Development and validation of a LC-MS/MS method based on a new 96-well Hybrid-SPE™-precipitation technique for quantification of CYP450 substrates/metabolites in rat plasma
- First Online:
- Cite this article as:
- Ardjomand-Woelkart, K., Kollroser, M., Li, L. et al. Anal Bioanal Chem (2011) 400: 2371. doi:10.1007/s00216-010-4618-3
- 389 Views
A rapid and selective high-throughput HESI-LC-MS/MS method for determining eight cytochrome P450 probe drugs in one-step extraction and single run was developed and validated. The four specific probe substrates midazolam, dextromethorphan, tolbutamide, theophylline and their metabolites 1-hydroxymidazolam, dextrorphan, hydroxyl(methyl)tolbutamide, 1,3-dimethyluric acid, together with the deuterated internal standards, were extracted from rat plasma using a novel 96-well Hybrid-SPE™-precipitation technique. The bioanalytical assay was based on reversed phase liquid chromatography coupled with tandem mass spectrometry in the positive ion mode using selected reaction monitoring for drug (-metabolite) quantification. All analytes were separated simultaneously in a single run that lasted less than 11 min. The intra- and inter-day precisions for all eight substrates/metabolites were 1.62–12.81% and 2.09–13.02%, respectively, and the relative errors (accuracy) for the eight compounds ranged from −9.62% to 7.48% and −13.84% to 8.82%. Hence, the present method provides a robust, fast and reproducible analytical tool for the evaluation of four major drug metabolising cytochrome P450 (3A4, 2C9, 1A2 and 2D6) activities with a cocktail approach in rats to clarify herb–drug interactions. The method can be used as a basic common validated high-throughput analytical assay for in vivo interaction studies.
Inhibition of cytochromes P450 enzyme (CYP) activity by herbal products can significantly increase exposure of co-administrated drugs that are metabolised by the same CYP enzyme. This can result in significant adverse events [1, 2]. Therefore, accurate models for predicting metabolic herb–drug interactions could be useful tools in order to avoid toxic adverse reactions. Most published information about herb–drug interactions is based on unvalidated case reports, reviews, theory, personal opinions or in vitro studies of limited relevance .
To address potential inhibition/induction of CYP enzymes due to various administered herbal drug species, dosages, herbal medicinal products or pure natural products, an assay for high-throughput analyses with a new Hybrid-SPE™-precipitation technique in rat plasma has been developed and validated. The novel Hybrid-SPE™-precipitation technology consists of a simple and generic 96-well plate designed for the gross level removal of endogenous protein and phospholipid interferences from plasma prior to liquid chromatography-mass spectrometry (LC-MS)/MS analysis with a one-step elution technique. So far, the sample clean-up methods for CYP enzyme substrates used as “cocktails” and their metabolites are mostly tedious and usually require separate extraction steps from either plasma or urine, and these generally involve a lengthy extraction procedure [4–7]. Solid-phase extraction (SPE), liquid–liquid extraction and protein precipitation (PPT), which are time consuming and subject to more matrix effects, are the common used methods for the preparation of plasma samples [8–11]. Since the increasing tendency to shorten analytical LC-MS run times and ballistic gradients, the removal of the co-extracted phospholipids of highly ionic nature, represents an extremely important step in the sample preparation process. Moreover, the Hybrid-SPE™-precipitation technique is well suited for simultaneous analysis of analytes possessing diverse polarities. Theophylline is a quite polar analyte while tolbutamide for example is relatively hydrophobic.
Hence, the Hybrid-SPE™-precipitation 96-well technology offers a less time consuming extraction of non-specific targeting compounds while removing phospholipids. So far, only a comparison of Hybrid-SPE™-Precipitation technology and classical protein precipitation has been carried out to evaluate the novel technique . This validated method was applied to an interaction study in rats using an enzyme cocktail with known CYP substrates (CYP3A4/midazolam, CYP2D6/dextromethorphan, CYP2C9/tolbutamide and CYP1A2/theophylline).
Chemicals and reagents
High-performance liquid chromatography
HPLC experiments were performed on a quaternary LC pump (Flux Instruments, Rheos 2000, Reinach, Switzerland) with a vacuum degasser (Flux Instruments, Rheos CPS-LC, CPS-Tune 25–100 bar, Reinach, Switzerland) for solvent delivery. Mobile phase A was 0.1% (v/v) formic acid in H2O and mobile phase B consisted of acetonitrile with 0.1% (v/v) formic acid. Mobile phase A and B were pumped through a SunFire™ C18 column (2.1 × 50 mm I.D., 3.5 μm; Waters Corporation, Milford, MA, USA) at a flow rate of 0.3 mL/min using a gradient from 10% to 100% B in 6 min as shown in the Electronic supplementary material (Fig. S1). The separation was performed at room temperature. Volumes of 10 μL were injected using an autosampler CTC Analytics HTC PAL system thermostated at 6 ± 2 °C. The total run time was 11 min. The autosampler needle was rinsed with methanol and water before and after each injection.
Analytical parameters (HESI/MS/MS) for CYP3A4, CYP2D6, CYP2C9 and CYP1A2 assays
Substrate/metabolite and internal standard (IS)
Concentration range (ng/mL plasma)
QC samples (low, medium and high) (ng/mL plasma)
Retention time (min)
10, 50 and 250
326.09 > 248.9/291.0
4, 20 and 100
342.07 > 202.9/324.0
2.5 ng/10 μL Injectionvolume
331.09 > 213.0/296.1
10, 50 and 250
272.19 > 147.0/215.1
10, 50 and 250
258.17 > 157.0/201.1
2.5 ng/10 μL Injectionvolume
275.20 > 171.0/215.1
4,000, 20,000 and 100,000
271.09 > 154.95
40, 200 and 1,000
287.09 > 89.0/170.9
2.5 ng/10 μL Injectionvolume
296.14 > 89.0/171.0
500, 2,000 and 5,000
181.06 > 96.03/123.9
25, 100 and 250
197.05 > 142.0/169.1
2.5 ng/10 μL Injectionvolume
225.08 > 124.0/181.0
Preparation of stock solutions, calibration samples and quality control samples
Initial stock solutions of mida, 1-OH-mida, dem, dor, tol, OH–CH3–tol and their internal standards (IS) were prepared by dissolving accurately weighed 1 mg in 1 mL methanol, and dissolving 1 mg of theo and 1,3dmu in 1 mL 30% (v/v) ethanol. The two standard mixtures were prepared by dissolution of individual compounds in methanol and 30% (v/v) ethanol to give required final stock concentrations. These standard mixtures were kept frozen at −20 °C. 100 μL of each standard mixture were added to 800 μL blank rat plasma to yield spiked desired calibration standards. On the day of analysis a calibration curve was constructed using 100 μL of each plasma standard. Quality control (QC) samples for determination of accuracy and precision at three concentrations for each calibration range (QCs low, medium and high) were prepared in batches with serial dilution in the same manner as the calibration standards, stored at −20 °C and were stable for at least 3 months. Plasma concentrations for mida, dem, dor and 1,3dmu were 5, 10, 25, 50, 100 and 250 ng/mL. For 1OH-mida, the plasma concentrations were 2, 4, 10, 20, 40 and 100 ng/mL, for theo 100, 500, 1,000, 2,000, 2,500 and 5,000 ng/mL, for tol 2, 4, 10, 20, 40 and 100 μg/mL, and for OH–CH3–tol 20, 40, 100, 200, 400 and 1,000 ng/mL. The investigated ranges of each CYP substrate/metabolite and quality control samples are summarized in Table 1. Calibration curves and QC samples were prepared during validation and each experimental run.
A mixture of internal standard stock solutions was prepared and diluted in acetonitrile/methanol (50:50) +0.5% citric acid to obtain a working solution that was used for sample pre-treatment. This internal standard working solution contained: midazolam-d5, dextromethorphan-d3, hydroxytolbutamide-d9 and β-hydroxyethyltheophylline at concentrations of 25 ng/300 μL.
A newly available Hybrid-SPE™-precipitation (Hybrid-SPE-PPT) 96-well plate (50 mg sorbent/well) from Sigma-Aldrich Co. (Supelco, Bellefonte, PA) was used for extraction of a series of CYP substrates/metabolites from rat plasma. Aliquotes (100 μL) of the blank, calibration standards, QC samples and Unkowns (pharmacokinetic plasma samples) were added to the 96 well Hybrid-SPE-PPT plate followed by 300 μL of a precipitating agent consisting of 0.5% citric acid in acetonitrile/methanol (50:50) including the internal standards (midazolam-d5, dextromethorphan-d3, hydroxytolbutamide-d9 and β-hydroxyethyltheophylline, 25 ng/300 μL). In-well precipitation was performed by vortexing the plate for 1 min. Accordingly the 96-well plate was transferred to a 96-well vacuum manifold (VacMaster-96, Biotage, Uppsala, Sweden) and vacuum (−10 to–15 inHg) was applied for extraction. It took less than 3 min for the precipitated sample to completely pass through the packed bed. The resulting eluent was collected in a 96-well collection plate and evaporated to dryness at 40 °C using a CentriVap concentrator (Enviro-tech, Duesseldorf, Germany). The residue was reconstituted in 100 μL of the mobile phase (acetonitrile/water 10:90), transferred to glass autosampler vials with glass inserts (250 μL) and 10 μL was injected into the chromatographic system for analysis.
Ten male Sprague–Dawley rats (weighing 310 ± 15 g) purchased from Harlan (Indianapolis, IN, USA) were used in this study. The animals were housed in a plastic cage and allowed to adapt to the environment for one week before being used as a control group within a pharmacokinetic herb–drug interaction study. They were maintained on 12/12 h light/dark cycle and received standard chow and water ad libitum during the study. The animal experiment was performed according to the policies and guidelines of the Institutional Animal Care and Use Committee (IACUC) of the University of Florida, Gainesville, FL, USA (NIH publication #85-23).
For the pharmacokinetic study, an enzyme cocktail containing theo (CYP1A2; 10 mg/kg), tol (CYP2C9; 20 mg/kg), dem (CYP2D6; 20 mg/kg) and mida (CYP3A4; 7.5 mg/kg) in 2% Tween 80, 3% Ethanol, 12% PEG and 1.5% Na-benzoat buffer, was administered orally by gavage. Plasma samples (300 μL per blood sample) were collected from the sublingual vein into Vaccuette® LH Li+ heparin tubes (Greiner Bio-One NA., Inc.) prior to dosing and at 15, 30 min, 1, 2, 4, 6, 10 and 24 h. Before blood collection, the rat was anaesthetized with 2% halothane and the blood loss was replaced with 500 μL normal saline. The blood samples were centrifuged for 15 min at 4,000 rpm at 4 °C. The resulting plasma was transferred into tubes and stored at −20 °C until analysis.
Pharmacokinetic parameter estimates were obtained by non-compartmental analysis using Phoenix™ WinNonlin® software package version 6.1 (Pharsight corporation, Mountain View, CA, USA).
A full validation (specificity, linearity, calibration range, precision, accuracy, limit of quantification and limit of detection) following the general principles of the FDA guidelines (US DHHS, FDA, CDER, 2001) was performed for the assay in rat plasma . For analysing the extraction recoveries and freeze–thaw stabilities only two QC levels were used. To each sample, the appropriate internal standard was systematically added with the precipitation agent as described in the experimental part.
Linearity and calibration curves
The calibration curves consisted of six points covering a broad range individual for the different CYP substrates/metabolites (Table 1). The linearity was evaluated by analyses of six independent calibration curve batches obtained over 6 days. The regression equations were calculated by weighted linear regression (weighing factor, 1/x) analysis of peak area ratios (analytes/IS) versus the concentrations using Xcalibur® software (QuanBrowser). For calibration curves, the coefficient of determination R² should be higher than 0.99.
The acceptance criteria were that the coefficient of variation (CV), the percent difference between specified and calculated amount, and accuracy must not exceed 20% for limit of quantitation (LLOQ) and 15% for other standards.
Accuracy and precision
Summary of precision and accuracy from QC samples of rat plasma (n = 6 days; four replicates/day; n = 10 for total)
Spiking plasma concentration (ng/mL)
Intra-assay (precision and accuracy)
Inter-assay (precision and accuracy)
Concentration measured (mean (ng/mL))
RSDa (%; n = 6)
Relative errorb (%)
Concentration measured (mean; ng/mL)
RSDa (%; n = 4)
Relative errorb (%)
Evaluation of process efficiency
Extraction recoveries: mida, 1-OH-mida, dem, dor, tol, OH–CH3–tol, theo and 1,3dmu acid in extracted blank rat plasma vs. spiked in elution solution (three extractions with three replicates, n = 9)
Nominal concentration (ng/mL)
Peak area in processed samplesa (I; mean ± SD)
Peak area in unprocessed samplesb (II; mean ± SD)
Extraction recoveryc (%; I/II; mean ± SD)
3.117.831 ± 179.901
4.288.788 ± 215.130
72.70 ± 5.37
13.267.656 ± 2.403.499
15.780.198 ± 2.051.443
84.08 ± 27.12
1.816.658 ± 201.879
2.623.308 ± 197.029
69.25 ± 10.68
8.929.497 ± 580.326
10.658.595 ± 762.282
83.78 ± 11.26
9.711.920 ± 827.163
9.943.171 ± 371.311
97.67 ± 9.72
39.700.963 ± 1.658.130
38.776.313 ± 1.786.425
102.38 ± 7.22
7.529.771 ± 858.646
8.383.157 ± 218.265
89.82 ± 10.29
31.767.959 ± 2.595.484
35.463.057 ± 4.421.022
89.58 ± 15.57
361.713.637 ± 26.799.955
375.994.893 ± 15.452.420
96.20 ± 6.34
1.328.956.291 ± 27.266.834
1.348.864.568 ± 63.476.754
98.52 ± 4.33
1.782.263 ± 101.775
1.864.014 ± 117.152
95.61 ± 10.36
8.463.362 ± 780.806
9.540.188 ± 740.653
88.71 ± 11.01
58.815.291 ± 3.584.655
80.651.678 ± 3.630.963
72.93 ± 6.07
121.465.196 ± 18.702.255
153.633.588 ± 15.980.528
79.06 ± 19.33
95.374 ± 15.498
1.135.433 ± 112.603
8.40 ± 0.92
202.821 ± 26.909
2.728.881 ± 785.189
7.43 ± 3.09
LLOQ and LOD
The limit of detection (LOD) was defined as the concentration that produced a signal-to noise ratio of 3 (S/N, 3). It was determined after extraction followed by injection of spiked plasma with substrates/metabolites in decreasing concentrations. The LLOQ was determined as the lowest concentration of the analytes in plasma that could be quantified with an inter-assay CV lower than 20% and accuracy between 80% and 120%.
Long- and short-term freezing and thawing stabilities
Stability of substrates/metabolites and internal standards in rat plasma was evaluated by using QC samples at two different concentrations for 6 h at room temperature, during 3 month of storage at −20 °C and after three freeze (−20 °C)/thaw cycles with a minimal interval of 24 h. The quality control samples that had been frozen and thawed three times were quantified to freshly processed standard samples. The re-injection reproducibility or autosampler stability was determined in the final extract after 24 h and 4 days storage period in the autosampler tray at nominally 4–8 °C with the quality control concentrations. Analytes were considered stable in the biological matrix, the rat plasma, when 85–115% of the initial concentration was recovered.
Stability of the deuterated internal standards was assumed to be equal to the corresponding undeuterated analytes.
Results and discussion
The aim of this study was to develop and validate a LC-ESI-MS/MS method for the analysis of eight CYP substrates and metabolites in rat plasma, namely midazolam (mida), dextromethorphan (dem), tolbutamide (tol), theophylline (theo), as well as their major metabolites, 1-hydroxymidazolam (1-OH-mida), dextorphan (dor), hydroxytolbutamide (OH-tol) and 1,3-dimethyl uric acid (1,3dmu), by a novel Hybrid-SPE™-precipitation technique for high-throughput analysis.
For the development of a robust method chromatographic conditions including mobile phase composition and column type had to be optimised. The feasibility of various mixtures of solvents such as acetonitrile and methanol, along with altered flow-gradient programmes and different kind of columns (reversed phase C18 or hydrophilic interaction liquid chromatography HILIC) were tested for complete chromatographic resolution of all substrates and metabolites (data not shown). The assessment resulted in the selection of a SunFire™ C18 column (2.1 × 50 mm I.D., 3.5 μm), since it resulted in the desired chromatographic resolution. The tested direct injection of the organic solvent (acetonitrile:methanol with 0.5% citric acid) after Hybrid-SPE™-PPT elution using a HILIC column was not effective due to the 25% water content from the biological sample. It resulted in a poor peak shape (data not shown).
The molecular ions ([M + H]+) were used as pre-cursor ions to generate product-ion spectra. The most intense product-ions (see Table 1) were optimised and used as SRM transitions to ensure high sensitivity and selectivity. In order to compensate the high concentration range of 2–100 μg/mL plasma for tolbutamide, the transition (m/z) was tuned only for the second intensive signal (see Table 1. transition 271.09 > 154.95). With a standard dose of 20 mg tolbutamide per kilogrammes rat such high amounts (cmax = 91.1 μg/mL) can be expected . In order to compensate the overall method variability, including extraction and ionization variations, deuterated IS of mida, dem and tol were employed. For quantitative analysis of theo and 1,3dmu, ß-hydroxyethyltheophylline has been used.
For plasma extraction, a novel Hybrid-SPE-Precipitation procedure was applied. The goal of this zirconia sorbent SPE-precipitation strategy is directed on a high affinity for phospholipids besides any selectivity toward a wide range of basic, neutral and acidic compounds. Pucci et al. performed a systematic comparison of the simple protein precipitation procedure to the new sample preparation platform Hybrid-SPE-Precipitation and observed a significant improvement in bioanalysis and achieved a feasible and fast way to ensure the avoidance of phospholipid-based matrix effects .
The results showed that there was no endogenous substance interferring with analytes and IS using a SRM function at the retention times. The chromatographic run was performed using a short (50 mm) SunFire column, which is convenient for a high throughput of samples.
Linearity, sensitivity, precision and accuracy
Slope, intercept and coefficient of determination (R²) for the calibration curves for mida, 1-OH-mida, dem, dor, tol, OH–CH3–tol, theo and 1,3dmu calculated by weighted (1/x) linear regression (y = intercept + slope × x) (n = 7)
Concentration range (ng/mL)
R² ± SD
0.9963 ± 0.0037
0.9964 ± 0.0025
0.9964 ± 0.0036
0.9966 ± 0.0023
0.9950 ± 0.0033
0.9975 ± 0.0025
0.9975 ± 0.0021
0.9957 ± 0.0022
Assay performance data (inter-assay accuracy and precision) of all analysed compounds are summarized in Table 2. The intra-batch (n = 6) relative standard deviations and relative errors were less than 12.8%, the inter-assay (n = 4) RSDs and REs were less than 13.8% at each concentration of QC sample tested. These results indicate that the method is reliable and reproducible within its analytical range.
The LLOQ were defined as the lowest concentration of calibration standards. The LODs were ca. 0.5, 0.5, 0.5, 10.0, 0.2, 0.5, 2.0 and 5.0 ng/mL plasma for mida, dem, tol, theo, 1OH-mida, dor, OH-tol and 1,3dmu, respectively.
The stability of all substrates/metabolites in rat plasma during sample storing and processing procedures was evaluated by analysing QC samples at two concentrations. After three cycles (24 h, 48 h and 3 month storage, at −20 °C) the measured concentrations were between 85.7% and 112.8%, indicating no significant substance loss during repeated freezing and thawing.
Autosampler and bench-top stability
During the assay, the QC samples placed at room temperature were stable for at least 6 h. The stability of processed QC samples indicated that substrates/metabolites kept in the autosampler between 4 and 8 °C are stable within 4 days. All the results are given in the Electronic supplementary material (Fig. S2).
The extraction efficiency was experimentally determined from drug-free plasma spiked to contain the analytes at two concentration levels. The absolute extraction recoveries were evaluated by comparing the peak areas of the analytes (substrates/metabolites) obtained from the spiked plasma samples to those obtained from the corresponding post-extraction spiked plasma samples at the same concentrations. Table 3 shows the recoveries of all substrates and metabolites from rat plasma. The recoveries at two concentrations ranged from 69.25 ± 10.68 to 102.38 ± 7.22% with the exception of the recovery of 1,3-dimethyluric acid, which was approximately 8%. For 1,3dmu the Hybrid SPE™-precipitation technique appeared to be not optimal. However, accuracies and precisions were within the predefined acceptation criteria of less than 13.8%, as well for the limit of quantification.
Application to study substrate/metabolite pharmacokinetics in rats
The newly designed analytical method and extraction using the novel Hybrid-SPE™-precipitate plates were approved to accelerate sample processing and clarify whether a range of CYP enzymes have been inhibited or inducted and cause considerable adverse events. The problem of co-administrated over the counter herbal medicinal products with prescribed drugs is still neglected, even for popular herbs like Echinacea species and Ginkgo biloba, and have to be enforced with established in vivo studies using chemically characterized preparations. The basic prerequisite is to use a common validated high-throughput analytical method, which is reported in this article.
Although previous enzyme cocktails have used combinations of several drugs , to the best of our knowledge, this is the first work combining these four probe drugs for these four CYPs and using the new Hybrid-SPE™ technology with one extraction in a single run. The method development led to optimised chromatographic conditions, and validated novel high-throughput extraction and sample preparation procedures, analytical recovery, method precision, storage stability, and method detection limits for analysis of CYP substrates and metabolites. Deuterated IS were used to compensate for the overall method variability, including extraction and ionization variations. The method may play a predominant role in studying herb–drug or drug–drug interactions. The use of novel Hybrid-SPE™-precipitation 96-well plates and HESI/MS/MS allows an accurate, precise and reliable measurement of eight compounds over a wide concentration range with different polarity in rat plasma with a simple one-step extraction technique. The fully validated method is now ready for high-throughput quantification of CYP substrates and metabolites in plasma samples with a less time consuming one-step elution technique.
The Austrian Science Fund (FWF) is thanked for the award of an Erwin-Schroedinger scholarship to K. Ardjomand-Woelkart (J2754-B05).