this compartment is indeed acidic (Baron et partment, i.e., the late endosome (Brown et al., 1986). al., 1985a). However, the situation is quite different in the osTo determine if this was the result of osteoclastic teoclast, because this cell secretes its newly synthe- activity, the cells were incubated in the presence of sized enzymes instead of targeting them to an intra- ammonium chloride, thereby dissipating the existing MOLECULAR MECHANISMS OF BONE RESORPTION 321 Fig. 3. Immunolocalization of the 100 kD membrane protein in an osteoclast. The reaction product is present along the ruffled-border membrane but is lacking at the sealing zone membrane (SZ); double arrow indicates clear-cut limit between these two membrane domains. Note the ampullar dilations of extracellular space (ECS)at the cytoplasmic end of the plasma membrane infoldings of the ruffled border, which might appear in sections as intracellular vacuoles. Small arrows show the reaction product exclusively at the luminal side of the membrane. Bar, 600 nm. x 20,000. Reproduced from Baron et al., 1985, by copyright permission of the Rockefeller University Press. pH gradient across the membrane. Upon reincubation of the osteoclasts in normal culture medium, the cells were able to reacidify the bone-resorbing compartment, demonstrating that acidification is an active process due to the osteoclast (Baron et al., 1985b). Data from other laboratories have indeed supported and confirmed these observations. First, fluid micropipetted from the osteoclast environment in vitro (a method that does not, however, permit exact localization of the sample), suggested the presence of a low pH (Fallon, 1984). Second, studies with isolated osteoclasts placed on a slice of bone in vitro demonstrated that these cells were more active in a slightly acidic culture medium (Arnett and Dempster, 1986) and suggested that under these conditions the cells more readily establish and maintain a pH gradient. It is therefore reasonably well accepted now that the osteoclast is able to acidify the bone-resorbing compartment formed at its apical surface, and this acidification is the result of an active process performed by the cell itself. sosome. Therefore, the hypothesis was tested that the ruff led-border membrane could share common antigenic determinants with lysosomal membranes. For this purpose, a number of antibodies known to recognize lysosomal andor endosomal membrane antigens (Baron et al., 1985a) were used, and immunocytochemical localization of these antigens in osteoclasts was performed. The first and most striking result was the presence of a 100 kD lysosomal membrane protein a t the ruffled border of the osteoclast (Baron et al., 1985b). Interestingly, this antigen was limited strictly to the ruffled-border membrane (Fig. 3) but was not present in detectable amounts at the sealing zone or the basolateral membrane of the osteoclast. This observation thus established that 1)the osteoclast, although not epithelial in nature, is able to establish and maintain distinct plasma membrane domains, i.e., polarized apical and basolateral domains; and 2) the membrane limiting the extracellular bone-resorbing compartment shares some common antigen(s) with the limiting membrane of lysosomes, thereby better establishing the functional parallel between the bone-resorbing lacuna and secondary lysosomes. Further studies of membrane proteins in the ruffled border area, however, indicated that some other lysosomal membrane antigens were absent (Baron et al., 1985a), suggesting some potentially important differences between these two membranes. Given the known role of a proton pump adenosine Characterization of the Apical Plasma Membrane (Ruffled Border) and of its Potential Role in Acidification The apical bone-resorbing compartment, which is characterized by a low pH, a high concentration of lysosomal enzymes, and the presence of the substrate for these enzymes (the bone matrix proteins), is highly reminiscent of the luminal content of a secondary ly- 322 R. BARON triphosphatase (ATPase) in establishing and maintaining the pH gradient across the lysosomal membrane (Ohkuma and Poole, 1978) and given the fact that most ATPases are in the 100 kD molecular weight range, Reggio et al. (1984), who first described this 100 kD lysosomal membrane protein, tested the possibility that it could be the protein transporting protons into lysosomes and found that on immunoblots their antibody did not cross-react with the alpha subunit of the sodiudpotassium or the calcium ATPase, both of which are also in the 100 kD range; however, they recognized the 100 kD proton/potassium ATPase purified from gastric mucosae. This strongly suggested, but did not demonstrate, a possible link between the presence of this membrane protein in lysosomes (and in the osteoclast ruffled border) and the fact that both membranes limit acidic compartments. According to these observations, the proton-pump ATPase a t the ruff led border could be of the gastric type. Some reports in the literature have since indicated that the drug omeprazole, which specifically inhibits these pumps, decreased bone resorption in organ cultures (Tuukkanen and Vaananen, 1986) and the pH gradient formed by osteoclasts (Anderson et al., 1986). More recent data, although still preliminary, has more convincingly indicated that these pumps could be of the kidney type (i.e., electrogenic), inhibited by N-ethyl maleimide (NEM),a compound known to inhibit specifically electrogenic proton pumps from kidney or lysosomal membranes, but not the (H + /K + )ATPase (Al-Awqati, 1986), and are not inhibited by vanadate (Ghiselli et al., 19871, a compound known to inhibit the protodpotassium ATPase. Preliminary data from our laboratory using isolated osteoclast membrane vesicles has also shown a strong inhibition of acidification by NEM (Fuchs et al., unpublished data). In addition, the coating of the inner surface of the ruff led-border membrane is strongly reminiscent of the coating observed in kidney vesicles, recently shown to be formed by crystalline arrays of proton pump subunits (Brown et al., 1988). All of these observations therefore strongly suggest that the ruffled-border membrane is responsible for the acidification of the bone-resorbing compartment via a specific enrichment in proton pumps, most likely of the NEM-sensitive electrogenic type. Characterizationof the Basolateral Membrane and Its Transport Mechanisms Potentially Associated With Bone Resorption Of all the plasma-membrane transport systems, two are particularly important in the basolateral membranes of cells that transport protons across their apical membrane. These are the sodium pump (Na+, K + IATPase, which often provides the primary force for proton transport or associated cotransport (Cantley, 19811, and the bicarbonate-chloride exchanger, secreting the bicarbonate generated by the enzyme carbonic anhydrase, which provides protons to the apical pumps from COZ and water (Al-Awqati, 1978) and is highly enriched in the osteoclast (Gay et al., 1974,1983;Vaananen and Parvinen, 1983). Studies were therefore initiated that aimed at determining if the basolateral membrane of the osteoclast was indeed enriched in these two transport molecules. Using monoclonal antibodies t o both the alpha and beta subunits of the (Na + ,K + )ATPase, immunocytochemistry showed the osteoclast to be highly enriched in sodium pumps (Baron et al., 1986a). Not only was the cell enriched, but this enrichment was SO pronounced that it became specific for the osteoclast relative to other cells present in sections of normal bone tissue. In order to establish and quantitate this enrichment in sodium pumps further, we then proceeded to measure the binding of radiolabeled ouabain, a specific ligand for the alpha subunits of these pumps. About 5 million binding sites per osteoclast (Baron et al., 1986a) were found, a relatively high number compared with most other cells. Although this enrichment could be related to functions other than bone resorption per se, recent data from our laboratory indicates that the activity of the (Na + , K + )ATPase is required for bone resorption to occur. Thus, ouabain inhibits the release of 45Ca from fetal rat long bones in culture as well as the formation of resorption pits by isolated osteoclasts (Prallet et al., 1988). As is the case for the proton pump ATPase, our most recent data (Baron et al., 1988b)suggests that the sodium pump of the osteoclast is of the kidney type (alpha 1) (Sweadner and Gilkeson, 1985) and that it is concentrated mostly in the basolateral domain of the plasma membrane. To date there is no clear immunocytochemical or immunochemical data to demonstrate that the bicarbonate-chloride exchanger is present in the osteoclast. However, recent transport data based on the measurement of intracellular pH in isolated cells indicates that the osteoclast possesses such a system (Teti et al., 1987). However, further work is necessary to characterize this transport mechanism both biochemically and physiologically. Cellular and Molecular Mechanisms of Bone Resorption The various observations discussed above have led us to reassess current views on the mechanisms by which the multinucleated osteoclast performs its highly specialized function of resorbing bone, i.e. digesting an extracellular calcified matrix. This review is concluded by a summary of our current concepts on the molecular aspects of osteoclast function. The osteoclast is a multinucleated cell that is actively engaged in the synthesis, vectorial transport, and polarized secretion of lysosomal enzymes toward its apical pole. These secretory products are targeted to the apical ruff led-border membrane by mechanisms that involve cation-independent mannose-6-phosphate receptors. These receptors bind to the mannose-6-phosphate recognition marker in the Golgi complex, and the receptor-ligand complex, carried within small, coated transport vesicles, dissociates upon reaching the low pH established in the bone-resorbing compartment by the osteoclast. The apical bone-resorbing compartment is sealed off by the attachment of the osteoclast to the calcified matrix and is actively acidified by the osteoclast. The plasma membrane of the cell is divided into distinct domains. The apical membrane a t the ruffled border shares common antigenic determinants with lysosomal and endosomal membranes, including a 100 kD protein and proton pumps. The basolateral membrane is highly enriched in sodium pumps. Both of MOLECULAR MECHANISMS O F BONE RESORPTION these membrane transport systems seem analogous to the kidney forms rather than to the stomach and brain forms, respectively. Finally, the cytoplasm of the osteoclast is highly enriched in carbonic anhydrase, and bicarbonate-chloride transport appear to regulate the intracellular pH of this cell. Once the osteoclast forms a sealed-off bone-resorbing compartment (subosteoclastic) and secretes both lysosomal enzymes and protons (producing a low pH environment), the mineral phase dissolves, exposing the organic matrix to the proteolytic action of the enzymes. Protons are provided to the proton pumps by the intracellular action of carbonic anhydrase and the generated bicarbonate is secreted basolaterally via a bicarbonate-chloride exchanger in the membrane. The chloride exchanged in this process might be secreted apically, in parallel with protons, via chloride channels. The calcium that would be generated in the resorbing compartment would be disposed of by transcytosis via calcium channels, calcium ATPases, or sodium-calcium exchangers; the high basolateral concentration of sodium pumps suggests this latter as the most likely mechanism (Krieger and Tashjian, 1980). In this view of osteoclast function, all constituents of the matrix, whether mineral or organic, would be reduced to their elemental forms (ions and amino acids) extracellularly, and no intracellular phagocytic events would be required for the complete degradation of the bone matrix. According to these concepts, membrane transport mechanisms become the most important molecular physiological aspect of bone resorption. ACKNOWLEDGMENTS The present work was supported primarily by a grant from the National Institutes of Health (DE04724). The author is very grateful to Mrs. Lynn Neff for her expert technical help. LITERATURE CITED Al-Awqati, Q. 1978 H + transport in urinary epithelia. Am. J . Physiol., 235:F77-F88. Al-Awqati, Q. 1986 Proton-translocating ATPases. Annu. Rev. Cell Bid., 2:179-199. Anderson, R.W., J.R.Falck, J.L. Goldstein, and M.S. Brown 1984 Visualization of acidic organelles in intact cells by electron microscopy. Proc. Natl. Acad. Sci. U.S.A., 81:4838-4842. Anderson, T.R., and S.U. Toverud 1981 Further studies on the separation and identification of two phosphatases with acid optima from rat bone. Calcif. Tissue Int., 33:261-267. Anderson, R.E., D.M. Woodbury and W.S.S. Jee 1986 Humoral and ionic regulation of osteoclast acidity. Calcif. Tissue Int., 39:252-258. Arnett, T.R., and S.W. Dempster 1986 The effect of pH on bone resorption by rat osteoclasts in vitro. Endrocinology, 119:119-124. Avioli, L.V., and S.M. Krane 1977 Metabolic Bone Disease. Academic Press, New York. Baron, R. 1973 Remaniement de la lame cribriforme et des fibres desmodontales a u cours de la migration physiologique. J . Biol. Buccale, 1:151. Baron, R., S. Kellokumpu, L. Neff, S. Jamsa-Kellokumpu, A. Billecocq, B. Prallet, J. Beresford, J . Emanuel, K. Sweadner, and R. Levenson 1988b Sodium pumps and bone resorption: presence of basolateral kidney type alpha subunits in osteoclasts, effects of ouabain on bone resorption and interactions with calcitonin in MDCK cells. Calcif. Tissue Int., 42: (abstract). Baron, R., and A. Vignery 1981 Behavior of osteooclasts during a rapid change in their number induced by high doses of parathyroid hormone or calcitonine in intact rats. Metab. Bone Tis. Res. 2~339-346. Baron, R., A. Vignery, and M. Horowitz 1984 Lymphocytes, macro- 323 phages and the regulation of bone remodeling. In: Bone and Mineral Research 2. W.J. Peck ed., Elsevier, New York, pp. 175-243. Baron, R.,L. Neff, J. Lippincott-Schwartz, D. Louvard, I. Mellman, A. Helenius, and M. Marsh 1985a Distribution of lysosomal membrane proteins in the osteoclast and their relationship to acidic compartments. J . Cell Biol., 101:53a (abstract). Baron, R., L. Neff, D. Louvard, and J.P. Courtoy 19851, Cell-mediated extracellular acidification and bone resorption: Evidence for a low pH in resorbing lacunae and localisation of a 100-KD lysosomal membrane protein at the osteoclast ruffled border. J . Cell Biol., 101:2210-2222. Baron, R., L. Neff, C. Roy, A. Boisvert, and M. Caplan 1986a Evidence for a high and specific concentration of (Na + ,K + )ATPase in the plasma membrane of the osteoclast. Cell, 46:311-320. Baron, R., L. Neff, P. Tran Van, J.R. Nefussi, and A. Vignery 1986b Kinetic and cytochemical identification of osteoclast precursors and their differentiation into multinucleated osteoclasts. Am. J . Pathol., 122:363-378. Baron, R., L. Neff, W. Brown, D. Louvard, P. Courtoy, and M.G. Farquhar 1988a Polarized secretion of lysosomal enzymes: co-distribution of cation independent mannose 6-phosphate receptors and lysosomal enzymes in the osteoclast exocytic pathway. J . Cell Biol., 106:1863-1872. Blair, H.C., A.J. Kahn, E.C. Crouch, J.J. Jeffrey, and S.L. Teitelbaum 1986 Isolated osteoclasts resorb the organic and inorganic components of bone. J . Cell Biol., 102:1164-1172. Brown, W.J., and M.G. Farquhar 1984 The mannose-6-phosphate receptor for lysosomal enzymes is concentrated in cis-Golgi cisternae. Cell, 36:295-307. Brown W.J., and M.G. Farquhar 1987 The distribution of 215 kD mannose 6-phosphate receptors within cis (heavy) and trans (light) Golgi subfractions varies in different cell types. Proc. Natl. Acad. Sci. U.S.A. (in press). Brown, W.J., J . Goodhouse, and M.G. Farquhar 1986 Mannose6-Phosphate Receptors for Lysosornal Enzymes Cycle between the Golgi Complex and Endosomes. J. Cell Biol., 103:1235-1247. Brown, D., S. Gluck, and J. Hartwig 1987 Structure of the novel coating material in proton-secreting epithelial cells and identification as a n H+ATPase. J. Cell Biol., 105~1637-1648. Burger, E.H., J.W.M. Van der Meer, J.S. Van de Gevel, J.C. Gribnau, C.W. Thesingh, and R. Van Furth 1982 In vitro formation of osteoclasts from long term cultures of bone marrow mononuclear phagocytes. J . Exp. Med., 156:1604-1634. Cantley, L.C. 1981 Structure and mechanism of the ( N a + , K + IATPase. Curr. Top. Bioenerg. 11:201-237. Creek, K.E., and W.S. Sly 1984 The role of the phosphomannosyl receptor in the transport of acid hydrolases to lysosomes. In: Lysosomes in Biology and Pathology. J.T. Dingle, R.T. Dean, and W.S. Sly, eds. Elsevier, Amsterdam, pp. 63-82. Doty, S.B., and B.H. Schofield. 1972 Electron microscopic localization of hydrolytic enzymes in osteoclasts. Histochem. J., 4:245-258. Fallon, M.D. 1984 Bone resorbing fluid from osteoclasts is acidic-An in vitro micropuncture study. In: Endocrine Control of Bone and Calcium Metabolism. Cohn D.V. et al. eds. Amsterdam, Elsevier, p. 144. Farquhar, M.G. 1985 Progress in unraveling pathways of Golgi traffic. Annu. Rev. Cell Biol., 1:447-88. Fishman, D.A., and E. Hay 1962 Origin of the osteoclast from regenerating newt limbs. Anat. Rec. 143:329-334. Frost, H.M. 1964 Mathematical Elements of Bone Remodeling. Charles C. Thomas, Springfield, Illinois. Gay, C.V., and W.J. Mueller 1974 Carbonic-anhydrase and osteoclasts: localization by labeled inhibitor autoradiography. Science, 183:432-434. Gay, C.V., M.B. Ito, and H. Schraer 1983 Carbonic anhydrase activity in isolated osteoclasts. Metab. Bone Dis. Rel. Res., 5:33-39. Ghiselli, R., H. Blair, S. Teitelbaum, and S. Gluck 1987 Identification of the osteoclast proton pump. J. Bone Min. Res. 2:Suppl. 1 (abstract). Gothlin, G., and J.L.E. Ericsson 1971 Fine structural localization of acid phosphomonosterase in the brush border region of osteoclasts. Histochemie, 28:337-344. Griffiths, G., and K. Simmons 1986 The trans Golgi Network Sorting at the Exit Site of the Golgi Complex. Science, 234:438-443. Holtrop, M.E., and G.J. King 1977 The ultrastructureofthe osteoclast and its functional implications. Clin. Orhop. Rel. Res., 123: 177-196. Jee, W.S.S., and P.D. Nolan 1963 Origin of osteoclasts from the fusion of phagocytes. Nature, 200:225-226. Kallio, D.M., P.R. Garant, and C. Minkin 1971 Evidence of coated 324 R. BARON membranes in the ruffled border of the osteoclast. J . Ultrastruct. Res., 37:169-177. Kimmel, D., and W.S.S. Jee 1980 A quantitative histologic analysis of the growing long bone metaphysis. Calcif. Tissue Int. 32: 113-122. King, G.J., and M.E. Holtrop 1975 Actin-like filaments in bone cells of cultured mouse calvaria as demonstrated by binding to heavy meromyosin. J . Cell Biol., 66:445-451. KO,J.S., and G.W. Bernard 1981 Osteoclast formation in vitro from bone marrow mononuclear cells in osteoclast-free bone. Am. J. Anat., 161:415 -425. Kornfeld, S. 1986 Trafficking of lysosomal enzymes in normal and disease states. J. Clin. Invest., 77:l-6. Krieger, N.S., and A.H. Tashjian 1980 Parathyroid hormone stimulates bone resorption via a Na-Ca exchange mechanism. Nature, 287:843-845. Lucht, U. 1971 Acid phosphatase of osteoclasts demonstrated by electron microscopic histochemistry. Histochemie, 28:103-117. Mainferme, F., R. Wattiaux, and K. von Figura 1985 Synthesis, transport and processing of cathepsin C in Morris hepatoma 7777 cells and rat hepatocytes. Eur. J. Biochem., 153:211-216. Marchisio, P.C., D. Cirillo, L. Naldini, M.V. Primavera, A. Teti, and A. Zambonin-Zallone 1984 Cell-substratum interaction of cultured avian osteoclasts is mediated by specific adhesion structures. J. Cell Biol., 99:1696-1705. Miller S., W.S.S. Jee, D.B. Kimmel, and L. Woodbury 1977 EHDP effects on incorporation and accumulation of osteoclast nuclei. Calcif. Tissue Res., 22:243-252. Minkin, C. 1982 Bone acid phosphatase: Tartrate resistant acid phosphatase as a marker of osteoclast function. Calcif. Tissue Int., 34:285-290. Mundy, G., and T.P. Spiro 1981 The mechanisms of bone metastasis and bone destruction by tumor cells. In: Bone Metastases. L. Weiss and H.A. Gilbert, eds. G.K. Hall, Boston, pp. 64-82. Neuman, W.F., B.J. Mulryan, and G.R. Martin 1960 A chemical view of osteoclasts based on studies with Yttrium. Clin. Orhop., 17:124-134. Ohkuma, S., and B. Poole 1978 Fluorescence probe measurement of the intralysosomal pH in living cells and the perturbation of pH by various agents. Proc. Natl. Acad. Sci. U.S.A., 75:3327-3331. Owen, M. 1978 Histogenesis of bone cells. Calcif. Tissue Res., 25:205-207. Prallet, B., J. Beresford, L. Neff, and R. Baron 1988 Inhibition of sodium pumps by ouabain decreases bone resorption in fetal rat long bones and isolated rat osteoclast cultures. J. Bone Min. Res., 3:Suppl. 1, (abstract). Reggio, H., D. Bainton, E. Harms, E. Coudrier, and D. Louvard 1984 Antibodies against lysosomal membranes reveal a 100 kD protein which cross-reacts with purified H + , K + ATPase from gastric mucosa, J . Cell Biol., 99:1511. Saffar, J.L., and R. Baron 1977 A quantitative study of osteoclastic bone resorption during experimental periodontal disease in the golden hamster. J . Periodont. Res., 12:387-394. Saffar, J.L., and G.P. Makris 1986 A morphological and quantitative study of osteoclast changes during the progress of periodontitis in the hamster. J. Biol. Buccale, 14:255-262. Schenk, R., D. Spiro, and J. Wiener 1967 Cartilage resorption in tibia1 epiphyseal plate of growing rats. J. Cell Biol., 34:275-291. Sweadner, K.J., and R.C. Gilkeson 1985 Two isozymes of the (Na+ ,K + )ATPase have distinct antigenic determinants. J. Biol. Chem., 260:9016-9022. Teti, A,, H. Blair, A. Kahn, J. Konsek, C. Koziol, A. Zambonin-Zallone. S. Teitelbaum. and P. Schlesinaer 1987 Intracellular PH regu'lation of isolated osteoclasts b y chloridefbicarbonate exchange. J. Bone Min. Res. 2:Suppl. 1 (abstract). Tran Van, P., A. Vignery, and R. Baron 1982a An electron microscopic study of the bone remodeling sequence in the rat. Cell Tiss. Res., 225~283,1982a. Tran Van, P., A. Vignery, and R. Baron 19821, Cellular kinetics of the bone remodeling sequence in the rat, Anat. Rec., 202:445. Tuukkanen, J., and H.K. Vaananen 1986 Omeprazole, a specific inhibitor of H + -K + ATPase, inhibits bone resorption in vitro. Calcif. Tissue Int., 38:123-125. Vaananen, H.K., and E.K. Parvinen 1983 High active isoenzyme of carbonic anhydrase in rat calvaria osteoclasts. Histochemistry, 78:481-485. Vaes, G. 1968 On the mechanisms of bone resorption: the action of parathyroid hormone on the excretion and synthesis of lysosomal enzymes and on the extracellular release of acid by bone cells. J . Cell Biol., 39:676-697. Vaes, A. 1980 Collagenase, lysosomes, and osteoclastic bone resorption. In: Collagenase in Normal and Pathological Connective Tissues. D.E. Woolley and J.M. Evanson, eds. J. Wiley & Sons, London. pp. 185-207. Vignery, A., and R. Baron 1980 Dynamic histomorphometry of alveolar bone remodeling in the adult rat, Anat. Rec., 196:191. von Figura, K., and A. Hasilik 1986 Lysosomal enzymes and their receptors. Annu. Rev. Biochem., 55:167-193.