Obtaining Exposures of Metabolites in Preclinical Species through

Chem. Res. Toxicol. 2009, 22, 311–322

311

Obtaining Exposures of Metabolites in Preclinical Species through Plasma Pooling and Quantitative NMR: Addressing Metabolites in Safety Testing (MIST) Guidance without Using Radiolabeled Compounds and Chemically Synthesized Metabolite Standards Karthick Vishwanathan, Kathlene Babalola, Jack Wang, Robert Espina, Linning Yu, Adedayo Adedoyin, Rasmy Talaat, Abdul Mutlib,* and JoAnn Scatina Biotransformation, Drug Safety and Metabolism, Wyeth Research, 500 Arcola Road, CollegeVille, PennsylVania 19426 ReceiVed September 4, 2008

The recent guidance on “Safety Testing of Drug Metabolites” issued by the U.S. Food and Drug Administration, Center for Drug Evaluation and Research (CDER) has highlighted the importance of identifying and characterizing drug metabolites as early as possible in drug discovery and development. Furthermore, upon identifying significant circulating metabolites in human plasma, it has become important to demonstrate that these metabolites are present at an equal or greater exposure level (area under the curve, AUC) in any one of the preclinical species used in safety testing. Frequently, synthetic standards of metabolites are not available, and hence, obtaining their AUC values can be a challenge. In this report, we demonstrate how combinations of nuclear magnetic resonance (NMR) spectroscopy, liquid chromatography/ultraviolet/mass spectrometry (LC/UV/MS), and plasma pooling methods were used to obtain reliable AUC values of metabolites present in the plasma of preclinical species from short-term safety studies. Plasma pooling methods were compared to the traditional approaches of obtaining quantitative information on the levels of circulating metabolites in preclinical species. The exposure values obtained via sample pooling were comparable to those obtained by traditional methods of analyzing samples individually. In the absence of synthetic chemical standards, calculations of AUC values of metabolites, using either sample pooling or traditional approaches, were achieved through the use of UV detectors. In cases where the UV properties of metabolites were significantly different from their parent compounds, NMR was used as a quantitative tool to obtain exposure values. NMR was found to be useful in quantitating biologically produced metabolites, which could subsequently be used as reference compounds for further quantitative studies. The limitations of UV detectors to obtain exposure estimates are discussed. A practical solution is presented that will enable us to obtain a quantitative assessment of metabolite exposure in humans and coverage in toxicology species, hence, circumventing the use of radiolabeled compounds or authentic chemically synthesized standards of metabolites.

Recent guidance from the Food and Drug Administration (FDA) to the pharmaceutical industry recommends nonclinical safety testing of metabolites that are unique or present in disproportionate levels in humans as compared to preclinical species used in safety studies (1). It is recommended that metabolites that are present in quantities greater than 10% of the parent systemic exposure in humans, as determined by the area under the curve (AUC)1 values, be present in approximately equivalent or greater quantities in at least one of the preclinical species subjected to toxicology testing of the parent drug, to avoid discrete safety testing. Minor circulating human metabolites, which are considered to be less than 10% of parent AUC,

are considered to be less important, and further studies to demonstrate their coverage in toxicity species may not be required. An occurrence of a circulating unique metabolite only in humans but not in animals is very rare. If such unique metabolites are found to be present at >10% of parent AUC, nonclinical testing with these metabolites is recommended. A more common scenario is the presence of circulating metabolites at disproportionately higher levels in humans than in preclinical species used in safety studies (1). The presence of disproportionate levels of metabolites in humans needs to be addressed through further nonclinical testing with these metabolites, as outlined in the decision tree presented in the guidance document (1).

* To whom correspondence should be addressed. Tel: 484-865-7525. E-mail: [email protected]. 1 Abbreviations: AUC, area under the curve; DMPK, drug metabolism and pharmacokinetics; ESI-LC/MS, electrospray ionization-liquid chromatography mass spectrometry; 1H NMR, proton nuclear magnetic resonance; HRMS, high-resolution mass spectrometry; MAD, multiple ascending dose; Mpk, milligram per kilogram; MIST, Metabolites in Safety Testing; MRM, multiple reaction monitoring; MS/MS, mass spectrometry/mass spectrometry; PDA, photodiode array detector; SAD, single ascending dose; XIC, extracted ion chromatography.

To implement the recommendations in this guidance, it is important that the systemic exposure of the parent and metabolite(s) is obtained from preclinical species and humans as early as possible so that the development of a drug progresses unhindered. As part of routine drug discovery and development efforts in pharmaceutical companies, short- and long-term safety studies are conducted in both rodent and nonrodent species. As a practice, the exposure of parent compound is obtained in these species after single and multiple doses. At the end of the safety

Introduction

10.1021/tx8003328 CCC: $40.75  2009 American Chemical Society Published on Web 12/10/2008

312

Chem. Res. Toxicol., Vol. 22, No. 2, 2009

studies, toxicological (e.g., histopathological and clinical chemistry) evaluations are conducted, and at the same time, the quantitative exposure values of the parent compound are obtained through parallel toxicokinetic studies. Quite frequently, the AUC values of the metabolites are not obtained, as neither the synthetic standards nor the relevant human in vivo metabolite profiles are available. Typically, determinations of the exposures of metabolites in preclinical species are deferred to later studies. In the existing paradigm within the pharmaceutical industry, the quantitation of metabolites in clinical and preclinical studies often takes place after the first in human studies is completed when the human in vivo plasma metabolite profile becomes available. However, quite frequently, some or all of the metabolites found in human plasma are not available to develop quantitative assays. Hence, none or a limited number of metabolites are quantitated at this stage. Until recently, the criteria used to quantitate metabolites present in human circulation have been unclear, and hence, case-by-case strategies were employed for compounds in development. A common approach has been to estimate the coverage of metabolites by radiolabeled studies conducted in both preclinical species and humans, but the ADME studies with radiolabeled compounds in humans are not conducted until fairly late in the development. The recent guidance from the FDA has set a threshold value for metabolites found in human circulation that we should take into consideration during our safety assessment studies. To meet the expected criteria, novel and creative approaches are needed to obtain quantitative information, preferably early in the development of a compound. Recently, it was suggested that a combination of mass spectrometry and radiomatic detection could be used to provide quantitative information on circulating human plasma metabolites whose synthetic standards were not available (2). However, while this approach appears reasonable, it still relies on the availability of radiolabeled compounds early in development. Recently, we suggested the use of NMR as a tool to obtain quantitative information on the levels of biologically generated metabolites that could be used as reference standards for further quantitative work by LC/MS (3). To implement this strategy effectively, one has to consider expeditious ways of generating exposure values of metabolites from early human and preclinical safety studies. Herein, we describe plasma sample pooling methods that we employed to obtain reasonable estimates of AUCs of metabolites via traditional and novel methods. Several investigators have described pooling of plasma samples to obtain AUC values of compounds in the past (4-11). One of the reasons that this approach was not utilized extensively was that certain PK parameters of interest (Cmax, Tmax, t1/2) to a pharmacokineticist could not be obtained. However, with the recent regulatory requirements on metabolites, the total exposure of metabolites is of greater consequence; hence, methods that would accelerate attainment of these AUC values would prove to be invaluable. An approach consisting of sample pooling combined with the application of modern quantitative tools such as LC/UV/MS and NMR were considered as viable options in obtaining AUC values of metabolites in the possible absence of authentic metabolite standards. Plasma pooling approaches either subject-based or time-based have been utilized for PK evaluations in a high-throughput environment. In the subject-based approach, equal aliquots of sample from the same time point from all of the subjects in a particular group are pooled and analyzed. The resulting concentration is the mean concentration of all of the subjects at that time, and a mean concentration-time profile can be constructed, thereby calculating the average AUC, average Cmax,

Vishwanathan et al.

average Tmax, etc. In the time-based approach, the samples from an individual subject are pooled in the time domain, thereby obtaining one sample per subject. The concentration measured from this one subject over the total time provides an estimate of the AUC with a single pooled sample that is theoretically equivalent to the linear trapezoidal method (4). This time-based approach is most commonly used and was first described by Hamilton et al. (4) to estimate the AUC of valproic acid. This procedure was further adopted and modified for a bioequivalence study of hydrocortisone tablets by Bouvier D’Yvoire et al. (5). A comparison of the traditional approach and the plasma pooling approaches was evaluated by Riad et al. (6), Hop et al. (7), Kuo et al. (8), Hsieh et al. (9), and Ashokraj et al. (10), who demonstrated the value and utility of this method in providing the AUC estimates. Cheung et al. (11) described the application of the plasma pooling approaches to determine the AUC, AUMC, and mean residence time from PK studies and showed the time-saving and high-throughput capability of this approach. The excellent correlation data obtained from these studies indicate that the time-based plasma pooling method uses minimal samples to estimate the AUC of the drug candidate. Here, we present a similar plasma pooling approach and compare it to the traditional approaches of obtaining quantitative information on the levels of circulating metabolites in preclinical species using nonradiolabeled compound and in the absence of chemically synthesized standards of metabolites. However, to validate this approach, we utilized synthetic standards of metabolites to compare and correlate with the traditional approach. In the absence of synthetic standards, calculations of AUC values of metabolites were achieved through the use of UV or NMR as quantitative detectors. The UV profiles were obtained for the parent compound and its metabolites and were shown to be identical or similar before the UV peak areas were used to calculate AUC values of metabolites. A number of compounds, which underwent various metabolic reactions, leading to unchanged or changed chromophores responsible for UV absorbance, were investigated in this study. In situations where the chromophores were altered due to metabolism, an alternative approach in estimating the AUC of the metabolite(s) is presented. Because of the potential changes that could take place in the chromophores of compounds undergoing metabolic reactions, the alternative method of using biologically isolated metabolites as possible reference standards was explored. Previously, we had demonstrated that reliable quantitative information on metabolites isolated from in vitro and in vivo sources could be obtained by NMR (3). We propose that these biologically isolated metabolites could be used as reference standards to generate LC/MS calibration curves for further quantitative assessment of metabolites in plasma.

Materials and Methods Chemicals and Supplies. All compounds (1-4) and metabolite(s) (M1A, M2A, M3A, M3B, M4A, and M4B) were synthesized in the Department of Medicinal Chemistry, Wyeth Research, and had molecular masses in the range of 250-600 Da. Because of the proprietary nature of the compounds, only the partial structures are shown in Figure 1. NADPH and phosphate buffer (pH 7.4) were purchased from Sigma Chemical Co. (St. Louis, MO). Rat and human liver microsomes were obtained from Xenotech (Kansas City, KS). Bond-Elut C18 cartridges (10 g/60 cc and 500 mg/10 mL) were obtained from Varian Sample Preparation Products (Harbor City, CA). Analytical HPLC columns were obtained from various commercial suppliers. HPLC grade water, methanol, and acetonitrile were

Metabolite Exposures through Sample Pooling and NMR

Chem. Res. Toxicol., Vol. 22, No. 2, 2009 313 Table 1. Details of in Vivo Studies Conducted with Compounds 1-4 in Preclinical Safety Studiesa compound

species

gender

doses

days

1

dog

2

dog

3

rat dogb dog

F M F M M M M F

10, 30, 60, 100 60 10, 90 30 50, 150, 550 300 200, 600, 2000 600

7 7 7 5 7 1 7 7

4

a Beagle dogs and Sprague-Dawley rats were administered various doses of compounds 1-4 in separate studies. Plasma samples (t ) 0 to t ) 24 h with at least eight time points) were obtained after the final dose. b Single dose study.

Figure 1. Partial structures of compounds 1-4 and their metabolites assessed in this study. R1 and R2 groups contain groups that have the chromophores responsible for UV absorption.

purchased from Mallinckrodt Chemicals (Phillipsburg, NJ). All deuterated solvents were purchased from Cambridge Isotope Laboratories, Inc. (Andover, MA). All solvents and chemical standards were of the highest grade commercially available with purities greater than 99%. NMR Spectroscopy. NMR data were recorded at 30 °C on a Bruker Avance III 600 MHz spectrometer operating at the basic frequency of 600.13 MHz. All spectra were collected using a 5 mm CPTCI CryoProbe (Bruker BioSpin Corp., Billerica, MA). The spectrometer was controlled using TopSpin (Bruker BioSpin, v 2.0 pl 5). Each sample contained the same volume (0.5 mL) of the solvent and was placed in the magnet at the same depth. All data collection was performed without spinning the samples. The probe was tuned and matched to the specific frequency using the automatic tuning method (ATMA) from Bruker BioSpin. The samples were field frequency locked using the 2H signal of the deuterated solvent. Shimming was performed on each sample prior to data acquisition using the TopShim automatic shimming method from Bruker BioSpin on the solvent peak. Each data set was collected under automation with the same parameter set using the ICON-NMR software (Bruker BioSpin, v 3.7) and manual sample changes. All spectra were referenced to the solvent signal. In Vivo Studies. Plasma samples from rats or dogs dosed with the respective parent compounds were obtained from the toxicokinetic portion of the nonclinical safety study conducted at Wyeth Research, Drug Safety and Metabolism (Chazy, NY). The details of compounds, species used in the studies, and dosing schedules are provided in Table 1. The plasma samples were obtained at various time points on days 1, 5, or 7 and were frozen at -70 or -80 °C until analyzed. Typical time points for plasma collection included predose (0 h), 0.5, 1, 2, 4, 6, 8, 12, and 24 h. At least 400 µL of plasma sample from each time point was available for analysis. Sample Preparation and Pooling Methods To Obtain AUC Values. Various approaches were employed to analyze the plasma samples from preclinical safety studies to obtain the AUC values. Method 1 (Traditional Method). The parent and its metabolite(s) were analyzed individually at all time points using the standard curves generated with synthetic standards. Quan-

titation was conducted using LC/mass spectrometry/mass spectrometry (MS/MS) methods (see below) based on electrospray ionization techniques. The analytical methods consisted of constructing standard curves with at least seven points in the concentration range (2-5000 ng/mL). To ensure the reliability of the assay, quality control (QC) samples were prepared in the low, mid, and high regions of the calibration curve. Following LC/MS/MS multiple reaction monitoring (MRM) analysis, pharmacokinetic parameters were calculated using WINNONLIN. This approach provided the traditional linear trapezoidal method of measuring AUC values of metabolites and/or parent compounds. Method 2 (Plasma Pooling for Each Individual Animal). In this method, the plasma samples were pooled based on time and volume from individual study animals in a particular dose group. Representative calculations used to derive the volumes of plasma aliquot from each time point are shown in Table 2. The procedure for the time based pooling technique is wellestablished and reported elsewhere (4-11). Briefly, the volume of aliquot taken at a particular time is proportional to the difference in the time interval on either side of the sampling time except for t0 (initial time point) and tf (final time point) where the difference is obtained from the next or the previously obtained sample. One pooled plasma sample per animal was obtained by this method. The parent and metabolite(s) concentrations were measured in triplicate from each pooled sample using LC/MS/MS MRM analysis. After obtaining the concentrations of parent compound and its metabolites from these pooled samples, the following equation was used to calculate the AUC values (4):

AUCmetabolite or parent)Cpool(tf - t0)

(1)

where AUC is the area under curve in ng h/mL or µg h/mL, Cpool is the measured concentration of the metabolite or parent in the pooled sample by LC/MS, tf is the final time point (usually 24 h), and t0 is the initial time point (usually 0 h). The AUC values were obtained by multiplying the measured concentrations by 24 h in most cases. Method 3 (Plasma Pooling from All Animals in a Dose Group). Equal aliquots (e.g., 50 µL) of plasma at each time point were first pooled from all of the animals in a particular dose group. This created samples, which had average concentrations of analytes at each time point from all of the animals in that dose group. Further pooling of samples across all of the time points (with pooled samples from all subjects) was done as described above for method 2. Hence, one pooled plasma sample per dose group was obtained by this method. The parent and metabolite(s) concentrations were measured in triplicate from this sample using LC/MS/MS MRM analysis. The AUC

314


Vishwanathan et al.

Table 2. Representative Example of “Method 2” Illustrating How Plasma Pooling Was Conducted for Each Animal (Dosed with Compound 1) from a Particular Dose Groupa

SAN #1 SAN #2 SAN #3

time (h)

0

0.5

1

2

4

6

12

24

Tf - T0 ) 24

∆tj volume ) (k∆tj) (µL) volume ) (k∆tj) (µL) volume ) (k∆tj) (µL)

0.5 5 5 5

1 10 10 10

1.5 15 15 15

3 30 30 30

4 40 40 40

8 80 80 80

18 180 180 180

12 120 120 120

total vol ) 480 total vol ) 480 total vol ) 480

a k ) proportionality constant, usually obtained after defining the volume at t ) 0. A value of 10 is used here, for pipetting precision. SAN# refers to study animal number.

Table 3. MRM Transitions (Q1 f Q3) and Mass Spectral Parameters Employed during the LC/MS/MS Analysis of Metabolites of Compounds 1-4a compound

Q1

Q3

DP

CE

CXP

r2b

M1A M2A M3A M3B M4A (weighed) M4A (NMR) M4B (NMR) IS

287.2 478.3 375.1 373.2 489.1 489.1 513.2 361.1

200.1 349.0 289.3 255.3 220.1 220.1 259.0 330.1

66 75 75 75 76 76 76 75

35 32 28 25 33 33 37 25

15 15 15 15 10 10 10 15

0.9987 0.9976 0.9900 0.9990 0.9993 0.9991 0.9984 NA

a The following mass spectrometer conditions were common to all quantitative LC/MS experiments performed. CAD gas, 5 arbitrary units; CUR, 25 arbitrary units; IS GAS1, 50 arbitrary units; IS GAS2, 50 arbitrary units; IS voltage, 5000 V; temperature, 500 °C; interface heater, on; dwell time, 50 ms. Abbreviations: CAD, collision activated dissociation; CE, collision energy; CUR, curtain gas; CXP, collision cell exit potential; DP, declustering potential; and Q1/Q3, quadrupoles 1 and 3, respectively. b The correlation coefficient was obtained for calibration curves generated for each metabolite in the concentration range 2-5000 ng/mL.

values were obtained in a similar manner as described above for method 2. Sample Analysis. Samples from Methods 1-3 for LC/ MS/MS MRM Analysis. All plasma samples were extracted by protein precipitation with acetonitrile. Briefly, 25-100 µL aliquots of plasma from standard calibration points, QC, and from samples obtained via methods 1-3 (described above) were protein precipitated with 400 µL of acetonitrile containing the internal standard. After vortexing and centrifugation, the supernatants were transferred to a new plate and evaporated to dryness under nitrogen. The samples were subsequently reconstituted with 300 µL of 20:80 methanol:water (v/v) for LC/MS/ MS analysis in the MRM mode. The MRM transitions employed to monitor the compound, and its metabolite(s) are provided in Table 3. All LC/MS/MS analysis was performed on a Sciex API 4000 mass spectrometer equipped with the turboion spray source and interfaced with an Agilent HPLC system consisting of a binary pump, column thermostat, and a LEAP LC-10 autosampler. The HPLC procedures were performed under gradient conditions using a mobile phase of 5 mM ammonium acetate in water (A) and acetonitrile (B) on a Gemini C18 50 mm × 2.0 mm column. The mobile phase gradient conditions were adjusted such that the analytes of interest eluted at ∼2.5 min and that there were no coelution of components. No interference was observed during the analysis. For all analytes, the instrument was tuned to provide the best possible MRM transitions. The peak area ratios of analytes to internal standards were calculated followed by regression analysis using powerfit without weighting to generate the standard curve. All standard curves showed linearity over the range of 2-5000 ng/mL (r2 > 0.99), and all QC values were within acceptable limits (95% purity as determined by LC/ UV/MS and NMR analysis. Using NMR, the exact concentrations of the isolated metabolites were measured using the calibration curve constructed with the parent compound 4 as described previously (3). Briefly, the NMR spectra of compound 4 (parent) were obtained at five different concentrations, and a calibration curve was created based on the integration of proton signals (average). The concentrations of metabolites M4A and M4B were calculated using this standard curve as described previously (3). The results are presented in Table 6. It has been shown by Espina et al. (3) during the validation of the quantitative NMR

316


Vishwanathan et al.

Table 6. NMR Determination of Concentrations of Metabolites M4A and M4B after Their Isolation from Rat Liver Microsomal Incubationsa,b compounds compound 4 (µM) day 1 day 2 mean SD % RSD

M4A

M4B

M4A standard

100

500

1000

3000

5000

unknown

unknown

2500

106 102 104 3 2.5

495 498 497 2 0.4

986 986 986 0 0.0

3024 3027 3025 2 0.1

4989 4987 4988 2 0.0

431 429 430 1 0.3

629 617 623 8 1.3

2506 2498 2502 6 0.2

a The concentrations of M4A and M4B were determined using the NMR calibration curve constructed with the parent compound 4. A known solution of synthetic standard of M4A was prepared as QC to confirm the quantitation of biologically isolated M4A and M4B. b Triplicate determinations were made on each sample on two separate days.

Table 7. Comparison of AUCs (µg h/mL) Obtained by Using Gravimetrically (A) and NMR (B) Determined Initial Concentrations of Metabolites of Compound 4a method 1

dose (mg/kg) (gender)

metabolite M4A

200 (M) 600 (M) 2000 (M) 600 (F) 200 (M) 600 (M) 2000 (M) 600 (F)

M4B

method 2

method 3

A

B

A

B

A

B

UV basedb

1.2 ( 0.2 1.7 ( 0.6 2.8 ( 0.3 3.4 ( 0.8

1.1 ( 0.1 1.5 ( 0.5 2.5 ( 0.3 3.1 ( 0.8 3.0 ( 1.5 3.9 ( 1.9 5.3 ( 0.5 6.3 ( 1.4

1.5 ( 0.4 1.6 ( 0.2 2.5 ( 0.4 3.2 ( 1.2

1.3 ( 0.3 1.4 ( 0.2 2.2 ( 0.4 2.9 ( 1.1 3.7 ( 0.7 3.6 ( 0.2 5.6 ( 1.5 7.2 ( 3.5

1.5 1.5 2.7 3.1

1.3 1.3 2.4 2.8 4.2 3.7 6.3 6.9

3.5 2.9 5.8 10.0 11.2 12.2 14.2 12.8

a A corresponds to LC/MS quantitation performed with a weighed synthetic metabolite standard to produce a calibration curve (gravimetric). B corresponds to LC/MS quantitation performed with a biologically isolated standard whose initial concentration was determined by NMR. This “reference standard” was subsequently used to generate a calibration curve as in approach A. Method 1, traditional method; method 2, plasma pooling for each individual animal; and method 3, plasma pooling from all animals in a dose group. b Mean of triplicate values.

Table 8. Comparison of MS Ionization and UV Responses of Compounds (1-4) and Their Respective Metabolites at 1 and 10 µg/mL UV responsea

compound M1A compound M2A compound M3A M3B compound M4A M4B

1 2 3 4

HRMS ionization efficienciesb

1 µg/mL (peak area)

% parent


% parent


% parent


% parent

5.5 × 105 5.3 × 105 7.1 × 105 6.2 × 105 8.3 × 105 5.8 × 105 8.2 × 105 2.4 × 104 5.2 × 104 3.6 × 104

100 96 100 88 100 69 98 100 218 14

5.6 × 106 5.3 × 106 7.1 × 106 6.2 × 106 8.4 × 106 5.9 × 106 8.3 × 106 4.8 × 105 8.0 × 105 6.7 × 105

100 94 100 87 100 69 98 100 165 140

4.4 × 108 2.7 × 108 1.7 × 108 1.2 × 108 4.6 × 108 3.3 × 108 4.1 × 108 8.5 × 106 6.9 × 107 2.4 × 106

100 62 100 69 100 72 88 100 809 28

1.8 × 109 1.0 × 109 9.4 × 108 5.8 × 108 2.4 × 109 1.0 × 109 1.7 × 109 2.1 × 108 9.0 × 108 5.9 × 107

100 56 100 62 100 44 74 100 429 28

a UV responses were obtained using the parent UV maxima. scan MS data.

b

HRMS efficiencies were obtained based on 5 ppm mass window using Orbitrap full

method that the concentrations of metabolites measured using this approach are within 10% of the gravimetrically determined nominal values. Hence, the measured concentrations of the isolated metabolites were used to obtain the absolute amount of metabolite that was isolated from a biological source. Subsequently, this sample (with the concentration of the metabolite known) was used to prepare a series of diluted samples that were used to prepare a calibration curve for LC/ MS/MS analysis of plasma samples. A comparison of AUC values obtained using synthetic (gravimetrically weighed) and biologically isolated standards of metabolite M4A is shown in Table 7. Furthermore, as no synthetic standard of metabolite 4B was available, an estimate of the AUC values for this compound was obtained using the biologically isolated material (Table 8). Comparison of MS Ionization and UV Responses for Compounds 1-4 and Their Corresponding Metabolites. Compounds 1-4 along with their metabolite(s) were prepared at concentrations of 1 and 10 µg/mL each and were evaluated

for their MS ionization and UV responses under identical experimental conditions. HPLC/UV/MS analysis was conducted with equimolar amounts of metabolite(s) and parent compound prepared in acetonitrile:water (1:9, v/v). The UV peak areas of the parent compounds and their corresponding metabolites present in same quantities (1 or 10 µg/mL) were obtained using a diode array detector placed in-line with the HPLC column. The HPLC system used for analysis was a Surveyor HPLC (Thermo Scientific, San Jose, CA) equipped with a diode array UV detector. The HPLC mobile phase used in this study is described above in “Method 3 for LC/UV/MS Analysis”. The mass spectral response for compounds 1-4 and their metabolites were evaluated after their elution from the UV detector that was placed in tandem with the mass spectrometer. The peaks were analyzed by a LTQ Orbitrap mass spectrometer (Thermo Scientific) equipped with an ESI interface. The instrument was operated in either positive or negative ionization modes. Calibration and tuning of the instrument were conducted with parent compound only and not the metabolites because in normal


Chem. Res. Toxicol., Vol. 22, No. 2, 2009 317

circumstances, the synthetic standards of metabolites will be unavailable. The high-resolution mass spectrometric responses were also recorded within a 5 ppm window based on theoretical m/z values (Table 8).

Results Four sets of compounds (1-4) and their metabolites (M1A, M2A, M3A, M3B, M4A, and M4B) with a range of physicochemical properties were evaluated in this study. Because of the proprietary nature of the compounds, partial structures depicting the site of biotransformation reactions are shown in Figure 1. The in vitro and in vivo metabolism of these compounds were studied in-house, and major metabolites were identified and characterized prior to requesting synthetic standards. The major metabolites were produced via several types of metabolic reactions including hydroxylation, ring scission, N-oxidation, and desaturation (Figure 1). Compounds 1-4 were administered to either dogs or rats, and the plasma samples were obtained after 1, 5, or 7 consecutive days of dosing (see Table 1). The plasma samples were subsequently processed via different methods to obtain the AUC values of metabolites as described in the Materials and Methods. The availability of metabolite standards enabled us to obtain and compare the quantitative AUC values obtained by various approaches. However, under normal circumstances, synthetic standards of these metabolites are usually not available in early development to develop quantitative LC/MS/MS assays to measure their concentrations in plasma. The AUC values of the parent compounds were available from these short-term safety studies (using validated GLP bioanalytical methods), and hence, these numbers were not regenerated in this study. It is expected that the AUC values of the parent compounds will be readily available from validated assays at the time of exposure determination for these metabolites. The results described below will demonstrate how comparable AUC values could be obtained via UV and NMR approaches in the absence of synthetic standards of metabolites. Furthermore, results are presented that illustrate the value of sample pooling in obtaining reliable AUC estimates, hence saving considerable resources. The photodiode array (PDA) UV spectra and LC/UV chromatograms of compound 1 and its metabolite M1A present in pooled plasma sample are shown in Figure 2a,b, respectively. The UV profiles for 1 and M1A were found to be identical. Calculations of the AUC values for M1A, using UV peak areas, are shown in Table 4. The comparison of AUCs obtained via different approaches for M1A is shown in Table 5. The AUC values obtained via the traditional method (method 1) were used as reference points against which all other numbers were compared. The mean AUC obtained via UV approach was within 20% of the values obtained via method 1. It should be noted that methods 1 and 2 provide data showing a range in exposure values, indicative of the interindividual variability among different subjects in the same dose group (Table 5). A comparison of AUC values obtained via methods 1 and 2 shows the value of sample pooling, as very similar numbers were produced by these two methods. Method 3, as it ultimately produces only one sample per dose group, provides a mean exposure value only. Similarly, the UV method, as it relies on pooled sample from method 3, yields only an average estimate of the exposure level of a metabolite in a particular dose group. Similarly, the AUC calculations were performed for compounds 2 and 3 and their respective metabolites M2A, M3A, and M3B. The PDA spectra of compounds 2 and 3 along with their metabolites M2A, M3A, and M3B are shown in Figures

Figure 2. (A) UV spectra of compound 1 and its metabolite, M1A. (B) LC/UV chromatograms of pooled samples produced via method 3 at different doses of compound 1. The wavelength (λ) was set in the range 320-360 nm.

3 and 4, respectively. The UV profiles of metabolites and their parent compounds were compared and shown to be similar prior to obtaining the UV peak area ratios for AUC determination. In all of these cases, the AUC values obtained by various approaches were within 20% of each other, indicating the suitability of the pooling approach and the use of UV-based quantitation to estimate the AUCs of metabolites (Table 5). The UV-derived AUC obtained from M3A was interesting, as it appeared to have underestimated the exposure level. Further analysis of the UV response from equimolar quantities of 3 and M3A showed a 30% difference in the response, even though the UV profiles (see Figure 4) looked identical. A possible explanation for this observation is that while λmax may not have changed for M3A, the molar absorptivity may have been reduced due to the unavailability of lone pair of electrons from the nitrogen that formed the N-oxide. One should be aware of this possible limitation of the UV-based approach for estimating AUCs. For compound 4 and its metabolites M4A and M4B, the UV spectra were very different from each other as shown in Figure 5. In this case, obtaining UV peak area ratios to calculate the AUC values was pointless, as it would have led to erroneous results. An alternate method for obtaining the AUC values of metabolites is highly desirable in situations where the chro-

318


Figure 3. (A) UV spectra of compound 2 and its metabolite, M2A. (B) LC/UV chromatograms of pooled samples obtained via method 3 at different doses of compound 2. The wavelength (λ) was set in the range 290-310 nm.

mophores of the metabolites are found to be different from their respective parent compounds. Metabolic reactions can lead to significant alterations in the UV-absorbing characteristics of the metabolites. In such cases, the proportionality constants derived from UV peak areas of metabolites and parent compounds cannot be used to obtain the AUC values. An alternate strategy, based on the use of NMR to quantify biologically isolated metabolites, was employed to obtain the AUC values of such metabolites. As described previously, isolation and quantitation of biologically generated metabolites by NMR were considered as an alternate method for obtaining quantitative values in the absence of synthetic standards (3). Once the concentrations of the isolated metabolites are determined by NMR, an LC/MS/ MS method is developed to obtain concentrations of the metabolites present in plasma extracts. Hence, in the case of metabolites of compound 4, sample preparation and analyses were performed as described previously but with the LC/MS/ MS quantitation performed using isolated metabolites. Metabolites 4A and 4B were isolated from rat liver microsomes and quantitated by NMR (Table 6). A synthetic standard of 4A was also available, and a known concentration (2.5 mM) of this metabolite was used as QC to ensure the accuracy of quantitative NMR (Table 6). Synthetic standard of M4B was not available; therefore, NMR was used to obtain quantitative information. Hence, AUC data were obtained using both biologically isolated metabolites (referred to as “M4A-NMR” and “M4B-NMR”)

Vishwanathan et al.

Figure 4. (A) UV spectra of compound 3 and its metabolites, M3A and M3B. (B) LC/UV chromatograms of pooled samples obtained via method 3 at different doses of compound 3. The wavelength (λ) was set in the range 265-275 nm.

and synthetic reference standard (referred to as M4A-STD”). This allowed cross-validation of the methods utilizing biologically isolated and synthetic reference standards to obtain the AUC values (Table 7). The studies conducted with weighed standard of M4A corresponded to the traditional method of conducting quantitative analysis whereby serially diluted solutions of weighed analyte are prepared to obtain the standard curve for LC/MS/MS analysis. Similarly, the biologically isolated metabolite samples, M4A-NMR and M4B-NMR, were used to prepare a series of solutions intended for the preparation of calibration curves by LC/MS analysis. The results from the calculation of AUC from these approaches are shown in Table 7, which indicate an excellent correlation of AUC values between the traditional and the pooled sample approach when using either biologically generated or synthetic standards of metabolite. Because the concentration measured by NMR was accurate for M4A, it was assumed the concentration measured by NMR for M4B was also valid and, hence, the AUC obtained for M4B. As seen from Table 7, the calculation based on UV approach overestimated the AUC values of the metabolites, as expected, since the UV spectra were very different for the parent and the metabolite(s). Compounds 1-4 and their metabolites (where available) were subjected to UV and mass spectral analysis to evaluate the similarities and differences in the analytical instruments’ response. As part of the decision-making process, the methods



are similar between the parent compound and its metabolites (compounds --3), the UV responses at both 1 and 10 µg/mL were identical ((20% deviation from parent). However, the HRMS responses varied greatly ((60% when compared with parent). For 4 and its metabolites M4A and M4B, the UV spectra were very different from each other, leading to significant differences in the UV responses. The HRMS ionization efficiency for these metabolites varied greatly as well (>800%).

Discussion

Figure 5. (A) UV spectra of compound 4 and its metabolites, M4A and M4B. (B) LC/UV chromatograms of pooled samples obtained via method 3 at different doses of compound 4. The wavelength (λ) was set in the range 200-360 nm.

that were subsequently employed to obtain the AUC values of metabolites in biological extracts were based on these initial results. Using the synthetic standards of compounds and their respective metabolites, the MS ionization and UV responses were evaluated at two concentrations (1 and 10 µg/mL). As shown in Table 8, for compounds 1-3, the UV responses were similar between the parent compounds and their metabolites. However, the comparison of UV peak areas of 4, M4A, and M4B analyzed at the same concentration and under identical experimental conditions indicated a significant (>400%) difference between the parent compound and its metabolites. Identifying such a significant differences in the UV profiles of metabolites as compared to parent molecule is a very important part of the strategy in estimating AUC values. Similarly, the high-resolution mass spectrometer (HRMS) ionization efficiencies performed with a 5 ppm window for each selected pseudo molecular ion indicated significant variation for the compounds and their metabolites. The differences in ionization potentials for metabolites produced as a consequence of a biotransformation modification on a parent molecule are very well-established. Hence, compounds present at equimolar concentrations, although structurally related, can have significant differences in the mass spectral response (Table 8). Therefore, utilizing mass spectral response to obtain peak area ratios of metabolites and parent compounds will also lead to unreliable estimates of AUC values. The results shown in Table 8 indicate that when the UV spectra

The recently published Guidance for Industry “Safety Testing of Drug Metabolites” clearly outlines the FDA’s position on the circumstances where the safety of human-disproportionate metabolites should be investigated (1). The subject of “metabolites in safety testing (MIST)” has been extensively debated and discussed, especially after it was highlighted by a joint publication sponsored by the American PhRMA group in 2002 (12). The MIST publication summarized the best practices within the U.S. pharmaceutical industry in assessing the role of drug metabolites as potential mediators of the toxicity of new drug products. Draft guidance from the FDA for the pharmaceutical industry on safety testing of drug metabolites was issued in 2005, resulting in further discussions on this subject (13-20). In this draft guidance, the threshold for defining a major metabolite was lowered to 10% or more of the exposure to circulating drugrelated material (13). Subsequently, in the recent finalized guidance, the regulatory agency redefined the criteria for major metabolites as being 10% of the total exposure of parent compound (1). The new criteria set forward by the regulatory agency has somewhat facilitated obtaining relevant exposure information without the obligatory use of radiolabeled compounds to determine total exposure of circulating drug-related material. Each metabolite can be assessed relative to the parent compound and defined easily as being minor or major using the 10% criterion set by FDA. Obviously, one of the major tasks in implementing this recommendation is determining the exposure values of all metabolites that are found in human circulation in early development. This will be a bigger challenge for compounds that are extensively metabolized and which show multiple metabolites in human plasma at steady state. The presence of low levels of parent compound, as a consequence of improved potency or extensive metabolism, can pose some challenges in assessing all metabolites that are present in greater than 10% of parent AUC. The purpose of this publication is to demonstrate how one could obtain the relevant exposure values of metabolites in plasma of preclinical species (and in plasma from early human studies). Herein, we propose a strategy to address the coverage of potential human-produced metabolites in preclinical species without the need for authentic certified metabolite standards or radiolabeled studies. This strategy calls for analyzing the plasma samples (at steady state) from early safety studies to obtain estimates of the AUC values of metabolites. Of course, one must realize that prior to the first in human (FIH) studies, human in vivo metabolite profiles are not available; hence, one could depend on a reliable in vitro system that would give a fairly good idea on the potential metabolites that need to be monitored in preclinical species. Alternatively, one could wait until the FIH studies before analyzing the preclinical plasma samples. As the compound progresses into FIH studies, the same approach can be applied to obtain exposure estimates of major metabolites from humans administered multiple doses of the

320


Scheme 1. Decision Tree Showing Various Approaches That Could Be Used in Obtaining the AUC Values of Metabolitesa

a On the basis of historical data, the second route, B, where synthetic standards are not available, is a more common scenario. Isolation of biologically synthesized metabolite and subsequent quantitation by NMR is a preferred option in many cases.

compound. Subsequently, comparison of metabolite exposures in humans and preclinical species can be made (3). This initial estimate and comparison of AUC values of metabolites can aid the decision-making process such as the need to have large quantities of a metabolite made for further safety testing, as suggested in the FDA guidance document (1). If initial data suggest that the metabolite is covered in one of the preclinical species (e.g., greater than the AUC calculated for humans), then the request for this metabolite could be deprioritized, as it is of less concern. However, if a metabolite is barely covered in any of the toxicity species or is present as a unique human metabolite, then it is important to have this metabolite chemically synthesized and tested in safety studies as recommended in the FDA guidance (1). This approach of comparing the exposure values of metabolites after multiple doses in animals and humans is more appropriate and desirable than comparisons made with single doses of radiolabeled compounds. To obtain exposure values of metabolites in the absence of chemically synthesized metabolite standards, a strategy as outlined in Scheme 1 was employed. The critical activities, which are discussed below, included sample pooling, comparing UV spectra of metabolites and their parent compounds, obtaining UV ratios to obtain AUC values, isolating biologically generated metabolites, and employing NMR as an alternate and a more desirable tool to assist in the AUC calculations. Pooling of plasma is suggested to reduce the number of samples for analyses as well as permitting an overall evaluation of relative levels of metabolites as compared to the parent compound. Plasma pooling methods, described previously in the literature (4-11), were shown to produce almost identical AUC values as compared to the data obtained via analyses of individual samples (Tables 5-7). In initial studies, the intent is to obtain reliable estimates of AUC values of metabolites in an expeditious manner. Sample pooling allows us to obtain these data rapidly, as only a few samples will need to be analyzed. Sample pooling also leads to consolidation of analytes in a limited volume; hence, quantitative analysis becomes much easier. For example, the AUCs obtained for M1A-M3B showed values ranging from 3.56 to 74.6 µg h/mL via method 2, where LC/MS/MS was used to obtain concentrations of pooled samples from individual animals administered different doses (Table 5).

Vishwanathan et al.

The corresponding concentrations that were measured (by LC/ MS) in these pooled samples ranged from 14.8 to 3110 ng/mL, respectively. Setting bioanalytical methods with LOQ of 10 ng/ mL is much easier than trying to perform quantitative analysis at 1 ng/mL, as done routinely in bioanalytical LC/MS laboratories. The low levels of metabolites in individual samples can at times pose real analytical challenges as we attempt to quantitate analytes present close to the LOQ (often 1 ng/mL) of an assay, especially with the critical late-time points. Because of the importance of concentrations of metabolites from late time points, one could potentially introduce significant error in the calculation of AUCs via the traditional method of sample analysis. On the other hand, this is not an issue with pooled samples as the metabolites are present well above the LOQ of the assay method. One disadvantage with sample pooling is that we do not get other PK parameters such as half-lives and Cmax values. However, with the intent of comparing only the exposure values to address metabolite coverage, it is very logical to pool samples and obtain the AUC in this manner. In studies where interindividual variability is likely (as in a heterogeneous population), method 2 can be used to pool samples. During the early stages of compound development, method 3 can be used to provide average AUC estimates of metabolites. However, as the compound moves into clinical setting, sample pooling by method 2 will be highly desirable so that interindividual variability among human subjects could be obtained. Another advantage of sample pooling is that we obtain the overall 0-24 h LC/UV/MS profiles (Figures 2-5) that display the relative levels of metabolites to the parent compounds at each dose level (if sample pooled by method 3). These profiles can provide a semiquantitative assessment of metabolite levels. Because sample pooling provides an overall metabolic profile for the entire time points pooled, another limited set of samples may be prepared from early and late time points to identify any metabolite(s) that could be accumulating or slowly eliminated with respect to parent. As part of the workflow process, an evaluation of UV characteristics of the metabolites is made before further quantitative estimates are made (Scheme 1). A decision tree showing the proposed workflow is shown in Scheme 1. Depending on the UV spectra of the parent and its metabolites, a simple pooling experiment can be performed to estimate the AUC of metabolites without the need for authentic metabolite standards. If the metabolite UV spectrum coincides with that of the parent, then the LC-UV profile from pooled samples could be used to estimate the 0-24 h AUC of the metabolites (Table 5). In most cases, the bioanalytical laboratory that routinely conducts validated analysis of samples from safety studies provides the AUC of the parent compound. As shown in this study, the chromophores (hence, the λmax and UV profile) of metabolites can at times be very different from the parent compound; therefore, this approach may not be applicable in all cases (Table 7). For compounds 1 and 2, the site of modification on the molecule was distant from the relevant chromophore that was being monitored. For compound 3, one of the metabolites, M3A, was produced by oxidation of the nitrogen, whose lone pair was part of the conjugated chromophore system (depicted as R1 in Figure 1). Although the UV profile of M3A looked similar to the parent compound, it appears that the formation of N-oxide (hence the lone pair becomes unavailable) led to a change in the molar absorptivity, while retaining the same λmax. The differences in the molar absorptivity between M3A and compound 3 were demonstrated by obtaining UV responses at approximately equimolar con-


centrations (Table 8); a 30% reduction in response was observed with M3A as compared to the parent compound. For compound 4, the creation of a double bond in conjugation with the existing chromophore (M4A) or scission of the piperazine ring system (M4B) created a significantly different UV profiles for these two metabolites as compared to the parent compound. The changes in the UV profiles of metabolites 4A and 4B are examples of where attempts to get AUC estimates from UV peak areas failed. During our studies to investigate the suitability of using the UV peak area ratios, we also had to make sure that the metabolite peaks were very well-resolved from each other and from other endogenous components. Therefore, a long gradient was often used to ensure good resolution of all components. Although it can be time-consuming developing the right separation method, this is not a significant problem, considering that only limited numbers of samples are analyzed once the right method is developed. Considering all of these potential limitations, an alternate method that could be more widely used in obtaining the AUC values was investigated. This alternate method is dependent on the availability of pure metabolite samples that could be used as “reference standards”. Traditionally, metabolite reference standards undergo a series of qualifications before they can be used for quantitative analysis as recommended by FDA (21, 22, and references cited therein). One of the challenging tasks in making an overall assessment of exposures of metabolites in early development has been the non- or limited availability of these fully certified/characterized synthetic metabolite reference standards. With the need to obtain quantitative information on metabolites much earlier, it is logical to explore and use metabolites that could be obtained readily, especially those derived via nonsynthetic chemical routes. The recent progress made in analytical techniques as well as the availability of various metabolite-generating systems has greatly facilitated isolation and characterization of drug metabolites early in drug discovery and development (23). Furthermore, as proposed before, one could make use of the biologically isolated metabolites so that they can be used as reference standards rather than relying on synthetic chemistry (3). NMR can be used to quantitate biologically isolated metabolites so that they could subsequently be used as reference standards in developing analytical methods (especially LC/MS-based) to analyze biological samples (3) (Scheme 1). This strategy eliminates the need for resource-consuming chemical synthesis of metabolites as well as the need for radiolabeled compounds in early development. Another major advantage of using these biologically generated metabolites is that a direct comparison of AUC values of metabolites could be obtained quickly between human subjects and preclinical species administered multiple doses of the compound. Currently, human ADME studies are conducted with a single dose of radiolabeled compound; hence, valid comparisons between clinical and preclinical exposures of metabolites cannot be really made. One of the dilemmas that we face with these limited quantities of biologically isolated metabolites is our ability to generate certificates of analysis. It has previously been proposed that quantitative 1H NMR along with mass spectral (LC/UV/MS) analysis should suffice in providing identity and purity information on chemically synthesized compounds (24, 25). This would enable considerable sample- and time-saving with a concomitant simplification of the analytical characterization process. It was suggested that traditional quantitative characterization tools such as elemental analysis, HPLC, loss on drying, and residue on ignition determination are not needed in today’s fast-paced and expensive


drug discovery environment. NMR is currently used routinely to obtain quantitative information on bioactive natural products, where chemical standards are nonexistent (26, 27). It is proposed that the biologically isolated metabolites, if shown to be pure by LC/UV/MS and NMR, should be acceptable as reference standards for initial quantitative studies, usually by LC/MS/ MS. A limited set of method verifications can be conducted with these metabolite standards after which they can be used in bioanalytical assays. A tiered approach to bioanalytical validation, as proposed before, could be implemented using these biologically produced metabolites (28). The tiered approach would allow quantitative information to be achieved in early development using bioanalytical methods with limited validation, with validation criteria increasing as a compound progresses further in development. This strategic approach will lead to considerable savings for the companies. Herein, we propose that sample pooling, especially by method 2 (to obtain subject variability) combined with a quantitative assay using chemically or biologically synthesized metabolites, can be used to obtain the AUC estimates of metabolites in early development. An approach such as this would permit a quick assessment of whether human-produced metabolites are adequately covered by preclinical species or additional safety testing of disproportionate metabolites would be required.

References (1) U.S. Food and Drug Administration (2008) Guidance for Industry: Safety Testing of Drug Metabolites, U.S. Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Rockville, MD. (2) Yu, C., Chen, L. C., Gorycki, F. L., and Neiss, T. G. (2007) A rapid method for quantitatively estimating metabolites in human plasma in the absence of synthetic standards using a combination of liquid chromatography/mass spectrometry and radiomatic detection. Rapid Commun. Mass. Spectrom. 21 (4), 497–502. (3) Espina, J. R., Yu, L., Wang, J., Tong, Z., Vashishtha, S., Talaat, R., Scatina, J., and Mutlib, A. E. (2008) Nuclear magnetic resonance as a quantitative tool to determine the concentrations of biologically produced metabolites: Implications in metabolites in safety testing. Chem. Res. Toxicol. 22, 299–310. (4) Hamilton, R. A., Garnett, W. R., and Kline, B. J. (1981) Determination of mean valproic acid serum level by assay of a single pooled sample. Clin. Pharmacol. Ther. 29, 408–413. (5) Bouvier d’Yvoire, M., Coquerel, A, Chretien, P., and Tremblay, D. (1993) Pooling plasmas for a rapid answer in pharmacokinetic studies: application to a bioequivalence study of hydrocortisone tablets. Fundam. Clin. Pharmacol. 7, 253–260. (6) Riad, L. E., Chan, K. K., and Sawchuk, R. J. (1991) Determination of the relative formation and elimination clearance of two major carbamezepine metabolites in humans: A comparison between traditional and pooled sample analysis. Pharmacol. Res. 8, 541–543. (7) Kuo, B. S., Noord, T. V., Feng, M. R., and Wright, D. S. (1998) Sample pooling to expedite bioanalysis and pharmacokinetic research. J. Pharm. Biomed. Anal. 16, 837–846. (8) Hop, C. E. C. A., Wang, Z., Chen, Q., and Kwei, G. (1998) Plasma pooling methods to increase throughput for in vivo pharmacokinetic screening. J. Pharm. Sci. 87, 901–903. (9) Hsieh, Y., Bryant, M. S., Brisson, J. M., Ng, K., and Korfmacher, W. A. (2002) Direct cocktail analysis of drug discovery compounds in pooled plasma samples using liquid-chromatography tandem mass spectrometry. J. Chromatogr. B 767, 353–362. (10) Ashokraj, Y., Singh, I., Kaur, K. J., Kohli, G., Bhade, S. R., Varma, M. V. S., Kaul, C. L., and Panchagnula, R. (2005) Plasma pooling: Utility in expediting bioequivalence assessment of rifampicin containing fixed dose combinations. Int. J. Tuberculosis Lung Dis. 9, 697– 699. (11) Cheung, B. W. Y., Cartier, L. L., Russlie, H. Q., and Sawchuk, R. J. (2005) Using sample pooling methods in the determination of AUC and AUMC in pharmacokinetic studies. Fundam. Clin. Pharmacol. 19, 347–354. (12) Baillie, T. A., Cayen, M. N., Fouda, H., Gerson, R. J., Green, J. D., Grossman, S. J., Klunk, L. J., LeBlanc, B., Perkins, D. G., and Shipley, L. A. (2002) Drug metabolites in safety testing. Toxicol. Appl. Pharmacol. 182, 188–196.

322


(13) U.S. Food and Drug Administration (2005) Guidance for Industry: Safety Testing of Drug Metabolites. Draft Guidance, U.S. Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Rockville, MD. (14) Smith, D. A., and Obach, R. S. (2005) Seeing through the mist: Abundance versus percentage. Commentary on metabolites in safety assessment. Drug Metab. Dispos. 33, 1409–1417. (15) Smith, D. A., and Obach, R. S. (2006) Metabolites and safety: What are the concerns, and how should we address them? Chem. Res. Toxicol. 19, 1570–1579. (16) Prueksaritanont, T., Lin, J. H., and Baillie, T. (2006) Complicating factors in safety testing of drug metabolites: Kinetic differences between generated and preformed metabolites. Toxicol. Appl. Pharmacol. 217, 143–152. (17) Davis-Bruno, K. L., and Atrakchi, A. (2006) A regulatory perspective on issues and approaches in characterizing human metabolites. Chem. Res. Toxicol. 19, 1561–1569. (18) Guengerich, F. P. (2006) Safety assessment of stable drug metabolites. Chem. Res. Toxicol. 19, 1559–1560. (19) Humphreys, W. G., and Unger, S. E. (2006) Safety assessment of drug metabolites: Characterization of chemically stable metabolites. Chem. Res. Toxicol. 19, 1564–1569. (20) Stevens, J. L. (2006) Future of toxicology. Mechanisms of toxicity and drug safety: Where do we go from here? Chem. Res. Toxicol. 19, 1393–1401. (21) U.S. Food and Drug Administration (2000) Guidance for Industry: Analytical Procedures and Methods Validation Chemistry, Manufacturing and Controls Documentation, U.S. Food and Drug Administra-

Vishwanathan et al.

(22) (23)

(24) (25) (26) (27)

(28)

tion, Center for Drug Evaluation and Research (CDER) and Center for Biologics Evaluation and Research (CBER), Rockville, MD. U.S. Food and Drug Administration (2001) Guidance for Industry: Bioanalytical Method Validation, U.S. Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Rockville, MD. Mutlib, A. E., and Shockcor, J. P. (2003) In Application of LC/MS, LC/NMR, NMR and Stable Isotopes in Identifying and Characterizing Metabolites in Drug Metabolizing Enzymes: Cytochrome P450 and Other Enzymes in Drug DiscoVery and DeVelopment (Lee, J. S., Obach, R. S., and Fisher, M. B., Eds.) Marcel Dekker, New York, NY. Pinciroli, V., Biancardi, R., Visentin, G., and Rizzo, V. (2004) The well-characterized synthetic molecule: A role for quantitative 1H NMR. Org. Process Res. DeV. 8, 381–384. Rizzo, V., and Pinciroli, V. (2005) Quantitative NMR in synthetic and combinatorial chemistry. J. Pharm. Biomed. Anal. 38, 851–857. Pauli, G. F., Jaki, B. U., and Lankin, D. C. (2005) Quantitative 1HNMR: Development and potential of a method for natural products analysis. J. Nat. Prod. 68, 133–149. Pauli, G. F., Jaki, B. U., Lankin, D. C., and Walter, J. A. (2008) Quantitative NMR of bioactive natural products. In BioactiVe Natural Products: Detection, Isolation and Structural Determination (Colgate, S. M., and Molyneux, R. J., Eds.) pp 113-141, CRC Press, Boca Raton, FL. Vishwanathan, C. T., Bansal, S., Booth, B., DeStefano, J., Rose, M. J., Sailstad, J., Shah, V. P., Skelly, J. P., Swann, P. G., and Weiner, R. (2007) Workshop/conference reportsQuantitative bioanalytical methods and implementation: Best practices for chromatographic and ligand binding assays. AAPS J. 9, E30–E42.

TX8003328

Obtaining Exposures of Metabolites in Preclinical Species through

Recommend Documents