Article Cite This: Anal. Chem. 2018, 90, 952−960
pubs.acs.org/ac
Single-Use Poly(etheretherketone) Solid-Phase Microextraction− Transmission Mode Devices for Rapid Screening and Quantitation of Drugs of Abuse in Oral Fluid and Urine via Direct Analysis in RealTime Tandem Mass Spectrometry Tijana Vasiljevic, Germán Augusto Gómez-Ríos, and Janusz Pawliszyn* Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1 S Supporting Information *
ABSTRACT: The analysis of oral fluid (OF) and urine samples to detect drug consumption has garnered considerable attention as alternative biomatrices. Efficient implementation of microextraction and ambient ionization technologies for rapid detection of target compounds in such biomatrices creates a need for biocompatible devices which can be implemented for in vivo sampling and easily interfaced with mass spectrometry (MS) analyzers. This study introduces a novel solid-phase microextraction−transmission mode (SPME-TM) device made of poly(etheretherketone) (PEEK) mesh that can rapidly detect prohibited substances in biofluids via direct analysis in real-time tandem MS (DART-MS/MS). PEEK mesh was selected due to its biocompatibility, excellent resistance to various organic solvents, and its ability to withstand relatively high temperatures (≤350 °C). The meshes were coated with hydrophilic−lipophilic-balance particle-poly(acrylonitrile) (HLBPAN) slurry. The robustness of the coated meshes was tested by performing rapid vortex agitation (≥3200 rpm) in LC/MSgrade solvents and by exposing them to the DART source jet stream at typical operational temperatures (∼250−350 °C). PEEK SPME-TM devices proved to be robust and were therefore used to perform ex vivo analysis of drugs of abuse spiked in urine and OF samples. Excellent results were obtained for all analytes under study; furthermore, the tests yielded satisfactory limits of quantitation (median, ∼0.5 ng mL−1), linearity (≥0.99), and accuracy (80−120%) over the evaluated range (0.5−200 ng mL−1). This research highlights plastic SPME-TM’s potential usefulness as a method for rapidly screening for prohibited substances in on-site/in vivo scenarios, such as roadside or workplace drug testing, antidoping controls, and pain management programs. ithin the field of forensic science, there is a great need for a highly robust and reliable tool that can rapidly detect the presence of drugs of abuse in biofluids.1 Driving under the influence of drugs (DUID) is a very serious criminal offense that can have dire consequences. According to the Canadian Centre on Substance Abuse (CCSA), substance use played a primary role in approximately 34.2% of fatal car crashes in 2010.2 Due to an increase in such accidents, many methods have been developed to detect commonly abused prohibited substances in oral fluid3−8 and urine.3,9,10 These methods generally use a chromatographic step as a confirmatory test, which places a time restraint on the whole procedure. This is especially problematic in DUID cases where obtaining reliable results rapidly is of prime importance. While still requiring extensive validation procedures for wide acceptance as a confirmatory method,1 it is unsurprising that ambient mass spectrometry (AMS)11,12 has received considerable attention from forensic scientists, as this group of technologies is capable of rapidly determining target compounds while also reducing or avoiding the need for sample preparation, eliminating the chromatographic step, requiring minimal or
W
© 2017 American Chemical Society
no solvent use, and decreasing the overall time required to complete the process. Among the diverse array of AMS techniques, direct analysis in real time (DART) has been of particular interest in forensic science. This technology uses heated and highly energetic helium (He) gas to desorb/ionize compounds from surfaces of various materials.13−15 Although DART can reveal important qualitative information about the forensic sample, its applicability for quantitative analysis tends to be more troublesome because it lacks the ability to normalize the sample (i.e., a homogeneous and reproducible mechanism for introducing the analytes into the MS system from one sample to the other). In an attempt to solve this issue, Fernandez’s group developed a mesh-like device with fixed geometrical characteristics (i.e., strand size, pitch-to-pitch distance, and percentage opening) that provided better control over sample introduction.16 This device, known as transmission mode (TM), allowed gas to flow more efficiently through the Received: September 29, 2017 Accepted: November 24, 2017 Published: November 24, 2017 952
DOI: 10.1021/acs.analchem.7b04005 Anal. Chem. 2018, 90, 952−960
Article
Analytical Chemistry
its great mechanical stability, a property arising from its aromatic structure,42 which can be seen in Supporting Information Figure S1. In addition, PEEK is a thermoplastic material, with a melting temperature of 343 °C,43, and with the exception of 98% sulfuric acid, it is chemically inert to most solvents and acids.44 It is important to note however that in certain cases target analytes can be highly nonpolar, so a careful investigation of potential analyte adherence to plastics should be done prior to their use.63 PEEK’s biocompatible feature has allowed it to be used in dental surgery45 and maxillofacial reconstruction,46 as well as in spinal,47 orthophedic,48 and cardiac implants.49 Given PEEK’s many benefits, we decided to apply SPME coating to a PEEK-based mesh to rapidly detect and quantitate drugs in oral fluid and urine via DART-MS/MS. Different parameters were evaluated, such as coating robustness under strong agitation or high temperatures, as well as mesh reusability. Encouraging results led us to apply disposable SPME-TM to quantitatively determine drugs of abuse in ex vivo OF and urine analysis. In addition, we attempted a semiquantitative analysis of caffeine concentration in an OF sample obtained from a volunteer who had recently consumed coffee. Certainly, the results obtained show that SPME has tremendous potential as a simple, effective device for roadside drug testing.
mesh, which permitted better ion transmission and reproducibility.16 Although this technology enabled the sample to be pragmatically positioned in front of the MS system (i.e., a spot of the sample is dried on the mesh prior to analysis), its lack of sample preparation prevented a dramatic enhancement in the limits of detection. In this regard, one of the sample preparation tools that has been successfully implemented with many AMS techniques is solid-phase microextraction (SPME).17−19 Depending on the experimental conditions, compounds of interest, and matrices under investigation, SPME can extract either a “small” amount of analyte (also known as negligible depletion) or it can be used in cases in which a significant amount of analyte is extracted (i.e., exhaustive extraction).20 SPME is cleverly designed, comprising the immobilization of a small amount of extractive material (coating) onto a designated substrate.20 This sample preparation tool has seen a rapid increase in bioanalytical applications over the past few years due to the development of matrix-compatible coatings21−25 that allows SPME devices to be directly introduced (DI) into complex matrices, which in turn leads to negligible biofouling and remarkable target analyte extraction. In addition, SPME features efficient sample cleanup, which reduces issues related to ionization suppression and ionization enhancement.26,27 Indeed, SPME is no stranger to DART, and diverse geometrical SPME formats (e.g., fiber, in-tube, mesh) have been interfaced with DART in numerous applications.21,28−31 Nonetheless, as our team has recently demonstrated,18,32,33 the best results are obtained when SPME is implemented as a TM substrate. For instance, we were able to quantify cocaine and diazepam in sub(ng mL−1) levels in urine and plasma samples using stainless steel (SS) meshes that were adequately coated only on the strands with an octadecylpoly(acrylonitrile) (C18-PAN) slurry.32 Furthermore, we have elsewhere documented how mesh coated with hydrophilic−lipophilic-balanced (HLB) particles can be used to efficiently detect pesticides in different food matrices.18,33 These meshes were also coupled to a portable MS system to execute rapid semiquantitative analyses of target analytes in complex matrices, as well as to profile milk samples from different species and farming systems.33 Over the past few years, various substrates have been used for the immobilization of extractive materials in SPME34−37 devices, the most recent example of which being poly(butylene terephthalate) (PBT) plastic support, which was developed by Reyes-Garcés et al.38 PBT proved to be a very useful material for single-use SPME, as the device provided satisfactory results for drug quantitation in plasma, urine, and blood, while also being very robust. PBT-based devices also showed no interferences from being used as a material in the method.38 It is evident that the use of alternative materials opens up a diverse array of potential applications for SPME; for example, on-site road testing. Given the rise in DUID accidents over the past few years, our primary goal was to develop a single-use device that can be interfaced with DART and that can eventually be used for rapid, in vivo, on-site drug detection. In order to fulfill such a requirement, it is essential to use a biocompatible material, which is a material that does not detrimentally affect living systems.39 While SPME is able to satisfy the above definition, it must also possess a barrier (binder) that prevents the coating from being fouled by the adherence of large biomolecules or by coextracted matrix components.40 As such, poly(etheretherketone) (PEEK) was selected in order to fulfill the substrate biocompatibility requirement.41 PEEK is a semicrystalline polymer known for
■
EXPERIMENTAL SECTION Reagents and Supplies. The following standards were obtained from Cerilliant (all standards had a concentration of 1000 mg L−1): cocaine, methamphetamine, nordiazepam, fentanyl, 3,4-methylenedioxymethamphetamine (MDMA), heroin, phencyclidine (PCP), oxazepam, methadone, oxycodone, lorazepam, lysergic acid dimethylamine (LSD), diazepam, caffeine, and nicotine. The respective internal standards were also ordered from Cerilliant (all at a concentration of 100 mg L−1), namely, cocaine-d3, methamphetamine-d5, MDMA-d5, PCP-d5, nordiazepam-d5, oxazepam-d5, methadone-d3, oxycodone-d3, lorazepam-d4, LSD-d3, heroin-d9, diazepam-d5, fentanyl-d5, caffeine-C13, and nicotine-d4. A pooled batch of urine and OF was collected by having 10 healthy individuals (5 male and 5 female) each expectorate into a 10 mL vial. Phosphate buffer saline (PBS) was then made in the laboratory according to a published procedure.50 In addition, a female volunteer provided 1 mL of OF to measure for caffeine levels. For this test, OF was collected after a 24 h caffeine fast, followed by collection at 5 min, 1 h, 3 h, and 5 h post-coffee consumption. Further details about suppliers can be obtained in the Supporting Information (see Supplier Information, page S3). DART-MS/MS and LC-MS/MS. A DART-standardized voltage and pressure (DART-SVP) model ion source (IonSense, Inc., Saugus, MA, USA) was coupled to a triple quadrupole mass spectrometer (TSQ Vantage) by Thermo Scientific (San Jose, CA, USA) via a Vapur interface (IonSense). The needle valve of the membrane pump used with the Vapur interface was adjusted to the blue indicator at position 4 in order to maintain adequate MS vacuum and provide sufficient sensitivity required for quantitative analysis via DART-TM. Adjustment was performed according to the Vapur Pump Optimization Protocol suggested by the manufacturer. The DART-SVP was fitted with a singledimensional motorized linear rail that was controlled through the DART-SVP web-based software in order to reproducibly and consecutively automatically position the SPME-TM devices 953
DOI: 10.1021/acs.analchem.7b04005 Anal. Chem. 2018, 90, 952−960
Article
Analytical Chemistry in front of the DART source (a speed of 0.2 mm s−1 was used). In order to guarantee good reproducibility and higher throughput of the desorption process, a custom-made holder (UW-12) able to allocate up to 12 SPME-TM devices was developed.32 The DART source was operated using the following conditions: positive ion mode; high-voltage (HV) electrode (−3000 V); discharge electrode (+350 V); and a grid voltage of +350 V. The gas heater was optimized at 350 °C in order to yield the optimum intensities for most of the analytes. LC-MS/MS conditions are detailed in the Supporting Information. Preparation of PEEK Mesh for SPME-TM Coating. A 12 × 12 cm mesh piece was cut and cleaned in isopropanol (IPA) and water (H2O) for 15 min by sonication. The meshes were then dried (100 °C) and etched for 90 s on both sides using an ATC-2020-IM ion mill (AJA International, Scituate, MA, USA) at a setting of 400 V and 190 mA. The meshes were then purged with nitrogen (N2) and left in a desiccator to ensure that the coating application remained unaffected by possible contaminants and moisture. A paper cutter was used to obtain mesh strips (2.5 × 0.5 cm, length by width). Images of the mesh were taken using an Olympus microscope (SZX10) with a SC30 digital camera (Olympus, Tokyo, Japan), and SEM images were acquired using a Zeiss FESEM 1530 (Carl Zeiss, Oberkochen, Germany). Application of HLB-PAN Coating to the Meshes. PAN solution was prepared by dissolving 7 g of PAN powder with 100 mL of dimethylformamide (DMF) via periodical vortexing to ensure uniform dissolution. Oasis HLB particles (Waters, Milford, MA, USA) were used to prepare the slurry for the coating application. Since the size of the particles (30 μm) was large for the application, the particles were ground using a ball mill for 2 h at 240 rpm. The ground particles were then collected, and 1 g was mixed with 10 mL of PAN solution to obtain a final PAN-HLB slurry. The slurry was then stirred for 12 h at 1800 rpm to provide adequate mixing of particles with PAN. The mixture was applied to the meshes by dip coating using an in-house method developed at the University of Waterloo. After the mixture had been applied, excess solvent was removed via gas pressure, and the coating was then cured at 100 °C. In order to facilitate the handling of the meshes, the meshes were attached to PBT with soldering tool to provide extra support (i.e., thermally attached; see Figure 1). The meshes were then cleaned twice in a solvent solution consisting of methanol, isopropanol, and acetonitrile (MeOH/IPA/ACN, 50:25:25) at 1500 rpm for 15 min to remove any leftover chemicals from the polymerization or coating procedures. To activate the extractive sites on the particles, the meshes were preconditioned in a mixture of methanol/water (MeOH/H2O, 50:50). Evaluation of Coating Mesh Endurance. The evaluation was conducted by exposing a new set of coated PEEK meshes to DART’s thermal desorption (350 °C, 0.2 mm s−1). After the first thermal desorption, the meshes (n = 5) were preconditioned and used to perform cocaine extraction from a PBS sample that had been spiked with 50 ng mL−1 cocaine. After extraction, the meshes were desorbed in solvent (MeOH/ ACN/formic acid, 80:20:0.1) for 20 min using an agitator at 1800 rpm. LC-MS/MS analysis was then used to determine the meshes’ ability to retain their extractive capacity after one exposure to the DART source. Once the solvent had been desorbed, the same meshes were cleaned with the MeOH/IPA/ ACN (50:25:25) mixture and exposed to thermal desorption
Figure 1. (a) Bare PEEK mesh without any coating attached; (b) mesh coated with the HLB particles; (c) SEM of bare mesh at 500× magnification; (d) SEM of coated mesh at 500× magnification; (e) close-up of HLB particles attached to mesh at 5K× magnification; and (f) coated mesh attached to the PBT support.
using the DART source (350 °C, 0.2 mm s−1) for a second time. The meshes were then preconditioned, and a PBS sample spiked with 50 ng mL−1 cocaine was once again used to perform the extractions. The amount extracted was determined using solvent desorption. LC-MS/MS analysis was conducted to determine the extractive ability of the meshes after a second exposure to the DART source. The entire procedure was repeated until the total number of thermal desorptions for the meshes was 5. The amount extracted after each thermal desorption was compared to the amount extracted from a mesh that was never exposed to thermal desorption. Extractive Procedure for Oral Fluid and Urine. Stock solutions for all studied analytes and their respective deuterated analogues were made, and the two matrices were spiked at the appropriate concentrations (0.5, 1, 2.5, 5, 10, 25, 50, 75, 100, 125, 150, 175, and 200 ng mL−1) for the calibration curve and (8, 40, 80, and 140 ng mL−1) for the validation points. The internal standard was spiked at a fixed level of 10 ng mL−1, and all spiking was conducted so as to keep the organic content in the matrices below 1% in order to simulate a “real” sample and to prevent any effects from the solvent during the extraction step.20 For the semiquantitative measurements of caffeine, a calibration curve for caffeine was made in PBS; this calibration curve consisted of nine calibration levels (i.e., 1, 5, 10, 50, 100, 250, 500, 750, and 1000 ng mL−1). Caffeine-C133, spiked at 50 ng mL−1, was used as internal standard. After spiking, both complex matrices, urine and OF, were gently agitated on a vortex at 200 rpm for at least 2 h to allow for proper equilibration between the analytes and the matrix. Prior to extraction, the meshes were briefly washed with H2O to remove the organic content which can affect the extractive capabilities 954
DOI: 10.1021/acs.analchem.7b04005 Anal. Chem. 2018, 90, 952−960
Article
Analytical Chemistry
deattachment occurred (Figure 1f). Another important point of consideration in assessing stability was whether the meshes would be able to withstand exposure to the DART source’s high temperatures. Thus, a stability test was conducted by exposing the meshes to the DART source at 250, 300, and 350 °C. Temperatures beyond 400 °C were not explored due the PEEK’s reported melting point of 343°.41−43 However, as noted by Fernandez’s group, the temperature indicated by the DART software is generally higher than the temperature of the region where the heated gas comes into contact with the coated surface;52 hence, it was not a concern that the material would be compromised at the set temperature of 350 °C. Indeed, the unravelling of mesh filaments was not observed (see Figure 1 and Figure S2f). Mesh repeatability was assessed by performing a rapid extraction (1 min) from a vial containing PBS solution spiked with 50 ng mL−1 cocaine. Analysis was performed via LC-MS/ MS to eliminate any errors associated with the desorption/ ionization step, and, as such, to enable a strict evaluation of the coating’s repeatability. To determine the amounts of cocaine extracted, the meshes were desorbed in a 80:20:0.1 MeOH/ ACN/formic acid solvent mixture already described elsewhere.24 After the first desorption, the meshes were cleaned twice using the above-mentioned cleaning solution mixture, and the obtained samples were additionally analyzed for carry-over. The results obtained for mesh repeatability (n = 5) can be seen in Figure S3. The amount of cocaine extracted by the meshes varied from 12.5 to 16 ng, with a mean of 14.4 ± 1.3 ng and a relative standard deviation (RSD) of 9.5% (without internal standard correction). These results were quite satisfactory given the short and nonautomated extraction process. The cleaning solutions were tested for carryover, but the amounts remaining on the mesh were below the LOQ of the experiment. The analytical performance of the PEEK meshes was further compared to that of stainless steel meshes32 (mesh opening, 250 μm; strand size, 90 μm; open area, 55%). For this comparison, both the SS and PEEK meshes were coated with the same ground particles and analyzed under the same DART parameters. A sample of PBS was spiked with 50 ng mL−1 heroin, and extraction was carried out for 1 min under the above-mentioned conditions (n = 3 for PEEK meshes; n = 3 for SS meshes). The two meshes were compared by monitoring the product ion of heroin (m/z 370 → 165), with the resultant ion chronograms revealing comparable analytical performance between PEEK and SS meshes (see Figures S4 and S5 in the Supporting Information). Assessment of Potential Reusability. Even though the plastic meshes are intended to be single use, the possibility of reusability was also investigated. As can be observed in Figure S6, extraction capability drops by approximately 37% after DART desorption. A one-tailed student’s t test and a one factor analysis of variance (ANOVA) was conducted (α = 0.05) to assess whether there is a significant difference between the meshes prior to and after exposure. The calculated p-values (
