Bobby P.C. Koeleman

Unbiased Application of the Transmission/Disequilibrium Test to Multilocus Haplotypes

When the transmission/disequilibrium test (TDT) is applied to multilocus haplotypes, a bias may be introduced in some families for which both parents have the same heterozygous genotype at some locus. The bias occurs because haplotypes can only be deduced from certain offspring, with the result that the transmissions of the two parental haplotypes are not independent. We obtain an unbiased TDT for individual haplotypes by calculating the correct variance for the transmission count within a family, using information from multiple siblings if they are available. An existing correction for dependence between siblings in the presence of linkage is retained. To obtain an unbiased multihaplotype TDT, we must either count transmissions from one randomly chosen parent or count all transmissions and estimate the significance level empirically. Alternatively, we may use missing-data techniques to estimate uncertain haplotypes, but these methods are not robust to population stratification. An illustration using data from the insulin-gene region in type 1 diabetes shows that the validity and power of the TDT may vary by an order of magnitude, depending on the method of analysis.

Differential association of the PTPN22 coding variant with autoimmune diseases in a Dutch population.

Protein tyrosine phosphatase PTPN22 is involved in the negative regulation of T-cell responsiveness. Recently, the association of a coding variant of the PTPN22 gene-R620W(1858C>T) with a number of autoimmune diseases has been described. Therefore, we tested the association of PTPN22 1858*T allele in Dutch early onset type 1 diabetes (T1D) and rheumatoid arthritis (RA) patients, as well as celiac disease (CD) patients, for which no previous study of PTPN22 has been reported. The PTPN22 variant was strongly associated with T1D in cases vs controls (P=2 × 10−7, OR=2.3, 95% CI=1.7–3.1) as well as in a transmission disequilibrium test in nuclear trios (P=9 × 10−9, OR=3.3, CI=2.1–5.0), RA (case/control: P=0.003, OR=1.8 CI =1.2–2.6), but not CD, in spite of a trend of increased homozygosity (P=0.05) and early age at onset (P=0.01). PTPN22 is not generally associated with T-cell mediated autoimmune diseases, although it might play a role in the CD patients with early clinical manifestation.

Genotype effects and epistasis in type 1 diabetes and HLA-DQ trans dimer associations with disease

Alleles of HLA class II genes DQB1, DQA1, and DRB1 in the MHC region are major determinants of genetic predisposition to type 1 diabetes (T1D). Several alleles of each of these three loci are associated with susceptibility or protection from disease. In addition, relative risks for some DR-DQ genotypes are not simply the sum or product of the single haplotype relative risks. For example, the risk of the DRB1*03-DQB1*02/DRB1*0401-DQB1*0302 genotype is often found to be higher than for the individual DRB1*03-DQB1*02 and DRB1*0401-DQB1*0302 homozygous genotypes. It has been hypothesized that this synergy or epistasis occurs through formation of highly susceptible trans-encoded HLA-DQ(α1, β1) heterodimers. Here, we evaluated this hypothesis by estimating the disease associations of the range of DR-DQ genotypes and their inferred dimers in a large collection of nuclear families. We determined whether the risk of haplotypes in DRB1*0401-DQB1*0302-positive genotypes relative to the DRB1*03-DQB1*02-positive genotypes is different from that of DRB1*01-DQB1*0501, which we used as a baseline reference. Several haplotypes showed a different risk compared to DRB1*01-DQB1*0501, which correlated with their ability to form certain trans-encoded DQ dimers. This result provides new evidence for the potential importance of trans-encoded HLA DQ molecules in the determination of HLA-associated risk in T1D.

Adaptation of the extended transmission/disequilibrium test to distinguish disease associations of multiple loci: the Conditional Extended Transmission/Disequilibrium Test

Linkage and association studies in complex diseases are used to identify and fine map disease loci. The process of identifying the aetiological polymorphism, the molecular variant responsible for the linkage and association of the chromosome region with disease, is complicated by the low penetrance of the disease variant, the linkage disequilibrium between physically‐linked polymorphic markers flanking the disease variant, and the possibility that more than one polymorphism in the most associated region is aetiological. It is important to be able to detect additional disease determinants in a region containing a cluster of genes, such as the major histocompatibility complex (MHC) region on chromosome 6p21. Some methods have been developed for detection of additional variants, such as the Haplotype Method, Marker Association Segregation Chi‐squares (MASC) Method, and the Homozygous Parent Test. Here, the Extended Transmission/Disequilibrium Test is adapted to test for association conditional on a previously associated locus. This test is referred to as the Conditional Extended TDT (CETDT). We discuss the advantages of the CETDT compared to existing methods and, using simulated data, investigate the effect of polymorphism, inheritance, and linkage disequilibrium on the CETDT.

CTLA4 is differentially associated with autoimmune diseases in the Dutch population.

Cytotoxic T lymphocyte-associated antigen 4 (CTLA4) is an important negative regulator of T-cell response and its genetic association with type 1 diabetes (T1D) has recently been demonstrated. The frequent co-association of autoimmune diseases (AID) and the implication from multiple genome scans that the CTLA4 gene region is a general autoimmune region, led us to study the role of CTLA4 in independent cohorts of T1D, coeliac disease (CD) and rheumatoid arthritis (RA) patients. We present independent data that confirm the association of CTLA4 in Dutch patients with juvenile onset T1D and show differential association of CTLA4 with CD and RA. The CTLA4 gene polymorphisms were tested for association in 350 T1D, 310 CD, 520 RA patients and 900 controls. In addition, 218 families were tested by the transmission disequilibrium test (TDT). T1D patients showed the highest association with the MH30*G: −1147*C: +49*G: CT60*G: JO37_3*G (haplotype 2) in both a case/control cohort (P=0.002, OR=1.42) and by TDT (P=0.02, OR=1.43). In contrast, this haplotype showed no association in the RA and CD cohorts. However, we observed an increased frequency of the MH30*G: −1147*T: +49*A: CT60*G: JO37_3*A (haplotype 3) in the CD patients diagnosed at a young age (OR=1.6, P=0.026, Pc=0.052). Furthermore, when T1D and CD patients were stratified based on the HLA risk, the T1D susceptible CTLA4 haplotype 2 was over-represented in the high HLA-risk T1D and CD groups. In conclusion, we confirmed association between CTLA4 haplotype 2 and T1D in the Dutch population. Association with another CTLA4 haplotype (haplotype 3) was confirmed for CD, but only in those patients who had an early age of expression. No effect was found between RA and CTLA4. The association of the CTLA4 haplotype 2 with the high-risk HLA genotype in T1D and CD, which share DQ2 as the one of high-risk alleles, might provide a clue to understanding the common genetic background of AID.

Genetic variants of RANTES are associated with serum RANTES level and protection for type 1 diabetes

RANTES (regulated on activation, normal T-cell expressed and secreted) is a T-helper type 1 (Th1) chemokine that promotes T-cell activation and proliferation. RANTES is genetically associated with asthma, sarcoidosis and multiple sclerosis. The concentration of RANTES is increased at inflammation sites in different autoimmune diseases. Type 1 diabetes (T1D) is a Th1-mediated disease with complex genetic predisposition. We tested RANTES as a candidate gene for association with T1D using three single-nucleotide polymorphism (SNP) variants (rs4251719, rs2306630 and rs2107538) to capture haplotype information. The minor alleles of all SNPs were transmitted less frequently to T1D offspring (transmission rates 37.3% (P=0.002), 38.7% (P=0.007) and 41.0% (P=0.01)) and were less frequently present in patients compared to controls (P=0.009, 0.03 and 0.04, respectively). A similar protective effect was observed for the haplotype carrying three minor alleles (transmission disequilibrium test (TDT): P=0.003; odds ratio (OR)=0.55; confidence interval (CI): 0.37–0.83; case/control: P=0.03; OR=0.74; CI: 0.55–0.98). Both patients and controls carrying the protective haplotype express significantly lower serum levels of RANTES compared to non-carriers. Subsequently, we tested a cohort of 310 celiac disease patients, but failed to detect association. RANTES SNPs are significantly associated with RANTES serum concentration and development of T1D. The rs4251719*A–rs2306630*A–rs2107538*A haplotype associated with low RANTES production confers protection from T1D. Our data imply that RANTES is associated with T1D both genetically and functionally, and contributes to diabetes-prone Th1 cytokine profile.

Conditional ETDT analysis of the human leukocyte antigen region in type 1 diabetes

Several studies have indicated that additional genes in the major histocompatibility complex (MHC) region, other than the class II genes HLA‐DQB1 and ‐DRB1 (the IDDM1 locus), may contribute to susceptibility and resistance to type 1 diabetes. The relative magnitude of these non‐ DR/DQ effects is uncertain and their map location is unknown owing to the extraordinary linkage disequilibrium that extends over the 3.5 Mb of the MHC. The homozygous parent test has been proposed as a method for detection of additional risk factors conditional on HLA‐DQB1 and ‐DRB1. However, this method is inefficient since it uses only parents homozygous for the primary disease locus, the DQB1‐DRB1 haplotype. To overcome this limitation, Conditional ETDT was used in the present report to test for association conditional on the DQB1‐DRB1 haplotype, thereby allowing all parents to be included in the analysis. First, we confirm in UK and Sardinian type 1 diabetic families that allelic variation at HLA‐DRB1 has a very significant effect on the association of DQB1 and vice versa. The Conditional ETDT was then applied to the HLA TNF (tumour necrosis factor) region and microsatellite marker D6S273 region, both of which have been reported to contribute to IDDM1 independent of the HLA‐DQB1‐DRB1 genes. We found no evidence for a major role for either of these two regions in IDDM1.

Association of interferon-γ and interleukin 10 genotypes and serum levels with partial clinical remission in type 1 diabetes

We studied whether serum interferon (IFN)‐γ or interleukin (IL)‐10 levels and their corresponding functional polymorphic genotypes are associated with partial remission of type 1 diabetes (T1D). A multi‐centre study was undertaken in patients with newly diagnosed T1D and matched controls. T1D patients were followed for 3u2003months and characterized for remission status. Partial clinical remission was defined as a daily insulin doseu2003≤u20030·38 units/kg/24u2003h with an HbA1cu2003≤u20037·5%. Thirty‐three patients and 32 controls were phenotyped for serum concentrations of IFN‐γ and IL‐10 and genotyped for functional polymorphisms of the IFN‐γ and IL‐10 genes. Sixteen of 25 informative patients (63%) remitted. Serum IFN‐γ concentrations were significantly decreased in remitters but increased in non‐remitters compared to controls, and did not change over time in any group. IFN‐γ genotypes corresponded with serum levels in controls and non‐remitters, but not in remitters who displayed the lowest serum IFN‐γ levels despite more often carrying high‐producing IFN‐γ genotypes. Neither the frequency of IL‐10 genotypes nor serum IL‐10 concentration differed between patients and controls. The combination of high‐producing IFN‐γ genotype together with low serum IFN‐γ concentration at the time of diagnosis provided a strong positive predictive value for remission. Serum IFN‐γ concentrations predicted by genotype and observed serum levels were discordant in remitters, suggestive of regulation overruling genetic predisposition. Although high‐producing genotypes were less frequent in remitters, they were predictive of remission in combination with low serum IFN‐γ levels. These data imply that remission is partially immune‐mediated and involves regulation of IFN‐γ transcription.

Association analysis of myosin IXB and type 1 diabetes.

To date, seven studies have provided evidence for an association between the gene encoding for myosin IXB (MYO9B) and celiac disease (CD), and inflammatory bowel diseases, including single nucleotide polymorphisms (SNPs) rs2305767, rs1457092, and rs2305764. We investigated whether MYO9B is associated with T1D. The three SNPs were genotyped in Dutch samples from 288 T1D patients and 1615 controls. The A allele of SNP rs2305767A>G showed some evidence of association with T1D (nominal p for genotype = 0.06; OR carrier = 1.51, 95% CI = 1.04-2.19), but not in British samples from 4301 case patients and 4706 controls (p = 0.53), or when the Dutch and UK data were pooled (N patients = 4582, N controls= 6224; Mantel-Hansel p = 0.83). Furthermore, the nonsynonymous rs1545620 C>A SNP that has been associated with the inflammatory bowel disease, showed no association with T1D in British case-control set (p = 0.57). We conclude that MYO9B might not be a strong determinant of T1D, although there was some association in our initial Dutch study. Further studies are needed to evaluate the role of MYO9B in T1D.