Plasmid DNA-Entrapped Nanoparticles Engineered from

Zhengrong Cui and Russell J. Mumper*. Division of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky,. Lexington, Kentucky 40536-008...
0 downloads 0 Views 171KB Size
Download PDF
Bioconjugate Chem. 2002, 13, 1319−1327

1319

Plasmid DNA-Entrapped Nanoparticles Engineered from Microemulsion Precursors: In Vitro and in Vivo Evaluation Zhengrong Cui and Russell J. Mumper* Division of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, Kentucky 40536-0082. Received June 10, 2002; Revised Manuscript Received July 29, 2002

Nonviral gene therapy has been a rapidly growing field. However, delivery systems that can provide protection for pDNA and potential targeting are still desired. A novel pDNA-nanoparticle delivery system was developed by entrapping hydrophobized pDNA inside nanoparticles engineered from oilin-water (O/W) microemulsion precursors. Plasmid DNA was hydrophobized by complexing with cationic surfactants DOTAP and DDAB. Warm O/W microemulsions were prepared at 50-55 °C with emulsifying wax, Brij 78, Tween 20, and Tween 80. Nanoparticles were engineered by simply cooling the O/W microemulsions containing the hydrophobized pDNA in the oil phase to room temperature while stirring. The nanoparticles were characterized by particle sizing, zeta-potential, and TEM. Nanoparticles were challenged with serum nucleases to assess pDNA stability. In addition, the nanoparticles were coincubated with simulated biological media to assess their stability. In vitro hepatocyte transfection studies were completed with uncoated nanoparticles or nanoparticles coated with pullulan, a hepatocyte targeting ligand. In vivo biodistribution of the nanoparticles containing I-125 labeled pDNA was monitored 30 min after tail-vein injection to Balb/C mice. Depending on the hydrophobizing lipid agent employed, uniform pDNA-entrapped nanoparticles (100-160 nm in diameter) were engineered within minutes from warm O/W microemulsion precursors. The nanoparticles were negatively charged (-6 to -15 mV) and spherical. An anionic exchange column was used to separate unentrapped pDNA from nanoparticles. Gel permeation chromatography of pDNAentrapped and serum-digested nanoparticles showed that the incorporation efficiency was ∼30%. Free ‘naked’ pDNA was completely digested by serum nucleases while the entrapped pDNA remained intact. Moreover, in vitro transfection studies in Hep G2 cells showed that pullulan-coated nanoparticles resulted in enhanced luciferase expression, compared to both pDNA alone and uncoated nanoparticles. Preincubation of the cells with free pullulan inhibited the transfection. Finally, 30 min after tail vein injection to mice, only 16% of the ‘naked’ pDNA remained in the circulating blood compared to over 40% of the entrapped pDNA. Due to the apparent stability of these pDNA-entrapped nanoparticles in the blood, they may have potential for systemic gene therapy applications requiring cell and/or tissuespecific delivery.

INTRODUCTION

Gene therapy has attracted a great deal of attention in recent years. The current gene delivery systems can be divided into two categories: (1) viral-based systems, and (2) non-viral-based systems. Delivery of interesting genes with viral vectors has proven to the most efficient method. However, potential toxicity, including mutagenesis and carcinogenesis, from the viral vectors presents a major limitation. In contrast, delivery of genes with plasmid DNA (pDNA) may be much safer, and perhaps more pharmaceutically acceptable. However, the efficiency of current non-viral-based systems is very poor (Rolland and Felgner, 1998; Liu and Huang, 2002). Moreover, due to the presence of serum nucleases in the blood and the rapid clearance rate of pDNA, improved formulations and delivery systems for pDNA are needed for successful systemic gene therapy (Kawabata et al., 1995; Mahato et al., 1995a; Yoshida et al., 1997; Houk et al., 2001). * Corresponding author: Russell J. Mumper, Ph.D., Assistant Professor of Pharmaceutical Sciences, Assistant Director, Center for Pharmaceutical Science and Technology, College of Pharmacy, University of Kentucky, Lexington, KY 40536-0082. Tel: (859) 257-2300 ext. 258, Fax: (859) 323-5985, e-mail: [email protected].

To date, many pDNA delivery systems, including lipoplexes (Clark and Hersh, 1999; Ishiwata et al., 2000), polyplexes (Nishikawa et al., 2000a and 2000b; Oh et al., 2001; Verbaan et al., 2001), and lipopolyplexes (Liu and Huang, 2002), have been developed. However, in general, these systems remain very sensitive to serum challenge. Moreover, these systems tend to aggregate due to extensive interaction with serum proteins and/or self-association in the blood (Chesnoy and Huang, 2000). Therefore, when administered by intravenous injection, these systems tend to accumulate into the lung. Significant work has been done to improve these systems and overcome the stated limitations. For example, for lipoplexes, complexes of cationic lipids or liposomes with pDNA, careful balance of the hydrophilicity/lipophilicity of the surface of the lipoplex particle, modification of the surface charge, and/or deposition of “shielding” agents such as poly(ethylene glycol) (PEG) on the surface of the lipoplexes, has proven to be able to extend the circulation of the complexes and therefore lead to increased distribution of pDNA to distal sites of interest (Ishiwata et al., 2000; Patel, 1992; Hong et al., 2002). Despite some promising leads using the complexes described above, delivery systems that protect pDNA from serum nuclease digestion, provide stability for

10.1021/bc0255586 CCC: $22.00 © 2002 American Chemical Society Published on Web 10/12/2002

1320 Bioconjugate Chem., Vol. 13, No. 6, 2002

prolonged circulation in blood, and that can be easily modified for potential cell or tissue targeting are still needed. Recently, researchers have attempted to encapsulate or entrap pDNA inside stable lipid-based particles for potential systemic administration. Wheeler et al. described a method to prepare stabilized plasmid-lipid particles (SPLP) (around 70 nm in diameter) using a detergent dialysis procedure (Wheeler et al., 1999). The particles allowed for the encapsulation of pDNA within a lipid envelope, which in turn protected the plasmid from digestion by serum nucleases. The particles were further stabilized in aqueous media by PEG coating. The particles demonstrated extended circulation with approximately 10% of the intravenously injected dose accumulating in a distal tumor. Further, Bailey and Sullivan have described the encapsulation of pDNA in neutral liposomes (Bailey and Sullivan, 2000). Lipoproteins, or emulsion particles that consist of lipids and apolipoproteins, have also been described by Hara et al. (1997b). This group took advantage of the hydrophobic interior of these natural emulsions to solubilize DNA. Negatively charged DNA was first complexed with cationic lipids containing a quaternary amine headgroup. The resulting hydrophobized complex was extracted by chloroform and then incorporated into reconstituted chylomicron remnant particles (∼100 nm in diameter) with an efficiency of ∼65%. When injected into the portal vein of mice, the level of transgene expression in the liver was ∼100-fold greater than that of mice injected with ‘naked’ pDNA. Histochemical examination revealed that a large number of parenchymal cells and other types of cells in the liver expressed the transgene. Unfortunately, very poor expression was found in the liver after tail vein injection. In addition, other pDNA encapsulated cationic emulsions have been reported (Hara et al., 1997a). For example, Hara et al. formed a hydrophobic complex of pDNA with DC-Chol (3β[N-(N′,N′-dimethylaminoethane)carbamoyl]cholesterol) and then encapsulated the complex into the oil core of emulsion prepared with castor oil, Tween 80, phosphatidylcholine, and DOPE (dioleoylphosphatidylethanolamine). More recently, Chesnoy et al. described stabilization of pDNA encapsulated cationic emulsions by coating with PEG (Chesnoy et al., 2001). Compared to the unmodified emulsions, stabilization with PEG (910%, w/w) significantly decreased the plasma clearance and liver uptake of the emulsion. Our laboratories have likewise been interested in the development of improved nanoparticle-based pDNA delivery systems. Previously, we reported on a novel method to engineer stable cationic nanoparticles (∼100 nm) from oil-in-water (O/W) microemulsion precursors, comprised of emulsifying wax as the oil phase and CATB (cetyltrimethylammonium bromide) as the cationic surfactant (Cui and Mumper, 2002a and 2002b). Using this method, uniform nanoparticles were engineered within minutes in a single vessel without the use of organic solvents or the application of high torque mechanical mixing. Plasmid DNA was then coated on the surface of these preformed cationic nanoparticles and administered to mice resulting in significant enhancements in humoral, Th1-type, and proliferative immune responses over ‘naked’ pDNA after both topical and subcutaneous administration to mice. In these present studies, we sought to further explore this novel nanoparticle engineering method by entrapping pDNA inside the nanoparticles. These pDNAentrapped nanoparticles were characterized and evaluated both in vitro and in vivo.

Cui and Mumper EXPERIMENTAL PROCEDURES

Reagents. Plasmid containing a CMV promoter luciferase gene (pCMV-luc, 4.7 kp) was a gift from Valentis, Inc. (The Woodlands, TX). I-125 labeled plasmid pBR332 was purchased from the Lofstrand Labs, Ltd. (Gaithersburg, MD). Emulsifying wax, Tween 20, and Tween 80 were from Spectrum Chemical (New Brunswick, NJ). Brij 78 was a free sample from Uniqema (Wilmington, DE). Dimethyldioctadecylammonium bromide (DDAB), diethylaminoethyl (DEAE) Sepharose CL-6B anion exchanger, and mouse serum were purchased from Sigma-Aldrich Corporation (St. Louis, MO). 1.2-Dioleoyl-3-trimethylammoniopropane (DOTAP) and dioleoyl phosphotidylethanolamine (DOPE) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). Sepharose CL4B was purchased from Amersham Pharmacia Biotech, Inc. (Piscataway, NJ). Luciferase assay kits were purchased from Promega Corporation (Madison, WI). {N-[2(Cholesterylcarboxyamino)ethyl]carbamorylmethyl}pullulan (Chol-pullulan) was purchased from Dojindo Molecular Technologies, Inc. (Gaithersburg, MD). Lipofectin reagent was from Gibco BRL, Life Technologies (Gaithersburg, MD). Bligh and Dyer Extraction. Bligh and Dyer extraction was performed as described by Reimer et al. with slight modification (Reimer et al., 1995). Briefly, DDAB or DOTAP and pDNA were solubilized in the monophase of chloroform/methanol/water (1:2.1:1 v/v/v). The pDNA concentration was fixed at 10 µg/mL while the concentrations of the lipids were varied so that the charge ratios of the cationic lipid to the pDNA were ranged from 0:1 to 2:1 ((). The mixtures were incubated at room temperature for 30 min in a total volume of 1 mL and then were partitioned into two phases by the addition of water and chloroform (250 µL each). The two phases were briefly mixed by vortexing. After a further incubation of 15 min, the lower organic phase and the upper aqueous phase were separated by centrifugation at 14000 rpm at room temperature for 5 min. The aqueous phase was removed and diluted with water, if necessary, for determination of the pDNA concentration using a F-2000 Fluorescence Spectrophotometer (Hitachi Instruments, Danbury, CT) after staining with PicoGreen (Molecular Probes, Eugene, OR). DDAB/pDNA and DOTAP/pDNA complexes with charge ratio of 1:1 (() were used to investigate the dissociation of these complexes by NaCl. Briefly, cationic lipid/pDNA complexes (1:1 () were prepared in Bligh and Dyer monophases. The required amount of NaCl (2 M in water) was then added into the monophase system. After 15 min of incubation at room temperature, Bligh and Dyer extractions and pDNA concentration determination were performed as described above. Engineering of pDNA-Entrapped Nanoparticles. Nanoparticles were cured from warm microemulsion precursors engineered as previously described with modifications (Cui and Mumper, 2002a and 2002b). Cationic lipid/pDNA complexes (1:1 (; pDNA concentration of 10 or 30 µg) were prepared in the monophase as described above. After Bligh and Dyer extraction, the lower organic phase was collected. Consequently, depending on the amount of pDNA formulated, either 2 or 6 mg of emulsifying wax was dissolved into this organic phase. The organic phase was then evaporated at room temperature to obtain a dried mixture of emulsifying wax and cationic lipid/pDNA complexes in glass scintillation vials. After melting of the mixture at 50-55 °C, a required volume of deionized and filtered (0.22 µm) water was

Plasmid DNA-Entrapped Nanoparticles

added to the vial. The suspension was vigorously stirred at 50-55 °C until a homogeneous milky slurry was formed. Then, required volumes of Brij 78 (100 mM in water), Tween 20 (1/4, w/w in water), and/or Tween 80 (1/4, w/w in water) were added while stirring. Stirring was maintained until a clear transparent microemulsion system was formed (within 3-5 min). The microemulsions, if formed, were then cooled to room temperature while stirring to cure nanoparticles. DOPE (5% w/w) was incorporated in the nanoparticles as previously described (Cui and Mumper, 2002a and 2002b). For comparison and as a control, empty cationic nanoparticles with no pDNA were also prepared as previously described while maintaining the identical concentrations of all components. Plasmid DNA was then added to the preformed suspension for adsorption on the surface of the nanoparticles. Physical Characterization of Nanoparticles. The nanoparticle size was measured by photon correlation spectroscopy (PCS) using a Coulter N4 Plus Submicron Particle Sizer (Coulter, Miami, FL) by scattering light at 90° at 25 °C for 120 s for each 1 mL sample. The zeta potentials of the nanoparticles were measured using a Zeta Sizer 2000 from Malvern Instruments (Southborough, MA). Gel Permeation Chromatography. Nanoparticle purification and confirmation of pDNA entrapment were completed using gel permeation chromatography (GPC). For all GPC, 100 µL samples were applied to either Sepharose CL-4B or DEAE Sepharose CL-6B columns (10 × 55 mm). The samples were eluted with 5 mM HEPES in 150 mM NaCl pH 7.4 (HBS), and the fractions (0.5 mL) were assessed for pDNA concentration using either a gamma counter (for I-125 labeled pDNA) or PicoGreen fluorescence assay (for unlabeled pDNA). If the samples applied were nanoparticle suspension, the counts per second (CPS, an indication of nanoparticle concentration) of all eluted fractions were also measured with Coulter N4 Plus Submicron Particle Sizer. Serum Stability Assays. The in vitro stability of the pDNA in the presence of mouse serum nucleases was investigated by mixing 50 µL of either pCMV-luc or nanoparticle suspensions (prepared with pCMV-luc and trace amount of I-125 labeled pBR332) with 150 µL of mouse serum (Sigma-Aldrich). The mixture was then allowed to remain at room temperature for 8 h for enzymatic digestion of the plasmid. One hundred microliters (100 µL) of the samples, before and after serum digestion, were applied to Sepharose CL-4B column as mentioned above to confirm the protection of pDNA from serum nucleases after entrapment in the nanoparticles. Intact pDNA eluded out in the void volume while digested pDNA was fractionated. Stability of the Entrapped pDNA Extracted from the Nanoparticles. To investigate the stability of the pDNA after entrapment, pDNA was extracted from purified nanoparticles and was applied on a 1% Seakem Gold Agarose Ready-to-use Gel from BioWhittaker Molecular Applications (Rockland, ME) for electrophoresis. Specifically, nanoparticles containing DOTAP complexed pDNA (30 µg/mL) were prepared (see nanoparticle Composition II). The nanoparticles were purified with DEAE Sepharose CL-6B column and reconcentrated using ultrafiltration (100 kDa molecular weight cutoff, Brinkmann Instruments, Inc., Westbury, NY). The purified and concentrated nanoparticle suspensions were air-dried in a 37 °C incubator to form a dry nanoparticle powder. A combination of 200 µL of chloroform, 200 µL of NaCl solution (150 mM), and 420 µL of methanol were added

Bioconjugate Chem., Vol. 13, No. 6, 2002 1321

to the nanoparticle dry powder. This mixed Bligh & Dyer system was then gently stirred for 30 min on hot plate maintained at 50-55 °C. Plasmid DNA was finally partitioned with water and chloroform (100 µL each) into the upper aqueous phase as mentioned above. The pDNA in the aqueous phase was then used for gel electrophoresis analysis. For comparison, pDNA was also extracted from a physical mixture of emulsifying wax with pDNA/ DOTAP complexes. The extraction procedure was the same as that used for the pDNA-entrapped nanoparticles. Chol-Pullulan Coated Nanoparticles. Chol-pullulan was deposited on the surface of the nanoparticles using a previously described method (Cui and Mumper, 2002a and 2002b). Briefly, 250 µg of chol-pullulan (suspension in water, 5 mg/mL) was added into the preformed warm microemulsions. After being stirred at 50-55 °C for 5 min, the microemulsions were cooled to room temperature while stirring to form nanoparticles. Overnight stirring at room temperature was continued for complete deposition of pullulan on the surface of the nanoparticles. Free surfactants and chol-pullulan were removed by GPC using a DEAE Sepharose CL-6B column. Stability of the Nanoparticles in Simulated Biological Media. The stability of the nanoparticles, coated or uncoated with pullulan, in either normal saline (0.9% w/v NaCl) or 10% (v/v) fetal bovine serum (FBS) in normal saline was monitored as previously described by diluting the nanoparticle suspensions in these media and measuring the particle size over 30 min at 37 °C. In Vitro Hep G2 Cell Transfection. In vitro cell transfection studies were completed as previously described (Cui and Mumper, 2002a and 2002b). Hep G2 cells were obtained from American Type Culture Collection (ATCC, Rockville, MD) and were maintained in Eagle’s Minimum Essential Medium (EMEM) (Gibco, BRL) containing 10% fetal bovine serum (Gibco, BRL) and 1% penicillin-streptomycin (Gibco, BRL). Transfections were performed on cells that were approximately 80% confluent. Cells were plated in 48 well plates at a density of 5 × 105 cells/well and incubated overnight. The cells were coincubated with the formulations to provide a final plasmid (pCMV-luc) concentration of 2 µg/well. After 4 h of coincubation, the medium was replaced with fresh EMEM medium. Cells were harvested 48 h later by removing the medium, washing with 1X PBS buffer (three times), and then adding 200 µL 1 × lysis buffer (Promega), leaving for 5-10 min and freeze-thawing three times. A 20 µL sample was assayed for luciferase activity by injecting 100 µL reconstituted luciferase assay solution (Promega) and measuring the light produced for a period of 10 s using a ML2250 Dynatech Luminometer (Dynatech Laboratories, Chantilly, CA). The protein content of a 20 µL sample of the supernatant was determined using the Coomassie Plus Protein Assay Reagent (Pierce). In one treatment, the Hep G2 cells were preincubated with 100 µg of chol-pullulan for 5 min before the addition of the pullulan-coated pDNA-entrapped nanoparticles. As recommended by Dynatech, luciferase expression data were reported as the ratio of the full integral of the samples to that of the negative control, divided by the total amount of protein in the 20 µL of samples assayed. Biodistribution of Nanoparticles after Tail-Vein Injection to Mice. Ten to twelve week old female Balb/C mice (average weight, 18 g) from Harlan SpragueDawley Laboratories were used for the animal studies. NIH guidelines for the care and use of laboratory animal were observed. The mice were anesthetized and then

1322 Bioconjugate Chem., Vol. 13, No. 6, 2002

Cui and Mumper

Figure 2. Particle size and zeta potential nanoparticles engineered from microemulsion precursors. DDAB ) Composition I (pDNA-entrapped nanoparticles); DDAB/Ad ) pDNA added to preformed nanoparticles having Composition I; DOTAP ) Composition II (pDNA-entrapped nanoparticles); DOTAP/Ad ) pDNA added to preformed nanoparticles having Composition II. (b) represents zeta potential measurements. All data points are shown as the mean ( SD (n ) 3).

Figure 1. (A) The effect of increasing amounts of either DDAB (9) or DOTAP (0) on the recovery of pDNA in the upper aqueous phase following Bligh and Dyer extraction of the lipid/pDNA complexes (see Materials and Methods). (B) The effect of increasing amounts of NaCl on the recovery of pDNA from the aqueous phase following Bligh and Dyer extraction. The pDNA used was 10 µg. The charge ratio for the lipid vs pDNA was 1:1 ((). All data points are shown as the mean ( SD (n ) 3).

injected via the tail vein with 200 µL (2 µg pDNA) of either pDNA alone in normal saline (n ) 5), pDNAentrapped nanoparticles (n ) 5), and pullulan-coated pDNA-entrapped nanoparticles coated with pullulan (n ) 7). The nanoparticles were purified using normal saline as the mobile phase before injection. In all formulations, a trace amount of I-125 labeled pDNA was included. Thirty minutes after tail-vein injection, blood (400 µL) was collected by cardiac puncture. The liver, lungs, heart, spleen, kidneys, and tail of the mice were also collected. Radioactivity in each sample was determined using a Cobra II Auto Gamma Counter from Packard BioScience Company (Meriden, CT). The total blood volume of a mouse was assumed to be 7.5% (v/w) of the total mouse weight (Mosqueira et al., 2001). RESULTS AND DISCUSSION

Hydrophobizing Plasmid DNA. Cationic lipids have previously been used to complex pDNA by electrostatic interaction and thereby increase the lipophilicity of pDNA (Reimer et al., 1995). As expected, if pDNA alone was added into the Bligh and Dyer monophase system, almost all of the pDNA was recovered in the aqueous phase (Figure 1A). However, by complexing the pDNA with an increasing amount of either DDAB or DOTAP, a correspondingly decreasing amount of pDNA was recovered in the aqueous phase. In fact, starting from a charge ratio of 0.5:1 ((), no detectable pDNA was

recovered from the aqueous phase as determined using the PicoGreen fluorescence DNA quantification assay. These results demonstrated that both DDAB and DOTAP can be used to hydrophobize pDNA. To confirm that the complex formation was due to electrostatic interaction, and to determine the concentration of NaCl required to disrupt the formation of cationic lipids/pDNA complexes in Bligh and Dyer monophase system, an increasing amount of NaCl was added into the monophase system formed with either DDAB or DOTAP at a charge ratio of 1:1 (() before extraction. As shown in Figure 1B, the addition of NaCl disrupted the cationic lipids/pDNA complexes and prevented the partition of pDNA into the lower organic phase. At a molar ratio of NaCl to cationic lipid of 8:1, binding of pDNA to both cationic lipids, was completely prevented, leading to almost no pDNA partitioning into the organic phase. This effective NaCl concentration was estimated to be only ∼8 mM. Preparation and Characterization of pDNAEntrapped Nanoparticles. In formulation optimization studies, a matrix of various excipient combinations were tested to identify those systems that formed a clear and transparent microemulsion at 50-55 °C. After varying the concentrations of surfactants including Brij 78, Tween 20, and Tween 80, two different nanoparticle compositions were selected for further studies by monitoring the clarity of the microemulsions and measuring the size and uniformity of the cured nanoparticles. The two nanoparticle compositions were Composition I: 1:1 charge ratio (() of either DDAB or DOTAP with pDNA (final pDNA concentration: 10 µg/mL), emulsifying wax (2 mg/mL), Brij 78 (5 mM), and Tween 20 (15 mM); and Composition II: 1:1 charge (() ratio of either DDAB or DOTAP with pDNA (final pDNA concentration: 30 µg/ mL), emulsifying wax (6 mg/mL), Brij 78 (14 mM), and Tween 20 (22 mM), and Tween 80 (2 mM). Shown in Figure 2 are the particle sizes and zeta potentials of two representative nanoparticles. The DDAB containing nanoparticles (Composition I) had particle size and zeta potential of 159 ( 4 nm and -13.5 ( 4 mV, respectively. The DOTAP-containing nanoparticles (Composition II) had particle size and zeta potential of 99 ( 7 nm and -7.9 ( 0.5 mV, respectively. Figure 3 shows the distribution of the nanoparticles obtained from the

Plasmid DNA-Entrapped Nanoparticles

Bioconjugate Chem., Vol. 13, No. 6, 2002 1323

Figure 3. The particle size distribution of pDNA-entrapped nanoparticles (Composition I and II).

Figure 4. Gel permeation chromatograph (GPC) of pDNA with Sepharose CL-4B (0) or DEAE Sepharose CL-6B (9). A trace amount of I-125 labeled pDNA was included in the formulations. Columns were equilibrated and eluted with 5 mM HEPES in 150 mM NaCl pH 7.4 (HBS).

particle sizer. The very narrow distribution demonstrated that the nanoparticles were uniform. Moreover, the nanoparticles were spherical as demonstrated by TEM (data not shown). Also shown in Figure 2 are nanoparticles (Composition I and II) as mentioned above except that the pDNA was added into preformed cationic nanoparticle suspensions for adsorption (“Ad”). Both particle size and zeta potential of these pDNA-coated nanoparticles were significantly different from that of the pDNA-entrapped nanoparticles. The particle size data showed that the entrapment of pDNA in nanoparticles resulted in larger nanoparticles, perhaps due to issues relating to packing of the lipid/ pDNA complexes in the nanoparticles. Confirmation of Plasmid DNA Entrapment in the Nanoparticles. The following studies were carried out to further confirm the entrapment of pDNA in the nanoparticles and to estimate the entrapment efficiency. Using a Sepharose CL-4B column, pDNA eluted out mostly in fraction number 3 (void volume) demonstrated by the highest CPM in this fraction (Figure 4). However, as expected, the positively charged DEAE Sepharose CL6B column efficiently bound most of the pDNA (>90%) (Figure 4). In contrast, after entrapment of the same pDNA into the nanoparticles, DEAE Sepharose CL-6B column was no long able to retard most of the pDNA from eluting out for both the Composition I nanoparticles (Figure 5A) and the Composition II nanoparticles (Figure 5B). Moreover, the nanoparticle eluent profiles, estimated by the counts per second (CPS), overlapped with the CPM strongly suggesting the association of pDNA with the nanoparticles. The presence of pDNA inside the nanoparticles was further confirmed by an in vitro mouse serum digestion

Figure 5. Gel permeation chromatograph of pDNA-entrapped nanoparticles. A trace amount of I-125 labeled pDNA was included in the formulations. One hundred microliters of nanoparticle suspensions was passed through the DEAE Sepharose CL-6B column. Eluent fractions (0.5 mL) were collected and measured in a gamma counter and particle sizer to determine radioactivity (counts per minute, CPM, )) and particle intensity (counts per sec, CPS, 9), respectively. (A) pDNA-entrapped nanoparticles (Composition I). (B) pDNA-entrapped nanoparticles (Composition II).

assay. As shown in Figure 6A, after 8 h of coincubation of the pDNA alone with mouse serum, all ‘naked’ pDNA was digested into small oligo- or mononucleotides as demonstrated by the shifted eluent profiles when applied to Sepharose CL-4B column. The same pDNA, when entrapped in the nanoparticles (Composition I and II), was not completely digested by the mouse serum nucleases (Figure 6B and 6C) demonstrating that the nanoparticles provided some protection against degradation. Serum digested plasmids were likely located in the bulk aqueous phase of the suspension or loosely associated on the surface of the nanoparticles. It was reported that cationic lipids, such as DDAB, may complex pDNA without condensing it. Therefore, it is likely that pDNA in these complexes were still susceptible to enzymatic degradation (Reimer et al., 1995). The entrapment efficiencies of pDNA in the nanoparticles were determined by area under curve (AUC) analysis using the Linear Trapezoidal Method. The entrapment efficiencies for the DDAB/DNA-entrapped nanoparticles (Composition I; Figure 6B) and the DOTAP/DNA-entrapped nanoparticles (Composition II; Figure 6C) were estimated to be 24.5 ( 3.4% and 29.4 ( 2.1%, respectively. As shown in Figure 7, entrapped pDNA remained stable after extraction from the nanoparticles. The entrapment efficiencies in nanoparticles were lower than expected based on the Bligh and Dyer extraction experiments wherein all

1324 Bioconjugate Chem., Vol. 13, No. 6, 2002

Cui and Mumper

Figure 7. Agarose gel electrophoresis of pDNA extracted from pDNA-entrapped nanoparticles (see Materials and Methods for the extraction procedure). Lane 1 and 5, Gibco BRL 1kb DNA ladder; lane 2, plasmid CMV-luc as a positive control; lane 3, CMV-luc extracted from the mixture of pDNA/DOTAP complexes and emulsifying wax; lane 4, CMV-luc extracted from pDNA-entrapped nanoparticles. O.C. indicates open circular plasmid; S.C. indicates supercoiled plasmid.

Figure 6. Gel permeation chromatograph of pDNA-entrapped nanoparticles with Sepharose CL-4B column for ‘naked’ pDNA alone (A) and pDNA-entrapped nanoparticles (B, C). A trace amount of I-125 labeled pDNA was included in the plasmids. One hundred microliters of nanoparticle suspensions, before (9) or after (0) mouse serum digestion, was passed through the column. Eluent fractions (0.5 mL) were collected and measured in a gamma counter to determine plasmid concentration (counts per min, CPM). (B) pDNA-entrapped nanoparticles (Composition I). (C) pDNA-entrapped nanoparticles (Composition II). In the case of ‘naked’ pDNA alone after digestion, the pDNA concentration was determined using PicoGreen and was reported as fluorescence intensity (A).

plasmid DNA was extracted into the organic (chloroform) phase when complexed with cationic lipid at a charge ratio of at least 0.5:1 ((). It is likely that the cationic lipid/pDNA complexes had lower affinity for the melted emulsifying wax oil phase in the microemulsion precursor as compared to the Bligh and Dyer organic (chloroform) phase. The reduced affinity of the complex in the melted

oil phase could therefore explain the lower than expected entrapment efficiencies of pDNA in the cured nanoparticles. In Vitro Hep G2 Cell Transfection. Pullulan has been successfully used as a ligand for liver hepatocyte targeting (Cui and Mumper, 2002a; Tabata et al., 1999; Xi et al., 1996; Kaneo et al., 2001). In vitro cell transfections studies were completed to further assess the integrity and functionality of the entrapped pDNA and to investigate whether pullulan, coated on the nanoparticle surface, could enhance nanoparticle uptake and transgene expression. After coating with pullulan, the size of the nanoparticles (DDAB, composition I) was significantly increased to 254 ( 72 nm (p ) 0.01, twosample t-test). The zeta potential of the pullulan-coated nanoparticles, (-12.4 ( 5 mV) was comparable to that of uncoated nanoparticles (-13.5 ( 4 mV) uncoated nanoparticles. As shown in Figure 8, pDNA alone or pDNA entrapped into nanoparticles did not result in measurable transfection. Incorporation of DOPE as an endosomolytic lipid enhanced the transfection by about 4-fold. Moreover, after deposition of pullulan on the surface of these pDNA-entrapped nanoparticles (DOPEfree), significantly enhanced transfection was observed. Preincubation of the Hep G2 cells with free chol-pullulan significantly reduced the transfection enhancement (Figure 8), suggesting that the uptake of the pullulan-coated pDNA-entrapped nanoparticles may be via a receptor mediated endocytic process. These results were comparable to our previously reported in vitro transfection results using pDNA-coated cationic nanoparticles with pullulan (Cui and Mumper, 2002a and 2002b). In these previous studies, coating of pDNA on the surface of the cationic nanoparticles significantly enhanced in vitro cell transfection, compared to that of ‘naked’ pDNA. Further, the incorporation of DOPE in the pDNA-coated nanoparticles enhanced the transfection efficiency by over 5-fold. The coating of nanoparticles with pullulan increased the transfection ability of the pDNA-coated nanoparticles by over 40-fold to levels that were comparable to that of the Lipofectin. A similar trend was observed in these present studies; however, the overall magnitude of expression compared to Lipofectin using pDNA-entrapped nanoparticles was significantly lower. The lower relative levels of expression with pDNAentrapped nanoparticles in these studies may have been

Plasmid DNA-Entrapped Nanoparticles

Figure 8. In vitro transfection of liver Hep G2 cells (500000 cells) (n ) 3) in the presence of 10% FBS after 48 h with pDNA alone (pDNA), pDNA-Lipofectin (Lipof), uncoated pDNAentrapped nanoparticles (NP), uncoated pDNA-entrapped nanoparticles with DOPE (NP-DOPE), or pullulan-coated pDNAentrapped nanoparticles (Pull-Entp). For the group of PullEntp*, Hep G2 cells were preincubated with free chol-pull 5 min before the addition of the Pull-NPs. The pDNA dose was 2 µg for all samples. A one-way ANOVA followed by pairwise comparison with Fisher’s protected least significant difference procedure (PLSD) was completed to analyze the data. *Indicates that the result of the Lipofectin was significantly different (p < 0.05) from that of the others. **Indicates that the result for NPDOPE was significantly different (p < 0.05) from that of NP and pDNA alone. ***Indicates that the result for Pull-NP was significantly different (p < 0.05) from that of all other groups, except for NP-DOPE. NP ) Composition II.

due to (1) the nanoparticles used in the present studies were net negatively charged while those in our previous studies were net positively charged even after pDNA coating, (2) the use of different nanoparticle compositions, and/or (3) issues relating to the kinetics of release of the entrapped pDNA from the nanoparticles. These factors are currently being assessed in ongoing experiments in our laboratories. Biodistribution of the pDNA-Entrapped Nanoparticles in Mice. A short-term stability study of the nanoparticles in stimulated biological media was completed prior to the in vivo biodistribution studies to predict the nanoparticle stability. As shown in Figure 9, the size of pullulan-coated or uncoated nanoparticles did not change after 30 min of incubation at 37 °C in either normal saline (0.9% NaCl w/v) or 10% (v/v) of fetal bovine serum (FBS) in normal saline. Thirty minutes after tail vein injection, 47.6 ( 7.5%, 77.3 ( 10.4%, and 69.3 ( 3.8% of the total radioactivity administered was recovered from the mice injected with ‘naked’ pDNA alone, pDNA-entrapped nanoparticles, and pullulan-coated pDNA-entrapped nanoparticles, respectively. The majority of the recovered radioactivity was located in the blood and liver (Figure 10). Compared to only 16% for the ‘naked’ pDNA, 35-40% of the nanoparticle-entrapped pDNA were still circulating in the blood 30 min after tailvein administration. The retention of pDNA in the blood at 30 min using both nanoparticle formulations was statistically significant over ‘naked’ pDNA alone. This suggests that the nanoparticles may have the potential for prolonged circulation and avoidance of the RES, problems commonly observed with other plasmid DNA delivery systems. Plasmid DNA, both free and entrapped, was also found in lung, heart, spleen, and kidney of the mice although in relatively low levels. The absence of a high level of radioactivity in the lung with both types of pDNA-entrapped nanoparticles strongly suggested that

Bioconjugate Chem., Vol. 13, No. 6, 2002 1325

Figure 9. The particle size of the uncoated nanoparticles (NPs) and pullulan-coated nanoparticles (Pull-NPs) in various media at 37 °C for 30 min. For particle size measurement, the nanoparticle preparations were diluted with either 0.9% (w/v) of NaCl (normal saline) or 10% (v/v) of FBS in normal saline. Statistical analyses (two-sample t-test assuming equal variances) demonstrated no difference between the particle size at 0 and 30 min for all groups.

Figure 10. In vivo biodistribution of ‘naked’ pDNA alone (black bars, n ) 5), pDNA-entrapped nanoparticles (white bars, n ) 5), and pullulan-coated pDNA-entrapped nanoparticles (gray bars, n ) 7) 30 min after tail vein injection to Balb/C mice. Statistical analyses were completed using a one-way ANOVA followed by pairwise comparison with Fisher’s protected least significant difference procedure (PLSD). *Indicates that, in blood, the results for nanoparticles are significantly different from that for the free ‘naked’ pDNA alone.

these nanoparticles did not aggregate in blood resulting in rapid lung accumulation, an event that is common for cationic lipid complexes with pDNA (Mahato et al., 1995a and 1995b; Bailey and Sullivan, 2000; Takakura et al., 2001). The biodistribution results for both free ‘naked’ pDNA and pDNA-entrapped nanoparticles agreed both qualitatively and quantitatively to those of Bailey and Sullivan (2000). Bailey and Sullivan (2000) entrapped pDNA in neutral liposomes (∼200 nm) with the aid of ethanol and calcium. Biodistribution studies with these pDNA-entrapped liposomes showed about 32-37% of the total dose remaining in the blood 1 h following tail vein injection to mice, with about 15-20% remaining in the liver. Interestingly, the total radioactivity recovered with the pDNA-entrapped liposomes (∼60%) agreed well with that for pDNA-entrapped nanoparticles in our studies. Other organs/tissues that were not collected in either study that may account for unrecovered radioactivity include thyroid, brain, urine, extravascular tissue, and/

1326 Bioconjugate Chem., Vol. 13, No. 6, 2002

or intestines. Blood retention after tail-vein injection using these pDNA-entrapped nanoparticles compare very favorably to published reports of other systems. For example, Chesonyl et al. (2001) showed that 30 min after iv injection, 14-22% of the injected pDNA/emulsion complexes stabilized by 9-10% of PEG remained in mouse plasma. Further, Minchin et al. (2001) reported that 20 min after intraarterial injection, about 30% of the injected pDNA complexed with polyinosinic acid and polycationic liposomes was found in the circulation. Finally, Tam et al. (2000) showed that 4 h after iv injection, 55% of the injected plasmid-lipid particles (SPLP) were remained in the plasma. It was also found in these present studies that the pullulan coating did not statistically enhance liver accumulation of pDNA entrapped in nanoparticles. The pullulan-coated nanoparticles were larger than the uncoated nanoparticles, thus it is possible that the larger pullulan-coated nanoparticles had reduced penetration through the sinusoidal endothelial barrier of the liver. Other possible reasons include (1) relatively low binding affinity of the pullulan ligand, (2) suboptimal concentration of the pullulan ligand on the nanoparticles. To this end, we are currently assessing the application of alternative hepatocyte-specific ligands and the optimization of the pullulan-coating on the nanoparticle surface. Nonviral gene therapy involving the delivery of plasmid DNA has gained considerable attention over the past decade. However, to administer pDNA by intravenous injection, improved delivery systems are needed that can (1) protect DNA from serum nuclease digestion, (2) circulate in the blood long enough to reach distal organs and tissues, and (3) be readily modified for cell- and tissue-specific targeting. In addition to delivery strategies employing microparticles, liposomes, or cationic condensing agents such as lipids, polymer, and peptides/proteins, researchers have attempted to hydrophobize DNA and then incorporate the hydrophobized DNA into micron or nanometer sized particles. Examples include the stabilized plasmid-lipid particles (SPLP) prepared using a detergent dialysis procedure by Wheeler et al. (1999), reconstituted chylomicron remnant particles reported by Hara et al. (1997a), and more recently the poly(ethylene glycol)-stabilized DNA/emulsion systems based on the reconstituted chylomicron remnant particles described by Chesnoy et al. (2001). The pDNA-entrapped nanoparticles reported in these present studies demonstrated prolonged circulation in blood (Figure 10). Although the mechanism(s) need further investigation, several possible explanations may be mentioned. The pDNA-entrapped nanoparticles were slightly net negatively charged. Therefore, possible association of negatively charged lipoproteins present in the blood with the nanoparticle surface may have been reduced and/or prevented (Nishikawa et al., 1998). Second, surfactants such as Brij 78, Tween 20, and Tween 80 may provide a more hydrophilic (‘stealth’-like) surface on the nanoparticles, thereby precluding the aggregation of the nanoparticles by self-association or the rapid elimination by the RES (Lundberg et al., 1996). Similarly, the particle size of the nanoparticles was relatively small, perhaps reducing elimination of the nanoparticles by the mononuclear phagocytes in the liver (Kupfer cells) (Sakadea and Hirano, 1998). It was demonstrated in this current study that pDNAentrapped nanoparticles having diameters from 100 to 160 nm could be engineered from oil-in-water (O/W) microemulsion precursors formed at increased temperatures (i.e., 40-55 °C) and then simply cooled to form

Cui and Mumper

nanoparticles. The engineering process used in the present studies for pDNA-entrapped nanoparticle preparation has the following advantages: (i) the nanoparticle engineering process is rapid and spontaneous and completed in a single vessel, (ii) stable pDNA-entrapped nanoparticles (∼150 nm) can be engineered within minutes, and (iii) cell-specific ligands can be incorporated into the nanoparticles during the engineering process. Nonionic emulsifying wax (comprised of cetyl alcohol and polysorbate 60 in a molar ratio of about 20:1) was selected as the oil phase since it is generally regarded as safe (GRAS) and has a melting point of ∼50-55 °C. The wax is typically used in cosmetics and topical pharmaceutical formulations and is generally regarded as a nontoxic and nonirritant material. For example, cetyl alcohol is currently used as an excipient in the marketed product Exosurf Neonatal. In addition, polysorbate 60 is used in many pharmaceutical products including parenteral products. In these present and previous studies using nanoparticles with similar compositions, no gross inflammatory, allergic, or toxic affects have been observed in mice after administration of these nanoparticles by the intravenous, subcutaneous, intramuscular, and topical routes (Cui and Mumper, 2002a and 2002b). ACKNOWLEDGMENT

This work was supported, in part, by NSF grant BES9986441, and from the AFPE and the Burroughs Wellcome Fund through the AACP New Investigators Program for Pharmacy Faculty. LITERATURE CITED (1) Bailey, A. L., and Sullivan, S. M. (2000) Efficient encapsulation of DNA plasmids in small neutral liposomes induced by ethanol and calcium. Biochim. Biophys. Acta 1468, 239-252. (2) Chesnoy, S., Durand, D., Doucet, J., Stolz, D. B., and Huang, L. (2001) Improved DNA/emulsion complex stabilized by poly(ethylene glycol) conjugated phospholipids. Pharm. Res. 18, 1480-1484. (3) Chesnoy, S., and Huang, L. (2000) Structure and function of lipid-DNA complexes for gene delivery. Annu. Rev. Biophys. Biomol. Struct. 29, 27-47. (4) Clark, M., and Hersh, E. M. (1999) Cationic lipid-mediated gene transfer: current concepts. Curr. Opin. Mol. Ther. 1, 158-76. (5) Cui, Z. R., and Mumper, R. J. (2002a) Genetic immunization using nanoparticles engineered from microemulsion precursors. Pharm. Res. 19, 939-946. (6) Cui, Z. R., and Mumper, R. J. (2002b) Topical immunization using nanoengineered genetic vaccines. J. Controlled Release 81, 173-184. (7) Hara, T., Tan, Y., and Huang, L. (1997a) In vivo gene delivery to the liver using reconstituted chylomicron remnants as a novel nonviral vector. Proc. Natl. Acad. Sci. U.S.A. 94, 14547-14552. (8) Hara, T., Liu, F., Liu, D., and Huang, L. (1997b) Emulsion formulation as a vector for gene delivery in vitro and in vivo. Adv. Drug Del. Rev. 24, 265-271. (9) Hong, M. S., Lim, S. J., Oh, Y. K., and Kim, C. K. (2002) pH-sensitive, serum-stable and long-circulating liposomes as a new drug delivery system. J. Pharm. Pharmocol. 54, 5158. (10) Houk, B. E., Martin, R., Hochhaus, C., and Hughes, J. A. (2001) Pharmacokinetics of plasmid DNA in the rat. Pharm. Res. 18, 67-74 (2001). (11) Ishiwata, H., Suzuki, N., Ando, S., Kikuchi, H., and Kitagawa, T. (2000) Characteristics and biodistribution of cationic liposome and their DNA complexes. J. Controlled Release 69, 139-148. (12) Kaneo, Y., Tanaka, T., Nakano, T., and Yamaguchi, Y. (2001). Evidence for receptor-mediated hepatic uptake of pullulan in rats. J. Controlled Release 70, 365-373.

Plasmid DNA-Entrapped Nanoparticles (13) Kawabata, K., Takakura, Y., and Hashida, M. (1995) The fate of plasmid DNA after intravenous injection in mice: involvement of scavenger receptors in its hepatic uptake. Pharm. Res. 12, 825-830. (14) Liu, F., and Huang, L. (2002) Development of nonviral vectors for systemic gene delivery. J. Controlled Release 78, 259-266. (15) Lundberg, B. B., Mortimer, B. C., and Redgrave, T. G. (1996) Submicron lipid emulsions containing amphipathic poly(ethylene glycol) for use as drug-carriers with prolonged circulation time. Int. J. Pharm. 134, 119-127. (16) Mahato, R. I., Kawabata, K., Nomura, T., Takakura, Y., and Hashida, M. (1995a) Physicochemical and pharmacokinetics characteristics of plasmid DNA/cationic liposome complexes. J. Pharm. Sci. 84, 1267-1271. (17) Mahato, R. I., Kawabata, K., Takakura, Y., and Hashida, M. (1995b) In vivo disposition characteristics of plasmid DNA complexed with cationic liposomes. J. Drug Target. 3, 149157. (18) Minchin, R. F., Carpenter, D., Orr, R. J., (2001) Polyinosinic acid and polycationic liposomes attenuate the hepatic clearance of circulating plasmid DNA. J. Pharmacol. Exp. Ther. 296, 1006-1012. (19) Mosqueira, V. C. F., Legrand, P., Morgat, J., Vert, M., Mysiakine, E., Gref, R., Devissaguet, J., and Barratt, G. (2001) Biodistribution of long-circulating PEG-grafted nanocapsules in mice: effects of PEG chain length and density. Pharm. Res. 18, 1411-4119. (20) Nishikawa, M., Takakura, Y., and Hashida, M. (1998) Biofate of fat emulsions, in Submicron emulsions in drug targeting and delivery (S. Benita, Ed.) pp 99-118, Harwood Academic Publishers, Amsterdam. (21) Nishikawa, M., Takemura, S., Yamashita, F., Takakura, Y., Meijer, D. K., Hashida, M., and Swart, P. J. (2000a) Pharmacokinetics and in vivo gene transfer of plasmid DNA complexed with mannosylated poly(L-lysine) in mice. J. Drug Target. 8, 29-38. (22) Nishikawa, M., Yamauchi, M., Morimoto, K., Ishida, E., Takakura, Y., and Hashida, M. (2000b) Hepatocyte-targeted in vivo gene expression by intravenous injection of plasmid DNA complexed with synthetic multi-functional gene delivery system. Gene Ther. 7, 548-555. (23) Oh, Y. K., Kim, J. P., Yoon, H., Kim, J. M., Yang, J. S., and Kim, C. K. (2001) Prolonged organ retention and safety

Bioconjugate Chem., Vol. 13, No. 6, 2002 1327 of plasmid DNA administered in polyethylenimie complexes. Gene Ther. 8, 1587-1592. (24) Pam, P., Monck, M., Lee, D., Ludkovski, O., Leng, E. C., Clow, K., Stark, H., and Scherrer, P. (2000) Stabilized plasmid-lipid particles for systemic gene therapy. Gene Ther. 7, 1867-1874. (25) Patel, H. M. (1992) Serum opsonins and liposomes: their interaction and opsonophagocytosis. Crit Rev. Ther. Drug Carrier Syst. 9, 1437-1446. (26) Reimer, D. L., Zhang, Y. P., Kong, S., Wheeler, J. J., Graham, R. W., and Bally, M. B. (1995) Formation of novel hydrophobic complexes between cationic lipids and plasmid DNA. Biochemistry 34, 12877-12883. (27) Rolland, A., and Felgner, P., Eds. (1998) Nonviral gene delivery systems. Adv. Drug Deliv. Rev. 30, 1-227. (28) Sakaeda, T., and Hirano, K. (1998) Effect of composition on biological fate of oil particles after intravenous injection of O/W lipid emulsion. J. Drug Target. 6, 273-284. (29) Tabata, Y., Matsui, Y., Uno, K., Sokawa, Y., and Ikada, Y. (1999) Simple mixing of IFN with a polysaccharide having high liver affinity enables IFN to target to the liver. J. Interferon Cytokine Res. 19, 287-292. (30) Takakura, Y., Nishikawa, F., Yamashita, F., and Hashida, M. (2001) Development of gene drug delivery systems based on pharmacokinetic studies. Eur. J. Pharm. Sci. 13, 71-76. (31) Verbaan, F. J., Oussoren, C., van Dam, I. M., Takakura, Y., Hashida, M., Crommelin, D. A. J., Hennink, W. E., and Storm, G. (2001) The fate of poly(2-dimethyl amino ethyl)methacrylate-based polyplexes after intravenous administration. Int. J. Pharm. 214, 99-101. (32) Wheeler, J. J., Palmer, L., Ossanlou, M., MacLachlan, I., Graham, R. W., Zhang, Y. P., Hope, M. J., Scherrer, P., and Cullis, P. R. (1999) Stabilized plasmid-lipid particles: construction and characterization. Gene Ther. 6, 271-281. (33) Xi, K., Tabata, Y., Uno, K., Yoshimoto, M., Kishida, T., Sokawa, Y., Ikada, Y. (1996) Liver targeting of interferon through pullulan conjugation. Pharm. Res. 13, 1846-1850. (34) Yoshida, M., Mahato, R. I., Kawabata, K., Takakura, T., and Hashida, M. (1997) Disposition characteristics of plasmid DNA in the single-pass rat liver perfusion. Pharm. Res. 12, 597-603.

BC0255586