Article pubs.acs.org/ac
Practical and Efficient Strategy for Evaluating Oral Absolute Bioavailability with an Intravenous Microdose of a Stable Isotopically-Labeled Drug Using a Selected Reaction Monitoring Mass Spectrometry Assay Hao Jiang,*,† Jianing Zeng,*,† Wenying Li,‡ Marc Bifano,§ Huidong Gu,† Craig Titsch,† John Easter,⊥ Richard Burrell,⊥ Hamza Kandoussi,† Anne-Françoise Aubry,† and Mark E. Arnold† †
Bioanalytical Sciences/Analytical and Bioanalytical Sciences, ‡Biotransformation/Pharmaceutical Candidate Optimization, §Clinical Pharmacology/Drug Metabolism and Clinical Pharmacology, Bristol-Myers Squibb, Princeton, New Jersey, United States ⊥ Discovery Chemistry Synthesis, Bristol-Myers Squibb, Wallingford, Connecticut, United States S Supporting Information *
ABSTRACT: A strategy of using selected reaction monitoring (SRM) mass spectrometry for evaluating oral absolute bioavailability with concurrent intravenous (IV) microdosing a stable isotopically labeled (SIL) drug was developed and validated. First, the isotopic contribution to SRM (ICSRM) of the proposed SIL drug and SIL internal standard (IS) was theoretically calculated to guide their chemical synthesis. Second, the lack of an isotope effect on drug exposure was evaluated in a monkey study by IV dosing a mixture of the SIL and the unlabeled drugs. Third, after the SIL drug (100 μg) was concurrently IV dosed to humans, at Tmax of an oral therapeutic dose of the unlabeled drug, both drugs in plasma specimens were simultaneously quantified by a sensitive and accurate SRM assay. This strategy significantly improves bioanalytical data quality and saves time, costs, and resources by avoiding a traditional absolute bioavailability study or the newer approach of microdoses of a radio-microtracer measured by accelerator mass spectrometry.
O
health authority regulations permitting the use of microdoses of a drug to obtain early pharmacokinetic information,7 absolute bioavailability has been determined by administering a microdose (≤100 μg and ≤1% of active dose) of a 14C-labeled drug, mitigating the requirement to develop and test the safety of an IV formulation. Quantification of the radiolabeled drug is done by application of accelerator mass spectrometry (AMS) which provides sensitivity at ∼1 pM level for many small molecule drugs.8−11 However, there is a significant drawback in this application in that only the 14C atom of the labeled drug in the samples can be detected with AMS, while the unlabeled PO drug has to be detected with another bioanalytical technique (typically LC-MS/MS). Systematic errors that may occur during sample analyses on two separate instrumental platforms increase the variability in the results obtained. In addition, there are disadvantages of using AMS in bioanalytical applications, which cannot be overcome due to the nature of the technique: (a) very complicated and labor intensive sample processing and, therefore, low sample analysis throughput, (b) not amenable to the use of an internal standard (IS) to improve
ral drug absolute bioavailability describes the actual or absolute amount of drug absorbed from an oral (PO) formulation, expressed as the area under curve (AUC) of a PO formulation relative to that of an intravenous (IV) formulation. The absolute bioavailability study provides valuable information on the fundamental pharmacokinetic parameters of absorption, volume of distribution, and clearance in drug development. Human absolute bioavailability information is required by some regulatory agencies as part of the drug registration. 1 Conducting an absolute bioavailability study in the conventional two-period crossover design doubles the cost and time compared to a single period study design, and day-to-day pharmacokinetic variability cannot be avoided. Most importantly, the traditional approach requires one to conduct preclinical toxicity tests to ensure adequate safety of the IV formulation in human and significant formulation efforts to overcome issues of solubility or stability of the drug in the IV formulation. The development of an IV formulation at a full therapeutic dose can be technically challenging, especially for low-solubility drugs.2 An alternative approach, using a stable isotopically labeled (SIL) drug as IV formulation concurrently dosed with the unlabeled PO drug at a similar dose and using mass spectrometry to distinguish the isotopes by mass difference was initially proposed to allow a single period absolute bioavailability clinical study.3−6 With the changes in © 2012 American Chemical Society
Received: August 28, 2012 Accepted: October 29, 2012 Published: October 29, 2012 10031
dx.doi.org/10.1021/ac3024558 | Anal. Chem. 2012, 84, 10031−10037
Analytical Chemistry
Article
experimental product ion scans in positive electrospray ionization (ESI) mode on a triple quadrupole mass spectrometer (AB SCIEX Triple Quad5500). Second, isotopic abundances (IA) of the neutral loss fragments and the product ion were calculated using online isotope pattern calculators.20 A series of labeling designs were proposed for labeling on either the neutral loss fragments or the product ion. The ICSRM were calculated17 by multiplying the isotopic abundances of the neutral loss (NL) isotope with n mass increase by the isotopic abundances of the product ion (P) isotope with p mass increase, i.e., ICSRM = IANLn × IAPp. Third, a solution of daclatasvir in acetonitrile/water (50:50, v/v) was analyzed by monitoring all proposed SRM channels (precursor ions → product ions), and the experimental ICSRM was compared to the corresponding calculated values. Fourth, concentration ratios of PO drug and IV drug at each time point were predicted using available pharmacokinetic data from oral clinical trials and the simulated IV pharmacokinetic data. Appropriate SIL compounds were selected as IV drug or the SIL internal standard based on the ICSRM impact%, which was calculated by multiplying the ICSRM by the predicted concentration ratios of the “light” compound to the “heavy” compound. The SIL drug with least ICSRM impact% was selected as IV drug and the SIL IS. Quantitative Determination of Daclatasvir and [15N4,13C2]Daclatasvir in Human Plasma. The mixture of daclatasvir and [15N4,13C2]daclatasvir was spiked into human blank plasma to obtain standard curve samples (at 0.02/2, 0.05/5, 0.1/100, 0.5/50, 1/100, 5/500, 8/800, and 10/1000 ng/mL for [15N4,13C2]daclatasvir/daclatasvir) and QC samples (at 0.02/2, 0.05/5, 0.5/50, 5/500, 8/800, and 20/2000 ng/ mL). The samples (200 μL) were extracted with methyl-tertbutyl ether (600 μL) after mixing with the IS (50 μL of [13C10]daclatasvir solution at 100 ng/mL) and buffering with 1.0 M ammonium bicarbonate in water (50 μL). The mixture was separated into two layers by centrifugation (2000 g for 3 min), and the upper organic solvent layer was transferred to clean 96-well polypropylene plates followed by evaporating under nitrogen at room temperature. The dry extracts were reconstituted with 100 μL of a reconstitute solution of acetonitrile/5 mM ammonium acetate in water/acetic acid (50:50:0.01, v/v/v) and injected (10 μL) into HPLC analytical columns (Atlantis dC18, 2.1 mm × 50 mm, 3 μm, Waters Co.). The two coeluted analytes (retention time at ∼2.7 min) were chromatographically separated from endogenous substances with a 3-min linear gradient of 20−60% mobile phase B. Mobile phase A was 5 mM ammonium acetate in water with 0.01% acetic acid, and mobile phase B was acetonitrile. The HPLC eluate was introduced into a positive electrospray ionization (ESI) source of the mass spectrometer and simultaneously monitored in three SRM channels, m/z 739.4 > 565.3 (daclatasvir), m/z 745.4 > 571.3 ([15N4,13C2]daclatasvir), and m/z 749.4 > 575.3 ([13C10]daclatasvir). The concentrations of two analytes were quantitatively determined by back-calculating the chromatographic responses (peak area ratios of the analyte to the internal standard) of each analyte from the regressed standard curves (linear regression model with a 1/x2 weighting) in every bioanalytical run. The SRM assay was validated according to Food and Drug Administration (FDA) Guidance for Bioanalytical Method Validation21 and the Standard Practice Procedures (SOPs) of Bristol-Myers Squibb. The assay recovery and the matrix factor of each analyte in human plasma during liquid−liquid extraction was determined at low
the precision and accuracy of the sample analyses, (c) need for complete chromatographic separations of the 14C-drug from all metabolites containing the 14C labeling, (d) high sample analysis cost (∼0.5 million dollars/study) from only a few laboratories in the world, and (e) relatively long data turnaround time due to the labor-intensive sample processing. With the recent significant improvements in mass spectrometer ionization efficiencies, ion transmission, and the latest generation detection systems, the detection limit of selected reaction monitoring (SRM) assays has been further lowered to a low pM level. This sensitivity makes it possible to utilize a microdose of a SIL drug as IV formulation instead of a 14C labeled drug in these single period absolute bioavailability studies. Both the SIL drug and the unlabeled drug can be simultaneously quantified by the SRM assay, which overcomes the concerns related to the application of AMS. Optimization of SRM assay conditions to reach a required LLOQ at low pg/mL level is essential during the SRM assay development. In addition to pushing mass spectrometer’s sensitivity to the limit by optimizing detection parameters, there are some other techniques that can be used to further improve the sensitivity, such as microflow liquid chromatography,12 captive spray ionization,13 chemical derivatization,14 turbulent flow chromatography,15 etc. However, sometimes the unmet sensitivity of an SRM assay is still a concern in the application of IV microdosing SIL drug to oral absolute bioavailability studies, if a drug’s volume distribution is too large which results in a very low drug’s plasma concentration. In this situation, AMS is still an alternative approach for detecting microdose 14C drug due to ∼10 times or better sensitivity, even though there are some disadvantages in this technique. In the present study, SRM was successfully applied in support of an oral absolute bioavailability study for daclatasvir, a potent inhibitor of hepatitis C virus NS5a protein, being developed by Bristol-Myers Squibb for the treatment of hepatitis C infections.16 In developing the SRM assay, the chemical structures of SIL drug for IV dosing and the SIL IS were designed to minimize isotopic interference from the unlabeled drug, based on a theoretical calculation of isotopic contribution to the SRM (ICSRM) of the SIL drug and IS.17 Unlike a radiolabeled drug, which contains a single labeled atom, a SIL drug typically contains 3 or more labeled atoms. In some instances, SIL drugs have been shown to have different absorption, distribution, metabolism, and excretion (ADME) properties than the unlabeled drug, in particular deuterated drugs.18,19 Since the successful outcome of the clinical study depends on the two drugs being absolutely identical in their metabolism, the SIL drug intended for IV dosing was dosed to monkeys to confirm the absence of an isotope effect on drug exposure. In the human absolute bioavailability study, the IV microdosing was commenced at the time of maximal concentration (Tmax) of the oral therapeutic dose of unlabeled drug. Plasma specimens were collected at the scheduled time points. The specimens were analyzed by a single SRM assay capable of measuring both the stable isotopically labeled and the unlabeled drug. The accurate concentration data in plasma specimens was used to calculate daclatasvir’s oral absolute bioavailability.
■
EXPERIMENTAL SECTION Calculation and Verification of ICSRM of Stable Isotopically Labeled Daclatasvir. First, a major fragmentation pathway of daclatasvir was chosen on the basis of 10032
dx.doi.org/10.1021/ac3024558 | Anal. Chem. 2012, 84, 10031−10037
Analytical Chemistry
Article
Figure 1. Chemical structures of daclatasvir, the proposed SIL compounds, and fragmentation pathways in mass spectrometry. (a) Three major fragmentation pathways were observed from product ion scans in atmosphere pressure ionization (API) positive electrospray mode. The fragmentation occurred on either one side of molecule or both sides of molecule due to the symmetric chemical structure. (b) The formula of product ions and neutral loss fragments for each fragmentation pathway was elucidated for calculating the isotopic contribution to the selected reaction monitoring (ICSRM).
and high concentrations (5.00 and 800 ng/mL for daclatasvir and 0.050 and 8.00 ng/mL for [13C2,15N4]daclatasvir) according to the published procedures.22 Isotope Effect on Daclatasvir Exposure in Vivo. A mixture solution of daclatasvir and [13C10]daclatasvir (1:1, w/ w) was administered IV (1.5 mg/kg) to three monkeys. Blood samples were collected at scheduled time points of 0.16, 0.25, 0.50, 0.75, 1, 2, 4, 6, 8, and 24 h. The concentrations of daclatasvir and [13C10]daclatasvir in plasma were determined by the SRM assay in which an analogue compound was used as the internal standard (the stable label IS not being available at the time of analysis). Human Oral Absolute Bioavailability Study for Daclatasvir. Eight healthy male and female subjects were administered a single oral dose of daclatasvir, 60 mg tablet, followed 1 h later by an intravenous (pushed over a 3-min period) dose of 100 μg [15N4,13C2]daclatasvir. Blood samples were collected at scheduled time points of 0.5, 1, 1.125, 1.25, 1.5, 1.75, 2, 2.5, 3, 4, 5, 6, 8, 10, 12, 16, 24, 36, 48, and 72 h. The concentrations of daclatasvir and [15N4,13C2]daclatasvir in plasma after liquid−liquid extraction were quantitatively determined by the validated SRM assay. The absolute bioavailability (F) of daclatasvir was estimated as a ratio of dosed-normalized exposure, i.e., PO dosing vs IV dosing, specifically
F=
[AUC(INF)]po /dose po [AUC(INF)]iv /doseiv
■
RESULTS Stable Isotopically Labeled Drug Design. Triple quadrupole mass spectrometer in SRM mode detection is a specific, sensitive analytical technique and the most popular bioanalytical tool for quantitative determination of drug concentrations in biological matrixes. However, quadrupole mass analyzers with unit mass resolution (separating each mass from the next integer mass) cannot distinguish different molecular isotopes with the same integer mass; therefore, the natural abundance isotopes in the unlabeled drug have the potential to interfere with the SRM detection of SIL drugs. In this study, daclatasvir (C40H50N8O6, Figure 1a) has significant isotope peaks (M + 1, 47.033%; M + 2, 12.054%; M + 3, 2.203%; M + 4, 0.318%; M + 5, 0.038%; M + 6, 0.004%, etc.) in its quadrupole mass spectrum due to the contributions of 13C (1.0816%), 2H (0.0115%), 15N (0.3694%), 17O (0.0381%), and 18O (0.2055%). These isotope peaks may significantly interfere with SRM detection of the SIL daclatasvir. Therefore, the ICSRM was calculated to guide the SIL compound design. There are publications23−25 and online tools20,26 for calculating isotopic abundances. However, the calculation result is only applicable for a single stage mass 10033
dx.doi.org/10.1021/ac3024558 | Anal. Chem. 2012, 84, 10031−10037
Analytical Chemistry
Article
spectrometer with detection of entire molecules. Recently, we have reported17 that the ICSRM can be accurately calculated on the basis of the isotopic abundances of fragment ions and the placement of the labeled atoms and proposed a strategy for minimizing these interferences. Briefly, the SRM is dependent on the formation of two fragments from a molecular ion, a neutral loss fragment and a product ion. On the basis of probability theory, the ICSRM can be simply calculated by multiplying the IA of the neutral loss (NL) isotope with n mass increase by the IA of the product ion (P) isotope with p mass increase, i.e., ICSRM = IANLn × IAPp, Table S1 and S2, Supporting Information. For example, the ICSRM of daclatasvir to the SRM detection at m/z 743.4 > 569.3 (M + 4, 0 mass increases on the neutral loss fragment and 4 mass increases on the product ion) was calculated to be 0.122727%, with labeling on the product ion fragment by any four atomic isotopes in the chemical structure, such as 13C, 15N, 2H, etc. As expected, the isotopic abundances of the precursor ion (Table S1, Supporting Information) were much higher than the calculated ICSRM (Table S2, Supporting Information), suggesting that using isotopic abundances of the precursor ion will overestimate the isotopic contribution in an SRM assay. Using this approach, the ICSRM for three proposed SIL drugs were calculated (Table S3, Supporting Information). The experimental results confirmed the calculations as shown in Table S3 (Supporting Information). In addition, the ICSRM is related to the drug fragmentation pattern, since it is determined by the isotopic abundances of both neutral loss isotopes and product ion isotopes. The ICSRM under different fragmentation patterns (Figure 1a,b) were calculated and verified by the experimental results (data not shown). To select appropriate SIL compounds as IV drug or the IS, the impact of ICSRM on the SRM assay accuracy was evaluated. This impact is related to the concentration ratio of the “light” compound (CL) to the “heavy” compound (CH) in specimens; i.e., the higher the concentration ratio is, the larger is the impact from the “light” compound, and the impact is expressed as:
Figure 2. Calculated ICSRM impact% for selecting different SIL compounds as the IV drug and the IS. The ICSRM impact% was calculated by multiplying the ICSRM by the corresponding concentration ratio of “light” compound to “heavy” compound (ICSRM impact% = ICSRM × CL/CH). PO (Cmax, 50 mg), the clinical PO maximum concentration at 50 mg dose; IV1(C0, F100%) and IV2(C0, F10%), the simulated IV time zero concentration at bioavailability (F) of 100% and 10%; IV3, the minimal IV drug concentration set as the LLOQ; IS, the predetermined concentration of the internal standard.
a 1:1 mixture of [13C10]daclatasvir and unlabeled daclatasvir was conducted. Even though [13C10]daclatasvir was eventually selected as the IS, based on the calculated ICSRM, it was used as the probe drug in the monkey study because the study was conducted before all SIL drugs had been synthesized. It represents the worse case for metabolism, with the labeling of four 13C on every pyrrolidine ring (the major metabolic site) and one 13C on every imidazole ring (Figure 1a). In the results, isotope effects on drug exposure were not observed as evidenced by the consistent drug exposure ratios over 24 h, Figure 3, indicating that the monkey’s distribution, metabolism, and excretion of the SIL and the unlabeled daclatasvir are equivalent. This result provided the basis for concurrently dosing a SIL drug and an unlabeled drug for the absolute bioavailability study in humans. Evaluation of ICSRM Impact% on Data Accuracy in the SRM Assay. On the basis of the SIL design described above, [15N4,13C2]daclatasvir, selected as IV drug, and [13C10]daclatasvir, selected as the IS, were synthesized with an isotopic purity of 99.6% (impurities: [M + 4] 0.4%, [M + 0] < 0.1%, others < 0.1%) and 93.2% (impurities: [M + 9] 6.5%, [M + 8] 0.3%, [M + 0] < 0.1%, others < 0.1%), respectively. The ICSRM impact% of unlabeled daclatasvir on the SRM of [13C2,15N4]daclatasvir at low and high QC levels (0.05/5 and 8/ 800 ng/mL, expressed as [13C2,15N4]daclatasvir/daclatasvir) was evaluated by adding daclatasvir to the low and high QC
ICSRM impact% = ICSRM × C L /C H
For daclatasvir, three SIL compounds, [15N4], [15N4,13C2], and [13C10]daclatasvir, were proposed as either the IV drug or the IS. The impacts of ICSRM were calculated on the basis of the calculated ICSRMs (Table S3, Supporting Information), the estimated concentrations of the PO drug and the IV drug in plasma samples, and the predetermined concentration of the IS (Figure 2). The ICSRM impact% of PO drug on the IV drug (0.51%) and the ICSRM impact% of the IV drug on the IS (0.02%) were not significant. However, in Figure 2a,b,c, the ICSRM impact% of the “light” compounds on IV drug was significant, 3.56%, 440.50%, 72.65%, respectively. This was because (a) the concentration ratio of the “light” compound to the IV drug (“heavy” compound) was large due to the very low IV drug concentrations; (b) the ICSRMs were significant (Figure 2b,c) due to the 4 mass unit differences between the “light” compound and the IV drug (“heavy” compound) in comparison to a 6 mass unit difference as shown in Figure 2a,d. Overall, the design in Figure 2d indicated the least ICSRM impact%, so [15N4,13C2]daclatasvir was selected as the IV drug and [13C10]daclatasvir was selected as the IS. Isotope Effect on SIL Drug Exposures. To demonstrate that the SIL daclatasvir has identical ADME properties to the unlabeled daclatasvir, a monkey study with IV administration of 10034
dx.doi.org/10.1021/ac3024558 | Anal. Chem. 2012, 84, 10031−10037
Analytical Chemistry
Article
accuracy at the LLOQ for IV postdose samples, because the maximal concentration ratios predicted in the study were less than 1000 which is 100 times less than 105 (data not shown). Thus, for the clinical plasma samples with concentration ratios less than 1000, the ICSRM impact% should be less than 1.05%, which is not significant, whereas the ICSRM was expected to be significant for IV predose samples which contained high concentrations of unlabeled daclatasvir in plasma, since IV dosing starts at Tmax of PO dosing. This impact was confirmed during study sample analysis in that [13C2,15N4]daclatasvir signals were present and greater than the LLOQ in some of the IV predose samples. However, these data were not included in the pharmacokinetic calculation as they were considered as predose zero concentrations. In addition, the impact on the quantitation of [13C2,15N4]daclatasvir for the terminal phase samples, whose concentrations were close to the LLOQ of 0.02 ng/mL, was not significant either. The impact is related to the concentration ratios of the PO drug vs the IV drug but not the absolute concentration. The concentration ratios in the terminal phase samples were predicted to be the same as other samples collected after drug absorption phase. SRM Assay Validation. Developing and validating an accurate and precise SRM assay to support plasma sample analyses for the IV and PO drugs are crucial to an absolute bioavailability study. The mixture of the unlabeled daclatasvir and [15N4,13C2]daclatasvir was spiked into human blank plasma for preparing standard curve samples and quality control samples (QC). There was a 100-fold difference between the two analytes due to consideration of different concentration ranges of two analytes in human plasma samples. This significantly avoided sample dilution and allowed one to simultaneously analyze both analytes in one analysis. [13C10]daclatasvir was added to the samples as the IS prior to the sample extraction. The samples were prepared by liquid− liquid extraction with methyl tert-butyl ether to remove plasma matrix components, followed by the SRM analysis. Assay selectivity and specificity were demonstrated by the results from six different lots of blank plasma with or without the spiked IS and with the spiked analytes at the LLOQ levels, 0.02 ng/mL for [15N4,13C2]daclatasvir and 2 ng/mL for daclatasvir. No analyte peaks were detected in the blank samples with or without spiked IS at the expected retention time, indicating no interference from endogenous substances or from the IS with the analyte detection, Figure S1 (Supporting Information). The two analytes at the LLOQ levels showed significant chromatographic peaks with signal-to-noise ratios greater than 5. The deviations of the calculated concentrations from the nominal values were within ±20.0% for all six LLOQ samples from different matrix lots for each analyte. On the basis of the validation results, the validated SRM assay was determined to be accurate and precise. The standard curves, which ranged from 0.02 to 10 ng/mL for [15N4,13C2]daclatasvir [response = 3.74459 × 10−2 × concentration + (−2.03822 × 10−6), R2 = 0.998825] and 2 to 1000 ng/mL for unlabeled daclatasvir [response = 4.00710 × 10−3 × concentration + (9.93042 × 10−5), R2 = 0.999655], were fitted to a 1/x2 linear regression model. The regression model was selected on the basis of a statistical analysis of data for calibration standards from all accepted validation runs. The intra-assay precisions, based on four levels of analytical QCs (low, mid, geometric mean of low and mid and high), were within 6.0% CV, and interassay precisions were within 3.2% CV for both analytes. The assay accuracy, expressed as percent deviation, was within ±2.5% of
Figure 3. Pharmacokinetic profiles of [13C10]daclatasvir (a) and daclatasvir (b) and concentration ratios (c) of [13C10]daclatasvir and daclatasvir after 24 h of an intravenous dose to monkeys. Three monkeys were administrated with a mixture (1:1, w/w) of [13C10]daclatasvir and unlabeled daclatasvir. The concentrations of [13C10]daclatasvir and daclatasvir in plasma specimens and the control samples spiked with the drug mixture were detected with the SRM assay. A consistent concentration ratio of [13C10]daclatasvir and daclatasvir over 24 h demonstrates that stable isotopic labeling has no significant isotope effect on the drug exposure. The broken line indicates the [13C10]daclatasvir/daclatasvir (13C/12C) drug concentration ratio in the control samples.
samples to reach a concentration ratio (daclatasvir/[13C2,15N4]daclatasvir) of 1000, i.e., 0.05/50 and 8/8000 ng/mL. The concentration of [13C2,15N4]daclatasvir was determined, and the percent deviation against the nominal concentration was within ±4.2%, indicating that there was no significant ICSRM impact% on the SRM of [13C2,15N4]daclatasvir when daclatasvir was present at a 1000-fold ratio to the SIL drug. In addition, ICSRM impact% on data accuracy at the lower limit of quantification (LLOQ) was evaluated by spiking unlabeled daclatasvir (2000 ng/mL), [13C2,15N4]daclatasvir (20 ng/mL), and [13C10]daclatasvir (100 ng/mL) into blank human plasma, respectively. The results suggested that there was no significant ICSRM impact% of [13C2,15N4]daclatasvir on the SRM of [13C10]daclatasvir. However, the unlabeled daclatasvir, at a concentration of 2000 ng/mL in plasma, showed a 105% contribution to the SRM of [13C2,15N4]daclatasvir at the LLOQ of 0.02 ng/mL. This contribution was from the ICSRM of the unlabeled daclatasvir with a concentration ratio of 105 (= 2000/ 0.02 ng/mL). However, this would not impact the data 10035
dx.doi.org/10.1021/ac3024558 | Anal. Chem. 2012, 84, 10031−10037
Analytical Chemistry
Article
would have good bioavailability which is in agreement with the results of this study. The detailed pharmacokinetic data will be reported separately. From a bioanalytical point of view, the concentration ratio of the PO drug vs the IV drug in each sample was important for evaluating ICSRM impact% as discussed above. The concentration ratios from the study samples are presented in Figure 4b, showing that the concentration ratios were within a range of 42−535. The concentration ratios of subject 7 kept increasing and were in a lower range. This was due to a lower PO drug exposure and comparable IV drug exposures compared to other subjects, possibly caused by a slow PO absorption. In addition, there was a ratio dip on the plot for subject 3, which was caused by IV drug concentration rises at 8 and 10 h. The reason for this concentration rise is still not clear, considering that the concentrations of the PO and the IV drugs in each of these two samples were quantified simultaneously in one SRM assay. In this study, a fixed PO/IV drug concentration ratio of 100 was applied to the standard curve and QC samples to avoid mass detector saturation. An interference test had been conducted in which the PO/IV drug concentration ratio was 1000 in the QC samples, and the results demonstrated no impact to the IV drug determination.
the nominal concentration values for both analytes. The stability of the analytes in solvents and human plasma was evaluated to ensure analyte integrity during sample processing and storage. The stability of both analytes in human plasma was demonstrated at room temperature, −20 °C frozen, and multiple freeze−thaw cycles; the stabilities in processed samples, fresh human blood, and spiking solutions were also confirmed. Assay recoveries for both analytes and the IS were in the range of 38.8−48.5%, and there was no significant matrix effect (matrix factors in the range of 0.96−1.01) on the SRM assay. Application of the SRM Assay to a Human Oral Absolute Bioavailability Study. The plasma specimens, collected at the scheduled time points over 72 h (>5 half-lives), were analyzed using the validated SRM assay for the concentrations of [15N4,13C2]daclatasvir and unlabeled daclatasvir. The accuracy (≤± 5.2% deviation from the nominal concentrations) and precision (≤3.7%CV) from the QC samples in the sample analysis runs demonstrated good bioanalytical assay performance. Incurred sample reanalysis (ISR) for assay reproducibility was conducted by reanalyzing 10% of the study samples. Each of a sample’s results (initial and repeat) was within 2.3% of the mean of the two values, which demonstrated the good data reproducibility. The area under the curve (AUC) of the concentration−time graph for each drug and subject was calculated for estimating the oral absolute bioavailability of daclatasvir. The average concentration−time is presented in Figure 4a. The AUC for each drug and subject was calculated for estimating the oral absolute bioavailability of daclatasvir. The absolute bioavailability in seven out of eight subjects ranged within 66.2−78.9% (with a relative standard deviation of 5.7%), except one subject (subject 7) with a value of 35.6%. The values for oral bioavailability in the animal species ranged from 38% to 123% which suggested that humans
■
DISCUSSION This is the first report to the authors’ knowledge that SRM was applied to a human oral absolute bioavailability study that was conducted by concurrent IV microdosing of a SIL drug and oral dosing of an unlabeled drug. The strategy for calculation of ICSRM, evaluation of isotope effect on SIL drug exposure, and development of an accurate and precise SRM assay has been fully validated for this study, which is very critical to a successful absolute bioavailability study. Calculation of ICSRM provides a good reference and guide for chemically synthesizing SIL compounds as IV drug and IS to eliminate isotopic contribution in an SRM assay; evaluation of isotope labeling effect on drug exposures confirms the SIL IV drug has identical distribution and elimination properties in vivo as the PO drug; an accurate and precise SRM method ensures accurate drug concentration data for calculating pharmacokinetic parameters. Worth mentioning, the method of calculating ICSRM is simple and accurate and is very helpful for selecting an appropriate SIL IS during SRM assay development, which can be widely applied in bioanalytical quantitation to help improve data accuracy and precision. It is recommended to conduct an animal study to confirm the lack of isotope effect on ADME. While 13C- or 15N-labeled drugs, as is this case in this paper, do not typically have an effect on ADME, a small study will demonstrate the absence of the effect. It is especially recommended when deuterated drugs are used, because deuterated drugs have been reported to have different ADME properties than the unlabeled drug. A species which has the closest drug metabolism to human is recommended. In this study, we applied a simple design that only 1:1 mixture of [13C10]-labeled and the unlabeled drugs was intravenously dosed to monkeys to confirm if the labeled analogue would be discriminated in vivo. The dosing route was chosen as the same as planned for the microdose in humans. The results demonstrated the labeled analogue was not recognized as a different form from the unlabeled drug and was identically distributed, metabolized, and excreted, indicating that the labeled drug was a good IV microtracer that had the same ADME as the PO unlabeled drug. In regards to the
Figure 4. Representative plots of concentration vs time (a) and concentration ratios (b) of daclatasvir (PO drug, 60 mg) and [15N4,13C2]daclatasvir (IV drug, 0.1 mg, dosed at estimated Tmax of 1 h) in 8 healthy subjects. 10036
dx.doi.org/10.1021/ac3024558 | Anal. Chem. 2012, 84, 10031−10037
Analytical Chemistry
Article
GuidanceComplianceRegulatoryInformation/Guidances/ucm078933. pdf. Accessed May 18, 2012. (8) Barker, J.; Garner, R. C. Rapid Commun. Mass Spectrom. 1999, 13, 285−293. (9) Garner, R. C. Curr. Drug Metab. 2000, 1, 205−213. (10) Lappin, G.; Stevens, L. Expert. Opin. Drug Metab. Toxicol. 2008, 4, 1021−1033. (11) Gao, L.; Li, J.; Kasserra, C.; Song, Q.; Arjomand, A.; Hesk, D.; Chowdhury, S. K. Anal .Chem. 2011, 83, 5607−5616. (12) Duan, X.; Weinstock-Guttman, B.; Wang, H.; Bang, E.; Li, J.; Ramanathan, M.; Qu, J. Anal. Chem. 2010, 82, 2488−2497. (13) Dahal, U. P.; Jones, J. P.; Davis, J. A.; Rock, D. A. Drug Metab. Dispos. 2011, 39, 2355−2360. (14) Niwa, M. Bioanalysis 2012, 4, 213−220. (15) Rudewicz, P. J. Bioanalysis 2011, 3, 1663−1671. (16) Gao, M.; Nettles, R. E.; Belema, M.; Snyder, L. B.; Nguyen, V. N.; Fridell, R. A.; Serrano-Wu, M. H.; Langley, D. R.; Sun, J. H.; O’Boyle, D. R, 2nd; Lemm, J. A.; Wang, C.; Knipe, J. O.; Chien, C.; Colonno, R. J.; Grasela, D. M.; Meanwell, N. A.; Hamann, L. G. Nature 2010, 465, 96−100. (17) Gu, H.; Wang, J.; Aubry, A. F.; Jiang, H.; Zeng, J.; Easter, J.; Wang, J. S.; Dockens, R.; Bifano, M.; Burrell, R.; Arnold, M. E. Anal. Chem. 2012, 84, 4844−4850. (18) Van Langenhove, A. J. Clin. Pharmacol. 1986, 26, 383−389. (19) Sharma, R.; Strelevitz, T. J.; Gao, H.; Clark, A. J.; Schildknegt, K.; Obach, R. S.; Ripp, S. L.; Spracklin, D. K.; Tremaine, L. M.; Vaz, A. D. Drug Metab. Dispos. 2012, 40, 625−634. (20) Isotope Pattern Calculator v4.0. http://yanjunhua.tripod.com/ pattern.htm. Accessed May 20, 2012. (21) Guidance for Industry: Bioanalytical Method Validation. http:// www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/ucm070107. pdf. Accessed May 21, 2012. (22) Jiang, H.; Zeng, J.; Kandoussi, H.; Liu, Y.; Wang, X.; Bifano, M.; Cojocaru, L.; Ryan, J.; Arnold, M. E. J. Chromatogr., A 2012, 1245, 117−121. (23) Yergey, J. A. Int. J. Mass Spectrom. Ion Phys. 1983, 52, 337−349. (24) Kubinyi, H. Anal. Chim. Acta 1991, 247, 107−119. (25) Rockwood, A. L.; Van Orden, S. L.; Smith, R. D. Anal. Chem. 1995, 67, 2699−2704. (26) Sheffield ChemPuter: Isotope Patterns Calculator. http:// winter.group.shef.ac.uk/chemputer/isotopes.html. Accessed May 26, 2012.
clinical study, since the IV drug was microdosed at Tmax of PO dosing, the IV drug (1/600 dose of the PO drug) as a small portion of total daclatsvir was disposed, metablized, and excreted in the same manner as the PO drug due to lack of isotope effect in vivo. Therefore, it is not required to demonstrate the pharmcokinetics is linear. Because the two analytes and the IS have the same chemophysical properties and were coeluted from the high performance liquid chromatography (HPLC) column, significant ionization competition between the analytes and the IS was found. The IS responses in all nonblank samples were expected to be the same due to the sample amount of the IS being spiked into the samples in the SRM assay. However, the IS response was suppressed by high concentration of daclatasvir in the samples. [15N4,13C2]daclatasvir concentration was about 100 times less than unlabeled daclatasvir, and therefore, its presence had a very smaller impact (data not shown). By increasing the values of the mass spectrometer’s ion source heating gas, temperature, ionization voltage, and declustering potential, the ionization competition was reduced ∼20%. On the basis of this finding, the standard curve range was truncated from 2−1000 ng/mL to 2−500 ng/mL to minimize the competition; the variability of IS responses decreased from ∼10-fold to ∼3-fold. This phenomenon suggests that a SIL IS is necessary for the SRM assay accuracy, because the response of trace amounts of [15N4,13C2]daclatasvir is suppressed by the high concentration daclatasvir and only the SIL internal standard can track the ionization competition in the SRM assay. In conclusion, this strategy makes it possible to simultaneously IV microdose a SIL drug with a PO therapeutic dosing and quantify both drugs to assess oral absolute bioavailability; it offers significant savings in both resources and time in the in-life and laboratory phases of the study.
■
ASSOCIATED CONTENT
S Supporting Information *
Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected] (H.J.);
[email protected]. (J.Z.). Notes
The authors declare no competing financial interest.
■
REFERENCES
(1) Ings, R. M. Bioanalysis 2009, 1, 1293−1305. (2) Wong, J.; Brugger, A.; Khare, A.; Chaubal, M.; Papadopoulos, P.; Rabinow, B; Kipp, J; Ning, J. Adv. Drug Delivery Rev. 2008, 60, 939− 954. (3) Wolen, R. L. J. Clin. Pharmacol. 1986, 26, 419−424. (4) Preston, S. L.; Drusano, G. L.; Glue, P.; Nash, J.; Gupta, S. K.; McNamara, P. Antimicrob. Agents Chemother. 1999, 43, 2451−2456. (5) Finglas, P. M.; Witthöft, C. M.; Vahteristo, L.; Wright, A. J.; Southon, S.; Mellon, F. A.; Ridge, B.; Maunder, P. J. Nutr. 2002, 132, 936−939. (6) Parr, A.; Gupta, M.; Montague, T. H.; Hoke, F. AAPS J. 2012, 14, 639−645. (7) Guidance for Industry, Investigators, and Reviewers: Exploratory IND Studies. U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), January 2006, http://www.fda.gov/downloads/Drugs/ 10037
dx.doi.org/10.1021/ac3024558 | Anal. Chem. 2012, 84, 10031−10037