Estimation of the Extent of in Vivo Formation of a Mutagenic Aromatic

ARTICLE pubs.acs.org/crt

Estimation of the Extent of in Vivo Formation of a Mutagenic Aromatic Amine from a Potent Thyromimetic Compound: Correlation of in Vitro and in Vivo Findings Kamelia Behnia,* Georgia Cornelius, Jian Wang, Petia Shipkova, Susan Johnghar, William Washburn, Robert Brigance, Paul Stetsko, Andrew Henwood,† James P. Wojciechowski,† Punit Marathe, A. David Rodrigues, and W. Griffith Humphreys Department of Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Research and Development, Pennington, New Jersey 08534, United States † Department of Genetic Toxicology, Bristol-Myers Squibb Drug Safety Evaluation, E. Syracuse, New York 13057, United States ABSTRACT: The development of compounds with the potential for genotoxicity poses significant safety risks as well as risks of attrition. Although genotoxicity evaluation of the parent molecule is routine and reasonably predictive, assessing the risk of commercialization when release of a genotoxic degradant and/or metabolite from a nongenotoxic parent molecule is suspected is much more challenging and resource intensive. Much of the risk of the formation of a genotoxic degradant/ metabolite can be discharged with the conduct of carcinogenicity studies in models where the compound is formed, but this approach requires a great deal of time and resources. In this manuscript, we investigated the contribution of various factors (pH, serum instability, and hepatic metabolism) to the formation of a mutagenic aromatic amine from a potent and highly selective thyromimetic compound ([3-(3,5-dibromo-4-(4-hydroxy-3isopropyl-5-methylphenoxy)-2-methylphenylamino)-3-oxopropanoic acid], compound 1), under in vitro conditions. The kinetic parameters obtained from in vitro experiments combined with the pharmacokinetics of 1 in vivo (e.g., plasma concentration
time profile and clearance) were used to estimate the extent of in vivo formation of [4-(4-amino-2,6-dibromo-3-methylphenoxy)-2isopropyl-6-methylphenol] (compound 2), in rats upon administration of a single oral dose of 1. The agreement between the predicted values (1.9% conversion of total administered dose) with the observed levels of 2 in rats (0.2%-2.2% of the 10 mg/kg dose, 10 mg/kg) further prompted the utilization of this approach to predict the extent of release of this mutagen in humans upon administration of 1. The projection of 0.13% conversion to 2 from an efficacious daily dose of 15 mg of 1 translated to the generation of 20 μg of 2 and provided the basis for the decision to terminate the development of 1.

’ INTRODUCTION Genotoxicity and mutagenicity are a significant cause of compound attrition during drug discovery and lead optimization.1
3 Early genotoxicity screening methods such as bacterial reversemutation (Ames assays) and bacterial SOS reporter assays (SOS chromotest) are effective methods to ascertain the induction of gene mutations and DNA damage of parent drug molecules.4,5 The challenge to the pharmaceutical industry continues to be devising cost beneficial procedures that can reliably reveal the inherent human risk associated with potentially genotoxic compounds at different stages of the development process. Many factors must be considered in the overall risk assessment, including therapeutic area and potential to treat life-threatening disease, life-expectancy of the patients, and whether viable alternative treatments exist. Adding to the challenge is that the full characterization of the genotoxic potential of new compounds must include the characterization of impurities, degradants, and metabolites. In this article, we describe a method r 2011 American Chemical Society

employed to assess the potential risk to humans which would be incurred by the development of a compound prone to conversion to yield trace levels of a genotoxic degradant/metabolite by estimating the extent of formation of this mutagen in vivo. There are three ICH documents that provide general guidance on identification and safety characterization of organic and inorganic impurities in drug substances,6 degradation products,7 and residual solvents.8 According to ICH, there should be at a minimum an assessment of general toxicity and genotoxicity, the latter involving tests for point mutations and chromosomal aberrations. While the Q3A(R2)5 states that identification of impurities at levels below the threshold level (the lower of either 0.1% or 1 mg/day for a drug dose of up to 2 g/day) is generally not necessary, the European Medicines Evaluation Agency (EMEA) guideline on limits of genotoxic impurities, which came Received: February 24, 2011 Published: May 16, 2011 905

dx.doi.org/10.1021/tx200087q | Chem. Res. Toxicol. 2011, 24, 905–912

Chemical Research in Toxicology

ARTICLE

its analogues on the basis of estimated levels of exposure to 2 in humans upon daily administration of 1. To determine the extent of formation of 2 in humans, a series of in vitro studies were conducted to determine the contribution of pH, serum, and hepatic hydrolytic metabolism to its formation. Physiological modeling and simulations were used to predict the levels generated in rats following the oral administration of 1. This article provides details of these studies that ultimately led to a prediction of the extent of formation of 2 in humans and became the basis for a decision regarding the development path of 1.

Figure 1. Structures of compounds 1 and 2.

into force in January 2007, suggest a generic limit of 1.5 μg/day for a genotoxic impurity8,9 and describes the “Threshold of Toxicological Concern” (TTC) as the “Threshhold of Regulation”. With significant improvements in analytical technology, it became clear that previously unquantifiable trace levels of numerous substances known to be animal carcinogens were now in fact detectable at low levels. For clinical drug development, the staged TTC10
13 represents an approach to control genotoxic impurities during the development phase by taking steps for the prevention of impurity formation, reduction of the impurity level to an acceptable threshold, and additional characterization of the genotoxic and carcinogenic risk. While there has not been the same level of guidance given to the formation of mutagenic metabolites, the same principles might be expected to be applicable.14,15 A carcinogenicity risk assessment scheme was proposed by Dobo et al.14 for human genotoxic metabolites using quantitative human and rodent ADME data to estimate the exposure to the metabolite of genotoxic concern at the intended human efficacious dose and the maximum dose used in the 2-year rodent carcinogenicity studies. These exposures were applied to this scheme, on the basis of known cancer potencies, to arrive at the probability of the genotoxic metabolite posing carcinogenic risk in excess of 1 in 100,000. The situation is additionally complicated when a process impurity is also formed as a metabolite through enzymatic conversion. Compound 1 (Figure 1) is a potent and highly selective thyroid agonist with the potential to be used for weight reduction. During the course of its development, structural integrity assays found a trace impurity in 1, which was later identified as an aromatic amine (2, Figure 1). Compound 2 was the precursor to 1 via acylation with malonic acid during the synthetic procedure. It was hypothesized that hydrolysis of the amide bond either due to chemical processes or metabolic enzymes released 2 (Figure 1). Compound 2 was found to be genotoxic and Ames positive following incubation in the presence of mixed function oxygenases from the liver tissue fraction from Aroclor-induced rats (S9). Aromatic amines are prone to form reactive metabolites such as conjugated N-hydroxylamine and/or nitroso derivatives, due to cytochrome P450 (CYP) metabolism. CYP mediated metabolism of aromatic amines has been demonstrated to generate N-hydroxylamines, which can subsequently be conjugated or form nitroso species that may then eliminate the O-conjugated moiety to produce a reactive nitrenium ion or react directly (nitroso species). Both reactive species are capable of covalently binding to cellular nucleophiles such as DNA.16
18 Although amounts of the aromatic amine impurity were subsequently reduced, concerns were raised whether hydrolysis of the amide bond of 1 would occur in vivo in humans leading to increased exposure to the mutagenic 2. Therefore, it was important to make a decision about the development of 1 and

’ EXPERIMENTAL PROCEDURES Materials. Compounds 1 [3-(3,5-dibromo-4-(4-hydroxy-3-isopropyl5-methylphenoxy)-2-methylphenylamino)-3-oxopropanoic acid] and 2 [4-(4-amino-2,6-dibromo-3-methylphenoxy)-2-isopropyl-6-methylphenol] were synthesized at Department of Medicinal Chemistry, Bristol-Myers Squibb (US patent # 7,342,127). Sodium taurocholate, potassium phosphate monobasic, potassium chloride, lecithin, acetic acid, hydrochloric acid, and citrate buffers were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). Fresh serum from five species (CD1 mouse, SD rat, beagle dog, cynomolgus monkey, and human, pooled) was purchased from Bioreclamation (Hicksville, NY). Liver microsome and liver S9 fractions from five species (CD1 mouse, SD rat, beagle dog, cynomolgus monkey and pooled human) were purchased from BD Gentest (Woburn, MA). For the Ames assays, positive controls, Aroclor1254 induced S9, salt, and cofactors were purchased from Molecular Toxicology, Inc., (Boone, NC). All other chemicals and organic solvents were of analytical grade and purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). The purity of 1 (with respect to presence of 2) and the levels of 2 in the dosing solution used in the in vivo studies are reported in the Results section related to BDC rat studies. The batch of 1 that was used in the Ames assay was >99.99% pure. The level of 2 in the stock solution of 1 used in the hepatic metabolism studies was 0.02%, and it was 0.001% in the stock solution that was used in the pH stability studies. Animals. Sprague
Dawley rats (200
300 g) were purchased from Charles River Laboratories (Wilmington, MA) and were housed and maintained under the standard conditions recommended by Animal Care and Use Committee (ACUC) in a 12 h light/12 h dark, constanttemperature environment with free access to rat chow and drinking water. Animals were acclimatized to this environment for 1 week prior to surgery. Animals were subjected to bile duct and/or jugular vein cannulation under light anesthesia. For the bile duct cannulated rat studies, bile was collected 24 h prior to experiment and was infused into the duodenum via a duodenal cannula during the experiment. For oral studies, rats were fasted overnight and fed 4 h postdose. Ames Assay. The Ames reverse-mutation plate incorporation assays were conducted using standard procedures.19 For the exploratory Ames assays, duplicate plates of strains TA98 and TA100 with and without S9 were tested. The S9 mixture contained 10% Aroclor induced rat liver S9 extract (v/v) in 33 mM KCl, 8 mM MgCl2, 5 mM glucose-6phosphate, and 4 mM NADP in 100 mM phosphate buffer. The test article carrier and vehicle control was dimethyl sulfoxide (DMSO). PH-Dependent Stability Studies. Phosphate and citrate buffers (0.01 M) and 0.08 N hydrochloric acid were used to attain solutions with pH ranging from 1 to 7. Simulated gastric (SIG) and intestinal fluids (SIF) were prepared according to the published literature.20,21 Solutions of 1 were prepared in the mixture of phosphate or citrate buffer (0.01 M, pH 4 to 7) and acetonitrile or HCl (0.01 and 0.08 M, pH 1 to 2) and acetonitrile containing a final concentration of 1 mg/mL (for pH 4
7) or 0.5 mg/mL (for pH 1
2). Additional stability studies were carried out in simulated gastric (SGF, pH 1.2) and intestinal fluids (SIF, pH 6.8) spiked with 1 at similar concentrations. Incubations were carried out for 906

dx.doi.org/10.1021/tx200087q |Chem. Res. Toxicol. 2011, 24, 905–912

Chemical Research in Toxicology

ARTICLE

24 h at 37 °C (n = 2 or 3). Serial samples were collected at different time points and analyzed for the concentration of 1 and 2 by LC/MS/MS analysis. Stability Studies in Fresh Serum. Stability of 1 was tested in fresh serum from five species (mouse, rat, dog, cynomolgus monkey, and human). Fresh serum was spiked with drug to achieve a final concentration of 10 or 100 μM and incubated at 37 °C for 4 h. Serial samples were collected and centrifuged after the addition of acetonitrile. The resulting supernatants were analyzed for the concentration of 1 and 2 by LC/MS/MS analysis.

tandem mass spectrometer (Applied Biosystems, USA) equipped with an electrospray ionization source. Data acquisition utilized selected reaction monitoring (SRM). Quantitation involved separate calibration curves to minimize contamination. Standard curves and quality control (QC) samples, defining the dynamic range of the bioanalytical method, were prepared in the respective biological matrix and processed in the same fashion as the study samples. All standard curve, QC, and study samples were precipitated with acetonitrile (1:2, v/v) containing an internal standard (IS), vortexed, and centrifuged. An aliquot of the resulting supernatant was injected and analyzed by LC/MS/MS. Calculations. The rate of formation of 2 by serum hydrolysis (dB/dt) was estimated from the in vitro incubations of the parent drug with fresh serum. The rate of conversion to 2 was assumed to follow first order kinetics (that is the concentrations of 1 used in these studies were below the Km value) and was estimated according to the following equation:

Stability Studies in Liver Microsomes and S9 Fractions. Compound 1 (10 or 100 μM) was incubated with NADPH (2 mM)fortified liver microsomes or with liver S9 fractions from five species (mouse, rat, dog, cynomolgus monkey, and human) at a final protein concentration of 1.0 mg/mL. The incubations were carried out for 2 h in a humidified (5% CO2) incubator at 37 °C. Reactions were stopped by the addition of equal volumes of acetonitrile. After centrifugation, the resulting supernatants were analyzed for the concentration of 1 and 2 by LC/MS/MS analysis. Bile Duct Cannulated (BDC) Rat Studies. A total of four BDC rat studies were conducted and the number of animals in each study ranged from 2 to 4 rats. The drug was administered orally (10 mg/kg, 5 mL/kg) in a mixture of PEG 400/water (1:1, study #1 and 2) or PEG 400/potassium phosphate buffer (100 mM, pH 7.0)/water (1:0.1:0.9, study #3 and 4) to BDC male Sprague
Dawley rats (200
300 g). One rat, in study #3, died immediately after dosing due to unknown reasons. Serial blood, urine, and bile samples were collected over a period of 9 h postdose. The entire gastrointestinal (GI) tract and livers were collected at the end of the study from some rats. Blood samples were collected via the jugular vein into K3EDTA-containing tubes and centrifuged to obtain plasma. Acetonitrile was added to plasma samples to precipitate the proteins. The GI tract and livers were homogenized with three volumes of water and then extracted using acetonitrile. Following the addition of acetonitrile, all of the samples were centrifuged, and the resultant supernatants were analyzed for the concentrations of 1 and 2 by LC/MS/MS analysis. Single Dose Pharmacokinetics in Rats. The pharmacokinetic profile of 1 was evaluated in male Sprague
Dawley rats (200
300 g) following a single intra-arterial (5 mg/kg) (n = 3) or oral (10 mg/kg) dose (n = 3). The vehicle used in these studies was PEG 400/water (1:1). Serial blood samples were collected via the jugular vein into K3EDTAcontaining tubes over a period of 24 h postdose and centrifuged at 4 °C (1500
2000g) to obtain plasma. Plasma samples were then treated with acetonitrile, and after centrifugation, supernatants were analyzed for the concentration of the drug by LC/MS/MS analysis. Pharmacokinetic parameters, such as total body clearance, were estimated by noncompartmental analysis using Kinetica software (KINETICA software, version 4.2, InnaPhase Corporation, Philadelphia, PA). HPLC/MS Methods. Data to support the pH-dependent stability assessment were collected on a Waters Ultima triple quadrupole mass spectrometer interfaced with a Waters 2790 HPLC system. Solvent A was 10 mM ammonium formate/0.1% formic acid in 15:85 acetonitrile/ water, and Solvent B was 10 mM ammonium formate and 0.1% formic acid in 95:5 acetonitrile/water. The HPLC column was a Luna C18, 2.0  30 mm with a flow rate of 200 μL/min without splitting prior to MS analysis. The response for 1 saturated the detector and, therefore, was not collected. Compounds 1 and 2 were chromatographically separated to avoid any contributions from in-source fragmentation of 1. Trace amounts of 2 were detected in all samples at time zero as an impurity. The HPLC system used for analysis of the rest of the samples consisted of two low pressure mixing flux pumps (Cohesive, USA) and a CTC Analytics HTS PAL autosampler (CTC Analytics, Switzerland) equipped with a cooling stack maintained at 5 °C. The column was a Waters Atlantis C18 (2.0  50 mm, 3 μm particle (Waters Corp., Milford, MA), and the mobile phase was 0.1% formic acid in acetonitrile/water. The HPLC was interfaced to a Sciex API3000

k

A sf B dB=dt ¼ kA

where A and B are concentrations of 1 and 2, respectively, and k is the first order rate constant for the formation of 2. Because of negligible changes in the concentration of the parent drug and slow rate of conversion to 2 during incubations with serum, the rate of formation of 2 (i.e., dB/dt) was estimated from the regression analysis of the concentration
time profile of 2 in these incubations. This rate was then used with the starting concentration of the parent drug (i.e., A, 100 μM) to arrive at rate constant k. In order to predict the contribution of serum hydrolysis to the formation of 2 from 1 in vivo in rats, the serum concentration
time profile of 1 in rats (following a single oral dose of 1 at 10 mg/kg) was used to calculate the average serum concentrations of 1 within the time intervals of 0
0.25 h, 0.25
0.5 h, 0.5
1 h, 1
2 h, 2
4 h, etc. assuming constant concentration of 1 during each time interval. Average serum concentrations of 1 within each time interval (for example, 0
0.25 h, 0.25
0.5 h, etc.) were multiplied by the total volume of serum in rats (i.e., 7.8 mL for 0.25 kg) to obtain amounts of 1 in serum. The in vitro rate constant k (calculated above) was used to calculate the amount of 2 formed from 1 during each time interval and to simulate the amount of 2 formed over time from 1. The percent of the administered dose converted to 2, by serum enzymes, was then calculated from the cumulative amount of 2 (over the course of 48 h) as calculated above and a dose of 1 (i.e., 2.5 mg for 0.25 kg rat) as follows: % dose converted to 2 by serum hydrolysis ¼ ðcum:amt:2Þ=dose To determine the contribution of hepatic metabolism to the formation of compound 2, the rate constant (k) for the formation of compound 2 in liver microsomes was determined by regression analysis. Following normalization for the volume of the incubation and amount of microsomal protein, the hepatic intrinsic clearance of 1 for conversion to 2 was calculated as follows:22
24 ! ! ! k 45 mg proteins g of liver weight   CLh , int ¼ Cprotein 1 g liver weight kg of body weight where Cprotein (mg/mL) is the final concentration of microsomal protein in the incubation. The liver weight-to-body weight in rats and humans was taken as 40 and 21 g/kg, respectively.21 The term CLhb (mL/min/kg), hepatic formation clearance of 2 from parent drug, was estimated using the well-stirred model as follows: CLh b ¼

Qh  CLint Qh þ CLint

where, Qh is the hepatic blood flow in rats and humans (55 and 21 mL/ min/kg, respectively).25,26 Free fractions were not accounted for in these 907

dx.doi.org/10.1021/tx200087q |Chem. Res. Toxicol. 2011, 24, 905–912

Chemical Research in Toxicology

ARTICLE

Figure 2. Mutagenic activity of compounds 1 and 2 in Salmonella strain TA98 assay (with and without S9).

Figure 3. Formation of 2 from 1 during incubations with different pH environments, including SGF and SIF.

calculations on the basis of both in house (not published) and literature data suggesting a better prediction of the in vivo clearances without adjustment for plasma protein binding values.27,28 The percent of the administered dose converted to 2, by hepatic enzymes, was calculated from the ratio of the hepatic formation clearance of 2 (CLhb) to the total body clearance (TBC) of 1 in rats as follows:

Table 1. Percent Conversion (%/h) of 1 to 2 in Serum, Liver Microsomes and Liver S9 Fractions (at 100 μM concentrations of 1)

% dose converted to 2ðby hepatic metabolismÞ ¼ CLh b=TBC The overall percent dose converted to 2 in rats, after a single oral dose at 10 mg/kg, was calculated as the sum of the % dose converted by serum and by hepatic metabolism. This final value was then compared to the in vivo values obtained from the BDC rat studies. Because of minimal contribution of pH to the conversion to 2, the pH effect was not accounted for in these calculations.

species

serum

liver microsomes

liver S9 fraction

mouse

0.19

0.33

0.22

rat

0.02

0.05

0.04

dog

stable

0.014

0.004

monkey

0.01

0.21

0.3

human

stable

0.04

0.03

Role of Enzymatic Pathways in the Formation of Compound 2. In vitro studies were conducted to investigate the

contribution of hydrolytic enzymes in serum to the formation of 2. The rate of formation of 2 was negligible when 1 was incubated with serum from five species was at 10 μM. Incubation at a higher concentration of parent drug (100 μM) suggested that the rate of hydrolysis in serum was species-dependent. Compound 1 was completely stable in dog and human serum but relatively unstable in the cyno, mouse, and rat serum (Table 1). The rate of hydrolysis of 1 was highest in the mouse serum (0.2% per h), followed by rat (0.02% per h) and cynomolgus monkey sera (0.01% per h). Compound 1 (Figure 1) was relatively stable in the incubations with liver microsomes. The only metabolite observed in these incubations was 2. Except for the dog, formation of 2 (Figure 1) in mouse, rat, cynomolgus monkey, and human microsomes was similar to that observed with the S9 fraction (Table 1). The overall formation of 2 in the mouse and cynomolgus monkey microsomes (0.2 to 0.3% and 0.22 to 0.26% per h, respectively) was higher than that in dog and human microsomes (0.004 to 0.014 and 0.028 to 0.036% per h, respectively) (Table 1). In Vivo Formation of Compound 2 in BDC Rats. Several BDC rat studies were carried out to investigate the extent of formation of 2 in vivo. Higher levels of 2 as an impurity were found in the dosing solution in study #1 (0.14%) as compared to studies #2, 3, and 4 (0.003 to 0.01%). Biliary excretion was the main route of elimination of 2 in rats. For example, the cumulative amount of 2 excreted in the bile

total % dose converted to 2 ¼ % dose converted to 2ðby hepatic metabolismÞ þ % dose converted to 2 by serum hydrolysis

’ RESULTS Ames Assay. Compound 1 was not mutagenic in the exploratory Ames assay in strains TA98 and TA100, with or without S9 activation, when tested at concentrations of 5 to 5000 μg/plate. In contrast, 2 was mutagenic in two independent experiments with tester strain TA98 in the presence of S9 metabolic activation (Figure 2). Increases in histidine revertants/plate (g2-fold concurrent vehicle control) were observed at concentrations g500 μg/plate (Figure 2). Nonenzymatic Formation of Compound 2. Results from the pH-dependent stability studies suggested that the release of 2 was pH-dependent and increased under acidic conditions (Figure 3). Rate of formation of 2 at pH 1 was 7-fold higher than at pH 7. Rate of hydrolysis of 1 at pH 1 was 0.001% per hour resulting in an increase in the concentration of 2 from 5 to 10 ng/ mg (0.001%, present in the stock solution) to 35
50 ng/mg. In simulated gastric (SGF) and intestinal fluids (SIF), the rate of release of 2 was similar to what was observed due to the effect of pH alone (Figure 3). 908

dx.doi.org/10.1021/tx200087q |Chem. Res. Toxicol. 2011, 24, 905–912

Chemical Research in Toxicology

ARTICLE

Table 2. Levels of 2 Found in Bile Duct Cannulated Rats (as % Dose)a study #

vehicle

plasmab

bile

urine

GI tract

liver

#1 (N=4)c

PEG 400: H2O (1:1)

1
16

0.15
2.1

< LLQ

0.16
0.4

ND

#2 (N=2)

PEG 400: H2O (1:1)

< LLQ

0.1
0.14

ND

ND

ND

#3 (N=1)

PEG 400: KPO3 buffer: water (1:0.1:0.9)

5
36

0.74

0.07

0.28

0.26

#4 (N=3)

PEG 400: KPO3 buffer: water (1:0.1:0.9)

< LLQ

0.12

< LLQ

0.16

ND

a

Dose was 10 mg/kg. LLQ: lowest limit of LC/MS quantitation (i.e. 1 ng/mL). ND: not detected. b Plasma levels are in ng/mL, not % dose. c A higher level of 2 impurity was found in dosing solution in study #1 (0.14%).

Figure 6. Plasma concentration
time profile of 1 after a 10 mg/kg oral dose to rats.

Figure 4. Plasma concentration
time profiles of 1 (parent) and 2 in bile duct cannulated rats (study number 3, N = 1).

solution did not have a significant impact on the amounts of 2 found in 2 out of the 3 rats, confirming minimal formation of 2 due to the pH effect alone (Table 2). Compound 2 was found only sporadically in the serum of BDC rats at concentrations ranging from 1 to 36 ng/mL. In one rat (study #3), a complete serum concentration
time profile of 2 could be characterized (Figure 4) with an area under the plasma concentration
time curve (AUC) of 161 ng 3 min/mL and a Cmax of 36 ng/mL. The ratio of the Cmax or AUC of 2 to its 1 was 0.01 (∼1%). Predicting the Formation of Compound 2 in Rats. Physiological modeling was used to estimate the extent of formation of 2 in vivo in rats. While 1 was stable in human serum, it was hydrolyzed to form 2 in rat serum. The rate of formation of 2 in rat serum was estimated to be 0.0207 μM/h (Figure 5). This rate, along with the concentration of 1 (100 μM), was used to calculate the first order rate constant (k) for the formation of 2 in rat serum (0.000207 h
1). Thereafter, the plasma concentration
time profile of 1 (Figure 6), total volume of plasma in rats, and the rate constant for the formation of 2 in rat serum (0.000207 h
1) were used to simulate the cumulative amount of 2 formed over the course of 48 h following a single dose of 1 to rats (Figure 7). On the basis of these calculations, the cumulative amount of 2 formed in rats by serum hydrolysis was estimated to be 1.7% of an administered dose of 10 mg/kg (i.e., 2.5 mg for 0.25 kg rat). The rate of formation of 2 in rat liver microsomes was used to calculate the intrinsic formation clearance and hepatic formation clearance of 2 in rats (Table 3). In vivo clearance in rats was determined independently following intravenous administration of 1 (3.8 mL/min/kg, Table 3). The contribution of hepatic formation clearance of 2 to the overall clearance of 1 was calculated to be 0.23% in rats (Table 3). The overall percent conversion to 2 through nonenzymatic (i.e., pH, considered to be negligible as compared to enzymatic

Figure 5. Rate of formation of 2 from 1 in incubations with rat serum.

ranged from 0.1 to 2.1% of the administered dose of 1 (Table 2). In comparison, the concentration of compound 2 in urine from study #1, 2, and 4 was below the limit of quantitation (