Review

Pathogenic bone loss in rheumatoid arthritis: mechanisms and therapeutic approaches

Rheumatoid arthritis (RA) is a systemic inflammatory disorder with progressive joint destruction being its hallmark feature. Conventional treatments, such as DMARDs, aim to inhibit the inflammation; however, over the past decade, major advances in our understanding of bone metabolism has given us the ability to directly treat the bone loss in RA. Although modern anti-inflammatory therapies, such as anti-TNF-a treatment, have resulted in a remarkable improvement in the treatment of RA, these treatments do not directly target bone destruction in the joint. The aim of this review is to demonstrate that antiresorptive therapies can prevent structural joint destruction, particularly in the early stages of anti-inflammatory treatment. A number of novel approaches targeting bone resorption are also becoming available.

1 KEYWORDS: anti-inflammatory treatments n antiresorptive treatments n B ligand Melissa D Cantley , n NFATc1 n nuclear factor of activated T cells n osteoclasts receptor activator of Malcolm D Smith2 & nuclear factor n rheumatoid arthritis David R Haynes1† †Author for correspondence: Rheumatoid arthritis (RA) is a systemic inflam- joints in the hands and feet as well as the knees. 1Discipline of Pathology, School matory disorder characterized by destruction of Other synovial joints such as the ankles, hips, of Medical Sciences, University the joint, and this has a major impact on its func- elbows and shoulders are relatively spared of Adelaide, Adelaide, South tion. Progressive joint destruction is the hall- and the spine is rarely affected, apart from the Australia, 5005, Australia mark feature of RA. The disease is characterized atlanto-axial joint in the cervical spine. RA Tel: +61 883 033 180 by joint inflammation, synovial hyperplasia and rarely, if ever, affects distal interphalaengeal Fax: +61 883 034 408 associated destruction of bone and cartilage [1]. joints even though these are synovial joints. [email protected] It is the most common autoimmune disease in It is unclear why RA affects some synovial- 2Rheumatology Research Unit, Australia and currently affects 1% of the world’s lined joints frequently and others rarely or Repatriation General Hospital, population [201]. Despite the prevalence of this never. The primary joints affected include the Adelaide, 5041, South disease, complete understanding of the disease metacarpo-phalangeal, proximal interphalan- Australia, Australia process is still lacking. Most studies have focused geal, mid-carpal, radio-carpal and distal radio- on inflammation in the soft tissues; however, ulnar joints in the hands [2]. In the feet, the research is now focusing on preventing bone and target sites include the metatarso-phalangeal, cartilage destruction. Conventional treatments proximal interphalangeal and the intertarsal for RA aim to inhibit the inflammation; how- joints [2]. The disease results in pain, stiffness ever, over the past decade, major advances in our and swelling of the joints, which impacts on understanding of bone metabolism have enabled their functional capacity. Recurring inflam- us to identify the mechanisms of joint destruc- mation leads to bone and cartilage destruction, tion in RA. This knowledge, in combination affecting the individual’s physical functioning, with the recent development of therapies inhib- and causes both short and long-term morbid- iting bone loss, gives us the ability to directly ity [3]. The synovium is the major target for the treat the bone loss in RA. Although modern inflammatory process where infiltration of the anti-inflammatory therapies have resulted in a tissue occurs with multiple immune cells and remarkable improvement in the treatment of cytokines resulting in tissue inflammation. The RA, this review aims to demonstrate that anti- synovial tissue expands and forms a pannus that resorptive therapies, in combination with mod- invades the bone and cartilage, destroying the ern anti-inflammatory treatments, are needed to tissue as it proceeds [4]. The inflammation of the better prevent structural joint destruction and synovial tissues promotes osteoclastogenesis that disease progression in RA. leads to both focal articular bone erosion at the Rheumatoid arthritis presents clinically as site of pannus formation as well as systemic bone a symmetrical polyarthritis that targets many loss, similar to osteoporosis [4,5]. The invasion of part of joints in the body. RA only affects synovial- inflammatory tissue into the subchondral bone lined joints and predominantly affects peripheral involves many cell types including fibroblasts,

10.2217/IJR.09.42 © 2009 Future Medicine Ltd Int. J. Clin. Rheumatol. (2009) 4(5), 561–582 ISSN 1758-4280 561 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

lymphocytes and monocytes. Monocytes are fracture [17]. It is important to note that other the precursors of osteoclasts that resorb bone factors may contribute to the increased risk of through the acidic disillusion of bone mineral fracture such as the development of osteoporosis and enzymatic destruction of bone matrix [6]. owing to decreased mobility of the RA patient. Resorption of the unmineralized cartilage does not appear to directly involve osteoclasts Mechanisms of disease but seems to be due to proteases that are syn- A major hallmark of inflammatory bone loss thesized by the synovial fibroblasts, neutrophils diseases, such as RA, is enhanced osteo­clastic and the chondrocytes [7]. The articular cartilage bone resorption without a corresponding is avascular and nonmineralized, consisting of increase in bone formation. Osteoclasts are the chondrocytes that are embedded in the extra­ multinucleated cells derived from the mono- cellular matrix. The extracellular matrix con- cyte/macrophage hematopoietic lineage that sists of a type II collagen network that forms are responsible for resorbing bone [18–20]. The the structural backbone of the matrix along with essential role that osteoclasts play in bone loss other proteins [7]. There is a large amount of in RA has been demonstrated in animal studies proteoglycan in cartilage that binds water and showing that mice without osteoclasts are resis- functions as a mechanical shock absorber [7]. tant to arthritis-induced bone loss [21,22]. The The cartilage damage in RA can be attributed crucial role of osteoclasts in RA bone loss is also to a number of catabolic factors, such as pro- supported by studies of human tissue obtained inflammatory cytokines, aggrecanases, matrix from patients with RA [23–26]. In animal models metalloproteinases and nitric oxide. The syno- of RA, and in human RA, large multinucleated vial fibroblasts express large amounts of matrix- osteoclastic cells resorbing subchondral bone degrading enzymes that invade the articular have been detected at sites of bone erosion in the cartilage resulting in cartilage destruction [7]. joints [27,28]. These osteoclasts form as a result of Many of the factors produced by the inflamed exposure to inflammatory cytokines present in synovium that are involved in the bone destruc- synovial tissue, and understanding this process tion also have an effect on chondrocyte func- is the key to developing treatments for bone loss tioning [8]. Inflammatory cytokines, IL-1 and in RA. Figure 1 demonstrates the inflammatory TNF-a are thought to play a major role in car- cytokines involved in both localized bone loss tilage loss [8]. Many studies have demonstrated within the joint and generalized bone loss. that IL-1 stimulates chondrocytes to increase production of matrix metalloproteinases that „„Osteoimmunology degrade the cartilage [9,10]. The involvement of the immune system is fun- While the localized destruction of the joint is damental to the progression of RA. In recent the major focus of treatment, generalized bone decades the involvement of the immune system loss, known as osteopenia, results in a loss of in bone metabolism has become apparent (as bone mineral density (BMD) and elevates the reviewed by Takayanagi et al. [29]). Therefore, risk of osteoporotic fractures in patients with it is not surprising that bone loss in RA appears RA [11]. Clinical studies have evaluated BMD to be controlled by immune factors and cells. in the hands of patients with early RA and found This relationship between the immune system an association between reduction in BMD and and bone metabolism has been termed ‘osteo­ disease severity [12,13]. This increased systemic immunology’, and an understanding of osteo­ bone loss is attributed to the effects of pro­ immunology is vital for developing ways of inflammatory cytokines released from the site preventing bone loss in RA. of synovial inflammation that act systemically and result in a generalized bone loss [14–16]. RA „„Macrophage colony has also been demonstrated to have an impact stimulating factor on increasing the risk of fractures as a result of Macrophage colony stimulating factor (M-CSF) the associated generalized bone loss. One study was one of the first factors to be recognized as found an increased risk of fracture, especially of playing an essential role in the process of osteo- the hip (relative risk: 2.0) and the spine (relative clast development. It is normally produced by risk: 2.4), in RA patients compared with age and osteoblasts or bone marrow stromal cells and sex-matched controls. Predictors included a long transmits its signal by binding to its specific duration of RA, low BMI and use of cortico­ receptor, cFMS, which is a member of the tyro- steroids which, apart from a long duration of sine kinase receptor superfamily. This binding RA, are known risk factors for osteoporotic activates early transcription factors, such as c-fos,

562 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

Monocyte Pre-osteoclasts Macrophages Osteoclasts

RANKL RANKL

Local bone loss Lymphocytes IL-1, TNF, IL-6, IL-17, IL-23, etc. Fibroblasts Bone Inflamed Inflammatory cytokines synovium in joint

Pre-osteoclasts

Systemic bone loss RANKL Osteoclasts Osteoblasts

Figure 1. The pathogenesis of osteoclastogenesis in RA. Inflammatory cytokines in the RA joint stimulate osteoclast formation locally and externally to the bone in several ways. First, inflammatory cytokines can stimulate RANKL production by lymphocytes (e.g., IL-6 [181]) and fibroblasts (e.g., IL-6 [182]) in the joint. Second, they can activate monocytes (e.g., TNF-a [37], IL-23 [183]) that arrive at the inflamed joint to become macrophages and pre-osteoclasts (e.g., IL-1 [184], TNF-a [36]). Third, they may have a direct effect on pre-osteoclasts, usually in concert with RANKL, to directly promote osteoclast formation (e.g., IL-1 [185], IL-6 [186], IL-23 [187], TNF-a [185,188]). Systemic bone loss can also be a feature of RA and is stimulated by promoting osteoclast formation within the bone. In this situation, inflammatory cytokines act systemically to stimulate RANKL production by osteoblasts (e.g., IL-1 [189] IL-6 [190], IL-17 [191], IL-23 [187], TNF-a [60,189]) and may also directly promote osteoclast formation in concert with RANKL released by osteoblasts. RA: Rheumatoid arthritis; RANKL: Receptor activator of NF-kB ligand.

which initiates osteoclastogenesis. The main role „„Receptor activator of NF-kB ligand/ of M-CSF in osteoclast formation is to enhance receptor activator of NF-kB pathway the proliferation and survival of both precursor Receptor activator of NF-kB ligand (RANKL) cells and mature osteoclasts [30]. The essential role is a membrane-bound protein of the TNF family of M-CSF (known as CSF-1 in mice) in osteo- that is expressed by osteoblasts, fibroblasts and clastogenesis was demonstrated by the observa- activated T cells and is a key mediator in the tion that M-CSF knockout mice (op/op) were process of osteoclast formation. In normal bone osteopetrotic owing to the inhibition of bone metabolism, RANKL is expressed by osteoblas- resorption. Daily injections of M-CSF reversed tic cells; however, in inflammatory disease states, the osteopetrosis, highlighting its importance in it is also produced by immune cells, such as T osteoclast formation [31]. Sarma and Flanagan lymphocytes [33]. RANKL is also involved in the also demonstrated that M-CSF induced substan- hormone regulation of bone metabolism, since it tial osteoclast formation and bone resorption in is expressed in response to a variety of molecules cultures of human bone marrow stromal cells [32]. including parathyroid hormone (PTH), vitamin

future science group www.futuremedicine.com 563 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

D3 and IL-11 [34,35]. RANKL binds to its recep- RANKL expression [33–37] and enhancing the tor, receptor activator of NF-kB (RANK), on the effects of RANKL [47] is thought to be impor- surface of osteoclast precursors and activates the tant. However, the data strongly supports the c-jun terminal kinase that activates NF-kB to view that it is the elevated levels of RANKL that stimulate the formation of osteoclasts that resorb largely results in enhanced osteoclast formation bone [21]. RANKL has been demonstrated to be and enhanced bone resorption in RA. essential for osteoclast bone resorption [21,22] although, under some circumstances, TNF has „„Osteoprotegerin been reported to induce osteoclast formation in Osteoprotegerin (OPG) is a soluble TNF recep- the absence of RANKL [36]. However, according tor-like molecule that is a natural inhibitor of to others, TNF alone was insufficient to induce RANKL. It acts as a decoy receptor and pre- osteoclastogenesis, suggesting an essential role vents the binding of RANKL to RANK, thereby of RANKL in osteoclast formation [37]. The preventing osteoclastogenesis. OPG knockout importance of RANKL in the process of osteo- mice have an osteoporotic phenotype and have a clast bone resorption was demonstrated using high incidence of fractures [48]. Conversely, over­ RANKL knockout mice that were found to expression of OPG in transgenic mice resulted exhibit an osteopetrotic pheno­type owing to an in the development of severe osteopetrosis absence of osteoclasts [21]. Conversely, excessive owing to the inhibition of osteoclastic bone administration of recombinant RANKL in mice resorption [49]. Administration of OPG to mice results in increased osteoclast formation that was shown to prevent TNF-a-mediated bone leads to bone resorption, and the mice develop destruction by reducing the number of osteo- osteoporosis [34]. clasts able to resorb bone [50]. Active osteoclasts Receptor activator of NF-kB ligand has can be isolated from the pannus joint from RA been demonstrated to form this important patients and there is a correlation between the link between immunology and bone physiol- ratio of RANKL mRNA:OPG mRNA and ogy since inflammatory cytokines, such as resorption ex vivo [24]. Furthermore, the level of TNF-a and IL-1, stimulate the production of OPG protein was markedly reduced in patients RANKL [38,39]. Previous studies have provided with RA [51], whilst higher levels of RANKL evidence for implicating RANKL in the patho- were expressed by lymphocytes and macro- genesis of the bone destruction that occurs in phages in tissues extracted from patients with RA. Elevated expression of RANKL has been active RA [41]. Figure 3 demonstrates the lower found in the synovium of patients with RA, and OPG expression levels observed in active RA tis- therefore may contribute to the excessive bone sue compared with normal tissue. These results loss observed [24,40,41]. We previously demon- were also confirmed by Petitt and colleagues strated higher levels of RANKL expression in [43] who demonstrated that OPG was expressed the mononuclear aggregates and fibroblast-like remotely in the synovium of RA patients and cells in the subintimal regions of the synovial was limited at sites of bone erosion, especially membrane from patients with active RA [42]. in regions that were associated with RANKL It was later confirmed that the high expression expression. These results confirm that the of RANKL is largely confined to sites of bone imbalance between RANKL and OPG plays resorption at the pannus–bone interface and a major role in bone destruction in RA, and subchondral bone erosion [43]. Figure 2 clearly that OPG expression is a crucial determinant demonstrates the enhanced RANKL expres- of the degree to which RANKL can stimulate sion in RA synovium compared with normal ­osteoclast development [43]. synovium. Elevated expression of RANKL by inflammatory cells, such as synovial fibroblasts „„Intracellular regulators and activated T-cells, has also been demon- of osteoclasts strated to be associated with RA [40–41,44,45]. Binding of RANKL to RANK activates a num- Using murine spleen cells, peripheral blood ber of intracellular factors including TRAF-6, mononuclear-derived T cells were demonstrated c-fos and calcium signaling pathways that regu- to promote osteoclast formation and activation late osteoclast differentiation and activation, as in vitro [45]. In a rat model of adjuvant arthritis, shown in Figure 1. These are all responsible for local T cell activation was also found to be asso- the induction and activation of nuclear factor ciated with RANKL expression and subsequent of activated T cells-1 (NFATc1), which has been joint destruction [46]. The involvement of a vari- found to be the master transcription factor for ety of inflammatory cytokines in stimulating osteoclastogenesis [52].

564 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

„ „TRAF-6 TRAF-6 is a member of the TNF receptor factor (TRAF) family of proteins that plays an impor- tant role in linking the interaction of RANK on the cell membrane with the adaptor proteins downstream. Binding to TRAF adaptor proteins within the cytoplasmic domain of RANK is the initial step in the process of RANK signaling [53]. In a gene knockout study, TRAF-6-deficient mice were demonstrated to develop a severe osteopetrotic phenotype and to be defective in osteoclast formation owing to defective signaling from RANK upon RANKL binding [54]. The downstream targets of TRAF-6 are transcrip- tion factors including NF-kB, activator pro- tein-1 (AP-1) and NFATc1, cascades of mitogen- activated protein kinases (MAPKs) such as p38 stress kinase, c-jun N-terminal kinase (JNK), ERK and Pi3K/AKT pathways [53].

„„NF-kB Figure 2. Expression of RANKL in synovium of an RA subject and a normal NF-kB is a pleiotropic transcription factor that subject. RANKL expression detected with mAb 626/immunoperoxidase and AEC is also critical in RANK signaling and the pro- (red) in synovial tissues from (A) a patient with active RA and cess of osteoclast formation and activation. It is (B) a normal subject. also likely to be a significant factor in regulat- AEC: 3-amino-9-ethylcarbazol; mAB: Monoclonal antibodies; RA: Rheumatoid arthritis, RANKL: Receptor activator of NF-kB ligand. ing synovial inflammation[55] . Its importance in osteoclastogenesis was confirmed by investiga- tions using NF-kB knockout mice, who devel- activating transcription factor (ATF) family oped osteopetrosis as a result of defective osteo- of proteins [60]. C-fos is a transcription factor clast formation [56]. NF-kB proteins reside in and is a member of the FOS family that has the cytoplasm of nonstimulated cells and rapidly been implicated in a number of metabolic pro- enter the nucleus when stimulated, for example, cesses, including both the skeletal and immune by RANKL. This classical NF-kB is activated systems. C-fos expression is induced by both by the binding of RANKL to RANK, which M-CSF and the RANKL/RANK interaction. activates the IkB kinase (IKK) complex [57]. The C-fos-deficient mice have been demonstrated IKK complex consists of two catalytic sites – the to develop a severe osteopetrotic phenotype, IKK-a and IKK-b (IKK-1 and IKK-2) – and the demonstrating its importance in the process of associated regulatory subunit IKK-g. The IKK-b osteoclast formation [61,62]. C-jun is a member of is essential for NF-kB activation, as it phosphor- the JUN family of proteins, some of which play ylates the inhibitor subunit IKB-a, which leads a key role in osteoclastogenesis. This has been to its degradation and subsequent activation of demonstrated through studies with transgenetic NF-kB [58]. In the alternate activation pathway, mice expressing dominant-negative c-jun, where IKK-a is required for NF-kB activation, result- it was demonstrated that severe osteopetrosis ing in phosphorylation and proteasome-induced results from impaired osteoclastogenesis [63]. processing of p100, which is cleaved to generate C-Jun was also demonstrated to play a crucial an active p52 product [57,59]. Both of these path- role in the regulation of the NFAT family; over ways that activate NF-kB are important in the expression of NFAT promoted osteoclast forma- process of osteoclastogenesis. tion that could be suppressed by overexpressing dominant-negative c-jun [63]. „„AP-1 family: c-fos & c-jun Interaction of RANK and RANKL activates the „„Nuclear factor of activated T cells transcription factor complex AP-1. The AP-1 NFATc1 is the key intracellular molecule transcription factor is a dimeric complex that that regulates terminal osteoclast forma- consists of the FOS family of proteins (c-fos, tion (Figure 4) [64]. RANKL binding to RANK FOSB, FRA1 and FRA2), the JUN family of recruits TRAF adaptor proteins to activate proteins (c-JUN, JUNB and JUND) and the key signaling cascades that promote the

future science group www.futuremedicine.com 565 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

cyclophosphamide and abatacept) were origi- nally tested in animal models and their effi- cacy in these models results in progression to clinical studies. The models used to study RA include adjuvant-induced arthritis (AIA), collagen-induced arthritis (CIA), collagen antibody-induced arthritis (CAIA), K/BxN Tg mouse, hTNFaTg mouse and the severe combined immunodeficiency (SCID) mouse model [4]. The advantages and disadvantages of using each of these models to investigate RA and test possible treatments are shown in Table 1. The most commonly used animal models are AIA in rats and the CIA and CAIA models in mice, as these have many features in common with human RA.

„„Rat adjuvant-induced arthritis model The rat adjuvant-induced arthritis (AIA) model was the first animal model of RA and is now Figure 3. OPG expression in synovium of an RA subject and a normal widely used to test new therapeutic agents. This subject. OPG staining (mAb 805) of synovial tissues from (A) a patient with active is a T-cell- and neutrophil-dependent disease [70] rheumatoid arthritis and (B) a normal subject. that is induced by a single intradermal injection mAb: Monoclonal antibodies; OPG: Osteoprotegerin. of complete Freund adjuvant, resulting in a rapid differentiation of monocytes into multinucle- progressive onset of the disease 7 days later. The ated mature osteoclasts (as reviewed by Asagiri disease is characterized by severe joint inflamma- and Takayanagi [64]). NFATc1 is the terminal tion and marked bone resorption with histology, factor that directly induces the expression of taken at the completion of the study (usually osteoclast genes such as the calcitonin receptor 14 days after adjuvant injection), demonstrat- (CTR), cathepsin K (Cath-K), tartrate-resistant ing neutrophil infiltration and subchondral bone acid phosphatase (TRAP) and the b3 integrin, destruction. Drugs currently approved for RA in addition to the osteoclast-associated recep- treatment that demonstrated therapeutic efficacy tor (OSCAR) [65–67]. The ligand for OSCAR is in the AIA model include methotrexate [71–75], yet to be identified. The essential requirement ciclosporin A [76,77], gold compounds [78], peni- for NFATc1 is evident by its ability to induce cillamine, prednisone, cyclophosphamide, indo- osteoclast-specific genes in the absence of methacin, anakinra [79,80], naproxen, celecoxib, RANKL [68]. Furthermore, NFATc1-deficient flobufen and etanercept[80,81] . embryonic stem cells are unable to differentiate into osteoclasts [69]. „„Collagen-induced arthritis model Animal models of disease Collagen-induced arthritis (CIA) is another Many developments in understanding the pro- commonly used arthritis model to test poten- cess of bone metabolism and the mechanisms tial treatments. It is induced by immuniza- of joint destruction in RA have been made tion with heterogeneous type II collagen in through the use of animal models. These mod- complete Freund’s adjuvant. Mice develop a els are widely used to investigate the mecha- polyarthritis with severe cartilage and bone nisms involved in the RA disease process and destruction. This animal model of disease also play a critical role in the development shares several clinical, histopathological and of drugs for treating RA. The models used immunological features with the human RA are known to share features with the human condition and can therefore be effectively used disease process and can therefore be used to to test therapies that target both inflammation investigate the mechanisms of actions and to and bone loss [70]. Drugs currently approved for determine any associated side effects of novel RA treatment that demonstrated therapeutic therapies. Many of the current treatments now efficacy in the CIA model include prednisone, on the market for RA (such as ciclosporin A, cyclophosphamide, indomethacin, anakinra, indomethacin, anakinra, celecoxib, etanercept, etanercept and abatacept.

566 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

„ „Collagen antibody-induced early and late phases of disease [85]. The ease of arthritis model disease induction and the close resemblance to Collagen antibody-induced arthritis is a newer the human disease has resulted in this model animal model that is a rapid, simple and versa- becoming commonly used to investigate RA. tile model of RA induced by systemic adminis- tration of a cocktail of type II collagen mono- „„TNF transgenic inflammatory clonal antibodies that are directed to conserved mouse model auto-antigenic epitopes of collagen type II, Transgenic mice are commonly used to study either alone or in combination with lipopoly- the roles that particular genes play in the RA saccharide (LPS) [82,202]. LPS and the auto- disease process. For example, in the TNF antibodies to type II collagen have a synergistic transgenic (TNFtg) mouse model, mice have action that induces arthritis [202]. This model a human TNF-a transgene modified in the 3´ is known to exhibit pathogenic features that region, and they therefore express high levels of are similar to those in human RA. It has pre- both soluble and membrane-bound TNF-a [86]. viously been demonstrated to result in severe These mice develop a symmetrical polyarthritis inflammation in the ankles and paws of mice approximately 4–6 weeks after birth. Multiple [83], with clinically apparent arthritis occurring lines of TNFtg mice have been generated with- by day 4 and active arthritis peaking at day 6 out any other obvious developmental defects [87]. [82,83]. Tissue from paw joints demonstrated Interestingly, no joints are spared, including the marked pathological changes with synovial tempomandibular joint [87,88]. hyper­plasia, a large number of infiltrating poly- morphonuclear and mononuclear cells, exten- „„Measuring bone loss in animal sive pannus formation at the cartilage–bone models of RA junction, severe cartilage destruction and bone The animal models of RA are commonly used erosion [83,84]. The development of CAIA was to test therapies that inhibit the bone destruc- originally thought to be independent of a T- tion, so ana­lysis techniques are required to and B-cell mechanism; however, it has recently quantitate the bone loss and to determine if been demonstrated that T-cell activation is antiresorptive drugs effectively suppress this important for the development of CAIA [85]. bone loss. Many of the current methods used, It is likely that T cells play an important role such histo­morphometry and 2D radiography, in the prolongation of CAIA. An inhibitor are not effective at giving a true indication of of T-cell activation, ciclosporin A, was dem- bone loss and can also be prone to measure- onstrated to reduce the severity of arthritis ment errors. One of the most effective methods developed using a CAIA model in both the that is now commonly used to determine the

FcRγ OSCAR-L RANKL

ITAM OSCAR

RANK

++ TRAF 6 Ca cFos Mitf OSCAR PU.1 NFATc1 transcription

Osteoclastogenesis

Figure 4. The important intracellular regulators of osteoclast formation. NFATc1 is the central intracellular molecule regulating terminal osteoclast formation. RANKL binds to its receptor, RANK, and the TRAF adaptor proteins are recruited, prompting the differentiation of monocytes into multinucleated mature osteoclasts via activation of key signaling cascades. RANK: Receptor activator of NF-kB; RANKL: Receptor activator of NF-kB ligand.

future science group www.futuremedicine.com 567 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review Ref. [79,80] [86,87] [86,88,89] [80,85,86] [53,71–75] [71,72,76–80] [79,81–84,202] days, and no pannus or erosions

Development periostitis, of bony ankylosis and most the of extra-articular manifestations Self limiting and not characterized exacerbations by and remissions The cell infiltrate is predominantly polymorphonuclear cells, in humans high proportion mononuclear of cells observed Mouse model has slower onset Preparation CII of is a time-consuming procedure sourcesCommercial purified of CIInative are expensive Unclear whether and B-cell it is T- dependent Genetic modificationsof mice can have unpredictable consequence the to immune system and also other to phenotypic features Genetic modificationsof mice can have unpredictable consequence the to immune system and also other to phenotypic features The arthritis induced is apparently as acute, is not reported past 7 were seen Disadvantages resorption and cartilage destruction T cell and neutrophil dependent disease beCan either or rat in mouse Short onset time Marked cartilage destruction, bone resorption, periosteal proliferation, marked synovitis and periarticluar inflammation Paws are characterized extreme by swelling and erythema and B-cellT- dependent and responses can be measured a cocktailof anti-type of II collagen monoclonal antibodies alone or with a combination lipopolysaccharide of The severity disease of can also be controlled changing by the dose the of monoclonal antibody cocktail Develop synovial hyperplasia with a large number infiltrating of polymorphonucelar and mononuclear cells, extensive pannus formation at the cartilage-bone junction, severe cartilage destruction and bone erosion Several mouse strains can develop arthritis without administration external any of antibody or antigen Mice spontaneously develop chronic progressive inflammation, occurs at 3 weeks ageof then progressively evolves chronic to inflammation A simple model in which mice develop a deforming arthritis approx 4–6 weeks after birth Mice develop symmetrical polyarthritis Muliple lines TNFtg of mice been have generated without obvious any developmental defects No joints are spared from inflammation and destruction, also include tempomandibular joint severe, erosiveChronic, arthritis is induced There are spontaneous remissions and exacerbations similar the to human disease investigating cartilage destruction in RA Many the of features RA of weeks last at for least 12 RatRat, mouse Reliable, rapid onset easily measurable polyarticular inflammation, marked bone Shares both immunological and pathological features human of RA Mouse Rapid, simple and versatile model RA of that is induced systemic by administration RatMouse A reliable well characterized model that is T-cell dependent Severe combined immunodeficiency mouse modelfeasible is a methodfor 1. The1. advantages and disadvantages of animal models of RA.

Tg modelTg Mouse Relatively fast and cost efficient a Table Model Adjuvant-induced arthritis AnimalsCollagen-induced arthritis Advantages Collagen antibody-induced arthritis K/BxN mouse Tg Mouse Relatively fast and cost efficient model involving B andcells T hTNF Streptococcal cell wall arthritis Severe combined immunodeficiency RA: Rheumatoid arthritis.

568 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

extent of bone loss in animal models is micro- computed tomography (CT). This has many distinct advantages as it enables the detailed 3D micro architecture of bone to be examined. Various studies have used the method of micro- CT to quantitate the volume of bone loss in models of arthritis [84,89–91]. New technology has now been developed that allows quanti- tative live animal micro-CT analyses to be performed over the duration of a time-course experiment [92]. This allows investigators to follow live animals throughout the disease process and to monitor changes in bone loss before and after drug administration. Animals can be repeatedly scanned throughout a study and this means that they are able to act as their own controls. This allows investigators to determine the effects that drugs have on bone loss in individual animals, thereby reducing sources of error, as one major problem encoun- tered in animal models of RA is the difference in susceptibility to disease development. The micro-CT scanners also have a high-resolu- tion camera (9 or 18 µM) that produces more detailed scans and means that subtle changes in bone are able to be detected. Although a detailed ana­lysis of cartilage destruction can not be investigated, this can be assessed via narrowing of the joint space. Using the CAIA model and the new method of live animal micro-CT, we have been able to demonstrate severe pitting and a loss of bone mass (Figure 5). Abnormal bone growth also demonstrates fusion of the smaller wrist and foot bones, which can also be seen in this late stage of disease that is similar to human RA. Therefore, this model can be effectively used to test potential inhibitors of bone destruction. Figure 5. 3D models of left-hand side front mouse paws. (A) Control paw, no disease. (B) Paw with collagen-antibody induced arthritis. Therapeutic approaches to treat RA The main risk of bone erosion in RA occurs „„Conventional RA therapies in the first 2 years, and earlier treatment tar- NSAIDs geted at the initial stages of bone loss is more NSAIDs have long been the first line of treat- likely to cause arrest [93]. The main goals of ments for pain and inflammation in RA. This RA treatment are to relive pain, control the group of drugs block inflammation in RA by inflammation and prevent joint destruction preventing prostaglandin release through inhi- [94]. The following section is divided into two bition of the cyclooxygenase enzymes (COX) parts. First, conventional therapies for RA that of which there are two isoforms – COX-1 and tend to focus on reducing inflammation in the COX-2. Aspirin is one of the commonly used soft tissues, and second therapies that target the NSAIDs for the treatment of RA but it has been bone destruction of RA. This review will focus reported to have no effect on preventing joint on the effects of these drugs on both local- destruction [94]. Since many of the traditional ized bone erosion and generalized bone loss, NSAIDs have been found to have adverse side with changes in bone markers and fractures effects, including major gastrointestinal selec- only briefly discussed since less is known about tive COX-2 inhibitors such as celecoxib, more these markers. recently, robecoxib and valdecoxib have been

future science group www.futuremedicine.com 569 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

used to treat RA [94]. Although conventional suppressing osteoblasts and bone formation; NSAIDs and selective COX-2 inhibitors have however, they may exert a moderate beneficial been demonstrated to reduce inflammation in effect through the suppression of inflammatory- RA, they can have ambiguous effects on bone. induced bone erosions [7]. Glucocorticoids may Under pathological conditions, prostaglandins play a role in suppressing early inflammation in

such as prostaglandin E2 can stimulate osteo- RA; however, with their adverse side effects from clast-mediated bone resorption [95–97]. On the long-term use and their inhibition of bone forma- other hand, prostaglandins produced as a result tion, they are not the most useful treatment to of COX-2 stimulation are reported to increase prevent structural damage in RA. bone formation by stimulating osteoblast for- mation and in association with hormones and DMARDs growth factors, prostaglandins have also been One of the most common conventional thera- demonstrated to play an important role in bone pies now widely used for the effective treatment healing [98,99]. Although NSAIDs have been of RA are DMARDs. There are a variety of demonstrated to have many benefits in treating DMARDs used to control RA including sul- the symptoms of RA, this group of drugs gener- fasalazine, antimalarials, penicillamine, gold, ally appears not to have a major benefit on bone methotrexate (MTX), azathioprine, leflu- destruction in RA. nomide and cyclophosphamide. These can be Modified NSAIDs could provide an alter- used as monotherapy or in combination with native treatment to conventional NSAIDs for other therapeutic agents. DMARDs are small the treatment of RA. Nitrosylated flurbiprofen molecular weight, orally available drugs that derivative HCT1026 has been demonstrated to have anti-inflammatory properties and are used inhibit bone resorption, both in vivo and in vitro. for long-term control of the disease process [7]. HCT1026 was demonstrated to strongly sup- MTX is a widely used DMARD and is the press osteoclast formation, activity and survival standard of care for patients with moderate-to- in murine osteoclast cultures by inducing osteo- severe arthritis owing to the fact that it is safe clast apoptosis; however, there was no effect on and well tolerated at therapeutic doses [103,104]. osteoblasts or macrophages. This derivative It appears to act through the induction of ade- compound also inhibited RANKL-induced nosine, which has anti-inflammatory proper- NF-kB and extracellular signal-regulated kinase ties. MTX is often used in combination with (ERK) pathways, whereas the parent compound, other agents such as etanercept [105,106], adalim- flurbiprofen, did not. Since this derivative is not umab [107], anakinra [108], leflunomide[109] and a COX inhibitor due to the presence of a nitric rituximab [110] to treat RA. MTX effectiveness is oxide group, this reduces the side effects related achieved by providing an additional anti-inflam- to gastrointestional toxicity [100]. matory mechanism and/or by acting synergisti- cally with other therapies [7]. DMARDs have Glucocorticoids resulted in improvements in symptoms in a sub- Glucocorticoids are a class of steroid hormones stantial proportion of RA patients; however, not that bind to the cortisol receptors and have a vari- all patients respond favorably. There have also ety of biological effects. They have been found to been issues relating to toxicity and adverse events have anti-inflammatory effects and are commonly [103] and for these reasons, patients tend to take used in the treatment of RA. Glucocorticoids time to find an effective DMARD that works were originally found to be reliable and have well for them, and when they do find a suitable rapid effects on controlling joint inflammation DMARD, it can take time to develop substan- [94]. However, long-term use of glucocorticoids tial anti-inflammatory effects. Unfortunately, has been demonstrated to be associated with during this time, irreversible bone and cartilage the induction of osteoporosis. Treatment with destruction can continue to occur. In addition, glucocorticoids has been found to be associated only a minority of patients continue with effec- with an increased risk of fractures, particularly tive DMARD treatment over the long term, as of the hip and vertebrae, and this is dose depen- the majority discontinue use as a result of adverse dent and occurs rapidly after treatment is com- side effects [111]. menced [101,102]. Other side effects of glucocorti- coid therapy include hypertension, peptic ulcer Anti-TNF-a therapy diseases, accelerated atherosclerosis, vascular dis- During the process of inflammation, a number ease and many others [94]. Glucocorticoids have of cytokines such as TNF-a, IL-1, IL-6 and also been found to have negative effects on bone, IL-17 are involved in the differentiation and

570 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

activation of osteoclasts that are responsible for The effects of anti-TNF-a treatment on gen- resorbing bone [112]. TNF-a, one of the major eralized bone loss have been investigated and mediators involved in RA, is produced by acti- small increases in BMD of the spine and hip vated macrophages as well as synoviocytes in have been found in patients on TNF blockers, the inflamed synovial tissue. TNF-a is reported while RA patients on MTX alone have small to both directly and indirectly induce osteoclast decreases in BMD [114,115]. However, many of formation and hence, it forms a link between these changes were not clinically significant. It the immune and bone systems. TNF-a-induced has been suggested that the mechanisms under- osteoclast formation may require both M-CSF lying localized bone loss around a joint, bone and RANKL, and its resorptive effects may be erosions and generalized osteoporosis are due, due to its enhancement of RANKL activity [48] at least in part, to TNF, and anti-TNF therapy and its ability to stimulate RANKL produc- may therefore have benefits on both localized tion by osteoblasts [7]. TNF-a has also been and generalized bone loss [114,115]. demonstrated to induce osteoclast formation independently of the RANKL/RANK interac- Interleukins: IL-1, IL-3, IL-6, IL-12, IL-15, tion [36]. The fact that TNF-a is able to induce IL-17 & IL-23 osteoclast formation in both the absence and In addition to TNF-a, other inflammatory cyto- presence of RANKL has resulted in it being a kines such as IL-1, IL-6, IL-12, IL-15, IL-17 and target for therapies aimed at suppressing both IL-23 play a role in the pathogenesis of RA, and the inflammation and bone destruction in RA. these have become targets for therapeutic agents. Anti-TNF-a therapy has become increasingly For example, IL-1 is an inflammatory cytokine popular for the treatment of RA owing to its rapid that plays a key regulatory role in the disease onset of action, which reduces joint destruction process of RA. Like TNF-a, IL-1 has been dem- compared with the slower acting drugs. There onstrated to increase expression of both RANKL are currently five TNF-a antagonists approved and RANK and to facilitate osteoclast differenti- for RA treatment: ation [116]. Anakinra, a recombinant IL-1 recep-

n Adalimumab – human monoclonal antibody tor antagonist (IL-1Ra), is currently approved for the treatment of RA [7]. The effectiveness of n Entanercept – TNF receptor (p75): FclIgG anakinra has been investigated as both a mono- construct therapy and in combination with MTX [108]. It

n Infliximab – chimeric monoclonal antibody has been demonstrated to effectively reduce focal bone erosion in arthritis [108,117–119]. Novartis n Golimumab (CNTO148) – fully human AG (Basel, Switzerland) have developed a monoclonal anti-TNF-a antibody fully humanized monoclonal antibody target- ® n Certolizumab pegol (Cimzia , UCB, Brussels, ing IL-1 that neutralizes the activity of human Belgium) – humanized anti-TNF-a antibody IL-1b, known as canakiumab (ACZ885). This is administered by the intravenous or subcutane- All of these treatments work through inhi- ous route and is currently in Phase III clinical bition of TNF-a activity with the main dif- development with promising early results [120]. ference being in the method of delivery and A number of new therapies are being devel- the frequency of administration. However, oped against other cytokines involved in the the greater efficacy of treatment with TNF-a inflammatory cascade. IL-3, a cytokine that is inhibitors comes at a much higher cost and with secreted by T-helper (Th) cells, stimulates the the fact that they do not work adequately on all proliferation, differentiation and survival of patients [103]. In addition, continuous parental pluripotent hematopoetic stem cells. In bone administration is associated with an increased marrow stem cells isolated from mice, IL-3 was risk of infection [103], with reactivation of TB demonstrated to irreversibly inhibit osteoclast being a major concern. As a result, anti-TNF-a formation and resorption in a dose-depen- therapies are not usually initiated until other dent manner [121]. It was also demonstrated less expensive treatments have been tried and to inhibit TNF-a-induced bone resorption erosions may have already occurred before anti- in vitro [121]. In a mouse model of monoclonal TNF-a therapy takes effect. Even treatments antibody/LPS-induced arthritis, pretreatment commenced at the onset of disease may not pre- with IL-3 also resulted in less inflammation in vent joint damage as recent evidence has demon- the paws, indicating that IL-3 treated mice are strated that bone damage can still occur during resistant to monoclonal antibody/LPS-induced anti-TNF-a therapy [113]. arthritis [121].

future science group www.futuremedicine.com 571 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

Other inhibitors of interleukins that are in (etidronate, tiludronate and clodronate) [127]. clinical trails include tocilizumab, a humanized The amino-BPs are known to have a stronger monoclonal antibody against the IL-6 receptor antiresorptive effect since they induce osteoclast with several Phase II clinical trials suggesting apoptosis. The mechanism of action of BPs var- that tocilizumab may be effective in treating ies and some may be better suited to the treat- RA [122]. Earlier this year, tocilizumab was ment of systemic bone loss whereas others may approved for clinical use [123]. HuMaxIL-15 is a be better for reducing focal bone erosions, such high-affinity fully human monoclonal antibody as that seen in RA [127]. In an AIA model, two against IL-15 that has been tested in Phase I– BPs, etidronate and alendronate, were demon- II clinical trials and has resulted in substantial strated to prevent both paw swelling and bone improvements in disease activity as well as loss [128]. Itoh and colleagues showed that both being well tolerated [124]. A monoclonal anti- clodronate and alendronate reduced systemic body against IL-17 (AIN457) and a monoclo- bone loss in a rat CIA, and clodronate also pro- nal antibody against p40 subunits of cytokines tected against focal bone loss [129]. Zoledronic IL-12 and IL-23 – ustekinumab [125] – are also acid is a new generation BP that is thought to in clinical trials. suppress bone loss by reducing the life span of While these novel anticytokine therapies osteoclasts [130]. In a study using mice trans- offer useful options for the treatment of RA, genic for human TNF that spontaneously they do not directly inhibit osteoclast-mediated develop arthritis, zoledronic acid was demon- bone destruction so the bone and cartilage loss strated to have no effect on inflammation, but may continue to progress despite treatment. The a single dose significantly retarded bone erosion involvement of a variety of inflammatory cyto- and repeated administration almost completely kines in stimulating RANKL expression [33–35] blocked bone erosion [131]. This was confirmed and enhancing the effects of RANKL [47] to in another study using a CIA model of arthritis induce the formation of osteoclasts that resorb in which zoledronic acid effectively suppressed bone is important. However, therapies that spe- structural joint damage. It significantly reduced cifically target osteoclast bone resorption com- focal bone erosions and juxtaarticular trabecu- bined with therapies targeting inflammation, as lar bone loss, even though inflammation was discussed above, are likely to more effectively slightly increased [132]. These observations of BPs stop the joint damage and disease progression having no effect on inflammation confirms the of RA. importance of combination therapy with anti- inflammatory agents such as DMARDs and „„Antiresorptive therapies antiresorptive therapy such as BPs. In a more Despite the use of conventional and emerging recent study, zoledronic acid was tested in a small therapies that target the inflammation, struc- group of 39 patients with RA. Treatment was tural damage can still progress that results found to reduce the number of MRI hand and in permanent joint damage in RA patients. wrist bone erosions by 61% compared with the Therefore, there is a need for effective drug placebo [130]. Both animal and clinical studies therapy that will preserve the joint structure [6]. indicate that BPs protect against generalized A number of options have been suggested that bone loss, but not all of them result in a signifi- may be used in combination with conventional cant reduction in focal bone damage [127]. The anti-inflammatory treatments. ability of a particular BP to protect the joint in RA also seems to be dependent upon the disease Bisphoshonates model used and also on the particular BP used. Bisphosphonates (BPs) are a group of drugs that are known to inhibit osteoclast activity and are Hormone replacement therapy used to treat a wide variety of bone disorders, The role of hormones in the regulation of bone including osteoporosis, tumor-associated osteo­ metabolism is well known and hormone replace- lysis and arthritis [112]. The first generation BPs ment therapy has been demonstrated to be were synthesized in the early 1970s and only beneficial in preventing postmenopausal osteo­ exhibited weak activity against osteoclasts, porosis [133,134]. The use of hormone replacement whereas the later generation agents exhibit stron- therapy as a treatment for RA is very controver- ger antiresorptive effects [126]. There are two sial as early studies indicated a protective effect main classes of BPs – amino-BPs (alendronate, of estrogen against RA [135], whereas more recent ibandronate, risedronate, pamidronate, zole- studies have disputed this observation [136]. dronic acid and icandronate) and nonamino-BPs Different effects have been reported with some

572 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

studies demonstrating a significant effect on reduce bone resorption for up to 6 months [144]. the disease process whilst others show no effect. Repeated injections resulted in significant As a result of the potentially severe side effects, increases in bone mass in the axial and periph- hormone replacement therapy is no longer rec- eral skeleton, in trabecular and cortical bone. ommended as a treatment for RA [137]. Animal To date, no adverse side effects have been studies using a rat CIA model demonstrated that reported [143]. Injections of denosumab twice estrogen treatment over a 28 day period delayed yearly with ongoing MTX treatment were found the time of disease onset and reduced the degree to inhibit structural damage in patients with RA of disease, thus demonstrating a chondroprotec- with no increased rate of side effects [145]. At the tive effect [138]. However, this is not supported 6 month time point, the increase in MRI ero- by human studies where in a randomized con- sion scores was significantly lower in the group trolled trial with 27,347 postmenopausal sub- treated with 180 mg of denosumab and there jects, hormone therapy was demonstrated to was also a significant difference in the modified have no significant effect on reducing the risk sharp erosion score at 6 months. In addition, of developing RA, or on reducing the severity there sustained suppression of the bone turn- of RA [136]. Raloxifene is a selective estrogen over markers (serum CTX-1 and PINP) was receptor modulator that has been approved for observed. Little effect on cartilage was noted the treatment of postmenopausal osteoporosis as cartilage turnover markers did not decrease [139]. Using a CIA model of arthritis, raloxifene at the 6 month time point [145]. Although deno- treatment (at 60 µg/mouse/day) was demon- sumab significantly reduced bone destruction, strated to significantly reduce the frequency denosumab did not, however, have an effect on and severity of the disease, with very little joint RA inflammatory activity observed as RA flares, destruction being observed in raloxifene-treated or on joint space narrowing, and had no signifi- animals. Levels of the proinflammatory cytokine cant effect on reducing cartilage erosion [145]. It IL-6 were also significantly lowered in treated was suggested that this lack of effect on joint mice, indicating that raloxifene may also have space narrowing could be related to the specific an anti-inflammatory effect [139]. These results action of denosumab, that is the inhibition of clearly demonstrate the conflicting and complex RANKL, or as a result of insufficient dosages effects that hormone replacement can have on used [145]. postmenopausal women. The role of the RANKL/OPG system in car- tilage destruction is still not known. Inhibition RANKL/RANK interactions of RANKL can protect the mineralized carti- The RANKL/RANK pathway is becoming an lage from osteoclast-mediated bone loss, but important target for the treatment of patho- nonmineralized cartilage is not directly affected logical bone loss. This was first tested with a by osteoclasts [6]. In some studies, RANKL single subcutaneous injection of OPG fused to inhibition has been demonstrated to protect the human IgG constant region in women with against proteoglycan loss and hence, cartilage postmenopausal osteoporosis [140]. This single damage, whereas in other studies, no protec- injection rapidly and profoundly reduced bone tion was observed. Romas and colleagues found turnover for a substantial period of time, indi- that while administration of OPG in a mouse cating that the inhibition of RANKL could be CIA model markedly reduced bone erosions, the an effective treatment for bone loss diseases effects on cartilage were modest in the context of such as osteoporosis [140]. However some risks severe synovitis [146]. On the other hand, Kong were identified and associated with chronic use and colleagues demonstrated that OPG largely of OPG, including the generation of anti-OPG preserved articular cartilage except at sites where antibodies, which could lead to cross reactivity cartilage was in direct contact with the pannus with the endogenous OPG and would neutral- [47]. It is suggested that OPG has chrondopro- ize the OPG activity [141]. There was also the tective effects in the early disease phase, but possibility of OPG binding to TNF-related once synovitis is established, this is no longer apoptosis-inducing ligand (TRAIL), which the case [146]. The effects of RANKL inhibition could affect the normal defense mechanisms on cartilage could also be indirect, as protection against tumors [141]. More recently, denosumab from subchondral bone loss could also protect (formally known as AMG 162), a mono­clonal the cartilage attached to that bone from being antibody to RANKL, has been studied in destroyed [6,147]. This supports the concept that clinical trials of postmenopausal women with there is a relationship between the cartilage low BMDs [142,143]. A single dose was found to and bone. RANK and OPG are expressed by

future science group www.futuremedicine.com 573 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

chondrocytes and may be highly expressed in rats treated with ML120B also demonstrated a RA; however, RANKL was not able to activate decrease in NF-kB-regulated gene expression human articular chondrocytes [148]. To date, of the inflammatory cytokines TNF-a, IL-1b the results of RANKL inhibitors such as denso- and iNOS. There was also a downregulation sumab indicate that inhibition of RANKL leads of NF-kB activity in joints of rats treated with to increased BMD and decreased bone resorp- this inhibitor [92]. Like many of these antiresorp- tion. However, the use of RANK/RANKL tive drugs, IKK-b inhibition may also inhibit inhibitors may have unwanted side effects, as immune responses. However, as yet, these side they will inhibit both inflammatory and physio- effects have not been reported with regard to the logical bone resorption. It is still unclear whether studies in models of RA. The beneficial effects RANKL inhibition has any positive effects on of these inhibitors on inflammation, bone and suppressing cartilage destruction. cartilage loss could indicate that these drugs may be particularly useful for the treatment of RA. Cathepsin K inhibitors Cathepsin K is an enzyme found predominantly „„Other approaches to treat RA in osteoclasts and plays a role in bone resorption Vitamin D through its activity in the resorption lacunae. Vitamin D, which is essential for the formation Since it is one of the few extracellular proteolytic of healthy bone, is obtained from dietary sources enzymes capable of degrading native fibrillar col- and the action of sunlight. There are conflicting lagen, it is thought to play an important role in results with regard to whether vitamin D levels joint destruction [149]. Cathepsin K is known to be have any effect on the risk of developing RA highly expressed by synovial fibroblasts and mac- and whether treatment with vitamin D can slow rophages in RA joints [150]. Since Cathepsin K is disease progression. One study found that vita- also thought to play an important role in osteo- min D intake was inversely associated with RA clast bone resorption, it has become a target for onset [155], whereas another study demonstrated novel antiresorptive drugs. Cathepsin K inhibi- that vitamin D deficiency does not increase the tors that are highly potent, selective and orally risk of RA [156]. Vitamin D may be a poten- applicable have been developed and some have tial treatment for RA as it can exert immuno­ been demonstrated to inhibit bone resorption modulatory effects. The oral administration of [151–153]. Selective Cathepsin K inhibitors have high-dose vitamin D has been found to be safe passed preclinical trials and are now in clinical and can rapidly correct vitamin D deficiency trials for the treatment of osteoporosis [154]. followed by regular lower doses to maintain adequate levels [157]. In selected patients with IKK-b inhibition RA who are at high risk of vitamin D defi- The activation of IKK-b occurs downstream ciency, vitamin D replacement may reduce RA of RANK/RANKL ligation and IKK inhibi- disease severity [157]. However, it is still not clear tors may offer a selective and effective approach whether vitamin D is effective for treating RA to inhibit NF-kB activation. IKK-b is essen- in all patients. tial for NF-kB activation in response to pro­ inflammatory cytokines[58] . IKK-b phosphory- Statins lates the inhibitor subunit IKB-a, which leads Statins, also known as 3-hydroxy-3-methyl­ to its degradation and subsequent activation glutaryl-CopA (HMG-CoA) reductase inhibi- of NF-kB. It was demonstrated that in vitro, tors, are generally used in controlling lipid bone marrow cells that were deficient in IKK-b metabolism; however, recent studies have dem- did not form osteoclasts when stimulated with onstrated a broader role for these drugs. One of RANKL, confirming that inhibitors of IKK-b the most commonly used statins in humans is could reduce bone degradation in RA [58]. Oral simvastatin, and recent studies in animals and administration of ML120B, a selective potent humans have shown promising results in the inhibitor of IKK-b, was demonstrated to inhibit treatment of RA. In a Th-1 driven CIA model of paw swelling and offered significant protection murine inflammatory arthritis, simvastatin dose against arthritis-induced cartilage and bone loss, dependently suppressed the incidence and sever- using an adjuvant arthritis model. The general ity of CIA development and it also significantly anti-inflammatory and chondro-protective inhibited the progression of RA within 3 days effects may indicate that IKK-b activation is of treatment [158]. Simvastatin was also effective also involved with these aspects of the disease. at blocking T-cell/macrophage interactions [159]. In support of this, ankle joints obtained from More recently, simvastatin was studied in a trial

574 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

involving 24 patients with RA and was found to ERK and JNK) and tyrosine kinases (JAK-3 and improve the clinical symptoms of RA patients Syk) [125]. JAK-3 is a cytoplasmic tyrosine kinase in a short period of time [159]. Treatment with that is activated by cytokines and involved in simvastatin significantly decreased the levels of signaling in cells such as B cells, T cells and C-reactive protein and rheumatoid factor and it natural killer cells [162]. JAK-3 is known to play also reduced Th1:Th2 and CD4:CD8 ratios. No a significant role in the pathogenesis of RA[133] . adverse effects of simvastatin were observed in A JAK-3 inhibitor (CP-690,550) is in Phase II this investigation [159]. In a recent study, rats were clinical trials in patients and so far, remission given an intraperitoneal injection of streptococcal has been achieved in a third of the individu- bacteria to induce bone loss. Simvastatin adminis- als [125]. SyK is an intracellular protein kinase tered at 20 mg/kg/day was found to reduce both stimulated by activation of the Fcg receptors and early and late joint inflammation, reduce 60% B-cell receptors. A SyK inhibitor (fostamtinib of monocyte/macrophage influx and also to sup- disodium) has been demonstrated to result in press periarticular bone destruction occurring remission in over half the patients included in late in the course of disease. It also inhibited the a RA clinical trial [125]. While the recent data increase in periarticular osteoclasts in the arthritic regarding novel kinase inhibitors is encourag- joints [160]. These studies indicate that simvastatin ing, their direct effects on joint destruction are may have therapeutic benefits for RA treatment, yet to be determined. both through the suppression of inflammation and reduced periarticular bone destruction [159]. B-cell targets Depletion of B cells is a new treatment strategy Physical activity for RA. Rituximab is a chimeric monoclonal Other options, aside from drugs, are available antibody targeting the CD20 molecule expressed to help reduce the progression of RA. Physical on developing B cells. Depletion of B cells has exercise is an option chosen by many patients. been demonstrated to inhibit inflammation by However, this only tends to have an effect on the abolition of B cells as antigen-presenting generalized bone loss associated with RA. A sed- cells, blocking the synthesis of proinflammatory entary life style is a risk factor for osteopenia in cytokines such as TNF-a, leading to impaired RA with a relative risk of 1.6 [161], and moderate formation of immune complexes [7]. Although physical activity was found to reduce the risk of these inhibitors target the inflammation, B cells osteopenia by 50% [161]. also express RANKL and may thus play a role in the differentiation of osteoclasts [163]. Therefore, „„Novel approaches depletion of B cells could contribute to reducing As our understanding of the mechanisms the amount of osteoclast bone resorption. involved in the regulation of bone metabo- lism and the RA disease process improves, new Wnt pathway approaches for the treatment of bone loss in RA The Wnt pathway is a family of glycoproteins are being developed. that are involved in the regulation of multiple cellular activities including bone formation and Kinase inhibitors remodeling. Dickkopf-1 (DKK-1) is an inhibi- These small intracellular molecules are new tar- tory molecule that regulates the Wnt pathway. gets for therapeutics to treat RA. These inhibi- DKK-1 is induced by TNF-a and increased tors are small molecular weight peptides that levels of DKK-1 have been demonstrated to be are less expensive than many current therapies associated with bone resorption and decreased and appear to have the same efficacy as other levels of bone formation [164]. This Wnt pathway biological agents with reduced side effects [125]. may form an important link between inflamma- Kinases are an important class of intracellular tion and bone metabolism and may be an effec- molecules involved in signal transduction in tive target for RA therapeutic agents. Several inflammation. Binding of extracellular factors of the Wnt family members have been demon- to surface receptors results in the activation of strated to modulate the inflammatory response cytoplasmic kinases that modulate the activ- in RA. Wnt7b was found to be expressed in ity of transcription factors, thus regulating the the synovium from patients with RA [165]. In expression of genes [125]. There are now small addition, immunohistochemistry revealed that molecular weight orally available inhibitors Wnt7b was also present in the articular cartilage, that target these enzymes. Two different types bone and synovium of RA [165]. The cytokines of kinases that can be targeted – MAPKs (p38, TNF-a, IL-1b and IL-6 were also increased in

future science group www.futuremedicine.com 575 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

RA synovium and Wnt-7b-transfected normal in the hyperacetylation of histones, and possibly synovial cells [165]. It is believed that the Wnt other proteins, that modify the expression of pathways could potentially promote synovial numerous genes [174]. This can result in upregu- hyperplasia, inflammation, pannus formation lation of cell-cycle inhibitors, downregulation of and bone/cartilage destruction [166]. Therefore, immune stimulators, repression of inflammatory the Wnt pathway provides a novel target for cytokines [175] and suppression of osteoclast bone treating RA by changing the balance between resorption [176]. In particular, several HDACs bone formation and bone resorption. are known to regulate the transcription of genes such as NF-kB that are crucial for osteoclast

PAR2 agonists/antagonists ­development and inflammation[172,173] . There has recently been increasing interest There are 11 human isoforms of the zinc- in the role that protease activated receptors containing HDACs (HDAC 1–11) [177] and + (PARs), particularly PAR2, play in mediating a further seven NAD -dependent enzymes chronic inflammation. The PARs are a fam- (SIRT 1–7) [178]. Class I HDACs include HDAC ily of G-protein-coupled seven transmembrane 1, 2, 3 and 8 enzymes that are found mainly in domain receptors that are activated by proteo- the nucleus. Class II HDACs include HDAC 4, lytic cleavage of the extracellular domain. Four 5, 6, 7, 9 and 10 and are found in both the nucleus PARs have similar mechanisms of action but and cytoplasm [179]. Most HDACi research is have different biological functions and tissue currently focusing on developing inhibitors that

distributions [167]. PAR2 is found throughout selectively inhibit different HDAC isoforms, the body, especially in the epithelium, endothe- since they may have quite different anti- or lium, fibroblasts, osteoblasts, neutrophils, myo- pro-inflammatory properties. It is still not clear cytes, neurons and astrocytes [168]. Expression in which HDAC class and which HDACs within osteoblasts has been demonstrated both in vitro these classes play a role in suppressing osteoclasts

and in vivo [169]. PAR2 is expressed by a small bone resorption. number of proteases including mast-cell tryptase, HDACi have previously been investigated in factor Xa and trypsin [170]. models of RA and have been demonstrated to There is conflicting evidence as to whether reduce bone destruction in both AIA and CIA

PAR2 activation or inhibition results in reduced models [83,176,180]. Two HDACi, suber­oxylanilide osteoclastogenesis. Recent data has indicated hydroxamic acid and MS-275 decreased bone

that PAR2 agonists can regulate bone loss and resorption; however, suberoxylanilide hydroxamic inhibit osteoclast differentiation by reducing the acid could not prevent the onset of arthritis while formation of RANKL [170]. In an adjuvant mono- MS-275 displayed dramatic antirheumatic activi- arthritis model, a reduction in inflammation ties. High doses prevented bone erosion and also

and paw swelling was found in PAR2-deficient delayed the onset of arthritis [180]. Two HDACi, mice compared with wild-type mice. However, phenylbutyrate and trichostatin A, were found to

activation of PAR2 in the knee joint has demon- reduce joint swelling, decrease subintimal mono- strated proinflammatory effects[171] . Injection of nuclear cell infiltration, inhibit synovial hyperpla- a native peptide, SLIGRL0NH2, or a synthetic sia and suppress pannus formation and cartilage

PAR2 agonist, ASKH95, was also demonstrated and bone destruction in a model of adjuvant to result in joint swelling and hyperemia [171], arthritis [175]. These studies clearly demonstrate

suggesting that PAR2 antagonists could therefore the potential of HDACi to suppress the bone play a role in inhibiting the inflammation and resorption in RA. HDACi provide an appropriate subsequent joint destruction in RA. Whether therapeutic agent for treating pathological bone

PAR2 agonists or antagonists can be effective in loss in RA and could potentially be administered RA treatment remains unclear and further stud- as combination therapy with a DMARD, such as

ies are necessary. PAR2 does remain a potential MTX, in which their mechanism of action is to novel target to treat RA. reduce inflammation.

Histone deacetylase inhibitors Conclusion Enzymes known as histone deacetylases (HDACs) This review has discussed the major mecha- are emerging targets for therapeutic intervention nisms involved in the osteoclast-mediated bone in a variety of diseases [172,173]. These enzymes loss in RA and has given an overview of current remove acetyl groups from the histone proteins and emerging treatments of joint destruction in that condense the chromatin and regulate gene RA. There is a strong case that antiresorptive repression [174]. HDAC inhibitors (HDACi) result treatments are needed in the treatment of RA.

576 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

First, antiresorptive therapy in combination with anti-inflammatory therapies currently used to anti-inflammatory therapies can better prevent treat RA. Protection of the bone as early as pos- structural joint damage than treatment with sible in RA will allow rheumatologists time to anti-inflammatory therapy alone. Second, early establish the most effective long-term treatments antiresorptive treatments administered at initial for individuals with RA, while preserving the diagnosis are likely to prevent the joint damage joint structure and function. that may occur before slower acting therapies demonstrate their effects or before an effective Financial & competing interests disclosure anti-inflammatory therapy can be identified for The authors have no relevant affiliations or financial involve- an individual. ment with any organization or entity with a financial interest in or financial conflict with the subject matter or materials Future perspective discussed in the manuscript. This includes employment, con- Effective targeted treatments for bone resorption sultancies, honoraria, stock ownership or options, expert by osteoclasts are, and will, become available in ­testimony, grants or patents received or pending, or royalties. the near future. These are likely to become valu- No writing assistance was utilized in the production of able additions for use in conjunction with the this manuscript.

Executive summary ƒƒ Rheumatoid arthritis (RA) is a significant disease that esultsr in the destruction of bone and cartilage. ƒƒ Bone loss also occurs around the joint and systemically. Mechanisms of disease ƒƒ Enhanced osteoclast bone resorption is a hallmark feature of RA. ƒƒ There is a relationship between the immune system and bone metabolism – osteoimmunology understanding is vital for developing ways to treat RA. ƒƒ Imbalance between RANKL and OPG plays a major role in the destruction of RA, and enhanced signaling through the Receptor activator of NF-kB ligand (RANKL)/ Receptor activator of NF-kB (RANK) pathway is seen in RA. Animal models of disease ƒƒ Animal models of RA have been useful to gain an understanding of bone loss in RA. ƒƒ Live animal micro-computed tomography is a new technique used to give a time-dependent ana­lysis of bone volume loss. Therapeutic approaches to treat RA ƒƒ Conventional treatments, such as DMARDs, inhibit the inflammation but have little direct effect on preserving joint structure. ƒƒ Antiresorptive therapies such as bisphosphonates, RANKL inhibitors and other emerging therapies, inhibit enhanced osteoclast bone resorption in RA. Conclusion ƒƒ Antiresorptive therapies in combination with modern anti-inflammatory treatments may better prevent structural joint destruction and disease progression in RA. ƒƒ Protection of the bone upon diagnosis of RA will allow rheumatologists time to establish the most effective long-term treatment for individuals with RA while preserving the joint.

5 Romas E, Gillespie MT: Inflammation-induced 8 Goldring SR: Pathogenesis of bone and Bibliography bone loss: can it be prevented? Rheum. Dis. cartilage destruction in rheumatoid arthritis. Papers of special note have been highlighted as: Clin. North Am. 32, 759–773 (2006). Rheumatology (Oxford) 42, II11–II16 (2003). n of interest nn Discusses the process of inflammation- 9 Goldring MB, Suen LF, Yamin R, Lai WF: nn of considerable interest induced bone loss and how it can be targeted Regulation of collagen gene expression by 1 Findlay DM, Haynes DR: Mechanisms of bone to treat rheumatoid arthritis (RA). prostaglandins and interleukin-1b in cultured loss in rheumatoid arthritis. Mod. Rheumatol. chondrocytes and fibroblasts. Am. J. Ther. 3, 6 Schett G: Erosive arthritis. Arthritis Res. Ther. 15, 232–240 (2005). 9–16 (1996). 9, S2 (2007). 2 Jacobson JA, Girish G, Jiang Y, Resnick D: 10 Stephenson ML, Goldring MB, Birkhead JR, n Discusses the close relationship between the Radiographic evaluation of arthritis: Krane SM, Rahmsdorf HJ, Angel P: inflammation and bone degradation in RA, inflammatory conditions. Radiology 248, Stimulation of procollagenase synthesis 378–389 (2008). and how blocking osteoclast formation is a key parallels increases in cellular procollagenase strategy for treatment. 3 Langs G, Peloschek P, Bischof H, Kainberger F: mRNA in human articular chondrocytes Automatic quantification of joint space 7 Schett G, Stach C, Zwerina J, Voll R, exposed to recombinant interleukin 1 b or narrowing and erosions in rheumatoid arthritis. Manger B: How antirheumatic drugs protect phorbol ester. Biochem. Biophys. Res. Commun. IEEE Trans. Med. Imaging 28, 151–164 (2009). joints from damage in rheumatoid arthritis. 144, 583–590 (1987). Arthritis Rheum. 58, 2936–2948 (2008). 4 Walsh NC, Crotti TN, Goldring SR, 11 Deodhar AA, Woolf AD: Bone mass Gravallese EM: Rheumatic diseases: the effects nn Focuses on the mechanisms of bone and measurement and bone metabolism in of inflammation on bone. Immunol. Rev. 208, cartilage destruction in RA and discusses the rheumatoid arthritis: 228–251 (2005). different antirheumatic drugs that can a review. Br. J. Rheumatol. 35, 309–322 (1996). prevent joint damage.

future science group www.futuremedicine.com 577 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

12 Guler-Yuksel M, Allaart CF, nn Receptor activator of NF-kB ligand (RANKL) 37 Lam J, Takeshita S, Barker JE, Kanagawa O, Goekoop-Ruiterman YP et al.: Changes in is an essential factor for osteoclast formation Ross FP, Teitelbaum SL: TNF-a induces hand and generalised bone mineral density in and osteoprotegerin (OPG) may prevent bone osteoclastogenesis by direct stimulation of patients with recent-onset rheumatoid erosion in RA. A significant correlation macrophages exposed to permissive levels of arthritis. Ann. Rheum. Dis. 68, 330–336 between the ratio of RANKL mRNA: OPG RANK ligand. J. Clin. Invest. 106, 1481–1488 (2009). mRNA and the number of pits formed by (2000). 13 Guler-Yuksel M, Bijsterbosch J, osteoclasts bone resorption was determined in 38 Hofbauer LC, Heufelder AE: Role of receptor Goekoop-Ruiterman YP et al.: Changes in this study. activator of nuclear factor-kB ligand and bone mineral density in patients with recent osteoprotegerin in bone cell biology. J. Mol. 25 Suzuki Y, Tsutsumi Y, Nakagawa M et al.: onset, active rheumatoid arthritis. Ann. Med. 79, 243–253 (2001). Osteoclast-like cells in an in vitro model of bone Rheum. Dis. 67, 823–828 (2008). destruction by rheumatoid synovium. 39 Caetano-Lopes J, Canhão H, Fonseca JE: 14 Guler-Yuksel M, Allaart CF, Rheumatology (Oxford) 40, 673–682 (2001). Osteoimmunology – the hidden immune Goekoop-Ruiterman YP et al.: Changes in regulation of bone. Autoimmun. Rev. 8, 26 Toritsuka Y, Nakamura N, Lee SB et al.: hand and generalised bone mineral density in 250–255 (2009). Osteoclastogenesis in iliac bone marrow of patients with recent-onset rheumatoid arthritis. patients with rheumatoid arthritis. 40 Crotti T, Smith MD, Hirsch R et al.: Receptor Ann. Rheum. Dis. 68, 330–336 (2009). J. Rheumatol. 24, 1690–1696 (1997). activator NF kB ligand (RANKL) and 15 Guler-Yuksel M, Bijsterbosch J, osteoprotegerin (OPG) protein expression in 27 Kuratani T, Nagata K, Kukita T, Hotokebuchi Goekoop-Ruiterman YP et al.: Changes in periodontitis. J. Periodont. Res. 38, 380–387 T, Nakasima A, Iijima T: Induction of bone mineral density in patients with recent (2003). abundant osteoclast-like multinucleated giant onset, active rheumatoid arthritis. Ann. cells in adjuvant arthritic rats with 41 Crotti TN, Ahern MJ, Lange K et al.: Rheum. Dis. 67, 823–828 (2008). accompanying disordered high bone turnover. Variability of RANKL and osteoprotegerin 16 Goldring SR: Periarticular bone changes in Histol. Histopathol. 13, 751–759 (1998). staining in synovial tissue from patients with rheumatoid arthritis: pathophysiological active rheumatoid arthritis: quantification 28 Suzuki Y, Nishikaku F, Nakatuka M, Koga Y: implications and clinical utility. Ann. Rheum. using color video image ana­lysis. J. Rheumatol. Osteoclast-like cells in murine collagen induced Dis. 68, 297–299 (2009). 30, 2319–2324 (2003). arthritis. J. Rheumatol. 25, 1154–1160 (1998). 17 van Staa TP, Geusens P, Bijlsma JW, Leufkens 42 Crotti TN, Smith MD, Weedon H et al.: 29 Takayanagi H, Sato K, Takaoka A, HG, Cooper C: Clinical assessment of the Receptor activator NF-kB ligand (RANKL) Taniguchi T: Interplay between interferon and long-term risk of fracture in patients with expression in synovial tissue from patients with other cytokine systems in bone metabolism. rheumatoid arthritis. Arthritis Rheum. 54, rheumatoid arthritis, spondyloarthropathy, Immunol. Rev. 208, 181–193 (2005). 3104–3112 (2006). osteoarthritis, and from normal patients: 30 Chambers TJ: Regulation of the differentiation 18 Boyle WJ, Simonet WS, Lacey DL: Osteoclast semiquantitative and quantitative ana­lysis. Ann. and function of osteoclasts. J. Pathol. 192, 4–13 differentiation and activation. Nature 423, Rheum. Dis. 61, 1047–1054 (2002). (2000). 337–342 (2003). 43 Pettit AR, Walsh NC, Manning C, 31 Kodama H, Nose M, Niida S, Yamasaki A: 19 Lerner UH: Osteoclast formation and Goldring SR, Gravallese EM: RANKL protein Essential role of macrophage colony-stimulating resorption. Matrix Biol. 19, 107–120 (2000). is expressed at the pannus-bone interface at sites factor in the osteoclast differentiation supported of articular bone erosion in rheumatoid 20 Rubin J, Greenfield E: Osteoclast: origin and by stromal cells. J. Exp. Med. 173, 1291–1294 arthritis. Rheumatology (Oxford) 45, 1068–1076 differentiation. In: Topics in Bone Biology – (1991). (2006). Bone Resorption. Bronner F, Farach- 32 Sarma U, Flanagan AM: Macrophage Carson MC, Rubin R (Eds). Springer–Verlag, nn RANKL protein was found to be expressed at colony-stimulating factor induces substantial London, UK, 2 (2005). sites of osteoclast-mediated bone resorption osteoclast generation and bone resorption in and OPG expression was limited at sites of 21 Pettit AR, Ji H, von Stechow D et al.: human bone marrow cultures. Blood bone erosion, especially in areas associated TRANCE/RANKL knockout mice are 88, 2531–2540 (1996). protected from bone erosion in a serum with RANKL expression. 33 Teitelbaum SL: Osteoclasts: what do they do transfer model of arthritis. Am. J. Pathol. 159, 44 Haynes DR: Emerging and future therapies for and how do they do it? Am. J. Pathol. 170, 1689–1699 (2001). the treatment of bone loss associated with 427–435 (2007). 22 Redlich K, Hayer S, Ricci R et al.: chronic inflammation. Inflammopharmacology 34 Lacey DL, Timms E, Tan HL et al.: Osteoclasts are essential for TNF-a-mediated 14, 193–197 (2006). Osteoprotegerin ligand is a cytokine that joint destruction. J. Clin. Invest. 110, 45 Horwood NJ, Kartsogiannis V, Quinn JM, regulates osteoclast differentiation and 1419–1427 (2002). Romas E, Martin TJ, Gillespie MT: Activated activation. Cell 93, 165–176 (1998). 23 Fujikawa Y, Sabokbar A, Neale S, T lymphocytes support osteoclast formation 35 Matsuzaki K, Udagawa N, Takahashi N et al.: Athanasou NA: Human osteoclast formation in vitro. Biochem. Biophys. Res. Commun. 265, Osteoclast differentiation factor (ODF) induces and bone resorption by monocytes and 144–150 (1999). osteoclast-like cell formation in human synovial macrophages in rheumatoid arthritis. n peripheral blood mononuclear cell cultures. Using murine spleen cells, peripheral blood Ann. Rheum. Dis. 55, 816–822 (1996). Biochem. Biophys. Res. Commun. 246, 199–204 mononuclear-derived T cells were 24 Haynes DR, Crotti TN, Loric M, Bain GI, (1998). demonstrated to promote osteoclast formation Atkins GJ, Findlay DM: Osteoprotegerin and and activation in vitro. 36 Kobayashi K, Takahashi N, Jimi E et al.: receptor activator of nuclear factor kB ligand Tumor necrosis factor a stimulates osteoclast 46 Kong YY, Feige U, Sarosi I et al.: Activated (RANKL) regulate osteoclast formation by differentiation by a mechanism independent of T cells regulate bone loss and joint destruction cells in the human rheumatoid arthritic joint. the ODF/RANKL-RANK interaction. J. Exp. in adjuvant arthritis through osteoprotegerin Rheumatology (Oxford) 40, 623–630 (2001). Med. 191, 275–286 (2000). ligand. Nature 402, 304–309 (1999).

578 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

47 Holding CA, Findlay DM, Stamenkov R 61 Johnson RS, Spiegelman BM, Papaioannou V: 74 Suzuki Y, Nakagawa M, Masuda C et al.: et al.: The correlation of RANK, RANKL and Pleiotropic effects of a null mutation in the Short-term low dose methotrexate ameliorates TNFa expression with bone loss volume and c-fos proto-oncogene. Cell 71, 577–586 abnormal bone metabolism and bone loss in polyethylene wear debris around hip implants. (1992). adjuvant induced arthritis. J. Rheumatol. 24, Biomaterials 27, 5212–5219 (2006). 62 Grigoriadis AE, Wang ZQ, Cecchini MG 1890–1895 (1997). 48 Bucay N, Sarosi I, Dunstan CR et al.: et al.: c-fos: a key regulator of osteoclast- 75 Magari K, Miyata S, Nishigaki F, Ohkubo Y, Osteoprotegerin-deficient mice develop early macrophage lineage determination and bone Mutoh S, Goto T: Differential effects of onset osteoporosis and arterial calcification. remodeling. Science 266, 443–448 (1994). FK506 and methotrexate on inflammatory Genes Dev. 12, 1260–1268 (1998). 63 Ikeda F, Nishimura R, Matsubara T et al.: cytokine levels in rat adjuvant-induced 49 Mizuno A, Kanno T, Hoshi M et al.: Critical roles of c-jun signaling in regulation of arthritis. J. Rheumatol. 30, 2193–2200 Transgenic mice overexpressing soluble NFAT family and RANKL-regulated (2003). osteoclast differentiation factor (sODF) osteoclast differentiation. J. Clin. Invest. 114, 76 Haynes DR, Hutchens MJ, Whitehouse MW, exhibit severe osteoporosis. J. Bone Miner. 475–484 (2004). Vernon-Roberts B: A comparison of the Metab. 20, 337–344 (2002). 64 Asagiri M, Takayanagi H: The molecular disease-modifying and cytokine-regulating 50 Redlich K, Hayer S, Maier A et al.: understanding of osteoclast differentiation. activities of tenidap, piroxicam and Tumor necrosis factor a-mediated joint Bone 40, 251–264 (2007). cyclosporin-A using the adjuvant-induced model of arthritis in rats. destruction is inhibited by targeting 65 Matsumoto M, Kogawa M, Wada S et al.: Inflammopharmacology 6, 193–202 (1998). osteoclasts with osteoprotegerin. Arthritis Essential role of p38 mitogen-activated protein Rheum. 46, 785–792 (2002). kinase in cathepsin K gene expression during 77 Haynes DR, Gadd SJ, Whitehouse MW, 51 Haynes DR, Barg E, Crotti TN et al.: osteoclastogenesis through association of Mayrhofer G, Vernon-Roberts B: Complete Osteoprotegerin expression in synovial tissue NFATc1 and PU.1. J. Biol. Chem. 279, prevention of the clinical expression of from patients with rheumatoid arthritis, 45969–45979 (2004). adjuvant-induced arthritis in rats by cyclosporin-A and lobenzarit: the regulation spondyloarthropathies and osteoarthritis and 66 Matsuo K, Irie N: Osteoclast–osteoblast of lymph node cell populations and cytokine normal controls. Rheumatology (Oxford) communication. Arch. Biochem. Biophys. 473, production. Inflamm. Res. 45, 159–165 (1996). 42, 123–134 (2003). 201–209 (2008). 78 Ghia M, Mattioli F, Novelli F, Minganti V: 52 Takayanagi H: Mechanistic insight into 67 Sharma SM, Bronisz A, Hu R et al.: Inhibitory properties of triethylphosphine osteoclast differentiation in osteoimmunology. MITF and PU.1 recruit p38 MAPK and goldlupinylsulfide in adjuvant-induced J. Mol. Med. 83, 170–179 (2005). NFATc1 to target genes during osteoclast arthritis in rats. Farmaco 50, 601–604 (1995). 53 Theoleyre S, Wittrant Y, Tat SK, Fortun Y, differentiation. J. Biol. Chem. 282, Redini F, Heymann D: The molecular triad 15921–15929 (2007). 79 Feige U, Hu YL, Gasser J, Campagnuolo G, Munyakazi L, Bolon B: Anti-interleukin-1 OPG/RANK/RANKL: involvement in the 68 Takayanagi H, Kim S, Koga T et al.: and anti-tumor necrosis factor-a orchestration of pathophysiological bone Induction and activation of the transcription synergistically inhibit adjuvant arthritis in remodeling. Cytokine Growth Factor Rev. 15, factor NFATc1 (NFAT2) integrate RANKL Lewis rats. Cell. Mol. Life Sci. 57, 1457–1470 457–475 (2004). signaling in terminal differentiation of (2000). 54 Naito A, Azuma S, Tanaka S et al.: Severe osteoclasts. Dev. Cell 3, 889–901 (2002). 80 Bendele AM, Chlipala ES, Scherrer J et al.: osteopetrosis, defective interleukin-1 69 Takayanagi H, Kim S, Koga T et al.: Combination benefit of treatment with the signalling and lymph node organogenesis in Induction and activation of the transcription cytokine inhibitors interleukin-1 receptor TRAF6-deficient mice. Genes Cells 4, factor NFATc1 (NFAT2) integrate RANKL antagonist and PEGylated soluble tumor 353–362 (1999). signaling in terminal differentiation of necrosis factor receptor type I in animal 55 Tak PP, Gerlag DM, Aupperle KR et al.: osteoclasts. Dev. Cell 3, 889–901 (2002). models of rheumatoid arthritis. Arthritis Inhibitor of nuclear factor kB kinase b is a key 70 Hegen M, Keith JC Jr, Collins M, Rheum. 43, 2648–2659 (2000). regulator of synovial inflammation. Arthritis Nickerson-Nutter CL: Utility of animal 81 Bendele AM, McComb J, Gould T et al.: Rheum. 44, 1897–1907 (2001). models for identification of potential Combination benefit of PEGylated soluble 56 Iotsova V, Caamano J, Loy J, Yang Y, therapeutics for rheumatoid arthritis. Ann. tumor necrosis factor receptor type I (PEG Lewin A, Bravo R: Osteopetrosis in mice Rheum. Dis. 67, 1505–1515 (2008). sTNF-RI) and dexamethasone or lacking NF-kB1 and NF-kB2. Nat. Med. 3, 71 Walz DT, DiMartino MJ, Misher A: indomethacin in adjuvant arthritic rats. 1285–1289 (1997). Adjuvant-induced arthritis in rats. II. Drug Inflamm. Res. 48, 453–460 (1999). 57 Soysa NS, Alles N: NF-kB functions in effects on physiologic, biochemical and 82 Khachigian LM: Collagen antibody-induced osteoclasts. Biochem. Biophys. Res. Commun. immunologic parameters. J. Pharmacol. Exp. arthritis. Nat. Protoc. 1, 2512–2516 (2006). 378, 1–5 (2009). Ther. 178, 223–231 (1971). 83 Nasu Y, Nishida K, Miyazawa S et al.: 58 Ruocco MG, Karin M: IKKb as a target for 72 Sakuma S, Nishigaki F, Magari K et al.: Trichostatin A, a histone deacetylase inhibitor, treatment of inflammation induced bone loss. FK506 is superior to methotrexate in suppresses synovial inflammation and Ann. Rheum. Dis. 64, IV81–IV85 (2005). therapeutic effects on advanced stage of rat subsequent cartilage destruction in a collagen adjuvant-induced arthritis. Inflamm. Res. 50, 59 Asagiri M, Takayanagi H: The molecular antibody-induced arthritis mouse model. 509–514 (2001). understanding of osteoclast differentiation. Osteoarthr. Cartil. 16, 723–732 (2008). Bone 40, 251–264 (2007). 73 Morgan SL, Chen DT, Carlee J, Baggott JE: 84 Oestergaard S, Rasmussen KE, Doyle N et al.: Effect of methotrexate therapy on bone 60 Takayanagi H: Osteoimmunology: shared Evaluation of cartilage and bone degradation mineral density and body composition in rat mechanisms and crosstalk between the in a murine collagen antibody-induced adjuvant arthritis. J. Rheumatol. 31, immune and bone systems. Nat. Rev. arthritis model. Scand. J. Immunol. 67, 1693–1697 (2004). Immunol. 7, 292–304 (2007). 304–312 (2008).

future science group www.futuremedicine.com 579 Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

85 Mitamura M, Nakano N, Yonekawa T et al.: 97 Take I, Kobayashi Y, Yamamoto Y et al.: in patients with rheumatoid arthritis treated T cells are involved in the development of Prostaglandin E2 strongly inhibits human with background methotrexate. Ann. Rheum. arthritis induced by anti-type II collagen osteoclast formation. Endocrinology Dis. 63, 1062–1068 (2004). antibody. Int. Immunopharmacol. 7, 1360–1368 146, 5204–5214 (2005). 109 Kremer J, Genovese M, Cannon GW et al.: (2007). 98 Li M, Ke HZ, Qi H et al.: A novel, non- Combination leflunomide and methotrexate 86 Kannan K, Ortmann RA, Kimpel D: prostanoid EP2 receptor-selective prostaglandin (MTX) therapy for patients with active Animal models of rheumatoid arthritis and E2 agonist stimulates local bone formation and rheumatoid arthritis failing MTX their relevance to human disease. enhances fracture healing. J. Bone Miner. Res. monotherapy: open-label extension of a Pathophysiology 12, 167–181 (2005). 18, 2033–2042 (2003). randomized, double-blind, placebo controlled 87 Li P, Schwarz EM: The TNF-a transgenic 99 Paralkar VM, Borovecki F, Ke HZ et al.: trial. J. Rheumatol. 31, 1521–1531 (2004). mouse model of inflammatory arthritis. An EP2 receptor-selective prostaglandin E2 110 Emery P, Fleischmann R, Springer Semin. Immunopathol. 25, 19–33 agonist induces bone healing. Proc. Natl Acad. Filipowicz-Sosnowska A et al.: The efficacy (2003). Sci. USA 100, 6736–6740 (2003). and safety of rituximab in patients with active 88 Ahlqvist E, Hultqvist M, Holmdahl R: 100 Idris AI, Ralston SH, van’t Hof RJ: rheumatoid arthritis despite methotrexate The value of animal models in predicting The nitrosylated flurbiprofen derivative treatment: results of a Phase IIB randomized, genetic susceptibility to complex diseases such HCT1026 inhibits cytokine-induced signalling double-blind, placebo-controlled, dose- as rheumatoid arthritis. Arthritis Res. Ther. 11, through a novel mechanism of action. Eur. ranging trial. Arthritis Rheum. 54, 1390–1400 226 (2009). J. Pharmacol. 602, 215–222 (2009). (2006). 89 Noguchi M, Kimoto A, Sasamata M, 101 De Vries F, Bracke M, Leufkens HG, Lammers 111 Klinkhoff A: Biological agents for rheumatoid Miyata K: Micro-CT imaging ana­lysis for the JW, Cooper C, van Staa TP: Fracture risk with arthritis: targeting both physical function and effect of celecoxib, a cyclooxygenase-2 inhibitor, intermittent high-dose oral glucocorticoid structural damage. Drugs 64, 1267–1283 on inflammatory bone destruction in adjuvant therapy. Arthritis Rheum. 56, 208–214 (2007). (2004). arthritis rats. J. Bone Miner. Metab. 26, 102 van Staa TP, Leufkens HG, Abenhaim L, 112 Herman S, Kronke G, Schett G: Molecular 461–468 (2008). Zhang B, Cooper C: Use of oral corticosteroids mechanisms of inflammatory bone damage: 90 Barck KH, Lee WP, Diehl LJ et al.: and risk of fractures. June, 2000. J. Bone Miner. emerging targets for therapy. Trends Mol. Quantification of cortical bone loss and repair Res. 20, 1487–1494 (2005) (Discussion 1486). Med. 14, 245–253 (2008). for therapeutic evaluation in collagen-induced 103 Bansback NJ, Regier DA, Ara R et al.: 113 Hoff M, Kvien TK, Kalvesten J, Elden A, arthritis, by micro-computed tomography and An overview of economic evaluations for drugs Haugeberg G: Adalimumab therapy reduces automated image ana­lysis. Arthritis Rheum. 50, used in rheumatoid arthritis: focus on tumour hand bone loss in early rheumatoid arthritis: 3377–3386 (2004). necrosis factor-a antagonists. Drugs 65, Explorative analyses from the PREMIER 91 Schopf L, Savinainen A, Anderson K et al.: 473–496 (2005). study. Ann. Rheum. Dis. 68(7), 1171–1176 (2008). IKKb inhibition protects against bone and 104 Cronstein BN: Low-dose methotrexate: cartilage destruction in a rat model of a mainstay in the treatment of rheumatoid 114 Haugeberg G: Focal and generalized bone loss rheumatoid arthritis. Arthritis Rheum. 54, arthritis. Pharmacol. Rev. 57, 163–172 (2005). in rheumatoid arthritis: separate or similar 3163–3173 (2006). concepts? Nat. Clin. Pract. Rheumatol. 4, 105 Emery P, Breedveld FC, Hall S et al.: 402–403 (2008). 92 Cantley MD, Bartold PM, Marino V et al.: The Comparison of methotrexate monotherapy with use of live-animal micro-computed tomography a combination of methotrexate and etanercept 115 Sambrook P: Tumour necrosis factor blockade to determine the effect of a novel phospholipase in active, early, moderate to severe rheumatoid and the risk of osteoporosis: back to the A2 inhibitor on alveolar bone loss in an in vivo arthritis (COMET): a randomised, double- future. Arthritis Res. Ther. 9, 107 (2007). mouse model of periodontitis. J. Periodont. Res. blind, parallel treatment trial. Lancet 372, 116 Zwerina J, Redlich K, Polzer K et al.: 44, 317–322 (2009). 375–382 (2008). TNF-induced structural joint damage is 93 O’Dell JR: Treating rheumatoid arthritis early: 106 Klareskog L, van der Heijde D, de Jager JP mediated by IL-1. Proc. Natl Acad. Sci. USA a window of opportunity? Arthritis Rheum. 46, et al.: Therapeutic effect of the combination of 104, 11742–11747 (2007). 283–285 (2002). etanercept and methotrexate compared with 117 Bresnihan B, Alvaro-Gracia JM, Cobby M 94 Doan T, Massarotti E: Rheumatoid arthritis: each treatment alone in patients with et al.: Treatment of rheumatoid arthritis with an overview of new and emerging therapies. rheumatoid arthritis: double-blind recombinant human interleukin-1 receptor J. Clin. Pharmacol. 45, 751–762 (2005). randomised controlled trial. Lancet 363, antagonist. Arthritis Rheum. 41, 2196–2204 95 Fujita D, Yamashita N, Iita S, Amano H, 675–681 (2004). (1998). Yamada S, Sakamoto K: Prostaglandin E2 107 Keystone EC, Kavanaugh AF, Sharp JT et al.: 118 Jiang Y, Genant HK, Watt I et al.: induced the differentiation of osteoclasts in Radiographic, clinical, and functional A multicenter, double-blind, dose-ranging, mouse osteoblast-depleted bone marrow cells. outcomes of treatment with adalimumab (a randomized, placebo-controlled study of Prostaglandins Leukot. Essent. Fatty Acids human anti-tumor necrosis factor monoclonal recombinant human interleukin-1 receptor 68, 351–358 (2003). antibody) in patients with active rheumatoid antagonist in patients with rheumatoid 96 Inada M, Matsumoto C, Uematsu S, Akira S, arthritis receiving concomitant methotrexate arthritis: radiologic progression and Miyaura C: Membrane-bound prostaglandin E therapy: a randomized, placebo-controlled, correlation of Genant and Larsen scores. synthase-1-mediated prostaglandin E2 52-week trial. Arthritis Rheum. 50, 1400–1411 Arthritis Rheum. 43, 1001–1009 (2000). production by osteoblast plays a critical role in (2004). 119 Nuki G, Bresnihan B, Bear MB, McCabe D: lipopolysaccharide-induced bone loss associated 108 Cohen SB, Moreland LW, Cush JJ et al.: Long-term safety and maintenance of clinical with inflammation. J. Immunol. 177, A multicentre, double blind, randomised, improvement following treatment with 1879–1885 (2006). placebo controlled trial of anakinra (Kineret), a anakinra (recombinant human interleukin-1 recombinant interleukin 1 receptor antagonist, receptor antagonist) in patients with

580 Int. J. Clin. Rheumatol. (2009) 4(5) future science group Review Cantley, Smith & Haynes Pathogenic bone loss in rheumatoid arthritis: mechanisms & therapeutic approaches Review

rheumatoid arthritis: extension phase of a 135 D’Elia HF, Mattsson LA, Ohlsson C, n Two injections a year of denosumab with randomized, double-blind, placebo-controlled Nordborg E, Carlsten H: Hormone ongoing methotrexate treatment was trial. Arthritis Rheum. 46, 2838–2846 (2002). replacement therapy in rheumatoid arthritis demonstrated to inhibit structural damage 120 Church LD, McDermott MF: Canakinumab, a is associated with lower serum levels of in patients with RA with no adverse events. soluble IL-6 receptor and higher insulin-like fully-human mAb against IL-1b for the 146 Romas E, Sims NA, Hards DK et al.: growth factor 1. Arthritis Res. Ther. 5, potential treatment of inflammatory disorders. Osteoprotegerin reduces osteoclast numbers R202–R209 (2003). Curr. Opin. Mol. Ther. 11, 81–89 (2009). and prevents bone erosion in collagen- 121 Yogesha SD, Khapli SM, Srivastava RK et al.: 136 Walitt B, Pettinger M, Weinstein A et al.: induced arthritis. Am. J. Pathol. 161, IL-3 inhibits TNF-a-induced bone resorption Effects of postmenopausal hormone therapy 1419–1427 (2002). on rheumatoid arthritis: the women’s health and prevents inflammatory arthritis. 147 Gravallese EM: Bone destruction in arthritis. initiative randomized controlled trials. J. Immunol. 182, 361–370 (2009). Ann. Rheum. Dis. 61, II84–II86 (2002). Arthritis Rheum. 59, 302–310 (2008). 122 Fonseca JE, Santos MJ, Canhao H, Choy E: 148 Komuro H, Olee T, Kuhn K et al.: 137 Stefanick ML: Estrogens and progestins: Interleukin-6 as a key player in systemic The osteoprotegerin/receptor activator of background and history, trends in use, and inflammation and joint destruction. nuclear factor kB/receptor activator of guidelines and regimens approved by the Autoimmun. Rev. 2, 2 (2009). nuclear factor kB ligand system in cartilage. US Food and Drug Administration. 123 Oldfield V, Dhillon S, Plosker GL, Arthritis Rheum. 44, 2768–2776 (2001). Am. J. Med. 118, 64–73 (2005). Tocilizumab: a review of its use in the 149 Salminen-Mankonen HJ, Morko J, Vuorio E: 138 Nielsen RH, Christiansen C, Stolina M, management of rheumatoid arthritis. Drugs 69, Role of cathepsin K in normal joints and in Karsdal MA: Oestrogen exhibits type II 609–632 (2009). the development of arthritis. Curr. Drug collagen protective effects and attenuates 124 Baslund B, Tvede N, Danneskiold-Samsoe B Targets 8, 315–323 (2007). collagen-induced arthritis in rats. Clin. Exp. et al.: Targeting interleukin-15 in patients with Immunol. 152, 21–27 (2008). 150 Hou WS, Li Z, Gordon RE et al.: rheumatoid arthritis: a proof-of-concept study. Cathepsin k is a critical protease in synovial 139 Jochems C, Lagerquist M, Hakansson C, Arthritis Rheum. 52, 2686–2692 (2005). fibroblast-mediated collagen degradation. Ohlsson C, Carlsten H: Long-term 125 Senolt L, Vencovsky J, Pavelka K, Ospelt C, Am. J. Pathol. 159, 2167–2177 (2001). anti-arthritic and anti-osteoporotic effects of Gay S: Prospective new biological therapies for raloxifene in established experimental 151 Stroup GB, Lark MW, Veber DF et al.: rheumatoid arthritis. Autoimmun. Rev. 25, 25 postmenopausal polyarthritis. Clin. Exp. Potent and selective inhibition of human (2009). Immunol. 152, 593–597 (2008). cathepsin K leads to inhibition of bone 126 Fleisch H: Bisphosphonates: mechanisms of resorption in vivo in a nonhuman primate. 140 Bekker PJ, Holloway D, Nakanishi A, action. Endocr. Rev. 19, 80–100 (1998). J. Bone Miner. Res. 16, 1739–1746 (2001). Arrighi M, Leese PT, Dunstan CR: The 127 Breuil V, Euller-Ziegler L: Bisphosphonate effect of a single dose of osteoprotegerin in 152 Robichaud J, Oballa R, Prasit P et al.: A novel therapy in rheumatoid arthritis. Joint Bone postmenopausal women. J. Bone Miner. Res. class of nonpeptidic biaryl inhibitors of Spine 73, 349–354 (2006). 16, 348–360 (2001). human cathepsin K. J. Med. Chem. 46, 3709–3727 (2003). 128 Harada H, Nakayama T, Nanaka T, Katsumata 141 Anandarajah AP, Schwarz EM: Anti- T: Effects of bisphosphonates on joint damage RANKL therapy for inflammatory bone 153 Kumar S, Dare L, JA Vasko-Moser et al.: and bone loss in rat adjuvant-induced arthritis. disorders: mechanisms and potential clinical A highly potent inhibitor of cathepsin K Inflamm. Res. 53, 45–52 (2004). applications. J. Cell. Biochem. 97, 226–232 (relacatib) reduces biomarkers of bone resorption both in vitro and in an acute 129 Itoh F, Aoyagi S, Kusama H, Kojima M, Kogo (2006). model of elevated bone turnover in vivo in H: Effects of clodronate and alendronate on 142 Lewiecki EM, Miller PD, McClung MR monkeys. Bone 40, 122–131 (2007). local and systemic changes in bone metabolism et al.: Two-year treatment with denosumab in rats with adjuvant arthritis. Inflammation 28, (AMG 162) in a randomized Phase 2 study of 154 Robichaud J, Black WC, Therien M et al.: 15–21 (2004). postmenopausal women with low BMD. Identification of a nonbasic, nitrile- containing cathepsin K inhibitor (MK-1256) 130 Jarrett SJ, Conaghan PG, Sloan VS et al.: J. Bone Miner. Res. 22, 1832–1841 (2007). that is efficacious in a monkey model of Preliminary evidence for a structural benefit of 143 McClung MR, Lewiecki EM, Cohen SB osteoporosis. J. Med. Chem. 51, 6410–6420 the new bisphosphonate zoledronic acid in early et al.: Denosumab in postmenopausal women (2008). rheumatoid arthritis. Arthritis Rheum. 54, with low bone mineral density. N. Engl. 1410–1414 (2006). J. Med. 354, 821–831 (2006). 155 Merlino LA, Curtis J, Mikuls TR, Cerhan JR, Criswell LA, Saag KG: Vitamin D intake 131 Herrak P, Gortz B, Hayer S et al.: Zoledronic 144 Bekker PJ, Holloway DL, Rasmussen AS is inversely associated with rheumatoid acid protects against local and systemic bone et al.: A single-dose placebo-controlled study arthritis: results from the Iowa Women’s loss in tumor necrosis factor-mediated arthritis. of AMG 162, a fully human monoclonal Health Study. Arthritis Rheum. 50, 72–77 Arthritis Rheum. 50, 2327–2337 (2004). antibody to RANKL, in postmenopausal (2004). 132 Sims NA, Green JR, Glatt M et al.: Targeting women. J. Bone Miner. Res. 19, 1059–1066 156 Nielen MM, van Schaardenburg D, Lems osteoclasts with zoledronic acid prevents bone (2004). WF et al.: Vitamin D deficiency does not destruction in collagen-induced arthritis. 145 Cohen SB, Dore RK, Lane NE et al.: increase the risk of rheumatoid arthritis: Arthritis Rheum. 50, 2338–2346 (2004). Denosumab treatment effects on structural Comment on the article by Merlino et al. damage, bone mineral density, and bone 133 Palacios S: Advances in hormone replacement Arthritis Rheum. 54, 3719–3720 (2006). therapy: making the menopause manageable. turnover in rheumatoid arthritis: a 157 Leventis P, Patel S: Clinical aspects of BMC Womens Health 8, 22 (2008). twelve-month, multicenter, randomized, double-blind, placebo-controlled, Phase II vitamin D in the management of rheumatoid 134 de Villiers T: The role of menopausal hormone clinical trial. Arthritis Rheum. 58, 1299–1309 arthritis. Rheumatology (Oxford) 47, therapy in osteoporosis. Climacteric 10, 71–73 (2008). 1617–1621 (2008). (2007).

future science group www.futuremedicine.com 581 Review Cantley, Smith & Haynes

158 Leung BP, Sattar N, Crilly A et al.: A novel 172 Chen LF, Greene WC: Shaping the nuclear 185 Hofbauer LC, Lacey DL, Dunstan CR, anti-inflammatory role for simvastatin in action of NF-kB. Nat. Rev. Mol. Cell Biol. 5, Spelsberg TC, Riggs BL, Khosla S: inflammatory arthritis. J. Immunol. 170, 392–401 (2004). Interleukin-1b and tumor necrosis factor-a, 1524–1530 (2003). 173 Trepel J, Birrer MJ: Link A between histone but not interleukin-6, stimulate 159 Kanda H, Yokota K, Kohno C et al.: Effects deacetylase inhibitors and NF-kB signaling. osteoprotegerin ligand gene expression in of low-dosage simvastatin on rheumatoid Cell Cycle 2, 452–453 (2003). human osteoblastic cells. Bone 25, 255–259 (1999). arthritis through reduction of Th1/Th2 and 174 Richon VM, O’Brien JP: Histone deacetylase CD4/CD8 ratios. Mod. Rheumatol. 17, inhibitors: a new class of potential therapeutic 186 Hashizume M, Hayakawa N, Mihara M: 364–368 (2007). agents for cancer treatment. Clin. Cancer Res. IL-6 trans-signalling directly induces 160 Funk JL, Chen J, Downey KJ, Clark RA: 8, 662–664 (2002). RANKL on fibroblast-like synovial cells and is involved in RANKL induction by TNF-a Bone protective effect of simvastatin in 175 Chung YL, Lee MY, Wang AJ, Yao LF: and IL-17. Rheumatology (Oxford) 47, experimental arthritis. J. Rheumatol. 35, A therapeutic strategy uses histone deacetylase 1635–1640 (2008). 1083–1091 (2008). inhibitors to modulate the expression of genes 161 Tourinho TF, Capp E, Brenol JC, Stein A: involved in the pathogenesis of rheumatoid 187 Chen L, Wei XQ, Evans B, Jiang W, Physical activity prevents bone loss in arthritis. Mol. Ther. 8, 707–717 (2003). Aeschlimann D: IL-23 promotes osteoclast formation by up-regulation of receptor premenopausal women with rheumatoid 176 Nakamura T, Kukita T, Shobuike T et al.: activator of NF-kB (RANK) expression in arthritis: a cohort study. Rheumatol. Int. 28, Inhibition of histone deacetylase suppresses myeloid precursor cells. Eur. J. Immunol. 38, 1001–1007 (2008). osteoclastogenesis and bone destruction by 2845–2854 (2008). 162 Bingham CO 3rd: Emerging therapeutics for inducing IFN-b production. J. Immunol. 175, rheumatoid arthritis. Bull. NYU Hosp. Jt Dis. 5809–5816 (2005). 188 Komine M, Kukita A, Kukita T, Ogata Y, Hotokebuchi T, Kohashi O: Tumor necrosis 66, 210–215 (2008). 177 Butler KV, Kozikowski AP: Chemical origins factor-a cooperates with receptor activator 163 Kong YY, Yoshida H, Sarosi I et al.: OPGL is of isoform selectivity in histone deacetylase of nuclear factor kB ligand in generation of a key regulator of osteoclastogenesis, inhibitors. Curr. Pharm. Des. 14, 505–528 osteoclasts in stromal cell-depleted rat bone lymphocyte development and lymph-node (2008). marrow cell culture. Bone 28, 474–483 organogenesis. Nature 397, 315–323 (1999). 178 Itoh Y, Suzuki T, Miyata N: Isoform-selective (2001). 164 Goldring SR, Goldring MB: Eating bone or histone deacetylase inhibitors. Curr. Pharm. 189 Quinn JM, Horwood NJ, Elliott J, adding it: the Wnt pathway decides. Nat. Des. 14, 529–544 (2008). Gillespie MT, Martin TJ: Fibroblastic Med. 13, 133–134 (2007). 179 Monneret C: Histone deacetylase inhibitors. stromal cells express receptor activator of 165 Nakamura Y, Nawata M, Wakitani S: Eur. J. Med. Chem. 40, 1–13 (2005). NF-kB ligand and support osteoclast Expression profiles and functional analyses of 180 Lin HS, Hu CY, Chan HY et al.: differentiation. J. Bone Miner. Res. 15, Wnt-related genes in human joint disorders. Anti-rheumatic activities of histone 1459–1466 (2000). Am. J. Pathol. 167, 97–105 (2005). deacetylase (HDAC) inhibitors in vivo in 190 Kotake S, Sato K, Kim KJ et al.: 166 Sen M: Wnt signalling in rheumatoid collagen-induced arthritis in rodents. Interleukin-6 and soluble interleukin-6 arthritis. Rheumatology (Oxford) 44, 708–713 Br. J. Pharmacol. 150, 862–872 (2007). receptors in the synovial fluids from (2005). 181 Wong PK, Quinn JM, Sims NA, rheumatoid arthritis patients are responsible 167 Holzhausen M, Spolidorio LC, Ellen RP van Nieuwenhuijze A, Campbell IK, for osteoclast-like cell formation. J. Bone et al.: Protease-activated receptor-2 activation: Wicks IP: Interleukin-6 modulates Miner. Res. 11, 88–95 (1996). a major role in the pathogenesis of production of T lymphocyte-derived 191 Kotake S, Udagawa N, Takahashi N et al.: Porphyromonas gingivalis infection. cytokines in antigen-induced arthritis and IL-17 in synovial fluids from patients with Am. J. Pathol. 168, 1189–1199 (2006). drives inflammation-induced rheumatoid arthritis is a potent stimulator of 168 Holzhausen M, Spolidorio LC, Vergnolle N: osteoclastogenesis. Arthritis Rheum. 54, osteoclastogenesis. J. Clin. Invest. 103, Proteinase-activated receptor-2 (PAR2) 158–168 (2006). 1345–1352 (1999). agonist causes periodontitis in rats. J. Dent. 182 Mihara M, Moriya Y, Kishimoto T, Ohsugi Y: Res. 84, 154–159 (2005). Interleukin-6 (IL-6) induces the proliferation „ „Websites 169 Abraham LA, Jenkins AL, Stone SR, of synovial fibroblastic cells in the presence of 201 Welfare AIoHa: A Picture of Rheumatoid Mackie EJ: Expression of the thrombin soluble IL-6 receptor. Br. J. Rheumatol. 34, 321–325 (1995). Arthritis in Australia. Ageing HA (Ed.), (2009). receptor in developing bone and associated www.aihw.gov.au/publications/index.cfm/ tissues. J. Bone Miner. Res. 13, 818–827 183 Yago T, Nanke Y, Kawamoto M et al.: IL-23 title/10524 (1998). induces human osteoclastogenesis via IL-17 in vitro, and anti-IL-23 antibody attenuates 202 Chondrex Inc.: Arthrogen-CIA Arthritogenic 170 Smith R, Ransjo M, Tatarczuch L et al.: Monoclonal Antibodies. 2007, 2009. Activation of protease-activated receptor-2 collagen-induced arthritis in rats. Arthritis Res. Ther. 9, R96 (2007). www.chondrex.com/uploads/2008_ leads to inhibition of osteoclast Catalog_2_Web.pdf differentiation. J. Bone Miner. Res. 19, 184 Jimi E, Nakamura I, Duong LT et al.: 507–516 (2004). Interleukin 1 induces multinucleation and 171 Ferrell WR, Lockhart JC, Kelso EB et al.: bone-resorbing activity of osteoclasts in the Essential role for proteinase-activated absence of osteoblasts/stromal cells. Exp. Cell receptor-2 in arthritis. J. Clin. Invest. 111, Res. 247, 84–93 (1999). 35–41 (2003).

582 Int. J. Clin. Rheumatol. (2009) 4(5) future science group