Congenital vertebral malformations are phenotypically and etiologically heterogeneous. Their estimated incidence is between 0.5 to 1/1000 1234. Vertebral malformations may represent an isolated finding, occur in association with other renal, cardiac, or spinal cord malformations, or occur as part of an underlying chromosome abnormality or syndrome. These include, but are not limited to, hemifacial microsomia, Alagille, Jarcho-Levin, Klippel-Feil, Goldenhar,, basal cell nevus, trisomy 18, diabetic embryopathy and VACTERL (vertebral, cardiac, renal, limb anomalies, anal atresia, tracheo-esophageal fistula) syndromes. Vertebral malformations most commonly include hemivertebrae, vertebral bars, supernumerary vertebrae, butterfly vertebrae, and wedge-shaped vertebrae.

Congenital scoliosis is caused by segmentation defects such as fused vertebrae, vertebral body formation defects, and mixed defects in which both types of lesions are encountered 34. Each of these conditions may cause development of a spinal curve based on asymmetric growth. The severity of the curve is related to the type of defect and whether or not the primary problem is accompanied by any compensatory developmental changes.

Vertebral segmentation proceeds through a clock and wave front mechanism under Notch signaling control. Negative feedback through the canonical Wnt/β catenin pathway via Axin 2 modulates the process 56. Based on mouse-human synteny analysis, a series of candidate genes known to cause vertebral malformations in the mouse have been identified 5678. One gene, Wnt3a, is necessary for generation of the posterior portion of the neuraxis, as knockout mice fail to develop a tailbud and are truncated from a point slightly anterior to the hindlimbs 9. This gene is a member of a moderate-sized multigene family comprised of at least 12 members in humans and the mouse. Genes in this family function both in establishing the body plan in development and as potential oncogenes 10. Wnt3a has been proposed to be a major controlling gene in the oscillation of Notch signaling that is necessary for segmentation to occur 5.

Since isolated congenital vertebral malformation most often represents a sporadic occurrence within a particular family, it is virtually impossible to utilize traditional linkage approaches to identify causative genes. This makes candidate gene analyses a viable alternative method to study this condition. We hypothesized that mutations in WNT3A may be associated with the development of isolated vertebral malformations and congenital scoliosis. In order to test this hypothesis we undertook a pilot study and performed DNA sequence analysis in a heterogeneous cohort of 50 patients with congenital vertebral malformations.

Vertebral malformations were classified as butterfly vertebra, segmentation defect, or hemivertebra/hypoplasia. The vertebral malformations spanned the length of the entire spine and were classified as butterfly vertebrae, segmentation defect, hypoplasia and hemivertebrae 4. Butterfly vertebrae were defined by the presence of a sagittal cleft. Segmentation defects include defects arising from intervertebral discs such as block vertebrae and unsegmented vertebral bars. The vertebral malformations spanned the length of the entire spine and are illustrated (figure 1). The malformation characteristics of the study sample are highlighted in Table 2.

To analyze the WNT3A gene in the selected vertebral malformation cohort, a total of four segments in the WNT3A gene were amplified and sequenced in each patient. The sequencing covered the four exonic regions, splice junctions, and 381 bp of the promoter region. A 23-month-old female child of Cape Verde African ancestry with an isolated T12-L1 hemivertebrae and congenital scoliosis with a 38° lumbar scoliotic curve was found to be heterozygous for a GCC to ACC missense variant at codon 134 (UCSC position chr1:224,545,178 May 2004 assembly) of the WNT3A gene (figure 2), resulting in the substitution of alanine by threonine. The patient's clinically asymptomatic father was also heterozygous for the missense variant, whereas her mother was not. This exon is highly conserved, whereas codon 134 is less conserved in this exon. Interestingly, the alanine amino acid at codon 134 is conserved in mouse, rat, dog and opossum, whereas in elephant, chicken, Xenopus tropicalis and tetradon, alanine is replaced by threonine. To check for the prevalence of this variant in reference population, a total of 890 chromosomes from 445 subjects were tested for the WNT3A mutation, of which 133 were of African origin, 299 were Caucasians, and 11 were Asians. Only three subjects of African origin (Yoruba, Mbuti, and Bantu N.E.) were heterozygous for the variant. The overall prevalence of the variant in the tested reference population is 0.35%.

Moreover, a novel C/G single nucleotide variant was also found in the promoter region of the WNT3A gene (UCSC position chr1:224,501,511 May 2004 assembly). The patient cohort had an average heterozygosity of 0.130 and allele frequencies of 0.93 and 0.07 for the C and G alleles, respectively. In order to determine the significance of this variant, we typed a panel of 23 anonymized reference DNA samples of Caucasian ancestry. The average heterozygosity for this variant in the reference cohort was calculated to be 0.159 and the allele frequencies were 0.91 and 0.09 for the C and G alleles, respectively. The genotype distribution for this variant did not differ significantly between the patient's and reference cohorts (Chi-squared = 0.001, p-value = 0.97) and were in Hardy-Weinberg equilibrium in both cohorts. We conclude that this variant likely represents a polymorphism.

In a series of 50 patients with a broad array of vertebral malformations encompassing the entire length of the spinal column, we only identified a non-synonymous variant in exon 3 of WNT3A, resulting in substitution of alanine for threonine. Our observation of this variant with an overall frequency of 0.35% in the control population studied, and a frequency of 1.1% of African chromosomes included in the control population suggest and the observation that the patient's clinically asymptomatic father harbors the same variant strongly suggest that this variant may represent a polymorphism in the African population. Based on the data obtained from our pilot study, mutations in the WNT3A gene do not appear to be a common etiologic factor in the development of congenital vertebral malformations.

Patient heterogeneity for vertebral malformations and heterogeneity of candidate genes represent major limitations in genetic analyses. It is possible that mutations in WNT3A would have been identified if we had sequenced a cohort of patients characterized by having caudal vertebral malformation phenotypes. Because congenital vertebral malformations represent a rare condition, collection of a phenotypically homogeneous sample is not practical and would require a greater amount of time. DNA sequence analysis was limited to the WNT3A promotor, exons, and intron-exon boundaries. Sequence variants in the introns that could potentially contribute to abnormal transcription would not have been detected. Copy number variations in WNT3A will not have been detected by this methodology. Other genes that regulate the expression of WNT3A have not been examined.

Evidence for genetically heterogeneous etiologies of vertebral malformations includes the identification of mutations in DLL3 in patients with spondylocostal dysostosis 1415 and mutations identified in JAG 1 in patients with Alagille syndrome 16. Recently, a few DNA coding sequence alterations with uncertain clinical significance in PAX1 and DLL3 have been identified in a cohort of patients with phenotypically characterized vertebral malformations 17181920. Our inability to detect mutations in WNT3A, as well as other candidate genes studied, may reflect differences in the regulation of somitogenesis in humans as compared to mice. It is possible that mutations in WNT3A are lethal in humans. Although our pilot study included 5 stillbirth samples and no mutations in WNT3A were detected, a larger cohort of stillbirth samples with vertebral malformations would be needed to determine whether mutations in WNT3A are prevalent among stillbirths with congenital vertebral malformations. Tissue mosaicism for a specific mutation and/or epigenetic factors may have an influence in the development of vertebral malformations in humans. Vertebral malformations may have greater environmental components than genetic components to their origin. A relationship between abnormal distribution of the intersegmental arteries and vertebral malformations was observed in 4 of 11 human embryos and fetuses with congenital vertebral malformations, suggesting the possibility that abnormalities in vascularization during the resegmentation process may contribute towards the development of congenital vertebral malformations 21.

Clinical relevance in identifying mutations associated with the development of vertebral malformations and congenital scoliosis is to develop a molecular classification system that could provide prognostic information regarding the severity of progression of congenital scoliosis and the presence of associated birth defects to practitioners and families. Although congenital scoliosis and idiopathic scoliosis represent dichotomous processes, namely a failure to develop normal vertebral structures for the former and a failure to lose normal vertebral structures for the latter, there is evidence that similar genetic factors may play a role in both conditions 22. Identification of gene mutations which are associated with the development of congenital scoliosis may also help improve our understanding of the pathogenesis of idiopathic scoliosis.

In summary, we have been unable to detect mutations in the WNT3A gene in a pilot study of 50 patients with vertebral malformations encompassing the entire spine. This study illustrates the problems posed by attempting to design studies aimed at identification of patterning genes contributing towards the development of vertebral malformations. Future studies aimed at investigating the etiology of congenital vertebral malformations should be aimed at the study of additional candidate genes in a more homogeneous patient population.

The author(s) declare that they have no competing interests.

PFG, NG, CLR, RDB, EM, JKB, IG, FSJ, TF and OBA designed or assisted in the study design. NG, LI, and JK performed laboratory analysis of DNA samples obtained from patients. NG, CLR, RDB, EM, JKB, LI, KR, and JK assisted in the interpretation of data. PFG recruited patients. PFG, CLR, RDB, EM, KR, FSJ, TF, RMP, and OBA assisted in the analysis of phenotypic data. RDB assisted in the analysis of molecular data and PFG, IG, RMP performed the analysis of data. All authors assisted in the preparation and/or review of the manuscript and have read and approved the final manuscript.