Polyelectrolyte Complex Membranes for Specific Cell Adhesion

Feb 9, 2008 - Specific cell adhesion is achieved by a three-tiered structure: a silica ... is made available by participants in Crossref's Cited-by Li...
3 downloads 0 Views 1MB Size
Download PDF
Langmuir 2008, 24, 2611-2617

2611

Polyelectrolyte Complex Membranes for Specific Cell Adhesion Andrew C. A. Wan,* Benjamin C. U. Tai, Karl M. Schumacher, Annegret Schumacher, Sau Yin Chin, and Jackie Y. Ying* Institute of Bioengineering and Nanotechnology, 31 Biopolis Way, The Nanos, Singapore 138669 ReceiVed August 21, 2007. In Final Form: NoVember 2, 2007 The presentation of bioactive ligands on biomaterial surfaces is often confounded by the adsorption of proteins present in the biological milieu, rendering any type of cellular response nonspecific. We have engineered a polyelectrolyte complex membrane that demonstrates specific adhesion of various cell types for both two-dimensional (2D) and three-dimensional (3D) cell culture systems. Specific cell adhesion is achieved by a three-tiered structure: a silica cross-linked polycation as the bottom (first) tier, a nonfouling polyanion-poly(ethylene glycol) (PEG) conjugate as the intermediate (second) tier, and the cell-adhesion ligand as the top (third) tier. Each tier of the membrane was characterized in terms of chemical composition and dimensions. Epithelial cells (primary human cortical renal cells and a hepatocellular carcinoma cell line) cultured on the membranes exhibited little cell attachment on the polyanionPEG second tier and good cell adhesion on the RGD-modified third tier. Thus, the second tier allowed the effect of cell adhesion due to the ligand (third tier) to be isolated and distinguished from nonspecific cell attachment to the first tier. For the culturing of cells in three dimensions, the three-tiered membrane system was applied using a highly swellable chitosan membrane as the first tier. The resulting cell-membrane construct was uniformly dispersed and centrifuged to form a matrix that interacted intimately with cells in the form of a pellet. Presentation of RGD in the latter format enhanced the viability of human mesenchymal stem cells (hMSCs) over controls without RGD.

Introduction The trend in biomaterials technology has been shifting away from one where new materials are developed based on trialand-error to one whereby the materials are engineered based on designed biological functionalities. This is particularly evident in the efforts to tailor biocompatible surfaces for medical devices and tissue engineering. In the early generations of biomaterials, many surfaces were found to support cell attachment. These observations were largely due to nonspecific charge interactions between the cells and the surfaces, which were often enriched by the adsorption of proteins from the culture media. Such protein interactions were, however, indiscriminating in terms of ligand type and orientation. In recent years, the emerging paradigm in biomaterials is the removal or reduction of any nonspecific physiological response to generate a bioinert material and then to endow that material with specific ligands to create a more predictable biological response.1 Several strategies have been employed for this purpose, primarily using poly(ethylene oxide) (PEO)/poly(ethylene glycol) (PEG) polymers or copolymers and polyelectrolyte multilayers in the form of coatings or hydrogels.2-5 The objective of this study was to design a membrane with a nonfouling surface upon which biological function, as exemplified by adhesion, could be presented at will. In addition, the versatility of the current system allowed it to be applied as a coating as well as a particulate matrix that could be intimately cultured with cells in three dimensions. The basic structure of the membrane consisted of an underlying chitosan layer (first tier), an intermediate polyanion-PEG layer * To whom correspondence should be addressed. E-mail: [email protected] (J.Y.Y.); [email protected] (A.C.A.W.). (1) Hubbell, J. A. Curr. Opin. Biotechnol. 1999, 10, 123-129. (2) Berg, M. C.; Yang, S. Y.; Hammond, P. T.; Rubner, M. F. Langmuir 2004, 20, 1362-1368. (3) Hern, D. L.; Hubbell, J. A. J. Biomed. Mater. Res. 1998, 39, 266-276. (4) Vermette, P.; Gengenbach, T.; Divisekera, U.; Kambouris, P. A.; Griesser, H. J.; Meagher L. J. Colloid Interface Sci. 2003, 259, 13-26. (5) Lutolf, M. P.; Hubbell, J. A. Biomacromolecules 2003, 4, 713-722.

(second tier), and a top layer comprising the conjugated biological ligands (third tier) (Figure 1). Chitosan is a cationic polysaccharide derived from the crustacean exoskeleton.6 The use of chitosan as a biomaterial for drug delivery and tissue engineering applications has been widely investigated due to its biocompatibility and biodegradability; in fact, it is one of the few known polycations that possess such properties.7 The intermediate layer (second tier) of the membrane was comprised of a polyanion-PEG conjugate. As the carboxyl groups of the polyanion were required for the electrostatic interaction with chitosan, we have used thiol addition to form the polyanionPEG conjugate. Excess maleimidyl groups that remained from the reaction between the polyanion and PEG could then be reacted with the ligand to be presented (third tier). Membranes of various types and compositions have been used as components in implants and medical devices, such as guided periodontal tissue regeneration in dentistry applications,7,8 kidney hemodialysis,9 and as an epithelial equivalent for the conjunctiva.10 Besides these potential applications, our polyelectrolyte membranes would be useful for the engineering of epithelial tissues that make up the kidney and liver. The coatings could be applied to both 2D and 3D surfaces, including porous structures that act as scaffolds for tissue engineering. There is a pressing need not only to develop nonfouling coatings and membranes for devices used in ViVo, but also to present appropriate biomolecular ligands on these surfaces to evoke the desired biological response. This work describes a three-tiered membrane system that meets these objectives and, furthermore, incorporates the flexibility afforded by polyelectrolyte multilayer systems. (6) Khor, E.; Lim, L. Y. Biomaterials 2003, 24, 2339-2349. (7) Owen, G. Rh.; Jackson, J.; Chehroudi, B.; Burt, H.; Brunette, D. M. Biomaterials 2005, 26, 7447-7456. (8) Stavropoulos, F.; Dahlin, C.; Ruskin, J. D.; Johansson, C. Clin. Oral Impl. Res. 2004, 15, 435-442. (9) Ajuria, J. L.; Kimmel, P. L. In Principles and Practice of Dialysis, 3rd ed.; Henrich, W. L., Ed.; Lippincott, Williams and Wilkins: Hong Kong, 2004. (10) Ang, L. P. K.; Cheng, Z. Y.; Beuerman, R. W.; Teoh, S. H.; Zhu, X.; Tan, D. T. H. InVest. Ophthalmol. Visual Sci. 2006, 47, 105-112.

10.1021/la7025768 CCC: $40.75 © 2008 American Chemical Society Published on Web 02/09/2008

2612 Langmuir, Vol. 24, No. 6, 2008

Wan et al.

Figure 1. Three-tiered membrane/coating structure and their respective chemistries. First tier: chitosan membrane/coating, Second tier: polyanion-PEG conjugate (e.g., alginate-PEG conjugate), which forms a polyelectrolyte complex with the first tier. Third tier: ligand (e.g., RGD) conjugated to the second tier by maleimidyl chemistry.

A new generation of biomaterials is being developed where growth factors, immobilized to PEG matrices via matrix metalloproteinase (MMP) degradable peptide cross-linkers, are released in response to cell demand.11 These biomimetic strategies complement other efforts such as the one described in this work, whereby ligands are presented in the immobilized form. It is known that immobilized ligands are sometimes more effective, a phenomenon that is true not only for adhesion factors.12 Therefore, in addition to cell-adhesive effects, the three-tiered membrane system presented herein could conceivably be used to investigate the influence of various bioactive ligands on cell processes such as proliferation and differentiation. Experimental Section Casting of Membranes. Membranes were cast from a solution of 1% (w/v) chitosan in 2% (v/v) acetic acid (HOAc) in polypropylene molds and allowed to coagulate and dry in the fume hood for 1-2 days. Hydrolyzed tetraethoxysilane (TEOS, Fluka) was prepared by mixing 1 part of TEOS in 9 parts of 0.15 M HOAc by volume and vortexing the mixture for 1 h or until only one phase was present. Typically, hydrolyzed TEOS was incorporated into the chitosan solution at a volume ratio of 1:3. A biopsy punch was used to cut the membrane into circular disks (6 mm diameter) for the swelling studies. For the cell adhesion studies, the membranes were clamped within Minutissue rings of 7 mm i.d., and the chemical reactions were performed on one surface. Swelling studies were performed by immersing membranes in deionized water and measuring the diameter at regular time intervals until no further swelling occurred, which was typically within 6 h. Polyanion-PEG Conjugate. For the second tier, the cysteinylated derivatives of both alginate and heparin were synthesized and reacted with MAL-PEG-MAL, a PEG that is bifunctional with respect to the thiol-reactive maleimidyl end group. The polyanion-PEG (11) Lutolf, M. P.; Weber, F. E.; Schmoekel, H. G.; Schense, J. C.; Kohler, T.; Muller, R.; Hubbell, J. A. Nat. Biotechnol. 2003, 21, 513-518. (12) Ito, Y.; Zheng, J.; Imanishi, Y.; Yonezawa, K.; Kasuga, M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 3598-3601.

conjugate was then layered onto the chitosan-based membrane or coating, followed by rinsing to remove the excess, unreacted conjugate. Cysteine-alginate was synthesized as reported by BernkopSchnu¨rch and co-workers.13 Briefly, a 1% (w/v) solution of low molecular weight alginic acid (Sigma) was prepared in deionized water. 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC, Merck) was added to the solution at a final concentration of 50 mM and allowed to react for 45 min. An equal volume of 0.5% (w/v) solution of L-cysteine monohydrate hydrochloride (Merck) was then added dropwise to the mixture under stirring, and the pH was adjusted to 4.0. The resulting mixture was stirred for 2 h at room temperature before the pH was raised to 6.0, and reacted for one more hour. Cysteine-alginate was purified by dialyzing (Spectrum Laboratories, MWCO 3500) the mixture against 1 mM hydrochloric acid (HCl, Merck) for 1 h at 4 °C. This was followed by dialyzing twice against 1 mM HCl containing 1% (w/v) sodium chloride (NaCl, Merck) for 1 h each at 4 °C and, finally, dialyzing overnight against 1 mM HCl at 4 °C. The purified product was isolated by lyophilization (VirTis BenchTop 4K freeze dryer). To determine the degree of cysteine substitution, 1 mg of the lyophilized product was dissolved in 1 mL of deionized water, and the pH was adjusted to between 2 and 3. A volume of 200 µL of 1% (w/v) starch (Merck) solution was added to the solution, and the mixture was titrated against an aqueous iodine (Merck) solution (1 mM) until a permanent pale blue color was observed. The preparation of cysteine-heparin followed a similar procedure as that of cysteine-alginate, except that heparin (Sigma) was used instead of alginic acid. The degrees of cysteine substitution in cysteine-alginate and cysteine-heparin were determined by iodometric titration to be 360 and 579 µmol/g of polymer, respectively. To form the polyanion-PEG conjugate, 5 mg of cysteine-alginate or cysteine-heparin was reacted with 5.5 mg of bis-maleimidylPEG (MAL-PEG-MAL) (Nektar) in 0.5 mL of deionized water by mixing equivolume solutions of the two reactants. Preparation of Three-Tiered Membranes. A volume of 50 µL of the polyanion-PEG conjugate was applied to the silica-cross(13) Bernkop-Schnu¨rch, A.; Kast, C. E.; Richter, M. F. J. Controlled Release 2001, 71, 277-285.

PEG Complex Membranes for Specific Cell Adhesion linked polycationic membrane clamped in a Minutissue ring at room temperature. After 1 h, the solution was evaporated, and the membrane was rinsed thrice with deionized water to remove the excess polyanion-PEG conjugate. The RGD peptide, GCGYGRGDSPG (Mimotopes), was conjugated by applying 50 µL of a peptide solution (1 mg/mL) uniformly to the surface of the membrane. After 1 h of reaction, the membrane was rinsed thrice with deionized water. To examine the distribution of RGD peptide on the surface of the membranes, a fluorescent RGD peptide was immobilized in the same manner and viewed using a confocal microscope (Olympus). Polyelectrolyte Complex Coatings. Hydroxyl groups could be generated on glass surfaces using one of the following two methods. Glass coverslips could be immersed in a “Piranha” solution (i.e., a mixture of 30% H2O2 and 70% concentrated H2SO4) for 1 h at 100 °C, rinsed with deionized water, and dried under an air stream. Alternatively, glass coverslips could be cleaned in a RBS 35 detergent solution at 50 °C for 30 min. Each glass coverslip (2.2 cm × 2.2 cm) was subsequently coated with chitosan by uniformly applying 100 µL of a 3:1 chitosan/ hydrolyzed TEOS solution (0.5% (w/v) chitosan solution in 2% (v/v) HOAc, 1:9 TEOS/0.15 M acetic acid) on its surface. Following this, the polyanion-PEG conjugate and RGD peptide were applied as described in the membrane preparation subsection above. Characterization of Three-Tiered Membrane/Coating. 29Si cross-polarization magic-angle spinning (CPMAS) nuclear magnetic resonance (NMR) spectra were obtained with a Bruker Avance 400 spectrometer at a spin rate of 10 K. The thickness of the chitosan membrane was measured using a micrometer screw gauge. To determine the thickness of the polyanion-PEG conjugate layer (second tier), thin films of chitosan followed by the polyanionPEG conjugate layer were cast on silicon wafers for ellipsometry. Silicon wafers of 1 × 1 cm2 were dip-coated in a 1:3 TEOS/chitosan solution and dried to obtain thin films of chitosan. The thickness of these films was then measured using a Gaertner LSE-WS ellipsometer. A volume of 100 µL of the polyanion-PEG conjugate solution was then placed on the chitosan-coated silicon wafers for ∼1 h. Excess polyanion-PEG conjugate was washed off with deionized water, and the silicon wafers were then left to dry. The increased thickness of the coated silicon wafers was then determined by ellipsometry to obtain the thickness of the second tier. Cell Culture. Primary human cortical renal cells were obtained from Cambrex (Walkersville, MD). The proximal tubule cells were cultured in a renal epithelial growth medium (REGM, Cambrex, Walkersville, MD) under 5% CO2 at 37 °C. HepG2 cells were obtained from the American Type Culture Collection (ATCC) and cultured in full Dulbecco’s modified Eagle’s medium (DMEM, Sigma, St. Louis, MO). Membranes and coverslips were sterilized by immersion in 70% ethanol for at least 30 min, followed by exposure to ultraviolet light for 30 min. Cells were seeded at a density of 7.5 × 104 cells per well in a 24-well plate. Cell Staining and Immunohistochemistry. The adhered cells were fixed in ice-cold ethanol for 10 min. After several rinses with phosphate-buffered saline (PBS), the samples were incubated with a blocking solution containing PBS, 10% fetal calf serum (FCS), and 1% bovine serum albumin (BSA) for 30 min. The aquaporin-1 (AQP-1) primary antibody (Santa Cruz Biotechnologies, Santa Cruz, CA) was diluted at a ratio of 1:100 and incubated with the cells for 2 h. After several rinses, the specimens were incubated for 45 min with a donkey-anti-rabbit-IgG-FITC-conjugated secondary antibody (Jackson Immunoresearch Laboratories, West Grove, PA), which had been diluted 200-fold in PBS containing 1% BSA. Nuclear staining was done with 4′,6-diamidino-2-phenylindole (DAPI, SigmaAldrich, Singapore). The specimens were then analyzed using an IX71 Olympus microscope (Tokyo, Japan). Pellet Culture. A swellable form of the membrane could be cast from low molecular weight chitosan (MW = 103). The same threetiered chemical conjugation strategy was used to immobilize RGD on the membranes. hMSCs (Cambrex) were cultured on the membranes overnight. Using a micropipet, the membrane with adhered cells was gently triturated into a membrane particulate/cell suspension. The mixture was then centrifuged at 560g for 5 min to

Langmuir, Vol. 24, No. 6, 2008 2613

Figure 2. Membrane swelling ratio as a function of TEOS/chitosan volume ratio. obtain a pellet. The supernatant was removed, leaving just 200 µL. A volume of 1 mL of fresh media was then carefully added to the particulate/cell pellet. AlamarBlue Assay for Cell Pellets. A volume of 1 mL of the existing media was removed, leaving about 200 µL behind. A total of 1 mL of 12% (v/v) alamarBlue in media was then added carefully to the cell pellet, and the mixture was incubated at 37 °C in a CO2 incubator for 6 h. After that, 1 mL of the reduced alamarBluecontaining media was then removed for fluorescence measurements and replaced with 1 mL of fresh media. Fluorescence readings were obtained using excitation and emission wavelengths of 545 and 590 nm, respectively.

Results and Discussion The relevant chemical reactions in the formation of the threetiered membrane and the schematic of the three tiers are illustrated in Figure 1. Chitosan was used as the base polymer (first tier). In applying chitosan as a membrane or coating, the need to modify it may arise so as to improve its physical properties. The most common method involves cross-linking chitosan by means of a dialdehyde (e.g., glutaraldehyde) via its amine functionalities. In the present work, the physical properties of the chitosan membranes were improved by the incorporation of hydrolyzed TEOS. The 29Si CPMAS spectrum of a chitosan membrane incorporating hydrolyzed TEOS is shown in Figure S2 of the Supporting Information (SI). The three peaks, Q2 (116.53 ppm), Q3 (107.06 ppm), and Q4 (97.66 ppm), correspond to the resonance frequency of the central Si in the species Si(OSi)2(OH)2, Si(OSi)3(OH), and Si(OSi)4, respectively. This study confirmed that a silica phase had formed within the membrane, which cross-linked the chitosan matrix by either condensation, H-bonding, or both. Significant membrane swelling was observed in the absence of hydrolyzed TEOS. Swelling was proportionately reduced when an increasing volume of hydrolyzed TEOS was introduced (see Figure 2). A hydrolyzed TEOS/chitosan volume ratio of 1:3 was used in our synthesis to avoid membrane swelling. This study illustrated that cross-linking with silica had physically strengthened the chitosan membrane. The second tier was composed of a conjugate between a polyanion (alginate or heparin) and bis-maleimidyl-PEG (MAL-PEG-MAL). The stoichiometry of the reaction between the polyanion, illustrated for the case of cysteine-alginate, and MAL-PEG-MAL was determined by monitoring the peak

2614 Langmuir, Vol. 24, No. 6, 2008

Wan et al.

Figure 3. Confocal micrograph of the fluorescent RGD peptide conjugated on a three-tiered membrane system.

absorbance at 260 nm in the UV-visible spectrum. By keeping the quantity of MAL-PEG-MAL constant at 1.6 µmol and adding varying quantities of cysteine-alginate, the consumption of MAL groups due to reaction with thiol could be determined (see the Supporting Information, Figure S1). In a typical membrane preparation where 5 mg of cysteine-alginate (1.8 µmol thiol) was added to 1.6 µmol of MAL-PEG-MAL, 58% of the MAL groups were consumed, leaving the excess for reaction with the thiolated RGD peptide. This corresponded well with the theoretically calculated value (56%). The concept of coating by electrostatic interaction between the first and second tiers is similar to the polyelectrolyte multilayer technique.14 While the thickness of our typical chitosan membrane (first tier) was 23 µm, the second tier was only 36 ( 10 nm in thickness, corresponding to several molecular layers. Successful conjugation of the RGD peptide was demonstrated by immobilization of the fluorescent isothiocyanate (FITC) conjugated molecule onto the membranes followed by confocal microscopy (Figure 3). The RGD peptide was found to be immobilized in the form of densely packed clusters, the smallest clusters of which were ∼5-10 µm. Methods to immobilize ligands on nonfouling surfaces are important, as they allow us to study the effect of individual factors on cell adhesion, proliferation, and differentiation without the confounding influence of protein adsorption from the serum or the extracellular matrices (ECM) produced by the cells themselves.15 The two prominent approaches to achieve noncell-adhesive surfaces involve the use of PEG in its various forms (branched and linear)3,16 and, more recently, polyelectrolyte complexes.2,17 While the use of PEG imparts a distinct nonfouling property, an advantage of employing polyelectrolyte multilayers is the potential of incorporating proteins such as growth factors to influence cells grown on the biomaterial surface. This work aims to combine both approaches and their advantages, whereby a conjugate of a polyanion with PEG was used to form a polyelectrolyte complex with a polycationic (chitosan) membrane or coating. While chitosan has been modified with RGD directly18 and copolymers have been synthesized from PEG that possess functionalized side chains that may exhibit similar character(14) Decher, G. Science 1997, 277, 1232-1237. (15) Mrksich, M. Curr. Opin. Chem. Biol. 2002, 6, 794-797. (16) Groll, J.; Fiedler, J.; Engelhard, E.; Ameringer, T.; Tugulu, S.; Klok, H.-A.; Brenner, R. E.; Moeller, M. J. Biomed. Mater. Res. 2005, 74A, 607-617. (17) Yang, S. Y.; Mendelsohn, J. D.; Rubner, M. F. Biomacromolecules 2003, 4, 987-994. (18) Ho, M. H.; Wang, D. M.; Hsieh, H. J.; Liu, H. C.; Hsien, T. Y.; Lai, J. Y.; Hou, L. T. Biomaterials 2005, 26, 3197-3206.

Figure 4. Relative adhesion of human renal cortical cells on (1) chitosan, (2) chitosan-alginate-PEG, (3) chitosan-heparin-PEG, (4) chitosan-alginate-PEG-RGD, and (5) chitosan-heparinPEG-RGD membranes after 1 day, as measured by the relative absorbance of alamarBlue (with bars indicating the standard error).

istics,19 the three-tiered system described in this work offers unique advantages. First, the three-tiered system enabled isolation of the effect of the ligand from the effect of the base polymer (i.e., chitosan in this case) by interposing a nonfouling second tier between the underlying substrate (first tier) and the ligand (third tier). Therefore, we could be certain that any observable effect on the cells (e.g., cell adhesion) was due to the ligand and not the polymer. Another advantage of having a layered modification scheme was the flexibility of varying the ligand without having to modify the base polymer. Three-Tiered System for Specific Cell Adhesion. Since its identification as a primary attachment cue by Pierschbacher and Ruoslahti,20 the RGD sequence has been widely applied in the biomaterials field.21-23 This sequence binds primarily to the R5β3 and R5β1 integrin receptors on a variety of cell types, including kidney epithelial cells. After 1 day of culture, the activity (reflective of cell number) of primary human cortical renal cells grown on the various membranes showed the trend in Figure 4, as revealed by alamarBlue assay. The third tier (RGD-modified) membranes exhibited significantly higher adhesion than the second tier (polyanion-PEG) membranes, while cell adhesion on the first tier (chitosan) membrane was intermediate. The poorer cell adhesion observed for the second tier (polyanion-PEG) (19) VandeVondele, S.; Voros, J.; Hubbell, J. A. Biotechnol. Bioeng. 2003, 82, 784-790. (20) Pierschbacher, M.; Ruoslahti, E. Nature 1984, 309, 30-33. (21) Rowley, J. A.; Madlambayan, G.; Mooney, D. J. Biomaterials 1999, 20, 45-53. (22) Woerly, S.; Pinet, E.; de Robertis, L.; van Diep, D.; Bousmina, M. Biomaterials 2001, 22, 1095-1101. (23) Chung, T. W.; Lu, Y. F.; Wang, S. S.; Lin, Y. S.; Chu, S. H. Biomaterials 2002, 23, 4803-4809.

PEG Complex Membranes for Specific Cell Adhesion

Langmuir, Vol. 24, No. 6, 2008 2615

Figure 5. Light micrographs of primary human cortical cells cultured on (a) chitosan-alginate-PEG-RGD, (b) chitosan-alginate-PEG, (c) chitosan, (d) chitosan-heparin-PEG-RGD, and (e) chitosan-heparin-PEG membranes.

membranes compared to the unmodified chitosan membrane confirmed the nonfouling properties of the second tier. After 2 weeks of culture, the renal cells had become confluent on the third tier (RGD-modified) membranes. In contrast, very few cells were observed growing on the second tier (alginate-PEG-MAL) membranes (Figure 5b). However, the second tier (heparinPEG-MAL) membrane (Figure 5e) had supported cell growth to a similar extent as the first tier (chitosan) membrane (Figure 5c). As heparin sulfate proteoglycans are a major component of kidney ECM, the heparin component may have provided binding sites for both cell surface receptors and growth factors to positively influence cell adhesion.24,25 Labeling of the water channel protein, AQP-1, characteristic of the proximal renal tubule cell phenotype, confirmed the presence of AQP-1 expressing cells on the membranes (see Figure 6). The trend in AQP-1 expression correlated well with the cell density, as indicated by the nuclear stain DAPI. The effect of the various modified coatings on the growth of HepG2 is shown in Figure 7. In this case, both the degree as well as the cell adhesion pattern were affected by the availability of the RGD ligand. While cells were attached and distributed evenly in the case of the uncoated glass surface (Figure 7a and b), they were hardly attached to the alginate-PEG-MAL surface (Figure 7c and d). In contrast, cells on the RGD-modified surface attached and proliferated in the form of islands, interconnected by a series of bridges (Figure 7e). Within each island, cells were observed to aggregate into a tight formation (Figure 7f), and each bridge was constituted of cords of cells. The cell aggregation behavior might have resulted from the presentation of RGD in the form of clusters (Figure 3). The difference in the cell adhesion pattern between the glass and RGD-modified surface indicated that RGD ligands were effectively presented to the cells to mediate their adhesion onto an otherwise nonfouling, nonadhesive second tier. This further demonstrated the benefit of interposing nonfouling polyanion-PEG between the first and second tiers; such an approach would ensure that any observed cell adhesion was due specifically to the ligand. While other approaches can be found in the literature, for example, synthesis of copolymers of PEG and poly(L-lysine),19 cell attachment to coatings of such copolymers (ligand-modified or not) can be due to nonspecific charge interaction between the cell membrane and the substrate (e.g., poly(L-lysine)).

Figure 6. Fluorescence micrographs of primary human cortical cells cultured on (a) chitosan-alginate-PEG-RGD, (b) chitosan-alginatePEG, (c) chitosan, (d) chitosan-heparin-PEG-RGD, and (e) chitosanheparin-PEG membranes. The red-stained cells express AQP-1, as indicated by the arrows. DAPI staining (blue) displays the nuclei of the attached cells. Scale bar ) 50 µm.

2616 Langmuir, Vol. 24, No. 6, 2008

Wan et al.

Figure 7. Light micrographs of HepG2 cells cultured on glass coverslips with (a,b) no coating, (c,d) chitosan-alginate-PEG coating, and (e,f) chitosan-alginate-PEG-RGD coating. Images (b), (d), and (f) were taken at a higher magnification.

Figure 8. hMSCs seeded onto the swellable RGD-modified membranes (third tier) exhibited good adhesion and spreading of cells (a) compared to cells seeded onto the alginate-PEG control (second tier) (b). The corresponding pellet cultures are shown in (c) and (d) as histological cross sections stained with hematoxylin and eosin. Cell nuclei are stained dark purple, and the membrane matrix is stained red.

In addition to being used in the form of membranes and coatings, the three-tiered system for ligand presentation could be extended to cell culture in three dimensions. In this work, epithelial cell lines were chosen for cell culture on the 2D membranes. 2D culture is deemed appropriate, since epithelial cells grow in contact with a 2D basement matrix, exhibit apical basolateral polarity, and possess a free apical surface in ViVo. Unlike epithelial cells, however, most other cells grow in a 3D

environment, and recent studies have illustrated the important physiological and phenotypic differences of cells grown in 3D versus 2D.26,27 To simulate the more relevant 3D environment, (24) Groffen, J.; Ruegg, M. A.; Dijkman, H.; van de Velden, T. J.; Buskens, C. A. J. Histochem. Cytochem. 1998, 46, 19-27. (25) Powell, K.; Fernig, D. G.; Turnbull, J. E. J. Biol. Chem. 2002, 277, 2855428563. (26) Abbott, A. Nature 2003, 424, 870-872.

PEG Complex Membranes for Specific Cell Adhesion

Langmuir, Vol. 24, No. 6, 2008 2617

and b). Cross sections of pellets obtained from these membranes/ cells are shown in Figure 8c and d. Hematoxylin and eosin staining revealed the presence of a well dispersed, granular membrane material in intimate contact with the cells. An alamarBlue assay to measure cell metabolism revealed higher activities for cells in the pellet culture with RGD-modified membranes (Figure 9). In this context, the membrane material could be considered to be an artificial ECM, which could be further endowed with signals to affect stem cell proliferation and/or differentiation. A similar concept has been investigated for a particulate system, whereby the growth factor was released rather than immobilized.28

Conclusions

Figure 9. Relative fluorescence of alamarBlue pellet cultures of hMSCs with (9) RGD-modified membranes (third tier) and ([) alginate-PEG control (second tier).

the three-tiered membrane system was applied to a highly swellable chitosan membrane, and cells were grown on the membrane, which was then dispersed to obtain a suspension of particulates and cells. This suspension was then subjected to centrifugation to give a pellet that could be maintained in culture. For the 3D model, we selected human mesenchymal stem cells (hMSCs), as this cell type is widely being investigated for tissue engineering applications. hMSCs seeded onto the RGD-modified membranes exhibited good adhesion and spreading of cells, as compared to the cells seeded onto the alginate-PEG (second tier) control (Figure 8a (27) Cukierman, E.; Pankov, R.; Stevens, D. R.; Yamada, K. M. Science 2001, 294, 1708-1712.

A three-tiered membrane system has been engineered with both physical and chemical modifications, invoking chemistries that are compatible to the retention of both membrane structure and biological activity of the pendant ligands. The system was successfully applied to 2D cultures of epithelial cells as well as to 3D cultures of mesenchymal stem cells. Such a system could provide a more defined chemistry for the new generation of biomaterials, which would not interact indiscriminately with elements of the biological milieu but rather provide a blank template upon which biological functionality could be presented at will. Acknowledgment. The authors thank Shujun Gao, Dr. Su Seong Lee, Ken Goh, and Gabriel Tjio for technical assistance and discussions. This work was supported by the Institute of Bioengineering and Nanotechnology (Biomedical Research Council, Agency for Science, Technology and Research, Singapore). Supporting Information Available: Plot of the reaction of MAL-PEG-MAL with cysteine-alginate as monitored by maleimidyl absorption, and CPMAS 29Si spectrum of a membrane prepared from a 3:1 chitosan/hydrolyzed TEOS solution. This material is available free of charge via the Internet at http://pubs.acs.org. LA7025768 (28) Mahoney, M. J.; Saltzman, W. M. Nat. Biotechnol. 2001, 19, 934-939.