A sensitive assay for the quantification of morphine and ... - Springer Link

Anal Bioanal Chem (2011) 400:715–728 DOI 10.1007/s00216-011-4775-z

ORIGINAL PAPER

A sensitive assay for the quantification of morphine and its active metabolites in human plasma and dried blood spots using high-performance liquid chromatography– tandem mass spectrometry Claudia F. Clavijo & Keith L. Hoffman & James J. Thomas & Brendan Carvalho & Larry F. Chu & David R. Drover & Gregory B. Hammer & Uwe Christians & Jeffrey L. Galinkin

Received: 15 September 2010 / Revised: 28 January 2011 / Accepted: 4 February 2011 / Published online: 12 March 2011 # Springer-Verlag 2011

Abstract Opioids such as morphine are the cornerstone of pain treatment. The challenge of measuring the concentrations of morphine and its active metabolites in order to assess human pharmacokinetics and monitor therapeutic drugs in children requires assays with high sensitivity in small blood volumes. We developed and validated a semiautomated LC-MS/MS assay for the simultaneous quantification of morphine and its active metabolites morphine 3β-glucuronide (M3G) and morphine 6β-glucuronide (M6G) in human plasma and in dried blood spots (DBS). Reconstitution in water (DBS only) and addition of a protein precipitation solution containing the internal standards were the only manual steps. Morphine and its metabolites were separated on a Kinetex 2.6-μm PFP analytical column using an acetonitrile/0.1% formic acid

Electronic supplementary material The online version of this article (doi:10.1007/s00216-011-4775-z) contains supplementary material, which is available to authorized users. C. F. Clavijo : K. L. Hoffman : J. J. Thomas : U. Christians : J. L. Galinkin (*) iC42 Clinical Research & Development, Department of Anesthesiology, University of Colorado Denver, 1999 N. Fitzsimons Parkway, Suite 100, Aurora, CO 80045–7503, USA e-mail: [email protected]

gradient. The analytes were detected in the positive multiple reaction mode. In plasma, the assay had the following performance characteristics: range of reliable response of 0.25–1000 ng/mL (r2 >0.99) for morphine, 1– 1,000 ng/mL (r2 >0.99) for M3G, and 2.5–1,000 ng/mL for M6G. In DBS, the assay had a range of reliable response of 1–1,000 ng/mL (r2 >0.99) for morphine and M3G, and of 2.5–1,000 ng/mL for M6G. For inter-day accuracy and precision for morphine, M3G and M6G were within 15% of the nominal values in both plasma and DBS. There was no carryover, ion suppression, or matrix interferences. The assay fulfilled all predefined acceptance criteria, and its sensitivity using DBS samples was adequate for the measurement of pediatric pharmacokinetic samples using a small blood of only 20–50 μL. Keywords Morphine . Morphine 3β-glucoronide . Morphine 6β-glucuronide . LC-MS/MS . Human plasma . Dried blood spots Abbreviations CV% Coefficient of variance DBS Dried blood spots LLOD Lower limit of detection LLOQ Lower limit of quantitation QC Quality control

B. Carvalho : L. F. Chu : D. R. Drover : G. B. Hammer Department of Anesthesia, Stanford University, Stanford, CA 94305, USA J. L. Galinkin The Children’s Hospital Denver, Anschutz Medical Campus, 13123 East 16th Avenue, Aurora, CO 80045, USA

Introduction Morphine is the most commonly used opioid medication to relieve moderate to severe acute and chronic pain in

716

patients of all ages [1, 2]. However, surprisingly little research has been carried out examining the pharmacokinetics of morphine and its metabolites in subjects at the extremes of age. The primary reason for this is a lack of sensitive drug assays that require very low sample volumes. These low-volume assays would make patient and parental consent much more probable. Morphine is a pure opioid receptor agonist that is highly selective for μ-opioid receptors. Opioid-related side effects include respiratory depression, drowsiness, decreased gastrointestinal motility, nausea, vomiting, and alterations of the endocrine and autonomic nervous systems [3, 4]. Following the administration of morphine in high doses, excitatory behavior, allodynia, myoclonus, and seizures have been reported [1]. The use of morphine has increased in past decades, improving pain management and quality of medical treatment especially in cancer patients [5]. Both active metabolites of morphine, morphine 6β-glucoronide (M6G) and morphine 3β-glucoronide (M3G; structures are shown in Electronic supplementary material (ESM) Fig. S1), play a role in pain relief and probably also in the undesired effects of morphine. Even when morphine itself is not detectable, its metabolites can often still be found in plasma [3]. M6G has higher analgesic potency than morphine, but fewer side effects. The potency ratio depends on the route of administration and has been found to be between 3:1 and 300:1 when compared to morphine [6]. This may be explained by M6G’s stronger affinity to μ1- than to μ2opioid receptors, although its binding to a distinct unknown receptor may be responsible [6]. The role of M3G in pain control is not as well understood. Studies have shown that although M3G has no analgesic activity, it evokes a range of dose-dependent excitatory behaviors. It has been hypothesized that M3G may be responsible for several undesired effects attributed to morphine [7] as well as the development of tolerance to morphine [8]. Epidural administration of morphine sulfate produces analgesia with minimal impairment of motor, sensory, or sympathetic function [9]. In recent years, several new morphine formulations have become available or are being developed. We herein report the development, validation, and comparison of specific assays for the simultaneous quantitation of morphine and its metabolites in human plasma and dried blood spots (DBS) using high-performance liquid chromatography (HPLC)–tandem mass spectrometry. This assay is sufficiently sensitive for the detection of very low drug concentrations such as those that follow epidural administration and for pharmacokinetics studies in pediatric patients in whom smaller sample volumes are especially important. The small-volume sampling technique allows for pharmacokinetic studies using only 20–50 μL of whole or

C.F. Clavijo et al.

capillary blood per time point compared with traditional blood sampling techniques which can require as much as 500–1,000 μL of blood.

Materials and methods Chemicals and reagents Solvents and reagents (HPLC grade acetonitrile, methanol, water, formic acid 88%, and zinc sulfate) used for sample preparation and as mobile phases were purchased from Fisher Scientific (Fair Lawn, NJ, USA) and were used without further purification. Morphine and its deuterated internal standard, d3-morphine, were purchased from Isotec, Sigma Aldrich (Miamisburg, OH, USA). M3G, M6G, and the deuterated internal standards d3-M3G and d3-M6G were from Lipomed (Cambridge, MA, USA). Whatman 903 filter paper cards were purchased from Fisher Scientific. Calibrators and quality control samples Stock solutions were purchased at a concentration of 1 mg/mL in methanol. Working solutions for quality control samples and standard curves were prepared by dilution of the stock solutions with human EDTA plasma or whole blood. Human plasma and blood used for assay development and validation purposes were obtained from Bonfils Blood Center (Denver, CO, USA). The use of deidentified blood products for the development, validation, calibration, and quality control of analytical assays was considered “exempt” by the Colorado Multi-institutional Internal Review Board (Aurora, CO, USA). The protein precipitation/internal standard solution (acetonitrile for plasma and methanol/0.2 M ZnSO4, 7:3 (v/v), for DBS containing deuterated internal standards for morphine, M3G and M6G, at a concentration of 10 ng/ mL for plasma and for DBS) was prepared freshly daily before extraction. The expiration time for the protein precipitation solution was set to 12 h, and the remaining solution was discarded thereafter. For plasma, calibration standards and quality control samples were prepared by enriching blank human EDTA plasma with morphine, M3G and M6G. In addition to blank and zero samples, the calibration curve was built using the following concentrations of the three analytes: 0, 0.1, 0.25, 0.5, 1, 2.5, 5, 10, 25, 50, 100, 250, 500, and 1,000 ng/mL. Quality control (QC) samples had the following concentrations: 1, 5, 50, 250, and 750 ng/mL. Calibration curves and quality control samples for DBS analysis were made in human whole blood. The concentration for morphine, M3G and M6G, were the same as

Quantification of morphine and its metabolites in human plasma and dried blood spots

described for plasma (0, 0.1, 0.25, 0.5, 1, 2.5, 5, 10, 25, 50, 100, 250, 500, and 1,000 ng/mL) for calibrators and (1, 5, 50, 250, and 750 ng/mL) for quality control samples. Filter paper cards were spiked with 50 μL of each concentration using a calibrated pipette. Samples were allowed to dry for 3 h. Sample extraction EDTA plasma samples The only manual step during plasma sample extraction was protein precipitation. Six hundred microliters of protein precipitation solution that also contained the three deuterated internal standards (10 ng/mL) was added to 100 μL of human plasma. After vortexing (2.5 min) and centrifuging (4 °C, 13,000×g, 8 min), the supernatant was transferred into glass HPLC vials (Agilent Technologies, Santa Clara, CA, USA). Dried blood spots DBS samples collected on Whatman 903 filter paper cards were punched (6.4-mm diameter, contained 20 μL blood) and reconstituted with 100 μL of HPLC grade water. Six-hundred microliters of the protein precipitation solution, methanol/0.2 M ZnSO4 (7:3, v/v), containing the internal standards at a concentration of 10 ng/mL was added to the sample. Samples were vortexed (2.5 min) and centrifuged (4 °C, 8,000×g, 8 min). After centrifugation, 200 μL of the sample supernatants was transferred into HPLC vials and 200 μL of HPLC grade water was added.

717

acetonitrile to 80% acetonitrile within 2.5 min. The flow rate was 1.0 mL/min and the column kept at room temperature. The Agilent 1100 Series HPLC system was interfaced with the Sciex API5000 tandem quadrupole mass spectrometer using a turbo electrospray source. The mass spectrometer was run in the positive multiple reaction monitoring (MRM) mode. The electrospray source and the mass spectrometer were adjusted to the following parameters (nomenclature as used in the Analyst 1.4.2. software): curtain gas, 40; gas 1, 60; gas 2, 40 with a temperature of 700 °C; CAD gas, 10 (all gases: arbitrary units); ion spray voltage, 5,000 V; declustering potential, 50 V for morphine and 40 V for the metabolites; entrance potential, 10 V for morphine and metabolites; collision cell exit potential, 10 V for morphine and 20 V for the metabolites; and collision energy, 79 V for morphine and 45 V for the metabolites. For morphine, the following ion transition was monitored: m/z=286.2 [M+H]+→152.0, for M3G and M6G m/z=462.2 [M+H]+→286.2; for the internal standard d3 morphine m/z=289.2 [M+H]+→152.1 and for d3 M3G and d3 M6G m/z=465.1 [M+H]+→289.2. The mass spectrometry signal was recorded starting after injection, and the total run time was 3.9 min. After the analysis was completed, peaks were integrated and the results were printed. Morphine, M3G, and M6G signals were corrected based on the internal standard and quantified using the calibration curves that were included in each batch. All calculations were carried out using the Applied Biosystems Analyst Software (version 1.4.2.). Validation procedures

Equipment The supernatants were analyzed using an LC-MS/MS system that consisted of the following components: one G1312A binary pump, one G1379A vacuum degassers, and one G1316A thermostatted column compartment (all Agilent 1100 series, Agilent Technologies) in combination with a CTC/PAL thermostatted autosampler (adjusted to +4 °C, Zwingen, Switzerland). A Sciex API 5000 triplestage quadrupole mass spectrometer was used as detector (Applied Biosystems, Foster City, CA, USA). The HPLC and the mass spectrometer were controlled by Analyst software (version 1.4.2). LC-MS/MS analysis Twenty microliters of the extracted sample was injected onto a Kinetex PFP, 2.6-μm particle size and 100×4.6-mm analytical column (Phenomenex, Torrance, CA, USA). Samples were eluted with a mobile phase of acetonitrile (containing 0.1% formic acid) and 0.1% formic acid in HPLC grade water. The gradient was run from 20%

The assay was validated using human plasma and whole blood samples (for DBS) from healthy volunteers enriched with the analytes following the guidelines for bioanalytical method validation as issued by the FDA Center for Drug Evaluation and Research [10]. Predefined acceptance criteria The performance of the assay was considered acceptable if precision at each concentration was ≤15% for intra-day (coefficient of variance, CV%) and inter-day variability (residual standard deviation in percent as estimated using one-way analysis of variance and “day” as grouping variable). Intra- and inter-day accuracy had to be within ±15% of the nominal values. The only exception was at the lower limit of quantitation (vide infra). Lower limit of detection, LLOQ, and linearity Linearity and the range of reliable response were assessed using the calibration samples (n=6 per concen-

718

tration per day). The lower limit of detection was the lowest morphine, M3G, and M6G concentrations in plasma and DBS that resulted in a peak-to-noise ratio of 4:1. The lower limit of quantitation (LLOQ) was the lowest morphine, M3G, and M6G concentrations in plasma and dried blood spots that consistently resulted in accuracies within ±20% of the nominal concentration and precisions ≤20%. Precision and accuracy Intra-day precision and accuracy were determined by analysis of the morphine, M3G, and M6G QC samples (1, 5, 50, 250, and 750 ng/mL; n=6, analyzed in a single run). Determination of inter-day precision and accuracy was based on the QC samples. Samples were extracted and analyzed on three different days (n=6 per concentration and day). Intra-day precision is reported as CV%, inter-day precision as relative standard deviations in percent (as estimated by one-way analysis of variance with “day” as independent variable), and accuracy in percent of the nominal concentrations. The PASW software (version 18.0, IBM/SPSS, Chicago, IL, USA) was used for the statistical analyses. Absolute recoveries The recoveries were determined by comparing the signals of QC samples (plasma or whole blood on DBS 1, 5, 50, 250, and 750 ng/mL; n=6) that were spiked with morphine, M3G, and M6G and then extracted with the signals of extracted blank samples that were enriched with the same concentrations of morphine, M3G, and M6G. Matrix interferences, ion suppression, and carryover effect Interferences caused by matrix signals were excluded by analysis of blank plasma and DBS collected from ten different individuals. To detect changes in ionization efficiency by coeluting matrix substances, we used a dual strategy. Again, blank human EDTA plasma and DBS samples from ten different healthy individuals were obtained. In the first approach, supernatants obtained from blank samples were enriched with 1, 5, 50, 250, or 750 ng/mL morphine, M3G, and M6G, and signal intensities were compared with those after injection of a corresponding amount of the compounds from a stock solution. The second approach was based on the method described by Müller et al. [11]. The supernatants obtained from blank samples of ten different individuals after extraction from plasma and DBS were injected onto the analytical column as described above, while morphine, M3G, and M6G or the internal standards (10 μg/mL dissolved in 0.1 formic acid/methanol, 1:1, v/v) were

C.F. Clavijo et al.

continuously infused post-column via T-piece at 10 μL/min using a syringe pump (Harvard Apparatus, Holliston, MA, USA). The extent of ion suppression was assessed by monitoring the intensity of the ion currents in MRM-mode for morphine (m/z=286.20 [M+H]+→152.0, for M3G and M6G m/z=462.2 [M+H]+→286.2; for the internal standard deuterated morphine m/z=289.2 [M+H]+→152.1 and for deuterated M3G and M6G m/z=465.1 [M+H]+→289.2) after injection of the blank extracted plasma and dried blood spot samples into the LC-MS/MS system. Potential carryover was assessed by alternately analyzing blood samples enriched with concentrations of morphine, M3G, and M6G at the upper limit of quantitation (1,000 ng/ mL, n=3 each) followed by extracted blank EDTA plasma and dried blood spot samples. Stability studies To test the stability of plasma and DBS samples, QC samples were freshly prepared and one set (n=6 per concentration level) was extracted and analyzed immediately (baseline) for plasma and 3 h later for DBS. The remaining samples were used to establish stability during three freeze–thaw cycles (n=6 per concentration level and cycle). Samples were kept frozen at −80 °C and thawed at room temperature. Benchtop and storage stability were also tested using QC samples. Samples were kept at either −80°C, −20 °C, and +4 °C or at room temperature. Plasma samples were kept for 7 days as well as for 3 and 6 months. Stability of DBS was tested up to 7 days. Samples at room temperature were processed and analyzed after 4, 8, 12, 24, 48, and 168 h (n=3 per time point), and the results were compared with the results of the baseline samples. Extracted sample/autosampler stability was tested by placing extracted QC samples (n=6 per concentration) into the thermostatted autosampler that was adjusted to +4 °C. Samples were injected immediately (baseline) and after 12, 24, and 48 h. In all cases, samples were considered stable when the results were within ±15% of the baseline values.

Results As a first step, MS and MS/MS spectra were recorded after direct infusion of morphine, M3G, and M6G and the deuterated morphine, M3G, and M6G into the electrospray source via a syringe pump (Harvard Scientific, Holliston, MA, USA). Compounds were dissolved at a concentration of 10 μg/mL in methanol/0.1% formic acid 80:20 (v/v) and were delivered at a rate of 10 μL/min. Figure 1 shows the product ion scan spectra of morphine and the metabolites. For the robust and reliable quantification of drugs in complex matrices, it is recommended to monitor more than one ion


719

Intensity [counts per second]

A

m/z


B

m/z

Fig. 1 MS/MS spectra of morphine (a), morphine-3-glucuronide (M3G) (b), and morphin-6-glucuronide(M6G) (c). For morphine, M3G, and M6G, the MS/MS spectra were scanned from m/z=100–1,200. Please note that the spectra in a–c do not line up

transition. However, as obvious from the representative full scan spectra in Fig. 1, MS/MS of morphine (a) and especially of its metabolites (Fig. 1b, c) results in only one ion transition with sufficient sensitivity. There were no other parameters that

resulted in an improvement of the fragmentation patterns. Due to the lack of other sufficiently sensitive ion transitions, we decided to monitor only the major ion transitions and not to sacrifice much needed sensitivity (especially for M6G) by

720

C.F. Clavijo et al.


C

m/z

Fig. 1 (continued)

monitoring additional ion transitions that, unless in the case of very high analyte concentrations, would not result in additional qualification information. The following ion transitions were the predominant transitions and thus were selected for quantification: m/z=286.20 [M+H]+→152.0 for morphine, m/z=462.2 [M+H]+→286.2 for M3G and M6G; m/z=289.2 [M+H]+→152.1 for the internal standard deuterated morphine and m/z=465.1 [M+H]+→289.2 for deuterated M3G and M6G. The absolute recoveries of morphine during the extraction of human plasma ranged from 93.9% to 107.7%. The absolute recoveries of the metabolites M3G and M6G were between 95.6% and 115% and between 95.9% and 120%, respectively. The recovery of morphine extracted from DBS was between 99.6% and 108.3%. For the metabolites, the absolute recoveries during extraction of DBS were between 95.6% and 102% for M3G and between 99.7% and 103.3% for M6G. Both tests confirmed the absence of ion suppression in human plasma and DBS from ten different individuals. Figure 2 shows a representative experiment following the method described by Müller et al. [11] indicating lack of ion suppression/enhancement for morphine. A potential carryover was assessed by analyzing extracted blank EDTA plasma and dried blood spot samples after the highest calibrators (1000 ng/mL). No carryover was detected in any of the blank samples.

In EDTA plasma, the lower limit of detection (LLOD) was 0.1 ng/mL for morphine and 0.5 ng/mL for M3G and M6G. The LLOQ was 0.25 ng/mL for morphine, 1 ng/mL for M3G, and 2.5 ng/mL for M6G. Figure 3 shows ion chromatograms of morphine samples in plasma (blank, at the LLOD, and at the LLOQ) and Fig. 4 ion chromatograms of M3G and M6G at the LLOQ in plasma. The assay was linear from 0.25 to 1,000 ng/mL for morphine, from 1 to 1,000 ng/mL for M3G, and from 2.5 to 1,000 ng/mL for M6G (r2 >0.99, n=6; ESM Table S1). Assay accuracy and precision were determined using five different concentrations of morphine and the metabolites in human plasma (1, 5, 50, 250, and 750 ng/mL). The results for intra-day and inter-day accuracies and precisions are listed in Table 1. In DBS, the lower limit of detection was 0.25 ng/mL for morphine and 0.5 ng/mL for M3G and M6G. The LLOQ was 1 ng/mL for both morphine and M3G and 2.5 ng/mL for M6G (Figs. 3 and 4). The assay was linear from 1 to 1,000 ng/mL for morphine and M3G and from 2.5 to 1,000 ng/mL for M6G (r2 >0.99, n=6; ESM Table S1). Assay accuracy and precision were determined using five different concentrations of morphine (1, 5, 50, 250, and 750 ng/mL) and four different concentrations of the metabolites (5, 50, 250, and 750 ng/mL). The results for intra-day and inter-day accuracies and precisions are listed and compared to those in plasma in Table 1. Interestingly,


721

a

Time (min)

b

Time (min) Fig. 2 Lack of ion suppression. Blank human EDTA plasma (a) and dried blood spot (b) samples from ten different healthy individuals were extracted and injected into the LC-MS/MS system. A representative experiment is shown. Morphine (10 μg/mL dissolved in 0.1 formic acid/methanol, 1:1, v/v) was infused post-column via T-piece at a rate of 10 μL/min using a syringe pump (Harvard Apparatus). The

extent of ion suppression was established by monitoring the signal intensity of the ion currents in MRM mode (m/z = 286.2 [M+H]+→152.0) at the retention times of the analyte (marked by arrow) after injection of blank extracted blood samples into the LCMS/MS system. Ion suppression would have caused a “dip” in the morphine signal produced by constant infusion

722

C.F. Clavijo et al.

Fig. 3 Representative ion chromatogram of blank (a), LLOD (b), and LLOQ (c) of morphine after extraction from plasma samples and blank (d), LLOD (e), and LLOQ (f) of morphine after extraction from dried blood spots


Fig. 3 (continued)

723

724

C.F. Clavijo et al.

assay performances for EDTA plasma and DBS were similar. Morphine, M3G, and M6G extracted from human plasma and DBS were stable for at least 5 days at +4 °C (autosampler stability). Morphine, M6G, and M3G were stable in human plasma or DBS during at least three freeze–thaw cycles. In human plasma, morphine stored on the bench at room temperature was stable for at least 12 h. The metabolites were only stable for