Quantitative Trait Locus (QTL) Linkage Analysis of a Large Pedigree with Von Willebrand Disease Identifies Novel Modifier Loci of Von Willebrand Factor Levels and Confirms Some Previously Reported Modifier Loci

Abstract 240

Von Willebrand disease (VWD) is the most common inherited bleeding disorder with a prevalence ranging from 0.8 to 1.3%. VWD is characterized by incomplete penetrance and variable expressivity, even in families with a unifying VWF mutation, suggesting the presence of modifier genes. We diagnosed and characterized a 1500-member multigenerational Amish family in which 132 individuals exhibit a single autosomal-dominant C4120T mutation (R1374C) in the A1 domain of the mature von Willebrand Factor (VWF) molecule. When accounting for the C4120T mutation, analyses of coagulation factor VIII levels (FVIII:C) and log transformed, sex- and age-adjusted VWF levels in SOLAR estimated a heritability (h²) of 0.301 for VWF antigen levels (VWF:Ag; p=2.7×10⁻⁶), 0.386 for VWF ristocetin cofactor activity (VWF:RCo; p=1.16×10⁻⁸), and 0.280 for FVIII:C (p=1.52×10⁻⁵). These results demonstrate strong evidence of residual heritability of the traits despite the obvious effect of the 4120 mutation. Among individuals with the wild-type VWF CC4120 genotype, ABO genotype, the only known major modifier of VWF levels, explained 6.5, 10.0 and 4.2% of the variability in VWF:Ag, VWF:RCo and FVIII:C levels, respectively, suggesting that a significant genetic component of the variability of these traits remains unaccounted for.

In order to search for modifier loci that contribute to the observed variability we then performed a primary genetic screen at an average 10 cM interval using a standard short tandem repeat polymorphism mapping set in a 384-member subset of the pedigree. Quantitative trait locus (QTL) linkage analysis was performed using SOLAR for VWF:Ag, VWF:RCo and FVIII:C. The highest LOD score generated demonstrated linkage of VWF:RCo to region 12p13 where VWF has been physically mapped (LOD=64.73). When accounting for the C4120T mutation, five other chromosomal regions were identified with LOD scores above 2.0 with linkage to FVIII:C, VWF:Ag, and/or VWF:RCo: 2q14 (LOD=2.48), 6p12 (LOD=2.22), 9q33 (LOD=2.84), 10q26 (LOD=2.29), and 16q11 (LOD=2.66). Notably, the ABO locus maps to chromosome 9q33, demonstrating by linkage the known effect of ABO on VWF levels.

Five chromosomal regions demonstrated LOD scores between 1.5 and 2.0 for the above phenotypes: 3p14 (LOD=1.56), 4q34 (LOD=1.64), 5q33 (LOD=1.8373), 15q14 (LOD=1.78), and 21q21 (LOD=1.85). Comparison to other linkage and association studies such as the Genetic Analysis of Inherited Thrombophilia (GAIT) and the Framingham Study identified several concurrent peaks in our analysis. Specifically, 22q11and 9q34 were reported in the GAIT study as chromosomal regions linked to VWF:Ag and 6p22 was reported in the Framingham study as well as in the GAIT study as linked to the same phenotype. Furthermore the 22q11 and 21q21 signals display orthology to Mvwf6, a murine modifier region of VWF levels recently reported. Two other signals, 5q33 and 6p12 also display orthology with regions closely syntenic to Mvwf4, another murine modifier. These comparisons strengthen the validity of this multigenerational pedigree as a tool to identify genetic modifiers of the phenotypic variability of VWF levels and provide the basis to which chromosomal regions will be followed-up to identify causative alleles. While genome-wide association studies (GWAS) are characterized by the identification of common variants of relatively modest effect, linkage studies such as the one we present here may have the ability to identify rare variants as well as allelic heterogeneity at a given modifier locus, which may explain a significant portion of the observed variability and likely account for a portion of the genetic component not identified by GWAS analyses.