Normal and abnormal development of the aortic wall and valve: correlation with clinical entities

During organogenesis the first functional organ to form is the heart. The first sign of valvulogenesis is the formation of endocardial cushions in the atrioventricular canal and outflow tract. The atrioventricular cushions contribute to the atrioventricular (mitral and tricuspid) valve leaflets, whereas the outflow tract cushions contribute to the semilunar (aortic and pulmonary) valve leaflets [4]. Development of semilunar valves is a complex process in which neural crest cells, second heart field (SHF) progenitors and endocardial cushion derived cells play a role (Fig. 1). The developmental origin of the endocardial cushion cells themselves has been a matter of debate in the past years. Recent lineage tracing studies with Nkx2.5 [5] have shown that SHF progenitor cells give origin to three specific cell lines: 1. vascular smooth muscle cells (VSMCs) of the great arteries, 2. outflow tract and right ventricular myocardium, and to 3. the much discussed endothelial-derived endocardial cushion cells, which are in part derived from the endothelium [5]. Recently, Harmon et al. presented data on the boundary where SHF-derived VSMC meet neural crest cell-derived VSMC at the base of the aorta [5]. The SHF contribution to the aortic media then forms a vertical seam complementary with neural crest-derived VSMC [5]. Next to contributing to the vascular wall, a population of neural crest cells migrates to the outflow tract cushions where they are important for semilunar valve formation and outflow tract septation [6, 7]. Preliminary data show a contribution of the arterial epicardium to the VSMCs of the ascending aorta [8].

During valvulogenesis several signalling pathways such as Wnt/β-catenin, NOTCH, transforming growth factor β (TGF-β), bone morphogenetic protein (BMP), vascular endothelial growth factor (VEGF), NFATc1 and MAPK, as well as transcription factors, including Twist1, Tbx20, Msx1/2, and Sox9, are necessary for the regulation of cell migration, proliferation, and extracellular matrix deposition in the developing valves [9‐12].

As a consequence, more than one cell population that contributes to both aortic wall and semilunar valve formation may be involved in the development of bicuspidy. Clinically, several BAV subtypes are distinguished on the basis of the fused commissure or raphe position. Type 1: raphe between right coronary cusp (RCC) and left coronary cusp (LCC), type 2: between RCC and non-coronary cusp (NCC) and type 3: raphe between LCC and NCC (Fig. 2). Clinical outcome differs between the valve types, supporting a different developmental background as underlying cause. Recent support for the role of deficient neural crest cell contribution in development of type 1 BAV was seen in the Rock 1,2 deficient mouse [6]. Fernandez et al. argued that type 1 and type 2 BAVs have a different pathogenesis [13]. An altered neural crest cell behaviour was suggested to be responsible for the development of type 1 BAVs and the endothelial nitric oxide (eNOS) mutation for type 2 BAV. eNOS is expressed by endocardial cells [14], cardiomyocytes [14] and VSMCs [15], all SHF-derived cell types/populations, indicating a role for SHF progenitors in the development of type 2 BAV. Another role for SHF in the development of type 2 BAV was demonstrated by endocardial specific deletion of the gene encoding for the activin type I receptor (ALK2) [16].

Other genetic defects leading to BAV have been identified with selective knockout of genes in murine models but have not focused on the differentiation in type 1 or type 2 BAVs. From the recognised genes, Pax3 is a marker of neural crest cells [17]. Furthermore, the identified SHF markers Nkx2.5 and GATA5 (endocardial cell-specific) are associated with the development of BAV [18].

Abnormalities of NOTCH signalling in the neural crest [19] or SHF can also contribute to the development of abnormal semilunar valves [17, 20]. Interestingly, inhibition of NOTCH in SHF impairs fibroblast growth factor 8 (Fgf8) signalling, which results in the development of BAV, but also in VSMC abnormalities of the great arteries [21]. Therefore, we postulate that a developmental defect of various progenitor cell lines may provide a common mechanism underlying aortic valvulopathy (BAV), as well as aortopathy. The next section focuses on genetic defects described in BAV and TAD in human and their origin in embryogenesis.

Consistent data have suggested a genetic cause of BAV disease [22, 23]. Despite high heritability it remains challenging to determine the underlying mechanism of BAV in the human population, supported by murine data, as it is probably due to interacting mutations in diverse genes encoding transcription factors, extracellular matrix (ECM) proteins and signalling pathways that regulate cell proliferation, differentiation, adhesion or apoptosis.

Although a remarkable reduction in eNOS levels was seen in BAV patients, this could not be correlated to a mutation in the eNOS gene [24]. Moreover, mutations in the NOTCH1 gene, which is expressed by both neural crest and SHF, and mapped to chromosome 9q34, have been associated with the development and progression of BAV [19, 25]. Further genetic haplotypes within the AXIN1-PDIA2 locus have been recognised that strongly associate with BAV. AXIN1 (Axis Inhibitor 1) is a critical member of the Wnt pathway, which regulates both heart valve formation [26] and cardiac neural crest development [27]. Another haplotype within the Endoglin gene (known as a co-receptor in the TGFβ pathway) is required for differentiation of neural crest cells into VSMCs that populate the aorta [28].

From these genetic defects we conclude that there is a clear link to defects in neural crest as well as SHF-derived cell populations with elements for BAV. Alterations in neural crest signalling are associated with the most common type 1 BAV [29], identified as the valve type with most severe aortic wall abnormalities as compared with the other valve types [13, 30, 31] while SHF-related genes seem to correlate more with BAV type 2.

This brings us to the next important question whether TAD, not necessarily accompanied by BAV, is related to defects in neural crest signalling or alternatively, is there a more specific role for the SHF-derived cells? The following section concentrates on genetics of TAD in syndromes including Marfan, Ehlers-Danlos, Smad3 mutations and Loeys-Dietz and their link to embryonic development.

Marfan syndrome (MFS) is a connective tissue disorder characterised by cardiovascular, skeletal and ocular manifestations. The progressive dilation of the aortic root culminating in dissection is a major cause of morbidity and mortality in MFS patients. This syndrome is the result of a defect in the fibrillin-1 (FBN1) gene that localises on chromosome 15q21.1 and is inherited in an autosomal dominant manner [32]. A second locus for Marfan syndrome (MFS2) has been mapped to chromosome 3p25-24.2, and a heterozygous mutation in TGFΒR2 was subsequently identified as the genetic defect [33]. The TGFBR2 mutations in MFS patients involve the serine-threonine kinase domain and reduce TGFβ-induced receptor signalling.

Loeys-Dietz (LDS) is caused by another defect in the TGFβ signalling pathway. In this syndrome TGFBR1 and TGFBR2 mutations are mapped to chromosome 9q33-34 and 3p24-25 respectively. Cardiovascular lesions in LDS include aortic valve regurgitation and aortic root dilation, aneurysm formation and dissection. Other phenotypic characteristics include craniosynostosis, cleft palate, bifid uvula, congenital heart disease and mental retardation.

Another syndrome presenting with aneurysms, dissections and tortuosity throughout the arterial tree in association with mild craniofacial features and skeletal and cutaneous anomalies has recently been described by Van de Laar et al. [34]. The genetic locus has been mapped to chromosome 15q22.2-24.2 and shows that the disease is caused by mutations in SMAD3, essential for propagation of the TGF-β signal to the nucleus and activation of downstream gene transcription.

An additional syndrome worth mentioning is Ehlers-Danlos (EDS). Patients with vascular EDS often present with dissection or rupture of the thoracic aorta. Aortic dissections have been reported in at least 10 % of patients with EDS [35]. This syndrome is attributed to a mutation in the gene encoding type III procollagen (COL3A1), mapped to 2q24.3-q31 [36]. Type I and type III collagen are the most abundant collagen fibres found in the media and adventitia of the aortic wall. In addition to providing mechanical strength, collagen has other functional properties, including activation of intracellular signalling cascades, storage of soluble factors, such as IL-2, and regulation of their local activity [37].

Furthermore, TAD occurs in association with an autosomal dominant disorder in the absence of syndromic features, termed FTAAD (familial thoracic aortic aneurysm and dissection). A variety of genetic loci have been identified in this regard, such as TAAD1 locus, on chromosome 5q 13-14 [38] and TAAD2 locus on chromosome 3p24-25 [39], the mutant gene associated with this locus is TGFBR2. Only 5 % of investigated families have this mutation, suggesting a relatively rare cause [39]. TAAD3 locus on chromosome 15q24-26 [38, 39] and TAAD4 locus on chromosome 10q23-24 are also related to the development of aortic dissection. In the latter ACTA2 encoding for VSMC a-actin is identified [40]. The mutations impair the function of the VSMCs and this affects the integrity of the vessel wall, making it prone to dilation. This mutation has also been associated with an increased activity of the TGFβ pathway in the aorta [41]. In TAAD5 the identified gene is TGFBR1, mapped to chromosome 9q33-34 [42].

In most of the above TAD cases an increased TGFβ activity has been identified. However, TGFβ is not specific for neural crest, as it is also clearly involved in the endothelium and the SHF-derived cell populations [43]. In syndromes as MFS, EDS, and LDS, bicuspidy is not an obligatory clinical manifestation, indicating that a defective TGFβ signalling is at least not the main factor causing BAV formation. Neural crest defects seem to cause the most frequently occurring type 1 BAV which is associated with most marked complications of the aortic wall, and often with an increased TGFβ activity [44]. However, the clinical course is not complicated with TAD in all BAV patients. Thus, neural crest involvement and TGFβ activity together are not sufficient to explain the variability within the pathogenesis of BAV and associated aortopathy. Additional pathogenetic factors need to be taken into account such as haemodynamics or a contribution of SHF.

The morphology of the bi-leaflet valve produces a nonaxial transvalvular turbulent flow jet within the aortic root [45]. This turbulent flow, along with other haemodynamic factors, as an increased stroke volume (for instance aortic regurgitation), have been suggested to facilitate developing aortic complications, as the created abnormal biomechanics and helical flow alterations lead to an uneven wall stress distribution. However, several studies have confirmed that ascending aortic aneurysms can develop in the absence of valve abnormality [46]. Moreover, Yasuda et al. have reported development of aortic dilation after surgical repair of the diseased bicuspid aortic valve [47]. These studies suggest that structural wall abnormalities at the cellular level may be important for the onset of dilation. Therefore, haemodynamic factors alone are not sufficient to explain the pathogenesis of aortopathy associated with bicuspid aortic valves.

The alternative hypothesis is that genetically determined abnormalities of the aortic wall lead to a defect in the cellular microenvironment, causing or at least contributing to the aortic pathology and render the wall vulnerable to haemodynamic stress.

Several studies have focused on differences between the dilated aortic wall in BAV and TAV. Histopathological features of the aortic wall in BAV show decreased medial inflammation, elastin fragmentation and cystic medial necrosis, when compared with TAV [48]. In addition, the aortic wall has a different composition in BAV, with a significantly thicker tunica media but significantly thinner tunica intima [49]. In recent years research on ECM composition mainly established differences in the aortic media of dilated aortic wall in BAV and TAV [25, 38]. To take a step further in unravelling a possible different pathogenetic mechanism of TAD in BAV and TAV, we investigated non-dilated aortic walls of both valve types. The aortic media was specifically studied for maturation of VSMCs and ageing characteristics. We concluded that TAD in TAV has aspects of ageing, whereas in bicuspidy there is a defective smooth muscle cell differentiation (Fig. 3) unrelated to ageing. These results suggest that the fundamental difference in the aortic wall make-up of BAV is found in less differentiated VSMCs as compared with TAV while still both neural crest cells and SHF contribute to the VSMCs in the aortic wall. Haemodynamic factors might play a role in the aortic complications, but superimposed on the already present structurally immature aortic wall seen in BAV.