Anal. Chem. 2009, 81, 3194–3198
Technical Notes Simultaneous Separation, Metering, and Dilution of Plasma from Human Whole Blood in a Microfluidic System Tomoya Tachi,† Noritada Kaji,†,‡ Manabu Tokeshi,*,†,‡ and Yoshinobu Baba†,‡,§,|,⊥ Department of Applied Chemistry, Graduate School of Engineering, MEXT Innovative Research Center for Preventive Medical Engineering, Plasma Nanotechnology Research Center, Nagoya University, Nagoya 464-8603, Japan, Health Technology Research Center, National Institute of Advanced Industrial Science and Technology (AIST), Takamatsu 761-0395, Japan, and Institute for Molecular Science, National Institutes of Natural Sciences, Okazaki 444-8585, Japan In about a 3 min period, we have simultaneously separated plasma from human whole blood and metered and diluted the plasma using a microchip with an interchannel microstructure. The plasma separation was based on both cross-flow filtration and sedimentation of red blood cells in the microchannels. Metering and diluting operations of the plasma were based on volume control of liquid in the microchannels by syringe pumps. On this microchip, we produced plasma diluted by a factor of 6 from whole blood containing theophylline and we observed very little hemolysis. It is possible to separate plasma from one or just several drops of whole blood by using this microchip. Sample preparation is necessary for analysis of biological complex samples in current clinical laboratories. In many analyses of markers in blood, either plasma or serum is prepared by centrifugation or sedimentation because blood corpuscle cells and hemoglobin interfere with analysis based on optical measurement techniques. Only limited microchip-based measurements are feasible using whole blood;1-5 most such measurements require the separation of plasma from whole blood, such as a single drop * To whom correspondence should be addressed. E-mail: tokeshi@ apchem.nagoya-u.ac.jp. Fax: +81-52-789-4498. † Department of Applied Chemistry, Graduate School of Engineering, Nagoya University. ‡ MEXT Innovative Research Center for Preventive Medical Engineering, Nagoya University. § Plasma Nanotechnology Research Center, Nagoya University. | National Institute of Advanced Industrial Science and Technology (AIST). ⊥ Institute for Molecular Science, National Institutes of Natural Sciences, Okazaki. (1) Toner, M.; Irimia, D. Annu. Rev. Biomed. Eng. 2005, 7, 77–103. (2) Vrouwe, E. X.; Luttge, R.; Olthuis, W.; van den Berg, A. Electrophoresis 2005, 26, 3032–3042. (3) Vrouwe, E. X.; Luttge, R.; van den Berg, A. Electrophoresis 2004, 25, 1660– 1667. (4) Vrouwe, E. X.; Luttge, R.; Vermes, I.; van den Berg, A. Clin. Chem. 2007, 53, 117–123. (5) 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.
3194
Analytical Chemistry, Vol. 81, No. 8, April 15, 2009
of whole blood obtained from a finger stick. In current point-ofcare devices such as blood glucose analysis kits, plasma is isolated from whole blood using glass fiber filters or microporous membranes. These filtration systems are widely used in test-strip assays but would not be easily integrated and combined with microfluidic systems which could be applied to complex assay systems.6 Many papers on plasma separation from whole blood using microchips have been reported recently.7-18 Plasma has been isolated from whole blood on a microchip by various methods, such as external forces like centrifugation,7 capillary force8 and acoustic standing wave force,9 filtrations by membranes10 and microbeads,11 and separations based on specially designed microchannel structures.8,12,13,16,18 For instance, serum was separated from whole blood using a CD-type microchip for centrifugation.7,8 One of these microchips is based on accelerating sedimentation of red blood cells by centrifugation,7 and the other is based on extracting plasma by capillary force and then metering the plasma by centrifugation.8 Plasma separation from whole blood is mostly performed using microfluidic devices. Partial separation (6) Yager, P.; Edwards, T.; Fu, E.; Helton, K.; Nelson, K.; Tam, M. R.; Weigl, B. H. Nature 2006, 442, 412–418. (7) Grumann, M.; Steigert, J.; Riegger, L.; Moser, I.; Enderle, B.; Riebeseel, K.; Urban, G.; Zengerle, R.; Ducre´e, J. Biomed. Microdevices 2006, 8, 209– 214. (8) Pugia, M. J.; Blankenstein, G.; Peters, R.-P.; Profitt, J. A.; Kadel, K.; Willms, T.; Sommer, R.; Kuo, H. H.; Schulman, L. S. Clin. Chem. 2005, 51, 1923– 1932. (9) Petersson, F.; Nilsson, A.; Holm, C.; Jo ¨nsson, H.; Laurell, T. Lab Chip 2005, 5, 20–22. (10) Thorslund, S.; Klett, O.; Nikolajeff, F.; Markides, K.; Bergquist, J. Biomed. Microdevices 2006, 8, 73–79. (11) Moorthy, J.; Beebe, D. J. Lab Chip 2003, 3, 62–66. (12) Crowley, T. A.; Pizziconi, V. Lab Chip 2005, 5, 922–929. (13) VanDelinder, V.; Groisman, A. Anal. Chem. 2006, 78, 3765–3771. (14) VanDelinder, V.; Groisman, A. Anal. Chem. 2007, 79, 2023–2030. (15) Wilding, P.; Kricka, L. J.; Cheng, J.; Hvichia, G.; Shoffner, M. A.; Fortina, P. Anal. Biochem. 1998, 257, 95–100. (16) Yang, X.; Hibara, A.; Sato, K.; Tokeshi, M.; Morishima, K.; Kikutani, Y.; Kimura, H.; Kitamori, T. Proc. µTAS 2004, Malmo ¨, Sweden, September 26–30, 2004; pp 120-122. (17) Yang, S.; Undar, A.; Zahn, J. D. Lab Chip 2006, 6, 871–880. ¨ ndar, A.; Zahn, J. D. ASAIO J. 2006, 52, 698–704. (18) Yang, S.; Ji, B.; U 10.1021/ac802434z CCC: $40.75 2009 American Chemical Society Published on Web 03/13/2009
Figure 1. (a) Schematic illustration of the upper and lower plates of the microchip and (b) images of the microchannels of the upper and lower plates.
of plasma from whole blood was demonstrated in laminar flow and in an acoustic standing wave in a microchannel.9 The wave had a node in the middle of the microchannel, leading to enrichment of red blood cells in the middle and depletion in the periphery of the microchannel. Membrane filters were combined with a microfluidic device for plasma separation from 10-30% whole blood solution.10 In another experiment, plasma separation was done by porous filtration using polyHEMA (2-hydroxyethyl methacrylate) microbeads and rabbit whole blood was diluted by a factor of 20 as a sample.11 In both membrane and porous filtrations, whole blood must be diluted to prevent rapid decrease in efficiency of filtration as a result of clogging with blood cells at the dead-end filter. Some papers have reported isolation of plasma from whole blood using a specially designed structure within a microchip. For example, plasma was separated from whole bovine blood in a cross-flow driven by capillary forces by using a planar microfilter structure.12 In this experiment, the amount of plasma collected with a microchip was limited to about 50 nL. Isolation of plasma from human whole blood was conducted in a continuous cross-flow on a molded microfluidic device by pressure.13 As well, separation and analyses of plasma from whole blood was performed in a cross-flow on a microchip with an interchannel microstructure.16 Diluted plasma was separated from whole blood and R-fetoprotein in the plasma was detected, but the plasma was not quantitatively diluted because the dilution rate depends on the viscosity of whole blood. The error of quantification was possibly caused by the different viscosities of the whole blood samples in the method. When loaded with blood diluted to 20% hematocrit to prevent clogging of the microchannel with blood cells, plasma was extracted, which was about 8% of the blood volume. In contrast, plasma separation was done in a microfluidic device using the Zweifach-Fung effect.17,18 The plasma selectivity was almost 100%, and the plasma separation volume percentage was near 20%. For microchip-based quantitative analysis using whole blood, it is important to meter and dilute plasma obtained from whole blood on a microchip so that the analysis can be carried out as a continuous process on another or the same microchip, but there
have been only a few reports on these operations. Pugia et al.8 demonstrated metering and diluting plasma after separation of plasma by centrifugation from whole blood using a CD-type microchip. Metering was achieved by use of a fixed volume area with micrometer-sized capillaries at the beginning and end of the metered area. Precise dilution was carried out by mixing of metered plasma and metered buffer. However, applications of this approach are limited to centrifugation by CD-type microchips and it is difficult to connect the other types of microfluidic devices. In this research, we designed and operated a microchip with an interchannel microstructure for the separation of plasma from human whole blood and for metering and diluting the plasma. The plasma separation was based on both cross-flow filtration and sedimentation of red blood cells in the microchannels. Metering and diluting the plasma was based on volume control of liquid in the microchannels by syringe pumps. Simultaneous operations to separate, meter, and dilute plasma obtained from whole blood were done in microchannels using this microchip, and plasma precisely diluted by a factor of 6 was generated without hemolysis in the microchannels in about 3 min. This microchip is easily connected with other microchannels and microfluidic devices for quantitative analyses. EXPERIMENTAL SECTION Microchip Design. The microchip used in this work was fabricated on Pyrex glass substrate using standard photolithographic and wet chemical etching techniques. A schematic illustration is shown in Figure 1. The microchip was specially designed for this work and was manufactured for us by the Institute of Microchemical Technology (IMT), Co., Ltd. (Kanagawa, Japan). The microchip was 30 mm long and 70 mm wide. Its microchannels consisted of two 70 mm long main channels linked to the inlets and outlets and shallow channels connecting the main channels. The main channels were about 310 µm wide and 155 µm deep. Each main channel consisted of 10 mm long curved regions near the inlet and outlet and a 50 mm long linear part at the middle. The linear parts of the two main channels were parallel and separated by a 30 µm distance, and they were Analytical Chemistry, Vol. 81, No. 8, April 15, 2009
3195
Figure 2. Principle of serum separation from whole blood in microchannels.
connected with the shallow channels laterally every 90 µm. These linear parts consisted of 20 mm long lengths without shallow channels near the inlets, 40 mm long lengths with shallow channels, and 10 mm lengths without shallow channels near the outlets. Shallow channels were 12 µm wide and 1 µm deep. These microstructures were given by combining upper and lower plates designed like Figure 1. Blood is a suspension of red blood cells, white blood cells, and platelets in plasma. Red blood cells are discoid, anuclear cells about 8 µm in diameter and 2.5 µm in thickness, white blood cells are spherical cells 8-12 µm in diameter, and platelets are discoid particles 1-3 µm in diameter.13 A schematic illustration of the principle of plasma separation by the microchip is shown in Figure 2. Separation of plasma from whole blood was based on both crossflow filtration and sedimentation of red blood cells by gravity in the microchannels. The main channels were readily passable for all types of cells. Cross-flow was formed in the shallow channels perpendicular to the main channels, and blood cells were not permeable through the shallow channels because they were only 1 µm deep. The sedimentation rate of red blood cells depends on several factors of whole blood, which is 1-10 µm/s for the nonturbulent condition.19 When flow is laminar in a microchannel, it is assumed that the sedimentation rate would be about 3 µm/s. Under this condition, it would take about 50 s for red blood cells to drop to the bottom of the main microchannels. The linear flow rate of whole blood would be about 0.4 mm/s for red blood cells to drop to the bottom within the 20 mm distance near the inlet, where there were no shallow channels coming from the main channels. Assuming that the cross section of the main channels was a rectangle, 0.4 mm/s corresponded to about 1.1 µL/min. In the main channels, the flow was predicted to be laminar because the Reynolds number was 0.28 (
