- Email: [email protected]
pH-dependent geometric isomerization of roxithromycin in simulated gastrointestinal fluids and in rats
pH-Dependent Geometric Isomerization of Roxithromycin in Simulated Gastrointestinal Fluids and in Rats SHUQIU ZHANG, JIE XING, DAFANG ZHONG Laboratory of Drug Metabolism and Pharmacokinetics, Shenyang Pharmaceutical University, Wenhua Road 103, Shenyang 110015, People’s Republic of China
Received 30 June 2003; revised 6 November 2003; accepted 13 November 2003
ABSTRACT: The biotransformation of roxithromycin in simulated gastrointestinal fluids at 378C and in rats was investigated by using liquid chromatography–tandem mass spectrometry. Roxithromycin degraded to its Z-isomer and decladinose derivative in simulated gastrointestinal fluids in vitro at pH 3, and followed pseudo first-order degradation with a rate constant (SD, standard derivation) of 0.1066 min1 (0.0014) at pH 1.0, 0.0994 min1 (0.0031) at pH 1.2, 0.0400 min1 (0.0003) at pH 1.3, 0.0136 min1 (0.0008) at pH 1.8, and 0.0022 min1 (0.0002) at pH 3.0, respectively. The ratio of Z-roxithromycin to roxithromycin (SD) was 0.21 (0.01) at pH 1.0, 0.19 (0.03) at pH 1.2, 0.18 (0.01) at pH 1.3, 0.15 (0.01) at pH 1.8, and 0.08 (0.02) at pH 3.0, respectively. Pepsin and NaCl added to gastric fluid had no effect on the transformation of roxithromycin. Roxithromycin underwent four metabolic routes such as geometric isomerization, demethylation, dealkylation, and hydrolysis of cladinose in rats after oral administration. The geometric isomerization in rats was neither observed after an intravenous dose, nor after an oral dose with Na2CO3 alkalization. The geometric isomerization between roxithromycin and its Z-isomer took place in gastric fluid both in vitro and in vivo. It was interconvertible and pH-dependent. The isomerization of roxithromycin to its Z-isomer was less than that of Z- to E-configuration both in vitro and in vivo. ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 93:1300– 1309, 2004
Keywords: roxithromycin; geometric isomerization; mass spectrometry; liquid chromatography; metabolism; degradation
INTRODUCTION Roxithromycin, an ether-oxime derivative of erythromycin, showed greater antibacterial potency and longer duration of action than erythromycin. It has a wide antibacterial spectrum for upper and lower respiratory infections, skin and soft-tissue infections, and urogenital and orodental infections.1,2 This might be attributed in part
Correspondence to: Dafang Zhong (Telephone: 86-2423902539; Fax: 86-24-23902539; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 93, 1300–1309 (2004) ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association
1300
to the fact that the presence of the oxime ether side chain at position 9 conferred to roxithromycin better acid stability and less potency to form inhibitory cytochrome P450 Fe (II)-metabolite complex than erythromycin.3 In earlier studies in humans and animals, four metabolites of roxithromycin were identified. They were N-mono- and N-di-demethylated derivatives, decladinose derivative, and erythromycin oxime.4–6 Our previous study using liquid chromatography– tandem mass spectrometry (LC-MSn), showed that roxithromycin mainly underwent four metabolic pathways in humans after oral administration, including geometric isomerization of roxithromycin (from E- to Z-configuration of 9-oxime ether
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
side chain), demethylation, dealkylation of oxime ether side chain, and hydrolysis of cladinose.7 Among them, the geometric isomerization was a novel biotransformation pathway. The metabolism study of roxithromycin in dogs showed that the geometric isomerization of roxithromycin occurred after oral administration, but not after intravenous injection.8 The isomerized derivatives of roxithromycin were detected neither in in vitro incubations with rat liver microsomes9 nor in the isolated perfused rat liver.10 Also, it was reported that the isomerization of roxithromycin in golden hamsters exhibited less extent than in humans and in dogs.11 This seemed to be related to the gastrointestinal environment. To further understand the mechanism of geometric isomerization of roxithromycin and to assess the role of gastrointestinal environment for the metabolism of roxithromycin, investigations of roxithromycin biotransformation in simulated gastrointestinal fluids in vitro and in rats in vivo were conducted.
EXPERIMENTAL Chemicals Roxithromycin and erythromycin oxime reference substances were generous gifts from the Huatai Drug Research Institute (Shenyang, China). Decladinose roxithromycin and Z-roxithromycin were synthesized by the Huatai Drug Research Institute and purified in our laboratory. Ndemethyl roxithromycin (RU44981) was obtained from Hoechst-Marion-Roussel (Romainville Cedex, France). O-demethyl roxithromycin was synthesized at the Laboratory of Pharmaceutical Chemistry, Shenyang Pharmaceutical University (Shenyang, China). All reference substances were structurally confirmed with 1H NMR spectra and mass spectra.7 Methanol and acetonitrile were of high-performance liquid chromatography (HPLC) grade (Yuwang Co., Shandong, China). All other chemicals were of analytical grade. Simulated gastric fluids (pH 1.2, consisting of HCl, and NaCl or pepsin) were prepared according to USP XXIII. The hydrochloric acid solutions of pH 1.0, 1.3, and 1.8, and phosphate buffers (0.1 M) of pH 3.0, 5.0, and 7.4, were used as simulated gastrointestinal fluids to assess the pH effect on the isomerization of roxithromycin. The pH values of all simulated gastrointestinal fluids were determined and calibrated by using a pH meter.
1301
Roxithromycin Degradation in Simulated Gastric Fluid (pH 1.2) Roxithromycin at the concentration of 5 mM was incubated at 378C in the simulated gastric fluid of pH 1.2 for 2 h. Either pepsin or NaCl, as a component of simulated gastric fluid, was added to other sets of the experiment to examine their effects on roxithromycin degradation. Aliquots of 200 mL were taken at predetermined times of 10, 20, 30, 45, 60, and 120 min, and mixed with 300 mL of 0.1 M Na2CO3 to terminate the reactions. Each set of experiments was conducted in triplicate. To further understand the mechanism of roxithromycin isomerization, the digestion of Zroxithromycin in simulated gastric fluid was also conducted as described above. pH Effect on the Transformation of Roxithromycin In Vitro A series of hydrochloride acid solutions (pH 1.0, 1.3, and 1.8) and phosphate buffers of (pH 3.0, 5.0, and 7.4), was used as simulated gastrointestinal fluids to assess the pH effect on the transformation of roxithromycin by incubation with 5 mM roxithromycin in vitro at 378C. Samples were taken at 10, 20, 30, 60, 120, 240, and 360 min. Additional samples were taken at 5 min for incubations in simulated gastric fluid of pH 1.0. The reactions were terminated by addition of 1.5fold of 0.1 M Na2CO3 to samples. Each set of experiments was conducted in triplicate. The stability of roxithromycin and Z-roxithromycin in termination medium was investigated by incubating in the mixture of 2 mL of simulated gastric fluid (pH 1.0) and 3 mL of 0.1 M Na2CO3. Animal Experiments Eighteen male Wistar rats, 250 20 g, were supplied by the Lab Animal Center of Shenyang Pharmaceutical University (grade II, certificate no. 042). Animals were fed a normal standard diet ad libitum and acclimatized at a 12-h light cycle for at least 5 days before being used. After being fasted for 12 h, animals were divided randomly into six groups, three in each, and were implanted with a PE-10 cannula into the common bile duct under anesthesia by ethyl ether. After being recovered for 2 h, animals of two groups were orally administered 20 mg/kg of either roxithromycin or Z-roxithromycin [suspended in 3% (w/v) CMC-Na solution at a concentration of 5 g/L]. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
1302
ZHANG, XING, AND ZHONG
Animals of another two groups were given by gavage 2 mL of 0.5 M Na2CO3 followed by the same oral dose of either roxithromycin or Zroxithromycin as described above at 2 min. Animals of the last two groups were given 3.5 mg/kg of either roxithromycin or Z-roxithromycin by intravenous injection (dissolved in 20% ethanol solution at a concentration of 2.5 g/L). Bile was collected for 24 h.
sheath gas at a flow rate of 0.75 L/min and auxiliary gas at a flow rate of 0.15 L/min. MSn spectra were obtained in mass range from m/z 100 to 1400, in positive ion mode, by collision induced dissociation using helium as collision gas. Data were analyzed by Xcalibur software (version 1.2; Finnigan). Metabolites were identified by composite comparisons of their LC behaviors and MSn spectra with those of synthesized references.
Sample Pretreatment
Analysis Calibration
The samples (0.2 mL) from in vitro incubations were mixed with 0.3 mL of 0.1 M Na2CO3, and followed by addition of 0.1 mL of clarithromycin (4 mg/L) as internal standard. After adding 3 mL of ethyl ether, the mixture was vortexed for 2 min, and centrifuged at 3000g for 5 min. The organic layer was taken and evaporated under a stream of nitrogen at an ambient temperature. The residues were reconstituted with 1.0 mL of mobile phase, and an aliquot of 10 mL was injected onto the liquid chromatograph for LC-MSn analysis. The bile sample (0.5 mL) was diluted with 0.5 mL of water, filtered through precut membranes (0.45 mm), and then applied to a preconditioned 1.5-mL Sep-Park C18 cartridge (J.T. Baker, Phillipsburg, NJ). The cartridge was washed with 2.0 mL of water and eluted with 2.0 mL of methanol. A 20-mL aliquot of the eluent was injected onto the liquid chromatograph for LCMSn analysis.
For in vitro incubation experiments, calibration curves were established at a range of 0.03–5.0 mM for roxithromycin and Z-roxithromycin based on the ratio of peak area (roxithromycin/IS) versus concentration by using weighted (w ¼ 1/c) linear least-squares regression analysis. The lower limit of quantification for roxithromycin and Zroxithromycin was 0.03 mM. The analytical intraand interassay coefficients of variation were below 11%.
LC-MSn Analysis The samples were analyzed by an HPLC system (Shimadzu Corp., Kyoto, Japan) equipped with a Finnigan LCQ ion trap mass spectrometer (San Jose, CA) via an electrospray ionization interface. A Kromasil ODS column (particle size 5 mm, 20 cm 4.6 mm inside diameter; Hi-Tech Scientific Instrument Corp., Tianjin, China) was used for sample separation. A mobile phase consisting of acetonitrile-methanol-10 mM ammonium acetate (50:10:35, v/v/v) was used at a flow rate of 0.4 mL/min for the analysis of in vitro incubation samples. Identification of roxithromycin metabolites in rat bile was performed by using another mobile phase system consisting of acetonitrilemethanol-10 mM ammonium acetate (43:10:47, v/ v/v) at a flow rate of 0.3 mL/min. The ionization was realized by applying spray voltage of 4.25 kV, capillary temperature of 1808C, and capillary voltage of 30 V. Nitrogen was used as both the JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
RESULTS Roxithromycin Transformation in Simulated Gastric Fluid (pH 1.2) In Vitro Roxithromycin was detected at a retention time of 12 min, with [M þ H]þ ion at m/z 837 (Fig. 1). It showed MS/MS fragment ions at m/z (relative abundance, %) 716, 679 (100), 558, 540, and 522, characteristic losses of oxime alkylether side chain (121 u), cladinose moiety (158 u), water (18 u), and desosamine (157 u) (Fig. 2). The
Figure 1. Representative chromatograms and MS/ MS spectra of roxithromycin and its metabolites in simulated gastric fluid (pH 1.2) after a 10-min incubation at 378C.
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
Figure 2. Proposed MS/MS fragmentation pathways of roxithromycin.
fragmentation profile was in accordance with that of the previously reported for roxithromycin.12 Two metabolites M1 and M2 were found in the samples incubated with simulated gastric fluid at 378C (Fig. 1). M1 was detected at LC retention time of 9 min, with the same [M þ H]þ and MS/MS fragment ions as those of roxithromycin. M1 was identified as Z-isomer of roxithromycin by comparison with the synthesized Z-roxithromycin reference. M2, with LC retention time of 8 min, showed quasi-molecular ion [M þ H]þ at m/z 679, which is 158 u less than the parent drug. This is a characteristic for the loss of cladinose moiety. Its MS/MS spectra gave fragment ions at m/z 661, 558, 540, and 522 (100), which resulted from loss of water, oxime side chain, and desosamine. M2 was identified as decladinose roxithromycin by comparison with synthesized reference. A minor metabolite, with the same [M þ H]þ and MS/MS fragmentation ions as M2, measured at a retention time of 7 min, was considered as the Z-isomer of M2. The degradation of roxithromycin in simulated gastric fluids followed pseudo first-order kinetics. The semilogarithmic curves of roxithromycin concentration versus time are shown in Figure 3. The rate constants of roxithromycin degradation were 0.0994 min1 (0.0031) in simulated gastric fluid, 0.0995 min1 (0.0030) in simulated gastric fluid containing NaCl, and 0.0993 min1 (0.0033) in simulated gastric fluid containing pepsin, respectively. The geometric isomer of roxithromycin was detected immediately after mixing of roxithromycin with gastric fluid. The ratio of Z-roxithromycin to the parent drug (SD) was constant at 0.19 (0.03) in gastric fluid, 0.18 (0.03) in gastric fluid containing NaCl, and 0.18 (0.03) in gastric fluid containing pepsin, respectively, during a 10–
1303
Figure 3. The concentration of roxithromycin in simulated gastric fluids (pH 1.2) versus time of incubation at 378C.
45-min incubation period. The level of roxithromycin was below the limit of quantification after 45 min of incubation. The level of decladinose roxithromycin was increased with time. To gain additional insight into the isomerization of roxithromycin, the incubation of Z-roxithromycin was also conducted in the simulated gastric fluid at 378C as described above. Zroxithromycin was transformed to roxithromycin and decladinose derivative in simulated gastric fluid. The isomerization of Z- to E-configuration was with greater extent than that of E- to Zconfiguration. The mean ratio of Z-roxithromycin to its E-isomer for 10–45 min of incubation (SD) was 0.16 (0.01) in the gastric fluid, 0.17 (0.01) in the gastric fluid containing NaCl, and 0.16 (0.01) in the gastric fluid containing pepsin, respectively. The degradation scheme of roxithromycin in the simulated gastric fluid at 378C is shown in Figure 4. pH Effect on the In Vitro Geometric Isomerization of Roxithromycin When either roxithromycin or Z-roxithromycin was incubated with the mixture (about pH 10) of 2 mL of pH 1.0 gastric fluid and 3 mL of 0.1 M Na2CO3, neither geometric isomer nor degradation derivative were detected. The 0.1 M Na2CO3 was used by 1.5-fold of sample volume to terminate the reaction. The degradation of roxithromycin was performed at 378C in the simulated gastrointestinal fluid of pH 1.0, 1.3, 1.8, 3.0, 5.0, and 7.4. Except for Z-isomer and decladinose derivatives of roxithroJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
1304
ZHANG, XING, AND ZHONG
Roxithromycin Metabolism in Rats after an Oral Dose
Figure 4. The proposed degradation of roxithromycin in the simulated gastric fluid at 378C.
mycin, no other degradation derivative was detected in simulated gastrointestinal fluids of pH 1.0–3.0. No degraded derivative of roxithromycin was detected in simulated gastrointestinal fluid of pH 5.0 and 7.4 during incubation. The decomposition rate of roxithromycin decreased with the increase in pH value from 1.0 to 3.0. The rate constant (SD) of roxithromycin degradation was 0.1066 min1 (0.0014) at pH 1.0, 0.0400 min1 (0.0003) at pH 1.3, 0.0136 min1 (0.0008) at pH 1.8, and 0.0022 min1 (0.0002) at pH 3.0, respectively. The extent of isomerization of roxithromycin to Z-isomer decreased with the increase of pH value. The mean ratio (SD) of Z-roxithromycin to roxithromycin was constant at 0.21 (0.01) at pH 1.0 during the sampling interval (5–30 min), 0.18 (0.01) at pH 1.3 during the sampling interval (10–60 min), 0.15 (0.01) at pH 1.8 during the sampling interval (10–240 min), and 0.08 (0.02) at pH 3.0 during the sampling interval (45–360 min), respectively. The degradation rate constants of roxithromycin and the ratios of Z-roxithromycin to roxithromycin at different pH’s are summarized in Table 1.
A number of roxithromycin-related derivatives, with quasi-molecular ion [M þ H]þ at m/z of 837, 823, 809, 749, 735, and 679, respectively, were detected in the rat bile after an oral dose of roxithromycin of 20 mg/kg. Their TIC and MS/MS full scan chromatograms are shown in Figure 5A. Their structures were confirmed or suggested based on HPLC behaviors and multiple-step MS data with reference substances. Roxithromycin and its Z-isomer M1 were detected in m/z 837 channel at retention times of 28 and 18 min, and were confirmed with reference substances, respectively. The ratio of Z-isomer to roxithromycin was about 0.3 at maximum in the bile collected during 2–4 h after administration of roxithromycin. Decladinose roxithromycin (M2) was detected at a retention time of 15 min. And another metabolite, with the same [M þ H]þ and MS/MS fragment ions as those of M2, was detected at a retention time of 12 min. It was proposed to be the Z-isomer of M2. Metabolites M3 and M4 were measured at m/z 823 channel at retention times of 21 and 15 min, respectively, with the same [M þ H]þ at m/z 823, which is 14 u less than the parent drug, a typical characteristic loss of methyl group. M3 showed MS/MS fragments of m/z 665 (100), 544, 522, and 508, resulting from the decomposition of cladinose (158 u), oxime alkylether side chain (121 u), demethyl desosamine (143 u), and two molecules of water (36 u). M3 was confirmed to be N-demethylated roxithromycin by comparison with reference standard. M4 gave MS/MS fragment ions at m/z 665 (100), 558, 540, 522, and 508, suggesting the losses of cladinose moiety, O-demethylated oxime alkylether side chain (107 u), two molecules of water, and desosamine, respectively. M4 was confirmed to be O-demethy-
Table 1. The Degradation Rate Constants and Extents of Isomerization of Roxithromycin in Simulated Gastric Fluids of Different pH Values at 378C pH Value
k (min1)a (SD) Z/Eb (SD)
1.0
1.2
1.3
1.8
3.0
0.1066 (0.0014) 0.21 (0.01)
0.0994 (0.0031) 0.19 (0.03)
0.0400 (0.0003) 0.18 (0.01)
0.0136 (0.0008) 0.15 (0.01)
0.0022 (0.0002) 0.08 (0.02)
a
First-order degradation rate constant. Ratio of Z-roxithromycin to roxithromycin.
b
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
1305
Figure 5. Total ion current and MS/MS full scan chromatograms of roxithromycin and its metabolites in rat bile collected during 2–4 h after an oral dose of roxithromycin (20 mg/kg) (A), or an oral dose of roxithromycin (20 mg/kg) with alkalization of Na2CO3 (B), or an intravenous dose of roxithromycin (3.5 mg/kg) (C).
lated roxithromycin by comparison with synthesized reference. Metabolites M5 and M6 were detected at m/z 809 with retention times of 19 and 12 min, respectively. Their same quasi-molecular ion [M þ H]þ at m/z 809, 28 u lower than roxithromycin, suggested they were di-demethylated metabolites of roxithromycin. M5 gave MS/MS fragment ions at m/z 651 (100), 530, 522, and 512, characteristic losses of cladinose, and then oxime alkylether side chain, di-demethyl desosamine (129 u), and water. M5 was tentatively assigned as N,N-di-demethylated roxithromycin based on its MSn data and by comparison with N-demethyl roxithromycin. M6 showed MS/MS fragment ions at m/z 651 (100), 544, 526, and 508, indicating the losses of cladinose, and then demethylated oxime alkylether side chain, water, and demethylated desosamine. M6 was proposed to be N,O-didemethyl roxithromycin. Metabolite M7, with an LC retention time of 14 min, showed [M þ H]þ at m/z 749, 88 u lower than that of roxithromycin, which is characteristic loss of alkylether side chain. Its MS/MS fragment ions were detected at m/z 591 (100), 558, 540, and 434, suggesting the decomposition of cladinose, oxime group (33 u), water, and desosamine. M7 was confirmed as erythromycin oxime by the comparison of LC behavior and MSn data with
reference standard. Another metabolite, with the same MSn spectra as M7 but higher level than M7, measured at retention time of 12 min, was considered as Z-isomer of M7. Metabolite M8 showed an LC retention time of 12 min, and [M þ H]þ at m/z 735 (14 u less than M7). Its MS/MS fragment ions at m/z 577 (100), 544, 526, and 434, were 14 u lower than the corresponding fragment ions of M7. It was tentatively suggested to be N-demethyl erythromycin oxime based on its mass data. Its Z-isomer, showing the same [M þ H] and MS/MS fragment ions as M8, was also detected at a retention time of 10 min. A similar metabolic profile was observed for Z-roxithromycin in rats after oral administration of Z-roxithromycin (Fig. 6A). Z-roxithromycin was more easily transformed to its E-isomer (roxithromycin) than E- to Z-configuration. The level of roxithromycin was higher than that of Z-isomer in the bile collected after 2 h of administration of Z-roxithromycin. The Metabolism of Roxithromycin after an Oral Dose with Na2CO3 When roxithromycin was orally administrated to rats with Na2CO3, demethylated metabolites (M3, M4, M5, and M6) and dealkylated metabolites JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
1306
ZHANG, XING, AND ZHONG
Figure 6. Total ion current and MS/MS full scan chromatograms of Z-roxithromycin and its metabolites in rat bile collected during 2–4 h after an oral dose of Z-roxithromycin (20 mg/kg) (A), or an oral dose of Z-roxithromycin (20 mg/kg) with alkalization of Na2CO3 (B), or an intravenous dose of Z-roxithromycin (3.5 mg/kg) (C).
(M7 and M8) were also detected in rat bile, but decladinose derivative (M1), and Z-isomers of parent drug and metabolites in the rat bile collected during 0–24 h after drug administration were not detected (Fig. 5B). When Z-roxithromycin was orally administrated to rats with Na2CO3 (Fig. 6B), no E-isomers of parent drug and metabolites were detected in the bile collected during 0–8 h after dosing, whereas roxithromycin was detected after 8 h of administration. The ratio of roxithromycin to its Z-isomer reached about 0.1 in the bile collected during the 8–10 h interval, and about 0.3 during the 12–24 h interval, after drug administration. The Z-isomers of metabolites M3 and M4 were detected at retention times of 15 and 11 min, respectively. The Z-isomers of M5 and M6 showed retention times of 14 and 10 min, respectively. Finally, the Z-isomers of M7 and M8 were detected at retention times of 12 and 10 min, respectively. The Metabolism of Roxithromycin after an Intravenous Dose When roxithromycin was intravenously injected to rats (Fig. 5C), neither Z-isomer nor decladinose derivative of roxithromycin was detected in the 0–24 h bile after administration. Other JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
metabolites including demethylated and dealkylated derivatives of E-configuration were detected. After an intravenous dose of Z-roxithromycin, only Z-isomers of demethylated metabolites and dealkylated metabolites were measured in the rat 0–24 h bile, but decladinose derivative and roxithromycin as well as E-isomers of metabolites were not detected (Fig. 6C). The metabolic profile of roxithromycin in rats is shown in Figure 7.
DISCUSSION Roxithromycin and its metabolites followed common fragmentation pathways in electrospray ionization MS/MS full scan, which was consistent with previous results.12 Z-isomers showed the same MSn spectra as those of the corresponding E-isomers, but their retention times under the reversed phase HPLC conditions were shorter than the corresponding E-isomers. In the simulated gastric fluid at 378C, roxithromycin transformed to its Z-isomer and decladinose derivative. The geometric isomerization in the gastric acid condition had not been reported, which
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
1307
Figure 7. The proposed metabolic pathways of roxithromycin in rats after an oral dose of 20 mg/kg of roxithromycin.
might be partly due to the lack of sensitivity and/or specificity of the previous assay methods. Roxithromycin degraded in simulated gastric fluids at 378C according to first-order kinetics with a half-life of 6.5 min at pH 1.0, 17 min at pH 1.3, 51 min at pH 1.8, and 315 min at pH 3.0, respectively. No degradation occurred at pH 5.0 and 7.4. Neither NaCl nor pepsin altered the degradation of roxithromycin. These results were generally consistent with a previous study for roxithromycin disintegration by HPLC with coulometric detection.13 The isomerization occurred immediately after mixing of either roxithromycin or its Z-isomer with gastric fluids. Actually, the ratios of Z-roxithromycin to roxithromycin were constant during sampling duration. The extent of isomerization of E- to Z-configuration decreased with the increase of pH value from 1.0 to 3.0. This suggested the pH dependence of roxithromycin isomerization. It was reported that the roxithromycin oxime was transformed to its Z-configuration in strong alkaline medium (1.2 mL of 2 M NaOH and 2.7 mL of methanol).3 In the present study, no isomerization of roxithromycin was observed at pH 5, 7.4, and 10 (in the mixture of 2 mL of pH 1.0 simulated gastric fluid and 3 mL of 0.1 M Na2CO3). It is interesting to note that the isomerization between E- and Z-configurations was interconvertible, but the extent of isomerization of Z- to E-configuration was much greater than that of E- to Z-configuration. That might be related to the differences in stereo characteristics between two configurations of roxithromycin. Gharbi-Benarous et al.14,15 reported that the oxime side chain in the roxi-
thromycin molecule was oriented above the macrocyclic lactone ring and the oxygen atoms [O (17) and O (19)] of the chain were engaged in tight hydrogen bonding with 6- and 11-hydroxyl groups of the macrocycle, based on 1H and 13C NMR data. This resulted in a globular structure for roxithromycin with less freedom for the macrocyclic lactone ring and sugar unit. We had compared 1H NMR spectra of roxithromycin and Z-roxithromycin (in CDCl3). No obvious signal was observed in Z-roxithromycin spectra at the corresponding chemical shift positions of 6- and 11-hydroxyl groups (2.48 and 4.38 ppm, respectively) as in roxithromycin spectra. This suggested that the intramolecular hydrogen bonding might be weak or disappeared in Z-configuration. Therefore, the E-configuration should be the preferred configuration to Z. Roxithromycin underwent four metabolic pathways such as geometric isomerization, hydrolysis of cladinose, demethylation, and dealkylation in rats after oral administration. Z-roxithromycin followed a similar metabolic profile to parent drug, but its isomerization was found to be faster than roxithromycin. No isomerized derivatives and decladinose derivative were detected during 0–24 h after oral administration of roxithromycin with Na2CO3 alkalization or after intravenous injection of roxithromycin. When Z-roxithromycin was given orally with alkalization, some E-isomerized derivatives were detected after 8 h of administration, which probably resulted from the accumulated secretion of gastric acid and greater liability of isomerization from Z- to E-configuration. The JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
1308
ZHANG, XING, AND ZHONG
above results suggested that the geometric isomerization of roxithromycin should take place in the stomach of animals, and it was dependent on gastric acid. It is also collateral evidence to support the previous studies that no isomerized derivatives were detected in incubations with rat liver microsomes9 and in perfused rat liver10 by mass spectrometry. The results also proposed that the metabolites of Z-configuration, detected after oral administration of roxithromycin, should be the metabolites of Z-roxithromycin, but were not the consequence of transformation from the metabolites of E-configuration. The levels of roxithromycin and its metabolites were normally higher than those of the Z-isomers after oral administration of roxithromycin, but the levels of dealkylated metabolites (M7 and M8) were lower than those of their Z-isomers (similar to reported results7), which suggested that the dealkylation of oxime ether side chain of Z-roxithromycin seemed more susceptible to occur than that of roxithromycin. Previous studies reported that the geometric isomerization of roxithromycin after oral administration was predominant in humans and in dogs,7,8 but insignificant in rats and in golden hamsters.11 The species differences in roxithromycin geometric isomerization might be attributable to the differences in the acidity and volume of gastric fluid among different animals. Additionally, it is likely that the effects of administration factors on gastric fluid acidity might be less in bigger animals than in smaller animals. Considering the weaker antibiotic potency of Zroxithromycin than that of roxithromycin,3,16,17 the pH-dependent geometric isomerization during absorption of roxithromycin should have some effects on roxithromycin bioavailability, consequently clinical efficacy. This could provide some inspiration for the design of dosage form and clinical use of roxithromycin. It seems to be beneficial to increase the pH of gastric juice for the roxithromycin absorption, considering both isomerization and decladinose. A study in humans showed that the roxithromycin concentration in the gastric juice and gastric tissue could be significantly increased by combined administration with proton pump inhibitor, which was considered to contribute to the synergic beneficial action in the eradication therapy of Helicobacter pylori.13
CONCLUSIONS Roxithromycin degraded to its Z-isomer and decladinose derivative in simulated gastric fluid, JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
and underwent the following metabolic pathways in rats: geometric isomerization, hydrolysis of cladinose, demethylation, and dealkylation. The geometric isomerization of roxithromycin was pH dependent and reversible both in vitro and in vivo. Z-roxithromycin was more liable to transformation to its E-isomer.
ACKNOWLEDGMENTS This research was supported by the National Natural Science Foundation of China, number 39930180.
REFERENCES 1. Young RA, Gonzalez JP, Sorkin EM. 1989. Roxithromycin. A review of its antibacterial activity. Pharmacokinetic properties and clinical efficacy. Drugs 37:8–41. 2. Markham A, Faulds D. 1994. Roxithromycin: An update of its antimicrobial activity, pharmacokinetic properties and therapeutic use. Drugs 48: 297–326. 3. Gase JC, d’Ambrieres SG, Lutz A, Chantot JF. 1991. New ether oxime derivatives of erythromycin A: A structure-activity relationship study. J Antibiot 44:313–330. 4. Esumi Y, Zin Y, Ban S, Ninomiya S, Hayashi K, Yokoshima T. 1988. Absorption, distribution, metabolism, and excretion of RU 28965 in animals. Chemotherapy 36(Suppl 4):148–163. 5. Koyama M, Tateno M, Shirotsuka M, Yamamoto T, Hirayama M, Saito K, Okui K. 1988. Absorption, metabolism, and excretion of RU 28965 in humans. Chemotherapy 36(Suppl 4):164–183. 6. McLean A, Sutton JA, Salmon J, Chatelet D. 1988. Roxithromycin: Pharmacokinetic and metabolism study in humans. Br J Clin Pract 42(Suppl 55): 52–53. 7. Zhong DF, Li XQ, Wang AM, Xu YJ, Wu SD. 2000. Identification of the metabolites of roxithromycin in humans. Drug Metab Dispos 28:552–559. 8. Zhang SQ, Zhang LF, Xing J, Zhong DF. 2003. The metabolism of roxithromycin in dogs. Yao Xue Xue Bao 38:374–379. 9. Zhong DF, Zhang SQ, Sun L, Zhao XY. 2002. Metabolism of roxithromycin in phenobarbitaltreated rat liver microsomes. Acta Pharmacol Sin 23:455–460. 10. Jarukamjorn K, Thalhammer T, Gollackner B, Pittenauer E, Jaeger W. 1998. Metabolism of roxithromycin in the isolated perfused rat liver. J Pharm Pharmacol 50:515–519.
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
11. Zhang SQ, Zhong DF. 2003. The metabolism of roxithromycin in golden hamsters. Asian J Drug Metab Pharmcokinet 3:195–202. 12. Zhong DF, Tian L, Li XQ, Xu YJ. 2000. Studies on the electrospray ion trap mass spectra of roxithromycin and its metabolites. Chem J Chin Univ 21: 31–35. 13. Kees F, Holstege A, Ittner KP, Zimmermann M, Lock G, Scho¨lmerich J, Grobecker H. 2000. Pharmacokinetic interaction between proton pump inhibitors and roxithromycin in volunteers. Aliment Pharmacol Ther 14:407–412. 14. Gharbi-Benarous J, Delaforge M, Artaud I, Girault J-P. 1990. Analysis of the 1H and 13C NMR spectra of the novel macrolide antibiotic roxithromycin.
1309
Structure and conformation in solution. Magn Reson Chem 28:846–855. 15. Gharbi-Benarous J, Delaforge M, Jankowski CK, Girault J-P. 1991. A comparative NMR study between the macrolide antibiotic roxithromycin and erythromycin A with different biological properties. J Med Chem 34:1117–1125. 16. Li XQ, Zhong DF, Huang HH, Wu SD. 2001. Demethylation metabolism of roxithromycin in humans and rats. Acta Pharmacol Sin 22: 469–474. 17. Huang HH, Yu LH, Sun TM, Li XQ, Zhong DF. 2000. In vitro antibacterial activity of roxithromycin and its major metabolites. Chin J Antibiot 25: 443–446.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
Received 30 June 2003; revised 6 November 2003; accepted 13 November 2003
ABSTRACT: The biotransformation of roxithromycin in simulated gastrointestinal fluids at 378C and in rats was investigated by using liquid chromatography–tandem mass spectrometry. Roxithromycin degraded to its Z-isomer and decladinose derivative in simulated gastrointestinal fluids in vitro at pH 3, and followed pseudo first-order degradation with a rate constant (SD, standard derivation) of 0.1066 min1 (0.0014) at pH 1.0, 0.0994 min1 (0.0031) at pH 1.2, 0.0400 min1 (0.0003) at pH 1.3, 0.0136 min1 (0.0008) at pH 1.8, and 0.0022 min1 (0.0002) at pH 3.0, respectively. The ratio of Z-roxithromycin to roxithromycin (SD) was 0.21 (0.01) at pH 1.0, 0.19 (0.03) at pH 1.2, 0.18 (0.01) at pH 1.3, 0.15 (0.01) at pH 1.8, and 0.08 (0.02) at pH 3.0, respectively. Pepsin and NaCl added to gastric fluid had no effect on the transformation of roxithromycin. Roxithromycin underwent four metabolic routes such as geometric isomerization, demethylation, dealkylation, and hydrolysis of cladinose in rats after oral administration. The geometric isomerization in rats was neither observed after an intravenous dose, nor after an oral dose with Na2CO3 alkalization. The geometric isomerization between roxithromycin and its Z-isomer took place in gastric fluid both in vitro and in vivo. It was interconvertible and pH-dependent. The isomerization of roxithromycin to its Z-isomer was less than that of Z- to E-configuration both in vitro and in vivo. ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 93:1300– 1309, 2004
Keywords: roxithromycin; geometric isomerization; mass spectrometry; liquid chromatography; metabolism; degradation
INTRODUCTION Roxithromycin, an ether-oxime derivative of erythromycin, showed greater antibacterial potency and longer duration of action than erythromycin. It has a wide antibacterial spectrum for upper and lower respiratory infections, skin and soft-tissue infections, and urogenital and orodental infections.1,2 This might be attributed in part
Correspondence to: Dafang Zhong (Telephone: 86-2423902539; Fax: 86-24-23902539; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 93, 1300–1309 (2004) ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association
1300
to the fact that the presence of the oxime ether side chain at position 9 conferred to roxithromycin better acid stability and less potency to form inhibitory cytochrome P450 Fe (II)-metabolite complex than erythromycin.3 In earlier studies in humans and animals, four metabolites of roxithromycin were identified. They were N-mono- and N-di-demethylated derivatives, decladinose derivative, and erythromycin oxime.4–6 Our previous study using liquid chromatography– tandem mass spectrometry (LC-MSn), showed that roxithromycin mainly underwent four metabolic pathways in humans after oral administration, including geometric isomerization of roxithromycin (from E- to Z-configuration of 9-oxime ether
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
side chain), demethylation, dealkylation of oxime ether side chain, and hydrolysis of cladinose.7 Among them, the geometric isomerization was a novel biotransformation pathway. The metabolism study of roxithromycin in dogs showed that the geometric isomerization of roxithromycin occurred after oral administration, but not after intravenous injection.8 The isomerized derivatives of roxithromycin were detected neither in in vitro incubations with rat liver microsomes9 nor in the isolated perfused rat liver.10 Also, it was reported that the isomerization of roxithromycin in golden hamsters exhibited less extent than in humans and in dogs.11 This seemed to be related to the gastrointestinal environment. To further understand the mechanism of geometric isomerization of roxithromycin and to assess the role of gastrointestinal environment for the metabolism of roxithromycin, investigations of roxithromycin biotransformation in simulated gastrointestinal fluids in vitro and in rats in vivo were conducted.
EXPERIMENTAL Chemicals Roxithromycin and erythromycin oxime reference substances were generous gifts from the Huatai Drug Research Institute (Shenyang, China). Decladinose roxithromycin and Z-roxithromycin were synthesized by the Huatai Drug Research Institute and purified in our laboratory. Ndemethyl roxithromycin (RU44981) was obtained from Hoechst-Marion-Roussel (Romainville Cedex, France). O-demethyl roxithromycin was synthesized at the Laboratory of Pharmaceutical Chemistry, Shenyang Pharmaceutical University (Shenyang, China). All reference substances were structurally confirmed with 1H NMR spectra and mass spectra.7 Methanol and acetonitrile were of high-performance liquid chromatography (HPLC) grade (Yuwang Co., Shandong, China). All other chemicals were of analytical grade. Simulated gastric fluids (pH 1.2, consisting of HCl, and NaCl or pepsin) were prepared according to USP XXIII. The hydrochloric acid solutions of pH 1.0, 1.3, and 1.8, and phosphate buffers (0.1 M) of pH 3.0, 5.0, and 7.4, were used as simulated gastrointestinal fluids to assess the pH effect on the isomerization of roxithromycin. The pH values of all simulated gastrointestinal fluids were determined and calibrated by using a pH meter.
1301
Roxithromycin Degradation in Simulated Gastric Fluid (pH 1.2) Roxithromycin at the concentration of 5 mM was incubated at 378C in the simulated gastric fluid of pH 1.2 for 2 h. Either pepsin or NaCl, as a component of simulated gastric fluid, was added to other sets of the experiment to examine their effects on roxithromycin degradation. Aliquots of 200 mL were taken at predetermined times of 10, 20, 30, 45, 60, and 120 min, and mixed with 300 mL of 0.1 M Na2CO3 to terminate the reactions. Each set of experiments was conducted in triplicate. To further understand the mechanism of roxithromycin isomerization, the digestion of Zroxithromycin in simulated gastric fluid was also conducted as described above. pH Effect on the Transformation of Roxithromycin In Vitro A series of hydrochloride acid solutions (pH 1.0, 1.3, and 1.8) and phosphate buffers of (pH 3.0, 5.0, and 7.4), was used as simulated gastrointestinal fluids to assess the pH effect on the transformation of roxithromycin by incubation with 5 mM roxithromycin in vitro at 378C. Samples were taken at 10, 20, 30, 60, 120, 240, and 360 min. Additional samples were taken at 5 min for incubations in simulated gastric fluid of pH 1.0. The reactions were terminated by addition of 1.5fold of 0.1 M Na2CO3 to samples. Each set of experiments was conducted in triplicate. The stability of roxithromycin and Z-roxithromycin in termination medium was investigated by incubating in the mixture of 2 mL of simulated gastric fluid (pH 1.0) and 3 mL of 0.1 M Na2CO3. Animal Experiments Eighteen male Wistar rats, 250 20 g, were supplied by the Lab Animal Center of Shenyang Pharmaceutical University (grade II, certificate no. 042). Animals were fed a normal standard diet ad libitum and acclimatized at a 12-h light cycle for at least 5 days before being used. After being fasted for 12 h, animals were divided randomly into six groups, three in each, and were implanted with a PE-10 cannula into the common bile duct under anesthesia by ethyl ether. After being recovered for 2 h, animals of two groups were orally administered 20 mg/kg of either roxithromycin or Z-roxithromycin [suspended in 3% (w/v) CMC-Na solution at a concentration of 5 g/L]. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
1302
ZHANG, XING, AND ZHONG
Animals of another two groups were given by gavage 2 mL of 0.5 M Na2CO3 followed by the same oral dose of either roxithromycin or Zroxithromycin as described above at 2 min. Animals of the last two groups were given 3.5 mg/kg of either roxithromycin or Z-roxithromycin by intravenous injection (dissolved in 20% ethanol solution at a concentration of 2.5 g/L). Bile was collected for 24 h.
sheath gas at a flow rate of 0.75 L/min and auxiliary gas at a flow rate of 0.15 L/min. MSn spectra were obtained in mass range from m/z 100 to 1400, in positive ion mode, by collision induced dissociation using helium as collision gas. Data were analyzed by Xcalibur software (version 1.2; Finnigan). Metabolites were identified by composite comparisons of their LC behaviors and MSn spectra with those of synthesized references.
Sample Pretreatment
Analysis Calibration
The samples (0.2 mL) from in vitro incubations were mixed with 0.3 mL of 0.1 M Na2CO3, and followed by addition of 0.1 mL of clarithromycin (4 mg/L) as internal standard. After adding 3 mL of ethyl ether, the mixture was vortexed for 2 min, and centrifuged at 3000g for 5 min. The organic layer was taken and evaporated under a stream of nitrogen at an ambient temperature. The residues were reconstituted with 1.0 mL of mobile phase, and an aliquot of 10 mL was injected onto the liquid chromatograph for LC-MSn analysis. The bile sample (0.5 mL) was diluted with 0.5 mL of water, filtered through precut membranes (0.45 mm), and then applied to a preconditioned 1.5-mL Sep-Park C18 cartridge (J.T. Baker, Phillipsburg, NJ). The cartridge was washed with 2.0 mL of water and eluted with 2.0 mL of methanol. A 20-mL aliquot of the eluent was injected onto the liquid chromatograph for LCMSn analysis.
For in vitro incubation experiments, calibration curves were established at a range of 0.03–5.0 mM for roxithromycin and Z-roxithromycin based on the ratio of peak area (roxithromycin/IS) versus concentration by using weighted (w ¼ 1/c) linear least-squares regression analysis. The lower limit of quantification for roxithromycin and Zroxithromycin was 0.03 mM. The analytical intraand interassay coefficients of variation were below 11%.
LC-MSn Analysis The samples were analyzed by an HPLC system (Shimadzu Corp., Kyoto, Japan) equipped with a Finnigan LCQ ion trap mass spectrometer (San Jose, CA) via an electrospray ionization interface. A Kromasil ODS column (particle size 5 mm, 20 cm 4.6 mm inside diameter; Hi-Tech Scientific Instrument Corp., Tianjin, China) was used for sample separation. A mobile phase consisting of acetonitrile-methanol-10 mM ammonium acetate (50:10:35, v/v/v) was used at a flow rate of 0.4 mL/min for the analysis of in vitro incubation samples. Identification of roxithromycin metabolites in rat bile was performed by using another mobile phase system consisting of acetonitrilemethanol-10 mM ammonium acetate (43:10:47, v/ v/v) at a flow rate of 0.3 mL/min. The ionization was realized by applying spray voltage of 4.25 kV, capillary temperature of 1808C, and capillary voltage of 30 V. Nitrogen was used as both the JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
RESULTS Roxithromycin Transformation in Simulated Gastric Fluid (pH 1.2) In Vitro Roxithromycin was detected at a retention time of 12 min, with [M þ H]þ ion at m/z 837 (Fig. 1). It showed MS/MS fragment ions at m/z (relative abundance, %) 716, 679 (100), 558, 540, and 522, characteristic losses of oxime alkylether side chain (121 u), cladinose moiety (158 u), water (18 u), and desosamine (157 u) (Fig. 2). The
Figure 1. Representative chromatograms and MS/ MS spectra of roxithromycin and its metabolites in simulated gastric fluid (pH 1.2) after a 10-min incubation at 378C.
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
Figure 2. Proposed MS/MS fragmentation pathways of roxithromycin.
fragmentation profile was in accordance with that of the previously reported for roxithromycin.12 Two metabolites M1 and M2 were found in the samples incubated with simulated gastric fluid at 378C (Fig. 1). M1 was detected at LC retention time of 9 min, with the same [M þ H]þ and MS/MS fragment ions as those of roxithromycin. M1 was identified as Z-isomer of roxithromycin by comparison with the synthesized Z-roxithromycin reference. M2, with LC retention time of 8 min, showed quasi-molecular ion [M þ H]þ at m/z 679, which is 158 u less than the parent drug. This is a characteristic for the loss of cladinose moiety. Its MS/MS spectra gave fragment ions at m/z 661, 558, 540, and 522 (100), which resulted from loss of water, oxime side chain, and desosamine. M2 was identified as decladinose roxithromycin by comparison with synthesized reference. A minor metabolite, with the same [M þ H]þ and MS/MS fragmentation ions as M2, measured at a retention time of 7 min, was considered as the Z-isomer of M2. The degradation of roxithromycin in simulated gastric fluids followed pseudo first-order kinetics. The semilogarithmic curves of roxithromycin concentration versus time are shown in Figure 3. The rate constants of roxithromycin degradation were 0.0994 min1 (0.0031) in simulated gastric fluid, 0.0995 min1 (0.0030) in simulated gastric fluid containing NaCl, and 0.0993 min1 (0.0033) in simulated gastric fluid containing pepsin, respectively. The geometric isomer of roxithromycin was detected immediately after mixing of roxithromycin with gastric fluid. The ratio of Z-roxithromycin to the parent drug (SD) was constant at 0.19 (0.03) in gastric fluid, 0.18 (0.03) in gastric fluid containing NaCl, and 0.18 (0.03) in gastric fluid containing pepsin, respectively, during a 10–
1303
Figure 3. The concentration of roxithromycin in simulated gastric fluids (pH 1.2) versus time of incubation at 378C.
45-min incubation period. The level of roxithromycin was below the limit of quantification after 45 min of incubation. The level of decladinose roxithromycin was increased with time. To gain additional insight into the isomerization of roxithromycin, the incubation of Z-roxithromycin was also conducted in the simulated gastric fluid at 378C as described above. Zroxithromycin was transformed to roxithromycin and decladinose derivative in simulated gastric fluid. The isomerization of Z- to E-configuration was with greater extent than that of E- to Zconfiguration. The mean ratio of Z-roxithromycin to its E-isomer for 10–45 min of incubation (SD) was 0.16 (0.01) in the gastric fluid, 0.17 (0.01) in the gastric fluid containing NaCl, and 0.16 (0.01) in the gastric fluid containing pepsin, respectively. The degradation scheme of roxithromycin in the simulated gastric fluid at 378C is shown in Figure 4. pH Effect on the In Vitro Geometric Isomerization of Roxithromycin When either roxithromycin or Z-roxithromycin was incubated with the mixture (about pH 10) of 2 mL of pH 1.0 gastric fluid and 3 mL of 0.1 M Na2CO3, neither geometric isomer nor degradation derivative were detected. The 0.1 M Na2CO3 was used by 1.5-fold of sample volume to terminate the reaction. The degradation of roxithromycin was performed at 378C in the simulated gastrointestinal fluid of pH 1.0, 1.3, 1.8, 3.0, 5.0, and 7.4. Except for Z-isomer and decladinose derivatives of roxithroJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
1304
ZHANG, XING, AND ZHONG
Roxithromycin Metabolism in Rats after an Oral Dose
Figure 4. The proposed degradation of roxithromycin in the simulated gastric fluid at 378C.
mycin, no other degradation derivative was detected in simulated gastrointestinal fluids of pH 1.0–3.0. No degraded derivative of roxithromycin was detected in simulated gastrointestinal fluid of pH 5.0 and 7.4 during incubation. The decomposition rate of roxithromycin decreased with the increase in pH value from 1.0 to 3.0. The rate constant (SD) of roxithromycin degradation was 0.1066 min1 (0.0014) at pH 1.0, 0.0400 min1 (0.0003) at pH 1.3, 0.0136 min1 (0.0008) at pH 1.8, and 0.0022 min1 (0.0002) at pH 3.0, respectively. The extent of isomerization of roxithromycin to Z-isomer decreased with the increase of pH value. The mean ratio (SD) of Z-roxithromycin to roxithromycin was constant at 0.21 (0.01) at pH 1.0 during the sampling interval (5–30 min), 0.18 (0.01) at pH 1.3 during the sampling interval (10–60 min), 0.15 (0.01) at pH 1.8 during the sampling interval (10–240 min), and 0.08 (0.02) at pH 3.0 during the sampling interval (45–360 min), respectively. The degradation rate constants of roxithromycin and the ratios of Z-roxithromycin to roxithromycin at different pH’s are summarized in Table 1.
A number of roxithromycin-related derivatives, with quasi-molecular ion [M þ H]þ at m/z of 837, 823, 809, 749, 735, and 679, respectively, were detected in the rat bile after an oral dose of roxithromycin of 20 mg/kg. Their TIC and MS/MS full scan chromatograms are shown in Figure 5A. Their structures were confirmed or suggested based on HPLC behaviors and multiple-step MS data with reference substances. Roxithromycin and its Z-isomer M1 were detected in m/z 837 channel at retention times of 28 and 18 min, and were confirmed with reference substances, respectively. The ratio of Z-isomer to roxithromycin was about 0.3 at maximum in the bile collected during 2–4 h after administration of roxithromycin. Decladinose roxithromycin (M2) was detected at a retention time of 15 min. And another metabolite, with the same [M þ H]þ and MS/MS fragment ions as those of M2, was detected at a retention time of 12 min. It was proposed to be the Z-isomer of M2. Metabolites M3 and M4 were measured at m/z 823 channel at retention times of 21 and 15 min, respectively, with the same [M þ H]þ at m/z 823, which is 14 u less than the parent drug, a typical characteristic loss of methyl group. M3 showed MS/MS fragments of m/z 665 (100), 544, 522, and 508, resulting from the decomposition of cladinose (158 u), oxime alkylether side chain (121 u), demethyl desosamine (143 u), and two molecules of water (36 u). M3 was confirmed to be N-demethylated roxithromycin by comparison with reference standard. M4 gave MS/MS fragment ions at m/z 665 (100), 558, 540, 522, and 508, suggesting the losses of cladinose moiety, O-demethylated oxime alkylether side chain (107 u), two molecules of water, and desosamine, respectively. M4 was confirmed to be O-demethy-
Table 1. The Degradation Rate Constants and Extents of Isomerization of Roxithromycin in Simulated Gastric Fluids of Different pH Values at 378C pH Value
k (min1)a (SD) Z/Eb (SD)
1.0
1.2
1.3
1.8
3.0
0.1066 (0.0014) 0.21 (0.01)
0.0994 (0.0031) 0.19 (0.03)
0.0400 (0.0003) 0.18 (0.01)
0.0136 (0.0008) 0.15 (0.01)
0.0022 (0.0002) 0.08 (0.02)
a
First-order degradation rate constant. Ratio of Z-roxithromycin to roxithromycin.
b
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
1305
Figure 5. Total ion current and MS/MS full scan chromatograms of roxithromycin and its metabolites in rat bile collected during 2–4 h after an oral dose of roxithromycin (20 mg/kg) (A), or an oral dose of roxithromycin (20 mg/kg) with alkalization of Na2CO3 (B), or an intravenous dose of roxithromycin (3.5 mg/kg) (C).
lated roxithromycin by comparison with synthesized reference. Metabolites M5 and M6 were detected at m/z 809 with retention times of 19 and 12 min, respectively. Their same quasi-molecular ion [M þ H]þ at m/z 809, 28 u lower than roxithromycin, suggested they were di-demethylated metabolites of roxithromycin. M5 gave MS/MS fragment ions at m/z 651 (100), 530, 522, and 512, characteristic losses of cladinose, and then oxime alkylether side chain, di-demethyl desosamine (129 u), and water. M5 was tentatively assigned as N,N-di-demethylated roxithromycin based on its MSn data and by comparison with N-demethyl roxithromycin. M6 showed MS/MS fragment ions at m/z 651 (100), 544, 526, and 508, indicating the losses of cladinose, and then demethylated oxime alkylether side chain, water, and demethylated desosamine. M6 was proposed to be N,O-didemethyl roxithromycin. Metabolite M7, with an LC retention time of 14 min, showed [M þ H]þ at m/z 749, 88 u lower than that of roxithromycin, which is characteristic loss of alkylether side chain. Its MS/MS fragment ions were detected at m/z 591 (100), 558, 540, and 434, suggesting the decomposition of cladinose, oxime group (33 u), water, and desosamine. M7 was confirmed as erythromycin oxime by the comparison of LC behavior and MSn data with
reference standard. Another metabolite, with the same MSn spectra as M7 but higher level than M7, measured at retention time of 12 min, was considered as Z-isomer of M7. Metabolite M8 showed an LC retention time of 12 min, and [M þ H]þ at m/z 735 (14 u less than M7). Its MS/MS fragment ions at m/z 577 (100), 544, 526, and 434, were 14 u lower than the corresponding fragment ions of M7. It was tentatively suggested to be N-demethyl erythromycin oxime based on its mass data. Its Z-isomer, showing the same [M þ H] and MS/MS fragment ions as M8, was also detected at a retention time of 10 min. A similar metabolic profile was observed for Z-roxithromycin in rats after oral administration of Z-roxithromycin (Fig. 6A). Z-roxithromycin was more easily transformed to its E-isomer (roxithromycin) than E- to Z-configuration. The level of roxithromycin was higher than that of Z-isomer in the bile collected after 2 h of administration of Z-roxithromycin. The Metabolism of Roxithromycin after an Oral Dose with Na2CO3 When roxithromycin was orally administrated to rats with Na2CO3, demethylated metabolites (M3, M4, M5, and M6) and dealkylated metabolites JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
1306
ZHANG, XING, AND ZHONG
Figure 6. Total ion current and MS/MS full scan chromatograms of Z-roxithromycin and its metabolites in rat bile collected during 2–4 h after an oral dose of Z-roxithromycin (20 mg/kg) (A), or an oral dose of Z-roxithromycin (20 mg/kg) with alkalization of Na2CO3 (B), or an intravenous dose of Z-roxithromycin (3.5 mg/kg) (C).
(M7 and M8) were also detected in rat bile, but decladinose derivative (M1), and Z-isomers of parent drug and metabolites in the rat bile collected during 0–24 h after drug administration were not detected (Fig. 5B). When Z-roxithromycin was orally administrated to rats with Na2CO3 (Fig. 6B), no E-isomers of parent drug and metabolites were detected in the bile collected during 0–8 h after dosing, whereas roxithromycin was detected after 8 h of administration. The ratio of roxithromycin to its Z-isomer reached about 0.1 in the bile collected during the 8–10 h interval, and about 0.3 during the 12–24 h interval, after drug administration. The Z-isomers of metabolites M3 and M4 were detected at retention times of 15 and 11 min, respectively. The Z-isomers of M5 and M6 showed retention times of 14 and 10 min, respectively. Finally, the Z-isomers of M7 and M8 were detected at retention times of 12 and 10 min, respectively. The Metabolism of Roxithromycin after an Intravenous Dose When roxithromycin was intravenously injected to rats (Fig. 5C), neither Z-isomer nor decladinose derivative of roxithromycin was detected in the 0–24 h bile after administration. Other JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
metabolites including demethylated and dealkylated derivatives of E-configuration were detected. After an intravenous dose of Z-roxithromycin, only Z-isomers of demethylated metabolites and dealkylated metabolites were measured in the rat 0–24 h bile, but decladinose derivative and roxithromycin as well as E-isomers of metabolites were not detected (Fig. 6C). The metabolic profile of roxithromycin in rats is shown in Figure 7.
DISCUSSION Roxithromycin and its metabolites followed common fragmentation pathways in electrospray ionization MS/MS full scan, which was consistent with previous results.12 Z-isomers showed the same MSn spectra as those of the corresponding E-isomers, but their retention times under the reversed phase HPLC conditions were shorter than the corresponding E-isomers. In the simulated gastric fluid at 378C, roxithromycin transformed to its Z-isomer and decladinose derivative. The geometric isomerization in the gastric acid condition had not been reported, which
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
1307
Figure 7. The proposed metabolic pathways of roxithromycin in rats after an oral dose of 20 mg/kg of roxithromycin.
might be partly due to the lack of sensitivity and/or specificity of the previous assay methods. Roxithromycin degraded in simulated gastric fluids at 378C according to first-order kinetics with a half-life of 6.5 min at pH 1.0, 17 min at pH 1.3, 51 min at pH 1.8, and 315 min at pH 3.0, respectively. No degradation occurred at pH 5.0 and 7.4. Neither NaCl nor pepsin altered the degradation of roxithromycin. These results were generally consistent with a previous study for roxithromycin disintegration by HPLC with coulometric detection.13 The isomerization occurred immediately after mixing of either roxithromycin or its Z-isomer with gastric fluids. Actually, the ratios of Z-roxithromycin to roxithromycin were constant during sampling duration. The extent of isomerization of E- to Z-configuration decreased with the increase of pH value from 1.0 to 3.0. This suggested the pH dependence of roxithromycin isomerization. It was reported that the roxithromycin oxime was transformed to its Z-configuration in strong alkaline medium (1.2 mL of 2 M NaOH and 2.7 mL of methanol).3 In the present study, no isomerization of roxithromycin was observed at pH 5, 7.4, and 10 (in the mixture of 2 mL of pH 1.0 simulated gastric fluid and 3 mL of 0.1 M Na2CO3). It is interesting to note that the isomerization between E- and Z-configurations was interconvertible, but the extent of isomerization of Z- to E-configuration was much greater than that of E- to Z-configuration. That might be related to the differences in stereo characteristics between two configurations of roxithromycin. Gharbi-Benarous et al.14,15 reported that the oxime side chain in the roxi-
thromycin molecule was oriented above the macrocyclic lactone ring and the oxygen atoms [O (17) and O (19)] of the chain were engaged in tight hydrogen bonding with 6- and 11-hydroxyl groups of the macrocycle, based on 1H and 13C NMR data. This resulted in a globular structure for roxithromycin with less freedom for the macrocyclic lactone ring and sugar unit. We had compared 1H NMR spectra of roxithromycin and Z-roxithromycin (in CDCl3). No obvious signal was observed in Z-roxithromycin spectra at the corresponding chemical shift positions of 6- and 11-hydroxyl groups (2.48 and 4.38 ppm, respectively) as in roxithromycin spectra. This suggested that the intramolecular hydrogen bonding might be weak or disappeared in Z-configuration. Therefore, the E-configuration should be the preferred configuration to Z. Roxithromycin underwent four metabolic pathways such as geometric isomerization, hydrolysis of cladinose, demethylation, and dealkylation in rats after oral administration. Z-roxithromycin followed a similar metabolic profile to parent drug, but its isomerization was found to be faster than roxithromycin. No isomerized derivatives and decladinose derivative were detected during 0–24 h after oral administration of roxithromycin with Na2CO3 alkalization or after intravenous injection of roxithromycin. When Z-roxithromycin was given orally with alkalization, some E-isomerized derivatives were detected after 8 h of administration, which probably resulted from the accumulated secretion of gastric acid and greater liability of isomerization from Z- to E-configuration. The JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
1308
ZHANG, XING, AND ZHONG
above results suggested that the geometric isomerization of roxithromycin should take place in the stomach of animals, and it was dependent on gastric acid. It is also collateral evidence to support the previous studies that no isomerized derivatives were detected in incubations with rat liver microsomes9 and in perfused rat liver10 by mass spectrometry. The results also proposed that the metabolites of Z-configuration, detected after oral administration of roxithromycin, should be the metabolites of Z-roxithromycin, but were not the consequence of transformation from the metabolites of E-configuration. The levels of roxithromycin and its metabolites were normally higher than those of the Z-isomers after oral administration of roxithromycin, but the levels of dealkylated metabolites (M7 and M8) were lower than those of their Z-isomers (similar to reported results7), which suggested that the dealkylation of oxime ether side chain of Z-roxithromycin seemed more susceptible to occur than that of roxithromycin. Previous studies reported that the geometric isomerization of roxithromycin after oral administration was predominant in humans and in dogs,7,8 but insignificant in rats and in golden hamsters.11 The species differences in roxithromycin geometric isomerization might be attributable to the differences in the acidity and volume of gastric fluid among different animals. Additionally, it is likely that the effects of administration factors on gastric fluid acidity might be less in bigger animals than in smaller animals. Considering the weaker antibiotic potency of Zroxithromycin than that of roxithromycin,3,16,17 the pH-dependent geometric isomerization during absorption of roxithromycin should have some effects on roxithromycin bioavailability, consequently clinical efficacy. This could provide some inspiration for the design of dosage form and clinical use of roxithromycin. It seems to be beneficial to increase the pH of gastric juice for the roxithromycin absorption, considering both isomerization and decladinose. A study in humans showed that the roxithromycin concentration in the gastric juice and gastric tissue could be significantly increased by combined administration with proton pump inhibitor, which was considered to contribute to the synergic beneficial action in the eradication therapy of Helicobacter pylori.13
CONCLUSIONS Roxithromycin degraded to its Z-isomer and decladinose derivative in simulated gastric fluid, JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
and underwent the following metabolic pathways in rats: geometric isomerization, hydrolysis of cladinose, demethylation, and dealkylation. The geometric isomerization of roxithromycin was pH dependent and reversible both in vitro and in vivo. Z-roxithromycin was more liable to transformation to its E-isomer.
ACKNOWLEDGMENTS This research was supported by the National Natural Science Foundation of China, number 39930180.
REFERENCES 1. Young RA, Gonzalez JP, Sorkin EM. 1989. Roxithromycin. A review of its antibacterial activity. Pharmacokinetic properties and clinical efficacy. Drugs 37:8–41. 2. Markham A, Faulds D. 1994. Roxithromycin: An update of its antimicrobial activity, pharmacokinetic properties and therapeutic use. Drugs 48: 297–326. 3. Gase JC, d’Ambrieres SG, Lutz A, Chantot JF. 1991. New ether oxime derivatives of erythromycin A: A structure-activity relationship study. J Antibiot 44:313–330. 4. Esumi Y, Zin Y, Ban S, Ninomiya S, Hayashi K, Yokoshima T. 1988. Absorption, distribution, metabolism, and excretion of RU 28965 in animals. Chemotherapy 36(Suppl 4):148–163. 5. Koyama M, Tateno M, Shirotsuka M, Yamamoto T, Hirayama M, Saito K, Okui K. 1988. Absorption, metabolism, and excretion of RU 28965 in humans. Chemotherapy 36(Suppl 4):164–183. 6. McLean A, Sutton JA, Salmon J, Chatelet D. 1988. Roxithromycin: Pharmacokinetic and metabolism study in humans. Br J Clin Pract 42(Suppl 55): 52–53. 7. Zhong DF, Li XQ, Wang AM, Xu YJ, Wu SD. 2000. Identification of the metabolites of roxithromycin in humans. Drug Metab Dispos 28:552–559. 8. Zhang SQ, Zhang LF, Xing J, Zhong DF. 2003. The metabolism of roxithromycin in dogs. Yao Xue Xue Bao 38:374–379. 9. Zhong DF, Zhang SQ, Sun L, Zhao XY. 2002. Metabolism of roxithromycin in phenobarbitaltreated rat liver microsomes. Acta Pharmacol Sin 23:455–460. 10. Jarukamjorn K, Thalhammer T, Gollackner B, Pittenauer E, Jaeger W. 1998. Metabolism of roxithromycin in the isolated perfused rat liver. J Pharm Pharmacol 50:515–519.
GEOMETRIC ISOMERIZATION OF ROXITHROMYCIN
11. Zhang SQ, Zhong DF. 2003. The metabolism of roxithromycin in golden hamsters. Asian J Drug Metab Pharmcokinet 3:195–202. 12. Zhong DF, Tian L, Li XQ, Xu YJ. 2000. Studies on the electrospray ion trap mass spectra of roxithromycin and its metabolites. Chem J Chin Univ 21: 31–35. 13. Kees F, Holstege A, Ittner KP, Zimmermann M, Lock G, Scho¨lmerich J, Grobecker H. 2000. Pharmacokinetic interaction between proton pump inhibitors and roxithromycin in volunteers. Aliment Pharmacol Ther 14:407–412. 14. Gharbi-Benarous J, Delaforge M, Artaud I, Girault J-P. 1990. Analysis of the 1H and 13C NMR spectra of the novel macrolide antibiotic roxithromycin.
1309
Structure and conformation in solution. Magn Reson Chem 28:846–855. 15. Gharbi-Benarous J, Delaforge M, Jankowski CK, Girault J-P. 1991. A comparative NMR study between the macrolide antibiotic roxithromycin and erythromycin A with different biological properties. J Med Chem 34:1117–1125. 16. Li XQ, Zhong DF, Huang HH, Wu SD. 2001. Demethylation metabolism of roxithromycin in humans and rats. Acta Pharmacol Sin 22: 469–474. 17. Huang HH, Yu LH, Sun TM, Li XQ, Zhong DF. 2000. In vitro antibacterial activity of roxithromycin and its major metabolites. Chin J Antibiot 25: 443–446.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004