Clinical relevance of genotype–phenotype correlations beyond vascular events in a cohort study of 1500 Marfan syndrome patients with FBN1 pathogenic variants

Pauline Arnaud,Olivier Milleron,Nadine Hanna,Jacques Ropers,Nadia Ould Ouali,Amel Affoune,Maud Langeois,Ludivine Eliahou,Florence Arnoult,Philippe Renard,Marlène Michelon-Jouneaux,Marie Cotillon,Laurent Gouya,Catherine Boileau,Guillaume Jondeau

doi:10.1038/s41436-021-01132-x

Abstract

PurposeMarfan syndrome (MFS) is a connective tissue disorder in which several systems are affected with great phenotypic variability. Although known to be associated with pathogenic variants in the FBN1 gene, few genotype–phenotype correlations have been found in proband studies only. MethodsIn 1,575 consecutive MFS probands and relatives from the most comprehensive database worldwide, we established survival curves and sought genotype–phenotype correlations. ResultsA risk chart could be established with clinical and genetic data. Premature termination codon variants were not only associated with a shorter life expectancy and a high lifelong risk of aortic event, but also with the highest risk of severe scoliosis and a lower risk for ectopia lentis (EL) surgery. In-frame variants could be subdivided according to their impact on the cysteine content of fibrillin-1 with a global higher severity for cysteine loss variants and the highest frequency of EL surgery for cysteine addition variants. ConclusionThis study shows that FBN1 genotype–phenotype correlations exist for both aortic and extra-aortic features. It can be used for optimal risk stratification of patients with a great importance for genetic counseling and personalized medicine. This also provides additional data for the overall understanding of the role of fibrillin-1 in various organs.

Highlights

Marfan syndrome (MFS [MIM 154700]) is remarkable for its phenotypic variability and its evolving definition.[1,2] Limiting the definition of MFS patients to individuals with heterozygous pathogenic variants in the FBN1 gene has not reduced phenotypic variability as much as anticipated.[1]
The 1,575 MFS patients carried 643 different pathogenic variants in the FBN1 gene categorized either as “in-frame” (n = 379 [59%]) or “premature termination codon (PTC)” [n = 264 (41%]), figures in keeping with those reported in the UMD-FBN1 reference database.[4]
Aortic event rate, and prevalence of extra-aortic features in a large population of MFS patients carrying a pathogenic variant in the FBN1 gene and benefiting from modern care

Summary

Introduction

Marfan syndrome (MFS [MIM 154700]) is remarkable for its phenotypic variability and its evolving definition.[1,2] Limiting the definition of MFS patients to individuals with heterozygous pathogenic variants in the FBN1 gene has not reduced phenotypic variability as much as anticipated.[1] Pathogenic variants in the FBN1 gene have been associated with phenotypes ranging from normal (incomplete penetrance is rare but possible) to severe MFS in childhood with cardiovascular, ophthalmologic, skeletal, cutaneous, and neurologic features, and shortened life expectancy This clinical variability and unpredictability remain limiting factors for personalized follow-up and effective genetic counseling.[3] To date, more than 3,000 different MFS patients carrying more than 1,800 different pathogenic variants in the FBN1 gene have been described, covering the full spectrum of phenotypes and variant types.[4]. Few genotype–phenotype correlations have been reported:[6] pathogenic variants within the neonatal region (corresponding to former exons 24–32, 25–33) have been associated with more severe phenotypes, including neonatal MFS,[7] while premature termination codon (PTC) variants (held responsible for haploinsufficiency) have been associated with more severe cardiovascular phenotypes than in-frame pathogenic variants (themselves being associated with a dominant negative effect).[8,9,10,11] Variants changing the cysteine content in fibrillin-1 have been associated with more ophthalmologic manifestations.[7,12] Most of these correlations were evaluated in studies of probands only, who usually display more severe phenotypes, thereby limiting the phenotypic spectrum and the possible identification of genotype–phenotype correlations

Methods

Results

Conclusion