Genetic associations and phenotypic heterogeneity in the craniosynostotic rabbit

Abstract

Craniosynostosis (CS) is a disorder that involves the premature ossification of one or more cranial sutures. Our research team has described a naturally occurring rabbit model of CS with a variable phenotype and unknown etiology. Restriction-site associated DNA (RAD) sequencing is a genomic sampling method for identifying genetic variants in species with little or no existing sequence data. RAD sequencing data was analyzed using a mixed linear model to identify single nucleotide polymorphisms (SNPs) associated with disease occurrence and onset in the rabbit model of CS. SNPs achieving a genome-wide significance of p ≤ 5 x 10-8 were identified on chromosome 2 in association with disease occurrence and on chromosomes 14 and 19 in association with disease onset. Genotyping identified a coding variant in fibroblast growth factor binding protein 1 (FGFBP-1) on chromosome 2 and a non-coding variant upstream of integrin alpha 3 (ITGA3) on chromosome 19 that associated with disease occurrence and onset, respectively. Retrospective analysis of patient data revealed a significant inverse correlation between FGFBP-1 and ITGA3 transcript levels in patients with coronal CS. FGFBP-1 and ITGA3 are genes with roles in early development that warrant functional study to further understand suture biology.

Fig 1. Graphical representation of p values obtained from MLM analysis of RAD sequencing data.

Panels correspond to disease occurrence (1a, 1c) and age-of-onset (1b, 1d). For Manhattan plots (1a, 1b) the x axis corresponds to the genomic position of the autosomes and the y axis shows the −log10 of the p value. The horizontal red line corresponds to the genome-wide significance threshold of p ≤ 5 × 10⁻⁸. For regional association plots (1b, 1d) the x axis corresponds to the chromosomal position and the y axis shows the −log10 of the p value. Genes in the region are shown below (not to scale).

Fig 2. Linkage disequilibrium pattern (D’) and minor allele frequencies (MAF) for disease occurrence (2a, 2b) and age-of-onset (2c, 2d).

Linkage blocks were identified in Haploview (2a, 2c). A standard color scheme is used to display LD with bright red color for very strong LD (LOD = 2 D' = 1), white color for no LD (LOD<2, D'<1), pink red (LOD = 2 D'<1), and blue (LOD<2 D' = 1) for intermediate LD. Analysis of Variance (ANOVA) revealed significant differences in allelic distribution for all groups within the linkage block on chromosome 2 and between EOS rabbits and all other populations within the linkage block on chromosome 19.

Fig 3. Pedigree of the CS rabbit colony.

Colony animals were retrospectively genotyped for an A → G transition located within the FGFBP-1 coding region at chromosomal position 2:6,430,107 and for a C → G transversion located upstream of ITGA3 at chromosomal position 19:37,599,083. Biallelic genotypes are reported for FGFBP-1/ITGA3. Color-coding for phenotype is summarized in the figure key.

Fig 4. Growth across the coronal suture by diagnosis and by genotype.

Radio-opaque amalgam markers anterior and posterior to the coronal suture during initial diagnosis at 10 days of age. Radiographic growth measurements were made at day 25 (■) and day 42 ($■$) as described in Methods. Two-way ANOVA revealed significant differences in growth rates when animals were grouped by diagnosis (4a). More granular distinctions were evident within the population when animals were grouped by genotype (4b). Progressive accumulation of variant alleles at the two loci correlated with increasing growth restriction. Significant growth restriction was evident in heterozygotes with the AG/CG genotype when compared to the reference AA/CC genotype (p < 0.001) or to the single variant AA/CG genotype (p < 0.05). Animals with the GG/CG genotype exhibited significantly different growth compared to all other groups (p ≤ 0.0001) and animals with the GG/GG genotype exhibited significantly different growth compared to all other groups (p ≤ 0.0001) as denoted by an asterisk (*).

Fig 5. FGFBP-1 expression negatively correlates with ITGA3 expression in patients with fusion of the coronal suture but not in patients with fusion of the metopic or sagittal sutures.

Scatter plots describing the correlation between FGFBP-1 and ITGA3 expression were prepared for a patient cohort consisting of 199 patients and controls [4]. Patient samples were associated with isolated fusion of the coronal, metopic, or sagittal sutures. No correlation was evident within the control population (5a). A significant (p ≤ 0.0155) negative correlation was evident in patients with isolated coronal suture fusion (5b). Correlation of gene expression with isolated metopic (5c) or sagittal (5d) suture fusion fell below the level of significance suggesting that this interaction may be unique to the coronal suture.