Oxidation of Selected Phenothiazine Drugs During Sample Preparation: Effects of Varying Extraction Conditions on the Extent of Oxidation

Oxidation of Selected Phenothiazine Drugs During Sample Preparation: Effects of Varying... Abstract Characterization of degradation products formed from selected phenothiazine drugs during standard solid-phase extraction (SPE) approaches is described. An analytical method for promethazine (PMZ), chlorpromazine (CPZ) and their respective N-desmethyl and sulfoxide metabolites in biological samples (bone tissue extract and blood) by ultra performance liquid chromatography-photodiode array detection, using mixed-mode SPE for basic drugs was developed. When ethyl acetate:isopropanol:ammonium hydroxide (80:17:3) was used as the elution solvent during method development, extraneous peaks were observed that were absent in the negative controls. Analysis of extracts of PMZ and CPZ individually showed extraneous peaks, including peaks with retention time and UV spectra suggesting the formation of the sulfoxide metabolites, amongst others. Analytes were then extracted individually and analyzed by ultra performance liquid chromatography-quadrupole time-of-flight mass spectrometry. The results confirmed the oxidation of PMZ to its sulfoxide and N-oxide metabolites and oxidation of CPZ to its sulfoxide metabolite. Oxidation was also observed in analysis of whole blood, and thus was not specific to bone tissue extract. To determine if extraction with minimal oxidation was possible, extractions using SPE with a different elution solvent system (dichloromethane:isopropanol:ammonium hydroxide) and filtration/pass through extraction (FPTE) with and without evaporation were evaluated. The results demonstrated that the sample preparation method highly influenced the extent of oxidation. FPTE without an evaporation step was the only method that did not measurably induce analyte oxidation. Introduction Methods for extraction of basic drugs from postmortem tissue samples are often based on liquid–liquid extraction (LLE) (1, 2) or solid-phase extraction (SPE) (2, 3), in order to provide sufficiently clean extracts so that they may be analyzed by gas chromatography-mass spectrometry (GC/MS) or high-performance liquid chromatography (HPLC) with optical detection, or to reduce matrix effects in methods based on liquid chromatography with tandem mass spectrometry (LC/MS/MS). Work in our laboratory is focused on analysis of drugs and metabolites in skeletal remains. Following a solvent extraction procedure to extract analytes from the bone matrix, SPE-based methods generically designed for basic drugs, with minor variations to improve cleanliness or recovery, are used to assay a range of analytes (4–6). At the onset of this work, we sought to examine the relative distribution of two basic drugs from the same class (phenothiazines) in skeletal tissues, with the intent of investigating the effect of slight structural differences on the relative distribution of each drug and its metabolites across the skeleton. Herein, we present analytical characterization and comparison of four different methods for analysis of selected phenothiazines and their metabolites in bone tissue extract (BTE). First, we describe our initial efforts to develop and characterize a method for the analysis of promethazine, chlorpromazine and their respective N-desmethyl and sulfoxide metabolites in skeletal remains using ultra performance liquid chromatography-photodiode array detection (UPLC-PDA). Throughout the analytical method development and characterization efforts, evidence of analyte degradation during sample preparation was observed. We also describe our characterization of the degradation as oxidation using UPLC-PDA and ultra performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-qTOF-MS), and the influence of modifications to the extraction process on the extent of oxidation. Lastly, we developed and characterized a simple sample preparation method for use in the analysis of promethazine, chlorpromazine and corresponding metabolites that did not induce any measurable oxidation of the analytes. Accordingly, the goal of this work is to illustrate the influence of different sample preparation methods on the oxidative degradation of selected phenothiazine drugs. Materials and Methods Chemicals Promethazine (PMZ), chlorpromazine (CPZ), desmethylpromethazine (DPMZ), desmethylchlorpromazine (DCPZ), promethazine sulfoxide (PMZSO), chlorpromazine sulfoxide (CPZSO), promethazine N-oxide (PMZNO) and promazine (PZ) were purchased from Toronto Research Chemicals (Toronto, ON). The internal standards PMZ-d3 and CPZ-d3 were purchased from Cerilliant (Round Rock, TX, USA). Methanol (MeOH), acetonitrile (ACN) isopropanol (iPrOH), ethyl acetate (EA), and dichloromethane (DCM) used in sample preparation and extraction were HPLC grade, and obtained from EMD chemicals (Gibbstown, NJ) and J.T Baker (Center Valley, PA), respectively. For UPLC-PDA mobile phases, HPLC-grade water was obtained through a Milli-Q water purification system. ACN and formic acid used for UPLC-PDA mobile phase were LC/MS grade and obtained from Fisher Scientific (Ottawa, ON). All solvents used for UPLC-qTOF-MS (ACN, MeOH and water) were LC/MS grade and purchased from OmniSolv (EMD Millipore, Gibbstown, NJ). Leucine enkephalin was used as a reference material and was obtained from Waters Corporation (Milford, MA). Sample preparation Solid-phase extraction The sample matrix used in method development was BTE, prepared by subjecting samples of drug-free decomposed bone to methanolic microwave-assisted extraction (1 g bone/5 mL MeOH), followed by evaporation to dryness and then reconstitution in 5 mL of phosphate buffer (PBS 0.1 M, pH 6). Each standard or calibrator was prepared in 1 mL of BTE, using 500 ng of PZ as internal standard. Lipids and proteins were precipitated by addition of 3 mL of ACN:MeOH (1:1) followed by storage at −20°C overnight. Samples were centrifuged for 10 min (4,000 rpm) and the supernatant was collected and evaporated to 1 mL using a Centrivap® vacuum concentrator (Labconco, Kansas City, MO, USA). Following evaporation, samples were diluted with 2 mL of PBS and then acidified with 100 μL of glacial acetic acid prior to SPE. Samples underwent SPE using Clean Screen XCEL I (130 mg) 48-well plates (United Chemical Technologies, Bristol, PA). Wells were conditioned using sequential additions of 3 mL MeOH, 3 mL water and 3 mL of PBS. After loading samples by gravity, wells were washed with 3 mL of PBS and then 3 mL of 0.1 M acetic acid. Wells were then dried under vacuum (~350 mmHg) for 5 min. After drying, wells were washed with 3 mL of MeOH and then sorbents were dried again under vacuum for 10 min (~350 mmHg). Two different elution solvent systems were used: a solvent mixture-based EA, consisting of EA:iPrOH: NH4OH (80:17:3) or one based on DCM consisting of DCM:iPrOH:NH4OH (80:17:3). Extracts were then evaporated to dryness at 40°C by vacuum centrifugation and reconstituted in 500 μL of Mobile Phase A (0.1% formic acid in 90:10 water:ACN). Samples were centrifuged for 10 min at 13,000×g and then transferred to autosampler vials. Filtration/pass through extraction (FPTE) Samples were prepared in 500 μL of BTE and underwent protein precipitation by addition of 1 mL of (ACN:MeOH, 1:1). Following incubation at −20°C overnight, the samples were centrifuged at 4,000 rpm and then poured directly into wells of the FASt® 96-well plate (United Chemical Technologies, Bristol, PA). After filtration, the samples were collected, evaporated to dryness by vacuum centrifugation, reconstituted in 500 μL of mobile phase and centrifuged for 10 min at 13,000 rpm and then transferred to autosampler vials. FPTE without evaporation Calibrators were prepared in 1 mL of BTE. A 200 μL volume was removed from the each calibrator and transferred to a clean test tube where 800 μL of ACN:H2O (1:1) was added for a final volume of 1 mL. The internal standards were added, the samples were vortexed and then poured directly into the wells of a FASt® 96-well plate. Once the samples were filtered, they were collected and centrifuged for 10 min at 13,000 × g. After centrifugation, the samples were transferred to autosampler vials. SPE of calibrators prepared in blood To prepare the blood calibrators, 250 μL of blood was combined with 700 μL of PBS and 50 μL of analyte stock solution was added for a final volume of 1 mL and a concentration of 1,000 ng/mL. The calibrators then underwent protein precipitation, centrifugation and SPE with ethyl acetate-based elution solution in the same manner as described above for BTE. UPLC-PDA settings and conditions An AcquityTM UPLC equipped with a photodiode array detector (UPLC-PDA; Waters Corp., Milford, MA) or with a quadrupole time-of-flight mass spectrometer (UPLC-QTOF-MS; Waters Corp., Milford, MA) was used for the analysis of extracts. The column used was a Raptor biphenyl column (150 mm × 2.1 mm, 2.7 μm particle diameter; Restek, Bellefonte, PA) with column temperature set to 50°C. A binary gradient elution (A: 0.1% formic acid in 90:10 water:ACN and B: 0.1% formic acid in 90:10 (ACN:water)) was used. The gradient was as follows: 95:5 A:B held for 1 min, followed by a linear increase to 70:30 A:B over 4 min and held for 1 min; followed by a linear increase to 20:80 A:B over 3 min; and finally, a reversion back to 95:5 A:B, for 1 min. The total run time was 10 min with a constant flow rate of 0.300 mL/min and the injection volume was 15 μL. The wavelength range was set from 210 to 400 nm, and for quantitative comparisons, the sulfoxide metabolites were monitored at 240 nm, while the remaining analytes were monitored at 250 nm. Data acquisition was performed using Waters MassLynx software version 4.1. UPLC-qTOF-MS settings and conditions The UPLC was operated under the same chromatographic conditions as described above in the section UPLC-PDA, with the exception that a 2 μL injection volume was used. Mass spectrometry was performed on a Waters Acquity UPLC equipped with a Waters Xevo G2-XS-qTOF-MS (Waters Corp., Milford, MA). Data were acquired in sensitivity mode under positive electrospray ionization with resolution > 20,000 at full width half maximum. The acquisition range was from m/z 50 to 601, using a scan time of 0.1 s. Capillary voltage and cone voltage were set to 0.5 kV and 25 V, respectively. The source temperature was 150°C, the desolvation gas flow was set to 1,000 L/h at a temperature of 500°C and the cone gas was set to 50 L/h. Data acquisition was achieved using MSe mode, with low collision energy set to 6.0 eV, and the high-energy ramp ranged from 10–40 eV. Nitrogen was used as both the drying and nebulizing gas. The collision gas was argon. Verification of calibration of the mass axis from m/z 50 to 601 was conducted daily with 5 mM sodium formate. Leucine enkephalin was used as the lockmass reference compound in positive mode at m/z 278.2641 and infused at a flow rate of 10 μL/min. Compound identification was based on retention time (±0.02 min), mass deviation (± 10 mDa) and appropriate isotope profile. MassLynx® Software (version 4.1) was used for data acquisition and processing. Characterization of analytical performance—UPLC-PDA of SPE extracts Characterization of analytical performance for the different SPE methods included measures of precision and bias (inter-day and intra-day, on at least five different days), linearity, matrix effects, and limits of detection (LOD) and quantitation (LOQ), similar to what is mandated in the guidelines of the Scientific Working Group For Forensic Toxicology (SWGTOX). Standard analyte calibrators were prepared in 1 mL of BTE in triplicate, at concentrations ranging from 0 to 10,000 ng/mL. Standard curves were prepared on at least five different days from extracted calibrators to assess precision, linearity, bias, LOD and LOQ. Characterization of analytical performance—FTPE without evaporation and UPLC-qTOF-MS Calibrators were prepared in 1 mL of BTE at concentrations ranging from 0 to 2,000 ng/mL in triplicate. The concentration dependence of six non-zero concentrations over the concentration range 10–2,000 ng/mL was assessed through quadratic regression (considered acceptable where R2 ≥ 0.99) of peak-area ratios versus concentration. Analytical precision was measured as the coefficient of variation (CV), of triplicate analyses over the assayed concentration range on five different days. Precision was considered to be acceptable when the CV of triplicate measurements was 20% or lower. Bias was determined through blinded analysis of triplicate samples at two different concentrations of each analyte per run. Bias was considered to be acceptable when the measured concentration was within 20% of the theoretical concentration. The LOD for a given analyte was defined as the lowest concentration assayed with S/N ≥3, but not subject to precision acceptability criteria, while the LOQ for a given analyte was defined as the lowest concentration assayed where precision (CV) was ≤ 20%. The matrix effect (ion enhancement or suppression experienced by analytes in an extract) was determined by the post-extraction spike method and was evaluated at three concentrations levels: low (25 ng/mL), mid-range (100 ng/mL) and high (1,000 ng/mL). The response of the analyte in neat solution was compared to the response of the analyte spiked at the same concentration into a blank matrix sample that underwent the sample preparation process. Matrix effects at a given concentration were represented as the percentage increase or decrease in peak area relative to that of the analyte in neat solution. Analyte recovery was defined as the ratio (expressed as a percentage) of analyte peak area for an extracted sample to the peak area of that analyte, spiked at the same nominal concentration, into a drug-free, matrix-matched extract. Autosampler stability—UPLC-PDA and UPLC-qTOF-MS The stability of analytes while on the UPLC-PDA autosampler (maintained at 25°C) was evaluated by repeated injection of the extracted samples at two different concentration levels (100 ng/mL and 2,000 ng/mL, n = 3) for 0, 6, 12, 18, 24, 30, 36 h. Analytes were considered stable if there was no deviation in analyte response in excess of 20% from the response of the corresponding sample at t = 0 h. UPLC-qTOF-MS autosampler stability was also (maintained at 10°C) by repeated injection of extracted analytes analyzed individually at two different concentration levels (100 ng/mL and 1,000 ng/mL per analyte, ni = 3) after 0, 12, 24 and 36 h incubation. Analytes were considered to be stable if there was no deviation in analyte response more than 20% from the response of the corresponding analyte at t = 0 h. Results Characterization of analytical performance—SPE using the EA-based elution solvent system Analytical performance data for UPLC-PDA analyses of samples prepared by SPE using the EA-based elution solvent are summarized in Table I. Response ratios were linear (R2 > 0.99) from 25 ng/mL to 10,000 ng/mL for all analytes. While precision criteria were not met in all cases, the data in Table I indicate that CV values in excess of 20% were observed in four or fewer cases of a total of 90 different sets of triplicate extractions for PMZ and its metabolites, and in seven or fewer cases for CPZ and its metabolites. The precision of analyses across all 90 extractions are summarized in the Supplementary Data. For the extractions that used the EA-based elution solvent system, 91–98% of extractions yielded CV values below 20%, and 62–82% of extractions yielded CV values below 10%. Table I. Summary of analytical performance parameters (LOD, LOQ, precision, linearity, bias)a Analyte  Retention time (min, ± 0.05)  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤ 20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  7.13  10  25  0.9–22.6 [4/90]  0.9990  −0.3−(−101.4) [8/90]  PMZSO  4.32  10  25  0.4–30.0 [2/90]  0.9986  −0.9−(−44.0) [6/90]  DPMZ  6.76  10  25  0.9–18.8 [2/90]  0.9987  0.8−(−18.1) [5/90]  CPZ  8.09  10  25  1.8–22.2 [2/90]  0.9994  0.1−(−23.6) [1/90]  CPZSO  5.32  10  25  0.8–29.2 [7/90]  0.9932  0.5−(−36.5) [6/90]  DCPZ  7.95  10  25  1.7–21.7% [3/90]  0.9988  −0.3−(30.5) [2/90]  Analyte  Retention time (min, ± 0.05)  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤ 20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  7.13  10  25  0.9–22.6 [4/90]  0.9990  −0.3−(−101.4) [8/90]  PMZSO  4.32  10  25  0.4–30.0 [2/90]  0.9986  −0.9−(−44.0) [6/90]  DPMZ  6.76  10  25  0.9–18.8 [2/90]  0.9987  0.8−(−18.1) [5/90]  CPZ  8.09  10  25  1.8–22.2 [2/90]  0.9994  0.1−(−23.6) [1/90]  CPZSO  5.32  10  25  0.8–29.2 [7/90]  0.9932  0.5−(−36.5) [6/90]  DCPZ  7.95  10  25  1.7–21.7% [3/90]  0.9988  −0.3−(30.5) [2/90]  aSPE using the EA-based elution solvent system was used in sample preparation for analysis by UPLC-PDA. Data were collected over nine different sets of extractions of analyte standard mixtures ranging from 25 to 10,000 ng/mL, where each standard concentration was analyzed in triplicate. Accuracy (bias) was assessed through blind analysis of standard samples prepared in BTE at concentrations ranging from 150 to 2,000 ng/mL. Bias was acceptable when the measured concentration deviated from the target concentration by no more than 20%. The data in Table I indicate that absolute bias values in excess of 20% were observed in 6/90 cases or less for PMZ and its metabolites, and in no more than 6/90 cases for CPZ and its metabolites. The stability of the analytes while resident on the autosampler tray was assessed at two different concentrations over a 36 h time period. For all analytes, there was no change in response ratio in excess of 20% of the initial response (t = 0 h), indicating that they remained stable while on the instrument waiting to be run. Appearance and putative identification of extraneous peaks in chromatograms of extracted standards A closer examination of the chromatograms from extracted calibrators revealed extraneous peaks that were not present in the negative control or neat standard mixture. Four minor chromatographic peaks (labeled 1–4) and drug standard peaks were detected in the chromatograms as shown in Figure 1. In analysis of multiple neat standard mixtures, analyte-free samples of Mobile Phase A and samples of extraction reagents that had been evaporated and reconstituted in Mobile Phase A, no extraneous peaks were observed, ruling out contamination of solvents as the source. To assess which, if any, analytes were undergoing chemical degradation during sample preparation, calibrators were extracted from BTE individually (ni = 3). Chromatograms of extracted calibrators initially containing only PMZ or CPZ produced multiple extraneous peaks not found in the drug-free negative control (Figures 2 and 3). The retention time of Peak 2 from Figure 2 and Peak 1 from Figure 3 (4.70 min and 5.32 min, respectively), from each extracted sample corresponded to those of PMZSO and CPZSO, respectively. Further, the UV spectrum of each sulfoxide standard was indistinguishable from the UV spectrum of the corresponding extraneous peak (Figures 2 and 3). These results support the putative identification of the compounds corresponding to Peak 1 (Figure 3) and Peak 2 (Figure 2) as CPZSO and PMZSO, respectively. Figure 1. View largeDownload slide Appearance of extraneous peaks in UPLC-PDA chromatograms of drug-positive extracts from BTE. (A) Neat analyte standard mixture for retention time verification. Extraction was by SPE using the EA-based elution solvent system. (B) Drug-free negative control. (C-E) UPLC-PDA chromatograms (240 nm, 4.50–9.00 min window) of extracted standards (n = 3) at a concentration of 2,000 ng/mL. Extraneous peaks are labeled 1–4. Figure 1. View largeDownload slide Appearance of extraneous peaks in UPLC-PDA chromatograms of drug-positive extracts from BTE. (A) Neat analyte standard mixture for retention time verification. Extraction was by SPE using the EA-based elution solvent system. (B) Drug-free negative control. (C-E) UPLC-PDA chromatograms (240 nm, 4.50–9.00 min window) of extracted standards (n = 3) at a concentration of 2,000 ng/mL. Extraneous peaks are labeled 1–4. Figure 2. View largeDownload slide Comparison of the retention time (4.70 min) and UV spectra of extraneous Peak “2” to the retention time (4.67 min) and spectrum of a PMZSO standard. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C) UPLC-PDA chromatogram (240 nm, 4.5–9.0 min window) of PMZ calibrator prepared at a concentration of 1,000 ng/mL in BTE, and extracted by SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–3. Figure 2. View largeDownload slide Comparison of the retention time (4.70 min) and UV spectra of extraneous Peak “2” to the retention time (4.67 min) and spectrum of a PMZSO standard. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C) UPLC-PDA chromatogram (240 nm, 4.5–9.0 min window) of PMZ calibrator prepared at a concentration of 1,000 ng/mL in BTE, and extracted by SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–3. Figure 3. View largeDownload slide Comparison of the retention time (5.32 min) and UV spectra of extraneous Peak “2” to the retention time (5.29 min) and spectrum of a CPZSO standard. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C) UPLC-PDA chromatogram (240 nm, 4.5–9.0 min window) of CPZ calibrator prepared at a concentration of 1,000 ng/mL in BTE, prepared using SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–4. Figure 3. View largeDownload slide Comparison of the retention time (5.32 min) and UV spectra of extraneous Peak “2” to the retention time (5.29 min) and spectrum of a CPZSO standard. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C) UPLC-PDA chromatogram (240 nm, 4.5–9.0 min window) of CPZ calibrator prepared at a concentration of 1,000 ng/mL in BTE, prepared using SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–4. Confirmation of extraneous products by UPLC-qTOF-MS: phenothiazene oxidation In order to confirm that PMZ and CPZ were being oxidized to their corresponding sulfoxides, high-resolution mass spectral data were acquired. The same set of individually extracted standards was analyzed by UPLC-qTOF-MS using the same column and chromatographic method to obtain accurate mass measurements for the compounds corresponding to the extraneous peaks. These data served as additional parameters for characterization and identification. The extraneous peaks that were observed in the chromatograms from the UPLC-PDA (Figures 2 and 3) were also observed in the total ion chromatograms (TICs) from the UPLC-qTOF-MS (Figure 4). The TIC for the extracted PMZ sample indicated the presence of compounds eluting at 4.71 min, 4.81 min and 7.7 min with measured masses of 317.1356, 301.1355 and 301.1426 Da, respectively (Figure 4C, 4MS-C). The TIC for the extracted CPZ sample indicated the presence of compounds eluting at 5.45 min, 5.82 min, 6.18 min and 8.23 min with measured masses of 335.1017, 351.1020, 376.1273 and 335.1054 Da, respectively (Figure 4D, 4 MS-D). The accurate mass data are summarized in Table II. For both extracted PMZ and CPZ samples, two of the extraneous compounds had the same mass but differed in retention time. If we consider the accurate mass (M) measured for the parent drug molecules (285.1419 Da for PMZ and 319.1090 Da for CPZ), a pattern of M + 16 or M + 32 was observed for the mass of the extraneous compounds, suggesting the occurrence of oxidation (Figure 3, Table II). As expected, comparison of the results with neat reference standards indicated that PMZ was oxidized to PMZSO (Peak 2) and PMZNO (Peak 3) while CPZ was oxidized CPZSO (Peak 1). The PMZNO standard was acquired after the samples had been analyzed to confirm the identity of Peak 3 in Figure 4C and as a result, is not included in the chromatograms for neat standard mixtures shown in any of the figures. Where reference standards were available, the parameters used for compound identification were retention time, accurate mass and fragmentation pattern. Possible candidates for the remaining oxidation products based on accurate mass data include the sulfone, sulfoxide-N-oxide or hydroxylated form of the parent drug. Table II. Summary of accurate mass data for labeled peaks in Figures 4 and 5 Extracted promethazine standard (Figure 4)  Extracted chlorpromazine standard (Figure 5)  Compound  Mass (Da, ± 0.005)  Compound  Mass (Da, ± 0.005)  Peak 1  317.136  Peak 1  335.102  Peak 2  301.136  Peak 2  351.102  Peak 3  301.143  Peak 3  376.127  Promethazine  285.142  Peak 4  335.105  Chlorpromazine  319.109  Extracted promethazine standard (Figure 4)  Extracted chlorpromazine standard (Figure 5)  Compound  Mass (Da, ± 0.005)  Compound  Mass (Da, ± 0.005)  Peak 1  317.136  Peak 1  335.102  Peak 2  301.136  Peak 2  351.102  Peak 3  301.143  Peak 3  376.127  Promethazine  285.142  Peak 4  335.105  Chlorpromazine  319.109  Figure 4. View largeDownload slide Total Ion Chromatograms (TICs) of Extracted PMZ and CPZ Calibrators in BTE Prepared Using SPE with EA-based Elution Solvent System. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C, D) UPLC-qTOF-MS TICs of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). Figure 4. View largeDownload slide Total Ion Chromatograms (TICs) of Extracted PMZ and CPZ Calibrators in BTE Prepared Using SPE with EA-based Elution Solvent System. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C, D) UPLC-qTOF-MS TICs of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). As with BTE, extraneous peaks were observed in the TICs for extracted PMZ and CPZ standard samples when blood was utilized as the sample matrix (Figure 5) that were not present in the analyte-free control. However, there was the presence of a new peak with retention time of 4.26 min and predominant ion with m/z 287.1553 in chromatogram C, and a new peak with retention time 4.94 min and predominant ion with m/z of 321.1142 in chromatogram D. Additionally, the peaks corresponding to PMZ and PMZNO were not detected. Figure 5. View largeDownload slide Appearance of extraneous peaks in extracts of PMZ and CPZ from blood. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms (TICs) of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in blood, prepared using SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding high-resolution mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D, respectively). Figure 5. View largeDownload slide Appearance of extraneous peaks in extracts of PMZ and CPZ from blood. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms (TICs) of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in blood, prepared using SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding high-resolution mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D, respectively). Semi-quantitative comparison of oxidation products—BTE vs. blood Table III summarizes the relative extent of oxidation of the parent compounds (PMZ and CPZ), indicated by increases in the peak-area ratio (Asulfoxide/Aparent analyte). A baseline level of oxidation products were present in the analytical drug standards, but the extent of oxidation, indicated by increases in the peak-area ratio (sulfoxide/parent analyte) increased drastically once the drug standards were subjected to these sample preparation and extraction conditions. Data in Table III showed strong differences in the relative extent of oxidation between extracts from different sample matrices, with the greatest extent of oxidation observed when blood was used as the sample matrix. Table III. A semi-quantitative comparison of the relative extent of formation of PMZSO and CPZSO from the corresponding parent compound, based on the sample matrixa   Promethazine  Chlorpromazine  Matrix  Mean peak-area ratio (APMZSO/APMZ × 100%)  Mean peak-area ratio (ACPZSO/ACPZ × 100%)  Mobile Phase A  3.9  3.9  BTE  70.8  71.3  Blood  10,013  9,947    Promethazine  Chlorpromazine  Matrix  Mean peak-area ratio (APMZSO/APMZ × 100%)  Mean peak-area ratio (ACPZSO/ACPZ × 100%)  Mobile Phase A  3.9  3.9  BTE  70.8  71.3  Blood  10,013  9,947  aMobile Phase A represents neat drug standard prepared in mobile phase and analyzed. The peak-area ratio is the ratio of the area of PMZSO or CPZSO relative to that of the parent analyte originally added (i.e., PMZ or CPZ). The peak-area ratio is represented as a mean percentage (n = 3). Relative extent of phenothiazine oxidation: influence of extraction conditions As described above, PMZ and CPZ were oxidized during sample preparation producing the corresponding sulfoxides as well as other oxidation products. When the analyte standards were extracted from BTE individually by SPE with an EA-based elution solvent system, extraneous peaks were present in the total ion chromatogram (Figure 6). The TIC for the extracted PMZ sample indicated the presence of compounds eluting at 4.71 min (m/z 317.1356, Peak 1), 4.81 min (m/z 301.1355, Peak 2) and 7.74 min (m/z 301.1426, Peak 3) (Figure 6C). Peak 2 has been identified as PMZSO and Peak 3 has been identified as PMZNO and is the most abundant peak and oxidation product formed. The TIC for the extracted CPZ sample indicated the presence of compounds eluting at 5.45 min (335.1017, Peak 1), 5.82 min (351.1020, Peak 2), and 6.18 min (335.1054, Peak 3) and 8.23 min (335.1054, Peak 4) (Figure 6D). CPZSO has been identified as Peak 1 and Peak 4 is the most abundant oxidation product formed and is presumed to be chlorpromazine N-oxide. Table IV. Summary of analytical performance parameters (LOD, LOQ, precision, linearity, bias)a Analyte  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤ 20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  10  25  0.9–18.7 [0/50]  0.9974  0.4–(197.6) [5/20]  PMZSO  10  25  0.8–24.1 [3/50]  0.9917  0.3–(180.3) [7/20]  DPMZ  10  25  0.7–21.9 [2/50]  0.9918  0.4–(55.6) [5/20]  CPZ  10  25  0.9–38.3 [1/50]  0.9960  −0.25–(61.8) [11/20]  CPZSO  10  25  0.9–22.9 [2/50]  0.9923  −1.9–(−306.9) [6/20]  DCPZ  10  25  0.3–20.8 [1/50]  0.9938  −2.6–(−85.4) [12/20]  Analyte  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤ 20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  10  25  0.9–18.7 [0/50]  0.9974  0.4–(197.6) [5/20]  PMZSO  10  25  0.8–24.1 [3/50]  0.9917  0.3–(180.3) [7/20]  DPMZ  10  25  0.7–21.9 [2/50]  0.9918  0.4–(55.6) [5/20]  CPZ  10  25  0.9–38.3 [1/50]  0.9960  −0.25–(61.8) [11/20]  CPZSO  10  25  0.9–22.9 [2/50]  0.9923  −1.9–(−306.9) [6/20]  DCPZ  10  25  0.3–20.8 [1/50]  0.9938  −2.6–(−85.4) [12/20]  aSPE using the DCM-based solution was used in sample preparation for analysis by UPLC-PDA. Data were collected over five different sets of extractions of analyte standard mixtures ranging from 25 to 10,000 ng/mL, where each standard concentration was analyzed in triplicate. Figure 6. View largeDownload slide TICs of extracted PMZ and CPZ calibrators in BTE prepared using SPE with EA-based elution solvent system. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). Figure 6. View largeDownload slide TICs of extracted PMZ and CPZ calibrators in BTE prepared using SPE with EA-based elution solvent system. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). In an effort to minimize analyte oxidation, various sample preparation methods were evaluated. SPE using a different elution solution (80:17:3 DCM:iPrOH:NH4OH), a commonly used solvent system for the elution of basic drugs, was assessed first (Table IV). In this case, the chromatograms of extracts prepared by SPE using the DCM-based elution solution bore extraneous peaks. However, as shown in Figures 6 and 7, it is apparent that the distribution and quantity of the extraneous peaks differs from experiments using SPE with an EA-based elution solvent system. The TIC for the extracted PMZ sample indicated the presence of compounds eluting at 4.83 min (m/z 317.1665, Peak 1), 4.90 min (m/z 301.1355, Peak 2) and 7.82 min (m/z 301.1497, Peak 3) (Figure 7C). PMZSO (Peak 2) and PMZNO were present (Peak 3) in the sample. However, PMZ was the most abundant compound and the sulfoxide is the most abundant oxidation product. The TIC for the extracted CPZ sample indicated the presence of compounds eluting at 5.54 min (m/z 335.1166, Peak 1), 7.61 min (m/z 285.1557, Peak 2). (Figure 7D). The peak corresponding to CPZ was the most abundant compound, and CPZSO was the most abundant oxidation product (Peak 1). When the elution solvent was based on DCM, PZ was formed as a degradation product (Peak 2) but this was not the case when ethyl acetate was utilized. Figure 7. View largeDownload slide TICs of extracted PMZ and CPZ calibrators in BTE prepared using SPE with DCM-based elution solvent system. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–3. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). Figure 7. View largeDownload slide TICs of extracted PMZ and CPZ calibrators in BTE prepared using SPE with DCM-based elution solvent system. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–3. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). The third sample preparation method involved Filtration/pass through extraction (FPTE) instead of SPE. FPTE does not require an elution step, which was presumed to be a factor contributing to the oxidation because of the strong organic nature of the elution solvent and presence of NH4OH. The TIC for the extracted PMZ sample only contained one extra peak at a retention time of 4.89 min (m/z 301.1264) which corresponds to PMZSO (Figure 8C). The TIC for extracted CPZ contained three extra peaks, Peak 1 (5.49 min) corresponds to CPZSO, Peak 2 (7.50 min) corresponds to the formation of PZ, and Peak 3 was a new extraneous peak at 7.73 min with an accurate mass of 315.1455 Da (Figure 8D). Although this method also resulted in analyte oxidation, the peak areas of the oxidation products were reduced. Figure 8. View largeDownload slide TICs of Extracted PMZ and CPZ calibrators in BTE Prepared Using FTPE. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS TICs of extracted PMZ and CPZ standards at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–3. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). Figure 8. View largeDownload slide TICs of Extracted PMZ and CPZ calibrators in BTE Prepared Using FTPE. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS TICs of extracted PMZ and CPZ standards at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–3. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). The final sample preparation method that was assessed was FPTE with no evaporation steps. This approach was a very simple preparation technique with a minimum number of steps that could influence or change the sample. Figure 9 demonstrates that this method did not cause any extraneous peaks to appear in the total ion chromatogram. Therefore, it is reasonable to conclude this method did not induce any measurable oxidation. Figure 9. View largeDownload slide TICs of Extracted PMZ and CPZ Standards in BTE Prepared Using FTPE Without Evaporation. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C–E) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ standards (n = 3) at a concentration of 1,000 ng/mL in BTE Below the TICs are the corresponding accurate mass spectra (MS) for the peaks at 7.76 and 8.35, which correspond to PMZ and CPZ, respectively. Figure 9. View largeDownload slide TICs of Extracted PMZ and CPZ Standards in BTE Prepared Using FTPE Without Evaporation. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C–E) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ standards (n = 3) at a concentration of 1,000 ng/mL in BTE Below the TICs are the corresponding accurate mass spectra (MS) for the peaks at 7.76 and 8.35, which correspond to PMZ and CPZ, respectively. Semi-quantitative comparison of oxidation products—comparing different extraction conditions Table VI compares the relative formation of the oxidized species produced across the different sample preparation methods used. As the method changed, the relative formation of PMZSO changed as indicated by a decrease in the PMZSO-PMZ peak-area ratio as steps were removed from the sample preparation method. FPTE resulted in less oxidation than was observed with SPE and, when the evaporation step was removed, oxidation products were not measurably formed. Furthermore, it is important to note that the deuterated internal standards were also converted to their corresponding oxidized form to the same extent as the analytes in the neat standards (~4%). Analytical characterization of the FPTE (without evaporation) method Analytical characterization of FTPE without evaporation included assessment of concentration dependence, LOD, LOQ, precision, accuracy, recovery and matrix effects in a manner similar to those proposed in the SWGTOX guidelines. All analytes were fit with quadratic regression lines and concentration dependence was assessed over the range 10 to 2,000 ng/mL. Strong correlations (R2 > 0.99) were observed over five different days. The LOD and LOQ were determined to be 10 ng/mL for all analytes. The precision and accuracy data are summarized in Table V. Precision and accuracy were acceptable ranging from 0.04% to 14% and 0.09% to 20%, respectively. Recovery ranged from 90% to 110%, and matrix effects (suppression or enhancement) were less than 25% for all analytes. Analyte stability while resident on the autosampler tray in extracted samples revealed there was no loss in analyte response in excess of 20% of the initial response ratio at t = 0 h for all analytes except for the PMZ-D3 internal standard. At the 24 h time interval, PMZ-D3 demonstrated loss in response exceeding the 20% threshold. However, this instability did not affect the precision or accuracy of the standard curves, this could be because the time required to analyze all curve samples on the instrument does not commonly exceed 15 h. Table V. Summary of analytical performance parameters (LOD, LOQ, precision, linearity, bias)a Analyte  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  10  10  0.04–8.2 [0/34]  0.998  1.4–(−17.4) [0/10]  PMZSO  10  10  0.37–13.9 [0/34]  0.998  −1.4–(18.5) [0/10]  DPMZ  10  10  0.8–10.8 [0/34]  0.998  −1.1–(8.4) [0/10]  CPZ  10  10  0.9–13.79 [0/34]  0.999  0.55–(19.9) [0/10]  CPZSO  10  10  0.7–11.9 [0/34]  0.998  0.66–(13.6) [0/10]  DCPZ  10  10  0.2–12.7 [0/34]  0.998  0.09–(18.1) [0/10]  Analyte  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  10  10  0.04–8.2 [0/34]  0.998  1.4–(−17.4) [0/10]  PMZSO  10  10  0.37–13.9 [0/34]  0.998  −1.4–(18.5) [0/10]  DPMZ  10  10  0.8–10.8 [0/34]  0.998  −1.1–(8.4) [0/10]  CPZ  10  10  0.9–13.79 [0/34]  0.999  0.55–(19.9) [0/10]  CPZSO  10  10  0.7–11.9 [0/34]  0.998  0.66–(13.6) [0/10]  DCPZ  10  10  0.2–12.7 [0/34]  0.998  0.09–(18.1) [0/10]  aFPTE without evaporation was used for sample preparation for analysis by UPLC-QTOF-MS. Data were collected over five different sets of extractions of analyte standard mixtures ranging from 10 to 2,000 ng/mL, where each standard concentration was analyzed in triplicate. Discussion The initial purpose of this research was to develop and validate a semi-quantitative method to evaluate the relative distribution of selected phenothiazine drugs (PMZ and CPZ), and their N-desmethyl and sulfoxide metabolites in skeletal remains. The method was intended for application to studies examining different PMZ and CPZ exposure patterns to understand the significance of drug and metabolite levels in toxicological analysis of bone, and to assess whether their small differences in chemical structure were associated with significant differences in the patterns of drug and metabolite distribution. During the characterization of analytical figures of merit, precision and bias were not consistently meeting the required criteria (≤ 20%). UPLC-PDA chromatograms from extracted standards showed extraneous peaks and further experiments revealed that the analytes were degrading during sample preparation. Consequently, the main objective of the work shifted to the characterization of analyte degradation. The data presented indicate that PMZ and CPZ underwent oxidation during sample preparation and extraction, where PMZ was oxidized to PMZSO and PMZNO while CPZ was CPZSO. Thus, some of the oxidation products included naturally occurring metabolites of the drugs. In addition to the identified oxidation products, other products were formed for which the putative identity include the sulfone, sulfoxide-N-oxide, chlorpromazine N-oxide or hydroxylated form of the parent drug. Oxidation was also observed in analysis of whole blood, and to a greater extent, indicating that the oxidation was not specific to analysis of BTEs, and the extent of oxidation may be expected to vary between different sample matrices. Importantly, the autosampler stability results provide strong evidence that the oxidation took place during the extraction, before the sample was placed on the instrument, and did not occur over time as the samples remained on the autosampler tray. Oxidation of phenothiazines The susceptibility of phenothiazines to oxidation has been reported (7–13), although the majority of such work was done between 1950 and 1990. The most commonly reported oxidation occurs at the sulfide linkage which first forms an unstable radical cation that leads to the generation of the sulfoxide. After the sulfoxide is formed, further oxidation can occur at the sulfur atom which results in the formation of the sulfone (14–16). Various factors have been noted to influence the oxidation reaction including acidity, concentration of oxidizing agents, time, temperature and the side-chain of the molecule (11–13). Much of the research describing phenothiazine oxidation has focused on the generation of oxidation products by chemical, electrochemical, enzymatic and catalytic means (13, 15, 17–19). The oxidation of phenothiazines remains a complex subject and debate about the mechanisms and products formed continues. To the authors’ knowledge, there have been no reports in the analytical toxicology literature describing oxidation of the phenothiazine drugs during the process of preparation of biological samples for analysis. It is possible that the formation of oxidation products may go unnoticed, depending on the sample preparation methods and analytical instrumentation employed. For example, with the use of targeted methods such as GC/MS in SIM mode, or LC/MS/MS in MRM mode, ions corresponding to the oxidation products may not arise in the appropriate time window, or they may be completely excluded from the list of ions or transitions used. For those methods that monitor sulfoxide metabolites, the presence of sulfoxide in any given calibrant or sample chromatogram is expected. Provided that the oxidation reaction occurs in a reproducible and concentration dependent manner, acceptable standard curves could be generated in a given assay. The results of this work suggest that the extent of oxidation is quite reproducible under these sample preparation conditions, as the measured CV values in replicate analyses of all analytes were less than 20% in over 90% of extractions done, and less than 10% in 58–82% (Table I, Supplementary Data). If the extent of oxidation was not reproducible, a wider variability in measured precision and bias values might be expected. This reproducibility represents another reason that analyte oxidation during sample preparation might go unnoticed. Effect of matrix on phenothiazine oxidation The data presented here suggest that the extent of oxidation and the products formed may vary between sample matrices, as shown in comparing data from samples prepared in BTE (Figure 4) to those prepared in blood (Figure 5). The phenothiazines showed a greater extent of oxidation in the blood than in BTE which may be due to the hemoglobin present in red blood cells. This is consistent with studies reporting the oxidation of sulfide functionalities due to various reactive molecules present in blood (20, 21). Furthermore, the authors noticed that CPZ was converted to CPZSO in whole blood, but much of the conversion was due to the set-up of the analytical procedure. Given the wide variation in the nature of sample matrices in postmortem toxicology, it is likely that the extent of phenothiazine oxidation could differ from sample to sample, due to variations in the nature of the matrix. While most bias estimations from the extractions performed were within acceptable limits, this may have been because the same blank matrix was used in preparation of calibrants and the positive control samples used in measurement of accuracy, resulting in a similar extent of oxidation between them. In casework, the matrix used for preparation of standard curves necessarily differs from that of a given sample, and the extent of oxidation may be expected to also differ. Hence, measured phenothiazine and metabolite concentrations may potentially be inaccurate, interfering with toxicological interpretation of the results. The data shown here suggest that all phenothiazine analytes (including drugs and metabolites) should have corresponding deuterated internal standards to minimize the likelihood of erroneous results, presuming that a given analyte and its deuterated analog would oxidize to the same extent. Effect of extraction conditions on phenothiazine oxidation The data presented here show that the analytes underwent oxidation during sample preparation, with oxidation products including naturally occurring metabolites of the drugs being assayed. In some cases, the oxidation resulted in unacceptable bias and precision. In efforts to redesign the method to minimize analyte oxidation, the first area that was investigated was the elution step in the SPE process. Experiments were done where PMZ and CPZ standards were dissolved directly in the EA-based elution solvent and then evaporated to dryness (i.e., with no extraction step from biological matrix), and the resulting data indicated extensive oxidation (see Supplementary Figures 1 and 2 in Supplementary Data). We theorized that the analyte oxidation was due to the specific elution solvent system used, and then evaluated the use of the DCM-based elution solvent system. As shown in Figures 6 and 7 and Table VI, the relative extent of oxidation, as well as the number of products formed, was higher when using the EA-based elution solvent system. Numerous experiments then assessed the effects of reducing the evaporation temperature from 70°C to 40°C, eliminating exposure to light, and evaporation under argon gas instead of using vacuum centrifugation. None of these alterations eliminated oxidation (see Supplementary Data). The subsequent sample preparation method that was employed removed the elution step completely by utilizing FPTE in place of SPE. The results were promising (Figure 8 and Table VI), with reduced oxidation of PMZ and CPZ. Table VI. A semi-quantitative comparison of the relative extent of formation of PMZSO and CPZSO from the corresponding parent compound, based on the sample preparation methoda   Promethazine  Chlorpromazine  Method  Mean peak-area ratio (APMZSO/APMZ × 100%)  Mean peak-area ratio (ACPZSO/ACPZ × 100%)  SPE EA elution solution  67.0  89.2  SPE DCM elution solution  38.9  48.9  Filtration with evaporation  16.3  13.8  Filtration without evaporation  3.6  3.0  Unextracted standards  3.9  3.6    Promethazine  Chlorpromazine  Method  Mean peak-area ratio (APMZSO/APMZ × 100%)  Mean peak-area ratio (ACPZSO/ACPZ × 100%)  SPE EA elution solution  67.0  89.2  SPE DCM elution solution  38.9  48.9  Filtration with evaporation  16.3  13.8  Filtration without evaporation  3.6  3.0  Unextracted standards  3.9  3.6  aThe unextracted drug standard is prepared in Mobile Phase A and analyzed. The peak-area ratio is the ratio of the area of PMZSO or CPZSO relative to that of the parent analyte originally added (i.e., PMZ or CPZ). The peak-area ratio is represented as a mean percentage (n = 3). Through comparison of the relative level of oxidation products in neat standards with those formed in these various experiments, it was hypothesized that the oxidation process was driven largely by the evaporation step. Thus, the FPTE method without evaporation was developed. By removing the evaporation step, no additional oxidation was induced, and the amount of oxidation products formed was indistinguishable from those in the unextracted neat standard (Figure 9, Table VI). Comparison of the results from the four sample preparation methods shows that sample preparation conditions heavily influenced the extent of oxidation, specifically the type and relative formation of the oxidized species produced. Conclusions We have demonstrated that oxidation of selected phenothiazines occurred during standard preparation of samples for analysis by methods typically used in analytical toxicology laboratories. The oxidation products observed included common metabolites of the parent drug and may confound toxicological interpretation. The incidence of oxidation may not be detected by certain analytical configurations. These results are of particular importance for laboratories employing tandem MS methods for analysis of phenothiazines based on MRM. Also, the variability in the extent of oxidation between different samples and calibrators may yield erroneous results. Our work also established the influence of changing the sample preparation method on the extent of oxidation. A new simple extraction method was developed and characterized for the analysis of phenothiazines in skeletal tissues that did not measurably generate any oxidation products. Supplementary Data Supplementary data are available at Journal of Analytical Toxicology online. Acknowledgments The authors would like to thank the Natural Sciences Engineering Research Council of Canada for their financial support of this work. References 1 Paterson, S., Cordero, R., Burlinson, S. ( 2004) Screening and semi-quantitative analysis of post mortem blood for basic drugs using gas chromatography/ion trap mass spectrometry. Journal of Chromatography B , 813, 323– 330. Google Scholar CrossRef Search ADS   2 Maurer, H.H. ( 2005) Multi-analyte procedures for screening for and quantification of drugs in blood, plasma, or serum by liquid chromatography-single stage or tandem mass spectrometry (LC-MS or LC-MS/MS) relevant to clinical and forensic toxicology. Clinical Biochemistry , 38, 310– 318. Google Scholar CrossRef Search ADS PubMed  3 Pirola, R., Mundo, E., Bellod, L., Bareggi, S.R. ( 2002) Simultaneous determination of clomipramine and its desmethyl and hydroxyl metabolites in plasma of patients by high-performance liquid chromatography after solid-phase extraction. Journal of Chromatography B , 772, 205– 210. Google Scholar CrossRef Search ADS   4 Watterson, J.H., Imfeld, A.B., Cornthwaite, H.,.C. ( 2014) Determination of colchicine and O-demethylated metabolites in decomposed skeletal tissues by microwave assisted extraction, microplate solid phase extraction and ultra-high performance liquid chromatography (MAE-MPSPE-UHPLC). Journal of Chromatography B , 960, 145– 150. Google Scholar CrossRef Search ADS   5 Fraser, C.D., Corthnwaite, H.M., Watterson, J.H. ( 2014) Analysis of dextromethorphan and dextrorphan in decomposed skeletal tissues by microwave assisted extraction, microplate solid-phase extraction and gas chromatography-mass spectrometry (MAE-MPSPE-GCMS). Drug Testing and Analysis , 7, 708– 713. Google Scholar CrossRef Search ADS PubMed  6 Watterson, J.H., Donohue, J.P. ( 2011) Relative distribution of Ketamine and Norketamine in skeletal tissues following various periods of decomposition. Journal of Analytical Toxicology , 35, 452– 458. Google Scholar CrossRef Search ADS PubMed  7 Davidson, A.G. ( 1978) The determination of sulphoxide in degraded phenothiazine formulations by difference spectrophotometry. Journal of Pharmacy and Pharmacology , 30, 410– 414. Google Scholar CrossRef Search ADS PubMed  8 Chignell, C.F., Motten, A.G., Buettner, G.R. ( 1985) Photoinduced free radicals from chlorpromazine and related phenothiazines: relationship to phenothiazine-induced photosensitization. Environmental Health Perspectives , 64, 103– 110. Google Scholar CrossRef Search ADS PubMed  9 Bahnemann, D., Asmus, K.D., Willson, R.L. ( 1983) Free radical induced one-electron oxidation of the phenothiazines chlorpromazine and promethazine. Journal of the Chemical Society, Perkin Transactions , 2, 1661– 1668. Google Scholar CrossRef Search ADS   10 Underberg, W.J.M. ( 1978) Oxidative degradation of pharmaceutically important phenothiaiznes I: isolation and identification of oxidation products of promethazine. Journal of Pharmaceutical Sciences , 67, 1128– 1131. Google Scholar CrossRef Search ADS PubMed  11 Karpinska, J., Starczewska, B., Puzanowska-Tarasiewicz, H. ( 1996) Analytical properties of 2- and 10-disubstitutied phenothiazine derivatives. Analytical Sciences : the International Journal of the Japan Society for Analytical Chemistry , 12, 161– 170. Google Scholar CrossRef Search ADS   12 Pankratov, A.N., Uchaeva, I.M., Stepanov, A.N. ( 1993) Chemical and electrochemical oxidation of phenothiazine. Canadian Journal of Chemistry , 71, 674– 677. Google Scholar CrossRef Search ADS   13 Kojlo, A., Karpinska, J., Kuzmicka, L., Misiuk, W., et al.  . ( 2001) Analytical study of the reaction of phenothiazines with some oxidants, metal ions, and organic substances (review article). Journal of Trace and Microprobe Techniques , 19, 45– 70. Google Scholar CrossRef Search ADS   14 Takahashi, D.M. ( 1980) Rapid determination of chlorpromazine hydrochloride and two oxidation products in various pharmaceutical samples using high-performance liquid chromatography and fluorescence detection. Journal of Pharmaceutical Sciences , 69, 184– 187. Google Scholar CrossRef Search ADS PubMed  15 Bosch, E., Kochi, J.K. ( 1995) Catality oxidation of chlorpromazine and related phenothiazines. Cation radicals as the reactive intermediates in sulfoxide formation. Journal of the Chemical Society, Perkin Transactions , 1, 1057– 1064. Google Scholar CrossRef Search ADS   16 Cheng, H.Y., Sackett, P.H., McCreery, R.L. ( 1978) Kinetics of chlorpromazine cation radical decomposition in aqueous buffers. Journal of the American Chemical Society , 100, 962– 967. Google Scholar CrossRef Search ADS   17 Wisniewska, J., Rzesnicki, R., Topolski, A. ( 2011) A mechanistic study on the disproportionation and oxidative degradation of phenothiazine derivatives by manganese(III) complexes in phosphate acidic media. Transition Metal Chemistry , 36, 767– 774. Google Scholar CrossRef Search ADS   18 Blankert, B., Haven, H., van Leeuwen, S.M., Karst, U., et al.  . ( 2005) Electrochemical, chemical and enzymatic oxidations of phenothiazines. Electroanalytical , 17, 1501– 1510. Google Scholar CrossRef Search ADS   19 Gasco, M.R., Carlotti, M.E. ( 1978) Oxidation of promazine and promethazine by ferric perchlorate. Journal of Pharmaceutical Sciences , 67, 168– 171. Google Scholar CrossRef Search ADS PubMed  20 McKay, G., et al.  . ( 1985) Therapeutic monitoring of chlorpromazine II: pitfalls in whole blood analysis. Therapeutic Drug Monitoring , 7, 472– 477. Google Scholar CrossRef Search ADS PubMed  21 Hawes, E.M., et al.  . ( 1986) Therapeutic monitoring of chlorpromazine III: minimal interconversion between chlorpromazine and metabolites in human blood. Therapeutic Drug Monitoring , 8, 37– 41. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Analytical Toxicology Oxford University Press

Oxidation of Selected Phenothiazine Drugs During Sample Preparation: Effects of Varying Extraction Conditions on the Extent of Oxidation

Loading next page...
 
/lp/ou_press/oxidation-of-selected-phenothiazine-drugs-during-sample-preparation-jDRMwljCHJ
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0146-4760
eISSN
1945-2403
D.O.I.
10.1093/jat/bkx067
Publisher site
See Article on Publisher Site

Abstract

Abstract Characterization of degradation products formed from selected phenothiazine drugs during standard solid-phase extraction (SPE) approaches is described. An analytical method for promethazine (PMZ), chlorpromazine (CPZ) and their respective N-desmethyl and sulfoxide metabolites in biological samples (bone tissue extract and blood) by ultra performance liquid chromatography-photodiode array detection, using mixed-mode SPE for basic drugs was developed. When ethyl acetate:isopropanol:ammonium hydroxide (80:17:3) was used as the elution solvent during method development, extraneous peaks were observed that were absent in the negative controls. Analysis of extracts of PMZ and CPZ individually showed extraneous peaks, including peaks with retention time and UV spectra suggesting the formation of the sulfoxide metabolites, amongst others. Analytes were then extracted individually and analyzed by ultra performance liquid chromatography-quadrupole time-of-flight mass spectrometry. The results confirmed the oxidation of PMZ to its sulfoxide and N-oxide metabolites and oxidation of CPZ to its sulfoxide metabolite. Oxidation was also observed in analysis of whole blood, and thus was not specific to bone tissue extract. To determine if extraction with minimal oxidation was possible, extractions using SPE with a different elution solvent system (dichloromethane:isopropanol:ammonium hydroxide) and filtration/pass through extraction (FPTE) with and without evaporation were evaluated. The results demonstrated that the sample preparation method highly influenced the extent of oxidation. FPTE without an evaporation step was the only method that did not measurably induce analyte oxidation. Introduction Methods for extraction of basic drugs from postmortem tissue samples are often based on liquid–liquid extraction (LLE) (1, 2) or solid-phase extraction (SPE) (2, 3), in order to provide sufficiently clean extracts so that they may be analyzed by gas chromatography-mass spectrometry (GC/MS) or high-performance liquid chromatography (HPLC) with optical detection, or to reduce matrix effects in methods based on liquid chromatography with tandem mass spectrometry (LC/MS/MS). Work in our laboratory is focused on analysis of drugs and metabolites in skeletal remains. Following a solvent extraction procedure to extract analytes from the bone matrix, SPE-based methods generically designed for basic drugs, with minor variations to improve cleanliness or recovery, are used to assay a range of analytes (4–6). At the onset of this work, we sought to examine the relative distribution of two basic drugs from the same class (phenothiazines) in skeletal tissues, with the intent of investigating the effect of slight structural differences on the relative distribution of each drug and its metabolites across the skeleton. Herein, we present analytical characterization and comparison of four different methods for analysis of selected phenothiazines and their metabolites in bone tissue extract (BTE). First, we describe our initial efforts to develop and characterize a method for the analysis of promethazine, chlorpromazine and their respective N-desmethyl and sulfoxide metabolites in skeletal remains using ultra performance liquid chromatography-photodiode array detection (UPLC-PDA). Throughout the analytical method development and characterization efforts, evidence of analyte degradation during sample preparation was observed. We also describe our characterization of the degradation as oxidation using UPLC-PDA and ultra performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-qTOF-MS), and the influence of modifications to the extraction process on the extent of oxidation. Lastly, we developed and characterized a simple sample preparation method for use in the analysis of promethazine, chlorpromazine and corresponding metabolites that did not induce any measurable oxidation of the analytes. Accordingly, the goal of this work is to illustrate the influence of different sample preparation methods on the oxidative degradation of selected phenothiazine drugs. Materials and Methods Chemicals Promethazine (PMZ), chlorpromazine (CPZ), desmethylpromethazine (DPMZ), desmethylchlorpromazine (DCPZ), promethazine sulfoxide (PMZSO), chlorpromazine sulfoxide (CPZSO), promethazine N-oxide (PMZNO) and promazine (PZ) were purchased from Toronto Research Chemicals (Toronto, ON). The internal standards PMZ-d3 and CPZ-d3 were purchased from Cerilliant (Round Rock, TX, USA). Methanol (MeOH), acetonitrile (ACN) isopropanol (iPrOH), ethyl acetate (EA), and dichloromethane (DCM) used in sample preparation and extraction were HPLC grade, and obtained from EMD chemicals (Gibbstown, NJ) and J.T Baker (Center Valley, PA), respectively. For UPLC-PDA mobile phases, HPLC-grade water was obtained through a Milli-Q water purification system. ACN and formic acid used for UPLC-PDA mobile phase were LC/MS grade and obtained from Fisher Scientific (Ottawa, ON). All solvents used for UPLC-qTOF-MS (ACN, MeOH and water) were LC/MS grade and purchased from OmniSolv (EMD Millipore, Gibbstown, NJ). Leucine enkephalin was used as a reference material and was obtained from Waters Corporation (Milford, MA). Sample preparation Solid-phase extraction The sample matrix used in method development was BTE, prepared by subjecting samples of drug-free decomposed bone to methanolic microwave-assisted extraction (1 g bone/5 mL MeOH), followed by evaporation to dryness and then reconstitution in 5 mL of phosphate buffer (PBS 0.1 M, pH 6). Each standard or calibrator was prepared in 1 mL of BTE, using 500 ng of PZ as internal standard. Lipids and proteins were precipitated by addition of 3 mL of ACN:MeOH (1:1) followed by storage at −20°C overnight. Samples were centrifuged for 10 min (4,000 rpm) and the supernatant was collected and evaporated to 1 mL using a Centrivap® vacuum concentrator (Labconco, Kansas City, MO, USA). Following evaporation, samples were diluted with 2 mL of PBS and then acidified with 100 μL of glacial acetic acid prior to SPE. Samples underwent SPE using Clean Screen XCEL I (130 mg) 48-well plates (United Chemical Technologies, Bristol, PA). Wells were conditioned using sequential additions of 3 mL MeOH, 3 mL water and 3 mL of PBS. After loading samples by gravity, wells were washed with 3 mL of PBS and then 3 mL of 0.1 M acetic acid. Wells were then dried under vacuum (~350 mmHg) for 5 min. After drying, wells were washed with 3 mL of MeOH and then sorbents were dried again under vacuum for 10 min (~350 mmHg). Two different elution solvent systems were used: a solvent mixture-based EA, consisting of EA:iPrOH: NH4OH (80:17:3) or one based on DCM consisting of DCM:iPrOH:NH4OH (80:17:3). Extracts were then evaporated to dryness at 40°C by vacuum centrifugation and reconstituted in 500 μL of Mobile Phase A (0.1% formic acid in 90:10 water:ACN). Samples were centrifuged for 10 min at 13,000×g and then transferred to autosampler vials. Filtration/pass through extraction (FPTE) Samples were prepared in 500 μL of BTE and underwent protein precipitation by addition of 1 mL of (ACN:MeOH, 1:1). Following incubation at −20°C overnight, the samples were centrifuged at 4,000 rpm and then poured directly into wells of the FASt® 96-well plate (United Chemical Technologies, Bristol, PA). After filtration, the samples were collected, evaporated to dryness by vacuum centrifugation, reconstituted in 500 μL of mobile phase and centrifuged for 10 min at 13,000 rpm and then transferred to autosampler vials. FPTE without evaporation Calibrators were prepared in 1 mL of BTE. A 200 μL volume was removed from the each calibrator and transferred to a clean test tube where 800 μL of ACN:H2O (1:1) was added for a final volume of 1 mL. The internal standards were added, the samples were vortexed and then poured directly into the wells of a FASt® 96-well plate. Once the samples were filtered, they were collected and centrifuged for 10 min at 13,000 × g. After centrifugation, the samples were transferred to autosampler vials. SPE of calibrators prepared in blood To prepare the blood calibrators, 250 μL of blood was combined with 700 μL of PBS and 50 μL of analyte stock solution was added for a final volume of 1 mL and a concentration of 1,000 ng/mL. The calibrators then underwent protein precipitation, centrifugation and SPE with ethyl acetate-based elution solution in the same manner as described above for BTE. UPLC-PDA settings and conditions An AcquityTM UPLC equipped with a photodiode array detector (UPLC-PDA; Waters Corp., Milford, MA) or with a quadrupole time-of-flight mass spectrometer (UPLC-QTOF-MS; Waters Corp., Milford, MA) was used for the analysis of extracts. The column used was a Raptor biphenyl column (150 mm × 2.1 mm, 2.7 μm particle diameter; Restek, Bellefonte, PA) with column temperature set to 50°C. A binary gradient elution (A: 0.1% formic acid in 90:10 water:ACN and B: 0.1% formic acid in 90:10 (ACN:water)) was used. The gradient was as follows: 95:5 A:B held for 1 min, followed by a linear increase to 70:30 A:B over 4 min and held for 1 min; followed by a linear increase to 20:80 A:B over 3 min; and finally, a reversion back to 95:5 A:B, for 1 min. The total run time was 10 min with a constant flow rate of 0.300 mL/min and the injection volume was 15 μL. The wavelength range was set from 210 to 400 nm, and for quantitative comparisons, the sulfoxide metabolites were monitored at 240 nm, while the remaining analytes were monitored at 250 nm. Data acquisition was performed using Waters MassLynx software version 4.1. UPLC-qTOF-MS settings and conditions The UPLC was operated under the same chromatographic conditions as described above in the section UPLC-PDA, with the exception that a 2 μL injection volume was used. Mass spectrometry was performed on a Waters Acquity UPLC equipped with a Waters Xevo G2-XS-qTOF-MS (Waters Corp., Milford, MA). Data were acquired in sensitivity mode under positive electrospray ionization with resolution > 20,000 at full width half maximum. The acquisition range was from m/z 50 to 601, using a scan time of 0.1 s. Capillary voltage and cone voltage were set to 0.5 kV and 25 V, respectively. The source temperature was 150°C, the desolvation gas flow was set to 1,000 L/h at a temperature of 500°C and the cone gas was set to 50 L/h. Data acquisition was achieved using MSe mode, with low collision energy set to 6.0 eV, and the high-energy ramp ranged from 10–40 eV. Nitrogen was used as both the drying and nebulizing gas. The collision gas was argon. Verification of calibration of the mass axis from m/z 50 to 601 was conducted daily with 5 mM sodium formate. Leucine enkephalin was used as the lockmass reference compound in positive mode at m/z 278.2641 and infused at a flow rate of 10 μL/min. Compound identification was based on retention time (±0.02 min), mass deviation (± 10 mDa) and appropriate isotope profile. MassLynx® Software (version 4.1) was used for data acquisition and processing. Characterization of analytical performance—UPLC-PDA of SPE extracts Characterization of analytical performance for the different SPE methods included measures of precision and bias (inter-day and intra-day, on at least five different days), linearity, matrix effects, and limits of detection (LOD) and quantitation (LOQ), similar to what is mandated in the guidelines of the Scientific Working Group For Forensic Toxicology (SWGTOX). Standard analyte calibrators were prepared in 1 mL of BTE in triplicate, at concentrations ranging from 0 to 10,000 ng/mL. Standard curves were prepared on at least five different days from extracted calibrators to assess precision, linearity, bias, LOD and LOQ. Characterization of analytical performance—FTPE without evaporation and UPLC-qTOF-MS Calibrators were prepared in 1 mL of BTE at concentrations ranging from 0 to 2,000 ng/mL in triplicate. The concentration dependence of six non-zero concentrations over the concentration range 10–2,000 ng/mL was assessed through quadratic regression (considered acceptable where R2 ≥ 0.99) of peak-area ratios versus concentration. Analytical precision was measured as the coefficient of variation (CV), of triplicate analyses over the assayed concentration range on five different days. Precision was considered to be acceptable when the CV of triplicate measurements was 20% or lower. Bias was determined through blinded analysis of triplicate samples at two different concentrations of each analyte per run. Bias was considered to be acceptable when the measured concentration was within 20% of the theoretical concentration. The LOD for a given analyte was defined as the lowest concentration assayed with S/N ≥3, but not subject to precision acceptability criteria, while the LOQ for a given analyte was defined as the lowest concentration assayed where precision (CV) was ≤ 20%. The matrix effect (ion enhancement or suppression experienced by analytes in an extract) was determined by the post-extraction spike method and was evaluated at three concentrations levels: low (25 ng/mL), mid-range (100 ng/mL) and high (1,000 ng/mL). The response of the analyte in neat solution was compared to the response of the analyte spiked at the same concentration into a blank matrix sample that underwent the sample preparation process. Matrix effects at a given concentration were represented as the percentage increase or decrease in peak area relative to that of the analyte in neat solution. Analyte recovery was defined as the ratio (expressed as a percentage) of analyte peak area for an extracted sample to the peak area of that analyte, spiked at the same nominal concentration, into a drug-free, matrix-matched extract. Autosampler stability—UPLC-PDA and UPLC-qTOF-MS The stability of analytes while on the UPLC-PDA autosampler (maintained at 25°C) was evaluated by repeated injection of the extracted samples at two different concentration levels (100 ng/mL and 2,000 ng/mL, n = 3) for 0, 6, 12, 18, 24, 30, 36 h. Analytes were considered stable if there was no deviation in analyte response in excess of 20% from the response of the corresponding sample at t = 0 h. UPLC-qTOF-MS autosampler stability was also (maintained at 10°C) by repeated injection of extracted analytes analyzed individually at two different concentration levels (100 ng/mL and 1,000 ng/mL per analyte, ni = 3) after 0, 12, 24 and 36 h incubation. Analytes were considered to be stable if there was no deviation in analyte response more than 20% from the response of the corresponding analyte at t = 0 h. Results Characterization of analytical performance—SPE using the EA-based elution solvent system Analytical performance data for UPLC-PDA analyses of samples prepared by SPE using the EA-based elution solvent are summarized in Table I. Response ratios were linear (R2 > 0.99) from 25 ng/mL to 10,000 ng/mL for all analytes. While precision criteria were not met in all cases, the data in Table I indicate that CV values in excess of 20% were observed in four or fewer cases of a total of 90 different sets of triplicate extractions for PMZ and its metabolites, and in seven or fewer cases for CPZ and its metabolites. The precision of analyses across all 90 extractions are summarized in the Supplementary Data. For the extractions that used the EA-based elution solvent system, 91–98% of extractions yielded CV values below 20%, and 62–82% of extractions yielded CV values below 10%. Table I. Summary of analytical performance parameters (LOD, LOQ, precision, linearity, bias)a Analyte  Retention time (min, ± 0.05)  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤ 20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  7.13  10  25  0.9–22.6 [4/90]  0.9990  −0.3−(−101.4) [8/90]  PMZSO  4.32  10  25  0.4–30.0 [2/90]  0.9986  −0.9−(−44.0) [6/90]  DPMZ  6.76  10  25  0.9–18.8 [2/90]  0.9987  0.8−(−18.1) [5/90]  CPZ  8.09  10  25  1.8–22.2 [2/90]  0.9994  0.1−(−23.6) [1/90]  CPZSO  5.32  10  25  0.8–29.2 [7/90]  0.9932  0.5−(−36.5) [6/90]  DCPZ  7.95  10  25  1.7–21.7% [3/90]  0.9988  −0.3−(30.5) [2/90]  Analyte  Retention time (min, ± 0.05)  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤ 20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  7.13  10  25  0.9–22.6 [4/90]  0.9990  −0.3−(−101.4) [8/90]  PMZSO  4.32  10  25  0.4–30.0 [2/90]  0.9986  −0.9−(−44.0) [6/90]  DPMZ  6.76  10  25  0.9–18.8 [2/90]  0.9987  0.8−(−18.1) [5/90]  CPZ  8.09  10  25  1.8–22.2 [2/90]  0.9994  0.1−(−23.6) [1/90]  CPZSO  5.32  10  25  0.8–29.2 [7/90]  0.9932  0.5−(−36.5) [6/90]  DCPZ  7.95  10  25  1.7–21.7% [3/90]  0.9988  −0.3−(30.5) [2/90]  aSPE using the EA-based elution solvent system was used in sample preparation for analysis by UPLC-PDA. Data were collected over nine different sets of extractions of analyte standard mixtures ranging from 25 to 10,000 ng/mL, where each standard concentration was analyzed in triplicate. Accuracy (bias) was assessed through blind analysis of standard samples prepared in BTE at concentrations ranging from 150 to 2,000 ng/mL. Bias was acceptable when the measured concentration deviated from the target concentration by no more than 20%. The data in Table I indicate that absolute bias values in excess of 20% were observed in 6/90 cases or less for PMZ and its metabolites, and in no more than 6/90 cases for CPZ and its metabolites. The stability of the analytes while resident on the autosampler tray was assessed at two different concentrations over a 36 h time period. For all analytes, there was no change in response ratio in excess of 20% of the initial response (t = 0 h), indicating that they remained stable while on the instrument waiting to be run. Appearance and putative identification of extraneous peaks in chromatograms of extracted standards A closer examination of the chromatograms from extracted calibrators revealed extraneous peaks that were not present in the negative control or neat standard mixture. Four minor chromatographic peaks (labeled 1–4) and drug standard peaks were detected in the chromatograms as shown in Figure 1. In analysis of multiple neat standard mixtures, analyte-free samples of Mobile Phase A and samples of extraction reagents that had been evaporated and reconstituted in Mobile Phase A, no extraneous peaks were observed, ruling out contamination of solvents as the source. To assess which, if any, analytes were undergoing chemical degradation during sample preparation, calibrators were extracted from BTE individually (ni = 3). Chromatograms of extracted calibrators initially containing only PMZ or CPZ produced multiple extraneous peaks not found in the drug-free negative control (Figures 2 and 3). The retention time of Peak 2 from Figure 2 and Peak 1 from Figure 3 (4.70 min and 5.32 min, respectively), from each extracted sample corresponded to those of PMZSO and CPZSO, respectively. Further, the UV spectrum of each sulfoxide standard was indistinguishable from the UV spectrum of the corresponding extraneous peak (Figures 2 and 3). These results support the putative identification of the compounds corresponding to Peak 1 (Figure 3) and Peak 2 (Figure 2) as CPZSO and PMZSO, respectively. Figure 1. View largeDownload slide Appearance of extraneous peaks in UPLC-PDA chromatograms of drug-positive extracts from BTE. (A) Neat analyte standard mixture for retention time verification. Extraction was by SPE using the EA-based elution solvent system. (B) Drug-free negative control. (C-E) UPLC-PDA chromatograms (240 nm, 4.50–9.00 min window) of extracted standards (n = 3) at a concentration of 2,000 ng/mL. Extraneous peaks are labeled 1–4. Figure 1. View largeDownload slide Appearance of extraneous peaks in UPLC-PDA chromatograms of drug-positive extracts from BTE. (A) Neat analyte standard mixture for retention time verification. Extraction was by SPE using the EA-based elution solvent system. (B) Drug-free negative control. (C-E) UPLC-PDA chromatograms (240 nm, 4.50–9.00 min window) of extracted standards (n = 3) at a concentration of 2,000 ng/mL. Extraneous peaks are labeled 1–4. Figure 2. View largeDownload slide Comparison of the retention time (4.70 min) and UV spectra of extraneous Peak “2” to the retention time (4.67 min) and spectrum of a PMZSO standard. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C) UPLC-PDA chromatogram (240 nm, 4.5–9.0 min window) of PMZ calibrator prepared at a concentration of 1,000 ng/mL in BTE, and extracted by SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–3. Figure 2. View largeDownload slide Comparison of the retention time (4.70 min) and UV spectra of extraneous Peak “2” to the retention time (4.67 min) and spectrum of a PMZSO standard. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C) UPLC-PDA chromatogram (240 nm, 4.5–9.0 min window) of PMZ calibrator prepared at a concentration of 1,000 ng/mL in BTE, and extracted by SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–3. Figure 3. View largeDownload slide Comparison of the retention time (5.32 min) and UV spectra of extraneous Peak “2” to the retention time (5.29 min) and spectrum of a CPZSO standard. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C) UPLC-PDA chromatogram (240 nm, 4.5–9.0 min window) of CPZ calibrator prepared at a concentration of 1,000 ng/mL in BTE, prepared using SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–4. Figure 3. View largeDownload slide Comparison of the retention time (5.32 min) and UV spectra of extraneous Peak “2” to the retention time (5.29 min) and spectrum of a CPZSO standard. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C) UPLC-PDA chromatogram (240 nm, 4.5–9.0 min window) of CPZ calibrator prepared at a concentration of 1,000 ng/mL in BTE, prepared using SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–4. Confirmation of extraneous products by UPLC-qTOF-MS: phenothiazene oxidation In order to confirm that PMZ and CPZ were being oxidized to their corresponding sulfoxides, high-resolution mass spectral data were acquired. The same set of individually extracted standards was analyzed by UPLC-qTOF-MS using the same column and chromatographic method to obtain accurate mass measurements for the compounds corresponding to the extraneous peaks. These data served as additional parameters for characterization and identification. The extraneous peaks that were observed in the chromatograms from the UPLC-PDA (Figures 2 and 3) were also observed in the total ion chromatograms (TICs) from the UPLC-qTOF-MS (Figure 4). The TIC for the extracted PMZ sample indicated the presence of compounds eluting at 4.71 min, 4.81 min and 7.7 min with measured masses of 317.1356, 301.1355 and 301.1426 Da, respectively (Figure 4C, 4MS-C). The TIC for the extracted CPZ sample indicated the presence of compounds eluting at 5.45 min, 5.82 min, 6.18 min and 8.23 min with measured masses of 335.1017, 351.1020, 376.1273 and 335.1054 Da, respectively (Figure 4D, 4 MS-D). The accurate mass data are summarized in Table II. For both extracted PMZ and CPZ samples, two of the extraneous compounds had the same mass but differed in retention time. If we consider the accurate mass (M) measured for the parent drug molecules (285.1419 Da for PMZ and 319.1090 Da for CPZ), a pattern of M + 16 or M + 32 was observed for the mass of the extraneous compounds, suggesting the occurrence of oxidation (Figure 3, Table II). As expected, comparison of the results with neat reference standards indicated that PMZ was oxidized to PMZSO (Peak 2) and PMZNO (Peak 3) while CPZ was oxidized CPZSO (Peak 1). The PMZNO standard was acquired after the samples had been analyzed to confirm the identity of Peak 3 in Figure 4C and as a result, is not included in the chromatograms for neat standard mixtures shown in any of the figures. Where reference standards were available, the parameters used for compound identification were retention time, accurate mass and fragmentation pattern. Possible candidates for the remaining oxidation products based on accurate mass data include the sulfone, sulfoxide-N-oxide or hydroxylated form of the parent drug. Table II. Summary of accurate mass data for labeled peaks in Figures 4 and 5 Extracted promethazine standard (Figure 4)  Extracted chlorpromazine standard (Figure 5)  Compound  Mass (Da, ± 0.005)  Compound  Mass (Da, ± 0.005)  Peak 1  317.136  Peak 1  335.102  Peak 2  301.136  Peak 2  351.102  Peak 3  301.143  Peak 3  376.127  Promethazine  285.142  Peak 4  335.105  Chlorpromazine  319.109  Extracted promethazine standard (Figure 4)  Extracted chlorpromazine standard (Figure 5)  Compound  Mass (Da, ± 0.005)  Compound  Mass (Da, ± 0.005)  Peak 1  317.136  Peak 1  335.102  Peak 2  301.136  Peak 2  351.102  Peak 3  301.143  Peak 3  376.127  Promethazine  285.142  Peak 4  335.105  Chlorpromazine  319.109  Figure 4. View largeDownload slide Total Ion Chromatograms (TICs) of Extracted PMZ and CPZ Calibrators in BTE Prepared Using SPE with EA-based Elution Solvent System. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C, D) UPLC-qTOF-MS TICs of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). Figure 4. View largeDownload slide Total Ion Chromatograms (TICs) of Extracted PMZ and CPZ Calibrators in BTE Prepared Using SPE with EA-based Elution Solvent System. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C, D) UPLC-qTOF-MS TICs of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). As with BTE, extraneous peaks were observed in the TICs for extracted PMZ and CPZ standard samples when blood was utilized as the sample matrix (Figure 5) that were not present in the analyte-free control. However, there was the presence of a new peak with retention time of 4.26 min and predominant ion with m/z 287.1553 in chromatogram C, and a new peak with retention time 4.94 min and predominant ion with m/z of 321.1142 in chromatogram D. Additionally, the peaks corresponding to PMZ and PMZNO were not detected. Figure 5. View largeDownload slide Appearance of extraneous peaks in extracts of PMZ and CPZ from blood. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms (TICs) of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in blood, prepared using SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding high-resolution mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D, respectively). Figure 5. View largeDownload slide Appearance of extraneous peaks in extracts of PMZ and CPZ from blood. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms (TICs) of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in blood, prepared using SPE with EA-based elution solvent system. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding high-resolution mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D, respectively). Semi-quantitative comparison of oxidation products—BTE vs. blood Table III summarizes the relative extent of oxidation of the parent compounds (PMZ and CPZ), indicated by increases in the peak-area ratio (Asulfoxide/Aparent analyte). A baseline level of oxidation products were present in the analytical drug standards, but the extent of oxidation, indicated by increases in the peak-area ratio (sulfoxide/parent analyte) increased drastically once the drug standards were subjected to these sample preparation and extraction conditions. Data in Table III showed strong differences in the relative extent of oxidation between extracts from different sample matrices, with the greatest extent of oxidation observed when blood was used as the sample matrix. Table III. A semi-quantitative comparison of the relative extent of formation of PMZSO and CPZSO from the corresponding parent compound, based on the sample matrixa   Promethazine  Chlorpromazine  Matrix  Mean peak-area ratio (APMZSO/APMZ × 100%)  Mean peak-area ratio (ACPZSO/ACPZ × 100%)  Mobile Phase A  3.9  3.9  BTE  70.8  71.3  Blood  10,013  9,947    Promethazine  Chlorpromazine  Matrix  Mean peak-area ratio (APMZSO/APMZ × 100%)  Mean peak-area ratio (ACPZSO/ACPZ × 100%)  Mobile Phase A  3.9  3.9  BTE  70.8  71.3  Blood  10,013  9,947  aMobile Phase A represents neat drug standard prepared in mobile phase and analyzed. The peak-area ratio is the ratio of the area of PMZSO or CPZSO relative to that of the parent analyte originally added (i.e., PMZ or CPZ). The peak-area ratio is represented as a mean percentage (n = 3). Relative extent of phenothiazine oxidation: influence of extraction conditions As described above, PMZ and CPZ were oxidized during sample preparation producing the corresponding sulfoxides as well as other oxidation products. When the analyte standards were extracted from BTE individually by SPE with an EA-based elution solvent system, extraneous peaks were present in the total ion chromatogram (Figure 6). The TIC for the extracted PMZ sample indicated the presence of compounds eluting at 4.71 min (m/z 317.1356, Peak 1), 4.81 min (m/z 301.1355, Peak 2) and 7.74 min (m/z 301.1426, Peak 3) (Figure 6C). Peak 2 has been identified as PMZSO and Peak 3 has been identified as PMZNO and is the most abundant peak and oxidation product formed. The TIC for the extracted CPZ sample indicated the presence of compounds eluting at 5.45 min (335.1017, Peak 1), 5.82 min (351.1020, Peak 2), and 6.18 min (335.1054, Peak 3) and 8.23 min (335.1054, Peak 4) (Figure 6D). CPZSO has been identified as Peak 1 and Peak 4 is the most abundant oxidation product formed and is presumed to be chlorpromazine N-oxide. Table IV. Summary of analytical performance parameters (LOD, LOQ, precision, linearity, bias)a Analyte  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤ 20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  10  25  0.9–18.7 [0/50]  0.9974  0.4–(197.6) [5/20]  PMZSO  10  25  0.8–24.1 [3/50]  0.9917  0.3–(180.3) [7/20]  DPMZ  10  25  0.7–21.9 [2/50]  0.9918  0.4–(55.6) [5/20]  CPZ  10  25  0.9–38.3 [1/50]  0.9960  −0.25–(61.8) [11/20]  CPZSO  10  25  0.9–22.9 [2/50]  0.9923  −1.9–(−306.9) [6/20]  DCPZ  10  25  0.3–20.8 [1/50]  0.9938  −2.6–(−85.4) [12/20]  Analyte  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤ 20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  10  25  0.9–18.7 [0/50]  0.9974  0.4–(197.6) [5/20]  PMZSO  10  25  0.8–24.1 [3/50]  0.9917  0.3–(180.3) [7/20]  DPMZ  10  25  0.7–21.9 [2/50]  0.9918  0.4–(55.6) [5/20]  CPZ  10  25  0.9–38.3 [1/50]  0.9960  −0.25–(61.8) [11/20]  CPZSO  10  25  0.9–22.9 [2/50]  0.9923  −1.9–(−306.9) [6/20]  DCPZ  10  25  0.3–20.8 [1/50]  0.9938  −2.6–(−85.4) [12/20]  aSPE using the DCM-based solution was used in sample preparation for analysis by UPLC-PDA. Data were collected over five different sets of extractions of analyte standard mixtures ranging from 25 to 10,000 ng/mL, where each standard concentration was analyzed in triplicate. Figure 6. View largeDownload slide TICs of extracted PMZ and CPZ calibrators in BTE prepared using SPE with EA-based elution solvent system. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). Figure 6. View largeDownload slide TICs of extracted PMZ and CPZ calibrators in BTE prepared using SPE with EA-based elution solvent system. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–4. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). In an effort to minimize analyte oxidation, various sample preparation methods were evaluated. SPE using a different elution solution (80:17:3 DCM:iPrOH:NH4OH), a commonly used solvent system for the elution of basic drugs, was assessed first (Table IV). In this case, the chromatograms of extracts prepared by SPE using the DCM-based elution solution bore extraneous peaks. However, as shown in Figures 6 and 7, it is apparent that the distribution and quantity of the extraneous peaks differs from experiments using SPE with an EA-based elution solvent system. The TIC for the extracted PMZ sample indicated the presence of compounds eluting at 4.83 min (m/z 317.1665, Peak 1), 4.90 min (m/z 301.1355, Peak 2) and 7.82 min (m/z 301.1497, Peak 3) (Figure 7C). PMZSO (Peak 2) and PMZNO were present (Peak 3) in the sample. However, PMZ was the most abundant compound and the sulfoxide is the most abundant oxidation product. The TIC for the extracted CPZ sample indicated the presence of compounds eluting at 5.54 min (m/z 335.1166, Peak 1), 7.61 min (m/z 285.1557, Peak 2). (Figure 7D). The peak corresponding to CPZ was the most abundant compound, and CPZSO was the most abundant oxidation product (Peak 1). When the elution solvent was based on DCM, PZ was formed as a degradation product (Peak 2) but this was not the case when ethyl acetate was utilized. Figure 7. View largeDownload slide TICs of extracted PMZ and CPZ calibrators in BTE prepared using SPE with DCM-based elution solvent system. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–3. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). Figure 7. View largeDownload slide TICs of extracted PMZ and CPZ calibrators in BTE prepared using SPE with DCM-based elution solvent system. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ calibrators at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–3. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). The third sample preparation method involved Filtration/pass through extraction (FPTE) instead of SPE. FPTE does not require an elution step, which was presumed to be a factor contributing to the oxidation because of the strong organic nature of the elution solvent and presence of NH4OH. The TIC for the extracted PMZ sample only contained one extra peak at a retention time of 4.89 min (m/z 301.1264) which corresponds to PMZSO (Figure 8C). The TIC for extracted CPZ contained three extra peaks, Peak 1 (5.49 min) corresponds to CPZSO, Peak 2 (7.50 min) corresponds to the formation of PZ, and Peak 3 was a new extraneous peak at 7.73 min with an accurate mass of 315.1455 Da (Figure 8D). Although this method also resulted in analyte oxidation, the peak areas of the oxidation products were reduced. Figure 8. View largeDownload slide TICs of Extracted PMZ and CPZ calibrators in BTE Prepared Using FTPE. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS TICs of extracted PMZ and CPZ standards at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–3. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). Figure 8. View largeDownload slide TICs of Extracted PMZ and CPZ calibrators in BTE Prepared Using FTPE. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C,D) UPLC-qTOF-MS TICs of extracted PMZ and CPZ standards at a concentration of 1,000 ng/mL in BTE. Extraneous peaks are labeled 1–3. Below the TICs are the corresponding accurate mass spectra (MS) for each labeled peak in chromatograms C and D (MS-C, MS-D). The final sample preparation method that was assessed was FPTE with no evaporation steps. This approach was a very simple preparation technique with a minimum number of steps that could influence or change the sample. Figure 9 demonstrates that this method did not cause any extraneous peaks to appear in the total ion chromatogram. Therefore, it is reasonable to conclude this method did not induce any measurable oxidation. Figure 9. View largeDownload slide TICs of Extracted PMZ and CPZ Standards in BTE Prepared Using FTPE Without Evaporation. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C–E) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ standards (n = 3) at a concentration of 1,000 ng/mL in BTE Below the TICs are the corresponding accurate mass spectra (MS) for the peaks at 7.76 and 8.35, which correspond to PMZ and CPZ, respectively. Figure 9. View largeDownload slide TICs of Extracted PMZ and CPZ Standards in BTE Prepared Using FTPE Without Evaporation. (A) Neat analyte standard mixture for retention time verification. (B) Drug-free negative control. (C–E) UPLC-qTOF-MS total ion chromatograms of extracted PMZ and CPZ standards (n = 3) at a concentration of 1,000 ng/mL in BTE Below the TICs are the corresponding accurate mass spectra (MS) for the peaks at 7.76 and 8.35, which correspond to PMZ and CPZ, respectively. Semi-quantitative comparison of oxidation products—comparing different extraction conditions Table VI compares the relative formation of the oxidized species produced across the different sample preparation methods used. As the method changed, the relative formation of PMZSO changed as indicated by a decrease in the PMZSO-PMZ peak-area ratio as steps were removed from the sample preparation method. FPTE resulted in less oxidation than was observed with SPE and, when the evaporation step was removed, oxidation products were not measurably formed. Furthermore, it is important to note that the deuterated internal standards were also converted to their corresponding oxidized form to the same extent as the analytes in the neat standards (~4%). Analytical characterization of the FPTE (without evaporation) method Analytical characterization of FTPE without evaporation included assessment of concentration dependence, LOD, LOQ, precision, accuracy, recovery and matrix effects in a manner similar to those proposed in the SWGTOX guidelines. All analytes were fit with quadratic regression lines and concentration dependence was assessed over the range 10 to 2,000 ng/mL. Strong correlations (R2 > 0.99) were observed over five different days. The LOD and LOQ were determined to be 10 ng/mL for all analytes. The precision and accuracy data are summarized in Table V. Precision and accuracy were acceptable ranging from 0.04% to 14% and 0.09% to 20%, respectively. Recovery ranged from 90% to 110%, and matrix effects (suppression or enhancement) were less than 25% for all analytes. Analyte stability while resident on the autosampler tray in extracted samples revealed there was no loss in analyte response in excess of 20% of the initial response ratio at t = 0 h for all analytes except for the PMZ-D3 internal standard. At the 24 h time interval, PMZ-D3 demonstrated loss in response exceeding the 20% threshold. However, this instability did not affect the precision or accuracy of the standard curves, this could be because the time required to analyze all curve samples on the instrument does not commonly exceed 15 h. Table V. Summary of analytical performance parameters (LOD, LOQ, precision, linearity, bias)a Analyte  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  10  10  0.04–8.2 [0/34]  0.998  1.4–(−17.4) [0/10]  PMZSO  10  10  0.37–13.9 [0/34]  0.998  −1.4–(18.5) [0/10]  DPMZ  10  10  0.8–10.8 [0/34]  0.998  −1.1–(8.4) [0/10]  CPZ  10  10  0.9–13.79 [0/34]  0.999  0.55–(19.9) [0/10]  CPZSO  10  10  0.7–11.9 [0/34]  0.998  0.66–(13.6) [0/10]  DCPZ  10  10  0.2–12.7 [0/34]  0.998  0.09–(18.1) [0/10]  Analyte  Limit of detection (LOD, ng/mL)  Limit of quantitation (LOQ, ng/mL)  Precision (CV, %) (acceptance criteria: ≤20%) [# failed]  Linearity (acceptance criteria: R2 ≥ 0.99)  Bias (%) (acceptance criteria: ≤20%) [# failed]  PMZ  10  10  0.04–8.2 [0/34]  0.998  1.4–(−17.4) [0/10]  PMZSO  10  10  0.37–13.9 [0/34]  0.998  −1.4–(18.5) [0/10]  DPMZ  10  10  0.8–10.8 [0/34]  0.998  −1.1–(8.4) [0/10]  CPZ  10  10  0.9–13.79 [0/34]  0.999  0.55–(19.9) [0/10]  CPZSO  10  10  0.7–11.9 [0/34]  0.998  0.66–(13.6) [0/10]  DCPZ  10  10  0.2–12.7 [0/34]  0.998  0.09–(18.1) [0/10]  aFPTE without evaporation was used for sample preparation for analysis by UPLC-QTOF-MS. Data were collected over five different sets of extractions of analyte standard mixtures ranging from 10 to 2,000 ng/mL, where each standard concentration was analyzed in triplicate. Discussion The initial purpose of this research was to develop and validate a semi-quantitative method to evaluate the relative distribution of selected phenothiazine drugs (PMZ and CPZ), and their N-desmethyl and sulfoxide metabolites in skeletal remains. The method was intended for application to studies examining different PMZ and CPZ exposure patterns to understand the significance of drug and metabolite levels in toxicological analysis of bone, and to assess whether their small differences in chemical structure were associated with significant differences in the patterns of drug and metabolite distribution. During the characterization of analytical figures of merit, precision and bias were not consistently meeting the required criteria (≤ 20%). UPLC-PDA chromatograms from extracted standards showed extraneous peaks and further experiments revealed that the analytes were degrading during sample preparation. Consequently, the main objective of the work shifted to the characterization of analyte degradation. The data presented indicate that PMZ and CPZ underwent oxidation during sample preparation and extraction, where PMZ was oxidized to PMZSO and PMZNO while CPZ was CPZSO. Thus, some of the oxidation products included naturally occurring metabolites of the drugs. In addition to the identified oxidation products, other products were formed for which the putative identity include the sulfone, sulfoxide-N-oxide, chlorpromazine N-oxide or hydroxylated form of the parent drug. Oxidation was also observed in analysis of whole blood, and to a greater extent, indicating that the oxidation was not specific to analysis of BTEs, and the extent of oxidation may be expected to vary between different sample matrices. Importantly, the autosampler stability results provide strong evidence that the oxidation took place during the extraction, before the sample was placed on the instrument, and did not occur over time as the samples remained on the autosampler tray. Oxidation of phenothiazines The susceptibility of phenothiazines to oxidation has been reported (7–13), although the majority of such work was done between 1950 and 1990. The most commonly reported oxidation occurs at the sulfide linkage which first forms an unstable radical cation that leads to the generation of the sulfoxide. After the sulfoxide is formed, further oxidation can occur at the sulfur atom which results in the formation of the sulfone (14–16). Various factors have been noted to influence the oxidation reaction including acidity, concentration of oxidizing agents, time, temperature and the side-chain of the molecule (11–13). Much of the research describing phenothiazine oxidation has focused on the generation of oxidation products by chemical, electrochemical, enzymatic and catalytic means (13, 15, 17–19). The oxidation of phenothiazines remains a complex subject and debate about the mechanisms and products formed continues. To the authors’ knowledge, there have been no reports in the analytical toxicology literature describing oxidation of the phenothiazine drugs during the process of preparation of biological samples for analysis. It is possible that the formation of oxidation products may go unnoticed, depending on the sample preparation methods and analytical instrumentation employed. For example, with the use of targeted methods such as GC/MS in SIM mode, or LC/MS/MS in MRM mode, ions corresponding to the oxidation products may not arise in the appropriate time window, or they may be completely excluded from the list of ions or transitions used. For those methods that monitor sulfoxide metabolites, the presence of sulfoxide in any given calibrant or sample chromatogram is expected. Provided that the oxidation reaction occurs in a reproducible and concentration dependent manner, acceptable standard curves could be generated in a given assay. The results of this work suggest that the extent of oxidation is quite reproducible under these sample preparation conditions, as the measured CV values in replicate analyses of all analytes were less than 20% in over 90% of extractions done, and less than 10% in 58–82% (Table I, Supplementary Data). If the extent of oxidation was not reproducible, a wider variability in measured precision and bias values might be expected. This reproducibility represents another reason that analyte oxidation during sample preparation might go unnoticed. Effect of matrix on phenothiazine oxidation The data presented here suggest that the extent of oxidation and the products formed may vary between sample matrices, as shown in comparing data from samples prepared in BTE (Figure 4) to those prepared in blood (Figure 5). The phenothiazines showed a greater extent of oxidation in the blood than in BTE which may be due to the hemoglobin present in red blood cells. This is consistent with studies reporting the oxidation of sulfide functionalities due to various reactive molecules present in blood (20, 21). Furthermore, the authors noticed that CPZ was converted to CPZSO in whole blood, but much of the conversion was due to the set-up of the analytical procedure. Given the wide variation in the nature of sample matrices in postmortem toxicology, it is likely that the extent of phenothiazine oxidation could differ from sample to sample, due to variations in the nature of the matrix. While most bias estimations from the extractions performed were within acceptable limits, this may have been because the same blank matrix was used in preparation of calibrants and the positive control samples used in measurement of accuracy, resulting in a similar extent of oxidation between them. In casework, the matrix used for preparation of standard curves necessarily differs from that of a given sample, and the extent of oxidation may be expected to also differ. Hence, measured phenothiazine and metabolite concentrations may potentially be inaccurate, interfering with toxicological interpretation of the results. The data shown here suggest that all phenothiazine analytes (including drugs and metabolites) should have corresponding deuterated internal standards to minimize the likelihood of erroneous results, presuming that a given analyte and its deuterated analog would oxidize to the same extent. Effect of extraction conditions on phenothiazine oxidation The data presented here show that the analytes underwent oxidation during sample preparation, with oxidation products including naturally occurring metabolites of the drugs being assayed. In some cases, the oxidation resulted in unacceptable bias and precision. In efforts to redesign the method to minimize analyte oxidation, the first area that was investigated was the elution step in the SPE process. Experiments were done where PMZ and CPZ standards were dissolved directly in the EA-based elution solvent and then evaporated to dryness (i.e., with no extraction step from biological matrix), and the resulting data indicated extensive oxidation (see Supplementary Figures 1 and 2 in Supplementary Data). We theorized that the analyte oxidation was due to the specific elution solvent system used, and then evaluated the use of the DCM-based elution solvent system. As shown in Figures 6 and 7 and Table VI, the relative extent of oxidation, as well as the number of products formed, was higher when using the EA-based elution solvent system. Numerous experiments then assessed the effects of reducing the evaporation temperature from 70°C to 40°C, eliminating exposure to light, and evaporation under argon gas instead of using vacuum centrifugation. None of these alterations eliminated oxidation (see Supplementary Data). The subsequent sample preparation method that was employed removed the elution step completely by utilizing FPTE in place of SPE. The results were promising (Figure 8 and Table VI), with reduced oxidation of PMZ and CPZ. Table VI. A semi-quantitative comparison of the relative extent of formation of PMZSO and CPZSO from the corresponding parent compound, based on the sample preparation methoda   Promethazine  Chlorpromazine  Method  Mean peak-area ratio (APMZSO/APMZ × 100%)  Mean peak-area ratio (ACPZSO/ACPZ × 100%)  SPE EA elution solution  67.0  89.2  SPE DCM elution solution  38.9  48.9  Filtration with evaporation  16.3  13.8  Filtration without evaporation  3.6  3.0  Unextracted standards  3.9  3.6    Promethazine  Chlorpromazine  Method  Mean peak-area ratio (APMZSO/APMZ × 100%)  Mean peak-area ratio (ACPZSO/ACPZ × 100%)  SPE EA elution solution  67.0  89.2  SPE DCM elution solution  38.9  48.9  Filtration with evaporation  16.3  13.8  Filtration without evaporation  3.6  3.0  Unextracted standards  3.9  3.6  aThe unextracted drug standard is prepared in Mobile Phase A and analyzed. The peak-area ratio is the ratio of the area of PMZSO or CPZSO relative to that of the parent analyte originally added (i.e., PMZ or CPZ). The peak-area ratio is represented as a mean percentage (n = 3). Through comparison of the relative level of oxidation products in neat standards with those formed in these various experiments, it was hypothesized that the oxidation process was driven largely by the evaporation step. Thus, the FPTE method without evaporation was developed. By removing the evaporation step, no additional oxidation was induced, and the amount of oxidation products formed was indistinguishable from those in the unextracted neat standard (Figure 9, Table VI). Comparison of the results from the four sample preparation methods shows that sample preparation conditions heavily influenced the extent of oxidation, specifically the type and relative formation of the oxidized species produced. Conclusions We have demonstrated that oxidation of selected phenothiazines occurred during standard preparation of samples for analysis by methods typically used in analytical toxicology laboratories. The oxidation products observed included common metabolites of the parent drug and may confound toxicological interpretation. The incidence of oxidation may not be detected by certain analytical configurations. These results are of particular importance for laboratories employing tandem MS methods for analysis of phenothiazines based on MRM. Also, the variability in the extent of oxidation between different samples and calibrators may yield erroneous results. Our work also established the influence of changing the sample preparation method on the extent of oxidation. A new simple extraction method was developed and characterized for the analysis of phenothiazines in skeletal tissues that did not measurably generate any oxidation products. Supplementary Data Supplementary data are available at Journal of Analytical Toxicology online. Acknowledgments The authors would like to thank the Natural Sciences Engineering Research Council of Canada for their financial support of this work. References 1 Paterson, S., Cordero, R., Burlinson, S. ( 2004) Screening and semi-quantitative analysis of post mortem blood for basic drugs using gas chromatography/ion trap mass spectrometry. Journal of Chromatography B , 813, 323– 330. Google Scholar CrossRef Search ADS   2 Maurer, H.H. ( 2005) Multi-analyte procedures for screening for and quantification of drugs in blood, plasma, or serum by liquid chromatography-single stage or tandem mass spectrometry (LC-MS or LC-MS/MS) relevant to clinical and forensic toxicology. Clinical Biochemistry , 38, 310– 318. Google Scholar CrossRef Search ADS PubMed  3 Pirola, R., Mundo, E., Bellod, L., Bareggi, S.R. ( 2002) Simultaneous determination of clomipramine and its desmethyl and hydroxyl metabolites in plasma of patients by high-performance liquid chromatography after solid-phase extraction. Journal of Chromatography B , 772, 205– 210. Google Scholar CrossRef Search ADS   4 Watterson, J.H., Imfeld, A.B., Cornthwaite, H.,.C. ( 2014) Determination of colchicine and O-demethylated metabolites in decomposed skeletal tissues by microwave assisted extraction, microplate solid phase extraction and ultra-high performance liquid chromatography (MAE-MPSPE-UHPLC). Journal of Chromatography B , 960, 145– 150. Google Scholar CrossRef Search ADS   5 Fraser, C.D., Corthnwaite, H.M., Watterson, J.H. ( 2014) Analysis of dextromethorphan and dextrorphan in decomposed skeletal tissues by microwave assisted extraction, microplate solid-phase extraction and gas chromatography-mass spectrometry (MAE-MPSPE-GCMS). Drug Testing and Analysis , 7, 708– 713. Google Scholar CrossRef Search ADS PubMed  6 Watterson, J.H., Donohue, J.P. ( 2011) Relative distribution of Ketamine and Norketamine in skeletal tissues following various periods of decomposition. Journal of Analytical Toxicology , 35, 452– 458. Google Scholar CrossRef Search ADS PubMed  7 Davidson, A.G. ( 1978) The determination of sulphoxide in degraded phenothiazine formulations by difference spectrophotometry. Journal of Pharmacy and Pharmacology , 30, 410– 414. Google Scholar CrossRef Search ADS PubMed  8 Chignell, C.F., Motten, A.G., Buettner, G.R. ( 1985) Photoinduced free radicals from chlorpromazine and related phenothiazines: relationship to phenothiazine-induced photosensitization. Environmental Health Perspectives , 64, 103– 110. Google Scholar CrossRef Search ADS PubMed  9 Bahnemann, D., Asmus, K.D., Willson, R.L. ( 1983) Free radical induced one-electron oxidation of the phenothiazines chlorpromazine and promethazine. Journal of the Chemical Society, Perkin Transactions , 2, 1661– 1668. Google Scholar CrossRef Search ADS   10 Underberg, W.J.M. ( 1978) Oxidative degradation of pharmaceutically important phenothiaiznes I: isolation and identification of oxidation products of promethazine. Journal of Pharmaceutical Sciences , 67, 1128– 1131. Google Scholar CrossRef Search ADS PubMed  11 Karpinska, J., Starczewska, B., Puzanowska-Tarasiewicz, H. ( 1996) Analytical properties of 2- and 10-disubstitutied phenothiazine derivatives. Analytical Sciences : the International Journal of the Japan Society for Analytical Chemistry , 12, 161– 170. Google Scholar CrossRef Search ADS   12 Pankratov, A.N., Uchaeva, I.M., Stepanov, A.N. ( 1993) Chemical and electrochemical oxidation of phenothiazine. Canadian Journal of Chemistry , 71, 674– 677. Google Scholar CrossRef Search ADS   13 Kojlo, A., Karpinska, J., Kuzmicka, L., Misiuk, W., et al.  . ( 2001) Analytical study of the reaction of phenothiazines with some oxidants, metal ions, and organic substances (review article). Journal of Trace and Microprobe Techniques , 19, 45– 70. Google Scholar CrossRef Search ADS   14 Takahashi, D.M. ( 1980) Rapid determination of chlorpromazine hydrochloride and two oxidation products in various pharmaceutical samples using high-performance liquid chromatography and fluorescence detection. Journal of Pharmaceutical Sciences , 69, 184– 187. Google Scholar CrossRef Search ADS PubMed  15 Bosch, E., Kochi, J.K. ( 1995) Catality oxidation of chlorpromazine and related phenothiazines. Cation radicals as the reactive intermediates in sulfoxide formation. Journal of the Chemical Society, Perkin Transactions , 1, 1057– 1064. Google Scholar CrossRef Search ADS   16 Cheng, H.Y., Sackett, P.H., McCreery, R.L. ( 1978) Kinetics of chlorpromazine cation radical decomposition in aqueous buffers. Journal of the American Chemical Society , 100, 962– 967. Google Scholar CrossRef Search ADS   17 Wisniewska, J., Rzesnicki, R., Topolski, A. ( 2011) A mechanistic study on the disproportionation and oxidative degradation of phenothiazine derivatives by manganese(III) complexes in phosphate acidic media. Transition Metal Chemistry , 36, 767– 774. Google Scholar CrossRef Search ADS   18 Blankert, B., Haven, H., van Leeuwen, S.M., Karst, U., et al.  . ( 2005) Electrochemical, chemical and enzymatic oxidations of phenothiazines. Electroanalytical , 17, 1501– 1510. Google Scholar CrossRef Search ADS   19 Gasco, M.R., Carlotti, M.E. ( 1978) Oxidation of promazine and promethazine by ferric perchlorate. Journal of Pharmaceutical Sciences , 67, 168– 171. Google Scholar CrossRef Search ADS PubMed  20 McKay, G., et al.  . ( 1985) Therapeutic monitoring of chlorpromazine II: pitfalls in whole blood analysis. Therapeutic Drug Monitoring , 7, 472– 477. Google Scholar CrossRef Search ADS PubMed  21 Hawes, E.M., et al.  . ( 1986) Therapeutic monitoring of chlorpromazine III: minimal interconversion between chlorpromazine and metabolites in human blood. Therapeutic Drug Monitoring , 8, 37– 41. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Journal

Journal of Analytical ToxicologyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from Google Scholar, PubMed
Create lists to organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off