Single and Double Mutations in gyrA but Not in gyrB Are Associated with Low- and High-Level Fluoroquinolone Resistance in Helicobacter pylori

Abstract

In one French hospital the rate of resistance to ciprofloxacin in Helicobacter pylori was 3.3% (2 of 60 strains) in 1999. The six resistant clinical strains (four from 1996 and two from 1999) and three ciprofloxacin-selected single-step mutants studied carried one gyrA mutation but none in gyrB. Clinafloxacin and garenoxacin were the most active fluoroquinolones against these mutants. Occurrence of a second gyrA mutation was associated with high MICs of all fluoroquinolones tested.

Failures of proton pump inhibitor-based triple therapies for eradication of Helicobacter pylori are frequently due to resistance to either clarithromycin or metronidazole. Furthermore, coresistance to both drugs is not exceptional, particularly after failure of first-line therapy (9). Alternative regimens in case of coresistance have thus to be developed. Among fluoroquinolones, ciprofloxacin and levofloxacin demonstrate good in vitro activity against H. pylori (12). In one study, levofloxacin was shown to be an effective alternative to clarithromycin in triple therapies (5). Several other fluoroquinolones, such as clinafloxacin or gemifloxacin, appear to be even more active in vitro than ciprofloxacin against wild-type H. pylori (10, 12).

Acquired ciprofloxacin resistance already exists in H. pylori, but the prevalence of this phenomenon is poorly documented, and no data for France exist. The mechanisms implicated have been examined in only two studies, one concerning clinical strains (11) and the other concerning laboratory mutants (15), linking resistance to the occurrence of a gyrA mutation. It is not known how gyrA mutations affect the activities of the newer fluoroquinolones. The eventual role of gyrB mutations has not yet been examined.

In this work, our goals were to determine the rate of primary resistance to ciprofloxacin in one French hospital, to appreciate the respective roles of gyrA and gyrB mutations in resistance, and to compare the in vitro activity of ciprofloxacin to those of seven newer fluoroquinolones against ciprofloxacin-susceptible strains as well as genetically characterized mutants.

Sixty clinical strains of H. pylori isolated in Henri Mondor hospital in 1999 were screened for resistance to ciprofloxacin (defined by an MIC higher than 1 μg/ml [6]) by plating 10 μl of a suspension calibrated at 3 McFarland units on horse blood-supplemented (10%) Mueller-Hinton agar containing 1 μg of ciprofloxacin per ml. Two strains grew after 72 h of incubation, growth involved the totality of the inoculum for both strains, and the rate of resistance was thus 3.3% (2 of 60 strains). Rates of resistance lower than 10% were also reported from northern, eastern, and southern European countries: The Netherlands (7), Germany (9), Bulgaria (3), Spain (14), and Portugal (4).

We then determined, by the agar dilution method as described previously (8), the MICs of ciprofloxacin and of seven newer fluoroquinolones for nine ciprofloxacin-susceptible clinical strains from 1999 chosen at random, for the two ciprofloxacin-resistant isolates from 1999 (328R and 361R), and for four other resistant strains isolated in 1996 (H14R, H23R, H67R, and HboR). Against susceptible strains, compared to ciprofloxacin, moxifloxacin was 2- to 4-fold less active, levofloxacin was as active, trovafloxacin, gatifloxacin, and garenoxacin were 1- to 4-fold more active, gemifloxacin was 2- to 4-fold more active, and clinafloxacin was 4- to 16-fold more active (Table 1). The resistant strains could be separated in two groups. The first group included four strains (H14R, H23R, H67R, and HboR) for which the MICs of garenoxacin, gemifloxacin, and clinafloxacin remained relatively low: 1 to 2, 1 to 4, and 0.5 to 1 μg/ml, respectively. The relative potency of garenoxacin against these resistant strains was higher than those of other fluoroquinolones (Table 1). The two strains of the second group (328R and 361R) presented a higher level of resistance to all the fluoroquinolones tested except ciprofloxacin (Table 1). Thus, the levels of resistance of these two strains to ciprofloxacin, garenoxacin, gemifloxacin, and clinafloxacin were similar.

TABLE 1.

Phenotypic and genotypic characteristics of fluoroquinolone-susceptible and -resistant strains of H. pylori

Strain	MIC (μg/ml)a of:								Substitution(s)c in GyrA
	CIP	LEV	MOX	GAT	TRO	GAR	GEM	CLI
Clinical strains
321S	0.25	0.5	0.5	0.12	0.25	0.25	0.12	0.03	0
322S	0.25	0.5	1	0.12	0.25	0.25	0.12	0.06	0
323S	0.5	1	2	0.5	0.5	0.25	0.25	0.12	0
325S	0.25	0.5	0.5	0.25	0.25	0.25	0.06	0.06	0
326S	0.25	0.5	0.5	0.12	0.25	0.12	0.06	0.03	0
327S	0.5	0.5	1	0.25	0.25	0.25	0.12	0.06	0
329S	1	0.5	2	0.25	0.25	0.25	0.25	0.06	0
330S	1	1	2	0.5	0.5	0.25	0.25	0.12	0
331S	0.5	1	2	0.5	0.5	0.25	0.25	0.12	0
328R	16	32	64	32	8	8	8	8	Asp91Gly
361R	16	32	64	32	8	8	8	8	Asp91Gly
H14R	16	8	16	4	4	2	4	1	Asp91Asn
H23R	16	8	16	4	4	1	4	1	Asn87Lys
H67R	8	8	8	2	4	1	2	0.5	Asp91Gly
HboR	16	8	8	2	4	1	1	1	Asp91Asn
In vitro mutantsb
325MR1	8	8	8	2	2	0.5	1	0.5	Asp91Asn
325MR2	8	8	16	4	1	1	2	1	Asp91Asn
325MR3	16	32	128	32	8	4	8	8	Asp91Asn
H23MR1	64	128	>128	128	64	16	16	16	Asn87Lys + Asp91Tyr
328MR1	32	>128	>128	128	64	16	16	32	Asn87Lys + Asp91Gly

^aMICs were determined by the agar dilution method. CIP, ciprofloxacin; LEV, levofloxacin; MOX, moxifloxacin; GAT, gatifloxacin; TRO, trovafloxacin; GAR, garenoxacin; GEM, gemifloxacin; CLI, clinafloxacin.

^bMutants 325MR1, 325MR2, and 325MR3 were obtained by plating susceptible strain 325S onto agar containing 1 μg of ciprofloxacin/ml. Mutants H23MR1 and 328MR1 were obtained by plating resistant strains H23R and 328R onto clinafloxacin-containing agar.

^c0, no substitution. There were no substitutions in GyrB for any strain tested.

We amplified by PCR and sequenced the quinolone resistance-determining regions (QRDRs) of the gyrA (from codon 38 to 154) and gyrB (from codon 392 to 500) genes from the six resistant strains. The gyrA primers used were 5′-TTTRGCTTATTCMATGAGCGT and 5′-GCAGACGGCTTGGTARAATA, and the gyrB primers were 5′-YGCAAAAGCCAGAGAAGCCA and 5′-ACATGCCCTTGTTCAATCAGC. The sizes of the amplified fragments of gyrA and gyrB were 428 and 444 bp, respectively. One gyrA mutation was found in all resistant strains, but no gyrB mutation was found (Table 1). The gyrA mutations detected had already been reported by Moore et al. (11). In agreement with their results, we found that the most common mutation in H. pylori was at Asp91, and we were able to confirm the role of these mutations in resistance by transformation of susceptible isolate 325S with the PCR-amplified gyrA fragments of resistant strains (data not presented). Only one gyrA mutation was detected in strains 328R and 361R, which presented a higher level of resistance to newer fluoroquinolones than the other resistant clinical strains, thus suggesting that an additional resistance mechanism could exist in these strains. They could carry a second mutation in the DNA topoisomerase IV parC or parE gene. However, this enzyme may not exist in H. pylori because the two genome sequences of H. pylori available (1, 13) do not contain the parC or parE gene. Active efflux also appears unlikely because in one study it was found that efflux did not play any role in resistance to antibiotics in H. pylori (2). These strains could perhaps carry a gyrA or gyrB mutation in a region distinct from the QRDRs.

By sequencing the QRDRs of the ciprofloxacin-susceptible strains, we observed that amino acid variations not implicated in resistance could exist in the QRDRs of both GyrA and GyrB. This was not reported before, but it is important to differentiate these variations from resistance-related alterations. Three were observed in GyrA: the most important is the variation Asn→Thr at resistance hot spot position 87, and the two others are Ala97→Val and Arg140→Lys. Three were also observed in GyrB: Ser479→Gly, Asp481→Glu, and Arg484→Lys.

Single-step mutants were obtained from ciprofloxacin-susceptible isolate 325S by plating 200 μl of a suspension calibrated at 5 McFarland units on horse blood-supplemented B2 agar (Oxoid, Dardilly, France) containing 1 μg of ciprofloxacin per ml, and three were chosen at random for study. We observed that resistant mutants with phenotypes and genotypes similar to those observed in clinical strains could be obtained in vitro in one selection step (Table 1). Furthermore, high-level-fluoroquinolone-resistant mutants were obtained by plating resistant strains H23R and H328R onto clinafloxacin-containing agar. Two of the mutants obtained, H23RM1 and 328RM1, deriving from H23R and 328R, respectively, were selected for further study. MICs of all fluoroquinolones for both mutants were high, and this was associated with the occurrence of a second gyrA mutation (Table 1). To our knowledge, the existence of highly fluoroquinolone-resistant H. pylori gyrA double mutants has never yet been reported.

In conclusion, our work shows that fluoroquinolone resistance of H. pylori already exists in France but remains rare and that gyrA single or double mutations have a critical role in this resistance. It also shows that among newer fluoroquinolones clinafloxacin, gemifloxacin, and garenoxacin are markedly more active than ciprofloxacin against susceptible strains and some but not all gyrA mutants.