BioMEMs Materials and Fabrication Technology - Semantic Scholar
Biomedical Microdevices 5:1, 11±19, 2003 # 2003 Kluwer Academic Publishers. Manufactured in The Netherlands.
BioMEMs Materials and Fabrication Technology: Section Editors Dr Shuvo Roy and Dr Mak Paranjape
Control of Mammalian Cell and Bacteria Adhesion on Substrates Micropatterned with Poly(ethylene glycol) Hydrogels Won-Gun Koh,1 Alexander Revzin,2 Aleksandr Simonian,3 Tony Reeves,3 and Michael Pishko1*
1 Department of Chemical Engineering and Department of Materials Science and Engineering, The Pennsylvania University, University Park, PA 16802-4420 E-mail: [email protected] 2 Department of Chemical Engineering, Texas A&M University, College Station, TX 77843-3122 3 Department of Biochemistry and Biophysics, Texas A&M University, College Station, TX 77843-3122
Abstract. A simple method for controlling the spatial positioning of mammalian cells and bacteria on substrates using patterned poly(ethylene glycol) (PEG) hydrogel microstructures is described. These microstructures were fabricated using photolithography on silicon, glass or poly (dimethylsiloxane) (PDMS) surfaces modi®ed with a 3-(trichlorosilyl) propyl methacrylate (TPM) monolayer. During the photogelation reaction, the resulting hydrogel microstructures were covalently bound to the substrate via the TPM monolayer and did not detached from the substrate upon hydration. For mammalian cell patterning, microwell arrays of different dimensions were fabricated. These microwells were composed of hydrophilic PEG hydrogel walls surrounding hydrophobic TPM ¯oors inside the microwells. Murine 3T3 ®broblasts and transformed hepatocytes were shown to selectively adhere to the TPM monolayer inside the microwells, maintaining their viability, while adherent cells were not present on the hydrogel walls. The number of cells inside one microwell could be controled by changing the lateral dimension of the microwells, thus allowing only a single cell per microwell if desired. In the case of 30 6 30 lm microwells, as many as 400 microwells were fabricated in 1 mm2. In addition, PEG hydrogel microstructures were also shown to effectively resist the adhesion of bacteria such as Escherichia coli. Key Words. poly(ethylene glycol) hydrogel microstructures, photolithography, cell patterning, microwells
1.
Introduction
Many mammalian cell-based bioassays for the screening of large libraries of potential pharmaceutical agents or the detection of toxic compounds and pathogens in the environment have been developed over the years and some are in use today. Although most of cell-based assays are still performed in 96 well plates, many researches are moving toward 384 well and higher density plate formats
to improve the performance and throughput of these assays. With further assay miniaturization in mind, a number of researchers have explored the patterned deposition of cells in microsystems through the control of cellular adhesion (Bousse, 1996; Makohliso et al., 1999; Sundberg, 2000; Kane et al., 1999; Jung et al., 2001; Chen et al., 1997, 1998; Lahiri et al., 1999; Lopez et al., 1993; Mrksich et al., 1997; Singhvi et al., 1994; Takayama et al., 1999; Whitesides et al., 1994; Zhang et al., 1999; Turner et al., 1999; John et al., 1998; Bhatia et al., 1997, 1998; Folch et al., 1998; den Braber and Toner, 1996). Similarly, the control of bacterial adhesion may be a valuable tool toward developing cell-based sensor arrays based on genetically engineered bacteria. Early studies of bacterial adhesion have mostly focused on reducing bacterial adhesion on implanted biomaterials, the critical event in the pathogenesis of foreign body infections (An and Friedman, 1998; Park et al., 1998). However, because analyte speci®city can be readily modi®ed by genetic engineering and because of the relatively robust nature of these micro-organisms as compared to mammalian cells, bacterial cells have been studied extensively for sensing applications (Rainina et al., 1996a, 1996b; Larsen et al., 1997; Belkin et al., 1997; Liao et al, 2001). The ability to selectively attach bacteria to micropatterned substrates could be used to create sensor arrays for applications such as rapid screening for infectious diseases or to detect toxic compounds (Cowan et al., 2001; Ostuni et al., 2001a).
*Corresponding author: Department of Chemical Engineering, The Pennsylvania State University, 104 Fenske Laboratory, University Park, PA 16802-4420.
11
12
Koh et al.
A number of studies reported that cellular adhesion is controled by the chemical and physical characteristics of surfaces such as hydrophobicity and hydrophilicity, surface charge, and surface roughness (Jung et al., 2001; Folch and Toner, 2000; Ito, 1999). These factors can be readily manipulated to control the spatial distribution of cells on a substrate. The rapid development of microfabrication and surface engineering techniques have stimulated the development of novel methods that may be used to control cell adhesion and spreading with micrometer-scale resolution. The most common approaches to de®ning and controlling regions that promote or resist the adhesion of cells on this length scale is to control the surface tension and wetting properties by patterning hydrophilic and hydrophobic regions on the surface using photolithography (Matsuda and Sugawara, 1995; Thomas et al., 1997; Dewez et al., 1998) or microcontact printing ( m-CP) patterns of selfassembly monolayers (SAM) on the surfaces (Kane et al., 1999; Chen et al., 1998; Amirpour et al., 2001; Ghosh et al., 1999; Franko et al., 2000; Mrksich et al., 1996; Singhvi et al., 1994a; Whitesides et al., 2001; Matsuda and Sugawara, 1995). Here we controlled cell adhesion using hydrogels of poly (ethylene glycol) (PEG), a non-biodegradable and hydrophilic polymer that has been previously used to minimize protein adsorption and mammalian cell adhesion onto biomaterial surfaces (Zhang et al., 1998; Tziampazis et al., 2000; Lu et al., 2001). PEG grafting to surfaces has been also used to render a variety of surfaces resistant to bacterial adhesion (Park et al., 1998; Ostuni et al., 2001b; Razatos et al., 2000; Cunliffe et al., 1999). The groups of Ingber and Whitesides demonstrated that SAMs containing hydrophilic ethylene glycol oligomers as the terminal segment resist cell adhesion while hydrophobic SAMs promote cell adhesion. These regions can be con®ned to 10±100 mm wide areas by patterning these SAMs using m-CP on a gold surface (Singhvi et al., 1994b). In this study we described the control of cell adhesion and the spatial con®nement of cell proliferation by means of micrometer scale PEG hydrogel structures on silicon, glass or poly (dimethylsiloxane) surfaces modi®ed with a chlorosilane coupling agent. Acrylate or methacrylate derivatives of PEG can be cross-linked into hydrogels and photolithographically patterned on silicon, glass, and plastic surfaces (Sirkar and Pishko, 1998; Ward et al., 2001; Revzin et al., 2001). Using these patterned substrates, we demonstrated the ability to capture and con®ne mammalian cells (®broblasts and hepatocytes) inside arrays of microwells fabricated from hydrogels of varying dimensions. In addition, we demonstrated the partial resistance of PEG hydrogel microstructures to the adhesion of a model bacterium, Escherichia coli.
2. Materials and Methods 2.1. Materials Poly(ethylene glycol) diacrylate (PEG-DA, MW 575), 2,20 -dimethoxy-2-phenyl-acetophenone (DMPA), anhydrous carbon tetrachloride, n-heptane and per¯uorooctane were purchased from Aldrich Chemical Co. (Milwaukee, WI). Dulbecco's modi®ed Eagle's medium (DMEM), fetal bovine serum (FBS), antibiotic/antimycotic solution, trypsin, ethylenediaminetetra-acetate (EDTA), sodium chloride, sodium phosphate and potassium phosphate monobasic were purchased from Sigma Chemical Co. (St. Louis, MO). 3-(trichlorosilyl) propyl methacrylate (TPM) was purchased from Fluka Chemicals (Milwaukee, WI). Hydrogen peroxide was purchased from EM Science (Gibbstown, NJ). Sulfuric acid was purchased from Fisher Scienti®c (Fair Lawn, NJ). E. coli bacteria (DH5) were obtained from Novagen Inc. (Madison, WI). Bacto-tryptone and yeast extract were obtained from Becton & Dickinson (Franklin Lakes, NJ). Gram Safranin stain was purchased from Fisher Diagnostics (Pittsburgh, PA). Murine 3T3 ®broblasts and murine hepatocytes (SV40 transformed) were obtained from American Type Culture Collection (Manssas, VA). Live/Dead Viability/Cytotoxicity Kit (L-7013) was obtained from Molecular Probes (Eugene, OR). Phosphate buffered saline (PBS, pH 7.4) solution consisted of 1.1 mM potassium phosphate monobasic, 3 mM sodium phosphate dibasic heptahydrate, and 0.15 M NaCl in 18 MO ? cm deionized water (E-pure, Barnstead, Dubuque, IW). Luria-Bertani (LB) broth for bacterial growth consisted of 10 g of bacto-tryptone, 5 g of yeast extract and 10 g of sodium chloride in 1 L of DI water ( pH 7). Silicon wafers were obtained from Wafer World Inc. (West Palm Beach, FL). Poly(dimethyl siloxane) (PDMS) prepolymer and curing agent (Sylgard 184) were obtained from Dow Corning (Midland, MI). 2.2. Equipment Photogelation reactions were performed using a 365 nm, 300 mW/cm2 light source (EFOS Ultracure 100ss Plus, UV spot lamp, Mississauga, Ontario). Scanning electron microscopy (SEM) was performed with a JEOL T330A at 15 kV (JEOL, Ltd., Peabody, MA). Before SEM characterization, substrates were attached aluminum mounts with carbon tape and sputter-coated with gold Ê . Bright ®eld optical microscopy to a thickness of 200 A was performed with a Zeiss Axiophot 135 microscope (Carl Zeiss Inc., Thornwood, NY). For the Live/Dead Viability assay, the same microscope was used, but in the ¯uorescence mode. The spin-coater used in these studies was obtained from Integrated Technologies (Acushnet,
Control of Mammalian Cell and Bacteria Adhesion on Substrates
MA). The plasma chamber was purchased from Harrick Scienti®c Co. (Ossining, NY). Contact angles were measured with a Rame Hart Model 100 goniometer (Mountain Lakes, NJ) using doubly distilled water. Chrome coated sodalime photomasks used for patterning of microwells were obtained from Advance Reproductions (Andover, MA). USAF negative test target mask was purchased from Melles Griot (Irvine, CA). 2.3. Functionalization of substrates With the goal of improving hydrogel adhesion on substrate surfaces, all substrates are modi®ed with a methacrylate-functionalized trichlorosilane. Silicon and glass surfaces were silanized as described previously (Revzin et al., 2001). Brie¯y, silicon wafers were cleaned in ``piranha'' solution consisting of 3:1 ratio of 30% w/v H2O2 and H2SO4 and subsequently immersed in 1 mM solution of 3-(trichlorosilyl) propyl methacrylate (TPM) in 80%/20% mixture of heptane/carbon tetrachloride. PDMS functionalization was performed according to a published protocol (Silver et al., 1999). In brief, the PDMS surface was ®rst treated in an oxygen plasma for 5 minutes. and then immersed in the silanizing solution containing 10±20 mg TPM per gram of per¯uorooctane. Formation of TPM monolayers on the substrates was con®rmed by ellipsometry and contact angle measurements. 2.4. Preparation of micropatterned surfaces PEG hydrogel micropatterns were fabricated from PEGDA (MW 575) by proximity photolithography. The gel precursor solution was prepared by dissolving 10 mg DMPA (an UV light photoinitiator) per 1 mL of PEG-DA solution. This precursor solution was spin coated at 3,000 rpm for 20 seconds onto the pretreated substrate, forming a uniform layer. A photomask with the desired pattern was brought into proximity of the gel precursor ®lm and illuminated with 365 nm UV light for 0.5 seconds. After UV photopolymerization, only exposed regions underwent free radical induced cross-linking and became insoluble in common PEG solvents such as water. As a result, desired patterns were obtained by washing away unreacted regions with water so that only the projection of the mask remained on the substrate surface. SEM was used to observe pattern morphology. 2.5. Cell culture and the seeding of substrates Mammalian cells were incubated at 37 C in 5% CO2 and 95% air. Murine ®broblasts were cultured in DMEM with 4.5 g/L glucose and 10% FBS. Murine hepatocytes were cultured in DMEM containing 1.0 g/L glucose, 200 nM dexamethasone and 4% FBS. Both phenotypes were cultured to con¯uence in 75 cm2 polystyrene tissue culture ¯asks and con¯uent cells were subcultured every
13
2 to 3 days by trypsinization with 0.25% (w/v) trypsin and 0.13% (w/v) EDTA. To seed patterned substrates, cells were trypsinized from routine culture and centrifuged at 1,000 rpm at 25 C for 5 minutes. The supernatant was removed, cells were resuspended in fresh culture medium containing serum, and an aliquot was obtained for cell counting in a hemocytometer to adjust seeding density. The cells then were seeded at a density of 1.0 6 105 cells/mL on micropatterned substrates which was sterilized by exposure to 365 nm UV light overnight before seeding. The seeded substrates were then cultured under conditions described earlier. Twenty four hours after seeding, the substrates were rinsed with PBS in order to remove non-adherent cells. Adherent cells were either ®xed and imaged by bright-®eld optical microscopy or assayed for cell viability. Substrates patterned with bacteria were also created. Isolated colonies of E. coli were used to inoculate 5 mL of LB broth and cultured at 37 C in roller tubes for 18 hours. Patterned substrates were then immersed in 5 mL of LB broth containing 100 mL of cultured E. coli. Hydrogel patterned substrates were then incubated for 6 hours and rinsed with fresh LB media and stained with Gram Safranin for imaging of adherent cells. 2.6. Cell viability A live/dead viability/cytotoxicity ¯uorescence assay was used to investigate the viability of adherent cells on micropatterned substrates. This assay uses SYTO 10 and Dead Red as ¯uorophores to distinguish living cells and dead cells. SYTO 10 stains live cells green, while Dead Red stains dead cells red. For this assay, 2 mL of the two ¯uorophores were added to 1 mL HEPES-buffered saline (HBS) to make the staining solution. After 24 hours of incubation, the substrates were washed with HBS, the staining solution was placed on the substrates and incubated for 15 minutes in darkness at room temperature. Substrates were washed with HBS again and ®xed with 4% glutaraldehyde in HBS for 60 minutes. The ®xative was then removed and samples were imaged using ¯uorescence microscopy.
3. Results and Discussions 3.1. Functionalization of substrates Surface modi®cation was found to be essential for the fabrication of well-de®ned, hydrogel micropatterns with good adhesion on silicon, glass and polymer substrates. Self-assembly of trichlorosilanes on substrates was used to create surface-tethered methacrylate groups capable of covalent bonding with PEG-DA during free radical induced photogelation. In order to create such an
14
Koh et al.
adhesion-promoting layer, silicon or glass substrates were ®rst treated with ``piranha'' solution. This pretreatment step was needed to hydroxylate the substrate surface. Introducing hydroxylated surfaces into the TPM solution resulted in the formation of a dense network of Si±O±Si bonds on the substrate surface and pendant methacrylate functionalities at the substrate/solution interface. This surface modi®cation was easily visualized by the increase in water contact angle associated with hydrophobic methacrylated alkylsilanes on hydrophilic silicon/glass surfaces (88 + 1 ). Ellipsometry measurements of modi®ed silicon surfaces indicated that the organosilane ®lms Ê , indicating the presence of a monolayer of was 14 + 3 A TPM on the substrate surfaces. Previous studies using time of ¯ight secondary ion mass spectroscopy (TOF SIMS) also con®rmed the presence of the monolayer (Revzin et al., 2001). Because of their potential advantages over silicon or glass substrates regarding ¯exibility and hemo-/biocompatibility, PDMS surfaces were also functionalized with TPM monolayers. Following a similar procedure to that used for glass and silicon, we ®rst treated PDMS in an oxygen plasma to hydroxylate the surface. The oxygen plasma introduced silanol groups (Si-OH) on the surface by the oxidation of methyl groups (Si-CH3) of PDMS at the plasma/polymer interface (McDonald et al., 2000). This oxygen plasma treatment changed the native hydrophobic PDMS surface to one that was more hydrophilic. The contact angle of water on PDMS oxidized for 5 minutes is almost zero, compare to 90 + 5 for untreated PDMS. Oxidized PDMS was then treated with TPM in per¯uorooctane to produce pendent methacrylate functionalities on the surface. After silanization, the contact angle for PDMS increased to 64 + 5 . Such glow discharge oxidation followed by silanization will likely be useful for attaching microstructures to other plastics besides PDMS. 3.2. Fabrication of micropatterned substrates Micropatterned substrates were fabricated with PEG hydrogel microstructures on the functionalized surfaces. The formation of the PEG hydrogels was based on the UV initiated free-radical polymerization of acrylate or methacrylate end groups on PEG derivatives. That is, when exposed to UV light in the presence of a photoinitiator, (meth)acrylate groups formed reactive free radical sites which reacted with each other, thus resulting in the formation of polyacrylate or polymethacrylate networks highly cross-linked with PEG (Mellott et al., 2001). The ability of PEG-DA to gel upon exposure to UV light was used to create negative patterns on substrates using photolithography. Using this tech-
nique, we fabricated various geometries of 3dimensional hydrogel microstructures by changing the design of the mask and encapsulated sensing agents such as living cells or ¯uorophore labeled proteins inside these microstructures (Revzin et al., 2001; Koh et al., 2002). To prepare hydrogel microstructures on substrates, a chrome-coated sodalime photomask containing two different arrays of wells with individual lateral dimensions of 60 6 60 mm and 30 6 30 mm arranged in 20 6 20 arrays were designed and fabricated. Each microwell was surrounded by hydrogel walls which were 20 mm wide and the height of this hydrogel wall was approximately 10 mm as measured by pro®lometry. The volume of the microwells may be easily controled by changing their depth and the area of the well, the former controled by the spin-coating speed which dictates the height of the hydrogel microstructure and the latter by the feature size of the mask. The photomask allowed PEG hydrogel walls to be polymerized upon exposure to UV light while residual unreacted macromer was removed elsewhere by developing the pattern in water. Thus, patterning created a clear contrast between adhesion resistant, hydrophilic hydrogel walls of the microwells and adhesion promoting, hydrophobic methacrylated surfaces inside the microwells. Microstructures were ®rmly anchored to the surface by TPM monolayers and did not delaminate during prolonged exposure to aqueous environments for several weeks. Figures 1(a) and 1(b) show micropatterned silicon substrates with two different sizes of microwells. Clearly de®ned 3-dimensional hydrogel patterns without any residual polymer inside microwells were observed with the PEG hydrogel forming the walls of the microwells. Even though the height of the hydrogel wall increased by nearly 30% by water-induced swelling, noticeable lateral swelling did not occur because of covalent bonding between hydrogel structures and substrates. While inorganic or metal substrates have been used extensively for micropatterning, polymeric substrates present a very attractive and a less explored alternative due to their inexpensive, chemically versatile, and sometimes elastic or biodegradable characteristics. Recently, several research efforts have been directed toward creating a patterned polymeric template (Ward et al., 2001; Hyun and Chilkoti, 2001; Ostuni, 2001b). Most of these efforts used PDMS and were a direct extension of micro-contact printing (``soft lithography'') and micro-molding methods developed by Whitesides and coworkers (Whitesides et al., 2001). In order to demonstrate the ¯exibility of the proposed hydrogel patterning methodology and to enable its future application in combination with soft lithography techniques, we functionalized PDMS surfaces with TPM and fabricated arrays of hydrogel microwells on this substrate. As
Control of Mammalian Cell and Bacteria Adhesion on Substrates
15
shown in Figure 1(c), high quality patterns of hydrogel microstructures could be generated on PDMS substrates. These micropatterns did not delaminate upon hydration or detach when these elastomeric PDMS substrates were ¯exed. 3.3. Cell adhesion on micropatterned substrates The patterning of PEG-based hydrogels on a TPM treated surface created hydrophobic and hydrophilic domains which interacted differently with adsorbing proteins and cells. To investigate the adhesion of cells on PEG hydrogels and TPM monolayers, relatively large hydrogel structures were fabricated on silicon and glass substrates functionalized with TPM, immersed in a suspension of murine ®broblasts, and cultured overnight. Figure 2 clearly shows that minimal cell adhesion occurred on the PEG hydrogel microstructures (walls) while large numbers of cells adhered, spread, and proliferated on the hydrophobic TPM monolayer on the surface of the glass substrate. The exclusionary effects of PEG hydrogels against cells or proteins may be attributed to good conformational ¯exibility, high polymer chain mobility and the hydrophilic nature of PEG (So®a and Merill, 1997). This difference in cell adhesion between TPM monolayers (the base of a well) and PEG hydrogel walls was expected to allow the spatial control of cell spreading in microwells (Figure 2). Figures 3(a) and 4(a) show ®broblasts immobilized on micropatterned substrates with individual microwells with lateral dimensions of 60 6 60 mm and 30 6 30 mm respectively, and a 20 mm wide hydrogel wall in both cases. As expected, for both dimensions, cells attached and spread only on the hydrophobic interior of the microwells with hydrogel walls serving as effective barriers to cell adhesion, proliferation and cross-over. Previous studies demonstrated that dimensions of the cell resistant barrier
Fig. 1. Scanning electron micrographs of a micropatterned substrate consisting of an array of microwells: (a) microwells having lateral dimensions of 60 6 60 mm on the silicon; (b) lateral dimension of 30 6 30 mm on the silicon; (c) hydrogel microstructures on PDMS.
Fig. 2. Optical micrograph of ®broblasts seeded onto a patterned glass substrate with PEG hydrogel microstructure walls and a hydrophobic TPM base.
16
Koh et al.
Fig. 3. Fibroblasts on hydrogel patterned silicon substrates: (a) optical micrograph and (b) live/dead ¯uorescent viability assay of cells cultured inside microwells having lateral dimensions of 60 6 60 mm.
are important in containing and con®ning cells (Chen et al., 1998). Here we demonstrated that a 20 mm hydrogel barrier was suf®cient to segregate cells incubated for 24 hours and we anticipate that the relatively high height of these hydrogel walls will be a more reliable method of con®ning cells inside microwells than other cell patterning technique based on SAMs. If the area of the cell-adherent region is equal to or less than area of maximum spreading of cells, the phenotypic shape and viability of adherent cells is thought to be compromised (Kane et al., 1999). Previous studies also demonstrated that endothelial cells underwent a high rate of apoptosis within 24 hours when the area of the cell-adhesive region was below 500 mm2, while cells con®ned to larger areas did not undergo apoptosis (Chen et al., 1998). Here we observed nearly identical morphologies for ®broblasts in both sizes of
Fig. 4. Fibroblasts cultured inside microwells (silicon substrate) having lateral dimensions of 30 6 30 mm: (a) optical micrograph; (b) live/dead ¯uorescent viability assay.
microwells (3,600 mm2 and 900 mm2) after 24 hours of incubation (see Figures 3(a) and 4(a)). The viability of cells con®ned within both corrals was also investigated using a ¯uorescent live/dead assay that stains live cells green and dead cells red by the difference in membrane permeability between living and dead cells after 24 hours of incubation. Figures 3(b) and 4(b) show ¯uorescent micrographs of stained cells which adhered inside both sizes of microwells. As is evident by the green emitted light, cells remained viable in both geometries. We also clearly observed that while large microwells contained as many as two or three cells in one microwell, we were able to spatially control cell placement at the single-cell level by reducing the size of the microwells. Based on these results, these microwells were large enough for cells to spread and remain viable, and the number of adherent cells could be controled, to the single cell level, by changing the size of microwells. Murine SV-40
Control of Mammalian Cell and Bacteria Adhesion on Substrates
17
Fig. 5. Hepatocytes cultured on a silicon surface patterned with PEG hydrogel microstructures. (a) optical micrograph of hepatocytes con®ned within microwells; (b) live/dead ¯uorescent viability assay.
transformed hepatocytes were also seeded onto identical substrates and similar results were observed (Figures 5(a) and 5(b)). 3.4. Patterning of bacteria on surfaces We also explored the possibility of patterning bacteria on substrates using PEG hydrogel microstructures fabricated on silicon. Patterned silicon substrates were incubated in LB broth containing suspended E. coli for 6 hours. This contact time was previously reported as suf®cient for the relatively rapid phenomenon of bacterial adhesion to occur (Sommer et al., 1999). After incubation, patterned substrates with adherent bacteria were removed and stained with Gram Safranin for imaging using optical microscopy or substrates were imaged directly using electron microscopy. An ESEM micrograph of E. coli anchored to a hydrophobic TPMmodi®ed silicon surface is shown in Figure 6(a). Most of the surface adherent cells are rod-shaped, the phenotype normally associated with E. coli. Figure 6(b) demonstrates the adhesion of bacterial cells presented with a
Fig. 6. Adhesion of E. coli on silicon substrates patterned with hydrogel microstructures: (a) scanning electron micrograph of adherent bacteria on a TPM-modi®ed silicon surface; (b) optical micrograph of three-dimensional hydrogel trenches con®ning bacteria.
choice of TPM-modi®ed silicon and hydrogel surfaces. Hydrogel microstructures were much more resistant to bacterial adhesion than TPM-modi®ed silicon as was evident by the difference in adherent cell density on the two surfaces. Here E. coli preferentially adhered to a silicon surface inside 3-dimensional hydrogel trenches. These bacteria patterning results are preliminary, with further work being necessary to quantify colony-forming units and optimize the hydrogel formulation for maximum adhesion resistance. However, these results clearly demonstrate the potential for creating substrates for selective bacterial adhesion which could be used in designing novel, cell-based biosensor arrays based on engineered bacteria in a microwell format.
18
4.
Koh et al.
Conclusion
Here we demonstrated a simple and general method using photolithography to fabricate patterned hydrogel microstructures and the ability of these structures to control the adhesion and proliferation of mammalian cells on various substrates. Micropatterned substrates were prepared by the photogelation of PEG macromers onto surfaces functionalized with a 3-(trichlorosilyl)propyl methacrylate monolayer. These functionalized surfaces not only ensured the attachment of the hydrogel to the substrate but also allowed cell adhesion in areas absent of hydrogel because of their hydrophobicity. Using photomasks, we created arrays of microwells and demonstrated that ®broblasts and hepatocytes adhered and spread only on the TPM monolayer without losing their viability. By changing the size of microwells, we could control the number of cells in one microwell. In the case of 30 6 30 mm microwells, as many as 400 microwells were fabricated in 1 mm2 and more than 3 million microwells can be fabricated in the same area as a 96 well plate. Mammalian cells placed in high density microwell plates described here could potentially be used in high content drug screening and may be combined with micro¯uidic devices to create multianalyte cellbased biosensors. Initial results demonstrating the ability of PEG hydrogel microstructures to control bacterial adhesion were also shown, opening the possibility to create genetically engineering cell-based biosensor arrays for the detection of a variety of analytes including pesticides and neurotoxins.
Acknowledgments We gratefully acknowledge ®nancial support from the National Aeronautics and Space Administration (NAG 9 1277) and the Texas Advanced Technology Program. We also thank Prof. Richard Crooks (Department of Chemistry, Texas A&M University) for the use of the contact angle instrument. MVP wishes to thank the Alfred P. Sloan Foundation for its support through a research fellowship.
References M.L. Amirpour, P. Ghosh, W.M. Lackowski, R.M. Crooks, and M.V. Pishko, Analytical Chemistry 73, 1560±1566 (2001). Y.H. An and R.J. Friedman, Journal of Biomedical Materials Research 43, 338±348 (1998). S. Belkin, D.R. Smulski, S. Dadon, A.C. Vollmer, T.K. Van Dyk, and R.A. Larossa, Water Research 31, 3009±3016 (1997).
S.N. Bhatia, M.L. Yarmush, and M. Toner, Journal of Biomedical Materials Research 34, 189±199 (1997). S.N. Bhatia, U. Balis, M. Yarmush, and M. Toner, Biotechnol. Prog. 14, 378±387 (1998). L. Bousse, Sensors and Actuators B 34, 270±275 (1996). C. Chen, M. Mrksich, S. Huang, G. Whitesides, and D. Ingber, Science 276, 1425±1428 (1997). C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, and D.E. Ingber, Biotechnology Progress 14, 356±363 (1998). S.E. Cowan, D. Lipmann, and J.D. Keasling, Biotechnology Letters 23, 1235±1241 (2001). D. Cunliffe, C.A. Smart, C. Alexander, and E.N. Vulfson, Applied and Environmental Microbiology 65, 4995±5002 (1999). E.T. den Braber, J.E. de Ruijter, H.T. Smits, L.A. Ginsel, A.F. von Recum, and J.A. Jansen, Biomaterials 17, 1093±1099 (1996). J.L. Dewez, J.B. Lhoest, E. Detrait, V. Berger, C.C. Dupont-Gillain, L.M. Vincent, Y.J. Schneider, P. Bertrand, and P.G. Rouxhet, Biomaterials 19, 1441±1445 (1998). A. Folch and M. Toner, Biotechnol. Prog. 14, 388±392 (1998). A. Folch and M. Toner, Annual Review of Biomedical Engineering 2, 227±256 (2000). M. Franko, P.F. Nealey, S. Campbell, A.I. Teixeira, and C.J. Murphy, Journal of Biomedical Materials Research 52, 261±269 (2000). P. Ghosh, M.L. Amirpour, W.M. Lackowski, M.V. Pishko, and R.M. Crooks, Angewandte Chemie 38, 1592±1595 (1999). J. Hyun and A. Chilkoti, Journal of the American Chemical Society 123, 6943±6944 (2001). Y. Ito, Biomaterials 20, 2333±2342 (1999). P.S. John, R. Davis, N. Cady, J. Czajka, C. Batt, and H. Craighead, Analytical Chemistry 70, 1108±1111 (1998). D.R. Jung, R. Kapur, T. Adams, K.A. Giuliano, M. Mrksich, H.G. Craighead, and D.L. Taylor, Critical Reviews in Biotechnology 21, 111±154 (2001). R.S. Kane, S. Takayama, E. Ostuni, D.E. Ingber, and G.M. Whitesides, Biomaterials 20, 2363±2376 (1999). W. Koh, A. Revzin, and M.V. Pishko, Langmuir 18, 2459±2462 (2002). J. Lahiri, L. Isaacs, J. Tien, and G. Whitesides, Anal. Chem. 71, 777± 790 (1999). L.H. Larsen, T. Kjaer, and N.P. Revsbech, Analytical Chemistry 69, 3527±3531 (1997). J.-D. Liao, S.-H. Wang, and D.-J. Hsu, Sensors and Actuators B 72, 167±173 (2001). G. Lopez, M. Albers, S. Schreiber, R. Carroll, E. Peralta, and G. Whitesides, J. Am. Chem. Soc. 115, 5877±5878 (1993). L. Lu, K. Nyalakonda, L. Kam, R. Bizios, A. Gopferich, and A.G. Mikos, Biomaterials 22, 291±297 (2001). S.A. Makohliso, D. Leonard, L. Giovangrandi, H.J. Mathieu, M. Ilegems, and P. Aebisher, Langmuir 15, 2940±2946 (1999). T. Matsuda and T. Sugawara, Journal of Biomedical Materials Research 29, 749±756 (1995). J.C. McDonald, D.C. Duffy, J.R. Anderson, D.T. Chiu, H. Wu, O.J.A. Schueller, and G.M. Whitesides, Electrophoresis 21, 27±40 (2000). M.B. Mellott, K. Searcy, and M.V. Pishko, Biomaterials 22, 929±941 (2001). M. Mrksich, C.S. Chen, Y. Xia, L.E. Dike, D.E. Ingber, and G.M. Whitesides, Proceedings of the National Academy of Sciences of the United States of America 93, 10775±10778 (1996). M. Mrksich, L. Dike, J. Tien, D. Ingber, and G. Whitesides, Experimental Cell Research 235, 305±313 (1997). E. Ostuni, R.G. Chapman, M.N. Liang, G. Meluleni, G. Pier, D.E. Ingber, and G.M. Whitesides, Langmuir 17, 6336±6343 (2001a). E. Ostuni, C.S. Chen, D.E. Ingber, and G.M. Whitesides, Langmuir 17, 2828±2834 (2001b).
Control of Mammalian Cell and Bacteria Adhesion on Substrates K.D. Park, Y.S. Kim, D.K. Han, Y.H. Kim, E.H.B. Lee, H. Suh, and K.S. Choi, Biomaterials 19, 851±859 (1998). E.I. Rainina, I.E. Badalin, O.V. Ignatov, A.Y. Fedorov, A.L. Simonian, and S.D. Varfolomeyev, Applied Biochemistry amd Biotechnology 56, 117±127 (1996a). E.I. Rainina, E.N. Efremenko, S.D. Varfolomeyev, A.L. Simonian, and J.R. Wild, Biosensors and Bioelectronics 10, 991±1000 (1996b). A. Razatos, Y.L. Ong, F. Boulay, D.L. Elbert, J.A. Hubbell, M.M. Sharma, and G. Georgiou, Langmuir 16, 9155±9158 (2000). A. Revzin, R.J. Russell, V.K. Yadavalli, W. Koh, C. Deister, D.D. Hile, M.B. Mellott, and M.V. Pishko, Langmuir 17, 5440±5447 (2001). J.H. Silver, J. Lin, F. Lim, V.A. Tegoulia, M.K. Chaudhury, and S.L. Cooper, Biomaterials 20, 1533±1543 (1999). R. Singhvi, A. Kumer, G.P. Lopez, G.N. Stephanopoulos, D.I.C. Daniel, I.C. Wang, G.M. Whitesides, and D.E. Ingber, Science 264, 696±698 (1994a). R. Singhvi, A. Kumar, G. Lopez, G. Stephanopoulos, D. Wang, G. Whitesides, and D. Ingber, Science 264, 696±698 (1994b). K. Sirkar and M.V. Pishko, Analytical Chemistry 70, 2888±2894 (1998). S.J. So®a and E. Merill, Poly(ethylene glycol): Chemistry and Biological Applications, Vol. 680, edited by J.M. Harris, American Chemical Society, Washington D.C., 1997.
19
P. Sommer, C. Martin-Rouas, and E. Mettler, Food Microbiology 16, 50±515 (1999). C.J. Sundberg, Current Opinion in Biotechnology 11, 47±53 (2000). S. Takayama, J. McDonald, E. Ostuni, M. Liang, P. Kenis, R. Ismagilov, and G. Whitesides, Proc. Natl. Acad. Sci. USA 96, 5545±5548 (1999). C.H. Thomas, C.D. McFarland, M.L. Jenkins, A. Rezania, J.G. Steele, and K.E. Healy, Journal of Biomedical Materials Research 37, 81±93 (1997). J. Turner, W. Shain, D. Szarowski, M. Andersen, S. Martins, M. Isaacson, and H. Craighead, Exp. Neurol. 156, 33±49 (1999). E. Tziampazis, J. Kohn, and P.V. Moghe, Biomaterials 21, 511±520 (2000). J.H. Ward, R. Bashir, and N.A. Peppas, Journal of Biomedical Materials Research 56, 351±360 (2001). G. Whitesides, A. Kumar, G. Lopez, H. Biebuyck, P. DiMilla, R. Haerter, and K. Prime, Journal of Cellular Biochemistry 18C, 241, 250, 270 (1994). G. Whitesides, E. Ostuni, S. Takayama, X. Jiang, and D.E. Ingber, Annual Review of Biomedical Engineering 3, 335±373 (2001). M.Q. Zhang, T. Desai, and M. Ferrari, Biomaterials 19, 953±960 (1998). S. Zhang, L. Yan, M. Altman, M. Lassle, H. Nugent, F. Frankel, D. Lauffenburger, G. Whitesides, and A. Rich, Biomaterials 20, 1213± 1220 (1999).
BioMEMs Materials and Fabrication Technology: Section Editors Dr Shuvo Roy and Dr Mak Paranjape
Control of Mammalian Cell and Bacteria Adhesion on Substrates Micropatterned with Poly(ethylene glycol) Hydrogels Won-Gun Koh,1 Alexander Revzin,2 Aleksandr Simonian,3 Tony Reeves,3 and Michael Pishko1*
1 Department of Chemical Engineering and Department of Materials Science and Engineering, The Pennsylvania University, University Park, PA 16802-4420 E-mail: [email protected] 2 Department of Chemical Engineering, Texas A&M University, College Station, TX 77843-3122 3 Department of Biochemistry and Biophysics, Texas A&M University, College Station, TX 77843-3122
Abstract. A simple method for controlling the spatial positioning of mammalian cells and bacteria on substrates using patterned poly(ethylene glycol) (PEG) hydrogel microstructures is described. These microstructures were fabricated using photolithography on silicon, glass or poly (dimethylsiloxane) (PDMS) surfaces modi®ed with a 3-(trichlorosilyl) propyl methacrylate (TPM) monolayer. During the photogelation reaction, the resulting hydrogel microstructures were covalently bound to the substrate via the TPM monolayer and did not detached from the substrate upon hydration. For mammalian cell patterning, microwell arrays of different dimensions were fabricated. These microwells were composed of hydrophilic PEG hydrogel walls surrounding hydrophobic TPM ¯oors inside the microwells. Murine 3T3 ®broblasts and transformed hepatocytes were shown to selectively adhere to the TPM monolayer inside the microwells, maintaining their viability, while adherent cells were not present on the hydrogel walls. The number of cells inside one microwell could be controled by changing the lateral dimension of the microwells, thus allowing only a single cell per microwell if desired. In the case of 30 6 30 lm microwells, as many as 400 microwells were fabricated in 1 mm2. In addition, PEG hydrogel microstructures were also shown to effectively resist the adhesion of bacteria such as Escherichia coli. Key Words. poly(ethylene glycol) hydrogel microstructures, photolithography, cell patterning, microwells
1.
Introduction
Many mammalian cell-based bioassays for the screening of large libraries of potential pharmaceutical agents or the detection of toxic compounds and pathogens in the environment have been developed over the years and some are in use today. Although most of cell-based assays are still performed in 96 well plates, many researches are moving toward 384 well and higher density plate formats
to improve the performance and throughput of these assays. With further assay miniaturization in mind, a number of researchers have explored the patterned deposition of cells in microsystems through the control of cellular adhesion (Bousse, 1996; Makohliso et al., 1999; Sundberg, 2000; Kane et al., 1999; Jung et al., 2001; Chen et al., 1997, 1998; Lahiri et al., 1999; Lopez et al., 1993; Mrksich et al., 1997; Singhvi et al., 1994; Takayama et al., 1999; Whitesides et al., 1994; Zhang et al., 1999; Turner et al., 1999; John et al., 1998; Bhatia et al., 1997, 1998; Folch et al., 1998; den Braber and Toner, 1996). Similarly, the control of bacterial adhesion may be a valuable tool toward developing cell-based sensor arrays based on genetically engineered bacteria. Early studies of bacterial adhesion have mostly focused on reducing bacterial adhesion on implanted biomaterials, the critical event in the pathogenesis of foreign body infections (An and Friedman, 1998; Park et al., 1998). However, because analyte speci®city can be readily modi®ed by genetic engineering and because of the relatively robust nature of these micro-organisms as compared to mammalian cells, bacterial cells have been studied extensively for sensing applications (Rainina et al., 1996a, 1996b; Larsen et al., 1997; Belkin et al., 1997; Liao et al, 2001). The ability to selectively attach bacteria to micropatterned substrates could be used to create sensor arrays for applications such as rapid screening for infectious diseases or to detect toxic compounds (Cowan et al., 2001; Ostuni et al., 2001a).
*Corresponding author: Department of Chemical Engineering, The Pennsylvania State University, 104 Fenske Laboratory, University Park, PA 16802-4420.
11
12
Koh et al.
A number of studies reported that cellular adhesion is controled by the chemical and physical characteristics of surfaces such as hydrophobicity and hydrophilicity, surface charge, and surface roughness (Jung et al., 2001; Folch and Toner, 2000; Ito, 1999). These factors can be readily manipulated to control the spatial distribution of cells on a substrate. The rapid development of microfabrication and surface engineering techniques have stimulated the development of novel methods that may be used to control cell adhesion and spreading with micrometer-scale resolution. The most common approaches to de®ning and controlling regions that promote or resist the adhesion of cells on this length scale is to control the surface tension and wetting properties by patterning hydrophilic and hydrophobic regions on the surface using photolithography (Matsuda and Sugawara, 1995; Thomas et al., 1997; Dewez et al., 1998) or microcontact printing ( m-CP) patterns of selfassembly monolayers (SAM) on the surfaces (Kane et al., 1999; Chen et al., 1998; Amirpour et al., 2001; Ghosh et al., 1999; Franko et al., 2000; Mrksich et al., 1996; Singhvi et al., 1994a; Whitesides et al., 2001; Matsuda and Sugawara, 1995). Here we controlled cell adhesion using hydrogels of poly (ethylene glycol) (PEG), a non-biodegradable and hydrophilic polymer that has been previously used to minimize protein adsorption and mammalian cell adhesion onto biomaterial surfaces (Zhang et al., 1998; Tziampazis et al., 2000; Lu et al., 2001). PEG grafting to surfaces has been also used to render a variety of surfaces resistant to bacterial adhesion (Park et al., 1998; Ostuni et al., 2001b; Razatos et al., 2000; Cunliffe et al., 1999). The groups of Ingber and Whitesides demonstrated that SAMs containing hydrophilic ethylene glycol oligomers as the terminal segment resist cell adhesion while hydrophobic SAMs promote cell adhesion. These regions can be con®ned to 10±100 mm wide areas by patterning these SAMs using m-CP on a gold surface (Singhvi et al., 1994b). In this study we described the control of cell adhesion and the spatial con®nement of cell proliferation by means of micrometer scale PEG hydrogel structures on silicon, glass or poly (dimethylsiloxane) surfaces modi®ed with a chlorosilane coupling agent. Acrylate or methacrylate derivatives of PEG can be cross-linked into hydrogels and photolithographically patterned on silicon, glass, and plastic surfaces (Sirkar and Pishko, 1998; Ward et al., 2001; Revzin et al., 2001). Using these patterned substrates, we demonstrated the ability to capture and con®ne mammalian cells (®broblasts and hepatocytes) inside arrays of microwells fabricated from hydrogels of varying dimensions. In addition, we demonstrated the partial resistance of PEG hydrogel microstructures to the adhesion of a model bacterium, Escherichia coli.
2. Materials and Methods 2.1. Materials Poly(ethylene glycol) diacrylate (PEG-DA, MW 575), 2,20 -dimethoxy-2-phenyl-acetophenone (DMPA), anhydrous carbon tetrachloride, n-heptane and per¯uorooctane were purchased from Aldrich Chemical Co. (Milwaukee, WI). Dulbecco's modi®ed Eagle's medium (DMEM), fetal bovine serum (FBS), antibiotic/antimycotic solution, trypsin, ethylenediaminetetra-acetate (EDTA), sodium chloride, sodium phosphate and potassium phosphate monobasic were purchased from Sigma Chemical Co. (St. Louis, MO). 3-(trichlorosilyl) propyl methacrylate (TPM) was purchased from Fluka Chemicals (Milwaukee, WI). Hydrogen peroxide was purchased from EM Science (Gibbstown, NJ). Sulfuric acid was purchased from Fisher Scienti®c (Fair Lawn, NJ). E. coli bacteria (DH5) were obtained from Novagen Inc. (Madison, WI). Bacto-tryptone and yeast extract were obtained from Becton & Dickinson (Franklin Lakes, NJ). Gram Safranin stain was purchased from Fisher Diagnostics (Pittsburgh, PA). Murine 3T3 ®broblasts and murine hepatocytes (SV40 transformed) were obtained from American Type Culture Collection (Manssas, VA). Live/Dead Viability/Cytotoxicity Kit (L-7013) was obtained from Molecular Probes (Eugene, OR). Phosphate buffered saline (PBS, pH 7.4) solution consisted of 1.1 mM potassium phosphate monobasic, 3 mM sodium phosphate dibasic heptahydrate, and 0.15 M NaCl in 18 MO ? cm deionized water (E-pure, Barnstead, Dubuque, IW). Luria-Bertani (LB) broth for bacterial growth consisted of 10 g of bacto-tryptone, 5 g of yeast extract and 10 g of sodium chloride in 1 L of DI water ( pH 7). Silicon wafers were obtained from Wafer World Inc. (West Palm Beach, FL). Poly(dimethyl siloxane) (PDMS) prepolymer and curing agent (Sylgard 184) were obtained from Dow Corning (Midland, MI). 2.2. Equipment Photogelation reactions were performed using a 365 nm, 300 mW/cm2 light source (EFOS Ultracure 100ss Plus, UV spot lamp, Mississauga, Ontario). Scanning electron microscopy (SEM) was performed with a JEOL T330A at 15 kV (JEOL, Ltd., Peabody, MA). Before SEM characterization, substrates were attached aluminum mounts with carbon tape and sputter-coated with gold Ê . Bright ®eld optical microscopy to a thickness of 200 A was performed with a Zeiss Axiophot 135 microscope (Carl Zeiss Inc., Thornwood, NY). For the Live/Dead Viability assay, the same microscope was used, but in the ¯uorescence mode. The spin-coater used in these studies was obtained from Integrated Technologies (Acushnet,
Control of Mammalian Cell and Bacteria Adhesion on Substrates
MA). The plasma chamber was purchased from Harrick Scienti®c Co. (Ossining, NY). Contact angles were measured with a Rame Hart Model 100 goniometer (Mountain Lakes, NJ) using doubly distilled water. Chrome coated sodalime photomasks used for patterning of microwells were obtained from Advance Reproductions (Andover, MA). USAF negative test target mask was purchased from Melles Griot (Irvine, CA). 2.3. Functionalization of substrates With the goal of improving hydrogel adhesion on substrate surfaces, all substrates are modi®ed with a methacrylate-functionalized trichlorosilane. Silicon and glass surfaces were silanized as described previously (Revzin et al., 2001). Brie¯y, silicon wafers were cleaned in ``piranha'' solution consisting of 3:1 ratio of 30% w/v H2O2 and H2SO4 and subsequently immersed in 1 mM solution of 3-(trichlorosilyl) propyl methacrylate (TPM) in 80%/20% mixture of heptane/carbon tetrachloride. PDMS functionalization was performed according to a published protocol (Silver et al., 1999). In brief, the PDMS surface was ®rst treated in an oxygen plasma for 5 minutes. and then immersed in the silanizing solution containing 10±20 mg TPM per gram of per¯uorooctane. Formation of TPM monolayers on the substrates was con®rmed by ellipsometry and contact angle measurements. 2.4. Preparation of micropatterned surfaces PEG hydrogel micropatterns were fabricated from PEGDA (MW 575) by proximity photolithography. The gel precursor solution was prepared by dissolving 10 mg DMPA (an UV light photoinitiator) per 1 mL of PEG-DA solution. This precursor solution was spin coated at 3,000 rpm for 20 seconds onto the pretreated substrate, forming a uniform layer. A photomask with the desired pattern was brought into proximity of the gel precursor ®lm and illuminated with 365 nm UV light for 0.5 seconds. After UV photopolymerization, only exposed regions underwent free radical induced cross-linking and became insoluble in common PEG solvents such as water. As a result, desired patterns were obtained by washing away unreacted regions with water so that only the projection of the mask remained on the substrate surface. SEM was used to observe pattern morphology. 2.5. Cell culture and the seeding of substrates Mammalian cells were incubated at 37 C in 5% CO2 and 95% air. Murine ®broblasts were cultured in DMEM with 4.5 g/L glucose and 10% FBS. Murine hepatocytes were cultured in DMEM containing 1.0 g/L glucose, 200 nM dexamethasone and 4% FBS. Both phenotypes were cultured to con¯uence in 75 cm2 polystyrene tissue culture ¯asks and con¯uent cells were subcultured every
13
2 to 3 days by trypsinization with 0.25% (w/v) trypsin and 0.13% (w/v) EDTA. To seed patterned substrates, cells were trypsinized from routine culture and centrifuged at 1,000 rpm at 25 C for 5 minutes. The supernatant was removed, cells were resuspended in fresh culture medium containing serum, and an aliquot was obtained for cell counting in a hemocytometer to adjust seeding density. The cells then were seeded at a density of 1.0 6 105 cells/mL on micropatterned substrates which was sterilized by exposure to 365 nm UV light overnight before seeding. The seeded substrates were then cultured under conditions described earlier. Twenty four hours after seeding, the substrates were rinsed with PBS in order to remove non-adherent cells. Adherent cells were either ®xed and imaged by bright-®eld optical microscopy or assayed for cell viability. Substrates patterned with bacteria were also created. Isolated colonies of E. coli were used to inoculate 5 mL of LB broth and cultured at 37 C in roller tubes for 18 hours. Patterned substrates were then immersed in 5 mL of LB broth containing 100 mL of cultured E. coli. Hydrogel patterned substrates were then incubated for 6 hours and rinsed with fresh LB media and stained with Gram Safranin for imaging of adherent cells. 2.6. Cell viability A live/dead viability/cytotoxicity ¯uorescence assay was used to investigate the viability of adherent cells on micropatterned substrates. This assay uses SYTO 10 and Dead Red as ¯uorophores to distinguish living cells and dead cells. SYTO 10 stains live cells green, while Dead Red stains dead cells red. For this assay, 2 mL of the two ¯uorophores were added to 1 mL HEPES-buffered saline (HBS) to make the staining solution. After 24 hours of incubation, the substrates were washed with HBS, the staining solution was placed on the substrates and incubated for 15 minutes in darkness at room temperature. Substrates were washed with HBS again and ®xed with 4% glutaraldehyde in HBS for 60 minutes. The ®xative was then removed and samples were imaged using ¯uorescence microscopy.
3. Results and Discussions 3.1. Functionalization of substrates Surface modi®cation was found to be essential for the fabrication of well-de®ned, hydrogel micropatterns with good adhesion on silicon, glass and polymer substrates. Self-assembly of trichlorosilanes on substrates was used to create surface-tethered methacrylate groups capable of covalent bonding with PEG-DA during free radical induced photogelation. In order to create such an
14
Koh et al.
adhesion-promoting layer, silicon or glass substrates were ®rst treated with ``piranha'' solution. This pretreatment step was needed to hydroxylate the substrate surface. Introducing hydroxylated surfaces into the TPM solution resulted in the formation of a dense network of Si±O±Si bonds on the substrate surface and pendant methacrylate functionalities at the substrate/solution interface. This surface modi®cation was easily visualized by the increase in water contact angle associated with hydrophobic methacrylated alkylsilanes on hydrophilic silicon/glass surfaces (88 + 1 ). Ellipsometry measurements of modi®ed silicon surfaces indicated that the organosilane ®lms Ê , indicating the presence of a monolayer of was 14 + 3 A TPM on the substrate surfaces. Previous studies using time of ¯ight secondary ion mass spectroscopy (TOF SIMS) also con®rmed the presence of the monolayer (Revzin et al., 2001). Because of their potential advantages over silicon or glass substrates regarding ¯exibility and hemo-/biocompatibility, PDMS surfaces were also functionalized with TPM monolayers. Following a similar procedure to that used for glass and silicon, we ®rst treated PDMS in an oxygen plasma to hydroxylate the surface. The oxygen plasma introduced silanol groups (Si-OH) on the surface by the oxidation of methyl groups (Si-CH3) of PDMS at the plasma/polymer interface (McDonald et al., 2000). This oxygen plasma treatment changed the native hydrophobic PDMS surface to one that was more hydrophilic. The contact angle of water on PDMS oxidized for 5 minutes is almost zero, compare to 90 + 5 for untreated PDMS. Oxidized PDMS was then treated with TPM in per¯uorooctane to produce pendent methacrylate functionalities on the surface. After silanization, the contact angle for PDMS increased to 64 + 5 . Such glow discharge oxidation followed by silanization will likely be useful for attaching microstructures to other plastics besides PDMS. 3.2. Fabrication of micropatterned substrates Micropatterned substrates were fabricated with PEG hydrogel microstructures on the functionalized surfaces. The formation of the PEG hydrogels was based on the UV initiated free-radical polymerization of acrylate or methacrylate end groups on PEG derivatives. That is, when exposed to UV light in the presence of a photoinitiator, (meth)acrylate groups formed reactive free radical sites which reacted with each other, thus resulting in the formation of polyacrylate or polymethacrylate networks highly cross-linked with PEG (Mellott et al., 2001). The ability of PEG-DA to gel upon exposure to UV light was used to create negative patterns on substrates using photolithography. Using this tech-
nique, we fabricated various geometries of 3dimensional hydrogel microstructures by changing the design of the mask and encapsulated sensing agents such as living cells or ¯uorophore labeled proteins inside these microstructures (Revzin et al., 2001; Koh et al., 2002). To prepare hydrogel microstructures on substrates, a chrome-coated sodalime photomask containing two different arrays of wells with individual lateral dimensions of 60 6 60 mm and 30 6 30 mm arranged in 20 6 20 arrays were designed and fabricated. Each microwell was surrounded by hydrogel walls which were 20 mm wide and the height of this hydrogel wall was approximately 10 mm as measured by pro®lometry. The volume of the microwells may be easily controled by changing their depth and the area of the well, the former controled by the spin-coating speed which dictates the height of the hydrogel microstructure and the latter by the feature size of the mask. The photomask allowed PEG hydrogel walls to be polymerized upon exposure to UV light while residual unreacted macromer was removed elsewhere by developing the pattern in water. Thus, patterning created a clear contrast between adhesion resistant, hydrophilic hydrogel walls of the microwells and adhesion promoting, hydrophobic methacrylated surfaces inside the microwells. Microstructures were ®rmly anchored to the surface by TPM monolayers and did not delaminate during prolonged exposure to aqueous environments for several weeks. Figures 1(a) and 1(b) show micropatterned silicon substrates with two different sizes of microwells. Clearly de®ned 3-dimensional hydrogel patterns without any residual polymer inside microwells were observed with the PEG hydrogel forming the walls of the microwells. Even though the height of the hydrogel wall increased by nearly 30% by water-induced swelling, noticeable lateral swelling did not occur because of covalent bonding between hydrogel structures and substrates. While inorganic or metal substrates have been used extensively for micropatterning, polymeric substrates present a very attractive and a less explored alternative due to their inexpensive, chemically versatile, and sometimes elastic or biodegradable characteristics. Recently, several research efforts have been directed toward creating a patterned polymeric template (Ward et al., 2001; Hyun and Chilkoti, 2001; Ostuni, 2001b). Most of these efforts used PDMS and were a direct extension of micro-contact printing (``soft lithography'') and micro-molding methods developed by Whitesides and coworkers (Whitesides et al., 2001). In order to demonstrate the ¯exibility of the proposed hydrogel patterning methodology and to enable its future application in combination with soft lithography techniques, we functionalized PDMS surfaces with TPM and fabricated arrays of hydrogel microwells on this substrate. As
Control of Mammalian Cell and Bacteria Adhesion on Substrates
15
shown in Figure 1(c), high quality patterns of hydrogel microstructures could be generated on PDMS substrates. These micropatterns did not delaminate upon hydration or detach when these elastomeric PDMS substrates were ¯exed. 3.3. Cell adhesion on micropatterned substrates The patterning of PEG-based hydrogels on a TPM treated surface created hydrophobic and hydrophilic domains which interacted differently with adsorbing proteins and cells. To investigate the adhesion of cells on PEG hydrogels and TPM monolayers, relatively large hydrogel structures were fabricated on silicon and glass substrates functionalized with TPM, immersed in a suspension of murine ®broblasts, and cultured overnight. Figure 2 clearly shows that minimal cell adhesion occurred on the PEG hydrogel microstructures (walls) while large numbers of cells adhered, spread, and proliferated on the hydrophobic TPM monolayer on the surface of the glass substrate. The exclusionary effects of PEG hydrogels against cells or proteins may be attributed to good conformational ¯exibility, high polymer chain mobility and the hydrophilic nature of PEG (So®a and Merill, 1997). This difference in cell adhesion between TPM monolayers (the base of a well) and PEG hydrogel walls was expected to allow the spatial control of cell spreading in microwells (Figure 2). Figures 3(a) and 4(a) show ®broblasts immobilized on micropatterned substrates with individual microwells with lateral dimensions of 60 6 60 mm and 30 6 30 mm respectively, and a 20 mm wide hydrogel wall in both cases. As expected, for both dimensions, cells attached and spread only on the hydrophobic interior of the microwells with hydrogel walls serving as effective barriers to cell adhesion, proliferation and cross-over. Previous studies demonstrated that dimensions of the cell resistant barrier
Fig. 1. Scanning electron micrographs of a micropatterned substrate consisting of an array of microwells: (a) microwells having lateral dimensions of 60 6 60 mm on the silicon; (b) lateral dimension of 30 6 30 mm on the silicon; (c) hydrogel microstructures on PDMS.
Fig. 2. Optical micrograph of ®broblasts seeded onto a patterned glass substrate with PEG hydrogel microstructure walls and a hydrophobic TPM base.
16
Koh et al.
Fig. 3. Fibroblasts on hydrogel patterned silicon substrates: (a) optical micrograph and (b) live/dead ¯uorescent viability assay of cells cultured inside microwells having lateral dimensions of 60 6 60 mm.
are important in containing and con®ning cells (Chen et al., 1998). Here we demonstrated that a 20 mm hydrogel barrier was suf®cient to segregate cells incubated for 24 hours and we anticipate that the relatively high height of these hydrogel walls will be a more reliable method of con®ning cells inside microwells than other cell patterning technique based on SAMs. If the area of the cell-adherent region is equal to or less than area of maximum spreading of cells, the phenotypic shape and viability of adherent cells is thought to be compromised (Kane et al., 1999). Previous studies also demonstrated that endothelial cells underwent a high rate of apoptosis within 24 hours when the area of the cell-adhesive region was below 500 mm2, while cells con®ned to larger areas did not undergo apoptosis (Chen et al., 1998). Here we observed nearly identical morphologies for ®broblasts in both sizes of
Fig. 4. Fibroblasts cultured inside microwells (silicon substrate) having lateral dimensions of 30 6 30 mm: (a) optical micrograph; (b) live/dead ¯uorescent viability assay.
microwells (3,600 mm2 and 900 mm2) after 24 hours of incubation (see Figures 3(a) and 4(a)). The viability of cells con®ned within both corrals was also investigated using a ¯uorescent live/dead assay that stains live cells green and dead cells red by the difference in membrane permeability between living and dead cells after 24 hours of incubation. Figures 3(b) and 4(b) show ¯uorescent micrographs of stained cells which adhered inside both sizes of microwells. As is evident by the green emitted light, cells remained viable in both geometries. We also clearly observed that while large microwells contained as many as two or three cells in one microwell, we were able to spatially control cell placement at the single-cell level by reducing the size of the microwells. Based on these results, these microwells were large enough for cells to spread and remain viable, and the number of adherent cells could be controled, to the single cell level, by changing the size of microwells. Murine SV-40
Control of Mammalian Cell and Bacteria Adhesion on Substrates
17
Fig. 5. Hepatocytes cultured on a silicon surface patterned with PEG hydrogel microstructures. (a) optical micrograph of hepatocytes con®ned within microwells; (b) live/dead ¯uorescent viability assay.
transformed hepatocytes were also seeded onto identical substrates and similar results were observed (Figures 5(a) and 5(b)). 3.4. Patterning of bacteria on surfaces We also explored the possibility of patterning bacteria on substrates using PEG hydrogel microstructures fabricated on silicon. Patterned silicon substrates were incubated in LB broth containing suspended E. coli for 6 hours. This contact time was previously reported as suf®cient for the relatively rapid phenomenon of bacterial adhesion to occur (Sommer et al., 1999). After incubation, patterned substrates with adherent bacteria were removed and stained with Gram Safranin for imaging using optical microscopy or substrates were imaged directly using electron microscopy. An ESEM micrograph of E. coli anchored to a hydrophobic TPMmodi®ed silicon surface is shown in Figure 6(a). Most of the surface adherent cells are rod-shaped, the phenotype normally associated with E. coli. Figure 6(b) demonstrates the adhesion of bacterial cells presented with a
Fig. 6. Adhesion of E. coli on silicon substrates patterned with hydrogel microstructures: (a) scanning electron micrograph of adherent bacteria on a TPM-modi®ed silicon surface; (b) optical micrograph of three-dimensional hydrogel trenches con®ning bacteria.
choice of TPM-modi®ed silicon and hydrogel surfaces. Hydrogel microstructures were much more resistant to bacterial adhesion than TPM-modi®ed silicon as was evident by the difference in adherent cell density on the two surfaces. Here E. coli preferentially adhered to a silicon surface inside 3-dimensional hydrogel trenches. These bacteria patterning results are preliminary, with further work being necessary to quantify colony-forming units and optimize the hydrogel formulation for maximum adhesion resistance. However, these results clearly demonstrate the potential for creating substrates for selective bacterial adhesion which could be used in designing novel, cell-based biosensor arrays based on engineered bacteria in a microwell format.
18
4.
Koh et al.
Conclusion
Here we demonstrated a simple and general method using photolithography to fabricate patterned hydrogel microstructures and the ability of these structures to control the adhesion and proliferation of mammalian cells on various substrates. Micropatterned substrates were prepared by the photogelation of PEG macromers onto surfaces functionalized with a 3-(trichlorosilyl)propyl methacrylate monolayer. These functionalized surfaces not only ensured the attachment of the hydrogel to the substrate but also allowed cell adhesion in areas absent of hydrogel because of their hydrophobicity. Using photomasks, we created arrays of microwells and demonstrated that ®broblasts and hepatocytes adhered and spread only on the TPM monolayer without losing their viability. By changing the size of microwells, we could control the number of cells in one microwell. In the case of 30 6 30 mm microwells, as many as 400 microwells were fabricated in 1 mm2 and more than 3 million microwells can be fabricated in the same area as a 96 well plate. Mammalian cells placed in high density microwell plates described here could potentially be used in high content drug screening and may be combined with micro¯uidic devices to create multianalyte cellbased biosensors. Initial results demonstrating the ability of PEG hydrogel microstructures to control bacterial adhesion were also shown, opening the possibility to create genetically engineering cell-based biosensor arrays for the detection of a variety of analytes including pesticides and neurotoxins.
Acknowledgments We gratefully acknowledge ®nancial support from the National Aeronautics and Space Administration (NAG 9 1277) and the Texas Advanced Technology Program. We also thank Prof. Richard Crooks (Department of Chemistry, Texas A&M University) for the use of the contact angle instrument. MVP wishes to thank the Alfred P. Sloan Foundation for its support through a research fellowship.
References M.L. Amirpour, P. Ghosh, W.M. Lackowski, R.M. Crooks, and M.V. Pishko, Analytical Chemistry 73, 1560±1566 (2001). Y.H. An and R.J. Friedman, Journal of Biomedical Materials Research 43, 338±348 (1998). S. Belkin, D.R. Smulski, S. Dadon, A.C. Vollmer, T.K. Van Dyk, and R.A. Larossa, Water Research 31, 3009±3016 (1997).
S.N. Bhatia, M.L. Yarmush, and M. Toner, Journal of Biomedical Materials Research 34, 189±199 (1997). S.N. Bhatia, U. Balis, M. Yarmush, and M. Toner, Biotechnol. Prog. 14, 378±387 (1998). L. Bousse, Sensors and Actuators B 34, 270±275 (1996). C. Chen, M. Mrksich, S. Huang, G. Whitesides, and D. Ingber, Science 276, 1425±1428 (1997). C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, and D.E. Ingber, Biotechnology Progress 14, 356±363 (1998). S.E. Cowan, D. Lipmann, and J.D. Keasling, Biotechnology Letters 23, 1235±1241 (2001). D. Cunliffe, C.A. Smart, C. Alexander, and E.N. Vulfson, Applied and Environmental Microbiology 65, 4995±5002 (1999). E.T. den Braber, J.E. de Ruijter, H.T. Smits, L.A. Ginsel, A.F. von Recum, and J.A. Jansen, Biomaterials 17, 1093±1099 (1996). J.L. Dewez, J.B. Lhoest, E. Detrait, V. Berger, C.C. Dupont-Gillain, L.M. Vincent, Y.J. Schneider, P. Bertrand, and P.G. Rouxhet, Biomaterials 19, 1441±1445 (1998). A. Folch and M. Toner, Biotechnol. Prog. 14, 388±392 (1998). A. Folch and M. Toner, Annual Review of Biomedical Engineering 2, 227±256 (2000). M. Franko, P.F. Nealey, S. Campbell, A.I. Teixeira, and C.J. Murphy, Journal of Biomedical Materials Research 52, 261±269 (2000). P. Ghosh, M.L. Amirpour, W.M. Lackowski, M.V. Pishko, and R.M. Crooks, Angewandte Chemie 38, 1592±1595 (1999). J. Hyun and A. Chilkoti, Journal of the American Chemical Society 123, 6943±6944 (2001). Y. Ito, Biomaterials 20, 2333±2342 (1999). P.S. John, R. Davis, N. Cady, J. Czajka, C. Batt, and H. Craighead, Analytical Chemistry 70, 1108±1111 (1998). D.R. Jung, R. Kapur, T. Adams, K.A. Giuliano, M. Mrksich, H.G. Craighead, and D.L. Taylor, Critical Reviews in Biotechnology 21, 111±154 (2001). R.S. Kane, S. Takayama, E. Ostuni, D.E. Ingber, and G.M. Whitesides, Biomaterials 20, 2363±2376 (1999). W. Koh, A. Revzin, and M.V. Pishko, Langmuir 18, 2459±2462 (2002). J. Lahiri, L. Isaacs, J. Tien, and G. Whitesides, Anal. Chem. 71, 777± 790 (1999). L.H. Larsen, T. Kjaer, and N.P. Revsbech, Analytical Chemistry 69, 3527±3531 (1997). J.-D. Liao, S.-H. Wang, and D.-J. Hsu, Sensors and Actuators B 72, 167±173 (2001). G. Lopez, M. Albers, S. Schreiber, R. Carroll, E. Peralta, and G. Whitesides, J. Am. Chem. Soc. 115, 5877±5878 (1993). L. Lu, K. Nyalakonda, L. Kam, R. Bizios, A. Gopferich, and A.G. Mikos, Biomaterials 22, 291±297 (2001). S.A. Makohliso, D. Leonard, L. Giovangrandi, H.J. Mathieu, M. Ilegems, and P. Aebisher, Langmuir 15, 2940±2946 (1999). T. Matsuda and T. Sugawara, Journal of Biomedical Materials Research 29, 749±756 (1995). J.C. McDonald, D.C. Duffy, J.R. Anderson, D.T. Chiu, H. Wu, O.J.A. Schueller, and G.M. Whitesides, Electrophoresis 21, 27±40 (2000). M.B. Mellott, K. Searcy, and M.V. Pishko, Biomaterials 22, 929±941 (2001). M. Mrksich, C.S. Chen, Y. Xia, L.E. Dike, D.E. Ingber, and G.M. Whitesides, Proceedings of the National Academy of Sciences of the United States of America 93, 10775±10778 (1996). M. Mrksich, L. Dike, J. Tien, D. Ingber, and G. Whitesides, Experimental Cell Research 235, 305±313 (1997). E. Ostuni, R.G. Chapman, M.N. Liang, G. Meluleni, G. Pier, D.E. Ingber, and G.M. Whitesides, Langmuir 17, 6336±6343 (2001a). E. Ostuni, C.S. Chen, D.E. Ingber, and G.M. Whitesides, Langmuir 17, 2828±2834 (2001b).
Control of Mammalian Cell and Bacteria Adhesion on Substrates K.D. Park, Y.S. Kim, D.K. Han, Y.H. Kim, E.H.B. Lee, H. Suh, and K.S. Choi, Biomaterials 19, 851±859 (1998). E.I. Rainina, I.E. Badalin, O.V. Ignatov, A.Y. Fedorov, A.L. Simonian, and S.D. Varfolomeyev, Applied Biochemistry amd Biotechnology 56, 117±127 (1996a). E.I. Rainina, E.N. Efremenko, S.D. Varfolomeyev, A.L. Simonian, and J.R. Wild, Biosensors and Bioelectronics 10, 991±1000 (1996b). A. Razatos, Y.L. Ong, F. Boulay, D.L. Elbert, J.A. Hubbell, M.M. Sharma, and G. Georgiou, Langmuir 16, 9155±9158 (2000). A. Revzin, R.J. Russell, V.K. Yadavalli, W. Koh, C. Deister, D.D. Hile, M.B. Mellott, and M.V. Pishko, Langmuir 17, 5440±5447 (2001). J.H. Silver, J. Lin, F. Lim, V.A. Tegoulia, M.K. Chaudhury, and S.L. Cooper, Biomaterials 20, 1533±1543 (1999). R. Singhvi, A. Kumer, G.P. Lopez, G.N. Stephanopoulos, D.I.C. Daniel, I.C. Wang, G.M. Whitesides, and D.E. Ingber, Science 264, 696±698 (1994a). R. Singhvi, A. Kumar, G. Lopez, G. Stephanopoulos, D. Wang, G. Whitesides, and D. Ingber, Science 264, 696±698 (1994b). K. Sirkar and M.V. Pishko, Analytical Chemistry 70, 2888±2894 (1998). S.J. So®a and E. Merill, Poly(ethylene glycol): Chemistry and Biological Applications, Vol. 680, edited by J.M. Harris, American Chemical Society, Washington D.C., 1997.
19
P. Sommer, C. Martin-Rouas, and E. Mettler, Food Microbiology 16, 50±515 (1999). C.J. Sundberg, Current Opinion in Biotechnology 11, 47±53 (2000). S. Takayama, J. McDonald, E. Ostuni, M. Liang, P. Kenis, R. Ismagilov, and G. Whitesides, Proc. Natl. Acad. Sci. USA 96, 5545±5548 (1999). C.H. Thomas, C.D. McFarland, M.L. Jenkins, A. Rezania, J.G. Steele, and K.E. Healy, Journal of Biomedical Materials Research 37, 81±93 (1997). J. Turner, W. Shain, D. Szarowski, M. Andersen, S. Martins, M. Isaacson, and H. Craighead, Exp. Neurol. 156, 33±49 (1999). E. Tziampazis, J. Kohn, and P.V. Moghe, Biomaterials 21, 511±520 (2000). J.H. Ward, R. Bashir, and N.A. Peppas, Journal of Biomedical Materials Research 56, 351±360 (2001). G. Whitesides, A. Kumar, G. Lopez, H. Biebuyck, P. DiMilla, R. Haerter, and K. Prime, Journal of Cellular Biochemistry 18C, 241, 250, 270 (1994). G. Whitesides, E. Ostuni, S. Takayama, X. Jiang, and D.E. Ingber, Annual Review of Biomedical Engineering 3, 335±373 (2001). M.Q. Zhang, T. Desai, and M. Ferrari, Biomaterials 19, 953±960 (1998). S. Zhang, L. Yan, M. Altman, M. Lassle, H. Nugent, F. Frankel, D. Lauffenburger, G. Whitesides, and A. Rich, Biomaterials 20, 1213± 1220 (1999).
