Garenoxacin (BMS-284756) and Moxifloxacin in Experimental Meningitis Caused by Vancomycin-Tolerant Pneumococci

Abstract

The emergence of multidrug-resistant strains of Streptococcus pneumoniae drives the development and evaluation of new antipneumococcal agents, especially for the treatment of bacterial meningitis. The aims of the present study were to assess the antibacterial effectiveness of two new quinolones, garenoxacin (BMS; BMS-284756) and moxifloxacin (MOX) in experimental meningitis caused by a vancomycin (VAN)-tolerant S. pneumoniae strain and to compare the results with those obtained by therapy with VAN and ceftriaxone (CRO) in combination. Meningitis was induced in young male New Zealand White rabbits by intracisternal inoculation of a VAN-tolerant pneumococcal strain (strain Tupelo) from a child with meningitis. Sixteen hours after inoculation, therapy was given by intravenous administration of BMS at 20 mg/kg of body weight, followed 5 h later by administration at a dosage of 10 mg/kg (n = 9 animals) or MOX as two doses of 20 mg/kg every 5 h (n = 8 animals). For comparison, we studied the following groups: (i) animals treated with VAN (20 mg/kg every 5 h, three doses) and CRO (125 mg/kg, one dose) (n = 9), (ii) animals infected with a VAN-tolerant strain but not treated (n = 8), (iii) animals infected with a VAN-tolerant pneumococcus isolated from the nasopharynx of a carrier and treated with BMS (n = 8), and (iv) animals infected with a cephalosporin-resistant type 6B S. pneumoniae strain and treated with BMS (n = 6). The MICs of penicillin, CRO, VAN, BMS, and MOX for the Tupelo strain were 2, 1, 0.5, 0.06, and 0.03 μg/ml, respectively. The rates of killing of strain Tupelo (the change in the log₁₀ number of CFU per milliliter per hour) in cerebrospinal fluid at 5 h were −0.70 ± 0.35, −0.61 ± 0.44, and −0.49 ± 0.36 for BMS, MOX, and VAN-CRO, respectively. Therapy with BMS and MOX was as effective as therapy with VAN-CRO against VAN-tolerant pneumococcal meningitis in rabbits.

The emergence of multidrug-resistant pneumococci and, more recently, of pneumococci tolerant to vancomycin (VAN) are important public health concerns worldwide (6, 12; R. M. Atkinson, J. Sublett, K. M. Edwards, and E. I. Tuomanen, 40th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1776, p. 105, 2000; A. Marchese, S. Gianoglio, M. Dolcino, E. Tonoli, E. Debbia, and G. C. Schito, 40th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1778, p. 105, 2000). The bactericidal effectiveness of therapy with VAN and an expanded-spectrum cephalosporin against meningitis caused by these VAN-tolerant strains may be compromised by the potential relapse of infection (9). Evaluation of 138 pneumococcal isolates from the Active Bacterial Core Surveillance activity of the Centers for Disease Control and Prevention in 1998 revealed that up to 5% of pneumococci causing meningitis are VAN tolerant (C. A. Rodriguez, C. G. Whitney, and E. I. Tuomanen., 40th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1777, p. 105, 2000).

Two new quinolones, garenoxacin (BMS; BMS-284756) and moxifloxacin (MOX), were studied as therapy against meningitis caused by VAN-tolerant Streptococcus pneumoniae. BMS, a des-F(6) quinolone, displays broad antimicrobial activity and enhanced in vitro activity against S. pneumoniae (5, 17). In animal studies MOX was found to have bactericidal activity against penicillin- and cephalosporin-resistant pneumococci causing meningitis (13, 15). The aims of our study were to assess the bacteriologic effectiveness of these agents against experimental meningitis caused by VAN-tolerant pneumococci and to compare the results with those obtained by conventional therapy with VAN and ceftriaxone (CRO) in combination.

(This study was partially presented at the 41st Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, Ill., 16 to 19 December 2001.)

MATERIALS AND METHODS

Bacterial strains.

The VAN-tolerant strains tested, a meningitic isolate (strain Tupelo) and a nasopharyngeal isolate, were generously provided by Elaine Tuomanen, St. Jude Children's Research Hospital, Memphis, Tenn. Strain Tupelo (serotype 14) was obtained from a child with recurrent meningitis caused by this isolate (9). The nasopharyngeal isolate (isolate P9802-020; serotype 29) was collected from a child at Vanderbilt University Hospital, Nashville, Tenn. (Atkinson et al., 40th ICAAC). For comparison, a VAN-susceptible, cephalosporin-resistant type 6B isolate was used. This strain was originally isolated from an infant with meningitis (4).

After intrathecal passage in rabbits, the strains were grown overnight on tryptic soy agar with 3% sheep blood. The plates were washed with endotoxin-free phosphate-buffered saline (PBS), and aliquots of the resultant suspension were frozen at −70°C. For preparation of the inoculum, aliquots were thawed and diluted in endotoxin-free PBS to a concentration of approximately 1 × 10⁵ to 5 × 10⁵ CFU/ml. The inoculum size was confirmed by quantitative cultures in each experiment. For growth in liquid medium, bacteria were grown in Mueller-Hinton broth supplemented with 5% lysed horse blood.

Susceptibility tests.

The MICs of the study antibiotics for these strains were determined by a microdilution method, as recommended by the National Committee for Clinical Laboratory Standards (NCCLS) (11). For strain Tupelo, the E-test was also used to perform susceptibility studies. The minimal bactericidal concentrations (MBCs) of the study antibiotics for the VAN-tolerant isolates and the VAN-susceptible type 6B isolate were also determined by the NCCLS microdilution methodology.

Meningitis model.

The rabbit meningitis model used was modified from the model originally described by Dacey and Sande (2). Approval of the study protocol was obtained from the Institutional Animal Care and Research Advisory Committee at our institution. Young male New Zealand White rabbits weighing 2 to 2.5 kg were anesthetized with intramuscular ketamine (50 mg/kg of body weight) and acepromazine (4 mg/kg) before every procedure and frequently while they were restrained in frames. After withdrawal of cerebrospinal fluid (CSF) (250 μl), meningitis was induced by intracisternal injection of 1 × 10⁵ to 5 × 10⁵ CFU of bacteria per ml in 250 μl of endotoxin-free PBS. Approximately 16 h later the anesthetized animals underwent intracisternal taps to quantify the baseline bacterial concentrations in the CSF. Antibiotic therapy was then administered via a marginal ear vein (0 h). The animals were immobilized in stereotactic frames, and a spinal needle remained in the cisterna magna for frequent CSF sampling during the first 3 h after the receipt of antibiotics. The rate of removal of CSF did not exceed the rate of CSF formation, which is approximately 0.4 ml/h (16). Blood samples were drawn from a central ear artery. Flunixin meglumine (1.1 mg/kg) was administered intramuscularly every 12 h for analgesia. The animals were euthanized with pentobarbital (120 mg/kg) at the end of each experiment or earlier if they were severely lethargic or not recumbent.

Treatment.

BMS (Bristol-Myers Squibb Co., Princeton, N.J.) was administered at an initial dosage of 20 mg/kg, followed 5 h later by administration at a dosage of 10 mg/kg (n = 9 animals). MOX (Bayer AG, Wuppertal, Germany) was administered twice at 20 mg/kg, with administration of the two doses separated by 5 h (n = 8 animals). For comparison, nine animals received the combination of VAN (20 mg/kg every 5 h for three doses) and CRO (125 mg/kg as a single dose).

These dosage regimens were selected on the basis of their abilities to achieve concentrations in serum comparable to those achieved in humans, as previously determined in our laboratory (1, 4, 7, 14) and by others (13). All antibiotics were prepared according to the instructions of the manufacturers. Eight animals that were inoculated with strain Tupelo but that did not receive antibiotic therapy were used as controls. Eight animals were inoculated with the nasopharyngeal VAN-tolerant isolate and received BMS therapy; in addition, six animals were inoculated with the type 6B pneumococci and received BMS therapy. In a second group of experiments, three groups of animals infected with strain Tupelo received therapy (VAN, n = 7; VAN-CRO, n = 5; BMS, n = 5) and were monitored for 72 h to evaluate the animals for CSF relapses.

Sample collection and processing.

After the antibiotic(s) was given, CSF (150 to 200 μl) and blood (700 μl) were obtained at 1, 2, 3, 5, 10, and 24 h (specimens were also collected at 48 and 72 h in the second group of experiments). Blood and CSF samples were centrifuged at 5,000 × g for 10 and 5 min, respectively, and the supernatants were stored at −70°C. An additional 100 to 150 μl of CSF was collected before therapy for bacterial quantification. The concentrations of bacteria were determined at 0, 3, 5, 10, and 24 h posttherapy by plating undiluted and serial dilutions of CSF (100 μl) on sheep blood agar and incubating the plates in 5% CO₂ at 35°C for 24 h. The lowest bacterial concentration detectable by this method was 10 CFU/ml. For purposes of analysis, specimens with <10 CFU/ml were assigned a value of 1 (0 log₁₀) CFU/ml. Bacterial killing rates (BKRs) were calculated as the difference between bacterial concentrations at the start of therapy and at 3, 5, 10, and 24 h divided by time and was expressed as the change in the log₁₀ number of CFU per milliliter per hour. The total changes in the log₁₀ number of CFU per milliliter were also determined at 5, 10, and 24 h.

VAN susceptibility and autolysis rates.

By a killing curve method we determined autolysis rates and viability using 10-ml cultures of S. pneumoniae exposed to 10 times the MIC of VAN when the optical density at 620 nm reached 0.25 to 0.3 (corresponding to 5 × 10⁷ CFU/ml) (12). After different exposure times, 100-ml aliquots were serially diluted in Mueller-Hinton broth and plated for determination of viable counts on tryptic soy broth agar. For 6 h, colony counts and optical densities were measured every 1 and 2 h, respectively.

DNA fingerprinting by pulsed-field gel electrophoresis.

The DNA fingerprint profiles observed by pulse-field gel electrophoresis determined that the two tolerant strains (the meningitic and nasopharyngeal strains) obtained after passage through the rabbits were similar to the original isolates and were considered genetically related (10). The serotype and penicillin-binding protein profiles remained the same for all the isolates.

Statistical analysis.

Comparisons between two groups were performed by the t test and the Mann-Whitney test if the data were normally distributed and not normally distributed, respectively. Comparisons among three or more groups were performed by the Kruskal-Wallis test, followed by Dunn's multiple-comparisons test among groups when the data were significantly different. A P value of <0.05 was considered significant. Data are expressed as means ± standard deviations.

RESULTS

In vitro susceptibility.

By the E-test, the MICs of penicillin, CRO, and VAN for strain Tupelo were 2.0, 0.75, and 0.5 μg/ml, respectively. By the NCCLS microdilution method, the MICs of the study antibiotics for the Tupelo strain were as follows: CRO, 1 μg/ml; VAN, 0.5 μg/ml; BMS, 0.06 μg/ml; and MOX, 0.03 μg/ml. The MBCs of BMS and MOX for strain Tupelo were 0.125 and 0.06 μg/ml, respectively. For the nasopharyngeal isolate the MICs were 1 μg/ml for CRO, 0.5 μg/ml for VAN, 0.03 μg/ml for BMS, and 0.125 μg/ml for MOX and the MBCs were 0.25 and 0.5 μg/ml for BMS and MOX, respectively.

The MICs and MBCs of the study antibiotics for the VAN-susceptible type 6B isolate were as follows: penicillin, 2 and 2 μg/ml, respectively; CRO, 4 and 4 μg/ml, respectively; VAN, 0.25 and 0.25 μg/ml, respectively; and BMS, 0.06 and 0.06 μg/ml respectively.

Antimicrobial activity in the 24-h experiments.

The levels of bacterial killing over a 24-h period after BMS, MOX, and VAN-CRO therapy for meningitis caused by VAN-tolerant and VAN-nontolerant pneumococci are shown in Table 1. The bacterial concentrations in the CSF of animals infected with strain Tupelo that were left untreated (controls) and those that were treated with BMS, MOX, or VAN-CRO are illustrated in Fig. 1. The baseline colony counts were similar for all groups. For strain Tupelo, significant differences in colony counts were detected only at 3 and 5 h for animals treated with BMS and VAN-CRO (P = 0.017 and 0.036, respectively). The BKRs and total killing at 10 and 24 h were not different. The BKR of strain Tupelo by BMS at 0 to 3 h was significantly lower than that of the VAN-tolerant nasoparyngeal isolate (P = 0.009).

TABLE 1.

BKRs and changes in bacterial counts at 5, 10, and 24 h in experimental meningitis caused by VAN-tolerant S. pneumoniae isolates treated with BMS, MOX, and VAN-CRO and comparison of results with those for a VAN-susceptible type 6B strain

Organism, therapy (n)a	BKR (change in log10 CFU/ml/h) atb:			Change in bacterial log10 CFU/ml atb:
	0-3 h	0-5 h	0-10 h	5 h	10 h	24 h
VTSP, Tupelo
BMS (9)	−0.79 ± 0.48	−0.70 ± 0.35	−0.50 ± 0.12	−3.50 ± 1.76	−5.02 ± 1.22	−5.07 ± 1.08
MOX (8)	−0.92 ± 0.87	−0.61 ± 0.44	−0.50 ± 0.14	−3.05 ± 2.22	−4.97 ± 1.39	−5.42 ± 0.82
VAN-CRO (9)	−0.70 ± 0.62	−0.49 ± 0.36	−0.48 ± 0.16	−2.47 ± 1.80	−4.77 ± 1.63	−6.09 ± 1.15
VTSP, NP, BMS (8)	−1.45 ± 0.42	−0.99 ± 0.14	−0.52 ± 0.07	−4.92 ± 0.69	−5.23 ± 0.69	−5.23 ± 0.69
CRSP type 6B, BMS (6)	−0.95 ± 0.3	−0.79 ± 0.28	−0.65 ± 0.07	−3.95 ± 1.41	−6.52 ± 0.68	−6.52 ± 0.68

^aAbbreviations: VTSP, VAN-tolerant S. pneumoniae; NP, nasopharyngeal; CRSP, cephalosporin-resistant S. pneumoniae.

^bThe results are means ± standard deviations.

FIG. 1.

Bacterial concentrations (means ± standard deviations) in CSF after the administration of different regimens in experimental meningitis caused by VAN-tolerant S. pneumoniae strain Tupelo. Animals were not treated (controls; solid diamonds) or were treated with BMS at 20 mg/kg, followed 5 h later by 10 mg/kg (solid circles), MOX at 20 mg/kg twice with the dosage administrations separated by 5 h (open triangles), or VAN at 20 mg/kg every 5 h for three doses in combination with CRO at 125 mg/kg as a single dose (open squares).

We compared the bacterial concentrations in the CSF of animals that were infected with strain Tupelo and that were left untreated (controls) and compared them with those in animals treated with the different regimens (Fig. 1). The BMS-treated animals had significantly greater reductions in bacterial titers in CSF than the untreated group at 3, 5, 10, and 24 h (P = 0.031, 0.029, 0.003, and 0.021, respectively). The counts for the controls at 10 and 24 h were significantly larger than those for animals treated with MOX (P = 0.005 and 0.021, respectively). The counts for controls at 10 and 24 h were significantly larger than those for animals treated with VAN-CRO (P = 0.014 and 0.01, respectively). The maximal BKR of −1.45 CFU/ml/h was observed with BMS therapy for the VAN-tolerant nasopharyngeal isolate.

Comparison of BMS therapy for meningitis caused by strain Tupelo versus that for meningitis caused by a type 6B isolate (VAN susceptible) showed that the BKR at 0 to 10 h and the changes in bacterial counts at 10 and 24 h were significantly (P = 0.009) greater against the type 6B isolate (Table 1 and Fig. 2).

FIG. 2.

Bacterial concentrations in CSF after therapy with BMS (20 mg/kg, followed 5 h later by 10 mg/kg) in rabbits with meningitis caused by VAN-tolerant S. pneumoniae strain Tupelo (solid circles) and nasopharyngeal strains (open squares) and comparison of the results with those obtained with a VAN-susceptible type 6B strain (open triangles).

Meningitis caused by strain Tupelo in the 72-h experiment.

Three groups of animals with meningitis received antibiotic therapy for only 1 day with VAN alone, VAN-CRO, or BMS. The CSF from these animals was evaluated up to 72 h after the start of therapy. The changes in bacterial counts at 3, 5, 10, 24, 48, and 72 h are shown in Table 2. In two of five rabbits treated with BMS, bacterial regrowth occurred at 48 h, with a lack of clearance at 72 h. The BKRs and total killing at 3, 5, 10, 24, 48, and 72 h were similar for the three regimens (VAN, VAN-CRO, and BMS).

TABLE 2.

Change in bacterial counts at various intervals after start of therapy in experimental meningitis caused by a VAN-tolerant pneumococcus (strain Tupelo) treated with different antibiotics

Therapy (n)	Change in bacterial log10 CFU/ml ata:
	3 h	5 h	10 h	24 h	48 h	72 h
VANb (7)	−1.98 ± 1.34	−4.65 ± 2.16	−6.04 ± 1.28	−6.04 ± 1.28	−6.04 ± 1.28	−6.04 ± 1.28
VAN-CROc (5)	−2.23 ± 1.40	−4.71 ± 1.20	−5.63 ± 1.40	−6.52 ± 1.77	−6.52 ± 1.77	−6.52 ± 1.77
BMSd (5)	−2.13 ± 1.55	−4.89 ± 1.62	−5.94 ± 0.75	−6.36 ± 0.99	−4.18 ± 2.96	−4.28 ± 2.85

^aTimes are from the start of treatment, and the data are expressed as means ± standard deviations.

^bVAN at 20 mg/kg every 5 h for three doses.

^cVAN at 20 mg/kg every 5 h for three doses and CRO at 125 mg/kg as a single dose.

^dBMS at 20 mg/kg followed 5 h later by 10 mg/kg.

Vancomycin susceptibility.

After 4 h of exposure to VAN at 10 times the MIC (5 μg/ml), the killing of both VAN-tolerant strains was ≤2 log₁₀ CFU/ml: 1.95 and 1.61 log₁₀ CFU/ml for strain Tupelo and the nasopharyngeal isolate, respectively. In contrast, the killing of the VAN-susceptible type 6B strain was 4.86 log₁₀ CFU/ml after 4 h of exposure to VAN at 10 times the MIC (2.5 μg/ml).

DISCUSSION

Antibiotic tolerance, a phenomenon distinct from antibiotic resistance, was first described in pneumococci in 1970 (18). Tolerance is the ability of bacteria to survive in the presence of an antibiotic without growing or undergoing lysis. Tolerance occurs if either the pneumococcal autolysin, which lyses the cell wall, is not triggered or the autolysin is not present. Recently, Novak et al. proposed a molecular mechanism for tolerance to VAN and other classes of antibiotics that involves the loss of function of the VncS histidine kinase of a two-component signal transduction system, which may regulate a basic pathway triggering autolysis. This mechanism includes the transport-sensor-regulator locus Vex-VncS/R (12; Atkinson et al., 40th ICAAC; L. L. Grinius, W. Coleman, B. Desai, H. Li, and D. A. Morrison, 41st Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1509, p. 96, 2001).

Because of their excellent antipneumococcal activities and good penetration into inflamed meninges, the new quinolones such as gatifloxacin, BMS, and MOX have been shown to be effective for the treatment of experimental meningitis caused by penicillin- and cephalosporin-resistant pneumococci (8, 14, 15). BMS demonstrated good bacterial killing in CSF in experimental VAN-tolerant S. pneumoniae meningitis, and MOX therapy was comparably effective in our model. Although two of five animals treated with BMS had regrowth of S. pneumoniae Tupelo at 48 to 72 h but none of the animals treated with VAN or VAN-CRO had regrowth, only two doses of BMS (given every 5 h) were administered, whereas three doses of VAN (given at 0, 5, and 10 h) were given with the VAN regimens.

We were unable to demonstrate the tolerance of strain Tupelo in our meningitis model. However, our animal studies were not optimally designed to confirm that strain Tupelo is a VAN-tolerant isolate, according to the in vitro definition. In in vitro studies VAN is usually added in the early exponential phase of growth (∼4 h). In our animal studies VAN and the other study drugs were given 16 h after inoculation to compare the different antimicrobial therapies. That time was chosen because the concentrations of organisms in CSF are comparable to those seen at the time of diagnosis of pneumococcal meningitis in children (3). In patients with meningitis the phase of growth of S. pneumoniae in CSF is highly variable and depends on the time of diagnosis and the start of treatment.

The optimal treatment of meningitis caused by VAN-tolerant pneumococci is unknown and will be difficult to establish clinically. This is because most clinical laboratories are unable to test properly for the VAN tolerance of pneumococci, universal immunization with conjugate pneumococcal vaccines will substantially reduce the rates of invasive pneumococcal disease, and meningitis caused by tolerant pneumococci is very uncommon. Thus, data derived from animal studies such as those described here will be necessary as a surrogate guide for patient management.