Off balance: T-cells in antineutrophil cytoplasmic antibody (ANCA)-associated vasculitides

Abstract

There is substantial evidence that T-cells are off balance in antineutrophil cytoplasmic antibody (ANCA)-associated vasculitides. Genetic risk factors may influence shaping of the TCR repertoire and regulatory control of T-cells in predisposed individuals. T-cells are found in inflammatory lesions. Vigorous Th1-type responses are seen in Wegener's granulomatosis and microscopic angiitis, whereas a Th2-type response predominates in Churg–Strauss syndrome. Oligoclonality and shortened telomers indicate antigen-driven clonal expansion and replicative senescence of T-cells in ANCA-associated vasculitides. Potent CD28^– Th1-type cells displaying an effector-memory/late differentiated, senescent phenotype are expanded in peripheral blood and are found in granulomatous lesions in Wegener's granulomatosis. Differences in proliferative peripheral blood T-cell responses to the autoantigens proteinase 3 (PR3)- and myeloperoxidase (MPO) have not consistently been detected between patients with ANCA-associated vasculitides and healthy controls in vitro. To recognize an autoantigen, break tolerance, and maintain autoimmune disease T- and B-cells require particular triggers and lymphoid structures. There is preliminary evidence of lymphoid-like structures and possible maturation of autoreactive PR3-ANCA-specific B-cells in granulomatous lesions in Wegener's granulomatosis. Alteration of the T-cell response and anomalous autoantigen-presentation in lymphoid-structures could facilitate development of autoimmune disease in ANCA-associated vasculitides.

Introduction

T-cells play a crucial role in regulating immune responses. Failure to adequately control T-cell activity may result in unwanted sequelae such as chronic inflammation and autoimmune disease, both of which are hallmarks of antineutrophil cytoplasmic autoantibody (ANCA)-associated vasculitides (Fig. 1). Based on their highly specific association with ANCA directed either against proteinase 3 (PR3) or myeloperoxidase (MPO), three diseases of as yet unknown aetiology are coined ANCA-associated: Wegener's granulomatosis (WG), microscopic Polyangiitis (MPA), and Churg–Strauss syndrome (CSS). All three diseases affect predominantly small vessels, i.e. small arteries, arterioles, capillaries, and venules. The vasculitic lesions and glomerulonephritis generally contain no or only few immune-complexes (‘pauci-immune’). PR3-ANCA is highly specific for WG. MPO-ANCA is detected in 60–80% of MPA patients. Both MPO-ANCA and, less frequently PR3-ANCA, are also detected in a number of CSS patients (Table 1) [1,2]. Apart from their diagnostic value and correlation with disease activity ANCA play a direct pathogenic role in inducing systemic vasculitis by interacting with PR3 or MPO on the surface of cytokine primed neutrophil granulocytes as numerous in vitro and several in vivo studies suggest. The interaction of ANCA and neutrophils results in premature neutrophil activation, subsequent endothelial cell damage, and further leucocyte recruitment [3].

Fig. 1.

(a) Necrotizing granulomatous inflammation in WG. (b) Fibrinoid necrotizing small vessel vasculitis with inflammatory cell infiltration in and around the vessel wall in WG. (c) Capillaritis with eosinophil granulocytes in CSS.

Table 1.

CHC-definition, symptoms, and frequency of ANCA in ANCA-associated vasculitides [reviewed in [1–3]].

Disease	CHC definition	Typical symptoms	Frequency of ANCA
Wegener's granulomatosis (WG)	Granulomatous inflammation involving the respiratory tract, and necrotizing vasculitis affecting small to medium-sized vessels (e.g., capillaries, venules, arterioles, and arteries). Necrorizing glomerulonephritis is common.	Pulmonary–renal syndrome	Localized WG: ≤50%Generalized WG: 95% (Strongly associated with PR3-ANCA)
Microscopic polyangiitis (MPA)	Necrotizing vasculitis, with few or no immune deposits, affecting small and medium sized vessels. Necrotizing glomerulonephritis is very common. Pulmonary capillaritis often occurs.	Pulmonary–renal syndrome	40–80% (Usually MPO-ANCA, few cases with PR3-ANCA)
Churg–Strauss syndrome (CSS)	Eosinophil-rich and granulomatous inflammation involving the respiratory tract, and necrotizing vasculitis affecting small to medium-sized vessels, and associated with asthma and eosinophilia.	Pulmonary infiltrates, eosinophilia, polyneuropathy, cardiac involvement	10–70% (Usually MPO-ANCA, few cases with PR3-ANCA)

While these studies support a direct pathogenic role of PR3- and MPO-ANCA in inducing vasculitis in predisposed individuals, the question remains how ANCA are induced. Further, two of the diseases (WG and CSS) are characterized by granuloma formation predominantly affecting the respiratory tract. In contrast, granuloma formation is absent in MPA [1,2]. Animal models of MPO- and PR3-ANCA induced vasculitis have failed to display granulomatous lesions typical of WG and CSS so far [4–6]. The only murine model of PR3-ANCA disease shows aggravation of TNF-α induced panniculitis, which is a rare disease manifestation hardly ever reported in WG [6]. Vasculitic and granulomatous lesions abundantly contain T-cells and macrophages in ANCA-associated diseases [7–10]. T-cells are activated, and there is evidence of an antigen-driven T-cell response in ANCA-associated vasculitides (Table 2) [11–20]. Furthermore, remission has been induced with therapeutics directed against T-cells in refractory patients [21,22]. Against this background this review will focus on the role of T-cells in inflammation, alterations of the T-cell compartment, and their potential role in breaking tolerance in ANCA-associated vasculitides.

Table 2.

Evidence of T-cell involvement in ANCA-associated vasculitides

Finding	Reference
T-cells infiltrate vasculitic and granulomatous lesions	[7–10]
Activation markers are up-regulated on T-cells even in remission and despite treatment	[11,12].
S(oluble)IL-2R, sCD4, sCD8, sCD30 are elevated, indicating T-cell activation	[13–16]
T-cells proliferate in the presence of PR3 and MPO	[17,18]
Predominance of IgG1-, IgG3-, and/or IgG4-ANCA suggests antigen-driven T-cell dependent immune response	[19]
T-cells are oligoclonally expanded	[20]
T-cell directed treatments induce remission in refractory disease	[21,22]

Genetic background

ANCA-associated vasculitides are generally adulthood diseases [1,2]. In a review on systemic lupus erythematosus Lipsky [23] discusses the inverse relationship between the intensity of a genetic predisposition and the time necessary to develop autoimmune disease. Thus, the delayed onset of an autoimmune disease indicates that the net genetic abnormality might be quite subtle.

Only a limited number of self-antigens become target of human autoimmune diseases. The HLA haplotype is one of the major genetic risk factors [24]. T-cell recognition of self and foreign peptide antigens is restricted by MHC polymorphism. Recognition of self-peptide/self-MHC molecules is important both for positive and negative selection during thymic T-cell development shaping the T-cell receptor (TCR) repertoire and for subsequent survival and proliferation of naïve T-cells [25]. However, under physiological conditions the majority of self-antigens are immunologically ignored. Sustained autoreactivity and immunopathology is only induced and maintained under certain inflammatory conditions [26]. Interestingly, MPO- and PR3-ANCA have been detected in some patients with severe bacterial endocarditis and secondary vasculitis. ANCA titres decline with the resolution of the infection and remission of the disease [27]. An example of how MHC alleles and an exogenous factor evoke an immune response directed against MPO is the minocycline-induced lupus-like syndrome. In this drug-induced disease patients are HLA-DQB1 positive and either HLA-DR2 or HLA-DR4 positive [28].

An association of HLA-DR2 and HLA-DR4 with ANCA-associated vasculitides has been reported [29–31]. However, it has been argued that the fact that approximately half of Caucasian individuals are HLA-DR2 positive might explain why so far no particular narrow HLA phenotype has been identified to convey susceptibility to ANCA-associated vasculitides. Other as yet unidentified genetic factors might be more important [31]. HLA-DRB1*0901 is strongly associated with MPA in Japanese patients [32]. A protective role of HLA-DRB1*13 has been postulated based on an under-representation of this allele in the population [30,33]. Recently a strong association of the HLA-DPB1*0401 allele with WG has been reported [34]. Susceptibility to another granulomatous disease, chronic beryllium disease, is also conveyed by HLA-DPB1 alleles [35].

Under physiological conditions induction of autoimmune disease is prevented by a number of mechanisms. Negative feedback stimulation via CTLA-4 during DC–T cell interaction and counter-regulatory action of cytokines may be important in down-regulating anti-self immune responses [26,36,37]. Diminished frequencies of the most effective allele for CTLA-4 expression may be one factor conferring susceptibility to WG [38,39]. A shift to the AA genotype of the bi-allelic polymorphism at position -1082 of the promoter region of the interleukin (IL)-10 gene associated with a reduced IL-10 release might be another factor favouring Th1-type responses, granuloma formation, and autoreactivity to PR3 in WG [40].

Cytokine response

Alterations of the T-cell phenotype and cytokine response have been analysed most intensely in WG [41–43]. PBMC exhibit increased IFN-γ secretion in WG. TNF-α secretion from PBMC and CD4⁺ T-cells is elevated compared to healthy controls. Both inactive and active patients produce increased amounts of monocytic IL-12 [44]. Analysis of CD4⁺ and CD8⁺ T-cells and T-cell lines and clones from peripheral blood, bronchoalveolar lavage, and granulomatous lesions from nasal biopsies with ELISA and RT-PCR disclosed predominant IFN-γ production by T-cells in WG. Few patients produce IL-4 [45]. Immunohistochemical staining of granulomatous lesions from nasal biopsies disclosed abundant IFN-γ and CD26 (optional Th1-type marker) expression in localized WG, i.e. very early WG restricted to the respiratory tract. PR3-ANCA is often not detected in this very early stage. IFN-γ and CD26 expression is less prominent in generalized WG, i.e. once the disease has progressed and affects many organs. PR3-ANCA is nearly always detected in generalized WG. The number of CD3⁺ T cells in granulomatous lesions ranges from 50 to 70%, whereas 30% of the cells are CD14⁺ monocytes and macrophages. PBMC from localized WG exhibit higher spontaneous IFN-γ production compared to generalized WG. IL-4 and IL-10 mRNA is found in higher amounts in PBMC in generalized WG compared to localized WG [46]. Another study found a marked T-cell infiltrate and increase of IL-4 expression but no IFN-γ in immunohistochemical analyses of nasal biopsies in generalized WG. However, it was not explicitly denoted whether granulomatous and/or vasculitic lesions were present. Further, a different antibody than in the aforementioned study was used for IFN-γ staining [47]. Flow-cytometric analysis discloses up-regulated expression of Th1-type CC-chemokine receptor CCR5 and Th2-type CCR3 on CD4⁺ and CD8⁺ T-cells both in localized and generalized WG suggestive of activation and migratory capacity of T-cells. However, in localized WG Th1-type CCR5 expression is stronger on circulating T-cells as well as in granulomatous lesions compared to generalized WG [48,49]. The CCR5- and CCR3-ligand RANTES (regulated on activation normal T cell expressed and secreted/CCL5) is expressed in granulomatous lesions from the respiratory tract in WG [50]. Predominance of CCR5 expression on T-cells in early, localized WG might favour stronger recruitment of Th1-type cytokine secreting cells into granulomatous lesions in localized WG as compared to later stages of generalized WG [49].

Taken together, these data suggest that an aberrant Th1-type response might play a role during initiation of WG [51]. After a variable period of time generalized PR3-ANCA positive WG usually develops. Disease progression is associated with a ‘switch’ or further complexity of the collective T-cell response with the appearance of another subset of Th2-type cells and a less prominent Th1-type cytokine production in granulomatous lesions of the upper respiratory tract. This ‘switch’ or increasing complexity of the cytokine profile might be a consequence of further B-cell expansion and T-cell dependent PR3-ANCA production during disease progression [41,43,45,46,48,49]. In generalized WG renal lesions have been reported to be polarized towards a Th1-type response, but contain Th2-type IL-4 producing and CCR3⁺ cells as well [47,52].

The question has been raised whether the Th1/Th2 dichotomy is overstated. Th1- as well as Th2-type cytokines support both cellular and humoral immune responses to adequately match an antigen-challenge [53]. However, triggering disease activity by infections as seen in many autoimmune diseases might in parts be a consequence of the cytokine response. An anecdotal report underscores the importance of cytokines for triggering disease activity. A patient with localized WG was treated with IFN-α for acute hepatitis c virus (HCV)-infection. IFN-α treatment resulted in HCV elimination, but WG activity unexpectedly deteriorated [54]. Longitudinal studies would further contribute to establish a role of Th1/Th2-type polarized responses during disease progression [43].

CSS is characterized by a marked blood-eosinophilia and eosiniphilic vasculitis and granulomatous inflammation. Similar to WG there is an evolution from an initial stage of the disease. The disease usually starts with bronchial asthma, progresses to hypereosinophilic syndrome, and finally ends with CSS. Prolonged survival of eosinophils due to inhibition of CD95-mediated apoptosis by soluble CD95 seems to contribute to the marked eosinophilia in CSS [55]. T-cell lines produce mainly IL-4 and IL-13 in CSS suggesting that T-cells may drive eosinophilc inflammation [56,57].

Altered T-cell phenotype

T-cells display an altered phenotype in ANCA-associated vasculitides. A subset of circulating T-cells lacking the costimulatory molecule CD28 is expanded in WG [58–60]. The expansion is independent of age. Similar to findings in rheumatoid arthritis not every patient is affected by this expansion, but the majority are. An effect of different immunosuppressive regimen on the expansion and phenotype of CD28^– T-cells has not been demonstrated so far [60,61]. The expansion of CD28^– T-cells starts early in the disease process and is already evident in localized WG. With disease progression to generalized WG further expansion of CD28^– T-cells is seen [49]. The expansion of CD28^– T-cells correlates with the organ involvement [60]. CD28^– T-cells are enriched in bronchoalveolar fluid and abundantly present in granulomatous lesions in WG [61]. Although they lack costimulatory CD28 expression, these T-cells are not inert. Instead, they display potent effector functions. Peripheral blood CD4⁺CD28^– T-cells and CD4⁺CD28^– T-cells within granulomatous lesions are a major source of Th1-type cytokine secretion which is mainly restricted to TNF-α and IFN-γ[62]. CD4⁺CD28^– T-cells express the differentiation marker CD57, the activation marker and adhesion molecule CD18 (β₂-integrin), and Th1-type CCR5. Furthermore, CD4⁺CD28^– T-cells show intracytoplasmic perforin expression indicating a cytotoxic potential of these cells [49,62]. Oligoclonality and shortened telomers suggest clonal expansion and replicative senescence of T-cells in ANCA-associated vasculitides [20,63–65].

Based on the analysis of cell surface markers and functional studies of antigen-experienced T-cells CD28^– T-cells have been coined effector memory or late differentiated T-cells [66]. The Th1-type CD28^– T-cell subset in WG displays features of such effector memory/late differentiated, senescent T-cells [62,67]. Skewing of the T-cell phenotype affects the whole CD4⁺ and CD8⁺ T-cell population and might reflect a profound generalized alteration in T-cell differentiation [62,68]. This alteration also affects antigen-specific responses as has been shown for CMV-specific CD8⁺ T-cells in WG [68].

Expansion of effector memory/late differentiated, senescent CD28^– T-cells may be a consequence of cytokine effects, genetically determined or due to an antigen-driven process. It is still unresolved what causes the expansion of CD28^– T-cells in WG. Escape of senescent T-cells from regulatory control, and failure of regulatory T-cells or cytokine networks might be factors either linked to an initial antigen-driven event or at least contributing to maintenance of chronic inflammation and autoreactivity in ANCA-associated vasculitides. In line with this notion a decrease in, potentially regulatory, Th2-type Vα24⁺Vβ11⁺ NKT-cells has been reported in WG recently [69].

Autoreactive PR3- and MPO-specific T-cells

Whereas the majority of PR3- and MPO-ANCA recognize conformational epitopes of the autoantigen, T-cells recognize linear peptide sequences [3,42,70]. A rather straight-forward assumption in ANCA-associated vasculitides would suggest that phenotypic alterations and telomeric loss indicative of T-cell senescence were a direct consequence of autoantigen stimulation. However, analyses of peripheral blood T-cell reactivity to PR3 and MPO do not seem to support this assumption on first look. Proliferation of PBMC and T-cells in response to crude granular extract of neutrophils, inactivated PR3 and MPO, single or overlapping linear PR3- and MPO-derived peptides is seen in ANCA-associated vasculitides. However, similar proliferative responses to PR3 and MPO are also detected in healthy controls. Most proliferation experiments used 7–10 days of T-cell stimulation with PR3 or MPO [17,18,71–76]. Although there is evidence that some patients preferentially reacted with certain peptide sequences, and PR3-derived peptides induce more frequently and stronger proliferative responses, no immunodominant epitopes have been consistently reported for ANCA-associated vasculitides so far [17,74,75].

Proliferation and IFN-γ mRNA expression in PR3-stimulated T-cells correlates over time. IL-4 mRNA expression is not detected with T-cell proliferation in response to PR3 [47]. Using HLA-DR and HLA-A and –B binding PR3-derived peptide sequences and cytokine flow-cytometry, the frequency of peripheral blood PR3-specific CD4⁺TNF-α⁺ (0·64 ± 0·09%, mean ± SEM) and CD8⁺TNF-α⁺ T-cells (0·65 ± 0·18%) has been determined recently. Similar to the aforementioned proliferation assays there was no significant difference in frequencies of PR3-specific T-cells between WG and healthy controls. While a cytokine T-cell response is readily detected in response to foreign antigen such as cytomegalovirus pp65, a response to autoantigens such as PR3 and MPO seems to take longer possibly reflecting lower affinity and special requirements for autoantigen recognition. Frequencies of PR3-specific T-cells were higher after 10 days of stimulation suggestive of a selective expansion of this cell fraction after stimulation with time. Therefore, the ‘real’ frequency of such cells in peripheral blood might be considerably lower [77]. In contrast to these findings, Popa et al. [76] measured high IL-10- but low IFN-γ secretion by ELISA after 7 days of stimulation with whole PR3. Mayet et al. [78] showed IL-4 secretion in an ELISA in response to PR3 after T-cell proliferation.

Obviously, in interpreting these results a number of immunological and methodical aspects have to be considered. The autoantigens PR3 and MPO seem to evoke a much less vigorous, delayed T-cell cytokine response compared to foreign antigen such as CMV for instance. The rather artificial nature of the stimulation may bias cytokine responses in vitro. Proliferation assays may selectively support the expansion of single antigen-specific T-cell as well as other subsets via bystander activation. ELISA only detects the composite cytokine response of a bulk population of specifically activated cells as well as unspecifically activated cells via bystander reaction, whereas flow-cytometric analysis of cells allows analysis of cytokine production at the individual cell level [77,79–81]. Further, the plasticity of T-cell cytokine responses has to be considered and this, again, might also depend on the antigen used. Whereas distinct IFN-γ and IL-4 producing effector T-cell populations have been determined by flow cytometry [79], sequential expression of cytokines by individual T-cells has also been demonstrated [80]. Individual Th1-type cells could first support their expansion by IL-2 production and finally down-regulate their activity by IL-10 production [80].

From autoreactivity to autoimmune disease in ANCA-associated vasculitides

The TCR repertoire is shaped on self-peptide/self-MHC complexes in the thymus. Thymic T-cell selection results in the eventual release of weakly self reactive T-cells. Their low affinity for self antigens prevents full activation. The released naïve T-cells continue to survive in a resting state by contact with self-peptide/self-MHC complexes in the periphery presented, most probably, on lymphoid dendritic cells in T-cell zones [25]. T-cells with low affinity for self-antigens and T-cells specific for self-antigens that do not access the thymus (such as CNS antigens) remain tolerant and as a consequence autoantigens are generally ignored [25,82]. Still autoreactive memory T-cells are detected in healthy individuals, but seem to be controlled by regulatory mechanisms and regulatory T-cells [83]. As discussed above, T-cell proliferation in response to PR3 and MPO is also seen in healthy controls [17,18,73–76]. A number of mechanisms such as molecular mimicry, bystander-activation, aberrant costimulation, influence of cytokines, failure of appropriate negative feedback stimulation in autoantigen-recognition have been implicated in the break of tolerance and generation of activated, autoreactive memory T-cells [23–26,53,82,83]. However, the aforementioned mechanisms should favour autoimmunity against a random set of self-antigens rather than against groups of particular self-antigens or sole autoantigens as seen in human autoimmune diseases such as ANCA-associated vasculitides. Additional mechanisms have to be at work to direct the inflammatory response towards a particular autoantigen. Autoantigen presentation and recognition has to be sustained in secondary (or tertiary) lymphoid tissues to maintain generation of autoreactive T- and B-cells. Recently, the ‘inflammatory status’ of the target organ mediated by cytokines and pattern recognition receptors has been shown to be crucial for the progression from autoreactivity against a particular antigen to overt autoimmune disease in a transgene mouse model of type I diabetes [84].

So, where and how do T- and B-cells ‘see’ PR3 or MPO and react to it in ANCA-associated vasculitides? And why is WG so closely associated with a response to PR3 and MPA to MPO, but not to all constituents of neutrophil granulocytes? And how is self-reactivity to PR3 and MPO persistently maintained? So far, there are only preliminary answers. Many authors presume that granulomatous lesions seen in WG and CSS are a consequence of random leucocyte recruitment subsequent to ANCA-induced premature activation and degranulation of neutrophil granulocytes and endothelial damage. In contrast, Wegener himself [85] and Fienberg [86] suggested that WG starts as granulomatous disease in the respiratory tract and systemic vasculitis develops subsequently. Moreover, it seems that relapse is related to persisting granulomata in WG [87]. Vasculitis is not present in all cases and necrotizing granulomatous inflammation apart from vessels seen in specimen of lung tissue in WG [49,88,89]. There is preliminary evidence of (tertiary) lymphoid-like structures and possible maturation of autoreactive PR3 – specific B-cells in granulomatous lesions in WG [90]. Neutrophil granulocytes and PR3 are found more frequently in such lesions compared to other granulomatous diseases [91]. The inactivation and processing of PR3 from apoptotic neutrophils via antigen-presenting cells such as dendritic cells is aberrant in patients with ANCA-associated vasculitis [92–94]. This might facilitate the recognition of the particular autoantigen (PR3) in tertiary (and secondary?) lymphoid structures in WG.

The initiating mechanisms driving granuloma formation in WG and CSS and/or triggering ANCA-mediated vasculitis in WG, MPA, and CSS have still not been clearly determined [87]. Chronic nasal carriage of Staphylococcus aureus is a risk factor for disease exacerbation of WG [95]. Infection has been shown to induce bronchial lymphoid tissue (iBALT) in an animal model recently [96]. Thus, infections might trigger granuloma formation resembling tertiary lymphoid tissue in WG. Apart from inhaled agents other potential inducing factors are drugs, in particular antithyroid drugs, and inorganic chemicals such as hydrocarbons and especially silica [87]. T-cells could contribute to granuloma formation as suggested by in vitro and in vivo studies (Table 3) [97–99]. Alteration of the T-cell response and anomalous autoantigen-presentation in lymphoid-structures facilitates development of autoimmune disease in ANCA-associated vasculides (Fig. 2).

Table 3.

Role of T-cells in granuloma formation in in vitro and in vivo models.

Model	Main findings	Reference
In vitro granuloma model	• Activated T-cells induce necrotic centres in granulomatous lesions	[97]
In vivo granuloma model	• Activated CD4 Th1-type cells induce the transformation from an unspecific microabscess to granuloma formation	[98]
	•T-cells and IFN-γ are crucial in granuloma formation	[99]

Fig. 2.

Principal illustration of the T-cell interaction with other cells in ANCA-associated vasculitis: alteration of the T-cell response with expansion of late-differentiation phenotype cells, vigorous cytokine responses, and aberrant costimulation might facilitate autoantigen recognition mediated by antigen-presenting cells (APC), and finally maturation of autoantigen-specific B-cells. PR3 and MPO may be released from polymorphnuclear cells (PMN).

Conclusion

There is substantial evidence that T-cells are off balance in ANCA-associated vasculitides. As outlined above, genetic risk factors (HLA-DR/HLA-DP, CTLA-4, IL-10) may influence shaping of the TCR repertoire and regulatory control of T-cells. T-cells are found in inflammatory lesions. Vigorous Th1-type responses are seen in WG and MPA and a Th2-type response in CSS. Oligoclonality and shortened telomers indicate antigen-driven clonal expansion and replicative senescence of T-cells in ANCA-associated vasculitides. Potent CD28^– Th1-type cells displaying an effector-memory/late differentiated, senescent phenotype are expanded in peripheral blood and are found in granulomatous lesions in WG, but it is not clear, what causes their expansion. Differences in peripheral blood T-cell proliferation and frequencies as measured in response to the autoantigens PR3- and MPO have not consistently been detected in patients with ANCA-associated vasculitides compared to healthy individuals in vitro. However, a closer look at the in vivo interaction of T-cells with (auto)antigen-presenting cells and B-cells in lymphoid structures is needed to understand why and how autoimmunity solely against PR3 or MPO is induced in ANCA-associated vasculitides. In the future much will depend on disclosing how (auto)antigen-presenting cells and T- and B-cells interact in vitro and in vivo in order to understand what kicks T-cells off balance in ANCA-associated vasculitdes.