[edit] Etiology of human heart defects

[edit] definition & classification

Congenital heart defects (CHDs) are a group of diseases characterized by a structural anomaly of the heart that is present at birth. They do not necessarily manifest themselves in the neonate and can remain benign throughout life. They are distinguished from the cardiomyopathies, which are diseases of the heart muscle, and the cardiac rhythm disorders, diseases of the heart's pacemaking and conduction system. These three cardiac diseases can and often will co-occur in one patient, and are sometimes interlinked. Structural heart defects can in itself cause heart muscle overgrowth or dilatation, which in turn can lead to rhythm abnormalities. These disorders can also co-occur due to a common underlying cause, so without one being the functional consequence of the other. For example, NKX2-5 mutations can provoke conduction abnormalities and/or structural defects, and ACTC mutations can cause cardiomyopathies and/or isolated septal defects.

CHDs arise from errors in cardiac development. Although attempts have been made to classify CHDs according to specific cardiac developmental processes that have gone awry, a descriptive anatomical classification is generally favored, since it does not invoke the assumption that one knows what part of development is at the base of the CHD. Anatomically, various defects in the septa, valves, inflow and outflow regions of the heart are distinguished. An example of such a classification system is the AEPC list of cardiac codes (link).

From a clinical perspective, one can distinguish syndromic and non-syndromic (or isolated) CHDs. This distinction is a practical one made by clinicians, as it reflects the risk for a patient to develop further complications unrelated to the CHD and since it is often the first step to reach an etiological diagnosis (Meberg et al., 2005). A first group of individuals considered to have a syndromic CHD are those with multiple congenital anomalies (MCA), i.e. having a second major congenital abnormality which is unrelated to the CHD (e.g. malformations the kidney, skeleton, brain, developmental delay, …). Major malformations have clear medical, functional or esthetic consequence for the individual Leppig et al., 1987). Other individuals present multiple (i.e. three or more) minor anomalies (=i.e. physical variants without functional or esthetic consequences, that are found rarely occur in the normal population (Carey et al., 1996).

Depending on what heart defects are included in the study (e.g. bicuspid aortic valve), syndromic CHDs make up 12-22% of all CHDs (Stephensen et al., 2004; Meberg et al., 2007). Syndromic and isolated CHDs are thought to have a different etiology: syndromes arise from a single and often genetic cause, whilst isolated CHDs often have a multifactorial cause, related to the interplay between multiple genetic and environmental factors. However, as will be outlined below, sometimes syndromes are caused by a teratogens or multiple genetic factors and also isolated CHDs can in rare instances be ascribed to a single genetic cause. Of note, the distinction between syndromic and isolated CHDs is purely clinical. For that reason, it is partly arbitrary since currently no universally used definitions for minor physical anomalies or firm criteria for ‘dysmorphism’ exist. Moreover, syndromic versus non syndromic do CHD not not necessarily implicate different genes: for example, TBX5 mutations cause CHDs and hand malformations in most patients (=syndromic CHD), but in some patients with the same mutation, a CHD is the only apparent expression (=isolated CHD).

[edit] environmental causes

In a minority of cases, environmental factors are shown to cause the cardiac defect. Well documented examples include development of CHDs upon antenatal exposure to teratogens such as alcohol or antiepileptic drugs. These have been reviewed extensively by Jenkins et al. 2007. One of the major teratogenic risk factors for cardiovascular malformation is pregestational diabetes, associated with up to an 18 fold increase in the risk of a varying spectrum of CHDs at birth (Jenkins et al. 2007). Outflow tract defects (like TGA and truncus arteriosus), complete AVSD and also cardiomyopathies are overrepresented heart diseases, while outflow tract obstruction defects are less typical (Loffredo et al., 2001; Wren et al., 2003). Also extra-cardiac malformations are overrepresented in children from diabetic mothers, often fitting in the VACTERL-like spectrum (Loffredo et al., 2001).

Another known risk factor in the fetal environment is the presence of a second fetus: twin pregnancies sharing the same chorion confer an increased risk for CHDs. This risk factor was the subject of a review by Manning. The frequency of CHDs in dizygotic twin pregnancies is thought to be 1% per fetus, similar to the population risk (Manning). However, the concordance for CHDs dizygotic twin pairs is estimated at 13.6% (Caputo), which is higher than the recurrence risk for the nearest non-twin sibling (4% in the same study). This increased recurrence risk can either be secondary to the twinning and the concomitant altered fetal environment (in utero 'crowding'), or a reflection of the equal effect of external, environmental factors on cardiac development.

In monozygotic twins, the risk is influenced by the chorionic and amniotic structures. The risk of CHDs in dichorionic monozygotic twins is not known. Monochorionic diamniotic (MC/DA) twins however are at an increased risk for CHDs (5-7%), and the concordance rate in MC/DA twins for the presence of a CHD is 25% (Manning & Archer,2006, Bahtiyar et al., 2007). One factor involved in this abnormal heart development are the altered hemodynamics during development. Placental vascular anastomoses leading to twin-to-twin transfusion (TTT) syndrome are strongly implicated in this changed hemodynamics. A systematic literature review on the TTT effect showed an increase in CHD frequency from 3.4% in MC/DA twins without TTT to 8% in MC/DA twins with TTT (Bahtiyar et al., 2007). In monoamniotic (MA) twins the risk is thought to be even higher. However, since they occur rarely, reliable risk estimates are lacking. Manning & Archer reported a CHD in 4 out of 7 MA twins. Interestingly, these included 2 right atrial isomerisms, a very rare defect. This suggests that certain CHDs can be caused by the late embryonic timing of the twinning event (Manning, 2008).

It is evident that studies to unveil the genetic contributions to CHDs using twin cohorts do not provide unambiguous answers. However, it remains a challenge to explain the discordance in cardiac development observed in dichorionic monozygotic twins where altered hemodynamics do not play a role. Possible explanations include postzygotic mutations including differences in copy number variations (CNV's) (Bruder et al., 2008), as well as epigenetic differences originating postzygotically, differences in X-inactivation patterns, and stochastic factors. The latter include both chance “catastrophic' events which may for instance lead to disorders such as oculo-auriculo-vertebral spectrum (Digilio et al., 2008), a condition suspected to be caused by vascular disruption (Hartsfield, 2007). Stochastic factors during morphogenesis are also thought to play a role (Kurnit et al., 1987; Samoilov et al., 2006)

[edit] genetic causes

[edit] population studies

Population studies of CHDs show that they occur at a rate of a little under 1%. Since they are more frequent in offspring of individuals with a CHD, genetic factors contribute to their etiology. This recurrence risk is higher for offspring (4.1 %) than for siblings (2.1 %) of CHD patients (Burn et al., 1998), suggesting a major autosomal dominant inheritance pattern. However, there are numerous studies where CHD patients are much more likely to have sibs affected with a CHD than parents affected with a CHD. One example is patients with a hypoplastic left heart (HLH): there are more CHD sibs of HLH patients (30.5%) than CHD parents (10.5%). Also for aortic coarctation patients, sibs more often also have a CHD (12.8%) than parents (6.5%) (Loffredo et al., 2004). This contrast between the incidence in children and in parents of CHD patients may seem contradictory, as it suggests different vertical transmission rates. However, it probably is a reflection of the reduced fecundity and increased mortality that is associated with CHDs such as HLH. It also suggests that, together with the increasing CHD prevalence (because the improved survival rates following advances in surgical management of CHDs) the incidence of CHDs will increase: more CHD patients will have children. Consequently, the genetic causes of CHDs will increase, rendering a thorough understanding of the etiology of CHDs even more crucial.

The CHD risk for children of mothers with a CHD (2.3%-5.7%) is consistently higher than for those of fathers with a CHD (1.3%-2.2%) (Romano-Zelekha et al., 2001; Gill et al., 2001; Burn et al., 2001). There are several possible explanations for the increased transmission of CHD through females.

• One explanation is an lower susceptibility of females to develop a CHD. Indeed, the frequency of CHDs in females in lower than in males (CHD male/ female ratio = 1.25 vs population male/female ratio = 1.06) (Pradat et al., 2003). Females apparently need an increased genetic/environmental burden to develop a CHD. Since this higher genetic burden is transferred to subsequent generations, and environmental factors are often also impacting subsequent generations, it confers an increased recurrence risk through a mother with a CHD compared to a father with a CHD.
• This may indicate a potential involvement of the mitochondrial genome, but no mitochondrial mutations have been identified so far.
• Also paternal imprinting of CHD causing genes could explain the increased transmission through the female germline. No evidence exists for such a mechanism.
• Potential confounders should however be taken into account:
  • these studies are often performed using mailing questionnaires, and the response rate for fathers is consistently lower than for mothers.
  • Also non-paternity might be a confounder. However, this cannot explain the entire increase in risk.

[edit] classifications of genetic causes

Several classifications exist for genetically caused CHDs. One can classify mutations based on:

• the genetic lesion: monogenic versus chromosomal CHD
• the inheritance pattern and the pathogenic mechanism of the mutation
• the number of genetic components: monogenic, oligogenic, polygenic

[edit] monogenic and chromosomal CHDs

One classification is based on the genetic lesion, and distinguishes monogenic and chromosomal CHDs. This distinction was made historically based upon the detection technique, i.e. the presence of a chromosomal aberration detected by karyotyping versus the detection of a small change on the base pair level by direct sequencing. This distinction based upon detection technique is rapidly becoming blurred. Novel techniques for genome-wide copy-number determination such as array-CGH (array comparative genomic hybridization) are enhancing the resolution of karyotyping and can detect chromosome imbalances that were previously only detectable by PCR and sequence analysis. Moreover, the advent of genome-wide sequencing promises to also move the boundary of sequencing techniques into the field of chromosome aberration detection. Despite this blurring boundary between detection techniques for monogenic and chromosomal CHDs, some features still warrant this distinction:

Chromosomal CHDs are caused large chromosome imbalances that are rare and affect multiple genes. Typical examples of chromosomal CHDs include:

• AVSDs in 40% of trisomy 21 (Down syndrome) (Freeman et al., 2008)
• ASD, VSD and PS in 33% of terminal deletions of chromosome 4p (Wolf-Hirschhorn syndrome) (Maas et al., 2008; Wiekzorec et al., 2000; Battaglia et al., 1999)
• aortic abnormalities (CoA, AS, BAV and transverse arch elongation) in 50% of monosomy X (Turner syndrome) (Sybert et al., 1998; Ho et al., 2004)

This type of genetic lesions is practically absent from the normal population: it is mostly incompatible with a normal development and therefore results in a syndromic presentation. Consequently, it occurs mostly de novo. It can cause contiguous gene syndromes, where different parts of the patients phenotype can be explained by different genes affected by the imbalance. A classic example is Williams-Beuren syndrome (OMIM:194050), caused by the recurrent deletion of chromosome 7q11.23. Haplo-insufficiency of one of the genes in the deleted region, ELN, is responsible for the supravalvar aortic stenosis that is commonly present in these patients (Li et al., 1997; Tassabehji et al., 1997), while one or more other genes (LIMK1, GTF2IRD1, GTF2I and CYLN2) are responsible for the behavioral and developmental characteristics and dysmorphic features of patients (Wu et al., 1998). Studies using array CGH have shown that insertions and deletions (indels) beyond the resolution of standard karyotyping can also cause syndromic CHDs (Thienpont et al., 2007). These indels can affect one or more genes, and demonstrate the limitations of this classification.

Monogenic CHDs are historically considered to be caused by small base pair changes (<1kb) that affect a single gene. Each individual carries such small nucleotide variants throughout his genome, on average 1 every 1000 bp, and these typically are polymorphisms. Only a subset of these are disease-causing, when they functionally affect a gene critical for heart development. When this gene is also critical for other developmental processes, mutations can result in a syndromic CHD.

[edit] inheritance pattern of the CHD

Another classification of genetic causes of CHDs is based on the pattern of inheritance. One can discriminate autosomal dominant, autosomal recessive, X-linked, (Y-linked), mitochondrial, and non-traditional patterns of inheritance (for example: mosaicism, genomic imprinting or unstable mutations (e.g. myotonic dystrophy). Such a classification assumes high penetrance of the CHD disorder; otherwise no pattern of inheritance can reliably be deduced from the pedigree.

[edit] Autosomal dominant

Most mutations identified in CHD patients are inherited in an autosomal dominant fashion (38/51 genes in CHDWIKI (July 2008)), meaning that the mutation affects a single allele residing on an autosome. Of these, the majority (22/38) probably affect gene function through haploinsufficiency: one single allele is insufficient to effectuate a normal function during development. Genes susceptible to haploinsufficiency ('dosage sensitive genes') typically belong to categories such as signaling molecules (4/21), transcription factors (11/21) and transcription modulators, e.g. through chromatin remodeling (6/21) where a precise dosage is critical for a normal function. This also explains why both an increased and a decreased dosage of the gene may lead to a phenotype, as for instance seen for TBX1: both deletions and duplications of the 22q11.2 region have been found in CHD patients (Portnoï et al., 2005), and similarly both mutations that decrease (loss-of-function) and increase (gain-of-function) TBX1 activity have been shown to cause DiGeorge syndrome (Zweier et al., 2007).

Dominant mutations can also function through a dominant negative effect, where the mutant protein inhibits the function of the protein formed by the normal allele, e.g. during a process of dimerization. This has been described for some JAG1 mutations in Alagille syndrome, (Boyer-Di Ponio et al., 2007) and certain mutations in TFAP2B causing Char syndrome. (Satoda et al., 2007). Other examples include genes encoding structural proteins of the heart muscle, such as MYH6, MYH11 or ACTC. Mutations in these sarcomere components can cause sarcomere disarray and the concomitant muscle dysfunction. It has been suggested that this reduced cardiac muscle function may cause hemodynamic changes during development, resulting in CHDs.

A third mechanism for the dominant effect of a mutation is through gain-of-function. In this case , the mutant protein acquires a new function causing it to be active at a time or on a place where it normally is not active. Activating mutations in different components of the Ras/MAPK pathway are known to cause Noonan syndrome and related disorders such as CFC and Costello syndrome. The mutant gene encodes a constitutively active form of the protein where the normal auto-inhibitive properties are alleviated.

[edit] Autosomal recessive

When both alleles of an autosomal gene need to be mutated for a phenotype to become manifest, the phenotype will segregate in an autosomal recessive fashion. The patient carries the mutation in a homozygous (or compound heterozygous) form and usually both parents carry a heterozygous mutation. Recessive disorders are more frequent in isolated populations where a few rare (mutant) alleles are frequently present (founder mutations), and in children from consanguineous relationships, as they are homozygous for multiple regions of their genome. Recessive mutations are usually loss-of-function mutations.

There are a few known forms of CHDs with a recessive inheritance pattern (8/51 genes), and practically all (7/8) are syndromic: EVC, EVC2, LBR, MGP, NPHP3, ROR2 and SLC2A10. One exception are NKX2-6 mutations, that were described by Heathcote et al. in 3 patients with a truncus arteriosus that were all from a single consanguineous family (Abushaban et al., 2003).

CHDs are more frequent in children from consanguineous parents, with a reported relative risk of 1.8 to 2.3 for first cousin parents (Nabulsi et al., 2003; Yunis et al., 2006). This can be explained by an increased genetic vulnerability of these children for any multifactorial malformation, or by the increased risk for an autosomal recessive disorder. Especially ASDs seem overrepresented in children from consanguineous marriages (Nabulsi et al., 2003; Yunis et al., 2006). Interestingly, ASDs are also more frequent in sibs than in parents of isolated ASD II patients (8.5% vs 1.6%), suggesting an important recessive component in the etiology of ASDs (Caputo et al., 2005).

[edit] X-linked

When a mutation affects a gene located on chromosome X and manifests a phenotype when a single allele is present (i.e. typically in an XY male), the disorder has an X-linked recessive inheritance. Most of these mutations are associated with a loss-of-function.

In CHDWiki (in July 2008), there are six genes on chromosome X that cause CHDs: ATRX, BCOR, CFC1, GPC3, MID1 and ZIC3. All are causing syndromic CHDs. Females mostly remain unaffected by these mutations as the functional copy on their other X chromosome is sufficient to maintain normal cardiac development, or because they preferentially inactivate the X chromosome carrying the mutation, resulting in a normal expression level. However, exceptions have been reported, including mildly affected females carrying a MID1 mutation (So et al., 2005) and full ATRX syndrome in a girl carrying a ATRX mutation on an X chromosome that was preferentially activated (Badens et al., 2006).

Of interest, CHDs are present in approximately 40% of cases with monosomy X (Turner syndrome) (Sybert et al., 1998), and other vascular abnormalities are even more common. The most frequently observed cardiovascular malformations in Turner syndrome belong to the left outflow tract abnormality spectrum. These typically are elongation of the transverse arch (50%), coarctation of the aorta (12%), aortic valve abnormalities (such as stenosis, BAV and regurgitation) (16%) and also HLH (2-3%) (Sybert et al., 1998; Ho et al., 2004). Whether a single gene or multiple genes are implicated is currently unknown. In a normal 46,XX female, one X chromosome is inactivated. This activation occurs before cardiac development. Decreased expression from the X chromosome prior to activation is therefore unlikely to explain the CHDs. On the other hand, the pseudo-autosomal region as well as certain genes across chromosome X escape X-inactivation (Brown et al., 1997). Each of these could contribute to cardiogenesis and the concomitant high incidence of CHDs in Turner syndrome. CHDs seem less common in women with a 46X,i(Xq) (10%) (Sybert et al., 1998), suggesting that the major genes responsible for CHDs lie on the long arm of the X chromosome. However, there is insufficient data for a conclusive genotype-phenotype correlation.

Sex differences in incidence of specific types of CHD have been observed (table 1, Adapted from Pradat et al). When males are more frequently affected than females, this may point to a role X-linked genes. However, other CHD occur more frequently in females, and this cannot readily be explained by differences in the sex chromosomes. Also for many other multifactorial disorders (e.g. cleft lip and palate, pyloric stenosis, …) differences are observed in males versus females. This points to sex-dependent differences in embryonic development that are associated with different susceptibilities to various disturbances of normal embryonic development. The biological basis of these differences is not known but X or Y linked genes could play a role. Also hormonal differences could have an influence during later stages of development by acting on hormone-responsive genes involved in cardiac development.

[edit] the number of loci involved

The number of genes contributing to congenital heart defects in a single individual may vary, and in unrelated families different combinations of genes are most likely involved. To what extent a specific mutation contributes to a CHD observed in a patient is depending on the size of the effect disturbing the normal process of heart development. This effect can vary, and thus one can discriminate a spectrum of effects from monogenic causes of CHDs (where a single mutation has a dramatic effect, leading to a disturbed development) over oligogenic to polygenic (where a single mutation is not sufficient, and interacts with a few or many other factors). This vocabulary thus reflects the number of genetic factors that are presumed to play a role in the development of the CHD. In the spectrum from monogenic–oligogenic-polygenic, each has a decreasing relative contribution (figure 1).

As mentioned earlier, the recurrence risk for CHDs is not compatible with a major monogenic component in most instances. Moreover, in sporadic (i.e. non-familial) patients with CHDs, the frequency of de novo mutations in known genes is low (Posch et al., 2008). These observat suggests that fully penetrant monogenic causes of CHDs are rare. An estimate of the penetrance of isolated CHDs may be obtained from syndromic CHDs. There are not many syndromes that are associated with a CHD penetrance higher than 50%. (examples : del22q11; Down, Turner,…) Assuming that the genetic causes for syndromic CHDs are a good representation of penetrance of these for isolated CHDs, the number of causes for isolated CHDs with a high penetrance (>50%) is low.

As for any classification system, this distinction between monogenic, oligogenic and polygenic causes for CHDs is arbitrary, and one cannot unambiguously classify all genetic causes in a category. We will classify gene mutations with an estimated population frequency higher than 1% as polygenic, since these are considered polymorphisms. Mutations with a population frequency lower than 1% can either be monogenic or oligogenic. The term oligogenic is used when a single locus is insufficient to explain the presence of the CHD, but rather when the CHD is caused or modulated by a small number of loci across the genome (Badano and Katsanis, 2002). It is therefore also used to explain nonpenetrance in disorders that were thought to be monogenic. Examples of clear oligogenic inheritance patterns are sparse, and most loci that are associated with nonpenetrance are for the moment still classified under monogenic conditions. We anticipate however that low-penetrant loci will in some cases turn out to be oligogenic.

[edit] monogenic

Mutations that are associated with an CHD but that are not found in the normal population will be considered monogenic. Most known mutations that cause isolated CHDs have a high penetrance. This probably reflects a severe bias, since most genes identified so far were found via linkage studies in large families, necessarily having a high penetrance. In most cases, isolated and syndromic CHD genes are associated with a variable expressivity, meaning that different CHD types can be caused by a mutation of the same gene.

This can sometimes be explained by the type of mutation that is present. For example, loss-of-function mutations in TBX5 cause septal defects in 70% of patients, while gain-of-function mutations have been described to cause atrial fibrillation but rarely septal defects.

However, often the variability can only be explained by modifying factors, either genetic or environmental. A typical example is the 22q11 deletion syndrome, where exactly the same genetics lesion can be found in multiple individuals that are affected to a variable extend. Polymorphisms in the VEGF-gene have been shown to explain part of the variability in susceptibility to a CHD in the del22q11.2 deletion syndrome (Stalmans et al., 2003). Similarly, the increased frequency of AVSD amongst female and black Down syndrome patients suggests sex and race-dependent genetic differences. This again points to a more complex pathogenesis of the phenotype, where genes on other chromosomes likely also interact with genes on chromosome 21 (Freeman et al., 2008).

Despite this variability, it is of interest that specific types of CHD are often associated with distinct syndromes. For clinicians this may aid in the diagnosis of syndromic CHD, and for researchers this can indicate that specific genes or genetic pathways are involved in the development of distinct components of the heart. For instance, in Down syndrome, 39 % of cases have an AVSD, whereas Tetralogy of Fallot, which is considered to have a different pathogenesis is observed in 6 % of cases (Freeman et al., 2008). For the del22q11.2 syndrome, we have observed in a cohort of 147 patients 89 with a CHDs (60%). The majority of these were conotruncal malformations (50% Tetralogy of Fallot and 11% truncus arteriosus). Pulmonary stenosis usually makes up 9% of CHD types but was only found in one del22q11 patient.

[edit] oligogenic

An emerging cause for CHDs are the so-called rare inherited variants. These mutations are usually present below 1% in the normal population but are overrepresented in selected patients. They are typically inherited from a asymptomatic parent and are therefore not believed to be the only cause of CHDs in these patients. Other mutations need to be present to cause CHDs. The few genes that have been implicated in this inheritance mechanism in CHDs are all from the same Nodal pathway: FOXH1, GDF1, CFC1 and TDGF1. Mutations in these genes have been found in 5% of patients from a group of patients with outflow tract septation defects like transposition of the great arteries, truncus arteriosus, double outlet right ventricle and tetralogy of Fallot (Roessler et al., 2008). As one would expect, a single patient was found to carry multiple alleles associated with reduced Nodal pathway strength. Given the suspected low penetrance of each of these mutations, segregation cannot be used to indicate pathogenicity of these mutations. By using the zebrafish as a biosensor, Roessler et al. demonstrated however that part of these mutations have a detrimental influence on protein function. They moreover showed that these deleterious mutations are significantly overrepresented in a population of CHD patients compared to normal controls.

These mutations were found by a candidate gene approach: all these genes play a role in Nodal signaling, a pathway disturbed in some cases of CHDs and involved in left-right asymmetry establishment (Peeters and Devriendt). In order to also impact genetic counseling of CHD patients, more data should be gathered to establish the penetrance of these mutations. The contribution of rare inherited variants in other genes and pathways to the etiology of CHDs similarly remains to be investigated. Given the sporadic nature of many CHDs, it is anticipated that many more of these oligogenic causes will still be identified.

polygenic

A third group of causes is called complex, since multiple factors contribute to the development of the disorder. It is assumed that in many instances multiple genetic polymorphisms interact in disturbing normal cardiac development and/or function. In addition, to a lesser extent, the environment may influence this process. Virtually nothing is currently known about genes implicated in this multifactorial model of CHD. One of the few genes implicated is VEGF. Low expression haplotypes of this gene are overrepresented in Tetralogy of Fallot patients.

It is unclear whether accumulation of deleterious polymorphisms in a single individual is as such is sufficient for developing CHDs. It is has however been shown that certain polymorphisms can modulate the penetrance of CHDs in monogenic disorders. In theory, any of the following situations can be envisaged (Zlotoroga et al., 2003):

• a polymorphism of the allele present in cis could influence the expression of the mutant protein, resulting in a modulation of phenotypic expression. This mechanism has thus far not been investigated in CHDs, but could be used for the identification of causal genes in common genetic lesions such as trisomy 21.

• a polymorphism of the allele present in trans could influence the expression or function of the wild-type protein. This was checked for the remaining TBX1 allele in del22q11 patients by Rauch et al. (2004). However, there was no association between any common variant and the presence of a CHD in the del22q11 patients

• a polymorphism elsewhere in the genome could contribute to the susceptibility of the patient to CHDs. It was shown that low expression VEGF haplotypes are overrepresented in del22q11 patients with a CHD (Stalmans et al., 2003).

To identify other polygenic causes of CHDs, we anticipate that genome-wide association studies will make a major contribution to identify novel relevant players in human heart development. Moreover, given the overlap between polygenic, oligogenic and monogenic causes of CHDs, such a genome-wide approach to identify polygenic causes may allow for the identification of mutations in the same gene with a much higher penetrance and clinical significance.

As mentioned,  the frequency a single gene mutation is associated with CHDs (i.e. the penetrance) will vary according to its effect size. The relative contribution of other influences such as environmental and stochastic factors increases as the effect size of the genetic factors decreases. A purely genetic polygenic cause is therefore a lot less likely than a multifactorial cause. In these cases, genetic and environmental influences both impact cardiac development, either independently or cumulatively. An example of such a cumulative impact is the association between polymorphisms in the NNMT gene (encoding a protein that catalyzes nicotinamide), nicotinamide intake and CHDs (van Driel et al., 2008). Presence of the NNMT A-allele combined with maternal low nicotinamide intake at peri-conception confers a significantly increased risk for CHDs.

[edit] confirming genetic causes

A major challenge remains the confirmation of the role of a specific positional or functional candidate gene in the pathogenesis of CHD. The traditional approach of mutations analysis in patients faces many limitations :

• given the heterogeneity of CHD, large cohorts of patients need to be screened. Moreover, since the expression of a mutation in a given gene can be variable, other types of CHD need to be included in the mutation screen than that of the index patient.
• The genetic component in CHD can be highly diverse, ranging from monogenic over oligogenic and polygenic to multifactorial. Current evidence shows that de novo mutations are exceptional in non-syndromic CHD. Rather, an increasing number of studies reports on a higher number of rare (inherited) variants in individuals with CHD compared to normal controls. Therefore, in the future, we envision not only mutation screens in large cohorts of patients, but also simultaneous screening of a large number of genes in the same patients.
• Finally, evidence for the involvement of a gene in a multifactorial disorder is obtained from association studies. There is a lack of such studies in the field of CHD at the present time.
• mutation screens should not only include sequence analysis but also screening for CNV’s.
• functional studies, either in vitro or in vivo (animal models) will provide further evidence of a functional role of the found sequence alterations.