Adhesion Assays of Endothelial Cells on Nanopatterned Surfaces

Mar 10, 2010 - Modulation of cell adhesion, proliferation and differentiation on materials designed for body implants. Lucie Bacakova , Elena Filova ,...
0 downloads 0 Views 2MB Size
Anal. Chem. 2010, 82, 3016–3022

Adhesion Assays of Endothelial Cells on Nanopatterned Surfaces within a Microfluidic Channel Se Yon Hwang,† Keon Woo Kwon,† Kyung-Jin Jang,‡ Min Cheol Park,† Jeong Sang Lee,*,§ and Kahp Y. Suh*,†,‡ School of Mechanical and Aerospace Engineering, Seoul National University, Seoul, 151-742, Korea, Interdisciplinary Program in Nano-Science and Technology, Seoul National University, Seoul, 151-742, Korea, and Department of Thoracic & Cardiovascular Surgery, College of Medicine, Boramae Medical Center, Seoul National University, Seoul, 156-707, Korea We present a simple analytical method to measure adhesion of human umbilical vein endothelial cells (HUVECs) and calf pulmonary artery endothelial cells (CPAEs) using nanopatterned, biodegradable poly(lactic-co-glycolic acid) (PLGA) surfaces for potential applications to artificial tissue-engineered blood vessel. Various nanostructured PLGA surfaces (350 nm wide ridges/350 nm grooves, 350 nm ridges/700 nm grooves, 350 nm ridges/1750 nm grooves, 700 nm ridges/350 nm grooves, 1050 nm ridges/350 nm grooves, 1750 nm ridges/350 nm grooves) and flat (unpatterned) surfaces were fabricated on the bottom of polydimethylsiloxane (PDMS) microfluidic channel of 2 mm width and 60 µm height by using thermal imprinting and irreversible channel bonding. To measure adhesion strength of HUVECs and CPAEs, the cells were exposed to a range of shear stress (12, 40, and 80 dyn/ cm2) within the channels for 20 min after a preculture for 3 days and the remaining cells were counted under each condition. The highest adhesion strength was found on the surface of 700 nm wide ridges, 350 nm wide grooves for both cell types. The enhanced adhesion on nanopatterned surfaces can be attributed to two aspects: (i) contact guidance along the line direction and (ii) clustered focal adhesions. In particular, the contact guidance induced cell alignment along the line directions, which in turn lowers wall shear stress applied to the cell surface, as supported by a simple hydrodynamic model based on cell morphology. Microfluidic technology offers great advantages for vascular research because of its ability to control shear stress, growth factor gradient, coculture, and cell migration.1-12 To understand the mechanisms that determine vascular function and dysfunction, it * To whom correspondence should be addressed. E-mail: [email protected] (K.Y.S.); [email protected] (J.S.L.). † School of Mechanical and Aerospace Engineering. ‡ Interdisciplinary Program in Nano-Science and Technology. § Department of Thoracic & Cardiovascular Surgery. (1) Kotsis, D. H.; Spence, D. M. Anal. Chem. 2003, 75, 145–151. (2) Song, J. W.; Gu, W.; Futai, N.; Warner, K. A.; Nor, J. E.; Takayama, S. Anal. Chem. 2005, 77, 3993–3999. (3) Tkachenko, E.; Gutierrez, E.; Ginsberg, M. H.; Groisman, A. Lab Chip 2009, 9, 1085–1095.

3016

Analytical Chemistry, Vol. 82, No. 7, April 1, 2010

is useful to culture endothelial cells (ECs) or smooth muscle cells (SMCs) within a microfluidic channel and expose the cells to relevant biological, chemical, or physical stimuli in vitro in a controlled fashion. In particular, ECs are exposed to hemodynamic shear stress in vivo and incorporation of such a fluidic environment into cell cultures is of paramount importance to understand the response of ECs to fluidic shear stress and potentially the mechanisms for various cardiovascular diseases such as atherosclerosis. Recent developments in artificial blood vessels have made great success using synthetic materials such as Dacron or expanded polytetrafluoroethylene (ePTFE) with a diameter larger than 6 mm. For artificial blood vessels with a diameter less than 6 mm, however, thrombotical events may occur to rapidly narrow the inner space of synthetic grafts. The main reason for this is a relatively low flow rate within small diameter blood vessels, not the aspect of material defects.13 Therefore EC-seeded synthetic grafts have been proposed as the next step of vascular tissue engineering to fabricate artificial blood vessels with diameters less than 6 mm. In a native blood vessel, there is a monolayer of ECs that provide nonthrombogenic interface between the blood flow and the vessel wall. Thus the emulation of a native blood vessel, through the fabrication of endothelial cell coated-synthetic vascular graft, is a promising concept in this research area. However, one challenge is the detachment of cells from the wall of artificial grafts after bypass surgery because of a continuous shear stress generated from blood flow. (4) Spence, D. M.; Torrence, N. J.; Kovarik, M. L.; Martin, R. S. Analyst 2004, 129, 995–1000. (5) Young, E. W. K.; Wheeler, A. R.; Simmons, C. A. Lab Chip 2007, 7, 1759– 1766. (6) Barkefors, I.; Le Jan, S.; Jakobsson, L.; Hejll, E.; Carlson, G.; Johansson, H.; Jarvius, J.; Park, J. W.; Jeon, N. L.; Kreuger, J. J. Biol. Chem. 2008, 283, 13905–13912. (7) Takayama, S.; McDonald, J. C.; Ostuni, E.; Liang, M. N.; Kenis, P. J. A.; Ismagilov, R. F.; Whitesides, G. M. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 5545–5548. (8) Westcott, N. P.; Lamb, B. M.; Yousaf, M. N. Anal. Chem. 2009, 81, 3297– 3303. (9) Oblak, T. D.; Root, P.; Spence, D. M. Anal. Chem. 2006, 78, 3193–3197. (10) Ku, C. J.; Oblak, T. D.; Spence, D. M. Anal. Chem. 2008, 80, 7543–7548. (11) Young, E. W. K.; Simmons, C. A. Lab Chip 2010, 10, 143–160. (12) Lu, H.; Koo, L. Y.; Wang, W. C. M.; Lauffenburger, D. A.; Griffith, L. G.; Jensen, K. F. Anal. Chem. 2004, 76, 5257–5264. (13) Nerem, R. M.; Seliktar, D. Annu. Rev. Biomed. Eng. 2001, 3, 225–243. 10.1021/ac100107z  2010 American Chemical Society Published on Web 03/10/2010

In constructing a cell-coated synthetic graft, cell adhesion is the most important factor because endothelial cell monolayer needs to be protected against separation from the graft. For this reason, high adhesion of the ECs should be taken into account when fabricating an artificial blood vessel. To this aim, we integrated a nanostructured surface within a microfluidic channel to culture ECs with strong cell alignment and cell-cell contacts. Cellular responses, such as adhesion, migration, and proliferation are highly affected by materials chemistry and topography of the constructed cell seeding surface. In particular, surface nanotopography has been shown to exert influence over various cellular responses, such as adhesion, growth, proliferation, and differentiation in many cell types.14-27 Therefore, it would be possible to construct a highly cell-adhesive surface by tailoring biomaterials interface such as shape and dimension of the nanostructure. Another consideration for structured biomaterials interface is that materials of tissue-engineered vascular graft should provide physical and mechanical support to cell culture and be implantable for functioning as a tissue in the body. To meet these needs, a thermoplastic polymer of poly(lactic-co-glycolic acid) (PLGA) is used here and its nanostructures were fabricated by thermal imprinting. PLGA is widely used for biodegradable synthetic vascular graft because it is biocompatible and its mechanical properties are appropriate for cell growth.28 Ease of fabrication is also a merit of the material; the degradation time of the polymer can be easily controlled, and it can be made into a sheet-type graft.16,28-32 The objective of this study is to fabricate nanopatterned PLGA surfaces that are then enclosed within a microfluidic channel and to perform adhesion assays of the cultured ECs under large shear stresses.2,5 The effect of nanostructures on cell morphology and (14) Dalby, M. J.; Riehle, M. O.; Johnstone, H.; Affrossman, S.; Curtis, A. S. G. Biomaterials 2002, 23, 2945–2954. (15) Chung, T. W.; Liu, D. Z.; Wang, S. Y.; Wang, S. S. Biomaterials 2003, 24, 4655–4661. (16) Miller, D. C.; Thapa, A.; Haberstroh, K. M.; Webster, T. J. Biomaterials 2004, 25, 53–61. (17) Kim, P.; Kim, D. H.; Kim, B.; Choi, S. K.; Lee, S. H.; Khademhosseini, A.; Langer, R.; Suh, K. Y. Nanotechnology 2005, 16, 2420–2426. (18) Sniadecki, N.; Desai, R. A.; Ruiz, S. A.; Chen, C. S. Ann. Biomed. Eng. 2006, 34, 59–74. (19) Kim, D. H.; Kim, P.; Song, I.; Cha, J. M.; Lee, S. H.; Kim, B.; Suh, K. Y. Langmuir 2006, 22, 5419–5426. (20) Dalby, M. J.; Gadegaard, N.; Tare, R.; Andar, A.; Riehle, M. O.; Herzyk, P.; Wilkinson, C. D. W.; Oreffo, R. O. C. Nat. Mater. 2007, 6, 997–1003. (21) Kwon, K. W.; Choi, S. S.; Lee, S. H.; Kim, B.; Lee, S. N.; Park, M. C.; Kim, P.; Hwang, S. Y.; Suh, K. Y. Lab Chip 2007, 7, 1461–1468. (22) Bettinger, C. J.; Zhang, Z. T.; Gerecht, S.; Borenstein, J. T.; Langer, R. Adv. Mater. 2008, 20, 99-+. (23) Kim, D. H.; Seo, C. H.; Han, K.; Kwon, K. W.; Levchenko, A.; Suh, K. Y. Adv. Funct. Mater. 2009, 19, 1579–1586. (24) Kim, D. H.; Han, K.; Gupta, K.; Kwon, K. W.; Suh, K. Y.; Levchenko, A. Biomaterials 2009, 30, 5433–5444. (25) Suh, K. Y.; Park, M. C.; Kim, P. Adv. Funct. Mater. 2009, 19, 2699–2712. (26) Bettinger, C. J.; Langer, R.; Borenstein, J. T. Angew. Chem., Int. Ed. 2009, 48, 5406–5415. (27) Kulangara, K.; Leong, K. W. Soft Matter 2009, 5, 4072–4076. (28) He, W.; Ma, Z. W.; Yong, T.; Teo, W. E.; Ramakrishna, S. Biomaterials 2005, 26, 7606–7615. (29) Ma, Z. W.; He, W.; Yong, T.; Ramakrishna, S. Tissue Eng. 2005, 11, 1149– 1158. (30) He, W.; Yong, T.; Ma, Z. W.; Inai, R.; Teo, W. E.; Ramakrishna, S. Tissue Eng. 2006, 12, 2457–2466. (31) Badami, A. S.; Kreke, M. R.; Thompson, M. S.; Riffle, J. S.; Goldstein, A. S. Biomaterials 2006, 27, 596–606. (32) Hashi, C. K.; Zhu, Y. Q.; Yang, G. Y.; Young, W. L.; Hsiao, B. S.; Wang, K.; Chu, B.; Li, S. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 11915–11920.

expression of membrane cytoskeletal protein were systematically investigated to find an optimum structure for high cell adhesion. As described shortly, the fraction of elongated and aligned ECs with respect to the flow direction was dramatically increased on nanopatterned surfaces with enhanced adhesion forces, as supported by fluorescent images of actin filaments and vinculin (focal adhesion protein). It appears that the nanostructured surfaces provide in-vivo-like environments for ECs, which is supported by the fact that the cell adhesion can be controlled by changes in the topography of nanostructured surfaces even under identical chemical composition of the material. The findings of this in vitro study are expected to have implications in the development of artificial tissue-engineered blood vessel. MATERIALS AND METHOD Fabrication of Poly(Urethane acrylate) (PUA) Mold. The PUA mold was composed of a functionalized prepolymer with an acrylate group for cross-linking, a monomeric modulator, a photoinitiator and a radiation-curable releasing agent. The liquid precursor was drop-dispensed onto a prepared silicon master that was fabricated by photolithography, and then a transparent polyethylene terephthalate (PET) film with 50 µm thickness was placed on the liquid precursor surface. Subsequently, the mold was exposed to ultraviolet (UV) light for 25 s and peeled off from the patterned silicon master. Details on the mold preparation can be found elsewhere.21 Fabrication of Various PLGA Nanostructures. Scheme 1 shows a schematic procedure for fabricating a nanopatterned poly(dimethyl siloxane) (PDMS) microfluidic channel. Slide glass was washed with isopropyl alcohol (IPA) for 1 min in a Petri dish, cleaned by using distilled water in a water bath and dried in a stream of nitrogen. A 3% (w/v) solution of PLGA prepared in chloroform (100 µL) was drop-dispensed on the slide glass and a flat PDMS block was placed to remove solvent and obtain a smooth PLGA layer. Subsequently, the slide glass was preheated on a hot plate for additional solvent removal as well as for increasing adhesion between PLGA and glass slide. Then, a nanopatterned PUA mold was placed on the coated PLGA film and embossed into the molten PLGA by applying a constant pressure (∼ 0.7 bar) while heating in a vacuum oven at 100 °C for 20 min (glass transition temperature (Tg) of PLGA ) 59 °C). After the thermal imprinting process, the assembly (PUA mold + glass substrate) was cooled to room temperature, and the mold was peeled off the substrate. Irreversible Bonding with PDMS Microfluidic Channel. PDMS microfluidic molds (2 mm width and 60 µm height, 4 cm length) were fabricated by casting PDMS prepolymer (Sylgard 184 Silicon elastomer, Dow Corning) on etched silicon master prepared by photolithography and cross-linking at 70 °C for 1 h with 10 wt % of the curing agent. Then the PDMS channel replicas were carefully detached from the silicon master and cut prior to use. Holes (1.5 mm) were punched on the PDMS channel for inlet and outlet. For channel bonding, the PDMS channel mold and nanopatterned glass substrate were treated in oxygen plasma for 60 s (60W, PDC-32G, Harrick Scientific, Ossining, NY), and then the channel was carefully aligned on the glass substrate and firmly pressed for conformal sealing. Cell Culture. Prior to cell seeding, the microfluidic channel was coated with fibronectin by injecting 10 µg/mL solution of Analytical Chemistry, Vol. 82, No. 7, April 1, 2010

3017

Scheme 1. Schematic Diagram of the Fabrication Procedure for PDMS Microchannel with Nanopatterned Substratea

a Nanopatterns were generated by thermal imprinting, and the patterned substrate and PDMS microfluidic mold were completely sealed by oxygen plasma treatment.

fibronectin for 1 h prepared in phosphate buffered saline (PBS) and then rinsed 3 times with PBS to remove residual fibronectin. After this, 20 µL culture medium was injected into the microchannel and incubated for 30 min before injection of HUVECs and CPAEs. The two cell lines were purchased from the American Type Culture Collection (ATCC). HUVECs were cultured in endothelial basal medium (EBM-2, clonetics) containing endothelial growth factor supplement (EGM-2 bullet kit, cambrex) and 100 m/mL penicillin-streptomycin. CPAEs were cultured in RPMI 1640 medium (Gibco BRL, Grand Island, NY) containing 10% fetal bovine serum (FBS) and 100 m/mL penicillin-streptomycin. Cells were grown in a humidified 5% CO2 incubator at 37 °C. When the cells were confluent at about 70% of culture flask, cells were incubated with Hoechst 33342 nuclear dye (1 µg/mL in supplemented RPMI 1640 medium) for 40 min in the same condition. Cell Adhesion Assays. After nanopatterns were constructed on the bottom of the PDMS microfluidic channel, the cells labeled with molecular fluorescence dye were precultured for 3 days. Cells were stained with 10 µg/mL Hoechst 33342 and injected into a microfluidic channel using a micropipet. Fibronectin was coated for 1 h before endothelial cell seeding at a coating concentration of 10 µg/mL. Cell density was controlled at ∼15 000/cm2 to make a confluent cellular monolayer. After preculture, various shear stresses were applied through the microfluidic system, and the detachment of the cells was measured by using ImagePro software. Microfluidic channel was connected to a syringe pump, and cell culture medium flowed into the channel with flow rates of 48, 160, and 320 µL min-1 (12, 40, 80 dyn/cm2) for 40 min. The strength of cell adhesion was represented by the percentage of adhered cells before and after application of shear stress. The number of adhered cells was counted over at least three different locations on each sample, and the data were averaged over the values. Immunofluorescence Labeling. For F-actin immunofluorescence staining, cells were washed with PBS and fixed inside the channel with 3.7% paraformaldehyde in PBS for 15 min at room 3018

Analytical Chemistry, Vol. 82, No. 7, April 1, 2010

temperature. Then cells were permeabilized in 1% BSA and 0.3% Triton X-100 in PBS for 10 min, incubated with antibodies conjugated pholloidin-TRITC (Sigma, St Louis, MO, 50 µg/mL) for 1 h at room temperature, and washed with PBS 3 times. For vinculin immunoflurescence staining, cells were incubated with a monoclonal mouse antivinculin antibody (Sigma-Aldrich) for 1.5 h, washed with PBS 3 times, and then incubated with the FITCconjugated secondary antibody goat antimouse IgG (Jacson) followed by washing with PBS 3 times. For nucleus staining cells were incubated with DAPI (Sigma-Aldrich) for 2 min and washed with PBS 3 times. After counterstaining with DAPI (Sigma) for nuclei, the cells were washed with PBS. The PDMS channel was removed and the cells were covered with mounting medium (Invitrogen, Carlsbad, CA). To obtain images, cells were photographed with a conventional fluorescence microscope (Olympus, Inc., Melville, NY). Measurement of Cell Dimensions on PLGA Nanostructures. The size and shape of the adhered cell were measured from fluorescence image using image analysis program and the height was measured using z stack mode of microscope (Olympus, Inc., Melville, NY) after immunofluorescence labeling. Scanning Electron Microscopy. High-resolution scanning electron microscopy (SEM) images of the PLGA nanostructures were obtained using a HITACHI S-4800 microscope (Hitachi, Japan) operating at an accelerating voltage of 5 kV. To avoid charging effects, substrates were sputter-coated with Au to the thickness of 20 nm prior to measurements. RESULTS AND DISCUSSION Fabrication of Microfluidic Channel with PLGA Nanopatterned Substrate. For cell culture substrates, various nanoline patterns were fabricated by thermal imprinting process by utilizing thermoplastic characteristic of PLGA (Tg ) 59 °C). To generate uniform patterns, it is important to perform imprinting without air trap, which otherwise results in a reduced height or a nonuniform height distribution. To alleviate this problem, a roller was mildly rounded several times on the sheet-type PUA

Figure 1. SEM images of various PLGA nanopatterns used in the experiment: (a) 350 nm ridges/350 nm grooves, (b) 350 nm ridges/ 700 nm grooves, (c) 350 nm ridges/1050 nm grooves, (d) 350 nm ridges/700 nm grooves, (e) 350 nm ridges/700 nm grooves, (f) 350 nm ridges/700 nm grooves, (g) 350 nm ridges/700 nm grooves, and (h) flat control. Scale bar represents 2 µm. (i) A tilted view of the nanopatterned microfluidic channel (2 mm width and 60 µm height).

mold surface, which helped expel trapped air prior to imprinting process. It was found that no surface treatment was necessary because PLGA is more adhesive to the glass substrate than to the PUA mold. Figure 1a-h shows SEM images of various PLGA nanoline structures. The dimensions were as follows: (a) 350 nm wide ridges/350 nm grooves, (b) 350 nm ridges/700 nm grooves, (c) 350 nm ridges/1050 nm grooves, (d) 350 nm ridges/1750 nm grooves, (e) 700 nm ridges/350 nm grooves, (f) 1050 nm ridges/ 350 nm grooves, and (g) 1750 nm ridges/350 nm grooves. Figure 1h shows a PLGA flat surface. As shown in the figure, the fabricated PLGA nanostructures were well-defined with good edge definition. The height of PLGA nanostructure was measured to be ∼500 nm, corresponding to the original height of PUA mold. Shown in Figure 1I is a tilted view of the nanopatterned microfluidic channel of 2 mm width and 60 µm height where four different substrates were enclosed on the bottom of the channel in a consecutive manner. As shown in the figure, the nanostructures were neatly fabricated over a large area without many defects (see different diffraction colors in Figure 1i). Adhesion Assays of HUVECs and CPAEs on Various Nanopatterned Substrates. Microfluidic channels have been used for cell adhesion assays for various cell types.5,12,21,33-38 In particular, they offer a simple and efficient tool to measure adhesion strength of ECs5,38 by utilizing mechanical forces applied (33) Gutierrez, E.; Groisman, A. Anal. Chem. 2007, 79, 2249–2258. (34) Green, J. V.; Kniazeva, T.; Abedi, M.; Sokhey, D. S.; Taslim, M. E.; Murthy, S. K. Lab Chip 2009, 9, 677–685. (35) Lee, J. H.; Kaplan, J. B.; Lee, W. Y. Biomed. Microdevices 2008, 10, 489– 498. (36) Wang, H. Y.; Kim, J. H.; Zou, M.; Tung, S.; Kim, J. W. IEEE Trans. Nanotechnol. 2008, 7, 573–579. (37) Zhang, X.; Jones, P.; Haswell, S. J. Chem. Eng. J. 2008, 135, S82–S88. (38) Shao, J. B.; Wu, L.; Wu, J. Z.; Zheng, Y. H.; Zhao, H.; Jin, Q. H.; Zhao, J. L. Lab Chip 2009, 9, 3118–3125.

Figure 2. Fraction of adherent cells using (a) HUVECs and (b) CPAEs after application of shear stress for 40 min. Two different shear rates were used to resemble in vivo like conditions. The presented values were averaged over 20 repetitions and error bars represent the standard errors of the mean (standard deviation/n, where n ) 20).

on the cell surface. In this study, we applied different shear stresses for HUVECs and CPAEs to resemble different in vivo conditions. In detail, we applied an average wall shear stress of 12 dyn/cm2 and a maximum wall shear stress of 40 dyn/cm2 for HUVECs and 40 dyn/cm2 and 80 dyn/cm2 for CPAEs, respectively. After application of a wall shear stress for 40 min, we measured cell adhesion for 7 different nanopatterns and flat control (Figure 2). Statistical analysis of the HUVECs demonstrates that almost 95% of the original cells remained at 12 dyn/cm2 for all PLGA surfaces tested including control (flat PLGA surface) (Figure 2a). When the shear stress was increased to 40 dyn/cm2, adhesion behaviors were much changed depending on the pattern geometry. First, the fraction of adherent cells decreased dramatically on the nanopatterned surfaces of 1:1, 1:2, 1:3, and 1:5 (ridge to groove ratio) ( 90% of cells were attached on the surface of 700 nm ridges/350 nm grooves (2:1), which showed the best performance among the patterns tested. For the rest two surfaces of 3:1 and 5:1, the fraction of adherent cell was relatively high and comparable to that of the flat surface (90% under the shear stress level of 40 dyn/cm2. When the shear stress was increased to 80 dyn/ cm2, the number of adhered cells was much decreased with a notable peak at the pattern of 700 nm wide ridges/350 nm grooves. Furthermore, the minimum adhesion was found on the flat surface. In general, the strength of cell adhesion was higher for CPAEs regardless of the pattern type (>∼60%). (39) Feugier, P.; Black, R. A.; Hunt, J. A.; How, T. V. Biomaterials 2005, 26, 1457–1466. (40) Anamelechi, C. C.; Truskey, G. A.; Reichert, W. M. Biomaterials 2005, 26, 6887–6896.

Analytical Chemistry, Vol. 82, No. 7, April 1, 2010

3019

Figure 3. Angle of orientation (difference of angles between pattern direction and cell alignment) for HUVECs for different nanopatterns and flat control (standard deviation/n, where n ) 20).

These observations suggest that the cell-substrate adhesion can be promoted by incorporating a suitable surface texture in the form of line pattern on the cell culture substrate. To account for different adhesion strength, we consider the interplay between adhesion contact area and anisotropic alignment of the cells. For the nanopatterned surfaces of 1:1, 1:2, 1:3, and 1:5 (ridge to groove ratio), the adhesion areas would be smaller than those of the patterns with larger ridges (2:1, 3:1, and 5:1), as the initial contact area is dictated by the ridges (protruding areas), which is fixed at 350 nm width in the current study. The reduced ridges will result in a decreased population of the focal adhesion sites at cell-substrate interface. For larger ridges, the adhesion strength would increase with the increase of the pattern width, which in turn leads to reduced elongation or alignment along the line direction. If the cells are less aligned, their fluidic resistance would

increase accordingly. Figure 3 shows the angle of orientation for HUVECs, which can be defined as the difference of angles between pattern direction and cell alignment. As shown in the figure, the angle slightly increases with increasing the pattern width, and it becomes ∼45° for flat control. Therefore, a simple phenomenological conclusion can be made at this stage as to why the pattern of 700 nm ridges/350 nm grooves could be optimum for the highest adhesion strength. In the remaining sections, we elaborate on the observed adhesion assays in terms of a simple hydrodynamic model and fluorescent staining of actin filaments and focal adhesion proteins. Hydrodynamic Force Applied on the Surface of HUVECs. According to a previous study,41 the applied hydrodynamic force (FH) on cell surfaces can be given by the vector summation of hydrodynamic shear force and torque generated on cell surface and can be expressed as follows:

1 + ( ac ) = 3√b + c √c + a τ 1 + ( ac )

FH ) τavAsurf

2

2

2

2

2

2

wall

(1)

where τav is the average shear stress, Asurf is the cell surface area, τwall is the wall shear stress, and a, b, and c are the length, width, and height of the adhered cell, respectively. Figure 4a shows dimensional parameters for the adhered cell before application of shear stress. Here, we assumed the cell shape as a triangular pyramid for calculation simplicity. FH is also affected

Figure 4. (a) Schematic illustration for the geometry of the adhered cell. For calculation simplicity, the cell shape is assumed as a triangular pyramid. (b) Plots of the calculated hydrodynamic forces for various nanopatterns and flat control along with adhesion assay results. Brackets indicate statistical significance, P < 0.005. 3020

Analytical Chemistry, Vol. 82, No. 7, April 1, 2010

Table 1. Dimensional Parameters of HUVECs Cultured for 3 Days on Various Nanopatterns and Flat Control before Applying Shear Stress: (a) Height, (b) Width, and (c) Length. These Values were Averaged Over 20 Repetitions and Error Bars Represent the Standard Errors of the Mean (Standard Deviation/n, where n ) 20). The Corresponding Calculated Hydrodynamic Force is also Shown on the Bottom

a (µm) b (µm) c (µm) FH (N)

350 nm 1:1

350 nm 1:2

350 nm 1:3

350 nm 1:5

350 nm 2:1

350 nm 3:1

350 nm 5:1

nll

83.7 ± 10.3 2.31 ± 0.20 7.12 ± 0.43 7.5713 × 10-9

88.9 ± 7.45 2.11 ± 0.23 7.78 ± 0.33 8.64 × 10-9

95.3 ± 18.1 2.15 ± 0.18 6.12 ± 0.30 7.44 × 10-9

97.1 ± 10.0 2.27 ± 0.19 6.02 ± 0.47 7.52 × 10-9

76.8 ± 8.50 2.51 ± 0.19 3.58 ± 0.26 4.04 × 10-9

79.3 ± 5.75 2.57 ± 0.23 4.6 ± 0.39 5.03 × 10-9

78.3 ± 7.40 2.8 ± 0.25 4.94 ± 0.42 5.39 × 10-9

69.6 ± 8.95 3.74 ± 0.45 2.6 ± 0.27 3.81 × 10-9

by microchannel shape factor (γ) that is defined as the ratio of cell height to channel height, especially in the case of small γ (90% of HUVECs were remained (45) Zamir, E.; Katz, B. Z.; Aota, S.; Yamada, K. M.; Geiger, B.; Kam, Z. J. Cell Sci. 1999, 112, 1655–1669.

Analytical Chemistry, Vol. 82, No. 7, April 1, 2010

3021

Figure 6. Fluorescence images of vinculin stains of HUVECs on (a) nanopatterned and (b) flat surfaces.

on the surface of 350 nm 2:1 ridges/grooves after applying a shear stress of 40 dyn/cm2. For comparison, the retention of HUVECs in some of previous works was below 50% after applying a shear stress of 20-30 dyn/cm2.39,40 Also, our results demonstrate that control of geometrical shape such as width, height, and spacing exerts a dramatic effect on the adhesion of ECs even using the same material composition, which has not been explored systematically with earlier studies. For example, several researches reported that surface topography could enhance adhesion of ECs without a detailed study on the effect of shear stress.14-16 Also, some researchers found that aligned nanofibers, when the cells were grown on the surface, induced an enhanced phenotype of ECs, expressing a high level of ICAM1 or PECAM1. However, a systematic study with various geometries and shear stress levels was not reported. CONCLUSIONS We have presented adhesion assays of human umbilical vein endothelial cells (HUVECs) and calf pulmonary artery endothelial cells (CPAEs) on various nanopatterned surfaces within a microfluidic channel. By controlling fluidic shear stress applied on the cell surface, the number of adhered cells was counted after medium flow for 40 min under different flow rates. To emulate different in vivo conditions, different shear stresses were generated for HUVECs and CPAEs: 12 (average wall shear stress) and 40 dyn/cm2 (maximum wall shear stress) for HUVECs and 40 (average wall shear stress) and 80 dyn/cm2 (maximum wall shear stress) for CPAEs. It has been demonstrated that the cell-substrate adhesion can be significantly enhanced by incorporating a suitable surface texture in the form of line pattern on the cell culture substrate.

3022

Analytical Chemistry, Vol. 82, No. 7, April 1, 2010

For the conditions used in our experiments, the nanopattern of 700 nm ridges/350 nm grooves exhibited the highest adhesion strength, that is, >90% of cells remained on the surface under the higher shear stress level for both cell types. To account for cell adhesion behaviors, we have considered the interplay between mechanical and biological contributions: fluidic shear stress and cell focal adhesion. On the basis of the calculation of hydrodynamic forces, the HUVECs cultured on the nanopattern of 700 nm ridges/350 nm grooves showed the lowest among the patterns tested, suggesting that mechanical stress can be minimized by elongated cell alignment along the groove direction. Also, the HUVECs grown on the nanopattern showed clustered and aligned vinculin stains parallel to the line direction with higher fluorescence intensity and density than those of the cells on flat control. Therefore, both mechanical and biological processes should be synergistically matched up to induce strong cell adhesion and cell-cell contacts under in vivo like environments. These findings would be useful toward the development of artificial tissueengineered blood vessel. ACKNOWLEDGMENT This work was supported by the alliance program between Engineering and Medical Schools of Seoul National University, World Class University program (R31-2008-000-10083-0), Engineering Reserach Institute, and Korea Research Foundation Grant (MOEHRD, KRFJ03003). This work was supported in part by Seoul National University Hospital and Boramae Hospital. Received for review January 14, 2010. Accepted February 24, 2010. AC100107Z