Anal. Chem. 2003, 75, 3122-3127
High-Performance Liquid Chromatography-Atmospheric Pressure Photoionization/Tandem Mass Spectrometric Analysis for Small Molecules in Plasma Yunsheng Hsieh,*,† Kara Merkle,‡ Ganfeng Wang,† Jean-Marc Brisson,† and Walter A. Korfmacher†
Drug Metabolism and Pharmacokinetics Department, Schering-Plough Research Institute, Kenilworth, New Jersey 07033, and Applied Biosystems, Foster City, California 94404
A generic high-performance liquid chromatography (HPLC) system interfaced with an atmospheric pressure photoionization (APPI) source and a tandem mass spectrometer was developed for the quantitative determination of small molecules in plasma in support of exploratory in vivo pharmacokinetics. This report summarizes the effects of variations in reversed-phase mode HPLC conditions such as mobile-phase flow rate, solvent composition, organic modifier content, and nebulizer temperature on the photoionization efficiency of both clozapine and lonafarnib. The matrix ionization suppression effect on this method was investigated using the postcolumn infusion technique. The procedure was used to quantitate plasma levels following oral administration of 42 drug discovery compounds to rats. The pharmacokinetic results of 42 drug discovery compounds in rats evaluated by both APPI and atmospheric pressure chemical ionization interfaces were found to be well correlated. Over the past decade, electrospray ionization (ESI) and atmospheric pressure chemical ionization (APCI) techniques have become the standard ionization interfaces for high-performance liquid chromatography (HPLC)-MS/MS systems used for qualitative or quantitative analysis of small molecules.1 ESI normally produces little fragmentation while forming both protonated and deprotonated ions of most polar compounds in the positive and negative ionization mode, respectively. ESI is also vulnerable to ionization suppression from biological matrixes resulting in inconsistent analytical outcomes.2-5 In contrast to ESI, APCI generates ions from less polar compounds up to ∼1500 Da in combination with a corona discharge assisted by a heated * Corresonding author. E-mail:
[email protected]. Phone: 908740-5385. Fax: 908-740-2966. † Schering-Plough Research Institute. ‡ Applied Biosystems. (1) Plumb, R. S.; Dear, G. J.; Higton, D. N.; Mallett, D. M.; Pleasance, S.; Biddlecombe R. A. Xenobiotica 2001, 31, 599. (2) Hsieh, Y.; Chintala, M.; Mei, H.; Agans, J.; Brisson, J.; Ng, K.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2001, 15, 2481. (3) Miller-Stein, C.; Bonfiglio, R.; Olah, T. V.; King, R. C. Am. Pharm. Rev. 2000, 3, 54. (4) Mei, H.; Hsieh, Y.; Nardo, C.; Xu, X.; Wang, S.; Ng, K.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2003, 17, 97. (5) Matuszewski, B. K.; Constanzer M. L.; Chavez-Eng, C. M. Anal. Chem. 1998, 70, 882.
3122 Analytical Chemistry, Vol. 75, No. 13, July 1, 2003
nebulizer and can be operated at higher LC flow rates of around 1-2 mL/min. APCI is generally less susceptible than ESI to matrix effects.3-5 However, APCI may produce in-source fragmentation of thermally unstable compounds and therefore is not preferred for qualitative assays such as metabolite characterization.6 Another atmospheric pressure ionization technique, photoionization (APPI), introduced by Bruins and co-workers,7 is a relatively new and novel ionization interface for the HPLC-MS/MS system. It offers an alternative method of introducing samples to mass spectrometers. The APPI interface is similar to that of APCI but uses a photoionization lamp and a dopant to form dopant radical cations producing protonated eluent molecules, followed by a protontransfer reaction with the analyte. It was reported that, for certain neutral compounds with low proton affinity such as testosterone,8 idoxifene, and its alcohol metabolites,9 APPI outperformed APCI and ESI in terms of ionization sensitivity and validation statistics. Due to the advancement of combinatorial chemistry and parallel synthesis, today’s pharmaceutical industry has substantially increased the number of new chemical entities produced each year. Consequently, there continues to be a need for developing sensitive and universal HPLC-API/MS/MS assays for detecting drug discovery compounds in large numbers of samples derived from various in vitro and in vivo experiments. In this work, we further investigate the potential for using APPI as an alternative to ESI and APCI and demonstrate its applicability to a range of real samples for the determination of small molecules. The postcolumn infusion technique was adapted for the investigation of the matrix ionization suppression effects on the HPLC-APPI/MS/MS system. Several critical factors such as the compositions of mobile phase used for the reversed-phase mode chromatography, which might affect the ionization efficiency of drug components when APPI is used, were explored using the flow injection analysis (FIA) method. The suitability of the APPI technique for small-molecule determinations was confirmed through correlation of rat pharmacokinetic results obtained from both APPI and APCI systems. (6) Cox, K. A.; Clarke, N.; Ridgen, D.; Korfmacher, W. A. Am. Pharm. Rev. 2001, 4, 45. (7) Robb, D. B.; Covey, T. R.; Bruins, A. P. Anal. Chem. 2000, 72, 365. (8) Alary, J.; Berthemy, A.; Tuong, A.; Uzabiaga, M. Procedings of the 50th ASMS Conference on Mass Spectrometry and Allied Topics, 2002. (9) Yang, C.; Henion, J. J. Chromatogr., A 2002, 970, 155. 10.1021/ac0300082 CCC: $25.00
© 2003 American Chemical Society Published on Web 05/13/2003
Figure 1. Schematic diagram of the postcolumn infusion system.
EXPERIMENTAL METHODS Reagents and Chemicals. Lonafarnib, a drug candidate for antitumor therapy, and the test compounds 1-42 with molecular weights ranging from 424 to 624 are new chemical entities from five different drug discovery programs produced by ScheringPlough Research Institute. The chemical structure of lonafarnib, known as SCH 66336, was reported previously.10 Acetonitrile and toluene (HPLC grade) were purchased from Fisher Scientific (Pittsburgh, PA). Formic acid (80%), clozapine, and ammonium acetate (99.999%) were purchased from Aldrich Chemical Co., Inc. (Milwaukee, WI). Deionized water was generated from a Milli-Q water purifying system purchased from Millipore Corp. (Bedford, MA), and house high-purity nitrogen (99.999%) was used. Drugfree rat plasma was purchased from Bioreclamation Inc. (Hicksville, NY). Mobile phases A and B were composed of 4 mM ammonium acetate in water-acetonitrile (90:10) and 4 mM ammonium acetate in water-acetonitrile (10:90), respectively. Equipment. HPLC-MS/MS analysis was performed using a PE Sciex (Concord, ON, Canada) model API 3000 triple quadrupole mass spectrometer equipped either with heated nebulizer (APCI) or PhotoSpray (APPI) probes. The APPI system is composed of a heated nebulizer to vaporize the sample prior to inducing ionization, a power supply for the krypton lamp for photoionization, a nitrogen supply for cooling the lamp and protecting the optical window, and a syringe pump for dopant delivery. The toluene solution used as dopant was continuously introduced into the heated nebulizer through a fused-silica capillary at 10 µL/min for quantitation of drug discovery compounds in support of PK studies. The temperature of the heated nebulizer was set at 450 °C for both APPI and APCI sources. The working principle and schematic representations of the HPLCAPPI/MS/MS system were depicted in detail elsewhere.7 The HPLC system consisted of a Leap autosampler with a refrigerated sample compartment (set to 10 °C) from Leap Technologies (Carrboro, NC), Shimadzu on-line degasser, LC-10ADVP pump, and LC-10AVP controller (Columbia, MD). A Synergi C18 column (2.0 × 30 mm, 4 µm) from Phenomenex Inc (Torrance, CA) was used as the analytical column. The Quadra 96 (Tomtec, Hamden, CT) system was used for semiautomated sample preparation with the protein precipitation method. A schematic diagram of the postcolumn infusion system for the matrix effect studies is shown in Figure 1. The experimental mass spectrometric conditions were determined using the same generic state file for both APCI and APPI interfaces without optimization for individual compounds. (10) Liu, M.; Bryant, M. S.; Chen, J.; Lee, S.; Yaremko, B.; Lipari, P.; Malkowski, M.; Ferrari, E.; Nielsen, L.; Prioli, N.; Dell, J.; Sinha, D.; Syed, J.; Korfmacher, W.; Nomeir, A.; Lin, C.; Wang, L.; Taveras, A.; Doll, R.; Njoroge, G.; Mallams, A.; Remiszewski, S.; Catino, J.; Girijavallabhan, V.; Kirschmeier, P.; Bishop, R. Cancer Res. 1998, 58, 4947.
Clozapine and lonafarnib were continuously infused into PEEK tubing between the analytical column and mass spectrometer through a tee using a Harvard Apparatus model 2400 (South Natick, MA) syringe pump. Either a protein precipitation extract of blank rat plasma or mobile phase B (10 µL) was injected into the HPLC column for comparison of ionization responses. The effluent from the HPLC column was mixed with the infused compounds prior to the APPI interface. Animal Dosing and Sample Collection. For each compound, two rats were dosed via oral administration at a dose of 10 mg/ kg producing 12 study plasma samples which were pooled across the two rats to generate six “single-pooled” samples at the six time points (0.5, 1, 2, 3, 4, and 6 h postdose).11 The single-pooled rat plasma samples were stored at -20 °C until analysis. The sample set was assayed via an abbreviated two-point calibration curve, which results in six samples plus eight standards (including two blank plasma samples) and two solvent blanks for all six test compounds as described elsewhere.11 Standard and Sample Preparation. Stock solutions of clozapine, lonafarnib (used as internal standard, ISTD), and all other test compounds were prepared as 1 mg/mL solutions in methanol. Analytical standard samples for rat PK studies were prepared by spiking known quantities of the standard solutions to blank plasma to provide an abbreviated three-point calibration curve at 25, 250, and 2500 ng/mL levels. Data are reported for samples within the analytical range established by two appropriate standards (depending upon their estimated plasma concentrations, e.g., 25-250 or 250-2500 ng/mL) and (2×) above and (0.4×) below this range (e.g., for 25 and 250 ng/mL, up to 500 ng/mL, and down to 10 ng/mL).11 Samples with calculated concentration values below 10 ng/mL were reported as zero. For the protein precipitation method, a 150-µL acetonitrile solution containing 1 ng/µL ISTD was added to 50 µL of the pooled study plasma samples and standard plasma samples in a 96-well plate. After vortexing and centrifugation, the supernatant was transferred to a new 96-well plate using the Tomtec Quadra 96 system. Aliquots of 10 µL were injected for HPLC-MS/MS analysis. Chromatographic Conditions. Chromatographic separation was achieved using mobile phases A and B. The HPLC conditions for rat plasma analyses of all test compounds are summarized in Table 1. The effluent from the HPLC systems was connected directly to the mass spectrometer when either the APPI or the APCI source was used. The retention times for the test compounds and ISTD were between 0.7 and 1.2 min. (11) Korfmacher, W. A.; Cox, K. A.; Veals, J.; K. Ng, K.; Hsieh, Y.; Wainhaus, S.; Broske, L.; Prelusky, D.; Nomeir, A.; White, R. E. Rapid Commun. Mass Spectrom. 2001, 15, 335.
Analytical Chemistry, Vol. 75, No. 13, July 1, 2003
3123
Figure 2. Relative APPI responses of clozapine (solid) and lonafarnib (open) as a function of the water-acetonitrile ratios of the solvent.
Figure 4. APPI peak responses of (A) clozapine and (B) lonafarnib as a function of delivery speed of dopant using flow injection analysis. Table 1. HPLC Conditions Used for Both APPI and APCI Interfaces
time (min) 0.00 0.30 0.35 0.55 0.60 1.5 1.6 Figure 3. Effects of LC eluent flow rates on photoionization efficiency for (A) clozapine and (B) lonafarnib.
Mass Spectrometric Conditions. The mass spectrometer was operated in the positive ion mode. The ion spray voltage for APPI was set at 1.3 kV. The rest of the mass spectrometric parameters for both APCI and APPI interfaces were set to be identical. The precursor [M + H]+-product ion MS/MS transitions selected to monitor clozapine and lonafarnib were m/z 327 f 270 and m/z 639 f 471, respectively, for both APPI and APCI sources. The protonated molecules were fragmented by collision-induced dissociation with nitrogen as collision gas at a pressure of instrument setting 5. The collision offset voltage was varied between 30 and 45 V depending on the mass spectra of the test compounds. Data were acquired and processed using Analyst 1.1 software (PE Sciex). RESULTS AND DISCUSSION For the HPLC-APPI/MS/MS system, the HPLC eluent is first vaporized and subjected to photoionization prior to mass spec3124 Analytical Chemistry, Vol. 75, No. 13, July 1, 2003
composition of mobile phase (%) A B 90 90 0 0 0 0 90
10 10 100 100 100 100 10
flow rates (mL/min)
valve positionsa
1.1 1.1 1.1 1.1 0.8 0.8 1.1
B B B B A A B
a B, HPLC flow diverted to waste; A, HPLC flow diverted to the mass spectrometer.
trometric detection. Ionization is mainly based on charge and proton transfer to the analytes from the protonated dopant molecules that have been ionized by the 10-eV photons produced by a vacuum-ultraviolet lamp.7 The photoionization reactions may also simultaneously or sequentially involve many steps such as photoexcitation, photodissociation, fluorescence, charge transfer, and proton transfer.12 The detection sensitivity of the analytes should be directly related to the photoionization efficiency, which could be also dependent on many factors including solvent composition, which can also affect sensitivity in both ESI and APCI interfaces.13-14 In this work, the effects of experimental conditions on the photoionization efficiency were investigated using the FIA (12) Wit, J. S.; Jorgenson, J. W. J. Chromatogr. 1987, 411, 201. (13) Rauha, J.; Vuorela H.; Kostianinen, R. J. Mass Spectrom. 2001, 36, 1269. (14) Leinonen, A.; Kuuranne, T.; Kostiainen, R. J. Mass Spectrom. 2002, 37, 693.
Figure 5. Effect of organic modifier compositions such as ammonium acetate (open) and formic acid (solid) on photoionization efficiency of (A) clozapine and (B) lonafarnib.
technique with a mixture solution of clozapine and lonafarnib. The effects of water-acetonitrile ratios on the peak areas of the extracted ion chromatograms of clozapine and lonafarnib at a constant flow rate of 1 mL/min are shown in Figure 2. Figure 2 indicates that the photoionization efficiency decreases as the aqueous contents in the solvents increase for both analytes at a constant probe temperature of 400 °C. The reduced sensitivities due to the higher water content of the mobile phase could be regained for both analytes by increasing the photospray nebulizer temperature (data not shown). This phenomenon suggests that incomplete vaporization occurred with higher aqueous contents at a probe temperature of 400 °C. The ionization potentials (IP) of most common solvents used in reversed-phase chromatography are greater than the photon energy emitted from the krypton discharge lamp (water, IP ) 12.6 eV; methanol, IP )10.8 eV; acetonitrile, IP ) 12.2 eV). In principle, the photon energy should ionize only the analytes and dopant (toluene, IP ) 8.83 eV) molecules, which results in the formation of radical cations. These ions generated through UV light react in turn with solvent molecules through collision in gas phase producing intermediate protonated charged clusters. The production of analytical ions is through the proton-transfer process if the proton affinity of the analytes is higher than that of solvent molecule.13,14 Other chemicals such as acetone have been used as the dopant in the APPI system.7 In this work, we selected toluene as dopant because it has a lower IP than acetone and had been shown to provide better detection sensitivity for the compounds tested previously.7 The production of the protonated clozapine and lonafarnib ions
in the APPI system was found to be 100 times greater when toluene was introduced than that without addition of the dopant. We also observed that the solvent combination of water-methanol doubled the photoionization efficiency for clozapine and lonafarnib as compared to a water-acetonitrile solvent system. A possible explanation for this better sensitivity could be due to more effective nebulization and vaporization processes provided by methanol over acetonitrile. The relationship of solvent eluent flow rate versus the photoionization responses of clozapine and lonafarnib at nebulizer temperatures of 400 and 500 °C are given in Figure 3A and B, respectively. As indicated in Figure 3, the relative responses of both analytes were reduced significantly (100% down to 20%) as the flow rate was increased from 0.1 to 0.6 mL/min at a consistent dopant delivery speed. The photoionization responses of lonafarnib obtained at two different temperatures were found to be no significant difference. This suggested that the heat generated at 400 °C should be sufficient to vaporize both lonafarnib and the solvent. However, the ion responses of clozapine with APPI source at 400 °C are higher than those at 500 °C. This indicated that clozapine molecules might be thermally degraded when nebulizer temperature above 400 °C were used. The lower sensitivity at higher mobile-phase flow rates was assumed to be the result of the dilution effect on the dopant. This hypothesis was further examined by the enhanced detection sensitivity of both analytes at a constant solvent flow rate of 1 mL/min when the delivery speed of toluene solvent was increased as shown in Figure 4. At each delivery speed (10-80 µL/min) of dopant solvent, 10 µL of mixture containing clozapine and lonafarnib was injected into FIA-APPI/MS/MS system three times. Both formic acid and ammonium acetate are commonly used modifiers for the HPLC-APCI/MS/MS system. We observed that the addition of ammonium acetate and formic acid had a slight increase on the photoionization efficiency at concentrations below 15 mM as demonstrated in Figure 5. At the concentrations of both modifiers over 15 mM, the ion production of clozapine and lonafarnib increased proportionally. In general, ammonium acetate contributed more than formic acid to the photoionization efficiency of both compounds tested in this work. Overall, the introduction of dopant solvent had shown the most impact on the photoionization efficiency of both test compounds. These findings indicate that the proton-transfer reaction is one of the rate-limiting steps in the ionization process and the presence of dopant is essential for photoionization detection. A recently understood concern about assay reliability when HPLC-MS/MS methods are being developed is the increased likelihood of encountering a matrix ionization suppression problem.4-5,15-18 In addition, the accuracy and reproducibility of the analytical results is often affected by the varying degree of the matrix effects due to different sample preparation methods or ionization techniques.4 The infusion HPLC-APPI/MS/MS chromatograms of clozapine and lonafarnib after a 10-µL injection of either mobile phase or plasma protein precipitation extract are (15) Bonfiglio, R.; King, R. C.; Olah, T. V.; Merkle, K. Rapid Commun. Mass Spectrom. 1999, 13, 175. (16) King, R.; Bonfiglio, R.; Fernandez-Metzler, C.; Miller-Stein, C.; Olah, T. J. Am. Soc. Mass Spectrom. 2000, 11, 942. (17) Wang, G.; Hsieh, Y.; Cheng, K.-C.; Ng, K.; Korfmacher, W. A. Spectrosc.,Int. J. 2003, 17, 511. (18) Hsieh, Y.; Brisson J.; Wang, G.; Ng, K.; Korfmacher, W. A. J. Pharm. Biomed. Anal., in press.
Analytical Chemistry, Vol. 75, No. 13, July 1, 2003
3125
Figure 6. Reconstructed infusion HPLC-APPI/MS/MS chromatograms of clozapine and lonafarnib following mobile-phase (solid line) and blank plasma precipitation extract injections (dot line). The region showing lower responses indicated the area of matrix ionization suppression.
Figure 7. Representative reconstructed HPLC-APPI/MS/MS chromatograms of lonafarnib, and test compounds 26-31 (from top to bottom) from standard rat plasma at a concentration of 250 ng/mL.
given in Figure 6. The differences in the infusion chromatograms between the mobile-phase injection and the rat plasma extract injection were considered to be caused by the matrix ion suppression effects due to plasma sample extract constituents eluting from the column. Figure 6 demonstrates that the degree of loss of APPI response and the length of time required for the APPI response to return to its presample injection sensitivity were consistent with both clozapine and lonafarnib in this study. The APPI responses of clozapine and lonafarnib with a given rat plasma 3126
Analytical Chemistry, Vol. 75, No. 13, July 1, 2003
protein precipitation extract were not vulnerable to the matrix ion suppression effect. However, we observed more severe matrix effects at the same chromatographic region when the dopant was not in use (data not shown). For reliable quantitative determination, it is suggested that the retention times of all test compounds should be in the region of little or no matrix ion suppression. The retention times of clozapine, lonafarnib, and other test compounds for both APPI and APCI assays were located in the matrix ionization suppression-free region. The representative extracted
Table 2. Comparison of Rat PK Results Obtained by APPI and APCI Interfaces
compds
mass transition ranges
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
583 f 412 570 f 399 552 f 381 588 f 417 596 f 425 550 f 395 593 f 219 624 f 453 509 f 338 622 f 320 677 f 350 541 f 320 619 f 350 544 f 219 569 f 397 514 f 460 603 f 379 519 f 348 441 f 192 439 f 192 554 f 365 487 f 373 537 f 163 560 f 323 504 f 379 513 f 348 458 f 192 484 f 192 474 f 365 454 f 373 470 f 163 574 f 320 563 f 392 597 f 426 559 f 219 525 f 219 485 f 314 542 f 371 424 f 192 452 f 358 453 f 359 548 f 441
a
Cmax (ng/mL) APPI APCI
AUC(0-6 h) (ng‚h/mL) APPI APCI
166 47 465 159 20 164 19 151 166 529 554 429 nda 109 232 579 106 128 628 1320 54 3460 nd nd 88 981 384 1480 502 460 2030 678 571 193 76 65 268 73 782 2650 1250 94
763 188 2211 523 72 255 23 486 399 2098 1427 1491 nab 267 990 1845 406 481 2761 3111 91 5313 na na 308 2118 1376 6285 1360 938 4567 1453 2510 905 273 157 1127 302 2809 5047 1913 275
189 56 497 159 17 141 16 161 132 603 546 329 nd 108 206 340 95 116 465 1102 45 2899 nd nd 65 900 441 1956 520 267 1889 768 747 214 72 54 267 60 739 2551 810 121
844 184 2234 466 65 263 21 499 413 1876 1080 1669 na 314 822 1389 354 436 2213 2609 73 4575 na na 283 1938 1421 7342 1423 598 4683 1650 3094 945 246 147 1129 251 2891 5877 1356 360
nd, not detected. b na, not available.
HPLC-APPI/MS/MS chromatograms of the ISTD and compounds 26-31 are shown in Figure 7. The peak areas and the retention times for all the test compounds in rat standard and study plasma were reproducible throughout the experiments. The same rat plasma standard and study samples were independently analyzed for all the drug discovery compounds using either APPI or APCI methods under the identical HPLC conditions. The rat PK results of compounds 1-42 in terms of Cmax and AUC(0-6 h) measured by the APPI method with those obtained by the APCI method are summarized in Table 2. Similar correlation coefficients were obtained for both Cmax and AUC(0-6 h) parameters, r2 ) 0.980 and r2 ) 0.982, respectively, using both approaches, as demonstrated in Figure 8. The Student t-test results indicated no significant difference of individual concentrations at
Figure 8. Correlation of analytical results obtained by HPLC-APCI/ MS/MS method vs the HPLC-APPI/MS/MS method in terms of (A) Cmax and (B) AUC(0-6 h).
each time point for those test compounds determined by both assays with 95% confidence (R ) 0.05). The above results demonstrated that this HPLC-APPI/MS/MS method was equivalent to the HPLC-APCI/MS/MS methods in terms of analytical accuracy. CONCLUSIONS A sensitive and fast turnaround method based on a HPLCAPPI/MS/MS system has been demonstrated for the quantitative determination of small drug components in rat plasma. Several major experimental factors employed in the described method such as the delivery speed of dopant, the composition of HPLC eluent, nebulizer temperature, and mobile-phase flow rate significantly affected the ionization efficiency of APPI system. The HPLCAPPI-MS/MS method should prove to be applicable for PK screening and provided analytical results that were equivalent to those obtained by our standard HPLC-APCI/MS/MS approach. It was further shown that APPI provides minimal matrix ionization suppression effect for two study compounds, which may suggest that the APPI source is less sensitive to matrix ion suppression than either APCI or ESI sources. ACKNOWLEDGMENT The authors thank our animal dosing group for planing PK studies and Medicinal Chemistry group for synthesizing the test compounds presented in this work. Received for review January 2, 2003. Accepted April 8, 2003. AC0300082
Analytical Chemistry, Vol. 75, No. 13, July 1, 2003
3127
