Pathogenesis of ANCA-Associated Vasculitis, an Update
- First Online:
- Cite this article as:
- Kallenberg, C.G.M. Clinic Rev Allerg Immunol (2011) 41: 224. doi:10.1007/s12016-011-8258-y
- 698 Views
Clinical observations, including a report of neonatal vasculitis occurring in a child born from a mother with anti-neutrophil cytoplasmic antibody directed to myeloperoxidase (MPO-ANCA)-associated vasculitis, suggest a pathogenic role for ANCA. Such a role is supported by in vitro experimental data showing that ANCA can activate primed neutrophils to the production of reactive oxygen species and lytic enzymes resulting in lysis of endothelial cells. An interplay between neutrophils, the alternative pathway of complement, and MPO-ANCA resulting in systemic vasculitis including necrotizing glomerulonephritis has clearly been demonstrated in animal models. An in vivo pathogenic role of ANCA directed to proteinase 3 (PR3-ANCA) has, however, not been substantiated. In PR3-ANCA-associated vasculitis, granulomatous inflammation points to involvement of cell-mediated immunity. In vitro studies, indeed, suggest that PR3-specific Th17 CD4-positive lymphocytes are operative in lesion development. The triggering role of microbial factors is becoming more clear. In particular Staphylococcus aureus carriage and infection with Gram-negative bacteria could contribute to induction and persistence of ANCA-associated vasculitis (AAV). Insight into the pathogenic pathways involved in AAV have opened and will further open new ways to targeted treatment.
KeywordsANCA-associated vasculitisPathogenesisPR3-ANCAMPO-ANCAStaphylococcus aureus
Disease associations of PR3-ANCA and MPO-ANCA
Idiopathic crescentic glomerulonephritis
Clinical Arguments for a Pathogenic Role of ANCA
The initial description of ANCA in WG already mentioned that titers of ANCA are lower or disappear during inactive disease compared to remission . This was followed by a small-sized prospective study in which titers of ANCA as detected by indirect immunofluorescence (IIF) were measured every month and related to the occurrence of relapse . This study showed that all relapses (n=17) were preceded by a significant rise in ANCA titer with a mean period of 50 days (range, 9–106 days) between rise in titer and clinical relapse.
Furthermore, treatment solely based on increasing titers of ANCA was able to prevent relapses of WG . A larger prospective study in which changes in ANCA, measured both by titration in the IIF-test and by PR3-specific ELISA, were related to possibly ensuing relapses was performed by Boomsma et al. . Eighty-six patients with PR3-ANCA WG were followed for 24 months with sampling of sera at least every 2 months. A rise of ANCA was defined as a 75% increase by ELISA or a fourfold increase in titer by IIF. Twenty-six out of the 33 relapses occurring during this period of observation were preceded by a rise as measured by ELISA (sensitivity of 79%) and 17 out of 33 relapses by a rise in IIF (sensitivity of 52%). Twelve out of 38 rises by ELISA were not followed by a relapse (specificity of 69%) and 13 rises by IIF were not related to relapse (specificity of 75%). Most rises were followed by a relapse within 6 months, but, infrequently, the relapse occurred later than 6 months. These data suggest a relationship between rises of ANCA and relapsing disease, but this relationship is far from robust. Data from the Wegener’s Granulomatosis Etanercept study, in which 180 patients with WG were followed, did not confirm an association between levels of ANCA and disease activity in WG . Although levels were lower during remission compared to active disease, decreases in levels of ANCA were not associated with a shorter time to remission and increases were not associated with ensuing relapses.
Other data, however, do point to a role of ANCA in the induction of disease activity. Stegeman et al.  showed that persistence of ANCA in WG, intermittently or continuously, after induction of remission is a strong risk factor for relapse with a relative risk of 9.0. More recently, Sanders et al.  observed that a level of PR3-ANCA >10 U/ml at 24 months after starting treatment was predictive of relapse within 5 years (RR 4.6, CI 1.2–6.3). The introduction of rituximab, a B-lymphocyte depleting monoclonal antibody, in the treatment of AAV has corroborated the assumption that ANCAs are involved in the pathogenesis of AAV. Two large clinical trials have shown that rituximab is as effective as cyclophosphamide for induction of remission in AAV [10, 11]. Even more interesting are the data on treatment with rituximab in 108 patients with refractory PR3-ANCA-associated AAV from the Mayo Clinic . They observed in this open series that all 108 patients treated with rituximab in combination with corticosteroids came into remission. All relapses that were noticed occurred after reconstitution of B cells, and all but one were preceded by a rise in ANCA. Relapses were treated with an additional course of rituximab with corticosteroids which resulted in remission in all cases. Pre-emptive treatment with rituximab based on B cell reconstitution accompanied by re-appearance or increase in titer of ANCA was able to prevent relapses.
Finally, a strong argument for a pathogenic role of MPO-ANCA comes from the observation of glomerulonephritis and pulmonary hemorrhage in a neonate born from a mother with MPO-ANCA-associated AAV. Transient occurrence of vasculitis in the neonate was thought to be caused by transplacental transfer of MPO-ANCA .
In conclusion, data from clinical observations do not prove that ANCAs are pathogenic in the AAV but certainly suggest that B lymphocytes, including those producing ANCA, play a major role in disease expression.
Are ANCAs Pathogenic? Data from In Vitro Studies
It has been noticed for a long time that ANCAs are able to activate neutrophils and monocytes once the cells are primed with low doses of cytokines such as tumor necrosis factor alpha (TNF-α), interleukin-1, and interleukin-18. Priming results, among others, in the surface expression of the targets of ANCA, that is PR3 and MPO, on the membranes of these cells allowing interaction with ANCA [14, 15].
Once neutrophils are primed, interaction with ANCA results in their production of reactive oxygen species and the release of lytic enzymes such as elastase [14, 15]. Not only binding of ANCA to their target antigens on the membrane of neutrophils is crucial but also interaction of the Fc part of the autoantibodies with Fcγ receptors on these cells. In particular, the Fcγ receptors IIa and IIIb are involved [16, 17]. Furthermore, ANCA-induced neutrophil activation occurs in vitro not in solution but on a surface only . This makes sense in the in vivo situation as full neutrophil activation in the circulation, as occurs in, e.g., septic shock, would be a life-threatening condition. Further studies using experimental flow systems have shown that ANCAs are able to convert rolling neutrophils into neutrophils that firmly adhere to the endothelium . This process can be blocked by antibodies against the Fcγ receptor IIa, again demonstrating the interaction of the Fc part of the autoantibodies with Fcγ receptors on the neutrophil. In addition, it can be inhibited by antibodies to CD11b, showing that integrins are involved in this process as well. Various studies have demonstrated that these adherent neutrophils are now further stimulated by ANCA and release reactive oxygen species and lytic enzymes, resulting in detachment and apoptosis of endothelial cells . More recently, the signal transduction pathways involved in ANCA-induced neutrophil activation have been elucidated, with, possibly, selective activation of diacylglycerol kinase .
The data from in vitro studies clearly demonstrate the potential of ANCA to induce necrotizing vasculitis and glomerulonephritis. As will be discussed in the next section, studies in experimental animals strongly support the pathogenetic concept derived from these in vitro studies.
Are ANCAs Pathogenic? Data from In Vivo Experimental Studies
Initial studies in Brown Norway (BN) rats were directed to induce an autoimmune response to MPO by immunizing these rats with human MPO in complete Freund’s adjuvant (CFA). Rats developed anti-MPO antibodies cross-reacting with rat MPO, but no lesions were detected in target organs such as kidneys and lungs. However, ex vivo perfusion of their kidneys with the products of activated neutrophils, that is MPO and H2O2, resulted in the development of pauci-immune necrotizing crescentic glomerulonephritis (NCGN) . Furthermore, induction of an autoimmune response to MPO in BN rats was shown to aggravate subclinical anti-glomerular basement membrane disease in these rats into full-blown glomerulonephritis . This demonstrated the phlogistic potential of anti-MPO antibodies in these rats.
A less artificial model of MPO-ANCA-associated AAV was developed by Xiao et al. . They immunized MPO-deficient mice with mouse MPO. As a result, these mice developed an immune response to mouse MPO. Next, splenocytes from these mice were transferred into immunodeficient and normal mice. These recipient mice developed pauci-immune NCGN and systemic necrotizing small-vessel vasculitis including hemorrhagic pulmonary capillaritis. Transfer of IgG alone from these MPO-deficient mice immunized with MPO into normal mice resulted in pauci-immune focal NCGN. Injection of lipopolysaccharide (LPS) in this latter model was able to augment this focal NCGN, and less severe lesions developed when anti-TNF-α antibodies were given . Further studies, using this model, showed that neutrophils  as well as the presence of MPO  are requirements for the induction of lesions. Following the in vitro observations, the role of the complement system has been further elaborated in this model. First, the model was developed in mice deficient for various components of the complement system. It was shown that mice lacking complement factor B and mice lacking factor C5 did not develop lesions, whereas mice deficient in complement C4, a crucial factor of the classical pathway of complement activation, developed pauci-immune NCGN identical to that in normal mice . This clearly demonstrates the role of the alternative pathway of complement activation in the development of MPO-ANCA-associated NCGN. Secondly, pre-treatment of mice with anti-C5 monoclonal antibodies was able to prevent disease development, whereas a strong reduction of lesions occurred when anti-C5 antibodies were administered once anti-MPO NCGN had already been induced . Finally, van Timmeren et al.  treated the IgG fraction containing anti-MPO antibodies from MPO-knockout mice immunized with mouse MPO with endoglycosidase which deglycosylated this IgG fraction resulting in non-reactivity with Fcγ receptors. Injection of deglycosylated IgG into recipient mice could largely prevent the development of NCGN.
Little et al.  induced anti-MPO necrotizing vasculitis in Wistar Kyoto rats by immunizing rats with human MPO in CFA. In this model, they showed, by intravital microscopy, that the antibodies are able in vivo to convert rolling neutrophils into firmly adhered neutrophils that invade into the vascular wall resulting in necrotizing vasculitis.
Taken together, the data from studies in experimental animals strongly suggest, if not prove, that MPO-ANCAs are pathogenic in vivo. The interplay of anti-MPO antibodies, neutrophils, and the alternative pathway of the complement system results in pauci-immune necrotizing vasculitis and glomerulonephritis. The elucidation of the various pathways involved enables testing of targeted treatment.
In contrast to MPO-ANCA-associated MPA, PR3-ANCA-associated WG is, in addition to necrotizing vasculitis, characterized by granulomatous inflammation. So, any model of PR3-ANCA-associated AAV should demonstrate granulomatous inflammation. In analogy to the model of anti-MPO-induced vasculitis, Pfister et al.  immunized PR3-deficient mice with recombinant murine PR3. The immunized mice developed antibodies to mouse PR3 which were able to recognize PR3 on the surface of mouse neutrophils. However, passive transfer of the IgG fraction of immunized mice containing anti-mouse PR3 antibodies into naive mice did not result in the development of vasculitic lesions in kidneys and lungs. The only observation they made was an increased local inflammation in the skin after local TNF-α injection in anti-PR3-transferred mice compared to control-transferred mice. Also, van der Geld et al.  induced antibodies to rat-PR3 in rats by immunization of rats with human-mouse chimeric PR3. Again, no lesions developed following immunization, even in rats systemically injected with LPS. These data suggest that in contrast to MPO-ANCA AAV, PR3-ANCA AAV is pathogenetically a more complex disease in which cell-mediated immunity plays an additional role as reflected by the occurrence of granulomatous inflammation.
Granulomatous Inflammation is Associated with PR3-ANCA
Irrespective of the underlying diagnosis, PR3-ANCA AAV is strongly associated with granuloma formation while MPO-ANCA AAV is not . Furthermore, PR3-ANCA AAV shows far more relapses than MPO-ANCA disease which should be explained. Granulomatous inflammation suggests cell-mediated immunity as an effector mechanism. Indeed, percentage and number of effector memory T cells, characterized by the CD45RO phenotype and the absence of CCR7, are increased in PR3-ANCA patients during remission together with an increased state of activation compared to controls . Surprisingly, percentages of effector memory T cells decreased during active disease. It was hypothesized that these cells migrate into lesional tissue during active disease. Indeed, effector memory cells could be detected in the urine of PR3-ANCA AAV patients with active renal involvement . Assessment of urinary effector memory T cells might be a very sensitive indicator of renal activity in AAV as has also been suggested in lupus nephritis . Next, we investigated the characteristics of autoimmune T cells in patients with PR3-ANCA AAV by stimulating peripheral blood mononuclear cells with PR3 and analyzing the proliferating CD4-positive T cells for their intracellular cytokine pattern. Interferon gamma (IFN-γ) was used as a marker for so-called Th1 cells, the effector cells of cell-mediated immunity towards intracellular microorganisms; interleukin-4 as a marker for Th2 cells, helper T cells involved in humoral immune responses; and interleukin-17 (IL-17) as marker for Th17 cells . Th17 cells, generated in a cytokine milieu with transforming growth factor β, interleukin-6, and particularly interleukin-23 (IL-23), are major effector cells in autoimmunity . They produce IL-17 which has an effect on various cells including endothelial cells, epithelial cells, and monocytes/macrophages and induce the latter to the production of pro-inflammatory cytokines and chemokines including CXCL-1, a strong chemoattractant for neutrophils . Increased levels of IL-17 have been detected in various autoimmune diseases, including AAV . Interestingly, PR3-stimulated proliferating CD4 T cells from PR3-ANCA AAV patients predominantly produced IL-17 and not IFN-γ . Thus, Th17 cells appear the major effector cells in PR3-ANCA disease in which granulomatous inflammation is a major characteristic.
Autoimmune diseases are characterized by a dysbalance in the immune system in which immune tolerance has been broken. Peripheral tolerance is, at least for a major part, maintained by regulatory T cells, characterized by high membrane expression of CD25 and the intracellular presence of the transcription factor FoxP3. Studying the presence of these regulatory T cells in WG, Abdulahad et al. , surprisingly, observed that percentages of these cells were increased in WG patients. However, these cells were functionally deficient in downregulating responsiveness of CD25-low- and CD25-negative CD4 T cells to mitogens and antigens. Recent data suggest plasticity of the CD4-positive T cell populations resulting in shifts from phenotypically regulatory T cells towards effector T cells under conditions created by (dys)balance in cytokines .
Are Microbial Factors Involved in the Pathogenesis of AAV?
Stegeman et al. , later on confirmed by others, observed that 60–70% of patients with WG are chronic carriers of Staphylococcus aureus in their nasal cavity. Whether this is a primary or secondary phenomenon, the latter being related to mucosal damage due to the disease itself, is not clear. Interestingly, chronic nasal carriage of S. aureus was associated with a strongly increased risk for relapsing disease (relative risk of 7.2). In addition, maintenance treatment with co-trimoxazole, in particular in patients with WG, could reduce the occurrence of relapses by 60% . Several hypotheses have been formulated to explain the role of S. aureus in the induction of relapses of WG. The first hypothesis suggests that S. aureus, probably by superantigens contained within the bacterium, stimulates the adaptive immune response in a non-antigen specific way. Indeed, Popa et al.  confirmed that 71% of WG patients were chronic nasal carriers of S. aureus, and that, within this population, 71% carried an S. aureus strain expressing at least one superantigen. They also observed that the presence of S. aureus containing the toxic-shock-toxin-1 superantigen was associated with a 14.5-fold increased risk for relapsing disease . The possible role of S. aureus superantigens in activating the adaptive autoimmune response was also suggested by studies of Voswinkel et al. [49, 50]. Single cell analysis of B lymphocytes from nasal biopsies of patients with WG showed that B cell receptors displayed affinity maturation within the granulomatous lesions. Analyzing VH-gene usage of these B cells, they found that VH-gene usage was compatible with S. aureus superantigen stimulation. We recently observed that unmethylated (bacterial) CpG is able to stimulate (autoreactive) B cells of WG patients , also pointing to a role of S. aureus in the pathogenesis of AAV.
Chronic carriage of S. aureus may also activate disease in AAV along other ways. S. aureus carriage may be responsible for low-grade infection in the upper airways with production of low levels of pro-inflammatory cytokines. The latter could locally prime neutrophils and monocytes/macrophages which can then be further activated by ANCA.
More recently, Kain et al.  also have pointed to a role of bacterial products in the induction and persistence of AAV. They observed that 93% (76 out of 84) of patients with pauci-immune necrotizing glomerulonephritis, most of them simultaneously positive for PR3-ANCA or MPO-ANCA, had autoantibodies against human LAMP-2, a heavily glycosylated type 1 membrane protein. This protein is present on the membrane of various cells including neutrophils and endothelial cells. The antibodies produce a cytoplasmic staining pattern on ethanol-fixed neutrophils. Anti-hLAMP-2 autoantibodies were present during active disease but hardly (in 7% of cases) during remission and not in healthy controls and disease controls. In vitro, the antibodies could activate neutrophils and endothelial cells. Anti-hLAMP-2 antibodies, induced in rabbits following immunization, were transferred into recipient rats which developed pauci-immune focal crescentic glomerulonephritis. These data suggest a pathogenic role of anti-hLAMP-2. Most interestingly, an immunodominant epitope of hLAMP-2 showed strong homology with FimH, an adhesive protein of Gram-negative bacteria such as Escherichia coli and Klebsiella pneumoniae. Immunization of rats with FimH induced antibodies to FimH which cross-reacted with hLAMP-2. This resulted in the development of pauci-immune focal necrotizing glomerulonephritis in these rats. These data, which have to be confirmed by others, strongly suggest that infection with Gram-negative bacteria in a susceptible host might induce pathogenic autoantibodies resulting in NCGN.
The role of microbial factors in the induction of autoimmune vasculitis will be further analyzed in a collaborative European–USA study called INTRICATE (infectious triggers of chronic autoimmunity).
Current data suggest that MPO-ANCA could play a pathogenic role in necrotizing small-vessel vasculitis and glomerulonephritis. The pathogenic role of PR3-ANCA is less clear. In PR3-ANCA-associated vasculitis, granulomatous inflammation points to involvement of cell-mediated immunity. Indeed, PR3-specific T cell responses, with a predominance of Th17 cells, seem to be operative in lesion development. The role of microbial factors is becoming more clear. In particular, S. aureus carriage and infection with Gram-negative bacteria could contribute to induction and persistence of ANCA-associated vasculitis. Insights into the pathogenetic pathways involved in AAV have opened and will further open new ways of targeted treatment.