SUBJECTS AND METHODS

We investigated 11 consecutive patients (seven men; four women) with predominant EL, eight of whom had a family history of dominantly inherited lens dislocation, and in whom a diagnosis of MFS was excluded according to the current diagnostic criteria.³ Physical examinations and investigations are reported in Table 1, and included echocardiography with measurement of aortic root diameter (adult upper limit 39 mm), assessment of valve morphology and function, skeletal features, skin extensibility and lumbar striae, assessment of visual acuity, slit lamp and fundus examination. In the absence of a suitable clinical indication, such as severe chronic low back or abdominal pain, magnetic resonance imaging (MRI) for dural ectasia was not performed. Peripheral blood samples were collected, with appropriate informed consent, from probands and available family members.

Mutational analysis

A set of 65 pairs of primers (Sigma-Genosys, Pampisford, Cambridgeshire, UK) was used for polymerase chain reaction (PCR) amplification of all 65 exons of the FBN1 gene. The oligonucleotide sequences were those previously described¹⁵ or reported in the literature.^16,17 All PCR reagents except primers were supplied by Gibco-BRL Life Technologies (Paisley, UK). The amplifications were performed in a DNA Thermal-Cycler (Hybaid, Ashford, Middlesex, UK) under conditions depending on the exon analysed.

Single strand conformation analysis (SSCA) was carried out as previously described.¹⁵ Electrophoresis and silver staining chemicals were from Merck (Poole, Dorset, UK) and Sigma-Aldrich (Poole, Dorset, UK).

Sequencing was carried out on an ABI310 using the Big Dye chemistry (Applied Biosystems, Warrington, Cheshire, UK).

Detailed protocols are available on request.

RESULTS

We analysed 11 patients, whose clinical findings are schematically reported in Table 1. The six FBN1 mutations, one of which recurred in two patients, are reported in Table 2, together with three DNA variants which might represent rare polymorphisms. Several known polymorphisms were identified, without any significant difference in distribution between patients and controls (data not shown). The known polymorphisms and the mutations found were tested in all family members available. Wherever possible to establish, the polymorphisms were not the results of de novo changes in the DNA of patients. The mutations were absent in 160 chromosomes from unrelated controls and in 280 chromosomes from unrelated patients affected by cardiovascular diseases, including thoracic aortic aneurysm and MFS (data not shown). This result led us to exclude the possibility of the mutations identified being polymorphisms.

Clinical summaries

Patients IBP, VW, and MZ demonstrated only FBN1 variants likely to represent polymorphisms, so that no further blood samples were taken from family members. No family members were available for patients OP and JL.

In patient BM the mutation was demonstrated in the proband’s younger sister, who has bilateral EL and glaucoma, and in this sister’s as yet unaffected 5 year old son. The proband’s unaffected older sister demonstrated the mutation, but is clinically unaffected, presumably an example of reduced penetrance, as reported previously in another EL family.¹² The proband’s unaffected mother and brother did not carry the mutation, presumably inherited from the proband’s deceased father, who appears to have large eyes in his photographs, although he is not known to have had EL or glaucoma.

The mutation discovered in patient GB was also found in an affected brother but not in the brother’s two unaffected adult daughters. Both parents of the proband, including an affected father, died of unrelated causes (cancer) and DNA samples were unavailable.

The echocardiogram of 10 year old patient NS was within normal limits for age. The father of proband NS carries the mutation but does not have dislocated lenses. He demonstrates slight facial asymmetry with simple ears, slight rib sulcus on the left anteriorly, and a non-progressive dilated aortic root (40 mm; upper normal limit 39 mm), unchanged at 42 and 45 years, with mild aortic regurgitation. He has slightly lax skin over the elbow and large hands and feet. The proband’s brother is unaffected on eye examination and does not carry the mutation. This family demonstrates intrafamilial phenotype variability.

The family of patient MG demonstrates six affected members in three generations, with dominantly inherited EL. The proband’s affected mother and two affected children also carry the mutation. Affected family members do not fulfil the criteria for MFS. No unaffected relatives were available for analysis. Two affected adult siblings of the mother were uncooperative. The proband’s mother died unexpectedly aged 72 of undiagnosed ruptured ascending thoracic aortic aneurysm.

DISCUSSION

Analysis of the literature reveals that mutations affecting cysteines, calcium binding amino acids, or residues conserved among similar domains, are usually associated with more severe phenotypes.^7–9 Considering the six different mutations reported in this study, three mutations (R122C, R240C, R545C) have already been reported, all of them involving the substitution of a cysteine for a non-conserved arginine. Mutation R122C has been reported four times,^13,18,19 in all cases the patient presenting with atypical MFS, due to the lack of serious cardiovascular manifestations. Mutation R240C has been reported twice, in a classic MFS patient¹³ and in an EL patient,¹⁴ reflecting the interfamilial phenotype variability. Mutation R545C has been reported twice, in patients with cardiovascular involvement.^13,17 Nevertheless, in our series, these three mutations were characterised in four patients without any serious cardiovascular involvement.

Overall, four out of seven mutations (57%) reported in our study involved the substitution of a cysteine for a non-conserved arginine. This could be driven by the highly probable C to T transitions at CpG dinucleotides, but conversely arginine to cysteine mutations have been reported in only 10 of the 137 entries (∼7%) of the Marfan database⁸ (p<0.005, Fisher’s test) and in only four of the 219 entries (∼2%) of the FBN1 Human Gene Mutation Database, Cardiff (HGMD; http://archive.uwcm.ac.uk/uwcm/mg/search/127115.html) (p<0.0001, Fisher’s test). This statistically significant difference might be indicative of a recurrent type of mutation in the group of patients affected by predominant EL.

Moreover, the same type of mutation recurs in two out of four reported cases of predominant EL.^12–14

Considering the three novel mutations described, C652Y causes a cysteine substitution in a cbEGF-like domain in patient MG. The severe nature of the mutation identified, the family history and the borderline aortic root dimension make long term echocardiogram follow up of all family members necessary. Could possible worsening of the condition be anticipated by the molecular diagnosis, or is this a mildly affected patient carrying an apparently severe mutation?

Mutation N164S (patient BM) causes an asparagine to serine substitution in an EGF-like domain. This type of mutation has been reported five times in the literature,^8,20 but in all cases it involved an asparagine part of a consensus sequence for calcium binding. The mutation in this study is instead a substitution of a variable residue and is conservative, with both amino acids polar but uncharged. It is not clear if this mutation is causative.

Mutation S634P (patient NS) does not affect a conserved or invariant amino acid for cbEGF-like domains. Although it is not considered classically severe, it remains to be established what is the real effect of a non-conservative substitution of a non-polar for a polar uncharged residue. It should be noted that the proband’s father could be described as a marfanoid patient, thus suggesting a case of variable penetrance.

FBN1 variant A/G 3963 (patient VW) affects the second to last nucleotide of exon 31 and it has previously been reported as a polymorphism.²¹ However, the A residue at this position is predominant in FBN1 exon “donor sequences” (56%) and the same mutation has been observed in a patient with late onset aortic aneurysm, EL and marfanoid build (P Johnson, personal communication), where it affects the splicing of exon 31 through activation of a cryptic splice site. The unavailability of RNA from patient VW does not permit further analysis.

IVS6 del T -26 (patient IBP) and IVS62 A+8 to C (patient MZ) are other FBN1 variants observed in this study. They represent rare polymorphisms already reported in the literature.^14,22

In this study we report a success rate of 7/11 (63%) in identifying causative or putative FBN1 mutations in patients affected by predominant EL with marfanoid habitus. These results are within the 23%–86% range of mutation identification rate in MFS and MFS related patients in recent investigations.^{13,14,16,17,21} These results confirm that SSCA, although not 100% successful, is a suitable method and a viable alternative to direct sequencing for mutations identification, because of its low cost, simplicity, and efficiency, once the optimal conditions have been established.

Although the group studied is small, it is the largest predominant EL consecutive series yet reported. All mutations described in this paper are in the first 15 exons of the gene, while the Marfan database mutations⁸ are distributed over the gene length, with just 21/137 (15%) identified in the same region (Fig 1) (p<0.0001, Fisher’s test). This statistically significant difference is maintained even when compared with the HGMD (p<0.0001, Fisher’s test), where 31/219 mutations (14%) are within the first 15 exons.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Schematic representation of the FBN1 mutations distribution. At the top are the FBN1 mutations reported in the Marfan database,⁸ while at the bottom are indicated the mutations identified in this study. The first 15 exons of the gene, encoding for the N-terminal domains of the fibrillin-1 protein, are highlighted in dark grey, with the others represented in light grey. The exons are numbered from 1 to 65.

The N-terminal domains of fibrillin-1 have a role in directing the formation of dimer intermediates which aggregate to form the functional microfibril.²³ Clustering of the mutations identified in this study at the N-terminus may underline its different role when compared with the middle region of the protein. In our series of patients a relatively mild phenotype, with late onset of stable cardiovascular features in some cases has been observed. In contrast, mutations in the central region, which includes the neonatal cluster of exons 24–32, are usually associated with more severe phenotypes.¹⁰ Furthermore, it has been reported that mutations in the C-terminal end of the protein could also be linked to less severe phenotypes,²⁰ thus stressing the possibility that involvement of different regions of the FBN1 gene might be partially responsible for phenotype variability. Further studies of patients representing the mild end of the MFS spectrum will help to clarify this issue.

Meanwhile, since a tendency to late onset aortic dilatation and/or dissection is an occasional feature, it is recommended that patients with predominant EL be screened with echocardiography initially and at regular intervals throughout their lifetime.