Detection and characterisation of genetic associations with canine skull shape and disease

Abstract

The brachycephalic head conformation is a morphological trait under human selection across many popular breeds such as the Pug. The conformation is characterised by the concurrent rostrocaudal shortening and mediolateral widening of the skull. Overwhelming evidence suggests this skull conformation predisposes breeds to a respiratory morbidity, Brachycephalic Obstructive Airway Syndrome (BOAS) which restricts airflow. This is a major welfare concern. To date, very little is known regarding the genetic factors underpinning canine craniofacial variation or how such factors may influence BOAS aetiology. This project utilised computer tomography scans of 560 dogs to generate high resolution three-dimensional reconstructions of the skull. Geometric morphometric analysis permitted the separation of the confounding influence of size on shape (allometry) at an individual-level – a study design never previously utilised in dogs. Genome-wide association studies (GWAS) using viscerocranial shape changes identified a QTL on chromosome 1 associated with canine brachycephaly. Haplotype mapping refined the critical interval to encompass the SPARC-related modular calcium-binding protein 2 (SMOC2) gene. Leveraging haplotype mapping and whole-genome sequencing, a long interspersed nuclear element (LINE-1) was discovered within intron eight of SMOC2. Transcriptomic analysis revealed the presence of alternative SMOC2 transcripts containing premature stop codons and a downregulation of the gene among brachycephalic carriers of the LINE-1. Models of phenotypic effect predicted that this structural variant explains up to 36% of the total craniofacial variation in the dog.

Despite many brachycephalic dogs being fixed for the SMOC2 LINE-1 mutation, heterogeneity in BOAS presentation suggests additional genetic factors contribute to disease risk and presentation. To address this, a respiratory distress syndrome with remarkable similarities to BOAS was assessed in the Norwich Terrier. Endoscopic examinations of 233 Norwich Terriers graded and characterised the respiratory syndrome for the use in a GWAS. A single QTL on chromosome 13 was associated with changes in the structure of the laryngeal saccules and cartilage of affected dogs.

The underlying disease-associated haplotype encompassed the disintegrin and metalloproteinase with thrombospondin motifs 3 (ADAMTS3) gene which was observed to harbour a missense mutation in severely affected dogs. Screening for the variant across 89 diverse dog breeds revealed it was enriched in brachycephalic breeds, suggesting that the brachycephalic conformation alone does not completely explain the presentation of BOAS in these breeds. Together, these results provide the opportunity to improve animal welfare by offering genetic screening tests and further our understanding of the condition which is driven by both skeletal and non-skeletal factors.