INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Helicobacter pylori colonizes the stomachs of approximately one-half of the world's population (6). The prevalence of infection is higher in developing countries (70 to 90%) than in the United States and Western Europe (25 to 50%) (6, 20). Infection with this bacterium results in chronic superficial gastritis which, in some cases, will progress to peptic ulceration, gastric carcinoma, and MALT lymphoma (reviewed in reference 6). Eradication of H. pylori results in ulcer healing and a drastic reduction in the rate of ulcer recurrence (6, 9).

In France, the triple regimen of amoxicillin, clarithromycin, and a proton pump inhibitor is recommended for use for eradication of H. pylori (4). The 5-nitroimidazole metronidazole is used as an alternative to either of the two antibiotics in cases of resistance (particularly to clarithromycin) or patient allergy (4). Despite this, the prevalence of resistance to metronidazole in H. pylori is relatively high in France (approximately 25%) (7; N. Broutet, F. Guillon, E. Sauty, D. Lethuaire, and F. Mégraud, Program Abstr. 18th Interdisc. Meet. Anti-Infect. Chemother., abstr. 130/P1, 1998), similar to what is observed in other western European countries or in the United States. This prevalence is far higher in developing countries and in certain immigrant populations, suggesting that metronidazole resistance is associated with the prior use of nitroimidazoles to treat anaerobic and parasitic infections (1, 6, 7). It has recently been unequivocally demonstrated that previous exposure of H. pylori to metronidazole in vivo results in the emergence of resistant strains (10).

Recently, resistance to metronidazole in H. pylori was demonstrated to be associated with mutational inactivation of the rdxA gene, which encodes an oxygen-insensitive NADPH nitroreductase (8). This enzyme reduces metronidazole to active metabolites that are directly toxic to the bacterium (8). The authors described certain frameshift and missense mutations in rdxA that were associated with resistance in a small number of clinical strains originating from Canada, Lithuania, and Peru (8). More recently, Jenks et al. (11), using an H. pylori mouse model, demonstrated that mutations in the rdxA gene were implicated in the development of metronidazole resistance in 25 of 27 isolates derived by treating a single metronidazole-susceptible strain with metronidazole.

There is little consensus on methods of susceptibility testing for H. pylori. It has been shown that disk diffusion tests for metronidazole correlate poorly with the results obtained by agar dilution determination of the MIC (16). The E-test method is more accurate, but important variations are observed depending on the medium, the inoculum, and the duration of incubation used (16). Furthermore, because H. pylori is a fastidious, slow-growing organism, these tests are difficult and time-consuming to perform. A better understanding of the underlying mechanisms of metronidazole resistance might lead to the development of molecularly based methods, as has been the case for clarithromycin susceptibility testing (15, 18).

The aim of this study was to examine further the role of the rdxA gene in the development of metronidazole resistance by clinical strains of H. pylori isolated from patients of French or North African origin. Because the sequences of rdxA genes from unrelated metronidazole-susceptible strains differ by approximately 5% at the nucleotide level, we compared the nucleotide sequences of paired sensitive and resistant clinical isolates. In addition, we also determined the nucleotide sequences of the rdxA genes of a series of individual, nonpaired, resistant strains.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacteria and growth conditions. Primary cultures of H. pylori from antral biopsies were stored at 80°C in glycerol-supplemented brain heart infusion (bioMérieux, Marcy l'Etoile, France). In order to identify metronidazole-resistant strains, these cultures were diluted and subcultured onto blood agar medium (Blood Agar Base no. 2; Oxoid, Lyon, France) supplemented with 10% horse blood (bioMérieux) and the Dent selective supplement (Oxoid). One hundred individual colonies from each biopsy were then subcultured in parallel onto the same medium with and without 8 µg of metronidazole (Rhône-Poulenc Rorer, Vitry-sur-Seine, France) per ml. The plates were incubated at 37°C under microaerobic conditions in an anaerobic jar (Oxoid) with a hydrogen and carbon dioxide generator (Oxoid) and a catalyst (Oxoid).

Metronidazole susceptibility testing. Susceptibility to metronidazole was assessed by the E-test method (AB Biodisk, Solna, Sweden) performed according to the instructions of the manufacturer and using Mueller-Hinton agar (Oxoid) supplemented with 10% horse blood (bioMérieux) and a cell suspension calibrated at 3 McFarland units. Plates were read after three days of incubation at 37°C as described above. Strains were considered resistant to metronidazole if the MIC was 8 µg/ml (23).

Genotyping of paired susceptible and resistant (S-R) isolates obtained from the same patient. Target chromosomal DNA was extracted from H. pylori strains by using the QIAamp Tissue Kit (Qiagen, Courtaboeuf, France). Random amplified polymorphic DNA (RAPD) analysis, with a single 11-bp oligonucleotide (5'-AGTTCAGCCAC-3'), was used to confirm the genetic relatedness of paired S-R isolates (13). PCR was performed in a 100-µl volume containing 10 mM Tris-HCl (pH 8.3) (Boehringer Mannheim, Meylan, France), 1.5 mM MgCl₂, 50 mM KCl, 0.25 mM each deoxynucleotide (Pharmacia Biotech, Uppsala, Sweden), 1 µl of DNA sample, 2.5 U of Taq DNA polymerase (Boehringer Mannheim), and 3 µM oligonucleotide primer. Amplification was carried out using a GeneAmp PCR System 9600 thermal cycler (Perkin-Elmer Biosystems, Courtaboeuf, France) programmed for 45 consecutive cycles consisting of a denaturation step of 94°C for 1 min, a primer-annealing step of 36°C for 1 min, and an extension step of 72°C for 2 min. Amplification products were analyzed by electrophoresis on a 2% agarose gel. Isolates were considered to be genetically related if their profiles did not differ by more than one band.

Amplification by PCR and direct sequencing of the rdxA gene. Two pairs of oligonucleotide primers, rdx1 (5' [position 1014242 in the H. pylori genome database {21}]-CGTTAGGGATTTTATTGTATGCTACA-[position 1014216] 3')-rdx2 (5' [position 1013751]-CCCCACAGCGATATAGCATTGCT-[position 1013774] 3') and rdx3 (5' [position 1013856]-GTTAGAGTGATCCCCTCTTTTGCTCA-[position 1013830] 3')-rdx4 (5' [position 1013451]-CACCCCTAAAAGAGCGATTAAAAC-[position 1013475] 3'), were used to amplify two overlapping PCR products (of 491 and 405 bp, respectively) that constituted a total of 789 bp containing the entire 630-bp rdxA gene (11). PCR was performed using the same reaction mixture as described above except that the oligonucleotide primers were used at a concentration of 2 µM. PCR was carried out with the same apparatus programmed for 35 consecutive cycles of 95°C for 1 min, 48°C (491-bp fragment) or 51°C (409-bp fragment) for 1 min, and 72°C for 1 min. Amplification products were analyzed by electrophoresis on a 2% agarose gel. Nucleotide sequences of the PCR products obtained were determined directly on both strands by the method of Sanger et al. (17) with a Abiprism 377 apparatus (Perkin-Elmer Biosystems), using the four oligonucleotide primers described above.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of S-R pairs of isolates. By subculturing individual colonies originating from primary cultures of antral biopsies onto medium with and without 8 µg of metronidazole per ml, we identified 15 metronidazole-resistant cultures originating from French patients with gastroduodenal ulcer (six patients), nonulcer dyspepsia (four patients), gastric cancer (one patient), mucosa-associated lymphoid tissue (MALT) lymphoma (one patient), and undetermined disease (three patients). We also identified 11 resistant cultures from North African patients with gastroduodenal ulcer (seven patients), nonulcer dyspepsia (two patients), and undetermined disease (two patients). None of these French or North African patients had received a metronidazole-containing eradication regimen. Eight of the 15 cultures originating from French patients contained metronidazole-resistant strains only, while seven consisted of metronidazole-susceptible and -resistant isolates. For the 11 cultures originating from North African patients, five contained metronidazole-resistant strains only and six contained mixed susceptible and resistant isolates. The MICs for the susceptible strains varied from 0.12 to 1 µg/ml, while those for the resistant strains were 32 µg/ml.

Nucleotide sequences of the rdxA genes for the French S-R pairs and individual resistant strains. Results for the French isolates are shown in Table 1. In the seven S-R pairs and eight resistant strains examined, two PCR-amplified rdxA-containing fragments of the expected size were obtained. A frameshift mutation in rdxA was found in seven cases (in the metronidazole-resistant isolates of three of the S-R pairs and in four resistant strains), resulting either in a truncated protein (six cases, i.e., pairs FP2, FP21, and FP22 and strains FR4, FR47, and FR245) or in an altered C-terminal amino acid sequence (one case, i.e., strain FR7). Five of these mutations were the result of the gain of a single adenine (A) or thymine (T) nucleotide in poly(A) tracts located at nucleotide positions 20 to 25 and 187 to 193. In a further case (pair FP11), the resistant allele of one S-R pair contained deletion of approximately one-third of the gene, and this was associated with a frameshift resulting in a truncated protein.

View larger version (36K): [in this window] [in a new window]   FIG. 1.   Amino acid sequences of rdxA gene products for French and North African matched susceptible-resistant (S/R) pairs of isolates and individual resistant strains. The resultant amino acid changes found in the gene products of the resistant members of the susceptible-resistant pairs are underlined.

Nucleotide sequences of the rdxA genes for the North African S-R pairs and individual resistant strains. Results for the North African strains are shown in Table 2. The rdxA gene was amplified and sequenced for all isolates (six S-R pairs and five resistant strains). A frameshift mutation resulting in truncated protein was found in three resistant strains (strains AR178, ARS, and AR13). For the resistant allele of one of the S-R pairs (pair APB), there was an insertion of a 268-bp DNA fragment in the rdxA gene, resulting in a frameshift mutation and a truncated protein. This fragment had the same size and showed 97.8% identity with a variant of insertion sequence IS605 (called is605) originally described by Censini et al. for the cag pathogenicity island of strain CCUG 17874 (2). is605 is composed of the two arms of IS605 without any associated open reading frames (2).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Goodwin et al. have demonstrated that certain cases of metronidazole resistance in H. pylori can be explained by null mutations in a gene (rdxA) that encodes an oxygen-insensitive nitroreductase (8). In this work, we have examined paired S-R isolates as well as individual resistant strains of H. pylori of French or North African origin for the presence of genetic alterations of rdxA. The prevalence of metronidazole resistance in H. pylori is higher in developing countries than in Western Europe and the United States, and this is likely to be due to prior usage of nitroimidazoles to treat anaerobic and parasitic infections (6, 20). We also wanted to determine if different mechanisms of resistance to metronidazole exist in isolates of H. pylori from different geographic locations, which might possibly reflect either low or high previous exposure to this antibiotic.

There is considerable allelic diversity of the nucleotide sequence of rdxA between different metronidazole-susceptible strains (8). In order to detect changes in rdxA involved in resistance, particularly missense mutations, we studied primary cultures that were mixed with respect to metronidazole susceptibility and resistance. This phenomenon of heteroresistance to metronidazole in H. pylori has been previously described for humans (5, 8, 12, 19, 22) as well as for experimentally infected mice (10). We found that approximately half of the metronidazole-resistant primary cultures that we examined contained isolates that were susceptible and resistant to metronidazole. These data confirm that heteroresistance to metronidazole is frequently encountered in H. pylori. It has recently been shown, in a mouse model of H. pylori infection, that a mixed population of susceptible 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 (10). Furthermore, Weel et al., studying the susceptibility to metronidazole by E-test of 152 H. pylori clinical isolates, found that 37 strains were homogeneously resistant whereas 28 were heteroresistant (22). The phenomenon of heteroresistance has important implications for the accuracy of in vitro susceptibility testing of H. pylori, and it is important that multiple colonies are tested.

Genotyping of the coinfecting susceptible and resistant isolates by RAPD analysis showed that these isolates were genetically related in all cases. This confirms the results of Goodwin et al. (8), namely, that resistance to metronidazole in H. pylori arises essentially by de novo mutation and appears not to be due to the coexistence of sensitive and resistant unrelated strains or to horizontal gene transfer between unrelated strains.

Goodwin et al. showed that both frameshift mutations resulting in a truncated protein and missense mutations were implicated in the development of metronidazole resistance by clinical isolates of H. pylori (8). In their work, the majority of changes in the rdxA gene were due to missenses mutations (8). In contrast, in the strains generated using the mouse model of infection, 78% (21 of 26) of the resistant strains contained a frameshift mutation resulting in a stop codon; they were found throughout the rdxA gene (11). In our work, frameshift mutations were implicated in 38% (10 of 26) of the cases, and the resulting stop codons were also at various locations. More than half (6 of 10) of these frameshift mutations occurred within two poly(A) tracts, located at nucleotide positions 20 to 25 and 186 to 192, confirming that slipped-strand mispairing may be an important mechanism in the regulation of the expression of this gene (11, 14). Interestingly, the two poly(A) tracts which were particularly unstable were also implicated in resistance in 61% (8 of 13) of the cases of frameshift mutations described by Jenks et al. (11).

In addition, we showed that a truncated RdxA protein could also result not only from a frameshift mutation but also either from deletion of a large portion of the gene (pair FP11) or from insertion of a 268-bp variant of insertion sequence IS605 (pair APB). The resistance to metronidazole of H. pylori strain NCTC 11638 has also been shown to be due to insertion into rdxA of another variant of IS605 made up of the 41-bp inverted repeat of this transposon (Y. J. Debets-Ossenkopp, A. Goodwin, C. M. J. E. Vandenbroucke-Grauls, R. G. J. Pot, D. E. Berg, P. S. Hoffmann, and J. G. Kusters, Abstr. 11th Workshop Eur. Helicobacter pylori Study Group, abstr. 01/2, 1998). We also identified another novel genetic alteration of rdxA that has yet not been reported in resistant isolates; the deletion of six nucleotides led to the loss of two amino acids in pair APK and to the development of resistance to metronidazole.

The comparison of the nucleotide sequences of the rdxA genes of the S-R pairs of isolates permitted us to identify missense mutations that resulted in single amino acid changes in the resistant strain in 6 out of 13 cases: Cys16His, Ser43Leu, and Ser79Ile for the French pairs and Thr58Ala, Ala67Val, and Ala187Asp for the North African ones. Only one of these amino acid substitutions (Ala67Val) had previously been described as being associated with metronidazole resistance (11). The Ser43Leu substitution was at a position within a region (SPSSYNTQPWHFVMV, amino acids 43 to 57) that is highly conserved among classical oxygen-insensitive NADPH nitroreductases (CNRs) of gram-negative bacteria (8). Another substitution (Thr58Ala) was at a position just downstream of this conserved region. Resistant strains harboring multiple missense mutations in rdxA have previously been described (8, 11), but this was not observed for the isolates that we studied. The MIC of metronidazole for all of these resistant mutants was 32 µg/ml, as determined by the E-test method, suggesting that there was no clear correlation between the types of mutations identified and the level of resistance, similar to what has been found by Jenks et al. (11).

Interestingly, the Cys16His substitution found for pair FP25 might not be involved in resistance, because a histidine at position 16 was also found for the two members of another S-R pair of isolates, pair APA (Fig. 1). However, as the peptide sequences of the RdxA proteins from FP25 and APA were not identical (Fig. 1), the His16 found in APA may be compensated for by another amino acid change to restore metronidazole susceptibility.

One to three putative missense mutations were also identified in the rdxA genes of 6 of the 13 individual resistant strains examined. Their significance is unclear, as the sequence of the susceptible isogenic isolate was not available for comparison. Thus, the role of rdxA in resistance could not be clearly demonstrated for approximately half of the individual resistant strains that we tested. This shows that more data have to be collected concerning the possible variations of the wild-type sequence of rdxA. However, one of these putative mutations (Ala80Thr) has previously been implicated in resistance to metronidazole (8), and another (Cys87Tyr) involves the loss of Cys87, an amino acid conserved in all CNR proteins of gram-negative bacteria so far characterized (8).

In total, we identified a variety of genetic alterations in rdxA associated with metronidazole resistance, most of which had not been described before. Because of the diversity of genetic changes involved (missense or frameshift mutations in various parts of the gene, deletion, and insertion of transposable elements), the possibility of detecting metronidazole resistance by molecular methods seems compromised. Comparison of the genetic alterations found in French strains with those in North African strains did not reveal significant differences, suggesting similar mechanisms for inactivation of rdxA in strains of different geographical origins.

Importantly, in 1 of the 13 S-R pairs examined (pair APA), the rdxA gene did not appear to be involved in resistance. Another study has also identified metronidazole-resistant strains in which the rdxA gene appears to be intact (11). One possible explanation is the presence of mutations in the promoter region of rdxA or in a gene responsible for the regulation of the expression of rdxA. The absence of expression of rdxA has recently been described for one resistant strain, although the genetic mechanism involved has not yet been characterized (D. H. Kwon, M. Kato, R. Reddy, M. Osato, D. Y. Graham, and F. A. K. El-Zaatari, Abstr. 99th Gen. Meet. Am. Soc. Microbiol., abstr. A-24, 1999). Another possibility is that alterations in other genes could be implicated in metronidazole resistance of H. pylori. The frxA gene, encoding an NAD(P)H flavin reductase which is also a CNR homologue and has 25% protein-level identity with RdxA, may be implicated (8). Diminished intracellular accumulation of metronidazole by active efflux or reduced uptake or overexpression of RecA leading to increased DNA repair (3) are other possible mechanisms that may result in the development of metronidazole resistance in some strains of H. pylori.