The diagnosis of Marfan syndrome (MFS) depends on a multidisciplinary clinical evaluation and identification of a set of diagnostic criteria, recently revised, known as the Ghent criteria 1. The spectrum of clinical manifestations is very wide, but typically the cardiovascular, musculoskeletal, and ocular systems are involved.

Molecular study to identify mutations in the FBN1 gene can establish a definitive diagnosis even with atypical or “incomplete” phenotypes. This gene, located on chromosome 15q21.1, codes for a glycoprotein, profibrillin-1, which is then converted to fibrillin-1, a major component of extracellular microfibrils. Over 600 different mutations have been described 2, almost all unique to an affected individual or family 3.

Mutations may be found anywhere on the FBN1 gene, and may be missense or lead to a premature termination codon (PTC) 3.

Of the few genotype-phenotype correlations that have been established, the most notable is the association of neonatal MFS, the most severe form of the syndrome, with mutations in exons 24-32.

Early identification of asymptomatic carriers is desirable given the high penetrance of mutations and the morbidity and mortality associated with MFS, mainly related to the risk of aneurysm rupture and dissection of the ascending aorta 4. There are treatments that can improve the prognosis of these patients if applied in time.

There are few published data concerning the molecular diagnosis of patients with MFS in the Portuguese population.

The aim of the present work was to evaluate the frequency and type of FBN1 gene mutations in a population of Marfan syndrome patients referred to a tertiary care center with cardiothoracic surgery.

The study population included 30 individuals with MFS from 14 unrelated families (14 index cases and 16 relatives), 57 % male, mean age 37±12 years.

Of these, 27 (90 %) presented major cardiovascular criteria; 13 (43 %) had undergone surgical aortic root replacement, 43 % had mitral valve prolapse, 23 % evidence of aortic dissection and 23 % dilatation of the descending aorta. Regarding medication, 19 patients (63 %) were taking beta-blockers, six (20 %) angiotensin-converting enzyme inhibitors, and seven (23 %) angiotensin II receptor blockers.

The prevalences of the Ghent criteria are shown in Table 1.

Genetic study was performed in the 14 index cases and in 22 relatives (16 with positive genetic test and 8 with no FBN1 mutation identified). The results of molecular study and their correlation with the presence of MFS criteria are shown in Tabl 2.

Twelve different mutations were identified in the 14 families: ten new and two known (cases 8 and 13, mutations Cys2111Arg 2 and Arg861X 9). One patient had two different mutations. The mutations were identified in functionally distinct domains: three in TGF-β-like domains, six in cbEGF-like domains, two in hybrid domains, and one in the terminal COOH region.

In three index cases that fulfilled the Ghent criteria, no FBN1 mutation was found.

To the best of our knowledge this is the first genotypic description of Portuguese patients with MFS.

The percentage of patients fulfilling the Ghent criteria in whom mutations were detected was slightly lower than in the literature 10,11 (79 % vs. 90 %, respectively), but this may be due to the wide confidence interval (64.4 %-93.6 %) resulting from the small sample size.

The high frequency of new mutations (10 out of 12) supports current thinking that most MFS patients are carriers of a genetic variant unique to their family.

Interestingly, the type and distribution of mutations in our patients were significantly different from those in published data 9,12. We found 36 % missense mutations and 50 % leading to a premature termination codon. As expected, most carriers of a PTC mutation did not present ectopia lentis 9. In the literature, missense mutations account for two-thirds of FBN1 mutations, most of which affect cysteine residues associated with cbEGF units and thus interfere with disulfide bonds and protein stability. Individuals with such mutations are more likely to have ectopia lentis 12. The rest usually have PTC mutations, leading to the formation of incomplete proteins that interfere with normal protein polymerization (dominant negative effect) 13 or result in functionally inactive proteins through RNA degradation, reducing the total quantity of fibrillin-1 (haploinsufficiency) 10,14. Individuals with these mutations tend to present more severe musculoskeletal alterations, particularly joint hypermobility, greater propensity for aortic dissection 9,15, and a significantly lower risk for ocular involvement 9, although there is no direct relationship between phenotype severity and the quantity of mutant transcript.

Our results may have been influenced by selection bias. As the series is from a tertiary care center, a large proportion of the patients fulfilled the Ghent criteria (both cardiovascular and musculoskeletal) and therefore nonsense mutations would be likely to be more prevalent 9,12.

The fact that we discovered two different mutations in one patient highlights the value of complete sequencing of the FBN1 gene, particularly in certain patients. The few cases in the literature of compound heterozygosity in MFS 16,17 are characterized by a more severe phenotype. Multiple mutations may also contribute to phenotypic variability within families.

We did not identify any asymptomatic carriers (those with FBN1 gene mutations but no major clinical manifestations). However, screening of relatives revealed individuals with mild morphological abnormalities that were insufficient for a diagnosis of MFS but for whom we were then able to provide appropriate clinical monitoring and genetic counseling. Excluding MFS in other relatives, meanwhile, meant that they did not require periodic clinical assessments or imaging studies.

In the three cases in which no mutation was found, it is possible that a mutation existed in the promoter or non-coding intronic regions of FBN1, or in large gene rearrangements undetectable by PCR; there may also have been mutations in other genes, particularly those coding for TGF-β type I and II receptors (TGF-βI and TGF-βII) 18.

In this study we have highlighted the need for multidisciplinary clinical evaluation of MFS patients and their relatives.

By describing 10 new mutations, we have broadened the spectrum of known FBN1 mutations associated with MFS.

A better understanding of the etiopathology of Marfan syndrome, including molecular study, may lead to the discovery of new prognostic markers and earlier identification of candidates for therapeutic interventions 9.

This work was funded by a grant from the Research Center of Hospital de S. Joao, Porto, Portugal.

The authors have no conflicts of interest to declare.