An ARHGEF10 deletion is highly associated with a juvenile-onset inherited polyneuropathy in Leonberger and Saint Bernard dogs

Abstract

An inherited polyneuropathy (PN) observed in Leonberger dogs has clinical similarities to a genetically heterogeneous group of peripheral neuropathies termed Charcot-Marie-Tooth (CMT) disease in humans. The Leonberger disorder is a severe, juvenile-onset, chronic, progressive, and mixed PN, characterized by exercise intolerance, gait abnormalities and muscle atrophy of the pelvic limbs, as well as inspiratory stridor and dyspnea. We mapped a PN locus in Leonbergers to a 250 kb region on canine chromosome 16 (Praw = 1.16×10-10, Pgenome, corrected = 0.006) utilizing a high-density SNP array. Within this interval is the ARHGEF10 gene, a member of the rho family of GTPases known to be involved in neuronal growth and axonal migration, and implicated in human hypomyelination. ARHGEF10 sequencing identified a 10 bp deletion in affected dogs that removes four nucleotides from the 3'-end of exon 17 and six nucleotides from the 5'-end of intron 17 (c.1955_1958+6delCACGGTGAGC). This eliminates the 3'-splice junction of exon 17, creates an alternate splice site immediately downstream in which the processed mRNA contains a frame shift, and generates a premature stop codon predicted to truncate approximately 50% of the protein. Homozygosity for the deletion was highly associated with the severe juvenile-onset PN phenotype in both Leonberger and Saint Bernard dogs. The overall clinical picture of PN in these breeds, and the effects of sex and heterozygosity of the ARHGEF10 deletion, are less clear due to the likely presence of other forms of PN with variable ages of onset and severity of clinical signs. This is the first documented severe polyneuropathy associated with a mutation in ARHGEF10 in any species.

Figure 1. Genome-wide association mapping of Leonberger PN.

(A) Manhattan plot of GWAS for PN with 93 (52 cases, 41 controls) Leonbergers. X-axis = chromosome, Y-axis = negative decadic logarithms of permutation-corrected P-values. (B) Manhattan plot of GWAS for PN with 93 Leonbergers. X-axis = chromosome, Y-axis = negative decadic logarithms of permutation-corrected P-values, after correction for population stratification and cryptic relatedness using a mixed model approach and genomic control. (C) The quantile-quantile (QQ) plot corresponding to (B).

Figure 2. SNP haplotypes and genes from the GWAS identified CFA16 region associated with Leonberger PN.

A representative sample of the observed CFA16 haplotypes is demonstrated using eleven Leonbergers, with both haplotypes shown for each dog as green or yellow bars. LEO 1-4 and LEO 9 - 11 are cases (indicated in red), while LEO 5-8 are controls (indicated in blue). The most significantly associated SNP (BICF2G630820235) is highlighted in orange. Green bars represent the most significantly associated haplotype, which, in some cases extended the entire 1.5 Mb region. However, a minimum associated haplotype of approximately 250 kb (enclosed in the dark outlined box) was observed. An alternate haplotype, demonstrated by the yellow bars, was present in many of the controls and some of the cases. Some PN cases (LEO 9-11) are either heterozygous or devoid of the “green” case-associated haplotype and a phenotypically normal dog (LEO 8) is heterozygous for the associated haplotype. CFA16 (Mb, CanFam2) ENSEMBL-annotated genes in the region are shown at the bottom of the figure.

Figure 3. ARHGEF10 mutation analysis.

(A) Genomic DNA. Electropherograms of the ARHGEF10 c.1955_1958+6delCACGGTGAGC mutation. Representative sequence traces of PCR products amplified from genomic DNA of 3 dogs with the different genotypes are shown. (B) Transcript. Electropherograms of RT-PCR using primers in exon 17 and 19. Representative sequence traces of RT-PCR products amplified from cDNA of 3 dogs with the different genotypes are shown. Note the underrepresentation of the mutant transcript in the heterozygous dog, which is probably due to nonsense mediated decay (NMD). Both the normal and the new splice site created by the deletion are underlined (Panel A); the cryptic splice site in the deletion allele creates a cDNA missing only four base pairs, which creates a frame shift and predicts a premature stop codon within seven amino acid residues (Panel B).

Figure 4. Age-of-onset based on ARHGEF10 genotype.

Age-of-onset of clinical signs for all well-phenotyped cases (where age-of-onset was reported) is shown for all cases from the original GWAS and the independent, validation group, by ARHGEF10 genotype. Not all 206 dogs are shown due to the fact that, while we sometimes knew the dog had clinical signs by a certain age, the exact onset of those signs before that age was unknown.