A high-density SNP genome-wide linkage search of 206 families identifies susceptibility loci for chronic lymphocytic leukemia

Abstract

Chronic lymphocytic leukemia (CLL) and other B-cell lymphoproliferative disorders display familial aggregation. To identify a susceptibility gene for CLL, we assembled families from the major European (ICLLC) and American (GEC) consortia to conduct a genome-wide linkage analysis of 101 new CLL pedigrees using a high-density single nucleotide polymorphism (SNP) array and combined the results with data from our previously reported analysis of 105 families. Here, we report on the combined analysis of the 206 families. Multipoint linkage analyses were undertaken using both nonparametric (model-free) and parametric (model-based) methods. After the removal of high linkage disequilibrium SNPs, we obtained a maximum nonparametric linkage (NPL) score of 3.02 (P = .001) on chromosome 2q21.2. The same genomic position also yielded the highest multipoint heterogeneity LOD (HLOD) score under a common recessive model of disease susceptibility (HLOD = 3.11; P = 7.7 x 10(-5)), which was significant at the genome-wide level. In addition, 2 other chromosomal positions, 6p22.1 (corresponding to the major histocompatibility locus) and 18q21.1, displayed HLOD scores higher than 2.1 (P < .002). None of the regions coincided with areas of common chromosomal abnormalities frequently observed in CLL. These findings provide direct evidence for Mendelian predisposition to CLL and evidence for the location of disease loci.

Figure 1

NPL scores across each chromosome. In each plot, ---- shows NPL statistics obtained using all SNPs (with LD; n = 9690), while ▔ shows NPL statistics obtained after exclusion of high-LD SNPs (n = 7495).

Figure 2

Plots of transformed linkage statistics (−log₁₀[P value]) after the removal of high-LD SNPs for chromosomes 2, 5, 6, 11, and 18. Transformed HLOD scores under the best dominant model are shown in black; under the best recessive model, in red; and transformed NPL values, in blue. The best recessive model for each of the 5 chromosomes was chromosomes 2, 6, and 18 (common recessive) and chromosomes 5 and 11 (rare recessive). The best dominant model for each of the 5 chromosomes was chromosomes 2, 6, and 11 (common dominant) and chromosomes 5 and 18 (rare dominant).