Pathogenesis of ANCA-Associated Vasculitis

Abstract

Antineutrophil cytoplasmic autoantibodies (ANCA) associated vasculitis (AAV) are a group of systemic vasculitis characterized by inflammation and necrosis of blood vessel walls. Genetic, epigenetic and environmental factors contribute to the etiology and pathogenesis of AAV. Based on currently available clinical and experimental evidence, it is reasonable to conceptualize that in predisposed patients, different triggers can lead to the production of autoantibodies (ANCA) that in the context of an inflammatory environment can cause tissue inflammation and vascular injury. Several different pathways and mechanisms in the pathogenesis of AAV are described in this contemporary review.

Introduction

Three systemic autoimmune small vessel vasculitis syndromes are associated with antineutrophil cytoplasmic autoantibodies (ANCA) and are collectively known as ANCA-associated vasculitides (AAV). (1). AAV comprises granulomatosis with polyangiitis (GPA, formerly known as Wegener’s granulomatosis), microscopic polyangiitis (MPA) and eosinophilic granulomatosis with polyangiitis (EGPA, formerly known as Churg-Strauss syndrome)(1). These syndromes have distinguishing clinical manifestations and histological findings. GPA and EGPA are characterized by granulomatous inflammation that usually affects the respiratory tract, which is – by definition – absent in MPA. The granulomatous inflammation of GPA is predominantly neutrophilic, whereas that of EGPA is eosinophil. In contrast, the clinical manifestations caused by necrotizing small vessel vasculitis and capillaritis, such as diffuse alveolar hemorrhage, mononeuritis multiplex or glomerulonephritis, are shared between all three syndromes.

Two different types of ANCA are associated with these syndromes(2). ANCA directed against the neutrophil serine protease, proteinase 3 (PR3) cause a cytoplasmic (C-ANCA) staining pattern on ethanol fixed neutrophils by indirect immunofluorescence microscopy. PR3-ANCA are the predominant ANCA type in GPA, less common in MPA, and the rare exception in EGPA(2). In contrast, ANCA directed against myeloperoxidase (MPO) generating a perinuclear (P-ANCA) fluorescence pattern are found in the majority of patients with MPA, but only in about 5–10% of patients with GPA(2). In EGPA, ANCA are usually of the MPO-ANCA type and can be detected in 30–70% of patients (3–5).

There is likely a contribution of genetic, epigenetic and environmental factors in the etiology and pathogenesis of AAV. In this review, we aim to summarize and discuss the current views and insights on the pathogenesis of AAV with particular focus on GPA and MPA.

Epidemiology of AAV

AAV have a reported incidence of 10–20 cases per million per year(6). The distribution of the specific syndromes varies geographically and with ethnicity(7). In different populations the annual incidence of GPA ranges between 4.9–10.6 cases/million and the reported point prevalence varies between 24–157 cases/million(6, 8–10). GPA is more common in Northern than in Mediterranean Europe, where MPA is more common(7). GPA is also exceedingly rare in African-Americans and in Japan, where it is almost exclusively encountered on the Northern island of Hokkaido (11, 12). A similar latitude-dependent incidence gradient has also been documented in New Zealand(13). Furthermore, in New Zealand as well as in France the incidence of AAV is higher in the populations of European ethnicity than among Non-Europeans, Asians or Pacific Islanders (13, 14).

Etiology and Pathogenesis

The etiology of the AAV syndromes remains unclear. Like for many other polygenic systemic autoimmune diseases, AAV are in all likelihood the result of complex interactions between genetic factors predisposing for the loss of self-tolerance and autoimmunity and triggering environmental exposures. These factors induce and maintain inappropriate lymphocyte activation and autoantibody production which lead to tissue injury and predispose to disease relapses. Table 1 summarizes different mechanisms and pathways that may contribute to the development of AAV.

Table 1.

Different potential factors in the pathogenesis of ANCA-associated vasculitis

Mechanism	Evidence
Genetic predisposition	Genetic association studies of different SNPs of the following genes: HLA, SERPINA1, CTLA 4, PRTN3 and PTPN22
Environmental trigger	Silica as inflammasome complex activator, drugs (propylthiouracil, hydralazine and penicillamine) which could have polyclonal B lymphocyte stimulatory properties that may induce the production of ANCA
Infectious trigger	Staphylococcus aureus eliciting molecular mimicry and T and B cell activation via superantigens.
B-cells	B-cells are precursors of ANCA forming plasma cells. Increased proportion and total number of B-cells have been identified in AAV patients. Elevated B-lymphocyte stimulator factors including BAFF have also been identified in AAV patients.
T-cells	T-cells are found within granulomas as well as in other lesions present in AAV. Elevated levels of markers of T-cell activity such as soluble interleukin-2 (IL-2) receptor, neopterin, and soluble CD30 have been shown to be associated with disease activity. Increased T-cell number and activity(CD4+CD25+ and CD4+CD25+CD134+/GITR+) have also been found. Functional defect in Tregs (CD4+CD25high FoxP3+CD127low) have been described in patients in remission. GPA patients in remission allow an increased amount of Th17 cells(CD4+IL-17+) reactive to PR3.
Autoantibodies (ANCA)	Pathogenic role of ANCA has been supported by several in vitro and also in vivo animal models. Activation of neutrophils and monocytes by ANCA antibodies causes generation of reactive oxygen species, release of proteases and cytokine production by which inflammation and tissue damage is created. ANCA antibodies also induce apoptosis in endothelial cells.
Complement pathway	Increased neutrophil activity via the alternative complement pathway→C5a primes neutrophils and enhances ANCA induced neutrophil activation. Complement activation and its resulting products promote inflammation and enhance tissue damage.

SNP: single nucleotide polymorphism; HLA: human leukocyte antigen;CTLA4: Cytotoxic T-lymphocyte antigen 4; ANCA: anti-neutrophil cytoplasmic antibody; AAV: ANCA-associated vasculitis; BAFF: B cell activating factor; GPA: granulomatosis with polyangiitis; PR3: proteinase 3

Genetic Predisposition

There are studies of rare cases in which single gene defects can cause AAV-like phenotypes(15–17); however, the vast majority of genetic association studies in AAV have been in the context of single nucleotide polymorphisms (SNPs) with small effect size. A familial aggregation study found a relative risk of 1.56 for GPA among first degree relatives of patients with GPA, which is similar to the risk seen in rheumatoid arthritis(18). Familial clustering has also been reported with other AAV syndromes (19, 20). Several candidate genes have been investigated in different cohorts of patients with AAV, and SNPs have been found in several genes coding for proteins involved in the immune response(21).

Variants most strongly and most reproducibly associated with AAV are found in the human leukocyte antigen (HLA) and PTPN22 genes (21). HLA genes, located on the short arm of chromosome 6, have been associated with various autoimmune disorders such as rheumatoid arthritis, giant cell arteritis, Behçet’s disease, and antiglomerular basement membrane disease(22, 23). Multiple HLA genes have been confirmed to be prevalent in different AAV populations; for example, in a Caucasian population of German patients with GPA, a group of investigators reported that HLA-DRB1*04 was over-represented in these individuals(24), whereas another group located the risk alleles as HLA-DPB1*0401 in a cohort of 282 subjects with GPA(25). In the Asian population, HLA-DRB1*0901 has been found to be associated with MPA and MPO-ANCA-positive vasculitis patients in a Japanese population (26, 27).

Most candidate gene studies in AAV comprise only GPA cohorts. MPA and EGPA are usually underrepresented in many genetic association studies. Nevertheless, interesting genetic differences are emerging between the different syndromes. For instance, the HLA-DPB1*0401 variant represents a strong and reproducible genetic risk factor for GPA, but not for MPA or EGPA (28–30). Interestingly, African Americans with PR3-ANCA associated vasculitis had a 36 fold higher likelihood to have the HLA-DRB1*15 genotype than community-based controls, and the HLA-DRB*1501 allele, which is of Caucasian rather than African descent, conveyed a 73.3-fold higher risk for PR3-ANCA associated vasculitis (31)

In contrast, an increase in the PTPN22 620W allele, which has been associated with other autoantibody associated autoimmune diseases and has thus been implicated in the regulation of B lymphocyte activity, was found to be associated with both GPA and MPA(32–34).

Cytotoxic T-lymphocyte antigen 4 (CTLA4), expressed mostly on CD4 positive T lymphocytes, exerts an inhibitory function on T lymphocytes by binding to CD80 and CD86 on antigen presenting cells. CTLA4 competes with the co-stimulatory molecule CD28, which exerts a stimulatory effect on T lymphocytes, for binding to CD80 and CD86. If T lymphocytes are activated via the T cell receptor and CD28, CTLA4 expression is increased, probably as a regulatory mechanism. Via these mechanisms CTLA4 is thought to maintain peripheral self-tolerance(35). Elevated levels of CTLA4 have been found in GPA, undoubtedly a reflection of T lymphocyte activation(36). CTLA4 is coded for by the CTLA4 gene, and CTLA4 polymorphisms, which may negatively affect expression or function of CTLA4, have been associated with autoimmune diseases including GPA (33, 37, 38). Recently, a meta-analysis of 7 studies showed an increased susceptibility for AAV associated with CTLA-4 polymorphisms in European patients (39).

Another gene of interest is PRTN3, which codes for PR3, the most prominent target antigen for ANCA in GPA. Eight SNPs have been identified in the PRTN3 promotor region and exons, but their function for PR3 expression is unclear. An increased frequency of one promotor polymorphism (A546G) was found among patients with GPA(40). PR3 is not only stored in granules and released during neutrophil activation, but it is also expressed on the neutrophil membrane, where it may be engaged by PR3-ANCA. The membrane expression of PR3 varies between individuals, remains constant over time, and appears genetically determined(41, 42). One study found a link between membrane PR3 expression and HLA antigens(43). Individuals with high membrane PR3 expression are significantly more frequent among patients with GPA than in the normal population, and patients with GPA and high membrane PR3 expression are at higher risk for relapses than those with low membrane PR3 expression(44, 45).

The major natural inhibitor for PR3 is α1-antitrypsin, coded for by the SERPINA1 gene(AAT). Of the many polymorphisms of AAT, the Z- and S alleles are responsible for the majority of α1-antitrypsin deficiency cases. An increased frequency of heterozygosity for the Z- and S-alleles in association with GPA, but not MPA, have been reported by several groups in small studies and recently confirmed in larger studies(46, 47). The clinical significance of this association is not entirely clear, but heterozygosity for the Z-allele has been associated with a worse prognosis of the disease(48). This may be because Z-allele carriers, but not S-allele carriers, have increased levels of pro-inflammatory polymers circulating in their blood (47). Deposits of these polymers have been documented in kidney biopsy specimens from patients with active GPA, and these polymers can prime neutrophils and augment the activation of neutrophils by PR3-ANCA in vitro(47).

Unconfirmed or conflicting results have been reported for the association with either GPA or MPA and polymorphisms of other genes including those coding for the high affinity soluble interleukin 2 receptor (IL-2RA), interleukin 10 (IL-10), leukocyte immunoglobulin-like receptor A2 (LILRA2) and CD226 (CD226) as well as for the Fc receptors FcRIIa and FcRIIIb(21).

Another, yet unclear, factor to be considered in the interpretation of genetic SNP association studies is the issue of copy number variation(49). Copy number variations as a result of gene duplication, triplication or exon shuffling may occur at a higher frequency in the genome than SNPs and may be more important for evolution(49). The study of effects of copy number variations on the disease phenotype of complex autoimmune diseases including the AAV syndromes is just beginning. Genome-wide association studies (GWAS) have produced valuable insights into the genetic background of several complex immunologic and non-immunologic disorders. Recently, the first GWAS in AAV (GPA and MPA, no EGPA) was performed in the United Kingdom in a discovery cohort of 1233 AAV patients and 5884 controls; and was replicated in 1454 northern European case patients and 1666 controls(50••). This study confirms previous preliminary data that favor heterogenic genetic backgrounds for GPA and MPA, even though the study was relatively underpowered for MPA cases. The strongest associations found in this study were with the antigenic specificity of ANCA (PR3-ANCA versus MPO-ANCA) rather than with the clinical diagnosis (GPA versus MPA). This GWAS confirmed strong associations with the HLA-DP region, the serpin A1 gene (SERPINA1, which encodes α1-antitrypsin), and with the PRTN3 gene (encodes proteinase 3) in patients with PR3 antibodies, and HLA-DQ in patients with MPO antibodies. Other previously reported associations were not confirmed in this study.

In summary, the predisposing genetic factors in AAV are most likely heterogeneous and differ between patients. Genetic factors may result in alterations in HLA mediated antigen presentation, global T- and B-lymphocyte activation, defective immune regulation and abnormal target antigen structure and/or function. Whether any of these predisposing genetic factors by itself or in combination is sufficient to cause the disease remains to be determined.

Environmental Exposures and Drugs as Triggers

Environmental triggers for the onset of AAV in susceptible individuals remain unknown for most patients. A significant association between MPA with MPO-ANCA and exposure to silica has been reported(51). Silica is a potential activator of the inflammasome complex that generates, among other mediators, interleukin 1 (IL-1)(52).

The onset of AAV (predominantly MPA) has also been reported following exposure to a variety of therapeutic agents including propylthiouracil, hydralazine and penicillamine (53, 54). These agents have polyclonal B lymphocyte stimulatory properties, which may induce the production of ANCA. However, there are several features distinguishing drug-induced AAV from typical AAV. First, the ANCA response is often targeting several different antigens at the same time, whereas in typical AAV; ANCA target either PR3 or MPO, but not both, and not other additional antigens such as human neutrophil elastase or lactoferrin. Second, drug-induced AAV usually subsides after discontinuation of the offending agent. Thus, the cases of drug-induced AAV support the hypothesis of the pathogenic role of ANCA in the development of vasculitis, but they do not explain the loss of tolerance to ANCA target antigens characterizing GPA and MPA.

Infectious Triggers

Infectious pathogens such as mycobacterial and fungal organisms represent the most frequent cause of granulomatous inflammation of the upper respiratory tract. Therefore the presence of granulomatous inflammation in patients with GPA resulted in an initial extensive search for a causative microorganism (55–57). So far almost all attempts to culture, stain or treat for infectious pathogens failed. There are however certain indicators that imply a role of microorganisms in GPA. Patients with GPA commonly report symptoms of upper respiratory tract infections immediately preceding their initial presentation or a disease relapse (58–61).

Infections have also been implicated as triggers as well as persistent drivers of various autoimmune diseases including AAV. A variety of different and often interrelated mechanisms by which infections trigger and perpetuate the disease in predisposed patients have been proposed. Many infectious agents have been reported to induce an ANCA response. In most instances these ANCA are directed against target antigens other than MPO or PR3. Furthermore, these infection-associated ANCA usually disappear once the infection resolves (2, 62, 63). Thus, it seems that two conditions need to be met for infections to trigger AAV. First, tolerance to self-antigens needs to be broken, i.e. host conditions permissive for the development of autoimmunity need to allow for the ongoing production of antibodies directed against self antigens (ANCA). Second, the persistent specific autoantibodies (ANCA) need to have pathogenic potential for the development of tissue injury characteristic of AAV.

One concept by which infections can elicit an autoimmune response in susceptible hosts is molecular mimicry(64). Subsequent diversification of T and B lymphocyte responses (“epitope spreading”) may lead to reactivities with different epitopes on the same target molecule (intramolecular spreading) or even extend to other molecules (intermolecular spreading)(65, 66). The only example of direct epitope mimicry leading to ANCA in patients with pauci-immune glomerulonephritis has been reported by Kain and co-workers(67). They found ANCA targeting lysosomal membrane protein-2 (LAMP-2) in the majority of patients with pauci-immune focal necrotizing glomerulonephritis (67). These ANCA recognized an epitope on LAMP-2 and cross-reacted with the homologous bacterial adhesin FimH (67). Antibodies to LAMP-2 transferred to rats caused pauci-immune glomerulonephritis in the recipients (67). Rats immunized with FimH developed antibodies to FimH that cross-reacted with human LAMP-2 and also developed pauci-immune glomerulonephritis (67). The frequency of LAMP-2 specific ANCA has recently been confirmed in different patient cohorts from Europe(68), but not from the United States(69). It also appears that the LAMP-2 specific ANCA disappear quickly after initiation of immunosuppressive therapy, and methods for their detection are not as robust as those used for the detection of PR3-ANCA and MPO-ANCA(68). Consequently, the clinical relevance of this finding remains controversial.

An indirect mechanism of molecular mimicry leading to typical PR3-ANCA has been proposed by Pendergraft and colleagues(70). They demonstrated that in selected patients ANCA may represent anti-idiotype antibodies(71). Antibodies are formed against complementary peptides (antisense peptide sequence) to PR3 (cPR3)(70). The cPR3 peptides, which may represent mimics of microbial peptide sequences, are the target of the primary immune response. Indeed, several bacterial peptide sequences including Staphylococcus aureus (S. aureus) sequences were found to have homologies with cPR3(70). True PR3-ANCA are subsequently the result of a secondary immune response mounted against the idiotype of these anti-cPR3 antibodies(70). Pendergraft et al. were able to show that mice immunized with cPR3 developed antibodies against both cPR3 and PR3, and they found antibodies against cPR3 in 7 of 34 PR3-ANCA positive patients(70). However, the same group was unable to demonstrate a similar scenario for MPO-ANCA(72). Furthermore, Tadema et al. were unable to reproduce the increased frequency of anti-cPR3 antibodies in patients with GPA compared to healthy volunteers or MPO-ANCA positive patients(73). Interestingly, portions of the cPR3 sequence have homology with portions of the plasminogen sequence, and anti-plasminogen antibodies were detected in some patients with AAV, potentially contributing, among other potential mechanisms, to the well recognized increased risk for thromboembolic events(74, 75).

Whereas direct or indirect epitope mimicry may account for the development of an antibody response, these phenomena do not fully explain the sustained loss of self tolerance observed in the majority of these patients. A variety of mechanisms by which infections promote the autoimmune response as well as the propensity for chronic relapses in AAV have recently been studied. Nearly two thirds of patients presenting with GPA are nasal carriers of S. aureus(76). This is much higher than in the general population. S. aureus carriers are at higher risk for relapse of GPA than non-carriers, and treatment with trimethoprim-sulfamethoxazole resulted in a significant reduction of the relapse rate (76–78). Multiple mechanisms may contribute to the increased risk of relapse conveyed by S. aureus. S. aureus produces superantigens known to be powerful non-specific (antigen-independent) T- and B lymphocyte activators able to induce significant cell proliferation and cytokine release(79, 80). Patients colonized with superantigen-producing strains of S. aureus are at higher risk for disease flares than those colonized with superantigen-negative strains (81). Compared to healthy controls, patients with GPA were found to have an expansion of T lymphocytes expressing Vbeta segments specific for S. aureus superantigen(82). However, a direct association between the presence of S. aureus producing superantigens and the expansion of T lymphocytes reactive to these superantigens in individual patients could not be confirmed(82).

S. aureus derived superantigens and peptidoglycans as well as fungal beta-glucans can induce the expansion of IL-17 producing CD4 positive T cells, so called Th17 cells, in an IL-23 dependent manner(83–85). Th17 cells, which are now recognized as central players in the development of autoimmunity. They are highly potent inflammatory cells that initiate and maintain tissue inflammation by recruiting other inflammatory cells while generating a milieu that makes them resistant to control by T-regulatory (Treg) cells(86, 87). Persistent IL-23 production by antigen-presenting cells appears necessary to maintain Th17 cells at the site of inflammation(87). Both elevated IL-23 and IL-17 levels have been found in patients with active disease and remained elevated despite treatment during clinical remission(88•). An increased frequency of Th17 cells responding to staphylococcal enterotoxin B was found in GPA patients in remission compared to healthy controls, regardless of ANCA status, whereas an increased frequency of PR3-responsive Th17 cells was restricted to PR3-ANCA positive patients(89). Chronic infections may thus set the stage for chronic ongoing inflammation and the loss of self tolerance.

Necrotizing granulomatous inflammation consisting of monocytes, macrophages, neutrophils, T cells, B cells and plasma cells, predominantly located in the respiratory tract, sets GPA clinically apart from MPA(1). Structures resembling germinal centers have been documented within the granulomatous inflammatory tissue of GPA, and the tissue also stained positive for PR3(90). The immunoglobulin (VH) gene mutational patterns obtained from granulomatous nasal tissue of patients with GPA suggest that the selection and maturation of PR3-ANCA producing B lymphocytes may start within the granulomatous lesions(90).

Cytosine-phosphate-guanine (CPG) motifs are pathogen-associated molecular patterns (PAMPs) that are recognized by the pattern recognition receptor Toll-like receptor 9 (TLR9). Unmethylated CPG oligodeoxynucleotides (CpG-ODN) are potent immune-stimulants. B lymphocytes isolated from patients with AAV with active disease as well as during remission can be induced to produce ANCA when exposed to CpG-ODN and IL-2 (91, 92•). Significantly, more PR3-ANCA patients than MPO-ANCA patients produced ANCA in vitro in response to CpG exposure (92•). These observations provide another link between S. aureus infection, granulomatous inflammation of the respiratory tract and the higher relapse rate of patients with PR3-ANCA or GPA compared to patients with MPO-ANCA or MPA.

Neutrophils are key players of the innate immune defense against microorganisms. An additional antimicrobial defense mechanism of neutrophils has been described recently – a unique type of cell death distinct from apoptosis and necrosis that is associated with the formation and release of neutrophil extracellular traps (NETs)(93). NETs are extracellular structures containing chromatin and granule proteins (including ANCA target antigens) in which invading microbes are trapped and killed. The formation of NETs depends on reactive oxygen species generated by NADPH oxidase (94). ANCA, which are known to induce a respiratory burst in primed neutrophils can also induce the formation of NETs, and ANCA antigens bound to NETs are accessible to ANCA(95••). NETs containing PR3 and MPO are detectable in the kidneys of patients with AAV in the absence of infection. Interestingly, S. aureus can rapidly and strongly induce NETs formation even without causing neutrophil death (96••). Chromatin-immunoglobulin complexes are thought to lead to the loss of tolerance and autoantibody (ANCA) production in a TLR9-dependent manner (97). Taken together, these observations suggest that ANCA may induce a vicious cycle of self-perpetuated NETs formation and more ANCA production, particularly in the presence of S. aureus infection. This hypothesis is further supported by observations of increased TLR9 expression by monocytes of patients with GPA who were S. aureus nasal carriers (98•).

To summarize, current evidence suggests that infections and/or colonization with infectious organisms such as S. aureus are contributing to the pathogenesis of AAV, especially GPA. Pathogens may provide the initial antigen sources (molecular mimicry), pro-inflammatory signals (e.g. stimulation of pathogen-associated molecular pattern receptors by CpG, peptidoglycans and fungal beta-glucans) and triggers for innate (NET formation) and adaptive (superantigens) immune cells activation. In genetically susceptible individuals and/or in the context of other environmental exposures these factors may trigger a break in immune tolerance and the development of sustained or recurrent autoimmunity.

Role of B and T Lymphocytes

The role of B lymphocytes has gained increased interest after efficacy in AAV of the B-cell depleting agent, rituximab, was demonstrated (99). B lymphocytes have been found in affected tissues including kidneys and most significantly in granulomatous lesions of the respiratory tract, where they are located in close proximity to numerous PR3-positive cells and where the selection and maturation into PR3-ANCA producing B lymphocytes may occur(90). Antigen-specific B lymphocytes are the progenitor cells of short-lived plasma cells thought to be the source of autoantibodies including ANCA (100, 101). The proportion of circulating activated B lymphocytes is increased in patients with GPA compared to healthy controls, and it is higher in patients with active disease versus those in remission and in patients with generalized disease versus those with limited disease(102). Moreover, elevated serum levels of B lymphocyte stimulator (BLyS, also known as B cell activating factor, BAFF), a cytokine belonging to the TNF superfamily and known to promote B cell survival, differentiation and proliferation, were found to be increased in patients with active AAV(103,104•,105).

T cells are usually found within granulomas as well as other lesions present in AAV. Circulating T cells in AAV have shown different abnormalities with disturbed homeostasis that is skewed towards memory and pro-inflammatory T-cell types. ANCA are high-affinity class-switched antibodies(106). This implies that ANCA production is dependent on T lymphocyte help, that autoreactive T lymphocytes are present, and that there is insufficient counter-regulation by regulatory T cells (Tregs). Indeed, T lymphocyte abnormalities have long been suspected to be the main reason for the chronic relapsing nature of GPA(102). Markers of T cell activity such as soluble IL-2 receptor, neopterin and soluble CD30 have been shown to be associated with disease activity(107, 108). Patients with GPA in remission have an increased percentage of circulating CD4 positive effector memory T lymphocytes as well as a functional defect of circulating CD4 positive CD25 positive Tregs (109–111••). Moreover, patients with reduced numbers of Tregs required more prolonged treatment to achieve remission, and their relapse rates were higher (111••).

As previously discussed, another T cell subset that has drawn attention recently in autoimmunity is the IL-17-producing T cell subset (Th17)(112). The number of Th17 cells reactive to the autoantigen PR3 was found to be increased in patients with GPA in remission (88, 113). Furthermore, Ordonez et al. found that AAV patients exhibit an expanded CD45RC positive T helper cell population that is a source of IL-17 (114). IL-17 acts on endothelial cells, epithelial cells and antigen presenting cells causing release of chemokines that will activate neutrophils, such as IL-8 and CXCL-1(115). IL-17 can also induce IL1-beta and TNF-alpha production and release by macrophages, and these cytokines, in turn, prime neutrophils and monocytes resulting in the expression of ANCA-target antigens on their surface(116).

Pathogenic Role of ANCA

Clinical observations and a large body of experimental data support a pathogenic role of ANCA for the development of small vessel vasculitis. The small vessel injury of AAV appears to be caused by activated leukocytes. The pro-inflammatory pathogenic effects of ANCA are all contingent on their interactions with their target antigens expressed on the surface of primed neutrophils and monocytes. When primed with inflammatory cytokines such as tumor necrosis factor (TNF)-α or microbial products in vitro, leukocytes express proteinase 3 (PR3) and myeloperoxidase (MPO) on their surface (117, 118). PR3-ANCA and MPO-ANCA can activate primed neutrophils and monocytes by binding directly to their antigens expressed on the surface or by Fc-receptor engagement; these interactions initiate signal transduction cascades via multiple pathways that are similar in neutrophils and monocytes (119–123).

Several pro-inflammatory effects derived from the activation of neutrophils and monocytes by ANCA are responsible for the tissue injury seen in AAV. Fully activated neutrophils degranulate and release toxic proteases and enzymes including elastase, PR3, MPO and others(117). ANCA also induce a respiratory burst resulting in the release of oxygen radical species (117, 124). In addition, ANCA induce the expression of cell adhesion molecules on neutrophils and endothelial cells leading to an increased adhesion of neutrophils to endothelial cells (125–129). Moreover, the binding of ANCA to primed leukocytes induces the production and release of chemotactic cytokines including IL-1, MCP-1 and IL-8(130–133). These cytokines attract more neutrophils and monocytes to the site of inflammation. Thus, when the ANCA-induced cytokine release occurs at the endothelial interface, the normal chemotactic gradient that draws neutrophils out of the vasculature into the tissues is lost. This causes further accumulation of fully activated neutrophils in the vessel wall, where they cause more injury. Figure 1 summarizes several complex interactions between ANCA and inflammatory and endothelial cells thought to be instrumental during the formation of granulomatous inflammation and vasculitis.

Figure 1.

A. Necrotizing granulomatous inflammation develops in response to an unknown insult. This inflammatory reaction as well as the original insult result in the release of various cytokines, proteinase 3 (PR3) and potentially bacterial superantigens. Granulomatous inflammation in conjunction with the original insult, secreted cytokines and possibly bacterial superantigen promote the production of ANCA by PR3 specific B-lymphocytes and the selection of PR3 specific T-cells in predisposed individuals. There is a dysbalance between effector T cells (Teffs) and regulatory T cells (Tregs) with further release of proinflammatory cytokines promoting neutrophil priming.

B. ANCA have been shown in vitro to bind to PR3 expressed on the surface of neutrophils and monocytes. ANCA binding results in cellular activation, release of pro-inflammatory cytokines and neutrophil degranulation.

C. In vitro ANCA also opsonize apoptotic neutrophils by binding to PR3 expressed on their cell surface. In addition in the presence of ANCA apoptotic neutrophils exhibit decreased phosphatidylserine expression on their cell surface. Neutrophil opsonization and perturbed phosphatidylserine expression result in the inappropriate release of pro-inflammatory cytokines by macrophages clearing apoptotic neutrophils.

D. Endothelial damage is promoted by various mediators released by neutrophils. In addition the increased expression of cell adhesion molecules and presence of ANCA facilitates the binding of neutrophils to endothelial cells. PR3 is internalized by endothelial cells and promotes apoptosis.

E. Neutrophil extracellular trap (NET) formation occurs in lesions as a consequence of neutrophil apoptosis and degranulation. NET-derived products activate dendritic cells and B-cells by sensing via Toll-like receptors. Interferon (IFN-a) production by dendritic might have an impact on local immune regulation and it has been shown to impair regulatory T cells function.

F. Endothelial damage and tissue inflammation caused by localized complement activation. Complement is activated predominantly by the alternative pathway resulting in cleavage of C5 into C5b which causes the assembly of MAC. C5a is able to prime neutrophils to enhance ANCA-induced neutrophil activation.

The ANCA target antigens released from activated or dying neutrophils can also directly bind to endothelial cells(134). This may result in apoptosis of endothelial cells and in localized immune complex formation with circulating ANCA (135). Low levels of localized immune-complex deposition, which has been documented in early vasculitic skin lesions as well as renal lesions, can in turn induce localized complement activation(136, 137).

In vitro studies have also suggested that ANCA modify the clearance of apoptotic cells. Opsonization of pre-apoptotic cells by ANCA is associated with an increased production of inflammatory cytokines by phagocytosing macrophages(138). Moreover, pre-apoptotic cells have a decreased cell surface expression of phosphatidylserine (the recognition signal for macrophages) in the presence of ANCA(139). Consequently, in the presence of ANCA the non-inflammatory clearance of apoptotic cells by macrophages may be perturbed in favor of inflammation and necrosis.

The pathogenic role of ANCA has also been supported by several in vivo animal models. These models are based on the transfer of antibodies generated against ANCA target antigens into healthy recipient animals. The transfer of anti-MPO IgG or splenocytes obtained from MPO-knockout mice, which were immunized with murine MPO, into Rag 2 knock-out mice (lacking mature T- and B-lymphocytes) as well as into wild-type mice resulted in pauci-immune crescentic necrotizing glomerulonephritis similar to that found in humans(140). A direct augmenting effect of MPO-ANCA on neutrophil - endothelial interactions causing microvascular injury was documented in a rat anti-MPO antibody transfer model (141). Murine anti-PR3 antibodies generated in a similar fashion only caused an increased inflammatory response at the site of tissue injury, but not a vasculitic phenotype, when transferred into wild-type mice(142).

The inflammatory lesions of AAV are referred to as “pauci-immune”, implying that only few immune complexes or complement factors can be identified by immunofluorescence microscopy. Furthermore, in contrast to classic immune-complex mediated disease, patients with active AAV have normal serum complement levels. Nevertheless, low grade localized immune-complex formation and complement activation may play a role, and there is growing evidence suggesting that the complement pathway is involved in the pathogenesis of AAV (137, 143–145). Moreover, complement activation might also contribute to the increased risk of venous thromboembolism observed in active AAV, as activated complement factors trigger the coagulation cascade(146, 147).

It has recently been recognized that activation of the alternative complement pathway by ANCA may represent an important amplification loop of inflammation that contributes to renal (and other tissue) injury in AAV. In the murine anti-MPO antibody transfer model the development of necrotizing glomerulonephritis is dependent on activation of the alternative complement pathway, and the development of lesions can be prevented and treated with an antibody that inhibited complement factor 5 (C5) activation(148, 149). Mice lacking the receptor for activated C5 on neutrophils also do not develop the renal lesions (150). In vitro studies showed that supernatants from ANCA-activated neutrophils can cause the production of C5a in normal serum, C5a receptor dependent priming of normal neutrophils, and the increased neutrophil membrane expression of PR3. Even though the renal lesions in humans are called “pauci-immune”, components of the alternative complement pathway can be detected in patients with AAV, but not normal controls or patients with minimal change disease(151).

Despite all this evidence, proof that ANCA alone can cause disease in humans has remained elusive. One case study in which an infant was born to a mother with active MPA is often quoted as such evidence (152). The infant developed a pulmonary-renal syndrome 48 hours after delivery and was found to have serum MPO-ANCA titers similar to the mother (152). The child was treated with glucocorticoids and plasma exchange and recovered. However, this observation is countered by another report of a case where MPO-ANCA were also transferred from the mother to the newborn, but the newborn remained perfectly healthy despite persistence of the transferred MPO-ANCA in the newborn for several weeks(153). The two contrasting case studies are consistent with observations made in large cohort studies. The development of severe vasculitic disease manifestations and severe flares usually do not occur in the absence of ANCA, but not all patients with persistent ANCA inevitably suffer such flares (76, 154–156).

Conclusions

Based on currently available clinical and experimental evidence it is reasonable to conceptualize that in predisposed patients, different triggers can lead to the production of autoantibodies (ANCA) that in the context of an inflammatory environment can cause tissue inflammation and vascular injury. However, many of the proposed mechanisms behind the pathophysiology of AAV may only apply to certain clinical subsets of patients. Moreover, despite the several different pathways and mechanisms described in this review, there is not a one-size-fits all cases, or even most of the cases. Despite the substantial advances in our understanding of the pathogenesis of AAV, many open questions still remain. As our knowledge regarding the pathogenesis of AAV evolves, new therapeutic targeted strategies are emerging for the continued quest to control disease activity with the minimum of adverse effects in individual patients.

Abbreviations

NG

necrotizing granuloma

H

histiocyte

PMN

polymorphonuclear neutrophil

T

T-lymphocyte

B

B-lymphocyte

DC

dendritic cell

GC

giant cell

P

plasma cell

Mono

monocyte

PR3

proteinase 3

LPS

lipopolysaccaride

ROS

reactive oxygen species

ANCA

anti-neutrophil cytoplasmic antibodies

HLE

human leukocyte elastase

NO

nitric oxide

MPO

myeloperoxidase

IL-8

interleukin 8

MCP-1

monocyte chemoattractant protein-1

TNFα

tumor necrosis factor α

IL-1β

interleukin 1β

IL-1

interleukin 1

ICAM

intercellular adhesion molecule

VCAM

vascular cell adhesion molecule

PGE2

prostaglandin E2

TxB2

thromboxane B2

MAC

Membrane attack complex

TLRs

Toll-like receptors