Anal. Chem. 1996, 68, 1658-1660
Rapid, Solid Phase Extraction Technique for the High-Throughput Assay of Darifenacin in Human Plasma Barry Kaye,* William J. Herron, Paul V. Macrae, Sylvia Robinson, David A. Stopher, Richard F. Venn, and William Wild
Department of Drug Metabolism, Pfizer Central Research, Sandwich, Kent CT13 9NJ, U.K.
A novel method has been developed for the rapid solid phase extraction of drugs and metabolites from biological fluids, prior to further analysis. The newly designed, 96tube micropreparation block facilitates high throughput by enabling the extraction of 96 samples simultaneously. The system is described, linked to HPLC/APCI-MS/MS, for the determination of darifenacin in human plasma. The resulting procedure, using deuterated darifenacin as internal standard, is validated over the concentration range 25-2000 pg/mL; accuracy (0.6-4.6%) and precision (3.6-18.8%) are considered acceptable and overall recovery was determined to be ∼50%. Along with other groups,1-3 we are developing the application of atmospheric pressure ionization (API) mass spectrometry in the field of bioanalysis, with particular reference to the quantitative determination of drugs and their metabolites in biofluids. Thus, we published an analytical procedure for the determination of abanoquil in blood, which exemplified the subnanogram per milliliter sensitivity and selectivity that can be achieved using an API source and the short run times (1-2 min) which are possible using the small HPLC columns (30 × 4.6 mm) linked into the system.4 Clearly, application of this technology will enable a high throughput in drug analysis, which is largely dependent on the rate of sample workup. This paper describes a newly designed micropreparation system that enables high throughput by the simultaneous workup of 96 samples by solid phase extraction, before analyzing these by APCI tandem mass spectrometry. The technique has been applied successfully to the sensitive determination of darifenacin in human plasma. Darifenacin [(S[-2-[1-[2(2,3-dihydrobenzofuran-5-yl)ethyl]-3-pyrrolidinyl]-2,2-diphenylacetamide; Figure 1] is a M3 selective muscarinic receptor antagonist under clinical development for urinary incontinence. To support the interpretation of clinical studies, the analytical procedure described was designed to determine the drug at relatively low clinical doses. EXPERIMENTAL SECTION Solid Phase Micropreparation System. The design of the solid phase micropreparation system was based around the format (1) Fouda, H.; Nocerini, M.; Schneider, R.; Gedutis, C. J. Am. Soc. Mass Spectrom. 1991, 2, 164. (2) Avery, M.; Mitchell, D.; Falkner, F.; Fouda, H. Biol. Mass Spectrom. 1991, 21, 353. (3) Huang, E. C.; Henion, J. D. Anal. Chem. 1991, 63, 732. (4) Kaye, B.; Clarke, M. W. H.; Cussans, N. J.; Macrae, P. V.; Stopher, D. A. Biol. Mass Spectrom. 1992, 21, 585.
1658 Analytical Chemistry, Vol. 68, No. 9, May 1, 1996
Figure 1. Structures of darifenacin and deuteriodarifenacin. A possible origin of fragment ion m/z 147 is indicated.
of the 96-well plates which have been used for conducting bioassays for many years. The system comprises three main components (see Figure 2): a micropreparation block (approximate size 12.5 × 8 × 4 cm), which incorporates 96 tubes containing solid phase sorbent arranged in the 8 × 12 format, volume extension strips (1.5-mL capacity), and a second block containing 96 deep collecting wells (1-mL capacity; approximate size 12.5 × 8 × 3 cm). Either block may be located in an aluminum base plate which is constructed so that a low vacuum may be applied to the system to enable solvent to percolate through the solid sorbent more rapidly. The blocks and extension strips are moulded in polypropylene. The tubes in the micropreparation block are filled with C18 Bondesil (Varian Ltd.; Harbor City, CA). The micropreparation block and the assembly of this block with the second collection block are covered by U.K. Patent 2,243,446 and its foreign equivalents. The complete assembly will be available, under licence, from Porvair Filtronics Ltd. (Shepperton, Middlesex, U.K.). Materials. Darifenacin [(S)-2-{1-[2-(2,3-dihydrobenzofuran-5yl)ethyl]-3-pyrrolidinyl}-2,2-diphenylacetamide] hydrobromide was obtained from Pfizer Central Research, Sandwich, U.K. Darifenacin-d5 (Figure 1) was synthesized and supplied by Dr. G. N. Maw, Pfizer Central Research. Isotopic purity was >99.6%. Other reagents were of either HPLC or analytical grade and were used without further purification. Sample Preparation. Plasma samples were thawed at 37 °C for 15-20 min or at 2-5 °C in a refrigerator overnight and centrifuged at 2000g for 5-10 min. The sorbent in each of the tubes in the micropreparation block was conditioned by washing with methanol (1 mL), followed by buffer (1 mL; 0.2 M Na2HPO4, pH 7.0). To each tube was added a mixture of plasma (1 mL) 0003-2700/96/0368-1658$12.00/0
© 1996 American Chemical Society
and internal standard solution (0.5 mL; acetonitrile/0.2 M Na2HPO4, 3:7, v/v, containing 1.0 ng of internal standard). Tubes were then washed with 0.2 M Na2HPO4 buffer (1 mL, pH 7.0) and aqueous methanol (1 mL, 1:1, v/v), and analytes were eluted with methyl tert-butyl ether (0.8 mL, containing 1% triethylamine by volume) into the block containing deep collecting wells, under slight vacuum. The residual fluid in these wells was evaporated to dryness under a nitrogen stream in a sample concentrator (79mm needles in 96-well format; Techne Ltd., Fisons, Loughborough, U.K.). To the residue in each tube was added mobile phase (150 µL), and following capping of the tubes with strips of polypropylene caps, the block was vortex mixed. Water (100 µL) was added to each tube and the block vortex mixed again prior to centrifuging at 2000g for 30-60 min at 4 °C (Jouan C312, Jouan Ltd., Tring, U.K.). The block was then located unsealed in an autoinjector for HPLC (Merck-Hitachi AS 4000, BDH-Merck, Poole, U.K.). The autoinjector was programmed to wash immediately after each injection to reduce time lag between injections. The initial dispensing of plasma and internal standard solution into the micropreparation block may be done manually or, for convenience, automatically using a robotic sample processor (Tecan RSP; Tecan U.K., Goring-on-Thames, U.K.). The validation data reported here were obtained using the robot. HPLC/APCI-MS/MS. HPLC was performed on a PerkinElmer 3 µM C8 column (30 × 4.6 mm) using a Shimadzu LC-6A pump (Dyson Instruments Ltd., Houghton-le-Spring, U.K.). Sample injection solution (150 µL) was introduced onto the column, and the flow rate used was 1 mL/min. The mobile phase was methanol/water (80:20, v/v) containing ammonium acetate (0.02 M) and triethylamine (0.005 M), adjusted to pH 7.0 with acetic acid. Mass spectrometric detection was carried out using a PE Sciex API III Plus triplet quadrupole instrument (Toronto, Canada) operating in the positive ion APCI mode, using the heated nebulizer with corona discharge. The nebulizer probe was run at 500 °C, with nitrogen nebulizing gas at a pressure of 80 psi. Multiple reaction monitoring was employed using argon as collision gas at a thickness of (260-280) × 1013 molecules/cm2, with collision energy of 30 eV. Parent to daughter transitions were monitored for m/z 427-147 and for m/z 432-147 for darifenacin and deuterated internal standard, respectively. The resolution of the quadrupoles was set to 2.2 amu at half-height. Dwell time for each transition was 100 ms. Peak area ratios for the selected ions were determined automatically using the PE Sciex software package MacQuan 1.1.2. Calibration Curve and Quantification. Calibration samples were prepared in plasma with each batch of test samples to cover the range 25-2000 pg/mL for darifenacin. Thus to 1.5-mL samples of plasma was added darifenacin to given concentrations of 25, 50, 100, 200, 500, 1000, 1500, and 2000 pg/mL drug. The compound was added in volumes of up to 60 µL of aqueous methanol (1:1, v/v). Calibration lines were constructed for the drug by plotting peak area ratios of analyte to internal standard against analyte concentration. A weighted (1/y2) linear regression line was fitted over the 80-fold concentration range for the analyte.5 Drug concentrations in test samples were interpolated from the calibration line. (5) Caulcutt, R.; Boddy, R. Statistics for Analytical Chemists; Chapman and Hall: London, 1989; Chapter 8.
Figure 2. Cross section through 96-well micropreparation solid phase extraction system.
RESULTS AND DISCUSSION Use of the API mass spectrometer in the multiple reaction monitoring mode, i.e., monitoring parent to daughter ion transitions, gives a very highly selective mode of detection of analytes.6 Thus, this powerful mass discrimination considerably reduces the need for lengthy chromatography prior to detection and we routinely use short columns of 30-mm length and retention times of ∼1 min to our quantitative work. Even with injection washing time added, there is only 2-3 min between injections, thus potentially allowing a high throughput of samples during normal working practices. However, in our experience, conventional workup of the analyte from the biofluid was rate-determining, whether by liquid/liquid or solid phase extraction, yielding a turn around of about 50-70 samples per batch, which is albeit appreciable when compared to conventional HPLC or GC methods where there may be up to 30 min between injections. However, some way of increasing workup rate would be a real advantage in assaying the large number of samples which often emanate from clinical studies. In other laboratories within Pfizer Central Research, an enzyme inhibition assay has been run successfully at relatively high throughput, using a similar assembly. We utilized the basic design for a solid phase extraction assembly for 1-mL plasma samples for drug analysis by API-MS. A second aim was, in order to reduce the number of manual transfer steps, that the second block into which the analytes were eluted should locate conveniently into a sample concentrator and an autoinjector linked to the mass spectrometer. The assembly used is shown in Figure 2. The deepened microtiter plate filled with sorbent over a frit, combined with the snap-fit volume extension tubes can accept up to 2 mL of fluid above the column. The spigots beneath the column prevent eluant solution creeping round the base of the columns with possible cross contamination when vacuum was applied to collect eluted analyte into the second deep well collection block. This collection block could be used in a Techne sample concentrator (modified with 96 tips) to dry samples and could be located in a Merck-Hitachi AS 4000 autoinjector for sample injection into the HPLC linked to the mass spectrometer, thus obviating further manual transfer other than the addition of the final injection solution. To achieve high sensitivity in the HPLC/MS analysis, conditions for chromatography and ionization were optimized essentially (6) Busch, K. L.; Glish, G. L.; McLuckley, S. A. Mass Spectrometry/Mass Spectrometry: Techniques and Applications of Tandem Mass Spectrometry; VCH Publishers Inc.: New York, 1988; Chapter 2.
Analytical Chemistry, Vol. 68, No. 9, May 1, 1996
1659
Figure 3. Daughter ion spectra obtained from the protonated molecular ions of darifenacin-d5 (top) and darifenacin (bottom). Spectra obtained in positive ion APCI mode. Table 1. Accuracy and Precision for the Assay of Darifenacina concn (pg/mL) prepared
measd mean (SD)
n
accuracy (%)
precision (%)
25 500 2000
26.4(5) 503(51) 1910(69)
22 24 24
5.4 0.65 4.6
18.8 10.2 3.6
a Combined data from three separate runs. Accuracy (mean of measured value - prepared value)/(prepared value)] × 100; precision, relative standard deviation ) SD × 100/mean.
as we described previously.4 Also, use of darifenacin-d5 as internal standard and using the triplet quadrupole in the multiple reaction monitoring mode was conducive to better detection; the transition, parent to daughter ion m/z 147 was monitored for both darifenacin and darifenacin-d5. The daughter mass spectra of the compounds are shown in Figure 3. Fragmentation of the parent ions is apparent with the daughters being prominent ion in each case. Using darifenacin-d5 as internal standard, overall recovery of the analyte was ∼50%. Darifenacin and darifenacin-d5 coelute with a retention time of 0.7 min. Response was linear over the range 25-2000 pg/mL and typical calibration curve parameters were as follows: slope, 1.419; intercept, 0.016; correlation coefficient, 0.999. Precision and accuracy of the procedure were determined by analyzing batches of replicate fortified samples on three separate occasions, and overall mean data are presented in Table 1. The accuracy and precision of the method was considered satisfactory; the mean relative standard deviation for replicates at 25 pg/mL was 18.8% and this concentation was defined as the lower limit of quantification for darifenacin in human plasma. Using the micropreparation system, 96 samples could be worked up, conveniently, in ∼1.5 h, compared with 50-60 samples in ∼3 h using manual procedures. The method has been applied successfully to determine plasma concentrations of darifenacin in subjects in clinical studies fol-
1660 Analytical Chemistry, Vol. 68, No. 9, May 1, 1996
Figure 4. Typical plasma concentration/time profile of darifenacin in a subject after the last dose of a 5-day multiple oral dose regimen of 2.5 mg t.i.d.
lowing both single and multiple doses. An example of a concentration versus time profile following the last dose of a 5-day multiple oral dose regimen of 2.5 mg t.i.d. is shown in Figure 4. Although in this report, the microprepartaion technology has been linked to HPLC/APCI-MS, it is clearly generally applicable to any analytical technique that requires the solid phase extraction of analytes from large numbers of samples. ACKNOWLEDGMENT The authors thank Mr. Tony Castleman of Porvair Filtronics for stimulating discussions and advice on the manufacture of the 96-tube assembly.
Received for review July 31, 1995. Accepted January 30, 1996.X AC9507552 X
Abstract published in Advance ACS Abstracts, March 1, 1996.
