Article pubs.acs.org/ac
Reusable Solid-Phase Microextraction Coating for Direct Immersion Whole-Blood Analysis and Extracted Blood Spot Sampling Coupled with Liquid Chromatography−Tandem Mass Spectrometry and Direct Analysis in Real Time−Tandem Mass Spectrometry Fatemeh S. Mirnaghi and Janusz Pawliszyn* Department of Chemistry, University of Waterloo, 200 University Avenue West, Waterloo, Ontario, N2L 3G1, Canada S Supporting Information *
ABSTRACT: Three different biocompatible polymers were tested and evaluated in order to improve the whole-blood biocompatibility of previously developed C18−polyacrylonitrile (C18−PAN) thin-film solid-phase microextraction (SPME) coating. Among all methods of modification, UV-dried thin PAN-over C18−PAN provided the best results. This coating presented reusable properties and reproducible extraction efficiency for at least 30 direct extractions of diazepam from whole blood [relative standard deviation (RSD) = 12% using external calibration and 4% using isotope dilution calibration]. The amount of absolute recovery for direct immersion analysis and based on the free concentration of diazepam in blood matrix was about 4.8% (desorption efficiency = 98%). The limit of quantitation (LOQ) for the developed solid-phase microextraction liquid chromatography−tandem mass spectrometry (SPME-LC−MS/MS) method for direct wholeblood analysis was 0.5 ng/mL. The optimized modification of the coating was then used for an extracted blood spot (EBS) sampling approach, a new sampling method which is introduced to address the limitations of dried blood spot sampling. EBS was evaluated using LC−MS/MS and direct analysis in real time (DART)−MS/MS, where, for a 5 μL blood spot, LOQs of 0.2 and 1 μg/mL, respectively, were achieved for extraction of diazepam.
S
application of biocompatible coatings can aid in minimizing the attachment of macromolecules on the surface of an open-bed SPME coating and in preserving the efficiency of the coating. The development of biocompatible SPME coating can be achieved through application of biocompatible extractive phases such as poly(dimethylsiloxane) (PDMS)3 and polypyrrole (PPY),4,5 or modification of the conventional SPME coatings with biocompatible polymers [i.e., polyacrylonitrile (PAN),6,7 and poly(ethylene glycol) (PEG)2,8]. The latter category provides the advantage of developing an unlimited number of extractive phases via modification of available SPME coatings. When SPME coating is used for whole-blood analysis, the phenomena of protein adsorption and cellular adhesion are two different issues which both should be addressed. Our studies indicated that, in spite of compatibility in plasma matrix, mostly all available lab-made or commercial biocompatible SPME coatings suffer from lack of reusability in whole blood due to irreversible blood cell attachment on the surface of the coating. The attached cells act as diffusion barriers on the coating surface and cause loss of extraction efficiency. However, several biocompatible coatings have been tested for long-term use in plasma;6,7,9,10 up to now “long-term
olid-phase microextraction (SPME) is an open-bed equilibrium-based system in which an extractive phase is coated on the outer surface of a substrate.1 One of the most beneficial features of an open-bed system is the opportunity of application of large sample volume without facing the limitation of breakthrough volume. This provides selective equilibriumbased extraction of analytes of interest, even those which are not successfully retained in an equivalent packed-bed system. Another beneficial aspect of an open-bed sorbent is the prevention of clogging and contamination of the extractive phase with macromolecules and particulate and elimination of additional time-consuming pretreatment steps before analysis (main limitations of packed-bed extractive phases). However, open-bed extractive phases may still face the challenge of adsorption of the macromolecules and particulates to the surface of the coating. The adhesion of macromolecules to the coating surface can affect the kinetics of extraction and manipulate the amount extracted by the extractive phase, even for one-time use. During recent years, this issue has been addressed through development of biocompatible SPME coatings. The term of biocompatibility in case of SPME coatings can be directed to two different terms: (i) preventing the adverse and/or toxic reactions in the living system (in vivo) and (ii) minimizing the adhesion of biological molecules such as proteins or blood cells to the surface of the coating (in vivo and in vitro).2 In both cases of in vivo and in vitro systems, © 2012 American Chemical Society
Received: June 30, 2012 Accepted: August 28, 2012 Published: August 28, 2012 8301
dx.doi.org/10.1021/ac3018229 | Anal. Chem. 2012, 84, 8301−8309
Analytical Chemistry
■
stability” in whole blood has not been studied. Our previous work reported on the development and optimization of thinfilm C18−polyacrylonitrile (C18−PAN) coating for at least 70 times reproducible usage for human plasma analysis,7 even though this coating showed irreversible blood cell attachments after several extractions from the whole blood. Therefore, the current study focuses on modification of C18−PAN SPME coating to improve compatibility with whole-blood matrix in long-term use. The surface chemistry and topography of the coating in contact with biological matrixes are important parameters that may impact protein adsorption and blood cell interaction.11 One common solution for minimizing the adsorption of blood cells is to fabricate a hydrophilic polymer layer as a high activating barrier on the surface which protects the surface against adsorption.12 For this purpose, in this study the modification of thin-film C18−PAN coating was evaluated using three different polymers [including polyacrylonitrile (PAN), 6,7 2-methacryloyloxy ethyl phosphorylcholine (MPC),13−15 and glycidol16). In addition, a new washing strategy after extraction was employed to improve reusability of the modified coatings in whole blood. Using an automated 96 thin-film SPME system, different modified coatings were evaluated in terms of the extent of blood cell attachment, reproducibility, extraction efficiency, and reusability for direct immersion whole-blood analysis in long-term usage. As another focus of the study, the optimal modified coating (UV-dried thin PAN-over C18−PAN) was used for evaluation and optimization of extracted blood spot sampling coupled with liquid chromatography−tandem mass spectrometry (LC−MS/ MS) and direct analysis in real time (DART)−MS/MS. Dried blood spot sampling (DBS) is a noninvasive sampling technique which presents remarkable advantages (e.g., uses a small blood volume and requires easy storage and shipment method);17−19 however, many studies have reported on the drawbacks of application of DBS.17,20−22 In order to overcome DBS limitations, the current work introduces an “extracted blood spot” (EBS) approach. EBS is the utilization of a biocompatible SPME coating (instead of filter paper/card in DBS) for blood spot (or other biofluids) sampling. In EBS, the three steps of preconcentration, cleanup, and extraction of analyte of interest all happen in one single step. The advantages of the EBS versus conventional DBS include (1) extraction of analytes and preservation against possible variation, (2) prevention of the adverse ion suppression/enhancement caused by filter paper and blood spot matrix, (3) no limitation with filter paper punching, hematocrit, and chromatographic effect, (4) simple sample preparation with less experimental steps, and (5) automation and high throughput. Direct analysis in real time is a fast method which provides immediate analysis and screening of samples without the need for sample pretreatment.23−25 However, the introduction of the real sample containing a complex mixture of matrix components may provide interferences and high background noise, thereby making qualitative and quantitative analysis difficult.26 For the first time, in this study the biocompatible modified SPME coating is utilized for cleanup and extraction of compounds (extracted blood spot sampling) from complex matrixes prior to DART. For the entire study, diazepam and diazepam-d5 were used as model analyte and isotopic-labeled analogue, respectively.
Article
EXPERIMENTAL SECTION
Chemicals and Materials. The details for chemicals and materials are described in the Supporting Information (section 1.1). Preparation of Modified Biocompatible C18−PAN SPME Thin-Film Coatings. C18−PAN SPME coatings were prepared with the method of spraying following the same formula and procedure reported in our previous work.7 C18− PAN coatings were then modified using different biocompatible polymers. Since the dipping method was found to be more uniform and reproducible (compared to spraying and brush painting methods), it was therefore used for covering the original C18−PAN for all methods of modifications. PAN-Over C18−PAN. PAN glue was prepared by dissolving 10% w/w PAN particles in dimethylformamide (DMF), followed by heating for 1 h at 90 °C. Different strategies were tested for modification of the C18−PAN coating using PAN glue: (i) a set of C18−PAN coatings (n = 6) were dipped in PAN glue (20 s) and dried at 180 °C for 2 min, (ii) two sets of C18−PAN coatings (n = 6 each) were immersed in PAN glue for short and long dipping times (20 and 60 s) and dried at 70 °C for 4 h, and (iii) two sets of thin and thick PAN-over coatings were prepared by dipping C18−PAN coatings (n = 6 each) in PAN glue for short and long dipping times (20 and 60 s), respectively. Then, they were dried for 30 min (each side) under ultraviolet (UV)] lamp light [emitting UVA (320−400 nm) and UVB (280−320) regions]. Methods of preparation of MPC-modified and glycidolmodified C18−PAN coatings are given in the Supporting Information (section 1.2). LC−MS/MS Condition. The LC−MS/MS condition is described in the Supporting Information (section 1.3). Automated 96 Thin-Film SPME System. The Concept 96 thin-film SPME device and 96 autosampler are discussed in detail elsewhere.7 In this study the static wash station of the original Concept 96 autosampler was modified to an agitating wash station. Figure S-1 in the Supporting Information illustrates the picture of modified Concept 96 autosampler. Automated SPME-LC−MS/MS Procedure for Direct Immersion Blood Analysis. The following procedure was performed for the evaluation of all types of modified C18−PAN coatings. The spiked whole-blood samples were incubated to reach equilibrium between the drug and the binding matrix in blood. The modified C18−PAN coatings were preconditioned for half an hour prior to extraction (agitating in methanol/water 50:50, v/v). Equilibrium extraction was performed from 1 mL of spiked whole blood or phosphate-buffered saline buffer (PBS), pH = 7.4, for 60 min (1000 rpm and 2.5 mm amplitude). After the extraction from blood, the coatings were washed for 20 s in Nanopure water while being agitated. Then, the thin-film SPME coatings were desorbed in 1 mL of acetonitrile/water 50:50 (v/v) for 40 min (1500 rpm, 1 mm amplitude). The 96 well plate containing the final extract was directly transferred into the LC−MS/MS autoinjector for analysis. Prior to the next extraction, the carryover of analyte on the coating was cleaned using a second desorption in 1 mL of acetonitrile/water 1:1 v/v (30 min, 1500 rpm, 1 mm amplitude). Extracted Blood Spot Sampling LC−MS/MS Procedure. The optimized UV-dried thin PAN-over C18−PAN coating was used for the extracted blood spot experiment. Using a micropipet, 5 μL of the spiked whole blood was spread 8302
dx.doi.org/10.1021/ac3018229 | Anal. Chem. 2012, 84, 8301−8309
Analytical Chemistry
Article
Figure 1. Direct immersion SPME-LC−MS/MS analysis using UV-dried thin PAN-over C18−PAN (thin) and MPC-modified C18−PAN for extraction of 100 ng/mL of diazepam from whole blood, n = 6 (quantitation based on external calibration). It should be noted that the reason for much lower recovery in whole-blood matrix compared to that of PBS (Figure S-5 in the Supporting Information) is due to the intensive diazepam− protein binding in whole blood.
ionization takes place (Figure S-2d in the Supporting Information). For both bare mesh and coated mesh, 5 μL aliquots of the sample were deposited on the meshes using a micropipet. After 5 min of extraction the coated meshes were rinsed for 5 s with Nanopure water. Then coated meshes containing sample spots were allowed to fully dry (using a blow drier) before being introduced to the DART source for analysis. The following was used as optimum parameters in DART: source gases, helium and nitrogen; source gases pressure, 80 psi; desorption temperature, 250 °C; discharge needle voltage, 3000 V; grid voltage, 350 V; sampling time, 30 s; gap time, 5 s; the distance between the DART gun and ceramic transfer tube, 11 mm; transmission module sampling speed, 0.2 mm/s. The details of MS/MS conditions for DART are described in the Supporting Information (section 1.4).
on the PAN-over C18−PAN coating. Next, the blood spots were left for 2 min to enable the extraction of analytes by the coating. Immediately following this, the blades were assembled into the 96 blade holders and mounted on the 96 autosampler to be automatically washed for 20 s in Nanopure water (along with agitation). The coatings were then desorbed in 1 mL of acetonitrile/water 1:1 v/v for 40 min (1500 rpm, 1 mm amplitude). The coatings could also be refrigerated and stored for a later desorption if any transportation is necessary. Finally, the samples were analyzed using the same LC−MS/MS conditions as described for direct immersion whole-blood analysis (Supporting Information, section 1.3). Before reusing the coating, a second desorption in 1 mL of acetonitrile/water 1:1 v/v (30 min, 1500 rpm, 1 mm amplitude) was performed to clean the carryover from the previous desorption. Coating Preparation for EBS-DART−MS/MS. Initially, the 74 × 74 stainless steel bare meshes were coated with a layer of C18−PAN using the method of brush painting as in the following procedure: The method of preparation of C18−PAN slurry is already reported elsewhere.7 The stainless steel bare meshes were cut to the dimensions of 1.2 cm × 15 cm to fit into the transmission module. Then, using a painting brush whose width was approximately the same as that of the mesh, a thin layer of the coating slurry was spread on the surface of the mesh. Next the meshes were dried, followed by drying in the oven (180 °C) for 2 min.7 In the next step, a layer of PAN was placed on top of the coatings (using brush painting) followed by drying under UV light for 30 min each side. Figure S-2, parts a and b, in the Supporting Information compares the bare mesh versus coated mesh. EBS-DART−MS/MS Condition. A DART standardized voltage and pressure (DART SVP) model ion source (IonSense, Inc., Saugus, MA) was coupled with a Applied Biosystems API 4000 triple-quadrupole mass spectrometer via a Vapur interface (IonSense, Inc.). For EBS-DART−MS/MS analysis, the optimization of the parameters and analysis of the samples were performed by spotting the samples on the PANover C18−PAN coated mesh that is held in a sampling module called a transmission module (IonSense, Inc.). The transmission module contains a series of 10 holes with 7 mm diameter (placed every 14 mm) which sandwiches and holds the mesh onto which samples can be pipetted (Figure S-2c in the Supporting Information).27 A linear rail was capable of moving the module through the gap between the DART source and the ceramic ion collection and transfer tube, where the
■
RESULTS AND DISCUSSION
Evaluation of Modified Biocompatible C18−PAN SPME Coating Using Direct Blood Analysis. Evaluation of modified C18−PAN coating using all types of biocompatible polymers showed that UV-dried thin PAN-over C18−PAN provided the best biocompatibility and reusability for at least 30 extractions from whole blood. PAN-Over C18−PAN. Many previous studies have reported on the application of PAN polymer for the development of biocompatible surfaces and reported significant reduction in the interaction of matrix media with the surface.28−31 Structure of PAN polymer is shown in Figure S-3a in the Supporting Information. Different drying strategies were tested for modification of the C18−PAN coating using a PAN layer: among all methods UV-dried thin PAN-over C18−PAN showed the best performance. The performance of all PANover coatings under different drying conditions was compared in the following. (i) Drying at 180 °C: this modification resulted in creation of small holes on the surface of the coatings, possibly due to vigorous evaporation of the solvents while drying at high temperature. Although the amount of blood cell attachment decreased by this modification, it resulted in localized red blood cell attachments in these small holes (Figure S-4, parts a and b, in the Supporting Information). (ii) Drying at 70 °C: modified coatings made with this drying strategy (for both 20 and 60 s dipping) presented improved prevention of red blood cell attachment on the surface of the coatings (up to 16 extractions). However, this drying strategy resulted in some unsatisfactory changes in the coating structure. 8303
dx.doi.org/10.1021/ac3018229 | Anal. Chem. 2012, 84, 8301−8309
Analytical Chemistry
Article
Figure 2. Scanning electron microscopy (SEM) micrographs of (a) original C18−PAN before exposure to blood and (b) UV-dried thin PAN-over C18−PAN before exposure to blood.
pictures of PAN-over C18−PAN coating before and after exposure to blood (after rinsing with water), indicating the effectiveness of the biocompatible modification and wash step for cleaning the blood cells. The dominant characteristic of the polyacrylonitrile [poly(vinyl cyanide)] layer is the presence of strongly polar nitrile groups. The CN functional groups are perfectly suited to participate in hydrogen bindings with water molecules, forming a hydration layer which inhibits interaction of blood cells and proteins with the coating surface.32 However, in the cases of MPC- and glycidol-modified C18−PAN coatings the protective layer on the surface does not uniformly exist (Figure S-6, parts b and c, in the Supporting Information). Optimization of SPME Parameters for UV-Dried Thin PAN-Over C18−PAN SPME Coating for Direct Immersion Whole-Blood Analysis. The optimized SPME parameters for the PAN-over C18−PAN coating were comparable with that of original C18−PAN coating for extraction of diazepam (30 min of preconditioning, 60 min of extraction, and 40 min of desorption), with the exception of the wash step.7 A detailed description of optimized parameters for UV-dried thin PANover C18−PAN SPME coating is discussed in the Supporting Information (section 2.2). The exposure of PAN-over C18− PAN coating to the whole blood results in loose attachment of the blood cells to the surface. Therefore, the high-speed agitation during the wash step was found necessary for complete rinsing of the blood cells from the surface. Optimization of the wash step showed that a 20 s wash in water is enough to clean the blood cells from the surface of the coating. A shorter and longer wash step resulted in red blood cell attachments and higher chance of analyte loss, respectively. In order to evaluate the possible analyte loss during 20 s, the wash solution was directly injected into the LC−MS/MS system. The amount of absolute loss of diazepam was found to be 0.3% (RSD = 13%, n = 6). Since the coatings are reusable, prior to the next extraction a second desorption was performed to clean the trace residual that remained in the coating from the previous desorption (carryover). The percentage of carryover was calculated based on the ratio of the mass (ng) extracted from the second desorption over the total mass (ng) from the first and the second desorption. The percent carryover was determined to be 2.0% (desorption efficiency = 98%) for PANover C18−PAN SPME coating. Evaluation of a third desorption did not result in any detectable signal, confirming that a second 30 min desorption is enough for cleaning the carryover residues from the coating.
(iii) Drying under UV light: in order to resolve the problem of unwanted coating changes at high temperatures, UV light (30 min each side) was used for curing of two sets of thin and thick PAN-over coatings. For both sets of coating (n = 6 for each) the improvement of biocompatibility in whole blood was significant. However, the thinner PAN layer provided much better coating efficiency and quality. As shown in Figure 1, thin PAN-over C18−PAN coating resulted in reproducible efficiency [relative standard deviation (RSD) = 12%] for at least 30 extractions from whole blood without any blood cell attachment (Figure S-4c in the Supporting Information). Absolute recovery was defined as the ratio of extracted mass versus the initial spiked mass in the sample, multiplied by 100. In addition to the improved blood stability, the thinner modified coating provided higher extraction efficiency: the absolute extraction recovery for thin PAN-over coating (95% ± 5% for PBS) is comparable with that of original C18−PAN.7 However, the thicker modification provided a drop in extraction efficiency. Figure S-5 in the Supporting Information compares the coating efficiency of all modified coatings with that of original C18−PAN. It is assumed that the thin PAN layer acts as a porous bag which holds the C18−PAN coating (Figure 2, parts a and b), without changing the original properties. However, the thicker PAN layer probably blocks the available surface pores on the coating and acts as a barrier for the analytes (Figure S-6a in the Supporting Information). Therefore, this changes the efficiency of the original coating and results in lower recoveries even when longer extraction time is provided. Detailed descriptions for evaluation of MPC- and glycidolmodified C18−PAN coatings are described in the Supporting Information (section 2.1) Coating Topography and Blood Biocompatibility. In case of the original C18−PAN coating, there is a high ratio of stationary phase particles to the polyacrylonitrile content, and PAN mainly works as a porous glue to bind the particles together. As a result, there is not a full coverage protection on top of the particles to prevent blood cell attachments (Figure 2a). In contrast, addition of another biocompatible layer on the coating can provide a protective layer on the coating’s outer surface. The scanning electron microscopy (SEM) micrographs of original C18−PAN coating and UV-dried thin PAN-over C18−PAN are shown in Figure 2. The SEM micrographs indicate that the UV-dried PAN-over modification properly provides the required protective layer on the C18−PAN coating surface. There was no difference between the SEM 8304
dx.doi.org/10.1021/ac3018229 | Anal. Chem. 2012, 84, 8301−8309
Analytical Chemistry
Article
Table 1. Recovery, Reproducibility, and Validation for Direct Immersion SPME-LC−MS/MS and Extracted Blood Spot LC− MS/MS
direct whole-blood LC−MS/ MS extracted blood spot LC− MS/MS (5 μL, 2 min) a
% absolute recovery
interday (RSD %) isotope dilution/external calibration
intraday (RSD %) isotope dilution/external calibration
% absolute matrix effect
LOD (ng/ mL)
LOQ (ng/ mL)
4.8a
4/12a
1/10a
105
0.2
0.5
2.1b
5/14b
2/12b
98
70
200
linearity (R2) 0.996 0.993
n = 4 for 100 ng/mL diazepam in whole blood. bn = 4 for 5 μL of 1 μg/mL diazepam in blood.
Reproducibility and Validation for Direct Immersion Whole-Blood Analysis LC−MS/MS. Reproducibility of the assay was evaluated based on interday and intraday relative standard deviation and was compared for both isotope dilution and external calibration methods. Absolute matrix effect was calculated based on the ratio of peak area of the blank blood extract spiked with the known concentration (50 ng/mL) of analyte to the peak area obtained from standard solution of analyte with the same concentration, multiplied by 100. Matrix effects larger than 120% and smaller than 80% represent significant ionization enhancement and suppression for a given analyte, respectively.33,34 For the entire study, the limit of detection (LOD) and the limit of quantitation (LOQ) were calculated based on 3 × S/N (signal-to-noise), and 10 × S/N, respectively. The S/N was measured manually on chromatogram printout and was confirmed based on four replicated analyses at the LOD and LOQ levels. Table 1 reports on absolute recovery, inter- and intraday reproducibility, absolute matrix effect, LOD, LOQ, and linearity for equilibrium extraction of diazepam for direct immersion whole-blood analysis. Extracted Blood Spot Sampling. Optimization of EBS Parameters for LC−MS/MS Analysis. Using the current geometry of the thin-film (blade) coating (Figure S-7 in the Supporting Information), a blood spot volume as low as 3 and up to 20 μL can be used for EBS sampling. However, this volume can be increased by increasing the surface area of the coating. In this work, a 5 μL sampling volume was used for the entire EBS study. It is important that the time period for extraction of analytes after deposition of the blood spots on the coatings be optimized. Since the coating is designed to be reusable, the extraction time should be short enough in order to prevent complete drying of the proteins and blood cells on the coating and long enough to obtain a reproducible recovery. Our studies showed that a 2 min extraction addresses both issues. The optimization of wash, desorption, and carryover steps for extracted blood spot sampling is similar to that of the direct immersion procedure: a 20 s wash step along with agitation is optimal for efficient washing of the matrix from the coating. In addition, 40 min of desorption in acetonitrile/water 1:1 v/v (1500 rpm, 1 mm amplitude) was used for efficient desorption of analytes with minimum carryover. Before reusing the coating, a 30 min second desorption was optimized to clean the 2.0% carryover from the previous desorption. (See above for the description and the formula for calculation of carryover.) Quantitation, Reproducibility, and Validation of EBS-LC− MS/MS. Since EBS deals with the concept of microextraction, it benefits from having a small pre-equilibrium extraction time, and consequently achieves a small but reproducible recovery. In order to obtain precise quantitative results, it is highly
recommended that the proper quantitative calibration method be applied for EBS-LC−MS/MS analysis. The reproducibility of the assay was compared using two quantitative methods (matrix-matched external calibration and isotope dilution calibration). As expected, isotope dilution provided significant improvement in reproducibility. Therefore, in order to obtain higher precision for pre-equilibrium sampling, isotope dilution was used as the quantitation method for this study. The recoveries of diazepam for tested spiking levels of 0.5 and 50 μg/mL were 97% ±2% and 102% ±3% when isotope dilution was employed for quantitation (n = 4). Table 1 demonstrates recovery, inter- and intraday reproducibility, LOD, LOQ, and linearity for EBS-LC−MS/MS analysis when using a 5 μL volume of blood and 2 min extraction time. The limit of detection and quantitation can be improved by using a larger volume of the sample. Direct Analysis in Real Time. Optimization of Coating Preparation for EBS-DART−MS/MS. For the DART application, the thickness of the coated mesh should be very thin in order to prevent complete blocking of the mesh holes with the coating and to allow transmission of heated gas through the coating for efficient desorption and ionization of analytes. In addition, the method of coating preparation should be reproducible. Among different methods of coating preparation (dipping, brush painting, spraying), brush painting was found to be the most applicable and reproducible method for preparation of a thin layer of coating on mesh for the EBSDART system. In order to ensure transmission pathways through the coating, 10 tiny holes were made on the spotting area of the coated mesh using a very small needle, keeping the hole patterns similar for all the coatings. Optimization of EBS for DART−MS/MS. An extraction time much longer than 5 min resulted in difficulties in washing the blood from the coated mesh surface. Therefore, 5 min was used as an optimal extraction time for the entire EBS-DART−MS/ MS studies. After 5 min of extraction, a minimum 5 s pressured stream of water (using a wash bottle) was enough to wash the blood from the surface of coated meshes. Since the application of the wash step for the bare mesh resulted in complete rinsing of the analyte along with blood from the mesh, no wash step was applied for bare mesh. Next, in order to prevent any variation in analysis, the meshes should be fully dried before introduction to the DART source. For the samples which were not let to become fully dried, higher RSD values were obtained. Four different drying methods were evaluated: drying at room temperature (30 min), drying with KimWipe paper tissue (wiping to absorb water), and drying with hot or cold air using blow dryer (3 min). In our study, all four methods of drying were comparable in terms of recovery and reproducibility; however, the last three methods were found to be much faster. Very short sampling time in DART (30 s) does not provide full desorption of analytes from the coating. Therefore, before 8305
dx.doi.org/10.1021/ac3018229 | Anal. Chem. 2012, 84, 8301−8309
Analytical Chemistry
Article
reusing the coatings, DART can be used for complete desorption of residual analytes in longer exposure time, or the coatings can be vortexed in acetonitrile/water 1:1 v/v for 40 min. Quantitation and Validation of EBS-DART−MS/MS. The quantification of analytes in DART−MS analysis is usually determined by matrix-matched external calibration. However, many studies have been reported on the limitation of ambient MS analysis in DART with relatively high RSD for signal intensities within repeated DART−MS measurements.26,35 Isotope-labeled internal standards can significantly improve reproducibility by compensating for signal fluctuation and unavoidable matrix effect which usually happens in ambient ionization techniques.26,36 In this study matrix-matched isotope dilution calibration was used for quantitation of analysis. Figure S-8 in the Supporting Information illustrates isotope dilution calibration for diazepam in whole blood using the EBS-DART− MS/MS system. In addition, Table 2 reports on linearity,
Table 3. Reproducibility for DART−MS/MS and EBSDART−MS/MS reproducibility (RSD, %) isotope dilution/external calibrationa bare mesh matrix solvent (acetonitrile/ water 1:1 v/v) PBS buffer whole blood
recovery (%) (2 μg/mL spiking level)a recovery (%) (50 μg/mL spiking level)a LOD (μg/mL) LOQ (μg/mL) linearity (R2) linearity (μg/mL)
108
102
93
0.3 1 0.9965 1−200
0.3 1 0.9981 1−50
intraday (n = 4)
interday (n = 4)
intraday (n = 4)
4/24
3/21
3/20
3/22
5/30 4/40
4/33 5/35
4/25 3/21
3/30 2/20
5 μL of 5 μg/mL diazepam (and diazepam-d5 in case of isotope dilution) in whole blood.
matrix components. The main goal for employing the coated mesh instead of bare mesh and the combination of EBS with DART was the normalization of the matrix and cleaning up of the sample in order to minimize the matrix effect. To evaluate the matrix effect, the full scan Q1 mass spectra of DART−MS analysis of 5 μL of a 10 ppm diazepam blood spot sample on coated mesh was compared with that of bare mesh. Figure 3a illustrates a significant suppression of diazepam (285.1) signal for bare mesh due to the existence of matrix inferences, indicating the necessities for application of a proper sample cleanup and selective mode of MS detection (such as multiple reaction monitoring as the applied mode in this study). In contrast, EBS sampling aids for elimination of interferences and results in much cleaner mass spectra (Figure 3b). Furthermore, as shown in Figure 4, there is a significant shift in the matrixmatched external calibration curve of DART−MS/MS (bare mesh) compared with that of EBS-DART−MS/MS (coated mesh). This can be explained due to adverse effect of the matrix component on ionization of the analyte on the bare mesh or possible preconcentration of the analyte on the coated mesh, or both. Further investigations on the extent of these aspects are under process in our laboratory.
bare mesh
96
interday (n = 4)
a
Table 2. Recovery and Validation for EBS-DART−MS/MS UV-dried PAN-over C18− PAN
PAN-over C18−PAN coated mesh
a
Recoveries were calculated based on matrix-matched isotope dilution calibration when extracting 5 μL of 5 μg/mL diazepam in whole blood (n = 4).
recovery, and limits of detection and quantitation of EBSDART−MS/MS sampling versus regular DART−MS/MS for extraction of diazepam. Recoveries were calculated based on matrix-matched isotope dilution calibration. As shown by the results, EBS provided improvement in the linear range when compared to bare mesh. It should be noted that the reported LOD and LOQ values are based on 5 μL of blood sample and considering >95% diazepam−plasma protein binding (amount extracted by SPME is proportional to the free concentration of analyte). Reproducibility of EBS-DART−MS/MS. Inter- and intraday reproducibility of the assay was evaluated for bare and coated mesh in different matrixes (acetonitrile/water 1:1 (v/v) solvent, PBS, and blood), using external and isotope dilution calibration. Table 3 demonstrates that, in all cases, isotope dilution has shown a significant improvement in reproducibility of results via normalization of analyte response. The reason for relatively higher RSD for external calibration for all three matrixes and both types of meshes could be explained due to the following reasons: (i) variations in position of the mesh in the transmission survey, (ii) random distribution of the sample spot on meshes, (iii) variation of the thickness of the coated meshes, and (iv) the possibility of a leaky Vapur interface on the MS. Even though, even using external calibration, the reproducibility of the assay is improved using coated mesh compared to bare mesh. Evaluation of Matrix Interferences in EBS-DART−MS/MS. Evaluation of matrix effect is crucial, since reliability, accuracy, and precision of the analysis could be influenced by interfering
■
CONCLUSION AND FUTURE DIRECTIONS Attachment of blood cells and macromolecules to the coating can affect the reliability of the extraction, even for single usage. The PAN-over C18−PAN 96 blade SPME system can be confidently reused (for 30 extractions) in whole blood for both screening applications and quantitative analyses without the risk of contamination. This biocompatible coating has the potential to be applied for analysis of other complex biological matrixes such as tissue and food samples. As a future approach, biocompatible PAN-over C18−PAN SPME coating offers the chance for reliable at vivo (instant sampling of biofluids in the field at the time they leave the living system) and in vivo applications using other geometries of coating. Furthermore, the reusability of the coating for 30 extractions provides improved cost efficiency per experiment. EBS benefits from advantages of DBS and presents solutions for limitations of DBS. EBS-LC−MS/MS integrates sampling and sample preparation into a single step, and an automated highthroughput system facilitates the preparation of 96 EBS samples in approximately 42 min which corresponds to
