The genetic landscape of cardiovascular left–right patterning defects

Abstract

Heterotaxy is a disorder with complex congenital heart defects and diverse left–right (LR) patterning defects in other organ systems. Despite evidence suggesting a strong genetic component in heterotaxy, the majority of molecular causes remain unknown. Established genes often involve a ciliated, embryonic structure known as the left-right organizer (LRO). Herein, we focus on genetic discoveries in heterotaxy in the past two years. These include complex genetic architecture, novel mechanisms regulating cilia formation, and evidence for conservation of LR patterning between distant species. We feature new insights regarding established LR signaling pathways, bring attention to heterotaxy candidate genes in novel pathways, and provide an extensive overview of genes previously associated with laterality phenotypes in humans.

Introduction

Although humans are externally symmetric, our internal organs are asymmetric in their placement within the body. Left–right (LR) patterning is an embryologic event after the development of anterior–posterior and dorsal–ventral axes during which the LR axis is specified, followed by the reinforcement and maintenance of left- and right-sided identities. Failure to properly develop and maintain LR patterning results in a spectrum of abnormalities characterized by defects in laterality and abnormal placement of internal organs. Laterality defects, also known as situs abnormalities, include situs inversus totalis as well as situs ambiguus, the latter term used here interchangeably with heterotaxy syndrome. The heart, which exists as a midline linear tube early in cardiogenesis, undergoes several looping steps to reach its final four-chambered configuration and is particularly sensitive to LR signaling cues for normal development. Abnormal LR patterning therefore frequently results in complex congenital heart defects (CHDs). Investigation of the developmental genetic mechanisms that drive LR patterning in model systems (such as zebrafish, medaka fish, frogs, and mice) has identified conserved genetic pathways and specific genes that are important causes or candidates for human laterality disorders. It is timely to review the genetic pathways known to be important for LR patterning in model organisms and provide information on genes that have been documented to cause human heterotaxy and the specific phenotypes identified (Supplemental Tables 1-3).

The genetic architecture of heterotaxy is complex and investigation has been complicated by its rarity, its genetic and phenotypic heterogeneity, and poor phenotypic description in the literature. The high relative risk for first-degree relatives of an individual with heterotaxy highlights the importance of genetic factors [1]. Autosomal dominant, recessive, and X-linked inheritance patterns have been identified in heterotaxy cases, and single-gene mutations as well as aneuploidy or pathogenic copy number variants (CNVs) have been identified as causative [2].

However, current clinical genetic testing identifies an underlying genetic cause in less than 40% of cases, indicating that additional mechanisms and contributors remain to be identified. This review will discuss advancements, mainly within the last two years, in the understanding of the genetic causes of heterotaxy. We highlight single genes that have been identified as causative or associated with heterotaxy, many of which have been identified by work in model organisms. We review new pathways that have been recently identified in laterality disorders and emphasize more complex genetic architecture including new evidence of digenic or multigenic inheritance and post-transcriptional mechanisms that contribute to heterotaxy. For each gene associated with heterotaxy in humans, we provide a detailed summary of the reported phenotypes in order to address a current gap in the literature.

Phenotypic findings in left–right patterning disorders

Situs inversus totalis, an LR patterning disorder in which laterality is disrupted resulting in mirror image reversal, is classically associated with isolated dextrocardia without other structural CHDs. Here, we focus primarily on heterotaxy, in which patients have severe morbidity and mortality mainly due to diverse CHDs present in ~95% of patients [3]. Heterotaxy, characterized by abnormal arrangement of the thoracic and abdominal organs due to disturbed LR asymmetry, occurs in 0.8 per 10,000 livebirths [3]. In addition to morbidity and mortality related to CHD, abdominal organ involvement can also impact the quality of life. For example, intestinal malrotation (~36% of heterotaxy cases [3]) can be a risk for volvulus. Patients with heterotaxy with asplenia (~38% of heterotaxy cases [3]) have a significantly higher incidence of infections requiring hospitalizations compared with patients with heterotaxy with normal spleen function [4]. Biliary atresia, when present, is a significant cause of morbidity. Some genetic causes of laterality disorders are shared between heterotaxy and situs inversus totalis, often related to mutations in cilia-related genes.

Mendelian inheritance, copy number variants and monogenic causes of heterotaxy

The genetics of heterotaxy and laterality disorders are complex. Although not discussed in this review, heterotaxy is non-randomly associated with Trisomy 13 and 22q11.2 deletion syndrome as well as other well-characterized genetic syndromes (Supplemental Tables 2 and 3). Additionally, copy number variants (CNVs) have been identified associated with heterotaxy [2]. Autosomal dominant causes identified to date are frequently related to developmental signaling pathways functioning at the LRO during early embryonic development. Autosomal recessive causes typically relate to abnormalities in cilia structure or function, and pathogenic variants in the transcription factor ZIC3 are identified in approximately 75% of families with X-linked heterotaxy and in about 3% of males with heterotaxy without a family history [2,5]. Monogenic causes of heterotaxy can be associated with decreased penetrance and variable expressivity. While a monogenic cause is not identified in many individual cases, an understanding of developmental networks for known causative genes, as highlighted in more detail below, has led to the identification of new candidate genes and is important for beginning to elucidate more complex genetic models.

Major signaling pathways for light–right patterning

It is well known that LR asymmetry is established at the LRO, a pit-like structure (Fig. 1a) [6]. The LRO contains centrally located pit cells with motile cilia and peripherally located crown cells with non-motile, sensory cilia. The motile cilia are titled toward the posterior and rotate clockwise producing a leftward fluid flow (Fig. 1b). This causes asymmetric calcium transients and ultimately results in left-sided expression of NODAL in the LRO (Fig. 1c). NODAL expression then propagates to the left lateral plate mesoderm (l-LPM) (Fig. 1e) resulting in left-sided expression of the transcription factor PITX2 and subsequently normal organ asymmetry (Fig. 1f). For more detailed information regarding LR patterning, we recommend the following review [6]:

Fig. 1.

Schematic representation of the establishment of LR asymmetry in an early-stage mouse embryo and overview of genetic pathways of heterotaxy. (a) Overview of the LRO. The LRO contains centrally located pit cells (yellow) and peripherally located crown cells (blue). Leftward fluid flow (green arrow) is generated by the pit cells. (b) Oblique view of the LRO showing centrally located pit cells containing motile cilia (grey) and peripherally located crown cells containing non-motile, sensory cilia (brown). Note the motile cilia are tilted towards the posterior (P) and rotate in a clockwise orientation (purple circular arrow around each cilia) which produces a leftward fluid flow. Genes discussed in this review affecting motile and immotile cilia are listed. * denotes that SHROOM3 is involved in PCP signaling during cardiac development but its role in PCP signaling in the LRO specifically is unknown. NEK8 appears to play a role in calcium sensing. (c) Top view of the LRO showing the major components of the Nodal/TGF-β signaling pathway at the LRO. Leftward fluid flow results in asymmetric calcium transients detected by PKD2-expressing sensory cilia in the crown cells. Through a mechanism not completely understood, this leads to reduced levels of the NODAL antagonist DAND5 on the left (L) side of the LRO and subsequently an increase in left-sided NODAL expression. Genes mutated in patients with heterotaxy and involved with Nodal/TGF-β signaling at the LRO are listed. (d) Illustration of a single pit cell with a single LRO motile cilium. Representative genes known to cause PCD that are also associated with heterotaxy are listed. The black rectangle depicts the location of a cross-section through the cilium. (d’) Cross-section of LRO motile cilium. These motile cilia have a characteristic 9 + 0 microtubule arrangement composed of 9 outer microtubule doublets on their periphery. In contrast to other motile cilia, LRO motile cilia lack a central pair of microtubule doublets and also lack radial spokes emanating from this central pair (not shown). (e) Overview of the l-LPM. (f) Detailed view of Nodal/TGF-β signaling occurring in the l-LPM. NODAL expression propagates from the LRO to the l-LPM by forming a heterodimer with GDF1. NODAL causes asymmetric expression of LEFTY1, LEFTY2, and PITX2. LEFTY1 and LEFTY2 act to restrict the expression of NODAL to the l-LPM while PITX2 acts as a downstream transcription factor for the expression of left-sided genes. An autosomal recessive variant in GDF1 was recently reported in an individual with heterotaxy. Fig. 1a, e, and f are adapted from [6]. Recent reviews discussing LR patterning and PCP signaling respectively include the following: [6,11]. For detailed information regarding genes mentioned within this figure, see Supplemental Table 2. A; anterior. L; left. L-LPM; left-lateral plate mesoderm. mt-related; mitochondria related. P; posterior. PCD; primary ciliary dyskinesia. PCP; planar cell polarity. R; right. TGF-β; transforming growth factor beta.

Motile cilia function during left–right patterning

After researchers definitively demonstrated the importance of cilia for leftward fluid flow at the LRO [7], it was revealed that mutations in genes responsible for motile cilia structure and/or function (i.e. DNAH11, DNAAF1, ODAD1; Fig. 1d and Supplemental Table 1-3), are associated with Primary Ciliary Dyskinesia, a disorder of chronic sinusitis, bronchiectasis, and situs inversus. These cases are usually autosomal recessive. However, recent evidence suggests digenic or oligenic interactions of motile cilia genes can also cause laterality defects [8-10]. The mechanisms underlying phenotypic pleiotropy are not yet clear though modifier genes impacting highly penetrant mutations likely play a role.

Wnt signaling, planar cell polarity, and actomyosin-based mechanisms for left–right patterning

Proper leftward fluid flow at the LRO requires motile cilia positioned towards the posterior of each cell, a process dependent on planar cell polarity (PCP) signaling (Fig. 1b) [6]. PCP signaling, a result of non-canonical Wnt signaling, is an important effector of cell shape and convergent extension. Disruption causes LR patterning and neural tube defects in multiple model organisms [11].

New studies have expanded our understanding of the role of two previously documented heterotaxy genes, SHROOM3 and ZIC3, in PCP signaling (Fig. 1b) [5,12,13]. While the disruption of PCP signaling in Shroom3 gene-trap mice has previously been reported to cause fully penetrant neural tube defects, a recent study reveals these mice also have partially penetrant CHDs due to SHROOM3’s interaction with PCP component DVL2. In the neural tube, SHROOM3 is believed to link PCP to actomyosin constriction to effect morphogenetic movement [14-16]. ZIC3 mutations have also been shown to affect convergent extension morphogenetic movements in frogs and fish, and PCP dysregulation has now been shown in mice [17,18]. ZIC3 is involved in several LR signaling pathways such as Nodal, Hedgehog, and canonical Wnt [13]. Notably, cilia in the LRO are abnormally positioned in Zic3 null mice and when bred with mice carrying null alleles of several PCP components, more severe or more complex heart defects result in embryonic offspring [18].

Unconventional myosin 1D, encoded by MYO1D, functions downstream of PCP for cell polarity (Fig. 1b). An autosomal recessive missense variant in the MYO1D was recently identified in a patient with heterotaxy [19]. Intriguingly, this gene was first implicated to regulate LR patterning in Drosophila which lacks a ciliated LRO [20,21]. However, MYO1D has now been revealed to also regulate leftward fluid flow and LR patterning in higher organisms such as frogs and zebrafish [22-24]. One study found MYO1D requires its tail homology-1 domain for the formation of the LRO itself via vacuolar trafficking [22]. Notably, the new heterotaxy MYO1D missense variant was located near this domain and predicted to alter the stability of the protein [19]. Others determined MYO1D interacts with the PCP component VANGL2 [23,24] and can regulate the orientation of cilia in the LRO [24]. The actomyosin-based mechanisms for LR patterning help better define shared mechanisms for the development of chirality that occur in species without a ciliated LRO such as Drosophila. In addition, this discovery highlights an additional layer of complexity in LR patterning in vertebrates that hints at a buffering capacity in the developmental signaling pathways underlying LR patterning. It is tempting to speculate that these complexities may partially explain the variable expressivity and reduced penetrance often seen.

Post-transcriptional control of the NODAL antagonist Dand5

Mutations in Nodal/TGF-β signaling pathway genes are often autosomal dominant [25-27]. However, autosomal recessive mutations in the NODAL co-ligand GDF1 and the NODAL antagonist DAND5 have recently been reported (Fig. 1c and f) [26,28•]. New evidence suggests the proximal DAND5 3’-UTR as a necessary component for its left-sided decay (Fig. 1c). Upon leftward fluid flow, a calcium-dependent mechanism causes the RNA binding protein BICC1 to bind to the 3’-UTR of Dand5 [29••,30••]. This leads to the left-sided decay of Dand5 mRNA in mice [29••] and left-sided translational inhibition of dand5 mRNA in frogs [30••]. In both cases, DAND5 is no longer able to antagonize NODAL on the left side allowing for left-sided NODAL expression [29••,30••]. Therefore, it is reasonable to speculate that mutations affecting the post-transcriptional control of DAND5 may alter Nodal/TGF-β signaling and cause heterotaxy.

Cilia formation and the importance of nucleoporins and never in mitosis A-related kinases

Nucleoporins (NUPs) are components of the nuclear pore complex, which is a ring-like structure responsible for transport between the cytoplasm and nucleoplasm. In addition to localizing to the nuclear periphery, NUP93, NUP188, and NUP205 have been found to localize to the cilium base forming a unique barrel-like structure [31-33•]. A duplication of NUP188 [34] as well as biallelic mutations in NUP205 and NUP210 [31], have recently been reported in patients with situs inversus and heterotaxy (Fig. 1b). Morpholino-induced knockdown on NUP93, NUP188, and NUP205 in frogs and zebrafish caused LR patterning defects [31-34] as well as a reduction of cilia in their LROs [32,33•]. Intriguingly, the role of these NUPs in cilia formation is specific to their function at the cilia and not the nucleus [33•].

Recent evidence also shows that these NUPs and several never in mitosis A-related kinases (NEKs) are potentially reciprocally regulated [31,35]. In addition, mutations in NEKs result in isolated CHD, situs inversus totalis, and heterotaxy (Fig. 1b) [34-36•]. Model systems also suggest NEKs are involved in LR patterning [34,37,38] and regulate cilia formation or calcium sensing via PKD2 (Supplemental Table 2) [35,37,38].

Mitochondria, cilia, and the development of heterotaxy

Exome sequencing analysis of 285 patients with heterotaxy and patients with heterotaxy-spectrum CHD identified 26 variants in mitochondrial-associated genes [39••]. Patients with heterotaxy had depleted mitochondrial DNA when compared with healthy controls and patients with isolated CHD without cilia dysfunction phenotypes. In addition, enhanced mitochondrial activity generated shorter cilia while impaired mitochondrial activity produced longer cilia. Both alterations disturbed fluid flow in the zebrafish LRO producing a heterotaxy phenotype. In particular, morpholino-induced knockdown of two mitochondrial genes (MTRR and TAZ) in zebrafish resulted in a heterotaxy phenotype and longer cilia in the LRO (Fig. 1b). RNA rescue experiments using wildtype human mRNA rescued the heterotaxy phenotype, while injecting patient variant mRNA did not. Mutations in mitochondria-associated genes disrupt cilia formation and therefore may act as susceptibility factors for non-monogenic, complex inheritance cases of CHD and heterotaxy [39••]. The identification of the relationship between energy production and cilia during development is novel and could be linked to the association of pregestational diabetes with heterotaxy (adjusted odds ratio 12.3) [40]. Further investigation will be required to elucidate potential gene-environment interactions for specific mitochondrial variants.

Ciliopathies and the development of left–right asymmetry

Genes classically associated with ciliopathies, disorders affecting both motile and immotile cilia, can also lead to LR asymmetry (i.e. INVS and NPHP3) (Fig. 1b, Supplemental Table 1-3). Biallelic variants in TTC21B, a critical component of the retrograde ciliary transport system, have been reported in several ciliopathy syndromes (Supplemental Table 2) including one case with situs inversus [41]. Recently, biallelic variants in TTC21B were first associated with heterotaxy in a family affected by multiple ciliopathy-spectrum disease (Fig. 1b) [42]. In addition, a copy number gain involving TTC21B has also been reported, suggesting this may be disease-causative [43]. Identifying additional evidence by accumulating more cases, a better understanding of segregation with disease, and information on the spectrum of phenotypes resulting from pathogenic variation in TTC21B is important. Nevertheless, these recent studies highlight the genetic complexity of disorders of LR development indicating that heterozygous and compound heterozygous variants as well as gene dosage alterations resulting from CNVs may lead to divergent laterality phenotypes.

Mutations in the ciliary transition zone gene TMEM107 generate fewer and abnormal cilia in patients with Meckel-Gruber, Orofaciodigital, and Joubert syndromes (Supplemental Table 2). TMEM107 is a transmembrane protein important for regulating protein content of primary cilia. LR patterning defects are absent in these patients and in Tmem107 hypomorphic mice (Supplemental Table 2). However, Tmem107 null mice are embryonic lethal and have left pulmonary isomerism [44•,45]. The phenotypic heterogeneity in these two mouse models is likely due to the location of ciliary defects: null mice have less cilia in the LRO, lateral plate mesoderm (LPM), and midline whereas hypomorphic mice have less cilia in the LPM only [44•]. This interesting finding indicates cilia formation exhibits tissue-specific differences in sensitivity to gene dysregulation and provides additional insight into a potential mechanism underlying phenotypic variability.

Novel heterotaxy candidate genes with unknown mechanisms

Two novel heterotaxy candidate genes have recently been reported: CC2D1A and CIROP. Zebrafish experiments found CC2D1A is involved in ciliogenesis and LR patterning while several model systems determined CIROP acts on the left side of the LRO downstream of leftward fluid flow and upstream of DAND5. However, both genes’ precise role in LR patterning is unknown [46,47•].

Summary and future directions

The genetic architecture of heterotaxy is complex. In addition to the wide variety of developmental pathways important for LR patterning discussed here, there is emerging evidence for digenic or multigenic inheritance. Investigation of individual rare variant enrichment in known and candidate genes my further elucidate important genetic associations. Investigation of multigenic and gene-environment interactions remains an important area for future research. In addition, the field has largely focused on exome sequencing to identify the genetic causes of heterotaxy. As technology has improved and whole-genome sequencing has become more prevalent, we are also optimistic that assessing non-coding regions for potential post-transcriptional modifications, alternative splicing, and gene-regulatory mechanisms may be fruitful to uncovering additional causes of heterotaxy.

The complex and heterogenous laterality phenotypic abnormalities seen in heterotaxy are such that almost no two patients have identical combinations of anatomic abnormalities. Several suggestive mechanisms highlighted in this review may underlie this phenotypic pleiotropy: the impact of modifier genes on highly penetrant mutations, digenic or multigenic inheritance, tissue and developmental-stage specific sensitivity of cilia or signaling pathways to heterotaxy gene network dysregulation, and buffering capacity and shared pathways for the development of LR patterning. Comprehensive review of LR patterning abnormalities in humans and model organisms (Supplemental Table 2) demonstrates that the field would benefit from significant improvement in phenotypic characterization. Combining deep phenotyping with robust genomic evaluation will continue to improve our knowledge of the genetic architecture of LR patterning defects and their spectrum of congenital anomalies.