Multilayered Films Fabricated from Plasmid DNA and a Side-Chain

Sep 22, 2007 - Doriane Richard , Isabelle Nguyen , Christine Affolter , Florent Meyer , Pierre Schaaf , Jean-Claude Voegel , Dominique Bagnard , Joell...
0 downloads 0 Views 657KB Size
Download PDF
Langmuir 2007, 23, 11139-11146

11139

Multilayered Films Fabricated from Plasmid DNA and a Side-Chain Functionalized Poly(β-amino Ester): Surface-Type Erosion and Sequential Release of Multiple Plasmid Constructs from Surfaces Jingtao Zhang, Sara I. Montan˜ez, Christopher M. Jewell, and David M. Lynn* Department of Chemical and Biological Engineering, UniVersity of WisconsinsMadison, 1415 Engineering DriVe, Madison, Wisconsin 53706 ReceiVed July 6, 2007. In Final Form: August 6, 2007 Hydrolytically degradable polyamines can be used to fabricate multilayered polyelectrolyte films that erode and release DNA in aqueous environments. Past studies have investigated films fabricated from poly(β-amino ester) 1 and the influence of polymer backbone structure on film erosion and the release of anionic polyelectrolytes. This investigation sought to characterize the influence of polymer side-chain structure on the stability of multilayered films in physiologically relevant media. Here, we report on the fabrication and characterization of multilayered films ∼150 nm thick assembled from plasmid DNA and side-chain functionalized polymer 2. We observed large differences in the behavior of films fabricated from polymer 2 as compared to films fabricated from polymer 1. Whereas films fabricated from polymer 1 erode and release DNA over ∼2 days when incubated in phosphate-buffered saline, films fabricated from polymer 2 erode and release DNA over ∼2 weeks. In addition, whereas films fabricated from polymer 1 undergo complex nanometer-scale physical transformations in aqueous media, characterization of the surfaces of films fabricated from polymer 2 by atomic force microscopy (AFM) demonstrates that the surfaces of these materials remain smooth and uniform during erosion. The apparent surface-type erosion of these materials permits the fabrication of ultrathin films with architectures that provide control over the timing and the order in which two different DNA constructs are released from surfaces. For example, the order in which two different DNA constructs are released from films and expressed by cells can be controlled to measurable extents by the relative order in which they are deposited during fabrication. These results suggest approaches to the localized and sequential release of multiple different DNA constructs to cells or tissues from the surfaces of tissue engineering scaffolds or implantable devices coated with multilayered films.

Introduction Methods for the layer-by-layer deposition of positively and negatively charged polyelectrolytes on surfaces provide powerful tools for the bottom-up assembly of ultrathin polymer films.1-4 Numerous past studies have demonstrated that these methods can be used to incorporate biological polyelectrolytes, such as DNA and proteins, into ultrathin films with nanometer-scale control over film thickness and composition.3,5,6 Layer-by-layer methods also permit the fabrication of thin films composed of multiple different layers of multiple different polyelectrolytes by controlling the relative order in which different polyelectrolyte layers are deposited during fabrication. The ability to fabricate films having such “stratified” architectures7-13 presents new opportunities to design polyelectrolyte assemblies that can be used to control the timing and/or the sequence with which multiple (1) Decher, G. Science 1997, 277, 1232-1237. (2) Bertrand, P.; Jonas, A.; Laschewsky, A.; Legras, R. Macromol. Rapid Comm. 2000, 21, 319-348. (3) Peyratout, C. S.; Dahne, L. Angew. Chem., Int. Ed. 2004, 43, 3762-3783. (4) Hammond, P. T. AdV. Mater 2004, 16, 1271-1293. (5) Ai, H.; Jones, S. A.; Lvov, Y. M. Cell Biochem. Biophys. 2003, 39, 23-43. (6) Tang, Z. Y.; Wang, Y.; Podsiadlo, P.; Kotov, N. A. AdV. Mater 2006, 18, 3203-3224. (7) Dubas, S. T.; Farhat, T. R.; Schlenoff, J. B. J. Am. Chem. Soc. 2001, 123, 5368-5369. (8) Cho, J.; Caruso, F. Macromolecules 2003, 36, 2845-2851. (9) Nolte, A. J.; Rubner, M. F.; Cohen, R. E. Langmuir 2004, 20, 3304-3310. (10) Garza, J. M.; Schaaf, P.; Muller, S.; Ball, V.; Stoltz, J. F.; Voegel, J. C.; Lavalle, P. Langmuir 2004, 20, 7298-7302. (11) Garza, J. M.; Jessel, N.; Ladam, G.; Dupray, V.; Muller, S.; Stoltz, J. F.; Schaaf, P.; Voegel, J. C.; Lavalle, P. Langmuir 2005, 21, 12372-12377. (12) Jessel, N.; Oulad-Abdelghani, M.; Meyer, F.; Lavalle, P.; Haikel, Y.; Schaaf, P.; Voegel, J. C. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 8618-8621. (13) Wood, K. C.; Chuang, H. F.; Batten, R. D.; Lynn, D. M.; Hammond, P. T. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 10207-10212.

different macromolecular species are released from or made available on surfaces coated with these materials.12,13 Thin films and coatings that provide spatial and temporal control over the delivery of multiple different DNA constructs to cells and tissues could have a significant impact on the development of functional scaffolds for tissue engineering or the development of localized gene-based therapies.12,14-16 In this paper, we report a step toward these broad goals by using layerby-layer methods to fabricate ultrathin polyelectrolyte multilayers using plasmid DNA and a hydrolytically degradable, side-chain functionalized poly(β-amino ester). We demonstrate that these assemblies erode in a manner consistent with a top-down, surfacetype erosion mechanism when incubated in physiologically relevant media and that this general approach can be used to fabricate multicomponent films capable of providing control over the timing and the order with which two different plasmid DNA constructs are released from surfaces. Several past studies have demonstrated that multilayered polyelectrolyte films can be fabricated using plasmid DNA12,17-22 and that these assemblies can be designed to erode in aqueous (14) Shea, L. D.; Smiley, E.; Bonadio, J.; Mooney, D. J. Nat. Biotechnol. 1999, 17, 551-554. (15) Klugherz, B. D.; Jones, P. L.; Cui, X.; Chen, W.; Meneveau, N. F.; DeFelice, S.; Connolly, J.; Wilensky, R. L.; Levy, R. J. Nat. Biotechnol. 2000, 18, 11811184. (16) Richardson, T. P.; Peters, M. C.; Ennett, A. B.; Mooney, D. J. Nat. Biotechnol. 2001, 19, 1029-1034. (17) Zhang, J.; Chua, L. S.; Lynn, D. M. Langmuir 2004, 20, 8015-8021. (18) Jewell, C. M.; Zhang, J.; Fredin, N. J.; Lynn, D. M. J. Controlled Release 2005, 106, 214-223. (19) Jewell, C. M.; Zhang, J.; Fredin, N. J.; Wolff, M. R.; Hacker, T. A.; Lynn, D. M. Biomacromolecules 2006, 7, 2483-2491. (20) Reibetanz, U.; Claus, C.; Typlt, E.; Hofmann, J.; Donath, E. Macromol. Biosci. 2006, 6, 153-160.

10.1021/la702021s CCC: $37.00 © 2007 American Chemical Society Published on Web 09/22/2007

11140 Langmuir, Vol. 23, No. 22, 2007

media and release DNA if the cationic components of these materials are selected or designed appropriately.17-19,21-24 We have demonstrated in a series of recent reports that hydrolytically degradable poly(β-amino ester) 1 can be used to fabricate ultrathin multilayered assemblies that erode gradually and release plasmid DNA when incubated in physiologically relevant environments.17-19,25,26 Our past studies have established that assemblies fabricated from polymer 1 erode through a mechanism that involves the hydrolysis of the esters in the polymer backbone25,27 and that objects coated with these erodible films can be used to promote surface-mediated cell transfection when placed in contact with mammalian cells in vitro.18,19 These past studies present promising approaches to the localized, surface-mediated delivery of DNA. However, there are several aspects of the behavior of assemblies fabricated from polymer 1 that could serve to limit the scope of applications for which these materials are suited. For example, assemblies fabricated from polymer 1 and plasmid DNA are, in general, limited to the release of DNA from surfaces over relatively short periods of time (e.g., ∼48 h).17,19 In addition, characterization of the surfaces of these films during erosion using atomic force microscopy (AFM) and scanning electron microscopy (SEM) demonstrated that these films undergo decomposition processes that lead to large-scale film reorganization upon exposure to physiologically relevant media.25,28 The tendency of these DNA-containing assemblies to reorganize in this manner has, thus far, prevented the use of polymer 1 to fabricate “stratified” or multicomponent assemblies that permit control over the sequential release of two or more DNA constructs.

Zhang et al.

to interact electrostatically with DNA in solution.31-33 Here, we report on the fabrication and physical characterization of ultrathin multilayered films fabricated from plasmid DNA and side-chain functionalized poly(β-amino ester) 2. The selection of this side chain amine-functionalized polymer for additional study and full characterization was made on the basis of an initial survey of the properties of assemblies fabricated using a small library of structurally related poly(β-amino ester)s. We report striking differences in the behavior of multilayered films fabricated from polymer 2 and plasmid DNA as compared to films fabricated using polymer 1. For example, whereas films fabricated from polymer 1 erode over a period of ∼2 days in physiologically relevant media,17 assemblies fabricated from polymer 2 and plasmid DNA encoding enhanced green fluorescent protein erode and release transcriptionally active DNA for ∼2 weeks. In addition, whereas films fabricated from polymer 1 undergo large-scale morphological transformations in aqueous media (as described above),25,28 we demonstrate here that films fabricated using polymer 2 remain smooth and uniform during erosion. We demonstrate further that the apparent top-down or surface-type erosion of films fabricated from polymer 2 makes possible approaches to the fabrication of ultrathin films that permit control over the timing and the order in which two different plasmid DNA constructs are released from surfaces (that is, the sequence with which two different plasmids are released from a surface can be controlled to measurable extents by the relative order in which they were deposited during fabrication). The work reported here suggests the basis of methods for the design of ultrathin films and coatings that could, with further development, be designed to provide broad control over the release of multiple different genes or growth factors to cells or tissues from the surfaces of tissue engineering scaffolds or other implantable devices. Materials and Methods

The work reported here arose from our interest in investigating the influence of poly(β-amino ester) structure on the fabrication, erosion profiles, and physical properties of multilayered films fabricated from DNA and other anionic polyelectrolytes.27-30 Whereas past studies have investigated the influence of the polymer backbone structure and hydrophobicity on film erosion and the release of anionic polyelectrolytes,27-29 the results reported here resulted from an investigation of the influence of the polymer side-chain structure on the behavior of these erodible assemblies. This work builds upon the results of past studies demonstrating that changes in the structure of the side chains of poly(β-amino ester)s can lead to large changes in the ability of these polymers (21) Blacklock, J.; Handa, H.; Soundara Manickam, D.; Mao, G.; Mukhopadhyay, A.; Oupicky, D. Biomaterials 2007, 28, 117-124. (22) Chen, J.; Huang, S.; Lin, W.; Zhuo, R. Small 2007, 3, 636-643. (23) Ren, K. F.; Ji, J.; Shen, J. C. Bioconjugate Chem. 2006, 17, 77-83. (24) Ren, K. F.; Ji, J.; Shen, J. C. Biomaterials 2006, 27, 1152-1159. (25) Fredin, N. J.; Zhang, J.; Lynn, D. M. Langmuir 2005, 21, 5803-5811. (26) Lynn, D. M. Soft Matter 2006, 2, 269-273. (27) Zhang, J.; Fredin, N. J.; Lynn, D. M. J. Polym. Sci., Part A: Polym. Chem. 2006, 44, 5161-5173. (28) Fredin, N. J.; Zhang, J.; Lynn, D. M. Langmuir 2007, 23, 2273-2276. (29) Zhang, J.; Fredin, N. J.; Janz, J. F.; Sun, B.; Lynn, D. M. Langmuir 2006, 22, 239-245. (30) Zhang, J.; Lynn, D. M. Macromolecules 2006, 39, 8928-8935.

General Considerations. 1H NMR spectra were acquired using a Bruker AC+ 300 spectrometer. Chemical shift values are reported in ppm and are referenced to residual protons from solvent. Gel permeation chromatography (GPC) was performed using a Waters 515 HPLC pump (Waters Corporation, Milford, MA), a Rheodyne model 7725 injector with a 20-µL injection loop, and two Waters Styragel HT 6E columns in series. THF containing 0.1 M triethylamine was used as the eluent at a flow rate of 1.0 mL/min. Data were collected using a Waters 2410 Refractive Index Detector and processed using the Waters Empower software package. Molecular weights are reported relative to monodisperse polystyrene standards. Silicon substrates (e.g., 0.5 × 2.0 cm) used for the fabrication of multilayered films were cleaned with acetone, ethanol, methanol, and deionized water and dried under a stream of filtered compressed air. Surfaces were then activated by etching with an oxygen plasma for 5 min (Plasma Etch, Carson City, NV) prior to film deposition. The optical thicknesses of films deposited on silicon substrates were determined using a Gaertner LSE ellipsometer (632.8 nm, incident angle ) 70°). Data were processed using the Gaertner Ellipsometer Measurement Program. Relative thicknesses were calculated assuming an average refractive index of 1.58 for the multilayered films. Thicknesses were determined in at least four different standardized locations on each substrate and are presented as an average (with standard deviation) for each film. UV-visible absorbance values for PBS solutions used to determine film release kinetics were recorded on a Beckman Coulter DU520 UV/vis (31) Lynn, D. M.; Anderson, D. G.; Putnam, D.; Langer, R. J. Am. Chem. Soc. 2001, 123, 8155-8156. (32) Akinc, A.; Lynn, D. M.; Anderson, D. G.; Langer, R. J. Am. Chem. Soc. 2003, 125, 5316-5323. (33) Anderson, D. G.; Lynn, D. M.; Langer, R. Angew. Chem., Int. Ed. 2003, 42, 3153-3158.

Plasmid DNA Multilayered Films Spectrophotometer (Fullerton, CA). Solution fluorescence measurements used to characterize concentrations of fluorophore labeled DNA were made using a Fluoromax-3 fluorometer (Jobin Yvon, Edison, NJ). The fluorescence of DNA labeled with Cy3 was measured using an excitation wavelength of 548 nm and an emission wavelength of 563 nm. The fluorescence of DNA labeled with Cy5 was measured using an excitation wavelength of 646 nm and an emission wavelength of 663 nm. Film topography and surface roughness were obtained from height data acquired in tapping mode on a Nanoscope Multimode atomic force microscope (Digital Instruments, Santa Barbara, CA). Silicon cantilevers with a spring constant of 40 N/m and a radius of curvature of less than 10 nm were used (model NSC15/Al BS, MikroMasch, Inc., Portland, OR). For each sample, at least two different scans were obtained at randomly chosen points near the center of the film. Height data were flattened using a second-order fit. Root-mean squared surface roughness (Rrms) was calculated over the scan area using the Nanoscope IIIa software package (Digital Instruments, Santa Barbara, CA). Materials. 1-(2-Amino ethyl) piperidine, poly(sodium 4-styrenesulfonate) (SPS, MW ) 70 000), and sodium acetate buffer were purchased from Aldrich Chemical Co. (Milwaukee, WI). 1,4Butanediol diacylate was purchased from Alfa Aesar Organics (Ward Hill, MA). Test grade n-type silicon wafers were purchased from Si-Tech, Inc. (Topsfield, MA). Linear poly(ethylene imine) (LPEI, MW ) 25 000) was purchased from Polysciences, Inc. (Warrington, PA). Phosphate-buffered saline was prepared by dilution of commercially available concentrate (EM Science, Gibbstown, NJ). Plasmid DNA [pEGFP-N1 (4.7 kb; encoding enhanced green fluorescent protein, EGFP, >95% supercoiled), pDsRed2-N1 (4.7 kb; encoding red fluorescent protein, RFP, > 90% supercoiled), or pCMV-Luc (6.2 kb; encoding firefly luciferase, >90% supercoiled)] was obtained from Elim Biopharmaceuticals, Inc. (San Francisco, CA) or the Waisman Clinical Biomanufacturing Facility at the University of Wisconsin-Madison. For experiments requiring fluorescently labeled plasmid, Cy3 and Cy5 Label-IT nucleic acid labeling kits were purchased from Mirus Bio Corporation (Madison, WI) and used according to the manufacturer’s instructions. Deionized water (18 MΩ) was used for washing steps and to prepare all buffer and polymer solutions. Compressed air used to dry films and coated substrates was filtered through a 0.4 µm membrane syringe filter. All materials were used as received without further purification unless noted otherwise. General Polymerization Procedure. Polymer 2 was synthesized in analogy to methods reported previously.31 Briefly, 1-(2-amino ethyl) piperidine (3 mmol) and 1,4 butanediol diacrylate (3 mmol) were weighed into a vial, and the reaction mixture was heated to 100 °C and stirred for 18 h. The resulting reaction products were dissolved in THF, precipitated into hexanes, and dried under vacuum. The number average molecular weight of the polymer was determined to be 2,100, relative to monodisperse polystyrene standards, with a polydispersity index (PDI) of 4.7. The polymer exhibited a single peak at 1732 cm-1 in the carbonyl region of the infrared spectrum. 1H NMR data: (CDCl , 300.135 MHz) δ (ppm) ) 4.09 (br t, 4H), 3 2.80 (br, t, 4H), 2.60 (br m, 2H), 2.43 (br m, 10H), 1.70 (br m, 4H), 1.56 (br m, 4H), 1.43 (br m, 2H). Preparation of Polyelectrolyte Solutions. Solutions of LPEI and SPS used for the fabrication of LPEI/SPS precursor layers (20 mM with respect to the molecular weight of the polymer repeat unit) were prepared using a 50 mM NaCl solution in water. LPEI solutions contained 5 mM HCl to aid polymer solubility. Solutions of polymer 2 (5 mM with respect to the molecular weight of polymer repeat units) and DNA (1 mg/mL) used for dipping were prepared in sodium acetate buffer (100 mM, pH ) 5.1). Fabrication of Multilayered Films. Films were deposited on planar silicon substrates pre-coated with a multilayered film composed of 10 bilayers of LPEI and SPS (terminated with a topmost layer of SPS) to ensure a suitably charged surface for the adsorption of polymer 2, as previously described.17 These precursor layers were fabricated using an automated dipping robot (Riegler & Kirstein GmbH, Potsdam, Germany). Multilayered films fabricated from DNA and polymer 2 were fabricated on these precursor layers manually

Langmuir, Vol. 23, No. 22, 2007 11141 using an alternate dipping procedure according to the following general protocol: (1) Substrates were submerged in a solution of polyamine for 5 min, (2) substrates were removed and immersed in an initial water bath for 1 min followed by a second water bath for 1 min, (3) substrates were submerged in a solution of DNA for 5 min, and (4) substrates were rinsed in the manner described above. This cycle was repeated until the desired number of polyamine/ DNA bilayers was reached. Films were either used immediately or were dried under a stream of filtered, compressed air and stored in a vacuum desiccator until use. All films were fabricated at ambient room temperature. Characterization of Film Erosion and Release Kinetics. Experiments designed to investigate film erosion and release kinetics were performed in the following general manner: Film-coated substrates were placed in a plastic UV-transparent cuvette, and 1.0 mL of phosphate-buffered saline (PBS, pH ) 7.4, 137 mM NaCl) was added to cover the film-coated portion of the substrate. The samples were incubated at 37 °C and removed at predetermined intervals to be examined by ellipsometry or AFM. Films were rinsed under deionized water and dried under a stream of filtered compressed air prior to measurement. Values were determined in at least four different predetermined locations on the substrate by ellipsometry, and the sample was returned immediately to the buffer solution. For experiments designed to monitor the concentration of DNA in solution, UV absorbance readings were made on the solution used to incubate the sample (at 260 nm, corresponding to the absorbance maximum of DNA). For experiments designed to monitor the concentration of fluorescently labeled DNA in solution, fluorescence measurements were made directly on the solution used to incubate the samples. For release experiments designed to produce samples of DNA for cell transfection experiments (see text), erosion experiments were conducted as described above with the following exceptions: at each predetermined time interval, substrates were removed from the buffer, placed into a new cuvette containing fresh PBS, and the original DNA-containing solution was stored for analysis. Cell Transfection Assays. COS-7 cells were grown in 96-well plates at an initial seeding density of 15 000 cells/well in 200 µL of growth medium (90% Dulbecco’s modified Eagle’s medium, 10% fetal bovine serum, penicillin 100 units/mL, streptomycin 100 µg/mL). Cells were grown for 24 h, at which time 50 µL of a Lipofectamine 2000 (Invitrogen, Carlsbad, CA) and plasmid mixture was added directly to the cells according to the general protocol provided by the manufacturer. The Lipofectamine 2000/plasmid transfection milieu was prepared by mixing 25 µL of the plasmid solution collected at each time point during release experiments (arbitrary concentrations but constant volumes) with 25 µL of diluted Lipofectamine 2000 reagent (25 µL stock diluted into 975 µL of water). Fluorescence images were acquired after 48 h using an Olympus IX70 microscope and analyzed using the Metavue version 4.6 software package (Universal Imaging Corporation).

Results and Discussion Fabrication of Multilayered Films Using Plasmid DNA and Polymer 2. We fabricated multilayered polyelectrolyte films using side-chain functionalized polymer 2, a plasmid DNA construct (pEGFP-N1) encoding enhanced green fluorescent protein (EGFP), and an alternate dipping process similar to that used in our past studies using polymer 1.17-19 Films were fabricated on planar silicon substrates to facilitate the characterization of film growth and erosion using ellipsometry. In all cases, silicon substrates were pre-coated with a multilayered film ∼20 nm thick composed of 10 alternating layers of linear poly(ethylene imine) (LPEI) and sodium poly(styrene sulfonate) (SPS) terminated with a layer of SPS to provide a surface suitable for the adsorption of polymer 2, as described previously for the fabrication of films using polymer 1.17-19 Figure 1 shows a plot of the average optical thickness versus the number of polymer 2/DNA layers (hereafter referred to as

11142 Langmuir, Vol. 23, No. 22, 2007

Figure 1. Plot of ellipsometric thickness versus the number of polymer 2/DNA bilayers deposited on a silicon substrate. Symbols represent average values and error for multiple independent measurements made on three different films.

Figure 2. Plot of change in film thickness (9) and solution absorbance at 260 nm ([) versus time for films fabricated from polymer 2 and plasmid DNA incubated in PBS at 37 °C. Error bars are shown and are in some cases smaller than the symbols used.

“bilayers”) deposited. Inspection of these data reveals film thickness to increase monotonically as a function of the number of bilayers deposited to yield a film ∼150 nm thick after the deposition of eight bilayers. Closer inspection of these data, however, reveals that this film growth profile is composed of two distinct growth regimes. Film thickness increases in a linear manner during the deposition of the first four bilayers, with a slope corresponding to an average bilayer thickness of ∼10.3 nm/bilayer. Film growth continues to increase in a linear manner upon deposition of the final four polymer 2/DNA layers but with a steeper slope corresponding to an average bilayer thickness of ∼22.4 nm/bilayer. This change in growth profile after the deposition of four polymer 2/DNA layers was repeatable over at least 10 different trials and may reflect changes in the influence of the foundation layers of LPEI and SPS on film growth as film thickness increases. Characterization of the surfaces of these films by AFM revealed that the surfaces of these assemblies were rough (rms roughness ∼37 nm), consistent with results reported previously for the surface characterization of films fabricated from polymer 1 and the pEGFP-N1 plasmid.25,28 We return to this observation again in the discussion below. Characterization of Film Erosion and DNA Release Profiles. Past studies have demonstrated that films fabricated from polymer 1 and plasmid DNA erode and release DNA over a period of ∼2 days when incubated in phosphate-buffered saline (PBS).17 We conducted a series of experiments to determine whether films fabricated from side-chain functionalized polymer 2 were capable of eroding when incubated in physiologically relevant media and, subsequently, whether differences in the structures of these two degradable polyamines would lead to differences in film erosion and release profiles. Figure 2 (filled squares) shows a plot of the optical thickness, as determined by

Zhang et al.

ellipsometry, of a polymer 2/DNA film eight bilayers thick incubated in PBS at 37 °C. These data demonstrate that film thickness decreases gradually and in a relatively linear manner, over a period of 300 h (12.5 days). Figure 2 (filled diamonds) also shows a plot of the absorbance (at 260 nm, the absorbance maximum of DNA) of the incubation buffer over this same time period. These data demonstrate that DNA is released into solution gradually and without an initial burst of DNA release, over the same 300-h incubation period. On the basis of these absorbance data, we estimate the amount of DNA in a film eight bilayers thick to be ∼9.1 µg/cm2. This value is approximately three times greater than that estimated for films eight bilayers thick fabricated from polymer 1 under otherwise identical conditions.18 This difference may result from differences in the structures and relative charge densities of polymers 1 and 2 (as discussed below). Characterization of released DNA using agarose gel electrophoresis suggested that ∼40% of the DNA was released as supercoiled DNA, with the remainder of the plasmid being released in an open circular topology (see Figure S1, Supporting Information). Additional cell-based transfection experiments demonstrated that the DNA released from these materials was capable of yielding high levels of EGFP expression in mammalian cells. These transfection results are consistent with the results of past studies using films fabricated using polymer 1 and are discussed in additional detail below (e.g., in Figure 5). The erosion and DNA release profiles shown in Figure 2 differ significantly from those of films fabricated using polymer 1, which generally erode and release DNA more rapidly under otherwise identical conditions (see above). We considered several possible explanations for these large differences based on differences in the molecular weights and chemical structures of these two degradable polyamines. The molecular weight of polymer 2 used in this current study (Mn ∼ 2100) is lower than the molecular weight of polymer 1 used in our past studies (Mn ∼ 10 000). We have thus far been unable to synthesize samples of polymer 2 having a substantially higher average molecular weight. Although polymer molecular weight likely influences certain aspects of the structure and stability of these materials, we note that higher molecular weight polymers would generally be expected to stabilize these ionically crosslinked films relative to lower molecular weight polyamines. As such, the differences in the molecular weights of polymers 1 and 2 noted above likely do not account for the large differences in erosion and release profiles. A more likely explanation arises from a consideration of differences in the relative densities and locations of the amine functionality in these two polymers. For example, polymer 2 has a higher density of protonatable amine functionality per unit mass as compared to polymer 1, and one of the amine groups in polymer 2 is located on the side chain, rather than in the backbone, of the polymer. Our current experiments do not permit quantification of the specific charge densities of these polyamines upon incorporation into multilayered films or relative differences in the ability of side-chain and backbone amines to interact electrostatically with DNA. We speculate, however, that the higher density of amine functionality in polymer 2 may lead to a higher density of ionic crosslinking in these materials, which should serve to stabilize these materials as polymer hydrolysis occurs (e.g., relative to films fabricated from polymer 1). The results of additional experiments using a structural analog of polymer 2 bearing a hydroxyl functionalized side-chain (rather than a protonatable tertiary amine-functionalized side chain) provide additional support for this proposition. Films fabricated from plasmid DNA and this hydroxyl-functionalized polyamine eroded and released DNA rapidly (e.g., over 48 h) when incubated in

Plasmid DNA Multilayered Films

Langmuir, Vol. 23, No. 22, 2007 11143

Figure 3. Representative tapping mode atomic force microscopy images (10 µm × 10 µm) of a polymer 2/DNA film eight bilayers thick deposited on a silicon substrate and subsequently incubated in PBS. Images were acquired under air at each indicated time after drying the film as described in the Materials and Methods section. Values of root-mean squared roughness (Rrms) and optical thickness (h, determined by ellipsometry) are shown below each image. The scale in the z direction is 300 nm for each image.

PBS at 37 °C (see Figure S2, Supporting Information). These results, when combined, suggest that a relative increase in ionic interactions serves to stabilize assemblies fabricated from polymer 2. We conducted an additional series of experiments using AFM to characterize the surface morphologies of polymer 2/DNA films as a function of film erosion. Figure 3 shows a series of nine representative AFM images (10 × 10 µm) and values of rootmean squared roughness (Rrms) acquired during the erosion of a film eight bilayers thick. These images reveal physical changes in the surface of the film from a morphology that was initially uniform and rough (Rrms ∼ 37 nm; as discussed above) to a morphology that was significantly more smooth (Rrms ∼ 15 nm) after 1 h of incubation. Further inspection of these data reveals that the surface of this film remains smooth and uniform as the thickness of the film decreases over a period of ∼160 h (e.g.,

Rrms ∼ 4-5 nm after ∼100 h of incubation; see accompanying values of film thickness, h, as determined by ellipsometry). The surface morphologies of the partially eroded films shown in Figure 3 vary significantly from the behavior of multilayered assemblies fabricated using plasmid DNA and polymer 1. As described above, characterization using AFM and SEM demonstrates that assemblies fabricated from polymer 1 and plasmid DNA undergo large-scale physical transformations upon incubation in PBS and decompose from uniform films into a complex, nanoparticulate morphology.25,28 By contrast, the data in Figure 3 suggest that films fabricated from polymer 2 and plasmid DNA erode and release DNA in a manner that does not involve such large-scale surface transformations. We note here that the data in Figure 3 were acquired using a film that was removed from buffer and dried prior to imaging by AFM and that this drying process could influence the observed morphologies of these

11144 Langmuir, Vol. 23, No. 22, 2007

Figure 4. Plots of DNA released versus time corresponding to the release of fluorescently labeled plasmid DNA from films having the general structure (A) (2/pEGFP-Cy3)4(2/pDsRed-Cy5)4 and (B) (2/ pEGFP-Cy3)2(2/pLuc)4(2/pDsRed-Cy5)2 incubated in PBS at 37 °C. Data points correspond to cumulative amounts of pDsRedCy5 (2) and pEGFP-Cy3 (9) in solution determined from solution fluorescence measurements. Dotted lines correspond to singleexponential fits to these experimental data.

materials. However, these data are consistent with a process of physical film erosion that proceeds in a top-down, surface-type manner that differs fundamentally and significantly from the behavior of films fabricated from polymer 1. Although the mechanism for the large-scale decomposition of films fabricated from polymer 1 is not yet completely understood, past studies suggest that this behavior can be understood in terms of a reduction in ionic interactions within these assemblies that results in increases in the mobility of DNA and polymer 1.28 The results of this current study suggest that the higher density of amine functionality in polymer 2, and the possible ability of the amine functionality in the side chain of the polymer to interact more strongly with DNA, provide a means to prevent large-scale film decomposition in these materials while still permitting gradual film erosion and the release of DNA into solution. Below, we demonstrate that the apparent top-down erosion profiles and physical behavior presented in Figure 3 present opportunities to fabricate multicomponent assemblies capable of providing control over the relative order in which two or more plasmid DNA constructs are released. Characterization and Erosion of Films Fabricated from Two Different Plasmid Constructs. The layer-by-layer procedure used to fabricate the materials above provides a straightforward approach to the assembly of ultrathin films containing multiple layers of multiple different plasmid DNA constructs. In principle, it is possible to fabricate films having the general structure (polyamine/X)n(polyamine/Y)m by deposit-

Zhang et al.

ing n bilayers of a polyamine and a first DNA construct X (deposited first to form the bottommost layers of the film) followed by m bilayers of a polyamine and a second DNA construct Y (deposited last to form the topmost layers of the film).7-13 Two recent studies have demonstrated that films fabricated to have a general hierarchical structure of this type provide opportunities to control the timing and the sequence with which two different types of DNA12 or two different anionic polysaccharides13 are released from multilayered films. For example, Jessel et al. demonstrated recently that multilayered films fabricated from poly(lysine), poly(glutamic acid), and different layers of two different plasmid DNA constructs could be used to provide control over the timing and the order in which these plasmid constructs were expressed by attached cells (e.g., by delaying the expression of DNA located in the lower layers of a film by ∼4 h relative to the expression of DNA located in the topmost portion of the film).12 In a second example, Wood et al. demonstrated that assemblies fabricated from degradable polyamine 1 and multiple different layers of two different anionic polysaccharides could be used to control the rates at which these two agents were released into solution.13 In this latter case, the ability to control the individual release profiles of these two polysaccharides was demonstrated to be dependent upon the incorporation of intermediate layers of chemically crosslinked polyelectrolytes to prevent the interdiffusion of individual polyelectrolyte layers during film assembly and erosion. We conducted a series of experiments to determine whether polymer 2 could be used to fabricate films using two different plasmid DNA constructs and, subsequently, whether the apparent surface-type erosion behavior of these films (as shown in Figure 3) could be used to provide control over the sequence in which two different plasmids were released from a film. To investigate this approach, we fabricated films using the pEGFP-N1 plasmid (as described above) and a second plasmid DNA construct (pDsRed-N1) encoding red fluorescent protein (RFP). The pDsRed2-N1 plasmid is the same size as the pEGFP-N1 plasmid (4.7 kb), and we have shown experimentally that these plasmids behave identically during fabrication (as determined by ellipsometry and UV spectrophotometry, data not shown). For these initial experiments, we labeled the pEGFP plasmid with a Cy3 fluorescent dye (denoted pEGFP-Cy3) and we labeled the pDsRed2 plasmid with a Cy5 fluorescent dye (denoted pDsRedCy5) to permit the characterization of each plasmid independently in subsequent experiments using fluorometry. Control experiments demonstrated that the conjugation of these fluorescent labels did not measurably influence film growth profiles. Figure 4A shows the results of a release experiment conducted using a film having the structure (2/pEGFP-Cy3)4(2/pDsRedCy5)4. Inspection of these data reveals that both plasmids are released into solution in approximately equal amounts over a period of ∼300 h. Further inspection of these data, however, reveals differences in the individual rates at which these two plasmids are released. For example, the pDsRed-Cy5 plasmid (filled triangles), which was incorporated into the final four bilayers deposited during the assembly of the film, is released more rapidly than the pEGFP-Cy3 plasmid (filled squares), which was incorporated into the bottom-most layers of the film. The dotted lines in Figure 4A correspond to single-exponential fits to these experimental data. Analysis of these fits reveals the apparent rate constant for the release of the pDsRed-Cy5 plasmid (kpDsRed ) 0.021 h-1) to be approximately twice that of the rate constant for the release of the pEGFP-Cy3 plasmid (kpEGFP ) 0.012 h-1).

Plasmid DNA Multilayered Films

Langmuir, Vol. 23, No. 22, 2007 11145

Figure 5. Representative low-magnification fluorescence microscopy images (4×) showing relative levels of enhanced green fluorescent protein (EGFP; green channel) and red fluorescent protein (RFP; red channel) expressed in COS-7 cells. Cells were transfected with samples of DNA released from films having the structures (2/pEGFP)2(2/pLuc)4(2/pDsRed)2 (columns I and II) and (2/pDsRed)2(2/pLuc)4(2/pEGFP)2 (columns III and IV), as indicated schematically. Cells were transfected by combining released DNA with Lipofectamine 2000 as a transfection agent (see text). The relative levels of EGFP and RFP observed correspond qualitatively to relative levels of each plasmid released and collected over each of the following time periods: 0-1 h, 1-12 h, 12-23 h, 23-34 h, 34-46 h, and 46-70 h.

The results in Figure 4A demonstrate that films fabricated from polymer 2 can be used to provide control over the relative rates at which two different plasmid DNA constructs are released from an ultrathin erodible film. We note here that, although it is possible to fabricate multilayered polyelectrolyte assemblies having ‘stratified’ architectures,7-13 the internal structures of these materials are generally characterized as having some degree of interpenetration between adjacent polymer layers.1 In addition, as described above, interlayer diffusion can also occur during or after film assembly to yield a more blended or homogeneous internal film structure.13 We conducted additional experiments

to determine whether the incorporation of additional polyelectrolyte layers between bilayers composed of the pDsRed-Cy5 and pEGFP-Cy3 plasmids would provide greater control over the kinetics with which these two plasmids were released into solution. Figure 4B shows the results of a release experiment conducted using a film having the structure (2/pEGFP-Cy3)2(2/pLuc)4(2/pDsRed-Cy5)2. This film was fabricated using the procedure described above, with the exception that four bilayers of polymer 2 and a third plasmid construct (pLuc) were deposited prior to the deposition of the final two polymer 2/pDsRed-Cy5 layers. We selected polymer 2 and another plasmid DNA construct

11146 Langmuir, Vol. 23, No. 22, 2007

for use in these experiments, as opposed to synthetic polyelectrolytes such as those described in past studies,13 because past experiments in our group suggest that synthetic anionic polymers such as SPS can displace plasmid DNA when deposited onto films fabricated from plasmid DNA and poly(β-amino ester)s such as polymer 1. The pLuc plasmid used in these experiments was not fluorescently labeled, and was chosen because it encodes for a protein product (firefly luciferase) that does not fluoresce and would not complicate subsequent characterization of these materials using fluorescence microscopy in cell-based transfection experiments (described below). The data in Figure 4B demonstrate that the incorporation of additional polymer 2/pLuc layers does not change significantly the rates at which these plasmids are released. As described above for the film having the structure (2/pEGFP-Cy3)4(2/ pDsRed-Cy5)4 (Figure 4A), the rate constant for the release of the pDsRed-Cy5 plasmid (kpDsRed ) 0.019 h-1) was approximately twice that of the rate constant for the release of the pEGFP-Cy3 plasmid (kpEGFP ) 0.010 h-1). It may prove possible in future experiments to incorporate additional layers of synthetic polyelectrolytes or other materials to provide additional control over the differences in the rates at which these two plasmids are released from these materials. Approaches based on the chemical crosslinking of intermediate polyelectrolyte layers13 were not pursued in this current study out of concern for the chemical and transcriptional integrity of the DNA in these assemblies. We demonstrate below, however, that the approach reported here is sufficient to exert a significant degree of control over the timing and the order in which the pEGFP and pDsRed plasmids are released and expressed in a functional cell transfection assay. Figure 5 shows a series of low-magnification micrographs of COS-7 cells 48 h after treatment with samples of plasmid DNA collected at various times during the erosion of a multilayered film having the structure (2/pEGFP)2(2/pLuc)4(2/pDsRed)2. This experiment was conducted in the manner described above and in Figure 4, with the exception that the pEGFP and pDsRed plasmids were not fluorescently labeled (see Materials and Methods for a complete description of cell transfection experiments). Inspection of these data (column I) reveals the expression of RFP in cells treated with samples of the incubation solution collected as early as 1 h after the immersion of the coated silicon substrate in PBS and that the number of cells expressing RFP increases significantly over time (time points beyond 70 h not shown). Further inspection of these data (column II) reveals the first significant levels of expression of GFP to be observed in cells treated with the sample of DNA collected over 23-34 h. The results in columns I and II of Figure 5, when combined, demonstrate that films having the structure (2/pEGFP)2(2/pLuc)4(2/pDsRed)2 can be used to provide control over the rates at which these two plasmids are released and expressed by cells at levels sufficient to delay the onset of observable levels of expression of GFP by ∼1 day. We note here that the results of these experiments provide only qualitative measures of the relative levels of each plasmid in solution at a given time. We conducted a final set of experiments to determine whether the delay in the release and expression of the pEGFP plasmid in column II of Figure 5 was a result of the relative location of this plasmid in the bottom of the (2/pEGFP)2(2/pLuc)4(2/ pDsRed)2 film or whether this delay in gene expression could simply result from differences in the relative rates at which the pEGFP and pDsRed plasmids are transcribed by cells. Columns III and IV of Figure 5 show the results of an experiment identical to that shown in columns I and II, with the exception that the bottom-most layers of the film were fabricated using the pDsRed

Zhang et al.

plasmid [i.e., using a film having the reverse structure: (2/ pDsRed)2(2/pLuc)4(2/pEGFP)2]. These data demonstrate clearly that the relative rates of release and expression of the different layers of plasmid DNA in these materials are dictated by the relative order in which the two plasmids are incorporated into these ultrathin assemblies during fabrication and not by the relative rates at which these two plasmids are processed by cells.

Summary and Conclusions We have reported the fabrication and characterization of ultrathin multilayered films fabricated from plasmid DNA and side-chain functionalized poly(β-amino ester) 2. We observed large differences in the behavior of films fabricated from plasmid DNA and polymer 2, as compared to films fabricated from plasmid DNA and polymer 1, when these films were incubated in physiologically relevant media. For example, whereas films fabricated from polymer 1 erode and release DNA over ∼2 days when incubated in phosphate-buffered saline, films fabricated from polymer 2 erode and release DNA over ∼2 weeks. In addition, whereas films fabricated from polymer 1 undergo complex nanometer-scale physical transformations in aqueous media, characterization of films fabricated from polymer 2 by AFM demonstrates that the surfaces of these materials remain smooth and uniform during erosion. A comparison of the structures of these polymers suggests that these large differences in film behavior may arise from differences in the relative densities and locations of the amine functionality in these two polymers. Our results suggest that the design of poly(β-amino ester)s having high densities of amine functionality provide a means to prevent large-scale film decomposition while still permitting gradual film erosion and the release of DNA into solution. The results of broader investigations of the influence of polymer side-chain structure and functionality on the erosion and release of DNA from multilayered films will be reported in a separate contribution. We have demonstrated that the apparent surface-type erosion of films fabricated from polymer 2 and plasmid DNA permits the fabrication of ultrathin films with architectures that provide control over the timing and the order in which two different DNA constructs are released from surfaces. For example, we demonstrated that the order in which two different DNA constructs were released from films could be controlled to a measurable extent by the relative order in which the two DNA constructs were deposited during fabrication. The results of this investigation suggest approaches to the design of ultrathin multilayered films that could be used to promote the localized and sequential release of multiple different DNA constructs from the surfaces of tissue engineering scaffolds or other implantable devices. Acknowledgment. Financial support was provided by the National Institutes of Health (R21 EB02746) and the Arnold and Mabel Beckman Foundation. S.I.M. participated in a NSF/REU summer program at the University of Wisconsin supported through the NSF Materials Research Science and Engineering Center (MRSEC) at UW. We are grateful to the NSF (CHE9208463) and the NIH (NIH 1 S10 RR0 8389-01) for support of the UW NMR spectroscopy facilities. D.M.L. is an Alfred P. Sloan Research Fellow. Supporting Information Available: Characterization of released DNA by agarose gel electrophoresis and results of control experiments for the erosion and release of DNA from a hydroxyl-functionalized derivative of polymer 2. This information is available free of charge via the Internet at http://pubs.acs.org. LA702021S