Enzymatic Tissue Digestion as an Alternative Sample Preparation

Feb 13, 2004 - compounds in dog and rat brain tissue. Comparison .... Canada) model API 3000 triple quadrupole mass spectrometer equipped with a PE ...
0 downloads 0 Views 114KB Size
Download PDF
Anal. Chem. 2004, 76, 1761-1767

Enzymatic Tissue Digestion as an Alternative Sample Preparation Approach for Quantitative Analysis Using Liquid Chromatography-Tandem Mass Spectrometry Chongwoo Yu, Lara D. Penn, John Hollembaek, Wenlin Li, and Lucinda H. Cohen*

Pfizer Global Research and Development, Bioanalytical Research, Department of Pharmacokinetics, Dynamics & Metabolism, 2800 Plymouth Road, Ann Arbor, Michigan 48105

Compound extraction from biological tissue often presents a challenge for the bioanalytical chemist. Labor-intensive homogenization or sonication of whole or powdered tissue is performed before compounds can be extracted and analyzed. Enzymatic digestion is commonly used for tissue dissociation and cell harvesting and offers the advantages of unattended sample preparation, potential automation, and low cost. The feasibility of enzymatic digestion as an alternate tissue preparation technique was evaluated for bioanalysis of drugs in conjunction with LC/MS/MS. Two different enzymes (collagenase and proteinase K) that are known to degrade connective tissues to allow tissue dissolution were chosen for evaluation, employing wellknown antidepressants desipramine and fluoxetine as test compounds in dog and rat brain tissue. Comparison between enzymatic digestion and conventional homogenization tissue preparation was performed, including investigation of matrix ionization suppression of both methods using a postcolumn infusion system. Results showed that enzymatic digestion has extraction efficiency comparable to homogenization. Matrix ionization suppression was not observed for either the test compounds evaluated or the sample extraction method. Test compound levels of incurred tissue samples prepared by enzymatic digestion were in good agreement with the values obtained by the conventional homogenization tissue preparation, indicating that enzymatic digestion is an appropriate tissue sample preparation method. Traditional biological sample preparation has been carried out for several years due to its low cost and straightforward methods. However, the demand for increased productivity and faster and high throughput assays has driven investigation of new technologies. These new techniques are expected to be faster and provide at least equivalent, if not superior, reproducibility and analyte extraction efficiency. A high demand still exists for improvement of biological sample preparation techniques, especially for tissues. For many types of tissue, sample preparation is a bottleneck in the bioanalytical process. The extraction of analytes from solid matrixes has been * Phone: 734-622-1803. Fax: 734-622-5115. E-mail: [email protected]. 10.1021/ac035077v CCC: $27.50 Published on Web 02/13/2004

© 2004 American Chemical Society

an active development area in modern sample preparation technology, especially in the past decade. Several new sample preparation methods for organic compounds in solid matrixes, including supercritical fluid extraction,1-3 modern Soxhlet extraction,4-5 microwave-assisted extraction,6-7 and pressurized fluid extraction,8-10 have been investigated; however, most of these techniques are more applicable to environmental analysis (i.e., soil, food, etc.). As a result, conventional homogenization or sonication is still the most widely accepted sample preparation method for the analysis of compounds in biological matrixes, such as tissue; however, these methods are labor-intensive and greatly lengthen the analysis time. In addition, extensive manual homogenization of tissue samples presents a serious noise hazard. One means to speed up the homogenization process was recently introduced11 and used a multiprobe, parallel processing approach. Throughput can be increased by a factor of 4-6 (depending on instrument configuration); however, this device still produces noise and may need manual attention during the sample preparation process. Enzymatic digestion is commonly used for tissue dissociation and cell harvesting and offers the advantages of unattended sample preparation, potential automation, and low cost. Although the enzymatic digestion technique has been utilized for decades, only a few papers12 have been published attempting to employ enzymatic digestion in tissue sample preparation of small molecules, rather than protein, DNA, or RNA harvesting. There have been no previous reports evaluating the feasibility of enzymatic (1) Howard, A. L.; Shah, M. C.; Ip, D. P.; Brooks, M. A.; Strode, J. T., 3rd.; Taylor, L. T. J. Pharm. Sci. 1994, 83 (11), 1537-1542. (2) Messer, D. C.; Taylor, L. T. Anal. Chem. 1994, 66 (9), 1591-1592. (3) Levy, J. M. LC-GC 1999, 17 (6S), S14-S21. (4) Randall, E. L. J. AOAC 1974, 57 (5), 1165-1198. (5) Arment, S. LC-GC 1999, 17 (6S), S38-S42. (6) Walter, P. J.; Chalk, S.; Kingston, H. M. In Microwave-Enhanced Chemistry; Kingston, H. M., Haswell, E. J., Eds.; American Chemical Society, Washington D. C., 1997, 55-222. (7) LeBlanc, G. LC-GC 1999, 17 (6S), S30-S37. (8) Hawthorne, S. B.; Yang, Y.; Miller, D. J. Anal. Chem. 1994, 66 (18), 29122920. (9) Draisci, R.; Marchiafava, C.; Palleschi, L.; Cammarata, P.; Cavalli, S. J. Chromatogr., B 2001, 753, 217-223. (10) Richter, B. E. LC-GC 1999, 17 (6S), S22-S28. (11) Wang, S.; Mei, H.; Ng, K.; Workowski, K.; Astle, T.; Korfmacher, W. Proc. 50th Ann. Conf. Am. Soc. Mass Spectrom. Allied Top. 2002; ThPJ 234. (12) Posyniak, A.; Zmudzki, J.; Semeniuk, S. J. Chromatogr., A 2001, 914, 8994.

Analytical Chemistry, Vol. 76, No. 6, March 15, 2004 1761

Figure 1. Chemical structures of desipramine, fluoxetine, and haloperidol.

digestion as an alternate tissue preparation technique for bioanalysis of drugs in conjunction with LC/MS/MS. Two different enzymes, collagenase and proteinase K, that are known to degrade connective tissues to allow tissue dissolution were chosen for evaluation using well-known antidepressants desipramine and fluoxetine (Figure 1) as test compounds in dog and rat brain tissue. Comparison between enzymatic digestion and conventional homogenization tissue preparation was performed for samples incurred from a rat in vivo study. EXPERIMENTAL SECTION Reagents and Chemicals. Desipramine, fluoxetine, haloperidol (internal standard), proteinase K, TES (tris(hydroxymethyl)methyl-2-aminoethane sulfonate), sodium acetate, ammonium acetate, Trizma hydrochloride, sodium dodecyl sulfate, glycerol, and ethylenediaminetetraacetic acid (EDTA) anhydrous were purchased from Sigma Chemical (St. Louis, MO). Purified collagenase (code: CLSPA) was obtained from Worthington Biochemical Corporation (Lakewood, NJ). Formic acid (88%) was supplied by J. T. Baker (Phillipsberg, NJ) and was diluted to desired concentration using deionized distilled water. Sodium chloride, calcium chloride anhydrous, HPLC-grade water, methanol, and acetonitrile were purchased from Mallinckrodt Baker Inc. (Paris, KY). Compounds A, B, C, and D were new chemical entities (NCE) obtained from Pfizer Global Research & Development (Ann Arbor, MI). Equipment. An American Scientific Products (McGraw Hill, IL) shaking water bath model YB-531 with a temperature control device was used for enzymatic tissue digestion, and a Polytron model PCU 11 homogenizer from Brinkmann Instruments Inc. (Westbury, NY) was used for tissue homogenization. Polypropylene transport mailing tubes (5-mL) from Elkay Products Inc. (Worcester, MA) were used as sample containers during the process of enzymatic tissue digestion. A standard multitube vortex1762

Analytical Chemistry, Vol. 76, No. 6, March 15, 2004

mixer from VWR Scientific Products (West Chester, PA) was used for vortex-mixing, and an Eppendorf centrifuge model 5810R from Brinkmann Instruments Inc. (Westbury, NY) was used for centrifugation. An Applied Biosystems/MDS Sciex (Concord, ON, Canada) model API 3000 triple quadrupole mass spectrometer equipped with a PE series 200 LC pump and autosampler were used for LC/MS/MS analysis. A MetaChem Polaris C18 column (50 × 2.0 mm, 5 µm) from ANSYS Technologies Inc. (Torrace, CA) was used as the analytical column. A Tomtec Quadra 96 model 320 (Hamden, CT) was used for semiautomated sample preparation for protein precipitation procedures. A Harvard Apparatus (South Natick, MA) pump 11 syringe pump with a 500-µL syringe from Hamilton Co. (Reno, NE) was employed for investigation of matrix ionization suppression. Animal Dosing and Sample Collection. Desipramine (3, 10, and 30 mg/kg in saline) and fluoxetine (10, 20, and 40 mg/kg in 4% (v/v) cremophor in saline) were administered intraperitoneally (IP) to 12 Sprague-Dawley rats per dose level, and whole brain samples were collected. In addition, control brain samples from eight rats receiving no desipramine or fluoxetine were obtained. Control female dog brain samples that received no desipramine or fluoxetine were obtained from MPI Research (Mattawan, MI). Tissue samples were stored at -80 °C until analysis. Enzymatic Tissue Digestion and Tissue Homogenization Sample Preparation. Enzymatic digestion conditions for tissue samples using collagenase or proteinase K were optimized. For tissue digestion using collagenase, collagenase was reconstituted in 0.05 M TES buffer with 0.36 mM calcium chloride (pH 7.5) to a concentration of 1 mg/mL. Desipramine and fluoxetine (both at 225 µg/mL) were spiked using Eppendorf micropipets into blank dog or rat brain tissue incubation mixtures, respectively, and placed in 5-mL polypropylene containers, then appropriate amounts of reconstituted collagenase were added to each sample to obtain an activity of 2.1 U/mg. Aliquots were taken at different time points during the digestion process at 37 °C. Digestion reactions were quenched using 0.2 M sodium citrate buffer (pH 5.0) and prepared for LC/MS/MS analysis. For tissue digestion using proteinase K, proteinase K was reconstituted in 10 mM Tris/HCl (pH 7.5) in 50% glycerol with 20 mM calcium chloride to a concentration of 10 mg/mL. Before adding proteinase K, desipramine and fluoxetine were volumetrically transferred into buffer solution containing blank dog or rat brain tissue chunks. Appropriate amounts of reconstituted proteinase K were added to each sample placed in 5-mL polypropylene containers with tissue digestion buffer that consisted of 100 mM Tris/HCl (pH 8.5), 1 mM EDTA, 0.2% sodium dodecyl sulfate, and 200 mM sodium chloride to obtain an activity of 0.099 U/mg. As a result, the final concentration of desipramine and fluoxetine were 1.5 ng/ mL. Aliquots were taken out at different time points during the digestion process at 55 °C and prepared for LC/MS/MS analysis. Compound stability studies were carried out by spiking 563 µg/ mL of desipramine and fluoxetine, respectively, into homogenized blank dog or rat brain tissue that were incubated without enzymes at two different temperatures, 37 and 55 °C, respectively. The incubation mixture was vortex-mixed prior to incubation. Aliquots were taken out at different time points during the incubation and were prepared for LC/MS/MS analysis. In the esterase activity assessment experiments, compounds A, B, and C, respectively,

were added (using compound D as an internal standard) to identically prepared collagenase or proteinase K digestion mixtures described above, giving final concentrations of 1.5 µg/mL as a result. Aliquots were taken at different time points during the digestion process at 37 or 55 °C and prepared for LC/MS/ MS analysis. All of the incubations were carried out in a shaking water bath at 37 or 55 °C, and all samples were vortex-mixed vigorously for 10 min following incubation. To homogenize tissue samples, tissues were weighed, and deionized distilled water was added to reach a concentration of 0.1 g tissue/mL. Tissue samples were homogenized by using the homogenizer described above, and test compounds were volumetrically added and vortex-mixed. For protein precipitation procedures following enzymatic tissue digestion or tissue homogenization, a 100-µL aliquot of each sample was mixed with 200 µL of 0.2 µg/mL of internal standard in acetonitrle, vortex-mixed, and centrifuged at 4000 rpm at 4 °C for 5 min. Supernatants were then transferred to a new 96-well plate using a Quadra 96 system. Analytical Standard Preparation. Stock solutions of desipramine and fluoxetine were prepared as 10 mg/mL, whereas the internal standard, haloperidol, was prepared as 0.5 mg/mL in a 50/50 (v/v) mixture of acetonitrile/methanol. Stock solutions were serially diluted using mobile phase as necessary. Analytical standards used to construct calibration curves were prepared separately for each type of extraction method. For homogenized samples, standards were prepared by spiking known quantities of the standard solutions to the blank brain tissue, and for enzymatic digestion, the standard with the highest concentration was spiked into the blank brain tissue before digestion. Serial dilutions were then carried out to desired concentrations after digestion was completed. Standards were processed in the same manner as samples. Accuracy (% RE) and precision (% CV) of the assay was assessed by analyzing 500, 5000, and 50000 ng/g quality control samples prepared identically to the analytical standards. Samples at each concentration level were analyzed in triplicate over three independent batch runs. Liquid Chromatograhy/Mass Spectrometry. Aliquots (1µL) of tissue extracts were injected directly onto the HPLC column. Chromatographic separation was achieved under isocratic conditions using a mobile phase of a 50/50 (v/v) mixture of acetonitrile/0.1% formic acid with 10 mM ammonium acetate using a MetaChem Polaris C18, 50 × 2.0-mm, 5-µm column. The column was maintained at room temperature, and the flow rate was 0.25 mL/min with a total run time of 2.5 min. Positive ion electrospray tandem mass spectra were recorded using an AB Sciex API 3000 triple quadrupole mass spectrometer equipped with Analyst (version 1.2) operating software. The ionspray voltage was set to 4500 V, and the probe temperature was set at 450 °C. Nitrogen was used as the collision gas, and the nebulizer, curtain, and collision gases were set to 6, 10, and 5, respectively. Multiple reaction monitoring parameters of test compounds were set as described in Table 1. Dwell times were set to 300 ms for each transition. Investigation of Matrix Ionization Suppression. Matrix ionization suppression in rat brain samples prepared by enzymatic digestion was compared with the effect in samples prepared by homogenization using a postcolumn infusion system reported elsewhere13-15 with the following modifications: Desipramine or

Table 1. LC/MS/MS MRM Parameters of Test Compounds compd desipramine fluoxetine haloperidol Pfizer A Pfizer B Pfizer C Pfizer D

ESI MRM transition collision declustering polarity (m/z) energy (eV) potential (eV) + + + + +

267.2 f 208.1 310.2 f 148.4 376.3 f 165.2 467.1 f 151.1 381.0 f 256.0 477.0 f 303.9 408.2 f 91.2

34 14 32 30 -40 -35 35

30 20 26 50 -50 -50 50

fluoxetine at two different concentrations, 10 ng/mL and 1 µg/ mL, respectively, were continuously infused into PEEK tubing between the analytical column and mass spectrometer through a tee using a syringe pump. Either 5 µL of protein precipitation extract of blank rat brain tissue or isocratic mobile phase was injected into the HPLC analytical column. The HPLC pump flow rate was set at 250 µL/min, which is the same flow rate used in sample analysis, while syringe pump flow rate was set at 25 µL/ min. Effluent from the HPLC analytical column was mixed with the infused test compounds and entered the electrospray interface. Recovery Assay. Recovery assays were carried out for both enzymatic digestion and homogenization at four different concentration levels: 2000, 25 000, 75 000, and 100 000 ng/g, respectively. The concentration of standards spiked into blank tissue before enzymatic digestion or homogenization was compared to the concentration of standards spiked into blank tissue after enzymatic digestion or homogenization for recovery calculations. RESULTS AND DISCUSSION Enzyme Selection and Optimization. Tissue dissociation and cell harvesting are two principal applications for enzymes in tissue culture research and cell biology studies. Despite the widespread use of enzymes for these applications over the years, their mechanisms of action for dissociation and harvesting are not wellunderstood. As a result, the choice of one technique over another is often arbitrary and based more on past experience than on an understanding of why the method works and which modifications could lead to even better results. The object of this study was to investigate the feasibility of enzymatic tissue digestion as an alternative sample preparation approach for LC/MS/MS analysis in small-molecule drug discovery. Two different enzymes were used in this study. Collagenase was chosen as an ideal candidate to begin feasibility studies, since it is a protease that targets and degrades the triple-helical native collagen fibers commonly found in connective tissue. Unlike some enzymes, collagenase should provide minimal degradation of most drug candidates. Trypsin and papain are more aggressive enzymes than collagenase; however, both exhibit esterase activity that may degrade some side chains of drug candidates.16,17 Chromatographically purified collagenase was used because it is known to be suitable for experiments requiring (13) Bonfiglio, R.; King, R. C.; Olah, T. V.; Merkle, K. Rapid Commun. Mass Spectrom. 1999, 13, 1175-1185. (14) King, R.; Bonfiglio, R.; Fernandez-Metzler, C.; Miller-Stein, C.; Olah, T. J. Am. Soc. Mass Spectrom. 2000, 11, 942-950. (15) Hsieh, Y.; Chintala, M.; Mei, H.; Agans, J.; Brisson, J.-M.; Ng, K.; Korfmacher, W. A. Rapid Commun. Mass Spectrom. 2001, 15, 2481-2487.

Analytical Chemistry, Vol. 76, No. 6, March 15, 2004

1763

more complete degradation of collagen. Visual inspection of ∼200-mg chunks of dog or rat brain tissue incubations with collagenase (2.1 U/mg) at 37 °C for 22 h resulted in nearly complete tissue dissolution. Proteinase K was the other enzyme chosen as a candidate for these studies. Proteinase K is a serine protease with a broad spectrum of action and the main proteolytic enzyme produced by the fungus Tritirachium album limber.18 Because of its broad substrate specificity, high activity, and its ability to digest native proteins, proteinase K has found considerable use in procedures in which the inactivation and degradation of proteins is required, particularly during the purification of nucleic acids. Proteinase K is very useful in the isolation of highly native, undamaged DNAs or RNAs, since most microbial or mammalian DNases and RNases are rapidly inactivated by the enzyme, particularly in the presence of sodium dodecyl sulfate.19 Proteinase K is an unusually stable enzyme, which perhaps can be attributed to the presence of calcium ions in solution.19 It is well-known and confirmed in our investigation that raising the temperature of the reactions from 37 to 50-60 °C, where proteinase K operates best, can increase the proteinase K activity several-fold. The pH vs activity curve of proteinase K, determined for the hydrolysis of urea-denatured hemoglobin, shows optimal activity in the pH range 7.5-12.0;19 however, the enzyme is normally used in pH range 7.5-9.0. A comparison of proteinase K digestion at different temperatures 37, 45, and 55 °C was carried out upon the basis of visual inspection using blank tissues. Tissue samples were digested more thoroughly as digestion time increased, and digestion was completed in a shorter time period than when collagenase was used. Among the three different temperatures investigated, tissue digestion at 55 °C provided complete digestion in the shortest time period. Nearly complete dissolution of ∼200 mg chunks of dog or rat brain tissue was obtained by digestion with proteinase K (0.099 U/mg) at 55 °C for 5 h. We noted that test compound responses started to decrease after 6 h of digestion, despite the fact that more complete digestion was obtained with longer digestion periods. As a result, digestion at the temperature of 55 °C for a period of 5 h was determined to be the optimal digestion conditions using proteinase K. To evaluate the feasibility of using proteinase K in a temperature range that gives its highest activity, a compound stability test of desipramine and fluoxetine was carried out at 37 and 55 °C. Figure 2 shows the MRM chromatograms of desipramine, fluoxetine, and haloperidol (internal standard). Test compounds remained stable throughout the 25-h incubation period, and there was no difference between responses at 37 and 55 °C. Although collagenase also provided good digestion of tissue, proteinase K was selected for further investigation in quantitative analysis of incurred rat brain tissue due to its higher digestion efficiency and shorter time required to complete digestion. Esterase Activity Evaluation. In the drug discovery process, it is essential to develop and implement the most efficient and (16) Tissue Dissociation Guide; Worthington Biochemical Corporation: Lakewood, NJ, 1999; pp 1-17. (17) http://worthington-biochem.com/ (18) Ebeling, W.; Hennrich, N.; Klockow, M.; Metz, H.; Orth, H. D.; Lang, H. Eur. J. Biochem. 1974, 47, 91-97. (19) Sweeny, P. J.; Walker, J. M. In Methods in Molecular Biology; Burrell, M. M., Ed.; Humana: Totowa, NJ, 1993, Vol. 16, 305-311.

1764 Analytical Chemistry, Vol. 76, No. 6, March 15, 2004

Figure 2. LC/MS/MS MRM chromatograms of desipramine, fluoxetine, and haloperidol.

Figure 3. Relative percent LC/MS/MS response of (A) compound A, (B) compound B, and (C) compound C at different time points in esterase activity assessment of enzymatic tissue digestion. Using ester group containing Pfizer compounds A, B, and C, the LC/MS/ MS responses of test compounds were normalized to the control incubation LC/MS/MS response where no enzymes were added.

reliable methods to screen and evaluate new chemical entities. Ester groups are often encountered as new chemical entities containing ester groups in the early drug discovery process. If the enzyme employed in enzymatic tissue digestion has esterase activity, this would adversely affect quantitation of ester-containing compounds. To evaluate the ability of the enzymatic digestion process to accommodate compounds containing the ester group, three test compounds (compounds A, B, and C) containing ester groups that were known to be unstable in rodent plasma were employed for an esterase activity assessment. This instability had been previously determined by comparing analyte concentration in rat plasma before and after 30 min incubation at 37 °C.

Figure 4. Comparison of average (A) desipramine and (B) fluoxetine concentrations in rat brain tissue extracted by enzymatic digestion or homogenization. Error bars represent standard deviation mean value and are symmetrical about the mean. Table 2. Typical Quantitation Parameters for Rat Brain Tissue Analysis of Desipramine and Fluoxetine Carried out by Employing Proteinase K Enzymatic Digestion or Homogenization enzymatic digestion

homogenization

desipramine fluoxetine desipramine accuracy (% RE) precision (% CV) LLOQ linear range

fluoxetine

(10%, back-calculated standards (20% at LLOQ, (15% at all other concentrations 100 ng/g tissue 100-100 000 ng/g tissue

calibration curve 0.000 018 5 0.000 015 9 0.000 020 4 0.000 009 72 slope calibration curve 0.001 53 0.001 60 0.001 27 0.001 63 intercept r2 0.999 4 0.999 2 0.999 0 0.993 1

Compounds A, B, and C exhibited at least a 50% decrease in concentration after incubation. Compound D was used as the internal standard in this investigation. Relative percent response was generated by normalizing the peak area ratio of the test compounds at each time point to the peak area ratio of control incubations of the test compounds to which no enzymes were added. As shown in Figure 3, the addition of enzymes appeared to increase the stability of these compounds, as compared to control incubations without enzymes. Responses of test compounds in collagenase-added incubations decreased as incubation time increased but showed similar response at time points after 4 h of incubation (75-93% response at 22 h), whereas proteinase K-added incubations showed similar response up to 6 h (86-92% response at 6 h). These results confirmed the validity of the optimal digestion period that was determined to be ∼22 h for collagenase and 5 h for proteinase K. Considering the fact that the investigated test compounds were extremely unstable in plasma, these results indicate that collagenase and proteinase K show no significant esterase activity, and the amount of enzyme

Table 3. Recovery Levels (%) of Desipramine and Fluoxetine in Rat Brain Tissue Analysis Carried Out by Proteinase K Enzymatic Digestion or Homogenization analyte, ng/g tissue

enzymatic digestion

homogenization

desipramine, 2000 desipramine, 25 000 desipramine, 75 000 desipramine, 100 000 fluoxetine, 2000 fluoxetine, 25 000 fluoxetine, 75 000 fluoxetine, 100 000

58.3 60.2 57.2 60.0 101.8 102.1 107.0 149.4

74.6 78.0 86.4 74.5 102.3 105.9 103.9 116.2

present, digestion time, and temperature are more important factors. Enzymatic Tissue Digestion vs Tissue Homogenization. Proteinase K was selected for further investigation due to its higher digestion efficiency and shorter time required to complete digestion, as compared to collagenase. Enzymatic tissue digestion was compared to conventional tissue homogenization by carrying out quantitative analysis of rat brain tissue samples from animals dosed with desipramine (n ) 26) or fluoxetine (n ) 33). Typical quantitation parameters are showed in Table 2. The test compound levels in rat brain tissue were calculated on the basis of the regression equation generated from calibration curves. In terms of quantitative assay performance parameters such as sensitivity and dynamic range, no significant differences were found between enzymatic digestion and homogenization. Recovery levels (percent) of desipramine and fluoxetine in rat brain tissue analysis carried out by employing enzymatic digestion or homogenization are presented in Table 3. Measurements at each concentration level were conducted in triplicate, and the average values are presented. For desipramine, somewhat higher recovery values were obtained for homogenization, as compared to enzymatic digestion. However, in the case of fluoxetine, recoveries from the two techniques were comparable. Analytical Chemistry, Vol. 76, No. 6, March 15, 2004

1765

Figure 5. Correlation of (A) desipramine and (B) fluoxetine concentrations in incurred rat brain tissue extracted by enzymatic digestion and conventional homogenization.

Figure 6. Comparison of effect of injecting blank rat brain tissue protein precipitation extracts prepared by (A) proteinase K enzymatic digestion and (B) conventional homogenization on column with postcolumn infusion of 1 µg/mL desipramine. Positive ion electrospray MRM transition of m/z 267.2 f 208.1 was monitored.

As shown in Figure 4, measured concentration levels of both desipramine and fluoxetine were dose-proportional, and enzymatic tissue digestion has extraction efficiency comparable to or greater than tissue homogenization. The incurred test compound levels were found to be generally higher with enzymatic digestion than the values obtained by the conventional homogenization tissue preparation. The reason for this observation is still unknown, but the role and reaction conditions (i.e., temperature) of the enzymes employed in the extraction process may contribute to this effect. As shown in Figure 5, measured concentrations of desipramine and fluoxetine in rat brain tissue extracted by enzymatic digestion and conventional homogenization, respectively, had good correlation with each other. All of the data generated from both individual extraction methods at each different concentration level passed the normality test, confirming that changes observed in each sample are consistent with a normally distributed population. Since 1766

Analytical Chemistry, Vol. 76, No. 6, March 15, 2004

all of the samples differ in size (milligrams of brain tissue) and analyte localization issues still exist, variability in analysis results from incurred in vivo brain tissue samples is both expected and acceptable. Investigation of Matrix Ionization Suppression. Matrix ionization suppression is a well-accepted concern about assay reliability when LC/MS/MS methods are being developed. There have been reports13-15,20 that matrix ionization suppression appears to be more likely a problem when protein precipitation is employed for sample preparation as compared to other methods, including liquid-liquid extraction or solid-phase extraction methods. Despite this concern, protein precipitation has been widely employed for sample preparation procedure for LC/MS/MS assays in drug discovery due to its speed and simplicity. To investigate and (20) Matuszewski, B. K.; Constanzer, M. L.; Chavez-Eng, C. M. Anal. Chem. 1998, 70, 882-889.

Figure 7. Comparison of effect of injecting blank rat brain tissue protein precipitation extracts prepared by (A) proteinase K enzymatic digestion and (B) conventional homogenization on column with postcolumn infusion of 1 µg/mL fluoxetine. Positive ion electrospray MRM transition of m/z 310.2 f 148.4 was monitored.

compare the matrix ionization suppression for extracts from rat brain tissue processed by enzymatic digestion or homogenization and followed by protein precipitation, the variability of the LC/ MS/MS responses for desipramine and fluoxetine was monitored using the postcolumn infusion system described by Bonfiglio et al.13 Any changes in the LC/MS/MS responses of the test compounds were assumed to be due to matrix ionization suppression from tissue sample extract constituents. The infusion LC/ MS/MS MRM chromatograms obtained through electrospray ionization of test compounds desipramine and fluoxetine using different sample preparation techniques are shown in Figure 6. The decrease in the LC/MS/MS response of test compounds at ∼0.6 min was attributable to the void volume. No significant matrix effects, either ionization suppression or enhancement, were observed at any other retention times in the infusion chromatograms. Matrix effects were not a concern for desipramine and fluoxetine. As shown in Figure 6, the LC/MS/MS responses of both test compounds desipramine and fluoxetine in rat brain tissue protein precipitation extracts prepared by enzymatic digestion or homogenization did not show any difference. CONCLUSIONS The feasibility of enzymatic digestion as an alternate tissue preparation technique for bioanalysis of drugs in conjunction with

LC/MS/MS was evaluated. Our investigation results indicate that enzymatic digestion using collagenase or proteinase K is an appropriate alternative sample preparation method for analysis of tissue samples. Proteinase K provided a faster, more thorough tissue digestion than collagenase. This method is simple and reliable and offers the advantages of unattended sample preparation and low cost. Further investigation in applications of this tissue sample preparation technique, particularly in those with a high extent of connective fibers, such as skin or tumors, is ongoing. ACKNOWLEDGMENT The authors thank Steven Michael and Cathy Knupp at Pfizer Global Research & Development for helpful review and discussion of this manuscript. The authors also thank Taewon Yoon and Youngjun Kim at the University of Illinois at Chicago for helpful discussions and suggestions. Preliminary data of this work was presented at the 51st American Society for Mass Spectrometry (ASMS) Conference on Mass Spectrometry and Allied Topics, Montreal, Quebec, Canada, June 8-12, 2003, Abstract ThPM 257.

Received for review September 14, 2003. Accepted January 13, 2004. AC035077V

Analytical Chemistry, Vol. 76, No. 6, March 15, 2004

1767