Rheumatoid arthritis and the concept of autoimmune disease

Corresponding Author:

Michel Neidhart
Center of Experimental Rheumatology Clinic of Rheumatology
University Hospital Zurich Zurich, Switzerland
E-mail: michel.neidhart@usz.ch

Abstract

We critically review the concept of autoimmunity in Rheumatoid Arthritis (RA). To classify a disease as autoimmune, it is necessary to demonstrate that the immune response to a self-antigen causes the observed pathology. In RA, neither autoantigens nor autoreactive T-cell clones were found. Molecular mimicry exists in some cases, e.g. after Proteus mirabilis infection. In some patients, cross-reactivity between microbial DnaJ and human HSP40 can occur. However, none of the autoantibodies described in RA, with perhaps exception of Anti-Citrullinated Peptides Antibodies (ACPAs), are responsible for the onset of disease and tissue damage. ACPAs appeared early in the disease process and could activate macrophages. In turn, this leads to the release of pro-inflammatory cytokines that stimulate synovial fibroblasts. In this disease, neither autoantibodies, nor leucocytes destroy bone and cartilage. This is caused by aggressive epigenetically modified synovial fibroblasts. Current treatments did not directly target those cells. The term “autoimmune disease” for RA should be used with caution until the link between immune phenomena and joint damage has been elucidated. However, we have to be open and to accept that RA is a heterogeneous disease. This is reflected by the large variability of responses to immunosuppressive drugs or biologicals.

Introduction

Rheumatoid Arthritis (RA) is the most common autoimmune disease with 0.5 to 1% prevalence worldwide. In many textbooks of medicine, RA is presented as a typical systemic autoimmune disease. This is no more so obvious in the latest edition of Harrison’s principles of internal medicine [1,2]. We will review here the reasons. Hallmarks of RA include immune dysfunctions, inflammation, synovial hyperplasia and joint destruction. Arguments in favour of the autoimmune disease concept are production of autoantibodies, genetic association with HLA-DRβ1 polymorphism, release of proinflammatory cytokines and chemokines, infiltration of leucocytes into the inflamed synovial tissue and beneficial effects of antiinflammatory and of some immunosuppressive therapies.

Autoimmunity refers to the presence of autoantibodies or T-cells that react with selfantigens. This doesn’t necessarily imply that the self-reactivity has pathogenic consequences [1]. Physiological autoimmunity is present in all individuals and increases with age. Polyreactive autoantibodies are essential e.g. to remove apoptotic debris. When autoimmunity is induced by an inciting event, e.g. tissue damage, it is in general limited.

Backed on Koch’s postulates of infection, clear criteria were developed to distinguish “autoimmune diseases” from physiological autoimmunity [3-5]. Thus, to classify a disease as autoimmune, it is necessary to demonstrate that the immune response to a self-antigen causes the observed pathology [1]. Classically, autoimmune diseases are attributed to a breakdown of tolerance when lymphocyte clones fail to discriminate between self and nonself [6]. This has necessarily to do with interactions of T-cells and antigen-presenting cells, recognition of one or more defined autoantigens and production of specific pathogenic autoantibodies [1].

Literature Review

Humoral immunity

To establish the role of an autoantibody in disease pathogenesis, it must be [3,4]:

• Detectable in the blood or injured tissue

• Responsible for the onset of disease and tissue damage

• Shown to correlate with the clinical manifestations

• Able to replicate the disease in an experimental model and/or to passively transfer the disease

In RA, prominent autoantibodies are IgMrheumatoid factors (IgM-RF), anti-type II collagen antibodies and Anti-Citrullinated Protein Antibodies (ACPA); none of them fulfil the last 3 criteria. RA is not alone with this problem, for example an important subset of patients with systemic sclerosis are positive for Scl-70 antibodies; in spite of a high specificity, the pathogenic role of those autoantibodies is hypothetical only [7]. This does not keep most authors from calling it an autoimmune disease.

IgM-RF and IgA-RF are diagnostic and prognostic markers, respectively. However, IgMRF can also appear during various infection, e.g. with influenza A virus [8]. ACPA are more specific and useful diagnostic tool to detect early RA [9]; however, ACPA also have been observed in lung diseases and in normal smokers. The naturally occurring epitopes recognised by these autoantibodies are not precisely characterised [10]. In addition, ACPA are very heterogeneous in terms of fine specificity in an individual patient and among different patients.

A natural “experiment” of passive transfer is the pregnant RA patient; however, a successful pregnancy, normal delivery, and a healthy infant is the usual and expected course [4]. Moreover, pregnancy has a known beneficial effect on the course of the disease.

Cellular immunity

Autoimmune diseases occur in those individuals in whom the breakdown of immune tolerance results in self-reactivity causing tissue damage [6]. The HLA-DRβ1 “shared epitope” genetic association is an important argument in favor of an autoimmune disease. However, it explains only 15-20% of the cases. To establish the role of autoreactive T-cells in disease pathogenesis, it must be [1]:

• Shown that T-cells specific for defined autoantigens are found in patients

• Shown that monoclonal expansions of T-cells occurs, as well as a restricted Vβ T-Cell Receptor (TCR) gene usage

• Able to attenuate the disease by an anti-Tcell therapy

• Able to transfer the disease.

Of course, these criteria are extremely dogmatic and can be discussed.

T-cell clones able to recognize type II collagen have been isolated from blood in both healthy subjects and RA patients [3]. Clearly, they belong to physiological autoimmunity, i.e., in humans are not pathogenic and unable to provoke cartilage damage. Hypotheses about antigens recognized by autoreactive T cells have stemmed from data on low affinity autoantibodies found in RA sera or from animal models. In recent years, the discovery of antibodies against peptides that have been modified post-translationally has opened new perspectives. Importantly, an array of citrullinated peptides fit avidly into the HLA-DR binding groove and activates T cells much more efficiently than the native protein [11]. The production of ACPA, however, probably represents a normal adaptive immune response against altered antigens rather than true autoimmunity [12]. Again, it's a question of definition.

Lymphocytes and monocytes infiltrate RA synovial tissues. However, a small minority of T cells only express activation markers that denote local stimulation, a fact consistent with a polyclonal origin of the T-cell infiltrate. The diversity of autoreactivity in RA suggests that it results from a general activation of the immune system [2]. Studies of the T-cell repertoire showed preferential use of the Vβ TCR families but also found that overexpressed TCRs varied across patients [13]. In contrast to the findings in animals, most studies in RA did not support the role of specific Vβ TCR genes [14]. Recently, however, the question reappeared. Thus, in synovial tissues of ACPA-positive RA patients, the T cell repertoire is specifically restricted [15]. However, the origin and role of these clonal alterations remain to be determined.

Adoptive transfers have been performed with RA mononuclear cells into SCID mice. High IgMRF production occurred. However, IgM-RF levels gradually decreased with time and synovial hyperplasia or joint injury were absent. In RA, IgM-RF immune complexes and complement activation are involved in vascular endothelial injury in rheumatoid nodules, but not in joint damage.

Despite the intensive efforts put into studying the effect of depleting and nondepleting anti- CD4 antibodies, no sufficient benefit was observed that would justify pursuing this option in patients with RA [16]. In a specific genetic background (HLA-B27+), asymmetric polyarthritis can even occur in HIV-positive patients [17]. Animal models suggested that a biased differentiation of pro-inflammatory Th1 and Th17 cells might occur. Unfortunately, in humans, an IL-17 receptor antibody, failed to show clinical benefit [2]. A provocative question is whether animal models of arthritis can represent human RA. Of course, It is unrealistic to expect a single model of arthritis to fully recapitulate human disease; however, the severe inflammatory component necessary to induce or to maintain arthritis in certain of those models can be criticized, e.g. the LPS boost in type II collagen-induced arthritis. Moreover, it has been claimed that mouse models may poorly reflect human genomic inflammatory responses [18]. However, this is a matter of debate [19].

Molecular mimicry

In animals, various joint-related molecules, among them type II collagen, glycoprotein gp39 and aggrecans/proteoglycans showed arthritogenic potentials [13]. In addition, in both animals and human, various infectious agents are associated with self-limited reactive arthritides. Molecular mimicry or cross-reactivity between a joint-related molecule or a microorganism and a self-antigen may lead to autoreactive T-cell clones, as e.g. in streptococcus-induced rheumatic fever. Animal models of arthritis have strongly suggested an association of autoimmunity with the human disease. This has caused investigators to study humans to validate an idea generated in mice, rather than vice-versa [5]. Molecular mimicry exists in some cases resembling RA, e.g. after infection with Proteus mirabilis [20]. A hypothesis also exists involving cross-reactivity between human Hsp40 and bacterial DnaJ [21]. At least, animal models showed that different ways can lead to arthritis. Similarly, in humans, the fact that various combinations of genes and types of environmental stress lead to arthritis suggests that we are not looking at a single disease, but at a process with multiple pathways [12]. Human RA should therefore be regarded as a syndrome, rather than a disease.

Neoepitopes

A new concept is the emergence of neoepitopes. Relationships of ACPA with smoking and Pophyromonas gingivalis were suggested by epidemiological studies. P. gingivalis is able to citrullinate arginine residues, thereby producing neoepitopes. The causal question remains at least for non-smoker and patients with proper dental hygiene.

Innate immunity

Alternatively, recent observations suggest that ACPAs directly influence monocytes/ macrophages, inducing M1 polarisation [22] and stimulating the release of IL-1β via phosphorylation of ERK1/2 and JNK pathways [23]. In addition, interferon-gamma released by T-lymphocytes responding to citrullinated neoepitopes can enhance the ability of macrophage to present antigens [24]. Monocyte-derived inflammatory macrophages, which have infiltrated the synovial tissue, are the main producers of pathogenic cytokines, in particular TNFβ, IL-1β and GM-CSF [25]. Recent findings suggest that specific microRNA is responsible for the maintenance of the inflammatory state [26]. An attractive hypothesis is that RA monocytes/ macrophages are “trained” [27].

Clinical clues

RA is designated as an autoimmune disease based on clinical clues for example the chronic inflammation and the favourable response to antiproliferative and immunosuppressive drugs. The number need to treat for new generation biologicals ranged from four to six treated patients to achieve one ACR50 response [28], meaning that a large proportion of RA patients are non-responders. Most efficient are etanercept, tocilizumab, and rituximab [29], suggesting roles for TNFβ, IL-6 and B-cells in subsets of patients. The most persuasive argument that RA has multiple pathways leading to the same phenotype is the diversity of responses to these highly specific immunotherapies [12].

Joint damage

Great efforts have been made in the last decade regarding early diagnosis and treatment of RA; this allows to substantially slow progression of joint damage, thereby preventing irreversible disability. However, it has to be noted that autoantibodies or leucocytes did not directly destroy bone and cartilage. The link between immune phenomena and joint damage has been called “a mystery” [1]. In RA, epigenetically modified synovial fibroblasts have an intrinsic aggressive behaviour and constitute the effector cells of joint destruction [30]. They release matrixdegrading enzymes, but also chemokines and survival signals attracting CD45RO+ memory T-cells and CD5+ IgM-RF-producing B-cells into the inflamed synovial tissue. Normally, synovial fibroblasts suppress T-cell proliferation, but in RA this regulatory mechanism is deficient [31]. This allows the appearance of lymphoid follicles and germinal centre-like structures. RA synovial fibroblasts are further activated by monocytes and macrophages releasing TNFβ and IL-6. Interestingly, transgenic TNF mice (TNFΔARE) back‐crossed to a RAG1−/− background still develop erosive polyarthritis, indicating a key role for this cytokine downstream of T and B cell responses [32,33]. Current treatments- mostly anti-inflammatory- did not directly target RA synovial fibroblasts.

Discussion and Conclusion

In RA, an infectious trigger or an autoantigen associated with T-cell monoclonal activation are unknown, neither autoantibodies nor leucocytes directly cause joint damage, a polyclonal rather than a monoclonal T-cell activation occurs, immunotherapy fails in many cases to cure the disease, and synovial fibroblasts are in fact the key effectors of cartilage and bone destruction. Probably neglected is the contribution of the innate immune system, i.e., epigenetically modified inflammatory macrophages. Taken together, regarding RA, this call for greater caution in the use of the term “autoimmune disease” until the link between immune phenomena and joint damage has been elucidated. This debate will require that we discuss the possible alternative to classify RA if the term “autoimmune” is to be abolished. A term like *autoinflammatory” would underline the contribution of innate immunity to the pathogenesis. It has been suggested that abnormal activation of pattern recognition receptors, like Toll-like receptors and receptors for advanced glycation end products by endogenous danger signals (alarmins/DAMPs) are crucial for perpetuating a vicious inflammatory cycle. However, the term “autoinflammatory” is already used for familial Mediterranean fever, TNF-related periodic fever syndrome, cryopyrin-associated periodic syndrome, adult-onset Still disease, Behcet's disease and gout. Again, it would reflect only a part of the RA pathogenesis. We have to be open to any suggestion.

References

1. Diamond B, Lipsky PE. Autoimmunity and autoimmune diseases. In: Harrison’s principles of internal medicine, 20th edition. 2, 2510-2514 (2018).
2. Shah A, St. Clair EW. Rheumatoid Arthritis. In: Harrison’s principles of internal medicine, 20th edition. 2, 2527-2540 (2018).
3. Rose NR, Bona C. Defining criteria for autoimmune diseases (Witebsky’s postulated revisited). Immunology. Today. 14(9), 426-430 (1993).
4. Putterman C, Naparstek Y. Koch’s postulates and autoimmunity: An opposing viewpoint. Immunol. Ther. Exp. 47, 25-32 (1999).
5. Pisetsky DS. Fulfilling Koch’s postulates of autoimmunity: Anti-NR2 Antibodies in mice and men. Arthritis. Rheum. 54(8), 2349-2352 (2006).
6. Cohen IR. Discrimination and dialogue in the immune system. Semin. Immunol. 12(3), 215-219 (2000).
7. Czömpöly T, Simon D, Czirják L et al. Anti-topoisomerase I autoantibodies in systemic sclerosis. Autoimmun. Rev. 8(8), 692-696 (2009).
8. Williams RC. Rheumatoid factors in subacute bacterial endocarditis and other infectious diseases. Scand. J. Rheumatol. 17(Suppl 75), 300-308 (1988).
9. van der Woude D, Rantapää-Dahlqvist S, Ioan-Facsinay A et al. Epitope spreading of the anti-citrullinated protein antibody response occurs before disease onset and is associated with the disease course of early arthritis. Ann. Rheum. Dis. 69(8), 1554-1561 (2010).
10. Kurowska W, Kuca-Warnawin EH, Radzikowska A et al. The role of Anti-Citrullinated Protein Antibodies (ACPA) in the pathogenesis of rheumatoid arthritis. Cent. Eur. J. Immunol. 42(4), 390-398 (2017).
11. Arend WP, Firestein GS. Pre-rheumatoid arthritis: predisposition and transition to clinical synovitis. Nat. Rev. Rheumatol. 8(10), 573-586 (2012).
12. Fierstein GS. The disease formerly known as rheumatoid arthritis. Arthr. Res. Ther. 16(3), 114 (2014).
13. Fournier C. Where do T cells stand in rheumatoid arthritis? Joint. Bone. Spine. 72(6), 527-532 (2005).
14. Garchon HH. Non-MHC-linked genes in autoimmune disease. Curr. Opin. Immunol. 4(6), 716-722 (1992).
15. Cantaert T, Brouard S, Thurlings RM et al. Alterations of the synovial T cell repertoire in anti-citrullinated protein antibody-positive rheumatoid arthritis. Arthritis. Rheum. 60(7), 1944-1956 (2009).
16. Hasler P. Biological therapies directed against cells in autoimmune disease. Springer. Semin. Immunopathol. 27(4), 443-456 (2006).
17. Reveille JD, Conant MA, Duvic M. Human immunodeficiency virus-associated psoriasis, psoriatic arthritis, and Reiter's syndrome: a disease continuum? Arthritis. Rheum. 33(10), 1574-1578 (1990).
18. Seok J, Warren HS, Cuenca AG et al. Inflammation and Host Response to Injury, Large Scale Collaborative Research Program. Genomic responses in mouse models poorly mimic human inflammatory diseases. Proc. Natl. Acad. Sci. USA. 110(9), 3507-3512 (2013).
19. Takao K, Miyakawa T. Genomic responses in mouse models greatly mimic human inflammatory diseases. Proc. Natl. Acad. Sci. USA. 112(4), 1167-1172 (2015).
20. Rashid T, Ebringer A, Wilson C. The link between Proteus mirabilis, environmental factors and autoantibodies in rheumatoid arthritis. Clin. Exp. Rheumatol. 35(5), 865-871 (2017).
21. Kotlarz A, Tukaj S, Krzewski K et al. Human Hsp40 proteins, DNAJA1 and DNAJA2, as potential targets of the immune response triggered by bacterial DnaJ in rheumatoid arthritis. Cell. Stress. Chaperones. 18(5), 653-659 (2013).
22. Zhu W, Li X, Fang S et al. Anti-citrullinated protein antibodies induce macrophage subset disequilibrium in RA patients. Inflammation. 38(6), 2067-2075, 2015.
23. Lai NS, Yu HC, Yu CL et al. Anti-citrullinated protein antibodies suppress let-7a expression in monocytes from patients with rheumatoid arthritis and facilitate the inflammatory responses in rheumatoid arthritis. Immunobiology. 220(12), 1351-1358 (2015).
24. James EA, Rieck M, Pieper J et al. Citrulline-specific Th1 cells are increased in rheumatoid arthritis and their frequency is influenced by disease duration and therapy. Arthritis. Rheumatol. 66(7), 1712-1722 (2014).
25. Kurowska-Stolarska M, Alivernini S. Synovial tissue macrophages: friend or foe? RMD. Open. 3(2), e000527 (2017).
26. Elmesmari A, Fraser AR, Wood C et al. MicroRNA-155 regulates monocyte chemokine and chemokine receptor expression in Rheumatoid Arthritis. Rheumatology (Oxford). 55(11), 2056-2065 (2016).
27. Neidhart M, Pajak A, Riksen NP et al. Oligomeric S100A4 is associated with monocyte innate immune memory and bypass of tolerance to subsequent stimulation with lipopolysaccharides. Front. Immunol. Inflammation. (2019).
28. Kristensen LE, Jakobsen AK, Bartels EM et al. The number needed to treat for second-generation biologics when treating established rheumatoid arthritis: a systematic quantitative review of randomized controlled trials. Scand. J. Rheumatol. 40(1), 1-7 (2011).
29. Tarp S, Furst DE, Dossing A et al. Defining the optimal biological monotherapy in rheumatoid arthritis: A systematic review and meta-analysis of randomised trials. Semin. Arthritis. Rheum. 46(6), 699-708 (2017).
30. Karouzakis E, Neidhart M, Gay RE et al. Molecular and cellular basis of rheumatoid joint destruction. Immunol. Lett. 106(1), 8-13 (2006).
31. Tykocinski LO, Lauffer AM, Bohnen A et al. Synovial Fibroblasts Selectively Suppress Th1 Cell Responses through IDO1-Mediated Tryptophan Catabolism. J. Immunol. 198(8), 3109-3117 (2017).
32. Kontoyiannis D, Pasparakis M, Pizarro T et al. Impaired on/off regulation of TNF biosynthesis in mice lacking TNF AU‐rich elements: implications for joint and gut-associated immunopathologies. Immunity. 10(3), 387–398 (1999).
33. Goh FG, Midwood KS. Intrinsic danger: Activation of Toll-like receptors in rheumatoid arthritis. Rheumatology (Oxford). 51(1), 7-23 (2012).