T-helper cells as new players in ANCA-associated vasculitides

In anti-neutrophil cytoplasmic autoantibody-associated vasculitides (AAV), several observations support a key role of T-helper cells (CD4+ T cells) in disease pathophysiology. An expanded population of effector memory CD4+ T cells in AAV patients may contribute to tissue injury and disease progression. In addition, functional impairment of regulatory T cells (TRegs) is reported in AAV patients. A fraction of TRegs have the capacity to differentiate into Th17 cells in the context of a proinflammatory environment. Therefore, nonfunctionality of TRegs described in AAV patients may be caused by their conversion into IL-17-producing cells that may contribute to granulomatous vasculitis. Further investigations directed at the plasticity of TRegs in AAV patients are warranted.


Introduction
Anti-neutrophil cytoplasmic autoantibody (ANCA)associated vasculitides (AAV) constitutes a group of disorders characterized by autoimmune necrotizing infl ammation of small blood vessels, which leads to systemic organ damage [1]. Th is group of systemic vasculitides includes Wegener's granulomatosis (WG), microscopic polyangiitis (MPA), and Churg-Strauss syndrome (CSS). Th ese disorders are predominantly associated with the presence of circulating ANCAs that are directed against proteins in the cytoplasmic granules of neutrophils. ANCAs with specifi city for proteinase-3 (PR3-ANCA) are associated with WG to a high degree, whereas ANCAs with specifi city for myeloperoxidase (MPO-ANCA) are predominant in MPA and to a lesser degree in CSS [2]. Although it remains unknown how these conditions develop, it has been postulated that ANCA in vivo bind to surface-expressed autoantigens (PR3 or MPO) on primed neutrophils, which enhances neutrophil degranulation and the release of toxic products that cause endothelial damage, ultimately leading to necrotizing vasculitis [2].
In vivo experimental studies have clearly demonstrated that MPO-ANCAs are pathogenic factors. Xiao and colleagues have shown that immunization of MPOdefi cient mice with mouse MPO results in an MPOdirected immune response, and transfer of splenocytes from these mice into immune-defi cient mice leads to development of pauci-immune necrotizing crescentic glomerulonephritis and systemic necrotizing vasculitis reminiscent of MPA [3]. Further support for the pathogenicity of ANCA comes from a recent study by van Timmeren and coworkers [4]. Th ey observed that adminis tration of anti-MPO antibodies hydrolyzed by the bacterial enzyme endoglycosidase S, which abolishes IgG binding to Fcγ receptors, attenuated both neutrophil infl ux and formation of glomerular crescents in the abovedescribed model of MPO-ANCA-induced glomerulonephritis. An immunopathogenic role for MPO-ANCA has also been strongly suggested by the occurrence of neonatal MPA in a child born to a mother with a history of MPO-ANCA-associated pulmonary renal syndrome [5].
In contrast to MPO-ANCA, in vivo evidence is still lacking for a direct vasculitic pathogenicity of PR3-ANCA. So far only PR3-induced and PR3-ANCA-induced enhancement of infl ammation has been demonstrated in an animal model [6,7]. Recent fi ndings by Primo and colleagues suggest that, under certain conditions, anti-PR3 antibodies can be pathogenic in rodents [8]. Th ey showed that adoptive transfer of splenocytes from PR3immunized mice into NOD-SCID mice resulted in the appearance of circulating anti-PR3 antibodies and crescentic glomerulo nephritis in the recipient mice. However, it is unclear whether glomerulonephritis in recipient mice is mediated by the humoral or the cellular arm of the anti-PR3 response.
Of note, infi ltrating T cells in granulomatous lesions as well as persistent T-cell activation have been reported in AAV patients [9][10][11][12]. Interestingly, T-cell-depleting therapy with anti-CD52 antibodies (alemtuzumab) and anti-thymocyte globulin can induce remission in refractory AAV patients [13,14]. Moreover, the IgG subclass distribution of ANCA, predominantly consisting of IgG 1 and IgG 4 , implies isotype switching of ANCA for which CD4 T-helper cells are required [15]. T-cellmediated immunity is thus thought to contribute to the patho genesis of ANCA-associated vasculitis. In the present review we will summarize the currently available data on the role of T cells in AAV. We shall fi rst discuss current thoughts about the contribution of T cells to tissue injury in AAV. Th e main emphasis will then be on the plasticity of regulatory T cells (T Regs ), their transition into Th 17 cells, and the involvement of Th 17 cells in granuloma formation and disease progression.

Involvement of CD4 T cells in AAV
Unlike other autoantibody-mediated diseases, AAV is characterized by an absence of deposited antibodies in aff ected tissue, in particular in glomeruli, designated as pauci-immune glomerular lesions [16]. Otherwise, immune eff ector cells such as CD4 + T cells, macrophages and granulocytes are enriched in granulomatous lesions [9,10,[17][18][19][20]. Th is suggests a primary role of cell-mediated immunity in initiating granuloma formation. Studies in mice and humans have demonstrated a key role of CD4 + T cells in the generation of a granulomatous response. For instance, Saunders and colleagues have shown that CD4-defi cient mice did not generate the typical mononuclear granulomatous lesions following Mycobacterium tuberculosis infection [21]. In humans, the extent of granuloma formation was correlated with peripheral CD4 T-cell counts in HIV patients with mycobacterial infection [22,23]. Th e important role of CD4 T cells in the expression of crescentic glomerulo nephritis has been demonstrated by Ruth and colleagues [24]. Th ey induced experimental anti-MPO-associated crescentic glomerulonephritis by immunizing C57BL/6 mice with human MPO followed by subsequent challenge with antiglomerular basement membrane antibodies. Mice depleted of CD4 + T cells at the time of adminis tration of anti-mouse glomerular basement membrane developed signifi cantly less glomerular crescent forma tion and less cell infl ux when compared with control mice. Th ese data provide convincing evidence that CD4 + T cells are crucial in granuloma formation and glomerulo nephritis.
Studies in AAV patients also support this notion. Proliferation of CD4 + T cells in response to the autoantigens PR3 and MPO have been reported in patients with AAV, although CD4 + T cells from healthy controls also proliferated in response to PR3 and MPO, albeit to a lesser extent [25]. Persistent CD4 T-cell activation has also been observed in peripheral blood from AAV patients [26][27][28]. Importantly, Marinaki and colleagues observed an association between persistent CD4 + T-cell activation and disease severity in both WG patients and MPA patients [29]. Recently, Seta and colleagues evaluated the eff ect of depleting CD4 or CD8 T cells on the proliferative response to MPO fragments of peripheral blood mononuclear cells isolated from MPA patients [30]. Strikingly, proliferation was completely lost after the depletion of CD4 + T cells, but not after depletion of CD8 + T cells. In our studies in WG patients, we observed a persistent expansion of a subset of memory CD4 + T cells, termed eff ector memory T cells (T EM ), with a reciprocal decrease in naïve CD4 + T cells [11]. Moreover, the CD8 + T-cell compartment also appears to be altered [31,32]. In accordance, infi ltrating T cells in lung lesions and glomeruli were shown to consist mainly of CD4 + T cells with a memory phenotype [9,33,34]. Also in CSS patients, given the allergic background and hypereosinophilia in this disease, activated CD4 + T cells producing Th 2 cytokines are believed to be the disease inducer [35]. Taken together, these fi ndings indicate that CD4 T cells can serve as eff ector cells in the pathogenesis of AAV.

CD4 + eff ector memory T cells: a key player in tissue injury in AAV
As mentioned above, several observations support the involvement of CD4 + T cells in the pathogenesis of AAV. Important evidence regarding their role in disease manifestations came also from the clinical observation that remission could be induced in WG patients by antibodies directed at T cells [36]. Indeed, an altered phenotype of CD4 + T cells has been found in AAV patients. An increased proportion of CD45RC Low CD4 + memory T cells was reported in peripheral blood of AAV patients [37]. In addition, an expanded population of CD4 + T cells lacking the co-stimulatory molecule CD28 was observed in peripheral blood and in granulomatous lesions of patients with WG [9,38]. Th ese CD28 -CD4 + T cells display upregulation of the T-cell diff erentiation marker CD57 and show intracytoplasmic perforin expression, indicating the cytotoxic potential of these cells [9]. Based on phenotype and functional characteristics, CD28 -T cells have been classifi ed as a T EM population that lacks the chemokine receptor CCR7 [39].
Consistent with these fi ndings, we observed a signifi cant increase in the frequency of circulating CD4 + T EM (CD45RO + CCR7 -) in WG patients in remission compared with healthy individuals [11]. In addition, we have shown that the number of these circulating CD4 + T EM decreases during active disease compared with that during complete remission, which is consistent with their migration towards infl amed tissues [11]. Indeed, our cross-sectional and follow-up studies confi rmed migration of CD4 + T EM during active renal disease into the diseased organs [40]. We observed a remarkable increase in CD4 + T EM in the urinary sediment with a concomitant decrease of circulating CD4 + T EM of WG patients with active renal involvement [40]. Th ese urinary CD4 + T EM decreased or disappeared from the urine during remission, which might refl ect their role in renal injury. In accordance with these fi ndings, Wilde and colleagues demonstrated that CD4 + T EM expressing CD134 are expanded in peripheral blood of patients with WG [41]. CD134 is thought to contribute to T-cell migration and tissue infi ltration through its interaction with OX40L on vascular endothelial cells [42]. Indeed, Wilde and coworkers have shown that CD134-expressing T cells were localized within the infl ammatory lesions of WG patients, supporting our hypothesis on migration of this T-cell subset to infl amed sites [41].
As mentioned, CD4 + T EM display natural killer (NK)like features such as cytotoxicity [39]. Th ey also mimic NK cells by their surface expression of the NKG2D molecule. NKG2D is an activating C-type lectin-like homodimeric receptor, which diff ers from other NKG2 members as it apparently lacks an antagonist and substitutes for CD28-mediated co-stimulatory signaling in CD28 -T EM [43]. One of the NKG2D ligands is the major histo compatibility complex class-I chain-related molecule A (MICA), which is expressed upon cellular injury and stress on target cells such as fi broblasts and epithelial cells [43]. Proof of concept for NKG2D-mediated tissue destruction was provided by Allez and colleagues [44], who demonstrated that NKG2D + CD4 + T-cell clones from patients with Crohn's disease kill target cells that express MICA via NKG2D-MICA interaction. Importantly, MICA is upregulated in peritubular endothelium and glomerular epithelial cells in AAV patients during active renal disease [45]. Strikingly, Capraru and colleagues have shown that NKG2D is preferentially expressed on expanded CD28 -CD4 + T cells in the peripheral blood of WG patients [31]. Next, they showed that both NKG2D and MICA are expressed in granulomatous lesions in WG, but not in disease controls. Killing mechanisms via NKG2D-MICA interaction therefore probably contribute to vessel injury and disease progression in AAV patients (Figure 1). Accordingly, selective targeting of NKG2D + CD4 + T EM or inhibiting MICA expression without impairing other parts of cellular immunity might have value in the treatment of AAV [46].

Th1/Th2/Th17 paradigm in AAV
Aberrant T-helper cell polarization has been described in AAV. Analysis of soluble markers for T-helper cell subsets in patient sera reveals a predominance of the Th 1 pattern with expression of IFNγ and sCD26 in patients with localized WG and in patients with MPA, whereas a shift towards a Th 2 pattern, with expression of IL-4, IL-5, IL-10, IL-13, sCD23, and sCD30, was observed in active generalized WG and CSS [47,48]. Th e same results were obtained from analysis of nasal granulomatous lesions in which abundant Th 1-associated markers (IFNγ, sCD26, CCR5) were seen during localized WG, whereas Th 1associated and Th 2-associated markers (IL-4 and CCR3) are found in generalized WG [20,49]. Besides the balance between Th 1 and Th 2, a recent breakthrough has revealed that IL-17-secreting T cells (Th 17) are another major pathogenic eff ector subset involved in the induction of infl ammation and autoimmunity [50,51]. It has been reported that induction of experimental autoimmune encephalomyelitis (EAE) was blocked in mice defi cient in either IL-17 or the Th 17 polarizing cytokine IL-23, whereas mice defi cient in either IFNγ or the Th 1 polarizing cytokine IL-12 show increased susceptibility to EAE [50,52,53]. Interestingly, Th 17 cells in EAE infi ltrate the brain prior to the onset of clinical symptoms, whereas Th 1 cells dominate the cellular infi ltrate thereafter when clinical disease develops [54]. It seems that T-cell-mediated disease manifestations are linked to Th 17 cells and not primarily to Th 1 responses.
Th e physiological role of Th 17 cells lies in bacterial defense -for example, against Staphylococcus aureusas shown in experimental pneumonia and the hyper-IgE syndrome [55,56]. Peptidoglycans as well as superantigens from S. aureus might have an immunomodulatory eff ect on dendritic cells by imprinting of a strong Th 17 polarizing capacity [57]. In addition, S. aureus α-toxin was shown to induce IL-17A secretion in CD4 T cells [58]. Intriguingly, nasal S. aureus co-localization has been reported to be related to relapse and correlates with endonasal activity in WG [59,60]. Infection with S. aureus might therefore drive a Th 17 response in AAV patients. Indeed, in patients with AAV we observed a skewing towards Th 17 cells following in vitro stimulation of peripheral blood samples [61]. In line with this observation, Ordonez and coworkers have shown that the expanded CD4 + memory T cells in AAV patients are a source of IL-17 [37]. Most importantly, we found a relative increase in autoantigen-specifi c Th 17 cells in ANCA-positive patients in com parison with ANCAnegative patients and controls [61]. Th is observation suggests involvement of Th 17 cells in the process of autoantibody production in AAV. Th ese results were corroborated by Nogueira and colleagues, who reported elevated levels of serum IL-17A and increased autoantigen-specifi c Th 17 cells in AAV patients during disease convalescence compared with healthy controls [62]. In addition, Saito and colleagues observed an increased frequency of circulating Th 17 cells in patients with active CSS compared with in patients with inactive disease and healthy controls [63].
IL-17 has been reported to promote the release of proinfl ammatory cytokines, which are essential for triggering the expression of PR3 and MPO on the surface of neutrophils (priming of neutrophils), and also to induce CXC chemokine release and expression of adhesion molecules responsible for the recruitment of neutrophils to the site of infl ammation [64][65][66]. Indeed, WG granulomata (site of infl ammation) are rich in neutrophils [20]. On the other hand, IL-17 has been shown crucial for the formation of an autoreactive germinal center in autoimmune BXD2 mice [67]. IL-17-producing T cells and B cells expressing the IL-17 receptor have also been reported to localize together in germinal centers [67]. Th is observation suggests that IL-17 not only induces tissue infl ammation but also could function on B cells to promote the germinal center reaction. Th e lymphocyte clusters in granuloma struc tures can resemble germinal center-like structures that might be induced by IL-17 and may provide a place for ANCA production. IL-17 thus seems to be an important player in disease development in AAV and in early granuloma formation in WG, whereas Th 1 and Th 2 cells might prevail in later stages [20,49]. Of note, Th 17 cells have not so far been demonstrated at infl amed sites in AAV.

Disturbance in the frequency and/or function of regulatory T cells in AAV
Natural T Regs , a subset of thymus-derived CD4 + T cells expressing a high level of IL-2Rα (CD25) and a unique transcription factor FoxP3, have been shown critical for preventing autoimmune responses. Defects in T Reg function or reduced numbers of T Regs have been documented in several autoimmune diseases [68]. Indeed, we found that the suppressive function of T Regs was defective in WG patients as compared with healthy controls [69]. In this group of patients, however, we observed a signifi cant increase of memory FoxP3 + CD25 High T Regs . In line with these fi ndings, Klapa and colleagues demonstrated an increased number of FoxP3 + T cells as well as phenotypical and functional alterations of T Regs in WG patients [70]. Th ey reported an increased number of interferon receptor I-positive T Regs in the peripheral blood of WG patients [70]. In addition, they showed that IFNα exaggerates functional T Reg impair ment ex vivo in response to the autoantigen PR3 [70]. T Regs in WG patients might thus display functional anergy in the context of an infl ammatory cytokine milieu.
Altered T Reg function in WG patients has also been reported by Morgan and coworkers [71]. Th ey observed that T Regs from healthy controls and from ANCA-negative patients were able to suppress T-cell proliferation to PR3, whereas T Regs from ANCA-positive patients failed to suppress this autoimmune response [71]. Dysfunction of T Regs is thus believed to play a role in the development of WG. In contrast, T Reg function in MPA patients was comparable with that in healthy controls, but FoxP3 levels were diminished in MPA patients [72]. MPA seems to be associated with a numerical defi ciency rather than a functional defi ciency of T Regs . More over, studies in CSS patients showed that both patients and controls have a similar number of CD25 + CD4 + T cells with an equal percentage of FoxP3-expressing cells. However, the suppressive function of T Regs in CSS patients still needs to be investigated [63,73].

Plasticity of T Regs in AAV: conversion towards Th17 eff ector cells within an infl ammatory milieu
A reciprocal relationship in the development of T Regs and Th 17 cells has recently been described. Th is may underlie the propensity of T Regs to convert to Th 17 cells in the context of proinfl ammatory stimuli, a phenomenon that has only recently been recognized [74][75][76]. Under neutral conditions in vitro, transforming growth factor beta can shift the balance towards functional FoxP3 + T Regswhereas in the context of an infl ammatory cytokine milieu (IL-1β, IL-2, IL-6, IL-15, IL-21, IL-23), functional T Regs convert towards IL-17-producing, nonfunctional T Regs . Th e rela tively novel notion of T-cell lineage plasticity is of interest in relation to many papers describing nonfunctional T Regs in several autoimmune conditions, including AAV. Our hypothesis is that these nonfunctional FoxP3 + T cells have lost their suppressive function due to co-expression of a second Th 17 lineage-associated transcription factor RORγt that interferes with Foxp3 activity [77]. Recently, diff erent isoforms of FoxP3 have been investigated in human T Regs that have been shown to impact T Reg function and lineage commitment. More specifi cally, the full-length isoform (FoxP3fl ) -but not the isoform lacking exon 2 (FoxP3Δ2) -interacts with RoRγt and inhibits the expression of genes that defi ne Th 17 cells [78][79][80]. Based on the aforementioned data, the putative nonfunctional T Regs described in AAV may lack their suppressive function due to upregulation of the FoxP3Δ2 isoform that fails to inhibit RORγt-mediated IL-17A mRNA trans cription. Upon stimulation in an infl amed context, these cells convert into IL-17producing eff ector T cells.
Evidence from several groups of investigators, including our own, support this hypothesis of conversion of T Regs into eff ector IL-17-secreting cells in AAV. As mentioned before, we found a signifi cant increase in the percentage of FoxP3 + CD25 High T Regs with a defective regulatory function in AAV patients in remission as compared with healthy controls. Furthermore, we demon strated a concurrent increase in the percentage of Th 17 cells upon in vitro stimulation of peripheral blood samples from AAV patients. Consistent with this, patients with AAV had signifi cantly higher serum levels of IL-17 compared with healthy controls [62]. Importantly, increased serum levels of IL-17 in AAV patients Infl ammatory cytokines (IL-1β, IL-6, transforming growth factor beta (TGFβ)) released due to bacterial or viral infections can promote skewing of a subset of functional regulatory T cells (T Regs ) towards IL-17-producing nonfunctional T Regs . These IL-17-producing cells play a key role in disease onset through their cytokine IL-17. This cytokine induces CXC chemokine release from the target tissue that will attract neutrophils to the site of infl ammation. In addition, IL-17 stimulates the release of IL-1β and TNFα from macrophages, which causes upregulation of the expression of endothelial adhesion molecules and induces translocation of proteinase-3 (PR3) and myeloperoxidase (MPO) to the neutrophil membrane (priming). Released PR3 and MPO can be processed and presented by antigen-presenting cells (APC) to T-helper cells. Since T Regs are converted into nonfunctional IL-17producing cells that fail to inhibit this autoimmune response, autoreactive T cells may undergo repeated stimulation by PR3-pulsed or MPO-pulsed APC, resulting in a pool of eff ector memory T cells (T EM ). In addition, PR3-stimulated T-helper cells act on B cells. The presence of IL-17 can enhance the production of anti-neutrophil cytoplasmic autoantibody (ANCA) by autoreactive B cells. Subsequently, ANCA binds to PR3 or MPO on primed neutrophils that adhere to endothelial cells, which enhances neutrophil activation resulting in degranulation and release of reactive oxygen species (ROS) and proteolytic enzymes that can damage vascular endothelial cells. Moreover, persistent activation of T-helper cells by PR3 or MPO, together with the breakdown of T Reg -mediated self-tolerance mechanisms, will induce autoreactive CD4 + T EM expansion. Expanded CD4 + T EM upregulate their killer immunoglobulin-like receptor (NKG2D) and interact with their ligand (major histocompatibility complex class-I chain-related molecule A (MICA)) on vascular endothelial cells, which in turn enhances their cytotoxic function and kills target cells in a perforin-dependent and granzymedependent way, ending up in vasculitis. correlated signifi cantly with increased levels of the cytokines that are involved in the conversion of T Regs into Th 17 cells; that is, IL-1β, IL-23 and IL-6 [62]. In addition, patients with active CSS showed an increased frequency of Th 17 cells with a decrease in the frequency of IL-10producing T Regs , whereas an inverse result was observed in CSS patients with inactive disease [63]. Th e aforementioned data appear to support a link between the conversion of T Regs into Th 17 cells and disease activity in AAV (Figure 1).

Therapeutic targets in patients with AAV
Because Th 17 cells contribute to infl ammation and granuloma formation, this T-cell subset could be a novel therapeutic target for AAV. Depletion of Th 17 cells by targeting specifi c surface proteins may be diffi cult as Th 17 cells share many surface markers with other T-cell subsets. A therapeutic approach targeting its cytokine (that is, IL-17) would therefore be more feasible. Indeed, neutralizing IL-17 by anti-IL-17 antibody or by soluble IL-17 receptors reduces infl ammation and bone erosion in various animal arthritis models [81]. Interestingly, humanized anti-IL-17 mAbs -including AIN457 and LY2439821, which neutralize the biologic activity of IL-17 -are in clinical trials. Th ese biologicals have been shown to induce clinically relevant responses in patients with psoriasis, rheumatoid arthritis, and non-infectious uveitis, compared with placebo without safety issues [82,83]. Neutralization of IL-17 could therefore represent a novel therapeutic approach for patients with AAV.
On the other hand, CD4 + T EM -supposed to act as a key trigger of disease expression and relapse in AAVmay also serve as a therapeutic target. Selective targeting of CD4 + T EM without impairing other parts of the humoral and cellular immune system could be a major step forward in the treatment of AAV. NKG2D blockade by anti-NKG2D antibodies has been reported to prevent autoimmune diabetes in NOD mice [84]. Blocking of NKG2D could be a new strategy in the treatment of AAV. Other studies have revealed that targeting of the voltagegated Kv1.3 channel, which is highly expressed on activated CD4 + T EM , provides a specifi c immunomodulatory approach [85,86]. Blockade of the Kv1.3 channel by ShK(L5) amide eff ectively prevented autoimmune disease in the EAE model of multiple sclerosis and suppressed delayed-type hypersensitivity in rats [85,86]. Th e selective targeting of CD4 + T EM using ShK(L5) amide and/or blocking the NKG2D-MICA interaction by anti-NKG2D antibodies may therefore hold therapeutic promise for AAV.

Conclusion
CD4 + T EM seem to be involved in tissue damage and renal injury in patients with AAV. Besides CD4 + T EM , impaired T Reg function and an increased Th 17 response are also reported in AAV patients. During the past 2 years, multiple studies indicate a link between T Regs and Th 17 cells. Indeed, in the context of an infl ammatory cytokine milieu, conversion of T Regs into IL-17-producing cells has been demonstrated. Evidence from several studies supports this conversion in AAV patients. Defective T Reg function in AAV patients can thus be explained by their conversion into eff ector Th 17 cells. Instead of suppressing autoreactive responses, these converted T Regs -through production of IL-17 -can participate in granuloma forma tion and tissue injury, which contribute to necrotiz ing granulomatous vasculitis in AAV patients. Th e mechanisms underlying the conversion of suppressive T Regs into nonfunctional T Regs in AAV await further investigation. Th is novel view into the role of converted T Regs in the pathophysiology of vasculitis will improve our understanding of AAV pathogenesis, which may lead to the identifi cation of new biomarkers and targets for therapeutic intervention.