Abstract
Olanzapine (OLZ) is a novel antipsychotic agent with a high affinity for serotonin (5-HT2), dopamine (D1/D2/D4), muscarinic (m1–m5), adrenergic (α1), and histamine (H1) receptors. The pharmacokinetics, excretion, and metabolism of OLZ were studied in CD-1 mice, beagle dogs, and rhesus monkeys after a single oral and/or intravenous dose of [14C]OLZ. After oral administration, OLZ was well absorbed in dogs (absolute bioavailability of 73%) and to the extent of at least 55% in monkeys and 32% in mice. The terminal elimination half-life of OLZ was relatively short in mice and monkeys (∼3 hr) and long in dogs (∼9 hr). In mice and dogs, radioactivity was predominantly eliminated in feces; but, in monkeys, the major route of elimination of radioactivity was urine. Dogs and monkeys excreted in urine, respectively, 38% and 55% of the dose over a 168-hr period, whereas the fraction of the dose excreted in urine of mice over the collection period (120 hr) was 32%. OLZ was subject to substantial first-pass metabolism; at the tmax, OLZ accounted for 19%, 18%, and 8% of the radioactivity, in mice, dogs, and monkeys, respectively. The ratio of AUC OLZ to AUC radioactivity was, respectively, 10%, 14%, and 4% in mice, dogs, and monkeys. The principal urinary metabolites in mice were 7-hydroxy OLZ glucuronide, 2-hydroxymethyl OLZ, and 2-carboxy OLZ accounting for ∼10%, 4%, and 2% of the dose. Metabolites that were present in urine in lesser amounts were 7-hydroxy OLZ, N-desmethyl OLZ, andN-desmethyl-2-hydroxymethyl OLZ. In dogs, the major metabolite accounting for ∼8% of the dose was 7-hydroxy-N-oxide OLZ. Other metabolites identified were 2-hydroxymethyl OLZ, 2-carboxy OLZ, N-oxide OLZ, 7-hydroxy OLZ, and its glucuronide and N-desmethyl OLZ. The major metabolite in monkey urine was N-desmethyl-2-carboxy OLZ, and accounted for ∼17% of the dose. In addition,N-oxide-2-hydroxymethyl OLZ, N-oxide-2-carboxy OLZ, N-desmethyl-2-hydroxymethyl, 2-carboxy OLZ, and 2-hydroxymethyl OLZ were identified in monkey urine. Thus, in mice and dogs, OLZ was metabolized through aromatic hydroxylation, allylic oxidation, N-dealkylation, and N-oxidation reactions. In monkeys, OLZ was biotransformed mainly through double oxidation reactions involving the allylic carbon and methyl piperazine nitrogen. Whereas the oxidative metabolic profile of OLZ in animals was similar to that of humans, animals were notable for not forming appreciable amounts of the principal human metabolite (i.e.10-N-glucuronide OLZ).
OLZ1(fig.1) is a new antipsychotic drug with a thienobenzodiazepinyl structure. OLZ displays a broad pharmacological profile with potent activity at dopamine (D1/D2/D4), serotonin (5-HT2A/2C), muscarinic (especially m1), histamine (H1) and adrenergic (α1) receptors (1, 2). The receptor binding profile of OLZ is very similar to clozapine, although OLZ is a more potent inhibitor of these receptors.
In clinical studies with patients suffering from schizophrenia or schizophreniform disorder, OLZ was effective in the treatment of both positive and negative symptoms of schizophrenia, with a low incidence of extrapyramidal side-effects (3-5). Antipsychotic efficacy of OLZ was demonstrated in the dose range of 5–20 mg/day.
The disposition and metabolism of OLZ after a single oral dose to healthy volunteers has recently been reported (6). OLZ was well absorbed and extensively metabolized. The primary metabolic route wasN-glucuronidation. OLZ also underwent oxidative metabolism through N-oxidation, N-demethylation, and 2-alkyl hydroxylation. This study describes the comparative absorption, pharmacokinetics, and metabolism of OLZ in mice, dogs, and monkeys. The studies were conducted after the administration of [14C]OLZ.
Materials and Methods
Reference Compounds and Other Materials.
The following compounds were synthesized at Lilly Research Laboratories: OLZ (2-methyl-4-(4-methyl-1-piperazinyl)-10H-thieno[2,3-B][1,5]benzodiazepine), [4,10a-14C2]OLZ ([14C]OLZ; radiochemical purity, 98.7%; specific activity, 26.2 μCi/mg), 4′-N-desmethyl OLZ (N-desmethyl OLZ, 2-methyl-4-(1-piperazinyl)-10H-thieno[2,3-B][1,5]benzodiazepine), 4′-N-oxide OLZ (N-oxide OLZ, 4-(2-methyl-10H-thieno[2,3-B][1,5]benzodiazepin-4-yl)-1-methylpiperazine-1-oxide), 2-hydroxymethyl OLZ (4-(4-methyl-1-piperazinyl)-10H-thieno[2,3-B][1,5]benzodiazepine-2-methanol), 2-carboxymethyl OLZ (methyl 4-(4-methyl-1-piperazinyl)-10H-thieno[2,3-B][1,5]benzodiazepine-2-carboxylate), 7-ethoxy OLZ (7-ethoxy-2-methyl-4-(4-methyl-1-piperazinyl)-10H-thieno〈2,3-B〉〈1,5〉benzodiazepine), and 4′-N-desmethyl-2-hydroxymethyl OLZ (4-(1-piperazinyl)-10H-thieno[2,3-B][1,5]benzodiazepine-2-methanol). NEE-154 Glusulase was purchased from the DuPont Company (Wilmington, DE). β-Saccharolactone was purchased from Sigma Chemical Co. (St. Louis, MO). Scintisol was supplied by Isolab, Inc. (Akron, OH). HPLC-grade ammonium acetate, acetonitrile, triethylamine, and reagent-grade boron tribromide were purchased from Fisher Scientific (Fair Lawn, NJ). 7-Hydroxy OLZ was prepared by deethylation of 7-ethoxy OLZ, and 2-carboxyl OLZ was prepared by hydrolysis of 2-carboxymethyl OLZ. Approximately 2 mg of each starting material was placed in separate siliconized tubes and dissolved in methylene chloride (2 ml). The solution was flushed with nitrogen and treated with boron tribromide solution (2 ml of 25% solution in methylene chloride). The reaction was allowed to proceed at room temperature for 2 hr. Approximately 90% of 7-ethoxy OLZ was converted to 7-hydroxy OLZ, whereas ∼50% of 2-carboxymethyl was converted to the corresponding acid as determined by HPLC and electrospray LC/MS.N-Desmethyl-2-carboxy OLZ was prepared by oxidizing the corresponding hydroxy compound using chromium trioxide (7).
Animal Experiments.
All animal experiments were conducted according to protocols approved by the Eli Lilly Animal Care and Use Committee. The dosing solution used for all animal studies was prepared by dissolving the required amounts of OLZ and [14C]OLZ in 1 M HCl and titrating the solution to approximately pH 6 by the addition of 0.1 M NaOH. The appropriate volume was then obtained by the subsequent addition of water.
Mouse.
Male CD-1 mice were obtained from Charles River Laboratories (Wilmington, MA) and acclimatized for 3 days before use. Food and water were supplied ad libitum at all times throughout the experiment. For radiolabeled excretion study, the mice were divided into three groups, with each group containing five mice. Each animal was administered a single oral (gavage) dose of OLZ (15 mg/kg containing 420 μCi/kg of [14C]OLZ). Urine and fecal samples were collected at 24-hr intervals for up to 120 hr. For pharmacokinetic study, four mice were used for each time point and dosed as described. Blood was collected and pooled from four mice at 0.5, 1, 2, 4, 7, 12, 24, 48, and 72 hr after the dose. Plasma was obtained by centrifugation and stored at −70°C until analysis. Metabolite identification was conducted in urine obtained from mice given a 20 mg/kg dose. Urine samples collected for 24-hr postdose from eight mice were combined and stored at −70°C until analyzed. Additional mice (10) were administered a single oral dose (15 mg/kg) of OLZ, and plasma was collected at ∼1 hr for metabolite identification.
Dog.
Four female beagle dogs (age: 2–4 years; weight: 8.7–13.1 kg) were obtained from stock animals maintained at Lilly Research Laboratories and placed in individual stainless-steel metabolism cages. Animals were fasted overnight before and 2 hr after drug administration. Animals were given a single oral (gavage) dose of OLZ (5 mg/kg containing 4 μCi/kg of [14C]OLZ). Urine and fecal samples were collected every 24 hr for 168 hr. Blood was drawn at 0, 0.5, 1, 3, 6, 12, 24, 48, 96, and 168 hr after dosing. Aliquots were withdrawn for determination of radioactivity, and the remainder was centrifuged to obtain plasma. Plasma was also obtained from three female dogs given a single intravenous dose of OLZ (5 mg/kg containing 5 μCi/kg [14C]OLZ) at 0, 0.08, 0.25, 0.5, 1, 3, 6, 12, 24, 36, 48, 72, 96, 120, 144, and 168 hr postdose. For identification of plasma metabolites, another group of three dogs were dosed orally with OLZ (5 mg/kg), and plasma was collected at 3 and 12 hr after the dose.
Monkey.
Young adult rhesus monkeys (2 males and 2 females) weighing between 3 and 7 kg were used in the study. Each animal was given a single oral (nasogastric) dose of OLZ (5 mg/kg containing 8.9 μCi/kg [14C]OLZ). Blood was collected at 0, 0.5, 1, 4, 8, 12, 24, 48, 96, 120, and 168 hr postdose, whereas urine and fecal samples were collected every 24 hr up to 168 hr.
Analysis of Radioactivity.
Total radioactivity in plasma and urine samples was determined using LSC after the addition of scintillation cocktail (Aquassure; New England Nuclear, Boston, MA). Feces was suspended in 5% aqueous sodium lauryl sulfate solution, and aliquots of the dried homogenate were combusted in a sample oxidizer. The fecal sample prepared in this manner was counted after addition of Aquassure. Mice carcasses were digested using alcoholic potassium hydroxide. The homogenate obtained in this manner was neutralized with acetic acid, and the radioactivity was determined by LSC. Quench correction was conducted by automatic external standardization.
HPLC Assay of Plasma OLZ.
Concentrations of OLZ in mouse, dog, and monkey plasma were determined by HPLC (8). In this assay, OLZ and the internal standard (2-ethyl homolog of OLZ) are isolated from plasma using a mixed-mode, solid-phase extraction, separated with a reversed-phase method, and detected electrochemically. The upper and lower limits of quantitation of the assay were 100 and 1 ng/ml, respectively.
Pharmacokinetics.
Noncompartmental analysis was used to determine the pharmacokinetics of OLZ and radioactivity. Cmax andtmax were assesed by visual inspection. The terminal elimination half-life was calculated using the relationship 0.693/k, where k is the elimination rate constant. The AUC-–t was calculated up to the last time point (t) by the trapezoidal rule.
In Vitro Protein Binding.
[14C]OLZ was dissolved in n-propyl alcohol at 0.2, 0.02, and 0.002 mg/ml, and an aliquot (15 μl) of each concentration was added into 2985 μl volume control plasma. Plasma samples were then placed in a water bath (∼37°C) for 1 hr. After ultracentrifugation (360,000 g at 37°C for 4 hr), the amount of OLZ in the supernatant was determined by LSC. The fraction of OLZ bound to protein was calculated from the radioactivity concentrations in the spiked sample and the supernatant.
Metabolite Isolation.
Aliquots of urine samples from each of the orally dosed dogs were combined (∼80 ml total urine) and made basic by the addition of 0.1 M ammonium hydroxide (8 ml). Ethyl acetate (300 ml) was added, and the phases were mixed by shaking vigorously. The ethyl acetate layer was separated and evaporated to dryness in a water bath (40°C) under a stream of nitrogen. The aqueous fraction was lyophilized to dryness, dissolved in water, and analyzed by HPLC. Pooled mouse urine was also extracted as described. For NMR analysis, the 7-hydroxyl-N-oxide metabolite was isolated from urine using column chromatography. Approximately 400 g of Amberlite XAD-2 resin was packed in a 2.5 × 30 cm glass column. The remaining urine samples from each dog were combined (∼1.5 liters) and passed through the column after preconditioning the column with methanol and purified water. The column was washed with water and the radioactivity eluted with methanol. Methanolic extracts were concentrated in vacuo at 25°C using a rotary evaporator, and the resulting residue was reconstituted in 70 ml methanol/water (1:1) for HPLC analysis. Aliquots (50 ml; 0–24 hr) of urine sample from each monkey were lyophilized to ∼2 ml. The residue was reconstituted in 5 ml water:methanol (4:1, v/v) and separated by HPLC.
Estimation of the Amount of Metabolites.
The amount of each metabolite in urine was estimated by LSC after isolation by HPLC. An aliquot of the ethyl acetate or aqueous extract (50–200 μl) was injected into HPLC, and each metabolite was collected as it eluted the column. Scintisol (15 ml) was added, and the amount of radioactivity was determined by LSC. The total radioactivity in the sample was determined by injecting an equal aliquot into the HPLC injector and collecting the entire sample before it reached the column.
Hydrolysis of Conjugates.
Glucuronide conjugates (∼2 μg) isolated from urine were hydrolyzed to the corresponding aglycone by incubating with Glusulase (containing 2,070 units of β-glucuronidase and 150 units sulfatase) at 37°C for up to 20 hr. Incubations were also conducted in the absence of Glusulase and in the presence of β-saccharolactone (0.0325 M).
HPLC Separation of Metabolites.
The HPLC system consisted of a Beckman pump, NEC controller, Waters Wisp autosampler, Applied Biosystem UV detector, and a Berthold radiodetector with 150 μl yittrium solid cell. Aliquots (≤200 μl) of concentrated urine or extract were analyzed on a Hypersil C18 column (5 μm particle size, 0.46 × 25 cm) using a gradient containing A (0.1 M ammonium acetate) and B (1% triethylamine in acetonitrile). The initial solvent composition was 90% A and 10% B. After 2 min, the pump was programmed to increase solvent B by 2.5%/min until a proportion of 40% A and 60% B was achieved. The mobile phase was maintained for 8 min at that composition. The flow rate was 1 ml/min. Metabolites were isolated by collecting the radioactive eluent as it eluted off the column. Several injections were made to obtain a sufficient amount of each metabolite for mass spectral identification.
LC-MS/MS.
Isolated metabolites were analyzed by LC/MS and LC-MS/MS on a Finnigan MAT TSQ700. Metabolites were introduced into the electrospray LC interface using a Waters Model 600 pump. Metabolites were separated on an Inertsil C18 column (5 μm particle size, 0.46 × 25 cm) using the same gradient as described with 0.05 M ammonium acetate and acetonitrile. Injection volumes ranged from 10 to 200 μl. The flow rate was 1 ml/min, and the effluent was split such that equal volumes were delivered into the ion source and a Raytest Ramona model 5LS radiodetector. MS spectra were obtained by scanning fromm/z 200 to 600 every second. For CID experiments, the collision gas (argon) pressure was maintained at 2.0 m torr, and the collision offset voltage was −20 eV. MS and MS/MS spectra were averaged for 1 min.
NMR Spectroscopy.
Proton and carbon-13 NMR spectra were recorded ind6-DMSO or CDCI3 on a Bruker AMX spectrometer operating at 500 MHz. Chemical shifts are reported in ppm relative to tetramethylsilane.
Results
Excretion of Radioactivity.
Mice administered a single oral dose (15 mg/kg) of OLZ eliminated 64.3 ± 3.4% (mean ± SD) and 31.9 ± 2.8% of the radioactivity, respectively, in feces and urine over a 120-hr period (table 1). The majority of the dose (>87%) was excreted during the first 48 hr of dosing. Less than 1% of the administered dose was recovered in the carcasses.
In dogs, ∼84% of the radioactivity was recovered after 168 hr, with slightly more radioactivity eliminated in the feces (45.6 ± 5.4%) than in the urine (38.4 ± 2.6%). Greater than 50% of the dose was recovered within 48 of dosing (table 1).
In monkeys, renal excretion was the primary mode of radiocarbon elimination accounting for 54.6 ± 3.7% of the dose. Another 28.5 ± 5.2% of the dose was eliminated via the feces over the same period. Greater than 50% of the dose was eliminated in the urine and feces 24 hr after the dose (table 1). There was no difference between males and females with respect to the amount of radioactivity in either the urine or feces.
Pharmacokinetics. Mice.
Pharmacokinetic parameters of OLZ and radioactivity in mice are shown in table 2. OLZ was quantitated in plasma using an HPLC assay with a lower limit of quantitation of 1 ng/ml. TheCmax of OLZ was 421 ng/ml and occurred at 0.5 hr after the dose. The corresponding value for radioactivity was 2,260 ng-eq/ml and was reached at a much later time (4 hr). At 0.5 hr, OLZ accounted for ∼19% of plasma radioactivity. This is indicative of the extensive metabolism of OLZ in the mouse. Similarly, OLZ accounted for 10% of the total 14C AUC. The plasma terminal half-life of OLZ was 3.2 hr. Radioactivity in plasma declined slowly with a half-life of 10.6 hr. The plasma radioactivity vs.time curve (fig. 2) showed elevated concentrations at both 0.5 and 4 hr, suggesting enterohepatic recycling.
Dogs.
The mean Cmax of OLZ was 172 ± 69 ng/ml and occurred between 1 and 3 hr in 3 of the 4 animals tested. The fourth animal had a tmax of 6 hr. The elimination of OLZ from plasma seemed to be biphasic (fig.3), with the terminal phase displaying a half-life of 9.2 ± 1.4 hr.
The mean tmax for radioactivity in plasma was 1 ± 0.0 hr, and the Cmax was 949 ± 296 ng-eq/ml. Plasma radioactivity declined with a mean half-life of 27.6 ± 12.0 hr. The ratio of AUC OLZ to AUC radioactivity was 0.14.
After a single IV dose of OLZ to three dogs, the meanCmax and AUC for OLZ were, respectively, 871 ± 241 ng/ml and 2,633 ± 1,041 ng * hr/ml. The corresponding values for plasma radioactivity were 1,145 ± 195 ng-eq/ml and 18,813 ± 2,598 ng-eq * hr/ml. Thus, after an IV administration, at the tmax OLZ accounted for 76% of the radioactivity, compared with a value of 18% after an oral dose. Because the amount of radioactivity excreted in urine after the oral and IV doses was almost the same (38.4% and 39.7% of the dose), the decreased bioavailability after oral administration is likely due to first-pass metabolism. The ratio of AUC OLZ to 14C AUC was the same as that obtained after oral dosing. The absolute oral bioavailability of OLZ was calculated to be 73%.
Monkeys.
The mean Cmax of OLZ and radioactivity were 60 ± 18 and 757 ± 169 ng eq/ml, and were reached on average within 1.5 hr postdose. Therefore, at the CmaxOLZ accounted for ∼8% of the plasma radioactivity. On the basis of AUC, the fraction of plasma radioactivity represented by OLZ was ∼4%.
The mean elimination half-life of OLZ was 3.4 ± 1.2 hr. The elimination of radioactivity from plasma was biphasic (fig.4), with the initial and terminal phases having half-lives of, respectively, 5.3 ± 0.7 and 98.7 ± 26.5 hr.
In Vitro Plasma Protein Binding.
The plasma protein binding of OLZ was similar in the three species studied, with mean binding being 77%, 75%, and 83% in mice, dogs, and monkeys, respectively. The binding was concentration-independent (10–1,000 ng/ml). The extent of protein binding was lower in these species than that reported for humans at 93% (6).
Metabolism. Mice.
Upon partitioning pooled urine (0–24 hr) between ethyl acetate and water, 19% of the radioactivity was extracted into the ethyl acetate, whereas 76% remained in the aqueous fraction. An aliquot of the aqueous fraction was separated by HPLC with radiochemical detection and yielded the chromatogram in fig. 5. The corresponding HPLC chromatogram from the ethyl acetate extract is shown in fig.6. The individual peaks were collected and analyzed by direct infusion electrospray MS and MS/MS. The following metabolites were identified in urine of mice by comparing their LC and LC-MS/MS properties to those obtained from synthetic standards.
The metabolite that eluted as peak 1 in fig. 5 was identified as 2-carboxy OLZ on the basis of the similarity of its HPLC retention time and product ion spectrum to those obtained from a sample of synthetic 2-carboxy OLZ. The positive ion electrospray mass spectrum of the major urinary metabolite (peak 2, fig. 5) exhibited an MH+ ion at m/z 505, which suggested that the metabolite was the glucuronide of a hydroxylated OLZ derivative (Mr OLZ = 312). The product ion spectrum ofm/z 505 was dominated by the fragment at m/z 329, which is likely due to loss of dehydroglucuronic acid from the conjugate. β-Glucuronidase hydrolysis of the conjugate resulted in 7-hydroxy OLZ, confirming the major metabolite in urine as 7-hydroxy OLZ glucuronide. Peaks 3 and 4 were characterized asN-desmethyl-2-hydroxymethyl OLZ and 2-hydroxymethyl OLZ, respectively, by comparison with authentic standards. Six metabolites (fig. 6) were isolated from the ethyl acetate extract for MS identification. 2-Hydroxymethyl OLZ, which was also present in the aqueous fraction, was identified as the component eluting as peak 1 in fig. 6. The metabolite that eluted as peak 2 had the same HPLC retention volume and MS/MS fragmentation as authentic 7-hydroxy OLZ. The metabolite shown as peak 3 (fig. 6) was identified asN-desmethyl OLZ. Unchanged OLZ was also excreted in urine (peak 4, fig. 6). The identities of the other radiolabeled components in fig. 6 were not confirmed, although MS/MS fragmentation indicated that they were OLZ metabolites. The peak eluting between peaks 2 and 3 had an apparent MH+ ion of m/z327, 2 Da less than that of 2-hydroxymethyl OLZ. This metabolite could possibly be the precursor of 2-carboxy OLZ, the 2-formyl derivative of OLZ.
In plasma, in addition to the parent compound, 2-hydroxymethyl OLZ,N-desmethyl OLZ and the glucuronide of a hydroxy OLZ metabolite were detected. Also, mass spectral data was obtained that indicated the presence of two isomeric glutathione conjugates of OLZ. The conjugates exhibited an MH+ ion at m/z 618, which upon CID fragmentation, gave MH-129+—a characteristic loss of glutathione conjugates (9), in addition to other fragments consistent with the glutathione conjugate of OLZ. The MS/MS data also suggested that the glutathione moiety was attached to one of the carbons of the benzene ring of OLZ.
Six metabolites of OLZ were identified in the urine of mice in addition to the parent compound. Based on the percentage of the radioactivity that was extracted into the ethyl acetate (19%) and the percentage remaining in the aqueous fraction (76%), the amount of urinary radioactivity accounted for by each metabolite was estimated as shown in table 3.
Dogs.
Pooled urine sample from the first 48 hr after dosing was used for metabolite identification. The partitioning of radioactivity between ethyl acetate and water was similar to that obtained for mouse urine with 15% extracted into ethyl acetate and 79% of the radioactivity remaining in the aqueous fraction. The HPLC separation of the radioactive components in the ethyl acetate and aqueous fractions is shown, respectively, in figs. 7 and 8. In the ethyl acetate extract, 2-hydroxymethyl OLZ, N-oxide OLZ, 7-hydroxy OLZ, N-desmethyl OLZ (peaks 1–4; fig.7) were identified in addition to the parent compound.
The major component in the aqueous fraction (peak 3, fig. 8) had a retention time that was different from the available standards. The electrospray MS of this metabolite gave an apparent protonated molecular ion of m/z 345. MS/MS experiments indicated that the metabolite was an N-oxygenated species with a hydroxyl group on the benzodiazepine moiety. Approximately 30 mg of the metabolite was isolated from urine using XAD-2 chromatography and further purified by HPLC fractionation. The 1H- and13C-NMR data obtained for the metabolite are shown in table4. The 1H-NMR of the metabolite showed a downfield shift of the 4′-CH3 to δ 3.08 (δ 2.21 for OLZ) and was identical to the value obtained for N-oxide OLZ. Similarly, the 13C-NMR exhibited a downfield shift of the 4′-CH3 resonance to δ 58.09 (δ 45.73 for OLZ). Two-dimensional nuclear Overhauser enhancement was used to confirm the exact position of the hydroxyl group on the benzene ring of OLZ. The absence of a C-7 proton (δ 6.83 for OLZ) and the fact that the C-9 proton showed ortho coupling only (J = 8 Hz) indicated the hydroxyl group was at the C-7 position. Thus, on the basis of combined MS and NMR data, the major urinary metabolite in dogs was identified as 7-hydroxy-N-oxide OLZ.
The aqueous fraction also contained a metabolite (peak 1, fig. 8) that was identified as 2-carboxy OLZ. In addition to 7-hydroxy OLZ glucuronide (peak 2, fig. 8), a glucuronide of OLZ was identified (peak 4, fig. 8) and characterized as the tertiary N-glucuronide, OLZ 10-N-glucuronide.
LC/MS analysis of the XAD-2 extract indicated the presence of apparent protonated molecular ions at m/z 432 and 448. The product ion spectra of these metabolites indicated that the metabolites might be the cysteine adducts of OLZ and N-oxide OLZ. CID analysis of the ion at m/z 432 resulted in fragment ions atm/z 345 and 311 that could be produced, respectively, from loss of 87 Da as a neutral [CH2⋕C—(NH2)COOH] from the cysteinyl moeity of the conjugate and complete cleavage of the cysteine residue. The ion at m/z 311 further fragmented to an ion atm/z 254. This transition is characteristic of OLZ and metabolites (6) and results from loss of 57 Da as CH2⋕CH—NH—CH3 from the methyl piperazine ring of the molecules. MS/MS analysis (precursor m/z 448) of the putative cysteine conjugate of N-oxide OLZ yielded a fragmentation pattern that was different from that obtained for the corresponding conjugate of OLZ. The fragment at m/z 261 perhaps resulted from the combined loss of 100 Da (scission of the methyl piperazine ring) and 87 Da [CH2⋕C—(NH2)COOH, from the cysteine residue]. The additional loss of possibly hydrogen sulfide resulted in a fragment at m/z 228. A weak ion at m/z 401 resulted from the loss of 47 Da from the methyl piperazine portion of the molecule that is a characteristic fragmentation pathway ofN-oxide OLZ (6).
After 3 hr postdose, plasma contained OLZ, 2-hydroxymethyl,N-oxide, N-desmethyl, and the 7-hydroxy metabolites, as well as the glucuronide of 7-hydroxy OLZ. After 12 hr, the plasma metabolite profile was similar to that obtained at 3 hr, except that the level of N-oxide was lower than that of the 7-hydroxy metabolite and no 7-hydroxy glucuronide was detected.
The relative amount of each metabolite and parent drug in urine was estimated by HPLC with radiochemical detection as detailed in theMaterials and Methods and is presented in table 3. The amount of 7-hydroxy OLZ was estimated from the ethyl acetate extract. The ethyl acetate extract contained 7-hydroxy OLZ; however, this metabolite was not detectable in the XAD-2 extract. The 7-hydroxy metabolite is fairly susceptible to air oxidation and could have decomposed during the lengthy XAD-2 extraction procedure.
Monkeys.
An aliquot of the first 24-hr urine sample from each monkey was concentrated ∼10-fold and used for metabolite identification. A typical HPLC radiochromatogram obtained from the concentrated urine from one animal is shown in fig. 9. The following metabolites were identified in monkey urine by LC-MS/MS and in comparison with authentic standards: N-desmethyl-2-carboxy OLZ (peak 2, fig. 9), N-oxide-2-hydroxymethyl OLZ (peak 4, fig. 9), N-desmethyl-2-hydroxymethyl OLZ (peak 6, fig. 9), and 2-hydroxymethyl OLZ (peak 7, fig. 9).
LC/MS analysis of the metabolite eluting as peak 3 (fig. 9) produced an apparent protonated molecular ion of m/z 505, the CID of which produced major ions at m/z 329, 272, and 84. Loss of 176 Da as a neutral to produce the fragment at m/z 329 suggested that the metabolite was the glucuronide of a hydroxylated OLZ derivative. The isolated material was hydrolyzed to 2-hydroxymethyl OLZ in the presence of Glusulase, confirming this metabolite as the glucuronide conjugate of 2-hydroxymethyl OLZ. The enzyme hydrolysis to the aglycone was inhibited by β-sacchrolactone. 2-Carboxy OLZ coeluted with the glucuronide of 2-hydroxymethyl OLZ (peak 3, fig. 9) under the HPLC conditions used. The 2-carboxy metabolite was identified after β-glucuronidase treatment of a sample of the material that eluted as peak 3 (fig. 9), which resulted in the production of 2-hydroxymethyl OLZ, a compound with a substantial difference in retention time from that of 2-carboxy OLZ.
The HPLC radiochromatogram from urine contained a metabolite (peak 1, fig. 9) with a retention time that was different from any of the available synthetic metabolite standards. LC/MS analysis of the metabolite yielded an apparent protonated molecular ion ofm/z 359 that afforded a product ion spectrum with major ions at m/z 312, 272, 259, 243, 199, 153, 135, and 85. The characteristic loss of 47 Da to produce the fragment at m/z312 indicated this metabolite was an analog of N-oxide OLZ (6). The additional 30 Da (Mr ofN-oxide OLZ = 328) suggested that the methyl group on the thiophene ring was oxidized to the carboxyl derivative. The fragment at m/z 199 is indicative of the presence of a carboxyl moiety at this position (unpublished observation). Thus, based on its distinct product ion spectrum and HPLC order of elution, this metabolite was tentatively assigned the structure ofN-oxide-2-carboxy OLZ.
The HPLC chromatogram also contained another metabolite (peak 5, fig. 9) that did not match the LC-MS/MS properties of any available standards. Analysis of the metabolite by LC/MS afforded an apparent protonated molecular ion at m/z 474, the CID of which yielded a spectrum with major ions at m/z 345, 311, 288, 254, and 84. This product ion spectrum suggested that the metabolite was an N-acetylcysteine conjugate of OLZ. The ions at m/z 345 and 311 are likely formed as a result of loss of 129 Da [CH2⋕C—(NHCOCH3)—COOH] from the cysteinyl moiety of the metabolite and complete cleavage of the cysteinyl residue, respectively. The ion at m/z 288 seems to result from combined loss of 57 Da [CH2⋕CH—NH—CH3] from the methyl piperazine ring and [CH2⋕C—(NHCOCH3)—COOH] from the cysteinyl moiety. The ion at m/z 311 further fragmented by losing 57 Da to give the ion at m/z 254; a typical fragmentation pathway for OLZ and metabolites (6).
The level of each metabolite was estimated by HPLC, and the result is shown in table 3. N-Desmethyl-2-carboxy OLZ was found to be the major metabolite in urine accounting for 36% of the radioactivity or 17% of the dose.
Discussion
Orally administered OLZ was well absorbed in dogs (absolute bioavailability: 73%) and to the extent of at least 55% in monkeys and 32% in mice, as demonstrated by the radioactivity recovered in urine. The elimination half-life of OLZ was relatively short in mice and monkeys (∼3 hr) and long in dogs (∼9 hr). The half-life of the drug in animals was much shorter than that reported for humans (∼27 hr). The longer half-life for radioactivity in the three animals (figs. 2, 3, 4) indicated the formation of long-lived metabolites in these species. In mice and dogs radioactivity was predominantly eliminated in feces; but, in monkeys, as with humans (6), the major route of elimination of radioactivity was urine. For the present study, male mice, female dogs, and male and female monkeys were used. This was not a concern, because there were no sex-related differences in the plasma concentrations of OLZ. Also, OLZ pharmacokinetics were found not to be dose-dependent over the range used in this study.2
OLZ was subject to substantial first-pass metabolism, as evidenced by the fraction of the plasma AUC accounted for by OLZ compared with total radioactivity. At the tmax, OLZ accounted for 19%, 18%, and 8% of the radioactivity, in mice, dogs, and monkeys, respectively. A similar relationship was evident on the basis of AUC comparisons also. OLZ accounted for 10%, 14%, and 4% of the plasma AUC, respectively, in mice, dogs, and monkeys, indicating that OLZ was metabolized most extensively by the monkey and least by the dog among the three species. This observation is also reflected in the elimination half-life of the compound in these species (table 2). The ratio of AUC OLZ to AUC radioactivity found for the dog was the same as that reported for humans (6). Although many metabolites were present in plasma, attesting to the varied and extensive metabolism, the parent compound was the largest single component in plasma of all species studied, including humans.
Of the dose administered to mice, ∼47% was excreted in the feces, whereas ∼25% appeared in the urine within 24 hr of dosing. The principal urinary metabolites were 7-hydroxy OLZ glucuronide (fig.10), 2-hydroxymethyl OLZ, and 2-carboxy OLZ accounting for ∼10%, 4%, and 2% of the dose, respectively. Metabolites that were present in urine in lesser amounts were 7-hydroxy OLZ,N-desmethyl OLZ, and N-desmethyl-2-hydroxymethyl OLZ. In mouse plasma, mass spectral evidence was obtained for the presence of two isomeric glutathione conjugates of OLZ, although the corresponding N-acetylcysteine or cysteine conjugates were not detected in urine. Thus, on the basis of urinary metabolites, the principal metabolic pathways of OLZ in mice were aromatic hydroxylation, allylic oxidation (2-alkyl oxidation), andN-demethylation.
In dog urine (urinary excretion accounted for ∼38% of the dose), the major urinary metabolite accounting for ∼8% of the dose was 7-hydroxy-N-oxide OLZ (fig. 10). Other metabolites identified in urine were 2-hydroxymethyl OLZ (∼3% of the dose), 2-carboxy OLZ (∼3% of the dose), N-oxide OLZ (∼1% of the dose), 7-hydroxy OLZ (∼1% of the dose), 7-hydroxy OLZ glucuronide (<1% of the dose), and N-desmethyl OLZ (<1% of the dose). The cysteine adducts of N-oxide OLZ and OLZ were also tentatively identified in urine extracts and were estimated to represent 2% and 1% of the dose, respectively. In dogs, therefore, OLZ underwent metabolic transformations mainly via aromatic hydroxylation, N-oxidation and 2-alkyl oxidation. The detection of the putative cysteine adducts implied the possible formation of their precursor glutathione conjugates. 2-Hydroxymethyl,N-oxide, N-desmethyl, 7-hydroxy OLZ, and its glucuronide were detected as circulating metabolites.
Approximately 55% of an oral dose given to monkeys was excreted in urine over a 7-day period, with ∼48% appearing within the first 24 hr. The major urinary metabolite accounting for ∼17% of the dose was a product of multiple oxidation, N-desmethyl-2-carboxy OLZ (fig. 10). Other metabolites identified in urine wereN-oxide-2-hydroxymethyl OLZ (∼6% of the dose),N-oxide-2-carboxy OLZ (∼4% of the dose),N-desmethyl-2-hydroxymethyl OLZ (∼4% of the dose), 2-carboxy OLZ (∼3% of the dose), 2-hydroxymethyl OLZ (∼3% of the dose), the N-acetylcysteine conjugate of OLZ (∼1% of the dose), and 2-hydroxymethyl OLZ glucuronide (∼1% of the dose). Thus, OLZ was biotransformed by the monkey mainly through double oxidation reactions involving the allylic carbon and the methyl piperazine nitrogen. Metabolism in the monkey was so efficient that intermediary metabolites such as N-oxide and N-desmethyl OLZ were not detected. Metabolism in the monkey was entirely driven by the allylic carbon attached to the thiophene ring, such that no oxidative metabolite was identified in which the methyl thiophene moiety was not modified.
Thus, in mice and dogs, OLZ was metabolized through aromatic hydroxylation (forming phenolic metabolites and/or their glucuronide conjugates), allylic (alkyl) oxidation, thiol conjugation, andN-oxidation reactions (fig. 10). However, there were substantial species differences in the biotransformation of OLZ. The N-oxidation pathway was absent in mice. The 7-hydroxy-N-oxide metabolite, which is a product of aromatic and N-oxidation reactions, was formed only in the dog. The monkey was studied in efforts to get an animal model that produced OLZ 10-N-glucuronide (a major metabolite in humans). Although the monkey did not seem to form 10-N-glucuronide, the pattern of oxidative metabolism was similar to that of humans. In contrast to the other species studied, including humans, no intact OLZ was excreted in monkey urine. Among the four animal species studied, the monkey was unique in apparently not forming metabolites resulting from the oxidative attack of the benzene ring of OLZ.
Similarities in the metabolic fate of OLZ in animals and humans include the 2-alkyl hydroxylation, N-dealkylation, andN-oxidation pathways. Notable differences were that, direct glucuronidation, producing mainly the 10-N-glucuronide and to a lesser extent 4′-N-glucuronide, was the principal metabolic pathway in humans (6). These N-glucuronides were absent in animal samples, with the exception of a trace amount of 10-N-glucuronide in dog urine. Also, aromatic hydroxylation that was found to be a principal determinant of the clearance of the drug in mice and dogs did not seem to be an important pathway in humans.
In summary, orally administered OLZ was well absorbed and extensively metabolized by all species studied. Greater than 20 metabolites were identified in the four species studied, with the major metabolite in urine being strictly species dependent.
Acknowledgments
We thank Larry Spangle for the NMR data of 7-hydroxy-N-oxide OLZ.
Footnotes
-
Send reprint requests to: Dr. Kelem Kassahun, Department of Drug Metabolism, Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Mail Drop 0825, Indianapolis, IN 46285.
-
↵2 R. Franklin, unpublished observations.
- Abbreviations used are::
- OLZ
- olanzapine
- LSC
- liquid scintillation counting
- Cmax
- maximum plasma concentration
- tmax
- time to maximum concentration
- AUC
- area under the plasma concentration-time curve
- LC-MS/MS
- liquid chromatography-tandem mass spectrometry
- CID
- collision-induced dissociation
- IV
- intravenous
- Received September 12, 1996.
- Accepted January 30, 1997.
- The American Society for Pharmacology and Experimental Therapeutics