Cytotoxic T lymphocyte–associated antigen 4 (CTLA-4) protein is a key negative regulator of the T cell immune response and is expressed predominantly in activated and regulatory T cells. Distinct functions of CTLA-4 include the following: setting the threshold for T cell activation, thereby contributing to the maintenance of peripheral tolerance (1); suppression of T cell proliferation and inflammatory cytokine production; and induction of apoptosis in activated T cells (2). Alterations in the expression or function of CTLA-4 are therefore likely to influence susceptibility to autoimmune diseases.
More than 100 single-nucleotide polymorphisms (SNPs) have been identified in the CTLA4 gene region (3), which maps to chromosome 2q33 (Figure 1A). However, only 1 coding-region SNP, the +49AG polymorphism in exon 1 resulting in a threonine-to-alanine conversion at codon 17, has been identified. Many autoimmune disease association studies of individual CTLA4 SNPs have yielded mixed results (4–7), although there is consistent evidence for an association with Graves' disease, autoimmune hypothyroidism, and type 1 diabetes mellitus. A recent study (3) that screened all known SNPs in the CTLA4 region demonstrated that the linkage with Graves' disease, autoimmune hypothyroidism, and type 1 diabetes mellitus was strongest with the G allele of the noncoding CT60A/G polymorphism, which was correlated with lower messenger RNA (mRNA) levels of a soluble, alternatively spliced form of CTLA4.
Figure 1. A, The CTLA4 locus is located at chromosome 2q33 and lies between the CD28 and inducible costimulator (ICOS) loci. The direction of transcription is from left to right, as shown by the arrow. The solid boxes represent the 4 exons, and the 5′- and 3′-untranslated regions are depicted by open boxes. The relative locations (not to scale) of some of the previously studied CTLA4 single-nucleotide polymorphisms (SNPs) in relation to the +49 and CT60 SNPs are shown. The +49 SNP in exon 1 is the only coding-region SNP identified, and the CT60 SNP is adjacent to the 3′-untranslated region, close to the (AT)n microsatellite repeat. B, Strong linkage disequilibrium across 8 of these SNPs, resulting in 3 common haplotypes, has been identified in Scandinavian Caucasians (13) and European Caucasians (14). These haplotypes can be identified or “tagged” by the +49;CT60 haplotype combinations identified in the present study.
Download figure to PowerPoint
Primary Sjögren's syndrome (SS) is a common systemic autoimmune inflammatory condition that is associated with exocrine failure of the salivary and lacrimal (tear) glands and predominantly affects women. Characteristic features of primary SS are severe dryness of the eyes and mouth (sicca) and a variety of extraglandular and systemic manifestations. Anti-Ro/SSA and/or anti-La/SSB autoantibodies are present in up to 80% of patient sera and are markers for both disease and disease severity. The etiology of SS is multifactorial, and the genetic and environmental risk factors are far from being elucidated (8). Studies performed to date have failed to identify any confirmed genetic associations of major significance, other than an association with HLA class II genes. Previously, we demonstrated that the degree of diversification of the anti-Ro/La autoantibody response is influenced by distinct HLA haplotypes (9).
The current study was undertaken to investigate genetic associations of the CTLA4 CT60 and +49G/A SNPs in Australian Caucasian patients with primary SS, and to determine CTLA4 haplotypes and their associations with susceptibility to primary SS and/or extraglandular and serologic features of the disease.
- Top of page
- PATIENTS AND METHODS
The CTLA4 +49 and CT60 SNPs were in strong linkage disequilibrium, and only 3 haplotypes were identified in both patients with primary SS and controls: +49A;CT60A, +49A;CT60G, and +49G;CT60G. These are identical to the +49;CT60 haplotypes observed in Spanish Caucasians (12) and effectively define the 3 extended haplotypes (Figure 1B) observed in Scandinavian and European Caucasians (13, 14).
The individual +49 and CT60 SNPs and their haplotypes were in Hardy-Weinberg equilibrium in both patients with primary SS and controls. Both allele and haplotype frequencies were therefore analyzed for both SNPs, but the haplotype analysis resulted in a more consistent interpretation of the data. There were significant differences in the frequency of the CTLA4 haplotype between patients and controls (P = 0.032) (Table 1). Relative to the +49G;CT60G haplotype, the frequency of both the +49A;CT60A haplotype (OR 1.53, 95% confidence interval [95% CI] 1.04–2.25) and the +49A;CT60G haplotype (OR 1.78, 95% CI 1.08–2.93) was increased in patients with primary SS compared with controls.
Table 1. CTLA4 +49A/G and CT60 allele and haplotype frequencies in patients with primary Sjögren's syndrome and controls*
| ||Controls (n = 156)||Patients (n = 111)||OR (95% CI)|
The CTLA4 haplotype frequencies in Ro/La autoantibody–positive versus Ro/La autoantibody–negative patients with primary SS were suggestive of some differences that did not reach statistical significance (P = 0.11), but the power was low because only 17 Ro/La-negative patients were available for analysis (Table 2). The frequency of the +49A;CT60G haplotype was increased in Ro/La-positive patients with primary SS compared with controls (OR 2.00, 95% CI 1.17–3.41) (Table 2), but not in Ro/La-negative patients compared with controls (OR 0.76, 95% CI 0.20–2.97) (Table 2). In contrast, the frequency of the +49A;CT60A haplotype was relatively increased in both autoantibody subgroups, but relatively more so in the Ro/La autoantibody–negative subgroup (Table 2). Therefore, although the frequency of both of these CTLA4 haplotypes was increased in SS, there were differences between their associations in terms of the Ro/La autoantibody response. There was no evidence of interaction between CTLA4 genotypes and HLA–DR2 or DR3 (both of which are associated with Ro/La autoantibodies in primary SS) (9), or any association with increasing diversification of the Ro/La response, IgG levels, rheumatoid factor titers, sex, or age at disease onset.
Table 2. CTLA4 +49G and CT60 allele and haplotype frequencies in Ro/La autoantibody–negative and autoantibody-positive patients with primary Sjögren's syndrome
| ||Ro/La status||OR (95% CI)*|
|Negative (n = 17)||Positive (n = 94)||Ro/La negative vs. controls||Ro/La positive vs. controls|
|+49A||0.706||0.691||1.66 (0.78–3.55)||1.52 (1.05–2.22)†|
|CT60G||0.382||0.537||0.49 (0.24–1.00)†||0.90 (0.63–1.27)|
|+49A;CT60A||0.618||0.463||1.89 (0.88–4.05)||1.40 (0.94–2.08)|
|+49A;CT60G||0.088||0.229||0.76 (0.20–2.97)||2.00 (1.17–3.41)†|
There were significant differences in the CTLA4 haplotype frequencies between patients with primary SS with and those without Raynaud's phenomenon (P = 0.046) (Table 3). The +49A;CT60A haplotype was protective against Raynaud's phenomenon (OR 0.49, 95% CI 0.27–0.91) (Table 3). There was, however, no association between Raynaud's phenomenon and Ro/La autoantibody status.
Table 3. Allele and haplotype frequencies showing that the CTLA4 +49A;CT60A haplotype is protective against RP in primary Sjögren's syndrome*
| ||RP negative (n = 64)||RP positive (n = 42)||OR (95% CI)|
We recently demonstrated that daytime sleepiness and urologic symptoms are more severe in patients with primary SS (11), and data regarding these symptoms were available for 63 of the 111 patients with primary SS who were typed for CTLA4 SNPs. The ESS measures daytime sleepiness and is a comprehensively validated, simple, self-administered questionnaire that addresses the likelihood of falling asleep in a variety of situations (15). The dose of the 49A;CT60G haplotype, which is associated with Ro/La autoantibody–positive primary SS, influenced the severity of daytime sleepiness in patients with primary SS (P = 0.036) (Figure 2), although there was no overall difference in the ESS score between Ro/La autoantibody–positive and Ro/La autoantibody–negative patients with primary SS. There were no observed associations between CTLA4 haplotypes and urologic symptom or fatigue severity within the primary SS cohort.
Figure 2. The Epworth Sleepiness Scale (ESS), a measure of daytime sleepiness. ESS scores have previously been shown to be more severe in patients with primary Sjögren's syndrome compared with controls with osteoarthritis (11). In 63 patients for whom data were available, the CTLA4 +49A;CT60G haplotype dose was significantly correlated with the ESS score (P = 0.036 [gamma correlation]). Values are the mean ± SEM.
Download figure to PowerPoint
- Top of page
- PATIENTS AND METHODS
The CTLA4 region is highly polymorphic, and the majority of previous studies addressing CTLA4 and autoimmune disease have focused on analysis of individual SNPs, with somewhat conflicting results. We and other investigators (12, 14, 16) have identified 3 common CTLA4 haplotypes. In the current study, we were able to reduce the analysis of haplotypes to 2 SNPs, identified by +49A;CT60A, +49A;CT6G, and +49G;CT60G SNP combinations (Figure 1B). Differences in CTLA4 function are most likely to be associated with the haplotype rather than with individual SNPs. Because individual SNPs may occur on multiple haplotypes, we believe that individual SNP analysis may result in failure to detect an association or inconsistent findings between studies, particularly if the haplotype frequencies vary. For example, we observed a decrease in the frequency of the +49G;CT60G haplotype in patients with primary SS compared with controls and a relative increase in the frequency of both the +49A;CT60A haplotype and the +49A;CT60G haplotype (primarily in Ro/La-positive patients). Therefore, a haplotype carrying the CT60G allele, previously associated with autoimmune disease (3), is clearly implicated in primary SS, yet analysis of the individual CT60 SNPs detected no association (Tables 1 and 2).
There is considerable clinical and serologic overlap between primary SS and systemic lupus erythematosus (SLE) (17), and the frequency of the +49A;CT60G haplotype, which was associated with Ro/La-positive primary SS in this study, was also increased in Spanish patients with SLE (12). This same haplotype is also associated with celiac disease (13), for which patients with primary SS are at increased risk (18). An increased prevalence of lymphoma in primary SS is well recognized (19), and, interestingly, both of the primary SS CTLA4 susceptibility haplotypes carry the +49A allele, which is associated with non-Hodgkin's lymphoma (20). Further study of the effect of CTLA4 haplotypes on the risk of lymphoma in patients with primary SS is clearly warranted.
The CTLA4 haplotypes associated with the systemic rheumatic diseases primary SS and SLE are probably different from the haplotypes associated with organ-specific autoimmune diseases. In Graves' disease, autoimmune hypothyroidism, and type 1 diabetes mellitus, associations with both CT60G and +49G SNPs were observed, with the association with CT60G being statistically stronger (3). This implies an association with the +49G;CT60G haplotype, which was protective against primary SS in the present study. Unlike these organ-specific autoimmune diseases, both primary SS and SLE are associated with systemic B cell activation, polyclonal hypergammaglobulinemia, antinuclear autoantibodies, and multisystem involvement.
CTLA4 exerts distinct independent effects during different phases of T cell responses, including setting the threshold for T cell activation, suppression of T cell proliferation, and induction of apoptosis in already-activated T cells (2). It may even contribute directly to the regulation of B cell responses, because B cells express CTLA4 after cell–cell contact with activated T cells (21). A recent study in patients with SLE demonstrated that CTLA4 genotypes influenced immune responsiveness to Epstein-Barr virus (EBV) (22). EBV has long been hypothesized to play a role in the etiology of primary SS and SLE, and there is evidence that the anti-Ro autoantibody response, which is common in these patients, may be a result of molecular mimicry between Epstein-Barr nuclear antigen 1 and the 60-kd Ro antigen (23). In this context, it is interesting that in the present study, individual CTLA4 haplotypes influenced the expression of Ro/La autoantibodies in primary SS.
These demonstrated differential effects of CTLA4 polymorphisms on the risk of organ-specific versus systemic autoimmune diseases imply differing pathways of autoimmune pathogenesis resulting from altered CTLA4 expression and/or regulation. Therefore, an understanding of the functional effects of the CTLA4 haplotypes may provide key insights into the pathogenesis of autoimmunity. Unfortunately, existing functional studies are difficult to interpret, because they have focused on individual SNPs and have not examined the combined functional effects of the haplotypes on which they occur. The +49G SNP may be associated with incomplete CTLA4 glycosylation, lower cell surface expression (24), and reduced control of T cell proliferation in vitro (25). The CT60G SNP is implicated in lower mRNA levels of the soluble alternatively spliced form of CTLA4 (3), although this was not confirmed in a more recent study (26). Binding of soluble CTLA4 to B7 ligands on antigen-presenting cells may be an important mechanism for blocking co-stimulation of T cell activation, and is the rationale behind the therapeutic use of CTLA-4Ig in human autoimmune diseases such as rheumatoid arthritis. CTLA-4Ig has demonstrated efficacy in a variety of animal models of autoimmunity (27); however, this may be an oversimplification, because transient CTLA-4Ig blockade has also been shown to enhance antigen-specific T cell responses against murine tumors (28).
The severity of infiltration of immune cells in primary SS salivary gland biopsy specimens is associated with both increased titers of Ro/La autoantibodies and an increased frequency of extraglandular manifestations (11). Although within-patient subgroup comparisons are hypothesis-generating rather than definitive, we observed some potentially interesting associations. The CTLA4 +49A;CT60A haplotype was protective against the development of Raynaud's phenomenon in patients with primary SS. Raynaud's phenomenon is associated with a variety of autoimmune rheumatic diseases, but the associated risk factors and pathogenesis are not well understood. A previous study linked interleukin-1 (IL-1) haplotypes with susceptibility to Raynaud's phenomenon in primary SS (29). Genetic associations with both CTLA4 and IL-1 thus implicate immune mechanisms in the development of Raynaud's phenomenon, although in our cohort Raynaud's phenomenon was not associated with other indicators of the severity of autoimmune disease, such as Ro/La autoantibodies.
We previously showed that the severity of daytime sleepiness (based on the ESS score) is significantly increased in patients with primary SS (11). In the present study, we observed that the +49A;CT60G haplotype was associated with predominantly Ro/La-positive primary SS, and also correlated with daytime sleepiness in a dose-dependent manner. Our working hypotheses to explain increased daytime sleepiness in patients with primary SS include upper airway collapse due to altered surface tension resulting from dry airways with consequent obstructive sleep apnea, or, alternatively, central nervous system circadian dysfunction mediated by antimuscarinic receptor autoantibodies (30). In addition, evidence is now accumulating for a direct relationship between sleep and the immune system (31), and the CTLA4 haplotype association with more severe daytime sleepiness observed in this study is further evidence of this link. We are currently evaluating the nature of the sleep disturbance in primary SS, as well as the role of CTLA4 haplotypes in patients with sleep apnea in the absence of autoimmune disease.
In conclusion, this study has demonstrated that CTLA4 haplotypes are associated with susceptibility to primary SS, as well as with autoantibody production and some extraglandular features in patients with primary SS. The +49A;CT60G haplotype is associated with susceptibility to primary SS, Ro/La autoantibodies, and increased daytime sleepiness. The +49A;CT60A haplotype is associated with susceptibility to primary SS and protection against Raynaud's phenomenon. This study is consistent with other studies implicating CTLA4 as an important susceptibility locus in autoimmune disorders; however, the associations observed for systemic rheumatic autoimmune diseases such as primary SS and SLE are different from those reported for organ-specific diseases such as Graves' disease, autoimmune hypothyroidism, and type 1 diabetes mellitus, which may imply separate mechanisms for the pathogenesis of autoimmunity.