Introduction

Among four common HLA-DR2 haplotypes observed in Caucasians, the DQA1*0102, DQB1*0602, DRB1*15 haplotype is negatively associated with type 1diabetes and is extremely rare among patients in most populations studied (reviewed in [1]). The available evidence suggests that the diabetes-protective effect associated with DR2 haplotypes may mostly map to the HLA-DQ locus and in particular to the DQB1*0602 allele. Protection appears to be dominant since DQB1*0602 protects from diabetes even in the presence of high-risk HLA alleles. It also confers dominant protection in autoantibody-positive first-degree relatives of affected individuals, who usually express fewer autoantibodies than relatives without DQB1*0602 [2]. Although diabetic patients carrying DRB1*15, DQB1*0602 are extremely rare, their existence indicates that the protective effect of this haplotype is not absolute. Sequence analysis of rare patients and autoantibody-positive first-degree relatives with DQB1*0602 have demonstrated that they carry normal alleles lacking any mutations that would affect the peptide binding site [2, 3].

In a cohort of patients from Sweden, a country with one of the highest incidences of type 1 diabetes, the DQA1*0102, DQB1*0602, DRB1*15 haplotype was found in 1% of patients. Most of these patients (87.5%) belonged to an older age group (14–34 years old) [3, 4], suggesting that the protection associated with this haplotype may attenuate with age in this population. Many studies have indicated that genes in the HLA region other than DR and DQ also contribute to diabetes susceptibility. It is possible that rare patients with DQB1*0602 carry a haplotype in which the protective effect of DQB1*0602 is overridden by another gene. If this is the case, it should be possible to identify markers that identify DQB1*0602 haplotypes with different associations with diabetes risk.

As part of the activities of the 13th International Histocompatibility Working Group we genotyped, in the Swedish case-control cohort described above, the DRB1 and DQB1 alleles, as well as eight microsatellite markers spanning the HLA complex, from D6S291 and mapping beyond the DPB1 gene to D6S2239, which is located near the haemochromatosis locus. Using this approach, we tested the hypothesis that microsatellite markers could help identify less protective DQB1*0602 haplotypes and help map additional loci that modify diabetes risk.

Subjects, materials and methods

Swedish case-control study

The population-based matched case-control study used here has been described elsewhere [3, 4]. Patients were diagnosed with type 1 diabetes according to the World Health Organization criteria. Two main age groups were defined in this study: the first group included patients who were younger than 15 years at the time of diagnosis, which lay between 1 September 1986 and 31 December 1987. Matched control subjects were children of the same sex, born on the same day and in the same county as the index case [5]. The second group comprised patients who were aged 15 to 34 years at diagnosis and developed diabetes between 1 January 1987 and 31 December 1988; matched control subjects were also ascertained. The combined data set includes a total of 971 type 1 diabetic patients and 702 control subjects. Of the diabetic patients with DQB1*0602, eight had sufficient DNA available and were included in this analysis. In these eight patients, the mean age at diagnosis was 23.1 years (SD=8.4, range 11–34). Information about the Swedish case-control study is available at the Multi-locus Age-dependent Genetic Effects On Type 1 Diabetes website (http://stat-db.stat.sfu.ca/magenta/). In total, data from this study included 552 patients and 433 control subjects with HLA DR-DQ and microsatellite marker genotypes. All subjects provided informed consent. The Ethics Committee at the Karolinska Institute, Stockholm, Sweden, approved the study, which was carried out in accordance with the principles of the Declaration of Helsinki.

HLA genotyping methods

Molecular HLA typing data were generated by previously described PCR methods using sequence-specific primers (SSP), RFLP and sequence-specific oligonucleotide probes (SSOP) [3, 6]. We did not identify subjects carrying DQB1*0602 in cis with DRB1*11, 13 or 14. Thus, all DQB1*0602-positive subjects identified carried DRB1*1501. The presence of DQB1*0602 was confirmed by PCR-SSP/RFLP and PCR-SSOP methods, both of which clearly distinguish DQB1*0602 from DQB1*0603.

Microsatellite genotyping

DNA fragments containing the polymorphic sequences were amplified independently using the PCR primers listed below. For each marker one of the primers was labelled with a fluorescent dye. The primer sequences and dyes used were: D6S291: 5′:gtttcttggggatgacgaattattcactaact, 3′:*ggcattcaggcatgcctggc, dye used FAM; D6S273 5′:*gcaacttttctgtcaatcca, 3′:gtttcttaccaaacttcaaattttcgg, dye used FAM ; TNFd: 5′:*catagtgggactctgtctccaaag, 3′:gtttcttagatccttccctgtgagttctgct, dye used NED; MIB 5′:gtttcttctaccatgacccccttcccc 3′:*ccacagtctctatcagtcca, dye used HEX; D6S265 5′:*acgttcgtacccattaacct 3′:gtttcttatcgaggtaaacagcagaaa, dye used HEX; D6S2222: 5′:*agtcatctgaagagttgg, 3′:gttttcttgcatgtcttctttgttaagg, dye used NED; D6S2223: 5′:gtttcttaataatgttaagtaacaaactagagtac, 3′:*actcaagcctgggcaatagagc, dye used HEX; and for D6S2239: 5′:*gttggaagcaatggattagatgtcc, 3′:gtttcttctacctgccaggaacaatatacac, dye used FAM; * indicates the fluorescently labelled primer. Dye-labelled fragments were separated according to size on 96-capillary sequencers (MegaBACE 1000; Amersham Biosciences, Piscataway, NJ, USA). Genotypes were called by using the Genetic Profiler software (version 1.1; Amersham Biosciences,) applied to the raw MegaBACE data. Microsatellite alleles were defined by fragment size analysis in a reference panel of 50 cell lines from the International Histocompatibility Working Group (http://www.ihwg.org). Alleles were given a numerical value designated as the normalised fragment index and based on a ranking of relative fragment size ranging from smallest to largest. The D6S265*15 allele corresponds to a cytosine-adenine (ca) core repeat motif (15 repeats) (http://www.ihwg.org/shared/micros.htm).

Haplotype estimation and statistical methods

Haplotype frequencies were computed on the basis of the probabilities for all possible haplotype phases for each individual, which were assessed using the Bayesian algorithm implemented by phase [7]. A permutation test weighting the probability of each haplotype assignment was used to compare haplotype frequencies. Allele and haplotype frequencies between groups were compared using a Pearson’s chi square test.

Results

We first compared the case and control allele frequencies of all markers among individuals who carried one or two copies of the DRB1*15, DQB1*0602 haplotype (Fig. 1a). The only significant difference was found at marker D6S265 (p<0.005). The difference was significant after correcting for multiple comparisons (p<0.045). One allele in particular was found to be more common in DRB1*15, DQB1*0602 carriers who had type 1 diabetes cases than in control subjects: thus 75% of the DRB1*15, DQB1*0602 carrier patients had allele D6S265*15 compared with 47% of control subjects. However, given the small sample size, this difference is not statistically significant (χ 2 [1 df]=2.40). We then estimated DRB1, DQB1, D6S265 haplotype frequencies in the patient and control sets (Table 1). A comparison of the haplotype frequency distribution of D6S265*15 between patients and control subjects who were carriers of DRB1*15, DQB1*0602 haplotypes resulted in a Pearson’s chi square value of 10.69 (1 df) (p<0.001); a Fisher’s exact test on the same contingency table yielded p=0.0013. Moreover, the haplotype odds ratios (OR) are significantly higher, namely ten-fold higher for DRB1*15, DQB1*0602 haplotypes that carry allele D6S265*15 than for those carrying other alleles at this marker (Fig. 1b). Thus, we have identified a locus that marks DQB1*0602 haplotypes associated with increased risk of type 1 diabetes.

Fig. 1
figure 1

a Sequence map position of the markers genotyped, relative to the position of some HLA genes. The p value (log 10 scale) was derived by comparing the allele frequencies between patients and control subjects at eight microsatellite markers in the HLA complex region among subjects that carried one or two copies of the DRB1*15 DQB1*0602 haplotype. Source of sequence map position: http://www.ncbi.nlm.nih.gov/mapview. b Odds ratio for type 1 diabetes of DRB1*15 DQB1*0602 haplotypes according to allelic variation at the D6S265 locus in the Swedish case-control study

Table 1 DRB1*15 DQB1*0602 D6S265 haplotypes in Swedish type 1 diabetes cases and control subjects

Discussion

Previous investigations had identified very rare type 1 diabetic patients who carried normal DQB1*0602 sequences, and thus suggested that the protective effect associated with DRB1*15, DQB1*0602 haplotype is remarkably strong, but not absolute [2, 8]. We have identified a less protective DRB1*15, DQB1*0602 haplotype in a population-based case-control study of patients from Sweden. This haplotype is marked by allele 15 at the D6S265 locus. The finding that allelic variation at this locus can override the dramatic protective effect associated with DQB1*0602 provides further evidence that this locus or a locus nearby may have an important effect on type 1 diabetes risk. It will be important to assess whether polymorphisms at D6S265 mark less protective haplotypes in populations with different DRB1*15, DQB1*0602 frequencies.

Marker D6S265 maps 100 kb telomeric of HLA-A, which has been previously associated with diabetes susceptibility [9]. However, earlier studies dealt mainly with class I alleles, which influence the disease susceptibility associated with the DRB1 and DQB1 predisposing haplotypes such as DR3 or DR4. Molecular typing of the HLA-A locus will be necessary to determine whether the influence of D6S265 on the risk associated with the DRB1*15 DQB1*0602 haplotype is independent of the class I HLA-A locus. However, associations between D6S265 and other autoimmune diseases have been reported (see references in [10]). Of particular interest is the observation by Harbo and co-workers [10] of an association between susceptibility to multiple sclerosis and D6S265 specifically on DRB1*15, DQB1*0602 haplotypes. Taken together, their data and our findings indicate that genetic variation at D6S265 can influence or is linked to a locus that can influence susceptibility to or protection from the autoimmunity conferred by DRB1*15, DQB1*0602 haplotypes. Known genes that may be marked by D6S265 include HLA-A, HLA-B, MICA, TNF and BAT1. Polymorphisms at these loci may have important effects on the function of cytotoxic T cells and cytokine secretion. Moreover, possible effects on transcriptional regulation may perhaps influence the expression of the HLA-DQ molecule encoded by DQA1*0102, DQB1*0602. Further characterisation of this region will be needed to identify the loci that contribute to the genetic protection from type 1 diabetes conferred by DRB1*15, DQB1*0602. This may also help understand the mechanisms preceding the onset of the disease.

In conclusion, our data identify a region of potential interest and provide a marker that helps identify less protective DQB1*0602 haplotypes. The ability to identify DQB1*0602 haplotypes with attenuated protective effects is relevant to the design of prevention studies. Currently, first-degree relatives of type 1 diabetic patients who are screened for prevention studies and found to carry DQB1*0602 are considered to be at extremely low risk and excluded from receiving treatment, although they are routinely offered follow-up. Typing for DS6265 in this population should allow testing of whether this marker identifies those rare cases that may have attenuated protection and in turn increased risk of developing type 1 diabetes.