B Cell Therapies for Rheumatoid Arthritis: Beyond B cell Depletion




Initially suggested by the presence of rheumatoid factor autoantibodies, multiple pathogenic roles for B cells (both antibody-mediated and antibody-independent) in rheumatoid arthritis (RA) now are supported by a growing body of experimental observations and human studies. The pathogenic significance of B cells in this disease also has been established conclusively by the proven benefit of Rituximab-induced B cell depletion in RA patients refractory to tumor necrosis factor (TNF) blockade. This article reviews the rationale for the use of B cell-targeting therapies in RA and discusses the caveats and limitations of indiscriminate B cell depletion as currently applied, ncluding incomplete depletion of pathogenic B cells and elimination of protective B cells. Finally, it presents alternative therapeutic strategies that exploit current knowledge of B cell activation, survival, and differentiation to provide more selective B cell and plasma cell targeting.


Although widely accepted that rheumatoid arthritis (RA) is mediated by inflammatory cytokines produced by macrophages and synovial fibroblasts in a T-dependent fashion, the success of B cell depletion (BCD) has prompted a re-evaluation of the place occupied by B cells in the pathogenesis and treatment of this disease. Indeed, a first glimpse of B cell abnormalities in RA was provided by the description more than 50 years ago of the presence in most patients of increased levels of autoantibodies directed against immunoglobulin G (IgG) (rheumatoid factor, RF), and more recently by the recognition of autoantibodies against neo-antigens created by the citrullination of cyclic peptides (CCP). It is still uncertain, however, whether these autoantibodies play a direct pathogenic role or are simply produced as a by-product of inflammation as illustrated by the production of RF in situations characterized by polyclonal B cell activation (including nonautoimmune bacterial endocarditis and hepatitis C infection). Nonetheless, multiple lines of evidence strongly indicate that the production of RF in RA is mediated by antigen-driven processes that are locally amplified in the synovium. Thus, early human work demonstrated that synovial antibody production was enriched for RF and did not reflect systemic immune responses induced by immunization, thereby suggesting enhanced migration or expansion of autoreactive B cells within the target organ. Subsequently, work in autoimmune prone MRL-lpr mice demonstrated that the production of RF was the result of strong antigen-driven selection of oligoclonal B cells. Finally, in keeping with the mouse data, clonal expansions of RF-producing synovial B cells also can be found in the human rheumatoid synovium. Collectively, the evidence indicates that antigen selection plays a critical role in the local expansion of synovial RF B cells and provides at least strong circumstantial evidence for a pathogenic role of autoantibodies or the B cells producing them.


Indeed, the rationale initially posited by Jonathan Edwards to pursue his pioneering studies of BCD in RA was based on the implications that the RF might have for disease pathogenesis. He astutely postulated that in addition to conventional immune complex-mediated inflammatory mechanisms, B cells expressing surface antibodies with RF properties could serve as universal antigen-presenting cells capable of amplifying autoimmune responses through RF-mediated capture and internalization of circulating immune complexes containing any type of autoantigens. This hypothesis had strong experimental underpinnings provided by the experiments of Roosnek and Lanzavecchia. More recent studies have unraveled other mechanisms whereby immune complexes could contribute to disease pathogenesis either through the costimulation of Toll-like receptors (TLR) and Fcγ receptors in dendritic cells and B cells and costimulation of synovial macrophages through Fcγ and C5a receptors.


The potential contributions of B cells have been significantly extended by numerous studies demonstrating that B cells also play many antibody-independent proinflammatory that are likely to the pathogenesis of RA. These functions include activation of CD4 and CD8 T cells, maintenance of CD8 T cell memory, induction of Th1 and Th17 effector cells, recruitment of dendritic cells (DC), suppression of T regulatory cells (Tregs), lymphangiogenesis and production of proinflammatory cytokines including tumor necrosis factor (TNF) and interleukin (IL)-6 and polyclonal B cell activators such as IL-6 and osteopontin. Of particular relevance to RA, synovial B cells are essential to sustain CD4 T cell activation in explanted RA synovium, an effect reversed by anti-CD20 mediated B cell depletion.


B cells are also critical for the organization of the lymphoid architecture through membrane-bound LTα1β2 induction of LTβR signaling. The latter function may be highly relevant to the pathogenesis of RA, since lymphoid infiltrates organized as B/T cell areas and in some cases with germinal center-like structure and function can be found in the RA synovium and have been postulated to play a pathogenic role through local amplification of autoimmune B cells and plasma cell generation. The actual significance of synovial GCs remains controversial due to the relatively small fraction of cases in which this configuration is detected (5% to 25%, possibly on the basis of sampling variability and patchy disease) and the lack of correlation in some studies between these findings and local autoimmunity. These studies stand in contrast to the recent demonstration that in 25% of RA patients analyzed, ectopic synovial lymphoid structures containing follicular dendritic cells express molecular markers of in situ somatic hypermutation and class switch recombination and are highly enriched for plasma cells locally producing antibodies against citrullinated fibrinogen. This study conclusively demonstrates the ability of synovial lymphoid structures to locally support the generation and diversification of RA-specific B cell autoreactivity. This is in keeping with earlier indication that synovial memory B cells could be generated locally, a concept later disputed by the finding of clonally related B cells in separate joints and in the systemic circulation, which suggests that, alternatively, pathogenic B cells may arise systemically and then migrate to the inflamed synovium where they could be diversified further in ectopic GC-like structures. Although the limited available data cannot establish whether pathogenic B cells could originate initially in the joint and then metastasize to other joints, the notion that a constant supply of recirculating B cells is important for maintaining the B cell microenvironment of inflamed joints bears significant connotations for the understanding of the mechanisms of action of B cell-targeted therapies in RA.


Current B cell therapies—limitations and challenges of universal B cell depletion


Rituximab-induced BCD is US Food and Drug Administration (FDA)-approved, in combination with methotrexate, for the treatment of RA patients who fail or are unable to tolerate anti-TNF drugs. Although significant favorable responses are attained in this difficult to treat group, only 50% to 65% of patients maintain ACR20 or moderate-to-good European League Against Rheumatism responses 6 months after treatment, and disease relapses frequently ensue after B cell repopulation, which typically occurs 6 to 9 months after treatment. As a result, systematic retreatment with Rituximab every 6 to 9 months is becoming a frequent practice. Repeated retreatment tends to be beneficial in patients who respond to the first treatment and has proven largely safe in patients receiving six to eight cycles of therapy. However, serum IgM levels below normal values and a downward trend in IgG levels are seen with repeated treatment and vaccine responses are likely to be compromised in chronically depleted patients. Moreover, the effect of chronic B cell depletion on protective T cell immunity also remains to be properly studied. Therefore, the need for indefinite retreatment should create significant concern that is further compounded by the report of progressive multifocal leukoencephalopathy (PML) in three RA patients treated with Rituximab (out of approximately 100,000 patients treated), including one recent case who had not received previous treatment with anti-TNF agents.


Typically, Rituximab achieves almost universal peripheral blood B cell depletion defined as less than 5 CD19+ cell/μL but less consistent profound depletion of less than 0.01 cells/μL, a level that seems to correlate with greater and more sustained clinical response. As it is also the case with systemic lupus erythematosus (SLE), B cell reconstitution in patients with good clinical responses is initiated by enhanced bone marrow generation of transitional B cells, whereas relapsing patients are characterized by relative expansion of memory cells that presumably Rituximab failed to deplete in the first place. In tissue, some studies have shown a correlation between clinical response to Rituximab, relatively early declines in numbers of synovial B cells, and decreased frequencies of plasma cells and immunoglobulin synthesis activity, whereas other studies have found no correlation between early synovial B cell decline and clinical response. All these studies have used limited numbers of patients, and their interpretation is confounded by the intrinsic variability and limitations of tissue sampling and disease heterogeneity. Nonetheless, the collective information available has been used by Silverman and Boyle to formulate the roadblock hypothesis, which postulates that the benefit of Rituximab could be mediated by interruption of the supply to the synovium of circulating B cells, leading to the progressive attrition of synovial B cells, plasma cells, and other inflammatory cells, particularly macrophages.


Therefore, the current goal of BCD is to achieve complete and sustained B cell depletion in blood and synovium of pathogenic B cells whose elimination might account for initial disease improvement even in the absence of significant autoantibody changes. This goal could be accomplished with either more intensive Rituximab regimens, more efficient anti-CD20 agents, or with combination therapy with other agents (including BAFF blockade to be discussed further). More extensive B cell depletion also could be achieved with antibodies that target surface antigens expressed more widely than CD20 whose expression begins in pre-B cells and is extinguished in most immature plasma cells (a pattern accounting for the limited efficacy of Rituximab to target plasma cells). In contrast, CD19 expression is initiated in pro-B cells and persists even in a fraction of mature CD138 + PC (F-EH Lee and I Sanz, unpublished data, 2010). Accordingly, anti-CD19 antibodies represent a promising alternative, and trials with humanized anti-CD19 antibodies are under development.


Potentially, more effective targeting of B cells and preplasma cells could be achieved by antibodies against CD79α/β, which in addition to a direct killing effect, also inhibit B cell receptor (BCR) activation, display clinical efficacy in lupus-prone MRL/lpr mice, and induce profound GC depletion in cynomolgus monkeys.


However, indiscriminate B cell depletion also may have a dark side, as this intervention also would compromise regulatory B cells capable of delaying progression in untreated patients or sustain the initial remission induced by the depletion of pathogenic B cells. Of note, B cell measurements used in published studies lacked the ability to differentiate between pathogenic and protective B cells. Protective B cell effects can be mediated by several direct mechanisms, including induction of T cell anergy, induction of Tregs or production of anti-inflammatory cytokines including IL-10 or TGFβ. The latter possibility is illustrated by the recent description of IL-10 producing B regulatory cells (Bregs) with a transitional phenotype capable of suppressing arthritis in animal models and by a growing body of knowledge regarding Breg functions in multiple other autoimmune diseases. It therefore could be envisioned that strategies to selectively deplete pathogenic B cells, expand in vivo Bregs, or infuse Breg cells previously expanded in vitro might represent future avenues for B cell therapies in RA.


Undoubtedly, there is a formidable challenge to understanding the different roles of B cells in RA (which could change at different times in the natural history of the disease or differ among disease subsets) and the relative balance between pathogenic and protective B cells at different times after B cell depletion. Collectively, these considerations suggest a model ( Fig. 1 ), in which the best risk/benefit ratio would be provided by induction therapies aimed at profound B cell depletion (or selective depletion of pathogenic B cells) followed by maintenance regimens that promote an environment dominated by protective B cells). The following sections provide a discussion of alternative strategies for B cell targeting (summarized in Table 1 and illustrated by Fig. 2 ).




Fig. 1


Design of therapeutic B cell targeting in rheumatoid arthritis (RA). The proposed therapeutic strategies are based on the duality of B cells, which can play both protective and pathogenic functions (selected functions are presented at the bottom). Throughout the figure, pathogenic and protective functions are color coded (red to blue shading, respectively). Current approaches use indiscriminate B cell deletion with initial response based on elimination of pathogenic functions. In this model, treatment failure is caused by insufficient depletion or preferential re-expansion of pathogenic cells. Such imbalance also would underlie the relapse commonly seen with B cell reconstitution and creates the need for additional cycles of B cell depletion. B cell reconstitution dominated by protective B cells might account for sustained responses in a small fraction of RA patients who would not need frequent retreatment. Alternative approaches include the integration of induction and maintenance strategies aimed at creating a favorable functional B cell balance. Thus, remission induction could be attained with profound B cell depletion (plus or minus plasma cell [PC] depletion); by selective depletion of pathogenic B cells (or blockade of pathogenic B cell cytokines); or by preferential expansion of regulatory B cells. The latter two approaches also could be used for maintenance therapy. In addition, maintenance could be accomplished by targeting pathways that promote survival, expansion or migration of pathogenic B cells and PC (BAFF, cytokines, chemokines).


Table 1

Mechanistic overview of current and potential B cell targeted therapies


























































































































Mechanism Target Effect Agents
B cell killing CD20 Depletion (pre-B cell to some PB) Rituximab, others
CD19 Depletion (pro-B cell to some mature PC) Antibodies
B cell killing + inhibition CD22 Depletion (transitional and B cells) Inhibition of activated cells
Depletion (pre-B cell to some PB, GC)
Antibodies
CD79a/b Inhibition of activated cells Antibodies
B cell signaling Syk
PI3 K
B cell inactivation and secondary death Inhibitors
B cell survival BAFF Depletion (transitional, naïve) Belimumab
BAFF+APRIL Depletion (transitional, naïve, SLPC) Atacicept
B cell costimulation and differentiation; inhibition of B cell effector functions TLR B cell inactivation, ↓ M cell homeostasis Inhibitors
TNF ↓ M cells, ↓ PC survival ETN, HMR, IFM, GLM
ICOS ↓ GC, ↓ M cell activation Antibodies
IL-6 ↓ PC generation, inhibit proinflammatory Beff cells Tocilizumab
IL-17 ↓ GC, ↓ autoantibody, ↓ PC generation Antibodies
IL-21 ↓ M cells maintenance, ↓ PC generation Antibodies
B cell homing and organization TNF, LT
LTβR
Inhibit formation—disrupts systemic and ectopic lymphoid architecture and GC ETN, ADM, IFM, GLM
Baminercept
CXCL13/CXCR5 As above; blocks B cell–F TH cell cross-talk, GC organization Antibodies
CXCL12/CXCR4 Disrupts GC organization, ↓ LLPC homing survival Antibodies, inhibitors
CCL10/CXCR3 ↓ Homing of proinflammatory Beff cells and PC Antibodies, inhibitors
CCL20/CCR6 ↓ Homing of CCR6 + Beff cells, ↓ IL-17 stimulation of B cells Antibodies, inhibitors
Breg expansion B cell depletion Promote dominance of Breg upon reconstitution B cell depleting agents
CD40 agonists Induce in vivo expansion of Breg cells Antibodies
GM-CSF/IL-15 Induce in vivo expansion of Breg cells GM-CSF/IL15 Fusokine
Plasma cell killing Proteasome Depletion of PC, activated B cells, ↓ proinflammatory cytokine secretion by Beff cells Proteasome inhibitors
CD38, CD27, others Depletion of PC ATG
BCMA Depletion of LLPC Antibodies
FcγR2b Depletion of PC, inhibit antibody production IVIG, antibodies
Plasma cell inhibition CCL2/CCR2 Suppress antibody production by PC Antibodies, inhibitors

This table summarizes different mechanisms of action whereby different agents could interfere with pathogenic B cell functions or promote protective B cell functions for the treatment of rheumatoid arthritis. For the sake of brevity, only the more prominent and immediate B cell effects of the corresponding interventions are included. To avoid redundancy, with the exception of B cell-depleting and antitumor necrosis factor (TNF) drugs, agents that could work through different mechanisms are mentioned only once. In most cases, the proposed effect of the mechanism in question is derived from mouse or human studies. In a few cases, highlighted in italics, the proposed mechanism is based on the speculative integration by the authors of known biologic effects. Finally, drugs currently used in clinical practice or in advanced stages of development for human autoimmune diseases are mentioned by name. Otherwise, they are generically identified as antibodies or inhibitors (the latter to indicate class agents such as specific enzymatic inhibitors or small molecules).

Abbreviations: ADM, adalimumab; ATG, antithymocyte globulin; Beff, cytokine-producing effector B cells; ETN, etanercept; F TH , follicular T helper cells; GC, germinal center; GLM, golimumab; IFM, infliximab; IL, interleukin; LLPC, long-lived plasma cells; M, memory; PB, plasmablasts; PC, plasma cells; SLPC, short-lived PC.



Fig. 2


The biologic basis for targeting pathogenic B cells in RA. Schematic representation of different molecular and cellular B cell targets. The recruitment of B cells into the pathogenic process could be interrupted at multiple levels including early bone marrow generation, survival, release, and censoring of autoreactive B cells. In the mature, postantigenic compartments, the pathogenic process could be blocked by inhibition of the B cell receptor (BCR) signaling complex or costimulation (the B cell-T H cell cross-talk is illustrated here for naïve B cells and for GC cells; memory cells may share similar pathways, but response to BCR/TLR/BAFF/APRIL and other stimulatory cytokines may differ significantly). GC reactions and their progeny of pathogenic memory and plasma cells can be attenuated either by disrupting microarchitecture (LTβR/CXCL13), migration/organization/duration (CXC13/CXCL12/interleukin [IL]-17/BAFF), or co-stimulation (IL-21). Blockade of different cytokines (IL-21, IL-6, IL-17) would decrease memory survival or differentiation into PC. Synovial generation of GC, memory, and plasma cells could be attenuated by similar interventions, and recruitment of systemic B cells into the inflamed synovium could be blocked by inhibition of chemo attraction by several chemokines including (but not limited to) CXCL12, CXCL13, CCL20, and CCL10. Finally, long-lived PC could be targeted either by depleting agents or by blocking survival factors as indicated in the figure and Table 1 . Pathogenic B cell functions also could be interrupted by depleting antibodies directed against surface antigens. The expression of the most relevant markers is indicated at the top. Also shown is the breadth of B cell populations whose survival depends on either BAFF or APRIL stimulation.




Targeting B cell activation and costimulation


In addition to anti-CD79 antibodies, multiple strategies could be pursued to inhibit B cell activation either through the BCR or costimulatory molecules central to the necessary cross-talk between B cells and T cells or dendritic cells. Potential targets include BCR signaling pathways, inducible T-cell co-stimulator (ICOS), and TLR signaling. The therapeutic potential of BCR inhibition has been suggested by an early clinical trial of the spleen tyrosine kinase (Syk) inhibitor R406 in RA patients. An Src family of nonreceptor kinase, Syk is required for proximal BCR signaling and mediates positive selection of immature B cells, a critical checkpoint for the censoring of autoreactive B cells. In this phase 1 placebo-controlled trial, clinical benefit was obtained as early as 1 week after treatment and was maintained through the 12 weeks duration of the study. Unfortunately, no B cell data were provided by this study, and therefore the actual mechanism of Syk inhibition remains to be clarified since Syk inhibition also interferes with T cell receptor and Fcγ and Fcε receptor signaling, thereby impacting other cells of relevance for the pathogenesis of RA including T cells, mast cells, neutrophils, and osteclast activation. These studies should provide the impetus to pursue the use of other BCR signaling inhibitors acting downstream of Syk such as PI3 K inhibitors and in particular specific PI3 Kδ inhibitors given the critical and nonredundant role played by this isoform in B cell development.


Upon interaction with activated T cells, B cells are themselves activated through several surface molecules including CD40, CD27, and the inducible costimulatory molecule ligand (ICOS-L or B7RP1). These molecules therefore represent attractive candidates for the inhibition of B cells, although a trial of anti-CD40L antibodies had to be discontinued in SLE because of thromboembolism. Targeting the ICOS pathway is of particular interest given the role played by ICOS+ T follicular helper cells (T FH ) in autoimmune B cell germinal center reactions, the expression of ICOS by rheumatoid synovial T cell and CD14+ monocytes, its ability to regulate IL-17 production, and the essential role it plays in collagen-induced arthritis. Anti-ICOS antibodies (AMG557) are under development for autoimmune indications.


Recent data showing that B7-1/2 (CD80/86) also can transmit activation signals to B cells suggest that other members of the B7 family could represent additional targets for inhibitory antibodies. These results also beg the question as to whether the interruption of CD28 signaling by CTLA4-Ig (abatacept, an effective, FDA-approved treatment), also could result in B cell inactivation and disruption of their antigen-presenting cell function. Finally, blocking of CD70-induced CD27 stimulation has been shown to improve mouse collagen-induced arthritis (CIA) disease and provides another example of the therapeutic potential of blocking costimulatory TNF superfamily members that participate in B cell–T cell costimulation.


TLRs represent important pharmacologic targets for the treatment of RA as evidenced by the benefit imparted in this disease by hydroxychloroquine, whose activity is mediated at least in part by relatively weak TLR9 inhibition. In addition to inducing receptor activator for nuclear factor κ B ligand in synovial fibroblasts (TLR2, 3 and 4), TLRs also provide critical costimulation to autoreactive B cells and at least in SLE, TLR3, 7, and 9 regulate the generation of RF as well as anti-DNA and anti-RBP (RNA-binding proteins such as Ro, LA, Sm/RNP). Other potentially important contributions of TLR9 activation of B cells in RA include the production of cytokines such as IL-6 and RANTES/CCL5. Powerful and specific TLR inhibitors, including a combined TLR9/7 inhibitor (DV1079) and TLR7/8/9 antagonists, are under development for RA and SLE. Potential drawbacks of this approach are illustrated by the exacerbation of murine lupus observed in the absence of TLR9 and by the dampening effect of TLR activation of B cells in autoimmune T cell responses.




Targeting B cell activation and costimulation


In addition to anti-CD79 antibodies, multiple strategies could be pursued to inhibit B cell activation either through the BCR or costimulatory molecules central to the necessary cross-talk between B cells and T cells or dendritic cells. Potential targets include BCR signaling pathways, inducible T-cell co-stimulator (ICOS), and TLR signaling. The therapeutic potential of BCR inhibition has been suggested by an early clinical trial of the spleen tyrosine kinase (Syk) inhibitor R406 in RA patients. An Src family of nonreceptor kinase, Syk is required for proximal BCR signaling and mediates positive selection of immature B cells, a critical checkpoint for the censoring of autoreactive B cells. In this phase 1 placebo-controlled trial, clinical benefit was obtained as early as 1 week after treatment and was maintained through the 12 weeks duration of the study. Unfortunately, no B cell data were provided by this study, and therefore the actual mechanism of Syk inhibition remains to be clarified since Syk inhibition also interferes with T cell receptor and Fcγ and Fcε receptor signaling, thereby impacting other cells of relevance for the pathogenesis of RA including T cells, mast cells, neutrophils, and osteclast activation. These studies should provide the impetus to pursue the use of other BCR signaling inhibitors acting downstream of Syk such as PI3 K inhibitors and in particular specific PI3 Kδ inhibitors given the critical and nonredundant role played by this isoform in B cell development.


Upon interaction with activated T cells, B cells are themselves activated through several surface molecules including CD40, CD27, and the inducible costimulatory molecule ligand (ICOS-L or B7RP1). These molecules therefore represent attractive candidates for the inhibition of B cells, although a trial of anti-CD40L antibodies had to be discontinued in SLE because of thromboembolism. Targeting the ICOS pathway is of particular interest given the role played by ICOS+ T follicular helper cells (T FH ) in autoimmune B cell germinal center reactions, the expression of ICOS by rheumatoid synovial T cell and CD14+ monocytes, its ability to regulate IL-17 production, and the essential role it plays in collagen-induced arthritis. Anti-ICOS antibodies (AMG557) are under development for autoimmune indications.


Recent data showing that B7-1/2 (CD80/86) also can transmit activation signals to B cells suggest that other members of the B7 family could represent additional targets for inhibitory antibodies. These results also beg the question as to whether the interruption of CD28 signaling by CTLA4-Ig (abatacept, an effective, FDA-approved treatment), also could result in B cell inactivation and disruption of their antigen-presenting cell function. Finally, blocking of CD70-induced CD27 stimulation has been shown to improve mouse collagen-induced arthritis (CIA) disease and provides another example of the therapeutic potential of blocking costimulatory TNF superfamily members that participate in B cell–T cell costimulation.


TLRs represent important pharmacologic targets for the treatment of RA as evidenced by the benefit imparted in this disease by hydroxychloroquine, whose activity is mediated at least in part by relatively weak TLR9 inhibition. In addition to inducing receptor activator for nuclear factor κ B ligand in synovial fibroblasts (TLR2, 3 and 4), TLRs also provide critical costimulation to autoreactive B cells and at least in SLE, TLR3, 7, and 9 regulate the generation of RF as well as anti-DNA and anti-RBP (RNA-binding proteins such as Ro, LA, Sm/RNP). Other potentially important contributions of TLR9 activation of B cells in RA include the production of cytokines such as IL-6 and RANTES/CCL5. Powerful and specific TLR inhibitors, including a combined TLR9/7 inhibitor (DV1079) and TLR7/8/9 antagonists, are under development for RA and SLE. Potential drawbacks of this approach are illustrated by the exacerbation of murine lupus observed in the absence of TLR9 and by the dampening effect of TLR activation of B cells in autoimmune T cell responses.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Oct 1, 2017 | Posted by in RHEUMATOLOGY | Comments Off on B Cell Therapies for Rheumatoid Arthritis: Beyond B cell Depletion

Full access? Get Clinical Tree

Get Clinical Tree app for offline access