Resolution of acute inflammation and the role of the antigen

Acute inflammation is the result of a physiological response to a signal of danger, which can be anything from tissue damage via chemical compounds and allergens to invading pathogens . Numerous immune cells and soluble factors such as complement components, cytokines, chemokines, and coagulation factors effectuate this response. It is controlled and fine-tuned by tissue-specific factors such as endothelial cells or specialized tissue-resident immune and stromal cells. Its primary aim is removal of the danger signal, that is, the triggering agent, be it in the form of clearance of pathogens or wound repair. To keep collateral tissue damage to a minimum and to prevent persistent inflammation, effective shutdown mechanisms exist that limit the inflammatory response, clear effector cells and detritus, and resolve inflammation. Failure of these mechanisms can be fatal, as exemplified by the destructive power of uncontrolled inflammation observed in situations such as sepsis or severe inflammatory response syndromes (SIRS).

In the case of chronic, persisting inflammation, the question arises as to the role of the agent that initially triggers the inflammatory process. Examples from allergy suggest that the trigger is crucial to maintain the response as most allergic inflammatory reactions, whether antibody or T-cell mediated, resolve completely in the absence of allergen. In autoimmune inflammation as seen in RA, agents triggering the inflammatory response are frequently self-antigens. Most self-antigens, however, are cell-specific antigens such as platelet antigens in immune thrombocytopenic purpura, or are expressed in various tissues. This makes their removal difficult, if not impossible. Exceptions are tissue-specific antigens expressed in (nonvital) organs that can be cleared by therapeutic intervention. Indeed, Graves’ disease ceases upon total thyroidectomy or radioactive ablation, or upon immune-mediated tissue destruction itself . In RA, however, citrullinated antigens are present ubiquitously as is IgG, the target of rheumatoid factor (RF). Both RF and anti-citrullinated protein antibodies (ACPAs) are involved in the formation of immune complexes that trigger inflammation, and the presence of both strongly associates with the development of RA . Moreover, genetic variants in genes that encode peptidyl arginine deiminases, the enzymes that generate citrullinated antigens, are risk factors for RA . Thus, it is likely that the continuous presence of antigen and consecutive antibody formation are crucial for chronicity in RA, but it is unknown whether the presence, availability, and amount of citrullinated antigens in synovial fluid (SF) or tissue are associated with active disease. In this context, it is interesting that neutrophils in RA were found to expel citrullinated antigens by the formation of extracellular traps, a process that is triggered by inflammatory cytokines but also by ACPAs themselves . This suggests a vicious circle in which antigen stimulates antibody formation which stimulates antigen formation. Thus, in analogy to allergy, one could postulate that in the absence of antigens, RA would remit.

In principle, however, inflammation can also resolve despite the persistence of the initially triggering agent. In gout, the prototype of acute arthritis, massive influx of neutrophils in response to monosodium urate crystals causes acute inflammation. The formation of neutrophil extracellular traps, however, leads to proteolytic degradation of cytokines and chemokines, which eventually limits inflammation and resolves arthritis despite the persistence of crystals in SF and tissue . In this case, failure to form neutrophil extracellular traps leads to persistent inflammation. Persistent inflammation due to failure to limit inflammation is thus different from the remitting relapsing form of gouty arthritis characterized by intervals free of detectable inflammation and suggests that not the trigger but defective control mechanisms determine chronicity. In fact, albeit infrequent, sustained remission is also observed in RA despite the presence of autoantibodies, and the mechanisms involved in the cessation of inflammation in these patients are poorly understood. Based on the issues discussed, it would be particularly interesting to understand whether the local, excessive formation of citrullinated antigens ceases or changes in these patients prior to achieving remission. It is, however, also possible that the antibodies themselves induce remission, as the potential of antibodies to elicit inflammatory responses depends on variable features such as glycosylation, as discussed further below.

Finally, inflammation can persist in the absence of its initially triggering agent. This is exemplified by persistent local inflammation in chronic obstructive pulmonary disease, which remains active despite the cessation of smoking . This latter observation, however, could involve tissue-remodeling and epigenetic effects operational in long-standing disease that are discussed in more detail below, and it is unclear whether these can also be observed in early disease.

Taken together, continuous presence of the triggering agent plays a crucial role in the persistence of inflammation, especially in allergy but also in diseases in which antibodies target self-antigens.

Role of humoral memory in driving chronicity

Most, but not all, antigens elicit a humoral immune response, that is, the formation of antibodies. Antibodies persist even if the antigen that triggered the immune response has long been cleared. This is a crucial mechanism of protection against pathogens and can be exploited in vaccination. However, it represents an important factor of pathogenicity in autoimmune diseases, in which a breach of tolerance mechanisms leads to the development of antibodies that target self-antigens. As in vaccination, the nature and availability of the antigen and the inflammatory environment in which the immune response develops are important determinants of the features of this response . For some antigens, antibody responses are rather short-lived, requiring repetitive boosting to achieve sustained protection. In principle, however, B-cell memory can persist for a human lifetime, as has been demonstrated in survivors of the 1918 influenza pandemic (the Spanish flu) and in individuals vaccinated against smallpox. Decades after antigenic challenge, both still harbored detectable levels of protective antibodies .

In autoimmune diseases, different patterns of autoantibody responses are observed. While the presence and levels of some autoantibodies are associated with activity of the respective disease (e.g., anti-double-stranded DNA antibodies, anti-neutrophil cytoplasmic antibodies), others are rather stably expressed (e.g., ACPAs, anti SS-A/B antibodies, anti-centromere antibodies) . Whether this reflects the degree of involvement of the individual autoantibody in disease-specific processes is not entirely clear and needs to be determined for each antibody and disease separately. However, it is becoming increasingly clear that autoantibody-secreting cells can be drivers of persistent inflammation and, therefore, could be relevant targets for therapeutic intervention . These antibody-secreting cells, termed plasma cells in their terminal stage of differentiation, compete for survival niches in bone marrow, spleen, or inflamed tissue . In these niches, they produce large amounts of antibodies and reside without proliferation. This latter aspect makes them extremely resistant to therapeutic depletion, as most current treatments such as methotrexate or cyclophosphamide target proliferating cells. Of interest, however, depletion of autoreactive plasma cells has recently been successful in lupus-prone mice by treatment with the proteasome inhibitor bortezomib . This depletion ameliorated disease, a concept that is currently also evaluated in human systemic lupus erythematosus (SLE). In human disease, successful depletion of autoreactive memory can also be achieved by myeloablative treatment regimens in bone-marrow transplantation settings, which is yet another proof of principle that underlines the important role of immunological memory in driving chronic inflammation in autoimmune diseases. In SLE, but also in systemic sclerosis, stem cell transplantation regimens have been able to achieve complete, sustained, drug-free remission in some patients, which are so far the only examples of “cure” . As relapses have also been observed, however, it is obvious that additional mechanisms could be involved in continuously triggering disease or in re-initiating autoantibody formation.

Nevertheless, autoantibodies and autoreactive memory B cells, especially autoantibody-producing plasma cells, are highly relevant players involved in chronicity of inflammatory diseases and, as such, involved in driving the development from UA to RA.

Epigenetic imprinting and role of the tissue environment

In recent years, immunologists increasingly recognize the role of tissue-specific factors for governing and determining immune responses . This includes specialized tissue-resident cells and stromal cells and their interaction with invading immune cells upon their recruitment by danger signals. With regard to chronicity of inflammatory responses, it is important to note that especially stromal cells can create a microenvironment that allows the survival of immune effector cells . As such, they can control whether immune cells remain at the site of inflammation after clearance of the danger signal, as withdrawal of appropriate survival signals will lead to apoptosis of effector cells . In a similar manner, stromal cells play a crucial role in the formation of survival niches for plasma cells and thereby determine their longevity . Also, T-cell development in the thymus strongly involves specialized endothelial cells and thymic fibroblasts. Both secreted factors such as cytokines and extracellular matrix molecules but also cell-to-cell contact of immune cells with stromal cells are relevant. Thus, next to the antigenic trigger and effector molecules such as antibodies produced by immune cells, tissue-specific factors could be involved in the transition of reversible UA to chronic RA. Also here, the question arises whether stromal cells acquire the ability to promote continuous survival of effector cells by, for example, being repetitively stimulated, or whether this is an intrinsic feature determined by, for example, genetic background. In the context of RA, synovial fibroblasts (SFs) have received much attention, as they appear to have a different phenotype and distinct functional properties as compared to fibroblasts derived from non-inflamed joints . Of interest, RA-SF keep these features in culture, that is, when taken out of the inflammatory environment, which suggests that a certain degree of imprinting has taken place prior to or during the inflammatory process. In fact, specific epigenetic modifications such as DNA hypomethylation have been detected in RA-SF, which could explain their sustained activated phenotype by which they not only create a specific environment but also actively participate in joint destruction . As inflammatory cytokines can influence DNA methylation, it is possible that chronicity of inflammation results from a vicious circle in which continuously triggered inflammation leads to imprinting of features in stromal cells that supports the local survival of immune effector cells, which again produce cytokines. Whether interference with autoantibody production before the inflammatory response reaches the target tissue can prevent initiation of this circle is one of the most intriguing questions for future study.

The development of RA

After this description of immunological mechanisms underlying the development of autoimmune disease in general, we would now like to focus on RA. The term “rheumatoid arthritis” was introduced in 1876 by Sir Alfred Garrod in an attempt to counteract the unsatisfactory use of designations such as “chronic rheumatism” and “rheumatic gout” . These terms also illustrate the lack of methods and knowledge in the past to differentiate between diseases such as gout and RA based on their pathophysiology. There is now ample evidence to suggest that RA, characterized by arthritis of mainly the small joints of the hands and feet, is the result of an autoimmune response.

In order to facilitate research into the preclinical and earliest clinically apparent phases of RA, the European League against Rheumatism (EULAR) has issued recommendations on terminology to be used to define specific subgroups during different phases of disease development . There are several risk factors known to be associated with the development of RA, which are described in more detail below in the section on autoantibodies. Furthermore, it is well established that increased acute-phase reactants and autoantibodies which characterize RA can be detected before clinical symptoms appear . The autoantibodies IgM rheumatoid factor (IgM RF) and ACPAs are even present several years before disease onset. Based on these findings, the following terminology has been recommended for differentiating between the different (pre-)clinical phases: (A) genetic risk factors for RA; (B) environmental risk factors for RA; (C) systemic autoimmunity associated with RA; (D) symptoms without clinical arthritis; (E) unclassified arthritis; and (F) RA. Although not every individual in phase A, B, or C may develop RA, and patients will generally present in phase E or F, this nomenclature exemplifies the current thinking about RA development and intends to provide a framework for future research.

Risk factors for progression of arthralgia to (rheumatoid) arthritis

Before the introduction of the nomenclature mentioned above, several studies already attempted to identify factors that could predict whether patients presenting with arthralgia or “early arthritis” would progress to RA. In a cohort of patients with arthralgia who were positive for ACPAs or IgM RF, the presence of ACPAs was highly associated with the development of arthritis . In another cohort of seropositive individuals who had either arthralgia or a positive family history for RA, smoking and being overweight were associated with developing RA . These risk factors are mostly in line with features which predict the development/presence of RA in patients who first present with inflammatory arthritis.

A large-scale attempt to identify discriminating factors for the development of RA (defined as persistent and/or erosive disease) among patients presenting with undifferentiated inflammatory synovitis was undertaken in the process of developing the 2010 ACR/EULAR classification criteria for RA . Based on data from 3115 patients from nine early arthritis cohorts, a list was derived of patient and disease characteristics associated with the initiation of methotrexate treatment within the first 12 months after presentation. This resulted in classification criteria in which especially the pattern of joint involvement (many small joints being most predictive) and the presence of IgM RF and ACPAs were highly predictive and therefore heavily weighted factors.

Autoantibodies and risk factors for autoantibody-positive RA

As discussed above, autoantibodies are the most potent predictive factor for the progression from UA to RA. RF, ACPAs, and anti-carbamylated protein antibodies (anti-CarP) have all been shown to be associated with the development of RA in patients with UA or arthralgia . Of these autoantibodies, ACPAs, commonly measured by the anti-cyclic citrullinated peptide (anti-CCP) assays, have the strongest predictive value. In an attempt to understand the risk factors associated with RA, many studies have focused on risk factors associated with autoantibody status and more specifically ACPA status. This review therefore proceeds with an overview of the risk factors associated with autoantibody-positive and -negative subsets of disease, before continuing with a discussion of the precise (development of) autoantibody characteristics.

Genetic risk factors

The risk of developing RA is known to be influenced by both environmental and genetic risk factors, of which the human leukocyte antigens (HLAs) are the most important. The association between RA and the HLA region was originally described more than 30 years ago . Subsequent studies revealed that several different HLA DR alleles are associated with RA, which led to the formulation of the so-called shared epitope (SE) hypothesis in 1987 . This hypothesis postulates that the observed associations between the HLA region and RA may be based on the fact that all HLA DR alleles which predispose to RA have the same or a similar amino acid sequence (the so-called shared epitope sequence) at positions 70–74 of the HLA-DRB1 molecule. This sequence is located in the peptide-binding groove of the HLA alleles and may therefore be directly involved in the presentation of peptides to arthritogenic T cells. More recently, the association between HLA DR alleles and seropositive RA was narrowed down to three single-amino-acid polymorphisms (at positions 11, 71, and 74) with the use of genome-wide single-nucleotide polymorphism (SNP) data . Despite this refinement on amino-acid level, the classification of HLA DR alleles as SE alleles is still commonly used, and can be useful for grouping alleles which predispose to seropositive RA.

After the first descriptions of ACPA, it soon became clear that the SE alleles are preferentially associated with ACPA-positive RA . Several reports showing that certain HLA SE alleles can bind and present citrullinated peptides have provided a possible biological explanation for the association between ACPA and the SE alleles. A chemical analysis of an induced murine immune response to citrullinated collagen type II demonstrated that the conformational change in the antigen induced by citrulline was critical for recognition by antibodies , again suggesting that citrulline plays an essential role in the initiation of the immune response.

In contrast to the strong predisposing effect of the HLA SE alleles on ACPA-positive RA with, for example, an odds ratio (OR) of 4.3 in a large study from the United Kingdom, a much weaker association with ACPA-negative RA was reported in the same study (odds ratio 1.4) which was mostly limited to the RF-positive ACPA-negative patients . This raises the question of whether there may be residual ACPA reactivities in the ACPA-negative subset which are not detected by the anti-CCP2-test , or whether the SE alleles may predispose to autoantibody-positive disease in general rather than to ACPA-positive RA.

Similar to the preferential risk association between the HLA SE alleles and ACPA-positive RA, the protective effects exerted by the HLA DR locus have also been found to be specific for ACPA-positive RA. In a large European meta-analysis, the HLA DRB1*13 alleles were found to be associated with a significant protective effect with an OR of 0.54 in ACPA-positive RA, after stratification for the SE alleles .

After these seminal findings with regard to the association between the HLA SE alleles and ACPA-positive RA, many of the genetic risk factors for RA have now been found to be specific for either ACPA-positive or ACPA-negative RA. The majority of genetic risk factors, such as PTPN22, the TRAF1-C5 locus, the OLIG3-AIP3-locus, and STAT4, are preferentially associated with ACPA-positive RA . Conversely, there are other genetic risk factors that have been described to be exclusively associated with ACPA-negative RA, such as HLA-DR3 and interferon regulatory factor (IRF) haplotypes .

In the context of ACPA-negative RA, it is also interesting to examine risk factors associated with anti-CarP autoantibodies, as these antibodies occur in part of the ACPA-negative RA population and are associated with more radiographic progression in this disease subset . Intriguingly, a recent study investigating anti-CarP in two large cohorts of RA patients found no striking associations with HLA DR alleles or PTPN22 which were independent of anti-CCP . This suggests that the pathophysiological pathways underlying the formation of ACPA and anti-CarP-autoantibodies are very different.

Environmental risk factors

In addition to genetic aspects, environmental risk factors are known to contribute to the etiology of RA. Many epidemiological studies have shown an association between cigarette smoking and the development of RA, and later studies revealed that this was mainly the case for RF-positive RA . Regarding RF-positive RA, the combined effect of the HLA SE alleles and smoking exceeds the additive effect of these single risk factors; a phenomenon known as statistical interaction . More recent data have shown a striking interaction between smoking and the SE alleles in conferring risk for ACPA-positive RA .

There are other environmental risk factors which have been reported to have differential effects on autoantibody-positive versus -negative RA. Silica dust exposure, for example, is another inhaled agent which has been shown to increase the risk of ACPA-positive RA . Periodontitis has received much attention as a possible risk factor for developing ACPA-positive RA on the basis of several lines of evidence. Epidemiological studies have consistently shown that RA patients more frequently suffer from periodontitis and have more severe periodontitis than controls . Furthermore, one of the bacteria involved in the pathogenesis of periodontitis, Porphyromonas gingivalis ( P. gingivalis ), is the only known bacterium with its own citrullinating enzyme (peptidyl arginine deiminase (PAD)). It therefore appears possible that P. gingivalis might citrullinate host proteins, thus generating citrullinated epitopes against which an immune response could be mounted . Despite the appealing elegance of this hypothesis, there are a number of findings that argue against a direct link between periodontitis and the development of ACPA-positive RA. The prospective Nurses’ Health study could find no association between periodontal disease and future RA . In patients with arthralgia, the presence of antibodies to P. gingivalis did not associate with the presence of ACPA, or the development of RA .

For ACPA-negative RA, parity has recently been described to contribute to the risk of disease in women between 18 and 44 years of age .

Taken together, the studies on risk factors have revealed that the pathophysiological basis of disease differs in autoantibody-positive versus autoantibody-negative RA. Despite the similarities in phenotype at the time of disease onset , this also suggests that the halting of disease progression may require different immunological interventions for these different disease subsets.

Heritability

The recent years have witnessed an immense gain of knowledge concerning the number of genetic risk factors for RA, thanks to the advent of new techniques which have enabled whole-genome association studies and deep sequencing. This has also led to various studies investigating the amount of variation in disease, which can be explained by genetic risk factors, also known as the heritability. Various different methods have been used to this end, which has led to substantial differences in the ensuing estimates. In a study using a large RA twin cohort, the heritability of ACPA-positive versus ACPA-negative RA was reported to be remarkably similar and approximately 66%, although the large confidence interval for ACPA-negative RA prohibited any firm conclusions about this subset . A more recent study from Sweden, on the other hand, used data on affected family members from a very large population-based case–control cohort to calculate the heritability . This yielded a heritability estimate of 50% for ACPA-positive RA and around 20% for ACPA-negative RA. These figures are easier to reconcile with the fact that the majority of genetic risk factors known for RA have been found to preferentially predispose to ACPA-positive RA.

In view of the many genetic risk factors which have now been identified for RA, an obvious question is how much of the total heritability/genetic variance can be explained by these risk factors. Recent studies estimate that all RA genetic risk factors together explain only approximately 16% of the total susceptibility to disease . However, a modeling approach, taking into account the hundreds of uncharacterized SNP associations which exist throughout the genome, has resulted in an estimate of 36% of RA disease risk explained by genetic risk factors . Compared to the 50% of ACPA-positive RA which is presumed to be caused by genetic factors as mentioned above, this would still leave approximately 15% of so-called missing heritability. Where to look for this missing heritability is still controversial. There has been much debate about a possible role of rare variants, but a recent large study seems to have brought this to a close in showing a negligible role for rare coding-region variants in the susceptibility to autoimmune diseases . Thus, the study of heritability in RA has overall provided us with a number for the total contribution of genetic risk factors (for seropositive disease probably 50–60%), but has also raised many questions about where to look for these genetic risk factors.

Multiple hit model for RA

The findings concerning genetic and environmental risk factors and their specific effects for mainly autoantibody-positive RA have led to the development of various models of disease development. This section first describes a popular pathophysiological model for ACPA-positive RA which was put forward several years ago, and then mention some important recent insights, which have led to some important modifications of the classic model.

The pathophysiology of RA can be viewed as a multistage process. In the first stage, a genetic predisposition along with environmental factors results in an adaptive immune response with antigen-dependent T- and B-cell activation. In the case of ACPA-positive RA, genetic risk factors such as the SE alleles together with smoking may lead to the development of an anti-citrullinated protein immune response. The association between smoking and the development of citrullinated antigens in bronchoalveolar lavage fluid cells could provide a pathogenetic link between smoking and the development of ACPA-positive RA. On the other hand, smoking most likely has much more effects than solely leading to citrullination, as is supported by data showing an association between smoking and the number of autoantibodies in RA patients, rather than the presence of ACPAs (unpublished confidential data van Wesemael et al.).

A nonspecific and as of yet unknown trigger could then function as a secondary event leading to joint inflammation, which manifests itself as UA. Other genetic risk factors involved in, for example, cytokine regulation play a role in determining the extent and duration of joint inflammation. The inflammatory process itself can subsequently further stimulate the adaptive immune response through the generation of new epitopes by, among other processes, citrullination, which has been shown to occur more readily in inflamed joints. In individuals who have previously developed an adaptive immune response (possibly with production of ACPA), the immune cells that now gain access to the joints can enhance the inflammation and lead to the increased production of cytokines and soluble inflammatory mediators. Large numbers of activated CD4-positive T cells accumulate in the joints of patients with RA. The abundant presence of these activated T cells in the joints has several different effects. Through production of these cytokines and through cell–surface interactions, the activated T cells stimulate monocytes, macrophages, and SFs to produce other cytokines such as tumor necrosis factor-alpha (TNFα) and to secrete matrix metalloproteinases . Furthermore, the activated T cells stimulate osteoclastogenesis . Finally, the autoreactive T cells are capable of providing help to autoreactive B cells, which can lead to the production of autoantibodies. B cells can also function as antigen-presenting cells and thus can promote T-cell activation and perpetuation of the autoimmune response . Activated B cells and autoantibody-producing plasma cells are present in RA synovium, although it is not completely clear if the original B-cell activation occurs in the joint or elsewhere. ACPAs bound to citrullinated antigens have also been detected in the joints of RA patients .

Autoantibodies may play a pathogenic role by fixing complement in the joint leading to the release of chemotactic factors such as C5a and the subsequent recruitment of inflammatory cells. Another mechanism by which immune complexes can contribute to joint damage is through the activation of monocytes and macrophages by binding to Fc-receptors on the surface of these cells. This could cause perpetuation of the synovial inflammation and progression of UA to RA and erosive disease.

In summary, this model assumes that genetic and environmental risk factors are most important at a very early stage in disease development, namely in the process of developing autoantibodies/anti-citrulline immunity. This is known to occur many years before disease onset, and which factors are involved in the progression from being just autoantibody-positive to having a clinical autoimmune disease, according to this model, remained unclear.

A recent publication using data from the Swedish twin cohort has however shed new light on the role of genetic and environmental risk factors in the multistep model of RA development . Totally, 12,590 twins were analyzed for the presence of genetic risk factors (such as the SE alleles), environmental risk factors (such as smoking), autoantibodies, and the presence of RA. By comparing monozygotic to dizygotic twin pairs, the influence of genetic risk factors as a whole and of the separate risk factors mentioned above could be calculated. Furthermore, the fact that the cohort consisted not only of RA patients meant that these calculations could not only be performed for RA but also for the risk of being autoantibody-positive without having RA. The results were surprising, with most of the risk of being ACPA-positive (without having RA) being determined by stochastic or nonshared environmental risk factors, rather than genetic or shared environmental risk factors.

This means that role of the risk factors described above should be reevaluated. In contrast to previous beliefs, the most potent genetic risk factors, the HLA SE alleles, appear to have their largest effect not during the development of anti-citrulline autoimmunity (and ACPA), but rather during the progression from systemic autoimmunity (manifested by the presence of ACPA) to RA.