Journal of Controlled Release 226 (2016) 205–216

Contents lists available at ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

Self-assembled ternary complexes stabilized with hyaluronic acid-green tea catechin conjugates for targeted gene delivery Kun Liang, Ki Hyun Bae, Fan Lee, Keming Xu, Joo Eun Chung, Shu Jun Gao, Motoichi Kurisawa ⁎ Institute of Bioengineering and Nanotechnology, 31 Biopolis Way, The Nanos, 138669, Singapore

a r t i c l e

i n f o

Article history: Received 6 October 2015 Received in revised form 8 January 2016 Accepted 3 February 2016 Available online xxxx Keywords: Hyaluronic acid Catechin Ternary complex CD44 targeting Gene delivery

a b s t r a c t Nanosized polyelectrolyte complexes are attractive delivery vehicles for the transfer of therapeutic genes to diseased cells. Here we report the application of self-assembled ternary complexes constructed with plasmid DNA, branched polyethylenimine and hyaluronic acid-green tea catechin conjugates for targeted gene delivery. These conjugates not only stabilize plasmid DNA/polyethylenimine complexes via the strong DNA-binding affinity of green tea catechin, but also facilitate their transport into CD44-overexpressing cells via receptor-mediated endocytosis. The hydrodynamic size, surface charge and physical stability of the complexes are characterized. We demonstrate that the stabilized ternary complexes display enhanced resistance to nuclease attack and polyanion-induced dissociation. Moreover, the ternary complexes can efficiently transfect the difficult-totransfect HCT-116 colon cancer cell line even in serum-supplemented media due to their enhanced stability and CD44-targeting ability. Confocal microscopic analysis demonstrates that the stabilized ternary complexes are able to promote the nuclear transport of plasmid DNA more effectively than binary complexes and hyaluronic acid-coated ternary complexes. The present study suggests that the ternary complexes stabilized with hyaluronic acid-green tea catechin conjugates can be widely utilized for CD44-targeted delivery of nucleic acid-based therapeutics. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Gene therapy has emerged as a potent therapeutic approach for the treatment of chronic diseases including cancer, monogenic and cardiovascular disease [1]. The success of gene therapy ultimately relies on the development of delivery systems capable of efficiently transferring exogenous genes into target cells. One major approach in gene delivery is to exploit non-viral vectors based on the polyelectrolyte complexes self-assembled through electrostatic interactions between cationic polymers and anionic nucleic acids. A variety of cationic polymers, such as poly(L-lysine), poly(β-amino ester) and poly(amido amine) dendrimers have been utilized to condense plasmid DNA (pDNA) into nanoscale polyelectrolyte complexes favorable for cell internalization [2–4]. Among the cationic polymers, polyethylenimine (PEI) has attracted significant interest because of its excellent transfecting ability with an intrinsic endosomolytic activity. It has been reported that the strong buffering capacity of PEI can facilitate the endosomal escape of pDNA by rupturing the endosome membrane via the “proton sponge” effect [5,6]. Despite the desirable features, there are several limitations to the practical applications of pDNA/PEI complexes: 1) strong cytotoxicity due to disruption of the cell membrane with positively charged PEI chains [7,8], 2) agglutination with blood components (e.g., erythrocytes ⁎ Corresponding author. E-mail address: [email protected] (M. Kurisawa).

http://dx.doi.org/10.1016/j.jconrel.2016.02.004 0168-3659/© 2016 Elsevier B.V. All rights reserved.

and complements) [9,10], and 3) dissociation of pDNA from the complexes due to competitive binding of polyanions (e.g., heparin and plasma proteins) and subsequent degradation of the liberated DNA by serum nucleases [11,12]. Electrostatic polymer coating has recently been proposed as a promising strategy to overcome the aforementioned limitations of pDNA/PEI complexes. Various anionic polymers such as poly(acrylic acid), poly(γglutamic acid) and carboxylated poly(ethylene glycol) derivatives have been electrostatically deposited on a cationic pDNA/PEI complex to alter its surface properties [13–15]. Kurosaki et al. have discovered that the poly(γ-glutamic acid) coating effectively reduces the cytotoxicity and agglutination of pDNA/PEI complexes by shielding their positive surface charges [14]. In addition, carboxylated poly(ethylene glycol) derivatives have been shown to protect pDNA/PEI complexes against non-specific interaction with plasma proteins [15]. Hyaluronic acid (HA), a natural polysaccharide, has also been investigated as a shielding material for pDNA/PEI complexes. HA is a linear non-sulfated glycosaminoglycan with non-immunogenic, biocompatible and biodegradable nature [16]. Since HA serves as a ligand for CD44, HA not only provides a protective coating on pDNA/PEI complexes, but also enables these complexes to target the cancer cells and liver endothelial cells that express high levels of CD44 [17,18]. HA-coated pDNA/PEI complexes have been reported to exhibit superior gene transfection efficiency compared to uncoated complexes in CD44-overexpressing cells via HA receptor-mediated uptake [19,20]. Recently, considerable efforts have been focused towards

206

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

developing chemically modified HA derivatives to further improve the stability, serum tolerance and transfecting ability. Ito et al. have described the transcription-enhancing activity of amphoteric spermine-HA conjugates mimicking the high-mobility-group (HMG) proteins essential for initiating transcription [21]. Furthermore, a disulfide-modified HA (HA-SS–COOH) has been explored to selectively trigger the release of pDNA in the intracellular reductive environment to enhance gene delivery efficiency [22]. In this study, we described the development of self-assembled ternary complexes constructed with pDNA, PEI and HA-epigallocatechin gallate (HA-EGCG) conjugates. EGCG is a major constituent of green tea catechins with beneficial properties including antioxidant [23], anti-inflammation [24] and anticancer effects [25]. Importantly, EGCG has been shown to bind strongly to biological macromolecules such as DNA and proteins via hydrogen bonding, π–π stacking and hydrophobic interactions [26–28]. Moreover, EGCG is known to inhibit the activity of various enzymes such as nucleases, collagenase and hyaluronidase by blocking their active sites irreversibly [29–31]. We have recently reported that macromolecular EGCG derivatives can form a compact nanocomplex with proteins and effectively protect them from degradation by serum proteases in vivo [27]. Inspired by the desirable properties of EGCG, we have designed nanoscale ternary complexes stabilized with HA-EGCG conjugates for CD44-targeted gene delivery. Not only did HAEGCG conjugate stabilize pDNA/PEI complexes via the strong DNAbinding affinity of EGCG, it also assisted their transport into CD44overexpressing cells via HA receptor mediated endocytosis. The stabilized ternary complexes exhibited enhanced protection of pDNA against nuclease attack and polyanion-induced dissociation. We demonstrated that these ternary complexes enable highly efficient transfection of the difficult-to-transfect HCT-116 colon cancer cell line even in serum-supplemented conditions. This was due to their enhanced stability, resistance to nuclease and CD44-targeting ability, which led to more effective nuclear transport of pDNA as compared to binary complexes and HA-coated ternary complexes. The current study suggested that the HA-EGCG-stabilized ternary complexes could be potentially applied for CD44-targeted delivery of nucleotide therapeutics to diseased cells or tissues. 2. Materials and methods 2.1. Materials HA (Mw = 90 kDa) was kindly donated by JNC Corporation (Tokyo, Japan). Branched PEI (Mw = 25 kDa), heparin sodium salt and 2,2diethoxyethylamine were purchased from Sigma-Aldrich (St. Louis, MO). EGCG was obtained from DSM Nutritional Products Ltd. (Basel, Switzerland). Green fluorescent protein (GFP)-encoding pDNA (pEGFP-C1, 4.7 kbp) was purchased from Clontech Laboratories (Mountain View, CA). SYBR® Safe DNA gel staining dye, AlamarBlue® cell viability assay reagent, Lipofectamine 2000 transfection reagent and LysoSensor® Green DND-153 (Life Technologies, Carlsbad, CA) were used according to the manufacturers' instructions. DNase I and DNase I reaction buffer were obtained from New England Biolabs (Ipswich, MA). Anti-human CD44 antibody, isotype control antibody and fluorescein isothiocyanate (FITC)-tagged secondary antibody were purchased from Bio-Rad Laboratories (Hercules, CA). Cy5 Label IT® Tracker intracellular nucleic acid localization kit was purchased from Mirus Bio (Madison, WI). Lab-Tek® II chamber slides were obtained from Thermo Fisher Scientific (Waltham, MA). All other chemicals and reagents were of analytical grade. 2.2. Synthesis of HA-EGCG conjugates HA-EGCG conjugates were synthesized in a two-step process established previously [30,32]. Firstly, EGCG was reacted with 2,2diethoxyethylamine (DA) to form ethylamine-bridged EGCG dimers.

In brief, 145 μL of DA (1 mmol) was added to 1.2 mL of cold MSA:THF (1:5) while stirring. The mixture was then added to EGCG (2.29 g, 5 mmol) dissolved in 3.8 mL of THF containing 1.7 μL of MSA and stirred overnight in the dark at room temperature. The unreacted EGCG was removed by multiple extraction cycles with ethyl acetate until no free EGCG was detected. The concentration of the purified ethylaminebridged EGCG dimer was determined by absorbance at 274 nm and was found to be 84 mg/ml (yield = 88%). In the second step, the ethylamine-bridged EGCG dimers were conjugated to HA via carbodiimide-mediated coupling reaction. In brief, HA (250 mg, 0.62 mmol) was dissolved in 19.8 mL of 0.4 M MES buffer (pH 5.2) containing 2.5 mL of DMF. NHS (89 mg, 0.78 mmol), ethylamine-bridged EGCG dimers (0.205 mmol in 2.7 mL of H2O) and EDC⋅HCl (150 mg, 0.78 mmol) were added successively and the pH of the mixture was adjusted to 4.7. The reaction mixture was purged vigorously with N2 for 10 min and then incubated overnight under N2. The HA-EGCG conjugates were then purified by three cycles of ethanol precipitation in the presence of NaCl. Subsequently, the precipitates were re-dissolved in 150 mL of H2O and dialyzed against H2O in N2 atmosphere overnight before lyophilization. The final yield was 74.4%. The degree of substitution (i.e., the number of EGCG dimers per 100 disaccharide units in HA) was determined by examining the absorbance of HA-EGCG conjugates at 274 nm using a Hitachi U-2810 spectrometer. The degree of substitution for HA-EGCG conjugates was determined to be approximately 2.5. 2.3. Preparation and characterization of polyelectrolyte complexes To prepare pDNA/PEI complexes, 1 μg of pDNA was gently mixed with branched PEI in 50 μL of deionized water at various N/P ratios (i.e., the ratios of the number of amino groups in PEI to the number of phosphate groups in pDNA). This mixture was incubated for 15 min at room temperature. The resulting pDNA/PEI complexes were mixed with HA-EGCG conjugates in 50 μL of deionized water at various C/P ratios (i.e., the ratios of the number of carboxylic groups in HA to the number of phosphate groups in pDNA), and then incubated for 30 min to form pDNA/PEI/HA-EGCG ternary complexes. For comparison, pDNA/ PEI/HA ternary complexes were prepared by mixing HA with the pDNA/PEI complexes at equivalent C/P ratios and subsequently incubating for 30 min. The hydrodynamic diameters of the polyelectrolyte complexes were measured using a particle size analyzer (Zetasizer Nano ZS, Malvern Instruments, UK). Each sample was diluted with deionized water to produce a final pDNA concentration of 1 μg mL−1. The zeta potential values were measured in 10 mM NaCl solution. All measurements were performed in triplicate. 2.4. Evaluation of pDNA condensation and physical stability of polyelectrolyte complexes The effect of ternary complex formation on pDNA condensation was assessed by agarose gel electrophoresis. Briefly, 10 μL of the polyelectrolyte complexes formulated with 1 μg of pDNA were applied to a 1.2% (w/v) agarose gel containing the SYBR® Safe DNA gel staining dye. The gel was subjected to electrophoresis at 100 V for 25 min in TAE buffer solution (40 mM Tris, 20 mM acetic acid and 1 mM EDTA). The gel image was taken under UV illumination by using a VersaDoc 4000 MP instrument (Bio-Rad Laboratories, Singapore). To examine the stability against polyanion-induced dissociation, the polyelectrolyte complexes (1 μg of pDNA) were incubated with heparin sodium salt solution (1 mg mL−1) for 30 min at 37 °C, and then analyzed by agarose gel electrophoresis [33] as described above. To evaluate nuclease resistance, the polyelectrolyte complexes (1 μg of pDNA) were incubated with 16 units of DNase I for 2 h at 37 °C [33]. The mixture was then treated with 0.5 mM EDTA solution for 15 min to inactivate DNase I. The resulting solution was treated with heparin sodium salt solution (2 mg mL−1) for 30 min to liberate pDNA from the complexes prior to agarose gel electrophoresis.

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

2.5. Cell culture Human colorectal carcinoma HCT-116 cells and human embryonic kidney HEK293 cells were obtained from American Type Culture Collection (Rockville, MD). They were propagated to confluence in T-75 flasks in 10 mL of McCoy's 5 A medium and Dulbecco's modified Eagle's medium respectively supplemented with 10% (v/v) fetal bovine serum and 1% (v/v) penicillin/streptomycin. The cells were maintained at 37 °C in a humidified atmosphere of 5% CO2. 2.6. Quantitative assessment of CD44 expression in cell lines To examine the CD44 expression levels, 5 × 105 cells were incubated with anti-human CD44 antibody or isotype control antibody (2 μg mL−1) for 15 min at room temperature. The cells were washed three times with ice-cold 10 mM phosphate-buffered saline (PBS, pH 7.4) containing 10% (v/v) FBS and then stained with FITC-tagged secondary antibody for 20 min. Afterwards, the cells were washed again and analyzed using a BD LSR II flow cytometer system (BD Biosciences, CA). 2.7. Gene transfection experiments HCT-116 cells and HEK293 cells were seeded on 24-well plates at a density of 1 × 105 cells per well and then cultivated for 24 h at 37 °C. These cells were incubated with 1 mL of the serum-supplemented media containing pDNA/PEI, pDNA/PEI/HA-EGCG or pDNA/PEI/HA complexes formulated with 1.5 μg of pDNA. The pDNA/PEI complexes formed at an optimal N/P ratio of 30 were used to prepare pDNA/PEI/ HA-EGCG and pDNA/PEI/HA ternary complexes at C/P ratios ranging from 0.1 to 10. Lipofectamine 2000 was used as a positive control and gene transfection was conducted according to the manufacturer's protocols at an optimal ratio of 1:2 (pDNA (in μg):Lipofectamine 2000 (in μL)). The transfection experiments were performed for 72 h (HCT-116 cells) or 48 h (HEK293 cells). The expression of GFP proteins in the cells was observed by an Olympus IX71 inverted fluorescence microscope equipped with a DP70 camera (Olympus Corporation, Japan). The transfection efficiency was quantitatively analyzed by flow cytometry. Briefly, the transfected cells were harvested, resuspended in PBS, and then analyzed by using a BD LSR II flow cytometer system (BD Biosciences, CA). The percentage of GFP-positive cells was determined by quantifying the fluorescent cells within the gated region from total 10,000 cells. The analytical gate was chosen as less than 2% of untreated control cells falling within the GFP-positive region. The mean fluorescence intensity was also determined by measuring the cellular fluorescence within the gated region. To examine the effect of CD44mediated uptake on the transfection efficiency, HCT-116 cells seeded on 24-well plates (1 × 105 cells per well) were incubated with 1 mL of media containing HA at various concentrations [22] or CD44 antibody [34] at a final concentration of 15 μg mL− 1. After 1 h pre-treatment, the cells were transfected with pDNA/PEI/HA-EGCG ternary complexes (1.5 μg of pDNA) formed at an optimal C/P ratio of 0.5 or pDNA/PEI binary complexes at N/P ratio 30 for 72 h, and then analyzed by flow cytometry as described above. 2.8. Cell viability assay HCT-116 cells and HEK293 cells were seeded on 96-well plates at a density of 1.5 × 104 cells per well and then cultivated for 24 h at 37 °C. These cells were incubated with 200 μL of the serumsupplemented media containing pDNA/PEI, pDNA/PEI/HA-EGCG or pDNA/PEI/HA complexes formulated with 0.3 μg of pDNA for 72 h (HCT-116 cells) or 48 h (HEK293 cells). Cell viability was evaluated by using the AlamarBlue® cell viability assay reagent which measures cellular metabolic reduction. Briefly, 100 μL of culture media containing 10% (v/v) AlamarBlue® reagent were added to each well of the 96-

207

well plates. After incubation for 2 h at 37 °C, the cellular fluorescence was measured using a Tecan Infinite microplate reader (Tecan Group, Switzerland) with an excitation wavelength at 545 nm and an emission wavelength at 590 nm. Cell viability values were expressed as percentages derived from the fluorescence intensity from the treated cells relative to untreated cells. 2.9. Flow cytometry analysis of cellular uptake of polyelectrolyte complexes For flow cytometry, pDNA was fluorescently labeled with Cy5 dyes by using a Cy5 Label IT® Tracker intracellular nucleic acid localization kit (Mirus Bio, Madison, WI) according to the manufacturer's instruction. The resulting Cy5-labeled pDNA was used to form polyelectrolyte complexes according to the aforementioned procedure. HCT-116 cells were seeded on 24-well plates at a density of 1 × 105 cells per well and then cultivated for 24 h at 37 °C. These cells were transfected with the polyelectrolyte complexes formulated with 1.5 μg of Cy5labeled pDNA as described above. At designated time points, the transfected cells were harvested, resuspended in PBS, and then analyzed by using a BD LSR II flow cytometer system (BD Biosciences, CA). The extent of cellular uptake was evaluated by measuring the mean fluorescence intensity from the cell population. 2.10. Intracellular trafficking of polyelectrolyte complexes Cy5-labeled pDNA was used to produce fluorescently labeled polyelectrolyte complexes to visualize their intracellular trafficking. HCT-116 cells were seeded on 8-well Lab-Tek® II chamber slides at a density of 2 × 104 cells per well and then cultivated for 24 h at 37 °C. These cells were transfected with the polyelectrolyte complexes formulated with 0.5 μg of Cy5-labeled pDNA as described above. At designated time points, the transfected cells were rinsed with PBS three times and then treated with 1 μM of LysoSensor® Green DND-153 for endolysosomal staining. The cells were then rinsed with PBS again and fixed with 4% (w/v) paraformaldehyde. After washing with PBS, the cell nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI) (1.5 μg mL−1). The cells were observed under a LSM 510 META confocal laser scanning microscope (Carl Zeiss, Germany). The clusters exhibiting Cy5 fluorescence signals were considered as Cy5-labeled pDNA condensed within the polyelectrolyte complexes. The fluorescent intensity of each cluster present in the organelles was measured by using the ImageJ software (National Institutes of Health). The nuclear localization of pDNA was determined by the following equation [35]: Nuclear localization of pDNA ð%Þ ¼

F nucleus  100 F cytosol þ F end=lys þ F nucleus

where Fcytosol, Fend/lys and Fnucleus represent the fluorescent intensities of Cy5-labeled pDNA clusters present in the cytosol, endosome/lysosome and nucleus, respectively. More than 30 cells in the confocal microscopic images were examined to calculate the average value for each group. 2.11. Intratumoral distribution of Cy5-labeled pDNA delivered by polyelectrolyte complexes Male 6-week old CrTac:NCr-Foxn1nu mice were inoculated subcutaneously with 1 × 107 HCT-116 cells each. After 2 weeks when the tumor reached approximately 200 mm3, various complexes containing Cy5labeled pDNA (pDNA/PEI complexes at N/P 30, pDNA/PEI/HA-EGCG or pDNA/PEI/HA ternary complexes at C/P 0.5 and pDNA/Lipofectamine 2000 complexes) were injected intratumorally at a dose of 12 μg DNA per mouse. Tumor tissues were harvested 6 h post-injection and fixed in 10% formalin buffer solution for sectioning. The tumor sections were stained with DAPI and visualized using a LSM 510 DUO confocal laser scanning microscope (Carl Zeiss, Germany). The mean fluorescent intensities of Cy5-labeled pDNA in the images were determined by

208

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

using ImageJ software. The care and use of laboratory animals were monitored according to the approved protocols of the Institutional Animal Care and Use Committee (IACUC) at the Biological Resource Center (BRC) in Biopolis, Singapore. 2.12. Statistical analysis All experiments were performed over triplicate. Data were reported as mean ± standard deviation. Statistical analysis was performed using the analysis of variance (ANOVA) and two-tailed, unpaired Student's t-tests. P b 0.05 was considered to be statistically significant. 3. Results 3.1. Design of self-assembled ternary complexes stabilized with HA-EGCG conjugates Fig. 1A shows the schematic illustration of the ternary complexes self-assembled from pDNA, PEI and HA-EGCG conjugates for targeted gene delivery. HA-EGCG conjugates were synthesized by conjugating ethylamine-bridged EGCG dimers to HA using carbodiimide-mediated coupling reaction [30]. The degree of substitution for HA-EGCG conjugates was estimated to be 2.5 by UV–visible spectroscopy. To prepare the ternary complexes, positvely-charged pDNA/PEI complexes were first formed through electrostatic interactions between anionic pDNA and cationic PEI molecules. Subsequent addition of HA-EGCG conjugates

induced the self-assembly of pDNA/PEI/HA-EGCG ternary complexes via electrostatic and other non-covalent interactions between HA-EGCG conjugates and the preformed binary complexes. Since EGCG is known to interact with DNA molecules via hydrogen bonding, π–π stacking and hydrophobic interactions [26–28], it was anticipated that HAEGCG conjugates having multiple EGCG moieties would bind to pDNA/ PEI complexes more strongly than HA alone through multivalent DNA binding. We hypothesized that the stabilized ternary complexes would exhibit enhanced protection of pDNA against degradation and polyanion-induced dissociation during their journey from the extracellular environment to the cell nucleus (Fig. 1B). Moreover, the surfaceexposed EGCG moieties would confer nuclease resistance on the ternary complexes because of its ability to inhibit nuclease activity [29]. The pDNA/PEI/HA-EGCG ternary complexes were expected to readily internalize into CD44-overexpressing cells via HA receptor mediated endocytosis and enable efficient transfection of the cells. Once internalized, these ternary complexes would effectively withstand nuclease attack, escape from the endosomal compartment via the proton sponge effect, and then mediate the transport of pDNA into the nucleus. 3.2. Formation and characterization of pDNA/PEI/HA-EGCG ternary complexes To determine the optimal condition required to produce binary pDNA/PEI complexes, we investigated the effect of N/P ratios on the extent of transgene expression and cell viability. The formation of pDNA/

Fig. 1. Formation of self-assembled ternary complexes constructed with plasmid DNA (pDNA), branched polyethylenimine (PEI) and hyaluronic acid-epigallocatechin gallate (HA-EGCG) conjugate. (A) Synthetic route of pDNA/PEI/HA-EGCG ternary complex formation for targeted gene delivery. (B) Schematic diagram illustrating the receptor-mediated cellular uptake, endosomal escape of pDNA/PEI/HA-EGCG ternary complex and nuclear localization of pDNA.

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

PEI complexes was first examined (Fig. S1). As N/P ratios increased, the size of the complexes gradually decreased with a concomitant increase in the zeta potential values. This result indicated the formation of more compact and positively charged pDNA/PEI complexes at escalating N/P ratios. Fig. 2A shows the transfection efficiency and cytotoxicity of the binary complexes in human embryonic kidney HEK293 cells. As N/P ratio increased from 10 to 30, the percentage of GFP-positive cells was proportionally increased with a concomitant decrease in the cell viability. The pDNA/PEI complexes formed at N/P ratio of 30 achieved the maximum transfection efficiency of 73.8 ± 5.2% while maintaining 80.2 ± 4.4% of cell viability. A further increase in N/P ratio up to 40 significantly decreased the transfection efficiency and cell viability. The severe cytotoxicity caused at N/P ratio of 40 was likely attributed to the formation of highly cationic pDNA/PEI complexes, which might bind strongly to the cell surface and disrupt the cell membrane [7,8]. A similar trend was observed for human colorectal carcinoma HCT-116 cells (Fig. 2B). Raising N/P ratios gradually increased the extent of GFP expression while reducing the cell viability. The maximum transfection efficiency (19.0 ± 0.8%) attained at N/P ratio of 30 was much lower than those obtained from HEK293 cells. This was consistent with a previous report, in which HCT-116 cells were transfected less efficiently by pDNA/PEI complexes than HEK293 cells [36]. Similarly, low transfection efficiency (19.8 ± 1.8%) was observed when transfection was conducted on MDA-MB-231 cells (Fig. S2A). This finding agreed well with an earlier study showing the poor transfection efficiency of PEI-based gene vectors on these cells [37]. Despite being poorly transfected by PEI-based vectors, octaarginine-modified liposome-based gene vectors have recently been reported to efficiently transfect HCT-116 cells [34]. Thus we used the commercially available Lipofectamine 2000 as a positive control for gene transfection and noted significantly higher transfection efficiency of 86.4 ± 0.8% (Fig. S2B). These results confirmed the poor transfection ability of pDNA/PEI complexes on HCT-116 cells. Based on the transfection and cell viability data, pDNA/PEI complexes formed at a fixed N/P ratio of 30 were chosen to construct pDNA/PEI/ HA-EGCG ternary complexes. The formation of pDNA/PEI/HA-EGCG ternary complexes was accomplished simply by mixing the binary pDNA/PEI complexes with HA-EGCG conjugates. To examine the effect of the amount of HA-EGCG conjugates on the ternary complex formation, we measured the size and surface charge of the ternary complexes formed at various C/P ratios (i.e., the ratios of the number of carboxylic groups in HA to the number of phosphate groups in pDNA). For comparison, pDNA/PEI/HA ternary complexes were prepared by mixing HA with the binary complexes at equivalent C/P ratios. As presented in Fig. 3A, pDNA/PEI complexes were produced with an average diameter of 167 ± 27 nm. We found that the sizes of pDNA/ PEI complexes were not significantly altered by coating with HA-EGCG conjugates. While pDNA/PEI complexes displayed a positive ζ potential of 31.3 ± 1.2 mV, the ζ potential value gradually decreased upon addition

209

of HA-EGCG conjugates at escalating C/P ratios (Fig. 3B). This phenomenon was likely triggered by the neutralization of cationic pDNA/PEI complexes by anionic HA-EGCG conjugates. Hence these results demonstrated that HA-EGCG conjugates were stably coated on the surface of pDNA/PEI complexes without disrupting the complex structure. At C/P ratio of 10, the ζ potential values of pDNA/PEI/HA-EGCG ternary complexes (−7.5 ± 1.9 mV) were significantly lower than those of pDNA/PEI/HA ternary complexes (1.2 ± 1.4 mV). The observed difference in the ζ potential values suggested that HA-EGCG conjugates exhibited a more effective anchorage on the surface of pDNA/PEI complexes than HA alone due to DNA-binding affinity of EGCG. 3.3. Stabilization of pDNA by pDNA/PEI/HA-EGCG ternary complexes We performed agarose gel electrophoresis assay to verify the physical stability of pDNA/PEI/HA-EGCG ternary complexes. Previously, it has been reported that addition of excess amounts of polyanions such as HA and heparin can cause the dissociation of polyelectrolyte complexes by competitive binding [12,38]. As shown in Fig. 4A, the electrophoretic migration of pDNA was completely impeded by PEI at N/P ratio of 30, indicative of the complete condensation of pDNA and PEI via electrostatic interactions. Importantly, no liberation of pDNA was observed from both pDNA/PEI/HA-EGCG and pDNA/PEI/HA ternary complexes at C/P ratios ranging from 0.1 to 10, suggesting that pDNA/PEI complexes were not destabilized by coating with HA or HA-EGCG conjugates. We next investigated whether the ternary complexes could protect pDNA from polyanion-induced dissociation. It is known that polyelectrolyte complexes are subjected to destabilization and/or dissociation in the bloodstream through competitive binding of polyanions (e.g., heparin and plasma proteins) [11,12]. As shown in Fig. 4B, pDNA/PEI complexes were disintegrated and the incorporated pDNA was liberated after treatment with heparin. We found that pDNA/PEI/HA ternary complexes were dissociated almost completely regardless of C/P ratios, indicating that HA coating could not prevent disintegration of the complexes. In contrast, no detectable pDNA release was observed for pDNA/PEI/HA-EGCG ternary complexes formed at C/P ratio of 10. Only a partial release of the incorporated pDNA was seen at C/P ratio of 5 whereas complete pDNA release was observed at C/P ratios of 0.1, 0.5 and 2. The above results demonstrated that HA-EGCG conjugates were able to stabilize pDNA/PEI complexes against polyanion-induced dissociation more effectively than HA alone. 3.4. CD44-targeted gene delivery by pDNA/PEI/HA-EGCG ternary complexes We examined gene transfection ability of pDNA/PEI/HA-EGCG ternary complexes in serum-supplemented medium to prove their

Fig. 2. Optimization of binary pDNA/PEI complexes for transfection. Viability and the extent of GFP expression of (A) HEK293 cells or (B) HCT-116 cells transfected with pDNA/PEI complexes formed at various N/P ratios.

210

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

Fig. 3. Characterization of ternary complexes. (A) Hydrodynamic diameters and (B) ζ potential values of pDNA/PEI/HA-EGCG and pDNA/PEI/HA ternary complexes formed at various C/P ratios. The pDNA/PEI complexes formed at an optimal N/P ratio of 30 were chosen to construct the ternary complexes.

potential applicability for CD44-targeted gene delivery. The in vitro transfection efficiency of the ternary complexes was assessed using two different cell lines. As presented in Fig. 5A, abundant CD44 expression was observed on HCT-116 cells while HEK293 cells expressed low levels of CD44. Fig. 5B depicts the fluorescent microscopic images of HCT-116 cells transfected with the ternary complexes formulated with GFP-encoding pDNA. We found that gene transfection efficiency of pDNA/PEI/HA-EGCG ternary complexes increased gradually with raising C/P ratios and then reached the maximum level at C/P ratio of 0.5. Lower levels of GFP expression were observed at C/P ratios above 0.5. On the other hand, pDNA/PEI/HA ternary complexes achieved the maximum gene transfection at C/P ratio of 5. Flow cytometric analysis revealed that pDNA/PEI/HA-EGCG ternary complexes enabled more efficient transfection of HCT-116 cells than pDNA/PEI/HA ternary complexes at C/P ratios ranging from 0.1 to 0.5 (Fig. 5C). For instance, pDNA/PEI/HA-EGCG ternary complexes at C/P ratio 0.5 exhibited a remarkably high transfection efficiency of 43.7 ± 1.5%, whereas the corresponding pDNA/PEI/HA ternary complexes transfected the cells to a much smaller extent (18.6 ± 0.6%) at an equivalent C/P ratio. Considering that both ternary complexes were similar in size and surface charge, the enhanced physical stability of pDNA/PEI/HA-EGCG ternary complexes was probably responsible for their superior transfection ability

in serum-supplemented conditions. Furthermore, despite the use of 10-fold lower amount of the coating polymer, the transfection efficiency of pDNA/PEI/HA-EGCG ternary complexes at C/P ratio of 0.5 was even higher than the maximum transfection efficiency achieved by pDNA/ PEI/HA complexes at C/P ratio of 5 (37.2 ± 0.4%, P b 0.01). Furthermore, the transfection efficiency of pDNA/PEI/HA-EGCG ternary complexes was almost 2.3-fold higher than those of pDNA/PEI complexes (19.0 ± 0.8%), which are commonly used as a benchmark for gene delivery. The efficient gene transfection was likely ascribed to enhanced cellular uptake of pDNA/PEI/HA-EGCG ternary complexes via CD44mediated endocytosis [19,20]. A similar trend was observed when we measured the mean fluorescence intensity of the transfected cells (Fig. 5D), indicating that HA-EGCG coating greatly improved the transfection efficiency of pDNA/PEI complexes as well as the transgene expression level per cell. We selected pDNA/PEI/HA-EGCG ternary complexes formed at C/P ratio of 0.5 for further studies because these complexes exhibited the highest gene transfection efficiency. To investigate whether the enhanced transfection ability of pDNA/ PEI/HA-EGCG ternary complexes was attributed to CD44-mediated endocytosis, we pre-incubated HCT-116 cells with either free HA or CD44 antibody to block CD44 prior to transfection with the ternary complexes. As shown in Fig. 6A, the transfection efficiency of pDNA/

Fig. 4. Stabilization of pDNA by pDNA/PEI/HA-EGCG ternary complexes. (A) Agarose gel electrophoresis of naked pDNA, pDNA/PEI complexes (N/P = 30), pDNA/PEI/HA-EGCG and pDNA/ PEI/HA ternary complexes formed at various C/P ratios. (B) Stability of polyelectrolyte complexes against polyanion-induced dissociation. The samples were treated with heparin (1 mg mL−1) for 30 min prior to gel electrophoresis.

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

211

Fig. 5. CD44-targeted gene delivery. (A) Flow cytometric analysis of CD44 expression levels in HEK293 cells and HCT-116 cells. The brown and gray histogram show the population of the cells labeled with anti-human CD44 antibody and isotype control antibody, respectively. (B) Fluorescent microscopic images of HCT-116 cells transfected with various polyelectrolyte complexes formulated with GFP-encoding pDNA; pDNA/PEI complexes (N/P = 30), pDNA/PEI/HA-EGCG and pDNA/PEI/HA ternary complexes formed at various C/P ratios. Lipofectamine was used as a positive control. Scale bars represent 500 μm. (C) The extent of GFP expression and (D) mean fluorescence intensity of HCT-116 cells transfected with the polyelectrolyte complexes. Statistically significant difference between two groups: ⁎⁎P b 0.01 and ⁎⁎⁎P b 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

PEI/HA-EGCG ternary complexes decreased gradually with increasing the concentration of free HA. After pre-incubation with 0.5 mg mL−1 of HA, the percentage of GFP-positive cells was reduced to 17.2 ± 3.1%. In contrast, HA pre-incubation had little influence on the transfection efficiency of pDNA/PEI complexes. Furthermore, in vitro inhibition of CD44 uptake by pre-treatment using CD44 antibody also led to a decline in the population of GFP-positive cells (Fig. 6B). These results revealed that HA-EGCG coating facilitated the transport of pDNA/PEI complexes into CD44-overexpressing cells via HA receptor mediated endocytosis, and thus increased their transfection efficiency. Unlike what was observed in HCT-116 cells, pDNA/PEI/HA-EGCG ternary complexes exhibited only a marginal improvement in transfection efficiency in HEK293 cells compared to pDNA/PEI/HA ternary complexes at C/P ratios ranging from 0.1 to 0.5 (Fig. 6C). Since the expression of CD44 in

HEK293 cells was much lower as compared to HCT-116 cells, it was conceivable that the internalization of pDNA/PEI/HA-EGCG ternary complexes by HEK293 cells was not as efficient as HCT-116 cells. Both pDNA/PEI/HA and pDNA/PEI/HA-EGCG ternary complexes transfected the cells to a lesser extent than pDNA/PEI complexes at all the C/P ratios tested. This suggested that surface modification of pDNA/PEI complexes with HA or HA-EGCG conjugate diminished their transfection efficacy in HEK293 cells. It is known that the positively charged surface of pDNA/ PEI complexes promotes their electrostatic interaction with the cell membrane and thus increases their cellular uptake efficiencies [8,39]. Accordingly, the decreased transfection efficiency of the ternary complexes was likely caused by the charge neutralization resulting from the surface coating of the cationic pDNA/PEI complexes with anionic HA or HA-EGCG conjugate.

212

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

Fig. 6. Evaluation of target specificity of pDNA/PEI/HA-EGCG ternary complexes. (A) Effect of HA pre-treatment on the transfection efficiency of pDNA/PEI/HA-EGCG ternary complexes (C/ P = 0.5). HCT-116 cells were incubated with media containing HA at various concentrations prior to the transfection. (B) Effect of CD44 antibody pre-treatment on the transfection efficiency of pDNA/PEI/HA-EGCG ternary complexes (C/P = 0.5). HCT-116 cells were incubated with media containing CD44 antibody (15 μg mL−1) prior to the transfection. ⁎⁎⁎P b 0.001. (C) The extent of GFP expression of HEK293 cells transfected with pDNA/PEI complexes (N/P = 30), pDNA/PEI/HA-EGCG and pDNA/PEI/HA ternary complexes formed at various C/P ratios.

We also examined the viability of HCT-116 cells and HEK293 cells treated with polyelectrolyte complexes. As shown in Fig. 7A, binary pDNA/PEI complexes formed at N/P ratio of 30 induced only a marginal reduction in the viability of HCT-116 cells to 84.8 ± 2.9%. The viability of HCT-116 cells treated with pDNA/PEI/HA ternary complexes was comparable to or slightly higher than those treated with the binary complexes. The same trend was observed from HEK293 cells (Fig. 7B). This finding indicated that the incorporation of EGCG moieties did not negatively impact the cytotoxicity of the ternary complexes. Since the cytotoxicity of cationic polyelectrolyte complexes is mainly caused by their surface charge density in general [40], there was no significant difference in cytotoxicity between pDNA/PEI/HA and pDNA/PEI/HA-EGCG ternary complexes with almost identical sizes and surface charges, as seen in Fig. 3. Collectively, the above results demonstrated that HAEGCG-stabilized ternary complexes enabled highly efficient gene transfection of CD44-overexpressing cells with minimal cytotoxicity. 3.5. Enhanced cellular uptake of pDNA/PEI/HA-EGCG ternary complexes We performed flow cytometric analysis to examine the intracellular uptake behavior of ternary complexes quantitatively. After HCT-116 cells were transfected with the ternary complexes formulated with Cy5-labeled pDNA, the extent of cellular uptake was determined by measuring the mean fluorescence intensity from the cell population. Fig. 8A shows the mean fluorescence intensity of the cells transfected with the ternary complexes formed at an optimal C/P ratio of 0.5 or

binary pDNA/PEI complexes. All the complexes exhibited increasing cellular uptake over time, which was consistent with other reports describing the cellular entry of polyplexes [41,42]. At 1 h and 4 h posttransfection, the cells treated with pDNA/PEI/HA-EGCG ternary complexes displayed markedly higher fluorescence intensity than those treated with pDNA/PEI complexes. This result confirmed that HAEGCG coating promoted the transport of pDNA/PEI complexes into CD44-overexpressing HCT-116 cells via CD44-mediated endocytosis [19,20]. Notably, pDNA/PEI/HA-EGCG ternary complexes entered the cells more efficiently than pDNA/PEI/HA ternary complexes although both complexes had similar sizes and surface charges. Next we investigated the influence of C/P ratios on the intracellular uptake of ternary complexes. As shown in Fig. 8B, the extent of cellular uptake of pDNA/PEI/HA-EGCG ternary complexes increased proportionally to C/P ratios. The highest uptake efficiency was attained at C/P ratio of 10. This suggested that the interactions between HA-EGCG conjugates and pDNA/PEI complex would become stronger with raising C/P ratios, leading to more effective stabilization of the ternary complexes. It was noteworthy that pDNA/PEI/ HA-EGCG ternary complexes exhibited more efficient cellular uptake than pDNA/PEI/HA ternary complexes. At C/P ratios ranging from 0.1 to 2, pDNA/PEI/HA ternary complexes also did not show any significant improvement in uptake efficiency compared to pDNA/PEI complexes. The low stability of pDNA/PEI/HA ternary complexes likely contributed to their poor intracellular uptake in serumsupplemented conditions. The above results demonstrated that HA-

Fig. 7. Cytotoxicity of polyelectrolyte complexes. Viability of (A) HCT-116 cells and (B) HEK293 cells transfected with pDNA/PEI complexes (N/P = 30), pDNA/PEI/HA-EGCG and pDNA/PEI/ HA ternary complexes formed at various C/P ratios.

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

213

Fig. 8. Quantitative assessment of intracellular uptake of polyelectrolyte complexes. (A) Mean fluorescence intensity of HCT-116 cells transfected for 1 h or 4 h with various polyelectrolyte complexes formulated with Cy5-labeled pDNA; pDNA/PEI complexes (N/P = 30), pDNA/PEI/HA-EGCG and pDNA/PEI/HA ternary complexes (C/P = 0.5). Significance levels were set at ⁎P b 0.05, ⁎⁎P b 0.01 and ⁎⁎⁎P b 0.001. (B) Effect of C/P ratios on the intracellular uptake of the ternary complexes. HCT-116 cells were transfected for 4 h with pDNA/PEI/HA-EGCG and pDNA/PEI/HA ternary complexes formed at various C/P ratios. All fluorescence intensities were normalized with respect to the mean fluorescence intensity of the cells transfected with pDNA/PEI complexes (N/P = 30). Statistically significant difference between two groups: ⁎P b 0.05 and ⁎⁎P b 0.01.

EGCG coating could be rationally controlled to maximize the physical stability, cellular uptake efficiency and transfection ability of the ternary complexes. 3.6. Evaluation of intracellular trafficking of polyelectrolyte complexes Endosomal escape and nuclear translocation of pDNA are known as prerequisites to transgene expression [40]. We studied the intracellular trafficking and nuclear entry of polyelectrolyte complexes by confocal laser scanning microscopy (Fig. 9A). Cy5-labeled pDNA, endolysosomal compartment and nucleus could be visualized in red, green and blue color, respectively. At 1 h post-transfection, binary pDNA/PEI complexes and pDNA/PEI/HA ternary complexes were distributed near the cell membrane. The observed cell binding was

probably ascribed to the electrostatic interactions between the positively charged complexes and the negatively charged cell membrane [8,39]. On the other hand, pDNA was visible in the cytoplasm and even in the nucleus after 1 h incubation with pDNA/PEI/HAEGCG ternary complexes. At 4 h post-transfection, most of pDNA/ PEI complexes and pDNA/PEI/HA ternary complexes still remained attached to the cell membrane. Although a small number of those complexes were internalized, they were mostly sequestered in the endolysosomal compartments, as evidenced by the appearance of yellow pigments (overlap of green and red pigments). In contrast, the majority of pDNA was accumulated in the nuclear and perinuclear region in the case of pDNA/PEI/HA-EGCG ternary complexes. These results revealed that pDNA/PEI/HA-EGCG ternary complexes facilitated more effective endosomal escape and nuclear

Fig. 9. Subcellular localization analysis. (A) Confocal microscopic images of HCT-116 cells transfected for 1 h or 4 h with various polyelectrolyte complexes formulated with Cy5-labeled pDNA; pDNA/PEI complexes (N/P = 30), pDNA/PEI/HA-EGCG and pDNA/PEI/HA ternary complexes (C/P = 0.5). Red, green and blue fluorescent regions show the intracellular distribution of Cy5-labeled pDNA, endolysosomal compartment and nucleus, respectively. Scale bars represent 20 μm. The close-up images (right column) show the nuclear localization of pDNA at 4 h post-transfection (Scale bars, 10 μm). (B) The extent of nuclear localization of pDNA assessed at 4 h post-transfection. Statistically significant difference between two groups: ⁎⁎P b 0.01 and ⁎⁎⁎P b 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

214

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

transport of pDNA than the binary complexes and pDNA/PEI/HA ternary complexes. We further conducted a quantitative evaluation on the extent of nuclear localization of pDNA by measuring the fluorescent signals from subcellular organelles [35]. As shown in Fig. 9B, pDNA/PEI/HA-EGCG ternary complexes enabled nuclear entry of pDNA to a far greater extent than pDNA/PEI complexes and pDNA/PEI/HA ternary complexes. We found that 41.9 ± 2.5% of the internalized pDNA was localized to the nucleus after 4 h treatment with pDNA/PEI/HA-EGCG ternary complexes. This result could probably be explained by the protective effect of EGCG moieties against cytoplasmic nucleases. It has been reported that cytoplasmic nucleases rapidly degrade pDNA before its translocation into the nucleus and thus decrease the efficiency of gene transfer [43,44]. It was anticipated that HA-EGCG coating would confer nuclease resistance on the ternary complexes because of the ability of EGCG to inhibit nuclease activity [29]. Indeed, pDNA/PEI/HA-EGCG ternary complexes exhibited enhanced protection of pDNA against DNase Imediated degradation compared to the binary complexes and pDNA/ PEI/HA ternary complexes (Fig. S3). 3.7. Intratumoral distribution of Cy5-labeled pDNA delivered by polyelectrolyte complexes To examine the intracellular distribution of the delivered pDNA in tumors, Cy5-labeled pDNA was encapsulated in pDNA/PEI/HA-EGCG ternary complexes and intratumorally injected into HCT-116 tumor bearing mice. The tumors were harvested 6 h post-injection and the tumor sections were then observed by confocal laser scanning microscopy (Fig. 10A). No pDNA was visible in the tumors treated with pDNA/PEI complexes or pDNA alone. Only a marginal level of

fluorescence was detected from the tumors treated with pDNA/PEI/HA ternary complexes. In contrast, administration of pDNA/PEI/HA-EGCG ternary complexes led to significantly enhanced distribution of pDNA in the tumors. Similar observations were noted in the positive control Lipofectamine. The extent of intratumoral uptake of pDNA was further confirmed by the quantitative analysis of the mean fluorescence intensities (Fig. 10B). This suggested that pDNA/PEI/HA-EGCG ternary complexes promoted efficient intracellular delivery of pDNA in the tumor regions as compared to pDNA/PEI complexes and pDNA/PEI/HA ternary complexes. These results supported the observation that HA-EGCG coating augmented in vitro cellular uptake and further demonstrated that HA-EGCG coating elicited more efficient intracellular delivery of pDNA in HCT-116 tumors in vivo. 4. Discussion In this study, we prepared ternary polyelectrolyte complexes by sequential addition of HA-EGCG conjugates to pre-formed pDNA/PEI binary complexes. These self-assembled ternary complexes were stabilized by the electrostatic interaction and other non-covalent interactions between HA-EGCG and pDNA/PEI. We observed that the addition of HAEGCG did not alter the size of pDNA/PEI binary complexes (Fig. 3A). This finding agrees with the literature reporting a marginal effect of electrostatic polymer coating on the size of pDNA/PEI complexes [45]. Notably, the sizes of pDNA/PEI/HA-EGCG ternary complexes were comparable to those of pDNA/PEI/HA ternary complexes, indicating that the incorporation of EGCG moieties did not adversely affect the selfassembly process. While excessive polyanion addition has been shown to cause instability [12,38], we noted that pDNA/PEI complexes were not destabilized by coating with HA-EGCG conjugates (Fig. 4A). We

Fig. 10. Intratumoral distribution of pDNA delivered by polyelectrolyte complexes. (A) Confocal microscopic images of HCT-116 tumor sections 6 h after intratumoral administration with various polyelectrolyte complexes formulated with Cy5-labeled pDNA; pDNA/PEI complexes (N/P = 30), pDNA/PEI/HA-EGCG and pDNA/PEI/HA ternary complexes (C/P = 0.5). pDNA alone and Lipofectamine were used as negative and positive controls, respectively. Red and blue fluorescent regions show the intracellular distribution of Cy5-labeled pDNA and nucleus, respectively. Scale bars represent 10 μm. (B) The mean fluorescence intensities of Cy5-labeled pDNA in the tumor sections as determined by using ImageJ software. Statistically significant difference between two groups: ⁎⁎P b 0.01 and ⁎⁎⁎P b 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216

hypothesized that this was probably because of the low concentrations of HA-EGCG that were introduced. Even at the highest C/P ratio of 10, the ratio of carboxylic groups in HA to amine groups in PEI was 0.33. This was reflected in the zeta potential measurements (Fig. 3B), where the ternary complexes formed with HA exhibited positive surface charges at all C/P ratios. This finding was consistent with a previous study [12], in which the dissociation of pDNA/PEI/polyanion complexes was detected only when the ratio of anionic group to amine group exceeded a critical value, the lowest of which was 0.5 in the case of heparin. As opposed to having a disruptive effect on complexation, HAEGCG coating was found to stabilize pDNA/PEI complexes against polyanion-induced dissociation instead, in particular at C/P ratios of 5 and 10 (Fig. 4B). This could be attributed to the enhanced binding interactions between EGCG and pDNA, leading to the effective stabilization of the ternary complexes at high C/P ratios. The in vitro transfection efficiency of the ternary complexes was assessed using two cell lines expressing vastly different levels of CD44–HEK293 (CD44 −) and HCT-116 (CD44 +) (Fig. 5A). We observed that pDNA/PEI/HA-EGCG ternary complexes enabled greater efficiency as compared to pDNA/PEI binary complexes in transfecting HCT-116 cells but not HEK293 cells (Figs. 5, 6C). This enhancement is a result of CD44-targeting mediated by the HA-EGCG conjugates and is in agreement with other reports of HA based complexes [19,20]. The involvement of CD44 pathway was further confirmed by the reduction in transgene expression due to free HA and CD44 inhibitor pretreatment (Fig. 6A and B). As expected, CD44-receptor mediated endocytosis led to a greater cellular uptake of the pDNA/PEI/HA-EGCG complexes than pDNA/PEI complexes (Fig. 8). But unlike the gradual increase in cellular uptake in accordance with increasing C/P ratios, the transfection efficiency of pDNA/PEI/HA-EGCG ternary complexes reached the maximum level at C/P ratio of 0.5 and then declined drastically when C/P ratio was raised to 10 (Fig. 5C and D). The observed discrepancy implied that another step apart from cellular uptake plays a crucial role in mediating the gene transfer. Besides cellular entry, unpacking of polyelectrolyte complexes to release pDNA has also been recognized as a key step in the gene delivery process [40,46,47]. The gel electrophoresis analysis revealed that raising C/P ratios increased the stability of pDNA/PEI/HA-EGCG ternary complexes against polyanion-induced dissociation (Fig. 4B). Hence it was evident that the ternary complexes formed at higher C/P ratios were more tightly condensed than those formed at lower C/P ratios. Inefficient liberation of pDNA from the tight complexes might be responsible for the decrease in transfection efficiency at C/P ratios above 0.5. Although HA coating also enhanced transfection efficiency possibly via CD44-targeting, HA-EGCG ternary complexes achieved significantly higher transfection efficiency than HA ternary complexes (Fig. 5C and D). This was attributed to the higher cellular uptake of pDNA/PEI/HAEGCG ternary complexes as compared to pDNA/PEI/HA ternary complexes (Fig. 8). Since the transfection experiment was conducted in the presence of serum, the enhanced physical stability of HA-EGCG ternary complexes was likely responsible for their higher cellular uptake. The improved stability was further demonstrated by the superior intracellular pDNA delivery mediated by HA-EGCG ternary complexes to HCT-116 tumors in vivo (Fig. 10). It was reasonable that HA-EGCG coating would effectively stabilize pDNA/PEI complexes against polyanioninduced dissociation in the extracellular environment, and thus increased their chance of cellular internalization. In addition, internalized pDNA/PEI/HA-EGCG ternary complexes could better withstand nuclease attack in the intracellular environment as a result of EGCGmediated nuclease inhibition (Fig. S3). This allowed for the successful escape of the complexes from the endosomal compartment via the proton sponge effect, and the subsequent delivery of pDNA to the nucleus. Taken together, this study demonstrated that HA-EGCG-stabilized ternary complexes conferred highly efficient transfection of CD44overexpressing cells under serum-supplemented conditions due to their enhanced stability, nuclease resistance and CD44-targeting ability.

215

5. Conclusion The current study described the design and application of nanoscale ternary complexes stabilized with HA-EGCG conjugates for CD44targeted gene delivery. In addition to stabilizing pDNA/PEI complexes via strong DNA-binding affinity of EGCG, HA-EGCG conjugate also facilitated their transport into CD44-overexpressing cells via receptormediated endocytosis. The stabilized ternary complexes displayed enhanced resistance to nuclease attack and polyanion-induced dissociation. At an optimal C/P ratio, these complexes efficiently transfected the difficult-to-transfect HCT-116 colon cancer cell line even in the presence of serum due to their enhanced stability, nuclease resistance and CD44-targeting ability. Subcellular trafficking analysis revealed that pDNA/PEI/HA-EGCG ternary complexes facilitated more efficient nuclear entry of pDNA than binary complexes and pDNA/PEI/HA ternary complexes. To the best of our knowledge, this is the first report on the development of green tea catechin-based gene delivery system. Our study also demonstrated that HA-EGCG-stabilized ternary complexes could be potentially applied for CD44-targeted delivery of nucleotide therapeutics to diseased cells. Acknowledgments This work was supported by the Institute of Bioengineering and Nanotechnology (Biomedical Research Council, Agency for Science, Technology and Research, Singapore (A*STAR)) and Science and Engineering Council (SERC) Personal Care Programme Grant, A*STAR. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jconrel.2016.02.004. References [1] A. Mountain, Gene therapy: the first decade, Trends Biotechnol. 18 (2000) 119–128. [2] K. Itaka, K. Yamauchi, A. Harada, K. Nakamura, H. Kawaguchi, K. Kataoka, Polyion complex micelles from plasmid DNA and poly(ethylene glycol)–poly(L-lysine) block copolymer as serum-tolerable polyplex system: physicochemical properties of micelles relevant to gene transfection efficiency, Biomaterials 24 (2003) 4495–4506. [3] D.M. Lynn, R. Langer, Degradable poly(β-amino esters): synthesis, characterization, and self-assembly with plasmid DNA, J. Am. Chem. Soc. 122 (2000) 10761–10768. [4] H.J. Kim, M.S. Kwon, J.S. Choi, S.M. Yang, J.K. Yoon, K. Kim, J.S. Park, Highly effective and slow-biodegradable network-type cationic gene delivery polymer: small library-like approach synthesis and characterization, Biomaterials 27 (2006) 2292–2301. [5] W.T. Godbey, K.K. Wu, A.G. Mikos, Poly(ethylenimine) and its role in gene delivery, J. Control. Release 60 (1999) 149–160. [6] O. Boussif, F. Lezoualc'h, M.A. Zanta, M.D. Mergny, D. Scherman, B. Demeneix, J.-P. Behr, A versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo: polyethylenimine, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 7297–7301. [7] S.M. Moghimi, P. Symonds, J.C. Murray, A.C. Hunter, G. Debska, A. Szewczyk, A twostage poly(ethylenimine)-mediated cytotoxicity: implications for gene transfer/ therapy, Mol. Ther. 11 (2005) 990–995. [8] D. Fischer, Y. Li, B. Ahlemeyer, J. Krieglstein, T. Kissel, In vitro cytotoxicity testing of polycations: influence of polymer structure on cell viability and hemolysis, Biomaterials 24 (2003) 1121–1131. [9] M. Ogris, P. Steinlein, M. Kursa, K. Mechtler, R. Kircheis, E. Wagner, The size of DNA/ transferrin-PEI complexes is an important factor for gene expression in cultured cells, Gene Ther. 5 (1998). [10] V.K. Sharma, M. Thomas, A.M. Klibanov, Mechanistic studies on aggregation of polyethylenimine-DNA complexes and its prevention, Biotechnol. Bioeng. 90 (2005) 614–620. [11] X.H. Luo, F.W. Huang, S.Y. Qin, H.F. Wang, J. Feng, X.Z. Zhang, R.X. Zhuo, A strategy to improve serum-tolerant transfection activity of polycation vectors by surface hydroxylation, Biomaterials 32 (2011) 9925–9939. [12] C. Wang, X. Luo, Y. Zhao, L. Han, X. Zeng, M. Feng, S. Pan, H. Peng, C. Wu, Influence of the polyanion on the physico-chemical properties and biological activities of polyanion/DNA/polycation ternary polyplexes, Acta Biomater. 8 (2012) 3014–3026. [13] V.S. Trubetskoy, S.C. Wong, V. Subbotin, V.G. Budker, A. Loomis, J.E. Hagstrom, J.A. Wolff, Recharging cationic DNA complexes with highly charged polyanions for in vitro and in vivo gene delivery, Gene Ther. 10 (2003) 261–271. [14] T. Kurosaki, T. Kitahara, S. Fumoto, K. Nishida, J. Nakamura, T. Niidome, Y. Kodama, H. Nakagawa, H. To, H. Sasaki, Ternary complexes of pDNA, polyethylenimine, and

216

[15]

[16] [17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25] [26]

[27]

[28] [29]

[30]

K. Liang et al. / Journal of Controlled Release 226 (2016) 205–216 gamma-polyglutamic acid for gene delivery systems, Biomaterials 30 (2009) 2846–2853. M. Sakae, T. Ito, C. Yoshihara, N. Iida-Tanaka, H. Yanagie, M. Eriguchi, Y. Koyama, Highly efficient in vivo gene transfection by plasmid/PEI complexes coated by anionic PEG derivatives bearing carboxyl groups and RGD peptide, Biomed. Pharmacother. 62 (2008) 448–453. D.A. Ossipov, Nanostructured hyaluronic acid-based materials for active delivery to cancer, Expert Opin. Drug Deliv. 7 (2010) 681–703. E.J. Oh, K. Park, K.S. Kim, J. Kim, J.-A. Yang, J.-H. Kong, M.Y. Lee, A.S. Hoffman, S.K. Hahn, Target specific and long-acting delivery of protein, peptide, and nucleotide therapeutics using hyaluronic acid derivatives, J. Control. Release 141 (2010) 2–12. K.Y. Choi, H. Chung, K.H. Min, H.Y. Yoon, K. Kim, J.H. Park, I.C. Kwon, S.Y. Jeong, Selfassembled hyaluronic acid nanoparticles for active tumor targeting, Biomaterials 31 (2010) 106–114. X. Sun, P. Ma, X. Cao, L. Ning, Y. Tian, C. Ren, Positive hyaluronan/PEI/DNA complexes as a target-specific intracellular delivery to malignant breast cancer, Drug Deliv. 16 (2009) 357–362. Y. Wang, Z. Xu, R. Zhang, W. Li, L. Yang, Q. Hu, A facile approach to construct hyaluronic acid shielding polyplexes with improved stability and reduced cytotoxicity, Colloids Surf. B 84 (2011) 259–266. T. Ito, N. Iida-Tanaka, T. Niidome, T. Kawano, K. Kubo, K. Yoshikawa, T. Sato, Z. Yang, Y. Koyama, Hyaluronic acid and its derivative as a multi-functional gene expression enhancer: protection from non-specific interactions, adhesion to targeted cells, and transcriptional activation, J. Control. Release 112 (2006) 382–388. Y. He, G. Cheng, L. Xie, Y. Nie, B. He, Z. Gu, Polyethyleneimine/DNA polyplexes with reduction-sensitive hyaluronic acid derivatives shielding for targeted gene delivery, Biomaterials 34 (2013) 1235–1245. T. Yokozawa, E.J. Cho, Y. Hara, K. Kitani, Antioxidative activity of green tea treated with radical initiator 2,2′-azobis(2-amidinopropane) dihydrochloride, J. Agric. Food Chem. 48 (2000) 5068–5073. M.E. Cavet, K.L. Harrington, T.R. Vollmer, K.W. Ward, J.Z. Zhang, Anti-inflammatory and anti-oxidative effects of the green tea polyphenol epigallocatechin gallate in human corneal epithelial cells, Mol. Vis. 17 (2011) 533–542. C.S. Yang, X. Wang, G. Lu, S.C. Picinich, Cancer prevention by tea: animal studies, molecular mechanisms and human relevance, Nat. Rev. Cancer 9 (2009) 429–439. X. Zheng, A. Chen, T. Hoshi, J. Anzai, G. Li, Electrochemical studies of (−)-epigallocatechin gallate and its interaction with DNA, Anal. Bioanal. Chem. 386 (2006) 1913–1919. J.E. Chung, S. Tan, S.J. Gao, N. Yongvongsoontorn, S.H. Kim, J.H. Lee, H.S. Choi, H. Yano, L. Zhuo, M. Kurisawa, J.Y. Ying, Self-assembled micellar nanocomplexes comprising green tea catechin derivatives and protein drugs for cancer therapy, Nat. Nanotechnol. 9 (2014) 907–912. T. Kuzuhara, Y. Sei, K. Yamaguchi, M. Suganuma, H. Fujiki, DNA and RNA as new binding targets of green tea catechins, J. Biol. Chem. 281 (2006) 17446–17456. T. Kuzuhara, Y. Iwai, H. Takahashi, D. Hatakeyama, N. Echigo, Green tea catechins inhibit the endonuclease activity of influenza A virus RNA polymerase, PLoS Curr. 1 (2009) RRN1052. F. Lee, J. Lim, M.R. Reithofer, S.S. Lee, J.E. Chung, C.A.E. Hauser, M. Kurisawa, Synthesis and bioactivity of a conjugate composed of green tea catechins and hyaluronic acid, Polym. Chem. 6 (2015) 4462–4472.

[31] T.T. Nguyen, Y.H. Moon, Y.B. Ryu, Y.M. Kim, S.H. Nam, M.S. Kim, A. Kimura, D. Kim, The influence of flavonoid compounds on the in vitro inhibition study of a human fibroblast collagenase catalytic domain expressed in E. coli, Enzym. Microb. Technol. 52 (2013) 26–31. [32] F. Lee, J.E. Chung, K. Xu, M. Kurisawa, Injectable degradation-resistant hyaluronic acid hydrogels cross-linked via the oxidative coupling of green tea catechin, ACS Macro Lett. 4 (2015) 957–960. [33] Y. Hu, D. Zhou, C. Li, H. Zhou, J. Chen, Z. Zhang, T. Guo, Gene delivery of PEI incorporating with functional block copolymer via non-covalent assembly strategy, Acta Biomater. 9 (2013) 5003–5012. [34] Y. Yamada, M. Hashida, H. Harashima, Hyaluronic acid controls the uptake pathway and intracellular trafficking of an octaarginine-modified gene vector in CD44 positive- and CD44 negative-cells, Biomaterials 52 (2015) 189–198. [35] H.S. Hwang, H.C. Kang, Y.H. Bae, Bioreducible polymers as a determining factor for polyplex decomplexation rate and transfection, Biomacromolecules 14 (2013) 548–556. [36] M. Breunig, U. Lungwitz, R. Liebl, A. Goepferich, Breaking up the correlation between efficacy and toxicity for nonviral gene delivery, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 14454–14459. [37] W.-C. Tseng, C.-H. Tang, T.-Y. Fang, The role of dextran conjugation in transfection mediated by dextran-grafted polyethylenimine, J. Gene Med. 6 (2004) 895–905. [38] Y. Koyama, E. Yamada, T. Ito, Y. Mizutani, T. Yamaoka, Sugar-containing polyanions as a self-assembled coating of plasmid/polycation complexes for receptor-mediated gene delivery, Macromol. Biosci. 2 (2002) 251–256. [39] W. Godbey, K.K. Wu, A.G. Mikos, Tracking the intracellular path of poly (ethylenimine)/DNA complexes for gene delivery, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 5177–5181. [40] D.W. Pack, A.S. Hoffman, S. Pun, P.S. Stayton, Design and development of polymers for gene delivery, Nat. Rev. Drug Discov. 4 (2005) 581–593. [41] Y. Matsumoto, K. Itaka, T. Yamasoba, K. Kataoka, Intranuclear fluorescence resonance energy transfer analysis of plasmid DNA decondensation from nonviral gene carriers, J. Gene Med. 11 (2009) 615–623. [42] Y. Wu, Y.P. Ho, Y. Mao, X. Wang, B. Yu, K.W. Leong, L.J. Lee, Uptake and intracellular fate of multifunctional nanoparticles: a comparison between lipoplexes and polyplexes via quantum dot mediated Forster resonance energy transfer, Mol. Pharm. 8 (2011) 1662–1668. [43] D. Lechardeur, K.J. Sohn, M. Haardt, P.B. Joshi, M. Monck, R.W. Graham, B. Beatty, J. Squire, H. O'Brodovich, G.L. Lukacs, Metabolic instability of plasmid DNA in the cytosol: a potential barrier to gene transfer, Gene Ther. 6 (1999) 482–497. [44] D. Pinto-Gonzalez Howell, R.J. Krieser, A. Eastman, M.A. Barry, Deoxyribonuclease II is a lysosomal barrier to transfection, Mol. Ther. 8 (2003) 957–963. [45] T. Kurosaki, T. Kitahara, S. Kawakami, K. Nishida, J. Nakamura, M. Teshima, H. Nakagawa, Y. Kodama, H. To, H. Sasaki, The development of a gene vector electrostatically assembled with a polysaccharide capsule, Biomaterials 30 (2009) 4427–4434. [46] H.C. Kang, K.M. Huh, Y.H. Bae, Polymeric nucleic acid carriers: current issues and novel design approaches, J. Control. Release 164 (2012) 256–264. [47] I.A. Khalil, K. Kogure, H. Akita, H. Harashima, Uptake pathways and subsequent intracellular trafficking in nonviral gene delivery, Pharmacol. Rev. 58 (2006) 32–45.