Determination of warfarin alcohols by ultra-high performance liquid chromatography–tandem mass spectrometry: Application to in vitro enzyme kinetic studies

Journal of Chromatography B, 944 (2014) 63–68

Contents lists available at ScienceDirect

Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb

Short Communication

Determination of warfarin alcohols by ultra-high performance liquid chromatography–tandem mass spectrometry: Application to in vitro enzyme kinetic studies Osama Y. Alshogran a,b , Andrew J. Ocque a,c , Jielu Zhao b , Billy W. Day b , Franc¸ois A. Leblond d , Vincent Pichette d,e , Thomas D. Nolin a,c,∗ a

Center for Clinical Pharmaceutical Sciences, School of Pharmacy, University of Pittsburgh, Pittsburgh, PA, United States Department of Pharmaceutical Sciences, School of Pharmacy, University of Pittsburgh, Pittsburgh, PA, United States c Department of Pharmacy and Therapeutics, School of Pharmacy, University of Pittsburgh, Pittsburgh, PA, United States d Service de Néphrologie et Centre de Recherche, Hôpital Maisonneuve-Rosemont, Montréal, Québec, Canada e Département de Pharmacologie, Université de Montréal, Montréal, Québec, Canada b

a r t i c l e

i n f o

Article history: Received 23 August 2013 Accepted 6 November 2013 Available online 14 November 2013 Keywords: Warfarin alcohols Warfarin Reduction LC–MS Enzyme kinetics

a b s t r a c t A sensitive, accurate, and reproducible ultra-high performance liquid chromatography–tandem mass spectrometry method was developed and validated for determination of warfarin and its alcohol metabolites (RS/SR- and RR/SS-warfarin alcohol) in 10 mM Tris–HCl incubation buffer (pH 7.4). Sample preparation involved acidification with 4% formic acid, followed by liquid–liquid extraction using methyl tert-butyl ether. Chromatographic separation was achieved using a Hypersil Gold C18 (2.1 mm × 100 mm, 1.9 ␮m) analytical column with gradient elution of solvent A (water containing 0.01% formic acid) and solvent B (acetonitrile containing 0.1% formic acid). The flow rate was 0.4 mL/min and the total run time was 5 min. Detection of analytes was performed using heated electrospray ionization (negative mode) and selected reaction monitoring. Excellent linearity was observed for all analytes over the standard curve concentration ranges of 100–10,000 ng/mL for warfarin, and 0.5–250 ng/mL for warfarin alcohols. The intra- and inter-day accuracy and precision for analytes were within ±10.0%. Excellent recovery and negligible matrix effects were observed. The method is robust, sensitive, accurate and reproducible, and was successfully applied to in vitro enzyme kinetic studies of warfarin. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Warfarin is the most commonly used oral anticoagulant for the treatment and prevention of thromboembolic disorders [1]. It is a technically difficult medication to use in clinical practice, due to its notoriously narrow therapeutic window, large interindividual variability, potential for food and drug interactions, and significant risk if suboptimally dosed [2]. Warfarin is administered clinically as a racemic mixture of both R- and S-enantiomers, and is highly metabolized, exhibiting regioselective and stereoselective metabolism [3,4]. It undergoes Phase I oxidation mediated mainly by cytochrome P450 (CYP) enzymes, CYP2C9 in particular, producing hydroxy metabolites that can be further metabolized by Phase II conjugation [3,4]. The effect of various factors (e.g., disease,

∗ Corresponding author at: Department of Pharmacy and Therapeutics, School of Pharmacy, University of Pittsburgh, 808 Salk Hall, 3501 Terrace Street, Pittsburgh, PA 15261, United States. Tel.: +1 412 624 2400; fax: +1 412 624 1850. E-mail address: [email protected] (T.D. Nolin). 1570-0232/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jchromb.2013.11.014

inflammation, genetic polymorphisms) on CYP2C9 function [5] and their subsequent impact on warfarin dose requirements has been extensively studied [6,7], and this information is now used to optimize warfarin therapy [8,9]. Another important metabolic pathway for warfarin elimination is reduction. Warfarin undergoes reduction by hepatic reductases, which generate warfarin alcohols of two diastereoisomers, namely RS/SR-warfarin alcohol (alcohol 1), and RR/SS-warfarin alcohol (alcohol 2) [10,11]. Warfarin reduction is catalyzed predominantly in the cytosol, producing warfarin alcohol 1 as the major metabolite [12,13]. Although reduction accounts for up to 20% of warfarin metabolism [11] and produces pharmacologically active alcohol metabolites [14], the effect of altered reductase function on the disposition of warfarin and its alcohol metabolites has not been studied to date. In order to do so, a robust and validated analytical method for the quantitative determination of each analyte is required. Several assays have been developed for measuring the enantiomers of hydroxywarfarin metabolites in biological samples including HPLC with UV [15], combination of UV/fluorescence and

64

O.Y. Alshogran et al. / J. Chromatogr. B 944 (2014) 63–68

circular dichroism [16] or MS detection [17–19], capillary zone electrophoresis with UV detection [20], and micellar electrokinetic chromatography with MS detection [21]. However, few analytical methods have been reported for the determination of warfarin alcohol metabolites. These include conventional HPLC methods with UV or fluorescence detection [22–25], and gas chromatography with mass spectrometric (GC–MS) detection [26]. In general, HPLC with UV or fluorescence detection methods are not as sensitive and selective as MS techniques and require larger sample volume to achieve high sensitivity. The GC–MS method is complex, and requires extensive sample preparation including a derivatization step [26]. In addition, each of the reported methods include minimal or no validation parameters for warfarin alcohols. To date, use of contemporary LC–MS techniques for the measurement of warfarin alcohols has not been reported. Thus, the goal of this work was to develop and validate a simple, rapid and robust ultra-high performance liquid chromatography–tandem mass spectrometry (UHPLC–MS/MS) assay for determination of warfarin and its alcohol metabolites. The method was used to examine the enzyme kinetics of warfarin in cytosolic fractions of rat liver tissue. 2. Experimental 2.1. Chemicals and reagents Warfarin (C19 H16 O4 ), d5-7-hydroxywarfarin (C19 H11 D5 O5 , used as internal standard for warfarin alcohols), formic acid, NADPH, magnesium chloride, and tris(hydroxymethyl)aminomethane (Trizma base) were purchased from Sigma (St. Louis, MO, USA). Warfarin alcohols (C19 H18 O4 ) were synthesized as previously described [10,27]. Racemic d5-warfarin (C19 H11 D5 O4 , used as internal standard for warfarin) was purchased from Toronto Research Chemical (North York, Ontario, Canada). Hydrochloric acid, methyl tert-butyl ether (MTBE), methanol, OPTIMA LC–MS grade water, and acetonitrile were purchased from Fisher Scientific (Pittsburgh, PA, USA). All chemicals were LC–MS grade or of the highest purity available. Nitrogen gas (ultra-pure, >99.9%) was produced by a Parker Balston nitrogen generator (Haverhill, MA, USA). Argon gas (ultra-pure, >99.9%) was provided by Valley (Wheeling, WV, USA).

0.01 m/z, scan time was set to 0.5 s, and full width at half maximum was set to 0.7 m/z for quadrupole one (Q1) and three (Q3). The ion transitions were m/z 307.1 → 161.1 for warfarin (collision energy = 17 V), m/z 312.2 → 161.2 for d5-warfarin (collision energy = 22 V), m/z 309.1 → 250.3 for warfarin alcohols (collision energy = 26 V), and m/z 328.1 → 177.1 for d5-7-hydroxywarfarin (collision energy = 22 V). Signal output was captured and processed with Xcaliber software v2.2 (Thermo Scientific, San Jose, CA, USA). 2.3. Preparation of calibration standard and quality control (QC) samples Each analyte (warfarin and warfarin alcohols) was dissolved in methanol to obtain a 1.0 mg/mL stock solution. These stock solutions were mixed with methanol to prepare intermediate stock solutions that were spiked into blank 10 mM Tris–HCl incubation buffer (pH 7.4) to create calibration standards at concentrations of 100, 200, 1000, 2000, 5000, and 10,000 ng/mL for warfarin, and 0.5, 5.0, 25, 50, 125, and 250 ng/mL for warfarin alcohols. Three quality control samples (LQC, MQC, and HQC) containing both analytes were made by spiking a separate 1.0 mg/mL stock solution into blank Tris–HCl solution at concentrations of 300, 4000, and 8000 ng/mL for warfarin, and 1.5, 100, and 200 ng/mL for warfarin alcohols. Preliminary experiments showed that recoveries of analytes were not affected by the presence of microsomal or cytosolic protein and cofactors. Therefore, standards and QCs were prepared in the incubation buffer alone. All stock solutions, standards and QC samples were stored at −80 ◦ C to simulate the storage conditions of study samples. 2.4. Sample preparation An internal standard working solution containing 5.0 ␮g/mL d5-warfarin and 1.2 ␮g/mL d5-7-hydroxywarfarin was prepared in methanol. A 100 ␮L aliquot of sample was spiked with 20 ␮L internal standard solution, and then acidified by adding 250 ␮L 4% formic acid in water. Sample extraction was carried out by adding 2 mL MTBE and placing on a shaker for 20 min at room temperature. The upper organic layer was removed and evaporated to dryness under a stream of nitrogen at 37 ◦ C. The dried residue was reconstituted with 150 ␮L of acetonitrile–water (25:75, v/v), and a 20 ␮L aliquot was injected onto the LC–MS system.

2.2. Equipment and UHPLC–MS/MS conditions 2.5. Assay validation Liquid chromatography was performed with an Accela series UHPLC system (Thermo Scientific, San Jose, CA, USA) including an autosampler and ultra-high performance binary pump. Chromatographic separation of the samples was achieved with a Thermo Hypersil Gold C18 (2.1 mm × 100 mm, 1.9 ␮m) analytical column, connected to 0.22 ␮m frit filter. The flow rate was 400 ␮L/min and composed of solvent A (water containing 0.01% formic acid), and solvent B (acetonitrile containing 0.1% formic acid). The gradient consisted of 60% solvent A and 40% B for 0–2.0 min, followed by 35% A and 65% B from 2.0 to 3.0 min. The column was then reequilibrated to initial conditions for 2 min. The total run time was 5 min. The autosampler was maintained at 10 ◦ C and the column temperature was held at 40 ◦ C. MS/MS detection was performed on a TSQ Quantum Ultra triple quadrupole mass spectrometer (Thermo Scientific, San Jose, CA, USA) equipped with a heated electrospray ionization source (HESI2). Analytes were detected in negative ionization mode using selected reaction monitoring. The spray voltage was set to 3000 V and the vaporizer temperature was set to 360 ◦ C. The sheath gas and auxiliary gas were set to 65 and 55 (arbitrary units), respectively. The ion transfer tube was set to 350 ◦ C. Collision gas (argon) pressure was set at 1.5 mTorr. Scan width was set to

2.5.1. Calibration and linearity Calibration curves were constructed using six concentrations of warfarin and warfarin alcohols. Each calibration level was run in duplicate for three days, except for the lower limit of quantification (LLOQ), which was run in triplicate. For each curve, the absolute peak-area ratios of the analyte to the internal standard were calculated and plotted against the nominal analyte concentration. Calibration curves were generated by equal weighting quadratic log-log regression analysis. 2.5.2. Accuracy and precision The accuracy and precision were assessed by analysis of six replicate QC samples for two days, followed by analysis of twelve replicate QC samples on the third day, for a total of n = 24 samples at each QC level. Intra-day accuracy and precision were determined from the twelve replicates on day 3, and inter-day accuracy and precision were calculated from all 24 QC samples. The calculated mean concentration relative to the nominal concentration was used to express accuracy (% bias). The relative standard deviation (% RSD) was calculated from the QC values and used to estimate the precision.

O.Y. Alshogran et al. / J. Chromatogr. B 944 (2014) 63–68

65

Table 1 Intra-and inter-day accuracy (% bias) and precision (% RSD) for LLOQ and quality controls. Analyte

Level

Nominal conc. (ng/mL)

Warfarin

LLOQ LQC MQC HQC

100 300 4000 8000

Warfarin alcohol 1

LLOQ LQC MQC HQC

0.5 1.5 100 200

Warfarin alcohol 2

LLOQ LQC MQC HQC

0.5 1.5 100 200

a b

Intra-daya

Inter-dayb

% bias

% RSD

−2.4 −10 −4.4 0.79

4.7 6.9 5.6 5.8

1.1 −4.4 5.2 2.2 0.27 5.7 8.8 7.2

% bias

% RSD

−5.4 −8.9 −4.7 −1.1

7.8 5.7 6.3 6.5

10.7 7.0 6.7 5.0

1.2 −2.9 5.8 0.53

6.5 8.2 6.3 5.1

11 5.7 4.8 4.0

0.4 3.9 7.7 3.5

8.8 6.6 5.0 5.7

Three replicates for LLOQ; 12 replicates for LQC, MQC, HQC. Nine replicates for LLOQ; 24 replicates for LQC, MQC, HQC.

Table 2 Extraction recovery, process efficiency and matrix effect of analytes in incubation buffer.a Analyte

Nominal conc. (ng/mL)

Extraction recovery (%, mean)

Process efficiency (%, mean)

Matrix effect (%, mean)

Warfarin

300 4000 8000

79.2 88.4 90.2

78.3 92.0 93.6

98.8 104 104

Warfarin alcohol 1

1.5 100 200

73.2 79.8 81.0

76.3 81.8 82.4

104 102 102

Warfarin alcohol 2

1.5 100 200

77.5 80.5 80.7

92.0 79.3 78.5

111 98.5 97.3

a

n = 4 for each QC level.

A dilution analysis was performed on samples spiked to twice the concentration of the highest standard, then diluted 1:4, 1:20, or 1:100 with blank Tris–HCl buffer before analysis. Each dilution level was processed in triplicate.

2.5.3. Recovery, process efficiency and matrix effect Extraction recovery was determined by comparing the response (peak areas) of the analytes spiked in Tris–HCl buffer before and after the extraction. Process efficiency was evaluated by comparing the response of analytes spiked in the buffer before extraction with that of neat solutions spiked in mobile phase, which was defined as 100% recovery. Matrix effect was assessed by comparing the response of the analytes spiked in the buffer after extraction with that of neat solutions spiked in mobile phase, representing 100% (no matrix effect). All experiments were conducted at three QC levels in four replicates.

2.5.4. Stability Stability was evaluated at low and high QC levels in triplicate. Bench-top stability was tested by analyzing samples that were left out on the bench top at room temperature for 4 h before analysis. Autosampler stability was determined for processed samples stored at 10 ◦ C, which were re-injected and analyzed against a new standard curve up to 24 h post-processing. Freeze–thaw stability was assessed for samples that were subjected to three freeze–thaw cycles at 24 h intervals from −80 ◦ C to room temperature prior to analysis. The results of all tested samples were compared with the recovery from samples that were freshly prepared (100% control). 2.6. Application of the method The analytical method was applied to the quantification of alcohol metabolites generated after in vitro warfarin reduction. Briefly, liver cytosol was isolated from rat liver tissue by a standard differential centrifugation procedure [28]. Cytosolic protein (0.5 mg/mL)

Table 3 Stability for quality controls (LQC and HQC) of warfarin and its alcohol metabolites.a Analyte

Target (ng/mL)

Stability Bench-top (4 h)

Warfarin

300 8000

Autosampler (24 h)

Three freeze/thaw cycles (24 h)

% of target

% RSD

% of target

% RSD

% of target

% RSD

106 111

1.8 0.45

103 95.1

8.6 2.3

99.6 111

2.5 0.85

Warfarin alcohol 1

1.5 200

112 103

3.5 3.6

103 101

5.9 0.82

111 95.9

7.4 1.1

Warfarin alcohol 2

1.5 200

102 104

3.0 2.7

111 101

7.9 0.86

106 101

9.2 2.5

a

Data are presented as means of n = 3 for LQC and HQC levels.

66

O.Y. Alshogran et al. / J. Chromatogr. B 944 (2014) 63–68

Fig. 1. Representative chromatograms of incubation buffer at LLOQ (left panel, A–D), and a cytosolic incubation sample at the lowest point of the Michaelis–Menten curve (right panel, E–H). (A) Warfarin: 100 ng/mL; (B and F) internal standard d5-warfarin; (C) warfarin alcohols: 0.5 ng/mL; (D and H) internal standard d5-7-hydroxywarfarin; (E) warfarin: 7941 ng/mL; (G) warfarin alcohol 2 and alcohol 1: 0.85, 7.97 ng/mL, respectively.

was incubated with warfarin (20–2000 ␮M) and 5 mM MgCl2 in 10 mM Tris–HCl buffer (pH 7.4). Reactions were started by the addition of 1 mM NADPH and conducted for 30 min at 37 ◦ C. Incubation conditions were optimized with respect to time and protein concentration. Following incubation, warfarin alcohols were measured as described above. A Michaelis–Menten model was used to fit the formation of alcohol metabolites, and total reductase enzyme activity was estimated (GraphPad Software Inc., San Diego, CA).

3. Results and discussion In order to support in vitro studies exploring warfarin reduction, we aimed to develop and validate a simple and robust UHPLC–MS/MS method for determination of warfarin and warfarin alcohol metabolites. To our knowledge, this is the first LC–MS/MS method developed for this purpose. The method was validated according to the U.S. Food and Drug Administration (FDA) guidelines for bioanalytical method validation [29]. The method is

O.Y. Alshogran et al. / J. Chromatogr. B 944 (2014) 63–68

67

ranged from 4.0% to 8.2% (Table 1). In all cases, bias and RSD values were within ±10% for all analytes. Acceptable accuracy and precision were also observed during the dilution analysis, as bias and RSD were within ±14.4%, and 4.8% in all samples, respectively. 3.2.3. Recovery, process efficiency and matrix effect The extraction recoveries and process efficiency of analytes ranged from 73.2% to 90.2%, and 76.3% to 93.6%, respectively. Overall, the recoveries and process efficiency of analytes were consistent and reproducible. The matrix effect was negligible, with measured concentrations deviating by −2.7% to 11% from neat samples (Table 2). 3.2.4. Stability Bench-top, autosampler, and freeze–thaw stability were assessed for analytes in the buffer at LQC and HQC levels. Samples were found to be stable for 4 h at room temperature (bench-top), and 24 h post analysis at 10 ◦ C in the refrigerated autosampler. Additionally, three freeze–thaw cycles had no effect on the stability of analytes in the buffer. The mean measured concentrations of analytes in the buffer ranged from 95% to 112% of freshly analyzed samples, indicating adequate stability under all conditions tested (Table 3). 3.3. Assay application

Fig. 2. Michaelis–Menten plots of warfarin alcohol 1 (A) and alcohol 2 (B), after incubation of a 0.5 mg/mL rat liver cytosolic protein with 1 mM NADPH, 5 mM MgCl2 and various concentrations of warfarin (20–2000 ␮M) for 30 min at 37 ◦ C. The y-axis represents the formation rate of the metabolites.

simple, accurate and precise, and is currently being used to support warfarin enzyme kinetic studies.

We are using the current method in the in vitro assessment of warfarin reduction. Michaelis–Menten plots of warfarin alcohols generated after incubation of rat cytosol with various concentrations of warfarin are shown in Fig. 2. The metabolic activities (Vmax ) for cytosolic reductases producing alcohol 1 and alcohol 2 were 33.2 and 9.9 pmol/mg protein/min, respectively. The affinity constant values (Km ) were 434.8 and 1583 ␮M for alcohol 1 and alcohol 2, respectively. The method is well suited for enzyme kinetic studies of warfarin requiring high-throughput quantitative determination of warfarin alcohol metabolites. 4. Conclusions

3.1. Chromatographic separation Representative chromatograms of analytes at the LLOQ and in a cytosolic reaction mixture are shown in Fig. 1. The analytes were identified based on retention time and mass spectra of individual analytical standards. The retention times were approximately 2.23, 2.61, and 2.86 min for warfarin alcohol 2, warfarin alcohol 1 and warfarin, respectively. The peaks of interest were well separated and free from evidence of ion suppression or enhancement. 3.2. Assay validation 3.2.1. Calibration and linearity Calibration curves were obtained over concentration ranges of 100–10,000 ng/mL for warfarin, and 0.5–250 ng/mL for warfarin alcohols, with a correlation coefficient (r2 ) greater than 0.996 for all curves. The LLOQ for each calibration curve demonstrated acceptable accuracy and precision (RSD and bias were within ±11% and ±5%, respectively, Table 1), and signal-to-noise was greater than 10:1. The intra- and inter-day accuracy and precision were within ±12% for all analytes calibration standards. 3.2.2. Accuracy and precision The intra-day and inter-day accuracy (% bias) and precision (% RSD) were determined at warfarin QC concentrations of 300, 4000, and 8000 ng/mL, and alcohols concentrations of 1.5, 100, and 200 ng/mL. Assay bias ranged from −10.0% to 8.8%, while the RSD

A simple, robust, rapid, accurate, and precise UHPLC–MS/MS method for the determination of warfarin and warfarin alcohols has been developed and validated. The short run-time and high sensitivity and specificity associated with LC–MS/MS renders this method valuable for studies relating to warfarin alcohols. References [1] J. Hirsh, J. Dalen, D.R. Anderson, L. Poller, H. Bussey, J. Ansell, D. Deykin, Chest 119 (2001) 8S. [2] M.J. Kim, S.M. Huang, U.A. Meyer, A. Rahman, L.J. Lesko, J. Clin. Pharmacol. 49 (2009) 138. [3] L.S. Kaminsky, Z.Y. Zhang, Pharmacol. Ther. 73 (1997) 67. [4] D.R. Jones, J.H. Moran, G.P. Miller, Drug Metab. Rev. 42 (2010) 55. [5] U.M. Zanger, M. Schwab, Pharmacol. Ther. 138 (2013) 103. [6] A.L. Jorgensen, S. Al-Zubiedi, J.E. Zhang, A. Keniry, A. Hanson, D.A. Hughes, D. Eker, L. Stevens, K. Hawkins, C.H. Toh, F. Kamali, A.K. Daly, D. Fitzmaurice, A. Coffey, P.R. Williamson, B.K. Park, P. Deloukas, M. Pirmohamed, Pharmacogenet. Genomics 19 (2009) 800. [7] L.H. Cavallari, T.Y. Langaee, K.M. Momary, N.L. Shapiro, E.A. Nutescu, W.A. Coty, M.A. Viana, S.R. Patel, J.A. Johnson, Clin. Pharmacol. Ther. 87 (2010) 459. [8] T.E. Klein, R.B. Altman, N. Eriksson, B.F. Gage, S.E. Kimmel, M.T. Lee, N.A. Limdi, D. Page, D.M. Roden, M.J. Wagner, M.D. Caldwell, J.A. Johnson, N. Engl. J. Med. 360 (2009) 753. [9] P. Lenzini, M. Wadelius, S. Kimmel, J.L. Anderson, A.L. Jorgensen, M. Pirmohamed, M.D. Caldwell, N. Limdi, J.K. Burmester, M.B. Dowd, P. Angchaisuksiri, A.R. Bass, J. Chen, N. Eriksson, A. Rane, J.D. Lindh, J.F. Carlquist, B.D. Horne, G. Grice, P.E. Milligan, C. Eby, J. Shin, H. Kim, D. Kurnik, C.M. Stein, G. McMillin, R.C. Pendleton, R.L. Berg, P. Deloukas, B.F. Gage, Clin. Pharmacol. Ther. 87 (2010) 572. [10] K.K. Chan, R.J. Lewis, W.F. Trager, J. Med. Chem. 15 (1972) 1265.

68

O.Y. Alshogran et al. / J. Chromatogr. B 944 (2014) 63–68

[11] R.J. Lewis, W.F. Trager, K.K. Chan, A. Breckenridge, M. Orme, M. Roland, W. Schary, J. Clin. Invest. 53 (1974) 1607. [12] T.A. Moreland, D.S. Hewick, Biochem. Pharmacol. 24 (1975) 1953. [13] J.J. Hermans, H.H. Thijssen, Biochem. Pharmacol. 38 (1989) 3365. [14] R.J. Lewis, W.F. Trager, A.J. Robinson, K.K. Chan, J. Lab. Clin. Med. 81 (1973) 925. [15] M. Miura, S. Okuyama, S. Kato, H. Kagaya, A. Murata, A. Komatsuda, H. Wakui, K. Sawada, Ther. Drug Monit. 33 (2011) 108. [16] H. Takahashi, T. Kashima, S. Kimura, N. Muramoto, H. Nakahata, S. Kubo, Y. Shimoyama, M. Kajiwara, H. Echizen, J. Chromatogr. B: Biomed. Sci. Appl. 701 (1997) 71. [17] D.R. Jones, G. Boysen, G.P. Miller, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 879 (2011) 1056. [18] T. Uno, T. Niioka, M. Hayakari, K. Sugawara, T. Tateishi, Ther. Drug Monit. 29 (2007) 333. [19] Z. Zuo, S.K. Wo, C.M. Lo, L. Zhou, G. Cheng, J.H. You, J. Pharm. Biomed. Anal. 52 (2010) 305.

[20] Q. Zhou, W.P. Yau, E. Chan, Electrophoresis 24 (2003) 2617. [21] X. Wang, J. Hou, M. Jann, Y.Y. Hon, S.A. Shamsi, J. Chromatogr. A 1271 (2013) 207. [22] M.J. Fasco, L.J. Piper, L.S. Kaminsky, J. Chromatogr. 131 (1977) 365. [23] M.J. Fasco, M.J. Cashin, L.S. Kaminsky, J. Liq. Chromatogr. 2 (1979) 565. [24] C. Banfield, M. Rowland, J. Pharm. Sci. 73 (1984) 1392. [25] Y.W. Wong, P.J. Davis, J. Chromatogr. 469 (1989) 281. [26] A.M. Duffield, P.H. Duffield, D.J. Birkett, M. Kennedy, D.N. Wade, Biomed. Mass Spectrom. 6 (1979) 208. [27] W.F. Trager, R.J. Lewis, W.A. Garland, J. Med. Chem. 13 (1970) 1196. [28] T.A. van der Hoeven, M.J. Coon, J. Biol. Chem. 249 (1974) 6302. [29] U.S. Food and Drug Administration, Guidance for Industry – Bioanalytical Method Validation, 2001, Available from: http://www.fda.gov/downloads/ Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM070107. pdf (accessed 23.08.13).