sign in sign up
<<
Home , Australian Mist, Bengal cat, British Saddleback, Cat genetics, Chillingham cattle, Egyptian Mau

September 27-September 29, 2013 Boston, MA

Thank you to our generous sponsors

Perpetual public-accessible conference proceedings website: www.vin.com/tufts/2013 2 6th Tufts’ Canine and Feline Breeding and Conference

Scientific Program

Saturday, September 28

Lecture Time: Title of Lecture: Speaker:

8:30-9:10 Unraveling the Sources of Genetic Structure Within Breeds Dr. Pam Wiener

9:10-9:50 Taking Advantage of Dog Breed Structure to Understand Dr. Elaine Ostrander Health

10:10-10:30 Genetics of Populations and Breeds: Implications for Dr. Leslie Lyons Breed Management for Health!

10:30-11:10 Breeding Practices According to Breeds; Time, Place, and Dr. Grégoire Leroy Consequences

11:10-11:30 , Outbreeding, and Breed Evolution Dr. Jerold Bell

1:15-1:55 Unraveling the Phenotypic and Genetic Complexity of Dr. Paula Henthorn Canine Cystinuria

1:55-2:35 How to Use and Interpret Genetic Tests for Heart Disease Dr. Kathryn Meurs in and Dogs

2:55-3:35 Update on Genetic Tests for Diseases and Traits in Cats: Dr. Leslie Lyons Implications for , Breed Management and Human Health

3:35-4:15 Hereditary Gastric Cancer in Dogs Dr. Elizabeth McNiel

3 The sources of genetic structure within breeds and its implications Pam Wiener, Ph.D. The Roslin Institute, R(D)SVS, University of Edinburgh, Division of Genetics and Genomics [email protected]

Genetic analyses of domesticated species have proved very useful for determining relationships between breeds (Wiener et al., 2004), for illuminating the processes underlying the domestication process (Wiener & Wilkinson, 2011), and for identifying associated with specific traits (Georges, 2007). An important tool is the use of clustering-based population genetic methods, in which populations are determined based on the genetic make- up of individuals, without prior population labelling. These techniques have been applied to domesticated animal species in a number of studies and in most cases, have demonstrated good correspondence between breeds and genetically-defined populations. Use of this approach has proven to be particularly useful for identifying that do not fit the general genetic profile of a given breed, for example, cross-bred or mis-classified individuals.

Within-breed genetic differentiation In some cases, however, clustering techniques have revealed population structure below the breed level, such that separate groupings are identified within breeds. This was demonstrated in an analysis of British breeds, in which the breed showed internal genetic structure (Wilkinson et al., 2008). There appeared to be greater differentiation between the two British Saddleback clusters than between some breed pairs (Figure 1). A similar finding was found for several British breeds (Wilkinson et al., 2011), in which within-breed differentiation was associated with different morphological types for some breeds and with different flocks in others (Figure 2). The latter pattern indicates restricted flow between breeders, which can lead to high rates of inbreeding.

Figure 2. Individual assignment based on clustering analysis at K=35. Histograms demonstrate the proportion of each individual’s genome that originated from each of 24 populations. Each individual is represented by a vertical line corresponding to its membership coefficient (q). Genetic structure is seen within breeds such as Araucana, Leghorn , eMaran, Silki and Sussex. Figure 1. A neighbour‐joining tree of British constructed Reproduced from Wilkinson et al. (2011). from allele‐sharing distances among all individuals. Bootstrap values greater than 500 are shown (out of 1000). British Saddleback individuals are found in two separate clusters. Reproduced from Wilkinson et al. (2008).

4 Several recent studies in dogs have also identified within-breed differentiation, which derives from several sources. Quignon et al. (2007) analysed American and European samples from four breeds and demonstrated a clear genetic separation of US and EU Golden retrievers. They also identified genetic differentiation within Bernese mountain dogs, but it was not clearly associated with geographical origin. Two other breeds in that study (Flat coated retrievers and Rottweilers) did not show evidence of genetic structure. In other cases, genetic differentiation is associated with phenotypic traits. Bjornfeldt et al. (2008) identified strong genetic differentiation in poodles due to size and coat colour. Standard poodles were clearly genetically distinct from all other poodles, while the smaller poodles were differentiated from each other based on a combination of size and coat colour. A study on Schnauzer breeds revealed a similar pattern of differentiation (Streitberger et al. 2011); the authors found that Giant Schnauzers were strongly differentiated from the other Schnauzer breeds, while the smaller Schnauzers clustered based on both coat colour and size. Mellanby et al. (2013) also demonstrated genetic structure within UK Cavalier King Charles spaniels, although the source of the differentiation was not clear. Preliminary analysis of UK Labrador retrievers indicates within-breed genetic differentiation related to the role of dogs (i.e. working gun dogs versus ) as well as phenotypic characteristics (unpublished results).

Implications for managing recessive diseases: Strong population structure may lead to high levels of inbreeding by creating partially independent sub-populations with relatively small effective population sizes, increasing the role of genetic drift. This can thereby increase the overall levels of homozygosity and thus, may also increase the numbers of individuals homozygous for recessive disease alleles. Management practices that increase mixing within the breed will reduce overall levels of inbreeding and therefore, may help reduce the levels of such diseases. Somewhat ironically, in rare breeds, management strategies that involve reduced breeding from a segment of the breed that carries known disease-associated variants may exacerbate the problem at other loci by reducing the effective population size (Collins et al., 2011) and thus these strategies must be designed with care and forethought.

Implications for genetic association studies and genetic evaluation: It is well established that the existence of genetic structure can lead to spurious associations in genome-wide association studies if the trait of interest is not evenly distributed with respect to genetic sub- groups (Lander & Schork, 1994; Price et al., 2006). Therefore, it is recommended that in such cases, stratification should be accounted for (Price et al., 2010). Population structure may also influence the implementation of genomic evaluation schemes, in which breeding decisions are based on genomic marker information; however, the implications of such structure are less clear in this case. For example, Daetwyler et al. (2012) conclude that the accuracy of prediction may be reduced by accounting for population stratification in some situations (e.g. low or medium density markers). Further study is required on this issue.

References Björnfeldt, S., F. Hailer, M. Nord & C. Vilà. (2008). Assortative mating and fragmentation within dog breeds. BMC Evolutionary Biology 8:28. Collins, L.M., L. Asher, J. Summers & P. McGreevy. (20110). Getting priorities straight: Risk assessment and decision-making in the improvement of inherited disorders in pedigree dogs. The Veterinary Journal 189: 147–154. Daetwyler, H.D., K.E. Kemper, J.J.J. van der Werf & B.J. Hayes. (2012). Components of the accuracy of genomic prediction in a multi-breed population. Journal of Animal Science 90: 3375- 3384. Georges, M. (2007). Mapping, fine mapping, and molecular dissection of quantitative trait loci in domestic animals. Annual Review of Genomics and Human Genetics 8: 131-162.

5 Lander, E.S. & N.J. Schork. (1994). Genetic dissection of complex traits. Science 265: 2037-2048. Mellanby, R.J., R. Ogden, D.N. Clements, A.T. French, A.G. Gow, et al. (2013). Population structure and genetic heterogeneity in popular dog breeds in the UK. The Veterinary Journal 196: 92-97. Quignon, P., L. Herbin, E. Cadieu, E.F. Kirkness, B. Hédan, et al. (2007). Canine population structure: assessment and impact of intra-breed stratification on SNP-based association studies. PLOS one 12: e1324. Price, A.L., N.J. Patterson, R.M. Plenge, M.E. Weinblatt, N.A. Shadick, et al. (2006). Principal components analysis corrects for stratification in genome-wide association studies. Nature Genetics 38: 904-909. Price, A.L., N.A. Zaitlen, D. Reich & N.J. Patterson. (2010). New approaches to population stratification in genome-wide association studies. Nature Reviews Genetics 11: 459-463. Streitberger, K., M. Schweizer, R. Kropatsch, G. Dekomien, O. Distl, et al. (2011). Rapic genetic diversification within dog breeds as evidenced by a case study on Schnauzers. Animal Genetics 43: 577-586. Wiener, P., D. Burton and J.L. Williams. (2004). Breed relationships and definition of British : a genetic analysis. Heredity 93: 597-602. Wiener, P. & S. Wilkinson (2011). Deciphering the genetic basis of animal domestication. Proceedings of the Royal Society B: 278: 3161-3170. Wilkinson, S., C.S. Haley, L. Alderson & P. Wiener. (2008). An empirical assessment of individual- based population genetic statistical techniques: application to British pig breeds. Heredity 106: 261-269. Wilkinson, S., P. Wiener, D. Teverson, C.S. Haley & P.M. Hocking. (2011). Characterization of the , structure and admixture of British chicken breeds. Animal Genetics 43: 552- 563.

6 T h e new journal o f medicine

review article

franklin h. epstein lecture Franklin H. Epstein, M.D., served the New England Journal of Medicine for more than 20 years. A keen clinician, accomplished researcher, and outstanding teacher, Dr. Epstein was Chair and Professor of Medicine at Beth Israel Deaconess Medical Center, Boston, where the Franklin H. Epstein, M.D., Memorial Lectureship in Mechanisms of Disease has been established in his memory. Both Ends of the Leash — The Human Links to Good Dogs with Bad Genes

Elaine A. Ostrander, Ph.D.

From the National Human Genome Re- or nearly 350 years, veterinary medicine and human medicine search Institute, National Institutes of have been separate entities, with one geared toward the diagnosis and treat- Health, Bethesda, MD. Address reprint requests to Dr. Ostrander at the National Fment in animals and the other toward parallel goals in the owners. However, Human Genome Research Institute, Na- that model no longer fits, since research on diseases of humans and companion tional Institutes of Health, 50 South Dr., animals has coalesced.1-4 The catalyst for this union has been the completion of the Bldg. 50, Rm. 5351, Bethesda, MD 20892, or at [email protected]. human genome sequence, coupled with draft sequence assemblies of genomes for companion animals.5,6 Here, we summarize the critical events in canine genetics and N Engl J Med 2012;367:636-46. DOI: 10.1056/NEJMra1204453 genomics that have led to this development, review major applications in canine Copyright © 2012 Medical Society. health that will be of interest to human caregivers, and discuss expectations for the future.

Human and Canine Genomics

In 2001, two independent draft versions of the human genome sequence and the con- comitant identification of approximately 30,000 genes were the seminal events that defined completion of the Human Genome Project.7,8 The genome was officially de- clared to be finished in 2004, with sequencing reported to include 99% of transcribing DNA.9 By comparison, the genome of the domestic dog, Canis lupus familiaris, was se- quenced twice, once to 1.5× density (i.e., covering the genome, in theory, 1.5 times) and once to 7.8× density (providing sequencing for more than 95% of base pairs) in the standard poodle and boxer, respectively.5,10 Subsequent contributions to the canine genome have focused on better annotation to locate missing genes,11 understanding structure,12 studying linkage disequilibrium,5,13 identifying copy-number variants,14-16 and mapping the transcriptome.17 The use of the canine genome to understand the genetic underpinning of dis- orders that are difficult to disentangle in humans has been on the rise for nearly two decades.1,2,18 The reason relates back to the domestication of dogs from gray (C. lupus), an event that began at least 30,000 years ago.19-21 Since their domestication, dogs have undergone continual artificial selection at varying levels of intensity, leading to the development of isolated populations or breeds5,22,23 (Fig. 1). Many breeds were developed during Victorian times24 and have been in existence for only a few hundred years, a drop in the evolutionary bucket.25 Most breeds are descended from small numbers of founders and feature so-called popular sires (dogs that have performed well at dog shows and therefore sire a large number of litters). Thus, the genetic character of such founders is overrepresented in the population.25,26 These facts, coupled with breeding programs that exert strong selection for particular

636 n engl j med 367;7 nejm.org august 16, 2012

The New England Journal7 of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. franklin h. epstein lecture

A B C

E F

D

G H

I J K

Figure 1. The Diversity of Dog Breeds. Breeds vary according to many traits, including size, leg length, pelage (coat), color, and skull shape. Shown are borzoi (Panel A), basset hound (Panel B), Chihuahua (Panel C), giant schnauzer (Panel D), bichon frise (Panel E), collie (Panel F), French bulldog (Panel G), dachshund (Panel H), German shorthaired pointer (Panel I), papillon (Panel J), and Neapolitan mastiff (Panel K). (Images courtesy of Mary Bloom, American Kennel Club.) physical traits, mean that recessive diseases are 22,27,28 The Genetic Power of Canine common in purebred dogs, and many breeds Families are at increased risk for specific disorders.2,29 We, and others, have chosen to take advantage of this One of the most striking features of canine fam- fact in order to identify genes of interest for hu- ilies is their large size, which makes them ame- man and canine health. nable to conventional linkage mapping. This fact

n engl j med 367;7 nejm.org august 16, 2012 637

The New England Journal8 of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

was particularly well illustrated in the search for disorder of American Staffordshire terriers with the canine gene for hereditary multifocal renal symptoms that are similar to a human adult- cystadenocarcinoma and nodular dermatofibro- onset form of the disorder known as Kuf’s disease. sis (RCND) in German shepherds.30 Although rare, In American Staffordshire terriers, neuronal ceroid RCND is a naturally occurring inherited cancer lipofuscinosis is caused by an R99H in syndrome that includes bilateral, multifocal tu- exon 2 of the gene encoding arylsulfatase G mors in kidneys and numerous, dense collagen- (ARSG), leading to a 75% decrease in sulfatase based nodules in the skin,31 a disorder that is activity. This study, therefore, both identified a new similar to the Birt–Hogg–Dubé syndrome (BHD) gene for consideration in human neuronal ceroid in humans.32 In dogs, the disease allele is highly lipofuscinosis and provided new information re- penetrant and transmitted in an autosomal dom- garding sulfatase deficiency and pathogenesis of inant fashion. The dog pedigree that was used for the disease. mapping the disease included one affected found- er male who sired several litters (Fig. 2). With DNA Breed Structure and Genetic available from nearly all dogs, this single pedigree Complexity Simplified had sufficient power to localize the disease gene to canine chromosome 5q12 with a logarithm of A recurring theme in the gene mapping of canine odds (LOD) score of 4.6, giving odds of more diseases is the power of the breed structure (Fig. than 10,000 to 1 that the mapping was correct.30 3). To be a registered member of a breed, the dog’s After the localization of RCND, the human ancestors must have been registered members as BHD locus was mapped to human chromosome well.26 In 2011, the American Kennel Club (www 17p12q11,33 which corresponds to canine chro- .akc.org) recognized 173 distinct dog breeds, with mosome 5q12. Both affected dogs and humans European clubs taking the number of established were found to carry in the same gene breeds to more than 400.24,43 encoding tumor-suppressor protein folliculin,34,35 Dog breeds offer the same advantage of reduc- which is hypothesized to interact with the energy ing locus heterogeneity that is gained by studying and nutrient-sensing signaling pathway consist- humans from geographically isolated countries ing of AMP-activated protein kinase (AMPK) and such as or Iceland.29 For any given com- mammalian target of rapamycin (mTOR).36 plex disease, a small number of genes and delete- Three issues about this example are striking. rious alleles will dominate the breed,3 much as the First, the single, large dog pedigree was collected 999del5 BRCA2 mutation does in Icelandic women and genotyped in a fraction of the time it took to with hereditary breast cancer.44 collect and characterize the many necessary hu- Epilepsy is a good example, since this disease man pedigrees. Second, BHD is associated with has been difficult to disentangle genetically in substantial variability in disease presentation in humans because of indistinct clinical phenotypes humans and may be hard to distinguish from and a high degree of locus heterogeneity. The similar disorders.37 In the case of the large ex- disease affects 5% of dogs and is reported in tended dog family, phenotyping was easy, since dozens of breeds. Remitting focal epilepsy in the every dog had the same genetic background and Lagotto Romagnolo breed45 is caused by variants the disease presentation was highly uniform. Also, in LGI2, a homologue of the human epilepsy LGI1 the dog locus was found before the human locus. gene. In contrast, miniature wire-haired dachs- Other disease genes that were first mapped in dogs hunds have a form of epilepsy reminiscent of the for which there is a close human proxy include progressive myoclonic disease known as Lafora’s narcolepsy,38 copper toxicosis,39,40 neuronal ceroid disease, which in humans is the most severe form lipofuscinosis,41 and ichthyosis,42 to name a few. of teenage-onset epilepsy. The similar disease in Each of such stories is illuminating in its own dachshunds is caused by an unusual expansion of way. In the case of narcolepsy in the Doberman a dodecamer repeat46 within the gene encoding pinscher, the identification of a mutation in the malin (EPM2B) that modulates gene expression by gene encoding hypocretin receptor 2 suggested a a factor of nearly 900. The presentation of epilepsy newly recognized pathway that is involved in the is expectedly unique in other breeds.47 Thus, one molecular biology of sleep. Another example is way to disentangle complex diseases like epilepsy canine neuronal ceroid lipofuscinosis, a late-onset is to study the disorder in different dog breeds.

638 n engl j med 367;7 nejm.org august 16, 2012

The New England Journal9 of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. franklin h. epstein lecture NT 2 2 NT 2 3 + 4 4 − 1 2 3 2 6 2 NT NT 5 2 1 NT 2 2 + 4 4 3 NT 1 4 4 3 + 4 4 7 2 3 4 4 4 3 + NT 3 3 4 4 4 4 3 + NT 5 2 2 NT 2 2 + 4 4 1 2 1 4 4 3 + 4 4 3 2 3 4 4 4 3 + 1 1 1 7 2 4 4 3 + NT 1 2 1 4 4 3 + 4 4 5 2 2 NT 2 2 + 4 4 2 2 1 5 3 3 + 4 4 3 2 3 4 4 4 4 3 + 3 2 3 4 4 4 3 + 3 2 3 4 4 4 4 3 + NT 1 2 1 4 4 5 + 4 4 5 2 2 NT 2 2 + NT 4 2 2 NT 6 3 3 + 1 4 − 1 1 1 7 2 2 8 2 − 1 1 1 7 2 3 − 1 1 7 2 2 6 3 NT NT NT 3 2 2 4 2 3 + 4 4 1 2 2 NT 2 2 + 4 4 3 2 3 4 4 4 6 2 + 1 2 1 5 3 3 − 1 4 3 2 3 4 4 4 3 + 3 2 3 4 4 4 4 3 + 3 2 2 NT 4 5 + 4 4 NT 47 48 49 50 51 52 53 − 1 1 1 7 2 6 2 1 2 2 NT 2 3 + 4 4 NT 1 2 1 4 4 3 + 4 4 3 2 3 4 4 4 3 + 3 2 2 NT 2 3 + 4 4 − 1 3 2 3 4 2 6 2 NT 1 7 3 4 4 4 3 + 2 2 2 NT 2 2 + 4 1 NT NT 2 2 NT 2 2 + 4 4 5 2 1 NT 3 3 + 4 4 2 2 1 4 4 3 + 4 4 2 4 2 3 + 4 4 1 2 54 55 56 57 58 59 60 − 1 1 3 2 2 6 2 1 1 4 4 4 3 + NT − 45 46 1 2 2 2 2 4 4 NT NT 3 2 3 4 4 4 4 3 + 3 2 3 4 4 4 6 2 + NT

2 2 2 NT 1 5 + 4 1 1 2 2 4 2 3 + 4 4 NT 2 1 5 3 3 + 1 4 RCND 1 1 1 7 4 4 3 + GLUT4 44 C02608 FH2140 FH2594 FH2383 C05.771 AHT141 NT 26 27 28 29 30 31 3 2 3 4 4 4 4 3 + ZuBeCa6 1 1 2 4 4 4 3 + NT NT 2 2 NT 2 2 + 3 1

3 2 2 4 2 3/5 + 4 4 2 2 1 5 3 3 + 1 4 − 2 3 4 4 4 3 1 1 2 4 4 4 4 + RCND 1 1 2 4 4 3 4 4 NT NT + GLUT4 C02608 FH2140 FH2594 FH2383 C05.771 AHT141 ZuBeCa6 2 2 2 NT 2 2 + 4 1 2 2 2 5 2 2 + 3 1 1 7 3 4 4 4 3 + 1 1 1 2 2 5 5 4 + NT 2 2 2 NT 1 5 + 4 1 4 2 1 5 3 3 + 1 4 1 1 1 2 2 2 2 4 + − 3 2 3 4 4 6 2 NT 5 1/2 1 3 1 2 + 6 2 4 2 1 5 3 3 + 1 4 2 2 1 NT 2 2 + 3 1 393837363534333218171615141312 393837363534333218171615141312 3 3 4 4 4 6 3 + 1 7 3 4 4 4 4 3 + 3 2 3 4 4 6 2 + NT 7 1 1 2 1 2 − 8 2 3 1/2 1 3 1 1 + 6 2 7 1 1 2 1 2 − 6 2 4 2 1 5 2 2 + 1 4 − 1 1 7 2 2 6 2 2 2 2 NT 2 2 + 4 1 1 1 1 4 2 3 5 6 + 3 2 3 4 4 4 6 2 + 3 2 3 4 4 4 4 3 + 1 1 1 7 4 4 3 + 5 1/2 1 3 1 2 + 5 5

NT 3 3 4 4 4 4 3 + 4 2 1 5 3 3 + 4 1 6 2 1 6 4 6 + 4 6 RCND 2 2 2 5 2 2 + 3 1 GLUT4 C02608 FH2140 FH2594 FH2383 C05.771 AHT141 − 1 1 1 7 2 2 6 2 ZuBeCa6 3 2 1 3 1 2 + 6 2 25 61 62 63 64 65 66 67 − 3 2 3 4 4 4 4 3 3 2 3 4 4 4 4 3 + 3 2 3 4 4 4 6 2 +

+ 1/1 1/1 1/2 3/3 2/1 3/5 5/8 5/2 4 2 1 5 3 3 + 4 1 2 2 2 5 2 2 − 3 1 40 41 42 43 3 1/2 1 3 1 2 + 6 2 RCND − 1 1 1 7 2 2 6 2 GLUT4 C02608 FH2140 FH2594 FH2383 C05.771 AHT141 ZuBeCa6 − 1 1 1 7 2 2 6 2 − 7 3 4 4 4 4 3 with the permission of the publisher. ? 30 1 2 2 1 2 2 + 4 1 7 1 1 2 1 2 − 8 2 3 2 1 3 1 1 + 5 5 2 2 2 5 1 5 + 4 1 3 2 3 4 4 4 6 2 + − 1 3 2 3 2 2 6 2 23 24 1 6 2 6 4 5 4 6 + − 1 1 1 7 2 2 6 2 5 1 1 3 1 2 − 6 2 4 2 1 5 3 3 + 1 4 1 1 1 7 2 4 4 3 + 2 2 2 NT 1 5 + 4 1 − 1 2 8 2 1 1 7 2 − 2 1 1 1 7 2 6 5 NT 1 3 1 2 + 6 2 NT 1 2 2 1 2 2 + 4 1 3 3 4 4 4 4 3 + NT 3 2 3 4 4 4 4 3 2 2 1 6 2 2 + 3 1 + 3 2 1 3 1 1 + 6 2 − 3 2 3 4 4 4 6 2 − 1 2 2 1 2 2 + 4 1 3 2 3 4 4 2 6 2 − 1 1 1 7 2 2 8 2 2 2 1 NT 1 2 + 3 1 5 1/2 1 3 1 2 + 5 5 − 1 1 1 7 2 4 3 − 1 3 3 4 2 6 2 NT 1 2 2 1 2 2 + 4 4 3 2 1 3 1 1 + 6 2 − 1 3 2 3 2 2 6 2 2 2 1 NT 1 5 + 4 1 − 1 1 3 2 2 2 6 2 − 1 1 1 7 2 6 2 4 2 NT 5 3 3 + 1 4 NT NT 1/2 1 3 1 2 + 6 2 3 2 4 4 4 4 3 + 3 4 4 4 4 3 + NT NT RCND GLUT4 C02608 FH2140 FH2383 FH2594 C05.771 AHT141 3 1/2 1 3 1 2 + 6 5 1 2 2 1 2 2 + 4 1 Not typed Phase unknown Mutant allele Wild-type allele ZuBeCa6 Unaffected Affected Diagnosis unknown 01 02 03 04 05 06 07 / ? − + − 1 1 7 2 2 8 3 3 2 3 4 4 4 6 2 + NT

???? 08 09 10 11 19 20 21 22 3 1/2 1 3 1 2 + 6 2 RCND GLUT4 1 7 2 4 4 4 3 C02608 + FH2140 FH2594 FH2383 C05.771 AHT141 ZuBeCa6 Mapping Pedigree for Canine Renal Cystadenocarcinoma and Nodular Dermatofibrosis (RCND).

? RCND GLUT4 C02608 FH2140 FH2594 FH2383 C05.771 AHT141 ZuBeCa6 Figure 2. A single affected male dog carrying an autosomal dominant allele for RCND sired five litters of pups with five and uniqueunaffected and unaffected dogs in females.white. Squares Affected indicate males, dogscircles are females,shown in black,and lines relationships. The portion of canine belowchromosome showingeach 5q14 squarelinkage or is circle.indicated Black as a barsrectangle indicate the portion of the affected parental chromosome inherited by portioneach offspring inherited from fromthe theaffected normal chromosome father,of theand whitefather. Allelesbars indicate for theeach marker are indicated as numbers. Breakpointsadjacent allow to themarker diseaseZuBeCa6. gene to Reprintedbe localized from to Jónasdóttira region et al.,

n engl j med 367;7 nejm.org august 16, 2012 639

The New England 10Journal of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine 6 3 2 4 12 5 7 1 8 10 11 9 Coyote Wolves Spaniels Retrievers Gun Dogs Toy Breeds Toy Breeds Small Terriers Sight Hounds Herding Dogs Scent Hounds Working Dogs Mastiff-like Dogs Ancient and Spitz Dogs B Dogs Working Dogs Scent Mastiff-like

Hounds

Toy poodle Toy Doberman pinscher Doberman Standard poodle Standard Giant schnauzer

Standard schnauzer Havanese dog

Portuguese water dog

Dachshund

Petit basset griffon vendeen griffon basset Petit Bloodhound Boston terrier

Boxer

Spaniels Beagle Bulldog

Basset hound Basset French bulldog German shorthaired pointer shorthaired German Miniature bull terrier

Brittany spaniel Brittany Staffordshire bull terrier Irish water spaniel water Irish Cavalier King Charles spaniel Charles King Cavalier Glen of Imaal

Bull mastiff

English springer spaniel springer English Mastiff

English cocker spaniel cocker English Jack Russell

American cocker spaniel cocker American Briard Miniature pinscher Miniature Australian terrier

Papillon Yorkshire terrier

Pug Cairn terrier

Brussels griffon Brussels West highland terrier

Small Shih-tzu Terriers

Pekingese Norwich terrier Toy Chihuahua

Dogs Bernese mountain dog

Pomeranian Saint Bernard Samoyed Great Dane Rottweiler

American EskimoSaluki dog Flat-coated retriever Flat-coated

Afghan hound retriever Golden

Siberian husky retriever Labrador

Newfoundland Shar-pei Collie Alaskan malamuteDingo Retrievers

Akita shepherd Australian

Chow-chow

Ancient and

Spitz Breeds Basenji collie Border

Shetland sheep dog sheep Shetland

Cardigan corgi Cardigan

Near East Spain Italy Borzoi Middle East corgi Pembroke

China

Balkans,

Coyote Whippet Kuvasz Greyhound

Eastern, and Old English sheep dog sheep English Old

thern Europe wolfhound Irish

Scottish deerhound Scottish Ibizan hound Ibizan

Nor Italian greyhound Italian Wolves Dogs Herding Sight Hounds A

640 n engl j med 367;7 nejm.org august 16, 2012

The New England 11Journal of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. franklin h. epstein lecture

causative genes. Although progressive rod–cone Figure 3 (facing page). Neighbor-Joining Tree of Domestic Dogs. degeneration was initially mapped in miniature On average, 10 to 12 dogs were genotyped for each of and toy poodles, the disorder appears in more approximately 80 breeds. Trees were constructed with than a dozen breeds and is phenotypically simi- the use of data from each genotyped dog individually lar to one form of human adult-onset, autosomal or by grouping the data from each member of a breed recessive retinitis pigmentosa. Analysis of addi- together, so each breed is represented as a single data tional SNPs allowed the investigators to reduce entry. Data were also analyzed in two ways: by consid- ering adjacent 10 single-nucleotide-polymorphism the disease locus to a 106-kb haplotype that is (SNP) windows or haplotypes or by considering each shared by affected dogs from 14 breeds. A muta- SNP alone. The two analytic methods provided similar tion in a novel gene was ultimately determined to results. Panel A shows the relationships among the cause the disease.50 Had there not been 14 af- various dog breeds. The color groupings indicate fected breeds sharing the founder mutation, which breeds that probably share common founders. Panel B shows the historical relationship of the breeds with the allowed the haplotype to be significantly reduced, same color coding used in Panel A. In each case, breeds only next-generation sequencing could have ulti- that share either common behaviors or morphologic mately localized the disease gene. traits are grouped together on the basis of DNA analysis, Although researchers could have correctly indicating that they probably share common ancestors. guessed a subset of the breeds that shared the A black dot indicates at least 95% bootstrap support (a measure of the likelihood that an evolutionary split same mutation at the causative locus for progres- occurred in a given location in an evolutionary tree) sive rod–cone degeneration by knowing about after the performance of 1000 replicates. Reprinted their shared heritage, common geographic origin, 23 from vonHoldt et al. with permission of the publisher. or shared morphologic features, in many cases the relationship among the breeds is too ancient to be obvious. With the use of both cluster analy- 51,52 23 Breed Structure and sis and neighbor-joining trees, a clear pic- Reducing Regions of Linkage ture is emerging regarding how breeds are re- Disequilibrium lated to one another genetically (Fig. 3). This type of information highlights groups of breeds The second way in which breed structure offers that probably share common founders (and hence unique advantages to genetic mapping is that the same disease alleles) and facilitates experi- when used judiciously, it allows researchers to mental design. move quickly from linked or associated markers to genes. In humans, linkage disequilibrium typ- Morphologic Features ically extends on the order of kilobases, whereas and Genetic Variation within dog breeds it can extend for megabases.5,13 Long linkage disequilibrium means that although The examples discussed thus far have focused on only a modest number of single-nucleotide poly- disease phenotypes. However, canine morpho- morphisms (SNPs) are needed for an initial map- logic studies have been informative for both dis- ping study, subsequent identification of the dis- covering new ways of perturbing the genome and ease mutation can be difficult. This task is suggesting candidate genes for related diseases. facilitated by leveraging interbreed relatedness. For instance, chondrodysplasia is a fixed trait for Haplotypes in the region of interest can be com- more than 20 breeds with disproportionately pared in related breeds with the same disorder, short legs recognized by the American Kennel with the goal of identifying a segment that is Club, including the dachshund, corgi, and basset shared by all affected dogs but absent in those hound (Fig. 5).53 lacking the trait (Fig. 4). A genomewide association study comparing Among the many investigators who have dem- 95 dogs from eight chondrodysplastic breeds with onstrated this principle are Goldstein et al.,48,49 702 dogs from 64 breeds lacking the trait identi- who had previously mapped a form of canine fied a single strong association (P = 1.0×10–102) progressive retinal atrophy called progressive rod– with canine chromosome 18. Although this very cone degeneration to a 30-mb region. Progressive low P value is probably exaggerated because of retinal atrophy is analogous to human retinitis the population structure, such a strong associa- pigmentosa, for which there are many forms and tion is not unusual when breeds sharing a trait

n engl j med 367;7 nejm.org august 16, 2012 641

The New England 12Journal of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

lian genomes, turns out to be important in simi- lar human diseases. Other canine morphologic traits that include German Shepherd * such characteristics as body size, leg width, and coat color have been mapped.22,28,54-58 Not sur- prisingly, loci that control both a morphologic trait and a disease have been identified. This may Collie * be a result of strong selection by breeders to propagate dogs of a certain appearance, which results in piggybacking of disease alleles, or in some cases, diseases are associated with the same genetic variants that create a morphologic effect. Pembroke Welsh Corgi * This is best illustrated by dermoid sinus, a neural- tube defect in the ridgeback breed that is caused by the same copy-number variant that produces the hair ridge characteristic of the Rhodesian ridgeback.59 Cardigan Welsh Corgi *

Mapping Multigenic Traits

When the dog genome sequence was published Giant Schnauzer * in 2005, Lindblad-Toh et al.5 hypothesized that breed structure would enable mapping of simple recessive traits in dogs with a genomewide as- Figure 4. Comparing Haplotypes as a Method for Reducing a Region sociation study of no more than 20 cases and con- of Association for a Given Mutation. trols each. They further reasoned that complex The mutation causing a hypothetical disease is indicated by a yellow star. The various breeds with the disease are shown on the left; the chromosome traits that are controlled by, for instance, five responsible for the disease is indicated by a horizontal bar. Within each genes could be mapped with 97% certainty on breed, meiotic breakpoints are indicated by the start and finish of the blue the basis of just 100 cases and 100 controls. This bar for each breed. When all breeds are considered together, the minimal was a bold prediction, since most genomewide associated region where the mutation must lie is between the red vertical association studies of complex human disorders lines. require thousands of samples. But the investiga- tors’ prediction proved to be correct, and many genomewide association studies in dogs have suc- from a common founder are compared with a cessfully mapped complex traits on the basis of no large number of unrelated control breeds. In this more than 50,000 SNPs and fewer than 200 dogs. case, the trait is caused by expression of an fgf4 Recent work by Wilbe et al.60 that identifies retrogene. This retrogene encodes fibroblast genes for systemic lupus erythematosus (SLE)– growth factor 4 in which all fgf4 exons are pres- related disease complex illustrates this point. Nova ent, but introns and regulatory signals are miss- Scotia -tolling retrievers have an abnormally ing (Fig. 5). The spliced copy of the gene is lo- high rate of autoimmune diseases, including SLE.61 cated a large distance away from the source gene. The breed is descended from a small number of Although such an arrangement is common in founders that survived two major outbreaks of insects, this was the first report of an expressed canine distemper virus in the early 1900s.62 It retrogene that alters a mammalian trait.53 Ex- has been hypothesized that autoimmune disor- pression studies showed that the fgf4 retrogene ders develop in these dogs because they have a was expressed in the long bones of 4-week-old particularly strong or reactive immune system, puppies, suggesting that mistimed expression, which helped them to survive the distemper out- incorrect RNA levels, or mislocalization of the breaks. In an analysis of 81 cases and 57 con- retrogene product caused premature closure of trols in a genomewide association study of the growth plates in the long bones of the car- 22,000 SNPs, investigators found five associated rier breeds. It will be interesting to see whether loci, three of which have already been validat- this gene, or this method of mutating mamma- ed.60 Candidate genes of particular interest in-

642 n engl j med 367;7 nejm.org august 16, 2012

The New England 13Journal of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. franklin h. epstein lecture

Figure 5. Mapping the Breed-Fixed Trait A Breeds with Risk of Chondrodysplasia of Chondrodysplasia. Panel A shows examples of breeds that are associated with chondrodysplasia, including the corgi, basset hound, and wire-haired dachshund. Panel B shows ob- served heterozygosity for breeds that are at increased risk for chondrodysplasia (red) and those that are not at increased risk (black) within the associated 34-kb region on canine chromosome 18. The x axis indicates the chromosomal position of association, and the B Observed Heterozygosity for Chondrodysplasia y axis indicates observed heterozygosity. The red and black lines indicate trends and highlight a 24-kb region with low heterozygosity in the dogs at risk for chondro- 0.4 dysplasia that is absent in dogs that are not at increased risk. Gene 1 is a pseudogene, a defective segment of 0.3 DNA that resembles a gene but cannot be transcribed, called txndc1 (similar to the gene encoding thioredoxin- related transmembrane protein 1), and gene 2 marks 0.2 the 3′ end of the gene encoding semaphorin 3C (SEMA3C). The green boxes are conserved in both sequence and context in all for which data 0.1

are available. A 5-kb insertion (red rectangle), which Observed Heterozygosity was observed only in dogs with an association with chondrodysplasia and was found between the two puta- 0.0 tive regulatory elements, contains an fgf4 retrogene. 23281978 23422559 23446056 23622780 LINE denotes long interspersed nuclear element, and Position on Chromosome 18 SINE short interspersed nuclear element. Panel C shows expression studies indicating that the fgf4 retrogene 23425000 23430000 23435000 23440000 23445000 is expressed in articular cartilage from the distal and Gene 1 Gene 2 proximal humerus isolated from a 4-week-old dog with Insert chondrodysplasia. The retrogene and source gene are Putative regulatory region Putative regulatory region distinguished by a single-nucleotide polymorphism, SINEs which is cut by restriction enzyme BsrB1 in comple- LINEs mentary DNA (cDNA) produced from the source gene, resulting in two bands on a 2% agarose gel, but uncut C Expression of Retrogene in the cDNA from the retrogene that is present in dogs Chondrodysplasia with chondrodysplasia, resulting in only one band. MW cDNA denotes molecular weight marker. The source of con- Retro- FGF4 gene 1 2 3 4 MW +Control trol material is DNA isolated from the testes of a dog 700 — 51 600 — with chondrodysplasia. Modified from Parker et al., 500 — — A with the permission of the publisher. 400 — 300 — — — G 200 — clude those associated with T-cell activation such as PPP3CA, BANK1, and DAPPI. 100 — No Chondrodysplasia Dogs and Cancer cDNA cDNA MW FGF4 5 6 FGF4 7 8MW +Control 700 — 600 — Of all the disorders for which dogs are likely to 500 — inform human health, canine cancer is likely to 400 — — A have the greatest effect.63 Cancers are the most 300 — — frequent cause of disease-associated death in 200 — — G dogs, and naturally occurring cancers are well 100 — described in several breeds.3,64,65 Although con- siderable effort has gone into the study of com- mon cancers, the dog has also served as a model for studies of rare tumors, including histiocytic exist: a localized variant, in which skin and sub- sarcomas, which are highly aggressive, lethal, cutical tumors develop in a leg and metastasize dendritic-cell neoplasms.66 In dogs, two forms to lymph nodes and blood vessels, and a dissem-

n engl j med 367;7 nejm.org august 16, 2012 643

The New England 14Journal of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

inated multisystem form, in which tumors affect of dogs and humans that is caused by a loss of the spleen, liver, and lungs.67 Histiocytic sarcomas the RPE65 protein owing to mutations in RPE65, will develop in approximately 20% of Bernese causing blindness shortly after birth. In a land- mountain dogs,68 and the condition is invariably mark study in 2001, Acland et al.76 used a recom- fatal.69 In humans, similar disorders such as binant adeno-associated virus carrying wild-type Langerhans’-cell histiocytosis have been well RPE65 to restore vision in a dog that was homo- characterized clinically, but the underlying cause zygous for the RPE65 mutation. Replication was is unknown.70 successful,77 and treated dogs maintained stable Recently, a genomewide association study for vision for at least 3 years.78 Humans with Leber’s histiocytic sarcoma was undertaken in dogs.71 congenital amaurosis are now being successfully Because the disorder occurs in so few breeds, treated for the disorder.79,80 Progressive retinal Bernese mountain dogs from France, the United atrophy occurs in more than 100 breeds of dogs, States, and the Netherlands were included, with suggesting dozens of naturally occurring models the idea that these independently propagating for additional study. So far, 18 genes for canine lines would offer the same advantages for reduc- retinal diseases have been found.81 ing a region of association that distinct, but re- lated, dog breeds provide.72 For this breed, this Dog Genetics and Behavior assumption proved to be true, and two loci were identified, one on chromosome 18. Fine mapping The canine system is valuable for mapping behav- and sequencing narrowed the locus to a single iors that are specific to both breed82 and species.23 risk-associated haplotype that spans the MTAP Abnormal behaviors, including separation anxiety, gene and contains one or more variants that alter dominance aggression, and obsessive–compulsive the expression of the nearby INK4A–ARF–INK4B disorder, are most amenable to genetic studies.83 locus but do not affect expression of MTAP itself. Partial success has been achieved with obsessive– Although 40% of a random sample of Bernese compulsive disorder in bull terriers and Doberman mountain dogs in the United States are homozy- pinschers.84,85 In Dobermans, the disease presents gous for the disease haplotype, histiocytic sar- as flank or blanket sucking and was recently coma develops in only about 20% of these dogs. mapped to a 1.7-Mb region of chromosome 7 near However, more than 60% of Bernese mountain the CDH2 gene. CDH2 mediates synaptic activity- dogs eventually die of cancer. The disease-asso- regulated neuronal adhesion, but to date no func- ciated portion of chromosome 11 corresponds to tional studies have illuminated these findings and human chromosome 9p21, which has been asso- no mutation has been reported.85 ciated with several types of cancer.73-75 We have hypothesized that multiple distinct cancers in Summary Bernese mountain dogs may be related to vari- ants within the MTAP–CDKN2A region and the What we most wish to understand about dog associated canine locus. Thus, studies of this health is the very same thing we wish to know naturally occurring dog model not only illumi- about ourselves. When will we, or they, get sick? nate a causative locus but also suggest a biologic How is the illness best treated? And what is the model for the study of germline variation in this likely outcome? Each half of a –human pair important cancer-susceptibility locus. wants to know what to expect from the other end of the leash and how to prolong the relationship. Dog Breeds and Gene Therapy Finally, as the end of life approaches, we seek to make both our canine companions and ourselves Although I have focused largely on the role of comfortable, settled in the knowledge that a full dogs in the identification of genes that are asso- life has been achieved. When considered in that ciated with disease, dogs have also served an im- frame, we are not so different from our canine portant role in the development of treatments. companions. As the scientific advances coalesce, One form of progressive retinal atrophy called joining us ever closer to the one family member Leber’s congenital amaurosis type 2 is a disease we actually get to choose, it is worth bearing in

644 n engl j med 367;7 nejm.org august 16, 2012

The New England 15Journal of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. franklin h. epstein lecture mind that though our methods may be different, Disclosure forms provided by the author are available with the our goals are the same: a healthy life well spent full text of this article at NEJM.org. in the best of company.

References

1. Tsai KL, Clark LA, Murphy KE. Un- dogs. Genome Res 2009;19:500-9. [Erra- Lingaas F. Renal microcystic tubular le- derstanding hereditary diseases using the tum, Genome Res 2009;19:690.] sions in two 1-year-old-dogs — an early dog and human as companion model sys- 16. Nicholas TJ, Cheng Z, Ventura M, sign of hereditary renal cystadenocarci- tems. Mamm Genome 2007;18:444-51. Mealey K, Eichler EE, Akey JM. The ge- noma? J Comp Pathol 2000;123:218-21. 2. Karlsson EK, Lindblad-Toh K. Leader nomic architecture of segmental duplica- 32. Birt AR, Hogg GR, Dubé WJ. Heredi- of the pack: gene mapping in dogs and tions and associated copy number vari- tary multiple fibrofolliculomas with other model organisms. Nat Rev Genet ants in dogs. Genome Res 2009;19:491-9. trichodiscomas and acrochordons. Arch 2008;9:713-25. 17. Briggs J, Paoloni M, Chen Q-R, Wen X, Dermatol 1977;113:1674-7. 3. Shearin AL, Ostrander EA. Leading Khan J, Khanna C. A compendium of ca- 33. Schmidt LS, Warren MB, Nickerson the way: canine models of genomics and nine normal tissue gene expression. PLoS ML, et al. Birt-Hogg-Dubé syndrome, a disease. Dis Model Mech 2010;3:27-34. One 2011;6(5):e17107. genodermatosis associated with sponta- 4. Lequarré A-S, Andersson L, André C, 18. Ostrander EA, Giniger E. Semper fi- neous pneumothorax and kidney neopla- et al. LUPA: a European initiative taking delis: what man’s best friend can teach us sia, maps to chromosome 17p11.2. Am J advantage of the canine genome architec- about human biology and disease. Am J Hum Genet 2001;69:876-82. ture for unravelling complex disorders in Hum Genet 1997;61:475-80. 34. Nickerson ML, Warren MB, Toro JR, both human and dogs. Vet J 2011;189:155- 19. Vilà C, Savolainen P, Maldonado JE, et et al. Mutations in a novel gene lead to 9. al. Multiple and ancient origins of the do- kidney tumors, lung wall defects, and be- 5. Lindblad-Toh K, Wade CM, Mikkelsen mestic dog. Science 1997;276:1687-9. nign tumors of the hair follicle in patients TS, et al. Genome sequence, comparative 20. Savolainen P, Zhang YP, Luo J, Lunde- with the Birt-Hogg-Dubé syndrome. Can- analysis and haplotype structure of the berg J, Leitner T. Genetic evidence for an cer Cell 2002;2:157-64. domestic dog. Nature 2005;438:803-19. East Asian origin of domestic dogs. Sci- 35. Lingaas F, Comstock KE, Kirkness EF, 6. Pontius JU, Mullikin JC, Smith DR, et ence 2002;298:1610-3. et al. A mutation in the canine BHD gene al. Initial sequence and comparative anal- 21. Sablin MV, Khlopachev GA. The earli- is associated with hereditary multifocal ysis of the cat genome. Genome Res 2007; est Ice Age dogs: evidence from Eliseev- renal cystadenocarcinoma and nodular 17:1675-89. ichi 1. Curr Anthropol 2002;43:795-9. dermatofibrosis in the German Shepherd 7. Lander ES, Linton LM, Birren B, et al. 22. Boyko AR, Quignon P, Li L, et al. A dog. Hum Mol Genet 2003;12:3043-53. Initial sequencing and analysis of the hu- simple genetic architecture underlies mor- 36. Hasumi Y, Baba M, Ajima R, et al. Ho- man genome. Nature 2001;409:860-921. phological variation in dogs. PLoS Biol mozygous loss of BHD causes early em- [Erratum, Nature 2001;412:565.] 2010;8(8):e1000451. bryonic lethality and kidney tumor devel- 8. Venter JC, Adams MD, Myers EW, et 23. vonHoldt BM, Pollinger JP, Lohmueller opment with activation of mTORC1 and al. The sequence of the human genome. KE, et al. Genome-wide SNP and haplo- mTORC2. Proc Natl Acad Sci U S A 2009; Science 2001;291:1304-51. [Erratum, Sci- type analyses reveal a rich history under- 106:18722-7. ence 2001;292:1838.] lying dog domestication. Nature 2010; 37. Smit DL, Mensenkamp AR, Badeloe S, 9. International Human Genome Se- 464:898-902. et al. Hereditary leiomyomatosis and re- quencing Consortium. Finishing the eu- 24. Fogle B. Encyclopedia of the dog. nal cell cancer in families referred for chromatic sequence of the human ge- London: DK Publishing, 1995. fumarate hydratase germline mutation nome. Nature 2004;431:931-45. 25. Parker HG. Genomic analyses of analysis. Clin Genet 2011;79:49-59. 10. Kirkness EF, Bafna V, Halpern AL, et modern dog breeds. Mamm Genome 38. Lin L, Faraco J, Li R, et al. The sleep al. The dog genome: survey sequencing 2012;23:19-27. disorder canine narcolepsy is caused by a and comparative analysis. Science 2003; 26. American Kennel Club. The complete mutation in the hypocretin (orexin) re- 301:1898-903. dog book. 19th ed., rev. New York: Howell ceptor 2 gene. Cell 1999;98:365-76. 11. Derrien T, Vaysse A, André C, Hitte C. Book House, 1998. 39. Yuzbasiyan-Gurkan V, Blanton SH, Annotation of the domestic dog genome 27. Patterson DF. Companion animal Cao V, et al. Linkage of a microsatellite sequence: finding the missing genes. medicine in the age of . marker to the canine copper toxicosis lo- Mamm Genome 2012;23:124-31. J Vet Intern Med 2000;14:1-9. cus in Bedlington terriers. Am J Vet Res 12. Breen M. Canine cytogenetics — from 28. Vaysse A, Ratnakumar A, Derrien T, 1997;58:23-7. band to basepair. Cytogenet Genome Res et al. Identification of genomic regions 40. van De Sluis B, Rothuizen J, Pearson 2008;120:50-60. associated with phenotypic variation be- PL, van Oost BA, Wijmenga C. Identifica- 13. Sutter NB, Eberle MA, Parker HG, et tween dog breeds using selection map- tion of a new copper gene by al. Extensive and breed-specific linkage ping. PLoS Genet 2011;7(10):e1002316. positional cloning in a purebred dog pop- disequilibrium in Canis familiaris. Ge- 29. Ostrander EA, Kruglyak L. Unleashing ulation. Hum Mol Genet 2002;11:165-73. nome Res 2004;14:2388-96. the canine genome. Genome Res 2000;10: 41. Abitbol M, Thibaud JL, Olby NJ, et al. 14. Nicholas TJ, Baker C, Eichler E, Akey 1271-4. A canine Arylsulfatase G (ARSG) muta- JM. A high-resolution integrated map of 30. Jónasdóttir TJ, Mellersh CS, Moe L, et tion leading to a sulfatase deficiency is copy number polymorphisms within and al. Genetic mapping of a naturally occur- associated with neuronal ceroid lipofusci- between breeds of the modern domesti- ring hereditary renal cancer syndrome in nosis. Proc Natl Acad Sci U S A 2010; cated dog. BMC Genomics 2011;12:414. dogs. Proc Natl Acad Sci U S A 2000; 107:14775-80. 15. Chen WK, Swartz JD, Rush LJ, Alvarez 97:4132-7. 42. Grall A, Guaguère E, Planchais S, et CE. Mapping DNA structural variation in 31. Moe L, Gamlem H, Jónasdóttir TJ, al. PNPLA1 mutations cause autosomal

n engl j med 367;7 nejm.org august 16, 2012 645

The New England 16Journal of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. franklin h. epstein lecture

recessive congenital ichthyosis in golden Spring Harb Symp Quant Biol 2009;74:327- Canine population structure: assessment retriever dogs and humans. Nat Genet 33. and impact of intra-breed stratification 2012;44:140-7. 57. Karlsson EK, Baranowska I, Wade on SNP-based association studies. PLoS 43. Wilcox B, Walkowicz C. The atlas of CM, et al. Efficient mapping of mendelian One 2007;2(12):e1324. dog breeds of the world. 5th ed. Neptune traits in dogs through genome-wide as- 73. Debniak T, Górski B, Huzarski T, et City, NJ: TFH Publications, 1995. sociation. Nat Genet 2007;39:1321-8. al. A common variant of CDKN2A (p16) 44. Tulinius H, Olafsdottir GH, Sigvalda- 58. Schmutz SM, Berryere TG. Genes af- predisposes to breast cancer. J Med Genet son H, et al. The effect of a single BRCA2 fecting coat colour and pattern in domes- 2005;42:763-5. mutation on cancer in Iceland. J Med tic dogs: a review. Anim Genet 2007; 74. Wrensch M, Jenkins RB, Chang JS, et Genet 2002;39:457-62. 38:539-49. al. Variants in the CDKN2B and RTEL1 45. Seppälä EH, Jokinen TS, Fukata M, et 59. Salmon-Hillbertz NHC, Isaksson M, regions are associated with high-grade al. LGI2 truncation causes a remitting fo- Karlsson EK, et al. Duplication of FGF3, glioma susceptibility. Nat Genet 2009; cal epilepsy in dogs. PLoS Genet 2011; FGF4, FGF19 and ORAOV1 causes hair 41:905-8. 7(7):e1002194. ridge and predisposition to dermoid sinus 75. Bishop DT, Demenais F, Iles MM, et 46. Lohi H, Young EJ, Fitzmaurice SN, et in Ridgeback dogs. Nat Genet 2007;39: al. Genome-wide association study identi- al. Expanded repeat in canine epilepsy. 1318-20. fies three loci associated with melanoma Science 2005;307:81. 60. Wilbe M, Jokinen P, Truvé K, et al. risk. Nat Genet 2009;41:920-5. 47. Ekenstedt KJ, Patterson EE, Mickelson Genome-wide association mapping iden- 76. Acland GM, Aguirre GD, Ray J, et al. JR. Canine epilepsy genetics. Mamm Ge- tifies multiple loci for a canine SLE-relat- Gene therapy restores vision in a canine nome 2012;23:28-39. ed disease complex. Nat Genet 2010;42: model of childhood blindness. Nat Genet 48. Acland GM, Ray K, Mellersh CS, et al. 250-4. 2001;28:92-5. Linkage analysis and comparative map- 61. Hansson-Hamlin H, Lilliehöök I. A 77. Bennicelli J, Wright JF, Komaromy A, ping of canine progressive rod-cone de- possible systemic rheumatic disorder in et al. Reversal of blindness in animal generation (prcd) establishes potential the Nova Scotia duck tolling retriever. models of Leber congenital amaurosis us- locus homology with retinitis pigmentosa Acta Vet Scand 2009;51:16. ing optimized AAV2-mediated gene trans- (RP17) in humans. Proc Natl Acad Sci 62. Strang A, MacMillan G. The Nova fer. Mol Ther 2008;16:458-65. U S A 1998;95:3048-53. Scotia duck tolling retriever. Loveland, 78. Acland GM, Aguirre GD, Bennett J, et 49. Goldstein O, Zangerl B, Pearce-Kel- CO: Alpine Publications, 1996. al. Long-term restoration of rod and cone ling S, et al. Linkage disequilibrium map- 63. Khanna C, Lindblad-Toh K, Vail D, et vision by single dose rAAV-mediated gene ping in domestic dog breeds narrows the al. The dog as a cancer model. Nat Bio- transfer to the retina in a canine model of progressive rod-cone degeneration inter- technol 2006;24:1065-6. childhood blindness. Mol Ther 2005;12: val and identifies ancestral disease-trans- 64. Bronson RT. Variation in age at death 1072-82. mitting chromosome. Genomics 2006;88: of dogs of different sexes and breeds. Am 79. Cideciyan AV. Leber congenital amau- 541-50. J Vet Res 1982;43:2057-9. rosis due to RPE65 mutations and its 50. Zangerl B, Goldstein O, Philp AR, et 65. Maquat LE. Defects in RNA splicing treatment with gene therapy. Prog Retin al. Identical mutation in a novel retinal and the consequence of shortened trans- Eye Res 2010;29:398-427. gene causes progressive rod-cone degen- lational reading frames. Am J Hum Genet 80. Kaplan J. Leber congenital amaurosis: eration in dogs and retinitis pigmentosa 1996;59:279-86. from darkness to spotlight. Ophthalmic in humans. Genomics 2006;88:551-63. 66. Moore PF, Affolter VK, Vernau W. Ca- Genet 2008;29:92-8. 51. Parker HG, Kim LV, Sutter NB, et al. nine hemophagocytic histiocytic sarco- 81. Miyadera K, Acland GM, Aguirre GD. Genetic structure of the purebred domes- ma: a proliferative disorder of CD11d+ Genetic and phenotypic variations of in- tic dog. Science 2004;304:1160-4. macrophages. Vet Pathol 2006;43:632-45. herited retinal diseases in dogs: the pow- 52. Parker HG, Kukekova AV, Akey DT, et 67. Affolter VK, Moore PF. Localized and er of within- and across-breed studies. al. Breed relationships facilitate fine map- disseminated histiocytic sarcoma of den- Mamm Genome 2012;23:40-61. ping studies: a 7.8-kb deletion cosegre- dritic cell origin in dogs. Vet Pathol 2002; 82. Spady TC, Ostrander EA. Canine be- gates with Collie eye anomaly across mul- 39:74-83. havioral genetics: pointing out the pheno- tiple dog breeds. Genome Res 2007;17: 68. Moore PF, Rosin A. Malignant histio- types and herding up the genes. Am J 1562-71. cytosis of Bernese mountain dogs. Vet Hum Genet 2008;82:10-8. 53. Parker HG, Ostrander EA. Canine ge- Pathol 1986;23:1-10. 83. Overall KL. Natural animal models of nomics and genetics: running with the 69. Fulmer AK, Mauldin GE. Canine his- human psychiatric conditions: assess- pack. PLoS Genet 2005;1(5):e58. tiocytic neoplasia: an overview. Can Vet J ment of mechanism and validity. Prog 54. Sutter NB, Bustamante CD, Chase K, 2007;48:1041-50. Neuropsychopharmacol Biol Psychiatry et al. A single IGF1 allele is a major deter- 70. Grogan TM, Pileri SA, Chan JKC, 2000;24:727-76. minant of small size in dogs. Science Weiss LM, Fletcher CDM. Histiocytic and 84. Moon-Fanelli AA, Dodman NH. De- 2007;316:112-5. [Erratum, Science 2007; dendritic cell neoplasms. 4th ed. Lyon, scription and development of compulsive 316:1284.] France: International Agency for Research tail chasing in terriers and response to 55. Cadieu E, Neff M, Quignon P, et al. on Cancer, 2008. clomipramine treatment. J Am Vet Med Coat variation in the domestic dog is gov- 71. Shearin AL, Hedan B, Cadieu E. et al. Assoc 1998;212:1252-7. erned by variants in three genes. Science The MTAP-CDKN2A locus confers sus- 85. Dodman NH, Karlsson EK, Moon- 2009;326:150-3. ceptibility to a naturally occurring canine Fanelli A, et al. A canine chromosome 7 56. Quignon P, Schoenebeck JJ, Chase K, cancer. Cancer Epidemiol Biomarkers locus confers compulsive disorder sus- et al. Fine mapping a locus controlling leg Prev 2012 May 23 (Epub ahead of print). ceptibility. Mol Psychiatry 2010;15:8-10. morphology in the domestic dog. Cold 72. Quignon P, Herbin L, Cadieu E, et al. Copyright © 2012 Massachusetts Medical Society.

646 n engl j med 367;7 nejm.org august 16, 2012

The New England 17Journal of Medicine Downloaded from nejm.org at NIH on August 22, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. Genetics of Cat Populations and Breeds: Implications for Breed Management for Health! Leslie A. Lyons, PhD, College of Veterinary Medicine, University of Missouri - Columbia [email protected]

Introduction Inclusive of feral domestic cats, the world-wide population of Felis silvestris catus is ~600 million, and represents on the order of 99% of all living individuals of the mammalian family (Dauphiné, N. I. C. O., and ROBERT J. Cooper 2009). In the United States 38.4 million households own cats, totaling 88 million individual animals, which incur $400 in expenditures per year per household4. Cats incur veterinary health care costs in a direct fashion, but they may be important vectors for pathogens which impact human behavior, such as Toxoplasma gondii. Latent with Toxoplasma gondii has been implicated in cultural variation. Associations with mental illness and infected status have also been reported. Thus the management of cat populations is increasing in the public interest. Understanding the genetic relationships of help to manage their health care and predict and prevent unwanted genetic diseases and traits.

Wildcat origins The domestic cat, Felis catus,25,38 is one of 38 species in the cat family Felidae, being a member of the Felis lineage.52 The Felis lineage is composed of three small African felids and four small felids that may be the progenitors of the domestic cat, including Felis lybica (), Felis silvestris (European wildcat), Felis ornata (Asian wildcat), and Felis bieti (Chinese desert cat).28,32 The domestic cat and the wildcat species can interbreed, producing fertile hybrids26; thus their demarcation as subspecies and even distinct species can be disputed. Because the common housecat is a domesticated derivative, the term Felis catus has been re-adopted and does not clearly denote the genetic relationship to the progenitor wildcats or their subspecies.25 The relationship of the African, European, and two Asian wildcats is somewhat controversial; currently, 21 subspecies are defined within these groupings.34 Other than the South African subspecies of African wildcat, Felis lybica cafra, most species of wildcat and their associated subspecies may be the progenitors of domestic cat populations,21,40 Felis lybica having the most scientific support.

Domestic cats likely participated actively in their own domestication; both humans and felines developed a symbiotic, commensal, mutual tolerance. Several independent sites of early civilizations are known to have developed between 8000 and 3000 BCE, including the Huang He River region of China; the Indus Valley in Pakistan; and the Fertile Crescent region, which extends from Iraq, into Turkey, south along the Levant region of the Mediterranean coast, and, arguably, into the Nile Valley of Egypt.8 As humans made the transition from hunter–gatherers to the more sedentary lifestyle of the farmer and permanent settlements subsequently developed, villages produced refuse piles and grain stores, attracting mice and rats,11 a primary prey species for the small wildcat. To obtain these easy meals, bold wildcats perhaps began to tolerate humans, and humans accepted the cat because of its utility in vermin control.

18 Domestic Populations Random-bred and feral cats represent the overwhelming majority of cats throughout the world, not fancy cat breed populations,3 although most genetic studies have focused on cat breeds to date. Considering the worldwide distribution of cats, the United States likely has the highest proportion of pedigreed cats. However, the proportion of pedigreed versus random-bred cats is still fairly low; only 10% to 15% of feline patients at the University of California, Davis Veterinary Medicine Teaching Hospital is represented by pedigreed cats.42 A general understanding of cat breed development and a more in-depth understanding of a limited number of foundation cat breeds will help predict health care problems on the basis of each cat’s genetic background.

Genetic studies of over a thousand cats from worldwide populations have allowed the definition of approximately ten genetically distinct cat populations from around the world. These populations can be used as the foundation genetic pools for specific breeds. The first documented that judged cats on their aesthetic value occurred in London, England, at the Crystal Palace in 1871.1 This competition presented only a handful of breeds, including the British, Persian, Abyssinian, Angora, and Siamese. Thus, these early documented cat breeds likely represented genetically distinct populations insofar as strict breeding programs were not established at the time. However, now they are genetically distinct breeds, but their genetic origins can be traced to their foundation populations.

Most worldwide cat fancy associations, such as the Cat Fanciers’ Association (CFA),16,17 The International Cat Association (TICA),61 the Governing Council of the Cat Fancy (GCCF),2 and the Fédération Internationale Féline (FIFe),22 recognize approximately 35 to 41 cat breeds, although only a few breeds overwhelmingly dominate the census of the registries. Persian cats and related breeds (e.g., Exotics, a shorthaired Persian variety) are among the most popular cat breeds worldwide and represent an overwhelming majority of pedigreed cats. Although not all cats produced by breeders are registered, perhaps only 20% to 30%, the CFA, one of the largest cat registries worldwide, generally registers approximately 40,000 pedigreed cats annually.18 Approximately 16,000 to 20,000 are Persians, and approximately 3000 are Exotics; thus the Persian group of cats represents more than 50% of the cat fancy population. Common breeds that generally have at least 1000 annual registrants are Abyssinians, Maine Coons, and Siamese. Other popular breeds include the and Burmese, which are more prevalent in other areas, such as the . Most of these popular breeds also represent the oldest and most established cat breeds worldwide. However, because of different breeding standards in different registries and population substructuring, not all cats identified as the same breed are genetically alike. Disease frequencies may be different for breeds in different parts of the world. For example, polycystic kidney disease has been shown to have about the same prevalence in Persian cats around the world,5,6,10,15 but hypokalemia in the Burmese is more limited to cats in the United Kingdom and Australia9,36 and not found in populations in the United States. Some lines of Burmese in the United States segregate for a craniofacial defect, which is not commonly found in Burmese cats outside the United States.50 The breed substructuring may be partially due to control measures that reduce migration of cats among countries, but it is also likely that

19 the known health concerns in the breeds have led to strong restrictions of imports and exports of fancy-breed cats.

A more recently developed cat breed, the Bengal,31 which is a between the Asian , Prionailurus bengalensis, and the domestic cat, has gained significant popularity throughout the world, even though some registries currently do not recognize the breed. Because of limited wildcat founders, the hybrid cats may have decreased genetic variation. These hybrid cats may also have allelic incompatibilities for a given gene; the genes between the two species, leopard cat and domestic cat, have millions of years of evolutionary divergence, which allows differences at the DNA sequence level of a gene. Hence an accumulation of different genetic variants that are functional within the species, but nonfunctional across the felid species, are likely present in some Bengal cats. Thus hybrid cat breeds may have unexpected health problems and infertility, creating a challenge for both genetic studies and primary health care.

Many modern cat breeds derived from an older “foundation” breed, thereby forming breed families or groups. Approximately 22 breeds can be considered foundation or “natural” breeds. Genetic studies have also shown that the foundation breeds have either significantly different genetic pools or sufficient selection and inbreeding that created significant genetic distinction (Figure 1). Cat breeds derived from the foundation breeds are often based on single gene variants, such as longhaired and shorthaired varieties, or even a hairless variety, as found in the Rex and Sphynx grouping. Color variants also tend to demarcate breeds, such as the “pointed” variety of the Persian, known as the Himalayan by many cat enthusiasts and as a separate breed by some associations, such as TICA.61 These derived breeds are not genetically significantly different and therefore share health concerns. , , and all use Persians to help define their structure; thus these breeds also suffer from polycystic kidney disease,43 and their genetic signatures are very similar to that of Persians, nearly obscuring their original population foundations of U.S. and UK cats.

A population case study: Turkish Cats The Lyons’ Feline Genetics laboratory has a standing interest in the dynamics of cat populations and domestic cat breeds. Through interactions with cat breeders, both in the United States and abroad, and also with collaborators from Turkish universities and animal shelters, the laboratory performed three studies on the genetics of cats reportedly and documented to be from Turkey.

Round 1 - The first study was published in a scientific journal in 2007 and analyzed 14 and 21 . These cats were primarily from breeders within the United States and cats were selected to have no grand-parents in common. Contributions from as many different breeders was attempted to properly survey the and genetic structure of the Turkish Angora and Turkish Van breeds in comparison to a variety of other breed cats from the USA. Random bred cats from collaborators at Turkish universities were also analyzed. The major outcomes of the first analyses of these breeds indicated:

20 1) Cats from the Mediterranean area, including Turkey, Israel, Cairo, Egypt and Italy are genetically distinct from cats of Western Europe, Asia, and the Eastern coast of Kenya, forming four major and distinct populations (races) of cats in the world. 2) Three cat breeds appear to have their ancient origins in the Mediterranean, including Turkish Angora, Turkish Van and potentially the . 3) The Turkish Van and Turkish Angora are genetically distinct breeds. 4) The Turkish Angora had more genetic diversity and a lower inbreeding level in comparison to Turkish Vans, suggesting they are slightly more genetically healthy. 5) Both Turkish Angora and Turkish Van were at the higher end of the spectrum of inbreeding levels amongst the cats evaluated, suggesting minimal outcrossing may be warranted. 6) The genetic variation of the random bred Turkish cats was amongst the highest of all cat populations, suggesting the region was the origins of cat domestication.

Round 2 (Figure 2) – At the request of various Turkish Van breeders and because of the interest to add genetic diversity to the Van breed by using cats from Turkey, the study was extended and analyzed an additional 30 cats. These cats represented individuals supplied by several different breeders from the USA, The Netherlands, , and Turkey. Four cats were included that were listed as crosses with cats noted as Vankedisi. These cats were genetically compared to the original 21 cats of the breed diversity study. The outcomes of this second study suggested: 1) Sixteen (16) of the 30 cats were highly significant similar genetically to the Turkish Vans from the USA, suggesting these cats constitute the same breed. These cats were designated Type A Turkish Vans (Fig. 2, red in Fig. 3).The three of four cats noted as crosses with Vankedisi cats were in this grouping. 2) One cat was significantly similar to a Turkish Angora – (Type C in Fig. 2, blue in Fig. 3) 3) Thirteen (13) cats had genetics that were significantly different from Turkish Vans, potentially from three different genetic sources designated at Type B, C and D.

Round 3 – After debate and complaints that breeders did not get to fairly contribute to the second study, even though submissions were accepted for over a year, an additional 130 cats were considered that were submitted by many different breeders. The breeders were asked to prioritize cats as again. Ninety-three (93) had sufficient DNA for the analysis. In addition, random bred cats from Cyprus, which were collected from the Malcolm Cat Sanctuary, as part of a study with National Geographic, were available for comparison. A larger analysis was performed that included Turkish Angoras, random bred cats from Turkey and Cyprus, all cats submitted for the previous studies, and the new 93 cats, for a database of 248 cats. All cats were considered in one large analysis. The analysis partitioned the cats based solely on genetic variation, not by any other identification. Three major genetic groupings of cats were observed. A cut-off value of 50% similarity was used to assign a cat to a group. The groupings were then inspected to see what cats they contained. The overall summary of the Turkish cat study suggested:

21 1) Results from the previous two studies are upheld and consistent. 2) Turkish Angora is a distinct breed and with significant contribution from Turkish random bred cats. The Turkish Angora breed contains the most representative cats of Turkey. 3) Turkish Vans are a distinct breed and show significantly less influence from Turkish random bred cats. 4) Cyprus cats are a distinct population within the Mediterranean. 5) Some limited migration of cats occurs between Cyprus and Turkey. 6) Type B, C and D cats from Round 2 were cats from Cyprus. 7) The Turkish Van is genetically similar to the four cats submitted as Vankedisi.

Conclusions The analysis of cat populations supports several aspects of genetic research but importantly also the management of cat breeds. Breeds that are genetically related all share the same health concerns. These breed “families” would be starting candidates for discussions of outcrossing to increase genetic diversity (Figure 4). In addition, by knowing the populations of origin, the health concerns of the foundation populations could be at risk and need to be considered for specific diseases and visa versa. Foundation, random bred cats could be used in outcorssing programs to increase gene pools but likewise need to be monitored for unwanted genetic traits.

References 1. The Cat-Show, Penny Illustrated Paper, The Naturalist: 511, July 22:22, 1871. 2. The Governing Council of the Cat Fancy (GCCF). http://www.gccfcats.org, 2010. 22 June 2011 3. American Pet Product Manufacturing Association: National pet owner's survey, Greenwich, Conn, 2008, The Association. 4. American Veterinary Medical Association: US pet ownership and demographics sourcebook, Schaumburg, Ill, 2007, The Association. 5. Barrs VR, Gunew M: Prevalence of autosomal dominant polycystic kidney disease in Persian cats and related-breeds in Sydney and Brisbane, Aust Vet J 79:257, 2001. 6. Barthez PY, Rivier P, Begon D: Prevalence of polycystic kidney disease in Persian and Persian related cats in France, J Feline Med Surg 5:345, 2003. 8. Bellwood P: First farmers: the origins of agricultural societies, Oxford, 2005, Blackwell Publishing. 9. Blaxter A, Lievesley P, Gruffydd-Jones T et al: Periodic muscle weakness in Burmese , Vet Rec 118:619, 1986. 10. Bonazzi M, Volta A, Gnudi G et al: Prevalence of the polycystic kidney disease and renal and urinary bladder ultrasonographic abnormalities in Persian and cats in Italy, J Feline Med Surg 9:387, 2007. 11. Bonhomme F, Martin S, Thaler L: Hybridation en laboratoire de Mus musculus L. et Mus spretus lataste, Experientia 34:1140, 1978. 15. Cannon MJ, MacKay AD, Barr FJ et al: Prevalence of polycystic kidney disease in Persian cats in the United Kingdom, Vet Rec 149:409, 2001. 16. CFA: The Cat Fanciers' Association cat encyclopedia, New York, 1993, Simon & Schuster.

22 17. CFA: The Cat Fanciers' Association complete cat book, ed 1, New York, 2004, Harper Collins Publishers. 18. CFA: Cat Fanciers' Association registration totals by color and breed—2003, and 1/1/58 to 12/31/03, Cat Fanciers' Almanac 20:72, 2004. 21. Driscoll CA, Menotti-Raymond M, Roca AL et al: The Near Eastern origin of cat domestication, Science 317:519, 2007. 22. FIFe. Federation Internationale Feline. http://fifeweb.org/index.php, 2010. 22 June 2011 25. Gentry AS, Clutton-Brock J, Groves CP: The naming of wild animal species and their domestic derivatives, J Archaeol Sci 31:645, 2004. 26. Gray AP: Mammalian hybrids: a check-list with bibliography, Farnham Royal, England, 1972, Commonwealth Agricultural Bureaux. 28. Hemmer H: The evolutionary systematics of living Felidae: present status and current problems, Carnivore 1:71, 1978. 31. Johnson G: The , Greenwell Springs, La, 1991, Gogees Cattery. 32. Johnson WE, Eizirik E, Pecon-Slattery J et al: The late Miocene radiation of modern Felidae: a genetic assessment, Science 311:73, 2006. 34. Kratochvil J, Kratochvil Z: The origin of the domesticated forms of the Genus Felis (Mammalia), Zoologicke Listy 25:193, 1976. 36. Lantinga E, Kooistra HS, van Nes JJ: [Periodic muscle weakness and cervical ventroflexion caused by hypokalemia in a ], Tijdschr Diergeneeskd 123:435, 1998. 38. Linneaus C: Systema naturae per regna tria naturae, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis, ed 10, Holmiae, 1758, Laurentii Salvii. 40. Lipinski MJ, Froenicke L, Baysac KC et al: The ascent of cat breeds: genetic evaluations of breeds and worldwide random-bred populations, Genomics 91:12, 2008. 42. Louwerens M, London CA, Pedersen NC et al: Feline lymphoma in the post- era, J Vet Intern Med 19:329, 2005. 43. Lyons L, Biller D, Erdman C et al: Feline polycystic kidney disease mutation identified in PKD1, J Am Soc Nephrol, 2004. 50. Noden DM, Evans HE: Inherited homeotic midfacial malformations in Burmese cats, J Craniofac Genet Dev Biol Suppl 2:249, 1986. 52. Nowak RM: Walker's carnivores of the world, Baltimore, 2005, Johns Hopkins University Press. 61. TICA. The International Cat Association, 2010. http://www.tica.org/ 22 June 2011 Figure 1. Genetic distinction of domestic cat breeds.

Different colors represent genetically distinct groups of cats. Different breeds of the same color imply they form a breed “family” in that these breeds are not genetically distinct from one another.

23 Figure 2. Genetic analysis of Turkish Cats

Genetic profiles of Turkish Van cats analyzed in the Turkish cat study. Colors indicate different genetic profiles. Each line is a different cat. Cats to the left are registered Turkish Vans in the USA.

Figure 3. Genetic analysis of Turkish cats in comparison to street cats from

Cyprus and Turkey. Genetic profiles of cats analyzed in the Turkish cat study. Colors indicate different genetic profiles. Each line is a different cat. Blue = Turkish street cats and Turkish Angoras, Red = Turkish Vans and Green = cats from Cyprus. Three genetic groupings were statistically significant.

24 Figure 4. Estimating cat breed health with genetics

a. Genetic variation in cat breeds. Ho is average heterozygosity, F is the fixation index.

a. Genetic variation in random bred cats. Ho is average heterozygosity, F is the fixation index.

25 Breeding practices according to breeds, time and place, and consequences Leroy G., Rognon X., INRA, UMR1313 Génétique Animale et Biologie Intégrative, France [email protected]

Abstract: With companion animals, there is a large diversity in breeding practices, which may impact genetic variability and health of selected breeds. Based on a survey, we illustrate the specificities of scent hound dog breeders regarding either their breeding goals or their mating practices. Mating between close relatives are also investigated in different cat breeds. Such mating are more or less frequent according to breeds and countries, but seems to occur more rarely in the last few years. Deleterious impact of inbreeding on traits is also investigated for four dog breeds. In most cases, litter size and longevity are found significantly reduced for individuals with large inbreeding levels. These results illustrate how some breeding practices, such as mating between close relatives, may have an impact on fitness and welfare in cat and dog breeds.

Introduction Selection in companion animals such as dogs or cats differs from other domesticated species as they are generally not raised by production or profits. A large majority of breeders are occasional or hobby breeders and it seems to exist a large diversity of breeding practices according to those breeders (Leroy et al. 2007). Recently, several reports have pointed out the potential deleterious impact that may have some of those selection practices regarding the welfare of purebred dogs (Nicholas 2011), in relation for instance to traits selection which, when taken to extremes, are deleterious to health (Collins et al. 2011). Inbreeding has also been shown to have deleterious impact on traits related to reproduction and occurrences of some diseases (Urfer 2009, Mäki et al. 2001).

We propose here to show, on a few examples in dog and cat, the diversity of breeding practices, between breeds, between countries, and over time. Impact of on litter size and longevity will also be illustrated based on some preliminary results on dog breeds.

Breeding practices according to breeds, results from a dog breeders survey

In dog species, where breeds shows a particularly large morphological diversity, in relation, among others, to their different uses (pets, hunting, herding…), one may expect to find also a large diversity of breeding goals according to breeds. To investigate if some differences could also be found in relation to selection tools, management methods, and reproduction tools, a survey composed of 55 questions was carried out in 2007 among 985 French dog breeders (Leroy et al. 2007). Two main explanatory variables were used to analyses the results, namely the number of litters produced, and the FCI group of the main breed raised.

26 Table 1: Average rank of breeding goals declared by dog breeders according to the FCI breed group, the lower number being the best goal (there was no significant effect of number of litters produced on the answers). Whole 1 2 3 6 7 8 9 4-5-10 sample Morphology*** 2.1 2.4 1.8 1.7 2.1 2.5 2.6 1.7 1.9 Behaviour*** 2.4 1.9 2.5 2.1 2.9 3.1 2.0 2.5 2.3 Health*** 2.9 3.1 2.1 3.0 4.2 3.5 2.5 2.1 2.8 Work*** 3.7 3.8 4.7 4.3 1.9 2.2 4.2 4.8 4.1 Others NS 4.9 4.9 4.9 4.8 4.9 4.8 4.8 4.9 4.9 NS non significant, *** P< 0.001 (From Leroy et al. 2007)

One of the main results of the study was related to the specificities of scent hound breeders, considering either breeding objectives or mating practices. Indeed, if “morphology” was in general considered as the first or the second breeding goal (see table 1), “working abilities” were more important for scent hounds (FCI group 6) pointing dog groups (FCI group 7) breeders. However, “health” as a selection goal appears completely secondary for scent hounds breeders, and only 27% of them indicated it as a breeding objective versus 56% for pointing dog breeders and 71% for overall breeders. It is difficult to interpret in what extent this result is related to a low number of health problems in scent hounds, which infers that health does not appear as a problem for breeders, or to the fact that scent hound breeders, who often raise their dogs in packs, paying less attention to the health of their dogs, relative to breeders raising dogs more “individually”. Scent hound breeders show other specificities in relation to mating practices. When paying for a mating made by a sire which does not belong to them, the main modality used is monetary payment, used by 85% of overall breeders, while in 6th group breeders, only 29% use this modality. Indeed 85% Scent hound breeders prefer to give a puppy of the litter instead, versus 28% on average. Scent hound breeders also indicate using less artificial insemination (AI) than other breeders: 85% of those breeders indicate they never used AI, versus 58% on average. Finally, according to the French Kennel Club, breeders of the 6th groups are also the only ones to regularly register dogs with unknown origin, as in 2012, those registrations represented 5% of the total registrations within the group, versus 0.1% for the other breeds. All those differences show how some breeds may have their own specificities regarding breeding practices. Those specificities may eventually be linked with specific patterns concerning within breed genetic structure and variability, which can have consequences for health and welfare of those breeds.

Inbreeding practices according to breeds, countries and time: examples in cat breeds Inbreeding practices correspond to intentional mating of related individuals, such as when breeders attempt to fix or maintain specific traits from a common ancestor. This constitutes a controversial practice, due to the eventual impact that inbreeding may have on the fitness of litters produced (inbreeding depression). As a consequence, mating between close relatives (full or half-sibs for instance) has been banned in several countries, such as the U.K. a few years ago. It is therefore particularly interesting to investigate differences that may exist according to these practices. In France, a recent study on 8 cat breeds and groups of breeds (Leroy et al. 2013a) have shown for instance that the % of individuals inbred when considering 2 generations (i.e. individuals which are the products of mating between sibs or direct parents), ranged from 2.7% (Main Coon) to 8.4% (Persian/ Exotic Shorthair), illustrating the differences according to breeds.

27 Those differences may also exist within a given breed. As an illustration we analysed an international pedigree database for Birman breed, provided by Jerold Bell. We computed the % of individuals inbred when considering different generations.

Table 2: % of individuals inbred considering 2 or 3 generations during the 1991-2010 period according to four countries Country USA UK Sweden Finland Number of individuals 1185 1481 2820 1508 considered % of individuals inbred 7% 4% 2% 0% after 2 generations % of individuals inbred 27% 26% 12% 7% after 3 generations

As illustrated by table 2, when comparing different countries over the 1991-2000 period, breeders from Nordic countries seem to make such mating rather rarely compared to the UK or the USA. In these two countries 26% and 27% of kittens born over this period of time are inbred when considering 3 generations, versus 7% and 12% in Finland and Sweden respectively.

% of individuals inbred considering

Figure 1: Evolution of % of Birman cats inbred according to different number of generations considered, over the 1970-2010 period

It appears also that such mating practices are less and less frequents (see figure 1): from the 70s to the 2000s, the percentage of individuals inbred considering 3 generations have decreased from 44% to less than 10%. These results are probably explained by the fact that welfare is a growing concern, which is particularly taken into account in Nordic countries.

Inbreeding consequences on litter size and longevity: examples in dog

It is not easy to quantify the impact of inbreeding on breed health, since they depend on the mating system, demographic history of the breed and the genetic mechanism involved (Ballou 1997). Here we propose to illustrate the consequences of inbreeding on prenatal and postnatal survival of purebred dogs, considering litter size and longevity, based on births and deaths declared for 4 breeds raised in France. Litters born over the 1990-2012 period as well as dogs

28 declared as dead over 2007-2012 were considered for this (see Leroy et al. 2013b). Here dogs were divided into three inbreeding classes, considering either individuals with inbreeding coefficient lower than 6.25% (corresponding to an inbreeding equivalent to a mating between cousins), between 6.25 and 12.5% (mating between half-sibs), and 12.5% and larger.

Litter size Longevity (in years) Inbreeding Figure 2: Evolution of litter size and longevity according to inbreeding coefficient for Bernese Mountain Dog (BMD), German Shepherd Dog (GSD), Epagneul Breton (EPB) and West Highland White Terrier (WHW) (95% standard error indicated)

Figure 2 shows the reduction in prolificacy and survival within dog breeds in relation to inbreeding depression. In all case, except for longevity in West Highland White Terrier, inbreeding classes were found to have a significant impact on the traits considered (P<0.001). For instance, in the German Shepherd Dog breed, the average litter size decreased from 5.1 for litters with low inbreeding coefficient, to 4.7 for litters with inbreeding coefficient larger than 12.5%. Similarly Epagneul Breton dogs with inbreeding coefficient lower than 6.25% showed an average around 11.5 years, while this longevity was reduced to 10.4 years for dogs with inbreeding larger than 12.5%. These results show that mating between close relatives clearly impact the fitness of litters produced, even if there are other factors that affect more largely the survival and the welfare of animals raised.

Discussion As illustrated above, breeders of companion animals show a large diversity of breeding practices, which may impact the genetic variability, as well as the health of populations and individuals selected. Inbreeding practices may have, in theory, positive effects at the population level. Indeed it is supposed to increase the exposure of recessive deleterious alleles to selection, increasing inbreeding purge and reducing the risk of dissemination of a specific defect (Leroy 2011). Yet, given the deleterious consequences that high level of inbreeding may have on traits related to fitness, namely the litter size and longevity, mating between close relatives should not be recommended in any case. In practice, it is quite difficult to avoid any level of inbreeding in a selection program, especially in breeds with . However, one may be recommend to limit rapid increase of inbreeding as, in theory, slow rates of inbreeding result in more efficient selection against deleterious defects (Fu et al. 1998). At the population scale, the over-use of some reproducers should also be avoided as it may increase the risk of dissemination of genetic disorders (Leroy and Baumung 2011). Finally, choosing reproducers unrelated, or eventually belonging to another breed, may constitute another option to introduce genetic variability within a given kennel or breed. To conclude, it has to be emphasized that the management of breed health have to be planned both at the breeder scale and at the breed club

29 scale. This is why to avoid health problem and get rid of inherited disease, the best chance for a dog or cat breed is to have breeders and clubs fully cooperating in this common goal.

References Collins LM, Asher L, Summers JF, McGreevy P (2011) Getting priorities straight: Risk assessment and decision-making in the improvement of inherited disorders in pedigree dogs. Vet J 189(2): 147- 154. Fu YB, Namkoong G, Carlson JE (1998) Comparison of breeding strategies for purging inbreeding depression via simulation. Conserv Biol 12: 856-864. Leroy G, Verrier E, Wisner-Bourgeois C, Rognon X (2007) Breeding goals and breeding practices of French dog breeders: results from a large survey. Rev Med Vet 158: 496-503. Leroy G (2011) Genetic diversity, inbreeding and breeding practices in dogs: Results from pedigree analyses. Vet J 189: 177-182. Leroy G, Baumung R (2011) Mating practices and the dissemination of genetic disorders in domestic animals, based on the example of dog breeding. Anim Genet 42(1): 66-74. Leroy G, Hedan B, Phocas F, Verrier E, Mary-Huard T (2013) Inbreeding impact on prolificacy and longevity in dogs. 64th annual EAAP meeting. Nantes. Leroy G, Vernet E, Pautet MB, Rognon X (2013a) An insight into population structure and gene flow within purebred cats. J Anim Breed Genet. Mäki K, Groen AF, Liinamo AE, Ojala M (2001) Population structure, inbreeding trend and their association with hip and elbow dysplasia in dogs. Anim Sci 73: 217-228. Nicholas FW (2011) Response to the documentary Pedigree Dogs Exposed: Three reports and their recommendations. Vet J 189(2): 123-125. Urfer SR (2009) Inbreeding and fertility in Irish Wolfhounds in Sweden: 1976 to 2007. Acta vet scand 51: 21.

30 Inbreeding, Outbreeding, and Breed Evolution Jerold S Bell DVM, Tufts Cummings School of Veterinary Medicine, North Grafton. MA [email protected]

Pure-bred dog and pedigreed cat breeds evolved over time through to standards. These standards may have been conformational, behavioral, or working standards. The standards were usually not organized and written at the inception of the breed, but instead written at a later date of breed organization. Written standards are often updated over time – sometimes to clarify, and sometimes to accommodate changes in the breed. Changes in breed standards may change the selective pressures on what was bred for in the past, or what may be bred for in the future.

The pedigree record of a breed at its inception may be muddled with individuals of unknown ancestry, or just individuals that fit the conformational or working standard of the breed. These are the breed’s foundation stock. It is only at a time after an official “establishment” of a breed that a stud-book is assembled, and soon closed to additional individuals of unknown ancestry. Some cat breeds maintained open stud books for a period of time that allowed for the continued registration of cats adhering to a conformational phenotype. This allowed added diversity to their gene pools.

Some breeds are formed through inbreeding on small kindreds of individuals who possess a particular phenotypic trait. When original breed records are discovered, it is found that several familial lines of ancestry during breed formation are often abandoned due to the expression of deleterious or undesirable traits. It is only the lines that produce the desired characteristics and thrive through matings and generations of breeding that become the mainstream ancestral “founders” of a breed.

Some breeds are formed through the cross-breeding of individuals from other established breeds. These individuals would be members of established breeds that have already gone through the original breeding and purging process. The new breed would still go through the typical expansion process.

The pedigree record of breeds shows that after formation, the breed will go through a significant population expansion associated with increased average inbreeding coefficients. The Birman cat breed and Cavalier King Charles Spaniel breeds are shown as examples.

Inbreeding coefficients show the genetic relatedness of the parents of individuals. Average inbreeding coefficients of breed populations show trends in breed evolution. You can look at coefficients two different ways – a total average inbreeding coefficient that accounts for all generations, and an average inbreeding coefficient based on a set number of generations. The total generational average inbreeding coefficient can only increase over time, unless importation from unrelated stock is added to the gene pool. A 10 generation average inbreeding coefficient calculated from generation to generation (based on decade of birth) will decrease in an expanding population where the average relatedness of breeding pairs is less than the previous generation. The single most important factor increasing average 10 generation inbreeding coefficients is the popular sire syndrome. With this, the breed gene pool truncates around a popular sire line, with the resultant loss of genetic influence of other quality male lines.

Molecular genetic studies of the chromosomal structure of dog breeds show large haplotype blocks (identical sections of ) and linkage disequilibrium (LD) representing the results of inbreeding and purging during breed development (vonHoldt BM et. al. Genome Res. 2011; 21:1294- 305). Studies of dog breeds estimate that they lose on average 35% of their genetic diversity through breed formation (Gray MM et. al. Genetics 2009; 181:1493-505).

31 Molecular genetic studies of populations over time mirror those of breed formation. A study of Finnish Grey Wolves showed significant genetic diversity early on, due to migration from Russian wolves. The population then went through a significant population expansion that coincided with increased average inbreeding coefficients, decreased heterozygosity, and increases in the number of family lines as well as effective breeding population size (Jansson E et. al. Mol Ecol. 2012; 21:5178-93).

Modern breeds of cats and dogs have gone through the above mentioned genetic selection, and are in various stages of expanding their breeding population and gene pools. Some breeds may have small effective population sizes and high homozygosity. However, if their offspring are generally healthy their population can grow and expand. They are at stages of breed development where more populous breeds were earlier in their development.

Population expansion is an important aspect of breed development and maintenance. It allows on average the successive mating of individuals less related than the prior generation. It allows the creation of new “family lines” and within-breed diversity. Population contraction is detrimental to breed maintenance due to the loss of breeding lines and genetic diversity. Maintaining adequate numbers of breeders and matings is important to breed vitality and survival.

As a consequence of breed formation dog and cat breeds have high homozygosity. This is the nature of breed formation. Homozygosity by itself is not detrimental to breeds unless they carry a high genetic load of deleterious receive genes. Some breeds may show decreased litter size, increased neonatal mortality, or shorter average life spans with increases in inbreeding coefficients. These “inbreeding depression” effects are due to the homozygous expression of specific deleterious genes that cause specific disease. Direct selection against these genes and phenotypes is required to improve breed health. If breed members are dying younger, what specific disease(s) is occurring in these individuals? If the breed shows issues with fertility and fecundity, then breeders should specifically select for increased fertility and fecundity.

Some advocates of dog and cat breeding call for organized outbreeding programs that mate the least related individuals to each other. These mirror the Species Survival Plans (SSP) formulated for rare and endangered species. The result of this effort will produce a randomized population and within-breed increases in heterozygosity regarding gene distribution. However, this will have no effect on the frequency of deleterious genes. Genes for breed-related genetic disorders that are already dispersed in the gene pool will continue to produce affected individuals in a random fashion. This type of breeding plan is also self-limiting, because as you remove the genetic differences between individuals it becomes increasingly harder to outbreed (find mates that are genetically unlike each other). A healthy and diverse breed gene pool should have many outbred clusters as well as different linebred families.

The genetic tools of linebreeding and outbreeding should be used for specific purposes. Breeders may use different breeding tools with each mating that are either closer (linebreeding) or more distant (outbreeding) than the average in the population based on their needs. Linebreeding concentrates the genes of specific ancestors. Outbreeding brings in genes that are not present in the mate. When breeders are each performing matings that are a little different from each other – some linebreeding in one line, some outbreeding, some linebreeding in another line, etc., it maintains a diverse breed population.

The only way to decrease the frequency of deleterious genes in a population (and increase the frequency of favorable genes) is through direct selection against (and for) those genes through and phenotypic evaluation. The rate and degree of genetic improvement through selection is directly

32 proportional to the amount of variation that exists between individuals within the breed. Randomizing a population through outbreeding decreases the ability to apply selective pressure for genetic improvement. Selective pressure requires lines of individuals who are unlike each other.

Some studies bemoan the homozygosity found in breeds, and call for selection to increase minor frequency alleles and haplotypes. Molecular genetic tools can identify these, but in most cases the phenotypic effects of increasing their frequency are unknown. It is possible that genetic selection for quality and against undesirable traits reduced the frequency of these genes. Blindly selecting for them without knowing their effect could significantly reverse selection-based breed improvement.

When breeds show high frequency of genetic disease, or significantly diminished fertility and fecundity, they may have too high a genetic load of disease liability genes. In extreme instances they may require; a SSP-type plan, opening the study book to importation, or cross-breeding to other related breeds. However, most breeds do not find themselves in such dire situations, and only require proper selection to improve their gene pools and genetic health.

The following conclusions can be made concerning breed evolution and health:

-The effects of inbreeding (homozygosity, large haplotype blocks and increased linkage disequilibrium) are a natural consequence of breed formation. -Healthy breed gene pools require expanding, or large stable populations. -Breed health should be measured based on regular surveys of health and reproduction. -Genetic selection for breed characteristics should avoid disease related phenotypes. -Genetic selection for breed health should be directed against specific disease liability genes and phenotypes. -Breeders should avoid the overuse of popular sires – the most significant factor in limiting breed genetic diversity.

33 Unraveling the Phenotypic and Genetic Complexity of Canine Cystinuria Paula S. Henthorn and Urs Giger, Section of Medical Genetics (PennGen), University of Pennsylvania, Philadelphia, PA [email protected]

Cystinuria is a disease of disrupted amino acid transport in the collecting ducts of the kidney fail to reclaim certain amino acids (cystine and the dibasic amino acids ornithine, lysine and arginine referred to as COLA). The increased urinary COLA concentrations reach saturation levels for cystine, which precipitates to form crystals and stones resulting in renal to urethral obstructions. Mutations in the SLC3A1 and SLC7A9 genes give rise to cystinuria in the vast majority of cystinuric humans, where the disease shows autosomal recessive or dominant inheritance (reviewed in Palacin et al., 2001; Chillaron et al., 2010).

Cystine calculi have been reported from at least 70 dog breeds, with increased incidence in several breeds (Ling et al., 1998; Osborne et al., 1999); in contrast cystinuria is rarely seen in cats. We previously demonstrate autosomal recessively inherited cystinuria in Newfoundland dogs (with less frequent urolithiasis in females due to anatomical urological differences) caused by a mutation in the SLC3A1 gene that precludes the expression of a functional protein (Casal et al. 1995; Henthorn et al. 2000). In addition we discovered a similar mutation in the SLC3A1 gene causing recessively inherited cystinuria, a dominantly inherited cystinuria due to a deletion in SLC3A1, and a missense mutation in SLC7A9 gene associated with persistent cystinuria and cystine stone formation in , Australian cattle, and (European) miniature pinscher dogs, respectively (Brons et al. 2013). These mutations and their consequences appear to be consistent to those seen in human cystinuria.

However, for a number of other breeds examined for mutations in the SLC3A1 and SLC7A9 gene protein-coding regions, no obvious mutations have been identified (Henthorn et al., 2000; Harnevik et al. 2006; PH, UG unpublished data). In addition, it appears that in some breeds (Mastiff and related breeds, Irish terriers), only adult, intact male dogs show elevated urine COLA concentrations. In these breeds, the average age of stone formation is later than seen in male Newfoundland dogs (Giger et al. 2011a,b; PH, UG unpublished data). Most importantly, in these breeds, urinary aminoaciduria normalized after , making neutering an effective treatment for cystinuria in some, but not all breeds. Neuter status has no effect on cystinuria in Newfoundlands, Labrador retrievers, Australian cattle dogs, and Miniature Pinschers (Brons et al. 2013).

For Mastiffs and related breeds, we have determined that a non-conserved amino acid substitution (Harnevik et al., 2006; PH unpublished data) as well as other DNA changes in the SLC3A1 gene that may affect the expression levels of that gene are associated with stone formation (PH, unpublished data). Intact male dogs that have two copies of this variant version of the SLC3A1 gene appear to form stones between 1 and 4 years of age (older than Newfoundlands, but younger than the average age of stone formation reported from the Minnesota stone laboratory; Osborne et al., 1999). However, not all stone-forming Mastiffs are homozygous (have two copies) of this variant allele. Additional genetic or environmental factors may play a role for cystinuria in Mastiffs. This variant SLC3A1 allele is not found in androgen- dependent cystinuric dogs of other breeds, several in which cystinuria has a relatively high incidence.

1

34 To simplify discussions of cystinuria, we have suggested a classification system for canine cystinuria that encompasses both discriminating aspects of the phenotype (for example, gender affected, androgen dependence, and mode of inheritance) and the gene associated with the disease (Brons et al. 2013; see table below). We designate type I cystinuria when the disease shows autosomal recessive inheritance, Type II when inheritance is autosomal dominant, and Type III for sex-limited/androgen-dependent inheritance (PH, UG, unpublished data). Additional types can be assigned if found. Specific mutations within each type should lead to phenotypes that are sufficiently similar that the same medical management and breeding advice applies to all cases within that type. Involvement of the SLC3A1 gene is indicated by adding –A, and similarly addendum of –B indicated involvement of mutations in SLC7A9.

Phenotype Type I - A Type II - A Type II - B Type III - Autosomal Autosomal Autosomal Inheritance Sex-limited recessive dominant dominant Gene SLC3A1 SLC3A1 SLC7A9 Unknown Males and Males and Males and Intact Adult Gender Females Females Females Males Androgen dependent No No No Yes Homozygou ≥ 8,000 ≥ 8,000 nd *COL s ≤ 4,000 A Heterozygou ≤ 500 ≥ 3,000 ≥ 700 s Newfoundland Aust. cattle Min. Pinscher Mastiff & Landseer dog related Breeds affected Labrador Scot. Deerhound Irish Terrier Newfoundland Aust. cattle Min. Pinscher †Mastiff & DNA-based genetic Landseer dog related (risk for test breeds Labrador earlier stone formation) * µmol/g creatinine, normal ≤ 500 † While we recommend DNA testing of Mastiffs and related breeds for cystinuria, be aware that this DNA test alone does not completely predict the cystinuria status of every dog (particularly for 1-2 dogs). Therefore, annual urinary nitroprusside testing is recommended for all adult intact male dogs.

While there is still much left to discover, these findings advance our understanding of this genetically complex disease. The characterization of the heterogeneity of cystinuria in different canine breeds and our proposed new classification system have important ramifications for the medical and genetic management of cystinuria in many dog breeds. Determining the molecular mechanism of cystinuria in Mastiffs and other breeds will provide insight into the genetically complex diseases. Most surprisingly, for cystinuria in some breeds, neutering can effectively cure the disease, but we caution clinicians to contact us for cases where no studies of cystinuria

2

35 have yet been performed in the breed. And finally, these and future studies will have an impact on the genetic control of cystinuria in future generations of dogs.

ACKNOWLEDGEMENTS Dr. Henthorn's cystinuria research is performed in collaboration with Dr. Urs Giger (University of Pennsylvania School of Veterinary Medicine) and Dr. Adrian Sewell (Department of Pediatrics, University Children’s Hospital, Frankfurt am Main, Germany) Contributors at the University of Pennsylvania include Dr. Ann-Kathrin Brons, Caitlin Fitzgerald, Michael Raducha, JunLong Liu, and Karthik Raj. This work was supported by the University of Pennsylvania School of Veterinary Medicine, the Canine Health Foundation, the National Institutes of Health (OD 010939), the Mastiff and Scottish Deerhound national breed clubs, and by individual breeders. We thank many veterinarians, dog owners and breeders for their participation in this work.

REFERENCES Brons A-K , Henthorn PS, Raj K, Fitzgerald CA, Liu J, Sewell AC, Giger U. SLC3A1 and SLC7A9 Mutations in Autosomal Recessive or Dominant Canine Cystinuria: A New Classification System. J Vet Internal Medicine, accepted for publication. Casal ML, Giger U, Bovee KC, Patterson DF. Inheritance of cystinuria and renal defect in Newfoundlands. J Am Vet Med Assoc 1995;207:1585-1589. Chillaron J, Font-Llitjos M, Fort J, Zorzano A, Goldfarb DS, Nunes V, Palacín M. Pathophysiology and treatment of cystinuria. Nat Rev Nephrol 2010;6:424-434. Giger U, Sewell AC, Lui J, Erat A, Sewell AC, Henthorn PS. Update on Fanconi Syndrome and Cystinuria in Dogs: Amino Acidurias. In: ACVIM Forum, , CO 2011a Giger U, Lee JW, Cait Fitzgerald et al, Characterization Of Non-Type I Cystinuria In Irish Terriers, J Vet Int Med, 2011b ACVIM Forum Abstracts, 2011b:25:718 Harnevik L, Hoppe A, Soderkvist P. SLC7A9 cDNA cloning and mutational analysis of SLC3A1 and SLC7A9 in canine cystinuria. Mamm Genome 2006;17:769-776. Henthorn PS, Liu J, Gidalevich T, Fang J, Casal ML, and Patterson DF. Canine cystinuria: polymorphism in the canine SLC3A1 gene and identification of a nonsense mutation in cystinuric Newfoundland dogs. Hum Genet 2000;107:295-303. Ling GV, Franti CE, Ruby AL, and Johnson DL. Urolithiasis in dogs. II: Breed prevalence, and interrelations of breed, sex, age, and mineral composition. Am J Vet Res 1998;59:630-642. Osborne CA, Sanderson SL, Lulich JP, Bartges JW, Ulrich LK, Koehler LA, KA, Swanson LL. Canine cystine urolithiasis. Cause, detection, treatment, and prevention. Vet Clin N Am:Sm An Pract 1999; Jan;29(1):193-211, xiii. Palacin M, Goodyer P, Nunes V, et al. Cystinuria. In: Scriver CR, ed. The metabolic and molecular bases of inherited disease, 8th ed. New York: McGraw-Hill; 2001:4909-4932.

3

36 How to Use and Interpret Genetic Tests for Heart Disease in Cats and Dogs Kathryn M. Meurs, DVM, PhD, Diplomate ACVIM (Cardiology), North Carolina State University College of Veterinary Medicine, Raleigh, NC [email protected]

Important definitions: Congenital heart disease- present since birth, may be inherited or may not be inherited Acquired heart disease- develops after the animal reached maturity, may be inherited or may not be inherited Heterozygote: has 1 copy of the mutated gene and 1 copy of a normal gene

Homozygote: has 2 copies of the mutated gene

Penetrance: Percentage of population with a mutation that show the disease

Expression: Severity of disease

Utilization of molecular information for screening and therapeutic issues Increasingly heart disease in dogs and cats is found to be of inherited origin. This seminar will discuss common testing for known genetic mutations for cats and dogs. Genetic tests are generally a PCR test that identifies either a marker for the disease or that identifies the actual genetic mutation. PCR is a method that takes a small amount of DNA provided by the clinician or owner and amplifies a region of interest so it can be carefully inspected. DNA can be usually provided in a blood sample in an EDTA tube, a buccal swab or even a semen sample. The DNA will be inspected for the abnormality by the lab and the presence or absence of the marker or mutation identified. However, breeders and owners should be cautioned and advised how to best use the information. The results should be carefully considered and should be weighed against the severity of the trait, the size of the breed’s gene pool, the mode of inheritance of the trait and the positive traits that this individual animal brings to a breed. In some cases, strict screening and removal programs may be very detrimental to small gene pools in specific breeds; breeding recommendations should be carefully designed. We will use two examples for illustration – Feline Hypertrophic Cardiomyopathy and Boxer Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC) Feline Hypertrophic Cardiomyopathy Hypertrophic cardiomyopathy is the most common form of heart disease in the cat. It is an adult onset and known to be inherited in the and breeds and thought to be inherited in Norwegian Forest, Siberian, Sphynx and Bengal cats among others. Causative genetic mutations have now been identified in the Maine Coon and Ragdoll. In the Maine Coon, a genetic mutation has been identified in the myosin binding protein C (MYBPC3) gene. The penetrance of the disease is fairly low for heterozygotes (about 30%

37 show disease) but high for homozygotes (about 80%). The Maine Coon mutation appears to be quite breed specific. It is unlikely to be associated with hypertrophic cardiomyopathy in other breeds of cats unless they are closely related to the Maine Coon breed. Additionally, although this mutation has been shown to lead to the development of this disease in this breed, not all cats with the disease have this mutation so it is clear that there is more than 1 mutation in the maine coon cat. A substitution mutation has also been identified in the myosin binding protein C gene in the Ragdoll cat. However, the Ragdoll mutation is different from the Maine Coon mutation and is at a different location. It is extremely unlikely that the Maine Coon and Ragdoll mutations were inherited from a common ancestor since the mutations are different and are located in such different regions of the gene. Additionally, it is very unlikely that other breeds of cats have the identical mutation. Genetic testing is now available to test a cat for either mutation by submitting a DNA sample to a reputable screening laboratory (http://www.ncstatevets.org/genetics/) . Good quality DNA samples can be obtained either from a blood sample in an EDTA tube or by brushing the oral gums of the cat with a special buccal swab, although many labs will even accept samples submitted on a cotton swab. The test results should verify that the cat is negative, heterozygous or homozygous for the mutation. Cats that test negative do not have the mutation. This does not mean that they cannot ever develop hypertrophic cardiomyopathy, it simply means that they will not develop the form of the disease caused by the specific genetic mutation. Due to an apparently fairly high prevalence of the mutation in both breeds, it would seem to be unwise to recommend that all cats with the mutation be removed from the breeding programs since this could result in dramatically altering the genetic makeup of these breeds. Additionally, it should be emphasized that not all cats that have the mutation, particularly if they are heterozygous, will develop a clinical form of the disease. Our current recommendations for both breeds are to not use cats that are homozygous for the mutation for breeding purposes since they will certainly pass on the mutation and they have the highest risk of developing the disease. Heterozygous cats should be carefully evaluated. Cats that have many strong positive breed attributes and are disease negative at time of breeding could be bred to a mutation negative cat. Their lack of clinical disease may suggest that they have a less penetrant form of the disease or that they just do not show evidence of this adult onset clinical disease yet. Therefore these cats should only be used if they are exceptional for the breed and they should be clinically evaluated for the disease every year. If they develop the clinical disease, they should no longer be kept in the breeding program. The offspring of the mating of a positive heterozygous and a negative should be screened for the mutation, and if possible, a mutation negative with desirable traits should be selected to replace the mutation positive parent in the breeding colony. Over a few generations this will decrease the prevalence of the disease mutation in the population, hopefully without greatly altering the genetic makeup of the breed too significantly. Finally, disease negative but mutation positive cats should be evaluated annually for presence of disease.

38 Arrhythmogenic right ventricular cardiomyopathy in the boxer Since the early 1980’s, the term boxer cardiomyopathy has been used to describe adult boxer dogs that present with ventricular arrhythmias, and sometimes, syncope. Recent studies have demonstrated that the disease has many similarities to a human disease called Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC). The similarities between the diseases include clinical presentation, etiology and a fairly unique histopathology that includes a fibrous fatty infiltrate of the right ventricular free wall. The disease is most commonly characterized by ventricular arrhythmias, syncope and sudden death. However, systolic dysfunction and ventricular dilation are seen in a small percentage of cases. Arrhythmogenic right ventricular cardiomyopathy is a familial disease in the boxer and appears to be inherited as an autosomal dominant trait. Unfortunately, the disease also appears to be a disease of variable genetic penetrance and affected dogs can have many different presentations including asymptomatic, syncope, sudden death and systolic dysfunction with CHF. A genetic mutation has now been identified for boxer arrhythmogenic right ventricular cardiomyopathy although it is not yet known if this will be responsible for all cases of the disease since in human beings there are several known mutations. Individuals interested in screening for the disease in breeding animals may do so with either a buccal swab or blood sample. (http://www.ncstatevets.org/genetics/) In human beings with ARVC, there are multiple genetic mutations that can lead to the development of the disease. At this time we do not know if this mutation is the only cause in the Boxer, therefore, clinical screening is still recommended. Since ARVC presents as an electrical abnormality more often than one of myocardial dysfunction, screening efforts should be based on annual Holter monitoring as well as annual echocardiography. Until a greater understanding of disease inheritance and disease progression exists, caution should be used when advising breeders to remove dogs from breeding programs. Overzealous removal of animals based on the results of a single Holter monitor may have a significant negative impact on the breed. Conclusions Genetic testing is becoming increasingly available for the pet population. It should be remembered that inherited disease is complex and that there is no perfect test. Results of the genetic test should be carefully considered and should be weighed against the severity of the trait, the size of the breed’s gene pool, the mode of inheritance of the trait and the positive traits that this individual animal brings to a breed. In some cases, strict screening and removal programs may be very detrimental to small gene pools in specific breeds; breeding recommendations should be carefully designed.

39 Update on Genetic Tests for Diseases & Traits in Cats: Implications for Cat Health, Breed Management & Human Health Leslie A. Lyons, PhD, College of Veterinary Medicine, University of Missouri - Columbia [email protected]

Below is an update of genetic tests for the domestic cat (30 July 2013). Text is modified from: Genetic testing in domestic cats. Lyons LA. Mol Cell Probes. 2012 Dec;26(6):224-30. doi: 10.1016/j.mcp.2012.04.004. Epub 2012 Apr 21.

Introduction Genetic testing has been available in the domestic cat since the 1960’s, but as like other species, over the past 50 years, the level of resolution has improved from the chromosome level to the sequence level. Knowing the direct causative mutation for a trait or disease assist cat breeders with the breeding programs and can help clinicians determine heritable presentations versus idiopathic versions of a health concern. Genetic tests cover all the various forms of DNA variants, including chromosomal abnormalities, mtDNA variation, gene loss, translocations, large inversions, small insertions and deletions and the simple nucleotide substitutions. Higher throughput technologies have made genetic testing cheaper, simpler and faster, thereby making cat genetic testing affordable to the lay public and small animal practice clinicians. The genetic resources for cats and other animal species have also opened the doors for animal evidence to be supportive in criminal investigations. This presentation will highlight the various tests available for the domestic cat and their specific capabilities and role’s in cat health and management.

Cat Chromosomes Some of the earliest genetic testing for any species was the examination of the chromosomes to determine the presence of the normal and complete genomic complement. Early studies of mitotic chromosomes of the domestic cat revealed an easily distinguishable karyotype consisting of 18 autosomal chromosomes and the XY sex chromosome pair, resulting in a 2N complement of 38 chromosomes for the cat genome [1]. Sex chromosome aneuploidies and trisomies of small acrocentric chromosomes were typically associated with cases of decreased fertility and syndromes that displayed distinct morphological presentations. Turner’s Syndrome (XO), Klinefelter’s Syndrome (XXY) and chimerism has been documented in the domestic cat. Because cat has a highly recognizable X-linked trait [2-5], Orange, and the X-inactivation process was recognized [6], tortoiseshell and calico male cats were the first feline suspects of chromosomal abnormalities. Karyotypic and now gene-based assays are common methods to determine if a cat with ambiguous genitalia [7] or a poor reproductive history has a chromosomal abnormality. Karyotypic studies of male tortoiseshell cats have shown that they are often mosaics, or chimeras, being XX/XY in all or some tissues [5, 8-15]. The minor chromosomal differences that are cytogenetically detectable between a domestic cat and an Asian leopard cat are likely the cause of fertility problems in the Bengal cat breed, which is a hybrid between these two species [16]. Other significant chromosomal abnormalities causing common “syndromes” are not well documented in the cat. Several research and commercial laboratories can perform cat chromosomal analyses when provided a living tissue, such as a fibroblast biopsy or whole blood for the analysis of white blood cells.

40 Inherited Disease Tests The candidate gene approach has been fruitful in domestic cat investigations for the identification of many diseases and trait mutations. The first mutations identified were for a gangliosidosis and muscular dystrophy, discovered in the early and mid-1990’s [17, 18], as these diseases have well defined phenotypes and known genes with mutations that were as found in humans. Most of the common diseases, coat colors, and coat types were deciphered in the cat following the same candidate gene approach.

Most of the identified disease tests in cats that are very specific to breeds and populations are available as commercial genetic tests (Table 1). Typically, diseases are identified in cat breeds, which are a small percentage of the cat population of the world, perhaps at most 10 – 15% in the USA [19]. However, some mutations that were found in a specific breed, such as mucopolysaccharidosis in the Siamese [20, 21], were found in a specific individual and the mutation is not of significant prevalence in the breed (Table 2). These genetic mutations should not be part of routine screening by cat breeders and registries, but clinicians should know that genetic tests are available for diagnostic purposes, especially from research groups with specialized expertise, such as at the University of Pennsylvania (http://research.vet.upenn.edu/penngen). Other diseases, such as polycystic kidney disease (PKD), are prevalent, PKD in Persians is estimated at 30 – 38% worldwide [22-24]. Because of cross breeding with Persians, many other breeds, such as British Shorthairs, American Shorthairs, Selkirk Rex, and Scottish Folds, also need to be screened for PKD [25-27]. As PKD testing begins to become less common, as breeders remove positive cats, other genetic tests are becoming more popular, such as coat color and other disease traits (Figure 1).

Genetic Testing Concerns in Hybrid Cat Breeds Several cat breeds were formed by crossing with different species of cats. The Bengal breed is acknowledged worldwide and has become a highly popular breed. To create Bengals, Asian leopard cats (Felis bengalensis) were and are bred with domestic cat breds like Egyptian Mau, Abyssinian and other cats to form a very unique breed in both color and temperament [28]. An Asian leopard cat had a common ancestor with the domestic cat about 6 million years ago, the about 8 million years ago, the Serval about 9.5 million years ago [29]. The is more closely related to a domestic cat than the leopard cat to the domestic cat. In addition, for some of these wild felid species, different subspecies were incorporated into the breed. The DNA sequence between a domestic cat and one of these wild felid species will have many genetic differences, less for the Jungle cat, more for Serval as compared to a domestic cat. The genetic differences are most likely silent mutations, but, the variation will interplay with genetic assays and may cause more allelic drop-out than what would be normally anticipated. No genetic tests are validated in the hybrid cats breeds, although the tests are typically used very frequently in Bengal cats. Thus, the accuracy for any genetic test is not known for hybrid cat breeds. A genetic test for the Charcoal coloration in Bengals will likely soon be available and is unique due to the hybridization with leopard cats.

Race and Breed Identification A newly developing test for the domestic cat is a race and breed identification panel. Based on the studies by Lipinski et al. (2008) [30], and Kurushima et al. (2012, submitted), STRs have been tested in a variety of random bred cats from around the world and a majority of the major

41 cat breeds of the USA and other regions. The genetic studies have been able to differentiate eight worldwide populations of cats – races – and can distinguish the major breeds. Analyses of the present day random bred cat populations suggest that the regional populations are highly genetically distinct, hence analogous to humans, different races of cats. The regional genetic differentiation is capture and displayed within the breeds that developed later from those populations. The foundation population (race) of the Asian breeds, such as Burmese and Siamese, are the street cats of Southeast Asia, whereas the foundation population (race) of the Maine Coon and are Western European cats. Phenotypic markers help to delineate breeds within specific breed families, such as the Persian, Burmese, and Siamese families. The cat race and breed identification tests are similar to tests that have been developed for the dog, such as the Mars, Inc. Wisdom Panel (http://www.wisdompanel.com/). Although similar, domestic cats are random bred cats and not a concoction of pedigreed breed cats. Cat breeds developed from the random bred populations that have existed in different regions of the world for thousands of years. Therefore, the claims of the cat race and breed identification tests are different than the dog tests, not claiming that most household cats are recent offspring of pedigreed cats.

Implications for Cat Health & Breeding To date, most cat genetic tests have been for traits that have nearly complete penetrance, having little variability in expression, and early onset. These aspects are important when considering management in the breed. If your cat has the PKD mutation – it will get kidney cysts – but the development of renal failure is variable (variable expression). Therefore, some recognized mutations in cats might be considered risk factors, predisposing an individual to health problem. Excellent examples of mutations that confer a risk in cats are the DNA variants associated with cardiac disease in cats. Hypertrophic cardiomyopathy (HCM) is a recognized genetic condition [31]. In 2005, Drs. Meurs, Kittleson and colleagues published that a DNA alteration, A31P, in the gene cardiac myosin-binding protein C 3 (MYBPC3) was strongly associated with HCM in a long-term research colony of Maine Coon cats at UC Davis [32]. Recent studies have shown that not all Maine Coon cats with the A31P mutation get HCM [33, 34] and one of those papers has mistakenly interpreted this lack of penetrance as being evidence that the A31P mutation is not causal [34]. This interpretation is misleading, causing debate as to the validity of the Maine Coon HCM test. As true in humans with cardiac disease, the finding that not all cats with the A31P mutation in MYBPC3 get HCM is actually usual in the field of HCM genetic testing.

Like cat HCM mutations, other disease mutations have shown variation in penetrance and expression, such as the CEP290 PRA mutation in Abyssinians and some cats with the pyruvate kinase deficiency can have very mild and subclinical presentations [35]. Thus, disease or trait causing mutations may not be 100% penetrant, thus, they do not always cause clinically detectable disease.

Conclusion Many aspects of the population and the specific mutation must be considered during management of a disease. Diseases with a low frequency in a large population could likely be eliminated. Diseases in a very high frequency or present in a very small population need to be slowly removed from the population with great care. Genetic testing is an important diagnostic tool for the veterinarian, breeder, and owner. Genetic tests are not 100% foolproof and the

42 accuracy of the test procedure and the reputation and customer service of the genetic testing laboratory needs to be considered. Some traits are highly desired and genetic testing can help breeders to more accurately determine appropriate breedings, potentially becoming more efficient breeders, thus lowering costs and excess cat production. Other traits or diseases are undesired, thus genetic testing can be used to prevent disease and potentially eradicating the concern from the population. Genetic tests for simple genetic traits are more consistent with predicting the trait or disease presentation, but, as genomics progress for the cat, more tests that confer risk will become more common.

43 Table 1. Common commercialized DNA tests for domestic cats.

Disease / Trait (alleles) MOI‡ Phenotype Breeds Gene Mutation

Agouti (A+, a)[36] AR Banded fur to solid All cats ASIP c.122_123delCA

Amber (E+, e)[37] AR Brown color variant Norwegian MC1R c.250G>A Forest

Brown (B+, b, bl)[38, 39] AR Brown, light brown color variants All cats TYRP1 b = c.8C>G, bl = c.298C>T

Color (C+, Cb, Cs, c)[39-41] AR Burmese, Siamese color pattern, All breeds TYR cb = c.715G>T, cs = c.940G>A, full albino c = c.975delC

Dilution (D+, d)[42] AR Black to grey / blue, All cats MLPH c.83delT

Orange to cream

Extension (E+, e) – Amber AR Brown/red color variant Norwegian MC1R c.250G>A [37] Forest

Fold (Fd, fd+) AD Ventral ear fold In Press

Gloves (G+, g)[43] AR White feet Birman KIT In Press

Hairless (Hr+, hr))[44] AR Atrichia Sphynx KRT71 c.816+1G>A

Inhibitor AD Absence of phaeomelanin All cats

Long fur (L+, l)[45, 46] AR Long fur All cats§ FGF5 c.356_367insT, c.406C>T, c.474delT, c.475A>C

Manx (M, m+) AD Absence/short tail Manx, American c.998delT, c.1169delC, and Bobtail, c.1199delC, PixieBob c.998_1014dup17delGCC

Rexing (R+, r)[47] AR Curly hair coat PYP2R5 c.250_253delTTTG

44 Rexing (Re+, re)[44] AR Curly hair coat KRT71 c.1108-4_1184del, c.1184_1185insAGTTGGAG, c.1196insT

Rexing[48] AD Curly hair coat Selkirk Rex KRT71 c.445-1G>C

Tabby[49] AR Blotched/classic pattern All cats TAQPEP S59X, T139N, D228N, W841X

AB Blood Type (A+, b)[50] AR Determines Type B All cats CMAH c.1del-53_70, c.139G>A

Craniofacial Defect AR Craniofacial Defect Burmese In Press

Gangliosidosis 1[51] AR Lipid storage disorder , Siamese GBL1 c.1457G>C

Gangliosidosis 2[52] AR Lipid storage disorder Burmese HEXB c.1356del-1_8, c.1356_1362delGTTCTCA

Gangliosidosis 2[18] AR Lipid storage disorder Korat HEXB c.39delC

Glycogen Storage Dis. IV[53] AR Glycogen storage disorder Norwegian GBE1 IVS11+1552_IVS12-1339 Forest del6.2kb ins334 bp

Hypertrophic AD Cardiac disease Maine Coon MYBPC c.93G>C Cardiomyopathy[32]

Hypertrophic AD Cardiac Disease Ragdoll MYBPC c.2460C>T Cardiomyopathy[54]

Hypokalemia[55] AR Potassium deficiency Burmese WNK4 c.2899C>T

Progressive Retinal AR Late onset blindness Abyssinian CEP290 IVS50 + 9T>G Atropy[56]

Progressive Retinal AD Early onset blindness Abyssinian CRX c.546delC Atropy[57]

Polycystic Kidney AD Kidney cysts Persian PKD1 c.10063C>A Disease[27]

45 Pyruvate Kinase Def.[58] AR Hemopathy Abyssinian PKLR c.693+304G>A

Spinal Muscular Atrophy[59] AR Muscular atrophy Maine Coon LIX1-LNPEP Partial gene deletions

‡ Mode of inheritance of the non-wildtype variant, § Long fur variants are more or less common depending on the breed. Not all transcripts for a given gene may have been discovered or well documented in the cat, mutations presented as interpreted from original publication.

46 Table 2. Other Mutations for Inherited Domestic Cat Diseases†.

Disease Gene Mutation‡ Disease Gene Mutation‡

11b-hydroxylase CYP11B1 Exon 7 G>A Mucopolysaccharidosis I[61] IDUA c. 1107_1109delCGA Def. (Congenital Adrenal or c. 1108_1110 Hypoplasia) [60] GAC

Dihydropyrimidinase Def. DPY8 c.1303G>A Mucopolysaccharidosis VI[21] ARSB c.1427T>C

Fibrodysplasia Ossificans ACVR1 G617A (R206H) Mucopolysaccharidosis VI[20, 62] ARSB c.1558G>A Progressiva

Gangliosidosis 1[63] GLB1 c.1448G>C

Gangliosidosis 2[64] HEXB c.1467_1491inv Mucopolysaccharidosis VII[65] GUSB c.1052A>G

Gangliosidosis 2[66] HEXB c.667C>T Niemann-Pick C[67] NPC c.2864G>C

Gangliosidosis 2[53] GM2A c.390_393GGTC Polydactyla[68] SHH c.479A>G

Hemophilia B[69] F9 c.247G>A Polydactyla[68] SHH c.257G>C, c.481A>T

Hemophilia B[69] F9 c.1014C>T Porphyria (congenital UROS c.140C>T, erythropoietic)[70] c.331G>A

Hyperoxaluria[71] GRHPR G>A I4 acceptor site Porphyria (acute intermittent)[72] HMBS c.842_844delGAG, c.189dupT, c.250G>A, c.445C>T

Lipoprotein Lipase Def.[73] LPL c.1234G>A Vitamin D Resistant Rickets[74] CYP27B1 c.223G>A, c.731delG

Mannosidosis, alpha[75] LAMAN c.1748_1751delCCAG Vitamin D Resistant Rickets[76] CYP27B1 c.637G>T

47 Mucolipidosis II[77] GNPTA c.2655C>T

† The presented conditions are not prevalent in breeds or populations but may have been established into research colonies. ‡ Not all transcripts for a given gene may have been discovered or well documented in the cat, mutations presented as interpreted from original publication.

48

Figure 1. Trends of genetic testing in the domestic cat. DNA-based genetic tests are presented for the cat. Parentage and individual identification (DNA) has not increased as cats do not require testing for registration. One of the most popular tests, PKD, is presented separately to show that the testing requests are decreasing as breeders are eliminating positive cats from breeding programs. Other disease tests and color tests are becoming more popular tests in the cat market.

49

Figure 2. The slippery slope of mutation lethality. Some mutations are so severe, they cause death in utero, such as PKD and taillessness. Some have high and severe penetrance while others have low and mild penetrance. All these factors and others should be considered when managing a cat population.

50 Table References

1. Eizirik, E., et al., Molecular genetics and evolution of melanism in the cat family. Curr Biol, 2003. 13(5): p. 448-53. 2. Peterschmitt, M., et al., Mutation in the melanocortin 1 receptor is associated with amber colour in the Norwegian Forest Cat. Anim Genet, 2009. 40(4): p. 547-52. 3. Lyons, L.A., et al., Chocolate coated cats: TYRP1 mutations for brown color in domestic cats. Mamm Genome, 2005. 16(5): p. 356-66. 4. Schmidt-Kuntzel, A., et al., and tyrosinase related protein 1 alleles specify domestic cat coat color phenotypes of the albino and brown loci. J Hered, 2005. 96(4): p. 289-301. 5. Imes, D.L., et al., Albinism in the domestic cat (Felis catus) is associated with a tyrosinase (TYR) mutation. Anim Genet, 2006. 37(2): p. 175-8. 6. Lyons, L.A., et al., Tyrosinase mutations associated with Siamese and Burmese patterns in the domestic cat (Felis catus). Animal Genetics, 2005. 36(2): p. 119-26. 7. Ishida, Y., et al., A homozygous single-base deletion in MLPH causes the dilute coat color phenotype in the domestic cat. Genomics, 2006. 8. Gandolfi, B., et al., Off with the gloves: Mutation in KIT implicated for the unique white spotting phenotype of Birman cats. . submitted, 2010. 9. Gandolfi, B., et al., The Naked Truth: Sphynx and Devon Rex cat breed mutations in KRT71. Mammalian Genome, 2010. in press. 10. Drogemuller, C., et al., Mutations within the FGF5 gene are associated with hair length in cats. Anim Genet, 2007. 38(3): p. 218-21. 11. Kehler, J.S., et al., Four independent mutations in the feline fibroblast growth factor 5 gene determine the long-haired phenotype in domestic cats. J Hered, 2007. 98(6): p. 555-66. 12. Gandolfi, B., et al., To the Root of the Curl: A Signature of a Recent Selective Sweep Identifies a Mutation That Defines the Cornish Rex Cat Breed. PLoS One, 2013. 8(6): p. e67105. 13. Gandolfi, B., et al., A splice variant in KRT71 is associated with curly coat phenotype of Selkirk Rex cats. Sci Rep, 2013. 3: p. 2000. 14. Kaelin, C.B., et al., Specifying and sustaining pigmentation patterns in domestic and wild cats. Science, 2012. 337(6101): p. 1536-41. 15. Bighignoli, B., et al., Cytidine monophospho-N-acetylneuraminic acid hydroxylase (CMAH) mutations associated with the domestic cat AB blood group. BMC Genet, 2007. 8: p. 27. 16. De Maria, R., et al., Beta-galactosidase deficiency in a Korat cat: a new form of feline GM1- gangliosidosis. . Acta Neuropathol (Berl), 1998. 96: p. 307-14. 17. Bradbury, A.M., et al., Neurodegenerative lysosomal storage disease in European Burmese cats with hexosaminidase beta-subunit deficiency. Mol Genet Metab, 2009. 97(1): p. 53-9. 18. Muldoon, L.L., et al., Characterization of the molecular defect in a feline model for type II GM2- gangliosidosis (Sandhoff disease). Am J Pathol, 1994. 144(5): p. 1109-18. 19. Martin, D.R., et al., Mutation of the GM2 activator protein in a feline model of GM2 gangliosidosis. Acta Neuropathol, 2005. 110(5): p. 443-50. 20. Meurs, K.M., et al., A cardiac myosin binding protein C mutation in the Maine Coon cat with familial hypertrophic cardiomyopathy. Hum Mol Genet, 2005. 14(23): p. 3587-93. 21. Meurs, K.M., et al., A substitution mutation in the myosin binding protein C gene in ragdoll hypertrophic cardiomyopathy. Genomics, 2007. 90(2): p. 261-4. 22. Gandolfi, B., et al., First WNK4-hypokalemia animal model identified by genome-wide association in Burmese cats. PLoS One, 2012. 7(12): p. e53173. 23. Menotti-Raymond, M., et al., Mutation in CEP290 discovered for cat model of human retinal degeneration. J Hered, 2007. 98(3): p. 211-20. 24. Menotti-Raymond, M., et al., Mutation discovered in a feline model of human congenital retinal blinding disease. Invest Ophthalmol Vis Sci. , 2010. 51(6): p. 2852-9. 25. Lyons, L.A., et al., Feline polycystic kidney disease mutation identified in PKD1. J Am Soc Nephrol, 2004. 15(10): p. 2548-55. 26. Grahn, R.A., et al., Erythrocyte Pyruvate Kinase Deficiency mutation identified in multiple breeds of domestic cats. BMC Vet Res, 2012. 8(1): p. 207. 27. Fyfe, J.C., et al., An approximately 140-kb deletion associated with feline spinal muscular atrophy implies an essential LIX1 function for motor neuron survival. Genome Res, 2006. 16(9): p. 1084-90.

51 28. Owens, S.L., et al., Congenital adrenal hyperplasia associated with mutation in an 11beta-hydroxylase-like gene in a cat. J Vet Intern Med, 2012. 26(5): p. 1221-6. 29. He, X., et al., Identification and characterization of the molecular lesion causing mucopolysaccharidosis type I in cats. Mol Genet Metab, 1999. 67(2): p. 106-12. 30. Yogalingam, G., et al., Feline mucopolysaccharidosis type VI. Characterization of recombinant N- acetylgalactosamine 4-sulfatase and identification of a mutation causing the disease. J Biol Chem, 1996. 271(44): p. 27259-65. 31. Yogalingam, G., et al., Mild feline mucopolysaccharidosis type VI. Identification of an N- acetylgalactosamine-4-sulfatase mutation causing instability and increased specific activity. J Biol Chem, 1998. 273(22): p. 13421-9. 32. Crawley, A.C., et al., Two mutations within a feline mucopolysaccharidosis type VI colony cause three different clinical phenotypes. J Clin Invest, 1998. 101(1): p. 109-19. 33. Uddin, M.M., et al., Identification of Bangladeshi domestic cats with GM1 gangliosidosis caused by the c.1448G>C mutation of the feline GLB1 gene: case study. J Vet Med Sci, 2013. 75(3): p. 395-7. 34. Martin, D.R., et al., An inversion of 25 base pairs causes feline GM2 gangliosidosis variant. Exp Neurol, 2004. 187(1): p. 30-7. 35. Fyfe, J.C., et al., Molecular basis of feline beta-glucuronidase deficiency: an animal model of mucopolysaccharidosis VII. Genomics, 1999. 58(2): p. 121-8. 36. Kanae, Y., et al., Nonsense mutation of feline beta-hexosaminidase beta-subunit (HEXB) gene causing Sandhoff disease in a family of Japanese domestic cats. Res Vet Sci, 2007. 82(1): p. 54-60. 37. Somers, K., et al., Mutation analysis of feline Niemann-Pick C1 disease. Mol Genet Metab. , 2003. 79: p. 99-103. 38. Lettice, L.A., et al., Point mutations in a distant sonic hedgehog cis-regulator generate a variable regulatory output responsible for preaxial polydactyly. Hum Mol Genet, 2008. 17(7): p. 978-85. 39. Goree, M., et al., Characterization of the mutations causing hemophilia B in 2 domestic cats. J Vet Intern Med, 2005. 19(2): p. 200-4. 40. Clavero, S., et al., Feline congenital erythropoietic porphyria: two homozygous UROS missense mutations cause the enzyme deficiency and porphyrin accumulation. Mol Med, 2010. 16(9-10): p. 381-8. 41. Goldstein, R., et al., Primary Hyperoxaluria in cats caused by a mutation in the feline GRHPR gene. J Hered, 2009. 100(Supplement 1): p. S2-S7. 42. Clavero, S., et al., Feline acute intermittent porphyria: a phenocopy masquerading as an erythropoietic porphyria due to dominant and recessive hydroxymethylbilane synthase mutations. Hum Mol Genet, 2010. 19(4): p. 584-96. 43. Ginzinger, D.G., et al., A mutation in the lipoprotein lipase gene is the molecular basis of chylomicronemia in a colony of domestic cats. J Clin Invest, 1996. 97(5): p. 1257-66. 44. Geisen, V., K. Weber, and K. Hartmann, Vitamin D-dependent hereditary rickets type I in a cat. J Vet Intern Med, 2009. 23(1): p. 196-9. 45. Berg, T., et al., Purification of feline lysosomal alpha-mannosidase, determination of its cDNA sequence and identification of a mutation causing alpha-mannosidosis in Persian cats. Biochem J, 1997. 328 ( Pt 3): p. 863-70. 46. Grahn, R., et al., No bones about it! A novel CYP27B1 mutation results in feline vitamin D-dependent Rickets Type I (VDDR-1). in preparation, 2011. 47. Mazrier, H., et al., Inheritance, biochemical abnormalities, and clinical features of feline mucolipidosis II: the first animal model of human I-cell disease. J Hered, 2003. 94(5): p. 363-73.

Other References

1. Wurster-Hill, D.H. and C.W. Gray, Giemsa banding patterns in the chromosomes of twelve species of cats (Felidae). Cytogenet Cell Genet, 1973. 12(6): p. 388-97. 2. Bamber, R.C., Herdman, E.C., The inheritance of black, yellow and tortoiseshell coat colour in cats. J Genetics, 1927. 18: p. 87-97. 3. Little, C.C., Colour inheritance in cats, with special reference to colours, black, yellow and tortoiseshell. Journal of Genetics, 1919. 8: p. 279-290. 4. Ibsen, H.L., Tricolor Inheritance. Iii. Tortoiseshell Cats. Genetics, 1916. 1(377-86).

52 5. Doncaster, L., On the inheritance of tortoiseshell and related colours in cats. Proc Philosophical Soc, 1904. 13: p. 35-38. 6. Lyon, M.F., Sex chromatin and gene action in mammalian X-chromosome. American Journal of Human Genetics, 1962. 14: p. 135-148. 7. Schlafer, D.H., et al., A Case of SRY-Positive 38,XY True Hermaphroditism (XY Sex Reversal) in a Cat. Vet Pathol, 2011. 8. Ishihara, T., Cytological studies on tortoiseshell male cats. Cytologia, 1956. 21: p. 391-398. 9. Chu, E.H.Y., H.C. Thuline, and D.E. Norby, Triploid-diploid chimerism in a male . Cytogenetics, 1964. 3: p. 1-18. 10. Thuline, H.C., Male tortoiseshell, chimerism and true hermaphroditism. Journal of , 1964. 4: p. 2-3. 11. Pyle, R.L., et al., XXY sex chromosome constitution in a with tortoise-shell points. J Hered, 1971. 62(4): p. 220-2. 12. Gregson, N.M., Ishmael, J., Diploid triploid chimerism in three tortoiseshell cats. Research in Veterinary Science, 1971. 12: p. 275-279. 13. Centerwall, W.R., Benirschke, K., Animal model for the XXY Klinefelter's syndrome in man: tortoiseshell and calico male cats. American Journal of Veterinary Research, 1975. 36: p. 1275-1280. 14. Kosowska, B., et al., Cytogenetic and histologic studies of tortoiseshell cats. Medycyna Weterynaryjna, 2001. 57: p. 475-479. 15. Kuiper, H., M. Hewicker-Trautwein, and O. Distl, [Cytogenetic and histologic examination of four tortoiseshell cats]. Dtsch Tierarztl Wochenschr, 2003. 110(11): p. 457-61. 16. Modi, W.S., et al., Chromosomal localization of satellite DNA sequences among 22 species of felids and canids (Carnivora). Cytogenet Cell Genet, 1988. 48(4): p. 208-13. 17. Winand, N.J., et al., Deletion of the dystrophin muscle promoter in feline muscular dystrophy. Neuromuscul Disord, 1994. 4(5-6): p. 433-45. 18. Muldoon, L.L., et al., Characterization of the molecular defect in a feline model for type II GM2- gangliosidosis (Sandhoff disease). Am J Pathol, 1994. 144(5): p. 1109-18. 19. Louwerens, M., et al., Feline lymphoma in the post-feline leukemia virus era. J Vet Intern Med, 2005. 19(3): p. 329-35. 20. Yogalingam, G., et al., Mild feline mucopolysaccharidosis type VI. Identification of an N- acetylgalactosamine-4-sulfatase mutation causing instability and increased specific activity. J Biol Chem, 1998. 273(22): p. 13421-9. 21. Yogalingam, G., et al., Feline mucopolysaccharidosis type VI. Characterization of recombinant N- acetylgalactosamine 4-sulfatase and identification of a mutation causing the disease. J Biol Chem, 1996. 271(44): p. 27259-65. 22. Cannon, M.J., et al., Prevalence of polycystic kidney disease in Persian cats in the United Kingdom. Vet Rec, 2001. 149(14): p. 409-11. 23. Barrs, V.R., et al., Prevalence of autosomal dominant polycystic kidney disease in Persian cats and related- breeds in Sydney and Brisbane. Aust Vet J, 2001. 79(4): p. 257-9. 24. Barthez, P.Y., P. Rivier, and D. Begon, Prevalence of polycystic kidney disease in Persian and Persian related cats in France. J Feline Med Surg, 2003. 5(6): p. 345-7. 25. Biller, D.S., D.J. Chew, and S.P. DiBartola, Polycystic kidney disease in a family of Persian cats. J Am Vet Med Assoc, 1990. 196(8): p. 1288-90. 26. Eaton, K.A., et al., Autosomal dominant polycystic kidney disease in Persian and Persian-cross cats. Vet Pathol, 1997. 34(2): p. 117-26. 27. Lyons, L.A., et al., Feline polycystic kidney disease mutation identified in PKD1. J Am Soc Nephrol, 2004. 15(10): p. 2548-55. 28. Johnson, G., The Bengal Cat. 1991, Greenwell Springs, LA: Gogees Cattery. 29. Johnson, W.E., et al., The late Miocene radiation of modern Felidae: a genetic assessment. Science, 2006. 311(5757): p. 73-7. 30. Lipinski, M.J., et al., The ascent of cat breeds: genetic evaluations of breeds and worldwide random-bred populations. Genomics, 2008. 91(1): p. 12-21. 31. Kittleson, M.D., et al., Familial hypertrophic cardiomyopathy in maine coon cats: an animal model of human disease. Circulation, 1999. 99(24): p. 3172-80. 32. Meurs, K.M., et al., A cardiac myosin binding protein C mutation in the Maine Coon cat with familial hypertrophic cardiomyopathy. Hum Mol Genet, 2005. 14(23): p. 3587-93.

53 33. Sampedrano, C., et al., Prospective Echocardiographic and Tissue Doppler Imaging Screening of a Population of Maine Coon Cats Tested for the A31P Mutation in the Myosin-Binding Protein C Gene: A Specific Analysis of the Heterozygous Status. . J Vet Intern Med 2009. 23: p. 91–99. 34. Wess, G., et al., Association of A31P and A74T polymorphisms in the myosin binding protein C3 gene and hypertrophic cardiomyopathy in Maine Coon and other breed cats. J Vet Intern Med, 2010. 24(3): p. 527- 32. 35. Kohn, B. and C. Fumi, Clinical course of pyruvate kinase deficiency in Abyssinian and Somali cats. J Feline Med Surg., 2008. 10(143-53). 36. Eizirik, E., et al., Molecular genetics and evolution of melanism in the cat family. Curr Biol, 2003. 13(5): p. 448-53. 37. Peterschmitt, M., et al., Mutation in the melanocortin 1 receptor is associated with amber colour in the Norwegian Forest Cat. Anim Genet, 2009. 40(4): p. 547-52. 38. Lyons, L.A., et al., Chocolate coated cats: TYRP1 mutations for brown color in domestic cats. Mamm Genome, 2005. 16(5): p. 356-66. 39. Schmidt-Kuntzel, A., et al., Tyrosinase and tyrosinase related protein 1 alleles specify domestic cat coat color phenotypes of the albino and brown loci. J Hered, 2005. 96(4): p. 289-301. 40. Imes, D.L., et al., Albinism in the domestic cat (Felis catus) is associated with a tyrosinase (TYR) mutation. Anim Genet, 2006. 37(2): p. 175-8. 41. Lyons, L.A., et al., Tyrosinase mutations associated with Siamese and Burmese patterns in the domestic cat (Felis catus). Animal Genetics, 2005. 36(2): p. 119-26. 42. Ishida, Y., et al., A homozygous single-base deletion in MLPH causes the dilute coat color phenotype in the domestic cat. Genomics, 2006. 43. Gandolfi, B., et al., Off with the gloves: Mutation in KIT implicated for the unique white spotting phenotype of Birman cats. . submitted, 2010. 44. Gandolfi, B., et al., The Naked Truth: Sphynx and Devon Rex cat breed mutations in KRT71. Mammalian Genome, 2010. in press. 45. Drogemuller, C., et al., Mutations within the FGF5 gene are associated with hair length in cats. Anim Genet, 2007. 38(3): p. 218-21. 46. Kehler, J.S., et al., Four independent mutations in the feline fibroblast growth factor 5 gene determine the long-haired phenotype in domestic cats. J Hered, 2007. 98(6): p. 555-66. 47. Gandolfi, B., et al., To the Root of the Curl: A Signature of a Recent Selective Sweep Identifies a Mutation That Defines the Cornish Rex Cat Breed. PLoS One, 2013. 8(6): p. e67105. 48. Gandolfi, B., et al., A splice variant in KRT71 is associated with curly coat phenotype of Selkirk Rex cats. Sci Rep, 2013. 3: p. 2000. 49. Kaelin, C.B., et al., Specifying and sustaining pigmentation patterns in domestic and wild cats. Science, 2012. 337(6101): p. 1536-41. 50. Bighignoli, B., et al., Cytidine monophospho-N-acetylneuraminic acid hydroxylase (CMAH) mutations associated with the domestic cat AB blood group. BMC Genet, 2007. 8: p. 27. 51. De Maria, R., et al., Beta-galactosidase deficiency in a Korat cat: a new form of feline GM1- gangliosidosis. . Acta Neuropathol (Berl), 1998. 96: p. 307-14. 52. Bradbury, A.M., et al., Neurodegenerative lysosomal storage disease in European Burmese cats with hexosaminidase beta-subunit deficiency. Mol Genet Metab, 2009. 97(1): p. 53-9. 53. Martin, D.R., et al., Mutation of the GM2 activator protein in a feline model of GM2 gangliosidosis. Acta Neuropathol, 2005. 110(5): p. 443-50. 54. Meurs, K.M., et al., A substitution mutation in the myosin binding protein C gene in ragdoll hypertrophic cardiomyopathy. Genomics, 2007. 90(2): p. 261-4. 55. Gandolfi, B., et al., First WNK4-hypokalemia animal model identified by genome-wide association in Burmese cats. PLoS One, 2012. 7(12): p. e53173. 56. Menotti-Raymond, M., et al., Mutation in CEP290 discovered for cat model of human retinal degeneration. J Hered, 2007. 98(3): p. 211-20. 57. Menotti-Raymond, M., et al., Mutation discovered in a feline model of human congenital retinal blinding disease. Invest Ophthalmol Vis Sci., 2010. 51(6): p. 2852-9. 58. Grahn, R.A., et al., Erythrocyte Pyruvate Kinase Deficiency mutation identified in multiple breeds of domestic cats. BMC Vet Res, 2012. 8(1): p. 207. 59. Fyfe, J.C., et al., An approximately 140-kb deletion associated with feline spinal muscular atrophy implies an essential LIX1 function for motor neuron survival. Genome Res, 2006. 16(9): p. 1084-90.

54 60. Owens, S.L., et al., Congenital adrenal hyperplasia associated with mutation in an 11beta-hydroxylase-like gene in a cat. J Vet Intern Med, 2012. 26(5): p. 1221-6. 61. He, X., et al., Identification and characterization of the molecular lesion causing mucopolysaccharidosis type I in cats. Mol Genet Metab, 1999. 67(2): p. 106-12. 62. Crawley, A.C., et al., Two mutations within a feline mucopolysaccharidosis type VI colony cause three different clinical phenotypes. J Clin Invest, 1998. 101(1): p. 109-19. 63. Uddin, M.M., et al., Identification of Bangladeshi domestic cats with GM1 gangliosidosis caused by the c.1448G>C mutation of the feline GLB1 gene: case study. J Vet Med Sci, 2013. 75(3): p. 395-7. 64. Martin, D.R., et al., An inversion of 25 base pairs causes feline GM2 gangliosidosis variant. Exp Neurol, 2004. 187(1): p. 30-7. 65. Fyfe, J.C., et al., Molecular basis of feline beta-glucuronidase deficiency: an animal model of mucopolysaccharidosis VII. Genomics, 1999. 58(2): p. 121-8. 66. Kanae, Y., et al., Nonsense mutation of feline beta-hexosaminidase beta-subunit (HEXB) gene causing Sandhoff disease in a family of Japanese domestic cats. Res Vet Sci, 2007. 82(1): p. 54-60. 67. Somers, K., et al., Mutation analysis of feline Niemann-Pick C1 disease. Mol Genet Metab. , 2003. 79: p. 99-103. 68. Lettice, L.A., et al., Point mutations in a distant sonic hedgehog cis-regulator generate a variable regulatory output responsible for preaxial polydactyly. Hum Mol Genet, 2008. 17(7): p. 978-85. 69. Goree, M., et al., Characterization of the mutations causing hemophilia B in 2 domestic cats. J Vet Intern Med, 2005. 19(2): p. 200-4. 70. Clavero, S., et al., Feline congenital erythropoietic porphyria: two homozygous UROS missense mutations cause the enzyme deficiency and porphyrin accumulation. Mol Med, 2010. 16(9-10): p. 381-8. 71. Goldstein, R., et al., Primary Hyperoxaluria in cats caused by a mutation in the feline GRHPR gene. J Hered, 2009. 100(Supplement 1): p. S2-S7. 72. Clavero, S., et al., Feline acute intermittent porphyria: a phenocopy masquerading as an erythropoietic porphyria due to dominant and recessive hydroxymethylbilane synthase mutations. Hum Mol Genet, 2010. 19(4): p. 584-96. 73. Ginzinger, D.G., et al., A mutation in the lipoprotein lipase gene is the molecular basis of chylomicronemia in a colony of domestic cats. J Clin Invest, 1996. 97(5): p. 1257-66. 74. Geisen, V., K. Weber, and K. Hartmann, Vitamin D-dependent hereditary rickets type I in a cat. J Vet Intern Med, 2009. 23(1): p. 196-9. 75. Berg, T., et al., Purification of feline lysosomal alpha-mannosidase, determination of its cDNA sequence and identification of a mutation causing alpha-mannosidosis in Persian cats. Biochem J, 1997. 328 ( Pt 3): p. 863-70. 76. Grahn, R., et al., No bones about it! A novel CYP27B1 mutation results in feline vitamin D-dependent Rickets Type I (VDDR-1). in preparation, 2011. 77. Mazrier, H., et al., Inheritance, biochemical abnormalities, and clinical features of feline mucolipidosis II: the first animal model of human I-cell disease. J Hered, 2003. 94(5): p. 363-73.

55 Hereditary Gastric Cancer in Dogs Elizabeth McNiel, DVM, PhD, Tufts Cummings School of Veterinary Medicine, Tufts Medical Center Molecular Oncology Research Institute [email protected]

Stomach cancer (gastric carcinoma) is considered a rare cancer in dogs, an impression that is reinforced by published literature on this disease. Most papers feature fewer than 20 cases, thus providing a very limited sketch of this disease. Several years ago, we reported that Chow Chows have a significantly increased risk for gastric cancer and began to study the disease in this breed. Subsequently, we have expanded our research to include other breeds at high risk including the Belgian Tervuren, Belgian Sheepdog, Keeshond, Irish Setter, Bouvier, Norwegian Elkhound, Akita, and Scottish Terrier. Other breeds may also be at risk. We suspect that difficulty in accurately diagnosing dogs with this cancer may result in underreporting of its prevalence in dogs of all breeds, although in certain breeds the disease is quite common. Several years ago we established the Canine Gastric Cancer Repository and Database to provide a tool to learn about canine gastric cancer and to develop strategies to prevent, diagnose and treat this aggressive and nearly uniformly fatal disease. What is gastric cancer? Cancer develops from cells that grow uncontrollably and invade normal tissues. Cancer in the stomach can derive from a number of different cell types, therefore many types of cancer can occur in the stomach. However, most cancers in the stomach derive from the epithelium or lining cells and are called gastric carcinoma (or adenocarcinoma).Therefore, most of the time, gastric cancer is considered synonymous with gastric carcinoma. A variety of classification systems that are based on microscopic appearance and position of the cancer in the stomach have been used to classify stomach cancer in people. One of the traditional systems (the Lauren System) consists of two groups: Intestinal Type and Diffuse Type. Intestinal type forms lumps or masses on the surface while diffuse type invades directly into the wall of the stomach causing thickness without the development of a surface mass. While dogs appear to be capable of developing both types of gastric carcinoma, it appears that diffuse type is most common. A systematic review of the histology from more than 100 canine gastric cancer cases is currently underway. What causes gastric cancer in dogs? The short answer to this question is that we don’t know. However, in humans both environmental causes and genetics play a role. Environmental contributors including diet (salt and nitrites) and a bacterial organism called Helicobacter pylori. H. Pylori does not appear to naturally infect dogs. The occurrence of gastric cancer in particular breeds, strongly suggests that genetics are important in the canine disease. We have found families with multiple individuals affected over multiple generations which also support this notion. Other evidence for this includes the high prevalence of diffuse carcinoma in dogs which is associated with familial gastric cancer in people. The mode of inheritance is not clear. We are collaborating with Elaine Ostrander’s lab at the NHGRI to identify gene(s) that cause canine gastric carcinoma.

56 What are the signs of gastric cancer in dogs? The signs of gastric cancer are usually insidious, particularly in the early stages. Consistently, we see loss of appetite and weight loss. In many cases, we also see vomiting, although it may be very intermittent and easy to ignore because all else seems normal. Occasionally there is diarrhea or dark tarry stool which indicates intestinal bleeding. Because the signs of stomach cancer are very vague and nonspecific, it is unusual for veterinarians to see a dog until the disease is quite advanced. How is a diagnosis of gastric cancer made? A presumptive diagnosis of gastric cancer can often be made based on abdominal ultrasonography. The difficulty is that gas in the stomach interferes with this imaging. Furthermore, a distinct mass is often lacking in these cases. Thickness of the gastric wall may be the best indication of cancer. This underscores the importance of routinely evaluating wall thickness, particularly in dogs of high risk breeds with gastro-intestinal signs. Definitive diagnosis of gastric cancer is based on biopsy. The least invasive way to obtain a biopsy is with endoscopic biopsy. However, in many cases the diagnosis is missed on endoscopy. The inaccuracies in endoscopic biopsy stem from the nature of most stomach cancers in dogs in which the surface lining may look relatively normal even though there is substantial infiltration of the stomach wall by cancer cells. Therefore many times veterinarians cannot determine where biopsies should be collecting. Furthermore, endoscopic biopsies are may miss the cancer cells that are more deeply embedded. Finally, the areas of affected stomach often become very firm and almost rubbery in consistency and biopsying these areas may yield inadequate tissue. Surgical biopsy has the best chance of providing an accurate diagnosis, but this is, of course, more invasive. However, in addition to biopsy, there may also be the opportunity to surgically remove the cancer. How is gastric cancer treated? Surgery is the treatment of choice for gastric carcinoma. However, the removal a cancer from the stomach is challenging. When the cancer is very advanced involving a large proportion of the stomach removal is not usually feasible. The location of the cancer is also a deciding factor. In our experience, removal of the cancer is not always possible and it is uncommon for the surgeon to be able to remove it completely. Even partial removal can provide some relief to the dog, though this is not always true. We are aware of a single dog that lived for an additional 5 years following removal of a gastric tumor and several others that survived for a year or more. However, the vast majority of dogs will die of stomach cancer in time. Several chemotherapy drugs have been used in dogs with stomach cancer although the effectiveness of these is questionable. The future of the management of stomach cancer will rely on development of genetic tests to identify at risk individuals, selective breeding in some cases, development of screening tests, and the development of new agents that target the molecular constitution of stomach cancer. These types of advances are only possible with better understanding of the genetics and molecular biology of canine stomach cancer.

57 6th Tufts’ Canine and Feline Breeding and Genetics Conference

Scientific Program

Sunday, September 29

Lecture Time: Title of Lecture: Speaker:

8:10-8:50 Half A Century with Canine Hip Dysplasia Dr. Åke Hedhammar

8:50-9:30 The Othopedic Foundation for Animals Hip Displasia Database: Dr. G. Greg Keller A Review

9:50-10:30 The genetics of hip dysplasia and implications for selection Dr. Tom Lewis

10:30-11:10 Genetic and Genomic Tools for Breeding Dogs With Healthy Dr. Rory Todhunter Hips

12:55-1:35 Holistic Management of Genetic Traits Dr. Anita Oberbauer

1:35-2:15 From FUS to Pandora Syndrome - The Role of Epigenetics Dr. Tony Buffington and Environment in Pathophysiology, Treatment, and Prevention

2:35-2:55 Breed Specific Breeding Strategies Dr. Åke Hedhammar

2:55-3:15 UK initiatives for breeding healthier pedigree dogs Dr. Tom Lewis

3:15-3:55 Genetic Tests: Understanding Their Power, and Using Their Dr. Jerold Bell Force for Good

58 Half a Century with Canine Hip Dysplasia

Åke A Hedhammar, DVM, M Sc, Ph D, Dipl. Internal Medicine -Companion Animals Dept. of Clinical Sciences, Swedish University of Agricultural Sciences, Uppsala, Sweden

Malformations of the hips in dogs was described by Dr Schnelle already in 1937, but it was not until about a century ago that it was made evident that it was a widely spread entity affecting not only German shepherds.

Since then great attention and efforts have been paid worldwide by researchers as well as breeders to reveal the mechanisms behind it and to decrease its prevalence.

The following is an attempt to briefly review knowledge attained and achievements made over these years.

The perspective is from a veterinarian, neither a surgeon or radiologist nor a geneticist. It’s the perspective of someone involved in some research on the effect of genes and environment and as consultant to the Swedish kennel club in the institution of screening and breeding programs to assist in the selection of suitable breeding stock.. Neither surgical nor medical treatment will be covered. With reference to lectures to follow by Drs Keller, Lewis and Todhunter on the OFA database, genetics and genomic tools respectively those aspects will not be covered very extensively.

ETIOLOGY AND RISK FACORS

Definition of CHD “varying degree of laxity of the hip joint permitting subluxation during early life, giving rise to varying degrees of shallow acetabulum and flattening of the femoral head, finally inevitably leading to osteoarthritis."

Contrary to human hip dysplasia it soon became evident that Canine Hip Dysplasia (CHD) is a developmental and degenerative disease rather than congenital /present already at birth.

Knowledge We have learned that its structural and functional properties during its development and its clinical course by degenerative processes are affected by genes as well as environmental factors.

Achievements Based on that knowledge we have got various tools to handle genetic as well environmental factors in individuals and breed populations.

To be further known and achieved The true etiology bye genes involved and their expression is still to be revealed as well as the interaction between genes and environmental factors.

SCREENING PROCEDURES AND REGISTRIES

Screening for early signs of CHD have been proven to predict clinical outcome as well as genetic transmission (disposition for early signs as well clinically manifest CHD).

59 Knowledge Various screening methods have been investigated and validated for its purpose. Radiographic methods in standardized stressed or non stressed positioning have proven to be more useful than palpation as practiced in human HD. The outcome of any radiological screening is strongly affected by age at screening, positioning and level of sedation calling for standardizations of this parameters.

Achievements Based on that knowledge screening programs have been extensively established by various bodies worldwide.

Registries on results from these programs nowadays most commonly contain positive as well as negative results on permanently identified individuals, open to the public and accessible on line and supporting computerized information on national breed populations.

To be further known and achieved Earlier and more simple and inexpensive screening methods would enhance a more extensive usage.. An ongoing dispute regarding the value of different screening methods hampers the inclination to screen by any method. By computerization of results from many individuals the prediction of the genotype is much more accurate than any screening of an individual dog. Exchange of results from national registries is hampered by differences in procedure and recordings and calls for an international harmonization.

EPIDEMIOLOGY AND PREVALENCE

CHD do affect almost all large sized breeds to variable extent and not just purebred/pedigree dogs. The prevalence is affected by type of dogs, mollosoid dogs to high extent and scent hounds to less than other.

Knowledge The prevalence is known and documented in populations with extensive screening but most commonly restricted to clinically unaffected animals at age of screening.

Achievements A decreased prevalence is proven to be achieved in populations in which breeding stock i routinely have been selected for hip status by standardized measures.

To be further known and achieved The prevalence in most populations - pedigree and non pedigree - is still unknown and scarce regarding clinically affected individuals.

GENETICS AND BREEDING PROGRAMS

Based on known genetics and established screening programs structured breeding programs have been instituted in a couple of countries on national breed populations.

Knowledge The advancement of tools in both population genetics and molecular genetics have enhanced our knowledge on how to handle the selection of breeding stock with reference to CHD.

60 Achievements Structured breeding programs have proven to be very effective in purpose bred populations of dogs for police, armed forces and as guide dogs as well as in national populations on condition that a significant fraction of the population is screened and that the result is taken into account.

To be further known and achieved Breed specific breeding programs would beneficially account for breed variations in prevalence, population structure and other traits to take into account.

The extent of breeding programs - not just screening is needed to significantly affect the true prevalence in most national breed populations. International breeding programs would enhance the effect by exchange of results from different screening programs.

NUTRITIONAL IMPACT

The detrimental effect of over nutrition, i.e. excessive amounts of food (overfeeding) and excessive amounts of specific nutrients (over supplementation) have been proven for many orthopedics conditions in large sized breeds including CHD.

Knowledge Already at an early stage it was proven how ad lib feeding increase prevalence and severity of skeletal disorders including CHD. Contrary to earlier believe no specific nutrient can prevent from CHD by given in excessive amounts.

Achievements Feeding practices of large sized dogs have to some extent changed from “the more-the better” to feeding moderate amount of complete and balanced diet resulting in optimal skeletal conformation rather than maximal rate of growth.

To be further known and achieved Despite extensive promotion of chondro-protective products very little is still known on how to prevent from arthritis in developmental disorders as CHD.

SUMMING UP

By attention and efforts by researchers as well as cynological organizations and breeders worldwide extensive knowledge have been accumulated and effective tools have been developed to control for CHD. The full effect of this is however hampered by lack of a wider implementation.

A wider implementation of current screening methods and thereon based breeding programs is much more important than any refinement to reveal more of the phenotypic expression of CHD.

References and further readings

Schnelle GB. Congenital dysplasia of the hip in dogs. As referred in J Am Vet Med Assoc. 1959; 135(4):234-5

61 Henricson B, Olson SE. Hereditary acetabular dysplasia in German shepherd dogs. J Am Vet Med Assoc. 1959; 135(4):207-10.

Hedhammar Å, Wu Fu-ming, Krook L, Schryver HF, Lahunta A, Whalen JP et al. Overnutrition and skeletal disease. An experimental study in growing Great Dane Dogs. Cornell Vet 1974; 64 Suppl 5.

Kasström H. Nutrition, weight gain and the development of hip dysplasia. Acta Radiol 1975; 344 Suppl: 135- 79.

Hedhammar A, Olsson SE, Andersson SA, Persson L, Pettersson L, Olausson A, Sundgren PE. Canine hip dysplasia: study of heritability in 401 litters of German Shepherd dogs. J Am Vet Med Assoc. 1979 May 1; 174(9):1012-6. http://www.ncbi.nlm.nih.gov/pubmed/570968

Swenson L, Audell L, Hedhammar A. 1997 Prevalence and inheritance of and selection for hip dysplasia in seven breeds of dogs in Sweden and benefit: cost analysis of a screening and control program. J Am Vet Med Assoc. 1997; Jan 15; 210(2):207-14. http://www.ncbi.nlm.nih.gov/pubmed?cmd=Retrieve&db=PubMed&list_uids=9018354&dopt=Abstra ct

Hedhammar 1998 ACTIVITIES BY FEDERATION CYNOLOGIC INTERNATIONAL (FCI) TO COMBAT ELBOW AND HIP DYSPLASIA at the Website of the International Elbow Working Group. http://www.vet-iewg.org/joomla/index.php/archive/13-1998-bologna/20-hedhammar-1998

Hedhammar A; Swensson L; Egenwall A 1999 Elbow arthrosis and hip dysplasia in Swedish dogs as reflected by screening programmes and insurance data. The European journal of companion animal practice 1999; 9:2.

Hedhammar A. Nutrition and selection of breeding stock with reference to skeletal health in large growing Dogs - Swedish experiences over 25 years. Presented at IAMS Large Breed Health Care Symposium, Venice, Italy November 17th, 2001.

Sallander M, Hedhammar Å, Trogen M. Diet, excercise and weight as risk factors in Hip Dysplasia and Elbow Arthrosis in Labrador Retrievers. Journal of Nutrition 2006; 136:2050S-2052S. http://jn.nutrition.org/content/136/7/2050S.full

Malm S, Strandberg E, Danell B, Audell L, Swenson L, Hedhammar A. Impact of sedation method on the diagnosis of hip and elbow dysplasia in Swedish dogs. Prev Vet Med. 2007 Mar 17;78(3-4):196- 209.

Hedhammar A. Canine hip dysplasia as influenced by genetic and environmental factors The European journal of companion animal practice 2007; 17(2):141-143. http://www.docstoc.com/docs/80460727/Canine-Hip-Dysplasia-as-influenced-by-genetic-and- environmental

Comhaire FH, Snaps F.Comparison of two canine registry databases on the prevalence of hip dysplasia by breed and the relationship of dysplasia with body weight and height. Am J Vet Res. 2008 Mar; 69(3):330-3.

Malm S, Fikse F, Egenvall A, Bonnett BN, Gunnarsson L, Hedhammar A, Strandberg E. Association between radiographic assessment of hip status and subsequent incidence of veterinary care and mortality related to hip dysplasia in insured Swedish dogs. Prev Vet Med. 2010 Feb 1;93(2-3):222-32.

62 Wilson B, Nicholas F, Thomson P. Selection against canine hip dysplasia: Success or failure? The Veterinary Journal 2011; 189 (2011) 160–168. http://actualidadveterinaria.files.wordpress.com/2011/08/selection-against-canine-hip-dysplasia- success-or-failure.pdf

Dennis R. 2012 Interpretation and use of BVA/KC hip scores in dogs. In Practice 2012; April Volume 34: 178–194. Down loaded from inpractice.bmj.com on May 2, 2012. http://actualidadveterinaria.files.wordpress.com/2011/08/selection-against-canine-hip-dysplasia- success-or-failure.pdf

Fikse WF, Malm S, Lewis TW. Opportunities for international collaboration in dog breeding from the sharing of pedigree and health data. Vet J. 2013 Aug 8. pii: S1090-0233(13)00197-4. doi: 10.1016/j.tvjl.2013.04.025. [Epub ahead of print]

Hazewinkel HAW, Goedegebuure SA, Poulos PW, Wolvekamp WThC. Influences of chronic calcium excess on the skeletal development of growing Great Danes. J Am Anim Hosp Assoc 1985; 21: 377- 91.

Lavelle RB. The effects of overfeeding of a balanced complete commercial diet to a group of growing Great Danes I: Burger IH, RiversJPS, red. Nutrition of the dog and cat. Cambridge: Cambridge University Press 1989:303- 14.

Comhaire FH, Snaps F 2008 Comparison of two canine registry databases on the prevalence of hip dysplasia by breed and the relationship of dysplasia with body weight and height. Am J Vet Res. 2008 Mar;69(3):330-3.

63 The Orthopedic Foundation for Animals Hip Dysplasia Database: A Review Greg Keller, D.V.M, DACVR, Orthopedic Foundation for Animals, Inc., Columbia, MO [email protected]

The Orthopedic Foundation for Animals, Inc. (OFA) is a private not-for-profit foundation which formed a voluntary hip dysplasia control database in 1966 with the following objectives:

1.) To collate and disseminate information concerning orthopedic and genetic disease of animals. 2.) To advise, encourage and establish control programs to lower the incidence of orthopedic and genetic diseases. 3.) To encourage and finance research in orthopedic and genetic disease in animals. 4.) To receive funds and make grants to carry out these objectives. The OFA’s voluntary database serves all breeds of dogs and has the world’s largest all breed hip databank on radiographic evaluations of the hip. Due to breed variation by size, shape and pelvic conformation the OFA hip evaluation is based on comparison among individuals of the same breed and approximate age. Like most hip schemes the OFA employs the hip extended ventrodorsal view of the pelvis. Hip phenotypes are categorized as normal (excellent-1, good-2 and fair-3), borderline-4 and dysplastic (mild-5, moderate-6 and severe-7). Unlike most hip schemes the dog must be at least 24 months of age and the consensus evaluation is derived from three independent evaluations by board certified radiologists (1). Breed improvement, the reduction in hip dysplasia, is dependent on the degree of genetic variation within the breed, the accuracy of identifying a superior phenotype and the selection of pressure exerted upon the trait by individual and/or the breed club. There are numerous reports of dramatic reduction in hip dysplasia in closed populations (2, 3 & 4). The OFA database, even though the submissions are voluntary, has seen a similar improvement in most breeds (5). There is a strong correlation between the hip phenotype of the sire, dam and grandparents with a reduction in the prevalence for hip dysplasia in the progeny (6). Figure one

64 represents hip phenotype data on 490,966 progeny where the hip phenotype is also known on sire and dam. It is assumed that radiographs submitted to OFA are generally prescreened by the veterinarian and the more obvious cases of hip dysplasia are probably not submitted. Therefore the actual frequency of hip dysplasia in the general population is unknown, but has been approximated by Corley and Rettenmaier to be higher than reported by OFA (7 & 8). However, the main objective of the OFA is to identify phenotypically normal animals as potential breeding candidates in order to reduce the frequency of hip dysplasia. A review of the OFA hip database using a minimum of 5,000 evaluations yielded 44 breeds (Table 1). Regardless of the breed the general trend is for an increase in the percent excellent phenotype and a reduction in the percent dysplastic. Breed differences in the trend rate could be due to initial breed variations in hip dysplasia, the size of the gene pool and the selection pressure exerted by individuals and/or breed clubs. OFA is approaching 1.6 million individual hip records and the real power in this data is the ability for the public to access data through the Canine Health Information Center (CHIC) www.caninehealthinfo.org and primarily from the OFA website (www.offa.org).

To 1980 1981 to 1985 1986 to 1991 to 1995 1996 to 2000 2001 to 2005 2006 to 2010 Total 1990

AFGHAN HOUND Ex % 24.4 27.6 31.9 37.0 34.5 35.5 38.1 30.0 Dys % Total 5.4 5.6 6.4 4.6 6.6 5.2 3.8 5.4 dogs 2703 695 787 736 714 620 452 6720

AIREDALE TERRIER Ex % 4.6 5.8 7.3 8.7 8.6 6.2 8.8 7.4 Dys % Total 13.8 17.4 15.0 10.5 6.8 8.4 8.5 10.9 dogs 484 604 923 1056 1115 1057 624 5886

AKITA Ex % 7.5 11.3 15.5 20.7 29.3 31.8 33.2 19.3 Dys % 17.2 17.2 14.4 9.8 8.9 5.6 6.1 12.2 Total dogs 2047 2529 3366 3464 2219 1412 964 16047

ALASKAN MALAMUTE Ex % 10.7 15.2 17.9 17.9 24.1 21.3 23.9 17.1 Dys % 13.7 12.3 11.4 9.9 7.4 8.8 8.1 10.9 Total dogs 3547 2012 2263 2068 1686 1292 869 13777

AUSTRALIAN SHEPHERD Ex % 10.6 10.6 13.1 14.7 19.3 17.8 21.4 16.8 Dys % Total 7.8 7.3 5.9 5.2 4.6 5.0 4.6 5.3 dogs 2028 1992 3523 5573 6711 6552 5311 31885

BELGIAN TERVUREN Ex % 14.3 16.5 21.7 26.2 32.6 31.6 32.9 26.5 Dys % Total 5.1 4.8 3.5 3.0 2.3 2.3 2.9 3.2 dogs 610 559 807 981 1040 1069 763 5859

BERNESE MOUNTAIN DOG Ex % 2.9 4.1 7.1 10.5 15.2 15.8 20.3 14.1 Dys % Total 31.2 25.0 19.7 14.4 12.5 13.7 12.5 15.0 dogs 554 929 1708 2557 2918 4519 4092 17525

BORDER COLLIE Ex % 8.0 9.5 10.7 14.1 13.7 16.9 13.4 Dys % Total 21.2 14.3 13.1 11.4 9.7 9.7 8.2 10.2 dogs 99 426 957 1750 2420 2843 2499 11137

BOUVIER DES FLANDRES Ex % 3.1 3.8 5.3 6.7 7.3 6.7 12.4 6.4 Dys % Total 19.1 19.0 17.0 11.4 10.9 12.0 10.1 14.1 dogs 768 1119 1428 1648 1312 976 742 8033

BOXER Ex % 1.2 .9 3.7 3.3 4.4 2.7 5.5 3.6 Dys % Total 16.9 16.1 13.4 8.2 8.6 10.6 10.0 10.4 dogs 242 217 536 1075 1292 1241 789 5411

BRITTANY Ex % 5.9 5.7 6.4 8.2 12.1 11.0 13.7 9.0 Dys % Total 19.9 18.6 16.7 11.7 11.6 10.9 8.3 13.8 dogs 2632 1916 2667 3048 2916 2896 1950 18109

BULLMASTIFF Ex % 1.6 .7 3.0 3.6 6.1 3.3 5.7 4.0 Dys % Total 30.5 32.4 29.3 21.4 22.3 22.7 20.1 23.6 dogs 367 299 543 1276 1204 1043 716 5470

65

To 1980 1981 to 1985 1986 to 1990 1991 to 1996 to 2000 2001 to 2005 2006 to 2010 Total 1995

CAVALIER KING CHARLES Ex % 7.4 3.8 3.1 4.7 4.1 4.2 4.2 SPANIEL Dys % Total 33.3 12.4 8.9 10.8 9.9 11.7 11.8 11.2 dogs 15 162 425 732 1277 1967 1639 6275

CHESAPEAKE Ex % 6.1 7.0 10.3 11.0 16.6 18.3 19.9 12.7 BAY RETRIEVER Dys % Total 24.9 22.8 23.0 20.6 17.9 13.9 13.1 19.5 dogs 1633 1629 1769 2196 2182 1895 1268 12622

CHINESE SHAR-PEI Ex % 5.2 6.5 8.8 8.6 13.3 12.3 13.1 9.3 Dys % Total 21.5 19.6 14.2 9.3 8.9 9.4 7.6 13.0 dogs 135 1660 3588 1885 1040 756 487 9563

CHOW CHOW Ex % 4.5 4.3 6.3 9.0 10.6 8.7 14.2 7.4 Dys % 22.0 23.7 21.7 14.4 13.7 13.5 15.9 18.9 Total dogs 673 1042 1236 812 584 541 366 5266

COCKER SPANIEL Ex % 12.8 9.2 9.5 8.5 12.8 11.2 13.4 10.9 Dys % 8.1 8.9 7.0 5.4 4.4 5.4 5.4 5.9 Total dogs 531 1164 1967 2328 2636 2689 1608 12991

DOBERMAN PINSCHER Ex % 13.1 13.7 17.9 19.6 22.8 19.0 19.1 18.2 Dys % Total 7.7 7.6 5.6 4.4 5.0 5.0 5.0 5.6 dogs 2251 1541 2263 2787 2475 2182 1856 15406

ENGLISH COCKER SPANIEL Ex % 12.1 10.7 16.5 17.5 21.7 24.4 28.4 19.1 Dys % Total 7.3 6.7 5.3 4.0 5.2 5.0 4.9 5.3 dogs 713 805 1124 1207 1057 1143 843 6916

ENGLISH SETTER Ex % 3.3 3.5 7.9 10.5 14.1 15.4 20.1 10.8 Dys % Total 28.0 17.6 15.7 12.7 12.2 11.1 9.7 15.1 dogs 1423 1214 1405 1668 1690 1616 1212 10272

ENGLISH Ex % 7.2 5.2 7.3 9.0 10.0 9.9 13.1 9.1 SPRINGER Dys % Total 19.7 18.0 15.7 10.8 9.5 9.1 8.0 12.3 SPANIEL dogs 1791 1352 2029 2433 2443 2678 2007 14792

FLAT-COATED RETRIEVER Ex % 9.0 8.7 19.0 18.4 21.7 24.8 26.8 20.1 Dys % Total 3.6 3.5 7.0 4.2 3.4 3.0 3.0 3.8 dogs 333 540 675 933 1029 1051 866 5464

GERMAN SHEPHERD DOG Ex % 2.6 1.9 3.1 3.7 4.8 5.3 7.2 4.1 Dys % Total 20.3 22.1 20.5 17.2 16.5 17.8 17.9 18.7 dogs 11723 11679 17243 22022 16766 14525 10968 105443

GERMAN Ex % 19.8 18.4 22.2 24.2 29.9 28.3 33.4 26.3 SHORTHAIRED Dys % Total 6.9 7.6 5.6 3.7 3.0 2.8 1.7 3.9 POINTER dogs 1371 1196 2059 2866 2916 3096 2126 15711

GOLDEN RETRIEVER Ex % 1.9 2.1 2.8 4.1 5.4 5.8 8.9 4.4 Dys % Total 23.1 23.2 22.5 17.8 16.1 16.1 13.2 18.8 dogs 16621 16290 19966 23139 22386 19099 13839 132110

GORDON SETTER Ex % 4.1 4.4 8.2 9.4 13.1 12.0 19.1 9.1 Dys % Total 25.9 20.8 20.4 17.0 12.0 15.7 9.7 18.4 dogs 1051 923 1128 959 786 669 476 6025

GREAT DANE Ex % 6.5 8.7 12.3 13.5 15.0 11.9 13.6 11.8 Dys % Total 13.0 15.1 11.4 9.7 10.0 11.8 10.5 114 dogs 2050 981 1396 1750 2108 2426 1844 12625

GREAT PYRENEES Ex % 9.2 11.7 12.9 16.2 16.5 16.0 18.0 14.3 Dys % Total 9.3 10.5 9.9 8.3 8.3 6.9 7.9 8.7 dogs 774 759 1031 1134 920 756 494 5890

IRISH SETTER Ex % 5.1 6.5 8.6 10.3 13.8 14.1 17.9 9.3 Dys % Total 14.8 11.3 12.2 9.9 8.8 7.2 6.4 11.3 dogs 3740 1257 1413 1463 1295 1200 825 11226

LABRADOR RETRIEVER Ex % 10.8 11.5 15.2 17.4 20.3 20.3 25.3 18.5 Dys % Total 14.2 14.7 13.3 11.9 11.0 9.8 8.0 11.3 dogs 14088 16564 28242 43244 51208 45458 27973 228094

MASTIFF Ex % 2.5 5.4 5.1 6.3 9.7 9.2 10.2 8.1 Dys % Total 19.9 22.7 23.2 17.2 16.5 18.1 18.0 18.2 dogs 322 503 986 2249 2778 2350 1398 10626

NEWFOUNDLAND Ex % 3.5 2.9 6.7 8.4 10.6 11.8 15.1 8.7 Dys % Total 31.6 31.4 25.8 21.4 23.1 21.3 20.3 24.4 dogs 1696 1709 2338 2514 2420 2561 1780 15094

OLD ENGLISH SHEEPDOG Ex % 7.8 9.4 13.8 15.5 17.7 17.2 21.8 11.9 Dys % Total 22.9 18.8 18.0 12.9 11.3 10.3 10.0 18.0 dogs 4434 1438 1299 1091 947 816 578 10637

66

To 1980 1981 to 1985 1986 to 1990 1991 to 1995 1996 to 2000 2001 to 2005 2006 to Total 2010

PEMBROKE WELSH CORGI Ex % 1.1 1.9 2.4 5.5 3.7 3.1 3.3 Dys % 19.2 20.7 18.1 14.3 13.7 17.3 18.1 16.5 Total dogs 99 787 1583 2056 2257 2275 1598 10733

POODLE Ex % 8.0 7.2 9.1 11.0 14.1 13.1 16.6 12.1 Dys % 17.4 14.1 14.2 9.9 9.9 9.6 8.7 11.2 Total dogs 2273 1821 2442 3401 4058 4945 3799 22908

PORTUGUESE WATER DOG Ex % 1.0 5.0 9.6 11.7 16.2 14.6 20.6 14.8 Dys % 26.2 22.8 16.1 11.3 10.1 9.5 8.7 11.1 Total dogs 103 303 795 1271 1539 1955 1739 7790

RHODESIAN RIDGEBACK Ex % 13.9 19.1 19.2 23.4 24.7 23.3 27.2 22.3 Dys % 11.8 8.5 5.9 2.9 2.9 2.9 2.5 4.6 Total dogs 1123 993 1450 1879 1968 2064 1575 11130

ROTTWEILER Ex % 4.3 5.2 7.4 9.2 13.0 11.9 14.2 8.4 Dys % 23.4 22.8 22.0 17.6 15.8 15.8 15.1 19.8 Total dogs 5901 16624 27339 22597 8973 6123 4921 92692

SAMOYED Ex % 8.3 7.4 9.1 10.9 13.1 14.4 19.1 10.8 Dys % 13.4 12.2 10.8 8.8 7.1 8.7 6.9 10.4 Total dogs 3767 2460 2593 2090 1787 1671 1416 15862

SHETLAND SHEEPDOG Ex % 26.0 29.7 25.8 27.6 28.7 25.3 32.5 27.9 Dys % 5.6 9.8 6.2 3.8 4.0 3.5 2.9 4.2 Total dogs 515 549 2676 4254 4640 4225 2876 19842

SIBERIAN HUSKY Ex % 24.6 28.7 37.6 38.0 41.6 39.1 41.4 34.0 Dys % 2.7 2.2 1.9 1.2 1.2 1.5 1.2 1.8 Total dogs 4453 2495 2588 2616 2102 1759 1396 17470

SOFT COATED WHEATEN Ex % 12.1 12.1 16.1 16.6 22.2 16.9 23.1 17.3 TERRIER Dys % 6.9 4.5 4.5 2.5 4.0 4.9 5.0 4.4 Total dogs 504 877 964 1030 1072 892 624 5992

VIZSLA Ex % 12.7 12.2 14.0 16.5 19.6 17.9 21.0 17.1 Dys % 10.4 8.0 7.4 4.7 6.0 5.9 4.4 6.3 Total dogs 1527 1068 1407 2004 2468 2865 2098 13547

WEIMARANER Ex % 13.1 12.6 19.8 22.4 25.8 24.7 28.1 21.6 Dys % 11.8 11.1 8.5 7.2 5.9 7.7 5.5 7.9 Total dogs 1586 1040 1514 2252 2410 1959 1138 11946

1.) Keller, GG et al: The Use of Health Databases and Selective Breeding 2.) Hedhammer A, Olsson SE, et al: Study of Heritability in 401 Litters of German Shepherd Dogs. JAVMA, Vol. 1974; 1012-1016, 1979. 3.) Swensen L et al: Prevalence and inheritance of and Selection for Hip Dysplasia in Seven Breeds of Dogs in Sweden and Benefit: Cost Analysis of a Screening and Control Program. JAVMA, 210:2, 1997, pp 207-214. 4.) Leighton EA: Genetics of Canine Hip Dysplasia. JAVMA, Vol. 210, No. 10, 1997, pp. 1474-1479. 5.) Kaneene JB et al: Update of a Retrospective Cohort Study of Changes in Hip Joint Phenotype of Dogs Evaluated by the OFA in the United States, 1989 – 2003. Veterinary Surgery, 38: 398-405, 2009. 6.) Keller GG, et al: How the Orthopedic Foundation for Animals (OFA) is tackling inherited disorders in the USA: Using hip and elbow dysplasia as examples. The Veterinary Journal (2011), doi: 10.1016/j. tvjl.2011.06.19 7.) Corley EA, et al: Reliability of Early Radiographic Evaluation for Canine Hip Dysplasia Obtained from the Standard Ventrodorsal Radiographic Projection. JAVMA, 211:9, 1997, pp. 1142-1146. 8.) Rettenmaier JL, Keller GG, et al: Prevalence of Canine Hip Dysplasia in a Veterinary Teaching Hospital Population. Vet. Rad. & Ultrasound, Vol. 43, No. 4, 2002, p. 313-318.

67 The Genetics of Hip Dysplasia and Implications for Selection Tom Lewis PhD, Animal Health Trust, Lanwades Park, Kentford, Newmarket, Suffolk, UK [email protected]

Hip dysplasia is a complex disease. When a trait or disease is described as ‘complex’ it is usually meant that the trait is influenced by both genetic and non-genetic, or environmental, effects. This makes it very difficult to determine the mode of inheritance, since the phenotype (the observable manifestation of the trait) is not necessarily an accurate indicator of the genetics; the genetics is only a part of the picture and is ‘overlaid’ by environmental influences (‘good’ genetics may be masked by detrimental environment and vice versa). Furthermore, the trait is often under the control of more than one (and usually many) genes meaning that we can no longer categorise individuals as clear, carriers or affected. In this lecture I will attempt to demonstrate the presumed genetic architecture of complex traits and show how we can achieve more effective selection in spite of the problems bequeathed by this complexity.

You are probably all familiar with Mendelian inheritance which Gregor Mendel demonstrated with a 3:1 ratio of yellow to green peas. This ratio allowed him to infer that the trait of pea colour was determined by 2 variants (alleles) at a single gene; the yellow allele (A) being ‘dominant’ and the green allele (a) being ‘recessive’. This meant that the two possible phenotypes of pea colour (yellow and green) were in fact produced by three possible genotypes. Homozygotes (so called as both alleles are the same variety) with two yellow alleles (AA) produced yellow peas and those with two green alleles (aa) produced green peas. Heterozygotes having one green and one yellow allele (Aa) were yellow in appearance (phenotype), the dominant yellow allele masking the recessive green. This is important since it allows phenotypic variations, in this example the green pea colour, to apparently disappear for a number of generations before suddenly reappearing.

Heterozygotes produce half their gametes (sex cells) with the A allele and half with the a allele. Therefore progeny of two heterozygotes will have the genotypes AA : Aa : aa in the ratio 1 : 2 : 1, but because the yellow allele A is dominant the phenotypic ratio is 3:1 (see figure 1). However, this 1 : 2 : 1 ratio is very important – because the ‘dominance’ we have encountered up to this point is not universal or complete across all traits or diseases.

Gametes from parent 1 (Aa) A a

A AA Aa (A )

a Aa aa Gametes from parent 2 parent Gametes from

Figure 1: Punnett square showing the genotypes and phenotypes from crossing two heterozygote parents.

68 Consider for a moment (hypothetically) that gene A (with 2 alleles A and a) determines the quantity of peas rather than their colour. So AA might yield 9 peas in each pod, while aa yields only 3. If the A allele is completely dominant, then we expect the heterozygote to show the same phenotype as the dominant homozygote; so in this hypothetical example Aa yields 9 peas per pod. However, as mentioned above, dominance is not universal or always complete. For example, imagine that instead the heterozygote yielded 6 peas per pod – half way between the two homozygotes. We can begin to look at things more quantitatively, plotting the number of peas per pod against the number of A alleles:

Figure 2: hypothetical examples of complete dominance (L) and completely additive (R) A allele.

These are two important examples; the chart on the left in figure 2 shows complete dominance, i.e. the heterozygote (Aa) is the same phenotypically as the AA homozygote. The chart on the right shows no dominance, or complete additivity (i.e. the heterozygote is the intermediate of the two homozygotes, and each A allele adds 3 peas per pod). Additivity is an important concept as we move on to consider ‘genetic variation’.

Complete additivity at a single gene will give us a 1 : 2 : 1 ratio of phenotypes (reflecting the genotype ratio). However, as stated earlier, many quantitative or complex traits are influenced by multiple genes. As we consider perfect additivity over an increasing number of genes (figure 3) we can see the phenotypic distribution (discounting non-genetic effects for a moment) approaching a ‘normal distribution’ (also known as the ‘bell curve’, and very important in statistics). Figure 3 shows (from left to right) the genetic distributions of traits controlled by 1, 3 and 6 genes respectively, followed by a normal distribution on the far right. Hopefully you can see that increasing the number of genes increases the number of phenotypic categories and begins to produce continuous genetic variation for the trait or disease in question. Thus, we have moved from thinking in terms of ‘clear’, ‘carrier’ and ‘affected’, to thinking in terms of a continuous scale of liability or risk.

69

Figure 3: (L to R) genotype frequency distributions for 1, 3 and 6 completely additive genes, and a normal distribution (far right).

This is probably not as novel a concept as it may appear; think about when you hear news reports about scientists having found a gene for cancer, heart disease, diabetes, Alzheimer’s etc. – it’s always a gene, not the gene. There isn’t a single gene for any of these diseases just as there isn’t a single gene for height or weight. So, for complex diseases like hip dysplasia we will have to deal with the concept of genetic variation and risk.

But the complexity doesn’t end here. As mentioned at the outset complex traits are influenced by both genetic and environmental factors. While the genes (and so the genetic risk) are determined at conception, this risk is subsequently modified by the effects of numerous known and unknown non-genetic or environmental influences. Think of heart disease; I may have a moderate genetic risk, but if I smoke, eat a poor diet, eat too much, take no exercise and have a stressful lifestyle my actual risk creeps up. My actual risk when I’m 50 may be higher than a 50 year old with a higher genetic risk, but who watches their weight, eats healthily, has never smoked and has a low stress lifestyle. The same is true for hip dysplasia, where known environmental effects include diet and early-life exercise regime.

Nevertheless, genetics makes an important contribution to the overall risk. The heritability of a trait tells us how important genetics is relative to non-genetic effects – strictly it is the proportion of phenotypic variation that is due to genetic variation. For hip dysplasia about 40% of the overall observable variation is due to genetic variation. This may not seem much - it is less than half after all - but it is by far the biggest single component.

However, when it comes to breeding, it is only the genetic risk we are concerned with, as it is only genetics that is passed across generations. This presents us with a problem – we are using phenotypes (hip scores) to guide our selections, but we know that they are not necessarily the best guide to genetics. We may unwittingly choose a dog with a good hip score, not knowing that this is actually more to do with a beneficial environment and that the genetic risk, which is passed to the progeny, is actually fairly high. But to date, hip scores are all that breeders have had to guide them.

This is where estimated breeding values, or EBVs, come in. EBVs are a quantitative estimate of the true genetic risk, or breeding value. We make the estimate using trait information, in the case of hip dysplasia using the hip score, on an individual and all its relatives. We are able to do this thanks to the availability of pedigree information, which allows us to quantify the relationship between all the dogs therein. Information on relatives, who share genes to a quantifiable degree, will allow us to make a better judgement on an individual’s genetics. For example, we may feel very differently about using a stud dog with a poor hip score if we knew that he had over 50 progeny scored with a very good average hip score. The performance of the progeny tells us about the genetics of the parent. In fact, this aspect has been key to the success of EBVs, which have been extensively used in the industry for over 20 years. Here we are concerned with milk production traits – traits that are only

70 expressed in females. Yet we have very accurate EBVs for dairy bulls based on the milking performance of thousands of their daughters. Somewhat paradoxically, we know more about a bull’s genetics with respect to milking traits than we do for any cow!

As with all estimates, it is useful to know how good an estimate the EBV really is. It is intuitive that we will have more confidence in the genetics of the stud dog mentioned above, with his own hip score and scores of 50 progeny known, than a stud dog with no information on itself or its progeny. Just as EBVs are a quantitatively formal way of taking account of relatives’ information in the assessment of an individual’s genetic liability, so we can formally calculate the accuracy of our estimate of true genetic liability (the EBV).

So, EBVs are a more accurate indicator of genetics than an individual phenotype – and are more abundant. Because we calculate the accuracy of each EBV we are able to quantify how much more accurate selection using EBVs will be than selection using phenotype; with more accurate selection delivering greater genetic progress. Results from research show that selection using EBVs is an average of 1.16 times more accurate than using hip scores, across 16 breeds (Lewis et al, 2013). Furthermore, EBVs are available for all animals in the pedigree. Selection using EBVs of dogs too young to have their own hip score is on average 1.30 times more accurate than selection using the parental phenotypes (Lewis et al, 2013). We also showed that a far greater proportion of animals had an EBV more accurate than knowing both parental hip scores, than actually had both parental hip scores known, demonstrating that EBVs are an effective way of providing more reliable information on a far greater proportion of the breed or population.

Further improvements in the accuracy of selection have been shown to be available from the way we use the phenotypes. For example, we have shown with selection index methodology that for Labradors EBVs for elbow dysplasia score are up to 10% more accurate when computed from a bivariate analysis of elbow and hip scores than from a univariate analysis of elbow scores alone. A positive genetic correlation between hip and elbow score means that hip score acts as a much more abundant, if slightly less accurate, indicator of elbow dysplasia (Lewis et al, 2011). This method was also used to determine a more effective combination of the nine component traits of the UK hip score than a simple aggregate total (Lewis et al, 2010b), again delivering more accurate selection.

Finally, it is important to remember that EBVs are simply a more effective way of using the hip score data in selection – they are NOT a direct replacement! Quality data is critical for the calculation of accurate EBVs. Furthermore, hip scores themselves have significant prognostic value for individual dogs.

References: Lewis, T.W., Blott, S.C., Woolliams, J.A.W. (2010a) Genetic evaluation of hip score in UK Labrador Retrievers. PLoS ONE, 5(10): e12797 http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0012797

Lewis, T. W., Woolliams, J.A.W, Blott, S.C. (2010b) Genetic evaluation of the nine component features of hip score in UK Labrador Retrievers. PLoS ONE, 5(10): e13610 http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0013610

Lewis, T.W., Ilska, J.J., Blott, S.C., Woolliams, J.A.W. (2011) Genetic evaluation of elbow scores and the relationship with hip scores in UK Labrador Retrivers. The Veterinary Journal 189: 227-233. http://www.sciencedirect.com/science/article/pii/S1090023311002383

71

Lewis, T.W., Blott, S.C., Woolliams, J.A.W. (2013) Comparative analysis of genetic trends and prospect of selection against hip and elbow dysplasia in 15 UK dog breeds. BMC Genetics 14:16 http://www.biomedcentral.com/1471-2156/14/16

72 Genetic and Genomic Tools for Breeding Dogs with Better Hips Rory J. Todhunter, BVSc, MS, PhD, DACVS, Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, Ithaca NY [email protected]

Canine hip dysplasia (CHD) is a developmental trait primarily affecting medium and large breed dogs. CHD is characterized by faulty conformation and laxity of the hip joint that usually affects both hips. Clinically, the osteoarthritis that results from hip dysplasia is characterized by hind limb lameness, reduced exercise tolerance, reluctance to jump, poor hind limb muscle mass, and laxity or pain in the hip joint. CHD can be detected radiographically as subluxation of the affected hip. CHD results in synovitis accompanied by effusion and osteoarthritis of the affected joint. Osteoarthritis is detected on a radiograph as osteophytes around the femoral neck (so-called Morgan’s line), femoral head and acetabulum and a flattening of the femoral head with a shallow, open acetabulum. Radiographs are insensitive to the presence of incipient osteoarthritis of upper limb joints like the hip. In addition, the radiographic alterations associated with hip dysplasia can be subtle and even an “unaffected” dog, as assessed by a radiograph, can still carry some of the mutations that contribute to the disorder.

Genetics

CHD in dogs is an inherited, polygenic trait in which mutations in several genes [the regions where these genes reside in the genome are called quantitative trait loci (QTLs)] are involved in its clinical expression. The molecular genetic basis of CHD is currently unknown. Many dogs with normal hips on radiographs carry at least a modicum of the trait-causing mutations but not all that are necessary to cause physical expression of the trait. CHD is a quantitative or complex trait that is expressed as a continuum from imperceptible to severe forms. This continuum in trait expression observed as the hip phenotype represented in the radiographic image is due to the additive nature of the genes and their alleles that underlie the trait. Some alleles increase trait expression and some contribute resistance to the trait. This continuum of trait expression is affected by environmental influences (such as plane of nutrition and exercise, and many unknown epigenetic factors) which interact with the genetic constitution to affect the degree to which the trait is manifested. CHD has a heritability between 0.25-0.7 depending on the pedigree in which it is estimated and the method used to measure the trait. We have found one gene, fibrillin 2, which has a deletion that segregates with CHD across several breeds represented in our genetic banking archive. Other candidate genes are under investigation. We recently genotyped about 1,000 dogs on the High Density Illumina canine mapping array in order to find markers and genes associated with CHD. Eventually, these genetic mapping experiments will lead to discovery of the mutations that contribute to CHD.

It will take a concerted effort to rid breeds of the genetic mutations that contribute to CHD expression or conversely, to introduce protective alleles at the loci that cause good hips. Breeding two dysplastic dogs can yield a 75% incidence of CHD in offspring, while mating two unaffected dogs can yield a 25% incidence. Selective breeding using normal dogs from normal parents and grandparents, as well as progeny testing, should decrease the incidence of CHD. The message here is that until we have a genetic test for CHD so we can detect genetically susceptible dogs, the best indication of a dog's genetic makeup is where it came from (its' parents and grandparents), what it produces (its' offspring), and the phenotype of its' siblings or half sibs.

73 To test whether a sire or dam carries mutations for CHD (even if the dog has OFA-good hips), it should be bred to sires or dams with good hips and the proportion of affected offspring recorded (progeny testing). As many as 20 offspring would be needed to be reasonably estimate a dog’s genetic value for CHD.

Hip Conformation Screening

Because it is an inherited trait, the traditional strategy to control CHD has been through establishment of registries. Registries can be voluntary or involuntary and each has its detractors. In the USA, the Orthopedic Foundation for Animals (OFA) has provided a standard for radiographic evaluation of hips based on breed, age, and conformation. Radiographs from dogs under 2 years of age are given a provisional assessment of hip status and a definitive hip certification is given to dogs 2 years or older. The OFA has far surpassed their one millionth radiographic submission. Radiographic changes related to the osteoarthritis associated with CHD may not be detected until two years of age or older. The sensitivity of the OFA radiograph at 12 months of age for detecting later development of osteoarthritis in affected hips ranges from 77- 99% depending on the severity of the CHD at the earlier age. Ninety to 95% of dysplastic dogs have changes associated with CHD at 12 months of age. However, another study showed that of all dogs developing hip osteoarthritis over their life span, only 53% had radiographic evidence of a CHD at 2 years of age. Radiographs are insensitive to the presence of incipient osteoarthritis in the hip. The joint is considered dysplastic when the femoral head conforms poorly to the acetabulum or there are remodeling boney changes at the capsular attachment to the acetabulum or femoral neck. Hip status is graded on a scale from excellent conformation to severe hip dysplasia; there are 7 grades in all.

The PennHIP™ radiographic method measures the maximum amount of lateral (distraction) hip joint laxity (distraction index). There is a positive relationship between the distraction index and subsequent development of osteoarthritis. PennHIP radiographs include an OFA style ventrodorsal projection, a compression and a distraction projection. The OFA style film is evaluated for hip congruity and osteoarthritis. Labrador Retrievers with a low distraction index (less than 0.3) at 8 months of age have about a 90% chance of being normal while those with distraction indices greater than 0.8 have about a 90% chance of being dysplastic and succumbing to secondary hip osteoarthritis. Most breeds have similar ranges and relationships between the distraction index and the development of hip osteoarthritis. When choosing between dogs for breeding, preferentially breed dogs with the lowest distraction indices of the available pool. The optimum age for PennHIP™ screening is at early maturity (8-12 months of age for medium to large breed dogs).

A radiographic imaging position called the dorsolateral subluxation (DLS) test was developed at Cornell with an eye to improving the accuracy of hip evaluation. The PennHIP™ method finds dogs with laxity (a risk factor for CHD) but not all dogs with hip laxity develop secondary hip osteoarthritis. Their hips presumably function normally when they ambulate. We developed a method in which the hips are imaged in their normal functional position. The dogs are imaged in ventral recumbency under heavy sedation (or general anesthesia). The stifles are flexed and positioned under the hips so that the ischiatic table is superimposed over the stifles. The DLS score equates with the proportion of the femoral head covered by the dorsal acetabular

74 rim. We compared the sensitivity (the percentage of dogs with osteoarthritis that were correctly identified) and specificity (percentage of dogs without osteoarthritis that were correctly identified) between the OFA-like extended-hip radiograph, the distraction index, and the dorsolateral subluxation score (DLS) score. For a single test, the DLS score is the most accurate in detection of both affected and unaffected dogs. A combination of the DLS score and Norberg angle gave the best estimate of a dog’s likelihood of developing subsequent osteoarthritis than any single test, including the DLS test. The Norberg angle is a measure of femoral head coverage on the OFA style extended hip radiograph. The Norberg angle ranges from near zero for severely subluxated hips to about 120º in the “best” hipped dogs. An angle over 105º seems to be preferable. So we are back to needing two methods to adequately describe hip conformation. This conclusion is supported by principal component analysis of the OFA score, the Norberg angle, the DI, and the DLS score to measure hip conformation.

My recommendation for selection of a young to early mature pet dog (not for breeding) with optimal hip quality is to examine the dog, palpate the hips for pain and Ortolani sign (under sedation), and to confirm physical findings with at least one radiograph – either a DLS or an extended hip radiograph. If the extended-hip (OFA style) image demonstrates subluxation (dysplasia) then no need to go further. If the dog has a “normal” extended hip projection but a positive Ortolani test, then that dog is at least susceptible to secondary osteoarthritis, if not dysplastic, and to document the laxity, should have a laxity imaging projection like the DLS method. A dog with optimal hip conformation should have palpated normally and have a DLS score over 55% or a DI below 0.4.

The dilemma is how to select puppies with optimum hip conformation. For those breeds with moderate to high risk for hip dysplasia, select pups from breeders where rigorous selection practices are employed (phenotypes recorded on both the sire and dam lines) so the buyer can review the breeding history. Both sides of the pedigree should be available. Information about results of previous breedings is very helpful (roughly 15-20 offspring of the same parents must be evaluated before one can have reasonable knowledge of the genetic quality of the same parents). Application of estimated breeding values (EBVs) for CHD will result in faster gain than basing breeding decisions on phenotype alone. Finally, marker-assisted selection will improve genetic quality for complex or polygenetic traits like hip dysplasia far faster them breeding “better than the average” and can replace the use of EBVs for those who can’t access them or estimate them with much accuracy (which is the case for most breeders not aligned with Service Organization like the Seeing Eye or Guiding Eyes for the Blind or military establishments).

Breeding Values for US Pure Breed Dogs Derived from the OFA Public Data Base

The breeding value in its earliest use was also called the selection index. The selection index was based on integration of genetic (pedigree relationships) and phenotypic information (OFA hip scores in our case) from each animal and its relatives and yields better results than phenotypic selection alone for desirable traits. The accuracy of the selection index of a subject increases when the OFA scores from its close relatives (e.g. progeny and ancestors) are included in the estimation. The selection index was developed into the Best Linear Unbiased Prediction (BLUP). The BLUP breeding strategy has been used successfully for genetic improvement, particularly in , and has also been applied in closed colonies of dogs with substantial

75 success. Variance components attributable to additive genetic and residual effects were estimated for the OFA hip and elbow scores and pedigrees. Genetic parameters, including the additive genetic variance and the residual variance were estimated using the REML procedure. Heritability (h2) is the proportion of additive variance over the total variance which is the sum of additive variance and residual variance. The general concept is to select dogs with the lower EBVs as these are the individuals with the lowest or best hip and elbow conformation.

We derived EBVs (a measure of a dog’s genetic potential to produce offspring with optimal characteristics for an inherited trait) and inbreeding coefficients for CHD in Labrador retrievers based on OFA hip scores in the OFA database and provided them to the public in 2010 (http://www.vet.cornell.edu/research/bvhip/). The OFA hip scores and pedigrees of the Labrador retrievers in the public data base were used for the genetic evaluation. Dogs were scored by the OFA radiologists into seven categories: excellent, good, fair, borderline, mild, moderate and severe hip dysplasia. The first three categories (excellent, good and fair) are generally considered “normal” dogs although they will carry some of the mutations that contribute to hip dysplasia. The last three categories (mild, moderate and severe) are considered “dysplastic” dogs. This analysis was undertaken independently of the OFA. The seven hip score categories were replaced with 7 numerical scores, starting with excellent as 1 and ending with severe as 7. A numerical value of 2 was assigned to the combined category of “normal”. Our analysis of the Labrador Retriever OFA hip breeding values over that period showed that there has been slow but consistent genetic improvement (Hou et al., PLoS One 2010). The explanation and methods that form the basis of the breeding values available in the search page of this web site were published in the American Journal of Veterinary Research in 2008 by Zhang et al. and a PDF of that paper is available in the publication section of my research home page in Clinical Sciences at Cornell University.

Since 1974, the Orthopedic Foundation for Animals (OFA) has provided a voluntary registry where the scores of hip and elbow radiographs of individual dogs and their pedigrees have been deposited. Following on from the research we reported on Labrador Retriever hip EBVs, we calculated estimated breeding values (EBVs) and inbreeding coefficients for a total of 1,264,422 dogs from 74 breeds which included at least 1,000 individuals. The analysis was performed with a bi-variate (used both hip and elbow scores) mixed model across these 74 breeds to improve the accuracy of the EBVs, to compensate for the deficiency in voluntarily reporting bias in the OFA public registry, and to provide an estimate of genetic correlation between the hip and elbow scores. There were 760,455 and 135,409 dogs with their own hip and elbow scores, respectively. The incidences of CHD and elbow dysplasia were 0.83% and 2.08% across the 74 breeds (21 breeds for elbow dysplasia) and ranged from 0.07% to 6%, and 0.5% to 8% within breeds, respectively. These incidences were far lower than the incidence reported in the hip and elbow dysplasia summary statistics by breed in the OFA web page (http://www.offa.org/stats_hip.html). The heritability of hip and elbow scores was estimated at 0.23 and 0.16, respectively. Over the 40 years since 1974, the genetic improvement for hip scores was 0.1 hip units or 16.4% of the average phenotypic standard deviation across the 74 breeds, which corresponded to a drop in the overall incidence of CHD of 3.37% clinically. For elbow scores, the genetic improvement was 0.0021 elbow units or 1.1% of the phenotypic standard deviation across the 21 breeds. Both genetic improvements were likely underestimated due to the inevitable bias against reporting osteoarthritic records. Genetic change in EBVs for

76 hip and elbow scores was breed specific; some breeds improved their genetic quality, some demonstrated little improvement, while in a few breeds, genetic quality deteriorated. We concluded that distinct breeding selection goals should be directed at improving the genetic quality based on each breeds’ genetic characteristics and breed requirements and we provide the first national hip and elbow EBVs by which to do so. The genetic and residual correlations between hip and elbow scores were 0.12 and 0.08, respectively. The weak positive genetic correlation suggested that selection based on hip scores would also slightly improve elbow scores but it is necessary to allocate effort toward improvement of elbow scores alone (Hou et al., 2013 PLoS One in press).

These estimated hip and elbow breeding values and inbreeding coefficients will be accessible in this Cornell hip dysplasia web site. The dogs with low breeding value (low OFA score means a better hip) and with higher accuracy (more related dogs measured, the higher the accuracy) are the most desirable for breeding purposes. Low accuracy means that not many dogs were available to estimate the breeding value.

Inbreeding

Inbreeding occurs when a mating is made with a relative or the parents shared common ancestors. The closer an individual dog is to its ancestors with other dogs and the more common ancestors, the stronger the inbreeding. The most severe inbreeding occurs in a sibling to sibling mating or offspring to their parents. These matings commonly occur in an effort to preserve features of a breed or line within a breed and is referred to as “line breeding”. The degree of inbreeding can be mathematically expressed as an inbreeding coefficient. The inbreeding coefficient of an individual is defined as the probability that any two homologous alleles (same forms of the genetic locus) are identical by descent. That is, they were transferred from an ancestor to the current generation. Inbreeding often occurs the deeper you trace a pedigree. It is almost impossible to avoid inbreeding in a limited population, especially when the population has experienced a bottle neck. Severe inbreeding could result in shorter lives and problems of fitness including hip dysplasia. The level of inbreeding has continuously accumulated in US pure breed dogs over the past 40 years with higher inbreeding occurring generally in the breeds with total populations and therefore smaller breeding populations.

Questions & Answers about the Application of Hip and Elbow Estimated Breeding Values and Inbreeding Coefficients to the Breeding and Selection of a Pup (taken from the Cornell Hip EBV web site for the Labrador Retriever)

Once the new EBVs for other breeds are uploaded, then similar strategies for purchase and breeding decisions will apply to other breeds.

Why is this search function to find Labrador Retrievers with better hip breeding values useful? The breeding values and inbreeding coefficients recorded in this web site enable me to find dogs with low hip score breeding values that belong to the current and recent generations. The use of the dogs in the lower part of the breeding value range for breeding will likely improve the hip quality of my breeding stock and puppies they produce. Purchase of puppies produced by the sires and dams with the lower breeding values will likely produce puppies with better hips

77 than if I based breeding decisions on hip radiographs alone. The reason is that the selection of dogs based on breeding values means that consideration has been given to both the dog's genetic (pedigree) information and hip radiographic information combined. Selection of dogs based on radiographs alone is very useful but faster genetic gain toward better hip conformation accrues when breeding decisions are made based on genetic information as well.

Why does negative breeding values means a better hip? The question arises due to the ambiguity of word “value”, which usually suggests the higher value the better. The breeding value is an indicator for the genetic basis of the hip score variation. Consequently, breeding values take the same unit and direction as the original phenotype – the OFA score. An OFA score of 1 is for an excellent hip and an OFA score of 7 is for the most severe hip dysplasia.

What is the difference between expected progeny difference (EPD) and breeding value? The breeding value is the prediction of the genetic basis of an individual OFA score. Half of the genetic basis is contributed from one parent and half from the other. If an individual is mated randomly, the expected difference of the progeny from the average (base) will be half of the breeding value. Therefore, half of the breeding value is called the EPD. For example, sire A and B have breeding values of -0.1 and 0.20, their EPDs will be -0.05 and 0.1. The progeny of sire A is expected to be 0.15 lower than the progeny of sire B.

Will an inbred dog definitely have progeny with high inbreeding? Not really. The progeny may not be inbred if the mate you select is not its relative. The inbreeding of an individual depends if the parents are relatives or not.

Why can a breeding value be negative? The current reported breeding values were the direct output of the solutions for each dog in the mixed linear model. The base of the breeding value is the average breeding value among all the dogs evaluated. The base is a “floating” base which can vary by adding new dogs which have better hips.

I wish to choose a pup from a litter and I know the parents who produced this litter? How should I use the information in the hip EBV data base? Once you decide the qualities of the parents you prefer, then gather the information about any inherited traits and diseases segregating in the pedigree that you can. For a pup’s genetic potential to grow up with good hip quality, go into the data base and look at the hip breeding values for the dogs you like. Then you can rank those dogs based on their potential to produce pups with good hip conformation (the lowest hip breeding value indicates the dog with the genetic potential to produce the best hip conformation based on the OFA score). If only one parent is found in the data base, then that’s the best you can do. Secondly, you can rank the parents according to their inbreeding coefficients. You should try to choose pups produced from litters whose parents have the lowest inbreeding coefficients.

I wish to choose a pup from a litter but I don’t have information about the hip scores of either parent? You can ask the breeder for any pertinent radiographic information they have about their dog. They may have PennHIP information. They may not use the OFA method. They may do no orthopedic screening at all. We also know that elbow dysplasia is a problem in the Labrador Retriever breed. If you can obtain no information about orthopedic disease in a dog’s pedigree, then I suggest you try another breeder.

78 I wish to choose or buy a male dog as a potential breeder? How should I use the information in the data base? Once you have selected the potential male dogs based on all the breed qualities you prefer, then rank the dogs based on their genetic potential to produce offspring with good hip conformation and on their inbreeding coefficients. Always breed to a female dog with the best hip conformation and lowest inbreeding coefficient you can find along with all the best qualities you can ascertain, orthopedic or otherwise.

I wish to choose or buy a female dog as a potential breeder? How should I use the information in the data base? Once you have selected the potential female dogs based on all the breed qualities you prefer, then rank them based on their genetic potential to produce offspring with good hip conformation and on their inbreeding coefficients. Always breed to a male dog with the best hip conformation and lowest inbreeding coefficient you can find along with all the best qualities you can ascertain, orthopedic or otherwise.

I bought a pup already but just found this web site. How should I use the information in the data base to decide if this puppy is at risk of hip dysplasia? If you can identify the parents in the data base, look at the OFA breeding values of the parents. If they are above 0, then the pup has a higher chance of developing hip dysplasia than if the breeding values are below 0. The closer the breeding value is to 1, the greater the susceptibility to develop hip dysplasia. If you decide the pup is susceptible, it should be examined regularly for hip instability by your veterinarian. Depending on the dog’s age, medical or surgical intervention may be an option. This is especially important if your dogs has clinical signs of hip dysplasia like reluctance to jump, bunny hopping gait behind at speed (both hind legs moving forward together), soreness or stiffness after exercise, a “wobbly” hind limb gait, poor muscle mass development behind compared to its forequarter, difficulty getting up, placing extra body weight on its fore limbs with a hunched back, a clicking sound when it walks, or reluctance to allow you to pet near its hips. Any pup susceptible to hip dysplasia or any developmental orthopedic disease should be watched for rapid body weight gain and if it is too fat, its food intake should be restricted under advice of your veterinarian.

If a puppy is at risk for hip dysplasia based on the breeding value of its parents, what should I do about it? Ask your veterinarian to examine your puppy’s hips regularly. This is especially important if your dog has clinical signs of hip dysplasia like reluctance to jump, bunny hopping gait behind at speed (both hind legs moving forward together), soreness or stiffness after exercise, a “wobbly” hind limb gait, poor muscle mass development behind compared to its forequarter, difficulty getting up, placing extra body weight on its fore limbs with a hunched back, a clicking sound when it walks, or reluctance to allow you to pet near its hips. Any pup susceptible to hip dysplasia or any developmental orthopedic disease should be watched for rapid body weight gain and if it is too fat, its food intake should be restricted under advice of your veterinarian.

I wish to choose a male dog for my female dog to produce a litter of pups with the best hips I can. How do I select a dog from this data base? Rank the male dogs based on their OFA hip breeding values scores and their inbreeding coefficients. Choose the dog with the qualities you like as well as the best genetic potential to produce offspring with good hip conformation and lower inbreeding co-efficient.

79 Once I have identified a litter for puppy selection or a dog to which I'd like to breed, how do I locate the owner or breeder? We can suggest trying Google, other Labrador Retriever owners, breed/trade magazines like "Just Labs", contacting the Labrador Retriever breed clubs, or the AKC, etc. You can also purchase a pedigree from the AKC and this will have an owner's name on it. Eventually estimated breeding values and inbreeding coefficients for OFA hip scores will be available for many breeds.

Genomic Reference Panel and Genomic Prediction

State-of-the-art for predicting the dogs that carry the best combination of alleles at the genes that contribute to hip dysplasia is called genomic prediction. By genomic, I mean a method that interrogates the whole genome of the individual dog. No gene has yet been identified that contributes substantially, say 20%, to the overall genetic variation of the full range of hip dysplasia. However, if the density of genetic markers or variants for which a dog is genotyped, is sufficient to capture every gene that “lives” near a marker, then we can use the marker genotypes as a surrogate for the genes. The marker(s) is so close to the gene that the form of its alleles is always inherited with the gene i.e. recombination does not interfere with this relationship. There are a couple of strategies that can be used to undertake the genomic prediction. A subset of genetic markers called single nucleotide polymorphisms (SNPs) that span the genome are jointly selected for their contribution to CHD (or any other complex trait). Or a set of SNPs that are each significantly associated with the trait are used to build a multivariate linear model in a forward or backward method keeping the markers in the model that accounts for the most variation but eliminating redundant markers.

We employed the joint marker or Bayesian approach for our first effort. We used the Norberg angle which is highly phenotypically and genetically correlated with the OFA hip score. A reference population was established of dogs belonging to breeds susceptible or resistant to hip dysplasia that have undergone genome wide SNP genotyping and that have accompanying estimated hip breeding values calculated. A new dog of a breed that is in the reference population is genotyped either across the genome or at the best subset of SNPs and its estimated breeding value for optimal hip quality is estimated from the dogs in the reference panel based on its own SNP genotypes. Modest correlations can also be made with the raw Norberg angle. The best estimates of the genetic potential of two dogs to produce offspring with optimal hip quality will be based on gene mutation tests but it will take resources and time to discover the genes that contribute to CHD. In the mean time, SNP based selection will have to suffice to which we will later add the mutations to improve the prediction model.

Currently, the largest reference population for genomic prediction we have available is for the Labrador Retriever (Guo et al., 2011). For 180 Labrador Retrievers genotyped on the Illumina version 1, 22K mapping array, genomic hip breeding values for the Norberg angle were calculated in a Bayesian framework (Guo et al., 2011). This statistical method uses all the available genotypes to explain the variability in the Norberg angle. The estimated hip breeding values of these Labrador Retrievers were correlated with their genomic breeding values using the most predictive (effective) 280 SNPs of the 22,000 markers in the version 1 array. 30% of the variation in the Norberg angle of 108 Labrador Retrievers not used to develop the reference genomic panel was explained by the genomic prediction. The accuracy for a true phenotype is

80 about as expected because the heritability of HD as measured by the Norberg angle is only about 0.2-0.3. The accuracy of the genomic prediction for estimated hip breeding values on a subset of the 108 naïve dogs was moderate at 57% of the variation. Ongoing research would combine genomic prediction with the true hip radiographs of a genotyped dog to improve the accuracy of the prediction by including newly genotyped and phenotyped dogs into the reference panel. Other breeds might be added on which to predict genetic quality of hips. Recalculation of the genomic prediction algorithm based on more individuals and denser genotyping using the Illumina HD array should improve accuracy of the prediction. This iteration would be repeated over and over.

81 Holistic Management of Genetic Traits Anita Oberbauer, Ph.D., Department of Animal Science, University of California, Davis, CA [email protected]

In breeding any species, first and foremost there should be goals and objectives. Breeders have different goals (improve the breed, optimize performance characteristics, win, financial remuneration) and some may view their goals as being more noble than the goals of others. Regardless, in any breeding endeavor one must strive for that specific goal(s) and in doing so make concessions. The hallmarks of a successful breeder include making progress toward the overall objective while minimizing the negative impact of tradeoffs.

Above, several possible breeding objectives were listed. An animal in a breeding pool is a composite of numerous elements that include desired type, health, performance, reproductive efficiency, structure, and temperament. When selecting breeding animals, these elements must be prioritized and their relative importance to one another weighed. For example, a dog or cat that is ideal in every category but lacks fertility fails to reach a breeding objective. A highly fertile dog or cat that lacks desired type likewise fails to meet a breeding goal. Thus, in any breeding program, one must achieve a balance blending often conflicting aspects.

Unfortunately, rather than looking at the long view and complexities of achieving breeding goals, the majority of claims against concerted breeding programs (purebred dog or cat for example) center on a perceived lack of concern by breeders to reduce harmful genetic conditions in order to win, make money, satisfy ego, or (fill in the blank). Yet when breeders and owners are asked to define “health” in relationship to dogs (or cats), definitions are many and varied. Definitions can be pragmatic (not needing visits to the veterinarian), focused solely on physical health, or focused solely on mental health; most commonly cited attributes of “health” are absence of disease or injury concomitant with the ability to perform normal/expected body functions and abilities. Most breeders or owners focus on the individual when considering health. In contrast, livestock producers also include population health (so called “herd health”) as a significant component of their concept of “health”. Herd health is especially critical for large numbers of animals and/or densely populated animal groupings. Stepping back and considering health in a broader perspective, herd health is definitely applicable to the dog population as a whole or to a particular breed. One can consider genetic health of the population as underpinning all the elements a breeder needs to attain a breeding goal.

An individual’s qualities (health, temperament, type, etc.) are a reflection of the population’s genetic potential. When selecting an individual for breeding, the breeder should balance the individual’s needs (a certain dog may need a mate who has a better shoulder assembly) with that of the population as a whole (excessive use of a popular sire can reduce the genetic diversity for future generations). Further, the breeder must make compromises. Even if the absolute perfect breed specimen is produced, to perpetuate that individual one must breed to a mate that has faults. What qualities should be emphasized in the less than perfect mate? One breeder will say type (and that includes every attribute ranging from , muzzle shape, ear placement, length of back, to bend of stifle and beyond!) whereas a second breeder will insist that temperament is most critical (and temperament also has a spectrum of qualifiers). Yet a third breeder will insist upon health (as discussed above, health means different things to different

82 people). Despite the varied opinions each breeder should have a prioritized and weighted view to a breeding program. Even then, the suite of traits that comprises the general element (type, performance, etc.) each needs to be prioritized and weighted. While no breeder would knowingly breed genetic defects, should one trade less than ideal eye color for better eye shape?

The domestication of the dog and cat reflected selection on traits that favored successful co- habitation with the human population. The inherent genetic diversity of the ancestral wolf permitted the expression of many traits that favored domestication. Yet the domestication process reduced some genetic diversity that was present in its ancestor; that is, bottlenecks in which limited numbers of individuals established a relationship with humans created subpopulations. Genetic diversity is critical to compensate for current and future challenges. For example, the restricted genetic diversity in the endangered black-tailed prairie dog has resulted in their susceptibility to an exotic, introduced pathogen that causes plague. Maintaining genetic diversity maintains the health of the population. Thus, the founding dog population represented a subset of the ancestral wolf and therefore dogs began with a smaller gene pool. The establishment of breeds within the dog population as a whole further reduced the gene pools for each breed.

The challenge in breeding is to fix the desirable traits while maintaining genetic diversity. Loss of genetic variety within a unique population (read “breed”) is considered highly detrimental to the overall genetic health of a breed. A population may begin with a limited gene pool. Developing a new breed and then closing the registry for that breed equates to a small gene pool. Using inbreeding schemes to fix desirable traits reduces genetic diversity by increasing the genetic homozygosity, that is making both copies of a gene identical. Increased homozygosity ensures that a particular desirable trait will be expressed it also potentiates the expression of genetic disorders that are recessively inherited. Furthermore, loss of heterozygosity is statistically correlated with greater autoimmune concerns. Taken together, although inbreeding, enhances uniformity within litters and fixes characteristic, breed-defining traits, it also has unintended consequences such as loss of rare alleles, increased homozygosity enabling expression of recessive disorders, and reducing effective population size. Thus, inbreeding has been the subject of much debate concerning the welfare and health of purebred dogs.

Similarly, extensive use of a popular sire also reduces heterozygosity effectively reducing the population size. The use of a popular sire also proves to be more effective at dispersing deleterious alleles within a breed than inbreeding (Leroy & Baumung, 2010) making disorders that occur in a popular sire (or one for which he carries the mutant alleles) more difficult to manage in the future. In humans the mutation rate resulting in random errors in DNA is one mutation in every 100 million base pairs equaling ~ 60 new mutations per generation and more mutations arise from the male (Conrad et al., 2011). Each human is estimated to carry approximately 1,000 deleterious mutations (Sunyaev et al., 2001). Also in humans, it has been demonstrated (Chun & Fay, 2011) that to eliminate some deleterious alleles may increase the frequency of others; a deleterious allele may hitchhike along with a desirable allele due to genetic linkage. In one review, all top 50 breeds the study evaluated had at least one associated with the conformation demanded by the standard (Asher et al., 2009). Deleterious mutations are difficult to eliminate from small populations and are likely to accumulate.

83 The association of deleterious with desirable traits has implications for proponents mandating only individuals clear of deleterious alleles are permitted to breed. When considering genetic health of an individual in relation to the population health, no single individual is free from all genetic mutation. A dog, any dog, when all genetic diseases have been characterized will fail at least one genetic test. Limiting breeding to those clear will further restrict the gene pool and introduce unintended health consequences. That does not mean that genetic testing is unwarranted. As Dr. Jerry Bell states, “breeding without genetic testing is irresponsible, and unethical.” Using available test results in a holistic approach is key to maintaining the overall genetic health of a breed.

In some cases the genetic test may indicate a risk, but not guarantee, of expression of a disease (for example, degenerative myelopathy, Chang et al., 2013). Utilizing that information to inform breeding decisions is critical but eliminating all dogs having a risk from the breeding population is unwise. In other cases, the presence of an allele may be viewed as deleterious or an asset. A particular allele for a behavioral trait is associated with highly productive working dogs although owners should emphasize non-confrontational training methods to achieve optimal performance; yet there is significant association between spontaneous episodic aggressive behaviors in dogs with that allele (Lit et al., 2013). Maintaining that diversity within the gene pool permits breeders to attain their individual goals.

A comment on crowd sourcing of health information: popular beliefs can be very wrong even if commonly held. An example from history, it was universally believed that the world was flat— even though there was consensus did not make that view factual. Just because “everyone” says it’s true does not make it so and sensible caution should be applied to health statements. Much is made of “healthy” mixed breeds; domesticated dogs carry deleterious mutations dating back to the original domestication step. Thus, there are health conditions that will be present in a dog, any dog, be it a purebred or mixed breed dog.

Concerted breeding to reduce unwanted traits is the only means to eliminate particular conditions. Wisdom and stewardship of a breed is essential. The genetic health of a breed depends upon wise sire and dam selection.

References Asher, L. Diesel, G., Summers, J.F., McGreevy, P.D., and Collins, L.M. (2009) Inherited defects in pedigree dogs. Part 1: Disorders related to breed standards. Veterinary Journal 182, 402-411.

Calboli, F.C.F., Sampson, J., Fretwell, N., and Balding D.J. (2008) Population structure and inbreeding from pedigree analysis of purebred dogs. Genetics 179, 593-601.

Chang HS, Kamishina H, Mizukami K, Momoi Y, Katayama M, Rahman MM, Uddin MM, Yabuki A, Kohyama M, Yamato O. (2013) Genotyping Assays for the Canine Degenerative Myelopathy-Associated c.118G>A (p.E40K) Mutation of the SOD1 Gene Using Conventional and Real-Time PCR Methods: A High Prevalence in the Pembroke Welsh Corgi Breed in Japan. Journal of Veterinary Medical Science. 75, 795-798

Chun S, Fay JC (2011) Evidence for Hitchhiking of Deleterious Mutations within the Human Genome. PLoS Genet 7(8): e1002240. http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1002240

84

Conrad et al. For the 1000 Genomes Project. (2011) Variation in genome-wide mutation rates within and between human families. Nature Genetics 43, 712-714.

Leroy, G. and Baumung, R. (2010) Mating practices and the dissemination of genetic disorders in domestic animals, based on the example of dog breeding. Animal Genetics doi 10.1111/j.1365- 2052.2010.02079.x.

Lit L, Belanger JM, Boehm D, Lybarger N, Haverbeke A, Diederich C, Oberbauer AM. (2013) Characterization of a dopamine transporter polymorphism and behavior in Belgian Malinois. BMC Genet. 2013 May 30;14:45. http://www.biomedcentral.com/1471-2156/14/45.

Sunyaev S, Ramensky V, Koch I, Lathe W, Kondrashov A, et al. (2001) Prediction of deleterious human alleles. Hum Mol Genet 10: 591–597.

OMIA. Online Medelian Inheritance in Animals. Reprogen, Faculty of Veterinary Science, University of Sydney, {December 2010}. World Wide Web URL: http://omia.angis.org.au/

85 From FUS to Pandora Syndrome - The Role of Epigenetics and Environment in Pathophysiology, Treatment, and Prevention C. A. Tony Buffington, DVM, PhD, DACVN, The Ohio State University Veterinary Hospital, Columbus, OH [email protected]

INTRODUCTION In an accurate clinical description of cats with lower urinary tract (LUT) disease published in 1925,1 the disorder was reported to be very common, the roles of confinement and highly nutritious food were discussed, and the common occurrence of the problem in Persian cats was identified. In 1970, the term feline urologic syndrome (FUS) was coined by Osbaldiston and Taussig to describe a problem, “characterized by dysuria, urethral obstruction, urolithiasis (although no stones were reported) and hematuria”.2 They concluded from a review of 46 cases, “the condition may not be a single disease entity, but rather a group of separate urologic problems.”

During the 1980s, Osborne, et al., suggested that FUS should be considered synonymous with feline lower urinary tract disease (FLUTD).3 Then, in 1995, the group4 suggested that the acronym FUS be redefined as feline urologic signs to emphasize that FUS is not an etiologic diagnosis of any particular LUT disease. They proposed that “when possible, refined diagnoses of lower urinary tract disease should encompass descriptive terms pertaining to the site (e.g., urethra, bladder), pathophysiologic mechanisms (e.g., obstructive uropathy, reflex dyssynergia), morphologic features (e.g., inflammation, neoplasia), and causes (e.g., anomalies, urolithiasis, bacteria, fungi),” and suggested that confirmed and suspected causes of LUT diseases in domestic cats be categorized as anatomic, iatrogenic, idiopathic, inflammatory (infectious and noninfectious), metabolic, neoplastic, neurogenic, or traumatic. The terms FUS and FLUTD have since been superseded by the ability of veterinarians to diagnose many distinct causes of the well-known clinical signs of dysuria, stranguria, pollakiuria, hematuria, and inappropriate urination (periuria) that, either individually or in some combination, cause clients to seek further evaluation of their cats.5

Retrospective studies suggest that the majority of non-obstructed cats with LUT signs have an idiopathic disorder, and that this percentage has not changed appreciably during the past 4 decades.2,6-9 The importance of LUT disorders to feline health is emphasized by the finding that elimination disorders (the vast majority of which are urinary) result in destruction of millions of cats in animal shelters in the United States every year.10 We defined idiopathic cystitis as an acute or chronic disease of waxing and waning signs of irritative voiding (dysuria, pollakiuria, hematuria, periuria), sterile urine, absence of cellular abnormalities suggesting neoplasia, and failure to identify an alternative cause for these signs after appropriate lower urinary tract (LUT) imaging procedures (combination of plain radiography, contrast radiography, contrast urethrography, ultrasonography) in the absence of cystoscopic evaluation.8 Feline interstitial cystitis (FIC), a subcategory of idiopathic cystitis, was defined as a chronic condition describing cats that have frequent recurrences or persistence of clinical signs and cystoscopic documentation of submucosal petechial hemorrhages (glomerulations) after bladder distension to 80 cm water pressure in the absence of an alternative explanation for these findings.9

1

86 Based on a series of studies conducted during the past two decades, a variety of problems beyond the urinary tract have been identified in cats with chronic, severe, recurrent LUT signs.11 These include epithelial, neurological, endocrine, immune and behavior abnormalities, as well as a variety of comorbid disorders (which often precede development of LUT signs) affecting many body systems. Enhanced central sympathetic drive in the face of inadequate adrenocortical restraint, which seems to be related to maintaining the chronic disease process, also has been identified. These systems appear to be driven by tonically increased activity of the central stress response system, which may represent the outcome of a developmental accident that durably sensitizes this system to the environment, possibly through epigenetic modulation of gene expression.12 The repeated observation that most of these problems resolve after exposure to an enriched environment provided additional evidence for a disorder of the central nervous system resulting in a chronic multi-system illness variably affecting the bladder and other organs, as opposed to a peripheral, organ-based problem.13-15

Diagnosis Based on the evidence outlined above, I believe that some cats with chronic LUT signs may have a “Pandora syndrome” (named for the Pandora myth, which reflects my experience in studying this problem, and my optimism that hope for effective treatment remains).16 Based on the currently available evidence, provisional criteria for diagnosis of a “Pandora syndrome” might include:

1. Chronicity – persistence or recurrence of the condition(s) over months to years. 2. Comorbidity - evidence of problems in other body systems (particularly preceding the presenting LUTS in the case of idiopathic cystitis. These may include behavioral, endocrine gastrointestinal, respiratory, dermatological, etc. 3. A history of early adverse experience (orphaned, bottle fed, rescued). 4. Evidence of familial involvement. That is, parents and or littermates have a similar illness profile.

Information about early experience and family members often cannot be obtained from owners, and none of these criteria can be considered pathognomonic for anything. They may serve only to raise one’s “index of suspicion” that a more systemic problem may be present. By taking the time to obtain a comprehensive review of the cat’s history and conduct a thorough physical examination before assuming that the cat has an isolated bladder (or other) disease, one may find that some cats appear to have a disease affecting more than the organ attributed to the presenting signs, which can helpfully inform one’s therapeutic recommendations. I urge others to test this hypothesis for themselves.

Treatment Based on current understanding of the role of the environment in chronic illness in cats, environmental enrichment is the first line of therapy to reduce the risk of recurrence of whatever clinical signs are present.13-15 Environmental enrichment for indoor-housed cats means provision of all “necessary” resources listed below, refinement of interactions with owners, a tolerable intensity of conflict, and thoughtful institution of change(s).5,17,18 The following areas all are considered based on their influence on the health and welfare of indoor-housed cats.

2

87 1. Food - Cats prefer to eat individually in quiet locations where they will not be startled by other animals, sudden movement, or activity of an air duct or appliance that may begin operation unexpectedly. Although canned food may be preferable for some cats due to the increased water content or a more natural “mouth feel”, some cats may prefer dry foods. If a diet change is appropriate, offering the new diet in a separate, adjacent container rather than removing the usual food and replacing it with the new food permits the cat to express its preferences. Natural cat feeding behavior also includes predatory activities such as stalking and pouncing. These may be simulated by hiding small amounts of food around the house, or by putting dry food in a container from which the cat has to extract individual pieces or move to release the food pieces, if such interventions appeal to the cat. Also, some cats seem to have specific prey preferences. For example, some cats prefer to catch , while others may prefer to chase mice or bugs. Identifying a cat’s “prey preference” allows one to buy or make toys that the cat will be more likely to play with. Specific ingredients or nutrients has been found to be of minor significance to patient outcome when an enriched environment is provided.13-15

2. Water - Cats also seem to have preferences for water that can be investigated. Water-related factors to consider include freshness, taste, movement (water fountains, dripping faucets or aquarium pump-bubbled air into a bowl), and shape of container (some cats seem to resent having their vibrissae touch the sides of the container when drinking). As with foods, changes in water-related factors should be offered in such a way that permits the cat to express its preferences. Additionally, food and water bowls should be cleaned regularly unless individual preference suggests otherwise.

3. Litter boxes - Litter boxes should be provided in different locations throughout the house to the extent possible, particularly in multiple cat households. Placing litter boxes in quiet, convenient locations that provide an escape route if necessary for the cat could help improve conditions for normal elimination behaviors. If different litters are offered, it may be preferable to test the cat’s preferences by providing them in separate boxes, since individual preferences for litter type have been documented. For cats with a history of urinary problems, unscented clumping litter should be considered. Litter boxes should be cleaned regularly and replaced; some cats seem quite sensitive to dirty litter boxes. Litter box size and whether or not it is open or covered also may be important to some cats.19

4. Space - Cats interact with both the physical structures and other animals, including humans, in their environment. The physical environment should include opportunities for scratching (both horizontal and vertical may be necessary), climbing, hiding and resting. Cats seem to prefer to monitor their surroundings from elevated vantage points, so climbing frames, hammocks, platforms, raised walkways, shelves or window seats may appeal to them. Playing a radio to habituate cats to sudden changes in sound and human voices also may be useful, and videotapes to provide visual stimulation are available.

5. Play - Some cats seem to prefer to be petted and groomed, whereas others may prefer play interactions with owners. Cats also can be easily trained to perform behaviors (“tricks”); owners just need to understand that cats respond much better to praise than to force, and seem to be more amenable to learning when the behavior is shaped before feeding. Cats also

3

88 may enjoy playing with toys, particularly those that are small, move, and that mimic prey characteristics. Many cats also prefer novelty, so a variety of toys should be provided, and rotated or replaced regularly to sustain their interest.

6. Conflict management - When cats’ perception of safety becomes threatened, they appear to respond by attempting to restore their “perception of control”.20 During such responses, some cats become aggressive, some become withdrawn, and some become ill.13 In our experience, intercat conflict commonly is present when multiple cats are housed indoors together and sickness behaviors are present in some of them.13 Signs of conflict between cats can be open or silent. Cats in open conflict may stalk each other, hiss, and turn sideways with legs straight and hair standing on end up to make themselves look larger. In contrast, signs of silent conflict can be easily missed; threatened cats may avoid other cats, decrease their activity, or both. They often spend increasingly large amounts of time away from the family, stay in areas other cats do not use, or attempt to interact with family members only when the assertive cat is elsewhere. Signs can result from two types of conflict; offensive and defensive. In offensive conflict, the assertive cat moves closer to the other cats to control the interaction. In defensive conflict situations, the threatened cat attempts to increase the distance between itself and the perceived threat. Although cats engaged in either type of conflict may spray or eliminate outside the litter box, we find that threatened cats are more likely to develop elimination problems.

A common cause of conflict between indoor-housed cats is competition for resources; space, food, water, litter boxes, perches, sunny areas, safe places where the cat can watch its environment, or attention from people. There may be no limitation to access to these resources apparent to the owner for conflict to develop; the cat's perceptions of how much control it wants over the environment or its housemates' behaviors determines the outcome of the situation.

Open conflict is most likely to occur when a new cat is introduced into the house, and when cats that have known each other since kittenhood reach social maturity. Conflict occurring when a new cat is introduced is easy to understand, and good directions are available from many sources for introducing the new cat to the current residents.21 Clients may be puzzled by conflict that starts when one of their cats becomes socially mature, or when a socially mature cat perceives that one of its housemates is becoming socially mature. When cats become socially mature, they may start to exert some control of the social groups and their activities. This may lead to open conflict between males, between females, or between males and females. And although the cats involved in the conflict may never be “best friends”, they usually can live together without showing signs of conflict or conflict- related illness. In severe cases, a behaviorist can be consulted for assistance in desensitizing and counter conditioning of cats in conflict so they can share the same spaces more comfortably if this is desired.

Treatment for conflict between cats involves providing a separate set of the listed resources for each cat; in locations where cats can use them without being seen by other cats if possible. This lets the cats avoid each other if they choose to without being deprived of any essential resource. Cats may require and use more space than the average house or

4

89 apartment affords them. The addition of elevated spaces such as shelves, “kitty condos”, cardboard boxes, beds, or crates may provide enough three-dimensional space to reduce conflict to a tolerable level. In severe situations, some cats may benefit from behavior- modifying medications. In our experience, however, medication can help when combined with environmental enrichment has occurred, but cannot replace it. Conflict also can be reduced by neutering all of the cats, and by keeping all nails trimmed as short as practicable. Whenever the cats involved in the conflict can not be directly supervised, they may need to be separated. This may mean that some of the cats in the household can stay together, but that the threatened cat is provided a refuge from the other cats. This space should contain all necessary resources for the cat staying in it.

Conflict with other animals, dogs, children, or adults is relatively straightforward. In addition to being solitary hunters of small prey, cats are small prey themselves for other carnivores, including dogs. Regardless of how sure the client is that their dog will not hurt the cat, to the cat the dog may represent a predator. To ensure the cat’s safety, it must be provided avenues of escape that can be used use at any time. For humans, it usually suffices to explain that cats may not understand rough treatment as play, but as a predatory threat. Most cats in urban areas in the United States are housed indoors and neutered, so conflict with outside cats can occur when a new cat enters the area around the house the affected cat lives in. To cats, windows offer no protection from a threatening cat outside. If outside cats are the source of the problem, a variety of strategies to make ones garden less desirable to them are available.

7. Pheromones - Pheromones are chemical substances that seem to transmit highly specific information between animals of the same species. Although the exact mechanism of action is unknown at this time and their effectiveness is not universally demonstrated,22 pheromones appear to effect changes in the function of both the limbic system and the hypothalamus to alter the animal’s emotional state. Feliway®, which contains a synthetic analogue of naturally occurring feline facial pheromone and , was developed to decrease anxiety- related behaviors of cats. Use of this product has been reported to reduce the amount of anxiety experienced by cats in unfamiliar circumstances, a response that may be helpful to these patients and their owners. Decreased spraying in multi-cat households, decreased marking, and a significant decrease in scratching behavior also has been reported subsequent to its use. Feliway is not a panacea for unwanted cat behaviors, its effectiveness may be improved by using it in combination with environmental enrichment, and/or drug therapies.

Because of the dearth of controlled trials, it currently is not possible to prioritize the importance of any of these suggestions, or to predict which would be most appropriate in any particular situation. Appropriately designed epidemiological studies might be able to identify particularly important factors, after which intervention trials could be conducted to determine their efficacy in circumstances where owners successfully implemented the suggested changes.

Follow-Up One of the critical keys to any successful therapy program is to follow the progress of the patient, which we generally delegate to a trained technician introduced to the client during the clinic appointment. We tell clients what our follow-up schedule is, and ask them to agree to a

5

90 preferred method and time to be contacted. Our first contact with the client occurs within a week after initial recommendations are made usually over the telephone, followed by repeat in-house “check-ins” at 3-6 weeks, 3 months, 6 months, and 1 year in uncomplicated cases (which need less follow-up). This allows one to monitor the patient’s progress, to make adjustments as needed, and to continue to coach the client. It also helps to determine when the owner is becoming frustrated or is having problems with the plan so that encouragement or suggestions to help them can be offered.

Conclusions Many indoor housed cats appear to survive perfectly well by accommodating to less than perfect surroundings. The neuro-endocrine-immune systems of some cats, however, do not seem to permit the adaptive capacity that healthy cats enjoy, so these cats may be considered a separate population with greater needs. Moreover, veterinarians are concerned more with optimizing the environments of indoor cats than with identifying minimum requirements for indoor survival. My current approach is to let the client choose the most appropriate intervention for their particular situation, and to let trained technicians do the enrichment implementation and follow-up (under veterinary supervision as appropriate).

Finally, the question of the relative merits of indoor housing to promote the welfare of cats (and the different opinions on what constitutes animal welfare in general) is beyond the present scope, and is a subject of controversy among experts. I hope to encourage extension of the welfare efforts of individuals working in zoos, who have recognized the effects of the quality of housing on the health on animals in their care and worked to enrich the environments of these animals, to all “captive” animals in our care. I believe that chronic idiopathic cystitis and a variety of related chronic health problems in cats may be better prevented than treated, and that we have a great opportunity to encourage this husbandry approach in veterinary clinical practice. Further information about environmental enrichment for indoor housed cats is available at: http://indoorpet.osu.edu/

References 1. Kirk H. Retention of urine and urine deposits In: Kirk H, ed. The Diseases of the Cat and its General Management. London: Bailliere, Tindall and Cox, 1925;261-267. 2. Osbaldiston GW, Taussig RA. Clinical report on 46 cases of feline urological syndrome. Vet Med/Small Anim Clin 1970;65:461-468. 3. Osborne CA, Johnston GR, Polzin DJ, et al. Redefinition of the feline urologic syndrome: feline lower urinary tract disease with heterogeneous causes. Vet Clin North Am Small Anim Pract 1984;14:409-438. 4. Osborne CA, Kruger JM, Lulich JP, et al. Feline Lower Urinary Tract Diseases In: Ettinger SJ,Feldman EC, eds. Textbook of Veterinary Internal Medicine. 4 ed. Philadelphia: W.B. Saunders, 1995;1805-1832. 5. Westropp J, Buffington CAT. Lower Urinary Tract Disorders in Cats In: Ettinger SJ,Feldman EC, eds. Textbook of Veterinary Internal Medicine. 7 ed. St. Louis: Elsevier-Saunders, 2010;2069-2086. 6. Kruger JM, Osborne CA, Goyal SM, et al. Clinical evaluation of Cats with lower urinary tract disease. Journal of the American Veterinary Medical Association 1991;199:211-216. 7. Barsanti JA, Brown J, Marks A, et al. Relationship of lower urinary tract signs to seropositivity for feline immunodeficiency virus in cats. Journal of Veterinary Internal Medicine 1996;10:34-38.

6

91 8. Buffington CA, Chew DJ, Kendall MS, et al. Clinical evaluation of cats with nonobstructive urinary tract diseases. Journal of the American Veterinary Medical Association 1997;210:46-50. 9. Buffington CAT, Chew DJ, Woodworth BE. Feline Interstitial Cystitis. Journal of the American Veterinary Medical Association 1999;215:682-687. 10. Patronek GJ, Glickman LT, Beck AM, et al. Risk factors for relinquishment of cats to an animal shelter. Journal of the American Veterinary Medical Association 1996;209:582-588. 11. Buffington CA. Idiopathic cystitis in domestic cats-beyond the lower urinary tract. J Vet Intern Med 2011;25:784-796. 12. Buffington CAT. Developmental Influences on Medically Unexplained Symptoms. Psychotherapy and Psychosomatics 2009;78:139-144. 13. Stella JL, Lord LK, Buffington CAT. Sickness behaviors in response to unusual external events in healthy cats and cats with feline interstitial cystitis. Journal of the American Veterinary Medical Association 2011;238:67-73. 14. Westropp JL, Kass PH, Buffington CA. Evaluation of the effects of stress in cats with idiopathic cystitis. Am J Vet Res 2006;67:731-736. 15. Buffington CAT, Westropp JL, Chew DJ, et al. Clinical evaluation of multimodal environmental modification (MEMO) in the management of cats with idiopathic cystitis. Journal of Feline Medicine and Surgery 2006;8:261-268. 16. Buffington CAT. Idiopathic Cystitis in Domestic Cats – Beyond the Lower Urinary Tract. JVIM 2011;doi: 10.1111/j.1939-1676.2011.0732.x. [Epub ahead of print]. 17. Herron ME, Buffington CAT. Environmental enrichment for indoor cats. Compend Contin Educ Pract Vet 2010;32:E1-E5. 18. Herron ME, Buffington CA. Environmental enrichment for indoor cats: implementing enrichment. Compend Contin Educ Vet 2012;34:E1-5. 19. Herron ME. Advances in understanding and treatment of feline inappropriate elimination. Top Companion Anim Med 2010;25:195-202. 20. Moesta A, Crowell-Davis S. Intercat aggression - general considerations, prevention and treatment. Tierarztliche Praxis Kleintiere 2011;39:97-104. 21. Overall KL, Rodan I, Beaver BV, et al. Feline behavior guidelines from the American Association of Feline Practitioners. Journal of the American Veterinary Medical Association 2005;227:70-84. 22. Gunn-Moore DA, Cameron ME. A pilot study using synthetic feline facial pheromone for the management of feline idiopathic cystitis. Journal of Feline Medicine and Surgery 2004;6:133-138.

7

92 Breed Specific Breeding Straegies

Åke A Hedhammar, DVM, M Sc, Ph D, Dipl. Internal Medicine -Companion Animals Dept. of Clinical Sciences, Swedish University of Agricultural Sciences, Uppsala, Sweden

Definition of a breeding program

• planned breeding of a group of animals ( or ) , usually involving at least several individuals and extending over several generations . • organized structure that is set up in order to realize the desired genetic improvement of the population • agreed strategy to influence prevalence of a defined phenotype in a defined population

Breeding programs for specific traits have been established in many countries. They are most commonly restricted to inherited disorders. Disease specific breeding programs have been instituted for disorders as hip dysplasia, hereditary eye defects and a number of other conditions possible to reveal by phenotypic or genotypic screening methods. Their values are indisputable but do not very well account for breed variations in prevalence, population structure and other traits to take into account. The goal for planned, organized and agreed breeding plans is broader than just a few specified genetic disorders

This presentation will review Swedish experiences to establish breed specific breeding programs taking into account not only disease specific breeding programs but also how to handle other undesired as well as desired traits and to adapt them to population structure and other differences between various breed populations.

Since 2004 the Swedish Kennel Club (SKC) have demanded every breed club to prepare a breed specific breeding program for their strategy to handle future development regarding desired as well as non-desired traits. It calls for a thorough description of current situation and to prioritize actions that should be taken to reach common agreeable goals for their national breed population.

Sources of information

To describe the breed population and the results from applicable screening programs for inherited disorders as well as behavior test data SKC have extensive material available on the web. Like in many other countries including US several breed clubs also have performed breed surveys on various health issues that form a good basis for the situation regarding many health issues.

In Sweden, more than 75 % of all Dogs are of known ancestral background and registered by SKC. Moreover over 50 % have insurance for life and veterinary care and the majority in one company- Agria Insurance.

Their database has been made available for population based epidemiological studies of a number of diseases.

The breed specific disease pattern in German Shepherds has recently been published and the breed specific disease patterns of more than 100 breeds are available as Agria Breed profiles.

93 Future perspective

As dog breeding is truly international breed specific breeding programs ideally should not only be prepared for national breed population. International breed specific programs would enhance exchange of breeding stock and vital breed populations. Country of origin would be the nucleus in such efforts and the International Breeding societies should take the lead in their preparation.

At The 1st International Workshop on Enhancement of Genetic Health in Purebred Dogs that was arranged by the Swedish Kennel Club in Stockholm on June 2-3, 2012.one of the key issues dealt with was Development of breed-specific breeding programs on national and international levels.

References and suggested further readings

Agria Dog Breed Profiles (ADBP) (2011) http://www.agria.se/agria/artikel/agria-dog-breed- profiles-1

Special Breed Specific Instructions (BSI) regarding exaggerations in pedigree dogs (2011) http://www.skk.se/Global/Dokument/Utstallning/special-breed-specific-instructions-A8.pdf

SKC (Swedish Kennel Club) (2011) Dog Health Workshop http://www.skk.se/in-english/dog-health-workshop-2012/

Hedhammar ÅA, Malm S, Bonnett B (2011) International and collaborative strategies to enhance genetic health in purebred dogs. Vet J. 189(2):189-96

BREEDING dogs in Sweden (2012) http://www.skk.se/Global/Dokument/Om- SKK/Breeding-dogs-in-Sweden-2012_webb.pdf

Code of Ethics for the Swedish Kennel Club (2013) http://www.skk.se/Global/Dokument/Om-SKK/Code-of-ethics_breeding-policy_ethical- guidelines_webb.pdf

Vilson A., Bonnett B., Hamlin H., Hedhammar A. (2013) Disease patterns in 32,486 insured German Shepherd Dogs in Sweden: 1995-2006, Vet. Record 2013 Aug 3;173(5):116

94 UK Initiatives for breeding healthier pedigree dogs Tom Lewis PhD, Animal Health Trust, Lanwades Park, Kentford, Newmarket, Suffolk, UK [email protected]

Selection against inherited disease is necessary for lasting and widespread improvement in many aspects of canine welfare. Successful genetic selection requires 1) the motivation to change a trait in the population; 2) data or information to differentiate between animals with respect to that trait; and 3) sufficient control of breeding animals to direct specific matings. These factors are demonstrated clearly in livestock, where the traits in question are associated with food production (e.g. milk yield). For farmers the motivation in changing (increasing) the milk yield in their herd is profit, since higher yields generate higher returns. They are easily able to differentiate animal performance through assiduous recording of yields (which are pretty much universal since payment is linked to quantity). Finally they have control over the breeding of the entire herd (often hundreds of animals). These three factors have resulted in the widespread genetic improvement in the performance of livestock and contributed to dramatic improvements in yields over the last 60 years. Furthermore, as a result of the overarching financial motivation, the abundance of data and complete control of breeding, it has been possible to identify strategies that maximize genetic gain while minimizing the risk of future problems due to inbreeding.

When it comes to breeding pedigree dogs the situation is less favourable to elicit widespread genetic change. First, consider the motivation – what are most breeders’ principle objectives? They are manifold; some breed primarily for success in the show ring or at field trials, some for working ability (working gun dogs, herding dogs, guide dogs, sniffer dogs), but I suspect most are hobby breeders and intend the puppies to go to pet homes. Thus, although health is likely a universal consideration among dog breeders it is only one of a multitude of selection objectives. Differentiation between breeding animals on ‘merit’ by dog breeders is highly subjective and has often been achieved by eye, experience or anecdote (reflecting the principle motivations). Finally, individual dog breeders have only very limited control over the breeding population, usually one or very few animals. Given that dog breeders are quite individualistic or self-reliant in terms of judgment of merit and scale of operation, the use of health information to elicit widespread improvement in health is often sub-optimal. DNA tests are an example of information that has been enthusiastically employed by dog breeders, possibly because they offer a simple and definitive result (and one impossible to evaluate visually), and they are consistent with individualistic operation.

The landscape of dog breeding means that some of the more sophisticated tools available to livestock breeders to maximise genetic gain will not be directly transferable to dogs. Nevertheless, there are measures that can be taken in several areas that will assist in improving the efficacy of selection for health, by focusing on the motivation, the information and the control. In this short talk I will highlight a few being undertaken in the UK.

Motivation It is important to stress that in the majority of cases, health already is one of the primary objectives of breeders. No one I’ve met explicitly intends to breed a dog with disease. However, in some cases primary motivations may supersede the motivation to breed for health; for example the trend for greater exaggeration of breed defining characteristics may have [inadvertently] led to compromising the health of some breeds (e.g. Brachycephalic airway disease in Bulldogs or Pugs, and skin conditions in Bassett Hounds).

95 If health can be linked to the primary motivations of breeders, then it will become a de facto selection objective. The introduction of vet checks at Crufts, barring progression of ‘Best of Breed’ winners failing the checks to group finals, is a way of linking health to success in the show ring. Health does appear to be a concern of puppy buyers. Raising awareness of and providing information on health to the general public could help to elicit changes in demand.

Information I covered a bit about the more effective use of health information in my earlier lecture (using EBVs for hip score to elicit more accurate selection). We have also heard about the importance of monitoring inbreeding in populations, and must consider appropriate breeding strategies when there are DNA tests for simple Mendelian recessive diseases, i.e. multi- objective selection often within limited genepools. Mate Select currently provides information on inbreeding coefficients of litters from potential matings, and shortly will include EBVs for hip and elbow score, and a simple population analysis for most of the breeds registered by the Kennel Club. The Kennel Club has a role to play, as the repository of health data and pedigree in the UK, in providing more accurate information regarding health and risks for both individual dogs and entire breeds.

Control Compared to livestock breeders, dog breeders have control over the breeding of far fewer animals. Coupled with a more individualistic or self-reliant ethos to dog breeding, possibly due to differing objectives and maybe even competition, the sharing and use of data to direct matings to meet common objectives is less widespread than in livestock sectors. However, health is a common objective (or should be, and is a universal if not the principle objective), and health information is increasingly available allowing breeders to be more discriminating in mate selection. Therefore, breeders will continue to benefit from a range of tools designed to allow them access to the most accurate information relating to health, and that will allow them to use it in their own way since ‘herd-wise’ solution are not realistic. The Kennel Club’s role in collating and presenting as much health information as possible is critical in coordinating the efforts of a multitude of breeders to meet a universal selection objective.

96 Genetic Tests: Understanding Their Power, and Using Their Force for Good Jerold S Bell DVM, Tufts Cummings School of Veterinary Medicine, North Grafton. MA [email protected]

Genetic tests are power tools, whose use can have a significant positive or negative impact on a breed’s gene pool. As with all power tools, they should come with an instruction manual on safety and their proper use.

The quantity and commercial availability of genetic tests offered for making breeding decisions are rapidly increasing. Breeders must understand the types of genetic tests that are available (phenotypic diagnostic tests, direct mutation DNA tests, linked marker-based DNA tests, susceptibility allele tests for complexly inherited disorders, pedigree and molecular genetic coefficients, EBVs and GBVs, etc.), and specifically what these tests tell them about the cats and dogs being tested. Along with the types of tests available, breeders must understand their proper use. Many of these issues are discussed in the article “Maneuvering the Maze of Genetic Tests: Interpretation and Utilization” (http://www.vin.com/proceedings/Proceedings.plx?CID=TUFTSBG2011&Category=10236&PID=6825 6&O=Generic)

The fact that a genetic test exists does not automatically qualify it for global utilization. With the plethora of genetic tests and their commercialization comes a realization that breeds can be tested into oblivion with selection that often has no bearing on health or quality. There are historical records of how improper use of genetic tests have reduced breed genetic diversity, as well as increased the frequency of other deleterious genes.

Selection is what created breeds, and selection is what will maintain breeds and improve their genetic health. Selection should be directed toward specific goals that directly improve the breed. Positive selection towards breed standards should ensure that they are not linked to disease liability. These may be conformational, behavioral, and/or working standards. Selection against disease liability should have a goal of preventing genetic disease without significantly eliminating breeding lines or restricting breed genetic diversity.

Genetic tests, pedigree and molecular genetic coefficients, and mating practices are tools that can allow the breeder to achieve defined breeding goals. When breeders begin to use these tools as the goals themselves, positive selective pressure is reduced, and breed gene pools will drift. Breeders must not lose sight of the fact that they are breeding entire individuals, and not a heart, an eye, a hip, or a coefficient number.

When evaluating an individual for breeding, the breeder must objectively assess the positive and negative traits and disorders displayed. Knowledge of the common hereditary disorders in the breed is important, as is their available genetic screening tests. For most dog breeds, these are listed in their breed page on the Canine Health Information Center website (www.caninehealthinfo.org/breeds). A similar website for cat breeds does not exist, however the Feline Advisory Bureau has a website detailing genetic disorders of cat breeds (www.fabcats.org/breeders/inherited_disorders).

Traits requiring selection in a mating should be listed and prioritized. Disorders that cause morbidity or mortality should have a high priority in selection. Traits and disorders caused by simple Mendelian genes can be changed and eliminated in a single generation. However, breeders should recognize that undesirable genes can be eliminated without eliminating breeding lines and affecting breed genetic diversity.

97

With testable simple Mendelian recessive genes causing genetic disorders, quality carriers can be breed to normal-testing mates and never produce the disorder. Quality normal-testing offspring should replace the carrier parent for breeding in the next generation to continue the breeding line. In this way, you lose the single testable gene, but continue the breeding line. Genetic tests should increase the options for breeding, and not limit them.

The typical response of a breeder on being informed of a carrier genetic test result is to remove the prospective breeding individual from a breeding program. If a majority of breeders do this, it can significantly limit the gene pool diversity of the breed. If an owner would breed an individual if it tested normal for a genetic disease, then a carrier result should not change that decision. A direct genetic test for a simple recessive trait does not alter WHO gets bred, only WHO THEY GET BRED TO (Henthorn P, personal communication).

Aside from preventing the production of affected individuals, breeders should select against placing new carrier-testing offspring into breeding homes. Carrier to normal matings produce on average, 50% carriers and 50% normal-testing offspring; a much higher carrier frequency than most breed-related disease liability genes. It is important to progressively decrease the frequency of deleterious genes in a breed, to increase breeding choices. This becomes especially important when there are several testable genes in a breed. With high carrier frequencies, selection can become more of an effort to prevent disease than to create the most desirable breed representative.

Complexly inherited traits will usually require more than one generation of selection to alter the genetic load of liability genes. Genetic selection should rely on genetic tests or phenotypic evaluations that are reflective and associated with causative genes. With complexly inherited traits (and with simple recessive traits that have no test for carriers), the phenotype of first-degree relatives (siblings, parents, and siblings of parents) best represent the range of liability genes that may be carried by the prospective breeding individual. This “breadth of pedigree” analysis can be evaluated through estimated breeding values (EBVs), or vertical pedigrees on the OFA website (www.offa.org).

Prospective mates should be listed and rated for the traits and disorders, in order to see which individuals might provide the greatest selective pressure for the most important traits. If an individual is highly desirable due to its traits and ability to pass them on, but also has several deleterious genes identified through genetic testing; then a parent, sibling, or prior-born offspring may provide the desired combination of traits and genetic test results.

Once a breeder has prioritized the traits and disorders that could undergo selection, (s)he must decide which will undergo selection in the next mating. The more traits that are undergoing selection; there will be less selective pressure that can be applied to any single trait.

As selection pressure is diminished by selection for test results that do not affect individual health and fitness, these should be avoided. Some commercial companies counsel to use genetic tests or coefficients as breeding goals. These include manipulation of MHC (major histocompatibility complex) haplotypes, or whole-breed outbreeding recommendations.

Certain specific MHC haplotypes are found to be linked to susceptibility for specific genetic disorders. However, general individual homozygosity or breed haplotype frequencies of the MHC loci by themselves have not been linked to disease or impaired health. In a study of semi-feral village dogs from

98 around the world, it was found that; 1) they share many of the same MHC haplotypes with pure-bred dogs, 2) they have many unique haplotypes that are not found in pure-bred dogs, and 3) pure-bred dogs also have many unique haplotypes that are not shared with village dogs. Pure-bred dogs do show increased homozygosity of MHC loci consistent with their large haplotype blocks and long linkage disequilibrium, however their predicted genetic depletion versus village dogs was not found (Kennedy LJ, et. al.: Do village dogs retain more major histocompatibility complex diversity compared to pedigree breed dogs? Poster presentation at the 7th International Conference on Advances in Canine and Feline Genomics and Inherited Disease, Cambridge, MA).

There is a movement to recommend generalized outbreeding programs for breeds to ostensibly retain genetic diversity. However, the types of matings used (linebreeding versus outbreeding) do not change gene frequencies. It is the selection of breeding animals that alters gene frequencies. The lecture notes “Inbreeding, Outbreeding and Breed Evolution” in the 6th Tufts Canine & Feline Breeding and Genetics Conference proceedings provide further depth to this issue.

Breeders must be wary of commercial offerings of genetic tests for genes that have not been proven to cause disease in their breed. This includes testing panels of collections of identified disease liability genes. Just because a gene is linked to disease in one breed does not automatically mean that it is linked to disease in all breeds. Causality or liability must be validated in each breed. If causality cannot be documented, then unwarranted selection just puts unnecessary pressure on the breed gene pool, and reduces the selective pressure on traits that are actually important to the breed.

Selection should be directed for specific desirable traits, and against disease liability genes. Efforts should be made to avoid the loss of quality breeding lines and genetic diversity in mating decisions. The most important aspect of maintaining breed genetic diversity is avoidance of the popular sire syndrome. Expanding or large, stable breeding populations are the best buffer against gene loss. Genetic tests provide excellent tools for breed improvement, and their proper utilization will allow breeders to see continued improvement in health and quality.

99 6th Tufts’ Canine and Feline Breeding and Genetics Conference

Poster Abstracts

Title: Name:

A Web Resource on DNA Tests for Canine and Feline Jeffrey Slutsky, Karthik Raj, Scott T Yuhnke, Hereditary Diseases and Urs Giger

Prevalence of Variant Alleles Associated with Meryl P. Littman, Michael G. Raducha, and Protein-losing Nephropathy in Soft Coated Wheaten Paula S. Henthorn Terriers

You’re getting on my nerves! The feline orofacial pain Barbara Gandolfi, Claire Rusbridge, Richard syndrome Malik and Leslie A. Lyons

The geographic diversification of domestic cats Razib Khan, Alejandro Cortes, Hasan Alhaddad, and Leslie Lyons

Who’s behind the mask and the cape? Asian Leopard Gershony LC, Cortes A, Penedo MCT, Davis Cat’s agouti allele affects coat colour phenotype in BW, Murphy WJ and Lyons LA Bengal cat breed

Genetic and Phenotypic Heterogeneity in Canine Aušra Milano, Gustavo D. Aguirre, Gregory M. Progressive Retinal Atrophy Acland, Orly Goldstein, Sue Pearce-Kelling

Publishing health data using open access, Nick Sutton, Aimee Llewellyn customised online platforms, and the benefits to researchers, breeders, and the public

Constrictive Myelopathy: a cause of hind limb ataxia Kathleen L. Smiler, Jon S. Patterson unique to Pug dogs?

Genetics and canine kidney disease: A risk locus in Andrew L. Lundquist, Noriko Tonomura, Ross Boxers with renal dysplasia identified by genome-wide Swofford, Michele Perloski, Katarina Tengvall, association Ake Hedhammar, Kerstin Lindblad-Toh

PennGen: Characterization of Metabolic and Molecular Caitlin A. Fitzgerald, Patricia O’Donnell, Karthik Genetic Defects in Dogs and Cats Raj, Michael Raducha, Ping Wang, Kate Berger, Margaret L. Casal, Peter J Felsburg, Paula S Henthorn, Mark E. Haskins, and Urs Giger

Congenital Hypothyroidism with Goiter in Cats due to Karthik Raj, Catherine V. Morrow, Anne Traas, a TPO Mutation Angela M. Erat, Marisa Van Hoeven, Hamutal Mazrier, Mark E. Haskins, and Urs Giger

Selection and the Co-Evolution of Breeds and Jerold S Bell Disease-Liability Genes

Population Genetic Studies and Gene Dynamics of Jerold S Bell Dog and Cat Breeds

100 A Web Resource on DNA Tests for Canine and Feline Hereditary Diseases

Jeffrey Slutsky, Karthik Raj, Scott T Yuhnke, and Urs Giger and the WSAVA Hereditary Disease Committee

Section of Medical Genetics (PennGen), School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA.

Following the first identification of a disease-causing mutation in dogs in 1989, and the more recent completion of the canine and feline genome sequences, much progress has been made in the molecular characterization of hereditary diseases in dogs and cats.

To increase access to information on diagnosing hereditary diseases in dogs and cats, a web application has been developed to collect, organize and display information on available DNA tests and other supporting information, including gene and chromosomal locations, mutations, primary research citations, and disease descriptions. The DNA testing information can be accessed at PennGen under the tab ‘Tests Available at Labs Worldwide’ at the URL: http://research.vet.upenn.edu/WSAVA-LabSearch. There are currently 170 molecular genetic tests available for hereditary diseases in dogs and cats offered by 54 laboratories worldwide.

This tool should provide clinicians, researchers, breeders and companion animal owners with a single comprehensive, up-to-date and readily searchable webpage for information regarding hereditary disease testing.

Supported in part by the WSAVA Hereditary Disease Committee, Waltham and NIH OD 010939.

101 Prevalence of Variant Alleles Associated with Protein-losing Nephropathy in Soft Coated Wheaten Terriers

Meryl P. Littman ([email protected]), Michael G. Raducha, and Paula S. Henthorn University of Pennsylvania School of Veterinary Medicine, Philadelphia, PA.

Variant alleles in NPHS1 and KIRREL2, the genes which encode the slit diaphragm proteins nephrin and filtrin/Neph3, respectively, were previously found associated with protein-losing nephropathy (PLN) in Soft Coated Wheaten Terriers (SCWT) by a genome-wide association study and subsequent gene sequencing of candidate genes in a statistically significant interval that differed among dogs with PLN compared with geriatric (14-18 year old) SCWT. Genotyping assays were developed for both of the single nucleotide polymorphisms (SNPs) in these genes that are in linkage disequilibrium in the breed. Homozygous positive dogs were shown to be at highest risk for the development of PLN, heterozygous dogs were at intermediate risk, and homozygous negative dogs were at low risk for the development of PLN.1

A prevalence study was performed to ascertain if breeders could safely remove carrier dogs in one generation. Cheek swab, blood, or semen samples were tested from 1549 SCWT dogs of all ages (median 4 yrs). Haplotypes are described as 1-1 (homozygous negative), 1-2 (heterozygous), and 2-2 (homozygous positive) for the PLN-associated variant alleles. The following table shows the frequencies found in various countries.

1-1 1-2 2-2 Variant Allele % % % Frequency (%) USA, n=1095** 34 47 19 43 (Hardy-Weinberg expected frequencies in the USA) (33) (49) (18) Canada, n=155 42.5 44.5 13 35 Total USA and Canada, n=1250** 35 47 18 42 Nordic Countries, n=125 42 44 14 36 UK/Ireland, n=119 66 24 10 22 Other (Australia, Poland, Argentina), n=55* 55.5 42 3.5 25 Total all countries, n=1549 (Unknown Sex, n=13) 39 44 17 39 Females, n=898** 39 44 17 39 Males, n=639 39 44 17 39 *Includes 1 Mi, undetermined NPHS1; 1-2 KIRREL2 **Includes 1 Fi, 1-2 NPHS1; 1-1 KIRREL2

Without genetic counseling with the knowledge of these haplotypes and assuming random breeding, the variant allele frequency would remain 43% in the USA. This high frequency indicates that it would be unwise to cull all carriers (1-2 or 2-2 dogs) of the variant alleles in one generation, thereby risking loss of genetic diversity, increased inbreeding, and the potential of increasing the incidence of other deleterious genetic traits. An approach to avoid producing high

102 risk homozygous positive (2-2) dogs would be to preferably breed desirable heterozygous (1-2) or homozygous positive (2-2) dogs to homozygous negative (1-1) dogs.

Instructions for DNA submissions are available at www.scwtca.org/health/dnatest.htm.

1. Littman MP, Wiley CA, Raducha MG, Henthorn PS. Glomerulopathy and mutations in NPHS1 and KIRREL2 in soft-coated Wheaten Terrier dogs. Mamm Genome 2013;24:119-126.

103 You’re getting on my nerves! The feline orofacial pain syndrome. Barbara Gandolfi, Claire Rusbridge, Richard Malik and Leslie A. Lyons The health of the Burmese breed is endangered by several diseases, such as hypokalemia, Burmese craniofacial defect, flat-chested kittens, an acute teething disorder, diabetes mellitus , and Feline Orofacial Pain Syndrome (FOPS). FOPS is characterized by an episodic, typically unilateral, discomfort with variable pain-free intervals. In many patients discomfort is triggered by movements of the mouth such as eating, drinking or grooming. Affected cats are most commonly presented with exaggerated licking and chewing movements, and pawing at the mouth. More severe cases develop self-mutilation of tongue, lips and buccal mucosa. Due to the severity of the lesions, many patients display anorexia. The syndrome is often recurrent and with time may become unremitting, with up to 10% of the cases being euthanized as a consequence of the condition. This condition is seen in a variety of feline populations, although Burmese cats make up the great majority of cases, suggesting a genetic basis for the syndrome. A genome- wide case-control association study that aimed to localize a the orofacial pain syndrome (FOPS), using the Illumina Infinium Feline 63K iSelect DNA array was performed on 24 cases and 50 healthy controls. The study resulted in the identification of a locus on cat chromosome C1 associated with FOPS. Preliminary data suggest an association on cat chromosome C1, within the low density lipoprotein receptor-related protein 1 gene (LRP1). The protein expressed in the central nervous system has been implicated in other pain syndromes and recent studies demonstrate that the gene is involved in migraine without aura. The length of the human transcript is 14,897 bp translated into 4544 amino acids, the gene contains 89 coding exons and is one of the largest genes in the human genome. Sequencing of the feline gene revealed several polymorphisms under consideration.

104 The geographic diversification of domestic cats

Razib Khan1, Alejandro Cortes1, Hasan Alhaddad1, and Leslie Lyons1 Department of Population Health and Reproduction1, University of California, Davis, CA 95616

Felis silvestris catus, the domestic cat, diverged ~10,000 years ago from populations of Felis silvestris lybica, the African wildcat. This result is supported by remains of cats inhumed with humans on the island of Cyprus and mtDNA phylogenies. More recently, within the last ~150 years there has been development of “fancy” breeds such as the Persian. But there are gaps in the evolutionary history of the cat between the initial domestication events in the Middle East, and the efforts of modern breed associations in developing specialized varieties.

To further explore variation in Felis silvestris with the aim of inferring historical dynamics, phylogenetic analysis were performed on over 3,000 individuals from 30 breeds and 30 regional populations using 38 autosomal microsatellites. These are inclusive of non-breed cats from six continents, breeds, wildcats, and hybrids. Genetic diversity and distance estimates were generated. Principle coordinate analysis was used to visualize distances. Analysis of population clustering utilizing the STRUCTURE package was performed. Finally, the TREEMIX package generated graphs of relationships across the populations, and migration events between lineages.

Over the data STRUCTURE analyses with >20 explicit clusters were less informative. The initial bifurcation occurred between domestic lineages and wildcats. Subsequent splits occurred between European, Middle Eastern, and Asian lineages. Known breeds’ attested histories were confirmed in terms of derivation from specific regional populations. Breed specific admixture events were identified. Geopolitical contours were recapitulated by genetic population structure. The cats of Iran and Iraq formed a distinct cluster from those of the Levant, possibly reflecting ancient divisions in the Middle East. Other genetic relationships are only comprehensible through understanding of local histories of colonialism. The population structure of domestic cats reflects local interactions with humans. Finally, preliminary replications of some of these analyses using 150 and 63,000 SNP data sets were examined.

105 Who’s behind the mask and the cape? Asian Leopard Cat’s agouti allele affects coat colour phenotype in Bengal cat breed

Gershony LC1, Cortes A1, Penedo MCT2, Davis BW3, Murphy WJ3 and Lyons LA1

1Department of Population Health and Reproduction, School of Veterinary Medicine, University of California - Davis, Davis, CA, USA 2Veterinary Genetics Laboratory, School of Veterinary Medicine, University of California - Davis, Davis, CA, USA 3Department of Veterinary Integrative Biosciences, College of Veterinary Medicine and Biomedical Sciences, Texas A&M University, College Station, TX, USA.

Coat colours and patterns are highly variable in cats and determined by several genes. The charcoal coat pattern inheritance in Bengal cats appears as an incomplete melanism, thus the agouti signalling protein gene (ASIP) was investigated as a candidate gene for this phenotype. DNA was isolated from buccal swabs obtained from 72 Bengal cats, where 49 were presumed to be charcoal. The coding region of ASIP was amplified by polymerase chain reaction and subsequently directly sequenced. The resulting sequences were compared to that of ten Asian leopard cats and three control domestic cats. Polymorphisms were investigated within the gene. Two non-synonymous SNPs were observed in exon 2 (c.41G>C and c.142T>C) when comparing the control domestic cat sequence with the leopard cat sequence, resulting in amino acid changes in the leopard cat (Cys14Ser and Ser48Pro, respectively). One synonymous single-nucleotide polymorphism (SNP) was found in exon 3, substituting a cytosine for adenine in the leopard cat (c.162C>A). Forty-three charcoal cats presented as compound heterozygotes at ASIP, consisting of an Asian leopard cat allele and a domestic cat non-agouti allele (a). The compound heterozygote state suggests that the interaction between the Asian leopard cat allele and the domestic cat allele allowed for the recessive non-agouti allele to influence the markings of the hybrid Bengal cat producing a darker, yet not completely melanistic, coat pattern. This study presents the first validation of a Leopard cat allele segregating in the Bengal breed affecting the overall phenotype of the pelage.

1) Further investigation should be conducted to assess similar interactions in other genes, and how they would affect the accuracy of genetic tests within this breed.

2) Further investigation should be performed to better illuminate the potential allelic interactions, and consequential phenotypic expression, within this hybrid breed.

106 Genetic and Phenotypic Heterogeneity in Canine Progressive Retinal Atrophy

Aušra Milano1, Gustavo D. Aguirre2, Gregory M. Acland3, Orly Goldstein3, Sue Pearce-Kelling1

1 Optigen, LLC, Ithaca, NY; 2 School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA 3 Baker Institute, Cornell School of Veterinary Medicine, Ithaca, NY

Mutations causing Progressive Retinal Atrophy (PRA) are the predominant cause of hereditary blindness in the domestic dog. Although over a dozen PRA mutations have been identified, including prcd-PRA which has been observed in over 25 breeds, many PRA mutations remain uncharacterized. During its nearly 15 years of operation, OptiGen has collected DNA, pedigrees and detailed phenotype descriptions from hundreds of dogs that have been diagnosed with PRA by veterinary ophthalmologists. This PRA research sample set includes over 500 samples and 100 breeds. DNA testing of these samples has revealed that many breeds harbor multiple forms of PRA, often with similar clinical symptoms. Here we present the distribution of prcd and other PRA-causing mutations that have been assayed within OptiGen’s PRA research sample set. Breeds in which multiple forms of PRA are known to segregate are presented as well as phenotypic variations in the PRA cases. Collaborative research projects that can make use of these samples are encouraged.

107 Publishing health data using open access, customised online platforms, and the benefits to researchers, breeders, and the public

Nick Sutton, Aimee Llewellyn The Kennel Club, 1-5 Clarges Street, Piccadilly, London W1J 8AB [email protected]

The Kennel Club has been recording and publishing health test results for DNA tests and the British Veterinary Association/Kennel Club Health Schemes (hip dysplasia, elbow dysplasia, and eye schemes) since 1965 in firstly the Kennel Club Gazette, and latterly the Breed Record Supplement. The Kennel Club initially launched Online Services to provide general health information. Then in May 2011 the bespoke online interface, Mate Select, was established specifically to publish and disseminate breed population and health data recorded on the Kennel Club Breed Register. Now, 2 years on, Mate Select, as an online publication resource is being reviewed with the objective to establish what, if any, impact this method of health data reporting has had on the accessibility of canine health information.

Prior to the launch of Mate Select, it had been recognised that while publishing health information was valuable to dog breeders, there were numerous practical limitations to doing so in hard-copy publications, such as the Breed Record Supplement. A primary limitation being that this form of publication is not open-access, or easily searchable - particularly over time. Records were published in the quarterly Breed Record Supplement, at an average of approximately 40,000 individuals each year. Conversely, Mate Select is a free unrestricted online interactive tool which receives approximately 300,000 searches each month, designed to provide breeders with free, accessible health information for individual dogs. This provides access to any breeder, enabling them to make informed choices which can have a positive impact on the health of any potential puppies produced, as well as the breed in general. The system was produced with expansion in mind and is able to accommodate advances in molecular and population genetics. All of the results published are linked to each individual dog’s record within the Kennel Club database, allowing imputation using customisable, defined criteria for on-going assessment and monitoring. This resource is particularly useful as guidance when prioritising health conditions or, establishing breeding restrictions such as Kennel Club DNA Control Schemes.

Mate Select in its current state, is divided into tools that reflect an individual dogs health (such as gene test results), and resources for considering breed-wide implications of individual mating selections, such as inbreeding. Together, this provides dog breeders with efficient and practical resources for reducing the risks of specific heritable condition and incorporating inbreeding and genetic bottle-neck mitigation strategies into their breeding plans, particularly in the selection of breeding stock. The Health Test Results Finder, which manages over 100,000 online searches each month, publishes all health results for approximately 80 breed-specific, individual single-gene mutation DNA tests. BVA/KC Health Schemes published records currently consist of over 260,000 hip scores, 21,000 elbow scores, 116 Chiari malformation/Syringomyelia (CM/SM) scores (introduced in 2012), as well as the results for over 134,000 clinical eye examinations. Recording of either DNA test results, or clinical examination “schemes” is expandable under the system and allows for improvements to the confirmation of data – such as parentage profiling (in the case of assigning hereditarily clear status) and notations where examined dogs have been microchip- confirmed for DNA tests. In addition, the data yielded from dogs undergoing clinical examination schemes allows for the development of tools for the future, such as Estimated

108 Breeding Values (EBVs). Linking the data to a pedigree or dog registration record adds confidence to the examination status recorded and allows population geneticist to review scheme uptake and results to calculate a threshold at which EBVs for a population can be developed with confidence. Publication of tools that require frequent updating and re- calculation, such as EBVs would be impossible without an online interface.

By definition, the development of a breed creates a population that can be increasingly limited without outcrossing or otherwise introducing new genes into the system. Therefore, tools that can provide a means to slow the rate of inbreeding, and/or reduce individual litter inbreeding coefficients are of value to dog breeders. Mate Select tools, developed in conjunction with the Kennel Club Genetics Centre at the Animal Health Trust, provides three coefficient of inbreeding (COI) “calculators”: Breed COI, Individual (dog) COI, and the Mating COI. For the dog breeder, the tool most practical is the Mating COI calculator. This allows dog breeders to perform hypothetical matings using a dam and sire they are considering to estimate the inbreeding coefficient for the resulting puppies. This number can then be compared to the breed average (which is provided for comparison after each search), to encourage breeding below the breed average and thus a decrease in the overall degree of inbreeding. This is, again, a resource that would be impossible without an interactive online interface. In the long term, breed-wide COI data can be assessed to monitor change, and encourage improvement.

In summary, by recording health test results against pedigree data Mate Select provides a robust, diverse and unique data resource that enables the public to make informed decisions. Using a freely accessible, searchable, and interactive interface has significant advantages over hard-copy publication. There is every indication that publishing health test results allows for the reduction or elimination of some heritable diseases, and therefore any robust method that makes this information more efficient and available is to every dog’s benefit. Although it is too early to determine the full impact that Mate Select has had on the health of the UK canine population, it is hoped that through improved accessibility and transparency of published test results, breeding trends towards the production of healthier dogs will occur more rapidly.

109 Constrictive Myelopathy: a cause of hind limb ataxia unique to Pug dogs?

Kathleen L. Smiler, DVM, DACLAM, Consultant, PO Box 429, Lakeville, MI 48366, Email [email protected] (248)-953-3182

Jon S. Patterson, DVM, PhD, DACVP, Michigan State University College of Veterinary Medicine, 163 DCPAH Building, 4125 Beaumont Rd., Lansing, MI 48910-8104, Email [email protected] (517) 353-9471

BACKGROUND Recently, a previously unreported condition termed “constrictive myelopathy” was described in 11 adult Pug dogs (J Am Vet Med Assoc 2013; 242:223-229). The paper reported a progressive incoordination and weakness of the hind limbs resulting from a constriction of the spinal cord at the thoracolumbar junction, and associated with malformations of the articulations of vertebrae in this area. The degenerative condition often progressed to paraplegia, with urinary and/or fecal incontinence. Despite surgical treatment, neurologic disease persisted or progressed. This myelopathy is seemingly unique to and reportedly rare in purebred Pug dogs, although anecdotal evidence suggests that the vertebral malformations (hypoplasia and/or aplasia of caudal articular processes) are relatively common in the breed, as supported by imaging studies. Authors of the published study hypothesize that the vertebral anomalies may represent a heritable condition in Pugs, and that instability at the thoracolumbar junction associated with the anomalies leads to the formation of a circumferential fibrous band which constricts the spinal cord. A case study of one Pug diagnosed with constrictive myelopathy at age 6.5 years, and euthanized at age 14 years is presented.

CASE DESCRIPTION A spayed female purebred Pug dog, was initially observed at age 6.5 years to have reluctance climbing stairs and urinary and fecal incontinence. Neurologic examination revealed bilateral hind limb weakness and ataxia, with increased tone in the left hind. Hind limb proprioceptive deficits were present bilaterally, and the cutaneous trunci response was absent caudal to T13. Radiographs and computed tomography (CT) suggested hypoplasia of caudal articular processes of T10-T12, and MRI suggested spinal cord compression at T12-T13. A diagnosis of "pug myelopathy" was made, and a dorsal laminectomy was performed in the area of compression. At surgery, a circumferential band of mature fibrous tissue, seen to compress the spinal cord, was removed. After surgery, the dog had improved hind limb function, and better control of urination and defecation. Approximately 6-7 months after surgery, however, the hind limb ataxia worsened, and a CT/myelogram suggested a demyelinating condition. The dog was treated with various doses of prednisone and underwent acupuncture therapy for 3 months at age 7.5 years. By age 8, the dog had complete urinary retention incontinence, and by age 9, would walk only if supported, and relied on front limbs to pull herself along. At age 12, a DNA sample was tested at University of Missouri for the degenerative myelopathy (DM) gene mutation, and results were negative. Approximately 1 week prior to euthanasia, the dog began having difficulty using one front limb, and euthanasia was elected.

Complete necropsy was done at the Michigan State University Diagnostic Center for Population and Animal Health (DCPAH). There was marked bilateral atrophy of the caudal thigh muscles, muscles over the pelvis, and epaxial muscles of the thoracic and lumbar spine. Slight scoliosis of

110 the vertebral column to the right was noted at the level of T6-T7, and there was mild bridging spondylosis on the ventral aspect of the vertebral bodies at theT6-T7 intervertebral space.

The entire vertebral column, containing the spinal cord, was placed in 10% neutral buffered formalin, and following fixation, the spinal cord was removed and vertebrae were disarticulated and examined. It was difficult to draw conclusions regarding the caudal articular processes of the T11, T12, and T13 vertebrae, at the site of surgery 8 years prior, but there appeared to be asymmetry with respect to size for the paired articular processes (right vs. left) of T12 and T13. Histologically, there was severe segmental chronic myelomalacia in the T12 and T13 spinal cord segments, with Wallerian degeneration cranial and caudal to this area. The leptomeninges were moderately to markedly thickened by dense fibrous tissue from T10-T13, with areas of arachnoid hyperplasia and dural fibrosis. Focal poliomyelomalacia in the C6 spinal cord segment was noted, and close inspection of the cervical vertebral column revealed dry, flaky intervertebral disc material at C5-C6 and C6-C7, suggesting a disc degeneration.

The final diagnosis was severe segmental chronic degenerative myelopathy at T12-T13, with meningeal fibrosis (T10-T13) and Wallerian degeneration. This appeared to be the major lesion, consistent with the 7 to 8-year history of progressive hind limb weakness, ataxia, and paralysis, and consistent with what was described by the surgeons who treated the dog. The more recent spinal cord lesion in the C6 segment involved primarily the gray matter and was consistent with an acute intervertebral disc extrusion that then became chronic.

CONCLUSIONS AND SIGNIFICANCE The Pug Dog Club of America (PDCA) has recognized the widespread anecdotal reports of hind limb ataxia and paralysis in Pugs and is committed to encouraging research to better understand spinal disease including “constrictive myelopathy,” and to effective strategies to manage the condition and reduce its incidence. The poster authors have initiated proposals to better characterize both the vertebral and neurological lesions, and to identify unique features which might distinguish constrictive myelopathy from other conditions with similar clinical presentations in Pugs. Blood and tissue samples will be banked for eventual DNA analysis as genetic components of this disease are considered.

To enhance awareness and accumulate data, a public outreach for case histories of Pugs with hind limb ataxia and weakness is ongoing, utilizing social media, announcements to Pug group media, presence at a national breed club dog show, and specific contacts with Pug rescue organizations. The Pug rescue organizations are increasingly burdened by the surrender of ataxic and paralyzed dogs, and it is difficult to find foster or permanent homes that will provide the skilled care required (especially those with urinary incontinence complications). The diagnostic procedures and long-term care will incur substantial costs for veterinary and rehabilitation palliative therapy. The complex of diseases causing hind limb ataxia and weakness in Pugs, possibly complicated by inherent vertebral malformations, is a formidable problem in the breed.

Figures in the poster will include various imaging results obtained for this case. MRI, CT myelogram, radiographs including post mortem; photographs of gross vertebrae after dissection, and photomicrographs of histopath of cord, etc.

111 Genetics and canine kidney disease: A risk locus in Boxers with renal dysplasia identified by genome-wide association

Andrew L. Lundquist1, Noriko Tonomura1,2, Ross Swofford1, Michele Perloski1, Katarina Tengvall3, Ake Hedhammar4, Kerstin Lindblad-Toh1,3

1 Broad Institute of Harvard and MIT, Cambridge, MA, USA, 2 Cummings School of Veterinary Medicine, Tufts University, North Gratton, MA, USA, 3 Science for Life Laboratory, Department of Medical Biochemistry and Microbiology, Uppsala University, Sweden, 4 Department of Clinical Sciences, Swedish University of Agricultural Sciences, Uppsala, Sweden Email: [email protected]

Kidney disease is common in dogs and many breeds are affected. Dogs can be affected early in life by various forms of inherited kidney disease or chronic kidney disease can manifest later in life from a variety of causes. Previous studies have shown that renal failure is among the top five causes of death in dogs and up to 30% of geriatric dogs have chronic kidney disease. Historically, certain breeds have been affected with a specific type of kidney disease, suggesting a genetic cause. Current canine genetic tests available include testing for hereditary nephritis in Samoyed and Cocker Spaniel (vetGen), cystinuria in Newfoundland (vetGen), and primary hyperparathyroidism in Keeshonden (Cornell University). Our group is focused on identifying the genetic cause of various forms of inherited canine nephropathy through genetic association studies. We are looking for collaborations with owners, breed clubs, and veterinarians to identify cases of canine kidney disease including: breed specific inherited nephropathies, isolated or litter specific cases of spontaneous kidney disease, and cases of adult dogs with chronic kidney disease. Previously, we helped identify the risk alleles for renal amyloidosis in Shar Peis and primary hyperparathyroidism in Keeshonden. Here we will discuss our efforts to identify risk alleles for renal dysplasia in Boxers.

Identification of genetic risk factors for renal dysplasia in dogs is essential as there is no treatment and affected dogs progress to renal failure and death at a young age. A genetic test for renal dysplasia is available, however its validity across species has come into question and the scientific community has called for additional validation of the test. We previously conducted a genome-wide association study using the Canine HD BeadChip comparing 17 US Boxers with renal dysplasia (age < 5) to 40 older Boxers (age > 10) with no known kidney disease. No association was detected at the locus defined by the currently available genetic test. Association analyses suggest a risk allele adjacent to a gene previously implicated in human hypodysplasia, a common cause of pediatric kidney disease. Sequencing the coding region of our candidate gene did not reveal a causative mutation, though variants nearby suggest a haplotype associated with disease. We are currently analyzing a 4 MB region surrounding the risk locus with targeted sequence capture to identify the causative variant(s) and we are working to acquire additional cases of renal dysplasia in Boxers and other breeds as these are essential to help validate our findings. These studies will help us dissect the genetics of canine renal dysplasia, improve our understanding of renal development in dogs and humans, and determine the appropriate genetic testing strategies for prevention.

112 PennGen: Characterization of Metabolic and Molecular Genetic Defects in Dogs and Cats

Caitlin A. Fitzgerald, Patricia O’Donnell, Karthik Raj, Michael Raducha, Ping Wang, Kate Berger, Margaret L. Casal, Peter J Felsburg, Paula S Henthorn, Mark E. Haskins, and Urs Giger Section of Medical Genetics, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA.

The Section of Medical Genetics at University of Pennsylvania School of Veterinary Medicine has actively pursued the diagnosis and management of hereditary diseases and genetic predispositions to disease in companion animals for the past 4 decades.

The specialty Pediatrics and Genetics Clinic and diagnostic and research laboratories have been characterizing many inherited traits in dogs and cats from the clinical features to the metabolic and molecular genetic defects.

The Metabolic Genetics Screening Laboratory supported by an NIH grant conducts routine analyses of amino acids, organic acids, and carbohydrates in urine samples for various inborn errors of metabolism such as many storage diseases, lactic and methylmalonic aciduria, cystinuria, and Fanconi syndrome. Particularly, the NIH grant also focuses on mucopolysaccharidosis (MPS), mannosidosis, and gangliosidosis, which are diagnosed by urinary spot tests and enzyme assays. Moreover, affected animals serve as excellent disease models of human disease.

Another area are hereditary blood disorders such anemia due to red cell defects (PK, PFK, osmotic fragility), bleeding disorders caused by coagulation factor (Factor VII and XI), and platelet disorders along with predisposition to infection resulting from white blood cell problems (X-SCID, LAD, avian tuberculosis). This laboratory also investigates canine and feline blood types and is offering typing service in case of incompatibility issues.

PennGen and the Josephine Deubler Laboratory, named in honor of Dr. Deubler (veterinarian, dog breeder, and dog show judge) were specifically established to provide genetic tests for veterinarians, breeders, and pet owners to assist in their effort to provide precise diagnosis and help with breeding of animals free of hereditary diseases known to particular breeds. The Laboratory offers DNA tests for genetic diseases found in dogs and cats mostly based upon the research performed by the investigators at Penn to identify affected, carrier (asymptomatic) or normal (clear) genotypes in pets. Tests offered by PennGen as well as other DNA testing laboratories worldwide can be found at http://research.vet.upenn.edu/WSAVALabSearch which is a searchable database by disease, breed, and laboratory.

PennGen provides various diagnostic genetic services and consultations for primary care veterinarians, veterinary specialists, breeders and pet owners in order to produce the healthiest dogs in each breed and to gain new knowledge and insight to these genetic diseases.

Supported in part by the National Institutes of Health (OD 010939), Canine Health, Winn Feline, and other foundations.

113 Congenital Hypothyroidism with Goiter in Cats due to a TPO Mutation

Karthik Raj, Catherine V. Morrow, Anne Traas, Angela M. Erat, Marisa Van Hoeven, Hamutal Mazrier, Mark E. Haskins, and Urs Giger

Section of Medical Genetics, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA.

Congenital hypothyroidism (CH) has been reported in many species; the hereditary forms can be divided into thyroid dysmorphogenesis and dyshormonogenesis. While thyroid hypoplasia has been described in dogs and cats, the molecular basis remains unknown. In contrast several breeds of dogs with goiterous CH, studied by Fyfe et al, were found to have deficient thyroid peroxidase (TPO) activity and disease-causing TPO gene mutations. The purpose of our study was to characterize a family of domestic shorthair cats with goiterous CH.

Clinical features included dwarfism and dullness, known as cretinism and seen with CH in all species, but also constipation and megacolon which are unique to cats with CH. Pedigree analysis documented an autosomal recessive mode of inheritance. Affected kittens developed a goiter and had low serum thyroxine (T4) and triiodothyronine (T3) when compared to controls, but high thyroid stimulating (TSH) hormone levels indicating thyroid dyshormonogenesis. Oral thyroid supplementation corrected the progression of clinical signs and prevented further constipation and reversed the megacolon.

The TPO enzyme activity was extremely low in hypothyroid cats when compared to that of normal cats. Genomic DNA and cDNA from affected, carrier, and normal cats were extracted and sequenced based upon primers developed from the feline genome database. A homozygous missense point mutation (c.1333G>A) in TPO, which results in an amino acid change (p.Ala445Thr), was discovered in affected cats and the mutant allele segregated within the family with goiterous CH. This is the first report of a TPO deficiency in cats. Unrelated domestic shorthair cats with goiterous CH did not have this same TPO mutation. The prevalence of this TPO mutation in the domestic cat population seems low, but CH is likely underreported in cats.

Supported in part by NIH OD 010939.

114 Selection and the Co-Evolution of Breeds and Disease-Liability Genes

Jerold S Bell, Tufts Cummings School of Veterinary Medicine, N. Grafton, MA USA [email protected]

Natural selection works against inherited traits and disorders that would reduce the ability to survive, thrive, and reproduce. Artificial selection can; reduce the frequency of disease-liability genes, be neutral to their propagation, or sometimes preferentially select for them. Selection must be appropriately applied in order to improve breed health.

Pure-bred dog and pedigree cat breeds evolved through selection for conformational, behavioral, and/or working standards. With extreme phenotypic selection, breeders have purposely selected for disease-liability, such as; the brachycephalic syndrome, excessive amounts of skin or skin folds, and overangulation.

Selection for traits has been linked to disease-liability, such as; hyperuricosuria (SLC2A9) in Dalmatians, cranio-facial defect (unpublished, Lyons) in Burmese, dermoid sinus (FGF3, FGF4, FGF19 and ORAOV1 duplication) in Ridgebacks, and osteochondrodysplasia (unidentified) in Scottish Folds. In some cases, the preferred trait can be genetically separated from the disease liability. In other cases, they are pleiotropic expressions of the same genotype.

Other disease liability genes are not linked to selection, but lay in the genetic background of breeds. Many of these are ancient mutations that preceded the separation of, and are shared by many breeds. These include complex disorders, such as; hip dysplasia, patella luxation, and diabetes mellitus (Types 1 & 2). Several ancestrally ancient mutations cause simple Mendelian disorders, such as; progressive rod-cone degeneration (prcd), multifocal retinopathy (cmr1), and hyperuricosuria (SLC2A9). Without direct selection, these can increase in frequency through the popular sire effect or genetic drift.

Some recommendations to improve the genetic health of breeds concentrate on selection to increase heterozygosity or minor allele frequencies. These methods; 1) do not select against disease-liability genes, 2) will not prevent the phenotypic expression of dispersed genes, and 3) may reverse the effects of positive selection through blind manipulation of minor alleles. Health- based selection should be specifically directed against deleterious traits and genes.

115 Population Genetic Studies and Gene Dynamics of Dog and Cat Breeds

Jerold S Bell DVM, Clinical Associate Professor, Dept. of Clinical Sciences, Tufts Cummings School of Veterinary Medicine [email protected]

(This article is based on a poster presented at the 7th International Conference on Advances in Canine and Feline Genomics and Inherited Diseases, Sweden 2012. It can be reproduced with the permission of the author.)

Breed Gene Dynamics

Each dog and cat breed has its own evolutionary history of founders, accumulated deleterious genes, population bottlenecks, popular sires, and geographical fragmentation. Some studies of dog and cat breeds focus on the inbreeding coefficients of individuals, and the effective population size of breeds as a measurement of their genetic vitality and ability to maintain themselves as pure breeds (Calboli et al. 2008, Genetics 179:593-601).

Most breeds started from a limited number of founders. As the population expands within a closed gene pool, it allows mating choices between individuals that are less closely related than the previous generation. This is shown by evaluating average 10 generation inbreeding coefficients (Mean 10 Gen IC). Early in breed development, inbreeding coefficients can be high due to inbreeding on a small founder population (as seen in the Borzoi and Burmese breeds), or breeding with a more diverse founder population (as seen in the Siberian Husky, Gordon Setter, and Cavalier King Charles Spaniel breeds).

116 As generational pedigrees extend beyond 10 generations, the IC Mean 10 Gen can decrease as populations utilize the breadth of their gene pool and the number of unique ancestors increase. When the Mean 10 Gen IC increases, it is usually because breeders are concentrating on popular sires. The Mean All Gen IC (homozygosity) necessarily goes up over time as a function of breed evolution. (The Mean All Gen IC of Burmese goes down in this example due to importation of Burmese with incomplete pedigrees.) The genetic health of dog and cat breeds is not a direct function of homozygosity or heterozygosity; but of the accumulation and propagation of disease liability genes.

Several researchers have found that dog breed genetic diversity is not a function of population size or average inbreeding levels (James 2011: Vet Journal 189:211-213, Bjornerfeld et al. 2008, BMC Evol Biol 8:28). Shariflou et al. (2011, Vet Journal 189:203-210 ) found that genetic diversity is not related to the

117 size of the breed, but to breeding practices and the even contribution of founding lines. The popular sire syndrome is the single most influential factor in restricting breed gene pool diversity.

Molecular genetic studies of cattle show limited genetic diversity in evolutionary founder populations (Bollongino et al. 2012, Mol Biol Evol. 2012 Sep;29(9):2101-4., The Bovine HapMap Consortium 2009 Science 324(5926):528-532). In spite of this, cattle breeds have propagated and are second only to dogs in mammalian genetic diversity.

Breed genetic health does not have to do with existing breed inbreeding coefficients, homozygosity, estimated number of founders, or other statistics. It has to do with reproductive ability and accumulated disease liability genes. Breed genetic health should be judged based on current breed health surveys.

Breeding Strategies

Some organizations have embraced the belief that close breeding is the cause of impaired breed health. They have adopted programs that restrict close breeding, and promote outbreeding to the least related individuals. This involves lowering mean inbreeding coefficients and/or increasing heterozygosity of SNPs or haplotypes.

Outbreeding programs are akin to a Species Survival Plan (SSP) that is utilized when attempting to “rescue” an endangered species. The vast majority of dog and cat breeds do not show evidence of genetic depletion such as; low reproductive success, and increased stillborn and neonatal mortality.

Recommendations to outbreed (only breed to those least related) homogenizes breeds and erases the genetic difference between individuals. It is a self-limiting process that requires matings be done between individuals who are genetically different from each other. Eventually there will be no more “lines” with differences. Everyone will be in the center, and no one at the periphery.

118 By erasing the genetic difference between individuals, this averts selective pressure for improvement. Breed gene pool diversity requires distinct lines in order to create selective pressure. A mix of breeding individuals from different lines within the breed maintains allelic polymorphism.

Breeders strive to select for healthy conformational, behavioral, and working standards for their breeds. Selection over time allows more individuals to conform to a standard.

Attempts to create heterozygosity for SNPs and haplotypes that have no defined positive or negative gene effect have as much a chance of reversing selection-based improvements as they have for being beneficial to a breed’s genetic health.

This has been shown in cattle breeds: Prioritization based on neutral genetic diversity may fail to conserve important characteristics in cattle breeds (Hall et al. 2012 J Anim Breed Genet 129(3):218- 225).

Prudent breeding practices allow some linebreeding, some outbreeding, and even occasional inbreeding; with different breeders maintaining breeding lines or crossing lines as they see fit. It is the different opinion and breeding actions of breeders that maintain breed diversity.

Genetic Health

We see increased genetic disease in pure-bred and cross-bred animals due to a lack of genetic testing and selection of breeding animals, and an associated increase in disease liability genes. Different mating types (inbreeding, linebreeding, outbreeding) are responsible for the expression of alleles in gene pairs, but not in allele propagation. Selection of breeding stock for the next generation, and their fecundity is what alters allele frequencies.

Genetic homozygosity is a function of speciation and breed formation. It is only detrimental if related to disease liability genes or impaired health. We must ensure that our selection recommendations improve breeds, and do not impede breeder efforts for progress in breed health, conformation, and function.

119 6th Tufts’ Canine and Feline Breeding and Genetics Conference

Articles

Title: Name:

A web resource on DNA tests for canine and feline Slutsky J, Raj K, Yuhnke S, Bell J, Fretwell N, hereditary diseases Hedhammar A, Wade C, Giger U.

Deciphering the genetic basis of animal domestication Wiener P & Wilkinson S

Both Ends of the Leash — The Human Links to Good Elaine A Ostrander Dogs with Bad Genes

Variation of cats under domestication: genetic Kurushima JD, Lipinski MJ, Gandolfi B, assignment of domestic cats to breeds and worldwide Froenicke L, Grahn JC, Grahn RA, Lyons LA random-bred populations

An insight into population structure and gene flow Leroy G, Vernet E, Pautet MB, Rognon X within purebred cats

Assessing the impact of breeding strategies on Leroy G & Rognon X inherited disorders and genetic diversity in dogs

How the Orthopedic Foundation for Animals (OFA) is Keller GG, Dziuk E, Bell JS tackling inherited disorders in the USA: Using hip and elbow dysplasia as examples

Comparative analyses of genetic trends and prospects Lewis TW, Blott SC and Woolliams JA for selection against hip and elbow dysplasia in 15 UK dog breeds

Prevalence of inherited disorders among mixed-breed Bellumori TP, Famula TR, Bannasch DL, and purebred dogs: 27,254 cases (1995-2010) Belanger JM, Oberbauer AM

Idiopathic Cystitis in Domestic Cats—Beyond the C.A.T. Buffington Lower Urinary Tract

120 The Veterinary Journal 197 (2013) 182–187

Contents lists available at SciVerse ScienceDirect

The Veterinary Journal

journal homepage: www.elsevier.com/locate/tvjl

A web resource on DNA tests for canine and feline hereditary diseases

Jeffrey Slutsky a, Karthik Raj a, Scott Yuhnke a, Jerold Bell b, Neale Fretwell c, Ake Hedhammar d, ⇑ Claire Wade e, Urs Giger a, a School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA, USA b Department of Clinical Sciences, Tufts Cummings School of Veterinary Medicine, North Grafton, MA, USA c UK Waltham Centre for Pet Nutrition, Freeby Lane, Fretwell, Leicestershire, UK d Department of Clinical Sciences, Swedish University of Agricultural Sciences, Uppsala, Sweden e Faculty of Veterinary Science, University of Sydney, New South Wales 2006, Australia article info abstract

Article history: Following the first identification of a disease-causing mutation in dogs in 1989 and the more recent com- Accepted 24 February 2013 pletion of canine and feline genome sequences, much progress has been made in the molecular charac- terization of hereditary diseases in dogs and cats. To increase access to information on diagnosing hereditary diseases in dogs and cats, a web application has been developed to collect, organize and dis- Keywords: play information on available DNA tests and other supporting information, including gene and chromo- Canine somal locations, mutations, primary research citations and disease descriptions. The DNA testing Feline information can be accessed at the URL: http://research.vet.upenn.edu/WSAVA-LabSearch. There are cur- Genetics rently 131 molecular genetic tests available for hereditary diseases in dogs and cats offered by 43 labo- Database Mutations ratories worldwide. This tool should provide clinicians, researchers, breeders and companion animal owners with a single comprehensive, up-to-date and readily searchable webpage for information on hereditary disease testing. Ó 2013 Elsevier Ltd. All rights reserved.

Introduction et al., 2005) and feline (Pontius et al., 2007) genome sequences, and their recent improved coverages and annotations (National Next to humans, the largest number of naturally occurring Center for Biotechnology Information, NCBI).1 Thus far, most of hereditary disorders and genetic predispositions to disease has the characterized hereditary disorders involve single gene defects been reported in dogs (Sargan, 2003; Giger et al., 2006; Bell with simple Mendelian inheritance and are mostly breed specific et al., 2012), followed by cats (Giger and Haskins, 2006; Pontius (Giger and Haskins, 2006; Giger et al., 2006). et al., 2007; Lyons, 2010, 2012). Notably, many hereditary disor- Knowing the specific molecular defect for a hereditary disease is ders in dogs and cats represent true homologues of genetic dis- valuable, since it offers the best opportunity to make a precise diag- eases in humans and thus serve as valuable naturally occurring nosis for an animal with clinical signs, helps to screen animals at risk disease models (Marschall and Distl, 2010; Mellersh, 2011). Since of developing the disease, permits identification of carrier animals many of these disorders are recessively inherited and occur with (heterozygous for a mutant allele but clinically healthy) and can high frequency in specific or related breeds due to common be used to test animals prior to breeding to assure that affected ani- inbreeding practices, they represent a serious health problem for mals are not produced in future generations (Giger et al., 2006; companion animals (Padgett, 1998; Vella et al., 1999; Giger et al., Lyons, 2010; Mellersh, 2011). The original research laboratories 2006; Asher, 2009; Hedhammar and Indrebø, 2011; Bell et al., where a disease-specific mutation is first discovered in a particular 2012). To address this issue, a thorough investigation of hereditary breed may or may not continue testing animals subsequent to the disorders, from clinicopathologic features to the molecular genetic completion of the relevant research. However, other university or basis of disease, has become a high priority. for-profit laboratories may offer these tests following the publica- Much progress has been made in the molecular characterization tion of the mutation, depending on patent and licensure restrictions. of hereditary diseases in dogs and cats since the initial identifica- The extent of information that is provided to the public varies from tion of the genetic basis for canine hemophilia B in 1989 (Evans one testing laboratory to another, but usually comprehensive infor- et al., 1989), aided by the completion of the canine (Lindblah-Toh mation on either the disease or mutation is unavailable.

⇑ Corresponding author. Tel.: +1 215 8988830. E-mail address: [email protected] (U. Giger). 1 See: http://www.ncbi.nlm.nih.gov/.

1090-0233/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tvjl.2013.02.021 121 J. Slutsky et al. / The Veterinary Journal 197 (2013) 182–187 183

Table 1 Sources of genetic disease information.

OMIA University of Sydney http://omia.angis.org.au/home; http://www.ncbi.nlm.nih.gov/omia/?term=omia CIDD University of Prince Edward Island www.upei.ca/cidd LIDA University of Sydney www.sydney.edu.au/vetscience/lida IDID Cambridge University http://server.vet.cam.ac.uk OFA Orthopedic Foundation of America http://www.offa.org CHF Canine Health Foundation http://www.akcchf.org Fabcats Feline Advisory Bureau http://www.fabcats.org/breeders/inherited_disorders

It is often daunting for veterinary clinicians, breeders and Table 2 researchers to keep up with rapid advances in diagnostic opportuni- Information available in the Canine and Feline Hereditary Disease (DNA) Testing ties. Despite a number of sources of genetic disease information cur- Laboratories web application. rently available on-line (Table 1), a few books (Bell et al., 2012), book Disease information Genetic information Laboratory information chapters, review articles and websites that have attempted to Disease name Chromosome Laboratory name gather information on genetic disease testing laboratories, the Related terms/synonyms Gene Website URL (hyperlink) number of disease-associated mutations, tests offered and laborato- Commonly used code Mutation description E-mail contact ries involved continue to grow and change, rapidly rendering many OMIA/OMIM number Research citation Mailing address Breeds affected Research hyperlink Country of these sources obsolete (Nicholas et al., 2011; Mellersh, 2012). Clinical disease To provide a comprehensive resource to find up-to-date, veri- description fied information on the currently available DNA tests for inherited diseases in dogs and cats, the Hereditary Disease Committee of the World Small Animal Veterinary Association (WSAVA) has devel- oped a web application featuring an interface that allows users to search the underlying database, which we describe below. by NCBI following consultation. Descriptions on each hereditary disease are contin- uously being collected from the Veterinary Information Network (VIN) Associate ebook for Hereditary Diseases.7 Materials and methods For the purposes of the data contained in this application, we defined a single heritable disease as an illness characterized by typical signs and routine laboratory The Canine and Feline Hereditary Disease (DNA) Testing Laboratories2 web tests and/or imaging abnormalities that occur due to a mutation in a particular application was developed using Microsoft ASP.net and a Microsoft SQL server data- gene. Therefore, if two breeds present with similar disease phenotypes, but differ base. The pages and database for this application are hosted on servers at the School in the gene mutated, the resulting disorders would be classified as separate dis- of Veterinary Medicine of the University of Pennsylvania (PennGen). eases. However, in the case where there are distinct mutations in the same gene We screened the scientific literature for the molecular characterization of in different breeds, causing the same illness, all these mutations would be listed hereditary diseases and genetic predispositions to disease in dogs and cats using as the same disease. PubMed3 and Commonwealth Agricultural Bureau (CAB) Abstracts.4 We also searched Only dog breeds recognized by the AKC, FCI and KC were included in the data- the Internet for laboratories that offer DNA testing for genetic diseases in dogs and base and we have not included information on mixed breeds unless they uniquely cats. We further checked the availability of DNA tests with dog and cat fancier asso- express a specific mutation not seen in any purebreds. Any disease seen in a pure- ciations, e.g. American Kennel Club (AKC), The Kennel Club (KC) UK, Fédération Cyno- bred dog or cat can, of course, occur in a mixed breed animal. For cats, we have in- logique Internationale (FCI), Cat Fancier Association (CFA) and The International Cat cluded domestic shorthair and domestic longhair cats as their own ‘breeds’, along Association (TICA), and organizations involved with genetic health issues in dogs or with the standard pure breeds, as stated by CFA and TICA. Since our data focuses cats, e.g. Canine Health Foundation (CHF), Orthopedic Foundation of Animals (OFA) on disease-specific mutations, tests for parentage and coat color, length and texture and Winn Feline Foundation. Each laboratory was contacted directly and asked for are excluded, unless directly associated with a disease. Finally, inclusion of affected specific information on each test, including which mutation(s) the laboratory tests breeds was limited to those backed by specific research, although on certain occa- for, which species and breeds are affected by each mutation tested for, if testing is still sions we have allowed a broader interpretation, where the mutation has been found available for each DNA test and if additional DNA tests are offered. through testing, but not confirmed in a published original study. No DNA mutation In addition to reviewing the published studies and research abstracts in which screen panels are included in the data. mutations were first described, we also verified unpublished information with re- search laboratories to identify additional disease-causing mutations and/or breeds affected by the same or different mutations in the same gene for which tests are Results now offered. The veracity of all unpublished information has not been verified by the authors, but generally the information is from established laboratories. Genetic information regarding the diseases listed, including gene affected, chromosome and The verified information on available DNA tests for hereditary mutation description, was obtained mainly through original research papers and diseases and genetic predispositions to diseases in dogs and cats published research on NCBI and PubMed. Mutations were described using the stan- is displayed on a website.8 We summarize here the information dard nomenclature as described by the Human Genome Variation Society.5 In addi- tion, genome and other databases in NCBI and the Genome Annotation Resource contained in the database to mid-2012 (Tables 2–6). It was discov- Fields – Felis catus (GARField) in the National Cancer Institute’s Laboratory of Genomic ered that four laboratories stopped offering DNA tests during the col- Diversity (Pontius and O’Brien, 2007)6 were used to describe the chromosomal loci of lection period and are therefore not included in the data. Forty-four the genes in dogs and cats, respectively. laboratories offered DNA tests for hereditary diseases in dogs and In some cases, the mutation in the database may be listed slightly differently to cats, 43 of which were included in the database and whose data that in the published literature due to new information on gene structure, release of updated genome assemblies, use of non-standard nomenclature and occasional er- we report on below; one corporate laboratory requested to be ex- rors in mutation descriptions. Online Mendelian Inheritance in Animals (OMIA) and cluded from the database. The name, address and website for each Online Mendelian Inheritance in Man (OMIM) numbers were collected from their laboratory, as well as details of each DNA test are provided. websites or based on information provided by laboratory responses. During the Twenty-two of the 43 testing sites are the laboratories and/or the analysis, it became evident that the NCBI used a different numbering system than OMIA for trait IDs, which caused confusion; fortunately, this has been corrected investigators that originally identified the mutation. These usually only test for a single mutation or a small group of (related) genetic diseases; 14 laboratories only test for a single disease and nine of 2 See: http://research.vet.upenn.edu/WSAVA-LabSearch. these only test samples from a single breed bearing the mutation. 3 See: www.ncbi.nlm.nih.gov/pubmed. 4 See: www.cabi.org. 7 See: http://www.vin.com/Members/Associate/Associate.plx?Book=1&Browse 5 See: http://www.hgvs.org/mutnomen. Chapter=&SpeciesID=5#Jump. 6 See: http://lgd.abcc.ncifcrf.gov. 8 See: http://research.vet.upenn.edu/WSAVA-LabSearch. 122 184 J. Slutsky et al. / The Veterinary Journal 197 (2013) 182–187

Table 3 Table 6 Information available from ‘View Disease Details’ link. Inheritance patterns of diseases with known mutations.

Disease name/synonyms Dogs Cats Total General description Autosomal recessive 107 19 126 Description in species Autosomal dominant 13 4 17 Mode of inheritance X-linked recessive 1 2 3 Etiology X-linked dominant 8 0 8 Breed, sex and age predilection Mitochondrial 1 0 1 Clinical findings and signs Diagnostic procedures Treatment and management Prevention focus of the laboratories and/or through a lack of demand to test Differential diagnosis for mutations that occur very rarely in a particular breed population Human disease homologue (Table 5). Available tests A total of 155 hereditary diseases (130 in dogs, 25 in cats) have Research references Contributor’s name and date been characterized at the molecular level and 125 currently can be assessed in laboratories (111 in dogs, 20 in cats). Although 94 dis- orders can be tested for by several laboratories (85 in dogs, 9 in Of 43 laboratories that offered DNA testing, 21 were commercial lab- cats), the rest are offered only by a single laboratory (Table 4), oratories that specialize in genetic disease testing. Twenty-eight lab- either due to patent and license restrictions, lack of published oratories offered DNA tests for dogs only, five for cats only and 10 for information and/or because the mutation is believed to occur very dogs and cats. No laboratory offers all available tests, due to rarely in a particular breed population. More than one mutation restrictions by patents, limited licensure, through a specific disease has been reported in the same gene for several disorders

Table 4 Summary of disease information in the database.

Dog Cat Total Number of disease tests 111 20 125a Diseases with a single mutation 87 15 102 Diseases with multiple mutations 24 5 29 Total mutations tested for 143 24 167 Single breed mutations 100 15 115 Mutations affecting multiple breeds 43 9 52 Total breed specific tests tested forb 361c 56d 417 Commercial breed specific tests 306 41 347 Non-profit breed specific test 176 35 211 Breed tests available at only one laboratory 123 13 136 Breed tests available at multiple laboratories 238 43 281 Maximum number of laboratories performing a test 10e 10f Maximum number of mutations in a single disease 6g 2h Maximum number of breeds tested for a single mutation 22i 16j Average number of laboratories testing a single breed specific mutation 3.6 3.0 Median number of laboratories 3 1 Average number of mutations for a specific disease 1.3 1.4 Median number of mutations 11 Average number of breeds for a specific mutation 2.3 2.9 Median number of breeds 11

a Includes six diseases where the mutation has been found in both species and a test is available in both species. b Total of the tests for each specific mutation available in a specific breed (i.e. a specific disease/mutation/breed combination). c There are 121 breed specific tests for dogs available at both commercial and non-profit laboratories. d There are 20 breed specific tests for cats available at both commercial and non-profit laboratories. e Multiple instances. f Blood type B mutation. g Factor IX deficiency (hemophilia B). h Multiple instances. i Primary lens luxation. j Progressive retinal atrophy (Rdac mutation), although Blood type B is offered for all breeds.

Table 5 Summary of laboratory information in the database.

Non-profit Corporate Total Number of laboratories 22 21 43 Average number of diseases tested by one laboratory 5.0 20.0 12.4 Median number of diseases tested by one laboratory 2 15 4.0 Maximum number of diseases tested by one laboratory 27 67 Minimum number of diseases tested by one laboratory 1 1 Average number of breed mutation tests by one laboratory 13.5 57.2 34.8 Median number of breed mutation tests by one laboratory 4.5 47 Maximum number of breed mutation tests by one laboratory 60 195 Minimum number of breed mutation tests by one laboratory 1 1

123 J. Slutsky et al. / The Veterinary Journal 197 (2013) 182–187 185

(24 disorders in dogs, five in cats); frequently, individual mutations Many mutations were found only in a single breed (69% of the are breed specific. The pattern of inheritance of the majority of mutations listed in the database), whereas some mutations have diseases in dogs and cats with known mutations is autosomal been found in multiple breeds, up to 22 for primary lens luxation. recessive; mutations that are inherited as autosomal dominant, Some disorders have only been identified in a single animal or fam- X-linked recessive, X-linked dominant or mitochondrial traits have ily and may not be present in the general breed population, e.g. X- also been identified (Table 6). Tests for several complex traits with linked severe combined immunodeficiency in dogs maintained in a multiple gene defects need to be investigated further. research colony (Henthorn et al., 1994); routine testing for such

Fig. 1. A sample disease test search for a coagulopathy in Beagles. (A) Searches can be done by disease/test, breed or laboratory. (B) Information regarding the selection is used to narrow down the results. (C) Information about the specific disease in this breed is displayed. (D) Information about the laboratories doing the specific test in this breed is displayed. 124 186 J. Slutsky et al. / The Veterinary Journal 197 (2013) 182–187 specific mutations usually is not offered. There are also cases example, we have chosen to search for factor VII deficiency, a com- where there are separate mutations affecting the same breed, caus- mon coagulopathy (Callan et al., 2006)(Fig. 1B). The application ing different forms of the disease, e.g. porphyria in domestic short- displays the pertinent genetic information regarding the heredi- hair cats (Clavero et al., 2010). tary disease (Fig. 1C), as well as the laboratories that offer the test (Fig. 1D). If further clinical details on the disease are desired, they may be accessed via the hyperlink through the ‘View Disease De- Discussion tails’ option to download a PDF file (Fig. 1C; Table 3). In the example shown in Fig. 1, three testing laboratories are In the past two decades, much progress has been made in the identified. The first laboratory listed will be the laboratory that characterization of disease-causing mutations in dogs and cats. originally identified the particular breed-specific disease mutation, Through DNA testing, this new information permits specific diag- if they are still testing for the mutation, or a laboratory that is di- nosis in an animal affected by a specific hereditary disease or rectly affiliated with the research group. The research article first allows an animal at risk of becoming ill because of a particular dis- describing the mutation may be accessed (Fig. 1C) through the tex- ease-causing mutation to be identified. Most genetic diseases are tual citation or through a hyperlink (in this case freely accessible inherited recessively and may occur commonly in one or more by the hyperlink to PubMed Central). This disease example also re- breeds due to particular breeding practices, such as deliberate veals that two other breeds have Factor VII deficiency caused by inbreeding or the extensive use of a popular sire (Wade, 2011). the same mutation (Alaskan Klee Kai and Scottish deerhound). Therefore, knowledge of the mutation allows screening of the While this coagulopathy has also been described in Great Pyrenees breeding stock and, by permitting selection of appropriate breed- and English springer spaniels, the disease-causing mutation(s) in ing animals, can eliminate the disease from future generations. these breeds have not yet been identified. Since the DNA test DNA tests are the most desirable tools for the detection of may not be helpful for these and other breeds, currently they are mutations causing hereditary diseases; they allow determination not contained in the database under this mutation test. of homozygosity and heterozygosity for a certain mutant/disease allele, only require small samples (such as blood or cheek swabs, Conclusions which can be shipped by regular mail), are relatively simple to per- form in the laboratory, are standardized and are potentially less This web-based application represents a source of up-to-date expensive than most other tests. There are many different tech- information on hereditary diseases in companion animals for vet- niques, from manual to robotically automated, for identification erinary clinicians looking for a laboratory to perform a test, of the normal and mutant allele for a disease. This web application researchers searching for information on hereditary diseases and does not provide information on these detailed laboratory tech- owners/breeders with affected animals or animals at risk of devel- niques, which often change with new technologies. Moreover, cur- oping a particular disease or passing on the mutant allele (carriers). rently there is no official quality control system for DNA testing in We intend to keep this web application updated by regular review veterinary medicine and the application presented here cannot as- of the pertinent literature, correspondence with testing laborato- sess the quality of testing of any laboratory listed. ries and through feedback from those involved in research on com- Although biochemical laboratory tests and imaging studies are parative medical genetics. This service will be continued by the used to diagnose some hereditary diseases in companion animals, WSAVA Hereditary Disease Committee. genomic DNA tests for single gene defects are considered to be the most accurate in clinical medicine and thus only DNA tests are in- Conflict of interest statement cluded here. Allowing for human errors from identifying animals, labeling and mixing up samples, these DNA tests are considered The authors from the University of Pennsylvania are associated to be accurate, assuming that regular laboratory standards, with with PennGen, one of the not-for-profit laboratories offering DNA appropriate positive and negative controls, are followed. tests, and the work was funded by the WSAVA through contribu- Current information on mutant allele frequency is limited, since tions from Waltham. the data generally are based upon a few rather small and fre- quently biased, rather than randomized, surveys or open registries. Also, common mutations may disappear from a population (breed) Acknowledgements due to the success of a DNA screening program. Recently, one company involved in canine disease testing has This study was supported in part by the WSAVA, Waltham and offered a multiple single nucleotide polymorphism (SNP) panel the USA National Institutes of Health grant NIH RR002152 and NIH analysis that screens for disease-causing mutations in mixed breed OD010939. The authors would like to acknowledge the assistance dogs (Mars Veterinary). This company was not included on the of VIN and especially Dr Linda Shell in the development of the dis- website, since panel analysis screens are not considered to be a ease information files in the Associate program, as well as many specific breed test. The results of the panel are not reported as a veterinary clinicians and scientists who provided valuable specific definitive diagnosis in affected animals, but alert the submitter if disease information. a mutation is found, so that further specific testing can be pursued at a DNA genetic disease testing laboratory. Unless patents and References licensures restrict its future use, such panel analyses may be used for all known DNA mutations in a species, making this method a Asher, L., 2009. Inherited defects in pedigree dogs. Part 1. Disorders related to breed standards. The Veterinary Journal 182, 402–411. simple and cost effective tool to screen for hereditary diseases in Bell, J.S., Cavanagh, K.E., Tilley, L.P., Smith, F.W.K., 2012. Veterinary Medical Guide to companion animals. Dog and Cat Breeds. Teton NewMedia, Jackson, Wyoming, USA, 705 pp. Our website is arranged by general categories: disease, breed, Callan, M.B., Aljamali, M.N., Margaritis, P., Griot-Wenk, M.E., Pollak, E.S., Werner, P., Giger, U., High, K.A., 2006. A novel missense mutation responsible for factor VII and laboratory, each of which can be searched separately deficiency in research Beagle colonies. Journal of Thrombosis and Haemostasis (Fig. 1A). After selecting an initial category to search, the users 4, 2616–2622. may select the specific disease, species (canine/feline) and breed Clavero, S., Bishop, D.F., Haskins, M.E., Giger, U., Kauppinen, R., Desnick, R.J., 2010. Feline acute intermittent porphyria: A phenocopy masquerading as an they are interested in. If there is more than one mutation known erythropoietic porphyria due to dominant and recessive hydroxymethylbilane to cause a disease, the specific mutation can be selected. As an synthase mutations. Human Molecular Genetics 19, 584–596. 125 J. Slutsky et al. / The Veterinary Journal 197 (2013) 182–187 187

Evans, J.P., Brinkhous, K.M., Brayer, G.D., Reisner, H.M., High, K.A., 1989. Canine Marschall, Y., Distl, O., 2010. Current developments in canine genetics. Berliner und hemophilia B resulting from a point mutation with unusual consequences. Münchener Tierärztliche Wochenschrift 123, 325–338. Proceedings of the National Academy of Sciences of the United States of Mellersh, C., 2011. DNA testing man’s best friend. The Veterinary Journal 168, 10– America 86, 10005–10099. 12. Giger, U., Haskins, M.E., 2006. Feline hereditary diseases. In: Horzinek, M.C., Mellersh, C., 2012. DNA testing and domestic dogs. Mammalian Genome 23, 109– Schmidt, V., Lutz, H. (Eds.), Krankheiten der Katze, Fourth Ed. Enke Verlag, 123. Stuttgart, Germany, pp. 589–602. Nicholas, F.W., Crook, A., Sargan, D., 2011. Internet resources cataloguing inherited Giger, U., Sargan, D.R., McNiel, E.A., 2006. Breed-specific hereditary diseases and disorders in dogs. The Veterinary Journal 189, 132–135. genetic screening. In: Ostrander, E., Giger, U., Lindblad-Toh, K. (Eds.), The Dog and Padgett, G.A., 1998. Control of Canine Genetic Diseases. Howell Book House, New its Genome. Cold Spring Harbor Laboratory Press, New York, USA, pp. 249–289. York, NY, USA. Hedhammar, Å.A., Indrebø, A., 2011. Rules, regulations, strategies and activities Pontius, J.U., Mullikin, J.C., Smith, D.R., Agencourt Sequencing Team, Lindblad-Toh, within the Fédération Cynologique Internationale (FCI) to promote canine K., Gnerre, S., Clamp, M., Chang, J., Stephens, R., Neelam, B., et al., 2007. Initial genetic health. The Veterinary Journal 189, 141–146. sequence and comparative analysis of the cat genome. Genome Research 17, Henthorn, P.S., Somberg, R.L., Fimiani, V.M., Puck, J.M., Patterson, D.F., Felsburg, P.J., 1675–1689. 1994. IL-2Rc gene microdeletion demonstrates that canine X-linked severe Pontius, J.U., O’Brien, S.J., 2007. Genome Annotation Resource Fields – GARFIELD: A combined immunodeficiency is a homologue of the human disease. Genomics genome browser for Felis catus. Journal of Heredity 98, 386–389. 23, 69–74. Sargan, D., 2003. IDID: Inherited Diseases in Dogs: Web-based information for Lindblah-Toh, K., Wade, C.M., Mikkelsen, T.S., Karlsson, E.K., Jaffe, D.B., Kamal, M., canine inherited disease genetics. Mammalian Genome 15, 503–506. Clamp, M., Chang, J.L., Kulbokas 3rd, E.J., Zody, M.C., et al., 2005. Genome Vella, C.M., Shelton, L.M., McGonagle, J.J., Stanglein, T.W., 1999. Robinson’s Genetics sequence, comparative analysis and haplotype structure of the domestic dog. for Cat Breeders and Veterinarians, Fourth Ed. Butterworth Heinemann, Oxford, Nature 438, 813–819. UK, 272 pp. Lyons, L.A., 2010. Feline genetics: Clinical applications and genetic testing. Topics in Wade, C.M., 2011. Inbreeding and genetic diversity in dogs: Results from DNA Companion Animal Medicine 25, 203–212. analysis. The Veterinary Journal 189, 183–188. Lyons, L.A., 2012. Genetic testing in domestic cats. Molecular Cellular Probes 26, 224–230.

126 Proc. R. Soc. B (2011) 278, 3161–3170 doi:10.1098/rspb.2011.1376 Published online 1 September 2011

Review Deciphering the genetic basis of animal domestication Pamela Wiener* and Samantha Wilkinson The Roslin Institute and Royal (Dick) School of Veterinary Studies, University of Edinburgh, Easter Bush, Midlothian EH25 9RG, UK Genomic technologies for livestock and companion animal species have revolutionized the study of animal domestication, allowing an increasingly detailed description of the genetic changes accompanying domestication and breed development. This review describes important recent results derived from the application of population and quantitative genetic approaches to the study of genetic changes in the major domesticated species. These include findings of regions of the genome that show between-breed differentiation, evidence of selective sweeps within individual genomes and signatures of demographic events. Particular attention is focused on the study of the genetics of behavioural traits and the impli- cations for domestication. Despite the operation of severe bottlenecks, high levels of inbreeding and intensive selection during the history of domestication, most domestic animal species are genetically diverse. Possible explanations for this phenomenon are discussed. The major insights from the surveyed studies are highlighted and directions for future study are suggested. Keywords: animal domestication; breed differentiation; selective sweep; population bottleneck

1. INTRODUCTION under selection are presented, as well as methods for elu- Understanding the history of domestication has been of cidating non-selective processes; many of these, if not all, interest to biologists at least since Darwin. He appreciated were first developed for human genetic analysis. This the wide variation within domesticated species, and article does not attempt to review all the relevant litera- throughout On the origin of species [1] (and later, in his two ture, but rather to use specific examples to illustrate volumes of Variation under domestication [2]) he used them common themes. The examples presented are primarily as examples of his theories. It is now well accepted that the taken from cattle, pigs, and dogs, where the process of animal domestication has involved a combination most advanced genetic resources are available. of human-imposed selection and non-selective forces, the latter including various forms of interference with the demography and mating programme of these species. 2. SELECTIVE FORCES It is only recently, with advances in genetic and statistical The assumption underlying the detection of signatures of technologies, that the genetic changes that have accompanied selection in the genome is that selection is locus-specific. animal domestication and breed development can be charac- By comparison, the effects of other evolutionary forces terized. A rapidly increasing number of species now have (random genetic drift, mutation and inbreeding) should full-genome sequences. High-density, genome-wide single be expressed genome-wide. Under this premise, the nucleotide polymorphism (SNP) panels have been produced methods for detecting selected loci attempt to identify for humans, as well as many other and animal species. those at which allele frequencies have changed in a pat- A variety of statistical techniques have concurrently been tern consistent with positive selection. The methods developed to analyse this data. One of the key aspects of differ in the information they use to find such loci (par- this analysis is to use genomic data in order to make infer- ticularly as there are very few data on historical allele ences about the selective and demographic forces that have frequencies), which will be outlined further. operated on individual species. This review discusses the contribution that genetic (a) Candidate gene studies data have made to our understanding of both selective One approach adopted in domestication genetics is to and non-selective processes of evolutionary change in examine patterns of diversity around candidate genes domesticated animal species, and the insights into the that, based on their function, are likely to have been targets domestication process that have been revealed by these of selection. Two such genes are growth differentiation studies. Applications of various population genetic- factor 8 (GDF-8), associated with muscle conformation, based methods for the detection of genomic regions and melanocortion 1 receptor (MC1R), associated with coat colour. GDF-8 (myostatin) is a negative regulator of skeletal * Author for correspondence ([email protected]). muscle growth, and naturally occurring mutations in this Electronic supplementary material is available at http://dx.doi.org/ gene have been associated with increased levels of muscle 10.1098/rspb.2011.1376 or via http://rspb.royalsocietypublishing.org. conformation in cattle, dogs, sheep and humans. There is

Received 5 July 2011 Accepted 12 August 2011 1273161 This journal is q 2011 The Royal Society 3162 P. Wiener & S. Wilkinson Review. Genetics of animal domestication

(a) (b)

1.0

0.8

0.6

0.4 heterozygosity 0.2

0 0123456 distance from GDF-8 (cM)

Figure 1. Selection for muscle conformation in cattle. (a). Belgian Blue cows from early (top) and late (bottom) in the 20th century (photos reprinted from Compere et al.[5]; stitches in the more recent photo indicate that the calf was delivered by caesarean section, which is common in Belgian Blue cattle and associated with the large size of double-muscled calves [5]). (b). Relationship between heterozygosity and genomic distance from the GDF-8 (myostatin) gene for Belgian Blue and homozygous for the 11 bp deletion (MH/MH) associated with double muscling (data from Wiener et al.[6] and Wiener & Gutierrez-Gil [7]). Blue circles, Belgian Blue; red circles, MH/MH South Devons. substantial diversity in GDF-8 across cattle breeds [3,4], phaeomelanin (yellow/red) pigments, and appears to have including, at the extreme, two independent loss-of- been a target for selection in pigs and other domesticated function mutations that are associated with the ‘double animal species. For example, there has been independent muscling’ phenotype, in which animals have highly exag- evolution of black coat colour in Asian and European gerated muscle conformation. At the beginning of the pigs, and selection for this phenotype appears to have twentieth century, the majority of Belgian Blue cattle had been particularly strong in Chinese pigs, where animals conventional conformation, and were used for both milk with the black coat were used preferentially in animal sacri- and production [5](figure 1a). However, after less fice rituals during the Neolithic period, because they were than a century of animal breeding, the double muscling considered sacred [13]. Asian wild boar (the closest relative phenotype is now nearly fixed in the breed, suggesting to the domestic pig) show extensive nucleotide variation at that there has been strong selection in favour of this trait, MC1R; however, nearly all European and Asian wild boar presumably owing to the increased amount of derived genotyped so far express the same MC1R protein, with meat [8,9]. Analysis of microsatellite diversity in the genotypes differing primarily by synonymous substitutions region flanking the GDF-8 gene revealed a significant [13,14]. This wild-type protein allows complete expression decrease in heterozygosity with increasing proximity to of both eumelanin and phaeomelanin pigments, and pro- GDF-8 in three primarily double-muscled breeds, includ- duces a coat in variable shades of brown [15]. In contrast, ing the Belgian Blue, as well as in the sub-group of in domestic pigs, there is reduced synonymous variation double-muscled South Devon cattle [6,7](figure 1b), relative to wild boar but at least nine different MC1R which was not seen in most non-double-muscled breeds. proteins in addition to the wild-type [14], which are associ- The pattern of heterozygosity in both Belgian Blue and ated with coat colour phenotypes ranging from red to black South Devon cattle is consistent with strong selection on and including a variety of spotting patterns (white coat this gene. While evidence of a signature of selection near and spotting are determined by a different gene, KIT) myostatin has not yet been published for other species, it (figure 2). Therefore, it appears that wild boar have been is likely to exist as variation in this gene has been shown subject to purifying selection for a camouflaged coat, to influence traits of economic interest in breeds of dogs whereas a relaxation of this form of natural selection in [10] and sheep [11]. combination with human-mediated selection for distinctive Coat colour and pattern are key traits in the develop- coat patterns has occurred in domestic pigs [14]. ment of livestock and companion animal breeds, as they were under selection well before breed development [12]. (b) Differentiation-based approaches A number of genes have been associated with coat colour Changes within breeds have occurred on an evolutio- in mammals, including the MC1R gene. MC1R influen- narily short timescale compared with natural animal ces the relative levels of eumelanin (black/brown) and populations; however, there is considerable phenotypic

Proc. R. Soc. B (2011) 128 Review. Genetics of animal domestication P. Wiener & S. Wilkinson 3163

Figure 2. Coat colour variation in pig breeds. Clockwise from top-left: Berkshire, British Saddleback, Gloucestershire Old Spots, Large Black, and Tamworth (photos: S. Wilkinson). variation between domesticated animal breeds, particu- breeds were found in genomic regions associated with larly in dogs. Recent studies in various species have morphological traits, including body size, skull and snout applied an approach where markers with strong evidence shape, coat characteristics and ear type [25]. of genetic differentiation (e.g. high levels of Wright’s FST, For cattle, the genetic-differentiation approach has high- a measure of genetic differentiation between populations, lighted genomic regions that include genes encoding coat or allele-frequency differences) are taken as signals of features or body size/conformation. Several studies have differential selection across populations. This approach identified high levels of between-breed genetic differen- originated in the days when genetic markers were limited tiation near coat colour loci, including MC1R (see §2a) and sparse, and the focus was on specific markers [16,17], and the Charolais dilution factor (Dc locus), indicating but in the current environment of dense, genome-wide that these genes have been important in the establish- markers for many species, genome scans of differentiation ment of cattle breeds [26,27]. Another gene that has have become a viable strategy to identify selected genes or been implicated as a possible target of selection based on genomic regions using the tails of the genome-wide FST allele-frequency differences between cattle breeds is the distribution to define the significance threshold [18]. growth hormone receptor (GHR) gene [26–28]. For this and other approaches, it has been recognized Although it is clear that large qualitative effects have that instead of using single-locus statistical values, a slid- been detected using these methods, there are known to ing window analysis removes the stochastic variation be limitations to FST-based methods for detecting genes between loci, and thus better highlights regions with with small or moderate effects. Wiener et al.[27] found signals of selection [19,20]. Although the population- the overall correlation between FST and the statistical differentiation approach was developed originally for signal from linkage mapping analysis (see §2e)tobelow analysis of human data (and is still used in this context in a study of two cattle breeds. While genes associated [18,21]), this technique is possibly even better suited to with coat colour could be detected as regions of large studies of domesticated animals because breeds are in allele-frequency differences, the signals for loci associated general genetically similar entities and the differences with quantitative traits were generally weaker. that do exist may reflect the relatively recent selection for breed-specific characteristics. Akey et al.[22] conducted an FST scan of the genome for (c) Frequency spectrum-based approaches 10 dog breeds and identified outliers, which they argued A common approach to test for selection in human and were candidates for targets of selection. This interpretation wild plant and animal populations is to use ‘frequency spec- of the results was supported by the fact that five genes that trum’ tests in which empirical allele distributions are had previously been mapped through association with compared with those predicted under a neutral model. ‘hallmark’ breed traits were among the 155 outlier SNPs One set of methods involves searching the genome for (including the insulin-like growth factor 1 gene—IGF1, regions with allele-frequency patterns that differ either associated with body size—and several coat colour from background (genome-wide) patterns or from those genes). Regarding the other outlier SNPs identified in predicted by a neutral model [29,30]. These methods their study, one of the highest FST values was only found involve calculation of a composite log likelihood (CLL) in the Shar-Pei breed, which is characterized by its distinc- for sliding window sets of genotypic data and testing signifi- tive skin-folding phenotype. The region where the high FST cance based on a likelihood ratio test [29,30]orby signal was found contains several genes, including HAS2, permutation testing [31]. This approach has recently the expression of which had previously been associated been applied to genome-wide SNP data for the 19 cattle with skin wrinkling in this breed [23]. A recently discovered breeds characterized by the Bovine HapMap Consortium duplication upstream of this gene appears to be responsible [32]. In a follow-up analysis of this dataset, Stella et al. for the wrinkling phenotype [24]. A separate study looking [31] calculated the difference for each SNP between the at genetic differentiation between 79 domestic dog breeds major allele frequency for a group of breeds defined by phe- found that the top 11 FST values measured across all notype and the overall frequency across all breeds. For

Proc. R. Soc. B (2011) 129 3164 P. Wiener & S. Wilkinson Review. Genetics of animal domestication black-coated breeds, there was a very strong signature of The extended haplotype-based methods have been selection on BTA18 for windows that include the MC1R applied mainly to human genetic data, but they have also coat colour locus (see §2a). A signature of selection was been implemented for several cattle datasets. Studies by also observed for polled (hornless) breeds on BTA1 Hayes et al.[28,39] found high values of iHS for SNPs in within a region previously associated with presence/absence several regions of bovine chromosome 6, including one of horns. For dairy breeds, 699 putative signatures of selec- region with the ABCG2 gene, associated with several tion were identified across the genome, with the highest dairy traits. The Bovine HapMap Consortium [32] also (negative) CLL value on BTA6 near the KIT gene, which applied the iHS test across the genomes of 19 breeds and is associated with the level of white coat spotting in cattle. found high iHS values in one or more breeds on most To make sense of the large number of significant results, chromosomes; these included regions on BTA2 near the authors looked for cases where genes from the same GDF-8, on BTA6 near ABCG2, and on BTA14 near a gene family were at the centre of the significant window region associated with intramuscular fat. There were (e.g. potassium channel genes, integrins and arginine-/ many other regions where a specific gene could not be serine-rich splicing factors), arguing that these gene families implicated as a selection target. More recently, Qanbari may have been under selection during dairy cattle breeding. et al.[40] applied the LRH test to denser (50 K SNP) Difficulties in applying frequency-spectrum-based tests data from Holstein dairy cattle. Although there were signifi- to SNP data have been raised because of the bias towards cant or nearly significant signals of selection for SNPs high-frequency alleles inherent in SNP ascertainment, associated with some dairy-related candidate genes (e.g. and thus interpretation of results can be problematic. the casein gene cluster encoding milk proteins and the While a number of solutions have been proposed to deal DGAT1 gene associated with milk fat percentage), of the with this issue [33], in the long term the best remedy will SNPs with greatest significance levels, none were found involve use of full-genome sequence data in place of SNP near these candidates. data. Developments in next-generation sequencing are The advent of whole-genome sequencing opens up new now making this a reality for many species (see §2d). possibilities for the detection of selection signatures. Rubin et al.[41] sequenced whole genomes of eight pools of chickens representing commercial lines, experimental (d) Extended homozygosity approaches lines and breeds selected for specific traits. The genome Another population-genetic approach for the detection of was searched for regions of low diversity by calculating a selective sweeps has been to look for extended homozygous normalized pooled heterozygosity measure in sliding win- genomic regions. This approach is based on ‘hitch- dows. One of the lowest statistics (suggestive of positive hiking’ theory [34], in which neutral variants increase in selection) was found in the region of the beta-carotene frequency owing to linkage disequilibrium (LD, the statisti- dioxygenase 2 (BCDO2) gene, which is associated with cal association between allele frequencies at different loci) skin colour in chickens. One or more regulatory mutations with alleles at a selected locus, resulting in reduced diversity that inhibit expression of the BCDO2 gene appear to be across the region. responsible for the yellow skin phenotype [42]. Most chick- One particularly convincing example of reduced diver- ens used for commercial egg and meat production in sity near a selected locus relates to chondrodysplasia industrialized countries (as well as many local breeds (shortened limbs) in dogs. A genome-wide SNP analysis worldwide) have the yellow skin phenotype and are homo- revealed a 24 kb region of reduced heterozygosity on zygous for the recessive yellow skin allele locus, whereas chromosome 18 in chondrodysplastic breeds (e.g. Dachs- other local chicken breeds have white skin and carry the hunds) relative to non-chondrodysplastic breeds [35]. dominant wild-type allele. The yellow skin allele appears This region includes an insertion of a retrogene encoding to have been derived from a different ancestral species fibroblast growth factor 4 (FGF4) in the chondrodysplas- (possibly the grey junglefowl) than most of the commercial tic dogs, the expression of which may result in altered chicken genome (for which the red junglefowl is the activation of one or more fibroblast growth factor recep- presumed wild ancestor), suggesting a hybrid origin of tors. A similar pattern of reduced heterozygosity near commercial chickens (see §3b)[42]. the IGF1 gene was observed in small dogs [36]. The region with the lowest heterozygosity score across A number of statistical methods aim to distinguish the all domestic lines included the locus-encoding thyroid length of homozygous segments generated by selection stimulating hormone receptor (TSHR) gene [41], which from those generated by neutral processes, which extends is involved in metabolic regulation and reproduction. the analysis beyond the heterozygosity of individual mar- This region was almost completely fixed over a 40 kb seg- kers. One of the first methods introduced to exploit the ment. Further analysis of this locus in domestic chickens hitch-hiking phenomenon in the context of high-density from a number of countries revealed that each animal car- genotype data was the long-range haplotype (LRH) test ried at least one copy of the derived haplotype (264/271 [37]. In this method, the age of each core haplotype in were homozygous). The role of TSHR in the domestication a genomic region is assessed using the length of extended of chickens is still unknown; however, the authors suggest haplotype homozygosity (EHH). Unusually, high EHH that it may be involved in the loss of seasonal reproduction values suggest a mutation that increased more quickly present in non-domesticated relatives. than expected under a neutral model. In an alterna- tive approach, the logarithm of the ratio of EHH for an ancestral allele to that for a derived allele (iHS) is (e) Genotype–phenotype association analyses used as the test statistic [38], such that large negative A powerful approach for gene mapping in livestock species (positive) values of iHS indicate selection for the derived is linkage mapping, in which regions of the genome associ- (ancestral) allele. ated with particular traits (quantitative trait loci, QTL) are

Proc. R. Soc. B (2011) 130 Review. Genetics of animal domestication P. Wiener & S. Wilkinson 3165 identified. Populations generated by breed or line crosses the original fox population showed continuous variation have proved to be particularly useful for identifying the for tameness/aggressiveness. A breeding programme was regions of the genome that distinguish the population foun- established with 100 females and 30 males, from which ders. Although this technique generally identifies fairly foxes were selected for their tameness using severe selec- wide intervals that include a large number of genes, in tion criteria [54]. The resulting population of tame foxes some cases it has led to the identification of individual behaved much like domestic dogs. Behavioural traits genes that influence physical traits related to domestication, other than tameness also evolved (e.g. tail wagging, lick- breed development or breed improvement (e.g. IGF2 in ing). Moreover, in addition to the changes in behaviour, pigs [43], DGAT1 [44] and GHR [45] in cattle). other morphological changes also occurred, some of QTL-encoding physical traits may also be associated which are reminiscent of dog breeds. For example, traits with behavioural traits. One such instance is the PMEL17 such as floppy ears, curly tails and shortened snouts gene encoding plumage colour in chickens, which was appeared in some foxes. Recent development of a linkage identified from a cross between red junglefowl and the com- map for the silver fox [55] has allowed QTL analysis of mercial White Leghorn. A 9 bp insertion in exon 10 acts in backcross and intercross populations derived from the a dominant fashion, such that birds homozygous for the tame population and an unselected (aggressive) popu- ancestral junglefowl allele (i) are black, whereas those car- lation. QTL for several tameness-related behavioural rying the White Leghorn allele (I) are white (heterozygotes traits map to fox chromosome 12; however, it is still sometimes have minor pigmentation). It has been demon- unclear whether these are associated with a single locus strated that there are substantial behavioural differences [56]. Furthermore, inconsistencies between results from between birds carrying the junglefowl and White Leghorn different crosses suggest a complex inheritance pattern alleles, such that i/i individuals birds are more vocal, have (e.g. strong epistatic interactions) for these traits. lower activity levels in a test measuring fear of humans, The study of silver foxes suggests that laboratory selec- and are more aggressive, social and explorative [46–48] tion for behavioural traits can emulate the process of than I/I birds, suggesting either that PMEL17 has pleio- domestication. Other researchers from Novosibirsk con- tropic effects on behaviour or the existence of a closely ducted an experiment selecting for reduced or enhanced linked behavioural locus [48]. This locus may also be aggression to humans in a population of wild-caught associated with feather-pecking, a bullying behaviour that rats [57]. Like the silver fox, this population has recently can result in severe damage to the victim [49]. Darker been exploited using genetic techniques to map regions of birds tend to suffer more from feather-pecking compared the genome associated with ‘tameness’ (as referred to with their white counterparts [46,50]. However, it remains above), defined by a linear combination of a set of behav- unresolved whether the effect on feather-pecking is due ioural traits [52]. QTL analysis indicates that more than solely to the plumage colour or whether the behaviour of one region is involved in the evolution of tameness in i/i birds makes them more likely to be targets of pecking. these rats [52] and that individual QTL may comprise The case of PMEL17 is particularly interesting in that it multiple sites [58]. demonstrates the possibility of selection for correlated traits Modification of behaviour is believed to have been one in domesticated animals. It is likely that the behavioural of the key aspects of animal domestication, including traits associated with PMEL17 were not the target traits selection for ‘reduced fear, increased sociability and in the development of the White Leghorn breed but may reduced anti-predator responses’ (p. 5 in [59]). As dog have been co-selected owing to selection for white plumage. breeders and owners know well, behaviour is also associ- Association between behavioural traits and coat colour ated with breed differences. In an investigation of four appears to be a common phenomenon. Genes in the mela- composite personality traits (playfulness, curiosity/fearless- nocortin system (including MC1R and the agouti gene) ness, sociability and aggressiveness) in 31 dog breeds, have been associated in mice and other vertebrate species Svartberg [60] found large differences between breeds for (e.g. lions, lizards and birds) with both coat colour and all traits. For example, popular pet breeds tended to have various behavioural traits, including aggressiveness, higher sociability and playfulness scores than less popular sexual behaviour and learning behaviour [51]. Eumela- breeds. nin-based coloration is generally associated with more aggressive behaviour. In her treatise on cattle breeds, Felius [12] claims that the Romans and later Europeans 3. NON-SELECTIVE FORCES also associated coat colour with cattle performance traits: While selection has clearly been an important force in the a red coat (the most common phenotype) was associated history of animal domestication, as with wild species, with a ‘fiery’ and hard-working character, whereas the other non-selective mechanisms have strongly influenced rare white coat was associated with a sluggish and lazy evolutionary change in these species. There are various disposition. However, the genetic association between approaches that allow inferences about demographic behavioural traits and coat colour is not universal, as and mating processes using genetic data. demonstrated by a study on rats in which ‘tameness’ QTL (see below) were on different chromosomes from a (a) Human-mediated modifications to QTL segregating for white coat spotting [52]. population size and structure In some cases, correlated selection appears to go in the One important advance with the advent of dense mar- other direction, such that selection for behavioural traits kers is the ability to exploit the relationship between may result in associated changes in more visible pheno- LD and effective population size (Ne, the number of indi- types, as has been seen in the well-described selection viduals in an idealized population that would have the experiment involving the silver fox (the ‘farm-fox exper- same rate of genetic drift as the actual population), such 2 iment’) [53]. Initiated in 1959 in Novosibirsk, Siberia, that Ne and r (the correlation between allele frequencies

Proc. R. Soc. B (2011) 131 3166 P. Wiener & S. Wilkinson Review. Genetics of animal domestication at two loci) are inversely related [61–63]. Hill [63] also There is clear evidence of declining Ne in commercial recognized that LD between tightly linked markers animal breeds, and in some cases this has resulted in extre- reflects older Ne than the LD between loosely linked mar- mely low variability. A feral British breed, Chillingham kers. Specifically, assuming linear population growth, cattle, was found to be homozygous at 24 out of 25 micro- LD between loci with recombination rate c reflects satellite loci [73], which is strikingly low when compared the Ne of 1/2c generations in the past [64]. With with other British cattle breeds [74]. The high levels of dense genotype data, this relationship can now be exploi- homozygosity in the Chillinghams presumably result ted to make inferences about population demographic from a very severe bottleneck and absence of immigration. history [64,65]. Looking over a longer timescale, ancient B. taurus DNA Using this approach, the Bovine HapMap Consortium has revealed a reduction in diversity at several cattle [32] found that LD declined rapidly with increasing phys- genes over the last 4000 years [75]. It is not yet clear ical distance between markers, but the rate of decline whether this is a genome-wide or loci-specific pattern. varied between cattle breeds. Overall LD levels for cattle were between those seen for humans (generally low) and (b) Introgression dogs. Ne appears to have declined recently for all breeds, Another human-related phenomenon that is manifested presumably owing to bottlenecks associated with domesti- in the architecture of genomes is that of introgression cation and breed formation. Comparing LD–distance between breeds. Animal breeders may practice cross- relationships across breeds can be used to understand the breeding to introduce certain desirable traits for breed different breed histories. Three Bos indicus (humped improvement. In the case of pig breeds, past human cattle, originating in the Indian subcontinent) breeds exam- activity has influenced the genetic composition of Euro- ined had lower LD than the Bos taurus (humpless cattle, pean breeds. In the 18th and 19th centuries, Asian originating in the Middle East) breeds at short distances alleles were introduced into certain British pig breeds to and intermediate values at long distances, indicating a rela- promote traits such as fattening and earlier maturation tively large ancestral population compared with the taurine [2]. Breeds that experienced genetic introgression breeds [32]. This characterization of B. indicus breeds is included Berkshire and Middle White, and Asian mor- consistent with findings of higher nucleotide diversity in phological characteristics such as the squashed face of B. indicus than in B. taurus breeds [32,66]. Estimates of the Middle White are still evident (see figure 2). Molecu- current Ne in several commercial taurine cattle breeds are lar studies have since provided genetic evidence of the very low (150), and the pattern of LD suggests a severe introgression from Asia to Europe. A study examining recent contraction consistent with breed formation and mitochondrial diversity in pigs revealed that a number modern breeding practices such as artificial insemination of European commercial pig breeds carry Asian-like [64,67,68]. mtDNA haplotypes [76]. The levels of Asian genetic Population contraction has also featured in the demo- introgression were highly variable, depending on the graphic history of dogs, as LD patterns suggest at least breed and commercial line, with an average of 29 per two bottlenecks: one at the time of domestication and cent frequency of Asian mtDNA haplotypes across Euro- another at the time of breed formation [69,70]. However, pean breeds. Genetic introgression can also be non- there are known difficulties in getting precise Ne estimates human-mediated, such as gene flow from wild relatives using LD patterns [71], and studies have therefore differed into the domestic pool and vice versa. For example, a in their estimates of the magnitude and timing of the dom- Chinese wild boar genotyped by Fang et al.[14] carried estication bottleneck. The study of Lindblad-Toh et al.[69] an MC1R allele common in European domestic pigs, suggests a substantial domestication-related bottle- which must have resulted from gene flow. It is not clear neck approximately 9000 generations ago, whereas that whether the introgression of grey junglefowl into the of Gray et al.[70] supports a more modest contraction primarily red junglefowl background of commercial approximately 5000 generations ago. In any case, the chickens, suggested by the presence of the yellow skin high level of LD over extended regions within dog breeds phenotype [42](see§2d), was a human-mediated event. is consistent with a more severe contraction at the time of breed-creation events [69,70]. Long runs of homo- zygosity (ROHs) are also common in most dog breeds, 4. LEVELS OF GENETIC DIVERSITY indicating recent inbreeding [25]. There is variation in One of the most interesting and somewhat surprising find- levels of LD between breeds of dogs. For example, ings arising from genetic studies of domesticated animals is Labrador retrievers have relatively low levels of LD (similar that despite the role of intensive selection, inbreeding and to that of some wolf populations), presumably because of population bottlenecks, many domesticated animal species high Ne [69,70]. are characterized by a high degree of genetic diversity. A severe contraction in size will also lead to a reduction Cattle, particularly B. indicus breeds, have substantial in the level of genetic diversity within populations. nucleotide diversity [32], indicating a large ancestral effec- Muir et al.[72] used high-density SNP data in chickens tive population size. There is also evidence from a number to estimate the proportion of ancestral alleles that are of individual genes that nucleotide variation is relatively absent from commercial chickens. In comparing the dis- high in domesticated pigs [77], where sustained gene flow tribution of alleles from commercial lines with that of with their wild boar relatives (see §3b) appears to play an various non-commercial and ancestral breeds, they esti- important role [78]. Despite the extensive bottleneck and mated that at least 50 per cent of the diversity in associated loss of alleles that accompanied the commercia- ancestral breeds is missing from commercial lines owing lization of broiler and layer lines [72], domestic chickens to bottlenecks early in the commercialization process, have extensive sequence diversity [79], again presumably continued inbreeding and industry consolidation. owing to a very large ancestral population which had even

Proc. R. Soc. B (2011) 132 Review. Genetics of animal domestication P. Wiener & S. Wilkinson 3167 greater levels of diversity (as also seen in present-day red highlighted demographic events, they also suggest diffi- junglefowl [80]). These high levels of genetic diversity con- culties in fully characterizing the history of animal tribute to the continuing ability of breeders to select for domestication using genetic data because of the concur- production traits. Despite the very low effective population rent action of multiple factors. Both selective and size of the Holstein, average milk yield has continued to non-selective forces have clearly played key roles in the increase [81]. Similarly, heritability for growth in broiler history of most domesticated species, and it may be diffi- chickens has remained at a similar level despite intensive cult to separate these factors. For example, extended selection over the last 50 years [82]. homozygosity and increased LD can derive from popu- Certain livestock breeds with particularly low popu- lation contraction and/or inbreeding as well as strong lation size (such as , discussed in §3a) selection, leading to problems distinguishing between and some purebred dogs appear to be exceptions to this these causes [84]. pattern. Many dog breeds were established with very low initial sizes, resulting in highly inbred populations and a high prevalence of inherited diseases (e.g. syringomyelia (b) Directions for further study in Cavalier King Charles Spaniels and atopic dermatitis Improvement and further development of statistical in various breeds [83]), presumably owing to the high fre- methods for identification of selection signals is an active quency of individuals homozygous for recessive alleles. area of investigation. In addition to the need for better This is reflected in the low level of nucleotide diversity ways of distinguishing between demographic and selection seen in the dog, when compared with chickens and cattle processes, new approaches may be required to adequately (electronic supplementary material, table S1). investigate the role of selection on quantitative traits such While there are many indicators to show that genetic as milk yield. Low power to detect selection on quantitative variation is being lost in domesticated animals, this traits [27] may help to explain the inconsistent picture of appears to be operating within an overall context of high selection signals seen in dairy cattle. levels of diversity in most cases, and therefore can be Another important area of further research is the identi- counteracted by informed breeding decisions. This is fication of the genes that have been selected for their impact not to suggest that conservation and breed management on tameness and other domestication-related behavioural is not required, but rather that animal breeding has not traits. While progress is currently being made in this direc- yet reached a point of no return. tion, the study of the genetic basis of these traits is still in its infancy. Long-term experimental selection for tameness in the silver fox has provided valuable insight into the domes- 5. CONCLUSIONS tication process and promises to provide even greater (a) Preliminary insights from genomic analyses understanding once genomic techniques are applied to Although identification of the genes important in animal this population. The loci underlying the rat and the fox domestication and breed development is still in its early tameness QTL do not map to orthologous regions [56], stages, some common themes have emerged. One is that and thus these studies have already demonstrated that there are clearly strong signatures of selection near a there are multiple genetic routes to evolving tameness. number of genes associated with coat colour and pattern As demonstrated by the silver fox and tame rat studies, (e.g. MC1R, KIT). This should not be surprising in that experimental populations may provide great insight into these visible phenotypes provide a clear-cut mechanism the process of domestication. There have been several for farmers and breeders to distinguish their animals recent studies examining genetic changes over the from others, and in some cases have served a cultural course of experimental selection on Drosophila [85] and role. Coat colour and pattern remain important features chicken [86] lines. More extensive analysis of this type of breeds and are still under selection. For example, of data, especially when genetic material is collected Red breeders have formed separate breed from different stages of the experiment, may allow infer- societies from Black Angus in a number of countries in ence of the processes of domestication that cannot be part because the red coat (an MC1R variant) is thought measured directly. A complementary and more direct to be more heat- and sun-tolerant than black. approach involves the analysis of ancient genetic material There are also genomic indicators that suggest selection from different historical periods. As techniques for work- on genes related to growth and body composition. There is ing with these samples improve, they will increasingly clear evidence in several cattle breeds for selection on the provide insights into the genetic changes that have myostatin gene, associated with muscle composition, and accompanied the domestication process [87]. several studies also suggest that there may have been selec- The Roslin Institute is supported by a core strategic grant tion on the GHR gene,associatedwithgrowthrateand from the UK Biotechnology and Biological Sciences various production traits. In dogs, there is also evidence of Research Council (BBSRC). S. Wilkinson is funded by a strongselectiononanumberofgenesassociatedwith CASE studentship from the BBSRC and Rare Breeds growth (e.g. IGF1) and skeletal traits, many of which are Survival Trust. related to breed-specific characteristics. The genomic picture of selection for dairy-related traits is somewhat cloudier than that seen for other cattle production characteristics. There REFERENCES is some indication of selection signals near the ABCG2 1 Darwin, C. 1859 On the origin of species by means of natural and DGAT1 genes, which have been associated with milk- selection, or the preservation of favoured races in the struggle production traits, but this is not consistent across studies. for life. London, UK: John Murray. Although these studies have indicated several genes 2 Darwin, C. 1868 The variation of animals and plants under that appear to have been under selection and have domestication, vols. I and II. London, UK: John Murray.

Proc. R. Soc. B (2011) 133 3168 P. Wiener & S. Wilkinson Review. Genetics of animal domestication

3 Grobet, L. et al. 1998 Molecular definition of an allelic regions. Genome Res. 15, 1468–1476. (doi:10.1101/gr. series of mutations disrupting the myostatin function 4398405) and causing double-muscling in cattle. Mamm. Genome 21 Coop, G. et al. 2009 The role of geography in human 9, 210–213. (doi:10.1007/s003359900727) adaptation. PLoS Genet. 5, e1000500. (doi:10.1371/ 4 Dunner, S., Miranda, M. E., Amigues, Y., Canon, J., journal.pgen.1000500) Georges, M., Hanset, R., Williams, J. & Menissier, F. 22 Akey, J. M., Ruhe, A. L., Akey, D. T., Wong, A. K., 2003 Haplotype diversity of the myostatin gene among Connelly, C. F., Madeoy, J., Nicholas, T. J. & Neff, breeds. Genet. Sel. Evol. 35, 103–118. M. W. 2010 Tracking footprints of artificial selection (doi:10.1186/1297-9686-35-1-103) in the dog genome. Proc. Natl Acad. Sci. USA 107, 5 Compere, G., Buren, R. & Hanset, R. 1996 The Belgian 1160–1165. (doi:10.1073/pnas.0909918107) Blue: the history of a cattle breed and its conquest of the 23 Zanna, G., Docampo, M. J., Fondevila, D., Bardagi, M., world. Tournai, Belgium: Casterman. Bassols, A. & Ferrer, L. 2009 Hereditary cutaneous 6 Wiener, P., Burton, D., Ajmone-Marsan, P., Dunner, S., mucinosis in Shar Pei dogs is associated with increased Mommens,G.,Nijman,I.J.,Rodellar,C.,Valentini,A.& hyaluronan synthase-2 mRNA transcription by cultured Williams, J. L. 2003 Signatures of selection? Patterns of dermal fibroblasts. Vet. Dermatol. 20, 377–382. (doi:10. microsatellite diversity on a chromosome containing a 1111/j.1365-3164.2009.00799.x) selected locus. Heredity 90,350–358.(doi:10.1038/sj.hdy. 24 Olsson, M. et al. 2011 A novel unstable duplication 6800257) upstream of HAS2 predisposes to a breed-defining skin 7 Wiener, P. & Gutierrez-Gil, B. 2009 Assessment of selec- phenotype and a periodic fever syndrome in Chinese tion mapping near the myostatin gene (GDF-8) in cattle. Shar-Pei dogs. PLoS Genet. 7, e1001332. (doi:10.1371/ Anim. Genet. 40, 598–608. (doi:10.1111/j.1365-2052. journal.pgen.1001332) 2009.01886.x) 25 Boyko, A. R. et al. 2010 A simple genetic architecture 8 Hanset, R. 1982 Muscular hypertrophy as a racial underlies morphological variation in dogs. PLoS Biol. 8, characteristic: the case of the Belgian Blue breed. In e1000451. (doi:10.1371/journal.pbio.1000451) Muscle hypertrophy of genetic origin and its use to improve 26 Flori, L., Fritz, S., Jaffrezic, F., Boussaha, M., Gut, I., beef production (eds J. W. B. King & F. Menissier), Heath, S., Foulley, J. L. & Gautier, M. 2009 The pp. 437–449. London, UK: Martinus Nijhoff Publishers. genome response to artificial selection: a case study in 9 Arthur, P.F. 1995 Double muscling in cattle: a review. Aust. dairy cattle. PLoS ONE 4, e6595. (doi:10.1371/journal. J. Agric. Res. 46, 1493–1515. (doi:10.1071/AR9951493) pone.0006595) 10 Mosher, D. S., Quignon, P., Bustamante, C. D., Sutter, 27 Wiener, P., Edriss, M. A., Williams, J. L., Waddington, N. B., Mellersh, C. S., Parker, H. G. & Ostrander, D., Law, A., Woolliams, J. A. & Gutierrez-Gil, B. 2011 E. A. 2007 A mutation in the myostatin gene increases Information content in genome-wide scans: concordance muscle mass and enhances racing performance in hetero- between patterns of genetic differentiation and linkage zygote dogs. PLoS Genet. 3, 779–786. (doi:10.1371/ mapping associations. BMC Genomics 12, 65. (doi:10. journal.pgen.0030079) 1186/1471-2164-12-65) 11 Clop, A. et al. 2006 A mutation creating a potential ille- 28 Hayes, B. J., Chamberlain, A. J., Maceachern, S., Savin, gitimate microRNA target site in the myostatin gene K., McPartlan, H., MacLeod, I., Sethuraman, L. & affects muscularity in sheep. Nat. Genet. 38, 813–818. Goddard, M. E. 2009 A genome map of divergent artifi- (doi:10.1038/ng1810) cial selection between Bos taurus dairy cattle and Bos 12 Felius, M. 1995 Cattle breeds: an encyclopedia. Doe- taurus beef cattle. Anim. Genet. 40, 176–184. (doi:10. tinchem, The Netherlands: Misset Uitgeverij BV. 1111/j.1365-2052.2008.01815.x) 13 Li, J., Yang, H., Li, J. R., Li, H. P., Ning, T., Pan, X. R., 29 Kim, Y. & Stephan, W. 2002 Detecting a local signature Shi, P. & Zhang, Y. P. 2010 Artificial selection of the mel- of genetic hitchhiking along a recombining chromosome. anocortin receptor 1 gene in Chinese domestic pigs Genetics 160, 765–777. during domestication. Heredity 105, 274–281. (doi:10. 30 Nielsen, R., Williamson, S., Kim, Y., Hubisz, M. J., 1038/hdy.2009.191) Clark, A. G. & Bustamante, C. 2005 Genomic scans 14 Fang, M. Y., Larson, G., Ribeiro, H. S., Li, N. & Anders- for selective sweeps using SNP data. Genome Res. 15, son, L. 2009 Contrasting mode of evolution at a coat 1566–1575. (doi:10.1101/gr.4252305) color locus in wild and domestic pigs. PLoS Genet. 5, 31 Stella, A., Ajmone-Marsan, P., Lazzari, B. & Boettcher, e1000341. (doi:10.1371/journal.pgen.1000341) P. 2010 Identification of selection signatures in cattle 15 Kijas, J. M. H., Wales, R., Tornsten, A., Chardon, P., breeds selected for dairy production. Genetics 185, Moller, M. & Andersson, L. 1998 Melanocortin receptor 1451–1461. (doi:10.1534/genetics.110.116111) 1 (MC1R) mutations and coat color in pigs. Genetics 150, 32 Gibbs, R. A. et al. 2009 Genome-wide survey of SNP 1177–1185. variation uncovers the genetic structure of cattle breeds. 16 Cavalli-Sforza, L. L. 1966 Population structure and Science 324, 528–532. (doi:10.1126/science.1167936) human evolution. Proc. R. Soc. Lond. B 164, 362–379. 33 Ramirez-Soriano, A. & Nielsen, R. 2009 Correcting esti- (doi:10.1098/rspb.1966.0038) mators of theta and Tajima’s D for ascertainment biases 17 Lewontin, R. C. & Krakauer, J. 1973 Distribution of caused by the single-nucleotide polymorphism discovery gene frequency as a test of theory of selective neutrality process. Genetics 181, 701–710. (doi:10.1534/genetics. of polymorphisms. Genetics 74, 175–195. 108.094060) 18 Akey, J. M., Zhang, G., Zhang, K., Jin, L. & Shriver, M. 34 Maynard Smith, J. & Haigh, J. 1974 Hitch-hiking effect D. 2002 Interrogating a high-density SNP map for signa- of a favorable gene. Genet. Res. 23, 23–35. (doi:10. tures of natural selection. Genome Res. 12, 1805–1814. 1017/S0016672300014634) (doi:10.1101/gr.631202) 35 Parker, H. G. et al. 2009 An expressed Fgf4 retrogene is 19 Endo, T., Ikeo, K. & Gojobori, T. 1996 Large-scale associated with breed-defining chondrodysplasia in dom- search for genes on which positive selection may operate. estic dogs. Science 325, 995–998. (doi:10.1126/science. Mol. Biol. Evol. 13, 685–690. 1173275) 20 Weir, B. S., Cardon, L. R., Anderson, A. D., Nielsen, D. 36 Sutter, N. B. et al. 2007 A single IGF1 allele is a M. & Hill, W. G. 2005 Measures of human popula- major determinant of small size in dogs. Science 316, tion structure show heterogeneity among genomic 112–115. (doi:10.1126/science.1137045)

Proc. R. Soc. B (2011) 134 Review. Genetics of animal domestication P. Wiener & S. Wilkinson 3169

37 Sabeti, P. C. et al. 2002 Detecting recent positive selec- of evolution of the dog. Russ. J. Genet. 40, 644–655. tion in the human genome from haplotype structure. (doi:10.1023/B:RUGE.0000033312.92773.c1) Nature 419, 832–837. (doi:10.1038/nature01140) 55 Kukekova, A. V. et al. 2007 A meiotic linkage map of the 38 Voight, B. F., Kudaravalli, S., Wen, X. Q. & Pritchard, silver fox, aligned and compared to the canine genome. J. K. 2006 A map of recent positive selection in the Genome Res. 17, 387–399. (doi:10.1101/gr.5893307) human genome. PLoS Biol. 4, 446–458. (doi:10.1371/ 56 Kukekova, A. V. et al. 2011 Mapping loci for fox domes- journal.pbio.0040446) tication: deconstruction/reconstruction of a behavioral 39 Hayes, B. J., Lien, S., Nilsen, H., Olsen, H. G., Berg, P., phenotype. Behav. Genet. 41, 593–606. (doi:10.1007/ Maceachern, S., Potter, S. & Meuwissen, T. H. E. 2008 s10519-010-9418-1) The origin of selection signatures on bovine chromosome 57 Naumenko, E. V.,Popova, N. K., Nikulina, E. M., Dygalo, 6. Anim. Genet. 39, 105–111. (doi:10.1111/j.1365-2052. N. N., Shishkina, G. T., Borodin, P. M. & Markel, A. L. 2007.01683.x) 1989 Behavior, adrenocortical activity, and brain 40 Qanbari, S., Pimentel, E. C. G., Tetens, J., Thaller, G., monoamines in Norway rats selected for reduced aggres- Lichtner, P., Sharifi, A. R. & Simianer, H. 2010 A siveness towards man. Pharmacol. Biochem. Behav. 33, genome-wide scan for signatures of recent selection in 85–91. (doi:10.1016/0091-3057(89)90434-6) Holstein cattle. Anim. Genet. 41, 377–389. 58 Albert, F. W. et al. 2011 Targeted resequencing of a geno- 41 Rubin, C. J. et al. 2010 Whole-genome resequencing mic region influencing tameness and aggression reveals reveals loci under selection during chicken domesti- multiple signals of positive selection. Heredity 107, cation. Nature 464, 587–591. (doi:10.1038/nature 205–214. (doi:10.1038/hdy.2011.4). 08832) 59 Jensen, P. 2006 Domestication: from behaviour to genes 42 Eriksson, J. et al. 2008 Identification of the yellow skin and back again. Appl. Anim. Behav. Sci. 97, 3–15. gene reveals a hybrid origin of the domestic chicken. (doi:10.1016/j.applanim.2005.11.015) PLoS Genet. 4, e1000010. (doi:10.1371/journal.pgen. 60 Svartberg, K. 2006 Breed-typical behaviour in dogs: 1000010) historical remnants or recent constructs? Appl. Anim. 43 Van Laere, A. S. et al. 2003 A regulatory mutation in IGF2 Behav. Sci. 96, 293–313. (doi:10.1016/j.applanim. causes a major QTL effect on muscle growth in the pig. 2005.06.014) Nature 425, 832–836. (doi:10.1038/nature02064) 61 Sved, J. A. 1971 Linkage disequilibrium and homozygos- 44 Grisart, B. et al. 2002 Positional candidate cloning of a ity of chromosome segments in finite populations. Theor. QTL in dairy cattle: identification of a missense mutation Popul. Biol. 2, 125–141. (doi:10.1016/0040-5809(71) in the bovine DGAT1 gene with major effect on milk 90011-6) yield and composition. Genome Res. 12, 222–231. 62 Hill, W. G. 1975 Linkage disequilibrium among multiple (doi:10.1101/gr.224202) neutral alleles produced by mutation in finite population. 45 Blott, S. et al. 2003 Molecular dissection of a quantitative Theor. Popul. Biol. 8, 117–126. (doi:10.1016/0040- trait locus: a phenylalanine-to-tyrosine substitution in the 5809(75)90028-3) transmembrane domain of the bovine growth hormone 63 Hill, W. G. 1981 Estimation of effective population-size receptor is associated with a major effect on milk yield from data on linkage disequilibrium. Genet. Res. 38, and composition. Genetics 163, 253–266. 209–216. (doi:10.1017/S0016672300020553) 46 Natt, D., Kerje, S., Andersson, L. & Jensen, P. 2007 Plu- 64 Hayes, B. J., Visscher, P. M., McPartlan, H. C. & mage color and feather pecking: behavioral differences Goddard, M. E. 2003 Novel multilocus measure of associated with PMEL17 genotypes in chicken (Gallus linkage disequilibrium to estimate past effective popu- gallus). Behav. Genet. 37, 399–407. (doi:10.1007/ lation size. Genome Res. 13, 635–643. (doi:10.1101/gr. s10519-006-9125-0) 387103) 47 Karlsson, A. C., Kerje, S., Andersson, L. & Jensen, P. 65 Tenesa, A., Navarro, P., Hayes, B. J., Duffy, D. L., 2010 Genotype at the PMEL17 locus affects social and Clarke, G. M., Goddard, M. E. & Visscher, P. M. 2007 explorative behaviour in chickens. Br. Poult. Sci. 51, Recent human effective population size estimated from 170–177. (doi:10.1080/00071661003745802) linkage disequilibrium. Genome Res. 17, 520–526. 48 Karlsson, A. C., Mormede, P., Kerje, S. & Jensen, P. (doi:10.1101/gr.6023607) 2011 Genotype on the pigmentation regulating 66 Murray, C., Huerta-Sanchez, E., Casey, F. & Bradley, PMEL17 gene affects behavior in chickens raised without D. G. 2010 Cattle demographic history modelled from physical contact with conspecifics. Behav. Genet. 41, autosomal sequence variation. Phil. Trans. R. Soc. B 312–322. (doi:10.1007/s10519-010-9379-4) 365, 2531–2539. (doi:10.1098/rstb.2010.0103) 49 Keeling, L., Anderson, L., Schutz, K. E., Kerje, S., 67 de Roos, A. P. W. 2008 Linkage disequilibrium and per- Fredriksson, R., Carlborg, O., Cornwallis, C. K., Pizzari, sistence of phase in Holstein–Friesian, Jersey and Angus T. & Jensen, P. 2004 Chicken genomics: feather-pecking cattle. Genetics 179, 1503–1512. (doi:10.1534/genetics. and victim pigmentation. Nature 431, 645–646. (doi:10. 107.084301) 1038/431645a) 68 Qanbari, S., Pimentel, E. C. G., Tetens, J., Thaller, G., 50 Bright, A. 2007 Plumage colour and feather pecking in Lichtner, P., Sharifi, A. R. & Simianer, H. 2010 The pat- laying hens, a chicken perspective? Br. Poult. Sci. 48, tern of linkage disequilibrium in German Holstein cattle. 253–263. (doi:10.1080/00071660701370483) Anim. Genet. 41, 346–356. 51 Ducrest, A. L., Keller, L. & Roulin, A. 2008 Pleiotropy 69 Lindblad-Toh, K. et al. 2005 Genome sequence, com- in the melanocortin system, coloration and behavioural parative analysis and haplotype structure of the domestic syndromes. Trends Ecol. Evol. 23, 502–510. (doi:10. dog. Nature 438, 803–819. (doi:10.1038/nature04338) 1016/j.tree.2008.06.001) 70 Gray, M. M., Granka, J. M., Bustamante, C. D., Sutter, 52 Albert, F. W. et al. 2009 Genetic architecture of tameness N. B., Boyko, A. R., Zhu, L., Ostrander, E. A. & Wayne, in a rat model of animal domestication. Genetics 182, R. 2009 Linkage disequilibrium and demographic history 541–554. (doi:10.1534/genetics.109.102186) of wild and domestic canids. Genetics 181, 1493–1505. 53 Belyaev, D. K. 1979 Destabilizing selection as a factor in (doi:10.1534/genetics.108.098830) domestication. J. Hered. 70, 301–308. 71 Wang, J. L. 2005 Estimation of effective population sizes 54 Trut, L. N., Plyusnina, I. Z. & Oskina, I. N. 2004 An from data on genetic markers. Phil. Trans. R. Soc. B 360, experiment on fox domestication and debatable issues 1395–1409. (doi:10.1098/rstb.2005.1682)

Proc. R. Soc. B (2011) 135 3170 P. Wiener & S. Wilkinson Review. Genetics of animal domestication

72 Muir, W. M. et al. 2008 Genome-wide assessment of 79 Wong, G. K. S. et al. 2004 A genetic variation map for worldwide chicken SNP genetic diversity indicates sig- chicken with 2.8 million single-nucleotide polymorph- nificant absence of rare alleles in commercial breeds. isms. Nature 432, 717–722. (doi:10.1038/nature03156) Proc. Natl Acad. Sci. USA 105, 17 312–17 317. (doi:10. 80 Granevitze, Z., Hillel, J., Chen, G. H., Cuc, N. T. K., 1073/pnas.0806569105) Feldman, M., Eding, H. & Weigend, S. 2007 Genetic 73 Visscher, P. M., Smith, D., Hall, S. J. G. & Williams, J. A. diversity within chicken populations from different 2001 A viable herd of genetically uniform cattle: deleter- continents and management histories. Anim. Genet. 38, ious alleles seem to have been purged in a feral strain of 576–583. (doi:10.1111/j.1365-2052.2007.01650.x) inbred cows. Nature 409, 303. (doi:10.1038/35053160) 81 Brotherstone, S. & Goddard, M. 2005 Artificial selection 74 Wiener, P., Burton, D. & Williams, J. L. 2004 Breed and maintenance of genetic variance in the global dairy relationships and definition in British cattle: a genetic cow population. Phil. Trans. R. Soc. B 360, 1479–1488. analysis. Heredity 93, 597–602. (doi:10.1038/sj.hdy. (doi:10.1098/rstb.2005.1668) 6800566) 82 Hill, W. G. 1998 Inferences from evolutionary biology to 75 Svensson, E. M., Anderung, C., Baubliene, J., Persson, livestock breeding In Proc. 6th World Congress on Genetics P., Malmstrom, H., Smith, C., Vretemark, M., Applied to Livestock Production, pp. 32–39. Armidale, Daugnora, L. & Gotherstrom, A. 2007 Tracing genetic Australia: Organising Committee. change over time using nuclear SNPs in ancient and 83 Rooney, N. & Sargan, D. 2009 Pedigree dog breeding modern cattle. Anim. Genet. 38, 378–383. (doi:10. in the UK: a major welfare concern? Hosham, UK: Royal 1111/j.1365-2052.2007.01620.x) Society for the Prevention of Cruelty to Animals. 76 Fang, M. Y. & Andersson, L. 2006 Mitochondrial diver- 84 Thornton, K. R., Jensen, J. D., Becquet, C. & Andol- sity in European and Chinese pigs is consistent with fatto, P. 2007 Progress and prospects in mapping recent population expansions that occurred prior to domesti- selection in the genome. Heredity 98, 340–348. cation. Proc. R. Soc. B 273, 1803–1810. (doi:10.1098/ 85 Burke, M. K., Dunham, J. P., Shahrestani, P., Thornton, rspb.2006.3514) K. R., Rose, M. R. & Long, A. D. 2010 Genome-wide 77 Ojeda, A., Rozas, J., Folch, J. M. & Perez-Enciso, M. analysis of a long-term evolution experiment with droso- 2006 Unexpected high polymorphism at the FABP4 phila. Nature 467, 587–590. (doi:10.1038/nature09352) gene unveils a complex history for pig populations. 86 Johansson, A. M., Pettersson, M. E., Siegel, P. B. & Genetics 174, 2119–2127. (doi:10.1534/genetics.106. Carlborg, O. 2010 Genome-wide effects of long-term 063057) divergent selection. PLoS Genet. 6, e1001188. (doi:10. 78 Ramirez, O. et al. 2009 Integrating Y-chromosome, mito- 1371/journal.pgen.1001188) chondrial, and autosomal data to analyze the origin of pig 87 Ludwig, A. et al. 2009 Coat color variation at the begin- breeds. Mol. Biol. Evol. 26, 2061–2072. (doi:10.1093/ ning of horse domestication. Science 324, 485. (doi:10. molbev/msp118) 1126/science.1172750)

Proc. R. Soc. B (2011) 136 Both Ends of the Leash — The Human Links to Good Dogs with Bad Genes Elaine A. Ostrander, Ph.D., National Human Genome Research Institute, National Institutes of Health, Bethesda, MD

N Engl J Med. 2012 August 16; 367(7): 636–646. doi: 10.1056/NEJMra1204453 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3508784/pdf/nihms421280.pdf (See also in Lecture notes for Dr. Ostrander’s lecture.)

Abstract. For nearly 350 years, veterinary medicine and human medicine have been separate entities, with one geared toward the diagnosis and treatment in animals and the other toward parallel goals in the owners. However, that model no longer fits, since research on diseases of humans and companion animals has coalesced.1–4 The catalyst for this union has been the completion of the human genome sequence, coupled with draft sequence assemblies of genomes for companion animals.5,6 Here, we summarize the critical events in canine genetics and genomics that have led to this development, review major applications in canine health that will be of interest to human caregivers, and discuss expectations for the future.

137 doi: 10.1111/age.12008 Variation of cats under domestication: genetic assignment of domestic cats to breeds and worldwide random-bred populations

J. D. Kurushima, M. J. Lipinski, B. Gandolfi, L. Froenicke, J. C. Grahn, R. A. Grahn and L. A. Lyons Department of Health & Reproduction, School of Veterinary Medicine, University of California – Davis, Davis, CA, 95616, USA.

Summary Both cat breeders and the lay public have interests in the origins of their pets, not only in the genetic identity of the purebred individuals, but also in the historical origins of common household cats. The cat fancy is a relatively new institution with over 85% of its 40–50 breeds arising only in the past 75 years, primarily through selection on single-gene aesthetic traits. The short, yet intense cat breed history poses a significant challenge to the development of a genetic marker–based breed identification strategy. Using different breed assignment strategies and methods, 477 cats representing 29 fancy breeds were analysed with 38 short tandem repeats, 148 intergenic and five phenotypic single nucleotide polymorphisms. Results suggest the frequentist method of Paetkau (single nucleotide polymorphisms = 0.78, short tandem repeats = 0.88) surpasses the Bayesian method of Rannala and Mountain (single nucleotide polymorphisms = 0.56, short tandem repeats = 0.83) for accurate assignment of individuals to the correct breed. Additionally, a post-assignment verification step with the five phenotypic single nucleotide polymor- phisms accurately identified between 0.31 and 0.58 of the misassigned individuals raising the sensitivity of assignment with the frequentist method to 0.89 and 0.92 for single nucleotide polymorphisms and short tandem repeats respectively. This study provides a novel multistep assignment strategy and suggests that, despite their short breed history and breed family groupings, a majority of cats can be assigned to their proper breed or population of origin, that is, race.

Keywords assignment testing, Felis catus, lineage, microsatellite, race, single nucleotide polymorphisms, short tandem repeat

nuances for breed standards and breeding practices. Fur- Introduction thermore, cat breed standards are defined by phenotypic Over the past 140 years, a plethora of pedigreed cat characteristics. Many of these phenotypes, such as hair varieties has developed due to mankind’s imposed artificial length, coat patterning and colours, are single-gene traits selection on the process of cat domestication. Since the first found at low to moderate levels in the general non- cat show in London in 1871, which showcased only five pedigreed cat population. Several commercial laboratories breeds, the development of pedigreed cats has increased in are marketing genetic tests to elucidate the breed ancestry popularity (Penny Illustrated Paper 1871). In the USA, the of dogs, ‘your best friend’ (Wisdom Panel, http://www. Cat Fanciers’ Association (CFA, http://www.cfa.org/) cur- wisdompanel.com/; Canine Heritage Breed Test, http:// rently recognises 41 breeds for competition, and The www.canineheritage.com/), prompting cat owners to International Cat Association (TICA, http://www.tica.org/) wonder about the ancestral origins of their own feline accepts 57 breeds. A majority of the breeds acknowledged companions. by these two large registries are also typical breeds around Because random-bred house cats have a different history the world; however, each has specific compared to dogs, genetic testing for breed and population assignments requires a slightly different approach. Whereas Address for correspondence the average canine found in the streets of most developed

L. A. Lyons, Department of Health & Reproduction, School of countries is more likely a cross-bred individual from Veterinary Medicine, 4206 VetMed 3A, University of California – multiple purebred lines, the average random-bred cat is Davis, Davis, CA 95616, USA. not a descendant of its pedigreed counterparts. For cats, the E-mail: [email protected] opposite scenario is more likely – pedigreed feline stocks are Accepted for publication 23 August 2012 the descendants of common street cats from discrete parts of

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324 311

138 312 Kurushima et al.

the world that have been selected for one or more distinct Two studies have evaluated the genetic distinction of cat traits (Table 1). Random-bred cats are the original popula- breeds. Lipinski et al. (2008) defined the connections tions from which the breeds developed, not a population of between the random-bred cat populations and their pedigreed cats gone feral. Also, converse to most dog descendant pedigreed lines using a DNA marker panel registries, to improve population health and reduce the containing two tetranucleotide and 36 dinucleotide effects of inbreeding depression, cat breeding associations microsatellites [a.k.a. short tandem repeat (STR)] markers. often seek to diversify their breed populations with random- Five hundred fifty-five individuals were demarcated into 20 bred cats from the breed’s presumed ancestral origin. For breeds. Four breeds remained unresolved as the selected this reason, most cat registries use the term ‘pedigreed’ and markers lacked sufficient power for demarcation, suggest- not ‘purebred’. ing the grouping of same cat breeds into breed families.

Table 1 Traditional cat breed origins.

Fixed or hallmark1 Date of Breed phenotype2 Origin establishment Derived breeds

Abyssinian Shorthair, ticked, agouti India, Africa 1868 Somali3 American Bobtail1 Bobtail Mutation-USA 1960 American Curl1 Rostral curl to pinnea Mutation-USA 1981 American Shorthair USA 1966 Wired hair Mutation-USA 1966 Mix-Australia 1990s Several breeds Birman Siamese points, gloves, Burma <1868 Snowshoe3 British Shorthair longhair England 1870s Burmese Non-agouti, Burmese Burma 1350–1767 Asian, Bombay, Tiffanie3, points Malayan, Dilute, non-agouti France XIV century Cornish Rex Curly coat Mutation-UK 1950 Devon Rex1 Curly coat Mutation-UK 1960 Sphynx (1966) Egyptian Mau Shorthair Egypt 1953 Europe Bobtail Japan VI–XII century Korat Dilute, non-agouti Thailand 1350–1767 LaPerm1 Curly coat Mutation-USA 1986 Maine Coon Longhair USA 1860s Manx1 No tail Isle of Man <1868 Cymric3 Munchkin1 Short legs USA 1990s Norwegian Forest Longhair Norway <1868 Ocicat Spots Crossbred 1964 Siamese 9 Abyssinian Blue eyes Mutation-USA 1980s Persian Longhair Persia <1868 Exotic3, Kashmir, Himalayan, Peke-faced, Burmilla Ragdoll Longhair USA 1960s Ragamuffin Dilute, non-agouti Russia <1868 Nebelung3 Scottish Fold1 Ventral fold to pinnea Mutation 1961 Highland Fold3 (Coupari) Selkirk Rex1 Curly coat Mutation-USA 1980s Siamese Siamese Points, Shorthair, Thailand 1350–1767 Colorpoint3, Javanese3, Non-agouti Balinese3, Oriental3 Siberian Longhair Russia <1868 Havana Brown, Don Sokoke Africa Sphynx, Sphynx Hairless Canada 1960s Devon Rex Tonkinese1 Shorthair, heterozygous Burmese Crossbred 1950s Siamese 9 Burmese and Siamese points Turkish Angora Longhair Ankara, Turkey XV century Turkish Van Longhair Van Lake, Turkey <1868

Origins are according to: Gebhardt (1991), The Royal Canin Encyclopedia (2000), TICA (http://tica.org/) and Australian Mist Breed Council (http:// www.australianmist.info/Home.html). 1Some breeds allow variants that do not have the hallmark trait, such as straight-eared American Curls or straight-coated Selkirk Rex. The Tonkinese has colour variants that produce Siamese and Burmese colorations. These variants are available for breeding but not for competition. 2Many breeds have limited colorations and patterns that vary between registries. Only the most definitive colourations and patterns across most registries are presented. 3Many derived breeds are long- or shorthaired varieties of the foundation breed but have different breed names; others are delineated by longhair or shorthair in the breed name. Several additional rex-coated cat populations have not developed into viable populations or are extinct.

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

139 Variation of cats under domestication 313

Furthermore, the breeds sampled by Lipinski et al. were in particular for closely related cat breeds that are demar- shown to be similar to the populations of street cats found cated by these single-gene traits. in Europe, the Eastern Mediterranean and Southeast Asia. Menotti-Raymond et al. (2008) used a panel of 11 Materials and methods tetranucleotide STR markers to characterise the delinea- tion of cat breeds. Using only the STR markers, 1040 Sample collection and genotyping individuals were demarcated into eight individual breeds and nine additional breed groups. Twenty breeds could not Twenty-nine breeds were represented by 477 cats. This be resolved at the breed level. These studies indicate that study included 354 cats from the work of Lipinski et al. distinct populations and breeds of cats can be defined (2008) that analysed 22 breeds. The 123 newly collected genetically, that breeds do have different worldwide samples represented seven additional breeds, including origins, tetranucleotide STRs do not perform as well as Scottish Fold, Cornish Rex, Ragdoll, Manx, Bengal, dinucleotide markers defining cat breeds, and some breeds and Australian Mist. All cats were representatives of their are so closely related that they cannot be distinguished breed as found within the USA, except for the Australian with even the rapidly evolving dinucleotide STRs. Mist Cats and a few Turkish Angora and Turkish Van The 38 highly polymorphic markers of Lipinski et al. samples from international submissions. Additionally, all (2008) and a recently developed panel of 148 intergenic cats were pedigreed and verified to be unrelated to the autosomal single nucleotide polymorphisms (SNPs) were grandparent level. Worldwide random-bred data (n = 944) recently applied to an extensive sample of random-bred were included from the previous study of Kurushima (2011) street cats collected throughout the world (Kurushima to assess the origins of each of the breed populations. New 2011). Nine hundred forty-four samples were collected samples were collected via a buccal (cheek) swab and from 37 locations spread throughout North and South extracted using the Qiagen DNeasy Blood and Tissue America, Europe, Africa and Asia. The study found both following the manufacturer’s protocol. marker sets to be efficient at distinguishing five long- Thirty-eight STRs were genotyped in the 123 newly established races; however, a few geographically close acquired cats following the PCR and analysis procedures of populations were better delineated with either SNPs or Lipinski et al. (2008). Unlinked non-coding autosomal SNPs STRs, most likely due to varying mutation rates between (n = 169) were selected to evenly represent all autosomes the markers. from the 1.99 coverage cat genomic sequence, which was Many methods of assignment testing have been devel- defined by one as resequencing data were oped using a variety of both genetic markers and statistical not available at the time of selection (Pontius et al. 2007). methods (Paetkau et al. 1995; Rannala & Mountain 1997; Primers were designed with the VeraCode Assay Designer Pritchard et al. 2000; Baudouin & Lebrun 2001). These software (Illumina, Inc.). Only SNPs that received a design techniques have been applied to various breeding popula- score of 0.75 or higher (with a mean design score of 0.95) tions including pigs, cattle and dogs (Schelling et al. 2005; (n = 162) were included in the analysis (Table S1). Five Negrini et al. 2009; Boitard et al. 2010). In cattle, Negrini additional phenotypic SNPs were also evaluated in all cats. et al. (2009) used 90 SNPs to both allocate and then The phenotypic SNPs consisted of a causative mutation for assign 24 breeds under both the Bayesian methods of the most common form of longhair in cats [AANG0202725 Pritchard et al. (2000), Rannala & Mountain (1997) and 0.1(FGF5):g.18442A>C] (Kehler et al. 2007), Burmese and Baudouin & Lebrun (2001), and the likelihood method of Siamese colour points [AANG02171092.1(TYR):g.11026G Paetkau et al. (1995). Negrini et al. (2009) concluded that >T and AANG02171093.1(TYR):g.1802G>A respectively] the Bayesian and frequentist methods, implemented (Lyons et al. 2005b) and the mutations for the colour respectively through Rannala & Mountain (1997; Bayes- variants chocolate and cinnamon [AY804234S6(TYRP1): ian) and Paetkau et al. (1995; frequentist), worked best g.593G>A and AANG02185848.1(TYRP1):g.10736C>T when attempting to assign unknown individuals to a respectively] (Lyons et al. 2005a). known database of representative samples from each Golden Gate Assay amplification and BeadXpress reads breed. were performed per the manufacturer’s protocol (Illumina, This article assesses the ability of a panel of 148 evenly Inc.) on 50–500 ng of DNA or whole-genome amplified dispersed genome-wide SNPs for population assignment of product. BEADSTUDIO software v. 3.1.3.0 with the Genotyping domestic cats. Different assignment techniques are exam- module v. 3.2.23 (Illumina, Inc.) was used to analyse the ined in a species exhibiting many recent and extreme data. Samples with a call rate <0.80 (n = 21) were removed population bottlenecks in addition to large numbers of from further clustering analysis. Additionally, only SNPs population migrants, also comparing the power and with a GenTrain Score >0.55 (n = 148) were included in the efficiency of this 148 SNP panel to fourfold fewer STRs. analysis (Table S1). Each run of the SNP assay contained The strength of phenotypic DNA variants is tested for both an internal positive and negative control to validate sensitivity and specificity to support individual assignment, repeatability and detect contamination.

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

140 314 Kurushima et al.

Population statistics (Piry et al. 2004). Breeds were assigned to the race that produced the highest log(likelihood) value. Hardy–Weinberg equilibrium (HWE) with associated chi- squared tests, as well as observed and expected heterozy- gosity, was calculated by breed with GENALEX v.6.3 (Peakall & Assignment testing Smouse 2006). Inbreeding coefficients (F ) within each IS Ten sets of 50 individuals were selected randomly from the breed and between-population variation values (F ) were ST sample set and assigned to a population of origin using the calculated with FSTAT v. 2.9.3.2 (Goudet 1995). Because of remaining samples as the reference populations using the predicted recent separation (co-ancestry) and small GENECLASS2 v.2.0.h (Piry et al. 2004). The Bayesian method population sizes of the breeds under consideration, Rey- of Rannala & Mountain (1997) and the frequentist method nold’s genetic distance was calculated between all pairs of suggested by Paetkau et al. (1995) were compared, as these breeds with the SNP data set (Reynolds et al. 1983). Nei’s methods performed best in the previous assignment study of genetic distance was used with the STR data set to Negrini et al. (2009) when compared to the Pritchard et al. accommodate the rapid mutation rate characteristic of (2000) and the Audoulin & Lebrun methods (2001). STRs (Nei 1972). Both distances were implemented through Average probabilities were computed using the Paetkau the software package PHYLIP v. 3.69 (Felsenstein 1989). et al. (2004) Monte Carlo resampling method through a simulation of 1000 individuals and a type I error rate (a)of Population structuring 0.01. Additionally, the assignment tests were performed in three iterations: intergenic SNPs, intergenic and phenotypic Bayesian clustering SNPs combined and STRs. Tallies of type I error (an Data sets were analysed with the Bayesian clustering individual not reassigned to its population of origin) and program STRUCTURE v.2.3.1 (Pritchard et al. 2000) under type II error (an individual assigned to the wrong popula- the admixture model with correlated allele frequencies and tion) were used to calculate the sensitivity and specificity of a burn in of 100 000 with 100 000 additional iterations. the assignment method (Negrini et al. 2009). Values of Q were calculated from K = 1toK = 33; each run The differences of the STR and SNP assignments also was replicated 10 times. Posterior log-likelihoods were used were compared, post-assignment, with and without the use to calculate ΔK to best estimate the number of ancestral of phenotypic SNPs. Cats were considered misassigned if populations through the program HARVESTER v.0.56.4 they had genotypes exclusionary for the breed, for example, (Evanno et al. 2005). All 10 iterations were then combined an individual assigned to the Exotic Shorthair breed was through the program CLUMPP v.1.1.2 (Jakobsson & Rosen- identified as misassigned if it was homozygous for longhair, berg 2007) to create a consensus clustering. To assess the a recessive trait in cats not found in that breed (see Table 1 effects of varying marker types on the final results, analysis for phenotypic diagnostic to breeds). using STRUCTURE was conducted with the two different data sets, SNPs and STRs. Results

Principal coordinate analysis Summary statistics Principal coordinate analyses were conducted on the Pedigreed cats (n = 477), representing 29 recognised Reynold’s (SNPs) and Nei’s (STRs) genetic distance matrices breeds, were included in this study (Table 2). Analysis of using the software GENALEX v.6.3 (Peakall & Smouse 2006). all cats from the previous Lipinski et al. (2007) study was For the PCA plots, both the data in the present manuscript attempted; however, DNA quality and quantity caused and data from the worldwide random-bred populations some sample loss, as did available SNP analysis resources. (Kurushima 2011) were considered to show the relation- The number of cats per breed ranged from 7 to 25 with an ship of the cat breeds and their random-bred population average of 16.4 individuals per breed. STRs had an average origins. call rate of 88.2%, and SNPs had a 94.0% average call rate. Although the chi-squared goodness-of-fit test indicated that 126 of the 148 SNPs and 36 of the 38 STRs were not in Breed race assignment HWE in at least one breed group, only one SNP marker Cat breed populations were assigned to the eight ancestral (AANG02147808.1:g.9376T>C) was not in HWE in more races of random-bred worldwide populations of cats than 50% of the breeds (Table S2). Twenty-seven breeds (Europe, Mediterranean, Egypt, Iraq/Iran, Arabian Sea, have 10–25 loci not in HWE; however, the Russian Blue India, Southeast Asia and East Asia) identified in the and Turkish Van breeds have 31 and 33 of the 186 genetic previous study by Kurushima (2011) by calculating log markers not in HWE respectively. The frequency of the (likelihood) values using the Bayesian population assign- genotypes and alleles for the phenotypic SNPs are indicated ment methods available in the software GENECLASS2 v.2.0.h in Table 3. The FGF5 mutation AANG02027250.1:

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

141 Variation of cats under domestication 315

Table 2 Population statistics of cat breeds.

Total Total

Breed n Alleles(SNP) Alleles(STR) PAB(STR) PAW(STR) Na(SNP) Na(STR) Ho(SNP) Ho(STR) FIS (SNP) FIS (STR)

Abyssinian 15 277 130 1 1 1.87 3.42 0.29 0.42 0.02 0.11 American SH 13 269 168 0 0 1.82 4.42 0.28 0.55 À0.02 0.04 Australian Mist 15 273 156 4 0 1.85 4.11 0.27 0.57 À0.01 À0.05 Bengal 18 274 192 10 2 1.85 5.05 0.24 0.58 0.07 0.03 Birman 20 247 133 3 0 1.67 3.50 0.17 0.44 0.13 0.03 British SH 18 276 192 2 0 1.87 5.05 0.24 0.55 0.10 0.06 Burmese 19 262 158 2 1 1.77 4.16 0.20 0.42 0.08 0.16 Chartreux 13 264 151 0 0 1.78 3.97 0.24 0.56 0.10 0.04 Cornish Rex 15 262 163 2 0 1.77 4.29 0.24 0.56 0.05 0.03 Egyptian Mau 14 268 160 1 0 1.81 4.21 0.25 0.50 0.03 0.11 Exotic SH 19 279 178 1 1 1.89 4.68 0.25 0.53 0.07 0.07 Havana Brown 14 245 113 1 0 1.66 2.97 0.17 0.42 0.12 À0.02 Japanese Bobtail 19 267 191 4 0 1.80 5.03 0.22 0.58 0.15 0.08 Korat 25 246 150 2 0 1.66 3.95 0.17 0.52 0.08 0.03 Maine Coon 19 282 210 2 1 1.91 5.53 0.26 0.60 0.11 0.04 Manx 17 282 233 6 2 1.91 6.13 0.30 0.70 0.00 À0.02 Norwegian Forest 16 284 248 8 0 1.92 6.45 0.28 0.67 0.06 0.02 Ocicat 10 264 142 3 2 1.78 3.74 0.24 0.50 0.04 0.05 Persian 15 276 181 1 0 1.87 4.76 0.29 0.50 À0.02 0.15 Ragdoll 15 265 178 4 0 1.79 4.68 0.29 0.62 À0.06 0.00 Russian Blue 17 259 146 2 1 1.75 3.84 0.19 0.45 0.16 0.06 Scottish Fold 17 269 180 2 1 1.82 4.74 0.26 0.57 0.00 0.05 Siamese 15 242 133 2 1 1.64 3.50 0.20 0.47 0.00 0.02 Siberian 17 275 227 4 2 1.86 5.97 0.26 0.71 0.09 À0.06 Singapura 17 232 94 1 0 1.57 2.47 0.18 0.34 0.06 0.02 7 222 92 0 0 1.50 2.42 0.17 0.37 0.00 0.00 Sphynx 17 277 178 2 0 1.87 4.68 0.27 0.55 0.05 0.05 Turkish Angora 21 284 275 10 1 1.92 7.24 0.25 0.67 0.11 0.06 Turkish Van 20 277 259 6 0 1.87 6.82 0.24 0.60 0.12 0.12 Total 477 296 490 1.79 4.54 0.24 0.53 0.06 0.04 n, number of samples; PAB, private alleles within breeds; PAW, private alleles within breeds and worldwide random-bred populations; Na, average effective number of alleles; Ho, observed heterozygosity; SNPs, single nucleotide polymorphisms; STRs, short tandem repeats; FIS, inbreeding coefficient. SNP statistics were calculated using intergenic SNPs only. g.18442A>C causing longhaired cats in the homozygous The average SNP-based observed heterozygosity was state was by far the most prevalent of the phenotypic SNPs, 0.24, ranging from 0.17 to 0.30, whereas the average which was found in all but eight of the breeds. In contrast, STR-based observed heterozygosity was 0.53, ranging from coat colour cinnamon, caused by AANG02185848.1 0.34 to 0.71 (Table 2, Fig. S1). FIS were lowest in the (TYRP1):g.10736C>T, was observed in only five breeds, Ragdoll (À0.06) and Siberian (À0.06) with SNPs and STRs two breeds having a frequency lower than 0.1. respectively and highest within the Australian Mist Cats (0.16) and Burmese (0.16). Between-population variation F values were 0.24 ± 0.01 with SNPs and 0.27 ± 0.02 Genetic diversity ST with STRs (data not shown). The population’s genetic data are presented in Table 2. Effective SNP alleles ranged from 1.50 to 1.92 with an across Breed clustering breed average of 1.79. The average effective number of STR alleles observed was 4.54 across breeds, ranging from 2.42 to The most likely value of K, the number of structured 7.23. Private STR alleles within breeds ranged from 0 to 10. groupings, could not be decisively determined. A significant However, when compared to worldwide random-bred popu- difference between the log-likelihoods was not evident for lations, private alleles within breeds dropped to between 0 either marker type between K = 17–33 (Fig. S2); however, a and 2 per breed (Table 2). No SNPs had private alleles in a plateau was suggested near K = 21 for STRs and near breed, although breeds had anywhere from 12 to 74 SNP K = 17 for SNPs; the STRUCTURE plots are presented in Fig. 1. alleles fixed within their population (Turkish Angora and As a result, a combination of the ΔK plots and common Sokoke respectively), and the minor allele frequency aver- sense directed selection of the most likely number of aged across all loci ranged from 0.22 in Bengal to 0.32 in populations. For STRs, at K > 24 (Fig. S3a), different Abyssinian with a mean of 0.25 (data not shown). lineages (breed lines) within specific breeds, such as

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

142 316 Kurushima tal. et

Table 3 Phenotypic SNP frequencies

Longhair Burmese Points Siamese Points Chocolate Cinnamon

© > > > + > > 02TeAtos nmlGenetics Animal Authors, The 2012 FGF5 475A C TYR 715G T TYR 940G A TYRP1 1373 5G A TYRP1 298C T

Breed n* AA AC CC Freq. C n1 GG GT TT Freq. T n1 GG GA AA Freq. A N1 GG GA AA Freq. A n1 CC CT TT Freq. T

Abyssinian 15 15 0 0 0 15 15 0 0 0 15 15 0 0 0 15 12 3 0 0.10 15 4 6 5 0.53 American SH 13 10 2 1 0.15 13 11 2 0 0.08 13 13 0 0 0 13 12 0 1 0.08 13 13 0 0 0 Australian Mist 13 11 2 0 0.08 15 2 0 13 0.87 12 10 1 1 0.13 15 6 6 3 0.40 15 7 7 1 0.30 Bengal 16 15 1 0 0.03 18 16 2 0 0.06 14 9 4 1 0.21 18 16 2 0 0.06 17 17 0 0 0 Birman 19 0 0 19 1.00 20 20 0 0 0 16 0 0 16 1.00 20 12 5 3 0.28 20 20 0 0 0 British SH 18 16 2 0 0.06 18 18 0 0 0 17 13 0 4 0.24 18 11 2 5 0.33 18 15 3 0 0.08 Burmese 19 19 0 0 0 19 0 1 18 0.97 16 16 0 0 0 19 9 4 6 0.42 19 18 0 1 0.05 Chartreux 10 5 5 0 0.25 13 13 0 0 0 11 11 0 0 0 13 13 0 0 0 13 13 0 0 0 Cornish Rex 15 14 1 0 0.03 15 15 0 0 0 14 3 4 1 0.21 14 13 1 0 0.04 15 15 0 0 0 143 © Egyptian Mau 12 12 0 0 0 14 14 0 0 0 12 12 0 0 0 14 14 0 0 0 14 14 0 0 0 02SihigItrainlFudto o nmlGenetics, Animal for Foundation International Stichting 2012 Exotic SH 17 5 10 2 0.41 19 19 0 0 0 17 14 2 1 0.12 19 15 3 1 0.13 19 19 0 0 0 Havana Brown 11 11 0 0 0 14 14 0 0 0 12 10 1 1 0.13 14 0 1 13 0.96 14 14 0 0 0 Japanese Bobtail 14 8 2 4 0.36 18 18 0 0 0 15 13 2 0 0.07 19 19 0 0 0 19 19 0 0 0 Korat 23 22 1 0 0.02 25 25 0 0 0 21 20 1 0 0.02 25 1 2 22 0.92 25 25 0 0 0 Maine Coon 14 0 0 14 1.00 18 18 0 0 0 17 17 0 0 0 19 16 2 1 0.11 19 19 0 0 0 Manx 15 8 6 1 0.27 17 17 0 0 0 16 16 0 0 0 17 16 1 0 0.03 17 17 0 0 0 Norwegian Forest 13 8 3 2 0.27 16 16 0 0 0 15 15 0 0 0 16 15 0 1 0.06 16 16 0 0 0 Ocicat 8 8 0 0 0 10 10 0 0 0 9 9 0 0 0 10 4 1 5 0.55 10 6 3 1 0.25 Persian 15 0 1 14 0.97 15 15 0 0 0 15 5 4 6 0.53 15 12 2 1 0.13 15 15 0 0 0 Ragdoll 15 4 3 8 0.63 15 15 0 0 0 15 0 0 15 1.00 15 13 2 0 0.07 15 15 0 0 0 Russian Blue 15 14 0 1 0.07 17 16 1 0 0.03 15 11 3 1 0.17 17 17 0 0 0 17 17 0 0 0 Scottish Fold 16 13 3 0 0.09 17 17 0 0 0 15 14 1 0 0.03 17 13 4 0 0.12 17 17 0 0 0 Siamese 15 15 0 0 0 15 15 0 0 0 13 0 0 13 1.00 15 2 6 7 0.67 15 15 0 0 0 Siberian 14 1 3 10 0.82 16 16 0 0 0 15 8 6 1 0.27 17 16 1 0 0.03 17 17 0 0 0 Singapura 16 16 0 0 0 15 0 0 15 1.00 14 14 0 0 0 17 17 0 0 0 17 17 0 0 0 Sokoke 6 6 0 0 0 7 7 0 0 0 4 3 0 1 0.25 6 5 1 0 0.08 7 7 0 0 0 Sphynx 16 9 1 6 0.41 16 6 6 4 0.44 12 9 1 2 0.21 17 8 5 4 0.38 17 17 0 0 0 Turkish Angora 20 0 0 20 1.00 21 21 0 0 0 20 17 1 2 0.13 20 15 5 0 0.13 21 21 0 0 0 Turkish Van 18 0 0 18 1.00 19 19 0 0 0 20 19 1 0 0.03 19 14 2 3 0.21 20 20 0 0 0

44 1All individuals were attempted for all phenotypic single nucleotide polymorphisms (SNPs); differing sample sizes are due to assay dropout. 311–324 , Variation of cats under domestication 317

Figure 1 Bayesian clustering of cat breeds. Clustering of breeds at K = 17 and K = 21 as calculated with single nucleotide polymorphisms (SNPs) and short tandem repeats (STRs) respectively. Each column represents an individual cat. The y-axis represents Q or the proportional estimate of genetic membership to the given cluster (K). Each K cluster is indicated by a unique colour.

Norwegian Forest Cat and Turkish Angora, became appar- Norwegian Forest Cat and Maine Coon, grouped with the ent before five other breed groups would delineate: Persian/ Western European random-bred cats; Turkish Angora and Exotic SH, British SH/Scottish Fold, Australian Mist/Bur- Turkish Van assigned to the Eastern Mediterranean cats mese, Birman/Korat and Siamese/Havana Brown. Similar and the Sokoke to the India/Arabian Sea region. Three results were found for the SNP-based analyses; however, the breeds showed regional variation depending on the marker associations of the Asian-based breeds varied (Fig. S3b). type used for assignment. When analysed with data from SNPs appear to resolve the Birman and Singapura breeds SNPs and STRs, the Turkish Angora was assigned to Europe from the other Asian breeds more readily. Considering both or to the Eastern Mediterranean, Bengal was assigned to SNPs and STRs, Persians appear to have influenced several Europe or to the Arabian Sea, and Ocicat was assigned to breeds: Exotic Shorthair, Scottish Fold, British Shorthair South Asia or Europe. and, to a lesser extent, Chartreux (Fig. 1). Within breeds of Asian heritage, Siamese have a strong influence on the Assignment testing Havana Brown, Korat and, to a lesser extent, Birman and Singapura (Fig. 1). The accuracy of assignment testing varied depending upon The principal coordinate analyses indicated the relation- not only the assignment method but also the marker type ship of the breeds and their likely closest random-bred used to differentiate the cat breeds. For example, when origins, that is, race (Fig. 2). The breeds that originated comparing the Bayesian method of Rannala & Mountain solely from European and American random-bred cats (1997) versus the frequentist method of Paetkau et al. clustered with the random-bred populations of Europe and (1995), the average sensitivity of assignment for the 148 America. Likewise, breeds with Asian descent grouped with non-phenotypic SNPs was 0.56 and 0.78 respectively South Asian populations of random-bred cats. The breeds (Table 4a and b). When the five phenotypic SNPs were that do not share similar coordinates with a random-bred included with the random SNPs, the average assignment population, such as Russian Blue, Ocicat, Singapura, sensitivity was 0.54 ± 1.4 and 0.83 ± 0.09 respectively. Australian Mist and Birman, have a strong influence from Overall, the STRs had higher average sensitivities of both Europe and Asia. 0.83 ± 0.05 and 0.88 ± 0.04 respectively. In six breeds, Using Bayesian clustering, the breeds were then assigned adding phenotypic SNPs into the frequentist assignment of back to the eight random-bred races of Kurushima (2011) individuals reduced the sensitivity of the test, and in six (Table S3a,b). Four regional areas seem to have contributed breeds, specificity was reduced. to the development of the considered cat breeds. Asian The post-assignment allocation using the five phenotypic breeds, such as Birman, Burmese and Siamese, grouped SNPs was able to correctly classify 57.5% of the 221 with Southern Asian cats; Western breeds, such as Persian, animals originally misassigned by the Bayesian method

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

144 318 Kurushima et al.

(a)

(b)

Figure 2 Principal coordinate analysis of cat breeds and worldwide random-bred cat populations. Colour shades indicate the population membership of the respective random-bred populations as determined by Kurushima (2011). Green, European or European-derived; light blue, Eastern Mediterranean; dark blue, Egypt; purple, Iraq/Iran; light pink, Arabian Sea; dark pink, India; light orange, Southeast Asia; dark orange, East Asia; white, pedigreed breed groups. (a) single nucleotide polymorphisms (SNPs) as calculated by Reynold’s genetic distance (Reynolds et al. 1983); (b) short tandem repeats (STRs) as calculated by Nei’s genetic distance.

with the intergenic SNPs and 50% of the 110 individuals effective in the STR assignments (identifying 27% and 32% originally misallocated by the frequentist method (Table 5). of the Bayesian and frequentist misassignments respec- The phenotypic-based corrections increased the sensitivity tively). The influence of recent breed development on the and specificity of the Bayesian method to 0.75 and 0.77 misassignment of individuals may be further visualised by respectively and the frequentist to 0.89 (both sensitivity plotting the crossed assignment rate as a function of the and specificity) and resulted in better resolution than did genetic distance between breeds (Fig. S4a,b). The crossed the use of intergenic SNPs alone (data not shown). The assignment rate increased as the genetic distance between effect of using phenotypic SNPs post-assignment was less breeds decreased.

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

145 Variation of cats under domestication 319

Table 4 Assignment accuracy of cats to breeds using the (a) Bayesian method, (b) frequentist method.

Intergenic SNPs Intergenic and phenotypic SNPs STRs

Ave. Ave. Ave.

Breed n EI EII Sens. Spec. Prob. EI EII Sens. Spec. Prob. EI EII Sens. Spec. Prob.

(a) Bayesian method Abyssinian 11 4 0 0.64 1.00 0.98 4 0 0.64 1.00 0.98 2 0 0.82 1.00 0.54 American SH 17 8 0 0.53 1.00 0.99 11 0 0.35 1.00 1.00 4 0 0.76 1.00 0.54 Australian Mist 20 0 1 1.00 0.95 1.00 0 2 1.00 0.91 1.00 2 15 0.90 0.55 0.92 Bengal 23 7 0 0.70 1.00 1.00 2 0 0.91 1.00 0.99 0 0 1.00 1.00 0.79 Birman 22 4 0 0.82 1.00 0.96 6 0 0.73 1.00 0.99 1 0 0.95 1.00 0.72 British SH 17 17 10 0 0 1.00 13 5 0.24 0.44 0.99 7 1 0.59 0.91 0.24 Burmese 16 4 1 0.75 0.92 1.00 2 0 0.88 1.00 1.00 4 1 0.75 0.92 0.86 Chartreux 11 1 13 0.91 0.43 1.00 1 7 0.91 0.59 1.00 1 1 0.91 0.91 0.61 Cornish Rex 23 12 0 0.48 1.00 0.97 14 0 0.39 1.00 0.98 5 0 0.78 1.00 0.58 Egyptian Mau 14 3 0 0.79 1.00 1.00 4 0 0.71 1.00 1.00 1 0 0.93 1.00 0.59 Exotic SH 22 16 8 0.27 0.43 1.00 17 6 0.23 0.45 1.00 6 1 0.73 0.94 0.69 Havana Brown 15 2 2 0.87 0.87 1.00 2 1 0.87 0.93 1.00 0 0 1.00 1.00 0.93 Japanese Bobtail 18 2 33 0.89 0.33 1.00 7 34 0.61 0.24 1.00 1 0 0.94 1.00 0.55 Korat 24 0 15 1.00 0.62 1.00 0 17 1.00 0.59 1.00 2 0 0.92 1.00 0.55 Maine Coon 27 3 21 0.89 0.53 1.00 10 32 0.63 0.35 1.00 6 1 0.78 0.95 0.61 Manx 22 20 1 0.09 0.67 1.00 21 1 0.05 0.50 1.00 4 16 0.82 0.53 0.48 Norwegian Forest 16 8 4 0.50 0.67 1.00 5 25 0.69 0.31 1.00 2 25 0.88 0.36 0.41 Ocicat 7 4 0 0.43 1.00 0.99 3 1 0.57 0.80 0.99 1 0 0.86 1.00 0.63 Persian 12 12 0 0 11 10 0 0.17 1.00 1 2 13 0.83 0.43 0.57 Ragdoll 16 16 0 0 11 16 0 0 11 5 0 0.69 1.00 0.6 Russian Blue 19 0 0 1 1.00 1.00 4 0 0.79 1.00 1.00 3 0 0.84 1.00 0.93 Scottish Fold 19 18 0 0.05 1.00 1.00 16 0 0.16 1.00 1.00 6 0 0.68 1.00 0.67 Siamese 19 19 0 0 11 19 0 0 11 1 0 0.95 1.00 0.63 Siberian 9 9 0 0 11 6 23 0.33 0.12 1.00 1 9 0.89 0.47 0.27 Singapura 19 1 0 0.95 1.00 1.00 0 0 1.00 1.00 1.00 2 0 0.89 1.00 0.86 Sokoke 5 0 0 1.00 1.00 1.00 0 0 1.00 1.00 1.00 0 0 1.00 1.00 0.81 Sphynx 25 16 0 0.36 1.00 0.99 14 0 0.44 1.00 0.99 3 0 0.88 1.00 0.34 Turkish Angora 18 5 125 0.72 0.09 1.00 11 134 0.39 0.05 1.00 11 2 0.39 0.78 0.46 Turkish Van 14 10 3 0.29 0.57 0.98 13 3 0.07 0.25 0.99 3 3 0.79 0.79 0.69

All Breeds 500 221 237 0.56 0.54 0.99 231 291 0.54 0.48 1.00 86 88 0.83 0.82 0.63 95% confidence 0.14 0.12 0.13 0.13 0.05 0.08 interval

(b) Frequentist method Abyssinian 11 0 0 1.00 1.00 0.32 0 0 1.00 1.00 0.32 2 0 0.82 1.00 0.33 American SH 17 1 0 0.94 1.00 0.53 4 0 0.76 1.00 0.60 2 0 0.88 1.00 0.27 Australian Mist 20 0 2 1.00 0.91 0.57 0 3 1.00 0.87 0.58 2 1 0.90 0.95 0.27 Bengal 23 2 0 0.91 1.00 0.43 2 0 0.91 1.00 0.43 0 0 1.00 1.00 0.21 Birman 22 1 0 0.95 1.00 0.39 1 0 0.95 1.00 0.38 1 0 0.95 1.00 0.34 British SH 17 10 6 0.41 0.54 0.45 5 4 0.71 0.75 0.33 5 3 0.71 0.80 0.16 Burmese 16 2 2 0.88 0.88 0.51 3 0 0.81 1.00 0.51 0 2 1.00 0.89 0.26 Chartreux 11 2 0 0.82 1.00 0.31 2 0 0.82 1.00 0.31 0 0 1.00 1.00 0.15 Cornish Rex 23 5 0 0.78 1.00 0.29 4 1 0.83 0.95 0.30 2 0 0.91 1.00 0.25 Egyptian Mau 14 1 0 0.93 1.00 0.29 2 0 0.86 1.00 0.32 3 0 0.79 1.00 0.18 Exotic SH 22 19 7 0.14 0.3 0.43 10 5 0.55 0.71 0.37 4 1 0.82 0.95 0.39 Havana Brown 15 1 0 0.93 1.00 0.48 2 1 0.87 0.93 0.49 0 0 1.00 1.00 0.37 Japanese Bobtail 18 4 0 0.78 1.00 0.29 3 0 0.83 1.00 0.26 1 0 0.94 1.00 0.29 Korat 24 1 0 0.96 1.00 0.41 0 0 1.00 1.00 0.42 0 0 1.00 1.00 0.45 Maine Coon 27 5 8 0.81 0.73 0.44 1 13 0.96 0.67 0.44 6 5 0.78 0.81 0.35 Manx 22 8 11 0.64 0.56 0.33 5 9 0.77 0.65 0.40 4 12 0.82 0.60 0.14 Norwegian Forest 16 1 46 0.94 0.25 0.33 3 20 0.81 0.39 0.37 1 3 0.94 0.83 0.06 Ocicat 7 0 1 1.00 0.88 0.27 0 2 1.00 0.78 0.30 1 1 0.86 0.86 0.10 Persian 12 9 19 0.25 0.14 0.39 6 10 0.50 0.38 0.45 1 6 0.92 0.65 0.26 Ragdoll 16 3 0 0.81 1.00 0.26 2 0 0.88 1.00 0.26 4 0 0.75 1.00 0.32 Russian Blue 19 0 0 1.00 1.00 0.31 1 0 0.95 1.00 0.32 3 0 0.84 1.00 0.39

(continued)

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

146 320 Kurushima et al.

Table 4 (continued)

Intergenic SNPs Intergenic and phenotypic SNPs STRs

Ave. Ave. Ave.

Breed n EI EII Sens. Spec. Prob. EI EII Sens. Spec. Prob. EI EII Sens. Spec. Prob.

Scottish Fold 19 10 0 0.47 1.00 0.84 10 0 0.47 1.00 0.85 2 0 0.89 1.00 0.45 Siamese 19 1 0 0.95 1.00 0.33 0 0 1.00 1.00 0.32 0 0 1.00 1.00 0.17 Siberian 9 5 3 0.44 0.57 0.19 4 3 0.56 0.63 0.22 0 18 1.00 0.33 0.11 Singapura 19 1 0 0.95 1.00 0.45 0 0 1.00 1.00 0.44 3 0 0.84 1.00 0.32 Sokoke 5 0 0 1.00 1.00 0.41 0 0 1.00 1.00 0.42 0 0 1.00 1.00 0.46 Sphynx 25 3 2 0.88 0.92 0.32 3 2 0.88 0.92 0.31 1 0 0.96 1.00 0.25 Turkish Angora 18 10 1 0.44 0.89 0.23 6 8 0.67 0.60 0.37 9 7 0.50 0.56 0.21 Turkish Van 14 5 3 0.64 0.75 0.27 4 2 0.71 0.83 0.37 2 3 0.86 0.80 0.18

All Breeds 500 110 111 0.78 0.78 0.39 83 83 0.83 0.83 0.39 59 62 0.88 0.88 0.27 95% confidence 0.09 0.10 0.06 0.07 0.04 0.06 interval

Bayesian method of Rannala & Mountain (1997). Frequentist method of Paetkau et al. (1995). 1 Essentially zero due to lack of sensitivity; n, number of samples from this breed tested over 10 iterations; EI, members of a breed that were incorrectly

assigned to another breed; EII, members of a different breed that were incorrectly assigned to the breed in question; Sens., sensitivity; SNPs, single nucleotide polymorphisms; STRs, short tandem repeats; Spec., specificity; Ave. Prob., average probability of assignment as calculated by the Paetkau et al. (2004) Monte Carlo resampling method.

Table 5 Total misassigned individuals identified post-assignment by phenotypic SNPs.

Assigned by SNPs Assigned by STRs

Bayesian Frequentist Bayesian Frequentist

Total Freq. Total Freq. Total Freq. Total Freq.

Longhair 105 0.49 37 0.34 11 0.13 11 0.18 Burmese Points 15 0.07 3 0.03 1 0.02 2 0.03 Siamese Points 15 0.07 16 0.15 6 0.07 3 0.05 Chocolate 8 0.04 0 0 2 0.02 0 0 Cinnamon 14 0.07 5 0.05 4 0.05 4 0.07 Total1 127 0.58 55 0.50 22 0.26 19 0.32

Frequency (SNPs: Bayesian = 221, Frequentist = 110 STRs: Bayesian = 86, Frequentist = 59); SNPs, single nucleotide polymorphisms; STRs, short tandem repeats. 1A few individuals were identified as misassigned with multiple phenotypic SNPs.

of cats between some countries. Overlapping niches Discussion between the wildcat progenitors, random-bred feral cats, The artificial selection and population dynamics of domestic random-bred house cats and fancy breeds likely produces cats and their associated fancy breeds are unique amongst continual, however limited, horizontal gene flow through- domesticated species. Cats are one of the more recent out the domestic cat world. mammalian domesticates, arguably existing in a unique The overall selection on the cat genome may be predicted quasi-domesticated state. Although domestication is an to be less intense than in other domesticated species. The cat ongoing process, the earliest instance of cat taming is fancy is <150 years old, and a majority of cat breeds were credited to a Neolithic burial site on Cyprus dated to 9500– developed in the past 50–75 years. Human selection in cats 9200 years ago (Vigne et al. 2004). Unlike other agricul- has focused on aesthetic qualities, such as coat colours and tural species and the domestic dog, until recently, cats have fur types, as opposed to complex behaviours and qualities, had minimal artificial selection pressures on their form and such as hunting skills and meat or milk production in dog function as they have naturally performed their required or in other livestock species. Many of the cat’s phenotypic task of vermin control. Barriers to gene flow are mitigated attributes, even those that affect body and appendage as cats are transported between countries via both pur- morphologies, are traits with basic Mendelian inheritance poseful and accidental human-mediated travel, although patterns. One simple genetic change, such as the longhair of recently rabies control legislation has reduced the migration the Persian versus the shorthair of Exotic Shorthairs, is the

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

147 Variation of cats under domestication 321 defining characteristic between these two breeds. Burmese Two of the most prevalent breeds are Persians and Bengals and Siamese points are found in a large metafamily of (http://www.tica.org/). Persians were one of the first breeds breeds that includes Burmese, Siamese, Javanese, Tonkinese to be recognised, and Bengals, although only introduced in and Birman, to name a few (Table 1). Brown colorations are the past 40 years, have risen to worldwide fame. Both breeds diagnostic in breeds such as the Havana Brown (chocolate) had moderate levels of heterozygosity and inbreeding. and the Abyssinian (cinnamon). These selective pressures Several less popular breeds, such as the Cornish Rex, are reflected in the causative SNP frequencies in Table 3. contained fairly high levels of variation and low inbreeding, Cat registries have recognised that some breeds are whereas two recently developed breeds, the Siberian and ‘natural’, such as the Korat and Turkish Van. These breeds Ragdoll, revealed high variation, perhaps a reflection of their are specific population isolates, and random-bred cats of recent development from random-bred populations. Thus, similar origins can be used to augment their gene pools. levels of variation and inbreeding cannot entirely be Other breeds are recognised as ‘hybrids’, developed from predicted based on breed popularity and breed age, implying purposeful cross-breeding of either different breeds or management by the cat breeders may be the most significant species. One such example is the Ocicat, an intentional dynamic for breed genetic population health. Abyssinian and Siamese cross. The Bengal is a unique breed The Bayesian cluster analysis supported the breed that is an interspecies hybrid between an Asian leopard cat demarcations from previous studies, especially the STR and various domestic breeds (Johnson-Ory 1991). As a analyses of Lipinski et al. (2008). Previously, 22 breeds, result, some cat breeds may be a concoction of various which included 15 of 16 ‘foundation’ cat breeds designated genetic backgrounds, including cats of different breeds but by the Cat Fanciers’ Association, delineated as 17–18 having the same racial origins, cats of different breeds from separate populations. This study added seven additional different racial origins and even different species. breeds, including the missing 16th ‘foundation’ breed, the The 29 breeds were selected to represent the major breeds Manx. However, the most likely value of K (number of of the cat fancy. Some breeds may have developed from structured groupings) could not be decisively determined by natural populations, while most cat breeds developed in the methods developed for wild populations. As STRUCTURE past 50 years. Several breeds that had clearly derived from creates a probability distribution of the breed populations another breed, such as Persians and Exotic Shorthairs, were by inferring the previous generation’s genotypic frequencies purposely chosen, whereas others were selected because through the principles of HWE, several practices in cat they were recently developed hybrid breeds, such as the breeding result in genetic populations that do not always Ocicat, Bengal and Australian Mist. Thus, STRs may be align with the inferences of STRUCTURE. Cat breeds have better for breaking up breed families, whereas intergenic variation in age of establishment and significantly different SNPs may give us more insight into the natural popula- genetic population origins, and the dissimilarity in breeding tions. More slowly evolving SNPs and relatively quickly practices can create distinct lines within a single breed that evolving STRs were examined to assess their power to may be as unique as one of the more recently established resolve cat breeds that have different patterns, origins and breeds. Additionally, many breeds were created through the ages of ancestry. crossing of two, often highly divergent, populations of Significant genetic variation is present in many cat breeds cats resulting in a hybrid of sorts, whereas other breeds and cannot be predicted entirely by effective population size still allow the introduction of cats from random-bred (popularity amongst cat breeders) or breeding practices populations. These instances confounded the log-likelihood alone. The Turkish Angora, originating from Turkey, an calculations, making an empirical determination difficult. area near the seat of cat domestication (Driscoll et al. 2007; As in previous studies, the breeds that were not deemed Lipinski et al. 2008), had the highest effective number of genetically distinct can be explained by the breed history alleles for both SNPs and STRs. A wide distribution of (Lipinski et al. 2008; Menotti-Raymond et al. 2008). The two heterozygosity levels and inbreeding values was found large breed families of Siamese and Persian types were re- throughout the remainder of the cat breeds. However, the identified, and the Persian family expanded with Scottish SNPs and STRs were not always concordant (as can be seen Folds. The Australian Mist was added to the previously in Fig. S1). A previous study found STRs often underesti- recognised grouping of the Siamese/Havana Brown/Bur- mate FST compared to SNPs, most likely due to a rapid STR mese, as this breed was created by cross-breeding with mutation rate, often leading to convergence (Sacks & Louie Burmese. More recent breeds, such as the Ragdoll and Bengal, 2008). An alternative hypothesis is that long isolated breeds are resolved as separate breed populations, suggesting STRs of a large population size have had sufficient time and alone can differentiate about 24 of 29 breeds, in addition to opportunity to increase STR heteorzygosity through muta- Turkish- versus USA-originating Turkish Angoras. At tion, but not so for SNPs. Regardless, SNPs and STRs have K = 17, SNPs could separate Birman from other Asiatic differing relative observed heterozygosity values for some of breeds but not the Singapura. Thus, both sets of markers the breeds (namely Abyssinians, Persians and Japanese provide valuable insight into the relationship of the breeds.

Bobtails) and is reflected in their FIS values. Because the breeds within the larger family groups are

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

148 322 Kurushima et al.

generally different by only a single-gene trait, an actual breed population should excel in assignment accuracy for inbred designation may not be appropriate and perhaps should be populations. consider varieties within a breed. The cat fancy has prece- Many breeds are defined by one genetic trait in the cat dence for this concept, the pointed Persian, a Himalayan, is fancy. Although many breeds can share a trait, such as considered a variety in the CFA but a breed by TICA. longhair, this same trait can exclude a breed (Table 3). Thus, Regardless of the marker assayed, the principal coordi- phenotypic traits were tested post-assignment, as many are nate and Bayesian assignment analyses clustered the not highly breed selective pre-assignment. Although the 38 majority of breeds with the random-bred population that highly polymorphic STRs consistently outperformed the was most influential to its creation, as suggested by popular SNPs, the addition of phenotypic SNPs as post-assignment breed histories. Sixteen breeds originated from European verification significantly improved the assignment rates. The populations, six breeds from South Asian populations, two reduction in sensitivity and specificity when combing the breeds from the Eastern Mediterranean and the Sokoke from phenotypic SNPs in the assignment may be due to the the India or Arabian Sea region. However, some marker- strength of selection imposed on these markers. In general, specific differences were noted. When SNP and STR results breeds that were more inbred, not open to outcrosses and not were compared through Bayesian assignment, the Turkish developed through the crossing of pre-existing breeds, had a Angora was assigned to Europe or the Eastern Mediterra- higher accuracy in reassignment; the Russian Blue, Sokoke nean respectively, Bengal was assigned to Europe or the and Abyssinian are examples. In contrast, breeds where Arabian Sea respectively and the Ocicat was assigned to outcrossing is common, either with other breeds or random- South Asia or Europe respectively. These dissimilarities were bred populations, tended to confuse the assignment algo- not reflected in the PCA results that were remarkably rithm and had a high probability of both type I and II error, similar in both SNPs and STRs. This was most likely due to such as the Persians, Turkish Angoras and Ragdoll. The most offsetting the mutation rate differences with distance common error in assignment by far was cross-assignment matrices that accommodate these attributes. between Exotic Shorthairs and Persians within this breed Nonetheless, the aforementioned breeds have unique family, a problem easily remedied by exploiting the FGF5 SNP histories that may explain the marker discrepancies with causing longhair in Persians. Bayesian assignment to random-bred populations. The Initially, cats could be localised to a regional population Turkish Angora breed was reconstituted from the Persian and breed family by STRs and/or SNPs. Secondary differen- (European) pedigree post-World Wars, and their genetic tiation within the breed family could be determined by diversity has recently been supplemented via outcrossing to genotyping mutations for phenotypic traits, especially Turkish random-bred cats. The identified subpopulations traits that are specific to or fixed within a breed. Some traits within the breed may reflect the latest influx of random-bred are required for breed membership; a Birman or Siamese cats. The Bengal and the Ocicat clustering could be a result must be pointed, implying homozygosity for the of the contribution of breeds from very different regional AANG02171093.1(TYR):g.1802G>A variant. Some traits origins such as Abyssinian, Egyptian Mau and the Siamese. are grounds for exclusion: all Korats are solid blue, and no Overall, the frequentist method of Paetkau et al. (1995) other colours or patterns are acceptable. Therefore, a outperformed the Bayesian method of Rannala & Mountain trait such as the longhair AANG02027250.1(FGF5): (1997) in assigning unknown individuals to their breed of g.18442A>C variant could be used as a means for identifying origin. Both methods rely on a frequency distribution to members of the Persian, Maine Coon, Turkish Angora, estimate the probability that an unknown arose in a given Turkish Van and Birman breeds and, likewise, a means for population. The differences lie in how that frequency discrimination as an exclusion marker for breeds such as the distribution is established. Paetkau’s frequentist method Abyssinian, Egyptian Mau, Sokoke and Ocicat. Other single- generates the frequency distribution based on the observed gene traits may be used to identify members of a small family alleles in each population, whereas the Bayesian method of cat breeds as well, such as the Burmese points, begins with an initial distribution in which every population AANG02171092.1(TYR):g.11026G>T, which is a prerequi- in the data set has an equal allele density and then site for membership to the Burmese and Singapura breeds. calculates a posterior probability distribution based on the The cinnamon mutation, AANG02185848.1(TYRP1): initial assumption given the observed data. Both methods g.10736C> T, is very rare in the general cat population, yet assume the populations are in HWE; however, the frequ- is a defining characteristic of the red Abyssinian. entist method is able to accommodate populations with Cat fancy registries may not agree with assignments due drastically different allele frequencies – populations such as to variations in breeding practices between the registries for those seen as a result of the cat fancy. Directed breeding, a given breed. The Tonkinese, which is genetically a such as that used in the development of pedigreed cats, compound heterozygote for the AANG02171092.1(TYR): inherently violates the assumptions of HWE. Therefore, a g.11026G>T and the AANG02171093.1(TYR):g.1802G>A frequentist method that identifies an individual’s origin variants, can produce both pointed and sepia cats; thus, based on the frequency of the genotypes in each potential Tonkinese can genetically resemble a Siamese or Burmese

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

149 Variation of cats under domestication 323 respectively at the TYR locus. However, in some cases, References registration restrictions do not allow these Tonkinese Baudouin L. & Lebrun P. (2001) An Operational Bayesian variants to be registered as Siamese or Burmese. In addition, Approach for the Identification of Sexually Reproduced Cross- some breed registries allow colour and hair variants that Fertilized Populations using Molecular Markers. In: Proc. Int. may not be permitted in another, confusing possible breed Symp. on Molecular Markers, pp. 81–94. assignments. Thus, the cats assigned in this study are more Boitard S., Chevalet C., Mercat M.J., Meriaux J.C., Sanchez A., Tibau likely specific to the cat fancy of the United States, and tests J. & Sancristobal M. (2010) Genetic variability, structure and for other breed populations that are registry- or regional- assignment of Spanish and French pig populations based on a specific may need to be developed. Since the development of large sampling. Animal Genetics 41, 608–18. this SNP panel, additional phenotypic SNPs have been Driscoll C.A., Menotti-Raymond M., Roca A.L. et al. (2007) The discovered in cats including the Norwegian Forest Cat Near Eastern origin of cat domestication. Science 317, 519–23. colour variant amber (Peterschmitt et al. 2009), three Evanno G., Regnaut S. & Goudet J. (2005) Detecting the number of additional longhaired mutations (Kehler et al. 2007) and clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14, 2611–20. the mutations responsible for hairlessness in Sphynx and Felsenstein J. (1989) PHYLIP – phylogeny inference package (version rexing of the Devon Rex (Gandolfi et al. 2010). These 3.2). Cladistics 5, 164–6. additional mutations, as well as disease mutations, could Gandolfi B., Outerbridge C., Beresford L., Myers J., Pimentel M., further delineate cat breeds. Alhaddad H., Grahn J., Grahn R. & Lyons L. (2010) The naked Aside from the public interest in knowing whether their truth: Sphynx and Devon Rex cat breed mutations in KRT71. prized family pet is descended from a celebrated pedigree, Mammalian Genome 21, 509–15. breed assignment is a vital tool in tracing the spread of Gebhardt R.H. (1991) The Complete Cat Book. Howell Book House, genetically inherited diseases throughout the cat world. New York. Much like humans and dogs, certain populations of cats are Goudet J. (1995) FSTAT (Version 1.2): a computer program to known to be at higher risk for particular diseases, such as calculate F-statistics. Journal of Heredity 86, 485–6. heart disease in the Maine Coon and Ragdoll (Meurs et al. Jakobsson M. & Rosenberg N.A. (2007) CLUMPP: a cluster matching and permutation program for dealing with label switching and 2005, 2007), polycystic kidney disease in the Persian multimodality in analysis of population structure. Bioinformatics (Lyons et al. 2004) and progressive retinal atrophy in the 23, 1801–6. Abyssinian (Menotti-Raymond et al. 2007). Knowing Johnson-Ory G. (1991) Getting to Know the Bengal Cat. Gogees whether a particular feline descended from one of these Cattery, Greenwell Springs, LA. at-risk populations may influence treatments in a clinical Kehler J.S., David V.A., Schaffer A.A., Bajema K., Eizirik E., Ryugo setting and help to better care for our animal companions. D.K., Hannah S.S., O’Brien S.J. & Menotti-Raymond M. (2007) In addition, understanding the population structuring of the Four independent mutations in the feline fibroblast growth factor 5 cat breeds can be of assistance to case–control studies for gene determine the long-haired phenotype in domestic cats. genome-wide association studies. The current study defined Journal of Heredity 98, 555–66. 24 of 29 cat breeds and an additional three breeds using Kurushima J.D. (2011) Genetic Analysis of Domestication Patterns phenotypic SNPs. With additional phenotypic and perhaps in the Cat (Felis catus): Worldwide Population Structure, and Human-mediated Breeding Patterns Both Modern and Ancient. disease-causing SNPs, the power of this STR/SNP panel to PhD dissertation, In: Genetics, p. 148. University of California, accurately assign individuals to specific cat breeds, in Davis, ProQuest Dissertations and Theses. (Publication No. AAT particular those breeds that are defined expressively by 11271.) single-gene traits, would be greatly increased. Lipinski M.J., Amigues Y., Blasi M. et al. (2007) An international parentage and identification panel for the domestic cat (Felis catus). Animal Genetics 38, 371–7. Acknowledgements Lipinski M.J., Froenicke L., Baysac K.C. et al. (2008) The ascent of We would like to thank the technical assistance of the cat breeds: genetic evaluations of breeds and worldwide random- Veterinary Genetics Laboratory of the University of Califor- bred populations. Genomics 91,12–21. nia – Davis and the University of California – Davis Genome Lyons L.A., Biller D.S., Erdman C.A., Lipinski M.J., Young A.E., Roe Center and those who graciously supplied us with buccal B.A., Qin B.F. & Grahn R.A. (2004) Feline polycystic kidney disease mutation identified in PKD1. Journal of the American swabs from their pets. Funding for this study was supplied in Society of Nephrology 15, 2548–55. part by National Geographic Expedition Grant (EC0360-07), Lyons L.A., Foe I.T., Rah H.C. & Grahn R.A. (2005a) Chocolate – National Institutes of Health National Center for Research coated cats: TYRP1 mutations for brown color domestic cats. Resources (NCRR) grant R24 RR016094R24, now the Mammalian Genome 16, 356–66. Office of Research Infrastructure Programs (ORIP) grant Lyons L.A., Imes D.L., Rah H.C. & Grahn R.A. (2005b) Tyrosinase R24OD010928, the University of California – Davis, Center mutations associated with Siamese and Burmese patterns in the for Companion Animal Health, the Winn Feline Foundation, domestic cat (Felis catus). Animal Genetics 36, 119–26. and a gift from Illumina, Inc. (LAL), and the University of Menotti-Raymond M., David V.A., Schaffer A.A., Stephens R., Wells California – Davis Wildlife Health Fellowship (JDK). D., Kumar-Singh R., O’Brien S.J. & Narfstrom K. (2007) Mutation

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

150 324 Kurushima et al.

in CEP290 discovered for cat, model of human retinal degener- Rannala B. & Mountain J.L. (1997) Detecting immigration by using ation. Journal of Heredity 98, 211–20. multilocus genotypes. Proceedings of the National Academy of Menotti-Raymond M., David V.A., Pflueger S.M., Lindblad-Toh K., Sciences of the United States of America 94, 9197–201. Wade C.M., O’Brien S.J. & Johnson W.E. (2008) Patterns of Reynolds J., Weir B.S. & Cockerham C.C. (1983) Estimation of the molecular genetic variation among cat breeds. Genomics 91,1–11. co-ancestry coefficient: basis for a short-term genetic-distance. Meurs K.M., Sanchez X., David R.M. et al. (2005) A cardiac myosin Genetics 105, 767–79. binding protein C mutation in the Maine Coon cat with familial Sacks B.N. & Louie S. (2008) Using the dog genome to find single hypertrophic cardiomyopathy. Human Molecular Genetics 14, nucleotide polymorphisms in red foxes and other distantly related 3587–93. members of the . Molecular Ecology Resources 8,35–49. Meurs K.M., Norgard M.M., Ederer M.M., Hendrix K.P. & Kittleson Schelling C., Gaillard C. & Dolf G. (2005) Genetic variability of M.D. (2007) A substitution mutation in the myosin binding protein seven dog breeds based on microsatellite markers. Journal of C gene in Ragdoll hypertrophic cardiomyopathy. Genomics 90, Animal Breeding and Genetics 122,71–7. 261–4. The Royal Canin Encyclopedia. (2000) The Royan Canin Encyclope- Negrini R., Nicoloso L., Crepaldi P. et al. (2009) Assessing SNP dia. Groupe Royan Canin, Paris, France. markers for assigning individuals to cattle populations. Animal Vigne J.D., Guilaine J., Debue K., Haye L. & Gerard P. (2004) Early Genetics 40,18–26. taming of the cat in Cyprus. Science 304, 259. Nei M. (1972) Genetic distance between populations. The American Naturalist 106, 283–92. Paetkau D., Calvert W., Stirling I. & Strobeck C. (1995) Microsat- Supporting information ellite analysis of population structure in Canadian polar bears. Additional supporting information may be found in the – Molecular Ecology 4, 347 54. online version of this article. Paetkau D., Slade R., Burden M. & Estoup A. (2004) Genetic Figure S1 Observed heterozygosity by breed. assignment methods for the direct, real-time estimation of Figure S2 Log likelihood and ΔK plots from the Bayesian migration rate: a simulation-based exploration of accuracy and power. Molecular Ecology 13,55–65. clustering of cat breeds. Peakall R. & Smouse P.E. (2006) GENALEX 6: genetic analysis in Figure S3 (a) Alternate plots of short tandem repeat (STR) Excel. Population genetic software for teaching and research. Bayesian clustering analysis of cat breeds; (b) Alternate Molecular Ecology Notes 6, 288–95. plots of single nucleotide polymorphisms (SNP) Bayesian Penny Illustrated Paper. (1871) Crystal Palace – Summer concert clustering analysis of cat breeds. today, Cat Show on July 13. In: Penny Illustrated Paper, p. 16, Figure S4 (a) Crossed assignment rate between breeds as a London. function of the Reynold’s genetic distance between popula- Peterschmitt M., Grain F., Arnaud B., Deleage G. & Lambert V. tions using single nucleotide polymorphisms (SNPs); (b) (2009) Mutation in the melanocortin 1 receptor is associated Crossed assignment rate between breeds as a function of the with amber colour in the Norwegian Forest Cat. Animal Genetics Reynold’s genetic distance between populations using short 40, 547–52. tandem repeats (STRs). Piry S., Alapetite A., Cornuet J.M., Paetkau D., Baudouin L. & Estoup Table S1 F by locus for genetic markers and design and A. (2004) GENECLASS2: a software for genetic assignment and first- ST generation migrant detection. Journal of Heredity 95, 536–9. GenTrain score for single nucleotide polymorphisms (SNPs). Pontius J.U., Mullikin J.C., Smith D.R. et al. (2007) Initial sequence Table S2 Chi-squared test for Hardy-Weinberg equilibrium and comparative analysis of the cat genome. Genome Research 17, of SNPs and STRs by cat breed. 1675–89. Table S3 (a) SNP assignment of cat breeds to random bred Pritchard J.K., Stephens M. & Donnelly P. (2000) Inference of cat populations; (b) STR assignment of cat breeds to random population structure using multilocus genotype data. Genetics bred cat populations. 155, 945–59.

© 2012 The Authors, Animal Genetics © 2012 Stichting International Foundation for Animal Genetics, 44, 311–324

151 J. Anim. Breed. Genet. ISSN 0931-2668

ORIGINAL ARTICLE An insight into population structure and gene flow within pure-bred cats G. Leroy1,2, E. Vernet3, M.B. Pautet3 & X. Rognon1,2

1 UMR1313 Gen etique Animale et Biologie Integrative, AgroParisTech, Paris, France 2 UMR1313 Gen etique Animale et Biologie Integrative, INRA, Jouy-en-Josas, France 3 LOOF, Pantin, France

Keywords Summary Cat; gene flow; genetic variability; inbreeding; Investigation of genetic structure on the basis of pedigree information population structure. requires indicators adapted to the specific context of the populations stud- Correspondence ied. On the basis of pedigree-based estimates of diversity, we analysed G. Leroy, UMR1313 Gen etique Animale et genetic diversity, mating practices and gene flow among eight cat popula- Biologie Integrative, AgroParisTech, 16 rue tions raised in France, five of them being single breeds and three consist- Claude Bernard, F-75231 Paris 05, France. ing of breed groups with varieties that may interbreed. When computed Tel: +33 (0) 1 44 08 17 46; Fax: +33 (0) 1 44 08 on the basis of coancestry rate, effective population sizes ranged from 127 86 22; E-mail: [email protected] to 1406, while the contribution of founders from other breeds ranged from 0.7 to 16.4%. In the five breeds, FIS ranged between 0.96 and 1.83%, with Received: 29 October 2012; this result being related to mating practices such as close inbreeding (on accepted: 27 April 2013 average 5% of individuals being inbred within two generations). Within

the three groups of varieties studied, FIT ranged from 1.59 to 3%, while FST values were estimated between 0.04 and 0.91%, which was linked to various amounts of gene exchanges between subpopulations at the paren- tal level. The results indicate that cat breeds constitute populations sub- mitted to low selection intensity, contrasting with relatively high individual inbreeding level caused by close inbreeding practices.

and a suboptimal management of genetic variability, Background such as popular sire effect, may lead to a dissemina- Genealogies constitute a profitable source of informa- tion of inherited disorders and an erosion of genetic tion to investigate breeding practices, diversity or diversity. A subsequent increase in inbreeding may genetic structure in livestock, and companion and eventually lead to an increased incidence of some dis- captive animal populations. Based on Mendelian seg- orders (Leroy & Baumung 2011) and a negative regation rules, pedigree analysis can be used to follow impact on fitness traits (Boakes et al. 2007). These gene transmission from generation to generation and issues have been well studied in dogs using pedigree between subsamples of an entire population, which files, with investigations into breeding practices may be particularly useful for recently created animal (Leroy & Baumung 2011), the characterization of breeds. genetic diversity (Leroy et al. 2006; Calboli et al. 2008; Cat breeds may constitute an interesting example of Shariflou et al. 2011) or inbreeding effects (Maki€ et al. recent populations submitted to various gene flow. 2001). Cat breeds are, as well as dog populations, Indeed, a majority of modern cat breeds has been threatened by genetic disorders with more than 250 developed over the past 50 years, on the basis of sim- inherited disorders reported by [Online Mendelian ple phenotypical variants, with one or several former Inheritance in Animals (OMIA), omia.angis.org.au]. populations (Lipinski et al. 2008). In companion ani- Yet, pedigree investigations have been less frequently mals, it has been found that some breeding practices conducted within this species (Mucha et al. 2011).

© 2013 Blackwell Verlag GmbH • J. Anim. Breed. Genet. (2013) 1–8 doi:10.1111/jbg.12043

152 Population structure within purebred cats G. Leroy et al.

The aim of this study was to analyse the genetic The first group (PES) involved two varieties: Persian diversity of eight cat pure-bred populations raised in and Exotic Shorthair. The Persian is one of the most France on the basis of pedigree data. There are two common breeds in the world and until 2010 showed main purposes: (i) to assess the level of genetic vari- the largest number of births among breeds raised in ability within cat breeds in relation to inbreeding France (among PES kitties born in 2010, 4934 were evolution and specific breeding practices, and (ii) to registered including 4209 declared as Persian). Cross- investigate the recent gene flow explaining current breeding is allowed with its shorthaired variety, the constitution and structure of cat populations. Exotic Shorthair breed (725 registrations for 2010). PES is also the only population among those analysed with a decrease in number of births (13%) between Material and methods 2003 and 2010 (Figure 1, Table 1). The second group (BSH) involved five varieties (outcrossing being Populations studied allowed between these populations in France): British In France, breed genealogies are managed in a unified Shorthair (1492 births in 2010), its longhair pheno- genealogical database handled by the Livre Officiel typical variant (295 births), the Scottish variety (504 des Origines Francßaises (LOOF). Among the 66 breeds births), the Highland variety (Scottish longhair vari- and varieties registered in France, five breeds and ant, 133 births) and the Selkirk variety (140 births). three groups of breeds/varieties were chosen, showing Finally, the Abyssinian population (288 births in both relatively good pedigree knowledge and a varia- 2010) and its long haired Somali variety (139 births in tion in population size or geographical origin. The five 2010) were also analysed together (ABS). breeds are Maine Coon, Bengal, Birman, Chartreux For each breed or group of varieties, current gener- and Devon Rex. Maine Coon and Bengal breeds have ation was defined based on individuals registered experienced a large population increase over the last between 2008 and 2010 with both parents known. 8 years (Figure 1): births increasing from 1325 to 4470 and from 136 to 1148, respectively, between Methods 2003 and 2010. Birman and Chartreux breeds, the only two populations of French origin among those We computed the number of equivalent complete studied, have a relatively large number of births generations traced (EqG) and generation intervals as (4015 and 2085 registrations in 2010, respectively). described in Leroy et al. (2006). Identity-by-descent By contrast, the Devon Rex breed was considered (IBD) estimators, that is, coefficients of inbreeding F here as an example of a breed with a small population and coancestry C, were computed and averaged over size (only 191 births in 2010). the current generations. To characterize genetic struc- The three groups include nine populations, which ture within breeds and varieties, we computed fixa- can be considered as either breeds or varieties depend- tion index FIS using the following equation (Leroy & ing on countries and the breeding rules of the associa- tions. For more clarity, subpopulations among groups Table 1 Demographic parameters of the breeds studied will be considered here as varieties. Reference population (individuals registered with both parents known Evolution over 2008–2010) Breed or of births group of (2003– Nb of Nb of Nb of Nb of varieties 2010) % breeders individuals sires dams

Abyssinian/ +61 115 1 307 163 297 Somali (ABS) Bengal +781 241 2896 367 645 Birman +86 1076 11 109 1087 2352 British Shorthair/ +198 452 6758 809 1401 / Highland/Scottish/ Selkirk (BSH) Chartreux +6 514 6494 477 1052 Devon Rex +96 42 469 82 128 Maine Coon +225 690 11 642 1215 2178 Persian/Exotic 13 1300 14 921 2201 3812 Figure 1 Evolution of births according to breeds over the 2003–2008 Shorthair (PES) period.

2 © 2013 Blackwell Verlag GmbH • J. Anim. Breed. Genet. (2013) 1–8

153 G. Leroy et al. Population structure within purebred cats

Baumung 2011), 2003 and 2010, the number of breeders ranged between 87 (Devon Rex) and 2,428 (PES). Between F C F ¼ : 2008 and 2010, current generation sizes ranged from IS 1 C 469 (Devon Rex) to 14,921 (PES). On average, sires For each group (AbS, BSH and PES), we differenti- produced 8.8 kittens, ranging from 5.7 (Devon Rex) F~ C~ C ated and averaged within all varieties, and as to 13.6 (Chartreux), and dams produced 4.7 kittens, coancestry averaged over the entire group (Caballero ranging from 3.7 (Devon Rex) to 6.2 (Chartreux). We F & Toro 2002), in order to compute -statistics, using found on average 1 sire for 1.8 dams. All the breeds the following equations, show good pedigree knowledge (EqG = 7.2 on aver- F~ C~ C~ C F~ C age), the highest values being found for Birman (8.1) F ; F ; F ; IS ¼ ST ¼ IT ¼ and Chartreux (8.3), while generation intervals ran- 1 C~ 1 C 1 C ged from 2.2 (Bengal) to 3.3 years (PES). Between The effective population size was estimated on the 2003 and 2010, the number of kittens born and used ΔF basis of individual rates of inbreeding i and coan- as reproducers followed the same trend as the number ΔC et al. F cestry ij (Cervantes 2011), considering i is the of births (Figure S1): on average overall breeds, 17% i inbreeding coefficient of individual ,Cij the coances- of kitties (including 5% of male and 12% of female) i j EqG try coefficient between individuals and , and i later became reproducers, the proportion ranging EqG and j their respective equivalent complete traced from 11% (Chartreux) to 22% (Bengal). generations: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi EqG 1 DFi ¼ 1 i ð1 FiÞ and qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Diversity indicators ðEqGiþEqGjÞ=2 DCij ¼ 1 ð1 CijÞ: Table 2 shows the IBD estimators for the eight breeds and groups of breeds/varieties. According to the breed, Effective population sizes were estimated by averag- F ranged from 2% (Maine Coon) to 4.4% (Chartreux). ing ΔFi over the current generation and ΔCij over These inbreeding levels could be explained by pedigree 100 000 pairs of individuals randomly sampled among knowledge, population size and also by mating between the current generation, using the following formulas: close relatives (close inbreeding): the proportion of inbred individuals, based on two and three generations, NeFi ¼ 1=2DF and NeCi ¼ 1=2DC: ranged from 2.7 (Maine Coon) to 8.4% (Devon Rex) Percentages of inbred individuals were computed and from 7.7 (Maine Coon) to 22.5% (Devon Rex), using the Van Raden (1992) method taking into respectively. For each breed considered, there was a account only two and three generations. The evolu- large increase in inbreeding coefficients during the first tion of average inbreeding coefficient according to the generations (Figure 2). Over the following generations, number of generations considered was also estimated inbreeding increase was smoother and more regular, for the current population. indicating no strong bottleneck event. On the basis of the breed origin of each founder Average coancestry C was always lower than F (ancestors of the current generation without parents inbreeding, which is illustrated by a positive IS value known), contribution of different breeds to each gene for all the breeds, ranging between 0.96% (Maine pool was computed for the different breeds and Coon) and 1.83% (Birman). In BSH and PES groups, groups of breeds, considering either founders or average C were lower than 1%, while the maximum parental origins, that is, origins of parents of individu- value was for the Chartreux breed (approximately als of the reference generations. The analyses 2.8%). Therefore, when using C instead of F to com- N were performed using the PEDIG software (http:// pute e, effective population size increased largely, dga.jouy.inra.fr/sgqa/article.php3?id_article=110,Boi- with NeCi and NeFi ranging between 127 (Devon Rex) chard 2002). and 1406 (PES) and between 64 (Devon Rex) and 161 (Maine Coon), respectively.

Results Gene flow within and between breeds and groups of Demographic and genealogical parameters varieties The eight breeds and groups of breeds studied showed By breed, the proportion of founders originating from a wide range of situations, regarding population size outside the breed was variable: based on founder or numbers of breeders (Figure 1, Table 1). Between approaches, the contribution of external origins ran-

© 2013 Blackwell Verlag GmbH • J. Anim. Breed. Genet. (2013) 1–8 3

154 Population structure within purebred cats G. Leroy et al.

Table 2 Genealogical parameters of the breeds considering current generation (2008–2010)

Inbreeding Coancestry

% of individuals inbred after Main founder

Breed or group Nb of 2 3 Overall FIS Out origins outside of varieties individuals EqG T F (%) NeFi generations generations generations C (%) NeCi (%) (%) the breed (%)

Abyssinian/Somali 1307 6.31 2.77 2.71 95 6 13.1 67.6 1.14 266 a 0.7 Unknown (0.5) (ABS) Bengal 2896 6.68 2.19 2.83 97 3.6 11 95.2 1.78 182 1.07 5.7 American Shorthair (2.2) Birman 11 109 8.07 3.18 2.93 115 4 13.5 95.1 1.12 365 1.83 5.3 Balinais (2.7) British Shorthair/ 6758 6.73 2.69 2.6 105 5.2 13.9 80.7 0.61 553 a 16 Persian (10.5) British Longhair/ Highland/ Scottish/ Selkirk (BSH) Chartreux 6494 8.29 2.9 4.41 78 4.5 13.5 98.5 2.78 146 1.68 11.5 British Shorthair (6.1) Devon Rex 469 6.47 2.44 4.29 64 5.1 22.5 65.1 2.54 127 1.79 5.9 Burmese (2.6) Maine Coon 11 642 7.27 2.41 1.98 161 2.7 7.7 91.3 1.03 363 0.96 1.2 Persian (0.9) Persian/Exotic 14 921 7.39 3.28 3.25 91 8.4 18 88.7 0.26 1406 a 1.7 British Shorthair (PES) Shorthair (1)

EqG, number of equivalent generations; T, generation intervals in years; F, average inbreeding coefficient; NeFi, inbreeding effective population sizes;

C, average coancestry coefficient; NeCi, coancestry effective population sizes; FIS, breed fixation index; Out, % of founder origins outside the breed. aSee Table 3.

in BSH, external origins contributed largely to the gene pool. Within this group, the British Shorthair constituted the largest population (64% of the total population group), and it was also the main origin for the different varieties of the group. Thus, founder ori- gins from the British Shorthair ranged from 59.6% (Selkirk) to 83.1% (British Shorthair) (Table S1). The contribution of the British Shorthair variety remained relatively important even considering its parental ori- gins, with its contribution ranging between 22.6% (Highland) and 91% (British Shorthair) (Figure 3). Within the group, all varieties but one (Selkirk) were involved as contributors of other ones. In the PES Figure 2 Evolution of average inbreeding coefficients for the current group, the Persian constituted by far the main origin, population according to the number of generations considered. All: all contributing to 86.6% of founder origins of the Exotic generations considered. Shorthair variety. However, considering the last gen- ged from 0.7 (ABS) to 16% (BSH) of the gene pool eration, 67.3% of parental origin in Exotic Shorthair (Table 2). Most of the time, those external origins belonged to the Exotic Shorthair. The two varieties of were mainly related to one breed: for instance, the the ABS group constituted more independent subpop- British Shorthair contributed up to 50% or more of ulations, with most of the founder and parental con- the external origins for two breeds (namely Chartreux tributions in the Abyssinian and the Somali coming and PES). However, considering the parental origins from Abyssinians and Somalis, respectively. of the current generation, gene flow was much more According to Table 3, within each of the three limited, and in each population, <1% of those origins groups, the average coancestry was relatively low belonged to external breeds (Figure 3). between each variety, ranging from 0.19 (Abyssinian As illustrated by Figure 3, the three groups of and Somali) to 0.67% (Scottish and Highland). As varieties show contrasting situations with regard to expected, the contrast between inbreeding and gene flow among subpopulations. As aforementioned, coancestry was lower when considering each variety

4 © 2013 Blackwell Verlag GmbH • J. Anim. Breed. Genet. (2013) 1–8

155 G. Leroy et al. Population structure within purebred cats

British Shorthair 91.0 (4342)

54.3 5.9 44.7 2.5

British 41.1 Longhair 0.6 44.7 Scottish 2.1 (1008) (768) 3.1 2.4 0.1 7.9

17.3 3.7 0.1 22.6 22.8 26.8 Selkirk Highland (383) (255)

73.4 BSH group 33.3 British Longhair 94.8 67.3 British Shorthair

Scottish Persian Exotic (12898) Shorthair 32.5 (2023) Highland 5.2 Selkirk 0.1 PES group 0.3 Persian

96.4 95.6 Exotic Shorthair

Abyssinian Figure 3 Founder contributions and parental Abys- Somali 4.4 Somali origins for BSH, PES and ABS groups. Circles sinian (412) (895) 3.6 indicate repartition of founder contribution Origins outside the according to the probability of gene origins, AbS group group while arrows represent parental origins (values in%). Sizes of arrows and circles are propor- tional to contributions and population size Within breed/ Contribution from another Origins outside (current generation size in parenthesis). variety contribution variety of the group the group independently than when considering groups of others, F-statistics adapted to pedigree analysis. Cat breeds. Indeed FIS values (1.76, 2.96 and 0.69% for breeds have rarely been investigated in the past, and the BSH, PES, and ABS, respectively, Table S2) were only study based on pedigree analysis (Mucha et al. lower than FIT (2, 3 and 1.59%, respectively), FST val- 2011) showed average inbreeding around 3%, conclud- ues being contrasted according to groups (0.24, 0.04 ing that cat populations are not threatened by negative and 0.91%, respectively). Yet we noticed that for Brit- effects of inbreeding. Considering coancestry as the ish Shorthair and Exotic Shorthair varieties, FIS was parameter to minimize for conservational purpose slightly higher (2.2 and 3.1%, respectively) than (Baumung & Solkner€ 2003), the breeds studied here when considering FIT for BSH and PES groups (2 and also show remarkably high levels of diversity, effective 3%, respectively), indicating the existence of a sub- population sizes computed based on coancestry (NeCi) structure remaining among those varieties. ranging between 127 and 1406. The average coancestry was indeed found to be quite low in comparison with dog breeds. As an illustration, for the eight breeds or Discussion groups of varieties, average coancestries ranged The aim of this study was to assess genetic diversity and between 0.3 and 2.8% (1.4% on average), with current gene flow within and between cat breeds, using, among generation sizes ranging from 469 to 14 921 (6949 on

© 2013 Blackwell Verlag GmbH • J. Anim. Breed. Genet. (2013) 1–8 5

156 Population structure within purebred cats G. Leroy et al.

Table 3 Average coefficients of inbreeding F C (%) within and between varieties and coancestry C within and between varieties British British of BSH, PES and ABS groups Group of varieties F (%) FIS (%) Longhair Shorthair Scottish Highland Selkirk

BSH British Longhair 2.08 1.24 0.85 0.57 0.5 0.68 0.25 British Shorthair 2.9 2.18 0.74 0.56 0.52 0.28 Scottish 2.12 1.30 0.83 0.67 0.21 Highland 2.18 0.82 1.37 0.22 Selkirk 1.74 0.20 1.94

Persian Exotic Shorthair

PES Persian 3.23 2.72 0.27 0.21 Exotic Shorthair 3.4 3.14 0.52

Abyssinian Somali

ABS Abyssinian 2.44 0.44 1.65 0.19 Somali 3.3 0.02 2.88

FIS, within variety fixation index. average), with the average EqG around 7.2. By compari- trated through F-statistics variations, can be explained son, in 24 dog breeds with an EqG larger than 6 (7.1 on by three non-exclusive phenomena: intentional mat- average), Leroy et al. (2009) found average coancestries ing between close relatives (close and line breeding), to be twice as high (2.8%, ranging between 0.6 and existence of subpopulations (Wahlund effect) and low 8.8%), with average current generation sizes approxi- effective population size. mately 8 times larger (54 645, ranging from 2167 to First, there is a tendency among breeders to plan mat- 156 492). This difference is probably related to the low ing between closely related cats. On average, approxi- number of offspring per reproducer in cat breeds. In this mately 5% of kittens were inbred after two generations, study, during a generation interval (around 3 years), meaning their parents were sharing at least one parent. sires and dams produced on average 8.8 and 4.7 kittens, According to an analysis of dog breeds and simulated respectively. By comparison, in dog breeds (Leroy & populations, an increase of approximately 0.7–1% of FIS Baumung 2011), sires and dams produced 16.4 and 8.3 could be expected for such a proportion of mating puppies, respectively, during a generation interval between half- and full-sibs (Leroy & Baumung 2011). (4 years). The average number of offspring produced This result was in agreement with the large inbreeding per breeder was also on average smaller in cats (12.6 increase observed considering the first generations, rel- estimated from Table 1) than in dogs (18, see Leroy ative to the following ones (Figure 2). et al. 2009). Therefore, in comparison with dog breed- Secondly, positive FIT values could also be explained ers, a large majority of cat breeders are occasional ones. by the existence of more or less differentiated subpopu- These breeders used their reproducers with low inten- lations within breeds or groups. Two of the three sity, the females producing on average one litter during groups of varieties (BSH, PES) show relatively high FIT the 2008–2010 period (litter size being found on aver- values, which could, at first sight, be explained through age around 3.4, data not shown). This has a clear posi- preferential mating within varieties. As illustrated by tive impact on genetic diversity, but does not mean that Table 3, between-subpopulation coancestry was regular bottlenecks do not occur within breeds, which always lower than within-subpopulation coancestry. may lead to the dissemination of inherited diseases However, Figure 3 shows that gene exchanges were (Wellmann & Pfeiffer 2009). relatively frequent among varieties of BSH and PES In comparison with coancestries (1.4% on average), groups, while in the AbS, only a small proportion of average inbreeding values were high (3.1% on aver- parents originated from the other variety. This was in age), leading to an underestimation of effective popu- agreement with the very low FST values estimated for lation sizes when using inbreeding instead of the BSH and the PES (0.24 and 0.04%) in comparison coancestry (Table 2). These differences, indicating with the ABS (0.91%) where the level of genetic differ- deviations from random-mating conditions and illus- entiation between Abyssinians and Somalis was larger.

6 © 2013 Blackwell Verlag GmbH • J. Anim. Breed. Genet. (2013) 1–8

157 G. Leroy et al. Population structure within purebred cats

Finally, FIT and FIS variations could also be explained the creation of the breeds. Several explanations can by the effective population size of breeds, a limited pop- be given for such gene flow. For instance, in the Char- ulation size decreasing fixation index and eventually treux breed, the large amount of British Shorthair leading to negative values. This phenomenon can be contribution (6.1%) is probably due to regular regis- interpreted considering the evolution of IBD estima- trations of blue British Shorthair individuals within tors. Indeed, in panmixia, inbreeding and coancestry the breed (breeders’ personal communication). In the are supposed to differ only by ΔIBD (i.e. 1/(2Ne)), the BSH group, breeders in the past have probably used average coancestry between reproducers corresponding Persian reproducers to improve the quality of their to the average inbreeding of the next generation. This coat, explaining their large contribution (10.5%) as is why, for a given generation and in random-mating founders. Today, on the basis of the pedigree file, such conditions at least, we should expect C to be larger than cross-breeding events rarely occur in France, where

F.Therefore,FIT (FIS respectively) should tend to they are only exceptionally allowed by LOOF, but decrease in breeds (varieties respectively) with a small may exist in other countries depending on different effective population size. It may explain the low aver- breeding rules. Development of DNA identification age FIS (0.69%) in AbS (related to the small population will help to monitor the occurrence of false parentage size of Abyssinian and Somali varieties) and therefore among cat breeds, as well as the level of introgression the moderate FIT (1.59%) within the group, despite the of unofficial outcrossings. Studies based on molecular large subpopulation differentiation FST index (0.91%). markers may also bring further information on breed A small effective population size also explains why the relationships. For instance, Lipinski et al. (2008) and Devon Rex breed, despite the largest proportion of indi- Kurushima et al. (2012) seem to confirm introgression viduals inbred after three generations (22.5%), showed of Persian individuals into British Shorthair popula- only a moderate FIS value (1.79%). tions, as well as British Shorthair individuals into the In a large group like the PES (considering popula- Chartreux breed. Using molecular markers, Menotti- tion size), the high FIT (3.0%) value estimated was Raymond et al. (2008) were not able to differentiate finally less due to the subpopulation differentiation Exotic/Persian, Abyssinian/Somali and British Short- ðFST ¼ 0:04Þ than to close inbreeding practices. hair/Scottish varieties. According to the same study, Indeed, 8.4 and 18% of individuals were found inbred Selkirk was, however, found to be different from Brit- after two and three generations, respectively, explain- ish Shorthair and Scottish varieties, in contradiction ing the large FIS value (2.96%). By contrast, in the with our results, given the amount of gene flow Maine Coon breed, where the smallest proportion of observed from British Shorthair to Selkirk varieties. individuals inbred after two and three generations Such discrepancies could eventually be explained by was found (2.7 and 7.7%, respectively), one of the the breeding rules existing in the USA, where only lowest FIS was also computed (0.96%). These different Persian and Exotic Shorthairs are permitted for cross- examples illustrate quite well how the fixation index breeding with Selkirk individuals. can be influenced by the breeding practice and the From a practical point of view, the NeCi values, found demographic situation of domestic populations. larger than 100 for each of the breeds studied, indicate A comparison between founder and parental origins that those populations are probably submitted to a lim- illustrates the variation in gene flow over time. When ited genetic drift. By contrast, the large inbreeding val- considering the parental origins, only a low amount of ues, connected to lower NeFi, may increase the outcrossing was detected within each breed (implying proportion of individuals affected by monogenic reces- <1% of parents). Based on these results, we can con- sive genetic disorders, in relation to their allele fre- sider each of the eight populations studied as almost quency (Leroy & Baumung 2011). Some measures closed, which justifies the grouping choices we made. should therefore be recommended to limit close However, the founder approach results highlight that inbreeding practices, at least for breeds with NeFi lower crossbreeding events have occurred in the past, with than 100, and particularly for Devon Rex, where 22.5% more or less important effect on genetic diversity, of individuals were inbred after three generations. depending on the breed studied. The French unified genealogical database was set Conclusions up in 2000, with founder individuals born during the 1980–2000 period. At this time, each of the breeds To conclude, we can state that cat breeds constitute and varieties studied were already recognized, which populations submitted to relatively low selection underlines the fact that external contributions are intensity, with various levels of genetic structure, mainly related to recurrent cross-breeding events after according to breeding practices and/or the existence

© 2013 Blackwell Verlag GmbH • J. Anim. Breed. Genet. (2013) 1–8 7

158 Population structure within purebred cats G. Leroy et al. of varieties, involving more or less important gene cats under domestication: genetic assignment of domestic flow within a given population. If at the population cats to breeds and worldwide random-bred populations. level, genetic drift is expected to be limited, high indi- Anim. Genet., 44(3), 311–324. vidual inbreeding level found by contrast led us to Leroy G., Baumung R. (2011) Mating practices and the dis- recommend that particular attention should be paid semination of genetic disorders in domestic animals, based to population structure and inbreeding practices. on the example of dog breeding. Anim. Genet., 42,66–74. Each of the breeds studied has been submitted to Leroy G., Rognon X., Varlet A., Joffrin C., Verrier E. cross-breeding events in the last 30 years, with different (2006) Genetic variability in French dog breeds assessed J. Anim. Breed. Genet. – impacts on breed genetic diversity. Yet, the eight popu- by pedigree data. , 123,1 9. lations studied are currently almost closed to foreign Leroy G., Verrier E., Meriaux J.C., Rognon X. (2009) Genetic diversity of dog breeds: within-breed diversity influence, with, however, regular gene flow remaining comparing genealogical and molecular data. Anim. among varieties. Studies like this one may provide use- Genet., 40, 323–332. ful information to define current population subdivi- Lipinski M.J., Froenicke L., Baysac K.C., Billings N.C., Leut- sions more clearly. They also give insight into former negger C.M., Levy A.M., Longeri M., Niini T., Ozpinar H., gene flow, which could be useful for gene association Slater M., Pedersen N.C., Lyons L.A. (2008) The ascent of et al. studies (Quignon 2007) or when considering cat breeds: genetic evaluations of breeds and worldwide authorization of new cross breed events. Cross-breeding random-bred populations. Genomics, 91,12–21. may constitute an interesting option for introducing Maki€ K., Groen A.F., Liinamo A.E., Ojala M. (2001) Popula- genetic diversity within a given breed and/or improving tion structure, inbreeding trend and their association with it, especially in relation to its health status. Further stud- hip and elbow dysplasia in dogs. Anim. Sci., 73,217–228. ies could consider more widely the potential impacts of Menotti-Raymond M., David V.A., Pfluegger S.M., Lind- those breeding practices (close breeding, line breeding blad Toh K., Wade C.M., O’Brien S., Johnson W.E. and outcrossing) on animal welfare and health. (2008) Patterns of molecular genetic variation among cat breeds. Genomics, 91,1–11. Mucha S., Wolc A., Gradowska A., Szwaczkowski T. Acknowledgements (2011) Inbreeding rate and genetic structure of cat popu- The authors would like to thank Emily Heppner and lations in Poland. J. Appl. Genet., 52, 101–110. Wendy Brand Williams for linguistic revision. Quignon P., Herbin L., Cadieu E., Kirkness E.F., Hedan B., Mosher D.S., Galibert F., Andre C., Ostrander E.A., Hitte C. (2007) Canine Population Structure: assessment and References Impact of Intra-Breed Stratification on SNP-Based Asso- PLoS One Baumung R., Solkner€ J. (2003) Pedigree and marker infor- ciation Studies. , 2, e1324. mation requirements to monitor genetic variability. Shariflou M.R., James J.W., Nicholas F.W., Wade C.M. Genet. Sel. Evol., 35, 369–383. (2011) A genealogical survey of Australian registered Vet. J. – Boakes E.H., Wang J., Amos W. (2007) An investigation of dog breeds. , 189, 203 210. inbreeding depression and purging in captive pedigreed Van Raden P.M. (1992) Accounting for inbreeding and populations. Heredity, 98, 172–182. crossbreeding in genetic evaluation of large populations. J. Dairy Sci. – Boichard D. PEDIG: A Fortran Package for Pedigree Analy- , 75, 305 313. sis Suited for Large Populations. In: 7th World Congress Wellmann R., Pfeiffer I. (2009) Pedigree analysis for con- Genet. Res. of Genetics Applied to Livestock Production. Montpel- servation of genetic diversity and purging. , – lier, 19–23 August 2002. 91, 209 219. Caballero A., Toro M.A. (2002) Analysis of genetic diver- sity for the management of conserved subdivided popu- Supporting Information lations. Conserv. Genet., 3, 289–299. Calboli F.C.F., Sampson J., Fretwell N., Balding D.J. Additional Supporting Information may be found in (2008) Population structure and inbreeding from pedi- the online version of this article: gree analysis of purebred dogs. Genetics, 179, 593–601. Figure S1 Evolution of number of reproducers Cervantes I., Goyache F., Molina A., Valera M., Gutierrez born and used according to breeds over the 2003– J.P. (2011) Estimation of effective population size from 2008 period. the rate of coancestry in pedigreed populations. J. Anim. Table S1 Origins of founders among varieties of Breed. Genet., 128,56–63. BSH, PES and ABS groups. Kurushima J.D., Lipinski M.J., Gandolfi B., Froenicke L., Table S2 Fixation index for BSH, PES and AbS Grahn J.C., Grahn R.A., Lyons L.A. (2013) Variation of groups.

8 © 2013 Blackwell Verlag GmbH • J. Anim. Breed. Genet. (2013) 1–8

159 The Veterinary Journal 194 (2012) 343–348

Contents lists available at SciVerse ScienceDirect

The Veterinary Journal

journal homepage: www.elsevier.com/locate/tvjl

Assessing the impact of breeding strategies on inherited disorders and genetic diversity in dogs ⇑ Grégoire Leroy , Xavier Rognon

AgroParisTech, UMR 1313, Génétique Animale et Biologie Intégrative, F-75231 Paris, France INRA, UMR 1313, Génétique Animale et Biologie Intégrative, F-78352 Jouy-en-Josas, France article info abstract

Article history: In the context of management of genetic diversity and control of genetic disorders within dog breeds, a Accepted 18 June 2012 method is proposed for assessing the impact of different breeding strategies that takes into account the genealogical information specific to a given breed. Two types of strategies were investigated: (1) eradi- cation of an identified monogenic recessive disorder, taking into account three different mating limita- Keywords: tions and various initial allele frequencies; and (2) control of the population sire effect by limiting the Canine number of offspring per reproducer. The method was tested on four dog breeds: Braque Saint Germain, Pedigree analysis Berger des Pyrénées, Coton de Tulear and Epagneul Breton. Breeding policies, such as the removal of all Genetic diversity carriers from the reproduction pool, may have a range of effects on genetic diversity, depending on the Popular sire effect Inherited disorders breed and the frequency of deleterious alleles. Limiting the number of offspring per reproducer may also have a positive impact on genetic diversity. Ó 2012 Elsevier Ltd. All rights reserved.

Introduction The Federation Cynologique Internationale (FCI) recommends that the number of offspring per dog should not be >5% of the num- Management of inherited diseases and genetic diversity in dif- ber of puppies registered in the breed population during a 5 year ferent breeds of dogs is a growing concern for breeders, owners period.2 In parallel, about 20% of disorders/traits reported in OMIA and the general public (Nicholas, 2011). According to Online Men- have been characterised at the molecular level (Nicholas et al., delian Inheritance in Animals (OMIA)1 more than 575 disorders/ 2011). However, even when a genetic test is available, members of traits have been reported in dogs and at least 200 have monogenic breed societies often do not know which is the best strategy to adopt determinism (Nicholas et al., 2011). The prevalence of a genetic dis- to reduce the prevalence of genetic disorders. This is especially order can be >50% within a given population (Collins et al., 2011) and important when considering the use of valuable stud animals that the consequences for canine health may vary substantially, depend- may be disease carriers. There is also a need for members of breed ing on the severity of the disorder and its frequency. societies to be aware of the impact of different policies on genetic Increases in inbreeding and widespread dissemination of genet- diversity. Windig et al. (2004) modelled the consequences of a policy ic disorders may have a deleterious impact on welfare of purebred for eradication of genetic disorders in sheep using simulated popula- dogs, as shown with hip dysplasia in German shepherd dogs and tions. There is a need to extend such studies to take into account the Golden retrievers (Mäki et al., 2001) or fertility in Irish wolfhounds level of complexity existing in real breeds, including non-random (Urfer, 2009). Founder effects and extensive use of popular sires are mating, importations and bottleneck events. considered to be the main reasons for the dissemination of genetic In this paper, we propose a method to assess the impact of disorders and are linked to a reduction in genetic diversity within a breeding strategies on the frequency of deleterious alleles and breed (Leroy and Baumung, 2011). It has been suggested that the genetic diversity, taking into consideration the genealogical infor- prevalence of genetic diseases could be reduced through careful mation available for a given breed. Two strategies were investi- selection and better management of genetic drift and inbreeding gated: (1) eradication of an identified monogenic recessive (Lewis et al., 2010). disorder using three different mating limitations and various initial allele frequencies; and (2) control of the popular sire effect through limitation of the number of offspring per reproducer.

⇑ Corresponding author. Tel.: +33 144081746. E-mail address: [email protected] (G. Leroy). 1 See: omia.angis.org.au/. 2 See: http://www.fci.be/uploaded_files/29-2010-annex-en.pdf.

1090-0233/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tvjl.2012.06.025 160 344 G. Leroy, X. Rognon / The Veterinary Journal 194 (2012) 343–348

Materials and methods mating restriction was applied from the first year of the programme. Three different thresholds for the number of offspring were considered for all breeds: 50 (ps50), To investigate the evolution of the frequency of a deleterious allele, we consid- 100 (ps100) and 200 (ps200). A limitation of 25 offspring per reproducer (ps25) ered a single gene with two alleles (A and a), homozygous individuals aa being re- was also considered for BSG and BRP, but could not be applied to COT or EPB, since garded as affected by the genetic defect. The initial allele frequencies of a were set sires of these breeds produce, on average, a number of offspring close to or >25 at 20% and 50%, respectively. Carriers were randomly distributed among founders (Table 1). (i.e. individuals without known parents) of a given pedigree, with alleles being In the two scenarios, we supposed a random replacement of reproducers. To transmitted according to Mendelian segregation rules. It was assumed that there test what would happen if new sires or dams were more or less related to those re- was no selection of the allele before the beginning of the breeding strategy. placed, we studied the possibility that, among the simulations, 50% of the replace- ment sires (or dams) were sampled among the 10th percentile of the most (or least) Breeds selected for analysis related sire (or dam) of individuals born in the same year. This procedure was tested for one breed (COT) considering two of the sub-scenarios (ps50 and erC for an initial Four French breeds of dogs with different population sizes were selected for frequency of 50%). Each scenario was programmed in Fortran 90, repeated and aver- analysis: Braque Saint Germain (BSG), Berger des Pyrénées (BRP), Coton de Tulear aged over 100 iterations (see Appendix A: Supplementary file 1). (COT) and Epagneul Breton (EPB) (Table 1). The numbers of dogs registered in France for each breed from 2006 to 2010 ranged from 283 (BSG) to 27,326 (EPB). Generation intervals (T) were computed for each breed for dogs born from 2001 Results to 2010. The number of equivalent complete generations (EqG), inbreeding coefficient (F) The four breeds had a high level of pedigree completeness, EqG and kinship coefficient (U; also known as ‘co-ancestry’, which corresponds to the values for the period 2006–2010 ranging from 6.98 (COT) to 9.33 degree of inbreeding of a potential offspring of a pair of individuals) were averaged for the 2006–2010 period (Leroy et al., 2006). Kinship was averaged over 10,000 (EPB) (Table 1). In the same period, F ranged from 0.056 (EPB) to pairs of dogs born during a given period. When considering simulated sub-scenar- 0.091 (BRP) and U ranged from 0.036 (EPB) to 0.103 (BSG). As illus- ios, kinship was averaged over 100 pairs sampled over 100 iterations. The evolution trated in Fig. 1, there was a global increase in kinship for each of genetic diversity was assessed considering the evolution of yearly average U. For breed over the whole period. each scenario, kinship rate was computed per generation DU using the formula

DUt =(Ut+1 À Ut)/(1 À Ut), considering Ut and Ut+1 as average kinship in 2000 (year before implementation of the breeding strategy) and 2010 (end of the period inves- Eradication of recessive disorder tigated) and correcting it by period considered and generation intervals.

Simulation process: ‘what if’ As illustrated in Fig. 2, the three breeding strategies had differ- ent impacts on the frequency of the deleterious allele. Removing all Given the genealogical file of a breed, the ‘what if’ simulation process investi- carriers from reproduction (sub-scenario erC) directly decreased gated ‘what’ would have happened ‘if’ a given breeding strategy had been applied the frequency to a value close to 0, whatever the initial frequency. over a 10 year period (2001–2010). Evolution of genetic diversity and allele fre- quencies were compared between the original and the modified pedigree files. Ped- Due to importations of some dogs (without known parents and igrees were modified using the rule that, for a litter born during the 2001–2010 considered here as founders), allele frequency was not exactly period, if its sire (or dam) was affected by the mating restriction corresponding equal to 0 during the period. When heterozygotes were allowed to the breeding policy (see below), the parent was replaced by the sire (or dam) to reproduce (sub-scenario erA), the consequences were limited; of another dog randomly sampled from dogs born in the same year and not affected for COT, the allele frequency decreased over 10 years to 22% when by the mating restriction. If all potential parents were affected by the mating restriction, then the sampling was made among dogs born in the preceding year. the initial frequency was 50% and to 13% when the initial fre- Mating restrictions were modelled according to two different breeding scenarios: quency was 20%. When heterozygote offspring of carriers were not allowed to Scenario ‘er’ reproduce (sub-scenario erI), the decrease in allele frequency was In this scenario, we analysed strategies aiming to eradicate a monogenic reces- amplified and reached values close to 0 after 10 years. These re- sive disorder, assuming that carriers may be identified early (e.g. through a genetic sults were similar for all four breeds (see Appendix A: Supplemen- test). We compared three sub-scenarios of breeding strategies with an increasing severity of selection against the disorder. For each sub-scenario, the two initial al- tary Fig. 1). lele frequencies were considered (20% and 50%): (1) sub-scenario erA, in which, When considering the impact of the different strategies on ge- from the first year of the programme, dogs affected by the disease (i.e. homozygote netic diversity, more reproducers were removed from the repro- aa) were removed from the reproductive pool; (2) sub-scenario erI,an‘intermediate’ ductive pool and the kinship increase was larger with increased policy in which, from the first year of the programme, dogs affected by the disease (i.e. homozygote aa) were also removed from the reproductive pool; heterozygote severity of selection against disorders and larger initial frequencies dogs (Aa) were allowed to reproduce, but their carrier offspring (i.e. heterozygote of the deleterious allele (see Appendix A: Supplementary Table 1). Aa or homozygote aa) were removed from the reproductive pool; and (3) sub-sce- Breeds with small populations were affected more than breeds nario erC, in which, from the first year of the programme, carriers (i.e. heterozygote with larger populations (Fig. 1). When the initial allele frequency Aa or homozygote aa) were removed from the reproductive pool. was set to 20%, kinship increase was, in general, limited. For exam- ple, when all carriers were removed from the reproductive pool Scenario ‘ps’ (sub-scenario erC) in 2010, U for BSG increased from 0.135 to The aim of this scenario was to control the popular sire effect through a limita- tion on the number of offspring allowed per sire. When a reproducer had exceeded 0.154 (+14%, P < 0.0001), while there was no change for EPB the maximum number of offspring, it was not allowed to reproduce any more; this (0.037 in each case, P > 0.05).

Table 1 Demographic and genealogical characteristics of the four breeds studied.

Breed name Number of dogs in pedigree file T 2006–2010 period Number of dogs FCI threshold Average number of offspring per EqG F U reproducer (maximal number observed) Sires Dams Braque Saint Germain 1999 4.69 283 14 14.8 (62) 7.6 (30) 7.73 0.073 0.103 Berger des Pyrénées 28,834 4.77 3630 182 13.2 (82) 7.4 (40) 7.22 0.091 0.054 Coton de Tulear 40,563 4.37 10,784 539 27.4 (233) 9.8 (39) 6.98 0.061 0.039 Epagneul Breton 183,181 4.88 27,325 1366 18.6 (297) 9.6 (54) 9.33 0.056 0.036

T, generation interval; FCI threshold: 5% of the number of dogs produced during the 2006–2010 period; EqG, number of equivalent generations; F, mean inbreeding coefficient; U, mean kinship coefficient. 161 G. Leroy, X. Rognon / The Veterinary Journal 194 (2012) 343–348 345

0.25 0.25

0.20 0.2

0.15 0.15 BSG 0.10 0.1

0.05 0.05 1990 1995 2000 2005 2010 1990 1995 2000 2005 2010

0.08 0.08 0.07 0.07 0.06 0.06

BRP 0.05 0.05 0.04 0.04 0.03 0.03 1990 1995 2000 2005 2010 1990 1995 2000 2005 2010

0.06 0.06

0.05 0.05

0.04 0.04 COT COT 0.03 0.03

0.02 0.02 1990 1995 2000 2005 2010 1990 1995 2000 2005 2010

0.05 0.05

0.04 0.04 EPB 0.03 0.03

0.02 0.02 1990 1995 2000 2005 2010 1990 1995 2000 2005 2010 Year Year Initial frequency: 20% Initial frequency: 50%

Fig. 1. Evolution of average kinship (U) over 10 years according to scenarios related to the eradication of a monogenic recessive disorder. BSG, Braque Saint Germain; BRP, Berger des Pyrénées; COT, Coton de Tulear; EPB, Epagneul Breton. Observed evolution; scenario erA; scenario erI; scenario erC.

0.25 0.60

0.2 0.40 0.15

0.1 0.20 0.05 Allele frequency 0 0.00 1990 1995 2000 2005 2010 1990 1995 2000 2005 2010 Year Year Initial frequency: 20% Initial frequency: 20%

Fig. 2. Evolution of the frequency of a deleterious allele over 10 years according to scenarios related to the eradication of a monogenic recessive disorder in the Coton de Tulear. Observed evolution; scenario erA; scenario erI; scenario erC.

When the initial allele frequency was set to 50%, kinship in- Limitation of popular sire effect crease was much higher. In 2010 for BSG, U increased from 0.135 to 0.154 for erA (+14%, P < 0.0001), 0.19 for erI (+41%, Application of sub-scenarios involving increasing constraints on P < 0.0001) and 0.21 for erC (+56%, P < 0.0001). However, the im- the number of offspring had an impact on the proportion of replaced pacts were limited for EPB when considering the last year of sim- parents (Table 2), as well as the average kinship (Fig. 3). When ulation. Proportionally to absolute kinship increase, DU limiting the number of offspring to 200 per reproducer (ps200), only computed from 2000 to 2010 was also affected (see Appendix A: a small proportion of matings were affected. There was no impact on Supplementary Table 2), e.g. for BSG, DU increased from 1.3% to U for BSG and BRP, while there were small decreases in U 5.3% per generation when the erC scenario was applied with an ini- from 0.038 to 0.036 (À6%, P < 0.0001) and from 0.038 to 0.037 tial allele frequency of 50%. (À4%, P < 0.0001) for COT and EPB, respectively, in 2010.

162 346 G. Leroy, X. Rognon / The Veterinary Journal 194 (2012) 343–348

Table 2 Proportion of sires and dams changed over the 2001–2010 period depending on the maximum number of offspring allowed per reproducer (ps scenarios).

Breed name Proportion of sires and dams changed (%) Threshold: 25 Threshold: 50 Threshold: 100 Threshold: 200 Sire Dam Sire Dam Sire Dam Sire Dam Braque Saint Germain 22.8 5.8 2.9 0 0 0 0 0 Berger des Pyrénées 67.0 10.0 19.0 0.1 2.8 0 0.3 0 Coton de Tulear – – 76.9 0.3 35.7 0 6.8 0 Epagneul Breton – – 49.0 1.3 22.7 0 4.6 0

Non-random replacement of reproducers 0.15 0.13 Fig. 4 illustrates the evolution of kinship in COT when breeders 0.11 tend to choose replacement sires and dams more related or less re- lated to the replaced one under two scenarios (erC initial allele fre- 0.09 BSG quency = 50% and ps50). The replacement of reproducers by related 0.07 animals tended to increase average kinship, while choosing unre- 0.05 lated reproducers tended to decrease kinship. 1990 1995 2000 2005 2010

0.07

0.06 Discussion

0.05 Management of genetic diversity constitutes an important issue BRP 0.04 for controlling the dissemination of inherited diseases and hence the welfare of dogs. In the present study, we used kinship to inves- 0.03 tigate the evolution of genetic diversity, since it is a key component 1990 1995 2000 2005 2010 of breed conservation (Baumung and Sölkner, 2003) and is directly 0.05 related to the number of founder genome equivalents, i.e., theoret- ical remaining alleles inherited from founders (Caballero and Toro, 0.04 2000). Therefore, the risk of spreading new inherited disorders is proportional to kinship increase. In an ideal closed population,

COT COT 0.03 average kinship increases steadily over time. However, in practice, fluctuations in its evolution may occur due to practices such as 0.02 importation of dogs without known pedigree. 1990 1995 2000 2005 2010 The ‘what if’ procedure developed in this study was used to 0.05 investigate the consequences of breeding practices based on real pedigree data. It takes into account parameters that are difficult 0.04 to include together in classical population simulations, such as overlapping generations, non-random mating and bottleneck

EPB 0.03 events. Using sub-scenario erI, in which heterozygotes were al- lowed to reproduce, but their carrier offspring were removed from 0.02 reproduction, it was estimated that a deleterious allele could be 1990 1995 2000 2005 2010 eliminated after 10 years of selection. Year In practice, the FCI recommendations concerning the number of offspring per reproducer are not applicable for the BRP, COT and Fig. 3. Evolution of average kinship (U) over 10 years according to scenarios related EPB (Table 1), since the maximum number of puppies produced to the limitation of the number of offspring allowed per reproducer. BSG, Braque by all reproducers in the period from 2006 to 2010 was less than Saint Germain; BRP, Berger des Pyrénées; COT, Coton de Tulear; EPB, Epagneul Breton. Observed evolution; scenario ps200; scenario ps100; scenario the recommended threshold specific to each breed. Furthermore, ps50; scenario ps25 (only for BSG and BRP). the FCI recommendation would be difficult to implement for the BSG breed, since sires currently produce more offspring on average than the recommended threshold. However, our simulation ap- When a smaller number of offspring was allowed, the propor- proach enables specific recommendations to be provided within tion of affected matings increased dramatically, modifying kinship the context of a given breed. evolution at the same time. When the number of permitted off- The approach used in this study relies on several hypotheses spring was limited to 50, sires were replaced for 77% of COT indi- and simplifications. We assumed that the current genetic structure viduals, leading to a decrease in U from 0.038 to 0.032 for this would be similar to that of 10 years previously, but this may lead breed in 2010 (À15%, P < 0.0001), while sires were replaced for to bias if the breed has undergone a large change in population 19% of BRP individuals, resulting in a decrease in U from 0.061 size. We also assumed that there was random replacement of to 0.055 (À10%, P < 0.0001). In BSG, there was little change in evo- reproducers, which seldom happens in real populations. As illus- lution of diversity in consecutive ps sub-scenarios. When the num- trated in Fig. 4, a non-random choice of replacement sires or dams ber of offspring per reproducer was limited to 25, there was an may have an effect on the evolution of diversity. It is difficult to unexpected increase of U from 0.135 to 0.149 in 2010 (+10%, estimate if, and at which level, breeders may choose reproducers P < 0.0001). more related or less related to the replaced ones; however, future

163 G. Leroy, X. Rognon / The Veterinary Journal 194 (2012) 343–348 347

Year Year Scenario erC (Initial allele frequency = 50%) Scenario ps50

Fig. 4. Evolution of average kinship (U) over 10 years for scenarios erC (initial allele frequency = 50%) and ps50 in the Coton de Tulear, according to the level of relatedness between replaced sires and dams and sires and dams chosen for replacement. U, Kinship; observed evolution; scenario with random replacement; scenario with 50% of replacement sires (or dams) sampled among the 10th percentile of the most related sire (or dam); scenario with 50% of replacement sires (or dams) sampled among the 10th percentile of the least related sire (or dam).

surveys could be implemented to give an indication about such 77% of the matings with respect to sire replacement. It would be choices. more reasonable to recommend a threshold around 100 (36% of On the basis of these results, some recommendations can be mating affected regarding sire pathway), even if the impact on ge- made for each of the four breeds included in this study, considering netic diversity will be more limited (a relative decrease of U of either an absolute increase in kinship or evolution of DU over the 10%). In the BRP, a threshold of 50 would allow kinship rate to 10 year period according to scenario. To limit the extent of decrease from 0.5% to 0.2%. inbreeding depression, it is generally considered that acceptable In the EPB, even with a high frequency of a deleterious allele, di- values of inbreeding (or kinship) rate per generation should not rect removal of carriers would not affect genetic diversity substan- be >0.5–1% (Bijma, 2000). This value could be somewhat larger tially and erC policy can be recommended in any case. Given the or smaller than the threshold for the BRP and COT, depending on large number of reproducers within the breed, even when a large the various scenarios considered in this study. Note that in scenar- number of individuals are removed from reproduction, the proba- ios aiming to eradicate a monogenic recessive disorder, a brief in- bility of a complete loss of genetically original families is small. crease of rate in kinship was followed by more stable kinship Therefore, the risk of occurrence of a bottleneck in relation to evolution once the disease had been removed. breeding strategies is more limited within the EPB breed. An off- Given its small population size, the situation with the BSG spring threshold of 100 should be adequate for the breed, since seems to be the most problematic. In order to remove a deleterious changing only 23% of sires in 2010 would have led to a predicted allele with a large frequency (50%) from the breed, the most effi- relative decrease of kinship of 12%. cient eradication policies (erI and erC) should be excluded, given their potential negative impact on genetic diversity. For a moder- Conclusions ate frequency of the allele (20%), it is more conceivable to use such policies, even if the predicted impact on genetic diversity (a rela- The simulation method developed here sought to assess the im- tive increase in kinship level of 14% in 2010) is not negligible. pact of different breeding strategies on the frequency of a deleteri- Otherwise, given the efforts already implemented for the manage- ous allele and on genetic diversity for four French dog breeds. By ment of genetic diversity within the breed, imposing a reasonable simulating changes occurring within a pedigree file after implemen- threshold of number of offspring will not improve the situation tation of a chosen breeding strategy, we have provided breed-spe- substantially. The two sires used the most in 2010 show a low level cific recommendations relating to issues such as the removal of an of kinship with the current population, explaining why kinship was inherited disease or limitation of number of offspring per repro- increased when applying the ps25 scenario. The recommendation ducer. The choice of a given strategy is also highly dependent on could be made to increase the number of reproducers or to imple- the existence of other traits to be selected, such as those related to ment more binding breeding schemes, for example minimising behaviour and to severity of the disease. For a same frequency, a dis- kinship (Fernandez et al., 2005). Outcrossing may be an interesting ease with a dramatic impact on viability will likely require a stricter option for the BSG and is periodically used by the breed society. breeding policy than a mildly deleterious one. Adaptation of the In the BRP and COT, the same recommendations could be given procedure to more complex situations (more complex inheritance, regarding eradication of a specific disease. For a large allele fre- segregation of several diseases) could be the subject of further quency (50%), directly removing all carriers (erC) is not desirable, studies. since DU computed over the period would increase from 0.5% to 1.2% and from 0.1% to 0.8%, respectively (see Appendix A: Supple- mentary Table 2), exceeding recommended thresholds. An inter- Conflict of interest statement mediate policy (erI) would have a moderate impact on genetic diversity (a relative increase in kinship level of 11% and 18% for None of the authors of this paper has a financial or personal BRP and COT in 2010, respectively). For an allele frequency close relationship with other people or organisations that could inappro- to 20%, direct removal of carriers (erC) would have a limited effect priately influence or bias the content of the paper. on kinship (a relative increase of 5% and 8% in 2010, respectively). A greater contrast may be observed between the BRP and COT Acknowledgements when limits are imposed on popular sire effects given a more ‘intensive’ use of reproducers in the COT. In this breed, in order The authors would like to thank the Société Centrale Canine for to have a relative decrease of kinship level of 15%, no reproducer the providing data, Michèle Tixier Boichard for useful discussions, should produce more than 50 offspring, which in turn would affect and Andrea Rau and Wendy Brand-Williams for linguistic revision. 164 348 G. Leroy, X. Rognon / The Veterinary Journal 194 (2012) 343–348

Appendix A. Supplementary material Leroy, G., Rognon, X., Varlet, A., Joffrin, C., Verrier, E., 2006. Genetic variability in French dog breeds assessed by pedigree data. Journal of Animal Breeding and Genetics 123, 1–9. Supplementary data associated with this article can be found, in Leroy, G., Baumung, R., 2011. Mating practices and the dissemination of genetic the online version, at http://dx.doi.org/10.1016/j.tvjl.2012.06.025. disorders in domestic animals, based on the example of dog breeding. Animal Genetics 42, 66–74. Lewis, T.W., Woolliams, J.A., Blott, S.C., 2010. Optimisation of breeding strategies to References reduce the prevalence of inherited disease in pedigree dogs. Animal Welfare 19, 93–98. Baumung, R., Sölkner, J., 2003. Pedigree and marker information requirements to Mäki, K., Groen, A.F., Liinamo, A.E., Ojala, M., 2001. Population structure, inbreeding monitor genetic variability. Genetics Selection Evolution 35, 369–383. trend and their association with hip and elbow dysplasia in dogs. Animal Bijma, P., 2000. Long-term genetic contributions: Prediction of rates of inbreeding Science 73, 217–228. and genetic gain in selected populations. PhD Thesis, Wageningen University, Nicholas, F.W., 2011. Response to the documentary Pedigree dogs exposed: Wageningen, The Netherlands, 225pp. Three reports and their recommendations. The Veterinary Journal 189, 123– Caballero, A., Toro, M.A., 2000. Interrelations between effective population size and 125. other pedigree tools for the management of conserved populations. Genetical Nicholas, F.W., Crook, A., Sargan, D.R., 2011. Internet resources cataloguing Research 75, 331–343. inherited disorders in dogs. The Veterinary Journal 189, 132–135. Collins, L.M., Asher, L., Summers, J.F., McGreevy, P., 2011. Getting priorities straight: Urfer, S.R., 2009. Inbreeding and fertility in Irish Wolfhounds in Sweden: 1976 to Risk assessment and decision-making in the improvement of inherited 2007. Acta Veterinaria Scandinavica 51, 21. disorders in pedigree dogs. The Veterinary Journal 189, 147–154. Windig, J.J., Eding, H., Moll, L., Kaal, L., 2004. Effects on inbreeding of different Fernandez, J., Villanueva, B., Pong-Wong, R., Toro, M.A., 2005. Efficiency of the use of strategies aimed at eliminating scrapie sensitivity alleles in rare sheep breeds in pedigree and molecular marker information in conservation programs. Genet The Netherlands. Animal Science 79, 11–20. 170, 1313–1321.

165 The Veterinary Journal 189 (2011) 197–202

Contents lists available at ScienceDirect

The Veterinary Journal

journal homepage: www.elsevier.com/locate/tvjl

How the Orthopedic Foundation for Animals (OFA) is tackling inherited disorders in the USA: Using hip and elbow dysplasia as examples ⇑ G. Gregory Keller a, , Edmund Dziuk a, Jerold S. Bell a,b a Orthopedic Foundation for Animals, Columbia, MO 65201-3806, USA b Department of Clinical Sciences, Tufts Cummings School of Veterinary Medicine, North Grafton, MA 01536-1895, USA article info abstract

Keywords: The Orthopedic Foundation for Animals (OFA) maintains an on-line health pedigree database for inher- Canine ited disorders of animals. With the American Kennel Club Canine Health Foundation, the OFA maintains Inherited disorders the Canine Health Information Center (CHIC) for parent breed clubs to identify breed-specific required Hip dysplasia health tests. Analysis of the results of OFA evaluations in the hip and elbow registries show that selection Elbow dysplasia based on phenotype improves conformation. Disorders with complex inheritance respond best to selec- Genetic registry tion based on depth (ancestors) and breadth (siblings) of pedigree health test results. This information can be derived from vertical pedigrees generated on the OFA website. Ó 2011 Elsevier Ltd. All rights reserved.

Introduction not only in orthopedic diseases but also for cancer, cardiac, hepatic, nephritic, neurologic, ocular and thyroid disease. A prominent businessman in the United States, John M. Olin, While the OFA’s initial focus was canine hip dysplasia, the was also an avid sportsman and recognized the impact of canine mission has broadened to include cats and other genetic hip dysplasia on his Labrador retrievers. Along with the Golden Re- diseases, including elbow dysplasia, patella luxation, autoim- triever Club of America, German Shepherd Club of America and the mune thyroiditis, congenital heart disease, Legg–Calve–Perthes veterinary community, he organized a meeting that eventually led disease, osteochondrosis dissecans (shoulder osteochondrosis), to the formation of the Orthopedic Foundation for Animals (OFA) in sebaceous adenitis and congenital deafness. The methodology 1966. The OFA is guided by the following four specific objectives: and criteria for evaluating the test results for each disorder are independently established by veterinary scientists from their (1) To collate and disseminate information concerning orthope- respective specialty areas and the standards used are generally dic and genetic diseases of animals. accepted throughout the world. Disorders present on the OFA (2) To advise, encourage and establish control programs to website include those that have a defined test for normalcy. Dis- lower the incidence of orthopedic and genetic diseases. orders such as epilepsy, gastric dilatation/volvulus and cancers (3) To encourage and finance research in orthopedic and genetic that do not have defined phenotypic or genotypic tests are not disease in animals. included. If genetic markers for disease liability are identified (4) To receive funds and make grants to carry out these in the future, these can be added as tools for genetic disease objectives. control. The power of the OFA genetic database lies in the compilation The OFA is governed by a voluntary Board of Directors. As a not- and integration of all health screening information in a single loca- for-profit organization, the revenue over expenses is either held in tion. For dogs with an existing OFA record, examination results the operating reserve or donated to support animal health-related from the Canine Eye Registry Foundation (CERF) are incorporated research. Most funding is channeled through the American Kennel in their OFA record. In addition, the results of genotypic tests that Club Canine Health Foundation (AKC-CHF)1 or Morris Animal Foun- are either submitted by the owner or through a cooperative agree- dation, with occasional direct funding. OFA has supported research ment with the parent club are also included in the OFA genetic database. Cutting-edge advancements in molecular genetics now account for over 90 DNA tests involving over 145 breeds of dogs and cats. ⇑ Corresponding author. Tel.: +1 800 4420418x223. E-mail address: [email protected] (G.G. Keller). The collection of such data is meaningless unless the data can 1 See: www.akcchf.org. be disseminated to parties of interest. The OFA maintains an

1090-0233/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tvjl.2011.06.019 166 198 G.G. Keller et al. / The Veterinary Journal 189 (2011) 197–202 on-line database of >1 million phenotypic and genotypic test re- Group, 2001). When a specific component of elbow dysplasia is ob- sults.2 All normal or grades of normal results in the OFA database served, it is reported in addition to the Grade as ununited anconeal are available on-line. Abnormal or grades of abnormal results are process, osteochondrosis or fragmented medical coronoid process. available on-line if released by the owner, or if the results are part Elbow radiographs are subjected to the same of-age or preliminary of a breed club program where all (normal and abnormal) test re- evaluation and certification process as hip radiographs. sults are published. Diseases with complex inheritance can respond to selective The Canine Health Information Center (CHIC)3 is a program that pressure based on phenotype (Keller, 2006; Pirchner, 1983). In this is dually sponsored by the OFA and the AKC-CHF. The parent clubs manuscript, the OFA hip and elbow registries are used to illustrate determine the breed-specific health issues for CHIC certification this response. and encourage breeder participation in the program. The CHIC pro- gram is not about normalcy; it is about health consciousness. Dogs Materials and methods receive CHIC certification if they have completed the required breed-specific health testing, regardless of the test results. Other The OFA hip registry of 1,187,831 evaluations was queried for hip ratings of requirements include permanent identification (tattoo or microchip) progeny where both parents also had known of-age hip ratings. Data were collected on progeny with of-age or preliminary hip confirmation ratings of normal (Excel- and release to the open database of abnormal results. CHIC encour- lent, 1; Good, 2; Fair, 3) or dysplastic (Mild, 5; Moderate, 6; Severe, 7). Progeny with ages health screening to improve the overall health of breeds. There Borderline (4) hip ratings were not included. The hip ratings of both parents were are presently over 139 parent breed clubs participating, with over recorded, including all seven grades. A hip Combined Parent Score (CPS) for each 64,500 dogs achieving CHIC certification. mating was determined by adding together the numbers corresponding to the The acceptance of the CHIC certification program by parent hip rating for each parent; for two OFA Excellent parents the CPS was 2 and for two OFA Severe parents the CPS was 14. Matings with the same CPS were combined breed clubs and breeders provides an avenue for the only proven together for analysis; e.g. Good mated to Borderline, Fair mated to Fair and Excel- method of genetic disease control: breed-specific phenotypic and lent mated to Mild all have a CPS of 6. genotypic screening of prospective breeding stock. The CHIC pro- The OFA elbow registry of 260,195 evaluations was queried for elbow ratings of gram provides a standard for breeders to practice health-conscious progeny where both parents had known of-age elbow ratings. Data were collected on progeny with preliminary or of-age elbow confirmation ratings of Normal (1) or breeding. It also allows pet owners to screen prospective purchases dysplastic (Grade I, 2; Grade II, 3; Grade III, 4). An elbow CPS for each mating was for evidence of health-conscious breeding. determined by adding together the numbers corresponding to the elbow rating for Another goal of the CHIC program is to collect and store canine each parent; for two OFA Normal parents the CPS was 2 and for two OFA Grade III DNA samples, along with corresponding genealogic and pheno- parents the CPS was 8. Matings with the same CPS were combined together for typic information, to facilitate future research and testing aimed analysis. Pearson correlation analysis was performed to compare the CPS of matings to at reducing the incidence of inherited disease in dogs. Researchers the observed percentages of hip dysplasia or elbow dysplasia in the progeny. have been hampered by the lack of appropriate DNA samples and the DNA repository addresses this need. To date, the CHIC DNA Results Repository contains DNA from over 12,500 dogs and has received 17 requests from researchers, resulting in the distribution of over Table 1 shows the hip ratings for 490,966 progeny in the OFA 2,200 DNA samples with their appropriate health and pedigree hip registry with known sire and dam hip ratings. The percentage information. of dysplastic progeny increased as the parental hip scores in- To evaluate hip dysplasia, the OFA employs the ventrodorsal creased. The total number of hip radiograph submissions from par- hip-extended positioning recommended by the American Veteri- ents with normal hip ratings was significantly greater than those nary Medical Association (AVMA Council on Veterinary Service, from parents with dysplastic hip ratings (P > 0.05). 1961). The in-house radiologist is the sole evaluator for prelimin- Fig. 1 shows the relationship between the CPS and the percent- ary evaluation of dogs <24 months of age. The reliability of preli- age of dysplastic progeny. Matings with the same CPS (on the diag- minary hip evaluations for predicting of-age OFA ratings was onal of Table 1) were strongly correlated with increasing demonstrated by Corley et al. (1997). Dogs or cats must be percentages of dysplastic progeny (Pearson correlation coefficient P24 months of age to receive OFA hip certification. Radiographs r = 0.96; P > 0.05). The single CPS that did not reflect this trend are independently evaluated by three board-certified veterinary was for matings between two severely dysplastic parents, where radiologists out of a pool of consultants maintained by the OFA. only 18 progeny were submitted for evaluation. The consensus rating of these three radiologists becomes the hip Table 2 shows the elbow ratings for 67,599 progeny in the OFA rating that is reported to the owner and referring veterinarian. elbow registry with known sire and dam elbow ratings. Matings There is a high degree of inter- and intra-reader correlation for including one normal parent had significantly lower percentages conventional and digital images (Corley, 1992; Essman and Sher- of progeny with elbow dysplasia (12.4%) than those between two man, 2006). parents with elbow dysplasia (45.4%) (P > 0.05). Matings involving Seven OFA hip ratings are reported: Excellent, Good, Fair, Bor- a parent with Grade I elbow dysplasia produced significantly more derline, Mild, Moderate or Severe. The first three ratings are con- elbow dysplasia (25.6%) than matings including a parent with nor- sidered to be normal, while the last three ratings are regarded as mal elbows ( 2 = 0.77, 6 df, P = 0.99). dysplastic. A Borderline rating is given when there is no clear con- v Fig. 2 shows the relationship between the CPS and the percent- sensus between radiologists to place the hips in a category of nor- age of progeny with elbow dysplasia. The Pearson correlation coef- mal or dysplastic. It is recommended that dogs with this rating ficient between the CPS and percentage of dysplastic progeny was have a repeat radiograph submitted after a minimum of 6 months. r = 0.06. The lack of correlation is due to the low percentage of dys- The OFA elbow dysplasia registry employs the protocol estab- 4 plasia in progeny of Grade III sires bred to Grade II dams, and Grade lished by the International Elbow Working Group (IEWG), which III parents bred to each other. The total number of progeny from consists of Normal or Grades I, II or III Dysplastic based on the sever- these matings numbered 14 and 3, respectively. ity of secondary osteoarthritis/degenerative joint disease present on an extreme flexed mediolateral view (International Elbow Working Discussion

2 See: www.offa.org. 3 See: www.caninehealthinfo.org. The OFA hip data and CPS demonstrate that hip dysplasia is 4 See: www.iewg-vet.org/. inherited in an additive and quantitative manner. This verifies 167 G.G. Keller et al. / The Veterinary Journal 189 (2011) 197–202 199

Table 1 Progeny results of matings between parents with known hip scores.

Sire rating Dam rating Total Excellent (1) Good (2) Fair (3) Borderline (4) Mild (5) Moderate (6) Severe (7) Excellent (1) Dysplastic (%) 3.6 6.1 9.6 12.3 13.4 18.7 18.5 Total 17,972 52,784 9039 155 1271 729 65 82,015 Good (2) Dysplastic (%) 5.8 9.6 14.6 17.5 18.9 23.0 31.5 Total 50,485 217,938 49,212 811 6930 3973 461 329,810 Fair (3) Dysplastic (%) 9.4 14.1 19.8 22.8 26.5 32.2 37.1 Total 6241 41,628 13,513 263 2301 1328 167 65,441 Borderline (4) Dysplastic (%) 8.9 17.7 20.2 22.2 30.8 50.0 50.0 Total 79 532 168 9 39 30 4 861 Mild (5) Dysplastic (%) 16.4 18.3 27.2 36.2 29.6 41.4 45.0 Total 807 4531 1532 47 459 239 40 7655 Moderate (6) Dysplastic (%) 18.9 22.8 31.6 34.4 35.0 38.0 65.3 Total 428 2618 896 32 266 213 49 4502 Severe (7) Dysplastic (%) 22.0 24.2 36.0 44.4 39.6 55.8 44.4 Total 59 360 136 9 48 52 18 682 Total 76,071 320,391 74,496 1326 11,314 6564 804 490,966

Fig. 1. Relationship of Combined Parent Score to percentage of hip dysplastic progeny.

the conclusions of other researchers that canine hip dysplasia is of OFA hip ratings in the breed over a 40 year period. These results inherited as a quantitative trait (Leighton, 1997; Zhu et al., 2009; validate the OFA recommendation that using parents with better Hou et al., 2010). Hou et al. (2010) analyzed all Labrador retrievers phenotypic hip conformation produces offspring with better hips. in the open-access OFA hip database and calculated an heritability It was expected that fewer radiographs would be submitted for of 0.21, which confirms hip dysplasia acting as a moderately heri- the progeny of two dysplastic parents, since fewer breeders per- table disease. They also confirmed a steady genetic improvement form such matings. The low numbers may also be due to pre- 168 200 G.G. Keller et al. / The Veterinary Journal 189 (2011) 197–202

Table 2 with an individual’s own rating, increases the accuracy of selection Progeny results of matings between parents with known elbow scores. and hence response to selection. Similarly, selection based on hor- Sire rating Dam rating Total izontal or breadth-of-pedigree hip ratings (siblings), when com- Normal Grade I Grade II Grade III bined with an individual’s own rating, increases accuracy of (1) (2) (3) (4) selection and hence response to selection (Pirchner, 1983; Keller, Normal (1) 2006). Dysplastic 10.1 24.1 29.4 28.1 Breeding schemes that employ estimated breeding values (%) (EBVs) that combine phenotypic ratings from all known relatives Total 55,867 4309 875 167 61,218 (weighted according to genetic relationship) provide the greatest Grade I (2) selective power, rather than single measurements on individual Dysplastic 22.0 41.0 46.9 52.2 (%) dogs (Zhu et al., 2009; Hou et al., 2010). EBVs that utilize molecular Total 3917 591 145 23 4676 genetic markers for liability genes would be even more beneficial Grade II (3) (Stock and Distl, 2010; Zhou et al., 2010). Dysplastic 32.6 55.4 65.8 57.1 The open-access OFA health database website provides breed- (%) Total 1121 222 38 14 1395 ers with the information that helps them to make informed breed- Grade III (4) ing decisions. When an individual dog’s record is accessed, detailed Dysplastic 23.9 38.1 14.3 0.0 information on all recorded health issues, including test results, (%) age at the time of testing and the resulting certification numbers, Total 251 42 14 3 310 are available. Sire and dam information are provided, as well as Total 61,156 5164 1072 207 67,599 information on full and half siblings and any offspring that may be in the database. A vertical pedigree can be generated from a link on the individual’s OFA page, providing traditional depth of pedi- gree and breadth of pedigree health information. This type of data screening of radiographs with obviously dysplastic hips by veteri- is extremely useful when trying to make selection decisions based narians; these radiographs may not be submitted to the OFA for on phenotypic data. evaluation (Paster et al., 2005). This would reduce the resultant fre- The vertical hip pedigree of the Golden retriever Champion (Ch.) quencies of dysplastic individuals. Prescreening of dysplastic Faera’s Starlight (Fig. 3) shows how parent, grandparent, offspring radiographs for OFA submission appears to be constant over time and sibling information are combined in a single graphic format for (Reed et al., 2000). evaluation. Whilst this dog had hips with an Excellent rating, he Traits such as hip dysplasia and elbow dysplasia are complexly was bred from Fair- and Good-rated parents, with three Fair- and (polygenically) inherited, with increasing incidence based on one Good-rated grandparents. While he produced 92.4% normal increasing frequencies of susceptibility alleles at loci that contrib- offspring with a preponderance of Good ratings, he produced more ute to variation in liability. Selection based on vertical or depth-of- Fair- than Excellent-rated offspring. The vertical pedigree provides pedigree hip ratings (parents and grandparents), when combined more information than the single individual rating. Vertical

Fig. 2. Relationship of Combined Parent Score to percentage of elbow dysplastic progeny. 169 G.G. Keller et al. / The Veterinary Journal 189 (2011) 197–202 201

Fig. 3. OFA vertical pedigree of Golden retriever Ch. Faera’s Starlight.

pedigrees of individual animals are available on the OFA website an evaluation of both subluxation on the ventrodorsal hip-ex- for the hip, elbow, cardiac, thyroid, patella, CERF (eye) and degen- tended view, as well as radiographic anatomy and secondary erative myelopathy registries. boney changes. EBV technology would combine all of the phenotypic informa- The PennHIP method recommends selection based on the DI tion in Fig. 3 into a single measurement that provides the most measurement of individual dogs. Based on PennHIP data of dogs accurate possible prediction of the average performance of the off- presented to the University of Pennsylvania School of Veterinary spring of the dog in question (Faera’s Starlight). However, the indi- Medicine, 100% of Golden retrievers and 89% of Labrador retrievers vidual’s OFA page and vertical pedigree allows the breeder to who received normal OFA ratings were deemed osteoarthritis-sus- determine where the liability comes from in the pedigree, the spe- ceptible by their DI (Powers et al., 2010). Powers et al. (2010) also cific results from each mating and each dog’s strengths and weak- raised the possibility that the Cardigan Welsh Corgi is genetically nesses. These are useful tools for selection and genetic fixed for hip dysplasia, based on DI measurements for the breed. improvement. However, the clinical presentation of disease in these breeds does The distraction index (DI) measurement of the PennHIP method not bear out these predictions, suggesting that there is a high false- for hip dysplasia control employs a mechanical distraction device positive rate for DI prediction of clinical disease. A study correlat- to measure maximal hip joint laxity as a predictor of future degen- ing ventrodorsal hip-extended radiographic ratings to later insur- erative joint disease and osteoarthritis (Smith et al., 1990). Penn- ance-related claims for hip dysplasia showed a strong association HIP studies show that the OFA rating and DI measurement are (Malm et al., 2010). Data correlating DI measurements to morbid- significantly associated (Powers et al., 2010) and DI measurements ity from clinical disease have not been published. submitted by their owners to the OFA are included in the hip dys- Dog breeds have closed stud books and dog breeders have con- plasia registry. cerns about genetic diversity and the effects of artificial selection While the DI provides a measurement of laxity, it does not take on their gene pools (Calboli et al., 2008). The removal of 89% or into account degenerative joint disease or osteoarthritic changes. more of possible breeding stock for a single genetic disorder Studies have shown that liability for hip dysplasia and liability (which would be required in order to breed only from those Labra- for osteoarthritis are controlled by separate genes (Clements dor retrievers with acceptable DI) will doom any breed to extinc- et al., 2006; Zhou et al., 2010). The OFA hip rating incorporates tion from genetic depletion. While breeding from only OFA 170 202 G.G. Keller et al. / The Veterinary Journal 189 (2011) 197–202

Excellent dogs will significantly improve hip ratings of progeny, Acknowledgement the elimination of the rest of the phenotypically normal dogs from breeding (most of which produce predominantly normal dogs) The authors thank Ms. Rhonda Hovan for allowing use of the would also severely restrict the gene pools of breeds. Pragmatic pedigree of Ch. Faera’s Starlight. breeding recommendations include breeding from normal dogs with increasing normalcy of parents, grandparents, siblings and References progeny, as shown on the OFA vertical pedigree, and through the use of EBVs. AVMA Council on Veterinary Service, 1961. Report of panel on canine hip dysplasia. Journal of the American Veterinary Medical Association 139, 791–798. The significant difference between progeny from one parent Calboli, F.C., Sampson, J., Fretwell, N., Balding, D.J., 2008. Population structure and with normal elbows and progeny from two parents with dysplastic inbreeding from pedigree analysis of purebred dogs. Genetics 179, 593–601. elbows suggests a qualitative trait. However, it is established that Clements, D.N., Carter, S.D., Innes, J.F., Ollier, W.E., 2006. Genetic basis of secondary elbow dysplasia is a polygenic (multifactorial) trait (Engler et al., osteoarthritis in dogs with joint dysplasia. American Journal of Veterinary Research 67, 909–918. 2009). Increasing CPS tended to increase the frequency of elbow Corley, E., 1992. Role of the Orthopedic Foundation for Animals in the control of dysplasia in the progeny, but low numbers of submissions for some canine hip dysplasia. Veterinary Clinics of North America Small Animal Practice mating types between dysplastic parents skewed the results, mak- 22, 579–593. Corley, E.A., Keller, G.G., Lattimer, J.C., Ellersieck, M.R., 1997. Reliability of early ing the correlation inconclusive. Again, pre-screening and non-sub- radiographic evaluations for canine hip dysplasia obtained from the standard mission to OFA of obviously dysplastic radiographs may have ventrodorsal radiographic projection. Journal of the American Veterinary affected the data. Medical Association 211, 1142–1146. Engler, J., Hamann, H., Distl, O., 2009. Schätzung populationsgenetischer parameter Grade I elbow dysplasia is a radiographic diagnosis that usually für röntgenologische befunde der ellbogengelenkdysplasie beim Labrador does not produce clinical disease or morbidity in the dog. Some retriever. Berliner und Münchener Tierärztliche Wochenschrift 122, 378–385. breed groups counsel owners to ignore the diagnosis of Grade I el- Essman, S., Sherman, A., 2006. Comparison of digitized and conventional radiographic images for assessment of hip joint conformation of dogs. bow dysplasia and to treat these dogs as if they were normal. How- American Journal of Veterinary Research 67, 1546–1551. ever, the data presented here demonstrates that progeny from a Hou, Y., Wang, Y., Lust, G., Zhu, L., Zhang, Z., Todhunter, R.J., 2010. Retrospective parent with Grade I elbow dysplasia, when bred to mates from analysis for genetic improvement of hip joints of cohort Labrador retrievers in the United States: 1970–2007. PLoS ONE 5, e9410. all other rating classifications, have a significantly increased fre- International Elbow Working Group, 2001. 2001 International Elbow Protocol quency of elbow dysplasia. These results are significantly different (Vancouver). www.iewg-vet.org/archive/protocol.htm (accessed 6 May 2011). from the results observed with progeny from one normal parent Keller, G.G., 2006. The Use of Health Databases and Selective Breeding: A Guide for bred to mates from all other rating classifications. Dog and Cat Breeders and Owners. Orthopedic Foundation for Animals, Columbia, Missouri, USA. www.offa.org/pdf/monograph2006web.pdf The data show that even two dogs with normal elbow radio- (accessed 6 May 2011). graphs may produce 10.1% progeny with elbow dysplasia. This is Leighton, E.A., 1997. Genetics of canine hip dysplasia. Journal of the American where consideration of depth and breadth of pedigree information Veterinary Medical Association 210, 1474–1479. Malm, S., Fikse, F., Egenvall, A., Bonnett, B.N., Gunnarsson, L., Hedhammar, A., becomes important. Any rating of elbow dysplasia in siblings of Strandberg, E., 2010. Association between radiographic assessment of hip status dogs with a normal elbow rating provides evidence that the normal and subsequent incidence of veterinary care and mortality related to hip dog may carry additional elbow dysplasia liability alleles. dysplasia in insured Swedish dogs. Preventive Veterinary Medicine 93, 222– 232. Selection for increasing normalcy of depth and breadth of pedi- Paster, E.R., LaFond, E., Biery, D.N., Iriye, A., Gregor, T.P., Shofer, F.S., Smith, G.K., gree information provides a better selection tool for complexly 2005. Estimates of prevalence of hip dysplasia in Golden retrievers and inherited disease. The use of the OFA vertical pedigree provides Rottweilers and the influence of bias on published prevalence figures. Journal of the American Veterinary Medical Association 226, 387–392. the information necessary to make informed breeding decisions. Pirchner, F., 1983. Population Genetics in Animal Breeding, Second Ed. Plenum The addition of EBVs that combine all of this information (Engler Press, New York, USA, 414 pp. et al., 2009) and that also include genotypes of DNA markers for Powers, M.Y., Karbe, G.T., Gregor, T.P., McKelvie, P., Culp, W.T., Fordyce, H.H., Smith, G.K., 2010. Evaluation of the relationship between Orthopedic Foundation for liability genes (Stock and Distl, 2010; Zhou et al., 2010) would be Animals’ hip joint scores and PennHIP distraction index values in dogs. Journal even more beneficial. of the American Veterinary Medical Association 237, 532–541. Reed, A.L., Keller, G.G., Vogt, D.W., Ellersieck, M.R., Corley, E.A., 2000. Effect of dam and sire qualitative hip conformation scores on progeny hip conformation. Conclusions Journal of the American Veterinary Medical Association 217, 675–680. Smith, G.K., Biery, D.N., Gregor, T.P., 1990. New concepts of coxofemoral joint The OFA data show that hip and elbow conformation improve stability and the development of a clinical stress-radiographic method for quantitating hip joint laxity in the dog. Journal of the American Veterinary with improving parental phenotypic ratings. The open access OFA Medical Association 196, 59–70. website provides health test results on individuals, as well as depth Stock, K.F., Distl, O., 2010. Simulation study on the effects of excluding offspring and breadth of pedigree health information on closely related indi- information for genetic evaluation versus using genomic markers for selection in dog breeding. Journal of Animal Breeding and Genetics 127, 42–52. viduals. This information provides the best means for making Zhou, Z., Sheng, X., Zhang, Z., Zhao, K., Zhu, L., Guo, G., Friedenberg, S.G., Hunter, L.S., breeding decisions for both complexly inherited and Mendelian Vandenberg-Foels, W.S., Hornbuckle, W.E., Krotscheck, U., Corey, E., Moise, N.S., disorders. Dykes, N.L., Li, J., Xu, S., Du, L., Wang, Y., Sandler, J., Acland, G.M., Lust, G., Todhunter, R.J., 2010. Differential genetic regulation of canine hip dysplasia and osteoarthritis. PLoS ONE 5, e13219. Conflict of interest statement Zhu, L., Zhang, Z., Friedenberg, S., Jung, S.W., Phavaphutanon, J., Vernier-Singer, M., Corey, E., Mateescu, R., Dykes, N., Sandler, J., Acland, G., Lust, G., Todhunter, R., 2009. The long (and winding) road to gene discovery for canine hip dysplasia. The authors are Chief of Veterinary Services (GGK), Chief Oper- The Veterinary Journal 181, 97–110. ating Officer (ED) and Director (JSB) of the not-for-profit Orthope- dic Foundation for Animals.

171 Comparative analyses of genetic trends and prospects for selection against hip and elbow dysplasia in 15 UK dog breeds Lewis et al.

Lewis et al. BMC Genetics 2013, 14:16 http://www.biomedcentral.com/1471-2156/14/16

172 Lewis et al. BMC Genetics 2013, 14:16 http://www.biomedcentral.com/1471-2156/14/16

RESEARCHARTICLE Open Access Comparative analyses of genetic trends and prospects for selection against hip and elbow dysplasia in 15 UK dog breeds Thomas W Lewis1*, Sarah C Blott1 and John A Woolliams2

Abstract Background: Hip dysplasia remains one of the most serious hereditary diseases occurring in dogs despite long-standing evaluation schemes designed to aid selection for healthy joints. Many researchers have recommended the use of estimated breeding values (EBV) to improve the rate of genetic progress from selection against hip and elbow dysplasia (another common developmental orthopaedic disorder), but few have empirically quantified the benefits of their use. This study aimed to both determine recent genetic trends in hip and elbow dysplasia, and evaluate the potential improvements in response to selection that publication of EBV for such diseases would provide, across a wide range of pure-bred dog breeds. Results: The genetic trend with respect to hip and elbow condition due to phenotypic selection had improved in all breeds, except the Siberian Husky. However, derived selection intensities are extremely weak, equivalent to excluding less than a maximum of 18% of the highest risk animals from breeding. EBV for hip and elbow score were predicted to be on average between 1.16 and 1.34 times more accurate than selection on individual or both parental phenotypes. Additionally, compared to the proportion of juvenile animals with both parental phenotypes, the proportion with EBV of a greater accuracy than selection on such phenotypes increased by up to 3-fold for hip score and up to 13-fold for elbow score. Conclusions: EBV are shown to be both more accurate and abundant than phenotype, providing more reliable information on the genetic risk of disease for a greater proportion of the population. Because the accuracy of selection is directly related to genetic progress, use of EBV can be expected to benefit selection for the improvement of canine health and welfare. Public availability of EBV for hip score for the fifteen breeds included in this study will provide information on the genetic risk of disease in nearly a third of all dogs annually registered by the UK Kennel Club, with in excess of a quarter having an EBV for elbow score as well. Keywords: Canine, Hip dysplasia, Elbow dysplasia, Estimated breeding value, Selection, Accuracy, Genetic correlation, Heritability, Welfare

Background femoral (hip) joint [6]. Over time, particularly in larger Hip dysplasia may be described as one of the most ser- and giant breeds, the malformation and laxity lead to the ious hereditary diseases occurring in pedigree dogs given abnormal wearing of bone surfaces and the appearance of the popularity of susceptible breeds and the prevalence the osteoarthritic signs of degenerative joint disease (DJD) therein [1,2]. It is also one of the most persistent, first [7]. The resultant osteoarthritis (OA) is irreversible and so having been described over 50 years ago [3-5]. Hip dyspla- the only way to effect a lasting and widespread improve- sia is a developmental orthopaedic disorder characterised ment in the welfare of susceptible breeds is through genetic by the formation of a dysmorphic, lax (loose) coxo- selection. Hip dysplasia remains a significant problem, despite the presence of several evaluation schemes across * Correspondence: [email protected] the world designed to provide an empirical phenotype for 1Kennel Club Genetics Centre at the Animal Health Trust, Lanwades Park, Kentford, Newmarket, Suffolk CB8 7UU, UK selection, partly due to its complexity; a polygenic back- Full list of author information is available at the end of the article ground and multiple environmental influences ensure no

© 2013 Lewis et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

173 Lewis et al. BMC Genetics 2013, 14:16 Page 2 of 11 http://www.biomedcentral.com/1471-2156/14/16

clear pattern of inheritance. Furthermore, the breeding increased accuracy and abundance of reliable informa- guidelines accompanying evaluation schemes have often tion that publication of EBV would provide. elicited only very weak selection [8,9]. In contrast elbow dysplasia, despite also being a devel- Methods opmental orthopaedic abnormality long recognised as a Data serious problem [10], has historically received less atten- Phenotype data comprised results of the British Veterinary tion than hip dysplasia. As a result, schemes evaluating Association (BVA)/UK Kennel Club (KC) hip and elbow elbow condition are younger than those examining hips, scoring schemes. Details of scoring protocols are given by and so data is less abundant. The term ‘elbow dysplasia’ Gibbs [26] and Lewis et al. [17]. In brief, radiographs of commonly describes a number of abnormalities associ- hips are scored bilaterally on 9 features according to the ated with developmental physiological incongruity of the degree of laxity and/or OA observed (8 features scored elbow joint that often result in OA [11]. 0 to 6, one feature scored 0 to 5). The aggregate of the This grouping of syndromes for both the pathology 18 scores reported ranges from 0 (indicating no malfor- and evaluation of elbow dysplasia may result in under- mation) to 106 (severe hip dysplasia). The BVA/KC elbow estimates of heritability [12]; which range from 0.10 to scoring scheme was launched in 1998 based on guidelines 0.38 [13-17] among various breeds. Analyses of more of the International Elbow Working Group (IEWG). specific elbow abnormalities have estimated higher heri- Elbow radiographs are scored according to the size of tabilities; for example 0.57 for fragmented coronoid detectable primary lesions and severity and extent of OA process in German Shepherd Dogs [9]. Estimates of her- observed, ranging from 0 (normal) to 3 (severe elbow dys- itability of hip condition generally have a smaller range plasia). The score of the worst elbow only is publically but appear moderate in magnitude, from 0.20 to 0.43 reported. Pedigree data was provided by the KC and linked across various breeds [8,14,16,18-20] despite using data to phenotype data via a unique registration number. from different international scoring schemes and hips Fifteen breeds (Akita [AKT], Bearded Collie [BEARD], being evaluated on both detectable laxity and OA. The Bernese Mountain Dog [BMD], Border Collie [BORD], reported genetic correlation between hip and elbow con- English Setter [ENG], Flat Coat Retriever [FCR], Gordon dition varies even more, from −0.09 to 0.42 [9,14,16,17]. Setter [GDN], Golden Retriever [GR], German Shepherd Many recent studies estimating the genetic parameters Dog [GSD], Labrador Retriever [LAB], Newfoundland of hip and elbow dysplasia score data have recom- [NEWF], Rottweiler [ROTT], Rhodesian Ridgeback [RR], mended selection using estimated breeding values (EBV; Siberian Husky [SHUSK] and Tibetan Terrier [TT]) [8,9,14,16,19-21]. EBV are the best linear unbiased pre- were included in the study. For 5 breeds (BMD, GR, dictor (BLUP) of every dog’s breeding value derived from GSD, LAB and ROTT) the genetic parameters of hip the pedigree information used in its calculation [1], and and elbow score were estimated using bivariate analyses. are a more accurate estimate of the genetic liability of a For the remaining 10 breeds, the genetic parameters of trait than the individual phenotype. However, attempts hip score only were estimated using univariate analyses. to quantify the potential benefit to the response to selec- For the ten breeds with hip score only, genetic param- tion against hip and elbow dysplasia that the increased eters and EBV were estimated simultaneously using data accuracy of selection using EBV would bring (compared from dogs evaluated at 365–1459 days old and between to phenotypic selection) are less common than param- 1990 and 2011 inclusive, and the entire KC electronically eter estimation, but have been made empirically by recorded pedigree extending back to the early 1980s; Lewis et al. [8], and via simulation by Stock and Distl hip score having undergone transformation to improve [22] and Malm et al. [23]. Improvements in the rate of normality (see below). For BMD and ROTT genetic pa- genetic progress (which is directly related to the accu- rameters and EBV were computed simultaneously for racy of selection, [24]) would be achieved not only hip and elbow data via bivariate REML analyses using through EBV acting as a more accurate predicator of evaluations from dogs of the same age and study period genetic risk (i.e. the true breeding value) than phenotype, and the entire KC electronic pedigree. The pedigrees of but also through enhanced opportunities to increase se- LAB, GSD and GR were too large to include in their en- lection intensity due to EBV being available for every tirety in bivariate parameter estimation on a desktop PC, dog in the pedigree [25]. EBV would effectively provide and so for parameter estimation in these breeds data a greater quantity of more reliable information with re- and/or depth of pedigree was truncated. For GSD and spect to breeding. This study, therefore, aims to estimate GR genetic parameters of hip and elbow score were esti- the genetic parameters of hip and elbow dysplasia in mated using data from all dogs of the same age and the UK registered breeds for which score data is most study period with a further 5 generations of pedigree. abundant, determine any genetic trends and evaluate For LAB genetic parameters of hip and elbow scores potential improvements in response to selection due to were estimated using data from all dogs evaluated at the

174 Lewis et al. BMC Genetics 2013, 14:16 Page 3 of 11 http://www.biomedcentral.com/1471-2156/14/16

same age and between 2000–2011, and 2 further genera- effects. To extend this univariate model to bivariate ana- 2 tions of pedigree. The genetic parameters for LAB, GSD lyses the variance terms such as σ a were replaced by and GR were then used in the calculation of BLUP EBV the appropriate bivariate covariance matrices (Σ) for the using hip and elbow data from 1990–2011 and the entire traits using the Kronecker product, such as A ⊗ ΣA.The 2 KC pedigrees of each breed (GR pedigree = 386,580 ani- phenotypic variance is denoted as σP, and heritability 2 mals; GSD pedigree = 572,552 animals; LAB data = 59,077 (h ) is calculated as the proportion of phenotypic vari- 2 2 evaluations, pedigree = 977,083 animals), undertaken by ance explained by the additive genetic variance (σA/σP). Edinburgh Genetic Evaluation Service (EGENES) using Phenotypic, additive genetic and residual correlations MiX99. The numbers of records used in the REML ana- (rP,rA,rE) were computed from the genetic (co)variances lyses of hip score for each breed are shown in Additional obtained. file 1: Table S1. Fixed effects included in the model were: sex, inbreed- Thus, data for EBV computation included 142,287 hip ing coefficient (as calculated using the entire KC elec- scores from all fifteen breeds, which have a total mean tronic pedigree), age in days at evaluation, absolute day of 82,118 registrations per year (2000 to 2010 data), and of birth (measured as days since 1st January 1980) and 13,908 elbow scores from BMD, GR, GSD, LAB and year of evaluation. Age in days and absolute day of birth ROTT; these breeds having a total mean of 70,363 regis- were fitted with random smoothing splines to model trations per year (2000–2010 data). temporal trends [8].

Analyses Meta-analysis of parameter estimates across breeds Mixed linear models were fitted using ASREML [27]. The spread of parameter estimates will be due to two For univariate analysis of hip score the model used was components: (i) sampling errors within a breed, and (ii) as per Lewis et al. (2010) [8]. For bivariate analysis of variation in the true parameter among breeds. A meta- hip and elbow score the model used was as per Lewis analysis of the parameter estimates was undertaken to et al. (2011) [17]. obtain the best estimate of the mean parameter for the Total hip score was log transformed (after adding 1 to population of breeds, together with a standard error to avoid necessitating the logarithm of zero) to improve account for both sampling and population variation. normality. Where applicable the untransformed mean of This followed the procedures of Corbin et al. [28]. The left and right elbow score was included as a y-variate. analysis provides an estimate of the variance of the true The possible transformation of observed values to more parameter among breeds, and if this is 0 then the pooled closely correspond to the underlying liability [17] was mean is identical to that obtained from using a weight not undertaken as the benefits were found to be small for each breed equal to the reciprocal of its sampling and because, importantly, the transformation depends variance. on the prevalence which may change over time. Data from 3 year old animals (1095–1459 days) were included Accuracy of estimated breeding values for consistency with hip data and after preliminary ana- The accuracy (r) of each animal’s EBV was calculated as: lysis using Labrador data showed the genetic correlation sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi – – PEV of elbow score at 365 1094 days and 1095 1495 days (i.e. r – ¼ 1 2 1 2 and 3 year olds) was indistinguishable from 1. ðÞ1 þ F σA The general form of the univariate linear model was as follows: (see Additional file 2), where PEV is the prediction error variance of each EBV, F is the inbreeding coeffi- Y Xb Za Wc e 2 ¼ þ þ þ cient for each animal and σ A is the estimated additive genetic variance obtained from the mixed model ana- where Y is the vector of observations, W, X and Z are lysis. ASREML provides both the estimates of the EBV known incidence matrices, b is the vector of fixed ef- and their associated PEVs. fects; a is the vector of random additive genetic effects Potential advantages of using EBV in future selection with the distribution assumed to be multivariate normal for lower hip/elbow scores were evaluated by compari- σ2 (MVN), with parameters (0, aA); c is the vector of ran- son of mean EBV accuracies with the predicted accuracy dom litter effects with the distribution assumed to be of phenotypic selection in all breeds. Firstly, the mean 2 MVN, with parameters (0, σ cIlitter), and e is the vector EBV accuracy of phenotyped animals born in 2010 (with 2 of residuals distributed MVN with parameters (0, σ eI). no progeny phenotypes) was compared to the accuracy I represents an identity matrix of an appropriate size, of phenotypic selection (h, [24]). Secondly, mean accur- A is the additive genetic relationship matrix and σ2 de- acy of EBV for animals born in 2011 (<365 days old and notes the variance of each of the respective random therefore without a phenotype), but for which both

175 Lewis et al. BMC Genetics 2013, 14:16 Page 4 of 11 http://www.biomedcentral.com/1471-2156/14/16

parental phenotypes were available, was compared to the Results accuracy of selection using these phenotypes (√(½).h, see Hips Additional file 3) to determine any potential improve- An average of between 6% (GSD) and 19% (GDN) of all ment in the response to selection of breeding animals dogs registered annually since 1990 had been hip scored. prior to obtaining their own scores. Finally, the propor- The rate of scoring is higher for breeding animals, with tion of animals born in 2011 (so without a phenotype) the mean percentage of breeding animals born annually with EBV accuracy exceeding √(½).h was calculated and since 1990 having undergone hip scoring ranging from compared to the proportion where both parental pheno- 27% of sires and 28% of dams (AKT) to 80% of sires types were available. (GDN) and 86% of dams (BMD), Figure 1. There was considerable variation in the distribution of total un- Assessment of genetic gain to date transformed hip scores (Figure 2 and Additional file 1: The genetic gain as a proportion of genetic standard Table S1), with mean hip score ranging from 7.89 deviation was calculated as: (mean EBVmaxyr-mean (SHUSK) to 23.35 (NEWF), mode from 6 to 10, median EBVminyr)/ σA. For hip score minyr = 1990, and for elbow from 8 to 14, and standard deviation from 4.38 (SHUSK) score minyr = 2000; maxyr = 2011 for both traits. The to 20.49 (NEWF). All distributions were highly skewed, trends in genetic disposition to hip/elbow score were with coefficient of skewness ranging from 1.46 (NEWF) discerned for each breed via regression of EBVs on date to 4.59 (FCR), reflecting the cumulative nature of the of birth, and intensity of selection (i) applied estimated scoring system [29]. by rearrangement of the following equation: The results of the analyses determined that the BEARD displayed the smallest phenotypic variation (0.219) in log

2 transformed total hip score and the NEWF the largest ΔG ¼ ih σP=L (0.605, Table 1). The FCR exhibited the smallest degree of additive genetic variation (0.073) of log transformed total where ΔG is the genetic trend determined by regression hip score and the NEWF the largest (0.279). Estimates of 2 of EBV on date of birth, h is the heritability, σP is the heritability of log transformed total hip score ranged from phenotypic standard deviation, and L is the generation 0.28 (FCR) to 0.48 (SHUSK). Estimates of litter variance interval. as a proportion of phenotypic variance (not shown)

Figure 1 Average proportion of breeding animals hip scored. Mean proportion of male and female breeding animals born annually from 1990–2010 that are hip scored for all 15 breeds.

176 Lewis et al. BMC Genetics 2013, 14:16 Page 5 of 11 http://www.biomedcentral.com/1471-2156/14/16

Figure 2 Hip score distribution for Newfoundland and Siberian Husky. Distribution of total hip score for the Newfoundland (top) and Siberian Husky (bottom) breeds, from dogs evaluated between 1990–2011 and 365–1459 days old.

Table 1 Parameter estimates of hip score ranged from 0.017 (AKT) to 0.141 (GDN), although litter 2 2 2 Breed σ P σ A h s.e. was not a significant effect in all models. Meta-analysis of AKT 0.478 0.187 0.39 0.053 estimates of heritability of hip score across the 15 breeds indicated only a small degree of heterogeneity among BEARD 0.219 0.100 0.46 0.048 breeds, with a mean estimate of heritability across breeds BORD 0.223 0.098 0.44 0.033 of 0.38 (s. e. 0.014). The estimate of variance of between ENG 0.295 0.104 0.35 0.049 breed heritability estimates was 1.8 x 10-3. FCR 0.257 0.073 0.28 0.032 Regression of EBV on date of birth showed recent im- GDN 0.450 0.194 0.43 0.062 proving (negative) genetic trends significantly different to NEWF 0.605 0.279 0.46 0.041 zero (P < 0.01) in all cases except that of the SHUSK, where the genetic disposition towards higher (unfavour- RR 0.445 0.146 0.33 0.048 able) hip score, while still determinable (P < 0.01), in- SHUSK 0.349 0.167 0.48 0.038 creased at a rate of 0.8% per year (Table 2). Those breeds TT 0.246 0.084 0.34 0.048 showing an improving genetic trend ranged from a decline BMD 0.355 0.129 0.36 0.040 in genetic propensity toward hip score of −0.13% per year GR 0.313 0.126 0.40 0.017 (FCR) to −1.98% per year (NEWF) on the untransformed GSD 0.390 0.138 0.35 0.015 scale. However, of those breeds showing an improving LAB 0.381 0.126 0.33 0.012 genetic trend the derived selection intensities are weak; equivalent to excluding between less than 2% (BEARD, ROTT 0.308 0.120 0.39 0.028 FCR and RR) and less than 18% (GDN) of the highest risk 2 2 Estimates of phenotypic and additive genetic variance (σ Pand σ A respectively) and heritability (h2, with standard error) of hip score for 15 breeds. The top panel animals from breeding. As a result the genetic progress shows parameters for 10 breeds derived from univariate analyses, while the made has been slow, with the difference in mean EBV bottom panel shows parameters for 5 breeds derived from bivariate analyses of from animals born in 1990 and 2011 equating to between hip and elbow score. Breed abbreviations: Akita [AKT], Bearded Collie [BEARD], Bernese Mountain Dog [BMD], Border Collie [BORD], English Setter [ENG], Flat only 0.12 (BEARD) and 0.82 (NEWF and GDN) of re- Coat Retriever [FCR], Gordon Setter [GDN], Golden Retriever [GR], German spective genetic standard deviations. Shepherd Dog [GSD], Labrador Retriever [LAB], Newfoundland [NEWF], Rottweiler [ROTT], Rhodesian Ridgeback [RR], Siberian Husky [SHUSK] and Tibetan The mean accuracies of EBV of phenotyped animals Terrier [TT]. born in 2010 were higher than the predicted accuracy of

177 Lewis et al. BMC Genetics 2013, 14:16 Page 6 of 11 http://www.biomedcentral.com/1471-2156/14/16

Table 2 Estimates of genetic progress and selection animals for which such phenotypes were actually avail- pressure for hip and elbow score able. The increment ranged from 2% (from 92.8% with -2 Hips Progress / σA b (x10 )ip excluded both parental phenotypes to 94.8% with EBV accuracy AKT −0.28 −0.66 −0.08 <0.04 > √½.h; ENG), to an increase of over three-fold (from BEARD −0.12 −0.16 −0.04 <0.02 19.5% with both parental phenotypes to 59.7% with EBV accuracy > √½.h; AKT). In some cases this jump was not BORD −0.36 −0.63 −0.13 <0.07 particularly large, ENG and GDN for example have in- − − − ENG 0.67 1.07 0.24 <0.13 crements of just 2% and 6% respectively, but in these FCR −0.17 −0.13 −0.04 <0.02 cases the increment in actual (mean) EBV accuracy com- GDN −0.82 −1.95 −0.32 <0.18 pared to √½.h is large (47% and 32% respectively). NEWF −0.82 −2.00 −0.22 <0.12 RR −0.19 −0.32 −0.06 <0.02 Elbows Since 2000 between 1% (GR, GSD, LAB) and 15% SHUSK 0.25 0.81 0.12 N/A (BMD) of all registered dogs of the 5 relevant breeds − − − TT 0.36 0.52 0.13 <0.06 have been elbow scored. The rate of scoring is higher for − − − BMD 0.30 0.68 0.12 <0.06 breeding animals, with the mean percentage of breeding GR −0.71 −1.20 −0.23 <0.13 animals born annually since 2000 having undergone GSD −0.48 −0.89 −0.16 <0.08 elbow scoring ranging from 8% of sires and 7% of dams LAB −0.77 −1.28 −0.28 <0.16 (ROTT) to 66% of sires and 77% of dams (BMD). There ROTT −0.59 −0.78 −0.14 <0.07 was variation in the distribution of untransformed elbow scores with mean elbow score ranging from 0.15 (LAB) Elbows to 0.61 (ROTT), standard deviation from 0.46 (LAB) to BMD −0.20 −0.72 −0.11 <0.06 0.87 (BMD) and coefficient of skewness from 0.92 − − − GR 0.13 0.31 0.09 <0.04 (ROTT) to 3.59 (LAB) (Additional file 4: Table S2). The GSD −0.14 −0.21 −0.14 <0.07 LAB displayed the smallest phenotypic variation (0.196) LAB −0.13 −0.18 −0.12 <0.06 and additive genetic variation (0.037) in elbow score and ROTT −0.21 −0.39 −0.15 <0.08 the BMD the largest (0.760 and 0.201 respectively). Esti- mates of heritability of untransformed mean elbow score Genetic progress was estimated in two ways: total change as mean EBV2011- mean EBV1990 (for hips, EBV2011-mean EBV2000 for elbows) as proportion of ranged from 0.14 (ROTT) to 0.30 (GR) (Table 4). Meta- σ genetic standard deviation ( A) and annually by the regression coefficient analysis of estimates of heritability of elbow score across (b) of EBV on date of birth. Selection pressure was described in two ways: standardised selection intensity (i) against hip/elbow score, and the equivalent the 5 breeds indicated only a small degree of heterogen- proportion of breeding individuals excluded required to achieve that intensity eity, with an across-breed estimate of heritability of by truncation of the distribution. The top panel shows parameters for 10 breeds derived from univariate analyses of hip score and the middle panel 0.218 (s.e. 0.026). The estimate of variance of between shows hip parameters for 5 breeds derived from bivariate analyses of hip and breed heritability estimates was similar to but smaller elbow score. The bottom panel shows elbow parameters for 5 breeds derived than that for hip score at 0.8 x 10-3. Estimates of litter from bivariate analyses of hip and elbow score. Breed abbreviations: Akita [AKT], Bearded Collie [BEARD], Bernese Mountain Dog [BMD], Border Collie variance as a proportion of phenotypic variance (not [BORD], English Setter [ENG], Flat Coat Retriever [FCR], Gordon Setter [GDN], shown) ranged from 0.007 (BMD) to 0.146 (ROTT), al- Golden Retriever [GR], German Shepherd Dog [GSD], Labrador Retriever [LAB], Newfoundland [NEWF], Rottweiler [ROTT], Rhodesian Ridgeback [RR], Siberian though litter was not a significant effect in all models. Husky [SHUSK] and Tibetan Terrier [TT]. The genetic correlation between hip and elbow scores ranged from 0.005 (BMD) to 0.550 (ROTT). However, selection on phenotype (h) for all breeds, ranging from the genetic correlation between the two traits was only an improvement of 8% (BEARD and SHUSK) to 24% determinable as significantly different from zero in LAB (FCR) (Table 3). The mean accuracies of un-phenotyped (P < 0.001). The deviation of the correlation from zero in animals born in 2011 but with phenotyped parents were ROTT approached significance (P = 0.055), suggesting higher than the anticipated accuracy of selection on par- that more data may have increased the power to detect ental phenotypes for all breeds by between 18% significance. Meta-analysis of estimates of genetic correl- (SHUSK) and 47% (ENG). Importantly the anticipated ation between hip and elbow score across the 5 breeds accuracy of selection on parental phenotypes (√(½).h) is indicated a greater degree of heterogeneity among optimistic since it ignores all potential biases from fixed breeds than found with the heritabilities, with an across- effects and changes in the addititive genetic variance breed estimate of genetic correlation of 0.216 (s.e. over generations due to selection [30]. The proportion 0.076). The estimate of variance of between breed gen- of all animals registered in 2011 with EBV accuracies etic correlation estimates was 13.1 x 10-3. greater than that anticipated from selection on parental Regression of EBV on date of birth showed a recent phenotypes was always greater than the proportion of slow but significantly (P < 0.05) improving genetic trend

178 Lewis et al. BMC Genetics 2013, 14:16 Page 7 of 11 http://www.biomedcentral.com/1471-2156/14/16

Table 3 Increment in accuracy of selection for low hip score using EBV versus phenotype Animals with phenotype Animals with parental phenotype Proportion with r > √½.h h mean r n incr √½.h mean r n incr EBV pheno incr AKT 0.62 0.74 23 1.18 0.44 0.64 129 1.45 0.597 0.195 3.05 BEARD 0.68 0.73 26 1.08 0.48 0.61 324 1.29 0.923 0.806 1.15 BORD 0.66 0.74 98 1.11 0.47 0.58 910 1.23 0.745 0.583 1.28 ENG 0.59 0.72 18 1.21 0.42 0.62 180 1.47 0.948 0.928 1.02 FCR 0.53 0.66 54 1.24 0.38 0.53 1067 1.39 0.998 0.866 1.15 GDN 0.66 0.73 30 1.12 0.46 0.61 145 1.32 0.928 0.873 1.06 NEWF 0.68 0.75 37 1.11 0.48 0.58 441 1.21 0.829 0.697 1.19 RR 0.57 0.66 20 1.16 0.40 0.52 521 1.28 0.844 0.502 1.68 SHUSK 0.69 0.74 12 1.08 0.49 0.58 288 1.18 0.478 0.209 2.29 TT 0.58 0.70 45 1.19 0.41 0.55 712 1.34 0.928 0.736 1.26 BMD 0.60 0.71 48 1.19 0.43 0.56 402 1.31 0.893 0.754 1.18 GR 0.63 0.73 277 1.15 0.45 0.54 5097 1.21 0.860 0.791 1.09 GSD 0.59 0.69 337 1.15 0.42 0.51 3343 1.21 0.571 0.441 1.29 LAB 0.57 0.70 1004 1.21 0.41 0.52 16160 1.28 0.685 0.494 1.39 ROTT 0.63 0.73 51 1.17 0.44 0.59 565 1.34 0.568 0.361 1.57 Mean 1.16 1.30 1.44 (Left panel) The mean accuracy (r) of EBV of phenotyped animals born in 2010 compared to accuracy of phenotypic selection (h), with the sample size (n) and increment in accuracy (incr). (Middle panel) The mean accuracy of EBV of unphenotyped animals born in 2011, but with parental phenotypes, compared to the accuracy of selection on parental phenotypes (√(½).h). (Right panel) The proportion of unphenotyped animals born in 2011 with EBV accuracy exceeding √(½).h (EBV) compared to the proportion of 2011 born animals with parental phenotypes available (pheno). The top panel utilised parameters for 10 breeds derived from univariate analyses, while the bottom panel utilised parameters for 5 breeds derived from bivariate analyses of hip and elbow score. Increments calculated prior to rounding. Breed abbreviations: Akita [AKT], Bearded Collie [BEARD], Bernese Mountain Dog [BMD], Border Collie [BORD], English Setter [ENG], Flat Coat Retriever [FCR], Gordon Setter [GDN], Golden Retriever [GR], German Shepherd Dog [GSD], Labrador Retriever [LAB], Newfoundland [NEWF], Rottweiler [ROTT], Rhodesian Ridgeback [RR], Siberian Husky [SHUSK] and Tibetan Terrier [TT].

(Table 2) in all 5 breeds, ranging from a decline in ge- The mean accuracies of EBV of phenotyped animals netic propensity toward elbow score of between −0.18% born in 2010 were higher than the predicted accuracy of per year (LAB) to −0.72% per year (BMD). The derived selection on phenotype (h) for all breeds, ranging from selection intensities were very weak; equivalent to ex- an improvement of 17% (GR) to 52% (ROTT) (Table 5). cluding between only less than 4-8% of the highest risk The mean accuracies of un-phenotyped animals born in animals from breeding. As a result the genetic progress 2011 but with phenotyped parents were similarly greater made has been slow, with the difference in mean EBV than the anticipated accuracy of selection on parental from animals born in 2000 and 2011 equating to be- phenotypes by between 23% (GR) and 71% (ROTT). The tween only 0.13 (LAB) and 0.21 (ROTT) of respective proportion of all animals registered in 2011 with EBV genetic standard deviations. accuracies greater than that anticipated from selection on parental phenotypes was greater than the proportion of animals for which both parental phenotypes were ac- Table 4 Parameter estimates of elbow score tually available in all 5 breeds, the increment ranging 2 2 2 from 23% (from 72.2% with both parental phenotypes to σP σA h s.e. rA s.e. rE s.e. 88.9% with EBV accuracy > √½.h; BMD) to a greater than BMD 0.760 0.201 0.26 0.054 0.005 0.134 0.122 0.051 10-fold increase (from 6% with both parental phenotypes GR 0.278 0.084 0.30 0.054 0.137 0.098 0.095 0.050 to 79.5% with EBV accuracy > √½.h; ROTT). GSD 0.265 0.048 0.18 0.062 0.203 0.140 −0.054 0.055 LAB 0.196 0.037 0.19 0.028 0.344 0.064 −0.003 0.024 The effect of inbreeding ROTT 0.533 0.073 0.14 0.106 0.550 0.299 −0.091 0.091 The effects of inbreeding coefficient were typically very

2 2 Estimates of phenotypic and genetic variance (σ Pand σ A respectively) and small and not significantly different to zero in all cases, 2 heritability (h ) of elbow score and genetic and residual correlations (rA and rE except on hip score in the RR (−0.69, s.e. = 0.350) and respectively, with standard errors) with hip score for 5 breeds. Breed abbreviations: Bernese Mountain Dog [BMD], Golden Retriever [GR], German on elbow score in the GR (0.83, s.e. = 0.316). In the RR Shepherd Dog [GSD], Labrador Retriever [LAB], Rottweiler [ROTT]. this corresponds to a decline of 0.75 points for the

179 Lewis et al. BMC Genetics 2013, 14:16 Page 8 of 11 http://www.biomedcentral.com/1471-2156/14/16

Table 5 Increment in accuracy of selection for low elbow score using EBV versus phenotype Animals with phenotype Animals with parental phenotype Proportion with r > √½h h mean r n incr √½ h mean r n incr EBV pheno incr BMD 0.51 0.66 46 1.28 0.36 0.52 385 1.43 0.889 0.722 1.23 GR 0.55 0.64 136 1.17 0.39 0.48 959 1.23 0.385 0.149 2.59 GSD 0.42 0.51 197 1.21 0.30 0.38 535 1.26 0.272 0.071 3.85 LAB 0.43 0.59 579 1.37 0.31 0.45 3411 1.45 0.600 0.104 5.76 ROTT 0.37 0.56 28 1.52 0.26 0.45 95 1.71 0.795 0.061 13.09 Mean 1.23 1.34 3.14 (Left panel) The mean accuracy (r) of EBV of phenotyped animals born in 2010 compared to accuracy of phenotypic selection (h), with the sample size (n) and increment in accuracy (incr). (Middle panel) The mean accuracy of EBV of unphenotyped animals born in 2011, but with parental phenotypes, compared to the accuracy of selection on parental phenotypes (√(½).h). (Right panel) The proportion of unphenotyped animals born in 2011 with EBV accuracy exceeding √(½).h (EBV) compared to the proportion of 2011 born animals with parental phenotypes available (pheno). Increments calculated prior to rounding. Breed abbreviations: Bernese Mountain Dog [BMD], Golden Retriever [GR], German Shepherd Dog [GSD], Labrador Retriever [LAB], Rottweiler [ROTT]. median hip score of 8 (or a 4.24 point decrease from a known). Using EBV owners of breeding bitches would hip score of 50) comparing coefficient of inbreeding of be able to more accurately assess the genetic merit of 0.125 to 0 (values obtained for offspring of a half-sib and potential sires resulting in an improved response to se- unrelated matings respectively). In the GR there is an in- lection, whether phenotypes are available or not. In crease of 0.1 points in elbow score comparing coefficient addition, EBV will be available for all registered animals of inbreeding of 0.125 to 0. of the breed, increasing selection intensity opportunities. For example: the projected time to achieve an improve- Discussion ment of 5 points in the median hip score via phenotypic The results from this study demonstrate the potential selection, under the guidelines which were in place for power of EBV to improve the predicted accuracy of the majority of the period covered by the data, range selection against hip and elbow dysplasia in many dog from 30 to over 300 years (NEWF and BEARD respec- breeds in the UK, including 3 of the 10 most popular tively) mainly due to weak selection intensity [8]. Al- breeds. The mean accuracies of EBV are always higher though these guidelines have now been amended to than would be obtained via selection on available pheno- promote selection from below the median rather than types (using either individual or both parental pheno- the mean phenotype, the opportunity to increase selec- types). Furthermore, a far greater proportion of juvenile tion intensity is more readily presented by EBV (their animals have EBV with a higher accuracy than can at universality within a breed removing the random sam- present be obtained by selection using both parental pling of genetic risk from the use of un-scored animals). phenotypes. Thus, reliable information is available on Selecting breeding stock with EBV below the breed much more of the population than currently exists, mean is projected to achieve such an improvement in which will allow breeders to make more accurate selec- between 9 years for NEWF and 18 years for BEARD. tions earlier in the life of the dog. The accuracy of selec- The increases in the proportion of animals with breeding tion is directly linked to genetic progress, meaning more value accuracies greater than that provided by parental accurate selection will lead to greater progress in breed- phenotypes illustrate that EBV provide, per phenotype, ing for health. We have demonstrated this to be the case more information on more animals, enabling wider com- in a wide range of breeds. The broader impact can be parison by breeders. An additional benefit from publish- realised by noting that the average annual number of ing EBV could be the indirect introduction of selection registrations of the 15 breeds included in this study, and pressure through potential pet owners more accurately so that will each have an EBV, is in excess of 80,000, ap- differentiating the genetic risk of hip (and elbow) dyspla- proximately 1/3 of all annual registrations with the UK sia among available litters. Kennel Club. It is crucial however that participation in the BVA/KC Substantially faster genetic improvement is expected screening schemes continues – the availability of EBV to come via both increased accuracy and greater selec- does not mean scoring is no longer necessary. Pheno- tion intensity, as the provision of EBV could have a types are the basis of accurate breeding values, and ac- major impact on the ways in which dogs are selected by curacies will rapidly decline if phenotypic information breeders and pet owners (accompanied by appropriate were to become sparse. Theory predicts that EBV accu- user information). Currently mate selection is based on racy would be expected to increase with participation, ancestral phenotypes and two dogs’ own phenotypes (if and a plot and regression of mean EBV accuracy at birth

180 Lewis et al. BMC Genetics 2013, 14:16 Page 9 of 11 http://www.biomedcentral.com/1471-2156/14/16

on the mean annual proportion of sires with hip scores parameter. The slightly higher estimate of between breed provides empirical support (Additional file 5: Figure S1). variance of heritability for hip score compared to that for Experience in livestock sectors reinforces the theory, elbow score may reflect the greater number of breeds in- where widespread and routine data collection and very cluded in the analysis for that trait, and inclusion of add- large family sizes (i.e. thousands of progeny) can yield itional breeds not currently in the sample may prove an EBV accuracies of >0.9, although, it must be noted, ac- outlier to this current collection. Nevertheless, results curacies are rarely so high at the time of selection. The from the meta-analysis suggest that the heritability of both resulting message to breeders is simple: continued scor- hip and elbow score are remarkably consistent across ing will maintain and further enhance the accuracy of breeds, and that most of the observable variation in esti- selection of breeding stock for healthy joints, as well as mates is due to sampling variation. The across breed esti- increasing the pool of animals with reliable information. mate of the residual correlation between hip and elbow Moreover, the phenotypic score is of value to breeders score is small (with a small s.e., 0.024 ± 0.035), and the and pet owners alike in providing an indicator of not meta-analysis revealed only small between breed variation only the genetics but the environmental influence on an in such estimates (Additional file 6: Table S3). This implies individual animal’s hip/elbow joints. While the EBV that across breeds there is a large degree of independence should guide breeding decisions, the phenotype is useful in non-genetic environmental risk factors on dysplasia of to inform the appropriate care of the dog that may the hip and elbow joint. This finding across multiple ameliorate the severity of hip and elbow dysplasia where breeds supports an earlier observation on the small envi- it occurs. ronmental correlation between hip and elbow score in The accumulation of phenotypes will be particularly LAB [17] and is somewhat surprising given that both dys- critical for future analyses of elbow dysplasia, where the plasias are developmental orthopaedic diseases. extent of recording is much less than for hip dysplasia, All breeds included in this study showed an improving and since elbow score is less heritable than hip score genetic trend with respect to hip and elbow score, except (possibly due in part to the collection of traits described the SHUSK, suggesting that phenotypic selection to date by the elbow score). This study only managed to detect has had a small but beneficial impact. The increasing a genetic correlation between hip and elbow scores with genetic propensity towards hip dysplasia in the SHUSK enough precision to be statistically significantly different was matched by the phenotypic trend (regression of total to zero in the LAB. Previously, we demonstrated that bi- hip score on date of birth showing a yearly rise of 0.075 variate analysis of hip and elbow data can confer signifi- score points), which has been observed previously [31]. cant benefits to the accuracy of EBV for elbow scores, However, the SHUSK had the best hip scores of all the where a favourable genetic correlation exists [17]. Add- 15 breeds analysed here. It may be that the historical itional elbow score data will be essential to determine role of the SHUSK as a sled dog has entailed de facto more precise genetic correlations between hip and elbow selection against lameness, but that increasing popularity score in BMD, GR, GSD and ROTT, although reported as pets or show dogs has weakened this tacit selection estimates from other studies indicate there may be wide pressure. The popularity of the breed in the UK has variation across breeds [9,14,16]. risen quickly recently, from 829 registered in 2000 to While genetic parameters are often (correctly) viewed 2,209 in 2010. While the general hip condition of the as specific to each breed, questions can arise as to SHUSK remains better than for many other breeds, 2 whether the genetic parameters (h and rA) from one breeders should be aware of the detrimental trend. It breed may be useful in BLUP analyses (EBV calculation) serves as an example that the transition to a popular pet of another. This is particularly relevant where small breed be accompanied by tools, such as EBV, that pro- population size means that breed-by-breed parameter tect the qualities of the breed for which it is valued. calculation is not feasible. The analysis of 15 breeds in this The results presented here indicate that the GDN has study using the same model provided a good opportunity been subject to the greatest selection intensity for reduc- to explore this matter. Results from the meta-analysis tion in hip score, equivalent to excluding the 18% of ani- indicate that there is more between breed variation in esti- mals with the worst hip scores from breeding. This is in mates of genetic correlation between hip and elbow score line with former breeding guidelines based on the mean than for heritability of elbow score, across the five breeds hip score and has been accompanied by a phenotypic for which both traits were analysed. While additional decline in hip score of over 0.6 points per year (from elbow scoring data will therefore be expected to result in regression of total hip score on date of birth) and a fall more consistent estimates of heritability across breeds as in the mean hip score from 24.35 in 1990 to 14.77 in sampling variance is reduced, the estimates of genetic 2010. The GDN is not a numerous breed, with a mean correlation between hip and elbow score are expected to of 324 dogs registered per year from 2000–2010, but ap- reflect the greater between breed variation in the true pears to have a large proportion of breeders committed

181 Lewis et al. BMC Genetics 2013, 14:16 Page 10 of 11 http://www.biomedcentral.com/1471-2156/14/16

to including health traits in selection objectives; for ex- through deterministic use of empirical data can then be ample over 80% of sires and dams undergo hip scoring. used in induction. The reported superiority of mean While slightly greater genetic progress was observed in EBV accuracies over the accuracy of selection on indi- the NEWF, a larger estimate of heritability and shorter vidual hip score phenotype reported here were smaller generation interval meant that the derived selection in- than reported by Malm et al. using simulation [23], how- tensity was smaller than for the GDN. However across ever there tended to be fewer animals with phenotypes all breeds and traits, regression of genetic gain on the in our data, implying less information. Comparison of proportion of breeding animals scored did not show sig- EBV accuracy with selection on parental phenotypes nificant association (P > 0.05). This demonstrates that shows the improvement was of similar magnitude. quantity of data alone does not guarantee genetic im- EBV for hip and elbow dysplasia are routinely com- provement, but that it must be accompanied by the appro- puted and published in Norway, Finland and Denmark priate breeding advice and the motivation by breeders to for up to 38 breeds and in Sweden for 5 breeds (K Maki, act upon it. Across comparable breeds, the rates of genetic personal communication), in Germany for GSD, and in progress calculated in this study were broadly typical of the USA for LAB. The public release of EBV described those that have been previously reported [16]. in this study is anticipated in the UK in 2013. The abun- Substantial improvements in the predicted accuracy of dance of EBV for hip and elbow dysplasia in so many selection, and therefore genetic progress, based on esti- countries raises the prospect of the globalisation of scor- mating breeding values have been quantifiably demon- ing and evaluation schemes. Analyses determining the strated here for a wide range of breeds, including a genetic correlations between individual scoring protocols number of the more uncommon breeds. For the more would enable dogs to be evaluated under any (participat- uncommon breeds, selection against diseases such as hip ing) scheme (UK registered dogs evaluated under the dysplasia is more problematic when based on pheno- FCI scheme and Scandinavian dogs participating in the types alone as there may be only a small number of the BVA/KC scheme for example) while still having an EBV candidates with a record, and so making a small breed in the country of registration [25]. It should be noted, smaller. Therefore an approach to increasing numbers of however, that not all scoring protocols may be equal in candidates with usable information, as demonstrated terms of predicting the lameness associated with hip and here, should be welcome. Rarer breeds are more likely to elbow dysplasia and consequential OA [33]. To address suffer the effects of genetic over-contribution of some this further research focussing on identifying OA and animals to future generations, usually through the wide- lameness later in the life of scored dogs would be wel- spread use of popular sires. Where selection does take come. Fortunately, the manner in which EBV for canine place in small populations (which it must do to improve health are presented offers an ‘outward continuity,allowing’ welfare where hip dysplasia is prevalent, as argued in the improvements to be made to the computational model or introduction) a balance must be struck between genetic to the evaluation protocol, as well as the utilisation of progress in reducing the burden of disease on the one international data, without noticeable disruption to the hand, and minimising the risk of the emergence of a end user [25]. novel genetic disease on the other, which can be mea- sured by the rate of inbreeding. The inbreeding coeffi- cient per se was found to be largely unrelated to, and Conclusion have only a small effect on, hip and elbow score in this The use of EBV by dog breeders is projected to facilitate study. However, one drawback with the use of EBV considerable improvements in the response to selection for based on pedigrees and phenotypes is that they too can healthier hip and elbow joints in a wide range of breeds, promote greater rates of inbreeding in the course of gen- through both enhanced accuracy and greater abundance of erating more progress [32]. This need not be inevitable, information. Across the 15 breeds analysed here estimates but instead places an emphasis on increasing awareness of heritability of hip and elbow score were remarkably con- of inbreeding among breeders, and making more tools sistent, and phenotypic selection has been successful in available to help them manage rates of inbreeding as eliciting genetic progress, albeit very slowly, in all breeds EBV are introduced. except the SHUSK. However, substantial improvement in In this study we elected to conduct a deterministic the accuracy of selection via use of EBV was demonstrated prediction of the superiority of EBV accuracy over that across all breeds, for both dogs with and without a pheno- of selection using phenotype. An alternative method type. The availability of EBV for hip score for 15 UK regis- would be to use simulation. However, simulations are tered pedigree dog breeds will provide information on the stochastic and can be prone to error in some situations. genetic risk of disease in nearly a third of all dogs annually A further disadvantage of simulation is a lack of insight registered by the UK KC, with in excess of a quarter having into the underlying causes, which when encountered an EBV for elbow score as well.

182 Lewis et al. BMC Genetics 2013, 14:16 Page 11 of 11 http://www.biomedcentral.com/1471-2156/14/16

Additional files 12. Bishop SC, Woolliams JA: On the genetic interpretation of disease data. PLoS One 2010, 5(1):e8940. 13. Beuing R, et al: Prevalence and inheritance of canine elbow dysplasia in Additional file 1: Table S1. Summary statistics of hip scores of all 15 German Rottweiler. J Anim Breed Genet 2000, 117(6):375–383. breeds. 14. Maki K, Groen AF, Liinamo AE, Ojala M: Genetic variances, trends and Additional file 2: Appendix 1. Derivation of accuracy of breeding mode of inheritance for hip and elbow dysplasia in Finnish dog values including F. populations. Anim Sci 2002, 75:197–207. Additional file 3: Appendix 2. Derivation of accuracy of mass 15. Janutta V, et al: Genetic analysis of three different classification protocols (phenotypic) selection. for the evaluation of elbow dysplasia in German shepherd dogs. J Small Anim Pract 2006, 47(2):75–82. Additional file 4: Table S2. Summary statistics of elbow scores for 5 16. Malm S, et al: Genetic variation and genetic trends in hip and elbow breeds. dysplasia in Swedish Rottweiler and Bernese Mountain Dog. J Anim Breed Additional file 5: Figure S1. Plot of EBV accuracy on proportion of Genet 2008, 125(6):403–12. sires with phenotypes. 17. Lewis TW, et al: Genetic evaluation of elbow scores and relationship with Additional file 6: Table S3. Summary of meta-analysis. hip scores in UK Labrador retrievers. Vet J 2011, 189:227–233. 18. Silvestre AM, et al: Comparison of estimates of hip dysplasia genetic parameters in Estrela Mountain Dog using linear and threshold models. Competing interests J Anim Sci 2007, 85(8):1880–4. TWL is fully funded and SCB partly funded by the UK Kennel Club Charitable 19. Hou Y, et al: Retrospective analysis for genetic improvement of hip joints Trust. The funders had no role in study design, data analysis, decision to of cohort Labrador retrievers in the United States: 1970–2007. PLoS One publish, or preparation of the manuscript. Hip and elbow score data and 2010, 5(2):e9410. pedigree was collated and provided by the UK Kennel Club. JAW declares no 20. Wilson BJ, et al: Heritability and phenotypic variation of canine hip competing interests. dysplasia radiographic traits in a cohort of Australian German Sheperd dogs. PLoS One 2012, 7(6):e39620. Author contributions 21. Ginja MM, et al: Diagnosis, genetic control and preventive management TWL, SCB & JAW conceived and designed the analyses; TWL performed the of canine hip dysplasia: a review. Vet J 2010, 184(3):269–76. analyses; TWL & JAW analysed the results; TWL, SCB & JAW wrote the paper; 22. Stock KF, Distl O: Simulation study on the effects of excluding offspring all authors read and approved the final manuscript. information for genetic evaluation versus using genomic markers for selection in dog breeding. J Anim Breed Genet 2010, 127:42–52. 23. Malm S, et al: Efficient selection against categorically scored hip dysplasia Acknowledgements in dogs is possible using best linear unbiased prediction and optimum The authors are grateful to the BVA hip and elbow scoring panellists for data contribution selection: a simulation study. J Anim Breed Genet 2012. http:// provided by their ongoing work, and to Dr M. Coffey and Dr K. Moore of onlinelibrary.wiley.com/doi/10.1111/j.1439-0388.2012.01013.x/abstract. EGENES at ’s Rural University College for provision of BLUP EBV on 24. Falconer DS, Mackay TFC: Introduction to Quantitative Genetics. 4th edition. the 3 most populous breeds. SCB and TWL gratefully acknowledge funding Longman: Edinburgh Gate, Harlow, Essex CM20 2JE; 1996. from the Kennel Club Charitable Trust. JAW gratefully acknowledges funding 25. Woolliams JA, Lewis TW, Blott SC: Canine hip and elbow dysplasia in UK from the BBSRC. Labrador retrievers. Vet J 2011, 189:169–176. 26. Gibbs C: The BVA/KC scoring scheme for control of hip dysplasia: Author details interpretation of criteria. Vet Rec 1997, 141(11):275–84. 1Kennel Club Genetics Centre at the Animal Health Trust, Lanwades Park, 2 27. Gilmour AR, et al (Eds): ASReml user guide release 3.0. UK: VSN International Kentford, Newmarket, Suffolk CB8 7UU, UK. The Roslin Institute and Royal Ltd, Hemel Hempstead, HP1 1ES; 2009. (Dick) School of Veterinary Studies, University of Edinburgh, Easter Bush 28. Corbin LJ, et al: Linkage disequilibrium and historical effective population Research Centre, Midlothian EH25 9RG, UK. size in the Thoroughbred horse. Anim Genet 2010, 41(Suppl 2):8–15. 29. Lewis TW, Woolliams JA, Blott SC: Genetic evaluation of the nine Received: 2 August 2012 Accepted: 25 February 2013 component features of Hip score in UK Labrador retrievers. PLoS One Published: 2 March 2013 2010, 5(10):e13610. 30. Bulmer MG: The effect of selection on genetic variability. Am Nat 1971, References 105(943):201–211. 1. Leighton EA: Genetics of canine hip dysplasia. J Am Vet Med Assoc 1997, 31. Willis MB: A review of the progress in canine hip dysplasia control in – 210(10):1474 9. Britain. J Am Vet Med Assoc 1997, 210(10):1480–2. 2. Asher L, et al: Inherited defects in pedigree dogs. Part 1: Disorders 32. Verrier E, Colleau JJ, Foulley JL: Long-term effects of selection based on – related to breed standards. Vet J 2009, 182:402 411. the animal model BLUP in a finite population. Theortetical and Applied 3. Schnelle GB: The veterinary radiologist: regional radiography - the pelvic Genetics 1993, 87:446–454. – region, Part 1. North Am Vet 1937, 18:53 56. 33. Wilson B, Nicholas FW, Thomson PC: Selection against canine hip 4. Schales O: Genetic aspects of dysplasia of the hip joint. North Am Vet dysplasia: Success or failure? Vet J 2011, 189:169–176. 1956, 38:476. 5. Henricson B, Olsson S-E: Hereditary acetabular dysplasia in German doi:10.1186/1471-2156-14-16 Shepherd dogs. J Am Vet Med Assoc 1959, 135:207–210. – Cite this article as: Lewis et al.: Comparative analyses of genetic trends 6. Brass W: Hip dysplasia in dogs. J Small Anim Pract 1989, 30:166 170. and prospects for selection against hip and elbow dysplasia in 15 UK 7. Maki K: Breeding against hip and elbow dysplasia in dogs.InPhD thesis. dog breeds. BMC Genetics 2013 14:16. University of Helsinki, Department of Animal Science; 2004. 8. Lewis TW, Blott SC, Woolliams JA: Genetic Evaluation of Hip Score in UK Labrador Retrievers. PLoS One 2010, 5(10):e12797. 9. Stock KF, et al: Genetic analyses of elbow and hip dysplasia in the German shepherd dog. J Anim Breed Genet 2011, 128:219–229. 10. Hodgman S: Abnormalities and defects in pedigree dogs 1. An investigation into the existence of abnormalities in pedigree dogs in the . Journal of small animal practice 1963, 4(6):447–456. 11. Hazewinkel HAW: Elbow Dysplasia, definition and known aetiologies. In 22nd annual meeting of the International Elbow Working Group. Munich, Germany:; 2007:6–17. http://www.vet-iewg.org/joomla/images/proceedings/ proceedings2007iewg.pdf (accessed 2 February 2011).

183 Prevalence of inherited disorders SMALL ANIMALS among mixed-breed and purebred dogs: 27,254 cases (1995–2010)

Thomas P. Bellumori, MS; Thomas R. Famula, PhD; Danika L. Bannasch, PhD, DVM; Janelle M. Belanger, MS; Anita M. Oberbauer, PhD

Objective—To determine the proportion of mixed-breed and purebred dogs with common genetic disorders. Design—Case-control study. Animals—27,254 dogs with an inherited disorder. Procedures—Electronic medical records were reviewed for 24 genetic disorders: hem- angiosarcoma, lymphoma, mast cell tumor, osteosarcoma, aortic stenosis, dilated cardio- myopathy, hypertrophic cardiomyopathy, mitral valve dysplasia, patent ductus arteriosus, ventricular septal defect, hyperadrenocorticism, hypoadrenocorticism, hypothyroidism, el- bow dysplasia, hip dysplasia, intervertebral disk disease, patellar luxation, ruptured cranial cruciate ligament, atopy or allergic dermatitis, bloat, cataracts, epilepsy, lens luxation, and portosystemic shunt. For each disorder, healthy controls matched for age, body weight, and sex to each affected dog were identified. Results—Genetic disorders differed in expression. No differences in expression of 13 ge- netic disorders were detected between purebred dogs and mixed-breed dogs (ie, hip dys- plasia, hypo- and hyperadrenocorticism, cancers, lens luxation, and patellar luxation). Pure- bred dogs were more likely to have 10 genetic disorders, including dilated cardiomyopathy, elbow dysplasia, cataracts, and hypothyroidism. Mixed-breed dogs had a greater probability of ruptured cranial cruciate ligament. Conclusions and Clinical Relevance—Prevalence of genetic disorders in both populations was related to the specific disorder. Recently derived breeds or those from similar lineages appeared to be more susceptible to certain disorders that affect all closely related pure- bred dogs, whereas disorders with equal prevalence in the 2 populations suggested that those disorders represented more ancient mutations that are widely spread through the dog population. Results provided insight on how breeding practices may reduce prevalence of a disorder. (J Am Vet Med Assoc 2013;242:1549–1555)

ogs are second only to humans in the number of ABBREVIATIONS Dhereditary diseases identified in the population.1 Information about the prevalence and etiology of dis- AKC American Kennel Club orders in dogs may provide insight into preventative CI Confidence interval measures and possible treatments for dogs with dis- IVDD Intervertebral disk disease eases as well as for humans sharing common disor- 2 ders. Although no single registry maintains a record influenced by recessive alleles in greater frequency than of genetic disease in dogs, it has been suggested that their crossbred counterparts. The common assump- purebred dogs are more prone to genetic disorders than 3 tion that a mixed-breed dog is healthier would not be are mixed-breed dogs. Breeding practices and selec- true if both parents carried deleterious mutations for tion pressures used by breeders of purebred dogs have the same disorder. Few data have been compiled to ac- been implicated in the perceived high frequency of ge- curately assess the question of whether purebred dogs netic disorders, whereas the random mating practices are at greater risk for genetic disorders, compared with of mixed-breed dogs have been suggested to increase 5 4 mixed-breed dogs. In a study of dogs affected with hip hybrid vigor (heterosis), resulting in healthier dogs. dysplasia, no significant difference in prevalence was The increased homozygosity expected in purebred observed between purebred and mixed-breed dogs. dogs offers the potential for these animals to have traits Domestic dogs are thought to be derived from 3 to 5 wolf lineages.6 Each lineage would be derived from a From the Department of Animal Science, College of Agricultural and few common ancestors; thus, one might expect some Environmental Sciences (Bellumori, Famula, Belanger, Oberbauer), disorders would be common to all dogs, regardless of and the Department of Population, Health and Reproduction, School breed. Genetic mutations that accompanied the domes- of Veterinary Medicine (Bannasch), University of California-Davis, Davis, CA 95616. tication process would be expected to be widely distrib- Address correspondence to Dr. Oberbauer (amoberbauer@ucdavis. uted throughout the dog population, affecting dogs of edu). any breed, including admixtures of breeds. In contrast

JAVMA, Vol 242, No. 11, June 1, 2013 Scientific Reports 1549 184 to more distant mutations, more recent selection pres- lowing categories were assessed: cancers (hemangiosar- sure (eg, in Europe during the Victorian era7) would coma, lymphoma, mast cell tumor, and osteosarcoma), influence the distribution of newer mutations, restrict- cardiac disorder (aortic stenosis, dilated cardiomyopathy, ing those to subsets of the overall dog population. It hypertrophic cardiomyopathy, mitral valve dysplasia, is likely that with breed refinement for specific tasks patent ductus arteriosus, and ventricular septal defect), and morphology, some mutations accompanied selec- endocrine disorders (hyperadrenocorticism, hypoad- tion for those traits. Rigorous selection pressures to re- renocorticism, and hypothyroidism), orthopedic dis- SMALL ANIMALS SMALL fine the breeds by inbreeding and bottlenecks4,8 would orders (elbow dysplasia, hip dysplasia, IVDD, patellar contribute to a loss of genetic diversity, thereby increas- luxation, and ruptured cranial cruciate ligament), and ing the likelihood of recessive disorders within a breed other (atopy or allergic dermatitis, bloat, cataracts, epi- population. lepsy, lens luxation, and portosystemic shunt). Mode The AKC registers purebred dogs and records of inheritance was not a factor in the selection of the ancestors. Although, in 2004, there were > 140 AKC- conditions under study. registered breeds, 10 breeds represented more than half of the reported AKC-registered dogs, whereas the 100 Medical records review—Patient records con- least popular breeds represented < 15% of all AKC reg- tained fields that included pertinent history, clinical istrations.9 The less popular breeds, with many fewer signs, clinical diagnosis, and other comments. Searches dogs registered each year, would be expected to have for keywords and any synonym or alternative represen- smaller effective gene pools. For example, the current tation for the genetic disorders were conducted in all population of Portuguese Water Dogs, ranked 56th fields. As an example, “Cushings,” “Cushing’s,” “Cush- in registrations as of 2011, has been traced back to a ing,” and “hyperadrenocorticism” were all keyword small number of dogs, mostly from 2 kennels, with ap- searches to extract data related to hyperadrenocorti- proximately 6 ancestors comprising 80% of the current cism. From each individual keyword search, a single gene pool.9 Breeds with smaller gene pools and reduced database of patients was created for each disorder. In genetic variation are more likely to phenotypically ex- addition to disorder status, patient identification num- press a recessive disorder.1 ber, breed, sex, species, body weight, date of birth, ad- Many studies have sought to describe the preva- missions date, discharge date, search-term field (eg, lence of disorders among individual breeds. Often, the pertinent history and clinical diagnoses), and keyword focus is on a single disorder and its inheritance pattern in context were captured. Each record was screened for in a particular breed to define possible mutations. Yet, accuracy, and only records with definitive confirmed more global studies designed to assess the proportion of diagnoses by the veterinary medical teaching hospi- mixed-breed and purebred dogs affected with heritable tal staff or the referring veterinarian were included for disorders can prove useful toward reducing the preva- analyses. Any record that referred to suspected diseases, lence of those disorders in the dog population. Describ- a presumptive diagnosis pending test results, rule-out ing disorders equivalently expressed within purebred diagnosis, or differential diagnosis or that included a and mixed-breed dogs may identify disorders com- diagnosis that was in any other way unconfirmed was mon in the overall population and suggest approaches omitted from analyses. For example, diagnoses of myx- to reduce the prevalence. In contrast, disorders more omatous mitral valvular disease were excluded from the prevalent to a particular breed may be reduced by use mitral valve dysplasia category. The sole exception was of concerted breeding practices. epilepsy, for which the disorder was classified into 1 of A recent study10 found a direct correlation between 3 categories (confirmed, probable, or suspect) on the disorders inherited in purebred dogs and the morpho- basis of the recorded information. Because of the nature logical characteristics specified in the . of the records explaining specific vertebral problems, Although that finding underscores the fact that pure- any dog with a laminectomy was considered to have bred dogs are considered at risk for disorders, it is un- IVDD, although laminectomy for cervical spondylomy- known whether mixed-breed dogs have the same risk elopathy was excluded. For each disorder, records were of genetic disorders that is suggested for purebred dogs. excluded such that only patients with a confirmed and The purpose of the study reported here was to describe reliable diagnosis of a particular disorder were retained. the prevalence of genetic disorders in the dog popula- Regardless of the number of visits, a given dog was tion as a whole. counted only once for a given disorder. To yield a com- parison of healthy or diseased dogs with dogs evaluated Materials and Methods at hospital for other reasons, a search for records of all dogs admitted after being hit by a car was also done. Case selection criteria—The data used in these The veterinary medical teaching hospital veteri- analyses were obtained by searching through the Uni- nary medical and administrative computer system was versity of California-Davis Veterinary Medical Teach- again searched to collect information on all of the dogs ing Hospital electronic records of all patients evaluated evaluated at the hospital from January 1, 1995, through from January 1, 1995, through January 1, 2010. The January 1, 2010. This data file contained all dogs evalu- genetic disorders selected for the study represented ated at the clinic, including those with and without the those expected to be present in the dog population at a disorders that were under study, yielding information measurable prevalence and to be debilitating, with con- for each of the 268,399 visits. Data from the confirmed fidence in the reliability of diagnosis. Additionally, dis- disorder files were matched to the full data file. In this orders that affected a variety of anatomic locations and way, individual patient records were matched so that all physiologic systems were chosen. Disorders in the fol- visit records for a single patient had the same diagno-

1550 Scientific Reports JAVMA, Vol 242, No. 11, June 1, 2013 185 sis and any patient that may not have had the disorder addition, by counting the number of data sets (of 50), listed for a specific visit was still classified as having the difference in disease risk between purebred and SMALL ANIMALS the disorder. A given dog could have been classified as mixed-breed dogs could be determined. having multiple disorders if > 1 disorder was confirmed All analyses were conducted via statistical softwarea via diagnostic evaluation. From this file containing all with a logit link function for analysis of the binomial unique dogs, control dogs were identified for use as variable of disease status. The model included terms for hospital controls in accordance with clinical research age class, weight class, and sex as well as a term for pure- designs.11 Specifically, none of the conditions under bred versus mixed-breed dog. Because each of the 50 study were diagnosed in these dogs. data sets was balanced for age, weight, and sex groups, Each patient had a breed designation. Dogs of the OR for any of these variables should be 1.0, and this AKC-recognized breeds, AKC miscellaneous breeds, was monitored in all analyses as a test of the sampling or Foundation Stock Service breeds were considered to process. The OR for purebred versus mixed-breed status be purebred dogs. All nondomesticated canine patients for each of the 50 data sets was saved, as were the lower (dingo or wolf) were removed. Pit bull–type dogs were and upper limits of the 95% CI for this estimate and its evaluated independently because of the inability to associated P value. Also counted were the number of validate purebred status. Any dog labeled as a mix was times (of 50 tests) the P value was less than or equal to considered to be a mixed-breed dog. From the records the commonly used type I error rate of 0.05. collected, age at each visit could be calculated. For each The number of dogs from each breed evaluated dog, the age of first recorded diagnosis at the veterinary at the veterinary medical teaching hospital was deter- medical teaching hospital for each disorder was calcu- mined as well as the number of dogs of each breed that lated and a mean age of first diagnosis was determined were defined as control (no disorder) or affected (hav- for each disorder. ing ≥ 1 disorder). The percentage of each breed that was control or affected was then calculated. Statistical analysis—For each disorder, appropri- ate population controls were identified from the com- Results plete data file containing all dogs evaluated at the vet- erinary medical teaching hospital in the 15-year time Of the 90,004 dogs examined at the veterinary frame. Because the number of dogs lacking a given con- medical teaching hospital small animal clinic that had dition far exceeded the number of dogs with the condi- an identified breed status (purebred, mixed, or pit bull– tion, to create the control population against which the type), 27,254 had ≥ 1 of the conditions under study dogs with the condition were compared, it was neces- and 62,750 were control dogs (Table 1). In terms of the sary to randomly sample the dogs lacking the condi- percentage of dogs of each breed with ≥ 1 disorder, 15 tion. Dogs were first stratified by body weight, sex, and breeds had < 20% of dogs with ≥ 1 disorder, 63 breeds age, and then each dog with a condition was matched had from 21% to 30%, 41 breeds had from 31% to 40%, to a randomly selected dog from the control group hav- and 10 breeds had > 40%. The mean age at the first visit ing the same weight, sex, and age classification. This (assessed as the first appointment at the hospital with sampling created control sets that represented the same a disorder diagnosis) was calculated for each disorder characteristics as the affected dogs except for breed sta- (Table 2). Patent ductus arteriosus and ventricular sep- tus. Control dogs were matched for age (0 to 2 years, tal defect were both diagnosed at a mean age of 1.32 > 2 to 7 years, or > 7 years), weight (0 to 12 kg [0 years. Hyperadrenocorticism was diagnosed at a mean to 26.4 lb], > 12 to 20 kg [26.4 to 44 lb], or > 20 kg age of 10.54 years, the oldest age of diagnosis for any [44 lb]), and sex (male, castrated male, female, or disorder. By comparison, dogs hit by a car had a mean spayed female) to each affected dog for each condi- age of 4.87 years. tion. The control dogs matched by the age, weight, and Of the 24 disorders assessed, 13 had no signifi- sex criteria were randomly selected from the complete cant difference in the mean proportion of purebred and data file, creating the control group for each disorder mixed-breed dogs with the disorder when matched for in accordance with clinical research designs.11 Thus, age, sex, and body weight (Table 2). Disorders without the controls were from the same population base from a significant predisposition included all the neoplasms which the dogs with disorders were derived. (hemangiosarcoma, lymphoma, mast cell tumor, and To enhance the reliability of the analyses, the sam- osteosarcoma), hypertrophic cardiomyopathy, mitral pling set of healthy control dogs was repeated 50 times valve dysplasia, patent ductus arteriosus, and ventricular for each condition investigated. That is, for any given septal defect in the cardiac category; hip dysplasia and condition, an equal number of healthy dogs, stratified patellar luxation in the orthopedic category; hypoadre- by the age, body weight, and sex of the affected dogs, were randomly selected 50 times to create repeated control data sets matched to the affected dogs. In this Table 1—Breed distribution of dogs with (Condition) and with- out (Control) inherited disorders evaluated at the Veterinary manner, the sole variable between the 50 randomly cre- Medical Teaching Hospital, University of California-Davis, in a ated data sets representing the control population was 15-year period. the number of mixed-breed or purebred dogs. In this way, 50 estimates (1 from each randomly selected set Breed Control Condition Total of controls) of the OR for the comparison of purebred Purebred 45,015 20,937 65,952 (73.3%) with mixed-breed dogs as well as the mean 95% CI of Mixed 16,693 5,990 22,683 (25.2%) Pit bull–type 1,042 327 1,369 (1.5%) this ratio and the mean P value used to test this ratio Total 62,750 27,254 90,004 (100%) against the null hypothesis of 1.0 were calculated. In

JAVMA, Vol 242, No. 11, June 1, 2013 Scientific Reports 1551 186 Table 2—Distribution and descriptive statistics of mixed-breed and purebred dogs with inherited conditions diagnosed over a 15-year period.

Mean age No. of times Mixed Purebred Total at first Mean breed was Disorder or injury (No. of dogs) (No. of dogs) (No. of dogs) diagnosis (y) Mean OR (95% CI) P value significant Cardiac Aortic stenosis* 33 357 390 3.0 3.03 (1.96–4.76) 0.000 50 SMALL ANIMALS SMALL Dilated cardiomyopathy* 32 329 361 7.23 3.45 (2.22–5.26) 0.000 50 Hypertrophic cardiomyopathy 3 33 36 6.51 2.04 (0.40–10.0) 0.336 9 Mitral valve dysplasia 40 180 220 4.09 1.85 (0.73–1.96) 0.446 5 Patent ductus arteriosus 81 329 410 1.32 0.85 (0.60–1.22) 0.480 3 Ventricular septal defect 16 117 133 1.32 1.72 (0.86–3.45) 0.168 15 Cancer Hemangiosarcoma 135 427 562 9.19 1.25 (0.95–1.64) 0.186 17 Lymphoma 392 1,182 1,574 8.0 1.11 (0.94–1.30) 0.271 8 Mast cell tumor 342 1,105 1,447 8.0 1.20 (1.01–1.43) 0.068 32 Osteosarcoma 187 522 709 8.23 1.09 (0.86–1.39) 0.449 3 Orthopedic Elbow dysplasia* 191 1,034 1,225 3.54 2.00 (1.63–2.50) 0.000 50 Hip dysplasia 500 1,431 1,931 3.89 1.05 (0.91–1.23) 0.473 4 IVDD* 833 3,658 4,491 7.35 1.41 (1.26–1.56) 0.000 50 Patellar luxation 466 1,710 2,176 5.16 1.04 (0.90–1.20) 0.490 0 Ruptured cranial cruciate ligament† 400 828 1,228 5.95 0.79 (0.67–0.94) 0.031 41 Endocrine Hyperadrenocorticism 281 808 1,089 10.54 1.02 (0.84–1.23) 0.593 0 Hypoadrenocorticism 67 228 295 8.72 1.23 (0.83–1.79) 0.354 5 Hypothyroidism* 326 1,369 1,695 6.86 1.56 (1.33–1.85) 0.000 50 Other Atopy or allergic dermatitis* 237 1,094 1,331 5.95 1.56 (1.30–1.89) 0.003 50 Bloat* 35 187 222 6.92 1.79 (1.10–2.94) 0.054 36 Cataracts* 734 2,822 3,556 9.21 1.27 (1.12–1.41) 0.000 50 Epilepsy total* 188 749 937 6.24 1.37 (1.10–1.69) 0.016 47 Epilepsy confirmed 146 565 711 6.57 1.33 (1.03–1.79) 0.062 28 Epilepsy probable 24 120 144 5.26 1.61 (0.88–2.94) 0.158 13 Epilepsy suspect 18 64 82 5.32 1.03 (0.48–2.22) 0.536 1 Lens luxation 64 251 315 9.07 1.14 (0.78–1.69) 0.478 2 Portosystemic shunt* 74 608 682 2.39 2.04 (1.49–2.77) 0.000 50 Hit by car† 569 1,069 1,638 4.87 0.59 (0.51–0.69) 0.000 50 Mean P value indicates comparison of purebred dogs with matched control sampling sets. Number of times breed was significant = Number of times (of 50) that comparison of affected dogs with matched control sampling sets indicated a significant (P < 0.05) difference in probability that mixed-breed and purebred categories differed in expression of the condition. Mean OR (95% CI) indicates comparison of purebred dogs relative to mixed-breed dogs. *Purebred dogs had a greater probability of expressing the condition. †Mixed breeds had a greater probability of expressing the condition. Epilepsy total consists of the sum of all 3 categories of epilepsy.

nocorticism and hyperadrenocorticism in the endocrine Ten genetic disorders had a significantly greater category; and lens luxation in the other category. probability of being found in purebred dogs. For aor- In contrast, 10 disorders were more prevalent in pure- tic stenosis, the top 5 breeds affected on the basis of bred dogs, compared with those found in mixed-breed the percentage of dogs of that breed affected and mixed dogs. Aortic stenosis and dilated cardiomyopathy in the breeds were Newfoundland (6.80%), Boxer (4.49%), cardiac category, hypothyroidism in the endocrine cat- Bull Terrier (4.10%), Irish Terrier (3.13%), Bouvier des egory, elbow dysplasia and IVDD in the orthopedic cate- Flandres (2.38%), and mixed breed (0.15%); for di- gory, and atopy or allergic dermatitis, bloat, cataracts, total lated cardiomyopathy, breeds included Doberman Pin- epilepsy, and portosystemic shunt were all diagnosed in scher (7.32%), Great Dane (7.30%), Neapolitan Mastiff a greater proportion of purebred dogs than mixed-breed (6.52%), Irish Wolfhound (6.08%), Saluki (5.88%), dogs (P < 0.05). The OR for these disorders ranged from and mixed breed (0.16%). Breeds affected with elbow 1.27 (cataracts) to 3.45 (dilated cardiomyopathy) for dysplasia included Bernese Mountain Dog (13.91%), purebred dogs, relative to mixed-breed dogs, indicating a Newfoundland (10.28%), Mastiff (6.55%), Rottwei- greater probability of the condition in purebred dogs. ler (6.31%), Anatolian Shepherd Dog (5.41%), and Cranial cruciate ligament rupture and being hit by mixed breed (0.90%); for IVDD, Dachshund (34.92%), a car were more likely to be observed in mixed-breed French Bulldog (27.06%), Pekingese (20.59%), Pem- dogs than purebred dogs, with a 1.3- and 1.7-fold prob- broke Welsh Corgi (15.11%), Doberman Pinscher ability of the condition, respectively. Whereas the per- (12.70%), and mixed breed (4.43%); for hypothyroid- centage of purebred dogs evaluated at the veterinary ism, Giant Schnauzer (11.45%), Irish Setter (7.69%), medical teaching hospital during this time frame was Keeshond (6.63%), Bouvier des Flandres (6.55%), Do- 73.3% and for mixed-breed dogs was 25.2%, the per- berman Pinscher (6.30%), and mixed breed (1.54%); centage of mixed-breed dogs evaluated after being hit for atopy or allergic dermatitis, West Highland White by a car was 35% and significantly (P < 0.05) greater Terrier (8.58%), Coonhound (8.33%), Wirehaired than expected (Table 2); a similar higher-than-expected Fox Terrier (8.16%), Cairn Terrier (6.91%), Tibetan percentage was observed for pit bull–type dogs. Terrier (5.86%), and mixed breed (1.08%); for bloat,

1552 Scientific Reports JAVMA, Vol 242, No. 11, June 1, 2013 187 Saint Bernard (3.76%), Irish Setter (3.42%), Blood- would cause an overrepresentation of some disorders SMALL ANIMALS hound (3.39%), Great Dane (2.80%), Irish Wolfhound in purebred dogs. Additionally, clients are willing to (2.70%), and mixed breed (0.20%); for cataracts, Silky pursue more extensive treatment at a referral hospital.13 Terrier (22.76%), Miniature Poodle (21.49%), Brussels Owners of purebred dogs are more likely to spend more Griffon (20.51%), Boston Terrier (19.61%), Tibetan on their dogs than are owners of mixed-breed dogs,14 Terrier (18.92%), and mixed breed (4.04%); for epi- which would result in a greater proportion of purebred lepsy (total), Catahoula Leopard Dog (3.90%), Beagle dogs, as seen in the present study. Some dogs in the (3.57%), Schipperke (3.42%), Papillon (3.40%), Stan- present study not classified as having a particular condi- dard Poodle (3.19%), and mixed breed (0.91%); and for tion may simply not have had that condition confirmed portosystemic shunt, Yorkshire Terrier (10.86%), Nor- because of the age of onset or the expense of definitive wich Terrier (7.41%), Pug (5.88%), Maltese (5.87%), diagnostic procedures. For example, epilepsy, atopy Havanese (4.35%), and mixed breed (0.35%). No single (allergic dermatitis), and hypothyroidism, all of which breed dominated the listings. Labrador Retrievers and have higher probability in purebred dogs, require more mixed-breed dogs were more frequently evaluated at the intensive diagnosis, and there may be sociological as- veterinary medical teaching hospital; therefore, those pects in which dog owners who own mixed-breed dogs dogs typically had a greater prevalence of every disorder. may have less incentive to confirm the diagnosis. However, the most frequent breeds affected by each dis- Data for an acute onset of a disorder may have order changed when adjusted for absolute numbers of been underrepresented in our data set if clients prefer- dogs of that breed evaluated at the clinic. Although some entially took the dog to their own veterinarian and not breeds appeared multiple times in different disorders, no a teaching hospital. Furthermore, the Veterinary Medi- breed dominated by the percentage of breed affected. cal Teaching Hospital of the University of California- Davis represents a dog population primarily from the Discussion west coast and may not represent dog populations in other geographic regions. However, for 1 condition in This study characterized the prevalence of genetic the present study (portosystemic shunt), the data and disorders among purebred and mixed-breed dogs eval- the breeds preferentially affected mirrored data for all of uated at the veterinary medical teaching hospital. The North America.15 study was designed specifically to evaluate purebred All of these biases would be expected equally dogs, compared with mixed-breed dogs in total, with- among mixed-breed and purebred dogs in the popula- out attempting to evaluate individual breed prevalence. tion under study, or a bias specifically against the pure- One concern with this approach is that a breed-specific bred dog population may have occurred; neither would disorder found in a high-population breed may inflate affect the objective of the study. Although these are po- the prevalence among purebred dogs, unduly influenc- tential limitations to the data, overall, the data set that ing interpretation of the results. This did not appear to was evaluated is, in the authors’ opinion, one of the best be the case because in those conditions with a differ- representations to include consistent diagnoses in large ence in prevalence between purebred and mixed-breed numbers of purebred and mixed-breed dogs. dogs, none of the top 5 breeds (as a percentage of dogs A previous study5 found no difference between evaluated at the hospital) were high-population breeds. purebred and mixed-breed dogs with hip dysplasia. The results indicated that genetic disorders were Our results, which corroborate the findings of the pre- individual in their expression throughout the dog pop- vious study,5 indicated that in addition to hip dyspla- ulation. Some genetic disorders were present with equal sia, several other disorders did not predominate among prevalence among all dogs in the study, regardless of purebred dogs. For genetic disorders that are found in purebred or mixed-breed status. Other genetic disor- multiple breeds or are equally present in mixed-breed ders were found in greater prevalence among purebred dogs, causal mutations may have arisen multiple times dogs. Every disorder was seen in the mixed-breed pop- or the progenitors of the affected dogs may have been ulation. Thus, on the basis of the data and analyses, the derived from a common distant ancestor carrying the proportion of mixed-breed and purebred dogs affected defect. Mutations introduced into the dog genome ear- by genetic disorders may be equal or differ, depending ly, in an ancestor closely associated with the wolf pro- on the specific disorder. genitor, would be spread through the dog population Although this study evaluated > 90,000 purebred at large. Perhaps the same desired traits that made dogs and mixed-breed dogs, there were limitations to the a favorable species for domestication16 were linked to study. The study population represented dogs evaluated alleles for hyperadrenocorticism, hypoadrenocorticism, at a teaching hospital, and the proportions of the disor- cancers, hip dysplasia, lens luxation, and some cardiac ders in the purebred and mixed-breed dogs may have disorders that were not found to be different between been different from that in the general canine popula- purebred and mixed-breed dogs. tion. However, the study population did reflect the pro- Alternatively, the selection for desirable morpho- portions of purebred and mixed-breed dogs evaluated logical traits may be linked to the presence of delete- at private veterinary hospitals in the United States.12 rious alleles. Patellar luxation and lens luxation are In a referral hospital, breeds that are considered pre- clear examples of size-oriented predisposition. These disposed to a certain condition may be evaluated with disorders did not differ in prevalence between purebred greater frequency and the condition may be diagnosed and mixed-breed dogs, yet appear to be more common at a higher rate than in other breeds or mixed-breed among smaller dogs. Another potential explanation for dogs that do not have a recognized predisposition. This a disorder’s equal prevalence in purebred and mixed-

JAVMA, Vol 242, No. 11, June 1, 2013 Scientific Reports 1553 188 breed dogs is that some tissues or organs may be less mixed-breed dogs having a 30% greater risk for this dis- resistant to genetic aberration and a number of different order than did purebred dogs. The increased risk may mutations may induce a similar phenotypic defect, even be caused by multiple musculoskeletal alleles from dif- though the precise mutations differ in the 2 dog popula- ferent physical conformations that, when combined, re- tions. Additionally, developmental abnormalities influ- duce the resilience of the ligament in the context of the enced by the environment or stochastic developmental joint, as has been suggested for humans.25 17

SMALL ANIMALS SMALL perturbations (eg, certain cardiac conditions) would Purebred dog owners, often devoted to a breed result in the same disease diagnosis. No significant dif- and seeking to track the health of that breed, may have ference was found for cancers between purebred and created the impression that purebred dogs are not as mixed-breed dogs. Genes for cancer expression may be healthy as mixed-breed dogs. Overall, the prevalence of spread widely among the dog population as a whole, disorders among purebred and mixed-breed dogs in the respond to environmental factors that affect all dogs, or present study depended on the condition, with some a combination of both. having a clear distinction between purebred and mixed- For disorders that affected purebred dogs in higher breed dogs and others having no difference. Our results proportions, the underlying causal mutations likely oc- confirmed those of other studies focused on hip dyspla- curred more recently, such as after the gene pools for sia5 and congenital portosystemic shunts15 and expand- particular purebred dogs were developed, or were char- ed the potential for future genetic studies to focus on acteristic of particular lineages. In this study, 4 of the several breeds when considering at-risk breeds to char- top 5 breeds (by percentage) affected with elbow dys- acterize the underlying genetic change. These results plasia are characterized as being from the Mastiff-like also gave insight on the potential effects of breeding dog lineage9: Bernese Mountain Dog, Newfoundland, practices to reduce prevalence. Reliable genetic tests or Mastiff, and Rottweiler. One could speculate that these screening at a young age may reduce some disorders breeds, having been derived from a common ancestor,18 in the dog population as a whole. Additionally, some share mutations. Transmission of genetic disorders may disorders may require breed registry intervention to re- not only occur within a single antiquity lineage, but also duce conformational selection pressures that contrib- may occasionally cross to another lineage as a result of ute to predisposing a breed to a disorder. desire for particular functional traits.8 A 1998 study19 supports this idea by revealing that certain disorders, a. Generalized linear function, R, version 12, R Foundation for Sta- such as elbow dysplasia and portosystemic shunt, oc- tistical Computing, Vienna, Austria. Available at: www.r-project. curred in clusters of highly related dogs, whereas clus- org/. Accessed Feb 21, 2012. ters of unrelated dogs were unaffected. Additionally, the purebred population was at greater risk for atopy than References was the mixed-breed dogs. The published literature in- 1. Brooks M, Sargan DR. Genetic aspects of disease in dogs. In: The dicates that certain breeds are more likely to have atopy genetics of the dogs. Wallingford, Oxfordshire, England: CABI than other breeds,20,21 suggesting that the high preva- Publishing, 2001;191–266. lence within individual breeds may result in the overall 2. Tsai KL, Clark LA, Murphy KE. Understanding hereditary dis- purebred population being at greater risk than the pop- eases using the dog and human as companion model systems. Mamm Genome 2007;18:444–451. ulation of mixed-breed dogs. Reports of mixed-breed 3. Karlsson EK, Lindblad-Toh K. Leader of the pack: gene mapping in dogs having equivalent atopy prevalence to subsets of dogs and other model organisms. Nat Rev Genet 2008;9:713–725. purebred dogs22 support the existence of such an effect 4. Leroy G. Genetic diversity, inbreeding and breeding practices in and underscore the concept of clustering of disorders dogs: results from pedigree analyses. Vet J 2011;189:177–182. among highly related dogs. 5. Rettenmaier JL, Keller GG, Lattimer JC, et al. Prevalence of ca- nine hip dysplasia in a veterinary teaching hospital population. Disorders may be associated with breed deriva- Vet Radiol Ultrasound 2002;43:313–318. tion or with breed bottlenecks. Such an example is the 6. Wayne RK, Ostrander EA. Origin, genetic diversity, and genome Irish Wolfhound, a breed with relatively few dogs regis- structure of the domestic dog. Bioessays 1999;21:247–257. tered annually. In the mid-1800s, the Irish Wolfhound 7. Hedhammar ÅA, Malm S, Bonnett B. International and collab- underwent a population bottleneck so severe that the orative strategies to enhance genetic health in purebred dogs. breed was thought to be extinct.23 The reduced effective Vet J 2011;189:189–196. 8. Lindblad-Toh K, Wade CM, Mikkelsen TS, et al. Genome se- population size suggests a relationship with the con- quence, comparative analysis and haplotype structure of the comitant increased risk of dilated cardiomyopathy in domestic dog. Nature 2005;438:803–819. Irish Wolfhounds. Indeed, as many as 1 in 3 Irish Wolf- 9. Parker HG, Ostrander EA. Canine genomics and genetics: run- hounds may be affected with this disorder.23 In the pres- ning with the pack. PLoS Genet 2005;1:e58. ent study, Irish Wolfhounds were in the top 5 purebred 10. Asher L, Diesel G, Summers JF, et al. Inherited defects in pedi- dog breeds with dilated cardiomyopathy, corroborating gree dogs. Part 1: disorders related to breed standards. Vet J 2009;182:402–411. the high prevalence, compared with other breeds. 11. Hulley SB. Designing clinical research. Philadelphia: Lippincott Other disorders appear to be more generalized and Williams & Wilkins, 2007. more frequently observed in mixed-breed dogs. For ex- 12. Trevejo R, Yang M, Lund EM. Epidemiology of surgical castra- ample, metabolic disturbances have been implicated in tion of dogs and cats in the United States. J Am Vet Med Assoc the onset of canine diabetes mellitus, for which the risk 2011;238:898–904. of development is higher in mixed-breed dogs.24 In the 13. Brown CM. The future of the North American veterinary teach- ing hospital. J Vet Med Educ 2003;30:197–202. present study, dogs with cranial cruciate ligament rup- 14. Dotson MJ, Hyatt EM. Understanding dog-human companion- ture included purebred dogs from at least 3 lineages (ie, ship. J Bus Res 2008;61:457–466. Mastiff, Akita, and German Wirehaired Pointer),9 with 15. Tobias KM, Rohrbach BW. Association of breed with the diag-

1554 Scientific Reports JAVMA, Vol 242, No. 11, June 1, 2013 189 nosis of congenital portosystemic shunts in dogs: 2,400 cases factors for atopic dermatitis in a Swedish population of insured (1980–2002). J Am Vet Med Assoc 2003;223:1636–1639. dogs. Vet Rec 2006;159:241–246. SMALL ANIMALS 16. Vaysse A, Ratnakumar A, Derrien T, et al. Identification of ge- 21. Lund E. Epidemiology of canine atopic dermatitis. Vet Focus nomic regions associated with phenotypic variation between dog 2001;21:32–33. breeds using selection mapping. PLoS Genet 2011;7:e1002316. 22. Schick RO, Fadok VA. Responses of atopic dogs to region- 17. Taussig HB. World survey of the common cardiac malforma- al allergens: 268 cases (1981–1984). J Am Vet Med Assoc tions: developmental error or genetic variant? Am J Cardiol 1986;189:1493–1496. 1982;50:544–559. 23. Parker HG, Meurs KM, Ostrander EA. Finding cardiovascular 18. Vonholdt BM, Pollinger JP, Lohmueller KE, et al. Genome-wide disease genes in the dog. J Vet Cardiol 2006;8:115–127. SNP and haplotype analyses reveal a rich history underlying dog 24. Guptill L, Glickman L, Glickman N. Time trends and risk fac- domestication. Nature 2010;464:898–902. tors for diabetes mellitus in dogs: analysis of veterinary medical 19. Ubbink GJ, Van de Broek J, Hazewinkel HA, et al. Cluster analy- data base records (1970–1999). Vet J 2003;165:240–247. sis of the genetic heterogeneity and disease distributions in 25. Wahl CJ, Westermann RW, Blaisdell GY, et al. An association purebred dog populations. Vet Rec 1998;142:209–213. of lateral knee sagittal anatomic factors with non-contact ACL 20. Nødtvedt A, Egenvall A, Bergval K, et al. Incidence of and risk injury: sex or geometry? J Bone Joint Surg Am 2012;94:217–226.

From this month’s AJVR

Efficacy of decontamination and sterilization of single-use single-incision laparoscopic surgery ports James G. Coisman et al

Objective—To determine the efficacy of decontamination and sterilization of a disposable port intended for use during single-incision laparoscopy. June 2013 Sample—5 material samples obtained from each of 3 laparoscopic surgery ports. Procedures—Ports were assigned to undergo decontamination and ethylene oxide sterilization See the midmonth issues without bacterial inoculation (negative control port), with bacterial inoculation (Staphylococcus au- reus, Escherichia coli, and Mycobacterium fortuitum) and without decontamination and sterilization of JAVMA (positive control port), or with bacterial inoculation followed by decontamination and ethylene oxide for the expanded sterilization (treated port). Each port underwent testing 5 times; during each time, a sample of the foam portion of each port was obtained and bacteriologic culture testing was performed. Bacterio- table of contents logic culture scores were determined for each port sample. for the AJVR Results—None of the treated port samples had positive bacteriologic culture results. All 5 positive control port samples had positive bacteriologic culture results. One negative control port sample had or log on to positive bacteriologic culture results; a spore-forming Bacillus sp organism was cultured from that port sample, which was thought to be an environmental contaminant. Bacteriologic culture scores avmajournals.avma.org for the treated port samples were significantly lower than those for the positive control port samples. Bacteriologic culture scores for the treated port samples were not significantly different from those for access for negative control port samples. to all the abstracts. Conclusions and Clinical Relevance—Results of this study indicated standard procedures for decontamination and sterilization of a single-use port intended for use during single-incision laparo- scopic surgery were effective for elimination of inoculated bacteria. Reuse of this port may be safe for laparoscopic surgery of animals. (Am J Vet Res 2013;74:934–938)

JAVMA, Vol 242, No. 11, June 1, 2013 Scientific Reports 1555 190 J Vet Intern Med 2011;25:784–796 Idiopathic Cystitis in Domestic Cats—Beyond the Lower Urinary Tract

C.A.T. Buffington

Signs of lower urinary tract (LUT) disease in domestic cats can be acute or chronic, and can result from variable combinations of abnormalities within the lumen of the LUT, the parenchyma of the LUT itself, or other organ system(s) that then lead to LUT dysfunction. In the majority of cats with chronic signs of LUT dysfunction, no specific underlying cause can be confirmed after standard clinical evaluation of the LUT, so these cats typically are classified as having idiopathic cystitis. A syndrome in human beings commonly known as interstitial cystitis (IC) shares many features in common with these cats, permitting com- parisons between the two species. A wide range of similarities in abnormalities has been identified between these syndromes outside as well as inside the LUT. A variety of potential familial and developmental risk factors also have been identified. These results have permitted generation of the hypothesis that some of these people have a disorder affecting the LUT rather than a disorder of the LUT. This perspective has suggested alternative diagnostic strategies and novel approaches to treatment, at least in cats. The purpose of this review is to summarize research investigations into the various abnormalities present in cats, to compare some of these findings with those identified in human beings, and to discuss how they might modify perceptions about the etiopathogenesis, diagnosis, and treatment of cats with this disease. Dedication: I dedicate this contribution to Professor Dennis J. Chew, whose collaboration, patience, and support made it all possible. Key words: Comorbidity; Developmental biology; Etiology; Phenotype; Syndrome.

igns of lower urinary tract (LUT) dysfunction in Sdomestic cats (Felis silvestris catus) include variable Abbreviations: combinations of dysuria, hematuria, periuria, poll- ACTH adrenocorticotropic hormone 1 akiuria, and stranguria. A review article published in FIC feline interstitial cystitis 2 1996 listed some 36 confirmed causes of LUT signs. GAG glycosaminoglycan These signs can be acute or chronic, and can result from IC interstitial cystitis variable combinations of abnormalities within the lumen KCl potassium chloride of the LUT (local external abnormalities), in the LUT LUT lower urinary tract itself (intrinsic abnormalities), or other organ system(s) SRS stress response system that then lead to LUT dysfunction (internal abnormali- UTI urinary tract infection ties). In the majority of cats with chronic signs of LUT dysfunction, however, no specific underlying cause can be confirmed after standard clinical evaluation of the as well as to considerable debate about the most appro- LUT. These cats typically are classified as cases of idio- priate name, diagnostic approach, and treatment pathic causation, hence the name idiopathic cystitis.1 recommendations. This reconsideration is ongoing, and Beginning in 1993, results of a series of studies using has resulted in the generation of new hypotheses related cats with chronic idiopathic LUT signs donated by owners to the etiopathogenesis of the signs and symptoms in for whom they no longer were acceptable pets have been both cats and human beings with this problem, as well as published. Initial studies of these cats focused on identifi- novel approaches to treatment, at least in cats. cation of abnormalities of the LUT because the affected The purposes of this review are to summarize some of cats were proposed to represent a naturally occurring the many research investigations into the external, intrin- model of a chronic LUT syndrome in human beings called sic, and internal abnormalities that are present in these cats interstitial cystitis (IC).3,4 These studies led to the proposal (this organization was chosen because it roughly parallels in 1996 that cats having chronic idiopathic LUT signs be the chronology of studies of the syndrome over the past 3 described as having ‘‘feline interstitial cystitis’’ (FIC).5 decades), to compare these findings with those identified in During the ensuing years, evidence also has accumu- human beings with IC during this time, and to consider lated that additional problems outside the LUT are how these results might modify perceptions about the commonly present in these cats, as well as in most diagnosis and treatment of cats with this problem. patients with IC. This evidence has led to reconsidera- tion of the cause(s) of the syndrome in these individuals, Nosology From the Department of Veterinary Clinical Sciences, College of Veterinary Medicine, The Ohio State University, Columbus, OH. Nosology refers to the naming of diseases. Diseases Corresponding author: C.A. Tony Buffington, Department of Vet- can be named according to etiology, pathogenesis, and erinary Clinical Sciences, College of Veterinary Medicine, The Ohio affected organ system(s), and by presenting signs and State University, 601 Vernon L. Tharp Street, Columbus, OH 43210- symptoms. A significant challenge to accurate nosology 1089; e-mail: buffi[email protected]. exists because diseases can be named based on prominent Submitted December 10, 2010; Revised January 26, 2011; Accepted March 28, 2011. signs and symptoms long before research identifies Copyright r 2011 by the American College of Veterinary Internal the etiology and pathogenesis. Whereas presenting Medicine signs sometimes result in naming a disease for the organ 10.1111/j.1939-1676.2011.0732.x associated with the signs, the disease might not originate

191 Idiopathic Cystitis in Cats 785 in the affected organ, and many diseases affect more than Regardless of the name eventually chosen to describe one organ. Thus, the name could reflect a subset of the cats with chronic idiopathic LUT and other clinical signs, problems associated with an underlying disease. This current evidence suggests that restriction of the descrip- could have affected the nosology describing cats with tion of these cats to their LUT signs does not capture all chronic idiopathic LUT signs6 and human beings with currently recognized features of the syndrome.15,16,18 IC.7 Feinstein8 recently concluded that ‘‘an important Regardless of agreement on an accurate descriptive term principle in naming apparently new ailments is to avoid for the syndrome, it seems appropriate for clinicians to etiologic titles until the etiologic agent has been suitably conduct a more comprehensive evaluation of cats pre- demonstrated. A premature causal name can impair a sented with these and other chronic idiopathic signs to patient’s recovery from the syndrome, and impede re- determine whether only these signs occur, or whether search that might find the true cause.’’ variable combinations of comorbid somatic and behav- Although terms such as ‘‘feline urological syndrome,’’9 ioral abnormalities also are present. Such an evaluation ‘‘feline lower urinary tract disease,’’10 and ‘‘feline inter- could result in a more complete diagnosis and implemen- stitial cystitis’’11 fairly accurately capture the currently tation of additional approaches to treatment for some recognized diagnostic criteria for LUT disorders, they cats, which has been associated with better outcomes.15 no longer seem to capture the extent of the problems For the purposes of this review, I will retain the termi- occurring in many cats. These terms all focus on the nology used to describe patients in the studies referenced, LUT, reflecting the prominent presenting signs and since it was what was used at the time the results were LUT-focused diagnostic testing rather than a thorough published. This is done with some trepidation because of evaluation of the entire cat. In human beings, more the risk of reinforcing the focus on the LUT rather than a comprehensive investigations of patients with IC and more comprehensive assessment of the problem list of the a variety of other chronic idiopathic disorders have patients, but such was, and to a greater or lesser extent resulted in the suggestion of names such as ‘‘medically still is, how studies have been reported. unexplained syndrome,’’12 ‘‘functional somatic syn- 13 14 drome,’’ or ‘‘central sensitivity syndrome’’ to Abnormalities Identified in FIC and IC describe the multiple abnormalities observed in these patients by physicians. The list of chronic disorders pro- Many features in common have been identified in cats posed to be covered by these names is long, and includes and human beings with the syndrome.19 Variable combi- problems addressed by most of the medical subspecial- nations of LUT abnormalities have been identified in ties. These names also seem to violate Feinstein’s patients of both species, who also often suffer multiple admonition, however, and it seems that some generic comorbid disorders.16,20 Moreover, the occurrence of co- umbrella term comparable to ‘‘cancer’’ or ‘‘infection’’ morbid disorders often precedes the occurrence of LUT might be more appropriate. One possibility, which I will signs and symptoms (C.A.T. Buffington, unpublished use in this review when it seems appropriate, is to adopt observation).21,22 These comorbid disorders also appear an interim name such as ‘‘Pandora’’ syndrome until the to occur more commonly in close relatives of human most biologically appropriate nosological term is identi- patients,23,24 and evidence of adverse early experiences fied. Tentative criteria for diagnosis of a ‘‘Pandora’’ has been reported in patients with FIC25 and IC.26 syndrome include: Two LUT forms of the syndrome have been reported, nonulcerative (Type I) and ulcerative (Type II); other 20 1. Presence of clinical signs referable to other organ sys- forms also could exist. Cats almost always present with tems in addition to the chronic idiopathic signs the Type I form, although the Type II form has been 27 prominently referable to a particular organ for which described, and in human beings, approximately 90% of 28 the patient is being evaluated. For example, variable patients have the Type I form. The etiopathogenesis of combinations of clinical signs referable to other organ these 2 forms differs. The Type II form appears to be an systems such as the gastrointestinal tract, skin, lung, inflammatory disease intrinsic to the bladder, whereas cardiovascular, central nervous, endocrine, and the Type I form might be neuropathic in origin. immune systems have been identified in cats with Owners commonly request evaluation of obvious LUT chronic idiopathic LUT signs.15,16 signs they observe in their cats, so a large amount of 2. Waxing and waning of severity of clinical signs asso- research has been directed toward the bladder, resulting ciated with events that (presumably) activate the in identification of a variety of abnormalities. The blad- 29,30 central stress response system (SRS).15,17,18 der is a deceptively sophisticated organ. Its internal 3. Resolution of signs associated with effective environ- covering consists of an epithelium with its underlying ne- mental enrichment.15,17,18 urovascular supporting tissue, which is surrounded by both smooth and striated muscle.31 These structures en- gage in complex neuroendocrine communication with A name like ‘‘Pandora’’ syndrome seems appropriate the rest of the body to determine the appropriate condi- for at least 2 reasons. First, it does not identify any spe- tions and timing for voiding. Bladder neural connections cific cause or organ, and second, it seems to capture the include sensory afferent, central, and somatic, sympa- dismay and dispute associated with the identification of thetic, and parasympathetic efferent neurons that so many problems (evils) outside the organ of interest of interact throughout the neuraxis between the urothelium any particular subspecialty. and the cerebral cortex.32 In addition to a variety of

192 786 Buffington neurotransmitters, bladder function also is influenced by such as the number of previous treatments, including both adrenocortical and sex hormones.33 catheterization of male cats, also might have influenced their results. Other studies have reported prevalence rates of UTI from 15 to 43% in cats with compromised urinary Local External Abnormalities tract defense mechanisms, and 1 study reported a pre- Toxic and Protective Factors. The presence of some valence of 22% in cats with no apparent predisposing toxin,34 abnormality of some protective factor,35,36 or factors.50 Some evidence suggests that colonization may presence of some microorganism37,38 in the urine has result from an underlying vulnerability in affected cats. been proposed to explain the LUT signs and symptoms Perineal urethrostomy did not lead to postoperative bac- in patients with FIC and IC. An abnormality of Tamm- terial infection in healthy cats, whereas it occurred Horsfall protein that results in loss of protection of the postoperatively in 22% of cats with histories of recurrent urothelium,39 the appearance of an ‘‘anti-proliferative or persistent urethral obstruction.51 factor’’ and local growth factor abnormalities that might There also might be a relationship between IC and disrupt cell signaling,40 and other changes in the urine of UTI in human beings. One recent study found evidence patients with IC have been identified and are being of UTI within the past 2 years in 38% of the IC/painful investigated.41,42 Whether these play causative roles in bladder syndrome patients they studied,52 although, ‘‘ . . . FIC or IC remains to be determined, although the rele- the infection domain was not associated with any vance of the Tamm-Horsfall protein abnormality was increased symptoms.’’ Additionally, retrospective data diminished by the report of absence of voiding dysfunc- suggest that a proportion, probably a minority, of tion or compatible histological abnormalities in Tamm- women had evidence of UTI or inflammation at the Horsfall protein knockout mice.43 onset of symptoms of IC/painful bladder syndrome.53 It Microbial Agents. Given the similarity in symptoms also has been speculated that intrinsic abnormalities between cystitis resulting from bacterial urinary tract make the LUT more vulnerable to microbial coloniza- infection (UTI) and FIC and IC, researchers have con- tion,38 which might be consistent with the observation of sidered infection to be a cause of the LUT signs and increased risk for bacterial UTI in these patients. symptoms for nearly 100 years. Guy Hunner, for whom the ‘‘Hunner’s ulcer’’ of the Type II form of the syn- drome was named, publically speculated that a bacterial Intrinsic Abnormalities infection was the cause of ‘‘a rare type of bladder ulcer in The Glycosaminoglycan (GAG) Layer. The internal women’’ in 1915.44 If microbes are associated with FIC surface of the LUT is coated by a GAG layer that might or IC, they could either cause the disorder, or be associ- be abnormal in patients with FIC or IC. A wide variety ated with it in some noncausal way. of sometimes-conflicting changes in the quantity and A role for infectious agents such as viruses in the LUT quality of the GAG layer in patients with IC is signs observed in cats has been investigated,37,45 reported.54–56 Decreased total GAG,35,57 and a specific although what relationship viruses play in the etiopatho- GAG known as GP-51,58 has been reported in cats with genesis of these signs in cats with naturally occurring FIC FIC. One group of investigators also found chondroitin remains unclear at this time.46 Moreover, investigations sulfate in the plasma of cats with feline urologic syn- of what role infectious agents might play in the systemic drome, leading them to conclude that the decreased manifestations of the syndrome are yet to be reported. chondroitin concentration they found in urine could have In human beings, 2 recent studies concluded that ‘‘IC resulted from reabsorption back across a more permeable is not associated with persistence of viral and bacterial urothelium.57 Limitations of most studies of urine GAG DNA in the bladder. A chronic infective etiology for the include the difficulty of the GAG assay and the variety of condition is excluded by these findings,’’47 and, ‘‘these methods used, so what role the GAG layer plays in these data suggest that the symptom flares of IC are not usu- disorders currently remains unresolved.59 ally associated with recurrent UTI and, therefore, are Experimental attempts to replenish the GAG layer likely due to a triggering of the other painful mechanisms also have been reported. In cats, 2 studies of the effects involved in IC patients who are culture-negative.’’48 of GAG replacement therapies have been investigated, Thus, the probability that an infectious agent commonly but no benefit beyond placebo was found in either causes the symptoms present in these patients seems study.60,61 In human beings, the beneficial effects of quite small. polysulfated62,63 and other GAGs64 on symptoms of IC Although microorganisms in the LUT might not com- or painful bladder syndrome/IC also appear to be small. monly cause FIC or IC, this does not mean that microbes As noted in a recent editorial commentary, the shift in have no association with the syndromes. A recent report perspective toward a more systemic view of IC ‘‘calls lo- of 134 cats in Norway evaluated for LUT signs found cal treatments into question.’’59 bacteriuria exceeding 103 CFU/mL in 44 (33%) cats, and Urothelium. A specialized epithelium called the uro- exceeding 104 in 33 (25%), either alone or with variable thelium lines the distal portion of the urinary tract, in- combinations of crystals and uroliths.49 These results cluding the renal pelvis, ureters, bladder, upper urethra, suggested a prevalence of bacteriuria higher than and glandular ducts of the prostate.65 The urothelium is reported previously, which the authors speculated might composed of a basal cell layer attached to a basement have resulted from differences between cases diagnosed membrane, an intermediate layer, and a superficial apical at primary and tertiary care facilities. Other variables, layer.66 Although healthy urothelium maintains a tight

193 Idiopathic Cystitis in Cats 787 barrier to ion and solute flux, factors such as altered pH or could be a neurally mediated byproduct of the stress electrolyte concentrations, mechanical, chemical, or neu- response associated with the disorder. One beneficial rally mediated stimulation, and infectious agents all can action of the tricyclic antidepressant amitriptyline (if impair the integrity of the barrier.67 such exists83) could be through inhibition of mast cell ac- Both functional and anatomical abnormalities of the tivation.84 In one recent report, however, no difference in urothelium have been reported in FIC and IC, although the degree of lymphocyte and mast cell infiltration, or in their cause and significance are unknown. In cats with neovascularization or staining for uroplakins, was found FIC, significantly higher bladder permeability to sodium between bladders of cats with feline idiopathic cystitis salicylate,68 as well as reduced transepithelial resistance and those with urolithiasis, and in this study urothelial and increased water and urea permeability after GAG staining was highest in tissues from affected cats.85 hydrodistention of the bladder, has been reported.69 A Detrusor Muscle. In contrast to the many abnormali- denuded urothelium with appearance of underlying cells ties found on the luminal side of the lamina propria, also was found in these cats by scanning and transmis- there is a paucity of data of etiopathogenic importance sion electron microscopy, leading the authors to implicating the bladder muscle in the pathophysiology of conclude that the urothelial damage and dysfunction FIC or IC.86 In cats with FIC, nonspecific inflammatory identified might ‘‘suggest novel approaches toward changes in the detrusor,87 as well as in vitro evidence to examining the etiology and therapy of IC.’’69 Ironically, suggest that the muscle functions relatively normally,79 a paper published the same month70 reported strikingly have been reported. similar electron microscopic findings—in healthy female Intrinsic Abnormalities—Summary. The etiopathogenic mice exposed to constant illumination for 96 hours, after significance of local bladder abnormalities occurring in pa- which they were returned to conventional day-night illu- tients with FIC and IC remains to be established. mination for 7 days before being killed. This report Moreover, in chronic diseases, clinical signs often do not 28 showed that comparable urothelial injury also could appear to correlate well with pathology in the bladder, or 88 occur in healthy animals exposed to stressful external elsewhere. For example, bladder lesions characteristically events. Neither of these studies examined any other associated with irritative voiding symptoms and pelvic tissues to determine if the observed abnormalities were pain in patients diagnosed with IC also have been observed 89 restricted to the bladder or had a more widespread in asymptomatic women undergoing tubal ligation. distribution. Some patients treated with cyclophosphamide also develop Recent studies have revealed that urothelial cells a hemorrhagic cystitis and voiding dysfunction without the 90 express a number of molecular ‘‘sensors’’ that confer pain often associated with IC. A similar situation also properties similar to both nociceptive and mechanosen- occurs in the bowel. In one study, rectal perception of dis- sitive type neurons on these cells. Thus, like superficial tention was actually attenuated in patients with ulcerative cells on other epithelial surfaces,71,72 urothelial cells pos- colitis,whereasitwasenhancedinpatientswithirritable 91 sess specialized sensory and signaling properties that bowel syndrome. To paraphrase the conclusion of the allow them to respond to their environment and to authors of this study, low-grade mucosal inflammation engage in reciprocal communication with neighboring alone is unlikely to be responsible for symptoms of func- urothelial and nerve cells.73 Alterations in the expression tional disorders. of various receptors, channels, and transmitters involved Most studies of FIC and IC also have failed to examine in both the ‘‘sensor’’ as well as ‘‘transducer’’ properties of tissues from other organs for comparison, so one cannot the urothelium at both gene and protein levels have been determine whether the identified changes are restricted to found in urothelial cells from both cats and human beings the LUT, or whether they also occur elsewhere in the with the syndrome.30 Alterations in stretch-mediated body of patients with the syndrome. Moreover, no tem- release of transmitters from the urothelium, including poral relationship has been established between these increased nitric oxide74 and adenosine triphosphate75 abnormalities and the onset of clinical signs. Finally, release also may influence urothelial integrity and cell-cell improvement in clinical signs has been reported to occur signaling. in the absence of cystoscopic or histological changes in Submucosa. Abnormalities also are present below the cats92 or human beings,93 and cystectomy does not urothelium, although the histological features of Type I resolve symptoms in human beings with the Type-I form FIC76 and IC77 are somewhat unusual. Vasodilatation of the syndrome.94 These findings suggest that important and vascular leakage in the general absence of any sig- parts of the problem lie elsewhere. nificant mononuclear or polymorphonuclear infiltrate is the most common finding, suggesting the presence of neurogenic inflammation.78,79 Increased numbers of Internal Abnormalities mast cells have been observed in biopsy specimens from Afferent Input. Sensory information is transmitted about 20% of patients with Type I FIC76 and IC,28 and from the bladder to the spinal cord by afferent neurons. are thought by some to be involved in the pathophysiol- Mechanosensitive bladder afferent neurons were found ogy of the syndrome.80 The finding of mast cells in to exhibit a small increase in sensitivity to distension with the bladder is by no means specific to these syndromes.81 154 mM saline in cats affected with FIC as compared The role of mast cells in IC and comorbid disorders, with normal cats, albeit at higher than normal spontane- especially those exacerbated by stress, was recently ous micturition pressures.95 The effect of increasing reviewed.82 It was concluded that mast cell activation concentrations (80–300 mM) of potassium chloride

194 788 Buffington

(KCl) on afferent firing also was examined, both because the rate-limiting enzyme of catecholamine synthesis, intravesical KCl has been used as a diagnostic probe for immunoreactivity have been identified in the pontine IC in human beings,96 and because it has been specu- locus coeruleus120 and the paraventricular nucleus of the lated,97 but never demonstrated, that the urine potassium hypothalamus of cats with FIC.121 concentration plays a role in the pathophysiology of IC. The locus coeruleus contains the largest number of Increased afferent firing similar to that seen with saline noradrenergic neurons, and is the most important source was observed during filling with KCl at concentrations of norepinephrine in the central nervous system. Afferent o150 mM; however, concentrations of 150–300 mM pro- input, including bladder distention, stimulates neuronal duced almost complete inhibition of afferent firing at activity in the locus coeruleus, which is the origin of the pressures between 30 and 80 cm of water, suggesting that descending excitatory pathway to the bladder.29 The lo- increased bladder permeability permits entry of suffi- cus coeruleus also is involved in such global brain ciently high concentrations of KCl into the submucosa functions as vigilance and arousal. Increased tyrosine to dampen neural activity. These data suggest that hydroxylase activity in the locus coeruleus also can occur afferent nerves become more sensitive to stimuli in cats in response to chronic external stressors,122 with accom- with FIC. panying increases in autonomic outflow.123 Moreover, A modest increase in Substance P, an 11 amino acid the locus coeruleus appears to mediate visceral responses sensory neurotransmitter peptide, immunoreactivity in to external as well as internal input.124 The increased unmyelinated neurons has been detected in bladder tissue immunoreactivity found in these nuclei might thus pro- from cats with FIC,98 and in some,99 but not all,100 stud- vide clues to the observation that the signs in cats117,125 ies of bladder tissue from human beings with IC. Bladder and symptoms in human beings126,127 follow a waxing Substance P receptor expression is significantly and waning course that can be influenced by external as increased in cats with FIC,101 and both increased102 and well as internal events. decreased103 in patients with IC. Clinical trials of External environmental events that activate the SRS are the therapeutic properties of Substance P antagonists in termed stressors.128 Examples of these events include sud- human beings to date have been disappointing, how- den movements, unknown or loud noises, novel and ever,104,105 and recent evidence suggests that Substance P unfamiliar places and objects, and the approach of strang- might limit the severity of inflammatory reactions,106,107 ers. Inadequate perception of control and predictability opening the possibility that the changes observed in alsocanactivatetheSRSinanimalsbecauseofinterfer- patients with these syndromes may reflect some protec- ence with attempts to cope with their environments.129 tive response. Depending on the frequency, intensity, and duration, A variety of abnormalities have been identified in dor- chronic activation of the SRS can overtax homeostatic sal root ganglion cell bodies of bladder-identified regulatory systems, resulting in diminished welfare,130 ab- neurons from cats with FIC. Cells from affected cats normal conduct, and sickness behaviors.131,132 were 30% larger, expressed altered neuropeptide pro- The acoustic startle response has been used as a probe of files, and exhibited slowly desensitizing, capsaicin- sensitivity to external events in patients with FIC and IC. induced currents related to increased protein kinase This response is a brainstem reflex that responds to unex- C-mediated phosphorylation of the transient receptor pected, loud stimuli, which has been shown to be increased potential vanilloid 1 receptor.108 Moreover, these abnor- by both fear and anxiety mediated by higher brain struc- malities were not restricted to cells from bladder- tures.133 The acoustic startle response in cats with FIC is identified neurons; similar findings were observed in greatest and most different from that of healthy cats during dorsal root ganglion cells throughout the lumbosacral stressful situations, but is still greater in cats with FIC than 108 (L4-S3) spinal cord. in healthy cats even when adapted to enriched housing con- Treatments targeting bladder sensory neurons have ditions.134 Exaggerated acoustic startle responses also have been tested, but without success to date.109 Resinifera- been reported in women with IC.135,136 toxin, a potent naturally occurring analog of capsaicin that activates transient receptor potential vanilloid 1 Efferent Output receptors on nociceptive sensory neurons, reduced blad- der compliance and capacity in a pilot study of Neural. Activation of the SRS by either internal or anesthetized cats with FIC.110 Controlled trials of both external stimuli can result in stimulation of peripheral capsaicin and resiniferatoxin in human beings with IC neural, hormonal, and immune responses. In addition to also have failed to find significant benefits over pla- increased activity in the locus coeruleus, plasma cat- cebo.111 As one expert recently concluded, ‘‘Intravesical echolamine concentrations are significantly (P o .05) instillation therapy has basically not changed during the higher in cats with FIC compared with healthy cats both at rest125 as well as during exposure to a moderate stress last few years, although some studies have disconfirmed 17 some regimens. Intensive research may hopefully result protocol. Furthermore, plasma catecholamine concen- in more effective treatments in the future.’’112 trations decreased in the healthy cats as they acclimated Brain. Exacerbations of LUT signs in response to to the stress, whereas even higher concentrations of plasma norepinephrine and epinephrine were found in external environmental challenges have been reported 17 both in laboratory studies17 and in client-owned cats cats with idiopathic cystitis. with FIC,113–117 as well as in patients with IC.118,119 In A functional desensitization of a-2 adrenergic recep- the brain, significant increases in tyrosine hydroxylase, tors in affected cats also has been identified by evaluating

195 Idiopathic Cystitis in Cats 789 their response to the selective a-2 adrenergic receptor These results, when combined with observations of in- agonist medetomidine in both in vivo137 and in vitro creased concentrations of corticotrophin-releasing studies.79 In vivo, heart rate decreased and pupil diame- factor121,149 and ACTH146 in response to stress in the ab- ter increased significantly in healthy cats compared with sence of a comparable increase in plasma adrenocortical cats with idiopathic cystitis, which also had significantly hormone concentrations, suggest the presence of mild lower respiratory rates than did healthy cats after intra- primary adrenocortical insufficiency or decreased adre- muscular administration of 20 mg medetomidine/kg body nocortical reserve in cats with FIC. Inappropriately low weight. No significant differences in blood pressure or plasma adrenocortical hormone concentrations also sedation level were observed. In vitro, electrical field have been observed in human beings with IC and chronic stimulation of bladder strips from cats with FIC revealed idiopathic prostatic pain syndrome.20,150 Potential mech- that atipamezole, an a-2 adrenergic receptor antagonist, anisms underlying the stress-related reductions in did not alter the relaxing effect of norepinephrine, fur- circulating adrenocortical steroid concentrations include ther suggesting downregulation of a-2 adrenergic endocrine,151 neural,152,153 and developmental influences receptors.79 on the adrenal gland.20 Abnormalities of efferent nerves also appear to be Immune. Studies of laboratory-housed17,154 and zoo- present. Bladder tissue from patients with FIC (A.J. confined cats155 have found that activation of the SRS is Reche and C.A.T. Buffington, unpublished observations, associated with a variety of sickness behaviors.18 Sick- 2001) and IC99,100 contains increased tyrosine hydroxy- ness behaviors refer to variable combinations of lase-immunoreactive neurons in both muscle and vomiting, diarrhea, anorexia or decreased food and urothelium. There is increased nitric oxide74 and norepi- water intake, fever, lethargy, somnolence, enhanced nephrine (but not acetylcholine) release from bladder pain-like behaviors, as well as decreased general activity, strips in cats with FIC.79 In addition, tyrosine hydroxy- body care activities (grooming), and social interac- lase-containing nerves occur in or near the bladder tions.156 Sickness behaviors are thought to reflect a mucosa, suggesting an interaction between noradrenergic change in motivation toward withdrawal to promote re- nerves and the urothelium. Urothelial cells can express covery by inhibiting metabolically expensive (eg, foraging) both a- and b-adrenergic receptors, and adrenergic ago- or dangerous (eg, exposure to predators) activities when nist stimulation of these receptors leads to nitric oxide the animal is in a relatively vulnerable state. Sickness be- release. These data support the view that the urothelium haviors are found across mammalian species, and their can be influenced by both afferent and efferent nerves, occurrence157 has been linked to immune activation and which in turn can influence the function of a variety of cell proinflammatory cytokine release,158 as well as to changes types and ultimately bladder function.138 Significant in- in mood and pathologic pain.132,159 Sickness behaviors creases in local nerve growth factor concentrations also can result both from peripheral (bottom-up) and central have been found in affected cats,139 and human be- (top-down) activation of immune responses. In a recent ings,140,141 which too can affect bladder nerve function,30 study of healthy cats and cats with FIC,18 (infra vide) although the finding in humans was not specific to IC.140 unusual environmental events, but not disease status, The specificity of the finding in cats is not known. resulted in a significant increase in total sickness behav- Activation of the SRS also can increase epithelial per- iors when the results were controlled for other factors. meability by neural mechanisms, permitting environ- Recent studies have begun to map the pathways that mental agents greater access to sensory neurons,142 which transduce activation of the SRS into cellular dysfunction. could result both in increased afferent firing and local Induction of the transcription factor nuclear factor-kBin inflammation. Thus, the effects of the emotional state of peripheral blood mononuclear cells was observed after the animal may modulate perceived sensations from environmental activation of the SRS.160 Only nor- peripheral organs, completing a loop that may be modu- epinephrine induced this response, which was reduced latedbybothcentralandperipheralneuralactivity.143 by both a(1)- and b-adrenergic inhibitors. The authors Hormonal. In addition to the sensory, central, and concluded that norepinephrine-mediated activation of efferent neural abnormalities identified, an ‘‘uncoupling’’ nuclear factor-kB represented a downstream effector of of SRS output, with a relative predominance of sympa- the response to stressful psychosocial events, linking thetic nervous system to hypothalamic-pituitary-adrenal changes in the activity of the SRS to a bewildering array activity,144 appears to be present in patients with FIC of cellular responses via cell surface receptors.161 Cyto- and IC. Sympathoneural outflow normally is restrained kines and a variety of other inflammatory and metabolic by adrenocortical output.145 In patients with FIC20,146 signals also can activate nuclear factor-kB by binding and IC,147,148 however, it increases without coactivation to different cell surface receptors, further complicat- of the adrenal cortex. Additionally, the adrenocortical ing interpretation of the source(s) of generation of cellu- response to adrenocorticotropic hormone (ACTH) stim- lar responses. Adrenocortical steroids tend to inhibit ulation during stressful circumstances is reduced, and activation of nuclear factor-kB.162,163 This and other cats with FIC often have small adrenal glands.25,146 adrenocortical steroid-related protective mechanisms164–166 Histopathological examination of these glands excluded might be less efficient in hypoadrenocortical states such as the presence of hemorrhage, inflammation, infection, FIC and IC. fibrosis or necrosis, and morphometric evaluation iden- Comorbid Disorders. The possibility of an internal tified reduced size of the fasciculata and reticularis zones cause in some patients with FIC and IC also is suggested of the adrenal cortex. by the presence of multiple comorbid disorders in many

196 790 Buffington patients, the absence of this pattern of comorbidity in abnormalities suggest a genetic or familial susceptibility, patients with other LUT diseases, and the unpredictable a developmental accident, or some combination of order of appearance of the comorbidities. Cats with FIC these.20,26 When a pregnant female is exposed to a suffi- can have variable combinations of comorbid disorders, ciently harsh stressor, or is unusually sensitive to including behavioral, cardiovascular, endocrine, and environmental stressors herself, the hormonal products gastrointestinal problems in addition to their LUT of the ensuing stress response may cross the placenta and signs.15,16,20,115,167 Most human beings with IC also affect the course of fetal development.172 The biological suffer from variable combinations of comorbid disorders ‘‘purpose’’ of transmitting this response to the fetus that affect a variety of other body systems.20,168–170 That might be to program the development of the fetal SRS patients with FIC and IC have variable combinations of and associated behaviors toward enhanced vigilance to other comorbid disorders raises the question of the increase the probability of survival.173 extent to which a different etiology affects each organ The effects of maternal hormones on the fetus seem to versus the extent to which some common disorder affects depend on the timing and magnitude of exposure in rela- all organs, which then respond in their own characteristic tion to the developmental ‘‘programs’’ that determine the ways. maturation of the various body systems during gestation External, intrinsic, or both, bladder abnormalities and early postnatal development.172 For example, if the could lead to development of these other disorders. fetus is exposed before initiation of a developmental Patients with extrinsic (eg, chronic UTI) or intrinsic program, there might be no effect on adrenal develop- (eg, bladder cancer or ‘‘overactive bladder’’) urological ment. Adrenal development might be reduced, however, disorders, however, have not been reported to be at if exposure occurs during the critical period when the comparable increased risk for development of the adrenocortical maturation program is running,20 or many comorbid disorders that afflict patients with IC. increased if exposure occurs after the period of adreno- Moreover, appearance of FIC (C.A.T. Buffington, un- cortical development.173 published observation) or IC21,22 does not predictably Postnatal stressors also can result in persistently precede development of other syndromes, further sug- increased central corticotrophin-releasing factor activity gesting that they are not a consequence but rather in animals.174 Behavioral abnormalities in adult rats can independent events or separate manifestations of a com- result from adverse events during the neonatal period.175 mon underlying disorder. These effects were mediated by epigenetic modification of Internal Abnormalities—Summary. In addition to the glucocorticoid receptor gene expression in the hippocam- variety of local bladder abnormalities identified in pus by DNA methylation and histone acetylation.176 patients with FIC and IC, examination of other tissues Adult mice subjected to chronic social stress have stress- for comparison has revealed that many of the identified induced epigenetic modulation of hippocampal gene changes are not restricted to the bladder, but also occur expression that is not restricted to the neonatal period.177 elsewhere in the body of patients with the syndrome. In addition, other studies of early environmental effects on Moreover, comorbid disorders apparently are as likely to rat pups have found alterations in autonomic emotional precede as to follow the onset of the syndrome. The num- motor circuits,178 as well as in monoamine, g-amino ber, order of onset, and extent of abnormalities identified butyric acid, and glutaminergic circuits in adulthood.179 outside the LUT in cats with FIC were unexpected, and it Studies in rodents also have shown that neonatal seems likely that more will be identified in the future. inflammation of the bladder can result in impaired blad- Moreover, many of the changes seem to be ‘‘functional,’’ der function in adults when the bladder is rechallenged.180 waxing and waning with disease activity, rather than Similar results also have been reported in the colon after structural. Disease activity also was found to change with neonatal manipulation181 or maternal deprivation.182 environmental circumstances, worsening during expo- These results support the hypothesis that events experi- sure to challenging (stressful) circumstances. enced during development may permanently affect Although a variety of internal abnormalities in tissues visceral sensory systems, representing an additional po- or systems distant to the bladder occur in patients with tential cause of chronic idiopathic disorders. Unfortu- FIC and IC, their etiopathologic significance has not nately, other organs were not evaluated in these studies, been established. Evidence also supports the observation so the full extent of the changes resulting from early that both external (environmental) as well as internal adverse experiences remains to be determined. (visceral) events can activate the SRS, leading to activa- Recent studies in human beings also have demon- tion of variable combinations of neural, hormonal, and strated that early adverse experience can result in immune responses. These responses might help explain durable alterations in endocrine and autonomic the number, location, and variability of subsequent responses to stress similar to those identified in IC.147,183 health problems.171 Although the dramatic adverse effects of abuse on the SRS of human beings are well known,184 less extreme Early Life Events parenting behaviors such as neglect, rejection, and hos- tility185 as well as a host of environmental events186 also The findings of increased corticotrophin-releasing fac- might play important mediating roles in the neuroendo- tor, ACTH, and sympathoneural activity in the presence crine abnormalities observed.187,188 of reduced adrenocortical response and small adrenal Early life events also can confer resilience to adverse fasciculata and reticularis zones without other apparent experience. Both genetic and environmental resilience

197 Idiopathic Cystitis in Cats 791 factors have been identified,189–191 and the effect of distal penis to attempt to dislodge any obstructions, external events on these factors on the developing ner- decompressive cystocentesis, and a darkened, low stress vous system might depend on the timing of exposure to environment that did not house any dogs resulted in res- them.192 Thus, research has demonstrated that early life olution of urethral obstruction, defined as spontaneous experience can have a multitude of effects on the exposed urination within 72 hours and subsequent discharge individual, from conferring susceptibility to reinforcing from the hospital, without the need for urethral catheter- resilience. Moreover, these effects can confer a suscepti- ization in 11/15 (73%) of male cats with urethral bility that might or might not eventually be unmasked by obstruction.195 And in a laboratory study, sickness later events,193,194 further complicating the story. behaviors were observed both in healthy cats and in cats with FIC in response to unusual external events for 18 Additional Findings 77 weeks after environmental enrichment. Increasing age and weeks when unusual external events occurred, The idea that a ‘‘Pandora’’ syndrome might be present but not disease status, resulted in a significant increase in in some cats with chronic idiopathic LUT signs devel- total sickness behaviors when controlled for other fac- oped from a number of clinical and laboratory studies. In tors. A protective effect of male sex on food intake in the late 1990s, a prospective, multicenter, double- healthy cats was observed, as well as a small increased blinded, placebo-controlled, randomized clinical trial risk of age for upper gastrointestinal (1.2) and avoidance designed to evaluate the efficacy of pentosan polysulfate behaviors (1.7). In contrast, unusual external events were for improving LUT signs in cats with FIC was con- associated with significantly increased risks for decreases in ducted.60 Cats with at least 2 episodes of LUT signs food intake (9.3) and elimination (6.4), and increases in within the past 6 months, cystoscopic findings of diffuse defecation (9.8) and urination (1.6) outside the litter box. glomerulations present in at least 2 quadrants of the These results suggest that some of the most commonly bladder, and the absence of an alternative diagnosis after observed abnormalities in client-owned cats occurred after appropriate clinical investigations were randomly as- unusual external events in both groups. Because all cats signed to receive either 0.0 (vehicle placebo), 2.0, 8.0, or were comparably affected by unusual external events, cli- 16.0 mg/kg pentosan polysulfate twice daily for 26 weeks. nicians may need to consider the possibility of exposure to Owners evaluated the cats weekly by rating hematuria, unusual external events in the differential diagnosis of cats stranguria, pollakiuria, periuria, and vocalization during presented for care for these signs. voiding attempts on a scale of 0–3 (none, mild, moderate, severe), and additional cystoscopic examination was per- Clinical Implications formed at the end of the study. All treatments were well tolerated by the cats; adverse events were rare and no Based on the evidence available to date, some cats consistent treatment-related pattern was evident. Aver- evaluated for chronic signs of LUT dysfunction might age owner-recorded scores of signs of LUT dysfunction instead have a ‘‘Pandora’’ syndrome. Given the comor- decreased by approximately 75% in all groups, although bid disorders sometimes found in cats with some other recurrent episodes occurred on some 35% of cats. While chronic disorders, other presentations of the syndrome these results suggest that nonspecific therapeutic seem likely. Based on these observations, and on the cur- responses might occur in cats with FIC, possibly by rent limited understanding of the many factors altering their perception of their surroundings, lack of a potentially involved, a reasonable diagnostic strategy ‘‘usual care’’ control group require that the study be for cats with chronic clinical signs referable to a particu- interpreted with caution. lar organ system might be to conduct a comprehensive The hypothesis that LUT signs might be responsive to investigation of the animal’s history, environment, and environmental influences, while not novel,113,114 led to other organ system function. Additional supportive data additional investigations. Laboratory studies revealed might include evidence of early adverse experience (or- that environmental enrichment was associated not only phaned, abandoned, etc.), presence of related signs in with reduction in LUT signs, but also with normalization family members, waxing and waning of signs related to of circulating catecholamine concentrations, bladder environmental threat, and the absence of evidence for an permeability, and cardiac function,17,137 and reduced re- alternative cause. Evidence for the presence of these sponses to acoustic startle.134 Based on these findings, additional factors would support diagnosis of ‘‘Pan- environmental enrichment was evaluated in a 10-month dora’’ syndrome, whereas evidence of absence of these prospective observational study of client-owned cats factors would argue for an organ-specific disorder. with moderate to severe feline idiopathic cystitis.15 In With regard to treatment, significant recovery from addition to their usual care, clients were offered individu- signs referable to the LUT and other systems has been alized recommendations for multimodal environmental reported in cats with LUT-predominant ‘‘Pandora’’ syn- modification based on a detailed environmental history. drome using tailored multimodal environmental In addition to significant reductions in LUT signs, modification.15 The effectiveness of environmental decreased fearfulness, nervousness, signs referable to the enrichment also suggests that pharmacological or other respiratory tract, and a trend toward reduced aggressive therapeutic interventions face an important barrier to behaviors were identified.15 demonstrate efficacy in the presence of the large thera- Most recently, a clinical study of pharmacologic ther- peutic response to this approach in cats with the apy, extrusion, inspection, and gentle massage of the syndrome. Moreover, pharmacological approaches that

198 792 Buffington require force, such as pilling, also might result in activation 5. Buffington CA, Chew DJ, DiBartola SP. Interstitial cystitis of the SRS. Given the lack of evidence for effectiveness of in cats. Vet Clin North Am Small Anim Pract 1996;26:317–326. most currently available pharmaceutical treatments for 6. Osborne CA, Kruger JM, Lulich JP, et al. Feline urologic cats with chronic idiopathic LUT signs at least, these syndrome, feline lower urinary tract disease, feline interstitial cysti- approaches should be undertaken with caution. tis: What’s in a name? J Am Vet Med Assoc 1999;214:1470–1480. The prognosis for recovery of cats with LUT-predom- 7. Hanno PM. Re-imagining interstitial cystitis. Urol Clin North Am 2008;35:91–99. inant ‘‘Pandora’’ syndrome appears to depend on the 8. Feinstein AR. The Blame-X syndrome: Problems and commitment of the owner, the modifiability of the envi- lessons in nosology, spectrum, and etiology. J Clin Epidemiol 2001; ronment, and the severity of the disorder in the cat. 54:433–439. Additionally, cats seem to retain the underlying vulnera- 9. Osbaldiston GW, Taussig RA. Clinical report on 46 cases bility, however, even after long periods of time without of feline urological syndrome. Vet Med/Small Anim Clin 1970;65: expressing clinical signs, if exposed to sufficiently severe 461–468. stressors. 10. Osborne CA, Johnston GR, Polzin DJ, et al. Redefinition of the feline urologic syndrome: Feline lower urinary tract disease with heterogeneous causes. Vet Clin North Am Small Anim Pract Summary and Perspective 1984;14:409–438. 11. Buffington CAT, Chew DJ, Woodworth BE. Feline inter- Currently available evidence suggests that many cases of stitial cystitis. J Am Vet Med Assoc 1999;215:682–687. chronic idiopathic LUT signs presently diagnosed as hav- 12. Schur EA, Afari N, Furberg H, et al. Feeling bad in more ing FIC actually may have a ‘‘Pandora’’ syndrome. The ways than one: Comorbidity patterns of medically unexplained and syndrome might result from early adverse experiences that psychiatric conditions. J Gen Intern Med 2007;22:818–821. sensitize the neuraxis to sensory input, increasing the fre- 13. Ablin K, Clauw DJ. From fibrositis to functional somatic quency and duration of activation of the SRS when syndromes to a bell-shaped curve of pain and sensory sensitivity: the individual is housed in a provocative environment. Evolution of a clinical construct. Rheum Dis Clin North Am 2009; The chronic ‘‘wear and tear’’ of persistent activation of 35:233–251. the SRS, when superimposed on the (possibly familial) 14. Yunus MB. Central sensitivity syndromes: A new paradigm and group nosology for fibromyalgia and overlapping conditions, variability of organ involvement, possibly explains the in- 171 and the related issue of disease versus illness. Semin Arthritis consistency of comorbid disorder presentation. Rheum 2008;37:339–352. The available data only suggest this scenario, however, 15. Buffington CAT, Westropp JL, Chew DJ, et al. Clinical and permit generation of the hypothesis. Many of the find- evaluation of multimodal environmental modification (MEMO) in ings are based on data obtained from small numbers of the management of cats with idiopathic cystitis. J Feline Med Surg severely affected animals recruited because of the severity 2006;8:261–268. of their disease, and have not been independently repli- 16. Buffington CAT, Westropp JL, Chew DJ, et al. A case-con- cated. One might imagine a number of additional trol study of indoor-housed cats with lower urinary tract signs. complementary or alternative ‘‘systemic’’ hypotheses J Am Vet Med Assoc 2006;228:722–725. related to variable combinations of genetic, epigenetic, 17. Westropp JL, Kass PH, Buffington CAT. Evaluation of the effects of stress in cats with idiopathic cystitis. Am J Vet Res and environmental influences; these remain to be explored. 2006;67:731–736. 18. Stella JL, Lord LK, Buffington CAT. Sickness behaviors in Acknowledgments response to unusual external events in healthy cats and cats with fe- line interstitial cystitis. J Am Vet Med Assoc 2011;238:67–73. The author expresses his heartfelt thanks to the many 19. Buffington CAT. Bladder pain syndrome/interstitial cysti- mentors, colleagues, and students with whom he has tis. In: Baranowski AP, Abrams P, Fall M, eds. Urogenital Pain in worked on research into this syndrome for their advice, Clinical Practice. New York: Informa Healthcare, USA; 2008: collegiality, and effort. 169–183. Supported by the National Institutes of Health P50 20. Buffington CAT. Comorbidity of interstitial cystitis with DK64539 Women’s Health and Functional Visceral Dis- other unexplained clinical conditions. J Urol 2004;172:1242–1248. 21. Wu EQ, Birnbaum H, Kang YJ, et al. A retrospective orders Center, and DK47538. claims database analysis to assess patterns of interstitial cystitis di- agnosis. Curr Med Res Opin 2006;22:495–500. References 22. Warren JW, Howard FM, Cross RK, et al. Antecedent nonbladder syndromes in case-control study of interstitial cystitis/ 1. Westropp J, Buffington CAT. Lower urinary tract disor- painful bladder syndrome. Urology 2009;73:52–57. ders in cats. In: Ettinger SJ, Feldman EC, eds. Textbook of 23. Weissman MM, Gross R, Fyer A, et al. Interstitial cystitis Veterinary Internal Medicine, 7th ed. St. Louis, MO: Elsevier-Saun- and panic disorder: A potential genetic syndrome. Arch Gen ders; 2010:2069–2086. Psychiatry 2004;I61:273–279. 2. Osborne CA, Kruger JM, Lulich JP. Feline lower urinary 24. Dimitrakov JD. A case of familial clustering of interstitial tract disorders. Definition of terms and concepts. Vet Clin North cystitis and chronic pelvic pain syndrome. Urology 2001;58:281vi–viii. Am Small Anim Pract 1996;26:169–179. 25. Westropp JL, Welk KA, Buffington CAT. Small adrenal 3. Baranowski AP, Abrams P, Berger RE, et al. Urogenital glands in cats with feline interstitial cystitis. J Urol 2003;170: pain-time to accept a new approach to phenotyping and, as a con- 2494–2497. sequence, management. Eur Urol 2008;53:33–36. 26. Buffington CAT. Developmental influences on medically 4. Hanno P, Lin A, Nordling J, et al. Bladder pain syndrome unexplained symptoms. Psychother Psychosom 2009;78:139–144. committee of the international consultation on incontinence. Ne- 27. Clasper M. A case of interstitial cystitis and Hunner’s ulcer urourol Urodyn 2010;29:191–198. in a domestic shorthaired cat. N Z Vet J 1990;38:158–160.

199 Idiopathic Cystitis in Cats 793

28. Hanno PM. Painful bladder syndrome/interstitial cystitis 51. Griffin DW, Gregory CR. Prevalence of bacterial urinary and related diseases. In: Wein AJ, Kavoussi LR, Novick AC, Partin tract infection after perineal urethrostomy in cats. J Am Vet Med AW, Peters CA, eds. Campbell-Walsh Urology, 9th ed. Philadel- Assoc 1992;200:681–684. phia, PA: W.B. Saunders Company; 2007:330–370. 52. Nickel JC, Shoskes D, Irvine-Bird K. Clinical phenotyping 29. Fowler CJ, Griffiths D, de Groat WC. The neural control of of women with interstitial cystitis/painful bladder syndrome: A key micturition. Nat Rev Neurosci 2008;9:453–466. to classification and potentially improved management. J Urol 30. Birder LA, Wolf-Johnston AS, Chib MK, et al. Beyond 2009;182:155–160. neurons: Involvement of urothelial and glial cells in bladder func- 53. Warren JW, Brown V, Jacobs S, et al. Urinary tract infec- tion. Neurourol Urodyn 2010;29:88–96. tion and inflammation at onset of interstitial cystitis/painful bladder 31. Kanai A, Andersson KE. Bladder afferent signaling: Re- syndrome. Urology 2008;71:1085–1090. cent findings. J Urol 2010;183:1288–1295. 54. Hurst RE, Parsons CL, Roy JB, et al. Urinary glyco- 32. Holstege G. Micturition and the soul. J Comp Neurol saminoglycan excretion as a laboratory marker in the diagnosis of 2005;493:15–20. interstitial cystitis. J Urol 1993;149:31–35. 33. Powell-Boone T, Ness TJ, Cannon R, et al. Menstrual cycle 55. Wei DC, Politano VA, Selzer MG, et al. The association of affects bladder pain sensation in subjects with interstitial cystitis. elevated urinary total to sulfated glycosaminoglycan ratio and high J Urol 2005;174:1832–1836. molecular mass hyaluronic acid with interstitial cystitis. J Urol 34. Rajasekaran M, Stein P, Parsons CL. Toxic factors in hu- 2000;163:1577–1583. man urine that injure urothelium. Int J Urol 2006;13:409–414. 56. Hauser PJ, Dozmorov MG, Bane BL, et al. Abnormal 35. Buffington CA, Blaisdell JL, Binns SP Jr, et al. Decreased expression of differentiation related proteins and proteoglycan core urine glycosaminoglycan excretion in cats with interstitial cystitis. proteins in the urothelium of patients with interstitial cystitis. J Urol J Urol 1996;155:1801–1804. 2008;179:764–769. 36. Parsons CL, Schmidt JD, Pollen JJ. Successful treatment of 57. Pereira DA, Aguiar JAK, Hagiwara MK, et al. Changes in interstitial cystitis with sodium pentosanpolysulfate. J Urol 1983; cat urinary glycosaminoglycans with age and in feline urologic syn- 130:51–53. drome. Biochim Biophys Acta-Gen Subj 2004;1672:1–11. 37. Kruger JM, Osborne CA. The role of uropathogens in fe- 58. Press SM, Moldwin R, Kushner L, et al. Decreased expres- line lower urinary tract disease. Clinical implications. Vet Clin sion of GP-51 glycosaminoglycan in cats afflicted with feline North Am Small Anim Pract 1993;23:101–123. interstitial cystitis. J Urol 1995;153:288A. 38. Keay SK, Warren JW. Is interstitial cystitis an infectious 59. Liebert M. The urothelial glycosaminoglycan layer revis- disease? Int J Antimicrob Agents 2002;19:480–483. ited. J Urol 2009;182:2103–2104. 39. Argade SP, Vanichsarn C, Chenoweth M, et al. Abnormal 60. Chew DJ, Bartges JW, Adams LG, et al Randomized, pla- glycosylation of Tamm-Horsfall protein in patients with interstitial cebo-controlled clinical trial of pentosan polysulfate sodium for cystitis. BJU Int 2009;103:1085–1089. treatment of feline interstitial (idiopathic) cystitis. In: ACVIM Fo- 40. Keay S. Cell signaling in interstitial cystitis/painful bladder rum, Montreal, Quebec, JVIM, 2009, 674. syndrome. Cell Signal 2008;20:2174–2179. 61. Gunn-Moore DA, Shenoy CM. Oral glucosamine and the 41. Erickson DR, Tomaszewski JE, Kunselman AR, et al. management of feline idiopathic cystitis. J Feline Med Surg Urine markers do not predict biopsy findings or presence of blad- 2004;6:219–225. der ulcers in interstitial cystitis/painful bladder syndrome. J Urol 62. Holm-Bentzen M, Jacobsen F, Nerstrom B, et al. A pro- 2008;179:1850–1856. spective double-blind clinically controlled multicenter trial of 42. Goo YA, Tsai YS, Liu AY, et al. Urinary proteomics sodium pentosanpolysulfate in the treatment of interstitial cystitis evaluation in interstitial cystitis/painful bladder syndrome: and related painful bladder disease. J Urol 1987;138:503–507. A pilot study. Int Braz J Urol 2010;36:464–478; discussion 63. Dimitrakov J, Kroenke K, Steers WD, et al. Pharmacologic 478–469, 479. management of painful bladder syndrome/interstitial cystitis: A sys- 43. Raffi HS, Bates JM Jr, Laszik Z, et al. Tamm-Horsfall pro- tematic review. Arch Intern Med 2007;167:1922–1929. tein acts as a general host-defense factor against bacterial cystitis. 64. Nickel JC, Egerdie RB, Steinhoff G, et al. A multicenter, Am J Nephrol 2005;25:570–578. randomized, double-blind, parallel group pilot evaluation of the 44. Hunner GL. A rare type of bladder ulcer in women: Report efficacy and safety of intravesical sodium chondroitin sulfate versus of cases. Boston Med Soc J 1915;172:247–292. vehicle control in patients with interstitial cystitis/painful bladder 45. Fabricant CG. The feline urologic syndrome induced by in- syndrome. Urology 2010;76:804–809. fection with a cell-associated herpesvirus. Vet Clin North Am Small 65. Khandelwal P, Abraham SN, Apodaca G. Cell biology and Anim Pract 1984;14:493–502. physiology of the uroepithelium. Am J Physiol Renal Physiol 46. Kruger JM, Osborne CA, Lulich JP. Changing paradigms 2009;297:F1477–F1501. of feline idiopathic cystitis. Vet Clin North Am Small Anim Pract 66. Apodaca G. The uroepithelium: Not just a passive barrier. 2009;39:15–40. Traffic 2004;5:117–128. 47. Al-Hadithi HN, Williams H, Hart CA, et al. Absence of 67. Birder LA, Kanai AJ, Cruz F, et al. Is the urothelium intel- bacterial and viral DNA in bladder biopsies from patients with in- ligent? Neurourol Urodyn 2010;29:598–602. terstitial cystitis/chronic pelvic pain syndrome. J Urol 2005;174: 68. Gao X, Buffington CA, Au JL. Effect of interstitial cystitis 151–154. on drug absorption from urinary bladder. J Pharmacol Exp Ther 48. Stanford E, McMurphy C. There is a low incidence of re- 1994;271:818–823. current bacteriuria in painful bladder syndrome/interstitial cystitis 69. Lavelle JP, Meyers SA, Ruiz WG, et al. Urothelial patho- patients followed longitudinally. Int Urogynecol J Pelvic Floor physiological changes in feline interstitial cystitis: A human model. Dysfunct 2007;18:551–554. Am J Physiol Renal Physiol 2000;278:F540–F553. 49. Eggertsdottir AV, Lund HS, Krontveit R, et al. Bacteriuria 70. Veranic P, Jezernik K. The response of junctional com- in cats with feline lower urinary tract disease: A clinical study of 134 plexes to induced desquamation in mouse bladder urothelium. Biol cases in Norway. J Feline Med Surg 2007;9:458–465. Cell 2000;92:105–113. 50. Litster A, Thompson M, Moss S, et al. Feline bacterial uri- 71. Paus R, Theoharides TC, Arck PC. Neuroimmunoendo- nary tract : An update on an evolving clinical problem. crine circuitry of the ‘brain-skin connection’. Trends Immunol 2006; Vet J 2011;187:18–22. 27:32–39.

200 794 Buffington

72. Funakoshi K, Nakano M, Atobe Y, et al. Differential de- 93. Thilagarajah R, Witherow RO, Walker MM. Oral cimeti- velopment of TRPV1-expressing sensory nerves in peripheral dine gives effective symptom relief in painful bladder disease: A organs. Cell Tissue Res 2006;323:27–41. prospective, randomized, double-blind placebo-controlled trial. Br 73. Birder LA. Urinary bladder urothelium: Molecular sensors J Urol 2001;87:207–212. of chemical/thermal/mechanical stimuli. Vasc Pharmacol 2006; 94. Peeker R, Aldenborg F, Fall M. The treatment of intersti- 45:221–226. tial cystitis with supratrigonal, cystectomy and ileocystoplasty: 74. Birder LA, Wolf-Johnston A, Buffington CA, et al. Altered Difference in outcome between classic and nonulcer disease. J Urol inducible nitric oxide synthase expression and nitric oxide produc- 1998;159:1479–1482. tion in the bladder of cats with feline interstitial cystitis. J Urol 95. Roppolo JR, Tai C, Booth AM, et al. Bladder A-delta 2005;173:625–629. afferent nerve activity in normal cats and cats with feline interstitial 75. Birder LA, Barrick SR, Roppolo JR, et al. Feline interstitial cystitis. J Urol 2005;173:1011–1015. cystitis results in mechanical hypersensitivity and altered ATP 96. Parsons CL, Greenberger M, Gabal L, et al. The role of release from bladder urothelium. Am J Physiol-Renal Fluid Elec- urinary potassium in the pathogenesis and diagnosis of interstitial trolyte Physiol 2003;285:F423–F429. cystitis. J Urol 1998;159:1862–1866. 76. Buffington CAT, Chew DJ, Woodworth BE. Animal model 97. Hohlbrugger G. Leaky urothelium and/or vesical ischemia of human disease—feline interstitial cystitis. Comp Path Bull 1997; enable urinary potassium to cause idiopathic urgency/frequency 29:3–6. syndrome and urge incontinence. Int Urogynecol J Pelvic Floor 77. Tomaszewski JE, Landis JR, Russack V, et al. Biopsy Dysfunct 1996;7:242–255. features are associated with primary symptoms in interstitial 98. Caito M. The presence and distribution of substance P and cystitis: results from the interstitial cystitis database study. Urology neuron-specific enolase in the feline bladder. Columbus, OH: The 2001;57:67–81. Ohio State University; 1995. MS Thesis Veterinary Anatomy. 78. Straub RH, Cutolo M. Involvement of the hypothalamic– 99. Hohenfellner M, Nunes L, Schmidt RA, et al. Interstitial pituitary–adrenal/gonadal axis and the peripheral nervous system in cystitis: Increased sympathetic innervation and related neuropep- rheumatoid arthritis: Viewpoint based on a systemic pathogenetic tide synthesis. J Urol 1992;147:587–591. role. Arthritis Rheum 2001;44:493–507. 100. Peeker R, Aldenborg F, Dahlstrom A, et al. Increased 79. Buffington CAT, Teng BY, Somogyi GT. Norepinephrine tyrosine hydroxylase immunoreactivity in bladder tissue from content and adrenoceptor function in the bladder of cats with feline patients with classic and nonulcer interstitial cystitis. J Urol 2000; interstitial cystitis. J Urol 2002;167:1876–1880. 163:1112–1115. 80. van de Merwe JP, Nordling J, Bouchelouche P, et al. Diag- 101. Buffington CAT, Wolfe SA High affinity binding sites for nostic criteria, classification, and nomenclature for painful bladder [3H]Substance P in urinary bladders of cats with interstitial cystitis. syndrome/interstitial cystitis: An ESSIC proposal. Eur Urol 2008; J Urol 1998;160:605–611. 53:60–67. 102. Marchand JE, Sant GR, Kream RM. Increased expression 81. Geddy PM, Gray S, Reid WA. Mast-cell density and pgp of substance P receptor-encoding mRNA in bladder biopsies from 9.5-immunostained nerves in angioleiomyoma—their relationship patients with interstitial cystitis. Br J Urol 1998;81:224–228. to painful symptoms. Histopathology 1993;22:387–390. 103. Sanchez Freire V, Burkhard FC, Kessler TM, et al. Micro- 82. Theoharides TC, Cochrane DE. Critical role of mast cells in RNAs may mediate the down-regulation of neurokinin-1 receptor inflammatory diseases and the effect of acute stress. J Ne- in chronic bladder pain syndrome. Am J Path 2010;176:288–303. uroimmunol 2004;146:1–12. 104. Rupniak NMJ, Kramer MS. Discovery of the anti- 83. Foster HE Jr, Hanno PM, Nickel JC, et al. Effect of am- depressant and anti-emetic efficacy of substance P receptor (NK1) itriptyline on symptoms in treatment naive patients with interstitial antagonists. Trends Pharmacol Sci 1999;20:485–490. cystitis/painful bladder syndrome. J Urol 2010;183:1853–1858. 105. Hill R. Reply: Will changing the testing paradigms show 84. Theoharides TC, Papaliodis D, Tagen M, et al. Chronic fa- that NK1 receptor antagonists are analgesic in humans? Trends tigue syndrome, mast cells, and tricyclic antidepressants. J Clin Pharmacol Sci 2000;21:465–465. Psychopharmacol 2005;25:515–520. 106. Sun NN, Wong SS, Keith I, et al. Tachykinin substance P 85. Specht A, Kruger JM, Fitzgerald SD, et al. Histochemical depletion by capsaicin exacerbates inflammatory response to side- and immunohistochemical features of chronic feline idiopathic stream cigarette smoke in rats. Toxicology 2004;201:39–50. cystitis. J Vet Intern Med 2004;18:416. 107. Okajima K, Harada N. Regulation of inflammatory 86. Elbadawi A. Interstitial cystitis: A critique of current responses by sensory neurons: Molecular mechanism(s) and possi- concepts with a new proposal for pathologic diagnosis and patho- ble therapeutic applications. Curr Med Chem 2006;13:2241–2251. genesis. Urology 1997;49:14–40. 108. Sculptoreanu A, deGroat WC, Buffington CAT, et al. Ab- 87. Buffington CAT, Chew DJ. Presence of mast cells in sub- normal excitability in capsaicin-responsive DRG neurons from cats mucosa and detrusor of cats with idiopathic lower urinary tract with feline interstitial cystitis. Exp Neurol 2005;193:437–443. disease. J Vet Intern Med 1993;7:126A. 109. Wein AJ. Resiniferatoxin in the treatment of interstitial 88. Hannan MT, Felson DT, Pincus T. Analysis of the discor- cystitis: A systematic review—editorial comment. J Urol 2009;182: dance between radiographic changes and knee pain in osteoarthritis 2328–2329. of the knee. J Rheumatol 2000;27:1513–1517. 110. March PA, Buffington CAT. Effects of resiniferatoxin on 89. Waxman JA, Sulak PJ, Kuehl TJ. Cystoscopic findings vesico-sympathetic reflexes and cortical activation in cats with feline consistent with interstitial cystitis in normal women undergoing interstitial cystitis. Soc Neurosci Abst 2001;27. tubal ligation. J Urol 1998;160:1663–1667. 111. Payne CK, Mosbaugh PG, Forrest JB, et al. Intravesical 90. Levine LA, Richie JP. Urological complications of resiniferatoxin for the treatment of interstitial cystitis: A ran- cyclophosphamide. J Urol 1989;141:1063–1069. domized, double-blind, placebo controlled trial. J Urol 2005;173: 91. Chang L, Munakata J, Mayer EA, et al. Perceptual re- 1590–1594. sponses in patients with inflammatory and functional bowel disease. 112. Toft BR, Nordling J. Recent developments of intravesical Gut 2000;47:497–505. therapy of painful bladder syndrome/interstitial cystitis: A review. 92. Chew DJ, Buffington CAT, Kendall MS, et al. Amitripty- Curr Opin Urol 2006;16:268–272. line treatment for idiopathic cystitis in cats (15 cases 1994–1996). J 113. Caston HT. Stress and the feline urological syndrome. Fe- Am Vet Med Assoc 1998;213:1282–1286. line Prac 1973;3:14–22.

201 Idiopathic Cystitis in Cats 795

114. Kirk H. Retention of urine and urine deposits. In: Kirk H, responsiveness to visceral related threat. J Urol 2009;181: ed. The Diseases of the Cat and its General Management. London: 2127–2133. Bailliere, Tindall and Cox; 1925:261–267. 136. Kilpatrick LA, Ornitz E, Ibrahimovic H, et al. Gating of 115. Buffington CAT. External and internal influences on dis- sensory information differs in patients with interstitial cystitis/pain- ease risk in cats. J Am Vet Med Assoc 2002;220:994–1002. ful bladder syndrome. J Urol 2010;184:958–963. 116. Cameron ME, Casey RA, Bradshaw JWS, et al. A study of 137. Westropp JL, Kass PH, Buffington CAT. In vivo evalua- environmental and behavioural factors that may be associated with tion of the alpha-2 adrenoceptors in cats with idiopathic cystitis. feline idiopathic cystitis. J Sm Anim Pract 2004;45:144–147. Am J Vet Res 2007;68:203–207. 117. Jones BR, Sanson RL, Morris RS. Elucidating the risk 138. Birder L. Role of the urothelium in bladder function. Scand factors of feline urologic syndrome. N Z Vet J 1997;45:100–108. J Urol Nephrol 2004;215(Suppl):48–53. 118. Koziol JA, Clark DC, Gittes RF, et al. The natural history 139. Wolf-Johnston AS, Buffington A, Roppolo JR, et al. Con- of interstitial cystitis: A survey of 374 patients. J Urol 1993;149: sequences of elevated NGF levels in urinary bladder and sensory 465–469. neurons in feline interstitial cystitis and cyclophosphamide induced 119. Rothrock NE, Lutgendorf SK, Kreder KJ, et al. Stress and cystitis. J Urol 2004;171:96. symptoms in patients with interstitial cystitis: A life stress model. 140. Jacobs BL, Smaldone MC, Tyagi V, et al. Increased nerve Urology 2001;57:422–427. growth factor in neurogenic overactive. Bladder and interstitial 120. Reche AJ, Buffington CAT. Increased tyrosine hydroxylase cystitis patients. Can J Urol 2010;17:4989–4994. immunoreactivity in the locus coeruleus of cats with interstitial 141. Birder LA, Wolf-Johnston A, Griffiths D, et al. Role cystitis. J Urol 1998;159:1045–1048. of urothelial nerve growth factor in human bladder function. 121. Welk KA, Buffington CAT. Effect of interstitial cystitis on Neurourol Urodyn 2007;26:405–409. central neuropeptide and receptor immunoreactivity in cats. In: 142. Birder LA, Hanna-Mitchell AT, Mayer EA, et al. Cystitis, Nyberg L, ed. Research Insights into Interstitial Cystitis. Alexan- epithelial dysfunction and co-morbid disorders. Neurourol Urodyn dria, VA: National institute of Diabetes and Digestive and Kidney 2011. In press. Diseases; 2003:74. 143. Fields H. State-dependent opioid control of pain. Nat Rev 122. Sands SA, Strong R, Corbitt J, et al. Effects of acute re- Neurosci 2004;5:565–575. straint stress on tyrosine hydroxylase mRNA expression in locus 144. Schommer NC, Hellhammer DH, Kirschbaum C. Dissoci- coeruleus of Wister and Wistar-Kyoto rats. Mol Brain Res 2000;75: ation between reactivity of the hypothalamus-pituitary-adrenal axis 1–7. and the sympathetic-adrenal-medullary system to repeated psycho- 123. Chrousos GP, Gold PW. The concepts of stress and stress social stress. Psychosom Med 2003;65:450–460. system disorders. J Am Med Assoc 1992;267:1244–1252. 145. Kvetnansky R, Fukuhara K, Pacak K, et al. Endogenous 124. Van Bockstaele EJ, Peoples J, Valentino RJ. A.E. Bennett glucocorticoids restrain catecholamine synthesis and release at rest Research Award. Anatomic basis for differential regulation of and during immobilization stress in rats. Endocrinology the rostrolateral peri-locus coeruleus region by limbic afferents. 1993;133:1411–1419. Biol Psychiatry 1999;46:1352–1363. 146. Westropp JL, Buffington CAT. Effect of a corticotropin re- 125. Buffington CA, Pacak K Increased plasma norepinephrine leasing factor (crf) antagonist on hypothalamic-pituitary-adrenal concentration in cats with interstitial cystitis. J Urol 2001;165: activation in response to crf in cats with interstitial cystitis. In: 2051–2054. Nyberg L, ed. Research Insights into Interstitial Cystitis. Alexan- 126. Ullrich PM, Turner JA, Ciol M, et al. Stress is associated dria, VA: National Institute of Diabetes and Digestive and Kidney with subsequent pain and disability among men with nonbacterial Diseases; 2003:73. prostatitis/pelvic pain. Ann Behav Med 2005;30:112–118. 147. Heim C, Ehlert U, Hellhammer DH. The potential role of 127. Whitmore K. Self-care regimens for patients with intersti- hypocortisolism in the pathophysiology of stress-related bodily dis- tial cystitis. Urol Clin N Am 1994;21:121–130. orders. Psychoneuroendocrinology 2000;25:1–35. 128. McEwen BS. Central effects of stress hormones in 148. Lutgendorf SK, Latini JM, Rothrock N, et al. Auto- health and disease: Understanding the protective and damaging nomic response to stress in interstitial cystitis. J Urol 2004;172: effects of stress and stress mediators. Eur J Pharmacol 2008;583: 227–231. 174–185. 149. Westropp JL, Buffington CAT. Cerebrospinal fluid 129. Ursin H, Eriksen HR. The cognitive activation theory of corticotrophin releasing factor and catecholamine concentrations stress. Psychoneuroendocrinology 2004;29:567–592. in healthy cats and cats with interstitial cystitis. In: Nyberg L, ed. 130. Broom DM. Animal welfare defined in terms of attempts to Research Insights into Interstitial Cystitis. Alexandria, VA: Na- cope with the environment. Acta Agric Scand Sect A-Anim Sci tional Institute of Diabetes and Digestive and Kidney Diseases; 1996;27(Suppl):22–28. 2003:72. 131. Korte SM, Koolhaas JM, Wingfield JC, et al. The Darwin- 150. Dimitrakov J, Joffe HV, Soldin SJ, et al. Adrenocortical ian concept of stress: Benefits of allostasis and costs of allostatic hormone abnormalities in men with chronic prostatitis/chronic pel- load and the trade-offs in health and disease. Neurosci Biobehav vic pain syndrome. Urology 2008;71:261–266. Rev 2005;29:3–38. 151. Fries E, Hesse J, Hellhammer J, et al. A new view on hypo- 132. Miller AH, Maletic V, Raison CL. Inflammation and its cortisolism. Psychoneuroendocrinology 2005;30:1010–1016. discontents: The role of cytokines in the pathophysiology of major 152. Buijs RM, Wortel J, Van Heerikhuize JJ, et al. Anatomical depression. Biol Psychiatry 2009;65:732–741. and functional demonstration of a multisynaptic suprachiasma- 133. Davis M. Neural systems involved in fear and anxiety mea- tic nucleus adrenal (cortex) pathway. Eur J Neurosci 1999;11: sured with fear-potentiated startle. Am Psychol 2006;61:741–756. 1535–1544. 134. Buffington CAT. Effect of feline interstitial cystitis on 153. Engeland WC, Arnhold MM. Neural circuitry in the regu- acoustic startle response. Program No. 670.14. 2002 Neuroscience lation of adrenal corticosterone rhythmicity. Endocrine 2005;28: Meeting Planner. Orlando, FL: Society for Neuroscience 2002. 325–332. Online. 154. Carlstead K, Brown JL, Strawn W. Behavioral and physi- 135. Twiss C, Kilpatrick L, Craske M, et al. Increased startle re- ological correlates of stress in laboratory cats. Appl Anim Behav Sci sponses in interstitial cystitis: Evidence for central hyper- 1993;38:143–158.

202 796 Buffington

155. Carlstead K, Shepherdson DS. Alleviating stress in zoos 176. Weaver IC, Champagne FA, Brown SE, et al. Reversal of with environmental enrichment. In: Moberg GP, Mench JA, eds. maternal programming of stress responses in adult offspring The Biology of Animal Stress: Basic Principles and Implications for through methyl supplementation: Altering epigenetic marking later Animal Welfare. New York: CABI Publishing; 2000:337–354. in life. J Neurosci 2005;25:11045–11054. 156. Dantzer R, O’Connor JC, Freund GG, et al. From inflam- 177. Tsankova NM, Berton O, Renthal W, et al. Sustained hip- mation to sickness and depression: When the immune system pocampal chromatin regulation in a mouse model of depression and subjugates the brain. Nat Rev Neurosci 2008;9:46–56. antidepressant action. Nat Neurosci 2006;9:519–525. 157. Marques-Deak A, Cizza G, Sternberg E. Brain-immune 178. Card JP, Levitt P, Gluhovsky M, et al. Early experience interactions and disease susceptibility. Mol Psychiatr 2005;10:239–250. modifies the postnatal assembly of autonomic emotional motor cir- 158. Raison CL, Miller AH. When not enough is too much: The cuits in rats. J Neurosci 2005;25:9102–9111. role of insufficient glucocorticoid signaling in the pathophysiology 179. Holmes A, le Guisquet AM, Vogel E, et al. Early life ge- of stress-related disorders. Am J Psychiatry 2003;160:1554–1565. netic, epigenetic and environmental factors shaping emotionality in 159. Strouse TB. The relationship between cytokines and pain/ rodents. Neurosci Biobehav Rev 2005;29:1335–1346. depression: A review and current status. Curr Pain Headache Rep 180. Randich A, Uzzell T, DeBerry JJ, et al. Neonatal urinary 2007;11:98–103. bladder inflammation produces adult bladder hypersensitivity. 160. Bierhaus A, Wolf J, Andrassy M, et al. A mechanism con- J Pain 2006;7:469–479. verting psychosocial stress into mononuclear cell activation. Proc 181. Al-Chaer ED, Kawasaki M, Pasricha PJ. A new model of Natl Acad Sci U S A 2003;100:1920–1925. chronic visceral hypersensitivity in adult rats induced by colon 161. Kumar A, Takada Y, Boriek AM, et al. Nuclear factor- irritation during postnatal development. Gastroenterology 2000;119: kappa B: Its role in health and disease. J Mol Med 2004;82:434–448. 1276–1285. 162. Wen Y, Yang SH, Liu R, et al. Estrogen attenuates nuclear 182. Barreau F, Ferrier L, Fioramonti J, et al. New insights factor-kappa B activation induced by transient cerebral ischemia. in the etiology and pathophysiology of irritable bowel syndrome: Brain Res 2004;1008:147–154. Contribution of neonatal stress models. Pediatr Res 2007;62: 163. Hayashi R, Wada H, Ito K, et al. Effects of glucocorticoids 240–245. on gene transcription. Eur J Pharmacol 2004;500:51–62. 183. Newport DJ, Heim C, Bonsall R, et al. Pituitary-adrenal 164. Turrin NP, Rivest S. Unraveling the molecular details in- responses to standard and low-dose dexamethasone suppression volved in the intimate link between the immune and neuroendocrine tests in adult survivors of child abuse. Biol Psychiatry 2004;55: systems. Exp Biol Med 2004;229:996–1006. 10–20. 165. Nadeau S, Rivest S. Glucocorticoids play a fundamental 184. Heim C, Newport DJ, Heit S, et al. Pituitary-adrenal and role in protecting the brain during innate immune response. J autonomic responses to stress in women after sexual and physical Neurosci 2003;23:5536–5544. abuse in childhood. J Am Med Assoc 2000;284:592–597. 166. Koulich E, Nguyen T, Johnson K, et al. NF-kappa B is in- 185. Lackner JM, Gudleski GD, Blanchard EB. Beyond abuse: volved in the survival of cerebellar granule neurons: Association of I The association among parenting style, abdominal pain, and soma- kappa beta phosphorylation with cell survival. J Neurochem tization in IBS patients. Behav Res Ther 2004;42:41–56. 2001;76:1188–1198. 186. Gunnar M, Quevedo K. The neurobiology of stress and de- 167. Freeman LM, Brown DJ, Smith FW, et al. Magnesium velopment. Annu Rev Psychol 2007;58:145–173. status and the effect of magnesium supplementation in feline hyper- 187. Heim C, Newport DJ, Wagner D, et al. The role of early trophic cardiomyopathy. Can J Vet Res 1997;61:227–231. adverse experience and adulthood stress in the prediction of neuro- 168. Rodriguez MA, Afari N, Buchwald DS. Evidence for over- endocrine stress reactivity in women: A multiple regression analysis. lap between urological and nonurological unexplained clinical Depress Anxiety 1999;15:117–125. conditions. J Urol 2009;182:2123–2131. 188. Oemler M, Grabhorn R, Vahlensieck W, et al. Psychosocial 169. Peters KM, Carrico DJ, Diokno AC. Characterization of a aspects of interstitial cystitis: Do biographical factors have a rele- clinical cohort of 87 women with interstitial cystitis/painful bladder vant impact on the disease course? Urologe A 2006;45:728–733. syndrome. Urology 2008;71:634–640. 189. Schmidt MV. Molecular mechanisms of early life stress-Les- 170. Clemens JQ, Meenan RT, O’Keeffe Rosetti MC, et al. sons from mouse models. Neurosci Biobehav Rev 2010;34:845–852. Case-control study of medical comorbidities in women with inter- 190. Bale TL, Baram TZ, Brown AS, et al. Early life programming stitial cystitis. J Urol 2008;179:2222–2225. and neurodevelopmental disorders. Biol Psychiatry 2010;68:314–319. 171. McEwen BS, Gianaros PJ. Central role of the brain in stress 191. Korosi A, Shanabrough M, McClelland S, et al. Early-life and adaptation: Links to socioeconomic status, health, and disease. In: experience reduces excitation to stress-responsive hypothalamic Adler NE, Stewart J, eds. Biology of Disadvantage: Socioeconomic neurons and reprograms the expression of corticotropin-releasing Status and Health. Oxford, UK: Blackwell Publishing; 2010:190–222. hormone. J Neurosci 2010;30:703–713. 172. Gluckman PD, Hanson MA The conceptual basis for the 192. Lupien SJ, McEwen BS, Gunnar MR, et al. Effects of stress developmental origins of health and disease. In: Gluckman PD, throughout the lifespan on the brain, behaviour and cognition. Nat Hanson MA, eds. Developmental Origins of Health and Disease. Rev Neurosci 2009;10:434–445. Cambridge, UK: Cambridge University Press; 2006:33–50. 193. Kirkengen AL, Ulvestad E. Heavy burdens and complex 173. Matthews SG. Early programming of the hypothalamo- disease—an integrated perspective. J Norwegian Med Assoc 2007; pituitary-adrenal axis. Trends Endocrinol Metab 2002;13:373–380. 127:3228–3231. 174. Coplan JD, Smith EL, Altemus M, et al. Variable foraging 194. Yehuda R, Flory JD, Pratchett LC, et al. Putative biolog- demand rearing: Sustained elevations in cisternal cerebrospinal fluid ical mechanisms for the association between early life adversity and corticotropin-releasing factor concentrations in adult primates. Biol the subsequent development of PTSD. Psychopharmacology (Ber- Psychiatry 2001;50:200–204. lin) 2010;212:405–417. 175. Meaney MJ, Szyf M. Maternal care as a model for experi- 195. Cooper ES, Owens TJ, Chew DJ, et al. A protocol for man- ence-dependent chromatin plasticity? Trends Neurosci 2005;28: aging urethral obstruction in male cats without urethral catheter- 456–463. ization. J Am Vet Med Assoc 2010;237:1261–1266.

203 6th Tufts' Canine & Feline Breeding and Genetics Conference

Friday Saturday Lectures Sunday Lectures 27-Sep 28-Sep 29-Sep 7:00-8:00 Breakfast & Registration 7:00-8:00 Breakfast 8:00-8:10 Introductions and Announcements 8:00-8:10 Introductions and Announcements 8:10-8:30 Illumina Presentation Breeds As Populations Canine Hip Dysplasia Unraveling the Sources of Genetic Structure Within Half A Century with Canine Hip Dysplasia 8:30-9:10 8:10-8:50 Breeds - Dr. Pam Wiener - Dr. Åke Hedhammar Taking Advantage of Dog Breed Structure to The Othopedic Foundation for Animals Hip Displasia 9:10-9:50 8:50-9:30 Understand Health - Dr. Elaine Ostrander Database: A Review - Dr. Greg Keller 9:50-10:10 Break 9:30-9:50 Break Genetics of Cat Populations and Breeds: Implications 10:10-10:30 The genetics of hip dysplasia and implications for for Breed Management for Health! - Dr. Leslie Lyons 9:50-10:30 selection - Dr. Tom Lewis Breeding Practices According to Breeds; Time, Place, 10:30-11:10 and Consequences - Dr. Grégoire Leroy Genetic and Genomic Tools for Breeding Dogs With 10:30-11:10 Inbreeding, Outbreeding, and Breed Evolution - Healthy Hips - Dr. Rory Todhunter 11:10-11:30 Dr. Jerold Bell

11:30-12:30 Panel Discussion 11:10-12:10 Panel Discussion 12:30-1:15 Lunch 12:10-12:55 Lunch Genetic Disorders Management of Genetic Disease Unraveling the Phenotypic and Genetic Complexity of Holistic Management of Genetic Traits - 1:15-1:55 12:55-1:35 Canine Cystinuria - Dr. Paula Henthorn Dr. Anita Oberbauer From FUS to Pandora Syndrome - The Role of How to Use and Interpret Genetic Tests for Heart 1:55-2:35 1:35-2:15 Epigenetics and Environment in Pathophysiology, Disease in Cats and Dogs - Dr. Kate Meurs Treatment, and Prevention - Dr. Tony Buffington 2:35-2:55 Break 2:15-2:35 Break Breed Specific Breeding Strategies - Update on Genetic Tests for Diseases and Traits in 2:35-2:55 Dr. Åke Hedhammar 2:55-3:35 Cats: Implications for Cat Health, Breed Management UK initiatives for breeding healthier pedigree dogs - and Human Health - Dr. Leslie Lyons 2:55-3:15 Dr. Tom Lewis Hereditary Gastric Cancer in Dogs - Genetic Tests: Understanding Their Power, and Using 3:35-4:15 3:15-3:55 Dr. Elizabeth McNiel Their Force for Good - Dr. Jerold Bell 4:15-5:15 Panel Discussion 3:55-4:55 Panel Discussion

6:00-8:00 Registration