Biomacromolecules 2003, 4, 987-994
987
New Class of Ultrathin, Highly Cell-Adhesion-Resistant Polyelectrolyte Multilayers with Micropatterning Capabilities Sung Yun Yang, Jonas D. Mendelsohn, and Michael F. Rubner* Department of Materials Science and Engineering, Massachusetts Institute of Technology, 77 Massachusetts Ave. Cambridge, Massachusetts 02139 Received February 3, 2003; Revised Manuscript Received April 2, 2003
Hydrogen-bonded multilayers comprised of polyacrylamide (PAAm) and a weak polyelectrolyte, such as poly(acrylic acid) (PAA) or poly(methacrylic acid) (PMA), were investigated for their surface-cell interactions. The assembled films were lightly cross-linked thermally or photochemically in order to render them stable in a physiological environment. Both PAA/PAAm and PMA/PAAm multilayers were found to exhibit a high resistance to the adhesion (cytophobicity) of mammalian fibroblasts, even with only a single bilayer coating. Protein adsorption to the multilayers, as revealed by surface plasmon resonance measurements, was greatly reduced for fibronectin and serum-containing medium. In situ swelling experiments indicate that the H-bonded multilayers are hydrogellike coatings capable of a high level of swelling in buffered solution. Utilizing the H-bonding nature of these multilayers, we were able to micropattern the films to create more complex cell-resistant/-adhesive surfaces. The long-term stability of the cell-resistant multilayers was found to be exceptionally good even under conditions (pH 7.4, buffered solution) where a high degree of swelling takes place. No degradation of the micropatterned films was observed over a period of a month, during which time the multilayer coatings remained highly resistant to cell-adhesion. Introduction Control over the manner in which proteins and cells interact with surfaces is critically important in manipulating and studying cellular responses such as adhesion, growth rate, motility, and differentiation as well as surface mediated apoptosis that may help to prevent the failure of implanted medical devices.1,2 In addition to being able to control the chemical and physical attributes of a surface to promote a specific cellular response, it is desirable to have practical methodologies available for creating large-area, chemically patterned surfaces and surfaces with controlled topography.3 Such heterogeneous surfaces have already provided new insights into the nature of cell-surface interactions and may lead to new opportunities for fabricating biosensor arrays and cell-based drug screening devices. Typically, suitable surface chemistries such as graft polymerization4 or monolayer self-assembly3 are used in conjunction with micronscale patterning techniques to create surfaces exhibiting both cell-resistant and cell-adherent domains. The most widely used “cell-resistant” coatings based on the ethylene glycol structural unit,5 however, have substrate limitations related to their preparation chemistry and problems associated with their long-term stability.6 Previous studies have suggested that ‘noncharged (ionically neutral), hydrophilic and highly hydrated surfaces’ are good candidates for protein or cell resistant coatings.7 For example, the reduced adsorption of proteins and cells onto poly(ethylene glycol) (PEG) surfaces has been explained by these characteristics. Likewise, thick hydrogels based on * To whom correspondence should be addressed. E-mail:
[email protected].
polyacrylamide (PAAm) have demonstrated good resistance to cell adhesion and are well established as bioinert platforms suitable for chemical functionalization to promote controlled cell adhesion.8,9 As a result, polyacrylamide cross-linked gels have been extensively used in biological applications such as nonfouling membranes, hydrogel drug delivery systems, gel electrophoresis, and flocculants in wastewater treatment.10,11 Despite the importance of PAAm hydrogels in the biomaterial area, there still does not exist a simple means for creating water insoluble thin film coatings. To present PAAm onto a desired surface, it typically has to be grafted onto the surface and cross-linked to form a hydrogel. These methods, however, are not conducive to the creation of conformally coated thin films with uniform surface coverage and nanoscale controllable thickness on many different substrates. In addition, they often involve multistep synthetic procedures that require extensive extractions of un-reacted toxic chemicals such as acrylamide monomer. Polyelectrolyte multilayers represent an attractive alternative approach for creating functionalized coatings. With this approach, ultrathin films are assembled layer-by-layer from the repetitive, sequential adsorption of oppositely charged polyelectrolytes from dilute aqueous solution.12 Layer-bylayer (LbL) deposition allows nanoscale control over the thickness, composition, and molecular structure of the deposited film. Furthermore, it can be utilized to fabricate conformal coatings without limitations on the size, geometry, or material type of the substrate. Because of the simplicity of the LbL process and the physical similarity it shares with natural biological systems, some groups have already investigated polyelectrolyte multilayers for biomedical applications
10.1021/bm034035d CCC: $25.00 © 2003 American Chemical Society Published on Web 05/07/2003
988
Biomacromolecules, Vol. 4, No. 4, 2003
by studying the interactions of multilayers with living cells.13-15 Although oppositely charged polyelectrolytes are most widely used in the LbL process, recently, we16,17 and others18,19 have demonstrated that polymer multilayers can also be assembled through hydrogen bonding (H-bonding) interactions. Because of the weak nature of H bonds, these multilayers are pH-sensitive; therefore, pH control of the dipping process and suitable posttreatment stabilizing processes are critical. Using this simple approach, it is possible to create thin film polyacrylamide blends in which the composition and physical properties are controlled at the molecular level by adjustments in the processing conditions. Of particular interest to this work is the layer-by-layer assembly of polyacrylamide with weak polyacids such as poly(acrylic acid) (PAA) and poly(methacrylic acid) (PMA). With these H-bonded multilayers, we describe here how we created practical bioinert surfaces. In addition, water-based patterning techniques to create cell-adhesive/resistant micropatterns using these multilayer films are also addressed in this paper. Materials and Methods Materials. Poly(acrylic acid) (PAA) (Mw ∼ 90 000, 25% aqueous solution) and poly(methacrylic acid) (PMA) (Mw ∼ 100 000) were purchased from Polysciences. Poly(allylamine hydrochloride) (PAH) (Mw ∼ 70 000) and the methylene blue dye were purchased from Aldrich. Lysozyme (from chicken egg white, E. C. 3.2.1.17) and bovine serum albumin (fraction V) were obtained from Sigma. Fibronetin (from human plasma, 440 000 Da) was purchased from Gibco/Invitrogen. The solutions of lysozyme, bovine serum albumin, and fibronectin were prepared as 1, 1, and 0.2 g/L solutions, respectively, in Dulbecco’s phosphate buffered saline (PBS) (pH ∼ 7.4, with calcium and magnesium). Polyelectrolyte Multilayer-Coated Tissue Culture Dishes. All polymer solutions were prepared with deionized water (0.01 M, based on the repeating unit molecular weight) without addition of salt. The pH of the rinsing baths as well as the polymer solutions was adjusted to 3.0 with dilute HCl (0.01 M). All polymer solutions were filtered with 0.45-µmpore size cellulose nitrate filters. The layer-by-layer process was carried out via the use of an automatic dipping machine (HMS programmable slide stainer from Zeiss Inc.). The substrates (tissue culture polystyrene dishes (TCPS)) were first immersed into a PAH solution (0.01 M, pH 3.0) for 15 min followed by rinsing with pH 3.0 water. This step introduces a primer/adhesion layer for the subsequent PAA and PAAm layers. Similar results were obtained without the use of the PAH primer layer. The substrates were then dipped into a polyacid (either PAA or PMA) solution followed by three rinses (2, 1, 1 min in separate bins) and then dipped into a PAAm solution followed by the same rinsing procedure. Dipping the substrate into a PAA solution followed by a PAAm solution completes the cycle for what we refer to as a single bilayer. This dipping cycle was repeated for producing the multilayers. The multilayer films were thermally cross-linked at 90 °C under vacuum for 8 h.
Yang et al.
With this treatment, no significant deformation of the TCPS takes place. Details concerning the assembly and crosslinking of these multilayer films can be found in a previous paper.17 Cell Culture. NR6WT mouse fibroblast cells were cultured on tissue culture polystyrene (TCPS) in R-minimal essential medium (R-MEM) supplemented with 7.5% fetal bovine serum (FBS), 1%(v/v) nonessential amino acids (10 mM), 1% (v/v) L-glutamine (200 mM), 1% (v/v) pyruvate sodium salt (100 mM), 1% (v/v) penicillin (10 000 U/ml, Sigma), 1% (v/v) streptomycin (10 mg/ml, Sigma), and 1% (v/v) Geneticin (G418) antibiotic (350 µg in 10 mL PBS) at 37 °C in a humidified atmosphere of 95% air and 5% CO2. The pH of the medium was adjusted to 7.4. For attachment and proliferation assays, fibroblasts were removed from their growth surface by trypsin and then spun down in a centrifuge at ∼1000 rpm for ∼5 min. The cells were then resuspended in fresh media, mixed in a 1:1 ratio with 0.4% trypan blue (Sigma) and counted with a hemocytometer with trypan blue exclusion to determine cell viability prior to seeding. The NR6WT fibroblasts were seeded onto the sterilized multilayer-coated TCPS dishes at a population density of ∼10 000 cells/cm2 for cell-adhesion investigations. Prior to cell adhesion studies, the multilayers were sterilized by spraying with 70% ethanol. A Nikon inverted phase contrast microscope with Openlab 3.0 software was used for all experiments to capture images of the cell density, morphology, and spreading on the multilayer surfaces over 30 days. The media was exchanged every other day, and cells were photographed daily. New cell batches with the same density were seeded onto the multilayer-coated dishes every week to examine the long-term cell resistance of the multilayer film. Quartz Crystal Microbalance (QCM). QCM crystals with silver and gold electrodes were used in this study. In both cases, very similar results were obtained. Gold electrodes were cleaned with hot water and ethanol and dried with N2 gas. Silver electrodes were cleaned with 1 wt % KOH solution (ethanol:water ) 3:2) for 2 min at 50 °C under weak sonication followed by a rinsing with deionized water and drying by flushing with N2 gas. A mass control QCM system20 comprised of a HP 53131A/ 225 MHz universal counter and computer was connected to an automated dipper. The frequency shift by molecule adsorption onto the QCM crystal was monitored every second by a computer using 232 cable. The amount of material deposited onto the multilayers and bare substrates was calculated using the relation between frequency shift and mass, as derived from the Sauerbrey equation:21 - ∆F )
2F02 AxFqµq
∆m
where F0 is the parent frequency of the QCM (10 MHz), A is the area of the electrode (0.177 cm2), Fq is the density of the quartz (2.65 g cm-3), and µq is the shear modulus (2.95 × 1011 dyne cm-2). Assuming the elastic property of the adsorbed film is close to that of quartz, we obtained the relationship between frequency shift and mass deposited as 1 Hz = 1 ng.20
Cell-Resistant Coating with Micropatterning Capabilities
Surface Plasmon Resonance (SPR). Protein binding experiments were performed by a BIAcore 2000 SPR instrument (Biacore, Inc.). Polyelectrolyte multilayers were assembled on SPR plain gold-coated glass sensor chips using the automated dipper system described above, and then the film-coated chips were heated at 90 °C for 8 h and mounted to plastic sensor chip supports for sample installation. After installation of the sensor chip to the SPR instrument, phosphate buffered saline (Ca2+, Mg2+ containing PBS from Dulbecco) was injected and flowed over the multilayercoated gold sensor chip for at least 1 h at a flow rate of 20 µL/min. Then, 100 µL of bovine serum albumin, fibronectin, lysozyme, and serum-containing cell culture medium were injected with a flow rate of 20 µL/min over separate flow channels. After injection of each protein, PBS was delivered to each channel to wash off any excess or poorly bound protein onto the film. A self-assembled monolayer (SAM) of ethylene glycol (EG) was prepared by immersion of a gold sensor chip into a 5% mercaptoundecyltriethyleneglycol (HS(CH2)11(OCH2CH2)3OH) ethanol solution overnight followed by a brief rinse with ethanol. Then the gold substrate was mounted to a plastic support, and the same protein injection method as the multilayer case was performed. As a control experiment, protein adsorption onto bare gold sensor chips was also measured with the same method as the above. The magnitude of the adsorption of the protein to each sample was then quantified graphically with the Biacore software, using the given relationship that an increase of 1000 response units (RU) ) 1 ng/mm2 of adsorbed protein. Swelling Experiments. The in situ swelling of PAA/ PAAm multilayer thin films assembled onto silicon substrates was obtained by using an atomic force microscope (AFM) with a fluid cell in contact mode with Si3N4 cantilevers under fluid (PBS with calcium and magnesium (pH ∼ 7.4)). A single layer of PAH was used as a primer/adhesion layer. The dry film thickness values were determined by using a Gaertner ellipsometer operating at 633 nm. AFM measurements were also performed to obtain dry thickness values of the films. Both methods were in good agreement (within 10%). To obtain the swollen film thickness of a sample, a scratch was made down to the silicon substrate with a razor blade or by subtractive patterning as described in a previous paper,17 followed by applying a drop of PBS over the scored area. Drops of PBS were added onto the film and allowed to equilibrate for at least an hour in order for the bufferexposed film to reach a stable swollen height, and then the drop area was scanned to find the swollen, “under fluid” thickness. The swelling information was obtained by comparing the relative differences between the dry film thickness and the “under fluid” sample thickness. Results and Discussion Highly uniform multilayer thin films of PAA/PAAm and PMA/PAAm can be assembled onto many different substrates as long as the pH of the dipping and rinsing solutions is maintained below about 4.0. At higher pH, the multilayer films dissolve because of ionization of the carboxylic acid
Biomacromolecules, Vol. 4, No. 4, 2003 989 Table 1. Some Physical and Chemical Characteristics of PAA/ PAAm and PMA/PAAm Multilayers PAA/PAAm PMA/PAAm (3.0/3.0)a (3.0/3.0)a average layer thickness (Å)b degree of ionization of acids as assembledc relative composition (PAA:PAAm, PMA:PAAm)d advancing/receding contact angles with water for acid-topped film (°)e advancing/receding contact angles with water for amide-topped film (°)e advancing/receding contact angles with PBS buffer (Ca2+, Mg2+ containing) for acid-topped film (°)e advancing/receding contact angles with PBS buffer (Ca2+, Mg2+ containing) for amide-topped film (°)e
55
