Supercritical Fluid Chromatography and High-Performance Liquid

Apr 19, 2007 - Lucie Nováková , Vincent Desfontaine , Federico Ponzetto , Raul Nicoli , Martial Saugy , Jean-Luc .... Supercritical fluids and green...
0 downloads 0 Views 131KB Size
Download PDF
Anal. Chem. 2007, 79, 3856-3861

Supercritical Fluid Chromatography and High-Performance Liquid Chromatography/Tandem Mass Spectrometric Methods for the Determination of Cytarabine in Mouse Plasma Yunsheng Hsieh,* Fangbiao Li, and Christine J. G. Duncan

Drug Metabolism and Pharmacokinetics Department, Schering-Plough Research Institute, Kenilworth, New Jersey 07033

The separation of cytarabine (ara-C) from the endogenous compounds in mouse plasma by packed-column supercritical fluid chromatography (pSFC) was achieved on bare silica stationary phase with an isocratic mobile phase composed of CO2/methanol solvent with addition of ammonium acetate. SFC is commonly assumed to be only applicable to nonpolar and relatively low-polarity compounds. In this work, a broader range of compound polarities amenable to pSFC with appropriate mobilephase modifiers and additives under normal-phase retention mechanism was demonstrated. The pSFC was integrated with an atmospheric pressure chemical ionization source and a tandem mass spectrometer (MS/MS) to enhance the sensitivity, selectivity, and speed of the assay. The influence of mobile-phase components on chromatographic performance and ionization efficiency of the test compounds was investigated for improving the sensitivity and separation for the analyte and the internal standard. The pSFC-MS/MS approach requiring ∼2.5 min/sample for the determination of ara-C at nanograms per milliliter in mouse plasma was partially validated with respect to stability, linearity, and reproducibility. The mouse plasma levels of ara-C obtained by the pSFC-MS/MS method were found to be consistent with those determined by various reversed-phase, high-performance liquid chromatography methods using a porous graphite carbon column, a mixed-mode column, or a C18 column in conjunction with an ion-pairing agent coupled to a tandem mass spectrometer. Cytarabine (ara-C) is an antineoplastic antimetabolite used in the treatment of leukemia. Ara-C is metabolized intracellularly into its active triphosphate form to damage DNA by multiple mechanisms, including inhibition of DNA repair through an effect on β-DNA polymerase and incorporation into DNA.1 To determine the pharmacodynamic significance of DNA-incorporated ara-C, a reliable analytical method is needed to accurately measure ara-C concentrations in plasma samples. Various analytical procedures * To whom correspondence should be addressed. E-mail: yunsheng.hsieh@ spcorp.com. Phone: 908-7405385. Fax: 908-7402966. (1) Jin, J.; Jiang, D. Z.; Mai, W. Y.; Meng, H. T.; Qian, W. B.; Tong, H. Y.; Huang, J.; Mao, L. P.; Tong, Y.; Wang, L.; Chen, Z. M.; Xu, W. L. Leukemia 2006, 20, 1361-1367.

3856 Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

using the reversed-phase high-performance liquid chromatography (HPLC) system with spectrometric detection had been reported for the quantitative determination of ara-C in plasma.2-4 However, these mobile-phase conditions are not compatible with HPLCMS systems that have become the instrument of choice for drug assay in modern pharmaceutical industry due to its inherent selectivity and sensitivity. For the HPLC-MS/MS assay, sufficient chromatographic retention for the quantitative determination of the drug components in biological samples is highly recommended to avoid possible interferences from drug-related biotransformation products or ionization suppression due to coeluted endogenous materials.5,6 To deal with small polar compounds such as ara-C, HPLC columns containing polar-end capped and polar-enhanced stationary phases have been utilized to retain these molecules under highly aqueous conditions. However, no substantial retention of ara-C using these reversed-phase chromatography columns under aqueous mobile-phase conditions was achieved. Hydrophilic interaction chromatography (HILIC) has been demonstrated to be a powerful technique for the retention of polar analytes, offering a difference in selectivity compared to the traditional reversedphase chromatography.7-9 HILIC separates compounds by eluting them with a strong organic mobile phase against a neutral hydrophilic stationary phase, resulting in solutes that are retained in order of increasing hydrophobicity. However, HILIC normally requires complex sample pretreatment procedures prior to analysis that are not desirable for higher throughput assays. Ion-pairing chromatography has been widely used as an effective way to obtain satisfactory retention of polar analytes.10-12 The ion-pairing (2) Wermeling, J. R.; Pruemer, J. M.; Hassan, F. M.; Warner, A.; Pesce, A. J. Clin. Chem. 1989, 35, 1011-1015. (3) Breithaupt, H.; Schick, J. J. Chromatogr. 1981, 225, 99-106. (4) Fahmy, O. T.; Korany, M. A.; Maher, H. M. J. Pharm. Biomed. Anal. 2004, 34, 1099-1107. (5) Hsieh, Y.; Chintala, M.; Mei, H.; Agans, J.; Brisson, J. M.; Ng, K.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2001, 15, 2481-2487. (6) Mei, H.; Hsieh, Y.; Nardo, C.; Xu, X.; Wang, S.; Ng, K.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2003, 17, 97-103. (7) Naidong, W.; Shou, W. Z.; Chen, Y. L.; Jiang, X. J. Chromatogr., B 2001, 754, 387-399. (8) Naidong, W.; Shou, W. Z.; Addison, T.; Maleki, S.; Jiang, X. Rapid Commun. Mass Spectrom. 2002, 16, 1965-1975. (9) Hsieh, Y.; Chen, J. Rapid Commun. Mass Spectrom. 2005, 19, 3031-3036. (10) Quintana, J. B.; Rodil, R.; Reemtsma, T. Anal. Chem. 2006, 78, 1644-1650. (11) Lee, X. P.; Kumazawa, T.; Fujishiro, M.; Hasegawa, C.; Arinobu, T.; Seno, H.; Ishii, A.; Sato, K. J. Mass Spectrom. 2004, 39, 1147-1152. 10.1021/ac062441s CCC: $37.00

© 2007 American Chemical Society Published on Web 04/19/2007

reagents added into the mobile phase are used to improve chromatographic retention of polar analytes on the lipophilic stationary phase. Volatile ion-pairing reagents are generally recommended for HPLC-MS systems to avoid ionization suppression.12 Using a mixed-mode, reversed-phase/ion-exchange column for HPLC is another useful approach to handle small molecules with lower octanol/water partition coefficient values.13 With an embedded ion-pairing group in the reversed-phase stationary, the mixedmode column requires no ion-pairing reagent in the mobile phase to retain ionizable polar compounds. With a carbon load stationary phase as a basis for interaction with the analytes, the mixed-mode column also offers a typical reversed-phase retention profile for neutral compounds. This surface modification enables retention and separation through the reversed phase but also anionexchange chromatography principles. It was reported that the porous graphite carbon (PGC) column was able to provide more efficient retention than other kinds of reversed-phase columns designed to trap very polar compounds.14 The surface of PGC is composed of flat sheets of hexagonally arranged carbon atoms, which is stable throughout the entire pH range and chemically inert to aggressive solvent systems to enable separation of analytes with extensive polarities. PGC chromatography commonly employs water, methanol, and acetonitrile as the mobile phase for the elution of the polar compounds but requires much stronger organic solvents such as dichloromethane and tetrahydrofuran for the elution of nonpolar analytes.15 In this work, we investigate the potential of the packed-column supercritical fluid chromatography (pSFC)-MS/MS method as a complimentary coverage to the HPLC-MS/MS method for the quantitation of ara-C in mouse plasma samples. A bare silica packed column with CO2/methanol mobile phase was employed for separation of ara-C and the endogenous compounds from mouse plasma samples. The impact of several factors such as the compositions of mobile-phase modifier, additives, and flow rate on the chromatographic performances and ionization efficiencies of the test components was explored. Matrix ionization suppression for the pSFC-MS/MS system was investigated, which is one of the common concerns on developing any new hyphenated MS method. The analytical accuracy for ara-C concentrations in mouse study plasma samples obtained by the pSFC-MS/MS assay was evaluated by comparison with those obtained by the reversedphase HPLC-MS/MS methods using a mixed-mode column, a PGC column, and a C18 column in combination with an ion-pairing reagent.

Figure 1. Chemical structures of (I) ara-C and (II) clofazimine.

EXPERIMENTAL METHODS Reagents and Chemicals. Ara-C, the probe drug as the analyte, nicotinic acid, and clofazimine as the internal standard (ISTD) were purchased from Sigma (St. Louis, MO). The chemical structures of ara-C and clofazimine are given in Figure 1. Methanol and acetonitrile (HPLC grade) were purchased from Fisher Scientific (Pittsburgh, PA). Formic acid (99.999%) (FA), trifluo-

roacetic acid (TFA), and ammonium acetate were purchased from Aldrich Chemical Co., Inc. (Milwaukee, WI). The volatile perfluorinated carboxylic-acid ion-pairing reagent, nonafluoropentanoic acid (NFPA), was purchased from Sigma (St. Louis, MO). Carbon dioxide (SFC size) was obtained from Airgas (Cheshire, CT). Deionized water was generated from a Milli-Q water purifying system purchased from Millipore Corp. (Bedford, MA), and inhouse high-purity nitrogen (99.999%) was used. Drug-free mouse plasma samples (with heparin) were purchased from Bioreclamation Inc. (Hicksville, NY). pSFC-MS/MS System. The pSFC experiments were performed on a Berger Analytical SFC system (Newark, DE) equipped with a SFC pump (pump A), a modifier pump (pump B), and a column oven as described previously.16-18 Tandem mass spectrometric detection was performed using an Applied Biosystems/ MDS Sciex (Concord, ON, Canada) model API 4000 triple quadrupole mass spectrometer equipped with a heated nebulizer atmospheric pressure chemical ionization (APCI) probe. The pSFC-MS/MS system consisted of a Leap autosampler with a refrigerated sample compartment (set to 10 °C) from Leap Technologies (Carrboro, NC). A silica column (100 × 4.6 mm, 10 µm) purchased from Mettler-Toledo AutoChem Inc. (Columbia, MD) was used as the separation medium and was maintained at 50 °C. The effluent from the analytical column was split to the mass spectrometer and waste. Two separate pumps were used for neat carbon dioxide and methanol as the modifier, respectively. Outlet pressure was set to 100 bar to maintain the CO2/methanol mixture in one phase throughout the column. The amount of modifier that was added into the carbon dioxide mobile phase was expressed in volume. HPLC-MS/MS Systems. HPLC-MS/MS analysis was performed using an Applied Biosystems/MDS Sciex model API 4000 triple quadrupole mass spectrometer equipped with the APCI source. The chromatographic system consisted of a Leap autosampler with a refrigerated sample compartment (set to 10 °C) from Leap Technologies, Shimadzu on-line degasser, LC-10AD VP

(12) Hsieh, Y.; Duncan, C. Rapid Commun. Mass Spectrom. 2007, 21, 573-578. (13) Hsieh, Y.; Duncan, C.; Liu, M. J. Chromatogr., B, in press. (14) Thiebaut, D.; Vial, J.; Michel, M.; Hennion, M. C.; Greibrokk, T. J. Chromatogr., A 2006; 1122, 97. (15) Hsieh, Y.; Duncan, C.; Brisson, J. Rapid Commun. Mass Spectrom. 2007, 21, 629-634.

(16) Hsieh, Y.; Favreau, L.; Schwerdt, J.; Cheng, K. C. J. Pharm. Biomed. Anal. 2006, 40, 799-804. (17) Hsieh, Y.; Favreau, L.; Cheng, K. C.; Chen, J. Rapid Commun. Mass Spectrom. 2005, 19, 3037-3041. (18) Chen, J.; Hsieh, Y.; Cook, J.; Morrison, R.; Korfmacher, W. A. Anal. Chem. 2006, 78, 1212-1217.

Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

3857

pump, and LC-10A VP controller (Columbia, MD). For the IPHPLC method using NFPA as an ion-pairing reagent, a MetaSil Basic C18 column (4.6 × 100 mm, 3 µm) from MetaChem Technology was used as the analytical columns. For the mixedmode HPLC method, a Primesep A column (3.2 × 50 mm) from Sielc (Holland, MI) was used as the analytical column. For the GPC HPLC method, a Hypercarb column (50 × 2.0 mm, 5 µm) from Thermo Electron Corp. (Waltham, MA), was used as separation media. The schematic diagram of the postcolumn infusion system for the matrix effect studies on the pSFC-MS/MS system is similar to that used for the HPLC-MS/MS systems shown elsewhere.5 The probe analyte and the ISTD dissolved in modifier solvent were continuously infused into Peek tubing in between the analytical columns and a tandem mass spectrometer through a tee-piece using a Harvard Apparatus model 2400 (South Natick, MA) syringe pump. Effluent from the analytical columns mixed with the infused compounds and then entered the APCI interface. Either a supernatant extract of blank rat plasma or modifier solvent (5 µL) (as a reference signal) was injected into the analytical column for comparison of ionization responses. Sample Collection. The animal dosing experiments were carried out in accordance with the National Institutes of Health Guide to the Care and Use of Laboratory Animals and the Animal Welfare Act. Study blood samples were collected at specified time points following oral administration to individual mice. After clotting on ice, serum was isolated by centrifugation and stored frozen (-20 °C) until analysis. Standard and Sample Preparation. Stock solutions of araC, nicotinic acid, and clofazimine were prepared as 1 mg/mL solutions in methanol. Analytical standard samples were prepared by spiking known quantities of the standard solutions into blank mouse plasma. The concentration range for ara-C in mouse plasma was 0.05-10 µg/mL. The mouse plasma samples were prepared using the protein precipitation technique. A 300-µL aliquot of a methanol solution containing 1 ng/mL nicotinic acid and clofazimine was added to 10 µL of plasma located in a 96-well plate. After mixing and centrifugation, the supernatant was automatically transferred to a second 96-well plate by the Quadra 96 instrument. Five-microliter aliquots of the extract were injected by the Leap autosampler to all hyphenated MS systems for quantitative analysis. Chromatographic and Mass Spectrometric Conditions. For the pSFC method, ara-C was eluted under isocratic condition using carbon dioxide modified with 40% methanol containing 1% water and 11 mM ammonium acetate. Total flow through the pSFC system was 4 mL/min. For the IP-HPLC analysis of ara-C, mobile phases A and B were composed of water and acetonitrile containing 0.1% NFPA and 0.1% FA as mobile-phase additives, respectively. Gradient chromatographic separation using mobile phases A and B was as follows: 0.3 (2% B), 3.6 (30% B), 4.0 (100% B), 4.8 (100% B), and 4.9 min (2% B) and finished at 5.0 min to achieve satisfactory resolution between ara-C and the endogenous compound from the spiked standard and study mouse plasma samples. The retention times for ara-C and nicotinic acid as the ISTD were 3.77 and 2.45 min, respectively, as shown in Figure 2. The chromatographic conditions using a mixed-mode column and a PGC column were described elsewhere.12,13 The effluent from 3858 Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

Figure 2. Reconstructed IP HPLC-MS/MS chromatograms of ara-C from (A) blank, (B) the spiked standard mouse plasma at 50 ng/mL, and the reconstructed IP HPLC-MS/MS chromatogram of nicotinic acid as the ISTD from (C) the study mouse plasma sample.

the HPLC systems was connected directly to the mass spectrometer. The mass spectrometer was operated in the positive ion mode. The heated pneumatic nebulizer conditions were as follows: 500 °C temperature setting, 80 psi nebulizing gas pressure, 1.0 L min-1 auxiliary gas flow, and 0.9 L min-1 curtain gas flow rate. The MS/ MS reaction selected to monitor ara-C, nicotinic acid and clofazimine were the transitions from m/z 240, 124, and 473, the [M + H]+ ions, to the product ions at m/z 112, 80, and 431, respectively. The protonated molecules were fragmented by collision-activated dissociation with nitrogen as the collision gas at a pressure of instrument setting 5. The collision offset voltages were set at 20 V for ara-C and nicotinic acid and 50 V for clofazimine. Data were acquired and calculated using Analyst 1.4.1 software (Applied Biosystems). RESULTS AND DISCUSSION Development of a pSFC-MS/MS Method. The search for bioanalytical methods for higher throughput analysis of new chemical entities has fueled many innovations in HPLC-MS/MS methods in combination with fast chromatography,19-21 direct plasma injection,22,23 and other technologies.24 Alternatively, carbon dioxide, the most commonly used mobile phase in SFC possessing low viscosity and high diffusivity has the potential for fast pharmaceutical analysis. There have been a few literature reports on interfacing SFC with mass spectrometry.26-31 Here, the unique CO2/methanol binary mixtures make the hyphenation of SFC to the APCI source compatible. The lack of polarity of neat CO2 as a mobile phase makes it difficult for elution of polar compounds (19) Wang, G.; Hsieh, Y.; Cui, X.; Cheng, K. C.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2006, 20, 2215-2221. (20) Hsieh, Y.; Fukuda, E.; Wingate, J.; Korfmacher, W. A. Comb. Chem. High Throughput Screening 2006, 9, 3-8. (21) Hsieh, Y.; Merkle, K.; Wang, G. Rapid Commun. Mass Spectrom. 2003, 17, 1775-1780. (22) Hsieh, Y.; Bryant, M. S.; Brisson, J. M.; Ng, K.; Korfmacher, W. A. J. Chromatogr., B 2002, 767, 353-362. (23) Hsieh, Y.; Brisson, J. M.; Ng, K.; White, R. E.; Korfmacher, W. A. Analyst 2001, 126, 2139-2143. (24) Hsieh, Y.; Korfmacher, W. A. Curr. Drug Metab. 2006, 7, 479-489.

Figure 4. Influence of the flow rate of mobile phase on (A) the retention times and (B) the APCI responses of ara-C and clofazimine.

Figure 3. Effects of (A) percentage of methanol in the mobile phase, (B) percentage of water in methanol, and (C) percentage of ammonium acetate in methanol on the retention times of ara-C and clofazimine.

on packed column with high efficiency.25 Consequently, SFC with CO2 requires the combination of more polar solvents as modifier to enhance the solvent strength of the mobile phase. The SFC retention is mainly governed by the polarity of the mobile phase. The modifier, methanol, employed in this work is a major polar elution solvent. The higher modifier content would shorten the retention of the analytes. A systematic study concerning the elution of the analyte and the ISTD with various concentrations of modifier and additives in the mobile phase was performed on a bare silica column. As shown in Figure 3A, the retention times of ara-C and clofazimine decreased as methanol concentration increased. This is due to the increase of solvent strength of the mobile phase resulting in deactivation of active sites on the surface of the packing material. The addition of a small concentration of additives such as TFA, ammonium acetate, and water was reported to improve the salvation power of methanol-modified CO2 to further enhance (25) Ibanez, E.; Senorans, F. J. J. Biochem. Biophys. Methods 2000, 43, 25-43. (26) Xu, X.; Roman, J. M.; Veenstra, T. D.; Van Anda, J.; Ziegler, R. G.; Issaq, H. J. Anal. Chem. 2006, 78, 1553-1558. (27) Coe, R. A.; Rathe, J. O.; Lee, J. W. J. Pharm. Biomed. Anal. 2006, 42, 573580. (28) Zheng, J.; Pinkston, J. D.; Zoutendam, P. H.; Taylor, L. T. Anal. Chem. 2006, 78, 1535-1545. (29) Chen, J.; Hsieh, Y.; Cook, J.; Morrison, R.; Korfmacher, W. A. Anal. Chem. 2006, 78, 1212-1217. (30) Hsieh, Y.; Favreau, L.; Cheng, K.-C.; Chen, J. Rapid Commun. Mass Spectrom. 2005, 19, 3037-3041. (31) Ventura, M. C.; Farrell, W. P.; Aurigemma, C. M.; Greig, M. J. Anal. Chem. 1999, 71, 4223-4231.

chromatographic performance.32,33 The effect of water was expected to be attributed to the deactivation of the active sites of the stationary phase.25 However, no change in retention time for both ara-C and clofazimine was observed with an increase of water content in modifier from 0.01 to 1% as shown in Figure 3B. The roles of ammonium salts as SFC additives were suggested as (1) charge neutralization via ion-pairing formation between the solute and additive and (2) charge introduction to the stationary phase by exchange of silanol hydrogen for ammonium ion followed by anion exchange of the analyte.32 Figure 3C shows that the higher concentration of ammonium acetate in the mobile phase led to shorter retention time for clofazimine. However, the retention time of ara-C remains constant as the concentration of ammonium acetate increases from 2.5 to 10 mM. The retention in SFC is also dependent upon other factors such as temperature and the flow rate of the mobile phase. Figure 4 A shows that the retention times of both ara-C and clofazimine reduce as the flow rate of mobile phase increases from 2 to 6 mL/min. Here, the gas-like properties of supercritical CO2 allow for faster optimum flow rates and the use of longer columns than with liquids. The experimental conditions such as the composition and the flow rate of eluent might alter chromatographic performances but also affect the ionization efficiencies of the analytes when hyphenating to an ionization source prior to mass spectrometric detection. Figure 4B reveals that the sensitivity of clofazimine reduced by ∼40% as the flow rate increased from 3 to 6 mL/min. However, for ara-C, the APCI responses increase as the flow rate increased from 2 to 4 mL/min, and then the signals started to decline at a flow rate higher than 4 mL/min. For the APCI interface, the effluent was vaporized through the heated nebulizer, which consists of a concentric pneumatic nebulizer and a heated quartz tube. The heated mixture of mobile phase and vapor was (32) Zheng, J.; Glass, T.; Taylor, L. T.; Pinkston, J. D. J. Chromatogr. A 2005, 1090, 155-164. (33) Zheng, J.; Taylor, L. T.; Pinkston, J. D.; Mangels, M. L. J. Chromatogr. A 2005, 1082, 220-229.

Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

3859

Figure 5. Effects of (A) percentage of methanol in the mobile phase, (B) percentage of water in methanol, and (C) percentage of ammonium acetate in methanol on the APCI responses of ara-C and clofazimine.

then introduced into the APCI reaction chamber. The plasma of air, solvent, and sample components then encounter a cloud of electrons emitted from the tip of a corona electrode pin held at 3-5 kV in atmosphere. The electric field is sufficiently strong to ionize nitrogen, oxygen, and solvent vapor by removal of an electron through electron impact ionization. The analytical ions could be produced by the proton-transfer process with the reagent gas plasma if the acidity of the analytes was less than that of protonated charged clusters. An increase in solvent flow rates resulting in appreciable negative impact on the ionization efficiency of the test compounds might be due to less effective heat transfer at higher flow rates. Normally, the organic solvent has a positive influence on the APCI sensitivity of the analytes due to the high volatility of modifier, which efficiently assists nebulization process. Figure 5A shows that the signals of all test compounds increased as the ratio of methanol in binary mixtures increased. However, the ionization efficiency of ara-C appeared to be more sensitive than that of clofazimine to the concentrations of methanol in the mobile phase. Ara-C, a nucleoside analogue, is a very polar compound (log P ) -2.712) that requires high desolvation energies to form gas-phase ions. The increased signal intensity at higher concentrations of methanol is likely a consequence of faster evaporation of solvent from the droplets. Hydrated gas-phase cluster ions can be produced with a sufficient number of water molecules attached. Figure 5B shows that there was no substantial impact on the 3860

Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

ionization efficiency of clofazimine in the positive ion mode as water content in methanol increased from 0.01 to 1%. However, there was a positive effect on the signal intensity for ara-C with increasing water concentration in methanol. This implied that water played an important role in the APCI process for the polar analyte. The additive molecules are often in large excess relative to the analytes and might assist or compete with the analytes for ionization. As shown in Figure 5C, there was no substantial effect on the ionization efficiencies of both ara-C and clofazimine on increasing the concentration of ammonium acetate in methanol. This indicated that the addition of ammonium acetate had no effect on the ionization efficiencies of the test compounds. Matrix Ionization Suppression Studies. The continuous postcolumn infusion experiments are the standard method for monitoring the ionization suppression for any new hyphenated approaches developed in this laboratory.5 The affected area of the chromatographic run was evaluated according to the differences in the infusion chromatograms between the mobile-phase injection and the mouse plasma extract injection. Any changes in the signals of the infused test compounds were assumed to be due to matrix ionization suppression caused by the sample-related materials eluting from the analytical column. The objectives of the postcolumn infusion experiments were to measure the extent of ionization suppression and to define the “safer” portion of the chromatographic window. The retention times of ara-C and the ISTD were observed to appear in the safe chromatographic window for the HPLC-MS/MS and pSFC-MS/MS systems (data not shown). In addition, the peak responses of ara-C and the ISTD spiked into the methanol solution and the supernatant after protein precipitation were found to be within 10% error. All of these findings suggested that there is no substantial matrix effect for both ara-C and the ISTD using these hyphenated-MS methods. Mouse Plasma Assays. The representative IP HPLC-MS/ MS chromatograms of ara-C from the blank and the spiked standard following a protein precipitation technique are shown in Figure 2A and B. The ara-C was eluted by gradient separation with increasing the organic content in the mobile phase. As shown in Figure 2A and B, a baseline separation for the endogenous compound (t ) 3.64 min) and ara-C (t ) 3.77 min) was achieved to avoid the mass spectrometric interference. One of the major goals in this work was to apply the pSFC-MS/MS method for the determination of ara-C in mouse plasma to demonstrate its suitability as an analytical method. The assay procedure involved a one-step protein precipitation procedure. The extracted pSFCAPCI-MS/MS chromatograms of ara-C from the blank, spiked standard, and study mouse plasma samples under a normal-phase chromatographic mode are shown in Figure 6C. As indicated in the IP HPLC-MS/MS method, there was a need for chromatographic separation of ara-C from the endogenous substances sharing the same mass range. As shown in Figure 6C, a baseline separation using pSFC for ara-C (t ) 1.31 min) and an endogenous substance observed in the study mouse plasma samples instead of the blank and standard mouse plasma samples was achieved in order to avoid the mass spectrometric interference. The calibration curves from the same transition ranges using the pSFC-APCI-MS/MS method were constructed by plotting the analyte/ISTD peak-to-area ratios against ara-C concentrations in mouse plasma. The calibration curve for ara-C using the APCI

Figure 7. Correlation between the mouse plasma concentrations of ara-C obtained by pSFC-MS/MS method and IP HPLC-MS/MS method.

those obtained by the pSFC-MS/MS method (data not shown). Student’s t-test results indicated that there were no significant differences of the mouse plasma concentrations of ara-C determined by HPLC-MS/MS methods using a PGC column, a mixedmode column, and a C18 column in conjunction with an ion-pairing reagent or the pSFC-MS/MS approach with 95% confidence (R ) 0.5). These results concluded that the pSFC-MS/MS method was equivalent with the other HPLC-MS/MS methods in terms of accuracy.

Figure 6. Reconstructed pSFC-MS/MS chromatograms of ara-C from (A) blank, (B) the spiked standard mouse plasma at 50 ng/mL, (C) the study mouse plasma sample, and the reconstructed pSFCMS/MS chromatogram of clofazimine as the ISTD from (D) the study mouse plasma sample.

interface from standard mouse plasma samples at each concentration level was linear with a correlation coefficient, r2, greater than 0.994 (graph is not shown). Accuracy (% bias) was less than 15% at all concentrations, from 50 to 10 000 ng/mL. The standard and study mouse plasma samples were independently analyzed for ara-C using the pSFC-APCI-MS/MS method. The retention times and peak shape for both ara-C and the ISTD in the spiked standard and study plasma samples were found to be reproducible during the course of the study. The interday precision and accuracy in the measurement of the spiked standard samples in replicates of three obtained for ara-C were within 15% relative standard deviation and the nominal values. The spiked standard mouse plasma samples were employed for all stability tests. Three freeze-thaw cycles before processing, benchtop stability for 6 h under room temperature, and autosampler stability for 24 h at 10 °C were determined. Ara-C was found to be stable under these experimental conditions. Figure 7 compares the mouse plasma levels of ara-C obtained by the IP HPLC-MS/MS method and the pSFC-MS/MS method. The mouse plasma levels of ara-C obtained by the other HPLC-MS/MS methods using either a mixed-mode column or a PGC column were also compared with

CONCLUSIONS A rapid and sensitive pSFC-APCI-MS/MS method was developed for monitoring ara-C in mouse plasma samples. The use of pSFC coupled to tandem mass spectrometry has facilitated the analysis of both polar and nonpolar compounds in biological fluids. The coupling of CO2/methanol normal-phase pSFC to the APCI interface with no derivatization prior to the analysis showed good potential for maximum separation efficiency and the detection of the analyte and the ISTD. The eluent compositions and flow rates were found to have substantial effects on the ionization efficiency and chromatographic performance of the test compounds. The ionization suppression in the proposed pSFC-APCI-MS/MS system was found to be negligible. Three reversed-phase HPLCAPCI-MS/MS assays using a mixed-mode column, a PGC column, or a C18 column in conjunction with an ion-pairing reagent were applied for the determination of ara-C in the study mouse plasma samples. These HPLC-MS/MS methods and the described pSFC-MS/MS method using a simple sample pretreatment procedure showed equivalent accuracy to the analytical results and have been proven to be reliable in support of an in vivo study. ACKNOWLEDGMENT The authors thank our animal dosing group for doing the in vivo studies presented in this work.

Received for review December 26, 2006. Accepted March 23, 2007. AC062441S

Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

3861