Effect of Aerobic and Anaerobic Environments on Antistaphylococcal Activities of Five Fluoroquinolones

Transcription

1 ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Feb. 1995, p Vol. 39, No /95/$ Copyright 1995, American Society for Microbiology Effect of Aerobic and Anaerobic Environments on Antistaphylococcal Activities of Five Fluoroquinolones RICHARD A. ZABINSKI, 1 KARLA J. WALKER, 2 ALISON J. LARSSON, 3 JULIA A. MOODY, 4 GLENN W. KAATZ, 5 AND JOHN C. ROTSCHAFER 3 * Miles Inc., West Haven, Connecticut ; MedTox Laboratories, New Brighton, Minnesota ; College of Pharmacy and The Antibiotic Pharmacodynamic Modeling Institute, University of Minnesota, 3 and Department of Anatomic and Clinical Pathology, 4 St. Paul-Ramsey Medical Center, St. Paul, Minnesota 55101; and Division of Infectious Diseases, Department of Internal Medicine, Wayne State University School of Medicine, Detroit, Michigan Received 18 October 1993/Returned for modification 2 March 1994/Accepted 11 December 1994 A previously established in vitro pharmacodynamic system was used to evaluate the antistaphylococcal activities of five fluoroquinolones under both aerobic and anaerobic conditions. Staphylococcus aureus ATCC was exposed to a 5- g/ml concentration of each of the following fluoroquinolones: ciprofloxacin, ofloxacin, temafloxacin, sparfloxacin, and clinafloxacin. Terminal elimination half-lives of 4, 6, 8, 8, and 13 h were simulated for the respective drugs. Each fluoroquinolone was bactericidal under both aerobic and anaerobic conditions. However, the bactericidal activity of each fluoroquinolone was delayed by anaerobiosis. This difference in fluoroquinolone activity under aerobic and anaerobic conditions could not be attributed to any particular parameter or physiochemical property but was most likely caused by a combination of factors (e.g., variations in hydrophobicity, intracellular ph, antibiotic concentration, and structure-activity relationships). Fluoroquinolone uptake studies were also performed to investigate the possibility of active, energydependent transport mechanisms in S. aureus ATCC Uptake studies indicated that active efflux does occur in S. aureus ATCC The antistaphylococcal activity of fluoroquinolones under aerobic conditions has been well established through numerous studies (13, 24). However, the antistaphylococcal activity of these antimicrobial agents under anaerobic conditions has not been well studied. As staphylococci are facultative by nature, there is a distinct likelihood that some clinical infections involving these organisms occur in an anaerobic or microaerophilic environment. Hence, there exists a need to evaluate the antistaphylococcal performance of fluoroquinolones under both aerobic and anaerobic conditions. Morrissey and Smith have proposed that fluoroquinolone uptake by Staphylococcus aureus is oxygen dependent (14, 17). Further, Mitsuyama et al. have demonstrated that fluoroquinolone uptake by porin-deficient bacteria is at least partially dependent upon hydrophobicity (16). The purpose of this investigation was to evaluate the antistaphylococcal activities of five fluoroquinolones with various degrees of hydrophobicity under both aerobic and anaerobic conditions. (This report was presented as an abstract at the 32nd Interscience Conference on Antimicrobial Agents and Chemotherapy, Anaheim, Calif. [abstract 43].) MATERIALS AND METHODS In vitro pharmacodynamic system. The pharmacodynamic model used in this investigation represents a modified version of that previously described by Garrison et al. (8) and has been previously described by Zabinski et al. (26). This system consisted of a 1,000-ml glass vessel with inflow and outflow ports, connective silicone tubing (Masterflex L/S thin-wall tubing; Cole-Parmer, Chicago, Ill.), a Masterflex peristaltic pump (Cole-Parmer), a fresh-medium reservoir, a stir-hot plate (Nuovo II; Barnstead/Thermolyne Corp., Dubuque, Iowa), a magnetic stir bar, and a thermometer. * Corresponding author. Mailing address: St. Paul-Ramsey Medical Center, Section of Clinical Pharmacology, 640 Jackson Street, St. Paul, MN Phone: (612) Fax: (612) A 1-ml volume of a quinolone-containing solution was injected into the system as a bolus to attain a starting antibiotic peak concentration of 5 g/ml. With the peristaltic pump, antibiotic-free medium was pumped into the system at a predetermined rate such that an equal volume of antibiotic-containing medium was displaced into a waste vessel, resulting in the simulation of a first-order, onecompartment pharmacokinetic process. An anaerobic environment was created by placing the pharmacodynamic system within a Bactron IV anaerobic chamber (Sheldon Mfg., Cornelius, Oreg.). The chamber operated with an anaerobic gas mixture of 5% hydrogen, 10% carbon dioxide, and 85% nitrogen. Positive pressure was maintained within the chamber at 0.25 in. (1 in. is 2.54 cm) of water to prevent air from leaking into the system. BBL GasPak (Becton Dickinson, Cockeysville, Md.) indicator strips were used to monitor anaerobiosis. These indicator strips are able to detect oxygen concentrations of 0.5% (1). Antibiotics. Ciprofloxacin hydrochloride was from Miles Pharmaceuticals (West Haven, Conn.), clinafloxacin hydrochloride (CI-960) was from Parke- Davis Pharmaceuticals (Ann Arbor, Mich.), ofloxacin was from Ortho Pharmaceuticals (Raritan, N.J.), temafloxacin hydrochloride was from Abbott Pharmaceuticals (Abbott Park, Ill.), and sparfloxacin (CI-978) was from Parke-Davis Pharmaceuticals. Organism. S. aureus ATCC was used as the test organism. Regrowth isolates and stock cultures of S. aureus were reconstituted from frozen stock ( 80 C) and subcultured at least twice before use. Standardized cultures were prepared by inoculating a 25-ml volume of cation-supplemented Mueller-Hinton broth (CSMHB; Difco, Detroit, Mich.) with two to five colonies of S. aureus. This suspension was grown at 35 C within the appropriate aerobic or anaerobic environment to the turbidity of a McFarland no. 1 standard, diluted 1:50, and again grown to the turbidity of a McFarland no. 1 standard. This standardized suspension was diluted 1:10 so that the initial inoculum in each experiment was to CFU of exponentially growing bacteria per ml. Media. CSMHB was used as the growth medium in each experiment. Calcium and magnesium were adjusted in accordance with National Committee for Clinical Laboratory Standards guidelines (20). Trypticase soy blood agar (Dimed, St. Paul, Minn.) plates were used for viable colony count determinations. Bacterial enumeration. Trypticase soy blood agar plates were inoculated with 100- l serially diluted samples. Plates from aerobic experiments were incubated at 35 C in room air. Plates from anaerobic experiments were incubated at 35 C within an anaerobic (5% H 2, 10% CO 2, 85% N 2 ) incubator which is a built-in feature of the Bactron IV anaerobic chamber. The bacterial colonies were then quantified to determine viable bacterial counts; the dilution that revealed between 30 and 300 colonies per plate was that which was used for constructing time-kill curves. Inoculation of plates with 100- l undiluted samples resulted in a lower limit of bacterial quantification of CFU/ml. Antibiotic carryover. The possibility of antibiotic carryover was eliminated 507

2 508 ZABINSKI ET AL. ANTIMICROB. AGENTS CHEMOTHER. through antibiotic removal with 1gofpolymeric binding resin (Amberlite XAD- 4/1090; Rohm & Haas, Philadelphia, Pa.) per ml of broth, as previously described (25). MIC determination. MICs and MBCs were determined under aerobic and anaerobic conditions in CSMHB by microtiter broth dilution by using final inocula of approximately 10 7 CFU/ml. These methods were consistent with National Committee for Clinical Laboratory Standards guidelines for susceptibility tests for bacteria that grow aerobically (20). MIC and MBC determinations were made for both unexposed organisms and organisms that had been exposed to antibiotics for 24 h in the pharmacodynamic system. Fluoroquinolone hydrophobicities. The hydrophobicities of the fluoroquinolones were derived by averaging partition coefficients that were obtained from the medical literature and the pharmaceutical industry (2, 10, 19, 21, 22). Partition coefficients derived from the literature were all determined by partitioning with 0.1 N phosphate buffer (ph 7.0 to 7.4) and n-octanol at 25 C. The partition coefficients are expressed as the ratio of the quinolone concentration in the organic phase to that in the aqueous phase. Aerobic and anaerobic time-kill kinetic studies. Time-kill studies and growth control experiments were carried out with the in vitro pharmacodynamic system. All experiments were single-dose studies that were simultaneously performed in duplicate. A peak concentration of 5 g/ml was chosen so that an equivalent intracellular-extracellular concentration gradient might be created for each quinolone. Terminal elimination half-lives of 4, 6, 6, 8, and 13 h were simulated for ciprofloxacin, clinafloxacin, ofloxacin, temafloxacin, and sparfloxacin, respectively. These fluoroquinolones represented a broad range of hydrophilic to hydrophobic properties. The initial inoculum in each experiment was 10 7 CFU of exponentially growing bacteria per ml. Temperatures were maintained between 35 and 37 C under aerobic or anaerobic incubation. Fifteen 1-ml samples were drawn from each system over 24 h at the following times: time zero; 30 s; 15, 30, and 45 min; and 1, 2, 3, 4, 6, 9, 12, 18, 21, and 24 h. Each sample was analyzed to determine the bacterial count, antibiotic concentration, ph, and temperature. Differential killing (defined as the difference in killing under aerobic and anaerobic conditions) was determined by subtracting the bacterial counts derived from anaerobic experiments from those derived from aerobic experiments. Differential killing after 1 h of exposure was used to compare the various quinolones, since 1 h seemed to allow enough time for the quinolones to differentiate themselves in terms of killing while still providing consistent results with a limited amount of inter- and intrarun variability. ph effect studies. Since the anaerobic environment reduced the medium, lowering its ph from 7.15 to 6.4, anaerobic time kill kinetic studies were repeated to determine the effect that reduced CSMHB had on fluoroquinolone activity. Two pharmacodynamic systems were run simultaneously, side by side, within the anaerobic chamber such that one system was run with reduced, nonbuffered CSMHB (ph 6.4) while the other was run with 0.1 M phosphate-buffered CSMHB (ph 7.15). The ph of the latter system was identical to that of nonbuffered CSMHB that had not been reduced by anaerobiosis. Experimental conditions (bacterial inoculum, temperature, half-life, and peak antibiotic concentration) were maintained as described above. Fluoroquinolone uptake studies. Since carbonyl cyanide m-chlorophenylhydrazone (CCCP) inhibits bacteria from actively transporting substances across their cell walls, bacteria were exposed to CCCP to investigate the possibility of an active, energy-dependent transport mechanism in S. aureus ATCC These uptake studies were performed by using methods previously described by Kaatz et al. (11). Stock [ 3 H]norfloxacin (47.2 mci/mg; provided by Merck, Rahway, N.J.) was diluted to a specific activity of 3.25 mci/mg by adding unlabeled drug. Uptake studies then were performed in triplicate by using [ 3 H]norfloxacin at a final concentration of 0.39 g/ml. These studies evaluated [ 3 H]norfloxacin uptake 15 min before (preincubation) and 15 min after addition of the metabolic inhibitor CCCP. Preincubating the bacteria with CCCP in this manner would prevent bacteria from actively taking up the antibiotic, while not preincubating the bacteria and treating them with CCCP at 15 min after antibiotic exposure would allow the bacteria to actively take up the drug. If, indeed, there was not a significant difference in antibiotic uptake between preincubated and nonpreincubated experiments, one could conclude that an active transport process does not exist in S. aureus ATCC Three time-kill kinetic studies were performed with CCCP. CCCP was used to mimic the effect that anaerobiosis has on fluoroquinolone active transport. Ofloxacin was used as the prototype fluoroquinolone, since it represented a midrange of fluoroquinolone hydrophobicities. First, control experiments were performed by exposing S. aureus ATCC to CCCP. In the next set of experiments, S. aureus was exposed to ofloxacin after 2 h of preincubation in CCCP. The third set of experiments was identical to the second, except that S. aureus was not preincubated with CCCP prior to exposure to ofloxacin. Rather, ofloxacin and CCCP were added to the broth culture of S. aureus simultaneously. The duration of each experiment was 8 to 10 h. Assay. A sample preparation and high-pressure liquid chromatography method which had been previously developed to analyze samples existing in CSMHB was used to measure concentrations of ciprofloxacin, clinafloxacin, ofloxacin, temafloxacin, and sparfloxacin (9). The isocratic mobile phase consisted of water, acetonitrile, phosphoric acid, sodium phosphate, sodium dodecyl sulfate, and acetohydroxamic acid. Separation on a Beckman Ultrasphere octyldecyl silane (5 m) column (4.6 mm by 15 cm), coupled with the detection TABLE 1. MICs and MBCs for S. aureus before and after antibiotic exposure Time and condition MIC/MBC ( g/ml) of: Ofloxacin Ciprofloxacin Clinafloxacin Temafloxacin Sparfloxacin Preexposure Aerobic 0.5/ / /2 0.5/ /0.12 Anaerobic 0.5/ /0.03 1/1 0.5/ /0.12 Postexposure Aerobic 1/1 0.12/0.12 2/2 0.5/0.5 ND a Anaerobic 0.5/ / /1 0.25/ /0.12 a ND, not done. Regrowth isolates were not recovered from aerobic sparfloxacin experiments because of contamination. capabilities of a Beckman 168 UV Detector (280 nm), produced assays which were linear for five successive serial dilutions ranging from to g/ml (R 2, 0.99). The lower limit of detection of the original assay was 0.5 g/ml, and the intraday coefficients of variation were less than 7.4%. We lowered the original curve levels to g/ml to encompass the lowest drug levels theoretically predicted and then actually observed in the experimental mode. The R 2 of the standard curves, the coefficients of variation of standards, and the coefficients of variation of the unknown analyses based upon an internal standard were as follows: ciprofloxacin, 0.992, 6.84%, and 5.16%; clinafloxacin, 0.997, 4.50%, and 6.33%; ofloxacin, 0.998, 2.45%, and 3.66%; temafloxacin, 0.998, 8.59%, and 8.91%; sparfloxacin, 0.998, 9.12%, and 4.54%. Analysis. Time-kill curves were plotted as logarithmic declines in CFU per milliliter versus time and were evaluated for (i) time to 3 log 10 killing, (ii) differential killing (defined as the difference between the averages of duplicate aerobic and anaerobic kill curves over the duration of the experiments), and (iii) extent of regrowth. Because of small sample sizes, inferential statistical methods were not applied to these data. Data from uptake studies were analyzed by using a one-way analysis of variance followed by Tukey s post-hoc test for multiple comparisons. RESULTS Susceptibility testing. MICs and MBCs of ciprofloxacin, clinafloxacin, ofloxacin, temafloxacin, and sparfloxacin against S. aureus ATCC are shown in Table 1. MICs and MBCs determined under aerobic conditions did not differ from those determined under anaerobic conditions. In general, MICs for isolates that regrew (24-h time point) during the experiments did not differ from those determined prior to antibiotic exposure, except for one notable exception. The MICs for isolates obtained from aerobic experiments done with clinafloxacin increased from to 0.12, an eightfold increase. Pharmacokinetic analysis. The first-order elimination of each drug was demonstrated in each experiment by a satisfactory fit to a log-linear relationship (R 0.99). Actual half-lives were similar to targeted values ( versus 4, versus 6, versus 6, versus 8, and versus 13 h for ciprofloxacin, clinafloxacin, ofloxacin, temafloxacin, and sparfloxacin, respectively). Actual mean peak concentrations were also similar to the target values. The target peak concentration was 5 g/ml, whereas the actual peak concentrations were , , , , and mg/ml for the respective quinolones. Time-kill kinetic studies. Bacteria grew exponentially from 10 5 to 10 9 CFU/ml throughout each growth control experiment, as indicated by a satisfactory fit to a log-linear relationship (R 0.99). Each fluoroquinolone demonstrated bactericidal activity under both aerobic and anaerobic conditions. However, the bactericidal activity of each fluoroquinolone was delayed by anaerobiosis. Ciprofloxacin was bactericidal after 1 h under aerobic conditions, whereas the drug required 4htoachieve a3log 10

3 VOL. 39, 1995 FLUOROQUINOLONE ACTIVITY UNDER ANAEROBIC CONDITIONS 509 FIG. 1. Time-kill kinetic curves of ciprofloxacin, clinafloxacin, ofloxacin, temafloxacin, and sparfloxacin versus S. aureus ATCC decline under anaerobic conditions. Corresponding results for clinafloxacin, ofloxacin, temafloxacin, and sparfloxacin were 0.08 versus 0.8, 0.5 versus 2.5, 0.3 versus 2, and 2 versus 4 h, respectively. The effects of anaerobiosis on fluoroquinolone activity are exemplified by the kill curves shown in Fig. 1 and summarized by the data in Table 2. Ciprofloxacin was the most hydrophilic fluoroquinolone studied, with a partition coefficient of The initial differential killing (defined as the difference in bacterial killing under aerobic and anaerobic conditions between time zero and 4 h) for ciprofloxacin was the largest of the five fluoroquinolones studied. Clinafloxacin is slightly more hydrophobic than ciprofloxacin, with a partition coefficient of The initial differential killing for clinafloxacin was slightly less than that for ciprofloxacin. Ofloxacin s partition coefficient is 0.34; ofloxacin demonstrated initial differential killing slightly less than that of clinafloxacin. Temafloxacin is more hydrophobic still, with a partition coefficient of 0.43, and the initial differential killing

4 510 ZABINSKI ET AL. ANTIMICROB. AGENTS CHEMOTHER. TABLE 2. Fluoroquinolone activity in terms of time to 3 log 10 killing and bacterial regrowth under aerobic and anaerobic conditions Drug Time to 3 log killing (h) 24-h regrowth (mean log 10 CFU/ml SD) O 2 No O 2 Aerobic Anaerobic Ciprofloxacin Clinafloxacin Ofloxacin Temafloxacin Sparfloxacin 2 4 ND a a ND, not done. for temafloxacin was the lowest of all of the fluoroquinolones tested. Sparfloxacin was the most hydrophobic fluoroquinolone studied, with a partition coefficient of The initial differential killing after 1 h of exposure to sparfloxacin was slightly less than that caused by temafloxacin. Differential killing (i.e., the difference in antimicrobial activity at each time point under aerobic and anaerobic conditions) caused by anaerobiosis is plotted in Fig. 2. As can be seen by the inserted graph in Fig. 2, differential killing was less predictable after 4 h of antibiotic exposure. It is interesting that even though clinafloxacin killed at a significantly more rapid rate than ciprofloxacin under both aerobic and anaerobic conditions, initial differential killing after 1 h of exposure was nearly the same for ciprofloxacin and clinafloxacin. These results are consistent with what one would expect on the basis of an evaluation of the relative hydrophobicities of these two agents. Bacterial regrowth was evident in each 24-h experiment. Anaerobiosis did not affect the extent or the slope of bacterial regrowth curves. Bacterial regrowth data are summarized in Table 2. ph effect studies. Through ph monitoring, it was found that reduction of CSMHB by the anaerobic environment resulted in a change in the ph of the medium from 7.15 under aerobic conditions to 6.4 under anaerobic conditions. The effect of this TABLE 3. Uptake of [ 3 H]norfloxacin by S. aureus ATCC exposed to CCCP prior to [ 3 H]norfloxacin (preincubated) or not exposed until 15 min after [ 3 H]norfloxacin addition (nonpreincubated) Time (h) Mean [ 3 H]norfloxacin uptake SD (ng/ mg of cell protein) Nonpreincubated Preincubated P value NS a NS NS NS NS NS a NS, no statistically significant difference. ph change was investigated to determine if it could explain the differential killing seen under anaerobic conditions. The differential killing at 1 h due to this ph change was 0.92 log 10 CFU/ml for ciprofloxacin. This differential of 0.92 log 10 CFU/ml represents the amount of differential killing that could be attributed to the drop in ph caused by anaerobic reduction of CSMHB. By buffering of the medium, this drop in ph was partially reversed, but it was still found that the ph change caused by the broth could not fully explain the total differential killing observed after 1 h of exposure. This ph change accounted for only 0.92 log 10 CFU/ml (or 55%) of the differential killing caused by the anaerobic environment for ciprofloxacin. The differential killing caused by ph was 46% of that caused by anaerobiosis for ofloxacin and 15% of that for temafloxacin. Fluoroquinolone uptake studies. Statistical analysis of uptake studies (Table 3 and Fig. 3) showed a statistically significant difference between uptake of [ 3 H]norfloxacin by S. aureus bacteria that were exposed to CCCP prior to [ 3 H]norfloxacin (preincubated) and that of those which were not exposed until 15 min after addition of [ 3 H]norfloxacin (nonpreincubated). Except for the 30-s time point, beyond 3 min, the amount of FIG. 2. Differential killing (defined as the difference in killing observed under aerobic and anaerobic conditions) from 0 to 4 h caused by anaerobiosis. The insert is a graph of differential killing over the entire 24 h of the experiment. The numbers in parentheses represent partition coefficients. FIG. 3. Uptake of [ 3 H]norfloxacin by S. aureus ATCC exposed to CCCP prior to [ 3 H]norfloxacin (preincubated) and bacteria not exposed until 15 min after addition of [ 3 H]norfloxacin (nonpreincubated).

5 VOL. 39, 1995 FLUOROQUINOLONE ACTIVITY UNDER ANAEROBIC CONDITIONS 511 FIG. 4. Effect of CCCP on the activity of ofloxacin against S. aureus ATCC , chemostatic dilution; E, S. aureus plus CCCP (growth control); F, S. aureus plus ofloxacin;, S. aureus plus ofloxacin plus CCCP;, S. aureus plus ofloxacin plus CCCP (preincubated for 2 h with CCCP). norfloxacin associated with preincubated cells was statistically greater than that associated with nonpreincubated cells. This is indicative of the presence of the low-level efflux process that is known to be present in S. aureus (12). Time-kill studies (Fig. 4) demonstrated that CCCP has an inhibitory effect on fluoroquinolone activity. Preincubation with CCCP slowed the rate and decreased the extent of killing by ofloxacin. When S. aureus was preincubated with CCCP, ofloxacin produced only 2 log 10 killing in 2 h, and when S. aureus was exposed to CCCP and ofloxacin without preincubation, ofloxacin produced 3 log 10 killing in 1 h. In contrast, when S. aureus was exposed to ofloxacin in the absence of CCCP, 3 log 10 killing was produced in 30 min with total killing of 4.5 log 10 CFU/ml. Growth control curves showed that a 50 M concentration of CCCP is bacteriostatic for S. aureus ATCC DISCUSSION Morrissey and Smith have concluded that fluoroquinolones require oxygen for bactericidal activity (14, 17, 18). In contrast, Cooper et al. have concluded that the lack of oxygen has no effect on the bactericidal activity of fluoroquinolones (4, 5). The results of the present study indicate that fluoroquinolones are bactericidal for S. aureus under anaerobic conditions. However, anaerobic conditions do diminish the rate at which fluoroquinolones kill S. aureus. The data from this and other studies (6, 7, 11, 15, 16, 19, 22, 23) support several explanations for the differential activity that the fluoroquinolones possess within aerobic and anaerobic environments. One explanation is that the intracellular accumulation of more hydrophobic quinolones is less affected by anaerobiosis than is that of those that are more hydrophobic. Another explanation involves self-promoted uptake and a difluorophenyl side chain. A third explanation involves intracellular ph changes, and still another explanation involves the paradoxic effect caused by inhibition of RNA synthesis. It is important to emphasize that the present study indicates that no single one of these explanations fully explains the differential killing caused by anaerobiosis. Rather, a number of factors must be considered, including hydrophobicity, intracellular ph, antibiotic concentration, and structure-activity relationships. The data in the present study indicate that fluoroquinolones with larger partition coefficients (more hydrophobic) are less affected by anaerobiosis than are fluoroquinolones that have lower partition coefficients (less hydrophobic). Fluoroquinolone hydrophobicity was not predictive of activity under aerobic or anaerobic conditions by itself, but when examined in relation to the initial (between 0 and 4 h) differential killing under aerobic versus anaerobic conditions, hydrophobicity was an important predictor of activity. However, hydrophobicity alone did not explain the variable effect that anaerobiosis had on quinolone activity (Fig. 2). Rather, other structure-activity relationships and physical properties undoubtedly contribute to the antimicrobial activity under anaerobic conditions in an undefined manner. Mitsuyama et al. have studied fluoroquinolone uptake in porin-deficient gram-negative bacilli and have concluded that hydrophobicity is an important predictor of activity against these organisms (16). However, the uptake of tosufloxacin, a quinolone that is only moderately hydrophobic (partition coefficient, 0.29), was found to be greater than that of more hydrophobic fluoroquinolones. Those researchers have speculated that the 2,4-difluorophenyl moiety of the naphthyridine nucleus of tosufloxacin plays an important role in outer membrane permeation and that the presence of this moiety is the reason why tosufloxacin was taken up to a greater extent than fluoroquinolones with higher partition coefficients. In the present study, temafloxacin was found to possess the lowest differential killing level of the five fluoroquinolones studied, even though temafloxacin (partition coefficient, 0.43) was significantly less hydrophobic than sparfloxacin (partition coefficient, 0.79). Temafloxacin, like tosufloxacin, possesses the 2,4- difluorophenyl moiety. These studies suggest that the presence of a difluorophenyl group at position 1, such as that found on temafloxacin (and tosufloxacin), may play an important role in bacterial uptake under conditions of anaerobiosis. Further, the similarities between the findings of the present study and those of the study done by Mitsuyama et al. suggest that similar uptake mechanisms were at work in both studies. This implies that the palliative membrane channels of S. aureus may be altered by anaerobiosis and that fluoroquinolone uptake under conditions of anaerobiosis may depend less upon diffusion through these channels and more upon diffusion or self-promoted uptake through the phospholipid bilayer (3, 16). Recently published work by Furet et al. supports another explanation for the differential effect caused by anaerobiosis (7). Those investigators demonstrated that intra- and extracellular phs have a profound effect on fluoroquinolone activity. Furet et al. explained that it is the negatively charged fluoroquinolone that interacts with DNA gyrase, producing the antimicrobial effect (7, 23). Further, this negatively charged species would become less prevalent as the intracellular ph decreased. An interesting speculation on the differential antimicrobial activities seen under aerobic versus anaerobic conditions would be that fluoroquinolone uptake by bacteria is influenced by an oxygen-dependent active transport process. Thus, less fluoroquinolone would be taken up by bacteria under anaerobic conditions, potentially explaining the observed differences in antibacterial activity. The uptake studies utilized in this investigation support the idea of an active efflux transport mechanism in S. aureus. By inhibiting active efflux, CCCP may prevent bacteria not only from eliminating the fluoroquinolone but also from eliminating metabolic waste products. Ultimately, accumulation of these waste products would cause

6 512 ZABINSKI ET AL. ANTIMICROB. AGENTS CHEMOTHER. intracellular acidification, which would diminish fluoroquinolone activity. The time-kill studies done with CCCP and anaerobiosis also support this supposition. That is, by inhibiting efflux, anaerobiosis may inhibit the clearance of metabolic by-products in the same way that CCCP does. The result may be similar in each case, i.e., intracellular acidification with diminished fluoroquinolone activity. However, examination of the pk a s of the fluoroquinolones studied still does not fully explain the variability in differential killing among the fluoroquinolones. Again, other structureactivity relationships (e.g., the difluorophenyl moiety) and physical properties (e.g., hydrophobicity) undoubtedly must also be considered. Crumplin and Smith have shown that a paradoxical effect on nalidixic acid activity occurs when concentrations that greatly exceed the MIC are used (6). Similar paradoxical effects have been described with ofloxacin, norfloxacin, and ciprofloxacin (15, 23). The paradoxical effect is thought to be caused by inhibition of RNA synthesis that precludes the action that the quinolone has on DNA gyrase (6). By applying these ideas to the present study, in which it was shown that anaerobiosis and CCCP may promote quinolone accumulation, one could theorize that anaerobiosis and CCCP may cause the accumulation of excessive intracellular concentrations of quinolones. These excessive concentrations may have caused a paradoxic effect that ultimately reduced the antibacterial activity of the quinolones. Oral broad-spectrum antibiotics with favorable pharmacokinetic profiles are highly desirable for a variety of infections often managed in an outpatient setting for extended periods of time. Often, many of these infections may harbor anaerobic or facultative bacteria. While the antibacterial activity of new antibiotics under aerobic conditions has been extensively researched, little attention has been directed at antibacterial activities under anaerobic or microaerophilic conditions. Data that have been generated have not explained the dynamic interactions among the bacteria, the environment, and the pharmacokinetics of the antibiotic. A more thorough understanding of these relationships and the physical and chemical properties of the antibiotic may allow more efficient antibiotic development. As new fluoroquinolones which possess potent activity against anaerobic bacteria are developed, further studies will be needed to investigate the pharmacodynamics of the antibiotic-bacterial interaction of these fluoroquinolones within anaerobic environments. ACKNOWLEDGMENTS We are grateful to David R. P. Guay, Kyle Vance-Byran, Diann Clarens, Frank Konstantinides, and Kristine A. Quart for support in developing the research set forth in the manuscript. We are also grateful to Josephine Ferry for assistance in preparing the manuscript. This study was supported in part by research grants from the ASHP Foundation and the RW Johnson Pharmaceutical Research Institute. When this study was conducted, R.A.Z. was a postdoctoral research fellow at the University of Minnesota and St. Paul-Ramsey Medical Center. REFERENCES 1. Becton Dickinson Microbiology Systems BBL GasPak Anaerobic Indicator, product package insert. Becton Dickinson & Co., Cockeysville, Md. 2. Bien, P. A. (Warner-Lambert/Parke-Davis) Personal communication. 3. Chapman, J. S., and N. H. Georgopapadakou Route of quinolone permeation in Escherichia coli. Antimicrob. Agents Chemother. 32: Cooper, M. A., J. M. Andrews, and R. Wise Bactericidal activity of sparfloxacin and ciprofloxacin under anaerobic conditions. J. Antimicrob. Chemother. 28: Cooper, M. A., J. M. Andrews, and R. Wise Letter. J. Antimicrob. Chemother. 29: Crumplin, G. C., and J. T. Smith Nalidixic acid: an antibacterial paradox. Antimicrob. Agents Chemother. 8: Furet, Y. X., J. Deshusses, and J.-C. Pechère Transport of pefloxacin across the bacterial cytoplasmic membrane in quinolone-susceptible Staphylococcus aureus. Antimicrob. Agents Chemother. 36: Garrison, M. W., K. Vance-Bryan, T. A. Larson, J. P. Toscano, and J. C. Rotschafer Assessment of effects of protein binding on daptomycin and vancomycin killing of Staphylococcus aureus by using an in vitro pharmacodynamic model. Antimicrob. Agents Chemother. 34: Herman, V. K., F. N. Konstantinides, R. A. Zabinski, J. C. Rotschafer, and F. B. Cerra HPLC analysis of ciprofloxacin in complex growth media, abstr. 336, p In Program and abstracts of the Annual Meeting of the Federation of the American Society of Experimental Microbiologists, Federation of the American Societies for Experimental Biology, Bethesda, Md. 10. Hirai, K., H. Aoyama, T. Irikura, S. Iyobe, and S. Mitsuhashi Differences in susceptibility to quinolones of outer membrane mutants of Salmonella typhimurium and Escherichia coli. Antimicrob. Agents Chemother. 29: Kaatz, G. W., S. M. Seo, and C. A. Ruble Mechanisms of fluoroquinolone resistance in Staphylococcus aureus. J. Infect. Dis. 163: Kaatz, G. W., S. M. Seo, and C. A. Ruble Efflux-mediated fluoroquinolone resistance in Staphylococcus aureus. Antimicrob. Agents Chemother. 37: Kayser, F. H., and J. Novak In vitro activity of ciprofloxacin against gram-positive bacteria. Am. J. Med. 82(Suppl. 4A): Lewin, C. S., I. Morrissey, and J. T. Smith The mode of action of quinolones: the paradox in activity of low and high concentrations and activity in the anaerobic environment. Eur. J. Clin. Microbiol. Infect. Dis. 10: Limb, D. I., D. J. W. Dabbs, and R. C. Spensor In vitro selection of bacteria resistant to the 4-quinolone agents. J. Antimicrob. Chemother. 19: Mitsuyama, J.-I., Y. Itoh, M. Takahata, S. Okamoto, and T. Yasuda In vitro antibacterial activities of tosufloxacin against and uptake of tosufloxacin by outer membrane mutants of Escherichia coli, Proteus mirabilis, and Salmonella typhimurium. Antimicrob. Agents Chemother. 36: Morrissey, I., and J. T. Smith Absence of bactericidal activity of ofloxacin and ciprofloxacin under anaerobic conditions, abstr. 155, p In 3rd International Symposium on New Quinolones. 18. Morrissey, I., and J. T. Smith Letter. J. Antimicrob. Chemother. 29: Nakanishi, N., S. Yoshida, H. Wakebe, M. Inoue, T. Yamaguchi, and S. Mitsuhashi Mechanisms of clinical resistance to fluoroquinolones in Staphylococcus aureus. Antimicrob. Agents Chemother. 35: National Committee for Clinical Laboratory Standards Approved standard. M7-A2. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically, 2nd ed. National Committee for Clinical Laboratory Standards, Villanova, Pa. 21. Shen, L. L. (Abbott Laboratories) Personal communication. 22. Shen, L. L., L. A. Mitscher, P. N. Sharma, T. J. O Donnell, D. T. W. Chu, C. S. Cooper, T. Rosen, and A. G. Pernet Mechanism of inhibition of DNA gyrase by quinolone antibacterials: a cooperative drug-dna binding model. Biochemistry 28: Smith, J. T., and C. S. Lewin Chemistry and mechanism of action of the quinolone antibacterials, p In V. T. Andriole (ed.), The quinolones. Academic Press, San Diego, Calif. 24. Wolfson, J. S., and D. C. Hooper Fluoroquinolone antimicrobial agents. Clin. Microbiol. Rev. 2: Zabinski, R. A., A. J. Larsson, K. J. Walker, S. S. Gilliland, and J. C. Rotschafer Elimination of quinolone antibiotic carryover through use of antibiotic-removal beads. Antimicrob. Agents Chemother. 37: Zabinski, R. A., K. Vance-Bryan, A. J. Krinke, K. J. Walker, J. A. Moody, and J. C. Rotschafer Evaluation of activity of temafloxacin against Bacteroides fragilis by an in vitro pharmacodynamic system. Antimicrob. Agents Chemother. 37: