Analysis of Biological Fluids by Direct Combination of Solid Phase

Jan 22, 2009 - Fax: +49 641 99 34809., †. Semmelweis University. , ‡. Justus-Liebig-Universität. Cite this:Anal. Chem. 2009, 81, 4, 1669-1675 ...
0 downloads 0 Views 630KB Size
Download PDF
Anal. Chem. 2009, 81, 1669–1675

Analysis of Biological Fluids by Direct Combination of Solid Phase Extraction and Desorption Electrospray Ionization Mass Spectrometry ´ da´m Hosszu´,† Noe´mi Czuczy,† and Zolta´n Taka´ts*,†,‡ Ju´lia De´nes,† Ma´ria Katona,† A Cell Screen Applied Research Center, Semmelweis University, Budapest, Hungary, and Institut fu¨r Analytische Chemie, Justus-Liebig-Universita¨t, Giessen, Germany A novel, solid phase extraction (SPE)-based sample preparation method was developed for desorption electrospray ionization (DESI) mass spectrometry. Conventional SPE sample preparation was followed by a custom elution procedure. The eluate was evaporated from the closing frit of the cartridge using a gas jet. Thus the analyte was concentrated on the surface of the frit, which is ideal for DESI analysis. Application of the above SPE protocol allowed the concentration of the analyte content of up to 1 L liquid sample into a 1 mm diameter circular spot. The sample preparation procedure can improve the overall sensitivity of the method by up to 6 orders of magnitude if the sample volume is sufficient. The device has been tested using aqueous solutions of Rhodamine 116; the limit of detection was comparable to the LOD of electrospray analysis. Methodology was tested for drug monitoring applications in human serum. Levels of Cyclosporine A were determined using a 0.1 mL serum sample. Dynamic range of the method exceeded 3 orders of magnitude; the detection limit was below the therapeutic serum concentration of the drug. Recent development of the group of ionization methods termed “direct ionization” or “ambient ionization” has opened a number of novel areas of application for mass spectrometric analysis. Ambient techniques such as thermal desorption/APCI were first described over 20 years ago, but most of them were never in widespread use.1,2 The recent renaissance of the ambient techniques started with the development of desorption electrospray ionization (DESI),3,4 which was quickly followed by “direct analysis in real time” (DART),5,6 desorption atmospheric pressure chemical * To whom correspondence should be addressed. E-mail: zoltan.takats@ anorg.chemie.unigiessen.de. Fax: +49 641 99 34809. † Semmelweis University. ‡ Justus-Liebig-Universita¨t. (1) Van Berkel, G. J.; Pasilis, S. P.; Ovchinnikova, O. J. Mass Spectrom. 2008, 43, 1161–1180. (2) Rommers, P.; Boumans, P. Fresenius’ J. Anal. Chem. 1996, 355. (3) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471–473. (4) Takats, Z.; Wiseman, J. M.; Cooks, R. G. J. Mass Spectrom. 2005, 40, 1261– 1275. (5) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297– 2302. (6) Petucci, C.; Diffendal, J.; Kaufman, D.; Mekonnen, B.; Terefenko, G.; Musselman, B. Anal. Chem. 2007, 79, 5064–5070. 10.1021/ac8024812 CCC: $40.75  2009 American Chemical Society Published on Web 01/22/2009

ionization (DAPCI),7 and a number of other methods. The common feature of these methods is the lack of the imperative need for sample preparation prior to analysis, which dramatically simplifies and thus accelerates the analytical scheme.8,9 The importance of high-throughput (HT) analytical methods has emerged with the advent of the combinatorial chemistry era in the 80s pharmaceutical industry. Semiquantitative methods were developed using mainly mass spectrometric methods with no or minimal chromatographic separation.10 The area has gotten further momentum from the development of modern HPLC-MS methods utilizing ESI and APCI ionization techniques.11-19 The development of HT bioanalytical methods based on electrophoresis,20 chromatography, and mass spectrometry was the basis of modern genomics, proteomics, metabolomics, and lipidomics. HT analytical techniques have gained high importance in medical diagnostics and clinical chemistry. The rapidly emerging field of personalized medicine also requires HT techniques for therapeutic drug monitoring, genetic and enzyme assays.21,22 HT DESI and DART were first demonstrated for the analysis of pharmaceutical formulations.23,24 DESI and DART analysis of tablets for process monitoring and the identification of counterfeit (7) Williams, J. P.; Patel, V. J.; Holland, R.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 2006, 20, 1447–1456. (8) Cooks, R. G.; Ouyang, Z.; Takats, Z.; Wiseman, J. M. Science 2006, 311, 1566–1570. (9) Harris, G. A.; Nyadong, L.; Fernandez, F. M. Analyst 2008, 133, 1297– 1301. (10) Kassel, D. B. Chem. Rev. 2001, 101, 255–268. (11) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64–71. (12) Horning, E. C.; Horning, M. G.; Carroll, D. I.; Dzidic, I.; Stillwell, R. N. Anal. Chem. 1973, 45, 936–943. (13) Kyranos, J. N.; Cai, H.; Wei, D.; Goetzinger, W. K. Curr. Opin. Biotechnol. 2001, 12, 105–111. (14) Alford, A. Biomed. Mass Spectrom. 1978, 5, 259–286. (15) Koh, H.-L.; Yau, W.-P.; Ong, P.-S.; Hegde, A. Drug Discov. Today 2003, 8, 889–897. (16) Lim, C.-K.; Lord, G. Biol. Pharm. Bull. 2002, 25, 547–557. (17) Richardson, S. D. Anal. Chem. 2008, 80, 4373–4402. (18) Vukelic, Z.; Zarei, M.; Peter-Katalinic, J.; Zamfir, A. D. J. Chromatogr. A 2006, 1130, 238–245. (19) Zo ¨llner, P.; Mayer-Helm, B. J. Chromatogr. A 2006, 1136, 123–169. (20) Garza, S.; Chang, S.; Moini, M. J. Chromatogr. A 2007, 1159, 14–21. (21) Jain, K. K. Curr. Opin. Mol. Ther. 2003, 5, 548–558. (22) Nunn, A. D. Cancer Biother. Radiopharm. 2007, 22, 722–739. (23) Chen, H.; Talaty, N. N.; Takats, Z.; Cooks, R. G. Anal. Chem. 2005, 77, 6915–6927. (24) Kauppila, T. J.; Wiseman, J. M.; Ketola, R. A.; Kotiaho, T.; Cooks, R. G.; Kostiainen, R. Rapid Commun. Mass Spectrom. 2006, 20, 387–392.

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

1669

medical formula have gained significance recently.25 The analysis of blood spots for screening for metabolic diseases by DESI-MS was demonstrated, as well as the HT analysis of biological fluids for the determination of drug levels.26 Although the applicability of ambient ionization methods for HT analysis has been successfully demonstrated on various fields, routine application of these techniques has not been reported so far. Ambient ionization methods initially appeared to bear the potential to revolutionize everyday application of mass spectrometry; however, all of them show a number of disadvantages compared to their conventional counterpart (i.e., DESI-ESI, etc.). Although comparison of spray ionization methods to desorption ionization methods is not straightforward, linearity, sensitivity, and other analytical parameters can directly be compared for liquid samples. In the case of sensitivity and LLOD, 2-3 orders of magnitude difference has been reported in favor of spray methods. While in the case of DESI and DART the absolute lower limit of quantitation (LLOQ) is in the mid pg range,27 fg level determination have been reported using nanospray ionization.28 The lack of sensitivity has been attributed to intrinsic factors, arising from the mechanism of ion formation.29-31 Although ambient mass spectrometric techniques are claimed to provide a “sample preparation-free” analytical scheme, this statement does not necessarily stand for regular analytical applications. On one hand, ambient ionization techniques do give mass spectrometric information on untreated, native samples, on the other hand, it is seldom the analytical task to obtain unspecified mass spectrometric data characteristic of the sample. Real life analytical problems generally tackle the qualitative and/ or quantitative determination of certain, pre-determined species of interest, and ambient ionization techniques often fail to accomplish this task. An ideal “sample-preparation free” technique should be able to detect and quantify species of interest in native samples in one step, at relevant concentrations. So far this has been accomplished only for a few combinations of analytes and matrixes, and no generally applicable MS-based technique has been described which fulfills these requirements. Furthermore, practically any treatment of sample prior to analysis may qualify as sample preparation (e.g., drawing of blood sample, preparation of plasma and deposition of plasma onto glass slide prior to DESI analysis) so complete elimination of sample preparation is often impossible, and even “sample preparation free” methods employ multiple steps of sample treatment prior to analysis. Complete elimination of sample preparation is necessary only, when in situ, real-time analysis is required, for example, in the case of airport security applications. The lack of proper sample preparation is usually reflected in poor sensitivity and interference from isobaric matrix peaks. The qualitative analysis of biological fluids by ambient mass spectrom(25) Fernandez, F. M.; Green, M. D.; Newton, P. N. Ind. Eng. Chem. Res. 2008, 47, 585–590. (26) Kauppila, T. J.; Talaty, N.; Kuuranne, T.; Kotiaho, T.; Kostiainen, R.; Cooks, R. G. Analyst 2007, 132, 868–875. (27) Cotte-Rodriguez, I.; Takats, Z.; Talaty, N.; Chen, H.; Cooks, R. G. Anal. Chem. 2005, 77, 6755–6764. (28) Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 1–8. (29) Takats, Z.; Cotte-Rodriguez, I.; Talaty, N.; Chen, H.; Cooks, R. G. Chem. Commun. 2005, 1950–1952. (30) Venter, A.; Sojka, P. E.; Cooks, R. G. Anal. Chem. 2006, 78, 8549–8555. (31) Pasilis, S. P.; Kertesz, V.; Van Berkel, G. J. Anal. Chem. 2007, 79, 5956– 5962.

1670

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

Figure 1. Schematics of (a) SPE cartridge and (b) the sampling card. (1) Polymer frits, (2) SPE packing, (3) polypropylene tubing, (4) polypropylene sheet, (5) polymeric membranes, (6) SPE disks.

etry, similarly to traditional methods, requires sample cleanup to eliminate background interference. Purification is actually more critical for ambient techniques than for LC-MS methods, since the former intrinsically do not involve chromatographic separation of the analytes prior to mass spectrometric detection. In the present study we report the development, characterization, and practical application of a universal sample preparation method for desorption ionization mass spectrometry, especially for desorption electrospray ionization-MS. The sample preparation method is based on the adsorption of the analyte molecules on solid phase extraction (SPE) packing and the elution of the analyte onto the membrane which seals the cartridge. Thus, matrix interference is eliminated, and the analyte molecules are concentrated on the confined surface of the closing frit of the SPE cartridge. Furthermore, instead of accepting the intrinsic surface characteristics of dried biological fluids (e.g., blood clot), the method offers a possibility to choose the optimal surface type for the desorption ionization method used. The sample preparation method was implemented for HT analysis, and it was demonstrated that the proposed sample preparation does not extend the overall time demand of the analytical procedure. EXPERIMENTAL SECTION Materials. Rhodamine 116 and Rhodamine 123, Sudan Red 6 B, Cyclosporine A (from Tolypocladium inflatum, g95%), Atrazine (PESTANAL, analytical standard), and Simazine (PESTANAL, analytical standard) were obtained from Sigma-Aldrich (St. Louis, MO). Cyclosporine D was obtained from BioMarker Ltd. (Godollo, Hungary). All solvents (HPLC-grade) were purchased from Merck (Nottingham, U.K.). SPE Cartridge. Custom built SPE cartridges consisting of polypropylene tubes, polymer frits, and SPE packing were used. The scheme of the cartridge is shown in Figure 1a. Polypropylene tubes were obtained from B. Braun (Melsungen, Germany), PTFE frits were purchased from Supelco (Bellefonte, PA), and PE frits were purchased from Biotage (Uppsala, Sweden). Various SPE packings were obtained from Biotage, Supelco, and Varian (Palo Alto, CA). The scheme of the sampling card is shown in Figure 1b. The cards consist of polypropylene sheets, with 5 mm diameter holes drilled into them. Five millimeter diameter disks of SPE material were embedded into polypropylene with one or two polymeric membranes covering the surfaces. SPE disks were obtained from Biotage (Uppsala, Sweden) and 3 M (St. Paul, MN); porous

Figure 2. 96-channel DESI ion source on the 2D moving stage. (1) Electro-sonic spray emitter, (2) SPE plate containing 96 individual SPE cartridges (depicted in Figure 1a.) in a 8 × 12 raster, (3) heated capillary, (4) home-built atmospheric interface housing, (5) temperature sensor, (6) 2D moving stage, (7) API 4000 QTRAP hybrid triple quadrupole/ion trap mass spectrometer. Samples are applied onto a 96 channel SPE plate and analyte is eluted onto top closing frits of individual cartridges. Prepared plate is secured onto aluminum frame (8), and individual cartridges are analyzed sequentially by moving them under sprayer (1). Ion source communicates with data acquisition software of mass spectrometer as an autosampler device.

polymeric membranes were purchased from Millipore (Billerica, MA). Filter paper (Whatman 903) was also used in the sampling cards. Instrumentation. SPE cartridges were eluted using a purposemade elutor device, consisting of a stainless steel cartridge holder, a gas heater unit, and a fluid delivery unit (Supporting Information, Figure S1a). The 96-channel elutor device (Supporting Information, Figure S1b) employs an 8 × 12 array of individual SPE cartridges inserted into a PPS plate. Application of samples and elution is carried out in parallel fashion on all cartridges, using custom-made vacuum manifold and elutor devices. (For details, see Supporting Information.) The 96-channel DESI ion source (Figure 2) consists of an electrosonic spray, the primary electrospray emitter mounted on a PTFE holder, which is secured onto a rotating stage (Parker, Cleveland, OH). The rotating stage is mounted on a threedimensional manual moving stage (Parker). The aluminum frame for the SPE plate is mounted on a computer controlled 2D moving stage (Newmark System, Mission Viejo, CA). Z dimensional (up-down) relative movement is implemented by a moving sprayer and a heated capillary inlet of the mass spectrometer in the z direction. Both moving stage systems are built onto a common aluminum platform, which is secured onto the mass spectrometer through a welded, square hollow section steel frame. Mass spectrometric experiments were performed on a TSQ Quantum Discovery (ThermoFinnigan, San Jose, CA) triple quadrupole mass spectrometer equipped with a slightly modified OmniSpray DESI ion source (Prosolia Inc., Indianapolis, IN) and on a modified API 4000 Q-TRAP hybrid triple quadrupole/ion trap mass spectrometer (Applied Biosystems/MDS Sciex, Concorde, Ontario, Canada). The atmospheric interface of the API 4000 instrument was modified; the original interface was replaced by a home-built heated capillary-type interface unit (Supporting Information, Figure S2).

Figure 3. Scheme of the SPE/DESI method. (1) analytes, (2) solvent, (3) upper porous membrane, (4) adsorbent, (5) lower porous membrane, (6) heated nozzle, (7) hot nitrogen, (8) eluent, (9) eluent vapor, (10) ion source, (11) spray, (12) desorbed ions.

Sample Preparation. C18 cartridges were wetted with methanol and conditioned with methanol/water 5:95. Sample was applied directly (Figure 3a); sample reservoirs connected to the cartridges were used when sample volume exceeded 0.5 mL. Following the application of the sample, the cartridges were washed with 0.5 mL methanol/water 5:95, and then were allowed to air-dry. The cartridges were transferred to the elutor device (see Supporting Information, Figure S1a) and were eluted with 150 µL of organic solvent at 40-200 µL/min flow rate (Figure 3b.). The organic solvent content of the eluate was evaporated from the closing frit of the cartridge using a 0.05-1 L/min hot (50-200 °C) nitrogen gas flow. The elution was concluded by replacing the organic solvent flow with 2 mL/min air flow by switching the 6-port valve to remove all residual solvent from cartridge. Skipping the latter step results in back-diffusion of the analyte into the wet SPE packing, which compromises the sensitivity and the reproducibility of the DESI method (Figure 3c). The 96-channel SPE plate was constructed by combining 96 individual cartridges in a 8 × 12 format, keeping the standard pattern and raster of 96-well microtiter plates. Samples were pipetted into the equilibrated cartridges using an 8-channel pipettor. Application of the samples was carried out using a homebuilt vacuum manifold. However, since the handling of 96 individual solvent lines is troublesome, an alternative approach was chosen for elution (Supporting Information, Figure S1b). In this case the cartridges are upside down compared to the single channel elutor. Appropriate volumes of eluent were pipetted into cartridges, and the eluting solvent was forced through the SPE packings by pressurized nitrogen (300-400 mbar). The eluate was evaporated from the closing frits by hot (100-150 °C) airflow provided by the elutor device. Analysis. Cartridges were placed into the DESI source and were analyzed using either DESI or DAPCI ionization. The ion source parameters and instrumental settings are summarized in Table 1. Geometrical parameters and flow rates of DESI were set to obtain a spray fingerprint size in the range of 0.1-1 mm2, to analyze cartridges in a single step, without moving during analysis. Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

1671

Table 1. Ion Source Parameters and Instrumental Settings parameters iIncident angle of the spray surface-spray distance surface-capillary inlet distance spray voltage nitrogen pressure spray solvent electrospray flow rate ion mode capillary voltagea capillary temperature tube lens offset a

TSQ Quantum

API 4000 QTRAP

60° 60° ∼2 mm ∼2 mm ∼2 mm ∼2 mm 4000 V 3000 V 7 bar 8 bar methanol-water 1:1 methanol-water 1:1 2 µL/min 10 µL/min positive positive 30 V 300 V 300 °C 180 °C 100 V

Declustering potential in the case of API 4000 QTRAP.

Figure 4. (a) Logarithmic calibration curves of Cyclosporine A from spiked human plasma samples with DESI and SPE/DESI methods. (b) LLOD of method depending on sample volume: the line indicates theoretical results with the DESI method; points indicate the results of real SPE/DESI measurements.

RESULTS AND DISCUSSION Comparison of DESI and SPE/DESI. The calibration curves for the determination of Cyclosporine A in human plasma by direct deposition DESI and SPE/DESI are shown in Figure 4a. In the case of direct deposition, 1 µL deproteinized plasma sample was dried onto a porous PTFE target, while in the case of SPE/DESI 100 µL sample was extracted. The SPE-based sample preparation results in the deposition of 2 orders of magnitude more analyte; thus, 2 orders of magnitude 1672

Table 2. Comparison of Analytical Parameters of DESI and SPE/DESI

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

absolute LLOD LLOQ time of the analysis linear range volume of the sample

DESI

SPE/DESI

10 pg-1 ng 10-1000 ng/mL 5s 1-100 µg/mL limited

10 pg-1 ng 100 fg/mL-100 pg/mL 10 s 10 ng-100 µg/mL unlimited

higher signal intensity is expected. In the range where both curves are linear (3-30 µg/mL), the ratio of signal intensities is close to the expected value (Figure 4a). However, the dynamic range of the direct deposition DESI is limited to this concentration range, while SPE/DESI is linear from 10 ng/mL to 30 µg/mL. At lower and higher concentrations, direct deposition DESI gives unreliable, practically qualitative information on the presence of Cyclosporine A in the samples. At low concentrations, the better linearity of SPE/DESI data can primarily be attributed to higher signal intensity and the partial elimination of the matrix-interference. The standard deviation of the SPE/DESI data is remarkably lower throughout the entire investigated concentration range. The observed better reproducibility was most likely due to the more even distribution of the analyte on the surface. Following direct deposition, the solvent evaporates from the droplet according to its intrinsic physico-chemical characteristics, resulting in heterogeneously distributed analyte clusters. During the elution step of SPE/DESI the eluate seeps through evenly the entire diameter of the closing membrane, and the solvent is evaporated immediately, resulting in homogeneous analyte distribution. In theory, initial sample size and hence concentration factor can be arbitrarily chosen for SPE/DESI methodology. In practice, though, saturation of the SPE packing and the usually limited sample volume (e.g., blood samples) set limits on the enhancement of sensitivity. Solution of atrazine in tap water was used to determine the LLOD of the method at different initial sample volumes. LLOD of the method improves as a function of initial sample volume in the case of tap water samples. At concentration factors higher than 10000, signal intensities tend to reach a maximum value (Figure 4b.). It has to be noted that in the case of spiked HPLC-grade water samples no saturation effects were observed in this range; the curve was linear throughout 6 orders of magnitude. Analytical parameters of DESI and SPE/DESI are summarized in Table 2. Systematic Characterization of SPE/DESI. The effect of various elution parameters was investigated by using dilute aqueous solutions of dyes (Rhodamine 116, Rhodamine 123, and Sudan Red 6 B) as samples and different eluting solvents including pentane, hexane, acetone, acetonitrile, dichloromethane, and methanol. Since elution volume for the dyes with water eluent is practically infinite on the octadecyl-silica packing used, the effects of sample volume and concentration were not studied. Investigated parameters included physico-chemical character of the eluent, elution flow rate, drying gas flow rate, drying gas temperature, and geometric parameters of the elutor device. Buffered aqueous solution (10 mL) of 100 ng dye was applied onto equilibrated SPE cartridges. The cartridges were dried and eluted with the appropriate solvent. Elution volumes depend on the polarity and H-bonding characteristics of solvents and analytes;

Figure 5. Comparison of elution process using solvent flow rate below and above maximum solvent evaporation rate. In the case of optimal elution (a) all solvent is evaporated, and analyte is crystallized evenly on frit surface. When evaporation is incomplete, (b) accumulating solvent redissolves analyte crystals, and eventually drying gas flow removes practically all analyte solution from the surface (c). Figure shows schematic representation of the processes, photo of resulting samples, and line scans obtained by DESI analysis.

minute deviations were associated with longitudinal diffusion effects. Since the elution volumes are determined by intrinsic characteristics of the solvent, the analyte, and the cartridge, optimal elution flow rate equals to the maximum evaporable solvent rate from the surface. This latter value was defined as the maximum flow rate, which does not result in the formation of a continuous liquid film on the closing frit of the cartridge (Figure 5a). If a continuous liquid film is formed on the frit surface, the eluted analyte is redissolved (Figure 5b) and crystallized in a circular pattern around the rim of the closing frit (Figure 5c). This phenomenon occurs at a well-defined solvent flow rate under given elutor geometry, drying gas flow rate, and gas temperature. The result of DESI line scan through the frit clearly indicates that the concentration of the analyte into a ring around the frit yields poor overall sensitivity. This effect was associated with the high local surface concentration of analyte, which exceeds the dynamic range of the DESI analysis. Since the general objective of systematic characterization was to find a global maximum for solvent evaporation rate, maximum evaporation rate (MER) values (defined above) were studied as a function of various parameters. Elutor geometry was found to have a well-defined effect on MER, since even analyte distribution was achieved only when drying gas emitter nozzle was in-line (in coaxial position) with the cartridge.

The dependence of MER on gas flow rate and temperature using methanol as solvent is depicted in Figure 6a. The figure shows that MER is a non-linear function of both gas flow rate and temperature. Under the given conditions MER shows a global maximum indicated on the figure, corresponding to ∼340 µL/ min. This in the case of Rhodamine 123 and methanol gives a minimum elution time of ∼15 s (Figure 6b). MER values give an inversely proportional function of evaporation enthalpy of the solvents used. In the case of more volatile solvents (isopentane, dichloromethane) MER as high as ∼800 µL/min is achievable, which results in 4-6 s elution times. Elution time under the above conditions is comparable to DESI analysis time, which means that sample preparation is not the rate-limiting step of the analytical procedure. Multiplexing. SPE/DESI was also designed in a 96-channel format for HT application. DESI-MS analysis of individual cartridges was carried out by moving the plate under the spray along the x axis, keeping the spray tip-surface distance constant.32 An analysis time of 2 s/sample was routinely achieved without compromising sensitivity. Reproducibility was found to highly depend on the sampling rate (Figure 7). Although sampling rates as high as 5 sample/s are theoretically achievable based on the linear motion parameters of the moving stage, reproducibility of results were found to be extremely poor at sampling rates higher than 0.5 sample/s. This feature was tentatively associated with (32) Kertesz, V.; Van Berkel, G. J. Anal. Chem. 2008, 80, 1027–1032.

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

1673

Figure 6. (a) Dependence of maximal evaporation rate on gas flow rate and temperature. (b) Dependence of maximal evaporation rate on liquid flow rate and elution time.

Figure 7. Reproducibility as a function of sampling rate. (RSD: relative standard deviation, ISTD: internal standard).

the proposed mechanism of DESI. Investigation of electrically nonconductive surfaces with DESI involves the formation of a liquid film and build-up of electrical charge on a confined surface area.4,33 Because of this mechanism, there is a well-defined equilibration time for DESI after a new surface sample is introduced. This initial (33) Costa, A. B.; Cooks, R. G. Chem. Commun. 2007, 3915–3917.

1674

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

equilibration period was estimated to be 100-300 ms for PTFE frit material, which time frame sets strong limitations on the maximum throughput for quantitative DESI analysis. Although there is ion formation during equilibration, quantitative analysis is not feasible because of signal fluctuation. Reproducibility and LLOD values obtained at 0.5 sample/s sampling rate clearly show that the analytical parameters of the method are comparable to those of HPLC-MS-based techniques routinely used for therapeutic drug monitoring. Sampling Card. An analogous method was developed for the analysis of samples adsorbed on sheet geometry adsorbent material. Sheet adsorbents include SPE discs and various filter papers. SPE discs are particularly widespread in environmental analysis, especially for the extraction of surface water samples, while filter paper-type sample collection devices are widely used in medical diagnostics for handling biological fluid samples. The analysis of samples present on sheet adsorbents by the means of desorption ionization mass spectrometry has been demonstrated in the case of dried spots of whole blood for drug monitoring and diagnostics of metabolic diseases.3 While the synergism between this sample format and desorption ionization (DI, especially DESI) mass spectrometry is evident, direct DI analysis of samples does not provide full availability of analyte molecules present in the bulk material of the adsorbent. DI-MS methods generally analyze molecules present on the condensed/gas interface; thus molecules “hidden” deep in solid phase adsorbent do not undergo ionization. Furthermore, in the case of DESI analysis, spray solvent tends to diffuse into the adsorbent material, which process decreases the availability of the analyte even further, and also hinders the formation of continuous liquid film on the surface. (Continuous liquid film has been considered being prerequisite for ion formation in DESI.33) Surface elution combined with sample card format (Figure 1b) provides a solution for the problems associated with DI-MS analysis of sheet adsorbents. To eliminate disadvantageous lateral diffusion effects, confined filter paper discs were embedded into a plastic card and were lined with hydrophobic PTFE membranes on one side. The sample format was termed SPE card. The PTFE membrane has a dual role in the depicted setup; it keeps the aqueous sample in the filter paper when the sample is applied and also provides an optimal surface for DESI analysis. Both application and elution of the sample occurs from the exposed surface of filter paper; hence eluted analyte molecules are concentrated on PTFE membrane prior to DESI analysis. Performance of the sampling device and the sample preparation method was tested using whole blood samples spiked with 10-1000 ng/ mL Cyclosporine A. Dried blood spots (DBS) on Whatman 903 filter paper were used with and without surface elution. Samples were applied directly onto filter paper, and excess blood was removed from card surface. In the case of surface elution of filter paper, eluted area was confined using a stainless steel clamp. Cyclosporine A was eluted using acetone/isopropanol 1:1. Results are summarized in Table 3. As it is shown in Table 3, surface elution gives almost an order of magnitude lower LLOD, which is further improved in the case of sampling card. There was also a considerable difference between the reproducibility of the different experimental setups. This latter feature was associated with the reproducibility of the

Table 3a native surface surface DBS on elution/ elution/ filter paper filter paper sampling card LLOQ (S/N ≈ 3) ng/mL RSD no ISTD Cyclosporine D ISTD

800

120

45% 26%

18% 12%

50 12% 4.2%

a DBS, dried blood spots; LLOQ, lower limits of quantitation; RSD, relative standard deviation; ISTD, internal standard; S/N, signal to noise ratio.

sampling volume. In the case of unprocessed filter paper the distribution of the analyte molecules on the surface depends on a number of factors including inhomogenity of the filter paper, exposure to environmental effects, and blood clotting pattern among others. When a confined area (d ) 2 mm) of dried blood spot undergoes surface elution, most of these factors are eliminated; however, the uneven distribution of the blood in the paper still results in poor reproducibility. The sampling card eliminates this effect, since in that case the adsorbent is completely saturated with the biological fluid, and the excess blood is simply wiped off from the card surface. Strictly quantitative analysis of dried blood spot samples is usually not feasible because of poor reproducibility of sample volume discussed above. The use of internal standard (ISTD) also involves problems, since internal standard can only be added to liquid phase sample prior to application onto the filter paper; however, in that case the simplicity of the sampling procedure is compromised. The well-defined adsorbed volume in case of the sampling card offers a solution for this problem. Since the volume of biological fluid retained by the well-defined filter paper disk is reproducible (standard deviation of blood volume was found to be less than 5% by weight measurement) internal standard solution can be dried onto filter paper prior to analysis. This technology, using Cyclosporine D as internal standard and heparinized blood containing known amounts of Cyclosporine A, resulted in highly reproducible results. Standard deviation varied between 3.1 and 6.8% for different concentration levels, compared to 10-15% SD of direct DESI analysis of the blood spots.

The described sampling card offers a robust, portable sample format, especially in the case of clinico-chemical applications. The sampling cards combined with described surface elution technology make a proper basis for a new type of centralized clinical diagnostics scheme. CONCLUSIONS The described sample preparation scheme for desorption ionization mass spectrometry offers a simple method for concentrating the analyte content of a large volume of liquid sample into small, confined surface areas. Although the method does not increase the absolute LOD of DESI-MS methods dramatically, almost arbitrarily low LLOQ values become achievable if a sufficient volume of sample is available. The presented method makes the application of recently developed, ambient ionization methods feasible for regular analytical purposes. DESI-MS based methods do not provide sufficient sensitivity, not even in the case of analytical problems that can be solved routinely using LC-MS methodologies. To fully utilize the advantageous features of DESI and related methods, enhancement of these methods with appropriate sample preparation steps is imperative in the case of quantitative HT applications. The presented method is able to increase the sensitivity of ambient ionization methods to a practical level, while the unique simplicity and rapidity of the techniques are not compromised. The method can also be combined with other desorption ionization methods of choice, including MALDI, and improve the sensitivity and reproducibility of these methods. ACKNOWLEDGMENT Work was supported by Hungarian National Office for Research and Technology under grants ComGenSE (OM-00240/ 2005) and SPEDI_007 (2006ALAP1-00146/2006). SUPPORTING INFORMATION AVAILABLE Further details are given in Figures S1 and S2. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review November 24, 2008. Accepted January 8, 2009. AC8024812

Analytical Chemistry, Vol. 81, No. 4, February 15, 2009

1675