Introduction

Autoimmune diseases are common disorders, affecting more than 5% of the world population. They share aspects of aberrant immunological tolerance toward self-antigens, and all are known to have genetic components. Several multiple, large-scale, genetic studies on autoimmune-disease-associated SNPs have been reported recently: peptidylarginine deiminase type 4 (PADI4) in rheumatoid arthritis (RA) (Suzuki et al. 2003); solute carrier family 22 members 4 and 5 (SLC22A4 and 5) in RA (Tokuhiro et al. 2003) and Crohn’s disease (CD) (Peltekova et al. 2004); programmed cell death 1 (PDCD1) in systemic lupus erythematosus (SLE) (Prokunina et al. 2002), type 1 diabetes mellitus (T1D) (Nielsen et al. 2003), and RA (Prokunina et al. 2004); and PTPN22 in T1D (Bottini et al. 2004), RA (Begovich et al. 2004), and SLE (Kyogoku et al. 2004).

Interestingly, the allele frequencies of some of these autoimmune-disease-associated SNPs vary substantially in different ethnic groups, as reported for CD-associated SNPs in the Caucasian and Japanese populations (Yamazaki et al. 2004). Information on ethnic differences in allele frequency of disease-associated variants is important for better understanding of the pathologic mechanisms of polymorphisms. We therefore investigated the allele frequency of nine SNPs in four autoimmune-disease-associated genes/loci in Caucasian, African-descent, and Japanese populations.

Subjects and methods

Japanese volunteers were recruited, and their informed consent to the study was obtained, as required by our ethical committee and as previously described (Suzuki et al. 2003). Human variation panels for African American and Caucasian samples were obtained from Coriell Cell Repositories. We genotyped 376 Japanese, 94 African American, and 94 Caucasian subjects for the following nine SNPs in four loci: padi_89, 90, 92, and 104 in PADI4 (Suzuki et al. 2003); slc2F1 and slc2F2 (Tokuhiro et al. 2003) and L503F in SLC22A4 (Peltekova et al. 2004); G-207C in SLC22A5 (Peltekova et al. 2004); PD-1.3A in PDCD1 (Prokunina et al. 2002), and R620W in PTPN22 (Bottini et al. 2004). Japanese data for the four SNPs in PADI4 were identical to those reported previously (Suzuki et al. 2003). Genotyping was performed using the Invader assay, TaqMan assay, or direct sequencing, and haplotype frequency was inferred with an EM algorithm and/or Haplotyper (Niu et al. 2002). The statistical significance of differences in allele frequency was determined (P value in Fisher’s exact test), and the genetic distance indicated by differences in allele frequency was quantitated as FST.

Results

Allele frequencies of the nine SNPs in the three ethnic groups are shown in Table 1. The largest difference in allele frequency among ethnic groups was 39% (L503F, C versus J). Three pairwise comparisons Caucasian versus Japanese (C versus J), African American versus Japanese (A versus J), and African American versus Caucasian (A versus C), were performed for the allele frequency of each SNP and CD-susceptible haplotype in the SLC22A4/5 locus. Of the nine SNPs in four loci, a significant difference in allele frequency (P < 0.01) was observed: five SNPs in three loci were significant in the C versus J comparison, five SNPs in three loci in the A versus J comparison, and two SNPs in two loci in the A versus C comparison. The FST values of more than 0.1 were obtained for four SNPs in two loci: one SNP in one locus was significant in the C versus J comparison, two SNPs in one locus in the A versus J comparison, and two SNPs in two loci in the A versus C comparison.

Table 1 Allele frequency of autoimmune-disease-associated variants in three ethnic groups
Full size table

Discussion

The allele frequencies of the four SNPs in PADI4 do not differ among the three ethnic groups, as indicated by a low FST value and an insignificant P value. However, SNPs in all the other three loci had allele frequencies that were significantly different in at least one of the comparisons among the three ethnic groups. The most remarkable difference was observed for the L503F SNP and haplotypes containing the SNP in SLC22A4/5 in a comparison between the Caucasian and Japanese populations. The SNP and the corresponding haplotypes were much rarer in the Japanese population and were also rare in African Americans. Although other differences in allele frequency less outstanding than L503F in SLC22A4/5, it was of note that PD-1.3A in PDCD1 and R620W in PTPN22 were only polymorphic in Caucasians and in the African-descent population, as for L503F in SLC22A4/5. Populations of African descent, Caucasians, and Japanese (an ethnic subgroup from far-east islands), should have different population histories of mutation, migration, isolation, and genetic drift, and there is no doubt that there should be substantial differences in allele frequencies among these ethnic groups. However, an explanation of the remarkable difference seen for L503F in SLC22A4/5 and others appears to require an additional mechanism. We note that Stefansson et al. (2005) reported that a 900-kb inversion polymorphism is frequently present in Caucasians but is rare in Africans and almost absent in East Asians, similar to the L503F SNP, and that prevalence of the variation in Caucasians was due to a positive effect in reproduction.

Beside the population genetics perspective, the difference in allele frequency among ethnic groups is important for association studies. Ioannidis et al. (2004) recently summarized multiple meta-analyses on common disease-susceptible polymorphisms from the perspective of ethnic variability. They concluded that the relative risk produced by susceptible variants seemed to be more similar among various ethnic groups compared with variability in their allele frequency. These findings would be correct in general but may not be applicable for polymorphisms that are common for some ethnic groups but absent in other groups. Ethnic uniformity and diversity in the identity of common disease-susceptible variants and their genetic contribution need to be extensively investigated before ethnic variations of polymorphisms are understood. Therefore, the details and differences in ethnic variations need to be incorporated into investigations of susceptibility variants for common diseases.