Anal. Chem. 2003, 75, 805-811
Demonstration of Direct Bioanalysis of Drugs in Plasma Using Nanoelectrospray Infusion from a Silicon Chip Coupled with Tandem Mass Spectrometry Jean-Marie Dethy,*,† Bradley L. Ackermann,‡ Claude Delatour,† Jack D. Henion,§ and Gary A. Schultz§
Lilly Development Center S.A., Eli Lilly and Company, Mont-Saint-Guibert, B-1348, Belgium, Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, Indiana 46285, and Advion BioSciences, Inc., Ithaca, New York 14850
Quantitative bioanalysis by direct nanoelectrospray infusion coupled to tandem mass spectrometry has been achieved using an automated liquid sampler integrated with an array of microfabricated electrospray nozzles allowing rapid, serial sample introduction (1 min/ sample). Standard curves prepared in human plasma for verapamil (r2 ) 0.999) and its metabolite norverapamil (r2 ) 0.998) were linear over a range of 2.5-500 ng/ mL. Based on the observed precision and accuracy, a lower limit of quantitation of 5 ng/mL was assigned for both analytes. Sample preparation consisted of protein precipitation with an organic solvent containing the structural analogue gallopamil as an internal standard. Protein precipitation was selected both to maximize throughput and to test the robustness of direct nanoelectrospray infusion. Aliquots of supernatant (10 µL) were transferred to the back plane of the chip using disposable, conductive pipet tips for direct infusion at a flow rate of 300 nL/min. Electrospray ionization occurred from the etched nozzles (30-µm o.d.) on the front of the chip, initiated by a voltage applied to the liquid through the pipet tip. The chip was positioned near the API sampling orifice of a triple quadrupole mass spectrometer, which was operated in selected reaction monitoring mode. Results are presented that document the complete elimination of system carryover, attributed to lack of a redundant fluid path. This technology offers potential advantages for MS-based screening applications in drug discovery by reducing the time for methods development and sample analysis. A developing trend in the analytical sciences is toward miniaturization. In addition to advantages, such as reduced reagent consumption and laboratory space conservation, lab-on-a-chip * Corresponding authors. E-mail:
[email protected]. Phone: 011-32-10476324. Fax: 011-32-1047-6925. † Eli Lilly and Company, Lilly Development Center S.A., Mont-Saint-Guibert, B-1348, Belgium ‡ Eli Lilly and Comapny, Lilly Research Laboratories, Lilly Corporate Center, Indianapolis, IN 46285 § Advion BioSciences, Inc., Ithaca, NY 14850. 10.1021/ac0260692 CCC: $25.00 Published on Web 01/17/2003
© 2003 American Chemical Society
technology offers new prospects for laboratory innovation and automation. Despite the introduction of numerous formats for microfluidics,1-3 lab-on-a-chip applications rely primarily on spectroscopic detection4,5 in part due to the lack of a viable interface between chip technology and mass spectrometry (MS). Recently, electrospray ionization (ESI) has been successfully integrated into a chip format allowing for an array of 100 individual ESI nozzles to reside on a single silicon chip.6 Applications of this technology, known as the ESI Chip, have been primarily directed toward proteomics7 and have been qualitative in nature. Quantitative applications, on the other hand, have yet to be fully explored. Liquid chromatography coupled to tandem mass spectrometry (LC/MS/MS) has emerged as the preferred technology for quantitative determination of drugs and metabolites in biomatrixes.8,9 Although routinely applied to clinical sample analysis, LC/MS/MS is now widely implemented in many applications of ADME (absorption, distribution, metabolism, excretion) in drug discovery. In vivo10,11 and in vitro12 applications of this technology have recently been reviewed. In the discovery setting, considerable emphasis has been placed on sample throughput. Several approaches have been reported to increase throughput including automated 96-well extraction,13 on-line sample introduction,14-16 (1) Quake, S. R.; Scherer, A. Science 2000, 290, 1536-1540. (2) Chien, R. L.; Parce, J. W. Fresenius J. Anal. Chem. 2001, 371, 106-111. (3) McKnight, T. E.; Culbertson, C. T.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 2001, 73, 4045-4049. (4) Gottschlich, N.; Culbertson, C. T.; McKnight, T. E.; Jacobson, S. C.; Ramsey, J. M. J. Chromatogr., B 2000, 745, 243-249. (5) Chabinyc, M. L.; Chiu, D. T.; McDonald, J. C.; Stroock, A. D.; Christian, J. F.; Karger, A. M.; Whitesides, G. M. Anal. Chem. 2001, 73, 4491-4498. (6) Schultz, G. A.; Corso, T. N.; Prosser, S. J.; Zhang, S. Anal. Chem. 2000, 72, 4058-4063. (7) Van Pelt, C. K.; Zhang, S.; Henion, J. J. Biomol. Technol. 2002, 13, 72-84. (8) Gilbert, J. D.; Olah. T. V.; Barrish, A.; Greber, T. F. Biol. Mass Spectrom. 1992, 21, 341-346. (9) Murphy, A. T.; Bonate, P. L.; Kasper, S. C.; Gillespie, T. A.; DeLong, A. F. Biol. Mass Spectrom. 1994, 23, 585-589. (10) Jemal, M. Biomed. Chromatogr. 2000, 14, 422-429. (11) Ackermann, B. L.; Berna, M. J.; Murphy, A. T. Curr. Top. Med. Chem. 2002, 2, 53-66. (12) Kearns, E. H. J. Pharm. Sci. 2001, 90, 1838-1858. (13) Janiszewski, J.; Schnieder, R.; Hoffmaster, K.; Swyden, M.; Wells, D.; Fouda, H. Rapid Commun. Mass Spectrom. 1997, 11, 1033-1037.
Analytical Chemistry, Vol. 75, No. 4, February 15, 2003 805
fast gradient elution,17 and monolithic stationary phases.18 Despite major advances toward increasing LC/MS throughput, the time needed for LC method development and chromatographic elution are significant compared to the actual time spent on MS detection. These factors can supply considerable overhead to screening applications which rely on MS detection. In this paper, direct bioanalysis is examined using nanoelectrospray infusion from a silicon chip. The primary motivation for direct bioanalysis is throughput, since the LC duty cycle as well as the time invested in establishing LC conditions are eliminated. In the interest of throughput, protein precipitation was used as the method for sample preparation. Not only does this approach represent a minimal procedure for sample preparation, it presents a demanding test case for this new technology. Direct bioanalysis suffers from two major disadvantages. The first is a greater likelihood of ion suppression.19 The second issue is a higher occurrence of interfering peaks, either from matrix components or drug metabolites.20 Despite these drawbacks, precedence exists for conducting bioanalysis without the benefit of LC separation. Chen and Carvey reported a validated assay for topiramate in human plasma using flow injection analysis (FIA) ESI-MS following liquid-liquid extraction.21 Direct sample introduction following solid-phase extraction (SPE) has also been accomplished. A novel example is the commercial apparatus, originally developed by Olech and co-workers, which couples 96-well disk SPE in a format allowing direct elution into the mass spectrometer.22 Off-line 96-well SPE followed by FIA sample introduction has also been considered. A recent example by Zheng and co-workers demonstrated the viability of this approach, relative to LC/MS, for the hepatic metabolic stability determination.23 The introduction of a microchip format for infusion bioanalysis has two distinct advantages over other direct approaches. First, miniaturization creates potential cost savings, in terms of both reagent consumption and waste disposal. Second, because the ESI Chip format uses an individual pipet tip and ESI nozzle for each sample, it is now possible to completely eliminate system carryover. Initial results are presented using this technology, which indicate that direct bioanalysis in the low-nanogram/milliliter regime is possible combining high throughput, minimal sample preparation and freedom from system carryover. (14) Zell, M.; Husser, C.; Hopfgartner, G. J. Mass Spectrom. 1997, 32, 23-32. (15) McLoughlin, D. A.; Olah, T. V.; Gilbert, J. D. J. Pharm. Biomed. 1997, 15, 1893-1901. (16) Ayrton, J.; Dear, G. J.; Leavens, W. J.; Mallett, D. N.; Plumb, R. S. Rapid Commun. Mass Spectrom. 1997, 11, 1953-1958. (17) Romanyshyn, L.; Tiller, P. R.; Hop, E. C. A. Rapid Commun. Mass Spectrom. 2000, 14, 1662-1668. (18) Wu, J. T.; Zeng, H.; Deng, Y.; Unger, S. E. Rapid Commun. Mass Spectrom. 2001, 15, 1113-1119. (19) King, R.; Bonfiglio, R.; Fernandez-Metzler, C.; Miller-Stein, C.; Olah, T. J. Am. Soc. Mass. Spectrom. 2000, 11, 942-950. (20) Jemal, M.; Xia, Y.-Q. Rapid Commun. Mass Spectrom. 1999, 13, 97-106. (21) Chen, S.; Carvey, P. M. Rapid Commun. Mass Spectrom. 2001, 15, 159163. (22) Olech, R. M.; Pranis, R. A.; Jacobson, J. R.; Perman, C. A.; Boman, B. A.; Soldo, J.; Speziale, R.; Astle, T. W.; Cole, M. J.; Janisewski, J. S.; Whalen, K. W. Proceedings of the 49th Conference on Mass Spectrometry and Allied Topics, Chicago, IL, May 27-31, 2001. (23) Zheng, J. J.; Lynch, E. D.; Unger, S. E. J. Pharm. Biomed. Anal. 2002, 28, 279-285.
806 Analytical Chemistry, Vol. 75, No. 4, February 15, 2003
Figure 1. Chemical structures and nominal molecular weights for verapamil, its active metabolite norverapamil, and the internal standard, gallopamil.
EXPERIMENTAL SECTION Reagents. Drug-free human plasma was obtained from the blood of healthy volunteers collected using ammonium heparin as the anticoagulant. All chemicals were of analytical grade and used without further purification. Verapamil hydrochloride and norverapamil hydrochloride (Figure 1) were purchased from Sigma Chemical Co. (St. Louis, MO). Gallopamil hydrochloride (Figure 1), used as internal standard, was obtained from Recordati (Milano, Italy). Acetic acid, acetonitrile, ethanol, and water were purchased from Acros Organics (Geel, Belgium). Stock solutions at 1 mg/mL were prepared in acetonitrile containing 0.1% acetic acid for all compounds. Dilution of the stock solution with acetonitrile (0.1% acetic acid) yielded the working solutions at concentrations of 100, 10, 1, 0.1, and 0.01 µg/mL. All stock and working solutions were stored in the refrigerator (4 °C). Procedure. Plasma aliquots (100 µL) were mixed with 100 µL of internal standard solution containing 50 ng of gallopamil followed by the addition of 400 µL of acetonitrile/ethanol/acetic acid (90/10/0.1; v/v/v). The samples were mixed in a vortex mixer for 30 s to ensure complete protein precipitation and then centrifuged for 10 min (10000g). The supernatant was transferred to glass test tubes and evaporated to dryness under a gentle stream of nitrogen at ambient temperature. The residue was redissolved in 200 µL of acetonitrile (0.1% acetic acid) followed by an additional centrifugation step (10 min at 10000g) to minimize particulate matter. The samples were transferred to 96-well ScreenMates plates (Matrix Technologies Corp., Cheshire, U.K.) for direct ESI-MS analysis using the described chip-based nanoelectrospray infusion system. Calibration curves were prepared by fortifying drug-free human plasma with both verapamil and norverapamil. Appropriate volumes of the working solutions were added to drug-free plasma to produce standard samples of the following concentrations: 2.5, 5.0, 10, 25, 50, 100, 250, and 500 ng/mL. The internal standard concentration was 500 ng/mL. All samples were treated as described above in preparation for analysis.
Table 1. Precision and Accuracy Data Obtained from the Nanoelectrospray Infusion Determination of Verapamil and Norverapamil in Human Plasma verapamil
norverapamil
std mean RSD accuracy mean RSD accuracy (ng/mL) (ng/mL) (%) (%) (ng/mL) (%) (%) N
Figure 2. Illustration showing the interface between the pipet tip sample delivery system and the ESI Chip. A robotic probe delivers sample (up to 10 µL) through a conductive pipet tip, which interfaces directly to the back plane of the ESI Chip. Voltage required for nanoelectrospray along with a slight positive pressure (N2) is delivered to the sample through the robotic probe. The ESI Chip was positioned near the atmospheric pressure ionization (API) sampling orifice of a triple quadrupole mass spectrometer.
Apparatus. Direct nanoelectrospray infusion analysis occurred using a NanoMate 100 incorporating ESI Chip technology (Advion BioSciences, Ithaca, NY). A schematic representation of the chip interface is shown in Figure 2. Each chip contained 100 nanoelectrospray nozzles (8 µm i.d. by 30 µm o.d.) oriented in a 10 × 10 array with a 2.25-mm pitch between each nozzle. The chips were fabricated from a monolithic silicon substrate using deep reactive ion etching (DRIE) and other standard microfabrication techniques.6 The ESI Chip was inserted into a NanoMate 100 robotic interface, which performed serial sample introduction using individual, disposable pipet tips (Advion BioSciences). In all cases, 10 µL of sample solution was delivered to the back plane of the ESI Chip through a conductive, graphite-containing pipet tip that mated directly to the chip centered in the region containing the through-chip channel. The sample solution was delivered at a flow rate of ∼300 nL/min under a slight positive pressure of nitrogen (0.6 psi) applied through the pipet tip to ensure constant sample flow to the chip. Electrospray ionization occurred from individual nozzles located on the opposite (front) side of the chip initiated by a +2000 V potential applied through the pipet tip. The NanoMate 100 was custom mounted onto the existing atmospheric pressure ionization (API) interface of a Micromass Quattro II triple quadrupole mass spectrometer (Manchester, U.K.). The ESI Chip was positioned ∼5 mm from the orifice of the external sampling cone held at a potential of +30 V to provide optimum transmission for the analyte ions of interest. A nitrogen cone gas flow rate of 100 mL/min was used along with an ion source temperature of 120 °C. The mass spectrometer was tuned to give unit mass resolution (peak width: 0.7 u, full width halfmaximum) for both quadrupoles Q1 and Q3. All MS/MS data were recorded in selected reaction monitoring (SRM) mode using a collision energy of 35 eV. Q1 was set to monitor the protonated molecules at m/z 455.0, 441.0, and 485.1 for verapamil, norverapamil, and gallopamil, respectively, while Q3 monitored the m/z 164.8 product ion for the three compounds. A dwell time of 400 ms per SRM channel was used with a 100-ms interchannel delay. Samples were injected at 1-min intervals using the ChipSoft v. 4.7.1 U software supplied with the NanoMate 100. This software provided complete control over the injection process as well as the ESI Chip, which was automatically repositioned between injections. A contact closure from the NanoMate 100 to the mass spectrometer was used to initiate SRM data acquisition,
2.5 5.0 10 25 50 100 250 500
2.6 4.6 9.8 24.2 51.0 99.4 252 497
33.3 10.9 7.1 9.2 8.8 6.8 6.9 3.6
102 93 98 97 102 99 101 99
2.7 4.9 8.9 20.6 43.9 92.9 234 445
27.7 19.6 5.7 18.7 16.8 14.1 9.7 6.3
109 98 89 83 88 93 93 89
9 8 9 8 9 6 7 10
which occurred for 30 s. Control of the mass spectrometer was accomplished using MassLynx v. 3.4 software. RESULTS Precision and Accuracy. The bioanalytical precision and accuracy of the ESI Chip technology were assessed through the analysis of multiple standard curves for verapamil and norverapamil prepared in human plasma by protein precipitation. Both analytes were present in each plasma sample analyzed. In this experiment, the first set of standards analyzed was used to define the calibration curve for verapamil and norverapamil. Subsequent analysis of replicate standards resulted in the precision and accuracy data found in Table 1. Figure 3 shows the selected ion current profiles acquired for the first curve obtained by sequential infusion analysis of eight standards spanning a plasma concentration range of 2.5-500 ng/mL. Each rectangular offset in the mass chromatograms represents a single injection from a separate, consecutive nozzle on the chip. These flat-top ion profiles, which are indicative of ESI infusion, were integrated using the MassLynx software. The data in Figure 3 confirm that the analysis duty cycle was ∼1 min/sample. Table 1 contains the precision and accuracy data obtained for verapamil and norverapamil. For both analytes, the precision was under 20% relative standard deviation, except for the lowest standard (2.5 ng/mL). Due to the high degree of imprecision at this level, the lower limit of quantitation (LLOQ) was assigned as 5 ng/mL for both verapamil and norverapamil. Overall the precision ranged from 3.6 to 10.9% RSD for verapamil and from 5.7 to 19.6% RSD for norverapamil. Accuracy values shown in Table 1 represent mean values for the replicates at a given concentration following interpolation of individual concentrations from the calibration curve. The accuracy values for verapamil ranged from 93 to 102% and from 83 to 98% for norverapamil. In evaluating these data, it is acknowledged that overall precision and accuracy found in Table 1 would not meet the more rigorous acceptance criteria associated with GLP validation.24 Nonetheless, these data are consistent with the expectations of discovery bioanalysis.25 Linearity. The calibration curves for verapamil and norverapamil were linear over the range tested (2.5-500 ng/mL) and (24) FDA Guidance for Industry. Center for Drug Evaluation and Research. Biopharmaceutics, Bioanalytical Method Validation. Issued 5/2001. http:// www.fda.gov/cder/guidance/index.htm. (25) Korfmacher, W. A.; Cox, K. A.; Ng, K. G.; Veals, J.; Hsieh, Y.; Wainhaus, S.; Broske, L.; Prelusky, D.; Nomeir, A.; White, R. E. Rapid Commun. Mass Spectrom. 2001, 15, 335-340.
Analytical Chemistry, Vol. 75, No. 4, February 15, 2003
807
Figure 3. (a) Series of eight nanoelectrospray infusion ion current profiles representing a single standard curve prepared by spiking drug-free human plasma with the internal standard, gallopamil (upper), verapamil (middle), and norverapamil (lower). Each flat-top deflection in the SRM ion current profiles corresponds to a single sample analyzed from a different ESI nozzle. The standard concentrations represented for verapamil and norverapamil are 2.5, 5.0, 10, 25, 50, 100, 250, and 500 ng/mL. The internal standard concentration was 500 ng/mL. (b) Expanded view of the lowest four standards analyzed in Figure 3a.
were constructed by plotting the peak area ratio of analyte to internal standard versus analyte plasma concentration. Individual calibration curves for verapamil and norverapamil were constructed using the first standard curve analyzed and fit by linear regression with 1/X weighting. The following straight-line equation and correlation coefficient were obtained for verapamil: y ) 0.00180x + 0.00141 (r2 ) 0.999). The corresponding results for norverapamil were y ) 0.00105x + 0.00017 (r2 ) 0.9998). Carryover. Prior to assessing system carryover, the selectivity of the method was first established for both verapamil and norverapamil through the analysis of blank (drug-free) human 808
Analytical Chemistry, Vol. 75, No. 4, February 15, 2003
plasma. Having demonstrated selectivity for both analytes, the question of system carryover was investigated. Figure 4 displays SRM ion current profiles for verapamil (top) and the internal standard gallopamil (bottom) for a series of three consecutive nanoelectrospray infusions. The first infusion period (12.5-13.3 min) indicates the signal observed from a 500 ng/mL verapamil plasma standard. Immediately after this high standard, a drug-free blank plasma sample was infused. Data from this second infusion appear in the period from 14.0 to 14.8 min (defined by the two dotted lines in Figure 4). It is noteworthy that the signal observed during this period did not exceed the level of
Figure 4. SRM ion current profiles for verapamil (upper) and gallopamil (lower) indicating the signal obtained from three sequential nanoelectrospray infusions used to assess system carryover. The first infusion (12.5-13.3 min) shows the signal observed from a 500 ng/mL verapamil human plasma standard. In this figure, the intensity scale for the verapamil trace was compressed 100-fold to permit observation of the baseline signal. This infusion was immediately followed by the analysis of a blank human plasma sample. The analysis period for this sample (14.0-14.8 min) is delineated by two dotted lines. The final infusion (15.5-16.2 min) corresponds to the analysis of a plasma blank containing internal standard.
background noise present during sample loading (i.e., no spray; 13.3-14.0 min). This infusion was followed by the analysis of a blank plasma sample containing the internal standard (15.5-16.2 min). The results from this experiment document the complete elimination of system carryover and are representative of the standard performance of this system. As expected, no carryover was detected in the concomitant analysis of norverapamil (data not shown). Sample Preparation. Although several sample preparation strategies (e.g., solid-phase or liquid-liquid extraction) are compatible with the ESI Chip technology, the present investigation was conducted using simple protein precipitation, as this would represent the most demanding test case. Two schemes for protein precipitation were investigated. In the first scheme (process 1), a 100-µL plasma sample fortified with 100 µL of internal standard was precipitated using 200 µL of acetonitrile/ethanol/acetic acid (90/10/0.1; v/v/v). In this case, the supernatant was infused immediately after the centrifugation step. Process 2 was similar to process 1, except that the supernatant from precipitation was dried down under nitrogen, reconstituted in 200 µL of acetonitrile/ ethanol /acetic acid (90/10/0.1; v/v/v), and recentrifuged as described in the Experimental Section. While process 1 obviously results in greater overall throughput, process 2 was found to yield superior data as well as greater robustness. Processes 1 and 2 are compared in Figure 5, which displays a series of standards at the low end of the calibration curve for verapamil. An approximate
5-fold enhancement in signal-to-noise ratio was observed using the expanded procedure (process 2), which was greater than the effect predicted from sample concentration alone (i.e., 2-fold). Although a number of factors may be proposed to account for this difference, the amount of water present in the sample is obviously important. Ionization Suppression. It is generally known that competition for electrospray ionization by matrix components can suppress the ion current signal for the analyte and can lead to poor precision and accuracy in bioanalytical data.26 Owing to the lack of on-line chromatography in the present study, an assessment of ionization suppression was undertaken. In this assessment, neat 50-ng aliquots of verapamil, norverapamil, and gallopamil were transferred to glass tubes, dried down, and reconstituted in 200 µL of acetonitrile (0.1% acetic acid). Alternatively, 50-ng aliquots were spiked into tubes containing dried supernatant obtained from the precipitation of 100 µL of drug-free human plasma. These samples were also reconstituted in 200 µL of acetonitrile (0.1% acetic acid). Since the recovery of the two methods could be considered equivalent, the decrease in analyte signal observed for the second set of samples was attributed to matrix ion suppression. The ion current signal decrease from ion suppression corresponded to 87, 91, and 86% for verapamil, norverapamil, and gallopamil, respectively. (26) Matuszewski, B. K.; Constanzer, M. L.; Chavez-Eng, C. M. Anal. Chem. 1998, 70, 882-889.
Analytical Chemistry, Vol. 75, No. 4, February 15, 2003
809
Figure 5. Comparison of the sensitivity for verapamil human plasma standards (2.5, 5.0 and 10 ng/mL) obtained using two sample preparation schemes. Process 1: 100 µL of plasma sample plus 100 µL of internal standard was precipitated using 200 µL of acetonitrile/ethanol/acetic acid (90/10/0.1; v/v/v), centrifuged, and directly infused using the NanoMate 100. Process 2: an expanded version of process 1 involving an evaporation of the original supernatant followed by reconstitution in 200 µL of acetonitrile (0.1% acetic acid) and recentrifugation. The sample was infused under the same analysis conditions as process 1.
DISCUSSION This preliminary investigation on the use of the ESI Chip technology for bioanalysis indicates the potential of this technology for drug discovery bioanalysis. Using direct nanoelectrospray infusion, low-nanogram/milliliter quantitation of a known drug and its metabolite were obtained with precision and accuracy acceptable for many screening applications in drug discovery.25 It is noteworthy that these results were obtained with minimal sample cleanup and employed a structural analogue internal standard in the presence of high matrix ionization suppression. Each analyte tested in this study experienced an approximate 10-fold loss in signal attributed to matrix ionization suppression. It is therefore reasonable to expect that subnanogram/milliliter quantitation could be achieved if an appropriate sample cleanup method, such as SPE or liquid-liquid extraction, was employed. This belief is reinforced by the fact that, in contrast to most extraction methods, no sample preconcentration occurred in the present investigation. The vintage of the mass spectrometer used in this report must also be considered. Although sample cleanup will likely be needed for many applications of nanoelectrospray infusion, the results presented herein suggest that, for certain applications, additional cleanup may not be necessary thereby saving time and expense. For example, an LLOQ of 5 ng/mL is certainly viable for early exposure assessment during lead generation. In addition, certain in vitro applications may also be amenable to direct nanoelectrospray infusion analysis due to higher analyte levels typically present in these experiments. The benefit or necessity for additional sample cleanup must obviously be evaluated on a caseby-case basis. The current sample throughput displayed by this technology (1 min/sample) is competitive with existing technologies utilizing LC/MS detection. Although faster analysis times can be achieved with the NanoMate 100, the rate is ultimately determined by the time needed for reliable MS data acquisition (currently 30 s). 810
Analytical Chemistry, Vol. 75, No. 4, February 15, 2003
Despite the legitimate focus on sample throughput, it is believed that the major benefit from nanoelectrospray infusion will be an overall simplification of the bioanalytical process. The fact that infusion can often occur at a higher organic solvent composition than is possible by reversed-phase LC is a further advantage of nanoelectrospray infusion. The influence of organic solvent composition on ESI sensitivity is well known and is believed to account for the superior results obtained using the extended sample preparation process referred to as process 2 (Figure 5). Despite this effect, the relative instability of the signal in process 1 was unexpected since applications involving highly aqueous compositions, such as those present in the study of noncovalent interactions,27 are readily sprayed using the ESI Chip. It is possible that elimination of water by process 2 also reduced the level of salt present in the final sample due to insolubility. The insertion of a second centrifugation step may also have had a favorable impact on signal stability by removing any suspended matter in the sample. The magnitude of these factors is not clear at the present time. The elimination of system carryover is an important attribute of this technology. According to the 2001 FDA Guidance on Bioanalytical Methods Validation, an interfering peak is defined as any signal that exceeds 20% of the LLOQ.24 This requirement often restricts the dynamic range of bioanalytical assays in drug development and results in considerable time and effort being expended to reduce system carryover to an acceptable level. The instrumentation used in this report eliminates this problem by avoiding a redundant fluid path. This advantage was not fully exploited in the present study, which only examined a range of 250-fold. A more detailed study awaits further investigation. The robustness of nanoelectrospray infusion was not rigorously investigated in the present study primarily because of the rapidly (27) Beverly, M. B.; Julian, R. K., Jr.; Schultz, G. A.; Henion, J. Determination of Aminoglycoside Dissociation Constants Using a Microfabricated Electrospray Chip. Submitted for publication in Rapid Commun. Mass Spectrom.
evolving nature of this technology. Nevertheless, the use of an internal standard is recommended for any quantitative measurement to control for nozzle-to-nozzle variability. An example in the present study is shown in Figure 4 where the signal for the internal standard (gallopamil) decreased by approximately half of its value from the 500 ng/mL verapamil standard. Although such phenomena occur, the data in Figure 3 give a more representative view of the consistency obtained with the ESI Chip technology. One of the biggest issues facing direct bioanalysis methods is the increased potential for interfering peaks, particularly from labile metabolites. Certain metabolite types (e.g., glucuronides, sulfates, N-oxides), which can potentially degrade to the parent during handling or fragment in the ion source to yield the protonated molecule of the parent drug, can give rise to spuriously high plasma concentrations. Experiments are currently being conducted to investigate this phenomenon. It is important to note that such metabolite “cross-talk” can often be reduced or eliminated by incorporating a sample cleanup step prior to infusion analysis. A precise examination of the cost savings of this technology is difficult owing to the situational nature of the question. Nonetheless, the issue of mobile-phase consumption and LC waste stream generation is straightforward. Conservatively speaking, nanoelectrospray infusion consumes 1000-fold less mobile phase compared to high-flow LC/MS methods. Additional savings from the elimination of HPLC equipment may be cited, although this would depend on the particular needs of a given laboratory. Finally, the influence that nanoelectrospray itself has on bioanalysis by direct infusion remains to be determined. On the basis of existing models for ESI,28,29 it is plausible to suggest that the additional droplet surface area present under nanoelectrospray (relative to conventional ESI) could lead to advantages with respect to sensitivity, linear dynamic range, or ion suppression. These factors will be studied as part of a subsequent investigation. (28) Kebarle, P.; Tang, L. Anal. Chem. 1993, 65, 972A-986A. (29) Enke, C. G. Anal.Chem. 1997, 69, 4885-4893.
CONCLUSIONS Despite the current interest in lab-on-a-chip technology in analytical chemistry, applications involving mass spectrometry have lagged behind other more conventional modes of detection. Utilizing a novel microchip format for nanoelectrospray infusion ionization, the feasibility of quantitative bioanalysis was demonstrated. Precision and accuracy obtained in the low-nanogram/ milliliter regime for verapamil and norverapamil, in human plasma, indicate the potential of this technology for bioanalytical screening applications in drug discovery. The instrumentation used in this study provided freedom from system-related carryover since each sample was analyzed using an individual ESI nozzle and was delivered to the chip using a separate pipet tip by a dedicated robotic interface. In addition, the success obtained using simple protein precipitation offers significant optimism for the use of this technology in combination with existing bioanalytical sample cleanup procedures to achieve subnanogram/milliliter quantitation in biological matrixes. Future studies will examine the usefulness of sample cleanup methods in conjunction with nanoelectrospray infusion to address the issues of matrix ionization suppression and metabolite cross-talk. Finally, it is hoped that this work will stimulate further investigation into the use of MS detection for quantitative applications involving microfluidic technology. ACKNOWLEDGMENT The authors extend their appreciation to the many individuals at Advion BioSciences, Inc. who contributed to the research and development of the ESI Chip technology and associated robotic infusion interface.
Received for review August 22, 2002. Accepted November 19, 2002. AC0260692
Analytical Chemistry, Vol. 75, No. 4, February 15, 2003
811
