A Novel Single-Step Fabrication Technique to Create Heterogeneous

Mar 18, 2005 - hydrogel structures, microfluidic devices were rotated 90° from the traditional xy axes and sealed with a cover slip. Two discreet flu...
0 downloads 0 Views 506KB Size
Download PDF
4168

Langmuir 2005, 21, 4168-4174

A Novel Single-Step Fabrication Technique to Create Heterogeneous Poly(ethylene glycol) Hydrogel Microstructures Containing Multiple Phenotypes of Mammalian Cells Jeanna C. Zguris,† Laura J. Itle,†,‡ Won-Gun Koh,† and Michael V. Pishko*,†,§,| Departments of Chemical Engineering, Chemistry, and Materials Science & Engineering and The Huck Institute for the Life Sciences, The Pennsylvania State University, University Park, Pennsylvania 16802-4420 Received December 3, 2004. In Final Form: February 1, 2005 In this study, a novel method for the one-step fabrication of stacked hydrogel microstructures using a microfluidic mold is presented. The fabrication of these structures takes advantage of the laminar flow regime in microfluidic devices, limiting the mixing of polymer precursor solutions. To create multilayered hydrogel structures, microfluidic devices were rotated 90° from the traditional xy axes and sealed with a cover slip. Two discreet fluidic regions form in the channels, resulting in the multilayered hydrogel upon UV polymerization. Multilayered patterned poly(ethylene glycol) hydrogel arrays (60 µm tall, 250 µm wide) containing fluorescent dyes, fluorescein isothiocyanate, and tetramethylrhodamine isothiocyanate were created for imaging purposes. Additionally, this method was used to generate hydrogel layers containing murine fibroblasts and macrophages. The cell adhesion promoter, RGD, was added to hydrogel precursor solution to enhance fibroblast cell spreading within the hydrogel matrix in one layer, but not the other. We were able to successfully generate patterns of hydrogels containing multiple phenotypes by using this technique.

Introduction Cell-based biosensors receive much attention because of their potential application in the detection of chemical and biological toxins as well as their application for high throughput drug screening.1,2 Cells are sensitive to a wide range of compounds; thus resultant sensors allow for detection to be based on cell viability and a wide variety of metabolic changes.3,4 The assessment of viability and metabolic change makes it possible to obtain information about physiological responses of cells to analytes of interest in a living system.3,5 Thus, a cell-based biosensor has the potential to minimize the need for costly experimentation on animals to determine toxicology and the efficacy of drug candidates. One important area of interest in developing a cellbased biosensor that simulates a laboratory animal is mimicking a cell’s in vivo environment. It is important that a cell-based biosensor not only provides a three* To whom correspondence should be addressed. Department of Chemical Engineering, The Pennsylvania State University, 104 Fenske Laboratory, University Park, PA 16802-4420. E-mail: [email protected]. Phone: (814) 863-4810. Fax: (814) 8657846. † Department of Chemical Engineering. ‡ The Huck Institute for the Life Sciences. § Department of Chemistry. | Department Materials Science & Engineering. (1) Khandurina, J.; Guttman, A. Microchip-based high-throughput screening analysis of combinatorial libraries. Curr. Opin. Chem. Biol. 2002, 6 (3), 359-66. (2) Sundberg, S. A. High-throughput and ultrahigh-throughput screening: solution- and cell-based approaches. Curr. Opin. Biotechnol. 2000, 11 (1), 47-53. (3) Bousse, L. Whole cell biosensors. Sens. Actuators, B 1996, 34, 270-275. (4) O’Connor, S. M.; et al. Immobilization of neural cells in threedimensional matrixes for biosensor applications. Biosens. Bioelectron. 2000, 14 (10-11), 871-81. (5) Park, T. H.; Shuler, M. L. Integration of cell culture and microfabrication technology. Biotechnol. Prog. 2003, 19 (2), 243-53.

dimensional extracellular matrix but also permits cell/ cell interactions. Traditionally, cell/cell interactions are measured in in vivo studies or through coculture techniques. Recent in vitro works have focused on the coculture of cells on polymer scaffolds or in multilayered hydrogels.6,7 While these methods provide information about cell/cell interactions in a three-dimensional, in vivo like environment, they often require time-intensive fabrication steps. The use of mammalian cells for high throughput drug screening and biosensing applications also requires the patterning of cells in spatially defined regions. Cell patterning can be accomplished by a variety of methods including direct deposition of cells onto surfaces through microcontact printing, lithography, or directed deposition using microfluidic channels.8-13 Hydrogel microstructures provide an ideal threedimensional, aqueous microenvironment for the encapsulation of biomolecules, including protein, nucleic acids, (6) Yamato, M.; et al. Thermally responsive polymer-grafted surfaces facilitate patterned cell seeding and co-culture. Biomaterials 2002, 23 (2), 561-7. (7) Liu, V.; Bhatia, S. N. Three-Dimensional Photopatterning of Hydrogels Containing Living Cells. Biomed. Microdevices 2002, 4 (4), 257-266. (8) Takayama, S., et al. Patterning cells and their environments using multiple laminar fluid flows in capillary networks. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 5545-5548. (9) Delamarche, E., et al. Microfluidic Networks for Chemical Patterning of Substrates: Design and Application to Bioassays. J. Am. Chem. Soc. 1998, 120, 500-508. (10) Folch, A.; Toner, M. Cellular micropatterns on biocompatible materials. Biotechnol. Prog. 1998, 14, 4 (3), 388-392. (11) Kane, R. S.; Takayama, Shuichi, Ostuni, Emanuele, Ingber, D. E.; Whitesides, G. M. Patterning proteins and cells using soft lithography. Biomaterials 1999, 20 (23-24), 2363-2376. (12) Singhvi, R., et al. Engineering cell shape and function. Science 1994, 264 (5159), 696-8. (13) Amirpour, M. L.; et al. Mammalian cell cultures on micropatterned surfaces of weak-acid, polyelectrolyte hyperbranched thin films on gold. Anal. Chem. 2001, 73 (7), 1560-1566.

10.1021/la0470176 CCC: $30.25 © 2005 American Chemical Society Published on Web 03/18/2005

Stacked Hydrogel Microstructures

and whole cells.14-20 Previously, we have shown that mammalian cells can be entrapped in poly(ethylene glycol) (PEG) microstructures using photolithography and photoreaction injection molding.14,15 Photoreaction injection molding utilizes microfluidic channels to deposit multiple mammalian cell lines simultaneously.15 This permits the fabrication of multiphenotypic arrays in one step, with the number of cell lines determined by the microfluidic mold used, and the number of array elements controlled by the placement of a photomask prior to exposure to ultraviolet light. Multiphenotypic cellular arrays are created with precisely controlled dimensions. This method of using a microfluidic mold to create a singlelayered hydrogel can be easily altered to create a multilayered hydrogel microstructure. Here, we utilize microfluidic channels and photolithography to generate patterns of stacked hydrogels containing multiple phenotypes in one step by utilizing the laminar flow regime common to microfluidic devices.21,22 By simple rotation of the microfluidic channels 90° from the traditional xy axis and introduction of multiple polymer solutions into the channel, a multilayered hydrogel can be created in one polymerization step. The polymer precursor solutions can be altered to contain fluorescent dyes, fluorescent-sensing chemistries, and viable mammalian cells. The small amount of precursor solution needed to fabricate these microstructures provides a distinct economic advantage.23,24 Utilizing a microfluidic device to create these multilayered hydrogels can become automated very easily, and the number of layers and the size of the layers can be changed depending on the ultimate use of the hydrogel structures. Experimental Methods Substrate Preparation. Glass microscope slides were obtained from VWR (Bristol, CT) and cleaned with 6 N sulfuric acid (VWR, Bristol, CT) and subsequently soaked in 0.1 M sodium hydroxide (Aldrich, Milwaukee, WI) for at least 4 h. Slides were then treated with 3-(trimethoxysilyl)propyl methacrylate (TPM) (Aldrich, Milwaukee, WI) in 3:1 mixture of n-heptane and carbon tetrachloride (Aldrich, Milwaukee, WI) for 5 min. Slides were washed with ethanol and deionized water and then dried with air. The silanization technique with TPM leaves free methacrylate (14) Koh, W. G.; Revzin, A.; Pishko, M. V. Poly(ethylene glycol) hydrogel microstructures encapsulating living cells. Langmuir 2002, 18 (7), 2459-2462. (15) Koh, W. G.; Itle, L. J.; Pishko, M. V. Molding of hydrogel microstructures to create multiphenotype cell microarrays. Anal. Chem. 2003, 75 (21), 5783-5789. (16) Russell, R. J.; et al. Poly(ethylene glycol) hydrogel-encapsulated fluorophore-enzyme conjugates for direct detection of organophosphorus neurotoxins. Anal. Chem. 1999, 71 (21), 4909-4912. (17) Seong, G. H.; Zhan, W.; Crooks, R. M. Fabrication of microchambers defined by photopolymerized hydrogels and weirs within microfluidic systems: application to DNA hybridization. Anal. Chem. 2002, 74 (14), 3372-3377. (18) Sirkar, K.; Revzin, A.; Pishko, M. V. Glucose and lactate biosensors based on redox polymer/oxidoreductase nanocomposite thin films. Anal. Chem. 2000, 72 (13), 2930-2936. (19) Zhan, W.; Seong, G. H.; Crooks, R. M. Hydrogel-based microreactors as a functional component of microfluidic systems. Anal. Chem. 2002, 74 (18), 4647-4652. (20) Heo, J.; et al. A microfluidic bioreactor based on hydrogelentrapped E. coli: cell viability, lysis, and intracellular enzyme reactions. Anal. Chem. 2003, 75 (1), 22-26. (21) Kenis, P. J. A.; et al. Fabrication inside Microchannels Using Fluid Flows. Acc. Chem. Res. 2000, 33 (12), 841-847. (22) Kamholz, A. E.; et al. Quantitative Analysis of Molecular Interaction in a Microfluidic Channel: The T-Sensor. Anal. Chem. 1999, 71 (23) 5340-5347. (23) Mitchell, P. Microfluidics - downsizing large-scale biology. Nat. Biotechnol. 2001, 19, 717-721. (24) Zhao, B.; Moore, J. S.; Beebe, D. J. Principles of Surface-Directed Liquid Flow in Microfluidic Channels. Anal. Chem. 2002, 74 (16), 42594268.

Langmuir, Vol. 21, No. 9, 2005 4169 groups on the glass surface which can react with poly(ethylene glycol) diacrylate (PEG-DA) during cross-linking with ultraviolet light exposure.25 Cell Culture. Murine 3T3 fibroblast, SV-40 transformed murine hepatocytes, and murine peritoneal macrophage cell lines were obtained from American Type Culture Collection (Manassas, VA). Cell lines were incubated at 37 °C in 5% CO2 and 95% humidified air. Murine fibroblasts and murine macrophages were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma, St. Louis, MO) with 4.5 g/L glucose and supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic solution (Sigma, St. Louis, MO). SV-40 transformed murine hepatocytes were cultured in DMEM containing 4.5 g/L glucose, 200 nM dexamethasone (Sigma Aldrich, St. Louis, MO), and 4% FBS. Fibroblasts and hepatocytes were cultured to confluence in 75 cm2 polystyrene tissue culture flasks, and confluent cells were subcultured every 2 to 3 days by trypsinization with 0.25% (w/v) trypsin and 0.13% (w/v) EDTA. Confluent murine macrophages were subcultured every 2 to 3 days by cell scraping. Preparation of Microfluidic Channels for Photoreaction Injection Molding. Microfluidic networks were formed from a 10:0.09 mixture of prepolymer and curing agents (Dow Corning Sylgard 184, Midland, MI). The resulting degassed mixture was poured over a master and cured at 60 °C for at least 2 h. The masters had negative micropatterns prepared using photolithography using SU-8 50 negative photoresist (Microlithography Chemical Corp, Newton MA) and a chrome sodalime photomask made by Advanced Reproductions (Andover, MA). To create the photoreaction injection mold, the master was placed in a Teflon mold. This mold was the width of the master, which ensured a clean edge for the microfluidic channel. Additionally, to preserve the edge of the microfluidic channel, coverslips were placed between the master and wall of the mold. After curing, the PDMS replica was removed from the master and the inlet and outlet ports were placed by piercing the replica with a blunted syringe needle through the backside of the network. The PDMS molds were used to make the stacked hydrogel microstructures. Hydrogel Preparation. Prior to hydrogel preparation, cells for encapsulation were stained with CellTracker Green CMFDA, CellTracker Blue CMAC, and CellTracker Orange CMTMR (Molecular Probes, Eugene, OR) for fluorescent imaging. Hydrogel precursor solutions for cell encapsulation were prepared using 10% v/v PEG-DA (MW 575, Aldrich, Milwaukee, WI) in phosphate-buffered saline (PBS) or serum-free cell culture media and 0.1% v/v 2-hydroxy-1-[4-(hydroxyethoxy)phenyl]-2-methyl-1-propanone (Irgacure 2959, Ciba, Tarrytown, NY) as a photoinitiator in ethanol. Cells were resuspended in sterile polymer precursor solution at a density of 1.0 to 2.0 million cells/mL. The Gly-Arg-Gly-Asp-Ser (GRGDS) (Calbiochem., San Diego, CA) adhesion peptide was incorporated into the PEG hydrogels. The peptide was conjugated to PEG by reacting the peptide with acryloyl-PEG-n-hydroxysuccinimide ester (acryloyl-PEG-NHS, MW 3400, Huntsville, AL) as previously described.15,26 Briefly, the peptide was dissolved to a concentration of 1 mg/mL in cell culture media and used to prepare a 10% w/v solution of acryloylPEG-NHS. The solution was allowed to react at room temperature for at least 2 h, before being incorporated into the hydrogel precursor solution described above. In the case where cells were not used, fluorescein isothiocyanate (FITC) and tetramethylrhodamine isothiocyanate (TRITC) (Aldrich, Milwaukee, WI) were used to distinguish the two layers. The precursor solutions were composed 10% v/v PEG-DA (MW 575, Aldrich, Milwaukee, WI) in PBS with 1% of 2-hydroxy-2methyl-1-phenyl-1-propanone (Darocur 1173, Ciba, Tarrytown, NY). The hydrogel microstructures were observed under an Axiovert Zeiss 200M fluorescent microscope with a mercury light source and FITC (exciter 480 ( 20 nm, emitter 535 ( 25 nm, beam splitter 505 long band-pass), TRITC (exciter 545 ( 15 nm, emitter 620 ( 30 nm, beam splitter 570 long band-pass), DAPI (25) Revzin, A.; et al. Fabrication of poly(ethylene glycol) hydrogel microstructures using photolithography. Langmuir 2001, 17 (18), 54405447. (26) Burdick, J. A.; Anseth, K. S. Photoencapsulation of osteoblasts in injectable RGD-modified PEG hydrogels for bone tissue engineering. Biomaterials 2002, 23 (22), 4315-4323.

4170

Langmuir, Vol. 21, No. 9, 2005

Zguris et al.

Figure 1. The fabrication of heterogeneous gels. (A) Side-by-side hydrogels are generated utilizing a two-step process in which the first precursor solution is placed in the microfluidic channel, polymerized, and removed. The second precursor solution is introduced into the channel and the same process is repeated. (B) One-step fabrication of stacked hydrogels. The PDMS mold is placed on a silanization surface and sealed with a coverslip. The precursor solutions are pumped into the chamber at the same flow rates and then cross-linked utilizing a UV light source. The mold can then be removed, and the excess precursor solution can be rinsed away with water. (exciter 360 ( 20 nm, emitter 460 ( 25 nm, beam splitter 400 long band-pass) filter cube sets (Zeiss, Thornwood, NY). Confocal images were obtained using an Olympus Fluoview confocal laser scanning microscope with UplanFL 40×/0.75 objective and an external halogen lamp for brightfield viewing (Olympus, Melville, NY). Fabrication of Heterogeneous Hydrogel Microstructures. Heterogeneous microstructures were generated using photoreaction injection molding. Microfluidic channels were filled with an initial precursor solution through loading with a microfluidic pump. A 100 µm circular photomask (Advanced Reproductions, Andover, MA) was positioned over the filled channels; polymer exposed to ultraviolet light for 30 s crosslinked to form an indissoluble array of cylindrical elements. The channels were flushed to rinse out any unreacted polymer precursor solution. A second polymer precursor solution was introduced into the channel, surrounding the hydrogel cylinders. A second, 200 × 100 µm rectangular photomask was positioned over the 100 µm cylinders, and UV light was shown through the mask for 30 s. The microfluidic mold was removed and unreacted precursor solution was rinsed away with water to reveal a heterogeneous hydrogel (Figure 1A). Fabrication of Stacked Hydrogel Microstructures. To create a mold for the fabrication of layered hydrogels, the PDMS mold was placed on a surface-modified substrate on its edge (flipped 90° from a normal channel placement). A cover slip was used to seal the channel creating the fourth wall. For dual layer hydrogel microstructures, solutions were introduced simultaneously into microfluidic chambers by using a syringe pump (PHD 2000, Infuse/Withdraw Pump, Harvard Apparatus, Holliston, MA) to apply equal flow rates to the solutions. Flow was discontinued when microfluidic channels were filled with the polymer precursor solution. A UV light was placed on the solution

for 2-5 s causing gelation. The mold was then removed, and the unreacted precursor solution was rinsed away with water (Figure 1B).

Results and Discussion Fabrication of Multiphenotypic Heterogeneous Hydrogel Microstructures. Heterogeneous hydrogel microstructures were generated as described above, using fluorescently stained fibroblasts (green) (Figure 2a) and hepatocytes (blue) (Figure 2b). This allows for the simulation of a native tissue environment in which multiple cell phenotypes can interact with one another, by patterning two different phenotypes in discrete spatial locations. However, this method of forming hydrogel microstructures still requires two distinct UV-polymerization steps. The first precursor solution containing cells, in this case fibroblasts, is added to the microfluidic channel and patterned and unreacted polymer is rinsed away. The second precursor solution is added, and a photomask must be precisely aligned over the original pattern before the second UV-polymerization step. The two distinct steps to generate a heterogeneous gel is a more time intensive process and increases the amount of UV light to which the cells are exposed. Additionally, cells in vivo exist in three-dimensional layers. Previously, to generate multiphenotypic stacked hydrogels, researchers have utilized a similar multistep fabrication method for tissue engineering developed by Liu and Bhatia.7 As with our side-by-side method, each cell layer is patterned

Stacked Hydrogel Microstructures

Langmuir, Vol. 21, No. 9, 2005 4171

Figure 2. A heterogeneous hydrogel microstructure encapsulating murine hepatocytes (blue) and murine fibroblasts (green) taken with different filter sets: (a) rectangular heterogeneous structures fabricated by laminar flow of cell-containing precursor solution inside a microchannel; (b) circle elements encapsulating hepatocytes were fabricated first and then surrounded with fibroblastcontaining rectangular hydrogel microstructures.

individually, with multiple applications of polymer precursor solutions and multiple ultraviolet light exposure times. A multistep method leads to longer ultraviolet light exposure times for cells and increases in the volume of solution needed for fabrication. To overcome these limitations, we utilized the laminar flow properties of microfluidic channels but altered the orientation of the microfluidic mold in relation to the substrate, thus generating a multilayered microstructure containing multiple phenotypes. Fabrication of PDMS Microfluidic Channels for Fabrication of Multilayered Hydrogels. The dimensions of microstructures obtained in this method are dependent on the template used to prepare the microfluidic mold. In this case, a master was made out of a negative photoresist, SU-8, on the edge of a substrate. The width of the microfluidic channel is controlled by controlling the film thickness of the photoresist, which can be affected by multiple variables, such as spin coating speed, soft- and postbake time, amount of UV exposure, and development time.27 The placement of the photoresist template on the edge of a silicon wafer or glass slide can cause some distortion of the feature due to edge effect commonly associated with photolithography and photoresist. This translated into slight distortions in the PDMS mold.28 This distortion in the mold can disturb the flow pattern of the precursor solutions that would affect the hydrogel structures.29 A more uniform template could be achieved by etching the feature into either a glass or silica substrate by using simple fabrication techniques. Etching the feature would also give better control on the height and width of the template, resulting in a more uniform mold. However, the use of photoresist in this case does not significantly hinder the creation of multilayer hydrogels using photoreaction injection molding and provides proof-of-concept for this technique. The mold used had three inlet ports, but only two of the ports were used when making the feature. All three ports may be used to create thinner layers or more layers. In the cases presented in this paper, the precursor solutions were identical, except for the addition of different cells or fluorescent dyes. This kept the viscosity the same, which in turn kept the velocity of the fluid through the channel the same. Keeping velocity the same in filling the mold (27) Microchem, Nano SU-8 Negative Tone Photoresists: Formulations 50 and 100; Microchem: Newton, MA, 2001; p 4. (28) McDonald, J. C.; Whitesides, G. M. Poly(dimethylsiloxane) as a Material for Fabricating Microfluidic Devices. Acc. Chem. Res. 2002, 35 (7), 491-499. (29) Hu, Y.; Werner, C.; Li, D. Electrokinetic Transport through Rough Microchannels. Anal. Chem. 2003, 75 (21), 5747-5758.

Figure 3. A side view of a stacked hydrogel microstructure. Panel A shows a brightfield image of the entire hydrogel structure, while panel B shows a layer containing FITC and panel C shows a TRITC layer.

is important to keep the layers the same size and predicting when the mold would be filled. Additionally, due to the low Reynold’s number in the microfluidic channels,30,31 fluids introduced into the channels through separate inlet ports will mix only through Brownian motion, thus keeping distinct fluid layers over long distances.32 By use of similar precursor solutions, any transport due to interfacial electrokinetic and intermolecular interactions is minimized.33 The result is a one-step method of fabricating two or more stacked layers of PEG hydrogel in intimate contact with one another. Several mold filling flow rates were examined: 5, 15, 25, 30, 40, and 80 µL/min. The flow rates that were the best for creating these features was 25-40 µL/min. The flow rate varies depending on the variations in the molds that are used. The normal ratio of curing agent to poly(dimethylsiloxane) (PDMS) prepolymer is 1:10. In this case, less curing agent is used to create a mold that is soft and has some adhesive properties that can be used to seal the mold with the cover slip. (30) Brody, J. P., et al. Biotechnology at Low Reynolds Numbers. Biophys. J. 1996, 71, 3430-3441. (31) Purcell, E. M. Life at Low Reynolds Number. Am. J. Phys. 1977, 45, 3-111. (32) Lutz, B. R.; Chen, J.; Schwartz, D. T. Microfluidics without microfabrication. Proc. Natl. Acad. Sci. U.S.A. 2003, 100 (8), 43954398. (33) Li, B.; Kwok, D. Y. Discrete Boltzman Equation for Microfluidics. Phys. Rev. Lett. 2003, 90 (12), 124502-1-124502-4.

4172

Langmuir, Vol. 21, No. 9, 2005

Zguris et al.

Figure 4. A top down view of a multilayer hydrogel microstructure encapsulating murine fibroblasts (blue) and murine macrophages (orange). (a) The bottom layer of the microstructure was augmented with RGD to enhance cell spreading (10× magnification). (b) Macrophages are visible in the upper layer, which contained no RGD (10× magnification). (c) After 24 h, visible cells are seen spreading in the bottom layer, while no spread cells are seen in the upper layer (40× magnification).

Multilayered Hydrogel Microstructures with Fluorescent Dyes. The multilayered hydrogel structures that were first created were made out of poly(ethylene glycol) diacrylate with two different fluorescent dyes, FITC and TRITC. The precursor solutions were the same composition so the convective transport of the solution was the same. The mold was exposed to UV light to promote crosslinking by a free radical reaction. The line hydrogel structure that was obtained was removed from the surface so the interface could be seen (Figure 3). The image shows the brightfield image of the hydrogel line with two distinct layers (Figure 3A). The individual layers were imaged under two different filter sets under the microscope, FITC (Figure 3B) and TRITC (Figure 3C). In the TRITC picture (Figure 3C), there is some activation in the FITC layer due to the fact that these dyes are a fluorescent resonance energy transfer pair. This lends to overlap in the emission and excitation wavelengths, which can lead to the appearance of FITC in the TRITC layer. Of larger interest is the optics with which the image was collected. Due to the instrumentation used, the automatic exposure time utilized for image acquisition reveals a high degree of internal reflection within the hydrogel. An evanescent wave is propagated through the hydrogel, which results in the appearance of TRITC at the surface of the hydrogel, in this case the FITC layer. The layers were not completely uniform; this could be due to the fact that the structure was removed from the substrate and was dehydrating. It was probably also due to the uneven mold that was formed from the slight distortion of the master by using the photoresist on the edge. To counteract some of these nonuniformities, primarily due to the use of a photoresist master, the master can be made by silicon-etching techniques. Additionally,

by carefully controlling fluid compositions and flow velocities, nonuniformities may also be minimized. Another area for slight variations in the layers of the hydrogel is due to the mixing of FITC and TRITC at the interface. The time (t) it takes for the dyes to fully mix is proportionate to the inverse of the diffusion coefficient (Deff) and the square of the width of the chamber (l)34

t ∝ l2/Deff

(1)

This information can be easily obtained, and the easiest way to alter the diffusion for minimal mixing to take place is by having a smaller channel length or change the velocity of the fluid. The length of the channel that was used was relatively small, and the beginning of the channel was polymerized to minimize this effect. Fabrication of Multilayered, Multiphenotypic Hydrogel Microstructures. Cell-containing precursor solutions were prepared as described above and used to form hydrogel microstructures. In Figure 4, RGD was added to the precursor solution used to resuspend fibroblast cells (Figure 4a), while no RGD was added to the layer containing macrophages (Figure 4b). Fibroblast spreading was easily seen in the layer containing RGD (Figure 4c), while no spreading was apparent in the layer containing macrophages, indicating that distinct extracellular environments can be maintained in different layers and confirming that no mixing of the precursor solution occurred during the fabrication procedure. The uneven distribution of cells within each layer is most likely due to cell settling and aggregation within the syringe prior to injection into microfluidic channels during (34) Johnson, T. J.; Ross, D.; Locascio, L. E. Rapid Microfluidic Mixing. Anal. Chem. 2002, 74 (1), 45-51.

Stacked Hydrogel Microstructures

Langmuir, Vol. 21, No. 9, 2005 4173

Figure 5. DIC images of hydrogel microstructures encapsulating murine fibroblasts and murine macrophages. (A) Macrophage cells are present in the uppermost layer, while fibroblasts are present in the bottom layer (B) as shown by white arrows. (C) Macrophage cells are shown in the uppermost layer, while fibroblast cells are present in the bottom layer (D). RGD was added to the fibroblast layer in (D), but not in (B), as evidenced by the enhanced cell spreading in (D). Panels E and F are 40× DIC images of a dual layer hydrogel. The cell number in each precursor solution was varied, such that the top layer shown in (E) has significantly fewer cells than that precursor solution used for (F).

the fabrication procedure, a problem easily overcome by the addition of a nontoxic surfactant to the precursor solution.35,36 Because of the short time (less than 2 min) between filling of the channels and UV polymerization, cells do not have time to settle within the microfluidic

channels. Additionally, while the cells were resuspended at the same density, more macrophages are visible than fibroblasts. Because of the three-dimensional nature of the structure, it was difficult to distinguish between the two layers using fluorescent microscopy. Thus, the top

(35) Toth, K.; Wenby, R. B.; Meiselman, H. J. Inhibition of polymerinduced red blood cell aggregation by poloxamer. Biorehology 2000, 37 (4), 301-312.

(36) Armstrong, J. K.; et al. Modulation of red blood cell aggregation and blood viscosity by the covalent attachment of Pluronic copolymers. Biorehology 2001, 38 (2-3), 239-247.

4174

Langmuir, Vol. 21, No. 9, 2005

layer was more readily seen, giving the appearance that macrophages had a higher seeding density than fibroblasts. Therefore, confocal microscopy was explored as a method of concretely resolving the different layers of the hydrogel structures. Figure 5 shows DIC images of hydrogel microstructures without RGD in either layer (Figure 5A,B) and microstructures containing no RGD in one layer (Figure 5C) and RGD in the other layer (Figure 5D). Because confocal microscopy allows us to readily distinguish between different focal planes, the multilayered structure was more readily seen. In parts A and B of Figure 5, the change in width of the structure serves as an identifying mark in the xy plane, the differing positions and quantity of cells in each layer (shown by white arrows) indicating a change in the z direction. In Figure 5D, cell spreading is readily apparent in one layer, but not the other (Figure 5C) indicating that RGD was present in one layer and not the other. To ensure that this was in fact a function of RGD inclusion and not differing cell types, the experiment was repeated with RGD in neither layer (Figure 5A,B). No cell spreading was observable in either layer. Higher magnifications of cells were included (Figure 5E,F) to show the presence of cells in each layer. A higher initial cell density was used in the precursor solution used to make the bottom layer (Figure 5F), while a lower cell density was used to make the top layer (Figure 5E). Finally, we were able to pattern stacked hydrogels by the insertion of a photomask between the UV-light source and the PDMS mold. In Figure 6, we see patterned hydrogels containing multiple fluorescent dyes. The bottom layer contains TRITC (Figure 6A), while the top layer contains FITC (Figure 6B). This illustrates the feasibility of generating spatially defined multilayer hydrogels. Conclusion In this study, we presented two methods for the fabrication of heterogeneous hydrogel structures containing multiple phenotypes. The first method utilized a twostep photoreaction injection molding process, while the second employed a novel one-step fabrication method for multilayered hydrogels. Both methods of creating heterogeneous structures utilized the advantages of microfluidic devices for hydrogel molding, including the small amount of material needed to fabricate the structures, the generation of micrometer scale, spatially defined structures, and inexpensive and easy fabrication. Ad-

Zguris et al.

Figure 6. Fluorescent images of patterned, multilayered hydrogel structures, 1 mm in height: (A) The top layer contains TRITC (red), while the bottom layer (B) contains FITC (green).

ditionally, the creation of multilayered heterogeneous structures has the added benefit of a single fabrication step, minimizing cellular exposure to ultraviolet light. The single-step method was used to generate multilayered hydrogel microstructures containing different fluorescent dyes and was extended to the encapsulation of two distinct cell phenotypes. Differing extracellular matrixes, with or without the cell adhesion promoter, RGD, could be created within one hydrogel microstructure. We were also able to successfully generate patterns of hydrogels containing multiple phenotypes. While only dual layered hydrogel microstructures were created, this technique can be easily extended to the creation of many layered hydrogels by adding more inlet ports to the mold. The addition of more ports would allow for additional precursor solutions to be introduced into the microfluidic mold, thus adding more layers to a single structure. Acknowledgment. The authors wish to thank NASA and the Commonwealth of Pennsylvania Center for Optical Technologies for providing funding for this research. J.C.Z. and L.J.I. wish to thank Elaine Kunze and Susan Magargee at the Center for Quantitative Cell Analysis, The Huck Institute for Life Sciences, Pennsylvania State University for assistance with confocal microscopy. LA0470176