Family-based analysis using a dense single-nucleotide polymorphism-based map defines genetic variation at PSORS1, the major psoriasis-susceptibility locus

Abstract

Psoriasis is a common skin disorder of multifactorial origin. Genomewide scans for disease susceptibility have repeatedly demonstrated the existence of a major locus, PSORS1 (psoriasis susceptibility 1), contained within the major histocompatibility complex (MHC), on chromosome 6p21. Subsequent refinement studies have highlighted linkage disequilibrium (LD) with psoriasis, along a 150-kb segment that includes at least three candidate genes (encoding human leukocyte antigen-C [HLA-C], alpha-helix-coiled-coil-rod homologue, and corneodesmosin), each of which has been shown to harbor disease-associated alleles. However, the boundaries of the minimal PSORS1 region remain poorly defined. Moreover, interpretations of allelic association with psoriasis are compounded by limited insight of LD conservation within MHC class I interval. To address these issues, we have pursued a high-resolution genetic characterization of the PSORS1 locus. We resequenced genomic segments along a 220-kb region at chromosome 6p21 and identified a total of 119 high-frequency SNPs. Using 59 SNPs (18 coding and 41 noncoding SNPs) whose position was representative of the overall marker distribution, we genotyped a data set of 171 independently ascertained parent-affected offspring trios. Family-based association analysis of this cohort highlighted two SNPs (n.7 and n.9) respectively lying 7 and 4 kb proximal to HLA-C. These markers generated highly significant evidence of disease association (P<10-9), several orders of magnitude greater than the observed significance displayed by any other SNP that has previously been associated with disease susceptibility. This observation was replicated in a Gujarati Indian case/control data set. Haplotype-based analysis detected overtransmission of a cluster of chromosomes, which probably originated by ancestral mutation of a common disease-bearing haplotype. The only markers exclusive to the overtransmitted chromosomes are SNPs n.7 and n.9, which define a 10-kb PSORS1 core risk haplotype. These data demonstrate the power of SNP haplotype-based association analyses and provide high-resolution dissection of genetic variation across the PSORS1 interval, the major susceptibility locus for psoriasis.

Figure 1

The target region, spanning 220 kb in the MHC class I interval. a, Position of repetitive elements and sequenced areas, plotted against distance (upper line). The positions of SNPs identified by resequencing are represented by circles (for SNPs genotyped in the cohort with psoriasis) and squares (for nontyped SNPs). b, Positions of known genes. Arrows indicate the direction of transcription. c, Outcome of PSORS1 refinement studies. Black boxes indicate published minimal intervals. Dotted lines have been added to illustrate how much farther these regions may extend according to more-conservative interpretations of data.

Figure 2

Patterns of LD conservation, as measured by analysis of 15 SNPs spanning 700 kb, at the boundary between MHC class I and class III loci. D′ decay from HLA-C SNP 290 (dashed line) and from CDSN1243 (continuous line) is plotted against distance and location of analyzed genes. The asterisk (*) and double asterisk (**) symbols denote the positions of HLA-C SNP 290 and of CDSN1243, respectively.

Figure 3

Output of TRANSMIT family-based association analysis undertaken using all 171 parent–affected offspring trios. a, Results of single-marker analysis. χ² values at each marker are plotted against distance and position of known genes. b, Results of three-marker analysis. χ² values for consecutive haplotypes are plotted against distance and position of known genes.

Figure 4

PSORS1 SNP haplotypes. Haplotypes are clustered on the basis of the presence of disease-associated alleles (indicated by a + or − sign before the locus name) at SNPs n.7 and n.9; at HCR269 (n.43) (Asumalahti et al. 2000), renamed as “HCR305” (Asumalahti et al. 2002); and at CDSN1243 (n.58) (Allen et al. 1999). The last four columns report haplotype frequencies (freq), observed (obs) versus expected (exp) transmissions, and haplotype status (neutral, overtransmitted [over], or undertransmitted [under]).