Novel Mutations in VANGL1 in Neural Tube Defects

INTRODUCTION

Neural tube defects (NTDs) are among the most common congenital malformations (1-2/1000 births) that involve genetic and environmental factors. They arise from partial or complete failure of the neural tube to close during embryogenesis (Kibar et al., 2007a). The most common forms of NTDs are open NTDs, including anencephaly and myelomeningocele (open spina bifida), where the affected nervous tissues are exposed to the body surface. Another rare form of open NTDs is craniorachischisis which results from failure of neural tube closure over the entire body axis. Among these conditions, only open spina bifida is compatible with life but with severe life-long disabilities. Closed NTDs are less severe conditions and include lipomyeloschisis, meningocele, caudal regression syndrome and split cord malformations (Rossi et al., 2006). The most relevant epidemiological finding in NTDs is the protective effect of periconceptional folic acid supplementation which reduces their incidence by 50-70% (MRC, 1991). Polymorphic variants in genes of the folate and homocysteine pathways have been associated with increased risk of NTDs, including a common variant (c.677C>T) in the MTHFR gene (5,10-methylene-tetrahydrofolate reductase) (Kibar et al., 2007a).

Mouse models offer a powerful tool to decipher the etiological complexity of NTDs (Harris and Juriloff, 2007). Particularly, the Loop-tail (Lp) mutant that develops craniorachischisis has been shown to carry missense mutations in the Vangl2 gene, which is the mammalian homolog of the Drosophila Strabismus/ Van gogh (Stbm/Vang) gene required for establishing planar cell polarity (PCP) in the developing eye, wing and leg tissues (Kibar et al., 2001; Murdoch et al., 2001). Studies of this mutant have provided the first line of evidence for involvement of the PCP pathway in NTDs in mammalians. PCP is controlled by a non-canonical Frizzled/Dishevelled (Fz/Dvl) signaling pathway that involves a number of additional core genes including Stbm/Vang, Flamingo (Fmi), Prickle (Pk), and Diego (Dgo) (Simons and Mlodzik, 2008). Members of the PCP pathway are highly conserved in vertebrates where they have been implicated in controlling the morphogenetic process of convergent extension (CE) during gastrulation and neurulation (Simons and Mlodzik, 2008). In mouse, combined mutations at Dvl1/Dvl2, Fz3/Fz6, and mutations at Celsr1 (a mouse Fmi homologue) and PTK7, like the Lp mutation, cause craniorachischisis (Kibar et al., 2007a).

Vangl1, a vertebrate homolog of Vangl2, encodes a membrane protein whose predicted features include 4 transmembrane (TM) domains and a PDZ-domain binding motif at the carboxy terminus involved in protein-protein interaction (Torban et al., 2004). Vangl1 shows a dynamic pattern of expression in the developing neural tube and genetically interacts with Vangl2 (Torban et al., 2008). In zebrafish, injection of Vangl1 mRNA partially suppresses the CE defect in tri/Vangl2 (tri, orthologue of Vangl2) mutant embryos, suggesting that Vangl1 and Vangl2 have similar biochemical activities (Jessen and Solnica-Krezel, 2004). Recently, we have identified in VANGL1 (MIM# 610132) three missense mutations in familial (p.Val239Ile, p.Arg274Gln) and sporadic (p.Met328Thr) cases of NTDs, including a de novo mutation (p.Val239Ile) appearing in a familial setting (Kibar et al., 2007b). These mutations affect highly conserved residues in the Vangl family and were not detected in 171 control individuals. Importantly, p.Val239Ile abolished interaction of VANGL1 with DVL proteins strongly suggesting a pathogenic effect on the protein function in the PCP pathway (Kibar et al., 2007b). In this current study, to better understand the role of VANGL1 in the development of NTDs, we analyzed this gene by direct sequencing in a large and well-characterized cohort of 673 patients affected with various forms of NTDs including cranial, open and closed spinal dysraphisms and of various ethnic origins.

PATIENTS AND METHODS

The cohort consisted of 284 Italian patients recruited at the Spina Bifida Center of the Gaslini Hospital in Genova, Italy and 389 American patients recruited at the Children's Memorial Hospital in Chicago, Illinois, United States. The Italian patients had an age range of 0-27 years and were recruited either at birth or in follow-up visits during the period of January 2006 -July 2007. The American patients were born from 1970-2007 and were recruited during the period of 1996-2007.The folate status is known only for mothers of Italian patients who were all confirmed to lack periconceptional folic acid supplementation. Clinical data for patients from both cohorts is summarized in Table 1. All patients included in this study are affected with non-syndromic or isolated NTDs. It is important to note that the patients group included 15 patients with cranial dysraphisms while all other were affected with various forms of open and closed spinal dysraphisms. One hundred twenty patients had a positive family history documented by clinical records (MRI and X-ray images) obtained from parents of cases.

The control group included 205 unrelated healthy random Italian individuals consisting of randomly selected children admitted to the Gaslini Children's Hospital for miscellaneous illnesses and healthy young adults who contributed samples to the blood bank of the Gaslini Institute. None had a first-degree relative with a neural tube defect. The study was approved by the local ethics committee and written informed consent was obtained from all patients, parents, and control individuals. The control group also included 44 individuals of Hispanic origin recruited as part of a case-control study conducted in a Hispanic population from the 14 counties along the Texas-Mexico border region. A detailed description of the study population has been previously reported (Barber et al, 2000). The control group also contained 55 healthy individuals of Tunisian origin, 90 CEU individuals of northern and western European ancestry from the International HapMap project (CEU, the Centre d'Etude du Polymorphisme Humain) (www.hapmap.org) and 1050 individuals from The Human Genome Diversity Project originating from 51 world populations throughout the world (www.cephb.fr/).

The coding exons of VANGL1 (GenBank NM_138959.2) were amplified from genomic DNA by PCR using primers flanking the exon-intron junctions. Direct dye terminator sequencing of PCR products was carried out using the ABI Prism Big Dye Systems. Samples were run on ABI 3700 automated sequencer and analyzed using the PhredPhrap software. Genotyping of the large panel of HapMap and HGDP controls was done using the ABI TaqMan® SNP Genotyping Assays or the Sequenom iPlex Gold technology. The latter genotyping reaction is based on template-directed single base extension using probes of various sizes. The products are then separated and detected by mass spectrometry (Ehrich et al., 2005).

RESULTS AND DISCUSSION

We detected in this cohort a total of 10 missense variants in VANGL1 that were absent in controls (Table 2). The potential pathogenic impact of the disease-associated missense variants was evaluated in view of a) the degree of evolutionary conservation of the affected residue in the Stbm/Vangl protein family, b) the nature of the amino acid replacement and its possible impact on protein function and c) possible co-segregation in additional family members.

Five of the 10 missense VANGL1 mutations detected only in affected cases, p.Ser83Leu, p.Phe153Ser, p.Arg181Gln, p.Leu202Phe, p.Ala404Ser, affect highly conserved amino acid residues and are absent from all ethnically-matched and/or HapMap and HGDP controls. These mutations are indicated in bold in Table 2 and are further described in Figure 1 and Table 3. We detected p.Ser83Leu and p. Arg181Gln in familial cases of NTDs and p.Phe153Ser, p.Leu202Phe and p.Ala404Ser in sporadic NTD cases.

The p.Ser83Leu variant was detected in three patients of Hispanic origin who are affected with a familial form of tight filum terminale with tethered cord (Table 3). Detailed clinical information and DNA from family members were unfortunately not available for these 3 patients. The variant p.Ser83Leu changes a highly conserved serine residue into a leucine in the predicted N-terminal cytoplasmic domain of the protein (Figure 1A). This serine residue is conserved across all members analyzed except in C. elegans where it is replaced by threonine (Figure 1B). While a threonine substitution preserves the hydrophilic nature at this residue position, a substitution to leucine is not conservative as it significantly reduces its hydrophilicity. It is interesting that this variant was detected in 3 NTD cases of the same Hispanic ethnic origin and who are affected with the same familial form of closed NTDs (filum terminale with tethered cord). Hispanics are the result of two-way admixture between Native American and European populations or of three-way admixture of Native American, European, and West African populations (Maox et al., 2007). Our HGDP control group included 127 subsaharan Africans, 161 Europeans and 108 Native Americans. The variant p.Ser83Leu was absent in all these 396 HGDP controls as well as 44 Hispanic and 205 Italian controls. This variant could act as an ethnic-specific predisposing factor to NTDs as it is absent in all ancestrally matched controls analyzed and dramatically affects a highly conserved amino acid residue.

The variant p.Phe153Ser was detected in a 2-years-old Caucasian white female of Tunisian origin affected with a closed spinal dysraphism (tethered cord) (Table 3). At birth, the proband showed a tuft of hair in the lumbar region and an irregular course of the gluteus cleft. MRI showed a low lying spinal cord (level L5) and a tight and thickened filum terminale at level S2-S3. A posterior schisis of the sacral vertebrae was also present. No NTD family history is present; however, the proband has a paternal first-degree cousin with congenital clubfoot, another one (born by a consanguineous mating) is affected by muscular dystrophy and 7 other first-degree cousins dead in the first month of life of unknown causes. The variant p.Phe153Ser was detected in the father of the affected proband and was absent from the mother. This variant was absent in 55 Tunisian controls, 357 Italian controls and in all 1050 individuals from the HGDP panel. This variant affects a moderately conserved Phe residue in the predicted TM2 of the protein (Figure 1A). Phe153 is replaced by leucine in mouse Vangl1, fly and worm Stbm and by tyrosine in zebrafish Vangl2 (Figure 1B). Phenylalanine, leucine and tyrosine are all hydrophobic. A substitution by a hydrophilic serine residue might affect the protein structure especially that it maps to a transmembrane domain.

The p.Arg181Gln variant was detected in an 8 years old Caucasian white male of Italian origin who suffers from lumbo-sacral myelomeningocele, severe Chiari II malformation, hydrocephalus, dilatation of right lateral and III ventricle and cervical hydromyelia (Table 3). In addition, he showed left traumatic parietal cephalocele due to cephalo-hematoma with evolutive fracture caused from mechanic insult during delivery, alteration of nervous structures near the fracture with severe cerebral suffering and right hemiparesis. Anomalies in other organs and apparatus include right inguinal hernia, phimosis, gastro-esophageal reflux, incontinent bladder, feeding and mastication problems and frequent episodes of vomiting. He represents a familial case where a distant maternal ancestor was reported to be affected by myelomeningocele. The p.Arg181Gln variant is present in the mother and absent from the father and unaffected sister (data not shown). It is interesting that this mutation was identified on the maternal side with a positive NTD family history. This variant was absent in 205 Caucasian white Italian controls, 90 Caucasian white CEU individuals and in all 1050 individuals from the HGDP panel. The p.Arg181Gln changes an arginine into a glutamine residue in the predicted intracellular loop defined by TM2 and TM3 (Figure 1A). Arg181 is absolutely conserved in all Vangl sequences analyzed (Fig. 2A). Arginine to glutamine is not a conservative substitution as it removes the ionic charge at this position.

The p.Leu202Phe variant was detected in a 9 years old Caucasian white female of Italian origin who suffers from lumbo-sacral myelomeningocele, Chiari II malformation, hydrocephalus, hydromyelia, low spinal cord, severe evolutive scoliosis with hip dysplasia, para-paresis, severe hypotonia and hypotrophy of the lower limbs. Anomalies in other organs and apparatus include neurological bladder, bowel and bladder incontinence, bilateral hydronephrosis (Table 3). This variant was absent in 207 Caucasian white Italian controls, 90 Caucasian white CEU individuals and in all 1050 individuals from the HGDP panel. The p.Leu202Phe variant changes a lysine into a phenylalanine in the predicted TM3 of the protein (Fig. 1A). Importantly, Leu202 forms part of a motif “WLF” that is absolutely conserved across evolution (Fig. 1B). The hydrophobicity of this residue position is not significantly affected by this substitution, as both of these amino acids are hydrophobic; however Leucine contains an aliphatic side chain while Phenylalanine has an aromatic side chain. The introduction of an aromatic ring at this position would introduce a bulkier amino acid that may affect the formation of this third transmembrane domain, resulting in a non-functional protein.

The p.Ala404Ser variant was detected in a 21 years old Caucasian white male of Italian origin affected with caudal regression syndrome characterized by partial sacro-coccygeal agenesis (type III, according to Pang's classification of sacral agenesis) and dysmarphism of spinal cord ending abruptly at L2, hydromyelia, schisis at S1 and reduction of the spaces between L3 and L4. He also shows scoliosis, knock knees, flat feet (Table 3). The p.Ala404Ser variant was absent in 204 Caucasian white Italian controls and in all 1050 individuals from the HGDP panel . This variant changes a highly conserved alanine residue in the predicted cytoplasmic domain of the protein (Figure 1 A). Alanine is conserved across all members analyzed except in worm Strabismus where it is replaced by proline (Figure 1B). While proline preserves the hydrophobic nature of alanine, this property is completely lost by the non conservative change to a hydrophilic serine residue.

The five other mutations that were only found in affected cases, p.Glu25Lys, p.Arg175Gln, p.Thr251Met, p.Tyr290His and p.Asp468Glu, affect poorly or moderately conserved amino acid residues across evolution (Table 2). The p.Glu25Lys variant was detected in one NTD Italian patient and changes a poorly conserved glutamate residue to lysine in the predicted N-terminal cytoplasmic domain of the protein (data not shown). The p.Arg175Gln variant was identified in one patient and changes a poorly conserved arginine residue to glutamine in the predicted intracellular loop defined by TM2 and TM3 (data not shown). The p.Thr251Met variant was detected in 2 NTD patients and changes a threonine to methionine in the cytoplasmic domain of VANGL1. This threonine is moderately conserved across evolution; in zebrafish Vangl2, it is replaced by cysteine that is closely related to methionine, suggesting a non pathogenic role for the p.Thr251Met substitution. The p.Tyr290His variant, caused by c.868T>C, changes a tyrosine residue into a histidine in the predicted cytoplasmic domain of the protein. This tyrosine is moderately conserved where it is replaced by histidine in Zebrafish Vangl1 and phenylalanine in Drosophila and C. elegans Strabismus (data not shown). The fact that Tyr290 is replaced by histidine in Zebrafish Vangl1 argues against a pathogenic role for this variant. The variant p.Asp468Glu changes a poorly conserved aspartate into glutamate in the cytoplasmic domain of VANGL1. In fact, Asp468 is replaced by glutamate in 6 homologues/orthologues of the Vangl family arguing against a pathogenic role for p.Asp468Glu (data not shown). All together, these data suggest that p.Glu25Lys, p.Arg175Gln, p.Thr251Met, p.Asp468Glu and p.Tyr290H are most likely not pathogenic.

We also identified in this cohort 11 VANGL1 mutations that were either silent or present in both NTD patients and controls. Of these, 10 variants were rare (<1%) and one variant represents a common coding SNP within the VANGL1 gene, rs4839469: G>A (p.Ala116Thr). These rare 10 variants and common SNP are shown in Table 1 and Table 2 of the Supplementary Appendix respectively. We evaluated the association of this SNP with the NTD phenotype in our cohort. The allele and genotype frequencies at this SNP showed an ethnic–specific variation in African Americans as compared to Hispanics and Caucasian whites. A larger sample size is needed to confirm and investigate this variation. We only conducted association analysis in the major Caucasian white group in our cohort where we included an additional set of 137 Italian patients from our previous study. In this group, rs4839469:G>A was detected at an allele frequency of 0.15 in the NTD patients and 0.14 in the controls. The genotypes GG, GA and AA were detected at a frequency of 0.72, 0.26 and 0.02 in the patients group and 0.74, 0.25, 0.01 in the control group respectively. We also conducted our association analysis after stratification by the type of NTD (open versus closed) (data not shown). Before and after stratification, we did not detect any statistically significant association between the SNP rs4839469:G>A and NTDs in our cohort (data not shown).

To determine the frequency of VANGL1 mutations in the control population, we sequenced the whole ORF and exon-intron junctions of this gene in a group of 205 Caucasian white controls of Italian origin (data not shown). Two silent (c.510T>C and c.546G>A) and 2 missense (c.1180G>A, c.1340C>T) mutations were detected only in controls. The c.1340C>T mutation changes a moderately conserved alanine residue to a valine residue (p.Ala447Val) at position 447 in the cytoplasmic domain of the predicted protein. Alanine and valine have similar physicochemical properties implicating a nonpathogenic effect of this variant. On the other hand, the c.1180G>A mutations changes an absolutely conserved alanine residue to a threonine at position 394 in the cytoplasmic domain of the predicted protein (p.Ala394Thr). Alanine to threonine substitution is not conservative as it eliminates the hydrophobicity at this position. The p.Ala394Thr variant was absent in all 204 Caucasian white Italian controls, 90 Caucasian white CEU individuals and in all 1050 individuals from the HGDP panel. This result should not be completely surprising as we detected each of three VANGL1 mutations, p.Arg181Gln and p.Phe153Ser in this study and p.Val239Ile previously (Kibar et al., 2007b), in the normal mother or father of the severely affected proband. We are most likely dealing with low penetrance alleles where the incidence and severity of the phenotype is modulated by the presence of other unknown genetic/environmental modifiers and where the genotype at a single locus is insufficient to explain phenotypic variability. Another possibility for this finding is that the control carrying p.Ala394Thr could be affected by a very mild form of NTDs such as vertebral schisis that is asymptomatic and has not been detected yet. It is also possible that the p.Ala394Th variant is not pathogenic. Functional validation of this variant is important to validate its effect on VANGL1 function and to further investigate its occurrence in an “apparently healthy” control.

Our present study provides additional evidence supporting the role of VANGL1 as a risk factor for development of spinal NTDs. We detected 7 patients that were heterozygous for 5 novel missense mutations, two of which, p.Ser83Leu and p.Arg181Gln, occur in familial settings. Four mutations, p.Phe153Ser, p.Arg181Gln, p.Leu202Phe and p.Ala404Ser, were “private” and one mutation, p.Ser83Leu, was recurrent in 3 familial cases of tethered cord syndrome. These mutations affect evolutionary conserved amino residues that are distributed along the entire length of the VANGL1 protein. Since these variants do not represent obvious null mutations (like stop codon, deletions), functional experiments are needed to investigate their effect on the protein function. All mutations detected so far in VANGL1 in NTDs are heterozygous, leading to the speculation that these variants may act as partial loss of function (LOF) alleles and interact with other environmental and genetic factors to cause the NTD phenotype. This is consistent with studies in mice where Vangl1 genetically interacts with Vangl2 to cause craniorachischisis and where it is postulated that heterozygosity for LOF alleles at Vangl1 may act as a sensitizing genetic lesion that predisposes to NTD in the presence of other environmental and/or genetic factors (Torban et al., 2008).

The phenotypic variability of NTDs as well as the limited number of affected patients with VANGL1 mutations has made it hard to correlate the type of mutations found with the clinical findings. It is important to note that we detected VANGL1 mutations in this study and previously (Kibar et al., 2007b) in both open and closed forms of spinal dysraphisms. One still undecided controversy is whether NTDs at different levels represent different defects caused by distinct genetic lesions. Studies of NTD recurrences in families demonstrated that NTDs tend to breed true within families but with a large degree of phenotypic variability (Detrait et al., 2005). In our previous study, we identified two mutations in VANGl1, p.Val239Ile and p.Arg274Gln, in 2 familial cases of spinal NTDs with intra-family phenotypic variability. The variant p.Val239Ile was detected in a proband affected with a severe form of caudal regression and in her brother affected with the milder dermal sinus NTDs. The variant p. Arg274Gln was detected in a female affected with an open myelomeningocele and in her mother affected with closed vertebral schisis. In mice, open and closed NTDs are caused by failure of neural tube closure at two modes of neurulation: failure of primary neurulation at any level of the body axis from the brain down to the upper sacral spine leads to “open” NTDs while failure of secondary neurulation at most of the sacral and coccygeal regions lead to the less common “closed” forms of NTDs (Kibar et al., 2007a). In particular, caudal regression is hypothesized as a notochordal defect that occurs early during gastrulation (Pang 1993). Fate mapping studies in the chick suggest that the neural plate gives rise to both the primary and secondary neural tubes and that both modes of neurulation represent a continuous program with similar molecular and cellular mechanisms. Consequently, it was hypothesized that the pathophysiology of the rostral and caudal NTDs does not necessarily implicate different mechanisms (Catala 2002). Our findings of VANGL1 mutations in both open and closed NTDs support this hypothesis of common underlying molecular mechanisms.

We have detected VANGL1 mutations only in a small fraction of NTD patients. This could be due to many factors. First, patients may have large heterozygous deletions at VANGL1 that could not be detected by PCR and sequencing. In fact, exciting data have recently emerged on the role of submicroscopic genomic imbalance or copy number variants as a frequent cause of birth defects (Feuk et al., 2006). Second, DNA-methylation changes at VANGL1 could be associated with NTDs. This is particularly interesting with the development of the methylation hypothesis which suggests that folic acid protects against NTDs by stimulating cellular methylation reactions (Blom et al., 2006). Third, mutations in VANGL1 in NTDs could be present in regulatory non-coding regions such as promoters that would affect the level of gene transcription. Finally, other genetic factors could be involved in the etiology of NTDs in this cohort, consistent with a multifactorial model for NTDs. Other members of the PCP pathway represent excellent candidates for involvement in spinal NTDs stressing the need for their systematic genetic and biochemical testing for their role in the etiology of NTDs. It is only when this testing is complete that we would begin to decipher the genetic complexity of spinal NTDs.