ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, JUlY 1993, P. 1497-15030066-4804/93/071497-07$02.00/0Copyright © 1993, American Society for Microbiology

Quinolones Potentiate Cefazolin-Induced Seizures inDBA/2 Mice

A. DE SARRO,1* M. ZAPPALA,2 A. CHIMIRRI,2 S. GRASSO,2 AND G. B. DE SARRO1

Institute ofPharmacology, School of Medicine, University ofMessina and Regio Calabria, 1 and DepartmentofMedicinal Chemistry, School ofPharmacy, University ofMessina, 98122 Messina, Italy

Received 10 August 1992/Accepted 12 April 1993

The behavioral and convulsant effects of cefazolin, a I-lactam derivative, were studied after intraperitonealadministration to DBA/2 mice, a strain genetically susceptible to sound-induced seizures, and Swiss mice.DBA/2 mice were more susceptible to seizures induced by cefazolin than were Swiss mice. The proconvulsanteffects of some quinolones on seizures evoked by intraperitoneal administration of cefazolin were also evaluatedin DBA/2 mice. Our study also demonstrated that the order of proconvulsant activity in our epileptic model waspefloxacin > enoxacin > ofloxacin > rufloxacin > norfloxacin > cinoxacin > ciprofloxacin > nalidixic acid.The relationships between the chemical structures and proconvulsant activities of quinolone derivatives were

studied. The relationship between lipophilicity and proconvulsant activity was also investigated.

The genetically epilepsy-prone mouse or Dilute BrownAgouti DBA/2J (DBA/2) mouse has been known since 1947to be a strain susceptible to audiogenic seizures (AGSs) (21).The characteristics and neurochemical abnormalities of thismouse strain have been described by various researchers (6,35, 36, 40). Between 16 and 30 days of age in response to aloud tone, this mouse strain shows ill-coordinated locomo-tion consisting of a wild running followed by rhythmic clonicjerking with the animal lying on one side, followed by tonicflexion and tonic extension of trunk, limbs, and tail. Thelatter phase may terminate with respiratory arrest and death.One investigator observed a longer period of susceptibility toAGSs of up to 39 days (45), and he has interpreted thisphenomenon as the consequence of differences in geneticcomposition among the strains. In addition, it has beenreported that DBA/2 mice have increased seizure suscepti-bility to a variety of nonaudiogenic convulsant treatments,including chemical and physical stimuli (6, 7, 19). At 8 weeksof age, when DBA/2 mice are not susceptible to AGSs, theystill showed major hyperexcitability in response to electri-cally induced convulsions, even after development of resis-tance to AGSs (6, 19). Extensive studies were done on thenature of AGSs in DBA/2 mice; a defect in uptake orutilization of glucose at the critical age was observed andproposed to be the primary cause of seizure susceptibility(37). Ca2"-ATPase pathology was observed and appeared tobe another factor which influences AGS susceptibility (31,32). Measurements of neurotransmitter concentrations andactivities of enzymes involved in neurotransmitter synthesisor further metabolism, as well as studies of receptor bindingsites, had showed no differences in the levels of the inhibi-tory and excitatory amino acids in the cerebellum or regionsconcerned with the auditory pathway (auditory cortex, co-

chlear nuclei, and inferior colliculi) (6, 9, 43). As far as

y-aminobutyric acid (GABA) is concerned, it was demon-strated that GABA inhibitory input is decreased in thecentral nervous systems of DBA/2 mice at 3 weeks, but notat 8 weeks, and that this effect is greater on the receptor thanon the ionophore. In addition, GABA- and benzodiazepine-binding sites exhibited regional and age-dependent variation

* Corresponding author.

in DBA/2 mice and a reduction of [3HJGABA-binding sites inwhole brain was described by several investigators (22, 23,42). A decrease in the number of benzodiazepine-bindingsites was reported for whole brains of DBA/2 mice at theseizure-susceptible age by Horton and coworkers (22). Themajor susceptibility to excitatory amino acid agonists, suchas kainate, N-methyl-D-aspartate, and homocysteine thiolac-tone, in this strain of mice in comparison with C57 micesuggests increased endogenous (central nervous system)excitatory responsiveness in DBA/2 mice (19). Thus, thisstrain of mice has been considered an excellent animal modelfor the study of certain kinds of human epilepsy and fortesting of new anticonvulsant drugs (6, 39).

Since cefazolin has been known to induce convulsions invarious animal species (12, 13, 29, 30), it was postulated thatthe patterns of responsiveness to this convulsant agent maybe different in AGS-susceptible DBA/2 mice and non-AGS-susceptible Swiss mice. The convulsant action of penicillinsand cephalosporins has been attributed to inhibition of theGABA system (2, 10, 21). The main intention of this studywas to examine cefazolin-induced seizures in AGS-proneDBA/2 mice and in Swiss mice. In addition, some experi-mental experiences with quinolones (1, 11) have demon-strated possible proconvulsant activity. Thus, the neuro-toxic effects of quinolones and the inhibitory effects ofcefazolin on GABA transmission prompted us to study theeffects of their concomitant administration. In the presentreport, the proconvulsant effects of several quinolones are

also described, compared, and discussed with particularregard to the structure-activity relationship and lipophilicity.

MATERIALS AND METHODS

Testing of anticonvulsant activity. DBA/2 mice (16 to 24 g,42 to 48 days old) and Swiss mice (16 to 24 g, 42 to 48 daysold) were purchased from Charles River (Calco, Como,Italy). Mice of each strain were injected with cefazolin (0.1to 2.0 mg/g of body weight intraperitoneally [i.p.]) dissolvedin sterile saline (0.1 ml/10 g of body weight). Animals were

placed in a Plexiglas box (40 by 40 by 30 cm) and observedfor 240 min after i.p. administration of cefazolin. For eachdose of cefazolin, 10 animals of each strain were used. Theintensity of the seizure response was scored on the following

1497

Vol. 37, No. 7

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 19

Dec

embe

r 20

21 b

y 45

.166

.156

.128

.

ANTIMICROB. AGENTS CHEMOTHER.

scale: 0, no response; 1, wild running; 2, clonus; 3, tonus; 4,respiratory arrest identical to that observed after auditorystimulation (14). A second group of animals (DBA only) wastreated with some quinolones at a dose of 250 nmol/g (whichcorresponds approximately to 100 p,g/g) or the vehicle (ster-ile saline), and 60 min later the same animals were injectedi.p. with various doses of cefazolin.

Pefloxacin mesylate, ciprofloxacin hydrochloride, ruflox-acin hydrochloride, and enoxacin sesquihydrate were dis-solved in sterile distilled water, while the other quinoloneswere dissolved in NaOH (0.1 N) and isotonic sterile saline.

Electrocortical (ECoG) analysis. ECoG activity was re-corded (eight-channel ECoG machine; OTE Biomedica,Florence, Italy) through four chronically implanted steelscrew electrodes inserted bilaterally in the frontoparietalarea. At least three mice treated with the largest dose ofquinolone derivatives plus cefazolin were studied forchanges in ECoG activity.

Statistical analysis. Statistical comparisons among groupsof control and drug-treated animals were done with Fisher'sexact probability test (incidence of the seizure phases) andthe Mann-Whitney U test. The percentage of incidence ofeach phase of the seizure was determined for the differentdoses of cefazolin administered, and dose-response curveswere fitted by the method of Litchfield and Wilcoxon (25).Fifty percent convulsant doses (CD50s; with 95% confidencelimits) for each compound and each seizure response phasewere estimated by the method of Litchfield and Wilcoxon(25) as adapted to an IBM computer system by Tallarida andMurray (41). Relative convulsant activities were also deter-mined by comparison of respective CD50s.

Lipophilicity measurements. The relative lipophilicity (Rm)of the compounds was measured by reversed-phase thin-layer chromatography as previously described (5). Briefly,silanized silica gel plates (Merck 60 F254) were used as thenonpolar stationary phase. The polar mobile phase was a30:70 (vol/vol) mixture of acetone and phosphate buffer atpH 7. Each compound was dissolved in chloroform (3mg/ml), and 5 ,ul of the solution was applied to the plate. Theexperiments were repeated five times with different disposi-tions of the compounds on the plate. Rf values were ex-pressed as means of five determinations. Rm values werecalculated from the experimental Rf values by the formulaRm = log (lIRf) - 1. Higher Rm values indicate higherlipophilicity.

Drugs. The sources of the drugs used were as follows:sodium cefazolin, Farmitalia Carlo Erba Laboratories, Mi-lan, Italy; enoxacin sesquihydrate, Zambeletti Laboratories,Milan, Italy; pefloxacin mesylate, Rhone-Poulenc PharmaS.p.A., Milan, Italy; cinoxacin, Eli Lilly & Co., Indianapo-lis, Ind.; ciprofloxacin hydrochloride, Bayer, Leverkusen,Germany; norfloxacin, I.S.F. Laboratories, Trezzano S/NMilan, Italy; ofloxacin, Sigma Tau Laboratories, Pomezia,Italy; rufloxacin hydrochloride, Mediolanum Farmaceutici,Milan, Italy; nalidixic acid, Sigma, St. Louis, Mo. Thestructures of the drugs used are shown in Fig. 1.

RESULTS

Comparative convulsant activities of cefazolin in Swiss andAGS-prone mice. As shown in Table 1, the intensities ofseizures were significantly different in the two strains of micestudied after i.p. administration of cefazolin. In particular,DBA/2 mice were more susceptible to convulsant doses ofcefazolin than were Swiss mice. Higher doses of cefazolin

C2C30

0

Cisoxaci.

HN- A

F 0 0 H

0

Emoxacis

IF N

0

Norfioxacim

Pefloxacia

HN/\7

H/N -")F COOHa

0

Ciprofloxacis

H3C

com0

Nalidixic Aad

H,CN ,-Y m

N 0~~~

F COOO"

Ofoxa

F: xcv0N

Rufloxacls

FIG. 1. Chemical structures of the quinolone derivatives stud-ied.

were necessary in Swiss mice for a similar incidence of theclonic and tonic seizure phases (Table 1).Combined treatment of DBA/2 mice with cefazolin and

quinolones. A second group of animals (DBA only) wastreated with some quinolones at a dose of 250 nmol/g (whichcorresponds to approximately 100 p,g/g) or the vehicle (ster-ile saline), and 60 min later the same animals were injectedi.p. with various doses of cefazolin. At 250 nmollg, eachquinolone was unable to elicit seizures in DBA/2 mice andelicited a mild decrease in spontaneous activity and explor-ative behavior. The effects of combined treatment of cefazo-lin with nalidixic acid, cinoxacin, ciprofloxacin, enoxacin,norfloxacin, ofloxacin, pefloxacin, or rufloxacin are shown inTable 2. Significant differences in the proconvulsant effectsof the quinolones tested were observed. The relative CD50s(with 95% confidence limits) of cefazolin alone and differentquinolones plus cefazolin are reported in Table 3.The concentration of cefazolin which produced a high

incidence of seizures was lower after administration ofequimolar doses of pefloxacin than with the other quinolonestested. The percentage of DBAI2 mice showing seizureswhen pretreated with 250 nmol of pefloxacin per g 1 h beforethey received cefazolin was, in some cases, significantlyincreased (P < 0.05) in comparison with the group of micethat received cefazolin alone. Similar and less evident effectswere observed in DBA/2 mice pretreated with an equimolardose (250 nmol/g) of enoxacin, norfloxacin, ofloxacin, ru-

1498 DE SARRO ET AL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 19

Dec

embe

r 20

21 b

y 45

.166

.156

.128

.

PROCONVULSANT ACTIVITY OF QUINOLONES 1499

TABLE 1. Incidence of seizures induced by cefazolin in DBA/2and Swiss micea

Strain and % Responsetreatment

(dose, 1Lg/g) WR Clonus Tonus RA SR

DBA/2Vehicle 0 0 0 0 00.25 30 0 0 0 0.30.5 50* 0 0 0 0.50.75 70 40 20 0 1.31.0 80 50 30 10 1.71.25 100* 70* 40 20 2.51.5 100 90* 60 50* 3.01.75 100 100 80 70 3.5

SwissVehicle 0 0 0 0 00.5 0 0 0 0 00.75 20 0 0 0 0.21.0 40 10 10 0 0.61.25 50 20 20 0 0.91.5 70 40 30 20 1.61.75 90 70 60 40 2.62.0 100 90 80 60 3.3a Groups of 10 DBA/2 or Swiss mice were injected i.p. with the stated doses

of cefazolin and observed for 240 min after drug injection. The incidence ofeach seizure phase is expressed as the percentage of mice in each groupdisplaying that phase. Significant differences (P < 0.05) in the incidence ofseizure phases between DBA/2 and Swiss mice are marked with asterisks.WR, wild running; RA, respiratory arrest; SR, arithmetic mean of themaximum individual responses for each animal in the group.

floxacin, or cinoxacin. Ciprofloxacin and nalidixic acid werestill less able to increase the convulsant properties of cefazo-lin. In particular, ciprofloxacin and nalidixic acid failed tomodify the occurrence of cefazolin-induced seizures signifi-cantly compared with the cefazolin group by both Fisher'sexact probability test and the Mann-Whitney U test.ECoG activity. In animals used to study ECoG activity,

electrocorticographic epileptic discharges appeared morerapidly in mice pretreated with quinolones than in those thatreceived cefazolin alone. In particular, these epileptic dis-charges appeared more rapidly in animals pretreated withpefloxacin than in those pretreated with ciprofloxacin ornalidixic acid. The animals pretreated with the other quino-lones showed an occurrence of electrocorticographic epilep-tic discharges with a latency intermediate between those ofpefloxacin and ciprofloxacin (data not shown). The electro-corticographic epileptic pattern was similar in DBA/2 micethat received cefazolin alone or a quinolone derivative pluscefazolin (Fig. 2 and 3).

Physicochemical parameters. The Rm and pK. values andmolecular weights of the quinolone derivatives tested aresummarized in Table 4.

DISCUSSION

The present study demonstrated marked differences be-tween Swiss and DBA/2 mice in response to seizures in-duced by cefazolin. In particular, seizure latency was con-sistently shorter in DBA/2 mice than in Swiss mice. In fact,a difference in seizure incidence was observed and Swissmice showed an increase in clonic seizure latency or thelatter phase of seizures failed to occur. These data suggestthat the neural mechanisms which delay or prevent the onsetof clonic seizures are not present in DBA/2 mice; this is inagreement with previous studies which showed that DBA/2

mice have increased seizure susceptibility to a variety ofnonaudio convulsant treatments (6, 7, 19).

Since previous studies have indicated that systemic ad-ministration of cefazolin to experimental animals producesepileptogenic effects (12, 13, 24, 29, 30) and that quinolonesmay be responsible for neurotoxic effects in humans andexperimental animals (1, 3, 4, 11), the present results furtherconfirm these effects. In fact, the combination of quinolonesand cefazolin induced more marked convulsant effects thancefazolin alone in experimental animals. The convulsantaction of 1-lactam derivatives has been related to reductionof GABA release from nerve terminals or inhibition of thebinding of GABA to its receptor sites (2, 10, 21). In aprevious study, we suggested that cefazolin is a tetrazolderivative which shows marked similarity to pentylenetet-razol, a well-known convulsant drug (12).Although the mechanism by which quinolones exhibit a

proconvulsant effect is still obscure, the compounds testedshowed different degrees of potency. Thus, we attempted toclarify the relationship between the structures and the epi-leptogenic activities of quinolones. Chemically, the newquinolones tested in this study are generally characterizedby having a fluorine atom at the C-6 position and a piperazinemoiety at position 7 of the quinoline or naphthyridine ring.Pefloxacin and enoxacin seem to be the most potent convul-sant compounds of our series (Tables 2 and 3). However,pefloxacin has a quinoline structure while enoxacin has anaphthyridine structure. Both pefloxacin and enoxacin pos-sess a piperazine moiety at position 7, like ofloxacin, cipro-floxacin, rufloxacin, and norfloxacin. We also examinedcinoxacin, which seems to be less potent than pefloxacin andenoxacin but more potent than ciprofloxacin. Cinoxacin hasno fluorine atom at position 6 and has a nitrogen atom inposition 2 of the quinoline ring. Nalidixic acid, a 1,8-naphthyridine derivative which does not possess a fluorineatom in its ring, showed a minimum convulsant effect. Theseresults suggest that either the presence of the fluorine atomor the presence of the quinoline and the naphthyridine ringseems to be significantly responsible for the proconvulsanteffects of quinolones. Since differences among the differentcompounds studied exist, we suspect that both the presenceof the fluorine atom and the presence of the quinoline andnaphthyridine structures might affect the capacity of quino-lones to penetrate the brain. Differences among the abilitiesof quinolones to enter the brain through the blood-brainbarrier exist, and recent data suggest that the concentrationsof ciprofloxacin in cerebral fluid are approximately 2 to 5%(up to 10%) of the concomitant concentrations in serum inpatients with noninflamed meninges and 2.5 times higherlevels have been detected in the presence of inflammation(8). Levels of ofloxacin and pefloxacin in cerebrospinal fluidwere higher than those of ciprofloxacin (28). In addition,concentrations of ofloxacin in cerebrospinal fluid were 50 to60% of those in serum in patients with bacterial meningitis,and pefloxacin also penetrated the brain well in patients withinflamed meninges (8). These differences among the abilitiesof various quinolones to penetrate the blood-brain barriermay, in part, explain the major neurotoxicity of somecompounds. Thus, the various degrees of proconvulsantactivity exhibited by these derivatives, which do not ionizeappreciably at physiological pH, may be partially related totheir lipophilicity (Table 4). Quinolones generally possessvery low lipophilicity, but it is possible that some quino-lones, such as pefloxacin, which has a long half-life, crossthe blood-brain barrier better than others. Pharmacokineticstudies using a method more sensitive than ours for calcu-

VOL. 37, 1993

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 19

Dec

embe

r 20

21 b

y 45

.166

.156

.128

.

ANTIMICROB. AGENTS CHEMOTHER.

TABLE 2. Effects of some quinolone derivatives on seizuresinduced by cefazolin in DBA/2 micea

Drug and treatment or % Responsecefazolin dose (,ug/g) WR Clonus Tonus RA SR

00102050100**100**

0

0

10204080*

100**100

0 00 0.30 1.1++

20 1.8++30 2.6++80** 3.8++

100** 4.0+

0 0

0 0.10 0.6+10 1.4++20 2.1+60** 3.3++80** 3.8+

100* 4.0

0 0 0

0 0 0

0 0 0.310 0 1.030 0 1.650 20 2.460 40 2.980 60 3.3

0 0 0

0 0 0

0 0 0.20 0 0.510 0 1.040 10 1.850 30 2.570 50 3.1

0

0

0

20

30

80*

90*

100

0

0

0

30

40

80*

90*

100

0

0

20

30

50

70*

100**

0 0

0 0.30 0.6+0 1.3++

20 1.960* 3.2++70* 3.680 3.8

0 0

0 0.30 0.8++10 1.6++30 2.3+60* 3.3++70* 3.680 3.8

0 0

0 0.30 1.0+ +

20 1.8++30 2.4+60** 3.2+80** 3.8+

Continued

TABLE 2-Continued

Drug and treatment or % Responsecefazolin dose (,ug/g) WR Clonus Tonus RA SR

NorfloxacinVehicle 0 0 0 0 00.1 0 0 0 0 00.25 30 20 0 0 0.50.5 50 30 10 0 0.90.75 80 60 30 0 1.71.0 100 80 50 30 2.61.25 100 90 70 50 3.11.5 100 100 90 60 3.5

a Groups of 10 DBA/2 mice were injected i.p. with some quinolones and 1h later with the stated doses of cefazolin and observed for 240 min aftercefazolin injection. The incidence of each seizure phase is expressed as thepercentage of mice in each group displaying that phase. Significant differencesin the incidence of seizure phases between concurrent control (cefazolinalone; see Table 1) and drug-treated (quinolones plus cefazolin) groups aremarked with the symbols * (P < 0.05) and ** (P < 0.01). WR, wild running;RA, respiratory arrest; SR, arithmetic mean of the maximum individualresponses for each animal in the group. The median seizure score ± thestandard error of the mean is reported for each dose level studied. Significantdifferences in the incidence of seizure phases between concurrent cefazolinalone (see Table 1) and quinolones plus cefazolin groups are marked with thesymbols + (P < 0.05) and ++ (P < 0.01) (Mann-Whitney U test).

lation of in vivo concentrations of drugs demonstrated thatpefloxacin was the most lipophilic quinolone (8). Moreover,we can suppose that at the concentrations reached by thedrug in the brain it may leave the cerebral area very slowly.A possible interaction of these compounds with GABAreceptor-binding sites has been reported (17, 38, 44).However, the fact that compounds with different poten-

cies as inhibitors of specific GABA binding, such as cipro-floxacin, rufloxacin, norfloxacin, and enoxacin, and verysimilar Rm values showed different degrees of proconvulsantactivity suggests the importance of other parameters. Tsujiand coworkers (44) have previously considered the rela-tionship between GABA receptor affinity (Ki values) andlipophilicity (apparent partition coefficient) for various quino-lones. They suggested that the increasing lipophilic charac-ter may be responsible for possible electrostatic interactionsbetween GABA receptors and quinolones. This may beresponsible for nonspecific adsorption on the lipoidal mem-branes, which increases with the increasing lipophilic char-

TABLE 3. Convulsant doses of cefazolin alone and variousquinolone derivatives plus cefazolin in DBA/2 mice

after i.p. administrationa

Compound(s) CD50s (,g/g) and 95% confidence limits

Clonic phase Tonic phase

Cefazolin 0.91 (0.72-1.15) 1.27 (1.03-1.55)Cinoxacin + cefazolin 0.51 (0.36-0.72)* 0.73 (0.46-1.17)*Ciprofloxacin + cefazolin 0.65 (0.56-0.75) 1.03 (0.86-1.23)Enoxacin + cefazolin 0.39 (0.24-0.64)* 0.69 (0.44-1.09)*Nalidixic acid + cefazolin 0.93 (0.74-1.17) 1.21 (0.98-1.47)Norfloxacin + cefazolin 0.56 (0.41-0.77)* 0.93 (0.74-1.17)Ofloxacin + cefazolin 0.39 (0.24-0.63)* 0.70 (0.55-0.89)*Pefloxacin + cefazolin 0.35 (0.21-0.57)* 0.86 (0.49-1.51)*Rufloxacin + cefazolin 0.57 (0.34-0.96)* 0.77 (0.63-0.93)*

a All data were calculated by the method of Litchfield and Wilcoxon (25).Significant differences between the CD50s of the group treated with cefazolinalone and groups treated with quinolones plus cefazolin are marked withasterisks (P < 0.01).

PefloxacinVehicle0.10.250.50.751.01.25

CinoxacinVehicle0.10.250.50.751.01.251.5

CiprofloxacinVehicle0.10.250.50.751.01.251.5

Nalidixic acidVehicle0.10.250.50.751.01.251.5

RufloxacinVehicle0.10.250.50.751.01.251.5

OfloxacinVehicle0.10.250.50.751.01.251.5

EnoxacinVehicle0.10.250.50.751.01.25

0 020 1060* 4080 60**100 80*100 100*100 100

0 010 030 2070 4090 60

100 90100 100100 100

0 00 0

30 060 3070 6090 80

100 90100 90

0 00 0

20 030 1060 3080 50100 70100 90

0 020 1040 2070 4090 50100 80100 100100 100

0 020 1050 3070 50*90 70100 90100 100100 100

0 020 1050 3080* 50*90 70100 90100 100

1500 DE SARRO ET AL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 19

Dec

embe

r 20

21 b

y 45

.166

.156

.128

.

PROCONVULSANT ACITIVITY OF QUINOLONES 1501

A Control

*cx

lCx ___J

Al Control Control

rCX

B 16 minaftu CEPHAZ.

rCX

rcxlCX O. %wA. %VIv

rcX

icx

C 251Min

D 66min

IN -VIrofmr-ll"lqrcx r r IrpiPiVIII

Icx B.. 1.-- 6. ' lw

Bi 16min fter CEPHAZ.

I Iji I i A LAli sof ai A

C1 25min

rDrew1 -- --

7i,_r1"r wpl

DI 66min

I 100 p VJ1pM,*-*lMwA41eI.1 I 100i V

i-iFIG. 2. ECoG patterns after cefazolin (CEPHAZ.) administra-

tion (1 mg/g, i.p.) to Swiss (A, B, C, and D) and DBA/2 (A1, B1, C1,and D1) mice. rCX, right cerebral cortex; lCX, left cerebral cortex.

icx %w*

A 12min after Peflox.

rCX

ICX _

B 12min after Cephaz.

rCX 44Plc A

C 6 0 min

acter of the drug. However, the concentration needed forinteraction of a quinolone and GABA is rather high andvaried among the different quinolones tested by a factor of-100. Thus, it appears questionable that a specific interac-tion of quinolones with the GABA receptor can explain theproconvulsant activity of these compounds (16).The proconvulsant activity of quinolone derivatives on

pentylenetetrazol-induced seizures in mice was recentlystudied (18), and the results obtained were, in part, similar tothose obtained with DBA/2 mice concomitantly adminis-tered quinolones and cefazolin.

Since some quinolones have been shown to be responsiblefor important pharmacokinetic interactions, for example,with xanthines, this possible mechanism also deserves con-sideration in this case (26, 33, 46). The differences in proteinbinding and penetration into the brain may also explain thetoxicity differences observed in the present study, which aresimilar to those reported by two recent studies of Dimpfeland coworkers (15, 16). Indeed, in both studies ciprofloxacinwas the quinolone with the least ability to induce neurotoxicchanges.

Nalidixic acid is well known as a drug which possesses alot of central nervous system side effects, and the recentstudies of Dimpfel and coworkers (15, 16) have demon-strated that this compound is highly active in hippocampalslices, while it appeared to be completely unable to enhanceseizures induced by cefazolin in our study. This phenome-non seems to be due to the high level of protein binding ofnalidixic acid and to its poor penetration of the brain (datanot shown). These differences between the study of Dimpfeland coworkers (15, 16) and ours may be due to the fact thatthe first two studies evaluated the effects of quinolones invitro by direct application of these compounds to hippocam-pal slices, while we used a whole-animal model in which wemay observe some pharmacokinetic effects of drugs in amore biologically relevant situation which usually appears inclinical experience. We suppose that quinolones may act byenhancing the central nervous system side effects induced bycefazolin. In addition, the presence of a common route ofexcretion for quinolones and cefazolin (glomerular filtrationand tubular secretion) might be considered (28). Quinolones

rCX

lcx

,1 J.LaJi.k

nyrIlII lrr pliN r" illwinw MllrRlljorlR.lily

D 130 min

A.- . ..-- ~.1 II 1.1rCX f1%1 .J IiO0 V

lcx IP1Al olol)Wl0I ov

FIG. 3. ECoG patterns after administration of pefloxacin (Pe-flox.; 100 Lg/g) and cefazolin (Cephaz.; 0.5 mg/g) i.p. to DBA/2mice. rCX, right cerebral cortex; ICX, left cerebral cortex.

and cefazolin are both organic acids and therefore maycompete for elimination if they are given concomitantly.Moreover, this interaction would be potentially responsiblefor increased concentrations in serum and prolonged effectsof one or both drugs.

Further experiments are necessary to clarify whether

TABLE 4. Molecular weights and pKa and Rm values ofquinolone derivativesa

Compound Mol wt pK. Rm

Cinoxacin 262.2 5.9 0.05Ciprofloxacin hydrochloride 385.8 6.0-8.8 0.32Enoxacin sesquihydrate 347.3 6.0-8.5 0.28Nalidixic acid 232.2 6.7 0.59Norfloxacin 319.3 6.4-8.7 0.32Ofloxacin 361.4 5.7-7.9 0.16Pefloxacin mesylate 465.5 6.3-7.6 0.31Rufloxacin hydrochloride 399.3 6.0-8.7 0.27

a Rm values were calculated as previously described (5). The pK. valuesshown were previously reported (8, 27, 34).

litIds,.I drarfr4rwm

VOL. 37, 1993

--V

.L,i,A,i A . Lb.. AdI-.JM.A.1

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 19

Dec

embe

r 20

21 b

y 45

.166

.156

.128

.

ANTIMICROB. AGENTS CHEMOTHER.

concomitant administration of quinolones and cefazolin re-sults in altered elimination of either compound or both.

ACKNOWLEDGMENTS

Financial support from the Italian Ministry of University andScientific and Technological Research (MURST) and the ItalianCouncil for Research (CNR, Rome) is gratefully acknowledged.Our thanks to Antonino Giacopello for skillful technical assis-

tance.

REFERENCES1. Akahane, K., M. Sekiguchi, T. Une, and Y. Osada. 1989.

Structure-epileptogenic relationship of quinolones with specialreference to their interaction with y-aminobutyric acid receptorsites. Antimicrob. Agents Chemother. 33:1704-1708.

2. Antoniadis, A., W. E. Muller, and U. Wollert. 1980. Inhibition ofGABA and benzodiazepine receptor binding by penicillins.Neurosci. Lett. 18:309-312.

3. Arcieri, G., E. Griffith, G. Gruenwaldt, A. Heyd, B. O'Brien, N.Becker, and R. August. 1987. Ciprofloxacin: an update onclinical experience. Am. J. Med. 82:381-386.

4. Ball, P. 1986. Ciprofloxacin: an overview of adverse experi-ences. J. Antimicrob. Chemother. 18:187-193.

5. Boyce, C. B. C., and B. V. Milborrow. 1965. A simple assess-ment of partition data for correlating structure and biologicalactivity using thin-layer chromatography. Nature (London) 208:537-539.

6. Chapman, A. G., M. J. Croucher, and B. S. Meldrum. 1984.Evaluation of anticonvulsant drugs in DBA/2 mice. Arzneim.Forsch. 34:1261-1264.

7. Chapman, A. G., G. B. De Sarro, M. Premachandra, and B. S.Meldrum. 1987. Bidirectional effect of 1-carbolines_Jn reflexepilepsy. Brain Res. Bull. 19:337-346.

8. Dalhoff, A. 1989. A review of quinolone tissue pharmacokinet-ics, p. 277-312. In P. B. Fernandes (ed.), Quinolones. J. R.Prous Science Publishers S.A., Barcelona, Spain.

9. Davies, W. E., and C. D. Owen. 1984. Amino acid levels in miceresistant and susceptible to audiogenic seizures (AGS). Br. J.Pharmacol. 82:254-262.

10. De Boer, T., J. C. Stoof, and H. Van Duyn. 1980. Effect ofpenicillin on transmitter release from rat cortical tissue. BrainRes. 192:296-300.

11. De Sarro, A., and G. B. De Sarro. 1991. Responsiveness ofgenetically epilepsy-prone rats to aminophylline-induced sei-zures and interactions with quinolones. Neuropharmacology30:169-176.

12. De Sarro, A., G. B. De Sarro, C. Ascioti, and G. Nistico. 1989.Epileptogenic activity of some P-lactam derivatives: structure-activity relationship. Neuropharmacology 28:359-365.

13. De Sarro, G. B., M. Cal6, G. Bagetta, R. Anfosso, E. Marmo, G.Nistic6, and V. Guarino. 1983. Comparative epileptogenic prop-erties of cefazolin and benzylpenicillin after intracaudate micro-injection in rats. Acta Pharm. Sin. 4:236-238.

14. De Sarro, G. B., V. Libri, C. Ascioti, R. Testa, and G. Nistic6.1987. Anticonvulsant effect of denzimol in DBA/2 mice. Neu-ropharmacology 26:1425-1429.

15. Dimpfel, W., A. Dalhoff, W. Hofmann, and G. Schluter. Synap-tic transmission in the hippocampus slice in the presence ofaminophylline alone and in combination with quinolones. Sub-mitted for publication.

16. Dimpfel, W., M. Spuler, A. Dalhoff, U. Hofmann, and G.Schluter. 1991. Hippocampal activity in the presence of quino-lones and fenbufen in vitro. Antimicrob. Agents Chemother.35:1142-1146.

17. Dodd, P. R., L. P. Davies, E. J. Watson, B. Nielsen, J. A. Dyer,L. S. Wong, and G. A. R. Johnston. 1989. Neurochemicalstudies on quinolone antibiotics: effects on glutamate, GABAand adenosine systems in mammalian CNS. Pharmacol. Toxi-col. 64:404-411.

18. Enginar, N., and L. Eroglu. 1991. The effect of ofloxacin andciprofloxacin on pentylenetetrazol-induced convulsions in mice.Pharmacol. Biochem. Behav. 39:587-589.

19. Engstrom, F. L., and D. M. Woodbury. 1988. Seizure suscepti-bility in DBA and C57 mice: the effects of various convulsants.Epilepsia 29(4):389-395.

20. Hall, C. S. 1947. Genetic differences in fatal audiogenic seizuresbetween two inbred strains of house mice. J. Hered. 38:2-6.

21. Hon, S., S. Kurioka, M. Matsuda, and J. Shimada. 1985.Inhibitory effect of cephalosporins on -y-aminobutyric acid re-ceptor binding in rat synaptic membranes. Antimicrob. AgentsChemother. 27:650-651.

22. Horton, R. W., S. A. Prestwich, and S. P. Jazrawi. 1984.Neurotransmitter receptor binding in genetically seizure-sus-ceptible mice, p. 191-200. In R. Fariello and P. Morselli (ed.),Neurotransmitter, seizures and epilepsy II. Raven Press, NewYork.

23. Horton, R. W., S. A. Prestwich, and B. S. Meldrum. 1982.-y-Aminobutyric acid and benzodiazepine binding sites in audio-genic seizure-susceptible mice. J. Neurochem. 39:864-870.

24. Kame, C., A. Sunami, and K. Tasaka. 1983. Epileptogenicactivity of cephalosporins in rats and their structure-activityrelationship. Epilepsia 24:431-439.

25. Litchfield, J. T., and F. W. Wilcoxon. 1949. A simplified methodof evaluating dose-effect experiments. J. Pharmacol. Exp. Ther.96:99-111.

26. Maesen, F. P. V., J. P. Teengs, C. Baur, and B. I. Davies. 1984.Quinolones and raised plasma concentrations of theophylline.Lancet ii:530.

27. Mc Evoy, G. K. 1985. American hospital formulary service: druginformation 85. American Society of Hospital Pharmacists,Bethesda, Md.

28. Neuman, M. 1990. Vademecum des antibiotiques et agentschimiotherapiques antiinfectieux (cinquieme edition). Maloine,Paris.

29. Nistic6, G., G. B. De Sarro, F. Naccari, R. Musolino, D. Rotiroti,G. Gallitto, and R. Di Perri. 1980. Cefazolin: a valid model ofexperimental epilepsy? Monogr. Neural Sci. 5:14-19.

30. Nistic6, G., G. B. De Sarro, D. Rotiroti, R. Silvestri, and E.Marmo. 1980. Antagonism by classical antiepileptic and sodiumvalproate of cefazolin induced experimental epilepsy in rats.Res. Commun. Chem. Pathol. Pharmacol. 29:429-444.

31. Palayoor, S. T., and T. N. Seyfried. 1983. Genetic study ofcationic ATPase activities and audiogenic seizure susceptibilityin recombinant inbred and congenic strains of mice. J. Neuro-chem. 42:526-533.

32. Palayoor, S. T., and T. N. Seyfried. 1984. Genetic associationbetween Ca ATPase activity and audiogenic seizures in mice. J.Neurochem. 29:1771-1774.

33. Raoof, S., C. Wollschlager, and A. K. Faroque. 1987. Ciproflox-acin increases serum levels of theophylline. Am. J. Med.82:115-118.

34. Ravizzola, G., G. Pinsi, F. Pirali, D. Colombrita, L. Peroni, andA. Turano. 1991. Pharmacokinetic properties and in vivo effi-cacy of rufloxacin (MF 934). Eur. J. Clin. Microbiol. Infect. Dis.10:482-483.

35. Schlesinger, K., W. Boggan, and D. X. Freedman. 1965. Genet-ics of audiogenic seizures. I. Relation to brain serotonin andnorepinephrine in mice. Life Sci. 4:2345-2351.

36. Schreiber, R. A. 1981. Developmental changes in brain glucose,glycogen, phosphocreatine and ATP levels in DBA/2J andC57BL/6 mice and audiogenic seizures. J. Neurochem. 37:655-661.

37. Schreiber, R. A., and A. L. Ungar. 1984. Glucose protectsDBA/2J mice from audiogenic seizures: correlation with brainglycogen levels. Psychopharmacology 84:128-131.

38. Segev, S., M. Rehavi, and E. Rubinstein. 1988. Quinolones,theophylline, and diclofenac interactions with the -y-aminobu-tyric acid receptor. Antimicrob. Agents Chemother. 32:1624-1626.

39. Seyfried, T. N., and G. H. Glaser. 1985. A review of mousemutants as genetic models of epilepsy. Epilepsia 26:143-150.

40. Suter, C., W. 0. Klingman, 0. W. Lacy, D. Boggs, R. Marks,and C. B. Complinger. 1958. Sound-induced seizures in animals.Neurology 8(Suppl. 1):117-120.

41. Tallarida, R. J., and R. B. Murray. 1986. Manual of pharmaco-

1502 DE SARRO ET AL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 19

Dec

embe

r 20

21 b

y 45

.166

.156

.128

.

PROCONVULSANT ACTIVITY OF QUINOLONES 1503

logical calculations with computer programs. II ed. Springer-Verlag KG, Berlin.

42. Ticku, M. K. 1979. Differences in -y aminobutyric acid receptorsensitivity in inbred strains of mice. J. Neurochem. 33:1135-1138.

43. Toth, E., A. Lajtha, S. Sarhan, and N. Seiler. 1983. Anticonvul-sant effects of some inhibitory neurotransmitter amino acids.Neurochem. Res. 8:291-302.

44. Tsu,ji, A., H. Sato, Y. Kume, I. Tamai, E. Okezaki, 0. Nagata,

and H. Kato. 1988. Inhibitory effects of quinolone antibacterialagents on -y-aminobutyric acid binding to receptor sites in ratbrain membranes. Antimicrob. Agents Chemother. 32:190-194.

45. Vicari, E. M. 1951. Fatal convulsive seizure in the DBA mousestrain. J. Psychol. 32:79-97.

46. Wjnands, W. J. A., C. L. A. Van Herwaarden, and T. B. Vree.1984. Enoxacin raises plasma theophylline concentrations. Lan-cet ii:108-109.

VOL. 37, 1993

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 19

Dec

embe

r 20

21 b

y 45

.166

.156

.128

.