Chapter 480. Development of the Cardiovascular System

Cardiogenesis involves a precisely orchestrated series of molecular and morphogenetic events that combines cell types from multiple lineages. Subtle perturbations of this process can result in life-threatening congenital heart defects (CHDs). As an organ essential for life, the heart is the first organ to form and must support the rapidly growing embryo before it has the opportunity to become a four-chambered organ. The combination of the complex morphogenetic events necessary for cardiogenesis and the superimposed hemodynamic influences may contribute to the exquisite sensitivity of the heart to perturbations as reflected in the estimated 10% incidence of severe cardiac malformations observed in early miscarriages. The fraction of congenital heart malformations that are compatible with development composes the spectrum of CHD observed clinically in nearly 1% of live births.1 An additional 1% to 2% of the population harbor more subtle cardiac developmental anomalies that only become apparent as age-dependent phenomena reveal the underlying pathology. With more than 1 million survivors of congenital heart disease (CHD) in the United States, it is becoming apparent that genetic disruptions that predispose to developmental defects can have ongoing consequences in maintenance of specific cell types and cellular processes over decades.2 A more precise understanding of the causes of CHD is imperative for the recognition and potential intervention of progressive degenerative conditions, such as heart failure, among survivors of CHD. Deciphering nature’s secrets of heart formation might lead to novel approaches to repair or regenerate damaged heart muscle. The potential of stem cells in regenerative medicine is enormous, and insights into the natural process of cardiogenesis from progenitor cells during embryogenesis will form the basis of reprogramming cells for therapeutic use.3

The anatomic features of most CHD in humans have been carefully cataloged. Although CHD was classified in the 19th century based on embryologic considerations, the advent of palliative procedures and clinical management led to a descriptive nomenclature founded on anatomic and physiologic features that governed surgical and medical therapy. However, seemingly unrelated CHD could be argued to share common embryologic origins from a mechanistic standpoint, suggesting that the etiology of CHD may be better understood by considering their developmental bases. More recent advances in genetics and molecular biology have stimulated a renaissance in seeking an embryologic framework for understanding CHD as alterations and null mutations in a wide array of genes have targeted the heart and vascular system and established abnormalities in cardiovascular ontogeny as a primary cause of embryonic demise.4,5 The ability to go beyond descriptions of the anatomic defects to developing an understanding of the genes responsible for distinct steps of cardiac morphogenesis has raised the prospects that the future of pediatric cardiology will involve more directed therapeutic and preventive measures.

Although human genetic approaches have been important in understanding CHD, detailed molecular analysis of cardiac development in humans has been difficult. The recognition that cardiac genetic pathways are highly conserved across vastly diverse species from flies to man has resulted in an explosion of information from studies in more tractable and accessible biologic models. These include fruit flies, zebrafish, chicks, and mice as model systems. Each biologic system offers unique opportunities to develop a deeper understanding of cardiogenesis. In this chapter, anatomic, molecular, and clinical aspects of cardiac embryology will be interwoven to develop a framework in which to consider the etiology of human CHD. Clinical lessons combined with experimental studies in mice, fish, and flies have led to a model, suggesting that unique regions of the heart have been added from distinct fields of progenitor cells in a modular fashion during evolution. In this model, defects in particular regions of the heart arise from unique genetic and environmental effects on distinct cell lineages during developmental windows of time. To simplify the complex events of cardiogenesis, regions of the developing heart will be considered individually in the context described previously. In addition to the classic review of cardiac development by Dehaan in 1966, more recent publications provide additional details into anatomic events that are required for normal cardiac morphogenesis.4,6,7

The more recent discovery of distinct pools of progenitors that contribute to individual chambers or regions of the heart provides an important basis for considering the causes of many forms of congen;ital hear disease (CHD).6 The early cardiac progenitors arise from mesodermal cells that migrate from the primitive streak toward the anterior portion of the developing embryo. Two distinct mesodermal heart fields that share a common origin organize themselves into a crescent shape. The well-studied “first heart field” is derived from cells in the anterior lateral plate mesoderm that align in a crescent shape at approximately embryonic (E) day 7.5 in the mouse embryo, roughly corresponding to week 2 of human gestation (Fig. 480-1). By mouse E8.0, or 3 weeks in humans, these cells coalesce along the ventral midline to form a primitive heart tube, with an interior layer of endocardial cells and an exterior layer of myocardial cells, separated by extracellular matrix necessary for reciprocal signaling between the two layers.

The heart tube provides a scaffold that enables a second population of cells to migrate and expand into cardiac chambers.6 These additional cells arise from an area often referred to as the “second heart field (SHF),” anterior and medial to the crescent-shaped first heart field7-9 (Fig. 480-1). Both heart fields appear to be regulated by complex positive and negative signaling networks involving members of the bone morphogenetic protein (Bmp), sonic hedgehog (Shh), fibroblast growth factor (Fgf), Wnt, and Notch proteins. Such signals often arise from the adjacent endoderm, although the precise nature and role of these signals remain unknown10-12 (reviewed in 13). SHF cells remain in an undifferentiated progenitor state until incorporation into the heart, and this may in part be due to closer proximity to inhibitory signals emanating from the midline of the embryo.

As the heart tube forms, the SHF cells also migrate into the midline and position themselves dorsal to the heart tube in the pharyngeal mesoderm. Upon rightward looping of the heart tube, SHF cells migrate through the pharyngeal mesoderm into the anterior and posterior portions, populating a large portion of the outflow tract, future right ventricle, and atria14 (Fig. 480-1). The left ventricle region is sparsely populated by the SHF and appears to be largely derived from the FHF. Once within the heart, FHF and SHF cells appear to proliferate in response to endocardial-derived signals such as neuregulin and epicardial signals dependent on retinoic acid, although the mechanisms through which these noncell autonomous events occur remain poorly understood.15-17 The distinct contributions of SHF and FHF cells to the right and left ventricles, respectively, suggest that some forms of hypoplastic right or left ventricle may be due to a specific loss of the progenitor cell types during development.

Regulation of the SHF involves numerous signaling and transcriptional cascades (Fig. 480-2). Factors secreted from the anterior portion of the heart tube may serve as chemoattractant signals to induce the migration of SHF cells, although the nature of such molecules remains unknown.8 Islet1 (Isl1), a transcription factor, is necessary for development of the SHF.14 Progeny of Isl1+ cells contribute to most of the heart except the left ventricle, but Isl1 expression is extinguished as progenitor cells begin to express markers of cardiac differentiation. Interestingly, Isl1+ cells mark niches of undifferentiated cardiac progenitor cells in the postnatal heart,18 suggesting that understanding the regulation of SHF-derived progenitor pools may be useful in developing approaches for cardiac repair (discussed later in this chapter).

Discovery of the SHF led to the reinterpretation of CHDs observed in model systems and in humans. More than a decade ago, the first evidence of distinct transcriptional regulation between the right and left ventricle was found through the discovery of Hand2, required for expansion of the right ventricle,19 and its relative, Hand1, active only in the left ventricle. We now know that Hand2 is highly expressed in the SHF, and the right ventricular hypoplasia that results from disruption of Hand2 likely represents a failure of SHF cells to expand into the right ventricle.20 Similarly, disruption of numerous genes that resulted in right ventricular hypoplasia or outflow tract defects were a direct result of loss of SHF cells.21 Conversely, mutation of genes such as TBX5, the Holt-Oram gene, affect only FHF cells.

Epigenetic factors also contribute to cardiomyocyte differentiation and chamber morphogenesis. Disruption of the muscle-specific chromatin remodeling proteins Smyd122 or Baf60c23 results in a phenotype reminiscent of Hand2 mutants: a small right ventricular segment (Fig. 480-2). Whether mutations in these or other epigenetic factors contribute to CHD remains unknown, but the recent discovery of a chromatin remodeling protein, CHD7, as the cause of CHARGE association suggests an important role for this class of regulators.

Whereas transcriptional and epigenetic events regulate many critical cardiac genes, translational control by small noncoding RNAs, such as microRNAs (miRNAs), has recently emerged as another mechanism to “fine-tune” dosages of key proteins during cardiogenesis.24-27 This is particularly important, as most known genetic causes of CHD involve heterozygous mutations that only affect protein dosage by 50% or less. MicroRNAs are genomically encoded 20 to 25 nucleotide RNAs that target mRNAs for translational inhibition or degradation.28 More than 650 human microRNAs have been identified,29 but the biologic function and mRNA targets are known in only few.

One of the better studied miRNAs is miR-1, which is specific to cardiac and skeletal muscle. The sequence of miR-1 is 100% conserved from worms to humans and is specifically expressed in developing cardiac and skeletal muscle progenitor cells as they differentiate.24 miR-1 and another miRNA, miR-133, are transcribed in a polycistronic message therefore sharing common regulation, and both are coexpressed in heart and skeletal muscle. miR-1 promotes the differentiation of cardiac progenitors and skeletal myoblasts, but interestingly, miR-133a has the opposite effect, inhibiting differentiation and promoting the proliferation of myoblasts.29 Other miRNAs have been identified that pattern the developing heart and help establish the vasculature, and it is likely that the dosage of most other cardiac developmental pathways will be controlled by yet unknown miRNA function. It will be important to determine if mutations in miRNAs or their target-binding sites contribute to human disease.

Although the pathways regulating individual cell lineages contributing to the heart are well understood, the subsequent complex events involved in integrating multiple cell types, formation of chambers, and patterning of the distinct regions of the heart are also now being elucidated. From an evolutionary standpoint, it appears that as organisms became more complex, a more elaborate cardiovascular system was required. Distinct cardiac chambers began to develop and adopt specialized functions.

Dorsal-Ventral Polarity

Beyond the asymmetric addition of SHF cells along the anterior-posterior axis, a discrete dorsal-ventral (DV) polarity occurs in the primitive heart tube. As the heart tube loops to the right, the ventral surface of the tube rotates to become the outer curvature of the looped heart, and the dorsal surface forms the inner curvature. The outer curvature becomes the site of active growth, while remodeling of the inner curvature is essential for the ultimate alignment of the inflow and outflow tracts of the heart. A model in which individual chambers “balloon” from the outer curvature in a segmental fashion has been proposed. Consistent with this model, numerous genes are expressed specifically on the outer or inner curvature of the heart.30-32 Remodeling of the inner curvature allows migration of the inflow tract to the right and the outflow tract to the left, facilitating proper alignment and separation of right- and left-sided circulations.

Left-Right Asymmetry and Cardiac Looping

Rightward looping of the heart tube is the first obvious sign of the break in left-right symmetry. The cellular mechanisms that drive cardiac looping remain poorly understood, but it has been postulated that differential rates of proliferation of cardioblasts or altered cell adhesion across the heart tube may be involved. When considering the mechanisms for cardiac looping, it is important to distinguish between the process of looping and the directionality of looping. Directionality of looping reflects overall asymmetry throughout the embryo, which is superimposed on the morphogenetic mechanisms for looping.

Abnormalities in the process of cardiac looping underlie a number of CHD. Folding of the heart tube positions the inflow cushions adjacent to the outflow cushions and involves extensive remodeling of the inner curvature of the looped heart tube. In the primitive looped heart, the segments of the heart are still in a linear pattern and must be repositioned considerably for alignment of the atrial chambers with the appropriate ventricles and the ventricles with the aorta and pulmonary arteries. The atrioventricular septum (AVS) begins to divide the common atrioventricular canal (AVC) into a right and left AVC that subsequently shifts to the right to position the AVS over the ventricular septum (Fig. 480-3). This allows the right AVC and the left AVC to be aligned with the right and left ventricles, respectively. Simultaneously, the conotruncal region becomes septated into the aorta and pulmonary trunks as the conotruncus moves toward the left side of the heart such that the conotruncal septum is positioned over the AVS (Fig. 480-3). The rightward shift of the AVS and leftward shift of the conotruncus converts the single-inlet, single-outlet heart into a four-chambered heart that has separate atrial inlets and ventricular outlets.

Arrest or incomplete movement of the AVS or conotruncus might result in malalignment of the inflow and outflow tracts (Fig. 480-3). A scenario in which the AVS fails to shift to the right would result in communication of the right and left AVCs with the left ventricle, a condition known as double-inlet left ventricle (DILV). Incomplete shifting may be the basis for “unbalanced” AVC defects, where the right AVC only partly communicates with the right ventricle. Similarly, if the conotruncal septum fails to shift to the left, both the aorta and pulmonary artery would arise from the right ventricle, causing a double-outlet right ventricle (DORV). From this embryologic perspective, it is not surprising that double-outlet left ventricles and double-inlet right ventricles are rarely seen clinically. In contrast, any abnormality in cardiac looping can be associated with DILV or DORV, along with other manifestations of improper alignment of specific regions of the heart.

The elegant molecular network regulating left-right (LR) asymmetry of the body plan has been reviewed33 and is not summarized here. However, it is worth highlighting several clues about the relationship between LR asymmetry and proper alignment of the cardiac chambers. The cascade of LR signals including Shh and Nodal converge on the transcription factor Pitx2. Pitx2 is initially LR asymmetric in the linear heart tube, but this asymmetry is translated into a dorsal-ventral polarity in the looped heart tube. Because Pitx2 regulates cell proliferation via cyclin D2 and also controls cell migration events,34 it is a potential link between signals regulating the direction and process of cardiac looping. Within certain subdomains, regulation of Pitx2 by the DiGeorge syndrome gene, Tbx1, integrates transcriptional pathways controlling morphogenesis and LR asymmetry of the aortic arch arteries.35

How do the insights into LR asymmetry affect our understanding of CHD? It is likely that patients with situs inversus totalis have a well-coordinated reversal of LR asymmetry and thus have a lower incidence of defects in visceral organogenesis. However, the majority of patients with LR defects has visceroatrial heterotaxy and thus has randomization of cardiac, pulmonary, and gastrointestinal situs where coordinated signaling is absent. Such patients can have defects in almost all aspects of cardiogenesis. Often, either the right or left side predominates with patients either having bilateral right-sidedness (asplenia syndrome) or bilateral left-sidedness (polysplenia syndrome). In such patients, features of the right or left side of the lungs, heart, and gut are duplicated. Disruption of cascades determining either the left or right side of the embryo might result in asplenia or polysplenia syndromes, respectively. Indeed, mutations in LR pathway members are found in some patients with heterotaxy. Familial examples of heterotaxy have also led to identification of mutations in a zinc-finger transcription factor, ZIC3, that result in LR axis abnormalities.

Cardiac Outflow Tract Regulation

Congenital cardiac defects involving the cardiac outflow tract, aortic arch, ductus arteriosus, and proximal pulmonary arteries account for 20% to 30% of all CHD. This region of the heart undergoes extensive and rather complex morphogenetic changes with contributions from neural crest cells and the SHF, as discussed previously. The cardiac outflow tract can be divided into the muscularized conus and the adjacent truncus arteriosus, collectively termed the conotruncus, as it arises from the primitive right ventricle. The conotruncus normally shifts to the left to override the forming ventricular septum. The truncus arteriosus then becomes septated by neural crest–derived mesenchymal cells into the aorta and pulmonary arteries, with a muscular ridge known as the conotruncal septum forming between the two vessels (summarized in Fig. 480-4). However, at this stage, the aorta communicates with the right ventricle and the pulmonary artery with the left ventricle. Subsequent rotation of the two vessels in a spiraling fashion places the aorta more dorsal and leftward and the pulmonary artery more ventral and rightward. This spiraling event achieves the normal alignment of the aorta and pulmonary artery to the left and right ventricles, respectively.

Mesenchyme cells originating from the crest of the neural folds are essential for proper septation and remodeling of the outflow tract and aortic arch.36 Such neural crest–derived cells migrate away from the neural folds and retain the ability to differentiate into multiple cell types (Fig. 480-4). The migratory path and ultimate fate of these cells depends on their relative position of origin along the anterior-posterior axis and are partly regulated by the Hox code.37 Neural crest cells differentiate and contribute to diverse embryonic structures, including the cranial ganglia, peripheral nervous system, adrenal glands, and melanocytes. Neural crest cells that arise from the otic placode to the third somite migrate through the developing pharyngeal arches and populate the mesenchyme of each aortic arch artery, and the mesenchyme necessary to septate the outflow tract septum (Fig. 480-4).

Abnormalities in septation or incomplete spiraling of the conotruncus result in many CHDs. For example, the conotruncal septum between the aorta and pulmonary artery forms in tetralogy of Fallot (TOF), but because of malalignment of the great vessels, the conotruncal septum and aorta are shifted to the right. This results in an overriding aorta and failure of the conotruncal septum to connect to the muscular ventricular septum, resulting in a ventricular septal defect (Fig. 480-3). Similarly, any malalignment of the conotruncus results in an obligatory VSD that, unlike muscular VSDs, does not have the potential to close spontaneously after birth.

The conotruncus gives rise to six bilaterally symmetric vessels known as aortic arch arteries. The aortic arch arteries arise sequentially along the AP axis, each traversing a pharyngeal arch before joining the paired dorsal aortae (Fig. 480-4). The first and second arch arteries involute, and the fifth arch artery never fully develops. The third, fourth, and six arch arteries undergo extensive remodeling to ultimately form distinct regions of the mature aortic arch and proximal pulmonary arteries. The majority of the right-sided dorsal aorta and aortic arch arteries undergo programmed cell death leading to a left-sided aortic arch. The third aortic arch artery contributes to the proximal carotid arteries and right subclavian artery. The left fourth aortic arch artery forms the transverse aortic arch between the left common carotid and left subclavian arteries. Finally, the sixth arch artery contributes to the proximal pulmonary artery and the ductus arteriosus (Fig. 480-4). Extrapolating from their embryologic origins, it is believed that aberrant right subclavian arteries and other subtle arch anomalies are the result of third aortic arch defects, interrupted aortic arch from fourth arch defects, and patent ductus arteriosus and proximal pulmonary artery hypoplasia/discontinuity from defects in sixth arch artery development.

Disruption of many signaling cascades that affect neural crest migration or development, including the endothelin and semaphorin pathways, cause outflow tract defects similar to those observed in humans.38 These include TOF, persistent truncus arteriosus, double-outlet right ventricle, ventricular septal defects, and defects of aortic arch patterning. Thus, abnormalities in neural crest migration or differentiation likely underlie many of the conotruncal and aortic arch defects seen in humans. Indeed, human mutations of the neural crest–enriched transcription factor TFAP2β result in persistent patency of the ductus arteriosus, a specialized aortic arch vessel essential for fetal cardiac physiology (Fig. 480-4). It is likely that other genetic mutations affect specific regions of the aortic arch.

Disruption of SHF development by mutation of genes such as Tbx1, Fgf8, and Islet1 results in defects similar to those observed with neural crest disruption (Fig. 480-2), including persistent truncus arteriosus (failure of outflow septation), malalignment of the outflow tract of the heart with the ventricular chambers, and ventricular septal defects.14,39-41 Because SHF-derived myocardial cells neighbor neural crest–derived cells and secrete growth factors such as Fgf8 in a Tbx1-dependent manner that influence neural crest cells,42 reciprocal interactions between the SHF and neural crest–derived cells in the outflow tract are likely essential for normal development. Consistent with this, humans with deletion or mutation of the SHF gene TBX1 who have DiGeorge syndrome (22q11 deletion syndrome)43 appear to have cell-autonomous defects of SHF development and non–cell-autonomous anomalies of neural crest–derived tissues. It will be interesting to determine if a large number of human cardiac outflow tract defects are a direct result of SHF migration, differentiation, or proliferation.

Cardiac Valve Formation

Appropriate placement and function of cardiac valves is essential for chamber septation and for unidirectional flow of blood through the heart. A molecular network involving Bmp2 and Tbx2 defines the position of the valves relative to the chambers.32,44,45 During early heart tube formation, “cushions” of extracellular matrix (ECM) between the endocardium and myocardium presage valve formation at each end of the heart tube. Reciprocal signaling, mediated in part by TGF-β family members, between the myocardium and endocardium in the cushion region induces endocardial cells to transform into mesenchymal cells that migrate into the ECM cushion.46-48 These mesenchymal cells differentiate into the fibrous tissue of the valves and are involved in septation of the common atrioventricular canal into right- and left-sided orifices.

Later in development, the endocardial cushions form condensed mesenchymal protrusions that represent the primitive valves. These condensed mesenchymal protrusions subsequently “elongate” to provide the true cardiac valve leaflets (Fig. 480-5). The elongation of primitive valves appears to be a result of restricted proliferation of endocardial cells overlying the mesenchymal projections on the vascular side of the valve and selective cell death under the expanding endocardial rim. The growth of the endocardial edge and evacuation of apoptotic cells underneath the proliferating endocardial rim sculpt the swollen mesenchymal primitive valves into a typical excavated shape and results in morphogenesis of the sinuses of Valsalva.

Many forms of congenital heart disease involve thickened, hyperplastic valve leaflets that can result in obstruction of blood flow. Study of genetic syndromes that affect valves have been informative regarding the mechanisms underlying this condition. Neurofibromatosis is associated with thickened valves and is caused by mutation of the NF1 gene. NF1 loss of function is mimicked by gain of function mutations in the Noonan syndrome gene, tyrosine phosphatase SHP2/PTPN11, which encodes a tyrosine phosphatase that activates Ras-Erk signaling, increased proliferation, and decreased apoptosis. The net result is semilunar valve and atrioventricular valve hyperplasia.49 Hypomorphic mutations in SOS1, an essential RAS guanine nucleotide-exchange factor (Ras-Gef), also result in enhanced RAS-ERK activation and can account for as many as 20% of the cases of Noonan syndrome not explained by SHP2 mutations.49

In humans, heterozygous NOTCH1 mutations disrupt normal development of the aortic valve and, occasionally, the mitral valve, with thickened valve leaflets.50 The Notch signaling pathway is required for cell fate and differentiation decisions throughout the embryo. The severity of valve disease associated with NOTCH1 mutations in humans varies widely from mild disease in which the aortic valve has two rather than three leaflets (bicuspid aortic valve) to severely thickened valves and defects in valve patency resulting in critical aortic stenosis in newborns or even left ventricular growth failure. Consistent with this genetic finding, 15% of “normal” relatives of children with hypoplastic left heart syndrome (HLHS) have subclinical bicuspid aortic valves,51 suggesting that disruption of the NOTCH signaling cascade may underlie a spectrum of aortic valve disease. Although not specifically affecting valves, human mutations in JAGGED1, a NOTCH ligand, also cause outflow tract defects associated with the autosomal-dominant disease, Alagille syndrome.52-54

The epicardium originates as a villous projection of cells in the area of the sinus venosus termed the “proepicardial organ.” This cluster of cells extends to the atrioventricular region and migrates out over the myocardial surface to completely encase the heart.55-59 Angiogenic sprouts from the subepicardial endothelial plexus form endothelial strands that grow into the aorta and develop multiple communications with all three cusps of the developing aortic valve.60-62 However, lumens develop only in two semilunar sinuses with resorption of the strands to the noncoronary cusp.63 These observations have implications for determining the factors that direct coronary artery anatomy in congenital heart diseases.

As human survivors of cardiac malformations ranging from simple atrial septal defects to more complex heart disease enter their third and fourth decades of life, new cardiac disease processes are becoming apparent, including abnormal conduction of electricity within the heart and diminished contractile function of the heart. The etiology of these “secondary” defects had in the past been ascribed to abnormal hemodynamics, but more recent evidence suggests that the same genes that cause early morphologic defects in heart development might later be directly involved in cardiac dysfunction and cell lineage disturbances in adulthood.2

For example, human and mouse mutations in NKX2.5 not only cause a developmental atrial septal defect, but also progressively disrupt electrical conduction through the cardiac chambers and can result in sudden death later in life.64 The atrioventricular (AV) node, which serves as the essential site of electrical communication between the atria and ventricles, is smaller than normal in adult Nkx2.5 mouse mutants; over time, the specialized muscle-derived conduction cells are lost and replaced by fibrotic tissue, resulting in progressive defects in electrical conduction.65

Another example of a congenital heart malformation causing disease later in life involves the aortic valve. Worldwide, 1% of the population is born with a bicuspid aortic valve, typically silent in childhood.1 However, two thirds of bicuspid aortic valves develop premature age-dependent calcific stenosis, resulting in poorly mobile, nonfunctioning valves, often in the fifth, sixth, or seventh decades of life. As a result, calcific aortic stenosis is the third leading cause of heart disease in adults and requires more than 30,000 surgical valve replacements per year in the United States. The more recent discovery that NOTCH1 mutations in humans cause bicuspid aortic valves and later calcification provides firm genetic evidence that the early developmental and later degenerative disease share a common genetic cause.66 Some family members with NOTCH1 mutations had normal tricuspid valves but still developed calcification, supporting the idea that the premature and severe calcification was primarily due to the genetic mutation itself rather than solely to hemodynamic effects on the valve leaflets.

The field of cardiac developmental biology has progressed rapidly over the past decade, and the heart is now one of the best-understood organs at the molecular, physiologic, and anatomic levels. Advances in human genetic tools have also led to a deeper understanding of the importance of developmental pathways in human disease. Coronary heart disease (CHD) can now be conceived not only as a defect of morphogenesis, but also in some cases as a failure of differentiation among subsets of lineages that contribute to the heart. We are now embarking on a phase in which knowledge of developmental pathways and high-throughput methods of genotyping rare and common gene variants should allow rigorous investigation into the causes of human heart disease. With the increasing recognition that CHD has a significant genetic contribution, we can now imagine that genetic variation underlies both the morphogenetic defect and the predisposition to long-term consequences that will affect clinical outcomes for the millions of CHD survivors. Thus, vigorous efforts to identify genetic variation associated with CHD and outcome will be essential as therapeutic or preventive measures to alter the course of disease may be possible throughout childhood and in the adult. It may even be conceivable to eventually predict genetic risk among parents and focus preventive strategies on those at greatest risk to vertically transmit disease. The efficacy of folic acid in prevention of neural tube defects provides hope for similar prevention of congenital heart disease.67

As the next phase of disease-related biology evolves, parallel advances in stem cell biology should usher in an era of new approaches. It is exciting to imagine future technologies that may enable generation of disease-specific embryonic stem cell lines for mechanistic studies of disease etiology and development of patient-specific stem cells for ultimate treatment modalities.