Canine diabetes mellitus: from phenotype to genotype

Authors

  • B. Catchpole,

    1. Department of Pathology & Infectious Diseases, Royal Veterinary College, University of London, London AL9 7TA
      *Centre for Integrated Genomic Medical Research, Faculty of Medical and Human Sciences, University of Manchester, Manchester M13 9PT
      †Department of Veterinary Medicine, University of Cambridge, Cambridge CB3 0ES
    Search for more papers by this author
  • L. J. Kennedy,

    1. Department of Pathology & Infectious Diseases, Royal Veterinary College, University of London, London AL9 7TA
      *Centre for Integrated Genomic Medical Research, Faculty of Medical and Human Sciences, University of Manchester, Manchester M13 9PT
      †Department of Veterinary Medicine, University of Cambridge, Cambridge CB3 0ES
    Search for more papers by this author
  • L. J. Davison,

    1. Department of Pathology & Infectious Diseases, Royal Veterinary College, University of London, London AL9 7TA
      *Centre for Integrated Genomic Medical Research, Faculty of Medical and Human Sciences, University of Manchester, Manchester M13 9PT
      †Department of Veterinary Medicine, University of Cambridge, Cambridge CB3 0ES
    Search for more papers by this author
  • W. E. R. Ollier

    1. Department of Pathology & Infectious Diseases, Royal Veterinary College, University of London, London AL9 7TA
      *Centre for Integrated Genomic Medical Research, Faculty of Medical and Human Sciences, University of Manchester, Manchester M13 9PT
      †Department of Veterinary Medicine, University of Cambridge, Cambridge CB3 0ES
    Search for more papers by this author

Abstract

inline image

Breed differences in susceptibility to diabetes mellitus in dogs suggest an underlying genetic component to the pathogenesis of the disease. There is little evidence for an equivalent of human type 2 diabetes in dogs, and it has been proposed that canine diabetes is more comparable to the type 1 form of the disease. Certain immune response genes, particularly those encoding major histocompatibility complex molecules involved in antigen presentation, are important in determining susceptibility to human type 1 diabetes. We tested the hypothesis that canine major histocompatibility complex genes (known as the dog leucocyte antigen) are associated with diabetes in dogs. A total of 530 diabetic dogs and more than 1000 controls were typed for dog leucocyte antigen, and associations were found with three specific haplotypes. The DLA-DRB1*009/DQA1*001/DQB1*008 haplotype shows the strongest association with diabetes in the UK dog population. This haplotype is common in diabetes-prone breeds (Samoyed, cairn terrier and Tibetan terrier) but rare in diabetes-resistant breeds (boxer, German shepherd dog and golden retriever), which could explain differences in the prevalence of diabetes in these different breeds. There is evidence that the DLA-DQA1*001 allele is also associated with hypothyroidism, suggesting that this could represent a common susceptibility allele for canine immune-mediated endocrinopathies.

Background

Historically, dogs have played a pivotal role in our understanding of the pathophysiology and treatment of diabetes mellitus. While studying the involvement of the pancreas in digestion, Joseph von Mering and Oskar Minkowski discovered that pancreatectomised dogs displayed signs of polydipsia and polyuria, associated with the presence of glucose in the urine (Von Mering and Minkowski 1890). They had inadvertently created an animal model of diabetes, still used today in some research laboratories, and correctly concluded that the pancreas must secrete an “antidiabetogenic factor”, later found to be insulin, which enables the body to utilise glucose. In 1921, Marjorie, a diabetic crossbreed dog, became the first recipient of insulin therapy, paving the way for treatment of human patients (Banting and others 1922). However, since that time, the non-obese diabetic mouse has become the preferred model of the human disease, and research into the pathogenesis and genetics of canine diabetes has been limited. In this review, we will describe our attempts to use the human disease as a “model” of canine diabetes, specifically to try to understand the genetic factors involved in breed susceptibility.

Classification of human diabetes

Human diabetes can be divided into various types, based on the underlying cause of the disease (Table 1). A major distinction is made between patients with type 1 and those with type 2 diabetes. Type 1 diabetes is typically a disease of juvenile onset and is associated with insulin deficiency, usually resulting from autoimmune destruction of pancreatic beta cells. Type 1 diabetic patients are prone to developing ketoacidosis and require insulin therapy to control their hyperglycaemia. Type 2 diabetes is typically a disease of adult onset (although it is becoming increasingly recognised in children) and is caused by a combination of defective insulin secretion with reduced insulin sensitivity of the tissues. This impaired glucose tolerance is usually associated with obesity and physical inactivity. Hyperglycaemia in type 2 diabetic patients can be often managed using a combination of dietary modification and oral hypoglycaemic drugs.

Table 1. Classification of human diabetes mellitus*
I. Type 1 diabetes
 Beta cell destruction, usually leading to absolute insulin deficiency
  A. Immune mediated
  B. Idiopathic
II. Type 2 diabetes
 May range from predominantly insulin resistance with relative insulin deficiency to a predominantly secretory defect with insulin resistance
III. Other specific types
 A. Genetic defects of beta cell function
 B. Genetic defects in insulin action
 C. Diseases of the exocrine pancreas
 D. Endocrinopathies
 E. Drug induced or chemical induced
 F. Infections
 G. Uncommon forms of immune-mediated diabetes
IV. Gestational diabetes

It has become increasingly recognised that there are some diabetic patients who develop diabetes as adults (usually older than 30 years of age), who are initially diagnosed as having type 2 diabetes but who are not overweight and who also show evidence of circulating autoantibodies. These patients are often managed initially using oral hypoglycaemic drugs, but they usually progress to requiring insulin therapy. This form of diabetes has now been classified as latent autoimmune diabetes of adults (LADA), sometimes referred to as type 1·5 diabetes. Juvenile-onset type 1 diabetes appears to result from an overwhelming autoimmune response against pancreatic beta cells that rapidly leads to their total destruction. In contrast, LADA seems to result from a more slowly progressive autoimmune process, and it takes several years before the beta cell mass has reduced to an extent whereby normoglycaemia cannot be maintained and clinical signs of diabetes become apparent.

Classification of diabetes in dogs

Canine diabetes is a heterogeneous disease, where several potential pathological mechanisms lead to hyperglycaemia. However, unlike the situation in human diabetes, there are no internationally accepted criteria for the classification of diabetes in dogs. It has been suggested that diabetes in veterinary species can be subdivided into either insulin-dependent diabetes mellitus (IDDM) or non-insulin-dependent disease (Feldman and Nelson 2004), although this is not particularly helpful in canine diabetes because virtually all diabetic dogs require insulin therapy. Furthermore, it would seem more appropriate to classify canine diabetes based on the underlying pathogenesis of disease rather than the clinical response to insulin treatment.

There are striking similarities between some forms of human diabetes and canine diabetes, but the human classification system is not easy to apply to dogs, as there are also clear differences in the mechanisms leading to hyperglycaemia, comparing the two species. Unlike the situation in human beings (and cats), there is little evidence that dogs suffer from a type 2 equivalent form of the disease. This is despite the fact that obesity is as much a problem in pet dogs as it is in the human population.

The authors currently use a classification system for canine diabetes that is based on the underlying cause of the hyperglycaemia (Table 2). Primary insulin deficiency diabetes (IDD) occurs when there is a lack of insulin production from the pancreas (absolute insulin deficiency), whereas primary insulin resistance diabetes (IRD) occurs when there is a lack of adequate insulin function in the tissues (relative insulin deficiency). Prolonged hyperglycaemia (that is, blood glucose more than 14 mmol/l) can, in itself, produce permanent beta cell dysfunction in dogs (Imamura and others 1988). Therefore, those animals with primary IRD will often progress to IDD, presumably as the result of secondary loss of beta cells that might be associated with glucose toxicity or beta cell exhaustion. Such cases might appear to have IDD at diagnosis, but control of hyperglycaemia with insulin therapy would prove difficult unless the underlying cause of insulin resistance has been addressed.

Table 2. Proposed classification system for canine diabetes mellitus
  • *

    Dogs can progress from IRD to secondary IDD as a consequence of beta cell loss associated with uncontrolled hyperglycaemia

Insulin deficiency diabetes (IDD) – absolute insulin deficiency
Primary IDD in dogs is characterised by a progressive loss of pancreatic beta cells. The aetiology of beta cell deficiency/destruction in diabetic dogs is currently unknown, but a number of disease processes are thought to be involved:
 • Congenital beta cell hypoplasia/abiotrophy
 • Beta cell loss associated with exocrine pancreatic disease (for example, pancreatitis)
 • Immune-mediated beta cell destruction
 • Idiopathic process
Insulin resistance diabetes (IRD) – relative insulin deficiency*
Primary IRD usually results from antagonism of insulin function by other hormones. These include the following:
 • Dioestrus/gestational diabetes
 • Secondary to other endocrinopathies
  ○ Hyperadrenocorticism
  ○ Acromegaly
 • Iatrogenic
  ○ Synthetic glucocorticoids
  ○ Synthetic progestagens

IDD, typically affecting dogs between five and 12 years of age, seems to be the most common type of diabetes in the UK dog population (Davison and others 2004). This is in contrast to the situation in human beings, where type 1 diabetes is typically of juvenile onset, affecting most individuals during adolescence. We have occasionally seen juvenile-onset diabetes in Labrador retrievers (six dogs from a total of 860 registered cases) that developed diabetes at three to six months of age. However, rather than resulting from an autoimmune process, this is likely to be caused by an inherited defect of beta cell development, leading to congenital beta cell aplasia/abiotrophy (Gepts 1965).

In the majority of canine diabetic patients, the cause of the insulin deficiency fails to be established, although it is believed to result from beta cell damage associated with either pancreatitis (Watson 2003) and/or immune-mediated beta cell destruction (Catchpole and others 2005). In one study, there was evidence of subclinical exocrine pancreatic disease in diabetic dogs, with two of 12 dogs showing evidence of exocrine pancreatic insufficiency (trypsin-like immunoreactivity [cTLI] less than 2·5 μg/l) and a further two dogs showing evidence of pancreatitis (canine pancreatic lipase immunoreactivity [cPLI] more than 200 μg/l) (Fleeman and others 2004).

There is evidence that autoantibodies are present in a proportion of diabetic dogs, suggesting that autoimmunity might be involved in the pathogenesis of disease in some animals. In one study, anti-islet-cell antibodies were detected in 50 per cent of newly diagnosed diabetic dogs (Hoenig and Dawe 1992). In our research, four of 30 newly diagnosed diabetic dogs tested positive for autoantibodies against the 65 KDa isoform of canine glutamic acid decarboxylase (GAD65), and two of these were also positive for antibodies against the insulinoma-associated antigen-2 (IA-2) autoantigen (Catchpole and others 2005). The onset of diabetes in middle age and the presence of circulating autoantibodies allow us to speculate that canine diabetes is more comparable to LADA rather than the classical juvenile-onset type 1 diabetes.

As there appears to be clear similarities between human and canine diabetes, we decided to use a candidate gene approach to determine whether these similarities would also be reflected at the genetic level. The first candidate genes that we studied were those of the major histocompatibility complex (MHC), as these immune response genes are the most important genetic risk factor for susceptibility to human type 1 diabetes and LADA (Kobayashi and others 2006).

Genetics of human diabetes: The role of MHC genes

There are around 20 different genetic loci that have been implicated in susceptibility to human type 1 diabetes (http://t1dbase.org/page/Loci/display/species/Human). However, IDDM1, containing MHC genes, is by far the most important disease susceptibility locus, conferring around 50 per cent of the total genetic risk (Todd and others 1987).

The MHC genes code for molecules responsible for antigen presentation to T lymphocytes, a process that is important for stimulating immune responses against infectious organisms but which is also likely to be involved in immune-mediated disease. There are two major classes of MHC molecules: MHC class I, responsible for presenting antigen located in the cytoplasm, and MHC class II, responsible for presenting antigen located in either the extracellular fluid or in intracellular vesicles. To ensure that mammals are able to respond to a wide variety of different pathogens, MHC genes have evolved in various ways to ensure efficient antigen presentation.

In human beings, there are multiple MHC class I genes, located in series on chromosome 6. The best characterised and most polymorphic of these genes are human leucocyte antigen (HLA)-A, HLA-B and HLA-C (Fig 1). Within the class II region, there is also a group of highly polymorphic genes called HLA-DR, HLA-DQ and HLA-DP. MHC class II molecules are heterodimers consisting of two different subunits (alpha and beta) (Fig 1), which are coded by different genes (for example, HLA-DRA and HLA-DRB). The genes coding for MHC class II molecules are highly polymorphic (more than 559 alleles for HLA-DRB) apart from HLA-DRA, which has a limited number of alleles. MHC class II genes are closely located on chromosome 6 and are inherited together as a set, which is referred to as a haplotype. This is often represented as HLA-DRB*n/DQA*n/DQB*n, where “n” represents the allele number.

Figure 1.

HLA genetics and MHC protein structure. Selected HLA genes are shown, located on human chromosome 6. HLA-A, HLA-B and HLA-C are MHC class I genes and HLA-DR, HLA-DQ and HLA-DP are MHC class II genes. The corresponding MHC proteins are shown as they would be expressed on the cell surface. Note that MHC class II molecules consist of two subunits (α and β), and therefore, there are two corresponding genes (for example, DRA and DRB). For antigen presentation, a peptide would be located in the binding groove of the MHC molecule as indicated. This MHC-peptide complex could then be recognised by a T cell receptor (CD4+ T cells detect MHC class II complex and CD8+ T cells detect MHC class I complex). HLA Human leucocyte antigen, MHC Major histocompatibility complex

As the MHC is both polygenic (that is, there is more than one MHC class I gene) and polymorphic (that is, each MHC gene can be highly variable), this gives rise to a huge diversity of MHC expression in an outbred population. Variation between different alleles impacts on the antigen-binding capacity of the MHC molecules they encode. Therefore, the MHC genes we inherit will dictate the antigen repertoire that we, as individuals, are able to present to the immune system and consequently influence our ability to fight infection. However, when the MHC gene pool becomes restricted in diversity (for example, in some ethnic groups in the human population and in some breeds of dog), this can lead to problems associated with the immune system.

As well as playing an important role in immune responses to infection, MHC genes can determine susceptibility to immune-mediated diseases, including allergy and autoimmunity. The association between MHC genes and diabetes was first reported more than 30 years ago (Nerup and others 1974), and it is clear now that certain MHC alleles/haplotypes are involved in susceptibility to diabetes, whereas others are protective. Initial studies demonstrated an association between HLA-DR3 and HLA-DR4 with type 1 diabetes (Platz and others 1981), with heterozygous (DR3/DR4) individuals most at risk, whereas individuals possessing HLA-DR2 are protected from developing diabetes. However, the particular MHC alleles that are linked to diabetes in the human population appear to be dependent upon ethnic origin (Zamani and Cassiman 1998). Thus, using human diabetes as a model, this is likely to be an important consideration when attempting to identify MHC susceptibility alleles in different dog breeds.

The molecular basis for the MHC genetic association with diabetes was shown to be likely because of differences in the structure of MHC proteins encoded by susceptibility alleles. Diabetes seems to be linked to the presence of an amino acid other than aspartic acid at position 57 of the HLA-DQ beta chain (DQβ non-Asp57) (Todd and others 1987) and also with the presence of arginine at position 52 of the HLA-DQ alpha chain (DQα Arg52) (Khalil and others 1990). Amino acid differences at these locations could impact on presentation of antigenic peptides by MHC molecules that will, in turn, influence immune responses, tolerance and autoimmunity, although the precise mechanism as to how this relates to diabetes still remains to be established.

The role of MHC genes in canine diabetes

Certain breeds of dog appear to be predisposed to developing diabetes. A database containing medical records of more than 6000 diabetic cases from veterinary schools in North America identified breeds including the miniature schnauzer, bichon frise, miniature poodle, Samoyed and cairn terrier as having an increased risk of the disease (Guptill and others 2003). A similar breed distribution was seen in our own database (Table 3) in which the Samoyed, Tibetan terrier and cairn terrier were the breeds at “high risk” for diabetes in the UK. In contrast, other breeds, including the boxer, German shepherd dog and golden retriever, appear to have a reduced risk of developing diabetes (Marmor and others 1982, Guptill and others 2003, Davison and others 2004). These breed differences in disease susceptibility would suggest that there is a significant genetic component to diabetes in dogs and that there may be, in fact, differences in genetic risk factors that are breed specific.

Table 3. Breed distribution of diabetic dogs in the UK
BreedNumber (per cent) of each breedOdds ratio (95 per cent confidence interval)
UK canine diabetes register (n=800)UK insured dog population* (n=46,593)
  • *

    Data courtesy of Pet Protect Insurance Ltd

Samoyed36 (4·5)101 (<0·1)21·7 (14·7-31·9)
Tibetan terrier17 (2)140 (0·3)7·2 (4·3-12·0)
Cairn terrier24 (3)220 (0·5)6·5 (4·3-10·0)
Miniature schnauzer17 (2)244 (0·5)4·1 (2·5-6·8)
Yorkshire terrier48 (6)819 (1·8)3·6 (2·6-4·8)
Border terrier15 (2)361 (0·8)2·4 (1·5-4·1)
Labrador retriever107 (13·4)6680 (14·3)0·9 (0·8-1·1)
Golden retriever10 (1·25)3050 (6·5)0·18 (0·1-0·33)
German shepherd dog8 (1)2814 (6·0)0·16 (0·08-0·31)
Boxer2 (<1)1906 (4·1)0·06 (0·01-0·24)

Research in our laboratory has indicated that autoantibodies can be detected in a proportion of diabetic dogs, suggesting that there is an immune-mediated component to the pathogenesis of diabetes in some cases (Catchpole and others 2005). We, therefore, predicted that susceptibility might be associated with immune response genes, particularly those encoding canine MHC class II molecules (known as dog leucocyte antigen [DLA] genes), located on canine chromosome 12 (Fig 2) (Debenham and others 2005).

Figure 2.

Canine DLA gene location and polymorphism. DLA-DR and DLA-DQ are closely located on canine chromosome 12 (CFA12). The number of alleles currently identified in the dog population is shown for each locus. These numbers are currently increasing as new alleles are discovered. DLA Dog leucocyte antigen

The DLA-DRA1 locus appears to be monomorphic in the dog population, but the other loci are polymorphic, with 102 DLA-DRB1, 26 DLA-DQA1 and 62 DLA-DQB1 alleles currently recognised (http://www.ebi.ac.uk/ipd/mhc/dla/index.html). Of particular note is that there is extensive interbreed, but often minimal intrabreed, DLA genetic variation in the dog population (Kennedy and others 2002). Some dog breeds in particular have become restricted in their DLA gene pool as a result of selective breeding. This could explain breed differences in immune responses to infection/vaccination, for example, Rottweiler dogs that have a very restricted DLA profile (Kennedy and others 2002) seem to respond poorly to canine parvovirus (Houston and others 1996). Additionally, selection for particular DLA alleles/haplotypes in a breed could lead to an increase in the prevalence of immune-mediated disease.

In a pilot study, funded by the Kennel Club Charitable Trust, 120 diabetic dogs were DLA genotyped, and one haplotype (DLA-DRB1*009/DQA1*001/DQB1*008) was found to be overrepresented in the diabetic group compared with breed-matched controls (Kennedy and others 2003). A grant from Petsavers allowed this research to be continued to investigate the DLA genetics of canine diabetes with a focus on trying to explain the differences in breed susceptibility to the disease. In total, 530 diabetic dogs and more than 1000 control dogs were DLA genotyped and analysed (Kennedy and others 2006a). Seventy diabetic dogs in this group were female entire at the time of diagnosis and, therefore, possibly suffering from dioestrus diabetes. As this is unlikely to have a particular DLA association, these female entire diabetic dogs were excluded from the analysis. The DLA-DRB1*009/DQA1*001/DQB1*008 haplotype was confirmed as being associated with for diabetes in this larger cohort of patients (Fig 3). In addition, two other haplotypes were found to have significantly increased prevalence in the diabetic population: DLA-DRB1*015/DQA1*006/DQB1*023 and DLA-DRB1*002/DQA1*009/DQB1*001. In contrast, one DLA-DQ haplotype, DLA-DQA1*004/DQB1*013, had significantly reduced prevalence in diabetic cases compared with controls, suggesting that this could be a protective haplotype, similar to that seen with HLA-DR2 in human diabetes.

Figure 3.

*DLA haplotypes associated with diabetes susceptibility and protection. DLA alleles for DRB1/DQA1/DQB1 are shown as a haplotype. “x” represents any DRB1 allele. Odds ratios, 95 per cent confidence intervals and statistical significance are shown on a logarithmic scale and were calculated comparing diabetic cases (n=460) with controls (n=1047). DLA Dog leucocyte antigen

As there is considerable interbreed, but often minimal intrabreed, variability in DLA haplotype frequencies, this can make it difficult to discriminate between alleles/haplotypes associated with disease and those merely associated with the breed profile of the sample groups used for the analysis. It could be argued that the DLA associations with diabetes that were identified are merely “markers” of susceptibility, and the true susceptibility loci lie elsewhere in the genome that are either in linkage disequilibrium with the DLA or associated with particular DLA haplotypes in some other way. There were insufficient numbers of samples to perform individual breed case-control analysis, so instead different breeds were grouped together for analysis, according to their diabetes risk status, in an attempt to determine whether particular DLA haplotypes segregated with susceptibility or resistance to the disease. To compare between breeds and as the DQB1 locus did not appear to contribute significantly to the DLA association, we focused on analysing DLA-DRB1/DQA1 two locus haplotypes. In the “high-risk” group of dogs (Samoyed, cairn terrier and Tibetan terrier), similar genotypes were seen (Table 4). It is clear that these three diabetes-prone breeds share similar DLA alleles, despite originating from disparate gene pools. Indeed, of the 44 diabetic dogs in this group, 33 dogs expressed two high-risk haplotypes, whereas only one dog failed to express any. In contrast, the DLA-DRB1*009/DQA1*001 susceptibility haplotype was not at all present in the “low-risk” group of control dogs (boxer, German shepherd dog and golden retriever).

Table 4. DLA genotypes of breeds at high risk for diabetes*
Haplotype 1 (DRB1/DQA1)Haplotype 2 (DRB1/DQA1)Samoyed (n=19)Cairn terrier (n=18)Tibetan terrier (n=7)Total (n=44)
  • DLA Dog leucocyte antigen

  • *

    Two locus DLA haplotypes are shown for diabetic dogs of the indicated breeds

009/001009/0011203
015/006015/00659317
002/009002/0090101
009/001015/00682111
009/001002/0090101
009/001Other haplotype2002
015/006Other haplotype3328
Other haplotypeOther haplotype0011

Two of the three diabetes susceptibility haplotypes (DLA-DRB1*009/DQA1*001 and DLA-DRB1*015/DQA1*006) appear to be common particularly in Samoyed dogs (in both diabetic cases and controls). This suggests that the breed as a whole carries MHC genes that predispose Samoyeds to developing diabetes, but, presumably, other genetic and environmental factors determine whether or not clinical disease occurs. This means that it is unlikely that DLA genotyping per se will allow breeding programmes to be established within the current “closed” gene pool that could reduce the prevalence of diabetes in this breed.

The molecular basis for the HLA association with type 1 diabetes in human beings is linked to variability in amino acids at key positions in the MHC class II molecule. Although there did not appear to be any association with the presence or absence of aspartic acid at position 57 on the canine DQ beta chain, there was an association with DQA alleles that coded for arginine at position 55 (Fig 3). This positively charged amino acid is likely to impact on antigen binding to the canine MHC molecule and might be analogous to the HLA-DQ alpha Arg52 association with human diabetes.

We investigated the possibility that diabetic dogs could have multiple endocrinopathies associated with autoimmunity. None of the 530 diabetic dogs analysed had a definitive diagnosis of hypothyroidism based on low serum thyroxine (T4) with concurrent high endogenous thyroid stimulating hormone (cTSH), although diabetes with concurrent hypothyroidism has been reported elsewhere (Hess and others 2000). There were five diabetic dogs with concurrent hypoadrenocorticism (based on a deficiency of cortisol after adrenocorticotropic hormone (ACTH) stimulation). These rare cases might be equivalent to human autoimmune polyendocrine syndrome type 2 (Betterle and others 2004). Four out of these five dogs expressed an Arg55-positive DQA allele. Additionally, two of the five dogs expressed the DLA-DRB1*009/DQA1*001/DQB1*008 high-risk haplotype.

Recent studies have demonstrated an association between canine hypothyroidism and the DLA-DQA1*001 allele (Kennedy and others 2006b), which is Arg55 positive and part of the DLA-DRB1*009/DQA1*001/DQB1*008 haplotype associated with diabetes. It is also interesting to note that the DLA-DRB1*009 allele, associated with diabetes, carries the “shared epitope”, a sequence of five amino acids in the third hypervariable region of DRB1, which has also been associated with rheumatoid arthritis in human beings (Wordsworth and others 1989) and also in the dog (Ollier and others 2001). The possibility exists that these particular DLA alleles constitute genetic risk factors for autoimmunity, with other genes dictating the type of autoimmune disease that might develop. This would be similar to the situation in human beings where the same HLA alleles are associated with several autoimmune disorders (Gregersen and Behrens 2006).

Conclusions

The results of our initial genetic analysis suggest that DLA genes play a role in determining susceptibility to canine diabetes. These results go some way to explaining the genetic basis for the breed differences we see in susceptibility to the disease. However, diabetes is a complex genetic disorder and there are likely to be many other genes involved. Furthermore, yet we have to determine what environmental risk factors are involved in triggering clinical disease in a genetically susceptible individual.

Acknowledgements

We are extremely grateful to The Kennel Club Charitable Trust and to Petsavers for funding this research and to the owners and veterinary surgeons who have contributed to the UK Canine Diabetes Register and Archive.

Ancillary