Quantifying the Genetic Basis of Marfan Syndrome Clinical Variability

Abstract

Marfan syndrome (MFS) is an autosomal dominant connective tissue disorder with considerable inter- and intra-familial clinical variability. The contribution of inherited modifiers to variability has not been quantified. We analyzed the distribution of 23 clinical features in 1306 well-phenotyped MFS patients carrying FBN1 mutations. We found strong correlations between features within the same system (i.e., ophthalmology vs. skeletal vs. cardiovascular) suggesting common underlying determinants, while features belonging to different systems were largely uncorrelated. We adapted a classical quantitative genetics model to estimate the heritability of each clinical feature from phenotypic correlations between relatives. Most clinical features showed strong familial aggregation and high heritability. We found a significant contribution by the major locus on the phenotypic variance only for ectopia lentis using a new strategy. Finally, we found evidence for the "Carter effect" in the MFS cardiovascular phenotype, which supports a polygenic model for MFS cardiovascular variability and indicates additional risk for children of MFS mothers with an aortic event. Our results demonstrate that an important part of the phenotypic variability in MFS is under the control of inherited modifiers, widely shared between features within the same system, but not among different systems. Further research must be performed to identify genetic modifiers of MFS severity.

Keywords: fibrillin; heritability; marfan; modifiers.

Figure 1

Cross correlations between 19 clinical features of MFS. Clinical features are represented as boxes. Red links between the boxes represent correlations significant after Bonferroni correction (p < 3 × 10⁻⁴). (E.L.: Ectopia lentis).

Figure 2

Agglomerative hierarchical clustering of clinical features of MFS. Agglomerative hierarchical clustering showed ocular (in green), cardiovascular (in red), and systemic clusters (in blue). Complete linkage clustering was chosen as linkage criteria, and the metric was Pearson correlations coefficient corrected for covariates. Node values indicate the minimal correlation between any couple of features within the corresponding subtree. On the right, the five clusters given by a 5% threshold are circled in red.

Figure 3

Heritability (H²) estimates of MFS clinical features. Heritability estimates were based on familial aggregation of clinical features among first-degree relatives. Error bars show standard deviations (SDs) and rely on asymptotic SDs of parent–offspring and siblings correlations provided by S.A.G.E. (Statistical Analysis for Genetic Epidemiology).

Figure 4

The Carter effect on the cardiovascular phenotype of MFS. (A) 61% of patients that had aortic surgery or dissection were males (p = 4.9 × 10⁻⁵). (B) Among MFS patients whose transmitting parent had aortic surgery or dissection, the prevalence of aortic surgery or dissection was significantly higher in the case of maternal transmission (p = 0.018). Standard deviations (SDs) are represented by error bars. d/N: number of patients with aortic surgery or dissection/sample size. (C) Among MFS patients whose transmitting parent did not have aortic surgery or dissection, the prevalence of aortic surgery or dissection was slightly higher in the case of maternal transmission (non-significant). SDs are represented by error bars. d/N: number of patients with aortic surgery or dissection/sample size. *: p < 0.05.