Introduction

First reported in 1965, granulomatous colitis (GC) is an uncommon form of inflammatory bowel disease (IBD) predominantly diagnosed in young Boxers and French Bulldogs.^{1 23–4} Dogs with GC show clinical signs of colitis, haematochezia and weight loss, which can be associated with anaemia, hypoproteinaemia, cachexia and euthanasia in severe cases.^{3 45–6} In severe cases, thickening of the colonic mucosa and ulceration are readily visualised on endoscopy, and sonography may reveal colonic thickening and regional lymphadenonathy. Histologically, GC of Boxers and French Bulldogs is characterised by infiltration of the lamina propria and submucosal layers by lymphocytes, granulocytes and macrophages containing periodic acid–Schiff-positive (PAS+) material. These cells are also found within, and surrounding lymphatics of, the tunica muscularis and serosal surface. The caecum and ileum are also involved with similar lesions, although often to a lesser degree.^{3,4,6 78910–11}

For many years, PAS+ GC in dogs was considered a severe and incurable idiopathic immune-mediated disease. The application of culture-independent methodologies to detect bacteria in fixed colonic biopsies from Boxers with GC led to the identification of multifocal clusters of mucosally invasive Escherichia coli within macrophages using immunocytochemistry and fluorescence in situ hybridisation (FISH).^1,3,4 FISH analysis uses fluorescently labelled oligonucleotide probes that hybridise to bacterial 16S or 23S ribosomal DNA to localise metabolically active bacteria within formalin-fixed tissues.^1,4 FISH utilising E coli/Shigella species-specific probes has become the definitive test for identifying mucosally invasive E coli in Boxers and French Bulldogs with GC. Clinical remission and cure of GC-affected dogs correlates with the eradication of mucosally invasive E coli by antibiotics that are capable of penetrating macrophages and killing intracellular E coli. Recently, genetic analysis of affected dogs has implicated a region on chromosome 38 that is involved in detection and killing of E coli in other species.¹² Thus, it is emerging that E coli-associated GC in Boxers and French Bulldogs is likely a heritable genetic defect in sensing or killing intracellular E coli.

In cats, only one case of GC has been previously been reported, which was published in 1979. It describes a 5-year-old Persian crossbred cat with clinical signs of colitis. Light and electronic microscopy of colonic biopsies showed mucosal colitis with characteristic PAS+ macrophages.¹³ This cat responded to chloramphenicol therapy and was healthy for at least 4 years after diagnosis. Even if at that time there was no identification of the underlying cause, the case report suggests involvement of an undiagnosed bacterial infection. It is against this background that we sought to determine the presence of mucosally invasive bacteria, including E coli, and clinical response to antimicrobial therapy that is effective in dogs in a cat with PAS+ GC.

Case description

A 4-year-old domestic shorthair male neutered cat was referred to the internal medicine service of Centre Hospitalier Vétérinaire Fregis (France) for chronic intermittent haematochezia and faecal incontinence of 7 months’ duration. No weight loss was reported and the cat had a good appetite. Therapeutic trials with fenbendazole (50 mg/kg PO for 5 days), metronidazole (10–15 mg/kg PO q12h for 10 days) and hyperdigestible gastrointestinal diet (Hill’s prescription diet i/d) were unsuccessful.

Physical examination, complete blood cell count and biochemistry panel (including serum folate and cobalamin concentration) were within normal limits.

Faecal sample for parasites by using direct smear evaluation, zinc sulfate centrifugal flotation techniques and PCR for Tritrichomonas and Giardia species were negative. Abdominal ultrasonography revealed a colonic wall thickening (2.5–3.0 mm) with attenuation of wall layering and hypo- to echoic multifocal nodules (2 mm diameter) in the submucosal layer. Colonoscopy showed an irregular and thickened colonic wall with multiple erosions, compatible with ulcerative colitis or infiltrative neoplasia (Figure 1). Colonic endoscopic biopsy samples were collected.

Figure 1

Colonoscopy showed thick irregular mucosa with multiple superficial ulcers (first image 11 o’clock). The mucosa was friable and bled easily during the procedure

Biopsies were fixed in 4% saline-buffered formalin and embedded in paraffin wax. Sections of 4–5 µm were stained with haematoxylin and eosin, PAS, Toluidine blue and Fite-Faraco, and submitted for routine histopathological examination. Histopathology revealed severe multifocal mucosal ulcerations and infiltration of the mucosal lamina propria by large numbers of macrophages, with scattered small lymphocytes and plasma cells (Figure 2). The macrophages had abundant eosinophilic granular cytoplasm that was strongly PAS+. Toluidine blue and Fite-Faraco stains did not show mast cell infiltration or acid-fast bacteria, respectively. Histological findings were consistent with severe PAS+ GC similar to that documented in Boxers and French Bulldogs.

Figure 2

Histopathology of colonic biopsies showed accumulation of macrophages with abundant cytoplasm containing periodic acid–Schiff (PAS)-positive material throughout mucosal lamina propria (original photos: LAPVSO). H&E = haematoxylin and eosin

FISH analysis of colonic biopsies using eubacterial (EUB338-6FAM) and E coli-specific probes (E coli–Cy3) and a previously described technique revealed multi-focal clusters of intracellular E coli (Figure 3).¹ More bacteria were visible with eubacterial vs E coli probes suggesting the possibility of a mixed infection.

Figure 3

Fluorescence in situ hybridisation of colonic biopsies showing multifocal clusters of invasive intracellular rods (EUB-338, upper row) that hybridised with a probe to Escherichia coli/Shigella species (lower row), similar to granulomatous colitis in dogs. Bacteria stain red (cy-3). Nuclei/DNA stain blue (4’,6-diamidino-2-phenylindole)

Colonic swab culture was positive for E coli and negative for Salmonella species, Yersinia species and Campylobacter species. Antimicrobial susceptibility profiles of E coli isolates showed broad susceptibility to macrophage-penetrating antimicrobials (Table 1). Treatment with enrofloxacin (5 mg/kg q24h for 6 weeks) led to progressive and complete resolution of clinical signs with remission sustained for 13 months to date.

Table 1

Bacterial culture results (from colonic mucosal swab)

Discussion

To our knowledge, this is only the second case report of PAS+ GC in an adult cat. Historically, a primary immune-mediated pathogenesis was presumed in dogs,^5,14 but the recent identification of invasive E coli in dogs with GC increased our suspicion that an infectious agent could be involved and led us to perform FISH analysis, which revealed the presence of multifocal clusters of invasive intracellular E coli. These results reinforce the suspicion of an ineffective phagocytic activity, which can be explained either by enhanced bacterial resistance to destruction, ineffective phagocytosis or a combination of these.¹³ In humans and experimental animals, it has emerged that genetic defects in bacterial sensing and killing (eg, NOD2, ATG16L) are frequently associated with IBD.^{15 16–17} Variants in these genes are associated with downstream effects on protein function and, subsequently, clinical disease.^{15 16–17} Recent studies in Boxer dogs and French Bulldogs with GC have identified a region in chromosome 38 encoding genes in the signaling lymphocyte activation molecule family that are involved in the sensing and killing of E coli in murine models.¹² GC in cats is extremely rare, with only two cases and two different breeds (Persian and DSH in this study) reported. It remains to be determined if a genetic basis exists.

The E coli strain isolated from this cat was susceptible to many different classes of antibiotics. We based anti-biotic selection on susceptibility and the ability to penetrate macrophages. Owing to their lipophilicity, fluoroquinolones attain high intracellular concentrations and are effective against susceptible E coli within macrophages in Boxers and French Bulldogs with GC.¹⁸ Therefore, fluoroquinolone therapy was selected. Even if several fluoroquinolones (eg, pradofloxacin or marbofloxacin) would have allowed safer use at higher concentrations, enrofloxacin was chosen as it is overall better documented in the therapeutic management of GC in dogs. However, we were aware that high doses of enrofloxacin are associated with acute retinal degeneration in cats,^19,20 although adverse effects occur generally above the current dosage of 5 mg/kg/day (dose administered in this case). We chose to administer for 6 weeks as short-term duration of therapy can be ineffective in dogs with GC and is thought to relate to the time taken for mucosal healing to prevent invasion of resident E coli.^1,4,21 The clinical improvement observed and the complete resolution of clinical signs with antibiotics closely paralleled findings in dog and support the direct causal role for E coli in feline GC.^1,4,21

This case report had several limitations. There was no culture of colonic biopsy and only a colonic wall swab culture was performed. Although it would be interesting to have a direct colonic culture, from our point of view the rectal wall culture seemed representative of the microbial flora involved.

A second limitation was the lack of clinical work-up performed to exclude fungal infection. However, fungal diseases are extremely rare in France. In addition, PAS staining was negative for fungal organism and clinical response was complete with enrofloxacin.

A third limitation was the lack of histopathological follow-up. In dogs, a study documented that the results of PAS staining remained positive for more than 6 months, despite clinical remission.^4,21 In cats, repeat histopathology evaluation after medical therapy has not been performed. Despite long-term clinical remission confirming the success of medical therapy, it would have been informative to repeat colonic biopsies, histopathology and FISH at least 6 months after the diagnosis to determine if clinical remission was associated with eradication of intracellular E coli and resolution of PAS+ colitis. However, in the face of successful medical therapy and long-term remission, a repeat biopsy was considered optional rather than essential.

Conclusions

This case report revealed that E coli-associated GC can also affect cats and should be considered in the differential diagnosis of chronic haematochezia in this species. Further studies are needed to assess molecular, genetic and immune pathways underlying intracellular invasion by E coli in cats.