Abstract
Gemifloxacin is a fluoroquinolone antibacterial compound with enhanced affinity for bacterial topoisomerase IV and is being developed for the treatment of respiratory and urinary tract infections. The disposition and metabolic fate of this antibiotic was studied in the rat and the dog, the animal species used in its toxicological evaluation. The investigations were carried out following oral and intravenous administration of gemifloxacin mesylate. Gemifloxacin is a racemic compound; therefore, the pharmacokinetics of its individual (+) and (−) enantiomers were characterized using a chiral high-performance liquid chromatography/tandem mass spectrometry assay. In both rat and dog, the pharmacokinetic profiles of the (+) and (−) enantiomers were essentially identical. The enantiomers were rapidly absorbed following oral administration of racemic gemifloxacin mesylate. They distributed rapidly beyond total body water, and their blood clearance values were approximately equal to one quarter of the hepatic blood flow in each species. Terminal phase elimination half-lives were ca. 2 h in the rat and 5 h in the dog. Gemifloxacin was metabolized to a limited extent following oral and intravenous administration of [14C]gemifloxacin mesylate, and all metabolites formed were relatively minor. The principal metabolites formed were the E-isomer (4–6% of dose) and the acyl glucuronide of gemifloxacin (2–6% of dose) in both species andN-acetyl gemifloxacin (2–5% of dose) in the rat. Data obtained following intravenous administration indicated that gemifloxacin-related material is eliminated from the body via urinary excretion, biliary secretion, and gastrointestinal secretion. Material was eliminated approximately equally by the three routes in the dog, whereas a slightly higher proportion of the dose was eliminated in the urine (46%) and a lower proportion in the bile (12%) of rats.
Gemifloxacin [(R,S)-7-(3-aminomethyl-4-syn-methoxyimino-1-pyrrolidinyl)-1-cyclopropyl-6-fluoro-1,4-dihydro-4-oxo-1,8-naphthyridine-3-carboxylic acid] is a fluoroquinolone antibacterial compound with enhanced affinity for bacterial topoisomerase IV and is being developed for the treatment of respiratory and urinary tract infections. The compound has a broad spectrum of activity against Gram-positive and Gram-negative bacteria (Oh et al., 1996; Cormican and Jones, 1997; Hohl et al, 1998). It has shown potent antibacterial activity against clinical isolates and reference strains in both in vitro studies and experimental models of infection in animals (Erwin and Jones, 1999; Johnson et al., 1999; Berry et al., 2000). The compound is particularly active against Gram-positive organisms including penicillin-, macrolide-, and quinolone-resistant Streptococcus pneumoniae (Hardy et al., 1999). It is ca. 4-fold more potent than moxifloxacin against S. pneumoniae (Johnson et al., 1999). Gemifloxacin has also shown potent activity against other major pathogens involved in respiratory tract infections, including Haemophilus influenzae andMoraxella catarrhalis and the atypical organisms,Legionella pneumophila, Chlamydia spp., andMycoplasma spp. (Felmingham et al., 1999; Hannan and Woodnutt, 2000). Furthermore, the compound has shown potent activity against many organisms that cause urinary tract infections (Cormican and Jones, 1997; Naber et al., 1999). This article provides an overview of a series of separate experiments conducted to determine the fate of gemifloxacin in rats and dogs, the species used in the toxicological evaluation of the compound. The compound was administered as its mesylate salt, and the doses selected were within the range of doses used in toxicology studies.
Materials and Methods
Chemicals.
[14C]Gemifloxacin mesylate, synthesized in the Chemical Development Department, SmithKline Beecham Pharmaceuticals, Harlow, UK (SB1) and nonradiolabeled gemifloxacin mesylate obtained from LG Chem Biotech Research Institute (Taejon, Korea) were used in these experiments. The 14C compound was labeled at position C-3 of the naphthyridine ring (Fig.1), and several batches with specific activities in the range 0.5 to 43.7 μCi/mg were used. The radiochemical purity of each batch was >98%. The chemical purities of both radiolabeled and nonradiolabeled compounds were at least 94%. SB-414000 (N-acetyl gemifloxacin), SB-414006 (O-desmethyl gemifloxacin), and SB-414007 (hydroxymethyl gemifloxacin) were used as chromatography reference standards and were also obtained from LG Chem Biotech Research Institute. All other chemicals used were reagent grade or better and were obtained from standard commercial suppliers.
Animals.
Separate groups (typically three to four animals per group) of male rats (Sprague-Dawley, CD strain, 210–455 g) and male Beagle dogs (10–20 kg) were used in these studies. The rats were obtained from either Charles River Ltd. (Margate, UK) or Harlan OLAC Ltd. (Bicester, UK). They had free access to standard laboratory pelleted diet (LAD 1, SQC rat and mouse maintenance diet A or SDS rat and mouse maintenance diet 1; SDS, Witham, UK) and water. The rats dosed orally with nonradiolabeled compound were fasted overnight and until 6 h after dosing. For the collection of excreta, they were housed individually in glass metabolism cages, and for other studies they were housed singly or in groups, in either plastic cages containing wood shavings or stainless steel cages with raised wire mesh floors to prevent coprophagy. The dogs were either supplied by SmithKline Beecham Pharmaceuticals (Hertfordshire, UK) or were obtained from either Bantin & Kingman (Hull, UK) or Ridglan Farms (Mt. Horeb, WI). The dogs were fed daily with ca. 400 g of either Harlan Teklad 9682 dog diet (Shaw's Farm, Bicester, UK) or Certified Canine Diet #5007 (PMI Feeds, Inc., Richmond, IN), but the animals were generally fasted overnight before dose administration and until up to ca. 6 h after dosing. Since emesis was observed in some animals following oral administration, the bile duct-cannulated dogs received half their ration of food immediately before oral dosing and the other half ca. 4 h after dosing. The dogs were housed individually in either stainless steel metabolism cages or in animal holding rooms. Standard conditions of temperature (18–29°C) and humidity (50 ± 20%) were maintained for both rats and dogs in these experiments. Room lighting was set to give 12 h of light per 24-h period.
Bile Duct Cannulation.
Rats
For excretion and metabolism studies, rats were surgically fitted with indwelling bile duct cannulae using a method similar to that described by Van Wijk et al. (submitted for publication). Each animal undergoing surgery received a subcutaneous injection of carprofen (Zenecarp, C-Vet VP, 50 mg/ml) at 5 mg/kg, before and ca. 24 h after surgery. Under general anesthesia using a mixture of isoflurane in oxygen/nitrous oxide, a cannula was inserted in the proximal bile duct and another one in the duodenum. Each cannula was passed through the abdominal wall, truncated subcutaneously, and exteriorized through the ventral surface of the tail. Each cannula was protected by means of a metal tail cuff overlying the exit site and a spring assembly. The animal was then placed in a glass metabolism cage modified to allow the free end of each cannula to be attached to a swivel joint fixed above the cage lid and allowed to regain consciousness. During a 5-day postoperative recovery period, physiological saline was infused through the duodenal cannula (ca. 0.5–0.6 ml/h) for the first day, after which the infusate was replaced with a 0.05% solution of artificial bile salts (Sigma, St. Louis, MO) in saline. After the recovery period, animals were selected for dose administration based on body weight gain and bile flow rate.
Dogs.
Dogs were surgically prepared using a method based on that described byKissinger et al. (1998). Before surgery, the dogs were fasted overnight and sedated with an intramuscular injection of acepromazine (10 mg/kg). Anesthesia was maintained during the procedure using isoflurane and oxygen. A T-piece apparatus, with three catheters running perpendicular from the T-piece (the bile collection catheter, the diaphragm catheter, and the distal flush catheter), was surgically inserted into the common bile duct between the last hepatic duct junction and the entrance to the duodenum and secured with 3–0 Prolene sutures. Each catheter was connected to a port implanted subcutaneously. By inflating the diaphragm, this device permitted the diversion of bile into an external collection apparatus and simultaneous infusion of a bile salts replacement solution. After bile collection was completed, the diaphragm was deflated to resume the bile flow into the duodenum. The dogs were treated orally with Clavamox (22 mg/kg b.i.d for 5 days) and Fentanyl, Duragesic-25 patches (one patch applied on the day before surgery and another one 2 days after surgery) to aid recovery. The animals were placed in jackets and tethers and allowed to recover for at least 10 days before dosing.
Dose Preparation and Administration.
Due to the photoinstability of the compound, dose preparation and administration was carried out under subdued lighting. For excretion/metabolism studies, single oral doses of [14C]gemifloxacin mesylate were given to rats, by gavage, at 168 mg of free base (fb)/kg in saline (10 ml/kg) and to dogs at 24 mg of fb/kg in gelatine capsules. Intravenous doses of the compound were given to both species at 10 mg of fb/kg, as a constant rate infusion over 30 min in sterile isotonic saline (10 and 5 ml/kg to the rat and dog, respectively). The dose was infused into a tail vein in the rat and into a cephalic vein in the dog. For tissue distribution, the radiolabeled compound was given orally to rats at 20 mg of fb/kg in saline (5 ml/kg).
In separate experiments, nonradiolabeled gemifloxacin mesylate was dosed orally and intravenously to rats and dogs in a crossover design to determine the pharmacokinetics of the individual enantiomers of gemifloxacin in plasma. The rats were surgically prepared before dose administration. Under isoflurane anesthesia, a cannula was inserted into a femoral vein (for dose administration) and another one was inserted into the jugular vein (for blood sampling). The cannulae were exteriorized at the back of the neck. On completion of surgery, each animal received a dose of Penbriten (ca. 150 mg/kg) and carprofen (ca. 20 mg/kg) subcutaneously to aid recovery. The animals were placed in jackets with tethers and allowed to recover for 5 days before dosing. Oral doses were administered in saline at 30 mg of fb/kg to the rat and in gelatine capsules to the dog at 10 mg of fb/kg. Intravenous doses were administered in sterile isotonic saline at 10 mg of fb/kg to both species. In the rat, the dose was infused into the femoral vein over 1 h (10 ml/kg), and in the dog into a saphenous vein over 20 min (5 ml/kg).
Collection of Samples.
Blood samples for the determination of concentrations of radioactivity and plasma radiometabolite patterns were collected after oral administration at 168 and 24 mg of fb/kg to the rat and dog, respectively. In the rat, terminal blood samples (by exsanguination under carbon dioxide narcosis) were obtained from three animals at each of 1-, 2-, and 6-h time points after dosing. Control samples were obtained from three undosed animals. In the dog, serial samples (5 or 20 ml) were collected via the jugular vein from each of three animals predose and at 2, 6, and 12 h after dosing. The blood samples were collected into tubes containing EDTA and the plasma separated by centrifugation. For each species, plasma was pooled per time point for the determination of radiometabolite patterns. Blood samples were also collected from additional groups of three to four animals dosed orally and intravenously with nonradiolabeled gemifloxacin mesylate to specifically assay for the individual enantiomers in plasma. In the rat, ca. 120-μl aliquots were collected at various times up to 24 h after dosing via the jugular vein of each animal. In the dog, 1-ml aliquots were withdrawn at various times up to either 24 h (i.v.) or 30 h (oral) after dosing from the cephalic vein of each animal. The samples were collected in tubes containing either heparin or EDTA as anticoagulant and the plasma separated by centrifugation.
For distribution studies in the rat, a blood sample was collected by cardiac puncture (under isoflurane/oxygen anesthesia) from one animal at each of 0.5, 1, 2, 4, 7, 24, 48, 72, 120, and 168 h after dosing with [14C]gemifloxacin mesylate at 20 mg of fb/kg. Each animal was then sacrificed and an extensive range of organs/tissues removed from the carcass.
Excreta and bile samples were collected following oral and intravenous administration of [14C]gemifloxacin mesylate to groups of three or four bile duct-cannulated animals. Control samples were obtained from the same animals before dose administration. Bile samples in the rat, and urine and bile samples from the dog, were collected over 0- to 24-h (i.v.) or 0- to 6-h and 6- to 24-h (p.o.) time periods during the first day and then daily for up to 96 h after dosing. Urine samples from the rat and feces samples from both species were collected over 24-h periods up to 96 h after dosing. Urine and bile samples from both species and fecal samples from the rat were collected in containers surrounded by solid carbon dioxide. Fecal samples from the dog were collected at ambient temperature. Cages were rinsed with water after each collection period and the washings were retained. In addition, cage debris was removed from dog cages before rinsing. Where possible, samples were protected from light during collection. At the end of the final collection period (96 h), the rats were killed and the gastrointestinal tract (G.I. tract) was excised from each animal. The residual carcass was retained for radioassay.
For the determination of radiometabolite patterns, portions of urine, bile, and plasma samples were treated with phosphoric acid to ca. pH 4 to 6 to prevent possible degradation of any acyl glucuronides present in the samples. Control samples were spiked with [14C]gemifloxacin mesylate at appropriate concentrations and treated in the same way as the test samples. All samples were subsequently stored in the dark at approximately −20°C or lower.
Sample Analysis.
Radioassay
Radioactivity was measured by liquid scintillation counting (LSC), using a Packard 1900TR (Packard Instrument Company, Downers Grove, IL), a Tri-Carb 1600TR (Packard Instrument Company, Meriden, CT) or a Wallac 1410 (Turku, Finland) spectrometer. Before scintillation counting, samples were prepared as follows: urine, plasma, cage washes, and HPLC samples were mixed with either Ultima Gold XR (Packard Instrument Company) or Quickszint 1 (Zinsser Analysis, Maidenhead, UK) scintillator (5 or 10 ml). Cage debris and fecal samples were homogenized in water (approximately 1:1 or 1:2, w/v), and aliquots were burned in oxygen using a Canberra Packard Ltd. automatic sample oxidizer (Pangbourne, UK). Samples of blood and tissues were also combusted using a sample oxidizer. Tissues were combusted whole or were either scissor-minced or homogenized in methanol/water (1:1, v/v), and duplicate weighed aliquots of each sample were taken for combustion. Combustion products for all samples were either absorbed in Opti-Sorb 1 and mixed with Opti-Sorb S scintillator (Fisons PLC, Loughborough, UK) or were absorbed in Carbosorb and mixed with Permafluor E+ scintillant (Canberra Packard Ltd. or Packard Instrument Company) before radioassay. Rat carcasses and G.I. tracts were solubilized in a mixture containing 12% (w/v) aqueous sodium hydroxide, methanol, and Triton X-100 (8:1:1, v/v). Triplicate aliquots of each digest were mixed with 10 ml of Quickszint 1 scintillator before LSC.
Specific assays.
Concentrations of the individual (+) and (−) enantiomers of gemifloxacin in rat and dog plasma were determined using a specific HPLC/MS-MS method. The compounds were extracted from plasma (50 μl) by precipitating the proteins using acetonitrile containing the internal standard (racemic [13C,2H3]gemifloxacin). Following centrifugation, an aliquot of the supernatant (20 μl) was injected onto a chiral HPLC column [Crownpak CR(+), Daicel, Cedex, France]. The enantiomers and the internal standard were detected by positive ion MS-MS using a Turbo IonSpray interface. Using this method, linear responses were observed over the concentration range of 10 to 2500 ng/ml. The lower limit of quantification for each enantiomer in both rat and dog plasma was 10 ng/ml, and accuracy was within 9%. Within-run coefficients of variation were <8%.
Pharmacokinetic parameters were derived using noncompartmental methods and calculated using either WinNonlin Professional (version 1.5) or an in-house validated computer program. Maximum concentrations (Cmax) and time to reach maximum concentrations (Tmax) were determined by visual inspection of the concentration versus time curves. The elimination half-lives were derived by linear regression analysis of the data deemed to visually lie in the terminal part of the log-transformed curve. Area under the plasma concentration versus time curve to the last data point [AUC(0–t)] was calculated using either a linear or a combined linear-logarithmic trapezoidal method. The extrapolated AUC from the time of the last quantifiable data point (t) to infinity [AUC(t–inf)] was estimated by dividing the predicted concentration at time tby the terminal phase rate constant. AUC(0–inf) was calculated as the sum of AUC(0–t) and AUC(t–inf). Plasma clearance was calculated by dividing the dose by AUC(0–inf) and converted to blood clearance using the blood:plasma concentration ratio (1.8 in both species; D. Kenworthy and G. Duncan, unpublished data). Volume of distribution at steady state (Vss) was calculated as the plasma clearance multiplied by the mean residence time [AUMC(0–inf)] divided by AUC(0–inf) minus T/2, where AUMC(0–inf) is the area under the moment curve extrapolated to infinity and T is the infusion time. Bioavailability (F) was calculated as the ratio of dose-normalized AUC(0–inf) after oral and intravenous administration.
Radiometabolite profiles.
Radiometabolite profiles were determined by HPLC and quantified using either a flow-through radiodetector or fraction collection followed by scintillation counting. Acid-stabilized urine and bile from both species and plasma from the rat were analyzed directly. Any particulate matter present in the samples was removed by centrifugation before analysis. Radioactivity in acid-stabilized dog plasma was extracted using Oasis HLB solid phase extraction cartridges (Waters Corporation, Milford, MA), whereas fecal samples from both species were extracted using acetonitrile/0.1 M hydrochloric acid (80:20, v/v). Dog fecal extracts were diluted with water (1:7, v/v) before analysis. An estimate of the recovery of radioactivity was determined by assaying aliquots of the extracts by LSC. Extraction efficiencies were in the range of 71 to 85% for all samples extracted, and these values were taken into account when calculating the proportions of the sample radioactivity accounted for by each metabolite. Neat samples or extracts were injected onto either a Prodigy ODS-3 or a YMC Pack ODS-AQ column (5 μm; 25 cm × 4.6-mm i.d.). The column was then eluted at ambient temperature, with 50 mM ammonium acetate, pH 5.0 (solvent A) and acetonitrile (solvent B) as the mobile phase, using the following solvent gradient program (expressed as percentage of solvent B): 10 to 20% over 0 to 10 min; held at 20% over 10 to 20 min; 20 to 40% over 20 to 40 min; 40 to 100% over 40 to 41 min; held at 100% over 41 to 45 min; and 100 to 10% over 46 to 55 min. Where necessary (rat plasma extracts and dog bile), the solvent gradient program was adjusted slightly to improve resolution between peaks. A flow rate of 1.0 ml/min was used. The column eluate was monitored continuously using a Jasco UV-785 absorbance UV/Visible spectrophotometer (Jasco Ltd., Great Dunmow, UK) set at 272 nm and a β-Ram radiochemical detector (LabLogic, Sheffield, UK). For samples that did not contain sufficient radioactivity for direct on-line radiodetection, 0.2-ml fractions of the eluate were collected and assayed for radioactivity by LSC (TopCount, Packard Ltd., Pangbourne, UK) using a solid scintillant (Yttrium silicate; Lumaplate). Control samples spiked with [14C]gemifloxacin mesylate were exposed to the same conditions as the test samples and were analyzed using the same procedures.
Identification of metabolites.
Where available, appropriate authentic reference compounds were used as chromatographic markers to aid the identification of metabolites. Positive identification of the metabolites was obtained by HPLC-MS and/or HPLC/MS-MS analyses of representative samples. Further confirmation of the identity of the acyl glucuronide of gemifloxacin was obtained by NMR of an isolate obtained from dog bile. Control samples spiked with [14C]gemifloxacin mesylate and aqueous solutions of reference compounds were analyzed using HPLC-mass spectrometry only. The HPLC method used for the isolation of metabolites was essentially identical to that used for the determination of the radiometabolite patterns. The column eluate was split such that ca. 0.7 ml/min was directed toward either a β-Ram radiodetector (LabLogic) for on-line radiodetection, to a Gilson 222 XL fraction collector (Anachem, Luton, UK) for off-line radiodetection, or to waste when radiodetection was not required. The remaining 0.3 ml/min was directed into the MS interface. Mass spectrometry was carried out using a Finnigan LCQ ion trap mass spectrometer (ThermoQuest, Hemel Hampstead, UK) with positive ion electrospray ionization. Before NMR, the acyl glucuronide was isolated from dog bile using 3-ml Oasis HLB solid phase cartridges and cleaned up by HPLC. The resulting methanolic extracts were combined, evaporated (TurboVap, Zymark Instruments, Cheshire, UK), and centrifuged (1500g, 10 min) before dilution with water and subsequent HPLC. Successive injections (1–3 ml) of the diluted extract were made onto an HPLC system using the conditions described previously, and appropriate fractions were collected and pooled. Desalting of the pooled fraction was carried out by HPLC but with water instead of ammonium acetate in the mobile phase. The desalted fraction was evaporated to dryness under a stream of nitrogen at room temperature and reconstituted in deuterium oxide before analysis by NMR using a Bruker DRX 500 pulse Fourier transform spectrometer, tuned to 500.13 MHz for proton nuclei (Bruker, Coventry, UK). Confirmation of the presence of gemifloxacin-N-oxide was carried out by treating appropriate samples with acidified titanium (III) chloride and analyzing the reduced product by mass spectrometry. Glucuronides were characterized by the treatment of selected samples with β-glucuronidase and monitoring the retention times of the liberated aglycones.
Results
Oral Absorption and Bioavailability.
It was not possible to accurately determine the proportion of the oral dose absorbed in either species since data obtained after intravenous doses of [14C]gemifloxacin mesylate to bile duct-cannulated rats and dogs indicated that there was a propensity for gastrointestinal secretion of drug-related material to occur. Approximately one third of the intravenously administered dose was excreted in the feces of these animals (Table1). In both species, the systemic oral bioavailability of the (+) and (−) enantiomers was similar, but in the dog interanimal variation in bioavailability was high (Table2). The systemic oral bioavailability of both enantiomers was ca. 11% in the rat at 30 mg of fb/kg, whereas it ranged from 28 to 99% in the dog at 10 mg of fb/kg. Peak concentrations of radioactivity and of gemifloxacin were generally observed within 2 h of oral administration in both species, indicating that absorption was relatively rapid.
Pharmacokinetics.
Similar plasma profiles (Figs. 2 and3) and pharmacokinetic parameters were obtained for the two enantiomers of gemifloxacin (Table 2) after both routes of administration. After oral administration, the compounds appeared rapidly in the systemic circulation reaching peak concentrations at 1.3 h in the rat and 1.5 h in the dog. The concentrations declined thereafter in an apparent bi-exponential manner. Data obtained after intravenous doses indicated that the mean blood clearance was low (ca. 1.3 and 0.4 l/h/kg in rat and dog, respectively) and the mean volume of distribution (Vss) was large in both species (ca. 4 l/kg). The mean apparent terminal phase elimination half-life of each enantiomer was ca. 2 h in the rat and 5 h in the dog.
Distribution.
Drug-related material was distributed widely throughout most organs and tissues following oral administration of [14C]gemifloxacin mesylate to the rat at 20 mg of fb/kg. Apart from the contents and the walls of the G.I. tract, highest concentrations of radioactivity (in terms of micrograms of free base equivalent per gram) were observed at 0.5 h in the adrenals (9.1), aorta (9.0), blood (2.0), brain (0.2), fat (0.6), heart (4.5), kidneys (20), liver (31), lungs (35), lymph nodes (7.1), pancreas (8.8), plasma (1.8), salivary gland (7.6), spinal cord (2.5), spleen (7.7), and the thyroid (150). The highest concentrations in the bladder (13), bone (2.7), bone marrow (6.5), the eye (0.8), lacrymal gland (6.4), pituitary (6.2), and the vena cava (5.7) were seen at 1 h, in muscle (8.4), prostate (3.9), and thymus (4.4) at 2 h, and in the skin (3.1) and the testes (1.4) at 7 h after dosing. The subsequent elimination of radioactivity was generally rapid and essentially complete by 24 h after dosing and, apart from the G.I. tract, low levels of radioactivity (generally <0.1 ugEq/g or ≤0.01% of dose) were only detectable in the bone, brain, fat, kidneys, lacrymal gland, liver, lungs, spinal cord, testes, and the thymus at 168 h after dosing.
Excretion.
In both species, the major route of elimination of drug-related material following oral administration of [14C]gemifloxacin mesylate was via the feces (Table 1). Renal and biliary elimination were minor routes. Following intravenous administration of the compound, approximately equal proportions of the dose were eliminated in the urine, bile, and feces in the dog, whereas in the rat, a slightly higher proportion was eliminated in the urine and a lower proportion was eliminated in the bile (Table 1). Together with the small amounts of radioactivity recovered in the carcasses (rats only) and cage washings and cage debris (dogs only), total recoveries of radioactivity were ≥95% of the dose after both routes of administration. The majority of the dose was recovered in the excreta and bile within the first 24 h.
Metabolite Profiles.
Gemifloxacin and its phase I metabolites were resistant to collision-induced dissociation and, therefore, mass spectrometry gave very weak MS-MS spectra. Although the spectra obtained were characteristic for the individual components, it was generally not possible to unequivocally postulate the structures for the fragment ions produced. Therefore, assignment of the structures was based on molecular weight data and a comparison of the chromatographic retention times and daughter ion fragmentation patterns of the metabolites with those of the available reference compounds (gemifloxacin,N-acetyl gemifloxacin, O-desmethyl gemifloxacin, and hydroxymethyl gemifloxacin). MS fragments and molecular ions (M + H+) of gemifloxacin and its metabolites are given in Table 3. The metabolic pathways are shown in Fig. 4. Reference standards were available for the N-acetyl, the hydroxymethyl, and theO-desmethyl metabolites, but none were available for gemifloxacin acyl glucuronide, the carboxylic acid derivative, and gemifloxacin glutamate. A reference standard for theE-isomer was also available, but because of rapid and extensive conversion of the compound to gemifloxacin, two components were observed for the standard on HPLC, one corresponding to gemifloxacin and the other minor component to the E-isomer. The minor component had identical chromatographic properties to the component in test samples assigned as the E-isomer.1H NMR data of the dog bile isolate were consistent with the presence of a glucuronide, with a slightly downfield resonance at 5.8 ppm, corresponding to the anomeric proton of an acyl type glucuronide.1H/13C heteronuclear spectra unequivocally confirmed this structure with a clear correlation from the glucuronide anomeric proton to the carbon of the carboxylic acid moiety. Comparison of HPLC retention times and mass spectral data confirmed the presence of the glucuronide in rat and other dog samples. For the carboxylic acid derivative, the molecular ion (m/z 405) and subsequent loss of carbon dioxide (loss of m/z 44–361) by HPLC/MS/MS confirmed its structure. The identity of gemifloxacin glutamate was based on molecular weight data only and, therefore, is regarded as tentative. Analyses of spiked control samples indicated that apart from a small amount (generally 2%) of isomerization of gemifloxacin to form theE-isomer, the compound was stable under experimental conditions used to determine the radiometabolite patterns.
Rat.
Gemifloxacin accounted for 54 to 69% of the radioactivity in the plasma at the time points examined (Table4). Other components detected were theE-isomer of gemifloxacin (7–12%), acyl glucuronide (3–5%), N-acetyl (2–3%), and hydroxymethyl gemifloxacin (1–5%). The radiometabolite patterns observed in the excreta and bile were consistent with those in plasma. Gemifloxacin accounted for totals of ca. 68 and 49% of the dose in the excreta and bile following oral and intravenous administration, respectively (Table5). The acyl glucuronide, theN-acetyl, and the E-isomer were the principal metabolites observed, and each component accounted for a total of 2 to 4% of the dose after oral administration and 4 to 6% of the dose after intravenous administration. Hydroxymethyl gemifloxacin and the glutamate conjugate of gemifloxacin each accounted for ≤0.3% of the oral dose and ca. 1% of the intravenous dose. Several other minor components were detected, including the carboxylic acid derivative and several conjugates (ether and acyl glucuronides, glutamate, glucoside, and glycine) of gemifloxacin and/or its phase I metabolites, each of which accounted for a total of no more than 1% of the dose in the excreta and bile after both routes of administration. There was no evidence of any saturation of the metabolism since similar data were obtained at a dose level approximately 8 times lower (20 mg of fb/kg; Boyle et al., unpublished data) than that used in the experiments described in this article (168 mg of fb/kg).
Dog.
As in the rat, gemifloxacin was the major component detected in dog plasma accounting for 55 to 64% of the radioactivity in the samples (Table 4). The principal metabolites detected were theE-isomer of gemifloxacin (5–9%), acyl glucuronide and glutamate conjugates (each 2–4%), and hydroxymethyl gemifloxacin (<2%). In the excreta and bile, gemifloxacin accounted for totals of ca. 51 and 45% of the dose following oral and intravenous administration, respectively (Table 6). The acyl glucuronide, the E-isomer, the carboxylic acid derivative of gemifloxacin, and O-desmethyl gemifloxacin were the principal metabolites observed, and each component accounted for a total of up to 4% of the dose after oral administration and up to 6% of the dose after intravenous administration. Several other minor components were detected, including hydroxymethyl gemifloxacin, the N-oxide of gemifloxacin, and several conjugates (ether, acyl and carbamyl glucuronides, glutamate, and glucoside) of gemifloxacin and/or its phase I metabolites, each of which accounted for a total of no more than 1.5% of the dose in the excreta and bile after both routes of administration.
Discussion
The studies reviewed in this article describe the disposition of gemifloxacin in the animal species used in its toxicological evaluation. In both rat and dog, drug-related material was rapidly absorbed following oral administration of its mesylate salt form, although gastrointestinal secretion of absorbed material confounded the measurement of the actual proportion of the dose absorbed.
Gemifloxacin is a racemic compound with its individual (+) and (−) enantiomers having equivalent antibacterial activities (S. F. Rittenhouse, B. Donald, and K. Coleman, unpublished data). In studies carried out using nonradiolabeled gemifloxacin mesylate, the plasma profiles and the pharmacokinetic parameters obtained for the individual enantiomers in each species were virtually identical after both oral and intravenous administration, indicating that the compound does not likely undergo stereospecific metabolism, although no formal studies have been carried out to prove this. The blood clearance values for the enantiomers were equivalent to only approximately one quarter of the hepatic blood flow in each species (4.8 and 1.9 l/h/kg in the rat and dog, respectively; Davies and Morris, 1993; Houston, 1994), indicating that they were not subject to extensive hepatic extraction. The large volume of distribution was consistent with a distribution in excess of total body water in both species. Terminal phase elimination half-lives were relatively short in the rat (ca. 2 h) but somewhat longer in the dog (ca. 5 h). The pharmacokinetic data for the enantiomers in these experiments are consistent with those reported previously for the racemate by Seo et al. (1996).
In the rat, consistent with the high volume of distribution of the parent drug, gemifloxacin-related material was shown to be widely distributed into tissues following oral administration of the radiolabeled compound and rapidly eliminated. There was no accumulation of material in any particular organs or tissues. The relatively high concentration of radioactivity observed in the thyroid at 0.5 h after dosing (150 ugEq/g) was considered to be an artifact since a subsequent study carried out following intravenous administration of the compound (M. C. Chalker and B. Whitby, unpublished data) did not indicate that gemifloxacin-related material had any particular affinity for this organ.
In both species, gemifloxacin was metabolized to only a limited extent; consequently, the compound accounted for the majority of the drug-related material in the systemic circulation after oral administration, and it was the major component detected in the excreta after both routes of administration. All metabolites formed were relatively minor, but the principal ones were the acyl glucuronide and the E-isomer of gemifloxacin in both species andN-acetyl gemifloxacin in the rat. The dog is known to be a poor acetylator (Timbrell, 1991) and, consequently, N-acetyl gemifloxacin was not observed in this species. Several other very minor components were detected in the two species, including hydroxymethyl gemifloxacin, O-desmethyl gemifloxacin (dog only), a carboxylic acid derivative, N-oxide (dog only), and several conjugates (ether and acyl glucuronides, glutamate, glucoside, and glycine) of gemifloxacin and/or its phase I metabolites.
Apart from the formation of the E-isomer, the primary metabolic routes for gemifloxacin were similar to those reported previously for quinolone antibiotics. Conjugation of the carboxylic acid group is a common metabolic pathway for these drugs (Outman and Nightingale, 1989; Sorgel, 1989), whereas N-acetylation in the rat occurs with compounds containing a primary amino group attached to a pyrrolidinyl ring, as in trovafloxacin (Dalvie et al., 1996), tosufloxacin (Dauphin et al., 1993), and BMY 43748 (Dauphin et al., 1993). Since the E-isomer was also detected in control samples spiked with [14C]gemifloxacin mesylate, some of this compound detected in test samples likely resulted from direct isomerization of gemifloxacin. However, the E-isomer, in part, may also have been formed metabolically, since interconversion of oxime isomers has been observed previously for roxithromycin (Zhong et al., 2000) and O-methylbenzophenone (Raman and Gorrod, 1991). Of the minor routes of metabolism observed for gemifloxacin, the formation of the hydroxymethyl is relatively novel and probably occurs via oxidative deamination to form an aldehyde and subsequent reduction by alcohol dehydrogenase (Testa, 1995). The formation of a similar hydroxylated metabolite has previously been reported for tosufloxacin and BMY 43748 (Dauphin et al., 1993).
The major route of elimination of gemifloxacin-related material following oral administration was via the feces. Urinary excretion and biliary secretion were minor routes. However, following intravenous administration approximately one third of the administered dose was excreted in the feces of bile duct-cannulated animals demonstrating the propensity for gastrointestinal secretion to occur. It is likely, therefore, that some of the orally administered dose recovered in the feces of bile duct-cannulated animals represents absorbed material.
Acknowledgments
We gratefully acknowledge the help of James Mack for the synthesis of the radiolabeled compound; Andrew Wells, David Ross, and Chung Kim for supplying nonradiolabeled gemifloxacin mesylate and reference compounds; David Kenworthy, David Prohaska, Sonia Alix, and Steven Porteous for determination of excretion profiles and radiometabolite patterns; Anthony Beck for NMR; and Andrew Hackett and Terry Forrest for the determination of tissue distribution in the rat.
Footnotes
-
Send reprint requests to: Mr. Jay V. Ramji, Drug Metabolism and Pharmacokinetics, SmithKline Beecham Pharmaceuticals, The Frythe, Welwyn, Hertfordshire, AL6 9AR, UK. E-mail:Jayant_Ramji{at}sbphrd.com
- Abbreviations used are::
- SB
- SmithKline Beecham
- HPLC
- high-pressure liquid chromatography
- LSC
- liquid scintillation counting
- MS
- mass spectrometry
- fb
- free base
- G.I. tract
- gastrointestinal tract
- AUC(0–t) area under the plasma concentration-time curve to the last quantifiable data point
- AUC(t–inf), area under the plasma concentration-time curve from the last quantifiable data point to infinity
- AUC(0–inf)
- area under the plasma concentration-time curve to infinity
- AUMC(0–inf)
- area under the moment curve to infinity
- Vss
- volume of distribution at steady state
- Received September 22, 2000.
- Accepted December 20, 2000.
- The American Society for Pharmacology and Experimental Therapeutics