Role of Bone Marrow in the Pathogenesis of Rheumatoid Arthritis

Rheumatoid arthritis (RA) is a chronic inflammatory disease characterized by persistent synovial proliferation. Thus, joints in RA consist of massive proliferating synovium, forming an invading tissue termed pannus, which results in the destruction of cartilage and bone. One of the most important histologic characteristics of the synovium in RA includes cellular proliferation in the lining layer as well as in the sublining layer (Tak, 2004). In the lining layer, both type A and type B synoviocytes, alternatively called intimal macrophages and fibroblast-like synoviocytes, respectively, are found to proliferate (Tak, 2004). In the sublining layer, there is infiltration of a variety of cells, including dendritis cells (DC), lymphocytes, plasma cells, and polymorphnuclear leukocytes. Notably, lymphoid cluster in RA synovium sometimes forms pseudo-germinal center, consisting of CD20+ B cells in the center surrounded by CD4+ T cells (Tak, 2004; Hirohata, 2004). In the synovium of RA, neovascularization is usually accompanied by lining cell proliferation and inflammatory cell infiltration (Firestein, 1999). In fact, lining cells and inflammatory cells have been found to produce angiogenic growth factors (Koch, 1998). It should be noted, however, that the synovium of RA also showed neovascularization in the areas without either lining cell proliferation or inflammatory cell infiltration, suggesting that the neovascularization might be one of the primary abnormal features that are most proximal to the etiology of RA (Hirohata & Sakakibara, 1999). A number of studies have suggested that abnormal activation of normal joint constituents, such as synovial lining cells, play a pivotal role in the synovial hyperplasia in RA (Shiozawa & Tokuhisa, 1992). However, increasing attention has emerged to the role of bone marrow in the pathogenesis of RA. The present article overviews an update on the role of bone marrow in the pathogenesis of RA.


Introduction
Rheumatoid arthritis (RA) is a chronic inflammatory disease characterized by persistent synovial proliferation.Thus, joints in RA consist of massive proliferating synovium, forming an invading tissue termed pannus, which results in the destruction of cartilage and bone.One of the most important histologic characteristics of the synovium in RA includes cellular proliferation in the lining layer as well as in the sublining layer (Tak, 2004).In the lining layer, both type A and type B synoviocytes, alternatively called intimal macrophages and fibroblast-like synoviocytes, respectively, are found to proliferate (Tak, 2004).In the sublining layer, there is infiltration of a variety of cells, including dendritis cells (DC), lymphocytes, plasma cells, and polymorphnuclear leukocytes.Notably, lymphoid cluster in RA synovium sometimes forms pseudo-germinal center, consisting of CD20+ B cells in the center surrounded by CD4+ T cells (Tak, 2004;Hirohata, 2004).In the synovium of RA, neovascularization is usually accompanied by lining cell proliferation and inflammatory cell infiltration (Firestein, 1999).In fact, lining cells and inflammatory cells have been found to produce angiogenic growth factors (Koch, 1998).It should be noted, however, that the synovium of RA also showed neovascularization in the areas without either lining cell proliferation or inflammatory cell infiltration, suggesting that the neovascularization might be one of the primary abnormal features that are most proximal to the etiology of RA (Hirohata & Sakakibara, 1999).A number of studies have suggested that abnormal activation of normal joint constituents, such as synovial lining cells, play a pivotal role in the synovial hyperplasia in RA (Shiozawa & Tokuhisa, 1992).However, increasing attention has emerged to the role of bone marrow in the pathogenesis of RA.The present article overviews an update on the role of bone marrow in the pathogenesis of RA. (Hematoxylin and eosin, original magnification x25) Fig. 2. Synovium-like tissue at pericardium lesion in an RA patient including proteoglycans, cytokines, arachidonic acid metabolites, and matrix metalloproteinases (MMPs), that lead to the destruction of joints (Firestein,1996).Unlike intimal macrophages, the precise origin of type B synoviocytes remains unclear, although they are thought to arise from the sublining tissue or other support structures of a joint (Firestein,1996).On the other hand, a number of studies have shown that peripheral blood dendritic cells (DC) accumulate in the synovium, where they undergo phenotypic and functional differentiation in situ (Zvaifler et al.,1985;Thomas et al.,1994).It has been also shown that synovial DC gradually lose their dinstinct morphologic appearance and become indistinguishable from fibroblasts in vitro (Hendler et al.,1985).Moreover, Kyogoku et al. identified the presence of DC-like cells that strongly express major histocompatibility complex (MHC) class II antigens and interact with T lymphocytes, in the sublining layers of the RA synovium (Kyogoku et al.,1992).They also showed that the sublining DC-like cells proliferate and differentiate into type A as well as type B synoviocytes to replace the lining layers (Kyogoku et al.,1992).

Generation of type B synoviocytes from bone marrow CD34+ cells in RA
Since it was shown that DC are derived from bone marrow CD34+ cells (Reid et al.,1992;Szabolcs et al.,1995;Chen et al.,2004), it was also likely that type B synoviocytes might be induced from bone marrow progenitors.In this regard, we previously demonstrated that bone marrow CD34+ cells from RA patients have abnormal capacities to respond to tumor necrosis factor-α (TNF-α) and to differentiate into fibroblast-like cells (FLC) producing MMP-1, suggesting that bone marrow CD34+ cells might generate type B synoviocytes and thus could play an important role in the pathogenesis of RA (Hirohata et al.,2001).Thus, CD34+ cells from the bone marrow of RA patients differentiated into cells with fibroblastlike morphology, which expressed prolyl 4-hydroxylase, in the presence of stem cell factor (SCF), GM-CSF, and TNF-α, much more effectively than CD34+ cells from the bone marrow of control subjects (Hirohata et al.,2001).

Capacity of bone marrow DC to differentiate into type B synoviocytes
We have recently demonstrated that bone marrow plasmacytoid DC (pDC) as well as myeloid DC (mDC), irrespective of their origin from RA bone marrow or osteoarthritis (OA) bone marrow, have prominent capacity to differentiate into FLC producing MMP-1 especially under influences of TNF-α (Hirohata et al.,2011).Of note, depletion of pDC from RA bone marrow CD34+ cells significantly diminished their capacities to differentiate into FLC, which were restored by addition of pDC in a dose-response manner (Hirohata et al.,2011).It should be pointed out that generation of FLC from RA bone marrow CD34+ cells or pDC was correlated with MMP-1 levels in culture supernatants (Hirohata et al.,2001;Hirohata et al.,2011).On the other hand, it has been demonstrated that cadherin-11 is abundantly expressed in type B synoviocytes (Chang et al.,2010) compared with lung or dermal fibroblasts (Vandooren et al.,2008).Accordingly, the FLC induced from RA and OA bone marrow pDC expressed comparable amounts of cadherin-11 mRNA to RA and OA synovial FLC (Hirohata et al.,2011).These results indicate that DC are one of the progenitors of type B synoviocytes irrespective of RA or OA and suggest that bone marrow CD34+ cells might differentiate into type B synoviocyte-like cells via DC, since DC have been demonstrated to originate from CD34+ cells (Reid et al.,1992;Szabolcs et al.,1995;Chen et al.,2004).It is also likely that expansion of immature DC from bone marrow CD34+ cells might be upregulated in RA compared with in OA, accounting for the enhanced capacity of RA bone marrow CD34+ cells to differentiate into FLC upon stimulation with TNF-α, although further studies are required to confirm this point.

Recruitment of type B synoviocytes and their precursors into RA joints: Role of DC
It is thus suggested that the presence of abnormal precursors within the bone marrow progenitor cells might play an important role in the pathogenesis of RA by providing a repopulating reservoir of type B synoviocytes, as has been also suggested in other recent studies (Sen et al.,2000).Notably, the numbers of mDC and pDC have been found to be significantly decreased in RA peripheral blood, whereas both mDC and pDC are present in synovial fluid from RA (Jongbloed et al.,2006).In fact, previous study showed that pDC are recruited to RA synovial tissues and contribute into the local inflammatory environment (Lande et al,2004;Cavanagh et al.,2005).Recent studies have disclosed that the characteristic clinical phenomenon of destructive arthritis spreading between joints is mediated, at least in part, by the transmigration of activated RA synovial fibroblasts (Lefèvre et al.,2009).Thus, RA synovial fibroblasts showed an active movement from human RA synovial tissue or human cartilage-sponge complex containing RA synovial fibroblasts implanted into the SCID mouse to the naive cartilage implanted at the contralateral flank via the vasculature, leading to a marked destruction of the target cartilage (Lefèvre et al.,2009).The movement of DC-like cells from the sublining layer to the lining layer in RA synovium (Kyogoku et al.,1992) and the presence of DC in synovial fluid (Jongbloed et al.,2006;Lande et al,2004) strongly suggest that DC might be also released from the joint to the systemic circulation via draining veins.Since bone marrow pDC as well as mDC have capacities to differentiate into type B synoviocyte-like cells, it is possible that DC, as precursors for synovial fibroblasts, also contribute to the spread of destructive arthritis between joints in RA.

RA as a disease of antigen-presenting cells
A number of studies have demonstrated that RA is strongly associated with HLA-DR4 or DR1 (Nepom,2001), which are involved in the presentation of antigens to T cells.These results suggest that the antigen presentation involving HLA-DR4 or the shared epitope might play a critical role in the development of synovitis in human RA.In fact, the interactions between DC and T cells, possibly through MHC class II antigens, have been disclosed in the sublining layers of the RA synovium (Kyogoku et al.,1992).If the antigens presented by APC to T cells are perpetuating antigens, such as autoantigens or antigens of persistently infected virus, that are presented through MHC class II molecules, continuing activation of APC might take place.Further studies to explore such antigens that involve persistent interactions between APC and T cells would be still important for the complete understanding of the pathogenesis of RA.As highlighted above, bone marrow derived monocytes and DC have been shown to be the precursors of type A synoviocytes and type B synoviocytes, respectively.It is therefore suggested that RA might be a disease of dysregulated activation of antigen-presenting cells (APC), leading to synovial proliferation.Lymphocytes activation in the synovium can also be triggered by the activation of APC, accounting for activation of T cells and B cells in the synovium.Triggering with arthritogenic antigens, followed by dysregulated generation of APC from the bone marrow might result in persistent recruitment of APC into the synovium.

Bone marrow abnormalities and angiogenesis in RA
A number of studies indicated that neovascularization is crucial to the synovial hyperplasia in RA (Koch,1998;Hirohata & Sakakibara,1999).Postnatal neovascularization has been attributed to so-called angiogenesis, a process characterized by the sprouting of new capillaries from preexisting blood vessels (Folkman & Shing,1992).However, recent studies have demonstrated that endothelial progenitor cells of bone marrow origin play a significant role in the de novo formation of capillaries without preexisting blood vessels, socalled vasculogenesis (Asahara et al.,1997;Gehling et al.,2000;Bhattacharya et al.,2000;Lin et al.,2000).
We also showed that RA bone marrow CD34+ cells have enhanced capacities to differentiate into endothelial cells in relation to synovial vascularization (Hirohata & Yanagida et al.,2004).Therefore, bone marrow CD34+ cells might contribute to synovial neovascularization by supplying endothelial precursor cells and, thus, play an important role in the pathogenesis of RA.Neovascularization of the synovium is not unique to RA.It has also been observed in OA synovium and has been shown to play an important role in the development of new cartilage and mineralization (Brown et al.,1980;Giatromanolaki et al.,2003;Haywood et al.,2003).Of note, recent studies have revealed that levels of expression of the angiogenic factors VEGF and platelet-derived endothelial cell growth factor are increased in RA as well as in OA, relative to normal subjects, whereas the presence of an activated synovial vasculature was high only in RA (Giatromanolaki et al.,2003).Moreover, the vascular activation by VEGF/KDR was significantly lower in OA than in RA patients, although the activation of the hypoxia inducing factor α (HIFα) pathway was comparable in OA and RA patients (Giatromanolaki et al.,2003).These observations suggest the presence of intrinsic abnormalities in synovial endothelial cells in RA patients.Of note, we have disclosed that the expression of VEGFR-2/KDR mRNA in RA bone marrow CD34+ cells was significantly higher than that in OA bone marrow CD34+ cells (Hirohata & Yanagida et al.,2004).It is therefore likely that the differences in VEGF/KDR vascular activation in bone marrow CD34+ cells might result in differences in their capacity to generate endothelial progenitor cells between RA and OA patients (Koch et al.,1994;Giatromanolaki et al.,2001).It has been shown that decreased numbers and impaired function of endothelial progenitor cells (EPCs) resulting in defective vasculogenesis are associated with RA, leading to premature atherosclerosis (Herbrig et al., 2006;Pakozdi et al., 2009).On the other hand, it has been recently disclosed that EPCs can be differentiated into 2 subpopulations, EPCs of monocytic versus hemangioblastic origin, which have been denoted as early-outgrowth and late-outgrowth EPCs, respectively (Jodon de Villeroché et al, 2010).More importantly, lateo u t g r o w t h E P C s h a v e b e e n f o u n d t o b e i n creased and have higher colony formation capacity in the active stage of RA.It is therefore likely that hemangioblastic EPC-dependent vasculogenesis might be associated with active inflammation and accelerated atherosclerosis in RA.

RA and hematopoietic stem cell transplantation
Although autologous hematopoietic stem cell transplantation (HSCT) has been used to treat severe RA in limited case reports (Joske,1997;Durez et al.,1998), a study with greater numbers of patients have disclosed that recurrence of RA is frequent after the autologous HSCT (Snowden et al.,2004;Bingham & Moore, 2004).Such frequent recurrence after autologous HSCT clearly indicates that abnormalities in bone marrow stem cells persist after the treatment.It has been proposed that bone marrow CD34+ progenitor cell reserve and function are defective in RA probably due, at least in part, to a TNF-α mediated effect, because significant restoration of the disturbed hematopoiesis was obtained following anti-TNF-α treatment (Papadaki et al.,2002;Porta et al.,2004).It should be noted, however, that blockade of TNF-α is not curative for RA in spite of its epoch-making impact on treatment of RA (Feldmann & Maini,2001).Thus, recurrence of RA is noted after discontinuation of blockade of TNF-α or even during anti-TNF-α therapy (Feldmann & Maini,2001).It is therefore likely that intrinsic abnormalities that were not secondary to the influences of TNF-α might be present in bone marrow progenitor cells, leading to recurrence of RA.In fact, although abnormal regulatory networks in the immune response and cell cycle categories were identified in bone marrow mononuclear cells from RA patients (Lee et al.,2011), it is possible that such changes might be secondary to systemic inflammation, presumably due to proinflammatory cytokines.In this regard, beyond its role in angiogenesis, the demonstration of the abnormal expression of VEGFR-2/KDR mRNA in RA bone marrow CD34+ cells (Hirohata & Yanagida et al.,2004) have brought an impact as the fist evidence for the intrinsic abnormality in RA bone marrow.Mesenchymal stem cells (MSCs) have been shown to have potent anti-inflammatory and immunomodulatory properties through suppression of Th1/Th17 response and induction of Treg response (Macdonald et al., 2011).However, it remains unclear whether MSC therapy is beneficial for treatment of RA.

Nuclear factor kappa B1 (NFkB1)
As mentioned above, bone marrow CD34+ cells from RA patients have abnormal capacities to respond to TNF-α and to differentiate into FLC producing MMP-1, suggesting that abnormalities in bone marrow CD34+ cells might play a role in the pathogenesis in RA (Hirohata et al.,2001).TNF-α is one of the first triggers to be found effective for the activation of NFκ B (Müller-Ladner et al.,2002).Of note, we have recently demonstrated that RA bone marrow CD34+ cells showed enhanced expression of NFκ B1 (p50), silencing of which resulted in prevention of their differentiation into FLC (Hirohata et al.,2006).

FK506-binding protein5 (FKBP5)
Nakamura et al. recently disclosed that the expression of several genes including amphiregulin (AREG), chemokine receptor 4 (CXCR4), and FK506-binding protein 5 (FKBP5), was augmented in bone marrow mononuclear cells from RA patients compared with those from OA patients (Nakamura et al.,2006).Interestingly, FKBP5 was found to be involved in nuclear translocation and activation of NFκB by degradation of inhibitor of NFκB alpha (IκBα) in a human megakaryoblastic leukemia cell line (Bouwmeester et al.,2004;Komura et al.,2005).It is therefore suggested that the up-regulated expression of both NFκB1 and FKBP5 mRNAs in bone marrow CD34+ cells from RA patients might be involved cooperatively in their abnormal responses to TNF-α to differentiate into type B synoviocyte-like cells.Although the expression of mRNAs for AREG, CXCR4 and FKBP5 has been shown to be augmented in RA bone marrow mononuclear cells (Nakamura et al.,2006),only FKBP5 mRNA expression was significantly upregulated in bone marrow CD34+ cells from RA (Matsushita et al.,2010).Therefore, it is suggested that the up-regulation of the expression of mRNAs for AREG and CXCR4 in RA bone marrow mononuclear cells might be sequelae of systemic inflammation of RA.By contrast, the up-regulation of FKBP5 mRNA expression in RA bone marrow CD34+ cells might not be secondary to systemic inflammation, but a primary abnormality in bone marrow CD34+ cells (Matsushita et al.,2010).It has been previously shown that TNF-α enhanced NFκB1 mRNA expression in bone marrow CD34+ cells from healthy individuals (Hirohata et al.,2006).However, TNF-α did not enhance FKBP5 mRNA expression in bone marrow CD34+ cells from healthy individuals (Matsushita et al.,2010).It is therefore confirmed that apart from NFkB1, the enhanced FKBP5 mRNA expression in RA bone marrow CD34+ cells is not secondary to systemic inflammation of RA.

Krüppel like factor 5 (KLF-5)
Krüppel like factor 5 (KLF-5), a zinc finger-containing transcription factor, activates many genes, including platelet-derived growth factor (PDGF) A/B, plasminogen activator inhibitor-1, inducible nitric oxide synthase and VEGF receptors (Shindo et al.,2002;Nagai et al.,2005).KLF-5 has been shown to cooperate with NFkB1 to activate PDGF-A gene expression (Nagai et al.,2005;Aizawa et al.,2004), which might be involved in synovial fibroblast-like cell proliferation (Ohba et al.,1996).KLF-5 mRNA expression in bone marrow CD34+ cells was significantly higher in RA patients than in OA patients (Hirohata et al.,2009)  Previous studies also demonstrated that the suppression of KLF-5 by silencing RNA resulted in a reduction of NFkB1 mRNA in IEC6 cells stimulated with lipopolysaccharide, indicating that KLF-5 is an upstream regulator for NFkB1 mRNA expression in IEC6 cells (Chanchevalap et al.,2006).Taken together, it is most likely that the upregulation of KLF-5 mRNA expression might lead to the enhanced expression of NFkB1 mRNA in bone marrow CD34+ cells, but not vice versa, in RA.In addition, the upregulation of KLF-5 mRNA as well as NFkB1 mRNA in RA bone marrow CD34+ cells might result in their abnormal capacities to differentiate into FLC.Although it is strongly suggested that KLF-5 might be an upstream regulator of NFkB1 mRNA in bone marrow CD34+ cells, further studies to explore the mechanism of abnormal expression of KLF-5 mRNA in BM CD34+ cells and its relation with FKBP5 would be important.

Conclusion
As summarized in Fig. 4, accumulating evidence has been provided for the involvement of bone marrow in the pathogenesis of RA.Thus, all the constituents in the proliferating synovial tissues might be supplied from bone marrow CD34+ cells.Apparently, RA bone marrow CD34+ cells have abnormal mRNA expression for several genes, possibly resulting in abnormal differentiation of monocytes and DC.Moreover, it is strongly suggested that DC, as precursors for synovial fibroblasts, might also contribute to the spread of destructive arthritis between joints in RA (Hirohata et al.,2011;Lefèvre et al.,2009).In the past decade, the importance of TNF-α in the pathogenesis of RA has come to be increasingly appreciated (Feldmann et al.,1996).We have revealed that CD34+ cells from bone marrow of RA patients have abnormal responsiveness to TNF-α (Hirohata et al.,2001).However, the precise sequelae of abnormal responses of CD34+ cells from bone marrow of RA patients to TNF-α remain unclear.KLF-5 might upregulate the expression of mRNAs for NFkB1 and VEGFR-2/KDR, whereas FKBP5 might enhance activation of NFkB, resulting in further upregulation of NFkB1 mRNA expression.Further studies that explore in detail the mechanism of abnormal expression of the genes, especially KLF-5 and FKBP5, in CD34+ cells from bone marrow of RA patients would be helpful in gaining a complete understanding of the etiology as well as the pathogenesis of RA.In this regard, we showed previously that GM-CSF-stimulated bone marrow CD34+ cells from 3 of 8 RA patients, but none from 7 OA patients, gave rise to spontaneous transformation of highly purified B cells of Epstein-Barr virus (EBV)-seronegative healthy donors, whereas neither GM-CSFstimulated bone marrow CD34+ cells alone nor highly purified B cells alone gave rise to spontaneously transformed B cell lines (Hirohata et al.,2000).All the transformed B cell lines Fig. 1.Transendothelial migration of monocyte into RA synovium.
. It is thus likely that the upregulation of VEGFR-2/KDR mRNA expression might be secondary to the enhanced KLF-5 mRNA expression in RA bone marrow CD34+ cells.Of note, TNF-α enhanced NFkB1 mRNA expression, but not KLF-5 m R N A e x p r e s s i o n , i n b o n e m a r r o w C D 3 4 + c e l l s f r o m n o r m a l i n d i v i d u a l s ( F i g .3 ) (Hirohata et al.,2009).

Fig. 3 .
Fig. 3. Effect of TNF-α on the expression of mRNAs for NFkB1 and KLF-5 in bone marrow CD34+ cells from a healthy donor

Fig. 4 .
Fig. 4. Schema for the suggested role of bone marrow in the pathogenesis of RA