MS method for the quantitation of disulfiram, its role in stabilized plasma and its application

Journal of Chromatography B, 937 (2013) 54–59

Contents lists available at ScienceDirect

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

A novel UPLC–ESI-MS/MS method for the quantitation of disulfiram, its role in stabilized plasma and its application Ling Zhang a , Ying Jiang a , Guanghui Jing a , Yilin Tang b , Xi Chen a , Dan Yang a , Yu Zhang a,∗ , Xing Tang a,∗ a b

Department of Pharmaceutics, College of Pharmacy, Shenyang Pharmaceutical University, Shenyang, Liaoning, PR China School of Materials and Engineering, Xi’an Jiaotong University, Xi’an, PR China

a r t i c l e

i n f o

Article history: Received 5 June 2013 Accepted 5 August 2013 Available online 20 August 2013 Keywords: Disulfiram Ultra performance liquid chromatography Electrospray ionization Mass spectrometry Pharmacokinetics

a b s t r a c t Disulfiram (DSF) has been used to treat alcoholism for many years and it has been suggested to play a key role in combatting many kinds of tumors. However, disulfiram has complex pharmacokinetics and is rapidly eliminated which limits its use as a tumor treatment. Therefore, a rapid and sensitive analytical method based on ultra performance liquid chromatography coupled to electrospray ionization-tandem mass spectrometry (UPLC–ESI-MS/MS) was developed and validated for the determination of disulfiram in rat plasma. Blood samples were pre-stabilized with a stabilizing agent and then plasma was obtained and subjected to solid phase extraction (SPE), and chromatographed on a Phenomenex Kinetex® XB C18 column with gradient elution using a mobile phase consisting of acetonitrile–water (containing 0.1% formic acid and 1 mM ammonium acetate) at a flow rate of 0.2 mL/min for 3 min. Multiple reactions monitoring in positive mode was carried out with disulfiram at 296.95/115.94 and diphenhydramine (internal standard, IS) at 256.14/167.02 over a linear range from 0.6 to 1200 ng/mL. The extraction recovery of disulfiram for different concentrations ranged from 75.7% to 78.3%. The intra- and inter-day precision was less than 8.93% and 12.39%, respectively, and the accuracy was within ±7.75%. The validated method was successfully applied to a pharmacokinetic study of disulfiram in rat plasma after oral administration of a dose of 180 mg/kg. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Disulfiram (DSF), a low molecular weight compound with the structure shown in Fig. 1, has been proven to be a safe and effective treatment of alcoholism for more than 60 years with only minor harmful side effects [1–4]. Moreover, an increasing number of investigations regarding the outcome of disulfiram on different kinds of tumor cells in vivo and in vitro suggest that disulfiram may act as an inhibitor of tumor growth and metastasis [5–7], including breast cancer [8–12], glioblastoma [13,14], melanoma [15–17], myeloid leukemia [18], and prostate cancer [19]. Previous studies have shown that disulfiram is unstable in gastric fluid and blood, and is rapidly converted to its monomer diethyldithiocarbamic acid (DDC) [4,21]. In addition, it is difficult to isolate, detect and quantify disulfiram in blood due to its dithiocarbamate structure, which is electrophilic and rapidly forms mixed disulfides with endogenous thiols [3,20,21]. However a few methods, including high performance liquid chromatography with UV

detection and gas chromatography (GC), have been developed to determine disulfiram in biological samples [3,4,21,22]. However, these methods have a number of shortcomings: poor sensitivity and reproducibility and time-consuming to carry out making it difficult to carry out efficient pharmacokinetic studies. Therefore, there is a need for a sensitive, reliable and quick method to measure plasma levels of disulfiram. Ultra performance liquid chromatography combines small particles, low system volume (dead volume) and rapid detection technology to enhance speed, sensitivity and separation while retaining the original principle and practicality of HPLC. Accordingly, a rapid and sensitive analytical method based on UPLC coupled to electrospray ionization-tandem mass spectrometry (UPLC–ESI-MS/MS) was developed and validated for the measurement of disulfiram in rat plasma.

2. Materials and methods 2.1. Chemicals and reagents

∗ Corresponding authors. Tel.: +86 24 23986343; fax: +86 24 23911736. E-mail addresses: [email protected] (Y. Zhang), [email protected] (X. Tang). 1570-0232/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jchromb.2013.08.009

Disulfiram was synthesized by the Department of Organic Synthesis of Shenyang Pharmaceutical University. Diphenhydramine (internal standard, IS) was supplied by the Department of Analytical

L. Zhang et al. / J. Chromatogr. B 937 (2013) 54–59

55

2.5. Preparation of calibration standards and quality control (QC) samples

Fig. 1. The structure of disulfiram (DSF).

Chemistry of Shenyang Pharmaceutical University. Diethylenetriaminepentaacetic acid (DTPA) was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Ammonium acetate, formic acid and acetonitrile (HPLC) were supplied by Dikma Company, Inc. (Richmond Hill, NY, USA). All other reagents were of chromatographic grade.

A stock solution of DSF was prepared by dissolving an accurately weighed quantity of DSF in methanol to obtain a high concentration (600 ␮g/mL). This stock solution was further diluted to prepare working standard solutions of the desired concentrations and quality control solutions at three different concentrations. The IS solution (100 ng/mL) was prepared by diluting the stock solution of diphenhydramine with methanol. All solutions were stored at −20 ◦ C but were allowed to warm up to room temperature before use. Calibration standards were prepared daily by spiking 200 ␮L of the stabilized blank plasma with 20 ␮L of appropriate working standard solutions and 20 ␮L IS solution to obtain final concentrations of 0.6, 1.2, 6, 12, 60, 120, 600 and 1200 ng/mL. QC samples were prepared in stabilized blank plasma at low, middle and high concentrations of 1.2, 60, 960 ng/mL. The standards and QCs were extracted in each analysis run using the same procedures for the samples as described below.

2.2. Instrumentation 2.6. Sample preparation The experiment was conducted using a Waters® XEVOTM TQ MS ACQUITY UPLC® System connected to a triple-quadrupole tandem mass spectrometer with electrospray ionization (ESI). Automated solid phase extraction (SPE) was performed using a Waters® OASIS HLB cartridge 1 cc/10 mg with an Auto Science AP-01P Vacuum Pump (Science Scientific, Inc.) equipped with twelve extraction modules operating in parallel. Water was filtered and purified using a Barnstead EASYpure IIRF/UV ultrapure water system (Dubuque, IA, USA).

Sample preparation involved SPE using a Waters Oasis 30 mg HLB cartridge. The cartridge was conditioned with 1 mL methanol followed by equilibration with 1 mL 0.05 M EDTA. Then, a 200 ␮L aliquot of stabilized plasma containing DSF with 20 ␮L IS solution was loaded onto a Waters Oasis cartridge and the cartridge was washed with 1000 ␮L water (containing 5% acetonitrile) followed by elution with 1000 ␮L acetonitrile. Then, 1000 ␮L eluent was collected in a clean glass tube and transferred to an autosampler vial, and 5 ␮L was injected via a partial loop with needle overfill into the UPLC–MS/MS system for analysis.

2.3. UPLC–MS/MS conditions 2.7. Method procedures Chromatographic separations were carried out on a Phenomenex Kinetex® XB C18 column (50 mm × 21 mm, 2.6 ␮m; Phenomenex, Torrance, CA, USA) with a run time of 3 min. The temperature of the column and auotosampler was held constant at 40 ◦ C and 4 ◦ C, respectively. The mobile phase used for gradient elution consisted of (A) acetonitrile and (B) water (containing 0.1% formic acid and 1 mM ammonium acetate) at a flow rate of 0.2 mL/min. The gradient conditions of the mobile phase were as follows: (A) was increased linearly from 20% to 80% during the first 0.8 min and held at that value for 1.4 min, then returned to the original ratio in the same way and re-equilibrated for 0.5 min. The injection volume was 5 ␮L. The mass spectrometer was operated in positive ion mode. The mass spectrometric parameters used for DSF and IS were as follows: the capillary voltage was set at 3.5 kV, the temperatures of the source and desolvation were 150 ◦ C and 400 ◦ C, respectively, nitrogen was used as the desolvation gas (600 L/h) and cone gas (50 L/h), argon was used as the collision gas, DSF was monitored at an m/z of 296.95/115.94 and IS at an m/z of 256.14/167.02, and the cone voltage and collision energy was set at 14 V and 10 V for DSF, and 14 V and 12 V for IS.

2.4. Stabilized plasma Blood samples from rats were collected in tubes which had been treated with sodium heparin and stabilized with the same volume of cold stabilizing agent, then vortexed for 30 s and centrifuged at 6000 rpm at 4 ◦ C for 5 min to obtain stabilized plasma. The stabilizing agent was composed of 0.9% sodium chloride, 0.64% sodium acetic and 0.8% DTPA at pH 4.5 and was stored at 4 ◦ C.

The method was validated in terms of its selectivity, linearity, accuracy and precision, recovery, matrix effect and stability according to US FDA guidelines [23]. The selectivity of the method was evaluated by analyzing blank plasma samples, spiked plasma samples at the LLOQ level from six different sources and rat plasma collected 1.5 h after oral administration of a dose of 180 mg/kg. The peak area of the coeluting interferences should be <20% of the peak areas of the LLOQ standards. The calibration standards were prepared and assayed in duplicate on three consecutive days. The calibration curves were constructed by assaying standard plasma samples at eight concentration levels over the range 0.6 –1200 ng/mL with 1/C2 as the weighting factor, where C stands for the nominal concentration. The accuracy and precision were evaluated by the determination of QC samples at three concentrations in six replicates over three consecutive days. The accuracy was expressed as the relative error (RE). The intra- and inter-day precision was expressed by the relative standard deviation (RSD). The RE and RSD should be within 15% except for the LLOQ, where it should not exceed 20%. Extraction recovery and matrix effect were evaluated by comparing the mean area response of each set of solutions at three concentrations of QC samples in six replicates. Set 1 QCs were prepared by spiking DSF (at low, medium and high QC concentrations) in stabilized blank plasma prior to extraction. Set 2 QCs were prepared by spiking the same amount of DSF as Set 1 in a 1000 ␮L aliquot of extracted stabilized blank plasma. Set 3 QCs were prepared by spiking the same amount of DSF as Set 1 in a 1000 ␮L aliquot of acetonitrile. The extraction recovery is calculated as the ratio of the mean peak area of an anaylte spiked prior to extraction

56

L. Zhang et al. / J. Chromatogr. B 937 (2013) 54–59

voltage, collision energy, flow rate of the desolvation gas and cone gas were optimized to maximize the MS response. 3.2. Optimization of chromatography

Fig. 2. MS/MS spectrum of disulfiram.

(Set 1) to that from a post-extraction solution (Set 2). The matrix effect is expressed as the ratio of the mean peak area of an analyte spiked post-extraction (Set 2) to that from neat solution (Set 3). The stability of each analyte after sample preparation was investigated by analyzing replicates (n = 3) of plasma samples at low and high QC concentrations, which were exposed to an autosampler temperature (4 ◦ C) for 24 h. All analytes were considered to be stable when 85–115% of the nominal concentration was found. 2.8. Application to a pharmacokinetic study in rats The validated UPLC–ESI-MS/MS method was applied to investigate the circulation profile of DSF after a single oral administration of a dose of 180 mg/kg to rats. The experimental protocol was approved by the University Ethics Committee for the Use of Experimental Animals and all animal studies were carried out according to the Guide for Care and Use of Laboratory Animals. Male Sprague-Dawley (SD) rats (n = 6, body weight range 200–220 g) were provided by the Laboratory Animal Center of Shenyang Pharmaceutical University (Shenyang, China). The rats were maintained in controlled conditions (temperature 20–25 ◦ C, relative humidity 55–60% and 12 h dark-light cycle) with free access to standard laboratory food and water for a one week acclimatization period, followed by fasting but with access to water overnight prior to the experiment to be carried out the next day. Venous blood (about 0.5 mL) was collected from the ophthalmic venous plexus 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 6, 8 and 12 h after dosing. The preparation of stabilized plasma and samples was as described above. 3. Results and discussion 3.1. Optimization of mass spectrometry In this assay, the mass spectrometry operation parameters were carefully optimized for the determination of DSF. A standard solution of DSF was directly introduced along with the mobile phase into the mass spectrometer with ESI as the ionization source. Also, the mass spectrometer was tuned in both positive and negative ionization modes to obtain an optimum response for DSF. It was found that the signal intensity of the positive ion was higher than that of the negative ion. In the precursor ion full-scan spectrum, the most abundant ion was the protonated molecule [M+H]+ m/z 296.95. In the product ion spectrum, a major fragment ion at m/z 147.91 was formed by rupture of the disulfide bond in the [M+H]+ ion. Further loss of the sulfur moiety from m/z 147.91 produced a fragment ion at m/z 115.94 (Fig. 2). The ion transition of m/z 296.95 → m/z 115.94 provided a better signal-to-noise (S/N) ratio compared with the ion transition of m/z 296.95 → m/z 147.91 and, accordingly, this was selected for the quantitation of DSF. Parameters such as the desolvation temperature, ESI source temperature, capillary and cone

The chromatographic conditions were modified to obtain high sensitivity, short retention times and symmetrical peak shapes. The separation and ionization of DSF and IS were affected by the composition of the mobile phase. Methanol and acetonitrile were both investigated as organic modifiers of the mobile phase. It was found that the peaks were more symmetric and the S/N ratio of DSF was clearly higher with acetonitrile than with methanol. Adding formic acid to the mobile phase increased the ionization but produced a leading peak for DSF while ammonium acetate produced a symmetric peak but with a lower ionization. As a result, 0.1% formic acid and 1 mM ammonium acetate in the mobile phase were used in combination to achieve a compromise. Gradient elution was used to improve the retention of the analyte and the separation from endogenous interference. The flow rate of mobile phase was determined based on the ionization efficiency of ESI source and optimal control range of flow rate of UPLC. High flow rate can reduce the ionization efficiency. Therefore, the flow rate was set at 0.2 mL/min. Under these optimal conditions, the run time of only 3 min for each sample fulfilled the requirement for a high sample throughput for analysis of samples in pharmacokinetic studies of DSF. 3.3. Selection of internal standard Deuterated standard would be a preferred IS in LC–MS. However, considering poor stability of disulfiram, we estimated that the stability of deuterated disulfiram would be poor likewise and the experiment would become more complex and difficult to carry out. We did not use deuterated disulfiram as IS, although it would have been the best and most logical choice. Diphenhydramine was selected as internal standard, because: (1) its retention time was short and reasonable; (2) its peak did not interfere with that of DSF (see Fig. 3); (3) diphenhydramine was stable and subjected to constant extraction recovery in the entire experiment (75%); (4) diphenhydramine had no interaction with DSF and did not affect the determination. 3.4. Optimization of plasma and sample preparation DSF is a hydrophobic compound, extremely unstable in blood and under oxidizing/reducing conditions. It has been suggested that its instability is related to its disulfide structure which may result in the formation of mixed disulfides with protein sulfhydryl groups [20]. Blood was processed via two steps: pre-stabilization with stabilizing agent and extraction by SPE. A stabilizing agent composed of 0.9% sodium chloride, 0.64% sodium acetic and 0.8% DTPA at pH 4.5 was suggested by Johansson [3]. We compared the effect of stabilizing agent on inhibiting DSF degradation in our experiment and found that the samples without stabilizer showed much lower plasma level of DSF than those of samples stabilized. Most of samples were below the limit of detection. The possible interaction mechanisms were inferred. DSF is a chelator of heavy metals such as cupric and zinc. Metal ions may act as an intermediary in the process that DSF forms chelates with the thiols of proteins. DTPA is a strong metal chelating agent, competitively inhibiting the interactions between DSF and metallic ions or thiol-containing molecules. We will further study the specific interaction mechanism later in our experiment. Acidic conditions were chosen because the disulfide exhibited greater acid-stability showed than parent thiols [3,20,21]. Approaches involving liquid–liquid extraction of DSF [21,22] were

L. Zhang et al. / J. Chromatogr. B 937 (2013) 54–59

57

Table 1 Precision and accuracy data for the analysis of disulfiram in plasma (on three validation days, six replicates per day). Concentration (ng/mL)

RSD (%)

Added

Found

Inter-day

Intra-day

RE (%)

0.60 1.20 60.0 960.0

0.59 1.21 62.6 992.2

8.82 8.89 8.93 4.66

8.73 9.02 12.39 5.62

−7.75 2.17 5.79 4.75

effects of the acid–base properties and volume of acetonitrile on the extraction recovery of DSF and IS were investigated and 1000 ␮L of neutral acetonitrile was found to be optimum. 3.5. Method validation 3.5.1. Selectivity Fig. 3 shows typical chromatograms of DSF and IS obtained from the analysis of blank plasma (Fig. 3A), a blank plasma sample spiked with DSF at the LLOQ and IS (Fig. 3B), and a plasma sample obtained 1.5 h after a single oral administration of 180 mg/kg DSF to a rat (Fig. 3C). The interferences caused by endogenous substances or from other sources were avoided by baseline chromatographic separation. The retention time of DSF and IS was 2.00 and 1.20 min, respectively. 3.5.2. Linearity and LLOQ Linear regression of the peak area ratios versus concentration was carried out over the concentration range 0.6–1200 ng/mL for DSF in plasma. A typical regression equation for the curve was y = 1.95C + 0.0141 (r = 0.9947), where y is the peak area ratio of DSF to IS, and C is the nominal concentration of DSF in plasma. The LLOQ for DSF was 0.6 ng/mL with a precision less than 20%. 3.5.3. Accuracy and precision The data for the intra- and inter-day accuracy and precision of the method are given in Table 1. The intra- and inter-day RSDs were not more than 8.93% and 12.39%, and the REs ranged from −7.75% to 5.79% at three QC levels, indicating acceptable accuracy and precision. Fig. 3. Example chromatograms of disulfiram (I) and IS (peak II). (A) Matrix blank sample; (B) LLOQ sample at a concentration of 0.6 ng/mL; and (C) plasma collected 1.5 h after oral administration of a dose of 180 mg/kg.

explored using chloroform, but these required a time-consuming sample preparation and failed to be reproduced in our prior experiments. In previous literature reports [3,4], on-line precolumn purification was chosen as the method of sample preparation. Although the purification and isolation of samples were accomplished simultaneously, a more complicated pump system was required. In addition, the extraction recovery and precision of online purification for a large batch of plasma samples would likely be reduced. This is because on-line precolumn purification systems fail to rapidly separate DSF from plasma as for many samples awaiting purification. The degradation of DSF caused by long-term contact with plasma leads to unreliable results. Thus, for superior extraction efficiency and consistently fast extraction of DSF from samples, off-line SPE processing sharing a similar extraction with on-line precolumn purification was explored to extract DSF from plasma. Twelve samples could be cleaned at one time using automated solid phase extraction equipment. The cartridge was flushed with 0.05 M EDTA to chelate free metal ions. Acetonitrile was selected as the elution solvent because it offered a higher recovery and better peak shape than methanol. The

3.5.4. Extraction recovery and matrix effect The extraction recovery of DSF from plasma was 75.7%, 77.4%, and 78.3% at concentrations of 0.6, 60, 960 ng/mL, respectively, as shown in Table 2. The mean extraction recovery of IS was 75%. Thus, the consistency in the recoveries of DSF and IS support the suitability of the extraction procedure for its application to routine sample analysis. No significant matrix effect for DSF and IS was observed, indicating that no co-eluting substances influenced the ionization of the analyte and IS. 3.5.5. Stability DSF is unstable, whose lability has been studied in some articles. DSF was found labile and rapidly reduce to its monomer within 4 min [21]. Meanwhile, even though stabilizing agent was added to the plasma, the detection and recovery of DSF decreased after cryostorage [3,25]. In our previous experiment, similar results were Table 2 Extraction recovery and matrix effect of disulfiram in plasma. Concentration (ng/mL)

Extraction recovery (%)

Matrix effect (%)

1.20 60.0 960.0

75.7 77.4 78.3

102.3 99.8 90.2

58

L. Zhang et al. / J. Chromatogr. B 937 (2013) 54–59

Table 3 Stability of disulfiram in an autosampler for 24 h at 4 ◦ C (n = 3). Concentration (ng/mL) Added 1.20 960.0

RSD (%)

RE (%)

4.35 10.6

8.75 −2.17

Found 1.31 939.2

found. Thus, we processed and analyzed immediately throughout the entire experiment. The total number of samples was 100 at maximum including 78 for each experiment group (six subjects and 11–13 blood points) and 22 for standard curves and QCs, which multiplied by the 3-min analysis time for each sample, the total waiting time required being about 6 h. For these reasons, the stability of DSF in autosampler for 24 h could meet the needs of our experiment and was tested. The results are listed in Table 3. DSF was stable in the elution solvent for 24 h at 4 ◦ C in the autosampler after SPE treatment. 3.5.6. Pharmacokinetic application The validated analytical method was successfully applied to a pharmacokinetic study involving administration of a single oral dose of 180 mg/kg DSF to rats. The profile of the mean plasma concentration–time curve of DSF is shown in Fig. 4. There was a marked individual variability in DSF plasma levels. The reason for this variability is uncertain, but might involve its high lipid solubility (log Kow = 3.88) [24], rapid reduction to thiols in vivo and variations in formation of mixed disulfides with endogenous thiols in rats. Johansson [3] reported that DSF was so labile that no detectable DSF was detected until the second week of repeated dosing of DSF. The plasma concentration–time profile exhibited multiple peaks. This might be due to the high lipid solubility, and high protein binding (96.1%) of DSF [25] as well as enterohepatic cycling [26]. Moreover, DDC, the plasma concentrations of which were measured and found to be a hundred times higher than DSF in prior testing, returned to DSF by reoxidation [27] which may also play a part in this phenomenon. To improve the instability and prolong the retention time of parent DSF in vivo, the injection of DSF lipid microspheres (DSF-LM) has been under investigation. Lipid microspheres, also known as lipid emulsions, are colloidal particles formed by an internal oil phase and an external water phase [28]. Using intravenous injection of lipid microspheres might prevent the irritation caused by conventional co-solvent systems or drugs [29–31], protein binding and the

Fig. 4. Mean plasma concentration (C)–time (t) curve of disulfiram after a single oral administration of disulfiram at a dose of 180 mg/kg to rats (n = 6).

hydrolytic degradation of drugs might be blocked to some extent due to loading the drug in the oil phase and/or oil–water interfacial film [28,32,33]. On one hand, DSF-LM injection may bypass the strong first-pass effect and the hydrolysis of DSF in gastric fluid if given by the intravenous route while, on the other hand, enzymolysis may be reduced by incorporating DSF in the internal oil phase to reduce its contact with endogenous thiols in blood and, thus, produce higher plasma concentrations in vivo. This idea was partly supported by our preliminary experiments. The results showed that the average plasma levels and Cmax produced by DSF-LM injection were dramatically higher than those of the DSF oral preparation, while the individual variability was greatly reduced. This validated analytical method could be also applied to a pharmacokinetic study of DSF following intravenous administration. The pharmacokinetics of DSF-LM will be published later. 4. Conclusions A UPLC–ESI-MS/MS method has been developed and validated for measuring DSF in rat plasma and plasma samples were processed by SPE. The validated method performed well in terms of selectivity, linearity, LLOQ, accuracy and precision and it was successfully applied to a pharmacokinetic study of DSF in rat plasma after oral administration of a dose of 180 mg/kg. Acknowledgements This work is supported by the National Science & Technology Major Project ‘Key New Drug Creation and Manufacturing Program’ (2013ZX09301305), National High-tech R&D Program of China 863 Program (2012AA020305) and the Doctoral Start-up Fund Program of Liaoning Province, China (20111141). References [1] J. Cobby, M. Mayersohn, S. Selliah, J. Pharmacokinet. Biopharm. 6 (1978) 369–387. [2] J.J. Helen, M. Pettinati, K.M. Kampman, C.P. O’Brien, J. Clin. Psychopharmacol. 25 (2006) 290–302. [3] B. Johansson, Clin. Chim. Acta 177 (1988) 55–64. [4] B. Johansson, J. Chromatogr. B: Biomed. Sci. Appl. 378 (1986) 419–429. [5] T.P. Kennedy, Method of treating cancer using tetraethyl thiuram disulfide, United States Patent No. 6,589,987, B2. [6] S. Verma, D.J. Stewart, J.A. Maroun, R.C. Nair, Am. J. Clin. Oncol. 13 (1990) 119–124. [7] F. Chu, C.A. O’Brain, Antioxid. Redox. Signal. 7 (2005) 855–862. [8] D. Chen, Q.C. Cui, H. Yang, Q.P. Dou, Cancer Res. 66 (2006) 10425–10433. [9] H. Zhang, D. Chen, J. Ringler, W. Chen, Q.C. Cui, S.P. Ethier, Q.P. Dou, G. Wu, Cancer Res. 70 (2010) 3996–4004. [10] X. Guo, B. Xu, S. Pandey, E. Goessl, J. Brown, A.L. Armesilla, J.L. Darling, W. Wang, Cancer Lett. 290 (2010) 104–113. [11] F. Wang, S. Zhai, X. Liu, L. Li, S. Wu, Q.P. Dou, B. Yan, Cancer Lett. 300 (2011) 87–95. [12] N.C. Yip, I.S. Fombon, P. Liu, S. Brown, V. Kannappan, A.L. Armesilla, B. Xu, J.L. Darling, W. Wang, Br. J. Cancer 104 (2011) 1564–1574. [13] J. Triscott, C. Lee, K. Hu, A. Fotovati, R. Berns, M. Pambid, M. Luk, R.E. Kast, E. Kong, E. Toyota, S. Yip, B. Toyata, S.E. Dunn, Oncotarget 3 (2012) 1112–1123. [14] P. Hothi, T.J. Martins, L. Chen, L. Deleyrolle, J.-G. Yoon, B. Reynolds, G. Foltz, Oncotarget 3 (2012) 1124–1136. [15] D. Cen, D. Brayton, B. Shahandeh, F.L. Meyskens, P.J. Farmer, J. Med. Chem. 47 (2004) 6914–6920. [16] S.S. Brar, C. Grigg, K.S. Wilson, W.D. Holder, A.J. Ghio, A.R. Whorton, G.W. Stowell, L.B. Whittall, R.R. Whittle, D.P. White, T.P. Kennedy, Mol. Cancer Ther. 3 (2004) 1049–1060. [17] B.W. Morrison, N.A. Doudican, K.R. Patel, S.J. Orlow, Melanoma Res. 20 (2010) 11–20. [18] J. Navratilova, P. Jungova, P. Vanhara, J. Preisler, V. Kanicky, J. Smarda, Int. J. Mol. Med. 24 (2009) 661–670. [19] K. Iljin, K. Ketola, P. Vainio, P. Halonen, P. Kohonen, V. Fey, R.C. Grafström, M. Perälä, O. Kallioniemi, Clin. Cancer Res. 15 (2009) 6070–6078. [20] J.H. Strömme, Biochem. Pharmacol. 14 (1965) 381–391. [21] J. Cobby, M. Mayersohn, S. Selliah, J. Pharmacol. Exp. Ther. 202 (1977) 724–731. [22] J.C. Jensen, M.D. Faiman, J. Chromatogr. B: Biomed. Sci. Appl. 181 (1980) 407–416.

L. Zhang et al. / J. Chromatogr. B 937 (2013) 54–59 [23] USPDA, 2001, http://www.fda.gov/downloads/Drugs/GuidanceCompliance RegulatoryInformation/Guidances/ucm070107.pdf (accessed 21.03.12). [24] C. Hansch, A. Leo, D. Hoekman, Exploring QSAR – Hydrophobic, Electronic, and Steric Constants, American Chemical Society, Washington, DC, 1995, pp. 80. [25] B. Johansson, J. Pharm. Pharmacol. 42 (1990) 806–807. [26] M.D. Faiman, C. Jensen, R.B. Lacourslere, Clin. Pharmacol. Ther. 36 (1984) 520–526. [27] J.H. Strömme, Biochem. Pharmacol. 14 (1965) 393–410. [28] A.G. Floyd, Pharm. Sci. Technol. Today 2 (1999) 134–143.

59

[29] W. Lixin, H. Haibing, T. Xing, S. Ruiying, C. Dawei, Int. J. Pharm. 323 (2006) 161–167. [30] S. Shi, H. Chen, X. Lin, X. Tang, Int. J. Pharm. 383 (2009) 264–270. [31] M. Su, M. Zhao, Y. Luo, X. Lin, L. Xu, H. He, H. Xu, X. Tang, Int. J. Pharm. 411 (2011) 188–196. [32] Y. Weng, F. Li, W. Chen, Y. Miao, X. Tang, Eur. J. Lipid Sci. Technol. 112 (2010) 1190–1199. [33] S. Shi, H. Chen, Y. Cui, X. Tang, Int. J. Pharm. 373 (2009) 147–155.