A role for humoral immunity is well established in the pathogenesis of Wegener's granulomatosis (WG), an antineutrophil cytoplasmic antibody (ANCA)–associated systemic vasculitis. Reports describing patients with active WG state that in 70–80% of cases ANCA are directed to proteinase 3 (PR3) and in ∼10% of cases to myeloperoxidase (MPO) (1, 2). Clinical and experimental evidence support a pathogenic role for ANCA IgG (3, 4).
Cellular immunity has been implicated as an important contributor to the development of WG as well. First, the characteristic granulomatous inflammation observed in different tissues strongly supports a role for T cells. Second, while PR3 is a target for the humoral response, it also plays a role in the cellular response (5). Third, the distribution of ANCA IgG subclasses suggests a T cell–dependent immune response, since IgG1 and IgG4 predominate in some studies (6). Finally, disease activity responds to some T cell–directed therapies (7–9).
Glomerulonephritis is an important disease manifestation of WG. In experimental systems, transfer of anti-MPO antibodies can induce necrotizing crescentic glomerulonephritis (4). Depletion of CD4+ T cells attenuates crescentic renal injury in experimental autoimmune anti-MPO glomerulonephritis (10). Effects of anti-PR3 antibodies were also studied in vivo (11–13), but induction of necrotizing crescentic glomerulonephritis was not described, likely due to differences between human and murine PR3. Presumably, there is interplay of humoral and cellular components in the development of glomerulonephritis in WG.
While acknowledging the extensive bidirectional communication between immune cells (e.g., between T and B cells), this review focuses specifically on T cell characteristics in WG. Much research has focused on features of peripheral blood T cells in this disease and is discussed here. Additionally, T cell abnormalities in the kidneys are addressed, and some attention is paid to lesions in the upper airways. As a guide, an overview of peripheral blood and renal T cell abnormalities is given in Figure 1.
Peripheral blood T cell abnormalities
Lymphopenia with markedly low numbers of CD4+ T helper cells.
Total lymphocyte numbers and absolute and relative numbers of CD4+ T helper cells are low in WG patients compared with controls, irrespective of disease activity (14). Lymphopenia in WG could result from treatment effects, but may also occur due to extensive recruitment of peripheral blood T cells into diseased tissues. Marinaki et al described several patients with reduced total lymphocyte and CD4+ T cell counts who had inactive vasculitis and were not treated with immunosuppressants (14). However, these results do not rule out therapy effects, since these patients did receive therapy during prior active disease. Therefore, immunosuppressants could be a major cause of the absolute reduction in total lymphocytes and, consequently, in CD4+ cell counts in WG. Nevertheless, the observation that a relative reduction in CD4+ cells occurs in WG patients before initiation of immunosuppressants supports the notion that the percentage of CD4+ cells seems to be less influenced by therapy and may be more disease-related (15).
While it is currently impossible to be conclusive about the origin of the lymphopenia in WG, it is clear that it is accounted for by a marked reduction in the CD4+ T cell population, characterized by decreased absolute numbers and percentages (14–19). Percentages and absolute numbers of CD8+ T cells are relatively increased (15, 16, 19). This shift in the CD4+:CD8+ T cell ratio seems to be of functional importance, since patients have a reduced ratio during all disease stages, and the ratio is particularly reduced in patients with long-lasting severe disease or renal involvement (19, 20). Presumably, the CD4+:CD8+ T cell ratio shift occurs mainly due to a marked reduction in the CD4+ T cell compartment, but theoretically a CD8+ T cell expansion could also result in this shift. In fact, recently an expanded subset of CD8+CD28+CD11b+ T cells was described in patients, and not in controls, that produced interferon-γ (IFNγ) ex vivo (19).
In conclusion, lymphopenia might result from a combined effect of therapy and disease. The relative reduction in CD4+ cells and the shift in the CD4+:CD8+ T cell ratio are probably disease-related. For the patient, however, lymphopenia is not necessarily bad, since marked lymphopenia has been associated with a lower incidence of relapse (21).
Skewing toward a memory population; effector memory T cells.
Various studies further investigated the phenotypes of the cells within the reduced CD4+ T cell population. Low absolute and relative numbers of naive CD4+CD45RBhigh cells were reported in patients compared with controls, indicative of skewing toward a memory population (14). The reduction in the total CD4+ cell population seems to be compensated for by a relative increase in the CD4+ memory cell subset. Even in WG patients with disease in remission, fewer naive CD4+ T cells and relatively more CD4+ effector memory T cells were found compared with controls, suggesting persistent skewing toward a memory phenotype regardless of disease activity (22).
Interestingly, WG patients with disease in remission had higher counts of peripheral blood CD4+ effector memory cells than did patients with active disease (22). Abdulahad et al hypothesized that the relative decrease in peripheral blood CD4+ effector memory cells during active disease resulted from selective migration of these cells into inflammatory areas, since these cells were detected in the urine of patients with active glomerulonephritis (23).
Priming by and persistence of a strong stimulus is needed for differentiation and maintenance, respectively, of CD4+ effector memory cells. In WG, a persistent antigenic trigger might result in prominent differentiation of naive T cells. The nature of this antigenic trigger can be manifold, but potentially Staphylococcus aureus–derived stimuli play a part, because this infectious agent has been implicated in WG (24, 25). Interestingly, antibodies to FimH, a protein carried by fimbriated organisms, cross-react with human lysosome-associated membrane protein 2 and have been correlated with development of necrotizing glomerulonephritis; it was not reported whether FimH activates CD4+ effector memory cells (26).
Activated CD4+ effector memory cells express CD25 and phenotypically resemble CD4+ Treg cells. Staining for FoxP3, however, a transcription factor expressed by Treg cells, revealed no selective increase of FoxP3+ cells within the CD4+ effector memory cell population. Therefore, it is unlikely that these cells are regulatory, and they presumably represent a persistently activated cell population (22). Persistent T cell activation is crucial for inducing autoimmunity in animal models (27) and has been correlated with disease severity (28).
In conclusion, CD4+ effector memory cells probably represent persistently activated T cells. The persistent presence of antigens, possibly extrinsic antigens derived from infectious agents, could trigger these T cells to proliferate.
Increased numbers of activated T cells.
Both CD4+ and CD8+ T cell populations were found to be activated in WG (16). Several markers, including CD25 and HLA–DR, characterize T cell activation. Correlations between disease activity and activation markers, such as CD69 and soluble CD30, were previously reported (15, 29). Recent investigations focused on CD25 and HLA–DR (14, 30).
CD25, the α-chain of the interleukin-2 receptor (IL-2R), is expressed on naive T cells upon activation and by Treg cells. CD25 expression is increased on CD4+ cells in WG irrespective of disease activity, likely signifying persistent activation since no increase in FoxP3 messenger RNA expression was observed (14, 30). Whereas the absolute number of CD4+CD25+ cells was decreased in patients, the percentage of CD4+ cells expressing CD25 was increased and was occasionally as high as ∼80–90% (14).
Apart from increased surface expression of CD25 on circulating T cells, plasma levels of soluble CD25, commonly referred to as soluble IL-2R (sIL-2R), are also elevated in patients (15, 31). Increased levels of sIL-2R, indicating T cell activation, correlated with disease activity in some studies (31, 32). Soluble IL-2R levels increased prior to and remained high during relapses. Major relapses were characterized by profound rises in sIL-2R (32, 33); however, elevated sIL-2R levels were also reported in patients with clinical remission of disease (34), supporting the notion of persistent T cell activation (14, 22). Notwithstanding contrasting results concerning the relationship of sIL-2R levels to disease activity, sIL-2R levels could be a parameter in monitoring disease activity.
HLA–DR is another well-known activation marker; it is expressed later than CD25 in the course of activation and persists longer on the cell surface. High expression of HLA–DR was described in expanded T cell subsets in patients, indicating chronic stimulation and activation (30). A high number of cells consistently expressed HLA–DR as well as CD25 regardless of the clinical status of patients. In WG, this discordance between the in vivo activation status of T cells and disease activity measured with clinical parameters has been reported by several investigators (35, 36). Expanding on this, Schlesier et al reported no correlation between T cell activation markers and disease duration, therapy, or laboratory parameters (16). However, those investigators did report a relationship between increased numbers of CD4+HLA–DR+ T cells and the degree of organ involvement.
To summarize, there is evidence of enhanced and persistent activation of T cells in the circulation in WG, suggesting an ongoing immune response during clinically inactive disease. The persistent pool of activated T cells may have the potential to reactivate the disease in the tissues and initiate relapses, but its clinical significance is incompletely understood. A recent study potentially adds to this understanding by demonstrating an increased percentage of CD134+ and glucocorticoid-induced tumor necrosis factor receptor–positive CD4+CD25+ lymphocytes in WG and by showing a correlation with disease activity (37). CD134+ T cells were also found in affected tissues, including the kidney (37).
Functional defect in Treg cells.
Treg cells are important inhibitors of T cell activation that function in peripheral tolerance and down-regulate immune responses to self antigens and non-self antigens, presumably at the T cell activation stage. Animal experiments demonstrated that Treg cell depletion and abnormalities can give rise to autoimmune diseases, such as glomerulonephritis (38). In the human autoimmune disease systemic lupus erythematosus (SLE), defective regulatory function of CD4+CD25+ Treg cells has also been observed (39).
In WG, little is known about Treg cell function. The previously described increase in T cell activation markers in WG could theoretically be associated with diminished Treg cell function. In the absence of adequate Treg cell activity, autoreactive T cells could be activated peripherally upon encountering self antigen, which would lead to increased expression of activation markers on circulating naive CD4+ cells. A functional defect of circulating Treg cells in WG patients with disease in remission was recently demonstrated (40). Patient CD4+CD25highFoxP3+ Treg cells induced a significantly lower level of suppression of proliferation of responder T cells compared with healthy control Treg cells. A limitation to that study, however, is that FoxP3 can be transiently up-regulated in activated CD4+CD25+ nonregulatory T cells as well (41). Furthermore, the functional defect demonstrated could be related to prior immunosuppressants.
Aberrant T cell costimulation
Reduced expression of costimulatory molecules.
T cells normally constitutively express the costimulatory molecule CD28 (42). During T cell receptor (TCR) stimulation by antigen, concomitant interaction of CD28 with its ligands CD80 and CD86 (B7-1 and B7-2) results in clonal T cell expansion and initiation of effector functions. When costimulatory signals for CD28 are absent during TCR engagement, CD28+ T cells become anergic or undergo apoptosis. CD28– T cells can potentially resist anergy/apoptosis following TCR engagement in the absence of CD28 costimulation. By this mechanism, these cells could escape important processes that function to prevent autoimmunity. Low CD28 expression could be linked to a skewing toward memory cells, since memory cells depend less on CD28 costimulation (15).
CD28 expression on peripheral blood T cells is low in WG patients compared with age-matched controls (15, 19) irrespective of therapy or disease phase, as demonstrated in treated and untreated patients with or without active disease. Interestingly, the low CD4+:CD8+ T cell ratio seen in patients, as previously described, reportedly relates to a high percentage of CD8+CD28– T cells (19). Lower fractions of CD28-expressing cells were observed in both CD4+ and CD8+ expanded T cell populations (30). Several interesting observations were reported regarding CD28– T cells in WG. First, low CD28 expression was reported in combination with high T cell CD80/CD86 expression after in vitro stimulation (43). Second, patients with relatively low CD28 expression had more severe disease (43). Third, high CD57 expression was described in expanded CD4+CD28– T cell populations, despite the fact that CD57 is not commonly expressed on CD4+ cells (30). Intracellular perforin was demonstrated in CD4+CD28–CD57+ cells, indicating cytotoxic potential (44). CD57 also marks proliferative inability, which occurs after many cell divisions. Underlining this, shortened telomeres were observed in patient T cells, supporting T cell activation with multiple cell divisions (45). Finally, CD57 and CD11b expression were increased on patient CD8+CD28– T cells. CD11b marks recently activated effector CD8+ cells (46). Cytotoxic CD8+CD28–CD57+CD11b+ cells might directly contribute to vascular endothelial damage and induce or aggravate vasculitis, although there is currently no evidence that this occurs in vivo (16).
Aberrant expression of inhibitory molecules.
The outcome of antigen-mediated TCR stimulation is not solely regulated by CD28 costimulatory signals, but also by inhibitory signals from, for example, CTLA-4 (47). Under normal conditions, CTLA-4 is expressed at the highest level after T cell activation (42), and the interaction between CTLA-4 on activated T cells with CD80/CD86 prevents or terminates T cell activation and inhibits proliferation.
In WG, CTLA-4 is highly expressed in expanded CD4+ T cell populations and even on B lymphocytes (30, 48), and these CTLA-4+ cells might dampen excessive T cell activation. The triad of low T cell CD28 expression, high T cell CD80/CD86 expression, and high T and B cell CTLA-4 expression in WG is intriguing. Loss of CD28 expression is possibly intrinsic to some potentially pathogenic T cell clones in patients. Concomitant high expression of CD80/CD86 and CTLA-4 on T and B cells might reflect a counteractive regulatory mechanism (Figure 2).
The precise role of CTLA-4 in vivo, apart from inhibitory signaling to T cells, is somewhat puzzling. For instance, CTLA-4 reportedly induced apoptosis of activated T cells (49, 50) while also preventing apoptosis via up-regulation of the Bcl-2 antiapoptotic protein (51). CTLA-4 is also more frequently expressed on Th2 cells and might thereby influence the susceptibility of Th2 populations to apoptosis (51). Furthermore, CTLA-4 was independently reported to have a role in inducing a Th1 response (52), making it difficult to predict the effects of high CTLA-4 expression in vivo.
Of note, WG has been associated with polymorphisms in the CTLA-4 gene. Specifically, a CTLA-4 single-nucleotide polymorphism in the promoter region at position –318 and aberrant numbers of (AT) repeats in the 3′-untranslated region of the CTLA-4 gene were associated with WG in some studies (53–55). These studies showed longer CTLA-4 (AT)n alleles to be more prevalent in patients than in controls (54, 55), and interestingly this polymorphism also prevailed in patients with rheumatoid arthritis (RA) and myasthenia gravis (56–58). However, data on associations with CLTA-4 gene polymorphisms are inconclusive, since investigators in other studies did not report these associations with WG (59, 60). These discrepancies between studies may result from insufficient power of individual studies, differences in patient selection and ethnicity between study populations, and disease heterogeneity (e.g., limited versus generalized disease).
In addition to CTLA-4, other inhibitors of T cell activation following TCR engagement might be important for disease development. Indeed, an association of combinations of polymorphisms in programmed death 1 and CTLA-4 has been reported in WG (59). Another negative T cell regulatory protein is protein tyrosine phosphatase N22 (PTPN22). One well-studied PTPN22 polymorphism in autoimmune diseases such as RA and SLE is the R620W polymorphism (61). A large study comparing WG patients and controls demonstrated an association of the R620W polymorphism with ANCA-positive WG (62). In mice lacking PTPN22 activity, proliferation of effector memory T cells was increased and germinal centers were formed in the spleens and Peyer's patches of these animals (63). Defective PTPN22 function could therefore facilitate autoimmune disease characterized by increased effector memory T cell activation and aberrant antibody production.
In conclusion, evidence supports the notion of aberrant inhibition of T cell activation in WG. Potentially, mutations in several molecules that are involved in regulating T cell responses predispose to development of autoimmune disease, including WG.
CD4+ T helper cell subsets
Th1 and Th2 cells.
CD4+ T helper cells can be classified as Th1 or Th2 cells based on their cytokine profile and function. Th1 responses activate macrophages and function in cellular immunity, while Th2 responses down-regulate Th1 responses and function in humoral immunity (64).
As mentioned previously, peripheral blood T cell CD28 expression is low in patients, and its ligands CD80 and CD86 are up-regulated. CD28 costimulation promotes production of Th2 cytokines, such as IL-4 and IL-5 (65, 66). Without CD28 costimulation, cells differentiate into Th1 cells (67), and thus reduced CD28 expression promotes a Th1 response characterized by high IFNγ production (43, 68). Th1 cells depend on CD80/CD86 costimulation for activation more than do Th2 cells (69); therefore, increased CD80/CD86 expression may further induce a Th1 response (43). High CTLA-4 expression may direct T helper cells toward both differentiation pathways, as previously described.
Investigators in several studies report that peripheral blood T cells of patients produce mainly IFNγ (36, 68, 70) and IL-2 (36), and that IL-4, IL-5, and IL-10 production is low (68, 70), in line with a Th1 response. What triggers this cytokine profile has not yet been elucidated. Activated cell-mediated immunity in the absence of immunoglobulin deposition, which characterizes pauci-immune glomerulonephritis, is unique to ANCA-associated disease and could imply a local role for a Th1-type response driven by IFNγ as well (71).
A Th1 cytokine pattern rich in IFNγ and poor in IL-4 has been found to predominate in granulomatous inflammation in WG (68). In that study, T cells from nasal granulomas of patients with systemic WG produced IFNγ but no IL-4. Introduction of IL-10 was demonstrated to dampen Th1 responses in WG, and in vitro studies showed that high levels of IL-10 suppressed proliferation of autoreactive T cells (5, 70), indicating a potential therapeutic value of IL-10 administration.
Despite these findings, it is a topic for debate whether a Th1 response is prototypical for WG. In direct contrast to the above-mentioned research, investigators in several studies report a predominant Th2 response in WG (5, 22), and stimulation with PR3 was reported to elicit a Th2-skewed cytokine pattern as well (5). It has been suggested that Th1 responses predominate in localized disease and that a shift toward a Th2 response occurs in generalized disease (72). Considering comparable disease manifestations in the nasal mucosa, both a local Th1 response (73) and Th2 response (73, 74) have been reported. In those studies the local Th1 response was found in patients with localized disease, while the Th2 response was demonstrated in patients with generalized disease. However, the previously mentioned study by Csernok et al (68) does not support this dichotomy, since they reported a Th1 response in the nasal mucosa in patients with generalized disease.
Overall, because of conflicting evidence, it remains unresolved whether a Th1 or Th2 response predominates in WG. It cannot be concluded that there is a shift from a Th1 response in localized disease to a Th2 response in generalized disease. Further studies might bring novel insights into the nature of T helper cell responses in WG and might provide an incentive for the development of novel therapies.
A third type of T helper cells are Th17 cells, which have received recent attention since they are possibly related to the development of autoimmunity. Th17 cells, thought to form a bridge between innate and adaptive immunity, originate from activated CD4+ cells and produce the cytokine IL-17 (64), which acts on endothelial cells and reportedly plays a role in RA and SLE (75, 76). A recent study of WG patients with disease in remission demonstrated an increased percentage of IL-17–secreting CD4+ Th17 cells in peripheral blood (77). Upon stimulation with PR3, higher percentages of Th17 cells were demonstrated in peripheral blood lymphocyte samples from ANCA-positive patients compared with samples from ANCA-negative patients and controls.
In conclusion, currently little is known about Th17 cells in WG. However, a recent study indicates their potential role in this autoimmune disease.
Abnormal expansions of T cells, especially CD4+ T cells expressing certain TCR Vα or Vβ genes, were found in WG patients compared with controls (78). For example, a significant increase in the mean percentage of Vβ 2.1 gene expression was reported in peripheral blood of patients (79).
Grunewald et al notably discovered a common dominating motif in the third complementarity-determining region (CDR3), the most variable CDR of the TCR, in 4 unrelated patients. A common motif implies recognition of a specific antigen in these patients, suggesting the existence of a common disease-associated antigen (80).
In search of specific T cell responses, much research has focused on PR3 and, to a lesser extent, MPO. Investigators in several studies report PR3- and MPO-specific T cell responses in vasculitis and, more specifically, WG (81–84). PR3 is thought to stimulate the development of autoreactive T cells in peripheral blood of patients. While PR3 promotes proliferation of patient CD4+ cells in vitro, it also cross-stimulates control cells (5). Investigators in several studies did report, however, that T cell responses to PR3 were higher in patients (81–83). A large study on peripheral blood T cell responses to ANCA antigens in patients, healthy controls, and disease controls showed higher PR3-specific responses in patients during all disease stages (active, remitting, treated, and untreated) (83). In contrast, Winek et al did not observe significant differences in frequencies of PR3-specific T cells between WG patients and controls (84).
Regarding the findings from several studies, PR3-specific T cell responses appear stronger in WG patients than in controls, but because this is not confirmed by all reports, presumably other disease-related antigens could also induce expansion of populations with specific TCRs. Underlining this, S aureus–specific CD4+ cells were identified in WG (85). These T cells were HLA–DR restricted and secreted Th2-type cytokines. Strikingly, several of these cells also recognized PR3. However, evidence is inconclusive, since T cell expansions that were significantly more frequent in WG patients than in controls were also reported not to be associated with S aureus or its superantigen. The data from that study, by Popa et al, suggest that S aureus does not exert its potential pathogenic function via superantigenic T cell activation (86).
A closer look at renal histology
So far, this review has highlighted data on T cell abnormalities from studies of peripheral blood mononuclear cells. Next to studies on peripheral blood T cells, there are also relevant studies on T cells in tissues, some of which were recently reviewed by Lamprecht et al (87) and Mueller et al (88). Importantly, they emphasized the similarities between WG granulomata and lymphoid follicle–like structures, suggesting that such structures might be the key to understanding the self-perpetuating inflammation and autoimmunity in WG. There is less evidence on T cell abnormalities in the kidney, but several studies have thoroughly investigated renal biopsy samples for the presence of T cells.
In the kidneys of patients with crescentic glomerulonephritis, interstitial T cell numbers are significantly increased (89) and correlate negatively with renal function (90). Particularly surrounding glomeruli with heavy crescent formation or sclerosis, CD3+ T cells are abundant (90, 91), and infiltrates comprised approximately even numbers of CD4+ and CD8+ cells (91).
While significant increases in interstitial, and often periglomerular, T cells are reported (Figure 3), T cells are relatively scarce in glomeruli (71, 90). Considering intraglomerular T cells, CD8+ cytotoxic T cells are present in higher numbers than CD4+ cells (90, 91). Within glomeruli, monocytes are the dominant cell type (71, 89–92).
As previously mentioned, CD25 expression on peripheral blood CD4+ T cells is high in WG, presumably denoting T cell activation. In contrast, immunohistochemical studies revealed that while CD4+ T cells were abundant in renal lesions, CD25 expression was weak (14). One explanation would be that mainly peripheral blood T cells are activated in WG. Alternatively, a subset of CD4+CD25+ cells consists of Treg cells, and lack of these cells in inflammatory lesions could indicate insufficient immunoregulation locally. A scarcity of Treg cells could indirectly result in tissue damage (14).
While no specific increase in CD25+ cells was reported at the tissue level, more activated CD45RO+ memory T cells were present in inflammatory lesions in pauci-immune glomerulonephritis than in “humorally mediated,” noncrescentic glomerulonephritis (71). This suggests that T cell activation does occur locally in pauci-immune glomerulonephritis.
WG patients are lymphopenic, which seems compensated for by an increase in activated and memory T cells. Expression of activation markers is high on peripheral blood T cells and is generally irrespective of disease activity or therapy. The skewing toward memory cells may explain the relapsing and remitting disease course, and these cells could be targeted for developing new therapies.
The role of Treg cells is not yet clear. Expanded CD4+CD25+ T cell populations are currently not thought to be predominantly regulatory but to be persistently activated effector cells. Persistent activation fits with a regularly relapsing disease and supports the hypothesis of a persistently present, strong antigenic trigger in WG. Furthermore, evidence supports the hypothesis of aberrant T cell costimulation in WG, and there seem to be disease-specific TCR responses, for example, to PR3. Regarding the predominant T helper cell response(s), the Th17 cells are of novel interest.
Evidence for activated cell-mediated immunity from studies on peripheral blood is substantial. Regarding pauci-immune glomerulonephritis, at sites of glomerular injury, effectors of cell-mediated immunity are also prominent. However, translation of peripheral blood T cell abnormalities to the tissue level remains difficult. Expression of activation markers like CD25 on peripheral blood T cells is generally high, but this is not reflected in renal biopsy samples.
In conclusion, in WG there are evident alterations in T cell characteristics in peripheral blood, even during quiescent disease. The origin of these abnormalities can likely be found on both a genetic and an environmental level. Homing of effector T cells into tissues probably drives the pathophysiology through interactions with resident and infiltrating cells that are more frequently present in tissue structures than in peripheral blood.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Ms Berden had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Berden, Kallenberg, Abdulahad, de Heer, Bruijn, Bajema.
Acquisition of data. Berden.
Analysis and interpretation of data. Berden, Kallenberg, Savage, Yard, Bajema.