Emerging therapies in rheumatoid arthritis: focus on monoclonal antibodies

Advances in the treatment of rheumatoid arthritis (RA) are attributed to several aspects such as new classification criteria enabling early diagnosis and intensive treatment with the application of treat-to-target principles as well as better understanding of the pathogenesis of RA contributing to the development of targeted therapies. However, reaching remission is still not achieved in most patients with RA, which is one of the driving forces behind the continuous development of novel therapies and the optimization of therapeutic strategies. This review will outline several new therapeutic antibodies modulating anti-inflammatory cytokines interleukin (IL)-2 and IL-10 and pro-inflammatory mediators granulocyte-macrophage colony-stimulating factor, fractalkine, and IL-6 that are in various stages of clinical development as well as the progress in manufacturing biotechnologies contributing to the next generation of antibodies and their potential to expand the therapeutic armamentarium for RA. In addition, the fate of unsuccessful therapies including agents targeting IL-15, the IL-20 family, IL-21, chemokine CXCL10, B-cell activating factor (BAFF), and regulatory T (Treg) cells or a novel concept targeting synovial fibroblasts via cadherin-11 will be discussed.


Introduction
In recent years, dramatic advances have emerged across medical disciplines in understanding the pathogenesis of immunemediated inflammatory diseases, which are accompanied by a significant shift in treatment options 1 . One of the most significant advances has been shown in the treatment of rheumatoid arthritis (RA), in which the discovery of glucocorticoid activity was among the first therapeutic milestones in the 1950s 2 , followed by widespread use of methotrexate since the turn of the 1980s and 1990s 3 . Later on, success in the treatment of RA can be attributed to the introduction of new classification criteria allowing earlier diagnosis and treatment and application of the treat-to-target principles with the aim to induce remission or at least low disease activity 4-6 . The biggest breakthrough in the treatment of RA, however, comes with biologic therapy, which has gradually begun to spread since the beginning of the new millennium with the introduction of tumor necrosis factor (TNF) inhibitors into clinical practice 7 .
Targeted disease-modifying antirheumatic drugs (DMARDs), consisting of biologic DMARDs (bDMARDs) and targeted synthetic DMARDs (tsDMARDs), antagonize soluble cytokines and their receptors and directly affect immune cell activity or intracellular signaling pathways [8][9][10] . The effect of targeted therapies is relatively rapid and often accompanied by a significant suppression of inflammation. In most cases, it can substantially improve quality of life and stop or significantly slow the progression of functional and structural impairment 11 . Currently, out of the biological therapies, five original TNF inhibitors and already even more biosimilar TNF inhibitors, two interleukin-6 receptor (IL-6R) inhibitors, and one IL-1 receptor inhibitor as well as B-lymphocyte (CD20) and co-stimulatory signal (CD28-B7 interaction) for T-cell activation inhibitors are available for the treatment of RA (Table 1). In addition to bDMARDs, two tsDMARDs, including Janus kinase inhibitors (JAKi) tofacitinib (pan-JAKi) and baricitinib (JAK1/2i) (Table 1), have expanded the treatment armamentarium 12 . Although the fate of patients with RA has improved significantly and a milder course of disease can be observed than in the past 13 , persistent remission, let alone drug-free remission or complete cure, is still elusive 14 . This review will outline several new directions for the potential new therapeutic antibodies and treatment strategies for RA, new biological drugs against new targets, and biological drugs against already-known targets that are currently in early or late stages of clinical development.

Cytokine and cell-targeted therapies
The development of biological therapies for RA has been facilitated by two main assumptions. The first prerequisite is an increasing understanding of the pathophysiological processes in which systemic immune mediators, immune cells, and signaling pathways play a role and have become targets for specific immune interventions 15 . The second important assumption is the dynamic development of the pharmaceutical industry and advances in genetic engineering and biotechnology 16 . Potential therapeutic antibodies that are in various stages of clinical development are summarized in Figure 1.

New targets in later phases of development
Granulocyte-macrophage colony-stimulating factor Granulocyte-macrophage colony-stimulating factor (GM-CSF) was first described in the 1970s and is considered a pro-inflammatory cytokine, which is upregulated upon stimulation with multiple mediators such as lipopolysaccharide (LPS) or TNF and is produced by a variety of immune as well as resident tissue cells 17 . As the expression of GM-CSF is increased in the synovial tissue and synovial fluid of patients with RA, it can stimulate the synthesis of adhesion molecules and pro-inflammatory cytokines, and it can also activate synovial fibroblasts and polarize macrophages into an inflammatory M1 phenotype, thus promoting rheumatoid synovial inflammation 18 . At present, several monoclonal antibodies targeting the GM-CSF pathway in patients with RA are being studied 19 .  Black text shows therapies that were effective in the treatment of rheumatoid arthritis, although some of the drugs are not further evaluated in rheumatoid arthritis owing to company prioritization. White text shows therapies that failed to prove efficacy or whose clinical trials were terminated owing to safety concerns. *However, anti-CD20 monoclonal antibody such as rituximab is effective and licensed for the treatment of rheumatoid arthritis. BAFF, B-cell activating factor; CXCL, C-X-C motif ligand; GM-CSF, granulocyte-macrophage colonystimulating factor; GM-CSFR, granulocyte-macrophage colony-stimulating factor receptor; IL, interleukin; mAb, monoclonal antibody; TNF, tumor necrosis factor; Treg, T regulatory response to at least one conventional synthetic DMARD (csD-MARD), significantly more mavrilimumab-treated (150, 100, and 30 mg) patients achieved American College of Rheumatology definition of improvement (ACR20) compared with placebo at week 24 (73.4%, 61.2%, 50.6% versus 24.7%, respectively [p<0.001]) and clinically meaningful response was observed as early as one week after treatment initiation 21 . Another phase IIb study evaluated the efficacy and safety of mavrilimumab and golimumab (anti-TNF) in patients with RA who have had an inadequate response to csDMARDs and anti-TNF agents 22 . Mavrilimumab demonstrated numerically lower response rates compared with golimumab in patients with inadequate response to csDMARDs and similar efficacy to that of golimumab in the anti-TNF non-responders. However, the study was not powered to provide a direct comparison between mavrilimumab and golimumab, and it is probable that a suboptimal dosage of mavrilimumab was used (100 mg every other week) 22 . Pulmonary safety assessments were performed because the inhibition of GM-CSF signaling has the potential to affect the clearance of surfactant by alveolar macrophages and thus be responsible for pulmonary alveolar proteinosis.
Although there was a potential risk of this rare disease, including neutropenia, infections, and other lung toxicities associated with the blockade of GM-CSF, none of these studies nor a longterm extension study up to 3.3 years revealed such concerns 23 . Mavrilimumab demonstrated sustained efficacy and an acceptable safety profile; however, due to future clinical development plans, the study was terminated (ClinicalTrials.gov Identifier: NCT01712399).
Otilimab (GSK 3196165 or MOR103) is a fully human highaffinity IgG monoclonal antibody against GM-CSF being developed by GlaxoSmithKline. In a phase IIb dose-ranging study (BAROQUE), patients with active RA were randomized to placebo or otilimab (22.5, 45, 90, 135, or 180 mg) administered weekly for five injections, and then every other week until week 50 24 . Dose-related treatment effect with the onset of clinical response as early as week one was observed for all doses. However, the primary endpoint (DAS28 remission at week 24) was not achieved; it was only numerically higher in patients at the highest dose of otilimab versus placebo (16% versus 3%; p=0.134). In addition, exposure to otilimab was lower than predicted owing to increased apparent clearance. In July 2019, a phase III clinical program ("Con-tRAst") was announced, in which otilimab (90 and 150 mg weekly doses) will be compared with two treatments with different modes of action: JAKi (tofacitinib) and anti-IL-6R (sarilumab) (https://www.gsk.com). Interestingly, GM-CSF-targeting therapies are also being studied in other rheumatic diseases. For instance, otilimab was studied in patients with hand osteoarthritis and showed numerical reduction in pain and functional impairment, while little difference was observed with placebo on tender and swollen joints, including magnetic resonance imaging (MRI) endpoints 27 . In patients with axial spondyloarthritis, a phase IIa proof-of-concept study evaluating the safety and efficacy of namilumab has been recruiting subjects since August 2018 (ClinicalTrials.gov Identifier: NCT03622658).

Old targets with the focus on interleukin-6
In Sirukumab (CNTO-136; proposed trade name Plivensia), a human monoclonal antibody against IL-6, was developed by Centocor, GlaxoSmithKline, and Janssen Biotech. Sirukumab was investigated at doses of 100 mg and 50 mg every two and four weeks, respectively, in five phase III trials (SIRROUND) in a broad spectrum of patients with RA 32 . Sirukumab showed significant improvements in disease activity, physical function, and quality of life along with inhibition of radiographic progression 33,34 . In a head-to-head study with adalimumab, sirukumab monotherapy showed greater improvements in disease activity (DAS28) but similar ACR50 response rates at week 24 35 . Although the long-term safety of sirukumab seems to be similar to that of other IL-6-inhibiting agents, drug regulatory authorities suggested re-evaluating the safety profile of sirukumab for increased mortality rates in post open-label studies 32 . Interestingly, the drug is currently under development for the treatment of major depressive disorder (ClinicalTrials.gov Identifier: NCT02473289).

New targets in early phases of development
Interleukin-2 IL-2, first cloned in the early 1980s, is predominantly produced by CD4 + T-cells and activated dendritic cells and has pleiotropic functions (for review, see 36). In high doses, IL-2 has been found to promote effector T cells, whereas low-dose IL-2 (ld-IL2) activates regulatory T cells (Tregs) and thus can have a broad therapeutic potential in many autoimmune and inflammatory diseases. Indeed, the results of a first prospective, phase I-IIa clinical trial recently demonstrated that ld-IL2 (1,000,000 IU/day) given for five days and then once every other week for six months selectively activated and expanded Tregs without activating effector T cells. Furthermore, the authors reported signals of efficacy without safety issues in 46 patients across 11 autoimmune and inflammatory chronic diseases, including RA, ankylosing spondylitis, systemic lupus erythematosus (SLE), and psoriasis 37 . In addition, in vitro and first-in-human data on a fusion protein of IL-2 mutein and human Fc (AMG 592) demonstrated dose-dependent, selective expansion of Tregs with no increase of major pro-inflammatory cytokines such as IL-6, TNF, or interferon-γ (IFN-γ) in healthy volunteers 38 . Based on these data, another phase Ib/IIa study evaluating the safety and efficacy of AMG 592 has been underway in patients with RA since May 2018 (ClinicalTrials.gov Identifier: NCT03410056) but also in patients with SLE (ClinicalTrials.gov Identifier: NCT03451422).
Interleukin-10 IL-10 is produced by virtually all leukocytes and inhibits the production of pro-inflammatory cytokines, e.g. TNF and IFN-γ, and abrogates antigen presentation and cell proliferation (for review, see 39). Despite the fact that it belongs to the most potent anti-inflammatory cytokine, limited efficacy with subcutaneously administered recombinant IL-10 was observed in a phase I trial in patients with RA in the past 40 .
Several reasons for this discrepancy can be speculated, e.g. complex mechanism of pathophysiological action of IL-10, including potential pro-inflammatory activity 41 , or short half-life of IL-10 hampering effective delivery of recombinant IL-10 to the sites of inflammation. To overcome these obstacles, Dekavil (F8IL10), a fully human anti-inflammatory immunocytokine composed of the vascular-targeting anti-fibronectin domain fused to IL-10, is under investigation in patients with RA 42 . In a phase II clinical trial, Dekavil (30-600 mg/kg) is administered once a week for eight consecutive weeks by subcutaneous injection in combination with MTX to RA patients who have previously failed at least one TNF inhibitor. Preliminary data have demonstrated some signs of efficacy, with 46% demonstrating ACR20 clinical response after eight weeks of drug administration. Dekavil was well tolerated; however, mild injection site reaction occurred in 60% of the patients 43 .
Fractalkine Fractalkine (FKN) is known as a CX3C chemokine that promotes cell adhesion and chemotaxis, but also angiogenesis and osteoclastogenesis, and increases the production of inflammatory mediators, thus playing a significant role in the pathogenesis of RA (reviewed in 44

Unsuccessful biological therapies in rheumatoid arthritis
Although several pro-inflammatory cytokines play a significant role in the pathogenesis of RA and their inhibition contributed to a significant reduction in synovial inflammation and joint damage in an experimental model of arthritis and proved to be effective in early phases of development in humans, further studies failed to confirm significant efficacy 46 (Table 2). For instance, IL-1 inhibitors are approved but only modestly effective in RA while highly effective in several autoinflammatory diseases 47 . An early phase study with IL-15 inhibitor therapy seemed to be efficient 48  Several strategies to inhibit B-cells were tested; for instance, humanized and human monoclonal antibodies ocrelizumab (trade name Ocrevus) and ofatumumab (trade name Arzerra) proved to be effective in patients with RA. However, after risk/benefit consideration, further studies were discontinued 54 . In order to reduce the molecule, promote penetration to sites of inflammation, and improve efficacy, single-chain antibody polypeptides known as SMIPs (small modular immunopharmaceuticals) have been developed that are approximately one-half to one-third of the size of the standard monoclonal antibody. A next-generation, humanized SMIP protein therapeutic against the CD20 antigen (SBI-087) was evaluated in patients with RA, but further development was terminated 55 . In addition, a further strategy to inhibit B-cell differentiation and survival via  anti-B-cell activating factor (anti-BAFF) antibody has also failed in the treatment of patients with RA 56 . Direct T-cell-targeting therapies such as alemtuzumab, keliximab, and clenoliximab did not show satisfactory results and were accompanied by a number of serious adverse reactions associated with severe CD4 + T-lymphopenia and rash 57 . Tregalizumab (BT-061) is a humanized agonist antibody that selectively activates Treg lymphocytes and was tested as an innovative treatment for RA; however, it also did not show significant clinical benefit 58 .
Recently, a novel approach targeting synovial fibroblasts in RA via the inhibition of cadherin-11 was presented 59 . Although an interesting concept, anti-cadherin-11 (RG6125) given on top of anti-TNF therapy in patients with active RA failed to be more effective than placebo and the study was discontinued 59 .

Next-generation antibodies
Conventional monoclonal antibodies (e.g. infliximab or adalimumab), Fc fusion proteins (etanercept and abatacept), and one antibody-derived molecule, a PEGylated anti-TNF Fab fragment (certolizumab pegol), are widely used for the treatment of patients with RA. Further advances in antibody engineering have led to the development of so-called next-generation therapeutic antibodies representing antibody fragments, dual/ bispecific targeting variants, and immunocytokines 60-62 . The purpose is to develop antibodies that are smaller, more stable, have higher affinity and improved tissue penetration, and have lower immunogenicity and toxicity than conventional antibodies. For instance, nanobodies are single-domain antibodies that were originally developed after the discovery of camelid antibodies consisting of heavy chain variable domain (lacking the Fcequivalent region) and can be engineered to bind to two (or even more) different antigens or epitopes on the cell surface, so-called bispecific (even multispecific) antibodies 61,62 . These antibodies are frequently used to treat cancer. Another therapeutic success was achieved by the development of immunocytokines, a fusion of an antibody and cytokine, which are widely studied as an anticancer therapy 63 and should extend cytokine half-life, reduce systemic adverse events, and deliver the cytokine specifically to sites of inflammation as mentioned above (Dekavil or AMG 592).
Belgian biopharmaceutical company Ablynx has developed unique new-generation bispecific therapeutic nanobodies ozoralizumab (ATN-103) and vobarilizumab (ALX0061) for the treatment of RA, the first humanized bispecific nanobodies, which neutralize TNF and IL-6R, respectively, and also bind to human serum albumin to increase the half-life. Ozoralizumab was tested in subcutaneous form at a once-monthly dosing interval in patients with RA and after 12 weeks of treatment, EULAR response was obtained in 97% of patients, an ACR20 response was achieved in 84%, and DAS28 remission was seen in 38% of patients with RA 64 . A phase II study evaluating long-term safety and tolerability was completed in 2012 (ClinicalTrials.gov identifier: NCT01063803), and no further reports of development have been made so far. An early phase study with vobarilizumab demonstrated promising results as well, with ACR20 response rates of up to 84% and DAS28 remission rates of up to 58% 65 . In addition, in a phase IIb monotherapy study (head-to-head versus tocilizumab) in RA patients 66 , a similar proportion of vobarilizumab-treated patients achieved ACR20 response at week 12 compared with tocilizumab (73-81% versus 78%). Although more patients achieved DAS remission on highest vobarilizumab dose (225 mg every other week), when using more stringent CDAI and SDAI remission criteria, a similar efficacy was obtained in the monotherapy with vobarilizumab (5-10%) and tocilizumab groups (9-11%).
Remtolumab (ABT-122), developed by AbbVie, is another bispecific antibody, a full-length IgG with dual-variable domains (DVD-Ig), targeting two important human cytokines: TNF and IL-17A. Based on a TNF-transgenic mouse model of arthritis, dual inhibition of both cytokines proved to be more effective in suppressing inflammation and bone/cartilage destruction than either cytokine inhibitor alone 67 . However, a phase II trial demonstrated that a strategy of dual inhibition of TNF and IL-17A with remtolumab does not appear to be significantly different from that of TNF inhibition alone with adalimumab in human patients with RA and psoriatic arthritis 68 . In addition to this finding, recent data proved that another DVD-Ig lutikizumab (ABT-981) targeting IL-1α/β did not improve pain or synovitis in patients with both erosive hand osteoarthritis and knee osteoarthritis 69,70 .

Conclusion
Biologicals have become an essential part of the therapeutic armamentarium for RA, significantly ameliorating disease activity and improving quality of life. Furthermore, the progress in our understanding of the pathogenesis of RA has significantly contributed to the development of novel intracellular targeted therapies such as JAK inhibitors tofacitinib and baricitinib that have already been implemented in routine clinical practice. Nowadays, more selective JAK inhibitors and other targeted synthetic therapies are in various stages of clinical development in RA, which is beyond the scope of this review and is reported elsewhere 71,72 . In addition to the development of new therapies, the patent protection of older original biological agents expires and biosimilars have been entering the market within the last few years, thus lowering the price and enabling the treatment of more patients 73 .
Despite current practice, early and intensive management with the treat-to-target approach, reaching remission is still not achieved in most patients, which is one of the driving forces behind the continuous development of novel therapies and the optimization of therapeutic strategies 74 . Thus, advances in antibody engineering and further therapeutic targets such as neuroimmune modulation, gene therapy, and epigenetic modification are being evaluated in the treatment of RA 75-77 . In addition, the two main treatment strategies can be potentially applied in the future to improve disease outcome, halt further progression, or even cure the disease either 1) by an induction-maintenance regime with biological therapies administered as early as the diagnosis is established 78 or 2) even before clinical arthritis becomes apparent 79 . Currently, several ongoing proof-ofconcept trials are testing whether targeting T-cells or B-cells can induce immunological reset and be used to prevent the onset of RA 80 . An advanced study demonstrated that a single infusion of 1,000 mg rituximab significantly delayed the development of arthritis in patients at risk of developing of RA 81 . Other treatment approaches for preventive strategies may represent, for instance, the use of tolerogenic dendritic cells 82 or targeting the IL-17/23 axis 83 in order to block arthritis development in pre-clinical phases of the disease. However, further research is needed to determine which therapeutic agent or strategy would best suit each individual patient.

Grant information
This work was supported by the Project for Conceptual Development for the institution of Ministry of Health Czech Republic -Institute of Rheumatology (number 023728).
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.