In association with other antibiotics, metronidazole is largely used in the first-line treatment of H. pylori infections, and an increased incidence of resistant strains has been observed in the last few years, with current rates varying from 17% in Europe to 44% in America. In contrast, almost no data are available concerning metronidazole resistance in H. bizzozeronii and other species of the H. heilmannii s.l. (such as H. salomonis and H. felis). The MICs of metronidazole for six H. bizzozeronii strains of animal origin have been estimated (by the agar dilution method) to range from 1 to 8 μg/mL, suggesting that this antibiotic could be efficiently applied to eradicate H. bizzozeronii infections. However, the analysis of isolates obtained from multiple biopsy samples from the same patient[7, 15] revealed the simultaneous presence of metronidazole-susceptible and resistant variants. These data suggest that the isogenic variation of H. bizzozeronii may lead to the accumulation of heteroresistant phenotypes for metronidazole, resulting in treatment failure.
Metronidazole resistance is a strong predictor of treatment failure for Helicobacter infections when the treatment contains metronidazole. However, due to insufficient information about the genetic background of the resistant phenotype, non-invasive detection of metronidazole resistance is not yet feasible[16, 26]. Several years of investigation have provided evidence that the main causes of metronidazole resistance in H. pylori are mutations that alter the correct function of the nitroreductases rdxA or frxA. However, this resistance has recently been shown to involve more complex changes than the simple inactivation of rdxA or frxA, including intracellular redox potential and global gene regulation[29, 30]. To investigate possible changes in H. bizzozeronii after the acquisition of metronidazole resistance, a comparative genome analysis of susceptible and resistant isogenic strains has been performed. Numerous single nucleotide polymorphisms (SNPs) and insertions or deletions (Indels) were detected in the metronidazole-resistant strain compared to the isogenic susceptible one, indicating that the modification of several genes could participate in the resistant phenotype. However, among the five nitroreductases identified in the H. bizzozeronii CIII-1GEN genome, only the homolog of H. pylori rdxA was affected in the resistant strain, and its C-terminal cysteine-containing conserved region (IACLXALGK) was disrupted. This mutation resulted from the extension (from C8 to C9) of a simple sequence cytosine repeat (SSCR) located in the 3’ region of the gene. A comparative analysis showed that a similar 3' SSCR is also present in the same position in the rdxA homolog of H. heilmannii s.s. but not in H. pylori. In this study, we showed that the extension of the 3′ SSCR of rdxA was the only mutation detected in the majority of in vitro spontaneous H. bizzozeronii metronidazole resistant mutants. In addition, we observed that an H. bizzozeronii ΔrdxA mutant strain showed the same MIC value of metronidazole that was observed in the spontaneous mutants. These data indicate that loss of function mutations in rdxA, and, in particular, the disruption of the conserved region IACLXALGK, are sufficient to produce clinically significant resistance to metronidazole in H. bizzozeronii. These findings extend our knowledge of the types of mutations that affect the functionality of rdxA in Helicobacter spp., and they support the idea that the C-terminal cysteine-containing conserved region plays a critical role in the ability of RdxA to catalyze nitroreduction. Moreover, slipped-strand mispairing of the rdxA 3′SSCR appears to be the most frequent mechanism leading to the inhibition of the H. bizzozeronii metronidazole-nitroreductase activity. Therefore, H. bizzozeronii rdxA represents the first example of a contingency gene being associated with metronidazole resistance.
Simple sequence repeats in Helicobacter and other bacterial genomes can mediate phase variation due to their high mutation rates and reversible mutations, providing a population-based mechanism for stochastic variation in expression of specific genes and rapid adaptation to environmental fluctuations[20, 32]. In this study, we demonstrated that H. bizzozeronii acquires resistance to metronidazole at a mutation rate similar to that described for phase-variable genes in Campylobacter jejuni and other bacterial species. In addition, we observed an increased level of susceptibility to metronidazole in a spontaneous mutant maintained in non-selective conditions after approximately 15 passages, indicating that the resistant phenotype is reversible. Based on these data, it is tempting to speculate that metronidazole resistance in H. bizzozeronii is a phase-variable phenotype due to the contingency nature of rdxA. However, several other unknown mechanisms could lead to metronidazole resistance in H. bizzozeronii, overcoming the stochastic variation. In addition, it is not clear how frequently the reversion of the mutation (C9 to C8) occurs. In fact, we recently detected that the larger fraction of the H. bizzozeronii population colonizing the stomach of a patient maintained the C9 allele in rdxA six months after the end of the therapy. Although it is clear that disruption of the C-terminal cysteine conserved region of rdxA decreases the ability of H. bizzozeronii to catalyze the metronidazole nitroreduction, the effect of this mutation on NAD(P)H-oxidase activity is unknown. Therefore, it may be possible that the same mutation that provides resistance to metronidazole does not alter the physiological activity of rdxA, or paradoxically, induces an increased fitness of the bacterium. This could lead to the fixation of the mutation in the population, as observed in the patient.
In conclusion, H. bizzozeronii, and potentially other species of H. heilmannii s.l., easily acquired clinically significant resistance to metronidazole due to the high mutation rate of SSCR located in the 3′ region of rdxA. Although we observed in vitro reversibility of the phenotype, there is evidence that the mutation can be maintained in vivo even six months after the end of therapy. Therefore, the contingency nature of rdxA should be carefully considered when metronidazole is used in the treatment of H. heilmannii-associated gastritis in humans.