Article pubs.acs.org/ac
Paper Spray and Extraction Spray Mass Spectrometry for the Direct and Simultaneous Quantification of Eight Drugs of Abuse in Whole Blood Ryan D. Espy,† Sebastiaan Frans Teunissen,‡ Nicholas E. Manicke,§ Yue Ren,∥ Zheng Ouyang,∥ Arian van Asten,*,‡,⊥ and R. Graham Cooks*,† †
Department of Chemistry and Center for Analytical Instrumentation, Purdue University, West Lafayette, Indiana 47907, United States ‡ Netherlands Forensic Institute, Department of Forensic Chemistry, Toxicology Laboratory, The Hague, The Netherlands § Department of Chemistry and Chemical Biology, Indiana University-Purdue University Indianapolis, Indianapolis, Indiana 47907, United States ∥ Department of Biomedical Engineering, Purdue University, West Lafayette, Indiana 47907, United States ⊥ Van ‘t Hoff Institute for Molecular Sciences, University of Amsterdam, Amsterdam, The Netherlands ABSTRACT: Determination of eight drugs of abuse in blood has been performed using paper spray or extraction spray mass spectrometry in under 2 min with minimal sample preparation. A method has been optimized for quantification of amphetamine, methamphetamine, 3,4-methylenedioxyamphetamine (MDA), 3,4-methylenedioxy-N-methylamphetamine (MDMA), 3,4-methylenedioxy-N-ethylamphetamine (MDEA), morphine, cocaine, and Δ9-tetrahydrocannabinol (THC) from a single blood spot. Sample to sample variations of 1−5% relative standard deviation were achieved using stable isotope-labeled internal standards and tandem mass spectrometry. Limits of detection for all drugs were below typical physiological and toxicological levels. Paper spray and extraction spray each used less than 10 μL of whole blood. These methods exhibit the potential for performing rapid and high-throughput assays for selective on-site multicompound quantitative screening of illicit drugs.
T
target these drugs with sufficient sensitivity and specificity. These drugs include marijuana, cocaine, amphetamines, opiates, PCP, benzodiazepines, and barbiturates, the first five constituting the “NIDA-5” of the U.S. National Institute of Drug Abuse. One or more of these drugs was reported to be found in over 30% of U.S. drivers stopped or arrested for driving under the influence.9 Norway is the first country planning to implement legislative limits for nonalcohol drugs corresponding to impairment seen at increasing blood alcohol concentrations. The background and justification for the suggested concentrations limits were recently published.10 With the implementation of impairment levels for driving under the influence of nonalcohol drugs also comes the challenge of enforcement by the police authorities. An instructive example is found in the case of quantitative alcohol testing which is used routinely on a large scale in many countries. These tests, used both in the field and in the laboratory, require robust, validated, certified, and high
he need for rapid assays for illicit drugs is becoming increasingly apparent as many countries in Europe and North America are paying increased attention to the problem of driving while drug-impaired, as outlined in a 2011 White Paper1 and the European Commission ROSITA (road-side testing assessment) initiative.2 Several research studies performed in various countries over the past few years correlate illegal drug use with traffic accidents and deaths.3 The need for road-side drug testing will become more urgent in the United States as more states legalize the consumption of marijuana. Marijuana is already the most commonly detected nonalcohol substance among drivers in accidents with several other drugs also being detected during recent studies.4−7 In situ instrumentation and analytical methods would greatly improve forensic efficiency in the field and reduce sample backlogs in the laboratory.8 Developing a suitable assay for “instant” and selective drug quantification would have several important applications not only in road-side testing for driving under the influence but also in forensics and toxicology, employment and workplace screening, and therapeutic drug monitoring. A small set of compounds constitutes a large portion of all drug cases, and therefore assays should be developed which © 2014 American Chemical Society
Received: May 4, 2014 Accepted: June 26, 2014 Published: June 26, 2014 7712
dx.doi.org/10.1021/ac5016408 | Anal. Chem. 2014, 86, 7712−7718
Analytical Chemistry
■
throughput methodology. Performing such robust and rapid screening for a set of eight drugs at very low cutoff levels in human biological matrices is, however, more challenging. Several methods currently exist for the measurement of illicit drugs. 11 The most common are immunoassays, 12 gas chromatography mass spectrometry (GC/MS),13 and liquid chromatography mass spectrometry (LC/MS/MS).14,15 The most popular immunoassay is enzyme-linked immunosorbent assay (ELISA), a biochemical test which uses antibodies and colorimetry to detect particular substances.16−18 The drawbacks of ELISA are well-known, with a lack of specificity (high falsepositives) being the primary issue. On the contrary, analytical methods utilizing tandem mass spectrometry are ideal as they provide multiple levels of analyte confirmation. An additional level of selectivity can be achieved when used in conjunction with chromatography, with liquid chromatography being the current gold standard. For monitoring and investigative purposes, there is a need for robust analytical methods that can be used to rapidly perform an accurate quantitative analysis of the most frequently encountered drugs of abuse and associated metabolites in human whole blood and preferably in other biological matrices including saliva and urine. Ideally, such methods would not require strict laboratory conditions and trained analytical experts. Road-side testing by police officers in combination with a minimally invasive sampling procedure would be best, but “out of laboratory” analysis at a police station could be equally beneficial allowing investigations to be processed swiftly and efficiently. This not only allows efficient use of resources within the criminal justice system but also minimizes uncertainty and duress on suspects and victims. It should in this respect be noted that the anticipated potential of “bringing the lab to the sample” approach may only be fully realized when the methods yield validated results that can be used as evidence in court. Currently, the available field methods often provide preliminary results thus requiring follow up analysis at the forensic laboratory. The forensic toxicological methods in use in the laboratory are highly accurate and sensitive and meet the formal quality criteria. However, state-of-the-art LC-MS analysis does not meet the criteria for in-field or point-of-care use by nonexperts in terms of mobility, robustness, or ease of operation. The recently introduced ambient mass spectrometry methodologies19,20 have the potential to bridge the gap between highly accurate and selective quantitative toxicological analysis and allow robust use outside the laboratory by nonexperts.21 Paper spray22,23 has been well-characterized for quantification of pharmaceutical drugs for whole blood,24−30 and this study describes its development with the purpose of applying it in a forensic setting. The work in this article employs ambient paper spray ionization mass spectrometry as well as extraction spray mass spectrometry31 to quantify a set of illicit drugs from blood in less than 2 min per sample. These methods, which use a small piece of filter paper, 2−10 μL of blood, and a few drops of solvent, provide the ability to perform online extraction and analysis with no sample preparation. Experiments were designed following recommendations by SWGTOX Standard Practices for Method Validation in Forensic Toxicology32 including bias, precision, carryover, interferences, as well as calibration and sensitivity guided by impairment-based limits.10
Article
EXPERIMENTAL METHODS
Chemicals and Reagents. All organic solvents (HPLC grade) were purchased from VWR Scientific (Chicago, IL, USA). Drug standards (in solution) and deuterated internal standards (in solution) were purchased from Cerilliant (Reston, VA, USA) and stored at −20 °C. The internal standards had specific deuteration sites which could differ from other manufacturers and affect fragment m/z values and ion ratios. Whatman grade 31ET-Chr paper was obtained as a sample from Whatman (Piscataway, NJ, USA). Human whole blood (20 individuals of varying age, gender, and race) with sodium heparin anticoagulant was purchased from Innovative Research Inc. (Novi, MI, USA) and stored at 4 °C. Vacuum tubes coated with sodium fluoride stabilizer were purchased from VWR Scientific (Chicago, IL, USA). Sample Preparation. Experiments were performed on the same day as working drug solutions were prepared which were carried out by serial dilution with methanol or acetonitrile. Blood was incubated to 37 °C prior to use and injected into sodium fluoride vacuum tubes. Internal standards were first diluted with water and then spiked into the blood. The working drug solutions were spiked into the liquid blood by 20-fold dilutions (10 μL into 190 μL blood) and vortexed at 3000 rpm for 15 s. Quality control samples were prepared separately using unique drug standard ampules. Paper Spray Ionization. Paper spray was performed using an automated ion source with disposable cartridges manufactured by Prosolia, Inc. (Indianapolis, IN). The cartridges consisted of a 6 cm2 teardrop-shaped paper substrate (Whatman chromatography paper, grade 31ET-Chr), a plastic casing, and a steel ball for electrical contact. Twelve microliters of spiked blood was spotted onto the paper and dried for 20 min in a 37 °C oven. Drying was performed for suitability purposes and to simulate a forensic laboratory setting, though the experiment may also be performed using fresh blood.27 Solvent was applied to the cartridge by three injections of 5 μL each on top of the blood spot, with a two-second delay between injections. Five injections of 15 μL each were then added behind the blood spot with a five-second delay between injections. This sequential solvent-injection program was designed to produce gradual solvent introduction and maximize analyte extraction. The solvent used was 80:20:0.1 methanol/dichloromethane/ hydroxylamine for all compounds other than THC. For THC, first an elution using the afore-mentioned solvent was sprayed, followed by the addition of 10 μL of methanol with 25 mM sulfuric acid being added on top of the blood spot and being sprayed again. The spray voltage was +3500 V. While the sample analysis was being performed, the next sample simultaneously received solvent injection. Extraction Spray Ionization. The functional principles of the extraction spray method are essentially identical to paper spray except for insertion of the piece of paper into a narrow, pulled glass capillary, as described: A paper strip (Whatman chromatography paper, grade 31ET-Chr, 6 × 0.5 × 0.7 mm) was preloaded with a 1.5 μL blood sample and dried in open air for 5 min. The loaded sample strip was inserted into a nanospray capillary (borosilicate glass pulled in-house) followed by an addition of 10 μL of solvent for extraction and spray. A DC voltage of 2000 V was added via a wire electrode (Warner Instrument Corp., Hamden, CT, USA). The solvent used was 80:20:0.1 methanol/dichloromethane/ 7713
dx.doi.org/10.1021/ac5016408 | Anal. Chem. 2014, 86, 7712−7718
Analytical Chemistry
Article
Table 1. MRM Transitions for Drugs and Corresponding Internal Standards MW
precursor m/z
fragment m/z (MRM transitions)
collision (arbitrary units)
tube lens (V)
ion ratioa
amphetamine
135.2
136.1
100:53
149.2
150.1
67
100:54
MDA
179.2
180.1
68
100:97
MDMA
193.2
194.1
65
100:25
MDEA
207.3
208.1
65
100:25
morphine
285.3
286.1
110
100:75
cocaine
303.3
304.1
74
100:6
Δ9-THC
314.5
315.1
91
100:34
amphetamine-D6
141.2
142.1
84
100:95
methamphetamine-D5
154.2
155.1
72
100:100
MDA-D5
184.2
185.1
88
100:100
MDMA-D5
198.3
199.1
75
100:30
MDEA-D5
212.3
213.1
71
100:25
morphine-D6
291.4
292.1
114
100:67
cocaine-D3
306.4
307.1
79
100:6
Δ9-THC-D3
317.5
318.1
17 7 19 11 23 19 12 26 13 28 45 42 20 25 26 29 19 10 25 13 26 20 15 24 16 26 25 39 20 25 29 39
75
methamphetamine
91.2 119.2 91.2 119.2 105.1 135.2 163.0 105.1 163.0 105.1 165.0 152.1 182.0 105.0 192.9 122.9 93.1 125.1 92.1 121.1 110.1 138.1 165.0 135.1 163.0 133.0 201.0 153.0 185.1 105.1 196.0 123.1
80
100:42
compound
a
Ion ratios refer to the ratio of the intensities of the MRM transitions.
Figure 1. Mass spectrometry analysis conditions for each sample. The method was performed automatically using the Xcalibur software. Each drug and deuterated internal standard had two product ions selected, one for quantitative and one for qualitative purposes.
Mass Spectrometry, Data Collection, and Data Processing. Paper spray and extraction spray were performed on a TSQ Quantum Access MAX (Thermo Scientific, San Jose, CA) in the multiple reaction monitoring (MRM) mode. Each monitoring experiment interrogated ions in a 0.010 m/z
hydroxylamine for all compounds other than THC. Methanol with 25 mM sulfuric acid was used to analyze THC. Extraction spray setup and analysis were performed manually because the automated paper spray source did not support this type of ionization source. 7714
dx.doi.org/10.1021/ac5016408 | Anal. Chem. 2014, 86, 7712−7718
Analytical Chemistry
Article
unstable below 2500 V, thus 3500 V was used in the positive ion mode. Most of the compounds performed well using standard organic solvents with an acidic modifier added for increased ionization efficiency, although morphine performed poorly under acidic conditions and THC required a special reaction with sulfuric acid to achieve adequate sensitivity. THC optimization (in the positive ion mode) is shown in Figure 3
window for a period of 75 ms, repeated 20 times for a total of 1.5 s to analyze each transition. The parameters used are shown in Table 1. Note that each of these values may differ between instruments and ionization sources. These values were determined automatically during instrument optimization. Abundant neutral losses which were not selective (e.g., NH3, 17 Da) were manually excluded. Inlet capillary temperature and voltage were 300 °C and 35 V, respectively. The most abundant fragment was used for quantification. Special MS conditions were required for morphine. Morphine performed best at a higher collision energy, optimally 2.1 mTorr, while the amphetamines fragmented too readily at this pressure. The optimized collision pressure for the amphetamines was 1.0 mTorr. In order to analyze all eight compounds from a single blood spot, a mass spectrometric method was developed to adjust the collision pressure on the fly. Figure 1 shows how the scan parameters were varied over the 90 s data acquisition period. The mass spectrometer required 10−15 s to stabilize after being switched between 1.0 and 2.1 mTorr. An additional segment was used for the eighth drug, THC, which, as described below, utilized a different spray solvent. Data were processed using the Xcalibur Quan Browser. Peaks were integrated, and quantification was performed using a ratio of the areas under the curve for the analyte and internal standard. Trend lines were constructed using 1/x2 weighted linear least-squares, and limits of detection were calculated as 3 times the standard deviation of the blank divided by the slope of the trend line.33 Reported precision values are the relative standard deviation of concentration determined from multiple blood spots taken from a single blood sample.
■
Figure 3. THC optimization: 10 μg/mL Δ9-THC by paper spray mass spectrometry with (top) 95:5:0.01 methanol/water/acetic acid and (bottom) methanol with 25 mM sulfuric acid. Bottom inset is the MS/ MS spectrum of protonated Δ9-THC.
RESULTS AND DISCUSSION Method Development. The primary objective and challenge in method development was to find a suitable set of conditions that would allow all drugs to be analyzed simultaneously from a single blood spot. The method was optimized in terms of solvent system, spray voltage, tube lens, collision energy, collision pressure, and scan speed. Tube lens and collision energy values were optimized prior to the experiment and adjusted in real-time by the instrument in MRM mode. The signal for morphine was greatly affected by collision pressure optimization, with an optimum pressure of 2.1 mTorr providing a 5-fold increase in sensitivity compared to 1.0 mTorr (Figure 2). Spray voltage was optimized to be as low as possible while still producing a stable spray and ion current. Lower voltages were also desirable for lowering the background produced by chemical noise. Paper spray became completely
where addition of 25 mM sulfuric acid improved sensitivity by 50 times. The cause of this improvement lies in the acidcatalyzed ring opening at the cyclohexyl ether, placing a permanent positive charge on THC. This process is a precursor step in the Ghamrawy reaction.34,35 The best solution for morphine was to use hydroxylamine as a solvent modifier. Although the mechanism is not entirely understood, we believe hydroxylamine to be a very mild source of protons without requiring adjustment of the solution pH. The sensitivities of the amphetamines and cocaine were unaffected using hydroxylamine instead of acetic acid. THC needed to be analyzed last after solvent modification or in a separate experiment because the sulfuric acid solvent system for THC was incompatible with the other drugs. For sequential analysis, THC may be analyzed from the same blood spot after the other drugs. The initial solvent (methanol/dichloromethane) gradually dried during analysis, after which at 70 s, 25 mM sulfuric acid was added to the paper. This addition was manual as the ion source did not yet have the capability of automating this step. The THC, which had already been extracted from the blood spot, reacted with the sulfuric acid. Upon reapplying spray voltage, the reaction product sprayed from the paper tip. It was important that the sulfuric acid solution was applied on top or ahead of the blood spot (not behind it), because this solution lysed the blood and caused it to run on the paper. Performing THC analysis separately or successively had no change on the limit of detection for THC.
Figure 2. Instrument optimization of collision pressure for morphine (m/z 286.1 to 165.0). 7715
dx.doi.org/10.1021/ac5016408 | Anal. Chem. 2014, 86, 7712−7718
Analytical Chemistry
Article
Table 3. Paper Spray Quality Control Dataa
Analytical Performance of Paper Spray MS. Calibration and Quality Controls. The use of isotopically labeled internal standards provided relative standard deviations of 1 to 10% at concentrations above the lower limit of quantification. Calibration curves were linear over the ranges tested (3 orders of magnitude). Detection limits were all at or well below the driver impairment limits previously established.10 Sensitivity for the amphetamines and cocaine were near or better than 1 ng/ mL. The LOD of THC was 4 ng/mL (Figure 4), while
analyte
nominal concentration (ng/mL)
interassay accuracy (% dev)
interassay precision (% CV)
40 120 400 3200 10 30 100 800 40 120 400 3200 10 30 100 800 10 30 100 800 40 120 400 3200 10 30 100 800 40 120 400 800
98.7 96.0 98.1 103.7 87.0 88.6 96.3 101.5 99.0 97.0 97.3 105.4 102.7 95.5 97.9 102.4 99.9 94.8 98.9 101.4 169.6 106.5 102.6 99.7 98.8 101.1 106.0 107.0 117.9 91.0 90.4 97.0
14.5 4.3 4.3 4.8 11.7 3.4 3.4 2.5 16.5 7.4 9.7 6.6 10.7 4.7 2.4 1.9 9.6 1.4 3.1 2.0 35.4 10.1 5.0 13.0 8.0 2.5 1.3 1.6 12.8 1.4 3.3 2.5
amphetamine
methamphetamine
MDA
MDMA
Figure 4. Calibration curve for THC from whole blood by paper spray mass spectrometry with THC-d3 internal standard. Quality control samples are indicated as red X’s.
MDEA
morphine provided better sensitivity for extraction spray (0.5 ng/mL) compared to paper spray (12 ng/mL). This is likely attributable to the stability of the spray and/or the smaller size of droplets from the nanospray emitter. Limits of detection for all compounds are shown in Table 2. Analyte carryover was also
morphine
cocaine
Table 2. Limits of Detection for the Eight Drugs Analyzed from One Sample drug
paper spray LOD (ng/mL)
extraction nano-electrospray LOD (ng/mL)
amphetamine methamphetamine MDA MDMA MDEA morphine cocaine Δ9-THC
1 0.3 2 0.04 0.3 12 0.05 4
0.9 1 2 0.3 0.08 0.5 0.06 2
Δ9-THC
a
n = 5 per run. Data are marked in bold when criteria were not met.
LLOQ all showed better than 5% RSD with excellent accuracies (Table 4). Table 4. Ion Ratios for Two SRM Transitions of Each Compound Using Paper Spray
monitored by running blanks after the highest calibration standards. Carryover was determined to be negligible (
