Genetic modifiers predisposing to congenital heart disease in the sensitized Down syndrome population

Abstract

Background: About half of people with Down syndrome (DS) exhibit some form of congenital heart disease (CHD); however, trisomy for human chromosome 21 (Hsa21) alone is insufficient to cause CHD, as half of all people with DS have a normal heart, suggesting that genetic modifiers interact with dosage-sensitive gene(s) on Hsa21 to result in CHD. We hypothesize that a threshold exists in both DS and euploid populations for the number of genetic perturbations that can be tolerated before CHD results.

Methods and results: We ascertained a group of individuals with DS and complete atrioventricular septal defect and sequenced 2 candidate genes for CHD: CRELD1, which is associated with atrioventricular septal defect in people with or without DS, and HEY2, whose mouse ortholog (Hey2) produces septal defects when mutated. Several deleterious variants were identified, but the frequency of these potential modifiers was low. We crossed mice with mutant forms of these potential modifiers to the Ts65Dn mouse model of DS. Crossing loss-of-function alleles of either Creld1 or Hey2 onto the trisomic background caused a significant increase in the frequency of CHD, demonstrating an interaction between the modifiers and trisomic genes. We showed further that, although each of these mutant modifiers is benign by itself, they interact to affect heart development when inherited together.

Conclusions: Using mouse models of Down syndrome and of genes associated with congenital heart disease, we demonstrate a biological basis for an interaction that supports a threshold hypothesis for additive effects of genetic modifiers in the sensitized trisomic population.

Figure 1. A threshold model for CHD

We hypothesize that the human population includes allelic variants in multiple genes that contribute to the risk of congenital heart disease, many of which have subtle or no effects by themselves. Additive effects of individual modifier genes can reach a threshold whereby heart septal development is disrupted (Eu with CHD), but the likelihood of inheriting many predisposing modifiers is small. Trisomy 21 is a significant risk factor for CHD, but alone is not sufficient to produce heart defects. However, those people with an extra copy of Hsa21 may require fewer disomic or trisomic modifiers to reach the threshold (Ts21 + Modifier). The relative contribution of the modifier in the sensitized DS population is therefore more readily detectable.

Figure 2. Resequencing strategy for CRELD1 and HEY2

We assessed coding regions of these genes including all exons with 50+ bp of sequence in the flanking introns, the 5′ UTR and 50+ bp of the 3′ UTR. A) Resequencing strategy for CRELD1; B) Resequencing strategy for HEY2. For exact coordinates, see Supplemental Table 1.

Figure. 3. A variety of septal defects were observed in mutant and trisomic mice at P0

A) Normal heart showing intact ventricular septum at P0; B) membranous VSD; C) muscular VSD ; D) normal heart showing atrial septum; E) ostium secundum ASD; F) ASD at higher magnification. For the incidence of defects in various models, see Table 2. Arrows indicate communication between the chambers. RA: right ventricle; LV: left ventricle; RA: right atrium; LA: left atrium; Scale bars: A-E, 400 μm, f, 150 μm.

Figure 4. Notch1, Creld1 and Hey2 interaction

A TaqMan assay was used to show coordinate regulation of Creld1 and Hey2 transcript levels in hearts of heterozygous null mice of either genotype. Notch1 was significantly down-regulated in Creld1^+/− hearts, as well. Fluorescent-(FAM-) labeled Creld1, Hey2, Notch1 and Gapdh were normalized to a VIC-labeled internal control, β-actin. Gapdh was used as a negative control. . The number of hearts tested was 4-6 per analysis as indicated on the dot plot. All statistical comparisons are relative to WT. * p<0.05; ** p<0.01 (Mann-Whitney test).

Figure 5. Hey2 can activate Creld1 expression in L-cells

A) Western blot showing Hey2 protein expression in the L-cells transfected with pCS2+/Hey2 but not in control cells transfected with pCS2+ alone . An antibody to β-actin was used as a loading control. B) TaqMan assay showed a 1.9-fold increase of Creld1 mRNA expression in pCS2+/Hey2 transfected L-cells compared to controls. Plotted values are the average ratios of Creld1 mRNA expression in seven experiments. Std. deviation is indicated. p value is indicated (Mann-Whitney test).