c mice treated with fluoroquinolone derivatives, followed by ultraviolet-A irradiation

Toxicology Letters 157 (2005) 203–210

Structure–phototoxicity relationship in Balb/c mice treated with fluoroquinolone derivatives, followed by ultraviolet-A irradiation Koichi Yabe ∗ , Koichi Goto, Toshimasa Jindo, Masayasu Sekiguchi, Kazuhisa Furuhama Drug Safety Research Laboratory, Daiichi Pharmaceutical Co. Ltd., 16-13, Kita-Kasai 1-Chome, Edogawa-ku, Tokyo 134-8630, Japan Received 9 October 2004; received in revised form 13 February 2005; accepted 13 February 2005 Available online 14 April 2005

Abstract We examined the structure–phototoxicity relationship for fluoroquinolone antimicrobial agents (quinolones) using female albino Balb/c mice. First of all, to obtain an optimum dosage level for induction of phototoxicity, the prototype phototoxicant sparfloxacin was intravenously administered once at 10 mg/kg, 30 mg/kg or 100 mg/kg to female mice, followed immediately by ultraviolet-A (UVA) irradiation for 4 h (21.6 J/cm2 ). The auricular thickness was measured at pre-dose (0 h), 4, 24, 48, 72 and 96 h post-dose, and then the histopathological examination of the auricle was performed. As results, the auricular thickness increased from 30 mg/kg, in conjunction with edema, cellular infiltration, epidermal necrosis and focal loss of the auricle. On the basis of these information, ciprofloxacin, enoxacin, fleroxacin, gatifloxacin, lomefloxacin, norfloxacin and ofloxacin were given intravenously to mice at a fixed dose of 100 mg/kg to compare their potential phototoxicities. Certain quinolones caused the auricular lesions in the following rank order (from lowest to highest): vehicle control (nonphototoxicity) = gatifloxacin = ofloxacin < ciprofloxacin = norfloxacin < enoxacin = fleroxacin < lomefloxacin = sparfloxacin. From the viewpoint of the structure–phototoxicity relationship, quinolones possessing the C-8 substituent with a fluorine or hydrogen and 1,8-naphthyridine derivative evoked phototoxicity in the mouse auricle. These results demonstrate that phototoxicity induced by quinolones would be related to the property of the eighth position. © 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Fluoroquinolone antimicrobials; Phototoxicity; Mouse; Ultraviolet-A

1. Introduction

∗ Corresponding author. Tel.: +81 3 3680 0151; fax: +81 3 5696 8335. E-mail address: [email protected] (K. Yabe).

Fluoroquinolone antimicrobial agents (quinolones) have been widely used in clinical fields because of their broad spectra and bactericidal activity. Some of the quinolones placed on the market, however, have been reported to evoke photosensitivity in humans

0378-4274/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2005.02.006

204

K. Yabe et al. / Toxicology Letters 157 (2005) 203–210

(Ferguson, 2003). Clinical manifestations include erythema in sun-exposed areas and extensively severe bullous eruptions (Lipsky and Baker, 1999). These adverse reactions have been considered phototoxic rather than photoallergic (Ferguson, 2003; Lipsky and Baker, 1999). The incidence of phototoxicity has been recognized to be approximately 10% of patients receiving fleroxacin, lomefloxacin and sparfloxacin (Ferguson, 2003). In contrast, the incidence of phototoxicity of ciprofloxacin, enoxacin, norfloxacin and ofloxacin has been reported to be less than 2.4% (Ferguson, 1995; Ferguson and Dawe, 1997). In experimental studies, Wagai et al. (1989) and Wagai and Tawara (1991b) have developed a method to detect phototoxicity using female albino Balb/b mice. This procedure has been used as a useful tool for predicting the phototoxic potential of quinolone derivatives because the in vivo system incorporates all the physiological aspects (Domagala, 1994). Nevertheless, there is little information dealing with the structure–phototoxicity relationship for quinolones in the in vivo systems under the same experimental conditions. In the present study, ciprofloxacin, enoxacin, fleroxacin, gatifloxacin, lomefloxacin, norfloxacin, ofloxacin and sparfloxacin were given intravenously to Balb/c mice, followed immediately by ultraviolet-A (UVA) irradiation to assess the possible structure–phototoxicity relationship. 2. Materials and methods 2.1. Drugs Eight quinolones (ciprofloxacin, enoxacin, fleroxacin, gatifloxacin, lomefloxacin, norfloxacin, ofloxacin and sparfloxacin) used for the present study were synthesized at Daiichi Pharmaceutical Co. Ltd. (Tokyo, Japan). These drugs were dissolved in 0.1N NaOH solution to obtain a constant administration volume of 10 ml/kg. 2.2. Animals Five-week-old female albino Balb/c mice (17–22 g) were purchased from Japan Charles River (Kanagawa, Japan). They were housed four to six animals per plastic cage in an air-conditioned room (temperature, 23 ± 2 ◦ C; relative humidity, 55 ± 20%; lighting cycle,

12 h) with free access to commercial laboratory chow (F-2, Funabashi Farm, Chiba, Japan) and tap water. After a 3-day acclimation, the animals were used for examinations. They were treated humanely, and the study protocol was in accordance with the institutional guidelines of Daiichi Pharmaceutical Co. Ltd. for use of laboratory animals. 2.3. Optimum dosage level of sparfloxacin for induction of phototoxicity Sparfloxacin as a prototype phototoxicant was intravenously administrated once at 10 mg/kg, 30 mg/kg or 100 mg/kg to groups of six mice each. Additional mice (n = 4) given 0.1N NaOH solution in the same way served as the vehicle control. The number of animals per group used in this experiment was selected on the basis of previous published data in our laboratory (Wagai et al., 1989; Wagai and Tawara, 1991a,b, 1992), and the intravenous administration was chosen as a dosing route capable of providing quinolone exposure at an identical level without bias due to a difference in the gastrointestinal absorption. The reproducibility of this test system has been confirmed in numerous experiments so far (Takayama et al., 1995; Wagai et al., 1989; Wagai and Tawara, 1991a,b, 1992). Immediately after administration, the animals were placed individually in partitioned chambers (4 cm × 8 cm × 4 cm) covered with a 3 mm pane of glass (Floatglass, Asahi Glass, Tokyo, Japan) to eliminate wavelength below 320 mm and irradiated with UVA at 1.5 mW/cm2 for 4 h (21.6 J/cm2 ) as described previously (Wagai et al., 1989). The light source was a bank of 10 black light tubes (FL20SB, diameter: 32.5 mm, length: 58 cm, Toshiba, Tokyo, Japan) emitting radiation within 300–400 nm (a peak: 352 nm). The irradiation condition was nearly equivalent to sunbathing for 4–5 h in the summer in Tokyo (latitude 35◦ 40 ). The intensity of UVA was measured at 365 nm with a UVX digital radiometer fitted with a UVX-36 sensor (UVP Inc., CA, USA). The auricular thickness was measured with a digital thickness micrometer gauge (IDC 543, Mitsutoyo, Tokyo, Japan) at pre-dose (0 h), and 4, 24, 48, 72 and 96 h post-dose. At termination (96 h), all animals were euthanized under the ether anesthesia. The auricular specimens were excised, fixed in 10% buffered formalin, embedded in paraffin wax, sectioned at 4 ␮m thickness,

K. Yabe et al. / Toxicology Letters 157 (2005) 203–210

205

treatment. At 100 mg/kg, however, the thickness could not be measured 72 h later because the auricle was sloughed. At 96 h later, the mean ± S.D. of the auricular thickness was 0.26 ± 0.01 mm for the vehicle control, 0.26 ± 0.02 mm for 10 mg/kg treatment, 0.54 ± 0.02 mm for 30 mg/kg treatment and 0.73 ± 0.05 mm (48 h later) for 100 mg/kg treatment. In histopathological examinations of the auricle (Table 1), edema, cellular infiltration and necrosis of epidermal cells were observed at 30 mg/kg (Fig. 1b), and further focal loss was noted at 100 mg/kg (Fig. 1c). Thickening of epidermis was observed around the remnant of auricular tissues at 100 mg/kg. No remarkable changes were noted in the 10 mg/kg and vehicle control groups (Fig. 1a).

stained with hematoxylin and eosin, and examined histopathologically. 2.4. Structure–phototoxicity relationships for quinolones From the sparfloxacin results, 7 quinolones (ciprofloxacin, enoxacin, fleroxacin, gatifloxacin, lomefloxacin, norfloxacin and ofloxacin) were intravenously administrated once at a fixed dose of 100 mg/kg to groups of four to six mice each. A total of 24 mice receiving 0.1N NaOH solution served as the corresponding control. All procedures were the same as described above. 2.5. Statistical analysis The quantitative auricular thickness is expressed as the mean ± standard deviation (S.D.). Statistical significance in the auricular thickness between the pre-dose (0 h) versus post-dose values was analyzed by paired Student’s t-test (EXSAS, Version 6.10, Arm Corporation, Osaka, Japan). A P-value of less than 0.05 was considered to be statistically significant.

3.2. Comparative phototoxic potential of quinolone derivatives From the sparfloxacin data (Section 3.1), the 96 h values of the auricular thickness was compared among the quinolones. In mice treated with quinolones and UVA irradiation, the mean ± S.D. of the auricular thickness 96 h later was 0.24 ± 0.01 mm for the vehicle control, 0.35 ± 0.05 mm for ciprofloxacin, 0.51 ± 0.04 mm for enoxacin, 0.62 ± 0.02 mm for fleroxacin, 0.24 ± 0.01 mm for gatifloxacin, 0.64 ± 0.02 mm (48 h later) for lomefloxacin, 0.27 ± 0.02 mm for norfloxacin and 0.24 ± 0.01 mm for ofloxacin (Fig. 2). Of the quinolones employed,

3. Results 3.1. Optimum dosage level of sparfloxacin for induction of phototoxicity Sparfloxacin at 30 mg/kg or more showed increases in the auricular thickness from 4 h after

Table 1 Incidence of the auricular lesions in female Balb/c mice receiving a single intravenous administration of quinolones, and followed by UVA irradiation for 4 h Drug

Vehicle Ciprofloxacin Enoxacin Fleroxacin Gatifloxacin Lomefloxacin Norfloxacin Ofloxacin Sparfloxacin a b c

Dose (mg/kg)

0 100 100 100 100 100 100 100 100

n

24 4 4 6 6 4 4 4 6

The animals were euthanized at 96 h post-dose. Number of animals showing changes. Not examined because the auricles were sloughed.

Auriclea Edema

Cellular infiltration

Epidermal necrosis

Focal loss

0 0 4b 4 0 –c 0 0 –c

0 4 4 6 0 –c 0 0 –c

0 0 0 6 0 –c 0 0 –c

0 0 0 0 0 4 0 0 6

206

K. Yabe et al. / Toxicology Letters 157 (2005) 203–210

noted in mice given ciprofloxacin or enoxacin, and further epidermal necrosis was observed in mice receiving fleroxacin. In the gatifloxacin, norfloxacin and ofloxacin groups, no remarkable lesions in the auricle were observed. 3.3. Structure–phototoxicity relationships for quinolones Based on the above result (Section 3.2), the phototoxic rank order was considered as follows (from lowest to highest): vehicle control (non-phototoxicity) = gatifloxacin = ofloxacin < ciprofloxacin = norfloxacin < enoxacin = fleroxacin < lomefloxacin = sparfloxacin. From the viewpoint of the relationship between chemical structure and phototoxic potential, a fluorine substituent at the C-8 position of the quinolone ring (fleroxacin, lomefloxacin or sparfloxacin) was moderate to severe in phototoxicity, and a hydrogen substituent (ciprofloxacin or norfloxacin) was mild in this lesion. Further, 1,8-naphthyridine (enoxacin) was moderate in phototoxicity. In contrast, a methoxy substituent at the C-8 position (gatifloxacin) or benzoxazine (ofloxacin) was non-phototoxic (Table 2). Furthermore, an amino substituent at the C-5 position (sparfloxacin) showed much severer alteration than a hydrogen substituent (gatifloxacin, ofloxacin, ciprofloxacin, norfloxacin, enoxacin or fleroxacin).

4. Discussion Fig. 1. Auricular micrographs of a female Balb/c mouse receiving a single intravenous administration of sparfloxacin, followed by UVA irradiation for 4 h. The animal was euthanized 96 h after treatment: (a) vehicle control, (b) 30 mg/kg. Edema, cellular infiltration and necrosis of epidermal cell were observed, (c) 100 mg/kg. Edema, cellular infiltration and thickening of epidermis (asterisk) were noted around the remnant auricular tissues.

a tendency of increased auricular thickness was noted from 4 h later in fleroxacin, from 24 h later in enoxacin, lomefloxacin and norfloxacin, from 48 h later in ciprofloxacin. Especially, lomefloxacin showed macroscopically focal loss of the auricle at 72 h post-dose. Meanwhile, the auricular thickness of norfloxacin recovered to the pre-dose level 72 h later. In histopathological examinations (Table 1), edema and/or cellular infiltration in the auricle were

Although, quinolone phototoxicity has been carried out to elucidate the mechanisms by using both in vitro and in vivo systems (Takayama et al., 1995), there is little information dealing with the structure–phototoxicity relationship evaluated in the in vivo systems under the same experimental conditions. We first explored the relative rank order of phototoxicity using Balb/c mice treated intravenously with eight quinolones, and then the structure–phototoxicity relationship was assessed. As results, a difference in phototoxic occurrence was associated with a difference in the substituent at the C-8 position of the quinolone ring. In brief, a fluorine substituent (fleroxacin, lomefloxacin or sparfloxacin) was most severely phototoxic, a hydrogen substituent (ciprofloxacin or norfloxacin) was mildly phototoxic, and a methoxy

K. Yabe et al. / Toxicology Letters 157 (2005) 203–210

207

Fig. 2. Auricular thickness in female Balb/c mice receiving a single intravenous administration of quinolones, followed by UVA irradiation for 4 h. Data are represented as the mean ± S.D. (n = 4–6). *,** P < 0.05, 0.01 vs. vehicle control. Letter ‘a’ indicates the auricles were sloughed and could not be measured after this timepoint: () vehicle control, (䊉) 100 mg/kg

208

K. Yabe et al. / Toxicology Letters 157 (2005) 203–210

Table 2 Structure–phototoxicity relationships in female Balb/c mice receiving a single intravenous administration of quinolones, and followed by UVA irradiation for 4 h

Drug

X8

Gatifloxacin

COCH3

R1

R5 H

R7

Phototoxicity −

−

Ofloxacin Ciprofloxacin

CH

Norfloxacin

CH

Enoxacin

H

+

C 2 H5

H

+

N

C 2 H5

H

++

Fleroxacin

CF

CH2 CH2 F

H

++

Lomefloxacin

CF

C 2 H5

H

+++

Sparfloxacin

CF

NH2

+++

Phototoxic potential was assessed by the results of the thickness and histopathological findings of the auricle at 96 h post-dose. The auricular thickness of lomefloxacin and sparfloxacin was estimated at 48 h post-dose because the auricles showed focal loss and could not be measured after this timepoint. (−) none, (+) mild, (++) moderate, (+++) severe.

substituent (gatifloxacin) was non-phototoxic. Further, 1,8-naphthyridine (enoxacin) and benzoxazine (ofloxacin) was moderately phototoxic and nonphototoxic, respectively. These phototoxic rankings correlated with the results of randomized controlled trials of phototoxicity due to quinolones in caucasian volunteers (Dawe et al., 2003; Ferguson, 2003). As a physicochemical property, certain quinolones have been known to be degraded by UV irradiation (Takayama et al., 1995). For example, Q-35, a quinolone derivative having a methoxy substituent at the C-8 position, was stable under UVA irradiation conditions. Meanwhile, Q-35 derivatives, which are substituted with a hydrogen or fluorine at the C-8

position, were degraded when exposed to UVA, and their photoproducts simultaneously provoked an increase in cytotoxicity (Matsumoto et al., 1992). Likewise, Marutani et al. (1993) reported that Q-35 did not induce phototoxic auricular inflammation in mice, while its hydrogen or fluorine derivative induced the inflammatory response. These results are essentially consistent with our observations. However, evidence that intra-auricular administration of photoproducts after UVA irradiation to mice did not induce phototoxic reactions suggests that these products in themselves did not possess a direct toxic action (Wagai and Tawara, 1991a). In other words, this phenomenon demonstrates that the existence of reactive oxygen

K. Yabe et al. / Toxicology Letters 157 (2005) 203–210

species generated from quinolones which absorbed UVA would play a crucial role in the onset of phototoxicity (Wagai and Tawara, 1991b, 1992). In fact, reactive oxygen species such as hydroxyradical, singlet oxygen or superoxide anion were directly detected in quinolone solutions exposed to light irradiation (Araki and Kitaoka, 1998; Umezawa et al., 1997). Of these reactive oxygen species, singlet oxygen has been thought to be an important factor responsible for induction of quinolone phototoxicity (Robertson et al., 1991). Araki and Kitaoka (1998) have reported that the yield of singlet oxygen from quinolone solutions is high in lomefloxacin or sparfloxacin, intermediate in ciprofloxacin or enoxacin, and low in levofloxacin (the s-isomer of ofloxacin). They have also recognized the existence of reaction pathways for generation of reactive oxygen species in lomefloxacin and sparfloxacin. Thus, the first step of the photosensitizing reaction is considered to be homolytic elimination of a fluorine substituent at the C-8 position of the quinolone ring. Afterward, this excited quinolone reacts with molecular oxygen and presumably generates a reactive oxygen species. This hypothesis was supported by the fact that 8-desfluorospafloxacin, a main photoproduct of sparfloxacin, was confirmed in the solution under UV irradiation conditions (Engler et al., 1998). From other viewpoint of structure–phototoxicity relationship, an amino substituent at the C-5 position (sparfloxacin) revealed much severer phototoxicity than a hydrogen substituent (gatifloxacin, ofloxacin, ciprofloxacin, norfloxacin, enoxacin or fleroxacin). However, Yoshida et al. (1996) have reported that an amino substituent at the C-5 position showed the ability to reduce the phototoxic potential in guinea pigs due to quinolones. Meanwhile, Miolo et al. (2002) have reported that a quinoloe derivative having an amino substituent at the C-6 position causes milder phototoxicity (hemolysis) in murine red blood cells than that having a fluorine or hydrogen substituent. Taken together, a substituent at the C-5 or C-6 position of the quinolone ring may be capable of modulating the severity of phototoxicity, but do not essential for its induction. As with our findings, Martinez et al. (1998) and Klecak et al. (1997) have reported that a physicochemical property of a fluorine substituent at the C-8 position of the quinolone ring (fleroxacin or lomefloxacin) contributes to the development of phototoxicity and

209

photocarcinogencity. Indeed, the formation of 8-oxo7,8-dihydro-2 -deoxyguanosine, a marker of oxidative DNA damage, in calf thymus DNA was high in lomefloxacin, an intermediate in ciprofloxacin, and low in moxifloxacin, a quinolone derivative having a methoxy substituent at the C-8 position (Spratt et al., 1999). However, their investigations were performed only by in the in vitro systems, except for the carcinogen study of Klecak et al. (1997). Their relative rank order concerning phototoxic potentials was made by citing the published in vivo data from other investigators, along with the in vitro results. In conclusion, our results in the present experiments demonstrate that phototoxicity due to quinolones would be related to the property of the eighth position. Acknowledgements We thank Dr. Tesuya Araki, Dr. Hiroaki Inagaki, Dr. Katsuhiro Kawakami, Dr. Michiyuki Kato and Dr. Makoto Takemura for their critical review and helpful advice. References Araki, T., Kitaoka, H., 1998. ESR detection of free radical and active oxygen species generated during photolysis of fluoroquinolones. Chem. Pharm. Bull. 46, 1021–1026. Dawe, R.S., Ibbotson, S.H., Sanderson, J.B., Thomson, E.M., Ferguson, J., 2003. A randomized controlled trial (volunteer study) of sitafloxacin, enoxacin, levofloxacin and sparfloxacin phototoxicity. Br. J. Dermatol. 149, 1232–1241. Domagala, J.M., 1994. Structure–activity and structure–side-effect relationships for the quinolone antibacterials. J. Antimicrob. Chemother. 33, 685–706. Engler, M., Rusing, G., Sorgel, F., Holzgrabe, U., 1998. Defluorinated sparfloxacin as a new photoproduct identified by liquid chromatography coupled with UV detection and tandem mass spectrometry. Antimicrob. Agents Chemother. 42, 1151–1159. Ferguson, J., 1995. Fluoroquinolone photosensitization: a review of clinical and laboratory studies. Photochem. Photobiol. 62, 954–958. Ferguson, J., 2003. Phototoxicity due to fluoroquinolones. In: Hooper, D.C., Rubinstein, E. (Eds.), Quinolone Antimicrobial Agents. ASM Press, Washington, DC, pp. 451–460. Ferguson, J., Dawe, R., 1997. Phototoxicity in quinolones: comparison of ciprofloxacin and grepafloxacin. J. Antimicrob. Chemother. 40 (Suppl. A), 93–98. Klecak, G., Urbach, F., Urwyler, H., 1997. Fluoroquinolone antibacterials enhance UVA-induced skin tumors. J. Photochem. Photobiol. B 37, 174–181.

210

K. Yabe et al. / Toxicology Letters 157 (2005) 203–210

Lipsky, B.A., Baker, C.A., 1999. Fluoroquinolone toxicity profiles: a review focusing on newer agents. Clin. Infect. Dis. 28, 352–364. Martinez, L.J., Sik, R.H., Chlgnell, C.F., 1998. Fluoroquinolone antimicrobials: singlet oxygen, superoxide and phototoxicity. Photochem. Photobiol. 67, 399–403. Marutani, K., Matsumoto, M., Otabe, Y., Nagamuta, M., Tanaka, K., Miyoshi, A., Hasegawa, T., Nagano, H., Matsubara, S., Kamide, R., Yokota, T., Matsumoto, F., Ueda, Y., 1993. Reduced phototoxicity of a fluoroquinolone antibacterial agent with a methoxy group at the 8 position in mice irradiated with long-wavelength UV light. Antimicrob. Agents Chemother. 37, 2217–2223. Matsumoto, M., Kojima, K., Nagano, H., Matsubara, S., Yokota, T., 1992. Photostability and biological activity of fluoroquinolones substituted at the 8 position after UV irradiation. Antimicrob. Agents Chemother. 36, 1715–1719. Miolo, G., Viola, G., Vedaldi, D., Dall’Acqua, F., Fravolini, A., Tabarrini, O., Cecchetti, V., 2002. In vitro phototoxicity properties of new 6-desfluoro and 6-fluoro-8-methylquinolones. Toxicol. In Vitro 16, 683–693. Robertson, D.G., Epling, G.A., Kiely, J.S., Bailey, D.L., Song, B., 1991. Mechanistic studies of the phototoxic potential of PD 117596, a quinolone antibacterial compound. Toxicol. Appl. Pharmacol. 111, 221–232. Spratt, T.E., Schultz, S.S., Levy, D.E., Chen, D., Schluter, G., Williams, G.M., 1999. Different mechanisms for the photoinduced production of oxidative DNA damage by fluoroquinolones differing in photostability. Chem. Res. Toxicol. 12, 809–815.

Takayama, S., Hirohashi, M., Kato, M., Shimada, H., 1995. Toxicity of quinolone antimicrobial agents. J. Toxicol. Environ. Health 45, 1–45. Umezawa, N., Arakane, K., Ryu, A., Mashiko, S., Hirobe, M., Nagano, T., 1997. Participation of reactive oxygen species in phototoxicity induced by quinolone antibacterial agents. Arch. Biochem. Biophys. 342, 275–281. Wagai, N., Tawara, K., 1991a. Important role of oxygen metabolites in quinolone antibacterial agent-induced cutaneous phototoxicity in mice. Arch. Toxicol. 65, 495–499. Wagai, N., Tawara, K., 1991b. Quinolone antibacterial-agentinduced cutaneous phototoxicity: ear swelling reactions in Balb/c mice. Toxicol. Lett. 58, 215–223. Wagai, N., Tawara, K., 1992. Possible direct role of reactive oxygens in the cause of cutaneous phototoxicity induced by five quinolones in mice. Arch. Toxicol. 66, 392–397. Wagai, N., Yamaguchi, F., Tawara, K., Onodera, T., 1989. Studies on experimental conditions for detecting phototoxic potentials of drugs in Balb/c mice. J. Toxicol. Sci. 14, 197– 204. Yoshida, T., Yamamoto, Y., Orita, H., Kakiuchi, M., Takahashi, Y., Itakura, M., Kado, N., Mitani, K., Yasuda, S., Kato, H., Itoh, Y., 1996. Studies on quinolone antibacterials. IV. Structure–activity relationships of antibacterial activity and side effects for 5- or 8-sunstituted and 5,8-disubstituted-7-(3-amino-1-pyrrolidinyl)1-cyclopropyl-1,4-dihydro-4-oxoquinoline-3-carboxylic acid. Chem. Pharm. Bull. 44, 1074–1085.