|Home | About | Journals | Submit | Contact Us | Français|
A new method for biolistic delivery of nucleic acids using a combination of cationic micro- and nanoparticles is reported. The new method is simpler to perform than the conventional calcium/spermidine-based formulations and shows eleven-fold improved nucleic acid binding capacity and dose-dependent performance both for in vitro and in vivo applications relative to either the conventional preparation or our recently reported direct cationic microparticle method. These features may enable higher throughput gene delivery and genetic immunization programs and open new venues for biolistic delivery method.
Gene-based therapeutics, such as genetically-modified plants, gene therapy and genetic immunization, holds great promise for improving crops and treating diverse diseases, ranging from inherited disorders to cancer. A major obstacle to their effectiveness remains the efficient introduction of bioactive nucleic acids (NA) into live tissue.1 The two most common vehicles for NA delivery are virus and virus-like-particles (VLPs). Practical applications of viral particle methods are limited by the difficulty of their preparation, very poor storability, and lack of tissue tropism.2 Among non-viral delivery methods, biolistics (biological ballistics) is unique because it avoids the perils of NA passage through medium (in vitro) or extracellular space (in vivo) by physically propelling it directly across the cellular membranes. Pioneered by J. Sanford and his colleagues over two decades ago,3 this method is broadly used for stable and transient transformation of cells,4 organelles,5 and organisms.6 For plastids and a number of cell and tissues, biolistic delivery is the only known route of successful NA administration. Biolistic delivery is also an important genetic immunization route with broad potential for human and animal applications. Clinical trials have shown that this approach elicits both humoral and cellular immune responses, and is one of the most consistently successful methods for delivering gene vaccines.7
The biolistic method is based on NA release from NA-coated gold microparticles kinetically propelled inside of the cells with a high-pressure blast. A number of the particle delivery devices, or gene guns, have been described to date and one is commercially available from Bio-Rad Laboratories (Helios® gene gun). Surprisingly, for the past two decades most of the efforts to optimize gene gun delivery have been focused on the refinement of the device itself without much attention paid to the microparticles that actually deliver the payload. The conventional biolistics protocol involves coating of spherical gold microparticles (1-2μm in diameter) with loose calcium-DNA-spermidine precipitate.4 The coating process is cumbersome, requires high degree of skill, and is highly sensitive to environmental factors and characteristics of the NA, particularly its size and purity. Namely, the conventional protocol cannot be applied to attach NA molecules smaller than 200bp, such as siRNA, because spermidine does not efficiently condense these smaller NA fragments. These disadvantages result in inconsistent and suboptimal performance of the biolistically delivered material and limit the scope of applications.
Recently, we reported a new method for preparation of gene gun microparticles in which gold microparticles are coated with positively charged polyethylenimine (PEI) polymer and DNA is directly bound to the surface. The PEI-coated gold microparticles (PEI-AuMP) were constructed by conjugating amine groups of PEI polymer to the carboxyl moiety of tiopronin linker self-assembled at the gold surface via thiol functionality. These new microparticles exhibited a DNA dose-dependent increase in luciferase (Luc) transgene activity.8
In a further attempt to increase the efficiency of this intercellular delivery vehicle, we argued that it may be advantageous to enable release of NA from the surface of PEI-AuMP in a form of a complex with nanoparticles (nanoplex) within the cytoplasm, which might serve not only as protection against degradation,10 but possibly assist DNA transfer into the nuclei by other nuclear localization mechanisms. We further hypothesized that some portion of the particles are likely to not directly enter the cytoplasm. These extracellular nanoplexes might dissociate from the microparticles and undergo endocytosis by the surrounding cells, thereby providing another opportunity for transfection. We tested the above concepts with the micronanoplexes (complexes of microparticles with nanoplexes) formed between cationic PEI-AuMP microparticles and anionic nanoplexes formed by complexation of DNA with cationic gold nanoparticles (AuNPs).
Unless otherwise noted, all chemicals were purchased from Sigma, Inc (Milwaukee, WI) and used without further purification. Spherical gold microparticles (d=1.5μm) were from Ferro, Inc (cat# J5G2000). Transmission electron microscopy (TEM) was done on Philips CM12S microscope. Scanning electron microscopy was performed on Leica-Cambridge 360FE microscope. In solution nano-sizing and zeta potential measurements were done on Zetasizer Nano-ZS instrument (Malvern Instruments, UK). All spectrophotometric measurements were carried out on NanoDrop® ND-1000 instrument.
The positively charged microparticles were prepared according to a slightly modified and more efficient procedure than the previously reported by us.8 The commercial grade gold microparticles (AuMP, d=1.5μm, Ferro, Inc) (20g) were cleaned in 40 mL of piranha solution (H2SO4:H2O2 = 3:1; Caution! Piranha solution is a corrosive and strongly oxidizing agent) for 1 hour and washed 4× with ddH2O and 2× with ethanol. A solution of 500mg 11-mercaptoundecanoic acid (catalog# 450561-5G, Sigma-Aldrich, Inc.) in 20mL of ethanol 200 proof was added and particles were shaken at 1400rpm at RT for 2 hours. The supernatant was withdrawn after brief centrifugation and the particles were washed 2× with ddH2O. The modified particles were reacted with 200mg of 1-ethyl-3-[3[dimethylaminopropyl]carbodiimide hydrochloride (EDC, catalog# 22980, Pierce, Inc) plus 300mg N-hydroxysuccinimide (NHS, catalog# 130672-5G, Sigma-Aldrich, Inc.) for 30 minutes with vigorous shaking at RT. The excess of reagents was removed by centrifugation/decantation. A solution of 2g of 50% polyethylenimine (PEI750, MW 750K, catalog# P3143, Sigma-Aldrich, Inc.) in 20ml of ddH2O adjusted to pH 9 was added and reaction allowed to react for 2 hours with vigorous shaking (1400rpm) at RT. The resulting PEI-modified microparticles were washed 3× with ddH2O, dried under vacuum overnight and stored in a dark cool place under nitrogen until further use. The presence of organic layer was confirmed directly by FT-IR and indirectly by DNA binding experiments.8
The nanoparticles were prepared according to a literature procedure9 with slight modifications. Briefly, 400μL of a freshly-made 213mM solution of cysteamine (catalog# 30070, Sigma-Aldrich, Inc) in argon-purged ultrapure water was added to 40mL of a solution of 1.42mM tetrachloroauric(III) acid (HAuCl4, catalog# 254169-500MG, Sigma-Aldrich, Inc) dissolved in argon-purged ultrapure water. The mixture turned yellow and translucent upon addition of cysteamine and was stirred vigorously for 10 min. A freshly prepared solution (10μL) of 10mM sodium borohydride (NaBH4) was quickly injected into the reaction mixture with a micropipette. After vigorous stirring for 15-30 min, the wine red solution was stored in the dark at 4°C under nitrogen blanket for up to two weeks.
A solution of 8 μg of pLUC DNA in water was mixed with 450μl of the AuNP solution (60μg/mL in Au) to give the ratio of Au/DNA = 20:1. The mixture instantly changed color from red to purple indicating formation of DNA-AuNP nanoplexes. A suspension of 8mg of PEI-AuMP microparticles in 48μL of pH6 MES (2-(N-morpholino)ethanesulfonic acid) buffer supplemented with 0.5M NaCl was added in 5 minutes. The color of the mixture rapidly cleared on gentle vortexing, indicating that the DNA-AuNP nanoplexes entirely settled onto the surface of PEI-AuMP microparticles. The suspension was briefly centrifuged at 4,000 rpm to a pellet and the supernatant discarded. The pellet was re-suspended in 165μl of n-butanol and cast into 8 bullets in Teflon® tubing which were dried under gentle stream of nitrogen. Caution! The n-butanol drying and shooting regimens should be performed in a chemical hood and while wearing an appropriate facial mask, respectively. The n-butanol and the microparticles dust may cause irritation of respiratory tract if inhaled.
NIH 3T3 cells were split into the 24-well tissue plate. The cells were grown to near confluence (~80%). Just before shooting the NIH 3T3 cells, the media was removed from the wells. The media was immediately added to the wells after the shooting. As a positive control NIH 3T3 cells were transfected with the pLuc plasmid using FuGENE 6 transfection reagent from Roche Applied Sciences (1μg of DNA per one well of the 24-well tissue plate). The cells were incubated for 24 hours (37°C, 5% CO2) and subjected to the Luciferase assay.
Two groups of mice (n=3) were shot with pLuc loaded microparticles in the ears. The ears were harvested after 24 hours and ground up in a manual homogenizer. The homogenized tissues were treated with 0.5ml of 1× reporter lysis buffer. The samples were vortexed for 15 min at room temperature and spun at 14,000 rpm for 10 min. The supernatants were assayed by using luciferase assay.
Assay for Luciferase activity was done using Promega's Luciferase Assay System. The media was completely removed from the wells and the cells were washed one time with 1× PBS (pH 7.4). The cells were covered with 1× Luciferase Cell Culture Lysis Reagent (Promega, Inc.) (500μl of the Luciferase Cell Culture Lysis Reagent per each well of the 24-well tissue plate). Then the cells and all liquid were transferred to a microcentrifuge tube. The tubes were placed on ice. 20μL of cell lysate were added per one well of the 96-well microplate, and the plate was placed into the luminometer with injector (Clarity™ Luminescence Microplate Reader, BioTek Instruments, Inc., Winooski, VT). The injector added 200μl of Luciferase Assay Reagent per well, then the well was read immediately. The light intensity of the reaction was measured for a period of 10 seconds. 1× Luciferase Cell Culture Lysis Reagent was used for the blank (20μl per each well of the 96-well microplate) in triplicate. All assays of Luciferase activity were done in triplicate as well.
Cationic AuNPs have been explored as non-viral gene delivery vehicles to deliver oligonucleotides, plasmids, and siRNA into cells.11 AuNPs are non-toxic, non-immunogenic, and offer advantages of easy preparation and multiple possibilities for further surface modifications. DNA is first adsorbed onto cationic gold nanoparticles in certain DNA/Au ratios to provide charged nanoparticle-DNA complexes (nanoplexes). By using such formulations, the transfection efficiencies could be increased many-fold compared to the introduction of DNA alone or complexed with other transfection reagents.9,12,13 Depending on the DNA/Au ratio, nanoplexes can be negatively or positively charged. In particular, we used negatively-charged nanoplexes that can be efficiently absorbed by the positively-charged PEI-AuMPs to form microparticle-nanoplex complexes (micronanoplexes), as schematically shown in Figure 1.
We prepared cationic cysteamine-modified AuNPs according to a literature procedure and followed process of their complexation with pLUC plasmid DNA by dynamic light scattering (DLS) and ζ-potential (surface charge) measurements in real time. The titration of cationic AuNPs (ζav=+46mV) with DNA (ζav=-47mV) was conducted until the reversal of the positive zeta potential to a value of ζav=-21mV was observed at the target 20:1 Au/DNA ratio (Figure 2). Figures 2A, B show the change in hydrodynamic diameter of the gold nanoparticles and gold nanoparticles complexed with pLUC plasmid. While the AuNPs on average are 36 nm in diameter, the nanoplexes are between 300-400 nm. Varying Au/DNA ratio from 1:1 to 80:1 led to almost linear increase in ζ-potential of the nanoplexes from -33mV to -5mV. The decreasing surface charge in turn resulted in sharp increase in the size of nanoplexes at approximately 60:1 ratio and above, which is in accordance with diminishing colloidal stability when surface charge approaches zero.14
The complexation of the AuNPs with DNA is optically manifested by a distinct blue shift of the surface plasmon band, a phenomenon associated with dispersion to aggregation transition of AuNPs.15 Figure 3A shows transmission electron microscopy (TEM) image of a typical nanoplex formed in the mixture of AuNPs with DNA. The dimension of the nanoplex agrees with the results of the DLS measurements. When the negatively charged nanoplexes are brought in contact with positively charged gold microparticles, the electrostatic attraction causes nanoplexes to be absorbed onto the surface of the microparticles creating micronanoplexes shown in Figure 3B, C. The absorption process is conveniently followed by rapid disappearance of the purple nanoplexes upon addition of the suspension of cationic gold microparticles (Figure 1). The color of the suspended microparticles changes from light yellow to dark purple as the complexation proceeds. Treatment of these micronanoplexes with 0.1M NH4OH solution releases the DNA and reconstitutes gold nanoparticles to their initial non-aggregated state (red color). A brief sonication is necessary in some cases to accelerate the dissociation process.
One of the most important advantages of using micronanoplexes is the ability to significantly increase DNA binding capacity. The increased capacity is attributed to the following two factors: (1) the larger surface area of nanoparticles provides more ample scaffolding for attaching more DNA and (2) the reduction of overall negative charge of the nanoplexes versus that of naked DNA allows an increased electrostatic absorption of nanoplexes per same PEI-AuMP surface charge density. The amount of DNA bound to the gold can be calculated spectrophotometrically as a difference between the input amount of NA and the amount of NA remaining in the supernatant following incubation.8 In the case of the micronanoplex formulations, this solution measurement is not possible because of the strong residual absorbance of by-products of gold nanoparticles synthesis.
To determine the DNA binding capacity of the microparticles, we measured the DNA remaining in the supernatant following incubation by gel electrophoresis. One milligram of microparticles was incubated with 1 to 16μg of DNA complexed with a fixed amount (56μL) of AuNPs. Once the micronanoplexes were formed and precipitated by centrifugation, the supernatant was assayed on an agarose gel to visualize unbound DNA. Figure 4A shows that when PEI-AuMPs were incubated with up to 11μg of DNA, no DNA was found in the supernatant. When 12μg of DNA was added to the microparticles, a small fraction of DNA remained in the supernatant. This indicates that the binding capacity of the PEI-AuMPs is ~11μg of DNA per milligram, which is significantly higher than either the conventional (1μg/mg) or directly-loaded PEI-AuMPs (3μg/mg).8 The binding was stable at hot and cold temperatures and did not dissociate by washing in water, ethanol, or n-butanol but, as mentioned above, could be reversed by treatment with dilute ammonia solutions at pH>9. In this range of pH the amino groups on the surface of PEI-AuMP microparticles become deprotonated and lose their binding affinity to DNA.
Conventionally loaded biolistic microparticles show dose-dependent expression of transgenes when loaded with 1μg or less of DNA, and linear dose responses are lost with any load above 0.5 μg. Loading DNA directly onto PEI-AuMPs represented a significant improvement by providing linear dose-dependent responses for particle loads up to 3μg.8 The higher binding capacity of the PEI-indirect microparticles described here suggests that they may provide further improvement in the range of dose-dependent responses. To test this, NIH 3T3 cells were transfected by gene gun using PEI-indirect microparticles loaded with increasing amounts of a luciferase reporter gene plasmid. Cells were harvested 24 hours later and assayed for luciferase activity. Figure 4B shows that there was a dose-dependent relationship between luciferase activity and transgene dose up to 12μg, which was the highest dose tested. Curve fitting of the data shows sigmoid dependence with increasing load, suggesting that saturation of cellular machinery is reached at around 10-12μg DNA. The intermediate rapid growth in luciferase activity can be explained by increasing size of the nanoplexes, which are easier released from the surface of the microparticles due to their decreasing surface charge density. These important findings highlight the fact that cellular gene expression machinery is far from being saturated by the currently used protocols.
Another set of experiments was performed in mice, using a linear expression construct expressing the same reporter gene. DNAs CMVi promoter, the LUC gene open reading frame (ORF), and the human growth hormone terminator were individually generated by PCR and assembled into an expression cassette by the early published procedure.16 This DNA was loaded onto the AuMP by using (1) the conventional calcium/spermidine protocol, (2) directly onto PEI-AuMP particles, and (3) indirectly using the micronanoplex formulation. These three formulations were delivered with the gene gun into mouse ear pinnas (1 shot per ear). Luc activity in the ear tissue was measured 24 hours later. Reporter gene activity measured in tissue shot with the micronanoplexes was remarkable eleven-fold higher than that of tissue shot with conventional AuMPs and nearly two-fold higher than that of ears shot with the previously reported directly loaded PEI-AuMP (Figure 5A).8 These results highlight the utility of charged microparticles for the delivery of smaller DNA fragments, such as linear expression elements (LEEs)16 or siRNA,8 where the conventional calcium/spermidine protocol fails to efficiently precipitate such fragments onto the surface of uncharged microparticles.
We also followed the durability of LUC gene expression following conventional and micronanoplex particle delivery to mice ear pinnas. Transgenes delivered with the micronanoplexes showed higher and more prolonged expression levels than those delivered by the conventional protocol (Figure 5B).
Finally, we evaluated the ability of the nanoplexes to produce secondary transfection by non-classical nuclear localization. We applied conventional calcium/spermidine/pLUC precipitates on uncharged AuMPs, AuNP/pLUC nanoplexes, AuNP/pLUC/PEI-AuMP micronanoplexes used in the biolistic delivery experiments, and control Lipofectamine® pLUC lipoplexes directly to dendritic cell cultures and saw no Luc expression in all cases (data not shown). This is not surprising since transformation of dendritic cells by using non-viral methods in general has proven to be difficult.17 In a control experiment with NIH 3T3 cells, only Lipofectamine® transfection produced moderate Luc activity. Based on these results, we conclude that biolistic delivery using micronanoplexes does not provide benefits of secondary transfection, instead the higher and more prolonged reporter gene activity are attributed to better nucleic acid protection against degradation using micronanoplex formulations than in the case of conventional calcium/spermidine/NA formulations.
At this point, the mechanism of intracellular DNA release from the micronanoplexes remains unclear and is in need of further elucidation. We can only speculate that the release of DNA-AuNP nanoplexes from the surface of PEI-AuMPs occurs followed by simultaneous or delayed free DNA release by a mechanism similar to the mechanism of DNA release from other nano-, poly-, and lipoplexes in the traditional chemical transfections or from histone-DNA complexes in nucleosomes.
In summary, we developed a new set of formulations for biolistic delivery of nucleic acids using cationic gold microparticles. The advantages offered by these formulations may open new possibilities for various applications of gene delivery such as genetic modifications of plants or immunization of animals. In addition to gold nanoparticles, the cationic surface of the microparticles can be used for the attachment of other anionic NA complexes, including polyplexes and lipoplexes currently used in chemical transfections. Layer-by-layer (LbL) alternating charge assembly using anionic, e.g. DNA, hyalouronate or alginate, and cationic, e.g. PEI, chitosan and polylysine, polyelectrolytes may enable new gene delivery applications where slow or controlled release is of critical importance. Furthermore, the outer layer polymers can be covalently modified to include specific membrane targeting agents in order to enhance the possible post-biolistic delivery and uptake of NA complexes. Finally, the technique described here offers unique opportunities for biolistic delivery of other charged species including proteins and polysaccharides.
We thank Drs. Sergey Sheleg, Jose Cano Buendia and Ms. Bee-Lian Quah for helping with the experiments. Mr. David Lowry from School of Life Sciences Bioimaging Facility at Arizona State University is acknowledged for TEM and SEM technical assistance. The work was funded by NIH grant AI057156 to KS.