Marfan Syndrome

A number sign (#) is used with this entry because all cases of the Marfan syndrome appear to be due to heterozygous mutation in the fibrillin-1 gene (FBN1; 134797) on chromosome 15q21.

Description

A heritable disorder of fibrous connective tissue, Marfan syndrome shows striking pleiotropism and clinical variability. The cardinal features occur in 3 systems--skeletal, ocular, and cardiovascular (McKusick, 1972; Pyeritz and McKusick, 1979; Pyeritz, 1993). It shares overlapping features with congenital contractural arachnodactyly (121050), which is caused by mutation in the FBN2 gene (612570).

Gray and Davies (1996) gave a general review. They published Kaplan-Meier survival curves for a cohort of British Marfan syndrome patients demonstrating greater survivorship in females than in males; a similar result had been reported by Murdoch et al. (1972) and by Silverman et al. (1995). Gray and Davies (1996) also proposed a grading scale for clinical comparison of the Marfan syndrome patients. The authors provided criteria for each grade and suggested uniform use of these scales may facilitate clinicomolecular correlations.

Clinical Features

Increased height, disproportionately long limbs and digits, anterior chest deformity, mild to moderate joint laxity, vertebral column deformity (scoliosis and thoracic lordosis), and a narrow, highly arched palate with crowding of the teeth are frequent skeletal features. Sponseller et al. (1995) evaluated spinal deformity in 113 patients with Marfan syndrome, 82 of whom were skeletally immature. Scoliosis was found in 52 of the 82 patients, with equal prevalence for the sexes. The thoracic portion of the curve was convex to the right in all but 2 patients.

Westling et al. (1998) quantitated the craniofacial morphology through evaluation of dental casts and lateral cephalograms in 76 Marfan patients. About 70% of the patients had been referred for orthodontic treatment, mostly because of crowded teeth or extreme maxillary overjet (overbite). The appearance of the palatal vault, crowded teeth, and overbite were illustrated.

Myopia, increased axial globe length, corneal flatness, and subluxation of the lenses (ectopia lentis) are ocular findings (Pyeritz and McKusick, 1979). Of 573 patients with Marfan syndrome, Izquierdo et al. (1994) found that 110 (19.2%) had strabismus; exotropia occurred in 67 (11.7%), esotropia in 12 (2.1%), vertical deviations in 8 (1.4%), and primary inferior oblique muscle overaction in 3 (0.5%).

Mitral valve prolapse, mitral regurgitation, dilatation of the aortic root, and aortic regurgitation are cardiovascular features. The major life-threatening cardiovascular complications--aneurysm of the aorta and aortic dissection--were reported in single cases by Baer et al. (1943) and Etter and Glover (1943). About one-third of affected persons have mitral valve prolapse, aortic root enlargement, or both on echocardiography, despite normal auscultatory findings on cardiac examination (Brown et al., 1975; Pyeritz and McKusick, 1979).

Other common or peculiar manifestations include striae distensae, pulmonary blebs (which predispose to spontaneous pneumothorax), and spinal arachnoid cysts or diverticula (Weir, 1973; Newman and Tilley, 1979; Cilluffo et al., 1981). By CT scanning, Pyeritz et al. (1988) found widening of the lumbosacral spinal canal (dural ectasia) in 36 of 57 patients with the Marfan syndrome and in none of 57 age- and sex-matched non-Marfan control patients. Severe changes were present in 13 patients, 2 of whom had associated neurologic signs, and included meningoceles or near-total erosion of a pedicle.

Yellin et al. (1991) described a father and 2 sons with multiple and bilateral episodes of pneumothorax.

Fattori et al. (1999) studied the frequency of dural ectasia in the Marfan syndrome, this feature having been classified as a major diagnostic criterion by Beighton et al. (1988) and De Paepe et al. (1996). MRI studies of the thoracic aorta and lumbosacral spine were performed in 83 patients with Marfan syndrome; 12 patients were younger than 18 years. The control group consisted of 100 individuals who underwent MRI of the lumbar spine for routine clinical indications; none of them had any potential causes for dural ectasia. Dural ectasia was identified in 76 (92%) patients and none of the control group. The severity of dural ectasia was related to age; the mean (SD) age of patients with mild dural ectasia was 26 years (14), whereas that of those with severe disease (meningocele) was 36 years (9) (p = 0.038). Dural ectasia was present in 11 of 12 patients younger than 18 years. No association was found between aortic dilatation and dural ectasia.

Ahn et al. (2000) studied dural ectasia in 32 volunteers with Marfan syndrome as diagnosed by the Ghent criteria (De Paepe et al., 1996). They concluded that MRI and CT diagnosed dural ectasia with high specificity and sensitivity, and suggested criteria for the diagnosis in adult Marfan patients.

Van den Berg et al. (1996) surveyed 135 patients with Marfan syndrome and found none that presented with the symptomatic intracranial aneurysm.

Morse et al. (1990) reported experience with 22 severely affected infants in whom the diagnosis of the Marfan syndrome was made in the first 3 months of life and reviewed the cases of 32 previously reported infants. They pointed out that serious cardiac abnormalities as well as congenital contractures may be present at birth. The most severely affected cases appear to be due to sporadic mutation; familial cases have milder manifestations which render them more difficult to detect during infancy.

Buntinx et al. (1991) described severe Marfan syndrome in a neonate with aortic dilatation and cardiac valve insufficiency who died from heart failure at 20 hours of age. Cardiomegaly and aortic dilatation had been noted at 35 weeks' gestation by means of routine ultrasonography. While the occurrence of contractures in the severe neonatal form raised a question about the distinctness of congenital contractural arachnodactyly (CCA; 121050) from Marfan syndrome, CCA has now been linked to the gene encoding fibrillin-2 (612570) on chromosome 5. The contractures tend to resolve with time.

Hanseus et al. (1995) described a case very similar to that of Buntinx et al. (1991): a newborn girl with extreme cardiomegaly discovered by fetal ultrasound at 34 gestational weeks. The girl died at the age of 10 hours. Postmortem examination showed cardiovascular lesions typical of Marfan syndrome. Immunofluorescence studies of cultured fibroblasts from the patient showed decreased amounts of immunostained fibrillin, supporting the clinical diagnosis of severe Marfan syndrome. Ruiz et al. (1996) described acute dissection of the aorta in a 5-year-old girl with Marfan syndrome. Surgical repair was successful. The patient's father had Marfan syndrome with previous surgery for aortic dissection. The mother made the diagnosis of acute dissection in her daughter.

The trisomy 8 syndrome (Pai et al., 1979) simulates the Marfan syndrome in its skeletal features. However, it does not show the ocular and aortic characteristics of the Marfan syndrome and does show unusual creases of the palms and soles and mental retardation which are not found in the Marfan syndrome. Most, if not all, cases of trisomy 8 have been mosaic, thus accounting for the relatively mild manifestations of trisomy of a large chromosome. James and Jacobs (1996) determined the parental origin in 8 cases of constitutional trisomy 8. In all 4 cases of spontaneous abortion, the additional chromosome was maternal in origin and there was evidence for nullichiasmate meiosis I as the basis of the trisomy. In contrast, all 4 cases of liveborn trisomy 8 studied appeared to have arisen by a mechanism consistent with the postzygotic mitotic gain of the additional chromosome, a mechanism consistent with mosaicism.

In a survey of adult genetic skeletal dysplasias in the Museum of Pathological Anatomy in Vienna, Beighton et al. (1993) pictured the skeleton of a man who died of pulmonary tuberculosis at the age of approximately 35 years. A detailed appraisal of the skeleton had been published in the pathology literature in 1904 as an example of macrosomia. The height was 191 cm, the hand length was 20 cm, and the upper segment/lower segment ratio was 0.82. The diagnosis was clearly Marfan syndrome.

Cistulli and Sullivan (1995) suggested that the Marfan syndrome is associated with a high prevalence of obstructive sleep apnea. They measured upper airway closing pressures during sleep in 12 patients and 6 age-, height-, and weight-matched controls. Obstructive sleep apnea was found in 10 of the patients. All 12 patients with the Marfan syndrome, including the 2 without obstructive sleep apnea, demonstrated increased upper airway collapsibility during sleep. In contrast, only 2 control subjects demonstrated upper airway closure, and that only at significantly higher suction pressures. Verbraecken et al. (1995) reported obstructive sleep hypopnea syndrome in a 35-year-old woman who had Marfan syndrome. They commented that 'such findings are not obvious because Marfan patients have a rather low weight and a tall and slender neck.' They presumed that an increased laxity of the pharyngeal wall was responsible for the complication.

Verbraecken et al. (2003) described a patient with Marfan syndrome and coexistent obstructive sleep hypopnea and restrictive lung disease, complicated by respiratory insufficiency. Successful treatment with nasal intermittent positive airway pressure and oxygen appeared to 'attenuate' the progressive dilatation of the aortic root, but did not reverse it. Verbraecken et al. (2003) hypothesized that a decrease of nocturnal intrathoracic pressures could at least partially explain the observation.

Murdoch et al. (1972) reported that life span in patients affected with the Marfan syndrome is markedly shortened and that most deaths are cardiovascular. Silverman et al. (1995) reported a study of survival in 417 patients from 4 referral centers with a definite diagnosis of the Marfan syndrome. They concluded that life expectancy had increased more than 25% since 1972 and suggested that the reasons for the increase might include (1) an overall improvement in population life expectancy; (2) benefits arising from cardiovascular surgery; (3) a greater proportion of mild cases due to increased frequency of diagnosis; and (4) medical therapy, including beta-blockers.

Gray et al. (1998) studied 206 Marfan syndrome patients ascertained through genetics clinics in Wales and Scotland during the period 1970 to 1990. Median survival was 53 years for males and 72 years for females. Mean age at death was 45.3 +/- 16.5 years. Severity was the best independent indicator of survival. The study was not able to assess the efficacy of beta-blockade.

Parida et al. (1997) described an infant with neonatal Marfan syndrome and a large hiatus/paraesophageal hernia with pronounced gastroesophageal reflux.

Whitelaw et al. (2004) reported 2 unrelated infant girls with neonatal Marfan syndrome and primary trabeculodysgenesis; FBN1 mutations were confirmed in both cases.

Tekin et al. (2007) identified a heterozygous mutation in the FBN1 gene (134797.0046) in a male infant with severe neonatal Marfan syndrome. Two older brothers were similarly affected, and all 3 sibs died at ages 2 to 4 months of cardiorespiratory insufficiency. Mosaicism for the mutation was identified in somatic cells and germ cells of the clinically unaffected father. Tekin et al. (2007) stated that this was the first report of familial neonatal Marfan syndrome.

Voermans et al. (2009) evaluated 10 patients with Marfan syndrome specifically for neuromuscular features. Four older patients had muscle weakness, 5 had mild to moderate reduction in vibration sense, and all older patients mentioned mild functional impairment compared to controls. Neurophysiologic studies showed axonal neuropathy in 4, and myopathic and neurologic changes in all. Imaging showed lumbosacral dural ectasia in 7 patients. Voermans et al. (2009) emphasized that neuromuscular features may be present in Marfan syndrome, particularly in older patients, and noted that muscle hypoplasia and myopathy had been reported by Marfan himself as a feature of the syndrome.

Lundby et al. (2012) found that 47 of 87 (54%) patients with Marfan syndrome had widening of the trunk of the pulmonary artery (30 mm or more) as determined by magnetic resonance or computed tomography imaging. Of these 47, 15% had no sign of disease of the ascending aorta.

Stheneur et al. (2014) studied 259 children who carried an FBN1 gene mutation and fulfilled Ghent criteria by comparing them with 474 non-Marfan syndrome children. The authors found that prevalence of skeletal features changed with aging: prevalence of pectus deformity increased from 43% at 0 to 6 years to 62% at 15 to 17 years, wrist signs increased from 28 to 67%, and scoliosis increased from 16 to 59%. Hypermobility decreased from 67 to 47% and pes planus decreased from 73 to 65% over the same interval. Striae increased from 2 to 84%. Prevalence of ectopia lentis remained stable, varying from 66 to 72%, similar to aortic root dilatation, which varied from 75 to 80%. When comparing Marfan syndrome children with non-Marfan syndrome children, height appeared to be a simple and discriminant criterion when it was greater than 3.3 standard deviations above the mean. Ectopia lentis and aortic dilatation were both similarly discriminating. Stheneur et al. (2014) found that mean aortic dilatation remained stable during follow-up in this population receiving beta-blocker therapy.

Biochemical Features

A number of abnormalities of connective tissue proteins or ground substance have been suggested as the biochemical basis of this disorder (Pyeritz and McKusick, 1981). Evidence has been presented for abnormalities of collagen primary structure (Byers et al., 1981) and crosslinking (Boucek et al., 1981) and of hyaluronic acid synthesis (Appel et al., 1979). See 120160.0020 for description of a mutation in the alpha-2 gene of type I collagen in a possible case of atypical Marfan syndrome.

Appel et al. (1979) showed that cell-free extracts of Marfan fibroblasts had 3 to 10 times more hyaluronic acid synthetase activity than did preparations from normal fibroblasts. No changes in the properties of this microsomal enzyme were found. In a Finnish family in which members of at least 3 generations showed the Marfan syndrome in somewhat atypical form, Pulkkinen et al. (1987) found evidence suggesting a defect in the alpha-1 chain of type I procollagen. Fibroblasts from the patients showed accumulation of collagen-specific tritium-labeled hydroxyproline. An extra band comigrating with the alpha-1(I) collagen chain of control cells could be detected intracellularly. Pulkkinen et al. (1987) suggested that a defect in the COL1A1 gene was the defect in this family. Atypical findings in this family included the occurrence of several stillbirths and spontaneous abortions.

By indirect immunofluorescence, Godfrey et al. (1990) demonstrated a consistent apparent deficiency of elastin-associated microfibrillar fibers in Marfan skin. Furthermore, there was deficient accumulation of related fibrous materials in cultures of Marfan fibroblasts. Both sets of observations were controlled with observations in normal subjects and patients with other heritable disorders of connective tissue. In 3 families, at least 1 affected individual in each of 2 generations showed the abnormality. In all cases, the abnormality of immunofluorescence cosegregated with the Marfan phenotype and unaffected family members were normal. The findings directed attention toward fibrillin, a glycoprotein of the microfibrillar component of the elastic fiber system (Hollister et al., 1985; Sakai et al., 1986).

McGookey et al. (1990) studied the synthesis, secretion, and incorporation of fibrillin into extracellular matrix by fibroblasts from 21 probands with the Marfan syndrome. The probands separated into 4 groups: (1) those whose cells synthesized about half the amount of fibrillin when compared to controls but secreted the protein normally; (2) those whose cells synthesized the same amounts of fibrillin as control cells but secreted the fibrillin less efficiently; (3) those whose cells synthesized and secreted fibrillin normally but did not incorporate the molecules into extracellular matrix; and (4) those in whom no abnormality in synthesis, secretion, or matrix incorporation could be detected. Family studies showed that all affected individuals from the same family had the same biochemical defect; unaffected family members showed no biochemical defect. The authors hypothesized that fibrillin is the major candidate protein responsible for the Marfan syndrome and that the majority of individuals with the syndrome have mutations affecting the synthesis, secretion, or matrix incorporation of the molecule. In the full report, Milewicz et al. (1992) presented studies of dermal fibroblasts from 26 patients. Cells from 7 patients synthesized approximately half the normal amount of fibrillin. Cells from an additional 7 patients synthesized a normal amount of fibrillin but secreted the protein less efficiently than control cells. In a further 8 probands, cells synthesized and secreted normal amounts of fibrillin but the protein was poorly incorporated into extracellular matrix. In only 4 probands were the synthesis and processing of fibrillin indistinguishable from those in control cells.

Superti-Furga et al. (1992) studied the synthesis of fibrillin and decorin in cultured fibroblasts of a patient with neonatally lethal Marfan syndrome. Fibrillin immunoreactivity in the extracellular matrix was markedly reduced and the fibrillar pattern was absent, in spite of normal amounts of fibrillin mRNA. Decorin mRNA, synthesis, and immunoreactivity were also reduced in the matrix. They suspected that a defect in fibrillin was primary and that the underexpression of decorin was secondary. The combination, however, may lead to the severe clinical phenotype. See 125255 for evidence of abnormality in the expression of decorin in the Marfan syndrome.

Inheritance

Variable expression in Marfan syndrome is the rule, but complete nonpenetrance has not been definitively documented. About one-quarter of affected individuals arise as new mutations; a paternal age effect is present, on average, in sporadic cases. Capotorti et al. (1959) described an Italian kindred in which 16 members of 3 generations showed the Marfan syndrome. The kindred contained an instance of marriage of affected first cousins. Of their 9 children, 4 were affected, 1 was normal, 3 died in infancy, and 1 was stillborn. Two of the 4 living affected children showed more severe manifestations than any of the other patients in the pedigree. Some members of this sibship may have been homozygous for the Marfan gene.

Homozygosity was also suggested by the families reported by Chemke et al. (1984) and Schollin et al. (1988). Karttunen et al. (1994) demonstrated that the presumably homozygous child in the family reported by Schollin et al. (1988) was, in fact, a compound heterozygote. The father had a trp217-to-gly mutation in heterozygous form, whereas the mother was heterozygous for a gly2627-to-arg substitution and the child had inherited both of these mutations. The infant had multiple skin folds all over the body at birth; at 1 week of age, eye examination demonstrated bilateral lens luxation. The baby had long limbs and died at the age of 4 months because of congestive heart failure. The father had ocular, skeletal, and cardiac manifestations as well as a positive family history; the mother had tall stature and myopia and her mother had died of a ruptured aortic aneurysm at the age of 51 years.

Burgio et al. (1988) described a girl in whom the clinical expression of the Marfan syndrome was limited mainly to the left side of the body. Subluxation of the lens was present only on the left, and the limbs were substantially longer on the left. The authors interpreted these findings as indicating postzygotic mutation. A mosaic condition may have arisen as a consequence of an early somatic mutation affecting 1 cell out of a very few, or it may have arisen through a single strand (half-chromatid) mutation, which, after the first mitotic division of the zygote, would lead to 1 normal cell and 1 cell carrying the mutation (Lenz, 1975). HLA typing of fibroblasts appeared to exclude chimerism, i.e., origin of the individual from 2 different zygotes, as the mechanism. Differences between the 2 sides of the body were demonstrated in immunohistopathologic studies of elastin-associated microfibrils (Hollister, 1988). Godfrey et al. (1990) reported in full on these histopathologic findings and also presented a clinical photograph of the patient who showed the longer left arm and leg with arachnodactyly on the left. Steinmann (1996) called attention to a possible similar case described as hemidolichostenomelia, or unilateral dolichostenomelia, reported by Maffei (1938).

Buntinx et al. (1991) reviewed cases of severe early-onset Marfan syndrome. Most of the cases have been sporadic and probably heterozygous.

Fried and Krakowsky (1977) reported a family as representing 'probable autosomal recessive Marfan syndrome': 2 girls with nonconsanguineous and ostensibly unaffected parents were affected.

De Vries et al. (2007) described what they considered to be the first family in which autosomal recessive inheritance of Marfan syndrome was confirmed molecularly. Two cousins from a consanguineous Turkish family were affected. The proband was a 22-year-old male with ectopia lentis, high-arched palate, pectus excavatum, spontaneous pneumothorax, mild scoliosis, and aortic aneurysm with dissection. His 13-year-old female cousin had ectopia lentis, high-arched palate, and lumbosacral ductal ectasia. Her aorta was normal. Neither cousin had striae or joint laxity, and thumb and wrist signs were negative. Both cousins were homozygous for an arg485-to-cys mutation (R485C; 134797.0047). All 4 healthy parents were heterozygous for the mutation, and none fulfilled the Ghent criteria for Marfan syndrome.

Mapping

Before the Marfan syndrome locus was definitively mapped to chromosome 15, Mace (1979) had reported a low positive lod score (1.17 at theta 0.30) for linkage with Rh.

Although the patient reported by Byers et al. (1981) appeared to have a change in the structure of the alpha-2 chain of type I collagen (COL1A2; see 120160), later studies excluded linkage with several collagen genes, including COL1A2. Borresen et al. (1985) studied 3 RFLPs at the COL1A2 locus in a large 4-generation family in which 21 persons had classic manifestations of Marfan syndrome. No linkage with COL1A2 could be found. Tsipouras et al. (1986) and Ahti et al. (1987) also reported observations excluding linkage with COL1A2. Dalgleish et al. (1987) demonstrated that neither the COL1A2 nor the COL3A1 (120180) gene is the site of the mutation responsible for the Marfan syndrome; in a large kindred, the Marfan syndrome segregated independently of RFLP markers of the 2 loci. Ogilvie et al. (1987) found that Marfan syndrome segregated independently of 4 major fibrillar collagen genes: COL1A1 (120150), COL1A2, COL2A1 (120140), and COL3A1. By means of RFLPs of the respective genes and the demonstration of recombinants, Francomano et al. (1988) also excluded COL1A2, COL2A1, and COL3A1 as the site of the mutation.

Kainulainen et al. (1990) demonstrated nonlinkage of Marfan syndrome to 5 genes on 2q that code for connective tissue components: fibronectin (135600), COL6A3 (120250), elastin (130160), COL3A1, and COL5A2 (120190).

On the basis of data presented at a workshop, Blanton et al. (1990) constructed an exclusion map for linkage of the Marfan syndrome.

By linkage analysis in 5 families, Kainulainen et al. (1990) mapped the Marfan mutation to chromosome 15. Three polymorphic markers gave a composite lod score of 3.92 at theta = 0.0. A lod score of 3.32 at theta = 0.0 +/- 0.12 was obtained with D15S45, which they had located in band q22.1.

In 4 large multiplex American families with classic Marfan syndrome, Dietz et al. (1991) confirmed the assignment of the MFS1 gene to chromosome 15 but established a more centromeric location, defined by markers D15S25 and D15S1; lod score = 12.1 at theta = 0.00. They suggested the location 15q15-q21.3.

From linkage studies, Tsipouras et al. (1991) concluded that the Marfan syndrome locus may be located between markers D15S48 and D15S49.

Magenis et al. (1991) and Lee et al. (1991) mapped the fibrillin gene to chromosome 15q21.1 by in situ hybridization. Linkage studies using polymorphic markers within the fibrillin locus demonstrated tight linkage between the Marfan phenotype and fibrillin (Lee et al., 1991; Dietz et al., 1991). Dietz et al. (1991) demonstrated the first fibrillin mutation in a patient with Marfan syndrome, thereby completing the proof that Marfan syndrome is caused by mutations in the fibrillin locus on chromosome 15.

In studies of chromosome 15 markers in 17 families from 5 different populations, Kainulainen et al. (1991) found no evidence inconsistent with the mutation being on chromosome 15 in all. Similarly, the large international collaborative study reported by Tsipouras et al. (1992) found no evidence for genetic heterogeneity in Marfan syndrome. Twenty-eight families were studied. Two families with presumably isolated ectopia lentis (129600) were also found to show linkage to chromosome 15, but 1 family with annuloaortic ectasia (607086) did not show linkage to the chromosome 15 fibrillin gene or to the fibrillin gene on chromosome 5. In a letter received February 4, 1992, Hayward et al. (1992) stated that of the 2 fibrillin mutations discovered up to that time, arg1137-to-pro had been found twice in 111 cases and cys2307-to-ser had been found once in 140 cases (mutations being scored once for each sporadic case and once for each family segregating the gene). They concluded that there are likely to be many different FBN1 mutations responsible for the Marfan syndrome.

Sarfarazi et al. (1992) reported on a study, involving 22 families with 225 affected persons, by an international consortium. One family in the study, that reported by Boileau et al. (1992, 1993), did not show linkage to loci on 15q. In this kindred (see LDS2, 610168), Collod et al. (1994) demonstrated linkage to markers in 3p25-p24.2 and they suggested that a second form of Marfan syndrome is located in that region. Subsequently, affected members of this family were shown to carry a mutation in the TGFBR2 gene (190182.0004).

Molecular Genetics

For a discussion of the molecular genetics of the Marfan syndrome, see the entry on fibrillin-1 (FBN1; 134797).

Genotype/Phenotype Correlations

Attias et al. (2009) compared clinical features and outcomes of 71 patients with TGFBR2 (190182) mutations to those of 243 patients with FBN1 mutations. Aortic dilation was present in a similar proportion of patients in both the TGFBR2 and FBN1 groups (78% and 79%, respectively) but was highly variable; the incidence and average age for thoracic aortic surgery and aortic dissection were also similar in the 2 groups. Mitral valve involvement was less frequent in the TGFBR2 than in the FBN1 group (p less than 0.05 for myxomatous valve, prolapse, or mitral regurgitation). Aortic dilation, dissection, or sudden death was the index event leading to genetic diagnosis in 65% of families with TGFBR2 mutations versus 32% with FBN1 mutations (p = 0.002). The rate of death was greater in TGFBR2 families before diagnosis, but similar once the disease was recognized. Most pregnancies were uneventful in both groups. Seven (10%) of the 71 patients with TGFBR2 mutations fulfilled the Ghent criteria for Marfan syndrome, including 2 with ectopia lentis, compared with 140 (58%) of 243 patients in the FBN1 group (p less than 0.0001); 3 patients in the TGFBR2 group fulfilled the diagnostic criteria for both Loeys-Dietz (see 610168) and Marfan syndromes. Noting that clinical outcomes were similar between treated patients from both groups, Attias et al. (2009) concluded that prognosis depends on clinical disease expression and treatment rather than simply the presence of a TGFBR2 mutation.

Using DNA samples from 300 patients with clinical features of MFS or a related phenotype that had been previously screened by DHPLC with no mutations found in the FBN1 gene, Hilhorst-Hofstee et al. (2011) performed multiplex ligation-dependent probe amplification (MLPA) and identified 9 patients from 5 families with deletion of 1 entire FBN1 allele. A tenth patient with complete deletion of FBN1 was identified by cytogenetic analysis and array CGH. All of the patients had facial and skeletal features of MFS, and 7 of the 10 patients fulfilled the Ghent criteria; the 3 patients who did not present the full clinical picture of MFS were young (5, 8, and 13 years of age, respectively). Aortic root dilation was present in 6 patients, 2 of whom underwent surgical repair at relatively young ages. Hilhorst-Hofstee et al. (2011) concluded that complete loss of 1 FBN1 allele does not predict a mild phenotype, and that their findings support the hypothesis that true haploinsufficiency can lead to the classic phenotype of Marfan syndrome.

Pathogenesis

For a detailed discussion of fibrillin-1 defects and their proposed pathogenic mechanism in the Marfan syndrome, see FBN1.

Abraham et al. (1982) suggested that aortic elastin has an abnormality, presumably primary in some cases of the Marfan syndrome. In 3 typical cases, they found reduced desmosine and isodesmosine (about one-half of control values) and a corresponding increase in lysyl residues. Also, the concentration of elastin per mg dry weight was reduced. The hydroxyproline content of elastin was increased in 2. Alkali treatment solubilized 46.2% of an elastin preparation in Marfan aortas compared with 23.7% of controls. The concentration and solubility of collagen were unchanged and the amino acid composition and genetic types of insoluble collagen were normal.

Reviewing the nature of the ocular zonule, Streeten (1982) pointed out that the zonular fibers closely resemble the microfibrils of elastic tissue in their staining characteristics, ultrastructural morphology, and amino acid composition. Furthermore, they share antigenic determinants with the microfibrils of elastic tissue. The dolichostenomelia and arachnodactyly, as well as the pectus excavatum and pectus carinatum, appeared to represent excessive longitudinal growth of tubular bones in the limbs, fingers, and ribs. An unleashing of the normal control on longitudinal growth as a result of a defect in the fibrous elements of the periosteum had been postulated (McKusick, 1956). When fibrillin was discovered (Sakai et al., 1986), its presence in suspensory ligament, aorta, and periosteum prompted Hollister (1988) to suspect its implication in the Marfan syndrome.

Aoyama et al. (1995) studied dermal fibroblasts from 57 patients with classical Marfan syndrome, 15 with equivocal MFS, 8 with single-organ manifestations, and 16 with other connective tissue disorders, including homocystinuria and Ehlers-Danlos syndrome (225310). Abnormal fibrillin metabolism was identified in 70 samples that were classified into 4 different groups based on quantitation of fibrillin synthesis and matrix deposition. Significant correlations were found for phenotypic features, including arachnodactyly, striae distensae, cardiovascular manifestations, and fibrillin groups II and IV, which included 70% of the MFS patients. In addition, these 2 groups were associated with 'shortened event-free' survival and more severe cardiovascular complications than groups I and III. Group III included most of the equivocal MFS/single manifestation patients with fibrillin abnormalities. In this study, a decrease in synthesis was defined as fibrillin production of less than 70% of normal. Likewise, decreased deposition was defined as less than 70% of fibrillin normally incorporated into the extracellular matrix. Group I was characterized by reduced synthesis and deposition between 35% and 70% of control values, in contrast to group II with values less than 35%. Samples in group III had normal synthesis, but fibrillin deposition values were between 35% and 70%, while in group IV normal synthesis was accompanied by severely reduced deposition to less than 35% of controls. The results were interpreted as indicating that fibrillin defects at the protein level per se are not specific for Marfan syndrome, but that the drastically reduced fibrillin deposition, caused by a dominant-negative effect of abnormal fibrillin molecules in persons defined as group II and IV, is of prognostic and possibly diagnostic significance.

Downing et al. (1996) described the NMR-derived solution structure of a covalently linked pair of calcium-binding epidermal growth factor (cbEGF)-like domains from human fibrillin-1. The 2 domains are in a rigid, rod-like arrangement, stabilized by interdomain calcium-binding and hydrophobic interactions. They proposed a model for the arrangement of fibrillin monomers in microfibrils that reconcile structural and antibody binding data, and they described a set of disease-causing mutations that provide the first clues to the specificity of the interactions of the calcium-binding EGF domains.

In a panel of 60 Marfan probands, Liu et al. (1996) identified 6 exon-skipping mutations. All skipped exons encoded cbEGF-like domains and maintained the reading frame. They demonstrated that the mutant mRNA missing an in-frame exon is stable and is only slightly shorter than normal fibrillin. They also demonstrated that both the normal and the abnormal forms are secreted and both forms of fibrillin molecules participate in the formation of microfibrils. Liu et al. (1996) concluded that the abnormal fibrillin forms have a dominant-negative effect.

Palz et al. (2000) analyzed the terminal 7 exons (exons 59-65) of the FBN1 gene in 124 unrelated patients with MFS and identified 5 novel mutations. Their results, together with the findings in a review of mutations by Collod-Beroud et al. (1998), showed that 40% (8/20) of the FBN1 mutations in exons 59-65 are associated with a mild phenotype characterized by a lack of aortic root pathology. In contrast, only 7% of mutations reported in the remainder of the gene resulted in a mild phenotype with no aortic root pathology, even when the mutation occurred in the same position in a cbEGF motif. Palz et al. (2000) suggested that the position of the affected motif, and perhaps the nature of the substituting residue, are of greater importance.

Tiecke et al. (2001) analyzed exons 24-40 of the FBN1 gene by temperature-gradient gel electrophoresis in 124 unrelated patients with Marfan syndrome and identified 12 probable disease-causing mutations, 10 of which were novel. A recurrent mutation (134797.0036) in exon 24 was found in 2 unrelated patients with atypically severe clinical manifestations. Their results, together with those in other reports, showed that 12 of 14 missense mutations in patients with atypically severe MFS clustered in exons 24-32, suggesting a critical functional role for this region. Atypically severe MFS was characterized by cardiovascular complications requiring surgery in childhood as well as by abnormal face and ears, with or without congenital contractures. Missense mutations associated with neonatal MFS were found primarily in exons 25 and 26. Despite the clustering of mutations associated with neonatal and atypically severe MFS, mutations associated with classic MFS occurred in the same region. Based on these findings, Tiecke et al. (2001) concluded that there was no way of predicting whether a given mutation in exons 24-32 would be associated with classic, atypically severe, or neonatal MFS.

A distinct subgroup of individuals with Marfan syndrome have distal airspace enlargement, historically described as emphysema, which frequently results in spontaneous lung rupture (pneumothorax). Neptune et al. (2003) showed that mice deficient in fibrillin-1 have marked dysregulation of transforming growth factor-beta (TGFB1; 190180) activation and signaling, resulting in apoptosis in the developing lung. Perinatal antagonism of TGF-beta with a TGF-beta-neutralizing antibody attenuated apoptosis and rescued alveolar septation in vivo. These data indicated that matrix sequestration of cytokines is crucial to their regulated activation and signaling and that perturbation of this function can contribute to the pathogenesis of disease. They raised the possibility that the pathogenetic mechanism uncovered by these experiments might underlie other manifestations of Marfan syndrome that are not easily reconciled by loss of mechanical tissue integrity, including myxomatous changes of valve leaflets and bone overgrowth. Furthermore, selective pathologies may be amenable to treatments that attenuate excessive cytokine activity.

In cultured human dermal fibroblasts treated with recombinant fibrillin-1 fragments containing the RGD (arg-gly-asp) integrin-binding motif of fibrillin-1, Booms et al. (2005) observed significant upregulation of the matrix metalloproteinases MMP1 (120353) and MMP3 (185250). They suggested that fibrillin fragments themselves might have pathogenic effects by leading to upregulation of MMPs, which in turn might be involved in the progressive breakdown of microfibrils thought to play a role in MFS.

Hutchinson et al. (2003) described an FBN1 deletion patient (46,XXdel(15)(q15q22.1)) whose fibrillin-1 protein and mRNA levels were significantly higher than expected for a single FBN1 allele. RNA analyses identified a variable reduction in total FBN1 transcript (78 to 27%) in 3 related individuals carrying a premature termination codon (PTC)-causing mutation compared with unaffected control individuals. Both pulse-chase analysis of fibrillin-1 biosynthesis and RNase protection analyses demonstrated that these differences were due to variation in the expression of the normal FBN1 allele and not nonsense-mediated decay of mutant RNA. The authors suggested that differences in normal FBN1 expression may contribute to the clinical variability seen within families with Marfan syndrome.

Diagnosis

The diagnosis of the Marfan syndrome is based on typical clinical features--skeletal, ocular, and cardiovascular--and a positive family history when available. The criteria established by Beighton et al. (1988) state that in the absence of an unequivocally affected first-degree relative, one should require involvement of the skeleton and at least 2 other systems with a minimum of 1 major manifestation (ectopia lentis, aortic dilatation/dissection, or dural ectasia). In the presence of an unequivocally affected first-degree relative, one requires only that 2 organ systems be involved.

Schott (1992) questioned the usefulness of 'span greater than height' in the diagnosis of Marfan syndrome. He pointed out that equality of span and height, which allows man to be portrayed within a square, is a concept that dates from the Roman architect Vitruvius and was, of course, immortalized by Leonardo da Vinci. However, anthropometry shows that span exceeds height in most (59-78%) normal adult white males.

Godfrey et al. (1993) reported prenatal diagnosis by the linkage method in a 4-generation Marfan kindred. The diagnosis was made using CVS material at 11 weeks' gestation. At birth the infant showed skeletal changes suggestive of the Marfan syndrome. The mutation involved a donor splice site in the FBN1 gene (134797.0014). In a pregnant affected female in a third generation, Rantamaki et al. (1995) succeeded in the prenatal diagnosis by chorionic villus sampling.

De Paepe et al. (1996) proposed a revision of diagnostic criteria for Marfan syndrome and related conditions. Most notable were the following: more stringent requirements for diagnosis of Marfan syndrome in relatives of an unequivocally affected individual; skeletal involvement as a major criterion if at least 4 of 8 typical skeletal manifestations are present; potential contribution of molecular analysis to the diagnosis of Marfan syndrome; and delineation of initial criteria for diagnosis of other heritable conditions with partially overlapping phenotypes. Of most concern were the misdiagnoses of relatives that arose by relying solely on Berlin Nosology (Beighton et al., 1988) after unequivocal diagnosis of a first-degree relative (Pereira et al., 1994; Dietz et al., 1995). Molecular evidence showed that the criterion of a positive family history could produce a bias in favor of overdiagnosis. De Paepe et al. (1996) reproduced the diagrams for normal range of aortic root dimensions versus body surface area in children and adults, as reported by Roman et al. (1989).

Rose et al. (2000) compared the Berlin (Beighton et al., 1988) and Ghent (de Paepe et al., 1996) nosologies. In a study at the NIH they found that 19% of patients diagnosed under the Berlin criteria failed to meet the Ghent standard. They also found that dural ectasia was the second most common major diagnostic manifestation, and that screening for dural ectasia established the diagnosis of Marfan syndrome in 23% of patients under the Ghent criteria.

De Backer et al. (2006) assessed the prevalence of minor cardiovascular manifestations in Marfan syndrome to evaluate their usefulness in a diagnostic setting. Seventy-seven patients with Marfan syndrome, aged 4 months to 55 years, underwent echocardiography to assess the presence of mitral valve prolapse and the diameter of the main pulmonary artery. A subset of 29 adult patients with Marfan syndrome also underwent MRI evaluation of the diameters of the thoracoabdominal aorta. Mitral valve prolapse was encountered in 66% of patients with Marfan syndrome, with an equal distribution of classic and nonclassic mitral valve prolapse. The main pulmonary artery diameter was significantly larger in patients with Marfan syndrome at all ages when compared with controls. In the adult group (older than 14 years), De Backer et al. (2006) provided a cutoff value of 23 mm to define pulmonary artery dilatation. The descending aorta was enlarged, but with substantial overlap with controls, thus precluding the use of a cutoff value. De Backer et al. (2006) concluded that mitral valve prolapse and main pulmonary artery dilatation are common findings in Marfan syndrome patients at all ages and are easy to assess with echocardiography. De Backer et al. (2006) recommended echocardiographic evaluation of mitral valve prolapse and main pulmonary artery diameter in patients referred for cardiovascular diagnostic assessment for Marfan syndrome.

Loeys et al. (2010) reported the establishment of a revised Ghent nosology for the diagnosis of Marfan syndrome, which emphasizes cardiovascular manifestations and in which aortic root aneurysm and ectopia lentis are considered cardinal clinical features. In the absence of any family history, the presence of those 2 manifestations is sufficient for the unequivocal diagnosis of MFS. In the absence of either of those 2 features, the presence of an FBN1 mutation or a combination of systemic manifestations is required, for which there is a revised scoring system. Loeys et al. (2010) noted that in the revised nosology, FBN1 testing is not mandatory but carries greater weight in the diagnostic assessment.

Hennekam (2012) commented on 2 papers reviewing the distinction between the Ghent and revised Ghent nosologies for the diagnosis of Marfan syndrome, one by Aalberts et al. (2012) in a Dutch population and the other by Yang et al. (2012) in a Korean population, in which no significant differences were found in the diagnostic yield except for some slight improvement in specificity for mitral valve prolapse (see 157700) and MASS syndrome (604308), respectively.

Clinical Management

The early mortality