Genetics of Congenital Heart Disease

Congenital heart disease (CHD) is one of the most common birth defects. Studies in animal models and humans have indicated a genetic etiology for CHD. About 400 genes have been implicated in CHD, encompassing transcription factors, cell signaling molecules, and structural proteins that are important for heart development. Recent studies have shown genes encoding chromatin modifiers, cilia related proteins, and cilia-transduced cell signaling pathways play important roles in CHD pathogenesis. Elucidating the genetic etiology of CHD will help improve diagnosis and the development of new therapies to improve patient outcomes.


Introduction
Congenital heart disease (CHD) is a form of birth defect that affects about 1% of infants born each year. Disturbances in heart development result in a variety of defects, and while CHD can be caused by environmental exposures to teratogens [1,2], a genetic underpinning for CHD is strongly supported by the observation of a high recurrence risk and familial forms of the disease, as well as the well-described association of CHD with chromosomal anomalies [3].
It is estimated that about 400 genes are associated with CHD pathogenesis. Mutations in genes encoding transcription factors, cell signaling transducers, and chromatin modifiers can interfere with cell type specification, differentiation, and patterning important in heart development causing perturbations in heart structure and function. As many of the proteins encoded by these genes work synergistically or are connected by functional networks, this suggests a broad interacting network may be associated with disease [4,5]. However,~60% of CHD cases remain unexplained, as studies into the genetic etiology of CHD have been confounded by the genetic diversity of human subjects [6]. Also confounding genetic inquiry is the genetic heterogeneity associated with CHD. Together this has resulted in variable expressivity where subjects with the same variants may exhibit different phenotypes, or variable penetrance where some individuals with a known pathogenic variant may have no disease. As a result, CHD largely has a non-Mendelian inheritance patterns and is best described as mediated by complex genetics.
There have been several studies utilizing targeted whole-exome or whole-genome sequencing to investigate the genetic basis for CHD. In trio studies, the proband is sequenced along with unaffected parents to identify pathogenic variants that may have arisen de novo. In familial studies, multiple members of a family are phenotyped and sequenced to identify variants that are inherited in diseased family members. In cohort studies, a large number of unrelated cases and healthy control samples undergo sequencing to determine if any single gene or set of genes is enriched for variants in the disease samples. Studies of de novo and rare inherited variants have revealed a higher burden of mutation in variants predicted to be damaging in genes associated with CHD, highly expressed in

Role of Transcription Factors
A combination of clinical studies and studies using mouse models have allowed the identification of transcription factors and cofactors involved in CHD and uncovered their roles in CHD pathogenesis ( Table 1). The further identification of novel variants and CNVs has emerged from large cohort studies [21][22][23]. Transcription factors in CHD patients are also observed to be enriched for de novo and loss of function mutations [10]. Proteins with such deleterious mutations displayed changes in transcriptional or synergistic activity, which can interfere with expression of downstream targets, causing the perturbation of cell type specification, and differentiation [21].

NKX2-5
NKX2-5 encodes a homeobox transcription factor that plays an important role in heart development. It is expressed at the earliest stages of cardiogenesis, regulating cardiomyocyte differentiation and proliferation [24]. NKX2-5 mutations were first identified to cause AV block and ASD [25,26], but have since been recovered in a wide spectrum of CHD. Moreover, the phenotype and penetrance of NKX2-5 mutations have been shown to be dependent on genetic background and interaction with other mutations in both mice and humans [22,25,27,28]. Together, these findings have complicated investigations into mechanisms by which NKX2-5 mutations cause CHD. In vitro mouse modeling of a heterozygous mutation in Nkx2-5 associated with AV block and ASD showed reduced NKX2-5 nuclear import, downregulation of BMP and Notch signaling, and ultimately dysregulation of genes involved in early cardiomyocyte differentiation and function and reduced cardiomyogenesis [29].

GATA Family
GATA4, 5, and 6 are zinc finger transcription factors that have been shown to be expressed in the developing heart and have roles in cardiogenesis [30]. Mutations in GATA4 that decrease transcriptional activity have been associated with BAV and VSD [31]. Mutations in genes that regulate GATA4, such as NEXN, have also been associated with CHD [32]. Gata4 has been shown to be required by Hh-responsive progenitors within the SHF involved in OFT development, with a heterozygous Gata4 mutation shown to cause VSD and OFT defects in mice, including DORV and AVSD [33]. Noncoding variants in GATA4 have also been associated with BAV, illustrating the importance of further research into noncoding and regulatory regions of the genome [34]. Heterozygous mutations in GATA6 also have been identified in CHD patients. Studies in mice showed Gata6 mutations can cause severe OFT defects through disruption of Sema3c and Plxna2 expression [35,36]. Mice that are double homozygous knockouts for Gata4/Gata6 exhibit acardia and only generate SHF progenitor cells [33]. Mutations in GATA5 have only more recently begin to be explored as a cause for CHD. Rare sequence variants have been reported in patients with TOF, VSD, familial atrial fibrillation, and BAV [37], and loss of Gata5 results in BAV in mice [38].

T-Box Family
The TBX transcription factors are expressed throughout the developing heart and play a key role in regulating cardiomyocyte identity [18]. Mutations in TBX1, which is expressed in outflow tract precursors, have been found in patients with DiGeorge syndrome, which is commonly associated with cardiac defects. Loss of transcriptional targets of Tbx1, such as Wnt5a, also cause severe hypoplasia of SHF-dependent structures in mice, similar to loss of Tbx1 [39]. In addition, CNVs affecting PRODH and DGCR6, which have been reported to affect TBX1 expression, have been associated with conotruncal defects in DiGeorge patients [40]. TBX5 and TBX20 activate gene expression in the cardiac chambers, TBX2 and TBX3 repress myocardial gene expression in the inflow and outflow tract precursors, and TBX18 is expressed in the venous pole. Deletion of these genes in mice result in a variety of cardiac defects [41]. TBX5 and TBX20 both drive chamber formation from FHF progenitors. Mutations in TBX5 are known to cause Holt-Oram syndrome, which is characterized by heart and upper limb deformities [42]. Studies in mice showed Tbx5 interacts with both Gata4 and Gata6, such that double heterozygous mutations with Gata6 result in neonatal lethality, and double heterozygous mutations with Gata4 result in more severe cardiac malformations and embryonic lethality. Mutations in TBX20 have also been associated with CHD such as TOF, and knockdown of Tbx20 in mice suggests that it plays a role in development of the SHF [41].

Forkhead Box Family
Several forkhead box (FOX) transcription factors also play important roles in heart development, with mutations leading to cardiac defects and embryonic lethality [43]. Deletion CNVs of the FOXF1, FOXC2, and FOXL1 are associated with CHD, particularly HLHS [44]. Mutations in FOXC2 are a well-characterized cause of TOF [45]. A mutation in FOXF1 was identified in one patient with AVSD, hypoplastic LV, bicuspid aortic valve, and also intestinal malrotation, indicating disturbance of left-right patterning. Another patient with VACTERL and HTX, was also identified with a mutation in FOXF1 as well as ZIC3, both of which regulate the specification of laterality [46,47]. FOXA2 has been shown to regulate TBX1 transcription and development of the outflow tract [43]. Mutations in FOXH1, a downstream target of the Nodal pathway signaling, have been identified in patients with VSD, TGA, and laterality defects [48,49]. Mutations in Foxj1, which is a regulator of ciliogenesis, were identified to cause complex CHD with HTX in a large-scale mouse mutagenesis screen [5].

Nuclear Receptor Family
A de novo mutation in the DNA binding domain of NR1D2, a nuclear receptor transcriptional repressor that acts in a heme-dependent manner, has been identified in a cohort of patients with AVSD [50]. It was shown to change transcriptional activity, and knockout mice were shown to have cardiovascular malformations. Another nuclear receptor, NR2F2 encodes a pleiotropic transcription factor shown to be required for normal development of the atria, coronary vessels, and aorta [51]. In a mouse model, cardiomyocyte-specific knockout of Nr2f2 resulted in ventricularized atria. A mutation in NR2F2 was found to segregate with disease in a family with DORV and VSD and absent in ethnically matched controls [52]. This mutant Nr2f2 protein has no transcriptional activity in a mouse model, eliminating synergistic transcriptional activation between NR2F2 and GATA4. Mutations that alter NR2F2 transcriptional activity with preserved repressor function were identified in patients with AVSD, TOF, aortic stenosis, CoA, and HLHS [53].

HAND Family
HAND1 and 2 are helix-loop-helix transcription factors that regulate, in a dose-dependent manner, the expansion of ventricular precursors [54]. In Hand1 null mice, heart development is arrested at the heart looping stage of development [55]. Hand1 conditional activation knock-in mice have increased expansion of the outer curvature of both ventricles but lack the interventricular groove and have a defect in formation of the septum. A mutation in HAND2 has been associated with VSD, and HAND2 may have synergistic activation effects with GATA4 and NKX2-5 [56]. Many other transcription factors have also been shown to cause CHD when mutated, and the phenotypes resulting from the disruption of many of these are described in Table 1.

Nodal Signaling
An important signaling pathway in cardiovascular development is the Nodal signaling pathway known to regulating left-right patterning. Central to left-right patterning is Nodal expression that is restricted to the left side of the developing embryo. This initiates a signaling cascade that establishes left-right asymmetry. In CHD patients, there is evidence of the enrichment of heterozygous damaging de novo and loss-of-function mutations in NODAL [10]. NODAL mutations were identified in patients with TGA and a family history of CHD [49]. De novo CNVs affecting NODAL were also identified in a cohort of patients with conotruncal defects or HLHS [82]. Mutations in ZIC3, a transcription factor that functions upstream of NODAL, were identified in the aforementioned study, as well as in a study of CHD patients with HTX [83,84]. Mutations in several downstream targets of NODAL-GDF1, CFC1, TDGF1, FOXH1, and SMAD-were also identified in a cohort of CHD patients. Another downstream target of NODAL, PITX2, encodes a paired-like homeobox domain transcription factor that is a core effector of left-right patterning. A nonsense mutation identified in a family with endocardial cushion defect and Axenfeld-Rieger syndrome, which is associated with OFT defects, eliminates its transcriptional activity and synergistic transcriptional activation with NKX2-5 [85].

Notch Signaling
Signaling through the Notch pathway regulates cardiac cell fate and morphogenesis of cardiac chambers and valves [86]. Notch regulates EMT of the AV cushion progenitor cells which later contribute to the AV septum [87]. Rare deleterious variants in NOTCH1 were identified in patients with strong family histories of disease [88]. NOTCH1 mutations have previously been associated primarily with left-sided lesions, but a study of NOTCH1 mutations in familial CHD identified individuals with right-sided and conotruncal defects [89]. While rare predicted loss-of-function and intronic variants in NOTCH1 increase risk for left ventricular outflow tract defects [90], rare or likely pathogenic variants in NOTCH1 have also been identified in a cohort of BAV patients requiring aortic root replacement [69], and de novo and rare variants were identified in patients with HLHS [91,92]. In addition, rare or novel protein-altering mutations in Notch pathway genes NOTCH1, ARHGAP31, MAML1, SMARCA4, JARID2, and JAG1 were shown to co-segregate with disease in families with left ventricular outflow tract defects, and an enrichment of pathogenic variants in these genes in patients vs. controls was observed [93]. Heterozygous rare coding mutations in MIB1, which activates the Notch pathway through promoting ubiquitination, endocytosis, and activation of Notch ligands, were identified in a Han Chinese CHD cohort. Two of these mutations were shown to reduce function, resulting in less JAG1 ubiquitination and the induction of Notch [94]. This upstream effector JAG1 is also associated with TOF [45]. Expression of Notch and its downstream targets are also reduced in mice with mutations in the Slit/Robo signaling pathway, resulting in membranous VSDs and BAV [95].

Wnt/β-Catenin Signaling
The Wnt/β-catenin pathway has an important role in many different aspects of heart development, including the regulation of cell proliferation in the SHF [96]. The recovery of candidate CHD genes in the Wnt pathway was observed in patients with bicuspid aortic valve (BAV) [37]. Enrichment for de novo variants in Wnt pathway genes has also been observed in CHD patients with neurodevelopmental defects, suggesting a shared genetic etiology [97]. Deletion of Apc, a negative regulator of canonical Wnt signaling, leads to ventricular hypoplasia in mice [98]. Context-dependent regulators of the Wnt pathway such as Bcl9 and Pygo are also associated with cardiac defects, such as AVSD in mice or TOF in humans [99]. Canonical Wnt signaling is regulated by interactions between Dkk1/2, and mice that are double knockouts for Dkk1 and Dkk2 exhibit myocardial and epicardial hypoplasia, as well as VSD in later stages of development [100]. Non-canonical Wnt signaling also has been shown to activate the planar cell polarity (PCP) pathway, which coordinates processes such as chamber remodeling through actomyosin polarization and also regulates ciliogenesis [101][102][103]. Several core members of the PCP pathway were identified to cause cardiac defects in a mouse forward genetic screen [5]. Together with the finding of enrichment in other cilia-related genes, they indicate the importance of the PCP pathway in heart development and disease.

Bmp Signaling
Bmp signaling is required for specification and differentiation of the cardiac mesoderm and it regulates Nkx2-5 expression through a negative feedback loop [96,104]. BMP4 deficiency can cause septal defects, defective endocardial cushion remodeling, and abnormal pulmonary valve formation, and common variants in BMP4 are associated with CHD in a Han Chinese cohort [105]. Nonsynonymous variants in SMAD6, an inhibitor of Bmp signaling, have been identified in CHD patients [106]. Furin deletion targeted to endothelial cells in mice can reduce Bmp4 and Et1, causing VSDs and valve malformations [107]. Also recovered were multiple de novo variants in SMAD2 [108,109], which transduces Bmp signaling by regulating downstream target gene transcription [109]. De novo protein-truncating, splicing, and deleterious missense variants in SMAD2 were identified in a cohort of CHD patients with a variety of defects including complex CHD with or without laterality defects and other congenital anomalies and late-onset vascular phenotype [110]. Mutations have also been recovered in GALNT1, a glycosyltransferase that can increase Bmp and Mapk signaling, causing aberrant valve formation due to increased cell proliferation in the outflow cushions [111]. Other studies suggest BMP10 plays a role in maintaining expression of NKX2-5 and other key cardiogenic factors to regulate cardiac growth [104]. HIC2 encodes a transcriptional repressor that may regulate BMP10 in the FHF lineage specified by NKX2-5 and MESP1. HIC2 is impacted by the 22q11 deletion associated with DiGeorge syndrome [112].

Sonic Hedgehog (SHH) Signaling
SHH signaling has been shown to play an important role in the development of the SHF, outflow tract septation, and proper outflow tract alignment [113,114]. SHH is secreted from the pharyngeal endoderm, and ligand is received by SHF cells, maintaining proliferation of these progenitor cells (Figure 1) [113]. GATA4 was shown to be required for proliferation of SHH-receiving cells and subsequent OFT alignment, and Gata4 mutations in mice cause DORV [115]. Signaling from BMP2 and BMP4 in the outflow tract myocardium, conversely, represses proliferation of SHH-receiving cells, with overexpression leading to premature differentiation of SHF cells and knockout resulting in embryonic lethality (Figure 1) [116]. SHH regulates development of SIX2+ progenitor cells, which contribute to the right ventricle, inflow tract, pulmonary trunk and ductus arteriosus [117]. Ablation of Six2+ cells in mice was shown to result in severe CHD such as common arterial trunk. SHH is also required for migration of cardiac neural crest cells to the OFT cushion, with SHH mutations in mice resulting in neural crest cell death and mislocalization ( Figure 1) [114]. Mutations in Megf8 can cause TGA or other complex CHD associated with HTX [118]. While Megf8 was previously proposed to regulate Tgfβ/Nodal signaling, a CRISPR screen recently identified Megf8 as a negative regulator of SHH signaling [119]. Moreover, another negative SHH regulator identified in the same screen, Mgrn1, was also previously shown to cause HTX with CHD, with the CHD comprising TGA [120]. In fact, the role of SHH in human CHD has not been systematically examined, but the recovery of other regulators of SHH signaling among mutations causing CHD from a large scale mouse mutagenesis screen would suggest this pathway is likely to play an important role in human CHD [5].

Sonic Hedgehog (SHH) Signaling
SHH signaling has been shown to play an important role in the development of the SHF, outflow tract septation, and proper outflow tract alignment [114,115]. SHH is secreted from the pharyngeal endoderm, and ligand is received by SHF cells, maintaining proliferation of these progenitor cells (Figure 1) [114]. GATA4 was shown to be required for proliferation of SHH-receiving cells and subsequent OFT alignment, and Gata4 mutations in mice cause DORV [116]. Signaling from BMP2 and BMP4 in the outflow tract myocardium, conversely, represses proliferation of SHH-receiving cells, with overexpression leading to premature differentiation of SHF cells and knockout resulting in embryonic lethality ( Figure 1) [117]. SHH regulates development of SIX2+ progenitor cells, which contribute to the right ventricle, inflow tract, pulmonary trunk and ductus arteriosus [118]. Ablation of Six2+ cells in mice was shown to result in severe CHD such as common arterial trunk. SHH is also required for migration of cardiac neural crest cells to the OFT cushion, with SHH mutations in mice resulting in neural crest cell death and mislocalization (Figure 1) [115]. Mutations in Megf8 can cause TGA or other complex CHD associated with HTX [119]. While Megf8 was previously proposed to regulate Tgfβ/Nodal signaling, a CRISPR screen recently identified Megf8 as a negative regulator of SHH signaling [120]. Moreover, another negative SHH regulator identified in the same screen, Mgrn1, was also previously shown to cause HTX with CHD, with the CHD comprising TGA [121]. In fact, the role of SHH in human CHD has not been systematically examined, but the recovery of other regulators of SHH signaling among mutations causing CHD from a large scale mouse mutagenesis screen would suggest this pathway is likely to play an important role in human CHD [5].

Ras/Mapk Signaling
The Ras/Mapk pathway, which regulates proliferation, growth, and other cell processes, is also known to play important roles in CHD. Thus, disruption of the Ras/Mapk pathway results in a number of related disorders collectively termed RASopathies, the most common of which is Noonan syndrome. Noonan syndrome has the highest incidence of CHD, particularly pulmonary stenosis, among RASopathy patients [3]. PTPN11, which encodes an upstream regulator of the Ras pathway, is well known to cause Noonan syndrome and is enriched for de novo mutations in a cohort of syndromic CHD patients [16]. A de novo mutation in MRAS, which contributes to ERK activation and downstream Mapk signaling, was identified in a patient with Noonan syndrome and cardiac

Ras/Mapk Signaling
The Ras/Mapk pathway, which regulates proliferation, growth, and other cell processes, is also known to play important roles in CHD. Thus, disruption of the Ras/Mapk pathway results in a number of related disorders collectively termed RASopathies, the most common of which is Noonan syndrome. Noonan syndrome has the highest incidence of CHD, particularly pulmonary stenosis, among RASopathy patients [3]. PTPN11, which encodes an upstream regulator of the Ras pathway, is well known to cause Noonan syndrome and is enriched for de novo mutations in a cohort of syndromic CHD patients [16]. A de novo mutation in MRAS, which contributes to ERK activation and downstream Mapk signaling, was identified in a patient with Noonan syndrome and cardiac hypertrophy [121]. Noonan syndrome patients were also identified with heterozygous de novo and inherited mutations in A2ML1, which may act upstream of the Ras signaling pathways. However, in cell lines, expression of A2ML1 did not activate the Ras/Mapk pathway [122].

Vegf Signaling
The Vegf signaling pathway is required for formation of the AV endocardial cushions and their morphogenesis into AV valves [123]. In a cohort of TOF patients, predicted damaging variants were identified in the Vegf-related genes FLT4, KDR, VEGFA, FGD5, BCAR1, IQGAP, FOXO1, and PRDM1. These variants are associated with absent pulmonary valve and right aortic arch [67]. In a cohort of patients with Down syndrome and AVSD, variants with the highest probability of being damaging in cases compared to Down syndrome patients without cardiac defects were in the VEGF-A pathway genes COL6A1, COL6A2, CRELD1, FBLN2, FRZB, and GATA5 [123]. Signaling pathway genes that have been shown to cause CHD when mutated in mice and humans, as well as their resulting phenotypes, are described in Table 2.

Myofilament and Extracellular Matrix Proteins
Proteins that compose the sarcomere and extracellular matrix are essential for proper structure and function of cardiac muscle. Mutations in ACTC1, DCHS1, TTN, ELN, MYH6, MYH7, and MYH11 are known to cause cardiac defects [3]. MYH6 mutations have been associated with atrial septal defects (ASD) and recently were shown to be significantly associated with CoA in a GWAS study of an Icelandic population [124]. TPM1, an essential component of the sarcomere, has been associated with cardiomyopathy [125]. Mutations in the cytoskeletal protein ACTC1 cause ASD that is thought to arise from cardiomyocyte apoptosis [126,127]. The actin-binding protein NEXN has also been associated with ASD [32]. Genes that regulate splicing of essential cardiac genes are also known to cause CHD. The splicing factor RBM20 regulates alternative splicing of genes associated with diastolic function and ion transport, as well as sarcomere assembly, particularly TTN where greater RBM20 expression is associated with the expression of shorter isoforms of TTN [128]. In mice, mutations in Rbm20 result in dilated cardiomyopathy (DCM) with similar severity to Ttn mutations, and arrhythmia that is more severe than Ttn mutations, indicating a role for other Rbm20 targets in disease [129].
Cells must be able to respond and adhere to other cells and the extracellular matrix to maintain structure and transduce intracellular signaling. In mice, deficiency in the matrix protein Ccn1, which regulates cell adhesion and migration, proliferation, survival, and differentiation, results in severe AVSD [130]. Mutations in BVES, a cell adhesion protein, were identified in TOF patients. One Bves mutation was shown to alter transcriptional activity in a cell based assay [131]. Pcdha9, encoding a protocadherin cell adhesion protein, was shown to have an essential role in valvular morphogenesis, as Pcdha9 mutation can contribute to the aortic hypoplasia/atresia in HLHS and also can cause bicuspid aortic valve (BAV) [79].

Chromatin Modifiers
Chromatin modifiers regulate the epigenetic marks that control DNA accessibility and transcriptional activity. Disruption of these processes can interfere with transcriptional programs important for orchestrating events in cardiovascular development. Chromatin modifiers were found to be enriched among genes with de novo mutations in a CHD cohort with diverse phenotypes including LVOTO, conotruncal defects, and HTX [10]. Several genes involved in the regulation of active H3K4me/inactive H3K27me histone marks were identified. KMT2D encodes one of these histone modifiers and is associated with Kabuki Syndrome with CoA, ASDs, and VSDs [152]. Mutations have also been recovered in CASZ1 encoding a zinc finger transcription factor that interacts with histones and is essential for cardiogenesis. A CASZ1 mutation associated with reduction in transcriptional activity caused VSD as a completely penetrant autosomal dominant trait [153].
The HDAC repressor complex plays a key role in many developmental processes, and several proteins that are associated with this complex are associated with CHD [154]. Thus, variants in SMYD4, a protein which interacts with HDAC1 and can modulate histone acetylation [155], were identified in patients with DORV and TOF. Genes regulating chromatin were also identified in Smarca4 and Prdm1 in a mouse forward genetic screen for CHD [5]. Another mutant recovered from the same screen harbored a CHD-causing mutation in Sap130, a Sin3A associated protein that is also part of the HDAC repressor complex. Mutation in Sap130 was shown to mediate left ventricular hypoplasia [156]. Double homozygous Pcdha9 and Sap130 mutations were shown to cause HLHS, with the Pcdha9 mutation found to drive the aortic valve phenotype associated with HLHS [79].

The Role of Cilia and Cilia-Transduced Cell Signaling During Cardiogenesis
The cilium is an organelle that protrudes from the cell surface and can be motile or nonmotile. Motile cilia are involved in cell motility and the generation of extracellular fluid flow, such as in mediating mucociliary clearance in the airway or cerebral spinal fluid flow in the brain. During early embryonic development, motile cilia in the embryonic node generate flow responsible for creating a gradient of signaling molecules, such as NODAL, that establishes left-right patterning. This is essential for normal cardiac morphogenesis, as disruption of left-right patterning causing HTX is associated with some of the most complex forms of CHD. Nonmotile cilia, known as primary cilia, can function as cell signaling transducers or serve as mechanosensors. Cilia and cilia-transduced cell signaling can modulate planar cell polarity and affect cytoskeletal organization involved in the regulation of EMT. This is essential for emergence of neural crest cells from the neural tube, epicardially-derived cells from the epicardium, and development of the cardiac cushion mesenchyme from endocardial EMT [157]. In addition, cell signaling pathways known to play essential roles in heart development, such as Wnt, Tgfb/Bmp, and SHH are all cilia-transduced ( Figure 2) [158].

The Role of Cilia and Cilia-Transduced Cell Signaling During Cardiogenesis
The cilium is an organelle that protrudes from the cell surface and can be motile or nonmotile. Motile cilia are involved in cell motility and the generation of extracellular fluid flow, such as in mediating mucociliary clearance in the airway or cerebral spinal fluid flow in the brain. During early embryonic development, motile cilia in the embryonic node generate flow responsible for creating a gradient of signaling molecules, such as NODAL, that establishes left-right patterning. This is essential for normal cardiac morphogenesis, as disruption of left-right patterning causing HTX is associated with some of the most complex forms of CHD. Nonmotile cilia, known as primary cilia, can function as cell signaling transducers or serve as mechanosensors. Cilia and cilia-transduced cell signaling can modulate planar cell polarity and affect cytoskeletal organization involved in the regulation of EMT. This is essential for emergence of neural crest cells from the neural tube, epicardially-derived cells from the epicardium, and development of the cardiac cushion mesenchyme from endocardial EMT [158]. In addition, cell signaling pathways known to play essential roles in heart development, such as Wnt, Tgfb/Bmp, and SHH are all cilia-transduced ( Figure 2) [159].  A central role for cilia in CHD pathogenesis was discovered via the use of forward genetics in mice with ENU mutagenesis to recover mutations causing CHD. Cardiovascular phenotype was assessed using fetal echocardiography, a noninvasive high-throughput phenotyping method that is also highly sensitive for the detection of CHD and allowed the screening of 100,000 fetal mice. While the screen was entirely phenotype-driven, surprisingly 50% of the mutations recovered causing CHD were cilia related. This encompassed mutations in 30 genes related to cilia and ciliogenesis ( Figure 3). Additionally, the screen also recovered many genes involved in cilia-transduced cell signaling ( Figure 2) and in vesicular trafficking (Figure 4), a cell process critical for ciliogenesis and cilia-transduced cell signaling [5]. A separate mouse screen also identified mutations in Dnah11, an axonemal protein, and Mks1, a basal body protein, to be associated with CHD [159]. Mutations in these same genes were also recovered in the large scale fetal mouse CHD screen. The ciliary gene Ift88 is an intraflagellar transport protein required for cilia formation, and Ift88 null mutant mice exhibited OFT defects. [160,161]. Cilia have also been shown to play a role in aortic valve disease, such as BAV [162]. Defects in development of the AV cushions in Cc2d2a mutant mice were associated with loss of cilia from the AV cushions ( Figure 5). In human studies, an enrichment of ciliary genes was observed among genes with damaging recessive variants in a CHD cohort [10]. We note analysis of the early mouse embryos has revealed primary cilia in the endocardium of the atria, the endocardial cushions, and the cushion mesenchyme, as well as in the epicardium [163]. Ciliary defects have only recently been identified as a cause of CHD, and their role in the developmental processes of the heart and the contribution to CHD pathogenesis warrants further studies. such as BAV [163]. Defects in development of the AV cushions in Cc2d2a mutant mice were associated with loss of cilia from the AV cushions ( Figure 5). In human studies, an enrichment of ciliary genes was observed among genes with damaging recessive variants in a CHD cohort [10]. We note analysis of the early mouse embryos has revealed primary cilia in the endocardium of the atria, the endocardial cushions, and the cushion mesenchyme, as well as in the epicardium [164]. Ciliary defects have only recently been identified as a cause of CHD, and their role in the developmental processes of the heart and the contribution to CHD pathogenesis warrants further studies.

Maternal Effects
Maternal genetics and behavior should also continue to be studied in relation to their effects on fetal cardiac development, as changes in the fetal environment have been associated with CHD. Congenital heart disease has been associated with maternal smoking, parental age, and maternal fertility and nonfertility medications [164], as well as maternal obesity [165], maternal alcohol consumption [166], and maternal viral infection [167]. CHD pathogenesis in these cases has been attributed to impacts on placental development [168], overactive maternal immune responses [169], and deficiency of folic acid, which is essential for fetal growth and development [170,171].

Future Directions
CHD is a heterogeneous disease with complex genetics underlying its pathogenesis. While a large body of evidence points to CHD being genetically heterogeneous, there may be a central role for cilia and chromatin modifiers in driving the complex genetics of CHD. However, the molecular mechanisms driving CHD pathogenesis are still not well understood. Mouse models with genetic mutations causing CHD is an invaluable resource for further mechanistic studies. Findings from these animal models may help guide assessments and validation of the role in disease of various sequence variants recovered from patients with CHD. Such pairing of animal studies with clinical findings may give novel insights not only into molecular mechanisms of human CHD, but the animal models generated may provide the means to develop therapies that may have improve outcome for patients with CHD. Similarly, large-scale studies of human cohorts will continue to reveal novel variants that are relevant to disease and their effects on phenotype and outcome. Stratification of analyses based on specific phenotypes, patient outcomes, and variant predictions will further reveal the genetic architecture underlying CHD. In addition, greater focus on common and non-coding variants can help uncover the role that these variants play in disease, particularly in the context of known rare and deleterious variants. Further investigation into epigenetics and the effects of maternal genetics will also be needed to obtain a full picture of the risk factors contributing to the penetrance and pathogenesis of CHD.
Funding: This research was funded by HL132024 and HL142788.

Conflicts of Interest:
The authors declare no conflict of interest.