Background:In newborns, severe congenital heart defects are rarer than mild ones. The reason why is unknown, but presumably related to a liability threshold that rises with the severity of a defect. Because the same genetic mutation can cause different defects, other variables may contribute to pushing an individual past a defect-specific liability threshold. We consider here how variables in the genetic architecture of a heart defect depend upon its fitness cost, as defined by the likelihood of survival to reproductive age in natural history studies.
Methods:We phenotyped ~10,000 Nkx2-5 +/newborn mice, a model of human congenital heart disease, from two inbred strain crosses. Genome-wide association analyses detected loci that modify the risk of an atrial septal defect, membranous or muscular ventricular septal defect, or atrioventricular septal defect. The number of loci, heritability and quantitative effects on risk of pairwise (G×GNkx) and higher-order (G×G×GNkx) epistasis between the loci and Nkx2-5 mutation were examined as a function of the fitness cost of a defect.Results: Nkx2-5 +/mice have pleiotropic heart defects; about 70% have normal hearts. The model recapitulates the epidemiological relationship between the severity and incidence of a heart defect. Neither the number of modifier loci nor heritability depends upon the severity of a defect, but G×GNkx and G×G×GNkx effects on risk do. Interestingly, G×G×GNkx effects are three times more likely to suppress risk when the genotypes at the first two loci are homozygous and from the same, rather than opposite strains in a cross. Syn-and anti-homozygous genotypes at G×G×GNkx interactions can have an especially large impact on the risk of an atrioventricular septal defect.
Conclusions:Given a modestly penetrant mutation, epistasis contributes more to the risk of severe than mild congenital heart defect. Conversely, genetic compatibility between interacting