TEXT

Top Abstract Text References

Fluoroquinolones (FQs) are inhibitors of DNA topoisomerases, mainly DNA gyrase and topoisomerase IV (8, 16). In most gram-negative bacteria, DNA gyrase is the primary target for FQs (4), but in gram-positive microorganisms, topoisomerase IV seems to have been the main target for most FQs (1). Nevertheless, target specificities can alter with various drugs, and some new molecules, such as sparfloxacin, have been shown to mainly target gyrase in some gram-positive bacteria (12).

Different mutations in DNA gyrase and topoisomerase IV subunits have been shown to confer FQ resistance (2, 18). We studied the in vitro activities of 13 FQs against Staphylococcus aureus clinical isolates in which the presence or absence of mutations in DNA gyrase and topoisomerase IV has been analyzed.

Ninety S. aureus clinical isolates, collected in the Hospital Universitario de Salamanca (in the mid-west of Spain) and in the Hospital General Universitario de Murcia (in the southeast of Spain), were included in the study. Isolates were selected on the basis of the MICs of ciprofloxacin for these strains. Our aim was to select some clearly susceptible isolates (MICs of around 0.1 to 0.2 µg/ml), some isolates for which the MICs were around the breakpoint (MICs of around 1 to 2 µg/ml), and highly resistant isolates (MICs of >4 µg/ml).

gyrA, gyrB, grlA, and norA and its promoter region were amplified by PCR (1, 6, 7, 15). The obtained DNA amplification fragments were then studied by the single-strand conformational polymorphism method (10). Several DNA fragments of each pattern obtained were sequenced as previously described (13).

Using previously described methods (17), we studied the accumulation of ciprofloxacin in four wild-type isolates for which the MICs were unusually high by using fluorimetric assays, and we then compared the results to those for other wild-type isolates.

Ciprofloxacin, ofloxacin, levofloxacin, difloxacin, pefloxacin, enoxacin, lomefloxacin, fleroxacin, tosufloxacin, sparfloxacin, clinafloxacin (CI-960), Bay y-3118, and grepafloxacin were provided by their respective manufacturers as standard powder.

The in vitro activities of the 13 FQs were evaluated by the agar dilution method, in accordance with the National Committee for Clinical Laboratory Standards guidelines (11).

The results from the genetic analysis of the isolates are summarized in Table 1. No significant mutations were found in gyrB or norA in any isolate. There were no isolates with a gyrA mutation alone.

Although isolates were selected only on the basis of MICs of ciprofloxacin, gyrA single mutations were much less frequent (6.7%) than gyrA-grlA double mutations or the wild type. Only less recently developed FQs (norfloxacin, ciprofloxacin, or ofloxacin) were being used in Spain when isolates were obtained (grepafloxacin, levofloxacin, and trovafloxacin had still not been marketed in Spain), and the FQs select first grlA mutants and then gyrA-grlA mutants. An explanation for the low frequency of grlA mutants, since all gyrA-grlA mutants are presumed to be grlA mutants first, might be that grlA mutants are a relatively unstable group and quickly evolve to become gyrA-grlA mutants when the antibiotic pressure is sufficient.

The MICs of each antibiotic tested in every genotype group appear in Table 2. According to the results obtained in this study, isolates with the same genomic profile for the gyrA, gyrB, and grlA genes show very homogeneous behavior against the quinolones tested. Bay y-3118, clinafloxacin, sparfloxacin, and grepafloxacin have similar MICs for strains with genotypes 1 and 2. 

Grepafloxacin, sparfloxacin, clinafloxacin, and tosufloxacin have been shown in previous studies to have enhanced activity against gram-positive bacteria compared to less recently developed FQs such as ciprofloxacin, ofloxacin, etc., while retaining strong activity against gram-negative bacteria (9). In gram-positive microorganisms, the main molecular target of less recently developed FQs is subunit A of topoisomerase IV (1), although some nonfluorinated quinolones, such as nalidixic acid, can target gyrase in S. aureus (3). Resistance seems to emerge from mutations in the "quinolone resistance determinant region" within grlA. Resistance to high levels of FQs appears when a second mutation occurs, usually in the quinolone resistance determinant region of gyrA (14). So far there is no direct evidence that any FQs primarily target gyrA in S. aureus. Data generated from studies using grepafloxacin suggest that grlA remains the primary target (5). In this case, both grlA single mutations and gyrA single mutations had a limited effect on the activity of the FQ against S. aureus, with double mutations being necessary for resistance (9). Our results show that grlA mutations have little effect on the MICs of newer FQs, as has been similarly shown for gyrA single mutations (14). Therefore, these quinolones are unlikely to be a decisive factor in selecting gyrA and grlA single mutants, the first step in the selection of highly resistant, gyrA-grlA double-mutant isolates. This suggests that development of clinical resistance to grepafloxacin and the other newer FQs would probably require mutations in both gyrA and grlA.

The isolation of such mutants would probably occur at a very low frequency, since the concentrations of quinolone required to kill both gyrA single mutants and grlA single mutants are likely to be much lower than the levels reached in serum and tissues following normal dosages of the drug.

MICs increased significantly when double-mutant isolates were tested. Clinafloxacin and Bay y-3118 were the most active FQs against these isolates.

Four isolates of thirty (13.3%) with genotype 1, one of six (16.6%) with genotype 2, and forty of fifty-four (83.3%) with genotype 3 were methicillin resistant. One isolate with genotype 4 and two isolates with genotype 5 were also methicillin resistant. The MICs of all the FQs tested were similar for methicillin-susceptible and methicillin-resistant isolates in every genotype group. Methicillin resistance was much more frequent in double-mutant isolates than in single-mutant and wild-type isolates. This higher frequency of methicillin resistance is believed to be due to epidemiological factors (e.g., higher frequency of treatment with FQs in these isolates).

Although in most cases there was a clear correlation between genotypes and MICs (the MICs of all the FQs were the same or very similar for almost all the strains in a given genotype group), the MICs of all the FQs for four isolates with no changes found in gyrA, gyrB, or norA were elevated relative to those for the other strains with no changes found in these regions. FQs had MICs for these four strains similar to those obtained for the grlA mutants. The study of these isolates revealed that they exhibit increased efflux activity, since treatment with carbonyl cyanide chlorophenylhydrazone increased the cell's accumulation of ciprofloxacin two to three times. This suggests that regions other than the promoter can control norA expression or that alternative efflux systems may exist in S. aureus. This efflux affected, to a greater or lesser extent, all the FQs studied. The most affected FQ was ofloxacin. The MICs of most of the less recently developed FQs for these four isolates were increased over those for the other isolates 4- to 32-fold. The MICs of the less recently developed FQs for these organisms were increased to around the breakpoint, thus making it inadvisable to use these quinolones for the treatment of infections by such isolates. The reaction of these four isolates to the newer FQs was also abnormal, although the MICs were lower than those of the less recently developed FQs (Table 3). Increased efflux affected the MICs of grepafloxacin, sparfloxacin, and clinafloxacin to a lesser extent than other FQs, confirming the results of previous studies conducted with these newer FQs (5). In conclusion, these results show that newer FQs show more activity than the less recently developed FQs against wild-type isolates and grlA mutants, the activity against gyrA-grlA mutants being more heterogeneous. Moreover, these newer FQs seem to be less affected by the efflux pumps found in these strains.