Pathogenesis of rheumatoid arthritis

Pathogenesis of rheumatoid arthritis

Author
Gary S Firestein, MD

Section Editor
Ravinder N Maini, BA, MB BChir, FRCP, FMedSci, FRS

Deputy Editor
Paul L Romain, MD

All topics are updated as new evidence becomes available and our peer review process is complete.

Literature review current through: May 2016. &#124 This topic last updated: Apr 14, 2016.

INTRODUCTION — Rheumatoid arthritis (RA) is the most common inflammatory arthritis and affects about one percent of the population [1]. It results from a complex interaction between genes and environment, leading to a breakdown of immune tolerance and synovial inflammation in a characteristic symmetric pattern. Distinct mechanisms regulate inflammation and matrix destruction, including damage to bone and cartilage [2]. Given the heterogeneous response to therapy, it is clear that RA in not just a single disease; instead, many pathways can lead to autoreactivity with similar clinical presentations.
The pathogenesis of RA is reviewed here. The etiology of this disorder, including putative genetic and environmental factors, is discussed separately. (See "Epidemiology of, risk factors for, and possible causes of rheumatoid arthritis".)
OVERVIEW — The initiation of rheumatoid arthritis (RA) is a combination of pre-determined (genetic) and stochastic (random) events. Susceptibility to RA is clearly defined by a pattern of inherited genes, with the human leukocyte antigen (HLA) major histocompatability (MHC) genes as the most important. However, scores of minor genes that include cytokine promoters, T cell-signaling genes, and many others, contribute to susceptibility and severity. It is equally clear that genes are not the sole influence, since the concordance rate for identical twins is only 12 to 15 percent. Of the environmental stimuli that contribute, the best defined is smoking, which can interact with genes to increase susceptibility up to 20- to 40-fold [3]. Epigenetic influences, such as hypomethylation of DNA, dysregulated histone marks, or expression of microRNAs, can also increase proinflammatory gene expression [4].
The most likely mechanism for the environmental component is repeated activation of innate immunity, especially at mucosal surfaces. This process can take many years, with evidence of autoimmunity increasing gradually until some unknown process tips the balance toward clinically apparent disease. For example, cigarette smoking is strongly associated with RA and induces peptidyl arginine deiminase (PAD) expression in alveolar macrophages [5]. These enzymes then convert arginine to citrulline in the airway, thereby creating neoantigens that can be recognized by the adaptive immune system. Increased citrullination is not specific to RA and occurs regularly with any environmental stress. What is unique to RA is the propensity for immune reactivity to the neoepitopes created by protein citrullination with the production of anti-citrullinated protein antibodies (ACPAs).
Other mucosal surfaces can also potentially contribute. The oral mucosa harbors Porphyromonas gingivalis in periodontal disease, which is also associated with RA [6]. These bacteria express PADs, which can citrullinate peptides in the mouth. The gut microbiome is also altered in early RA, with a preponderance of Prevotella species [7]. The influence of the bacterial environment of the gut is not well defined, but clearly affects arthritis susceptibility and severity in many pre-clinical models [8].
In “pre-RA,” ACPAs and autoantibodies like rheumatoid factors (RFs) can appear more than 10 years before clinical arthritis [9]. Other antibody systems, such as to carbamylated proteins, can also occur in response to proteins where lysine is non-enzymatically converted to homocitrulline. Note that ACPAs and other antibodies against altered peptides are not truly autoantibodies, since they are actually antibodies that recognize modified proteins such as vimentin and enolase. True autoimmunity probably occurs somewhat later due to epitope spreading. The antibodies may contribute to the initiation or exacerbation of synovitis, but do not necessarily cause RA by themselves [10]. A pattern of systemic inflammation is also apparent in pre-RA patients, as determined by multiplex analysis of cytokines in the serum [11]. Like autoantibodies, levels of multiple cytokines gradually increase in the years before RA symptoms occur.
Synovial biopsies in ACPA/RF-positive patients with arthralgias, which could be construed as “pre-RA,” are essentially normal [12]. A “second hit” is probably required to convert the predisposition to disease to clinically apparent synovial inflammation. In mouse models, this process can be mediated by immune complexes that engage the synovial innate immune system, especially mast cells [13]. Subsequent increases in vascular leakage provide access to the joint and permit complement fixation, recruitment of immune cells, and inflammation.
Once the inflammatory process is fully established, the synovium in RA organizes itself into an invasive tissue that can degrade cartilage and bone. The rheumatoid synovium has many characteristics of a locally-invasive malignancy, but it never becomes completely unresponsive to antiinflammatory and antiproliferative factors. Fibroblast-like synoviocytes (FLS) in the rheumatoid synovium can migrate from joint to joint, perhaps accounting for the symmetric and diffuse distribution.
GENETICS AND EPIGENETICS — An individual’s genome clearly plays a role in susceptibility and severity of rheumatoid arthritis (RA). The concordance rate in identical twins is 12 to 15 percent, while first-degree relatives of RA patients also have increased risk. The most prominent genes are the class II major histocompatibility complex (MHC), most notably human leukocyte antigen (HLA)-DR [14]. However, there are now over 100 different gene polymorphisms associated with RA. Most of these individually have a rather limited contribution, with odds ratios of 1.05-1.2. However, a few, like peptidyl arginine deiminases (PADs) and the T-cell signaling phosphatase encoded by PTPN11, can increase risk by about twofold. The risk genes are generally implicated in immune response, matrix regulation, and inflammation, suggesting that the mutations are causative. The genetics of RA are described in detail separately. (See "HLA and other susceptibility genes in rheumatoid arthritis".)
DNA methylation is a key epigenetic mechanism that silences gene expression when promoter regions are methylated. A variety of studies have demonstrated abnormal DNA methylation in RA peripheral blood mononuclear cells as well as synovial cells [15]. Studies evaluating the methylome of RA fibroblast-like synoviocytes (FLS) demonstrate a pattern of abnormal methylation in genes that have been implicated in the disease, including the complement cascade, focal adhesion, cytokine interactions, and others [16]. The methylome signature is stable in these cells, persists for many months in culture, and occurs very early in the disease [17]. Cytokines in the rheumatoid synovium could contribute by regulating enzymes, such as DNA methyl transferases [18]. Other epigenetic features, such as abnormal regulation of microRNAs, have also been observed [19]. Histone marks have also been implicated in RA FLS phenotype, and histone deacetylase inhibitors block cytokine gene expression and suppress inflammation in animal models of arthritis [20].
T LYMPHOCYTES — Activation of innate immunity is probably the earliest process in rheumatoid arthritis (RA), followed by citrullination or carbamylation, loading of antigen-presenting cells (APCs) with autoantigens or altered native peptides in the joint, then migration to central lymphoid organs. Once there, APCs present an array of antigens to T cells, which can then activate B cells and/or migrate back to the synovium.
Initial antigen(s) and the T lymphocyte — T cells constitute about 50 percent or more of cells in most RA synovia; most of these are CD4+ with a memory phenotype, but only 5 percent or fewer are B lymphocytes or plasma cells. RA synovial T lymphocytes display an activated surface phenotype, with high expression of human leukocyte antigen (HLA)-DR antigens and CD27. CD27+ CD4+ T cells provide B cell help that can potentially increase synovial antibody production. There appears to be a preponderance of T cells of the Th1 and Th17 subset, with deficiency of Th2 and regulatory T cells (Tregs).
It is unlikely that a single “rheumatoid antigen” exists. Instead, a broad spectrum of joint specific antigens, such a type II collagen, or nonspecific citrullinated antigens, is responsible [21]. Examples of citrullinated peptides that have been implicated include fibrinogen, vimentin, enolase, and collagen, each of which can elicit immune responses more efficiently than the unmodified proteins. The initiating antigen(s) probably vary from patient to patient, perhaps from joint to joint, and from early to late disease in the same patient. This concept has important implications in the search for pathogenic T cells and the likelihood that a single approach to tolerizing lymphocytes might not be effective in all patients.
For maximal T cell responses, second signals are generally required. Two of the co-stimulatory molecules through which such signals are provided are CD28 and CD40 ligand; both are highly expressed by synovial T cells in RA [22,23]. Synergistic effects of CD40-CD40 ligand interaction and interleukin (IL)-1 on cytokine production may occur [24].
An analysis of the physical structure of the "shared epitope" (SE) in the third hypervariable region of the HLA-DRB chains expressed on macrophages and B cells, the APCs, has shed some light on the initial antigens likely to be responsible for RA. The risk for RA is encoded by amino acid substitutions at positions 72 to 74 and modulated by the amino acids at positions 70 and 71 of the HLA-DRB1 molecule [25-27]. (See "HLA and other susceptibility genes in rheumatoid arthritis", section on 'Individual alleles and the shared epitope'.)
The three-dimensional reconstruction of the HLA class II antigen structure suggests that the SE may not restrict the precise structure of antigens selected for presentation to T cells. Rather, the epitope alters the conformation of the antigen-DR complex presented to the T cell receptor. As a result, it probably is not relevant for the T cell which peptide is presented in the susceptibility epitope; recognition is likely to be peptide-dependent, but not peptide-specific. These “arthritogenic” antigen(s) may carry a net negative charge [28], and frequently the citrullinated peptides bind with greater affinity than the unmodified peptide. Fine structure analyses unexpectedly suggest that amino acids 11 and 13, which are in the antigen-binding groove of the HLA molecule, have the closest association with RA in patients with the susceptibility epitope. These amino acid residues could play a role by binding to arthritogenic citrullinated peptides [29]. In fact, the citrullinated peptides bind more avidly to RA-associated alleles than the native peptides and are presented more efficiently to T cells [30]. The selective binding of altered antigens suggests that citrullination creates epitopes recognized by T cells that were not deleted during development.
Costimulation — Costimulation is an important aspect of T cell activation during the immune response in RA [31,32]. Presentation of antigen to T cells by antigen-presenting cells (eg, dendritic cells, macrophages, B cells) without costimulation by receptor/coreceptors (such as CD28/CD80, CD86, intercellular adhesion molecule-1/lymphocyte function-associated antigen-1), leads to anergy and death of insufficiently activated T cells. A B7-binding molecule, CTLA-4 fused with an immunoglobulin fragment (CTLA4-Ig or abatacept), is used in the treatment of patients with RA. (See "Major histocompatibility complex (MHC) structure and function" and "T-cell targeted therapies for rheumatoid arthritis", section on 'Abatacept'.)
Intriguing evidence suggests that costimulatory molecules are present in excess within rheumatoid tissue, thereby implying that T cell activation may take place without specific antigen [33-35]. As a result, self-perpetuating cycles of T cell proliferation sufficient to sustain autoimmunity may occur.
Apoptosis — Apoptosis (programmed cell death) has been investigated as a possible underlying defect among T cells in RA. As an example, deficiencies in normal programmed cell death due to an underlying T cell defect are a cause of autoimmune lymphoproliferation and arthritis in MRL lpr/lpr mice [36]. In addition, there is overexpression of a tumor suppressor gene p53 in rheumatoid synovium [37]. Mutations that have been identified in p53 may prevent normal DNA repair and interfere with apoptosis processes regulated by this gene product [38]. Other abnormalities that might contribute to increased survival of synovial cells include dysregulated phosphatase and tensin homolog (PTEN) and sentrin in the intimal lining [39]. hydroxymethylglutaryl CoA (HMG-CoA) reductase inhibitors (statins) block protein geranylgeranylation, and some of the drugs can induce apoptosis in RA synoviocytes through mitochondrial- and caspase 3-dependent pathways [40].
T cell-mediated adaptive immunity — A variety of antigens recognized by T cells may contribute to the synovial inflammatory response, either through the subsequent generation of autoantibodies by B cells or through the activation of T helper subsets like Th17 cells. Several autoantigen systems have potential as pathogenic factors. (See 'Type II collagen' below and 'Cartilage antigen glycoprotein-39 (gp39)' below and 'Immunoglobulin G' below and 'Citrullinated proteins and peptides' below and 'Carbamylated proteins and peptides' below.)
Type II collagen — One commonly implicated antigen is type II collagen, which is uniquely found in articular cartilage and the vitreous of the eye. Autoreactivity to type II collagen occurs in RA, although anti-collagen antibodies are present in a minority of patients. Some evidence suggesting autoimmunity to collagen includes:
●Citrullinated collagen activates T cells more effectively than native collagen in patients with the RA-associated class II major histocompatibility (MHC) genes [41].
●Chronic joint inflammation can be induced in experimental models by immunization with type II collagen [42], and by administration of monoclonal antibodies with affinity for type II collagen plus an adjuvant (E. Coli lipopolysaccharide) [43].
●Synovial fluid from patients with RA is enriched with activated T-cells that proliferate in vitro in response to type II collagen [44].
●Type II collagen that has been modified by oxidants present in the rheumatoid joint (eg, hydroxyl radical, hypochlorous acid and peroxynitrite) express neoantigenic epitopes that bind to antibodies in rheumatoid serum [45].
Cartilage antigen glycoprotein-39 (gp39) — This glycoprotein is secreted by both synovial cells and chondrocytes, and is selectively induced at sites of inflammation and tissue injury. Peptides derived from this protein (T cell epitopes of gp-39) that were defined in HLA-DR4 transgenic mice also stimulated T cells from human subjects that carried RA-associated HLA-DR4 alleles [46]. Cytokines produced subsequently by these activated T cells could help drive the chronic inflammatory process. SE-positive dendritic cells presenting an immunodominant human gp39 peptide (amino acids 263-275) may be more frequently found in the synovium of patients with RA than those with other inflammatory joint diseases or with osteoarthritis (61 versus 3 percent, respectively) [47].
Immunoglobulin G — Antibodies to immunoglobulin G (IgG), called rheumatoid factors (RFs), have long been implicated in the pathogenesis of RA. Circulating lymphocytes from patients with RA recognize oxidatively modified IgG in vitro by initiating a proliferative response and secreting IL-2 [48]. Reactive oxygen and nitric oxide products secreted by inflammatory cells generate covalent crosslinked IgG aggregates with biologic properties of true immune complexes [49].
Citrullinated proteins and peptides — As discussed above, anti-citrullinated protein antibodies (ACPAs) are relatively specific for RA [50]. Citrullination is catalyzed by peptidyl arginine deiminase; arginine residues on alpha enolase, vimentin, fibrin, and fibrinogen, as well as many other proteins, occur due to deimination within rheumatoid joints [51-54]. Intracellular citrullinated proteins colocalized with the deamidase in most of RA synovial samples [52]. However, citrullinated proteins are also found in the synovium of other forms of arthritis, in non-synovial tissue from patients with RA (eg, pulmonary rheumatoid nodules), in the lungs of patients with interstitial pneumonitis, in brain from patients with multiple sclerosis, and in normal brain [55,56]. Synovial citrullinated peptides are also found in most animal models of arthritis.
Comparisons of the SE frequencies on HLA-DRB1 alleles in healthy populations with RA patients who do or do not harbor ACPAs have shown that the SE is associated only with ACPA-positive disease and not with ACPA-negative disease. This indicates that the HLA-DRB1 alleles encoding the SE do not associate with RA as such, but rather with a particular phenotype, disease with ACPA [57]. (See "Biologic markers in the diagnosis and assessment of rheumatoid arthritis".)
A strong association between cigarette smoking, a known risk factor for RA, and the presence of HLA-DRB1*0404 or other HLA alleles comprising the SE has been noted for RA patients who have ACPA [58,59]. One epidemiologic study showed that the relative risk of developing RA is increased 20-fold in those who had two alleles for the SE, had ever smoked cigarettes, and were ACPA-positive [58]. Citrullinated proteins were present in the bronchoalveolar lavage fluid from the lungs of cigarette smokers, but were not demonstrated by immunostaining of fluid from nonsmokers. This study connects two important risk factors for RA, namely smoking and genetic predisposition conferred by carriage of the SE. It also raises the possibility that smoking-induced citrullinated proteins may serve as a link in the process, possibly as neoantigens.
The lack of an association between smoking and risk of RA in those who are ACPA-negative suggests that these disease subsets (ACPA-positive versus ACPA-negative) differ in their pathogenesis. However, a large collaborative study of Caucasian RA patients from North America confirmed a strong association between the presence of ACPA and the shared epitope, but found only a weak association between ACPA formation and smoking [60]. On the other hand, another study demonstrated a moderately strong association between ACPA and tobacco exposure, irrespective of the presence of the SE [61]. The most likely explanation for the propensity to form ACPAs in individuals with the SE is the avid binding of the altered protein to the RA-associated MHC molecules.
Carbamylated proteins and peptides — The citrullination system has been widely studied in RA, but it is probably one of many mechanisms that can create neoepitopes that predispose to the development of inflammatory arthritis. In addition, antibodies that recognize carbamylated proteins have subsequently also been discovered in patients with RA. Rather than converting arginine to citrulline, stressed cells can also convert lysine to homocitrulline. In one study, about 30 percent of ACPA negative patients with RA have detectable levels of anti-carbamylated protein antibodies [62]. Like ACPAs, these antibodies can also precede the development of clinical disease in RA [63]. The MHC associations for this subset will probably differ from the traditional ACPA positive RA if the carbamylated peptides bind more avidly to different alleles.
Glucose-6-phosphate isomerase — The serendipitous discovery of a strain of T-cell receptor transgenic mice that spontaneously develop chronic arthritis with histologic features similar to RA led to identification of a widely distributed enzyme, glucose-6-phosphate isomerase (GPI), as an autoantigen [64]. Antibodies to this protein probably do not play a major role in RA, although they provide lessons on how autoreactivity against a ubiquitous antigen might result in joint specific inflammation. In this case, the protein probably coats the surface of cartilage and is displayed in a way that optimizes complement fixation in the presence of antibody. Anti-GPI antibodies are present in a relatively low percentage of RA patients [65].
Some data suggest that the prevalence of anti-GPI antibodies may correlate with extraarticular manifestations of RA. As an example, in a study of 131 Dutch patients, the proportions with anti-GPI antibodies were as follows: in uncomplicated disease: 5 percent; in those with rheumatoid nodules: 18 percent; rheumatoid vasculitis: 45 percent; and Felty syndrome: 92 percent [66].
B LYMPHOCYTES — Rheumatoid factors (RFs) are immunoglobulins (Ig) reactive against epitopes on the Fc portion of IgG (see 'Immunoglobulin G' above). Compared with those with "seronegative" rheumatoid arthritis (RA), patients with polyarticular symmetrical arthritis who have a persistently positive test for RF are likely to have more erosions of bones and joints, more extraarticular manifestations, and worse function [67]. (See "Clinical manifestations of rheumatoid arthritis".)
The production of RFs results in part from the help provided from a specific subset of T cells to rheumatoid factor precursor B cells. Since T cells reactive with autologous IgG have not been identified in patients with RA, it is likely that these T cells react with antigen(s) and then bind to specific B lymphocytes which proliferate.
One such putative antigen is p205, a protein of unknown function that is present in synovium and synovial fluid. It has been reported to be a very effective stimulator of T-cells of patients with RA [68]. p205 possesses sequences of peptides with a high degree of similarity to those of the 3rd and 4th constant regions of the heavy chain of immunoglobulins [69]. This "molecular mimicry" may promote the development of RFs.
Interleukin (IL)-10 is generally considered to be a Th2 cell product that, in concert with IL-4, can downregulate inflammatory arthritis (see 'Cytokine networks in rheumatoid arthritis' below). IL-10 is a potent B cell-stimulating factor, in addition to its cytokine-suppressing activity. The levels of this cytokine are elevated in RA, although it is probably made by macrophages in the synovium. Thus, the cytokine milieu could contribute to activating B cells by inducing their proliferation and immunoglobulin production in the absence of specific antigen [70,71]. Other cytokines, like IL-6, BLyS (B lymphocyte stimulator, also known as BAFF, B cell activating factor), and APRIL (a proliferation-inducing ligand), are present in RA synovium and can influence the differentiation and activation of B cells [72].
As discussed in more detail elsewhere, anti-citrullinated peptide antibodies (ACPA) and possibly anti-carbamylated peptide antibodies are relatively specific for RA (see "Biologic markers in the diagnosis and assessment of rheumatoid arthritis", section on 'Anti-citrullinated peptide antibodies') Citrulline is formed by the action of peptidylarginine deiminases (PADIs) on peptidylarginine. PADIs are expressed at high levels in inflamed rheumatoid synovium; one major substrate appears to be fibrin. The immunoreactivity of citrullinated fibrin with IgA and IgM in the RA synovium and the colocalization of PADI and citrullinated peptides supports the notion that citrullinated fibrin is a potential autoantigen of RA [52,73].
ACPAs are present in the earliest stages of disease in almost 70 percent of rheumatoid patients [74]. B cell precursors for ACPAs are present in patients with RA and in healthy controls and can be stimulated by activation to produce these antibodies [75]. Thus, it follows that such deiminated antigens formed in the synovium could be involved in driving the local antibody response.
Affinity maturation of autoantibodies — Somatic rearrangement of germline genes is a crucial factor in the eventual development of RFs with a higher affinity for IgG than the original RFs. This "affinity maturation" is probably an antigen-driven process [76]. During this process, RFs change from being polyreactive with a low affinity for many antigens (eg, insulin, tetanus toxoid, beta-2-microglobulin) to being monoreactive with a much higher affinity for human Fc [77]. In addition, the tight complexes between monoreactive RF and IgG can activate complement. Additional polyreactive RFs are generated as a secondary response, and via reactivity with epitopes such as type II collagen, may focus and enhance joint inflammation.
Rheumatoid factor and immune complexes — Another factor that amplifies the inflammatory potential of RFs is the propensity for IgG RF to self-associate into large lattice-like complexes. These complexes can be found in all tissues of the rheumatoid joint, and may help concentrate additional material within this structure. As an example, within the superficial layers of articular cartilage in rheumatoid joints are RF-IgG complexes, antibodies against native and denatured type II cartilage collagen, and activated components of complement [78]. Cartilage destruction is facilitated by proteolytic degradation of matrix components by enzymes such as neutrophil elastase [79]. In addition, immune complexes isolated from synovial fluids may stimulate the production of tumor necrosis factor (TNF) from monocytes/macrophages [80]. (See 'Cytokine networks in rheumatoid arthritis' below.)
Galactose and IgGs — Although no data implicate it as a causative factor for RA, IgG from rheumatoid patients have fewer galactose residues at certain sites in the constant region at the Fc position. This finding is a probable consequence of a B cell deficiency of galactosyl-transferase [81,82], and may be associated with more active disease [83].
Clinical effect of B lymphocyte depletion — Additional support for a role of B lymphocytes in the pathogenesis of RA comes from beneficial effects on joint inflammation observed in clinical trials of B lymphocyte depletion. A more detailed discussion of the clinical application of B cell depletion is presented elsewhere. (See "Rituximab and other B cell targeted therapies for rheumatoid arthritis".)
Synovial biopsy studies before and after rituximab therapy show that synovial B cell depletion is less effective than in the blood and that tissue depletion is necessary but not sufficient for clinical efficacy [84]. The mechanism of benefit is uncertain, as B cell depletion in the tissue does not correlate well with local autoantibody production or cytokine expression. It is possible that the benefit of B cell depletion in RA relates instead to other B cell functions such as antigen presentation, especially in central lymphoid organs.
ANGIOGENESIS AND INFLAMMATORY CELL RECRUITMENT
New blood vessel growths — One of the earliest histopathologic responses in rheumatoid arthritis (RA) is the generation of new synovial blood vessels. This event is accompanied by the transudation of fluid and the transmigration of both lymphocytes into the synovium and of polymorphonuclear leukocytes into the synovial fluid. In the mature RA synovium, the mass of tissue is too much for even the multiple new capillaries to nourish, and local tissue ischemia is the result. The mean PO2 in rheumatoid synovial fluid is usually 30 mmHg, and occasionally is less than 15 mmHg.
Relative synovial hypoxia is associated with an increased production of the transcription factor hypoxia-inducible factor 1 (HIF-1) that activates transcription of genes that are of fundamental importance for angiogenesis, including those for vascular endothelial growth factor (VEGF) and the VEGF receptor [85].
Without new blood vessels, there would be no nutrients to support the highly catabolic synovium in RA [86-88]. The angiogenic response is precise, reproducible, and requires a series of angiogenesis factors that are produced in the rheumatoid joint [89]. New capillary formation is induced and/or stabilized by the transcription factor HIF-1 and by angiogenic factors generated by synovial cells. Such factors include:
●Heparin-binding growth factors (HBGF) [90-92]
●Macrophage angiogenic factor (MAF) [93]
●Vascular endothelial growth factor (VEGF) [92,94]
●Prostaglandins E1 and E2 [95]
●Interleukin (IL)-8 [96]
●Epithelial neutrophil activating peptide 78 (ENA-78) [97]
●Angiopoietin-1 [98]
●Heparanase [99]
When antibodies against IL-8 and ENA-78 were added to rheumatoid synovial tissue in vitro, virtually all the angiogenic activity in these explants was inhibited [96,97].
The enhanced expression of specific adhesion molecules, such as the integrin alpha-v beta-3, is also required for effective angiogenesis [100].
One of the key proinflammatory cytokines, tumor necrosis factor (TNF) may indirectly stimulate angiogenesis. As an example, endothelial cells exposed to TNF upregulate the production of an angiopoietin 1 receptor (Tie2) [101]. A complementary in-vitro finding is that the amount of angiopoietin 1 produced by cultured synovial cells is also increased by exposure to TNF. (See 'Cytokine networks in rheumatoid arthritis' below.)
Balancing this angiogenic response are factors which tend to inhibit neovascular formation. These include interferon-gamma (IFN-gamma) [102], transforming growth factor-beta (TGF-beta), IL-1 [103], angiostatin [104], endostatin [92,105], and low-molecular weight substances in articular cartilage (one of the few avascular tissues in the body) [89]. While anti-angiogenesis approaches are effective in pre-clinical models of arthritis, at least one therapeutic intervention (anti-alpha-v integrin antibody) has not been pursued after a phase II study [106]. It is possible that other pathways are responsible or that therapeutic approaches that focus on blood vessel growth need to be used in combination with other agents.
Cell migration — As the new vessels develop, cytokines produced in the synovium in response to the driving force of TNF (including IL-1, IL-6, IFN-gamma, and substance P) activate endothelial cells to produce adhesion molecules, such as intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1), P-selectin, and E-selectin [107,108]. These cellular adhesion molecules expedite activation-dependent sticking of leukocytes, thereby facilitating diapedesis and extravasation into the synovium. It appears that IL-15 and IL-18 have a major role in stimulating production of TNF, which, as mentioned above, has a broad capacity to trigger biosynthesis of multiple effector proteins.
Although there appears to be redundancy within the cellular adhesion system, certain pairs of ligand/receptors are very active in RA. (See "Leukocyte-endothelial adhesion in the pathogenesis of inflammation".)
Once binding occurs, chemokines, including CXCL8/IL-8 and CXCL5/ENA-78, are involved in enhancing transmigration of leukocytes [107,109,110]. Accumulation of mononuclear cells, particularly T lymphocytes of the Th1 subtype, may involve an interaction between CXCL16 derived from synovial macrophages and a receptor (CXCR6) [111,112]. CCL2 (MCP1) has also been implicated as a key factor that recruits monocytes into the rheumatoid synovium [113].
Although much attention has been focused on the presence of lymphocytes in the synovium, similar lymphoid accumulation also occurs in the juxtaarticular bone marrow. This was illustrated in a study of bone sample and synovium from patients with RA obtained at the time of joint replacement surgery [114]. The degree of inflammatory infiltration of the marrow was correlated with the density of osteoclasts. Increased osteoclast number and activity may account for the radiographic finding of juxtaarticular demineralization that occurs in RA.
Targeting chemokines has been attempted in RA, including anti-chemokine antibodies or chemokine receptor antagonists, but few have been successful. This is most likely due to the redundant nature of the chemokine system, which makes it difficult to block cell recruitment. In one phase II trial, an anti-CXCL10 antibody demonstrated efficacy when used in combination with methotrexate [115]. Multiple small molecule chemokine receptor antagonists have failed to demonstrate efficacy, including compounds that block CCR2 and CCR1.
Other potential approaches might include blocking signaling molecules that regulate a broad range of chemokine functions, such as PI3Kgamma [116]. Chemokine bind to G-protein-coupled receptors, which then activate PI3Kgamma. This kinase serves as the convergence point for multiple chemokines and is an attractive target to overcome the redundant chemokine system. In pre-clinical models, PI3Kgamma blockade is effective in RA and systemic lupus erythematosus (SLE) models [117]. Strategies that block PI3Kgamma, often in combination with another isoform that is required for B cell activation (PI3Kdelta), might be useful for a number of inflammatory diseases, but a phase 2 randomized trial failed to show sufficient efficacy of a dual PI3Kgamma/delta inhibitor [118]. It is not clear whether the lack of evident benefit was related to inadequate exposure to the compound or whether the mechanism is not critical to the pathogenesis of disease.
CYTOKINE NETWORKS IN RHEUMATOID ARTHRITIS
Role of cytokines in synovitis — Autocrine and paracrine communication through the elaboration of proinflammatory cytokines play a key role in initiation and perpetuation of rheumatoid arthritis (RA). (See "Role of cytokines in the immune system".)
A cascade network of cytokines has a pivotal role in synovitis, including granulocyte macrophage colony-stimulating factor (GM-CSF), interleukin (IL)-2, IL-15, IL-13, IL-17, IL-18, interferon-gamma (IFN-gamma), tumor necrosis factor (TNF), and transforming growth factor-beta (TGF-beta):
●TNF – TNF is a key cytokine in a subset of patients. It induces GM-CSF production by antigen-presenting cells; enhances proliferation of T cells; increases proliferation and differentiation of B cells; induces expression of adhesion molecules on endothelium; generates expression of collagenase, stromelysin (matrix metalloproteinase-3), and prostaglandins by synovial cells [119]; and may interfere with antigen-specific and nonspecific suppressive effects of regulatory T cells [120]. Although synergistic in many roles with IL-1 beta, TNF induces synthesis of IL-1 beta in synovial cells, whereas the reverse is not true. The administration of anti-TNF antibodies or a TNF receptor fusion protein decreases the activity of the disease. In addition, anti-TNF therapies produce a favorable decrease in the ratio of two osteoclast modulating factors: receptor activator of nuclear factor kB ligand (RANKL) and osteoprotegerin (OPG), a soluble receptor that competitively binds RANKL [121]. The resulting decrease in RANKL/OPG ratio is consistent with a decrease in local bone resorption. (See "Cytokine networks in rheumatic diseases: Implications for therapy", section on 'Inhibition of cytokine action' and "Normal skeletal development and regulation of bone formation and resorption".)
●IL-1 Family members: IL-1, IL-18 and IL-33 – IL-1 is a critical cytokine that was originally defined by its ability to induce fever, but is involved in almost every phase of immune responses [122]. In addition to regulating innate immunity and T cell differentiation, it also plays a key role in matrix regulation as a potent inducer of matrix metalloproteinases. IL-18 and IL-33 are members of the IL-1 family and can sustain the Th1 phenotype that is associated with RA. The former is synergistic with IL-12 and IL-15 in production of proinflammatory cytokines. Unlike many such destructive cytokines, IL-18 may actually inhibit osteoclast formation [123]. It may drive the local production of TNF and IL-1 beta, as a result of its ability to augment monocyte activation [124]. Expression of IL-18 in rheumatoid synovial tissue correlates with the acute-phase response [125]. IL-33 is an “alarmin,” which is a class of inflammatory mediators that alert cells to tissue damage, and is expressed in RA synovium. Blocking this cytokine is effective in animal models of arthritis [126]. Another alarmin, high-mobility group box protein 1 (HMGB1) has also been implicated in synovial inflammation, and blocking this mediator suppresses disease severity in animal models of arthritis [127].
●Colony-stimulating factors:
•GM-CSF – GM-CSF, produced constitutively by synovial macrophages, may be sufficient to activate expression of human leukocyte antigen (HLA)-DR molecules on antigen-presenting cells (APCs), thereby initiating T cell activation in the presence of antigen or superantigen and the beginnings of an immune response [128]. Clinical trials have confirmed that blocking GM-CSF with monoclonal antibodies decreases disease severity in RA [129,130].
•Macrophage colony-stimulating factor (M-CSF) – This cytokine plays a major role in osteoclast development. In the presence of parathyroid hormone, vitamin D, and prostaglandin E2 (PGE2), osteoblasts produce Receptor Activator for NfkB Ligand (RANKL) that, in the presence of M-CSF, binds to its receptor on monocyte/macrophages and induces these cells to become osteoclasts [131].
●Prototypical Th1 cytokines IL-2 and IFN-gamma – IL-2 and IFN-gamma, the characteristic lymphokines generated by activated T cells, are found in small quantities in rheumatoid tissues [132]. These prototypical Th1 cytokines are present in much higher concentrations in established Th1-mediated diseases, such as tuberculous pleuritis. Instead, cytokines associated with Th17 cells appear to play a more prominent role (see below).
●Prototypical Th2 cytokines IL-13 and IL-4 – Interleukin-13 (IL-13) and IL-4 are prototypical Th2 cytokines that play a role in antibody class switching and allergic diseases [133]. Both cytokines are either absent or present in only very low concentrations in RA synovium.
●Prototypical Th17 cytokine IL-17 – IL-17 is produced by Th17 cells, and, in human synovial tissue in culture, added IL-17 enhanced IL-6 production and collagen destruction. Cyclooxygenase-2 (COX-2) expression and prostaglandin E2 production by cultured fibroblast-like synovial (FLS) cells and freshly collected monocytes are increased by exposure to IL-17 [134,135]. This cytokine also increased bone resorption by enhancing osteoclast activation and decreased bone formation [136]. Increased levels of IL-17 in the synovium of patients with RA may be associated with an increased risk of radiographic progression despite treatment with traditional (nonbiologic) disease-modifying antirheumatic drugs (DMARDs) [137]. Of interest, a significant portion of synovial IL-17A is actually produced by mast cells [138].
●IL-15 – IL-15, produced by macrophages, can induce TNF production through activation of synovial T cells in RA via a cell-contact-dependent mechanism [139]. Local IL-15 production can lead to increased TNF synthesis in an autocrine and antigen-independent fashion. IL-15 RNA is also abundant in synovial specimens from patients with early RA [140]. IL-15 neutralization has some efficacy in RA, but less than most other available biologic agents [141].
●TGF-beta – TGF-beta can be considered a "reparative cytokine." As a member of the growth factor family, this cytokine inhibits T cell activation and proliferation, downregulates B cell proliferation and differentiation, inhibits biosynthesis of metalloproteinases (except for activation of biosynthesis of IL-1 beta by chondrocytes), protects articular cartilage from the degradative influences of IL-1, inhibits TNF secretion from macrophages, and accelerates wound repair [142].
Systemic administration of TGF-beta diminished acute and chronic phases of inflammation in experimental arthritis [143]. As yet, however, there have been no therapeutic trials of recombinant TGF-beta in patients with RA, in part because active TGF-beta (in contrast to the latent form) has a very short half-life in vivo.
Cytokine signaling — There are several well-described mechanisms involved with cytokine receptor ligation and subsequent cell activation. Many (but not all) of the cytokines implicated in RA signal via the Janus kinases (JAKs). The system includes four isoforms (JAK1, 2, and 3 and Tyk2), which then phosphorylate the signal transducers of activators of transcription (STATs) [144]. JAKs have been successfully targeted in RA, demonstrating their importance to cytokine function in the disease. One of these agents, tofacitinib, blocks JAK1, 2, and 3 and is approved for use in the United States for RA. A synovial biopsy study showed that tofactinib-induced decreases in phospho-STAT1 and -3 are good predictors of subsequent clinical response [145]. These two STATs are downstream of JAK1 and play a role in interferon (IFN) and IL-6 signaling. Because IL-6 inhibitors are effective in RA, it is possible that the mechanism of action for JAK inhibitors is related to blocking IL-6 function.
TISSUE REACTION AND MATRIX REMODELING
Fibroblast-like synoviocytes in the intimal lining — The relative contribution of resident fibroblasts in the synovial intimal lining, fibroblast-like synoviocytes (FLS), is now well-defined [146,147]. Data from a K/BxNmouse model indicate that far from serving as an innocent bystander in immune-mediated arthritis, the synovium regulates the entry and behavior of inflammatory cells and also its own capacity to damage specific parts of its environment. A membrane protein known as cadherin-11 appears to play a critical role in forming the intimal lining and promoting self-aggregation of FLS [147,148]. For example:
●Cadherin-11 mediates the organization of the intimal lining, including the migration of macrophages into the lining.
●Cadherin-11 mediates the migration of FLSs over the articular cartilage, leading to damage.
●In the absence of cadherin-11, the synovium is disorganized and demonstrates diminished production of extracellular matrix molecules.
●Mice deficient in cadherin-11 have decreased arthritis severity.
The impact of multiple cytokines, induced via transcription factors such as nuclear factor kappa B (NF-kB) and MAP kinase, could potentially imprint more aggressive behavior by rheumatoid arthritis (RA) synovial lining cells [149,150]. This phenotype also arises in part by the induction of Fos and Jun (by NF-kB) which act as transactivators and induce the biosynthesis of metalloproteinases that destroy cartilage, tendons, and bone [151]. Mitogen-activated protein kinase (MAPK) cascades are involved in inflammation and tissue destruction in RA, as well [152]. In particular, c-Jun N-terminal kinase (JNK) is a critical MAPK pathway for interleukin (IL)-1-induced collagenase gene expression, and therefore could be an important therapeutic target for RA [153]. (See "Principles of molecular genetics", section on 'RNA transcription'.)
MAPK activators (eg, c-Raf-1) and one of its downstream transcription factors, c-Myc, appear to act additively to induce the growth and invasiveness of rheumatoid synovium [154]. Other cytokine-independent pathways of synovial cell activation include endogenous retroviral elements and Toll-like receptors connected by a complex network of autocrine and paracrine-acting factors [155].
Although exhibiting some evidence reminiscent of transformed cells, the synoviocytes are distinguished from truly transformed cells by the fact that they are not immortalized. However, they have a powerful capacity to invade connective tissue of cartilage and tendon, may stimulate the differentiation and activation of osteoclasts, and can migrate from joint to joint in pre-clinical models [156-158].
Unlike synovial fibroblasts from healthy people, the cells from patients with RA, when transferred to immunodeficient mice, invade and destroy cartilage and bone. Despite evidence for damaged DNA in synovial cells, sensitive techniques only rarely show evidence for apoptosis (programmed cell death) in synovial lining cells [159]. Limited apoptosis may contribute to the synovial lining cell hyperplasia in the rheumatoid tissues. This possibly is related to somatic mutations in p53 tumor suppressor, a key regulator of DNA repair and cell replication [38]. The loss of p53 function enhances synoviocyte proliferation in animal models and increases the invasiveness of inflamed synovium into cartilage explants [160].
When fibroblast-like synovial cells are cultured on articular cartilage, matrix metallopeptidase (MMP)-13 (collagenase 3) is produced by the cells; this may be the mechanism by which the rheumatoid synovial pannus is attracted to and begins invading cartilage at the periphery of inflamed joints [161] (see 'Cartilage destruction' below) MMP-13 has close homologies of structure and similarities of action with the first collagenase described, MMP-1. Because of its avidity for type II collagen as a substrate, however, and its induction by cell attachment to cartilage, it may be the primary MMP involved in joint destruction. MMP-13 is produced by synovial fibroblasts, not macrophages [162]. A chemokine, stromal cell-derived factor (SDF)-1, which is produced in RA synovium, increases the secretion of MMP-13 in cultured human chondrocytes [163].
As with many other biologic systems, there is a metalloproteinase cascade. For example, collagenase-1 (MMP-1) is released from cells as a pro-enzyme. Subsequently, MMP-3 (stromelysin), which has no collagenolytic activity but can degrade multiple other substrates including break-down products of collagen and type IV (basement membrane) collagen, activates procollagenase to collagenase, often in concert with plasmin [164].
Cartilage destruction — Activated rheumatoid synovium eventually destroys cartilage at the cartilage-pannus junction. Cartilage can be damaged by direct invasion by the synovial cells, most notably by aggressive RA FLS. Using electron microscopy, a 0.5 to 1.0 micron amorphous zone can be found between the cell processes and relatively intact cartilage, indicating that proteases released from the invasive cells are bringing about the degradation of the glycosaminoglycans and collagen [165]. Proteoglycans (chondroitin sulfate, keratan sulfate, etc) are degraded early in the inflammatory process; the loss of collagen takes longer. Rare are areas of rapid chondrolysis; these areas include synovial lining cells which have phagolysosomes containing engulfed fragments of cartilage collagen [166]. In addition, chondrocytes can produce proteases that can degrade cartilage from within, in addition to the exogenous destructive effects of pannus and proteases produced by neutrophils in synovial fluid.
After a critical mass of synovitis has accumulated within a joint, the synovium invades the edge of cartilage at the joint capsule. Some mechanisms for this invasion include:
●Immune complexes, including antibodies to citrullinated peptides, deposited with activated complement components within superficial layers of cartilage that attract synovial tissue [167].
●Inflammatory mediators (principally metalloproteinases) deplete the superficial layer of proteoglycans in cartilage, thereby exposing collagen and fibronectin that readily bind synovial cells to the residual cartilage matrix and induce MMP-13 production [161,168].
●The strong reactivity of T lymphocytes for chondrocyte membranes (exposed by action of proteases on the cartilage) may attract these cells along with accompanying connective tissue cells and synovial matrix [169].
●Gene transfer studies in vitro have shown that a cell surface-targeted plasmin inhibitor can retard synovial fibroblast-dependent cartilage degradation, confirming the importance of the plasminogen activation system in synovial fibroblasts in the invasion of articular cartilage in RA [170].
●Fibroblast-like synoviocytes that express cadherin-11 bind to and invade into cartilage. Animal models of arthritis show that depletion of these cells prevents cartilage damage even though bone erosions continue to progress.
Simultaneous with degradation of cartilage is the cellular destruction of subchondral bone. Osteoclasts activated by synovial cytokines, including RANKL and cathepsins B, K, and L, add to the destructive capacity of metalloproteinases in destroying bone. The role of immune cells in osteoclast activation is now widely recognized. Activated T cells and bone marrow stromal cells produce receptor activator of nuclear factor kB ligand (RANKL), that is essential for the differentiation, activation, and survival of osteoclasts [171]. IL-18, mainly produced by macrophages in RA, has similar proinflammatory effects as does tumor necrosis factor (TNF)-alpha, and also induces RANKL [172].
Monocyte migration is also enhanced by RANKL an effect that could favor the accumulation of osteoclast precursors in the synovium [173]. A modulator of this is osteoprotegerin (OPG), a soluble receptor that competitively binds RANKL. An anti-RANKL antibody decreases bone erosion in RA, but has no effect on clinical signs and symptoms or cartilage damage in RA [174].
Metalloproteinases — The destruction of cartilage, bone, and tendons in RA is initiated largely by metalloproteinases. As noted above, stromelysin (metalloproteinase-3, MMP-3) is an important protease because it degrades cartilage proteoglycans, fibronectin, and type IV collagen in basement membrane, and activates collagenase. Patients who are homozygous for a polymorphism in the promoter region of the gene for MMP-3 have more rapid radiographic progression of erosive disease than those without this genotype [175]. The role of MMP-13 has already been discussed (see 'Fibroblast-like synoviocytes in the intimal lining' above) Macrophage elastase (MMP 12) and mRNA for this MMP are elevated in rheumatoid synovium in comparison with osteoarthritis tissues [176].
These enzymes are generated principally by the proliferating synovial lining cells at the invasive margin of synovium. Invasive tenosynovium may produce larger amounts of metalloproteinases than adjacent but noninvasive synovial tissue [177]. Increased expression of receptors for immunoglobulin (Ig) is associated with enhanced production of these enzymes [178]. Despite abundant evidence suggesting that metalloproteinases are critical for matrix destruction, several inhibitors have been ineffective in RA. Stromelysin-deficient mice also have normal amounts of bone destruction in animal models of arthritis [147].
Glycosidases and aggrecanase — Enzymes that are capable of removing portions of the carbohydrate side-chains of glycosaminoglycans (glycosidases) might also play a role in cartilage destruction in RA. In vitro, exposure of human articular cartilage to glycosidases results in depletion of glycosaminoglycans to a similar extent as does digestion with MMPs 1 and 9 [179]. ADAMTS-4 (a disintegrin and metalloproteinase with thrombospondin motifs 4) is the major aggrecanase that degrades the interglobular domain of glycosaminoglycans, producing an aggrecanase-specific neoepitope. It is a valid target for inhibitory molecules for future therapy in joint diseases [180]. Aggrecanase deficiency protects mice from osteoarthritis, suggesting that this might be a better target than metalloproteinases for protecting the extracellular matrix in RA [181].
Bone erosions — The course of joint pathology in RA is one of progressive bone and joint destruction with the absence of any sign of bone repair in response to inflammation [182,183]. At sites of active RA, there is a dramatic imbalance of bone turnover in which local bone resorption outweighs bone formation.
This pattern contrasts dramatically with that observed in other forms of destructive arthritis (eg, ankylosing spondylitis and psoriatic arthritis), which are characterized by new bone formation even as joint destruction progresses [184].
Several major contributors to the destruction of bone in RA have been elucidated:
●Tumor necrosis factor – TNF promotes the destruction of bone by increasing the number of osteoclasts and decreasing the number of osteoblasts at the site of inflammation [185].
●RANK ligand – The RANK ligand regulates osteoclast-mediated destruction of the joint architecture. The demonstration that an anti-RANK ligand antibody, denosumab, inhibits erosions in RA confirms the pivotal role of this regulatory pathway and osteoclasts in bone damage [174].
●Cathepsin K – This protease is active at acidic pH and is unusual for its ability to degrade native collagen fibrils as do certain of the MMP family. It is found in osteoclasts and is induced in osteoarthritic cartilage and in inflamed synovial tissue [186]. Cathepsin K inhibitors suppress bone destruction in mouse models of RA, suggesting that they might have clinical utility in human disease.
●Wnt pathway – The Wnt signaling pathway serves as a master controller of bone formation [187,188]. Wnt signaling represses the differentiation of mesenchymal progenitors into adipocytes and chondrocytes and promotes their differentiation into osteoblasts, the cells that lay down bone.
Wnt also inhibits osteoclast formation by positively regulating osteoblast expression of OPG, a soluble inhibitor of RANKL [184,189,190]. Thus, the Wnt pathway both boosts bone formation by fostering osteoblast activity in RA and impairs bone resorption through inhibition of osteoclasts [191].
●The Dickkopf (DKK) family – The DKK family, particularly DKK-1, is a group of secreted proteins that inhibits the Wnt pathway by binding to the Wnt receptor and a cell surface coreceptor, Kremen-1/2. DKK binding induces internalization of the receptor complex, thereby reducing Wnt signaling [187,188].
In a study of three different laboratory models of RA, a rat monoclonal antibody to mouse DKK-1 was employed to inhibit DKK-1 [184]. In all three models, DKK-1 inhibition protected against bone erosions and structural damage of the joint, even in the presence of ongoing inflammation.
In human disease, DKK-1 levels are elevated in the sera of patients with RA and within inflamed RA synovium. In contrast, serum levels of DKK-1 in ankylosing spondylitis are low compared with those of healthy controls. The well-established demonstration of inhibition of bone erosions by anti-TNF therapy in RA may in part be explained by the observation that circulating DKK-1 levels are normalized by this intervention, with consequential restoration of Wnt signaling [184].
SYNOVIAL FLUID IN RHEUMATOID ARTHRITIS — Synovial fluid characteristics are much different from those in the synovium and represent a complex mixture of lubricants, synovial tissue mediators, and constituents of serum [192]. For example, neutrophils are rarely seen in rheumatoid arthritis (RA) synovium but are abundant in RA synovial effusions. Also, the CD4:CD8 cell ratio is reversed in synovial effusions, with a predominance of CD4+ T cells in the tissue while CD8+ cells are more prevalent in the fluid.
Polymorphonuclear leukocytes are attracted to the joint, penetrate through synovial blood vessels, and quickly move into the joint space. Neutrophils subsequently deposit gene products (released from the cells after being synthesized) and the contents of granules (from three types: azurophil, specific, and "C") in the synovial fluid. These include [193]:
●Myeloperoxidase
●Elastase
●Lysozyme
●Collagenase
●Acid hydrolases
●Matrix metalloproteinases
●Interleukin (IL)-1 beta
●Prostaglandins
●Platelet activating factor (PAF)
●Leukotrienes, including leukotriene B4
These products can potentiate inflammation in the adjacent synovium. In addition, since leukotriene B4 and PAF are among the most potent chemoattractants known, activated neutrophils are able to recruit additional neutrophils in an autocrine fashion [194].
Activated leukocytes also produce oxygen-derived free radicals. The most toxic, hydroxyl ion, is very short-lived. Ferrous iron, abundant in synovium because of microhemorrhages, helps catalyze hydroxyl ion formation and enhances metalloproteinase biosynthesis [195]. Additional reactive species, including N-chloramines and hypochlorous acid (eg, bleach) are also formed, and may have deleterious effects upon cells and proteins [196].
Rheumatoid synovial fluid is an ischemic environment with very low oxygen tension. Hypoxia is an important factor in aggravating the inflammatory lesion in RA through increased production of cyclooxygenase-2 (COX-2)-derived nociceptive eicosanoids, increased production by synovial cells of matrix metalloproteinases (MMPs) [197], and induction of genes regulated by hypoxia-inducible factor (HIF)-1alpha [198].
COMPLEMENT ACTIVATION AND IMMUNE COMPLEXES — Complement activation and its interactions with immune complexes are important in rheumatoid arthritis (RA), especially in synovial effusions and at the cartilage interface. They could also play a role in the initiation of RA through complement fixation by ACPAs in the tissue. Their importance is implied by at least two observations [199]:
●Activated complement components have intrinsic inflammatory activity. (See "Overview and clinical assessment of the complement system".)
●Activation of complement often indicates the presence of immune complexes of sufficient size to independently activate the entire system (see "Complement pathways") Immune complexes can also generate secondary antibody responses towards T cell-dependent antigens, and activate memory B cells and plasma cells.
●Animal models of RA are often complement-dependent and are either prevented or have decreased severity in complement-deficient mice [200].
●Complement proteins are often depleted in RA synovial fluid, suggesting local complement consumption.
Specific receptors for the Fc portion of immunoglobulin (Ig) (eg, Fcg-RI, -RII, and -RIII) are found on most inflammatory cells. These complexes help clear immune complexes from the plasma after the immune complexes have fixed complement.
C3a is a product of C3 activation which increases capillary permeability. C3a is inactivated by an enzyme that cleaves its terminal arginine residue (C3adesArg). Levels of both C3a and C3adesArg are elevated in rheumatoid synovial fluids; furthermore, the levels correlate with C-reactive protein (CRP) levels, erythrocyte sedimentation rate (ESR), and disease activity indices [201]. Complement activation mediated by CRP may be decreased during treatment with infliximab [202]. As mentioned above, the deposition of rheumatoid factor (RF)-IgG complexes along with complement in superficial layers of articular cartilage may be a major force in attracting invasive synovial cells into cartilage [78].
ADDITIONAL INFLAMMATORY MEDIATORS — Additional factors, including nitric oxide, neuropeptides, and arachidonic acid metabolites, may play a contributory role in the pathogenesis of rheumatoid arthritis (RA).
Nitric oxide — Nitric oxide (NO) is an uncharged molecule with an unpaired electron generated by NO synthase that catalyzes the conversion of L-arginine and O2 into citrulline and NO. NO synthesis is induced by cytokines in macrophages, synovial cells, and chondrocytes. Although well-recognized as endothelial-derived relaxing factor, NO has a less well-defined role in inflammatory states. NO may reduce endothelial expression of endothelial adhesion molecules. On the other hand, NO also may interact with superoxide to produce toxic peroxynitrite and, in concert with neutrophils, may be toxic to endothelial cells [203,204].
Articular chondrocytes may produce and respond to NO. Induction of NO synthase (iNOS) and subsequent production of NO is stimulated in vitro by fragments of the matrix protein, fibronectin [205]; an effect that requires the presence of CD44, a receptor for hyaluronan on the chondrocyte surface.
Neuropeptides — Substance P (SP) is one of several neuropeptides that are released during reverse (antidromic) sensory nerve activation within joints [206]. SP can activate macrophages, stimulate B cell differentiation, cause fibroblast proliferation, attract neutrophils, and increase expression of cytokines, prostaglandins, and metalloproteinases. SP acts by binding to the neurokinin-1 receptor (NK-1R). Increased levels of SP are found in the synovial fluid and serum of RA patients, and NK-1R mRNA is upregulated in RA synoviocytes [207]. Continuous external application of capsaicin can deplete SP from nerve endings, and suppress neurogenic inflammation, thereby benefiting occasional patients with joint pain of diverse etiology.
Corticotropin-releasing hormone (CRH) may play a local, proinflammatory role in rheumatoid synovium. Higher concentrations of CRH are found in synovial fluid of patients with RA compared with those with osteoarthritis and a higher density of CRH receptors in synovium is also present in those with RA [208,209].
Calcitonin gene-related peptide (CGRP), in contrast to SP and CRH, may have antiinflammatory properties, and CGRP-containing nerve fiber density is lower in samples of synovium of patient with RA compared with those with OA [210].
Arachidonic acid metabolites — Arachidonic acid (AA) is a cell membrane fatty acid that is oxidized to prostaglandins, leukotrienes, and lipoxins; the enzyme responsible for releasing AA from membrane phospholipids is phospholipase A2 [211]. The wide variety of these lipid mediators in synovial fluid, which include lipoxin A4 and resolvin D5, have been characterized using mass spectrometry [212]. When injected into animal joints, phospholipase A2 causes a marked inflammatory and proliferative synovitis. Phospholipase A2 is increased in rheumatoid synovial fluids and is inhibited by glucocorticoids; in addition, its synthesis is induced by interleukin (IL)-1 and tumor necrosis factor (TNF).
Prostaglandins are generated by the actions of cyclooxygenase (COX) on AA. A major role of prostaglandin E2 in RA is to stimulate periarticular bone resorption early in disease, before the development of erosions of bone or cartilage. In one animal model of erosive arthritis, the effect of prostaglandin E2 on bone and cartilage resorption appeared to be mediated through the PGE2 type 4 receptor [213].
In addition, the prostaglandins enhance the manifestations of acute inflammation in joints. They incite pain, increase vascular permeability, and potentiate inflammatory effects of other mediators.
COX-1 is constitutively expressed by many cells, but COX-2 is inducible by IL-1, TNF, and other cytokines in rheumatoid synovial cells [214]. Knockout mice lacking the COX-2 gene are relatively resistant to the development of collagen-induced arthritis, while those without the COX-1 gene have similar characteristics to wild type mice [215]. More selective COX-2 inhibitors are used in some patients because they do not inhibit the formation of prostaglandins that protect the stomach and duodenum from acid, and are less likely to produce ulcers or upper gastrointestinal bleeding. (See "Overview of selective COX-2 inhibitors".)
5-lipoxygenase also uses arachidonic acid as a substrate. One of its final products, LTB4, is a potent neutrophil chemoattractant (along with IL-8, C5a, and platelet activating factor [PAF]) [216]. However, targeting LTB4 receptors has, thus far, not proven useful in RA [217].
Clotting factors and fibrinolysis — Factors responsible for clotting and fibrinolysis may have a role in RA. In addition to serving as the final activator of clot formation by cleaving fibrinogen to fibrin, thrombin is mitogenic for synovial cells, has angiogenic properties [218], enhances endothelial adhesion molecules and arachidonic acid synthesis, and promotes platelet aggregation [219]. Fibrin itself may facilitate cell growth and adhesion within the synovial pannus. The products of platelet aggregation, including formation of microparticles from platelets, are also implicated in RA. Microparticles can be detected in RA synovial effusions and mediate their proinflammatory functions, in part, through the generation of IL-1 [220].
Serine proteases that are mediators of fibrinolysis, including plasminogen activators and plasmin, may also contribute to cartilage degradation. Plasminogen activator is present in synovial fluid from patients with RA [221]. This results in the generation of plasmin which has a major role in the activation of synovial metalloproteinases.
Thus, minimizing fibrinolysis with various inhibitors may be beneficial in RA. As an example, the administration of tranexamic acid to animals with type II collagen-induced arthritis and patients with RA decreased excretion of collagen crosslink fragments [222]. In addition, the in vitro transfer of a gene for a serine protease inhibitor to synoviocytes from patients with RA decreased the destructive and invasive capacity of such cells when incubated with human cartilage in vitro or in immunosuppressed (SCID) mice [170].
SUMMARY
●The initiation of rheumatoid arthritis (RA) results from a combination of predetermined (genetic) and stochastic (random environmental) events. The human leukocyte antigen (HLA) major histocompatability (MHC) genes are the most important, but many other genes are involved and contribute to susceptibility and severity. The most likely mechanism for the environmental component is repeated activation of innate immunity. There is a propensity in RA for immune reactivity to develop to the neoepitopes created by the protein modification, such as citrullination, that results from environmental stressors like smoking; this leads to the production of anti-citrullinated protein antibodies (ACPA) that could initiate inflammation by fixing complement in the tissues. Like antibodies, levels of multiple cytokines gradually increase in the years before RA symptoms occur. (See 'Overview' above.)
●Once the autoimmune process is established, the synovium in RA organizes into an invasive tissue that can degrade cartilage and bone. The rheumatoid synovium has many characteristics of a locally-invasive tumor. (See 'Overview' above.)
●Activation of innate immunity is probably the earliest process in RA, followed by citrullination, loading of antigen-presenting cells (APCs) with either native or modified proteins in the joint, and then migration to central lymphoid organs. Once there, APCs present an array of antigens to T cells, which can then activate B cells and/or can migrate back to the synovium. It is unlikely that a single “rheumatoid antigen” exists. Instead, a broad spectrum of joint-specific antigens, such a type II collagen or nonspecific citrullinated antigens, is responsible. (See 'T lymphocytes' above.)
●One of the earliest histopathologic responses in RA is the generation of new synovial blood vessels, which is accompanied by the transudation of fluid and the transmigration of both lymphocytes into the synovium and of polymorphonuclear leukocytes into the synovial fluid. As the new vessels develop, cytokines produced in the synovium in response to tumor necrosis factor (TNF) activate endothelial cells to produce adhesion molecules, which expedite activation-dependent sticking of leukocytes, thereby facilitating diapedesis and extravasation into the synovium. (See 'Angiogenesis and inflammatory cell recruitment' above.)
●Autocrine and paracrine communication through the elaboration of a cascading network of proinflammatory cytokines plays a key role in initiation and perpetuation of RA and result in a “transformed phenotype” of synovial lining cells. Inflammatory cells are recruited to the bland synovium by the actions of interleukin (IL)-17A, TNF, IL-1, IL-6, IL-18, vascular endothelial growth factor, alarmins like IL-33 and high-mobility group box protein 1 (HMGB1), and chemokines. Retention in the synovium is facilitated by inhibition of apoptosis and other innate immunity mechanisms by interferon (IFN)-alpha and -beta, IL-15, and TNF. The T cells become organized and activated in the presence of IL-23, IL-27, IL-12, IL-15, IL-18, and chemokines. Simultaneously, the proliferative/destructive component of synovitis is generated by TNF, IL-17, bone morphogenic proteins, and transforming growth factor (TGF)-beta. Mast cell products may also have an important role. (See 'Cytokine networks in rheumatoid arthritis' above.)
●Cadherin-11, a synovial fibroblast membrane protein, mediates the organization and invasion of fibroblast-like synoviocytes (FLS) into synovial tissue. Activated rheumatoid synovium eventually destroys cartilage at the cartilage-pannus junction. The destruction of cartilage, bone, and tendons in RA is initiated largely by metalloproteinases. At sites of active RA, there is a dramatic imbalance of bone turnover in which local bone resorption outweighs bone formation. (See 'Tissue reaction and matrix remodeling' above.)
●The role of B cells is supported by the therapeutic effect of B cell-targeted biologic agents. The mechanism of action for B cells is uncertain, but could involve either pathogenic antibody projection, cytokine production, or antigen presentation. (See 'B lymphocytes' above.)
●The cellular components of rheumatoid synovial fluid differ from those in the synovium. As examples, neutrophils are rarely seen in RA synovium but are abundant in RA synovial effusions, and there is a predominance of CD4+ T cells in the tissue while CD8+ cells are typically more prevalent in the fluid. Products of neutrophils that are released into the synovial fluid can cause considerable damage and can potentiate inflammation in the adjacent synovium. (See 'Synovial fluid in rheumatoid arthritis' above.)
●Complement activation and its interactions with immune complexes are important in RA, especially in synovial effusions and at the cartilage interface. Additional factors, including nitric oxide (NO), neuropeptides, and arachidonic acid metabolites, may play a contributory role in the pathogenesis of RA. (See 'Complement activation and immune complexes' above and 'Additional inflammatory mediators' above.)