Autoantibody Biomarkers in Rheumatic Diseases

Autoantibodies encountered in patients with systemic rheumatic diseases bear clinical significance as a biomarker to help or predict diagnosis, clinical phenotypes, prognosis, and treatment decision-making. Furthermore, evidence has accumulated regarding the active involvement of disease-specific or disease-associated autoantibodies in the pathogenic process beyond simple association with the disease, and such knowledge has become essential for us to better understand the clinical value of autoantibodies as a biomarker. This review will focus on the current update on the autoantibodies of four rheumatic diseases (rheumatoid arthritis, myositis, systemic sclerosis, and anti-neutrophil cytoplasmic antibody associated vasculitis) where there has been a tremendous progress in our understanding on their biological effects and clinical use.


Introduction
Biomarkers are not only helpful for understanding distinct pathophysiological processes of the disease but also useful for patient care in terms of screening and early diagnosis of the disease, prediction of prognosis, and treatment decision-making. Autoantibodies in systemic rheumatic diseases have been biomarkers largely due to significant and sometimes exclusive association with the disease itself or certain clinical phenotypes of the disease. In addition, recent studies have shown active involvement of these autoantibodies in the disease process beyond simple association. This review will discuss updated information on the pathogenic roles of autoantibodies in four systemic rheumatic diseases: rheumatoid arthritis (RA), myositis, systemic sclerosis (SSc), and anti-neutrophil cytoplasmic antibody associated vasculitis (AAV) and their values to guide diagnosis, prognosis, and treatment of corresponding diseases.

Autoantibodies in Rheumatoid Arthritis
RA is an autoimmune-mediated chronic systemic inflammatory disease that primarily affects synovial joints but often shows extraarticular manifestations. Autoantigens targeted by autoantibodies found in RA display a wide spectrum of cellular components, suggesting that RA is characterized by accumulated autoreactivities in both B and T cells [1]. Among diverse autoantibodies found in patients with RA, rheumatoid factor (RF) and anti-citrullinated protein antibody (ACPA) are the two most remarkable autoantibodies in RA that provide useful clinical and pathophysiological information. Currently, they are included in the 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) classification criteria for RA [2].

Pathogenic Roles of ACPAs
ACPAs are a group of autoantibodies found in 70%-90% of RA patients [4]. They are directed towards citrullinated proteins generated by post-translational modification on arginine residues and, therefore, can recognize a variety of articular and non-articular antigens, including type II collagen, fibrinogen, vimentin, and others [4]. Importantly, citrullination results from peptidyl-arginine deiminase (PAD) induced by inflammation [4]. Compared to RFs, ACPAs have a higher disease specificity (70%-80% versus 90%-95%) [4]. They are primarily generated among genetically susceptible individuals or shared epitope (HLA-DR4) carriers as a result of environmental stimuli, such as smoking, that induce a series of events: chronic inflammation, PAD induction, citrullinated autoantigen expression, and high-affinity loading of citrullinated proteins on shared epitope molecules of antigen-presenting cells [14]. Similar to RFs, they are associated with the radiologic progression of RA [7]. The pathogenic role of ACPAs has become more evident by an animal model where transferred antibodies directed to citrullinated fibrinogen or citrullinated collagen led to disease exacerbation in mice with mild experimental arthritis [15,16]. According to the current knowledge, the pathogenic roles of ACPAs can be summarized as follows.

Fc Receptor Binding
IgG antibodies exert an effect via multimeric interactions with the Fc receptor (FcR) by forming ICs [17]. Likewise, evidence suggests that ACPAs contribute to RA inflammation via IC formation with citrullinated proteins [18][19][20]. For example, citrullinated fibrinogen-containing ICs induce tumor necrosis factor α (TNFα) secretion from macrophages in FcγR and Toll-like receptor 4 (TLR4)-dependent manners, where both protein citrullination (compared to native protein) and IC formation greatly potentiates TNFα secretion from macrophages [18]. These results support the hypothesis that ICs containing citrullinated peptides may serve as danger-associated molecular patterns (DAMPs) to stimulate innate immune cells. Considering that RA autoreactivity also targets other potential DAMPs of endogenous origin (e.g., heat shock proteins) [21], the roles of other pattern recognition receptors than TLR4 and the differences between citrullinated and native DAMPs than fibrinogen need future research.
In addition to IC formation with citrullinated proteins and induction of cytokine secretion, FcR engagement by ACPAs are also able to polarize macrophages to the pro-inflammatory M1 subset

Assays to Detect ACPAs
ACPAs are usually measured by anti-cyclic citrullinated peptide (anti-CCP) assays. Based on the finding that the anti-perinuclear factor, an autoantibody highly specific for RA, is directed to citrullinated filaggrin, the first-generation anti-CCP (anti-CCP1) utilized filaggrin-derived cyclic rather than linear citrullinated peptides to improve antibody recognition but was not widely used due to low sensitivity [35]. Unlike anti-CCP1, the second and third generation tests (anti-CCP2 and anti-CCP3) are no longer based on filaggrin-derived peptides but on a mixture of synthetic cyclic peptides selected from libraries of citrullinated peptides that mimic conformational epitopes [36]. It is well-established that the anti-CCP2 possesses a superior sensitivity (around 70%-80%) than anti-CCP1 while maintaining very high specificity (98%-99%) [35]. Until now, anti-CCP2 has been considered as the gold standard of testing for ACPA [35]. In recent years, new ACPA tests, such as the anti-CCP3 assay, anti-mutated citrullinated vimentin, or others, have been commercially introduced. However, they have not shown a clear advantage over the anti-CCP2 assay [35].

RFs and ACPAs as Biomarkers
The most common use of RFs or ACPAs as biomarkers is to divide RA patients into seropositive (RF and/or ACPA positive) versus seronegative subsets that differ in risk factors, clinical outcomes, and treatment response. The risk factors of the seropositive RA and those of ACPA formation are in fact overlapped (e.g., genetic risk factors such as shared epitope and smoking) [4]. While both RF and ACPA associate with radiographic damage compared to the seronegative status, the former associates more with extra-articular manifestations than the latter [37]. As expected from the B cell involvement in autoantibody formation, treatment response to rituximab, a B cell-depleting agent, is more favorable with seropositive than seronegative RA [38]. Notably, abatacept, a T cell inhibitor, was also found to associated with a better response among seropositive than seronegative patients in both clinical trials and real-world data [39,40].

Autoantibodies in Myositis
Myositis represents a systemic autoimmune disease that primarily targets muscles and often involves extra-muscular organs. There are distinct clinical subsets of myositis that can be defined by unique clinical phenotypes and associated autoantibody profiles, suggesting a heterogeneity of disease pathophysiology [41]. To date, various autoantibodies have been identified in myositis, some specific to myositis (myositis-specific autoantibodies, MSAs), and others also found in other autoimmune diseases (myositis-associated autoantibodies, MAAs). A high specificity for myositis and a striking association between individual MSAs and distinct clinical phenotypes make MSAs an excellent tool for subsetting patients and predicting disease courses and outcomes. Despite their usefulness, the new 2017 classification criteria for adult and juvenile myositis by the European League Against Rheumatism/American College of Rheumatology Classification (EULAR/ACR) criteria included only anti-Jo-1 to identify patients with myositis [41]. Table 1 summarizes various MSAs in terms of their counterpart autoantigens, prevalence, and associated clinical phenotypes. The most well-known MSAs are anti-aminoacyl-tRNA synthetase (anti-ARS) antibodies directed to the enzymes that catalyze the binding of amino acids to corresponding tRNA to form aminoacyl-tRNA, with anti-Jo-1 or anti-histidyl-tRNA-synthetase being the most common (Table 1). Transition from CADM to overt myositis; dysphagia in both Caucasians and Asians but ILD only in Asians.
Many of myositis-specific autoantigens, including ARS, are ubiquitously expressed intracellular proteins involved in various cell functions, particularly transcription or translation (Table 1). They are overexpressed and modified in apoptotic and regenerating muscle fibers [42], some of which are also expressed in tumor cells showing a link between cancer and autoimmunity to muscles [43].
While the associations of MSAs with myositis or certain clinical phenotypes of myositis are well-known, only limited information is available about the roles of MSAs in myositis pathogenesis. Animal models and in vitro studies showed that Jo-1 or histidyl-tRNA-synthetase might function as a chemokine and play a role in inducing myositis [44]. A mouse model immunized by Jo-1 further showed inflammation in the muscles and lungs, supporting a pathogenic role of Jo-1 in myositis development [45]. However, it is unknown whether or how anti-Jo-1 antibody regulates chemokine activity of Jo-1. Rather, anti-Jo1 can form IC that induces type I interferon production by plasmacytoid dendritic cells [46].
Recently, a new subset of myositis, referred to as immune-mediated necrotizing myopathy (IMNM), has been identified [47]; this subset of myositis is pathologically characterized by marked muscle fiber necrosis and regeneration with minimal inflammatory infiltrates and serologically by anti-signal recognition particle (anti-SRP) or anti-3-hydroxy-3-methylglutaryl-coenzyme A reductase (anti-HMGCR) antibodies. The latter antibody was particularly positive among IMNM patients with previous statin exposure [48]. While the roles of many other MSAs remain largely unclear, the pathogenic roles of these two antibodies have been better understood by animal models, as well as in vitro studies [49][50][51].
Muscle atrophy is one of the typical features of IMNM, together with necrosis and regeneration. Incubation with anti-SRP or anti-HMGCR antibodies led to muscle fiber atrophy and increased expression of atrophy pathway proteins in vitro [49]. In either anti-SRP or anti-HMGCR-positive muscle sections, greater sarcolemmal MAC deposits were observed in lesional areas than in anti-Jo-1-positive muscle tissues, correlating with the degree of muscle necrosis [50]. The pathogenic role of anti-SRP and anti-HMGCR IgG was directly shown in an in vivo study where passive transfer of anti-SRP or anti-HMGCR IgG led to significant muscle weakness, histologic deposits of MAC, and necrotic muscle fibers [51].

Assays to Detect MSAs
Historically, MSA identification has been done by protein or RNA immunoprecipitation that compared the molecular weights of precipitated peptides in reference to the positive control. For recently identified novel MSAs, including anti-MDA5, anti-NXP2, and anti-TIF1 antibodies, enzyme linked immunosorbent assays (ELISAs) were developed but their commercial use is still limited due to low prevalence of individual MSAs among patients with myositis. Recently, the line blot assay (LBA; e.g., Euroline Myositis Profile kit by Euroimmun Lübeck, Germany), in which antigens are placed on nitrocellulose as narrow lines, has been introduced particularly to detect multiple autoantibodies simultaneously. Parallel detection of multiple MSAs by the LBA has advantages, considering the low prevalence of each MSA. In a study by Espinosa-Ortega et al. [52], LBA was compared with immunoprecipitation among patients with myositis and showed a good correlation, except for the anti-TIF1γ antibody, for which LBA showed a lower sensitivity than immunoprecipitation. Moreover, LBA detected true positive anti-Jo1 samples that were missed by immunoprecipitation.

MSAs as a Biomarker
The most well-established role of MSAs as a biomarker would be based on their highly specific association with myositis and/or distinct clinical subsets (Table 1). Since patient prognosis is different according to individual MSAs and MSA-associated clinical subsets, MSAs are useful, not only for diagnosis, and but also for prognosis prediction and treatment decision of myositis patients.

Anti-Jo-1 and Other Anti-ARS
Among anti-ARS antibodies, anti-Jo-1 is the most common, found in up to 40% of all myositis patients, while the other anti-ARS antibodies individually constitute less than 5% (Table 1). The most striking feature associated with anti-ARS or anti-Jo-1 is observed as the anti-synthetase syndrome, where interstitial lung disease, myositis, and the presence of anti-ARS are typically co-present with or without other clinical findings, such as the Raynaud phenomenon, arthritis, or mechanic's hands. Interstitial lung disease (ILD) occurs in up to 90% of individuals possessing anti-Jo-1 antibodies [53]. Since ILD harbors a substantial impact on the survival of patients with myositis, the presence of anti-Jo-1 mandates a careful screening for ILD.
While patients with individual anti-ARS antibodies share a clinical phenotype called anti-synthetase syndrome, there might be slight differences among patients with different anti-ARS [54,55]. For example, other anti-ARS than anti-Jo-1 antibodies might be associated with ILD in the absence of myositis [55,56]. While other MSAs, such as anti-melanoma differentiation-associated protein 5 (anti-MDA5) antibodies, are also associated with ILD, ILD associated with anti-ARS are often slowly progressive compared to rapidly progressive ILD associated with anti-MDA5 [57].

Anti-MDA5
Anti-MDA5 antibodies are DM specific, found in up to 50% of adult DM patients [58]. They can be seen in both classical and clinically amyopathic DM (CADM) but more preferentially in CADM [59,60]. Due to their association with CADM, anti-MDA5 antibodies were also called anti-CADM-140 antibodies before target antigens were identified [59]. Anti-MDA5 antibodies were in fact found to be an independent indicator of rapidly progressive ILD in both classic DM and CADM [60]. Since the prognosis of anti-MDA5-associated ILD is invariably fatal due to respiratory failure, early aggressive treatment has been encouraged [61]. Recently, successful treatment against humoral immunity using rituximab or plasma exchange has been reported against anti-MDA5-positive rapidly progressive ILD among DM patients [61,62].

Anti-Mi2
Anti-Mi2 is a DM-specific autoantibody characterized by DM rashes (V neck or shawl sign) and good prognosis due to the absence of internal organ involvement [64].

Anti-Transcription Intermediary Factor 1 (Anti-TIF1)
Anti-TIF1 precipitates TIF1γ of 155 Kda with or without TIF1α of 140 KDa, thus formerly called anti-p155 or anti-155/140 antibodies [65]. While they are preferentially associated with both adult and juvenile DM [65,66], their clinical implications are much different according to age at onset of the affected population.
3.3.6. Anti-Small Ubiquitin-Like Modifier Activating Enzyme (Anti-SAE) Anti-SAE antibodies are directed to small ubiquitin-like modifier (SUMO) activating enzyme specific for DM. Anti-SAE-positive patients often present as CADM that transitions to overt myositis within months [79]. Although they are linked with a high frequency of systemic features, clinical phenotypes can be different across different ethnicities. For example, frequent dysphagia (30%-78% among anti-SAE-positive patients) has been reported in both UK and Asia [79][80][81], but high prevalence of ILD has only been reported in Asians, with ILD involving two-thirds of anti-SAE-positive patients [80,81].

Anti-SRP and Anti-HMGCR
These two MSAs are associated with IMNM and poor treatment response [47]. The latter is particularly noted among patients who had statin exposure prior to myositis onset [48].

Pathogenic Roles of SSc Specific Autoantibodies
While SSc-specific ANAs are associated with SSc or certain clinical phenotypes of SSc, the pathogenic roles of these autoantibodies are largely unknown, particularly considering the nuclear location of the antigens, with limited access by circulating autoantibodies. However, recent studies observed that SSc patients also have non-ANA autoantibodies, particularly those targeting receptors and signaling pathways, with functional consequences (Table 2). In addition, certain SSc-specific autoantibodies suggest that SSc may develop as a consequence of an immune response to the aberrant expression of self-antigens by tumor cells [86].
Although not specific for SSc, anti-endothelial cell antibodies (AECAs) have long been recognized in patients with SSc [89]. A potential pathogenic role of AECA has been suggested by their ability to induce apoptosis of human dermal microvascular endothelial cells in the presence of activated NK cells via the Fas pathway [90]. Recent evidence suggests that AECAs at least include those directed against (intercellular adhesion molecule-1) ICAM-1 [91]. Exposure to IgG from either AECA-positive patients or to anti-ICAM-1 antibodies was similarly found to induce generation of reactive oxygen species (ROS) and vascular cell adhesion molecule-1 (VCAM-1) expression from endothelial cells, indicating a possible contribution of AECAs to SSc vascular lesions [91]. In line with this, AECAs are known to associate with more severe disease phenotypes in terms of internal organ involvement, such as lung fibrosis, pulmonary arterial hypertension (PAH), and ischemic digital lesions [89,92].
In addition to AECA, SSc patients were found to have autoantibodies for renin-angiotensin and endothelin systems, both of which receptors are widely expressed throughout the vascular system and immune cells [93]. Patients with anti-angiotensin II type I receptor (AT1R) and/or endothelin-1 type A receptor (ETAR) antibodies were found to be at high risk for vascular complications, including PAH and ischemic digital lesions [94][95][96]. Increased signaling through the angiotensin and endothelin receptors has been linked to vasculopathy and fibrosis by increasing tissue growth factor β (TGFβ), interleukin-18, and VCAM-1 expression, and ROS generation from human dermal microvascular endothelial cells [96,97]. When healthy C57BL-6 mice were treated with repeated intravenous administrations of IgG from SSc patients positive for anti-AT1R or anti-ETAR antibodies, the lung tissue showed thickened airway vessels and elevated cell density in interstitial tissue [97]. However, the study did not confirm whether these histological changes were specifically induced by anti-AT1R or anti-ETAR antibodies, since it could be due to other agnostic autoantibodies, such as anti-platelet-derived growth factor receptor (anti-PDGFR) antibodies that can be co-present in SSc patient sera.
There are two types of anti-PDGFR among patients with SSc: nonagnostic and agonistic [98]. Only agonistic anti-PDGFR antibodies are SSc-specific, directed for stimulatory epitope of PDGFRα. Agonistic antibodies were found to cause receptor tyrosine phosphorylation of fibroblast PDGFR, which ultimately leads to activation of transcriptional factors to produce extracellular matrix (ECM) and reactive oxygen species (ROS) [99]. Of note, these antibodies stimulated PDGFR for an aberrantly longer time than PDGF itself [99]. In SCID mice, transferred SSc patient IgG or agonistic anti-PDGFR antibodies induced increased collagen deposition and fibroblast activation from the 3D bioengineered skin grafts that contained human keratinocytes and fibroblasts from healthy donor skins [100]. Similar to agonistic anti-PDGFR antibodies, anti-fibrillin-1 antibodies have also been reported to increase ECM via increased TGFβ production [101].
Anti-acetylcholine receptor antibodies have been suggested to associate with the gastrointestinal (GI) motility dysfunction of SSc. Certain SSc patients were found to have autoantibodies against myenteric neurons of the GI tract [102], and passive transfer of IgG from SSc patients with high titer anti-myenteric neuronal antibodies led to a decreased myenteric action potential formation [103]. In addition, inhibitory antibodies for type 3 muscarinic receptors (M 3 Rs) were found to have a role in GI dysmotility by working not only on myenteric neurons but also on GI smooth muscles [104].
Similarly observed in myositis, certain autoantibodies found in patients with SSc indicates a high risk of concurrent malignancies. For instance, temporal clustering of SSc and cancer onset among patients with anti-RNAP I/III has been noted [105]. Among patients who had cancer diagnoses and produced anti-RNAP I/III antibodies, SSc tends to develop within 2 years of the cancer diagnosis. Interestingly, tumor expression of nucleolar RNAP III was enhanced exclusively in patients with anti-RNAP I/III antibodies than those without. Coupled with later findings that genetic alteration of the POLR3A locus was exclusively found in anti-RNAP III-positive SSc patients with cancer than without and that POLR3A mutation triggered cellular immunity and cross-reactive humoral immune responses [86], the current hypothesis is that malignancy may initiate the scleroderma-specific immune response and drive disease in a subset of scleroderma patients. Anti-RNPC3 antibody directed for the minor spliceosome complex is another autoantibody associated with cancer-associated SSc, particularly among triple negative (negative for ACA, ATA, and anti-RNAP III) SSc patients [87,88]. Anti-RNPC3-positive SSc patients showed a 4-fold increased risk for cancer diagnosis within 2 years of SSc onset, and 89% of them possessed severe ILD at baseline [88].

Assays to Detect SSc Specific Autoantibodies
Very similar to MSAs, SSc-specific ANAs have traditionally been detected by immunoprecipitation, and the recent LBA method is also commercially available for SSc (e.g., Euroline ANA Profile 3 by Euroimmun). When LBA was used in an Australian cohort of 505 patients with SSc, subclassification and disease stratification by autoantibody profile was correlated with characteristic clusters of clinical phenotypes [106], suggesting a usefulness of LBA.
Functional antibodies have been detected for research purposes and not standardized yet, possibly due to uncertainties of pathogenic epitopes [98,107]. However, due to their association with functional consequences, development of commercial assays is particularly important for these antibodies in terms of prognosis prediction and treatment decision-making on B cell depletion therapy.

Classic SSc Specific ANAs
While ACA, ATA, and anti-RNAP III are the most common SSc-specific ANAs [84], the prevalence of each antibody substantially varies depending on the ethnic groups and detection methods [108]. The presence of SSc-specific ANAs among patients with Raynaud's phenomenon indicates a high risk for future SSc [109]. ACA ACA, first described as an indicator of CREST (Calcinosis, Raynaud's phenomenon, Esophageal hypomotility, Sclerodactyly, and Telangiectasia) syndrome, is now considered as a variant of limited cutaneous SSc [110]. According to the European Scleroderma Trials and Research group (EUSTAR) database, ACA is the most common autoantibody for limited cutaneous SSc found in up to 50% of them [84]. Typically, ACA-positive SSc patients tend to have limited cutaneous SSc and better prognosis due to a lack of internal organ involvement, such as ILD, cardiomyopathy, or renal crisis [111], except for PAH, which can occur later in the course of the limited disease [112]. In this regard, the evidence-based DETECT algorithm for the identification of SSc patients at high risk for developing PAH includes ACA as one of the useful screening items [113].

ATA
Detected in 20%-30% of SSc patients, ATAs are typically associated with diffuse cutaneous SSc and internal organ involvement [111]. Approximately two-thirds of ATA-positive SSc patients were found to have diffuse cutaneous SSc, and ILD occurs in 70% of them [111]. Since the main cause of death among SSc is ILD and PAH, ATA is a marker of poor prognosis [114]. Other important organ involvements associated with ATA include cardiomyopathy and peripheral vascular complications such as digital ulcer (DU) and gangrene, particularly early in the disease course [115].

Anti-RNAP III
The frequency of Anti-RNAP III in SSc patients varies among ethnic groups: higher frequency in North American Caucasian and UK patients (~25%) in comparison with French (~9%) or Japanese patients (~5%), with a pooled prevalence of 11% [116]. Majority of anti-RNAP III-positive patients show diffuse cutaneous SSc with rapid progression followed by regression of skin thickening [116,117].
Typically, anti-RNAP III positivity has been associated with a higher risk of renal crisis, but severe ILD was rare [115,118]. Another clinical phenotype associated with anti-RNAP III is gastric antral vascular ectasia (GAVE), observed in the EUSTAR database where 48% of patients with GAVE were anti-RNAP III-positive, compared with 16% of those without [119], suggesting anti-RNAP III involvement in SSc-related vasculopathy. While a recent study showed that anti-RNAP III antibody response is triggered by neoantigens by cancer-related genetic alteration, the prevalence of cancers among Anti-RNAP III-positive patients is generally low [86].
Anti-U3 RNP (Anti-Fibrillarin) Anti-U3 RNP recognizes a 34kD component of the U3 RNP complex called fibrillarin. They are detected in −10% of SSc patients, with a predilection for male sex and African-American patients [115,120,121]. Anti-U3 RNP has been suggested as a marker for severe SSc, reportedly linking with diffuse cutaneous SSc and diverse internal organ derangements of cardiac, renal, musculoskeletal, pulmonary, and gut systems [115,120,122,123], particularly in the form of pulmonary arterial hypertension; thus, poor survival [120].

Anti-Th/To
Anti-Th/To antibodies primarily bind to two RNA processing enzymes, RNase P and RNase MRP. They are relatively specific for SSc, seen in 1%-10% among patients with SSc. In general, they are associated with limited cutaneous SSc but have a high frequency of internal organ involvements, such as ILD, PAH, pericarditis, and/or renal crisis [115,124,125]; hence, a poor prognosis [124].

Anti-Eukaryotic Initiation Factor 2B
Observed in approximately 1% of British patients with SSc (7/548), anti-eukaryotic initiation factor 2B antibody is a cytoplasmic antibody strongly associated with diffuse cutaneous SSc and ILD [127].

Functional Antibodies
As mentioned above, functional antibodies are typically directed for receptors and signaling pathways involved in fibrosis, the vascular system, and neurotransmission. The exact prevalence of these antibodies is not known due to the unavailability of commercial assays, but the successful treatment of SSc skin and lung using rituximab supports the clinical importance of functional autoantibodies of SSc [128][129][130].

Autoantibodies in ANCA-Associated Vasculitis
Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitides (AAV) are a spectrum of systemic necrotizing vasculitis predominantly affecting small vessels of diverse organs. They are subdivided into three distinctive disease entities of granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA), and eosinophilic GPA (EGPA), which are associated with ANCA in more than 90% of cases under current clinical assays [131]. GPA and EGPA share extravascular necrotizing granulomatous inflammation or granulomatosis that most often affects respiratory tracts but potentially any organs [132]. While autoantibodies or ANCAs are directly involved in the pathogenesis of the disease [133,134], AAV is featured by a lack of immune deposits (pauci-immune).
Antineutrophil cytoplasmic antibodies (ANCAs) are autoantibodies directed against diverse antigens within the primary granules of neutrophils and the lysosomes of monocytes [132]. ANCAs may develop against any of the granular proteins, but the clinically relevant ones are those directed against myeloperoxidase (MPO) and proteinase 3 (PR3). While both MPO and PR3 are cytoplasmic antigens, ethanol fixation leads to dissolution of proteins within the neutrophil primary granules, leaving a perinuclear pattern for MPO-ANCA and a cytoplasmic pattern for PR3-ANCA [135]. The distinction between different AAVs based on ANCA patterns is far from perfect; PR3-ANCAs were present in about two-thirds of patients with GPA but also in one-quarter of patients with MPA, whereas MPO-ANCAs were in the majority of patients with MPA but also in up to one-quarter of patients with GPA [136]. Since the large-scale genome-wide association study demonstrated that the strongest genetic association comes with ANCA antigen specificity, rather than with a specific clinical syndrome [137], differentiating between patients with MPO-ANCAs or PR3-ANCAs might be more relevant, instead of distinguishing between patients with GPA, MPA, and EGPA [136].
Although current classification systems do not include ANCA as a diagnostic tool for AAV, the 2010 EULAR have pointed to considering ANCA in diagnostic and classification criteria for systemic vasculitis [138], and the 2012 revised Chapel Hill Consensus Conference Nomenclature of Vasculitides (CHCC 2012) called for adding a prefix of ANCA specificity to each AAV (i.e., MPO-ANCA GPA, and ANCA-negative EPA) [135], acknowledging the clinical relevance of ANCA in AAV.

Pathogenic Roles of ANCA
Direct evidence for ANCA pathogenicity comes from an animal model in which the passive transfer of MPO-ANCA resulted in pauci-immune crescentic GN and necrotizing vasculitis in small vessels [133]. Effective disease prevention by neutrophil depletion in this model further indicates a central role of neutrophils in ANCA-induced crescentic GN [139]. Unlike an animal model by MPO-ANCA, the development of PR3-ANCA-induced animal models has been less successful in earlier studies, possibly due to lack of PR3 expression on murine neutrophils and different FcR affinities for IgG across species in mice/rodents [140]. However, pathogenic PR3-ANCA has been also demonstrated in mice with a humanized immune system [134]. Figure 1 summarizes the pathogenic processes elicited by ANCAs. While MPO and PR3 are intracellular antigens, they can be accessed by antigen-presenting cells under a specific condition called priming [141]. Neutrophils primed with inflammatory stimuli (cytokines, lipopolysaccharides, complement fragments, etc.) release PR3 or MPO from primary granules to the cell surface [142], which permits ANCAs to interact with them. Cell surface displays of ANCA antigens can be achieved by their binding to the cell surface after release from primed or activated neutrophils [143] or by direct translocation of these antigens towards cellular membrane under inflammatory microenvironments [144]. Importantly, surface displays of ANCA antigens can be induced on apoptotic neutrophils independent of priming [145,146], which provides an alternative mechanism beyond neutrophil activation to allow interaction between ANCAs and ANCA antigens.
ANCA acts via antigen crosslinking by ANCA Fab or via FcγR engagement [147,148]. ANCA-activated neutrophils generate a respiratory burst with a production of oxygen radicals, degranulate destructive enzymes, and extrusion of NETs [149]. NETosis contributes to AAV by causing endothelial damage, lymphocyte stimulation, and activation of alternative complement pathways [150]. Since NETs further expose ANCA antigens to ANCA, the vicious cycle forms to perpetuate inflammation [151,152]. Besides neutrophil activation, ANCA also promotes neutrophil adherence on capillary endothelium by neutrophil integrin activation and upregulation [153]. Vascular injury occurs via ANCA-activated neutrophils adherent to endothelial cells [149] or via direct effects of internalized ANCA antigens by endothelial cells [154]. When internalized, PR3 induces endothelial apoptosis, whereas MPO induces intracellular oxidants [154]. While PR3 ANCA or MPO ANCA utilize a number of similar mechanisms to cause disease [132], different clinical phenotypes in association with ANCA specificity indicates there could be different inflammatory pathways for two different ANCAs [154,155]. However, it is largely unknown if and how PR3 and MPO ANCAs behave differently.
which permits ANCAs to interact with them. Cell surface displays of ANCA antigens can be achieved by their binding to the cell surface after release from primed or activated neutrophils [143] or by direct translocation of these antigens towards cellular membrane under inflammatory microenvironments [144]. Importantly, surface displays of ANCA antigens can be induced on apoptotic neutrophils independent of priming [145,146], which provides an alternative mechanism beyond neutrophil activation to allow interaction between ANCAs and ANCA antigens.

Assays to Detect ANCAs
Laboratories have used indirect immunofluorescence (IIF) assays as a screening to detect ANCA, semi-quantitatively reporting serum dilution factors. However, the positive result should be confirmed by antigen-specific assays. Moreover, atypical ANCAs that do not target PR3 or MPO (positive IF but negative ELISA) can be found in a range of nonvasculitic conditions, such as inflammation (drugs, chronic infection, cancer, and other autoimmune diseases). Since it is technically difficult to differentiate perinuclear ANCA (or MPO ANCA) patterns from antinuclear antibodies (ANA) based on ethanol fixation, some laboratories use both formalin and ethanol fixation to distinguish ANCAs from ANAs [156].
The first-generation ANCA ELISA developed in the late 1990s was a direct method immobilizing ANCA antigens to the surface of an ELISA plate, which often has a lower sensitivity than IIF, despite its high specificity (>95%) [157]. The second-generation capture ELISA uses secondary antibodies attached to the plate that specifically capture ANCA antigens, showing a better sensitivity than the direct ELISA [158,159]. The third-generation anchor ELISA immobilizes antigens to the ELISA plate via a bridging molecule to preserve epitope binding sites that could have been masked by capturing antibodies. The anchor ELISA was better than the direct ELISA in terms of sensitivity [160,161], which, in some studies, showed a superior sensitivity to capture ELSIA and IIF [162].
While the 1999 international consensus statement recommends a testing strategy that adopts IIF as an initial screening tool followed by confirmation with ELISA [163], the updated 2017 international consensus statement of ANA testing proposes that high-quality immunoassays can be used as the primary screening method for patients suspected of having the AAV without the categorical need for IIF, acknowledging the excellent sensitivity of the recent capture and/or anchor assays [136]. However, approximately 10% of patients with clinical features and pathology of AAV are negative for ANCA IIF [156].

ANCA as Biomarkers
As mentioned above, the positivity of MPO or PR3 ANCA in AAV ranges above 90% in generalized GPA and MPA, particularly in their active phase [131], and thus, may improve diagnostic sensitivity when used under high suspicion of vasculitis. In particular, rapid ANCA tests, when urgent therapeutic decisions are required, might be helpful as an adjunct before biopsy results are pending or not available.
In addition to diagnostic benefits, studies have shown that ANCA specificity is the major determinant of clinical manifestations and outcomes. In the analysis on 502 ANCA-positive patients with biopsy-proven AAV, MPO-ANCAs were present in >80% of patients with isolated crescentic glomerulonephritis, whereas PR3-ANCAs were in >80% of patients with lung cavities or upper airway involvement [164]. Similarly, the RAVE trial (rituximab versus cyclophosphamide for ANCA-associated vasculitis), the largest randomized controlled trial in AAV so far, has shown a higher rate of renal involvement (79% versus 59%) and peripheral neuropathy (27% versus 15%) among patients with MPO-ANCAs than with PR3-ANCAs and the opposite pattern with constitutional symptoms (53% versus 71%), ear/nose/throat involvement (33% versus 70%), and lung nodules or cavities (12% versus 28%) [165,166]. According to a study on 673 patients with GPA or MPA from five different randomized controlled trials, it was found that the addition of ANCA specificity to a clinical-only clustering strategy resulted in a better separation of patients with different clinical phenotypes and outcomes [167]. The study showed a higher relapse with PR3-ANCA but higher mortality without PR3-ANCA, indicating different treatment outcomes and relapse rates according to the ANCA specificity. A consistent finding was observed by a systematic review in which patient survival was worse in MPO-AAV [168]. However, outcomes on renal survival are controversial, with some studies reporting a better prognosis with MPO-ANCA [169] and others with PR3-ANCA [170,171].
In the RAVE trial [165] that compared cyclophosphamide and rituximab against AAV, ANCA specificity was not predictive of complete remission at six months in either the entire cohort or in each of the two treatment groups. However, a post hoc analysis showed that complete remission at six months was achieved with a similar frequency in both treatment groups among 66 MPO-ANCA-positive patients (61% versus 64%), whereas rituximab was significantly more effective than cyclophosphamide among 131 PR3-ANCA-positive patients (65% versus 48%; p = 0.04) [166].
According to the WEGENT trial that compared azathioprine versus methotrexate for maintenance therapy of AAV, only 30% of patients with GPA or MPA remained relapse-free during approximately 12 years of follow-up [172]. The relapse of AAV was often preceded by rising titers of ANCA, and there was rarely a relapse among patients with undetectable ANCAs [173]. The predictability of the ANCA titer change was particularly relevant for renal than nonrenal relapse, regardless of ANCA specificity [174]. While certain patients remain in clinical remission despite high titers of ANCA, the epitopes recognized by ANCAs were found to be different between inactive versus active phases, suggesting that ANCAs reacting with pathogenic epitopes are more relevant than titers [175].

Conclusions
The presence of disease-specific or disease-associated antibodies is the hallmark of the autoimmunity of the given rheumatic disease. Each rheumatic disease exhibits a unique pattern of antibody profiles that provides useful information for diagnosis, clinical phenotypes, and prognosis. Since the antibody titers are often correlated with disease activity or relapse, it is expected that they are actively involved in disease pathogenesis. Although exact pathophysiologic roles of the disease-specific or disease-associated antibodies remain largely unknown, tremendous progress has been made on this issue during the past decade. RA, myositis, SSc, and AAV are the prototype diseases in which disease-specific, or sometimes, disease-associated, autoantibodies are not only acknowledged as biomarkers to help diagnosis, predict prognosis, and guide treatment decision-making, but also play important roles to perpetuate and aggravate the disease process. Therefore, values of these autoantibodies as a biomarker would be potentiated by understanding their pathogenic roles closely linked with clinical observations.