Rheumatoid arthritis (RA) is a systemic, chronic autoimmune condition characterised by joint inflammation, that affects 1% of the UK adult population (1). The disease can have a huge impact on patient quality of life and healthcare costs. In 2010 the National Rheumatoid Arthritis Society estimated the annual cost of RA to the NHS at £700 million, but to the UK economy as a whole, £8 billion due to sick leave and falls in productivity (2). We know that RA patients do less physical activity and this can contribute to co-morbidities such as cardiovascular disease, depression, type 2 diabetes as well as a reduced life expectancy (3).
The disease often presents itself in individuals around middle age, with most common complaints being painful, swollen and stiff joints of the metacarpophalangeal (MCP) and proximal interphalangeal (PIP) joints, in a symmetrical distribution. Other commonly affected joints include the wrist, metatarsophalangeal (MTP) joints, elbows and ankles, while atypical presentations include a large joint monoarthritis of the hip or knee joint. The affected joints are warm to touch due to the underlying inflammatory process, and joint stiffness is worst after periods of inactivity and in the mornings, usually lasting for more than 30 minutes. Along with the clinical picture, diagnosis relies on the detection of autoantibodies such as rheumatoid factor (RF) or anti-cyclic citrullinated peptide antibodies (ACPAs) in the patient’s blood – seropositive RA. RF is a non-specific antibody against the fragment crystallizable (Fc) region on immunoglobulin G (IgG), present in 70% of those with RA, while ACPAs are highly specific for RA at 98%. If these autoantibodies are not present but the clinical picture fits, the disease is classified as seronegative RA.
Aside from synovitis, there are other underlying processes that contribute to the pain and disability experienced by those with the disease. Cells in the synovium proliferate rapidly to cause synovial hyperplasia, pannus formation, angiogenesis and erosion of cartilage and bone.
The aim of treatment is to reduce chronic pain and inflammation, slow down disease progression and induce remission, whereby patients can carry on functioning in their normal daily lives. When acute flare-ups occur, higher concentration and faster acting drugs can be applied to provide short-term pain relief. Currently, RA is treated first line using two disease-modifying antirheumatic drugs (DMARDs), one of which is usually methotrexate (MTX). Corticosteroids are considered in acute flare-ups, while biological treatment is reserved for DMARD resistant cases. Problems lie with the huge side effect profile of many of the drugs, treatments displaying efficacy in only around two thirds of patients, the high cost of biologicals and the inability to maintain drug concentrations at therapeutic levels locally in affected joints. There is clearly a need to develop new drugs that are targeted, can provide a sustained response and offer some sort of efficacy in patients that are treatment resistant.
In this review, we will examine the origins of RA and its pathogenesis, looking at which molecules can potentially act as targets for RA therapeutics. We will look at the current treatment options, then look at the evidence base for the latest alternative and novel approaches to RA therapeutics. This will include studies looking at growth factors, transcription factors, the role of nanotechnology in drug delivery and potential gene therapies.
RHEUMATOID ARTHRITIS AETIOLOGY
The exact cause of RA has yet to be established with most hypotheses and mechanisms citing both genetic and environmental influence. It is important to remember that autoimmunity and systemic inflammation are present before clinical presentation, hence to understand disease pathophysiology, we should examine the genetics of the disease and how this translates into a preclinical RA environment.
Differentiating genetic and environmental components of a disease has classically been investigated through twin studies, by comparing the concordance rates of RA in monozygotic (MZ) twins and dizygotic (DZ) twins. MZ twins have identical genotypes whereas DZ twins share only around 50% of their genes, so by observing disease concordance rates in both twin groups, it can be determined the extent to which genetics influences RA development. In a Danish study, probandwise concordance was 9.1% for MZ twins, 6.4% in DZ same sex twins and 2.2% in DZ opposite sex twins (4). It estimated that only 12% of RA is due to genetics, although it suggests a higher genetic component in ACPA positive RA. Heritability was estimated at 39% for ACPA positive RA in a more recent study (5), far lower than historical estimates (6, 7).
RA displays a sex bias in favour of females in a 3:1 ratio to males suggesting X-linked genetic factors and hormonal factors are involved (8, 9). In females, the ratio of paternal and maternal X-chromosome expression tends to be around 50%, but sometimes this can be skewed, offering a potential autoimmunity mechanism whereby X-linked self-antigens can evade negative selection in the thymus and escape to peripheral sites (10).
Major histocompatabilty complex (MHC) class II molecules are strongly associated with RA. MHC molecules, encoded by the human leukocyte antigen (HLA) gene complex on chromosome 6, help regulate the immune system in humans. Class II molecules aid presentation of both self and foreign peptide antigens by professional antigen presenting cells (APCs) to CD4+ helper T cells. Many HLA-DR1 and -DR4 serotypes have been linked to RA susceptibility and disease severity, in particular a shared epitope which consists of a sequence of 5 amino acids in positions 70-74 of the β1 chain of HLA-DR which includes the HLA-DRB1*0401 and *0404 alleles (11-13). Susceptibility and severity varies according to which shared epitope alleles are present among different ethnic populations, with some alleles even having a protective effect. Genome wide association studies have revealed several non-MHC genes that increase susceptibility to RA including PTPN22, CTLA4, PADI4, TRAF1, IRF5, CD40 and BDKR1 (14-16). While the association is apparent, the mechanisms to cause RA are unknown. It has been therefore postulated that while genetic factors are likely to determine susceptibility, epigenetic, post-translational modification and environmental factors are essential to triggering RA.
Despite the human genome project that sequenced human DNA being completed 15 years ago, it has been difficult to pinpoint the genetic basis of RA, in part due to its polygenetic and epigenetic nature. Epigenetics involves studying how gene expression and activity changes, despite the original DNA sequences remaining the same, through DNA methylation, histone modification and non-coding RNA (ncRNA). For example, highly significant changes in DNA methylation have been observed at 32 sites in T-cells and 20 sites in B cells, in RA patients (17). Half the T-cell sites were hypermethylated whereas all the B-cell sites were hypomethylated, with some of these sites corresponding to the MGMT DNA repair gene and CCS gene. Hypermethylation was found at 6 sites in the genes for T-cells, which corresponded to the gene coding for DUSP22. DUSP22 can negatively regulate the IL-6/STAT3 pathway, hence if intrinsically modified, may result in upregulation of IL-6 production, a cytokine partly responsible for the autoimmunity and chronic inflammation seen in RA patients (17).
Twin studies observing MZ twins over a period of time are useful for identifying environmental triggers. If only one of the MZ twins develops RA, this will largely be due to a single or combination of unique environmental factors, since their genotype is almost identical. Smoking is one of the most strongly associated risk factors for RA development in MZ twins (OR 12, 95% CI 1.78-513) and DZ twins (OR 2.5, 95% CI 0.92-7.87) (18). The mucosal surface microbiome compositions of the gut, lung and oral cavity (in particular, porphyronomas gingivalis)have been implicated in RA onset (19, 20).
In the early stages, it is thought that an environmental or genetic trigger elicits an immune response at a mucosal site, causing co-stimulation of T cells, B cells and dendritic cells, resulting in the production of RFs, ACPAs and other autoantibodies. Autoantibodies have been shown to be present for up to 9 (ACPA) and 22 years (RF) before symptom onset (21). B lymphocytes inside and outside of the joint can produce ACPAs which target citrullinated fibrinogen, type II collagen, vimentin, fibronectin and α-enolase, and on binding form immune complexes (22, 23). RF can bind to these complexes, and lead to the activation of the complement cascade (24). How autoimmunity results in intra-articular inflammation remains uncertain but mechanisms involving spread of epitopes from the initial mucosal site response via the lymphatic system to secondary lymphoid tissues, deposition of immune complexes within intra-articular microvessel endothelium and synovial tissue, and dissemination of activated T cells have been suggested (20).
Joint swelling caused by inflammation of the synovial membrane occurs with infiltration of leucocytes into the synovial compartment. Chronic activation of fibroblast-like synoviocytes (FLS) and most prominently resident macrophages, by T cells results in the production of pro-inflammatory cytokines including interleukin (IL)-1, IL-6, tumour necrosis factor α (TNFα), regulatory cytokines such as IL-10, and matrix metalloproteases (MMPs) (25). Neutrophils produce prostaglandins and reactive oxygen species. Upregulation of Th17 helper cells and suppression of regulatory T cells further promotes an inflammatory environment. Combined with hypoxic conditions, vascular endothelial growth factor (VEGF) is produced, and supported by cell adhesion molecules, MMPs and cytokines, it drives proliferation of endothelial cells and angiogenesis (26). Other growth factors (PDGF, FGF) produced in the synovium help stabilise the formation of these new blood vessels. The microenvironment facilitates migration of FLS and macrophages into the synovium and their proliferation, causing synovial hyperplasia and pannus formation. How these cells evade apoptosis is unclear but tumour suppressor gene mutations and modulation of the NF-KB signalling pathway remain plausible theories (27).
Cells of myeloid and lymphoid lineage, in particular macrophages, express granulocyte-macrophage colony stimulating factor (GM-CSF) in the synovium, while its levels have also found to be elevated in synovial fluid. (28) GM-CSF facilitates macrophage differentiation and activation in a positive feedback loop. Macrophages and FLS contribute to the production of destructive proteases that degrade the cartilage. Activation of osteoclasts by TNFα, macrophage colony stimulating factor (M-CSF) and RANKL leads to bone erosion (27).
The sustained inflammatory microenvironment within the joint is maintained by a complex interaction between cells, cytokines, transcription factors and genes, encompassing many feedback loops and signalling pathways. This makes the end goal of disease remission extremely challenging, however greater understanding of the unique signature of the RA joint has enabled researchers to make advancements in managing the condition.
CURRENT THERAPEUTIC STRATEGIES
At the moment, drug regimens for RA management consist of DMARDs, glucocorticoids, biological drugs and non-steroidal anti-inflammatory drugs (NSAIDs).
DMARDs are now the mainstay of treatment regimens, with RA patients being started on this class of drugs almost immediately after diagnosis. DMARDs have been around for decades, starting around 80 years ago with injectable gold and the anti-malarial hydroxychloroquine, discovered to have immunomodulatory effects in the 1950s. Commonly used DMARDs in the present day such as MTX and sulfasalazine (SSZ) first came to light in the early 60s. MTX is considered the gold standard treatment, due to its efficacy, safety, tolerability, long-term clinical experience and relatively low costs (29, 30). MTX is an antifolate but works in a number of ways to treat RA, including accumulation of adenosine and inhibition of polyamines, both of which contribute to its anti-inflammatory effects (31). Clinical responses take time to manifest, with improvements becoming apparent from a couple of weeks in some cases, with maximal efficacy sometimes taking up to several months. MTX is usually co-prescribed with another DMARD providing a synergistic effect (32) or started as monotherapy. Failing this, substitution to SSZ or leflunomide is considered due to the similar clinical and functional efficacy. Biological drugs are commenced in severe RA or if there is no response to conventional DMARDs. DMARDs are limited due their side effect profile consisting of interstitial lung disease, bone marrow suppression, retinopathy and hepatoxicity, and hence require frequent monitoring.
Glucocorticoids are often used in acute flare-ups to provide short term pain relief and rapid reduction in inflammation, while a treatment with a slower onset is started, e.g. a DMARD. Low-dose glucocorticoids can be used long term, sometimes as an adjunct to a DMARD, but usually only if all other treatment (DMARDs and biologicals) have failed to control the disease process. Long-term use has shown to slow down progression of bone erosions (33). Most of these patients use a dose equivalent to ≤7.5mg a day of prednisolone, but even this dose is still associated with systemic side effects (34).
In the mid-90s, one of the first pieces of evidence for biological drugs demonstrated the use of a chimeric monoclonal antibody to block TNF-α, paving the way for a new direction of treatment (35). They are on of the latest drug classes to come to market for the treatment of RA. They are proteins that have been genetically engineered using largely human DNA to target and block specific inflammatory mediators, usually cells (B cells, T cells) or cytokines (TNF-α, IL-1, IL-6) (see table 1).
More recently, synthetic JAK inhibitors that block the JAK-STAT signalling pathway have been developed, as this pathway has been associated with regulation of cytokine production, including IFNα, IFNγ, IL-6, IL-12 and IL-23 (36, 37). At the moment only two have been approved for use in RA. Biologicals require administration by intravenous infusion or subcutaneous injection, whereas JAK inhibitors are taken orally. When combined with MTX, they have a high efficacy, however they are also very expensive and have numerous adverse effects. Biologicals have been associated with reactivation of tuberculosis, hypercholesterolaemia, recurrent and serious infections and they have been linked to neurological disorders and cancer development (38). They are also ineffective for a third of RA patients.
NSAIDs are used as required in RA patients, for symptomatic pain relief. They do not slow down joint destruction or disease progression. NSAIDs block COX-1 and COX-2 enzymes, reducing the production of prostaglandins that cause pain and inflammation. Side effects include increased risk of gastrointestinal (GI) bleeds and nephrotoxicity.
GROWTH AND DIFFERENTIATION FACTORS – GM-CSF INHIBITORS
GM-CSF is a glycoprotein that acts as a proinflammatory cytokine in RA. It promotes maturation of macrophages and mast cells, and their migration from bone marrow to the synovium (27). Once bound to its receptor, it leads to downstream signalling of the JAK-STAT, MAPK, PI3K and NFkB pathways (28). GM-CSF can upregulate Th17 cell differentiation, which plays a role in inflammation in RA. Administration of recombinant GM-CSF has been shown to worsen the disease (39), and if produced by cells from bone marrow, has shown to be essential for the development of collagen-induced arthritis (CIA) in mice models (40). Mavralimumab is the novel therapy under investigation as a GM-CSF inhibitor for the treatment of RA. It works as an antibody against GM-CSF receptor α. The latest phase IIb study recruited patients that had failed DMARD and anti-TNFα treatment (EARTH EXPLORER 2) (41). Patients were randomised into two groups; to receive either golimumab every 4 weeks with placebo every 4 weeks starting at week 2; or mavrilimumab every 2 week; with both groups receiving background MTX. Both treatment groups demonstrated efficacy and tolerability. Although the golimumab group showed better clinical outcomes, the EARTH EXPLORER 1 trial showed that 100mg of mavrilimumab was a suboptimal dose, and greater efficacy could be achieved at 150mg (42). The EARTH EXPLORER 1 trial, another phase IIb study also showed significantly reduced disease activity, with improvements in clinical responses seen as early as 1 week after treatment onset. Long-term follow up data from these two trials and an open-label extension study looked at safety of the drug over a cumulative 899 patient-years, with a median of 2.5 years of treatment (43). In terms of rate per 100 patient-years, nasopharyngitis (7.68) and bronchitis (5.68) were the two most common adverse events. GM-CSF is known as a haematopoietic growth factor, hence there are concerns that blockade of this cytokine could lead to neutropenia. Data from this trial showed that the incidence of neutropenia was low at 0.45 per 100 patient-years. Interestingly, the incidence of pulmonary events was 9.24, but the changes in pulmonary function (FVC, FEV1 ad CLco) was deemed not significant.
MOR103 and Namilumab are monoclonal antibodies to GM-CSF. Doses of MOR103 at 0.3, 1.0 and 1.5 mg/kg were tested during a phase Ib/IIa trial in 96 RA patients and compared to placebo (44). Nasopharyngitis was the most common adverse event. The 1.0 mg/kg dose proved most efficacious, reducing baseline DAS-28 scores from weeks 4 through to week 10 after treatment onset, with improvements seen as early as 1 week. Only patients on this dose maintained a DAS-28 score below baseline at 16 weeks. Namilumab has been evaluated in a phase Ib study, where RA patients on a stable MTX dose were treated with 150mg or 300mg of the drug subcutaneously, or placebo, on days 1, 25 and 29 (45). The drug demonstrated efficacy against placebo at all times throughout the 12 week follow up period, and similarly with MOR103, the most common adverse event was nasopharyngitis.
INTRACELLULAR SIGNALLING AND TRANSCRIPTION FACTORS
Janus kinase (JAK) is a family of tyrosine kinases, consisting of JAK1, JAK2, JAK3 and tyrosine kinase 2 (TYK2). JAK is involved in the JAK-STAT signalling pathway, which partially regulates cytokine expression. When cytokines bind to their receptors, JAK phosphorylates the tail of the receptor, which attracts and binds STAT proteins. JAK then phosphorylates the STATs, allowing them to form a dimer which can travel to the cell nucleus, bind to DNA and regulate gene expression (46).
Tofacitinib is currently the only JAK inhibitor that has been approved for RA treatment. It is used as treatment in resistant RA or where other DMARD treatment has failed to elicit a response. It has shown to inhibit the common gamma chain interleukins (IL-2, IL-4, IL-7, IL-15 and IL-21), indicating selectivity for JAK3 inhibition (46). Within CD4+ T cells, it also disrupts IL-6 signalling, blocking phosphorylation of STAT1 and STAT3, suggesting partial JAK1 inhibition. An initial phase III randomised clinical trial of tofacitinib monotherapy against placebo showed significant improvements in ACR20, ACR50 and ACR70 response, and physical function as determined by HAQ-DI (Health Assessment Questionnaire – Disability Index), at 3 months (47). Improvements in ACR20 and HAQ-DI were seen in at little as 2 weeks. Tofacitinib administration was accompanied by an increase in the rate of serious infection, compared to placebo. Out of 330 patients on the drug, 54.1% had 701 adverse events, most commonly upper respiratory tract infection, headache and diarrhoea. The drug also increased LDL levels and raised liver aminotransferase levels. Unfortunately, the number of patients that achieved disease remission (score of <2.6 on DAS-28-4 [ESR]) was not significant, despite significant reductions disease activity. A systematic review with network meta-analysis concluded that oral tofacitinib has similar efficacy to other DMARD monotherapies, similar adverse event profile and comparable efficacy to biological DMARDS + MTX, when combined with MTX (48). There is a lack of long-term follow up data on efficacy, RA disease progression and safety.
Baricitinib and Ruxolitinib (INCB018424) are selective JAK1 and JAK 2 inhibitors. Network meta-analysis of the former, synthesised from 7 RCTs, suggested that baricitinib 4mg + DMARD was the most effective treatment for RA over baricitinib monotherapy, 2mg + DMARD and adalimumab + MTX (49). Baricitinib had an acceptable adverse event profile, however clinical outcome data was limited to only 3 months of follow-up. Clinical trials for ruxolitinib in RA have stagnated at the phase II stage. Decernotinib (VX-509) is a JAK3 inhibitor in testing. It has shown efficacy against placebo as monotherapy (50) and when combined with MTX (51). It has similar side effects to tofacitinib.
The DARWIN 1 and DARWIN 2 trials investigated Filgotinib, a selective JAK1 inhibitor, for RA treatment (52, 53) . This drug has demonstrated efficacy, safety, and has rapid onset of action with improvements in ACR20 and ACR50 by week 2. Filgotinib is undergoing further testing in the DARWIN 3 trial and FINCH trials. Upadacitinib is another selective JAK1 inhibitor in development (54).
Spleen tyrosine kinase (Syk) is an intracellular protein tyrosine kinase (55). It plays a role in inflammatory cell signalling, immunoreceptors including Fc receptor (FcR) and B cell receptor (BCR), and cytokines IL-1 and TNFα. It is present in phosphorylated form in the synoviocytes of RA patients. When fibroblast-like synoviocytes (FLS) were stimulated with TNFα, there was a sustained increase in Syk activation (55). Likewise, when the synoviocytes were incubated with the Syk inhibitor R406, before being stimulated by TNFα, kinase function was reduced. R406 has been shown to inhibit mast cell activation via FcR and it can also decrease IL-6 and MMP-6 expression. Fostamatinib is the prodrug of the experimental compound R406, which has been under investigation for the treatment of RA and other various autoimmune diseases including IgA nephropathy and lymphoma. Administration of R406 or Fostamatinib slowed RA progression and reduced severity in CIA rat models (56). Efficacy of the drug in humans is questionable. Human trials have mixed results; some demonstrated improvements when the drug was combined with a DMARD (57, 58); no significant differences between the drug and placebo as determined by ACR20, ACR50 and ACR70 response at 3 months (59); and improvements in ACR20 but failure to slow structural disease progression (60, 61). A systematic review concluded that there is a higher incidence rate of hypertension, neutropenia, diarrhoea, raised transaminases and rashes in fostamatinib, than in control groups (62). These side effects have been extracted from only 4 studies and one pooled analysis. Since Syk is involved in a number of signalling pathways, there are concerns that Syk inhibition may lead to many unintended side effects, and this needs to be addressed in future research.
Nanomedicine is an expanding branch of nanotechnology, with a vast array of healthcare applications and potential uses. Currently, two areas of interest are nanomaterials and nanoparticles (NPs). NPs aim to replicate the function of a naturally existing delivery system, that of circulating cells in vivo. There are many advantages to using NPs for drug delivery. A drug can be encapsulated in a NP shell to avoid being broken down or deactivated before reaching its target, be transported and targeted to specific sites where it is needed using ligands, undergo controlled release by altering the structure and composition of the NP and they can be produced on a large scale (63). Drugs are targeted to sites of inflammation, reducing the risk of unintended systemic effect on other parts of the body. Since delivery is targeted, less of the drug needs to be administered, which should mean reduced costs for treatment. The main problem faced by NPs is avoiding opsonization, whereby opsonin binds to all foreign bodies, allowing NPs to be recognised by macrophages and then undergoing phagocytosis before being cleared from circulating blood.