- Email: [email protected]
Selective isolation of magnetic nanoparticle-mediated heterogeneity subpopulation of circulating tumor cells using magnetic gradient based microfluidic system
Biosensors and Bioelectronics 88 (2017) 153–158
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Selective isolation of magnetic nanoparticle-mediated heterogeneity subpopulation of circulating tumor cells using magnetic gradient based microfluidic system Bongseop Kwak a,n,1, Jaehun Lee a,b,1, Dongkyu Lee a, Kangho Lee a, Ohwon Kwon a, Shinwon Kang b, Youngwoo Kim a a Korea Institute of Machinery and Materials, Daegu Research Center for Medical Devices and Rehab. Engineering, Department of Medical Device, 330 Techno Sunhwan-ro, Yuga-myeon, Dalsung-gun, Daegu, 42994 Republic of Korea b Kyungpook National University, College of IT Engineering, School of Electronics Engineering, 80 Daehak-ro, Buk-gu, Daegu, 41566 Republic of Korea
art ic l e i nf o
a b s t r a c t
Article history: Received 13 June 2016 Received in revised form 29 July 2016 Accepted 1 August 2016 Available online 2 August 2016
Relocation mechanisms of the circulating tumor cells (CTCs) from the primary site to the secondary site through the blood vessel network cause tumor metastasis. Despite of the importance to diagnose the cancer metastasis by CTCs, still it is formidable challenge to use in the clinical purpose because of the rarity and the heterogeneity of CTCs in the cancer patient’s peripheral blood sample. In this study we have developed magnetic force gradient based microfluidic chip (Mag-Gradient Chip) for isolating the total number of CTCs in the sample and characterizing the state of CTCs simultaneously with respect to the epithelial cell adhesion molecule (EpCAM) expression level. We have synthesized magnetic nanoparticles (MNPs) using hydrothermal method and functionalized anti-EpCAM on their surface for the specific binding with CTCs. The Mag-Gradient Chip designed to isolate and classify the CTCs by isolating at the different location in the chip using magnetic force differences depending on the EpCAM expression level. We observed 95.7% of EpCAM positive and 79.3% of EpCAM negative CTCs isolated in the MagGradient Chip. At the same time, the 71.3% of isolated EpCAM positive CTCs were isolated at the first half area whereas the 76.9% of EpCAM negative CTCs were collected at the latter half area. The Mag-Gradient Chip can isolate the 3 ml of heterogeneous CTCs sample in 1 h with high isolating yield. The EpCAM expression level dose not means essential condition of the metastatic CTCs, but the Mag-Gradient Chip can shorten the date to diagnose the cancer metastasis in clinic. & 2016 Elsevier B.V. All rights reserved.
Keywords: Circulating tumor cells (CTCs) Magnetic gradient chip (Mag-Gradient Chip) Heterogeneity Magnetic activated cell sorting (MACS) Epithelial cell adhesion molecule (EpCAM)
1. Introduction Circulating tumor cells (CTCs) is the single cancer cell which circulated human body through the blood vessel network from the primary tumor site after lack of nutrient and oxygen due to the overpopulation. The CTCs in the blood vessel keep trying to find the secondary site where they can survive with abundant nutrient and oxygen supplement. This relocation mechanism is called metastasis of tumor. This tumor cells are well known as the important clinical cue for diagnosis or prognosis of tumor metastasis. The single cancer cells escaped from the primary tumor site is mostly filtered at lymph node where they met the first immune system of our body. Because of this primary filtration process, the detection of CTCs is extremely hard because of its rarity (typically n
Corresponding author. E-mail address: [email protected] (B. Kwak). 1 These authors contributed equally to this paper.
http://dx.doi.org/10.1016/j.bios.2016.08.002 0956-5663/& 2016 Elsevier B.V. All rights reserved.
1–100 in 5 109 blood cells, 1 ml volume of peripheral blood from human cancer patient) (Hyun and Jung, 2014) in cancer patient's blood sample. Many researchers have developed their own microfluidic system to isolate the CTCs using biological (Nagrath et al., 2007; Stott et al., 2010), magnetical (Kang et al., 2012), mechanical (Park et al., 2014; Tang et al., 2014), and hydro-dynamical (Sollier et al., 2014; Park and Jung, 2009; Choi et al., 2011; Murlidhar et al., 2014) methods as well as combined method (Ozkumur et al., 2013; Moon et al., 2011) with activated sorting mechanisms. Recently, not only for the rarity of CTCs, but also the characterization of heterogeneity of CTCs are raising as main challengeable problem on microfluidic research field. The metastatic CTCs in the blood vessel required certain transforming step for extravasating known as de-differentiation for enabling invasion and dissemination into other organs. (Prang et al., 2005) Up to now, most of CTCs researches has focused on the direct cell culture of CTCs on their own microfluidic system for the further downstream analysis such as DNA sequencing to characterize the
154
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
heterogeneity of CTCs. (Kang et al., 2012) Although few researchers have tried to isolate and specify the heterogeneity of CTCs simultaneously using the microfluidic system, it is still a formidable challenge due to the low isolating yield, purity of heterogeneous CTCs and complex fluidic system (Mohamadi et al., 2015). To solve the problem with simple microfluidic system, we have developed the magnetic force gradient-based microfluidic system (Mag-Gradient Chip) which can specify the heterogeneity subpopulation of CTCs based on the EpCAM expression differences while the CTCs is being isolated by magnetic nanoparticles (MNPs). We have synthesized and surface modified the MNPs for the specific binding with the EpCAM on the CTCs. The different number of EpCAM on the EpCAM positive or negative CTCs is influenced by different amount of magnetic driven force. The magnetic force difference let the CTCs isolates at the different location on the Mag-Gradient Chip. Our experimental results show that the Mag-Gradient Chip promises to isolate and specify the heterogenic CTCs at the same time with high yield.
2. Materials and methods 2.1. Microchannel design The microchannel for the Mag-Gradient Chip is shown in Fig. 1. The microchannel has inlets, sample preparation segments, CTCs trapping segments, and a broadened outlet. Multiple micro-post arrays in the sample preparation segments designed to break down the aggregated cells or filtering the debris before CTCs reached the trapping segments. Five serpentine microfluidic trapping segments were fabricated for the CTCs trapping as shown in Fig. 1. Each trapping segment has 30 rectangle sub-segments
which are located on perpendicular direction in the fluid flow (Kang et al., 2012). From the serpentine trapping segment no.5 to 1, the magnetic flux density is square inverse proportion to the distance from the permanent magnet bar and it can generate the magnetic force gradient to isolate the EpCAM-positive or negative CTCs at different location of the segment channel. The distance between magnetic bar and each trapping segment channel no. 1, 2, 3, 4, and 5 is positioned in 500 mm, 1550 mm, 2600 mm, 3650 mm, and 4700 mm gap from the magnet bar, respectively. The designed depth of Mag-Gradient Chip is 80 mm. The detailed Mag-Gradient Chip fabrication process is described in Supplementary materials. 2.2. Magnetic nanoparticle conjugation The MCF-7 and MDA-MB-231 cells from same origin carcinoma (i.e. human breast cancer cells), which have different EpCAM expression level, were selected. (Hyun and Jung, 2014) The MCF-7 and MDA-MB-231 have 222.7 103 binding sites/cell and 1.7 103 binding sites/cell on the membrane surface, respectively. (Prang et al., 2005) The immuno-fluorescence stained cell images were shown in supplementary materials Fig. S1(a, b) To conjugate the magnetic particles with CTCs, Fe3O4 MNPs have been synthesized by a simple hydrothermal treatment of FeCl3, citrate, polyacrylamide, and urea. (Cheng et al., 2010). The characteristics of synthesized MNPs were shown in supplementary materials Fig. S2. Hydroxylated MNPs were functionalized with (3-Glycidyloxypropyl)tri-methoxysilane (Sigma-Aldrich, Missouri) for immobilization of EpCAM antibody (Polyclonal, Rabbit/IgG, Thermo Scientific, Massachusetts). After the washing process with ethanol and Dulbecco's phosphate buffered saline (DPBS, Thermo Scientific, Massachusetts), the surface modified MNPs were conjugate with the EpCAM antibody. 2 mg/ml of this EpCAM antibody-
Fig. 1. (a) Schematic diagram, and (b) Optical and microscopic images of the Mag-Gradient Chip.
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
conjugated MNPs was gently mixed with fixed CTCs (formaldehyde solution, Sigma-Aldrich, Missouri) for 30 min. Remove unbound MNPs using centrifuge down process. Goat Serum (Jackson ImmunoResearch Laboratories, Inc., Pensylvania) is added for blocking to prevent the nonspecific binding. The MNP conjugated CTCs morphologies are shown in supplementary materials Fig. S1(c, d).
155
3. Theory and mechanism To validate the overall magnetic forces on the MNPs labeled CTCs, the net force on a single magnetic particle is calculated. A single MNP in liquid medium within the microchannel would have to be magnetized due to the presence of the magnetic field, B. Assuming that a magnetized particle may be represented as an equivalent dipole, the magnetic force acting on a spherical MNP with radius R, magnetic susceptibility χm and magnetic permeability of vacuum m0 is given by (Song et al., 2010):
Fig. 2. (a) Magnetic nanoparticle conjugation scheme between surface modified MNPs and EpCAM on CTCs, (b) Magnetic force difference scheme between EpCAM negative/ positive CTCs. Magnetic micro-particles(MMPs) selective isolation experimental results under various of (c) the magnetic flux density, and (d) the fluid flow velocity. The green and red fluorescence represents 5 mm dia. and 1 mm dia. of magnetic particle, respectively. The normalized particle distribution of (e) 5 mm dia. and (f) 1 mm dia. of MMPs in Mag-Gradient Chip under 4.75 kG magnetic flux density and 41.6 mm/s fluid flow velocity conditions, respectively.
156
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
→2 Vparticleχm ∇B → Fmag = μ0
(1)
The moving direction of MP can be estimated by solving the motion equation which net forces acting on the particle. There are two dominant forces can be act on the particle which including the magnetic force of Eq. (1), Fmag, and the medium fluid drag force, Fdrag. The summation of net force acting on a single MP under magnetic field can be derived by following Newton's law of motion.
→ → Fmag = Fdrag
(2)
→ → Fdrag = 6πηRv
(3)
→2 Vparticleχm ∇B μ0
→ = 6πηRv
→2 2 → 2R χm ∇B v = 9ημ 0
(4)
of total fluorescence signal from 5 mm diameter of MMPs were isolated between segments 5-3. On the contrary to this, over 84% of 1 mm dia. of MMPs was isolated between segments 3-1 which magnet bar close location. Among the various magnetic flux densities, 4.75 kG of magnetic flux density was selected for the high throughput of CTCs isolation experiments. The maximum irregularity of magnetic flux density on the along the magnet surface is measured at 70.07 kG (n¼3). Fig. 2(d) shows that the MMP isolating results image under different fluid flow velocity. The magnetic flux density is fixed at 4.75 kG. At 41.6 mm/s fluid flow velocity, averaging 75% of 5 mm of MMPs were isolated between segments 5-3. On the contrary to this, 79% of total fluorescence signal from 1 mm dia. of MMPs was isolated between segments 3-1 which magnet bar close location. Over 41.6 mm/s, it is too fast to selectively isolating of different MMPs in Mag-Gradient Chip. Analogous to this, at the 20.8 mm/s fluid flow condition is too slow to put the particle into the microfluidic channel. Instead of the particle movement, the particles got stuck in the tubing or sample preparation segment due to the magnetic forces. Fig. 2(e, f) shows the normalized particle distribution under optimized conditions. From the feasibility test results using MMPs, we fixed the optimized conditions and performed all following CTCs isolation experiments under 4.75 kG of magnetic flux density, and 41.6 mm/s fluid flow velocity. The original MMPs isolating results are shown in supplementary materials Fig. S3.
(5)
Where η is the viscosity of the fluid medium, R is the radius of MP, → v is the vector component of the particle's instantaneous velocity under these forces. Vparticle is the particle volume. The overall velocity of single MNP to the magnet direction derived by electromagnetic field force can be derived shown in Eq. (5). The escape velocity from the main fluid flow stream is getting faster as inter magnetic-particle distance decreased. As illustrated in Fig. 2(b), using this theoretical net magnetic force differences exerted on the MNP, the EpCAM positive and negative cells can be distinguished by trapping location in Mag-Gradient Chip. From the addressed set of equations, the MNP trapping position under specific magnetic forces is simulated.
4. Results and discussion 4.1. Magnetic micro-particle feasibility test To prove the addressed hypothesis, feasibility tests have been performed using total magnetic force on CTCs-mimicked magnetic micro-particles (MMP). Both EpCAM positive and negative cell have certain number of EpCAM expression level on the cell membrane. The average EpCAM binding sites per cell of MCF-7 and MDA-MB-231 are 222.1 103, and 1.7 103, respectively. (Prang et al., 2005) We assumed that 200 nm diameter of MNPs can bind on the EpCAM on the surface of CTCs with 30% of binding yield (Arakaki et al., 2004). The overall magnetic force exerted on the MNPs labeled MCF-7 and MDA-MB-231 are equivalent to the 5.0–5.9 mm (5 mm diameter, Green fluorescent carboxyl magnetic particle, Spherotech, Illinois) diameter and 1.0–1.4 mm (1 mm diameter, Red fluorescent carboxyl magnetic particle, Spherotech, Illinois) MMPs. The selective isolation test was performed using two different size based MMPs under various magnetic flux densities and fluid flow velocities as shown in Fig. 2(c–f). The magnetic flux density and fluid flow velocity are the important boundary conditions for the CTCs isolating efficiency, and throughput, respectively. Fig. 2(c) shows that the MMP isolation results under different magnetic flux density. The fluid flow velocity is fixed at 41.6 mm/s. Over 3.36 kG magnetic flux density, approximately 75%
4.2. Single cell line isolation Before we applied this Mag-Gradient Chip for the mixed heterogeneous CTC sample, isolation experiments of each cell line (EpCAM positive: MCF-7, and EpCAM negative: MDA-MB-231) were conducted. Those two cell lines were labeled by 200 nm diameter MNPs before it tested with the Mag-Gradient Chip. An each cell concentrations were fixed at 1 105 cells/ml. Fig. 3(a, c) and (b, d) show that the isolated and distributed images of MCF-7 and MDA-MB-231, respectively. The isolating efficiency of MCF-7 and MDA-MB-231 is 97.3% 71.15% and 85.3% 72.52%, respectively. The capturing efficiency is calculated using the ratio between input number of cells versus number of cells in the waste. The isolating efficiency of CTCs in Mag-Gradient Chip is enough to apply on diagnosis of the tumor metastasis compared with other researches. (Hyun and Jung, 2014) Moreover, Fig. 3(a, b) shows that the 92.2% of MCF-7% and 78.0% of MDA-MB-231 of total captured CTCs is isolated at the different locations as we designed. This result proves that Mag-Gradient Chip can isolate the total CTCs and characterize the heterogenic CTCs simultaneously. 4.3. Selective isolation of mixed heterogeneity subpopulation of CTCs Premixed heterogeneous CTCs sample were conducted to selective isolation experiment with respect to the EpCAM positive/ negative cell difference. Prior to make a mixed sample, the plasma membrane of MCF-7 and the MDA-MB-231 were stained as green and deep red fluorescence as shown in Fig. 3(e), respectively. (CellMask™ plasma membrane stain, Thermo Scientific, Massachusetts) Each cell lines were conjugated with 200 nm diameter of MNPs separately. The 5 different final cell concentrations (1 103 cells/ml, 3 103 cells/ml, 1 104 cells/ml, 3 104 cells/ml, and 1 105 cells/ml) of mixed heterogenic CTCs samples were prepared to test. The 1:1 number ratio of MCF-7 and MDA-MB-231 is applied. Each concentration samples were tested 3 times repeatedly. The Fig. 3(e) shows that the experimental result of 1 105 cells/ml sample. The isolating efficiency of MCF-7 and MDA-MB-231 is 96.7% 72.0% and 77.2% 7 5.5%. The inset image of Fig. 3(e) shows that the different kinds of CTCs were isolated with respect to the distance from the magnet bar. As the distance is
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
157
Fig. 3. The isolation results of (a, c) EpCAM positive CTCs (MCF-7) and (b, d) EpCAM negative CTCs (MDA-MB-231). (n¼ 3) The selective isolation result images of (e) mixed heterogeneous CTCs. The MCF-7 and MDA-MB-231 were green and deep red fluorescence stained on cell plasma membrane. The right inset images were enlarged of each trapping sub-segment. (f) CTCs distribution at each segment (n¼ 15), and (g) CTCs isolating efficiency under different cell concentrations. (n ¼ 3, each sample was tested 3 times) All cell nuclei were stained by DAPI (4′,6-diamidino-2-phenylindol).
158
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
close to the magnet bar, the MDA-MB-231, which deep red fluorescence stained cells are dominant in the trapping sub-segment. The total average isolating efficiency of MCF-7 and MDA-MB-231 through 5 different cell concentrations is 95.7% 71.6%, 79.3% 73.8%, respectively. (n ¼15, each sample was tested 3 times) Also, the 71.3% of isolated MCF-7 is located between segment no. 5 to 3 where far site from the magnet bar. On the contrary to this, the average 76.9% of isolated MDA-MB-231 is located between segment no. 3 to 1 where close site from magnet bar. This result clearly shows that the heterogenic CTCs can be isolated and sorted by their EpCAM characteristic simultaneously.
5. Conclusion The CTCs is the important clinical evidence for the prognosis or diagnosis for tumor metastasis because of its relocation mechanisms. Not all the CTCs are get involved to this relocation mechanisms but still the total number of CTCs in the patient's peripheral blood is believed that the major parameter for diagnose of the tumor metastasis due to the hardness or absence of easy further downstream analysis methods or system. The EpCAM expression level is one of the big differences between metastatic and non-metastatic cancer cell. In this study, we have designed and developed the Mag-Gradient Chip which can generate the magnetic flux density gradient with respect to the distance from the magnet bar. Moreover, we have optimized the MNPs surface conditions to conjugate with the EpCAM on the CTCs surface membrane with respect to the expression level. The Mag-Gradient Chip can selectively isolate the heterogeneous CTCs at the different location in the system. This simple and multi-functional MagGrdient Chip can isolate the 3 ml of heterogeneous CTCs sample in 1 h. We assumed that the Mag-Gradient Chip can be powerful system to provide the opportunities to accurate diagnose of the tumor metastasis because of its high isolating yield and selective isolation function.
Acknowledgments This work was supported by the “Robotic point-of-care clinic technologies for neglected class of people” project of Korea Institute of Machinery and Materials under the auspices of the
Ministry of Science, ICT and Future Planning, South Korea (SC1140).
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.bios.2016.08.002.
References Arakaki, A., Hideshima, S., Nakagawa, T., Niwa, D., Tanaka, T., Matsunaga, T., Osaka, T., 2004. Biotechnol. Bioeng., 88(4), pp. 543–546. Cheng, W., Tang, K., Qi, Y., Sheng, J., Liu, Z., 2010. J. Mater. Chem. 20, 1799–1805. Choi, S., Ku, T., Song, S., Choi, C., Park, J.K., 2011. Lab. Chip 11 (3), 413–418. Hyun, K.A., Jung, H.I., 2014. Lab. Chip 14, 45–56. Kang, J.H., Krause, S., Tobin, H., Mammoto, A., Kanapathipillai, M., Ingber, D.E., 2012. Lab. Chip 12 (12), 2175–2181. Mohamadi, R.M., Besant, J.D., Mepham, A., Green, B., Mahmoudian, L., Gibbs, T., Ivanov, I., Malvea, A., Stojcic, J., Allan, A.L., Lowes, L.E., Sargent, E.H., Nam, R.K., Kelley, S.O., 2015. Angew. Chem. Int. Ed. Engl. 54 (1), 139–143. Moon, H.S., Kwon, K., Kim, S.I., Han, H., Sohn, J., Lee, S., Jung, H.I., 2011. Lab. Chip 11 (6), 1118–1125. Murlidhar, V., Zeinali, M., Grabauskiene, S., Ghannad-Rezaie, M., Wicha, M.S., Simeone, D.M., Ramnath, N., Reddy, R.M., Nagrath, S., 2014. Small 10 (23), 4895–4904. Nagrath, S., Sequist, L.V., Maheswaran, S., Bell, D.W., Irimia, D., Ulkus, L., Smith, M.R., Kwak, E.L., Digumarthy, S., Muzikansky, A., Ryan, P., Balis, U.J., Tompkins, R.G., Haber, D.A., Toner, M., 2007. Nature 450, 1235–1239. Ozkumur, E., Shah, A.M., Ciciliano, J.C., Emmink, B.L., Miyamoto, D.T., Brachtel, E., Yu, M., Chen, P.I., Morgan, B., Trautwein, J., Kimura, A., Sengupta, S., Stott, S.L., Karabacak, N.M., Barber, T.A., Walsh, J.R., Smith, K., Spuhler, P.S., Sullivan, J.P., Lee, R.J., Ting, D.T., Luo, X., Shaw, A.T., Bardia, A., Sequist, L.V., Louis, D.N., Maheswaran, S., Kapur, R., Haber, D.A., Toner, M., 2013. Sci. Transl. Med. 5 (179), 179ra47. Park, J.M., Kim, M.S., Moon, H.S., Yoo, C.E., Park, D., Kim, Y.J., Han, K.Y., Lee, J.Y., Oh, J. H., Kim, S.S., Park, W.Y., Lee, W.Y., Huh, N., 2014. Anal. Chem. 86 (8), 3735–3742. Park, J.S., Jung, H.I., 2009. Anal. Chem. 81 (20), 8280–8288. Prang, N., Preithner, S., Brischwein, K., Göster, P., Wöppel, A., Müller, J., Steiger, C., Peters, M., Baeuerle, P.A., DaSilva, A.J., 2005. Br. J. Cancer 92, 342–349. Sollier, E., Go, D.E., Che, J., Gossett, D.R., O’Byrne, S., Weaver, W.M., Kummer, N., Rettig, M., Goldman, J., Nickols, N., McCloskey, S., Kulkarni, R.P., Di Carlo, D., 2014. Lab. Chip 14 (1), 63–77. Song, S.H., Choi, J.H., Jung, H.I., 2010. Electrophoresis 31, 2762–2770. Stott, S.L., Hsu, C.H., Tsukrov, D.I., Yu, M., Miyamoto, D.T., Waltman, B.A., Rothenberg, S.M., Shah, A.M., Smas, M.E., Korir, G.K., Floyd Jr., F.P., Gilman, A.J., Lord, J. B., Winokur, D., Springer, S., Irimia, D., Nagrath, S., Sequist, L.V., Lee, R.J., Isselbacher, K.J., Maheswaran, S., Haber, D.A., Toner, M., 2010. Proc. Natl. Acad. Sci. USA 107 (43), 18392–18397. Tang, Y., Shi, J., Li, S., Wang, L., Cayre, Y.E., Chen, Y., 2014. Sci. Rep. 4, 6052. http://dx. doi.org/10.1038/srep06052.
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Selective isolation of magnetic nanoparticle-mediated heterogeneity subpopulation of circulating tumor cells using magnetic gradient based microfluidic system Bongseop Kwak a,n,1, Jaehun Lee a,b,1, Dongkyu Lee a, Kangho Lee a, Ohwon Kwon a, Shinwon Kang b, Youngwoo Kim a a Korea Institute of Machinery and Materials, Daegu Research Center for Medical Devices and Rehab. Engineering, Department of Medical Device, 330 Techno Sunhwan-ro, Yuga-myeon, Dalsung-gun, Daegu, 42994 Republic of Korea b Kyungpook National University, College of IT Engineering, School of Electronics Engineering, 80 Daehak-ro, Buk-gu, Daegu, 41566 Republic of Korea
art ic l e i nf o
a b s t r a c t
Article history: Received 13 June 2016 Received in revised form 29 July 2016 Accepted 1 August 2016 Available online 2 August 2016
Relocation mechanisms of the circulating tumor cells (CTCs) from the primary site to the secondary site through the blood vessel network cause tumor metastasis. Despite of the importance to diagnose the cancer metastasis by CTCs, still it is formidable challenge to use in the clinical purpose because of the rarity and the heterogeneity of CTCs in the cancer patient’s peripheral blood sample. In this study we have developed magnetic force gradient based microfluidic chip (Mag-Gradient Chip) for isolating the total number of CTCs in the sample and characterizing the state of CTCs simultaneously with respect to the epithelial cell adhesion molecule (EpCAM) expression level. We have synthesized magnetic nanoparticles (MNPs) using hydrothermal method and functionalized anti-EpCAM on their surface for the specific binding with CTCs. The Mag-Gradient Chip designed to isolate and classify the CTCs by isolating at the different location in the chip using magnetic force differences depending on the EpCAM expression level. We observed 95.7% of EpCAM positive and 79.3% of EpCAM negative CTCs isolated in the MagGradient Chip. At the same time, the 71.3% of isolated EpCAM positive CTCs were isolated at the first half area whereas the 76.9% of EpCAM negative CTCs were collected at the latter half area. The Mag-Gradient Chip can isolate the 3 ml of heterogeneous CTCs sample in 1 h with high isolating yield. The EpCAM expression level dose not means essential condition of the metastatic CTCs, but the Mag-Gradient Chip can shorten the date to diagnose the cancer metastasis in clinic. & 2016 Elsevier B.V. All rights reserved.
Keywords: Circulating tumor cells (CTCs) Magnetic gradient chip (Mag-Gradient Chip) Heterogeneity Magnetic activated cell sorting (MACS) Epithelial cell adhesion molecule (EpCAM)
1. Introduction Circulating tumor cells (CTCs) is the single cancer cell which circulated human body through the blood vessel network from the primary tumor site after lack of nutrient and oxygen due to the overpopulation. The CTCs in the blood vessel keep trying to find the secondary site where they can survive with abundant nutrient and oxygen supplement. This relocation mechanism is called metastasis of tumor. This tumor cells are well known as the important clinical cue for diagnosis or prognosis of tumor metastasis. The single cancer cells escaped from the primary tumor site is mostly filtered at lymph node where they met the first immune system of our body. Because of this primary filtration process, the detection of CTCs is extremely hard because of its rarity (typically n
Corresponding author. E-mail address: [email protected] (B. Kwak). 1 These authors contributed equally to this paper.
http://dx.doi.org/10.1016/j.bios.2016.08.002 0956-5663/& 2016 Elsevier B.V. All rights reserved.
1–100 in 5 109 blood cells, 1 ml volume of peripheral blood from human cancer patient) (Hyun and Jung, 2014) in cancer patient's blood sample. Many researchers have developed their own microfluidic system to isolate the CTCs using biological (Nagrath et al., 2007; Stott et al., 2010), magnetical (Kang et al., 2012), mechanical (Park et al., 2014; Tang et al., 2014), and hydro-dynamical (Sollier et al., 2014; Park and Jung, 2009; Choi et al., 2011; Murlidhar et al., 2014) methods as well as combined method (Ozkumur et al., 2013; Moon et al., 2011) with activated sorting mechanisms. Recently, not only for the rarity of CTCs, but also the characterization of heterogeneity of CTCs are raising as main challengeable problem on microfluidic research field. The metastatic CTCs in the blood vessel required certain transforming step for extravasating known as de-differentiation for enabling invasion and dissemination into other organs. (Prang et al., 2005) Up to now, most of CTCs researches has focused on the direct cell culture of CTCs on their own microfluidic system for the further downstream analysis such as DNA sequencing to characterize the
154
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
heterogeneity of CTCs. (Kang et al., 2012) Although few researchers have tried to isolate and specify the heterogeneity of CTCs simultaneously using the microfluidic system, it is still a formidable challenge due to the low isolating yield, purity of heterogeneous CTCs and complex fluidic system (Mohamadi et al., 2015). To solve the problem with simple microfluidic system, we have developed the magnetic force gradient-based microfluidic system (Mag-Gradient Chip) which can specify the heterogeneity subpopulation of CTCs based on the EpCAM expression differences while the CTCs is being isolated by magnetic nanoparticles (MNPs). We have synthesized and surface modified the MNPs for the specific binding with the EpCAM on the CTCs. The different number of EpCAM on the EpCAM positive or negative CTCs is influenced by different amount of magnetic driven force. The magnetic force difference let the CTCs isolates at the different location on the Mag-Gradient Chip. Our experimental results show that the Mag-Gradient Chip promises to isolate and specify the heterogenic CTCs at the same time with high yield.
2. Materials and methods 2.1. Microchannel design The microchannel for the Mag-Gradient Chip is shown in Fig. 1. The microchannel has inlets, sample preparation segments, CTCs trapping segments, and a broadened outlet. Multiple micro-post arrays in the sample preparation segments designed to break down the aggregated cells or filtering the debris before CTCs reached the trapping segments. Five serpentine microfluidic trapping segments were fabricated for the CTCs trapping as shown in Fig. 1. Each trapping segment has 30 rectangle sub-segments
which are located on perpendicular direction in the fluid flow (Kang et al., 2012). From the serpentine trapping segment no.5 to 1, the magnetic flux density is square inverse proportion to the distance from the permanent magnet bar and it can generate the magnetic force gradient to isolate the EpCAM-positive or negative CTCs at different location of the segment channel. The distance between magnetic bar and each trapping segment channel no. 1, 2, 3, 4, and 5 is positioned in 500 mm, 1550 mm, 2600 mm, 3650 mm, and 4700 mm gap from the magnet bar, respectively. The designed depth of Mag-Gradient Chip is 80 mm. The detailed Mag-Gradient Chip fabrication process is described in Supplementary materials. 2.2. Magnetic nanoparticle conjugation The MCF-7 and MDA-MB-231 cells from same origin carcinoma (i.e. human breast cancer cells), which have different EpCAM expression level, were selected. (Hyun and Jung, 2014) The MCF-7 and MDA-MB-231 have 222.7 103 binding sites/cell and 1.7 103 binding sites/cell on the membrane surface, respectively. (Prang et al., 2005) The immuno-fluorescence stained cell images were shown in supplementary materials Fig. S1(a, b) To conjugate the magnetic particles with CTCs, Fe3O4 MNPs have been synthesized by a simple hydrothermal treatment of FeCl3, citrate, polyacrylamide, and urea. (Cheng et al., 2010). The characteristics of synthesized MNPs were shown in supplementary materials Fig. S2. Hydroxylated MNPs were functionalized with (3-Glycidyloxypropyl)tri-methoxysilane (Sigma-Aldrich, Missouri) for immobilization of EpCAM antibody (Polyclonal, Rabbit/IgG, Thermo Scientific, Massachusetts). After the washing process with ethanol and Dulbecco's phosphate buffered saline (DPBS, Thermo Scientific, Massachusetts), the surface modified MNPs were conjugate with the EpCAM antibody. 2 mg/ml of this EpCAM antibody-
Fig. 1. (a) Schematic diagram, and (b) Optical and microscopic images of the Mag-Gradient Chip.
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
conjugated MNPs was gently mixed with fixed CTCs (formaldehyde solution, Sigma-Aldrich, Missouri) for 30 min. Remove unbound MNPs using centrifuge down process. Goat Serum (Jackson ImmunoResearch Laboratories, Inc., Pensylvania) is added for blocking to prevent the nonspecific binding. The MNP conjugated CTCs morphologies are shown in supplementary materials Fig. S1(c, d).
155
3. Theory and mechanism To validate the overall magnetic forces on the MNPs labeled CTCs, the net force on a single magnetic particle is calculated. A single MNP in liquid medium within the microchannel would have to be magnetized due to the presence of the magnetic field, B. Assuming that a magnetized particle may be represented as an equivalent dipole, the magnetic force acting on a spherical MNP with radius R, magnetic susceptibility χm and magnetic permeability of vacuum m0 is given by (Song et al., 2010):
Fig. 2. (a) Magnetic nanoparticle conjugation scheme between surface modified MNPs and EpCAM on CTCs, (b) Magnetic force difference scheme between EpCAM negative/ positive CTCs. Magnetic micro-particles(MMPs) selective isolation experimental results under various of (c) the magnetic flux density, and (d) the fluid flow velocity. The green and red fluorescence represents 5 mm dia. and 1 mm dia. of magnetic particle, respectively. The normalized particle distribution of (e) 5 mm dia. and (f) 1 mm dia. of MMPs in Mag-Gradient Chip under 4.75 kG magnetic flux density and 41.6 mm/s fluid flow velocity conditions, respectively.
156
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
→2 Vparticleχm ∇B → Fmag = μ0
(1)
The moving direction of MP can be estimated by solving the motion equation which net forces acting on the particle. There are two dominant forces can be act on the particle which including the magnetic force of Eq. (1), Fmag, and the medium fluid drag force, Fdrag. The summation of net force acting on a single MP under magnetic field can be derived by following Newton's law of motion.
→ → Fmag = Fdrag
(2)
→ → Fdrag = 6πηRv
(3)
→2 Vparticleχm ∇B μ0
→ = 6πηRv
→2 2 → 2R χm ∇B v = 9ημ 0
(4)
of total fluorescence signal from 5 mm diameter of MMPs were isolated between segments 5-3. On the contrary to this, over 84% of 1 mm dia. of MMPs was isolated between segments 3-1 which magnet bar close location. Among the various magnetic flux densities, 4.75 kG of magnetic flux density was selected for the high throughput of CTCs isolation experiments. The maximum irregularity of magnetic flux density on the along the magnet surface is measured at 70.07 kG (n¼3). Fig. 2(d) shows that the MMP isolating results image under different fluid flow velocity. The magnetic flux density is fixed at 4.75 kG. At 41.6 mm/s fluid flow velocity, averaging 75% of 5 mm of MMPs were isolated between segments 5-3. On the contrary to this, 79% of total fluorescence signal from 1 mm dia. of MMPs was isolated between segments 3-1 which magnet bar close location. Over 41.6 mm/s, it is too fast to selectively isolating of different MMPs in Mag-Gradient Chip. Analogous to this, at the 20.8 mm/s fluid flow condition is too slow to put the particle into the microfluidic channel. Instead of the particle movement, the particles got stuck in the tubing or sample preparation segment due to the magnetic forces. Fig. 2(e, f) shows the normalized particle distribution under optimized conditions. From the feasibility test results using MMPs, we fixed the optimized conditions and performed all following CTCs isolation experiments under 4.75 kG of magnetic flux density, and 41.6 mm/s fluid flow velocity. The original MMPs isolating results are shown in supplementary materials Fig. S3.
(5)
Where η is the viscosity of the fluid medium, R is the radius of MP, → v is the vector component of the particle's instantaneous velocity under these forces. Vparticle is the particle volume. The overall velocity of single MNP to the magnet direction derived by electromagnetic field force can be derived shown in Eq. (5). The escape velocity from the main fluid flow stream is getting faster as inter magnetic-particle distance decreased. As illustrated in Fig. 2(b), using this theoretical net magnetic force differences exerted on the MNP, the EpCAM positive and negative cells can be distinguished by trapping location in Mag-Gradient Chip. From the addressed set of equations, the MNP trapping position under specific magnetic forces is simulated.
4. Results and discussion 4.1. Magnetic micro-particle feasibility test To prove the addressed hypothesis, feasibility tests have been performed using total magnetic force on CTCs-mimicked magnetic micro-particles (MMP). Both EpCAM positive and negative cell have certain number of EpCAM expression level on the cell membrane. The average EpCAM binding sites per cell of MCF-7 and MDA-MB-231 are 222.1 103, and 1.7 103, respectively. (Prang et al., 2005) We assumed that 200 nm diameter of MNPs can bind on the EpCAM on the surface of CTCs with 30% of binding yield (Arakaki et al., 2004). The overall magnetic force exerted on the MNPs labeled MCF-7 and MDA-MB-231 are equivalent to the 5.0–5.9 mm (5 mm diameter, Green fluorescent carboxyl magnetic particle, Spherotech, Illinois) diameter and 1.0–1.4 mm (1 mm diameter, Red fluorescent carboxyl magnetic particle, Spherotech, Illinois) MMPs. The selective isolation test was performed using two different size based MMPs under various magnetic flux densities and fluid flow velocities as shown in Fig. 2(c–f). The magnetic flux density and fluid flow velocity are the important boundary conditions for the CTCs isolating efficiency, and throughput, respectively. Fig. 2(c) shows that the MMP isolation results under different magnetic flux density. The fluid flow velocity is fixed at 41.6 mm/s. Over 3.36 kG magnetic flux density, approximately 75%
4.2. Single cell line isolation Before we applied this Mag-Gradient Chip for the mixed heterogeneous CTC sample, isolation experiments of each cell line (EpCAM positive: MCF-7, and EpCAM negative: MDA-MB-231) were conducted. Those two cell lines were labeled by 200 nm diameter MNPs before it tested with the Mag-Gradient Chip. An each cell concentrations were fixed at 1 105 cells/ml. Fig. 3(a, c) and (b, d) show that the isolated and distributed images of MCF-7 and MDA-MB-231, respectively. The isolating efficiency of MCF-7 and MDA-MB-231 is 97.3% 71.15% and 85.3% 72.52%, respectively. The capturing efficiency is calculated using the ratio between input number of cells versus number of cells in the waste. The isolating efficiency of CTCs in Mag-Gradient Chip is enough to apply on diagnosis of the tumor metastasis compared with other researches. (Hyun and Jung, 2014) Moreover, Fig. 3(a, b) shows that the 92.2% of MCF-7% and 78.0% of MDA-MB-231 of total captured CTCs is isolated at the different locations as we designed. This result proves that Mag-Gradient Chip can isolate the total CTCs and characterize the heterogenic CTCs simultaneously. 4.3. Selective isolation of mixed heterogeneity subpopulation of CTCs Premixed heterogeneous CTCs sample were conducted to selective isolation experiment with respect to the EpCAM positive/ negative cell difference. Prior to make a mixed sample, the plasma membrane of MCF-7 and the MDA-MB-231 were stained as green and deep red fluorescence as shown in Fig. 3(e), respectively. (CellMask™ plasma membrane stain, Thermo Scientific, Massachusetts) Each cell lines were conjugated with 200 nm diameter of MNPs separately. The 5 different final cell concentrations (1 103 cells/ml, 3 103 cells/ml, 1 104 cells/ml, 3 104 cells/ml, and 1 105 cells/ml) of mixed heterogenic CTCs samples were prepared to test. The 1:1 number ratio of MCF-7 and MDA-MB-231 is applied. Each concentration samples were tested 3 times repeatedly. The Fig. 3(e) shows that the experimental result of 1 105 cells/ml sample. The isolating efficiency of MCF-7 and MDA-MB-231 is 96.7% 72.0% and 77.2% 7 5.5%. The inset image of Fig. 3(e) shows that the different kinds of CTCs were isolated with respect to the distance from the magnet bar. As the distance is
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
157
Fig. 3. The isolation results of (a, c) EpCAM positive CTCs (MCF-7) and (b, d) EpCAM negative CTCs (MDA-MB-231). (n¼ 3) The selective isolation result images of (e) mixed heterogeneous CTCs. The MCF-7 and MDA-MB-231 were green and deep red fluorescence stained on cell plasma membrane. The right inset images were enlarged of each trapping sub-segment. (f) CTCs distribution at each segment (n¼ 15), and (g) CTCs isolating efficiency under different cell concentrations. (n ¼ 3, each sample was tested 3 times) All cell nuclei were stained by DAPI (4′,6-diamidino-2-phenylindol).
158
B. Kwak et al. / Biosensors and Bioelectronics 88 (2017) 153–158
close to the magnet bar, the MDA-MB-231, which deep red fluorescence stained cells are dominant in the trapping sub-segment. The total average isolating efficiency of MCF-7 and MDA-MB-231 through 5 different cell concentrations is 95.7% 71.6%, 79.3% 73.8%, respectively. (n ¼15, each sample was tested 3 times) Also, the 71.3% of isolated MCF-7 is located between segment no. 5 to 3 where far site from the magnet bar. On the contrary to this, the average 76.9% of isolated MDA-MB-231 is located between segment no. 3 to 1 where close site from magnet bar. This result clearly shows that the heterogenic CTCs can be isolated and sorted by their EpCAM characteristic simultaneously.
5. Conclusion The CTCs is the important clinical evidence for the prognosis or diagnosis for tumor metastasis because of its relocation mechanisms. Not all the CTCs are get involved to this relocation mechanisms but still the total number of CTCs in the patient's peripheral blood is believed that the major parameter for diagnose of the tumor metastasis due to the hardness or absence of easy further downstream analysis methods or system. The EpCAM expression level is one of the big differences between metastatic and non-metastatic cancer cell. In this study, we have designed and developed the Mag-Gradient Chip which can generate the magnetic flux density gradient with respect to the distance from the magnet bar. Moreover, we have optimized the MNPs surface conditions to conjugate with the EpCAM on the CTCs surface membrane with respect to the expression level. The Mag-Gradient Chip can selectively isolate the heterogeneous CTCs at the different location in the system. This simple and multi-functional MagGrdient Chip can isolate the 3 ml of heterogeneous CTCs sample in 1 h. We assumed that the Mag-Gradient Chip can be powerful system to provide the opportunities to accurate diagnose of the tumor metastasis because of its high isolating yield and selective isolation function.
Acknowledgments This work was supported by the “Robotic point-of-care clinic technologies for neglected class of people” project of Korea Institute of Machinery and Materials under the auspices of the
Ministry of Science, ICT and Future Planning, South Korea (SC1140).
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.bios.2016.08.002.
References Arakaki, A., Hideshima, S., Nakagawa, T., Niwa, D., Tanaka, T., Matsunaga, T., Osaka, T., 2004. Biotechnol. Bioeng., 88(4), pp. 543–546. Cheng, W., Tang, K., Qi, Y., Sheng, J., Liu, Z., 2010. J. Mater. Chem. 20, 1799–1805. Choi, S., Ku, T., Song, S., Choi, C., Park, J.K., 2011. Lab. Chip 11 (3), 413–418. Hyun, K.A., Jung, H.I., 2014. Lab. Chip 14, 45–56. Kang, J.H., Krause, S., Tobin, H., Mammoto, A., Kanapathipillai, M., Ingber, D.E., 2012. Lab. Chip 12 (12), 2175–2181. Mohamadi, R.M., Besant, J.D., Mepham, A., Green, B., Mahmoudian, L., Gibbs, T., Ivanov, I., Malvea, A., Stojcic, J., Allan, A.L., Lowes, L.E., Sargent, E.H., Nam, R.K., Kelley, S.O., 2015. Angew. Chem. Int. Ed. Engl. 54 (1), 139–143. Moon, H.S., Kwon, K., Kim, S.I., Han, H., Sohn, J., Lee, S., Jung, H.I., 2011. Lab. Chip 11 (6), 1118–1125. Murlidhar, V., Zeinali, M., Grabauskiene, S., Ghannad-Rezaie, M., Wicha, M.S., Simeone, D.M., Ramnath, N., Reddy, R.M., Nagrath, S., 2014. Small 10 (23), 4895–4904. Nagrath, S., Sequist, L.V., Maheswaran, S., Bell, D.W., Irimia, D., Ulkus, L., Smith, M.R., Kwak, E.L., Digumarthy, S., Muzikansky, A., Ryan, P., Balis, U.J., Tompkins, R.G., Haber, D.A., Toner, M., 2007. Nature 450, 1235–1239. Ozkumur, E., Shah, A.M., Ciciliano, J.C., Emmink, B.L., Miyamoto, D.T., Brachtel, E., Yu, M., Chen, P.I., Morgan, B., Trautwein, J., Kimura, A., Sengupta, S., Stott, S.L., Karabacak, N.M., Barber, T.A., Walsh, J.R., Smith, K., Spuhler, P.S., Sullivan, J.P., Lee, R.J., Ting, D.T., Luo, X., Shaw, A.T., Bardia, A., Sequist, L.V., Louis, D.N., Maheswaran, S., Kapur, R., Haber, D.A., Toner, M., 2013. Sci. Transl. Med. 5 (179), 179ra47. Park, J.M., Kim, M.S., Moon, H.S., Yoo, C.E., Park, D., Kim, Y.J., Han, K.Y., Lee, J.Y., Oh, J. H., Kim, S.S., Park, W.Y., Lee, W.Y., Huh, N., 2014. Anal. Chem. 86 (8), 3735–3742. Park, J.S., Jung, H.I., 2009. Anal. Chem. 81 (20), 8280–8288. Prang, N., Preithner, S., Brischwein, K., Göster, P., Wöppel, A., Müller, J., Steiger, C., Peters, M., Baeuerle, P.A., DaSilva, A.J., 2005. Br. J. Cancer 92, 342–349. Sollier, E., Go, D.E., Che, J., Gossett, D.R., O’Byrne, S., Weaver, W.M., Kummer, N., Rettig, M., Goldman, J., Nickols, N., McCloskey, S., Kulkarni, R.P., Di Carlo, D., 2014. Lab. Chip 14 (1), 63–77. Song, S.H., Choi, J.H., Jung, H.I., 2010. Electrophoresis 31, 2762–2770. Stott, S.L., Hsu, C.H., Tsukrov, D.I., Yu, M., Miyamoto, D.T., Waltman, B.A., Rothenberg, S.M., Shah, A.M., Smas, M.E., Korir, G.K., Floyd Jr., F.P., Gilman, A.J., Lord, J. B., Winokur, D., Springer, S., Irimia, D., Nagrath, S., Sequist, L.V., Lee, R.J., Isselbacher, K.J., Maheswaran, S., Haber, D.A., Toner, M., 2010. Proc. Natl. Acad. Sci. USA 107 (43), 18392–18397. Tang, Y., Shi, J., Li, S., Wang, L., Cayre, Y.E., Chen, Y., 2014. Sci. Rep. 4, 6052. http://dx. doi.org/10.1038/srep06052.