|Home | About | Journals | Submit | Contact Us | Français|
Congenital heart disease (CHD) is the most common congenital anomaly in newborn babies. Cardiac malformations have been produced in multiple experimental animal models, by perturbing selected molecules that function in the developmental pathways involved in myocyte specification, differentiation or cardiac morphogenesis. In contrast, the precise genetic, epigenetic or environmental basis for these perturbations in humans remains poorly understood. Over the past few decades, researchers have tried to bridge this knowledge gap through conventional genome-wide analyses of rare Mendelian CHD families and by sequencing candidate genes in CHD cohorts. While yielding few, usually highly penetrant, disease gene mutations, these discoveries provided three notable insights. First, human CHD mutations impact a heterogeneous set of molecules that orchestrate cardiac development. Second, CHD mutations often alter gene/protein dosage. Third, identical pathogenic CHD mutations cause a variety of distinct malformations, implying that higher order interactions account for particular CHD phenotypes. The advent of contemporary genomic technologies including SNP arrays, next-generation sequencing, and CNV platforms are accelerating the discovery of genetic causes of CHD. Importantly, these approaches enable study of sporadic cases, the most common presentation of CHD. Emerging results from ongoing genomic efforts have validated earlier observations learned from the monogenic CHD families. In this review, we explore how continued use of these technologies and integration of systems biology is expected to expand our understanding of the genetic architecture of CHD.
Congenital heart disease (CHD) defines a large set of structural and functional deficits that arise during cardiac embryogenesis (Figure 1). CHD is the most common type of birth defect, accounting for one third of all major congenital anomalies. Worldwide, 1.35 million infants are born with CHD each year. CHD is also identified in 10% of stillbirths1 and is presumed to be a substantive cause of early fetal demise. The prevalence of CHD varies across countries and continents.2 In North America, CHD occurs in 8.1 per 1,000 live births3 while in Asia the prevalence is 9.3 per 1000 live births, a difference that is attributed in part to higher rates of parental consanguinity.2 Because the full spectrum of congenital heart defects includes mild lesions that are clinically quiescent for decades (e.g., BAV with a population prevalence ranging from 0.5% to 0.9%4–6), the worldwide prevalence of all CHD may exceed these estimates.
Until recently, nearly half of the deaths due to CHD occurred during infancy, but with remarkable advances in prenatal diagnosis, corrective strategies, and longitudinal care, infantile mortality has substantially declined. Today, more than 75% of CHD children who survive the first year of life, including those with complex malformations, will live into adulthood.7,8 Recent estimates define the prevalence of CHD in adults at approximately 3000 per million,9 a figure that predicts that there are 21 million adults living with CHD. Moreover, this unique cardiovascular disease population has been increasing by almost 5% per year.10 Life-long CHD can pose substantial physiological, emotional and socioeconomic challenges for patients, families, and society. As such, discovery of the causes for CHD is not only a fundamental research endeavor - it is vital to the healthcare of this growing community.
Causes of CHD are often partitioned into genetic and non-genetic categories. Well-recognized non-genetic etiologies of CHD include environmental teratogens (dioxins, polychlorinated biphenyls, pesticides11), maternal exposures (alcohol, isotretinoin, thalidomide, anti-seizure medications12), and infectious agents (e.g., rubella13). Despite decades of international efforts to combat these factors, the compendium of non-genetic causes of CHD continues to increase and to diversify. Anti-retroviral medications14 that are taken by eight million people worldwide,15 and the epidemic of obesity16 with associated phenotypes of diabetes17 and hypercholesterolemia18 are recognized as emerging risk factors for CHD.
The genetic landscape of CHD is also changing. The renowned pediatric cardiologist Dr. Helen Taussig speculated that since “common cardiac malformations...occur in otherwise “normal” individuals...these malformations must be genetic in origin.”19 Yet discovery that gene mutations cause CHD began only decades after her death. Initial human genetics methodologies had poor resolution, which restricted analyses to inherited forms of CHD. Given the historical rates of poor reproductive fitness in CHD patients and high mortality, early genetic studies of familial CHD were often biased towards uncomplicated malformations such as ASDs and VSDs. The improved health of CHD patients and major advances in genomic technologies has shifted this paradigm.
Contemporary methodologies provide robust opportunity for comprehensive genomic analyses of all CHD patients, including those with sporadic and complex malformations. An accompanying manuscript (The Congenital Heart Disease Genetic Network Study (CHD GENES): Rationale, Design, and Early Results20) details approaches being spearheaded by the NHLBI to study genetic etiologies in thousands of CHD patients. Comparable efforts are underway around the world.21–23
Deciphering the contributions of genetic and non-genetic causes of CHD has benefited from extensive model organism studies that have provided a wealth of insights into cardiac developmental biology. Molecular pathways have been identified that orchestrate formation of primordial cardiogenic fields, that shape the cardiac crescent and linear heart tube, and that drive atrial, ventricular, inflow and outflow tract morphogenesis.24–27 Within these pathways, details have emerged about molecules that promote lineage specification, differentiation, cell growth, and migration and that orchestrate temporal and spatial patterns of gene expression.28–32 Positioning previously discovered and novel CHD genes onto this blueprint presents remarkable opportunities to further extend the knowledge base of cardiac embryogenesis and to fully understand the causes and mechanisms of CHD.
In this review, we examine historical and recent genetic discoveries in CHD, focusing on malformations identified in infancy (Figure 1) that inform developmental themes in heart development. In addition, we examine strategies (Figure 2) that can expand the discovery of new CHD genes and explore relationships between genetic and non-genetic etiologies. Given the current pace of human CHD genetics and genomics research, these efforts can only be considered a preview. Readers should expect that rapidly emerging data will provide a much fuller understanding of the genetic architecture of heart development and CHD.
Familial CHD mutations occur as autosomal dominant, autosomal recessive, or X-linked traits that are expressed with high penetrance and with variable clinical manifestations. CHD is genetically heterogeneous. That mutations in different genes cause an identical malformation underscores the highly interdependent roles of molecules involved in heart development. Moreover, the spectrum of heart malformations that arise for an identical gene mutation implicates genomic context,33, 34 maternal-fetal environment,12 cardiac biomechanics,35 and other factors as important influences that impact the clinical consequences of CHD mutations.
Hundreds of autosomal dominant or X-linked mutations have been identified in familial forms of CHD. Information that has ensued from these discoveries is reviewed below. An evolutionary perspective of CHD mutations predicts that reduced reproductive fitness and early mortality would cause substantial negative selection that eliminates CHD mutations from human populations. If autosomal dominant or X-linked mutations make a significant contribution to the population prevalence of CHD, many must be new (de novo) mutations that initially result in sporadic CHD. Autosomal dominant de novo mutations should cause high recurrence rates in the offspring of sporadic CHD cases. However, a recent large analysis of 1.7 million Danes identified only 2.2% of individuals with CHD had an affected first-degree relative,36 data that challenges the model that dominant, de novo mutations are major contributors to CHD.
Autosomal recessive or somatic mutations and polygenic variants pose alternative genetic models to account for the population prevalence of CHD. In comparison to dominant gene mutations, far less is known about these genetic models in CHD. Epidemiologic data that parental consanguinity (especially first-cousin marriages37–39) significantly increases CHD risk40, 41 provides compelling evidence that recessive mutations cause CHD. Discovery of CHD mutations in genetically closed populations42, 43 and ascertainment of the burden of compound recessive mutations in out bred populations44, 45 should inform the contribution of recessive genes to CHD.
Somatic mutations in monogenic genes that arise during the early development of cardiac progenitor cells might cause some cases of CHD. Contemporary sequencing strategies provide estimates that in each generation a few (<10) de novo rare deleterious mutations occur.46, 47 While these data cannot be extrapolated to estimates of the frequency of new somatic mutations in rapidly proliferating and differentiating cells, they indicate that de novo mutations occur not uncommonly and imply that analyses of somatic mutations in malformed cardiac tissues from CHD patients may be informative.
The population prevalence of CHD is not dissimilar from other “common” disorders. By extrapolation, an alternative genetic model for CHD is that multiple variants, which individually contribute small risks that can be maintained throughout evolution, collectively cause CHD. Genome-wide association studies (GWAS) of large cohorts (≥1000 cases) are typically used to explore the common disease-common variant hypothesis. The prevalence and viability of some forms of CHD (e.g., ASD, VSD) and development of large CHD registries, such as CHD GENES, should enable testing of this genetic model. An alternative approach, which capitalizes on the evidence that rare monogenic mutations cause CHD, has been analyses of variants in candidate genes as polygenic risk factors of CHD. For example, an NKX2-5 gene variant that has functional consequences when assayed by in vitro experiments has been associated with CHD in five independent studies.48 As one percent of the population carries this NKX2-5 variant,49 this variant may predispose to, but not directly cause, CHD. Associations with other CHD variants in selected loci50, 51 also support a polygenic model of CHD and hint that genome-wide association studies in large CHD cohorts will be informative.
Recognizing these issues, below we review loci, genes, and mutations that cause CHD, and indicate how contemporary technologies continue to advance genetic models and mechanisms of CHD.
Chromosomal aneuploidy, the first recognized genetic cause of CHD, continues to be a major etiology today (Table 1). CHD occurs in approximately 40–50% of trisomy 21 (1 in 600 births52), 20–50% of Turner syndrome (1 in 2500 female births53), and in almost all cases of both trisomy 13 and trisomy 18.54 Although almost any cardiac malformation can occur with aneuploidy syndromes, prototypic lesions are observed in trisomy 21 (AVSD52) and Turner syndrome (CoA53) while other lesions (e.g., TGA) are strikingly underrepresented. An important early conclusion from these genotype-phenotype observations was that cardiac malformations are not due to a global change in genomic content but rather from altered dose of specific genes.
This concept gained more clarity with the development of methodologies to define sub-chromosomal changes in genome structure, denoted today as copy number variants (CNVs). CNVs are large deletions or amplifications of DNA segments that arise principally from inappropriate recombination, due to flanking region-specific repeat sequences or from highly homologous genes (such as ancestral duplication sites) that misalign during meiosis. CNVs that encompass millions of bases (Mb) can be identified by cytogenetic analyses, often in combination with fluorescence in situ hybridization (FISH). Smaller CNVs (affecting as few as several hundred bases) can be detected using high-resolution array-comparative genomic hybridization (array-CGH) or genomic microarrays that assess single nucleotide polymorphisms (SNPs) and copy number probes. These CNVs and insertions or deletions of size less than a hundred bases (collectively denoted as indels) are identified by sequence-based approaches.
As CNVs alter the dosage of contiguous genes, they can produce syndromic CHD (Table 1). A 3-Mb CNV on chromosome 22q11 causing CHD, craniofacial abnormalities, neurocognitive disabilities, absent or hypoplastic thymus, hypocalcemia/ hypoparathyroidism (velocardiofacial or DiGeorge syndrome, now denoted as 22q11 deletion syndrome) is the most common CHD CNV. Occurring in 1 in 4000 live births, chromosome 22q11 CNVs account for 15% of TOF cases.55 Although more than 30 genes are impacted by this CNV, sequence analyses of some of these candidates56 and animal models57 indicate that altered dose of one gene, TBX1, a T-box transcription factor that promotes cell proliferation in the secondary heart field,58, 59 from which the outflow tract and right ventricle develop60 accounts for most of the observed clinical features. CHD can also occur from a 1.5-Mb deletion on chromosome 7q11.23 that alters the dosage of over 25 genes and causes Williams-Beuren syndrome (SVAS, developmental delays, gregarious personality, elfin facies, and hypercalcemia.61 While disruption of the elastin gene (ELN) accounts for the cardiovascular abnormalities in Williams-Beuren syndrome and for isolated cases of nonsyndromic SVAS,62 unidentified genes impacted by the chromosome 7q11.23 CNV account for other phenotypes in this syndrome.
Recent analyses have identified multiple CNVs that contribute to isolated (non-syndromic) CHD (Table 2 and Online Table I). Large de novo CNVs (present in the probands but absent in both parents) have been reported in TOF,63, 64 left-sided lesions (e.g., AS, BAV, CoA65, 66, HLHS67) and other sporadic cases of CHD.23, 68–72 These studies estimate that 5–10% of sporadic, non-syndromic CHD, in patients with normal karyotype and FISH analyses, is due to a rare (≤1% population frequency) CNV.
Some CNVs encompass previously identified CHD genes or genes known to participate in heart development from the study of model organisms. For example, recurrent CNVs identified in CHD cases that occur at chromosome 8p23.1 impact the cardiac transcription factor GATA4,23 and CNVs at chromosomes 20p12.2, and 9q34.3, impact members of the Notch signaling pathway, JAG1 and NOTCH1.63
Other CNVs identified in CHD cases provide opportunities for the discovery of new disease genes, efforts that usually require multidisciplinary approaches (Figure 2). One strategy for defining the culprit gene is to demonstrate independent pathogenic mutations in CHD cases without the CNV, an approach that successfully identified TBX1 as the critical gene responsible for chromosome 22q11 deletion syndromes56 and CHD773 on chromosome 8q12.1 that causes CHARGE syndrome74 (coloboma of the eye, heart defects, atresia of the choanae, retardation of growth and/or development, genital and/or urinary abnormalities, and ear abnormalities and/or deafness). Other approaches harness bioinformatic strategies to prioritize genes and capitalizing on conserved pathways of heart development across species. A recent study of a recurrent CHD CNV at chromosome 6q24.3–25.1 that involved over 100 genes illustrates this strategy. Investigators annotated genes encoded within the critical CNV interval using datasets of cardiac developmental expression in the mouse. Selected candidate genes were then analyzed for dosage-sensitivity using morpholinos in zebrafish and monitoring cardiac development. This approach implicated TAB2 that encodes TGF-b-activated kinase 1 (also known as MAP3K7 binding protein-2), a kinase complex member which participates in activation of nuclear factor kappa b and activator protein-1.75 The identification of a chromosomal translocation (t(2;6)(q21;q25) involving TAB2 in a family with CHD provided further support that TAB2 participates in signal transduction during cardiac development.
In addition to defining novel CHD genes, CNVs can be used to assess developmental networks using bioinformatic repositories of biological interactions and functional annotations as well as gene-gene and protein- protein relationships. For example, bioinformatic analyses of rare CNVs identified in 2500 CHD cases23 showed that CNVs impacted genes that were significantly enriched for participation in Wnt signaling, which regulates cellular processes involved in proliferation and differentiation. Although Wnt signaling in cardiac development has been identified in model organisms, this study provided the first evidence for this pathway in human CHD.76, 77
Discovery of genes with point mutations that caused CHD (Table 3) was initially undertaken in familial cases using classical linkage analyses to identify CHD loci, and sequence analyses of candidate genes or positional-cloned genes to define pathogenic mutations. Contemporary strategies bypass steps that define CHD loci and instead identify CHD mutations by direct next-generation sequencing at high read depths (≥20 reads per base) of the exome (the 1% of the genomic sequence that encodes protein) or the whole genome. These approaches identify tens of thousands of SNPs per exome47 and multiple fold more SNPs per genome. As most of these SNPs will be unrelated to CHD, extensive post-sequencing filters are employed to focus on novel or rare SNPs (occurring in ≤1% populations matched for race and ethnicity) that are predicted to have deleterious functional consequences (e.g., non-synonymous SNPs that alter evolutionarily conserved residues), and that occur in genes that are expressed during heart development. Additional evidence for pathogenicity of rare, deleterious non-synonymous SNP includes 1) statistically significant co-segregation in familial CHD; 2) identification of recurrent deleterious non-synonymous SNPs that arise de novo in unrelated cases of sporadic CHD; 3) genetic complementation (e.g. CHD caused by deleterious non-synonymous SNPs in genes that participate in different steps of a cardiac developmental pathway); and 4) recapitulation of CHD in model organisms.
A survey of the current compendium of definitive CHD gene mutations predicts that the mechanism by which these perturb heart development is through haploinsufficiency, or a reduction in the dosage of the encoded proteins. Haploinsufficiency occurs through gene inactivation (e.g., nonsense or frameshift mutations), by altering gene expression (e.g., non-coding regulatory mutations) or by encoding nonfunctional or loss-of-function (LOF) proteins (e.g., missense mutations). CHD mutations that produce a gain in gene dosage (e.g., duplications or non-coding regulatory mutations) or increase protein activity (e.g., missense mutations that enhance protein function) are less common. The disproportionate numbers of haploinsufficiency/LOF CHD mutations may reflect inherent difficulties in recognizing sequence variants that increase gene expression or protein function. Alternatively, this imbalance may be biologically meaningful and indicate that a minimum threshold of expression of genes involved in heart development is more critical than excess levels.
Point mutations that increase or decrease the dosage of genes functioning in developmental pathways that are broadly used in organogenesis cause syndromic CHD (Table 1). Alagille syndrome (TOF, PS, and other CHD, cholestasis, skeletal abnormalities, distinctive facies, and ocular disease78, 79 is caused by dominant mutations in the JAG1 gene (which encodes the Notch receptor-1 ligand) in over 90% of cases or in the NOTCH2 gene.80 The broad mutational spectrum (frameshifts, nonsense, disrupted or cryptic splice signals, missense) in either gene reduces Notch signaling, a highly conserved pathway involved in lineage specification and cell-fate decision during development.
Holt-Oram Syndrome (ASDs, VSDs, conduction system disease, upper arm malformations,) can arise from dominant LOF mutations in TBX5, a member of the T-box gene family,81 that encode transcription factors that contain a conserved DNA-binding motif. T-box proteins function in regulating cell fate decisions and early pattern formation and different gene family members contribute to organogenesis.82 TBX5 is expressed in the upper limbs and heart.81 CHD mutations have been identified that disrupt 5’ regulatory sequences and that perturb residues in the T-box DNA-binding motif83 are predicted to reduce the levels of functional TBX5 protein.
Noonan syndrome and related disorders (PS, ASD, CoA, facial dysmorphism, short stature, pectus deformity, cubitus valgus, neck webbing, developmental delays) are caused by dominant gain-of-function mutations in one of eleven genes: PTPN11, SOS1, RAF1, KRAS, BRAF, MEK1, MEK2, HRAS, NRAS, SHOC2, and CBL. These genes encode molecules that function in the RAS-MAPK signal transduction pathway84, 85 that communicates extracellular signals to the nucleus by modulating a GDP/GTP-regulated protein kinase cascade. The RAS-MAPK pathway is implicated in cell proliferation, differentiation and survival by directly regulating transcriptional activation and indirectly by chromatin modification.86
Mutations in genes that cause syndromic CHD can occasionally produce isolated heart malformations.56, 62, 87 Possible explanations for the absence of extra-cardiac manifestation might include subclinical phenotypes and/or tissue-specific mechanisms for dosage compensation.
The list of gene mutations that cause isolated CHD (Table 3) is rapidly expanding. Rather than providing details about each gene, below we discuss three broad functional categories, transcriptional regulation, signal transduction, and cardiac structural proteins, into which isolated CHD genes can be parsed.
The critical importance of transcriptional regulation of gene expression for normal heart development was first identified by the discovery of CHD mutations in NKX2-5,34, 88–90 NKX2-6,42 GATA4,70, 91–93 GATA594, 95, GATA696, IRX4,97 TBX20,98, 99 and ZIC3.100, 101 CHD is caused by dominant mutations in each of these genes that are predicted to reduce physiologic levels of the encoded protein by mutations that inactivate one allele, or cause LOF by disrupting DNA interactions102 or perturbing the combinatorial interactions of transcription factors34, 91, 103 and transcriptional cofactors (e.g., FOG2, which encodes Friend of Gata-2104, 105). Definitive evidence that haploinsufficiency of cardiac transcription factors causes CHD is predicated on independent LOF mutations that have been identified in unrelated CHD cases and on mouse models with heterozygous gene deletions that recapitulate the cardiac malformations found in patients.106–108
Expression of cardiac transcription factors occurs in highly specified temporal-spatial patterns throughout development - a level of regulation that might predict there would be strong correlations between genotype and phenotype in CHD. In contrast, the clinical spectrum of malformations that arise from mutations in cardiac transcription factors is strikingly broad. Despite this generalization, the integration of insights from developmental biology has informed why some human CHD mutations produce specific clinical phenotypes. For example, LOF mutations in ZIC3 cause cardiac laterality defects that are often accompanied by visceral heterotaxy, an association that is explained by evidence that ZIC3 transcriptionally activates NODAL, a critical morphogen that is required for left-right patterning throughout the embryo.109 LOF mutations in GATA4 typically cause ASDs, but because a subset of these mutations disrupts GATA4-SMAD4 interactions that are critical for valve development, some patients have atrioventricular canal defects.103 Mutations in NKX2-5 and TBX5 cause cardiac malformations that are associated with electrophysiologic deficits, presumably because these transcription factors have been demonstrated to function in molecular specification of the myocytes in the conduction system.110
CHD occurs from LOF mutations in a variety of genes that encode molecules that participate in developmental signaling pathways. Establishment of a left– right axis during embryogenesis is predicated on a laterality signaling pathway that results in asymmetric placement of organs about the midline as well as in cardiac looping.111 Mutations in ZIC3, NODAL, and in LEFTY2, which encodes a molecule that restricts the expression of Nodal-responsive genes to the left side of embryos, disrupt normal laterality signals that direct cardiac looping and cause a spectrum of heart malformations.112–118
The Notch signaling pathway is implicated in multiple developmental processes. Mutations in NOTCH1, NOTCH2 and JAG1 are all predicted to reduce ligand-induced signaling,119 albeit with strikingly different consequences. As discussed above, mutations in NOTCH2 and JAG1 cause diverse phenotypes in Alagille syndrome. In contrast, NOTCH1 mutations typically cause malformations of the aortic valve.78, 79, 120 As Notch signaling participates in epithelial-to-mesenchymal transformation,121 a process that is critical for normal valvulogenesis, NOTCH1-dependent signals appear to be particularly important in this cellular transformation process.
The identification of independent LOF mutations in developmental signaling factors in unrelated CHD cases and evidence for genetic complementation (e.g. NOTCH2 and JAG1 cause Alagille syndrome) strongly supports the pathogenicity of these mutations. Although mice engineered to heterozygous LOF mutations in Jag1122 or Notch2123 lack CHD, homozygous deficiency of either gene causes an amalgam of defects and embryonic lethality. More recent studies of regional-specific depletion of these molecules124, 125 have demonstrated extensive abnormalities in cardiovascular morphogenesis,126 therein substantiating that mutations in these genes are definitive causes of human CHD.
Genes that encode cardiac structural proteins comprise the smallest category and least definitive monogenic cause of CHD. LOF mutations in ELN (which encodes elastin) cause CHD in the context of Williams-Beuren syndrome (described above) and less commonly in isolated cases of SVAS.62 Rare missense mutations and premature termination mutations in MYH6, MYH7 (encoding the a, b cardiac myosin heavy chains, respectively), and ACTC (a cardiac actin) have been reported as rare causes of autosomal dominant ASDs,127 Ebstein anomaly,128 and other CHD.129–132 As most missense mutations in cardiac sarcomere proteins cause human cardiomyopathy, and mice with haploinsufficiency of Myh6133 or Actc134 have normal heart structure, the evidence that sarcomere protein genes mutations cause CHD is not definitive. Rare missense mutations in MYH11 (encoding smooth muscle myosin heavy chain) are reported to cause dominant thoracic aortic aneurysm that is sometimes accompanied by PDA.135, 136 As Myh11-null mice have delayed closure of the ductus arteriosus,137 human MYH11 mutations associated with PDA may cause LOF.
Systems biology, which integrates complex datasets obtained from model organisms and humans into cogent pathways that operate in multidimensional spaces, provides new avenues to elucidate CHD.138, 139 Systems biologic approaches capitalize on the conservation of heart development genes and processes across species,140 molecular networks of heart development,141 with genetic and environmental risks for CHD. Using bioinformatics and computational algorithms to elucidate molecular pathways and interactions in heart development and CHD these strategies have the potential to predict the pathogenicity and consequences of individual CHD mutations.
Two examples illustrate new concepts that have emerged from systems-based analyses of CHD. A human dataset of CHD genes identified by CNVs, sequencing, or expression analyses142 was used to construct a cardiac developmental network that was enriched for functional gene-ontogeny terms indicative of crucial biological processes. Twelve dysfunctional modules in these networks were perturbed by CHD that informed the clinical phenotypes found in CHD patients better than predictions based on existing pathways. These data also predicted CHD candidate genes based upon “guilt-by-association”.142
Another approach built upon developmental programs and functional molecular networks involved in distinct anatomical cardiac structures (e.g., valves, septa, inflow, and outflow tracts.141 Analyses of genetic and environmental risks for CHD in the context of these datasets showed significant convergence of these heterogeneous risk factors on these molecular networks.143 An important conclusion from these analyses is that, although genetic and environmental factors involved in CHD impacted distinct genes involved in different pathways, these converged onto larger interaction network that collaborate to develop specific anatomical structures of the heart.143
The current repertoire of CHD genes can be epitomized by “the glass half empty”. Collectively these genes are still unable to account for the population prevalence of CHD. Despite this limitation, this dataset has illuminated one important genetic mechanism for CHD: altered levels of developmental signaling molecules involved in cardiogenesis. Physiologic levels can be perturbed by mutations that impact gene dosage, inactivate/enhance gene transcription, or that activate/inactivate a developmental pathway. This mechanism for disease is distinct from other cardiovascular genetic pathologies (e.g., hypertrophic or dilated cardiomyopathy, long QT syndrome, or Marfan syndrome) that arise from mutations in structural proteins with distinct and restricted functions in cardiovascular biology.
Discovery of other causes of CHD that affirm “the glass half empty” model may inform clinical insights. Instead of a monogenic mutation that alters gene dosage, CHD might arise from a “two-hit” model, as has been proposed in autism-spectrum disorders144 based on finding CNVs that are inherited from an unaffected parent and a de novo CNV in affected children. Inherited genetic variants that predispose to heart malformations and that require a second genetic hit to cause CHD could account for the recurrence rates observed in the offspring of CHD patients9 that are far less than recurrence rates for monogenic traits. A corollary to this hypothesis is that the collective burden or mitigating potential of altered levels of all molecules that participate in a common heart developmental pathway could account for variable clinical expression of CHD. Robust analyses of exomes, genomes and RNA expression in malformed heart tissues have the potential to test these hypotheses.
Other mechanisms for regulating gene expression and levels of encoded proteins may also contribute to CHD. Histone modifications and chromatin remodeling have substantial roles in activating or silencing gene expression.138, 145, 146 Recent studies147 that demonstrated changing patterns of chromatin modification as mouse embryonic stem cells differentiate, and that correlate with the expression of cardiac developmental transcription factors implicate epigenetic etiologies for CHD.
MicroRNAs (miRs) that regulate cardiac growth, remodeling and contribute to specific myocyte properties148–150 might promote CHD by post-transcriptional regulation of protein levels. Mice engineered to lack miR-1–2 had decreased levels of the cardiac transcription factor Hand2150 and heart malformations (VSDs) similar to those observed in Hand2-deficient mice.151, 152 Studies of malformed heart tissues from CHD patients153–155 have demonstrated dysregulation of miRs, but to date there is no evidence that sequence variation in miRs nor altered levels directly cause CHD.
Greater understanding of the fundamental mechanisms that regulate cardiac gene and protein dosage can be expected to define new causes of CHD and to empower new therapies. Development of intrauterine fetal surgery over the past decade has fostered early interventions to attenuate critical heart defects.156, 157 In the future, these approaches might incorporate molecular interventions – to improve CHD. The considerable redundancy and complexity of transcriptional regulation poses multiple levels for compensatory interventions - epigenetic manipulations, targeting transcriptional partners, supplementing downstream molecules, or reducing post-translational modifiers to increase protein levels.
So while today the glass of CHD remains half empty, ongoing research efforts will soon change this scenario. Modern genomic technologies, experimental models, and system-based approaches hold the promise for a fuller understanding of causes and mechanisms of CHD. When combined with the development of new strategies to prevent or repair heart malformations, CHD patients worldwide can anticipate a full glass of life.
This Review is in a thematic series on Cardiovascular Genetics, which includes the following articles:
Ali J. Marian, Hugh Watkins, Christine Seidman , Editors
Sources of Funding
This work was supported in part by grants from the Dubai Harvard Foundation for Medical Research (A.F., C.E.S.), the Howard Hughes Medical Institute (C.E.S.) and the National Institutes of Health (B.D.G., J.G.S and C.E.S).
In December 2012, the average time from submission to first decision for all original research papers submitted to Circulation Research was 14.5 days.
B.D.G. receives royalties from genetic testing for Noonan syndrome and related disorders from GeneDx, Correlegan, Prevention Genetics, the Baylor College of Medicine, and Harvard/Partners.