- Email: [email protected]
Gene Transfer by DNA–Gelatin Nanospheres
Archives of Biochemistry and Biophysics Vol. 361, No. 1, January 1, pp. 47–56, 1999 Article ID abbi.1998.0975, available online at http://www.idealibrary.com on
Gene Transfer by DNA–Gelatin Nanospheres Vu L. Truong-Le,* Scott M. Walsh,† Erik Schweibert,‡ Hai-Quan Mao,† William B. Guggino,‡ J. Thomas August,* ,§ ,1 and Kam W. Leong† *Department of Pharmacology and Molecular Sciences, †Department of Biomedical Engineering, ‡Department of Physiology, and §Department of Oncology, Johns Hopkins School of Medicine, Baltimore, Maryland 21205
Received April 6, 1998, and in revised form October 5, 1998
A DNA and gelatin nanoparticle coacervate containing chloroquine and calcium, and with the cell ligand transferrin covalently bound to the gelatin, has been developed as a gene delivery vehicle. In this study, the coacervation conditions which resulted in the formation of distinct nanoparticles are defined. Nanospheres formed within a narrow range of DNA concentrations and achieved incorporation of more than 98% of the DNA in the reaction. Crosslinking of gelatin to stabilize the particles does not effect the electrophoretic mobility of the DNA. DNA in the nanosphere is partially resistant to digestion with concentrations of DNase I that result in extensive degradation of free DNA but is completely degraded by high concentrations of DNase. Optimum cell transfection by nanosphere DNA required the presence of calcium and nanospheres containing transferrin. The biological integrity of the nanosphere DNA was demonstrated with a model system utilizing DNA encoding the cystic fibrosis transport regulator (CFTR). Transfection of cultured human tracheal epithelial cells (9HTEo) with nanospheres containing this plasmid resulted in CFTR expression in over 50% of the cells. Moreover, human bronchial epithelial cells (IB-3-1) defective in CFTR-mediated chloride transport were complemented with effective transport activity when transfected with nanospheres containing the CFTR transgene. © 1999 Academic Press
Synthetic gene delivery vehicles that have the required efficiency and safety for use in human gene therapy are being widely investigated as possible alternatives to biological vectors of viral and bacterial origin (1, 2). While the current synthetic systems are less efficient than viral vectors, rapid advances in cat1
To whom correspondence should be addressed.
0003-9861/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
ionic lipid designs (3, 4), amphiphilic DNA-binding peptides (5, 6), and polymers (7, 8) suggest that this gap may be closing. Moreover, synthetic DNA delivery systems that incorporate biological properties such as ligands for receptor-mediated endocytosis (9), agents that promote endosomal disruption (10), and cytoplasmic self-replication (11, 12) have achieved efficient levels of gene transfer. We previously introduced a novel gene delivery vehicle made of crosslinked DNA– gelatin nanosphere coacervates (13, 14). Nanosphere formation is driven by a combination of electrostatic and entropic forces with sodium sulfate employed as a desolvating reagent to facilitate greater charge– charge interactions between plasmid DNA and gelatin (15, 16). Sodium sulfate effects phase separation by influencing the degree of hydration of the two ionic species and thus increasing the degree of inter- and intracoulombic (attractive) forces between the ion pairs (17, 18). In addition, it has been suggested that sodium sulfate exerts a chargeshielding effect on oppositely charged sites of gelatin, which reduces the repulsive forces among the gelatin molecules and their intramolecular segments (15–17). Sodium sulfate was also crucial in forming chitosan– DNA nanoparticles (19). Because DNA is employed as the limiting reagent in the coacervation reaction, high DNA encapsulation efficiency (.98%) and loading level (25–30%, w/w) were obtained. The condition for nanosphere formation is mild, requiring neither the contact with high temperatures nor organic solvents, and permits the coencapsulation of labile agents. Cellular uptake of nanospheres was greatly enhanced by conjugating transferrin on the nanosphere surface, a phenomenon that may be attributed to endocytosis of the nanospheres. In addition, transfection of transferrincoated nanospheres was increased by the inclusion of chloroquine, presumably by its interference with the acidification of endosomal compartments, and thereby 47
48
TRUONG-LE ET AL.
improving the intracellular half-life of nanosphere DNA or by contribution to the stability of the DNA (20). The present report further characterizes the biochemical features of this DNA delivery system. The principal questions addressed are the reaction conditions necessary for particle formation, the protection of nanosphere DNA against DNase I digestion, the effect of encapsulated calcium on transfection, and the bioactivity of the DNA and its encoded protein as indicated by use of DNA encoding the cystic fibrosis transport regulator (CFTR). 2 EXPERIMENTAL PROCEDURES Plasmid constructs, antibodies, and cell lines. pSA306 is a plasmid containing the truncated CFTR gene with a 26-amino-acid epitope tag at the carboxy terminus (21). pRELuc contains a luciferase reporter gene inserted in the RSV promoter-driven Invitrogen pREP7 plasmid vector (14). mLAMP-1pcDNA is a RSV promoterdriven pcDNA I plasmid (Invitrogen) with full-length murine LAMP-1 gene (22). HIVgp160/LAMP is a pcDNA1 plasmid containing the HIV gp160/LAMP-1 insert (23). The AAV-CFTR is recombinant AAV virion containing the truncated CFTR cassette (24). Monoclonal antibodies are 934, chicken anti-human CFTR epitope tag antibody (25); and CHA, a nonspecific IgG with no known in vivo mouse epitope (26). 9HTEo is a human tracheal epithelial cell line (25), and IB-3-1 is a human bronchial epithelial cell line bearing the delta508 mutation (21). Synthesis of nanospheres. Porcine type A gelatin (175 bloom, Sigma) was boiled in distilled water (5% w/v) for 2 h, adjusted to pH 5.5 with NaOH, and then sterilized by filtering through a 0.22-mm filter. Gelatin–DNA nanosphere coacervates were synthesized by mixing 100 ml of a solution containing 5% (w/v) gelatin and 4 mM chloroquine with 100 ml of a solution containing 20 mg of plasmid DNA and 4.3 mM Na 2SO 4. Nanospheres without chloroquine were synthesized similarly except for a higher Na 2SO 4 concentration (45 mM). The reaction mixture was vortexed for 1 min at the highest speed at 55°C in a 0.5-ml Eppendorf tube. The unreacted components were removed by centrifuging the mixture on a 100-ml sucrose step gradient (35, 55, and 80%, w/w) at 40,000g for 7 min. The sucrose fraction containing the nanospheres (55% layer) was diluted with water to 200 ml, and 1 ml human transferrin (Sigma) solution (20 mg/ml in H 2O) and 22 ml of a 0.2 M 4-morpholineethanesulfonic acid buffer solution, pH 4.5, containing 0.02 mg/ml 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) were added. The reaction was carried out at room temperature for 30 min and then stopped by the addition of sodium acetate, pH 6.0, to 0.1 M final concentration. CaCl 2-treated nanospheres were generated by incubating freshly synthesized nanospheres for 12 h at 4°C with CaCl 2 and NaN 3 to a final concentration of 0.5 M CaCl 2 and 1% NaN 3 and then subsequently dialyzing against saline using a 300 kDa MWCO membrane for 12 h with two solution changes. Unless otherwise stated, this formulation is referred to as “standard condition” or “Nsp/cq/Ca/trf.” In vitro transfection. Nanospheres and the other DNA preparations (2 mg total DNA/well) were incubated for 4 h at 37°C and 5% CO 2 with 1 3 10 5 human kidney epithelial 293 cells grown on 12-well 2 Abbreviations used: RSV, Rous sarcoma virus; AAV, adeno-associated virus; CFTR, cystic fibrosis transmembrane regulator; Nsp, nanospheres; CQ, chloroquine; trf, transferrin; DMEM, Dulbecco’s modified Eagle’s medium; EDC, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride; PLL, poly-L-lysine; PBS, phosphatebuffered saline.
plates in medium containing Dulbecco’s modified eagle’s medium (DMEM), 2% fetal bovine serum, 2 mM L-glutamine, 50 units/ml penicillin, 50 mg/ml streptomycin, and 10 mg/ml gentamycin. Unbound nanospheres were removed by washing with DMEM, and the cells were further incubated in fresh medium containing 10% serum for 3 days. Luciferase gene expression was measured by assaying for enzyme activity in permeabilized cell extracts with the use of a luminometer (14). Transfection using Lipofectin and CaPO 4–DNA coprecipitation was carried out as described by Felgner (27) and by Graham and van der Eb (28), respectively. 36 Chloride efflux assay. Cells were seeded at 30 –50% confluence and transfected as above. Cells were washed 33 with CaMg-free phosphate-buffered saline (Gibco-BRL) to remove serum. Thirty microliters of 36Cl 2 solution (NEN-Dupont; 1 mCi/ml) was diluted in 9 ml of Ringers solution, and 1.5 ml of this loading solution was added to each well of a six-well plate. The plate was incubated for 2– 4 h in a 37°C warm room. The Ringers solution for these experiments was standard HCO 32-free, Hepes- and phosphate-buffered 140 mM NaCl Ringers solution supplemented with 5 mM glucose and titrated to pH 7.45 with 1 N NaOH. All efflux runs were performed in a 37°C warm room. Each well served as its own control. At Time 0, Ringers without cAMP agonists was added and removed immediately. A fresh aliquot of Ringers was added immediately and the efflux run was started. This process was repeated every 15 s until the time point at 1 min, at which time, Ringers with forskolin (2.5 mM), 8 bromo-cAMP (250 mM), and CPT-cAMP (250 mM) was added and removed and the sampling continued for the remaining 4 min of the efflux run. At the end of the run, 0.5 N NaOH was added in two aliquots to lyse and recover all of the cell lysate to determine how much labeled chloride had remained in the cells to standardize the data. Each sample was diluted in scintillation cocktail, counted in a scintillation counter, and normalized on a Microsoft Excel spreadsheet as the rate of labeled chloride lost from the cells per minute.
RESULTS
Complex coacervation of gelatin and DNA to form nanospheres. Coacervation reactions carried out using a range of concentration combinations of DNA, gelatin, and Na 2SO 4 showed that the formation of distinct DNA– gelatin nanospheres occurred over a narrow range of conditions. Diagrams of typical conditions for nanosphere formation with plasmids in the size range of 7 to 12 kbp are shown in Fig. 1. For example, with HIVgp160/LAMP pcDNA1, nanoparticles were formed with concentrations (w/v) of 0.0025 to 0.005% DNA, 1.5 to 5% gelatin, and 0.2 to 1.2% Na 2SO 4. In general, the required concentration of DNA was more confined than those of other components. Outside of these conditions, soluble gelatin–DNA complexes, aggregates, or loosely associated flocculates were obtained. Other conditions important for particle formation include a reaction temperature range of 55 6 7°C, the absence of other buffers or salts in the coacervation reaction, and the extent of mixing—a function both of the speed of vortexing and the dimensions of the reaction vessel. The conditions for coacervation may change if the mixing conditions are modified and if other charged molecules are included or the mixing conditions are modified and must be independently determined for any variation in the reaction protocol.
GENE TRANSFER BY DNA–GELATIN NANOSPHERES
49
FIG. 1. Three-dimensional contour plots of the combination of HIV-1gp160/LAMP DNA, gelatin, and Na 2SO 4 concentrations in the standard reaction mixture that resulted in nanosphere formation. A permutation of coacervation reaction conditions employing different DNA, gelatin, and Na 2SO 4 concentrations [all axis are given as percentages (w/v)] was carried out to determine the reaction conditions required for the formation of distinct, aggregate-free particles. The following concentration range were used: DNA [0, 0.00125, 0.0025, 0.005, and 0.01% (w/v)]; gelatin [0, 1, 2, 4, and 5% (w/v)]; and Na 2SO 4 [1.42, 1.136, 0.0852, and 0.284% (w/v)]. All concentration are final in the reaction mixture. Outside the delineated region, only gelatin–DNA complexes or loosely associated flocculates were observed.
A detailed synthesis procedure for making the optimized nanospheres formulation containing plasmid in the 7- to 12-kb size range is shown in Fig. 2. The formation of uncrosslinked nanoparticles occurred within seconds as gelatin is added to DNA at a high stirring rate. More than 98% of the DNA added to the reaction enters the coacervate. The nanoparticles, which may be viewed by phase-contrast microscopy, are separated from the unreacted components by centrifuging the nanosphere solution on a three-step sucrose density gradient solution. Although relatively stable at acidic pHs, these particles will dissociate in high-ionic-strength solutions at neutral pH, but are stabilized by crosslinking by the use of EDC, a water soluble carbodiimide that conjugates primary amino groups to carboxyl side chains of amino acids. Ligands or other reagents may also be included in the crosslink-
ing reaction at this point, after which sodium acetate or glycine is added to quench any unreacted carbodiimide. DNA crosslinking to itself or to chloroquine under the conditions of the reaction was ruled out because the electrophoretic mobility of DNA remained unchanged from that of untreated DNA when DNA was incubated with EDC or with chloroquine and EDC (Fig. 2b). When EDC was incubated with an equimolar mixture of DNA and gelatin, a DNA/gelatin ratio at which no gel retardation occurred, the electrophoretic mobility of DNA also remained unchanged. The DNA– gelatin complex as visualized by transmission electron microscopy in multiple fields was a spherical particle of solid core structure with a rough surface texture and composed of a matrix of interweaved gelatin and DNA (Fig. 3). Nanospheres made under standard conditions exhibited a mean diameter
50
TRUONG-LE ET AL.
FIG. 2. Schematic of the synthesis procedure for DNA– gelatin nanospheres. (a) Standard synthesis procedures for the optimal formulation of nanospheres. (b) Agarose gel electrophoresis of LAMP-1 pcDNA: lane 1, DNA alone; lane 2, DNA 1 EDC 1 chloroquine; lane 3, DNA 1 EDC 1 gelatin (1:1 mole ratio of DNA to gelatin).
of 412 6 120 nm, as determined by light scattering size analysis, and a zeta potential of 4.5 mV at pH 7.4. The size of the particle was influenced by the temperature of the reaction, the size of the plasmid, the Na 2SO 4 concentration, and the speed of mixing. Micro- as well as nanoparticle size ranges could be produced. In general, nano-size particles were obtained at a temperature of 55°C or lower, a Na 2SO 4 concentration higher than 10 mM, and with smaller plasmids (,12 kb). Protection of DNA against nuclease digestion. A major barrier to gene delivery in vivo is the vulnerability of DNA to nucleases that are found in abundance in serum and the extracellular matrix. A possible advantage of the encapsulated DNA approach to gene delivery is the protection against nucleases. We previously showed that nanosphere-encapsulated DNA was better protected against degradation in serum than was naked DNA (14). In this study the effect of DNase I on nanosphere DNA compared to other forms of DNA complexes was analyzed. Free plasmid DNA, nanospheres, poly-L-lysine–DNA complex (PLL–DNA), and
Lipofectamine–DNA complex were incubated for 15 min with 10 mU/ml DNase I. After incubation, the nanosphere samples were pelleted to separate the particles from the supernatant and the samples were analyzed by agarose gel electrophoresis (Fig. 4). Nanosphere DNA, Lipofectamine–DNA, and, to a lessor extent, PLL–DNA made at a 1:1 weight ratio were partially resistant to degradation during incubation with DNase I for 15 min. In contrast, naked DNA was extensively degraded. This protection of DNA by nanospheres was only partial, since nanospheres incubated in a higher DNase I solution (200 U/ml) showed complete digestion of nanosphere DNA. Interestingly, nanospheres crosslinked with twice the EDC concentration did not differ in DNA protection level (data not shown). The lack of complete protection against nuclease digestion suggests that DNA degradation occurred by a form of surface erosion of DNA that was dynamically mobile in the nanosphere, or, possibly more likely, that the nanosphere was porous enough to allow accessibility of DNase to the particle’s interior, or both.
GENE TRANSFER BY DNA–GELATIN NANOSPHERES
FIG. 3. The gelatin–DNA nanosphere as visualized by electron microscopy. 9HTEo cells were incubated with nanospheres containing pSA306 pDNA (2 mg/well) in complete medium for 15 min. The cells were washed twice with PBS, fixed by incubation with a PBS solution containing 2.5% glutaraldehyde and 2% paraformaldehyde for 30 min at 25°C, washed with 0.1 M cacodylate buffer, incubated with 1% osmium tetraoxide, treated with 1% uranyl acetate, and then dehydrated in graded ethanol. The cells were embedded in Epon, thin sectioned, and viewed on a Jeol electron microscope. The arrow shows a nanosphere at the surface of an 9HTEo cell after 15 min incubation (75,0003). Bar, 100 nm.
Effect of nanosphere calcium on transfection. Divalent metal cations such as Mg 21, Ba 21, and Mn 21 can form ionic complexes with the helical phosphates on DNA (29). Ca 21 exhibits similar affinity to DNA (K d of 1.1 3 10 23 M 21) and forms CaPO 4 complexes with the nucleic acid backbone and thus may impart a stabilization function to certain DNA structures (30, 31). It has been suggested that CaPO 4, when used in complexes with plasmid DNA, exerts its positive effect on gene transfer by stimulating cellular uptake of DNA in a process involving either endocytosis of the membrane-bound DNA complex or enhanced permeabilization of the plasma membrane to facilitate DNA entry (32). In addition, some studies have correlated intracellular elevation of calcium with transcription and gene expression (33). These observations suggest the possible benefit on nanosphere gene transfer by the codelivery of calcium. Transferrin-coated nanospheres added to 293 cells transfected better in the presence of exogenously added CaPO 4 (100 mM) (Fig. 5). Such transfection enhancement required the presence of transferrin because CaPO 4 failed to stimulate transfection when nanospheres alone (without transferrin) were added to 293 cells. Because the CaPO 4 effect on nanosphere-mediated transfection required the presence of transferrin on nanospheres, presumably to initiate surface binding, the results suggested that the
51
added CaPO 4 may act via stimulation of cellular uptake of bound nanospheres. The CaPO 4 effect could be mimicked when the nanospheres were loaded with calcium by incubating newly synthesized transferrincoated nanospheres with a concentrated calcium solution. Transfection of 293 cells with crosslinked nanospheres made with different calcium loading levels showed the highest transfection was achieved with those containing a 1.8% (w/w) calcium loading level. The possibility that the CaCl 2-mediated increase in transfection was due to the extracellular release of DNA and calcium ions from the nanosphere was investigated. In that case, the resultant free DNA and CaPO 4 (formed with PO 42 in the medium) could have complexed so that the enhancement in transfection might have resulted from formation of free DNA– CaPO 4 complexes and not from intact nanospheres. This possibility was tested by measuring transfection of 293 cells by nanosphere supernatant in the presence or absence of CaPO 4 (Table I). The level of transfection
FIG. 4. Protection of DNA against nuclease digestion. Plasmid DNA, DNA– gelatin nanospheres, poly-L-lysine–DNA complex (PLL), or Lipofectamine–DNA complex was incubated with 10 mU/ml of DNase I for 15 min at 37°C. The nanosphere samples were pelleted to separate the nanospheres from the supernatant, the pellet fraction incubated with trypsin to release the DNA, and the resulting DNA from the supernatant and pellet fractions was subjected to 1% agarose gel electrophoresis. The DNA from other samples was extracted twice with phenol:chloroform:isoamyl alcohol (50:49:1) prior to gel electrophoresis. Lanes: 1, untreated control DNA; 2, plasmid DNA treated with 10 mU/ml DNase; 3, Nsp–DNA treated with 10 mU/ml DNase, pellet fraction; 4, Nsp-DNA treated with 10 mU/ml DNase, supernatant fraction; 5, PLL–DNA at 1:1 weight ratio treated with 10 mU/ml DNase; 6, Lipofectamine–DNA at 2:1 weight ratio treated with 10 mU/ml DNase; 7, Nsp–DNA incubated in 200 U/ml DNase I for 15 min at 37°C, pellet fraction. Degradation of DNA was apparent with naked DNA and the PLL–DNA complex treated with 10 mU/ml DNase and with nanosphere DNA treated with 200 U/ml DNase I. In contrast, DNA in nanospheres (lane 3) and Lipofectamine–DNA complex (lane 6) exhibited greater resistance against nuclease degradation at a concentration of 10 mU/ml DNase.
52
TRUONG-LE ET AL.
FIG. 5. The effect of calcium on nanosphere transfection of 293 cells in vitro. Nanospheres synthesized containing pRELuc DNA (Nsp), containing transferrin coating (Nsp/trf ), and containing transferrin and calcium (Nsp/ Trf/Ca) were added to cells in a 12-well dish (2 mg total DNA/5 3 10 4 cells per well) for 4 h, washed to remove excess unbound nanospheres, and then assayed for luciferase expression at Day 3. Nsp 1 CaPO 4 and Nsp/ Trt 1 CaPO 4 represent, respectively, Nsp and Nsp/ Trf samples that were incubated with cells for 2 h, washed, pulsed with CaPO 4 (100 mM) for two additional hours, and assayed for luciferase expression at Day 3. Data are average values from triplicate determinations and are representative of two different experiments, 6SD.
of supernatant from complete, crosslinked nanospheres was barely above background, indicating that there was very little, if any, free DNA in the nanosphere sample. In comparison, cells transfected with supernatant of the uncrosslinked nanosphere sample containing free DNA released from the nanosphere showed a significant level of luciferase activity in the presence of added CaPO 4. It was also possible that proteases in the medium may partially degrade the nanosphere during the period of cell culture medium incubation, releasing free DNA and calcium which
could also form DNA–CaPO 4 complexes. To address this, the nanospheres were incubated with culture medium (without cells) for 4 h, and the resultant medium was used to transfect cells under similar conditions described above for the nanosphere supernatant (Table I). Again, there was a significant level of transfection only with the uncrosslinked nanosphere sample. Higher transfection activity was observed in the uncrosslinked nanosphere group (using CaPO 4 complexation) than in the nanosphere pellet group, which suggested that either crosslinking decreased the transfection efficiency or that the amount of DNA released from the crosslinked nanospheres by Day 3 was small. Thus, we concluded that little DNA was released from the crosslinked nanospheres into the cell culture medium during the time of incubation, and that the transfection can be attributed to intact nanospheres and not to free DNA and calcium contaminants. Transfection with nanosphere DNA encoding the human CFTR protein. Our studies have shown that the optimal nanosphere formulation, as defined in vitro transfection studies carried out on 293 human kidney epithelial cells, contained calcium in addition to chloroquine and surface-bound transferrin (14). Further documentation of the transfection with nanosphere DNA using different cell lines and a therapeutically relevant model gene was carried out. The bioactivity of DNA delivered to cells by nanospheres was studied by transfecting 9HTEo cells with plasmid pSA306 expressing a chimeric human CFTR protein containing a 26-amino-acid epitope tag at the carboxy terminus. Gene expression analysis, as measured by antibody binding to the epitope tag, showed approximately 50 to 60% of the cells incubated with the nanospheres were positive for the expression of the tag (Fig. 6). The frequency of stained cells following nanosphere transfection was significantly greater than that of cells transfected by the equivalent DNA dose given
TABLE I Luciferase activity (light units) a Nanosphere preparation
Nsp pellet
Nsp supernatant
Nsp medium 1CaPO 4
Complete Noncrosslinked
101,800 380
1450 550
30 13,830
1CaPO 4 910 450
4490 418,010
Note. Complete nanospheres (Nsp/trf/CQ/Ca) or noncrosslinked nanospheres (Nsp/CQ) containing the pRELuc plasmid were centrifuged (15,000g, 12 min) to separate the particulate fraction (Nsp pellet) from the supernatant fraction (Nsp supernatant). The nanosphere pellet fraction was resuspended in water and incubated with 293 cells (2 mg DNA/10 5 cells), while the nanosphere supernatant (2 mg DNA/10 5 cells) was incubated with 293 cells in the presence or absence of CaPO 4. In a separate experiment, nanospheres were incubated with fresh culture medium (without cells) for 4 h and then centrifuged to separate the particulate fraction from the supernatant fraction (Nsp medium). The supernatant fraction was subsequently incubated with 293 cells in the presence or absence of CaPO 4. Luciferase assay was carried at Day 3 to determine gene transfection. a Background subtracted.
GENE TRANSFER BY DNA–GELATIN NANOSPHERES
53
FIG. 6. Expression of the CFTR gene in 9HTEo human tracheal epithelial cells. Nanospheres containing pSA306 DNA encoding the CFTR gene with a 26-amino-acid epitope tag were prepared as described under Experimental Procedures. Transfection with the CFTR gene was carried out with 9HTEo cells plated on a cover glass in a 12-well plate. The cells were transfected by the addition of pSA306, 2 mg total DNA/3 3 10 5 cells/well, in nanospheres, as a Lipofectamine complex, or as naked DNA, washed after 4 h, and then incubated for 3 days. The cells were then fixed with 4% paraformaldehyde, blocked by adding 1% bovine serum albumin for 30 min, permeabilized with 0.1% saponin PBS for 10 min, incubated with the anti-epitope tag antibody mAb934 (1:100 dilution) for 1 h, washed three times with PBS, and then incubated with a 1:1000 dilution of fluorescein isothiocyanate-labeled anti-chicken IgG antibody (Jackson Lab, Bar Harbor, ME). The field of view shown contained cells at ;75– 80% confluency (the corresponding bright-field image is not shown). (A) Free DNA transfected cells showed only background cellular autofluorescence; the occasional cells with a higher level of staining were not investigated. (B) Nanospheretransfected cells showed specific staining in approximately 50% of the cells. (C) Lipofectamine–DNA-transfected cells showed staining in approximately 20% of the cells.
as Lipofectamide–DNA complex (.50% vs ;23%). This transfection frequency was the highest observed for any cell types tested using nanospheres as the transfectant agent. Other cells (293 human kidney epithelial, HeLa cells, IB-3-1 human bronchial epithelial, COS-7 monkey cells, and U937 human histiocytic– monocytic cells) typically showed a range of 2–15% transfection frequency. Variations in the transfection frequency were also observed with Lipofectamine– DNA and CaPO 4–DNA transfection methods. None of the nonviral gene transfection methods tested could match the results achieved by use of AAV. The majority of HTE cells infected under similar conditions with 10 9 AAV encoding the CFTR gene were transfected, representing an efficiency based on the amount of DNA that was generally 10 to 100 times greater in cell transfection than that achieved with the DNA plasmid delivery systems (data not shown). Cells transfected with nanosphere DNA encoding the CFTR gene form a functional chloride channel. Additional studies with plasmid pSA306 expressing a chimeric human CFTR protein were conducted with IB3-1 cells, a human bronchial epithelial cell line bearing the delta508 mutation which results in a defective cAMP-stimulated chloride transport phenotype. This system was used to study the efficacy of transfection with DNA– gelatin nanospheres in reconstituting chloride transport, comparing the effect of transfection with free DNA, DNA– gelatin nanospheres, or infection of the cells with an AAV CFTR vector system (Fig. 7). In this standard assay (34), the function of CFTR is assessed by cAMP stimulation of 36Cl 2 efflux and is expressed as the percentage rate of efflux per minute
normalized to 100%, as a function of time after cAMP stimulation. Because the efflux is followed with multiple samples and cAMP is added at the 1-min time period after the beginning of the experiment, each culture dish serves as its own control. Efflux of Cl 2 from the nontransduced cells (background cells), was not stimulated by cAMP. Similarly, in cells treated with free DNA there is no detectable increase in the rate of Cl 2 release following cAMP addition. The small fluctuations in the chloride efflux rate as seen with the nontransduced cells following cAMP addition are attributed to mechanical perturbations in the cells by adding cAMP. In contrast, when cAMP was added to cells transfected with the gene encoding the CFTR carrier by nanosphere DNA or by the AAV CFTR vector, there was a rapid and transient increase in 36Cl 2 efflux, indicative of the action of functional CFTR. In the nanosphere-treated cells, the rate jumped from less than 20% per minute just prior to cAMP addition to approximately 50% per minute after addition of cAMP. Likewise, in the AAV CFTR-treated cells the rate increased from about 20% per minute prior to 40% after. This rapid and transient increase in efflux demonstrated that CFTR was functioning in the transduced cells compared to nontransfected controls. Because the absolute magnitude of the response depends not only on the expression of CFTR in an individual cell but also on the number of cells transduced, a quantitative comparison of the absolute magnitude of the response among groups is not possible. It may be noted, however, that a comparable result in this assay was obtained with 10 9 AAV, which contains much less DNA than the 5 mg added as nanosphere DNA. We previ-
54
TRUONG-LE ET AL.
ously have shown that the transduction efficiency of nanosphere DNA in vitro, while greater than that of naked DNA, was 10 to 1000 times less than that of AAV virions (14). DISCUSSION
FIG. 7. 36Chloride efflux assay of chloride transport across human bronchial epithelial IB-3-1 cells transfected with pSA306 DNA as free DNA, nanosphere–DNA, or AAV CFTR viral particles. The IB3-1 cells are a human CF bronchial epithelial cell line with defective transport of Cl 2. pSA306 DNA (5 mg total DNA/5 3 10 5 cells/per 35-mm well of a multiwell plate), nanospheres containing pSA306 DNA (5 mg total DNA/5 3 10 5 cells/well), or AAV CFTR virions (10 9 viral particles/5 3 10 5 cells/well) were incubated with IB-3-1 cells for 4 h as described above and then analyzed for chloride efflux at Day 3 as described in detail elsewhere (34). The transfected cells were labeled with 36Cl 2 for 4 h, washed to remove excess 36Cl 2 (t 5 0 is the time after the last wash) and loaded with cAMP at t 5 1 min to activate the CFTR 36Cl 2 current. The medium was sampled for 36Cl 2 effluxed from the cells at various time points. The data for each experiment are averages of six separate assays from six individual dishes of cells run on the same day and are representative of two experiments done on two separate occasions. The results are expressed as the percentage rate of 36Cl 2 efflux per minute normalized to 100% and plotted as a function of time. Bkgd, background 36Cl 2 currents in 36chloride-loaded, untransfected cells.
Complex coacervation has been widely used to formulate controlled release microspheres. We had previously described the use of gelatin– chondroitin sulfate and collagen– chondroitin sulfate complex coacervates to deliver therapeutic agents (26, 34). In this system, DNA and gelatin are used as the oppositely charged biopolymers. Since there was no evidence of crosslinking of the DNA to gelatin or to any other nanosphere components, the nanosphere can be thought of as a solid core structure composed of gelatin entangled with DNA. The spherical nature of the coacervate is a result of the gelatin–DNA complex adopting the minimal energy architecture in a process governed by the relative interfacial tension. One of the advantages of the nanosphere approach to gene delivery is the enhanced protection of DNA against nuclease degradation, which should improve in vivo bioavailability of the DNA. Our data showed the significant improvement in DNA protection when the nanospheres were incubated with a low concentration of DNase I (10 mU/ml) compared to the rapid degradation of free DNA and a DNA–PLL complex. Higher degrees of nanosphere crosslinking further stabilized the gelatin matrix, which in turn improved the half-life of DNA treated with DNase; however, at high crosslinking density, the transfectability of the nanosphere was compromised, possibly because of a reduced rate of release of the DNA (unpublished results). In addition to increasing the crosslinking density of the gelatin matrix, we have attempted to increase the resistance of nanosphere DNA to DNase by coencapsulating DNase inhibitors. Encapsulation of sodium iodoacetate (35) and aurintricarboxylic acid (36), both known to be potent DNase I inhibitors, showed a minor enhancement in DNA protection, but failed to show consistent improvement in transfection level (data not shown). Other materials added to the coacervation reaction can be coencapsulated in the nanosphere by either covalent conjugation, entrapment, or ionic interactions with DNA or gelatin. However, these nanospheres are permeable to low-molecular-weight compounds as shown by the partial loss of chloroquine and calcium during dialysis. This suggests that larger and more charged compounds should enjoy a higher loading level. In principle, this may include peptides, proteins, and a variety of smaller but charged compounds. Chloroquine was added to enhance transfection by DNA, presumably by interfering with the acidification of en-
GENE TRANSFER BY DNA–GELATIN NANOSPHERES
dosome compartments (37) or reducing DNase degradation (38). A chloroquine loading level of 2% (w/w) into the nanospheres was achieved by reducing the Na 2SO 4 concentration 10-fold in the coacervation reaction. This high uptake of chloroquine was attributed to its affinity for polynucleotides (K d of 10 24 M 21) via intercalation in the DNA helix and electrostatic interactions with the phosphate backbone (39). The level of transgene expression in 293 human kidney epithelial cells incubated with nanospheres was also markedly increased by the presence of transferrin on the nanosphere and to a lesser extent by calcium. In the absence of transferrin there was no detectable transfection of 293 cells with as much as 10 mg total DNA either as gelatin–DNA nanospheres or gelatin–DNA complexes. Transferrin has been demonstrated to be efficient in mediating binding and uptake of DNA–polymer or –protein complexes in several studies (40), presumably by acting as a ligand to the corresponding cell surface receptor and thereby enhancing the binding and/or the uptake of the nanosphere into the cell. Transfection enhancement by calcium was attributed to calcium that was associated with the nanospheres and not to those resulting from leakage. Our data also suggested that a possible role of calcium is to facilitate release of DNA from the gelatin matrix by its competition for electrostatic interactions with the gelatin and DNA. However, other explanations that account for the enhancement in transfection such as stimulated uptake of nanospheres, perhaps as a result of the formation of CaPO 4 complexes on the nanospheres, or specific intracellular effects are also plausible. These in vitro transfection experiments using the CFTR gene have shown that a large percentage of 9HTEo cells could be transfected by use of the DNA– gelatin nanospheres. In addition, successful complementation of defective chloride transport in mutant IB-3-1 cells was demonstrated. While similar results have been accomplished using a variety of viral and nonviral gene delivery methods in vitro (41), the therapeutic advance to successful in vivo CF gene therapy remains a great challenge. Our studies have shown that under in vitro cell culture conditions, both gene transfer efficiency and efficacy were poorer than with the AAV viral counterpart. However, the microencapsulated DNA approach may offer some advantages that are unique compared to other approaches, such as targeted gene delivery by the use of monoclonal antibodies or cell ligands and codelivery of therapeutic compounds in a controlled fashion. Studies of the delivery of the CFTR gene to rabbit airways by use of nanosphere– DNA have shown a correspondingly high efficiency of transfection of airway epithelial cells as was observed in these in vitro studies, and an evaluation of the codelivery of sodium 4-phenylbutyrate, an agent that has been found to enhance the CFTR function by stim-
55
ulating the mutant DF508 CFTR activity (42), is underway. ACKNOWLEDGMENTS This study was supported by NIH CA 68011, NIH 1 RO1 A141908, and the Cystic Fibrosis Foundation.
REFERENCES 1. Ledley, F. D. (1994) Curr. Opin. Biotech. 5(6), 626 – 636. 2. Tomlinson, E., and Rolland, A. P. (1996) J. Controlled Rel. 39, 357–372. 3. Lee, R. J., and Huang, L. (1997) Crit. Rev. Ther. Drug Carrier Sys. 14(2), 173–206. 4. Hong, K., Zheng, W., Baker, A., and Papahadjopoulos, D. (1997) FEBS Lett. 400(2), 233–237. 5. Gottschalk, S., Sparrow, J. T., Hauer, J., Mims, M. P., Leland, F. E., Woo, S. L., and Smith, L. C. (1996) Gene Ther. 3(5), 48 –57. 6. Niidome, T., Ohmori, N., Ichinose, A., Wada, A., Mihara, H., and Hirayama, T. (1997) J. Biol. Chem. 272(24), 15307–15312. 7. Boussif, O., Lezoualch, F., Zanta, M. A., Mergny, M. D., Scherman, D., Demeneix, B., and Behr, J. P. (1995) Proc. Natl. Acad. Sci. USA 92(16), 7297–7301. 8. Haensler, J., and Szoka, F. C. (1993) Bioconj. Chem. 4, 372–379. 9. Goldman, C. K., Rogers, B. E., Douglas, J. T., Sosnowski, B. A., Ying, W., Siegal, G. P., Baird, A., Campain, J. A., and Curiel, D. T. (1997) Cancer Res. 57(8), 1447–1451. 10. Plank, C., Oberhauser, B., Mechtler, K., Koch, C., and Wagner, E. (1994) J. Biol. Chem. 269, 12918 –12924. 11. Noguiez-Hellin, P., Meur, M. R., Salzmann, J. L., and Klatzmann, D. (1996) Proc. Natl. Acad. Sci. USA 93(9), 4175– 4180. 12. Johanning, F. W., Conry, R. M., LoBuglio, A. F., Wright, M., Sumerel, L. A., Pike, M. J., and Curiel, D. T. (1995) Nucleic Acids Res. 23(9), 1495–1501. 13. Truong-Le, V. L., Mao, H-Q., Walsh, S. W., Leong, K. W., and August, J. T. (1997) Proc. Intern. Symp. Control Rel. Bioact. Mater. 24, 75–76. 14. Truong-Le, V. L., August, J. T., and Leong, K. W. (1998) Hum. Gene Ther. 9, 1709 –1717. 15. Marty, J. J., Oppenheim, R. C., and Speiser, P. (1978) Pharm. Acta Helv. 53, 17–23. 16. Bugenberg de Jong, H. G. (1949) in Colloid Science (Kruyt, H. R., Ed.), Vol. II, pp. 232– 480, Elsevier, New York. 17. Burgess, D. J. (1990) J. Colloid Interfacial Sci. 140, 227–239. 18. Khalil, S. A. H., Jr, and Carless, J. E. (1981) J. Pharm. Pharmacol. 20, 6978 – 6987. 19. Leong, K. W., Mao, H.-Q., Truong-Le, V. L., Roy, K., Walsh, S. W., and August, J. T. (1998) J. Contr. Rel. 53, 183–193. 20. Wu, P., and Schurr, J. M. (1989) Biopolymers 24, 1695–1703. 21. Flotte, T. R., Afione, S. A., Solow, R., Drumm, M. L., Markakis, D., Guggino, W. B., Zeitlin, P. L., and Carter, B. J. (1993) J. Biol. Chem. 268(5), 3781–3790. 22. Guarnieri, F. G., Arterburn, L. M., Penno, M. B., Cha, Y., and August, J. T. (1993) J. Biol. Chem. 268(3), 1941–1946. 23. Ruff, A. L., Guarnieri, F. G., Staveley-O’Carroll, K. F., Siliciano, R. F., and August, J. T. (1996) J. Biol. Chem. 272(13), 8671– 8678. 24. Egan, M., Flotte, T., Afione, S., Lolow, R., Zeitlin, P. L., Carter, B. J., and Guggino, W. B. (1992) Nature 358, 581–584. 25. Flotte, T. R., Afione, S. A., Conrad, C., McGrath, S. A., Solow, R., Oka, H., Zeitlin, P. L., Guggino, W. B., and Carter, B. J. (1993) Proc. Natl. Acad. Sci. USA 90(22), 10613–10617.
56
TRUONG-LE ET AL.
26. Truong, V. L., Williams, J. R., Hildreth, J. E. K., and Leong, K. W. (1995) Drug Deliv. 2, 166 –174. 27. Felgner, P. L., Gadek, T. R., Holm, M., Roman, R., Chan, H. W., Wenz, M., Northrop, J. P., Ringold, G. M., and Danielsen, M. (1987) Proc. Natl. Acad. Sci. USA 84, 7413–7417. 28. Graham, F. L., and Van Der Eb, A. J. (1973) Virology 52, 456 – 467. 29. Shack, J., Jenkins, R. J., and Thompsett, J. M. (1953) J. Biol. Chem. 203, 373–387. 30. Wiberg, J. S., and Neuman, W. F. (1957) Arch. Biochem. Biophys. 72, 66 – 83. 31. Herzog, M., and Soyer, M. O. Eur. J. Cell Biol. 30(1), 33– 41. 32. Orrantia, E., and Chang, P. L. (1990) Exp. Cell Res. 190(2), 170 –174. 33. Negulescu, P. A., Shastri, N., and Cahalan, M. D. (1994) Proc. Natl. Acad. Sci. USA 91, 2873–2877. 34. Devidas, S., Yue, H., and Guggino, W. B. (1998) J. Biol. Chem., in press.
35. Shao, W., and Leong, K. W. (1995) J. Biomaterials Sci. Polymer Ed. 7(5), 389 –399. 36. Nagae, S., Nakayama, J., Nakano, I., and Anai, M. (1982) Biochemistry 21, 1339 –1344. 37. Hallick, R. B., Chelm, B. K., Gray, P. W., and Orozco, E. M. (1977) Nucleic Acids Res. 4(9), 3055–3064. 38. Pless, D. D., and Wellner, R. B. (1996) J. Cell. Biochem. 62, 27–39. 39. Kurnick, N. B., and Radcliffe, M. S. (1962) J. Lab. Clin. Med. 60, 669 – 688. 40. Cohen, S. N., and Yielding, K. L. (1965) J. Biol. Chem. 240, 3123–3131. 41. Cotten, M., and Wagner, E. (1993) Curr. Opin. Biotechnol. 4, 705–710. 42. Flotte, T. R., and Carter, B. J. (1997) Adv. Pharmacol. 40, 85–101. 43. Walsh, S. W., Flotte, T. R., Truong-Le, V. L., Rubenstein, R., Beck, S., August, J. T., Zeitlin, P., and Leong, K. W. (1997) Proc. Int. Symp. Contr. Rel. Bioact. Mater. 24, 39 – 40.
Gene Transfer by DNA–Gelatin Nanospheres Vu L. Truong-Le,* Scott M. Walsh,† Erik Schweibert,‡ Hai-Quan Mao,† William B. Guggino,‡ J. Thomas August,* ,§ ,1 and Kam W. Leong† *Department of Pharmacology and Molecular Sciences, †Department of Biomedical Engineering, ‡Department of Physiology, and §Department of Oncology, Johns Hopkins School of Medicine, Baltimore, Maryland 21205
Received April 6, 1998, and in revised form October 5, 1998
A DNA and gelatin nanoparticle coacervate containing chloroquine and calcium, and with the cell ligand transferrin covalently bound to the gelatin, has been developed as a gene delivery vehicle. In this study, the coacervation conditions which resulted in the formation of distinct nanoparticles are defined. Nanospheres formed within a narrow range of DNA concentrations and achieved incorporation of more than 98% of the DNA in the reaction. Crosslinking of gelatin to stabilize the particles does not effect the electrophoretic mobility of the DNA. DNA in the nanosphere is partially resistant to digestion with concentrations of DNase I that result in extensive degradation of free DNA but is completely degraded by high concentrations of DNase. Optimum cell transfection by nanosphere DNA required the presence of calcium and nanospheres containing transferrin. The biological integrity of the nanosphere DNA was demonstrated with a model system utilizing DNA encoding the cystic fibrosis transport regulator (CFTR). Transfection of cultured human tracheal epithelial cells (9HTEo) with nanospheres containing this plasmid resulted in CFTR expression in over 50% of the cells. Moreover, human bronchial epithelial cells (IB-3-1) defective in CFTR-mediated chloride transport were complemented with effective transport activity when transfected with nanospheres containing the CFTR transgene. © 1999 Academic Press
Synthetic gene delivery vehicles that have the required efficiency and safety for use in human gene therapy are being widely investigated as possible alternatives to biological vectors of viral and bacterial origin (1, 2). While the current synthetic systems are less efficient than viral vectors, rapid advances in cat1
To whom correspondence should be addressed.
0003-9861/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
ionic lipid designs (3, 4), amphiphilic DNA-binding peptides (5, 6), and polymers (7, 8) suggest that this gap may be closing. Moreover, synthetic DNA delivery systems that incorporate biological properties such as ligands for receptor-mediated endocytosis (9), agents that promote endosomal disruption (10), and cytoplasmic self-replication (11, 12) have achieved efficient levels of gene transfer. We previously introduced a novel gene delivery vehicle made of crosslinked DNA– gelatin nanosphere coacervates (13, 14). Nanosphere formation is driven by a combination of electrostatic and entropic forces with sodium sulfate employed as a desolvating reagent to facilitate greater charge– charge interactions between plasmid DNA and gelatin (15, 16). Sodium sulfate effects phase separation by influencing the degree of hydration of the two ionic species and thus increasing the degree of inter- and intracoulombic (attractive) forces between the ion pairs (17, 18). In addition, it has been suggested that sodium sulfate exerts a chargeshielding effect on oppositely charged sites of gelatin, which reduces the repulsive forces among the gelatin molecules and their intramolecular segments (15–17). Sodium sulfate was also crucial in forming chitosan– DNA nanoparticles (19). Because DNA is employed as the limiting reagent in the coacervation reaction, high DNA encapsulation efficiency (.98%) and loading level (25–30%, w/w) were obtained. The condition for nanosphere formation is mild, requiring neither the contact with high temperatures nor organic solvents, and permits the coencapsulation of labile agents. Cellular uptake of nanospheres was greatly enhanced by conjugating transferrin on the nanosphere surface, a phenomenon that may be attributed to endocytosis of the nanospheres. In addition, transfection of transferrincoated nanospheres was increased by the inclusion of chloroquine, presumably by its interference with the acidification of endosomal compartments, and thereby 47
48
TRUONG-LE ET AL.
improving the intracellular half-life of nanosphere DNA or by contribution to the stability of the DNA (20). The present report further characterizes the biochemical features of this DNA delivery system. The principal questions addressed are the reaction conditions necessary for particle formation, the protection of nanosphere DNA against DNase I digestion, the effect of encapsulated calcium on transfection, and the bioactivity of the DNA and its encoded protein as indicated by use of DNA encoding the cystic fibrosis transport regulator (CFTR). 2 EXPERIMENTAL PROCEDURES Plasmid constructs, antibodies, and cell lines. pSA306 is a plasmid containing the truncated CFTR gene with a 26-amino-acid epitope tag at the carboxy terminus (21). pRELuc contains a luciferase reporter gene inserted in the RSV promoter-driven Invitrogen pREP7 plasmid vector (14). mLAMP-1pcDNA is a RSV promoterdriven pcDNA I plasmid (Invitrogen) with full-length murine LAMP-1 gene (22). HIVgp160/LAMP is a pcDNA1 plasmid containing the HIV gp160/LAMP-1 insert (23). The AAV-CFTR is recombinant AAV virion containing the truncated CFTR cassette (24). Monoclonal antibodies are 934, chicken anti-human CFTR epitope tag antibody (25); and CHA, a nonspecific IgG with no known in vivo mouse epitope (26). 9HTEo is a human tracheal epithelial cell line (25), and IB-3-1 is a human bronchial epithelial cell line bearing the delta508 mutation (21). Synthesis of nanospheres. Porcine type A gelatin (175 bloom, Sigma) was boiled in distilled water (5% w/v) for 2 h, adjusted to pH 5.5 with NaOH, and then sterilized by filtering through a 0.22-mm filter. Gelatin–DNA nanosphere coacervates were synthesized by mixing 100 ml of a solution containing 5% (w/v) gelatin and 4 mM chloroquine with 100 ml of a solution containing 20 mg of plasmid DNA and 4.3 mM Na 2SO 4. Nanospheres without chloroquine were synthesized similarly except for a higher Na 2SO 4 concentration (45 mM). The reaction mixture was vortexed for 1 min at the highest speed at 55°C in a 0.5-ml Eppendorf tube. The unreacted components were removed by centrifuging the mixture on a 100-ml sucrose step gradient (35, 55, and 80%, w/w) at 40,000g for 7 min. The sucrose fraction containing the nanospheres (55% layer) was diluted with water to 200 ml, and 1 ml human transferrin (Sigma) solution (20 mg/ml in H 2O) and 22 ml of a 0.2 M 4-morpholineethanesulfonic acid buffer solution, pH 4.5, containing 0.02 mg/ml 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) were added. The reaction was carried out at room temperature for 30 min and then stopped by the addition of sodium acetate, pH 6.0, to 0.1 M final concentration. CaCl 2-treated nanospheres were generated by incubating freshly synthesized nanospheres for 12 h at 4°C with CaCl 2 and NaN 3 to a final concentration of 0.5 M CaCl 2 and 1% NaN 3 and then subsequently dialyzing against saline using a 300 kDa MWCO membrane for 12 h with two solution changes. Unless otherwise stated, this formulation is referred to as “standard condition” or “Nsp/cq/Ca/trf.” In vitro transfection. Nanospheres and the other DNA preparations (2 mg total DNA/well) were incubated for 4 h at 37°C and 5% CO 2 with 1 3 10 5 human kidney epithelial 293 cells grown on 12-well 2 Abbreviations used: RSV, Rous sarcoma virus; AAV, adeno-associated virus; CFTR, cystic fibrosis transmembrane regulator; Nsp, nanospheres; CQ, chloroquine; trf, transferrin; DMEM, Dulbecco’s modified Eagle’s medium; EDC, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride; PLL, poly-L-lysine; PBS, phosphatebuffered saline.
plates in medium containing Dulbecco’s modified eagle’s medium (DMEM), 2% fetal bovine serum, 2 mM L-glutamine, 50 units/ml penicillin, 50 mg/ml streptomycin, and 10 mg/ml gentamycin. Unbound nanospheres were removed by washing with DMEM, and the cells were further incubated in fresh medium containing 10% serum for 3 days. Luciferase gene expression was measured by assaying for enzyme activity in permeabilized cell extracts with the use of a luminometer (14). Transfection using Lipofectin and CaPO 4–DNA coprecipitation was carried out as described by Felgner (27) and by Graham and van der Eb (28), respectively. 36 Chloride efflux assay. Cells were seeded at 30 –50% confluence and transfected as above. Cells were washed 33 with CaMg-free phosphate-buffered saline (Gibco-BRL) to remove serum. Thirty microliters of 36Cl 2 solution (NEN-Dupont; 1 mCi/ml) was diluted in 9 ml of Ringers solution, and 1.5 ml of this loading solution was added to each well of a six-well plate. The plate was incubated for 2– 4 h in a 37°C warm room. The Ringers solution for these experiments was standard HCO 32-free, Hepes- and phosphate-buffered 140 mM NaCl Ringers solution supplemented with 5 mM glucose and titrated to pH 7.45 with 1 N NaOH. All efflux runs were performed in a 37°C warm room. Each well served as its own control. At Time 0, Ringers without cAMP agonists was added and removed immediately. A fresh aliquot of Ringers was added immediately and the efflux run was started. This process was repeated every 15 s until the time point at 1 min, at which time, Ringers with forskolin (2.5 mM), 8 bromo-cAMP (250 mM), and CPT-cAMP (250 mM) was added and removed and the sampling continued for the remaining 4 min of the efflux run. At the end of the run, 0.5 N NaOH was added in two aliquots to lyse and recover all of the cell lysate to determine how much labeled chloride had remained in the cells to standardize the data. Each sample was diluted in scintillation cocktail, counted in a scintillation counter, and normalized on a Microsoft Excel spreadsheet as the rate of labeled chloride lost from the cells per minute.
RESULTS
Complex coacervation of gelatin and DNA to form nanospheres. Coacervation reactions carried out using a range of concentration combinations of DNA, gelatin, and Na 2SO 4 showed that the formation of distinct DNA– gelatin nanospheres occurred over a narrow range of conditions. Diagrams of typical conditions for nanosphere formation with plasmids in the size range of 7 to 12 kbp are shown in Fig. 1. For example, with HIVgp160/LAMP pcDNA1, nanoparticles were formed with concentrations (w/v) of 0.0025 to 0.005% DNA, 1.5 to 5% gelatin, and 0.2 to 1.2% Na 2SO 4. In general, the required concentration of DNA was more confined than those of other components. Outside of these conditions, soluble gelatin–DNA complexes, aggregates, or loosely associated flocculates were obtained. Other conditions important for particle formation include a reaction temperature range of 55 6 7°C, the absence of other buffers or salts in the coacervation reaction, and the extent of mixing—a function both of the speed of vortexing and the dimensions of the reaction vessel. The conditions for coacervation may change if the mixing conditions are modified and if other charged molecules are included or the mixing conditions are modified and must be independently determined for any variation in the reaction protocol.
GENE TRANSFER BY DNA–GELATIN NANOSPHERES
49
FIG. 1. Three-dimensional contour plots of the combination of HIV-1gp160/LAMP DNA, gelatin, and Na 2SO 4 concentrations in the standard reaction mixture that resulted in nanosphere formation. A permutation of coacervation reaction conditions employing different DNA, gelatin, and Na 2SO 4 concentrations [all axis are given as percentages (w/v)] was carried out to determine the reaction conditions required for the formation of distinct, aggregate-free particles. The following concentration range were used: DNA [0, 0.00125, 0.0025, 0.005, and 0.01% (w/v)]; gelatin [0, 1, 2, 4, and 5% (w/v)]; and Na 2SO 4 [1.42, 1.136, 0.0852, and 0.284% (w/v)]. All concentration are final in the reaction mixture. Outside the delineated region, only gelatin–DNA complexes or loosely associated flocculates were observed.
A detailed synthesis procedure for making the optimized nanospheres formulation containing plasmid in the 7- to 12-kb size range is shown in Fig. 2. The formation of uncrosslinked nanoparticles occurred within seconds as gelatin is added to DNA at a high stirring rate. More than 98% of the DNA added to the reaction enters the coacervate. The nanoparticles, which may be viewed by phase-contrast microscopy, are separated from the unreacted components by centrifuging the nanosphere solution on a three-step sucrose density gradient solution. Although relatively stable at acidic pHs, these particles will dissociate in high-ionic-strength solutions at neutral pH, but are stabilized by crosslinking by the use of EDC, a water soluble carbodiimide that conjugates primary amino groups to carboxyl side chains of amino acids. Ligands or other reagents may also be included in the crosslink-
ing reaction at this point, after which sodium acetate or glycine is added to quench any unreacted carbodiimide. DNA crosslinking to itself or to chloroquine under the conditions of the reaction was ruled out because the electrophoretic mobility of DNA remained unchanged from that of untreated DNA when DNA was incubated with EDC or with chloroquine and EDC (Fig. 2b). When EDC was incubated with an equimolar mixture of DNA and gelatin, a DNA/gelatin ratio at which no gel retardation occurred, the electrophoretic mobility of DNA also remained unchanged. The DNA– gelatin complex as visualized by transmission electron microscopy in multiple fields was a spherical particle of solid core structure with a rough surface texture and composed of a matrix of interweaved gelatin and DNA (Fig. 3). Nanospheres made under standard conditions exhibited a mean diameter
50
TRUONG-LE ET AL.
FIG. 2. Schematic of the synthesis procedure for DNA– gelatin nanospheres. (a) Standard synthesis procedures for the optimal formulation of nanospheres. (b) Agarose gel electrophoresis of LAMP-1 pcDNA: lane 1, DNA alone; lane 2, DNA 1 EDC 1 chloroquine; lane 3, DNA 1 EDC 1 gelatin (1:1 mole ratio of DNA to gelatin).
of 412 6 120 nm, as determined by light scattering size analysis, and a zeta potential of 4.5 mV at pH 7.4. The size of the particle was influenced by the temperature of the reaction, the size of the plasmid, the Na 2SO 4 concentration, and the speed of mixing. Micro- as well as nanoparticle size ranges could be produced. In general, nano-size particles were obtained at a temperature of 55°C or lower, a Na 2SO 4 concentration higher than 10 mM, and with smaller plasmids (,12 kb). Protection of DNA against nuclease digestion. A major barrier to gene delivery in vivo is the vulnerability of DNA to nucleases that are found in abundance in serum and the extracellular matrix. A possible advantage of the encapsulated DNA approach to gene delivery is the protection against nucleases. We previously showed that nanosphere-encapsulated DNA was better protected against degradation in serum than was naked DNA (14). In this study the effect of DNase I on nanosphere DNA compared to other forms of DNA complexes was analyzed. Free plasmid DNA, nanospheres, poly-L-lysine–DNA complex (PLL–DNA), and
Lipofectamine–DNA complex were incubated for 15 min with 10 mU/ml DNase I. After incubation, the nanosphere samples were pelleted to separate the particles from the supernatant and the samples were analyzed by agarose gel electrophoresis (Fig. 4). Nanosphere DNA, Lipofectamine–DNA, and, to a lessor extent, PLL–DNA made at a 1:1 weight ratio were partially resistant to degradation during incubation with DNase I for 15 min. In contrast, naked DNA was extensively degraded. This protection of DNA by nanospheres was only partial, since nanospheres incubated in a higher DNase I solution (200 U/ml) showed complete digestion of nanosphere DNA. Interestingly, nanospheres crosslinked with twice the EDC concentration did not differ in DNA protection level (data not shown). The lack of complete protection against nuclease digestion suggests that DNA degradation occurred by a form of surface erosion of DNA that was dynamically mobile in the nanosphere, or, possibly more likely, that the nanosphere was porous enough to allow accessibility of DNase to the particle’s interior, or both.
GENE TRANSFER BY DNA–GELATIN NANOSPHERES
FIG. 3. The gelatin–DNA nanosphere as visualized by electron microscopy. 9HTEo cells were incubated with nanospheres containing pSA306 pDNA (2 mg/well) in complete medium for 15 min. The cells were washed twice with PBS, fixed by incubation with a PBS solution containing 2.5% glutaraldehyde and 2% paraformaldehyde for 30 min at 25°C, washed with 0.1 M cacodylate buffer, incubated with 1% osmium tetraoxide, treated with 1% uranyl acetate, and then dehydrated in graded ethanol. The cells were embedded in Epon, thin sectioned, and viewed on a Jeol electron microscope. The arrow shows a nanosphere at the surface of an 9HTEo cell after 15 min incubation (75,0003). Bar, 100 nm.
Effect of nanosphere calcium on transfection. Divalent metal cations such as Mg 21, Ba 21, and Mn 21 can form ionic complexes with the helical phosphates on DNA (29). Ca 21 exhibits similar affinity to DNA (K d of 1.1 3 10 23 M 21) and forms CaPO 4 complexes with the nucleic acid backbone and thus may impart a stabilization function to certain DNA structures (30, 31). It has been suggested that CaPO 4, when used in complexes with plasmid DNA, exerts its positive effect on gene transfer by stimulating cellular uptake of DNA in a process involving either endocytosis of the membrane-bound DNA complex or enhanced permeabilization of the plasma membrane to facilitate DNA entry (32). In addition, some studies have correlated intracellular elevation of calcium with transcription and gene expression (33). These observations suggest the possible benefit on nanosphere gene transfer by the codelivery of calcium. Transferrin-coated nanospheres added to 293 cells transfected better in the presence of exogenously added CaPO 4 (100 mM) (Fig. 5). Such transfection enhancement required the presence of transferrin because CaPO 4 failed to stimulate transfection when nanospheres alone (without transferrin) were added to 293 cells. Because the CaPO 4 effect on nanosphere-mediated transfection required the presence of transferrin on nanospheres, presumably to initiate surface binding, the results suggested that the
51
added CaPO 4 may act via stimulation of cellular uptake of bound nanospheres. The CaPO 4 effect could be mimicked when the nanospheres were loaded with calcium by incubating newly synthesized transferrincoated nanospheres with a concentrated calcium solution. Transfection of 293 cells with crosslinked nanospheres made with different calcium loading levels showed the highest transfection was achieved with those containing a 1.8% (w/w) calcium loading level. The possibility that the CaCl 2-mediated increase in transfection was due to the extracellular release of DNA and calcium ions from the nanosphere was investigated. In that case, the resultant free DNA and CaPO 4 (formed with PO 42 in the medium) could have complexed so that the enhancement in transfection might have resulted from formation of free DNA– CaPO 4 complexes and not from intact nanospheres. This possibility was tested by measuring transfection of 293 cells by nanosphere supernatant in the presence or absence of CaPO 4 (Table I). The level of transfection
FIG. 4. Protection of DNA against nuclease digestion. Plasmid DNA, DNA– gelatin nanospheres, poly-L-lysine–DNA complex (PLL), or Lipofectamine–DNA complex was incubated with 10 mU/ml of DNase I for 15 min at 37°C. The nanosphere samples were pelleted to separate the nanospheres from the supernatant, the pellet fraction incubated with trypsin to release the DNA, and the resulting DNA from the supernatant and pellet fractions was subjected to 1% agarose gel electrophoresis. The DNA from other samples was extracted twice with phenol:chloroform:isoamyl alcohol (50:49:1) prior to gel electrophoresis. Lanes: 1, untreated control DNA; 2, plasmid DNA treated with 10 mU/ml DNase; 3, Nsp–DNA treated with 10 mU/ml DNase, pellet fraction; 4, Nsp-DNA treated with 10 mU/ml DNase, supernatant fraction; 5, PLL–DNA at 1:1 weight ratio treated with 10 mU/ml DNase; 6, Lipofectamine–DNA at 2:1 weight ratio treated with 10 mU/ml DNase; 7, Nsp–DNA incubated in 200 U/ml DNase I for 15 min at 37°C, pellet fraction. Degradation of DNA was apparent with naked DNA and the PLL–DNA complex treated with 10 mU/ml DNase and with nanosphere DNA treated with 200 U/ml DNase I. In contrast, DNA in nanospheres (lane 3) and Lipofectamine–DNA complex (lane 6) exhibited greater resistance against nuclease degradation at a concentration of 10 mU/ml DNase.
52
TRUONG-LE ET AL.
FIG. 5. The effect of calcium on nanosphere transfection of 293 cells in vitro. Nanospheres synthesized containing pRELuc DNA (Nsp), containing transferrin coating (Nsp/trf ), and containing transferrin and calcium (Nsp/ Trf/Ca) were added to cells in a 12-well dish (2 mg total DNA/5 3 10 4 cells per well) for 4 h, washed to remove excess unbound nanospheres, and then assayed for luciferase expression at Day 3. Nsp 1 CaPO 4 and Nsp/ Trt 1 CaPO 4 represent, respectively, Nsp and Nsp/ Trf samples that were incubated with cells for 2 h, washed, pulsed with CaPO 4 (100 mM) for two additional hours, and assayed for luciferase expression at Day 3. Data are average values from triplicate determinations and are representative of two different experiments, 6SD.
of supernatant from complete, crosslinked nanospheres was barely above background, indicating that there was very little, if any, free DNA in the nanosphere sample. In comparison, cells transfected with supernatant of the uncrosslinked nanosphere sample containing free DNA released from the nanosphere showed a significant level of luciferase activity in the presence of added CaPO 4. It was also possible that proteases in the medium may partially degrade the nanosphere during the period of cell culture medium incubation, releasing free DNA and calcium which
could also form DNA–CaPO 4 complexes. To address this, the nanospheres were incubated with culture medium (without cells) for 4 h, and the resultant medium was used to transfect cells under similar conditions described above for the nanosphere supernatant (Table I). Again, there was a significant level of transfection only with the uncrosslinked nanosphere sample. Higher transfection activity was observed in the uncrosslinked nanosphere group (using CaPO 4 complexation) than in the nanosphere pellet group, which suggested that either crosslinking decreased the transfection efficiency or that the amount of DNA released from the crosslinked nanospheres by Day 3 was small. Thus, we concluded that little DNA was released from the crosslinked nanospheres into the cell culture medium during the time of incubation, and that the transfection can be attributed to intact nanospheres and not to free DNA and calcium contaminants. Transfection with nanosphere DNA encoding the human CFTR protein. Our studies have shown that the optimal nanosphere formulation, as defined in vitro transfection studies carried out on 293 human kidney epithelial cells, contained calcium in addition to chloroquine and surface-bound transferrin (14). Further documentation of the transfection with nanosphere DNA using different cell lines and a therapeutically relevant model gene was carried out. The bioactivity of DNA delivered to cells by nanospheres was studied by transfecting 9HTEo cells with plasmid pSA306 expressing a chimeric human CFTR protein containing a 26-amino-acid epitope tag at the carboxy terminus. Gene expression analysis, as measured by antibody binding to the epitope tag, showed approximately 50 to 60% of the cells incubated with the nanospheres were positive for the expression of the tag (Fig. 6). The frequency of stained cells following nanosphere transfection was significantly greater than that of cells transfected by the equivalent DNA dose given
TABLE I Luciferase activity (light units) a Nanosphere preparation
Nsp pellet
Nsp supernatant
Nsp medium 1CaPO 4
Complete Noncrosslinked
101,800 380
1450 550
30 13,830
1CaPO 4 910 450
4490 418,010
Note. Complete nanospheres (Nsp/trf/CQ/Ca) or noncrosslinked nanospheres (Nsp/CQ) containing the pRELuc plasmid were centrifuged (15,000g, 12 min) to separate the particulate fraction (Nsp pellet) from the supernatant fraction (Nsp supernatant). The nanosphere pellet fraction was resuspended in water and incubated with 293 cells (2 mg DNA/10 5 cells), while the nanosphere supernatant (2 mg DNA/10 5 cells) was incubated with 293 cells in the presence or absence of CaPO 4. In a separate experiment, nanospheres were incubated with fresh culture medium (without cells) for 4 h and then centrifuged to separate the particulate fraction from the supernatant fraction (Nsp medium). The supernatant fraction was subsequently incubated with 293 cells in the presence or absence of CaPO 4. Luciferase assay was carried at Day 3 to determine gene transfection. a Background subtracted.
GENE TRANSFER BY DNA–GELATIN NANOSPHERES
53
FIG. 6. Expression of the CFTR gene in 9HTEo human tracheal epithelial cells. Nanospheres containing pSA306 DNA encoding the CFTR gene with a 26-amino-acid epitope tag were prepared as described under Experimental Procedures. Transfection with the CFTR gene was carried out with 9HTEo cells plated on a cover glass in a 12-well plate. The cells were transfected by the addition of pSA306, 2 mg total DNA/3 3 10 5 cells/well, in nanospheres, as a Lipofectamine complex, or as naked DNA, washed after 4 h, and then incubated for 3 days. The cells were then fixed with 4% paraformaldehyde, blocked by adding 1% bovine serum albumin for 30 min, permeabilized with 0.1% saponin PBS for 10 min, incubated with the anti-epitope tag antibody mAb934 (1:100 dilution) for 1 h, washed three times with PBS, and then incubated with a 1:1000 dilution of fluorescein isothiocyanate-labeled anti-chicken IgG antibody (Jackson Lab, Bar Harbor, ME). The field of view shown contained cells at ;75– 80% confluency (the corresponding bright-field image is not shown). (A) Free DNA transfected cells showed only background cellular autofluorescence; the occasional cells with a higher level of staining were not investigated. (B) Nanospheretransfected cells showed specific staining in approximately 50% of the cells. (C) Lipofectamine–DNA-transfected cells showed staining in approximately 20% of the cells.
as Lipofectamide–DNA complex (.50% vs ;23%). This transfection frequency was the highest observed for any cell types tested using nanospheres as the transfectant agent. Other cells (293 human kidney epithelial, HeLa cells, IB-3-1 human bronchial epithelial, COS-7 monkey cells, and U937 human histiocytic– monocytic cells) typically showed a range of 2–15% transfection frequency. Variations in the transfection frequency were also observed with Lipofectamine– DNA and CaPO 4–DNA transfection methods. None of the nonviral gene transfection methods tested could match the results achieved by use of AAV. The majority of HTE cells infected under similar conditions with 10 9 AAV encoding the CFTR gene were transfected, representing an efficiency based on the amount of DNA that was generally 10 to 100 times greater in cell transfection than that achieved with the DNA plasmid delivery systems (data not shown). Cells transfected with nanosphere DNA encoding the CFTR gene form a functional chloride channel. Additional studies with plasmid pSA306 expressing a chimeric human CFTR protein were conducted with IB3-1 cells, a human bronchial epithelial cell line bearing the delta508 mutation which results in a defective cAMP-stimulated chloride transport phenotype. This system was used to study the efficacy of transfection with DNA– gelatin nanospheres in reconstituting chloride transport, comparing the effect of transfection with free DNA, DNA– gelatin nanospheres, or infection of the cells with an AAV CFTR vector system (Fig. 7). In this standard assay (34), the function of CFTR is assessed by cAMP stimulation of 36Cl 2 efflux and is expressed as the percentage rate of efflux per minute
normalized to 100%, as a function of time after cAMP stimulation. Because the efflux is followed with multiple samples and cAMP is added at the 1-min time period after the beginning of the experiment, each culture dish serves as its own control. Efflux of Cl 2 from the nontransduced cells (background cells), was not stimulated by cAMP. Similarly, in cells treated with free DNA there is no detectable increase in the rate of Cl 2 release following cAMP addition. The small fluctuations in the chloride efflux rate as seen with the nontransduced cells following cAMP addition are attributed to mechanical perturbations in the cells by adding cAMP. In contrast, when cAMP was added to cells transfected with the gene encoding the CFTR carrier by nanosphere DNA or by the AAV CFTR vector, there was a rapid and transient increase in 36Cl 2 efflux, indicative of the action of functional CFTR. In the nanosphere-treated cells, the rate jumped from less than 20% per minute just prior to cAMP addition to approximately 50% per minute after addition of cAMP. Likewise, in the AAV CFTR-treated cells the rate increased from about 20% per minute prior to 40% after. This rapid and transient increase in efflux demonstrated that CFTR was functioning in the transduced cells compared to nontransfected controls. Because the absolute magnitude of the response depends not only on the expression of CFTR in an individual cell but also on the number of cells transduced, a quantitative comparison of the absolute magnitude of the response among groups is not possible. It may be noted, however, that a comparable result in this assay was obtained with 10 9 AAV, which contains much less DNA than the 5 mg added as nanosphere DNA. We previ-
54
TRUONG-LE ET AL.
ously have shown that the transduction efficiency of nanosphere DNA in vitro, while greater than that of naked DNA, was 10 to 1000 times less than that of AAV virions (14). DISCUSSION
FIG. 7. 36Chloride efflux assay of chloride transport across human bronchial epithelial IB-3-1 cells transfected with pSA306 DNA as free DNA, nanosphere–DNA, or AAV CFTR viral particles. The IB3-1 cells are a human CF bronchial epithelial cell line with defective transport of Cl 2. pSA306 DNA (5 mg total DNA/5 3 10 5 cells/per 35-mm well of a multiwell plate), nanospheres containing pSA306 DNA (5 mg total DNA/5 3 10 5 cells/well), or AAV CFTR virions (10 9 viral particles/5 3 10 5 cells/well) were incubated with IB-3-1 cells for 4 h as described above and then analyzed for chloride efflux at Day 3 as described in detail elsewhere (34). The transfected cells were labeled with 36Cl 2 for 4 h, washed to remove excess 36Cl 2 (t 5 0 is the time after the last wash) and loaded with cAMP at t 5 1 min to activate the CFTR 36Cl 2 current. The medium was sampled for 36Cl 2 effluxed from the cells at various time points. The data for each experiment are averages of six separate assays from six individual dishes of cells run on the same day and are representative of two experiments done on two separate occasions. The results are expressed as the percentage rate of 36Cl 2 efflux per minute normalized to 100% and plotted as a function of time. Bkgd, background 36Cl 2 currents in 36chloride-loaded, untransfected cells.
Complex coacervation has been widely used to formulate controlled release microspheres. We had previously described the use of gelatin– chondroitin sulfate and collagen– chondroitin sulfate complex coacervates to deliver therapeutic agents (26, 34). In this system, DNA and gelatin are used as the oppositely charged biopolymers. Since there was no evidence of crosslinking of the DNA to gelatin or to any other nanosphere components, the nanosphere can be thought of as a solid core structure composed of gelatin entangled with DNA. The spherical nature of the coacervate is a result of the gelatin–DNA complex adopting the minimal energy architecture in a process governed by the relative interfacial tension. One of the advantages of the nanosphere approach to gene delivery is the enhanced protection of DNA against nuclease degradation, which should improve in vivo bioavailability of the DNA. Our data showed the significant improvement in DNA protection when the nanospheres were incubated with a low concentration of DNase I (10 mU/ml) compared to the rapid degradation of free DNA and a DNA–PLL complex. Higher degrees of nanosphere crosslinking further stabilized the gelatin matrix, which in turn improved the half-life of DNA treated with DNase; however, at high crosslinking density, the transfectability of the nanosphere was compromised, possibly because of a reduced rate of release of the DNA (unpublished results). In addition to increasing the crosslinking density of the gelatin matrix, we have attempted to increase the resistance of nanosphere DNA to DNase by coencapsulating DNase inhibitors. Encapsulation of sodium iodoacetate (35) and aurintricarboxylic acid (36), both known to be potent DNase I inhibitors, showed a minor enhancement in DNA protection, but failed to show consistent improvement in transfection level (data not shown). Other materials added to the coacervation reaction can be coencapsulated in the nanosphere by either covalent conjugation, entrapment, or ionic interactions with DNA or gelatin. However, these nanospheres are permeable to low-molecular-weight compounds as shown by the partial loss of chloroquine and calcium during dialysis. This suggests that larger and more charged compounds should enjoy a higher loading level. In principle, this may include peptides, proteins, and a variety of smaller but charged compounds. Chloroquine was added to enhance transfection by DNA, presumably by interfering with the acidification of en-
GENE TRANSFER BY DNA–GELATIN NANOSPHERES
dosome compartments (37) or reducing DNase degradation (38). A chloroquine loading level of 2% (w/w) into the nanospheres was achieved by reducing the Na 2SO 4 concentration 10-fold in the coacervation reaction. This high uptake of chloroquine was attributed to its affinity for polynucleotides (K d of 10 24 M 21) via intercalation in the DNA helix and electrostatic interactions with the phosphate backbone (39). The level of transgene expression in 293 human kidney epithelial cells incubated with nanospheres was also markedly increased by the presence of transferrin on the nanosphere and to a lesser extent by calcium. In the absence of transferrin there was no detectable transfection of 293 cells with as much as 10 mg total DNA either as gelatin–DNA nanospheres or gelatin–DNA complexes. Transferrin has been demonstrated to be efficient in mediating binding and uptake of DNA–polymer or –protein complexes in several studies (40), presumably by acting as a ligand to the corresponding cell surface receptor and thereby enhancing the binding and/or the uptake of the nanosphere into the cell. Transfection enhancement by calcium was attributed to calcium that was associated with the nanospheres and not to those resulting from leakage. Our data also suggested that a possible role of calcium is to facilitate release of DNA from the gelatin matrix by its competition for electrostatic interactions with the gelatin and DNA. However, other explanations that account for the enhancement in transfection such as stimulated uptake of nanospheres, perhaps as a result of the formation of CaPO 4 complexes on the nanospheres, or specific intracellular effects are also plausible. These in vitro transfection experiments using the CFTR gene have shown that a large percentage of 9HTEo cells could be transfected by use of the DNA– gelatin nanospheres. In addition, successful complementation of defective chloride transport in mutant IB-3-1 cells was demonstrated. While similar results have been accomplished using a variety of viral and nonviral gene delivery methods in vitro (41), the therapeutic advance to successful in vivo CF gene therapy remains a great challenge. Our studies have shown that under in vitro cell culture conditions, both gene transfer efficiency and efficacy were poorer than with the AAV viral counterpart. However, the microencapsulated DNA approach may offer some advantages that are unique compared to other approaches, such as targeted gene delivery by the use of monoclonal antibodies or cell ligands and codelivery of therapeutic compounds in a controlled fashion. Studies of the delivery of the CFTR gene to rabbit airways by use of nanosphere– DNA have shown a correspondingly high efficiency of transfection of airway epithelial cells as was observed in these in vitro studies, and an evaluation of the codelivery of sodium 4-phenylbutyrate, an agent that has been found to enhance the CFTR function by stim-
55
ulating the mutant DF508 CFTR activity (42), is underway. ACKNOWLEDGMENTS This study was supported by NIH CA 68011, NIH 1 RO1 A141908, and the Cystic Fibrosis Foundation.
REFERENCES 1. Ledley, F. D. (1994) Curr. Opin. Biotech. 5(6), 626 – 636. 2. Tomlinson, E., and Rolland, A. P. (1996) J. Controlled Rel. 39, 357–372. 3. Lee, R. J., and Huang, L. (1997) Crit. Rev. Ther. Drug Carrier Sys. 14(2), 173–206. 4. Hong, K., Zheng, W., Baker, A., and Papahadjopoulos, D. (1997) FEBS Lett. 400(2), 233–237. 5. Gottschalk, S., Sparrow, J. T., Hauer, J., Mims, M. P., Leland, F. E., Woo, S. L., and Smith, L. C. (1996) Gene Ther. 3(5), 48 –57. 6. Niidome, T., Ohmori, N., Ichinose, A., Wada, A., Mihara, H., and Hirayama, T. (1997) J. Biol. Chem. 272(24), 15307–15312. 7. Boussif, O., Lezoualch, F., Zanta, M. A., Mergny, M. D., Scherman, D., Demeneix, B., and Behr, J. P. (1995) Proc. Natl. Acad. Sci. USA 92(16), 7297–7301. 8. Haensler, J., and Szoka, F. C. (1993) Bioconj. Chem. 4, 372–379. 9. Goldman, C. K., Rogers, B. E., Douglas, J. T., Sosnowski, B. A., Ying, W., Siegal, G. P., Baird, A., Campain, J. A., and Curiel, D. T. (1997) Cancer Res. 57(8), 1447–1451. 10. Plank, C., Oberhauser, B., Mechtler, K., Koch, C., and Wagner, E. (1994) J. Biol. Chem. 269, 12918 –12924. 11. Noguiez-Hellin, P., Meur, M. R., Salzmann, J. L., and Klatzmann, D. (1996) Proc. Natl. Acad. Sci. USA 93(9), 4175– 4180. 12. Johanning, F. W., Conry, R. M., LoBuglio, A. F., Wright, M., Sumerel, L. A., Pike, M. J., and Curiel, D. T. (1995) Nucleic Acids Res. 23(9), 1495–1501. 13. Truong-Le, V. L., Mao, H-Q., Walsh, S. W., Leong, K. W., and August, J. T. (1997) Proc. Intern. Symp. Control Rel. Bioact. Mater. 24, 75–76. 14. Truong-Le, V. L., August, J. T., and Leong, K. W. (1998) Hum. Gene Ther. 9, 1709 –1717. 15. Marty, J. J., Oppenheim, R. C., and Speiser, P. (1978) Pharm. Acta Helv. 53, 17–23. 16. Bugenberg de Jong, H. G. (1949) in Colloid Science (Kruyt, H. R., Ed.), Vol. II, pp. 232– 480, Elsevier, New York. 17. Burgess, D. J. (1990) J. Colloid Interfacial Sci. 140, 227–239. 18. Khalil, S. A. H., Jr, and Carless, J. E. (1981) J. Pharm. Pharmacol. 20, 6978 – 6987. 19. Leong, K. W., Mao, H.-Q., Truong-Le, V. L., Roy, K., Walsh, S. W., and August, J. T. (1998) J. Contr. Rel. 53, 183–193. 20. Wu, P., and Schurr, J. M. (1989) Biopolymers 24, 1695–1703. 21. Flotte, T. R., Afione, S. A., Solow, R., Drumm, M. L., Markakis, D., Guggino, W. B., Zeitlin, P. L., and Carter, B. J. (1993) J. Biol. Chem. 268(5), 3781–3790. 22. Guarnieri, F. G., Arterburn, L. M., Penno, M. B., Cha, Y., and August, J. T. (1993) J. Biol. Chem. 268(3), 1941–1946. 23. Ruff, A. L., Guarnieri, F. G., Staveley-O’Carroll, K. F., Siliciano, R. F., and August, J. T. (1996) J. Biol. Chem. 272(13), 8671– 8678. 24. Egan, M., Flotte, T., Afione, S., Lolow, R., Zeitlin, P. L., Carter, B. J., and Guggino, W. B. (1992) Nature 358, 581–584. 25. Flotte, T. R., Afione, S. A., Conrad, C., McGrath, S. A., Solow, R., Oka, H., Zeitlin, P. L., Guggino, W. B., and Carter, B. J. (1993) Proc. Natl. Acad. Sci. USA 90(22), 10613–10617.
56
TRUONG-LE ET AL.
26. Truong, V. L., Williams, J. R., Hildreth, J. E. K., and Leong, K. W. (1995) Drug Deliv. 2, 166 –174. 27. Felgner, P. L., Gadek, T. R., Holm, M., Roman, R., Chan, H. W., Wenz, M., Northrop, J. P., Ringold, G. M., and Danielsen, M. (1987) Proc. Natl. Acad. Sci. USA 84, 7413–7417. 28. Graham, F. L., and Van Der Eb, A. J. (1973) Virology 52, 456 – 467. 29. Shack, J., Jenkins, R. J., and Thompsett, J. M. (1953) J. Biol. Chem. 203, 373–387. 30. Wiberg, J. S., and Neuman, W. F. (1957) Arch. Biochem. Biophys. 72, 66 – 83. 31. Herzog, M., and Soyer, M. O. Eur. J. Cell Biol. 30(1), 33– 41. 32. Orrantia, E., and Chang, P. L. (1990) Exp. Cell Res. 190(2), 170 –174. 33. Negulescu, P. A., Shastri, N., and Cahalan, M. D. (1994) Proc. Natl. Acad. Sci. USA 91, 2873–2877. 34. Devidas, S., Yue, H., and Guggino, W. B. (1998) J. Biol. Chem., in press.
35. Shao, W., and Leong, K. W. (1995) J. Biomaterials Sci. Polymer Ed. 7(5), 389 –399. 36. Nagae, S., Nakayama, J., Nakano, I., and Anai, M. (1982) Biochemistry 21, 1339 –1344. 37. Hallick, R. B., Chelm, B. K., Gray, P. W., and Orozco, E. M. (1977) Nucleic Acids Res. 4(9), 3055–3064. 38. Pless, D. D., and Wellner, R. B. (1996) J. Cell. Biochem. 62, 27–39. 39. Kurnick, N. B., and Radcliffe, M. S. (1962) J. Lab. Clin. Med. 60, 669 – 688. 40. Cohen, S. N., and Yielding, K. L. (1965) J. Biol. Chem. 240, 3123–3131. 41. Cotten, M., and Wagner, E. (1993) Curr. Opin. Biotechnol. 4, 705–710. 42. Flotte, T. R., and Carter, B. J. (1997) Adv. Pharmacol. 40, 85–101. 43. Walsh, S. W., Flotte, T. R., Truong-Le, V. L., Rubenstein, R., Beck, S., August, J. T., Zeitlin, P., and Leong, K. W. (1997) Proc. Int. Symp. Contr. Rel. Bioact. Mater. 24, 39 – 40.