Photochemical Properties of Ofloxacin Involved in Oxidative DNA

Publication Date (Web): March 29, 2003 ... A Fluorine 1,2-Migration via Aryl Cation/Radical/Radical Anion/Radical Sequence .... Photochemistry and Pho...
0 downloads 0 Views 176KB Size
562

Chem. Res. Toxicol. 2003, 16, 562-570

Photochemical Properties of Ofloxacin Involved in Oxidative DNA Damage: A Comparison with Rufloxacin M. Consuelo Cuquerella,† Francisco Bosca´,† Miguel A. Miranda,*,† Alessandra Belvedere,‡ Alfio Catalfo,‡ and Guido de Guidi‡ Departamento de Quı´mica/Instituto de Tecnologı´a Quı´mica UPV-CSIC, Universidad Polite´ cnica de Valencia, Valencia, Spain, and Dipartimento di Scienze Chimiche, Universita` di Catania, Catania, Italy Received January 10, 2003

Photodegradation of ofloxacin (OFX) under aerobic conditions gives rise to N-demethylation, mainly involving coupling of radical cation OFX•+ with superoxide radical anion. Although H2O2 is produced as a byproduct, oxidative damage to DNA to give 8-OH-dGuo is associated with a type II mechanism. When the photosensitizing potentials of OFX and rufloxacin (RFX) are compared under the same conditions, the latter is shown to produce a much higher degree of DNA oxidation despite the close structural similarity. This is explained by a decrease of the triplet energy when sulfur instead of oxygen is attached to position 8 of the fluoroquinolone ring system. As a consequence, phosphate anions are able to quench OFX triplet but not RFX triplet; this reveals that the reaction medium has a strong influence on the photochemistry of OFX and hence on its photobiological properties.

Introduction

Chart 1

FLQs1

are antibacterial agents whose pharmacological action involves inhibition of the bacterial topoisosomerase DNA gyrase that controls the shape of DNA (1). However, these drugs can produce adverse effects in the presence of light. A number of reports have shown that FLQs may be efficient photosensitizers (2); some of them are responsible for undesired cutaneous reactions (3-10), while others can operate as photomutagenic and photocarcinogenic agents (11-14). In this context, photodehalogenation seems to be one of the main processes involved in the phototoxic effects of FLQs (15). Nevertheless, as this photodegradation route is not always significant, the photosensitizing properties of FLQs have also been attributed to the formation of ROS (16-19). As DNA is an important pharmacological target for FLQs, it makes sense to investigate FLQ-photosensitized DNA damage. Several authors have reported photoinduced guanine oxidation associated with FLQs (17, 18, 20, 21); however, most of the information available in the literature about the reaction mechanism is contradictory or incomplete. For example, the use of ROS scavengers and quenchers * To whom correspondence should be addressed. E-mail: [email protected]. † Universidad Polite ´ cnica de Valencia. ‡ Universita ` di Catania. 1 Abbreviations: Α, absorbance; , molar absorption coefficient; φ , F fluorescence quantum yield; φ∆, singlet oxygen quantum yield; φe, photoionization quantum yield; φISC, intersystem crossing quantum yield; (E)OFX, ofloxacin and its methyl ester; 8-OH-dGuo, 8-hydroxy2′-deoxyguanosine; BPH, benzophenone; ct-DNA, calf thymus DNA; DAD, photodiode array detector; dGuo, 2′-deoxyguanosine; dSp, diastereomeric mixture of spiroiminodihydantoin nucleosides; ROS, reactive oxygen species; ECD, electrochemical detector; FLQ, fluoroquinolone; HPLC, high performance liquid chromatography; MW, molecular weight; NP, naproxen; NQS, 1,2-naphthoquinone sulfonic acid; oxazolone, 2,2-diamino-4-[(2-deoxy-β-D-erythro-pentofuranosyl)amino]2,5dihydrooxazol-5-one; PB, phosphate buffer; PMT, photomultiplier tube; RFX, rufloxacin; dThd, thymidine; TRIS, R,R,R-tris-(hydroxymethyl)methylamine.

did not provide unambiguous evidence with regard to the pathway (type I or type II) involved in the process (17). In addition, the ability of FLQs to produce oxygen radical species as well as singlet oxygen cannot be directly associated with their phototoxicity (19, 22-24). Also, it was reported that the photooxidative potential cannot be correlated with the photodegradation quantum yield (17, 18). With this background, it appears necessary to devote further efforts to correlate the photophysical/photochemical properties of FLQs with their adverse photosensitizing side effects. The ultimate goal is to gain some predictive ability through the establishment of structureactivity relationships. Recently, we have reported on the photophysical and photochemical properties of RFX involved in the oxidative DNA damage photosensitized by this drug (25, 26). Ofloxacin (OFX) is a closely related FLQ (see Chart 1); the main structural difference is the presence of an oxygen atom (instead of sulfur) attached to position 8 of the quinolone ring system. This antibiotic exhibits both phototoxic and photocarcinogenic properties (12, 27). The aim of the present work has been to investigate the photophysical/photochemical processes of OFX involved in DNA damage, with special emphasis on the possible influence of the heteroatom (oxygen vs sulfur) on the key features of the FLQ chromophore and on their photobiological consequences.

10.1021/tx034006o CCC: $25.00 © 2003 American Chemical Society Published on Web 03/29/2003

Photochemical Properties of Ofloxacin

In this context, the photochemical and photobiological properties of OFX have been the subject of previous studies (18, 19, 28-31). Although such studies have provided valuable information on OFX, a fully satisfactory understanding of the photosensitizing ability of this drug is still missing. Thus, it has been observed that OFX photosensitizes singlet oxygen formation, but two different quantum yields have been reported, namely, 0.076 (19) and 0.13 (28). On the other hand, in the photolysis of OFX, fluoride release has been detected only under certain conditions (15, 29). Although OFX appears to induce photooxidation of DNA more efficiently than most of the photocarcinogenic FLQs, it is not able to produce cyclobutane thymine dimers, typical energy transfer products (18). It has been suggested that photooxidation of DNA by OFX is mainly due to singlet oxygen generation (type II mechanism); however, other photocarcinogenic FLQs with different photooxidizing potentials have a similar capability to generate singlet oxygen (19). Therefore, it appeared interesting to investigate the role played by type I (radical) and type II (singlet oxygen) mechanisms in the OFX photoinduced production of 8-OH-dGuo in DNA and to compare the results with those previously obtained for RFX (25). As in the case of RFX, parallel studies have been undertaken on (E)OFX for two reasons: (i) the complications associated with equilibrium between the carboxylic acid group and its conjugate base can be circumvented and (ii) esterification increases solubility in organic solvents and facilitates chromatographic analysis of the reaction mixtures. The most important results are that (i) photodegradation of (E)OFX under aerobic conditions involves Ndemethylation, explained by coupling of (E)OFX•+ and superoxide radical anion; (ii) although oxidative species such as H2O2 are generated in the aerated irradiations of (E)OFX, photoinduced production of 8-OH-dGuo in DNA is mainly associated with a type II mechanism; and (iii) the phosphate anions act as OFX triplet state quenchers. When comparing the above results on (E)OFX with those obtained for RFX under the same conditions, the most remarkable difference is that the latter produces a much higher degree of DNA oxidation. This is explained by a decrease of the triplet energy when sulfur instead of oxygen is the substituent at position 8 of the FLQ ring system. In PB, there is quenching of (E)OFX triplet, while in the case of RFX quenching by phosphate occurs to a much lower extent. Thus, the nature of the reaction medium has a strong influence on the photochemical behavior of (E)OFX and hence on its photosensitizing properties.

Experimental Procedures Chemicals. Sonicated ct-DNA was purchased from Pharmacia (Uppsala, Sweden). OFX, dGuo, 8-OH-dGuo, nuclease P1, alkaline phosphatase, perinaphthenone, NQS, guanidine, TRIS, sodium citrate dihydrate, sodium acetate trihydrate, ammonium formiate, acetic acid, trimethylamine (redistilled before use), CDCl3, and CD3OD were obtained from Sigma Chemical Company (St. Louis, MO). Methanol, acetonitrile (HPLC grade), and acetic acid were from Scharlau (Barcelona, Spain). Other chemicals were of reagent grade and used as received. Two different types of sodium PB (10 and 100 mM) were prepared from reagent grade products using deionized water; the pH of the solutions was measured through a glass electrode and adjusted with NaOH to be pH 7.3. EOFX was synthesized by the reaction of OFX with diazomethane in dichloromethane at room temperature during 24 h.

Chem. Res. Toxicol., Vol. 16, No. 4, 2003 563 Analytical Instrumentation. Ultraviolet spectra were recorded on a Shimadzu UV/vis scanning spectrophotometer (2101PC) with a slit width of 5 nm. The HPLC analyses of (E)OFX and their photoproducts were performed on a HPLC Varian Systems equipped with a 9012Q pump and a DAD (Varian 9065). The 1H NMR and 13C NMR spectra were measured by means of a Varian Gemini 300 MHz instrument; CDCl3 and CD3OD were used as solvents, and the signal corresponding to the deuterated solvent in each case was taken as the reference. Analysis of DNA hydrolysates and type II photoproducts of dGuo was performed with a Hewlett-Packard 1100 chromatograph equipped with on-line DAD and an ESA Coulochem ECD model 5100 (ESA, Inc., Bedford, MA) with a 5011 high sensitivity analytical cell (porous graphite) and a model 5020 guard cell, run in the screen mode. To evaluate the formation of type I photoproducts of dGuo, a HPLC Kontron system model 420 equipped with a 432 UV detector, a SFM fluorescent detector, and a SA 360 autosampler was employed. Fluorescence Measurements. The steady state fluorescence spectra were obtained with a FS900 Edinburgh Analytical Instruments apparatus, equipped with a 450 W xenon lamp. The procedure to determine the fluorescence quantum yields of (E)OFX was the same as that reported for RFX (25). The fluorescence quantum yield of quinine bisulfate in 1 N H2SO4 (φF ) 0.546) was used as standard (28). Laser Flash Photolysis Measurements. The equipment and procedure used for laser flash photolysis have been described elsewhere (25). Briefly, a pulsed Nd:YAG SL404G-10 Spectron Laser Systems was used for the excitation at 355 nm. The single pulses were ∼10 ns in duration, and the energy was from 1 to 10 mJ/pulse. Unless otherwise stated, OFX and EOFX were dissolved in 10 mM PB aqueous solutions. The pH was adjusted to 7.3 (with NaOH), and the absorbance at 355 nm was adjusted to 0.4. The solutions were deaerated (when specified) by bubbling nitrogen or N2O. Determination of φe was carried out using the comparative method with BPH in acetonitrile as standard. Thus, φe was obtained by application of eq 1 (32)

φe ) φISC(BPH) × ∆A[e (720 nm)] × [3BPH (520 nm)]/ ∆A[3BPH (520 nm)] × [e (720 nm)] (1) where ∆A[e (720 nm)] refers to the difference between the transient absorbance of OFX (or EOFX) at 720 nm under nitrogen and under N2O atmosphere, while ∆A[3BPH (520 nm)] refers to the absorbance of 3BPH at 520 nm. The BPH triplet molar absorption coefficient and quantum yield in acetonitrile were taken to be [3BPH (520 nm)] ) 6500 M-1 cm-1 and φISC(BPH) ) 1, respectively. The molar absorption coefficient of solvated electron was taken to be [e (720 nm)] ) 18 500 M-1 cm-1 (28, 32). The molar absorption coefficient of 3EOFX was determined by monitoring the energy transfer from the EOFX triplet excited state to NP (a naphthalene derivative). This experiment was performed using deaerated aqueous solutions (10 mM PB at pH ) 7.3) of 0.9 × 10-4 M EOFX containing different concentrations of NP. Thus, the [3EOFX(610 nm)] was calculated using eq 2

k2/k2 - k1 × ∆A[3NP (440 nm)] × [3EOFX (610 nm)] ) ∆A[3EOFX (610 nm)] × [3NP (440 nm)] (2) where ∆A[3EOFX (610 nm)] refers to the transient absorbance of EOFX at 610 nm at the beginning of the reaction, ∆A[3NP(440 nm)] refers to the NP triplet state (440 nm) at the end of the reaction, k1 is the 3EOFX decay rate constant in the absence of NP, and k2 is the different 3EOFX decay rate constant obtained with varying concentrations of NP. The molar absorption coefficient of 3NP in aqueous solutions was taken to be [3NP(440 nm)] ) 9500 M-1 cm-1 (33). This led to a value of 3900 ( 300 M-1 cm-1 for the molar absorption coefficient of 3EOFX at 610 nm.

564

Chem. Res. Toxicol., Vol. 16, No. 4, 2003

Cuquerella et al.

In the case of the OFX triplet excited state, two procedures have been used to determine the value of : (i) the same procedure described above for EOFX and (ii) the singlet depletion method, where it was first assumed that the triplet molar absorption coefficient varies linearly between 355 and 370 nm and afterward eq 3 was applied at the above wavelengths (λ ) 355 and 370 nm) and also at the intermediate wavelength of λ ) 360 nm.

∆Ai ) (Ti - Si) [3M] l

(i ) λ)

(3)

where [3M] is the concentration of triplet state, l is the optical path length of the monitoring beam, Ti is the triplet molar absorption coefficient at a given λ value, and Si is the molar absorption coefficient of the OFX ground state at a given λ. The Ti values at 610 nm determined by methods i and ii resulted to be essentially coincident (5900 ( 330 and 6000 ( 360 M-1cm-1, respectively). Determination of the φISC of OFX and EOFX was carried out using the comparative method described above for the photoionization process. Steady State Photolysis of (E)OFX. Irradiations were performed as previously described (25) using a Rayonet photochemical reactor equipped with eight black light phosphor lamps emitting in the 310-390 nm range, with a maximum at 350 nm. The fluence rate at the irradiation position was about 450 µW/cm2. The light intensity measurements have been described previously (34). The photoreactions of (E)OFX in PB aqueous solutions 60 µM in 10-100 mM PB (pH 7.3) were performed under both aerobic and anaerobic conditions and monitored by reverse phase HPLC using an analytical Kromasil 100 C18 column (Tracer, 25 cm × 0.4 cm, mean particle size 5 µm), a flow rate of 1 mL/min, and a mixture of acetonitrile/water/ trifluoracetic acid 15/84.9/0.1 (OFX). In addition, methylene blue (ca. 2 × 10-5 M) photosensitized degradation of OFX (0.1 mM) in 10 mM PB aqueous solution was performed with the filtered light (visible long pass filter with cutoff wavelength at 420 nm) from an Osram-HQL 125 W medium pressure Hg lamp located inside an immersion well photoreactor (Applied Photophysics model 3230). Isolation and Identification of the Photoproducts. The free carboxylic acid 1 (photoproduct 1) was obtained by two methods: (i) irradiation of OFX in aerated aqueous solutions and (ii) irradiation of EOFX in aerated aqueous solutions, followed by extraction with CH2Cl2, isolation of the N-demethylated photoproduct 2 (column chromatography on silica gel using dichloromethane/methanol 8/2 as eluent), and subsequent treatment with 1 M NaOH. Compound 1 was known (35), while EOFX and its demethylated photoproduct 2 have not been previously reported. Identification was mainly based on 1H NMR and 13C NMR spectroscopy. The spectral data for the new compounds EOFX and 2 follow. EOFX: exact mass found, 375.1587; calcd for C19H22N3O4F, 375.1594. 1H NMR: δ 1.54 (d, J ) 6.5 Hz, 3H, CH3), 2.36 (s, 3H, NCH3), 2.55 (m, 4H, MeNCH2CH2N), 3.34 (m, 4H, MeNCH2CH2N), 3.87 (s, 3H, OCH3), 4.4 (m, 3H, OCH2CHN), 7.49 (d, JH-F ) 12.5 Hz, 1H, H-5), 8.23 (s, 1H, H-2). 13C NMR (CDCl3): δ 18.5, 46.8, 50.9, 52.2, 55.1, 56.1, 68.5, 105.7 (JC-F ) 23 Hz), 109.6, 123.7, 124.1, 132.0 (JC-F ) 15 Hz), 140.0, 145.6, 156.0 (JC-F ) 247 Hz), 166.4, 173.0. Photoproduct 2: exact mass found, 361.1436; calcd for C18H20N3O4F, 361.1438. 1H NMR: δ 1.57 (d, J ) 6.9 Hz, 3H, CH3), 3.01 (m, 4H, MeNCH2CH2N), 3.31 (m, 4H, MeNCH2CH2N), 3.92 (s, 3H, OCH3), 4.35 (m, 3H, OCH2CHN), 7.72 (d, JH-F ) 12.5 Hz, 1H, H-5), 8.35 (s, 1H, H-2). 13C NMR (CDCl ): δ 17.3, 45.8, 51.1, 51.3, 53.9, 67.3, 105.2 (J 3 C-F ) 24 Hz), 108.9, 123.0, 125.1, 131.3 (JC-F ) 15 Hz), 138.7, 144.6, 158.5 (JC-F ) 247 Hz), 165.8, 172.1. Determination of H2O2 from Aerated Irradiations of OFX. The method is based on the reaction of H2O2 with Fe2+ to produce Fe3+, which in the presence of xylenol orange gives rise to the formation of a complex absorbing at 540 nm (36). Thus, aliquots (1 mL) of 60 µM OFX aqueous solutions in PB (10 mM)

were irradiated for different times (0-40 min) and then added to a solution prepared by mixing 1 mL of acidified 0.4 mM [Fe(SO4)2(NH4)2]‚6H2O (0.1 N H2SO4) with 1 mL of 1 mM xylenol orange in the same buffer. The absorbance of the resulting solution at 540 nm was measured after 1 h of incubation at room temperature in the dark. The same experiments under the same conditions were performed for RFX as a reference. Determination of 8-OH-dGuo in DNA. ct-DNA was incubated at a concentration of 0.1 mg/mL in 1720 µL of 60 µM (E)OFX in PB solution and then irradiated for 15 min. Subsequently, DNA was precipitated by addition of 170 µL of 3 M NaOAc and 5.1 mL of cold (-20 °C) absolute ethanol. The resulting mixture was stored at -20 °C during 2 h in the dark and subsequently centrifuged for 15 min at 13 000g at -10 °C. The DNA pellet was separated from the solution, washed with 4 mL of cold 80% ethanol, and dried under a stream of nitrogen. To obtain the free nucleosides for injection into the HPLC system, the following procedure was employed. The pellet was dissolved in 200 µL of TRIS-HCl 0.01 M (pH ) 7.00), 5.7 µL of 0.5 M NaOAc (pH ) 5.1), and 5 units of nuclease P1 in order to break the 3′-5′ sugar-phosphate bonds of DNA. After the solution was vortexed and incubated at 37 °C for 60 min in the dark, 80 µL of 0.4 M TRIS-HCl (pH ) 7.5) and 3 units of alkaline phosphatase were added to this solution. The resulting hydrolysates were centrifuged for 20 min at 13 000g within an Eppendorf microfuge, using a 10 000 MW cutoff Millipore filter, to remove enzymes and other polymeric residues such as undigested DNA (37). For analysis of DNA hydrolysates, 20 µL of the filtrate was injected into a reverse phase Supelcosil LC-18-S (5 µm, 4.6 mm i.d. × 25 cm), protected by a Supelguard LC-18-S. Elution was done using 12.5 mM aqueous sodium citrate, 25 mM aqueous sodium acetate, and acetic acid (pH ) 5.1) with 7% methanol (v/v) at a rate of 1 mL/min. The UV traces were monitored at 254 and 293 nm, which are the absorption maxima of dGuo and 8-OH-dGuo, respectively. The parameters of the ECD were set as follows: guard cell, 430 mV; detector I, 120 mV; detector II, 380 mV; response time, 0.1 s; gain, 10 × 100. Standard curves for dGuo and 8-OH-dGuo were obtained using authentic samples (38). Results are reported as (residues of 8-OH-dGuo vs residues dGuo) × 105 (39). Retention times for dGuo and 8-OH-dGuo were ca. 18 and 27 min, respectively. The same experiments under the same conditions were performed for RFX as a reference. The degradation kinetics of photosensitizers (E)OFX and RFX were done as described in the photolysis section. Determination of Type II dGuo Photoproducts. dGuo (1 mM) was irradiated in the presence of 60 µM (E)OFX in 3 mL of PB for various periods of time. Formation of a dSp was monitored with the HPLC chromatograph equipped with the DAD. Thus, 20 µL of the irradiated solution was injected in this system, and separation was achieved with a Supelcosil LC-NH 2 aminopropil substituted column (5 µm, 4.6 mm i.d. × 25 cm) protected by a Supelguard LC-NH2. Elution was done in an isocratic mode with a mixture of 25 mM aqueous ammonium formiate/acetonitrile (20:80 v/v) as the mobile phase at a rate of 1 mL/min. The UV traces were monitored at 230 and 254 nm, which correspond to the absorption maxima of the diastereomers (produced mainly by a type II mechanism) (40-43) and dGuo, respectively. A calibration curve for type II photoproducts was constructed using adequate standards. The same experiments under the same conditions were performed for RFX as a reference. The degradation kinetics of photosensitizers (E)OFX and RFX were done as described in the photolysis section. Determination of Type I dGuo Photoproducts. A more sensitive method for the determination of type I photoproducts is based on the reaction of guanidine produced by alkaline hydrolysis of dGuo photoproducts with NQS to form a fluorescent product (44). In this case, 250 mL of the dGuo/(E)OFX irradiated solution described above was treated with 250 mL of 1 N NaOH for 15 min at 60 °C. After this treatment, a 3 × 10-2 M aqueous solution of NQS (50 mL) was added, and the mixture was heated for 5 min at 60 °C and subsequently

Photochemical Properties of Ofloxacin

Chem. Res. Toxicol., Vol. 16, No. 4, 2003 565 Table 1. Quantum Yields of Fluorescence (OF), Photoionization (Oe), and Intersystem Crossing (OISC) of (E)OFX in 10 mM PB Aqueous Solutions at pH ) 7.3

φF φe φISC

OFX

EOFX

0.1 0.13 0.32

0.035 0.07 0.32

Table 2. Quenching Rate Constants (kq) of 3(E)OFX and 3RFX in 10 mM PB Aqueous Solutions at pH ) 7.3 by Different (Bio)Molecules kq (M-1 s-1) quencher

3OFX

3EOFX

3RFXa

O2 NP dGuo DNA dThd phosphate

3× 3 × 109 3 × 108 1 × 108