ABSTRACT
The Helicobacter pylori SS1 mouse model was used to characterize the development of resistance in H. pyloriafter treatment with metronidazole monotherapy and to examine the effect of prior exposure to metronidazole on the efficacy of a metronidazole-containing eradication regimen. Mice colonized with the metronidazole-sensitive H. pyloriSS1 strain were treated for 7 days with either peptone trypsin broth or the mouse equivalent of 400 mg of metronidazole once a day or three times per day (TID). In a separate experiment,H. pylori-infected mice were administered either peptone trypsin broth or the mouse equivalent of 400 mg of metronidazole TID for 7 days, followed 1 month later by either peptone trypsin broth or the mouse equivalent of 20 mg of omeprazole, 250 mg of clarithromycin, and 400 mg of metronidazole twice a day for 7 days. At least 1 month after the completion of treatment, the mice were sacrificed and their stomachs were cultured for H. pylori. The susceptibilities of isolates to metronidazole were assessed by agar dilution determination of the MICs. Mixed populations of metronidazole-resistant and -sensitive strains were isolated from 70% of mice treated with 400 mg of metronidazole TID. The ratio of resistant to sensitive strains was 1:100, and the MICs for the resistant strains varied from 8 to 64 μg/ml. In the second experiment, H. pylori was eradicated from 70% of mice treated with eradication therapy alone, compared to 25% of mice pretreated with metronidazole (P < 0.01). Mice still infected after treatment with metronidazole and eradication therapy contained mixed populations of metronidazole-resistant and -sensitive isolates in a ratio of 1:25. These results demonstrate thatH. pylori readily acquires resistance to metronidazole in vivo and that prior exposure of the organism to metronidazole is associated with failure of eradication therapy. H. pylori-infected mice provide a suitable model for the study of resistance mechanisms inH. pylori and will be useful in determining optimal regimens for the eradication of resistant strains.
Helicobacter pylori is a gram-negative, microaerobic, spiral bacterium that colonizes the stomachs of approximately half the world’s population (44). Infection with H. pylori is associated with severe gastrointestinal disease, including peptic ulceration, and eradication of the organism from the stomach facilitates duodenal ulcer healing and reduces ulcer relapse (16, 28). Although the 5-nitroimidazole metronidazole is an important component of many currently used H. pylori eradication regimens, resistance to this class of antibiotics is relatively common. It has been estimated that 10 to 30% of clinical strains isolated in Western Europe and the United States are metronidazole resistant (9, 11, 13, 36), and this prevalence is far higher in developing countries and in certain immigrant populations (2, 9, 11, 13). Metronidazole is commonly used to treat anaerobic and parasitic infections, and there is epidemiological evidence that metronidazole resistance in H. pyloriis associated with prior use of this antibiotic in certain patient groups (2, 4, 11, 13, 37). However, previous use of metronidazole is frequently not reported by patients (11), and consequently the development of resistance inH. pylori after metronidazole monotherapy is poorly characterized.
The relationship between previous exposure to metronidazole, the development of resistance to this antibiotic, and the eventual outcome of H. pylorieradication regimens remains unclear (30, 36). Many studies have demonstrated that infection with metronidazole-resistant strains is an important predictor of failure of metronidazole-containing eradication regimens, even when quadruple therapy is used (3, 5-7, 17, 19, 22, 31, 34, 37, 40-42, 47). However, there is also evidence that resistance to metronidazole can be partly overcome when antisecretory drug-based triple or quadruple therapies are used, and such regimens may achieve eradication in around 75% of patients infected with resistant strains (3, 6, 7, 17, 19, 22, 30, 32, 37, 40, 41, 48).
The aims of this study were to use a well-validated mouse model ofH. pylori infection (24) to (i) characterize the development of resistance in H. pylori after treatment with metronidazole monotherapy at doses normally used to treat anaerobic and parasitic infections and (ii) examine the impact of prior exposure of H. pylori to metronidazole on the efficacy of a metronidazole-containing eradication regimen.
MATERIALS AND METHODS
Bacteria and growth conditions. H. pylori SS1 is a mouse-adapted strain originally isolated from a patient with peptic ulcer disease (24). H. pylori SS1 was routinely cultured on a blood agar medium (Blood Agar Base no. 2; Oxoid, Lyon, France) supplemented with 10% horse blood (bioMérieux, Marcy L’Etoile, France) and the following antibiotics: 10 μg of vancomycin (Dakota Pharmaceuticals, Creteil, France)/ml, 2.5 IU of polymyxin (Pfizer Laboratories, Orsay, France)/liter, 5 μg of trimethoprim (Sigma Chemicals, Saint-Quentin Fallavier, France)/ml, and 4 μg of amphotericin B (Bristol-Myers Squibb, Paris, France)/ml. The plates were incubated at 37°C under microaerobic conditions in an anaerobic jar (Oxoid) with a carbon dioxide generator (CampyGen; Oxoid) without a catalyst. For the selection of metronidazole-resistant colonies and their subsequent subculture, the medium was additionally supplemented with metronidazole at 8 μg/ml (Sigma).
To determine viable counts of H. pylori, samples to be tested were serially diluted in sterile saline and then plated in duplicate onto blood agar plates supplemented with either 10% horse blood or fetal calf serum (Gibco BRL, Cergy Pontoise, France), 10 g of agar (Bacteriological Agar no. 1; Oxoid)/liter, 200 μg of bacitracin/ml, and 10 μg of nalidixic acid (Sigma)/ml. After 5 days of incubation, colonies with H. pylori morphology were identified according to standard criteria (morphology on Gram staining and the presence of catalase, oxidase, and urease enzyme activities) and enumerated (12).
Infection of mice with H. pylori SS1.Six-week-old specific-pathogen-free Swiss mice (Centre d’Elevage R. Janvier, Le-Genest-St-Isle, France) were housed in polycarbonate cages in isolators and fed a commercial pellet diet with water ad libitum. All animal experimentation was performed in accordance with institutional guidelines. Mice were inoculated intragastrically with a suspension of H. pylori SS1, which had been harvested directly from 48-h plate cultures into peptone trypsin broth (Organotéchnique, La Courneuve, France). Each animal was administered a single 100-μl aliquot of an inoculating suspension of 105 CFU/ml (equivalent to 100 times the 100% infectious dose [12]). This was administered with polyethylene catheters (Biotrol, Paris, France) attached to 1-ml disposable syringes. A control group of mice (n = 10) was given peptone trypsin broth alone.
Three groups of mice were used in the treatment study. The mice in group 1 were used in a pilot study to determine if metronidazole resistance was induced in H. pylori after treatment with two different doses of metronidazole monotherapy. Group 2 animals were used to further characterize the acquisition of resistance after metronidazole monotherapy. The mice in group 3 were used to examine the effect of prior exposure to metronidazole on the efficacy of a metronidazole-containing H. pylorieradication regimen.
Antimicrobial chemotherapy.All solutions were administered intragastrically in a final volume of 100 μl via polyethylene catheters as previously described. Seven weeks after infection, theH. pylori-colonized mice in group 1 were treated for 7 days either with peptone trypsin broth (n = 5) or 0.171 mg of metronidazole (Rhône-Poulenc Rorer, Vitry sur Seine, France) (n = 5) in a single daily dose or with peptone trypsin broth (n = 4) or 0.171 mg of metronidazole (n = 5) three times daily (TID) (see Table 1). The dose of 0.171 mg in a 30-g mouse is the body weight equivalent of a 400-mg dose in a 70-kg human. Fifteen weeks after infection, the H. pylori-colonized mice in group 2 were given either peptone trypsin broth (n = 10) or 0.171 mg of metronidazole TID for 7 days (n = 10) (see Table 1).
The H. pylori-colonized mice in group 3 were administered two treatment regimens (see Table 2). Treatment 1 was administered 15 weeks after infection and consisted of either peptone trypsin broth (n = 20) or 0.171 mg of metronidazole TID for 7 days (n = 17) . Treatment 2 was administered 1 month after the completion of treatment 1 and consisted of either peptone trypsin broth (n = 19) or the mouse body weight equivalent of a recommendedH. pylori eradication regimen of 20 mg of omeprazole (0.0086 mg; Astra Hassle AB, Molndal, Sweden), 250 mg of clarithromycin (0.107 mg; Abbott Laboratories, Saint-Rémy-sur-Avre, France), and 400 mg of metronidazole (0.171 mg) twice daily for 1 week (n = 18) (10).
Assessment of H. pylori infection in mice.Colonization with H. pylori was assessed at least 1 month after the completion of each treatment regimen as recommended by recent guidelines (46). The animals were sacrificed, the stomach of each mouse was removed, and serum was recovered in microtubes (Sarstedt France, Orsay, France). The presence ofH. pylori infection was determined by biopsy urease, quantitative culture, and serology. Stomachs were washed in physiological buffered saline and divided longitudinally into tissue fragments so that each fragment contained cardia, body, and antrum. For each stomach, one fragment was immediately placed in urea-indole medium and another was placed in peptone trypsin broth. The presence of urease activity in tissue fragments was detected in urea-indole medium incubated for 24 h at room temperature (12). For the performance of quantitative bacterial cultures on stomach samples, tissue fragments were homogenized in peptone trypsin broth by using disposable plastic grinders and tubes (PolyLabo, Strasbourg, France). The homogenates were serially diluted in sterile saline and plated directly onto blood and serum plates for enumeration and onto a selective plate containing 8 μg of metronidazole/ml. To increase the sensitivity of detection of metronidazole-resistant strains, all colonies that grew on the two enumeration plates were pooled and subcultured onto plates containing 8 μg of metronidazole/ml. H. pylori colonies were identified according to standard criteria and enumerated as described above.
Serum samples were tested for H. pyloriantigen-specific immunoglobulin G antibody by a previously described enzyme-linked immunosorbent assay technique (12). Briefly, 96-well Maxisorb plates (Nunc, Kamstrup, Denmark) were coated with 25 μg of a sonicated whole-cell extract of H. pyloriSS1. Serum samples were diluted 1:100 and were added in 100-μl aliquots to coated microtiter wells. To allow for nonspecific antibody binding, samples were also added to uncoated wells. Bound H. pylori-specific antibodies were detected by using biotinylated goat anti-mouse immunoglobulin and streptavidin-peroxidase conjugate (Amersham, Les Ulis, France).
The readings for uncoated wells were subtracted from those of the respective test samples. A cutoff value was determined from the mean optical density value ± 2 standard deviations for the corresponding samples from naive uninfected mice. Samples with optical density readings greater than this cutoff value were considered positive for H. pylori-specific antibodies.
Analysis of isolates.Five colonies identified as metronidazole-resistant H. pylori were selected from each mouse for further analysis. The susceptibility to metronidazole of each colony was assessed by agar dilution determination of the MIC. Inoculates yielding 104CFU/spot were inoculated onto plates of IsoSensitest agar (Oxoid) enriched with 10% horse blood containing doubling dilutions of metronidazole. The MIC was defined as the lowest concentration of metronidazole inhibiting growth when the plates were read after 72 h of incubation under microaerobic conditions (generated as described above) at 37°C. Isolates were considered resistant to metronidazole if the MIC was ≥8 μg/ml (46). To assess if resistance to metronidazole was stable in vitro, resistant strains were subcultured three times on nonselective media before redetermination of the MIC.
To quantify the proportion of metronidazole-resistant and -sensitive colonies isolated from mice that received each of the different combinations of treatments in group 3, 200 colonies isolated on nonselective blood agar plates were subcultured in parallel onto media with and without metronidazole (8 μg/ml). The ratio of metronidazole-resistant to metronidazole-sensitive colonies was assessed after 72 h of incubation under microaerobic conditions at 37°C.
Intracage transmission of H. pylori.Two groups of animals were used in order to exclude the possibility of transmission of H. pylori between mice housed in the same cage. To study short-term transmission, two mice infected withH. pylori were kept in the same cage as six uninoculated mice for 9 weeks. In a further experiment, two mice infected with H. pylori were kept in the same cage as eight uninoculated mice for the period covered by the treatment studies (7 months).
Statistical analysis.Differences in the eradication rates between the groups of mice were determined by the chi-square test. A P value of <0.05 was considered significant.
RESULTS
Development of metronidazole resistance after metronidazole monotherapy.In group 1, quantitative culture of gastric tissue samples 1 month after completion of treatment was positive for H. pylori in 18 of 19 mice (Table 1). The bacterial counts obtained ranged from 2.0 × 103 to 1.8 × 106 CFU/g of tissue and were similar in mice treated with peptone trypsin broth and those treated with metronidazole. H. pylori was not cultured from one mouse treated with metronidazole TID. The H. pylori-specific humoral response in this mouse was of a magnitude similar to that in other infected mice, suggesting that infection had been established and subsequently eradicated by metronidazole in this animal. A mixed population of metronidazole-resistant and metronidazole-sensitive H. pyloristrains was isolated from one mouse treated with 0.171 mg of metronidazole once daily and from two mice treated with 0.171 mg TID (Table 1). The MIC for all the resistant strains tested from these three mice (n = 15) was 32 μg/ml (compared to the MIC for the parental SS1 strain, which was 0.064 μg/ml).
Development of metronidazole resistance after metronidazole monotherapy
In group 2, 19 of 20 mice were infected with H. pylori1 month after the completion of treatment (Table 1). The H. pylori bacterial loads were similar in mice treated with peptone trypsin broth and those treated with metronidazole and ranged from 4.0 × 103 to 3.1 × 106CFU/g of tissue. Infection was eradicated from one mouse treated with metronidazole TID (Table 1). A mixture of metronidazole-resistant and metronidazole-sensitive colonies were isolated from the other nine mice treated with 0.171 mg of metronidazole TID. The MICs for the resistant strains isolated from these nine mice (n = 45) ranged from 8 to 64 μg/ml (Table 1). Strains with varying degrees of resistance to metronidazole were isolated from seven of these nine mice. The MIC for each resistant isolate was unchanged after three consecutive subcultures on nonselective medium. No resistant bacteria were isolated from the mice treated with peptone trypsin broth.
Prior exposure to metronidazole and the efficacy of eradication therapy.In group 3, none of the 10 mice inoculated with peptone trypsin broth were infected with H. pylori2 months after the completion of treatment 2 (Table2). In contrast, quantitative culture of gastric tissue samples was positive for H. pylori in all 10 SS1-inoculated mice that were treated twice with peptone trypsin broth (Table 2), with bacterial counts of 3.1 × 104to 1.0 × 107 CFU/g of tissue. No metronidazole-resistant strains were isolated from these mice. H. pylori was cultured from seven of nine SS1-infected mice that received metronidazole monotherapy followed by peptone trypsin broth (Table 2). Mixed populations of metronidazole-resistant and metronidazole-sensitive H. pylori was isolated from six of these mice. The range of MICs for the resistant strains tested (n = 30) was 16 to 32 μg/ml, and the ratio of metronidazole-resistant to -sensitive isolates was 1:100 (Table 2).
Effect of prior exposure to metronidazole on the efficacy of H. pylori eradication therapy
H. pylori was eradicated by a recommended triple-therapy regimen (omeprazole, clarithromycin, and metronidazole) from 25% of mice pretreated with metronidazole compared to 70% of mice not pretreated with this antibiotic (P < 0.01) (Table 2). The bacterial counts in the three mice still infected after treatment with peptone trypsin broth and eradication therapy were similar to those observed in nontreated mice (between 6.5 × 104 and 5.3 × 106 CFU/g of tissue). Two of these mice were infected by mixed populations of metronidazole-resistant and -sensitive isolates. The MIC for all the resistant isolates tested (n = 10) was 16 μg/ml, and the ratio of metronidazole-resistant to -sensitive isolates was <1:200 (Table 2). The stomachs of the six mice still infected with H. pylori after receiving both metronidazole monotherapy and eradication treatment contained 1.9 × 105 to 1.0 × 107CFU/g of tissue. All six of these mice harbored mixed populations of metronidazole-resistant and -sensitive isolates. The range of the MICs for the resistant isolates tested (n = 30) was 16 to 32 μg/ml, and the ratio of metronidazole-resistant to -sensitive isolates was 1:25 (Table 2). All MICs were unchanged after three consecutive subcultures on nonselective medium. At the time of sacrifice (2 months), serological testing was not predictive of successful eradication of H. pylori.
Intracage transmission of H. pylori.Transmission from H. pylori-colonized mice to uninoculated mice was not observed. At 9 weeks and 7 months, the two infected mice were confirmed as colonized with H. pylori SS1 (by culture, biopsy urease, and serology), while the uninoculated mice remained uncolonized.
DISCUSSION
Acquired resistance of H. pylori to metronidazole and other 5-nitroimidazole drugs has been reported worldwide. There is epidemiological evidence suggesting that variation in the use of metronidazole is responsible for the uneven distribution of resistance among different populations. It has been proposed that the high level of resistance in developing countries is associated with the prior use of metronidazole to treat parasitic infections (2, 4, 11, 13). It has also been suggested that the higher prevalence of primary metronidazole resistance found inH. pylori isolated from female patients correlates with the use of metronidazole to treat bacterial vaginosis and other infections of the female genital tract (4, 11, 13, 37). However, many of these studies have been unable to demonstrate a direct association between prior use of metronidazole and the presence of resistant strains, presumably because such use often goes unrecognized by the patient (11, 13, 37). Although resistance to metronidazole in H. pylori has been shown to arise readily in vitro (18, 43), the development of resistance after metronidazole monotherapy in vivo remains poorly characterized. Given the potential implications of metronidazole resistance for the selection of optimal eradication regimens, this is an important phenomenon calling for further study.
The emergence of a single metronidazole-resistantH. pylori strain after exposure to metronidazole was recently reported in a euthymic mouse model (29). However, the relevance of this model to human infection can be questioned, particularly because only low levels of colonization occur (24). In another study, metronidazole-resistant bacteria were isolated after treatment of H. pylori-infected gnotobiotic piglets (23). Although this model mimics many aspects of human H. pylori disease, it is necessary, for practical reasons, to use short study periods (3 weeks). In addition, the parental H. pylori strain used to colonize piglets (26695) already possesses intermediate resistance to metronidazole (MIC, 6.25 μg/ml; intermediate resistance has been defined as a MIC of 4 to 8 μg/ml [47]). It has been demonstrated that immunocompetent mice inoculated with H. pylori SS1 develop chronic infection (≥16 months), with gastric bacterial loads and host inflammatory responses similar to those found in humans (12, 24). We therefore selected the SS1 H. pylorimouse model for the study of the emergence of acquired resistance to metronidazole in this bacterium in vivo and for the examination of the effect of this resistance on the efficacy of a metronidazole-containing H. pylorieradication regimen. All experimentation was performed on long-term-colonized mice and, as far as was possible, in accordance with recent guidelines for clinical trials in H. pyloriinfection (46).
Using the H. pylori SS1 mouse model, we have demonstrated that H. pylori readily develops stable resistance to metronidazole in vivo after treatment with metronidazole monotherapy. In total, 73% of mice in groups 1 and 2 harbored resistant strains after receiving the TID treatment regimen. In a separate experiment, 67% of mice in group 3 that were treated with metronidazole TID followed by peptone trypsin broth developed resistant strains. The high numbers of mice colonized with resistant isolates cannot be explained by transmission of metronidazole-resistant strains between mice. We demonstrated that transmission ofH. pylori SS1 does not occur between mice housed in the same cage, and similar results have recently been presented by other workers (26). These results provide direct evidence that exposure of a clonal population of H. pylori to doses of metronidazole normally used to treat parasitic and anaerobic bacterial infections does result in the emergence of resistance to this antibiotic. Moreover, the development of resistance appears to occur at a high frequency at the individual level. The fact that H. pylori survives beneath the mucous layer and in the gastric pits in both the human and murine stomachs may explain its propensity to develop resistance to metronidazole (21, 24). The antibiotic penetrates such sites poorly, exposing the bacterium to sublethal doses and encouraging the development of resistance.
The phenomenon of heteroresistance, in which susceptible and resistant bacteria may be isolated from the same patient, has previously been described for H. pylori (8, 21, 39, 45). Our study clearly demonstrates that a mixed population of sensitive and resistant bacteria may arise after exposure of a clonal, metronidazole-susceptible strain of H. pylori to either metronidazole monotherapy or a metronidazole-containing eradication regimen. Accurate determination of the proportion of resistant strains demonstrated that after metronidazole monotherapy, the ratio of metronidazole-resistant to metronidazole-sensitive isolates was 1 in 100. This ratio rose to 1 in 25 in mice treated with both metronidazole monotherapy and eradication treatment, indicating that repeated exposure to metronidazole in vivo selects for a resistant population of H. pylori in the stomach. It is recognized that current methods of susceptibility testing may underestimate the frequency of metronidazole resistance in patients coinfected with metronidazole-resistant and -sensitive H. pylori strains (8, 45). If heteroresistance after treatment with metronidazole is confirmed to occur frequently in H. pylori-infected patients, this may have important implications for the susceptibility testing of this organism.
We also observed that, despite the fact that they originated from the same metronidazole-sensitive parental strain, the degree of susceptibility to metronidazole of the emergent resistant isolates varied, with MICs ranging from 8 to 64 μg/ml. Recently, resistance to metronidazole inH. pylori was demonstrated to be associated with mutational inactivation of the rdxA gene, which encodes an oxygen-insensitive NADPH nitroreductase (14). This model will allow us to evaluate the contribution of rdxA to the development of resistance in these strains and to evaluate if other potential resistance mechanisms might exist in H. pylori.
Clinical trials of the efficacy of omeprazole, clarithromycin, and metronidazole have yielded conflicting results for the successful eradication of metronidazole-sensitive and -resistant strains. While a number of studies have demonstrated that eradication is significantly reduced for metronidazole-resistant strains (3, 19, 22, 27, 33, 35, 47), others have shown that this regimen is equally effective for both sensitive and resistant isolates (1, 25). In this study, prior exposure of H. pylori to metronidazole had a considerable negative influence on eradication; this is the first time that a direct link between prior exposure to metronidazole and the outcome of eradication therapy has been clearly demonstrated. The magnitude of this effect is likely to reflect the high frequency of resistance induced by the pretreatment of mice with metronidazole monotherapy and is consistent with the observation that the effectiveness of metronidazole-containing eradication regimens is dependent on the prevalence of metronidazole-resistant H. pylori in the population (17). In addition, secondary resistance to metronidazole, which is recognized to arise during the course of H. pylori eradication therapy (15, 20, 37, 38), was observed to arise in strains isolated from 20% of mice that received eradication therapy only.
These data establish the H. pylori SS1 mouse model as a suitable system for the study of resistance mechanisms in this organism. It will also be useful for examining factors that influence the efficacy of H. pylori eradication and for determining the optimal eradication regimens for resistant strains.
ACKNOWLEDGMENTS
P. J. Jenks is supported by a Research Training Fellowship in Medical Microbiology from the Wellcome Trust, London, United Kingdom (reference no. 044330). R.L.F. acknowledges the support of the Association Charles Debré. Financial support was provided in part by Pasteur-Mérieux-Connaught (Lyon, France) and OraVax Inc. (Boston, Mass.).
We are grateful to Jani O’Rourke for helpful advice on the treatment protocols used and to Rhône-Poulenc Rorer, Astra Hassle AB, and Abbott Laboratories for the gifts of the pharmaceutical agents used in the treatment protocols.
FOOTNOTES
- Received 9 October 1998.
- Returned for modification 25 November 1998.
- Accepted 25 January 1999.
- Copyright © 1999 American Society for Microbiology
