INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

Helicobacter pylori was first cultured in Western Australia in 1983. This gram-negative spiral bacterium has since been shown to be a common inhabitant of the human gastric mucosa and areas of gastric metaplasia in the duodenum (11, 22, 24, 31). In developing countries, more than 80% of the population is infected by age 20, whereas up to about 50% of people 50 years of age and above are infected in developed countries (10). The organism is now a recognized etiologic agent of gastritis (20-22) and has been suggested to play a role in the pathogenesis of gastroduodenal ulcers and gastric carcinoma (1, 9, 27). In particular, the vacuolating cytotoxin (VacA) and cytotoxin-associated antigen (CagA) produced by type I but not by type II H. pylori have been causally associated with the development of severe forms of gastroduodenal disease (4, 29). While the function of the CagA protein is unknown, epithelial vacuolation caused by VacA may produce tissue damage and ulceration. Up to this time, the search for an effective cure for gastroduodenal disease induced by H. pylori has been hampered by the difficulty of reproducing infection, with its accompanying pathogenic sequelae, in easy-to-manage animal models. Since a small animal is generally preferred as a model for human disease, there have been attempts to infect germ-free mice and rats (2, 5), but without much success, possibly due to host specificity restriction of the human-derived H. pylori. To overcome the species specificity problem, Helicobacter species, such as Helicobacter felis and Helicobacter mustelae, were used to generate mouse (17) and ferret (8) infection models, respectively. However, the usefulness of these species is limited, since they do not express VacA and other virulence factors required for the induction of gastric pathology, such as ulcers and inflammation (29). Although successful colonization has been achieved with germ-free (14) or athymic nude mice (30), these models are relatively difficult to handle compared to conventional mice. The search for a suitable animal model was extended to other species, such as gnotobiotic piglets (6, 15, 16), gnotobiotic dogs (25), Mongolian gerbils (32), and monkeys (7). It is obvious that handling these animals in large numbers is rather difficult. Recently, successful colonization and induction of gastritis was achieved in specific-pathogen-free (SPF) and conventional BALB/c and CD-1 mice (19) by using fresh clinical H. pylori strains for challenge. H. pylori CFU were recovered from gastric mucus, suggesting that mucus was a favored niche for this organism. In the present study, we describe a new strain of mouse as a model of H. pylori infection and extend the previous observation on mouse mucus colonization by demonstrating that H. pylori adheres strongly to gastric epithelial cells in vivo. Such bacterial adherence is persistent and may give rise to gastritis, depending on the challenge dose and the H. pylori phenotype.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Animals. For routine challenge purposes, we used 7- to 8-week-old mice from a hairless but euthymic SPF strain (NS:Hr/ICR) from the animal breeding facility of Nisshin Flour Milling Co., Ltd., with genetic stocks deposited at the Imamichi Institute for Animal Reproduction (deposit no. IAR-NHI-9701), Toxicology Research Center, Ibaraki, Japan. The NS:Hr/ICR strain of mice is immunocompetent and not deficient in thymus. The thymus constitutes 0.12 to 0.21% of total body weight at 21 to 23 weeks of age (unpublished data). For comparative experimental infection purposes, the following animals were used: SPF Mongolian gerbils (MON/Jms/Gbs; Nippon SLC, Hamamatsu-shi, Shijuoka-ken, Japan), commercial hairless mice (HOS:HR-1; Hoshino Animal Breeder, Yashio-shi, Saitama-ken, Japan) and SPF BALB/c mice. All animals were housed in a lightweight stainless steel autoclavable pentagonal isolator (13) and given commercial food pellets and water ad libitum. The animal facility maintained a 12-h light-dark schedule.

Bacterial strains and cultivation conditions. H. pylori NSP335, NSP305, and NSP355 were isolated from gastric biopsy samples of patients with gastritis. NSP335 and NSP305 belonged to the type I phenotype (VacA and CagA positive) while NSP355 was type II (negative for the two antigens). NSP333 and NSP394 were laboratory-passaged strains (types II and I, respectively). The clinical strains were isolated from gastric biopsies by inoculation of Hp selective medium (Eiken Co., Ltd., Tokyo, Japan) and incubation under microaerobic conditions with the use of activated CampyPak Plus (BBL Microbiology Systems). The organism was identified by Gram staining; production of oxidase, catalase, urease, and H₂S; and reduction of nitrate. For challenge, the strains were grown in brain heart infusion broth containing 10% horse serum (Koujinbaiyo Co., Ltd., Saitama, Japan) under microaerobic conditions for 72 h in a temperature-controlled shaker.

Challenge conditions and experimental conditions for establishment of an animal model of gastric colonization. Prior to challenge, all test animals were left without food for two consecutive days. On challenge days 1 and 2, the mice were administered 10⁷ CFU orally in a single dose. After challenge, the mice were given food and water ad libitum. At specific time points after challenge, the mice were bled for serological testing by enzyme-linked immunosorbent assay (ELISA) and sacrificed for bacteriological isolation and histopathological examination. The ages and number of animals, the H. pylori strains used, the inoculum dose, and necropsy time postchallenge are indicated in Tables 1 to 3. The experiment was divided into three phases: (i) animal model selection (Table 1), (ii) H. pylori strain selection (Table 2), and (iii) high-challenge dosing for gastritis induction (Table 3). Since a positive finding should yield a bacterial count of at least 10 CFU per g of gastric tissue or 10 CFU per g of gastric mucus, these were taken as the minimum detection limits for our test.

Enumeration of H. pylori from gastric epithelial and mucus washings. During necropsy, half of the glandular region of the stomach was obtained under aseptic conditions. The excised gastric tissue was weighed and placed in 9 ml of phosphate-buffered saline (PBS), pH 7.0, containing 0.05% cysteine HCl and 0.1% agar. To confirm and quantify gastric colonization based on gastric epithelial cell adhesion by CFU of H. pylori, the gastric tissue section was washed by vortexing it 10 times at top speed in the above-mentioned PBS medium (about 2 s each time). This was done in a total of eight consecutive tubes for a total of 80 cycles of vortexing. Finally, the serially washed gastric tissue was dabbed with sterile filter paper to absorb water and the weight of the mucus-free tissue was taken. The tissue was then homogenized with a glass homogenizer in 9 ml of PBS medium. About 0.1 ml obtained at each washing step and from the supernatant of the gastric homogenate was inoculated on Hp selective agar plates and incubated for 5 days at 37°C under microaerobic conditions. CFU of H. pylori were identified according to the procedures described in "Bacterial strains and cultivation conditions." The presence of H. pylori CFU in the mucus wash or gastric tissue homogenate was considered indicative of colonization.

Histopathology. The other half of the glandular region of the stomach remaining after bacteriological testing was fixed in 10% buffered formalin and processed for histopathological examination by standard methods for hematoxylin-eosin (HE)-stained sections.

Antigen preparation for immunoassay plates and antibody titration by ELISA. Antigen was prepared from the NSP335 strain of H. pylori by a procedure previously described (18) for Helicobacter hepaticus ELISA. For coating microwells, the antigen was diluted to 5 µg/ml in carbonate buffer, pH 9.6, and 100 µl was dispensed into each well. After overnight incubation at 4°C, the plates were washed three times with PBS-Tween 20 (0.05% Tween 20) and blocked with 3% bovine serum albumin (fraction V; Miles) for 1 h at 37°C. The plates were washed again three times with the same wash buffer, and 100 µl of the test mouse serum prediluted 1:100 was dispensed into each well. After 1 h at 37°C, the wells were washed three times and 100 µl of horseradish peroxidase-anti-mouse immunoglobulin G (IgG) conjugate in suitable dilution was dispensed into each well. After 1 h at 37°C, the plates were washed five times and 100 µl of substrate (ortho-phenylenediamine) was dispensed into each well. After 10 to 15 min of incubation at room temperature, the reaction was stopped with 50 µl of 3 N H₂SO₄/well. Absorbance was read at 490 nm. Two wells per plate, not coated with antigen but treated similarly to the test wells, were used as the controls, and their optical density (OD) was subtracted from all test well readings. An absorbance of twice, or more than twice, that of the negative control was taken as positive. Differences in mean OD between groups were tested for significance by the Student t test.

Immunoperoxidase staining of paraffin sections. Paraffin sections of mouse stomach tissue were deparaffinized with xylene and rehydrated with decreasing concentrations of ethanol. For immunostaining, a Histofine kit (Nichirei, Tokyo, Japan) was used. All steps were performed at room temperature according to the manufacturer's instructions. Briefly, after a 5-min wash in PBS, the slides were blocked sequentially with 3% hydrogen peroxide in methanol and goat serum. The primary antibody consisted of a 1:4,000 dilution of rabbit anti-whole-cell H. pylori, while the second antibody was biotinylated goat anti-rabbit IgG. The horseradish peroxidase-streptavidin conjugate was allowed to bind with the biotinylated rabbit IgG, and diaminobenzidine (Histofine kit; Nichirei) was used as the substrate to detect the enzyme.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

The susceptibility of different candidate animal models to H. pylori colonization was examined as a first step, using a type I phenotype expressing both CagA and VacA antigens as a challenge strain. Table 1 shows that strain NS:Hr/ICR of hairless mice showed the highest degree of susceptibility to colonization by H. pylori. All test mice of this strain were colonized by 1 and 2 weeks postchallenge, while other test strains of mice had lower frequencies of positive gastric bacterial isolation, which further declined at 2 weeks postchallenge. The strain of Mongolian gerbils tested was negative at both 1 and 2 weeks of testing.

The difficulty in infecting SPF or conventional mice with H. pylori for use as an animal model of gastritis that mimics the human disease could be overcome by using a method recently described by Marchetti et al. (19). In that study, SPF or conventional BALB/c or CD-1 mice were deprived of food for 24 h before being challenged with fresh clinical isolates of H. pylori at days 0, 3, and 5. In the present study, we used a hairless but euthymic SPF mouse strain and a slightly different challenge protocol to produce results similar to those obtained by the above group. To colonize hairless mice, we routinely deprived the mice of food for 2 days and challenged them twice, on consecutive days, using fresh clinical isolates of H. pylori. The two-day fast was expected to bring down the number of antagonistic microflora, specifically Lactobacillus spp. (28). Our own unpublished observations showed that in mice, an initial Lactobacillus titer of 4.2 × 10⁷ CFU/100 mg of gastric tissue was reduced to 2.0 × 10⁷ CFU after 1 day of fasting and to 7.2 × 10⁶ CFU after the second day of fasting. Moreover, the population level of the Lactobacillus spp. that usually predominate in the stomachs of mice and are attached to epithelial cells (12, 26) may determine the outcome of infection with certain gastric pathogens, including H. pylori. An inverse relationship was found between the population levels of H. pylori and Lactobacillus spp. (12) in the mouse stomach. These data support the logic behind fasting before challenge, as was done in this study, in order to bring down the number of lactobacilli that may inhibit H. pylori colonization.

In our experience, the population of lactobacilli in NS:Hr/ICR hairless mice was at least 50 to 100 times lower than that in conventional BALB/c mice. This may also have contributed to the higher susceptibility of this mouse strain to H. pylori. Thus, seven of seven NS:Hr/ICR mice were infected at weeks 1 and 2 postchallenge with 10⁷ CFU of type I H. pylori, compared to two of seven and one of seven BALB/c mice infected at the same two time points, respectively (Table 1). The infection rates in BALB/c mice at weeks 1 and 2 in the present study were slightly lower than those in an earlier study, where two of four and four of four were infected at weeks 1 and 2, respectively (19). The difference may be related to the number of challenges performed (two in this study and three in the earlier study) and/or the bacterial strain and dose used (10⁷ CFU in this study and 10⁹ CFU in the earlier study). However, we cannot rule out other, yet-unknown genetic properties of NS:Hr/ICR that may account for the relatively high frequency of infection in this mouse strain compared to those in BALB/c or other strains or species tested.

Using the mouse strain NS:Hr/ICR as a colonization model, we examined the ability of different H. pylori strains to colonize the gastric mucosa. The results (Table 2) showed that the laboratory-passaged strains, NSP394 and NSP333, failed to establish gastric infection beyond 1 week postchallenge. All fresh clinical isolates consistently colonized the gastric mucosa, regardless of phenotype. This infection pattern was observed up to the fourth week postchallenge. However, we failed to observe evidence of gastric pathology at 4 and 8 weeks (data not shown) of infection among HE-stained sections of gastric tissue from mice infected with about 10⁷ CFU. We decided, therefore, to increase the challenge dose by centrifuging bacterial cells in brain heart infusion broth culture and using the resulting cell pellet for challenge. This gave a 10- to 100-times-more-concentrated bacterial load for oral dosing. With this inoculum, and with type I or type II clinical isolates of H. pylori as the challenge strain, the NS:Hr/ICR hairless mice were CFU positive from week 2 through week 8 postchallenge (Table 3). A H. pylori count conducted on these mice during week 8 yielded relatively high titers for both the gastric mucus and washed gastric epithelial cell homogenates. The titers recovered were 10^4.8 CFU/100 mg of gastric homogenate for the NSP335 strain (phenotype I) and 10^6.4 CFU/100 mg of gastric homogenate for the NSP355 strain (phenotype II). These titers were significantly different from those obtained with 10⁷ CFU as the challenge inoculum (zero count at 8 weeks postinfection). Clinically, culture-positive mice did not show any sign of digestive disturbance until the end of the observation period. Body weights were similar for both infected mice and noninfected control mice (data not shown).

Gastric colonization of the NS:Hr/ICR mice invariably gave rise to mucus-resident and epithelium-adherent H. pylori. We failed to completely detach adherent cells by extensive washing of gastric tissue, indicating that the physical association was intimate, perhaps involving cytoplasmic invaginations harboring H. pylori cells, as had been observed in a recent study on polarized T₈₄ human intestinal cell monolayers infected in vitro with H. pylori (3). Such adherent forms were shown in that study to be resistant to gentamicin, and susceptibility was restored when the cells were lysed to release the embedded H. pylori organisms from cytoplasmic pouches. This finding may explain the remarkable tenacity with which the bacteria resisted 80 cycles of vortex washing at top speed in the present study. The close association with epithelial cells persisted for as long as 8 weeks, which was the last observation time point in this study. Despite their small number (about 2 to 4% of the total gastric population), such adherent bacterial cells have the potential to induce inflammation because of their proximity to gastric epithelial cells. It will be of interest to evaluate the in vivo susceptibility of such adherent cells to antimicrobial therapy to better explain their role in persistence or relapse of infection after treatment.

Mice given 10⁸ to 10⁹ CFU of H. pylori during challenge exposure were tested for serum antibody with whole-cell lysate of H. pylori as the ELISA capture antigen at 4 and 8 weeks postchallenge. There was a progressive increase of relatively low antibody titers, indicating access to the systemic circulation by H. pylori antigen, possibly in the form of bacterial breakdown products or secretions. At 2 weeks postchallenge, antibody was barely detectable. At this time point, mean ELISA ODs and corresponding standard deviations were 0.142 ± 0.003 (n = 4) for the type I-infected group and 0.180 ± 0.008 (n = 3) for the type II-infected group, which were almost the same as that of the control (0.115 ± 0.002; n = 2). At 8 weeks postchallenge, the mean ELISA ODs and corresponding standard deviations were 0.395 ± 0.280 (n = 4) for the type I-infected group and 0.213 ± 0.007 (n = 4) for the type II-infected group versus 0.113 ± 0.002 (n = 2) for the control. Although the infected groups had higher mean ODs (about twice or more) than the control group, the difference was not statistically significant (P > 0.05).

Mice challenged with the lower dose of 10⁷ CFU of either strain of H. pylori and necropsied at 8 weeks postchallenge did not show any histologic lesions (not shown) in the gastric mucosa. However, mild-to-severe inflammatory cell infiltration indicative of gastritis was observed histologically at the eighth week in three of five mice challenged with the higher dose of 10⁸ to 10⁹ CFU of type I H. pylori but not in the type II-challenged mice (Table 3). Histological examination at the sixth week in the group of mice given the high dose of the type I strain revealed negative results (data not shown), indicating that gastritis emerged at some time between the seventh and eighth weeks postchallenge. The negative histopathologic findings in two of five mice may be technical in origin (half of the glandular stomach was used for bacterial isolation) or may represent a late onset (after 8 weeks) of inflammation. It will be interesting to see how long infection will persist beyond 8 weeks and whether chronic H. pylori infection will give rise to neoplastic changes. A gastric cancer model for H. pylori has been elusive up to this time.

The failure to induce gastritis in mice given a challenge dose of 10⁷ CFU of the type I strain and in those given a challenge dose of 10⁸ to 10⁹ CFU of the type II strain indicates that gastritis is a function of the bacterial load and the phenotype of the infecting strain. The absence of inflammatory cell infiltration in the gastric tissue of mice challenged with the higher dose of type II H. pylori also indicates a causal relationship between inflammatory changes and VacA and/or CagA phenotype, as had been observed in humans (4) and experimental animals (19). Among gastritis-positive mice challenged with the higher dose of type I strain organisms, histologic lesions consisted of mild-to-severe lymphocytic infiltration of the lamina propria and submucosal layer in the glandular region of the stomach (Fig. 1). Gross changes in the mucosa were not evident, nor were they expected, since intact gastric epithelial lining was visualized microscopically. By immunoperoxidase staining with rabbit anti-whole-cell H. pylori as a primary antibody, gastritis-positive tissue sections revealed the presence of bacterial antigen, visible as dark brownish granules in the glandular isthmus, which could often be seen distal to the area of submucosal inflammatory infiltration (Fig. 2). Type II-infected mice had no observable microscopic lesions (Fig. 1C), similar to the control mice (not shown).

View larger version (102K): [in this window] [in a new window]   FIG. 1.   (A and B) HE-stained sections of glandular stomach from hairless mice challenged 8 weeks earlier with 4 × 108 and 1 × 109 CFU of type I H. pylori (NSP335 strain). The tunica submucosae was infiltrated with mostly mononuclear cells (arrows). (C) HE-stained section of stomach from a hairless mouse challenged 8 weeks earlier with 2.6 × 109 and 3 × 109 CFU of type II H. pylori (NSP355 strain). The tunica submucosa, devoid of mononuclear cell infiltration in this mouse, was histologically indistinguishable from those of noninfected mice. TS, tunica submucosae; MM, muscularis mucosae; E, gastric epithelial cells; ME, muscularis externa (the lamina propria is not recognizable at this magnification). (Magnification, ×50.)

View larger version (166K): [in this window] [in a new window]   FIG. 2.   Immunostained section of the glandular area of the stomach of a hairless mouse showing the region distal to the focus of mononuclear cell infiltration shown in Fig. 1A. Immunoperoxidase staining was used, with rabbit anti-whole-cell H. pylori as the primary antibody. The dark granular bodies (arrows) represent H. pylori antigen. S, gastric mucosal surface; E, gastric epithelial cells. (Magnification, ×50.)

The natural susceptibility of a host to infection by a pathogenic organism results from interplay among the host, parasite, and environment, with the net effect of allowing the invading pathogen to occupy a niche inside the targeted host. From this favorable location, the pathogen may multiply and cause cell or organ damage and/or release other virulence factors. The present study addressed this issue in the development of a mouse model for H. pylori by prechallenge reduction of lactobacilli by dietary stress, selection of a susceptible strain of mice, and identification of fresh clinical isolates as colonizing strains at suitable challenge doses. The use of a convenient small animal model, such as the NS:Hr/ICR mouse, for evaluation of therapy aimed at eliminating mucus-resident and epithelium-adherent H. pylori cells should prove useful in the quest to eliminate this common bacterial infection from humans worldwide.