Analysis of dual active fluoroquinolones in Streptococcus pneumoniae.

Sir, A recent review by Smith et al.1 and associated correspondence2,3 raised important unresolved issues regarding which quinolones exert dual activity through gyrase and topoisomerase IV in Streptococcus pneumoniae. Dual action involves the substantial engagement of both enzyme killing pathways and is of particular interest in potentially limiting the emergence of resistance. Smith et al.3 concur with us2 that genetic studies are very important in identifying dual active drugs and agree that clinafloxacin is one such agent.1 Our original assignment of clinafloxacin as the archetypal dual action quinolone was based on genetic criteria:4 (i) that gyrA or parC resistance mutations each had minimal (∼two-fold) effects on clinafloxacin MICs but higher level resistance was seen for gyrAparC mutants, and (ii) that gyrA mutants could be selected with drug but only at the MIC and at low frequency (10–9 to 10–10). These findings are consistent with both gyrase and topoisomerase IV contributing substantially to drug action in vivo. We wish to point out that gemifloxacin, gatifloxacin and moxifloxacin share these same features5–9 and therefore, based on the clinafloxacin paradigm, should also be considered as dual active. The view of Smith et al.3 that gemifloxacin, gatifloxacin and moxifloxacin are not dual active derives from two arguments involving laboratory strains and clinical isolates, both of which are problematic. In the case of laboratory strains, Smith et al.3 report their unpublished observations that gyrA or parC mutants can be obtained by challenge with each of the quinolones. They comment that such mutants should not be recoverable (for dual action drugs) based on our suggestions ‘that with clinafloxacin, gatifloxacin, gemifloxacin and moxifloxacin, there is no increase in MIC observed with S. pneumoniae isolates that have either a parC or gyrA mutation alone’ and that ‘the mutants must have mutations in both gyrA and parC’. We do not make the attributed statements.2 Indeed, as with clinafloxacin,4 it is well known that either a gyrA or a parC change confers a small (∼two-fold) increase in MIC for gemifloxacin, gatifloxacin and moxifloxacin.5–8 Presumably, even when both targets are comparably engaged by a dual active drug, a mutation in one target increases the drug concentration needed to register drug effects through the other target, e.g. by requiring increased cleavable complex levels. Moreover, as with clinafloxacin,4 it is documented that gemifloxacin, gatifloxacin and moxifloxacin select single-step gyrA mutants displaying an ∼two-fold MIC increase.5–8 Recovery of gyrA (rather than parC) mutants may indicate that growth inhibition through gyrase is marginally more favoured for these agents. At all events, the crucial issue for dual activity is that parC and gyrA mutations have similarly small effects on MICs and that mutants are recovered only in a narrow drug concentration range (at or near the MIC), at low frequency, and exhibit minimal resistance. At least from published accounts, these features are met by gemifloxacin, gatifloxacin and moxifloxacin.5–9 The unpublished work of Smith et al.3 selecting single gyrA mutants ‘highly resistant’ to gatifloxacin and moxifloxacin and of parC mutants with gemifloxacin is intriguing but will require evaluation in the context of other studies. Interestingly, further support for dual activity comes from the finding that S79F ParC or S81F GyrA mutations reduce enzymic DNA cleavage induced by clinafloxacin, gemifloxacin, gatifloxacin and moxifloxacin by some 8to 64-fold in vitro (see Ref. 2, K. A. Gould and L. M. Fisher, unpublished results) but in terms of resistance in vivo, these effects are moderated to ∼two-fold by drug action on the other target. The second argument of Smith et al.3 against dual activity is based on the MICs of resistant S. pneumoniae clinical isolates, but is unpersuasive. By directly correlating quinolone MICs with parC/gyrA status, they conclude that a single parC mutation gives a spread of MIC increases to gatifloxacin and moxifloxacin in clinical strains, and moreover, that the presence of a single gyrA mutation results in dramatic MIC increases. Unfortunately, clinical isolates are heterogeneous and may carry multiple resistance mutations that confound the interpretation of MICs. Therefore, large MIC increases observed for clinical strains cannot be confidently attributed wholly to parC or gyrA alterations and used to argue against dual activity. Though we agree it is very important to understand quinolone resistance arising in the clinical setting,3 there are difficulties in drawing unambiguous conclusions about drug action from such studies. Stepwise-selected laboratory mutants bearing well-characterized resistance mutations will continue to play a key role in guiding the analysis of dual activity and other mechanistic aspects of quinolone action.