Fully Automated Electro Membrane Extraction Autosampler for LCâMS

Subscriber access provided by UCL Library Services

Article

A fully automated electro membrane extraction autosampler for LCMS systems allowing soft extractions for high throughput applications David Fuchs, Stig Pedersen-Bjergaard, Henrik Jensen, Kasper D. Rand, Steen Honoré Hansen, and Nickolaj Jacob Petersen Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b01243 • Publication Date (Web): 29 May 2016 Downloaded from http://pubs.acs.org on May 31, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 1 321x571mm (300 x 300 DPI)

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 7 123x76mm (300 x 300 DPI)


Page 2 of 11

Page 3 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


A fully automated electro membrane extraction autosampler for LCMS systems allowing soft extractions for high throughput applications David Fuchs1, Stig Pedersen-Bjergaard1,2, Henrik Jensen1, Kasper D. Rand1, Steen Honoré Hansen1 and Nickolaj Jacob Petersen1 1

Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, 2100 Copenhagen, Denmark 2 School of Pharmacy, University of Oslo, PO Box 1068 Blindern, 0316 Oslo, Norway ABSTRACT: The current work describes the implementation of electro membrane extraction (EME) into an autosampler for high throughput analysis of samples by EME-LC-MS. The extraction probe was built into a luer lock adapter connected to a HTC PAL® autosampler syringe. As the autosampler drew sample solution, analytes were extracted into the lumen of the extraction probe and transferred to a LC-MS system for further analysis. Various parameters affecting extraction efficacy were investigated including syringe fill strokes, syringe pull up volume, pull up delay and volume in the sample vial. The system was optimized for soft extraction of analytes and high sample throughput. Further, it was demonstrated that by flushing the EME-syringe with acidic wash buffer and reverting the applied electric potential, carry-over between samples can be reduced to below 1 %. Performance of the system was characterized (RSD: < 10 % R2: 0.994) and finally, the EME-autosampler was used to analyze in-vitro conversion of methadone into its main metabolite by rat liver microsomes and for demonstrating the potential of known CYP3A4 inhibitors to prevent metabolism of methadone. By making use of the high extraction speed of EME, a complete analytical workflow of purification, separation and analysis of sample could be achieved within only 5.5 minutes. With the developed system large sequences of samples could be analyzed in a completely automated manner. This high degree of automation makes the developed EME-autosampler a powerful tool for a wide range of applications where high throughput extractions are required before sample analysis.

INTRODUCTION A typical bioanalytical workflow involves a sample preparation step followed by separation and detection of analytes of interest. For many analytical procedures, separation and detection of a large number of samples can routinely be done without manual handling by the introduction of an autosampler, thus leaving the sample preparation step as a main bottleneck. A sample preparation step preceding the analysis of samples is often required to remove matrix components that are detrimental for the analysis of compounds of interest or to enrich the anlayte. Conventional and most commonly used sample preparation methods involve solid phase extraction (SPE), solid phase micro-extraction (SPME), protein precipitation (PP) and liquid-liquid extraction (LLE). Electro membrane extraction (EME) is a relatively new sample preparation method first described by PedersenBjergaard et al. in 20061. It makes use of an electric potential applied across a supported liquid membrane (SLM) to extract analytes from a sample solution through the SLM and into an acceptor solution. The SLM consists of an organic solvent

such as nitrophenyl octyl ether (NPOE) or octanol immobilized in a porous polymeric support. Not at least due to its high applicability for miniaturization 2-4 and its fast extraction speed 5-7 , EME has received growing attention as a viable alternative to conventional sample preparation methods. Various attempts towards automation and inline coupling of EME and related sample preparation methods to analytical systems have been reported. They include direct coupling of polymer inclusion membranes (PIMs) and SLMs to CE 8,9, electro extraction (EE) and SLM microchip extraction coupled to LC and LC-MS 10,11 as well as various studies describing the direct coupling of EME to CE, MS and LC-MS 4,12-16. In a more recent publication Drouin et al. describes an EME device used for the extraction of peptides into a dynamic acceptor phase. In Drouin’s setup acceptor phase is constantly renewed and electrolysis in the acceptor phase can therefore be avoided leading to enhanced recovery of analytes.17 Although the aforementioned approaches represent valuable contributions towards a higher degree of automation of EME, they show limited applicability for high throughput applica-



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

tions. This is mainly caused by relatively long extraction times, limited re-usability of the extraction membranes or the need for manual exchange of sample solutions between extractions. The current work represents a big step towards complete automation of an entire analytical workflow. It describes the implementation of an EME probe into a commercial autosampler. The EME-autosampler was directly coupled to a LC-MS system and therefore allowed completely automated purification and analysis of a large number of samples without any manual sample handling. The various parameters affecting extraction performance of the EME-autosampler were investigated and optimized for soft extractions of target analytes and high sample throughput. For soft extractions, only small amounts of analytes are extracted and therefore re-analysis of the same sample solution is possible where necessary. Finally, the EME-autosampler was applied to study in-vitro the conversion of methadone into its main metabolite 2-ethylidene1,5-dimethyl-3,3-diphenylpyrrolidine (EDDP) as well as to measure the potential of various CYP3A4 inhibitors to prevent the metabolism of methadone.

Page 4 of 11

lock adaper the EME probe was located (Figures 1B and1C). A hollow fiber made of porous polypropylene (Plasmaphan P1LX, Membrana, Wuppertal, Germany) was used as support for the liquid membrane. The hollow fiber had a 330 µm i.d., 150 µm wall thickness and 0.4 µm pore size and was on both ends connected to 530 mm long fused silica capillaries (Polymicro Technologies, 100 µm i.d., 245 µm o.d.). Seal tight connection of hollow fiber and fused silica capillaries was achieved by heat shrinking of the intersection of fiber and capillary leaving approximately 3 mm of intact porous fiber in between the two capillaries. The heat applied to each end of the probe (approximately 200 ○C) shrunk the diameter of the hollow fiber by closing the pores, providing seal tight connection of the fiber to the capillaries.

EXPERIMENTAL SECTION Chemicals and Reagents Ketoconazole, omeprazole, fluoxetine hydrochloride, βnicotinamide adenine dinucleotide 2′-phosphate reduced tetrasodium salt hydrate (NADPH) and NPOE were obtained from Sigma-Aldrich (St. Louis, MO). Methadone hydrochloride was obtained from Nordisk Droge og Kemikalie A/S (Copenhagen, Denmark). Rat liver microsomes (male SpragueDawley, pooled; 20 mg/mL) were purchased from Corning/BD Biosciences (Woburn, MA). Formic acid (HCOOH) was obtained from Merck (Darmstadt, Germany) and LC grade acetonitrile (ACN) was purchased from TH. Geyer (Stockholm, Sweden). All used water was deionized (dH2O, 18 MΩ cm) with a Millipore Direct-Q 3 UV system (Billerica, MA). 100 mM potassium phosphate buffer/5 mM MgCl2 (pH 7.4), 1 mg/mL fluoxetine stock solutions, 10 mM of NADPH, 10 mM of HCOOH, 0.05 % HCl, 0.5 % HCl, 0.05 % HCOOH and 0.5 % HCOOH solutions were prepared in dH2O. 10 mg/mL ketoconazole stock solutions were prepared in methanol and 1 mg/mL omeprazole stock solutions were prepared in 50 % methanol. 1 mg/mL methadone stock solutions were prepared in 10 % ethanol. Working solutions of methadone, omeprazole, fluoxetine and ketoconazole were prepared by adequate dilutions of stock solutions in 100 mM potassium phosphate buffer/5 mM MgCl2 (pH 7.4). EME syringe preparation A schematic of the EME-syringe’s luer lock adapter is shown in Figure 1A and 1B. The used adapter was a female luer to female 10-32 fitting adapter (IDEX Health and Science, Oak Harbor, WA). As EME-autosampler syringe needle, a 51 mm long fused silica capillary tubing (Polymicro Technologies, Phoenix, AZ, i.d. 250 µm, o.d. 350 µm) was connected via a flush nut ferrule and a microtight peek tubing sleeve (color code: orange, i.d. 330 µm, IDEX, Oak Harbor, WA) to the luer lock adapter. A non-conducting fused silica capillary instead of a metal needle was used to ensure electrical insulation between the syringe and the autosampler. Inside the luer

Figure 1: A: Side view illustration of the electro membrane luer lock adapter. B: Top view illustration of the electro membrane luer lock adapter. C: Light microscopy photograph of a porous hollow fiber used as support for the electro membrane connected on both ends to fused silica capillaries.

The probe was placed inside the luer lock adapter through 2 holes (500 µm diameter each) on opposite ends of each other.


Page 5 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


A third hole in the luer lock adapter was used to place the end of a metal wire (80 % nickel/20 % chromium, Omega Engineering, Stamford, CT) close to the SLM. The metal wire was required to apply electric potential close to the SLM. Hot melt glue (Scotch-Weld hot melt adhesive 3764-TC-Q, 3M, St. Paul, MN) was used for leak tight closing of the holes. A small droplet of NPOE, used as supported liquid membrane, was applied onto the hollow fiber inside the luer lock adapter. NPOE immediately filled the pores of the fiber by capillary forces which could be visually inspected as the hollow fiber’s color changed from white to transparent when NPOE was applied. Excess of organic solvent used as liquid membrane was removed from the fiber using a medical wipe. The luer lock adapter was then connected to a 1 mL glass syringe (Gastight 1001, Hamilton, Bonaduz, Switzerland). Due to the hydrophobic nature of the porous hollow fiber, immobilization of NPOE in the pores of the hollow fiber proved to be very stable. At the applied acceptor phase flow rate (20 µL/min) the same liquid membrane could be used for a whole working day (> 50 extractions) without any apparent change in extraction performance. To remove NPOE from the hollow fiber, the inside of the probe was flushed with ethanol at the end of each working day applying a sufficiently high flow rate to push the liquid membrane out of the pores of the hollow fiber. For all performed experiments one single probe was used. This was done to avoid possible variations in the manual probe production process (such as slightly different membrane lengths) which could potentially impair repeatability of extractions. Instrumental setup The EME syringe was mounted to a HTC PAL autosampler (CTC Analytics, Zwingen, Switzerland) and connected to a LC-MS (Agilent 1100 series LC/MSD ion trap, Agilent Technologies, Santa Clara, CA) system via a 10 port-switching valve (Valco instruments Co. Inc., Schenkon, Switzerland) as illustrated in Figure 2. In addition to the injection unit to which the EME-syringe was mounted, the HTC PAL autosampler was equipped with a control terminal, a temperature controlled microplate stack (sample tray holder) and a wash station. Coupling of the EME syringe to the 10 port switching valve was done by connecting the two fused silica capillaries from the EME-syringe to the switching valve with microtight peek tubing sleeves (color code: blue, i.d. 280 µm, IDEX, Oak Harbor, WA). In addition, a syringe pump for pumping of acceptor phase (10 mM HCOOH) through the capillaries was connected to the 10 port switching valve. Further, a 20 µL sample loop was used to which anaytes were transferred after extraction via the acceptor phase flow. The 10 port switching valve switched between 2 positions: a sample extraction/injection position in which acceptor phase bypassed the EME syringe thereby leaving a stagnant acceptor phase in the lumen of the electro-membrane and a sample load position in which acceptor phase was pumped with a flow rate of 20 µL/min through the fused silica capillaries (Figure 2). In this position sample previously extracted (while the switching valve was in the extraction/injection position) was loaded to the sample loop. As illustrated in Figure 2, the EME-syringe was further coupled to a HVS448, high voltage sequencer (Labsmith, Livermore, CA) via an electric wire to establish electric contact required for EME.

Figure 2: Setup of the EME-autosampler. The EME-syringe was mounted on a HTC PAL autosampler and coupled to a LC-MS system via a 10 port switching valve. Electric potential was applied using a voltage sequencer capable of switching between positive and negative electric potential.

Communication of the LC-MS system with the HTC PAL autosampler, 10 port switching valve and high voltage sequencer was established via an external relay contact interface board (Agilent G1351A-66500), installed to the HPLC pump module. The contact board was used to send out sequenced TTL (transistor-transistor logic) signals to the HTC PAL autosampler, 10 port switching valve and high voltage sequencer thereby synchronizing the operational steps of the instruments. As part of the LC-MS system, the relay contact interface board was controlled by the HPLC software Chemstation (B.01.03). Sample analysis with the EME-LC-MS system For better illustration of a sample analysis sequence, Video file S-1 and Table S-1 (supplementary information) describe the various operational steps carried out by the instruments. In a first step, the EME-autosampler was moved to a vial in the sample tray holder and extracted analytes from the sample solution through the supported liquid membrane and into the acceptor phase in the lumen of the EME-probe. During the extraction step an electric potential of + 200 V was applied by the voltage sequencer at the electrode in the luer lock adapter relative to the 10 port switching valve which was grounded. An electric field across the SLM was thereby established as the metal parts of the 10 port valve were in contact with the acceptor phase. Extraction was done by pulling sample solution up and down the EME-syringe for a number of cycles. During this process the sample solution passed the electro membrane and analytes were extracted. After extraction was finished, extract was moved to the sample loop by switching the 10 port valve to the sample load position. Simultaneously,



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the EME-syringe was moved to the autosampler wash station. Washing of the electro-membrane was necessary to backextract analytes trapped in the liquid membrane which would otherwise lead to substantial carry-over between samples. Washing was done by reversing the applied electric potential to – 200 V and flushing the EME syringe with acidic wash solution. During washing of the EME syringe, loaded extract was injected to the LC-MS system by switching the 10 port valve back to the sample extract/inject position. After sample analysis by the LC-MS system, the whole procedure was automatically reinitiated by extracting analytes from a new sample vial. LC-MS method For chromatographic separation of extracted analytes a reversed phase column (Zorbax SB-C18, Agilent, 2.1 x 30 mm, 1.8 µm particle size) coupled to an Agilent 1100 series HPLC system was used. Separation was performed for 5 minutes at a flow rate of 350 µL/min by isocratic elution using 30 % ACN/0.1 % HCOOH as mobile phase. The column was operated at 40 ○C. Eluting analytes were analyzed using an Agilent 1100 series LC/MSD ion trap operated in full scan mode. Sample loop loading time investigation For investigation of the time required to transfer the extract from the lumen of the SLM inside the EME-syringe to the sample loop of the 10 port switching valve, extractions from 10 µM methadone solutions were repeatedly performed. After each sample extraction, the extract was transferred for varying time periods (15 s – 105 s) to the sample loop by switching the 10 port valve to the sample load position. The optimal sample loop fill time (time the 10 port switching valve was in the sample load position) at an acceptor phase flow rate of 20 µL/min was determined by integrating the eluting methadone peaks and comparing the obtained peak areas. Optimization of various extraction parameters on sample recovery Several extraction parameters, as described in Table 1, were investigated with respect to extraction efficiency. Table 1: Investigated extraction parameters Extraction parameter

Elucidation of extraction parameter

Experimental range

fill strokes

number of repetitions the EMEsyringe draws and pushes back sample into the sample vial

2, 4, 8 and 16 pull up cycles

pull up volume

volume of sample drawn and pushed back into the sample vial by the EME syringe

50 µL, 100 µL

pull up delay

time, sample rested in the EMEsyringe after each pull up

0 s, 5 s, 10 s

sample volume

volume of sample in the sample vial

150 µL, 250 µL, 500 µL

For investigation of the various extraction parameters, methadone (10 µM) was extracted using standard extraction parameters and changing one extraction parameter at a time. Standard extraction parameters were defined as 2 fill strokes,

Page 6 of 11

150 µL sample volume in the vial, 100 µL pull up volume and 0 s pull up delay. Extraction efficiency was investigated by comparing obtained peak areas and by calculating recoveries relative to direct injections of 10 µM methadone solution to the LC-MS system via the sample loop. Recovery (R) was calculated using the equation: % =

∗ ∗ 100 ∗

Where VSL is the volume of the sample loop (20 µL) and VSS the volume of the sample solution in the vial. For direct injections, the 20 µL sample loop coupled to the 10 port switching valve was loaded with 10 µM methadone in 10 mM HCOOH solution and sample was injected to the LC-MS system. For analysis of directly injected samples the same LC-MS method as used for analysis of compounds extracted with the EME autosampler was applied. To determine the amount of methadone trapped in the SLM a mass balance was calculated. Therefore sample solutions were analyzed by LC-MS prior and after EME (5 µL injection volume, same LC-MS method was used as for EME-LC-MS) and UA (%) (percentage of unextracted analyte) was determined using the equation: % =

∗ 100 !

The percentage of analyte trapped in the SLM was thereafter calculated using the equation: % " #$% % & '" ()* = 100 − % − UA (%) The equation is valid assuming that all extracted analytes were collected in the sample loop of the 10 port valve. The used sample loop had a very large volume compared to the extraction volume in the lumen of the extraction probe (20 µL sample loop volume compared to < 1 µL volume of extraction probe lumen). Therefore it can be assumed that under optimized sample loop loading time conditions all extracted analytes were collected in the sample loop. Minimization of sample carry-over – membrane cleaning To minimize carry-over between samples, optimal wash buffer composition and optimal number of wash cycles (number of repetitions the EME-syringe is flushed with wash buffer) were investigated. Optimal wash buffer composition was determined by extracting methadone (10 µM) applying standard extraction parameter followed by washing the EMEsyringe with dH2O, 0.05 % HCl, 0.5 % HCl, 0.05 % HCOOH or 0.5 % HCOOH respectively. For the washing procedure, one wash cycle was performed by the autosampler wash station and an electric potential of – 200 V was applied. After each sample analysis, carry-over was measured by analysis of a blank extraction (100 mM potassium phosphate buffer/5 mM MgCl2 (pH 7.4)). Relative carry-over was determined by calculating the ratio of the carry-over peak area (from blank extraction) and the sample peak area when cleaning the membrane with different wash buffers. After optimization of the wash buffer composition, the effect of the number of wash cycles on carry-over was determined. Extraction of methadone (10 µM) from 100 mM potassium phosphate buffer/5 mM MgCl2 (pH 7.4) followed by varying numbers of wash cycles of the EME-syringe with the optimized wash buffer composition was done. After each


Page 7 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


sample analysis, carry-over was measured by analysis of a blank extraction (100 mM potassium phosphate buffer/5 mM MgCl2 (pH 7.4)). Relative carry-over was determined by calculating the ratio of the carry-over peak area (from the blank extraction) and the sample peak area when cleaning the membrane with varying numbers of wash cycles. During the wash cycles, an electric potential of - 200 V was applied. Performance characterization under optimized extraction parameter settings To ensure proper performance of the EME-autosampler for measuring the conversion of methadone to EDDP by rat liver microsomes, the system was characterized with respect to repeatability, linearity, limit of detection (LOD) and limit of quantitation (LOQ). For determination of repeatability and linearity, serial two-fold dilutions of methadone were prepared in rat liver microsomes (3.2 µM, 6.3 µM, 12.5 µM, 25.0 µM) and analyzed in triplicates using optimized extraction parameters. Repeatability was determined by calculating the relative standard deviations (RSD) of the triplicates. Linearity was determined by doing a regression analysis for the tested concentration range. For determination of the LOD and LOQ methadone samples with a concentration close to the expected limit of detection (1.6 µM) were measured in triplicates and the following formulas were used for calculation of the LOD and LOQ: 3.3 ∗ 0 10 ∗ 0 ),- = ),1 = ( ( Where 0 is the standard deviation of the measured sample triplicates and S the slope of the standard curve.18 Monitoring metabolic conversion of methadone to EDDP For studying the in-vitro conversion of methadone to EDDP by rat liver microsomes the sample stack of the EMEautosampler and all reagents were preheated to 37 ○C. 700 µL of 100 mM potassium phosphate buffer/5 mM MgCl2 (pH 7.4), 100 µL of a 100 µM methadone solution (final concentration: 10 µM), 100 µL of rat liver microsomes (20 mg/mL) and 100 µL of a 10 mM NADPH solution were mixed in a sample vial and conversion of methadone into its main metabolite EDDP was monitored by repeated analysis of the sample with the EME-autosampler. NADPH is required as co-factor for the metabolic reaction. Likewise, MgCl2 is commonly added to the metabolic reaction mixture to increase the reaction rate of certain CYP enzymes. Analysis was automatically performed every 5.3 minutes by extraction from the sample vial. In order to minimize the effect of sample depletion when extracting methadone from the metabolic reaction solution, the extraction was performed under soft conditions. Therefore the following extraction parameters were applied: pull up cycles: 2; pull up volume: 100 µL, pull up delay: 0 s. The experiment was repeated 3 times.

omeprazole, fluoxetine or ketoconazole respectively (final concentration: 100 µM). For the positive control, 50 µL of 100 mM potassium phosphate buffer/5 mM MgCl2 (pH 7.4) instead of chemical inhibitor was added to the sample solution. Solutions were incubated for 45 minutes in a water bath at 37 ○ C during constant shaking. After 45 minutes 3 aliquots of 150 µL of each sample solution were transferred to sample vials and put into the EME-autosampler sample stack pre-cooled to 1 ○C. Samples were left in the cooled sample stack for 15 minutes before analysis with the EME-autosampler was started. Decreasing the temperature of the samples was done to ensure that the metabolism was terminated before analysis. The degree of inhibition (inh. %) of formation of EDDP was determined using the equation:

'"ℎ. % = 1 −

3--4 56!7 578 ∗ 100 3--4 !59 7

RESULTS AND DISCUSSIONS Optimal sample loop loading time Determination of the optimal sample loop loading time was required as a too long loading time would drain parts of the extract to waste. Likewise, a too short loading time would leave parts of the extract before the sample loop. Highest peak areas where obtained when loading the sample loop for 45 s (data not shown). Hence, for all consecutive experiments a sample loop loading time of 45 s was used. Investigation of effect of various extraction parameters The effect of general extraction parameters affecting EME were discussed in a number of previous publications7,19-22 where conventional offline EME setups were used. These parameters include extraction voltage, chemical composition of the liquid membrane, surface volume of the membrane and pH in sample solution and acceptor phase. The EME-autosampler introduces a set of additional parameters that can be used to optimize extraction efficiency including: • • • •

fill strokes sample pull up volume pull up delay sample volume in vial

Extraction parameters were optimized with respect to high sample throughput (short extraction times) and soft sample extraction. Soft sample extraction allows, where necessary, to re-analyze sample solutions as only small amounts of analyte get removed from the sample. Optimization was done by defining standard extraction parameters (2 fill strokes, 150 µL sample volume in the vial, 100 µL pull up volume and 0 s pull up delay) and measuring the effect of the parameters on extraction efficiency by changing one parameter at a time.

Measuring the effect of chemical inhibitors on the metabolism of methadone The EME-autosampler was used to measure the effect of the CYP3A4 inhibitors omeprazole, fluoxetine and ketoconazole on the metabolic conversion of methadone to EDDP. 300 µL of 100 mM potassium phosphate buffer/5 mM MgCl2 (pH 7.4) were mixed with 50 µL of 100 µM methadone (final concentration: 10 µM), 50 µL of a 10 mM NADPH solution, 50 µL rat liver microsomes (20 mg/mL) and 50 µL of 1000 µM of



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3: Effect of the sample pull up volume, fill strokes (number of repetitions sample is drawn and pushed back into the sample vial), pull up delay (time, sample rested in the EME-syringe after each pull up) and sample volume in the test vial on the amount of extracted analyte and analyte recovery; measurements were done using standard extraction parameters: 100 µL pull up volume, 2 fill strokes, 0 s pull up delay, 150 µL sample volume in vial and changing one parameter at a time

In a first experiment the effect of a pull up delay (time, sample rested in the EME-syringe after each pull up) on extraction efficiency was determined. As observable in Figure 3 C pull up delays of 5 and 10 seconds showed only moderate effects on the amount of extracted analyte compared to no pull up delay. Peak areas increased from 3.2E7 to 4.1E7 (5 seconds pull up delay, relative increase 18 %) and 4.5E7 (10 seconds pull up delay, relative increase 25 %) respectively. When optimizing the setup for high sample throughput the pull up delay is therefore a factor that should be kept low as it strongly increases the extraction time while increasing only slightly the amount of extracted sample. The result suggests that analyte in the sample solution that is rested in the EME-syringe for 5 or 10 seconds is not in close enough spatial proximity with the electro membrane in the luer lock adapter to be extracted efficiently. The moderate increase in extracted sample can most probably rather be attributed to a small amount of sample solution in the dead volume of the luer lock adapter around the electro membrane from which analyte could be extracted efficiently during the sample resting time of 5 or 10 seconds. In a next set of experiments, the effect of sample volume in the test vial and the effect of the sample pull up volume were investigated. As illustrated in Figure 3 D, the sample volume in the vial had no observable effect on the amount of extracted sample. However, as the ratio between sample loop volume and sample volume in the test vial increases, recovery decreases proportionally. When optimizing sample extraction with respect to soft extraction, sample volume can therefore be considered to be an important parameter. As shown in Figure 3 A, the amount of extracted sample increased when the sample pull up volume was raised from 50 µL (peak area 2.2E7) to 100 µL (peak area 3.3E7, relative increase: 32 %). When increasing the pull up volume, more sample passes the electro membrane leading to a higher

Page 8 of 11

amount of extracted analyte. Both, the sample volume in the vial and the sample pull up volume have no or only minor influence on the extraction time and high sample throughput can therefore be achieved irrespective the sample volume in the vial and the sample pull up volume. As main factor to influence extraction efficiency, the number of fill strokes (number of repetitions sample is drawn and pushed back into the sample vial by the EME-syringe) was identified. As observable in Figure 3 B, extraction efficiency increased approximately 3 fold when increasing the number of fill strokes from 2 to 16. When applying 16 fill strokes obtained peak areas were in a similar range as peak areas obtained from direct injections. Direct injections were done by injecting 10 µM methadone in 10 mM HCOOH through the 20 µL sample loop coupled to the 10 port switching valve. Similar to signal intensity, recovery increased from 4 % when applying 2 fill strokes to 13 % when 16 fill strokes were applied (Figure 3B and Figure 4). Similar to an increase in sample pull up volume, more sample passed the electro membrane when the number of fill strokes was increased, leading to a higher amount of extracted analyte and thus higher recoveries. Further, calculation of a mass balance demonstrated that only a very small amount of sample was trapped in the SLM (– 4 % to 5 %, negative values are due to experimental inaccuracies). The result of the mass balance calculation is shown in Figure 4. The results demonstrate that the number of fill strokes is an important optimization parameter for applications where high sensitivity and therefore higher recovery values are desirable. As for the current work the setup was optimized with respect to high sample throughput and soft sample extraction, the optimized sample extraction parameters were defined as: 2 fill strokes, 100 µL pull up volume, 0 seconds pull up delay and 150 µL sample volume. The parameters constitute a compromise between high sample throughput and soft extraction on one hand side and sufficiently high sensitivity for the chosen applications on the other hand.

Figure 4: Distribution ratio of methadone in sample phase, acceptor phase and membrane after EME applying varying numbers of fill strokes; extractions parameters: 150 µL sample volume in vial, 0 s pull up delay, 100 µL sample pull up volume

Membrane cleaning to minimize carry-over effect To minimize carry-over between samples, the autosampler was programmed to clean the EME-syringe after each sample extraction using the EME-autosampler wash station. The wash step was performed to back-extract analytes trapped in the SLM by reversing the applied electric potential to – 200 V and simultaneous flushing of the EME-syringe with wash buffer. Carry-over was measured when washing with dH2O, 0.05 % HCl, 0.5 % HCl, 0.05 % HCOOH or 0.5 % HCOOH. All four


Page 9 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


acidic wash buffers showed strongly reduced carry-over compared to dH2O. Signal intensities of the carry-over peaks after washing with acidic wash buffers compared to carry-over peaks after washing with dH2O were reduced by 73 % ± 3 % (0.05 % HCl), 67 % ± 5 % (0.5 % HCl), 76 % ± 7 % (0.05 % HCOOH) and 81 % ± 4 % (0.5 % HCOOH). As lowest carryover was determined using 0.5 % HCOOH all consecutive experiments were performed using 0.5 % HCOOH as wash buffer.Further, the effect of the number of wash cycles (number of filling and emptying the EME-syringe with wash buffer) was investigated.

Figure 5: Carry-over of methadone applying varying numbers of wash cycles with 0.5 % HCOOH wash buffer As shown in Figure 5, relative carry-over of 27 % was measured when no membrane wash was carried out. Carryover decreased with increasing number of wash cycles and was < 1 % (0.9 %) when 8 wash cycles were applied. In order to keep carry-over at a minimum, 8 wash cycles were applied in all consecutive experiments. Characterization of the EME-autosampler at optimized extraction parameters To ensure proper performance of the developed system for measuring the conversion of methadone to EDDP by rat liver microsomes, the EME-autosampler was characterized with respect to LOD, LOQ, repeatability and linearity in the relevant concentration range. For all experiments optimized extraction parameters were used (2 fill strokes, 100 µL pull up volume, 0 seconds pull up delay, 150 µL sample volume). The same parameters were also used for subsequent measurement of the conversion of methadone into EDDP by rat liver microsomes and for measuring the potential of various drugs to inhibit formation of EDDP. Based on the results from the investigation of the extraction parameters the settings were chosen as a compromise between high sample throughput/soft extraction and a sufficiently high sensitivity of the system. Using these extraction parameters a LOD of 1.27 µM and a LOQ of 3.86 µM was determined. It should however be noted that much lower LOD values can be achieved when using extraction parameters more in favor of high extraction recoveries. Within the tested concentration range from 3.1 µM to 25 µM good repeatability (4 % - 10 % RSD) and linearity (R2: 0.994, y = 1.3E6x + 3.5E6) was obtained. It should be noted that all samples for characterization of the EME-autosampler were extracted and analyzed in one sequence. A sequence of 16 samples (sample triplicates at 4 analyte concentrations plus blank samples) was thereby analyzed without any manual sample handling in less than 90 minutes.

Automated monitoring of in-vitro conversion of methadone to EDDP by rat liver microsomes The EME-autosampler was used for fully automated monitoring of the conversion of methadone into its main metabolite EDDP. Methadone was mixed with rat liver microsomes and NADPH and incubated in the EME-autosampler sample stack heated to 37 ○C. Every 5.3 minutes analytes (methadone and EDDP) were extracted from the reaction vial and analyzed by LC-MS.

Figure 6: Automated monitoring of conversion of methadone into its main metabolite EDDP by rat liver microsomes; methadone and EDDP signal intensity: left y-axis; EDDP/methadone ratio: right y-axis

After the first extraction (approximately 40 seconds after initiation of the metabolic reaction) only methadone was present in the extract. As the metabolic reaction proceeded, methadone was converted into EDDP resulting in continuously decreasing signals for methadone and increasing signals of EDDP and consequently increasing EDDP/methadone ratios (Figure 6). Automated analysis of the effect of chemical inhibitors on the metabolism of methadone by rat liver microsomes Conversion of methadone into EDDP is mainly triggered by the enzyme CYP3A423,24. Inhibitors of CYP3A4 can affect the metabolism of methadone by increasing its bioavailability in the body. Knowledge about the inhibitory potential of coadministered drugs is therefore of crucial importance in drug development.

Figure 7: Analysis of the effect of chemical inhibitors on the conversion of methadone into EDDP by rat liver microsomes

The potential of the known CYP3A4 inhibitors25,26, ketoconazole omeprazole and fluoxetine, to inhibit the metabolism of methadone was investigated using the EME-



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

autosampler. 10 µM methadone were incubated for 45 minutes with rat liver microsomes, NADPH and an excess amount of inhibitor (100 µM). Formation of EDDP in the presence of ketoconazole, omeprazole and fluoxetine, compared to a positive control not containing CYP3A4 inhibitors was inhibited by 68 %, 38 % and 60 %respectively (Figure 7). The results were in good accordance with a prior study of the inhibitory potential of various drugs on the metabolism of methadone by human liver microsomes24. In this reference work EDDP formation was inhibited by 75 % (ketoconazole), 29 % (omeprazole) and 56 % (fluoxetine) . The obtained results therefore suggest that the tested drugs inhibit the metabolism of methadone by rat liver microsomes to a similar extend to human liver microsomes. CONCLUSION The current work describes the automation of a complete analytical workflow of sample preparation, separation and analysis by implementing an EME probe into a commercial autosampler and coupling it to LC-MS. High extraction speed as an inherent attribute of EME and the absence of any manual sample handling before analysis make the introduced EMEautosampler a promising approach for a wide range of high sample throughput applications. It was demonstrated that carry-over between samples caused by analytes trapped in the SLM can be minimized by flushing the EME-syringe with acidic wash buffer while reversing the applied electric potential. Further, it was shown that by adjusting sample extraction parameters that are unique to the EMEautosampler, such as the number of the EME-autosampler fill strokes, extraction can be optimized with respect to analyte recovery and sample throughput. In the current work the EME-autosampler was used for automated monitoring of the conversion of methadone into its main metabolite EDDP by rat liver microsomes. In order not to affect the metabolic reaction by repeated extractions from the same sample, soft extraction parameters were used. For other applications where higher sensitivity and recovery is desirable, extraction parameters can be adjusted accordingly. Future work will further broaden the analytical potential of the EME-autosampler by dealing with further optimization of the extraction parameters towards exhaustive extractions required for the analysis of trace amounts of analytes.

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Video file S-1: Video illustrating the various analysis steps during EME-LC-MS (MPG) Table S-1: Description of action steps carried out by the various setup components during analysis of a sample (PDF)

AUTHOR INFORMATION Corresponding Author * E-mail: [email protected]

ACKNOWLEDGMENT

Page 10 of 11

We acknowledge ARIADME, a European FP7 ITN Community’s Seventh Framework Program under Grant Agreement No. 607517 and the Core Facility for Integrated Microscopy, Faculty of Health and Medical Sciences, University of Copenhagen

REFERENCES (1) Pedersen-Bjergaard, S.; Rasmussen, K. E. Journal of chromatography. A 2006, 1109, 183-190. (2) Kuban, P.; Bocek, P. Journal of chromatography. A 2014, 1346, 25-33. (3) Kuban, P.; Bocek, P. Analytica chimica acta 2014, 848, 43-50. (4) Payan, M. D.; Li, B.; Petersen, N. J.; Jensen, H.; Hansen, S. H.; Pedersen-Bjergaard, S. Analytica chimica acta 2013, 785, 60-66. (5) Gjelstad, A.; Andersen, T. M.; Rasmussen, K. E.; Pedersen-Bjergaard, S. Journal of chromatography. A 2007, 1157, 38-45. (6) Eibak, L. E.; Gjelstad, A.; Rasmussen, K. E.; Pedersen-Bjergaard, S. Journal of chromatography. A 2010, 1217, 5050-5056. (7) Seip, K. F.; Jensen, H.; Sonsteby, M. H.; Gjelstad, A.; Pedersen-Bjergaard, S. Electrophoresis 2013, 34, 792799. (8) Pantuckova, P.; Kuban, P.; Bocek, P. Analytica chimica acta 2015, 887, 111-117. (9) Pantuckova, P.; Kuban, P.; Bocek, P. Journal of chromatography. A 2013, 1299, 33-39. (10) Lindenburg, P. W.; Tempels, F. W.; Tjaden, U. R.; van der Greef, J.; Hankemeier, T. Journal of chromatography. A 2012, 1249, 17-24. (11) Li, B.; Petersen, N. J.; Payan, M. D.; Hansen, S. H.; Pedersen-Bjergaard, S. Talanta 2014, 120, 224-229. (12) Dugstad, H. B.; Petersen, N. J.; Jensen, H.; Gabel-Jensen, C.; Hansen, S. H.; Pedersen-Bjergaard, S. Analytical and bioanalytical chemistry 2014, 406, 421-429. (13) Petersen, N. J.; Foss, S. T.; Jensen, H.; Hansen, S. H.; Skonberg, C.; Snakenborg, D.; Kutter, J. P.; PedersenBjergaard, S. Analytical chemistry 2011, 83, 44-51. (14) See, H. H.; Hauser, P. C. Analytical chemistry 2014, 86, 8665-8670. (15) Fuchs, D.; Jensen, H.; Pedersen-Bjergaard, S.; Gabel-Jensen, C.; Hansen, S. H.; Petersen, N. J. Analytical chemistry 2015, 87, 5774-5781. (16) Fuchs, D.; Gabel-Jensen, C.; Jensen, H.; Rand, K. D.; Pedersen-Bjergaard, S.; Hansen, S. H.; Petersen, N. J. Analytica chimica acta 2016, 905, 93-99. (17) Drouin, N.; Rudaz, S.; Schappler, J. Analytical chemistry 2016. (18) Validation of an Analytical Method - ICH Guideline Q2(R1), Validation of Analytical Procedures: Text and Methodology, 2005 (19) Eskandari, M.; Yamini, Y.; Fotouhi, L.; Seidi, S. Journal of Pharmaceutical and Biomedical Analysis 2011, 54, 11731179. (20) Rezazadeh, M.; Yamini, Y.; Seidi, S. Journal of Chromatography B-Analytical Technologies in the Biomedical and Life Sciences 2011, 879, 1143-1148. (21) Middelthon-Bruer, T. M.; Gjelstad, A.; Rasmussen, K. E.; Pedersen-Bjergaard, S. Journal of separation science 2008, 31, 753-759. (22) Gjelstad, A.; Rasmussen, K. E.; PedersenBjergaard, S. Journal of chromatography. A 2007, 1174, 104-111. (23) Moody, D. E.; Alburges, M. E.; Parker, R. J.; Collins, J. M.; Strong, J. M. Drug Metabolism and Disposition 1997, 25, 1347-1353. (24) Foster, D. J. R.; Somogyi, A. A.; Bochner, F. British Journal of Clinical Pharmacology 1999, 47, 403-412.


Page 11 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(25) Ferrari, A.; Coccia, C. P.; Bertolini, A.; Sternieri, E. Pharmacol Res 2004, 50, 551-559. (26) Katz, H. I. British Journal of Dermatology 1999, 141, 26-32.


Fully Automated Electro Membrane Extraction Autosampler for LCâMS

Recommend Documents

Fully Automated Electro Membrane Extraction Autosampler for LCâMS