| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Endocrinology, Diabetes, and Bone Diseases, Mount Sinai School of Medicine (Y.B., E.S.C., R.V., T.F.D., Y.T.), New York, New York 10029; and Division of Statistical Genetics, Columbia University (D.A.G.), New York, New York 10032
Address all correspondence and requests for reprints to: Dr. Yaron Tomer, Division of Endocrinology, Diabetes, and Bone Diseases, Box 1055, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, New York 10029. E-mail: yaron.tomer{at}mssm.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A large number of GD cases are familial [reviewed by Tomer and Davies (5)], as evidenced by a sibling risk ratio of 11.6 (6). However, GD is also frequently seen as isolated cases with no evidence of familiality even in the extended family. Because there is a significant environmental component to the risk for GD, the observation of no family involvement in a high percentage of cases suggests that genetic influences may differ in familial and sporadic forms. Furthermore, finding increased risk due to specific genes in the familial group would lend support to the idea that those genes influence familial clustering in GD. To date, two immune regulatory genes have consistently been shown to be associated with GD, the human leukocyte antigen (HLA), and cytotoxic T lymphocyte antigen-4 (CTLA-4) genes. However, whether these genes contribute mostly to the familial form of GD or to both the sporadic and familial forms of GD is unclear.
HLA-DR3 was consistently found to be associated with GD (7, 8, 9, 10, 11). However, most linkage studies for HLA were negative (12, 13, 14, 15, 16, 17), albeit some workers did report linkage (18). CTLA-4 has been shown to be associated with GD (19, 20, 21, 22, 23, 24, 25, 26), but convincing evidence of linkage has been shown with thyroid autoantibodies (TAbs) (27, 28), and the relative risk (RR) is approximately 1.5 (29), suggesting that CTLA-4 is a minor contributing gene for familial GD.
Although GD is seen in both sporadic and familial forms, as noted above, no data have been reported to date comparing the influence of HLA and CTLA-4, or any other immune regulatory genes, on familial vs. sporadic GD. If HLA and CTLA-4 were major genes that contribute significantly to the familial clustering of GD, as suggested by others (10, 18), one would expect their influence to be more pronounced in familial GD than in sporadic GD. Therefore, the aim of our study was to examine the contributions of HLA and CTLA-4 to the familial clustering of GD by comparing the influence of these genes on the familial and sporadic forms of GD. The CTLA-4 gene region on chromosome 2q33 contains two additional immune regulatory genes that could potentially contribute to the familial clustering of GD, CD28, and inducible costimulator (ICOS). Even though recently it has been shown that CTLA-4 is most likely the GD susceptibility gene in this locus (28, 29), CD28 and ICOS could possibly be associated only with the familial form of GD. Therefore, we included CD28 and ICOS in our analysis of familial and sporadic GD patients.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The project was approved by the Mount Sinai School of Medicine institutional review board. All study participants singed an informed consent. We studied 160 Caucasian GD patients (69 GD probands from a dataset of GD families and 91 sporadic GD patients), and 150 age- and sex-matched healthy Caucasian controls. For the transmission disequilibrium test (TDT) analysis, we used the 69 families with familial GD. In all families both parents and the proband were available for analysis.
Clinical assessment. The diagnosis of GD was determined on the basis of documented clinical and biochemical evidence of past or present hyperthyroidism requiring treatment (antithyroid medications, radioactive iodine, or surgery) plus the presence of at least one of the following: positive TSH receptor antibodies, exophthalmos, and/or a diffuse thyroid scan. TSH receptor antibodies were measured by a radioreceptor assay (Kronus, Boise, ID).
Familial autoimmune thyroid disease (AITD). Individuals with familial AITD were defined as individuals with GD who had at least one other family member with GD or Hashimoto’s thyroiditis (HT) (17). All family members of these probands were examined by an endocrinologist, and the diagnosis was confirmed. There were a total of 69 familial AITD patients.
Familial GD (multiplex GD). It was possible that among the familial AITD group, the subset of GD probands who had relatives with GD was different from the subset of GD probands who had relatives with HT. To examine this possibility we also analyzed separately the group of GD probands who had first degree relatives with GD. This subset of familial GD (multiplex GD) included 48 individuals.
Sporadic GD. Sporadic GD was defined as GD in an individual who reported having no relatives with either GD or HT or any other thyroid problem or other autoimmune diseases including type 1 diabetes. There were a total of 91 sporadic GD patients.
Controls. One hundred and fifty age- and sex-matched healthy Caucasian volunteers served as controls in our studies.
DNA preparation
DNA was extracted from whole blood using the Puregene kit (Gentra Systems, Minneapolis, MN).
Human leukocyte antigen (HLA) typing
Molecular typing of HLA-DR was carried out according to the requirements of the American Society for Histocompatibility (30). Patients and controls were typed for HLA-DR using the technique of group-specific PCR amplification, followed by restriction enzyme digestion, as previously described (31).
CTLA-4 single nucleotide polymorphism (SNP) genotyping
CTLA-4 A/G49 SNP. We genotyped the A/G SNP in the leader peptide of CTLA-4 gene at position 49 in exon 1 (CTLA-4 A/G49 SNP), which was shown to be associated with GD (22, 23, 24, 32). This was performed using an automated, fluorescent-based, restriction fragment length polymorphism (RFLP) analysis as previously described (24). Briefly, DNA was amplified using the following primers: forward primer, GCTCTACTTCCTGAAGACCT; and reverse primer, AGTCTCACTCACCTTTGCAG. The forward primer was fluorescence labeled. PCRs were performed in a 20-µl reaction mixture containing 50 ng genomic DNA; 5 pmol of each primer (one of which was fluorescence labeled); PCR buffer containing 50 mmol/liter KCl; 10 mmol/liter Tris-HCl (pH 8.3); 1.5 mmol/liter MgCl2; 200 µmol/liter of each deoxy (d)-ATP, dGTP, dTTP, and dCTP; and 1 U AmpliTaq DNA polymerase (Applied Biosystems, Foster City, CA). Reaction mixtures were heated to 94 C for 7 min, then cycled 30 times as follows: 30 sec at 94 C, 30 sec at 55 C, and 30 sec at 72 C. Fluorescence-labeled PCR products were incubated at 37 C with the restriction enzyme BbvI for 2 h. The digested PCR product was diluted 1:25 in double distilled H2O, denatured, and separated on an ABI-310 genetic analyzer (Applied Biosystems). The two alleles were easily separated using this method; the A allele resulted in an undigested PCR product of 162 bp, and the G allele resulted in a digested PCR product with two fragments of 90 and 72 bp. Because the 90-bp fragment contained the fluorescence-labeled forward primer, it was visualized on the ABI-310, whereas the 72-bp fragment was not. Allele typing was performed using Genotyper 2.0 software.
CTLA-4 3'-untranslated region (3'UTR) CT60 SNP. The CT60 SNP was typed by the same fluorescent-based RFLP method (24) using the following primers: forward primer, ATAATGCTTCATGAGTCAGCTT; and reverse primer (fluorescent labeled), GAGGTGAAGAACCTGTGTTAAA. The PCR product was incubated with the restriction enzyme MaeII at 40 C for 5 h. The A allele resulted in a fluorescence-labeled undigested PCR product of 178 bp, and the G allele resulted in a digested PCR product with two fragments: a fluorescence-labeled 107-bp fragment and a nonlabeled 71-bp fragment.
CD28 SNP genotyping
The CD28 C/T SNP at position 17 of intron 3 (CD28 IVS3+17C/T SNP) was typed by the same fluorescent-based RFLP method (28) using the following primers: forward primer, (fluorescent labeled) CACAAGGAAGGAAATGCACT; and reverse primer, AAATAAACCACATAGGCAAA, and the restriction enzyme AciI. The T allele resulted in a fluorescence-labeled undigested PCR product of 280 bp, and the C allele resulted in a digested PCR product with two fragments: a fluorescence-labeled 208-bp fragment and a nonlabeled 72-bp fragment.
ICOS (GT)n microsatellite typing
We analyzed the ICOS gene using a guanine-thymine dinucleotide repeat polymorphism [(GT)n] in intron 4 of the ICOS gene (33). DNA was amplified using the following primers: forward primer, GGTGTTGAAGCATAAAGATG; and reverse primer, TCCCCTCTCCATTGCCTTTC. The forward primer was fluorescence labeled. PCR was performed in 15-µl reaction volume as previously described (28). The PCR products were separated on an ABI 310 genetic analyzer (Applied Biosystems), as described above. Allele calling was performed using Genotyper 2.0 software (Applied Biosystems).
Statistical analysis
All marker genotyping data were exported to a database (Oracle), where they were integrated with existing phenotype information and prepared for association analyses. Case-control association analyses were performed using 
2 and Fisher’s exact tests with Yates correction. The relative risk (RR) was calculated by the method of Woolf (34). P < 0.05 was considered significant. We grouped alleles with less than 10% frequency together for this analysis.
TDT
We performed the TDT using Genehunter version 2.0. The TDT compares the rate of transmission of parental alleles to affected offspring with the rate expected if there is no preferential transmission. A significantly increased transmission of a certain allele to affected offspring indicates association of that allele with the disease.
Linkage disequilibrium (LD) and haplotype analyses
LD and haplotype analyses were performed using SNPhap software (www_gene.cimr.cam.ac.uk/clayton/software/snphap.txt) with some modifications.
Power calculations
We used CDC simulation software (Epi Info, version 6.03; Centers for Disease Control, Atlanta, GA) to define the information limits of our experimental results. Our power calculations assumed n = 50 for the first analysis group (similar to the number of patients in our familial GD subset) and n = 90 for the second analysis group (similar to the number of patients in our sporadic GD subset). We assumed the population frequency of the susceptibility allele to be 15% for HLA-DR3 and 32% for CTLA-4 based on our data (Table 2
). Our power calculations indicated that a sample size of 50 familial and 90 sporadic GD patients would give our study 80% power to detect a difference between the groups resulting in a RR of more than 3.0 with an 
of 0.05 for HLA-DR3; for CTLA-4 we had 80% power to detect a difference resulting in a RR of more than 2.6 ( = 0.05).
|
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
HLA-DR.
HLA-DR typing was performed in 156 of our patients. The allele frequencies of the HLA-DR in all GD patients and controls are shown in Table 1. As expected, our results showed that DR3 was positively associated with GD (P = 9.0 x 10–7; RR = 3.8). Additionally, DR7 was negatively associated with GD, i.e. its frequency was increased in the controls compared with the GD patients (P = 0.01; RR = 0.4), suggesting that it had a protective effect. TDT analysis showed increased transmission of HLA-DR3 to GD-affected offspring (P = 0.03).
|
. As previously reported (22, 23, 24, 29, 32), GD was associated with the G alleles of the CTLA-4 A/G49 and the CT60 SNPs. Haplotype analysis showed an increased frequency of the A/G49 SNP-CT60 GG haplotype in patients vs. controls (34.6% and 26.6%, respectively; 2 = 9.42; P < 0.01). TDT analysis showed increased transmission of the G allele of the A/G49 SNP to GD-affected offspring, but it did not reach statistical significance (P = 0.2), most likely due to the small number of families (n = 69).
CD28.
There was no significant association between GD and CD28 IVS3+17C/T SNP (Table 2). The C allele was present in 22.4% of patients and 21.2% of controls.
ICOS.
Seven alleles were observed for the ICOS (GT)n microsatellite marker with sizes ranging from 100–114 bp. There was no statistically significant difference between the allele frequencies in the patients and controls (Table 2).
Comparison of the effects of immune regulatory genes in the familial AITD, familial GD, and sporadic GD subsets
HLA-DR.
The frequencies of HLA-DR alleles in familial and sporadic GD patients are shown in Table 1. The HLA-DR3 allele was present in 38.5% of the familial AITD patients and 40.7% of the sporadic GD patients (P = 0.78). The HLA-DR7 allele was present in 7.7% of the familial AITD patients and 7.7% of the sporadic GD patients (P = 1.0). The frequencies of these two alleles were also not significantly different in the familial GD (multiplex GD) group (35.4% and 6.3%, respectively). Thus, there was no significant difference between frequencies of the associated HLA-DR alleles in the familial and sporadic GD patients.
CTLA-4.
The frequencies of the CTLA-4 A/G49 and CT60 SNP alleles in familial and sporadic GD patients are shown in Table 2. The G allele of A/G49 was present in 39.1% of the familial AITD patients, 38% of the familial GD patients, and 43.2% of the sporadic GD patients (P = 0.46). Thus, there was no significant difference between the frequency of the G allele of the CTLA-4 A/G49 SNP in the familial and sporadic forms of GD. Similarly, there was no statistically significant difference among the frequency of the G allele of the CT60 SNP in the familial AITD patients (64.4%), the familial GD patients (66.7%), and the sporadic GD patients (62.8%).
CD28 and ICOS.
As shown in Table 2, there was no significant difference between the allele frequencies of the CD28 IVS3+17C/T SNP and the ICOS (GT)n microsatellite marker in the familial AITD, familial GD, and the sporadic GD patients.
Interaction between HLA-DR3 and the CTLA-4 A/G49 SNP
We also analyzed the joint risk between HLA-DR3 and the CTLA-4 A/G49 SNP. Case-control association analysis of the HLA-DR3 in combination with the CTLA-4 A/G49 G allele showed evidence for additive influence on disease expression of both HLA-DR3 and the CTLA-4 A/G49 SNP, giving an RR of 5.9 (Table 3). However, there was no evidence that the interaction of HLA-DR3 and CTLA-4 A/G49 SNP in GD risk resulted in a stronger association with the familial or sporadic form of GD (Table 3).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
It is well known that GD is genetically influenced, but it also has a strong environmental component, as can be inferred from the approximately 30% penetrance derived from twin studies (35, 36) and family linkage studies (17). It might also be assumed that when a marker is associated and linked with a disease, it will show a stronger association in probands from families with the disease than in sporadic patients on the presumption that the sporadic patients are more likely to be affected by environmental factors than the familial cases. However, our study demonstrated lack of a more powerful genetic influence of HLA-DR3 and CTLA-4 in the familial form of GD, and this result, combined with the low RRs conferred by these genes, continue to support the idea that HLA-DR3 and CTLA-4 have a modulatory role in GD.
HLA class II molecules (DR, DQ, and DP) play a key role in the immune response by binding peptide antigens and presenting them to T cell receptors (37). Caucasian GD patients have been consistently shown to have an increased frequency of HLA-DR3 (7, 10, 11). The RR for people with DR3 has been shown to be approximately 2–4 (reviewed in Ref.5). Recently, we have shown that the association of HLA-DR3 with GD was largely determined by the presence of arginine at position 74 of the DR ß1-chain (38). However, linkage studies of HLA in GD have been largely negative (13, 14, 15, 31), suggesting that HLA-DR3 conferred a generalized increased risk for GD, but did not have a major contribution to the familial clustering of the disease (31). The results of the current study extend the previous findings, showing that HLA-DR3 was associated with both the sporadic and familial forms of GD, giving similar RRs, and thus again suggested that HLA-DR3 conferred a generalized risk for GD and was not linked with GD in families.
The T cell-specific cell surface receptors CD28 and CTLA-4 are important regulators of T cell activation by antigen-presenting cells (39). CD28 enhances T cell functions essential for effective antigen-specific immune responses, whereas CTLA-4 counterbalances the CD28-mediated signals and thus prevents overstimulation of the immune system (39). Another member of this family of costimulatory molecules, ICOS, was identified by Hutloff et al. (40). Unlike the constitutively expressed CD28, ICOS is induced on the T cell surface and does not up-regulate the production of IL-2, but induces the synthesis of IL-4 (41). There have been many reports demonstrating an association between the CTLA-4 gene and the AITDs (reviewed in Ref.5) as well as other autoimmune diseases (22, 42). This association has been consistent across populations of different ethnic backgrounds, such as Caucasians (19, 20, 22, 23, 24), Chinese (21), Japanese (25, 43), and Koreans (44). However, the RR conferred by CTLA-4 is modest (29), suggesting that CTLA-4 is a minor modulating gene for GD. Our results support this hypothesis, because we did not see an increased effect of CTLA-4 in familial GD. Thus, most likely, CTLA-4 causes a modest generalized increased risk for GD as well as for other autoimmune diseases.
We have previously shown that the CTLA-4 gene region was linked in some families with the propensity to secrete TAbs; therefore, we examined whether the influence of the CTLA-4 gene was stronger in these linked families. Indeed, our results showed a higher frequency of the G allele of the A/G49 SNP in probands of the linked families compared with probands of unlinked families and sporadic GD patients (27). Therefore, we concluded that CTLA-4 conferred a general modest increased risk for GD (as evidenced by the similar and low RRs in sporadic and familial GD), but that there existed a subgroup of families in which it had a major effect, as demonstrated by the linkage with TAbs and the higher RR.
Interestingly, CD28, CTLA-4, and ICOS form a gene cluster on a 300-kb contig on chromosome 2q33 (33) in man and chromosome 1 in mice (29). Because CTLA-4, CD28, and ICOS form a gene cluster, the association of autoimmune diseases with the CTLA-4 locus could represent the effects of any of these three genes alone or in combination due to linkage disequilibrium. Recently, we (28) and others (29) were able to show that CTLA-4 and not CD28 or ICOS was the susceptibility gene at this locus. However, these studies were performed in nonselected GD patients, and it was possible that in the subset of familial GD patients CD28 and/or ICOS may play a role in the genetic susceptibility. Our results demonstrated that this was not the case. Even in the subset of familial GD patients, neither CD28 nor ICOS was associated with GD, demonstrating again that they most likely do not play a role in the genetic susceptibility to GD. At present it is unknown which CTLA-4 polymorphism is the causative polymorphism. Recent data have suggested that the causative SNP is located in the 3'UTR of the gene (29); however, it cannot be conclusively determined whether the 3'UTR CT60 SNP or the 3'UTR AT microsatellite is the causative polymorphism, because these two sequence variants are 816 bp apart and are in tight linkage disequilibrium (29). Our analysis of the 3'UTR CT60 SNP has shown an association with GD, giving a similar RR as conferred by the A/G49 SNP.
Additional explanations for our findings are also possible. For example, the existence of significant genetic heterogeneity in the familial cases could decrease the association signal in the familial cases. If genetic heterogeneity exists in the familial GD patients, only a subset of the familial cases is influenced by the associated alleles at the tested marker locus, whereas the remaining familial cases are influenced by other genes. Therefore, when the whole familial group is analyzed together, the strength of the association signal decreases to a level similar to that seen in the sporadic cases. Because genetic heterogeneity has been shown to occur in GD (17, 45), this may be the reason why both HLA and CTLA-4 were associated with GD with the same RR in sporadic and familial cases. Indeed, as previously mentioned, the association of the CTLA-4 A/G49 SNP was found to be stronger in probands from linked families, demonstrating that CTLA-4 exerted a major effect in these linked families, but not in others (27). Another possible explanation is that HLA-DR3 and CTLA-4 conferred a general susceptibility to GD, whereas other genes may influence the penetrance of the disease. According to this hypothesis, in a family in which the HLA/CTLA-4 susceptibility alleles were inherited along with a gene(s) increasing GD penetrance, several individuals would develop GD (familial GD). In contrast, in families in which only the HLA/CTLA-4 susceptibility alleles were inherited, but the penetrance was low, few or only one individual would develop GD (sporadic GD). In addition, shared environmental exposures (e.g. low iodine diet) may contribute to the familial clustering of GD and, therefore, could further complicate the comparison of familial and sporadic GD. Alternatively, we cannot exclude the possibility that our dataset was not large enough to detect minor differences between the frequencies of DR3 and the CTLA-4 A/G49 SNP in familial and sporadic GD. Our power calculations showed that using our dataset we had 80% power to reject a difference between the groups with a RR of 2.6 or more for CTLA-4 and 3 or more for HLA-DR3. Hence, if there are differences between familial and sporadic GD not detected by our dataset, they are relatively small and of minor significance. Larger datasets may be needed to detect such small differences in the frequencies of alleles between sporadic and familial GD.
GD is most likely caused by many common genes with small effects (5). According to this model, inheriting a combination of these genes increases susceptibility to GD. Indeed, our results showed evidence for an interaction between HLA-DR3 and the CTLA-4 A/G49 SNP, giving a combined RR of 5.9. Thus, inheriting the susceptibility alleles for HLA-DR and CTLA-4 resulted in a stronger risk for GD than inheriting each one alone.
In conclusion, both familial and sporadic GD were associated with HLA-DR3 and CTLA-4 with the same RRs. These results suggest that HLA-DR3 and CTLA-4 conferred a general increased risk for GD in both the sporadic and familial forms, whereas additional genes may contribute to the penetrance of the disease in families and/or to the familial clustering of the disease.
|
Footnotes |
|---|
This work was presented in part at the 75th Annual Meeting of the American Thyroid Association, September 2003, Palm Beach, FL.
Abbreviations: AITD, Autoimmune thyroid disease; CTLA-4, cytotoxic T lymphocyte antigen-4; GD, Graves’ disease; HLA, human leukocyte antigen; HT, Hashimoto’s thyroiditis; ICOS, inducible costimulator; LD, linkage disequilibrium; RFLP, restriction fragment length polymorphism; RR, relative risk; SNP, single nucleotide polymorphism; TAb, thyroid autoantibody; TDT, transmission disequilibrium test; UTR, untranslated region.
Received September 30, 2003.
Accepted June 14, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  F. K. Kavvoura, T. Akamizu, T. Awata, Y. Ban, D. A. Chistiakov, I. Frydecka, A. Ghaderi, S. C. Gough, Y. Hiromatsu, R. Ploski, et al. Cytotoxic T-Lymphocyte Associated Antigen 4 Gene Polymorphisms and Autoimmune Thyroid Disease: A Meta-Analysis J. Clin. Endocrinol. Metab., August 1, 2007; 92(8): 3162 - 3170. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Vigano, D. Lattuada, E. Somigliana, A. Abbiati, M. Candiani, and A. M. Di Blasio Variants of the CTLA4 gene that segregate with autoimmune diseases are not associated with endometriosis Mol. Hum. Reprod., October 1, 2005; 11(10): 745 - 749. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. I. Shin, J. M. Park, J. S. Lee, D. H. Kim, and H. J. Jeong Development of Graves' disease during cyclosporin treatment for severe Henoch-Schonlein nephritis Nephrol. Dial. Transplant., September 1, 2005; 20(9): 2014 - 2015. [Full Text] [PDF]
|
|
|
|
|
|
B. Golden, L. Levin, Y. Ban, E. Concepcion, D. A. Greenberg, and Y. Tomer Genetic Analysis of Families with Autoimmune Diabetes and Thyroiditis: Evidence for Common and Unique Genes J. Clin. Endocrinol. Metab., August 1, 2005; 90(8): 4904 - 4911. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-C. Weng, M.-J. Wu, and W.-S. Lin CT60 Single Nucleotide Polymorphism of the CTLA-4 Gene Is Associated with Susceptibility to Graves' Disease in the Taiwanese Population Ann. Clin. Lab. Sci., January 1, 2005; 35(3): 259 - 264. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
