Pharmacogenetics is the study of how an individual's genetic composition determines his or her response to drugs. In humans, variations in toxic effects and efficacy of drugs among individuals are mainly determined by genetic polymorphisms in drug-metabolizing enzymes, drug receptors, or drug transporters.1 In dogs, polymorphisms in drug-metabolizing enzymes, drug receptors, and drug transporters have been reported,2–4 and these polymorphisms frequently segregate along breed lines. For example, Greyhounds have low activity of a particular cytochrome P450 enzyme, CYP2B11; following administration of propofol in these dogs, plasma concentrations of the drug are sustained and recovery from anesthesia is delayed, compared with findings in mixed-breed dogs.2,3 A polymorphism in the canine dopamine receptor D4 gene has been detected, and its allelic frequency varies among breeds.4 Perhaps the most striking example of pharmacogenetics in dogs involves the drug transporter P-gp. P-glycoprotein, the product of the ABCB1 (formerly MDR1) gene, acts as an important barrier to the distribution of substrate drugs to selected tissues and restricts access of xenobiotics through the blood-brain barrier, the blood-testes barrier, and the placenta; it also has important excretory functions in enterocytes, biliary canalicular cells, and renal tubular epithelial cells.
The ABCB1-1Δ polymorphism in dogs consists of a 4–base pair deletion mutation. This deletion results in a shift of the reading frame that generates several premature stop codons.5 Because protein synthesis is terminated before as much as 10% of the protein product is synthesized, dogs with 2 mutant alleles have a P-gp null phenotype. Such dogs are approximately 100 times as susceptible to ivermectin-induced neurologic toxicosis as ABCB1 wild-type dogs. These dogs also appear to be highly susceptible to the neurologic adverse effects of loperamide6 and other avermectins, including milbemycin, selamectin,7 and moxidectin.8 Heterozygous dogs appear to have an intermediate phenotype with respect to responses to avermectins. In a recent case report9 of a dog with the ABCB1-1Δ mutation, hematologic and gastrointestinal toxic effects developed following administration of recommended and reduced doses of vincristine; however, the dog tolerated full doses of cyclophosphamide.
The ABCB1-1Δ mutation primarily affects herding breeds of dogs. Results of earlier studies10–13 involving small populations of dogs have indicated that approximately 75% of Collies in the United States, France, and Australia have at least 1 mutant allele. Other breeds that have been reported to be affected with the ABCB1-1Δ mutation include Old English Sheepdogs, Australian Shepherds, Shetland Sheepdogs, English Shepherds, Border Collies, Silken Windhounds, McNabs, and Longhaired Whippets.12 The purpose of the study reported here was to evaluate the breed distribution of the ABCB1-1Δ polymorphism in a large number of dogs in North America, including dogs of several herding breeds in which this polymorphism has been detected and other breeds in which this polymorphism has not yet been identified.
Materials and Methods
DNA samples—Swabs containing buccal cells were obtained from dogs by owners or veterinarians and submitted to the Veterinary Clinical Pharmacology Laboratory at Washington State University for ABCB1 genotyping. The reasons for submission of samples for ABCB1 genotyping varied widely, but included a potential need for treatment of the dog with a P-gp substrate drug (most commonly ivermectin, vincristine, or doxorubicin), screening for breeding purposes, evaluation as part of a larger breed-specific screening program, or investigation of a prior adverse reaction to a P-gp substrate drug; in some instances, owners had been encouraged by a breeder or rescue society personnel to have ABCB1 genotyping performed on their pet. Owners or veterinarians were asked to report the dog's breed, but no effort was made to confirm the reported breed. More than 90% of dogs in the study were living in the United States or Canada. ABCB1 genotyping was performed by the Veterinary Clinical Pharmacology Laboratorya according to published methods.14 Dogs with 2 ABCB1 wild-type alleles were designated as ABCB1 wt/wt, dogs with 1 ABCB1 wild-type allele and 1 ABCB1-1Δ allele were designated as ABCB1 mut/wt, and dogs with 2 ABCB1-1Δ alleles were designated as ABCB1 mut/mut. Genotype and allele frequencies were tallied for individual breeds or breed groupings (mixed breed [parentage either unknown or known not to include a herding breed], herding breed mix [at least 1 parent that was known to be a herding breed], and unknown).
Results
Results of ABCB1 genotyping performed on DNA samples from 5,368 dogs were included in the study (Table 1). Approximately half of the submissions were samples collected from Australian Shepherds (1,421 [26.5%]) and Collies (1,424 [26.5%]). Collectively, samples from Shetland Sheepdogs (448 [8.3%]), herding breed mixes (312 [5.8%]), Border Collies (306 [5.7%]), and German Shepherd Dogs (166 [3.1%]) accounted for approximately 23% of submissions. Among the 5,368 dog breeds represented by ≥ 10 dogs, dogs having an ABCB1-1Δ wt/wt genotype included Bearded Collie (n = 189), Skye Terrier (79), Welsh Corgi (32), Australian Cattle Dog (23), Golden Retriever (22), Labrador Retriever (18), Bulldog (16), Boxer (15), Shih Tzu (14), American Pitbull Terrier (10), Akabash (10), and Doberman Pinscher (10). Two hundred twenty-one samples from 124 other breeds (< 10 dogs in each breed category) were analyzed, with each dog having the ABCB1-1Δ wt/wt genotype. Samples from nonpurebred dogs were separated into 2 categories: herding breed mixes (dogs for which at least 1 parent was known to be a herding breed) and mixed-breed dogs (dogs for which the parentage was either unknown or known not to include a herding breed).
Frequency of ABCB1 genotypes (No. of dogs [%]) determined via analysis of DNA in buccal swab samples collected from 5,368 dogs in North America between May 1, 2004, and September 30, 2007.
| Breed | No. of dogs | Genotype | ||
|---|---|---|---|---|
| ABCB1 wt/wt | ABCB1 mut/wt | ABCB1 mut/mut | ||
| Australian Shepherd | 1,421 | 754 (53) | 525 (37) | 142 (10) |
| Border Collie | 306 | 301 (98) | 4 (1) | 1 (0.003) |
| Collie | 1,424 | 322 (23) | 598 (42) | 504 (35) |
| English Shepherd | 28 | 28 (100) | 0 (0) | 0 (0) |
| German Shepherd | 166 | 149 (90) | 14 (8) | 3 (2) |
| Herding—breed mix* | 312 | 276 (89) | 32 (10) | 4 (1) |
| Longhaired Whippet | 24 | 10 (42) | 14 (58) | 0 (0) |
| Miniature Australian Shepherd | 285 | 180 (63) | 96 (34) | 9 (3) |
| McNab | 1 | 1 (100) | 0 (0) | 0 (0) |
| Mixed breedt | 238 | 212 (89) | 19 (8) | 7 (3) |
| Old English Sheepdog | 40 | 39 (97.5) | 1 (2.5) | 0 (0) |
| Shetland Sheepdog | 448 | 395 (88) | 47 (11) | 6 (1) |
| Silken Windhound | 16 | 11 (69) | 5 (31) | 0 (0) |
| Other purebreeds | 659 | 659 (100) | 0 (0) | 0 (0) |
ABCB1 genotyping was performed according to previously published methods.13 ABCB1 wt/wt = Dogs with 2 ABCB1 wild—type alleles. ABCB1 mu1/wt= Dogs with 1 ABCB1 wild—type allele and 1 ABCB1—1Δ allele. ABCB1 mul/mut = Dogs with 2 ABCB1—1Δ alleles.
Herding breed mix was a dog in which at least 1 parent was known to be a herding breed. †Dogs were not identified specifically as a herding breed mix or a nonherding breed mix; the proportion of the former in this group is not known.
The ABCB1-1Δ allele was detected in 10 breeds including the German Shepherd Dog, a breed in which the ABCB1-1Δ allele had not been previously identified, to our knowledge (Table 1). More than half of the Collies and Longhaired Whippets evaluated had the mutant ABCB1 allele (ie, either ABCB1 mut/mut or ABCB1 mut/wt genotypes). Other breeds of dogs in which the ABCB1-1Δ allele was detected in a high proportion (> 25%) of dogs included Australian Shepherd (667/1,421), Miniature Australian Shepherd (105/285), and Silken Windhound (5/16). Analysis of the study data did not reveal the presence of the ABCB1-1Δ allele in dogs of 2 breeds (McNabs and English Shepherds) in which the ABCB1-1Δ allele had been previously identified.12 However, the numbers of dogs of these 2 breeds that were evaluated were quite low (1 McNab and 28 English Shepherds). A relatively large number of Bearded Collies (n = 189 dogs), a herding breed that originated in Great Britain as a descendent of working sheepdogs, were evaluated, and the ABCB1-1Δ allele was not identified in any of those dogs.
The ABCB1-1Δ mutation was identified in a large proportion of herding breed mixes (36/312 [11.5%]) as well as mixed-breed dogs (26/238 [10.9%]) in this study. Samples from several mixed-breed dogs were analyzed because the dogs had reacted adversely to ivermectin during treatment for demodectic mange. Adverse effects generally developed after a single dose of ivermectin (300 to 600 μg/kg administered PO or via SC injection). Genotyping of these dogs prior to administration of ivermectin had not been performed by the attending veterinarians because the dogs did not appear to be herding breeds (Figure 1).
Photographs of 2 mixed-breed dogs for which DNA analysis of buccal swab samples yielded a positive result for the ABCB1-1Δ mutation in a study of samples collected from 5,368 dogs in North America. Subjectively, these dogs do not appear to have herding breed ancestry. A—Dog that developed signs of neurologic toxicosis after treatment with ivermectin (extralabel dose) for mange. The genotype of this dog is ABCB1 mut/mut (dogs with 2 ABCB1-1Δ alleles). B—Dog that sired a puppy that developed signs of neurologic toxicosis secondary to treatment with ivermectin for mange. The genotype of this dog is ABCB1 mut/wt (dogs with 1 ABCB1 wild-type allele and 1 ABCB1-1Δ allele).
Citation: Journal of the American Veterinary Medical Association 233, 6; 10.2460/javma.233.6.921
Discussion
In the present study, it was not surprising that a high percentage of samples submitted for ABCB1 genotyping were from dogs of herding breeds such as Collies and Australian Shepherds because the ABCB1-1Δ mutation appears to occur primarily in these breeds. Furthermore, the allelic frequency of ABCB1-1Δ is high in those breeds, compared with the allelic frequencies in most other breeds.12 Allelic frequencies of ABCB1-1Δ have been reported for Collies, Old English Sheepdogs, Shetland Sheepdogs, and Miniature Australian Shepherds.12 Despite the potential for sampling bias in our study, the distribution of ABCB1-1Δ genotypes was quite similar to that previously reported for these breeds.12 In contrast, there was a higher proportion of Australian Shepherds with the ABCB1 mut/mut genotype (142/1,421 [10%]) in our study, compared with that reported previously (3/178 [1.7%]).12 Our study population was substantially larger than that in the previous study, suggesting that it may be a better representative sample of Australian Shepherds in North America. Alternatively, sampling bias such as submission of samples from dogs (or their relatives) that had developed signs of toxic effects after receiving a P-gp substrate drug or dogs that were related to individuals in which the ABCB1-1Δ allele had been identified may have skewed our results. From a clinical perspective, results from both studies support the advisability of assessing the ABCB1 genotype of an Australian Shepherd prior to administration of P-gp substrate drugs.
The ABCB1-1Δ mutation was previously identified in 2 sighthound breeds: the Silken Windhound and the Longhaired Whippet.12 It has been speculated that the introduction of the ABCB1-1Δ mutation in Long-haired Whippets likely resulted from interbreeding with Shetland Sheepdogs to improve the long-haired phenotype.12 The Silken Windhound was developed in the 1980s through crossbreeding of several sighthound breeds, including the Longhaired Whippet. In the previous study,12 the percentages of ABCB1 mut/mut dogs of the Silken Windhound and Longhaired Whippet breeds were higher than findings in our study. The numbers of dogs of those breeds evaluated in the previous study (89 Longhaired Whippets and 84 Silken Windhounds) were greater than the numbers in the study reported here (29 Longhaired Whippets and 15 Silken Wind-hounds); the former may therefore be a more accurate representation of the breed populations.
To our knowledge, German Shepherd Dogs had not previously been identified as a breed that carries the ABCB1-1Δ mutation. This is an interesting finding because, on the basis of marker alleles, the ABCB1-1Δ mutation was derived from 1 dog12 and is identical by descent; it did not occur via separate mutations. The dog with the original ABCB1-1Δ mutation was proposed to be a working sheepdog that lived in Great Britain in the mid 1800s, prior to the formal establishment of the Collie and Old English Sheepdog as breeds of dog. It is likely that this dog was a breeding dog and therefore sired many of the dogs that eventually formed the Collie breed and, to a lesser extent, the Old English Sheepdog breed. With the exception of the German Shepherd Dog, the other herding breeds affected by the ABCB1-1Δ mutation can be traced back to working sheepdogs in Great Britain. In the present study, only 3 German Shepherd Dogs were homozygous for the ABCB1-1Δ mutation; 2 of these 3 dogs were white German Shepherds, and the other was a white-factored dog (ie, it had a white-colored parent or grandparent). Several of the ABCB1 mut/wt German Shepherd dogs in our study were also white-factored dogs. Whether the ABCB1-1Δ mutation accompanies an allele for white hair is not known.
McNabs12 and Border Collies15 have been identified as breeds that carry the ABCB1-1Δ mutation. The allelic frequency for McNabs was 17.1% (n = 35 dogs).13 In our study, only 1 McNab was evaluated and it did not have the ABCB1-1Δ mutation. The distribution of the ABCB1-1Δ genotype in Border Collies in our study and in another study of Border Collies living in Germany was similar.15 In both samples, a limited proportion of Border Collies (< 2%) had the ABCB1-1Δ allele.
Several purebred dogs of nonherding origin, such as Labrador Retrievers, Skye Terriers, and others, were evaluated in the present study because of reported drug reactions. Although these dogs did not have the ABCB1-1Δ polymorphism, it is possible that another mutation that affects P-gp function was present. However, the authors are not aware of any other polymorphisms in the canine ABCB1 gene, and the potential presence of additional ABCB1 gene polymorphisms was not investigated in the study of this report.
Results of our study identified a relatively high percentage of mixed-breed dogs with the ABCB1 mut/mut (7/238 [3%]) or ABCB1 mut/wt (19/238 [8%]) genotype. These percentages are likely to be lower in the general population because of sampling bias in our data collection. Specifically, samples from several of the mixed-breed dogs were submitted for analysis because the dog had developed signs of toxicosis after receiving a P-gp substrate drug. What was surprising to the authors, however, is that many of these dogs had no physical resemblance to herding breed dogs that are known to have the ABCB1-1Δ allele. For practicing veterinarians, this means that one cannot simply assess the appearance of a dog to decide whether it can be treated safely with P-gp substrate drugs, such as macrocyclic lactones (at doses used for treating mange), vincristine, doxorubicin, or loperamide.
The ABCB1-1Δ polymorphism is present in a high percentage of herding breed dogs that are examined by veterinarians. The ABCB1-1Δ polymorphism is associated with increased frequency of adverse drug reaction development5,6,9,14 and with suppression of the hypothalamic-pituitary-adrenal axis.16 Veterinarians should be familiar with dog breeds that have the ABCB1-1Δ polymorphism to make appropriate pharmacologic choices for these patients. Prior to administration of vincristine, doxorubicin, loperamide, or doses of ivermectin or other macrocyclic lactones for extralabel treatment of mange, it is warranted to determine the ABCB1-1Δ genotype of dogs of the 10 herding breeds and 2 sighthound breeds as well as mixed-breed dogs to avoid severe adverse drug reaction events. Conversely, assessment of the ABCB1 genotype in purebred dogs of other breeds may not be necessary.
ABBREVIATION
| P-gp | P-glycoprotein |
Veterinary Clinical Pharmacology Laboratory, Washington State University, Pullman, Wash. Available at: www.vetmed.wsu.edu/vcpl. Accessed Jun 2, 2008.
References
- 1.↑
Fujita K, Sasaki Y. Pharmacogenomics in drug-metabolizing enzymes catalyzing anticancer drugs for personalized cancer chemotherapy. Curr Drug Metab 2007;8:554–562.
- 2.
Court MH, Hay-Kraus BL & Hill DW, et al. Propofol hydroxylation by dog liver microsomes: assay development and dog breed differences. Drug Metab Dispos 1999;27:1293–1299.
- 3.
Hay Kraus BL, Greenblatt DJ & Venkatakrishnan K, et al. Evidence for propofol hydroxylation by cytochrome P4502B11 in canine liver microsomes: breed and gender differences. Xenobiotica 2000;30:575–588.
- 4.↑
Niimi Y, Inoue-Murayama M & Kato K, et al. Breed differences in allele frequency of the dopamine receptor D4 gene in dogs. J Hered 2001;92:433–436.
- 5.↑
Mealey KL, Bentjen SA & Gay JM, et al. Ivermectin sensitivity in collies is associated with a deletion mutation of the mdr1 gene. Pharmacogenetics 2001;11:727–733.
- 6.↑
Sartor LL, Bentjen SA & Trepanier L, et al. Loperamide toxicity in a collie with the MDR1 mutation associated with ivermectin sensitivity. J Vet Intern Med 2004;18:117–118.
- 8.↑
Tranquilli WJ, Paul AJ, Todd KS. Assessment of toxicosis induced by high-dose administration of milbemycin oxime in Collies. Am J Vet Res 1991;52:1170–1172.
- 9.↑
Mealey KL, Northrup NC, Bentjen SA. Increased toxicity of P-glycoprotein-substrate chemotherapeutic agents in a dog with the MDR1 deletion mutation associated with ivermectin sensitivity. J Am Vet Med Assoc 2003;223:1434,1453–5.
- 10.
Mealey KL, Munyard KA, Bentjen SA. Frequency of the mutant MDR1 allele associated with multidrug sensitivity in a sample of herding breed dogs living in Australia. Vet Parasitol 2005;131:193–196.
- 11.
Mealey KL, Bentjen SA, Waiting DK. Frequency of the mutant MDR1 allele associated with ivermectin sensitivity in a sample population of Collies from the northwestern United States. Am J Vet Res 2002;63:479–481.
- 12.↑
Neff MW, Robertson KR & Wong AK, et al. Breed distribution and history of canine mdr1-1Δ, a pharmacogenetic mutation that marks the emergence of breeds from the collie lineage. Proc Natl Acad Sci U S A 2004;101:11725–11730.
- 13.↑
Hugnet C, Bentjen SA, Mealey KL. Frequency of the mutant MDR1 allele associated with multidrug sensitivity in a sample of collies from France. J Vet Pharmacol Ther 2004;27:227–229.
- 14.↑
Henik RA, Kellum HB & Bentjen SA, et al. Digoxin and mexiletine sensitivity in a Collie with the MDR1 mutation. J Vet Intern Med 2006;20:415–417.
- 15.↑
Geyer J, Döring B & Godoy JR, et al. Frequency of the nt230 (del4) MDR1 mutation in Collies and related dog breeds in Germany. J Vet Pharmacol Ther 2005;28:545–551.
- 16.↑
Mealey KL, Gay JM & Martin LG, et al. Comparison of the hypothalamic–pituitary–adrenal axis in MDR1–1Ä and MDR1 wildtype dogs. J Vet Emerg Crit Care 2007;17:61–66.
