Published online before print January 5, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Juvenile Diabetes Research Foundation/Wellcome Trust (JDRF/WT) Diabetes and Inflammation Laboratory, Cambridge Institute for Medical Research, University of Cambridge, Addenbrooke’s Hospital, Cambridge, UK
2 Correspondence: JDRF/WT Diabetes and Inflammation Laboratory, Cambridge Institute for Medical Research, University of Cambridge, Wellcome Trust/MRC Building, Addenbrooke’s Hospital, Cambridge CB2 0XY, UK. E-mail: john.todd{at}cimr.cam.ac.uk
ABSTRACT
Gene-gene interaction analyses have been suggested as a potential strategy to help identify common disease susceptibility genes. Recently, evidence of a statistical interaction between polymorphisms in two negative immunoregulatory genes, CBLB and CTLA4, has been reported in type 1 diabetes (T1D). This study, in 480 Danish families, reported an association between T1D and a synonymous coding SNP in exon 12 of the CBLB gene (rs3772534 G>A; minor allele frequency, MAF=0.24; derived relative risk, RR for G allele=1.78; P=0.046). Furthermore, evidence of a statistical interaction with the known T1D susceptibility-associated CTLA4 polymorphism rs3087243 (laboratory name CT60, G>A) was reported (P<0.0001), such that the CBLB SNP rs3772534 G allele was overtransmitted to offspring with the CTLA4 rs3087243 G/G genotype. We have, therefore, attempted to obtain additional support for this finding in both large family and case-control collections. In a primary analysis, no evidence for an association of the CBLB SNP rs3772534 with disease was found in either sample set (2162 parent-child trios, P=0.33; 3453 cases and 3655 controls, P=0.69). In the case-only statistical interaction analysis between rs3772534 and rs3087243, there was also no support for an effect (1994 T1D affected offspring, and 3215 cases, P=0.92). These data highlight the need for large, well-characterized populations, offering the possibility of obtaining additional support for initial observations owing to the low prior probability of identifying reproducible evidence of gene-gene interactions in the analysis of common disease-associated variants in human populations.
Key Words: CBLB • CTLA4 • Type 1 diabetes • rs3087243 • rs3772534
Diabetes in the KDP rat [a spontaneous animal model of human type 1 diabetes (T1D)] can be attributed to two unlinked loci. In addition to the MHC, another locus, Iddm/kdp1, was mapped to a nonsense mutation in Cblb (Casitas B-lineage lymphoma b, or Cas-Br-M murine ecotropic retroviral transforming sequence b) on rat chromosome 11, resulting in a truncated protein [1 ]. CBLB is known to function in the control of T cell activation [1 ]. Consequently, using a tag SNP approach [2 ], we previously tested for an association between T1D and the human ortholog CBLB on chromosome 3q11–q13.1 in 754 multiplex T1D families from the UK and USA. We found no statistical evidence for an association (multilocus test P=0.69 in 1416 parent-child trios) [3 ].
A recent publication from Bergholdt et al. [4 ] has also examined CBLB. They genotyped eight SNPs (all with a MAF greater than 0.01) in a panel of 253 T1D families and found that a synonymous coding SNP in exon 12 (rs3772534 G>A; MAF=0.03) showed some evidence of an association with T1D [26 transmissions of the G allele vs. 11 nontransmissions; derived relative risk (RR) for G allele=2.36, 95% confidence interval (CI)=1.17–4.48; P=0.03]. In an attempt to extend support for the result, rs3772534 was further genotyped in a second collection of 227 families, but no evidence of an association was found (15 transmissions of the G allele vs. 12 nontransmissions; derived RR for G allele=1.25, 95% CI=0.59–2.67, P =0.29) [4 ]. After combining the families, some evidence of an association remained (41 transmissions of the G allele vs. 23 nontransmissions; derived RR for G allele=1.78, 95% CI=1.07–2.97, P=0.046) [4 ].
The regulation of CBLB by the known T1D susceptibility gene CTLA4 on chromosome 2q33 makes the possibility of an interaction between these genes plausible and consequently, Bergholdt et al. [4 ] genotyped the CTLA4 SNP rs3087243 G>A (CT60; MAF not reported, but from previous studies is 0.46 in European populations) in their first collection (253 families) and further analyzed the CBLB SNP rs3772534, stratifying according to rs3087243 genotype. Affected individuals were divided into two groups; those homozygous for the rs3087243 susceptibility allele, G, in the first and G/A heterozygotes and A/A homozygotes in the second. They found that the CBLB SNP rs3772534 G allele was overtransmitted to offspring with the CTLA4 rs3087243 G/G genotype; the transmissions between the two rs3087243 genotype subgroups were significantly different, (P<0.0001). Unexpectedly, they did not try to reproduce this result in their second collection [4 ].
Previously, we had not directly tested for an association between T1D and the CBLB SNP rs3772534, because the SNP was well captured (R2=0.996) by the selected tag SNPs [3 ]. To exclude the suggested [4 ], but unlikely, possibility that we might have missed important information and also in an attempt to replicate in a larger collection, the reported interaction between rs3772534 and rs3087243, we have genotyped both SNPs in a large family collection (2162 parent-child trios from the UK, USA, Norway, and Romania). Consistent with our previous analysis, we found no evidence for an association between T1D and the CBLB SNP rs3772534 (RR for allele G=1.14 and P=0.33 in families; Table 1 ).
|
View this table: [in a new window] |
Table 1. Analysis of the CBLB SNP rs3772534 in 2162 Parent-Child Trios, 3453 Cases, and 3655 Controls
|
Bergholdt et al. [4
] only tested for a statistical interaction between CBLB rs3772534 and CTLA4 rs3087243 in their first collection of 253 multiplex/simplex Danish families. They did not attempt to reproduce the result in their second independent sample collection (227 simplex families). Interestingly, they found no evidence for an association between T1D and rs3087243 (P=0.18). This is probably due to a lack of sufficient power to detect the small effect of CTLA4 rs3087243 (G allele frequency=0.58 and RR for G allele=1.19 in simplex families [6
] ) in a collection the size of their first collection (an estimated 408 trios). At least 1100 parent-child trios would have to be genotyped in order to have 80% power to detect the rs3087243 RR for allele G of 1.19 at the type 1 error rate, 
= 0.05. In contrast, we found statistical evidence for an association between T1D and CTLA4 rs3087243 in both collections (families: P=6.0x10–4; RR=1.17, 95% CI=1.07–1.28; case-control: P=1.3x10–7; OR=1.21, 95% CI=1.13–1.30).
To test for a statistical interaction between the CBLB SNP rs3772534 and CTLA4 rs3087243, we performed a case-only locus-locus interaction analysis [5 ] with rs3087243 genotypes grouped as defined by Bergholdt et al. (G/G and A/G plus A/A) [4 ]. We were unable to find any evidence for an interaction, defined as deviation from the multiplicative model for the joint effects of rs3772534 and rs3087243 in cases and T1D offspring (1994 T1D offspring from 1367 families and 3215 cases: F1,3214=0.01 and P=0.92; Table 2 ). We also tested the previously reported set of tag SNPs [3 ] for an interaction between the CBLB region and CTLA4 rs3087243, again finding no statistical evidence to support an interaction (1148 T1D offspring from 663 families, F2, 662=1.23 and P=0.29).
|
View this table: [in a new window] |
Table 2. The CTLA4 SNP rs3087243 Genotype by the CBLB SNP rs3772534 Genotype in 3215 Cases and 1994 T1D Offspring
|
ACKNOWLEDGEMENTS
This work was funded by the Wellcome Trust and the Juvenile Diabetes Research Foundation International. We thank the Human and Diabetes UK for USA and UK multiplex families, respectively; Kjersti Rønningen and Dag Undlien for the collection of the Norwegian families; Cristian Guja and Constantin Ionescu-Tirgoviste for the collection of the Romanian families; David Savage, Chris Patterson, Dennis Carson and Peter Maxwell for Northern Irish samples; the British Society for Paediatric Endocrinology and Diabetes for the TID case samples; and David Strachan, Susan Ring, Wendy McArdle, Marcus Pembrey, and Richard Jones for DNA from the 1958 British Birth Cohort; and we acknowledge use of DNA from the 1958 British Birth Cohort collection, funded by the Medical Research Council grant G0000934 and Wellcome Trust grant 068545/Z/02. DNA samples were prepared by Tasneem Hassanali, Gillian Coleman, Sarah Field, Trupti Mistry, Kirsi Bourget, Sally Clayton, Matthew Hardy, Jennifer Keylock, Erna King, Pamela Lauder, Meeta Maisuria, William Meadows, Meera Sebastian, and Sarah Wood.
FOOTNOTES
1 Present address: Sulston Laboratory, The Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambs CB10 1SA, UK. 
Received September 19, 2006; revised December 4, 2006; accepted December 5, 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
