Genetic investigation of recurrent exertional rhabdomyolysis risk and performance in racehorses

Abstract

Recurrent exertional rhabdomyolysis (RER) is a disease in Thoroughbred and Standardbred racehorses involving repeated episodes of skeletal muscle cell breakdown in response to exercise in otherwise physically fit horses. Based on epidemiological and heritability studies, RER has been hypothesized to be a complex polygenic disease with significant sex and environmental influences. Horses affected by RER are often managed differently from their unaffected counterparts, commonly with changes to their diets and exercise regimens having success in preventing or reducing the frequency of ER episode. Early identification of horses at risk for RER therefore has the potential to reduce the frequency of ER episodes, as high risk individuals could be managed more appropriately prior to ever experiencing ER. An in-depth understanding of the genetic basis of RER may provide insight into the disease’s pathophysiology, which can eventually lead to more targeted management and therapies. In addition, there is evidence to suggest that RER in Standardbreds is associated with better racing performance. Thus, additional insight into the nature of this association and any genetic similarities could be helpful in guiding breeding strategies in this population to decrease the prevalence of this common disease while maintaining a high standard of racing performance. In this study, we evaluated risk factors for RER and identified differences in racing careers for North American Standardbred racehorses with and without RER. Prevalence for RER is much higher in our study than in previous studies of STDB populations outside of North America. We found that mares are at significantly higher risk for RER than stallions. An additional level of sex-associated risk in geldings was identified in the trotter subpopulation but not in the pacer subpopulation of STDBs. We also replicated results suggesting an association between RER and better racing performance in the Standardbred. This association between performance outcomes in horses with RER indicates that selective breeding for performance may be a driving force in maintenance of this disease at a high frequency in the population. Next, we assessed the genetic basis of several performance traits in Standardbred racehorses using high density genome-wide SNP data. Heritability for fastest mile times, fastest last quarter mile times, and career earnings in a large population of North American STDB racehorses was moderately heritable, with the greatest variability in heritability estimates coming from career earnings. Our genetic architecture analysis revealed that variability in these performance traits can be explained by many SNPs located throughout the genome with varying effect sizes. Our GWAS did not reveal many regions of the genome with high confidence for any of the performance metrics evaluated. The goals of our investigation of the genetic basis of RER in Thoroughbred and Standardbred racehorses were to identify variants predictive of RER in both breeds, and to discover putative functional variants in plausible biological candidate genes which may be associated with the disease. Our genetic architecture analyses further demonstrate the complex nature of RER in Thoroughbreds and Standardbreds. The majority of the heritability was explained by SNPs with large and medium effect sizes, providing evidence for the feasibility of a multi-marker assay for RER risk prediction in these two breeds. Single- and multi-marker GWAS and haplotype analyses revealed 16 regions of the genome associated with RER susceptibility in either or both breeds, one of which was previously reported in TBs. Random forest (RF) feature selection resulted in 3,547 loci predictive of RER case-control status. We designed a custom genotyping assay with 28,514 variants to develop a predictive risk model for RER and validate loci associated with the disease in an independent cohort of racehorses and other related horse breeds. Although these loci were separately selected in each breed, only 3,325 (12%) variants were chosen that were unique to either the TB or STDB, providing further evidence that both breeds likely share a significant genetic component to RER risk. The primary goal for follow-up from the results of this project will be to continue genotyping additional samples using our custom genotyping assay. We plan to extend this work beyond simple case-control classification by developing a risk model that estimates the probability that an individual will develop RER. Afterwards, we will validate this predictive risk model in an independent cohort of TBs and STDBs phenotyped for RER. The models developed in our follow-up studies will be further evaluated for use in other horse breeds with significant RER risk.