1. Top of page
  2. Abstract


The precipitating event(s) that triggers Wegener's granulomatosis (WG) is unknown. Cytokines, costimulatory molecules, and counterregulatory molecules control the quality and intensity of immune responses. Thus, they are relevant candidates for genetic studies of immune dysregulation in WG, the pathogenesis of which may be facilitated by multiple acquired and/or inherited factors. This study was undertaken to investigate possible genetic associations of various proinflammatory cytokines and CTLA-4, a receptor for T cell inhibition, with WG.


Using polymerase chain reaction–based DNA genotyping, we investigated the polymorphisms located in the genes encoding a variety of proinflammatory cytokines and CTLA-4 in 117 American patients with WG and 123 ethnically matched healthy controls.


Compared with controls, patients with WG had a significantly lower frequency of homozygosity for the shortest allele (designated allele 86) of the Ctla4 microsatellite polymorphism (AT)n located in the 3′-untranslated region (3′-UTR) of exon 3 (47.0% versus 69.9%; P = 0.0005). Significant differences between patients and controls in the allelic and genotypic frequencies of polymorphisms in the other cytokine and cytokine receptor genes studied (tumor necrosis factor α [TNFα], TNF receptor I [TNFRI], TNFRII, interleukin-1β [IL-1β], IL-6) were not found.


The Ctla4 (AT)n 86 allele has been previously demonstrated to be crucial for maintenance of normal levels of CTLA-4 expression and balance between T cell activation and inhibition. Our results in American patients confirm findings from a Scandinavian cohort in which a positive association between WG and longer alleles of (AT)n in the Ctla4 3′-UTR was demonstrated. Diminished frequencies of the most effective allele for CTLA-4 expression may represent a WG-related susceptibility mutation that accounts, in part, for increased T cell activation and clonal expansion in WG. Blockade of T cell costimulation using CTLA-4Ig might be a useful therapeutic intervention, providing an alternative or complementary approach to conventional treatment with immunosuppressive agents.

Ligation of the T cell receptor with peptide bound on the major histocompatibility complex provides the first signal for T cell activation. However, activation does not occur until a second signal is provided by costimulatory molecules. Expressed on T cells, positive costimulators such as CD28 enhance activation, while CTLA-4 (cytotoxic T lymphocyte–associated antigen 4) down-regulates T cell activation (1, 2). CTLA-4 ligation leads to T cell quiescence or anergy, terminating T cell activation (3). CTLA-4 represents an essential checkpoint for immune function in vivo. In mice, CTLA-4 deficiency leads to death due to lymphoproliferative disease (4–6).

The gene encoding human CTLA-4 (Ctla4) contains multiple polymorphisms (7). Certain of them affect the expression and function of the protein. To date, 3 Ctla4 polymorphisms have been studied extensively (8–10) (Figure 1). A single-nucleotide polymorphism (SNP) at position −318 (relative to the transcriptional start site) in the promoter region has been shown to influence gene expression in in vitro gene reporter assays (11). The higher-secretor allele T of this SNP at −318 is protective against autoimmunity and inflammatory responses, but occurs with low frequency in Caucasian populations (11, 12). Carriers of allele G in another SNP coding sequence 1 (CDS1) at position +49 have reduced levels of messenger RNA (mRNA) and T cell surface protein expression of CTLA-4. Allele G is associated with increased susceptibility to autoimmune/inflammatory diseases (13, 14). There is strong linkage disequilibrium between the +49 SNP and a dinucleotide repeat polymorphism, (AT)n, located at position +642 of the 3′-untranslated region (3′-UTR) (10, 13, 15). It is already recognized that allele A in CDS1 is strongly linked to the shortest allele, 86, of the (AT)n polymorphism (7, 14, 16), while allele G is linked to longer alleles (alleles that have polymerase chain reaction [PCR] products longer than 86 bp) (17).

thumbnail image

Figure 1. Common polymorphisms in the human gene encoding CTLA-4. Multiple CTLA-4 polymorphisms have been found; shown are the 3 that have been reported early and studied extensively. The genetic variations and locations of the polymorphisms are indicated by arrows.

Download figure to PowerPoint

We have previously shown that individuals with long alleles of Ctla4 (AT)n have hyperreactive T cells and that the length of the AT repeat in Ctla4 has a significant linear correlation with levels of soluble interleukin-2 (IL-2) receptor in the circulation (18). Heterozygotes have less CTLA-4 mRNA from the longer allele than from the shorter allele in vivo, in part because the longer allele, which contains a higher number of AT repeats, is responsible for reduced mRNA stability, thus resulting in less protein synthesis (19). Microsatellite (AT)n in the 3′-UTR might be the major, primary genetic mutation that affects CTLA-4 gene expression. While haplotypes, rather than single mutations, may account for the altered gene expression (20), haplotypes of CTLA-4 have not yet been studied in healthy individuals or patients with Wegener's granulomatosis (WG). Nevertheless, the shortest allele of microsatellite (AT)n in Ctla4 serves as a marker for high-level CTLA-4 expression.

WG is a systemic disorder that most often affects the upper and lower airways and the kidneys. However, any organ in the body may be involved. The classic pathologic triad includes necrotizing inflammation with granulomas and vasculitis. Without treatment, patients with generalized WG have a mean survival of only 5 months (21). The etiology of WG is largely unknown. Evidence suggests that the disease is multifactorial (22), with both environmental and genetic factors contributing to its pathogenesis. Previous studies revealed a genetic association with certain Ctla4 polymorphisms in a Swedish population (16, 23). In view of the vital function of CTLA-4 in immune regulation and the ample evidence of immune dysregulation in WG, we hypothesized that variation in Ctla4 might play a role in the pathogenesis of WG. We therefore conducted this genetic association study of CTLA-4 and candidate proinflammatory cytokine genes in American patients with WG.


  1. Top of page
  2. Abstract

Patients and healthy controls.

One hundred seventeen unrelated American WG patients from the Center for Vasculitis Care and Research, Department of Rheumatic and Immunologic Diseases, Cleveland Clinic Foundation, and 123 age-, sex-, and ethnicity-matched healthy, unrelated controls were enrolled in the current study. The controls included friends or spouses who accompanied patients, and volunteers from the hospital staff. All of the WG patients and controls were Caucasian, and all were from the same general geographic area. The patient group consisted of 57 men and 60 women, and the control group included 59 men and 64 women; WG has an equal prevalence in men and women. The American College of Rheumatology 1990 criteria (24) were used in the diagnosis of WG. Forty-five patients had glomerulonephritis in addition to other features of WG (generalized WG), and 72 did not have renal involvement (limited WG). The study was approved by the Institutional Review Board of the Cleveland Clinic Foundation.

DNA extraction.

Peripheral blood samples were collected from all subjects. DNA was prepared from EDTA-preserved peripheral blood leukocytes, using standard proteinase K digestion and phenol–chloroform extraction.

Ctla4 genotyping.

A fluorescence-based genotyping method, employing an ABI Prism 377 DNA Sequencer (Perkin-Elmer Applied Biosystems, Foster City, CA), was used for determination of (AT)n polymorphisms in the 3′-UTR of Ctla4, as described previously (17).

Tumor necrosis factor α −308 promoter restriction fragment length polymorphism.

A dimorphism is located at position −308 in the promoter region of the tumor necrosis factor α (TNFα) gene. The rare TNFα allele 2 (TNF2) is responsible for the high-secretor phenotype of TNFα. Associations of TNF2 with a number of inflammatory/autoimmune disorders have been documented (25). We performed TNFα −308 restriction fragment length polymorphism (RFLP) genotyping, using primers and methods that have been described previously (23).

TNF receptor I and II RFLP.

An SNP at position +36 in exon 1 of the TNF receptor I (TNFRI) gene creates a digestion site recognized by restriction enzyme Msp AII (New England Biolabs, Beverly, MA) for allele G, but not allele A. Primer sequences and PCR amplification parameters have been described previously (26). A T-to-G substitution polymorphism at position +196 in exon 6 of the TNFRII gene was genotyped, using primers described by Al-Ansari et al (27). Allele T was recognized by restriction enzyme Nla III (New England Biolabs). Allele G is not recognized by this restriction enzyme and has recently been shown to be associated with rheumatoid arthritis in Caucasian patients in the UK (28).

IL-1β Taq I RFLP.

A Taq I RFLP located in exon 5 of the IL-1β gene has been shown to influence the gene expression of IL-1β (29). Allele 2 is associated with the high-secretor phenotype. We performed PCR-based RFLP genotyping, using methods that have been described previously (29, 30).

IL-6 −174 promoter RFLP.

The IL-6 −174 promoter RFLP has been shown to govern the rate of IL-6 expression, and allele C in this SNP (G to C) is associated with low levels of circulating IL-6 in healthy subjects (31). The primers and PCR amplification methods for the polymorphism located at −174 in the promoter or IL-6 gene have been described previously (32).

Measurement of antineutrophil cytoplasmic antibody.

Serum samples were examined for antineutrophil cytoplasmic antibody (ANCA) by myeloperoxidase (MPO)– and proteinase 3 (PR3)–specific enzyme-linked immunosorbent assays (Scimedx, Denville, NJ) and by standard indirect immunofluorescence on ethanol-fixed neutrophils. Patients with detectable levels of circulating ANCAs at any time during the disease course were classified as ANCA positive, while those with undetectable ANCA (measured on at least 2 occasions) were regarded as ANCA negative.

Statistical analysis.

Allelic and genotypic frequencies were compared between patients and controls and between subgroups of patients, using a chi-square test with Yates' correction. Calculations of odds ratios and 95% confidence intervals for relative risks were performed, after application of Fisher's exact test if appropriate. Some authors have proposed that correction of P values is not necessary in the context of previously demonstrated disease association and functional influence of gene expression (33). Nevertheless, although associations of (AT)n in CTLA-4 with autoimmune/inflammatory diseases have been demonstrated by various independent groups and there is evidence indicating the functional relationships between this polymorphism and gene expression, we took a cautious approach to correction of the P values: P values were corrected (Pcorr) for the number of comparisons made on each locus. All P values were 2-sided. Pcorr values less than 0.05 were considered significant.


  1. Top of page
  2. Abstract

Association of Ctla4 with WG.

The Ctla4 (AT)n in the 3′-UTR is highly polymorphic, consisting of multiple alleles. Although the evolution of AT expansion is unknown, recent studies have demonstrated a correlation between higher numbers of AT repeats and 1) low stability and rapid degradation of CTLA-4 mRNA, 2) low levels of CTLA-4 protein expression, and 3) enhanced T cell activation in vivo due to impaired inhibitory costimulatory signals (14, 15, 18). The net effect would be low thresholds for T cell activation. We therefore divided the study subjects into 2 groups according to the alleles of (AT)n in Ctla4: individuals homozygous for the shortest allele, 86, and individuals carrying longer alleles (designated as alleles xx). Results in WG patients and matched healthy controls are displayed in Table 1. The percentage of individuals carrying longer alleles was significantly increased among patients with WG compared with controls. The frequency of patients homozygous for the shortest allele, 86, which is most effective in providing CTLA-4 function, was significantly reduced in the patient group (47.0%, versus 69.9% in controls; P = 0.0005, Pcorr = 0.001).

Table 1. (AT)n microsatellite polymorphisms in the 3′-untranslated region of Ctla4 in patients with Wegener's granulomatosis (n = 117) and healthy controls (n = 123)*
  • *

    Values are the number (%). 86/86 denotes homozygosity for the shortest allele that has polymerase chain reaction (PCR) products of 86 bp in length; xx/xx denotes the presence of at least 1 allele that has PCR products longer than 86 bp.

  • P = 0.0005, corrected P = 0.001, 95% confidence interval 0.2–0.6, versus controls.

86/8655 (47.0)86 (69.9)
xx/xx62 (53.0)37 (30.1)

There was no significant difference in genotypic and allelic frequencies of Ctla4 (AT)n between patients with generalized WG and those with limited WG or between male and female patients with WG. Sera from 114 patients with WG were assayed for ANCA during periods of clinically apparent active disease. Eighty-nine patients (78%) had detectable levels of circulating ANCA, while 25 were ANCA negative. There was no significant difference in the distribution of the genotypic and allelic frequencies of Ctla4 (AT)n between ANCA-positive and ANCA-negative WG patients. Eighty-four of the 89 ANCA-positive patients (94%) were positive for PR3 ANCA, and 7 (8%) were positive for MPO ANCA. Two of the ANCA-positive patients had both PR3 ANCA and MPO ANCA. Due to the small number of MPO ANCA–positive patients, we were unable to make meaningful comparisons between ANCA subsets and associations with genetic variation in Ctla4.

Lack of association between TNFα −308 RFLP and WG.

The genotypic and allelic distributions of the TNFα −308 RFLP were similar in patients and controls (Table 2). No relationship between the TNFα −308 RFLP and clinical variables (generalized versus limited WG) or ANCA titers was found.

Table 2. Frequencies of TNFα −308, TNFRI, TNFRII, IL-1β Taq I, and IL-6 −174 RFLP in patients with Wegener's granulomatosis and healthy controls*
  • *

    Values are the number (%). There were no statistically significant differences between the patient group and the control group for any of these parameters. TNFα = tumor necrosis factor α; TNFRI = TNF receptor I; IL-1β = interleukin-1β; RFLP = restriction fragment length polymorphism.

TNFα −308  
 A1/A190 (76.9)92 (74.8)
 A1/A225 (21.4)28 (22.8)
 A2/A22 (1.7)3 (2.4)
 A/A36 (30.8)40 (32.5)
 A/G62 (53.0)63 (51.2)
 G/G19 (16.2)20 (16.3)
 T/T59 (50.4)64 (52.0)
 T/G52 (44.4)53 (43.1)
 G/G6 (5.1)6 (4.9)
IL-1β Taq I  
 A1/A176 (65.0)83 (67.5)
 A1/A238 (32.5)39 (31.7)
 A2/A23 (2.6)1 (0.8)
IL-6 −174  
 C/C24 (20.5)26 (21.1)
 C/G64 (54.7)66 (53.7)
 G/G29 (24.8)31 (25.2)

Lack of association of polymorphisms in TNFα receptor genes.

Table 2 shows the genotypic frequencies of TNFRI exon I Msp AII and TNFRII exon 6 Nla III SNPs in patients with WG and healthy controls. No difference between patients and controls was observed. Stratification of patients by sex, ANCA status, and renal involvement did not reveal any correlation between these 2 SNPs and subsets of patients.

IL-1β Taq I RFLP and IL-6 −174 RFLP.

Polymorphisms located in the genes coding IL-1 and IL-6 were also investigated, and no differences between patient and control groups were found (Table 2). There was no relationship between either of these 2 polymorphisms and subgroups of patients stratified by sex, ANCA status, or renal involvement.


  1. Top of page
  2. Abstract

Compared with analyses that utilize genetic markers of the entire genome, candidate gene approaches, as applied in this study, select polymorphisms situated close to or within genes coding proteins that have functional significance in the pathogenesis of a disease. Candidate gene studies have the advantages of being guided by established knowledge of pathogenesis and having lower cost and greater efficiency. Therefore, despite the introduction of genetic map markers many years ago and the disadvantages of possible investigator bias and incomplete evaluation of what might be important unexplored genetic factors, the candidate gene approach remains a powerful, widely used strategy for the study of disease associations and pathogenesis.

The results of the current studies provide evidence of a genetic association between Ctla4 polymorphisms and WG in an American population. This work extends our previous findings of a positive association between longer alleles of (AT)n in the 3′-UTR of Ctla4 and WG (23). These results demonstrate that genetic variations in this T cell–related gene confer susceptibility to WG and that such influence is not limited to our previously reported WG patients from Sweden (23).

Since the discovery of Ctla4, numerous experiments from different groups using both murine and human systems have demonstrated a critical role of CTLA-4 in down-regulating T cell activation, which has a profound impact on inflammation and autoimmunity (1). There are multiple sites of immune dysfunction in WG, including excessive activation of macrophages, T cells, and B cells. CTLA-4 appears to play an important role in down-regulating immune responses. The underrepresentation of the shortest allele in Ctla4 in both Swedish and American patients with WG indicates that the overrepresented longer alleles of Ctla4 (AT)n might contribute to perpetuation of immune dysfunction, especially T cell activation, in WG (34–36).

TNFα is capable of priming neutrophils, leading to a preactivated state, oxidative burst, and neutrophil degranulation. Levels of TNFα in serum and expression and secretion of TNFα by peripheral blood mononuclear cells and T cells are increased in patients with active WG (35, 37, 38). TNFα concentrations are also increased in affected renal tissue (38). The polymorphism located at nucleotide position −308 of the promoter region has been studied extensively (39), and the majority of the published data from independent groups supports the notion that TNFα −308 allele 2 or its haplotype plays a direct role in the increased production of TNFα (20, 39–42).

Our previous preliminary results in Swedish patients did not support the idea that TNF2 plays a major role in the pathogenesis of WG (23). However, that study was limited by its small patient cohort. Moreover, disease associations with specific genes may vary in different ethnic populations. For example, different TNFα gene polymorphisms are associated with differences in susceptibility to severe malaria in East and West Africa (42), and interethnic differences in severity of disease are well recognized in systemic lupus erythematosus (43). In order to study the role of genetic determinants of TNFα in the pathogenesis of WG, we included analysis of the TNFα −308 RFLP in the current study. Results from both studies suggest that enhanced production of TNFα in WG does not derive from this genetic variation. It is likely that increased circulating and tissue levels of TNFα might be determined by posttranscriptional events in WG. Similar conclusions might be drawn with regard to IL-1β and IL-6 in WG (Table 2).

TNF exerts its wide spectrum of proinflammatory effects via 2 receptors, TNFRI and TNFRIII (TNFR p55 and TNFR p75). Both receptors can produce a soluble form of TNFR that binds to TNFα in vivo, limiting its proinflammatory functions. Improvement in scores on the WG modification of the Birmingham Vasculitis Activity Score (44) after 6 months of treatment in a clinical trial of etanercept, a fusion protein consisting of 2 molecules of TNFRII conjugated with the Fc portion of human IgG (45), underscores the significance of TNFα and TNF receptors in the disease. However, no association of polymorphisms in TNF receptor genes was found in the present study. Taken together, these findings indicate that in genetically predisposed patients with WG, e.g., those who are positive for Ctla4 (AT)n longer alleles, certain triggering factor(s), such as yet-unknown microorganisms, might elicit heightened T cell responses, while proinflammatory cytokines, such as TNFα and IL-1β, augment autoimmune/inflammatory injury pathways.


  1. Top of page
  2. Abstract