Direct Plate-Reader Measurement of Nitric Oxide Released from

Aug 3, 2010 - This device enables the use of a standard high-throughput tool (the plate ... An Integrated Microfluidic Device for Monitoring Changes i...
0 downloads 0 Views 231KB Size
Download PDF
Anal. Chem. 2010, 82, 7492–7497

Direct Plate-Reader Measurement of Nitric Oxide Released from Hypoxic Erythrocytes Flowing through a Microfluidic Device Stephen T. Halpin and Dana M. Spence* Department of Chemistry, Michigan State University, East Lansing, Michigan 48824 The ability to perform a fluorescence-based quantitative determination of a biologically important analyte directly released from mammalian cells using a standard microtiter plate reader to measure wells integrated into a microfluidic device is reported. Specifically, the amount of nitric oxide (NO) released from flowing erythrocytes (ERYs) exposed to a hypoxic buffer is measured using a fluorescein-based probe. The ERYs are pumped through channels in one layer of the poly(dimethylsiloxane) (PDMS) device; as these cells release NO, it flows through a porous polycarbonate membrane to the probe. The device is then placed into a standard microtiter plate reader for measurement, with the entire calibration and analyte determination occurring simultaneously. Using this method, NO release from hypoxic ERYs was determined to be 6.9 ( 1.8 µM, a significantly increased value in comparison to that from normoxic ERYs of 0.60 ( 0.04 µM (p < 0.001, n ) 4 rabbits). Furthermore, the reproducibility (reported as a %RSD) of measuring fluorescence standards was 3.5%. Detection limits, dynamic range, and optimal membrane pore diameters are also reported. This device enables the use of a standard highthroughput tool (the plate reader) to measure analytes in a microfluidic device, the ability to improve the quantitative determination of a relatively unstable molecule (NO), and the incorporation of a flow component and blood constituent into a system that can be combined with microtiter plate technology. Microfluidic systems for biological analyses, especially cellular analyses, is a research area that has greatly expanded.1-7 Microfluidic-based systems offer many desirable features for cellular analyses including fast analysis times (on the order of seconds if direct detection is used), the possibility for mass

production, small channel volumes (on the order of nanoliters), and the ability to inject and manipulate sample volumes as small as picoliters. Traditionally, these devices are fabricated in glass8 or other materials such as plastics, low temperature ceramics, and poly(dimethylsiloxane) (PDMS).9 PDMS has gained widespread use in cellular analyses due to the ease of fabrication and the fact that PDMS is gas permeable and can be reversibly sealed to a variety of substrates.10 Micron-sized channels are made by soft lithography techniques, where a “master” of a positive relief structure is first fabricated on a silicon wafer and then a mixture of PDMS prepolymer/curing agent is poured against the master. Important to this work, our group11,12 and others13-15 have shown that PDMS can be used to create 3-dimensional fluidic devices. Initial works showed that PDMS-based devices can be used to pattern both proteins and cells, and because PDMS is permeable to gases, the cells behave normally.16-18 It has been shown that using PDMS devices to culture cells, allows not only patterning of cells but also delivery of agents to the adhered cells and monitoring of events that occur over select cells or a portion of a cell.18 More recent work has been aimed at operations such as cell culture,2,19 cell sorting and handling,20 and single cell analysis.7,21-23 (8) (9) (10) (11) (12) (13) (14) (15) (16)

(17) (18)

* To whom correspondence should be addressed. E-mail: dspence@ chemistry.msu.edu. (1) Young, E. W. K.; Beebe, D. J. Chem. Soc. Rev. 2010, 39, 1036–1048. (2) El-Ali, J.; Sorger, P. K.; Jensen, K. F. Nature 2006, 442, 403–411. (3) Martin, R. S.; Root, P. D.; Spence, D. M. Analyst (Cambridge, U. K.) 2006, 131, 1197–1206. (4) Price, A.; Culbertson, C. T. Anal. Chem. 2007, 79, 2614–2621. (5) Sims, C. E.; Allbritton, N. L. Lab Chip 2007, 7, 423–440. (6) Wu, M. H.; Huang, S. B.; Lee, G. B. Lab Chip 2010, 10, 939–956. (7) Wang, Y.; Chen, Z. Z.; Li, Q. L. Microchimica Acta 2010, 168, 177–195.

7492

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

(19) (20) (21) (22)

Dolnik, V.; Liu, S.; Jovanovich, S. Electrophoresis 2000, 21, 41–54. Becker, H.; Gartner, C. Electrophoresis 2000, 21, 12–26. McDonald, J. C.; Whitesides, G. M. Acc. Chem. Res. 2002, 35, 491–499. Genes, L. I.; Tolan, N. V.; Hulvey, M. K.; Martin, R. S.; Spence, D. M. Lab Chip 2007, 7, 1256–1259. Tolan, N. V.; Genes, L. I.; Subasinghe, W.; Raththagala, M.; Spence, D. M. Anal. Chem. 2009, 81, 3102–3108. Hulvey, M. K.; Martin, R. S. Anal. Bioanal. Chem. 2009, 393, 599–605. Kuo, T.-C.; Cannon, D. M.; Shannon, M. A.; Bohn, P. W.; Sweedler, J. V. Sens. Actuators, A 2003, 102, 223–233. Wu, H.; Odom, T. W.; Chiu, D. T.; Whitesides, G. M. J. Am. Chem. Soc. 2003, 125, 554–559. Chiu, D. T.; Jeon, N. L.; Huang, S.; Kane, R. S.; Wargo, C. J.; Choi, I. S.; Ingber, D. E.; Whitesides, G. M. Proc. Natl. Acad. Sci. 2000, 97, 2408– 2413. Mata, A.; Boehm, C.; Fleischman, A. J.; Muschler, G.; Roy, S. Biomed. Microdevices 2002, 4, 267–275. Takayama, S.; McDonald, J. C.; Ostuni, E.; Liang, M. N.; Kenis, P. J. A.; Ismagilov, R. F.; Whitesides, G. M. Proc. Natl. Acad. Sci. 1999, 96, 5545– 5548. Gomez-Sjoberg, R.; Leyrat, A. A.; Pirone, D. M.; Chen, C. S.; Quake, S. R. Anal. Chem. 2007, 79, 8557–8563. Moehlenbrock, M. J.; Price, A. K.; Martin, R. S. Analyst 2006, 131, 930– 937. Huang, B.; Wu, H.; Bhaya, D.; Grossman, A.; Granier, S.; Kobilka, B. K.; Zare, R. N. Science 2007, 315, 81–84. Wu, H.; Wheeler, A.; Zare, R. N. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 12809–12813. 10.1021/ac101130s  2010 American Chemical Society Published on Web 08/03/2010

Despite these successes, use of microfluidic devices for biological analyses has not become mainstream and there are several possible reasons for this. First, many of the reported chip systems involve complicated fabrication, thereby prohibiting widespread use due to cost and/or time for production. Another problem in the field of microfluidic-based cell analyses is the demonstration of cell culture alone and a lack of integrating analysis steps to measure intra- or extra-cellular molecules or pathways once cells are ready for analysis. Finally, most biomedical researchers are familiar with the use of microtiter plates, well readers, fluid handlers, etc. that are based on 96-well plate technology. Often, it is difficult to couple microchip systems to such technology in high use by the biomedical fields. Continuous demands for high throughput analyses, largely spurred by drug discovery, have created an industry centered on the 96-well plate and equipment for handling these plates in an automated manner. In order to more accurately test drug candidates, an in vitro system which more closely represents in vivo conditions (e.g., incorporation of blood flow, interactions of multiple cell types, real-time measurement of the analytes of interest) should be more useful in preventing false conclusions in the drug development process. There has been other progress toward automating analyses by combining high throughput technologies with microfluidic devices. For example, much work has been done in the area of purifying and sequencing DNA, such as radial systems developed for separation and sequencing that can perform 96 analyses at once,24 and others that can provide sample-in, answer-out capabilities for DNA analysis.25 Advances have been made in utilizing existing high throughput technologies implementing solid phase purification of DNA from blood using polycarbonate devices.26-28 Other attempts at observing enzyme kinetics utilizing absorbance measurements29 and surface plasmon resonance-based detection have been successfully performed.30 Recent attempts by our lab31 to mimic certain characteristics of blood vessels while performing analysis in a high throughput manner have shown that multiple analyte detection in the presence of erythrocytes (ERYs) is possible utilizing a fluorescence macrostereomicroscope with a CCD camera as a detector; however, these studies were limited in throughput by the imaging area of the microscope. Furthermore, the wells of the device were set up in a format that did not replicate those of the 96-well plate. Here, we present results from studies designed to develop a microfluidic device that utilizes a microplate reader as a detector to analyze the nitric oxide release (NO) from hypoxic ERYs. In (23) Wheeler, A. R.; Throndset, W. R.; Whelan, R. J.; Leach, A. M.; Zare, R. N.; Liao, Y. H.; KFarrell, K.; Manger, I. D.; Daridon, A. Anal. Chem. 2003, 75, 3581–3586. (24) Paegel, B. M.; Emrich, C. A.; Weyemayer, G. J.; Scherer, J. R.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 574–579. (25) Easley, C. J.; Karlinsey, J. M.; Bienvenue, J. M.; Legendre, L. A.; Roper, M. G.; Feldman, S. H.; Hughes, M. A.; Hewlett, E. L.; Merkel, T. J.; Ferrance, J. P.; Landers, J. P. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 19272–19277. (26) Witek, M. A.; Hupert, M. L.; Park, D. S. W.; Fears, K.; Murphy, M. C.; Soper, S. A. Anal. Chem. 2008, 80, 3483–3491. (27) Shadpour, H.; Hupert, M. L.; Patterson, D.; Liu, C. G.; Galloway, M.; Stryjewski, W.; Goettert, J.; Soper, S. A. Anal. Chem. 2007, 79, 870–878. (28) Park, D. S. W.; Hupert, M. L.; Witek, M. A.; You, B. H.; Datta, P.; Guy, J.; Lee, J. B.; Soper, S. A.; Nikitopoulos, D. E.; Murphy, M. C. Biomed. Microdevices 2008, 10, 21–33. (29) Kang, J. H.; Park, J. K. Sens. Actuators, B:Chem. 2005, 107, 980–985. (30) Choi, C. J.; Cunningham, B. T. Lab Chip 2007, 7, 550–556. (31) Tolan, N. V.; Genes, L. I.; Subasinghe, W.; Raththagala, M.; Spence, D. M. Anal. Chem. 2009, 81, 3102–3108.

the device described here, the NO that is released from the ERYs is able to immediately (limited by diffusion rate of the gaseous NO to the probe above the porous membrane) react with a fluorescent probe, thus resulting in a quantitative determination of NO that is closer to the time point at which it was actually released from the ERY. EXPERIMENTAL SECTION Preparation of ERYs. ERYs used in this study were obtained from animals following protocols approved by the Animal Investigation Committee at Michigan State University. Male New Zealand white rabbits (2.0-2.5 kg) were anesthetized using ketamine (8 mL/kg, im) and xylazine (1 mg/kg, im) followed by pentobarbital sodium (15 mg/kg, iv). Rabbits were ventilated with room air at a rate of 20 breaths/min using a tidal volume of 20 mL/kg by placing a cannula in the trachea. A catheter was then placed into the carotid artery for administration of heparin (500 units, iv) prior to exsanguination through the same catheter. Approximately 80 mL of whole blood is collected from each animal. Whole blood was then centrifuged three times at 500g at 25 °C for 10 min. After each centrifugation, the plasma and buffy coat were collected for other experimentation before the remaining solution was resuspended and washed twice in a physiological salt solution (PSS) (containing in mM, 4.7 KCl, 2.0 CaCl2, 140.5 NaCl, 12 MgSO4, 21.0 tris(hydroxymethyl)aminomethane, and 5.6 glucose with 5% bovine serum albumin at a final pH of 7.4). All samples were prepared and analyzed within 8 h of harvesting from the animal. Fabrication of PDMS Arrays. In this device, PDMS slabs were fabricated using the established techniques of soft lithography.32,33 Briefly, masters were fabricated by spin coating piranha cleaned silicon wafers with SU-8 50 photoresist (Microchem Corporation, Newton MA) at 500 rpm for 15 s and then 1000 rpm for 30 s, producing features that were 100 µm tall. Caution: Piranha solutions are extremely corrosive. The coated wafer is then exposed, through a transparency mask containing 200 µm channels, to ultraviolet light to induce cross-linking of the photoresist, producing the master after development. The completed device, as described pictorially in Figure 1, consisted of two slabs of PDMS, with a 0.2 µm pore diameter polycarbonate membrane (Steriltech Inc., Kent, WA) sealed between the layers. While other groups34,35 have used PDMS membranes in NO sensing, here, we employ polycarbonate membranes due to success involving cell culture on polycarbonate, and the potential future utility to detect multiple analytes12 that may not be able to permeate silicone-based membranes. To facilitate the thermocuring of the completed device, a 20:1 ratio of bulk polymer to curing agent of Sylgard 184 (Ellsworth Adhesives, Germantown, WI) was used on surfaces to be sealed to the membranes or PDMS. A 5:1 mixture of bulk polymer to curing agent was then used to coat this surface, adding rigidity to the device. Inlets to the 200 µm wide by 100 µm tall channels were punched using 20 gauge tubing, (32) McDonald, J. C.; Duffy, D. C.; Anderson, J. R.; Chiu, D. T.; Wu, H. K.; Schueller, O. J. A.; Whitesides, G. M. Electrophoresis 2000, 21, 27–40. (33) Duffy, D. C.; McDonald, J. C.; Schueller, O. J. A.; Whitesides, G. M. Anal. Chem. 1998, 70, 4974–4984. (34) Mizutani, F.; Yabuki, S.; Sawaguchi, T.; Hirata, Y.; Sato, Y.; Iijima, S. Sens. Actuators, B:Chem. 2001, 76, 489–493. (35) Reynolds, M. M.; Frost, M. C.; Meyerhoff, M. E. Free Radical Biol. Med. 2004, 37, 926–936.

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

7493

Figure 1. Assembly of Device. The device consists of two layers as shown in (a). The bottom contains L-shaped 200 µm wide × 100 µm channels. Syringe pumps are fitted to the device through 20 gauge stainless tubing. Between the layers are polycarbonate membranes with 0.02 µm pore diameters. The top layer contains registration marks molded from a master. The probe wells and waste wells are punched with a 7/32 in. hollow punch. In (b), the assembled device with probe solutions and lines plumbing the inlets is shown.

and wells were punched using a 7/32 in. hollow punch. The two slabs that make up the device were then irreversibly thermally cured together with the polycarbonate track etched membrane sealed between the slabs. Flow to the channels was then supplied using syringe pumps at 1.0 µL/min. After irreversible sealing, the device was cut to the size of a 96-well plate. Alignment of Device for Use in Plate Reader. A critical aspect of the work reported here is the ability to place the microfluidic device in the plate reader such that the internal robotics in the reader align the wells with the light paths uniformly and with high precision. While multiple approaches were explored, the most straightforward method used registration crosses on the masters, which were then aligned to a guide plate. The guide plate was fabricated by attaching a guide transparency with the same registration marks as the device to a glass plate that was cut to the size of a 96-well plate. The locations of the wells on this transparency were then measured and programmed into the software controlling the plate reader (Spectramax M4, Molecular Devices). Careful alignment was then required prior to analysis to be sure the internal robotics of the plate reader positioned the plate properly. Membrane Pore Diameter Optimization. Devices were constructed containing polycarbonate track-etched membranes having pore sizes of 0.6, 0.4, 0.2, and 0.1 µm. A minimum of four devices were employed for each pore size investigated. Five standards containing diethylamine nonoate (DEANO, Cayman Chemical, Ann Arbor, MI) concentrations ranging from 0 µM DEANO to 9 µM DEANO were measured in each device. Precision, limits of detection, and bulk transport of fluid from the flow channels to the analytical wells were analyzed. DAF-FM Concentration Optimization. After determination of the optimal membrane pore size, the 4-amino-5-methylamino2′,7′-difluorescein (DAF-FM, Molecular Probes, Carlsbad, CA) concentration in the analytical wells was varied from 10 to 40 µM, to measure five standards of DEANO that were pumped through four different devices to determine an optimal probe concentration. The optimal probe concentration was determined by an evaluation of detection limit and precision. Fluorescence Determination of Nitric Oxide Standards. Nitric oxide standards were prepared from DEANO in Hanks buffered salt solution (HBSS) at a pH of 7.4. DEANO standards 7494

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

ranging from 0 to 10 µM were prepared and pumped (with syringe pumps) through the device containing 0.2 µm pore diameter membranes at 1.0 µL/min for 30 min underneath wells that contained 40 µL of the optimal concentration of 20 µM DAF-FM in phosphate buffered saline at pH 7.4. After flowing for 30 min, the device was removed from the pumps, aligned on the glass plate, and fluorescence intensity was determined with an excitation wavelength of 488 nm, emission of 520 nm, and an emission filter of 515 nm. Determination of NO Release from Hypoxic/Normoxic ERYs. A 7% hematocrit sample of ERYs was prepared in either hypoxic or normoxic PSS. Hypoxic buffer was prepared with the aid of Oxyrase (Oxyrase Inc. Mansfield, OH), a commercially available enzyme that consumes oxygen in solution. The provided Oxyrase broth suspension (1.5 mL) was added to 15 mL of PSS and allowed to incubate for 30 min. The 30 min incubation was shown to reduce the dissolved oxygen in the buffer to approximately 3% of the saturated concentration, as determined with a Clark type electrode. This method of deoxygenation of the buffer was determined to be superior to purging with a noble gas because lower oxygen saturation was achievable, and foaming of the buffer does not occur with the Oxyrase enzyme. ERY suspensions prepared in hypoxic or normoxic buffer were then pumped through the device. RESULTS Initial adaptation of the described microfluidic device, previously investigated for readout under a macrostereomicroscope,11 required a few key optimizations, most significant being the need for larger wells for proper analysis in the plate reader. Three well diameters were investigated: 1/8, 7/32, and 1/4 in. It was found that while the 1/8 in. diameter well was advantageous in decreasing reagent consumption, it was too small to permit appropriate alignment of the device in the plate reader. The larger 1/4 in. diameter was then investigated, as it was the approximate size of a well on a 96-well plate. In this case, the larger surface prohibited a reproducible and firm seal between the membrane and channel, thus the membrane would “pop up” from the channel, increasing the exposed area of the membrane to the analyte and irreproducibly affecting the mass transport of analyte and analytical signal. A compromise of 7/32 in. diameter wells was selected as

Figure 2. Verification with fluorescein (Fn). Samples are pumped into the device as shown in Figure 1, arranged as indicated in (a). Fn standards were pumped under wells containing phosphate buffered saline for 30 min, aligned in the plate reader, and then analyzed for fluorescence intensity (b). Error bars are n ) 3 samples on the one device, with an r2 value of 0.98 and a detection limit of 30 nM Fn (c).

it could be reproducibly sealed, yet large enough to be analyzed by the plate reader. In contrast to our previous reported device,11 where multiple wells were addressed by a single fluidic channel, the larger size of the wells in the device to be placed in the plate reader only allowed one analytical well per channel, as the signal on a single channel between two wells was not reproducible. We believe that the reduced precision was due to the larger exposed surface area ratio of the channel to membrane reducing the resistance to flow through the membrane, thus resulting in pumped solution to flow across the membrane. In order to test the device shown in Figure 1 for precision and proper alignment in the microplate reader, measurements were performed in the absence of the complexities presented by deoxygenation of buffers and the free radical molecule NO. Specifically, fluorescein solutions from 1.2 to 3.0 µM in water were pumped under wells containing buffer for 30 min at 1.0 µL/min. These concentrations were selected to yield signals of similar magnitudes to those expected with the DAF-FM probe. The device was then aligned with the guide plate and analyzed. The fluorescein standards were pumped through the device channels to the wells as shown in Figure 2a. Next, this device was placed in the microplate reader, and the actual emission values for some of the wells are shown in Figure 2b. A calibration curve based on the values in Figure 2b was constructed and shown in Figure 2c. The linearity (r2 ) 0.98) and precision of the measurements (as indicated by the error bars representing standard deviation associated with each concentration) demonstrates the functionality of the device and, more importantly, the proper alignment of the plate in the reader. To demonstrate precision

Table 1. Optimization of Membrane Pore Diametera membrane pore y intercept slope size (µm) (Fl. U) (µM/Fl U) 0.6 0.4 0.2 0.1

6100 ± 120 5800 ± 370 6240 ± 490 7340 ± 410

110 ± 40 82 ± 13 113 ± 23 173 ± 21

r2

RSD of 5 µM standard (%)

0.67 ± 0.13 0.88 ± 0.01 0.98 ± 0.01 0.90 ± 0.02

21.0 ± 8.0 11.0 ± 1.0 3.0 ± 0.7 9.0 ± 0.9

a Four pore diameters in polycarbonate track etched membranes were investigated in four devices each, and figures of merit were analyzed by measuring triplicate DEANO standards in each device. The ideal pore diameter was found to be 0.2 micrometers. Errors are represented as a standard deviation of four devices.

between wells, an experiment was performed where a 3.0 µM fluorescein standard was pumped under each well, and a relative standard deviation of 3.5% was measured for n ) 16 wells. These experiments were performed on multiple devices to confirm the functionality of the alignment scheme. After alignment and function of the device was verified with fluorescein measurements, a study was performed to determine the optimal pore diameter and DAF-FM concentration. To determine the optimal pore diameter, DEANO standards were pumped into devices containing different membrane pore diameters. Calibration curves were prepared to investigate figures of merit for each pore diameter, which are summarized in Table 1. For each of these experiments, a constant DAF-FM concentration of 20 µM was used with a flow rate of 1.0 µL/min. When the higher diameter pore sizes were used, poor reproducibility and calibration were observed, likely due to the inconsistencies of mass transport across the membrane. As the Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

7495

Table 2. Optimization of DAF-FM Concentrationa probe concentration detection correlation RSD of 1 µM RSD of 9 µM (µM) limit (µM) coefficient standard (%) standard (%) 10 20 30 40

1.32 ± 0.18 0.51 ± 0.08 1.82 ± 0.13 28 ± 13

0.96 ± 0.04 0.97 ± 0.02 0.92 ± 0.03 0.77 ± 0.14

3.1 ± 0.9 3.6 ± 0.3 7.9 ± 1.3 8.0 ± 1.5

7.2 ± 1.3 0.6 ± 0.2 4.0 ± 0.8 6.0 ± 1.1

a DAF-FM concentrations were evaluated with triplicate standards ranging from 1 to 9 µM in n ) 3 devices. The optimal concentration was 20 µM, as it provided the lower detection limit and sufficient reproducibility at the higher concentration. Error is reported as standard deviation.

wells in this device are larger than those previously investigated by our group, there is a lower pressure drop across these wells and a higher tendency for the fluid being pumped to cross the membrane as opposed to flowing through the channel. This occurs inconsistently, resulting in large relative standard deviations for devices containing these membranes, reducing their utility. Meanwhile, the smallest pore size investigated, 0.1 µm diameter, resulted in a higher average y-intercept, likely due to less dilution as a result of bulk transport across the membrane in the smaller membrane pore diameter. Perhaps also due to the smaller pore size, reproducibility was also less ideal because there was less diffusional transport as well. After investigating all of the parameters, an ideal pore diameter of 0.2 µm was selected, as it provided a good combination of reproducibility and sensitivity. Once determination of optimal membrane pore diameter was complete, investigations to determine optimal DAF-FM concentration were performed. Four concentrations were investigated on three devices each. The highest concentration investigated, 40 µM probe, did not demonstrate sufficient correlation to be considered. The remaining three were investigated in terms of their detection limits and dynamic range, which was investigated by considering the reproducibility at the 1 and 9 µM DEANO

standards. As shown in Table 2, the lower concentration of DAFFM exhibited sufficient reproducibility at the lower concentration, though dynamic range was limited at the higher concentrations. From this data, the optimal DAF-FM concentration was determined to be 20 µM. After completion of the optimization of the device parameters, a series of NO standards ranging from 0 µM through 10 µM were pumped under wells containing 20 µM DAF-FM in PBS in three replicate channels per device. These standards were pumped at the same time as two solutions of ERYs. One of the ERY samples was prepared in normoxic buffers, while the other was prepared in an identical buffer having a reduced dissolved oxygen concentration. The overall schematic of the device, with ERYs flowing underneath wells containing the fluorescence probe, is shown in Figure 3. The device containing both NO standards and ERY samples was then aligned in the microplate reader, and the fluorescence readings for each well were recorded. On the basis of the NO standards, it was determined that the NO released from the ERYs under normoxic conditions was 0.60 ± 0.04 µM while those exposed to hypoxic buffers released 6.9 ± 1.8 µM. This value is significantly different from that determined when ERYs are exposed to hypoxia in a static system, centrifuged, and then examined for NO release by mixing the supernatant with DAFFM in a cuvette and measuring with a standard benchtop fluorometer. In the static system, the NO release from normoxic ERYs was 0.69 ± 0.17 µM while those exposed to hypoxic buffers released 2.1 ± 0.1 µM. While both methods (both calibrated with DEANO in their respective system) resulted in increases in NO release for ERYs exposed to hypoxic buffers, we suspect that the difference in the extent of these increases in NO release as measured by the two techniques is due to NO degradation prior to measurement. For those measurements using the benchtop fluorometer, a certain amount of time (∼15-20 min) elapses between exposure of the cells to the NO release stimulus and measurement, which is primarily due to centrifugation of cells to

Figure 3. Determination of NO from ERYs. Nitric oxide release from hypoxic and normoxic ERYs was determined utilizing our device. Shown in (a) is the release of NO from ERYs flowing underneath wells containing DAF-FM. After pumping is complete, the wells are analyzed with a fluorescence excitation of 488 nm and an emission of 520 nm. Shown in (b) is the NO released calibrated to DEANO standards. There is a significant difference (p < 0.01) for n ) 4 rabbits. Error bars are SEM. 7496

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

obtain supernatant. In contrast, in the microfluidic device, the NO reacts with the DAF-FM immediately after diffusion through the porous polycarbonate membrane and is stabilized. CONCLUSION Here, a microfluidic device was employed to determine the release of the reactive NO molecule from ERYs using a microplate reader in the detection step of the analysis. The plate reader requires about 20 s from the time the device is loaded into the instrument to the time emission intensity values are displayed on the screen of the computer. While the microscope takes a single image in less than one second, further analysis of these images to get analytical values adds to the time significantly (multiple minutes). Furthermore, the microscope is limited in the area that it can image at a single time. While a plate reader has a similar limitation (i.e., that it can only analyze a single device at a time), greater throughput with existing equipment could be achieved using racks of plates and robots, the technology for which already exists. Finally, the plate reader’s raw output is a number proportional to fluorescence signal, which can be quickly and easily processed by Laboratory Information Management System (LIMS) systems, as opposed to the microscope, whose output is an image, requiring image processing, which if automated would require significantly more computer resources. In addition to improving throughput, the ability to measure analytes secreted from cells in a near-real time manner is another (36) Sprague, R. S.; Ellsworth, M. L.; Stephenson, A. H.; Lonigro, A. J. Am. J. Physiol. 1996, 271, H2717–H2722. (37) Stamler, J. S.; Jia, L.; Eu, J. P.; Mcmahon, T. J.; Demchenko, I. T.; Bonaventura, J.; Gernert, K.; Piantadosi, C. A. Science 1997, 276, 2034– 2037.

advantage of performing such measurements using a microfluidic device. For example, a potent vasodilator, NO, is known to relax the smooth muscle cells surrounding our blood vessels, enabling their relaxation and subsequent enhancement in blood flow through the vessel.36 However, multiple mechanisms have been proposed to explain the origin of the NO that reaches the smooth muscle layer. One of these proposed mechanisms suggests that the NO is carried by the ERY as a nitrosothiol and, upon exposure of the ERY to hypoxic conditions, releases NO into the bloodstream.37 However, quantitatively determining such a release from the ERY is made somewhat challenging due to the instability of this reactive molecule. As shown, the values for the NO release are significantly different depending on the timing of the measurement of the analyte (which is dramatically increased using the described device). The device described here represents a step forward toward integrating microfluidic systems with existing high throughput technology such as multipipetters and plate readers. Moreover, this device presents a connection between the extensive work of the lab automation community in developing high throughput handling systems for 96 well plates, and the advantages of cell analysis are presented by utilizing microfluidics and can potentially benefit drug discovery by incorporating more accurate in vitro models with high throughput technology.

Received for review April 29, 2010. Accepted July 16, 2010. AC101130S

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

7497