Micropatterning of Polymer Brushes: Grafting from Dewetting Polymer

Aug 10, 2012 - We anticipate this approach will be suitable for the patterning of brushes, especially for biomedical applications such as in the study...
0 downloads 9 Views 7MB Size
Article pubs.acs.org/Biomac

Micropatterning of Polymer Brushes: Grafting from Dewetting Polymer Films for Biological Applications A. M. Telford,†,§ L. Meagher,*,‡ V. Glattauer,‡ T. R. Gengenbach,‡ C. D. Easton,‡ and C. Neto*,† †

School of Chemistry, The University of Sydney, F11, NSW 2006 Australia CSIRO Materials Science and Engineering, Bag 10, Clayton South, Victoria 3169, Australia § CSIRO Future Manufacturing National Research Flagship, Clayton, Victoria 3168, Australia ‡

S Supporting Information *

ABSTRACT: In this novel platform, a micropatterned polymer brush was obtained by grafting poly(poly(ethylene glycol) methyl ether methacrylate) (poly(PEGMA)) from a thin macroinitiator film using atom transfer radical polymerization (ATRP). A pattern of holes was formed in the macroinitiator film by taking advantage of its spontaneous dewetting above the glass transition temperature from a bottom polystyrene film, driven by unfavorable intermolecular forces. Patterning by dewetting can be achieved at lengthscales from a few hundred nanometers to several tens of micrometers, by simply thermally annealing the bilayer above the glass transition temperature of the polymer. This approach is substrate-independent, as polymer films can be cast onto surfaces of different size, shape, or material. As a demonstration of its potential, proteins, and individual cells were attached on targeted bioadhesive polystyrene areas of the micropatterns within poly(PEGMA) protein-repellent brushes. We anticipate this approach will be suitable for the patterning of brushes, especially for biomedical applications such as in the study of single cells and of cell cocultures.



INTRODUCTION Grafted polymer brushes are widely used as a chemically robust and versatile way to tailor the functional properties of surfaces, such as wettability, biocompatibility, corrosion resistance, and friction.1,2 Recently, the patterning of grafted polymer brushes on surfaces has attracted great attention; patterned brushes combine tailored chemical functionality with topographical features,3 which is very useful in a wide range of applications in the biological and medical fields, all benefiting from the ability to control the position and scale of surface domains that are adhesive to biological material (proteins, DNA, cells).4 The control of cell density and spreading on a surface, which in turn controls the mechanical stress within the cell, can help to direct their fate toward death, proliferation and even differentiation, which is the basis of tissue engineering in vitro.5,6 Furthermore, the control over cell localization can lead to the manipulation of the degree of homotypic and heterotypic contact between cells in cocultures.4,7 In different applications, the isolation of single cells or small colonies of cells in arrays of separated microbioreactors (microwells) can allow fundamental studies on the interaction of a cell with its microenvironment and with its neighboring cells. This is particularly important for growing stem cells in populations that maintain their pluripotency,8,9 which is still a challenging objective, as the field’s understanding of the mechanisms that govern their natural tendency to specialize on a surface is still poor. We present an approach for creating micropatterned grafted polymer brushes with regions of chemical and topographical © 2012 American Chemical Society

contrast, combining the spontaneous dewetting of bilayers of thin polymer films with polymer grafting-from via atom transfer radical polymerization (ATRP). The resulting polymer coating was successfully employed to pattern proteins and cells on a surface. The most successful method for the covalent tethering of polymer chains onto surfaces in a brush conformation is the “grafting-from” technique.10 Polymerization initiators are generally covalently attached to a solid substrate, and polymerization of chains is initiated through the formation of radical species onto which monomers can add. The patterning of polymer brushes, especially for biomedical applications, has been carried out using a wide range of methods, such as patterning the surface initiators by microcontact printing11 or photopolymerization through a mask,12,13 and by scanning probe microscopy-assisted writing.14 The success of these methods is counteracted by the need for lengthy lithographic or scanning procedures and for specialized and expensive equipment to fabricate masks and molds. Dewetting, on the other hand, is a spontaneous patterning process by which unstable thin (≈100 nm) liquid films break apart on a substrate, driven by unfavorable intermolecular forces at the interface.15,16 The film breaks apart into holes that grow with time, eventually transforming into a series of isolated droplets. Unstable Received: July 9, 2012 Revised: August 8, 2012 Published: August 10, 2012 2989

dx.doi.org/10.1021/bm3010534 | Biomacromolecules 2012, 13, 2989−2996

Biomacromolecules

Article

tube was then immersed in an oil bath at 65 °C and left to react with stirring for 22 h. The reaction was stopped by cooling down the vial to room temperature and exposing the contents to air. The viscous product was diluted in toluene and precipitated from cold methanol. The polymer powder was dried under vacuum at room temperature. 1 H NMR and size exclusion chromatography (SEC), calibrated with poly(methyl methacrylate) standards in THF, were used to characterize the polymer chains in solution. The thermal behavior of the macroinitiator was studied by differential scanning calorimetry (DSC). Details on the characterization are reported in the Supporting Information (SI), Figures S1−3. Sample Preparation. Sample preparation was performed in a class-100 laminar flow cabinet. All tools and substrates were thoroughly cleaned to remove dust and organic contaminants, as described elsewhere.26 Polystyrene (PS, Mn 96000 g mol−1, PDI 1.04; Polymer Standards, Germany) was spin-cast in films on clean silicon wafers from a 15 mg mL−1, filtered toluene solution (3000 rpm for 1 min). Macroinitiator films were then spin-cast on top of the PS films from a 10 mg mL−1 acetonitrile (>99.9%, Merck) solution (4000 rpm for 1 min), as acetonitrile is a nonsolvent for PS. The average thickness of the films obtained was 58 ± 1 nm for PS and 28 ± 1 nm for the macroinitiator, as determined by spectroscopic ellipsometry. Both asprepared films were smooth, with root-mean-square (rms) roughness values less than 0.5 nm, as measured by atomic force microscopy. Where required, the bilayers were dewetted by annealing on a hot plate with fine temperature control (ATV Technologie GmbH, Muenchen) in air to obtain the desired pattern of holes. The annealing temperature used was 100 °C. The dewetting process was monitored by optical microscopy (Nikon LV-150). Grafting of Poly(PEGMA) Brushes. The grafting of poly(PEGMA) from the macroinitiator film was carried out using AGET ATRP. In a typical run, PEGMA (1.425 g, 3 mmol), CuCl2·H2O (3.4 mg, 0.02 mmol), tris-(pyridyl methyl) amine (TPMA; 5.8 mg, 0.02 mmol; Sigma-Aldrich), ethyl 2-bromopropanoate (3.6 mg, 0.02 mmol), and isopropyl alcohol/water 50% (12.7 mL) were thoroughly mixed in a 22 mL glass vial. Isopropyl alcohol (IPA; >99.5%), CuCl2·H2O (>99%), and ethyl 2-bromopropanoate were purchased from Merck. PEGMA (Mn = 475 g mol−1) was purified using activated basic alumina just before use. The silicon wafer coated with the macroinitiator/PS bilayer was introduced in the mixture, and the vial was sealed with a rubber septum. The mixture was deoxygenated by purging with nitrogen for 30 min at 0 °C. After equilibrating at room temperature, the reducing agent was added using a microsyringe. The reaction was stopped by exposing the catalyst to air and quenching in ice. Size exclusion chromatography (SEC) calibrated with poly(methyl methacrylate) standards in DMF was used to characterize the free polymer chains in solution (details in the SI). Surface Characterization. XPS analysis was performed using an AXIS Ultra DLD spectrometer (Kratos Analytical Inc.) with a monochromated Al Kα source at a power of 150 W (15 kV × 10 mA), a hemispherical analyzer operating in the fixed analyzer transmission mode and the standard aperture (analysis area: 0.3 × 0.7 mm). Each specimen was analyzed at an emission angle of 0° as measured from the surface normal. Assuming typical values for the electron attenuation length of relevant photoelectrons in organic compounds, the XPS analysis depth (from which 95% of the detected signal originates) was between 5 and 10 nm. All elements present were identified from survey spectra (acquired at a pass energy of 160 eV). To obtain more detailed information, such as chemical structure and oxidation states, high resolution spectra were recorded from individual peaks at 40 eV pass energy (yielding a typical peak width for polymers of 1.0−1.1 eV). Binding energies were referenced to the aliphatic hydrocarbon peak at 285.0 eV. Further detail on the XPS composition analysis can be found in the SI. XPS images were acquired at a pass energy of 160 eV using the field of view 2 lens (∼400 × 400 μm) in high resolution mode. Each image was collected at 530 eV (O 1s) and the background image collected at 515 eV. Ellipsometry characterization was performed on a multiwavelength J. A. Woollam Co. Inc. M-2000 instrument, using an angle of incidence

polymer films spontaneously dewet when heated above the glass transition temperature, Tg. When cooled below Tg, the resultant pattern is “frozen in,” preserving the desired topography. A few different strategies for grafting polymer brushes from polymeric films have been reported in the literature.17−24 The approach has the advantage that brushes can be patterned on surfaces of any material and on nonflat objects. This is the first time that dewetting of a macroinitiator film has been carried out, so that the polymer brushes could be spontaneously patterned. Figure 1 is a schematic illustration of our approach,

Figure 1. Schematic showing the patterning approach for proteins and cells. (1) A bilayer of polymers is formed: a film of the macroinitiator (bearing ATRP initiating groups) is spin-coated on top of a polystyrene (PS) film. (2) The macroinitiator film dewets upon thermal annealing, to form holes. (3) Poly(PEGMA) chains are grafted from the macroinitiator film, producing a patterned brush. (4) The patterned brush is exposed to a protein solution, which adsorb only within the PS holes and facilitate attachment of fibroblasts within the holes.

consisting of four simple steps: formation of a bilayer of two polymer films, dewetting of the top film, grafting of brushes on the patterned top film by surface-initiated ATRP, and adsorption of proteins and cells from solution. ATRP is a reversible-deactivation polymerization technique that can produce polymers with controlled molecular weight, dispersity and architecture, and bearing “living” chain ends, which can be readily functionalized.25 Our system was designed to induce the controlled and selective adsorption of proteins and cells on adhesive domains within patterns (step 4, Figure 1). We investigated the interaction of extra-cellular matrix proteins with the developed patterns, and the subsequent attachment of cells onto the protein-coated domains.



MATERIALS AND METHODS

Synthesis of the Macroinitiator. A dry 4 mL glass tube was charged with 2-(2-bromopropanoyloxy)ethyl methacrylate (0.7 mL, 3.77 mmol), toluene (0.7 mL; >99.9%, Sigma-Aldrich), doubly recrystallized azobis(isobutyronitrile) (AIBN; 0.6 mg, 3.6 μmol), and 2-cyanopropan-2-yl dithiobenzoate (CPDB; 8.0 mg, 36.3 μmol). The tube was sealed with a rubber septum, and the mixture was deoxygenated by bubbling nitrogen for 5 min at 0 °C. The glass 2990

dx.doi.org/10.1021/bm3010534 | Biomacromolecules 2012, 13, 2989−2996

Biomacromolecules

Article

The dewetting of the macroinitiator film on top of a PS film was typical for a bilayer system (Figure 2). At the temperature

of 75° in a wavelength range of 370−1000 nm. The data was collected in at least three points on each sample and averaged. Atomic force microscopy characterization was performed on an Asylum Research MFP-3D, using Tapping Mode in air. The data was collected in at least three points on each sample and averaged. Biological Studies. Bovine serum albumin and human fibrinogen were tagged with fluorescein isothiocyanate (FITC, all Sigma Aldrich) using the procedure recommended by Invitrogen Molecular Probes. A drop (about 100 μL) of the 1−2 mg mL−1 solution was deposited on each patterned surface, and left to incubate for 3−4 h at room temperature in the dark. The surfaces were then rinsed with PBS buffer to remove the nonadsorbed protein, then in Milli-Q water to remove excess salts from the buffer used, and finally gently blown dry with nitrogen. The presence of adsorbed protein was then established by optical fluorescence microscopy (Olympus BX61) within 24 h. Visualization of the patterns was carried out using both visible light, and in fluorescence mode (excitation wavelength 494 nm, emission wavelength 518 nm, exposure time 10 s). For cell experiments, more robust coatings were employed. The PS used had much higher molecular weight (Mn of 3850000 g mol−1), and both the PS and macroinitiator films were thicker (approx 130 and 72 nm, respectively). L929 mouse fibroblasts (cell line ATCC−CCL-1, Rockville, MD) were used to investigate attachment and spreading on the samples. Cells were cultured in minimum essential medium containing 10% fetal bovine serum and 1% nonessential amino acids. Where vitronectin precoating was required, the samples were covered with 50 μL of bovine vitronectin solution in PBS (600 μg mL−1) and incubated at 37 °C for 4 h, then rinsed with PBS. All samples were soaked in phosphate buffer saline (PBS) containing double strength antibiotic-antimycotic reagent (Invitrogen) for 2 h at room temperature prior to cell culture. Where prestaining of cells was performed, a suspension of cells at a density of 1 × 105 cells mL−1 was reacted with a fluorescent DiLC12 membrane stain (Molecular Probes) at 1 μg mL−1 for 1 h at 37 °C. Cells were then seeded at 2.8 × 104 cells cm−2 in a nontreated tissue culture plate (Nunc, Denmark) containing the patterned sample, which was then incubated at 37 °C, 5% CO2 in air for 24 h. For the live/dead assay (Molecular Probes), cells were seeded on the substrate as above, and the assay carried out after 24 h using staining with calcein and ethidium homodimer at concentrations of 2 and 4 μM, respectively. The samples were imaged using a fluorescent Nikon inverted microscope (Eclipse T2000-U).

Figure 2. Optical micrographs of the dewetting of a 28 ± 1 nm thick macroinitiator film on top of a 58 ± 1 nm thick PS substrate, upon annealing at 100 °C. The time from the start of annealing is indicated in each micrograph. The holes nucleate all at the same time and grow at a rate determined by the annealing temperature. The desired pattern can be “frozen-in” by simply interrupting the annealing.

employed, the dewetting was relatively quick (less than 5 min to produce the surfaces employed for the patterning of proteins and cells). Holes in the macroinitiator film appeared in a random distribution on the PS substrate (Figure 2a), grew with time (Figure 2b), coalesced with neighboring holes (Figure 2c), and the liquid polymer ribbons eventually ruptured into droplets (Figure 2d). The dewetting macroinitiator film retracted from the substrate, exposing the underlying PS substrate, and accumulated in raised rims around the holes (light colored rims around the holes in Figure 2). The hole size distribution was narrow at any one time (relative variation = 6%), as the holes all nucleate at the same time.28 Similar results may be achieved at relatively low temperatures (just above approximately 51 °C) but in longer annealing times. This becomes relevant in the preparation of these coatings onto devices that are sensitive to high temperatures (e.g., plastic medical implants or catheters). Ordered patterns can be produced by dewetting,29 for example, by guiding the dewetting through the use of a stamped self-assembled monolayer.30 Nevertheless, for the applications summarized in the introduction, such as cell studies, growth of cell cocultures for tissue engineering, and for the control of attached cell density onto the surface of implantable biomaterials, lateral order is often not necessary. By dewetting, hole size, density, and shape can be easily controlled by controlling dewetting time, preannealing and polymer molecular weight, and film thickness, respectively.31 Grafting of Poly(PEGMA) Brushes from Macroinitiator Films. When surface-initiated activators generated by electron transfer atom transfer radical polymerization (AGET ATRP) was used,25 poly(poly(ethylene glycol)methyl ether methacrylate) (poly(PEGMA)) chains were grafted from the dewetted macroinitiator film in a rapid reaction lasting approximately 30 min in an isopropanol/water mixture. AGET ATRP is a particularly robust and simple process in the ATRP family, as it only requires stable forms of the copper catalyst, has little



RESULTS AND DISCUSSION Patterning of Macroinitiator Films by Dewetting. The macroinitiator employed in this work was synthesized by reversible addition−fragmentation chain transfer polymerization (RAFT) to control the molecular weight and dispersity, which strongly influence the density of dewetted holes. RAFT is a robust and straightforward technique to produce welldefined polymeric architectures.27 The macroinitiator, with an average molecular weight (Mn) of 20700 g mol−1 and a polydispersity index (PDI) of 1.28, was designed to be structurally similar to poly(methyl methacrylate), which is known to readily dewet from PS.16 The reaction was stopped at approximately 83% conversion (determined by NMR) to avoid termination by a combination of two growing chains, which would lead to the formation of a second population of polymer chains, with about double the molecular weight. The RAFT reaction scheme, the NMR structure of the macroinitiator, and differential scanning calorimetry data are reported in Figures S1−3 in the SI. The DSC data shows a glass transition at around 27 °C and a melting point around 51 °C. This indicates that films of the macroinitiator should be solid at room temperature and able to dewet at temperatures above 51 °C. Macroinitiator films were spin-coated onto PS films, annealed at 100 °C, and the dewetting monitored by optical microscopy. 2991

dx.doi.org/10.1021/bm3010534 | Biomacromolecules 2012, 13, 2989−2996

Biomacromolecules

Article

Figure 3. (a) First order kinetics of the poly(PEGMA) polymerization. Estimated solution polymer average molecular weight (Mn) measured by SEC (filled squares) and polydispersity index (PDI, empty circles, right vertical axis) plotted vs conversion (the straight line is the theoretical prediction for Mn). (b) Thickness of grafted poly(PEGMA) film measured by ellipsometry versus solution polymer Mn (empty squares), fitted with a second order polynomial function (solid line), and results of two independent chain extension experiments (arrows i and ii). (c) XPS high resolution C 1s spectra of poly(PEGMA) brushes with different thickness, grafted from macroinitiator films. The ether carbon contribution (C3, red), more abundant in poly(PEGMA), and the ester carbon contribution (C5, green), more abundant in the macroinitiator, are highlighted in each spectrum. (d) Two-dimensional XPS map (oxygen O 1s signal) of a 40 μm holes dewetted pattern, after grafting with poly(PEGMA) (approximately 60 nm thick). The dark round areas correspond to the dewetted holes, while the bright background is the oxygen-rich poly(PEGMA) brush.

(PEGMA) chains from deep within the film, as well as on its surface. The ability of the monomer to penetrate the macroinitiator film and polymerize from within it has been observed before in the case of polyelectrolyte multilayers containing ATRP initiators.37 The composite brush prepared in this work was likely composed of a hierarchical architecture of macroinitiator chains onto which poly(PEGMA) chains were grafted, which were themselves composed of a methacrylate backbone bearing relatively short (8−9 units) poly(ethylene oxide) side chains. In the biological study, this deeply intermixed, dendrimer-like brush was shown to be extremely effective in preventing protein adsorption and cell attachment. Two chain extension experiments (two arrows, Figure 3b) confirmed the presence of living groups on the terminal ends of the grafted polymer chains: the grafting reaction was allowed to proceed to a given conversion, stopped by exposure to air, and later reinitiated, thanks to the presence of living initiator ends on the grafted poly(PEGMA) chains. The presence of such living groups is fundamental for the postfunctionalization of the polymer brush, a feature that can greatly enhance the versatility of this platform. X-ray photoelectron spectroscopy (XPS) survey spectra were collected from the three separate layers constituting the platform (PS, macroinitiator, and poly(PEGMA), Figure S4 and Table S1 in the SI). The atomic composition data clearly indicate the presence of bromine (from living ATRP initiator groups) within the top ∼10 nm (the penetration depth of XPS) of the brush at any given thickness (values decreasing from 3.1 to 0.53% with increasing brush thickness), confirming that the poly(PEGMA) chains bore living ends available for postfunctionalization. Presented in Figure 3c is a collection of high resolution XPS C 1s spectra for samples with different brush thickness. With

sensitivity to traces of radical scavengers such as oxygen, and therefore, does not require the use of a Schlenk line and complex deoxygenating procedures. All that is required is to bubble nitrogen in the reaction mixture for a short time to remove the bulk of the dissolved oxygen and to work in a sealed vessel. Finally, our particular selection of solvent allowed to obtain the desired polymer brush in a remarkably short time (about 45 min for the thickest brush prepared) while retaining the integrity of the macroinitiator film. The use of a soluble sacrificial initiator led to the formation of large quantities of free chains in solution, which provide information on the molecular weight and dispersity of the surface-grafted chains.32−34 The sacrificial initiator in solution also allows for better control of the polymerization on the surface by maintaining the optimal ratio between the different catalyst species in proximity of the surface.35,36 Size exclusion chromatography (SEC) on the chains grown in solution (Figure 3a, filled squares) showed that the molecular weight of poly(PEGMA) increased linearly with conversion following the theoretical prediction. The molecular weight polydispersity index was below 1.2 for all samples (Figure 3a, empty circles). These features prove that the solution polymerization was controlled, and following Ohno and co-workers,32,33 we can assume that the polymerization of the surface-grafted brushes was also controlled. The thickness of the grafted poly(PEGMA) brush increased gradually, in concordance with the reaction conversion (Figure 3b). The increase in brush thickness was controllable and reproducible and followed a second order polynomial trend. The growth did not exhibit the linear trend expected for polymer brushes grafted from rigid substrates, likely because of swelling of the macroinitiator film and growth of poly2992

dx.doi.org/10.1021/bm3010534 | Biomacromolecules 2012, 13, 2989−2996

Biomacromolecules

Article

increasing brush thickness, the signal characteristic of the poly(PEGMA) brush (ether carbon, component C3, red peaks) became increasingly intense, and the signal related to the macroinitiator (ester carbon, component C5, green peaks) became less intense because of the smaller number of ester groups in poly(PEGMA) compared to the macroinitiator. The spectra were normalized to the component peak with the highest intensity in each spectrum, which was component C1 when the brush thickness was low and component C3 when the brush thickness was above 7.5 nm, and the poly(PEGMA) contribution was dominant over the macroinitiator one. The XPS profiles for the thickest poly(PEGMA) brushes analyzed matched those of the plain poly(PEGMA) reference. The results confirmed the successful growth of a poly(PEGMA) brush grafted from the macroinitiator film. The atomic composition data relative to the components C3 and C5, grafted against brush thickness, can be found in Figure S6 in the SI. Comparison of the atomic composition data (Table S2 in the SI) derived from the high resolution C1 spectra on the three separate layers constituting the platform (PS, macroinitiator, and poly(PEGMA), Figure S5 in the SI) and from poly(PEGMA) brushes with different thickness (Figure 3c) provided more information on the brush structure. The ester carbon component (C5) was expected to level at approximately 8% as soon as the poly(PEGMA) brush reached a thickness larger than the XPS penetration depth; instead, this value was higher than expected for brush thicknesses up to 54 nm. This indicates the presence of macroinitiator chains within the top ∼10 nm at any given brush thickness (except for the thickest brush prepared), confirming the hierarchical architecture of the polymer brush suggested by the nonlinear growth with conversion (Figure 3b). The success of chemical patterning of the composite poly(PEGMA) brush was established by a two-dimensional XPS map of oxygen on a dewetted macroinitiator film grafted with approximately 60 nm of poly(PEGMA) (Figure 3d). The dark areas correspond to the dewetted holes, where the oxygenpoor PS is exposed. The bright area around the holes corresponds to the poly(PEGMA) brush, which is rich in oxygen. Selective Immobilization of Extracellular Matrix Proteins onto the Patterned Polymer Brush. The dewetted, grafted surfaces were employed to pattern proteins in specific locations. Shown in Figure 4a−d are optical micrographs of two macroinitiator films, dewetted to different extents, one containing holes of around 20 μm in diameter and one completely dewetted, containing only droplets of macroinitiator on top of flat PS. The dewetted macroinitiator films were grafted with approximately 60 nm of poly(PEGMA) composite brush and incubated in a 1 mg mL−1 solution of bovine serum albumin tagged with fluorescein isothiocyanate (FITC-BSA). FITC-BSA adsorbed only onto the adhesive PS microdomains (Figure 4b,d) and not on the surrounding poly(PEGMA)-grafted layer. Fluorescence microscopy experiments carried out using FITC-tagged human fibrinogen (FITC-FGN) gave the same result obtained for BSA (Figure S8a,b, SI). A control experiment carried out on a dewetted macroinitiator, without grafted poly(PEGMA), showed that FITC-FGN adsorbed on the entire surface, with no preference for the dewetted PS holes (Figure S8c,d, SI). Confirmation of BSA patterning was obtained also by AFM imaging (Figure S7, SI). The roughness

Figure 4. (a, c) Optical and (b, d) fluorescence micrographs of two dewetted patterns grafted with a 60 nm thick poly(PEGMA) brush, incubated with FITC-BSA. The fluorescent proteins adsorbed selectively in the areas where the underlying polystyrene was exposed (bottom of holes in (a) and flat layer underneath macroinitiator droplets in (b)).

of the PS domains was shown to increase dramatically after protein adsorption. Surface roughness has been quantitatively related to adsorbed protein mass. 38 The fluorescence microscopy and AFM characterizations with proteins further confirmed that grafting was successful on the dewetted pattern and clearly demonstrated that the presence of a poly(PEGMA) layer greatly reduced protein adsorption on the grafted areas, as expected for this low-fouling polymer.39 Poly(PEGMA) chain grafting did not occur in the dewetted domains (due to the absence of ATRP initiating groups), which allowed for the localization of the protein in these regions. Selective Immobilization of Cells onto the Patterned Polymer Brush. Finally, the patterned surfaces were employed to selectively immobilize cells. In preliminary tests, the interaction of fibroblasts with the PS film, the macroinitiator film and substrates with uniform poly(PEGMA) brush coatings was tested, with and without vitronectin preadsorption (Figure S9, SI). The poly(PEGMA) brush was effective in preventing fibroblast adhesion with and without vitronectin preadsorption, while cell adhesion occurred on both PS and macroinitiator surfaces both with and without vitronectin preadsorption. Cell adhesion on untreated PS was poor, which is consistent with the trend in the field of using plasma discharge treatments to produce PS surfaces with better cell adhesion properties.40−42 The presence of a high concentration of vitronectin preadsorbed onto the PS films guaranteed a good cell attachment. The micropatterned brush surfaces were hence exposed to vitronectin before the seeding of fibroblasts. The vitronectin would collect onto the PS domains, but not onto the poly(PEGMA) background. Fibroblasts were allowed to attach in serum onto two different patterned substrates grafted with a 60 nm thick poly(PEGMA) brush, containing holes of either 20 or 40 μm diameter. After 24 h, selective cell attachment was observed on the pattern with holes of 40 μm diameter (Figure 5a). The fibroblasts attached almost exclusively inside the vitronectin-coated PS holes, individually or in groups of three or four. Obvious signs of spreading within the holes were observed in some of the cells, which appeared elongated or 2993

dx.doi.org/10.1021/bm3010534 | Biomacromolecules 2012, 13, 2989−2996

Biomacromolecules

Article

Figure 6. (a) Optical micrograph and (b, c) fluorescence micrographs of fibroblasts adhering inside 40 μm diameter PS holes within a micropatterned poly(PEGMA) grafted brush (60 nm thick), after incubation with vitronectin. Images (a) and (b) were collected on the same area and clearly show that the cells selectively attached inside the dewetted holes. The Live/Dead staining revealed live cells in green and dead cells in red. All cells, except for a few exceptions, were alive after 24 h. Image (c) shows how fibroblasts exhibited extensive spreading when not clustered too densely within the same vitronectincoated PS domain.

Figure 5. Fluorescence micrographs showing fibroblasts adhering on patterns of (a) 40 μm diameter and (b) 20 μm PS holes within a micropatterned poly(PEGMA) grafted brush (60 nm thick), after incubation with vitronectin. The cells selectively attached inside the dewetted holes when the available area was big enough (a), while they randomly deposited onto the surface when the dewetted holes were too small and exhibited a spherical shape, typical of nonadhered cells (b). In such a case, the fibroblasts could only spread when two or more dewetted holes were adjacent (red arrows). The images were converted to gray scale to better visualize both the hole rims and the fluorescent cells.

which stains viable cells in green, and ethidium homodimer, which stains dead cells in red. Almost all of the attached cells fluoresced green, indicating that they were viable after 24 h. The test confirmed the selective attachment of viable fibroblasts within the dewetted PS domains (compare Figure 6a and b, which were collected on the same area). The test also indicated that the attached fibroblasts could spread on the vitronectincoated domains when they were isolated and the available adhesive area was sufficient, while they would conserve a more rounded shape when the adhesive domains were too crowded (Figure 6c). Regardless of the number of cells attached in each domain, all the cells exhibited good viability. This result is very promising for the potential use of this platform in controlled cell culturing.

flattened. Only a small number of cells were present on the poly(PEGMA) brush areas and were of a rounded morphology, consistent with a low degree of interaction with the surface. Cell adhesion was also observed on the patterns with 20 μm holes; however, the cells were rounded (not spread) due to the restricted area available for interaction (Figure 5b). The only exceptions were those areas where the protein-coated domains were close enough to each other for the cells to bridge between them (see arrows in Figure 5b). In such cases, the fibroblasts would elongate greatly to increase their contact area with the adsorbed vitronectin. In their seminal work, Chen et al.43 demonstrated that cells require a minimum spreading area in order to survive after attachment. The fibroblasts used here were of similar size to the holes and spread only when they could bridge over adjacent 20 μm holes. The viability of the fibroblasts patterned on 40 μm diameter holes was ascertained by a live/dead staining assay (Figure 6). In this technique, two different dyes were used, calcein-AM,



CONCLUSIONS In conclusion, for the first time, thin film dewetting and polymer grafting were combined to create micropatterned grafted brushes. AGET ATRP allowed the growth of polymer graft layers in a robust, simple, and reproducible way and, for the first time, was used to graft poly(PEGMA) brushes from a polymer film under mild conditions. Dewetting has intrinsic characteristics that make this approach versatile. By using appropriate polymers, this patterning method could be made to suit many applications, such as tuning the mechanical properties, cytotoxicity and biodegradability of surfaces. This process is easily up-scalable 2994

dx.doi.org/10.1021/bm3010534 | Biomacromolecules 2012, 13, 2989−2996

Biomacromolecules and substrate-independent, as polymer thin films may be applied and patterned on large and nonflat surfaces. The coating of substrates that are not wetted by the bottom PS layer may be achieved by precoating with an appropriate polymer that wets both the substrate and the PS film. The coating of substrates with high roughness, which can influence the dewetting process, may be achieved by employing a thick PS film, so as to smooth the surface. A possible limitation of this approach would be the coating of objects with macroscopic sharp features, as they may interfere with the formation of a homogeneous coating. By carefully selecting experimental parameters, it is possible to control the size and density of the holes on the surface. By preparing a bilayer of macroinitiator films, brushes of different composition could be grown inside and outside the dewetted holes, for example, combining orthogonal polymerization techniques such as RAFT and ATRP. This would be a simple approach to produce polymer brush surfaces with chemically distinct areas, with independent control over the composition and thickness of each area. Ordered patterns can be produced by dewetting, but for applications such as cell studies or for improving the biocompatibility of biomaterials, lateral order is not necessary. The patterned grafted brushes investigated here present outstanding protein and cell patterning capabilities. The control over protein localization on the substrates guided the adhesion of fibroblasts inside the dewetted holes. The cell adhesion experiments illustrated an interesting relationship between the physical area available for cell attachment and spreading: the size of the dewetted, protein-coated domains, determined whether cells attached in selected locations and their morphology (rounded or spread). Poly(PEGMA) brushes synthesized by AGET ATRP have living, functionalizable ends, which can be tagged with specific moieties, such as a peptide for the selective attachment of only one type of cell. In this manner, two different cell types could be arranged inside and outside the dewetted holes, leading to cell cocultures with controlled homotypic and heterotypic interactions, fundamental in the formation of tissues in vitro. On a more general note, this brush patterning approach could contribute significantly to studies of chemical patterning, regulation of wetting behavior, and biosensing.



ACKNOWLEDGMENTS



REFERENCES

The authors thank Prof Sebastién Perrier for useful discussions and for providing the AIBN, Ms. Suzanne Pereira for providing 2-(2-bromopropanoyloxy)ethyl methacrylate, Dr. San Thang for CPDB, Ms. Aeysha Hussein for TPMA, when not available from Sigma, and Ms. Penny Bean for vitronectin. The authors acknowledge the facilities, and the scientific and technical assistance, of the Australian Microscopy and Microanalysis Research Facility at the Australian Centre for Microscopy and Microanalysis, The University of Sydney. Funding from The Australian Research Council, The University of Sydney, and the CSIRO Future Manufacturing National Research Flagship is gratefully acknowledged.

(1) Advincula, R. C.; Brittain, W. J.; Caster, K. C.; Ruhe, J. Polymer Brushes: Synthesis, Characterisation, Applications; Wiley-VCH: Weinheim, Germany, 2004. (2) Barbey, R.; Lavanant, L.; Paripovic, D.; Schuwer, N.; Sugnaux, C.; Tugulu, S.; Klok, H. A. Polymer brushes via surface-initiated controlled radical polymerization: Synthesis, characterization, properties, and applications. Chem. Rev. 2009, 109 (11), 5437−5527. (3) Ducker, R.; Garcia, A.; Zhang, J.; Chen, T.; Zauscher, S. Polymeric and biomacromolecular brush nanostructures: progress in synthesis, patterning and characterization. Soft Matter 2008, 4, 1774− 1786. (4) Kaji, H.; Camci-Unal, G.; Langer, R.; Khademhosseini, A. Engineering systems for the generation of patterned co-cultures for controlling cell-cell interactions. Biochim. Biophys. Acta, Gen. Subj. 2011, 1810 (3), 239−250. (5) Mammoto, T.; Ingber, D. E. Mechanical control of tissue and organ development. Development 2010, 137 (9), 1407−1420. (6) Lehnert, D.; Wehrle-Haller, B.; David, C.; Weiland, U.; Ballestrem, C.; Imhof, B. A.; Bastmeyer, M. Cell behaviour on micropatterned substrata: limits of extracellular matrix geometry for spreading and adhesion. J. Cell Sci. 2004, 117 (1), 41−52. (7) Hannachi, I. E.; Yamato, M.; Okano, T. Cell sheet technology and cell patterning for biofabrication. Biofabrication 2009, 1 (2), 022002. (8) Tang, J.; Peng, R.; Ding, J. The regulation of stem cell differentiation by cell-cell contact on micropatterned material surfaces. Biomaterials 2010, 31 (9), 2470−2476. (9) Charnley, M.; Textor, M.; Khademhosseini, A.; Lutolf, M. P. Integration column: microwell arrays for mammalian cell culture. Integr. Biol. 2009, 1 (11−12), 625−634. (10) Prucker, O.; Rühe, J. Synthesis of poly(styrene) monolayers attached to high surface area silica gels through self-assembled monolayers of azo initiators. Macromolecules 1998, 31, 592−601. (11) Chen, T.; Jordana, R.; Zauscher, S. Polymer brush patterning using self-assembled microsphere monolayers as microcontact printing stamps. Soft Matter 2011, 7, 5532−5535. (12) Iwata, R.; Suk-In, P.; Hoven, V. P.; Takahara, A.; Akiyoshi, K.; Iwasaki, Y. Control of nanobiointerfaces generated from well-defined biomimetic polymer brushes for protein and cell manipulations. Biomacromolecules 2004, 5 (6), 2308−2314. (13) Larsson, A.; Du, C. X.; Liedberg, B. UV-patterned poly(ethylene glycol) matrix for microarray applications. Biomacromolecules 2007, 8 (11), 3511−3518. (14) Maynor, B. W.; Filocamo, S. F.; Grinstaff, M. W.; Liu, J. Directwriting of polymer nanostructures: poly(thiophene) nanowires on semiconducting and insulating surfaces. J. Am. Chem. Soc. 2001, 124 (4), 522−523. (15) Thickett, S. C.; Neto, C.; Harris, A. T. Biomimetic surface coatings for atmospheric water capture prepared by dewetting of polymer films. Adv. Mater. 2011, 23 (32), 3718−3722.

ASSOCIATED CONTENT

S Supporting Information *

The synthesis and characterization of the macroinitiator, additional details on the characterization of poly(PEGMA) in solution, the XPS characterization of plain PS, macroinitiator, and poly(PEGMA) films, additional protein adsorption experiment on the patterned brushes, and cell adhesion control experiments on plain PS, macroinitiator, and poly(PEGMA) films. This material is available free of charge via the Internet at http://pubs.acs.org.





Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; [email protected]. au. Notes

The authors declare no competing financial interest. 2995

dx.doi.org/10.1021/bm3010534 | Biomacromolecules 2012, 13, 2989−2996

Biomacromolecules

Article

transfer radical polymerization techniques. Macromolecules 1998, 31 (17), 5934−5936. (36) Tsujii, Y.; Ohno, K.; Yamamoto, S.; Goto, A.; Fukuda, T. Structure and properties of high-density polymer brushes prepared by surface-initiated living radical polymerization. Adv. Polym. Sci. 2006, 197, 1−45. (37) Edmondson, S.; Vo, C. D.; Armes, S. P.; Unali, G. F.; Weir, M. P. Layer-by-layer deposition of polyelectrolyte macroinitiators for enhanced initiator density in surface-initiated ATRP. Langmuir 2008, 24 (14), 7208−7215. (38) Pamula, E.; De Cupere, V.; Dufrene, Y. F.; Rouxhet, P. G. Nanoscale organization of adsorbed collagen: Influence of substrate hydrophobicity and adsorption time. J. Colloid Interface Sci. 2004, 271 (1), 80−91. (39) Gunkel, G.; Weinhart, M.; Becherer, T.; Haag, R.; Huck, W. T. S. Effect of polymer brush architecture on antibiofouling properties. Biomacromolecules 2011, 12 (11), 4169−4172. (40) Pitt, W. G. Fabrication of a continuous wettability gradient by radio-frequency plasma discharge. J. Colloid Interface Sci. 1989, 133 (1), 223−227. (41) Steele, J. G.; Dalton, B. A.; Johnson, G.; Underwood, P. A. Polystyrene chemistry affects vitronectin activity - An explanation for cell attachment to tissue-culture polystyrene but not to unmodified polystyrene. J. Biomed. Mater. Res. 1993, 27 (7), 927−940. (42) Tamada, Y.; Ikada, Y. Cell-adhesion to plasma-treated polymer surfaces. Polymer 1993, 34 (10), 2208−2212. (43) Chen, C. S.; Mrksich, M.; Huang, S.; Whitesides, G. M.; Ingber, D. E. Geometric control of cell life and death. Science 1997, 276, 1425−1428.

(16) Neto, C. A novel approach to the micropatterning of proteins using dewetting of polymer bilayers. Phys. Chem. Chem. Phys. 2007, 9, 149−155. (17) Ohno, K.; Kayama, Y.; Ladmiral, V.; Fukuda, T.; Tsujii, Y. A versatile method of initiator fixation for surface-initiated living radical polymerization on polymeric substrates. Macromolecules 2010, 43 (13), 5569−5574. (18) Edmondson, S.; Armes, S. P. Synthesis of surface-initiated polymer brushes using macro-initiators. Polym. Int. 2009, 58 (3), 307− 316. (19) Tsai, W. B.; Chien, C. Y.; Thissen, H.; Lai, J. Y. Dopamineassisted immobilization of poly(ethylene imine) based polymers for control of cell-surface interactions. Acta Biomater. 2011, 7 (6), 2518− 2525. (20) Teare, D. O. H.; Schofield, W. C. E.; Roucoules, V.; Badyal, J. P. S. Substrate-independent growth of micropatterned polymer brushes. Langmuir 2003, 19 (6), 2398−2403. (21) Nystrom, D.; Malmstrom, E.; Hult, A.; Blakey, I.; Boyer, C.; Davis, T. P.; Whittaker, M. R. Biomimetic surface modification of honeycomb films via a “grafting from” approach. Langmuir 2010, 26 (15), 12748−12754. (22) Genua, A.; Alduncin, J. A.; Pomposo, J. A.; Grande, H.; Kehagias, N.; Reboud, V.; Sotomayor, C.; Mondragon, I.; Mecerreyes, D. Functional patterns obtained by nanoimprinting lithography and subsequent growth of polymer brushes. Nanotechnology 2007, 18, 21. (23) von Werne, T. A.; Germack, D. S.; Hagberg, E. C.; Sheares, V. V.; Hawker, C. J.; Carter, K. R. A versatile method for tuning the chemistry and size of nanoscopic features by living free radical polymerization. J. Am. Chem. Soc. 2003, 125 (13), 3831−3838. (24) Coad, B. R.; Lu, Y.; Meagher, L. A substrate-independent method for surface grafting polymer layers by atom transfer radical polymerization: Reduction of protein adsorption. Acta Biomater. 2012, 8, 608−618. (25) Braunecker, W. A.; Matyjaszewski, K. Controlled/living radical polymerization: Features, developments, and perspectives. Prog. Polym. Sci. 2007, 32 (1), 93−146. (26) Telford, A. M.; James, M.; Meagher, L.; Neto, C. Thermally cross-linked PNVP films as antifouling coatings for biomedical applications. ACS Appl. Mater. Interfaces 2010, 2 (8), 2399−2408. (27) Moad, G.; Rizzardo, E.; Thang, S. H. Living radical polymerization by the raft process - a second update. Aust. J. Chem. 2009, 62 (11), 1402−1472. (28) Jacobs, K.; Herminghaus, S.; Mecke, K. R. Thin liquid polymer films rupture via defects. Langmuir 1998, 14, 965−969. (29) Xue, L.; Han, Y. Pattern formation by dewetting of polymer thin film. Prog. Polym. Sci. 2011, 36 (2), 269−293. (30) Ghezzi, M.; Thickett, S. C.; Neto, C. Early and intermediate stages of guided dewetting in polystyrene thin film. Langmuir 2012, 28 (27), 10147−10151. (31) Seemann, R.; Herminghaus, S.; Jacobs, K. Gaining control of pattern formation of dewetting liquid films. J. Phys.: Condens. Matter 2001, 13 (21), 4925−4938. (32) Ohno, K.; Akashi, T.; Huang, Y.; Tsujii, Y. Surface-initiated living radical polymerization from narrowly size-distributed silica nanoparticles of diameters less than 100 nm. Macromolecules 2010, 43 (21), 8805−8812. (33) Ohno, K.; Ma, Y.; Huang, Y.; Mori, C.; Yahata, Y.; Tsujii, Y.; Maschmeyer, T.; Moraes, J.; Perrier, S. Surface-initiated reversible addition-fragmentation chain transfer (RAFT) polymerization from fine particles functionalized with trithiocarbonates. Macromolecules 2011, 44 (22), 8944−8953. (34) Oh, J. K.; Min, K.; Matyjaszewski, K. Preparation of poly(oligo(ethylene glycol) monomethyl ether methacrylate) by homogeneous aqueous AGET ATRP. Macromolecules 2006, 39 (9), 3161−3167. (35) Ejaz, M.; Yamamoto, S.; Ohno, K.; Tsujii, Y.; Fukuda, T. Controlled graft polymerization of methyl methacrylate on silicon substrate by the combined use of the Langmuir-Blodgett and atom 2996

dx.doi.org/10.1021/bm3010534 | Biomacromolecules 2012, 13, 2989−2996