Gene Transfection into Adherent Cells Using Electroporation on a

Langmuir 2008, 24, 13525-13531

13525

Gene Transfection into Adherent Cells Using Electroporation on a Dendrimer-Modified Gold Electrode S. Koda, Y. Inoue, and H. Iwata* Institute for Frontier Medical Sciences, Kyoto UniVersity, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan ReceiVed July 7, 2008. ReVised Manuscript ReceiVed September 30, 2008 Gene transfection into adherent cells from plasmid DNA (pDNA)-arrayed substrates known as gene transfection arrays appears to be a promising tool for the high-throughput analysis of gene functions and protein-protein interaction networks. We tested the ability of electric pulse-stimulated gene transfection from a substrate to overcome low expression efficiency and cross contamination between spots on arrays. We prepared the electrodes used for electric pulse-stimulated gene transfection by sequentially loading a gold thin layer, a self-assembled monolayer of a carboxylic acid-terminated alkanethiol (COOH-SAM), and poly(amidoamine) (PAMAM) dendrimers, either through electrostatic interactions or by covalent linkage to COOH-SAM and then to pDNA. When dendrimers were loaded onto the electrode using electrostatic interactions, the gene-expression efficiency of adherent cells increased as the generation numbers of the dendrimers that we used increased. Gene expression was rarely observed in adherent cells when dendrimers were covalently immobilized onto the electrode. Additionally, we successfully demonstrated site-specific gene transfer using a dendrimer-array electrode with no cross contamination between spots on the electrode.

Introduction Gene transfection arrays were first described by Sabatini et al. in 2001;1 since then, they have been the focus of intense study regarding their utility for high-throughput analysis of gene-gain or -loss functions.2-12 In particular, nonviral substrate-mediated gene delivery, which is based on the cellular endocytic uptake of nucleic acid complexes with the assistance of cationic liposomes (lipoplex) or cationic polymers (polyplex), has been employed, because of ease of handling and low toxicity. However, several obstacles, such as low expression efficiency and cross contamination between spots, remain to be solved for high-density analysis.13 Several methods have been employed to immobilize plasmid nucleic acids (pDNA) on substrates, such as the gelation of a gelatin or collagen and pDNA mixture,1-3 or electrostatic4-7 and biological interactions8 with the surface. The efficiency of gene delivery into adherent cells depends on the method used to immobilize and release the pDNA. In most cases, gene transfection to cells has been accomplished by spontaneously releasing pDNA from a lipoplex- or polyplex-adsorbed substrate; unfortunately, this often results in low expression efficiency or * Corresponding author. E-mail: [email protected]. (1) Ziauddin, J.; Sabatini, D. M. Nature 2001, 411, 107. (2) Delehanty, J. B.; Shaffer, K. M.; Lin, B. Biosens. Bioelectron. 2004, 20, 773. (3) Chang, F.-H.; Lee, C.-H.; Chen, M.-T.; Kuo, C.-C.; Chiang, Y.-L.; Hang, C.-Y.; Roffler, S. Nucleic Acids Res. 2004, 32, e33. (4) Luo, D.; Saltzman, W. M. Nat. Biotechnol. 2000, 18, 893. (5) Yamauchi, F.; Kato, K.; Iwata, H. Biochim. Biophys. Acta 2004, 1672, 138. (6) Painnier, A. K.; Anderson, B. C.; Shea, L. D. Acta Biomater. 2005, 1, 511. (7) Zhang, J.; Lynn, D. M. AdV. Mater. 2007, 19, 4218. (8) Segura, T.; Volk, M. J.; Shea, L. D. J. Controlled Release 2003, 93, 69. (9) Ovcharenko, D.; Jarvis, R.; Smith, S. H.; Kelnar, K.; Brown, D. Method 2005, 11, 985. (10) Moffat, J.; Grueneberg, D. A.; Yang, X.; Kim, S. Y.; Kloepfer, A. M.; Hinkle, G.; Piqani, B.; Eisenhaure, T. M.; Luo, B.; Grenier, J. K.; Carpenter, A. E.; Foo, S. Y.; Stewart, S. A.; Stockwell, B. R.; Hacohen, N.; Hahn, W. C.; Lander, E. S.; Sabatine, D. M.; Root, D. E. Cell 2006, 124, 1283. (11) Yoshikawa, T.; Uchimura, E.; Kishi, M.; Funeriu, D. P.; Miyake, M.; Miyake, J. J. Controlled Release 2004, 96, 227. (12) Ohtake, N.; Niikura, K.; Suzuki, T.; Nagakawa, K.; Sawa, H.; Ijiro, K. Bioconjugate Chem. 2008, 19, 507. (13) Hook, A. L.; Thissen, H.; Voelcker, N. H. Trends Biotechnol. 2006, 24, 471.

cross contamination. In our previous studies,14-16 we immobilized pDNA on the surface through electrostatic interactions and then released it from the surface by applying an electric pulse. Our electrically triggered transfection, called substrate-based electroporation, appears to be superior in expression efficiency, cell viability, and transfection reproducibility, and there is minimal cross contamination among adjacent spots. Furthermore, we can control the timing of DNA transfection.14 When high-throughput analyses are performed, higher geneexpression efficiency is required because the size of a single spot decreases as the overall spot density on the transfection array increases. A large amount of pDNA must be retained on the spot surface to enhance the quantity of pDNA delivered into adherent cells. Cationic polymers play an important role in our substratebased electroporation. They determine the amount of pDNA immobilized on the spots and the efficiency with which pDNA is released from the surface (Scheme 1A). In previous studies, we examined polyethyleneimines of various molecular weights, which have high amino-group densities within a polymer chain.14 Recently, dendrimers have been suggested to be a superior polycation for gene trasfection.17-20 Dendrimers have weak interactions with solid substrates due to their spherical structure; large amounts of protein or DNA can be retained on dendrimermodified substrates because of the large surface area and high amino-group density21-23 (Scheme 1B). In substrate-based electroporation, dendrimers are expected to result in the retention (14) Yamauchi, F.; Kato, K.; Iwata, H. Nucleic Acids Res. 2004, 32, e187. (15) Yamauchi, F.; Kato, K.; Iwata, H. Langmuir 2005, 21, 8360. (16) Inoue, Y.; Fujimoto, H.; Ogino, T.; Iwata, H. J. R. Soc. Interface 2008, 5, 909. (17) Tomalia, D. A.; Naylor, A. M.; Goddard, W. A., III Angew. Chem., Int. Ed. Engl. 1990, 29, 138. (18) Kukowska-Latallo, J. F.; Bielinska, A. U.; Johnson, J.; Spindler, R.; Tomalia, D. A.; Baker, J. R., Jr. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 4897. (19) Bielinska, A. U.; Chen, C.; Johnson, J.; Baker, J. R., Jr. Bioconjugate Chem. 1999, 10, 843. (20) Zhang, X.; Intra, J.; Salem, K. Bioconjugate Chem. 2007, 18, 2068. (21) Benters, R.; Niemeyer, C. M.; Wohrle, D. ChemBioChem 2001, 2, 686. (22) Benters, R.; Niemeyer, C. M.; Drutschmann, D.; Blohm, D.; Wohrle, D. Nucleic Acids Res. 2002, 30, e10. (23) Pathak, S.; Singh, A. K.; McElhanon, J. R.; Dentinger, P. M. Langmuir 2004, 20, 6075.

10.1021/la8021358 CCC: $40.75  2008 American Chemical Society Published on Web 11/08/2008

13526 Langmuir, Vol. 24, No. 23, 2008

Koda et al.

Scheme 1. (A) Systematic Representation of the Apparatus of an Electro-stimulated Transfection Array Modified with PAMAM Dendrimers: (a) Cell-Adherent Electrode (-), (b) COOH-SAM, (c) Cationic Polymer (PAMAM Dendrimer), (d) pDNA, (e) Cell, and (f) Counter Electrode (+); (B) Chemical Structure of PAMAM Dendrimer (Generation 1)

of large amounts of pDNA on electrodes and to promote the effective release of pDNA into adherent cells on the substrate. In this study, we prepared electrodes for electroporation with ethylene diamine core poly(amidoamine) (PAMAM) dendrimers. Using either electrostatic interactions or covalent immobilization, we loaded the dendrimers onto self-assembled monolayer membranes carrying carboxyl groups to investigate the influence of dendrimer-substrate interactions on reporter-gene expression. We quantitatively estimated the amount of pDNA adsorbed onto the dendrimer-loaded electrodes and then compared the reporter protein expression efficiency between the two dendrimer-loading methods. We also examined the dependence of expression efficiency on the generation numbers of the dendrimers. Finally, we performed site-specific gene transfection with plasmid-arrayed electrodes prepared using a photopatterning method for the purpose of comparison.

Materials and Methods Cell Culture, pDNA, and Dendrimers. Human embryonic kidney cells HEK293 were obtained from the Health Science Research Resources Bank (Osaka, Japan). The cells were routinely maintained in minimal essential medium (MEM; Invitrogen, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 0.1 mg/mL nonessential amino acids, 100 U/mL penicillin, and 0.1 mg/mL streptomycin at 37 °C in a 5% CO2 atmosphere. The pDNAs used in this study were pEGFP-C1 (4.7 kbp) and pDsRed2-C1 (4.7 kbp), encoding an enhanced green fluorescent protein (EGFP) and Discosoma sp. red fluorescent protein (DsRed2), respectively, with a cytomegalovirus (CMV) promoter (Clontech Laboratories, Palo Alto, CA). The pDNAs were expanded in Escherichia coli and purified using the EndoFree Plasmid Maxi Kit (Qiagen, Hilden, Germany) in accordance with the manufacturer’s instructions. The purified pDNAs were dissolved in Tris-EDTA (TE) buffer (10 mM Tris and 1 mM EDTA; pH 8.0) and stored at -20 °C until use. PAMAM dendrimers with generations of 1, 3, 5, and 7 (G1, G3, G5, and G7, respectively) in methanol were purchased from SigmaAldrich. After the methanol was evaporated, the PAMAM dendrimers were dissolved in Dulbecco’s phosphate-buffered saline (PBS) at a concentration of 0.1 mg/mL, and the pH was adjusted to 7.4 with 1 M HCl. Preparation of Dendrimer-Loaded Electrodes. A glass plate (22 mm × 26 mm × 0.5 mm, Matsunami Glass Industry, Osaka, Japan) was treated with piranha solution (7:3 by volume ratio of sulfuric acid and hydrogen peroxide, respectively) at room temperature for 5 min to remove organic impurities. The glass plate was rinsed with double-deionized water (DDW) and 2-propanol and dried under a stream of nitrogen gas. A layer of chromium (1 nm thickness) was deposited onto the glass surface; the next layer (gold, 49 nm thickness) was deposited on the top of the chromium layer in a continuous process using a thermal evaporator (V-KS200, Osaka Vacuum Instrument, Osaka, Japan) operated at a pressure of (3.0-8.0)

× 10-4 Pa. Immediately after deposition, the plate was immersed in a 1 mM 11-mercaptoundecanoic acid (Aldrich Chemical Co., Milwaukee, WI) solution in ethanol at room temperature overnight to form a self-assembled monolayer (SAM) containing carboxylic acid-terminated alkanethiols on the gold-evaporated glass plate (COOH-SAM). The SAM plates were washed with ethanol and DDW and dried under a stream of nitrogen gas. Two types of PAMAM dendrimer-loaded electrodes were prepared using different interactions between the dendrimers and the COOHSAM surface. The first, the dendrimer-adsorbed electrode, was prepared using electrostatic adsorption. The second, the dendrimerimmobilized electrode, was prepared using covalent immobilization. The COOH groups in the SAM were activated using a mixed solution of 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride (WSC, 0.1 mol/L) and N-hydroxysuccinimide (NHS, 0.05 mol/L) in DDW for 10 min. Activated or nonactivated COOH-SAM was exposed to the PAMAM dendrimer solution in PBS at a concentration of 0.1 mg/mL for 30 min at room temperature to covalently immobilize or to electrostatically adsorb the PAMAM dendrimer onto the electrode surface, respectively. Finally, the dendrimerloaded electrodes were washed with DDW to remove weakly adsorbed dendrimers. Preparation of pDNA-Adsorbed Electrodes for Physicochemical Analysis. Adsorption of pDNA onto the dendrimer-loaded electrodes was carried out in sterile conditions. The dendrimerloaded electrodes were rinsed with 70% ethanol and air-dried in a laminar flow hood. The electrodes were then exposed to pEGFP-C1 in PBS solution (5 µg/mL, pH 7.4) for 3 h at room temperature to allow pEGFP-C1 to electrostatically adsorb onto the electrode surfaces. The resulting pEGFP-C1 adsorbed electrodes were extensively washed with PBS to remove any weakly bound pEGFPC1 from the electrode surfaces, and the remaining salts were washed from the electrodes with DDW. Fourier Transform Infrared-Reflection Absorption Spectroscopy (FTIR-RAS). Infrared absorption spectra of sample surfaces were collected by the reflection-absorption method using a Spectrum One (PerkinElmer, Boston, MA) spectrometer equipped with a Reflector (Harrick Scientific Co., NY) and a mercury-cadmium telluride (MCT) detector cooled by liquid nitrogen (FTIR-RAS). Glass plates (22 mm × 30 mm × 1.0 mm) coated with a 199 nm thick layer of gold were used for FTIR-RAS analyses. Spectra were obtained using a p-polarized infrared laser beam at an incident angle of 75° in a chamber purged with dry nitrogen gas (Taiyo Toyo Sanso Co., Osaka, Japan) for 128 scans at 4 cm-1 resolution from 4000 to 750 cm-1. FTIR-RAS analyses were performed for the COOHSAM and dendrimer-loaded electrodes with different generation numbers, and for the pEGFP-C1 adsorbed dendrimer-loaded electrodes, using a bare gold electrode as the reference surface. Surface Plasmon Resonance Measurement. We assembled a surface plasmon resonance (SPR) instrument with a Kretschmann configuration24 and utilized it to follow the kinetics of surface (24) Hirata, I.; Morimoto, Y.; Murakami, Y.; Iwata, H.; Kitano, E.; Kitamura, H.; Ikada, Y. Colloids Surf., B 2000, 18, 285.

Gene Transfer from Dendrimer-Loaded Electrodes

Langmuir, Vol. 24, No. 23, 2008 13527

Table 1. PAMAM Dendrimer Properties PAMAM dendrimer

generation number

molecular weight

number of amino group

G1 G3 G5 G7

1 3 5 5

1430 6909 28 825 116 490

8 32 128 512

modification. For SPR analyses, a COOH-SAM was formed on goldevaporated BK7 glass plates (1.515 refractive index and 25 mm × 25 mm × 1.0 mm dimensions; Sigma Koki, Tokyo, Japan) (49 nm thick gold layer), as described above. The beam of a HeNe laser (λ ) 632.8 nm, 5 mW; Sigma Koki) was linearly p-polarized and entered through a hemicylinder prism coupled to the gold layer on the glass plate with immersion oil (refractive index ) 1.515; Cargille Laboratories, Ceder Grove, NJ). All of the experiments were performed at 30 °C. Dendrimers were covalently immobilized on the COOH-SAM surface in an SPR measurement cell. The COOH-SAM surface was first exposed to DDW, and then to WSC/NHS solution (0.1 mol/L WSC and 0.05 mol/L NHS in DDW) for 10 min, followed by a stream of DDW for an additional 10 s to wash the WSC/NHS solution from the surface. Next, 0.1 mg/mL dendrimers in PBS solution flowed into the SPR cell for 20 min, followed by a stream of DDW for 5 min to remove weakly adsorbed dendrimers. After the DDW was replaced with PBS by allowing PBS to flow into the cell for 5 min, pEGFP-C1 in PBS (5 µg/mL) was perfused into the cell for 25 min, followed by a stream of PBS for 5 min to wash the surface. To measure by SPR the immobilization of pEGFP-C1 on the electrostatically adsorbed dendrimers on the COOH-SAM surface, the procedures were the same as those described above, except that the WSC/NHS activation step was skipped. The amount of adsorbed dendrimers and pEGFP-C1 was determined by SPR angle shifts, using the following relationship:24 amount of adsorbed dendrimer and pEGFP-C1 (ng/cm2) ) 500 × increase of the resonance angle (deg). Transfection of pDNA to Cells on the pDNA-Loaded Electrode. A pDNA transfection setup is schematically depicted in Scheme 1. A sterilized silicone frame (14 mm × 14 mm inner area, 0.5 mm thickness) was attached to the electrodes to confine the cell-culture region. A suspension of HEK293 cells was seeded onto the pEGFPC1-adsorbed electrodes at a density of 2.0 × 104 cells/cm2, and the HEK293 cells were cultured at 37 °C in a 5% CO2 atmosphere for 24 h to ensure cell attachment onto the pEGFP-C1-adsorbed electrode surface. The attached HEK293 cells were washed with PBS, and a counter electrode (gold-evaporated glass plate; 1 nm chromium and 199 nm gold) was placed on the frame with the gold surface facing down and above the pEGFP-C1-adsorbed electrode. The space between the two gold electrodes was filled with PBS. An electric pulse was applied twice at an interval of 1 s using a pulse generator (GenePulser Xcell, Bio-Rad, Hercules, CA) between the electrode with adherent cells and the counter electrode at a pulse strength of 280 V/cm for a 5 ms duration at 4 °C. After the pulse application, the cells were cultured in MEM with 10% FBS for an additional 48 h. The cells in PBS containing calcium and magnesium ions (1 mM CaCl2 and 1 mM MgCl2) were monitored for EGFP expression using an epifluorescence microscope (BX-71, Olympus, Tokyo, Japan). The cells adhering to the electrodes were collected by trypsinization (2.5 mg/mL trypsin for 3 min at 37 °C), and the fraction of cells expressing EGFP was quantitatively evaluated by using a fluorescent-activated cell sorting (FACS) apparatus (FACSCalibur, Becton Dickinson, Franklin Lakes, NJ). Twenty thousand cells were counted be FACS. A threshold value of fluorescence intensity between EGFP-positive and -negative cells was set at the EGFP-negative region, which contained 99.5% of the nontransfected cells. Site-Specific Gene Transfection. pDNA spots were arrayed on dendrimer-loaded electrodes for the site-specific gene transfer. Goldevaporated glass plates were immersed into an ethanol solution of 1-hexadecanthiol (Tokyo Kasei Kogyo Co., Tokyo, Japan) to prepare the SAM-carrying methyl groups (CH3-SAM). The CH3-SAM surface

Figure 1. FTIR-RAS spectra of electrode surfaces. (a) COOH-SAM, (b) G3 dendrimer-adsorbed electrode, (c) G5 dendrimer-adsorbed electrode, (d) G7 dendrimer-adsorbed electrode, (e) G7 dendrimer-immobilized electrode, (f) pEGFP C1-adsorbed G7 dendrimer-adsorbed electrode, and (g) pEGFP-C1-adsorbed G7 dendrimer-immobilized electrode.

was irradiated with ultraviolet light (180 mW/cm2, Ushio Tokyo, Japan) through a photomask with an array of transparent and circular dots (1 mm φ, 5 × 5 dots). The plates were washed with ethanol to remove photodegradation products and then immersed in an 11mercaptundecanoic acid solution for 1 h to form a COOH-SAM within the irradiated regions. A G7-dendrimer solution in PBS was applied to the COOH-SAM regions by manual pipetting. After sterilization with 70% ethanol and air-drying of the G7 dendrimeradsorbed electrode in a laminar flow hood, 5 µg/mL PBS solutions of either pEGFP-C1 or pDsRed2-C1 were applied to the dendrimercontaining spots by manual pipetting. Gene transfection with an electric pulse was performed as described above except that the pulse duration time was 2.5 ms. Microscopic observation was performed with an epifluorescence microscope after the cells were cultured in MEM with 10% FBS for 48 h.

Results The characteristics of the PAMAM dendrimers employed in this study are summarized in Table 1. IR-RAS spectra of the COOH-SAM, dendrimer-adsorbed electrodes with different generation numbers, a G7 dendrimer-immobilized electrode, and pEGFP-C1-adsorbed dendrimer-loaded electrodes are shown in Figure 1. The peaks at 1640 cm-1 and 1515-1570 cm-1, attributed to amide I absorption band of the CdO stretching region and amide II adsorption band of the N-H stretching region, respectively, were observed for the dendrimer-loaded electrodes, but not for the COOH-SAM surface. The appearance of these two peaks indicates that the dendrimers were successfully loaded onto the COOH-SAM surface. The peak areas of amide I or II absorption increased with the increase in dendrimer generation numbers on the electrode. Furthermore, the peak at 1742 cm-1 for the dendrimer-loaded surface, attributed to the CdO stretching mode of the carboxylic acids in the COOH-SAM surface, decreased with the increase in the dendrimer generation numbers. The decrease in the peak intensity at 1742 cm-1 and the appearance of a new carboxylate peak at 1560 cm-1 is due to salt formation between the carboxylic acid groups and the amino groups of the dendrimers.25 These results indicate that the dendrimers were effectively loaded onto the COOH-SAM surface and a larger number of amino groups was introduced onto the electrodes (25) Yoo, D.; Shiratori, S. S.; Rubner, M. F. Macromolecules 1998, 31, 4309.

13528 Langmuir, Vol. 24, No. 23, 2008

Koda et al.

Figure 2. SPR sensorgrams during deposition of G7 dendrimers and pEGFP-C1 on (a) the COOH-SAM surface and (b) the COOH-SAM surface activated with WSC/NHS solution.

when higher-generation dendrimers were used. The peak intensities of amide I or II absorption for the G7 dendrimer-immobilized electrode were weak when compared to that for the G7 dendrimeradsorbed electrode. This result indicates that the number of amino groups on the G7 dendrimer-adsorbed electrode was larger than that on the G7 dendrimer-immobilized electrode. Clear absorbance peaks appeared at 1095, 1235, and 1649 cm-1 when pEGFP-C1 was applied to the G7 dendrimer-loaded electrode (Figure 1f and g). These peaks were attributed to the symmetric (1095 cm-1) and antisymmetric (1235 cm-1) stretching vibrations of phosphate groups in the DNA backbone, and to overlapping of the vibration modes of amino groups of dendrimers (1640 cm-1) and nucleotides (1649 cm-1).26 These IR-RAS spectral data confirm the adsorption of pEGFP-C1 onto the dendrimer-loaded electrode. To quantitatively estimate the amounts of dendrimers and pEGFP-C1 adsorbed onto the electrodes, we performed SPR analyses using dendrimers of different generations and pEGFPC1. Figure 2 shows the typical SPR profiles of G7 dendrimers with or without activation of the COOH-SAM by WSC/NHS solution. An SPR profile of the electrostatic adsorption of G7 dendrimer onto the COOH-SAM surface and the subsequent adsorption of pEGFP-C1 onto the surface is shown in Figure 2a. The angle value difference between times 0 and 30 min was 400 mDA. This angle shift was caused by the adsorption of G7 dendrimer onto the COOH-SAM surface. After the surface was washed with PBS, pEGFP-C1 in PBS was applied for 37-62 min, and then the surface was washed again with PBS. Adsorption of pEGFP-C1 onto the dendrimer-adsorbed surface induced a 600 mDA angle shift. Figure 2b shows the SPR profile of the covalent immobilization of the G7 dendrimer onto the COOHSAM surface activated with WSC/NHS solution, and the subsequent adsorption of pEGFP-C1 onto the dendrimer-loaded surface. Figure 2a and b shows that the SPR profiles of these two electrodes, prepared by the electrostatic adsorption or covalent immobilization of G7 dendrimers, were quantitatively the same. The density of amino groups, calculated from the amounts of dendrimer and pEGFP-C1 on the electrodes, is summarized in Figure 3. As the generation number of the dendrimer increased, the density of amino groups on the electrode increased. Moreover, the density of amino groups on the G7 dendrimer-immobilized electrode was smaller than that on the G7 dendrimer-adsorbed electrode. These results are also supported by the IR-RAS spectra shown in Figure 1. Similarly, the amount of pEGFP-C1 on the dendrimer-loaded electrode increased as the generation number of dendrimers increased. The amount of pEGFP-C1 adsorbed onto the dendrimer-loaded electrode strongly depended on the (26) Zhou, Y.; Li, Y. Biophys. Chem. 2004, 107, 273.

Figure 3. Relationship between dendrimer generation numbers and the density of amino groups (line) or adsorbed amounts of pEGFP-C1 (bars) on electrodes calculated from SPR angle shift values. To calculate the density of amino groups on electrodes, we used the dendrimer properties shown in Table 1. *p < 0.05.

surface density of amino groups on the electrode and could be controlled by the generation number of dendrimers used. A suspension of HEK293 cells was applied to the pEGFPC1-adsorbed electrodes at a density of 2.0 × 104 cells/cm2, and the cells were cultured for 24 h to adhere to the electrode. Complexes of dendrimers and pDNA or pDNA only were released from the electrode and delivered into adherent cells by the application of an electric pulse. Figure 4 shows the phase contrast and fluorescent microscopic images of cells 48 h after the electric pulse. From the phase contrast microscopic images, it appears that most of the cells were viable on the dendrimer-loaded electrodes after the pulse application. This was also supported by propidium iodide staining (>95% viable). Fluorescent microscopic images of the cells showed that EGFP was successfully expressed after the electric pulse application. Adsorbed pEGFP-C1 on the dendrimer-adsorbed electrode was transfected into the cells by the electric trigger. On the other hand, very few cells expressed EGFP without the pulse application (Figure 4e). Without the electric pulse application, either pores were not made in the cell membrane or only very small amounts of pEGFP-C1 were released from the electrode. The number of cells expressing EGFP on the dendrimer-adsorbed electrode at 48 h after pulse application increased with higher-generation dendrimers (Figure 4a-d). Figure 4 includes the results of pEGFP-

Gene Transfer from Dendrimer-Loaded Electrodes

Langmuir, Vol. 24, No. 23, 2008 13529

Figure 4. Phase contrast and fluorescent microscopic images of HEK293 cells adhering to pEGFP-C1 adsorbed electrodes. (a) G1 dendrimer-adsorbed electrode, (b) G3 dendrimer-adsorbed electrode, (c) G5 dendrimer-adsorbed electrode, (d) G7 dendrimer-adsorbed electrode, (e) G7 dendrimeradsorbed electrode without pulse application, and (f) G7 dendrimer-immobilized electrode. Scale bar: 500 µm.

Figure 5. FACS analyses of HEK293 cells transfected with pEGFP-C1 on electrodes. (A) Histogram of FACS of live cells transfected on G7 dendrimer-adsorbed electrodes (a) and naive cells (b). (B) Relationship between the amount of adsorbed pEGFP-C1 on dendrimer-loaded electrodes and EGFP expression efficiency. O, Dendrimer-adsorbed electrode; b, Dendrimer-immobilized electrode.

C1 transfection into cells on an electrode on which G7 dendrimer was covalently immobilized. As shown in Figure 4f, only a few cells expressed EGFP. Release of pEGFP-C1 as a complex with dendrimers is essentially important for EGFP expression in these cells. EGFP expression efficiency was quantitatively estimated with FACS analysis. Figure 5A shows an example of a histogram plotting cell numbers against the fluorescent intensity of EGFP for cells transfected with pEGFP-C1 on a G7 dendrimer-adsorbed electrode, and against naive cells. A threshold value for fluorescence intensity between EGFP-positive and -negative cells was set at the EGFP-negative region, which contained 99.5% of the naive cells. We used FACS to examine EGFP expression

efficiency on electrodes with different generation dendrimers. The expression efficiency was plotted as a function of the amount of plasmid on the electrodes (Figure 5B). Our analyses demonstrated that the EGFP expression efficiency for cells on the dendrimer-adsorbed electrode increased with higher-generation dendrimers, as was seen by fluorescence microscopy (Figure 4a-d). On the other hand, the expression efficiency resulting from the dendrimer-immobilized electrode was quite low (