The mutant selection window and antimicrobial resistance

Transcription

1 Journal of Antimicrobial Chemotherapy (2003) 52, DOI: /jac/dkg269 Advance Access publication 12 June 2003 The mutant selection window and antimicrobial resistance Karl Drlica* Public Health Research Institute, 225 Warren Street, Newark, NJ 07103, USA The mutant selection window is an antimicrobial concentration range extending from the minimal concentration required to block the growth of wild-type bacteria up to that required to inhibit the growth of the least susceptible, single-step mutant. The upper boundary is also called the mutant prevention concentration (MPC). Placing antimicrobial concentrations inside the window is expected to enrich resistant mutant subpopulations selectively, whereas placing concentrations above the window is expected to restrict selective enrichment. Since window dimensions are characteristic of each pathogen antimicrobial combination, they can be linked with antimicrobial pharmacokinetics to rank compounds and dosing regimens in terms of their propensity to enrich mutant fractions of bacterial populations. For situations in which antimicrobial concentrations cannot be kept above the window, restricting the enrichment of mutants requires combination therapy. Keywords: mutant selection window, antimicrobial resistance Introduction During the last decade, most bacterial pathogens have exhibited antimicrobial resistance, and considerable effort has been directed at developing new agents to replace those whose usefulness has been eroded by resistance. At the same time, a better understanding of resistance has been sought to lengthen the useful lifespan of antimicrobials. The latter approach frequently separates genetic resistance into two types: acquired, and de novo. Acquired resistance involves genes or gene systems that move into a pathogen population from an external source. These resistance factors, which are frequently carried by mobile genetic elements, can lower susceptibility so much that resistance arises in a single step. Slowing the dissemination of acquired resistance generally requires case-by-case consideration and is only mentioned briefly in the present work. De novo resistance often develops in a gradual, stepwise manner, usually from the accumulation of mutations that individually lower susceptibility by modest increments. In principle, blocking the growth of mutants should interrupt the development of de novo resistance. Conversely, agents and treatment protocols that readily enrich mutant subpopulations should lead to resistance more rapidly than ones that do not, even though the regimens may be equally effective at curing disease. The problem is to identify mutant-prone regimens prior to widepread clinical use. The present review updates work on the mutant selection window hypothesis. 1,2 This hypothesis, which supersedes the mutant prevention concentration (MPC) concept, serves as an in vitro framework for identifying antimicrobial pathogen situations that are most likely to lead to genetic resistance, and it provides a rationale for deciding whether a compound should be administered as part of a combination regimen rather than being used alone. Readers are reminded that the hypothesis still awaits in vivo validation. Mutant selection window In the 1990s, Baquero suggested that a dangerous concentration range exists in which mutants are selected most frequently. 3,4 Later we defined the boundaries of the range when we noticed that mycobacterial mutants recovered from agar plates displayed a characteristic response to fluoroquinolone concentration. 5 Increasing concentration initially causes a sharp drop in colony recovery, as the growth of wild-type cells is inhibited. A distinct plateau is then observed, and finally a second sharp decline in mutant recovery occurs. The plateau arises from the outgrowth of subpopulations of resistant mutants. 5,6 The second sharp decline takes place when drug concentrations are reached that block the growth of all single-step mutants. 7 Thus mutants are enriched selectively at fluoroquinolone concentrations between the two sharp drops in colony recovery. This concentration range is termed the mutant selection window. 1,2 The lower boundary of the window is the lowest concentration that blocks the growth of the majority of drug-susceptible cells, since below that concentration the mutant cells do not have a growth advantage. The lower boundary can be approximated by the MIC for half the cells in the population (MIC (50) ); however, inhibition of 99% of the cells (MIC (99) ) is a more suitable boundary since it is measured more accurately. The standard MIC is less satisfactory in approximating the lower boundary because some selective pressure is exerted when so many cells (10 4 to 10 5 ) are used in the measurement. Indeed, resistant mutants can be enriched by repeated passage of cells at concentrations just below MIC. 8 Nevertheless, MIC must be near the... *Tel: ; Fax: ; drlica@phri.org The British Society for Antimicrobial Chemotherapy

2 K. Drlica bottom of the window because treatment of Staphylococcus aureus with moxifloxacin in a dynamic model shows that mutant enrichment occurs above the MIC, not below it. 9 Placing MIC near the lower boundary of the selection window contradicts traditional medical teaching, in which resistant mutants are thought to be enriched selectively at concentrations below MIC This distinction is important because traditional dosing recommendations to exceed MIC 13 are likely to place drug concentrations inside the selection window where they will enrich resistant mutant subpopulations. Whereas low drug concentrations do not enrich resistant mutants, they do allow pathogen population expansion; consequently, low drug doses indirectly foster the generation of new mutants that will be enriched by subsequent antimicrobial challenge. The upper limit of the window is the drug concentration that blocks the growth of the least susceptible, single-step mutant. Above this concentration, cell growth requires the presence of two or more resistance mutations. Since two concurrent mutations are expected to arise rarely, few mutants will be amplified selectively when a susceptible population is exposed to drug concentrations that exceed the upper boundary. For example, with fluoroquinolones the mutation frequency for resistance due to target (topoisomerase) mutations can be less than 10 7 ; consequently, more than bacteria would be required to find a cell with two concurrent, independent fluoroquinolone-resistant target mutations. In clinical cases, bacterial populations may reach cells within an infected individual, but is unlikely. Thus, resistance is expected to develop rarely when drug concentrations are kept above the upper boundary of the mutant selection window. This expectation led to the upper boundary being designated the MPC. 5 MPC is approximated experimentally as the lowest concentration that allows no colony growth when more than cells are applied to drug-containing agar plates. The choice of cells is based on several considerations. First, is large enough for mutant subpopulations to be present for testing. Second, infections rarely contain more than organisms. Third, testing more cells is often logistically difficult. In the two cases that have been investigated, a correlation exists between MPC and concentrations that inhibit growth of the least susceptible, first-step mutant. 7,24 The measurement of MPC is performed in two general ways. In one, cells are applied to multiple agar plates at several antimicrobial concentrations such that the total number of cells tested for a given drug concentration exceeds When narrow concentration increments are used, isolated colonies can be found and counted to show that their number progressively approaches zero as drug concentration increases (mutant selection curves become steeper as MPC is approached). In a second method, more than cells are placed on single agar plates that differ in drug concentration by two-fold increments. This method, which allows large numbers of isolates to be surveyed, often gives confluent growth or no growth owing to the large concentration increment. With some bacteria, the large inoculum may affect the apparent susceptibility. Correction factors for inoculum effects can be obtained by carrying out the same experiment with smaller inocula distributed to many more plates. 25 For both methods, growth at antimicrobial concentrations below MPC is confirmed by retesting colonies for growth on agar containing the selecting concentration of drug. To assure that the mutants are stable, they are grown on drug-free agar prior to retesting. Strong experimental support for the window hypothesis was provided recently by fluoroquinolone treatment of S. aureus in a dynamic in vitro model. 9 Moxifloxacin enriched mutants only with a dosing regimen that placed the antimicrobial concentration inside the selection window for at least 20% of the treatment time. In this case, the lower boundary of the window was approximated by MIC so standard pharmacodynamic parameters could be related to the window. For several fluoroquinolones, the MPC corresponded to AUC 24 /MIC > 200 h (AUC is the area under the 24 h time-concentration curve measured after administration of the antimicrobial). The concentration boundaries of the selection window, MIC (99) and MPC, place no restriction on the types of mutant selected. Indeed, many resistance types are expected (e.g. uptake, efflux, degradation and drug target mutants). Fluoroquinolone treatment of mycobacteria serves as an example: more than 20 different gyrase alleles have been recovered using concentrations inside the selection window. 6 These mutants differ in fluoroquinolone susceptibility, and as drug concentration is raised, the fraction of cells that form colonies decreases. Since the upper boundary of the window (MPC) is estimated using a large population of susceptible cells containing spontaneous mutant subpopulations, the most resistant mutant need not be available in pure culture to determine the boundary. It is important to emphasize that the mutant selection window applies to any step in the process of gradual accumulation of mutations. In that context, MPC correlates with the MIC of the least susceptible, nextstep mutant. The window is usually narrower when bacteria have two intracellular targets rather than one. 5 For example, most mycobacteria have only gyrase as a target for the fluoroquinolones, and the large difference between wild-type and mutant susceptibility produces a broad window. Most other bacterial pathogens have both DNA gyrase and DNA topoisomerase IV as targets of the fluoroquinolones. The presence of two targets means that drug concentrations high enough to inhibit both susceptible targets require two mutations for growth. If the two targets have different MICs, colony recovery would drop sharply at the first MIC, exhibit a plateau, and then drop sharply again when the second MIC is reached. If the two targets have similar MICs, the curve will drop sharply without a plateau. 24 Development of de novo resistance Depiction of the mutant selection window in terms of pharmacokinetic profiles (Figure 1) provides a framework for considering initial stages in the development of resistance. Antimicrobials are Figure 1. Pharmacodynamic depiction of the mutant selection window. A hypothetical pharmacokinetic profile is shown in which MIC and MPC are arbitrarily indicated. Double-headed arrow indicates the mutant selection window. 12

3 The mutant selection window usually administered to produce tissue concentrations above the MIC. This allows time for host defences to reduce the pathogen population to where bacterial outgrowth and disease symptoms do not occur after treatment is stopped. When defence systems are inadequate, drug action is required to eliminate the pathogen. Drug costs and potential side effects tend to keep concentrations low while still providing a favourable patient outcome. However, hundreds, and perhaps thousands of resistant cells can be present prior to administration of antibiotic (mutation frequencies are often in the order of 10 6 to 10 8, whereas bacterial infections can contain organisms). Doses of antimicrobial that are inside the selection window can allow growth of the mutant portion of the population. When episodes of infection are brief, mutant enrichment in an individual patient may not be detected easily. Nevertheless, passage of a pathogen through many treated patients is expected to increase the mutant fraction of the bacterial population gradually. Even if infection arises from a single pathogen cell, over time the probability increases for a given cell to be resistant. Thus an antimicrobial agent may cure 99% of the cases, but when millions are considered, the development of resistance is an inevitable consequence of dosing strategies that place drug concentrations inside the mutant selection window. Two general scenarios can lead to de novo resistance. In the first, the presence of more than one mutation is required for a cell to be considered resistant. With this pattern, resistant populations develop stepwise through the gradual accumulation of mutations that individually reduce susceptibility by low-to-moderate increments. When individual cells are tested, some are found to have intermediate levels of susceptibility. An example of this pattern is the development of fluoroquinolone resistance in Streptococcus pneumoniae: recent clinical isolates of S. pneumoniae contain a variety of target and nontarget resistance mutations If a strain already contains a resistance mutation, the next mutational step is achieved more readily. 24 Thus the development of resistance accelerates with the accumulation of mutations. Since many alleles can accumulate (at least seven with S. pneumoniae), 28 incremental improvements in fluoroquinolone activity are likely to be neutralized by selective enrichment of mutants. The second scenario is illustrated by treatment of Mycobacterium tuberculosis with most agents, and treatment of Escherichia coli and S. aureus with rifampicin. In these situations, antimicrobial resistance arises in a single step a mutation reduces susceptibility so much that no tolerable concentration of drug can block mutant growth. In this situation, the upper boundary of the window (MPC) is above the maximum tolerable drug concentration; 2,30 individual organisms in a bacterial population are either very susceptible or highly resistant. 31 Examples of both situations can be identified from the data shown in Table 1 (in stepwise resistance C max is greater than MPC, whereas the reverse is true for single-step resistance; the table lists total drug concentrations for simplicity and because corrections for protein binding effects are controversial). 32,33 Most plasmid-borne resistance is expected to fall in the single-step category, since many rounds of selective pressure are likely to have occurred prior to plasmid entry into the bacterial population in question. Promoting and restricting resistance Loss of bacterial population susceptibility is expected to occur most readily with agents and treatment regimens that place the antimicrobial concentration inside the window for long periods of time. 1,3 As pointed out above, first-line anti-tuberculosis agents fall in this category: no standard agent is administered so that concentations exceed MPC. 30 Consequently, the agents are used in multidrug regimens, since that requires the acquisition of multiple mutations for bacterial growth. 34 The same solution is indicated for rifampicin treatment of S. aureus and for drug bacterial combinations in which plasmidborne resistance is likely to be present in subpopulations of otherwise susceptible bacteria. Resistance is also expected to occur readily when bacterial populations contain a diverse array of resistant mutants associated with many different levels of susceptibility, because overall susceptibility can then be reduced in small steps. Examples are seen with the fluoroquinolones. 6,24,35,36 With these compounds, the use of concentrations near the bottom of the selection window is more likely to enrich mutants, because in that range the mutants are present at a much higher frequency (low concentrations of fluoroquinolone allow the recovery of so many non-target resistance mutants that the target topoisomerase mutants are difficult to obtain). 6 Several types of clinical data are readily interpreted within the selection window framework. In one, Thomas and associates 37 determined (i) the pharmacokinetics of ciprofloxacin for patients hospitalized with pneumonia caused by Pseudomonas aeruginosa, (ii) the MIC for each infecting strain, and (iii) whether treatment failure was associated with bacterial resistance. Low values of the pharmacodynamic parameter AUC/MIC, which correspond to dosing in the lower portion of the window, invariably resulted in resistance and treatment failure. Higher values that placed concentrations near the middle of the window (AUC/MIC > 100) were associated with only 25% resistance and failure [in these experiments the maximal concentration reached 5 MIC, or about 50% MPC (J. Blondeau, personal communication)]. We expect that analysis of mutants recovered from this type of study will show that non-target mutants predominate among the low AUC/MIC cases, whereas gyrase (target) mutants will be found in the high ones. Consistent with these ideas, mutants selected at low AUC/MIC conditions in a rat pseudomonal pneumonia model are largely of the efflux type. 38 A second example is quinolone resistance among isolates of S. aureus. With this organism, ciprofloxacin concentrations fall inside the mutant selection window for most of the dosing period. 39 Once resistant clones are enriched selectively, they probably disseminate rapidly within institutional settings, since widespread fluoroquinolone resistance developed soon after the introduction of ciprofloxacin. 40 A more complex example may soon be seen with S. aureus and the new 6-desfluoroquinolone garenoxacin. The pharmacokinetic profile of this agent falls inside the window with ciprofloxacinresistant mutants. Consequently, additional resistance determinants are expected to be selected readily when this compound is used for methicillin-resistant S. aureus, because many isolates are already ciprofloxacin-resistant. 41 However, serum concentrations are well above MPC with ciprofloxacin-susceptible S. aureus. These organisms should exhibit resistance rarely if serum concentrations reflect those in tissues where mutants are enriched. With S. pneumoniae, the pharmacokinetic profile for levofloxacin also falls inside the selection window. 39 Both case reports 28,42 45 and surveillance studies are showing increases in resistance, which in Hong Kong has reached the dissemination phase. 11,46 Newer fluoroquinolones exhibit lower values of MPC, which leads to a ranking of compounds in terms of time that serum concentrations exceed MPC (moxifloxacin > gemifloxacin > gatifloxacin > levofloxacin). 50 Clinical comparisons of fluoroquinolones, with respect to the development of resistance in both S. aureus and S. pneumoniae could provide tests of the window hypothesis: compounds having concentrations 13

4 K. Drlica Table 1. Mutant selection window for various organisms and antimicrobial agents Organism Compound MIC (99) (mg/l) MPC (mg/l) MPC/ MIC (99) C max (mg/l) MPC/C max Reference E. coli Norfloxacin Rifampicin 7 >4000 > >42 2 Tobramycin Chloramphenicol Penicillin G a M. smegmatis Gatifloxacin Ciprofloxacin Moxifloxacin ,65 Chlorampheniol ,66 Penicillin G a ,67 Erythromycin ,68 Tetracycline ,69 M. tuberculosis Rifampicin 0.02 >80 > > Streptomycin 0.2 >320 > > Isoniazid Capreomycin Kanamycin 1.5 >800 > >38 30 Cycloserine Ciprofloxacin Moxifloxacin Gatifloxacin S. aureus Strain RN450 Garenoxacin Moxifloxacin ,65 Gatifloxacin Levofloxacin Ciprofloxacin Erythromycin ,68 Penicillin G a ,67 Chloramphenicol ,66 Tetracycline ,69 Clinical isolates Ciprofloxacin Garenoxacin S. pneumoniae ATCC Moxifloxacin Levofloxacin Clinical isolates b Moxifloxacin Gemifloxacin Gatifloxacin Trovafloxacin Levofloxacin a specific activity 1600 U/µg. b Measured with more than 150 clinical isolates as standard MIC 90 and provisional MPC 90. In this work, single agar plates were used for each drug concentration. above MPC for the longest times should be least likely to enrich resistant mutants. The number of patients involved in treatment is likely to be important. For example, the probability of a bacterial mutant being present is related to the number of bacteria per patient times the number of patients. Thus, keeping concentrations above MPC may not prevent resistance when very large numbers of patients are considered. The number of patients also bears on whether it is necessary to maintain drug concentrations above MPC for infections that contain few bacterial cells, since with few cells the probability that mutants will be present is small. The latter consideration may be relevant to short-term prophylactic use of antimicrobial agents. Traditional pharmacodynamic approach The MPC-based strategy differs from other pharmacodynamic approaches advocated for restricting the development of resistance MPC represents a conceptual concentration threshold for restricting the development of resistance that has been estimated in vitro; thus, it can be used to restrict the development of resistance from the time an agent is first introduced into clinical practice. However, no animal or clinical study is yet available to provide a quantitative relationship between in vitro and in vivo values. In contrast, traditional pharmacodynamic strategies based on time above MIC, AUC/MIC or C max /MIC reflect the cumulative attack of susceptible 14

5 The mutant selection window cells in vivo; these parameters correlate empirically with patient and animal outcome. 54,55 Use of patient outcome as an end point allows standard pharmacodynamic parameters to bypass uncertainties associated with determining available drug concentration at the site of infection. However, the standard methods require examination of large numbers of patients, or the use of heavy inocula, to define the drug concentration that will prevent an overall increase in resistance prevalence. So far, the traditional measurement has revealed reduction in resistance with increasing dose, but not the elimination of resistance. 37,55 In early work using in vitro pharmacodynamic models, several bacterial species were treated with the fluoroquinolone enoxacin, and a C max /MIC ratio of 8 was found to restrict the outgrowth of resistant mutants. 56 Since the total number of cells tested (10 7 ) was low, it is likely that the system measured growth of efflux mutants and other mutants of low-to-intermediate susceptibility. Consequently, the work probably underestimated the pharmacodynamic parameters needed to block the development of resistance. Even values of C max /MIC greater than 20 are insufficient for lomefloxacin to prevent P. aeruginosa from killing rats. 55 While it is not known that lomefloxacin treatment failure is the result of outgrowth of resistant mutants, that is currently the most straightforward assumption. Dual targeting and closing the mutant selection window Ng et al. 15 and Pan et al. 16 pointed out that an antimicrobial agent that inhibits two different targets with equal efficacy would require a cell to acquire two concurrent mutations for growth; only rarely would resistant mutants be recovered. Within the context of the selection window hypothesis, the window would be closed (MIC = MPC). Dual targeting compounds offer many of the advantages of combination therapy without the problems associated with pharmacokinetic mismatches 57 and increased adverse events associated with the use of two agents. Since many bacterial species have two intracellular targets for the fluoroquinolones (DNA gyrase and DNA topoisomerase IV), these agents have been investigated for dual targeting. 58 Several new compounds (moxifloxacin, gatifloxacin, gemifloxacin and clinafloxacin) approach the dual target situation with S. pneumoniae, as judged (i) by very low mutation frequencies, 59 (ii) by decreased susceptibility of both gyra and parc resistance mutants (clinafloxacin), 59 and (iii) by a diminished plateau (inflection point) in plots of mutant recovery versus fluoroquinolone concentration (moxifloxacin). 24 Thus dual targeting appears to be a good approach for refining compounds to restrict resistance. Concluding remarks Consideration of the mutant selection window leads to the suggestion that antimicrobial concentrations between MIC (99) and MPC enrich mutant subpopulations selectively (standard MIC is often quantitatively similar to MIC (99) ). Such conditions may suppress most infections, 39,60 especially when host defences effectively eliminate pathogens. However, when large numbers of patients are treated at concentrations inside the selection window, susceptibility decreases gradually. Eventually a point is reached at which the antimicrobial agent becomes ineffective. According to these ideas, restricting the development of resistance requires that antimicrobial concentrations at the site of infection be kept above MPC. If that cannot be done for a given agent pathogen combination, the agent should be used as part of a combination therapy involving agents with different targets. Such an approach is likely to be required for plasmid-borne resistance. The effects of lethal action have not been integrated into the mutant selection window hypothesis, although they are certainly important to the development of resistance. The most obvious effect is the reduction in bacterial load. This facilitates host defencemediated removal of bacteria and reduces the probability that new mutants will arise. However, an agent that kills susceptible cells but not mutants will enrich pre-existing mutants. Thus desirable agents should also kill resistant mutants. 24,61 63 For such compounds, therapeutic drug concentrations may not need to be kept above MPC throughout therapy to restrict mutant enrichment. Whether exceeding the MPC is sufficient to restrict the development of resistance requires clinical testing. Such tests are important because numerical considerations, such as mutation frequencies and relevant drug concentrations, could depend significantly on whether the microbes are growing on agar plates or in host organisms. Moreover, fluctuations in antimicrobial pharmacokinetics could require dosing adjustments to make MPC an effective threshold. Animal and clinical studies now seem justified, since a mutant selection window can be measured for many pathogen antimicrobial combinations. 64 Acknowledgements I thank the following for critical comments on the manuscript: Marila Gennaro, Glenn Tillotson, and Xilin Zhao. The work was supported by grants from Bayer AG, Bristol- Myers-Squibb, and the National Institutes of Health (AI 35257). References 1. Zhao, X. & Drlica, K. (2001). Restricting the selection of antibioticresistant mutants: a general strategy derived from fluoroquinolone studies. Clinical Infectious Diseases 33, Suppl. 3, S Zhao, X. & Drlica, K. (2002). Restricting the selection of antibioticresistant mutants: measurement and potential uses of the mutant selection window. Journal of Infectious Diseases 185, Baquero, F. & Negri, M. C. (1997). Strategies to minimize the development of antibiotic resistance. Journal of Chemotherapy 9, Suppl. 3, Baquero, F. (1990). Resistance to quinolones in Gram-negative microorganisms: mechanisms and prevention. European Urology 17, Suppl. 1, Dong, Y., Zhao, X., Domagala, J. et al. (1999). Effect of fluoroquinolone concentration on selection of resistant mutants of Mycobacterium bovis BCG and Staphylococcus aureus. Antimicrobial Agents and Chemotherapy 43, Zhou, J.-F., Dong, Y., Zhao, X. et al. (2000). Selection of antibiotic resistance: allelic diversity among fluoroquinolone-resistant mutations. Journal of Infectious Diseases 182, Sindelar, G., Zhao, X., Liew, A. et al. (2000). Mutant prevention concentration as a measure of fluoroquinolone potency against mycobacteria. Antimicrobial Agents and Chemotherapy 44, Dalhoff, A. (2001). Comparative in vitro and in vivo activity of the C-8 methoxy quinolone moxifloxacin and the C-8 chlorine quinolone Bay y Clinical Infectious Diseases 32, Suppl. 1, S Firsov, A., Vostrov, S., Lubenko, I. et al. (2003). In vitro pharmacodynamic evaluation of the mutant selection window hypothesis: four fluoroquinolones against Staphylococcus aureus. Antimicrobial Agents and Chemotherapy 47, Ambrose, P., Zoe-Powers, A., Russo, R. et al. (2002). Utilizing pharmacodynamics and pharmacoeconomics in clinical and formulary 15

6 K. Drlica decision making. In Antimicrobial Pharmacodynamics in Theory and Clinical Practice (Nightingale, C., Murakawa, T. & Ambrose, P., Eds), pp Marcel Dekker, New York, NY, USA. 11. Ho, P. L., Tse, W. S., Tsang, K. et al. (2001). Risk factors for acquisition of levofloxacin-resistant Streptococcus pneumoniae: a casecontrol study. Clinical Infectious Diseases 32, Schentag, J. J., Gilliland, K. K. & Paladino, J. A. (2001). What have we learned from pharmacokinetic and pharmacodynamic theories? Clinical Infectious Diseases 32, Suppl. 1, S Heffelfinger, J., Dowell, S., Jorgensen, J. et al. (2000). Management of community-acquired pneumonia in the era of pneumococcal resistance. Archives of Internal Medicine 160, Zhao, X., Xu, C., Domagala, J. et al. (1997). DNA topoisomerase targets of the fluoroquinolones: a strategy for avoiding bacterial resistance. Proceedings of the National Academy of Sciences, USA 94, Ng, E. Y., Trucksis, M. & Hooper, D. C. (1996). Quinolone resistance mutations in topoisomerase IV: relationship to the flqa locus and genetic evidence that topoisomerase IV is the primary target and DNA gyrase is the secondary target of fluoroquinolones in Staphylococcus aureus. Antimicrobial Agents and Chemotherapy 40, Pan, X.-S., Ambler, J., Mehtar, S. et al. (1996). Involvement of topoisomerase IV and DNA gyrase as ciprofloxacin targets in Streptococcus pneumoniae. Antimicrobial Agents and Chemotherapy 40, Iseman, M. (1994). Evolution of drug-resistant tuberculosis: a tale of two species. Proceedings of the National Academy of Sciences, USA 91, Bingen, E., Lambert-Zechovsky, N., Mariani-Kurkdjian, P. et al. (1990). Bacterial counts in cerebrospinal fluid of children with meningitis. European Journal of Clinical Microbiology and Infectious Diseases 9, Feldman, W. (1976). Concentrations of bacteria in cerebrospinal fluid of patients with bacterial meningitis. Journal of Pediatrics 88, Mitchison, D. A. (1984). Drug resistance in mycobacteria. British Medical Bulletin 40, Wimberley, N., Faling, L. & Bartlett, J. (1979). A fiberoptic bronchoscopy technique to obtain uncontaminated lower airway secretions for bacterial culture. American Review of Respiratory Diseases 119, Fagon, J., Chastre, J., Trouillet, J. et al. (1990). Characterization of distal bronchial microflora during acute exacerbation of chronic bronchitis. American Review of Respiratory Diseases 142, Low, D. (2001). Antimicrobial drug use and resistance among respiratory pathogens in the community. Clinical and Infectious Diseases 33, Suppl. 3, S Li, X., Zhao, X. & Drlica, K. (2002). Selection of Streptococcus pneumoniae mutants having reduced susceptibility to levofloxacin and moxifloxacin. Antimicrobial Agents and Chemotherapy 46, Blondeau, J., Zhao, X., Hansen, G. et al. (2001). Mutant prevention concentrations (MPC) of fluoroquinolones with clinical isolates of Streptococcus pneumoniae. Antimicrobial Agents and Chemotherapy 45, Bast, D., Low, D., Duncan, C. et al. (2000). Fluoroquinolone resistance in clinical isolates of Streptococcus pneumoniae: contributions of type II topoisomerase mutations and efflux on levels of resistance. Antimicrobial Agents and Chemotherapy 44, Pestova, E., Millichap, J., Noskin, G. et al. (2000). Intracellular targets of moxifloxacin: a comparison with other fluoroquinolones. Journal of Antimicrobial Chemotherapy 45, Urban, C., Rahman, N., Zhao, X. et al. (2001). Fluoroquinoloneresistant Streptococcus pneumoniae associated with levofloxacin therapy. Journal of Infectious Diseases 184, Jones, M. E., Sahm, D. F., Martin, N. et al. (2000). Prevalence of gyra, gyrb, parc, and pare mutations in clinical isolates of Streptococcus pneumoniae with decreased susceptibilities to different fluoroquinolones and originating from worldwide surveillance studies during the respiratory season. Antimicrobial Agents and Chemotherapy 44, Dong, Y., Zhao, X., Kreiswirth, B. et al. (2000). Mutant prevention concentration as a measure of antibiotic potency: studies with clinical isolates of Mycobacterium tuberculosis. Antimicrobial Agents and Chemotherapy 44, Murray, B. E. (1997). Antibiotic resistance. Advances in Internal Medicine 42, Craig, W. A. & Andes, D. R. (2000). Correlation of the magnitude of the AUC 24 /MIC for 6 fluoroquinolones against Streptococcus pneumoniae (SP) ith survival and bactericidal activity in an animal model. In Programs and Abstracts of the Fortieth Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Abstract 289, p. 7. American Society for Microbiology, Washington, DC, USA. 33. Ernst, E. J., Klepser, M., Petzold, R. et al. (2002). Evaluation of survival and pharmacodynamic relationships for five fluoroquinolones in a neutropenic murine model of pneumococcal lung infection. Pharmacotherapy 22, Fox, W., Elklard, G. & Mitchison, D. (1999). Studies on the treatment of tuberculosis undertaken by the British Medical Research Council Tuberculosis Units, , with relevant subsequent publications. International Journal of Tuberculosis and Lung Disease 3, S Nakamura, S. (1997). Mechanisms of quinolone resistance. Journal of Infection and Chemotherapy 3, Drlica, K. & Zhao, X. (1997). DNA gyrase, topoisomerase IV, and the 4-quinolones. Microbiology and Molecular Biology Reviews 61, Thomas, J., Forrest, A., Bhavnani, S. et al. (1998). Pharmacodynamic evaluation of factors associated with the development of bacterial resistance in acutely ill patients during therapy. Antimicrobial Agents and Chemotherapy 42, Join-Lambert, O., Michéa-Hamzehpour, M., Köhler, T. et al. (2001). Differential selection of multidrug efflux mutants by trovafloxacin and ciprofloxacin in an experimental model of Pseudomonas aeruginosa acute pneumonia in rats. Antimicrobial Agents and Chemotherapy 45, Tillotson, G., Zhao, X. & Drlica, K. (2001). Fluoroquinolones as pneumococcal therapy: closing the barn door before the horse escapes. Lancet 1, Acar, J. & Goldstein, F. (1997). Trends in bacterial resistance to fluoroquinolones. Clinical Infectious Diseases 24, Suppl. 1, S Schmitz, F.-J., Fluit, A., Brisse, S. et al. (1999). Molecular epidemiology of quinolone resistance and comparative in vitro activities of new quinolones against European Staphylococcus aureus isolates. FEMS Immunology and Medical Microbiology 26, Fishman, N. O., Suh, B., Weigel, L. M. et al. (1999). Three levofloxacin treatment failures of pneumococcal respiratory tract infections. In Programs and Abstracts of the Thirty-ninth Interscience Conference on Antimicrobial Agents and Chemotherapy, San Francisco, CA, American Society for Microbiology, Washington, DC, USA. 43. Weiss, K., Restieri, C., Gauthier, R. et al. (2001). A nosocomial outbreak of fluoroquinolone-resistant Streptococcus pneumoniae. Clinical and Infectious Diseases 33, Davidson, R., Cavalcanti, R., Brunton, J. et al. (2002). Resistance to levofloxacin and failure of treatment of pneumococcal pneumonia. New England Journal of Medicine 346, Kays, M., Smith, D., Wack, M. et al. (2002). Levofloxacin treatment failure in a patient with fluoroquinolone-resistant Streptococcus pneumoniae pneumonia. Pharmacotherapy 22, Ho, P., Yung, R., Tsang, D. et al. (2001). Increasing resistance of Streptococcus pneumoniae to fluoroquinolones: results of a Hong Kong multicentre study in Journal of Antimicrobial Chemotherapy 48, Kays, M. & Denys, G. (2001). Fluoroquinolone susceptibility, resistance, and pharmacodynamics versus clinical isolates of Streptococcus 16

7 The mutant selection window pneumoniae from Indiana. Diagnostic Microbiology and Infectious Disease 40, Quale, J., Landman, D., Ravishankar, J. et al. (2002) Streptococcus pneumoniae, Brooklyn, New York: fluoroquinolone resistance at our doorstep. Emerging Infectious Diseases 8, Nagai, K., Appelbaum, P. C., Davies, T. A. et al. (2002). Susceptibilities to telithromycin and six other agents and prevalence of macrolide resistance due to L4 ribosomal protein mutation among 992 Pneumococci from 10 central and Eastern European countries. Antimicrobial Agents and Chemotherapy 46, Hansen, G., Metzler, K., Drlica, K. et al. Mutant prevention concentration for gemifloxacin with clinical isolates of Streptococcus pneumoniae. Antimicrobial Agents and Chemotherapy 47, Schentag, J. (1999). Antimicrobial action and pharmacokinetics/ pharmacodynamics: the use of AUIC to improve efficacy and avoid resistance. Journal of Chemotherapy 11, Drusano, G. (2000). Fluoroquinolone pharmacodynamics: prospective determination of relationships between exposure and outcome. Journal of Chemotherapy 12, Suppl. 4, Craig, W. (2001). Does the dose matter? Clinical and Infectious Diseases 33, Suppl. 3, S Preston, S., Drusano, G., Berman, A. et al. (1998). Pharmacodyamics of levofloxacin: a new paradigm for early clinical trials. Journal of the American Medical Association 279, Drusano, G. L., Johnson, D., Rosen, M. et al. (1993). Pharmacodynamics of a fluoroquinolone antimicrobial agent in a neutropenic rat model of Pseudomonas sepsis. Antimicrobial Agents and Chemotherapy 37, Blaser, J., Stone, B., Groner, M. et al. (1987). Comparative study with enoxacin and netilmicin in a pharmacodynamic model to determine importance of ratio of antibiotic peak concentration to MIC for bactericidal activity and emergence of resistance. Antimicrobial Agents and Chemotherapy 31, Drlica, K. (2001). A strategy for fighting antibiotic resistance. ASM News 67, Fisher, L. M. & Heaton, V. J. (2003). Dual activity of fluoroquinolones against Streptococcus pneumoniae. Journal of Antimicrobial Chemotherapy 51, Pan, X. & Fisher, L. M. (1998). DNA gyrase and topoisomerase IV are dual targets of clinafloxacin action in Streptococcus pneumoniae. Antimicrobial Agents and Chemotherapy 42, Fogarty, C., Greenberg, R., Dunbar, L. et al. (2001). Effectiveness of levofloxacin for adult community-acquired pneumonia caused by macrolide-resistant Streptococcus pneumoiae: integrated results from four open-label, multicenter, phase III clinical trials. Clinical Therapeutics 23, Lu, T., Zhao, X., Li, X. et al. (2001). Enhancement of fluoroquinolone activity by C-8 halogen and methoxy moieties: action against a gyrase resistance mutant of Mycobacterium smegmatis and a gyrasetopoisomerase IV double mutant of Staphylococcus aureus. Antimicrobial Agents and Chemotherapy 45, Dong, Y., Xu, C., Zhao, X. et al. (1998). Fluoroquinolone action against mycobacteria: effects of C8 substituents on bacterial growth, survival, and resistance. Antimicrobial Agents and Chemotherapy 42, Zhao, X., Wang, J.-Y., Xu, C. et al. (1998). Killing of Staphylococcus aureus by C-8-methoxy fluoroquinolones. Antimicrobial Agents and Chemotherapy 42, Lu, T., Zhao, X., Li, X. et al. (2003). Effect of chloramphenicol, erythromycin, moxifloxacin, penicillin, and tetracycline concentration on the recovery of resistant mutants of Mycobacterium smegmatis and Staphylococcus aureus. Journal of Antimicrobial Chemotherapy 52, Sullivan, J. T., Woodruff, M., Lettieri, J. et al. (1999). Pharmacokinetics of a once-daily oral dose of moxifloxacin (Bay ), a new enantiomerically pure 8-methoxy quinolone. Antimicrobial Agents and Chemotherapy 43, Burke, J., Wargin, W., Sherertz, R. et al. (1982). Pharmacokinetics of intravenous chloramphenicol sodium succinate in adult patients with normal renal and hepatic function. Pharmacokinetics and Biopharmocology 10, Visser, I., Arnouts, P., vanfurth, R. et al. (1993). Clinical pharmacokinetics of continuous intravenous administration of penicillins. Clinical and Infectious Diseases 17, Mather, L. E., Austin, K. L., Philpot, C. R. et al. (1981). Absorption and bioavailability of oral erythromycin. British Journal of Clinical Pharmacology 12, Healy, D. P., Dansereau, R., Dunn, A. et al. (1997). Reduced tetracycline bioavailability caused by magnesium aluminum silicate in liquid formulations of bismuth subsalicilate. Annals of Pharmacotherapy 31, Zhao, X., Eisner, W., Perl-Rosenthal, N. et al. (2003). Mutant prevention concentration of garenoxacin (BMS ) for ciprofloxacinsusceptible or -resistant Staphylococcus aureus. Antimicrobial Agents and Chemotherapy, 47,