Gastric dilatation-volvulus is a life-threatening condition that occurs at high frequency in many large and giant breeds of dogs, particularly those with deep, narrow thoracic cavities.1–4 The lifetime risk of GDV exceeds 5% for all large-breed dogs combined5 and rises to 37% in Great Danes specifically.6
The disease involves accumulation of gas in the stomach (dilatation) and torsion of the stomach on its axis (volvulus), which blocks the escape of gas. In combination, these events result in rapid expansion of the stomach, causing severe compression of vital gastric, cardiac, and other blood vessels. Tissue damage, shock, and death rapidly ensue without aggressive treatment. Decompression and surgical reorientation of the stomach can ameliorate the immediate problem, and gastropexy can prevent subsequent torsion events.1,2,6 Future episodes of dilatation may still occur after gastropexy,6 but this recurrence is not typically life-threatening.
Gas accumulation results primarily from bacterial fermentation in the stomach.7 The torsion is often considered a secondary and passive response to the increased gastric pressure and space restrictions, but the highly contorted state of the stomach at full volvulus2 appears more consistent with an active process involving abnormal and uneven contractions of the gastric musculature. Although dilatation can occur independently of volvulus,1,2,6 particularly after gastropexy, data indicate that most, if not all, dogs with dilatation also have some degree of volvulus, suggesting that volvulus may be the primary driver of GDV.8
The causes of GDV are not clearly understood, but several risk factors have been identified, including advancing age,4,5 dietary factors,9–11 behavioral factors,12,13 preexisting health conditions,14 and genetic factors.2–5,15–17 Dietary factors such as food particle size18 and oil content,11 frequency and size of meals,12,19 and elevation of the food dish5 have been reported, but the importance of these factors is unclear. Behavioral factors such as vigorous exercise shortly after a meal13 or stress12,19 have also been linked to GDV, but these factors are difficult to quantify. Temperament or stress appears in multiple reports as associated with GDV. Dogs described as nervous or stressed are reportedly more likely to have gastric bloat than are dogs described as happy or calm.1,5,8,12,19–21 Studies have either established22 or excluded17 prior splenectomy as a predisposing factor.
The most important risk factors for GDV may be genetic, considering that strong associations with GDV exist for dog breeds,3 families,15,16,20 and sex.17 To date, no specific genes have been identified as associated with this disease. A major component of the genetic link to GDV could involve the shape and size of the thorax, given that large dogs with a deep narrow thoracic cavity appear to be predisposed.1,2,4,16,20 It could be argued that this specific morphology might predispose dogs to volvulus on strictly structural grounds,16 although this argument becomes less compelling for intrabreed and intrafamilial comparisons. Another possible genetic influence on GDV could involve dysfunction of gastrointestinal motility and its hormonal regulation, although whether altered motility is a cause or a result of GDV is unclear.23
We consider an alternative hypothesis for a genetic link to GDV: that variations in the genes of the immune system may predispose dogs to this condition, possibly through modulation of the gastrointestinal microbiome, autoimmune mechanisms, or both. This hypothesis is derived from 3 types of evidence: that GDV may be associated with IBD in dogs, that IBD is associated with dysbiosis of the gastrointestinal microbiome, and that specific variants of certain immune genes play an important role in both microbiome dysbiosis and the etiology of IBD.
In a retrospective study14 of 23 dogs treated for GDV and from which intestinal biopsy specimens were collected during corrective surgery, at least 60% of dogs also had detectable IBD. Although the sample size in that study was small, the findings suggest that IBD might be a predisposing condition for GDV or that the 2 conditions may originate from a common etiologic pathway. Indeed, a study24 of risk factors for IBD in dogs revealed many of the same large, deep-chested dogs at highest risk for GDV.24
Over the last 2 decades, the gastrointestinal microbiome has emerged as a major player in the health of humans and other animals. All mammals, including humans and dogs, host a similar community of microorganisms in their gastrointestinal tract, and this community is known as the gut microbiome.25–34 This highly diverse, complex community includes hundreds of genera and thousands of bacterial species.27,31 Hosts and their microbiomes have evolved together to form a dynamic interdependent ecosystem in which each requires the other for optimal health. The diet, genetics, and health of the host impact the makeup of the gut microbiome, either encouraging or inhibiting growth of different bacterial species.28,30,35–41 Conversely, the microbiome impacts the host in various ways, including competition for nutrients, contribution to digestion by metabolizing dietary materials that the host cannot metabolize, production of metabolites with beneficial or toxic effects to the host, and inhibition of colonization by pathogenic species.38,39,42,43
In humans, perturbations of the gut microbiome have been linked to many gastrointestinal diseases, including colon cancer44 and IBD.45–47 Similarly, in dogs, dysbiosis of the gut microbiome is associated with IBD.48–50 The IBD-associated changes observed in the gut microbiomes of both humans51 and dogs33,52 are remarkably similar.
Two classes of genes play important roles in regulating the gut microbiome by targeting certain microorganisms for immune destruction: antigen-presenting genes of the MHC and innate immunity genes. These genes and the gut microbiome have codeveloped a complex dynamic of selective tolerance that facilitates survival of beneficial bacterial species while defending against pathogenic species.53,54 Mutations and certain polymorphisms in these genes can have negative impacts on health, through failure to detect some pathogens (immunity gaps), failure to tolerate beneficial microorganisms, or elicitation of an inappropriate response against host tissues and cells (autoimmune response).55 These genes can also impact health indirectly by their effect on the gut microbiome.
Mutations or specific combinations of alleles of the aforementioned genes result in changes in the gut microbiome, and these changes can lead to specific gastrointestinal-related diseases.28 For example, mutations in the human genes NOD256–59 and TLR460 have been associated with IBD, particularly with Crohn disease, in humans. In dogs, mutations in the genes TLR4 and TLR561–63 as well as in NOD264–66 have been implicated in IBD. Similarly, mutations of the autophagy-related 16-like 1 gene (ATG16L1), which plays a role in clearance of pathogenic bacteria targeted by intracellular sensor NOD2,67 are also associated with IBD in humans.68–70 Among the MHC genes, a specific human allele of the class IB locus, HLA*B27, is associated with various human diseases, including IBD.46 Although the effect of HLA*B27 on IBD may involve an autoimmune activity against gastrointestinal tissues, this allele also alters the gut microbiome, and this change also appears to play a role in the disease.47,71
The purpose of the study reported here was to explore the genetic component of our hypothesis that variations in the genes of the immune system predispose dogs to GDV. Specifically, we sought to characterize the innate immunity genes NOD2 and TLR5, the autophagy gene Atg16L1, and the most variant MHC genes DLA88 and DRB1 in groups of Great Danes with or without a history of GDV. We hypothesized that specific alleles of these genes would be more common in Great Danes with GDV than in healthy Great Danes. Such alleles may be useful as genetic markers for screening of Great Danes for a predisposition to GDV.
Materials and Methods
Ethics statement
All procedures, authorization forms, information packets, and questionnaires used in this study were submitted to and approved by the Institutional Animal Care and Use Committee at Fred Hutchinson Cancer Research Center (protocol No. 50836). Owner consent was obtained for all participating dogs.
Animals
Great Danes were identified for inclusion in the study through an email network of breeders and owners. All interested owners were sent a questionnaire regarding their dog's age, sex, diet, temperament, coat color, exercise level, medical history, and family history of GDV. Two groups of Great Danes were established on the basis of whether GDV was identified in their history. Dogs chosen for the GDV group had at least 1 episode of GDV that required emergency intervention by a veterinarian. Dogs chosen for the control group had never had severe gastric dilatation or torsion, nor did they have a history of any other major gastrointestinal-related problems. All dogs that had received prophylactic gastropexy were eliminated from the study. No attempt was made to restrict participation on the basis of sex, diet, exercise level, coat color, or age.
To qualify for inclusion, all dogs were required to have complete questionnaire data and confirmation of specific group criteria. Owners were also required to be willing to collect (or have collected by their veterinarian) and send blood and buccal swab samples to the investigators.
Sample collection
Owners of enrolled dogs were provided an EDTA blood tubea preloaded with 0.5 mL of a DNA stabilizerb or a buccal swabc and asked to either have their veterinarian collect a blood sample into the tube or swab the buccal surface of their dog's mouth themselves with the provided swab. For blood sample collection, 1 to 2 mL was obtained by the veterinarian, injected into the provided tube, and stored at room temperature (varied) until shipment. Alternatively, owners swabbed their dog's cheeks with the provided swab, then stored it in a sealed container in a freezer until shipment. All samples were returned cold by overnight express mail to the investigators' laboratory.
DNA processing
Genomic DNA was purified from blood or buccal swab samples on silica-coated magnetic beadsd in accordance with the bead manufacturer's instructions. For buccal swab samples, the reagent volumes were doubled to ensure complete immersion of the swab during lysis. The DNA was eluted in 200 μL of 10mM Tris (pH, 7.9) and quantified by UV absorption.e Final concentrations ranged from 10 to 100 ng/μL and were used, unadjusted, as template for amplification.
Target amplification
Specific target regions of candidate genes associated with GDV were amplified by means of PCR assay.f The 30-μL PCR reactions contained 20 to 100 ng of genomic DNA and primers at 200nM each in a commercial polymerase mix.g Specific primers and conditions for each target were reported separately for each gene. All primers used for amplification and sequencing were summarized (Appendix), and their positions within the target gene were illustrated (Supplementary Figure S1, available at http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.78.8.934).
Genomic target sequences were derived from the Boxer genome made available through an online genome search tool.h The target exons were identified by a search of the Boxer genome with appropriate mRNA sequences listed in the National Institutes of Health genetic sequence database (GenBank)i and identified by accession number as follows: DLA88, exons 2 and 3 (AF100610, AF101487); DRB1, exon 2 (DQ056276); TLR5, exon 1 (NM_001197176); NOD2, exon 3 (NM_001287039); and ATG16L1, exon 8 (XM_005635866). The amplified target sequences were separated by electrophoresis on 1% (DLA88) or 2% (all other targets) agarose gels, the predicted DNA bands were excised and purified on gel extraction columns,j and DNA was eluted in 50 μL of 10mM, Tris (pH, 7.9).
DLA88—Amplification of the polymorphic region of DLA88 (exons 2 and 3) was performed essentially as described elsewhere,72 with minor modifications to primer length and PCR conditions. Typically, primers 2 and 6 (200nM each; Appendix; Supplementary Figure S1) were used to amplify a 1.1-kilobase region containing exons 2 and 3. Cycling conditions included 2 holds at 95°C for 5 minutes, then 72°C for 15 seconds, followed by 35 cycles at 95°C for 15 seconds, 66°C for 30 seconds, and 72°C for 1 minute, and a final hold at 72°C for 2 minutes. When primers 2 and 6 failed to yield product, alternative primers were used to avoid suspected mutations in the primer-binding regions of the template. If none of the primer sets amplified a product, a larger interfering mutation was assumed and the gene was designated CA for cannot amplify.
DRB1—Amplification of the polymorphic region of DRB1 (exon 2) was performed essentially as described.73 Primers 9 and 10 were used routinely, and alternative primers 8 and 11 were used in difficult situations (Appendix; Supplementary Figure S1). Cycling conditions included a hold at 95°C for 5 minutes, followed by 35 cycles at 95°C for 15 seconds, 60°C for 30 seconds, and 72°C for 30 seconds, and a final hold at 72°C for 1 minute.
TLR5—The 5’ region of TLR5, containing the positions of the 2 most important nonsynonymous SNPs linked to IBD in dogs, G22A and C100T61,63 (Supplementary Figure S1), were amplified in a single fragment by use of primers 14 and 15 (Appendix). The following cycling conditions were used: hold for 5 minutes at 95°C, 38 cycles of 95°C for 20 seconds and 68°C for 1 minute, and a final hold at 68°C for 2 minutes.
NOD2—In German Shepherd Dogs, 4 cosegregated nonsynonymous SNPs in exon 3 of NOD2 are associated with IBD.66 Similarly, an R675W mutation in humans74 and a T1949C (F-S) mutation in pigs,75 which are also linked to IBD, occur in this same region (Supplementary Figure S1). A 730-bp region from exon 3 of the NOD2 gene, containing all of these SNP locations, was amplified separately by use of primers 12 and 13 (Appendix) and the same cycling conditions as for TLR5.
Atg16L1—Exon 8 and flanking DNA of the canine Atg16L1 gene, which is homologous to the region in the human Atg16L1 gene harboring the IBD-risk polymorphism T300A,68–70 was amplified with primers 16 and 17 (Appendix; Supplementary Figure S1) by use of the same cycling conditions as for TLR5.
Sequence analysis
Each gel-purified PCR product was sequenced directly to observe ambiguous bases at heterozygous positions. Fluorescent dideoxynucleotide chain termination reactions were performed on a thermocyclerf by use of a commercially available reagent kitk and 2 to 4 μL of gel-purified PCR product. Reaction products were separated and analyzed by capillary electrophoresis.l Each product was sequenced in both directions by use of the amplification primers, and heterozygous base positions were recorded only when observed in both directions. For longer DLA88 templates, internal primers were also used (Appendix; Supplementary Figure S1). In situations in which the 2 alleles were not clearly identifiable by direct sequencing, the PCR products were cloned,m and at least 8 clones were sequenced for confirmation of specific alleles.
Sequences were compared by use of a bioinformatics search tooln to our in-house database of DLA alleles accumulated from online sources.i,o Non-matching alleles were assigned a temporary designation by use of the closest matching allele followed by the symbol A. These new alleles have been submitted to GenBank, and their respective accession numbers were indicated (Supplementary Figures S2–S4, available at http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.78.8.934).
Statistical analysis
The primary goal was to identify whether any specific alleles modified a dog's risk of GDV. Summary statistics were computed to characterize the included dogs. Initial screening for association of specific alleles with GDV was performed by χ2 analysis considering 2 possible variations, allele X or not allele X, without consideration of the number of possible alleles or the confounding effects of other variables such as age, sex, temperament, or exercise level. Values of P < 0.05 combined with ORs > 2 or < 0.5 were considered significant.
To adjust for the strong relationship between attained age and risk of GDV, Cox proportional hazards regression models with GDV as the outcome (dependent) variable were fit to determine HRs and associated 95% CIs for allele and dog data, with age used as the time scale.76 Time began at age 0 and was truncated at the age at which GDV was diagnosed or age at questionnaire completion (censored). The HR can be interpreted as the ratio of the likelihood of developing GDV between dogs of the same age with versus without the characteristic (ie, allele) of interest. A key model assumption, which was tested, was that the hazard remained constant across age, so that it represented a summary measure of relative risk, carefully accounting for absolute risk differences across age. Because dogs sometimes originated from the same family, correlation among members of the same family was accounted for by use of a robust sandwich estimator.77 Each dog was coded as having none, 1, or 2 of each allele to evaluate any possible dose-response effect of alleles on risk of GDV. In addition, this variable was collapsed to form an indicator variable for any presence of a specific allele.
Univariate associations with GDV were evaluated for each candidate allele and dog characteristic (sex, activity level, and temperament). Each allele was also evaluated in a multivariable model adjusted for significant dog characteristics. All tests were 2-tailed. Given that 42 alleles were examined, we considered adjusting P values for multiple comparisons, but use of such an adjustment (0.05/42 = 0.001) would have been overly conservative because of the nature of correlations between alleles. Therefore, unadjusted P values were reported, but we advise caution in interpreting results for which the P values were between 0.05 and 0.01.
Results
Animals
Owners of 178 Great Danes volunteered to enroll their dogs in the study. Of these dogs, 81 met the full criteria for inclusion and were enrolled, including 39 dogs in the GDV group and 42 dogs in the control group. All dogs in the GDV group had a history of at least 1 GDV episode that involved torsion and required immediate veterinary intervention. None of the dogs in the control group had a history of GDV or preventative gastropexy surgery, as reported by owners.
Data on individual dogs (Supplementary Table S1, available at http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.78.8.934) were summarized for each group (Table 1). All dogs were at least 1 year of age. Enrolled dogs included 9 family sets that contributed > 1 dog to the study; 6 of these family sets contributed 2 dogs, and 1 family set each contributed 3, 6, and 8 dogs. Cases of GDV did not appear to cluster within the families. Each family represented in the study by > 2 dogs contributed to both the control and GDV groups, and the potential nonindependence imposed by these relationships was accounted for in the statistical analysis.
Characteristics of Great Danes with a history of at least 1 GDV episode (GDV group) and Great Danes with no such history (control group).
| Characteristic | GDV group (n = 39) | Control group (n = 42) |
|---|---|---|
| Age | ||
| At study entry | 6.4 (2.9) | 4.6 (2.6) |
| After first GDV episode | 4.6 (2.8) | NA |
| Sex | ||
| Female | 26 (67) | 27 (64) |
| Male | 13 (33) | 15 (36) |
| Exercise level | ||
| Low | 10 (26) | 6 (14) |
| Medium | 21 (54) | 19 (45) |
| High | 8 (21) | 17 (40) |
| Temperament | ||
| Nervous | 18 (46) | 6 (14) |
| Typical for the breed | 15 (38) | 24 (57) |
| Calm | 6 (15) | 12 (29) |
Values for age are mean (SD), and values for all other characteristics are number (%) of dogs in the group that had the specified characteristic.
NA = Not applicable.
At the time of questionnaire completion, the mean age of dogs in the GDV group (6.4 years) differed significantly from that of dogs in the control group (4.6 years). However, the mean age of dogs in the GDV group at the time of the first GDV episode (4.6 years) was comparable to the mean age of dogs in the control group.
Genetic variation
DLA88—The polymorphic domain of DLA88 (exons 2 and 3; Supplementary Figure S1) was amplified, sequenced, and compared with the database of known canine alleles. Among the 81 dogs included in the study, 23 distinct alleles were identified (Table 2). These included 16 previously reported alleles and 7 novel variants (Supplementary Figure S3). Five of the novel variants were similar to established alleles (designated in the report with the symbol Δ). Two new variants, designated herein as X and Y, differed by several SNPs from any reported DLA88 alleles. The target DLA88 region of 3 dogs was not detected after several attempts with several primer sets, presumably because of considerable disruption of the sequence in this region. In these situations, the allele was designated as CA (cannot amplify). This was not because of poor DNA quality given that, in all situations, other genes were successfully amplified from the same DNA. The CA allele was included in the analysis on the grounds that it may represent a specific or null allele of DLA88.
Number (%) of the Great Danes in Table 1 carrying specific alleles of polymorphic MHC gene DLA88.
| Allele | No. of alleles | GDV group (n = 39) | Control group (n = 42) |
|---|---|---|---|
| 1001 | 0 | 38 (97) | 39 (93) |
| 1 | 1 (3) | 3 (7) | |
| 101 | 0 | 39 (100) | 40 (95) |
| 1 | 0 (0) | 2 (5) | |
| 1201 | 0 | 23 (59) | 30 (71) |
| 1 | 12 (31) | 9 (21) | |
| 2 | 4 (10) | 3 (7) | |
| 1301 | 0 | 37 (95) | 42 (100) |
| 1 | 2 (5) | 0 (0) | |
| 1501 | 0 | 39 (100) | 41 (98) |
| 1 | 0 (0) | 1 (2) | |
| 01501Δ | 0 | 39 (100) | 41 (98) |
| 2 | 0 (0) | 1 (2) | |
| 1601 | 0 | 38 (97) | 42 (100) |
| 1 | 1 (3) | 0 (0) | |
| 2201 | 0 | 38 (97) | 42 (100) |
| 1 | 1 (3) | 0 (0) | |
| 2501 | 0 | 38 (97) | 40 (95) |
| 1 | 1 (3) | 2 (5) | |
| 02501Δ | 0 | 38 (97) | 42 (100) |
| 1 | 1 (3) | 0 (0) | |
| 402 | 0 | 36 (92) | 40 (95) |
| 1 | 3 (8) | 1 (2) | |
| 2 | 0 (0) | 1 (2) | |
| 0402Δ | 0 | 38 (97) | 37 (88) |
| 1 | 1 (3) | 4 (10) | |
| 2 | 0 (0) | 1 (2) | |
| 4301 | 0 | 35 (90) | 37 (88) |
| 1 | 4 (10) | 4 (10) | |
| 2 | 0 (0) | 1 (2) | |
| 04301Δ | 0 | 29 (74) | 27 (64) |
| 1 | 6 (15) | 9 (21) | |
| 2 | 4 (10) | 6 (14) | |
| 4401 | 0 | 39 (100) | 41 (98) |
| 1 | 0 (0) | 1 (2) | |
| 501 | 0 | 39 (100) | 40 (95) |
| 1 | 0 (0) | 2 (5) | |
| 50901 | 0 | 38 (97) | 40 (95) |
| 1 | 1 (3) | 2 (5) | |
| 5101 | 0 | 25 (64) | 37 (88) |
| 1 | 10 (26) | 5 (12) | |
| 2 | 4 (10) | 0 (0) | |
| 601 | 0 | 33 (85) | 38 (90) |
| 1 | 5 (13) | 2 (5) | |
| 2 | 1 (3) | 2 (5) | |
| CA | 0 | 39 (100) | 39 (93) |
| 2 | 0 (0) | 3 (7) | |
| X | 0 | 38 (97) | 42 (100) |
| 1 | 1 (3) | 0 (0) | |
| Z | 0 | 38 (97) | 42 (100) |
| 2 | 1 (3) | 0 (0) | |
| 201 | 0 | 39 (100) | 41 (98) |
| 1 | 0 (0) | 1 (2) | |
| 2 | 0 (0) | 0 (0) | |
| 1601Δ | 0 | 38 (97) | 42 (100) |
| 1 | 1 (3) | 0 (0) | |
| 2 | 0 (0) | 0 (0) |
CA = Cannot amplify.
Seven novel variants were identified. Five (designated by Δ) were similar to established alleles. Two (designated X and Z) differed by several SNPs from any reported alleles for the gene.
DRB1—The polymorphic domain of DRB1 (exon 2; Supplementary Figure S1) was amplified, sequenced, and compared with the database of known canine alleles. Ten distinct alleles were identified (Table 3), all of which have been previously reported. Variants from 2 dogs were listed as CA because no PCR product was detected after repeated attempts with multiple primer sets.
Number (%) of the Great Danes in Table 1 carrying specific alleles of polymorphic MHC gene DRB1.
| Allele | No. of alleles | GDV group (n = 39) | Control group (n = 42) |
|---|---|---|---|
| 101 | 0 | 17 (44) | 13 (31) |
| 1 | 17 (44) | 15 (36) | |
| 2 | 5 (13) | 14 (33) | |
| 102 | 0 | 38 (97) | 42 (100) |
| 1 | 1 (3) | 0 (0) | |
| 1201 | 0 | 20 (51) | 32 (76) |
| 1 | 13 (33) | 8 (19) | |
| 2 | 6 (15) | 2 (5) | |
| 1301 | 0 | 36 (92) | 42 (100) |
| 1 | 2 (5) | 0 (0) | |
| 2 | 1 (3) | 0 (0) | |
| 1501 | 0 | 36 (92) | 39 (93) |
| 1 | 3 (8) | 3 (7) | |
| 1502 | 0 | 36 (92) | 39 (93) |
| 1 | 3 (8) | 3 (7) | |
| 1601 | 0 | 33 (85) | 38 (90) |
| 1 | 6 (15) | 4 (10) | |
| 1801 | 0 | 38 (97) | 41 (98) |
| 1 | 1 (3) | 1 (2) | |
| 2201 | 0 | 39 (100) | 41 (98) |
| 1 | 0 (0) | 1 (2) | |
| 601 | 0 | 33 (85) | 29 (69) |
| 1 | 6 (15) | 11 (26) | |
| 2 | 0 (0) | 2 (5) | |
| CA | 0 | 38 (97) | 41 (98) |
| 2 | 1 (3) | 1 (2) |
CA = Cannot amplify.
TLR5—A 5’ region of TLR5 (Supplementary Figure S1) containing the amino-terminal 40 codons of the protein as well as a significant IBD-associated SNP site in German Shepherd Dogs61,63 was amplified and sequenced. Four DNA variants were identified among the study dogs, and these encoded just 2 distinct variants at the amino acid level (Table 4; Supplementary Figure S3). Variant A corresponded to the canonical form of TLR5 and represented 84% of the alleles observed. Variant B, which contained a T8A substitution, corresponded to the IBD-associated SNP described by Kathrani et al.61,63 These variants were referred to henceforth as alleles TLR5*A and TLR5*B, respectively.
Number (%) of the Great Danes in Table 1 carrying specific alleles of canine innate immunity genes NOD2 and TLR5.
| Allele | No. of alleles | GDV group (n = 39) | Control group (n = 42) |
|---|---|---|---|
| TLR5*A | 0 | 1 (3) | 1 (2) |
| 1 | 15 (38) | 6 (14) | |
| 2 | 23 (59) | 35 (83) | |
| TLR5*B | 0 | 24 (62) | 36 (86) |
| 1 | 15 (38) | 6 (14) | |
| TLR5*CA | 0 | 38 (97) | 41 (98) |
| 2 | 1 (3) | 1 (2) | |
| NOD2*A | 0 | 24 (62) | 24 (57) |
| 1 | 11 (28) | 15 (36) | |
| 2 | 4 (10) | 3 (7) | |
| NOD2*B | 0 | 9 (23) | 6 (14) |
| 1 | 14 (36) | 21 (50) | |
| 2 | 16 (41) | 15 (36) | |
| NOD2*C | 0 | 30 (77) | 34 (81) |
| 1 | 7 (18) | 8 (19) | |
| 2 | 2 (5) | 0 (0) |
NOD2—A region of exon 3 in NOD2 (Supplementary Figure S1) was amplified, containing the positions of 4 SNPs associated with IBD in German Shepherd Dogs66 as well as positions homologous to a human SNP associated with Crohn disease74 and a porcine loss-of-function mutation.75 All 4 canine SNPs were observed, generating 3 nonsynonymous variants within the study group (Table 4; Supplementary Figure S4).
Atg16L1—Exon 8 of Atg16L1 (Supplementary Figure S1), containing the homologous position of a single human T300A SNP associated with IBD,68–70 was amplified and sequenced. This region did not vary among the Great Danes tested, and all dogs were homozygous for the T300 form. Although a role for ATG16L1 in GDV could not be ruled out, the target SNP did not appear to be important. Therefore, Atg16L1 was not analyzed further.
Unadjusted associations of genetic and nongenetic variants with GDV
Univariate analysis of simple associations of genetic and nongenetic variants with GDV in Great Danes revealed 3 genetic variants with potential associations: DLA88*5101 (OR,4.74; P = 0.004; 95% CI, 1.62 to 13.84), DRB1*1201 (OR, 2.83; P = 0.02; 95% CI, 1.19 to 6.75), and TLR5*B (OR, 3.10; P = 0.007; 95% CI, 1.37 to 7.00; Table 5). Two DRB1 alleles, 101 and 601, had a slightly protective association (OR values of 0.68 and 0.43, respectively) but did not meet the criteria for significance (P < 0.05 and OR < 0.50). None of the NOD2 alleles were associated with GDV. Stratification of the analysis by dog family had no effect on these findings. When unrelated dogs, or the combined families of related dogs, were analyzed separately, the results were the same: all 3 GDV-associated alleles retained an OR ≥ 2 (data not shown).
Selected results of univariate (χ2) analysis of associations between GDV and genetic variants or nervous temperament for the Great Danes in Table 1.
| Prevalence (%) | Frequency (%)† | |||||||
|---|---|---|---|---|---|---|---|---|
| Variable | GDV | Control | Total | GDV | Control | OR | 95% CI | P value‡ |
| DLA88*5101 | 18 | 5 | 23 | 23 | 6 | 4.74 | 1.62–13.84 | 0.004 |
| DRB1*1201 | 25 | 12 | 37 | 32 | 14 | 2.83 | 1.19–6.75 | 0.02 |
| DRB1*101 | 27 | 43 | 70 | 35 | 51 | 0.05 | 0.25–1.03 | 0.06 |
| DRB1*601 | 6 | 15 | 21 | 8 | 18 | 0.38 | 0.15–0.98 | 0.05 |
| TLR5*B | 18 | 7 | 25 | 19 | 7 | 3.10 | 1.37–7.00 | 0.007 |
| Nervous temperament | 18 | 6 | 24 | 47 | 14 | 5.14 | 1.80–14.71 | 0.002 |
In prevalence calculations, homozygous alleles were counted twice. Values of P < 0.05 were considered significant.
Frequency = Percentage of dogs in group with indicated variant.
The P values were determined on the basis of a 2-variant assumption (allele X or not allele X).
Similar analysis of nongenetic data acquired from questionnaires revealed no significant association between GDV and familial incidence of GDV or dog sex, activity level, or coat color (data not shown). However, temperament and age varied between GDV and control groups. When owners were asked if they considered their dog's temperament to be nervous, typical for the breed, or calm, nervousness was highly associated with GDV (OR, 5.14; P = 0.002; 95% CI, 1.80 to 14.71; Table 5).
Age-adjusted association of genetic and nongenetic variants with GDV
A near-linear accumulation of first GDV episodes with age was identified in Great Danes with GDV, indicating a fairly constant probability of a first event at any given age (Figure 1). Therefore, age at questionnaire completion may have been an important contributor to the probability that a predisposed dog would have had a GDV episode. To control for this possibility, Cox proportional hazards analysis was performed, with dog age used as the time scale, thereby ensuring that comparisons were made between same-age dogs. This model was also designed to account for family relationships that existed among some of the dogs.
Accumulative incidence of first GDV episode with age in 39 Great Danes. A steady increase is apparent in the incidence of GDV with age, and there is no evidence for any age-specific risk.
Citation: American Journal of Veterinary Research 78, 8; 10.2460/ajvr.78.8.934
After the aforementioned adjustments, 3 of the 4 risk factors identified in the unadjusted analysis (nervous temperament, DRB1*1201, and TLR5*B) retained significant association with GDV (P < 0.05; HR > 2.0; Table 6), although the P value for DRB1*1201 exceeded 0.01 (a more conservative approach to significance testing).
Results of Cox proportional hazards analysis of associations between GDV and genetic or nongenetic variants for the Great Danes in Table 1, controlling for age and dog family.
| Variable | HR | 95% CI | P value |
|---|---|---|---|
| Nervous (vs typical for the breed) | 2.32 | 1.20–4.46 | 0.01 |
| Calm (vs typical for the breed) | 1.08 | 0.40–2.88 | 0.88 |
| DLA88*50901 | 0.39 | 0.21–0.72 | 0.003 |
| DLA88*5101 | 1.39 | 0.70–2.74 | 0.35 |
| DRB1*101 | 0.47 | 0.26–0.83 | 0.01 |
| DRB1*1201 | 2.20 | 1.07–4.51 | 0.03 |
| DRB1*1502 | 4.4 | 1.16–16.75 | 0.03 |
| DRB1*601 | 0.50 | 0.26–0.97 | 0.04 |
| TLR5*B | 2.64 | 1.44–4.84 | 0.002 |
For genes, the referent category is no such allele. A separate model was created for each gene and allele combination.
Data shown are restricted significant (P < 0.05) associations or loss of associations previously identified by χ2 analysis (Table 5).
On the other hand, DLA88*1501 lost significance as a risk factor in this analysis. This particular allele was identified in dogs of both groups that were older than their cohorts by a mean of 0.7 years (GDV group) or 0.5 years (control group). Alleles DRB1*00101 and DRB1*00601 both retained their protective association with GDV after age adjustment, although the P value for DRB1*601 exceeded 0.01.
Two alleles acquired significance in hazard models after age adjustment: DLA88*50901 had a protective association with GDV (HR, 0.39; P = 0.003), and DRB1*1502 was associated with an increased risk of GDV (HR, 4.40; P = 0.03). However, because few dogs had either of these alleles, the importance of this finding remains to be verified with larger numbers and should be interpreted with caution.
A dose-response effect of these alleles was also examined when numbers were sufficiently large (Table 7). The only variant with such an effect was the protective DRB1*101 allele, by which the HR decreased from 0.61 for heterozygotes to 0.24 for homozygotes. For DRB1*1201, the HR decreased somewhat for homozygotes versus heterozygotes, but values were not significant. Homozygous status for TLR5*B was not observed. Age adjustment resulted in a significant association of NOD2*C with GDV for homozygotes; however, because only 2 dogs with GDV had this allele, this association should be interpreted cautiously.
Results of Cox proportional hazards modeling of dose-response effects of alleles of various genes on the hazard of GDV for the Great Danes in Table 1, controlling for age and dog family.
| Allele | Copy No. | HR | 95% CI | P value |
|---|---|---|---|---|
| DLA88*5101 | 0 | Referent | — | — |
| 1 | 1.25 | 0.58–2.71 | 0.57 | |
| 2 | 1.88 | 0.86–4.12 | 0.11 | |
| DRB1*101 | 0 | Referent | — | — |
| 1 | 0.61 | 0.35–1.09 | 0.09 | |
| 2 | 0.24 | 0.09–0.67 | 0.006 | |
| DRB1*1201 | 0 | Referent | — | — |
| 1 | 2.36 | 1.02–5.50 | 0.046 | |
| 2 | 1.93 | 0.84–4.47 | 0.12 | |
| TLR5*B | 0 | Referent | — | — |
| 1 | 2.64 | 1.44–4.84 | 0.002 | |
| 2 | — | — | — | |
| NOD2*C | 0 | Referent | — | — |
| 1 | 0.80 | 0.34–1.88 | 0.60 | |
| 2 | 4.38 | 1.47–13.03 | 0.008 |
— = Not applicable.
See Table 6 for remainder of key.
Temperament-adjusted association of genetic variants with GDV
To determine whether the significant differences identified through univariate modeling were confounded by other factors, multivariable models were also fit to evaluate adjusted associations for alleles. First, nonallele variables associated with GDV were identified, and in a multivariable model including all available variables, only temperament had a significant association, with nervous dogs having almost 3 times the risk of developing GDV as dogs with a more typical temperament for the breed (HR, 2.70; P = 0.002; 95% CI, 1.43 to 5.08; Table 8).
Results of multivariable Cox proportional hazards analysis of associations between GDV and nongenetic variants for the Great Danes in Table 1, controlling for age and dog family.
| Variable | HR | 95% CI | P value |
|---|---|---|---|
| Male (vs female) | 1.00 | 0.50–1.98 | 0.99 |
| Exercise level | |||
| Low (vs high) | 1.09 | 0.50–2.37 | 0.83 |
| Medium (vs high) | 1.57 | 0.75–3.28 | 0.23 |
| Temperament | |||
| Nervous (vs typical for the breed) | 2.70 | 1.43–5.08 | 0.002 |
| Calm (vs typical for the breed) | 1.20 | 0.46–3.15 | 0.71 |
See Table 1 for key.
We proceeded to fit adjusted multivariable models for each allele, adjusting for temperament (Table 9). For all models shown, the association of temperament with GDV was similar to that previously identified (HRs ranged from 2.1 to 3.2) and remained significant. For brevity, only the allele effects are shown, each adjusted for temperament. Only alleles with marginal (P < 0.10) or significant associations are reported, although others were tested.
Selected results of multivariable Cox proportional hazards analysis of associations between GDV and genetic variants for the Great Danes in Table 1, controlling for temperament, age, and dog family.
| Allele | HR | 95% CI | P value |
|---|---|---|---|
| Allele present (yes vs no) | |||
| DLA88*4301Δ | 0.43 | 0.21–0.88 | 0.02 |
| DLA88*50901 | 0.52 | 0.25–1.08 | 0.08 |
| DRB1*101 | 0.45 | 0.26–0.77 | 0.004 |
| DRB1*1201 | 2.44 | 1.22–4.89 | 0.01 |
| DRB1*1502 | 6.73 | 1.83–24.70 | 0.004 |
| TLR5*B | 2.43 | 1.31–4.50 | 0.005 |
| No. of alleles present DLA88*4301Δ | |||
| 0 | Referent | — | — |
| 1 | 0.65 | 0.30–1.42 | 0.28 |
| 2 | 0.27 | 0.11–0.64 | 0.003 |
| DLA88*5101 | |||
| 0 | Referent | — | — |
| 1 | 1.51 | 0.67–3.38 | 0.32 |
| 2 | 2.96 | 1.29–6.79 | 0.01 |
| DRB1*101 | |||
| 0 | Referent | — | — |
| 1 | 0.62 | 0.35–1.12 | 0.11 |
| 2 | 0.21 | 0.08–0.54 | 0.001 |
| DRB1*1201 | |||
| 0 | Referent | — | — |
| 1 | 2.53 | 1.14–5.57 | 0.02 |
| 2 | 2.29 | 0.96–5.45 | 0.06 |
| TLR5*B | |||
| 0 | Referent | — | — |
| 1 | 2.43 | 1.31–4.50 | 0.005 |
Three DLA88 alleles were significantly associated with GDV. Allele 5101 reappeared as increasing the risk of GDV when present in 2 copies, and 50901 had a marginal (P = 0.08) protective effect but was identified in only 3 dogs. Allele 4301Δ appeared de novo as significantly protective against GDV.
Three DRB1 alleles had significant associations with GDV: 01201 and 1502 increased the risk of GDV, whereas 101 had a protective effect. The newly emerged risk association of allele 1502, not identified in the unadjusted or univariate models, appeared to be strong in this model (HR, 6.73; P = 0.004). However, this estimate was based on small, evenly distributed observations (3 dogs/group) and should therefore be considered tenuous. As in both previous models, the TLR5*B variant continued to have a significant (P = 0.005) association with GDV after adjustment for temperament (HR, 2.43).
Discussion
The hypothesis tested in the present study was that canine innate immunity and MHC genes might play a role in predisposing Great Danes to GDV. Findings indicated that specific variants of DLA88, DRB1, and TLR5 were associated with an increased risk of this condition. These alleles are heretofore referred to as risk alleles, although passive genetic linkage to the responsible genes could not be ruled out. Two candidate risk alleles, DRB1*1201 and TLR5*B, that were identified in the unadjusted models remained as significant risk alleles after adjustments for dog age and temperament. Both retained HRs > 2.0 and P values < 0.05 through all 3 test strategies, indicating an association of these alleles with GDV that was independent of age and temperament. A third candidate risk allele, DLA88*5101, lost significance in the age-adjusted model but reappeared as a risk allele after adjustment for temperament, suggesting a more complex relationship between this gene and GDV.
These 3 alleles (DRB1*1201, TLR5*B, and DLA88*5101) were fairly abundant in the Great Danes of the present study, particularly those with GDV, accounting for 24%, 34%, and 24%, respectively, of the alleles observed at their gene locus within the GDV group. In fact, 77% of dogs in the GDV group had at least 1 of these alleles, compared with 36% of dogs in the control group. The frequencies of these alleles were sufficiently high to posit that they could contribute to a considerable percentage of GDV episodes in Great Danes.
Several relatively uncommon alleles had no significant association with GDV in the initial univariate analyses but were associated with an increased risk (DRB1*1502 and NOD2*C) or protective relationship (DLA88-50901) after age and temperament adjustments. None of these alleles constituted > 10% of the total alleles observed in either group. The associations of these alleles with GDV must be viewed with some skepticism because of the small numbers of observations of each allele in the study groups and the statistical adjustments that restricted the sizes of the groups compared in each model. Furthermore, uncommon alleles could only have a marginal effect on a population that has 37% lifetime risk of GDV. If consideration were restricted to those alleles that occurred at frequencies > 10% in dogs with GDV, and with significant associations in at least 2 of the 3 analysis models described here, 3 genetic risk factors with potential for high impact would remain: DLA88*5101, DRB1*1201, and TLR5*B.
When considered independently, each of the identified risk alleles was associated with an approximately 2-fold greater risk of GDV in Great Danes. When considered as a group, these 3 risk alleles were associated with a 3-fold increased risk of GDV. The proportion of dogs with ≥ 1 GDV episode increased from 20% for dogs with no risk alleles to 62% for dogs with ≥ 1 risk allele. Little evidence was found for a significant dose-response effect of these risk alleles, whether for a given gene or among the 3 genes. These genetic associations do not suggest that these genes play a direct role in GDV. However, the finding that 3 immune-related genes were associated with the disease supports an immune component to GDV. Regardless of the specific roles that these risk genes may play in GDV, the genes may be useful for genetic screening of Great Danes for a predisposition to the disease.
The statistical power of the study reported here was limited by the size and structure of the enrolled groups. By limiting the study to Great Danes, no direct conclusions could be drawn regarding other breeds. The less prevalent risk alleles would require larger study groups to determine their impact, if any, on GDV. Furthermore, this cross-sectional study involved only dogs that survived to the time of study conclusion, and indeed, for dogs in the GDV group, only those that survived their GDV episode. As such, it is possible that dogs with the most severe cases of GDV were not included.
Inflammatory bowel disease and GDV may be associated in dogs, and findings of the present study confirmed that at least 2 genes were associated with both disorders. One reason for focusing on immune-related genes was the report14 of co-occurrence of GDV with IBD, a disease with an etiology that clearly involves dysregulation of the immune system. Specific alleles or haplotypes of the DLA genes have been associated with several disorders in dogs, including exocrine pancreatic insufficiency,72,78 hypoadrenocorticism,79 diabetes mellitus,73 polymyositis,80 lymphocytic thyroiditis,78 symmetrical lupoid onychodystrophy,81 chronic superficial keratitis,82 anal furunculosis,83 and systemic lupus erythematosus–related disease complex.84 Similarly, mutations in TLR5 have been associated specifically with IBD in both dogs63 and humans.60 Furthermore, in the present study the same TLR5 allele linked to IBD in dogs was also linked to GDV. Humans with the MHC class I allele, B27,46 or the class II allele, HLADRB*01:03,85 are predisposed to IBD. Although similar associations of MHC genes with IBD in dogs have not yet been established, our findings indicated that a class I MHC antigen, DLA88*5101, and a class II antigen, DRB1*1201, were associated with GDV. These relationships support the argument that IBD and GDV are related in dogs and that the 2 diseases may have common genetic roots.
Great Danes were used in the present study to leverage the high incidence of GDV in this breed and to reduce the additional variables inherent to a multibreed study. Therefore, it remains unclear whether risk alleles identified for Great Danes would also apply to other breeds. Whereas it would be useful, in this regard, to compare the relative frequencies of these risk alleles with GDV incidence in other breeds, such information is still incomplete. From findings of the present study, the overall frequency of risk alleles in Great Danes can be calculated as 31% for DRB1*1201, 21% for DLA88*5101, and 23% for TLR5*B on the basis of the frequency in the GDV and control groups and the published frequency of GDV in Great Danes (37%6).
In contrast, ongoing DLA typing across multiple breeds by our research group has yielded much lower frequencies of 7% for DRB1*1202 and 4% for DLA88*5101 among approximately 1,000 tested dogs (unpublished data). We have also analyzed the DLA typing data regarding > 1,000 Beagles and Beagle-mix dogs evaluated at the Fred Hutchinson Cancer Research Center. The prevalence of DRB1*1201 in 1,003 tested dogs was 1.3%, and that of DLA88*5101 among 338 tested dogs was 0%, with both values far less than in Great Danes. This is consistent with the low incidence of GDV in Beagles.3 As reported by other investigators,15 the incidence of GDV in a mixed dog population (approx 3.7%) was less than a tenth that in Great Danes. Thus, the high incidence of GDV in Great Danes generally corresponds with a higher than average incidence of the DLA risk alleles.
Assessment of this relationship in specific breeds is challenged by the lack of statistically robust breed-specific information on both allele frequency and GDV incidence. However, it is clear that the relationship does not hold for at least 1 breed (ie, Golden Retriever). Golden Retrievers have a high frequency of DRB1*1201 (25%86) and DLA88*5101 (47%87), yet this breed is considered at low risk for GDV.3 Therefore, the risk effect of the DLA alleles reported here may depend on a genetic context that varies among breeds. No breed-specific data for TLR5 are available that are germane to this possibility. It will be important to examine these genetic markers in several breeds before making any generalizations as to their linkage or role in GDV.
In the study reported here, focus was placed on MHC and innate immunity genes because of a working hypothesis that these genes might influence the gastrointestinal tract through regulation of the microbiome. It remains to be determined whether the risk of GDV associated with these genes corresponds with changes in the microbiome, as per our hypothesis, or through some other mechanism.
Acknowledgments
The DLA typing was performed by the Canine Large Animal Core of the Core Center of Excellence in Hematology at Fred Hutchinson Cancer Research Center (National Institute of Diabetes and Digestive and Kidney Disease; grant No. P30 DK56465).
Supported by grants from Robert and Rebecca Pohlad and the Van Sloan Foundation. Dr. Venkataraman was also supported by a grant from the American Kennel Club Canine Health Foundation.
The authors thank Gretchen Johnson, Deborah Higginbotham, Gil Guday, and Penelope Blossom for technical assistance and Helen Crawford for assistance with manuscript and figure preparation.
ABBREVIATIONS
| CI | Confidence interval |
| DLA | Dog leukocyte antigen |
| GDV | Gastric dilatation-volvulus |
| HR | Hazard ratio |
| IBD | Inflammatory bowel disease |
| MHC | Major histocompatibility complex |
| SNP | Single base polymorphism |
Footnotes
Vacutainer K2-EDTA blood tube, Becton, Dickinson and Co, Franklin Lakes, NJ.
DNAGard, Biometrica Systems Inc, San Diego, Calif.
Isohelix SK-2S, Cell Products Ltd, Kent, England.
BioSprint 96 blood kit, Qiagen, Valencia, Calif.
NanoDrop spectrophotometer, Fisher Scientific, Pittsburgh, Pa.
Gene Amp 9700 thermocycler, Applied Biosystems/Thermo-Fisher Scientific, Grand Island, NY.
MyFi high-fidelity polymerase mix, Bioline USA Inc, Taunton, Mass.
BLAT search genome, University of California–Santa Cruz Genome Informatics Group, Santa Cruz, Calif. Available at: genome.ucsc.edu/cgi-bin/hgBlat. Accessed Jun 30, 2016.
GenBank, National Institutes of Health, Bethesda, Md. Available at: www.ncbi.nlm.nih.gov/genbank/. Accessed Jun 30, 2016.
QIAquick gel extraction column, Qiagen, Valencia, Calif.
BigDye Terminator cycle sequencing reagent, Applied Biosystems, Waltham, Mass.
ABI 3730xl DNA analyzer, Applied Biosystems, Waltham, Mass.
TOPO TA PCR cloning kit, Invitrogen, Carlsbad, Calif.
BLAST, National Center for Biotechnology Information, National Institutes of Health, Bethesda, Md. Available at: blast. ncbi.nlm.nih.gov/. Accessed Jun 30, 2016.
IPD-MHC database, EMBL-EBI, Hinxton, Cambridgeshire, England. Available at: www.ebi.ac.uk/ipd/mhc/dla. Accessed Jun 30, 2016.
References
1. Broome CJ, Walsh VP. Gastric dilatation-volvulus in dogs. N Z Vet J 2003; 51:275–283.
2. Monnet E. Gastric dilatation-volvulus syndrome in dogs. Vet Clin North Am Small Anim Pract 2003; 33:987–1005 (vi.).
3. Evans KM, Adams VJ. Mortality and morbidity due to gastric dilatation-volvulus syndrome in pedigree dogs in the UK. J Small Anim Pract 2010; 51:376–381.
4. Hendriks MM, Hill KE, Cogger N, et al. A retrospective study of gastric dilatation and gastric dilatation and volvulus in working farm dogs in New Zealand. N Z Vet J 2012; 60:165–170.
5. Glickman LT, Glickman NW, Schellenberg DB, et al. Incidence of and breed-related risk factors for gastric dilatation-volvulus in dogs. J Am Vet Med Assoc 2000; 216:40–45.
6. Ward MP, Patronek GJ, Glickman LT. Benefits of prophylactic gastropexy for dogs at risk of gastric dilatation-volvulus. Prev Vet Med 2003; 60:319–329.
7. Van Kruiningen HJ, Gargamelli C, Havier J, et al. Stomach gas analyses in canine acute gastric dilatation with volvulus. J Vet Intern Med 2013; 27:1260–1261.
8. Bell JS. Inherited and predisposing factors in the development of gastric dilatation volvulus in dogs. Top Companion Anim Med 2014; 29:60–63.
9. de Battisti A, Toscano MJ, Formaggini L. Gastric foreign body as a risk factor for gastric dilatation and volvulus in dogs. J Am Vet Med Assoc 2012; 241:1190–1193.
10. Raghavan M, Glickman N, McCabe G, et al. Diet-related risk factors for gastric dilatation-volvulus in dogs of high-risk breeds. J Am Anim Hosp Assoc 2004; 40:192–203.
11. Raghavan M, Glickman NW, Glickman LT. The effect of ingredients in dry dog foods on the risk of gastric dilatation-volvulus in dogs. J Am Anim Hosp Assoc 2006; 42:28–36.
12. Glickman LT, Glickman NW, Schellenberg DB, et al. Multiple risk factors for the gastric dilatation-volvulus syndrome in dogs: a practitioner/owner case-control study. J Am Anim Hosp Assoc 1997; 33:197–204.
13. Lauten SD. Nutritional risks to large-breed dogs: from weaning to the geriatric years. Vet Clin North Am Small Anim Pract 2006; 36:1345–1359 (viii.).
14. Braun L, Lester S, Kuzma AB, et al. Gastric dilatation-volvulus in the dog with histological evidence of preexisting inflammatory bowel disease: a retrospective study of 23 cases. J Am Anim Hosp Assoc 1996; 32:287–290.
15. Glickman LT, Glickman NW, Perez CM, et al. Analysis of risk factors for gastric dilatation and dilatation-volvulus in dogs. J Am Vet Med Assoc 1994; 204:1465–1471.
16. Schellenberg D, Yi Q, Glickman NW, et al. Influence of thoracic conformation and genetics on the risk of gastric dilatation-volvulus in Irish setters. J Am Anim Hosp Assoc 1998; 34:64–73.
17. Grange AM, Clough W, Casale SA. Evaluation of splenectomy as a risk factor for gastric dilatation-volvulus. J Am Vet Med Assoc 2012; 241:461–466.
18. Theyse LF, van de Brom WE, van Sluijs FJ. Small size of food particles and age as risk factors for gastric dilatation volvulus in Great Danes. Vet Rec 1998; 143:48–50.
19. Elwood CM. Risk factors for gastric dilatation in Irish Setter dogs. J Small Anim Pract 1998; 39:185–190.
20. Glickman LT, Glickman NW, Schellenberg DB, et al. Nondietary risk factors for gastric dilatation-volvulus in large and giant breed dogs. J Am Vet Med Assoc 2000; 217:1492–1499.
21. Pipan M, Brown DC, Battaglia CL, et al. An Internet-based survey of risk factors for surgical gastric dilatation-volvulus in dogs. J Am Vet Med Assoc 2012; 240:1456–1462.
22. Sartor AJ, Bentley AM, Brown DC. Association between previous splenectomy and gastric dilatation-volvulus in dogs: 453 cases (2004–2009). J Am Vet Med Assoc 2013; 242:1381–1384.
23. Gazzola KM, Nelson LL. The relationship between gastrointestinal motility and gastric dilatation-volvulus in dogs. Top Companion Anim Med 2014; 29:64–66.
24. Kathrani A, Werling D, Allenspach K. Canine breeds at high risk of developing inflammatory bowel disease in the southeastern UK. Vet Rec 2011; 169:635.
25. Suchodolski JS, Xenoulis PG, Paddock CG, et al. Molecular analysis of the bacterial microbiota in duodenal biopsies from dogs with idiopathic inflammatory bowel disease. Vet Microbiol 2010; 142:394–400.
26. Arumugam M, Raes J, Pelletier E, et al. Enterotypes of the human gut microbiome (Erratum published in Nature 2011;474:666). Nature 2011; 473:174–180.
27. Handl S, Dowd SE, Garcia-Mazcorro JF, et al. Massive parallel 16S rRNA gene pyrosequencing reveals highly diverse fecal bacterial and fungal communities in healthy dogs and cats. FEMS Microbiol Ecol 2011; 76:301–310.
28. Spor A, Koren O, Ley R. Unravelling the effects of the environment and host genotype on the gut microbiome. Nat Rev Microbiol 2011; 9:279–290.
29. Swanson KS, Dowd SE, Suchodolski JS, et al. Phylogenetic and gene-centric metagenomics of the canine intestinal microbiome reveals similarities with humans and mice. ISME J 2011; 5:639–649.
30. Wu GD, Chen J, Hoffmann C, et al. Linking long-term dietary patterns with gut microbial enterotypes. Science 2011; 334:105–108.
31. Garcia-Mazcorro JF, Dowd SE, Poulsen J, et al. Abundance and short-term temporal variability of fecal microbiota in healthy dogs. Microbiologyopen 2012; 1:340–347.
32. Hooda S, Minamoto Y, Suchodolski JS, et al. Current state of knowledge: the canine gastrointestinal microbiome. Anim Health Res Rev 2012; 13:78–88.
33. Suchodolski JS, Markel ME, Garcia-Mazcorro JF, et al. The fecal microbiome in dogs with acute diarrhea and idiopathic inflammatory bowel disease. PLoS One 2012; 7:e51907.
34. Schloissnig S, Arumugam M, Sunagawa S, et al. Genomic variation landscape of the human gut microbiome. Nature 2013; 493:45–50.
35. Li F, Hullar MA, Schwarz Y, et al. Human gut bacterial communities are altered by addition of cruciferous vegetables to a controlled fruit- and vegetable-free diet. J Nutr 2009; 139:1685–1691.
36. Beloshapka AN, Dowd SE, Suchodolski JS, et al. Fecal microbial communities of healthy adult dogs fed raw meat-based diets with or without inulin or yeast cell wall extracts as assessed by 454 pyrosequencing. FEMS Microbiol Ecol 2013; 84:532–541.
37. Klosterbuer AS, Hullar MA, Li F, et al. Gastrointestinal effects of resistant starch, soluble maize fibre and pullulan in healthy adults. Br J Nutr 2013; 110:1068–1074.
38. Koeth RA, Wang Z, Levison BS, et al. Intestinal microbiota metabolism of l-carnitine, a nutrient in red meat, promotes atherosclerosis. Nat Med 2013; 19:576–585.
39. Mulle JG, Sharp WG, Cubells JF. The gut microbiome: a new frontier in autism research. Curr Psychiatry Rep 2013; 15:337.
40. Panasevich MR, Kerr KR, Dilger RN, et al. Modulation of the faecal microbiome of healthy adult dogs by inclusion of potato fibre in the diet. Br J Nutr 2015; 113:125–133.
41. Panasevich MR, Kerr KR, Serao MC, et al. Evaluation of soluble corn fiber on chemical composition and nitrogen-corrected true metabolizable energy and its effects on in vitro fermentation and in vivo responses in dogs. J Anim Sci 2015; 93:2191–2200.
42. Kanauchi O, Matsumoto Y, Matsumura M, et al. The beneficial effects of microflora, especially obligate anaerobes, and their products on the colonic environment in inflammatory bowel disease. Curr Pharm Des 2005; 11:1047–1053.
43. Li F, Hullar MA, Beresford SA, et al. Variation of glucoraphanin metabolism in vivo and ex vivo by human gut bacteria. Br J Nutr 2011; 106:408–416.
44. Flint HJ, Duncan SH, Scott KP, et al. Interactions and competition within the microbial community of the human colon: links between diet and health. Environ Microbiol 2007; 9:1101–1111.
45. Friswell M, Campbell B, Rhodes J. The role of bacteria in the pathogenesis of inflammatory bowel disease. Gut Liver 2010; 4:295–306.
46. Rashid T, Ebringer A. Gut-mediated and HLA-B27-associated arthritis: an emphasis on ankylosing spondylitis and Crohn's disease with a proposal for the use of new treatment. Discov Med 2011; 12:187–194.
47. Rosenbaum JT, Davey MP. Time for a gut check: evidence for the hypothesis that HLA-B27 predisposes to ankylosing spondylitis by altering the microbiome. Arthritis Rheum 2011; 63:3195–3198.
48. Suchodolski JS, Dowd SE, Wilke V, et al. 16S rRNA gene pyrosequencing reveals bacterial dysbiosis in the duodenum of dogs with idiopathic inflammatory bowel disease. PLoS One 2012; 7:e39333.
49. Honneffer JB, Minamoto Y, Suchodolski JS. Microbiota alterations in acute and chronic gastrointestinal inflammation of cats and dogs. World J Gastroenterol 2014; 20:16489–16497.
50. Junginger J, Lemensieck F, Moore PF, et al. Canine gut dendritic cells in the steady state and in inflammatory bowel disease. Innate Immun 2014; 20:145–160.
51. Hotte NS, Salim SY, Tso RH, et al. Patients with inflammatory bowel disease exhibit dysregulated responses to microbial DNA. PLoS One 2012; 7:e37932.
52. Xenoulis PG, Palculict B, Allenspach K, et al. Molecular-phylogenetic characterization of microbial communities imbalances in the small intestine of dogs with inflammatory bowel disease. FEMS Microbiol Ecol 2008; 66:579–589.
53. Round JL, Mazmanian SK. The gut microbiota shapes intestinal immune responses during health and disease. Nat Rev Immunol 2009; 9:313–323.
54. Round JL, O'Connell RM, Mazmanian SK. Coordination of tolerogenic immune responses by the commensal microbiota. J Autoimmun 2010; 34:J220–J225.
55. Murphy K. Part V: the immune system in health and disease. In: Janeway's immunobiology. 8th ed. New York: Garland Science (Taylor & Francis Group), 2012; 509–668.
56. Russell RK, Drummond HE, Nimmo EE, et al. Genotype-phenotype analysis in childhood-onset Crohn's disease: NOD2/CARD15 variants consistently predict phenotypic characteristics of severe disease. Inflamm Bowel Dis 2005; 11:955–964.
57. Yazdanyar S, Weischer M, Nordestgaard BG. Genotyping for NOD2 genetic variants and Crohn disease: a metaanalysis. Clin Chem 2009; 55:1950–1957.
58. Barreau F, Madre C, Meinzer U, et al. Nod2 regulates the host response towards microflora by modulating T cell function and epithelial permeability in mouse Peyer's patches. Gut 2010; 59:207–217.
59. Yazdanyar S, Nordestgaard BG. NOD2/CARD15 genotype and common gastrointestinal diseases in 43,600 individuals. J Intern Med 2010; 267:228–236.
60. Cario E. Toll-like receptors in inflammatory bowel diseases: a decade later. Inflamm Bowel Dis 2010; 16:1583–1597.
61. Kathrani A, House A, Catchpole B, et al. Polymorphisms in the TLR4 and TLR5 gene are significantly associated with inflammatory bowel disease in German Shepherd Dogs. PLoS One 2010; 5:e15740.
62. Kathrani A, House A, Catchpole B, et al. Breed-independent toll-like receptor 5 polymorphisms show association with canine inflammatory bowel disease. Tissue Antigens 2011; 78:94–101.
63. Kathrani A, Holder A, Catchpole B, et al. TLR5 risk-associated haplotype for canine inflammatory bowel disease confers hyper-responsiveness to flagellin. PLoS One 2012; 7:e30117.
64. Hugot JP, Gonzalez JP, Denys C. Evolution of the Old World Arenaviridae and their rodent hosts: generalized host-transfer or association by descent? Infect Genet Evol 2001; 1:13–20.
65. Catchpole B, Allenspach K. Canine inflammatory bowel disease: does innate immunity fail to discriminate between friend and foe? Vet J 2012; 194:7–8.
66. Kathrani A, Lee H, White C, et al. Association between nucleotide oligomerisation domain two (Nod2) gene polymorphisms and canine inflammatory bowel disease. Vet Immunol Immunopathol 2014; 161:32–41.
67. Travassos LH, Carneiro LA, Ramjeet M, et al. Nod1 and Nod2 direct autophagy by recruiting ATG16L1 to the plasma membrane at the site of bacterial entry. Nat Immunol 2010; 11:55–62.
68. Lassen KG, Xavier RJ. An alteration in ATG16L1 stability in Crohn disease. Autophagy 2014; 10:1858–1860.
69. Zhang J, Chen J, Gu J, et al. Association of IL23R and ATG16L1 with susceptibility of Crohn's disease in Chinese population. Scand J Gastroenterol 2014; 49:1201–1206.
70. Sadaghian Sadabad M, Regeling A, de Goffau MC, et al. The ATG16L1–T300A allele impairs clearance of pathosymbionts in the inflamed ileal mucosa of Crohn's disease patients. Gut 2015; 64:1546–1552.
71. Hacquard-Bouder C, Ittah M, Breban M. Animal models of HLA-B27-associated diseases: new outcomes. Joint Bone Spine 2006; 73:132–138.
72. Tsai KL, Starr-Moss AN, Venkataraman GM, et al. Alleles of the major histocompatibility complex play a role in the pathogenesis of pancreatic acinar atrophy in dogs. Immunogenetics 2013; 65:501–509.
73. Kennedy LJ, Davison LJ, Barnes A, et al. Identification of susceptibility and protective major histocompatibility complex haplotypes in canine diabetes mellitus. Tissue Antigens 2006; 68:467–476.
74. Hugot JP, Chamaillard M, Zouali H, et al. Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease. Nature 2001; 411:599–603.
75. Jozaki K, Shinkai H, Tanaka-Matsuda M, et al. Influence of polymorphisms in porcine NOD2 on ligand recognition. Mol Immunol 2009; 47:247–252.
76. Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. New York: John Wiley & Sons, 1980.
77. Therneau TM, Grambsch PM. Modeling survival data: extending the Cox model?. New York: Springer-Verlag Inc, 2000.
78. Evans JM, Tsai KL, Starr-Moss AN, et al. Association of DLADQB1 alleles with exocrine pancreatic insufficiency in Pembroke Welsh Corgis. Anim Genet 2015; 46:462–465.
79. Massey J, Boag A, Short AD, et al. MHC class II association study in eight breeds of dog with hypoadrenocorticism. Immunogenetics 2013; 65:291–297.
80. Massey J, Rothwell S, Rusbridge C, et al. Association of an MHC class II haplotype with increased risk of polymyositis in Hungarian Vizsla dogs. PLoS One 2013; 8:e56490.
81. Wilbe M, Ziener ML, Aronsson A, et al. DLA class II alleles are associated with risk for canine symmetrical lupoid onychodystropy (SLO) (Erratum published in PLoS One 2010;5:10.1371/annotation/6f88cccd-2f41-464d-904b-24356b5d87c9). PLoS One 2010; 5:e12332.
82. Barrientos LS, Zapata G, Crespi JA, et al. A study of the association between chronic superficial keratitis and polymorphisms in the upstream regulatory regions of DLA-DRB1, DLA-DQB1 and DLA-DQA1. Vet Immunol Immunopathol 2013; 156:205–210.
83. Barnes A, O'Neill T, Kennedy LJ, et al. Association of canine anal furunculosis with TNFA is secondary to linkage disequilibrium with DLA-DRB1*. Tissue Antigens 2009; 73:218–224.
84. Wilbe M, Jokinen P, Hermanrud C, et al. MHC class II polymorphism is associated with a canine SLE-related disease complex. Immunogenetics 2009; 61:557–564.
85. Goyette P, Boucher G, Mallon D, et al. High-density mapping of the MHC identifies a shared role for HLA-DRB1*01:03 in inflammatory bowel diseases and heterozygous advantage in ulcerative colitis. Nat Genet 2015; 47:172–179.
86. Angles JM, Kennedy LJ, Pedersen NC. 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 2005; 66:173–184.
87. Ross P, Buntzman AS, Vincent BG, et al. Allelic diversity at the DLA-88 locus in Golden Retriever and Boxer breeds is limited. Tissue Antigens 2012; 80:175–183.
Appendix
Primers used for amplification and sequencing of target gene regions in Great Danes.
| Primer No. | Primer | Sequence (5′ to 3′) |
|---|---|---|
| 1 | DLA88-Fa | GCGGCGACGGCCAGTGTCCCCGGAG |
| 2 | DLA88-Fb | CATTGGGTGCAGCGTTTCTGGAG |
| 3 | DLA88-Fc | ACCCCGGGAAGCCGCCTCTCC |
| 4 | DLA88-Fd | TCCGGGTCCGGGCGTCACC |
| 5 | DLA88-Ra | GACCCTGAGTCCATATTCCCTTCC |
| 6 | DLA88-Rb | CCTAGTAGGACTATCAACACCCAG |
| 7 | DLA88-Rc | GTAAAACCTCCCGGGAGTTCC |
| 8 | DRB1-Fa | GTGCCGCCGTCGGTGTCTTCC |
| 9 | DRB1-Fb | GGAAACACGAACCGTGGGTGCTG |
| 10 | DRB1-Ra | GACAGTGCCCCTCCGGGACAG |
| 11 | DRB1-Rb | TGAAATCGGGCTCTCAGAGGGAC |
| 12 | NOD2-F | CACGGACATGTACCTGCTGATCC |
| 13 | NOD2-R | GCCAGCCGCTCCTCCTGCATCTC |
| 14 | TLR5-F | GCTTGCACGGCTGTGTTTCCGTC |
| 15 | TLR5-R | GATGTAGTTGAAGCTCAGCAGGAG |
| 16 | ATG16L1-F | AGAAGTCATCTTTGAATCCCTCTGG |
| 17 | ATG16L1-R | CCTTCTATCAGACACAGTCAACAGG |
Positions of primers on canine genes are shown in Supplementary Figure S1. Primers 2 and 6 amplify a 1,100-bp fragment of DLA88. Primers 9 and 10 amplify a 406-bp fragment of DRB1. Primers 14 and 15 amplify a 324-bp fragment of TLR5. Primers 12 and 13 amplify a 730-bp fragment of NOD2. Primers 16 and 17 amplify a 280-bp fragment of ATG16L1. All other primers were used for sequencing and alternative amplification strategies.
