In vitro non-viral gene delivery with nanofibrous scaffolds
http://www.100md.com
《核酸研究医学期刊》
1Department of Chemistry, Stony Brook University Stony Brook, NY 11794-3400, USA 2Department of Biomedical Engineering, Stony Brook University Stony Brook, NY 11794-2580 USA 3Department of Materials Science and Engineering, Stony Brook University Stony Brook, NY 11794-2275, USA
*To whom correspondence should be addressed. Benjamin Chu, Department of Chemistry, Stony Brook University, Stony Brook, NY 11794-3400, USA. Tel: +631 632 7928; Fax: +631 632 6518; Email: bchu@notes.cc.sunysb.edu
ABSTRACT
Extracellular and intracellular barriers typically prevent non-viral gene vectors from having an effective transfection efficiency. Formulation of a gene delivery vehicle that can overcome the barriers is a key step for successful tissue regeneration. We have developed a novel core-shelled DNA nanoparticle by invoking solvent-induced condensation of plasmid DNA (?-galactosidase or GFP) in a solvent mixture and subsequent encapsulation of the condensed DNA globule in a triblock copolymer, polylactide-poly(ethylene glycol)-polylactide (L8E78L8), in the same solvent environment. The polylactide shell protects the encapsulated DNA from degradation during electrospinning of a mixture of encapsulated DNA nanoparticles and biodegradable PLGA (a random copolymer of lactide and glycolide) to form a nanofibrous non-woven scaffold using the same solution mixture. The bioactive plasmid DNA can then be released in an intact form from the scaffold with a controlled release rate and transfect cells in vitro.
INTRODUCTION
The search for viable alternative approaches to deal with problems associated with viral DNA delivery (i.e. safety, toxicity and fate of transfected cells, etc.) has prompted many researchers to develop non-viral means of delivering genes in vivo. Typical non-viral approaches include naked DNA, cationic lipids formulated into liposomes and subsequent complex formation with DNA (lipoplexes), complex formation of cationic polymers with DNA (polyplexes), and collagen- or hyaluronan-based DNA gels (1). Recently, new biomaterials, such as the gene activated matrix (GAM) and biodegradable polymers (i.e. polylactide-co-glycolide, PLGA) have been developed that can serve as scaffolds for DNA delivery and tissue engineering (2,3) Even though non-viral gene delivery systems offer enhanced safety over viral systems, they are plagued with comparatively low transfection efficiencies. Non-viral gene delivery encounters many barriers, including degradation of the DNA in plasma, uptake of DNA by the reticuloendothelial system, lysosomal degradation of the DNA and lack of translocation to the nucleus (1,4). To overcome these barriers, the critical component of design formulation must be based on the fundamental understanding of the interactions between DNA and various chemical species in solution. Proper characterization of vectors thus becomes an essential step in the development of efficient non-viral gene delivery systems.
The process of DNA condensation has attracted a great deal of attention in recent years due to its biological importance in viral DNA packing, as well as in the development of gene delivery vehicles (5,6). The extended DNA chain could effectively be collapsed into a compact globule through charge neutralization by cationic polymers (7,8), polyamines and other multivalent cations, such as spermidine (9) and cobalt hexamine [ (10). In this study, we chose the approach of solvent-induced condensation of DNA. There are two advantages of using a poor (or non-) solvent for DNA to undergo coil-to-globule transition: (i) no contamination by other unwanted species, and (ii) easy removal of solvent through evaporation. The chosen poor solvent was 94% (in volume) N,N-dimethylformamide (DMF) + 6% 1x TE buffer. It was found that DNA could remain intact in DMF for at least several days and its bioactivity was maintained.
We have designed a unique scheme that can overcome some barriers in non-viral gene delivery. Our scheme is as follows. Encapsulated DNA nano-particles with a core-shell structure were first created by condensing plasmid DNA in a poor solvent mixture, followed by encapsulation of the condensed DNA in a triblock copolymer of poly(lactide)-b-poly(ethylene glycol)-b-poly(lactide) (LEL) that formed micelles under the same solvent environment. This mixture was designed as a suitable solvent for random copolymers of lactide and glycolide (PLGA). The mixtures of encapsulated DNA and PLGA were then electrospun to form a non-woven nanofibrous and nanocomposite scaffold, whereby the solvent mixture was rapidly removed by evaporation in the jet stream during fiber formation in the electrospinning process. Thus, the plasmid DNA was preserved in the nanofibrous scaffold. As a gene delivery vehicle, the polylactide shell protected the DNA core from degradation and maintained its bioactivity during electrospinning. The encapsulated DNA could be released from the scaffold by controlled degradation of the biodegradable scaffold component (PLGA). Transfection of adhered cells could then be enhanced by sustained DNA release and immobilization in contact with adhered cells.
MATERIALS AND METHODS
Preparation of DNA and triblock polymer
pCMV? plasmid DNA (7164 bp) encoding ?-galactosidase (Clontech, Palo Alto, CA) was extracted from cultured Escherichia coli utilizing a Qiagen (Valencia, CA) GigaPrep DNA isolation kit. PLA-PEG-PLA triblock copolymer was synthesized by copolymerizing PEG (3.4K) and D,L-lactide at 130°C for 15 h with stannous octoate being the catalyst. The molecular weight of purified reaction product was determined by 1H nuclear magnetic resonance (1H NMR) and gel permeation chromatography (GPC).
Laser light scattering (LLS)
A LLS spectrometer equipped with a BI-9000 AT digital correlator and a solid-state laser (DPSS, Coherent, 200 mW and 532 nm) was used to perform static and dynamic light scattering (DLS) studies over an angular range of 20–120°. In static LLS, the angular dependence of the excess absolute time-averaged scattered intensity, known as the Rayleigh ratio Rvv(), was measured. For a very dilute solution, the weight-averaged molecular weight (Mw), the root mean square radius of gyration (Rg), and the second virial coefficient (A2) can be obtained on the basis of
(1)
where H = 42n2(dn/dC)2/(NA4) and q = (4n/)sin(/2) with NA, n, dn/dC and being the Avogadro constant, the solvent refractive index, the specific refractive index increment and the wavelength of light in vacuum, respectively.
In dynamic LLS, the intensity–intensity time correlation function G(2)(t) in the self-beating mode was measured.
(2)
where A is the measured base line, ? is a coherence factor, t is the delay time and g(1)(t) is the normalized first-order electric field time correlation function. g(1)(t) is related to the line width distribution G() by
(3)
By using a Laplace inversion program, CONTIN, the normalized distribution function of the characteristic line width G() was obtained. The line width is a function of both C and q, which can be expressed as
(4)
with D, kd and f being the translational diffusive coefficient, the diffusion second virial coefficient, and a dimensionless constant, respectively. When the concentration is extremely dilute and Rgq << 1, /q2 is approximately equal to D. D can be further converted into the hydrodynamic radius Rh by using the Stokes–Einstein equation:
(5)
where kB, T and are the Boltzmann constant, the absolute temperature, and the viscosity of the solvent, respectively.
Measurement on dn/dC
Differential refractometer (Phoenix Precision Instruments, NY) was used to measure the dn/dC value of PLA and PEG in 94% DMF + 6% TE at 25°C. The instrument was calibrated in advance by using aqueous potassium chloride solutions. At a wavelength of 546 nm, the measured dn/dC of PLA and PEG in 94% DMF + 6% TE exhibited a value of 0.030 ml/g and 0.044 ml/g, respectively. The dn/dC values of plasmid DNA in 1x TE buffer and in 94% DMF + 6% TE were 0.17 ml/g and 0.10 ml/g, respectively.
Scaffold preparation and DNA release
DNA incorporation into electrospun scaffolds and DNA release assays from scaffold sections have been reported previously by our laboratory (11). The scaffold preparation procedures in this study were similar to those reported before. Briefly, PLA-PEG block copolymer was dissolved in N,N-dimethylformamide, to which pCMV? plasmid (5 mg/ml, in TE Buffer) and PLGA (LA/GA = 75/25) were added. The final solution contains 94% DMF + 6% TE. Solutions and scaffolds without block copolymer (only PLGA and DNA) were prepared identically except for the omission of block copolymer. Scaffolds were electrospun at 25 kV with a solution flow rate of 20 μl/min. The spinneret (anode) was fixed at 15 cm above the aluminum-covered rotating collection drum (cathode).
The electrospun scaffolds were cut into 1.5 x 1.0 cm sections and each section was incubated at 37°C with 1 ml TE buffer in Eppendorf tubes. The amount of DNA released into solution was quantified by using the Pico Green Assay, where solutions were excited at 485 nm, and the emission was measured at 530 nm in a microplate reader (CytoFluor Series 4000, Perseptive Biosystems). Integrity of released DNA was determined by 0.8% agarose gel electrophoresis and visualized by ethidium bromide staining.
Cell culture and transfection
Transfection studies were conducted using the MC3T3-E1 pre-osteoblastic cell line. MC3T3 cells were maintained in log growth phase using -MEM supplemented with 10% fetal bovine serum (FBS) (Life Technologies, Grand Island, NY). Cells were plated at an initial density of 1 x 105 cells/well into 6-well plates, 24 h before transfection. The negative control was designated as 2 μg naked DNA added directly to cell culture solution. A positive control was constructed using the commercially available Fugene 6 transfection reagent . Cells were incubated with two 1.5 x 1 cm sections of DNA loaded scaffolds, with and without the block copolymer. For the other set of transfection experiments with the GFP plasmid, 1 x 105 cells (20 μl) were placed directly in the center of 1.5 x 1 cm sections of PLGA/LEL and PLGA/LEL/GFP plasmid DNA scaffold (n = 3/scaffold type). Initially, the 20 μl droplet remained intact in the center of the scaffold, but it spread out to cover the entire scaffold with time (1 h). After 1 h, the scaffold was transferred into a 12-well tissue culture plate, containing cell culture medium. After 24 h, the scaffolds were removed from the tissue culture plates and fixed with 1% formaldehyde before staining with 0.1% Nuclear Fast Red (Sigma) prepared in 5% aluminum sulfate. The scaffolds were then washed twice with distilled water, and dehydrated with an ethanol gradient (70, 95 and 100%). Finally, the scaffolds were placed on microscope slides and cells (present only on the exterior surface of the scaffold) were imaged under epifluorescence for the presence of GFP using a Zeiss Axiovert microscope.
RESULTS AND DISCUSSION
LLS
Both static light scattering (SLS) and DLS, was employed to monitor the changes in DNA size and conformation. Figure 1A and B show Zimm plots of pCMV? plasmid DNA (7164 bp) encoding ?-galactosidase in 1x TE buffer and in 94% DMF + 6% TE, respectively. From the Zimm plot, the weight average molecular weight (Mw), the radius of gyration (Rg) and the second virial coefficient (A2) were obtained (Table 1). In 94% DMF + 6% TE, Mw of plasmid DNA was only slightly larger than that in 1x TE buffer. However, the Rg value was significantly reduced from 135 nm (in 1x TE buffer) to about 15 nm, suggesting the occurrence of chain collapse. A2 was used to describe the interactions between solute and solvent. In general, A2 is positive when a sample was dissolved in a good solvent. As expected, the value of A2 (–3.7 x 10–4 mol cm3 g–2) became negative in 94% DMF + 6% TE because this mixture was a poor solvent for DNA. The hydrodynamic radius, Rh, obtained from the CONTIN (12) analysis of dynamic light scattering data, was also dramatically decreased from 87 nm in 1x TE buffer to about 17 nm in 94% DMF + 6% TE (Figure 1C). The relationship between the conformation of polymer chains and the Rg/Rh ratio has been well established. Typically, a random coil has an Rg/Rh value of 1.5, and a solid sphere has a value of 0.775 (13). As shown in Table 1, plasmid DNA had an Rg/Rh ratio of 1.55 in 1x TE buffer, suggesting essentially a random-coil conformation; while in 94% DMF + 6% TE, the value was about 0.88, very close to 0.775, indicating a relatively compact solid-sphere conformation.
Figure 1 Condensation of plasmid DNA in 94% DMF + 6% TE studied by laser light scattering. (A and B) Shows the Zimm plots of plasmid DNA in 1x TE buffer, and in 94% DMF + 6% TE, respectively. (C) Shows the size distribution of DNA measured by DLS at 90° and 1.0 x 10–5 g/ml DNA concentration.
Table 1 Physicochemical properties of plasmid DNA measured by LLS
The triblock copolymer used in this study (LEL), was synthesized according to a known scheme (14). In brief, PEG and D,L-lactide with known ratio were mixed together and the polymerization was accomplished at 130°C for 15 h using stannous octoate as the catalyst. Before mixing plasmid DNA and LEL together, the behavior of LEL itself in 94% DMF + 6% TE was studied. With the known molecular weight of PEG block (3.4 k), the number-averaged molecular weight of PLA block could be estimated from proton NMR (data not shown), which was about 0.6 k. Therefore, the average repeating unit of each block was estimated to be L8E78L8. GPC also confirmed this as the weight average molecular weight of LEL was about 5.9 k.
Figure 2A shows the excess scattered intensity of LEL at different copolymer concentrations in the same solvent mixture. No sharp increase in the scattered intensity over a range of concentrations was observed, suggesting either the absence of micelle formation or a very low critical micelle concentration. At concentrations higher than 1 x 10–3 g/ml, LEL showed Mw of about 180 k (Figure 2B), a value at almost 30 times larger than that measured by GPC. The measured Rg value was 39 nm, significantly larger than the size of a polymer chain with an Mw of only 5.9 k. The hydrodynamic radius, Rh, measured from DLS (data not shown), had a value of 44 nm, very close to Rg. All these results confirmed that LEL chains were aggregated in 94% DMF + 6% TE at concentrations above 1 x 10–3 g/ml.
Figure 2 Encapsulation of plasmid DNA by triblock copolymer of LEL in 94% DMF + 6% TE. (A) Excess scattered intensity of LEL at different concentrations measured at 90°; (B) Zimm plot of LEL at concentrations of 1.25, 6.24 and 31.2 mg/ml; (C and D) CONTIN analysis of plasmid DNA (hollow circles), LEL (hollow squares) and their mixtures (solid diamonds) at different DNA/LEL weight ratio. The concentration of DNA used in panels C and D is 1 x 10–5 g/ml.
DNA (at 1 x 10–5 g/ml) was subsequently mixed with LEL at two different concentrations, 3.12 x 10–4 g/ml and 6.25 x 10–3 g/ml. At 3.12 x 10–4 g/ml, where LEL did not aggregate, the size and distribution of the DNA/LEL mixture were almost the same as those of plasmid DNA itself (Figure 2C). In other words, there was an insufficient amount of LEL to encapsulate the condensed plasmid DNA. At LEL concentration of 6.25 x 10–3 g/ml, a value where LEL already aggregated, only one ‘size’ with a relatively broad size distribution was observed for DNA/LEL in the solvent mixture (Figure 2D). The average size, however, became even smaller than that of the pure LEL aggregate, but larger than that of the plasmid DNA itself, indicating that a fusion of plasmid DNA and LEL aggregates had occurred.
The molecular scale interaction of DNA and LEL is depicted schematically in Figure 3. The figure also included experimental evidence for the condensation of plasmid DNA, the aggregation of LEL and the encapsulation of DNA by LEL in 94% DMF + 6% TE. Initially, after being air dried from 1x TE buffer, plasmid DNA was rod-like with a length >200 nm (Figure 3A), very close to the shape of a compact random coil with high super-helicity. Depending on the chain length and the degree of super-helicity, the DNA chain could be loose, compact or even branched. In 94% DMF + 6% TE, the DNA chain was condensed to form a much smaller globule with a diameter of 30 nm (Figure 3B). By design, LEL was also aggregated in 94% DMF + 6% TE above a certain concentration, which can be explained as follows. TE is a good solvent for PEG but not for PLA. In the solution mixture, water molecules prefer to stay together with PEG blocks instead of PLA blocks. Even though DMF is a good solvent for both PEG and PLA, their solubility in the solvent mixture could be different, causing the triblock copolymer molecules to aggregate. The average diameter of the aggregates was around 80 nm at concentrations above 1 x 10–3 g/ml (Figure 2B). Given that the bond lengths of C-C and C-O are 0.154 nm and 0.143 nm, respectively, the maximum backbone length of L8E78L8 should be <30 nm even if the polymer chain were fully stretched. Therefore, each LEL aggregate could contain more than one PEG domain, while the PEG domains should contain most of the water molecules.
Figure 3 Schematic presenting the condensed plasmid DNA, the aggregation of LEL and the encapsulation of DNA by LEL. (A–C) Shows TEM images of DNA in 1x TE buffer, DNA in 94% DMF + 6% TE and encapsulated DNA in 94% DMF + 6% TE, respectively.
Upon mixing with condensed DNA, which is hydrophilic, such PEG/water domains could serve as a ‘shelter’ for the plasmid DNA. As DNA and PLA were essentially incompatible, one expect that the LEL chains could accommodate the condensed DNA particles to form a core-shell structure. In other words, the DNA particles were encapsulated by LEL, which was exactly what we observed by laser light scattering results as well as TEM imaging at various stages (super coiled plasmid, condensed plasmid and encapsulated DNA nano-composite particle, Figure 3A–C). After mixing with 6.25 x 10–3 g/ml LEL, the size increased and the edge became blurred mainly due to the short LEL chains acting as a protective shell for the condensed DNA (Figure 3C). All of these experimental results were in agreement with our hypothesis. The core-shell structure, as shown in Figure 3, was quite different from the DNA structure, typically toroids (15,16) and rods (17), formed in aqueous solutions upon mixing with multivalent cations, including cationic polymers and polyamines (6).
Formulation of DNA particles into electrospun nanofibers
DNA nanoparticles with a core-shell structure were created in two steps: condensation and encapsulation. The main objective for the PLA shell was to produce an effective protective shell for the DNA globules during the electrospinning process. Moreover, such DNA particles could be incorporated into non-woven nanofibrous PLGA scaffolds for controlled release (11). The electrospun nanofibrous scaffolds possess many unique properties, including high surface-to-volume ratio, appropriate porosity and malleability, to conform to a wide variety of sizes and shapes, which make them superior substrates for tissue engineering (18) as well as for the delivery of cells, and bioactive agents, including drugs (19,20), protein (21) and DNA (11,22).
Based on laser light scattering results, PLGA (Mw = 75 k, LA/GA = 75/25) and plasmid DNA with or without LEL were formulated into a viscous solution, and the scaffold was produced by electrospinning. Following construction of the PLGA/DNA scaffolds with and without LEL, a comparison of scaffold morphology was made via SEM. Figure 4A shows an example of PLGA scaffolds having 10% w/w LEL and DNA with a LEL/DNA weight ratio of more than 150. The scaffolds contained thin and fairly uniformly distributed submicron diameter fibers. The SEM image of PLGA scaffold without LEL (data not shown) showed no visual difference in the overall morphology, porosity and fiber diameter. Thus, the presence or absence of a block copolymer has little effect on the overall structural morphology of these scaffolds. Upon the removal of remaining trace amounts of solvent (94% DMF + 6% TE) under vacuum, the ability of scaffolds to support cellular (pre-osteoblastic MC3T3) adhesion was tested, which results are illustrated in Figure 4B.
Figure 4 Incorporation and release of plasmid DNA from electrospun nanofibrous PLGA-based scaffolds. (A) SEM image of electrospun PLGA scaffold with LEL added; (B) Cellular adhesion on PLGA scaffold containing LEL and encapsulated DNA; (C) Gel electrophoresis of DNA samples, lane 1 ladder, lane 2 control plasmid, lanes 3 and 4 released (4 h) plasmid DNA from PLGA scaffold without/with LEL; (D and E) Shows the controlled release of DNA from PLGA scaffold with LEL.
DNA released from the scaffold was analyzed via the use of a Pico Green assay and agarose gel electrophoresis. Results showed that the DNA-containing scaffolds without the block copolymer contained degraded DNA (as a result of the electrospinning process), which is shown in lane 3 of Figure 4C where the degraded DNA appeared as a ‘smear’ spanning the length of the gel lane. In contrast, in the presence of block copolymer, the DNA released from the scaffold was essentially intact structurally, with no apparent degradation (lane 4 of Figure 4C). Lane 2 indicates unincorporated intact plasmid DNA, which was used as a positive control. Figure 4C demonstrates that the core-shell structure has protected the condensed DNA globule from degradation during the electrospinning process. More importantly, the structural integrity of the plasmid DNA was preserved.
To clarify how the DNA/LEL nanoparticles were incorporated in the PLGA fibers, the sustained release of intact DNA from PLGA scaffold was measured over a 7 day time period via the Pico Green assay. Figure 4D and E illustrate the amount of DNA released from the scaffold decreased with time. After 7 days, a large fraction (75%) of the DNA remained in the partially degraded scaffold. Thus, the DNA release profile could be primarily coupled with the scaffold degradation profile. It is noted that the structural integrity of the encapsulated and dried (after evaporation of solvent in the scaffold from electrospinning) DNA remained intact. During the release process, where a dry scaffold was immersed into an aqueous solution, LEL aggregates were disassembled and the LEL chains would likely form a flower-like micelle with PLA being the core and PEG being the corona (data not shown). This inside-out transformation would instantly cause the release of the encapsulated DNA. Therefore, the DNA release rate could also be influenced by the diffusion of aqueous solution into the PLGA scaffold. The biodegradation of the scaffold is also partially affected by the presence of encapsulated DNA. As only 25% DNA was released in 7 days, a large amount of the encapsulated DNA was successfully incorporated inside the ‘dry’ PLGA fiber. We could modulate the DNA release profile by tuning the chemical composition of PLGA, i.e. the LA/GA ratio (23).
After forming the scaffold, a small portion of the encapsulated DNA could be attached on and near the fiber surface. These DNA/LEL particles were quickly released upon dipping the scaffold into an aqueous solution. The release of encapsulated DNA inside the PLGA could depend on its location in the fiber, as clearly illustrated by the two release rates (Figure 4D). The remaining portion of these DNA/LEL particles could not be released until the biodegradable scaffold, whose degradation rate could be controlled, was degraded further.
Transfection of released DNA
There was a difference in the structural quality of the released DNA from the two types of scaffolds (with and without block copolymer). We sought to determine whether these differences could be translated into the bioactivity of the released DNA. Results from transfection experiments using pre-osteoblastic MC3T3 cells and sections of scaffolds immersed in culture medium, are presented in Figure 5. The results clearly indicate that no transfection was evident in cell populations incubated with PLGA and DNA scaffolds without LEL (Figure 5A). This lack of transfection could be correlated with the fact that the DNA released from the scaffold without the block copolymer was degraded (Figure 4C, lane 3). In contrast, transfection was observed from scaffolds where 10% LEL block copolymer was incorporated (Figure 5B). Compared with naked DNA that showed no transfected cells, (Figure 5C), the PLGA/LEL scaffold with encapsulated DNA enhanced transfection, albeit at very low levels, is <1%. An equivalent amount of DNA complexed with the transfection reagent, Fugene, resulted in 35% transfection efficiency (Figure 5D).
Figure 5 Transfection of MC3T3 cells by nanofibrous electrospun scaffolds. (A–D) shows the transfection of DNA released from PLGA without LEL (A), with LEL (B) and the negative (C) and positive (D) controls. Panels E–H show the increased transfection by plating MC3T3 cells directly on a GFP plasmid DNA containing scaffold. (E) Fluorescent image of cells on GFP plasmid DNA containing PLGA/10% block copolymer scaffold 24 h post-plating. Arrowheads indicate representative transfected GFP-expressing cells and arrows indicate non-transfected cells. (F) Light micrograph of cells (shown in E) stained with nuclear fast red. Arrowheads and arrows indicate identical cells in E. (G) Fluorescent image of cells on a control scaffold containing no DNA and indicating no fluorescent (green) cells (arrow). (H) Light micrograph of cells (shown in G) stained with nuclear fast red. Arrow indicates identical cells in G. This transfection efficiency reflected that of cells present only on the exterior surface of the scaffold and not for all cells within the scaffold.' Scale bar: 100 μm in Panels A–D, 10 μm in panels E–H.
The low level (<1%) transfection efficiency obtained could be due to the fact that the DNA/LEL particles disassembled in an aqueous solution. Thus, we reasoned that a more appropriate way of indicating transfection efficiency was warranted. Instead of immersing sections of scaffolds into a culture dish with a growing monolayer of MC3T3 cells, we plated cells directly onto a scaffold containing a green fluorescence protein (GFP) DNA plasmid. Results from this experiment showed a dramatic increase in transfection efficiency as seen in Figure 5E and F (note that this transfection efficiency reflects that of cells present on only the exterior surface of the scaffold and not for all cells within the scaffold). In contrast, no green cells were apparent in a scaffold without any GFP DNA, as expected (Figure 5G and H). The increase in DNA transfection occurred probably in two ways: (i) the scaffold brought cells closer to the released DNA by allowing cell adhesion directly onto the nanofibers, and (ii) the released DNA maintained a relatively higher concentration around the cell surface due to its immobilization (24), enabling greater levels of DNA to enter the cells.
Together with our previous studies (11), these results demonstrate that by understanding and capitalizing on the molecular interactions of block copolymers and plasmid DNA in solution, novel structures and additional functionality can emerge. A deeper understanding on the molecular self-assemblies and the electrospinning process has enabled us to develop an effective gene delivery vehicle. In the present case, in vitro gene delivery to cells was accomplished by attracting the cells to a scaffold that was capable of supporting cellular adhesion. Thus, the incorporation of a homing device could be of secondary importance. The volume contraction of a condensed DNA molecule was of the order of a 1000-fold, although less than what a virus could accomplish. Nevertheless, it permitted us to encapsulate and to protect the DNA plasmid for incorporation into electrospun nanofibrous scaffolds. Finally, these scaffolds with their network structures reduced the mobilization of released DNA before transfection. Thus, control of the fabricated gene delivery scaffold could be developed in terms of the scaffold morphology, material properties and release profiles, ultimately leading towards improved transfection efficiency.
ACKNOWLEDGEMENTS
We thank Christine Falabella for help with tissue culture. Ben Chu gratefully acknowledges the support of this work by the National Science Foundation (DMR 9984102 and 0454887), and in particular, for the creativity extension award that permits him to pursue this new direction. Funding to pay the Open Access publication charges for this article was provided by NSF (DMR 0454887).
REFERENCES
Brown, S.A. and Uchegbu, I.F. (2001) Gene delivery with synthetic (non-viral) carriers Int. J. Pharm., 229, 1–21 .
Shea, L.D., Smiley, E., Bonadio, J., Mooney, D.J. (1999) DNA delivery from polymer matrices for tissue engineering Nat. Biotechnol., 17, 551–554 .
Fang, J., Zhu, Y.Y., Smiley, E., Bonadio, J., Rouleau, J.P., Goldstein, S.A., McCauley, L.K., Davidson, B.L., Roessler, B.J. (1996) Stimulation of new bone formation by direct transfer of osteogenic plasmid genes Proc. Natl Acad. Sci. USA, 93, 5753–5758 .
Pouton, C.W. and Seymour, L.W. (2001) Key issues in non-viral gene delivery Adv. Drug Deliv. Rev., 46, 187–203 .
Bloomfield, V.A. (1996) DNA condensation Curr. Opin. Struct. Biol., 6, 334–341 .
Vijayanathan, V., Thomas, T., Thomas, T.J. (2002) DNA nanoparticles and development of DNA delivery vehicles for gene therapy Biochemistry, 41, 14085–14094 .
Wagner, E., Zenke, M., Cotten, M., Beug, H., Birnstiel, M.L. (1990) Transferrin-polycation conjugates as carriers for DNA uptake into cells Proc. Natl Acad. Sci. USA, 87, 3410–3414 .
Dunlap, D.D., Maggi, A., Soria, M.R., Monaco, L. (1997) Nanoscopic structure of DNA condensed for gene delivery Nucleic Acids Res., 25, 3095–3101 .
Saminathan, M., Thomas, T., Shirahata, A., Pillai, C.K.S., Thomas, T.J. (2002) Polyamine structural effects on the induction and stabilization of liquid crystalline DNA: potential applications to DNA packaging, gene therapy and polyamine therapeutics Nucleic Acids Res., 30, 3722–3731 .
Thomas, T.J. and Bloomfield, V.A. (1983) Collapse of DNA caused by trivalent cations: pH and ionic specificity effects Biopolymers, 22, 1097–1106 .
Luu, Y.K., Kim, K., Hsiao, B.S., Chu, B., Hadjiargyrou, M. (2003) Development of a nanostructured DNA delivery scaffold via electrospinning of PLGA and PLA-PEG block copolymers J. Control Release, 89, 341–353 .
Provencher, S.W. (1982) CONTIN: a general purpose constrained regularization program for inverting noisy linear algebraic and integral equations Comput. Phys. Commun., 27, 229–242 .
Burchard, W. (1983) Static and dynamic light scattering from branched polymers and biopolymers Adv. Polym. Sci., 48, 1–124 .
Kim, K., Chung, S., Chin, I., Kim, M., Yoon, J. (1999) Crystallization behavior of biodegradable amphiphilic poly(ethylene glycol)-poly(L-lactide) block copolymers J. Appl. Polym. Sci., 72, 341–348 .
Golan, R., Pietrasanta, L.I., Hsieh, W., Hansma, H.G. (1999) DNA toroids: stages in condensation Biochemistry, 38, 14069–14076 .
Vijayanathan, V., Thomas, T., Antony, T., Shirahata, A., Thomas, T.J. (2004) Formation of DNA nanoparticles in the presence of novel polyamine analogs: a laser light scattering and atomic force microscopic study Nucleic Acids Res., 32, 127–134 .
Rouzina, I. and Bloomfield, V.A. (1998) DNA bending by small, mobile multivalent cations Biophys J., 74, 3152–3164 .
Yoshimoto, H., Shin, Y.M., Terai, H., Vacanti, J.P. (2003) A biodegradable nanofiber scaffold by electrospinning and its potential for bone tissue engineering Biomaterials, 24, 2077–2082 .
Verreck, G., Chun, I., Rosenblatt, J., Peeters, J., Dijck, A.V., Mensch, J., Noppe, M., Brewster, M.E. (2003) Incorporation of drugs in an amorphous state into electrospun nanofibers composed of a water-insoluble, nonbiodegradable polymer J. Control Release, 92, 349–360 .
Kim, K., Luu, Y.K., Chang, C., Fang, D., Hsiao, B.S., Chu, B., Hadjiargyrou, M. (2004) Incorporation and controlled release of a hydrophilic antibiotic using poly(lactide-co-glycolide)-based electrospun nanofibrous scaffolds J. Control Release, 98, 47–56 .
Amsden, B.G. and Goosen, M.F.A. (1997) An examination of factors affecting the size, distribution and release characteristics of polymer microbeads made using electrostatics J. Control Release, 43, 183–196 .
Zeng, J., Xu, X., Chen, X., Liang, Q., Bian, X., Yang, L., Jing, X. (2003) Biodegradable electrospun fibers for drug delivery J. Control Release, 92, 227–231 .
Zong, X., Kim, K., Ran, S., Fang, D., Hsiao, B.S., Chu, B. Biomacromolecules, (2003) 4, 416–423 .
Segura, T., Volk, M.J., Shea, L.D. (2003) Substrate-mediated DNA delivery: role of the cationic polymer structure and extent of modification J. Control Release, 93, 69–84 .(Dehai Liang1, Yen K. Luu2, Kwangsok Kim1)
*To whom correspondence should be addressed. Benjamin Chu, Department of Chemistry, Stony Brook University, Stony Brook, NY 11794-3400, USA. Tel: +631 632 7928; Fax: +631 632 6518; Email: bchu@notes.cc.sunysb.edu
ABSTRACT
Extracellular and intracellular barriers typically prevent non-viral gene vectors from having an effective transfection efficiency. Formulation of a gene delivery vehicle that can overcome the barriers is a key step for successful tissue regeneration. We have developed a novel core-shelled DNA nanoparticle by invoking solvent-induced condensation of plasmid DNA (?-galactosidase or GFP) in a solvent mixture and subsequent encapsulation of the condensed DNA globule in a triblock copolymer, polylactide-poly(ethylene glycol)-polylactide (L8E78L8), in the same solvent environment. The polylactide shell protects the encapsulated DNA from degradation during electrospinning of a mixture of encapsulated DNA nanoparticles and biodegradable PLGA (a random copolymer of lactide and glycolide) to form a nanofibrous non-woven scaffold using the same solution mixture. The bioactive plasmid DNA can then be released in an intact form from the scaffold with a controlled release rate and transfect cells in vitro.
INTRODUCTION
The search for viable alternative approaches to deal with problems associated with viral DNA delivery (i.e. safety, toxicity and fate of transfected cells, etc.) has prompted many researchers to develop non-viral means of delivering genes in vivo. Typical non-viral approaches include naked DNA, cationic lipids formulated into liposomes and subsequent complex formation with DNA (lipoplexes), complex formation of cationic polymers with DNA (polyplexes), and collagen- or hyaluronan-based DNA gels (1). Recently, new biomaterials, such as the gene activated matrix (GAM) and biodegradable polymers (i.e. polylactide-co-glycolide, PLGA) have been developed that can serve as scaffolds for DNA delivery and tissue engineering (2,3) Even though non-viral gene delivery systems offer enhanced safety over viral systems, they are plagued with comparatively low transfection efficiencies. Non-viral gene delivery encounters many barriers, including degradation of the DNA in plasma, uptake of DNA by the reticuloendothelial system, lysosomal degradation of the DNA and lack of translocation to the nucleus (1,4). To overcome these barriers, the critical component of design formulation must be based on the fundamental understanding of the interactions between DNA and various chemical species in solution. Proper characterization of vectors thus becomes an essential step in the development of efficient non-viral gene delivery systems.
The process of DNA condensation has attracted a great deal of attention in recent years due to its biological importance in viral DNA packing, as well as in the development of gene delivery vehicles (5,6). The extended DNA chain could effectively be collapsed into a compact globule through charge neutralization by cationic polymers (7,8), polyamines and other multivalent cations, such as spermidine (9) and cobalt hexamine [ (10). In this study, we chose the approach of solvent-induced condensation of DNA. There are two advantages of using a poor (or non-) solvent for DNA to undergo coil-to-globule transition: (i) no contamination by other unwanted species, and (ii) easy removal of solvent through evaporation. The chosen poor solvent was 94% (in volume) N,N-dimethylformamide (DMF) + 6% 1x TE buffer. It was found that DNA could remain intact in DMF for at least several days and its bioactivity was maintained.
We have designed a unique scheme that can overcome some barriers in non-viral gene delivery. Our scheme is as follows. Encapsulated DNA nano-particles with a core-shell structure were first created by condensing plasmid DNA in a poor solvent mixture, followed by encapsulation of the condensed DNA in a triblock copolymer of poly(lactide)-b-poly(ethylene glycol)-b-poly(lactide) (LEL) that formed micelles under the same solvent environment. This mixture was designed as a suitable solvent for random copolymers of lactide and glycolide (PLGA). The mixtures of encapsulated DNA and PLGA were then electrospun to form a non-woven nanofibrous and nanocomposite scaffold, whereby the solvent mixture was rapidly removed by evaporation in the jet stream during fiber formation in the electrospinning process. Thus, the plasmid DNA was preserved in the nanofibrous scaffold. As a gene delivery vehicle, the polylactide shell protected the DNA core from degradation and maintained its bioactivity during electrospinning. The encapsulated DNA could be released from the scaffold by controlled degradation of the biodegradable scaffold component (PLGA). Transfection of adhered cells could then be enhanced by sustained DNA release and immobilization in contact with adhered cells.
MATERIALS AND METHODS
Preparation of DNA and triblock polymer
pCMV? plasmid DNA (7164 bp) encoding ?-galactosidase (Clontech, Palo Alto, CA) was extracted from cultured Escherichia coli utilizing a Qiagen (Valencia, CA) GigaPrep DNA isolation kit. PLA-PEG-PLA triblock copolymer was synthesized by copolymerizing PEG (3.4K) and D,L-lactide at 130°C for 15 h with stannous octoate being the catalyst. The molecular weight of purified reaction product was determined by 1H nuclear magnetic resonance (1H NMR) and gel permeation chromatography (GPC).
Laser light scattering (LLS)
A LLS spectrometer equipped with a BI-9000 AT digital correlator and a solid-state laser (DPSS, Coherent, 200 mW and 532 nm) was used to perform static and dynamic light scattering (DLS) studies over an angular range of 20–120°. In static LLS, the angular dependence of the excess absolute time-averaged scattered intensity, known as the Rayleigh ratio Rvv(), was measured. For a very dilute solution, the weight-averaged molecular weight (Mw), the root mean square radius of gyration (Rg), and the second virial coefficient (A2) can be obtained on the basis of
(1)
where H = 42n2(dn/dC)2/(NA4) and q = (4n/)sin(/2) with NA, n, dn/dC and being the Avogadro constant, the solvent refractive index, the specific refractive index increment and the wavelength of light in vacuum, respectively.
In dynamic LLS, the intensity–intensity time correlation function G(2)(t) in the self-beating mode was measured.
(2)
where A is the measured base line, ? is a coherence factor, t is the delay time and g(1)(t) is the normalized first-order electric field time correlation function. g(1)(t) is related to the line width distribution G() by
(3)
By using a Laplace inversion program, CONTIN, the normalized distribution function of the characteristic line width G() was obtained. The line width is a function of both C and q, which can be expressed as
(4)
with D, kd and f being the translational diffusive coefficient, the diffusion second virial coefficient, and a dimensionless constant, respectively. When the concentration is extremely dilute and Rgq << 1, /q2 is approximately equal to D. D can be further converted into the hydrodynamic radius Rh by using the Stokes–Einstein equation:
(5)
where kB, T and are the Boltzmann constant, the absolute temperature, and the viscosity of the solvent, respectively.
Measurement on dn/dC
Differential refractometer (Phoenix Precision Instruments, NY) was used to measure the dn/dC value of PLA and PEG in 94% DMF + 6% TE at 25°C. The instrument was calibrated in advance by using aqueous potassium chloride solutions. At a wavelength of 546 nm, the measured dn/dC of PLA and PEG in 94% DMF + 6% TE exhibited a value of 0.030 ml/g and 0.044 ml/g, respectively. The dn/dC values of plasmid DNA in 1x TE buffer and in 94% DMF + 6% TE were 0.17 ml/g and 0.10 ml/g, respectively.
Scaffold preparation and DNA release
DNA incorporation into electrospun scaffolds and DNA release assays from scaffold sections have been reported previously by our laboratory (11). The scaffold preparation procedures in this study were similar to those reported before. Briefly, PLA-PEG block copolymer was dissolved in N,N-dimethylformamide, to which pCMV? plasmid (5 mg/ml, in TE Buffer) and PLGA (LA/GA = 75/25) were added. The final solution contains 94% DMF + 6% TE. Solutions and scaffolds without block copolymer (only PLGA and DNA) were prepared identically except for the omission of block copolymer. Scaffolds were electrospun at 25 kV with a solution flow rate of 20 μl/min. The spinneret (anode) was fixed at 15 cm above the aluminum-covered rotating collection drum (cathode).
The electrospun scaffolds were cut into 1.5 x 1.0 cm sections and each section was incubated at 37°C with 1 ml TE buffer in Eppendorf tubes. The amount of DNA released into solution was quantified by using the Pico Green Assay, where solutions were excited at 485 nm, and the emission was measured at 530 nm in a microplate reader (CytoFluor Series 4000, Perseptive Biosystems). Integrity of released DNA was determined by 0.8% agarose gel electrophoresis and visualized by ethidium bromide staining.
Cell culture and transfection
Transfection studies were conducted using the MC3T3-E1 pre-osteoblastic cell line. MC3T3 cells were maintained in log growth phase using -MEM supplemented with 10% fetal bovine serum (FBS) (Life Technologies, Grand Island, NY). Cells were plated at an initial density of 1 x 105 cells/well into 6-well plates, 24 h before transfection. The negative control was designated as 2 μg naked DNA added directly to cell culture solution. A positive control was constructed using the commercially available Fugene 6 transfection reagent . Cells were incubated with two 1.5 x 1 cm sections of DNA loaded scaffolds, with and without the block copolymer. For the other set of transfection experiments with the GFP plasmid, 1 x 105 cells (20 μl) were placed directly in the center of 1.5 x 1 cm sections of PLGA/LEL and PLGA/LEL/GFP plasmid DNA scaffold (n = 3/scaffold type). Initially, the 20 μl droplet remained intact in the center of the scaffold, but it spread out to cover the entire scaffold with time (1 h). After 1 h, the scaffold was transferred into a 12-well tissue culture plate, containing cell culture medium. After 24 h, the scaffolds were removed from the tissue culture plates and fixed with 1% formaldehyde before staining with 0.1% Nuclear Fast Red (Sigma) prepared in 5% aluminum sulfate. The scaffolds were then washed twice with distilled water, and dehydrated with an ethanol gradient (70, 95 and 100%). Finally, the scaffolds were placed on microscope slides and cells (present only on the exterior surface of the scaffold) were imaged under epifluorescence for the presence of GFP using a Zeiss Axiovert microscope.
RESULTS AND DISCUSSION
LLS
Both static light scattering (SLS) and DLS, was employed to monitor the changes in DNA size and conformation. Figure 1A and B show Zimm plots of pCMV? plasmid DNA (7164 bp) encoding ?-galactosidase in 1x TE buffer and in 94% DMF + 6% TE, respectively. From the Zimm plot, the weight average molecular weight (Mw), the radius of gyration (Rg) and the second virial coefficient (A2) were obtained (Table 1). In 94% DMF + 6% TE, Mw of plasmid DNA was only slightly larger than that in 1x TE buffer. However, the Rg value was significantly reduced from 135 nm (in 1x TE buffer) to about 15 nm, suggesting the occurrence of chain collapse. A2 was used to describe the interactions between solute and solvent. In general, A2 is positive when a sample was dissolved in a good solvent. As expected, the value of A2 (–3.7 x 10–4 mol cm3 g–2) became negative in 94% DMF + 6% TE because this mixture was a poor solvent for DNA. The hydrodynamic radius, Rh, obtained from the CONTIN (12) analysis of dynamic light scattering data, was also dramatically decreased from 87 nm in 1x TE buffer to about 17 nm in 94% DMF + 6% TE (Figure 1C). The relationship between the conformation of polymer chains and the Rg/Rh ratio has been well established. Typically, a random coil has an Rg/Rh value of 1.5, and a solid sphere has a value of 0.775 (13). As shown in Table 1, plasmid DNA had an Rg/Rh ratio of 1.55 in 1x TE buffer, suggesting essentially a random-coil conformation; while in 94% DMF + 6% TE, the value was about 0.88, very close to 0.775, indicating a relatively compact solid-sphere conformation.
Figure 1 Condensation of plasmid DNA in 94% DMF + 6% TE studied by laser light scattering. (A and B) Shows the Zimm plots of plasmid DNA in 1x TE buffer, and in 94% DMF + 6% TE, respectively. (C) Shows the size distribution of DNA measured by DLS at 90° and 1.0 x 10–5 g/ml DNA concentration.
Table 1 Physicochemical properties of plasmid DNA measured by LLS
The triblock copolymer used in this study (LEL), was synthesized according to a known scheme (14). In brief, PEG and D,L-lactide with known ratio were mixed together and the polymerization was accomplished at 130°C for 15 h using stannous octoate as the catalyst. Before mixing plasmid DNA and LEL together, the behavior of LEL itself in 94% DMF + 6% TE was studied. With the known molecular weight of PEG block (3.4 k), the number-averaged molecular weight of PLA block could be estimated from proton NMR (data not shown), which was about 0.6 k. Therefore, the average repeating unit of each block was estimated to be L8E78L8. GPC also confirmed this as the weight average molecular weight of LEL was about 5.9 k.
Figure 2A shows the excess scattered intensity of LEL at different copolymer concentrations in the same solvent mixture. No sharp increase in the scattered intensity over a range of concentrations was observed, suggesting either the absence of micelle formation or a very low critical micelle concentration. At concentrations higher than 1 x 10–3 g/ml, LEL showed Mw of about 180 k (Figure 2B), a value at almost 30 times larger than that measured by GPC. The measured Rg value was 39 nm, significantly larger than the size of a polymer chain with an Mw of only 5.9 k. The hydrodynamic radius, Rh, measured from DLS (data not shown), had a value of 44 nm, very close to Rg. All these results confirmed that LEL chains were aggregated in 94% DMF + 6% TE at concentrations above 1 x 10–3 g/ml.
Figure 2 Encapsulation of plasmid DNA by triblock copolymer of LEL in 94% DMF + 6% TE. (A) Excess scattered intensity of LEL at different concentrations measured at 90°; (B) Zimm plot of LEL at concentrations of 1.25, 6.24 and 31.2 mg/ml; (C and D) CONTIN analysis of plasmid DNA (hollow circles), LEL (hollow squares) and their mixtures (solid diamonds) at different DNA/LEL weight ratio. The concentration of DNA used in panels C and D is 1 x 10–5 g/ml.
DNA (at 1 x 10–5 g/ml) was subsequently mixed with LEL at two different concentrations, 3.12 x 10–4 g/ml and 6.25 x 10–3 g/ml. At 3.12 x 10–4 g/ml, where LEL did not aggregate, the size and distribution of the DNA/LEL mixture were almost the same as those of plasmid DNA itself (Figure 2C). In other words, there was an insufficient amount of LEL to encapsulate the condensed plasmid DNA. At LEL concentration of 6.25 x 10–3 g/ml, a value where LEL already aggregated, only one ‘size’ with a relatively broad size distribution was observed for DNA/LEL in the solvent mixture (Figure 2D). The average size, however, became even smaller than that of the pure LEL aggregate, but larger than that of the plasmid DNA itself, indicating that a fusion of plasmid DNA and LEL aggregates had occurred.
The molecular scale interaction of DNA and LEL is depicted schematically in Figure 3. The figure also included experimental evidence for the condensation of plasmid DNA, the aggregation of LEL and the encapsulation of DNA by LEL in 94% DMF + 6% TE. Initially, after being air dried from 1x TE buffer, plasmid DNA was rod-like with a length >200 nm (Figure 3A), very close to the shape of a compact random coil with high super-helicity. Depending on the chain length and the degree of super-helicity, the DNA chain could be loose, compact or even branched. In 94% DMF + 6% TE, the DNA chain was condensed to form a much smaller globule with a diameter of 30 nm (Figure 3B). By design, LEL was also aggregated in 94% DMF + 6% TE above a certain concentration, which can be explained as follows. TE is a good solvent for PEG but not for PLA. In the solution mixture, water molecules prefer to stay together with PEG blocks instead of PLA blocks. Even though DMF is a good solvent for both PEG and PLA, their solubility in the solvent mixture could be different, causing the triblock copolymer molecules to aggregate. The average diameter of the aggregates was around 80 nm at concentrations above 1 x 10–3 g/ml (Figure 2B). Given that the bond lengths of C-C and C-O are 0.154 nm and 0.143 nm, respectively, the maximum backbone length of L8E78L8 should be <30 nm even if the polymer chain were fully stretched. Therefore, each LEL aggregate could contain more than one PEG domain, while the PEG domains should contain most of the water molecules.
Figure 3 Schematic presenting the condensed plasmid DNA, the aggregation of LEL and the encapsulation of DNA by LEL. (A–C) Shows TEM images of DNA in 1x TE buffer, DNA in 94% DMF + 6% TE and encapsulated DNA in 94% DMF + 6% TE, respectively.
Upon mixing with condensed DNA, which is hydrophilic, such PEG/water domains could serve as a ‘shelter’ for the plasmid DNA. As DNA and PLA were essentially incompatible, one expect that the LEL chains could accommodate the condensed DNA particles to form a core-shell structure. In other words, the DNA particles were encapsulated by LEL, which was exactly what we observed by laser light scattering results as well as TEM imaging at various stages (super coiled plasmid, condensed plasmid and encapsulated DNA nano-composite particle, Figure 3A–C). After mixing with 6.25 x 10–3 g/ml LEL, the size increased and the edge became blurred mainly due to the short LEL chains acting as a protective shell for the condensed DNA (Figure 3C). All of these experimental results were in agreement with our hypothesis. The core-shell structure, as shown in Figure 3, was quite different from the DNA structure, typically toroids (15,16) and rods (17), formed in aqueous solutions upon mixing with multivalent cations, including cationic polymers and polyamines (6).
Formulation of DNA particles into electrospun nanofibers
DNA nanoparticles with a core-shell structure were created in two steps: condensation and encapsulation. The main objective for the PLA shell was to produce an effective protective shell for the DNA globules during the electrospinning process. Moreover, such DNA particles could be incorporated into non-woven nanofibrous PLGA scaffolds for controlled release (11). The electrospun nanofibrous scaffolds possess many unique properties, including high surface-to-volume ratio, appropriate porosity and malleability, to conform to a wide variety of sizes and shapes, which make them superior substrates for tissue engineering (18) as well as for the delivery of cells, and bioactive agents, including drugs (19,20), protein (21) and DNA (11,22).
Based on laser light scattering results, PLGA (Mw = 75 k, LA/GA = 75/25) and plasmid DNA with or without LEL were formulated into a viscous solution, and the scaffold was produced by electrospinning. Following construction of the PLGA/DNA scaffolds with and without LEL, a comparison of scaffold morphology was made via SEM. Figure 4A shows an example of PLGA scaffolds having 10% w/w LEL and DNA with a LEL/DNA weight ratio of more than 150. The scaffolds contained thin and fairly uniformly distributed submicron diameter fibers. The SEM image of PLGA scaffold without LEL (data not shown) showed no visual difference in the overall morphology, porosity and fiber diameter. Thus, the presence or absence of a block copolymer has little effect on the overall structural morphology of these scaffolds. Upon the removal of remaining trace amounts of solvent (94% DMF + 6% TE) under vacuum, the ability of scaffolds to support cellular (pre-osteoblastic MC3T3) adhesion was tested, which results are illustrated in Figure 4B.
Figure 4 Incorporation and release of plasmid DNA from electrospun nanofibrous PLGA-based scaffolds. (A) SEM image of electrospun PLGA scaffold with LEL added; (B) Cellular adhesion on PLGA scaffold containing LEL and encapsulated DNA; (C) Gel electrophoresis of DNA samples, lane 1 ladder, lane 2 control plasmid, lanes 3 and 4 released (4 h) plasmid DNA from PLGA scaffold without/with LEL; (D and E) Shows the controlled release of DNA from PLGA scaffold with LEL.
DNA released from the scaffold was analyzed via the use of a Pico Green assay and agarose gel electrophoresis. Results showed that the DNA-containing scaffolds without the block copolymer contained degraded DNA (as a result of the electrospinning process), which is shown in lane 3 of Figure 4C where the degraded DNA appeared as a ‘smear’ spanning the length of the gel lane. In contrast, in the presence of block copolymer, the DNA released from the scaffold was essentially intact structurally, with no apparent degradation (lane 4 of Figure 4C). Lane 2 indicates unincorporated intact plasmid DNA, which was used as a positive control. Figure 4C demonstrates that the core-shell structure has protected the condensed DNA globule from degradation during the electrospinning process. More importantly, the structural integrity of the plasmid DNA was preserved.
To clarify how the DNA/LEL nanoparticles were incorporated in the PLGA fibers, the sustained release of intact DNA from PLGA scaffold was measured over a 7 day time period via the Pico Green assay. Figure 4D and E illustrate the amount of DNA released from the scaffold decreased with time. After 7 days, a large fraction (75%) of the DNA remained in the partially degraded scaffold. Thus, the DNA release profile could be primarily coupled with the scaffold degradation profile. It is noted that the structural integrity of the encapsulated and dried (after evaporation of solvent in the scaffold from electrospinning) DNA remained intact. During the release process, where a dry scaffold was immersed into an aqueous solution, LEL aggregates were disassembled and the LEL chains would likely form a flower-like micelle with PLA being the core and PEG being the corona (data not shown). This inside-out transformation would instantly cause the release of the encapsulated DNA. Therefore, the DNA release rate could also be influenced by the diffusion of aqueous solution into the PLGA scaffold. The biodegradation of the scaffold is also partially affected by the presence of encapsulated DNA. As only 25% DNA was released in 7 days, a large amount of the encapsulated DNA was successfully incorporated inside the ‘dry’ PLGA fiber. We could modulate the DNA release profile by tuning the chemical composition of PLGA, i.e. the LA/GA ratio (23).
After forming the scaffold, a small portion of the encapsulated DNA could be attached on and near the fiber surface. These DNA/LEL particles were quickly released upon dipping the scaffold into an aqueous solution. The release of encapsulated DNA inside the PLGA could depend on its location in the fiber, as clearly illustrated by the two release rates (Figure 4D). The remaining portion of these DNA/LEL particles could not be released until the biodegradable scaffold, whose degradation rate could be controlled, was degraded further.
Transfection of released DNA
There was a difference in the structural quality of the released DNA from the two types of scaffolds (with and without block copolymer). We sought to determine whether these differences could be translated into the bioactivity of the released DNA. Results from transfection experiments using pre-osteoblastic MC3T3 cells and sections of scaffolds immersed in culture medium, are presented in Figure 5. The results clearly indicate that no transfection was evident in cell populations incubated with PLGA and DNA scaffolds without LEL (Figure 5A). This lack of transfection could be correlated with the fact that the DNA released from the scaffold without the block copolymer was degraded (Figure 4C, lane 3). In contrast, transfection was observed from scaffolds where 10% LEL block copolymer was incorporated (Figure 5B). Compared with naked DNA that showed no transfected cells, (Figure 5C), the PLGA/LEL scaffold with encapsulated DNA enhanced transfection, albeit at very low levels, is <1%. An equivalent amount of DNA complexed with the transfection reagent, Fugene, resulted in 35% transfection efficiency (Figure 5D).
Figure 5 Transfection of MC3T3 cells by nanofibrous electrospun scaffolds. (A–D) shows the transfection of DNA released from PLGA without LEL (A), with LEL (B) and the negative (C) and positive (D) controls. Panels E–H show the increased transfection by plating MC3T3 cells directly on a GFP plasmid DNA containing scaffold. (E) Fluorescent image of cells on GFP plasmid DNA containing PLGA/10% block copolymer scaffold 24 h post-plating. Arrowheads indicate representative transfected GFP-expressing cells and arrows indicate non-transfected cells. (F) Light micrograph of cells (shown in E) stained with nuclear fast red. Arrowheads and arrows indicate identical cells in E. (G) Fluorescent image of cells on a control scaffold containing no DNA and indicating no fluorescent (green) cells (arrow). (H) Light micrograph of cells (shown in G) stained with nuclear fast red. Arrow indicates identical cells in G. This transfection efficiency reflected that of cells present only on the exterior surface of the scaffold and not for all cells within the scaffold.' Scale bar: 100 μm in Panels A–D, 10 μm in panels E–H.
The low level (<1%) transfection efficiency obtained could be due to the fact that the DNA/LEL particles disassembled in an aqueous solution. Thus, we reasoned that a more appropriate way of indicating transfection efficiency was warranted. Instead of immersing sections of scaffolds into a culture dish with a growing monolayer of MC3T3 cells, we plated cells directly onto a scaffold containing a green fluorescence protein (GFP) DNA plasmid. Results from this experiment showed a dramatic increase in transfection efficiency as seen in Figure 5E and F (note that this transfection efficiency reflects that of cells present on only the exterior surface of the scaffold and not for all cells within the scaffold). In contrast, no green cells were apparent in a scaffold without any GFP DNA, as expected (Figure 5G and H). The increase in DNA transfection occurred probably in two ways: (i) the scaffold brought cells closer to the released DNA by allowing cell adhesion directly onto the nanofibers, and (ii) the released DNA maintained a relatively higher concentration around the cell surface due to its immobilization (24), enabling greater levels of DNA to enter the cells.
Together with our previous studies (11), these results demonstrate that by understanding and capitalizing on the molecular interactions of block copolymers and plasmid DNA in solution, novel structures and additional functionality can emerge. A deeper understanding on the molecular self-assemblies and the electrospinning process has enabled us to develop an effective gene delivery vehicle. In the present case, in vitro gene delivery to cells was accomplished by attracting the cells to a scaffold that was capable of supporting cellular adhesion. Thus, the incorporation of a homing device could be of secondary importance. The volume contraction of a condensed DNA molecule was of the order of a 1000-fold, although less than what a virus could accomplish. Nevertheless, it permitted us to encapsulate and to protect the DNA plasmid for incorporation into electrospun nanofibrous scaffolds. Finally, these scaffolds with their network structures reduced the mobilization of released DNA before transfection. Thus, control of the fabricated gene delivery scaffold could be developed in terms of the scaffold morphology, material properties and release profiles, ultimately leading towards improved transfection efficiency.
ACKNOWLEDGEMENTS
We thank Christine Falabella for help with tissue culture. Ben Chu gratefully acknowledges the support of this work by the National Science Foundation (DMR 9984102 and 0454887), and in particular, for the creativity extension award that permits him to pursue this new direction. Funding to pay the Open Access publication charges for this article was provided by NSF (DMR 0454887).
REFERENCES
Brown, S.A. and Uchegbu, I.F. (2001) Gene delivery with synthetic (non-viral) carriers Int. J. Pharm., 229, 1–21 .
Shea, L.D., Smiley, E., Bonadio, J., Mooney, D.J. (1999) DNA delivery from polymer matrices for tissue engineering Nat. Biotechnol., 17, 551–554 .
Fang, J., Zhu, Y.Y., Smiley, E., Bonadio, J., Rouleau, J.P., Goldstein, S.A., McCauley, L.K., Davidson, B.L., Roessler, B.J. (1996) Stimulation of new bone formation by direct transfer of osteogenic plasmid genes Proc. Natl Acad. Sci. USA, 93, 5753–5758 .
Pouton, C.W. and Seymour, L.W. (2001) Key issues in non-viral gene delivery Adv. Drug Deliv. Rev., 46, 187–203 .
Bloomfield, V.A. (1996) DNA condensation Curr. Opin. Struct. Biol., 6, 334–341 .
Vijayanathan, V., Thomas, T., Thomas, T.J. (2002) DNA nanoparticles and development of DNA delivery vehicles for gene therapy Biochemistry, 41, 14085–14094 .
Wagner, E., Zenke, M., Cotten, M., Beug, H., Birnstiel, M.L. (1990) Transferrin-polycation conjugates as carriers for DNA uptake into cells Proc. Natl Acad. Sci. USA, 87, 3410–3414 .
Dunlap, D.D., Maggi, A., Soria, M.R., Monaco, L. (1997) Nanoscopic structure of DNA condensed for gene delivery Nucleic Acids Res., 25, 3095–3101 .
Saminathan, M., Thomas, T., Shirahata, A., Pillai, C.K.S., Thomas, T.J. (2002) Polyamine structural effects on the induction and stabilization of liquid crystalline DNA: potential applications to DNA packaging, gene therapy and polyamine therapeutics Nucleic Acids Res., 30, 3722–3731 .
Thomas, T.J. and Bloomfield, V.A. (1983) Collapse of DNA caused by trivalent cations: pH and ionic specificity effects Biopolymers, 22, 1097–1106 .
Luu, Y.K., Kim, K., Hsiao, B.S., Chu, B., Hadjiargyrou, M. (2003) Development of a nanostructured DNA delivery scaffold via electrospinning of PLGA and PLA-PEG block copolymers J. Control Release, 89, 341–353 .
Provencher, S.W. (1982) CONTIN: a general purpose constrained regularization program for inverting noisy linear algebraic and integral equations Comput. Phys. Commun., 27, 229–242 .
Burchard, W. (1983) Static and dynamic light scattering from branched polymers and biopolymers Adv. Polym. Sci., 48, 1–124 .
Kim, K., Chung, S., Chin, I., Kim, M., Yoon, J. (1999) Crystallization behavior of biodegradable amphiphilic poly(ethylene glycol)-poly(L-lactide) block copolymers J. Appl. Polym. Sci., 72, 341–348 .
Golan, R., Pietrasanta, L.I., Hsieh, W., Hansma, H.G. (1999) DNA toroids: stages in condensation Biochemistry, 38, 14069–14076 .
Vijayanathan, V., Thomas, T., Antony, T., Shirahata, A., Thomas, T.J. (2004) Formation of DNA nanoparticles in the presence of novel polyamine analogs: a laser light scattering and atomic force microscopic study Nucleic Acids Res., 32, 127–134 .
Rouzina, I. and Bloomfield, V.A. (1998) DNA bending by small, mobile multivalent cations Biophys J., 74, 3152–3164 .
Yoshimoto, H., Shin, Y.M., Terai, H., Vacanti, J.P. (2003) A biodegradable nanofiber scaffold by electrospinning and its potential for bone tissue engineering Biomaterials, 24, 2077–2082 .
Verreck, G., Chun, I., Rosenblatt, J., Peeters, J., Dijck, A.V., Mensch, J., Noppe, M., Brewster, M.E. (2003) Incorporation of drugs in an amorphous state into electrospun nanofibers composed of a water-insoluble, nonbiodegradable polymer J. Control Release, 92, 349–360 .
Kim, K., Luu, Y.K., Chang, C., Fang, D., Hsiao, B.S., Chu, B., Hadjiargyrou, M. (2004) Incorporation and controlled release of a hydrophilic antibiotic using poly(lactide-co-glycolide)-based electrospun nanofibrous scaffolds J. Control Release, 98, 47–56 .
Amsden, B.G. and Goosen, M.F.A. (1997) An examination of factors affecting the size, distribution and release characteristics of polymer microbeads made using electrostatics J. Control Release, 43, 183–196 .
Zeng, J., Xu, X., Chen, X., Liang, Q., Bian, X., Yang, L., Jing, X. (2003) Biodegradable electrospun fibers for drug delivery J. Control Release, 92, 227–231 .
Zong, X., Kim, K., Ran, S., Fang, D., Hsiao, B.S., Chu, B. Biomacromolecules, (2003) 4, 416–423 .
Segura, T., Volk, M.J., Shea, L.D. (2003) Substrate-mediated DNA delivery: role of the cationic polymer structure and extent of modification J. Control Release, 93, 69–84 .(Dehai Liang1, Yen K. Luu2, Kwangsok Kim1)