CHD7 gene polymorphisms are associated with susceptibility to idiopathic scoliosis

Abstract

Idiopathic scoliosis (IS) is the most common spinal deformity in children, and its etiology is unknown. To refine the search for genes underlying IS susceptibility, we ascertained a new cohort of 52 families and conducted a follow-up study of genomewide scans that produced evidence of linkage and association with 8q12 loci (multipoint LOD 2.77; P=.0028). Further fine mapping in the region revealed significant evidence of disease-associated haplotypes (P<1.0 x 10-4) centering over exons 2-4 of the CHD7 gene associated with the CHARGE (coloboma of the eye, heart defects, atresia of the choanae, retardation of growth and/or development, genital and/or urinary abnormalities, and ear abnormalities and deafness) syndrome of multiple developmental anomalies. Resequencing CHD7 exons and conserved intronic sequence blocks excluded coding changes but revealed at least one potentially functional polymorphism that is overtransmitted (P=.005) to affected offspring and predicts disruption of a caudal-type (cdx) transcription-factor binding site. Our results identify the first gene associated with IS susceptibility and suggest etiological overlap between the rare, early-onset CHARGE syndrome and common, later-onset IS.

Figure 1.

Pedigree of family IS14. Blackened symbols indicate affected individuals, and the arrow indicates the proband. Checkerboard patterns denote individuals with mild scoliosis (<15° Cobb angle) who were scored as “unknown” in subsequent analyses. We note that individual IS14–19 was originally reported as affected but, upon reevaluation, was considered to be of unknown affection because of borderline curve measurements.

Figure 2.

Results of genomewide scan in family IS14. Distance across chromosomes is plotted on the X-axis versus results of linkage analysis along the Y-axis. Resulting LOD scores for parametric analyses in which we considered only affected individuals and dominant inheritance are plotted as solid lines for each chromosome. Maximal results were obtained from chromosomes 1, 8, and 10. The top three NPL scores also occurred for chromosomes 1, 8, and 10 and are overlaid and plotted as dashed lines.

Figure 3.

IS in a representative proband from the 52-family set. Standing posteroanterior radiograph reveals a right thoracic curve in an otherwise healthy adolescent female.

Figure 4.

Analyses of linkage and transmission disequilibrium for 8q microsatellite loci in 52 families with IS. Polymorphic microsatellites spaced at 5–10 cM were genotyped in all members of the 52 families. The method of KAC was used to compute linkage (dashed line), and family-based association was measured using the TDTae (solid line). For reporting consistency, results are shown as P values (−log transformed) versus position (in cM) for the two methods.

Figure 5.

Fine-mapping results for the CHD7 gene. A, CHD7 genomic region is shown with exons indicated in blue and intronic conserved sequence blocks shown in red. B, Plot of single-point linkage and transmission disequilibrium P values for 23 SNPs in the CHD7 gene. −log₁₀P values are plotted along the Y-axis versus physical position along the X-axis for each SNP. Results for the ASP, HHRR, and TDTae statistics are shown by diamonds (red), squares (blue), and triangles (orange), respectively. The TDTae method is more significant than the ASP method for markers in a region of high pairwise LD, as expected, given that the TDT method was originally developed to increase evidence of linkage when marker and trait loci are in high LD. C, Graphical representation of the pairwise LD (Δ²) values for all 23 SNPs. This panel may be thought of as a “heat map” of the pairwise LD values (Δ²). Both the horizontal and vertical axes represent the 23 SNPs shown by tick marks and ordered as in table 1, and places where the axes intersect indicate the pairwise Δ² values for those marker pairs. In this panel, pairs of markers with larger Δ² values (close to or equal to 1, indicating complete LD) are denoted in red, whereas pairs with smaller Δ² values (closer to or equal to 0, indicating linkage equilibrium) are denoted in blue, as illustrated by the vertical color bar. Panels A, B, and C have been aligned so that results for each of the markers correspond among the three panels.

Figure 6.

Maximum multilocus TDT results for each set of four SNPs. For SNPs that appear in more than one set of overlapping windows, we report the average of the log-transformed P values for the two maximum multilocus TDT statistics. P values were computed using a bootstrap sample of 50,000.