REVIEWS Emerging cell and cytokine targets in rheumatoid arthritis Gerd R. Burmester, Eugen Feist and Thomas Dörner Abstract | Despite major progress in the treatment of rheumatoid arthritis (RA), strong unmet medical need remains, as only a minor proportion of patients reach sustained clinical remission. New approaches are therefore necessary, and include manipulation of regulatory T cells, which might be able to restore the disturbed immune system and could even lead to a cure if this restored regulation were to prove sustainable. Logistical and conceptual problems, however, beset this attractive therapeutic approach, including difficulties with ex vivo expansion of cells, specificity of targeting and the optimal time point of administration. Therefore, alternative avenues are being investigated, such as targeting B‑cell effector functions and newly identified proinflammatory cytokines. On the basis of success with B‑cell depleting therapy using anti-CD20 agents, further treatment modalities are now exploring direct or indirect interference in B‑cell-mediated immunity with the use of agents directed against other B‑cell surface molecules. Novel approaches target intracellular B‑cell signalling and regulatory B cells. New cytokine-directed therapies target important proinflammatory mediators such as GM‑CSF, new members of the IL‑1 family, IL‑6 and its receptor, IL‑17, IL‑20, IL‑21, IL‑23 as well as synovium-specific targets. This article reviews these emerging cell and cytokine targets with special focus on biologic agents, some of which might reach the clinic soon whereas others will require considerable time in development. Nevertheless, these exciting new approaches will considerably enhance our repertoire in the battle against this potentially devastating disease. Burmester, G. R. et al. Nat. Rev. Rheumatol. advance online publication 12 November 2013; doi:10.1038/nrrheum.2013.168
Introduction
Department of Rheumatology and Clinical Immunology, Charité– Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany. Correspondence to: G. R. Burmester gerd.burmester@ charite.de
Rheumatoid arthritis (RA) is a disease of unknown aetiology that is characterized by changes in the synovial tissue, prompted by unknown initiating events, potentially involving infections and tissue injury. The subsequent inflammatory process is reflected by joint pain and swelling as well as systemic manifestations, caused by metabolites of arachidonic acid and various inflammatory cytokines. Abnormalities in the cellular and humoral immune response lead to the occurrence of autoantibodies (rheumatoid factors [RF] and antibodies against citrullinated peptides/proteins [ACPA]) as well as the immigration of T cells and B cells into the synovium. In the effector phase, cartilage is destroyed by invading fibroblasts and the juxta-articular bone by activated osteoclasts.1 Synovial hyperplasia is both a hallmark of RA and the main contributor to the formation of invasive pannus tissue. The observation of T‑cell accumulation in the synovium has led to the hypothesis that a T‑cell dependent inflammatory reaction to an unknown antigen underlies the pathology. This assumption is supported by Competing Interests G. R. Burmester declares associations with the following companies: AbbVie, BMS, MedImmune, MSD, Pfizer, Roche/ Chugai and UCB. E. Feist declares associations with the following companies: BMS, Pfizer, Novartis and Roche/Chugai. T. Dörner declares associations with the following companies: Elli Lilly, NovoNordisk, Roche/Chugai, Takeda, Janssen/J&J, UCB and Sanofi. See the article online for full details of the relationships.
immunogenetic data implicating the HLA class II system,2 by findings derived from animal models,3 as well as by observations of RA disease remission in patients with AIDS;4 furthermore, disease improvement occurs for a substantial proportion of patients after treatment with modulators of the T‑cell co-stimulation pathway, influencing the interaction between antigen presenting cells (APCs) and T cells.5 B cells also have a major pathogenic role starting from (auto)antigen presentation followed by autoantibody production leading to immune complex formation and cytokine release.6 Mechanisms that ought to downregulate the disturbed T‑cell and B‑cell hyper reactivity seem not to operate sufficiently in patients with RA, suggesting defects in the regulatory T (TREG)-cell and B-cell compartments. At the inflammatory site, the synovial lining becomes remarkably thickened due to an invasion of macrophages and the proliferation of resident synovial fibroblasts.7 The degree of synovial hyperplasia correlates with the severity of cartilage erosions resulting in inflammatory pannus formation, which attaches to and invades into joint cartilage, while osteoclast activation leads to parallel bone destruction.8,9 Synoviocytes in this region secrete substantial amounts of matrix degrading enzymes such as collagenase, stromelysin, and gelatinase.10 Thus, at the stage of fully manifested RA, the destructive process is dominated by activated cells of macrophage and fibroblast origin,11 in an environment containing pro-inflammatory cytokines
NATURE REVIEWS | RHEUMATOLOGY
ADVANCE ONLINE PUBLICATION | 1 © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS Key points ■■ Rheumatoid arthritis (RA) is caused by an intricate interplay between T cells, macrophages and B cells; these cells or their products are, therefore, important targets of therapy ■■ T-cell directed approaches target co-stimulation as well as molecules involved in T‑cell activation and regulation (such as CD4 antigen) ■■ Novel therapeutic approaches might include harnessing the action of regulatory T cells ■■ Compounds targeting GM‑CSF and IL‑17 are already in an advanced stage of clinical development in RA ■■ Cytokines such as IL‑20 and IL‑21 contribute to the pathogenesis of RA and have promise as potential targets for treatment ■■ B-cell directed therapies comprise anti-CD20 therapies, CD19-directed and CD52-directed depletion strategies; alternative approaches include CD22mediated ligation of inhibitory B‑cell receptor and modification of adhesion molecule expression by B cells
such as IL‑6 and TNF.12 Accordingly, treatment regimens that selectively antagonize inflammatory cytokine signalling have been successfully applied in patients with RA.13 In summary, therefore, RA is characterized by several stages: initial activation of the (auto)immune system leading to the inflammatory cascade (initiation phase); establishment of chronic pathology (sometimes referred to as ‘chronification’) with perpetuation of inflammatory processes in joints and certain extra-articular sites (transformation phase); and, finally, the destruction of the target tissues resulting in irreversible organ damage (effector phase) (Figure 1). Emerging therapies in RA address molecular and cellular targets in all of these phases. These new approaches are necessary, as despite the remarkable success of modern targeted therapy in RA, a considerable need for new treatment modalities remains. Indeed, in established disease, treatment efficacy is considered on the basis of the ‘ACR 60/40/20 rule’, which decrees the proportion of patients who should see ACR response to therapy rates of 20%, 50% and 70%, denoting improvements from baseline in joint structure and functional as well as laboratory data from patients.14 This rule essentially implies that only 20% of patients will reach a major treatment response threshold (ACR70) with any given biologic treatment. Whereas these response rates are considerably higher in patients with early RA, in patients refractory to antiTNF therapy only 10–15% reach this point.15 Moreover, adherence to treatment with biologic agents is only in the order of 60% over a 1–2-year period, thus frequently necessitating a switch to another form of treatment.16 Finally, patients are increasingly seen in the clinic who have ‘cycled’ through treatments of all available modes of action (that is, conventional DMARDs, anti-TNF agents, IL‑6-inhibitors, T‑cell co-stimulation modulators and B‑cell depletion therapies) and who are thus in desperate need of new treatment modalities. This Review will address some of these emergent new approaches. Another fast-developing approach in RA therapy is the specific intracellular targeting of pathways using small molecules (as reviewed elsewhere in 2013).17,18 These agents are a new generation of drugs that are able to block the function of specific intracellular enzymes such
as Janus kinases (JAKs) or spleen tyrosine kinase (SYK). Thus, although not the focus of this Review, it is note worthy that these compounds interfere with the signalling of various proinflammatory cytokines and, therefore, can in fact be considered as indirect cytokine blockers. 17,18
Targeting the disturbed T‑cell response in RA Implicated cell types A number of cell types have been identified that could be used to (down)regulate a disturbed T‑cell response in RA, foremost the TREG-cell subset, but also regulatory B (BREG) cells, regulatory dendritic cells, suppressive macrophages and CD8+ suppressor T cells.19 Studies in animal models have shown that defects in CD4+CD25+FOXP3+ TREG cells can contribute to the development of autoimmunity that is reversible by adoptive transfer of TREG cells.20 Moreover, studies of TREG cells in human peripheral blood samples, and their ability to suppress T‑cell proliferation, have suggested that this mechanism is also important in human autoimmunity, which is consistent with a defect in T‑cell regulation seen in patients with the IPEX syndrome (immunodysregulation, polyendocrinopathy, enteropathy, X‑linked syndrome), which is caused by loss-of-function mutations of the FOXP3 gene.21 Notably, exacerbation of collagen-induced arthritis (CIA) has been shown subsequent to depletion of TREG cells.22 Conversely, adoptive transfer of TREG cells ameliorated ongoing arthritis in several mouse models of the disease, highlighting the therapeutic potential of TREG cells.23,24 Moreover, TREG cells whose development was induced in vitro from conventional CD4+ T cells (iTREG cells) seemed to be more effective in suppressing osteoclastogenesis and the development of bone erosions in mice with CIA than thymic derived, natural TREG (nTREG) cells.24 Challenges in delineating TREG-cell functions These preclinical data notwithstanding, studying TREG cells and other implicated regulatory cell types in patients with RA represents a major challenge. The cell compartment most accessible to analysis is the peripheral blood, which might not actually be representative of tissue sites ‘where the real action is’. Indeed, homing of regulatory cells to such sites in an attempt to regulate a disturbed immune reaction might lead to an (erroneously) low number of peripheral regulatory cells measured in circulation, or re-circulation of the cells at higher numbers might occur, thus correctly representing extravascular tissue conditions in the blood. Consideration of this uncertainty does not even begin to encompass challenges with regard to different states of disease activity or the influences of therapies including biologic agents. Furthermore, regulatory cells in people are usually identified by the expression of intracellular molecules such as FOXP3 or zinc finger protein Helios,25 which precludes isolation of the entire TREGcell population alive for study of their ex vivo functions. Functional properties, however, are important—effector T cells at certain states of development and TREG cells alike can be CD4+CD25+FOXP3+.26,27 Thus, no uniform finding regarding the proportions and properties of TREG cells can yet be found in the literature.
2 | ADVANCE ONLINE PUBLICATION
www.nature.com/nrrheum © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS a
b Inflammation
Induction Autocrine activation
Local injury, hyperplasia Synovial membrane
Infections
Self-antigen presentation (citrullinated peptides)
Cytokine network (TNF, IL-6, IL-23)
GM-CSF
APC
T cell
Macrophage infiltration Circulating monocytes potentially activated in the bone marrow
Modification of autoantigens incl. citrullination
TH1
TH17
TH17
TH
Impaired TREG cell
B cell
Critical event Activation of APC and subsequent activation of adaptive immune cells
Innate immune activation
c
T-cell and B-cell polyclonal activation Inflammatory cytokine production Synovial invasion Germinal centre and B-cell follicle formation
Adaptive immunity in susceptible hosts
d Destruction
Self perpetuation Cartilage Autoantigens
TEFF
TH17
TREG
TNF IL-6 IL-17
TH17
Membrane antigens
Synovial macrophage
B cell
Collagenase
TH17
B cell
Genotoxic environment Proteases Cytokines
Chondrocyte
RA
TH
Other cartilage Proteoglycans products
Pannus formation
Osteoclast activation
Collagen Autoantibodyinduced activation of osteoclasts
Bone
Figure 1 | Stepwise development of arthritis in RA. a | Induction phase: initial activation of the (auto)immune system leads to an inflammatory cascade. Possible triggers are injuries, infections and exposures to toxic substances (smoking). These events, which involve APCs and the citrullination of relevant proteins, might occur outside of the joints as well as within them. Along with monocyte/macrophage infiltration into the synovium, local synovial cells, notably fibroblasts and macrophages, are activated leading to the secretion of proinflammatory cytokines of both the innate and adaptive immune systems. b | Inflammation phase: self antigens, notably citrullinated proteins, are presented in the context of HLA class II molecules that are characteristic of RA. This presentation leads to polyclonal activation of T cells and B cells, and the formation of germinal-like centres in the synovial tissue. This process is insufficiently controlled by T REG cells. c | Self perpetuation: cartilage autoantigens, which are not normally accessible to the immune system, become exposed by damage and are presented, activating the immune system against cartilage tissue with further infiltration of pannus into the joints resulting in further destruction. d | Destruction phase: synovial fibroblasts and osteoclasts are activated by proinflammatory cytokines such as TNF and IL‑6. Destruction of bone and cartilage ensues. Abbreviations: APC, antigenpresenting cell; GM-CSF, granulocyte-macrophage colony-stimulating factor; RA, rheumatoid arthritis; T EFF, effector T (cell); TH, helper T (cell); TH1, type 1 helper T (cell); TH17, type 17 helper T (cell); TREG, regulatory T (cell).
Studies of TREG-cell numbers in the peripheral blood and tissue of synovial samples from patients with RA have shown that in established disease, CD4+CD25hi cells were comparable in number to control samples,28,29 whereas a moderate decrease of TREG cells was described in untreated patients with early RA.30 One study,31 however, reported an increase in the relative and absolute numbers of CD4+CD25hiFOXP3+ cells in the peripheral blood of patients with RA. Overall in synovial fluid, reports generally agree that the percentage of TREG cells is higher in patients with RA than controls.28–31 Whereas several
studies (published 2003–2006)28–30 of the suppressive function of TREG cells from both the peripheral blood and the synovium found no defects in suppression, a careful 2004 analysis by Ehrenstein et al.32 identified defects of TREG cells of patients with RA related to impaired suppression of IFN-γ and TNF production in co-culture assays. Interestingly, TNF inhibition (using infliximab) in patients with RA led to increased numbers of peripheral TGF-βproducing TREG cells, accompanied by decreased levels of C‑reactive protein (CRP).32 Overall, current data suggest that defects in immune regulation that occur in RA are
NATURE REVIEWS | RHEUMATOLOGY
ADVANCE ONLINE PUBLICATION | 3 © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS Tregalizumab therapy CD4
FOXP3+ TREG cell
TREG cell transfer
Potential gene transfer
CD25
CTLA-4 CD80/86 Tregitope
Inhibitory cytokines IL-2 treatment
DC TCR Induction of IDO
Effector T cell
IL-2 consumption
In vivo
Expansion
In vitro
Figure 2 | Potential interventions using TREG cells in RA. IL‑2 treatment might directly restore and/or initiate the growth and/or function of CD25-positive TREG cells in vivo. TREG cells might be expanded in vitro and potentially be modified through genetic manipulation to target antigens present in the inflamed joint. These cells could then be transferred back to the patient. Another approach would be to specifically activate TREG cells using the anti-CD4 monoclonal antibody tregalizumab. Finally, Tregitopes—peptides derived from human IgG—might be used to specifically activate TREG cells, dampening an autoimmune response. Abbreviations: DC, dendritic cell; IDO, indolamin‑2,3-dioxygenase; RA, rheumatoid arthritis; TCR, T‑cell receptor; TREG, regulatory T (cell).
in Figure 2. These developmental pathways include the use of IL‑2 to enhance the functionality and potentially number of TREG cells, as has been shown in several mouse models of autoimmunity and in patients with graft versus host disease37 and hepatitis C‑associated vasculitis. 38 Another interesting approach is the use of the monoclonal antibody (mAb) tregalizumab (BT‑061), which is claimed to provide an activation signal to naturally occurring TREG cells, but not to conventional T cells, and to bind to a unique epitope of CD4+ cells.39 This antibody has been tested in an early clinical programme in patients with psoriasis and RA, with results sufficiently encouraging to embark on a larger, currently ongoing phase II programme.40,41 Of special interest are ‘Tregitopes’—peptides derived from the Fc region of human IgG sequences. 42 These epitopes reportedly have high affinity in binding to human HLA class II molecules, are conserved across IgG isotypes and induce the expansion of CD4+/CD25hi/ FOXP3+ T cells, suppressing antigen-driven T‑cell activation responsiveness in vitro. It has been hypothesized that Tregitopes are responsible for the immunomodulatory action of polyclonal intravenous immunoglobulin (IVIg) treatment. Once sufficiently characterized and purified, these peptides could be potential therapeutics in many autoimmune situations.42
B-cell targets in RA probably related to defects intrinsic to TREG cells and to the inflammatory environment in the RA synovium.
Challenges in targeting TREG cells Besides incomplete understanding of their functions, another problem in targeting regulatory cell abnormalities in RA is that, unlike in mouse models of the disease, the inciting antigen in patients is not clear among the many antigens recognized by autoantibodies.33,34 Isolation and expansion of highly specific TREG-cell subsets is therefore currently unfeasible, as the inducing antigen (or indeed, antigens plural) has not been clearly defined. Thus, as a compromise, polyspecific TREG cells will have to be employed, presumably with less potency than antigenspecific cells. Alternatively, gene therapy approaches will be required,35 with numerous attendant regulatory issues (such as a current lack of good manufacturing process [GMP] guidelines). Moreover, many of the antigenic specificities recognized by pathogenic autoantibodies can be found years prior to disease onset,36 indicating an early (clinically unnoticed) loss of regulatory capacity that might be difficult to restore after full disease onset. Another important issue is uncertainty regarding the in vivo stability of transferred TREG cells in an inflammatory environment. Therefore, drug development in at least the foreseeable future will necessarily involve ‘simple’ approaches to harnessing TREG cell functions in order to downregulate a disturbed immune reaction. Potential approaches to harness TREG cells An overview of current approaches to potentially harness TREG-cell functions in RA therapeutics is shown
Various avenues to targeting pathogenic B‑cell functions inspired by clinical experiences with rituximab are in development for RA (Table 1), including the use of depleting and non-depleting B‑cell-directed antibodies (including modulation of B‑cell signalling), and potential consideration of BREG cells. Development of mature B cells involves numerous stages that provide rigorous control against autoreactivity. These developmental steps are usually accompanied by changes in the expression of surface markers that might serve as targets of therapeutic antibodies.43 However, the mechanisms that enable autoreactive B cells to escape negative selection and mature into (auto)antibody-producing plasma cells are poorly understood. In RA, underlying tolerance mechanisms are considered to be dysregulated or overwhelmed, enabling the production of characteristic autoantibodies, ACPA and RF. Detection of this ‘serologic signature’ of RA clearly identifies patients with broken tolerance mechanisms (defective immune regulation) within their adaptive immune system.44 Currently, we do not have a clear understanding of how genetic and environmental factors result in the loss of B‑cell self-tolerance, but certain MHC class II alleles (particularly shared epitope alleles) and smoking represent substantial risk factors.45 However, the pathogenic role of ACPA as evidence of ultimate B-cell autoreactivity becomes evident based on recent observations that these autoantibodies have been generated in inflamed synovium via T-cell-mediated activation of B cells carrying somatically mutated BCR gene rearrangements.46 Current direct and indirect surface and extracellular B‑cell targets in RA are summarized in Figure 3.
4 | ADVANCE ONLINE PUBLICATION
www.nature.com/nrrheum © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS Table 1 | Direct targeting of B cells in rheumatic diseases Target; putative and/or intended mechanism of action
Agent
Developmental status*
CD19; B-cell depletion
MDX-1342 (anti-CD19 fully human mAb)
Phase I trial in patients with RA (of single i.v. dose in combination with methotrexate) completed; no results posted121
CD20 type I; B-cell depletion
Rituximab (chimeric anti-CD20 mAb)
Approved for use in RA48,49,66 and ANCA-associated vasculitis50
Ocrelizumab (humanized anti-CD20 mAb)
Discontinued for RA and SLE67,68
Veltuzumab (humanized anti-CD20 mAb)
Phase II study in patients with RA terminated (“trial re-design; no safety issues identified”); no results posted122
Ofatumumab (fully human anti-CD20 mAb)
Phase I/II reported69 Phase II trial in patients with RA ongoing; but currently not recruiting (support progression to phase III)123
CD22; peripheral reduction of B cells, inhibition of B-cell activation (BCR via phosphorylation of SYK and PLCγ2) and proliferation
Epratuzumab (humanized anti-CD22 mAb)
Phase III trials in patients with SLE ongoing and recruiting124,125 Phase II data reported126
CD52; depletion of T cells and B cells
Alemtuzumab (humanized anti-CD52 mAb)
Phase I/II in RA, no further studies72.73
*As of October 2013. Abbreviations: ANCA, anti-neutrophil cytoplasmic antibody; mAb, monoclonal antibody; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus.
Direct targeting using antibodies Subsequent to the successful application of rituximab, an anti-CD20 B‑cell-targeting mAb, in patients with RA, anti-neutrophil cytoplasmic antibody-associated vasculitis and other refractory autoimmune diseases,47–50 B‑cell directed therapies and underlying B‑cell functions have gained renewed interest. Targeting a specific surface molecule to efficiently deplete B‑cell populations is also the basis of anti-CD19 and anti-CD52 strategies (Table 1). Furthermore, a substantially different approach to modu lating B‑cell behaviour also exists: the use of inhibitory receptors (such as the B-cell receptor [BCR], CD22, FcγRIIb) to employ their natural function in order to control immune and B-cell activation. Anti-CD20 mAbs can be differentiated into type I (rituximab-like) and type II (tositumomab-like) subsets, which recognize different domains of the surface molecule and induce different effector-cell profiles.51 CD20 is a prototypic example of a therapeutic target to which binding to different loops of the protein can result in distinct downstream effects. Independently, Fc-mediated functions of anti-CD20 agents define the immunological and clinical consequences of this antibody class. Certain type II anti-CD20 mAbs efficiently induce programmed cell death, in contrast to rituximab-like type I anti-CD20 mAb. The prototypic anti-CD20 mAb, GA101 with non-glycoengineered Fc regions (which is currently in development for patients with malignancies refractory to rituximab, but not in patients with rheumatic diseases), triggers nonapoptotic programmed cell death via actin reorganization and lysosomal degradation with improved clearance of B cells.52 The B‑cell depleting effect of rituximab, a chimeric IgG1 antibody, can also be achieved by humanize d (ocrelizumab, veltuzumab), or fully human (ofatumumab) anti-CD20 mAbs (Table 1). Direct targeting is also possible against other cell-surface antigens
expressed on uniquely on B cells (CD19, CD22) or also expressed by B cells (CD52). Anti-CD20 and antiCD19 antib odies, as well as the anti-CD52 antibody ale mtuzumab—which also targets T cells—bind to the target antigen and deplete the cell by initiating a mixture of apoptosis, c omplement-dependent cyto toxicity, and antibody-dependent cell-mediated cellular cytotoxicity.53 Nevertheless, with the exception of antiCD20 (type I) antibodies and alemtuzumab, clinical experience and ongoing trials of B‑cell-depleting anti bodies in RA are lacking, including of type II anti-CD20 agents and the anti-CD22 antibody epratuzumab; in this section, therefore, B‑cell targets CD19 and CD22 are only briefly introduced. CD19 CD19, a cell-surface regulator of the response to antigens, establishes signalling thresholds that are crucial for B‑cell development and activation.54 CD19 is expressed from the early stages of B‑cell maturation onward, is present on memory B cells, and—although substantially down regulated—is expressed on some plasma cells, which is in contrast to CD20.55–57 Although anti-CD19 antibodies have been mainly investigated in the treatment of lymphoma and leukaemia,58 a phase I trial has been conducted in patients with RA (Table 1). A single-chain bispecific antibody to CD19 and CD3, blinatumomab, can cause lysis of human lymphoma cells in vivo and has exhibited substantial clinical efficacy in early-stage clinical trials in patients with lymphoma.59 Overexpression of CD19 has been linked with development of auto immunity,60 but has not yet been thoroughly investigated in RA. With regard to the expression of CD19, it is noteworthy that a fraction of CD19– bone marrow plasma cells exists that seems likely to be unaffected by antiCD19 antibodies.54 However, the safety aspects of these approaches are not clear yet.
NATURE REVIEWS | RHEUMATOLOGY
ADVANCE ONLINE PUBLICATION | 5 © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS ongoing (Table 1); however, anti-CD22 is not currently under clinical investigation in RA.
IL-21 T cell Anti-CD19
CD40 CD19
Epratuzumab
TCR
MHC II
CD40L
CD22
DC CTLA-4 Abatacept
CD80 and CD86
Rituximab
BAFF
BAFF-R
CD20
Belimumab
B cell
Ocrelizumab
TACI
Tabalumab
CD52 BCMA
Alemtuzumab FcγRIIb
Plasma cell
Atacicept APRIL
Autoantibodies
Immune complexes
Figure 3 | Direct and indirect B‑cell targeting. Direct targets of anti‑B cell therapy comprise surface molecules expressed by B cells or their subsets, such as type I CD20 (target of rituximab and ofatumumab), CD22 (target of epratuzumab, which is not in development for RA), CD19, CD52 (targeted by alemtuzumab, which also cotargets T cells), and receptor–ligand pairs in co-stimulatory pathways, such as CD40–CD40L, CD80/CD86–CTLA‑4 (inhibited by abatacept) and inducible T-cell costimulator (ICOS)–ICOS ligand. FcγRIIb is thought to be an inhibitory receptor that controls B‑cell activation. Inhibiting cytokines involved in B‑cell differentiation and maintenance is an indirect anti‑B cell principle. These molecules include BAFF (specifically targeted by belimumab and tabalumab), APRIL (atacicept simultaneously blocks BAFF and APRIL), IL‑21, IL‑6–IL-6R, IL‑1 and others, which fuel the inflammatory response via activation of other immune cells (T cells, macrophages, neutrophils) resulting into co-activation of B cells and effector functions. Abbreviations: APRIL, a proliferation-inducing ligand (TNF ligand superfamily member 13); BAFF, B-cell activating factor (TNF ligand superfamily member 13B); BAFF‑R, BAFF receptor (TNF receptor superfamily member 13C); BCMA, B-cell maturation protein (TNF receptor superfamily member 17); CTLA‑4; cytotoxic T‑lymphocyte protein 4; DC, dendritic cell; FcγRIIb, Fcγ receptor IIb; TACI, transmembrane activator and CAML interactor (TNF receptor superfamily member 13B); TCR, T‑cell receptor. Adapted from Gregersen, J. W. & Jayne, D. R. W. Nat. Rev. Nephrol. (2012).120 © NPG.
CD22 As we have mentioned, therapeutic antibody binding to some B‑cell-specific cell-surface targets does not induce B‑cell depletion, as exemplified by CD22—a member of the lectin-like immunoglobulin superfamily. CD22 is expressed to a higher extent on naive versus memory B cells,61 and is part of the CD19–CD21–CD22 BCR complex. Its function is to modulate BCR antigen-binding strength and CD19-mediated signal transduction, and it also provides essential survival signals and is involved in B-cell adhesion.62 As with CD20, cell-surface expression of CD22 ceases when mature B cells differentiate into CD22– plasma cells.62 A study of epratazumab function in patients with systemic lupus erythematosus (SLE) suggested that ligation with this anti-CD22 antibody leads to moderate peripheral B‑cell reduction (of approximately 30%), and inhibition of B‑cell proliferation;63 furthermore, it has been shown to substantially reduce levels of BCR downstream signalling kinase activity (SYK and phospholipase C‑γ2 [also known as 1‑phosphatidylinositol 4,5-bisphosphate phosphodiesterase γ‑2]) and intracellular Ca2+.64 Phase III trials in patients with SLE are
CD20 In comparison with CD19 and CD22, a wealth of data on CD20 defines its suitability as a target of immunotherapy, including its specific expression on more than 95% of B cells in blood and lymphoid organs.56 CD20 is not expressed on pro‑B cells or antibody-producing plasma cells.65 Depletion of CD20+ cells, therefore, removes B cells of intermediate stages of development, but does not directly target pro‑B cells and their precursors, or long-lived plasma cells. Among anti-CD20 antibodies, rituximab is at the most advanced clinical stage. Data from phase III double-blind randomized, placebo-controlled, trials (RCTs) and subsequent studies have demonstrated substanstially higher response rates and inhibition of structural joint damage with rituximab, in comparison with placebo, in patients with RA of various disease stages.48,49,66 Ocrelizumab, a humanized anti-CD20 mAb, is no longer in development for use in RA. Although phase III clinical trials have demonstrated efficacy and tolerability of ocrelizumab (SCRIPT, FEATURE, FILM, STAGE), recent analyses of these studies determined that safety risks currently outweigh any benefits of their use for the treatment of RA.67,68 Currently under investigation in RA is ofatumumab. A double-blind, phase I/II RCT in patients with inadequate response to DMARDs reported significantly higher ACR20 response rates at 24 weeks with three doses of ofatumumab (300, 700, or 1,000 mg) compared with placebo (ACR20 rates of 40%, 49%, and 44%, respectively, versus 11% for placebo, P
Introduction
Department of Rheumatology and Clinical Immunology, Charité– Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany. Correspondence to: G. R. Burmester gerd.burmester@ charite.de
Rheumatoid arthritis (RA) is a disease of unknown aetiology that is characterized by changes in the synovial tissue, prompted by unknown initiating events, potentially involving infections and tissue injury. The subsequent inflammatory process is reflected by joint pain and swelling as well as systemic manifestations, caused by metabolites of arachidonic acid and various inflammatory cytokines. Abnormalities in the cellular and humoral immune response lead to the occurrence of autoantibodies (rheumatoid factors [RF] and antibodies against citrullinated peptides/proteins [ACPA]) as well as the immigration of T cells and B cells into the synovium. In the effector phase, cartilage is destroyed by invading fibroblasts and the juxta-articular bone by activated osteoclasts.1 Synovial hyperplasia is both a hallmark of RA and the main contributor to the formation of invasive pannus tissue. The observation of T‑cell accumulation in the synovium has led to the hypothesis that a T‑cell dependent inflammatory reaction to an unknown antigen underlies the pathology. This assumption is supported by Competing Interests G. R. Burmester declares associations with the following companies: AbbVie, BMS, MedImmune, MSD, Pfizer, Roche/ Chugai and UCB. E. Feist declares associations with the following companies: BMS, Pfizer, Novartis and Roche/Chugai. T. Dörner declares associations with the following companies: Elli Lilly, NovoNordisk, Roche/Chugai, Takeda, Janssen/J&J, UCB and Sanofi. See the article online for full details of the relationships.
immunogenetic data implicating the HLA class II system,2 by findings derived from animal models,3 as well as by observations of RA disease remission in patients with AIDS;4 furthermore, disease improvement occurs for a substantial proportion of patients after treatment with modulators of the T‑cell co-stimulation pathway, influencing the interaction between antigen presenting cells (APCs) and T cells.5 B cells also have a major pathogenic role starting from (auto)antigen presentation followed by autoantibody production leading to immune complex formation and cytokine release.6 Mechanisms that ought to downregulate the disturbed T‑cell and B‑cell hyper reactivity seem not to operate sufficiently in patients with RA, suggesting defects in the regulatory T (TREG)-cell and B-cell compartments. At the inflammatory site, the synovial lining becomes remarkably thickened due to an invasion of macrophages and the proliferation of resident synovial fibroblasts.7 The degree of synovial hyperplasia correlates with the severity of cartilage erosions resulting in inflammatory pannus formation, which attaches to and invades into joint cartilage, while osteoclast activation leads to parallel bone destruction.8,9 Synoviocytes in this region secrete substantial amounts of matrix degrading enzymes such as collagenase, stromelysin, and gelatinase.10 Thus, at the stage of fully manifested RA, the destructive process is dominated by activated cells of macrophage and fibroblast origin,11 in an environment containing pro-inflammatory cytokines
NATURE REVIEWS | RHEUMATOLOGY
ADVANCE ONLINE PUBLICATION | 1 © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS Key points ■■ Rheumatoid arthritis (RA) is caused by an intricate interplay between T cells, macrophages and B cells; these cells or their products are, therefore, important targets of therapy ■■ T-cell directed approaches target co-stimulation as well as molecules involved in T‑cell activation and regulation (such as CD4 antigen) ■■ Novel therapeutic approaches might include harnessing the action of regulatory T cells ■■ Compounds targeting GM‑CSF and IL‑17 are already in an advanced stage of clinical development in RA ■■ Cytokines such as IL‑20 and IL‑21 contribute to the pathogenesis of RA and have promise as potential targets for treatment ■■ B-cell directed therapies comprise anti-CD20 therapies, CD19-directed and CD52-directed depletion strategies; alternative approaches include CD22mediated ligation of inhibitory B‑cell receptor and modification of adhesion molecule expression by B cells
such as IL‑6 and TNF.12 Accordingly, treatment regimens that selectively antagonize inflammatory cytokine signalling have been successfully applied in patients with RA.13 In summary, therefore, RA is characterized by several stages: initial activation of the (auto)immune system leading to the inflammatory cascade (initiation phase); establishment of chronic pathology (sometimes referred to as ‘chronification’) with perpetuation of inflammatory processes in joints and certain extra-articular sites (transformation phase); and, finally, the destruction of the target tissues resulting in irreversible organ damage (effector phase) (Figure 1). Emerging therapies in RA address molecular and cellular targets in all of these phases. These new approaches are necessary, as despite the remarkable success of modern targeted therapy in RA, a considerable need for new treatment modalities remains. Indeed, in established disease, treatment efficacy is considered on the basis of the ‘ACR 60/40/20 rule’, which decrees the proportion of patients who should see ACR response to therapy rates of 20%, 50% and 70%, denoting improvements from baseline in joint structure and functional as well as laboratory data from patients.14 This rule essentially implies that only 20% of patients will reach a major treatment response threshold (ACR70) with any given biologic treatment. Whereas these response rates are considerably higher in patients with early RA, in patients refractory to antiTNF therapy only 10–15% reach this point.15 Moreover, adherence to treatment with biologic agents is only in the order of 60% over a 1–2-year period, thus frequently necessitating a switch to another form of treatment.16 Finally, patients are increasingly seen in the clinic who have ‘cycled’ through treatments of all available modes of action (that is, conventional DMARDs, anti-TNF agents, IL‑6-inhibitors, T‑cell co-stimulation modulators and B‑cell depletion therapies) and who are thus in desperate need of new treatment modalities. This Review will address some of these emergent new approaches. Another fast-developing approach in RA therapy is the specific intracellular targeting of pathways using small molecules (as reviewed elsewhere in 2013).17,18 These agents are a new generation of drugs that are able to block the function of specific intracellular enzymes such
as Janus kinases (JAKs) or spleen tyrosine kinase (SYK). Thus, although not the focus of this Review, it is note worthy that these compounds interfere with the signalling of various proinflammatory cytokines and, therefore, can in fact be considered as indirect cytokine blockers. 17,18
Targeting the disturbed T‑cell response in RA Implicated cell types A number of cell types have been identified that could be used to (down)regulate a disturbed T‑cell response in RA, foremost the TREG-cell subset, but also regulatory B (BREG) cells, regulatory dendritic cells, suppressive macrophages and CD8+ suppressor T cells.19 Studies in animal models have shown that defects in CD4+CD25+FOXP3+ TREG cells can contribute to the development of autoimmunity that is reversible by adoptive transfer of TREG cells.20 Moreover, studies of TREG cells in human peripheral blood samples, and their ability to suppress T‑cell proliferation, have suggested that this mechanism is also important in human autoimmunity, which is consistent with a defect in T‑cell regulation seen in patients with the IPEX syndrome (immunodysregulation, polyendocrinopathy, enteropathy, X‑linked syndrome), which is caused by loss-of-function mutations of the FOXP3 gene.21 Notably, exacerbation of collagen-induced arthritis (CIA) has been shown subsequent to depletion of TREG cells.22 Conversely, adoptive transfer of TREG cells ameliorated ongoing arthritis in several mouse models of the disease, highlighting the therapeutic potential of TREG cells.23,24 Moreover, TREG cells whose development was induced in vitro from conventional CD4+ T cells (iTREG cells) seemed to be more effective in suppressing osteoclastogenesis and the development of bone erosions in mice with CIA than thymic derived, natural TREG (nTREG) cells.24 Challenges in delineating TREG-cell functions These preclinical data notwithstanding, studying TREG cells and other implicated regulatory cell types in patients with RA represents a major challenge. The cell compartment most accessible to analysis is the peripheral blood, which might not actually be representative of tissue sites ‘where the real action is’. Indeed, homing of regulatory cells to such sites in an attempt to regulate a disturbed immune reaction might lead to an (erroneously) low number of peripheral regulatory cells measured in circulation, or re-circulation of the cells at higher numbers might occur, thus correctly representing extravascular tissue conditions in the blood. Consideration of this uncertainty does not even begin to encompass challenges with regard to different states of disease activity or the influences of therapies including biologic agents. Furthermore, regulatory cells in people are usually identified by the expression of intracellular molecules such as FOXP3 or zinc finger protein Helios,25 which precludes isolation of the entire TREGcell population alive for study of their ex vivo functions. Functional properties, however, are important—effector T cells at certain states of development and TREG cells alike can be CD4+CD25+FOXP3+.26,27 Thus, no uniform finding regarding the proportions and properties of TREG cells can yet be found in the literature.
2 | ADVANCE ONLINE PUBLICATION
www.nature.com/nrrheum © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS a
b Inflammation
Induction Autocrine activation
Local injury, hyperplasia Synovial membrane
Infections
Self-antigen presentation (citrullinated peptides)
Cytokine network (TNF, IL-6, IL-23)
GM-CSF
APC
T cell
Macrophage infiltration Circulating monocytes potentially activated in the bone marrow
Modification of autoantigens incl. citrullination
TH1
TH17
TH17
TH
Impaired TREG cell
B cell
Critical event Activation of APC and subsequent activation of adaptive immune cells
Innate immune activation
c
T-cell and B-cell polyclonal activation Inflammatory cytokine production Synovial invasion Germinal centre and B-cell follicle formation
Adaptive immunity in susceptible hosts
d Destruction
Self perpetuation Cartilage Autoantigens
TEFF
TH17
TREG
TNF IL-6 IL-17
TH17
Membrane antigens
Synovial macrophage
B cell
Collagenase
TH17
B cell
Genotoxic environment Proteases Cytokines
Chondrocyte
RA
TH
Other cartilage Proteoglycans products
Pannus formation
Osteoclast activation
Collagen Autoantibodyinduced activation of osteoclasts
Bone
Figure 1 | Stepwise development of arthritis in RA. a | Induction phase: initial activation of the (auto)immune system leads to an inflammatory cascade. Possible triggers are injuries, infections and exposures to toxic substances (smoking). These events, which involve APCs and the citrullination of relevant proteins, might occur outside of the joints as well as within them. Along with monocyte/macrophage infiltration into the synovium, local synovial cells, notably fibroblasts and macrophages, are activated leading to the secretion of proinflammatory cytokines of both the innate and adaptive immune systems. b | Inflammation phase: self antigens, notably citrullinated proteins, are presented in the context of HLA class II molecules that are characteristic of RA. This presentation leads to polyclonal activation of T cells and B cells, and the formation of germinal-like centres in the synovial tissue. This process is insufficiently controlled by T REG cells. c | Self perpetuation: cartilage autoantigens, which are not normally accessible to the immune system, become exposed by damage and are presented, activating the immune system against cartilage tissue with further infiltration of pannus into the joints resulting in further destruction. d | Destruction phase: synovial fibroblasts and osteoclasts are activated by proinflammatory cytokines such as TNF and IL‑6. Destruction of bone and cartilage ensues. Abbreviations: APC, antigenpresenting cell; GM-CSF, granulocyte-macrophage colony-stimulating factor; RA, rheumatoid arthritis; T EFF, effector T (cell); TH, helper T (cell); TH1, type 1 helper T (cell); TH17, type 17 helper T (cell); TREG, regulatory T (cell).
Studies of TREG-cell numbers in the peripheral blood and tissue of synovial samples from patients with RA have shown that in established disease, CD4+CD25hi cells were comparable in number to control samples,28,29 whereas a moderate decrease of TREG cells was described in untreated patients with early RA.30 One study,31 however, reported an increase in the relative and absolute numbers of CD4+CD25hiFOXP3+ cells in the peripheral blood of patients with RA. Overall in synovial fluid, reports generally agree that the percentage of TREG cells is higher in patients with RA than controls.28–31 Whereas several
studies (published 2003–2006)28–30 of the suppressive function of TREG cells from both the peripheral blood and the synovium found no defects in suppression, a careful 2004 analysis by Ehrenstein et al.32 identified defects of TREG cells of patients with RA related to impaired suppression of IFN-γ and TNF production in co-culture assays. Interestingly, TNF inhibition (using infliximab) in patients with RA led to increased numbers of peripheral TGF-βproducing TREG cells, accompanied by decreased levels of C‑reactive protein (CRP).32 Overall, current data suggest that defects in immune regulation that occur in RA are
NATURE REVIEWS | RHEUMATOLOGY
ADVANCE ONLINE PUBLICATION | 3 © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS Tregalizumab therapy CD4
FOXP3+ TREG cell
TREG cell transfer
Potential gene transfer
CD25
CTLA-4 CD80/86 Tregitope
Inhibitory cytokines IL-2 treatment
DC TCR Induction of IDO
Effector T cell
IL-2 consumption
In vivo
Expansion
In vitro
Figure 2 | Potential interventions using TREG cells in RA. IL‑2 treatment might directly restore and/or initiate the growth and/or function of CD25-positive TREG cells in vivo. TREG cells might be expanded in vitro and potentially be modified through genetic manipulation to target antigens present in the inflamed joint. These cells could then be transferred back to the patient. Another approach would be to specifically activate TREG cells using the anti-CD4 monoclonal antibody tregalizumab. Finally, Tregitopes—peptides derived from human IgG—might be used to specifically activate TREG cells, dampening an autoimmune response. Abbreviations: DC, dendritic cell; IDO, indolamin‑2,3-dioxygenase; RA, rheumatoid arthritis; TCR, T‑cell receptor; TREG, regulatory T (cell).
in Figure 2. These developmental pathways include the use of IL‑2 to enhance the functionality and potentially number of TREG cells, as has been shown in several mouse models of autoimmunity and in patients with graft versus host disease37 and hepatitis C‑associated vasculitis. 38 Another interesting approach is the use of the monoclonal antibody (mAb) tregalizumab (BT‑061), which is claimed to provide an activation signal to naturally occurring TREG cells, but not to conventional T cells, and to bind to a unique epitope of CD4+ cells.39 This antibody has been tested in an early clinical programme in patients with psoriasis and RA, with results sufficiently encouraging to embark on a larger, currently ongoing phase II programme.40,41 Of special interest are ‘Tregitopes’—peptides derived from the Fc region of human IgG sequences. 42 These epitopes reportedly have high affinity in binding to human HLA class II molecules, are conserved across IgG isotypes and induce the expansion of CD4+/CD25hi/ FOXP3+ T cells, suppressing antigen-driven T‑cell activation responsiveness in vitro. It has been hypothesized that Tregitopes are responsible for the immunomodulatory action of polyclonal intravenous immunoglobulin (IVIg) treatment. Once sufficiently characterized and purified, these peptides could be potential therapeutics in many autoimmune situations.42
B-cell targets in RA probably related to defects intrinsic to TREG cells and to the inflammatory environment in the RA synovium.
Challenges in targeting TREG cells Besides incomplete understanding of their functions, another problem in targeting regulatory cell abnormalities in RA is that, unlike in mouse models of the disease, the inciting antigen in patients is not clear among the many antigens recognized by autoantibodies.33,34 Isolation and expansion of highly specific TREG-cell subsets is therefore currently unfeasible, as the inducing antigen (or indeed, antigens plural) has not been clearly defined. Thus, as a compromise, polyspecific TREG cells will have to be employed, presumably with less potency than antigenspecific cells. Alternatively, gene therapy approaches will be required,35 with numerous attendant regulatory issues (such as a current lack of good manufacturing process [GMP] guidelines). Moreover, many of the antigenic specificities recognized by pathogenic autoantibodies can be found years prior to disease onset,36 indicating an early (clinically unnoticed) loss of regulatory capacity that might be difficult to restore after full disease onset. Another important issue is uncertainty regarding the in vivo stability of transferred TREG cells in an inflammatory environment. Therefore, drug development in at least the foreseeable future will necessarily involve ‘simple’ approaches to harnessing TREG cell functions in order to downregulate a disturbed immune reaction. Potential approaches to harness TREG cells An overview of current approaches to potentially harness TREG-cell functions in RA therapeutics is shown
Various avenues to targeting pathogenic B‑cell functions inspired by clinical experiences with rituximab are in development for RA (Table 1), including the use of depleting and non-depleting B‑cell-directed antibodies (including modulation of B‑cell signalling), and potential consideration of BREG cells. Development of mature B cells involves numerous stages that provide rigorous control against autoreactivity. These developmental steps are usually accompanied by changes in the expression of surface markers that might serve as targets of therapeutic antibodies.43 However, the mechanisms that enable autoreactive B cells to escape negative selection and mature into (auto)antibody-producing plasma cells are poorly understood. In RA, underlying tolerance mechanisms are considered to be dysregulated or overwhelmed, enabling the production of characteristic autoantibodies, ACPA and RF. Detection of this ‘serologic signature’ of RA clearly identifies patients with broken tolerance mechanisms (defective immune regulation) within their adaptive immune system.44 Currently, we do not have a clear understanding of how genetic and environmental factors result in the loss of B‑cell self-tolerance, but certain MHC class II alleles (particularly shared epitope alleles) and smoking represent substantial risk factors.45 However, the pathogenic role of ACPA as evidence of ultimate B-cell autoreactivity becomes evident based on recent observations that these autoantibodies have been generated in inflamed synovium via T-cell-mediated activation of B cells carrying somatically mutated BCR gene rearrangements.46 Current direct and indirect surface and extracellular B‑cell targets in RA are summarized in Figure 3.
4 | ADVANCE ONLINE PUBLICATION
www.nature.com/nrrheum © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS Table 1 | Direct targeting of B cells in rheumatic diseases Target; putative and/or intended mechanism of action
Agent
Developmental status*
CD19; B-cell depletion
MDX-1342 (anti-CD19 fully human mAb)
Phase I trial in patients with RA (of single i.v. dose in combination with methotrexate) completed; no results posted121
CD20 type I; B-cell depletion
Rituximab (chimeric anti-CD20 mAb)
Approved for use in RA48,49,66 and ANCA-associated vasculitis50
Ocrelizumab (humanized anti-CD20 mAb)
Discontinued for RA and SLE67,68
Veltuzumab (humanized anti-CD20 mAb)
Phase II study in patients with RA terminated (“trial re-design; no safety issues identified”); no results posted122
Ofatumumab (fully human anti-CD20 mAb)
Phase I/II reported69 Phase II trial in patients with RA ongoing; but currently not recruiting (support progression to phase III)123
CD22; peripheral reduction of B cells, inhibition of B-cell activation (BCR via phosphorylation of SYK and PLCγ2) and proliferation
Epratuzumab (humanized anti-CD22 mAb)
Phase III trials in patients with SLE ongoing and recruiting124,125 Phase II data reported126
CD52; depletion of T cells and B cells
Alemtuzumab (humanized anti-CD52 mAb)
Phase I/II in RA, no further studies72.73
*As of October 2013. Abbreviations: ANCA, anti-neutrophil cytoplasmic antibody; mAb, monoclonal antibody; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus.
Direct targeting using antibodies Subsequent to the successful application of rituximab, an anti-CD20 B‑cell-targeting mAb, in patients with RA, anti-neutrophil cytoplasmic antibody-associated vasculitis and other refractory autoimmune diseases,47–50 B‑cell directed therapies and underlying B‑cell functions have gained renewed interest. Targeting a specific surface molecule to efficiently deplete B‑cell populations is also the basis of anti-CD19 and anti-CD52 strategies (Table 1). Furthermore, a substantially different approach to modu lating B‑cell behaviour also exists: the use of inhibitory receptors (such as the B-cell receptor [BCR], CD22, FcγRIIb) to employ their natural function in order to control immune and B-cell activation. Anti-CD20 mAbs can be differentiated into type I (rituximab-like) and type II (tositumomab-like) subsets, which recognize different domains of the surface molecule and induce different effector-cell profiles.51 CD20 is a prototypic example of a therapeutic target to which binding to different loops of the protein can result in distinct downstream effects. Independently, Fc-mediated functions of anti-CD20 agents define the immunological and clinical consequences of this antibody class. Certain type II anti-CD20 mAbs efficiently induce programmed cell death, in contrast to rituximab-like type I anti-CD20 mAb. The prototypic anti-CD20 mAb, GA101 with non-glycoengineered Fc regions (which is currently in development for patients with malignancies refractory to rituximab, but not in patients with rheumatic diseases), triggers nonapoptotic programmed cell death via actin reorganization and lysosomal degradation with improved clearance of B cells.52 The B‑cell depleting effect of rituximab, a chimeric IgG1 antibody, can also be achieved by humanize d (ocrelizumab, veltuzumab), or fully human (ofatumumab) anti-CD20 mAbs (Table 1). Direct targeting is also possible against other cell-surface antigens
expressed on uniquely on B cells (CD19, CD22) or also expressed by B cells (CD52). Anti-CD20 and antiCD19 antib odies, as well as the anti-CD52 antibody ale mtuzumab—which also targets T cells—bind to the target antigen and deplete the cell by initiating a mixture of apoptosis, c omplement-dependent cyto toxicity, and antibody-dependent cell-mediated cellular cytotoxicity.53 Nevertheless, with the exception of antiCD20 (type I) antibodies and alemtuzumab, clinical experience and ongoing trials of B‑cell-depleting anti bodies in RA are lacking, including of type II anti-CD20 agents and the anti-CD22 antibody epratuzumab; in this section, therefore, B‑cell targets CD19 and CD22 are only briefly introduced. CD19 CD19, a cell-surface regulator of the response to antigens, establishes signalling thresholds that are crucial for B‑cell development and activation.54 CD19 is expressed from the early stages of B‑cell maturation onward, is present on memory B cells, and—although substantially down regulated—is expressed on some plasma cells, which is in contrast to CD20.55–57 Although anti-CD19 antibodies have been mainly investigated in the treatment of lymphoma and leukaemia,58 a phase I trial has been conducted in patients with RA (Table 1). A single-chain bispecific antibody to CD19 and CD3, blinatumomab, can cause lysis of human lymphoma cells in vivo and has exhibited substantial clinical efficacy in early-stage clinical trials in patients with lymphoma.59 Overexpression of CD19 has been linked with development of auto immunity,60 but has not yet been thoroughly investigated in RA. With regard to the expression of CD19, it is noteworthy that a fraction of CD19– bone marrow plasma cells exists that seems likely to be unaffected by antiCD19 antibodies.54 However, the safety aspects of these approaches are not clear yet.
NATURE REVIEWS | RHEUMATOLOGY
ADVANCE ONLINE PUBLICATION | 5 © 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS ongoing (Table 1); however, anti-CD22 is not currently under clinical investigation in RA.
IL-21 T cell Anti-CD19
CD40 CD19
Epratuzumab
TCR
MHC II
CD40L
CD22
DC CTLA-4 Abatacept
CD80 and CD86
Rituximab
BAFF
BAFF-R
CD20
Belimumab
B cell
Ocrelizumab
TACI
Tabalumab
CD52 BCMA
Alemtuzumab FcγRIIb
Plasma cell
Atacicept APRIL
Autoantibodies
Immune complexes
Figure 3 | Direct and indirect B‑cell targeting. Direct targets of anti‑B cell therapy comprise surface molecules expressed by B cells or their subsets, such as type I CD20 (target of rituximab and ofatumumab), CD22 (target of epratuzumab, which is not in development for RA), CD19, CD52 (targeted by alemtuzumab, which also cotargets T cells), and receptor–ligand pairs in co-stimulatory pathways, such as CD40–CD40L, CD80/CD86–CTLA‑4 (inhibited by abatacept) and inducible T-cell costimulator (ICOS)–ICOS ligand. FcγRIIb is thought to be an inhibitory receptor that controls B‑cell activation. Inhibiting cytokines involved in B‑cell differentiation and maintenance is an indirect anti‑B cell principle. These molecules include BAFF (specifically targeted by belimumab and tabalumab), APRIL (atacicept simultaneously blocks BAFF and APRIL), IL‑21, IL‑6–IL-6R, IL‑1 and others, which fuel the inflammatory response via activation of other immune cells (T cells, macrophages, neutrophils) resulting into co-activation of B cells and effector functions. Abbreviations: APRIL, a proliferation-inducing ligand (TNF ligand superfamily member 13); BAFF, B-cell activating factor (TNF ligand superfamily member 13B); BAFF‑R, BAFF receptor (TNF receptor superfamily member 13C); BCMA, B-cell maturation protein (TNF receptor superfamily member 17); CTLA‑4; cytotoxic T‑lymphocyte protein 4; DC, dendritic cell; FcγRIIb, Fcγ receptor IIb; TACI, transmembrane activator and CAML interactor (TNF receptor superfamily member 13B); TCR, T‑cell receptor. Adapted from Gregersen, J. W. & Jayne, D. R. W. Nat. Rev. Nephrol. (2012).120 © NPG.
CD22 As we have mentioned, therapeutic antibody binding to some B‑cell-specific cell-surface targets does not induce B‑cell depletion, as exemplified by CD22—a member of the lectin-like immunoglobulin superfamily. CD22 is expressed to a higher extent on naive versus memory B cells,61 and is part of the CD19–CD21–CD22 BCR complex. Its function is to modulate BCR antigen-binding strength and CD19-mediated signal transduction, and it also provides essential survival signals and is involved in B-cell adhesion.62 As with CD20, cell-surface expression of CD22 ceases when mature B cells differentiate into CD22– plasma cells.62 A study of epratazumab function in patients with systemic lupus erythematosus (SLE) suggested that ligation with this anti-CD22 antibody leads to moderate peripheral B‑cell reduction (of approximately 30%), and inhibition of B‑cell proliferation;63 furthermore, it has been shown to substantially reduce levels of BCR downstream signalling kinase activity (SYK and phospholipase C‑γ2 [also known as 1‑phosphatidylinositol 4,5-bisphosphate phosphodiesterase γ‑2]) and intracellular Ca2+.64 Phase III trials in patients with SLE are
CD20 In comparison with CD19 and CD22, a wealth of data on CD20 defines its suitability as a target of immunotherapy, including its specific expression on more than 95% of B cells in blood and lymphoid organs.56 CD20 is not expressed on pro‑B cells or antibody-producing plasma cells.65 Depletion of CD20+ cells, therefore, removes B cells of intermediate stages of development, but does not directly target pro‑B cells and their precursors, or long-lived plasma cells. Among anti-CD20 antibodies, rituximab is at the most advanced clinical stage. Data from phase III double-blind randomized, placebo-controlled, trials (RCTs) and subsequent studies have demonstrated substanstially higher response rates and inhibition of structural joint damage with rituximab, in comparison with placebo, in patients with RA of various disease stages.48,49,66 Ocrelizumab, a humanized anti-CD20 mAb, is no longer in development for use in RA. Although phase III clinical trials have demonstrated efficacy and tolerability of ocrelizumab (SCRIPT, FEATURE, FILM, STAGE), recent analyses of these studies determined that safety risks currently outweigh any benefits of their use for the treatment of RA.67,68 Currently under investigation in RA is ofatumumab. A double-blind, phase I/II RCT in patients with inadequate response to DMARDs reported significantly higher ACR20 response rates at 24 weeks with three doses of ofatumumab (300, 700, or 1,000 mg) compared with placebo (ACR20 rates of 40%, 49%, and 44%, respectively, versus 11% for placebo, P
