pubs.acs.org/Langmuir © 2010 American Chemical Society
Characterization of Collagen Fibrils Films Formed on Polydimethylsiloxane Surfaces for Microfluidic Applications Tighe A. Spurlin,* Samuel P. Forry, Gregory A. Cooksey, and Anne L. Plant Biochemical Science Division, National Institute of Standards and Technology, 100 Bureau Drive, Gaithersburg, Maryland 20899 Received May 26, 2010. Revised Manuscript Received July 18, 2010 Type I collagen fibrillar thin films have been prepared on hydrophobic recovered poly(dimethylsiloxane) (PDMS) surfaces and inside of irreversibly sealed PDMS microfluidic devices. Fibrillar films prepared on PDMS surfaces have been characterized with optical microscopy and atomic force microscopy and compared with films prepared using more traditional bulk methods on thiol-coated gold substrates. Collagen fibril films formed after 18 h of incubation on PDMS surfaces were observed to have similar underlying film thicknesses (15 nm), fibril size (67 nm), fibril coverage (45%), and physiologically supermolecular structure when compared to films on gold substrates. Collagen fibrils formed within devices were also determined to be usable across physiologically relevant cell perfusion rates. To validate the utility of these collagen fibril thin films for cell culture applications, vascular smooth muscle cells are shown to attach to collagen fibrils and exhibit cell spread areas equivalent to those seen on collagen fibrils created via bulk cell culture methods on thiol-coated gold substrates. These results extend the use and benefits of collagen fibril thin films into microfluidic-based cellular studies.
Introduction Microfluidic devices are increasingly being used as alternatives to cell culture dishes for cell assays because they can be used to control the solution microenvironments around cells,1-4 they can be interfaced with multiple complementary analysis techniques,5-9 and they offer the potential for data to be collected in a high throughput fashion.10-12 In both cell culture dishes and microfluidic systems it is challenging to provide cells with relevant extracellular matrices (ECM) that elicit reproducible and physiologically relevant cell behaviors.13 Microfluidic devices for cell experiments have been coated with fibronectin,14 collagen monomer,15 gelatin,13 and Matrigel.16 With the exception of fibronectin, these ECM materials are either not found in in vivo ECM (collagen monomer or gelatin) or are an ill-defined mixture of ECM proteins and growth factors secreted by mouse tumor *Corresponding author: phone (301) 975-5088, e-mail
[email protected]. (1) Squires, T. M.; Quake, S. R. Rev. Mod. Phys. 2005, 77, 977–1026. (2) Kim, L.; Vahey, M. D.; Lee, H. Y.; Voldman, J. Lab Chip 2006, 6, 394–406. (3) Kim, L.; Toh, Y. C.; Voldman, J.; Yu, H. Lab Chip 2007, 7, 681–694. (4) Sia, S. K.; Whitesides, G. M. Electrophoresis 2003, 24, 3563–3576. (5) Vilkner, T.; Janasek, D.; Manz, A. Anal. Chem. 2004, 76, 3373–3385. (6) El-Ali, J.; Sorger, P. K.; Jensen, K. F. Nature 2006, 442, 403–411. (7) Reich, C.; Hochrein, M. B.; Krause, B.; Nickel, B. Rev. Sci. Instrum. 2005, 76. (8) Cen, E. G.; Dalton, C.; Li, Y. L.; Adamia, S.; Pilarski, L. M.; Kaler, K. J. Microbiol. Methods 2004, 58, 387–401. (9) Thompson, D. M.; King, K. R.; Wieder, K. J.; Toner, M.; Yarmush, M. L.; Jayaraman, A. Anal. Chem. 2004, 76, 4098–4103. (10) Gomez-Sjoberg, R.; Leyrat, A. A.; Pirone, D. M.; Chen, C. S.; Quake, S. R. Anal. Chem. 2007, 79, 8557–8563. (11) Melin, J.; Quake, S. R. Annu. Rev. Biophys. Biomol. Struct. 2007, 36, 213–231. (12) Thorsen, T.; Maerkl, S. J.; Quake, S. R. Science 2002, 298, 580–584. (13) Paguirigan, A.; Beebe, D. J. Lab Chip 2006, 6, 407–413. (14) Lu, H.; Koo, L. Y.; Wang, W. C. M.; Lauffenburger, D. A.; Griffith, L. G.; Jensen, K. F. Anal. Chem. 2004, 76, 5257–5264. (15) Young, E. W. K.; Wheeler, A. R.; Simmons, C. A. Lab Chip 2007, 7, 1759– 1766. (16) Chaw, K. C.; Manimaran, M.; Tay, F. E. H.; Swaminathan, S. Biomed. Microdevices 2007, 9, 597–602. (17) Koster, S.; Leach, J. B.; Struth, B.; Pfohl, T.; Wong, J. Y. Langmuir 2007, 23, 357–359. (18) Lee, P.; Lin, R.; Moon, J.; Lee, L. P. Biomed. Microdevices 2006, 8, 35–41. (19) Lanfer, B.; Freudenberg, U.; Zimmermann, R.; Stamov, D.; Korber, V.; Werner, C. Biomaterials 2008, 29, 3888–3895.
Langmuir 2010, 26(17), 14111–14117
cells (matrigel). Several groups17-19 have reported on the ability to generate collagen Type I fibrils in microfluidic channels under conditions of flow. However, these previous publications have focused on conditions that affect the alignment of fibrils not attached to a surface, and so they do not examine the ability of collagen films to form on the poly(dimethylsiloxane) (PDMS) surface, provide quantitative data regarding fibril size, or quantitatively compare cell response to fibrils formed in the device and fibrils formed using bulk culture methods. Type I collagen fibrils are prevalent in a variety of tissues,20 have been widely utilized in in vitro model systems,20 and allow attachment of many different cell lines. For these reasons we have previously21-23 developed collagen fibrillar thin films on traditional cell culture dishes that provoke cell responses similar to those seen in 3D collagen gels but provide unique advantages over gels in terms of handling, imaging with optical microscopy, and characterization with surface analysis techniques. Here, we provide salient details concerning the formation of collagen fibrillar films in PDMS microfluidic devices that are similar to collagen fibrillar thin films we have formed in traditional culture plates in fibril size, collagen film thickness, the presence of native collagen supermolecular structure, and the ability of the fibril films to invoke reproducible cell responses. In addition, we provide data that shows collagen fibril films are removed from the PDMS substrate at shear stresses exceeding 200 dyn/cm2 to demonstrate the acceptable fluid flow conditions3,15,24-26 that (20) Nimni, M. Collagen; CRC Press: Boca Raton, FL, 1988; Vol. I. (21) Elliott, J. T.; Tona, A.; Woodward, J. T.; Jones, P. L.; Plant, A. L. Langmuir 2003, 19, 1506–1514. (22) McDaniel, D. P.; Shaw, G. A.; Elliott, J. T.; Bhadriraju, K.; Meuse, C.; Chung, K. H.; Plant, A. L. Biophys. J. 2007, 92, 1759–1769. (23) Elliott, J. T.; Halter, M.; Plant, A. L.; Woodward, J. T.; Tona, A. Biointerphases 2008, 3, 19–28. (24) Wagner, C. T.; Durante, W.; Christodoulides, N.; Hellums, J. D.; Schafer, A. I. J. Clin. Invest. 1997, 100, 589–596. (25) Lee, A. A.; Graham, D. A.; Dela Cruz, S.; Ratcliffe, A.; Karlon, W. J. J. Biomech. Eng. 2002, 124, 37–43. (26) Papadaki, M.; McIntire, L. V.; Eskin, S. G. Biotechnol. Bioeng. 1996, 50, 555–561.
Published on Web 07/28/2010
DOI: 10.1021/la102150s
14111
Article
Spurlin et al.
Figure 1. (a) Microfluidic devices used for cell culture experiments. (b) Microfluidic devices used to determine the wall shear stress that would remove collagen films.
can be used during rinsing protocols and cell studies on microfluidic devices. The studies detailed in this report show that collagen Type I fibril films can be formed on PDMS surfaces and validate that the fibril films can be used in microfluidic devices for cell culture applications.
Experimental Section Materials. PDMS (Sylgard 184, Dow Corning, Midland, MI) prepolymer, silcon wafers (Monto Silicon Technology, Spring City, PA), and photoresist (SU-8 2050, Microchem, Newton, MA, or AZ9260 photoresist, Mays Chemical Co., Indianapolis, IN) were used in the fabrication of PDMS devices. Acid-solubilized purified collagen from bovine skin (Purecol, Nutacon BV, Netherlands), Dubelco’s phosphate buffered saline (DPBS) (Sigma, St. Louis, MO), and sodium hydroxide (Sigma) were used in the formation of collagen fibrils. Rat aortic vascular smooth muscle cells (A10 vSMCs), line A10 (American Type Culture Collection, Manassas, VA), were cultured in Dulbecco’s Modified Eagles Medium (DMEM; Mediatech, Herndon, VA), glutamine, nonessential amino acids, penicillin (100 U/ml), streptomycin (100 mg/mL), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer (25 mM), and 10% (v/v) fetal bovine serum (FBS, Invitrogen, Carlsbad, CA). For fluorescence imaging of cells, Texas Red-C2-Maleimide (Sigma) was used to label the cell body and DAPI (Sigma) was used to label the cell nuclei. Fluorescent images of collagen fibrils were obtained by labeling fibrils with rabbit antibovine collagen Type I primary antibody (Millipore, Temecula, CA) and goat antirabbit Alexa Fluor 488 (Invitrogen) secondary antibody. Fabrication of PDMS Devices. Masters to create channels with the desired dimensions for cell culture devices (Figure 1A) were made by spin-coating SU-8 (Microchem) onto a silicon wafer (100 rpm/s acceleration to 2000 rpm for 60 s) for a thickness of 50 μm.27-29 SU-8-coated wafers were then soft baked (first step: 6 min, 65 °C; second step: 20 min, 95 °C), exposed, postbaked (first step: 1 min, 65 °C; second step: 5 min, 95 °C), and developed (Microchem). Masters for the smaller devices (Figure 1B) were made by spin-coating two coats of AZ9260 photoresist (Mays Chemical Co., Indianapolis, IN) onto a silicon wafer (100 rpm/s acceleration to1250 rpm for 60 s) for a final thickness of 30 μm. The photoresist was then exposed to UV light (2700 mJ/cm2, MA-6, Karl Suss, Palo Alto, CA) through a transparency mask (CAD Art Services, Inc., Bandon, OR) and developed (Mays Chemical Co.). Glass microscope slides that served as the base of the device were spin-coated with degassed (10:1) poly(dimethylsiloxane) (PDMS, Sylgard 184, Dow Corning) and cured in a 60 °C oven for at least 4 h. Microfluidic channels were created by curing (10:1) PDMS on a master with the desired dimensions for 4 h in a 60 °C oven. PDMS channels and PDMS coated slides were irreversibly (27) McDonald, J. C.; Duffy, D. C.; Anderson, J. R.; Chiu, D. T.; Wu, H. K.; Schueller, O. J. A.; Whitesides, G. M. Electrophoresis 2000, 21, 27–40. (28) Becker, H.; Gartner, C. Electrophoresis 2000, 21, 12–26. (29) Love, J. C.; Wolfe, D. B.; Jacobs, H. O.; Whitesides, G. M. Langmuir 2001, 17, 6005–6012.
14112 DOI: 10.1021/la102150s
bonded through exposure to air plasma.30,31 PDMS pieces were placed into a desiccator jar, with a steel plate bottom, and pumped down for 60 s with a Maxima Cplus vacuum pump (Fischer Scientific, Waltham, MA). Air plasma was generated inside of the jar for 5 s through the use of a Panasonic 1100 W (Panasonic, Secaucus, NJ) microwave oven operating at 30% of its duty cycle. PDMS pieces were then placed into contact to form a irreversible bond that has been reported to withstand up to 50 psi.30,32 Hydrophobic recovery of the surfaces was ensured by exposing bonded devices to ambient conditions for at least 4 days.27 Microfluidic devices were sterilized with ethanol and rinsed extensively with sterile water before being filled with collagen solutions or a suspension of cells. Chambers used for fibril characterization and cell culture studies had dimensions as indicated in Figure 1a, while chambers used for shear flow experiments had dimensions as indicated in Figure 1b. The larger chambers were used for cell culture because the surface area is similar to the dimensions of a well within a 96 well plate. The large chambers also allowed us to cut open devices and examine the collagen fibrils and cells formed within the device. The height (h) of each chamber listed in Figure 1 was confirmed by optical profilometry (NT9000,Veeco, Santa Barbara, CA), while the length and width were confirmed through the used of an optical microscope (IX81, Olympus, Melville, NY) that had been calibrated through the use of a standard grid (Ted Pella, Redding, CA). Collagen Fibril Formation. Collagen Type I fibrils were prepared on 1-hexadecanethiol-coated gold surfaces and PDMS slabs as previously described.21,33 A similar process was utilized to create collagen fibrils within a microfluidic device from neutralized collagen solutions. Briefly, an acid-stabilized collagen Type I monomer solution (Purecol, Nutacon BV, Netherlands) with a concentration of 3.1 mg/mL was diluted to 300 μg/mL using DPBS and neutralized through the addition of 1.25% (v/v) 0.1 M sodium hydroxide. The neutralized collagen solution was drawn into a 1 mL glass syringe that was then attached to the inlet of the microfluidic device and mounted on a syringe pump (74900, Cole Palmer, Vernon Hills, IL) to aid in hands-free delivery of solution to the microfluidic device. This allowed the user to freely examine the device during filling through the use of an inverted microscope (IX 71, Olympus, Melville, NY). Microfluidic devices filled with neutralized collagen solution were placed into an incubator at 37 °C for 16 h to allow collagen fibrils to form on the surface. For cell experiments that required sterility, we prepared collagen solutions, filled the syringe, and attached fluid inlets/outlets to the device under aseptic conditions in a class 2 biosafety lab hood. Following incubation, samples were rinsed with DPBS below fluid flow rates that removed collagen films from the surface of the device and stored for less than 1 week before use. Atomic Force Microscopy. Collagen fibrils formed on 1-hexadecanethiol-coated gold surfaces,21,22,34-37 on masked PDMS slabs, and in PDMS microfluidic devices were imaged with an atomic force microscope (AFM) to determine if significant differences in fibril morphology existed. To determine that collagen films grow on a PDMS surface, we masked half of a PDMS slab during collagen film growth so a clear demarcation line existed between bare PDMS and collagen film growth could be visualized through the use of an optical microscope. Images of fibrils formed in devices were obtained after carefully opening a (30) Hui, A. Y. N.; Wang, G.; Lin, B. C.; Chan, W. T. Lab Chip 2005, 5, 1173–1177. (31) McDonald, J. C.; Whitesides, G. M. Acc. Chem. Res. 2002, 35, 491–499. (32) Chaudhury, M. K.; Whitesides, G. M. Langmuir 1991, 7, 1013–1025. (33) Spurlin, T. A.; Bhadriraju, K.; Chung, K. H.; Tona, A.; Plant, A. Biomaterials 2009, 30, 5486–5496. (34) Elliott, J. T.; Woodward, J. T.; Umarji, A.; Mei, Y.; Tona, A. Biomaterials 2007, 28, 576–585. (35) Amyot, F.; Small, A.; Boukari, H.; Plant, A.; McDaniel, D.; Gandjbakhche, A. Biophys. J. 2007, 484A–484A. (36) Elliott, J. T.; Woodward, J. T.; Langenbach, K. J.; Tona, A.; Jones, P. L.; Plant, A. L. Matrix Biol. 2005, 24, 489–502. (37) Elliott, J. T.; Tona, A.; Jones, P. L.; Plant, A. L. Mol. Biol. Cell 2002, 13, 68A–68A.
Langmuir 2010, 26(17), 14111–14117
Spurlin et al.
Article
dried device by cutting three sides off the device with a razor blade and peeling the remaining PDMS away from the main chamber. Fibrils were imaged with a Bioscope II AFM (Veeco, Santa Barbara, CA) that was attached to a Zeiss 100 TV microscope (Carl Zeiss, Munich, Germany). DNP-20 tips (Veeco) with spring constants of ∼0.05 N/m in contact mode in air were used for data collection. Images were collected at scan rates of 1 Hz at a resolution of 512 512 pixels. AFM images were flattened and plane fit before image analysis. High-resolution images were captured at scan rates of 0.8 Hz at a resolution of 2048 2048 pixels to better visualize banding on collagen fibrils.38 Collagen fibril heights and collagen film thicknesses were determined through cross-section analysis of 100 individual fibrils on three different samples using Nanoscope V software (Veeco). For calculation of the individual fibril heights we selected only fibrils lying on the surface and not on top of one another for height measurements. Additionally, for determination of individual fibril heights we did not use areas near regions of the sample that had been masked to form regions without fibril formation. Film thicknesses were calculated from height profiles taken across a region of the sample that had been masked to form regions with and without fibril film formation. At certain instances in these height profiles we observed that a large spike in height occurred. For the purposes of data analysis these spikes in height were ignored because they are an artifact from the masking and peeling procedure that was used to create regions devoid of collagen fibrils. The percentage of the surface covered by the collagen fibrils was determined through the use of manual thresholding in WxSM software (Nanotec Electronica, Madrid, Spain) of three randomly selected 50 μm 50 μm image areas.
Calculation of the Fluid Flow Induced Shear Stress That Removes Fibrils. To determine the wall shear stress that removed fibrils from the bottom of microfluidic devices, we positioned the microscope objective in the middle of the device shown in Figure 1B and collected phase microscopy movies over 5 min (1 frame/2 s) while generating the following flow rates: 0.5, 1, 5, 10, 25, 50, 75, 100, and 200 μL/min. An IX 71 (Olympus) microscope with a Photometrics (Tuscon, AZ) 512 EMCCD camera was used to capture each image frame. To calculate the wall shear stress above which sections of the fibril thin film were removed in the device, we used the Purday approximation (eq 1), where τw is the wall shear stress, Q is the volumetric flow rate, m and n are constants related to the channel aspect ratios (m = 1.7 þ 0.5(h/w), n = 2), μ is the viscosity of the solution (0.001 kg s-1 cm-1), w is the width of the channel (995 μm), and h is the height of the channel (18 μm).15,39 Equation 1 was selected to facilitate the conversion of applied flow rates into values of shear stress, which can be related to physiological processes such as interstitial flow and stress-induced protein expression.40,41 Here, we use this relation to determine if collagen films formed within a microfluidic device can withstand physiologically relevant shear stresses. A more complete description of the origins of eq 1 with a full summary of wall shear stress values can be found in the Supporting Information. τ ¼
2μQ m þ 1 ðn þ 1Þ wh2 m
ð1Þ
Cell Culture. Cells from culture flasks were centrifuged at 100 G for 5 min to obtain a concentrated cell suspension (300 000 cells/ mL) that was confirmed through the use of a cell multisizer (Multisizer 3, Beckman Coulter, Fullerton, CA). Cell suspensions (38) Kadler, K. E.; Holmes, D. F.; Trotter, J. A.; Chapman, J. A. Biochem. J. 1996, 316, 1–11. (39) Shah, R.; London, A. Laminar Flow Forced Convection in Ducts; Academic Press: New York, 1978; pp 296-202. (40) Tada, S.; Tarbell, J. M. Am. J. Physiol. 2000, 278, H1589–H1597. (41) Alshihabi, S. N.; Chang, Y. S.; Frangos, J. A.; Tarbell, J. M. Biochem. Biophys. Res. Commun. 1996, 224, 808–814.
Langmuir 2010, 26(17), 14111–14117
Figure 2. (a) Bare PDMS surface that was not coated with a collagen film, (b) height profile on bare PDMS surface, (c, e, g) collagen was allowed to adsorbed only on right half of the PDMS surface for the indicated time, (d, f, h) sharp jump in height profiles indicates the demarcation region between bare PDMS and physisorbed collagen film, (i) collagen adsorbed onto right half of selfassembled monolayer-coated gold substrate for 18 h, and (j) height profile illustrating thickness of collagen film formed on gold substrate. Scale bar represents 6 μm. Images were taken in contact mode in air. Black lines indicate region of surface where height profiles are taken from. were introduced to microfluidic devices via syringe, incubated for 24 h, fixed with 4% paraformaldehyde, and permeabilized with 0.5% Triton X-100 in PBS. Cells were plated onto collagen fibrils adhered to gold substrates by diluting the concentrated cell suspension (5300 cells/mL) in order to match the approximate cell density (1100 cells/cm2) on the microfluidic device. Incubation, fixing, and permeabilization of cells on gold surfaces were performed in a manner identical to that performed in microfluidic devices. Cells were stained for cell morphology experiments42 with Texas Red-C2-Maleimide (Tx-Red; 75 μg/mL; Sigma) to label the cell body and 1 μg/mL DAPI (Sigma). After rinsing dye from the microfluidic device or the culture dish using phosphate buffer, we imaged each device on an automated microscope to collect images for cell morphology analysis.42 Optical Microscopy. An inverted microscope (Olympus IX71) with Ludl (Hawthorne, NY) automated shutter, x,y,z stage, and excitation/emission filter wheels described previously42 (42) Elliott, J. T.; Tona, A.; Plant, A. L. Cytometry, Part A 2003, 52A, 90–100.
DOI: 10.1021/la102150s
14113
Article
Spurlin et al.
Table 1. Summary of AFM Data Collected from Collagen Fibril Films Formed on PDMS- and SAM-Coated Au Surfacesa sample
film thickness (nm)
fibril height (nm)
fibril coverage (%)
bare PDMS N/A N/A N/A PDMS (10 min) 5.2 ( 0.2 N/A N/A PDMS (1 h) 9.6 ( 1.2 17.2 ( 1.1
