The mechanisms of phototoxicity induced in mice by five quinolone antibacterial agents were investigated using mouse 3T3 fibroblast cells and Balb/c mice. In the in vitro study, the cultured cells were exposed to ultraviolet-A (UVA) in the presence of the five quinolones lomefloxacin, enoxacin, ciprofloxacin, ofloxacin and DR-3355 (the s-isomer of ofloxacin). Cytotoxicity after irradiation was assayed by the neutral red and MTT assay methods, both of which revealed dose-dependent phototoxicity for all five quinolones. Phototoxicity was inhibited by the addition of catalase, and was augmented by the addition of superoxide dismutase. Dimethylthiourea (a hydroxyl radical scavenger) protected against phototoxicity induced by four quinolones, but not against that by enoxacin. These results indicated that superoxide anions, hydrogen peroxide and hydroxyl radicals were generated in solutions of these quinolones under UVA irradiation. In the in vivo study, mice were injected in the auricle with hydrogen peroxide. Ear swelling reactions appeared dose dependently. When irradiated, these reactions were significantly augmented. These data suggested that cutaneous phototoxicity in Balb/c mice is initiated by the generation of reactive oxygens in the target tissue, especially of hydroxyl radicals.

The phototoxic effects of quinolone antimicrobial agents on mouse auricular skin and retina were examined histologically. Sparfloxacin at 50 or 100 mg/kg, which alone causes no histologic change, was orally administered to albino Balb/c mice, which were irradiated with ultraviolet A for 4 hr immediately after administration. In the auricle, degeneration of basal epidermal cells was sporadically observed at 2 hr (during the irradiation). Foci of slight edema with degenerated fibroblasts were seen in the dermis at 4 hr. Edema and neutrophil infiltration in the dermis became severe up to 96 hr. Initial changes in the retina were observed at 2 hr. Vacuolation of the photoreceptor segments (particularly the inner segment) was occasionally associated with swelling of retinal pigment epithelial cells. The segments became disorganized with time, and the outer nuclear layer showed reduced cellularity. The segments and layer were partially thinned and lost 96 hr later. Enoxacin at 400 and 800 mg/kg induced similar lesions to those of sparfloxacin. Levofloxacin caused similar lesions in the auricle but no change in the retina. The combination of oral administration of quinolone and ultraviolet A irradiation, which never caused apparent morphological changes alone, was shown to be able to induce phototoxic lesions in albino mice. Therefore, this method is thought to be useful to examine morphological changes caused by quinolone phototoxicity.

We investigated whether or not the generation of reactive oxygens and toxic photoproducts participated in the cutaneous phototoxicity mechanisms induced by the quinolone derivatives, ofloxacin (OFLX), enoxacin, lomefloxacin, ciprofloxacin and DR-3355 (the s-isomer of OFLX) in a mouse model. Pretreatment of Balb/c mice with allopurinol, soybean trypsin inhibitor, catalase and beta-carotene gave significant protection against ear swelling reactions induced by oral administration of quinolones and following ultraviolet-A (UVA) irradiation. Pretreatment with diethyldithiocarbamate augmented the swelling. No swelling was observed with direct injection into the auricle of UVA-pretreated photoproducts of the quinolones. These results showed that cutaneous phototoxicity did not depend on the generation of toxic photoproducts and suggested that oxygen metabolites generated in the xanthine oxidase pathway participated in the toxicity.

The reason for the differences in phototoxic potential between the 5 quinolone antibacterial agents lomefloxacin, enoxacin, ciprofloxacin, ofloxacin and DR-3355 (the s-isomer of ofloxacin) in mice was investigated. Superoxide anion, hydrogen peroxide (H2O2), and bleaching of p-nitrosodimethylaniline (B-NDMA) were detected in quinolone solutions during irradiation with ultraviolet-A (UVA). Apparent levels of H2O2 and the B-NDMA per mole of quinolone paralled the phototoxic potentials in the mice. The N-NDMA induced by quinolones and UVA was inhibited partially by treatment with D-mannitol and dimethylsulfoxide, and also with diethylenetriamine-pentaaceticacid (DTPA), suggesting that Haber-Weiss and Fenton reactions occurred. UVA concentration-dependently increased the level of the B-NMDA in H2O2 solution and the swelling in the ear pretreated by intra-auricular injection of H2O2. Both augmentations were inhibited by DTPA or DMSO. The swelling induced by the 5 quinolones and UVA was completely inhibited by pretreatment with dimethylsulfoxide. Oxygen consumption was detectable during the photodegradation, and increased with time. These results showed that the phototoxic potentials of the 5 quinolones were probably related to the amounts of toxic oxygens generated in the target cells during irradiation.

No abstract