Michael J. Day, BSc, BVMS(Hons), PhD, DSc, DECVP, FASM, FRCPath, FRCVS
There are three major canine chronic inflammatory enteropathies, which may share elements of pathology and pathogenesis. These are 1) idiopathic inflammatory bowel disease (IBD), 2) antibiotic-responsive diarrhoea (ARD; previously small intestinal bacterial overgrowth, SIBO), and 3) food-responsive diarrhoea (FRD; dietary hypersensitivity). Over the last decade there have been significant advances in understanding the pathogenesis of these disorders which should, in time, lead to the development of novel therapeutic approaches.
The Immunopathogenesis of Chronic Intestinal Inflammation
Increased understanding of the immunopathogenesis of canine enteropathy has come about as a consequence of investigations in equivalent human and rodent model diseases. The mucosal immune system maintains a delicate balance between responsiveness and tolerance, and disruption of this balance is fundamental to initiating chronic intestinal inflammation. The major factors involved in the maintenance of intestinal homeostasis are the mucosal barrier (epithelium), an appropriately functioning mucosal immune system and the presence of endogenous microflora. The requirement for a protective barrier is demonstrated by genetically modified mice which lack epithelial cohesion mediated by E-cadherin. In areas where the mutant gene is expressed, epithelial disruption occurs and there is localised intestinal inflammation. Similarly, mice with targeted disruption of a range of genes encoding immunologically active molecules (e.g., cytokines, TCR α chain) spontaneously develop intestinal inflammation, and alteration of the balance of regulatory T lymphocyte subsets permits expression of colitis in rodent models. These studies demonstrate that T cells are pivotal in maintaining gut immune homeostasis, and that the protective effects of T cells are mediated via the release of immunoregulatory cytokines.
The critical importance of the presence of an endogenous flora is demonstrated by the fact that intestinal inflammation does not develop when the mice in these model systems are reared in a germ free environment. This further suggests that loss of tolerance and an aberrant immune response to components of the endogenous bacterial flora may underlie inflammatory enteropathy. The initial contact between microbial antigens and the gut immune system is via the pattern-recognition receptors (e.g., Toll-like receptors; TLRs) expressed by antigen presenting cells (e.g., dendritic cells).
The increasing prevalence of human IBD (Crohn's disease and ulcerative colitis) in the western world is likely to be related to lifestyle factors in a genetically susceptible population. Genetic associations are an area of active research in this field.
Inflammatory Bowel Disease
Canine idiopathic IBD is characterised by persistent clinical signs of GI disease, associated with histological evidence of inflammation, of undetermined cause, in the small and/or large intestinal mucosa. IBD likely covers an overlapping spectrum of diseases, the most common of which is idiopathic lymphocytic-plasmacytic enteritis (LPE). Forms of IBD unique to particular breeds have been described, including immunoproliferative enteropathy of basenjis, a diarrhoeal syndrome in Lundehunds, gluten-sensitive enteropathy of Irish setters, and protein-losing enteropathy / nephropathy in soft-coated wheaten terriers (SCWTs). German shepherd dogs (GSDs) are predisposed to IBD. Until recently, histiocytic ulcerative colitis (HUC) in boxer dogs would have been included in this group of diseases, but we now know that this entity is associated with infection by adherent and invasive E. coli and can be readily treated by appropriate antimicrobials. The genetic component of HUC is suggested by the finding of single nucleotide polymorphisms (SNPs) in the NCF2 autophagy-related gene that encodes molecules involved in intracellular killing of bacteria.
Immunohistochemical investigation of endoscopic intestinal biopsies has shown no overall difference in total lamina propria (LP) cell counts in dogs with IBD versus controls, but differences were documented in the phenotype of LP immune cell subsets. There were increases in T cells (particularly those expressing the αβTCR and CD4), IgG+ plasma cells, macrophages and granulocytes, but a decreased number of mast cells. Increased numbers of intraepithelial T cells were also noted. In dogs with lymphocytic-plasmacytic colitis there are increases in IgA+ and IgG+ plasma cells, as well as T cells in both LP and epithelial compartments. Analysis of the clonality of T- and B-cell receptors has shown reduced diversity in lymphocytes within the intestine of dogs with IBD; suggesting selective expansion of lymphocytes with particular antigenic specificities.
Expression of nitrite, a metabolite of the inflammatory mediator nitric oxide (NO), has been measured in colonic lavage fluid from dogs with IBD, and there is increased mucosal expression of mRNA encoding inducible nitric oxide synthase (iNOS), the enzyme responsible for NO production. Serum acute phase protein inflammatory markers (e.g., CRP, haptoglobin) also correlate with disease activity. Detection of serum antibodies to the peri-nuclear cytoplasmic region of neutrophils (pANCA) and the yeast Saccharomyces cerivisiae may have diagnostic utility as in human enteropathy, but studies in dogs have not proven a significant association other than for pANCA in SCWTs. Faecal and serum concentrations of calprotectin (a calcium-binding protein in phagocytic cells) are also elevated in dogs with IBD, providing another potential biomarker.
Studies of cytokine gene expression in mucosal biopsy tissue have been disappointing. Initial work using conventional semi-quantitative RT-PCR suggested upregulation of particular genes, but when these studies were repeated using quantitative real-time RT-PCR, no specific cytokine expression profile was identified. The most recent such published study shows no elevation in the expression of genes associated with Th1 (IFN-γ) or Th17 cells (IL-17A, IL-22) or regulatory cells (IL-10, TGF-β). Upregulation of the expression of genes encoding chemokines (CCL-2, -20, -25 and -28, and CXCL8) has been shown in intestinal mucosa of dogs with IBD, but there was no upregulation of associated chemokine receptor genes. A recent investigation has used gene expression microarray to identify the broad sweep of genes that are either up- or downregulated in canine IBD. Upregulated genes mostly relate to immune function or tissue remodelling. Further molecular studies have examined the expression of genes encoding TLRs, but there is variation in the published data as to which TLRs may be over- or under-expressed with one study suggesting elevations in TLR-2, -4 and -9 and another only TLR-2. SNPs in the canine NOD2, TLR-4 and TLR-5 genes have been associated with IBD in GSDs.
A key element in the pathogenesis of IBD may be the qualitative nature of the intestinal microbiota. High-throughput molecular methodology has allowed characterization of the broad microbial content of the microbiota of individual animals. There are distinct differences between dogs with IBD and normal dogs, which enable these populations to be clearly stratified.
Some dogs with IBD are resistant to glucocorticotherapy, which may relate to increased expression of the drug efflux pump P-glycoprotein (PGP), encoded by the MDR1 gene, by target cells. Increased PGP expression has been shown by both enterocytes and lamina propria lymphocytes in dogs with IBD.
Antibiotic Responsive Diarrhoea
Dogs with ARD may, or may not, have an increased number of bacteria in the small intestine above that which is considered normal (105 CFU/ml duodenal juice, but this figure is controversial). Young animals are principally affected, and there is a strong breed predisposition for the GSD. Minimal histological changes are evident, but increased intestinal permeability has been demonstrated.
The predisposition of the GSD to ARD may be related to an abnormality in IgA biology. Some studies have suggested a relative serum IgA deficiency in healthy GSDs, whilst in others the wide variance in serum IgA concentration was more consistent with IgA dysregulation. Low tear and faecal IgA concentrations have also been documented in healthy GSD, suggesting defective mucosal IgA secretion. GSD with ARD may have relatively low serum and duodenal juice IgA concentrations, and there is a relative deficiency of IgA secretion from the small intestinal mucosa of GSDs with small intestinal diseases. Despite this, duodenal IgA+ plasma cell numbers in GSDs with ARD are either normal or increased. A study of a GSD colony revealed an association between reduced serum and faecal IgA concentrations, increased serum IgG concentrations and colonisation of the gut by enteropathogenic E coli.
Reduced IgA concentrations within the duodenal lumen in the presence of normal to increased LP IgA+ plasma cells in affected GSDs, suggests a complex defect in IgA metabolism in this breed. There are several points in the pathway of IgA translocation, from LP plasma cell to duodenal lumen that may be affected, but GSD with enteropathy have normal expression of genes encoding key molecules in this pathway: the IgA α chain, J chain, pIgR and secretory component.
Immunohistochemistry of the duodenal mucosa of dogs with ARD has shown increases in LP CD4+ cells in the presence of minimal histological change, and by conventional semi-quantitative RT-PCR there is increased mucosal mRNA expression for a variety of cytokines (most notably TNF-α and TGF-β). In situ hybridisation techniques have demonstrated increases in IL-10 and IFN-γ mRNA expression in dogs with numerically larger duodenal bacterial populations. These findings all suggest a role for disturbed intestinal immunological homeostasis in canine ARD.
Antibiotic treatment of dogs with ARD may lead to resolution of clinical signs, but not necessarily a reduction in bacterial numbers. GSDs with ARD treated with oxytetracycline had resolution of clinical signs, reductions in cytokine mRNA for TNF-α and TGF-β, but bacterial numbers (both total and anaerobic) did not decline. This might suggest that the effect of antibiosis is by modifying the balance of composition the enteric flora, or that these antibiotics have an immunomodulatory effect.
Food Responsive Diarrhoea
Advances in the study of FRD have been fewer than those for IBD and ARD, perhaps due to the difficulty in confirming a diagnosis of this entity. It is of note, however, that there is increasing evidence that a proportion of dogs which may have been considered to have IBD can be managed by dietary manipulation. Serological studies have revealed the poor diagnostic utility of detection of food allergen-specific IgE or IgG in serum, although faecal allergen-specific IgE is associated with clinical signs in research studies.
In a recent study, 10 of 13 GSDs with chronic enteropathy were categorized as having diet-responsive disease (and 3 as having ARD). Collectively, these dogs had elevated TLR-4 (but not TLR-5) expression and alterations in the nature of the intestinal microbiota. Increased expression of nuclear factor-kappaB (a transcription factor involved in cytokine gene expression) has been shown in the intestinal mucosa of dogs with FRD. Two studies have examined T-cell infiltration and cytokine gene expression in the skin and duodenal mucosa of dogs with cutaneous adverse reactions to food (without diarrhoea). Lesional skin from these dogs had an infiltrate of CD8+ T cells with upregulation of genes encoding IL-4, IL-13, IFN-γ and Foxp3. In contrast, no significant differences were found in T-cells or cytokine genes in duodenal samples from these dogs given either an elimination or provocation diet. This may suggest that the primary immunological changes in cutaneous adverse reactions to food occur in the skin rather than the gut. In contrast, our own studies of dogs with FRD have shown significant reductions in gene expression for a range of molecules (cytokines, chemokines and TLRs) in duodenal biopsy samples after dietary modification.