DNA and Polylysine Adsorption and Multilayer Construction

Aug 1, 2007 - Cite This:Langmuir200723189401-9408 ..... Qiaofeng Jin, Zhiyong Wang, Fei Yan, Zhiting Deng, Fei Ni, Junru Wu, Robin Shandas, Xin Liu, ...
0 downloads 0 Views 603KB Size
Download PDF
Langmuir 2007, 23, 9401-9408

9401

DNA and Polylysine Adsorption and Multilayer Construction onto Cationic Lipid-Coated Microbubbles Mark A. Borden,*,†,‡ Charles F. Caskey,† Erika Little,† Robert J. Gillies,‡ and Katherine W. Ferrara† Biomedical Engineering Department, UniVersity of California, DaVis, California 95616, and Arizona Cancer Center, Tucson, Arizona 85724 ReceiVed March 28, 2007. In Final Form: June 15, 2007 We report on a novel application of the layer-by-layer (LbL) assembly technique to attach multiple layers of DNA and poly-L-lysine (PLL) onto preformed lipid-coated microbubbles to increase the DNA loading capacity. We first measured the effects of the cationic lipid fraction and salt concentration on the microbubble stability. Microbubble production and stability were robust up to a cationic lipid fraction of 40 mol % in 10 mM NaCl. DNA adsorption was heterogeneous over the microbubble shell and occurred primarily on the condensed phase domains. The amount of adsorbed DNA, and subsequently adsorbed PLL, increased linearly with the fraction of cationic lipid in the shell. DNA loading was further enhanced by the LbL assembly method to construct polyelectrolyte multilayers (PEMs) of DNA and PLL. PEM buildup was demonstrated by experimental results from ζ potential analysis, fluorescence microscopy, UV spectroscopy, and flow cytometry. The PEMs exhibited two growth stages and were heterogeneously distributed over the microbubble surface. The DNA loading capacity onto the microbubbles was enhanced by over 10-fold by using five paired layers. However, the PEM shell did not prevent oscillation or destruction during ultrasound insonification. These results suggest that the surface can be compartmentalized to make multifunctional, high-payload ultrasound contrast agents for targeted gene therapy.

Introduction Ultrasound-targeted microbubble destruction (UTMD) is currently being developed to enhance in vivo transfection in gene therapy.1,2 This technique involves intravenous or intraarterial injection of a suspension of echogenic microbubbles with DNA. Ultrasound is focused deep within the body to destroy the circulating microbubbles in the target tissue. UTMD has been shown to enhance local vascular permeability, presumably through the creation of pores by microjets, shock waves, and other convective phenomena emanating from the inertia of oscillation. Fluorescent polymers and particles coadministered with microbubbles, for example, have been observed to permeate several hundred micrometers into tissue after insonification.3 Thus, in the broadest sense, the microbubble serves as a remotely triggered actuator of vascular permeability. In gene therapy, UTMD has been shown to significantly increase DNA transfection efficiency over that of administration of naked DNA alone, with high specificity for the insonified organ.4 Recent research has focused on another function for the microbubble in UTMD: as a vehicle to load and protect genetic material during transit to the target site. Nucleic acids can be coupled directly onto the microbubble shell by electrostatic interactions. This affords protection of the genetic material from * To whom correspondence should be addressed at the Chemical Engineering Department, Columbia University, 500 W. 120th St., New York, NY 10027. E-mail: [email protected]. † University of California. ‡ Arizona Cancer Center. (1) Bekeredjian, R.; Grayburn, P. A.; Shohet, R. V. Use of ultrasound contrast agents for gene or drug delivery in cardiovascular medicine. J. Am. Coll. Cardiol. 2005, 45 (3), 329-335. (2) Bekeredjian, R.; Katus, H. A.; Kuecherer, H. F. Therapeutic use of ultrasound targeted microbubble destruction: A review of non-cardiac applications. Ultraschall Med. 2006, 27 (2), 134-140. (3) Price, R. J.; Skyba, D. M.; Kaul, S.; Skalak, T. C. Delivery of colloidal, particles and red blood cells to tissue through microvessel ruptures created by targeted microbubble destruction with ultrasound. Circulation 1998, 98 (13), 1264-1267.

premature clearance and enzymatic degradation.5 Results have shown that the DNA remains intact, even after insonification.6,7 In this scenario, microbubble destruction results in simultaneous release of functional genetic payload and enhancement of local vascular permeability. This dual capability has been shown to increase DNA transfection efficiency in vivo.6-9 Shells comprising a synthetic lipid monolayer are of particular interest for UTMD due to a high degree of compliance during insonification. The monolayer shell readily reseals following rarefaction-induced rupture and respreads following compressioninduced buckling to maintain the overall shell integrity over multiple oscillations.10,11 This advantageous behavior is contrasted with that of rigid shells, such as those formed by proteins or bulk polymers, which require higher amplitude insonification and experience localized rupture through which the gas core escapes.11,12 Additionally, the lipid monolayer shell is better suited to control release of associated material. The lipid composition and microstructure can be engineered to shed a variety of colloidal (4) Porter, T. R.; Xie, F. Therapeutic ultrasound for gene delivery. Echocardiagraphy 2001, 18 (4), 349-353. Li, T. L.; Tachibana, K.; Kuroki, M.; Kuroki, M. Gene transfer with echo-enhanced contrast agents: Comparison between Albunex, Optison, and Levovist in micesInitial results. Radiology 2003, 229 (2), 423-428. Akowuah, E. F.; Gray, C.; Lawrie, A.; Sheridan, P. J.; Su, C. H.; Bettinger, T.; Brisken, A. F.; Gunn, J.; Crossman, D. C.; Francis, S. E.; Baker, A. H.; Newman, C. M. Ultrasound-mediated delivery of TIMP-3 plasmid DNA into saphenous vein leads to increased lumen size in a porcine interposition graft model. Gene Ther. 2005, 12 (14), 1154-1157. Tsunoda, S.; Mazda, O.; Oda, Y.; Iida, Y.; Akabame, S.; Kishida, T.; Masaharu, S. Y.; Asada, H.; Gojo, S.; Imanishi, J.; Matsubara, H.; Yoshikawa, T. Sonoporation using microbubble BR14 promotes pDNA/ siRNA transduction to murine heart. Biochem. Biophys. Res. Commun. 2005, 336 (1), 118-127. (5) Lentacker, I.; De Geest, B. G.; Vandenbroucke, R. E.; Peeters, L.; Demeester, J.; De Smedt, S. C.; Sanders, N. N. Ultrasound-responsive polymer-coated microbubbles that bind and protect DNA. Langmuir 2006, 22 (17), 72737278. (6) Christiansen, J. P.; French, B. A.; Klibanov, A. L.; Kaul, S.; Lindner, J. R. Targeted tissue transfection with ultrasound destruction of plasmid-bearing cationic microbubbles. Ultrasound Med. Biol. 2003, 29 (12), 1759-1767. (7) Bekeredjian, R.; Chen, S. Y.; Frenkel, P. A.; Grayburn, P. A.; Shohet, R. V. Ultrasound-targeted microbubble destruction can repeatedly direct highly specific plasmid expression to the heart. Circulation 2003, 108 (8), 1022-1026.

10.1021/la7009034 CCC: $37.00 © 2007 American Chemical Society Published on Web 08/01/2007

9402 Langmuir, Vol. 23, No. 18, 2007

structures, from submicrometer micelles to micrometer-sized vesicles.13 Current formulations involve electrostatic coupling of DNA onto microbubble shells comprising cationic lipids. Christiansen et al.6 showed that plasmid DNA remains intact after being loaded onto and subsequently released from microbubbles by insonification. The loading capacity is limited by the surface area, and saturation has been reported to occur at approximately 0.0010.005 pg/µm2.6,9 Such a low capacity could be responsible for the high microbubble doses that have been reported in preclinical trials;1,2 these doses often exceed by orders of magnitude the accepted safe dose of microbubbles for humans in contrast echocardiography. It is possible that a 10-fold increase in DNA loading capacity could lower the effective microbubble dose (i.e., increase the therapeutic index) by an order of magnitude or more and thus enable this technology for human clinical trials. Below, we report on a novel, simple, and robust method to increase the DNA loading capacity of preformed lipid-coated microbubbles. We start by first investigating the influence of the lipid charge and salt concentration on the microbubble colloidal stability. We then reveal new insights into the adsorption of DNA onto the multicomponent, microstructured lipid shell. We also demonstrate a method to increase the DNA loading capacity through the use of a layer-by-layer (LbL) assembly to form polyelectrolyte multilayers (PEMs). Our technique is shown schematically in Figure 1. The cationic microbubble serves as the template for the sequential adsorption of DNA and polyL-lysine (PLL) to form the PEM shell. Finally, we show that insonified PEM-shelled microbubbles oscillate and fragment similarly to unmodified lipid-shelled microbubbles. Methods and Materials Materials. DSPC and DSTAP were purchased from Avanti Polar Lipids (Alabaster, AL). PEG40S, PLL (30 kDa), and FITC-PLL (15-30 kDa) were purchased from Sigma-Aldrich (St. Louis, MO). (8) Shohet, R. V.; Chen, S. Y.; Zhou, Y. T.; Wang, Z. W.; Meidell, R. S.; Unger, T. H.; Grayburn, P. A. Echocardiographic destruction of albumin microbubbles directs gene delivery to the myocardium. Circulation 2000, 101 (22), 2554-2556. Unger, E. C.; Hersh, E.; Vannan, M.; McCreery, T. Gene delivery using ultrasound contrast agents. Echocardiagraphy 2001, 18 (4), 355361. Vannan, M.; McCreery, T.; Li, P.; Han, Z. G.; Unger, E.; Kuersten, B.; Nabel, E.; Rajagopalan, S. Ultrasound-mediated transfection of canine myocardium by intravenous administration of cationic microbubble-linked plasmid DNA. J. Am. Soc. Echocardiogr. 2002, 15 (3), 214-218. Chen, S. Y.; Shohet, R. V.; Bekeredjian, R.; Frenkel, P.; Grayburn, P. A. Optimization of ultrasound parameters for cardiac gene delivery of adenoviral or plasmid deoxyribonucleic acid by ultrasound-targeted microbubble destruction. J. Am. Coll. Cardiol. 2003, 42 (2), 301-308. Korpanty, G.; Chen, S.; Shohet, R. V.; Ding, J.; Yang, B.; Frenkel, P. A.; Grayburn, P. A. Targeting of VEGF-mediated angiogenesis to rat myocardium using ultrasonic destruction of microbubbles. Gene Ther. 2005, 12 (17), 1305-1312. (9) Chen, S. Y.; Ding, J. H.; Bekeredjian, R.; Yang, B. Z.; Shohet, R. V.; Johnston, S. A.; Hohmeier, H. E.; Newgard, C. B.; Grayburn, P. A. Efficient gene delivery to pancreatic islets with ultrasonic microbubble destruction technology. Proc. Nat. Acad. Sci. U.S.A. 2006, 103 (22), 8469-8474. (10) Chomas, J. E.; Dayton, P. A.; May, D.; Allen, J.; Klibanov, A.; Ferrara, K. Optical observation of contrast agent destruction. Appl. Phys. Lett. 2000, 77 (7), 1056-1058. Chomas, J. E.; Dayton, P.; Allen, J.; Morgan, K.; Ferrara, K. W. Mechanisms of contrast agent destruction. IEEE Trans.: Ultrason. Ferroelectr. Freq. Control 2001, 48 (1), 232-248. (11) Bloch, S. H.; Wan, M.; Dayton, P. A.; Ferrara, K. W. Optical observation of lipid- and polymer-shelled ultrasound microbubble contrast agents. Appl. Phys. Lett. 2004, 84 (4), 631-633. (12) Leong-Poi, H.; Song, J.; Rim, S. J.; Christiansen, J.; Kaul, S.; Lindner, J. R. Influence of microbubble shell properties on ultrasound signal: Implications for low-power perfusion imaging. J. Am. Soc. Echocardiogr. 2002, 15 (10), 12691276. Bouakaz, A.; Versluis, M.; de Jong, N. High-speed optical observations of contrast agent destruction. Ultrasound Med. Biol. 2005, 31 (3), 391-399. (13) Borden, M. A.; Kruse, D.; Caskey, C.; Zhao, S.; Dayton, P.; Ferrara, K. Influence of lipid shell physicochemical properties on ultrasound-induced microbubble destruction. IEEE Trans.: Ultrason. Ferroelectr. Freq. Control 2005, 52 (11), 1992-2002. Pu, G.; Borden, M. A.; Longo, M. L. Collapse and shedding transitions in binary lipid monolayers coating microbubbles. Langmuir 2006, 22 (7), 2993-2999.

Borden et al.

Figure 1. Schematic representation showing LbL assembly of a PEM shell on a preformed lipid-coated microbubble. In this case, the shell contains a cationic lipid species. DNA is the polyanion, and PLL is the polycation. Details of the assembly procedure are given in the text. PEG40S was chosen as the emulsifier because it is nonionic. PFB was purchased from SynQuest (Alachua, FL). SytoxOrange and YoYo-1 fluorescent dyes for DNA labeling and the membrane probe DiIC18 and DiOC18 were obtained from Invitrogen (Carlsbad, CA). Dubelcco’s phosphate-buffered saline (PBS) was used. All solutions were prepared with analytical grade reagents dissolved in doubly distilled water. Preparation of DNA. The salmon sperm DNA (Sigma-Aldrich) was degraded in water (2 mg/mL) by sonication at low power for 5 min with a cell disrupter (Branson 2510, Danbury, CT). Gel electrophoresis of the sonicated linear DNA (slDNA) resulted in a broad smear that indicated a wide range of molecular weights. We used Promega’s (Madison, WI) pGL4.13 plasmid, which carries a firefly luciferase reporter gene driven by an SV40 promoter. Plasmids were grown in JM109 chemically competent cells from Promega and were purified using Promega’s PureYield MaxiPrep kits. The plasmid preparations were pooled, ethanol precipitated (1/10 volume of 5 M NaOAc, 2.5 volumes 100% EtOH), and resuspended in a volume of pure water required to yield a concentration of 1 mg/mL. The total DNA concentration was measured using a NanoDrop-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). OD260:280 for each preparation measured 1.90-2.00. Gel electrophoresis of plasmid DNA (pDNA) showed intact plasmid only, with no indication of genomic DNA contamination. Three bands were often observed that corresponded to the circular, coiled, and supercoiled topographies. Microbubble Formation. Microbubbles were prepared by agitating a lipid solution with a perfluorobutane headspace, as previously described.14,15 Briefly, a lipid suspension (1-3 mg/mL), with molar percentages of 10% PEG40S, X% DSTAP, and (90 X)% DSPC, was prepared by film hydration at 65 °C, followed by mild sonication with the cell disrupter to form small, unilamellar vesicles. Microbubbles were formed by applying high-amplitude sonication (near the microtip limit) just below the gas-solution interface, with PFB flowing over the solution. Alternatively, microbubbles were formed by enclosing 2 mL of lipid suspension in a serum vial with PFB in the headspace and then shaking the vial in a vial activator (Mixtura, ImaRx Therapeutics, Tucson, AZ) for 45 s. Both methods were found to produce microbubbles with equivalent size distributions and yields. Sonication was preferred for producing larger batches (>10 mL), whereas shaking was preferred for smaller batches. Microbubbles were sized and counted using an Accusizer optical sizer (Particle Sizing Systems, Santa Barbara, CA). PEM Construction. Preformed microbubbles were washed by centrifugation-flotation as previously described.16 Briefly, mi(14) Borden, M. A.; Longo, M. L. Dissolution behavior of lipid monolayercoated, air-filled microbubbles: Effect of lipid hydrophobic chain length. Langmuir 2002, 18 (24), 9225-9233. (15) Borden, M. A.; Martinez, G. V.; Ricker, J. V.; Tsvetkova, N. M.; Longo, M. L.; Gillies, R. J.; Dayton, P. A.; Ferrara, K. W. Lateral phase separation in lipid-coated microbubbles. Langmuir 2006, 22 (9), 4291-4297. (16) Zhao, S.; Borden, M. A.; Bloch, S.; Kruse, D.; Ferrara, K. W.; Dayton, P. A. Radiation-force assisted targeting facilitates ultrasonic molecular imaging. Mol. Imaging 2004, 3 (3), 135-148.

DNA Loading onto Lipid Microbubbles crobubbles contained in a syringe were spun in a bucket-rotor centrifuge at 200g to 500g for 5-10 min. This produced a foam cake, under which the infranatant could be removed. Fresh solution then was pulled into the syringe, and the syringe was gently agitated to resuspend the microbubbles. Each washing step removed >98% of the soluble (or denser than water) contents. Mild flotation (∼1g for 5-10 min) was used to remove larger microbubbles (>5 µm diameter) for quantifying DNA and PLL loading. Following an initial wash, a primary layer was adsorbed by simply suspending the microbubbles in a DNA solution (0.5-2.0 mg/mL), incubating the resulting solution at room temperature, and washing off the residual DNA. DNA-coated bubbles then were suspended in a PLL solution (0.5-2.0 mg/mL) and washed. To make PEM-shelled microbubbles, these steps were repeated in sequence to reach the desired number of paired layers (m). Microbubble Characterization. The ζ potential of microbubbles diluted in 10 mM NaCl solution to ∼107 microbubbles/mL was determined with a ZetaSizer 3000 (Malvern, Worcestershire, U.K.). Confocal laser scanning microscopy of microbubble samples was performed on a PCM 3000 (Nikon, Japan) using Simple PCI software (Compix, Sewickley, PA). Maximum projection images were constructed from z stacks of the microbubble upper hemispheres. Epifluorescence imaging was performed on a custom upright microscope with an a 100× oil-immersion objective and an automated x-y-z positioning stage (Mikron Instruments, San Marcos, CA) driven by Simple PCI software. Flow cytometry of the microbubbles was performed with a FACScan (BD Biosciences, San Jose, CA). DNA Loading Measurement. The loading of sonicated salmon sperm DNA onto the microbubbles was quantified by first isolating 0.5 mL of microbubble suspension after mild flotation (resulting in ∼108 microbubbles/mL) and then sizing and counting to determine the total microbubble surface area by integration of the size distribution and assuming a spherical shape. Following counting, the total DNA mass was determined by UV absorption. A 1.0 mL sample of saturated NaCl solution was added to dissolve the PEM shells, and the suspension was heated to 70 °C for several hours to destroy the microbubbles, resulting in a clear solution. A 100 µL portion of each sample was transferred into a 96-well plate (in triplicate), and the absorbance at 260 nm was measured by a plate reader (BioTek, Winooski, VT). The DNA concentration was converted to mass units using a standard curve. In the case of plasmid DNA, the microbubbles were destroyed in a sonic bath (Branson 2510, Danbury, CT) heated to 65 °C, and DNA was released into the solution by addition of NaCl to saturation (as determined by the onset of salt precipitation). Plasmid DNA was quantified by measuring the adsorption at 260 nm with a NanoDrop spectrophotometer. The use of surface area as the basis (i.e., pg/µm2) allowed correction for variations in size distribution from batch to batch. High-Speed Video Microscopy. Images of oscillating microbubbles were acquired using a high-speed camera (DRS Hadland, Cupertino, CA) adapted to a microscope (IX70, Olympus, Melville, NY). Details of the experimental system were described previously.17 Briefly, a 1.0 or 2.25 MHz transducer was aligned so that it shared the focal region with the microscope (IL0108HP, Valpey-Fisher, Hopkinton, MA). The microbubbles were flowed into a transparent cellulose tube (Spectrum Laboratories, Rancho Dominguez, CA) for optical observation. The contrast agent was diluted to approximately 1 sphere/µL to avoid multibubble interactions. The transducer was driven by an arbitrary waveform generator (AWG 2021, Tektronix, Wilsonville, OR) and a 55 dB power amplifier (3100LA, ENI, Rochester, NY). The peak negative pressure and acoustic waveform characteristics were measured using a needle hydrophone (HNZ-0400, Onda Corp., Sunnyvale, CA). The initial and maximum diameters were measured with MATLAB (Mathworks, Natick, MA). Microbubbles were insonified with two different pulse sequences. For high mechanical index (MI ) peak negative pressure over the square root of the frequency) stimulation, the microbubbles were insonified with a sinusoidal, five-cycle pulse at 1.0 MHz and (17) Chomas, J. E.; Dayton, P.; May, D.; Ferrara, K. Threshold of fragmentation for ultrasonic contrast agents. J. Biomed. Opt. 2001, 6 (2), 141-150.

Langmuir, Vol. 23, No. 18, 2007 9403

Figure 2. Microbubble yields and size distributions at production and after washing as a function of the surface charge and ionic strength. (A) Total particle yields (mean ( SD, n > 6). (B, C) Average size distributions (n > 6) of microbubble suspensions as a function of the charged lipid fraction in 2 M NaCl. “Fresh” samples were taken straight from the vial within 1 h of formation. “Washed” samples underwent three cycles of flotation-centrifugation with infranatant exchange to remove submicrometer bubbles. Microbubbles were formed by the shaking method. a 800 kPa peak negative pressure (PNP). For low MI stimulation, the microbubbles were insonified with a 20-cycle pulse at 2.5 MHz and a 500 kPa PNP.

Results and Discussion Colloidal Stability of Charged Microbubbles. Inclusion of charged species in the microbubble shell could disrupt lipid packing, which is known to be crucial for microbubble stability.14,18 Even the slightest increase in surface tension is predicted to cause rapid microbubble dissolution.19 Unger and co-workers previously reported a maximum charged lipid concentration of 20 mol % for stable microbubbles.20 However, the relationship

9404 Langmuir, Vol. 23, No. 18, 2007

between surface charge and microbubble stability, defined here as the yield of microbubbles at production and following washing, was not well characterized prior to the present study. To address this, we measured the microbubble concentration and size distribution as a function of the DSTAP concentration (0-40 mol %). The PEGylated lipid concentration was held constant at 10 mol %. Figure 2 shows the effect of the DSTAP molar concentration on the microbubble yield and size for suspensions in high salt (2.0 M NaCl), where charge effects are expected to be minimal. The addition of a charged lipid (0-20 mol % DSTAP) significantly reduced the total bubble production yield (Figure 2A). However, the production yield was stable in the range of 20-40 mol % DSTAP. A similar trend was observed after washing, although the difference in yield between charged and uncharged was much less. For example, we observed only a 59% decrease in bubble concentration from 0 to 40 mol % and only a 3% decrease in microbubble concentration from 20 to 40 mol %. A closer look at the size distribution further elucidates the effect of a charged lipid on microbubble production. As seen in Figure 2B, the production of submicrometer particles was reduced in monotonic fashion with increasing cationic lipid. These submicrometer particles appeared to be bubbles due to their buoyancy21 and rapid destruction in a sealed syringe under applied pressure (data not shown). Apparently, electrostatic lateral repulsion between the lipid headgroups destabilized the monolayer shell at submicrometer radii of curvature. Figure 2C shows that, following washing to remove submicrometer bubbles, the size distributions were independent of the DSTAP content. We conclude that the effect of incorporating a charged lipid is to reduce the production of submicrometer bubbles, but that the impact on microbubble (i.e., >1 µm diameter) production is minimal. Figure 2A shows an interesting result: the average bubble concentration was 80% and 104% greater for 20 and 40 mol % DSTAP, respectively, when washing was performed at a lower ionic strength (1.0 mM). We further investigated the effect of the ionic strength on the bubble size and yield for shells composed of 20 mol % DSTAP. Figure 3 shows that bubble production was depleted at low salt (0 and 0.1 mM NaCl). Significantly more bubbles were produced at higher ionic strengths. Surprisingly, the submicrometer bubble population was much greater for 10 mM NaCl than for 2 M NaCl, although the microbubble populations were approximately equal. Similar results were observed for 15 mol % DSTAP (data not shown). The optimum in total cationic bubble production fortuitously occurred at physiological ionic strength. Taken together, these data suggest a complex role for electrostatics in lipid packing for this system. DNA and PLL Loading. To determine the maximum loading capacity of a microbubble without multilayers, we first measured the DNA surface density for a range of charged lipid fractions. Figure 4A shows that the DNA loading capacity increased linearly from 0.0009 ( 0.0006 pg/µm2 for 5 mol % DSTAP to 0.006 ( 0.0009 pg/µm2 for 30 mol % DSTAP. These values are close to those calculated from reports of plasmid DNA on cationic lipid(18) Klibanov, A. L. Targeted delivery of gas-filled microspheres, contrast agents for ultrasound imaging. AdV. Drug DeliVery ReV. 1999, 37 (1-3), 139157. (19) Duncan, P. B.; Needham, D. Test of the Epstein-Plesset, model for gas microparticle dissolution in aqueous media: Effect, of surface tension and gas undersaturation in solution. Langmuir 2004, 20 (7), 2567-2578. (20) Unger, E. C.; Porter, T.; Culp, W.; Labell, R.; Matsunaga, T.; Zutshi, R. Therapeutic applications of lipid-coated microbubbles. AdV. Drug DeliVery ReV. 2004, 56 (9), 1291-1314. (21) Borden, M. A.; Martinez, G. V.; Ricker, J.; Tsvetkova, N.; Longo, M.; Gillies, R. J.; Dayton, P. A.; Ferrara, K. W. Lateral phase separation in lipidcoated microbubbles. Langmuir 2006, 22 (9), 4291-4297.

Borden et al.

Figure 3. Size distributions (mean ( SD, n > 6) for freshly shaken suspensions containing 20 mol % DSTAP. Samples were formed by the shaking method described in the text and sampled within 1 h of formation. Microbubbles were formed by the shaking method.

Figure 4. DNA and PLL loading onto microbubbles in 10 mM NaCl as a function of cationic lipid fraction. (A) Sonicated salmon sperm DNA surface concentration (mean ( SD, n ) 3) obtained from UV adsorption and size analysis. Linear least-squares regression yielded a slope of 0.0002056 ( 0.00001284 (p < 0.0001). (B) Mean fluorescence intensities obtained from flow cytometry analysis (arbitrary fluorescence units, mean ( SEM, n g 3) after FITC-PLL was loaded onto DNA-coated microbubbles (primary layer only). Linear least-squares regression yielded a slope of 1.028 ( 0.06432 (p < 0.0001). Microbubbles were formed by the sonication method.

coated microbubbles by Chen et al.9 (0.004 pg/µm2) and Christiansen et al.6 (0.001 pg/µm2), although they are lower than that reported by Bekeredjian et al.7 (0.024 pg/µm2). The linear dependence suggests that the DNA binds to DSTAP in a stoichiometric manner. Assuming an average lipid molecular area of 40 Å2, a linear least-squares fit to the data in Figure 4A yields a ratio of 7.5 DNA base pairs per DSTAP molecule. The excess negatively charged phosphate moieties were responsible for charge reversal during LbL assembly, as discussed below. Our next task was to confirm the deposition of the cationic polymer PLL onto the adsorbed DNA. Fluorescence microscopy

DNA Loading onto Lipid Microbubbles

Langmuir, Vol. 23, No. 18, 2007 9405

Figure 6. ζ potential (mean ( SD, n g 3) versus polyion deposition step. Odd numbers represent PLL deposition, and even numbers represent DNA deposition. The microbubbles contained 20 mol % DSTAP and were formed by sonication. PEM construction was performed with sonicated salmon sperm DNA (2 mg/mL) and FITCPLL (2 mg/mL) in 10 mM NaCl.

Figure 5. Maximum projection fluorescence images of DNA and FITC-PLL adsorbed onto cationic microbubbles. (A) Epifluorescence images showing preferential adsorption of DNA onto condensed phase domains. Large domains were made by heating the suspension in a sealed vial to 65 °C and then cooling it to room temperature on the bench top. The left image shows the location of YoYo-1 labeled plasmid DNA. The right image shows the location of DiIC18 (5 min). These effects were expected owing to attractions between oppositely charged patches on neighboring microbubbles. Finally, aggregation was minimized by dispersing the microbubbles from the cake before adding the polyion and then maintaining adequate mixing (e.g., on a rotator) during the incubation steps. Under these optimal conditions, we (38) Steitz, R.; Leiner, V.; Siebrecht, R.; von Klitzing, R. Influence of the ionic strength on the structure of polyelectrolyte films at the solid/liquid interface. Colloids Surf., A 2000, 163 (1), 63-70. (39) Dubas, S. T.; Schlenoff, J. B. Swelling and smoothing of polyelectrolyte multilayers by salt. Langmuir 2001, 17 (25), 7725-7727.

9408 Langmuir, Vol. 23, No. 18, 2007

Figure 11. Maximum expansion ratios (maximum diameter during rarefaction over the initial resting diameter) determined from analysis of high-speed video microscopy streak images of microbubbles during insonification. The microbubbles (20 mol % DSTAP) contained three layers of plasmid DNA and PLL. The ultrasound pulse consisted of 20 cycles at 2.5 MHz and a 500 kPa PNP.

were able to achieve multilayer buildup without observing any change in microbubble size distribution (Figure 9). Oscillation in the Ultrasound Field. Finally, we used highspeed video microscopy to confirm that PEM-shelled microbubbles were acoustically active. Figure 10 shows the response of microbubbles to the ultrasound field. The top panel shows a microbubble coated with three paired DNA/PLL layers oscillating and dissolving during the low-MI pulse (20 cycles at 2.5 MHz and 500 kPa). The bottom panel shows a microbubble coated with five paired layers being fragmented and destroyed by the high-MI pulse (five cycles at 1.0 MHz and 800 kPa). Similar images of PEM-shelled microbubbles and unmodified, lipidcoated microbubbles were analyzed to determine the effect on damping. In general, the PEM-shelled microbubbles behaved similarly to lipid-only microbubbles, which have been described previously in detail.10,17,40 Figure 11 shows a plot of the maximum expansion ratio (taken as the maximum diameter achieved during rarefaction divided by the initial resting diameter) as a function of the initial diameter for PEM-shelled microbubbles (m ) 3) at the low-MI pulse (20 cycles at 2.5 MHz and 500 kPa). The microbubbles fragmented and were no longer visible by the end of the pulse in 53% of the examined cases. In cases where the microbubble did not fragment and was visible, as in Figure 10, the diameter of the bubble decreased on average by 46 ( 14% by the end of the pulse due to acoustically driven dissolution. Chomas et al.17 showed that, for a given ultrasound frequency, pressure, and phase, there is a threshold diameter below which most microbubbles will fragment. Our results showing fragmentation of small bubbles and dissolution of larger bubbles indicate that all microbubbles will fragment if multiple break pulses are transmitted with the parameters used here. Logistic analysis by Chomas et al.17 performed at 2.25 MHz and 500 kPa showed the fragmentation threshold of unmodified, lipid-coated microbubbles to be approximately 3 µm. In our experiments at 2.5 MHz, the initial diameter threshold for (40) Morgan, K. E.; Allen, J. S.; Dayton, P. A.; Chomas, J. E.; Klibanov, A. L.; Ferrara, K. W. Experimental and theoretical evaluation of microbubble behavior: Effect of transmitted phase and bubble size. IEEE Trans.: Ultrason. Ferroelectr. Freq. Control 2000, 47 (6), 1494-1509. Dayton, P. A.; Chomas, J. E.; Lum, A. F. H.; Allen, J. S.; Lindner, J. R.; Simon, S. I.; Ferrara, K. W. Optical and acoustical dynamics of microbubble contrast agents inside neutrophils. Biophys. J. 2001, 80 (3), 1547-1556.

Borden et al.

fragmentation was approximately 1.5 µm. The difference in threshold diameter likely is due to the PEM shell, since the transmitted center frequency was increased only slightly in comparison with that of Chomas et al.,17 and in each case the fractional bandwidth was ∼20%. In high-MI ultrasound experiments (1.0 MHz and 800 kPa), all PEM-shelled (m ) 5) and lipid-only microbubbles were observed to fragment within the five-cycle pulse (n > 20). However, we observed an effect of the PEM shell on the timing of fragmentation. Microbubbles coated with five paired layers tended to fragment on the second or third cycle rather than the first cycle, as was the case for lipid-only microbubbles. Thus, in this preliminary study, the PEM shell slightly increased microbubble stability to ultrasound over the first few cycles, although the consequence is probably trivial since applications in UTMD often use several thousand cycles over multiple pulses.1 It is possible that more uniform PEM shells could cause a greater damping effect on oscillation.

Conclusions The aspects that make LbL assembly on microbubbles more complex than previous particle systems (e.g., polystyrene, silica, emulsions, etc.), such as the fragility of the gas core and microstructure of the monolayer shell, also impart more versatility. For example, the gas core yields an ultrasound stimulus-responsive construct, and the lipid shell can be compartmentalized for different functions. We found that colloidal stability, DNA adsorption, and PEM construction each depend on lipid mixing, which in turn depends on the lipid composition and salt concentration. For example, DNA can be loaded directly onto DSTAP-enriched domains, and the size and shape of the domains can be controlled through heating and cooling. Fortuitously, microbubble stability and PEM deposition occurred at an optimal salt concentration that is physiological. Also, the 10-fold increase in DNA loading and the overall increase in surface homogeneity with increasing layer deposition are promising. The PEM shell mildly dampened oscillation during insonification, but the effect did not prevent fragmentation at typical UTMD parameters. Our results suggest that this method could be useful for designing the next generation of ultrasound contrast agents and triggered release vehicles. Acknowledgment. We are very grateful to Brenda Baggett at the Arizona Cancer Center and Heather Lindfors and William Moroski at the University of California, Davis, for providing plasmid DNA and to Terry Matsunaga and Nikhil Pargaonkar at ImaRx Therapeutics for use of the Accusizer and ZetaSizer instruments. This work was funded by NIH Grants CA 103828 and 76062 to K.W.F. and NIH Grant CA 077575 to R.J.G. Definitions DiIC18 ) 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate DiOC18 ) 3,3′-dioctadecyloxacarbocyanine perchlorate DMPC ) 1,2-dimyristoyl-sn-glycero-3-phosphocholine DMTAP ) 1,2-dimyristoyl-3-(trimethylammonio)propane DOPE ) 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine DOTAP ) 1,2-dioleoyl-3-(trimethylammonio)propane DPPC ) 1,2-dipalmitoyl-sn-glycero-3-phosphocholine DPTAP ) 1,2-dipalmitoyl-3-(trimethylammonio)propane DSPC ) 1,2-distearoyl-sn-glycero-3-phosphocholine DSTAP ) 1,2-distearoyl-3-(trimethylammonio)propane PEG40S ) PEG-40 stearate FITC ) fluorescein isothiocyanate PLL ) poly-L-lysine LA7009034