Materials and Methods

Sample population—To clarify the heritability and predict the mode of inheritance of cataracts and primary lens luxation in Jack Russell Terriers, a study was initiated in conjunction with regional and national Jack Russell Terrier clubs in the United States. Owners of Jack Russell Terriers voluntarily submitted survey questionnaires that requested registered name, sire and dam, date of birth, sex, and results of an ophthalmic evaluation provided on a CERF form. The CERF evaluation requires the use of pharmacologic mydriasis, slitlamp biomicroscopy, and indirect ophthalmoscopy and that the evaluation be performed by a board-certified veterinary ophthalmologist. Dogs were designated as affected with cataracts or lens luxation in the database when they had clinical signs consistent with the American College of Veterinary Ophthalmologists definition of heritable cataracts or lens luxation during the CERF examination. A dog was considered unaffected with these disorders when the eyes were certified as clinically normal during the CERF examination. When dogs were examined more than once, data from the most recent examination were used. Pedigree information was obtained from American Kennel Club records. In addition to the survey data, owners submitted buccal swab specimens obtained from each dog, which were used for isolation of genomic DNA.²² The study protocol was reviewed and approved by the Institutional Animal Care and Use Committee at the University of California, Davis.

Phenotypic data—Phenotypes for cataracts and lens luxation were determined by board-certified veterinary ophthalmologists, which enabled an accurate and repeatable assessment of ophthalmic disease. Prevalence data were collected on 872 dogs. Because of the need for pedigree information, an additional 1,898 dogs were included in the analyses. These additional dogs did not have a notation of cataracts or lens luxation in the records, but they were used to complete pedigree relationships. Of all dogs included in the analyses, 125 were inbred, with a mean inbreeding coefficient of 0.051.

No additional phenotypic information (eg, coat color) was collected in conjunction with the diagnosis of ophthalmic disease. Sex was the only additional variable included in the analysis. Because of the nature of the onset of cataracts and lens luxation, and to prevent inclusion of dogs that had not yet expressed the disorder because of age, dogs < 6 years old that were judged to be free of ophthalmic disease were included in the data analyses, but their phenotypic classification was defined as unknown. Therefore, there were 137 females > 6 years old with a recorded diagnosis of cataracts (114 unaffected and 23 affected) and 88 males > 6 years old with a recorded diagnosis of cataracts (68 unaffected and 20 affected). A diagnosis of lens luxation was recorded for 137 females (111 unaffected and 26 affected) and 80 males (62 unaffected and 18 affected) that were > 6 years of age.

Estimation of heritability and genetic correlation—For the objective of estimating the heritability of each disease, in addition to the genetic correlation between lens luxation and cataracts in Jack Russell Terriers, threshold models were used to assess the risk of disease. This method assumed that a dog could be assigned to a specific disease class (unaffected or affected) when an underlying, unobservable risk (or liability) for disease exceeded τ = 0, where τ is the unobservable threshold separating unaffected and affected dogs.

For a specific disease, the unobservable liability was assumed to have a multivariate normal distribution. The correlation in liability for a given ophthalmologic disease trait of 2 dogs (i and j, respectively) was modeled by use of the following equation:

View Expanded

Expansion of this single trait model to a model of 2 potentially correlated traits can be accommodated by defining the r_g and r_e between 2 traits (ie, lens luxation and cataracts). Then the correlation in liability between lens luxation and cataracts of 2 dogs (i and j, respectively) was modeled by use of the following equation:

View Expanded

Complex segregation analysis—The possibility that lens luxation or cataracts in Jack Russell Terriers could be influenced by the action of a segregating locus of large effect was also examined. Complex segregation analysis²⁴ is intended to integrate Mendelian transmission genetics at a single locus, with the patterns of covariance expected in polygenic inheritance. Criteria to establish the necessary evidence for the acceptance of a single locus model while reducing the risk of false-positive discovery of a major locus are provided elswhere.²⁵ Evaluation of the models necessary for complex segregation analysis in each of these binary disease traits was conducted with a Bayesian software package^b by analyzing 1 trait at a time. The Bayesian software was an extension of a package of subroutines for genetic analyses with Gibbs sampling²⁶ that was rewritten to accommodate complex segregation analysis in binary traits for pedigrees that include inbreeding. In contrast to the other computer program^a used in our study, the Bayesian software package could not accommodate multiple-trait complex segregation analyses. Accordingly, each analysis was conducted separately for each ophthalmologic disease.

DNA sequencing—To determine whether the mutation that has been identified as being associated with primary cataracts in 2 terrier breeds was also correlated with cataracts in Jack Russell Terriers,⁸ we sequenced exons 8 and 9 of HSF4 from the genomic DNA extracted from the buccal swab specimens obtained from 2 Jack Russell Terriers that expressed cataracts by 30 months of age (affected), 1 Jack Russell Terrier known to produce offspring that developed early onset cataracts yet did not develop cataracts itself (as determined by results of CERF examination) when > 7 years old (presumed carrier), and 2 Jack Russell Terriers that had not produced any offspring that developed cataracts and were considered clinically normal during CERF examinations conducted when both dogs were ≥ 7 years old (unaffected) for use in the DNA analyses. Exons 8 and 9 were amplified by use of primers and conditions described elsewhere.⁸ The amplified product was purified by use of a gel-purification kit^c and sequenced at an automated DNA sequencing facility.^d

Results

Estimates of heritability, r_g, r_e, and sex differences in liability to disease for each of the traits investigated were calculated. Lens luxation has a considerably (P = 0.002) large genetic component, a conclusion that was equally justifiable for cataracts, on the basis of heritability estimates of 0.99 and 0.73 for lens luxation and cataracts, respectively. Moreover, the impact of sex (difference between males and females) on expression of lens luxation (estimate [SE]; 0.07 [0.19]) and cataracts (0.18 [0.20]) was considered inconsequential given the magnitude of the SE (ie, estimates for the impact of sex were well within 1 SE of the value of 0). Thus, there was no differential contribution to expression of lens luxation or cataracts (P = 0.20) on the basis of the sex of a dog. There was a significant r_g (r_g = 0.56; P = 0.01) between lens luxation and cataracts. However, the r_e was not significantly different from 0 because the SE exceeded the point estimate and encompassed 0.

Several analyses were conducted for the complex segregation analysis of lens luxation. When only the analyses with fixed Mendelian transmission probabilities were considered, the data provided evidence for a segregating major locus (Table 1). Yet those results would have been misleading without consideration of the non-Mendelian transmission analysis. Thus, although estimates of the additive and dominance contributions as well as the total genetic variance contributed by the putative major locus differed significantly from 0 in the Mendelian analyses, when the probability of transmitting the putative disease allele was estimated from the data (rather than fixed at the expected Mendelian values of 1.0, 0.5, and 0), the pattern for cataracts did not readily conform to expectations of a major locus. In addressing criteria established in another study,²⁵ our estimates of the probabilities for allele transmission differed significantly from the expected Mendelian values of 1.0, 0.5, and 0 for transmission of the A allele from AA, AB, and BB genotypes, respectively. Specifically, the highest-density region for each of the transmission probabilities revealed considerable overlap among the 3 putative genotypes, which indicated that the transmission of this allele did not behave by the established rules put forth by Mendel. Therefore, results for the 2 analyses for transmission probabilities together cannot provide convincing evidence for the segregation of a locus with a substantially large impact on lens luxation.

Table 1—

Values of model parameters for lens luxation in Jack Russell Terriers in a Bayesian mixed-inheritance model with a completely recessive major locus, with and without Mendelian transmission of the putative major allele.

View Table

A similar conclusion for cataracts was reached by use of the same criteria. In analyses with fixed Mendelian transmission probabilities, the major locus variance overlapped with 0, which indicated that there was no evidence for a major locus of large effect segregating with an influence on cataracts in Jack Russell Terriers (Table 2). The non-Mendelian analyses provided further support to this conclusion in that the estimation of the transmission probability for the heterozygous genotype did not overlap with the expected value of 0.5.

Table 2—

Values of model parameters for cataracts in Jack Russell Terriers in a Bayesian mixed-inheritance model with a completely recessive major locus, with and without Mendelian transmission of the putative major allele.

View Table

Although the data did not support a segregating major locus, they did reveal a significant genetic contribution to cataracts. An insertion in exon 9 of HSF4 has been associated with presumed autosomal recessive cataracts in 2 terrier breeds. To determine whether that same mutation was directly correlated with cataracts in Jack Russell Terriers, exon 9 was sequenced in affected dogs, a dog that had produced offspring that developed cataracts yet that did not develop cataracts itself, and unaffected dogs. The sequence data for all 3 classes of dogs were identical, which indicated that the putative mutation responsible for cataracts in Boston Terriers and Staffordshire Bull Terriers was not detected in Jack Russell Terriers (Table 3). Furthermore, no deletions in exon 9 were detected in any of the 3 categories of Jack Russell Terriers, thus ruling out the mutation identified as associated with cataracts in Australian Shepherds.⁸ Two conservative DNA substitutions immediately outside the mutated region were detected in all Jack Russell Terriers that were not detected for reference Boxers,²⁷ Boston Terriers, Staffordshire Bull Terriers, or Australian Shepherds.

Table 3—

Genomic sequence of HSF4 exon 9 in a region (bp 3,627 to 3,656)²⁷ associated with cataracts in Staffordshire Bull Terriers (SBT), Boston Terriers (BT), and Australian Shepherds (AS) for 3 classes of Jack Russell Terriers (JRT; affected with cataracts, presumed carrier,* and unaffected) and for Boxers, which are a breed not predisposed to development of cataracts.

View Table

Discussion

Similar to results reported for other breeds, cataracts and lens luxation in Jack Russell Terriers had a substantial genetic component and were not influenced by sex. The exceedingly high heritability estimate for lens luxation must be viewed with caution. Values of heritability in excess of 0.90 are certainly not routinely detected. The possibility exists that the combination of a discrete trait and recorded patterns of inbreeding may have led to an estimate near the theoretic boundary of 1.0.

In the study of Jack Russell Terriers reported here, lens luxation and cataracts were positively, genetically correlated with each other. This implies that selection criteria applied to reduce the incidence of condition should also benefit reduction of the incidence for the other condition. In other studies,^28-30 investigators have found other concurrent ocular disorders, such as microphthalmia, in association with presumed inherited cataracts. Those findings of genetic association among various ocular disorders, coupled with the data reported here, suggest that real progress can be made toward improvements in ocular health.

Although the major locus model for lens luxation and cataracts must be rejected because the estimates of allele transmission failed to conform to Mendelian patterns of inheritance as evidenced by the large overlapping highest-density region for the 3 putative genotypes, which thereby indicated that criteria for a major locus have not been met,²⁵ the criteria also have not been violated. Thus, although the major locus model must be rejected with the data reported here, additional data may modify that conclusion if the additional data points reduce variation of the parameter estimates.

An insertion that introduces an early stop codon in the HSF4 gene has been implicated in the development of cataracts in Boston Terriers and Staffordshire Bull Terriers. A deletion in the same region of the HSF4 exon 9 also introduces an early stop codon and is associated with cataracts in Australian Shepherds, with affected dogs being heterozygous or homozygous for the mutation, which is suggestive of an autosomal dominant mode of inheritance.⁸ Comparing the sequence of exon 9 of the HSF4 gene for cataractous Jack Russell Terriers with the DNA sequence in Staffordshire Bull Terriers and Boston Terriers was reasonable, and lack of a causal mutation in the Jack Russell Terriers was not surprising. The lack of statistical support for the existence of a major locus affecting expression of cataracts in Jack Russell Terriers made the existence of a single causal mutation improbable. The ancestral history of Staffordshire Bull Terriers is similar to that of Boston Terriers, with both considered to have been derived from Bull Terriers during the late 1800s.^31,32 In contrast, Jack Russell Terriers are considered to have been derived from Fox Terriers.³³ Furthermore, cluster analysis reveals that terriers, as a group, are quite diverse in terms of their genetic ancestry,³⁴ which implies that common causal mutations may be rare among terrier breeds. The complex mode of inheritance possibly aligns cataracts of Jack Russell Terriers with 2 other breeds that have inherited cataracts (ie, German Shepherd Dogs and Cocker Spaniels), both of which lack mutations in this gene.⁸

In the study reported here, we ascertained that cataracts and lens luxation were both characterized as highly heritable in Jack Russell Terriers. Moreover, we concluded that the pattern of inheritance for these 2 ocular disorders is polygenic, with no convincing evidence that a major locus of large effect contributes to the expression of either disorder. The 2 ocular disorders were genetically correlated, which indicated that the genetic liability for cataracts is associated with the risk for lens luxation. Considered together, these findings can be used as encouragement for breeders because the high heritabilities suggest that selection against these diseases will be quickly rewarded with improvements in ocular health. Moreover, there was a positive r_g. Had a strong negative correlation been uncovered, breeders would have been forced to confront the unsettling genetic conclusion that selection against 1 ocular disease would lead to an increase in the liability of the other. Instead, the recommendation is that breeders must remain vigilant for both disorders when choosing the sires and dams of future generations of Jack Russell Terriers.