Genetica Pedigree

AJVR, Vol 63, No. 7, July 2002 1029

Canine hip dysplasia (CHD) is a common developmental

trait that affects primarily large breed dogs

and is characterized by poor hip joint congruity, joint

laxity and subluxation, and development of secondary

coxofemoral joint osteoarthritis (OA).1-3 Hip dysplasia

is a quantitative trait, the expression of which is influenced

by genetic,3,4 nutritional,5 and possibly hormonal

factors.6 Heritability estimates for CHD range from

0.11 to 0.68.7-9 The complex pattern of inheritance of

CHD suggests that expression of the trait is controlled

by genes located at several quantitative trait loci

(QTL).10 Alleles that contribute to the development of

a complex trait such as CHD may act in an additive or

dominant fashion, and the magnitude of the effect of

an individual locus is independent of its mode of

inheritance. Development of CHD undoubtedly results

from complex interactions among multiple genetic loci

and environmental factors. Nevertheless, a few major

QTL are likely to be involved in trait expression, and a

single major locus may have a substantial influence on

this trait on the basis of biometric methods outlined by

Leighton.4

Hip dysplasia is most commonly diagnosed by

examination of ventrodorsal hip-extended radiographic

views of the pelvis. Radiographic criteria for subjective

grading of dogs on the basis of identification of

dysplastic conformational features and OA have been

proposed.11 Selective breeding programs determined on

the basis of this form of radiographic assessment have

been applied in a number of dog populations; however,

because of the modest sensitivity of this test in

immature dogs, success in reducing the prevalence of

CHD has been limited.4,9,12-14 Alternative methods for

measurement of the dysplastic phenotype include the

distraction index (DI),2,15 dorsolateral subluxation

(DLS) score,16 and radiographic determination of the

age of detection of femoral capital epiphyseal ossification

(OSS).17,18 These tests measure different features

of hip conformation and differ with respect to their

sensitivity and specificity as predictors of development

of hip joint OA in an experimental pedigree.19

Because neither the genetic potential to develop

CHD nor the carrier status of an individual dog can be

unequivocally inferred from its phenotype, selective

breeding programs determined on the basis of phenotypic

evaluation of adult dogs are likely to remain inefficient.

Genetic testing would aid breeders and

Received Oct 26, 2002.

Accepted Jan 8, 2002.

From the Department of Clinical Sciences (Bliss, Todhunter,

Hamilton, Dykes, Yeager, Gilbert), James A. Baker Institute for

Animal Health (Lust, Williams, Burton-Wurster), and Center for

Canine Genetics and Reproduction (Acland), College of Veterinary

Medicine, and Animal Breeding (Quaas) and Biometrics (Casella)

Units, College of Agricultural and Life Sciences, Cornell

University, Ithaca, NY 14853; and Department of Statistics,

University of Florida, Gainesville, FL 32611 (Casella, Wu).

Supported by a grant from the Morris Animal Foundation, NIH

grant AR36554, the Consolidated Research Program, and College

of Veterinary Medicine Unrestricted Alumni Funds.

Address for correspondence to Dr. Bliss.

Quantitative genetics of traits associated

with hip dysplasia in a canine pedigree

constructed by mating dysplastic Labrador

Retrievers with unaffected Greyhounds

Stuart Bliss, DVM; Rory J. Todhunter, BVSc, PhD; Richard Quaas, PhD; George Casella, PhD;

Rongling Wu, PhD; George Lust, PhD; Alma Jo Williams, BS; Samuel Hamilton, BVSc;

Nathan L. Dykes, DVM; Amy Yeager, DVM; Robert O. Gilbert, BVSc, MMedVet;

Nancy I. Burton-Wurster, PhD; Gregory M. Acland, DVM

Objective—To determine the genetic influence on

expression of traits associated with canine hip dysplasia.

Animals—193 dogs from an experimental canine

pedigree.

Procedure—An experimental canine pedigree was

developed for linkage analysis of hip dysplasia by mating

dysplastic Labrador Retrievers with nondysplastic

Greyhounds. A statistical model was designed to test

the effects of Labrador Retriever and Greyhound alleles

on age at detection of femoral capital epiphyseal

ossification, 8-month distraction index, and 8-month

dorsolateral subluxation score.

Results—The additive effect was significant for age

at detection of femoral capital epiphyseal ossification.

Restricted maximum likelihood estimates (± SD)

for this trait were 6.4 ± 1.95, 10.2 ± 2.0, 10.8 ± 3.1,

11.4 ± 2.1, and 13.6 ± 4.6 days of age for

Greyhounds, Greyhound backcross dogs, F1 dogs,

Labrador Retriever backcross dogs, and Labrador

Retrievers, respectively. The additive effect was also

significant for the distraction index. Estimates for this

trait were 0.21 ± 0.07, 0.29 ± 0.15, 0.44 ± 0.12, 0.52

± 0.18, and 0.6 ± 0.17 for the same groups, respectively.

For the dorsolateral subluxation score, additive

and dominance effects were significant. Estimates

for this trait were 73.5 ± 4.1, 71.3 ± 6.5, 69.1 ± 6.0,

50.6 ± 12.9, and 48.4 ± 7.7%, respectively, for the

same groups.

Conclusions—In this canine pedigree, traits associated

with canine hip dysplasia are heritable. Phenotypic

differences exist among founder dogs of each breed

and their crosses. This pedigree should be useful for

identification of quantitative trait loci underlying the

dysplastic phenotype. (Am J Vet Res 2002;63:

1029–1035)

prospective owners in selection of immature dogs that

do not carry susceptibility alleles for CHD.

Unfortunately, the molecular genetic basis of CHD is

unknown, and elucidation of the QTL that underlie

expression of a complex trait such as CHD is a formidable

task. Linkage analysis is a statistical method for

mapping and identification of QTL that has been used

extensively in experimental settings as well as for

genetic analysis of several human diseases.20,21 In veterinary

medicine, linkage analysis has led to the identification

of loci22-25 or mutations26 underlying a number

of monogenic disorders. However, linkage analysis for

identification of QTL associated with a complex canine

trait of clinical veterinary importance has not been

reported.

Detection of linkage between a set of polymorphic

genetic markers and a complex trait requires that a

pedigree be informative with respect to mapped genetic

markers and the QTL associated with the trait of

interest.27 Informativeness refers to the overall confidence

with which the parental source of a marker allele

or a QTL can be determined for members of a pedigree.

The presence of significant measurable phenotypic differences

between members of a pedigree is a necessary

condition for observation of linkage and thus is an

indication of QTL informativeness. We have developed

an outcrossed pedigree for linkage analysis of CHD.28

The purpose of the study reported here was to describe

a biometrical model for evaluation of the differences

among the founder breed dogs and their crosses in this

pedigree with respect to the traits OSS, DI, and DLS

score. Our results demonstrate that the pedigree is

informative with respect to the QTL underlying

expression of these traits. This information provides a

necessary foundation for the development of statistical

models for linkage analysis that incorporate molecular

data from genome-wide analysis of genetic markers.

Materials and Methods

Pedigree—A canine pedigree was constructed for linkage

analysis of CHD by outcrossing dysplastic Labrador

Retrievers (LR) with unaffected Greyhounds (GH).28 Briefly,

trait-free GH (2 males and 5 females) were purchased from

racing stock. Dysplastic LR (5 males and 3 females) were

selected from an inbred pedigree that has been maintained at

the James A. Baker Institute for Animal Health, Cornell

University, for the study of CHD since 1968.1 Seven dogs

from the first filial generation (F1) were bred back to GH

and LR parents of other F1 litters or intercrossed to encourage

maximum recombination across the entire genome

(Fig 1).28 Two GH founder dogs were bred to each other and

produced a litter of 9 puppies; measurements from these

dogs were used to estimate the phenotype of GH founder

dogs at 8 months of age. Siblings of 6 LR founder dogs

(n = 22) were similarly used to improve the accuracy of 8-

month trait value estimates for this breed.

Dogs were bred by artificial insemination on the basis of

an increase in serum progesterone concentrations. Dogs were

inseminated between days 3 and 6 after their serum progesterone

concentrations reached 1.0

...