Steven L. Teitelbaum - US grants

The loss of bone in periodontal disease is largely the result of increased osteoclastic activity. While substantial new insight has been gained in identifying the factors involved in evoking this increased activity, the fundamental process of bone resorption remains incompletely understood. Much of this confusion reflects limitations inherent in most existing techniques for analyzing the resorptive process. In our original application, we proposed to explore new and less equivocal approaches to bone resorption. Our efforts have lead to the development of assay systems which employ essentially pure populations of resorptive cells (monocytes, macrophages and, more recently, multinucleated macrophages). These assays permit kinetic and quantitative analysis of two of the major events in the bone resorption sequence; the attachment of resorptive cells to the bone surface and bone matrix degradation. With these techniques in hand, along with some recently acquired biochemical skills, we are now in a position to address some of the fundamental issues of bone resorption. Specifically, we propose to: 1) examine the mechanisms by which resorptive cells recognize and bind to bone; 2) establish the temporal relationships between the binding process and onset of bone resorption; 3) determine the role of collagenase, other neutral proteinases and selected acid hydrolases in bone matrix degradation; 4) explore the functional relationship between the acquisition of multiple nuclei (as in osteoclastic development) and increased resorptive efficiency.

Implant loosening remains a major cause of delayed failure in joint replacement. Because such loosening occurs as a consequence of excess bone resorption resolution of this problem pivots on identifying the cells and resorptive mechanisms involved in bone matrix degradation. Recent work from this, and other laboratories, has shown that (i) osteoclasts resorb bone collagen using a lysosomal acid protease, (ii) osteoblasts may be involved in the resorptive phase of remodeling sine, in response to hormones, they produce both neutral collagenase and a potent collagenase inhibitor, and (iii) macrophages, which accumulate at sites of cemented implants, secrete a number of factors, including Interleukin-1, and prostaglandin E2, capable of stimulating resorptive activity. In the present application, experiments are described aimed at expanding and integrating these observations. Specifically, we propose to : (1) ISOLATE AND CHARACTERIZE THE ENZYME RESPONSIBLE FOR OSTEOCLAST-MEDIATED COLLAGENOLYSIS, (2) ASSESS THE INTRACELLULAR PROCESSING AND TRAFFICKING OF THE OSTEOCLAST COLLAGENOLYTIC ENZYME IN RESORPTIVE AND NON- RESORPTIVE OSTEOCLASTS, AND (3) EVALUATE THE CAPACITY OF MACROPHAGES TO (A) MODULATE OSTEOCLAST RESORPTIVE ACTIVITY AND (B) INFLUENCE SYNTHESIS OF OSTEOBLAST NEUTRAL COLLAGENASE AND COLLAGENASE INHIBITOR.

Inappropriate enhancement and diminution in the activity of bone resorbing cells are important causal factors in the pathophysiology of most metabolic diseases of bone. For example, osteopenia and osteopetrosis are, respectively, associated with increased or decreased levels of bone matrix degradation. It follows, therefore, that the establishment of rational therapies for the treatment of these disorders will depend upon developing a fuller understanding of the resorption mechanism. Bone resorption is a complex, multistage process initiated by the attachment of resorptive cells (or their immediate precursors) to the bone matrix. Previous studies of the latter phenomenon have revealed that: a) attachment and resorption are properties of relatively well differentiated cells; b) resorptive efficiency is paralleled by cell-bone binding; c) carbohydrates (oligosaccharides) on both cells and bone matrix are essential for attachment; d) glucocorticoids stimulate resorption by macrophages (and, perhaps, related cell types) by increasing binding efficiency; e) glucocorticoids increase binding by altering the exposure of specific membrane sugars. We now seem to be in particularly favorable position to extend our understanding of attachment/resorption mechanisms. The essential experimental techniques are largely in hand, at least one major component of the attachment mechanism has been identified (cell membrane sugars), and chemical agents (glucocorticoids) have been assessed which specifically modify both cell-bone binding and the cell membrane. Given this base, we propose to: 1) determine if osteoclasts and macrophage polykaryons attach to bone by the same mechanisms as macrophages; 2) isolate and identify the "core" protein(s) of bone "attachment" glycoproteins; 3) characterize the oligosaccharides associated with such glycoproteins; 4) establish the mechanisms by which glucocorticoids modulate binding and, in particular, the molecules associated with attachment; and 5) determine whether changes in membrane lipid composition and fluidity play a major role in glucocorticoid-stimulated cell-bone binding.

Osteoporosis always reflects enhanced osteoclastic relative to osteoblastic activity and hence, the long term goals of this project have been elucidation of the cellular mechanisms of bone resorption. Significant progress has been made in the last five years reflecting, in particular, the development of models whereby highly enriched populations of osteoclasts can be studied in vitro. Using such models, we have shown that bone resorption is a multistep process initiated by osteoclast-matrix recognition and attachment. In fact, it is now clear that the capacity of the cell to degrade bone ultimately depends upon its matrix binding ability. Recent evidence implicates integrins as the major class of molecules mediating cell-matrix recognition and attachment and we and others have shown that such is the case as regards the osteoclast. In particular, the integrin alphav/beta3 whose expression, we find, is regulated by the osteoclastogenic hormone, 1,25-dihydroxyvitamin D3, appears to be the major modulator of osteoclast-bone binding and bone resorption. It is also now evident that integrins, in addition to their attachment functions, are capable of initiating intracellular signalling when interacting with specific matrix proteins or peptides. As evidenced by our preliminary data, such is true as regards the osteoclast. Thus, our efforts will be aimed at studying the biology of integrins in the osteoclast as well as determining mechanisms whereby their function in vivo modulates osteoclast recruitment and bone resorption. We therefore specifically propose to: 1) investigate the mechanisms by which 1,25-dihydroxyvitamin d3 regulates integrin expression during osteoclastogenesis, 2) explore the relationship between bone matrix recognition by osteoclasts and delivery of a collagenolytic enzyme to the resorptive microenvironment; and 3) determine if regulation of integrin function modulates osteoclast recruitment and resorptive activity in vivo.

DESCRIPTION (provided by applicant): Inflammatory osteolysis, in conditions such as prosthetic implant loosening, reflects accelerated osteoclast (OC) recruitment, by products of inflammation. Thus, discovering the mechanisms by which local inflammation recruits OCs is central to preventing implant osteolysis and other forms of inflammation-induced bone loss. We have shown that tumor necrosis factor-alpha (TNF) is central to the osteoclastogenesis (OCgn) mediating experimental implant osteolysis. Furthermore, TNF directly induces macrophages to assume the OC phenotype, but must do so in the context of at least permissive levels of the essential osteoclastogenic molecule, RANK ligand (RANKL). RANKL is produced, in turn, by marrow stromal cells under the aegis of TNF. While observations, made during the last funding period, establish TNF and RANKL as central to inflammatory osteolysis, how marrow stromal cells, macrophages and T-lymphocytes mediate TNF-induced OCgn, and implant osteolysis, is unknown. Equally enigmatic are the mechanisms by which TNF enhances RANKL expression by marrow stromal cells, and the structural domains of RANKL, interacting with its receptor, RANK, which promote OCgn. We have developed a series of tools to address these issues. First, we have a chimeric mouse model which enables us to assess the roles of stromal cells, macrophages and T-lymphocytes in TNF-induced OCgn. Second our murine particle osteolysis model permits us to ask these questions in the context of implant loosening. Third, we have extensive experience in assessing regulation of osteoclastogenic genes and identifying the functional components of their promoters. Finally, we have solved the crystal structure of RANKL, which positions us to do the same in the context of the RANKL-RANK complex. Thus, we hypothesize that: (1) marrow stromal cells, macrophages and/or T-lymphocytes directly or indirectly mediate TNF-induced OCgn and implant osteolysis; (2) TNF regulates marrow stromal cell RANKL expression by specific molecular events; and (3) specific structural regions of the RANKL-RANK complex mediate OCgn. Our Specific Aims are therefore to: (1) determine how marrow stromal cells, macrophages and/or T-lymphocytes directly or indirectly mediate TNF- induced OCgn and implant osteolysis; (2) identify the specific molecular events by which TNF regulates marrow stromal cell RANKL expression; and (3) identify the structural regions of the RANKL-RANK complex which mediate OCgn.

DESCRIPTION (provided by applicant): Prompted by our observations and those of others, the ?v?3 integrin is a current therapeutic anti-resorptive target for diseases such as osteoporosis and inflammatory osteolysis. The purpose of this grant, from its inception, has been to expand the potential of therapeutically targeting the integrin by characterizing its associated molecules and the bone-degrading signals they transmit. The success of this exercise is underscored by the fact that in addition to those inhibiting ?v?3, drugs targeting signaling molecules, such as c-Src and Syk, which are effectors of the integrin, in OCs, are in clinical trial for osteolytic diseases. We have, in the past funding period, fulfilled our specific aims. While our efforts continue to define the mechanisms by which ?v?3 regulates the OC, particularly in the context of derivative signals which organize its cytoskeleton, our current attention turns to the integrin, itself. We address the mechanisms by which the integrin assumes an activated conformation which permits it to recognize ligand and transmit its bone-resorptive signals. Our findings and those of others, indicate M-CSF plays a central role in this regard. We have established that ?v?3 and M-CSF collaborate in organizing the OC cytoskeleton. Our data indicate, however, that the principal means by which M-CSF exerts its cytoskeletal effect is to prompt its receptor c-Fms to transmit intracellular signals to the cytoplasmic domain of the ?3 integrin subunit. These signals, in turn, transit the integrin from its default, resting state, to its activated conformation, permitting ligand recognition and cytoskeleton-organizing events. Our findings also suggest that talin activates OC ?v?3 and the signals derived thereof, by interacting with specific residues in the ?3 integrin cytoplasmic domain. We hypothesize, therefore that 1) M-CSF, liganding its receptor, c-Fms, transmits intracellular signals which activate the ?v?3 integrin in OCs; 2) talin and its recognition sequences in the ?3 cyoplasmic domain mediate M-CSF-induced OC cytoskeletal organization and function and 3) inhibiting ?v?3 activation, in vivo, diminishes pathological bone resorption. Our specific aims are therefore to determine 1) the mechanism by which M-CSF, interacting with its receptor c-Fms, transmits intracellular signals which activate the ?v?3 integrin in OCs; 2) the role of talin and its recognition sequences in the ?3 integrin cytoplasmic domain in mediating M-CSF-induced OC cytoskeletal organization and function and 3) the impact of inhibiting ?v?3 activation, in vivo, on pathological bone resorption.

[unreadable] DESCRIPTION (provided by applicant): Treatment of osteopenic disorders has relied, to-date, on anti-resorptive drugs such as estrogens and bisphosphonates. While these agents often retard progressive bone loss, they are not effective in reversing the established osteoporotic lesion, nor are they typically capable of curing patients already afflicted with the disease. The dramatic effect of parathyroid hormone as a potential clinical bone anabolic drug underscores the hypothesis that substantial enhancement of skeletal mass requires stimulation of bone formation. Thus, identification of molecules, which promote systemic osteogenesis, is a major focus of anti-osteoporosis research. We have made the surprising observation that the key osteoclastogenic cytokine, RANK ligand (RANKL), when administered subcutaneously as a GST-fusion protein, is a potent bone anabolic agent. This compound dramatically enhances osteoblastogenesis and, within one week, stimulates exuberant bone formation, as detected radiographically, histologically and densitometrically. Importantly, GST-RANKL, at doses inducing as much as a 25-fold increase in osteoblast (OB) number, does not promote osteoclastogenesis in vivo. We also have established that OBs, and their precursors, are direct targets of GST-RANKL and have shown that collagen type I synthesis, by these cells, is greatly accelerated when they are exposed to the fusion protein. These data position GST-RANKL, or its derivatives, as potential bone anabolic, anti-osteoporosis agents. We therefore hypothesize that (1) GST-RANKL enhances OBs function by distinct signal pathways; (2) GST-RANKL, transcriptionally and/or post-transcriptionally, induces collagen type I synthesis by OBs; and (3) GST-RANKL prevents and/or reverses osteoporosis. Our Specific Aims are therefore to: (1) identify the signal pathways by which GST-RANKL enhances OB function; (2) identify the mechanism by which GST-RANKL induces collagen type I synthesis by OBs; and (3) determine if GST-RANKL prevents and/or reverses osteoporosis.

DESCRIPTION (provided by applicant): Inflammatory osteolysis, as occurs in rheumatoid arthritis and orthopedic implant loosening, reflects accelerated osteoclast (OC) recruitment and activation. Hence, discovering the means by which local inflammation recruits and activates OCs is central to preventing this crippling complication. Our overriding goal has been to detail the mechanisms by which cytokines, such as TNF, RANKL and M-CSF mediate bone loss. We have achieved the aims of our previous application by characterizing the contributions of various TNF target cells to the osteoclastogenic process and detailing many of the intracellular events by which TNF promotes osteoclastogenesis. In the present proposal we turn to the cytokine essential for all pathological bone resorption, namely RANKL, whose crystal structure, in complex with its receptor, RANK, we resolved in the current funding period. While RANKL is essential for OC differentiation, it also enhances the bone resorptive activity of the mature OC, a process we find depends upon the cytokine's regulation of the cytoskeletal adaptor protein, paxillin and the small GTPase, Rac. Furthermore, our resolution of the RANKL/RANK crystal structure and that of RANKL with the cysteine-rich domains of OPG, the decoy receptor, positions us to develop structure-based RANKL antagonists to arrest RANK signaling in the OC. Given that inhibition of the RANK signaling pathway has proven therapeutic benefit, these novel RANKL inhibitors may impact the treatment of pathological osteolysis. Thus, we hypothesize that: 1) RANKL activation of the OC is mediated via paxillin 2) RANKL activation of the OC is mediated via Rac and 3) Structure-based RANKL antagonists will arrest OC development and function. Our Specific Aims are therefore to 1) Determine the mechanisms by which paxillin mediates RANKL activation of the OC 2) Determine the mechanisms by which RANKL activates Rac in the OC and 3) Engineer variants of RANKL and OPG that disrupt RANK signaling in the OC. PUBLIC HEALTH RELEVANCE. Osteoclasts are the cells which destroy bone, and thus, their increased activity is responsible for most forms of pathological bone loss such as osteoporosis or that attending rheumatoid arthritis. The key molecule which activates ostoclasts is known as RANK ligand (RANKL). The purpose of this proposal is to gain insight into the mechanism by which RANKL activates osteoclasts and to design candidate drugs which will inhibit this event and consequently, prevent pathological bone loss.

Various members of the Ras superfamily of small GTPases, which comprise over 200 proteins, are expressed selectively in all cells, where they perform a range of cellular functions. A well-studied subfamily is that of the Rho GTPases, with the three best-understood members being Rho, Rac and cdc42. These molecules, including three isoforms of Rac, a number of Rho proteins and cdc42, a single gene product, regulate cell proliferation, differentiation, survival and many aspects of the cytoskeleton. We find cdc42 a particularly important regulator of osteoclast (OC) number and function. While increased cdc42 activity causes osteoporosis in mice, animals lacking the active GTPase specifically in their OCs have enhanced bone mass. Thus, cdc42 is a candidate therapeutic target for states of accelerated bone resorption. cdc42 deletion in OCs enhances their number by suppressing apoptosis, a process accompanied by increased amounts of the proapoptotic protein Bim. However, the impact of osteoclastic Bim on cell number and resorptive capacity in the context of cdc42 presence or absence is unknown. Moreover, Bim is also expressed in osteoblasts (OBs), where it regulates their lifespan and function. Finally, while Bim structure/function studies have been performed in a number of cell types, the residues that control OC and OB apoptosis are unknown. Based on these facts we hypothesize that: 1) absence of cdc42 in OCs arrests pathological bone loss;2) cdc42 and Bim act in concert in OCs to regulate apoptosis of bone resorptive cells, while Bim controls the OB and 3) specific residues in Bim govern its capacity to regulate OC and OB apoptosis. Given that we have generated mice with gain or loss of cdc42 function in OCs and have the capacity to selectively delete Bim in OCs or OBs in vivo, we are positioned to address the following specific aims: 1) determine the role of osteoclastic cdc42 in pathological bone loss, 2) determine how cdc42 and Bim, expressed in OCs and OBs, regulate apoptosis and bone mass in resorptive cells and 3) identify the specific residues in Bim that govern its capacity to regulate OC and OB apoptosis.

DESCRIPTION (provided by applicant): Maintenance of bone mass requires the integrated activity of osteoclasts (OCs) and osteoblasts (OBs). While they are functionally distinct, both are polarized and execute their activities by regulated secretion of bone-degrading and -synthesizing molecules, respectively. Bone resorption necessitates extracellular proton transport as a result of insertion of the vacuolar H+ATPase into the bone-apposed OC plasma membrane and targeted secretion of the collagenolytic enzyme cathepsin K. Similarly, skeletal synthesis involves secretion of specific matrix proteins such as type 1 collagen onto existing bone surfaces. Thus, the exocytic capacity of bone cells is fundamental to skeletal homeostasis. Regulated secretion, also called exocytosis, requires cell polarity in which intracellular, cargo-containing vesicles are delivered, by organization of the cytoskeleton, to an exocytic plasma membrane domain, which in OCs and OBs is that apposed to bone. Upon arrival at the site of exocytosis, the cargo-containing vesicles fuse with the plasma membrane via a number of linked steps, including high affinity binding of a synaptotagmin, a family of vesicle-associated adaptors, with several plasma membrane-residing proteins. We show data that synaptotagmin VII regulates exocytosis by OCs and OBs in a cell-autonomous manner. In consequence, mice lacking synaptotagmin VII have decreased remodeling and are a model for type II or low turnover osteoporosis. Cell polarity, a pre-requisite for exocytosis, is regulated by a conserved signaling pathway wherein activated cdc42 recruits an atypical protein kinase C, which phosphorylates essential downstream targets, including atypical PKCs such as PKC;. We find that RANKL-mediated organization of the OC cytoskeleton, and the cell's capacity to resorb bone, involve activation of the cdc42/PKC;pathway. Providing additional support for this posture, mice whose OCs express increased levels of active cdc42 are osteoporotic and those lacking the small GTPase have osteopetrosis. Similarly, cell-specific deletion of PKC;generates OCs that fail to polarize or resorb bone. These observations suggest a model in which 1) PKC;, activated by cdc42, promotes OC polarization and 2) following polarization, synaptotagmin VII controls exocytosis of bone- regulating molecules by OCs. Additionally Syt VII regulates OB exocytosis. Hence, our specific aims are to determine the mechanisms by which 1) PKC;signaling promotes OC polarization and 2) synaptotagmin VII controls exocytosis by OCs and OBs. PUBLIC HEALTH RELEVANCE: Maintenance of bone mass requires optimal and coordinated function of the cells regulating this parameter. We have identified novel proteins that individually control bone turnover. We propose to identify the mechanisms by which these two proteins control bone mass, with the view of generating novel drugs to treat diseases such as osteoporosis.

Abstract Obesity and osteoporosis are endemic in our society yet their relationship is perplexing. While obesity has long been considered beneficial for skeletal health, recent studies suggest bone mass is diminished in a substantial subset of obese individuals. Thus, despite its demographic importance, the influence of fat on bone remains enigmatic. Although controversial, studies of the effect of fat-produced molecules, such as leptin and adiponectin, indicate these selected adipokines impact bone. Adipose tissue is, however, a complex organ and there is little mechanistic insight as to how fat, per se, and which variety of fat, regulates the skeleton. Such information is clinically relevant as individuals with a predominance of visceral fat are osteopenic whereas subcutaneous and brown fat may positively influence bone mass. Determination of how fat, in its various forms, targets bone cells will provide the framework for ameliorating the skeletal complications of obesity. Resolution of this issue, in patients, is limited, however, by the absence of an animal model in which manipulation of fat abundance eventuates in a robust skeletal phenotype. To this end, we generated mice completely lacking visceral, subcutaneous and brown fat. Despite the hypogonadal state of these fat free (FF) mice, trabecular bone volume is strikingly increased (400-500%) due to enhanced osteoblast activity. Unexpectedly in face of its marked increase in bone mass, osteoclastogenesis in FF mice is also markedly enhanced. This observation raises the possibility that visceral fat diminishes bone mass by arresting osteoclast-induced remodeling. Our observations establish that, by mechanisms to be determined, fat signals to bone and decreased adiposity may greatly increase bone mass, challenging the concept that obesity generally improves skeletal health. Most importantly, the skeletal phenotype of FF mice is completely rescued by adipocyte precursor transplantation. This transplantation-mediated normalization of FF bone provides the opportunity to directly explore the impact of deleting various adipocyte products on bone accrual and how visceral, subcutaneous and/or brown adipose tissue targets the skeleton. Hence, we hypothesize that fat diminishes bone accrual in an osteoblast- and osteoclast-dependent manner.