|Home | About | Journals | Submit | Contact Us | Français|
Congenital heart malformations are a major cause of morbidity and mortality especially in young children. Failure to establish normal left-right (L-R) asymmetry often results in cardiovascular malformations and other laterality defects of visceral organs.
To identify genetic mutations causing cardiac laterality defects.
We performed a genome-wide linkage analysis in patients with cardiac laterality defects from a consanguineous family. The patients had combinations of defects that included dextrocardia, transposition of great arteries, double outlet right ventricle, atrio-ventricular septal defects and caval vein abnormalities. Sequencing of positional candidate genes identified mutations in NPHP4. We performed mutation analysis of NPHP4 in 146 unrelated patients with similar cardiac laterality defects. Forty-one percent of these patients also had laterality defects of the abdominal organs. We identified eight additional missense variants that were absent or very rare in controls. To study the role of nphp4 in establishing L-R asymmetry, we used antisense morpholinos to knockdown nphp4 expression in zebrafish. Depletion of nphp4 disrupted L-R patterning as well as cardiac and gut laterality. Cardiac laterality defects were partially rescued by human NPHP4 mRNA, whereas mutant NPHP4 containing genetic variants found in patients failed to rescue. We show that nphp4 is involved in the formation of motile cilia in Kupffer’s vesicle (KV), which generate asymmetric fluid flow necessary for normal L-R asymmetry.
NPHP4 mutations are associated with cardiac laterality defects and heterotaxy. In zebrafish, nphp4 is essential for the development and function of KV cilia and is required for global L-R patterning.
Laterality defects refer to a broad group of disorders caused by the disruption of normal left-right (L-R) asymmetry of the thoracic or abdominal visceral organs 1. Situs inversus totalis is the mirror image reversal of all visceral organs, whereas heterotaxy is the abnormal orientation of one or more organs along the L-R axis 2. In heterotaxy, congenital heart malformations result in major morbidity and mortality 3. Although heterotaxy most often occurs as a sporadic condition, familial clustering has been documented with pedigrees suggesting autosomal recessive, autosomal dominant and X-linked inheritance 4–7.
L-R patterning of vertebrate embryos occurs prior to organ formation and is conducted by a conserved signalling cascade that includes asymmetric expression of the NODAL, LEFTY, and PITX2 genes in left lateral plate mesoderm (LPM) 8. Motile cilia are involved in establishing this L-R asymmetric signaling. Laterality defects have been linked to ciliary motility by the observation that 48% of individuals with primary cilia dyskinesia also had situs inversus totalis and 6% had heterotaxy 9.
Animal models have assisted our understanding of L-R patterning and the role of cilia. The inversus viscerum (iv) mouse has a mutation in the ciliary left-right dynein (lrd) gene and often develops laterality defects 10. Lrd was found to be required for normal motility of monocilia on an embryonic structure called the node. These node cilia generate a leftward fluid flow that is necessary for normal asymmetric Nodal-Lefty-Pitx2 signaling 11. In zebrafish, Kupffer’s vesicle is a ciliated organ analogous to the mouse node that is essential for normal L-R patterning 12. Asymmetric fluid flow generated by the monocilia may move signaling factors 11, 13 and/or bend mechanosensory cilia 14 to initiate asymmetric signaling.
Dysfunction of ciliary proteins gives rise to a wide range of human disorders known as ciliopathies. They can lead to a variety of defects including craniofacial, skeletal, respiratory, reproductive, renal, visual, olfactory and auditory abnormalities 15–17. The nephronophthisis (NPHP) and associated ciliopathies - Senior-Loken syndrome, Joubert syndrome, Meckel-Gruber syndrome - are characterized by cilia-related defects, including cystic kidney disease, retinal degeneration, liver fibrosis and brain malformations 18–19. Mutations in 18 genes are known to cause nephronophthisis and associated ciliopathies20–21. Interestingly, mutations in NPHP2/INVS and NPHP3 can also lead to heterotaxy, situs inversus and isolated congenital heart malformations 22–24.
Protein network analysis has shown that several of these proteins form an interaction network organized in at least three connected modules: NPHP1-4-8, NPHP5-6 and MKS 25. Ciliary localization analysis of eight nephrocystins (NPHP1-6, 9 and 10) indicates that they are present in the primary cilia, the basal body and/or the centrioles and suggest that they participate in ciliary assembly and trafficking 25–28.
In this study, a genome-wide linkage analysis identified nephronophthisis-4 (NPHP4) variants in patients with cardiac laterality defects. Functional studies indicated that loss of zebrafish nphp4 resulted in cardiac laterality defects. In addition, nphp4 depletion disrupted asymmetric Nodal expression in the LPM, indicating nphp4 is required for global L-R patterning of the embryo. Analysis of cilia in Kupffer’s vesicle revealed that loss of nphp4 reduced cilia length and disrupted asymmetric fluid flow. Our results establish the importance of nphp4 in cilia development and function. Furthermore, our findings suggest that malfunction of NPHP4 contributes to a wide range of congenital heart malformations and more complex defects within the heterotaxy spectrum.
An expanded Methods section is available in the Online Data Supplement.
We identified a consanguineous Iranian family including five patients with congenital heart malformations. Three patients (IV-1, IV-8 and IV-12; Figure 1a) were born with similar cardiac laterality defects (Table 1). Patient IV-1 had dextrocardia, atrial situs solitus, complete atrioventricular septal defect and discordant ventriculo-arterial connection with dextro-transposition of the great arteries (d-TGA). In addition, he had an interrupted inferior caval vein and a severe pulmonary valve stenosis (PS). He had no surgical correction and died suddenly at the age of 22 years. No autopsy was performed. Patient IV-8 had dextrocardia, dextrorotation and atrial situs solitus. She had an azygos continuation of the right infrahepatic part of the inferior caval vein draining into the right superior caval vein and the suprahepatic part of the inferior caval vein draining into the right atrium. She had a cor triatriatum with the right pulmonary veins draining into the right part of the left atrium and the left pulmonary veins into the left part of the left atrium. A persistent left inferior and superior caval vein also drained into the left part of the left atrium. She had a secundum atrial septal defect (ASD) and perimembranous ventricular septal defect (VSD). She had also left bronchial isomerism. Patient IV-12 had atrial situs solitus, atrio-ventricular concordance and ventriculo-arterial discordance namely, double outlet right ventricle and d-TGA. He also had a subpulmonary VSD, patent foramen ovale and patent ductus arteriosus (PDA).
In addition, patient IV-7 died shortly after birth due to an unspecified congenital heart malformation (Figure 1a). The fifth patient (IV-16) had mild congenital heart malformations consisting of a small VSD, PS and PDA, which was ligated at one year of age (Table 1).
Physical examination revealed no dysmorphisms and all patients had normal psychomotor development. CT/MRI or ultrasound of the abdomen revealed no kidney cysts and all individuals have reached adulthood at the time of their last evaluations. No abdominal laterality defects such as asplenia or polysplenia, malrotation of the gut or midline liver were detected in any of these cases. None of the patients had signs of abnormal mucociliary clearance. No visual problems or night blindness were reported.
The genome-wide linkage analysis was performed using Affymetrix SNP arrays. Two unaffected parents, three patients and one healthy sibling (Figure 1a) were included in the analysis. Multipoint linkage analysis revealed five regions on chromosomes 1, 2, 3, 9 and 11 with LOD scores above 2.5 (Figure 1b). The maxLOD scores (2.7) were located on chromosome 1p36 and 11p15. Subsequently, microsatellite markers mapping to all candidate regions were tested. The loci on chromosomes 2, 3, and 9 were excluded, based on heterozygosity observed in the patients (data not shown).
On the chromosome 1p36 locus, all three patients with cardiac laterality defects (IV-1, IV-8 and IV-12) shared a homozygous region covered by 59 SNPs from rs4845835 to rs1203695. Haplotype analysis showed recombinations that delimited the borders of the region from rs2722782 (5.26 Mb) to rs1203696 (14.21 Mb) (Online Figure Ia). Thus, the candidate region spanned 9 Mb and contained 152 genes (NCBI build 37.1).
These patients also showed homozygous genotypes for 49 consecutive SNPs on chromosome 11, from rs16905816 to rs10500752. Further fine mapping in the 11p area delineated the borders of the linkage region between markers D11S4188 (telomeric) and rs2896598 (centromeric) (Online Figure Ib). The chromosome 11 locus extended only 3 Mb (9.1–12.1 Mb), containing 34 genes (NCBI build 37.1).
Haplotypes from both loci were examined in all available family members. A normal person (IV-14, with normal MRI of the thorax/abdomen) had homozygous haplotypes on the chromosome 1 locus (Online Figure Ia). In addition, individuals IV-3 and IV-4 were homozygotes for the chromosome 11 locus (Online Figure Ib); both persons were reported as unaffected, but medical examinations could not be performed. Patient IV-16, exhibiting a mild cardiac phenotype and no laterality defects, carried heterozygous haplotypes at both loci (data not shown). Only the three patients with laterality heart defects had homozygous haplotypes on both loci. Since these were the only genomic regions where the three patients showed extended homozygosity, we further investigated these loci.
A total of 109 genes on the chromosome 1p36 locus had a known reference sequence (NCBI build 37.1). Selection of genes for sequence analysis was based on available expression and/or functional information. The data was analyzed through the use of Ingenuity pathway analysis (Ingenuity® Systems). Thirty-six candidate genes were selected from the chromosome 1 locus. Direct sequencing of their coding regions identified two novel homozygous missense variants in the NPHP4 gene present in three patients from the index family: c.3131G>A (p.Arg1044His) and c.3706G>A (p.Val1236Met) (Online Table I). These non-synonymous variants are extremely rare in the Iranian (Kurdish) population (allele frequency of 0.2% and 0.1% in 1232 control chromosomes). Moreover, the variants were absent in 270 Caucasian/Dutch and 178 Hispanic control chromosomes.
From the 34 genes mapping to the chromosome 11 locus, 19 genes had a well annotated reference sequence. Sequence analysis of their coding regions and exon-intron boundaries revealed only one novel DNA missense variant (Online Table I). In the AMPD3 gene (adenosine monophosphate deaminase 3), the homozygous c.2240 G>A (p.Arg747Gln) variant was found. This variant was not present in 626 control chromosomes. Mutations in the AMPD3 gene lead to (asymptomatic) deficiency of erythrocyte AMP deaminase 29 (OMIM 612874).
We sequenced all 30 exons of NPHP4 in three cohorts of patients with cardiac laterality defects - with or without other situs abnormalities. Patient samples were collected at the Erasmus Medical Center, Rotterdam, the Department of Clinical Genetics, Leuven and from the Baylor College of Medicine, Houston. All 146 patients had a variety of cardiac laterality defects. Transposition of the great arteries was the most frequently found (49% of the patients). In addition, complete atrio-ventricular septal defect, double outlet right ventricle and abnormal pulmonary venous return were often reported. Dextrocardia was present in 33% of patients. Moreover, 41% had documented laterality defects of the abdominal organs, including abdominal situs inversus, asplenia or polysplenia, midline liver and intestinal malrotation.
Nine missense variants were found in 10 patients (Figure 1c, d and Table 1). The population frequency of each allele was tested by sequencing ethnically matched controls. A variant was considered as likely non-pathogenic if the allele frequency in healthy individuals was higher than 1%. Thus, p.Pro1160Leu with a frequency of 2.1% in control chromosomes was excluded from further analysis. In addition, we investigated the frequency of these variants in available databases (dbSNP135, 1000Genomes, NHLBI exome project). All variants were very rare or absent in controls (allele frequency ≤0.8%, Table 1).
These rare NPHP4 variants were significantly more frequent in heterotaxy cases (6%, 9 of 146 cases) than in controls (1.2%, 3 of 250, Fisher’s exact test p=0.006). In silico evaluation was performed using five prediction computer programs. This assessment predicted the impact of amino acid substitutions on the structure and function of human proteins. The variants were classified as probably pathogenic if at least 3 programs considered them as damaging (Table 1). Seven variants satisfied this criterion. Interestingly, p.Phe91Leu, p.Arg961His and p.Arg1192Trp have been reported in patients with Senior-Loken syndrome type 4 or nephronophthisis type 4 30.
A zebrafish nphp4 ortholog was taken from the Ensembl database (Online Figure II). To determine the pattern of nphp4 expression during embryogenesis, we performed reverse transcription PCR (RT-PCR) and RNA in situ hybridization experiments at several developmental stages. Consistent with a recent report 31 we found nphp4 expression was maternally supplied and ubiquitously expressed during the first 24 hours of zebrafish development (Online Figure IIIb-g). RT-PCR detected nphp4 expression at all stages tested between 4-cell stage and 100 hours post-fertilization (hpf) (Online Figure IIIa). This early and ubiquitous expression pattern suggested a role for nphp4 during early development.
To assess the function of nphp4 during embryonic development, we utilized antisense morpholino oligonucleotides (MO) to knockdown expression of zebrafish Nphp4 protein. Embryos injected with a MO designed to block nphp4 mRNA translation (nphp4 TB-MO) developed dose-dependent morphological abnormalities reminiscent of embryos with cilia defects 32–34, including a curved body axis (Online Figure IVd) and otolith formation defects at 2 days post-fertilization (dpf) (Online Figure Va, b).
In addition, RNA in situ hybridization staining of the heart-specific marker cmlc2 revealed heart laterality defects. Uninjected controls and embryos injected with a standard control MO showed normal rightward looping of the heart at 2 dpf (Figure 2a, b). However, heart looping in nphp4 TB-MO injected embryos was significantly altered, as the heart often looped in the reverse orientation or failed to loop (Figure 2a, b).
To test whether heart laterality phenotypes were specific to knockdown of nphp4, we designed two additional MOs to interfere with nphp4 mRNA splicing at exon 4 (nphp4 SB-MO1) or exon 9 (nphp4 SB-MO2) (Online Figure IVa, b). Quantitative real time PCR (qPCR) analysis indicated nphp4 SB-MO1 reduced nphp4 mRNA levels by 90% (Online Figure IVc) and caused heart laterality defects without inducing body axis defects (Figure 2a, b and Online Figure IVd). This indicates heart L-R phenotypes are separable from axial defects. nphp4 SB-MO2 reduced the amount of normally spliced nphp4 mRNA by 50% (Online Figure IVc) and resulted in curved body axis and heart looping defects (Online Figure IVd and Figure 2a, b, respectively), similar to nphp4 TB-MO injected embryos. Injecting a lower dose of nphp4 SB-MO2 (0.4 ng) also altered heart looping, but with reduced penetrance (Figure 2b), suggesting partial loss of nphp4 can cause cardiac laterality defects.
Other abnormalities such as hydrocephalus or gross eye defects were not observed. At 5 dpf, pronephric cysts were observed with a low penetrance in embryos injected with TB-MO (11%) or SB-MO2 (8%) (Online Figure Vc, d). No pronephric cysts were observed in SB-MO1 injected embryos. Our results using three independent MOs suggested a role for nphp4 that is required for normal heart laterality in zebrafish.
To further confirm that defects observed in zebrafish embryos were specifically due to nphp4 depletion, we conducted rescue experiments using human wild-type (wt) NPHP4 mRNA. Co-injecting nphp4 TB-MO with wt NPHP4 mRNA resulted in a partial, but significant, rescue of heart looping defects (% of normal embryos improved from 43% to 60%, p=0.03; Figure 2c). Next, we co-injected nphp4 TB-MO with human NPHP4 mRNA containing either the c.3131G>A (p.Arg1044His) or c.3706G>A (p.Val1236Met) missense variant identified in the index family. In contrast to wt NPHP4, these NPHP4 variants failed to rescue heart looping defects (Figure 2c). These results suggest that these variants are pathogenic and are involved in human laterality defects.
To determine whether nphp4 plays a role in heart laterality specifically or is involved in establishing global L-R patterning of the embryo, we analyzed additional markers of L-R asymmetry. RNA in situ hybridization, using foxa3 probes to label the embryonic gut, showed that nphp4 knockdown significantly altered laterality of the liver and pancreas in nphp4 MO injected embryos (Figure 3a, b). We next analyzed expression of the Nodal-related gene southpaw (spaw), the earliest asymmetrically expressed gene in lateral plate mesoderm (LPM) in zebrafish 35. Control embryos exhibited normal left-sided spaw expression (Figure 3c, d). In contrast, nphp4 MO injected embryos showed a significant disruption of spaw expression, which was often reversed, bilateral or absent (Figure 3c, d). Altered asymmetric gene expression can result from defects in the embryonic midline 36. However, analysis of the midline markers no tail and sonic hedgehog revealed that midline structures were intact in nphp4 MO injected embryos (Online Figure VI). These results indicate nphp4 functions independent of midline development to control spaw expression and global L-R patterning of the embryo.
In zebrafish, Kupffer’s vesicle (KV) is a transient organ that generates cilia-driven asymmetric fluid flow necessary to bias spaw expression to the left LPM. Examination of live embryos at the 8 somite stage showed that the KV organ appeared normal in control MO (Figure 4a) and nphp4 MO injected embryos (Figure 4b, c). However, analysis of cilia in KV by fluorescent immunostaining with acetylated Tubulin antibodies revealed that the cilia were significantly shorter in nphp4 MO injected embryos (Figure 4e-g) as compared to controls (Figure 4d, g). We did not observe a significant difference of KV cilia number between control and nphp4 MO injected embryos (Figure 4h). To analyze KV cilia function, we injected fluorescent beads into KV of live embryos and used video microscopy to record fluid flow 12. Most control embryos showed strong counter-clockwise asymmetric fluid flow (Figure 4i, l; Movie S1). In contrast, flow was often absent (Figure 4j, l; Movie S2) or reduced (Figure 4k, l; Movie S3) in nphp4 MO injected embryos. Consistent with dose-dependent effects of nphp4 SB-MO2 on heart looping (Figure 2b), we observed more severe flow defects in embryos injected with a higher nphp4 SB-MO2 dose (Figure 4l). Together, these results show that nphp4 knockdown results in short KV cilia and compromises asymmetric fluid flow that is necessary for normal L-R patterning.
We found homozygous missense NPHP4 variants in a consanguineous family containing three patients with cardiac laterality defects, bronchial isomerism and normal abdominal situs. Interestingly, though NPHP4 is a cilia related gene that is mutated in patients with autosomal recessive juvenile nephronophthisis (NPHP type 4, OMIM 606966) 37 and Senior-Loken syndrome (SLSN4, OMIM 606996) 38, our patients did not show signs of nephronophthisis or retinitis pigmentosa, which are distinctive features of these diseases.
Because of the known interaction between NPHP1, NPHP2/INVS, NPHP3 and NPHP4 proteins 23–24, 37, it is obvious that mutations in one or more of these genes disrupt the same pathway(s) and can lead to similar phenotypes (i.e. nephronophthisis). Conversely, mutations within the same gene can lead to various phenotypic outcomes in different patients. Mutations in NPHP2 result in nephronophthisis with or without situs inversus and mild cardiac defects 23 whereas NPHP3 mutations lead to isolated nephronophthisis or retinal degeneration 39. Alternately, NPHP3 mutations can cause a broad clinical spectrum of early embryonic patterning defects comprising of situs inversus, congenital heart defects, central nervous system malformations and renal-hepatic-pancreatic dysplasia 24. The NPHP6 gene (CEP290) is another good example. The phenotypic spectrum of the mutations ranges from isolated blindness, SLSN, nephronophthisis, Joubert syndrome, Bardet-Biedl syndrome, to the lethal Meckel-Grüber syndrome 40.
We investigated the presence of NPHP4 variants in 146 sporadic patients having cardiac laterality defects, with or without involvement of other thoracic or abdominal organs. In 6% of the patients, we identified heterozygous missense variants compared to 1.2% of the ethnically matched controls, indicating mutation excess in the patients (p<0.006). No compound heterozygous or homozygous variants were detected in these sporadic cases. Similarly, single heterozygous NPHP4 variants were found in the majority of patients with autosomal recessive nephronophthisis type 4 30. A second mutation might be located in an area not covered by exon sequencing or in another (cilia-related) gene. The latest, a complex genetic model with combined effects of multiple genes seems the most plausible explanation. In fact, di- or oligogenic inheritance have been demonstrated in several ciliopathies, including the nephronophthisis 21, 41, Joubert syndrome 42 and Bardet Biedl syndrome 43–44.
The findings in our study are entirely consistent with a complex, oligogenic disease model. The rare heterozygous variants identified in the sporadic cases have probably an epistatic effect with additional genetic modifiers. Even in the index consanguineous family, we cannot exclude the existence of other genetic variants that explain the complex cardiovascular malformations and heterotaxy and the lack of renal/visual disease.
In congenital heart malformations and heterotaxy, the NODAL signaling pathway is a paradigm for oligogenic inheritance. Some patients with heterotaxy and/or conotruncal defects such as double outlet right ventricle (DORV) and transposition of great arteries (TGA), show several mutations in genes belonging to the NODAL signaling pathway 45–46. As functional significance of mutations in these genes were demonstrated, the cumulative effects of multiple mutations may lead to reduced NODAL signaling eventually resulting in congenital heart malformations. In addition, a combinatorial role between the NODAL signaling pathway and ZIC3 gene has been demonstrated in familial TGA patients 47. These studies support the notion that genetic variants or susceptibility alleles within one or more developmental pathways may dysregulate signaling in a synergistic fashion and cause congenital heart malformations or heterotaxy.
Studies in humans, zebrafish and mice indicate that NPHP2 and NPHP3 play a role in L-R axis determination 22–24, 39. To investigate the role of NPHP4 in establishing L-R asymmetry, we used antisense MOs to knockdown expression of zebrafish nphp4. Depletion of nphp4 in zebrafish resulted in abnormal heart and gut orientation, closely resembling the (cardiac) laterality defects observed in the patients. Co-injection of nphp4 TB-MO and human wt NPHP4 mRNA significantly ameliorated the phenotypic spectrum due to nphp4 depletion. In contrast, co-injection of nphp4 TB-MO and human NPHP4 mRNA containing genetic variants found in patients failed to rescue the laterality defects suggesting that these variants are pathogenic. Furthermore, analysis of asymmetric gene expression revealed that nphp4 knockdown alters asymmetric Nodal expression in the LPM without affecting expression of midline markers.
Our analyses in zebrafish have confirmed that knockdown of nphp4 results in shortened motile cilia 48. For first time we show that nphp4 depletion leads to disruption of cilia-driven fluid flow within KV which most likely cause laterality defects. Similarly, nphp3 knockdown in zebrafish leads to situs inversus and heterotaxy due to defective (fewer and shorter) KV cilia 49.
In conclusion, we identified NPHP4 mutations in patients with cardiac laterality defects and other malformations within the heterotaxy spectrum. In zebrafish, our results demonstrate that nphp4 is required for global L-R patterning of the embryo via regulation of Nodal signaling and plays a role that is essential for the development and function of KV cilia.
The linking of NPHP4 to L-R axis determination and laterality defects will help dissect the complex genetic composition of heterotaxy and related cardiovascular malformations.
We are grateful to the family and patients and who participated in the study.
We thank Fiona Foley, Chunlei Gao and Herma van der Linde for excellent technical assistance and Tom de Vries Lentsch for the artwork. We acknowledge Prof. Peter van der Spek for the use of Ingenuity Systems.
Sources of funding
This work was partially funded by the Dutch Heart Foundation, the Netherlands (2006T006) to I.M.B.H. L, an Erasmus MC grant and an Erasmus Fellowship (Erasmus Medical Center, The Netherlands) to A.M. B-A, a grant from the Center for Biomedical Genetics (CBG), the Netherlands to B.A.O., and a grant from the National Heart Lung and Blood Institute, USA, (R01HL095690) to J.D.A.