Fused-Core Silica Column High-Performance Liquid Chromatography

Drug Metabolism and Pharmacokinetics Department, Schering-Plough Research Institute, Kenilworth, New Jersey 07033. A high-performance liquid ...
0 downloads 0 Views 269KB Size
Anal. Chem. 2007, 79, 5668-5673

Fused-Core Silica Column High-Performance Liquid Chromatography/Tandem Mass Spectrometric Determination of Rimonabant in Mouse Plasma Yunsheng Hsieh,* Christine J. G. Duncan, and Jean-Marc Brisson

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

A high-performance liquid chromatography (HPLC) method using a fused-core silica particle packing was evaluated to allow fast and efficient separation for the analysis of pharmaceutical compounds. Fused-core particles are produced by “fusing” a porous silica layer onto a solid silica particle. The efficiencies of columns packed with 2.7 µm “fused-core” particles (a 0.5 µm porous shell fused to a solid 1.7 µm silica core particle) and 1.7 µm porous particles were compared in reversed-phase HPLC using rimonabant as an analyte. The fused-core silica materials providing the shorter diffusional mass transfer path for solutes are less affected in resolving power by increases in mobile-phase velocity than the sub-2 µm porous silica packings resulting in faster separations and higher sample throughput. This fast HPLC technology is comparable with ultrahigh-pressure liquid chromatography (UHPLC) in terms of chromatographic performance but demands neither expensive ultra-high-pressure instrumentation nor new laboratory protocols. The column effluent was directly connected to the atmospheric pressure chemical ionization (APCI) source prior to tandem mass spectrometric detection. In this work, the described fast HPLC-MS/MS and UHPLC-MS/MS approaches requiring approximately 1.5 min per sample were applied and compared for the determination of the rimonabant in mouse plasma samples at the low nanograms per milliliter region in support of a pharmacodynamic study. HPLC coupled with a tandem mass spectrometer (MS/MS), which provides excellent sensitivity and selectivity for monitoring new chemical entities, has become standard equipment in support of various in vitro and in vivo experiments.1-3 The large numbers of samples derived from various drug discovery experiments have yielded continuing challenges for developing high-speed chromatographic techniques to enable higher throughput sample quantification during the lead optimization processes.4 In most cases, HPLC-MS/MS methods demonstrate no chromatographic * To whom correspondence should be addressed. E-mail: yunsheng.hsieh@ spcorp.com. Phone: 908-740-5385. Fax: 908-740-2966. (1) Korfmacher, W. A. Using Mass Spectrometry for Drug Metabolism Studies; Korfmacher, W. A., Ed.; CRC Press: 2005; Chapter 1. (2) Korfmacher, W. A. Drug Discovery Today 2005, 10, 1357-1367. (3) Hsieh, Y.; Korfmacher, W. A. Curr. Drug Metab. 2006, 7, 479-489. (4) Hsieh, Y. The Encyclopedia of Mass Spectrometry; Niessen, W. M. A., Ed.; Elsevier Ltd.: 2006; Vol. 8, chapter 8.

5668 Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

interference from the endogenous components in complex biological fluids while maintaining appropriate separation power to avoid ion suppression liability and potential mass spectrometric interferences from the biotransformation products.5-7 The van Deemter equation describing the relationship between linear velocity (µ) and plate height (H) is one common way to evaluate the column efficiency of a new stationary phase as follows:

H ) A + B/µ + Cµ ) 2λdp + 2GDm/µ + ωdp2µ/Dm + Rdf2µ/Ds

Here, A, B, and C terms are the van Deemter coefficients representing eddy-diffusion, longitudinal diffusion, and resistance to mass transfer, respectively, which contribute to band broadening. G, ω, and R are constants. λ, dp, Dm, df, and Ds represent particle shape, particle diameter, the diffusion coefficient of the mobile phase, the film thickness, and the diffusion coefficient of the stationary phase, respectively. According to the van Deemter equation, reducing particle size of the stationary phase is one of the effective ways to directly enhance the column efficiency. For example, reducing the particle diameter from 5 to 1.7 µm will result in a 3-fold increase in column efficiency but 8.6-fold increase in the column backpressure. One solution to overcome the high backpressure is to shorten the lengths of columns packed with small particles (∼3 µm) to provide a rapid separation.8-10 This so-called microcolumn technique offers chromatographic resolution similar to standard regular analytical columns (50 mm × 4.6 mm) methods and includes benefits such as the reduction of both run cycle times and solvent consumption. Monolithic silica columns made from a single piece of porous silica gel can be operated at higher flow rates than conventional LC columns due to their lower back pressure.11,12 The low (5) Mei, H.; Hsieh, Y.; Nardo, C.; Xu, X.; Wang, S.; Ng, K.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2003, 17, 97-103. (6) Matuszewski, B. K.; Constanzer, M. L.; Chavez-Eng, C. M. Anal. Chem. 1998, 70, 882-889. (7) Hsieh, Y.; Chintala, M.; Mei, H.; Agans, J.; Brisson, J. M.; Ng, K.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2001, 15, 2481-2487. (8) Hsieh, Y.; Brisson, J. M.; Wang, G.; Ng, K.; Korfmacher, W. A. J. Pharm. Biomed. Anal. 2003, 33, 251-261. (9) Miao, X. S.; Metcalfe, C. D. J. Mass Spectrom. 2003, 38, 27-34. (10) Hsieh, Y.; Fukuda, E.; Wingate, J.; Korfmacher, W. A. Comb. Chem. High Throughput Screening 2006, 9, 3-8. (11) Hsieh, Y.; Wang, G.; Wang, Y.; Chackalamannil, S.; Brisson, J.; Ng, K.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2002, 16, 944-950. 10.1021/ac070343g CCC: $37.00

© 2007 American Chemical Society Published on Web 07/03/2007

backpressure despite higher mobile-phase flow rates is due to their higher permeability of monolithic silica. This makes higher speed separation possible without a noticeable effect on chromatographic resolution. Ultra-high-pressure liquid chromatography (UHPLC) is another technical advance that allows the liquid handling system to handle the high backpressure resulting from the stationary phase with sub-2 µm particles. UHPLC offers theoretical advantages in chromatographic resolution, speed, and sensitivity over conventional HPLC systems.13,14 When operating at higher than optimal mobile-phase linear velocity in order to achieve fast chromatographic separations, the C term of the van Deemter equation becomes the major contributing factor for decreased column efficiency. This is because the higher the velocity of the mobile phase leads to the greater axial dispersion and the broader peak. The superficially porous silica column employing fused-core particle technology provides the shorter path of diffusion for a solute into and out of the stationary phase to minimize peak broadening at a higher mobile-phase flow rate. In this work, chromatographic performances such as pressure drop, separation power, and ruggedness with six pharmaceuticals were evaluated and compared for fused-core and porous C18 particles under practical conditions. The column effluent was directly connected to the atmospheric pressure chemical ionization (APCI) source as part of the MS/MS system. The impact of mobile-phase flow rates on both the chromatographic performances and the APCI ionization efficiencies of the analyte was explored. Simultaneous selective reaction monitoring (SRM) of the test compounds in the positive ion mode was used for the quantitative determination. The matrix effects for quantitative HPLC-MS/MS and UHPLC-MS/MS methods for the determination of rimonabant in mouse plasma samples were investigated using a postcolumn infusion technique. Furthermore, the assay accuracy was demonstrated by a direct comparison of the mouse plasma levels of rimonabant obtained by these fast hyphenatedMS methods. EXPERIMENTAL METHODS Reagents and Chemicals. Ketoconazole, thioridazine, amiodarone, felodipine, rimonabant, and clofazimine as the internal standard (ISTD) were purchased from Sigma (St. Louis, MO). Methanol and acetonitrile (HPLC grade) were purchased from Fisher Scientific (Pittsburgh, PA). Formic acid (99.999%) (FA) was purchased from Aldrich Chemical Co., Inc. (Milwaukee, WI). Deionized water was generated from a Milli-Q water purifying system purchased from Millipore Corporation (Bedford, MA), and in-house high-purity nitrogen (99.999%) was used. Drug-free mouse plasma samples (with heparin) were purchased from Bioreclamation Inc. (Hicksville, NY). Chromatographic Conditions. Both HPLC and UHPLC were carried out on a Waters ACQUITY UPLC system, without column switching, using a 50 mm × 2.1 mm HALO C18 2.7 µm column (MAC-MOD Analytical, Inc., Chadds Ford, PA) or a 50 mm × 2.1 mm Waters ACQUITY UHPLC BEH (bridged ethyl hybrid) C18 (12) Hsieh, Y.; Wang, G.; Wang, Y.; Chackalamannil, S.; Korfmacher, W. A. Anal. Chem. 2003, 75, 1812-1818. (13) O’Connor, D.; Mortishire-Smith, R.; Morrison, D.; Davies, A.; Dominguez, M. Rapid Commun. Mass Spectrom. 2006, 20, 851-857. (14) Yu, K.; Little, D.; Plumb, R.; Smith, B. Rapid Commun. Mass Spectrom. 2006, 20, 544-552.

1.7 µm column (Waters Corporation, Milford, MA), respectively, as a separation media. The same mobile phases A and B composed of 100% water containing 0.1% FA and 100% acetonitrile containing 0.1% FA, respectively, were used for both HPLC and UHPLC. The same HPLC column was run on the UHPLC system using the same solvents and chromatographic conditions. The ACQUITY system is capable of pumping mobile phase at pressures up to 15 000 psi and includes an autosampler that can hold ten 96-well plates. The UHPLC instrument had a needle-in-needle injection design as well as two separate injection wash solvents to reduce sample carryover. For HPLC-MS/MS and UHPLC-MS/MS methods, a generic gradient chromatographic separation using mobile phases A and B was employed for the determination of rimonabant in mouse plasma samples as follows: 0.1 min (25% B), 0.7 min (100% B), 0.9 min (100% B), 0.91 min (25% B), and finished at 1.0 min. The mobile-phase flow rates for HPLC and UHPLC were maintained at 1.2 mL/min and 1 mL/min, respectively. For the HPLC-MS/ MS method, the retention times for the analyte and the ISTD were 0.77 and 0.53 min, respectively. For the UHPLC-MS/MS method, the retention times for the analyte and the ISTD were 0.82 and 0.51 min, respectively. The run cycle times for both methods were around 1.5 min. van Deemter plots were generated by analyzing the test mixture at a variety of mobile-phase flow rates. The height equivalent of a theoretical plate (H) was calculated from each peak in each chromatogram using the relationship

H ) L/N where L is the length of the column. N is number of theoretical plates calculated with the equation N ) 16(tR/W)2, where tR is the retention time of the peak and W is the peak width at a given peak height. The use of MS detection might seriously affect the true value of the plate height efficiency of a column due to the relatively large MS inlet “volume”. The schematic diagram of the postcolumn infusion system for the matrix effect studies on the HPLC-MS/MS system was reported elsewhere.7 The probe analyte and the ISTD dissolved in the mobile-phase solvent were continuously infused into Peek tubing in between the analytical columns and an MS/MS 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 acetonitrile (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 to 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 all test compounds 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 rimonabant in the spiked standard mouse Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

5669

plasma was 1-1000 ng/mL level. The mouse plasma samples were prepared using the protein precipitation technique. A 300 µL aliquot of acetonitrile/methanol (90/10) solution containing 1 ng/ µL of the ISTD was added to 50 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 ACQUITY autosampler to all hyphenated-MS systems for quantitative analysis. Mass Spectrometric Conditions. MS/MS analysis was performed using a quadrupole linear ion trap mass spectrometer model API 4000 QTRAP system from Applied Biosystems/MDS Sciex (Concord, ON, Canada) equipped with the APCI source. The hybrid MS/MS was operated in the positive ion mode. The APCI instrumental settings for temperature, ion gas 1, nebulizer current, collision gas, curtain gas, declustering potential, entrance potential, and collision cell exit potential were as follows: 500 °C, 50, 5, 6, 12, 100 V, 15 V, and 10 V, respectively. (The numbers without units are arbitrary values set by the Analyst software.) The MS/MS reaction selected to monitor ketoconazole, thioridazine, clofazimine, amiodarone, felodipine, and rimonabant were the transitions from m/z 531, 371, 473, 646, 384, and 463, the [M + H]+ ions, to the product ions at m/z 489, 126, 431, 100, 338, and 363, respectively. The protonated molecules were fragmented by collision-activated dissociation (CAD) with nitrogen as the collision gas at a pressure of instrument setting 5. The dwell time was set at 20 ms. The collision energies were set at 45, 30, 50, 45, 30, and 35 eV for ketoconazole, thioridazine, clofazimine, amiodarone, felodipine, and rimonabant, respectively. Data were acquired and calculated using Analyst 1.4.1 software (Applied Biosystems). RESULTS AND DISCUSSION Development of Fast HPLC-MS/MS Methods. An ideal fast chromatography method would reduce separation time while maintaining column efficiency. One common approach for reaching this goal is to reduce the particle size of packing materials to improve eddy diffusion and mass transfer resistance in the mobile phase.15,16 The use of smaller particles shortens the path length of the diffusion process of solutes between the stationary and mobile phases that is one of the sources of band broadening in HPLC. van Deemter plots are a common way to compare the column efficiencies of different stationary phases. As an example shown in Figure 1, a column packed with 1.7 µm porous silica particles is more efficient (lower H values) for separation of rimonabant than one packed with 2.7 µm fused-core silica particles at optimum flow rate under the same isocratic conditions. However, as shown in Figure 2, congruent with this improvement in column efficiency by reduction of particle size is a substantial increase in column backpressure which is proportional to the inverse of the particle size squared.17 The use of short columns with small particles for rapid separation would reduce column backpressure but at the expense of inferior column plate number. To improve the rates of mass transfer, one of the potential approaches is to use nonporous packings which are typically in the 1.5-2.5 µm range to allow faster rates of mass transfer and (15) Wu, N.; Liu, Y.; Lee, M. L. J. Chromatogr., A 2006, 1131, 142-510. (16) Kirkland, J. J. J. Chromatogr. Sci. 2000, 38, 535-544. (17) Majors, R. E. LC‚GC Eur. 2003, 16, 20-23.

5670 Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

Figure 1. van Deemter curves for the 50 mm × 2.1 mm columns packed with 2.7 µm fused-core particles and 1.7 µm porous particles. The mobile phase was 60/40 acetonitrile/water, 0.1% FA.

Figure 2. Comparison of backpressure for the 50 mm × 2.1 mm columns packed with 2.7 µm fused-core particles and 1.7 µm porous particles as a function of mobile-phase flow rates.

separation. However, the drawbacks of nonporous particles are their insufficient surface area relative to porous particles of comparable size resulting in decreased loading capacity but also the greater backpressure due to their small particle size than those with microparticulate HPLC porous packings of popular particle sizes (3 and 5 µm).18 Superficially porous packings are another conformation similar to nonporous silica particles providing increased sample capacity and lower backpressure that were prepared for rapid reversed-phase HPLC separation.19 Superficially porous silica particles with solid cores and thin porous outer shells, so-called “poroshell”, were proven to be able to offer substantial advantages over porous particles when operating at higher mobilephase flow rates. Poroshell are 5 µm particles with a 0.25 µm thick porous shell of 30 nm pores specifically designed for separating macromolecules such as polypeptides and proteins.19 The concept of using superficially porous silica packing through fused-core technology was further extended for the design of a novel stationary phase for separation of small molecules. Both poroshell and fused-core materials are superficially porous particles with solid cores. However, the fused-core particles are a layer of porous 0.5 µm particles fused to a solid 1.7 µm silica core of 9 nm pores specifically designed for separating smaller molecules (