Revisiting B cell tolerance and autoantibodies in seropositive and seronegative autoimmune rheumatic disease (AIRD)

Autoimmune rheumatic diseases (AIRD) are categorized as seropositive or seronegative, based on the presence or absence of a defined set of autoantibodies. Autoantibodies in AIRD are produced after loss of B cell tolerance, as a consequence of defects incurred during early B cell development, genetic variants in products regulating B and T cell peripheral tolerance, environmental factors, or immunologic triggers. Emerging evidence now suggests that AIRD that were previously considered seronegative do have autoreactive B cells and autoantibodies that contribute to disease.

Summary

Autoimmune rheumatic diseases (AIRD) are categorized seropositive or seronegative, dependent upon the presence or absence of specific autoreactive antibodies, including rheumatoid factor and anti‐citrullinated protein antibodies. Autoantibody‐based diagnostics have proved helpful in patient care, not only for diagnosis but also for monitoring of disease activity and prediction of therapy responsiveness. Recent work demonstrates that AIRD patients develop autoantibodies beyond those contained in the original categorization. In this study we discuss key mechanisms that underlie autoantibody development in AIRD: defects in early B cell development, genetic variants involved in regulating B cell and T cell tolerance, environmental triggers and antigen modification. We describe how autoantibodies can directly contribute to AIRD pathogenesis through innate and adaptive immune mechanisms, eventually culminating in systemic inflammation and localized tissue damage. We conclude by discussing recent insights that suggest distinct AIRD have incorrectly been denominated seronegative.

Introduction

Autoimmune rheumatic diseases (AIRD) are heterogeneous musculoskeletal disorders accompanied by substantial morbidity and mortality. AIRD mainly, although not exclusively, affect joints and muscles and are characterized by the presence of specific autoantibodies [1]. Traditionally, AIRD are classified as ‘seropositive’ or ‘seronegative’, according to whether or not autoantibodies are a known important feature [2]. Examples of seropositive AIRD include rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), systemic sclerosis (SSc), primary Sjögren’s syndrome (pSS) and idiopathic inflammatory myopathies (IIM) [3]. Importantly, although RA is classified within the seropositive category, approximately 30% of patients lack the presence of the classic autoantibodies rheumatoid factor (RF) and anti‐citrullinated protein antibodies (ACPAs) and have ‘seronegative RA’ [4].

Autoantibodies are secreted immunoglobulins (Igs) of isotype IgM, IgG, IgA or IgE [5]. A loss of immune tolerance to self‐antigens, including nucleic acids, lipids, proteins and tissue‐specific antigens, can elicit autoantibody production [6]. Autoantibodies can be directed against self‐antigens inside the nucleus, in the cytoplasm, at the cell surface and extracellular products [6, 7]. It has become increasingly evident that autoantibodies contribute to AIRD pathogenesis by promoting systemic inflammation as well as local tissue damage, involving both innate and adaptive immune mechanisms [6, 8, 9]. Although AIRD share pathophysiological mechanisms through which self‐reactive antibodies cause damage, the pathogenicity of autoantibodies varies per antibody isotype, target antigen and clinical disease phase [6, 9].

In clinical practice, the monitoring of autoantibodies has proved helpful in diagnostic processes, for classification purposes and as biomarkers for disease activity, prognosis and, in the light of personalized medicine, therapeutic responses of AIRD [1, 4, 10]. From a pathogenesis perspective, it is relevant that autoantibodies can be detected years prior to diagnosis [10]. A selection of studies that reported presence of autoantibodies years before clinical AIRD diagnosis is highlighted in Table 1. The antigen‐specific autoantibodies may show isotype‐switch and a progressive accumulation before disease onset [4, 5, 11]. These findings raise an important question: how can these self‐reactive molecules persist for years, without causing clinical signs or symptoms of autoimmunity?

Table 1.

Autoantibodies can be measured years before AIRD diagnosis (selected studies)

AIRD	Prevalence a	References	n	Autoantibody	Positive before diagnosis (% patients) b	Time from detection to diagnosis (years)
						Mean/median c	Upper limit
SLE	24/100.000	Arbuckle 2003 [114]	130	ANA	78	3·0	9·2
				Anti‐dsDNA	55	2·2	9·3
				Anti‐RNP	26	0·9	7·2
RA	860/100.000	Majka 2008 [115]	83	ACPA	61	5·4	13
				RF	57	6·0	14
SSc	4/100.000	Burbelo 2019 [116]	46	≥ 1 including anti‐Topo1, ‐RNAP III, centromere proteins, ‐Scl‐75, ‐Scl‐100	52	7·4	27·1
pSS	14/100.000	Theander 2015 [117]	117	ANA	n.r.	4·6	18·8
				SS‐A	n.r.	4·5‐5·1	18·8
				SS‐B	n.r.	3·5	16·1
				≥ 1	75	n.r.	19·5
IMM	5.1/100.000	Miller 1990 [118]	1	Anti‐Jo	100	–	0·4
		Abe 2017 [119]	105	Anti‐MDA5	2	n.r.	2
		Targoff 1992 [120]	5	Anti‐EJ	20	–	0·33
		Vulsteke 2020 [121]	1	Anti‐Mi‐2	100	–	0·25

ACPA = anti‐citrullinated peptide antibodies; AIRD: autoimmune rheumatic disease; anti‐EJ: anti‐gylcyl antibody; anti‐Jo‐1 = anti‐histidyl‐tRNA synthetase antibody; anti‐MDA5 = anti‐melanoma differentiation‐associated gene 5 antibody; anti‐RNAP = RNA polymerase; anti‐Scl = anti‐scleroderma antibody; anti‐Sm = Smith antibody; anti‐topoI = topoisomerase I antibody = anti‐Scl‐70 antibody; dsDNA: double‐stranded deoxyribonuclease; Ig = immunoglobulin; IIM: idiopathic inflammatory myopathy; Mi‐2 = Mi‐2 nuclear antigen antibody; n = number of patients that participated in the study; n.r. = not reported in published data; pSS = primary Sjögren’s syndrome; RA = rheumatoid arthritis; RF = rheumatoid factor; SLE = systemic lupus erythematosus; SS‐A = Sjögren’s syndrome antigen A = anti‐Ro antibody; SS‐B = Sjögren’s syndrome antigen B = anti‐La antibody; SSc = systemic sclerosis.

^aCooper G. S., Stroehla B. C. [3]. For IMM pooled prevalence shown of polymyositis and dermatomyositis.

^bInterpret results with care, as most reported studies are case–controls performed in either confirmed patients or patients with risk factors for AIRD.

^cMean time reported in studies by Arbuckle and Burbelo. Median time reported in studies by Majka and Theander. No mean or median reported because autoantibodies were detected only in one patient before diagnosis in studies by Miller, Targoff and Vulsteke.

Extensive investigation is ongoing to improve our understanding of how autoantibodies are generated and through which mechanisms autoantibodies initiate and perpetuate disease. This holds true even for traditionally classified seronegative AIRD, including psoriatic arthritis (PsA), ankylosing spondylitis (AS), seronegative RA and reactive arthritis [10, 12]. In this study we provide a literature overview of how loss of B cell tolerance results in autoantibody production in AIRD, and discuss the contribution of genetic predisposition, immune mechanisms and environmental factors. Next, we describe overarching innate and adaptive immune mechanisms that induce systemic inflammation and localized tissue damage. Finally, we revisit the diseases that were previously classified as seronegative AIRD in light of recently discovered autoantibody specificities that allow further differentiation of AIRD groups.

History

The discovery of antibodies dates back to the 19th century when Emil von Behring and Shibasabura Kitasato, in 1890, used serum from immunized animals to cure animals suffering from diphtheria [13]. Another key insight came from Julius Donath and Karl Landsteiner, who found that under cold circumstances blood components in serum from paroxysmal cold hemoglobinuria patients could break down their own (or other human) erythrocytes [14]. The concept of autoantibodies was born. The first report on self‐reactive antibodies in AIRD was published in 1940, when Erik Waaler discovered the presence of RF in a patient with RA [15]. RF, an immunoglobulin directed towards the Fc part of IgG, would later instigate a major leap in this field of research as a clinically relevant diagnostic and prognostic biomarker. Since then, technical advances have improved the detection rate of disease‐specific autoantibodies, resulting in the wide array of AIRD‐associated autoantibodies with various clinical implications known today [8].

Antibodies and B cell tolerance

Effective immune protection requires a broad, diverse and specific antibody repertoire [16, 17]. Initially, all naive B cells express an unique IgM‐type antigen receptor at the cell surface. After productive encounter of antigen, the naive B cell repertoire is refined by somatic hypermutation of the variable regions of heavy and light chain gene loci of the B cell antigen receptor and by class‐switch recombination to produce IgA, IgG and IgE isotypes [16, 17]. The resultant refined antibody repertoire enables recognition of a wide range of epitopes with high affinity, but has the inherent risk of recognizing innocuous self‐antigens [16, 17, 18]. Notably, a substantial part of the antibody repertoire of healthy individuals shows some level of self‐reactivity [19, 20]. Through incompletely understood mechanisms, the retention of a confined number of proto‐autoreactive naive B cells can improve the protective antibody response against foreign antigens. One mechanism was suggested from mouse‐based studies [21]. Here, transgenic B cells producing antibodies cross‐reactive to a foreign and self‐antigen underwent anergy upon encounter of self‐antigen, but upon encounter of high‐density foreign antigen increased its foreign‐specific affinity by directed hypermutation and selection [21]. However, to prevent excessive amounts of autoreactive B cells that may contribute to autoimmune disease, the immune system has incorporated checkpoints during central and peripheral B cell development to ensure that the number of proto‐autoreactive B cells gradually decreases during maturation [6, 16]. Importantly, this negative selection cannot be too stringent, because this would result in a limited antibody diversity unable to recognize all potential noxious antigens.

Despite the mechanisms of the immune system to ensure B cell tolerance, most AIRD are characterized by high titers of serum autoantibodies [18]. Autoantibody production is explained by a multi‐factorial process that involves the failure of the immune system to both eliminate and control autoreactive B cells. Mechanisms that contribute to this loss of tolerance include the persistence of autoreactive B cells through defective central and peripheral B cell development checkpoints (see section below: B cell development: transition through checkpoints), genetic predisposition (see section below: Genetic predisposition), environmental factors (see section below: Environmental factors) and immunological mechanisms, including an important role for cognate T cells (see section below: Immunological triggers) (Fig. 1) [6, 22, 23].

Fig. 1.

Breaches in B cell tolerance that contribute to autoantibody production in autoantibodies in autoimmune rheumatic diseases (AIRD) are generated if the immune system fails to eliminate and control (proto)autoreactive B cells. Involved in this multi‐factorial process are deficient B cell development checkpoints (a,b) and additional mechanisms that breach B cell tolerance (1c). (a) Central checkpoints of B cell development in bone marrow include positive selection of cells with a functional pre‐B cell receptor (BCR) (checkpoint 1), negative selection of immature B cells with an autoreactive pre‐BCR (checkpoint 2) and immunoglobulin (Ig) receptor ligand‐mediated apoptosis of immature B cells with an autoreactive BCR (checkpoint 3). (b) Peripheral checkpoints include apoptosis of immature proto‐autoreactive B cells in the spleen (checkpoint 4), anergy and follicular exclusion of mature proto‐autoreactive B cells upon autoantigen encounter (checkpoint 5) and prevention of recirculation of autoreactive B cells that emerged after somatic hypermutation in secondary lymphoid organs (checkpoint 6). (c) Additional mechanisms that contribute to a breach in B cell tolerance are the genetic predisposition, environmental factors and immunological triggers. Figure created using images fromhttp://smart.servier.com.HLA = human leukocyte antigen.

B cell development: transition through checkpoints

B cells develop from common lymphoid progenitors in a stepwise fashion to establish a repertoire of cells that is mainly non‐self‐reactive. For a detailed review of this subject we refer to published work [16]. Defects in the process of B cell development and maturation can contribute to the pathogenesis of autoantibody production in AIRD [24, 25]. In AIRD, increased numbers of proto‐autoreactive mature naive B cells, that have the potential to produce self‐reactive antibodies, persist after key developmental stages (indicated as checkpoints 2, 3 and 4 in Figure 1a,b) [18, 20, 26]. A study conducted in SLE patients describes that 20–50% of mature naive B cells produce self‐reactive antibodies, compared to 5–20% in healthy controls [26]. In an RA study these percentages were 35–52% in RA versus 20% in healthy controls [24]. These studies, although modest in size, underscore a contribution to autoantibody production of defective central B cell receptor (BCR) signaling in bone marrow and impaired receptor editing [18, 20, 26, 27, 28].

After checkpoint 4, naive B cell antigen exposure results in activation which requires help from cognate CD4⁺ T follicular helper (Tfh) cells to mount T cell‐dependent high‐affinity antibody responses and memory [29, 30]. At this developmental stage, proto‐autoreactive naive B cells become anergic upon autoantigen encounter and are excluded from migration into lymphoid follicles, which results in rapid cell death [31, 32]. These processes comprise checkpoint 5, and depend upon both continuous exposure to self‐antigen and competition for the follicle between autoreactive and ‘normal’ competitor mature B cells, which have different specificities [32, 33]. Next, the activated B cell undergoes class‐switch recombination via which the constant region of the antibody is substituted by the class‐switched Ig isotype [34]. Subsequently, the activated B cell migrates away from the T–B cell border, becomes a B cell blast and gives rise to a germinal center (GC), supporting clonal proliferation, somatic hypermutation and selection of higher‐affinity B cell clones [29].

An important feature of autoantibody production in AIRD is that it is not strictly limited to GCs in follicular regions of secondary lymphoid tissues. Extrafollicular antibody‐forming cells and ectopic lymphoid structures (ELS) were discovered in the synovium in RA, in salivary and lacrimal glands in pSS, in kidneys in SLE and in a small minority of muscles of dermato‐ and polymyositis patients [35, 36, 37, 38, 39]. These ELS are characterized by GC‐resembling organized lymphoid aggregates that contain autoantibody‐producing plasmablasts and even long‐lived plasma cells, normally only present in the bone marrow [40]. Several studies have reported associations of ELS with autoantibody titer, antibody status, circulating inflammatory cytokines and disease severity, suggesting their implication in the perpetuation of disease within target organs [38].

The sixth and last‐known checkpoint ensures that no autoreactive cells emerge after somatic hypermutation. These cells are prevented from recirculation in the long‐lived repertoire by undergoing apoptosis [6, 41]. Data support that autoreactive B cells in AIRD arise despite the multiple checkpoints discussed [6, 42]. In SLE, for example, immunoglobulins produced by memory B cells are highly reactive compared to germline‐encoded antibodies. These highly reactive antibodies can only have resulted from affinity maturation after antigen encounter [43]. Furthermore, high‐specificity IgG anti‐phospholipid antibodies show accumulation of mutations, suggesting affinity maturation [44]. Also, the high‐affinity binding of autoantibodies to nucleosomes and anti‐dsDNA is acquired by somatic hypermutation [42]. In RA, evidence for the persistence of autoreactive B cells after the last checkpoint comes from highly somatically mutated ACPA‐producing IgG secreting cells from synovial fluid, indicative of past encounter with autoantigens [45]. Altogether, failure to eliminate (proto)autoreactive B cells at these B cell developmental checkpoints is suggested to contribute to autoantibody production in AIRD [16].

Beyond checkpoints

Genetic predisposition

Genetic variants in products regulating B and T cell peripheral immune tolerance can contribute to a breach in B cell tolerance that contributes to generation of autoantibodies (Fig. 1c). These variants include molecules of the antigen presentation machinery [4, 6]. Autoantibody status and titers in AIRD are associated with specific human leukocyte antigen (HLA) class II haplotypes and several non‐HLA genes. The association with HLA class II supports an important role of a T cell‐dependent antigen‐driven response in autoantibody production. This is because B cells, after internalization and processing of BCR‐bound antigen, act as professional antigen‐presenting cells for CD4⁺ T cells. B cells present antigens to T cells as peptide–HLA complexes (class II HLA‐DR, ‐DP, ‐DQ) on the B cell surface that interact with the α/β T cell receptor on the T cell surface [46, 47]. This interaction activates cognate CD4⁺ T cells that, in turn, allow the furthering of an antigen‐specific B cell response to linked epitopes [48]. In RA, HLA‐DRB1 alleles that code a ‘shared epitope’ – an amino acid sequence QKRAA, QRRAA or RRRAA in residues 70–74 of HLA‐DRβ chain – are strongly associated with ACPA production [49, 50]. Moreover, in SLE and pSS, multiple autoantibodies are strongly correlated with specific DR and DQ haplotypes [51, 52]. In myositis, one of the well‐known associations is that of the DRB1*03 haplotype with anti‐Jo1 production [53]. Lastly, in SSc, DPB1*13:01 and DRB1*07:01 alleles are strongly associated with anti‐topoisomerase and ‐centromere status [54].

In addition to HLA molecules, additional known gene variants contribute to changes in B cell tolerance. These include molecules involved in BCR downstream signaling, antigen processing, lymphocyte proliferation and differentiation and clearance pathways of apoptotic material [4, 6, 37]. In SLE, SSc and RA, variants in gene products involved in BCR signaling pathways were shown to associate with autoreactive B cell development [18, 55]. Furthermore, abnormalities in genes encoding proteins involved in removal of self‐antigens from the extracellular milieu, or sensing the presence of RNA and DNA in endosomes, have been implicated in AIRD development [56, 57]. Single nucleotide polymorphisms in genes coding for Toll‐like receptors (TLR) – a family of pattern recognition receptors that recognize a wide range of pathogen‐associated molecular patterns (PAMPs) and are expressed by stromal cells, B cells, dendritic cells and macrophages – even correlate with the pathogenesis of AIRD, including SLE, RA and SSc (further discussed below, in the Immunological triggers section) [58].

Environmental factors

The association of specific environmental factors with the autoantibody response in AIRD suggests their contribution to loss of B cell tolerance (Fig. 1c). One example is cigarette smoking, which associates with ACPA positivity in RA, with autoantibody development in myositis and with anti‐topoisomerase I positivity in SSc [53, 59, 60]. Moreover, exposure to different toxic substances has been associated with autoantibodies, such as the association of silica with significantly higher ANA levels in murine SLE models [61]. Infection with pathogens as bacteria or viruses should also be included as an environmental factor associated with autoantibody response. Infections can induce a breach in B cell tolerance in at least three ways: first, via direct actions of the invading pathogen. In RA, for example, studies showed that oral P. gingivalis infection induces the citrullination of proteins, thereby generating neoepitopes on self‐antigens that trigger autoantibody production [62]. A second mechanism is molecular mimicry: an immune response initially directed towards a pathogen is perpetuated because of cross‐reactivity with foreign and self‐antigens [63, 64]. Molecular mimicry was reported, for example, in SLE for dsDNA with a dominant pneumococcal cell wall hapten, and also for SS‐A (anti‐Ro) antibodies with a latent viral protein Epstein–Barr virus (EBV) nuclear antigen‐1 [63, 64]. In RA, EBV and human endogenous retrovirus K have been suggested to share multiple epitopes with self‐antigens to induce cross‐reactivity, resulting in autoantibody production against interleukin (IL)‐2 and fibrin [65, 66]. Thirdly, autoreactive B cells can be directly prompted to proliferate and produce autoantibodies through innate stimulation, after encounter of PAMPs expressed by pathogens [67]. Ligation of innate receptors by PAMPs is also suggested to stimulate autoantibody production in RA indirectly through synovial fibroblasts [68]. When stimulated with TLR‐3 ligand poly(I:C), for example, RA fibroblast‐like synoviocytes induced class‐switch recombination of RA patient B cells but not B cells from healthy individuals [68]. This supports the role of TLR‐stimulated synoviocytes in promoting immunoglobulin class‐switch in RA synovium.

Immunological triggers

In addition to genetic variants and environmental factors, immunological triggers may contribute to a breach in B cell tolerance, resulting in autoantibody production in AIRD (Fig. 1c). The most important triggers are discussed below.

A carefully orchestrated collaboration between T and B cells is essential for an effective and rapid affinity‐matured protective antibody response. However, T and B cell interaction can also cause harm through triggering autoimmune responses [69]. The important role for cognate CD4⁺ T cells in AIRD is underlined by the strong association of class II HLA alleles with affinity‐matured autoantibodies, supporting a role for antigen presentation via specific HLA alleles by dendritic cells (DCs) and B cells to cognate CD4⁺ T helper cells in the autoantibody response. In pSS, SSc, SLE and RA increased Tfh frequencies are reported, which are specialized CD4 T cells that control B cell proliferation, isotype‐switch and somatic hypermutation [70, 71]. A second follicular CD4 T cell subset, regulatory (T_freg) cells, exerts further control on (auto)antibody responses and Tfh [30]. In SLE, the ratio of circulating Tfh to T_freg cells correlates with disease activity and anti‐dsDNA antibody level, suggesting their importance in autoantibody production [72]. In pSS, this ratio also strongly correlates with autoantibody production and T cell infiltration in salivary glands [73, 74].

As well as cellular determinants, molecular determinants on cells regulate autoantibody responses in SLE, RA and pSS, such as TLRs. Especially relevant are the endosome‐localized TLR‐7 and TLR‐9 that recognize RNA and DNA [6, 75]. Murine lupus models demonstrated that TLR‐7 and TLR‐9 are required for the generation of RNA and dsDNA‐specific autoantibodies, respectively [76]. However, a more complex role for TLR‐9 is suggested by autoimmune‐prone mouse models [77]. While B cell knock‐out experiments support that TLR‐7 drives autoantibody production, TLR‐9 instead appears protective against systemic autoimmunity through not completely understood mechanisms [76, 78]. The encounter of TLR ligands can trigger autoreactive B cell responses in at least two ways. First, co‐engagement of BCR and TLR can induce autoantibody production in proto‐autoreactive B cells [6, 79, 80]. Co‐engagement is induced by immune complexes that contain nuclear antigens, which elicit synergistic responses that recruit TLRs to internalized BCRs in auto‐phagosomes [67, 81, 82]. The presence of these TLRs and their nuclear ligands in auto‐phagosomes might explain the preponderance of reactivity of many autoantibodies with nuclear antigens in autoimmune diseases [67]. Secondly, TLR ligation can result in T cell‐independent B cell autoreactivity in the presence of B cell‐activating factor of the tumor necrosis factor family (BAFF) [83]. BAFF is a cytokine essential for maturation, proliferation and survival of peripheral B cells, and increased circulating BAFF is associated with pSS, SLE and RA [83, 84]. In the presence of BAFF, TLR ligation promotes B cell activation, class‐switch, somatic hypermutation and plasma cell differentiation that can all promote harmful autoantibody generation [85, 86, 87].

As well as T cells and TLRs, deficiency in the complement system is important to include as a factor contributing to the existence of autoantibodies in AIRD. Complement is an essential component of the humoral immune system that plays a role in both innate and adaptive responses. Up‐regulation, down‐regulation and dysregulation of complement can all contribute to autoimmune disease [88]. Well‐recognized anomalies associated with autoantibody production in AIRD are primary deficiencies in complement system pathways and regulators, including C1q, C2, C4, mannose‐binding lectin and C1‐inhibitor [88]. Furthermore, secondary deficiencies of complement by autoantibodies have been described. In SLE, 30–60% of the patients have anti‐C1q autoantibodies which are strongly associated with severe hypocomplementemia and lupus nephritis [7, 89]. In patients with SSc (26%), pSS (14%) and extra‐articular RA (> 30%), circulating C1q antibodies are also detected [89, 90]. Complement deficiency can reduce B cell tolerance as follows [88, 91]: first, C1q deficiency specifically intervenes with effective negative selection of autoreactive B cells in bone marrow [89]; and secondly, via insufficient elimination of immune complexes, apoptotic and necrotic cell material [92]. Under healthy conditions debris is opsonized by immunoglobulins and complement factors, and then rapidly cleared from the circulation via binding to CR1 on erythrocytes and through engulfment by phagocytes [7, 93]. When complement fails to eliminate debris, including various self‐antigens, exposure of the immune system to these antigens can increase the propensity that autoreactive B cells are triggered to be activated and to produce autoantibodies [6, 7, 92]. In addition, complement serum protein deficiencies, dysfunction of linked and complementary pathways involved in the clearance of apoptotic cells can contribute to AIRD [94, 95]. Here we provide one example, which is the deficiency of scavenger receptor type F family member 1, involved in the recognition and engulfment of apoptotic cells via complement component C1q, which was shown to induce lupus‐like disease and autoantibody production in mice [96].

The last important trigger that contributes to autoantibody production in AIRD is antigen modification [97]. Both apoptosis and inflammatory responses can initiate proteolysis of self‐proteins, and can cause post‐translational protein modifications such as phosphorylation, citrullination, carbamylation and deamination. These changes in protein appearance are relevant to AIRD: antigen modification can alter the immunogenicity of these molecules and thereby result in recognition by autoreactive B cells [97, 98, 99]. ACPAs, for example, can recognize various citrullinated proteins including α‐enolase, vimentin, fibrinogen and myelin‐binding protein [98].

Pathogenicity

The pathogenicity of autoantibodies varies per antibody type, self‐antigen specificity and clinical phase of disease. However, the pathogenic mechanisms through which autoantibodies contribute to localized tissue damage and systemic inflammation overlap and share involvement of both the innate and adaptive immune system (Fig. 2) [6, 9]. Collectively, the indicated mechanisms – immune complex deposition, antibody‐dependent cellular cytotoxicity, FcyR‐mediated cell activation and complement activation – create a proinflammatory environment. In the inflamed tissues, both immune and parenchymal cells are damaged through the release of reactive oxygen species, matrix‐degrading enzymes and proteolytic enzymes [9]. Moreover, systemic inflammation further impairs tolerance mechanisms of both B and T cells, which leads to more antibody production and a downward spiral of ongoing autoimmunity [6].

Fig. 2.

Key pathogenic effects of autoantibodies in autoantibody production in autoantibodies in autoimmune rheumatic diseases (AIRD). Autoantibodies can induce tissue damage and create a proinflammatory microenvironment through multiple components of the innate and adaptive immune system. Localized tissue damage mediated by autoantibodies involves three mechanisms. First, antibody‐dependent cellular cytotoxicity (top right): killing of antibody coated target cells by binding of the Fc domain of IgG autoantibody by Fcγ receptor (FcγR)‐expressing effector cells, most notably natural killer (NK) cells, granulocytes and macrophages. Secondly, through antibody‐induced activation of the complement pathway (lower right). Complement activation can cause cell lysis through assembly of the membrane‐attack complex, can induce phagocytosis of complement C3 proteolytic fragment‐coated (opsonized) damaged cells, and can recruit innate inflammatory cells through release of small complement fragments C3a and C5a (anaphylatoxins). Thirdly, immunpoglobulin (Ig)G autoantibodies can activate various FcγR‐expressing innate immune cells (lower left). For example, plasmacytoid dendritic cells (DC) that produce type I interferon (IFN), macrophages that produce tumor necrosis factor (TNF)‐α and mast cells that release granules with degrading enzymes and produce proinflammatory cytokines. Systemic inflammation is primarily mediated through deposition of circulating immune complexes (IC) (top left). These ICs contain autoantibodies bound to self‐antigens (such as DNA), Toll‐like receptor (TLR) ligands and post‐translationally modified proteins. IC deposition induces the systemic and synergistic activation of cells of the innate immune system via FcγR and TLR ligation. As a result, T helper cell responses are amplified and trigger the release of degrading enzymes and production of proinflammatory cytokines such as TNF‐α, interleukin (IL)‐1β and IL‐6. Together, these mechanisms contribute to a proinflammatory microenvironment with cytokines that further enhance inflammation and damage, through activation of parenchymal and immune cells, production of matrix‐degrading and proteolytic enzymes and release of reactive oxygen species. Figure created using images fromhttp://smart.servier.com. References listed in Supporting information, Table C.

In SLE the harmful effects of autoantibodies are extensively proven. Numerous data support the potency of ICs containing ANAs to initiate lupus nephritis [63, 100]. However, in other AIRD the pathogenicity of autoantibodies is less well understood [9, 37]. In RA, a debate is ongoing regarding whether ACPAs, ACPA‐producing B cells or T helper cells are especially responsible for transition from preclinical phases to arthritis [9, 23]. A main argument against a major role for autoantibodies in disease onset is the fact that they are present long before disease onset. However, prominent increase of autoantibody titers before onset of symptoms is supportive of such a driver role [11]. Moreover, seropositive RA patients have more severe disease and radiographic damage. Also, ACPAs were shown to promote arthritis in murine models, activate complement, induce cytokine production and activate Fcγ‐receptor‐expressing immune cells [9, 12, 98]. With an increasing number of studies reporting a role for ACPAs in RA pathogenesis, we anticipate that the pathogenic role of existing and new autoantibodies will be demonstrated in other AIRD.

Revisiting seronegative AIRD

Some AIRD are traditionally classified as seronegative, because of the low number of patients positive for ACPA, RF, SS‐A, SS‐B, anti‐dsDNA and other prototypical autoantibodies [2]. Common seronegative AIRD include reactive arthritis, undifferentiated spondyloarthritis, PsA, AS and seronegative RA [4, 12, 101]. We argue that the classification of seronegative AIRD needs revisiting, because increased frequencies of IgM‐, IgG‐ and IgA‐producing plasma cells and plasmablasts are detected in the circulation and joints of patients [102, 103]. Moreover, increasing numbers of autoantibodies are detected in these ‘seronegative’ diseases, and emerging evidence suggests that plasma cells and autoantibodies are involved in their disease course [4, 12]. Most data supporting the seropositive nature of these diseases are available for AS and PsA, of which we will now discuss recently identified autoantibodies, pathogenicity and breaches in B cell tolerance.

In PsA, several autoantibodies have been identified in plasma, serum and synovial fluid (SF) (Table 2). Indicative of the role of autoantibodies in disease pathogenesis is that titers and seropositive status of certain ‘new’ autoantibodies associate with disease activity [99]. Moreover, their involvement in PsA pathogenesis is suggested by the fact that some autoantibodies were significantly higher in PsA patients compared to patients with psoriasis alone [104]. Multiple autoantibodies are also reported in AS, such as anti‐CD74 (CLIP), anti‐oxidized collagen type II and antibodies against various extracellular matrix proteins. These are discussed in depth elsewhere [12, 105, 106]. These autoantibodies are suggested to be produced in tertiary organized lymphoid structures – an important source of autoantibodies in seropositive AIRD – as ectopic lymphoid structures were identified in both AS and PsA synovium [12, 107].

Table 2.

Autoantibodies detected in psoriatic arthritis (selected studies)

Antibody	Antigen a	n	Present (% patients)	Clinical association	References
Anti‐20s proteasome	20s proteasome	36	28%	No (tested for disease duration, nail involvement, dactylitis, ANA/ RF/ACPA status, articular phenotype)	Colmegna 2008 [110]
Anti‐MCV	MCV	46	24%	Association with presence of tender knee joints and nail psoriasis	Dalmady 2013 [123]
Anti‐PsA peptide b	TNRRGRGSPGAL	100	85%	n.r.	Dolcino 2014 [113]
Anti‐CarP	Carbamylated proteins	30	53%	Positive correlation with age, disease duration, ESR and PGA. Negative correlation with functional status.	Chimenti 2015 [99]
Anti‐α6‐integrin	α6‐integrin	46	28%	No (tested for early onset PsA)	Gal 2017 [124
Anti‐LL37	Cathelicidin LL37 Native / Carbamylated / Citrullinated	PL: n.r. / 32 / 29 SF: 19 / 17 / 21	PL: 0% / 52% / 32% SF: 37% / 47% / 57%	PL anti‐carbamylated: positive correlation with DAS44. SF anti‐native: positive correlation with CRP, ESR, swollen joints and DAS44.	Frasca 2018 [108]
Anti‐ADAMTSL5	ADAMTSL5	22	n.r. c	Positive correlation with skin disease activity	Yuan 2019 [125]

ACPA = anti‐citrullinated peptide antibody; ADAMTSL5 = a disintegrin‐like and metalloprotease domain with thrombospondin type 1 motifs like 5; ANA = anti‐nuclear antibody; CarP = carbamylated protein; CRP = C‐reactive protein; DAS44 = disease activity score 44; ESR = erythrocyte sedimentation rate; MCV = mutated citrullinated vimentin; n = number of patients with PsA diagnosis that participated in the study; n.r. = not reported in published data; PGA = patient global assessment of disease activity; PL = plasma; PsA = psoriatic arthritis; RF = rheumatoid factor; SF = synovial fluid.

^aDetected in serum, unless otherwise specified.

^b‘Anti‐PsA peptide’ (TNRRGRGSPGAL) antibodies recognize epitopes of self‐antigens in skin and joints.

^cNo cut‐off value for positivity reported, only graphical results of enzyme‐linked immunosorbent assay (ELISA) [immunoglobulin (Ig)G levels, µg/ml] for PsO n = 32, PsA n = 22 and autoantigen array [IgG levels (mean fluorescence intensity) for healthy controls (HC) n = 20, systemic lupus erythematosus (SLE) n = 7, PsO n = 73].

Similar to seropositive AIRD, several breaks of immune toleranceconsidered necessary to result in generation of autoantibodies in ‘seronegative’ disease, and similar mechanisms are identified. For example, post‐translational antigen modification is necessary for the generation of antibodies against citrullinated and carbamylated cathelicidin LL37, citrullinated vimentin, oxidized collagen type II and other carbamylated proteins [12, 105, 108, 109]. Also, anti‐protease antibodies might alter the cleavage pattern of the proteasome, which potentially results in the generation of immunogenic self‐antigens [110]. Next to antigen modification, there is evidence for a role of molecular mimicry leading to autoantibody development [12, 110]. Furthermore, neutrophils are important players in breaching tolerance. First, neutrophil extracellular trap (NET)‐derived complexes with self‐antigens may contribute to autoantibody production by interaction with self‐reactive B lymphocytes [12, 110, 111]. Secondly, neutrophil activation by granulocyte–macrophage colony‐stimulating factor (GM‐CSF) and complement fragments such as C5a is suggested to initiate autoimmunity through neutrophil degranulation. Degranulation can result in the release and post‐translational modification of autoantigens such as LL37, an anti‐microbial peptide with immune‐modulating properties [112]. This hypothesis is supported by a correlation of GM‐CSF and complement factor levels in serum and SF with autoantibody reactivity [104, 108]. Notably, not all mechanisms that are described in seropositive AIRD have been identified thus far in seronegative disease, which might be explained by either the absence of these particular mechanisms or by the infancy of this field of research.

The contribution of these newly identified autoantibodies in seronegative AIRD is currently under extensive investigation, and recent reports suggest significant pathogenicity. For example, ‘TNRRGRGSPGAL’ peptide antibodies, present in 85% of PsA patients, cross‐react with epitopes expressed in both skin and enthuses [113]. Moreover, these autoantibodies bind TLR‐2, which has an important role in activation of the innate immune system [4, 113]. Another example concerns the modification of carbamylation, which is suggested to trigger oxidative stress and to contribute to systemic inflammation [99]. It is hypothesized that AS autoantibodies directly damage bony structures by inducing osteoclastogenesis [12]. Neutrophils may play a role in autoantibody‐mediated tissue damage in tertiary lymphoid tissues in PsA and AS synovium. Deposited IgG ICs in these ELS co‐localize with infiltrating activated neutrophils that mediate inflammatory synovial damage [12] and, as known from seropositive AIRD, the presence of ELS is implicated in the perpetuation of disease [38].

Overall, these emerging insights support the notion that thus far we may have oversimplified distinct AIRD as being seronegative, by considering only a limited set of autoantibodies. The newly identified autoantibodies and their pathogenic effects support the concept of these disorders as falling within the spectrum of what was previously termed ‘seropositive’ autoimmune diseases, which opens up new avenues for investigating disease pathogenesis, identification of disease biomarkers or even new therapeutic targets.

Concluding remarks

Autoantibodies in AIRD can develop through sequential antigen‐driven events that ultimately cause a loss of B cell tolerance, which include defects in B cell development, genetic variants and specific immunological triggers. In this study we have summarized a current view of the role of autoantibodies in the pathogenesis and perpetuation of AIRD. Especially in recent years, detection technologies have advanced and have now been refined, allowing for the simultaneous assessment of multiple antibody specificities in unbiased non‐hypothesis‐driven approaches. As prices drop, we now anticipate the implementation of multiplex‐based approaches in diagnostic use to allow for simultaneous detection of autoantibody types and improved differentiation of AIRD groups. We believe that the validation of known antibodies and especially identification of relevant autoantibody specificities will provide insights into disease pathogenesis that can be applied in precision medicine. We anticipate that future research will further unravel the role of autoantibodies in AIRD pathogenesis, with the largest gains to be obtained in traditionally classified seronegative AIRD.

Disclosures

The authors state no conflicts of interest and have no disclosures. This research did not receive any specific grant from funding agencies in the public, commercial or not‐for‐profit sectors.

Author contributions

J. N. P., E. F. A. L. and M. B. were responsible for conception and writing of the manuscript. All authors contributed to substantial discussion of content, reviewing and revising the manuscript before submission. We apologize to those authors whose paper we were unable to cite in this review, which is due to page and reference limitations.

Supporting information

Table S1. References Table 1.

Table S2. References Table 2.

Table S3. References Figure 2.

Data Availability Statement

Not applicable.