Ultrasound in Med. & Biol., Vol. 28, No. 6, pp.

817 822, 2002 Copyright 2002 World Federation for Ultrasound in Medicine & Biology Printed in the USA. All rights reserved 0301-5629/02/$see front matter

PII: S0301-5629(02)00518-5

Original Contribution
DNA-LOADED ALBUMIN MICROBUBBLES ENHANCE ULTRASOUNDMEDIATED TRANSFECTION IN VITRO PETER A. FRENKEL,* SHUYUAN CHEN,* TO THAI,* RALPH V. SHOHET* and PAUL A. GRAYBURN*
*Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX, USA; and VA Medical Center, Dallas, TX, USA
(Received 17 October 2001; in nal form 27 March 2002)

AbstractUltrasound (US), with or without microbubbles, enhances gene transfer in cultured cells, but the effect is modest. We tested if attaching DNA to albumin microbubbles during bubble synthesis could enhance gene expression. Plasmid DNA was loaded on the albumin shell over a range of concentrations (500 to 10,000 g/mL). Optimal gene expression occurred with loading doses of 4000 g DNA/mL (4k-loaded bubbles). These microbubbles had diameters of 2.4 0.7 m and carried 40 pg DNA/microbubble. DNA-loaded microbubbles had optimal transfection at higher delivered doses of DNA than unloaded bubbles mixed with plasmid. The 4k-loaded bubbles demonstrated a vefold (p 0.0003) increase in luciferase reporter expression over that with unloaded bubbles. Similarly, transfection efciency was better for 4k-loaded microbubbles than unloaded microbubbles (41 3% vs. 9 3%, p < 0.0001). DNA loading of microbubbles enhances gene expression and transfection efciency in US-targeted transfection in vitro and may represent an improved avenue for therapeutic gene delivery in vivo. (E-mail: peter.frenkel@utsouthwestern.edu) 2002 World Federation for Ultrasound in Medicine & Biology. Key Words: Ultrasound, Microbubbles, Gene transfer, Ultrasound-targeted microbubble destruction.

INTRODUCTION The use of ultrasound (US)-mediated microbubble destruction as a gene therapy tool holds considerable promise. US by itself has been shown to enhance gene transfer in vitro (Bao et al. 1997; Unger et al. 1997) and in vivo (Miller et al. 1999; Huber and Psterer 2000). The mechanism by which this occurs is incompletely understood, but may include cavitation (Koch et al. 2000), and other effects that may cause sonoporation, a term that describes the formation of transient openings in the cell membranes; thereby, allowing entry of protein and other macromolecules. This latter phenomenon is evidenced in vivo by the fact that destruction of gas-lled microbubbles by US causes pericapillary extravasation of red blood cells with delivery of microbubble fragments and colloid particles deep into the pericapillary tissues (Price et al. 1998; Skyba et al. 1998). Our laboratory has taken advantage of the destructive effects of US on albumin microbubbles to deliver
Address correspondence to: Peter A. Frenkel, M.D, Cardiology, 111-A, Dallas VA Medical Center, 4500 S. Lancaster Road, Dallas, TX 75204 USA. E-mail: peter.frenkel@utsouthwestern.edu 817

and express reporter genes in the heart using an adenoviral transgene (Shohet et al. 2000). Unfortunately, adenovirus is an imperfect vector because it causes an immune response and hepatotoxicity. Although we have successfully delivered plasmid DNA to the heart (Chen et al. 2001) using US-targeted microbubble destruction (UTMD), we observed a substantial amount of interanimal variability in gene expression. We believed that part of this variability was due to poor colocalization of the genetic material with the microbubble. We, therefore, have tested if we could improve transfection by attaching DNA to the bubble. Here, we demonstrate optimized delivery of plasmid DNA by UTMD. Specically, we have developed a method of loading albumin microbubbles with plasmid DNA. Our approach had two purposes: 1. to increase the total amount of DNA delivered by the microbubbles, and 2. to ensure that the DNA is colocalized to the site of microbubble destruction. METHODS Plasmid DNA, DNA synthesis and DNA analysis SV-Luc is a plasmid carrying a rey luciferase reporter gene driven by an SV40 promoter (Agah et al.

818

Ultrasound in Medicine and Biology

Volume 28, Number 6, 2002

1997). CMV-LacZ is a plasmid carrying a nuclear localized -galactosidase (LacZ) reporter gene driven by the CMV promoter (Herz and Gerard 1993). Both plasmids are high copy number plasmids and were propagated in XL-1 Blue bacteria (Stratagene, La Jolla, CA). Large quantities (25 to 50 mg/L culture) of puried plasmid DNA were generated using a modication of the alkaline lysis protocol (Maniatis et al. 1989) that included a proteinase K digest after the RNase digest (Templeton et al. 1997). DNA was analyzed using OD260:280 ratios, and 1% agarose gel electrophoresis in tris-borate buffer and stained with ethidium bromide. In each case, the OD260:280 ratio was 2.0 and gel electrophoresis showed supercoiled plasmid DNA with no genomic DNA contamination. Albumin microbubble synthesis (PESDA, perfluoropropane-exposed sonicated dextrose albumin) In a modication of a previously described technique (Porter et al. 1997), albumin-coated microbubbles were generated using a sterile solution of 5% human serum albumin (ICN, Costa Mesa, CA), 10% dextrose (Sigma-Aldrich, St Louis, MO) in phosphate-buffered saline (PBS) overlaid with peruoropropane gas (C3F8, Air Products, Inc., Allentown, PA) and sonicated at 20 kHz using an XL2020 (Heat Systems Inc., Farmingdale, NY) for 8 s. DNA loading of albumin microbubbles A sterile solution of 5% human serum albumin, 10% dextrose in PBS was prepared as above and additional puried plasmid DNA added to the solution to make nal concentrations of one of the following in terms of g DNA/mL: 500, 1000, 2000, 3000, 4000, 5000 or 10000 g/mL (variously referred to as 0.5k-, 1k-, 2k-, 3k-, 4k-, 5k- and 10k DNA loading). The nal synthetic solution with DNA was then overlaid with peruoropropane gas and sonicated at 20 kHz for 6 to 12 s, as above. Fluorescent microscopy DNA-loaded bubbles were synthesized with the addition of uorosceine-labeled albumin at 0.5 mg/mL nal concentration and propidium iodide (Sigma-Aldrich) at 0.1 mg/mL nal concentration. Bubbles were washed twice with PBS to reduce background uorescence from the solution and visualized. At least 10 elds at a magnication of 680X were scanned. Representative microbubbles were digitally captured at 25 nm increments in Openlab 3.0.3 image acquisition and analysis software using a Zeiss Axioplan 2 Imaging motorized microscope (Carl Zeiss, Germany). Cell culture and cell transfection Cells from the 293 cell line (undifferentiated human embryonic kidney cells) were obtained as a gift from

Roger Unger, M.D. (UT Southwestern Medical School). Cells were grown to a density of 70% conuency in DMEM (GIBCO Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (GIBCO Invitrogen). Cells were digested with trypsin and suspended in DMEM with 5% horse serum (GIBCO Invitrogen) and diluted to a concentration of 1 106/mL. All transfec6 tions were carried out on 2 10 cells. Lipofection was performed using lipofectamine (GIBCO Invitrogen) per manufacturers instructions for cells in suspension (10 g plasmid DNA used 2 106 cells). Ultrasound transfections were carried out using 250 L of microbubbles (loaded or unloaded). Unloaded microbubble transfections received 10 g plasmid DNA. Cell/microbubble suspensions were insonied as follows. The suspensions were placed in a sterile, round bottomed, polysterene test tube (17 10 0 mm, VWR Scientic, Plaineld, NJ) that was lowered 5 cm into a 37C water bath so that the water surrounded the sample without contaminating the inside of the tube. The test tube was held 5 cm from the submersed US transducer (S3 transducer, Sonos 5500, Agilent Technology, Andover, MA) set to give a continuous 1.3-MHz US beam at a mechanical index of 1.6. Each sample received 120 s of US exposure in the water bath. During the 120-s insonation, the sample tube was continuously swirled and rotated manually at approximately 30 rpm. Suspensions were immediately plated simply by pippetting the entire insonied suspension onto 35 mm polysterene tissue cultures plates (Falcon, from BectonDickison Labware, Franklin Lakes, NJ) and incubated for 4 h at 37C. After 4 h, cells had attached to the plate surface and the media was changed to serum-free DMEM media supplemented with insulin, transferrin and selenium-A (GIBCO Invitrogen). Cells were maintained at 37C for another 44 h before harvesting (for a total of 48 h after insonation). Transfection data were based on samples that were carried out in triplicate and each experiment was repeated for a total of 3 times. Parallel transfections were set up to have identical replicates for transfection efciency estimates and viability assays. Cell harvest and luciferase and protein assays Cell lysates were made by incubating 150 L lysis buffer with each 35-mm well of transfected cells at room temperature for 5 min (lysis buffer contained: 25 mM tris-phosphate (pH7.8, 2 mM dithiothreitol, 2 mM diaminocylohexane-N,N,N,N,tetra-acetate, 10% glycerol, 0.1% Triton-X). Cell lysates and debris were scraped and collected into Eppendorf tubes and centrifuged at 13,000 g for 30 min at 4C. Supernatants were separated from the pellets and assayed for protein content and luciferase activity. Luciferase activity was normalized to protein content. Protein assays were performed using the Brad-

Attachment of plasmid DNA to microbubbles q P. A. FRENKEL et al.

819

ford method with BCA Protein Assay (Pierce; Rockford, IL) (Bradford 1976). Luciferase assays were carried out with 25 L of lysate and 100 L of luciferase assay buffer (1.07 mM magnesium carbonate hydroxide, 2.67 mM MgSO4, 0.1 mM EDTA, 33.3 mM DTT, 270 M coenzyme A, 470 M luciferin, 530 M ATP) in TD-20/20 luminometer (Turner Designs, Sunnyvale, CA) equipped with autoinjector. Luciferase activity is reported as the relative light units (RLU) measured by the TD-20/20 luminometer corrected for sample protein content. Transfection efficiency Transfection efciency is the proportion of transfected cells relative to the total number of treated cells. Transfection efciency was calculated with transfections performed with a CMV-enhanced -galactosidase vector (CMVLacZ). The 293 cell line was transfected in the same conditions as above. At 48 h after transfection, cells were stained blue to reveal LacZ expression. Transfection efciency was dened as (blue cells/total cells) 100. To stain the cells, each well was briey rinsed with PBS and lightly xed with 0.5% gluteraldehyde. Cells were stained blue with 5-bromo-4chloro-3-indolyl-b-Dgalactopyranoside (X-gal; Sigma-Aldrich) in 0.1 mol/L phosphate (pH 7.3), 2 mM MgCl2, 5 mM EGTA, 5 mM potassium ferricyanide and 5 mM potassium ferrocyanide (1 mg/mL Xgal) at 37C overnight and rinsed with PBS. Transfection efciency was tabulated by counting total cells and blue-stained cells in four high-power elds (Nikon TE300). The cells were counted by an experienced technician, who was blinded to the content of the samples as well as the results of the transfection. Viability assays To assess viability, cells in suspension were collected immediately after exposure to the US treatment and pelleted briey at 1000 g and resuspended in 1:1 mixture of PBS and 0.4% trypan blue stain (GIBCO Invitrogen). Four aliquots of each replicate were counted on a hemocytometer and the percentage of viable cells was dened as [(total cells blue-stained cells)/total cells] 100 and then corrected for the viability of untreated cells. The cells were counted by an experienced technician, who was blinded to the content of the samples as well as the results of the transfection. Coulter analysis Coulter counter analyses of microbubble size and concentration were performed at the Beckman Coulter Particle Characterization Laboratory (Hialeah, FL). Statistics All data were subjected to ANOVA statistical analysis using StatView (SAS) statistical software and reported 1 SD.

Fig. 1. Gel electrophoresis of DNA extracted from DNAloaded albumin bubbles. Lane 1 marker DNA; Lane 2 1k-loaded, DNA extracted from bubbles synthesized with 1000 g/mL; Lane 3 4k-loaded, DNA extracted from bubbles the plasmid used to synthesized with 4000 g/mL; Lane 4 load the DNA-loaded bubbles linearized with digestion. The closed-headed arrow points to linear DNA and the open-headed arrow points to supercoiled DNA.

RESULTS Sonication to incorporate DNA into albumin shells had no effect on DNA quality. DNA extracted from the loaded bubbles transformed E. coli bacteria as efciently as fresh unsonicated, super-coiled plasmid (data not shown). Gel electrophoresis of DNA extracted from loaded bubbles showed 1. no smearing indicative of shearing, and 2. that the DNA migrated as super-coiled DNA (Fig. 1). The amount of DNA incorporated into the bubbles increased in a linear fashion up to the loading dose of 4000 g plasmid DNA per mL synthetic solution (Fig. 2). Of total DNA used in the loading process, 17% could be extracted from the bubbles. Thus, almost one fth the DNA loading dose used was actually incorporated into the bubble shell, and that DNA was intact, super-coiled plasmid DNA. DNA loading had a minor effect on the diameter of the albumin microbubble relative to similarly synthesized unloaded microbubbles. Microbubbles had about a 20% increase in diameter (Table 1), which would not impart any signicant alteration in acoustical or rheological properties (Chomas et al. 2001a, 2001b). We

820

Ultrasound in Medicine and Biology

Volume 28, Number 6, 2002

Fig. 3. Immunofuorescence of DNA-loaded microbubbles. 4kDNA-loaded albumin microbubbles were synthesized in the presence of uoresceine-labeled albumin and propidium iodide (PI) to stain for DNA. (A) Immunouorescence for PI; (B) immunouorescence for uoresceine; (C) immunouorescence for both. Bubbles shown range in size from 2 to 4 m.

Fig. 2. Graph showing the amount of DNA recovered from the microbubbles (y-axis) for the different DNA concentrations used in loading the microbubbles during bubble synthesis (xaxis). A linear increase was observed up to a concentration of 5000 g/mL.

calculate that each microbubble carries approximately 40 pg of DNA. DNA is incorporated uniformly into the albumin shell as demonstrated by uorescent microscopy (Fig. 3). Nearly all DNA-loaded bubbles showed evidence of DNA integrated in the albumin shell. Propidium iodide staining was symmetrical and diffuse on each microbubble, suggesting DNA is incorporated evenly and not in aggregates. Although most microbubbles carry DNA to varying degrees, some carry a higher degree of DNA than others. Thus, although, on average, 40 pg of DNA is carried in the shell, the amount of DNA carried on any individual microbubble may vary. DNA-loaded bubbles enhanced US-mediated transfection in vitro. Cells in suspension were mixed with either DNA plus unloaded microbubbles or with DNAloaded microbubbles, and transfected by exposing the cell bubble mixture to a US beam from a submerged transducer in a water bath. Unloaded bubbles had optimal transfection in terms of reporter expression when 10

g of plasmid DNA was admixed with constant levels of both microbubbles and cells. Reporter expression fell rapidly and was undetectable at higher doses in the unloaded preparations (Fig. 4). DNA-loaded albumin microbubbles showed progressive linear increases in luciferase reporter expression (Fig. 5), which began to plateau with 4000 g plasmid DNA/mL microbubble synthetic solution. Luciferase expression was enhanced with 4000 g/mL (4k-loaded microbubbles) and 10,000 g/mL (10k-loaded microbubbles) by 5.1-fold and 6.4fold over that of unloaded microbubbles (p 0.0003) that were admixed with an optimal dose of plasmid DNA. Standard lipofection with Lipofactamine in comparison, had a 15-fold increase over unloaded microbubbles (p 0.0001). DNA-loaded microbubbles appear to deliver DNA more successfully than unloaded microbubbles mixed with optimal doses of plasmid DNA. Unloaded micro-

Table 1. Coulter counter data showing the effects of DNA loading with different concentrations of plasmid DNA on microbubble size and concentration
DNA loading No DNA 1000 g/mL 4000 g/mL 10,000 g/mL *p Diameter ( m) 2.0 2.2 2.4 2.3 0.03 0.16 0.07* 0.3* ( Concentration 109 bubbles/mL) 8 13 17 17 3.4 4.1 1.3 2.1

0.05 compared to microbubbles without DNA.

Fig. 4. Transfection of the 293 cell line in vitro by the luciferase gene. Luciferase activity is shown in relative light units (RLU) indexed for protein content on the y-axis. Unloaded albumin microbubbles were mixed with increasing doses of plasmid DNA (shown on the x-axis) before transfection. Both cell number and microbubble concentration were constant.

Attachment of plasmid DNA to microbubbles q P. A. FRENKEL et al.

821

Fig. 5. Transfection of 293 cell line in vitro by the luciferase gene. Luciferase activity is shown in relative light units (RLU) indexed for protein content on the y-axis. The x-axis represents transfections carried out with either lipofectamine, unloaded microbubble, or with DNA-loaded microbubbles over a range of loading concentrations. Plasmid DNA (10 g) was added to the unloaded microbubble samples just before transfection. DNA-loaded bubbles were made with albumin at 5% (see Methods section). The unloaded bubbles made with both 1% and 5% albumin were tested. The loading for DNA-loaded bubbles ranged from 500 to 10,000 g of plasmid DNA/mL of solution used to make bubbles as indicated.

bubbles showed maximal reporter expression at 10 g plasmid DNA/2 106 cells. Reporter expression with DNA doses greater than 50 had no detectible reporter activity. The 4k-loaded microbubbles carried about 170 g DNA, which is 15-fold greater than the 10 g used in unloaded samples. This increase in DNA delivered by the loaded microbubbles was not toxic, as evident in viability assays. There was no difference between the viability of cells treated with DNA-loaded microbubbles (93 1%) and those with microbubbles merely admixed with DNA (93 1%), and these two were less toxic than lipofection (87 4%, p 0.0044) (Fig. 6). In fact the increase in gene expression in the DNA-loaded samples

Fig. 7. Plot showing transfection efciency in 293 cell line transfections under the following treatment conditions: Bubble control unloaded microbubble with no DNA; Unloaded unloaded microbubble with 10 g DNA; DNA Loaded either 2k-loaded microbubble or 4k-loaded microbubble as indicated on the x-axis; Lipofection cells transfected with lipofectamine and 10 g DNA.

correlated with an increase in transfection efciency (Fig. 7). Thus, the DNA-loading process allows albumin microbubbles to deliver DNA more successfully to the US-targeted cells, enhancing gene expression without increasing toxicity. The efcacy of the loaded bubbles occurs over a different plasmid DNA dose range than that considered optimal in unloaded bubbles. DISCUSSION These data show that plasmid DNA is easily incorporated directly into the albumin microbubble shell. DNA-loaded microbubbles are not signicantly changed in their structure, but demonstrate enhanced gene expression in vitro. Thus, the plasmid DNA remains functional after conjugation to the albumin shell and after USmediated microbubble destruction. At optimal doses, DNA-loaded albumin microbubbles transfect cells more effectively than unloaded bubbles, showing increased transfection efciency and increased reporter expression. Interestingly, DNA-loaded bubbles show optimal results in a different dose range for plasmid DNA when compared with unloaded bubbles. We do not believe that DNA loading affects the structure of the DNA because DNA extracted from the loaded microbubbles remains in

Fig. 6. Plot showing viability of cells immediately after the following treatment conditions: Bubble control unloaded microbubble with no added DNA; Unloaded unloaded microbubble with 10 g DNA; DNA Loaded 4k-loaded microbubble; Lipofection cells transfected with lipofectamine and 10 g DNA.

822

Ultrasound in Medicine and Biology

Volume 28, Number 6, 2002 sion of Cre recombinase provokes cardiac-restricted, site-specic rearrangement in adult ventricular muscle in vivo. J Clin Invest 1997;100:169 179. Bao S, Thrall BD, Miller DL. Transfection of a reporter plasmid into cultured cells by sonoporation in vitro. Ultrasound Med Biol 1997; 23:953959. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of proteindye binding. Anal Biochem 1976;72:248 254. Chen S, Shohet RV, Frenkel PA, Mayer S, Unger RH, Grayburn PA. Successful expression of plasmid DNA in rat myocardium by ultrasound targeted microbubble destruction. J Am Coll Cardiol 2001;37:1A 648A. Chomas JE, Dayton P, Allen J, Morgan K, Ferrara KW. Mechanisms of contrast agent destruction. IEEE Trans Ultrason Ferroelec Freq Control 2001a;48:232248. Chomas JE, Dayton P, May D, Ferrara K. Threshold of fragmentation for ultrasonic contrast agents. J Biomed Opt 2001b;6:141150. French BA, Mazur W, Geske RS, Bolli R. Direct in vivo gene transfer into porcine myocardium using replication-decient adenoviral vectors. Circulation 1994;90:2414 2424. Herz J, Gerard RD. Adenovirus-mediated transfer of low density lipoprotein receptor gene acutely accelerates cholesterol clearance in normal mice. Proc Natl Acad Sci USA 1993;90:28122816. Huber PE, Psterer P. In vitro and in vivo transfection of plasmid DNA in the Dunning prostate tumor R3327-AT1 is enhanced by focused ultrasound. Gene Ther 2000;7:1516 1525. Koch S, Pohl P, Cobet U, Rainov NG. Ultrasound enhancement of liposome-mediated cell transfection is caused by cavitation effects. Ultrasound Med Biol 2000;26:897903. Maniatis T, Fritsch EF, Sambrook J. Extraction and purication of plasmid DNA, molecular cloning manual. Cold Spring Harbor, NY: Cold Spring Harbor Press, 1989:1.38 1.39. Marshall E. Gene therapy death prompts review of adenovirus vector. Science 1999;286:2244 2245. Miller DL, Bao S, Gies RA, Thrall BD. Ultrasonic enhancement of gene transfection in murine melanoma tumors. Ultrasound Med Biol 1999;25:14251430. Porter TR, Li S, Kilzer K. Smaller intravenous peruorocarbon-containing microbubbles produce greater myocardial contrast with intermittent harmonic imaging and better delineation of risk area during acute myocardial ischemia. J Am Soc Echocardiogr 1997; 10:792797. Price RJ, Skyba DM, Kaul S, Skalak TC. Delivery of colloidal particles and red blood cells to tissue through microvessel ruptures created by targeted microbubble destruction with ultrasound. Circulation 1998;98:1264 1267. Shohet RV, Chen S, Zhou YT, Wang Z, Meidell RS, Unger RH, Grayburn PA. Echocardiographic destruction of albumin microbubbles directs gene delivery to the myocardium. Circulation 2000; 101:2554 2556. Skyba DM, Price RJ, Linka AZ, Skalak TC, Kaul S. Direct in vivo visualization of intravascular destruction of microbubbles by ultrasound and its local effects on tissue. Circulation 1998;98:290 293. Templeton NS, Lasic DD, Frederik PM, Strey HH, Roberts DD, Pavlakis G.N. Improved DNA: Liposome complexes for increased systemic delivery and gene expression. Nat Biotechnol 1997;15: 647 652. Unger EC, McCreery TP, Sweitzer RH. Ultrasound enhances gene expression of liposomal transfection. Invest Radiol 1997;32:723 727.

a supercoiled state. Two possible explanations for this disparity are that the DNA loading exerts its effect by delivering more plasmid DNA to each transfected cell or that DNA loading merely transfects more cells with similar amounts of plasmid delivered per cell. The latter possibility appears more likely given that the degree of increase in reporter expression is of a similar magnitude to the increase in transfection efciency. We believe that DNA loading in microbubble synthesis will have a signicant effect on gene transfer in vivo for several reasons. First, transfection efciency is on the same order of magnitude as lipofection, but with slightly less toxicity, based on the viability assays. Second, the DNA is incorporated into the shell of the albumin, which has two important effects. Plasmid DNA will be present and not diluted at the site of microbubble destruction. Also, we believe that the shell will protect the DNA from circulating DNAses that can destroy plasmid DNA rapidly. In data utilizing plasmid only mixed with ultrasonic microbubbles, we had signicant expression of reporter gene in cardiac tissue (Chen et al. 2001), but had signicant animal-to-animal variation, which we presume in part is due to variable dissociation of the plasmid DNA from microbubbles in animals with different heart rates. Thus, DNA-loading should guarantee delivery of high-concentration DNA to the targeted tissue with the UTMD technique. Finally, DNA-loaded albumin microbubbles have signicant advantages over adenovirus and are not toxic to cells in vitro. Although the adenovirus can produce high levels of gene expression, its duration of expression is brief, often associated with profound inammation (French et al. 1994). It has a high spillover expression in hepatic tissue, and its use was associated with a fatal case of acute fulminate hepatitis in a patient (Marshall 1999). We have shown here that simple modication of microbubble synthesis can directly incorporate plasmid DNA into the microbubble shell, resulting in greater transfection efciency. Further studies are in progress to determine the optimal method of attaching plasmid DNA to microbubble shells comprised of other materials, such as phospholipids, polymers and acrylics. REFERENCES
Agah R, Frenkel PA, French BA, Michael LH, Overbeek PA, Schneider MD. Gene recombination in postmitotic cells. Targeted expres-