Article pubs.acs.org/ac
Negative Enrichment of Circulating Tumor Cells Using a Geometrically Activated Surface Interaction Chip Kyung-A Hyun,† Tae Yoon Lee,*,‡ and Hyo-Il Jung*,† †
School of Mechanical Engineering, Yonsei University, 262 Seongsan-no Seodaemun-gu, Seoul, 120-749, Republic of Korea Materials and Components R&D Laboratory, LG Advanced Research Institute, Seoul, 137-724, Republic of Korea
‡
ABSTRACT: Circulating tumor cells (CTCs) have attracted a great deal of attention, as they can be exploited to investigate metastasis. The molecular and cellular characteristics of these cells are little understood because they are rare and difficult to isolate. Many methods of isolation have centered on affinitybased positive enrichment (i.e., capturing target cells and eluting nontarget cells) using epithelial cell adhesion molecule (EpCAM) antibodies. It is known, however, that not all CTCs express the EpCAM antigen because they are heterogeneous by nature. In addition, negative enrichment (i.e., capturing nontarget cells and eluting target cells) has advantages over positive enrichment in isolating CTCs since the former can collect the target cells in an intact form. In this paper, we introduce a geometrically activated surface interaction (GASI) chip with an asymmetric herringbone structure designed to generate enhanced mixing flows, increasing the surface interaction between the nontarget cells and the channel surface. CD45 antibodies were immobilized inside the channel to capture leukocytes and release CTCs to the outlet. Blood samples from breast, lung, and gastric cancer patients were analyzed. The number of isolated CTCs varied from 1 to 51 in 1 mL of blood. Because our device does not require any labeling processes (e.g., EpCAM antibodies), intact and heterogeneous CTCs can be isolated regardless of EpCAM expression.
T
phenotype of the target cell, as there exist many rare cells that have not yet been discovered. Even when the phenotype of the target cell is known, using a specific antibody to capture the target cells may lose a large fraction of the cells, as antigen expression varies depending on physiological conditions. For example, EpCAM expression tends to change during the epithelial−mesenchymal transition (EMT) process in the metastatic cascade.9 A second limitation is that the isolated cells can be modified by antibody attachment, which is not desirable for downstream studies such as cellular and molecular analyses.10,11 In contrast, negative enrichment specifically eliminates the nontarget cells and collects all rare cells regardless of their immune−cytochemical expression level. Thus, the negative enrichment is more suitable for rare cell isolation and further analysis. Negative enrichment is usually carried out with immune−magnetic systems. There have been several attempts to isolate rare circulating cells using immune− magnetic systems.12 While Kang et al. pretreated CTCs with EpCAM-conjugated micromagnetic beads as mentioned above, Chen et al. pretreated leukocytes with CD45-conjugated micromagnetic beads and collected the CTCs near the outer rim by pulling out immune−magnetically activated leukocytes using a disk-based microfluidic platform.11 However, the pretreatment procedure causes cell loss and can cause low isolation yields. For this reason, a novel negative enrichment
he dissemination of circulating tumor cells (CTCs) from primary tumor sites is believed to involve tethering, rolling, and firm adhesion steps before their eventual extravasation, which leads to the formation of secondary tumor sites. The idea that CTCs are one of the major causes of metastasis has driven many researchers to isolate CTCs because the number of CTCs isolated can be useful for cancer diagnosis and prognosis.1,2 The isolation of intact CTCs is challenging due to their extremely low concentrations in blood, as few as one cell per 109 hematologic cells, which hampers any medical and molecular analysis.3 There are various strategies for cell enrichment using physical properties (size, density, electric charge, deformability) or biological properties (surface protein expression, viability, and invasion capacity). Affinity-based cell enrichment methods are often used due to high specificity and separation when cells have different affinities for capture molecules immobilized in the surfaces of the devices.4,5 There are two types of affinity cell enrichment methods (positive and negative enrichment). Positive enrichment isolates the target cells using the interactions between target cell surface antigens and antibodies. This technique can separate rare cells with high purity in suspensions of various types. In the case of CTCs, most positive enrichment is grounded on epithelial cell adhesion molecule (EpCAM) antibodies.6,7 For example, Kang et al. reported a microfluidic−micromagnetic cell separation device that immunomagnetically isolated CTCs with EpCAM antibody-coated magnetic microbeads.8 However, this technique has significant limitations. One of the major limitations is lack of information about the © 2013 American Chemical Society
Received: December 27, 2012 Accepted: March 25, 2013 Published: March 25, 2013 4439
dx.doi.org/10.1021/ac3037766 | Anal. Chem. 2013, 85, 4439−4445
Analytical Chemistry
Article
Figure 1. (a) Pictorial presentation of negative enrichment (side view). When the blood sample enters the inlet, the nontarget cells (i.e., leukocytes) are specifically captured on the CD45 antibody-coated channels, while the target cells (i.e., CTCs) are eluted through the outlet. The ridges on the top of the main straight channel create a helical flow that creates considerable interaction between the cells and the antibody-coated channel surface. (b) A photograph of the fabricated GASI chip. The GASI chip consists of an inlet, eight channels, and an outlet. Each channel is 2100 μm in width and 50 mm in length. To prevent breakdown of the PDMS channel due to their flexibility, a wall 400 μm in width bears the pressure. (c) The herringbone structure and key parameters used in this study.
degassing under a vacuum chamber, the mixture was cured on a hot plate at 95 °C for 50 min. The cured PDMS replica was removed from the master template, perforated at the channel inlet and outlet using a punch, and bonded to slide glass using oxygen plasma to form the microfluidic channels. Immobilization of Capture Molecules in the Channel. CD45 antibodies were used to capture the leukocytes inside the channel and consequently collect the CTCs in the outlet. To coat the channel with the CD45 antibody, an affinity surface was prepared using established protocols with slight modifications. First, for degassing of the channels, they were filled with 70% ethanol because the low-polarity solvent wetted the hydrophobic PDMS surface more easily. Then, the channels were rinsed with deionized water. A solution of 1 mg/mL biotin−bovine serum albumin (BSA) in 10 mM Tris−HCl buffer with 50 mM NaCl was introduced into the channel manually using a syringe, followed by incubation for 45 min and then rinsing with the 10 mM Tris−HCl buffer. To form the avidin−biotin complex, a solution of 0.2 mg/mL neutravidin in 10 mM Tris buffer was loaded into the channel and incubated for 15 min, followed by rinsing with Tris buffer and deionized water in steps. Finally, the biotinylated CD45 antibody was added to complete the affinity surface conjugation and incubated for 30 min, followed by two rounds of rinsing with phosphate-buffered saline (PBS) containing 1% of BSA to prevent nonspecific binding. All incubation of reagents was performed at room temperature. Sample Preparation. To evaluate the device’s enrichment efficiency, we used MCF-7 cells and MDA-MB-231 (a human breast adenocarcinoma cell line) and Jurkat cells (a human T cell lymphoblast like cell line). Cells were incubated under ordinary cell culture conditions and subsequently resuspended
strategy that does not require pretreatment is necessary for proper performance and further analysis of the CTCs. In this study, we developed a geometrically activated surface interaction (GASI) chip using a herringbone shape to efficiently capture a large number of hematological cells rather than CTCs (Figure 1a). The herringbone structure was patterned on the channel to produce transverse flow that facilitates effective contact between antigens on the cells and antibodies on the channel surface (Figure 1, parts b and c).13,14 This was the first attempt to enrich CTCs from metastatic cancer patients using a negative enrichment microfluidic chip. The GASI chip successfully enriched CTCs and will be a powerful tool for isolation of other types of circulating rare cells whose phenotypes are poorly understood.
■
MATERIALS AND METHODS Device Fabrication. Conventional two-layer soft-lithography was used to pattern the herringbone microchannels. First, a negative photoresist (SU-8 3025, MicroChem, Newton, MA, U.S.A.) was deposited onto silicon wafers to form the main channel and was subsequently subjected to several physicochemical processes: soft baking, UV light exposure, and postexposure baking. Then, before developing the photoresist, a negative photoresist was added on top of that to pattern the herringbone structure. The same physiochemical process was repeated to develop the main channel and herringbone channel at the same time. As a result, the two-layer herringbone structure was formed and was used as a master template for replica molding. A poly(dimethylsiloxane) (PDMS) prepolymer and a curing agent were mixed at 10:1 weight ratio (Sylgard 184, Dow Corning Corporation, MI, U.S.A.). Next, the mixture was poured onto the master template. After 4440
dx.doi.org/10.1021/ac3037766 | Anal. Chem. 2013, 85, 4439−4445
Analytical Chemistry
Article
Figure 2. Determination of the optimal conditions. Jurkat cells (1 × 106 cells/mL) were injected into the conventional HB chip, and the cells collected were counted using a hemocytometer. (a) Flow rate effects on capture efficiency. The capture efficiency decreased at high flow rates due to the high shear force. (b) Determination of the minimal concentration of anti-CD45 antibodies necessary for high capture efficiency. The capture efficiency shows a plateau starting at 10 μg/mL CD45 antibody. (c) Initial cell concentration effects on capture efficiency. The capture efficiency decreased as the initial concentration of cells increased because the attached cells seemed to be partially detached by interacting with nonadhered cells.
herringbone between half cycles was 1/3w and 2/3w; the number of half cycle grooves was 10; the height of the main channel was h = 50 μm; the height of the herringbone was hh = 45 μm; the groove width was wg = 50 μm; the groove pitch was wp = 100 μm (Figure 1c). Using the same dimensions of conventional HB chips, we determined the experimental conditions for negative enrichment by capturing the Jurkat cells. The inner surface of the channel was coated with antiCD45 antibody. Like peripheral white blood cells, Jurkat cells normally express CD45 at high levels, and an individual cell expresses multiple isoforms simultaneously.15 A high flow rate is required to treat a large volume of body fluid in a short period because CTCs are rare, as mentioned in the introduction. On the other hand, at high flow rates, cells interact little with the antibody on the channel surface and tend to be detached from the channel surface due to the high shear force. We loaded the Jurkat cells at a fixed concentration of 1 × 106/mL into the inlet and then counted the number of cells that came out through the outlet using a hemocytometer in order to estimate the capture efficiency, which was best [92.71% ± 1.59% (n = 3)] at a 10 μL/min inlet flow rate. As the flow rate was increased, the capture efficiency decreased from 84.10% ± 6.06% to 67.96% ± 4.20% (n = 3) (Figure 2a). However, we observed that the collected cell membranes were broken down at a 10 μL/min inlet flow rate. When the flow rate increased to 20 μL/min in the inlet, the cell membrane was observed to be intact. Because many CTCs rapidly undergo apoptosis, the cells have to be isolated as quickly as possible.16 So, we determined the optimal flow rate as 20 μL/min. To handle the large volume of blood, negative enrichment may not be cost-effective, so it is important to find the smallest concentration of anti-CD45 antibodies to achieve maximum efficiency. Figure 2b shows the Jurkat cell capture efficiency at various concentrations of anti-CD45 antibodies. The Jurkat cells are injected at a fixed concentration of 1 × 106/mL and a 20 μL/min flow rate into the inlet. At 10 μg/mL, the capture efficiency reached the saturation point [from 84.89% ± 1.63% to 87.81% ± 4.36% (n = 3)]. Following the optimal experimental conditions (flow rate, 20 μL/min; concentration of anti-CD45 antibodies, 10 μg/mL), we evaluated the capture efficiency of the HB chip at various concentrations of cells in order to judge whether the HB chip can handle the large amounts of cells in clinical trials. As predicted, the capture
at a specific concentration in 1 mL of PBS containing 1% BSA to reduce the cell aggregation. Blood samples of metastatic breast, lung, and gastric cancer patients were used for CTC analysis. Patients with cancer were recruited according to a separate protocol approved by the IRB. Human blood samples (1 mL) were collected in vacutainer tubes containing EDTA and were processed within 6 h. The erythrocytes were removed by hemolysis. Hemolysis buffer (Qiagen, Chatsworth, CA) was added to whole blood in a 1:10 v/v ratio and incubated for 10 min at room temperature. Following centrifugation at 600g for 10 min, the supernatant was removed. For sufficient hemolysis, 5 mL of hemolysis buffer was added to the pellet and incubated for 5 min at 4 °C. Then it was centrifuged at 1750g for 2 min, and the pellet was resuspended in 1 mL of PBS containing 1% BSA. Experimental Setup. A syringe pump (KDS210, KDS Scientific Incorporated, MA, U.S.A.) was used to inject the prepared sample continuously into the microchannel. The flow rate was from 10 to 40 μL/min.To observe the enrichment efficiency, after the enrichment process, the collected cells from the outlet were spin-coated on a slide glass at 1500 rpm for 5 min. Then, to identify the CTCs, the cells on the slide glass were immunostained with DAPI for DNA content, CD45 for leukocytes, and cytokeratin for CTCs. An image of the entire area was captured using an autostage-coupled fluorescence microscope. Using the CTC identification criteria (DAPI positive/CD45 negative/cytokeratin positive), the captured images were examined. Cell size and shape were also considered for the identification of CTCs.
■
RESULTS AND DISCUSSION Determination of Experimental Conditions for Negative Enrichment. The herringbone design has advantages for cell affinity chromatography because it produces a microvortex that improves the surface interactions between the cells and the antibody-coated channel at a low Reynolds number and low shear force.13 A herringbone (HB) chip designed for positive enrichment of CTCs (i.e., the target CTCs are captured on the chip surface) was previously reported by Stott et al.14 The dimensions of the herringbone structure were determined to produce maximum transverse flow. The angle between the herringbone and the axis of the channel was θ = 45°; the width of the herringbone was w = 300 μm; the orientation of the 4441
dx.doi.org/10.1021/ac3037766 | Anal. Chem. 2013, 85, 4439−4445
Analytical Chemistry
Article
Figure 3. Microscopic images of Jurkat cells captured on the (a) HB chip, (b) enhanced HB chip, and (c) GASI chip. According to the evolution of the microfluidic negative enrichment chip from the conventional HB chip to the GASI chip, the captured cells cover a broader area of the channel, even though they were still focused to the apex of the herringbone structure. (d) Jurkat cell capture efficiency in the HB chip, the enhanced HB chip, and the GASI chip at high initial cell concentrations. At 1 × 106 and 1 × 107/mL, the GASI chip captured 98.94% and 97.64% of cells, respectively, compared with the HB chip and the enhanced HB chip which captured 87.96%, 87.03% and 98.36%, 94.63% of cells, respectively, in the same experiment conditions. (e) Enrichment yields of the HB chip, the enhanced HB chip, and the GASI chip.
Jurkat cells was high, similar to real clinical samples. The new dimensions of the herringbone structure were as follow: h = 50 μm; hh = 50 μm; wg = 125 μm; wp = 200 μm (Figure 1c). The capture efficiency increased up to 98.36%. Although the captured cells are dispersed over a wider area than that of the original HB chip, most of the cells are still observed at each apex of the herringbone (Figure 3, parts a and b). In the herringbone channel, the behavior of the cells is determined by a unique combination of the drag force due to fluid flow, gravity, and buoyancy. When the magnitude of the zcomponent of the drag force is smaller than the net force of buoyancy and gravity, the cells are focused to the apex of the herringbone in the interfaces of the adjacent microvortices by the lateral flow.19,20 Since most mammalian cells are denser than the cell buffer solutions (density of a mononuclear cell, 1.077 g/cm3; density of PBS, 1.00 g/cm3), the cells are focused to the apex and captured. Therefore, we successfully designed a geometrically activated surface interaction chip. The basic dimensions of the GASI chip was the same as the “enhanced HB chip”, but the apex of the herringbone was increased by two (Figure 3c). The length of one ridge in herringbone structure changed from w = 300 μm in the enhanced HB chip to w = 150 μm in the GASI chip (Figure 1c). Because the length between one apex and the next is relatively shorter than in the enhanced HB chip, the cells were captured evenly, though they were focused on the apexes (Figure 3c). When the GASI chip was compared with other HB chips, the GASI chip significantly improved the capture efficiency to
efficiency tended to decrease at higher concentrations of cells because the adhered cells partially interact with cells traveling in the flow stream (Figure 2c).17 Creation of a Geometrically Activated Surface Interaction Chip to Maximize Capture Efficiency. When handling clinical samples using negative enrichment, it is important to capture as many nontarget cells as possible inside the microchannels. As shown in Figure 2c, the capture efficiency of nontarget cells was more than 92% at a cell population of 104−105/mL, whereas it was less than 88% at a cell population of 106−107. Since the real number of leukocytes ranges from 106 to 107, the HB chip is thought to be barely suitable for negative enrichment. The use of herringbone mixers to enhance particle−surface interactions in microfluidic devices is growing in a range of biological and clinical applications. Forbes and Kralj simulated the flow within a herringbone mixer to understand (1) where particles are most likely to interact with surfaces and (2) what physical parameters have the greatest impact on overall particle−surface interaction.18 These simulations are very important to determine how to modify the distribution of contact points toward specific target surfaces. Their computational model and theoretical framework provide guidelines for the geometrical optimization of surface interactions in the herringbone, as well as the location of those contacts. Using their computationally optimized dimensions for enhanced surface interaction, we fabricated an “enhanced HB chip” and evaluated the capture efficiency when the concentration of 4442
dx.doi.org/10.1021/ac3037766 | Anal. Chem. 2013, 85, 4439−4445
Analytical Chemistry
Article
Moreover, we evaluated the separation yield of the GASI chip using the MDA-MB-231 cells in order to clarify that the GASI chip collects all kinds of CTCs regardless of their EpCAM expression level. MDA-MB-231 cells have similar physical properties such as size and shape in comparison with MCF-7 cells. However, they have less EpCAM surface densities (1.7 × 103 EpCAM binding sites) than MCF-7 cells (2.22 × 105 binding sites).21 We successfully separated 90.67% of the MCF7 cells and 87.43% of MDA-MB-231 cells by an individual experiment with 20 μL/min. (Data not shown.) Isolation of Circulating Tumor Cells Using the GASI Chip. For CTC isolation, the GASI chip was tested on blood samples from patients with several kinds of metastatic cancer. Blood samples (1 mL) were collected from breast, lung, and gastric cancer patients. After the samples passed through the GASI chip, the collected cells were fixed on a slide glass and stained with DAPI for DNA content, PE-conjugated cytokeratins for CTCs, and FITC-conjugated CD45 for leukocytes. After imaging the slide glass, the captured images were examined carefully according to predefined criteria (DAPI positive, cytokeratin positive, CD45 negative). Because the negative enrichment system does not use any characteristics of target cells such as size or molecular surface markers but specifically removes nontarget cells, it can separate CTCs with a variety of features. (Figure 5) There have been reports that the level of EpCAM expression on a cell surface decreases in some types of tumor cells.9 Therefore, the enrichment efficiency of CTCs using positive enrichment that uses anti-EpCAM affinity molecules depends on the level of EpCAM expression, and this approach may lose considerable amounts of EpCAM-negative expressed CTCs. This can be a very crucial limitation in CTC analysis, as cancer cells are very heterogeneous. The number of CTCs isolated from individual patients is summarized in Table 1. When the 12 blood samples from several types of metastatic cancer patients were subjected to the GASI chip, at least one CTC was identified, regardless of the primary tumor. Generally, fluorescent CD45 markers are used as a positive control of leukocytes to exclude them from CTCs. However, several researchers have reported that a significant number of cells
98.94% (Figure 3d) and enrichment yield to 130.94-fold (Figure 3e). Although the number of cancer cells (MCF-7) collected from the GASI chip was not proportional to the number of cancer cells spiked into leukocyte (Jurkat) samples across the physiologically relevant range (50−1000 cancer cells/106 leukocytes) because of nonspecific binding, the purity of the spiked cancer cells increased remarkably to 3.64% in the GASI chip, compared with 0.94% in the HB chip (Figure 4). Therefore, the GASI chip is more suitable for negative enrichment of CTCs.
Figure 4. Separation yields of MCF-7 cells from Jurkat cells depending on the number of cancer cells. For the conventional HB chip, the purity of the MCF-7 cells was less than 0.65% due to insufficient surface interaction between the Jurkat cells and the CD45 antibody modified channel surface. For the GASI chip, although the recovery of the MCF-7 cells (from 94.67% to 42.10%) decreased more than with the HB chip (from 94% to 64.85%) due to nonspecific binding as the number of MCF-7 cells increased, the purity of the MCF-7 cells increased from 0.65% to 3.64%.
Figure 5. Fluorescent images of cells isolated using the GASI chip from metastatic cancer patients. CTCs were identified by predefined criteria (DAPI positive, CD45 negative, cytokeratins positive) For gastric cancer, all cytokeratins-positive cells appear dual-positive with CD45. 4443
dx.doi.org/10.1021/ac3037766 | Anal. Chem. 2013, 85, 4439−4445
Analytical Chemistry
Article
heterogeneous circulating rare cells continuously collected by the device is expected to give researchers many opportunities to investigate the molecular nature of rare cells.
Table 1. Number of CTCs in the Blood of Metastatic Cancer Patients Counted Using Immunofluorescence Staininga breast cancer patients P7 P8 P11 P12 a
CTC no. 3 51 6 3
(1)* (2)* (4)* (1)*
lung cancer patients P5 P6 P9 P10
CTC no. 1 2 3 7
(1)* (2)* (1)* (4)*
gastric cancer patients P2 P3 P4 P13
■
CTC no. 2 1 1 5
(2)* (1)* (1)* (5)*
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected] (H.-I.J.);
[email protected] (T.Y.L.). Phone: +82-(0)2-2123-5814 (H.-I.J.); +82-(0)2-5264950 (T.Y.L.). Fax: +82-(0)2-312-2159.
The number of dual-positive cells (CD45 and cytokeratins positive)*.
Notes
The authors declare no competing financial interest.
■
appear dual-positive for CD45 and cytokeratins.9 Although the identity of these dual-positive cells is not well-understood, we cautiously hypothesize that they are due to the heterogeneity of CTCs. These cells were included in the CTC enumeration and are described as the ratio of dual-positive cells (in parentheses of Table 1). Interestingly, all of the isolated CTCs from the blood of gastric cancer patients appeared dual-positive, while isolated CTCs from the blood of breast and lung cancer patients appeared in various distribution from 3.92% to 100% (Table 1). The efficiency for anti-CD45 to capture the cells (leukocytes or any other circulating cells expressing the CD45) can never be 100%. The capture efficiency in our experiments means a kind of probability that the CD45 antibodies inside the microfluidic channel can hold as many CD45-expressing cells as possible, but not all of them. Therefore, dual-positive CTCs could be escaped from the channel and observed in the collection tube. The molecular characterization of the dualpositive CTCs will be carried out for the further work.
ACKNOWLEDGMENTS This study was supported in part by a research program of the National Research Foundation of Korea (NRF) (Grant No. 2011-0016731) and a Grant from the National R&D Program for Cancer Control (Grant No. 1120290) as well as the Korea Health Technology R&D Project (Grant No. A121986), Ministry of Health and Welfare, Republic of Korea.
■
REFERENCES
(1) Kim, S. I.; Jung, H. I. J. Breast Cancer 2010, 13, 125−131. (2) Geng, Y.; Marshall, J. R.; King, M. R. Ann. Biomed. Eng. 2012, 40, 790−805. (3) Dharmasiri, U.; Witek, M. A.; Adams, A. A.; Soper, S. A. Annu. Rev. Anal. Chem. 2010, 3, 409−431. (4) Li, P.; Gao, Y.; Pappas, D. Anal. Chem. 2012, 84, 8140−8148. (5) Alix-Panabieres, C.; Schwarzenbach, H.; Pantel, K. Annu. Rev. Med. 2012, 63, 199−215. (6) Nagrath, S.; Sequist, L. V.; Maheswaran, S.; Bell, D. W.; Irimia, D.; Ulkus, L.; Smith, R.; Kwak, E. L.; Digumarthy, S.; Muzikansky, A.; Ryan, P.; Balis, U. J.; Tompkins, R. G.; Haber, D. A.; Toner, M. Nature 2007, 450, 1235−1239. (7) Wang, S.; Liu, K.; Liu, J.; Yu, Z. T.-F.; Xu, X.; Zhao, L.; Lee, T.; Lee, E. K.; Reiss, J.; Lee, Y. K.; Chung, L. W. K.; Huang, J.; Rettig, M.; Seligson, D.; Duraiswamy, K. N.; Shen, C. K.-F.; Tseng, H.-R. Angew. Chem., Int. Ed. 2011, 50, 1−6. (8) Kang, J. H.; Krause, S.; Tobin, H.; Mammoto, A.; Kanapathipillai, M.; Ingber, D. E. Lab Chip 2012, 12, 2175−2181. (9) Gorges, T. M.; Tinhofer, I.; Drosch, M.; Röse, L.; Zollner, T. M.; Krahn, T.; von Ahsen, O. BMC Cancer 2012, 12, 178−190. (10) Lara, O.; Tong, X.; Zborowski, M.; Chalmers, J. J. Exp. Hematol. 2004, 32, 891−904. (11) Chen, C. L.; Chen, K. C.; Pan, Y. C.; Lee, T. P.; Hsiung, L. C.; Lin, C. M.; Chen, C. Y.; Lin, C. H.; Chiang, B. L.; Wo, A. M. Lab Chip 2011, 11, 474−483. (12) Zborowski, M.; Chalmers, J. J. Anal. Chem. 2011, 83, 8050− 8056. (13) Stroock, A. D.; Dertinger, S. K. W.; Ajdari, A.; Mezic, I.; Stone, H. A.; Whitesides, G. M. Science 2002, 295, 647−651. (14) Stott, S. L.; Hsu, C. H.; Tsukrov, D. I.; Yu, M.; Miyamoto, D. T.; Waltman, B. A.; Rothenberg, M. S.; Shah, A. M.; Smas, M. E.; Korir, G. K.; Floyd, F. P., Jr.; Gilman, A. J.; Lord, J. B.; Winokur, D.; Springer, S.; Irimia, D.; Nagrath, S.; Sequistd, L. V.; Lee, R. J.; Isselbacher, K. J.; Maheswaran, S.; Haber, D. A.; Toner, M. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 18392−18397. (15) McKenney, D. W.; Onodera, H.; Gorman, L.; Mimura, T.; Rothstein, D. M. J. Biol. Chem. 1995, 270, 24949−24954. (16) Pantel, K.; Alix-Panabières, C.; Riethdorf, S. Nat. Rev. Clin. Oncol. 2009, 6, 339−351. (17) Sin, A.; Murthy, S. K.; Revzin, A.; Tompkins, R. G.; Toner, M. Biotechnol. Bioeng. 2005, 91, 816−826. (18) Forbes, T. P.; Kralj, J. G. Lab Chip 2012, 12, 2634−2637. (19) Williams, M. S.; Longmuir, K. J.; Yager, P. Lab Chip 2008, 8, 1121−1129.
■
CONCLUSION Circulating tumor cells have great potential for fundamental studies of cancer and clinical applications. Although isolation methods for circulating tumor cells have been extensively explored, the isolation and characterization of CTCs is still difficult because they are extremely rare. Most approaches are grounded in EpCAM antibody-based detection. For example, CellSearch (Veridex, U.S.A.) immunomagnetically captures CTCs from 7.5 mL of blood using ferrofluidic nanoparticles conjugated to a monoclonal antibody against EpCAM.22 The microchip detects the CTCs through immunocapture on microposts or herringbone mixing channels coated with the same antibody.23 Because EpCAM expression on a cell surface is down-regulated in some types of tumor cells, detection techniques that use the anti-EpCAM affinity molecule may lose considerable amounts of EpCAM-negative CTCs. This can be a very crucial limitation in CTC analysis, as cancer cells are very heterogeneous. Therefore, a new method in which CTCs are purely isolated without any antibody tagging is necessary. We designed and fabricated a geometrically activated surface interaction chip in which almost all of the blood cells stuck on the antibody-conjugated surface and the free CTCs escaped. The GASI chip generates enhanced mixing flows increasing the surface interaction between nontarget cells and the channel surface. Since many methods for CTC isolation are affinitybased (utilizing the EpCAM capturing capacity), almost all isolation has been limited to the EpCAM expression levels of CTCs. Our approach is novel because the GASI chip enables the isolation of various types of circulating rare cells, such as circulating endothelial cells (CECs), cancer stem cells (CSCs), circulating progenitor cells (CPCs), and CTCs, including nucleated red blood cells (nRBC). The number of intact and 4444
dx.doi.org/10.1021/ac3037766 | Anal. Chem. 2013, 85, 4439−4445
Analytical Chemistry
Article
(20) Hsu, C. H.; Carlo, D. D.; Chen, C.; Irimia, D.; Toner, M. Lab Chip 2008, 8, 2128−2134. (21) Prang, N.; Preithner, S.; Brischwein, K.; G̈ oster, P.; Ẅ oppel, A.; M̈ uller, J.; Steiger, C.; Peters, M.; Baeuerle, P. A.; DaSilva, A. J. Br. J. Cancer 2005, 92, 342−349. (22) den Toonder, J. Lab Chip 2011, 11, 375. (23) Yu, M.; Stott, S.; Toner, M.; Maheswaran, S.; Haber, D. A. J. Cell. Biol. 2011, 192, 373−382.
4445
dx.doi.org/10.1021/ac3037766 | Anal. Chem. 2013, 85, 4439−4445
