Autosomal dominant inheritance of cardiac valves anomalies in two families: Extended spectrum of left-ventricular outflow tract obstruction^†

INTRODUCTION

Anomalies of the atrioventricular and semilunar heart valves and associated structures account for 25–30% of all congenital cardiovascular malformations (CVM) [Loffredo, 2000]. Most occur sporadically in a single patient without affected family members, and unassociated with other malformations (nonsyndromic). On the other hand, well-defined syndromes with autosomal dominant inheritance, such as Noonan syndrome (caused by PTPN11, KRAS, SOS1, or RAF1 mutations), and Alagille syndrome (caused by JAGGED1 and NOTCH2 mutations) are often associated with valve defects [McElhinney et al., 2002; McDaniell et al., 2006; Sznajer et al., 2007].

Only a limited number of families with clear monogenic inheritance of nonsyndromic forms of congenital valve defects have been described. Familial nonsyndromic valve anomalies often include either the left-sided heart valves (aortic and mitral valve) or the right-sided heart valves (tricuspid and pulmonary valve). Left-sided valve anomalies can be part of a spectrum of anomalies of the left-ventricular outflow tract referred to as LVOTO (left-ventricular outflow tract obstruction; also known as obstructive anomalies of the left-heart and aorta) [McBride et al., 2005; Wessels et al., 2005; Garg et al., 2006], but in some families they occur without other LVOTO anomalies [Rao et al., 1969; McDonald and Maurer, 1989; Menahem, 1990]. In two families with LVOTO anomalies including bicuspid aortic valve (BAV) with calcification, Garg et al. 2006 documented truncating mutations in NOTCH 1. Two patients in these families also exhibited right-sided heart malformations, including double outlet right-ventricle and tetralogy of Fallot. Bicuspid aortic valve underlies the majority of patients with aortic valve disease and familial BAV has been described in single families suggesting autosomal inheritance [Emanuel et al., 1978; Glick and Roberts, 1994; Clementi et al., 1996; Huntington et al., 1997]. Studies on heritability of BAV support that genetic factors play a major role in BAV and demonstrate that BAV is often associated with other cardiovascular malformations, in particular LVOTO anomalies and thoracic aortic aneurysm (TAA) [Cripe et al., 2004; McBride et al., 2005; Loscalzo et al., 2007]. Locus heterogeneity for familial BAV has been established in several studies [Goh et al., 2002; Ellison et al., 2007; Martin et al., 2007], and genome-wide scans in families with BAV and/or associated CVM has demonstrated linkage to chromosomes 15q25-26, 18q, 5q, and 13q [Goh et al., 2002; Martin et al., 2007].

Another frequent left-sided heart valve anomaly, mitral valve prolapse, can be inherited as an autosomal dominant trait, and has been linked to chromosome 13 [Freed et al., 2003; Nesta et al., 2005]. Familial right-sided valve anomalies frequently represent syndromic forms of CVM and only a few nonsyndromic families are reported. Pulmonary stenosis (PS) is common in families with Noonan syndrome (PTPN11 mutations), Watson syndrome (NF1 mutations), and Alagille syndrome (JAGGED1 mutations). Mutations in PTPN11, however, have not been convincingly shown to be present in patients with nonsyndromic PS [Sarkozy et al., 2003], although a few cases of possible nonsyndromic PS with JAGGED1 mutations have been described [Krantz et al., 1999]. Nonsyndromic familial right-sided valve anomalies have only been described in a few small families. PS has been reported in some families with clear autosomal dominant inheritance [David, 1974; Ciuffo et al., 1985], and in some families with unknown mode of inheritance [Coblentz and Mathivat, 1952; Lamy et al., 1957; McCarron and Perloff, 1974; Klinge and Laursen, 1975; El-Said et al., 1979; Udwadia et al., 1996]. PS combined with ASD in a large pedigree has been shown to be due to a mutation in the GATA4 gene [Garg et al., 2003]. Familial pulmonic valve atresia and familial occurrence of tricuspid anomalies are very uncommon, and have only been described in a few families [DiChiara et al., 1980; Chitayat et al., 1992; Kumar et al., 1994; Grant, 1996; Grossfeld et al., 1997; Lin and Rosti, 1998; Bonnet et al., 1999].

Only a few families with combined left- and right-sided heart valve anomalies with monogenic inheritance have been described. A large family with autosomal dominant inheritance of mainly atrioventricular valve defects including Ebstein anomaly and atrioventricular canal has been described by Schunkert et al. 1997. Atrioventricular septal defects (AVSD) (also known as endocardial cushion defects) can be inherited as an autosomal dominant trait with variable expression and incomplete penetrance. These valve anomalies can be due to mutations in the gene encoding the cell adhesion molecule CRELD1 [Robinson et al., 2003], and a second locus is located on chromosome 1p31-p21. A few families with X-linked valvular dysplasia, a condition characterized by myxomatous degeneration, valvular regurgitation and secondary calcification affecting all four heart-valves have been described [Newbury-Ecob et al., 1993; Kyndt et al., 1998]. Recently mutations in the FLNA gene encoding filamin A were identified in these families [Kyndt et al., 2007].

The paucity of multiplex families with a clear Mendelian inheritance pattern of nonsyndromic cardiac valve malformation has precluded the identification of human genes specifically involved in cardiac valve morphogenesis and malformation. We describe two families with a similar autosomal dominant form of congenital heart malformation mainly consisting of cardiac valve anomalies.

CLINICAL REPORTS

Family 1

The pedigree of family 1 is shown in Figure 1.

Patient 1

The index patient (II-1) visited our Department of Clinical Genetics for genetic counseling. She had aortic valve replacement at the age of 42 because of a severely stenotic BAV. She received two new aortic valve prostheses in the following 30 years, and at the age of 60 she developed atrial fibrillation. Three of the five children of patient II-1 were healthy, and cardiologic examination including ECG and echocardiography revealed no abnormalities. Two other children (Patients 2 and 3) are affected.

Patient 2

One of the five children (III-4) of Patient II-1 was asymptomatic until he presented with progressive dyspnea at the age of 44. Echocardiography revealed a BAV, a dilated left-ventricle with thickening of the posterior wall, and left-ventricular dysfunction.

Patient 3

Another son (III-5) of Patient II-1 was diagnosed with a BAV with mild stenosis and regurgitation, and mild dilatation of the aorta at the age of 26 years. An X-ray of the thorax showed an elongated aorta. At the age of 36 years the aortic valve was stenotic, calcified and thickened, and mild tricuspid regurgitation was present. His left-ventricular function remained good. Chromosomal analysis showed a normal male karyotype. A microdeletion of the 22q11 (TBX1 gene) and 7q11.23 (elastin gene) region was excluded by fluorescent in situ hybridization (FISH).

Patient 4

A sister (II-2) of Patient II-1 was diagnosed with valvular aortic stenosis and regurgitation, mitral stenosis with regurgitation, and tricuspid regurgitation. She underwent three operation for aortic and mitral valve prostheses between the age of 40 and 44 years. At the age of 45 years a tricuspid valve correction was performed. She died at the age of 50. In four of her five children cardiologic examination including ECG and echocardiography revealed no abnormalities.

Patient 5

A daughter (III-7) of Patient II-2 was diagnosed with moderate mitral valve regurgitation at the age of 50 years.

Patient 6

A brother (II-4) of Patient II-1 was diagnosed with aortic valve regurgitation and received an aortic valve replacement.

Patient 7

The son (patient III-14) of Patient II-4 was diagnosed with aortic stenosis. Surgical correction was performed at the age of 9 years. He died 1-day after surgery.

Patient 8

One of the three daughters of Patient II-5 (patient III-18) underwent valvulotomy for a severely stenotic BAV at the age of 9 years. She also received a mitral valve prosthesis for severe mitral valve regurgitation.

Patient 9

The brother of Patient III-18 (patient III-21) was operated at the age of 5 years for a VSD and valvular PS. He developed a re-stenosis and pulmonary valve regurgitation.

Patient 10

A son (IV-1) of an asymptomatic sister of Patient III-18 was diagnosed with a perimembranous VSD at the age of 4 months.

MOLECULAR STUDIES

Linkage analysis using polymorphic microsatellites D9S1826, D9S158, and D9S1838 flanking the NOTCH1 gene was performed in both families. This excluded NOTCH1 as the disease gene in both families since multiple recombinants were found. Additionally, sequence analysis of all coding exons and intron–exon boundaries of the NOTCH1 gene was normal in one affected patient in each family.

DISCUSSION

We describe two families with autosomal dominant inheritance of isolated CVM mainly involving the cardiac valves. None of the affected family members showed signs of a connective tissue disorder or malformation syndrome. Among the 10 affected family members of the first family, 6 were diagnosed with an abnormal stenotic and/or insufficient aortic valve. In three patients, regurgitation of the mitral valve was present. Pulmonary valve abnormality was diagnosed in one family member. In the second family, abnormal semilunar valves were present in nine family members, seven with abnormal aortic valves, and two with defects of the pulmonary valves (Table I). Different cardiologists evaluated patients and BAV may not always be reported if present. In both families, septal defects were present in several patients, which we believe may be part of the spectrum since they co-existed with valve anomalies in several family members. In the first family, one patient had a VSD and PS. In the second family, an ASD was present in a patient with PS and in a patient with pulmonary atresia. No other congenital abnormalities or dysmorphic features were present in any of the patients, indicating that the CVM in these families is nonsyndromic. In both families, autosomal dominant inheritance is well supported since there are three affected generations with male–male inheritance and expression in both females and males. Nonpenetrance is present in one obligate carrier in both families. In family 2, patient II-2 was unaffected at the age of 43 years but showed moderate aortic valve regurgitation 14 years later at the age of 57. Patient II-8 showed no abnormalities at echocardiography at the age of 53. She has a tri-leaflet aortic valve and normal function of cardiac valves and the left-ventricle.

Autosomal dominant inheritance of nonsyndromic congenital valve anomalies has only been described in a limited number of families. In most cases consistence of valve anomalies of either predominantly left- or right-sided structures of the heart is present. In both presented families predominantly aortic valve anomalies were observed, although right-sided malformations such as pulmonary and tricuspid valve anomalies were present in some patients. This observation is also documented in other studies with smaller families (for instance only two persons affected) where patterns of inheritance are not always clear [Gill et al., 2003; Lewin et al., 2004; Martin et al., 2007]. In a recent study demonstrating high heritability of hypoplastic left-heart syndrome (HLHS) a high percentage of HLHS probands had both left- and right-sided valve dysplasia suggesting that HLHS is a severe form of valve malformation and anomalies of the left- and right-sided valves may have a common etiology [Hinton et al., 2007].

Human genes known to be involved in valvulogenesis include genes associated with elastogenesis and collagen synthesis, as elastin and collagen are major components of semilunar and atrioventricular valves. As mutations in most of these genes are associated with syndromic forms of CVM they are unlikely to be involved in our families with nonsyndromic CVM. Familial BAV and aortic valve stenosis in association with TAA is a well-recognized entity. Loscalzo et al. 2007 suggested that altered TGFB signaling might play a role in BAV with TAA as several aneurysm syndromes, including Marfan syndrome [Neptune et al., 2003], Loeys Dietz syndrome [Loeys et al., 2005], and arterial tortuosity syndrome [Coucke et al., 2006] are associated with upregulation of the TGFB pathway leading to loss of elastic fiber integrity. However, TGFBR1 and TGFBR2 gene analysis in 13 families with BAV and TAA revealed no mutations in [Loscalzo et al., 2007]. Recently mutations in the gene ACTA2, encoding vascular smooth muscle cell α- actin, were identified as a major cause of autosomal dominant inherited TAA. Interestingly multiple familymembers in 4 out of 14 described families with ACTA2 mutations showed BAV, indicating that genes encoding sacromeric proteins might be good candicate genes for a subset of familial LVOTO [Guo et al., 2007].

Elastin mutations in our families are unlikely as they predominantly cause supravalvular aortic stenosis, although aortic valve stenosis can also occur. Furthermore, right-sided valve anomalies and ASD/VSD are rarely described in patients with elastin mutations [Metcalfe et al., 2000; Eronen et al., 2002].

Only a few genes have been involved in nonsyndromic CVM with a monogenic mode of inheritance; these include the NOTCH1, Elastin, NKX2.5, NKX2.6, GATA4, CRELD1, MYH6, ACTA0, TBX20, and FLNA genes [for reviews: Bruneau, 2008; Ransom and Srivastava, 2007; Weismann and Gelb, 2007]. Mutations in NOTCH1 have been found in two families with BAV, aortic valve stenosis, aortic valve calcification and other LVOTO anomalies [Garg et al., 2006]. Interestingly one patient with AS and BAV also had ascending aortic dilatation. In both our families NOTCH1 was excluded as the disease gene by linkage analysis using polymorphic microsatellites (D9S1826, D9S158, D9S1838) flanking the NOTCH1 gene that showed multiple recombinants in both families. Additionally, sequence analysis of all coding exons and intron–exon boundaries was normal in an affected family member in both families.

Human NKX2.5 mutations can cause a number of different cardiac phenotypes [McElhinney et al., 2003; Elliott et al., 2003], including ASD/VSD and LVOTO, but anomalies of the semilunar valves, as observed in 17/21 of the patients in our families, are not often described [Majumdar et al., 2006]. A NKX2.5 mutation was excluded in both our families by mutation analysis in one affected family member. GATA4 mutations can lead to nonsyndromic ASD and other CVM including PS [Garg et al., 2003], whereas gross deletions of the 8p23 region encompassing the GATA4 gene are associated with a variety of cardiac anomalies, mainly PS and ASD. However, in contrast to our families left-sided cardiac anomalies are uncommon in these patients, although aortic/mitral regurgitation was reported in 1/18 patient of the families described by Garg et al. 2003. So far, no good candidate genes have been reported for our families, therefore a genome-wide linkage analyses has been initiated in both families.

Mature valve structures arise from endothelial cells of the endocardial cushions [Lincoln et al., 2004]. The endocardial cushions are formed by endothelial–mesenchymal transdifferentiation of a subset of endothelial cells that invade the extracellular matrix and differentiate into mesenchymal cells [Armstrong and Bischoff, 2004]. Valve leaflets eventually consist of a single endothelial cell layer and a central layer consistent of collagen, elastin, and glycosaminoglycans [Maron and Hutchins, 1974]. The endocardial cushion tissue contributes not only to the formation of valves, but also to the formation of membranous septa [Schroeder et al., 2003]. This might explain why some patients in our families have septal defects apart from valve anomalies.

In contrast to the sparse knowledge about the genes involved in human valve formation and malformation much more is known about valvulogenesis in mice. In mice signal transduction pathways including Wnt/β–catenin, Vegf, Notch, Bmp–Tgfβ, and Erb, and transcription factors including different GATA, FOX and SOX transcription factors have been implicated in heart cushion/valve formation (Table II, Fig. 2). These different signaling pathways exhibit extensive cross-talking, resulting in a complex integrated process of cardiac valve morphogenesis [for review: Schroeder et al., 2003; Armstrong and Bischoff, 2004]. These mouse models could provide functional candidate genes for families with cardiac valve anomalies once positional genetics approaches have localized the human disease gene.

The two multiplex families in this report may facilitate identification of human genes specifically involved in cardiac valve morphogenesis and aortic wall disease. The co-occurrence of aortic and pulmonary valve abnormalities and aortic aneurysms in these families supports a common genetic etiology in some forms of left- and right-sided valve anomalies, and expands the phenotype of LVOTO.