Interleukin-6 in juvenile idiopathic arthritis

Abstract

Juvenile idiopathic arthritis (JIA) is a chronic childhood arthritis. Its pathogenesis is very complicated, with the involvement of not only immune cells but various types of parenchymal cells, and is affected by both genetic and environmental predispositions. The clinical spectrum from inflammation to related conditions is largely mediated by cytokines including interleukin (IL)-6. Fluctuations in IL-6 and its related molecules can modulate the pathogenesis and the clinical presentation positively or negatively. The recent clinical impact of IL-6 blockade on JIA has begun a therapeutic paradigm shift. This review describes the characteristics of JIA, mainly focused on IL-6 with the current therapeutic perspective.

Keywords:

• Interleukin-6
• juvenile idiopathic arthritis
• JIA subtypes
• tocilizumab
• biologic DMARDS

Introduction

Juvenile idiopathic arthritis (JIA) is a disease occurring in childhood characterized by chronic inflammation in joints and multiple organs [1]. Like other inflammatory disorders, the immunologic pathogenesis of JIA is due to a multiphasic dysregulated immune system, affecting both innate and adaptive branches. Cytokines are centrally involved throughout this multistep, multiphasic cascade reaction, shaping this complicated disease state.

Interleukin (IL)-6 is a pleiotropic cytokine that mediates acute phase reactions [2]. It is produced not only by immune cells as a mediator of cell proliferation, differentiation, activation and survival, but by various type of parenchymal cells, such as endothelial cells, keratinocytes, adipocytes, and mesangial cells, as an innate response through pattern recognition receptors. It also modulates the production of acute phase proteins, stimulates collagen production from fibroblasts and exerts vascular endothelial activation and osteoclast differentiation.

IL-6 acts on cells through cell membrane gp130 via binding to membrane-bound IL-6 receptor, which is limitedly engaged in leukocytes and hepatocytes [3]. Alternately, IL-6 can act similarly through soluble type IL-6 receptor and gp130. Since gp130 is ubiquitously expressed, IL-6 has the capacity to act on all cells. IL-6 stimulation with adapter protein binding to gp130 results in activation of the JAK/STAT pathway and of the JAK-SH2-domain-containing protein tyrosine phosphatase-2-mitogen-activated protein kinase pathway, leading to cytokine production. This signal transduction is positively regulated by ADAM 17 causing shedding of membrane protein-type IL-6 receptor (IL-6R) [4] and by splicing variant soluble (s)IL-6R in circulating microvesicles [5] and is negatively regulated by soluble gp130 [6] and SOCS3.

Tocilizumab, an IL-6 inhibitor, has already been adapted for standard JIA management. Clinical findings in tocilizumab-treated patients revealed a part of the IL-6 role that had not been identified by in vitro or animal model studies. In this review, I describe the characteristics of JIA, focused on IL-6 and related molecules, with the current state of IL-6-blocking therapy centered on tocilizumab for JIA management.

Definition and classification of JIA

JIA is defined as arthritis of unknown etiology that begins before the 16th birthday and persists for at least 6 weeks. It includes various disease types (Figure 1), which are classified by the International League of Associations for Rheumatology into seven subtypes based on the clinical findings at disease onset (Table 1): systemic-onset type (soJIA), oligoarthritis type (oligo-JIA), rheumatoid factor-negative polyarthritis type (RF-negative poly-JIA), rheumatoid factor-positive polyarthritis type (RF-positive poly-JIA), psoriatic arthritis type (PsA-JIA), enthesitis-related type (ERA), and other [7]. soJIA is an arthritis presenting with distinct systemic inflammation, like fever, characteristic eruption, and lymphadenopathy. This type is the most likely to combine with macrophage activation syndrome (MAS), which is a lethal complication of JIA. Oligo-JIA, RF-negative poly-JIA, and RF-positive poly-JIA commonly exhibit arthrosynovitis, although they differ in the number and site of affected joints, susceptible age, prognosis, and complications. RF-positive poly-JIA closely resembles the clinical findings of rheumatoid arthritis (RA) in its clinical symptoms, biomarkers, and genetic profiling [8] as well as in its effective managements. PsA-JIA is very similar to juvenile-onset psoriatic arthritis. ERA is a disorder primarily involved in enthesis regions; some cases are HLA B27-positive and are complicated with gastro-intestinal, skin, eye, and/or heart valve disorders. ERA is likely to appear as juvenile-onset spondyloarthritis and ankylosing spondylitis.

Figure 1. Juvenile idiopathic arthritis as a heterogenous disease. Juvenile idiopathic arthritis is an umbrella term of chronic arthritis of unknown etiology. It consists of seven types; systemic-onset arthritis, oligoarthritis, rheumatoid factor-negative polyarthritis, rheumatoid factor-positive polyarthritis, psoriatic arthritis, enthesitis-related arthritis and others. They are basically classified by the presence of arthrosynovitis or enthesitis. They have some common features but different characteristics in pathogenesis, genetic and epigenetic profile, clinical presentation, therapeutic response and comorbidity and prognosis.

Table 1. ILAR classification criteria for juvenile idiopathic arthritis (The edmonton revision 2001).

Category	Inclusions	Exclusions
Systemic-onset arthritis	Arthritis plus more than 2 week's fever and at least one of the following: Evanescent rash. Generalized lymphadenopathy. Enlargement of liver or spleen. Serositis.	(a) (b) (c) (d) (f)
Oligoarthritis	Arthritis in less than 5 joints during the first 6 months. *Arthritis with more than 4 joints after the first 6 months termed as extended Oligoarthritis. Otherwise, termed as persistent Oligoarthritis.	(a) (b) (c) (d) (e) (f)
Polyarthritis, RF-negative	Arthritis with 5 or more joints during the first 6 months: IgM rheumatoid factor negative.	(a) (b) (c) (d) (e) (f)
Polyarthritis, RF-positive	Arthritis with 5 or more joints during the first 6 months: IgM rheumatoid factor positive on at least two occasions at least 3 months apart.	(a) (b) (c) (e) (f)
Psoriatic arthritis	Arthritis plus psoriasis or Arthritis plus at least two of the following: Dactylitis. Nail pits or onycholysis. A positive family history of psoriasis in a first-degree relative.	(b) (c) (d) (e) (f)
Enthesitis-related arthritis	Arthritis plus enthesitis or Arthritis/Enthesitis plus at least two of the following: A positive past/present history of sacroiliac pain or inflammatory back pain. Presence of HLA-B27. A positive family history of ankylosing spondylitis, enthesitis-related arthritis, sacroiliitis with inflammatory bowel disease, reactive arthritis, or acute anterior uveitis in a first-degree relative. Acute symptomatic anterior uveitis. Onset of arthritis in a boy after 6 year of age.	(a) (d) (e) (f)
Undifferentiated	Arthritis fulfilling no, or two or more categories.	none

Juvenile idiopathic arthritis is defined as arthritis of unknown etiology that begins before the 16^th birthday and persists for at least 6 weeks; other known conditions are excluded. Exclusion items: (a) Psoriasis or a positive history of psoriasis in the patient or a first-degree relative. (b) An HLA-B27 positive male beginning after the sixth birthday. (c) Ankylosing spondylitis, enthesitis-related arthritis, sacroiliitis with inflammatory bowel disease, reactive arthritis, or acute anterior uveitis, or a positive history of one of these disorders in a first-degree relative. (d) IgM rheumatoid factor on at least two occasions at least 3 months apart. (e) The presence of systemic-onset juvenile idiopathic arthritis in the patient. (f) Arthritis fulfilling the two categories.

IL-6 production and its pathologic role in JIA

IL-6 protein and transcript are detectable in the blood [9–11] and tissues [12] in all JIA subtypes. In soJIA, IL-6 forms a complex with sIL-6R in blood [13]. The IL-6 concentration in synovial fluid is much higher than that in blood [14]. It increases markedly before treatment and correlates with CRP, iron, hemoglobin, and platelet levels. It also correlates with the number of affected joints and the degree of disability [15,16]. A positive feedback loop via the epiregulin-ErbB1 pathway [17] is known to act as an IL-6 amplifier in RA joints [18]. Given that the ErbB-signaling pathway is activated in all types of JIA [19] as well as in RA, an amplified IL-6 system is expected to be locally involved in the pathogenesis.

IL-6 induces inflammatory cell differentiation and activation, osteoclast activation and periarticular inflammation in collaboration with other pro-inflammatory cytokines. In arthrosynovitis, IL-6 acts on monocytes in the peripheral blood and synovial fluid to markedly amplify inflammatory responses and to produce various cytokines and chemokines [20]. Meanwhile, IL-6-expressing cells are poorly detected within entheses or subchondral bone lesions in spondyloarthritis or ankylosing spondylitis [21]. These indicate that the pathological role of IL-6 in PsA-JIA and ERA is totally distinct from that in oligo/poly-JIA; this is the likely cause of the poor efficacy of IL-6 inhibitors in PsA-JIA and ERA. However, in early spondyloarthritis, IL-6 expression is detected in the connective tissue region of the sacroiliac joint [22], suggesting the role of cytokines may differ from phase to phase. Additionally, innate immune activation, mainly by IL-6 signaling, is key to PsA development in psoriasis patients [23]. Indeed, annexin A2, which is directly involved in the ossification of ligament fibroblasts, is upregulated by IL-6 [24]. Thus, IL-6 is expected to be differently involved in each JIA disease type.

IL-6 has implicated in both adaptive immunity and innate immunity. In oligo/poly-JIA, Th17 cells are considered as a major pathogenic driver, which could be differentiated by IL-6 instead of IL-23. Foxp3⁺CD4⁺ T cells in synovial fluid have an IL-6-producing ability [25] with antigen-specific inhibitory manner, whereas anti-IL-6 antibody could improve the inhibitory function of regulatory T cells in oligo-JIA [26]. In soJIA, NK cell dysfunction [27] is noted with a decreased expression of perforin and granzyme B in NK cells, similarly observed in IL-6 transgenic mice [28]. The involvement of IL-6 is supported by the clinical finding that treatment with IL-6 inhibitor restored NK cytotoxic activity in JIA patients [29]. In monocytes, IL-6 signaling is inhibited with reduced expression of membranous IL-6R and gp130, mainly due to suppression of p38 MAPK signaling by IL-1β stimuli [30]. Besides, IL-6 modifies cytokine profile in JIA synoviocytes with pro-inflammatory and anti-inflammatory manners [31].

JIA has genetic susceptible loci common to each type, suggesting the existence of a potential ‘pan-JIA’ gene [32] with subtype-dependent individual expression profiles and genetic architecture [33]. In oligo/poly-JIA, recent investigations using the Immunochip genotyping array revealed an ATP8B2-IL6R lesion at chromosome 1 with genomewide significance and an IL-6 lesion at chromosome 7 as highly associated loci [34]. In IL-6 promotor polymorphisms, Woo and colleagues found the proportion of soJIA patients with a − 174 C allele was smaller than that of other groups, especially in cases with patients under 5 years of age [35]. Meanwhile, in Iranian oligo/poly-JIA patients a significant association between the −174 G allele and disease susceptibility was demonstrated [36]. Thus, genetics information along with epigenetic findings will disclose the pathogenic significance of IL-6 signaling molecules in JIA.

Treatment

The therapeutic goal of JIA is an abrogation of inflammation and prevention of comorbidity [37]; the strategy is similar to the ‘treat-to-target’ strategy used for RA. IL-6 blockade therapy should be considered when IL-6 is a potential target. The treat-to-target strategy also recommends the adjustment of treatment along with the moving target. If the target has been changed from TNF to IL-6, then TNF inhibitor treatment should be switched to IL-6 inhibitor treatment. Continuous monitoring is critically important for better treatment to achieve cure, remission or low disease activity. As evidenced by the evolution of the mRNA and miRNA expression of neutrophils in oligo/poly-JIA [38], a cure or healing never mean reverting to the initial state unlike in infectious diseases. Instead, the system is rewritten afresh and the molecular parts are rewired with a novel design.

IL-6 inhibitors

Despite the relatively low association between JIA and IL-6 from the etiologic and genetic perspective [39], IL-6 inhibitors show drastic clinical effectiveness that has resulted in a crucial paradigm shift in JIA management [40], distinct from TNF inhibitors. However, the nature of the paradigm shift differs among disease types. In soJIA, the prototype IL-6 inhibitor tocilizumab improves both the mortality prognosis and the quality of life and, consequently, it is now the first-choice biologic disease-modifying antirheumatic drug (DMARD) in conjunction with IL-1 inhibitors. Even in oligo/poly JIA, the obstructing effect of IL-6 inhibition on joint destruction is clearly recognized in tocilizumab-treated patients; therefore, this treatment is on par with TNF inhibitors for treating JIA. Although tocilizumab was remarkably effective in a few psoriatic arthritis and spondyloarthritis/ankylosing spondylitis cases [41–43], most were early-phase so additional studies are needed to confirm its broader utility. In randomized placebo-controlled trials for adult psoriatic arthritis patients, tocilizumab did not show any effectiveness [44,45]. Thus, IL-6 inhibitors should not be considered as a standard therapy for PsA-JIA and ERA. However, a recent study using clazakizumab, another IL-6 inhibitor, showed effectiveness in psoriatic arthritis [46], indicating IL-6 blockade therapy might be an optional treatment for PsA-JIA in the future [47].

Currently, IL-6 inhibitors are commercially available as an anti-IL-6R antibody, such as tocilizumab (ACTEMRA or RoACTEMRA) or as an anti-IL-6 antibody, such as sarilumab (KEVZARA) and siltuximab (SYLVANT). Three other anti-IL-6 antibodies, sirukumab, olokizumab, and clazakizumab, are currently in clinical trials. Only tocilizumab is approved as a therapeutic agent for soJIA and for JIA with chronic active arthritis refractory to one or more DMARDs. Tocilizumab is an antagonistic, humanized, monoclonal antibody against both soluble and membrane-bound IL-6R. A half-life extension formulation of tocilizumab, SA237, is being tested in a clinical trial for RA but not for JIA. Currently, sarilumab [48] is under investigation in phase-II clinical trials for use in soJIA (ClinicalTrials.gov NCT 02991469) and polyJIA (ClinicalTrials.gov NCT 02776735). The treatment guidelines for soJIA by ACR first published in 2011 recommended tocilizumab as the second line drug after failure of glucocorticoid or anakinra because of its limited availability [49]. In the 2013 update [50], tocilizumab was described as a treatment tool for active arthritis; it was recommended as the second line drug after failure of glucocorticoid or anakinra in patients with active systemic features and was placed in the same position as TNF inhibitors and abatacept for cases without active systemic features and refractory to aggressive agents, such as glucocorticoid, MTX, leflunomide, or anakinra. Key milestones in the development of IL-6 inhibitors for JIA are shown in Figure 2.

Figure 2. Key milestones in the development of IL-6 inhibitors for juvenile idiopathic arthritis. JIA: juvenile idiopathic arthritis; soJIA: systemic onset JIA; poly-JIA: polyarticular JIA; FDA: US Food and Drug Administration; EMA: European Medicines Agency; ACR: American College of Rheumatology; TCZ: tocilizumab; TCZ-SQ: TCZ with subcutaneous injection; SARIL: sarilumab.

Tocilizumab

Tocilizumab was first examined in soJIA, then in poly-JIA. The preliminary results for soJIA were opened in 2005 [51,52], and later confirmed by two randomized, double-blind placebo-controlled trials conducted in Japan [53] and Europe and USA [54]. For poly-JIA, a preliminary tocilizumab study using 19 patients was conducted in Japan [55]. The effectiveness of tocilizumab was confirmed by the CHERISH trial [56], a randomized, double-blind placebo-controlled trial in Europe and the USA.

Phase II and III studies for soJIA

Yokota and colleagues conducted a phase-II tocilizumab escalating-dose trial on 11 soJIA patients who had received high-dose, long-term corticosteroids [51]. All patients achieved CRP-negative at 2 weeks after three doses of tocilizumab infusion. Ten of 11 patients achieved an improvement from baseline of ≥30% or ≥50% in at least three of the six core JIA criteria [57] (ACR Pedi 30 and 50, respectively), and seven patients achieved ACR Pedi 70. There was no study cessation due to adverse events. Meanwhile, Woo and colleagues also obtained a similar result with an open-label trial of single, ascending doses of tocilizumab in 18 Caucasian children with active soJIA [52]. Eleven of 18 patients completed the study achieving at least ACR Pedi 30, and eight of them achieved ACR Pedi 50 within 8 weeks. No severe adverse events were noted, and no one discontinued the treatment because of an adverse event. Both preliminary studies concluded that 8 mg/kg tocilizumab can be safely infused at two-week intervals with efficacy in soJIA.

Subsequently, as a phase-III study, a multicenter double-blind placebo-controlled, withdrawal study on 43 soJIA patients who achieved ACR Pedi 30 response by tocilizumab in the open-label lead-in phase was conducted in Japan [53]. Eighty percent of the tocilizumab-treated group sustained negative CRP value with ACR Pedi 30 response at the end of 48 weeks, compared with 17% of the placebo-controlled group. The sustained period of remission after lead-in phase was significantly longer in the tocilizumab-treated group than in the placebo-controlled group. Meanwhile, DeBenedetti and colleagues reached the same conclusion with a double-blind, placebo-controlled study using 112 soJIA patients in Europe, Australia, and the USA [54]. The tocilizumab-treated group showed a clear improvement in systematic symptoms and ACR Pedi score compared with the placebo-treated group at both week 12 (the primary endpoint) and week 52 (the end of this study). A greater glucocorticoid tapering effect was also observed in the tocilizumab-treated group. Based on these data, tocilizumab was approved in Japan, Europe and the USA for the treatment of soJIA in patients who inadequately responded to or became intolerant of previous therapy with one or more DMARD.

In succession to the phase-II and phase-III trials in Japan, an extension study was conducted to evaluate the efficacy and safety of long-term biweekly tocilizumab infusion [58]. With an average observation period of 3.4 years, 58 of the 67 patients remained on continuous tocilizumab treatment. At the final observation, >80% patients showed ≥50% improvement in ACR Pedi score, 61% showed 90% improvement, and 33% of patients were free from glucocorticoid therapy. This long-term follow-up study clearly indicates that tocilizumab is safe, well-tolerated, and has potential for maintaining long-term soJIA remission.

Preliminary and phase-III study for polyarticular-course JIA

The effectiveness of tocilizumab for polyarticular-course JIA was first revealed in Japan from an open label, multicenter study with a single-arm setting [55]. Eighteen of 19 patients who presented active arthritis refractory or intolerant to other medications including MTX achieved a > 50% improvement in ACR Pedi score after the initial 12 weeks by tocilizumab infusion every four weeks. ACR Pedi 70 achievement was 94.1% at week 24 and ACR Pedi 90 achievement steadily increased from 10.5% at week 12 to 64.7% at week 48. DAS28 was also improved from 5.8 at baseline to 1.7 at the end of the study. Seventeen of 19 patients completed the full 48 weeks of tocilizumab treatment without severe adverse events.

A phase-III, randomized, international, double-blind withdrawal trial termed CHERISH study on 166 patients who achieved ACR Pedi 30 response by tocilizumab in the open-label lead-in phase was conducted by the Paediatric Rheumatology International Trials Organization [56]. During the 40-week double-blind period, flare was found in 48.1% of the placebo-treated group compared with in 25.6% of the tocilizumab-treated group. At the endpoint of week 40, 64.6% of patients in the tocilizumab-treated group achieved >70% improvement and 45.1% of patients achieved 90% improvement in their ACR Pedi score. Serious adverse event was 11.1 per 100 person-years.

A long-term extension study of the original CHERISH trial was conducted in Poland and Russia [59]. Forty-one patients who had achieved ACR Pedi 70 without adverse events at the endpoint of the original CHERISH study were observed for another 27–89 weeks, receiving 8 mg/kg tocilizumab every 4 weeks. At 24 weeks after entering this extension study, the ACR Pedi 90 response rate reached at 80.5% from 73.2% at baseline. The inactive disease rate reached 76% from 63% at baseline. Notably, 48% of patients achieved clinical remission with tocilizumab. Severe adverse events occurred at a rate of 6.46/100 person-years. Only one case discontinued tocilizumab infusion due to severely decreased neutrophil count as a drug-related adverse event.

Practical use of tocilizumab for soJIA

According to post-marketing surveillance in Japan [60], fever and skin rash disappeared in 94% of patients and CRP was negative in 99% of patients by 52 weeks after tocilizumab induction. According to the German AID registry [61], tocilizumab had a clinical response in 35% of soJIA patients within the first 12 weeks and in 52% of patients after >1 year of tocilizumab usage. Of these, 75% of patients achieved remission or inactive stage, compared with the 9% of patients categorized as nonresponders. At the final observation point, tocilizumab treatment without any agent was administered in only 11% of patients; around 40% of patients needed regular glucocorticoid treatment.

There is no clear evidence on the ideal timing of tocilizumab introduction. A retrospective observational study of tocilizumab for soJIA conducted in Thailand found that early intervention was significantly superior to later intervention in reducing disease activity, increasing remission rate, improving joint function, allowing steroid withdrawal, and achieving growth [62]. Whether the optimal timing is 6 months or earlier remains unknown, but early control of IL-6 signaling appears to be important for improving soJIA prognosis.

Intravenous infusion of 8 mg/kg tocilizumab at two-week intervals is the basic protocol for treating soJIA. However, a pharmacokinetic study indicated that 12 mg/kg is optimal for patients weighing <30 kg [63]. This is clinically supported by the preliminary result that a 12 mg/kg infusion was more effective than an 8 mg/kg infusion for patients weighing <30 kg [54]. In polyarticular course JIA, 8 mg/kg tocilizumab at four-week intervals is the standard because adverse events like infections were increased with 10 mg/kg tocilizumab treatment.

Radiographic improvements in soft tissue swelling, osteoporosis, epiphyseal irregularity, subchondral bone cysts, joint space narrowing, bone erosion, and growth abnormality were reported in tocilizumab-treated soJIA patients [64]. The Larsen score was also improved, but several cases showed progressive, persistent joint destruction [65]. A retrospective observation study found tocilizumab had a risk of progressive joint destruction in patients who retained abnormal MMP-3 values even after achieving clinical remission and in cases with an older age of onset [66]. Meanwhile, one radiological repaired case following tocilizumab treatment was reported in poly-JIA [67].

There is little information on the tapering and withdrawal of tocilizumab for JIA treatment. A retrospective study of tocilizumab for soJIA reported that four-week intervals of tocilizumab therapy could be available in cases with milder clinical courses that present in the inactive stage with >30% improvement in the ACR Pedi score just 2 weeks after the first tocilizumab infusion [68]. Using this tapering method, seven of 25 cases finally terminated tocilizumab therapy without increasing glucocorticoids, suggesting that in certain types of soJIA, tocilizumab can be administered with longer intervals and eventually be discontinued. An extensive analysis revealed that, with four-week tocilizumab therapy intervals, successful cases have distinct baseline characteristics in laboratory variables, such as white blood cell count and levels of hemoglobin, total protein, albumin, CRP, ESR, ferritin, and LDH, compared with unsuccessful cases. This analysis also revealed differences in the MAS prevalence and in clinical findings, such as hepatomegaly and splenomegaly, as well as in the reduction rate of peripheral granulocyte count at 2 weeks post-tocilizumab induction. Furthermore, several baseline status and follow-up data were identified as factors predicting achievement of tocilizumab-off remission, referring to the condition without fever, lymphadenopathy, CNS involvement, or abnormal coagulation test results at baseline in addition to a lack of history of MAS and persistent low disease activity in clinical course [69].

Safety

Tocilizumab-related adverse events in JIA patients are similar to those in RA patients. Infections, for example, respiratory tract infections, and laboratory data abnormalities, for example, altered liver function, neutropenia, and elevated cholesterol levels, are detected but are generally milder than those in RA. These events are much more frequent in soJIA than in oligo/poly JIA, probably due to the natures of the disease and medications such as glucocorticoid.

The most common adverse event in tocilizumab-treated JIA patients is infection, most frequently pneumonia [70]. As in adults, glucocorticoid use in JIA raises the infection risk; however, TNF inhibitor use also increases this risk according to an observation study conducted over the ten years since the adaptation of biologic DMARDs [71]. There is not yet any comparable data about tocilizumab.

The high risk of opportunistic infection such as tuberculosis is well known in patients receiving biologics. In JIA, opportunistic infections during tocilizumab therapy are extremely infrequent. No new tuberculosis-infected JIA cases have been reported in any trials or surveys [72]. Compared with abatacept, etanercept, and adalimumab, tocilizumab has the highest infection risk [73]; meanwhile, there was no difference in infection risk between tocilizumab and IL-1 inhibitor, canakinumab, or anakinra in soJIA [74].

There are many reports about viral infections in rheumatic disease patients. Varicella zoster virus (VZV) is a problematic infectious agent for RA. In JIA, VZV infection has been reported in registry data in Germany [75] and the USA. The VZV incidence ratio in JIA is significantly higher than that in healthy people, and the VZV risk may be high particularly when etanercept is used [76]. VZV infection has also been reported for tocilizumab treatment; although it is unknown whether the IRR is high, VZV infections in JIA rarely become severe. Epstein–Barr virus (EBV), which is strongly relevant in MTX-treated RA, typically has no change in viral load following tocilizumab administration [77], although an exceptional case was reported with deterioration of chronic active EBV infection. There is no evidence concerning the occurrence of lymphoma due to EBV because only one case with Hodgkin lymphoma was reported in the German AID registry. For flu infection, mild symptoms, such as low-grade fever and a shortened febrile period, were noticed in tocilizumab-treated patients [78].

The gastrointestinal perforation observed in adults with rheumatic diseases is extremely rare in children, and only one patient with HLA-B27-positive ERA was reported in Germany [79].

A reduction in blood cell count, specifically neutropenia and thrombocytopenia, is a dose-dependent adverse event of tocilizumab treatment, frequently observed for 12 mg/kg treatment. As it is a transient, reversible change, tocilizumab can be resumed after a short break until recovery to normal levels [80]. Mostly, mild cytopenia is acceptable and not considered to be an immune-compromised condition because of the low increased risk of infection. Transient elevations of alanine aminotransferase, aspartate aminotransferase, total cholesterol, LDL cholesterol, and triglyceride levels are frequently observed.

Hypersensitivity reaction is frequently observed at young ages, particularly at the third tocilizumab dose. A high soluble TNF-R II value before the third dose is a risk for this hypersensitivity reaction in soJIA [81]. There have been a few cases with anaphylaxis. Hypersensitivity reaction is not a common reason for termination of tocilizumab therapy. Interestingly, a successful case of desensitization with recovery in adulthood of the lost effect on arthritis was reported in poly JIA [82]. Anti-tocilizumab antibody development is not a rare event. In a soJIA open-label extension study of 56 patients in Japan, 3 and 5 patients were anti-tocilizumab antibody-positive by week 48 and week 144, respectively [83], most cases of which were related to infusion reaction. In poly JIA, anti-tocilizumab antibody developed in 1 of 19 patients by week 48 [55].

Vaccinations

In general, vaccination with live-attenuated vaccines is prohibited in patients treated with biologic DMARDs, including tocilizumab, because of the risk of infection. A retrospective, multi-center survey detected three tocilizumab-treated soJIA patients with inactive disease who received live-attenuated vaccines [84]. One patient who received an oral polio vaccine had diarrhea 10 days after the vaccination. Another soJIA patient who was injected with varicella vaccine experienced a mild disease flare. The last patient received an MMR booster without any adverse event. Inactivated vaccines are safely injected into patients with soJIA receiving biologic DMARDs. Inoculation of flu vaccine during tocilizumab therapy can have the same immune response as healthy subjects without adverse events [85].

Pregnancy

Prospective and retrospective studies revealed 20 JIA patients who received tocilizumab just before or during pregnancy; they had more preterm births but no increased malformation risk [86].

Comparison of tocilizumab with other biologic DMARDs in JIA

No head-to-head comparative study has been conducted between tocilizumab and other biologic DMARDs in JIA. However, several retrospective studies using registries have recently been published.

In a systematic review and meta-analysis using five short-term randomized controlled trials in soJIA [74], there was no difference in effectiveness among various biologic DMARDs including tocilizumab, and there was no difference in adverse events between tocilizumab and IL-1 inhibitors. A short-term cohort of 66 soJIA patients in a real-world UK setting found that tocilizumab was superior in drug survival, while the efficacy of both were comparable 1 year after treatment [87]. Additionally, a retrospective observation study in 71 tocilizumab-treated and 60 IL-1 inhibitor-treated soJIA cases using the German BIKER registry indicated tocilizumab had greater effectiveness in systemic symptom resolution and active joint count reduction compared with IL-1 inhibitors [88]. Meanwhile, there were no differences in the JADAS-10 score, JADAS remission with a score of <1, minimal disease activity with a score of <3.8, or ACR inactive disease between tocilizumab and IL-1 inhibitors at the final observation. The drug-cessation rate due to insufficient effect was 8.5% for tocilizumab and 22% for IL-1 inhibitors; however, there was more cancellation due to intolerance for tocilizumab than for IL-1 inhibitors. The drug-discontinuation rate at remission was 25% for tocilizumab and 18% for IL-1 inhibitors. There were more adverse events in tocilizumab-treated cases than in IL-1 inhibitor-treated cases. MAS occurred similarly in both groups.

According to a meta-analysis of five randomized clinical trials for poly-JIA, including the CHERISH study, tocilizumab did not differ in effectiveness or safety compared with abatacept, adalimumab, anakinra, and etanercept [89]. In the German BIKER registry, there was no difference in the therapy response between three biologic DMARDs, and drug survival for these DMARDs was long in the order of tocilizumab, etanercept, and adalimumab [90]. The drug-cessation rate due to insufficient effect was 12% in both etanercept and tocilizumab and 22% in adalimumab. The drug-discontinuation rate at remission was 13% in etanercept, 9.3% in adalimumab, and 2.7% in tocilizumab.

Two prospective observational cohorts of biologic use in the United Kingdom demonstrated tocilizumab was infused in only 35 cases by 2014 [91], compared with 206 etanercept-treated cases. In soJIA, tocilizumab was used as the first biologic DMARD in 30 cases, whereas tocilizumab was applied as the first biologic DMARD for other JIA types in only 5 out of 292 cases. In soJIA, 64% of IL-1 inhibitor users and 80% of TNF inhibitor users switched to tocilizumab as the second biologic treatment. The inefficacy rate was 28% for tocilizumab compared with 64% for IL-1 inhibitors and 80% for TNF inhibitors. Notably, there were many cases where an adverse event was the reason for the switch. However, the French CEMARA register included <10 cases of tocilizumab-treated patients but recommended tocilizumab use for soJIA patients resistant to other biologic DMARDs [92].

IL-6 in other biologic therapies

In soJIA, canakinumab treatment induced a significant decrease in the serum IL-6 level at 3 days post-administration [93]. In contrast, for anakinra-treated patients, SOCS3 was upregulated in a whole transcriptome analysis using whole blood samples [94]. Together, these findings suggest the existence of cross-talk between IL-1 signaling and IL-6 signaling in soJIA. Additionally, regardless of tocilizumab use, a > 50% reduction in serum IL-6 levels within 3 months post-treatment and normalization within 6 months is related to clinical remission after 2 years [95]. In oligo/poly-JIA and PsA-JIA, serum IL-6 levels are known to decrease when TNF inhibitor treatment is effective [96], at which point the IL-6 value correlates with the JADAS and PASI scores in PsA-JIA. Meanwhile, ustekinumab, an IL12/23 inhibitor, causes no change in serum IL-6 levels, even during clinical remission [97]. IL-6R mRNA expression is significantly suppressed in ankylosing spondylitis patients effectively treated with a TNF inhibitor [98].

Complications

JIA has some complications, for example, MAS and uveitis. The importance of management of these complications is stated in the T2T strategy for JIA [37]. IL-6 is definitely related with these conditions.

Macrophage activation syndrome (MAS)

MAS is a representative lethal complication in pediatric rheumatic disease. It is a hemophagocytic syndrome, classified as secondary hemophagocytic lymphohistiocytosis. MAS occurs in 10%–20% of soJIA patients during the active stage. Currently, who develops MAS and when it emerges remain unpredictable.

The cause of MAS is thought to be an uncontrolled and persistent immune response by activated T lymphocytes, NK cells, and macrophages with impaired apoptotic cell clearance. Various cytokines, including IL-6, are presumably involved in the pathogenesis. Serum IL-6 is more highly elevated in MAS than in soJIA; however, this elevation is not considered to be essential for MAS pathogenesis because IL-6 inhibition therapy cannot treat or prevent MAS. A recent analysis reported that IFN-γ might be at the center of MAS [99], but the detailed mechanism was not revealed. It is well-known that MAS can be reproduced by adding LPS stimulation to IL-6-transgenic mice, suggesting the sustained IL-6 elevation and toll-like receptor activation might be important in MAS development [100]. Besides, a recent analysis revealed the elevation of Follistatin-like protein 1, which is a mediator for producing IL-6 under stimulation by IL-17 and TNFα in bone marrow stromal cells, in MAS [101]. These findings suggest IL-6 plays a key role in the pathogenesis of MAS as well as in JIA.

MAS can be induced by various drugs for soJIA. Twenty-one of 627 tocilizumab-treated patients developed MAS [102]. Yokota reported that 3 and 12 cases of definite and probable MAS, respectively, occurred in 394 soJIA patients [103]. MAS also occurred in one tocilizumab-treated patient with cytopenia [80]. Notably, MAS symptoms are masked or alleviated when MAS is generated under tocilizumab therapy; in such cases, fever is not observed and there is minimal hepatomegaly and less elevation of CRP and ferritin [104,105].

Uveitis

Uveitis is a rare complication of JIA, observed only in patients with oligo-JIA, PsA-JIA, or ERA. The frequency varies from <1% to 16%, depending on ethnicity and country. This complication is categorized as noninfectious uveitis; in some cases, IL-6 is elevated in the vitreous body and hydatoid as well as in the serum, suggesting IL-6 plays a role in uveitis pathogenesis. This idea is supported by the efficacy of anti-IL-6R antibody in uveitis patients and the experimental autoimmune uveitis findings in mice that uveitis is abrogated by IL-6 inhibitor treatment as well as in IL-6-deficient mice [106].

Immunosuppressive drugs have long been used for treating uveitis, but treatment with biologic DMARDs, including IL-6 inhibitors, has now been successfully introduced against refractory cases [107]. In JIA, IL-6 inhibitors had an improving effect on visual acuity and central foveal thickness by OCT in TNF inhibitor-resistant cases [108]. Tappeiner and colleagues reported that 10 of 17 MTX or TNF inhibitor-resistant JIA-uveitis patients displayed positive effects of tocilizumab treatment, with the disappearance of macular edema [109]. Seven of these patients reportedly maintained remission without recurrence, including an HLA-B 27-positive boy who was resistant to MTX, oral glucocorticoid, infliximab, adalimumab, abatacept, and rituximab [110]. A multicenter study for JIA uveitis refractory to TNF inhibitor reported tocilizumab was effective for ocular parameters, such as anterior chamber cell number, macular thickness, and visual acuity, with maintained remission for >1 year [111].

Growth failure

In the era before tocilizumab, the height SD score in soJIA patients was known to correlate with the disease duration [83]. Furthermore, 30% of soJIA patients could not catch up to normal height, even after termination of glucocorticoid therapy [112]. The cause of growth failure is thought to be influenced by IL-6 via an inhibitory effect on the GH–IGF-1 axis [113], supported by the observation of growth retardation in IL-6 transgenic mice [114]. IL-6 overexpression is well known to affect bone metabolism, causing bone resorption enhancement and ossification inhibition [115]. It could be also explained by the direct influence of IL-6 on the growth aspect of JIA, as an improvement in height was noted in >70% of JIA patients treated with tocilizumab for 2 years [116]. Furthermore, tocilizumab-treated patients showed catch-up growth along with improvement in IGF-1, osteocalcin, C-telopeptide of type-I collagen [117], COMP (as a cartilage marker), and BAP (as a bone metabolism marker) [118], indicating that IL-6 has a pathogenic role in growth apart from its role in inflammation in JIA.

Conclusion

IL-6 is involved in both the local and systemic inflammation of JIA pathogenesis as well as in JIA complications, as in RA and PsA. Although the detailed mechanism is still unknown, IL-6 inhibitor selection should be made based on type, time, disease phase, and case, considering biologic and genetic information. Future research will reveal adequate therapeutic strategy by answering questions whether IL-6 signaling blockades affects these roles, how they function in JIA, and how the pathogenesis of JIA can be altered via IL-6.

Conflict of interest

None.

Acknowledgments

We thank Katie Oakley, PhD, from Edanz Group for editing a draft of this manuscript.