INTRODUCTION

Top Abstract Introduction Case Report Materials and Methods Results Discussion References

The bacterial genus Helicobacter contains at least 18 species (15). These organisms colonize the gastrointestinal tracts of several mammalian and avian hosts. The type species, Helicobacter pylori, and the ferret gastric pathogen H. mustelae have well-documented causal roles in the development of peptic ulceration and neoplasia (9, 11, 12, 18, 29, 43). Other Helicobacter spp. have been associated with enteritis (5, 40) and inflammatory bowel disease (IBD) (3, 41). Some helicobacters, such as H. canis (2), H. pullorum (37), H. heilmannii (35), and "H. rappini" and H. cinaedi (10), may be zoonotic. The original descriptions of H. canis were from the feces of healthy and diarrheic dogs (38) and a child with enteritis (2). H. canis also has been isolated from a dog with hepatitis (14). H. canis is closely related genetically to H. hepaticus, an enteric helicobacter which can produce hepatitis, hepatic neoplasia, and IBD in some mouse strains (8, 13). Because IBD is a common clinical finding in domestic cats (39) and because Helicobacter spp. are associated with IBD in mice and rats, the possible relationship between helicobacters and IBD in cats should be explored (3, 20, 34). The purpose of this study was to ascertain whether Helicobacter spp. could be isolated from cats with and without diarrhea.

	CASE REPORT

Top Abstract Introduction Case Report Materials and Methods Results Discussion References

Four Bengal cats (Asian leopard cat-domestic cat hybrids) were evaluated from a cattery with 20 other cats, 75% of which had episodic severe, watery, projectile, and mucus- and blood-tinged diarrhea during the preceding 6 months. Each of the four hospitalized cats examined in this study had a previous history of diarrhea. The younger of the four cats, cats 1 and 2, were 8-month-old intact females presenting clinically with vomiting, lack of appetite, weight loss, and severe dehydration, were underweight (23), and were aware of but uninterested in their surroundings. Cat 3 was a 2-year-old intact female with vesicular faucitis, bilateral serous ocular discharge, and mild conjunctivitis consistent with calicivirus infection (31) and mild feline acne. Cat 4 was a 4-year-old intact male. Both of the adult cats were hydrated, in good condition, and normally responsive. Abdominal palpation of all four cats revealed fluid- and gas-filled small intestines and no mesenteric lymphadenopathy. Diarrhea in cats 1 and 2 was persistent, with the consistency of water, a foul odor, and more than 10 bowel movements per day. Adult cats 3 and 4 had intermittent diarrhea; their diarrheic feces were soft to liquid and malodorous. Gastrointestinal endoscopy was performed on cats 3 and 4 by standard techniques. Fresh gastrointestinal biopsy samples were obtained during endoscopy of the stomachs, proximal small intestines, and colons of the cats. Feces also were collected from nine asymptomatic cats for Helicobacter sp. isolation.

	MATERIALS AND METHODS

Top Abstract Introduction Case Report Materials and Methods Results Discussion References

Clinical pathology. Serum antibodies against viruses were evaluated by an indirect immunofluorescence assay with feline infectious peritonitis virus UCD1-infected Felis catus whole fetal (fcwf-4) cells as a substrate for feline enteric coronavirus (32) and feline immunodeficiency virus (FIV) Petaluma-infected fcwf-4 cells as a substrate for FIV (45). Feline leukemia virus was evaluated by an enzyme-linked immunosorbent assay for p27 antigen (25).

Gross pathology. Biopsy samples from cats 3 and 4 were either rapidly frozen in OCT medium (Sakura Finetech, Torrance, Calif.) in a 2-methylbutane bath within liquid nitrogen or collected into 10% formalin, fixed overnight, embedded in paraffin, cut into 5-µm thin sections, and stained with hematoxylin and eosin and Warthin-Starry stains. Cats 1 and 2 were euthanized with an intravenous overdose of barbiturates and immediately necropsied. Necropsy tissues were processed as described for biopsy specimens.

Bacterial cultures. Fecal specimens (two each from the four cats) were plated onto seven different agars for bacterial culturing. These were MacConkey (PML, Rancho Cordova, Calif.), CVA (cefoperazone-vancomycin-amphotericin B; Remel Laboratories, Lenexa, Kans.), fresh BHI (brain heart infusion with 2.5 mg of trimethoprim, 5 mg of vancomycin, 1.25 IU of polymyxin B, and 2 mg of amphotericin B per ml), fresh brucella (fetal calf serum, trimethoprim, vancomycin, polymyxin B, and amphotericin B; Anaerobe Systems, San Jose, Calif.), prereduced anaerobically sterilized modified agar with cefoxitin, cycloserine, and fructose (CCFA; Anaerobe Systems), egg yolk, and FDA (5,000 mg of polymyxin B per ml, 10% vancomycin in water, 10% trimethoprim in ethanol, and amphotericin B). CVA agar was incubated at 37 and 42°C under microaerobic conditions in vented jars containing N₂, H₂, and CO₂ (90:5:5). An additional nine cats without diarrhea were surveyed from the same colony approximately 6 months after the initial evaluation of the four cats with diarrhea. These nondiarrheic cats were specifically screened for Helicobacter spp. The biochemical tests and phenotypic characterization used were based on previously described media and methods used by us and others for identifying, characterizing, and naming other Helicobacter spp. (16). Colony morphology and Gram staining of bacteria were determined by use of organisms obtained after 72 h of incubation of blood agar under microaerobic conditions. Bile tolerance was determined by growth of organisms on 1.5% bile media (1.5% desiccated ox bile [Oxoid] in 5% blood agar). A sample was also placed in selenite broth. After overnight incubation, the selenite broth was subcultured onto XLT4 (xylose-lysine-Tergitol 4) agar (PML). FDA agar was incubated under microaerobic conditions at 42°C. Brucella and CCFA agars were incubated microaerobically at 37°C. When colonies appeared on FDA agar, they were subcultured to fresh brucella agar and kept at 37°C.

DNA extraction for PCR analysis. DNA was extracted from cultured organisms with a High Pure PCR template preparation kit (Boehringer Mannheim Biochemicals, Indianapolis, Ind.) according to the manufacturer's directions. Briefly, the samples were lysed and incubated with 40 µl of proteinase K for 1 h at 55°C. Binding buffer (200 µl) was added to each sample and incubated for 10 min at 72°C before 100 µl of isopropanol was added (16). The samples were placed in filter tubes and centrifuged at 8,000 rpm for 1 min. The flowthrough was discarded, 500 µl of wash buffer was added to the samples, and the samples were centrifuged as before. This washing step was repeated three times. Elution of the DNA was achieved by adding 200 µl of elution buffer to the filter tubes and centrifuging the samples for 1 min at 8,000 rpm.

PCR amplification of bacterial DNA. The primer sequence chosen for PCR amplification recognizes a region of the 16S rRNA gene specific for members of the Helicobacter genus. The sets of primers used amplified a product of 1.2 kb. PCR amplification was achieved by use of a previously described method (16). Briefly, 20 µl of the DNA preparation was added to 100 µl of a reaction mixture containing 1× Taq polymerase buffer (supplied by Pharmacia, Uppsala, Sweden, but supplemented with 1 M MgCl₂ to a final concentration of 2.25 mM), 0.5 µM each primer, 200 µm each deoxynucleotide, and 200 µg of bovine serum albumin per ml. Samples were heated at 94°C for 4 min, briefly centrifuged, and cooled to 61°C. At this time, 2.5 U of Taq polymerase (Pharmacia) and 1.0 U of polymerase enhancer (Perfect Match; Stratagene, La Jolla, Calif.) were added, and then 100 µl of mineral oil was laid over the samples. For amplification of the 1.2-kb fragment, the following conditions were used: 35 cycles of denaturation at 94°C for 1 min, annealing at 58°C for 3 min, and elongation at 72°C for 3 min, followed by an elongation step of 8 min at 72°C. A 15-µl aliquot of each sample was electrophoresed through a 1% agarose gel, stained with ethidium bromide, and viewed by UV illumination.

RFLP analysis. The 1.2-kb PCR-amplified Helicobacter DNA from the two strains of H. canis isolated from the diarrheic cats, the two strains from the nondiarrheic cats, an H. canis strain from a dog, and H. canis ATCC 51401 was subjected to restriction fragment length polymorphism (RFLP) analysis. DNA digestion was accomplished by the addition of 10 U each of the restriction endonucleases HhaI and BfaI (New England Biolabs, Beverly, Mass.) and 1 µl of restriction buffer (New England Biolabs) to 16 µl of DNA and incubation at 37°C for 3 h. The samples were separated on a 6% Visigel separation matrix (Stratagene), stained with ethidium bromide, and viewed by UV illumination.

Genomic DNA extraction for 16S rRNA gene sequencing. Bacteria isolated from both fecal samples from one cat were cultured on blood agar plates, and the cells were harvested and washed twice with 1 ml of double-distilled H₂O. The pellets were suspended in STET buffer (8% sucrose, 50 mM EDTA, 0.1% Triton X-100, 50 mM Tris-HCl [pH 8.0]), and lysozyme (hen egg white; Boehringer) was added to a final concentration of 3 mg/ml. The suspension was incubated for 12 min at 37°C and then lysed with 1% sodium dodecyl sulfate. RNase A (bovine pancreas; Boehringer) was added to a final concentration of 0.05 mg/ml, and the solution was incubated for 1 h at 37°C. Then, 0.1 volume of a 5% cetyltrimethylammonium bromide-0.5 NaCl solution (Sigma Chemical Co., St. Louis, Mo.) was added, and the solution was gently mixed and incubated at 65°C for 10 min. DNA was extracted with an equal volume of phenol-chloroform (1:1, vol/vol), precipitated overnight in 0.3 M sodium acetate with 2 volumes of absolute ethanol at 20°C, and pelleted by centrifugation at 13,000 × g for 1 h at 4°C. The ethanol was decanted, and the pellet was air dried and suspended in sterile distilled water.

16S rRNA gene sequencing. The sequences of the 16S rRNA genes of the two H. canis strains (MIT 98-152 and MIT 98-153) isolated from two fecal samples from one cat were determined. For the amplification of 16S rRNA cistrons, 16S rRNA gene sequencing, and 16S rRNA data analysis, we used the methods described by Fox et al. (16). Briefly, primers C70 and B37 (16) were used to amplify the 16S rRNA genes. The amplicons were purified and directly sequenced by use of a TAQuence cycle sequencing kit (U.S. Biochemicals, Cleveland, Ohio). The 16S rRNA gene sequences were entered into a program for the analysis of 16S rRNA data in Microsoft Quikbasic for use with IBM PC-compatible computers and aligned as previously described (30). The database used contains approximately 100 Helicobacter, Wolinella, Arcobacter, and Campylobacter sequences and more than 900 sequences for other bacteria. Similarity matrices were constructed from the aligned sequences by use of only base positions for which 90% of the strains had data and were corrected for multiple base changes by the method of Jukes and Cantor (21). Phylogenetic trees were constructed by the neighbor-joining method (33).

Nucleotide sequence accession number. The 16S rRNA sequences for strains MIT 98-152 and MIT 98-153 have been deposited in GenBank under one accession number, AF177475, given that the two isolates are identical.

	RESULTS

Top Abstract Introduction Case Report Materials and Methods Results Discussion References

Clinical pathology. Complete blood counts and serum chemistries were within normal limits for all four cats. The cats were serologically negative for feline leukemia virus antigen and antibodies against FIV. All four were seropositive for feline enteric coronavirus at a minimum dilution of 1:100.

Gross pathology. Gross lesions were similar in both adult cats on endoscopy and included fluid distention of the large and small intestines, thickened erythematous mucosal tissue in the duodenum, and diffuse, patchy mottling in the proximal colon. The distal colon, stomach, and esophagus were all within normal limits. Gross lesions observed during necropsy of the young cats were limited to fluid and gas distention of the small and large intestines. Other abdominal organs were within normal limits.

Bacterial cultures. Gram staining of the fecal smears from cat 4 showed mixed flora, with numerous large gram-positive rods consistent with Clostridium spp. There was growth of a small number of Clostridium colonies, but no clostridial toxins were detected in toxin assays. Cat 3 had mixed flora on fecal smear Gram staining, with large gram-positive rods with spores and large numbers of small and large spiral gram-negative rods. The feces were culture and toxin A positive for C. difficile.

PCR. Bacteria isolated from seven of the eight fecal samples from the four diarrheic cats amplified a 1.2-kb product specific for Helicobacter spp. (Fig. 1). All nine fecal samples from the nondiarrheic cats had a similar Helicobacter-specific 1.2-kb product.

View larger version (70K): [in this window] [in a new window]   FIG. 1.   Electrophoresis of DNA amplified by Helicobacter-specific PCR on a 6% Visigel separation matrix with Helicobacter-specific primers. Lanes: M, 1-kb DNA ladder; 1, reagent control; 2, H. canis ATCC 51401; 3 to 10, DNAs extracted from feline fecal isolates.

RFLP analysis. With the restriction enzymes HhaI and BfaI, the two strains from the diarrheic cats had patterns identical to those of an H. canis strain isolated from the liver of a dog (14) and H. canis ATCC 51401 (Fig. 2). Helicobacter spp. isolated from two cats without diarrhea had RFLP patterns consistent with H. canis isolated from two diarrheic cats (data not shown).

View larger version (123K): [in this window] [in a new window]   FIG. 2.   Comparison of RFLP patterns of the 1.2-kb Helicobacter-specific PCR product with restriction enzymes BfaI and HhaI. Lanes: M, 100-bp DNA ladder; 1, H. canis MIT dog isolate (14); 2, H. canis ATCC 51401; 3, MIT 98-152; 4, MIT 98-153.

16S rRNA analysis. The sequence of the cat isolates (MIT 98-152 and MIT 98-153) differed from that of the type strain of H. canis (L13464) by 4 bases (Fig. 3). There is an intervening sequence (IVS) in the 16S rRNA gene in about one-third of H. canis strains (24). However, the two feline H. canis isolates did not have an IVS. Whether the H. canis strains with IVSs represent one or more distinct subspecies has not been determined.

View larger version (41K): [in this window] [in a new window]   FIG. 3.   Phylogenetic tree constructed on the basis of 16S rRNA sequence similarity values. The two H. canis strains from diarrheic cats are in bold. Scale bar, 5% difference in nucleotide sequences, as determined by measuring the lengths of the horizontal lines connecting two species.

	DISCUSSION

Top Abstract Introduction Case Report Materials and Methods Results Discussion References

To our knowledge, this is the first report of H. canis in cats, and it is significant for several reasons, including the possibility that the organism has zoonotic potential. This organism is related to other helicobacters previously associated with colitis and proctitis, including H. hepaticus (3, 17), H. cinaedi (40), H. fennelliae (5, 40), and H. bilis (20, 34). The murine IBD produced by H. hepaticus and H. bilis is most pronounced in immunosuppressed rodents (3, 8, 20, 34, 41). However, in immunocompetent mouse strains, H. hepaticus also can induce enterocolitis and typhlitis (17, 44). Mice also can develop hepatitis, hepatic adenomas, and hepatocellular carcinoma as a result of H. hepaticus infection (42). H. canis has been previously reported in a child with gastroenteritis (2), dogs with and without diarrhea (38), and a puppy with necrotizing hepatitis (14).

The clinically ill Bengal cats in the present report had severe diarrhea, enterocolitis, and mild portal hepatitis associated with multiple primary and opportunistic pathogens, including H. canis, Cryptosporidium spp., C. perfringens, C. difficile, and C. helveticus. A gram-negative organism similar to H. heilmannii was observed in the stomachs of these cats. Although gastric helicobacters are common in cats and are associated with mild to moderate inflammation, definite clinical signs have not been linked to helicobacter-associated gastritis in cats (27, 28). Feline enteric coronavirus antibody also was detected in the serum, but this is a common finding in cattery cats (7) and is not typically associated with clinical enteritis (31). Because each of the protozoal and bacterial organisms isolated from these affected cats has been associated with diarrhea, it is not clear which if any was the primary pathogen and how significant synergistic interactions among the intestinal flora were in causing clinical disease.

Cryptosporidium spp. cause diarrhea in humans and animals, but the condition is self-limiting unless the subject is immunosuppressed (19). Cryptosporidiosis with oocyst shedding occurs in normal and diarrheic cats, and experimental attempts to infect cats with C. parvum commonly result in shedding with no clinical signs (31). Similarly, C. perfringens is frequently detected in stools of normal cats, although toxin production may suggest that C. perfringens is at least partially responsible for some clinical signs. Outbreaks of C. perfringens enteritis have been reported for cats maintained in a cattery (6) and captive cheetahs (4). Pseudomembranous colitis due to C. difficile has been reported for humans with underlying disease or undergoing antibiotic therapy; the infection often results from nosocomial exposure (22). As with C. perfringens, the diarrhea and intestinal inflammation are due to an exotoxin produced by C. difficile. C. helveticus is a catalase-negative or weakly positive Campylobacter sp. which was first reported for normal cats but has also been isolated from the feces of diarrheic cats (26, 36). The diversity of pathogens in the cats in this study suggests that they may have been immunocompromised. However, the cats were not infected with retroviruses and appeared healthy except for the enteritis. Alternatively, one or several of the bacterial organisms may have caused primary bacterial colitis and may, under certain conditions, be considered primary pathogens.

As noted for the cats in this study and other cats screened by us (10a), coinfection with Campylobacter spp. and Helicobacter spp. is commonly observed. Such coinfection also has been documented for humans with diarrheic feces (1). This situation poses a particular diagnostic challenge in correctly identifying both campylobacters and helicobacters in one fecal specimen, given the similarities in their phenotypic and biochemical profiles. As in our study, the use of specific and sensitive PCR primers distinguishing the two genera may be required.

The cats in the present report had enteritis and periportal hepatitis. However, cats without diarrhea were also colonized with H. canis. The data indicates that H. canis is endemic in this cat colony, given that H. canis was isolated from multiple cats at two time points (6 months apart). It will be important to further evaluate H. canis for causal roles in IBD and hepatitis. Additionally, studies are needed to determine whether or not enterohepatic disease associated with H. canis infection occurs more commonly in certain hosts of a particular genotype (e.g., particular breeds of dogs or cats), in immunosuppressed hosts, and in association with certain other enteric infections. The study of H. canis in cats with IBD may also provide insight into the etiopathogenesis of IBD in humans.