Introduction

Rheumatoid arthritis (RA) is a chronic inflammatory disease, characterized by a breach of immune regulation in the joint synovial membrane leading to severe damage and destruction of cartilage and bone. This disease does not only cause multiple systemic symptoms, such as fever, anemia, osteoporosis, or muscle weakness but also increases cardiovascular morbidity and risk of some cancers [77, 218]. At the local inflammatory site, namely the joint, disturbances in cytokine and chemokine pathways lead to an immune infiltrate that contributes to the increased proliferation of joint fibroblasts, named synoviocytes, and to the chronicity of inflammation [73, 74, 164].

The critical role of cytokines in many immune processes involved in RA pathogenesis [164] has led to the use of novel therapeutics that target specific cytokines with both biologic agents and small molecules [208, 50]. Even if these biologic agents are currently used as common RA therapies, they are prescribed mostly when conventional therapy has failed [216, 220]. Some of these novel biologic agents are now targeted in the standard treatment of RA patients. Indeed, tumor necrosis factor (TNF) or interleukin-6 (IL-6) receptor inhibitors are well-established RA treatments. An interleukin-1 receptor antagonist (IL-1Ra) also belongs to the Food and Drug Administration (FDA)-approved biologic agents. Other cytokines involved in RA pathogenesis are now the targets of second-generation biologic agents, including IL-17 and GM-CSF inhibitors [50].

There are many cytokines involved in RA pathogenesis. In this review, we focus our interest on a few that are central and targeted in RA treatment (TNF, IL-6, IL-1, IL-17, and GM-CSF). For each cytokine, we will describe its signaling pathway and key functions, then its involvement in RA pathogenesis. Finally, we will focus on current and new inhibitors.

TNF

Signaling pathway and functions

Signaling pathway

TNF is a pleiotropic cytokine that exerts functions in homeostasis and in pathogenesis of several diseases. TNF is mainly produced by monocytes and macrophages and also by some T cells. The versatile bioactivities of TNF come from the binding to and the activation of two distinct receptors, TNF-receptor 1 (TNFR1) and TNF-receptor 2 (TNFR2) [152, 3]. If TNFR1 is ubiquitously expressed, TNFR2 is restricted to several specific cell types, mainly immune cells, endothelial cells, and others as synoviocytes [70]. The signaling of TNF is activated by a TNF homotrimer that binds a homotrimeric TNFR (Fig. 1). The TNFR1 signaling is activated by soluble and transmembrane TNF while the TNFR2 signaling is activated primarily by transmembrane TNF. TNFR1 carries a death domain that recruits the adaptor protein TNFR1-associated death domain protein (TRADD). This leads to the assembly of different complexes and distinct functional outcomes [24]. TNFR1-complex I activates nuclear factor κB (NFκB) and mitogen-activated protein kinases (MAPKs) that induce inflammation, tissue degeneration, host defense, cell proliferation, and cell survival. TNFR1-complex IIa and b activate caspase-8 that induces apoptosis. TNFR1-complex IIc activates mixed lineage kinase domain-like protein (MLKL) that induces inflammation and necroptosis [132]. TNFR2 does not have a death domain and thus is unable to directly induce programmed cell death. TNFR2 recruits TNFR-associated factor 2 (TRAF2) to form TNFR2-complex I. This complex activates MAPKs, NFκB and protein kinase B (PKB also known as AKT) that induce the same functions as TNFR1-complex I.

Fig. 1
figure 1

Signaling pathways of TNF can bind two receptors, TNFR1 and TNFR2, and its signaling requires a homotrimer. TNFR1 recruits the protein TRADD to its death domain, lacking in TNFR2, and this leads to MLKL and caspase 8 activation mediating cell death. The recruitment of TRAF2 leads to the activation of other pathways (MAPK, NFκB, or PKB), mediating the different functions of TNF. IL-6 signaling goes through two ways. Classic signaling is mediated by mIL-6R and trans-signaling by sIL-6R. Both mechanisms lead to the activation of JAK which promotes STAT or Ras-MAPK activation, inducing the IL-6 functions. IL-1 can bind two receptors. IL-1R1 signaling goes through MyD88 that activates IRAK4. Then, IRAK1 which is associated with TRAF6 is activated by IRAK4, leading to the activation of NFκB and AP-1. IL-1R2 acts as a decoy receptor and thereby IL-1 binding does not induce signal transduction. The binding of IL-1Ra on IL-1R also does not induce signal transduction. IL-17 binds a heterodimeric receptor composed by IL-17RA and IL-17RC. Their conserved signaling domain SEFIR recruits the ubiquitin ligase Act-1. In turn, Act-1 recruits TRAF6 leading to the activation of NFκB, MAPK, and PI3K. GM-CSF binds a heterodimeric receptor, composed by two chains, GM-CSFRα and GM-CSFRβ that activates JAK2. This leads to the activation of STAT5 or PI3K, NFκB, and MAPK pathways, promoting GM-CSF functions

Full size image

Biological functions of TNF

One model of TNFR signaling suggests that TNFR1 mainly promotes inflammation and tissue degradation while TNFR2 primarily mediates local homeostatic effects such as cell survival or tissue regeneration [132, 202], giving to TNF its versatile bioactivities.

Homeostatic functions

TNF is a key cytokine in inflammation. An optimal defense against pathogens requires TNF. TNF plays a crucial role in host immunity against Mycobacterium tuberculosis (TB) infection. The control of TB infection requires the formation of granulomas, which are highly organized aggregates of macrophages and Th1 cells, and TNF is mandatory to the formation and maintenance of granulomas [172]. In line with this effect in defense, TNF is also involved in organogenesis and development of lymphoid organ architecture and germinal center formation. This cytokine plays a role in the induction of tissue repair, such as neuronal remyelination, cardiac remodeling, or cartilage regeneration [132]. Furthermore, TNF contributes to the resolution of inflammation by inducing apoptosis of inflammatory cells and could also play a role in the inhibition of tumorigenesis through cytostatic effects such as senescence [224, 21]. As mentioned, the homeostatic potential of TNF is at least partially mediated by TNFR2 [132].

Pathogenic functions

One of the major pathogenic functions of TNF is its inflammatory role associated with its chronic uncontrolled production. It induces the production of pro-inflammatory mediators by many cells. For example, TNF induces IL-6 production by synoviocytes from RA patients [145]. Furthermore, TNF orchestrates tissue recruitment and survival of inflammatory immune cells and promotes tissue destruction [72]. TNF plays also a role in autoimmunity, mainly through the inhibition of regulatory T cells [32]. In addition, TNF has a critical role in cardiovascular diseases. Indeed, TNF is cardiotoxic for healthy myocardium by inducing the apoptosis of cardiomyocytes [104]. During tumorigenesis, depending on the inflammatory environment, TNF is able to switch from tumor-inhibition to induction of genotoxic effects and to promotion of tumor-cell survival and proliferation and tumor metastasis [66]. TNF has been involved in the nociceptive response, which is the physiological process underlying the sensation of pain. Indeed, in arthritic patients, the blockade of TNF reduces the increased nociception before inducing anti-inflammatory effects [110].

Thus, TNF is clearly a major pro-inflammatory cytokine, and its uncontrolled production or function has been linked to the pathogenesis of many different inflammatory diseases, such as RA, inflammatory bowel disease, or psoriasis [5].

TNF and RA

Involvement in the pathogenesis

TNF is clearly a central cytokine in RA pathophysiology [73, 74, 164, 23, 132]. TNF displays a wide variety of effector functions relevant to the pathogenesis of RA and thus has a pro-inflammatory influence at several levels. TNF induces leukocyte and endothelial activation, synoviocyte activation and survival, cytokine and chemokine amplification, angiogenesis, and nociceptor activation [164, 23, 163]. In cultures of synovial cells from RA patients, the blockade of TNF significantly decreases the production of other pro-inflammatory cytokines and chemokines, such as IL-1, IL-6, IL-8, or GM-CSF [105, 22, 31]. Combined with the pro-inflammatory cytokine production, TNF regulates osteoclastogenesis via several mechanisms [192]. TNF is a potent driver of osteoclast formation [142], but it also inhibits osteoblast differentiation and function [20], promoting the imbalance between bone destruction and formation as seen in RA [192]. However, this activity first described in whole bone depends on local cell interactions, and TNF promotes osteogenic differentiation of isolated human mesenchymal stem cell-derived osteoblasts, with a potent synergistic interaction with IL-17 [193]. The results of studies in several arthritis animal models also provide further evidence of the importance of TNF in RA [12]. For example, transgenic mice that express a modified human TNF gene spontaneously develop an inflammatory and destructive peripheral arthritis characterized by many hallmarks of RA (joint inflammation, bone erosion, and cartilage destruction). The treatment of these animals by a monoclonal anti-TNF antibody prevents disease development [134]. In the collagen-induced arthritis (CIA) model, the blockade of TNF by an anti-TNF antibody or by a soluble TNF-receptor fusion protein improves inflammation and joint damage [244, 245]. More importantly, the crucial role of TNF in RA has been confirmed by results from clinical trials in patients treated with TNF inhibitors [68, 67, 177]. Furthermore, clinical inhibition of TNF is associated with the suppression of neovascularization [194]. TNF blockade also reverses the increased nociception long before its effect on inflammation becomes obvious. Indeed, RA patients receiving an intravenous infusion of anti-TNF (infliximab) rapidly show a decreased central nervous system activity linked to nociceptive stimuli. This appears as early as 24 h after its administration, while the anti-inflammatory effects such as the improvement of joint swelling and DAS28 are detected later at 14 days [110]. In addition, the inhibition of TNF in RA patients promotes the emergence of a regulatory T cell population that could suppress effector T cells [178] and also reduce the leukocyte trafficking to the joints [229].

Treatments directed against TNF

Anti-TNF therapy is now fully part of RA treatment. Since 1999, three anti-TNF therapies have been approved: two monoclonal antibodies, infliximab (chimeric human/mouse) and adalimumab (human) and a fusion protein comprising human soluble TNFR linked to Fc component of human IgG1, etanercept. Since 2009, two other anti-TNF treatments have been authorized, certolizumab pegol and golimumab. Although all drugs share their clinical benefit, some differences between them have been observed in terms of rate of response, survival rate, tolerance, including the risk of TB [111, 135]. Currently, biosimilars, defined as highly similar to an already approved biological product, are starting to be approved and available: Inflectra and Remsima for infliximab, Erelzi for etanercept, and Amjevita for adalimumab.

Although anti-TNF treatments have provided a clear improvement in RA care, only two thirds of RA patients respond to anti-TNF leaving space for other treatments [111]. Furthermore, considering the biological role of TNF, adverse events, mainly an increased incidence of infections have been observed. Some studies have showed an increased incidence of some malignancies, but the direct effect of the TNF inhibitors remains controversial [204]. The use of TNF inhibitors is associated with an increased risk of TB reactivation [241]. As mentioned above, TNF plays a key role in the maintenance of granulomas, which is critical for protection from TB. Thus, the blockade of TNF promotes the deregulation of granulomas and the diffusion of TB infection. In addition, RA patients are more sensitive to infections notably to TB. Chronic inflammation induces a defect in the Th1 pathway, which is critical for protection from TB. Furthermore, IL-17 has an inhibitory effect on the Th1 pathway by inhibiting the expression of a functional IL-12 receptor [232]. Despite differences in the risk of TB reactivation between the different TNF inhibitors, this explains in part why the risk of reactivation is maximum at the beginning of TNF inhibition when the disease is still active. In addition, the use of TNF inhibitors has been associated with the development of autoimmunity. In most cases, this is limited to the induction of the production of autoantibodies [13].

IL-6

Signaling pathway and functions

Signaling pathway

IL-6 is a pleiotropic cytokine displaying very diverse activities and produced by almost all stromal and immune cells [120]. IL-6 exerts its biological effects through a fully competent IL-6 receptor (IL-6R), composed by two subunits: a type I cytokine α-receptor subunit (IL-6R, also known as CD126) and a common signal-transducing-β-receptor subunit (gp130, also known as CD130) [205] (Fig. 1). Fully active IL-6R expression is largely restricted to hepatocytes, leukocytes, and megakaryocytes, while gp130 expression is ubiquitous. The function of the IL-6 receptor requires the formation of an IL-6-IL-6R-gp130 complex. This complex is clustered into a dimer structure [217]. IL-6R displays two distinct modes of signaling, classical and trans-signaling.

The classical signaling is relevant only to cells that express the two IL-6R subunits, IL-6R and gp130, and it is mediated by the membrane-bound IL-6R subunit (mIL-6R) [127]. IL-6 binds first mIL-6R that leads to the homodimerization of gp130 and to a high-affinity functional receptor. In contrast, trans-signaling involves the soluble form of the IL-6R subunit (sIL-6R). This sIL-6R binds secreted IL-6 to form a complex displaying a higher circulating half-life than IL-6 alone and promoting its bioavailability [197]. The complex sIL-6R-IL-6 binds gp130 and thus could theoretically stimulate most cells, as gp130 is ubiquitously expressed. For example, synoviocytes express gp130 but not IL-6R and respond to IL-6 only in the presence of sIL-6R [171]. The sIL-6R subunit, which lacks the intracytoplasmic portion of mIL-6R, is produced by the enzymatic cleavage of mIL-6R or by alternative splicing. Furthermore, this trans-signaling is highly regulated by the soluble form of gp130 (sgp130), which is a natural inhibitor of IL-6 signaling. Indeed, exogenous sgp130 inhibits the binding of the complex sIL-6R-IL-6 to the mgp130 [181, 129].

In both mechanisms, the formation of the functional complex leads to signal transduction through gp130. Two pathways mediate this signal transduction, the Janus family tyrosine kinase-signal transducer and activator of transcription (JAK-STAT) pathway and the Ras-MAPK pathway. The negative regulation of IL-6 signaling is mediated by suppressor of cytokine signals (SOCS) [120].

Biological functions of IL-6

In health

IL-6 is an essential cytokine for the regulation of the immune system. In innate immunity, IL-6 modulates key functions including hematopoiesis, accumulation of neutrophils at infection sites through the control of granulopoiesis [151, 48]. In adaptive immunity, IL-6 is critical for immunoregulation. Furthermore, IL-6 plays a role in the regulation of T cells and also on B cells. IL-6 acts at a checkpoint in the differentiation of naïve T cells, promoting the pro-inflammatory Th17 cells and inhibiting the regulatory T cells. IL-6 is involved in the proliferation, survival commitment of T cells and in the modulation of their effector cytokine production [120]. IL-6 also controls the survival, expansion, and maturation of B cells and the production of IgG, IgM, and IgA [5, 120]. The biological role of IL-6 in innate and adaptive immunities is confirmed by an impaired response to viral, parasitic, and bacterial infections during IL-6 deficiency [141, 114, 185]. The central homeostatic processes and immunological outcomes controlled by IL-6 are mainly through the classical signaling [120].

In disease

If the classical signaling is more involved in health, IL-6 trans-signaling is important in disease. In different animal models (colitis, tissue fibrosis, inflammatory arthritis, neuroinflammation, etc.), the trans-signaling plays a key role in the recruitment and apoptosis of leukocytes, in the maintenance of effector T cells, and in the inflammatory activation of stromal tissues [126, 37]. For example, at sites of the disease, the CD4+ T cells lack the IL-6R subunit but they are still responsive to IL-6 through the trans-signaling pathway [125, 120]. Furthermore, in very different diseases, such as cardiac myxoma, a benign heart tumor, in RA, or in Castelman’s disease, an overproduction of IL-6 is detected, partly explaining the inflammatory symptoms [130, 112, 248]. Thus, IL-6 has a protective role in many acute situations such as infections, but its over-production in inflammatory diseases such as RA drives disease progression and chronicity [103].

IL-6 and RA

Involvement in the pathogenesis

IL-6 is a key cytokine in RA pathogenesis and mediates effects rather similar to those of TNF in the synovial environment. IL-6 is involved in various inflammatory effector pathways through activation of immune cells, endothelial cells, synoviocytes, or osteoclasts and production of acute-phase proteins such as CRP [103, 165]. In antigen-induced arthritis (AIA) and CIA, IL-6 plays a key role in the development of the disease [191, 6]. The deletion of IL-6 protects mice from the development of CIA, and the neutralization of IL-6 by an antibody improves the disease [6, 84]. In a monkey CIA model, the blockade of IL-6 by an antibody (tocilizumab (TCZ)) decreases joint swelling and the infiltration of inflammatory cells [234]. Furthermore, an overproduction of IL-6 and of sIL-6R is found in the synovial fluid and blood of RA patients and this excess correlates with disease activity and joint destruction [158, 206]. In cultured cells, IL-6 increases the production of chemokines by several cell types (endothelial cells, mononuclear cells, or synoviocytes) but also the expression of adhesion molecules [225]. This leads to an aggravation of the local inflammation by amplifying the inflammatory cell infiltrate. In the presence of sIL-6R, IL-6 increases the proliferation of synoviocytes and induces RANKL expression that promotes osteoclastogenesis [171, 102] and MMP (Matrix MetalloProteinases) production that contributes to cartilage degradation [226]. Moreover, the crucial proof of concept for a pivotal role for IL-6 in RA pathogenesis is provided by clinical trials in which the anti-IL-6R antibody TCZ suppresses disease activity and erosive progression in RA patients [219, 186]. In addition, first studies targeting the IL-6R demonstrate a significant decreased inflammation and clinical disease activity [187, 49]. Therefore, IL-6 is clearly a key cytokine in RA pathogenesis and thus a well-identified target for RA treatment.

Treatments involving IL-6 targeting

Currently, the anti-IL-6 treatment approved for RA is tocilizumab, which is a recombinant humanized anti-human IL-6R monoclonal antibody [8]. TCZ prevents IL-6 from binding to both mIL-6R and sIL-6R, thereby inhibiting the IL-6-mediated inflammatory effects [211]. In 2010, TCZ was first approved for the treatment of RA patients with active moderate-to-severe disease and an inadequate response to at least one disease-modifying antirheumatic drugs (DMARD) [211]. In 2012, TCZ was approved as a first-line biological agent. Adverse events include infections as for TNF inhibitors and also some more specific to IL-6 inhibition such as colon perforation or liver changes [211].

There are currently four other IL-6 inhibitors under development [165, 230, 227]. Sarilumab is a fully human monoclonal antibody against IL-6Rα, and it has been studied in phase IIb-III clinical trials [96, 119, 81]. Sirukumab is a humanized monoclonal anti-IL-6 antibody in phase II/III clinical trials [221] and phase III [144]. Olokizumab is a humanized anti-IL-6 monoclonal antibody that blocks the final assembly of the IL-6 signaling complex, and it has been investigated in a phase IIb study [95]. Clazakizumab is a humanized monoclonal antibody that potently binds IL-6 and has completed a phase IIb study [166, 242].

IL-1 and IL-1Ra

Signaling pathway and functions

Signaling pathway

IL-1, secreted mainly by macrophages, monocytes, and dendritic cells, was the first interleukin to be identified; it was described as a protein-inducing fever with two major proteins, IL-1α and IL-1β encoded by two distinct genes [64, 85, 90]. IL-1α and IL-1β are synthetized as precursor proteins, pro-IL-1α and pro-IL-1β. Pro-IL-1α is an already biologically active form while pro-IL-1β has no biological activity until it undergoes a proteolytic cleavage by the inflammasome-mediated caspase-1. IL-1α and IL-1β exert similar biological activities through the same receptor IL-1 type 1 receptor (IL-1R1) (Fig. 1). Activated by the ligand binding, the combination with the IL-1R accessory protein (IL-R1AcP) results in a fully active IL-1R1. IL-1R1 binds the adaptor protein myeloid differentiation primary response gene 88 (MyD88) that activates the protein kinase IL-1 receptor-associated kinase (IRAK4). Then IRAK4 phosphorylates and activates IRAK1 that is associated with TRAF6. This induces the oligomerization and the activation of TRAF6 that leads to the activation of the transcription factors NFκB and activator protein 1 (AP-1). In addition to IL-1R1, IL-1α and IL-1β can bind the IL-1 type II receptor (IL-1R2), which acts as a decoy receptor without signal transduction. Another important member of the IL-1 family is the IL-1 receptor antagonist (IL-1Ra). IL-1Ra can bind IL-1R1, but this binding inhibits the IL-1 signaling as IL-1Ra does not attract the IL-1R accessory protein-interacting domain, necessary for signal transduction. Thus, the effects of IL-1 are tightly controlled by the natural inhibitors IL-1R2 and IL-1Ra [85, 90].

Functions of IL-1

Health

IL-1 is a classic endogenous pyrogen, inducing fever, and was first identified as responsible for resistance to infections [237]. In addition, IL-1 modulates the differentiation and the function of innate and adaptive lymphoid cells [90, 215]. IL-1α is released upon cell necrosis and not actively secreted by cells [51]. This implies that IL-1α acts as an alarmin by rapidly mediating the early phases of sterile inflammation [47]. In contrast to necrosis where IL-1α triggers an inflammatory response by inducing pro-inflammatory cytokines such as IL-1β, IL-6, or TNF, during apoptosis IL-1α moves to the nucleus preventing the induction of inflammation [51]. If IL-1α is present in homeostatic conditions, IL-1β is produced upon inflammatory signals. IL-1β has a pro-inflammatory activity at tissue level, leading to vasodilation or activation of innate immune cells such as neutrophils. As such, it is an important component of the host defense against infections [184, 5]. IL-1β also acts on T cells, mainly by supporting the Th17 cell differentiation [2]. IL-1Ra is expressed in almost all tissues. Its major role is to prevent uncontrolled activation of IL-1R1 and thus IL-1β-mediated inflammation [90].

Disease

IL-1 is a key pro-inflammatory cytokine affecting a large target panel of cells like immune cells, endothelial cells, and of organs such as brain or liver. It is a major pathogenic factor of autoinflammation, autoimmunity, or infections. For example, an overproduction of IL-1β characterizes several genetically determined autoinflammatory disorders, such as Familial Mediterranean fever, associated with a mutation of the MEFV gene encoding an inflammasome component or as cryopyrin-associated periodic syndromes (CAPS) linked to a gain-of-function mutation in NLRP3 [222, 207].

At high concentrations, IL-1 is clearly associated with inflammatory pathogenesis and tumorigenesis. IL-1 plays a key role in chronic intestinal inflammation, as it promotes the production of many pro-inflammatory cytokines [63, 153]. Furthermore, high circulating levels of IL-1β are found in patients with infectious or inflammatory conditions [153]. IL-1β also promotes tumor angiogenesis and metastasis of tumors [38, 9]. IL-1 also plays a role in joint inflammation and cartilage degradation by activating osteoclastogenesis [124] and in pathogenesis of heart disease [26]. Thus, IL-1 is a pleiotropic pro-inflammatory cytokine involved in pathogenesis of several diseases, including RA.

IL-1 and RA

Involvement in the pathogenesis

IL-1 plays a major role in RA pathogenesis, mainly in bone and cartilage destruction, by stimulating the production of RANKL, involved in osteoclastogenesis and MMP, involved in cartilage degradation [59]. In RA joints, high levels of IL-1 (IL-1α and IL-1β) are detected in the synovial membrane and fluid [58, 235], with an imbalance between the levels of IL-1Ra and IL-1 [10]. Furthermore, several animal model studies demonstrate the importance of IL-1 in joint inflammation and subsequent tissue damage. The injection of IL-1α or IL-1β into the knee joints of rabbits or rats induces the accumulation of leukocytes in synovial fluid, cartilage degradation, and finally, chronic features of arthritis [107, 71, 46, 198]. The expression of human IL-1β following a gene transfer method leads to a severe and aggressive form of arthritis, similar to some features of human RA [98, 99]. In addition, IL-1 is also found in the inflamed synovium from mice with AIA and CIA [236, 86]. According to their background, IL-1Ra-deficient mice exhibit a more severe form of CIA [157] or develop spontaneously severe destructive arthritis [116]. The administration of IL-1Ra in immune complex-induced arthritis [238] and CIA [128] in mice and AIA in rabbits [11] limits cartilage and bone destruction. Thus, animal models provide a major evidence that IL-1 plays a key role in arthritis, mainly in tissue damage. In humans, several studies using an anti-IL-1 therapy (anakinra) demonstrate a significant but moderate improvement of clinical signs of RA [167, 52]. Anakinra exhibits a protective effect against bone and cartilage damage, but this effect is lower than in animal models.

Treatment

Anakinra is a recombinant IL-1R antagonist. In 2001, it was approved by the FDA for the treatment of moderate-to-severe RA with at least one failed DMARD therapy. Its overall clinical response is moderate and less potent than anti-TNF therapies in most patients [167, 33], possibly linked to its short half-life in plasma [52]. In contrast, IL-1 inhibition is highly effective in some auto-inflammatory diseases. Nevertheless, anakinra treatment has been shown to be safe, with no increased risk of infections in contrast to anti-TNF [25]. Another anti-IL-1 agent, canakinumab which is an anti-IL-1β antibody, has been investigated in a phase II study showing moderate improvement of signs and symptoms related to joint swelling [7]. Gevokizumab, an anti-IL-1β monoclonal antibody has also been investigated in a phase IIa study [91]. In addition, soluble type I IL-1 receptor has been tested with modest effect [65], and rilonacept, another IL-1 inhibitor, showed no effect in RA [162].

IL-17

Signaling pathway and functions

Signaling pathway

IL-17 is a pro-inflammatory cytokine mainly secreted by Th17 and other T cells and which exerts a major role in adaptive immunity. IL-17 binds and acts through a heterodimeric receptor composed of two chains, IL-17RA and IL-17RC [233]. The binding of IL-17 to IL-17RA induces the recruitment of IL-17RC that forms a complex (Fig. 1). Both chains contain a conserved signaling domain, the SEF/IL-17R (SEFIR) [189] which recruits the ubiquitin ligase Act1 through its own SEFIR [87, 203]. In turn, Act1, which also contains a TRAF-binding domain, recruits TRAF6 [150, 203] leading to the activation of the classical NFκB and the MAPK pathway and possibly the phosphoinositide-3 kinase (PI3K) pathway [88]. Such activation induces many inflammatory genes, mainly the neutrophil-specific chemokines [89, 250].

Functions of IL-17

Protective functions

IL-17 plays a key role in adaptive immunity, mainly in host defense against extra-cellular bacterial and fungal infections [140, 123, 176]. In these cases, the major function of IL-17 is neutrophil chemotaxis at the infection site which is critical for the microorganism clearance [121]. This is mediated by the induction of CXC chemokines, including IL-8 and growth factors such as GM-CSF [83]. The synergy of IL-17 with other pro-inflammatory cytokines, including TNF, IL-6, and IL-1, also promotes neutrophil activation. In addition to neutrophil recruitment, IL-17 induces the production of chemo-attractants for other immune cells such as lymphocytes, dendritic cells, or monocytes [176].

Pro-inflammatory functions

Although IL-17 has a crucial role in host defense, an excess of IL-17 is found in many chronic inflammatory and auto-immune diseases. This results in excessive pro-inflammatory cytokine and chemokine secretion, including TNF, IL-6, IL-1, and IL-8 by cells present at the inflammatory site notably monocytes and mesenchymal cells, leading to chronic inflammation. In addition, IL-17 stimulates the production of antimicrobial peptides required for skin protection, and MMP involved in tissue destruction [175, 42]. Thus, IL-17 contributes to the tissue destruction that occurs in chronic inflammatory diseases such as psoriasis, multiple sclerosis (MS), or RA. IL-17 also acts directly on phenotype and cell survival. In RA, IL-17, in synergy with TNF, increases the survival of synoviocytes and contributes to their aggressive and invasive phenotype [231, 117]. IL-17 contributes by different mechanisms to chronic inflammation, and thus, it is a key cytokine in several inflammatory diseases, notably in RA [17].

IL-17 and RA

Role of IL-17 in RA pathogenesis

Several studies in RA patients and in animal models demonstrate the major involvement of IL-17 in RA pathogenesis. IL-17 is spontaneously produced by RA synovium [40, 1, 249], and these increased IL-17 levels correlate with more severe joint damage and disease activity [137, 168]. Through synergy with other cytokines, such as TNF, IL-1, or IL-6, IL-17 leads to synoviocyte activation and cytokine production contributing to chronic inflammation [174, 17] and IL-17 also plays a role in neovascularization [115]. The addition of an anti-IL-17 in RA human synovium or bone explants reduces synovial inflammation and bone resorption [44]. In the CIA model, the intra-articular injection of IL-17 promotes arthritis with symptoms similar to human RA [43, 154] while in IL-17-deficient mice, the disease induction is suppressed [180]. Furthermore, the inhibition of IL-17 by a specific antibody decreases and slows down the arthritis progress [155, 138]. In the SKG model, the arthritis is inhibited in IL-17-deficient mice [113], and the administration of anti-IL-17 antibody decreases its progress [214]. In the streptococcal cell wall (SCW) model, the chronicity of the destructive synovitis requires the IL-17 pathway [156]. Thus, IL-17 participates to the induction and also to the progress of arthritis as a major factor in chronicity [17].

Targeting of IL-17

Secukinumab is a fully human monoclonal antibody against IL-17A. In 2010, a phase I clinical trial showed positive results in RA patients [118]. Current data from phase II trials are not fully convincing. Indeed, in a placebo-controlled phase II study, the difference between the two groups did not reach statistical significance [94]. Nevertheless, a long-term safety and efficacy evaluation in a 52-week follow-up showed an improvement and a sustained response to secukinumab [93]. More recently, various phase III studies have been completed or are still in progress [139]. The results are not as good as in psoriatic arthritis and ankylosing spondylitis where the drug is approved. Another anti-IL-17A monoclonal antibody ixekizumab has reached a similar conclusion [97]. In contrast, a monoclonal antibody directed against IL-17RA, brodalumab had no effect [195, 161], perhaps due to an inhibition of the anti-inflammatory cytokine IL-25, which also binds IL-17RA.

The situation might be more complex because the response to anti-IL-17 treatment displays a high heterogeneity between patients [118, 97, 28]. Some biomarkers such as the association with HLA-DRB1 alleles with clinical responses failed to predict the treatment response [28]. Thereby, there is a need of new biomarkers. A bioassay detecting circulating bioactive IL-17 could be used as a selection biomarker as not all RA patients have an IL-17-driven disease [183].

GM-CSF

Signaling pathway and functions

Signaling pathway

GM-CSF is a secreted cytokine, belonging to the colony-stimulating factor family of hematopoietic growth factors. GM-CSF binds the heterodimeric GM-CSF receptor (GM-CSFR) composed of a specific, low-affinity ligand-binding α-chain (GM-CSFRα), and a signal-transducing β-chain (GM-CSFRβ) which is analogous to gp130 discussed above [78, 101, 243]. The GM-CSF binding to GM-CSFRα induces the assembly of the signaling GM-CSFR complex (Fig. 1). GM-CSFR activation leads to the phosphorylation of JAK2 which is associated to GM-CSFRβ [101]. This initiates the JAK2/STAT5 signaling pathway [108]. The activation of JAK2 also leads to the phosphorylation of tyrosine residues on GM-CSFRβ that mediates the activation of other signaling pathways, including MAPK, PI3K-Akt, and NFκB [109, 108]. Src-like adaptor proteins negatively regulate GM-CSFR signaling by interacting with the GM-CSFRα [147]. SOCS also negatively regulate GM-CSFR signaling by targeting GM-CSFRβ for proteasome-mediated degradation [27].

Functions

Homeostatic functions

GM-CSF is produced by many cell types, including myeloid cells such as monocytes, immune cells such as lymphocytes, or tissue-resident cells such as fibroblasts [243]. GM-CSF is responsible for the proliferation and differentiation of myeloid cells from bone marrow progenitors [170]. Specifically, GM-CSF promotes bone marrow production of neutrophils, eosinophils, and monocytes; induces cell activation; and extends cell survival [170, 169, 57].

GM-CSF is also essential for the regulation of mature myeloid cell functions such as chemotaxis and cell adhesion, of dendritic cell functions, of pro-inflammatory cytokine production, and of phagocytosis [78, 54, 243]. It has a central role in regulating innate immunity notably by the polarization of macrophages into M1 pro-inflammatory phenotype [80, 160, 79, 212]. Thus, GM-CSF is not only a hematopoietic growth factor but is also involved in innate and adaptive immunities [239].

Pathologic functions

Patients with arthritic and inflammatory disorders have an increase of GM-CSF levels and GM-CSFR expression [146, 76, 35, 100, 75]. A high level of GM-CSF is also found in the serum of patients with acute aortic dissection or in the cerebrospinal fluid of patients with active MS [39]. In MS, GM-CSF also promotes the breakdown of the blood-brain barrier and the entry of inflammatory cells [201, 136]. Furthermore, GM-CSF acts as a pain mediator. GM-CSF activates the signal transduction in nociceptors, and leads to pain [209] in tumor and arthritis models, in a prostaglandin-dependent manner [55, 56]. In addition, GM-CSFR is expressed by sensory neurons, and GM-CSFR signaling sensitizes nociceptors in these neurons, leading to hyperalgesia and peripheral nerve remodeling [223].

GM-CSF and RA

Involvement in the pathogenesis

Different studies involve GM-CSF in RA pathogenesis. First, GM-CSF levels are increased in synovial fluid and blood from RA patients [246, 69, 16, 75], and GM-CSFRα is also up-regulated in inflamed synovial tissue and in circulating mononuclear cells [76, 18]. Moreover, the administration of GM-CSF to treat neutropenia in RA patients has been followed by disease flare [196, 106, 60]. In animal models of arthritis, while deficiency or blockade of GM-CSF is protective, the overexpression or injection of GM-CSF promote disease. GM-CSF-deficient mice are protected from CIA development [36] or from methylated BSA/IL-1-induced arthritis [143]. In addition, the blockade of GM-CSF improves arthritis in the CIA model [53], mBSA/IL-1-induced arthritis [247], or in SKG models [214] and reduces inflammation and cartilage degradation in the SCW model [199]. Conversely, the administration of GM-CSF exacerbates arthritis in the CIA model [34]. Furthermore, as GM-CSF acts as a pain mediator, the high levels found in inflammatory synovium could be linked to joint pain, and GM-CSF-deficient mice are protected from this type of pain [55, 56].

Targeting of GM-CSF

Mavrilimumab is a fully human antibody that binds GM-CSFRα and thus inhibits GM-CSF signaling. It has been successfully tested in phase I/II trials with no serious adverse events [29, 30]. Another fully human antibody targeting GM-CSF itself, MOR103, has shown preliminary evidences of efficacy in a phase Ib/IIa clinical trial [15, 239]. Early studies on three other anti-GM-CSF are currently ongoing with lenzilumab, namilumab, and MORAb-022 in phase I [213]. Despite rapid treatment responses to mavrilimumab and MOR103 which provide evidence of clinical efficacy, further clinical investigation is necessary to determine the position of anti-GM-CSF therapy specifically in patients who did not respond to other available DMARDs.

Cytokine interactions and regulation

Interactions between cytokines

During RA pathogenesis, cytokines clearly play a central role. In addition to their own specific action, they act in synergy to increase the inflammatory processes and to maintain their own production (Fig. 2). These cytokines also interact with each other at the production level, as an example, IL-17 increasing IL-1 and TNF production by monocytes [174]. IL-17, IL-1, and TNF stimulate stromal cells to produce IL-6, and the combination of these cytokines leads to synergistic or additive effects [83, 41]. This synergistic effect is also extended to chemokine production such as CCL20, which in turn attracts Th17 cells [45]. The synergy between IL-1 and TNF induces the expression of endothelial adhesion molecules or GM-CSF by synovial cells [243]. Synergistic effects between IL-17, TNF, and IL-1 have been demonstrated in many cell types like synoviocytes, chondrocytes, osteoblasts, or myoblasts [174]. Furthermore, a complementary action of GM-CSF and IL-17 leads to a more severe pathology in experimental arthritis [240] and also further stimulates granulopoiesis and dendritic cell expansion [149]. Thereby, the key cytokines involved in RA pathogenesis, act often in synergy to increase and maintain the inflammation. This synergy could lead to new therapeutics by combining cytokine blockade [200, 19].

Fig. 2
figure 2

Interactions between cells and cytokines during RA pathogenesis. During RA, immune cells migrate to the inflammatory site, where they interact with local synoviocytes. This interaction leads to the production of cytokines that in turn activate immune and mesenchymal cells and promote their own secretion. For example, monocytes secrete IL-6 and IL-1β which promote Th17 differentiation. Th17 cells secrete IL-17 which contributes to chronicity, activating synoviocytes to produce IL-6 and monocytes to produce IL-1 and TNF. IL-1 and TNF act in synergy on synoviocytes to produce IL-6 and GM-CSF and synergize with IL-17 increasing IL-6 production. Synoviocytes secrete IL-6, which exerts a positive feedback contributing to hyperplasia. IL-6 also activates Th17 cells to produce IL-17. In addition, IL-6 acts on osteoblasts to produce RANKL which stimulates osteoclasts, promoting bone destruction. Osteoclasts are also activated by IL-1 and TNF. The stimulation of synoviocytes by the different cytokines leads to the production of chemokines, which recruit leukocytes amplifying the inflammatory response, and the production of MMP promoting cartilage damage. Thereby, cell interactions and cytokines themselves regulate cytokine production and contribute to the pro-inflammatory environment and its maintenance

Full size image

Regulation of cytokine production

As described above, cytokines act in synergy and have a regulatory effect on their production. Furthermore, another key point of regulation at the local site is the cell-cell interactions. During RA, immune cells migrate to the inflammatory site in the synovium and interact with local mesenchymal cells, synoviocytes, which then modulate the synovial cytokine profile [14]. These interactions alone lead to the pro-inflammatory cytokine production, notably IL-6 and IL-1β. The cell contact between activated immune cells and synoviocytes also leads to a high IL-17 secretion that could explain the presence of IL-17 in patient joints [188]. Thereby, cellular interactions are a key point in the induction and maintenance of massive cytokine production.

Another connecting point for cytokines is the JAK/STAT pathway. The activation of STAT proteins by cytokine-receptor binding results in JAK phosphorylation. This pathway is a conserved signaling pathway used by many cytokines, including IL-6 and GM-CSF [159, 190]. In RA patients, abnormalities in JAK/STAT pathway have been observed which correlate with high levels of cytokines, such as IL-6 [122]. Thereby, this pathway has become an attractive therapeutic target. Currently, tofacitinib, a small molecule JAK1/3 inhibitor has been approved by the FDA for the treatment of moderate-to-severe RA and several other JAK inhibitors are undergoing phase II trials [190]. These new therapeutics could represent a valuable addition to the current DMARDs and biologic agents (Fig. 3).

Fig. 3
figure 3

Current and future agents in RA targeting cytokines. Biologic agents target cytokines or their receptors. Infliximab, etanercept, adalimumab, certolizumab, and golimumab bind TNF, thereby inhibiting TNF-TNFR binding. Canakinumab and gevokizumab are monoclonal antibodies under development, and both bind IL-1β. Anakinra inhibits also IL-1 signaling by increased IL-1R occupancy without signaling. IL-17 is targeted by two antibodies under development, secukinumab and ixekizumab. Brodalumab targets IL-17RA and prevents IL-17 binding. Several biologic agents targeting GM-CSF itself are under development: MOR103, lenzilumab, namilumab, and MORAb-022. Mavrilimumab inhibits GM-CSF signaling by acting on GM-CSFRα. The only approved biologic agent inhibiting IL-6 signaling is tocilizumab which binds to mIL-6R and sIL-6R. Sarilumab, under development, also targets IL-6R. Three antibodies targeting directly IL-6 are under development, sirukumab, olokizumab, and clazakizumab. A new generation of small molecules targeting JAK/STAT pathway are under development. One of them, tofacitinib, which inhibits JAK, is now approved in some countries

Full size image

Unresolved aspects

The induction of RA remission is now the therapeutic objective. Despite the improvement of clinical symptoms and joint damage, molecular and cellular heterogeneity between patients impacts the response to therapies. Cytokines are involved in each phase of RA pathogenesis by contributing to autoimmunity, chronic inflammation, and tissue damage [164]. Nevertheless, choosing the right cytokine target for a given patient with a given disease is not easy as their hierarchical relationship is not yet clear. A linear model is proposed where TNF would drive downstream cytokines, including IL-6 or IL-1. This model seems to explain the dominance of TNF in synovial culture systems and thus the dominance of anti-TNF treatments [73]. Although biologic therapies have improved the management and outcome, the patients who achieve disease remission remain a minority [210, 133] and the biological DMARDs induce more adverse events, specifically infections than conventional synthetic DMARDs [204, 148].

Therefore, new strategies to reach higher rates of response and better safety with an optimum risk benefit are needed. In this way, combination therapy or bispecific molecules are options. Indeed, based on their synergistic interactions, bispecific molecules against TNF and IL-17 are under development [19]. The combination therapy remains to be clarified as the first studies showed failures, as for the combination of etanercept and anakinra [92]. Another way is to identify predictive biomarkers of response and tolerance. For example, the presence of autoantibodies against pro-inflammatory cytokines, such as IL-1α or more recently IL-17 has been described as a marker for a good prognostic in RA [131, 173, 182]. In addition, the synovial phenotype could correlate with the response to biologic therapies [62]. Such advance could allow a stratification of patients according to their disease and lead to a personalized treatment for each patient. Several studies have focused on genetic factors associated with treatment response [228]. For example, several single nucleotide polymorphisms in the promoter region of TNF are related to clinical responses to treatment [228, 61, 82]. At this stage, however, these markers are not yet established companion biomarkers.

Conclusion

Targeting of cytokines has been a major progress in RA treatment. Despite such advances made over the last years, many issues remain open. The use of new powerful tools and increased knowledge may lead to a better understanding of RA pathogenesis to select appropriate treatments for each patient and reach remission. Furthermore, little is known about RA patients who achieve a state of remission. Some evidence suggests that early and aggressive treatment is an important factor to reach remission [4, 179]. Most patients who achieve a remission state display less deterioration of function and radiographic progression. Nevertheless, some of them still have radiographic progression despite persistent clinical remission. Thereby, understanding the mechanisms of remission as well of chronicity may provide new means to reach a final cure.