Accepted Manuscript Title: Simultaneous determination of capecitabine and its three nucleoside metabolites in human plasma by high performance liquid chromatography–tandem mass spectrometry Author: Pan Deng Cheng Ji Xiaojian Dai Dafang Zhong Li Ding Xiaoyan Chen PII: DOI: Reference:

S1570-0232(15)00155-5 http://dx.doi.org/doi:10.1016/j.jchromb.2015.03.002 CHROMB 19359

To appear in:

Journal of Chromatography B

Received date: Revised date: Accepted date:

4-11-2014 2-3-2015 4-3-2015

Please cite this article as: P. Deng, C. Ji, X. Dai, D. Zhong, L. Ding, X. Chen, Simultaneous determination of capecitabine and its three nucleoside metabolites in human plasma by high performance liquid chromatographyndashtandem mass spectrometry, Journal of Chromatography B (2015), http://dx.doi.org/10.1016/j.jchromb.2015.03.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ip t cr us

Simultaneous determination of capecitabine and its three nucleoside metabolites in human plasma by high

M an

performance liquid chromatography–tandem mass spectrometry Pan Denga, 1, Cheng Jia, b, 1, Xiaojian Daia, Dafang Zhonga, Li Dingb, Xiaoyan Chena,*

Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 501 Haike Road, Shanghai 201203, P.R. China

b

Department of Pharmaceutical Analysis, China Pharmaceutical University, 24 Tongjiaxiang, Nanjing 210009, P.R. China



Corresponding author at: Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 501 Haike Road, Shanghai 201203, P.R. China.

Tel.: +86 21 50800738

ce pt

ed

a

E-mail address: [email protected]

These authors contributed equally to this work.

Ac

1

Page 1 of 34

ip t cr us M an

Highlights

 Simultaneous determination of Cape and three metabolites in human plasma.  The interferences from isotopic species were chromatographically separated. The ex vivo conversion of DFCR to DFUR in human blood was investigated.



The validated method has been successfully applied to a bioequivalence study.

ce pt Ac

Abstract

ed



Capecitabine (Cape) is a prodrug that is metabolized into 5'-deoxy-5-fluorocytidine (DFCR), 5'-deoxy-5-fluorouridine (DFUR), and 5-fluorouracil (5-FU) after oral administration. A liquid chromatographytandem mass spectrometry method for the simultaneous determination of capecitabine and its three metabolites in human plasma was developed and validated. The ex vivo conversion of DFCR to DFUR in human

Page 2 of 34

ip t cr us

blood was investigated and an appropriate blood sample handling condition was recommended. Capecitabine and its metabolites were extracted

M an

from 100 μL of plasma by protein precipitation. Adequate chromatographic retention and efficient separation were achieved on an Atlantis dC18 column under gradient elution. Interferences from endogenous matrix and the naturally occurring heavy isotopic species were avoided. Detection was performed in electrospray ionization mode using a polarity-switching strategy. The method was linear in the range of 10.0-5000 ng/mL for Cape, DFCR, and DFUR, and 2.00-200 ng/mL for 5-FU. The LLOQ was established at 10.0 ng/mL for Cape, DFCR, and DFUR, and 2.00

ed

ng/mL for 5-FU. The inter- and intra-day precisions were less than 13.5%, 11.1%, 9.7%, and 11.4%, and the accuracy was in the range of

ce pt

13.2% to 1.6%, 2.4% to 2.5%, 7.1% to 8.2%, and 2.0% to 3.8% for Cape, DFCR, DFUR, and 5-FU, respectively. The matrix effect was negligible under the current conditions. The mean extraction recoveries were within 105-115%, 92.6-101%, 94.0-100%, and 85.1-99.9% for Cape, DFCR, DFUR, and 5-FU, respectively. Stability testing showed that the four analytes remained stable under all relevant analytical

Ac

conditions. This method has been applied to a clinical bioequivalence study. Keywords Capecitabine; 5-Fluorouracil; Nucleoside metabolites; LCMS/MS

Abbreviations: 5-FU, 5-fluorouracil; Cape, capecitabine; DFCR, 5'-deoxy-5-fluorocytidine; DFUR, 5'-deoxy-5-fluorouridine; EMA: European Medicines Agency; FDA, Food and Drug Administration; HQC, high quality control; IS, internal standard; LLOQ, lower limit of quantification; LQC, low quality control; ME, matrix effect; MF, matrix factor; MQC, middle quality control; MRM, multiple reaction monitoring; SIL, stable isotope labeled; T1/2, half life; TIS, TurboIonspray; ULOQ, upper limit of quantification

Page 3 of 34

ip t cr us

1. Introduction

M an

Capecitabine (Cape) is a fluoropyrimidine carbamate used in the treatment of metastatic breast and colorectal cancers. As a prodrug, Cape is converted into the active agent 5-fluorouracil (5-FU) through a three-step enzymatic process after oral drug administration [1, 2]. First, hydrolysis by carboxylesterases leads to the formation of 5'-deoxy-5-fluorocytidine (DFCR). Second, cytidine deaminase catalyzes the conversion of DFCR to 5'-deoxy-5-fluorouridine (DFUR). Finally, further catabolism by thymidine phosphorylase produces 5-FU [1, 2].

ed

The pharmacokinetic parameters of Cape and its metabolites exhibited marked inter-individual variability [1]. Therefore, to completely

ce pt

evaluate the safety and efficacy of the Cape formulation, the determination of the plasma concentrations of the parent drug, active metabolite, and metabolites closely related to the generation of the active metabolite is necessary. The simultaneous determination of Cape and its metabolites in human plasma remains challenging. Among these three metabolites, DFCR and DFUR are major ones with plasma concentrations

Ac

near the parent drug level [1, 3-7], whereas 5-FU is present in plasma at a significantly lower level [1, 3, 5-7]. Therefore, linearity should be obtained at markedly different analyte concentration ranges. The differences in hydrophobicity between Cape and its metabolites also pose a challenge for the simultaneous analysis of these compounds using LC. In addition, the interference between DFCR and DFUR should be considered because the molecular weights of these compounds differ by only one unit. Moreover, ex vivo transformation of DFCR to DFUR in

Page 4 of 34

ip t cr us

blood by cytidine deaminase should be considered during blood collection and processing. To date, several LCMS/MS methods have been

M an

reported for the determination of Cape and its metabolites in human plasma [3, 8-12]. In some studies, tetrahydrouridine, a cytidine deaminase inhibitor, was added into blood to prevent ex vivo conversion of DFCR to DFUR [8, 13]. In another study, it was suggested that obtained blood samples should be immediately cooled on ice-water, and plasma should be stabilized using tetrahydrouridine for an extended period of storage [9].

ed

In the present study, a sensitive and selective LCMS/MS method with a polarity-switching strategy was developed for the simultaneous

ce pt

quantification of Cape, DFCR, DFUR, and 5-FU in human plasma. LC conditions were optimized to avoid isotopic interferences among DFCR, DFUR, and their isotope-labeled internal standard (IS) 13C, 15N2-DFCR. In addition, ex vivo conversion of DFCR to DFUR was quantitatively investigated using fresh human blood. This method was fully validated according to the US Food and Drug Administration (FDA) [14] and

Ac

European Medicines Agency (EMA) guidelines [15], and successfully applied to characterize the pharmacokinetic profiles of Cape and its three main metabolites in cancer patients after a single oral administration of 2000 mg of Cape. 2. Materials and methods

2.1. Chemical and reagents

Page 5 of 34

ip t cr us

Cape was kindly provided by Qilu Tianhe Pharmaceutical Co., Ltd. (Shandong, China). DFCR, DFUR, d11-Cape,

C,15N2-DFCR, and

C,15N2-5-FU were purchased from TLC Pharmachem., Inc. (Toronto, Ontario, Canada). 5-FU was supplied by the National Institute for the

M an

13

13

Control of Pharmaceutical and Biological Products (Beijing, China). HPLC-grade methanol, acetonitrile, and ammonium hydroxide were obtained from Sigma-Aldrich (St. Louis, MO, USA). Other regents were of analytical grade and were supplied by Tedia (Fairfield, OH, USA).

ed

Deionized water was generated by a Millipore Milli-Q Gradient Water Purification System (Molsheim, France). 2.2. Instrumentation

ce pt

An HPLC system consisting of a DGU-20A3 vacuum degasser, a LC-20AD pump, a CTO-20A column oven, and a SIL-20AC autosampler (Shimadzu, Kyoto, Japan) was used for solvent and sample delivery. Mass spectrometry detection was conducted on an AB Sciex API 4000 triple quadrupole mass spectrometer (Applied Biosystems, Concord, Ontario, Canada) equipped with a TurboIonspray (TIS) interface. AnalystTM

Ac

Version 1.5.2 (Applied Biosystems, Concord, Ontario, Canada) was used for data acquisition.

2.3. Liquid chromatography-tandem mass spectrometry conditions The analytes were separated on an Atlantis dC18 column (100  4.6 mm, 3 μm; Waters, Milford, MA, USA) maintained at 40oC with a C18 guard column (4.0  3.0 mm, 5 μm; Phenomenex, Torrance, CA, USA). The mobile phase consisted of (A) a mixture of 0.025% acetic acid and

Page 6 of 34

ip t cr us

0.0025% ammonium hydroxide solution (pH 3.8), and (B) methanol. A gradient elution method was used (Table 1).

M an

The electrospray ionization (ESI) source was operated with polarity switching during two periods in a single run. The first period, in negative ionization (ESI) mode, was applied between 0 min and 7.5 min. This mode allowed the detection of DFCR, DFUR, 13C,15N2-DFCR, 5-FU, and 13C,15N2-5-FU. The second period, in positive ionization (ESI+) mode, was used for the detection of Cape and its IS d11-Cape between 7.5 min and 10.5 min. Quantification was performed using multiple reaction monitoring (MRM). In the ESI mode, the MS/MS setting

ed

parameters were as follows: 25 psi curtain gas; 50 psi nebulizer gas (GS1); 50 psi turbo gas (GS2); 4500 V ion spray voltage; 550oC source

ce pt

temperature; and 200 ms dwell time. In the ESI+ mode, the MS/MS setting parameters were as follows: 25 psi curtain gas; 50 psi GS1; 50 psi GS2; +4000 V ion spray voltage; 550oC source temperature; and 200 ms dwell time. The MRM transitions and the optimized MS parameters for each analyte are summarized in Table 1.

Ac

2.4. Preparation of standards and quality control (QC) samples Stock solutions of Cape, DFCR, DFUR, and 5-FU at concentrations of approximately 1.00 mg/mL were prepared in methanol. The solutions were serially diluted with 50% aqueous methanol to obtain Cape/DFCR/DFUR/5-FU working solutions of 1.00/1.00/1.00/0.20, 3.00/3.00/3.00/0.50, 10.0/10.0/10.0/1.00, 30.0/30.0/30.0/2.00, 80.0/80.0/80.0/5.00, 200/200/200/10.0, and 500/500/500/20.0 g/mL. Calibration

Page 7 of 34

ip t cr us

standard samples were prepared by spiking 10.0 μL of working solutions into 1000 μL of blank plasma to obtain the final concentrations of 30.0/30.0/30.0/5.00,

100/100/100/10.0,

300/300/300/20.0,

800/800/800/50.0,

2000/2000/2000/100,

and

M an

10.0/10.0/10.0/2.00,

5000/5000/5000/200 ng/mL for Cape/DFCR/DFUR/5-FU. QC solutions were prepared through the separate weighing of standard references. QC samples were independently prepared in blank plasma at four concentrations: 10.0/10.0/10.0/2.00 ng/mL (LLOQ), 25.0/25.0/25.0/5.00 ng/mL (low quality control, LQC), 200/200/200/20.0 ng/mL (middle quality control, MQC), and 4000/4000/4000/160 ng/mL (high quality control,

ed

HQC) for Cape/DFCR/DFUR/5-FU. Stock solutions of d11-Cape,

13

C,15N2-DFCR, and

13

C,15N2-5-FU were prepared in methanol and diluted

13

ce pt

with 50% aqueous methanol to obtain the final IS working solution at 20.0/2000/400 ng/mL for d11-Cape/13C, 15N2-DFCR/13C, and 15N2-5-FU. C,15N2-DFCR was used as the IS for the determination of DFCR and DFUR. All solutions were stored at 4°C and brought to room temperature

before use. Standard and QC samples were stored at 20°C.

Ac

2.5. Conversion of DFCR to DFUR in human fresh blood The stock solution of DFCR prepared in methanol was immediately spiked with K2EDTA anti-coagulated fresh human blood to reach final concentrations of 25.0, 200, and 4000 ng/mL (0.2% of methanol in blood). These spiked blood samples were maintained at three different conditions, i.e., 0oC in an ice-bath, 4oC, and room temperature (21oC). Aliquots of blood (500 μL) at each storage condition were taken at T0, 15

Page 8 of 34

ip t cr

us

min, 30 min, 1 h and 2 h, and centrifuged at 0 oC immediately to separate plasma. Each set consisted of triplicate samples at each level. The plasma concentrations of DFCR and DFUR were determined by the validated LCMS/MS method, and the percentage of conversion was

M an

calculated. 2.6. Sample preparation

To a 100 μL aliquot of the plasma sample, 25 μL of the IS working solution, and 200 μL of methanol were added. The mixture was

ed

vortex-mixed for 1 min and centrifuged at 14000  g for 5 min at 4oC. The supernatant was transferred to another tube and evaporated to dryness at 40oC under a stream of nitrogen in a TurboVap evaporator (Zymark, Hopkinton, MA, USA). The residue was reconstituted in 200 μL of the

2.7. Method validation

ce pt

initial mobile phase, and an aliquot of 20 μL was injected into the LCMS/MS system for analysis.

The validation was performed in order to evaluate the method in terms of selectivity, linearity, precision and accuracy, matrix effect (ME),

Ac

recovery, stability, dilution integrity, and carry-over.

2.8. Application of the method to a clinical bioequivalence study The validated method was applied to investigate the plasma profiles of Cape and its three main metabolites in a pivotal bioequivalence

Page 9 of 34

ip t cr us

study, which included 61 male cancer patients who received a single oral administration of 2000 mg of Cape test or reference formulation under

M an

food intake conditions. This multi-center, randomized, two-way crossover study was performed according to the Declaration of Helsinki. Prior to the start of the study, informed consent was obtained from the participant and the protocol was approved by the local ethical review boards. Blood samples (3 mL) were collected into ice-cooled K2EDTA vacutainers at 0 (pre-dose), 0.33, 0.67, 1.0, 1.33, 1.67, 2.0, 2.5, 3.0, 3.5, 4.0, 5.0, 6.0, and 8.0 h after oral administration, and were immediately centrifuged at 3500  g for 10 min at 0oC to separate the plasma fraction. The

ed

plasma samples were stored at 80°C until analyzed. A total number of 1708 plasma samples were collected and analyzed. The pharmacokinetic

ce pt

parameters of Cape, DFCR, DFUR, and 5-FU were calculated by non-compartmental analysis using WinNonlin 1.5.2 (Pharsight, St. Louis, Missouri, USA). The maximum plasma concentrations (Cmax) and time to reach this value (Tmax) were directly obtained from the experimental data. The elimination rate constant (ke) was calculated using log-linear regression of the terminal portion of each curve. The elimination half-life

Ac

(T1/2) was calculated as 0.693/ke. The area under the curve [AUC(0t), from 0 to the last measurable plasma concentration (Ct)] was calculated using the linear trapezoidal method and was extrapolated to infinity [AUC(0∞)] using the following formula: AUC(0∞) = AUC(0t) + Ct/ke. 2.9. Incurred sample reanalysis (ISR) In this study, 120 plasma samples (7% of study sample size) at the maximal concentration (0.671.67 h) and around the elimination phase

Page 10 of 34

ip t cr us

(3.08.0 h) were selected for ISR to further evaluate the reproducibility of the analytical method. The difference between the ISR data and the

M an

original assay data should be within 20% of the mean for at least 67% of the repeats. 3. Results and discussion 3.1. Mass spectrometric conditions

ed

Cape and its metabolites can be regarded as basic to neutral molecules, with predicted pKa values of 9.5,,  and 8 for Cape, DFCR, DFUR, and 5-FU, respectively (calculated using Advanced Chemistry Development software, ADME Suite, Version 5.0). During the method

ce pt

optimization stage, the MS responses of the analytes were investigated under both ESI+ and ESI ionization modes. Each standard solution was individually infused into the mass spectrometer, and an LC flow with 50% of mobile phase A was also introduced into the mass spectrometer using a post-column “T” connection. Mobile phase A, consisting of water containing 0.1% acetic acid, was used in ESI+ ionization mode, and

Ac

water was used in ESI ionization mode.

For Cape, ESI+ mode produced a much higher MS signal intensity and more stable MS responses than that of ESI. The MS response of DFCR was higher under ESI+ mode, and the MS responses of DFUR and 5-FU under both modes were nearly comparable. These results are in accordance with their individual pKa values, as Cape and DFCR are basic compounds with relatively high pKa values for which protons are

Page 11 of 34

ip t cr us

easily obtained under acid mobile phase condition. In contrast, DFUR and 5-FU are neutral compound that could either lose or gain protons. The

M an

MS responses of the isotope-labeled ISs under ESI+ and ESI behaved similarly to those of the corresponding analytes. In conventional LC–MS methods, in which only one ionization mode (i.e., ESI+ or ESI) is used in a single run, the analysis of components with distinct properties might require double injection. In the current study, polarity switching between ESI+ and ESI modes was used in a single run. To avoid the loss of sensitivity during polarity switch, the MS detection process should be divided into different periods based on the optimized ionization polarity

ed

and LC retention times of the analytes. For these considerations, the analytes were separated into two groups based on their chemical properties

ce pt

(Table 2). The logP values of DFCR, DFUR and 5-FU were similar and much lower than that of Cape. Therefore, similar LC retention characteristics could be expected between DFCR, DFUR, and 5-FU. Further optimization of the MS conditions revealed that chemical noise in the mass spectra for these small molecule metabolites is lower under ESI compared with that under ESI+. Consequently, these three metabolites

Ac

along with their ISs were classified as one group that used ESI detection, and the deprotonated molecules were produced at m/z 244, m/z 245, and m/z 129 for DFCR, DFUR, and 5-FU, respectively. The parent drug was analyzed under ESI+ mode with protonated molecule at m/z 360. The product ion spectra of the analytes and ISs under ESI+ or ESI, as well as their fragmentation profiles, are presented in Fig. 1. The most abundant and characteristic fragment ions were selected in the MRM transitions (Table 1).

Page 12 of 34

ip t cr us

The main challenge in quantifying multiple compounds with different MS signal intensities and dynamic ranges in a single assay by

M an

LC–MS/MS is obtaining linearity and the required sensitivity without signal saturation in the MS. In the present study, peak saturation was observed for Cape. To reduce the ion current at the detector and avoid signal saturation, simple adjustments such as sample injection volume reduction or sample dilution prior to injection may be applied. However, in the present case, these options were not feasible because of the LLOQ required for 5-FU. Therefore, the collision energy setting for Cape was reduced to 5, which lead to the stable and reproducible production

ed

of fragment ions from Cape without signal saturation. For the other analytes, the effects of capillary voltage, source temperature, declustering

ce pt

potential, and collision energy were examined and optimized to achieve optimal sensitivity. Thus, signal saturation for Cape was avoided, and the sensitivity for the metabolites still satisfied the requirements. 3.2. Liquid chromatographic conditions

Ac

The development of an LC method for the simultaneous determination of Cape and its metabolites remains a challenge. The metabolites of Cape are low mass hydrophilic compounds with low logP values (0.9, 1.55, and 1.78 for DFCR, DFUR, and 5-FU, respectively), which could not be readily retained on conventional reversed phase HPLC columns. Moreover, because of differences in the hydrophobicity of Cape and its metabolites, the HPLC conditions should be optimized to ensure the adequate resolution of each analyte as well as a reasonable total

Page 13 of 34

ip t cr

us

chromatographic run time. Furthermore, the naturally occurring heavy isotope ion originating from DFCR ([M  H + 1] m/z 245) may interfere

M an

with the detection of DFUR ([M  H] m/z 245). These two ions could not be distinguished with a unit-resolution triple-quadrupole instrument, and isobaric fragment ions could be produced from the two precursor ions, which might lead to substantial inaccuracy in the DFUR quantification. In addition, crosstalk from the same neutral loss was also a possible contributor to the interference between DFCR and DFUR. Approaches used to avoid such MS interferences might rely on LC separation. Similarly, the heavy isotopic ion ([M  H + 2] m/z 247) of DFUR

and

13

ed

might interfere with the MS detection of 13C,15N2-DFCR ([M  H] m/z 247), further necessitating adequate LC resolution among DFCR, DFUR, C,15N2-DFCR. Based on these considerations, the HPLC conditions were optimized to achieve baseline separation between DFCR and

ce pt

DFUR.

In this experiment, the HPLC columns with different sorbents that are suitable for the analysis of polar compounds were evaluated, including Synergi Hydro-RP C18, Atlantis dC18, ASB C18, Capcell PAK C18 AQ, and Gemini C18. Enhanced retention of the polar hydrophilic metabolites was achieved on the Atlantis dC18 column. The mobile phase composition was optimized to improve the peak shape and the

Ac

separation of the analytes. A gradient elution program was used to achieve adequate resolution while minimizing the chromatographic run time. The results showed that Cape readily produced a tailing peak with the use of ammonium acetate in mobile phase regardless of the pH value adjusted. In addition, the response of 5-FU decreased dramatically when ammonium acetate was added in the mobile phase, and its response decreased inversely with the buffer concentration. Acetic acid was added into the aqueous mobile phase instead, which led to a symmetric peak

Page 14 of 34

ip t cr

us

for Cape. Meanwhile, acetic acid stabilize the MS response of Cape under ESI+ detection. On the other hand, ammonium hydroxide solution could stabilize as well as improve the MS signals of DFCR, DFUR, and 5-FU under ESI detection. It was finally proved that an acetic

M an

acid/ammonium hydroxide solution (pH 3.8) and methanol were adequate to balance peak shape, retention time, and resolution of the four analytes. When the plasma samples were analyzed, an interference peak was observed at the retention time of 5-FU, therefore, the mobile phase was further optimized by adjusting the percentage of aqueous mobile phase, which resolved 5-FU from the interfering peak (Fig. 2). Under the

ed

optimized HPLC method, the retention times of Cape, DFCR, DFUR, and 5-FU were 8.2, 6.4, 6.6, and 4.6 min, respectively (Fig. 2), and mutual interferences were avoided (Fig. 3). The total chromatographic run time was 10.5 min. Although ammonium hydroxide and acetic acid

ce pt

are regarded as high volatile reagents, the retention time as well as the MS response for each analyte remained stable in each analytical run, therefore testified the method robustness.

Based on the optimized LC conditions, the MS detection process was divided into two periods. Each period was operated under one ionization mode to provide the maximum MRM sensitivity and stable detection. DFCR, DFUR, 5-FU,

13

C,15N2-DFCR, and

13

C,15N2-5-FU

Ac

were detected in negative mode from 0 min to 7.5 min. Next, the ESI source was switched to positive mode for Cape and d11-Cape detection from 7.5 min until the end of the detection process. Under the optimized LCMS/MS method, each analyte was well retained and resolved with ideal detection sensitivity. The LLOQ was established at 10.0 ng/mL for Cape, DFCR, and DFUR, and at 2.00 ng/mL for 5-FU. These values meet the bioanalysis requirements, especially for 5-FU, the plasma level of which is the lowest among the four analytes.

Page 15 of 34

ip t cr us

3.3. Conversion of DFCR to DFUR in human blood

M an

Cytidine deaminase in blood could transform DFCR into DFUR ex vivo [8]. It was reported that cytidine deaminase inhibitor, tetrahydrouridine, was added into blood to avoid such unwanted conversion [8, 13]. During the early stage of method development, we attempted to add tetrahydrouridine to blood samples at the recommended concentrations (1 μM and 5 μM). However, deteriorated peak shapes of metabolites were observed after such pretreatment, and the quantification results were irreproducible. Therefore, the stability of DFCR was

ed

investigated in human blood without cytidine deaminase inhibitor. It was found that the conversion of DFCR to DFUR was the lowest at 0oC. The conversion percentage of DFCR to DFUR was 5.70 ± 0.02% when spiked blood samples were kept at 0oC for no more than 15 min, and

ce pt

the values were consistent at three concentration levels. The formation of DFUR was increased to 8.4  10.5% when blood samples were incubated for 2 h at 0oC (Fig. 4). Given that the plasma level of DFCR is comparable to that of DFUR, the conversion percentage of DFCR to DFUR at 5.7% will not influence the reliable quantification of both analytes in human plasma. Therefore, it was recommended that blood sample

Ac

should be kept at 0oC (ice-bath) for less than 15 min before plasma separation. 3.4. Sample preparation

The sample preparation methods reported to determine Cape and its metabolites included solid phase extraction (SPE) [7, 10, 16], liquid-liquid extraction (LLE) [8], and protein precipitation [3, 9, 12]. In the present study, protein precipitation was chosen for sample

Page 16 of 34

ip t cr

us

preparation because of its excellent recovery rates and simplicity. To solve the problem of undesirable ion suppression or enhancement of target analytes caused by the co-eluting endogenous matrix, stable isotope-labeled (SIL) compounds were used as the ISs. Because the

M an

physicochemical properties of SIL compounds are almost identical to those of the unlabeled analytes, they can compensate for the variations during sample preparation and ME. Given the unavailability of the commercial isotope counterpart of DFUR, 13C, 15N2-DFCR was used as the IS for DFUR quantification. Because of the adequate sensitivity of the optimized LCMS/MS method, the plasma sample volume was reduced to

ed

100 L, which is lower than that used in previously reported methods (200-1000 L of plasma) [3, 7, 8, 10, 16]. In published methods utilizing similar protein precipitation method [9, 12], the LLOQ values for 5-FU were established at 50.0 ng/mL, which are higher than the present

ce pt

method. This may be due to substantial dilution of sample before MS analysis [9], specifications of the mass spectrometers, and high aqueous content of the mobile phase at the retention time of 5-FU [12]. 3.5. Method validation

Ac

3.5.1. Assay selectivity and carry-over

The typical chromatograms are shown in Fig. 2. An endogenous interference peak eluted at 3.97 min was observed in the MRM channel of 5-FU, and it was well resolved from 5-FU, whose retention time was 4.62 min. Meanwhile, the isotopic interference peak observed in the MRM channels of DFUR and 13C,15N2-DFCR was separated from the target analyte under the current chromatographic conditions (Fig. 3).

Page 17 of 34

ip t cr us

To assess carry-over, a double blank sample was injected after the ULOQ. Carry-over was deemed insignificant because no signal greater

3.5.2. Linearity of calibration curve and LLOQ

M an

than 20% of the LLOQ and 5% for the ISs was observed at the retention times for any of the analytes or ISs .

Linear regression curves were obtained over the concentration ranges of 10.0−5000 ng/mL for Cape/DFCR/DFUR and 2.00−200 ng/mL for 5-FU, respectively. The LLOQ for Cape/DFCR/DFUR/5-FU was 10.0/10.0/10.0/2.00 ng/mL. The sensitivity of this method was superior or

ed

equivalent to previous published methods for the determination of Cape, DFCR, and DFUR [3, 4, 8, 9, 12]. For 5-FU, the sensitivity was higher than the reported results [3, 8, 9, 11, 12], except for one method, in which the LLOQ was established at 1.00 ng/ml [10]. With the present LLOQ,

ce pt

the plasma concentrations of Cape/DFCR/DFUR/5-FU could be determined for up to 8 h after oral administration of 2000 mg of Cape, which was sensitive enough to allow for the investigation of the pharmacokinetic behavior of Cape, DFCR, DFUR, and 5-FU.

Ac

3.5.3. Precision and accuracy

The intra- and inter-day precision and accuracy values for the QC samples are summarized in Table 3, which were in the acceptable ranges. 3.5.4. Dilution using a blank matrix

A sample was spiked at 40000/40000/40000/1600 ng/mL for Cape/ DFCR/DFUR/5-FU to evaluate the 10-fold dilution at an over-curve

Page 18 of 34

ip t cr

us

level. The results indicated that a 10-fold dilution of the human plasma sample above the ULOQ was acceptable (RSD ≤ 6.2%, RE ranged from

M an

5.7% to 1.4%). 3.5.5. ME and extraction recovery

For each analyte and the IS, the matrix factor (MF) was calculated for each lot of matrix (blank plasma from six different lots), by calculating the ratio of the peak area in the spiked plasma post-extraction samples to the peak area in the water-substituted samples. It was found

ed

that the MFs from six different lots of blank plasma ranged from 84.1% to 116% for all compounds. The inter-subject variability of the IS-normalized MF, as measured by their RSD, was lower than 2.4%. Thus, ion suppression or enhancement from the plasma matrix was

ce pt

negligible under the current conditions.

The mean extraction recoveries obtained from three concentrations of QC plasma samples were 105%, 104% and 115% for Cape, 95.5%, 92.6% and 101% for DFCR, 94.0%, 96.1% and 100% for DFUR, 85.1%, 88.2% and 99.9% for 5-FU, respectively. The mean extraction recovery

3.5.6. Stability

Ac

rates of ISs were between 90.5% and 109%. The results indicated sufficient extraction efficiency.

The results of the stability experiments are presented in Table 4. The results demonstrated the good stability of Cape, DFCR, DFUR, and 5-FU during plasma sample preparation and analysis, as well as under different storage conditions.

Page 19 of 34

ip t cr us

Although several LC–MS/MS methods have been developed to analyze capecitabine and its metabolites, there are still some limitations.

M an

The advantages of the current approach over those reported previously are: (1) the sensitivity was improved, especially for 5-FU [3, 8, 9, 11, 12]; (2) the protein precipitation preparation was employed, which was simpler and faster than the reported SPE [7, 10, 16] and LLE [8] methods; (3) more analytes were determined simultaneously in shorter chromatographic run time, and the isotopic interference between DFCR and DRUR was avoided under the optimized HPLC conditions; (4) It was found that blood sample should be kept at 0oC (ice-bath) for less than 15 min

3.6. Clinical application

ce pt

presented at different levels in blood.

ed

before plasma separation to ensure minimum and consistent ex vivo conversion (5.70 ± 0.02%) of DFCR to DFUR, even though DFCR was

The validated LCMS/MS method was successfully applied to a bioequivalence study of Cape formulations in 61 male cancer patients. Up

Ac

to 1708 plasma samples were analyzed during a period of 15 days. Fig. 5 shows the mean plasma concentration-time profiles of Cape, DFCR, DFUR, and 5-FU following a single oral administration of 2000 mg of the test and reference formulations. The T1/2 values for Cape and its metabolites were approximately 1 h, and the LLOQ of the present method was sufficient to monitor at least eight half-lives of Cape, DFCR, DFUR, and 5-FU concentrations with good intra- and inter-assay reproducibility. The pharmacokinetic profiles of Cape, DFCR, DFUR, and

Page 20 of 34

ip t cr us

5-FU showed high inter-patient variability, which may be attributed to the different enzymatic activities involved in the metabolism process of

M an

Cape and its metabolites. 3.7. ISR

The percent difference between the initial concentration and the concentration measured in the repeat analysis was less than 20% of their mean for 93% of the repeats, therefore indicating good reproducibility of the current validated method.

ed

4. Conclusion

ce pt

This study developed a sensitive and reliable polarity switching LCMS/MS method for the simultaneous determination of Cape, DFCR, DFUR, and 5-FU in human plasma. The conversion of DFCR to DFUR was investigated, and plasma sample collection method was proposed to ensure minimum and consistent ex vivo conversion of DFCR to DFUR. The LLOQ was established at 10.0 ng/mL for Cape, DFCR, and DFUR,

Ac

and 2.00 ng/mL for 5-FU, using a 100-L of plasma sample. This method offers advantages over those previously reported methods, in terms of improved sensitivity, lower sample requirements, simplicity of the extraction procedure, and short chromatographic run time. The reproducibility in the measurement of subject samples was confirmed by ISR.

Page 21 of 34

ip t cr us

References

M an

[1] B. Reigner, K. Blesch, E. Weidekamm, Clin. Pharmacokinet., 40 (2001) 85-104. [2] C.M. Walko, C. Lindley, Clin. Ther., 27 (2005) 23-44.

[3] L.D. Vainchtein, H. Rosing, J.H. Schellens, J.H. Beijnen, Biomed. Chromatogr., 24 (2010) 374-386. [4] Y. Xu, J.L. Grem, J. Chromatogr. B, 783 (2003) 273-285.

ce pt

Cancer Res., 4 (1998) 941-948.

ed

[5] B. Reigner, J. Verweij, L. Dirix, J. Cassidy, C. Twelves, D. Allman, E. Weidekamm, B. Roos, L. Banken, M. Utoh, B. Osterwalder, Clin.

[6] J. Cassidy, C. Twelves, D. Cameron, W. Steward, K. O'Byrne, D. Jodrell, L. Banken, T. Goggin, D. Jones, B. Roos, E. Bush, E. Weidekamm, B. Reigner, Cancer Chemoth. Pharm., 44 (1999) 453-460.

Ac

[7] B. Reigner, S. Clive, J. Cassidy, D. Jodrell, R. Schulz, T. Goggin, L. Banken, B. Roos, M. Utoh, T. Mulligan, E. Weidekamm, Cancer Chemoth. Pharm., 43 (1999) 309-315. [8] D. Montange, M. Berard, M. Demarchi, P. Muret, S. Piedoux, J.P. Kantelip, B. Royer, J. Mass Spectrom., 45 (2010) 670-677. [9] M.J. Deenen, H. Rosing, M.J. Hillebrand, J.H.M. Schellens, J.H. Beijnen, J. Chromatogr. B, 913 (2013) 30-40.

Page 22 of 34

ip t cr us

[10] A. Salvador, L. Millerioux, A. Renou, Chromatographia, 63 (2006) 609-615.

M an

[11] H. Licea-Perez, S. Wang, C. Bowen, J. Chromatogr. B, 877 (2009) 1040-1046. [12] S.M. Guichard, I. Mayer, D.I. Jodrell, J. Chromatogr. B, 826 (2005) 232-237.

[13] T. Besnard, N. Renee, M.C. Etienne-Grimaldi, E. Francois, G. Milano, J. Chromatogr. B, 870 (2008) 117-120. [14] US Department of Health and Human Services, Food and Drug Administration, Guidance for Industry: Bioanalytical Method Validation,

ed

2001 http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM070107.pdf.

www.ema.europa.eu.

ce pt

[15] European Medicines Agency, Guideline on validation of bioanalytical methods,in: EMEA/CHMP/EWP/192217/2009, 2009,

Ac

[16] P. Buchner, E. Mihola, A. Sahmanovic, T. Steininger, C. Dittrich, M. Czejka, Anticancer Res., 33 (2013) 881-886.

Page 23 of 34

ip t cr us

Figure captions

their proposed fragmentation patterns.

Fig. 2. Typical MRM chromatograms of 5-FU (I),

13

M an

Fig. 1. Product ion mass spectra of Cape (A), DFCR (B), DFUR (C), 5-FU (D), d11-Cape (E), 13C,15N2-DFCR (F), and 13C,15N2-5-Fu (G) and

C 15N2-5-FU(II), DFCR (III), DFUR (IV),

13

C,15N2-DFCR (V), Cape (VI), and d11-Cape

ed

(VII). (A) a blank plasma sample; (B) a blank plasma sample spiked with IS (5.00/500/100 ng/mL d11-Cape/13C,15N2-DFCR/13C,15N2-5-FU in plasma); (C) a blank plasma sample spiked with Cape, DFCR, DFUR and 5-FU at the LLOQ (10.0/10.0/10.0/2.00 ng/mL for

Cape.

ce pt

Cape/DFCR/DFUR/5-FU) and IS; and (D) a plasma sample obtained from a patient at 0.67 h after a single oral administration of 2000 mg of

Ac

Fig. 3. Chromatographic separation of DFCR (A), DFUR (B), and 13C, 15N2-DFCR (C) using the optimized LC–MS/MS method.

Fig. 4. Plot of the conversion percentage of DFCR to DFUR as a function of incubation time in pooled fresh human blood at three concentration levels. (A) incubated at 0oC in an ice-bath, (B) incubated at 4oC, (C) incubated at room temperature (21oC).

Page 24 of 34

ip t cr us

Fig. 5. Mean plasma concentration versus time profiles for Cape (A), DFCR (B), DFUR (C), and 5-FU (D) following a single oral administration

Ac

ce pt

ed

M an

of 2000 mg of Cape (T: test formulation; R: reference formulation).

Page 25 of 34

Ac ce p

te

d

M

an

us

cr

ip t

Figure 1

Page 26 of 34

Ac ce p

te

d

M

an

us

cr

ip t

Figure 2

Page 27 of 34

Ac ce p

te

d

M

an

us

cr

ip t

Figure 3

Page 28 of 34

Ac ce p

te

d

M

an

us

cr

ip t

Figure 4

Page 29 of 34

Ac

ce

pt

ed

M

an

us

cr

i

Figure 5

Page 30 of 34

cr

ip t

Tables

129.2 132.2

Time (min) 0.0 2.0

Collision energy (eV) 5 20 20 24 20

Declustering potential (V) 90 60 50 50 50

42.2 44.2

33 33

50 50

Flow rate (mL/min)

Solvent A* (%)

Solvent B* (%)

0.60

100

0

0.60

100

0

an

5-FU 13 15 C, N2-5-FU

Product ion 244.2 255.1 107.0 108.0 131.0

M

Precursor ion 360.2 371.2 244.0 245.0 247.0

d

Gradient elution conditions

Compound Cape d11-Cape DFCR DFUR 13 15 C, N2-DFCR

ep te

MS parameters

us

Table 1. MS parameters and gradient elution conditions for the analysis of Cape, DFCR, DFUR, and 5-FU

Ac c

2.5 0.60 60 40 3.8 0.60 60 40 4.5 0.60 10 90 7.5 0.60 10 90 7.6 0.60 100 0 9.3 1.50 100 0 10.5 1.50 100 0 *: Solvent A is a mixture of 0.025% acetic acid and 0.0025% ammonium hydroxide solution, and solvent B is methanol.

Page 31 of 34

ip t cr

us

Table 2. Group assignment based on the chemical properties of the analytes (the logP and pKa values are predicted using Advanced Chemistry

Parameters

Group1 DFUR

5-FU

Group2 Cape

0.9 11.9

1.55 7.4

1.78 7.8

0.63 9.5

Ac c

LogP pKa

ep te

d

M

DFCR

an

Development software, ADME Suite, Version 5.0)

Page 32 of 34

ip t cr

DFUR

5-FU

9.27 ± 0.96 24.7 ± 1.3 203 ± 9 3791 ± 60 9.76 ± 1.03 25.0 ± 1.5 203 ± 7 4051 ± 92 10.2 ± 0.4 26.9 ± 1.0 214 ± 10 3775 ± 104 2.00 ± 0.19 5.06 ± 0.34 19.9 ± 0.6 161 ± 3

13.5 6.8 12.0 2.3 4.5 8.7 4.5 0.9 9.7 5.6 3.2 4.6 11.4 7.3 7.1 4.7

intra-day (n = 6 on 3 different days)

an

10.0 25.0 200 4000 10.0 25.0 200 4000 10.0 25.0 200 4000 2.00 5.00 20.0 160

inter-day (n = 6 in a single run)

9.9 4.8 2.0 1.4 11.1 5.6 3.1 2.4 2.7 3.2 4.8 2.4 9.4 6.7 1.6 1.3

M

Calculated

d

DFCR

Spiked

RSD (%)

ep te

Cape

Concentration (ng/mL)

Ac c

Analyte

us

Table 3. Precision and accuracy data for the analysis of Cape, DFCR, DFUR, and 5-FU in human plasma (3 days with six replicates per day). RE (%) inter-day (n = 6 in a single run)

intra-day (n = 6 on 3 different days)

13.2 4.3 4.0 4.9 1.1 2.0 2.5 1.1 3.9 8.2 7.9

7.3 1.1 1.6

7.1 3.8

5.6 0.2 1.2

2.0 1.1 0.1

5.2 2.4 0.2 1.3 1.3 2.2 7.7 7.0

0.7 0.3

Page 33 of 34

Table 4. Stability of Cape, DFCR, DFUR, and 5-FU in human plasma under various storage conditions (n = 3). Conditions

Analyte

Concentration (ng/mL) Nominal Calculated

Cape

25.0 4000 25.0 4000 25.0 4000 5.00 160

27.4 ± 0.9 3777 ± 22 24.5 ± 1.7 4182 ± 21 25.5 ± 0.4 3810 ± 38 5.17 ± 0.16 156 ± 4

25.0

26.5 ± 1.3

Bias (%)

Post-preparative 48 h in the autosampler tray at 4 oC

Cape

DFCR

5-FU

DFCR

Ac

DFUR

Long-term 20oC/34 days

5-FU

Cape DFCR DFUR 5-FU

cr 4.6 2.5 2.5 6.3 3.7 7.1 1.1

25.0 4000 25.0 4000 25.0 4000 5.00 160

22.2 ± 0.84 3667 ± 18 25.3 ± 1.6 4020 ± 136 25.2 ± 0.7 4061 ± 392 4.80 ± 0.23 158 ± 1

11.3 8.3 1.3 0.5 0.6 1.5 4.0 1.0

25.0

24.7 ± 0.80

0.1

4000 25.0 4000 25.0 4000 5.00 160

3605 ± 23 26.1 ± 0.63 3791 ± 110 25.9 ± 0.8 3826 ± 32 4.79 ± 0.19 144 ± 1

9.9 4.3 5.2 3.5 4.4 4.3 10.0

ce pt

Three freezer-thaw cycles Cape

6.0

3817 ± 16 25.6 ± 0.5 4101 ± 63 26.6 ± 1.3 3851 ± 93 5.35 ± 0.57 162 ± 4

4000 25.0 4000 25.0 4000 5.00 160

ed

DFUR

us

5-FU

an

DFUR

M

DFCR

9.4 5.6 2.2 4.5 2.2 4.7 3.4 2.3

ip t

Room temperature 6h

Page 34 of 34

Simultaneous determination of capecitabine and its three nucleoside metabolites in human plasma by high performance liquid chromatography-tandem mass spectrometry.

Capecitabine (Cape) is a prodrug that is metabolized into 5'-deoxy-5-fluorocytidine (DFCR), 5'-deoxy-5-fluorouridine (DFUR), and 5-fluorouracil (5-FU)...
518KB Sizes 3 Downloads 22 Views

Recommend Documents