Niels Pedersen, 1 Hongwei Liu, Lee Millon, Kimberly Greer

J Vet Diagn Invest 23:68 76 (2011) Dog leukocyte antigen class II associated genetic risk testing for immune disorders of dogs: simplified approaches using Pug dog necrotizing meningoencephalitis as a model Niels Pedersen, 1 Hongwei Liu, Lee Millon, Kimberly Greer Abstract. A significantly increased risk for a number of autoimmune and infectious diseases in purebred and mixed-breed dogs has been associated with certain alleles or allele combinations of the dog leukocyte antigen (DLA) class II complex containing the DRB1, DQA1, and DQB1 genes. The exact level of risk depends on the specific disease, the alleles in question, and whether alleles exist in a homozygous or heterozygous state. The gold standard for identifying high-risk alleles and their zygosity has involved direct sequencing of the exon 2 regions of each of the 3 genes. However, sequencing and identification of specific alleles at each of the 3 loci are relatively expensive and sequencing techniques are not ideal for additional parentage or identity determination. However, it is often possible to get the same information from sequencing only 1 gene given the small number of possible alleles at each locus in purebred dogs, extensive homozygosity, and tendency for disease-causing alleles at each of the 3 loci to be strongly linked to each other into haplotypes. Therefore, genetic testing in purebred dogs with immune diseases can be often simplified by sequencing alleles at 1 rather than 3 loci. Further simplification of genetic tests for canine immune diseases can be achieved by the use of alternative genetic markers in the DLA class II region that are also strongly linked with the disease genotype. These markers consist of either simple tandem repeats or single nucleotide polymorphisms that are also in strong linkage with specific DLA class II genotypes and/or haplotypes. The current study uses necrotizing meningoencephalitis of Pug dogs as a paradigm to assess simple alternative genetic tests for disease risk. It was possible to attain identical necrotizing meningoencephalitis risk assessments to 3-locus DLA class II sequencing by sequencing only the DQB1 gene, using 3 DLA class II linked simple tandem repeat markers, or with a small single nucleotide polymorphism array designed to identify breed-specific DQB1 alleles. Key words: Dog leukocyte antigen class II association; genetic risk testing; immune disorders; linked simple tandem repeat markers; linked single nucleotide polymorphism markers; necrotizing meningoencephalitis; Pug dog encephalitis. Introduction Specific dog leukocyte antigen (DLA) class II genotypes and haplotypes have been associated with an increasing number of autoimmune disorders of purebred dogs. A Vogt Koyanagi Harada-like syndrome of Akita dogs has a strong association with certain alleles and allele combinations of DLA class II genes, 1 as does autoimmune hemolytic anemia, 20 immune arthritis, 28 hypothyroidism in several breeds, 25,26,38 type I diabetes mellitus in Samoyed From the Center for Companion Animal Health and the Koret Center for Veterinary Genetics of the Veterinary Genetics Laboratory, School of Veterinary Medicine, University of California, Davis, CA (Pedersen, Liu, Millon), and the School of Natural Sciences and Mathematics, Indiana University East, Richmond, IN (Greer). 1 Corresponding Author: Niels Pedersen, Veterinary Genetics Laboratory, 213 CCAH Building, University of California Davis, One Shields Avenue, Davis, CA 95616. ncpedersen@ucdavis.edu dogs and Cairn and Tibetan Terriers, 23 anal furunculosis of German Shepherd Dogs, 5 and systemic lupus erythematosus related complex of Nova Scotia Tolling Retrievers. 37 Necrotizing meningoencephalitis (NME) of Pug dogs 10 and Addison disease in Nova Scotia Duck Tolling Retrievers 12 have been recently added to this growing list. Dog leukocyte antigen class II associations have also been described for visceral leishmaniasis 30 in domestic dogs in Spain and for generalized demodicosis in Boxer, Argentinean Mastiff, and mixed-breed dogs in Argentina. 16 It appears, therefore, that immune disorders of dogs follow the pattern observed in humans (i.e., that most will occur in context of certain alleles within genes of the major histocompatibility complex [MHC]). 34 Unlike human beings, the incidence of immune disorders in dogs has increased with purebreeding, 29 which often involves inbreeding on both maternal and paternal sides. 3,36 Reports on the risk association of certain DLA class II types and autoimmune disorders and chronic 68

Genetic tests for canine immune diseases 69 Figure 1. Nucleotide alignment of exon 2 of the 6 dog leukocyte antigen DQB1 alleles detected in Pug dogs. The nucleotides in bold are locations of single-nucleotide polymorphisms identified by Sequenom MassArray Design software e that can be used to identify each of the 6 alleles. infectious diseases in dogs have frequently mentioned the potential value of direct sequencing for allele assignment of the 3 DLA class II genes: DRB1, DQA1, and DQB1. However, direct sequencing of alleles at 3 different loci is time consuming, is expensive, and cannot easily be combined with genetic tests for parentage and/or identification. Therefore, the objective of the current study was to evaluate simpler and more versatile alternative test procedures to 3-locus DLA class II gene sequencing that might be applicable to immune disorders in dogs that are known to have DLA class II gene associated disease risk. The tests should be applicable to buccal swab DNA and, if possible, easily combined with parentage and/or identification markers. The disease chosen for the current study was NME of Pugs, which, like a spectrum of immune disorders of other purebreds, takes advantage of limited DLA diversity, strong DLA class II allele linkages, and a high degree of genetic homozygosity. Although many immune diseases of dogs in a number of purebreds have similar DLA class II gene associated risks, NME of Pugs was chosen in the current study to test the feasibility of simplifying genetic risk assessment. Necrotizing meningoencephalitis in Pugs, formerly known as Pug dog encephalitis, is a progressive and fatal disease affecting 1 2% of Pugs, usually before 7 years of age. 9,27 There has been no laboratory evidence for an infectious cause, 27,32,35 and the etiology is now presumed to be immunologic. 9,35 Pug dog NME has been likened to fulminate atypical forms of human multiple sclerosis in its form and strong association to a specific human leukocyte antigen class II genotype. 10 A whole genome scan recently associated risk for the disease solely with a specific region on canine chromosome 12 coding for the genes of the DLA complex. 10 Fine mapping studies revealed a region of high homozygosity that included the DLA class II DRB1, DQA1, and DQB1 genes. 10 Sequencing of these 3 genes confirmed both a high degree of homozygosity and a strong linkage between certain alleles at each loci, yielding a high-risk haplotype DLA-DRB1*010011/ DQA1*00201/DQB1*01501. Risk-protective and neutral allele combinations were also identified. These strong linkages also allow for several simple alternative approaches to direct sequences of alleles at all 3 DLA class II gene loci, including sequencing a single DLA class II locus (DQB1 in Pugs), the use of a threesome of DLA class II linked simple tandem repeat (STR) markers, and a 4 single-nucleotide polymorphism (SNP) based array capable of identi-

70 Pedersen et al. fying breed-specific DLA class II alleles (DQB1 in Pugs). Materials and methods Case material Purified DNA or DNA-containing tissue specimens were obtained from NME-affected and healthy Pugs. Some samples, in the form of liver or blood, were from Pugs previously reported. 10 Additional samples in the form of blood were from the collections of Dr. Karen Vernau, School of Veterinary Medicine, University of California, Davis. A final group of samples in the form of buccal swabs was also obtained from healthy Pugs sampled as part of a large collection of purebred dogs from the Veterinary Genetics Laboratory, University of California, Davis. The exact numbers of animals tested for each procedure are listed in the results and appropriate tables; exact numbers of samples and origins varied for each test procedure, depending on sample availability at the time. DNA was isolated from whole blood and liver samples using a standard proteinase K digestion procedure. 31 DNA isolation from the buccal swabs was performed as previously described. 1,15 DLA-DRB1/DQA1/DQB1 allele sequencing for NME risk Exon 2 of DLA class II genes DRB1, DQA1, and DQB1 was sequenced using internationally recognized intronic primers according to a standardized protocol. 17,22,24 The 270 base pair polymorphic sequence in exon 2 of the DLA- DRB1 was amplified using primers DRBF (59-GAT CCCCCCGTCCCCACAG-39) and DRBR (59-TGTGTCA CACACCTCAGCACCA-39). The 307 base pair polymorphic sequence in exon 2 of DLA-DQA1 was amplified using primers DQAF (59-TAAGGTTCTTTTCTCCCTCT-39)and DQAR (59-GGACAGATTCAGTGAAGAGA-39). The 267 base pair polymorphic sequence in exon 2 of the DLA- DQB1 was amplified using primers DQBF (59-TCACT GGCCCGGCTGTCTCC-39) and DQBR (59-GGTGC GCTCACCTCGCCGCT-39). Polymerase chain reaction (PCR) was carried out in 20-ml reaction volumes containing 20 50 ng of DNA, 0.2 mm of deoxyribonucleotide triphosphate, 0.25 mm of forward and reverse primers in 13 PCR buffer, and 1 unit of DNA polymerase. a The mixture was then incubated for 5 min at 93uCfollowedby30 cycles of 45 sec at 93uC, 45 sec at 61uC fordrb1,57uc for DQA1, and 66uC for DQB1; for 1 min at 68uC; and then a final extension of 10 min at 68uC. The PCR products were electrophoresed on 2% agarose gel and purified with a gel purification kit a according to the manufacturer s instructions. Sequencing of exon 2 regions of the DLA-DRB, -DQA, and -DQB genes was performed in both directions for the NME-affected Pugs and in one direction for healthy control dogs. For one-direction sequencing, the reverse primers for DRB1 and DQB1 and forward primer for DQA1 were used. The purified PCR products were sequenced with a cycle sequencing kit b in a 15-ml reaction containing 1 ml of terminator mix, b 2.2 ml of reaction buffer (53), 35 ng of sequencing primer, and 3 ml of purified PCR product. The sequencing reaction was incubated at 93uC Table 1. Polymerase chain reaction (PCR) and extension primer sequences used for single-nucleotide polymorphism (SNP) testing e for specific Pug dog leukocyte antigen DQB1 alleles. SNP no. PCR primer 1 PCR primer 2 Extension primers AGAGGATTTCGTGTACCAGT ACGTTGGATGTCCCCCACGTC GCTGTCGAA SNP1 ACGTTGGATGTCTCCCCGC AGAGGATTTC TGGTGAAATAGCACTCG ACGTTGGATGTCTCCCCGCAG AGGATTTC SNP2 ACGTTGGATGTCCCCCACG TCGCTGTCGAA GGACGGAGCGGGTGCGG ACGTTGGATGTCCCCCACGTCGC TGTCGAA SNP3 ACGTTGGATGTCTCCCCG CAGAGGATTTC agcctcccggttatagatg ACGTTGGATGTCTCCCCGCAG AGGATTTC SNP4 ACGTTGGATGTCCCCCACGTC GCTGTCGAA TCTATAACCGGGAGGAGT ACGTTGGATGTCCCCCACGTCGC TGTCGAA SNP5 ACGTTGGATGTCTCCCCGCAGA GGATTTC gggaacggagctcgggcggcc ACGTTGGATGTGCGCTCACCTCG CCGCTG SNP6 ACGTTGGATGGGCGGTCACGGA GCTCGG

Genetic tests for canine immune diseases 71 Table 2. Dog leukocyte antigen class II alleles and their haplotype frequency in 51 necrotizing meningoencephalitis affected and 57 healthy control Pug dogs. Alleles Haplotype % DRB1 DQA1 DQB1 Affected (n 5 102) Control (n 5 114) P value* 010011 00201 01501 78.43 28.07,0.0001 Increased 01501 00601 02601 9.8 16.67 0.23 Neutral 01501 00601 02301 1.96 12.28 0.0081 Decreased 01502 00601 02301 3.92 29.82,0.001 Decreased 00601 005011 00701 5.88 5.26 1.0 Neutral 01501 00601 05701 0 7.02 0.012 Decreased 01702 00201 01304 0 0.88 1.0 Neutral * Proportion of haplotypes in affected dogs: (102/216) 5 47.22%. Binomial test used to assess probability. Risk for 2 min and amplified for 40 cycles at 93uC for 20 sec, 50uC for 20 sec, and 60uC for 4 min. Unincorporated dye terminators and deoxyribonucleotide triphosphates were removed by gel filtration, c and the amplified products were analyzed by capillary electrophoresis on a genetic analyzer. Sequence analysis software d was used for alignment of sequence data. Homozygous and heterozygous (mixed) sequences were resolved, and allele nomenclature was assigned according to an international standard. 18 All DLA-DRB1, -DQA1, and -DQB1 alleles identified among Pugs 10 were previously known from this and other breeds. Linked STR marker test for NME risk Three STRs (12-005A_CA, 12_005A_CT, 12_005B_CA) on Canis familiaris autosome (CFA12) were mined from the canine genome (nucleotide 5,018,175; 5,361,083; and 5,677,936). These 3 markers were closely linked to the DLA class II region and certain alleles strongly associated with risk for developing NME in Pugs. 10 The original primer sequences used for the present study were derived from a set of 3,000 STR markers 40 and listed in DOGSET 39 (http://www.vgl.ucdavis.edu/dogset). The primer sequences and fluorescent tags used in the present study were as follows: 5ACA-F (FAM) TATTGCACCCTGGTGTCT GC, 5ACA-R TTGCCCTGGGTGGTAAAATC; 5ACT-F (NED) GGAACCCCCTGTAAAATTTCTT, 5ACT-R CAGCCAAGACCTTAGGAGCAA; 5BCA-F CCCTGG ATATGTGGCAGTCA, 5BCA-R (NED) TGCCCTCTT CCACTTCACCT. The PCR-amplified products were analyzed using a capillary instrument. b Single-nucleotide polymorphism testing for DQB1 genotype Six SNPs based on the sequences of 6 known Pug DLA- DQB1 alleles were identified using assay design software e (Fig. 1). The PCR and extension primer sequences of the 6 SNPs are given in Table 1. The SNPs were assayed using an automated system. e Although all 6 SNPs were used to test each sample, SNPs 1, 2, 5, and 6; 1, 2, 4, and 6; 2, 3, 4, and 6; and 2, 3, 5, and 6 were equally useful in allele identification. Results from SNPS 1, 2, 5, and 6 were used primarily for this study. The additional SNPs were used to confirm the genotype and to allow for occasional allele dropout in the primary SNP panel. Results Three-locus DLA class II haplotypes and relationship to NME risk Exon 2 of DLA-DRB1, -DQA1, and -DQB1 genes were sequenced from a total of 51 NME-affected and 57 healthy Pugs. Results from 35 normal and 38 NME-affected Pugs from this group were previously reported, 10 and 13 new affected and 22 new normal animals were added for this study. Five DRB1 (010011, 01501, 01502, 00601, and 01702), 3 DQA1 (00201, 00601, and 005011), and 6 DQB1 (01501, 02301, 02601, 00701, 01304, and 05701) alleles were identified (Table 2). These findings were identical to those previously reported, 10 with the exception of DQB1*05701, which was not observed among the original group of Pugs. This allele differs from DQB1*02301 by 2 nucleotides, each of which causes a change in amino acid (Fig. 1). It has been observed previously in a small number of Pugs and Husky dogs and has now been recognized as a new allele (L. J. Kennedy, University of Manchester, personal communication, 2010). The DQB1*05701 allele was interchangeable with DQB1*02301 in the DRB1*01501/ DQA1*00601/DQB1*02301 haplotype. The 5 DRB1, 3 DQA1, and 6 DQB1 alleles formed 7 3-locus haplotypes, the frequency of which is given in Table 2. These 7 haplotypes conferred high, neutral, or decreased risk for the development of NME, depending on zygosity (Tables 2, 3). A significant risk for disease was apparent when the predominant 010011/00201/01501 haplotype was present in a homozygous (odds ratio [OR] 15.58, relative risk [RR] 5.9) but not heterozygous state (OR 0.7, RR 0.78; Table 3). The 01502/00601/02301 haplotype was highly protective in both homozygous (OR 0, RR 0) and heterozygous (OR 0.14, RR 0.2) forms. The 01501/00601/02301 haplotype was protective in heterozygotes (OR 0.13, RR 0.16). The 01501/ 00601/02601 haplotype was protective in homozygotes (OR 0, RR 0) but had no effect in heterozygotes (OR 0.83, RR 0.85). The 00601/005011/00701 haplo-

72 Pedersen et al. Table 3. Odds ratio (OR) and relative risk (RR) for affected and normal Pug dogs that were either homozygous (homo) or heterozygous (hetero) for the alleles across DRB1/ DQA1/DQB1 haplotype. DRB-1/DQA-1/DQB-1 haplotype 01702/00201/ 01304 00601/005011/ 00701 01501/00601/ 05701 01501/00601/02601 01501/00601/ 02301 010011/00201/01501 01502/00601/02301 Homo Hetero Homo Hetero Hetero Hetero Homo Hetero Hetero Hetero Affected 33/51 (0.65) 14/51 (0.27) 0/51 (0) 4/51 (0.08) 2/51 (0.04) 0/51 (0) 0/51 (0) 10/51 (0.2) 6/51 (0.12) 0/51 (0) Normal 6/57 (0.11) 20/57 (0.35) 6/57 (0.1) 22/57 (0.39) 14/57 (0.25) 8/57 (0.14) 3/57 (0.05) 13/57 (0.23) 6/57 (0.11) 1/57 (0.02) RR 5.9 0.78 0 0.2 0.16 0 0 0.85 1.1 0 OR* 15.58 0.7 0 0.14 0.13 0 0 0.83 1.13 0 * OR significant at 95% confidence interval. type had no effect on disease in heterozygotes (OR 1.13, RR 1.1). Single-locus DLA class II genotype and relationship to NME risk Fifty-one affected and 57 healthy Pugs were tested in the current study. Linkage between DRB1/DQA1/ DQB1 alleles was such that NME risk could be defined solely by specific DQB1 alleles and their zygosity (Table 3). DQB1*01501 was associated with the 010011/00201/01501 haplotype, DQB1*02301 with the 01501/00601/02301 or 01502/00601/02301 haplotypes, DQB1*05701 with the 01501/00601/05701 haplotype, DQB1*02601 with the 01501/00601/02601 haplotype, DQB1*00701 with the 00601/005011/ 00701 haplotype, and DQB1*01304 with the 01702/ 00201/01304 haplotype. DQB1*01501 was also associated with high risk when in a homozygous state and neutral when heterozygous, whereas DQB1*02301 and *05701 were protective in both heterozygous and homozygous states. DQB*02601 was protective when homozygous and neutral in a heterozygous state; DQB*00701 was neutral in a heterozygous state, whereas DQB*01304 was too infrequent to assess risk. To validate the concept that the risk for NME can be determined by the status of DQB1 alone, the data set of Table 3 was expanded by sequencing only the DQB1 genes from an additional 16 affected and 66 healthy appearing Pugs (Table 4). These data indicate that specific DQB1 alleles and their zygosity had the same predictive risk value as the DRB1/ DQA1/DQB1 haplotypes that they defined (compare Tables 3 and 4). Correlation between high-risk STR testing and NME risk Pugs from among the total sample collection were tested using 3 STR markers specific for the DLA class II region; 54 had NME and 229 were healthy at the time of sampling. Thirty-eight to 39 of the 54 NMEaffected Pugs were homozygous for the predominant STR alleles 341, 327, and 280, but only 18 of the 229 healthy Pugs were homozygous for these same alleles (RR 6.6 9.0, OR 22.2 30.5; Table 5). There was a strong, although not across-the-board, linkage between STR and DLA class II gene haplotypes (Table 6). The 341-327-280 STR haplotype defined the high-risk 3-locus DLA class II haplotype (010011/00201/01501), and Pugs homozygous for this STR haplotype were also homozygous for the high-risk 3-locus DLA class II haplotype and by inference for the high-risk DQB1*1501 genotype. The 01502/00601/02301 and 01501/00601/02301 haplotypes, which were protective against NME in both homozygous and heterozygous states, were not defined by a single STR haplotype. Rather, they were

Genetic tests for canine immune diseases 73 Table 4. Relative risk (RR) for affected and normal Pug dogs that were either homozygous (homo) or heterozygous (hetero) for the alleles at DQB1 locus with expanded data. DQB1 allele 01501 02301/05701 02601 00701 01304 Homo Hetero Homo Hetero Homo Hetero Hetero Hetero Affected 45/67 (0.67) 17/67 (0.25) 0/67 (0) 7/67 (0.11) 0/67 (0) 13/67 (0.19) 7/67 (0.11) 0/67 (0) Normal 10/123 (0.08) 52/123 (0.44) 34/123 (0.28) 58/123 (0.47) 8/123 (0.07) 38/123 (0.31) 11/123 (0.09) 1/123 (0.008) RR 8.38 0.6 0 0.23 0 0.61 1.22 0 correlated with several different STR haplotypes, including 339/325/276, 343/324/284, and 343/324/282. The risk-neutral 00601/005011/00701 haplotype was defined by the 339/323/284 STR haplotype and was only found in a heterozygous state. Single-nucleotide polymorphism testing for DLA-DQB1 allele and NME risk A 6-SNP panel, employing an automated system, e was used to identify DQB1 alleles in 99 healthy and 41 NME-affected Pugs from among the group that had been previously genotyped for DQB1 (Table 7). DQB1 allele assignments were based on SNPs 1, 2, 5, and 6, with SNPs 3 and 4 used for confirmation and backups for allele dropout. Five of the 6 known DQB1 alleles were automatically called by the software. The exception was DQB1*00701, which was automatically read as C at SNP6 and therefore erroneously designated DQB1*1304. Sequence similarities between DQB1 and DRB1 in this region resulted in a triple amplification of the C SNP and a single amplification of the G SNP, resulting in a large unanticipated C peak and a smaller anticipated G peak (Fig. 2B). This duplication was not evident with the other 5 DQB1 alleles because they all possessed C at SNP6 and were therefore called correctly on automatic read. There was concordance between DQB1*00701 sequence and SNP assignments when SNP6 peaks were manually examined and the C peak ignored (Table 7). Therefore, a simple SNP test e can be used to determine the risk for NME with the same accuracy as 1- or 3-locus DLA class II gene sequencing or a threesome of linked STR markers. Discussion Several genetic test approaches were used to assess DLA class II associated disease risk using NME of Pugs as a model. The tests took advantage of limited genetic diversity and high-linkage disequilibrium in the DLA and a disease association with specific DLA class II alleles. Three-locus exon 2 sequencing provided exact identification of the alleles involved and their zygosity, allele haplotypes, and disease risk. However, because of strong allele linkages (haplotyping) between loci, sequencing of only the DQB1 alleles provided identical information on zygosity and disease risk as 3-locus sequencing. Given similar conditions in the DLA class II region among all purebreds, single DLA class II gene sequencing should also be applicable to other breeds with different DLA class II associated immune disorders. The number of DLA class II alleles and haplotypes among dogs in general and purebreds in particular is very limited. 2,19,24 At least 143 DRB1, 26 DQA1, and 66 DQB1 alleles have been formally designated to date in all dogs, with a limited number still to be identified and/or named (L. J. Kennedy, personal communication, 2010). 17,18,21 Although the potential for allelic variation within the DLA class II region is great, this potential is limited by natural linkage disequilibrium in the DLA and by more recent inbreeding. Thus, DLA class II diversity among Table 5. Pug dogs.* Correlation of homozygosity of the simple tandem repeat (STR) markers with necrotizing meningoencephalitis (NME) in NME Unaffected STR Predominant homozygous allele in NME dogs CFA12 map position{ Dogs with predominant homozygous allele % Dogs with predominant homozygous allele % 12_005A_CA 341/341 5,018,175 39/54 0.72 18/229 0.08 9 30.48 12_005A_CT 327/327 5,361,083 38/54 0.70 18/229 0.08 8.75 27.84 12_005B_CA 280/280 5,677,936 39/54 0.72 24/229 0.11 6.55 22.21 *RR5 relative risk; OR 5 odds ratio. { Nucleotide position on Canis familiaris autosome 12. RR OR

74 Pedersen et al. Table 6. Correlation of simple tandem repeat (STR) and DRB1/DQA1/DQB1 haplotypes and the relative risk (RR) and odds ratio (OR) for necrotizing meningoencephalitis in Pug dogs. Haplotype DRB1 DQA1 DQB1 STR haplotype* Affected Control RR OR 010011 00201 01501 341-327-280 Homo 38/54 (0.7) 18/229 (0.08) 8.95 27.84 Hetero 13/54 (0.24) 85/229 (0.37 0.65 0.54 01502 00601 02301 339-325-276 Homo 0/54 (0) 12/229 (0.05) 0 0 Hetero 1/54 (0.02) 68/229 (0.3) 0.06 0.04 01501 00601 02301 343-324-284 Homo 0/54 (0) 4/229 (0.02) 0 0 Hetero 1/54 (0.02) 38/229 (0.17) 0.11 0.09 01501 00601 02301 343-324-282 Homo 0/54 (0) 3/229 (0.01) 0 0 01502 02601 Hetero 3/54 (0.06) 47/229 (0.21) 0.27 0.23 01501 00601 02601 339-324-282 Homo 0/54 (0) 19/229 (0.08) 0 0 Hetero 8/54 (0.15) 68/229 (0.3) 0.5 0.41 01501 00601 02601 343-324-280 Hetero 1/16 (0.02) 16/229 (0.07) 0.27 0.27 05701 00601 005011 00701 339-323-284 Hetero 5/54 (0.09) 23/229 (0.1) 0.92 0.91 01702 00201 01304 339-323-280 Hetero 0/54 (0) 1/229 (0.004) 0 0 * Homo 5 homozygous; hetero 5 heterozygous. 6,434 purebred and mixed-breed dogs tested to date has been limited to 189 3-locus haplotypes (L. J. Kennedy, personal communication, 2010). Furthermore, each purebred dog possesses only a small number of specific 3-locus haplotypes, 2,19 similar to Pugs, with only 7 haplotypes identified to date. A linked STR marker test also worked well for determining DLA class II associated risk for NME in Pugs. The linked STR marker test for Pug dog NME was quick, was relatively inexpensive, and could be easily incorporated into a larger panel including parentage and/or identification markers. 6 8 Nevertheless, association studies would have to be repeated with the present linked STR markers for other reported autoimmune and infectious diseases in dogs to ascertain their utility. The use of such a STR panel is not novel; a less DLA class II specific linked STR marker test has proved effective in assessing risk for generalized demodicosis in Boxers, Argentinean Mastiffs, and mixed-breed dogs. 16 The STRs also undergo more genetic drift than coding genes (calculated at 1.1 3 10 22 for dogs 8 ), although such events are uncommon in the time periods in question and are easily recognized when parentage testing is included. Single-nucleotide polymorphism testing e proved to be relatively simple in the Pug model and is applicable to DLA class II associated immune disorders in other breeds. A small battery of SNPs was derived using the known exon 2 sequences of the 6 DLA-DQB1 alleles in Pugs. The small SNP panels used in the current study were able to predict the DQB1 genotypes and, by extension, the DRB1/DQA1/DQB1 haplotypes. The SNP-based testing may prove to be a simple and highly cost-effective method for determining DLA class II associated genotypes and/or haplotypes in autoimmune and infectious disorders in purebreds. However, as would be the case for linked STR markers, each SNP panel would have to be customized for breed based on the disease and riskassociated DLA class II alleles. Care would also be required to read around occasional sequence similarities between DRB1 and DQB1 sequence homologies. The SNP testing, like STR assays, can also be easily incorporated with identification and parentage panels using a larger number (20 40) of SNPs identified for this purpose. 33 All of the assays described herein worked extremely well with relatively crude sodium hydroxide extracted, buccal swab derived DNA. This significantly reduces the cost of testing by eliminating the need for the collection of whole blood and more expensive extraction procedures. Buccal swabs are also easily stored in a dry state, and the DNA lasts for many years. It is important that tests such as this, which measure disease risk, not be used for diagnosing a specific immune disease. This is because immune disorders often involve multiple genes as well as nongenetic factors. 11,13 Autoimmunity, as is now known from human genetics, involves a number of genes related to the immune system, both within and outside of the MHC region, which work in concert with certain MHC class I and II gene variations. 4,14,41 Epigenetic factors are also important in disease risk. This is best exemplified by comparing the disease phenotype and associated genotype in NME of Pugs. The incidence of NME in Pugs based on previous studies is approximately 1.2%, whereas the incidence of the high-risk haplotype in a homozygous state is 6 11%. 9 When these figures are used, only 1 in 5 10 Pugs at genetic risk for NME will develop the disease

Genetic tests for canine immune diseases 75 Table 7. DQB1 genotype determined by single-nucleotide polymorphism (SNP) testing. e SNP1 SNP2 SNP5 SNP6 DQB1 genotype determined by SNP DQB1 genotype determined by exon 2 sequence Affected Control RR* G C T C 01501/01501 01501/01501 29/40 5/79 11.9 GT GC AT C 01501/02301 01501/02301 1/40 21/79 0.09 GT GC T C 01501/02601 01501/02601 6/40 7/79 1.69 GT TC T CG{ 01501/01304{ 01501/00701 3/40 1/79 5.77 T GT AT CG{ 02301/01304{ 02301/00701 1/40 2/79 1 T G A C 02301/02301 02301/02301 0 16/79 0 T GC AT C 02301/05701 02301/05701 0 4/79 0 T G AT C 02301/02601 02301/02601 0 13/79 0 T GC T C 05701/02601 05701/02601 0 2/79 0 T G T C 02601/02601 02601/02601 0 6/79 0 T GT T CG{ 02601/01304{ 02601/00701 0 2/79 0 *RR 5 relative risk. { SNP6 called automatically as C and therefore was incorrectly designated DQB1*01304; the smaller G peak of the correct allele DQB1*00701 was identified on manual reading (see Fig. 2). in their lifetime. Therefore, tests for susceptibility and protective genotypes and haplotypes described herein should be used only for assessing disease risk and not for diagnosing a specific disorder. Genetic testing can also be used to select breeding pairs that will produce puppies at low or negligible risk. The first inclination is to reduce risk by breeding only dogs possessing protective haplotypes. Elimination of the high-risk haplotype and increasing the incidence of protective haplotypes would lead to a significant loss of genetic diversity. The susceptibility haplotype is present in a heterozygous or homozygous state in approximately 50% of the Pug population. However, it is fortunate that the susceptibility haplotype in Pugs only increases risk for NME when in the homozygous state. This allows breeders to maintain the susceptibility haplotype, albeit in a heterozygous state. The ability to maintain a susceptibility haplotype while reducing the incidence of a specific autoimmune disease will vary from breed to breed and from one disorder to another. If homozygosity for a specific allele or haplotype greatly increases risk, than selective mating against homozygosity may be preferred. Acknowledgements Funding for this study was provided by the Center for Companion Animal Health and the Veterinary Genetics Laboratory, School of Veterinary Medicine, University of California, Davis. The authors are grateful for the assistance of Dr. Phil Kass, University of California, Davis, with some of the statistical analyses and for the many Pug dog owners and breeders who contributed DNA samples from affected and healthy dogs. Sources and manufacturers a. HotStarTaqH Plus, QIAquickH; Qiagen Inc., Valencia, CA. b. Big DyeH Terminator, ABI 3730; Applied Biosystems, Foster City, CA. c. Performa DTR Ultra, Edge BioSystems, Gaithersburg, MD. d. Vector NTI Advance TM, Invitrogen Corp., Carlsbad, CA. e. MassARRAYH version 4.0.0.2, MassARRAYH Compact 96 System featuring iplexh Gold technology; Sequenom Inc., San Diego, CA. Figure 2. A, all DQB1 genotypes are homozygous (C/C) at single-nucleotide polymorphism 6 except for those containing DQB1*00701 (see panel B); B, the DQB1 genotypes containing *00701 are heterozygous (C/G) at single-nucleotide polymorphism 6 as a result of an uneven coamplification of closely related DRB1 and DQB1 sequences. References 1. Angles JM, Famula TR, Pedersen NC: 2005, Uveodermatologic (VKH-like) syndrome in American Akita dogs is associated with an increased frequency of DQA1*00201. Tissue Antigens 66:656 665. 2. Angles JM, Kennedy LJ, Pedersen NC: 2005, Frequency and distribution of alleles of canine MHC-II DLA-DQB1, DLA- DQA1 and DLA-DRB1 in 25 representative American Kennel Club breeds. Tissue Antigens 66:173 184.

76 Pedersen et al. 3. Bannasch DL, Bannasch MJ, Ryun JR, et al.: 2005, Y chromosome haplotype analysis in purebred dogs. Mamm Genome 16:273 280. 4. Baranzini SE: 2009, The genetics of autoimmune diseases: a networked perspective. Curr Opin Immunol 21:596 605. 5. Barnes A, O Neill T, Kennedy LJ, et al.: 2009, Association of canine anal furunculosis with TNFA is secondary to linkage disequilibrium with DLA-DRB1*. Tissue Antigens 73: 218 224. 6. Binns MM, Holmes NG, Marti E, Bowen N: 1995, Dog parentage testing using canine microsatellites. J Small Anim Pract 36:493 497. 7. DeNise S, Johnston E, Halverson J, et al.: 2004, Power of exclusion for parentage verification and probability of match for identity in American Kennel Club breeds using 17 canine microsatellite markers. Anim Genet 35:14 17. 8. Eggleston ML, Irion DN, Schaffer AL, et al.: 2002, PCR multiplexed microsatellite panels to expedite canine genetic disease linkage analysis. Anim Biotechnol 13:223 235. 9. Greer KA, Schatzberg SJ, Porter BF, et al.: 2009, Heritability and transmission analysis of necrotizing meningoencephalitis in the Pug. Res Vet Sci 86:438 442. 10. Greer KA, Wong AK, Liu H, et al.: 2010, Necrotizing meningoencephalitis of pug dogs associates with dog leukocyte antigen class II and resembles acute variant forms of multiple sclerosis. Tissue Antigens 76:110 118. 11. Handunnetthi L, Ramagopalan SV, Ebers GC, Knight JC: 2009, Regulation of major histocompatibility complex class II gene expression, genetic variation and disease. Genes Immun 11:99 112. 12. Hughes AM, Jokinen P, Bannasch DL, et al.: 2010, Association of a dog leukocyte antigen class II haplotype with hypoadrenocorticism in Nova Scotia Duck Tolling Retrievers. Tissue Antigens 75:684 690. 13. International MHC and Autoimmunity Genetics Network, et al.: 2009, Mapping of multiple susceptibility variants within the MHC region for 7 immune-mediated diseases. Proc Natl Acad Sci U S A 106:18680 18685. 14. Invernizzi P, Gershwin ME: 2009, The genetics of human autoimmune disease. J Autoimmun 33:290 299. 15. Irion DN, Schaffer AL, Famula TR, et al.: 2003, Analysis of genetic variation in 28 dog breed populations with 100 microsatellite markers. J Hered 94:81 87. 16. It V, Barrientos L, López Gappa J, et al.: 2010, Association of canine juvenile generalized demodicosis with the dog leukocyte antigen system. Tissue Antigens 76:67 70. 17. Kennedy LJ: 2007, 14th international HLA and immunogenetics workshop: report on joint study on canine DLA diversity. Tissue Antigens 69(suppl. 1):269 271. 18. Kennedy LJ, Barnes A, Happ GM, et al.: 2002, Evidence for extensive DLA polymorphism in different dog populations. Tissue Antigens 60:43 52. 19. Kennedy LJ, Barnes A, Happ GM, et al.: 2002, Extensive interbreed, but minimal intrabreed, variation of DLA class II alleles and haplotypes in dogs. Tissue Antigens 59:194 204. 20. Kennedy LJ, Barnes A, Ollier WER, Day MJ: 2006, Association of a common dog leukocyte antigen class II haplotype with canine primary immune-mediated haemolytic anaemia. Tissue Antigens 68:502 508. 21. Kennedy LJ, Brown JJ, Barnes A, et al.: 2008, Major histocompatibility complex typing of dogs from Russia shows further dog leukocyte antigen diversity. Tissue Antigens 71:151 156. 22. Kennedy LJ, Carter SD, Barnes A, et al.: 1998, Nine new dog DLA-DRB1 alleles identified by sequence-based typing. Immunogenetics 48:296 301. 23. Kennedy LJ, Davison LJ, Barnes A, et al.: 2006, Identification of susceptibility and protective major histocompatibility complex haplotypes in canine diabetes mellitus. Tissue Antigens 68:467 476. 24. Kennedy LJ, Hall LS, Carter SD, et al.: 2000, Identification of further DLA-DRB1 and DQA1 alleles in the dog. Eur J Immunogenet 27:25 28. 25. Kennedy LJ, Huson HJ, Leonard J, et al.: 2006, Association of hypothyroid disease in Doberman pinscher dogs with a rare major histocompatibility complex DLA class II haplotype. Tissue Antigens 67:53 56. 26. Kennedy LJ, Quarmby S, Happ GM, et al.: 2006, Association of canine hypothyroidism with a common major histocompatibility complex DLA class II allele. Tissue Antigens 68:82 86. 27. Levine JM, Fosgate GT, Porter B, et al.: 2008, Epidemiology of necrotizing meningoencephalitis in Pug dogs. J Vet Intern Med 22:961 968. 28. Ollier WE, Kennedy LJ, Thomson W, et al.: 2001, Dog MHC alleles containing the human RA shared epitope confer susceptibility to canine rheumatoid arthritis. Immunogenetics 53:669 673. 29. Pedersen NC: 1999, A review of immunologic diseases of the dog. Vet Immunol Immunopathol 69:251 342. 30. Quinnell RJ, Kennedy LJ, Barnes A, et al.: 2003, Susceptibility to visceral leishmaniasis in the domestic dog is associated with MHC class II polymorphism. Immunogenetics 55:23 28. 31. Sambrook J, Russell DW: 2001, Molecular cloning. In: A laboratory manual, 3rd ed., pp. 6.24 6.26. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 32. Schatzberg SJ, Haley NJ, Barr SC, et al.: 2005, Polymerase chain reaction screening for DNA viruses in paraffinembedded brains from dogs with necrotizing meningoencephalitis, necrotizing leukoencephalitis, and granulomatous meningoencephalitis. J Vet Intern Med 19:553 559. 33. Seddon JM, Parker HG, Ostrander EA, Ellegren H: 2005, SNPs in ecological and conservation studies: a test in the Scandinavian wolf population. Mol Ecol 14:503 511. 34. Svejgaard A: 1996, MHC and disease association. In: Weir s handbook of experimental immunology, 5th ed., pp. 1 13. Blackwell Scientific, New York, NY. 35. Talarico LR, Schatzberg SJ: 2010, Idiopathic granulomatous and necrotising inflammatory disorders of the canine central nervous system: a review and future perspectives. J Small Anim Pract 51:138 149. 36. Webb KM, Allard MW: 2009, Mitochondrial genome DNA analysis of the domestic dog: identifying informative SNPs outside of the control region. J Forensic Sci 54: 275 288. 37. Wilbe M, Jokinen P, Hermanrud C, et al.: 2009, MHC class II polymorphism is associated with a canine SLE-related disease complex. Immunogenetics 61:557 564. 38. Wilbe M, Sundberg K, Hansen IR, et al.: 2010, Increased genetic risk or protection for canine autoimmune lymphocytic thyroiditis in Giant Schnauzers depends on DLA class II genotype. Tissue Antigens 75:712 719. 39. Wong AK, Neff MW: 2009, DOGSET: pre-designed primer sets for fine-scale mapping and DNA sequence interrogation in the dog. Anim Genet 40:569 571. 40. Wong AK, Ruhe AL, Dumont BL, et al.: 2010, A comprehensive linkage map of the dog genome. Genetics 184:595 605. 41. Zenewicz LA, Abraham C, Flavell RA, Cho JH: 2010, Unraveling the genetics of autoimmunity. Cell 140:791 797.