Gemifloxacin Is Effective in Experimental Pneumococcal Meningitis

Abstract

In a rabbit model of Streptococcus pneumoniae meningitis, 5 mg of gemifloxacin mesylate (SB-265805) per kg/h reduced the bacterial titers in cerebrospinal fluid (CSF) almost as rapidly as 10 mg of ceftriaxone per kg/h (Δlog CFU/ml/h ± standard deviation [SD], −0.25 ± 0.09 versus −0.38 ± 0.11; serum and CSF concentrations of gemifloxacin were 2.1 ± 1.4 mg/liter and 0.59 ± 0.38 mg/liter, respectively, at 24 h). Coadministration of 1 mg of dexamethasone per kg did not affect gemifloxacin serum and CSF levels (2.7 ± 1.4 mg/liter and 0.75 ± 0.34 mg/liter, respectively, at 24 h) or activity (Δlog CFU/ml/h ± SD, −0.26 ± 0.11).

Pneumococci moderately or highly resistant to penicillin G and other β-lactam antibiotics are a worldwide challenge. A reduced sensitivity for penicillin is paralleled by increases of the MICs for all β-lactam and carbapenem antibiotics, and clinical failures of cefotaxime and ceftriaxone in the treatment of meningitis caused by penicillin-resistant isolates of Streptococcus pneumoniae have been observed (1). Therefore, treatment options with antibacterials not belonging to the groups of β-lactams and carbapenems appear highly desirable.

The activity of older quinolones such as ciprofloxacin, ofloxacin, and levofloxacin against S. pneumoniae is not high enough to use these compounds in the treatment of pneumococcal meningitis (10). Newer compounds, however, such as trovafloxacin, moxifloxacin, grepafloxacin, and gatifloxacin possess improved in vitro and in vivo antipneumococcal activity (7, 10, 16, 17). Gemifloxacin (SB-265805) is highly active against S. pneumoniae, its MIC at which 90% of the isolates tested are inhibited (MIC₉₀) lying approximately one order of magnitude below the MIC₉₀s of trovafloxacin, moxifloxacin, grepafloxacin, gatifloxacin, and sparfloxacin (7; D. M. Johnson, R. N. Jones, D. J. Biedenbach, M. A. Pfaller, G. V. Doern, and The Quality Control Group, 38th Intersci. Conf. Antimicrob. Agents Chemother., poster F-103, 1998; L. M. Kelly, M. R. Jacobs, and P. C. Appelbaum, 38th Intersci. Conf. Antimicrob. Agents Chemother., poster F-087, 1998). Since resistance to β-lactam antibiotics is not associated with a reduced sensitivity to quinolones (Kelly et al., 38th ICAAC) and gemifloxacin has low MICs (0.03 to 1 μg/ml) against quinolone-resistant pneumococci (T. Davies, L. M. Kelly, M. R. Jacobs, and P. C. Appelbaum, Abstr. 39th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1497, 1999), gemifloxacin may prove useful for the treatment of both penicillin-sensitive and -resistant strains of S. pneumoniae.

Treatment of pneumococcal meningitis with the β-lactam antibiotic ceftriaxone leads to a rapid release of proinflammatory cell wall components and an increase of tumor necrosis factor alpha and interleukin-1 beta in cerebrospinal fluid (CSF) (13, 19). Cell wall components and cytokines are thought to contribute to neuronal damage in bacterial meningitis (5).

The present study addresses whether the new quinolone gemifloxacin possesses adequate in vivo activity for the treatment of S. pneumoniae meningitis and whether it modulates the inflammatory host response occurring after initiation of therapy. This study also examines whether the penetration of gemifloxacin into CSF is affected by the coadministration of dexamethasone.

(These data were presented, in part, as a poster at the 9th European Congress of Clinical Microbiology and Infectious Diseases, Berlin, Germany, 21 to 24 March, 1999.)

In vitro activity.

The MICs and minimum bactericidal concentrations (MBCs) of gemifloxacin and ceftriaxone for the S. pneumoniae type 3 strain used in this and previous studies (10, 13, 16, 20, 22) were determined by the macrodilution method in tryptic soy broth.

Rabbit model.

After intramuscular induction of anesthesia with ketamine (25 mg/kg) and xylazine (5 mg/kg), New Zealand White rabbits (approximately 2.5 kg) were anesthetized with intravenous (i.v.) urethane for 24 h and were placed in a stereotaxic frame. A 22- by 3.5-in. spinal needle (Spinocan; Braun, Melsungen, Germany) was placed in the cisterna magna.

Meningitis was induced by intracisternal injection of 10⁶ CFU of S. pneumoniae type 3. Blood (3 ml) and CSF (300 μl) were previously drawn and at 12, 14, 17, 20, and 24 h after infection. Beginning 12 h after infection, nine rabbits received a maintenance dose of 5 mg of gemifloxacin mesylate per kg/h i.v. for 12 h, and nine rabbits received the same dose and an adjunctive treatment of 1.0 mg of dexamethasone per kg (Fortecortin; Merck, Darmstadt, Germany) 15 min prior to initiation of antibiotic therapy. Eight animals received 1 mg of gemifloxacin mesylate per kg/h. Gemifloxacin therapy was started with an i.v. bolus dose of twice the maintenance dose per hour. Animals treated with 20 mg of ceftriaxone per kg (Rocephin; kindly provided by Hoffmann-LaRoche, Grenzach-Wyhlen, Germany) i.v. bolus followed by a 10 mg/kg/h maintenance dose served as controls (n = 12).

Sample processing.

CSF leukocytes were counted in a Fuchs-Rosenthal hemocytometer. After coagulation, blood was centrifuged at 3,000 × g for 5 min, and the supernatant was immediately frozen at −80°C. Pneumococcal CSF titers were counted by plating 10 μl of undiluted CSF and serial 10-fold dilutions on blood agar plates, which were then incubated overnight at 37°C in an atmosphere of 5% CO₂. To diminish carryover phenomena, 500 μl of 1:100-diluted CSF was plated onto a separate blood agar plate. Bacterial titers at 12, 14, 17, 20, and 24 h served for log-linear regression analysis. The first sterile sample was assigned a value of 2 (log₁₀ 100). The remaining CSF was centrifuged at 3,000 × g for 5 min, and the supernatants were stored at −80°C for less than 3 months. Storage in biological fluids at −20°C for 3 months did not result in a loss of activity (supporting data available from the manufacturer). The concentrations of gemifloxacin and ceftriaxone in serum and CSF were determined by the agar well diffusion technique in Mueller-Hinton agar with Bacillus subtilis (ATCC 6633) spores and Escherichia coli 108 (collection of H. Hof, Department of Medical Microbiology, University of Heidelberg, Mannheim, Germany) (13). For serum and CSF samples, different standard curves were constructed by using undiluted and 1:20-diluted rabbit serum. To avoid interassay variation, serum and CSF samples were measured in one assay each. The quantification limit of the gemifloxacin bioassay was 0.2 μg/ml in serum and 0.1 μg/ml in CSF, and its intraassay coefficient of variation was below 10% at concentrations of 0.4 and 3.1 μg/ml in CSF and serum, respectively. The detection limits of the ceftriaxone assay were 0.5 μg/ml in CSF and 1 μg/ml in serum. The three major metabolites of gemifloxacin are present in low concentrations in serum and are at least 16 times less active than the parent compound; therefore, they should not have any impact on bioassays (supporting data available from the manufacturer). Area under the concentration-time curve from 12 to 24 h (AUC_12–24h) in serum and CSF were calculated by the linear trapezoidal rule by using the gemifloxacin concentrations measured at 14, 17, 20, and 24 h. The 24-h AUC-MIC ratio was estimated by multiplying the AUC_12–24h by 2 and dividing it by the MIC. The CSF-to-serum concentration ratio at 24 h was taken as an approximation of steady state.

Neuron-specific enolase (NSE) concentrations in CSF, which have been shown to correlate with clinical outcome in children with bacterial meningitis (4), were determined by an immunoluminometric method (LIA-mat NSE Prolifigen; Byk-Sangtec, Dietzenbach, Germany). Lactate was measured enzymatically (Biosen, Dreieich, Germany), and the CSF protein concentration was measured photometrically (BCA-protein-Test; Pierce, Rockford, Ill.).

Lipoteichoic and teichoic (LTA/TA) CSF concentrations were measured in the CSF of rabbits receiving 5 mg of gemifloxacin mesylate per kg/h and 10 mg of ceftriaxone per kg/h by enzyme immunoassay (18, 19).

Measurement of neuronal damage in the hippocampal formation.

In situ tailing to detect DNA double-strand breaks was performed with paraffin-embedded tissue (22). Apoptotic neurons in the dentate gyrus of the hippocampal formation were counted on in situ tailing-stained sections and were related to the area of the granular cell layer measured on adjacent hematoxylin-and-eosin-stained sections. The density of apoptotic neurons was expressed as the number of marked neurons per square millimeter.

Statistics.

Data were described as means ± standard deviations (SD). Several groups were compared by two-tailed analysis of variance for independent samples. The Bonferroni method was used to correct for multiple comparisons. Two groups were compared by two-tailed t test.

Gemifloxacin and ceftriaxone had respective MICs of 0.015 and 0.03 μg/ml and respective MBCs of 0.015 and 0.06 μg/ml for S. pneumoniae type 3. The bacterial titer in CSF 12 h after infection (i.e., prior to the initiation of therapy) did not differ significantly among the treatment groups (Table 1). The gemifloxacin concentrations in serum and CSF during the infusion of different doses are shown in Table 1. Dexamethasone (1 mg/kg) did not reduce gemifloxacin concentrations in serum and CSF. The CSF-to-serum concentration ratio at 24 h as an approximation of steady state during i.v. application of 5 mg of gemifloxacin mesylate per kg/h was 0.28 ± 0.17 without and 0.33 ± 0.18 with coadministration of dexamethasone (difference not significant) (Table 1). At a dose of 1 mg/kg/h, the serum and CSF concentrations were approximately three- to fivefold lower, and the CSF-to-serum concentration ratio was unchanged (Table 1). The CSF-to-serum concentration ratio of ceftriaxone at 24 h after infection was 0.12 ± 0.13 (means ± SD) (P < 0.05 versus gemifloxacin).

TABLE 1.

Pharmacokinetics of gemifloxacin and ceftriaxone in serum and CSF and CSF leukocyte density and protein content in experimental meningitis caused by S. pneumoniae (means ± SD)

Antibiotic (dose in mg/kg/h)a	n	Antibiotic concentration (μg/ml) in serum at:		Antibiotic concentration (μg/ml) in CSF at:		CCSF/CS at 24 h	AUC12–24h (mg × h/liter) in serum	AUC12–24h (mg × h/liter) in CSF	AUC24h/MIC ratio (extrapolated from the AUC12–24h) in serum	AUC24h/MIC ratio (extrapolated from the AUC12–24h) in CSF
		14 h	24 h	14 h	24 h
CRO (10)	12	107.9 ± 38.1	135.2 ± 50.5	4.9 ± 3.1	14.6 ± 10.5	0.11 ± 0.13	NDc	ND	ND	ND
GEM (1)	8	0.7 ± 0.4	0.6 ± 0.4	0.17 ± 0.13	0.17 ± 0.09	0.29 ± 0.12b	7.1 ± 3.2	1.9 ± 1.7	952 ± 425	259 ± 233
GEM (5)	9	3.0 ± 2.0	2.1 ± 1.4	0.27 ± 0.07	0.59 ± 0.38	0.28 ± 0.17b	26.5 ± 16.7	5.4 ± 2.9	3535 ± 2228	713 ± 384
GEM (5) + DXM	9	2.7 ± 1.3	2.7 ± 1.4	0.29 ± 0.19	0.75 ± 0.34	0.33 ± 0.18b	31.1 ± 12.3	5.8 ± 2.0	4145 ± 1637	771 ± 260

Bacterial CSF titer (log CFU/ml) at 12h	Bactericidal activity (δlog CFU/ml/h)	Log CSF leukocytes at:		CSF protein (g/liter) at:		CSF lactate (mmol/liter) at:		Apoptotic neurons in the dentate gyrus (per mm2)	NSE in CSF (ng/ml)
		12 h	24 h	12 h	24 h	12 h	24 h
7.74 ± 0.72	−0.38 ± 0.11	2.61 ± 1.11	3.88 ± 0.52	0.93 ± 1.10	3.90 ± 1.81	4.3 ± 2.0	12.5 ± 7.3	165 ± 49	73.3 ± 84.0
7.50 ± 0.61	−0.15 ± 0.19b	2.34 ± 0.71	3.31 ± 0.14b	1.63 ± 0.93	5.37 ± 1.65	3.4 ± 1.1	14.5 ± 3.0	175 ± 134	134.2 ± 43.5
7.67 ± 0.46	−0.25 ± 0.09	2.10 ± 0.58	3.50 ± 0.26	1.41 ± 1.19	3.80 ± 2.10	3.1 ± 1.1	11.3 ± 1.8	142 ± 52	94.5 ± 67.2
7.57 ± 0.82	−0.26 ± 0.11	2.44 ± 0.49	3.53 ± 0.21	0.97 ± 0.36	2.85 ± 1.42	2.8 ± 0.8	11.8 ± 2.7	179 ± 100	99.1 ± 76.0

^aCRO, ceftriaxone; GEM, gemifloxacin; DXM, dexamethasone, 1 mg/kg. 

^bP < 0.05 versus animals receiving ceftriaxone. 

^cND, not determined. 

The bactericidal activity of 5 mg of gemifloxacin mesylate per kg/h without and with adjunctive treatment with dexamethasone was almost as high as the bactericidal rate of ceftriaxone (−0.25 ± 0.09 ΔlogCFU/ml/h and −0.26 ± 0.11 ΔlogCFU/ml/h versus −0.38 ± 0.11 ΔlogCFU/ml/h [means ± SD, differences not significant]). Gemifloxacin mesylate (1 mg/kg/h) was less effective (P < 0.05 versus ceftriaxone) (Table 1). Before the initiation of antibiotic therapy, LTA/TA CSF concentrations were 2.36 ± 0.37 log ng/ml in rabbits receiving 5 mg of gemifloxacin mesylate per kg/h and were 2.56 ± 0.61 in those receiving ceftriaxone (difference not significant). At 14 h, the CSF of gemifloxacin-treated rabbits contained 2.35 ± 0.46 log ng/ml versus 2.83 ± 0.52 log ng/ml (P < 0.05). No significant differences were observed at 17 h (2.55 ± 0.67 versus 2.72 ± 0.58 log ng/ml), 20 h (2.41 ± 0.60 versus 2.65 ± 0.59 log ng/ml), and 24 h (2.42 ± 0.66 versus 2.61 ± 0.47 log ng/ml).

The density of apoptotic neurons in the dentate gyrus was slightly lower in animals receiving 5 mg of gemifloxacin mesylate per kg/h, and the NSE concentrations in CSF as parameters of neuronal damage were not significantly different among the single treatment groups (Table 1).

In the rabbit model of experimental meningitis, gemifloxacin mesylate at a dose of 5 mg/kg/h was almost as active as 10 mg of ceftriaxone per kg/h against a penicillin-sensitive strain of S. pneumoniae. Since gemifloxacin retains relatively low MICs even in highly β-lactam- and quinolone-resistant pneumococci, it may be suitable for multiresistant pneumococci (3, 17; T. Davies et al., 39th ICAAC; D. M. Johnson et al., 38th ICAAC). Therefore, these results are relevant for the treatment of meningitis caused by penicillin-sensitive and -resistant S. pneumoniae strains. An i.v. form suitable for critically ill humans will be developed (supporting data available from the manufacturer). We administered gemifloxacin by continuous infusion to facilitate the comparison with other compounds which have been studied by using a continuous i.v. infusion in this model (10, 13, 16, 20). In clinical practice, the use of continuous instead of bolus infusions at the same daily dose would reduce the maximum concentrations in serum and might be a strategy to decrease the incidence of neurotoxic side effects occurring with high serum concentrations of several quinolones (6). The dose of 5 mg of gemifloxacin per kg/h was chosen with respect to serum concentrations achieved in humans after single doses of up to 800 mg (4.33 ± 0.63 μg/ml) without serious side effects (supporting data available from the manufacturer). The serum levels observed by us during the infusion of 5 mg of gemifloxacin per kg/h were slightly lower than these concentrations. Since the anticipated daily dose for use in humans is lower than 800 mg of gemifloxacin mesylate, the dose administered in rabbits was not increased beyond 5 mg/kg/h.

The entry of gemifloxacin into the CSF was similar to the CSF penetration of other quinolones in the rabbit model of meningitis in steady state during continuous infusion (10, 13, 16) and with the AUC_CSF/AUC_serum ratio after bolus administration (7). In contrast to the hydrophilic β-lactam antibiotics and glycopeptides (2, 15), the moderately lipophilic quinolones readily enter the CSF (11), and they are poorer substrates than β-lactam antibiotics for pumps involved in the removal of drugs from the CSF (14). Trovafloxacin disposition in the central nervous compartment appears to be completely independent of probenecid-sensitive exit pumps (N. L. Jumble, W. Liu, G. L. Drusano, A. Louie, and M. H. Miller, Abstr. 39th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1924, 1999). Quinolone CSF concentrations are not substantially influenced by the state of the blood-CSF barrier or by the coadministration of dexamethasone (16). The AUC_CSF/AUC_serum ratio of ciprofloxacin (9) and ofloxacin (8) in humans with minor impairment of the blood-CSF barrier was almost identical to the CSF-to-serum ratios of these antibacterials in steady state in the rabbit model of pneumococcal meningitis (10). Conversely, dexamethasone reduced the entry of ceftriaxone and vancomycin into the cerebrospinal fluid during S. pneumoniae meningitis by 30 to 50%, leading to a decreased antibacterial activity in CSF against strains with a reduced sensitivity to these compounds (2, 15).

LTA and TA are potent proinflammatory components of the cell wall of S. pneumoniae (21). The concentrations of free LTA/TA in CSF were slightly lower during treatment with 5 mg of gemifloxacin mesylate per kg/h than during treatment with ceftriaxone. The difference reached statistical significance at 14 h, i.e., immediately after the initiation of antibiotic therapy. Similarly, trovafloxacin and moxifloxacin delayed the in vitro and in vivo release of LTA/TA from S. pneumoniae. Yet, they were less effective than rifamycins and quinupristin/dalfopristin in inhibiting the total amount released after 12 h of therapy (18, 19). In experimental S. pneumoniae meningitis, both gemifloxacin and moxifloxacin (16) did not substantially influence the CSF leukocyte count or the protein and lactate content and did not reduce the NSE in CSF or the density of apoptotic neurons in the dentate gyrus as measures of neuronal damage in meningitis compared to ceftriaxone (4, 22) (Table 1). Therefore, quinolones may not be the ideal compounds to attenuate the inflammatory host response and to reduce neuronal damage in bacterial meningitis. Rifampin, which leads to a stronger inhibition of the release of LTA/TA than quinolones (18), reduced the mortality in a mouse model of S. pneumoniae meningitis in comparison to mice treated with ceftriaxone (12).

In conclusion, at concentrations well tolerated by humans, the bactericidal effect of gemifloxacin was almost as rapid as that of ceftriaxone in the rabbit model of meningitis using a penicillin-sensitive S. pneumoniae strain. The entry into the CSF and the activity of gemifloxacin was not reduced by the coadministration of dexamethasone. Two hours after the initiation of therapy, the CSF concentrations of proinflammatory LTA were slightly lower during gemifloxacin therapy than during ceftriaxone therapy. Yet, a substantial attenuation of the inflammatory response or a reduction of parameters of neuronal damage by gemifloxacin was not observed. Since penicillin resistance is not associated with a decreased susceptibility to quinolones, its in vivo activity suggests that gemifloxacin may be useful in the treatment of meningitis caused by penicillin-sensitive and -resistant strains of S. pneumoniae.