- Email: [email protected]
In vitro and in vivo transfection of primary phagocytes via microbubble-mediated intraphagosomal sonoporation
Journal of Immunological Methods 371 (2011) 152–158
Contents lists available at ScienceDirect
Journal of Immunological Methods j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j i m
Technical Note
In vitro and in vivo transfection of primary phagocytes via microbubble-mediated intraphagosomal sonoporation Jason C.M. Lemmon a, Ryan J. McFarland b, Joanna M. Rybicka a, Dale R. Balce a, Kyle R. McKeown b, Regina M. Krohn a, Terry O. Matsunaga b, Robin M. Yates a, c,⁎ a b c
Department of Comparative Biology and Experimental Medicine, Faculty of Veterinary Medicine, University of Calgary, AB, Canada Department of Radiology Research, University of Arizona, AZ, USA Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Calgary, AB, Canada
a r t i c l e
i n f o
Article history: Received 23 February 2011 Received in revised form 9 May 2011 Accepted 1 June 2011 Available online 14 June 2011 Keywords: Phagocytosis Macrophage Dendritic cell Gene therapy DNA vaccine
a b s t r a c t The professional phagocytes, such as macrophages and dendritic cells, are the subject of numerous research efforts in immunology and cell biology. The use of primary phagocytes in these investigations however, are limited by their inherent resistance to transfection with DNA constructs. As a result, the use of phagocyte-like immortalized cell lines is widespread. While these cell lines are transfection permissive, they are generally regarded as poor biological substitutes for primary phagocytes. By exploiting the phagocytic machinery of primary phagocytes, we developed a non-viral method of DNA transfection of macrophages that employs intraphagosomal sonoporation mediated by internalized lipid-based microbubbles. This approach enables the transfection of primary phagocytes in vitro, with a modest, but reliable efficiency. Furthermore, this methodology was readily adapted to transfect murine peritoneal macrophages in vivo. This technology has immediate application to current research efforts and has potential for use in gene therapy and vaccination strategies. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Professional phagocytes, such as macrophages and dendritic cells, function in a diverse array of physiologic, pathologic and immunologic roles in health and disease. Hence, the cellular biology of these lineages is a focus for numerous research areas. A major barrier to these research efforts is the high resistance of primary macrophages and dendritic cells to DNA transfection, together with the biological insufficiencies of transfectionpermissive macrophage- and DC-like cell lines. Transfection resistance of the primary phagocytes is thought to be due to the innate ability of these cells to degrade foreign nucleic acids within the endolysosomal system (Burke et al., 2002). Methods
⁎ Corresponding author at: Department of Comparative Biology and Experimental Medicine, Faculty of Veterinary Medicine, University of Calgary, 3330 Hospital Dr. NW, Calgary, AB, Canada, T2N 4N1. Tel.: + 1 403 210 6249; fax: + 1 403 210 7882. E-mail address: [email protected] (R.M. Yates). 0022-1759/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2011.06.001
that circumvent DNA degradation within these compartments, such as electroporation, also result in lower transfection efficiencies in phagocytes, indicating that there are additional factors contributing to this resistance (Takahashi et al., 1992; Burke et al., 2002). While some success has been achieved using viral methods of gene transfer, these are typically limited by the size of foreign DNA and their suitability and selectivity for in vivo applications. In addition, the construction of recombinant viruses, particularly for a large number of constructs, is often time-prohibitive. Hence non-viral methods of gene delivery to primary phagocytes, although typically yielding lower transfection efficiencies and cell viabilities, are still preferred in certain scenarios. In particular, non-viral transfection of ex vivo and in vivo macrophages and dendritic cells has generated interest for the potential to use these cells as targets or vehicles in gene therapy (Burke et al., 2002). Electroporation of macrophages can be achieved using systems such as the commercially available Nucleofector™ kit marketed by Lonza. This method, although popular for macrophage-like cell lines, is
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
limited to in vitro applications and is associated with low viability and altered cellular morphology (Van De Parre et al., 2005). Due to the limitations of current methodologies for transfection of primary bone marrow derived macrophages (BMMØs), we sought to develop a novel, non-viral, phagocytespecific method of targeted gene delivery. The method we have developed utilizes intraphagosomal sonoporation mediated by internalized lipid-based microbubbles. Originally developed as contrast agents for use in diagnostic ultrasonography, gas-filled encapsulated microbubbles (herein referred to as microbubbles) have recently received attention for their potential use as gene delivery vectors in genetic therapy (Newman and Bettinger, 2007). Microbubbles are typically 1–3 μm in diameter and consist of heavy-gas cores (perfluorocarbons or sulphur hexafluoride) with stabilizing shells composed of lipid, protein or synthetic polymer (Lindner, 2004). In addition to the acoustic properties, which allow microbubbles to be used as contrast agents, inertial cavitation of microbubbles following insonation results in a large, localized release of mechanical energy (Sundaram et al., 2003; Schlicher et al., 2006). This energy can be released in the form of hydrodynamic shock waves with induced pressures of 4 mPa, microjets with a velocity of 5.5 m s − 1, and pressures of 15 kPa. Both shockwaves and microjets are capable of transiently breaching biological membranes which have an estimated critical pressure threshold of 3 kPa (Prentice et al., 2005; Zhao et al., 2008). This phenomenon has been termed “microbubble-mediated sonoporation”, and has been shown to produce pores up to 100 nm in diameter with half lives of a few seconds (Newman and Bettinger, 2007). Introduction of microbubble-associated plasmid DNA into non-phagocytic eukaryotic cells through microbubble-induced sonoporation has been demonstrated both in vitro and in vivo (Newman and Bettinger, 2007). We hypothesized that we could utilize the phagocytic machinery of primary phagocytes to facilitate their transfection through intraphagosomal sonoporation by the cavitation of internalized microbubbles. Furthermore, we reasoned that sonoporation of perinuclear phagosomes would yield a higher efficiency of transfection due to their proximity to the nucleus and be better tolerated than sonoporation of the plasma membrane. To specifically target phagocytes and to stimulate phagocytic uptake, we developed a cationic-, biotinylated-, lipid-coated microbubble that could be opsonized with an anti-biotin immunoglobulin (IgG) antibody and to which plasmid DNA could be electrostatically bound to the cationic shell. Following phagocytosis, cavitation of the microbubble could be induced by external insonation, which in some cases could result in transfection of the primary phagocyte. This method of transfection we term “microbubble-mediated intra-phagosomal sonoporation (MIPS)”. Although in vitro transfection efficiencies using this novel methodology were modest, transfected phagocytes were morphologically indistinguishable from untransfected controls. Moreover, using this approach we achieved the in vivo transfection of murine peritoneal macrophages. This not only demonstrated the potential of this and similar reagents in future genetic therapies targeting phagocytes, but more immediately provided a research tool that permits transfection and expression of DNA constructs within phagocytes of animal models.
153
2. Materials and methods 2.1. Mice, cells and constructs C57BL/6 (WT) mice were purchased from Charles River Laboratories. All animal experiments were conducted according to protocols approved by the University of Calgary Animal Care and Use Committee. Murine bone marrow-derived macrophages (BMMØs) were derived from 8 to 12 week old mice, as previously described (Yates et al., 2005). Murine peritoneal macrophages (PMØs) were isolated by peritoneal lavage with cold phosphate buffered saline pH 7.2 (PBS) of euthanized 8– 12 week old mice without elicitation. The macrophage and DClike cell lines RAW 264.7 (ATCC) and DC 2.4 cells (kindly provided by Dr. Yan Shi, University of Calgary) were maintained as previously described (Desrosiers et al., 2007). The 4.7-kb plasmid pmaxFP-Green-(C/N) (Lonza) was used in all optimization experiments and contains the fluorescent green protein pmaxFP-Green open reading frame under a CMV IE promoter. Plasmids were prepared using E.N.Z.A. endo-free plasmid maxi kits (Omega bio-tek). 2.2. Microbubble preparation Microbubbles were formed by the dispersion of a mixture of the phospholipids 1,2-dipalmitoyl-sn-glycero3-phosphocholine (DPPC), 1,2-dipalmitoyl-sn-glycero-3phosphoethanolamine-N-[methoxy(polyethylene glycol)2000] (ammonium salt) (DPPE-PEG2000), 1,2-dipalmitoyl3-trimethylammonium-propane (chloride salt) (TAPs), and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[biotinyl(polyethylene glycol)-2000] (ammonium salt) (biotinyl-PEG2000) (60.2:10:28.2:1.5, m:m:m:m) (Avanti Polar Lipids, Alabaster, Ala.), in a diluent comprising normal saline, glycerol, and propylene glycol (8:0.5:1.5, v:v:v). Samples were suspended in the diluent mixtures at a final lipid concentration of 2 mg/mL followed by heating to 55 °C until dissolution occurred. The sample was then allowed to cool to room temperature. The colloidal solution was then added to 2 mL vials (Wheaton, St. Louis, Mo.), which were then capped and sealed. The air was purged from the vials and replaced with N90% by volume perfluorobutane gas (Fluoromed, Round Rock, Tx.) using an in-house manifold. Microbubbles were activated utilizing a vial-mix agitator (Bristol-Myers Squibb Medical Imaging Inc.) at 4000– 4500 rpm for 45 s. Once activated the bubbles were washed with PBS and opsonized with the addition of 1 mL PBS containing 0.8 mg anti-biotin IgG (rabbit) (Rockland Immunochemicals) and incubation for 10 min. Unbound IgG was removed by washing with growth medium. Endotoxinfree plasmid DNA (pmaxFP-Green) was added to the solution followed by a 10 minute incubation period at the concentration indicated. 2.3. In vitro transfection Differentiated cells were transferred to 35 mm tissue culture treated Petri dishes at a concentration of 2 × 106 cells/ dish, and incubated overnight. Following incubation, the growth medium was removed from each dish and replaced with a 200 μL microbubble mixture consisting of growth
154
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
medium containing 7.5–15 × 107 opsonized microbubbles/mL. The dishes were then inverted and incubated at 37 °C for 5 min in 7% CO2 to facilitate the cellular binding of microbubbles. Growth medium (2 mL) was added and cells were further incubated for 5 min to ensure complete phagocytosis. Ultrasound was applied to the bottom of each dish using a fixed 1 MHz transducer (PMT medical) at 0.5 W cm− 1 for 3 s, unless otherwise stated. Insonation with higher frequencies was achieved using a 2.5-8 MHz transducer on a M5 portable diagnostic ultrasound (Mindray). 2.4. In vivo transfection Control and treated mice were injected intraperitoneally with 200 μL prepared microbubble solution in PBS (1.5–3 × 10 8 microbubbles/mL) with or without 20 μg adsorbed DNA. After 20 min, 2.5 MHz ultrasound was applied to the abdomen of each mouse for 45 s using a standard diagnostic ultrasound transducer with mechanical index (MI) of 1.0 to cavitate phagocytosed microbubbles. Mice were sacrificed at 24–48 h post injection and the peritoneal cavity imaged or flushed. 2.5. Flow cytometry and microscopy Transfection efficiencies were determined by flow cytometry using a Cell Lab Quanta™ cytometer with Kaluza software v1.0 (Beckman Coulter). A total of 30,000 cells were analysed per sample with viable cells gated by Coulter volume. Images of transfected in vitro or ex vivo cells were generated by an SP5 scanning laser confocal microscope (Leica). Fluorescent and phase max projections were then constructed using Metamorph software and the projections superimposed. Images of microbubbles or microbubble-containing cells were taken using an Olympus IX-70 phase contrast-equipped microscope. Secondary detection of rabbit IgG on the surface of microbubbles was achieved with incubation of Alexa Fluor 488 goat anti-rabbit IgG (Molecular Probes) (10 μg/mL) in PBS containing 5% bovine serum albumin (m/v). Visualization of endosomes and lysosomes was achieved by pre-incubation of macrophages with the non-toxic, membrane impermeable fluorescent tracer Alexa Fluor 594 hydrazide (Molecular Probes) (10 μg/mL in growth media) for 18 h prior to a 1 hour chase and phagocytosis of microbubbles (Yates et al., 2005). Images of GFP transfected peritoneal cells in situ were taken using an Olympus OV100 microscopy system. Control and transfected images were generated using identical settings. 3. Results and discussion Preliminary “trial and error” experimentation with microbubble formulation, opsonization, handling and delivery to BMMØs (not described) led to the base-line BMMØ transfection protocol using MIPS. In brief, BMMØs were seeded in 35 mm tissue culture treated Petri dishes at a concentration of 2 × 10 6 cells/dish and allowed to adhere overnight. A biotinylated, cationic lipid-based microbubble suspension was synthesized and activated by agitation in perfluorobutane. Microbubbles were washed with PBS, opsonized with rabbit anti-biotin IgG, and incubated with endotoxin-free pmaxFP-Green plasmid DNA (Fig. 1A). IgG opsonization could
be confirmed by detection of the opsonizing IgG with a secondary anti-rabbit IgG fluorescent antibody (Fig. 1A insert). Growth medium was removed from BMMØs and replaced with a 200 μL of growth medium containing approximately 1.5–3 × 10 7 microbubbles (approximately 8–15 microbubbles per BMMØ). To promote BMMØ contact with the buoyant microbubbles, Petri dishes were inverted for 5 min at 37 °C in 7% CO2. The meniscus formed by the microbubble suspension was sufficient to protect the inverted BMMØ monolayer from desiccation over this period. The Petri dishes were subsequently incubated in the upright position with additional growth medium for a further 5 min to ensure complete phagocytosis of adhered microbubbles (Fig. 1B). Microbubble localization within phagolysosomes was demonstrated in live BMMØs pulsed with the lysosomal tracer Alexa fluor 594 hydrazide (Fig. 1C, D). To promote inertial cavitation of the intraphagosomal microbubbles, Petri dishes were insonated for 3 s using a 1 MHz transducer. This was achieved by placing an ultrasound transducer under the sonolucent Petri dish with ultrasound gel to ensure transfer of acoustic energy to the cells. This treatment resulted in the complete cavitation of all intracellular microbubbles (Fig. 1E). The BMMØs were then incubated at 37°C in 7% CO2 to allow expression of introduced constructs. This treatment was generally well tolerated by the BMMØs and resulted in transfection efficiencies between 2 and 8% (Fig. 1F). Insonation with plasmid DNA alone did not result in any transfection, indicating that microbubble-mediation was necessary for membrane sonoporation. Additionally, the absence of IgG opsonization resulted in significantly reduced transfection efficiencies (44.5 ± 10% reduction), suggesting that IgGmediated phagocytosis was necessary to obtain optimal transfection efficiencies. Following the establishment of the basic intraphagosomal microbubble-mediated transfection protocol, we next sought to optimize parameters such as the concentration of plasmid DNA, microbubble:BMMØ ratios and the frequencies of insonation. Transfection efficiencies of BMMØs with pmaxFPGreen were determined after 48 h using flow cytometry with cells gated by Coulter volume. As predicted, the transfection efficiencies correlated to the amount of plasmid DNA loaded onto the microbubbles up to 10 μg/plate (Fig. 2A). The number of phagocytosed microbubbles was also found to correlate with the transfection efficiencies, but was inversely correlated with cell viability, which we reasoned was due to unrecoverable membrane or organellar damage (Fig. 2B). This could have been exacerbated by plasma membrane damage resulting from the cavitation of bound, but not internalized excess microbubbles. Frequency of insonation between 1 and 8 MHz, with a constant mechanical index, generally produced comparable levels of transfection and cell death (Fig. 2C). In general, 10 μg of DNA per 3.5 mm plate, 2–6 microbubbles/cell and insonation at 1–5 MHz produced optimal transfection efficiencies with maximal cellular viability. We next determined whether transfection via MIPS could be used to transfect primary bone marrow derived dendritic cells (BMDCs) and the macrophage- and dendritic cell-like cell lines RAW264.7 and DC-2.4. Using the parameters previously optimized for BMMØs, we successfully transfected these phagocytes with pmaxFP-Green (Fig. 3A–E). Intriguingly, the MIPS approach was less efficient at transfecting the
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
155
Fig. 1. Phase contrast micrographs depicting MIPS transfection of BMMØs in vitro. (A) Micrograph of washed, opsonized microbubbles following loading with plasmid DNA. Bar represents 10 μm. (A insert) Micrographs depicting presence of opsonizing rabbit anti-biotin IgG on the surface of microbubbles as detected by fluorescently conjugated anti-rabbit IgG secondary antibody. (Ai, ii) Microbubble opsonized with anti-biotin IgG. (Aiii, iv) Microbubble not opsonized with antibiotin IgG (negative control). (Ai, iii) Phase contrast and (Aii, iv) fluorescent micrographs. (B) BMMØs following phagocytosis of microbubbles. Arrows indicate internalized microbubbles. (C, D) Phase contrast and fluorescent images of microbubbles within BMMØ phagolysosomes 15 min after internalization. The red fluorescent tracer Alexa fluor 594 hydrazide had been previously pulsed and chased into endolysosomes before microbubble internalization. (C) Phase-fluorescent image overlay. (D) Red fluorescent image depicting lysosomal contents within microbubble-containing phagosomes. Arrows indicate internalized microbubbles. (E) BMMØs directly after insonation showing absence of all internalized microbubbles. (F) Phase-GFP image overlay depicting BMMØ expression of GFP 24 h following MIPS transfection.
macrophage- and DC-like cell lines which are generally permissive to commercially available transfection methodologies. We reasoned that this was most likely due to the lower phagocytic index of these immortalized cell lines, although cell-specific optimization of the protocol could improve trans-
fection efficiencies. Nonetheless, transfection of these phagocyte-like cell lines can be achieved through MIPS. Since the RAW264.7 and DC2.4 cell lines, as opposed to primary BMMØ and BMDCs, can be currently transfected using commercially available transfection approaches (such as electroporation), a
156
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
Fig. 2. Optimization of MIPS transfection parameters for BMMØ transfection. The efficiencies of transfection with pmaxFP-Green are expressed as the percentage of GFP positive cells as detected by flow cytometry with cells gated by Coulter volume. Viability was determined by trypan blue exclusion directly following insonation. (A) BMMØ transfection efficiencies achieved using 1, 2.5, 5, 10 and 20 μg plasmid DNA per 35 mm dish. (B) BMMØs transfection efficiencies and viabilities were determined as a function of the number of microbubbles added to each 35 mm dish. (C) BMMØs transfection efficiencies and viabilities as a function of frequency of insonation with a constant mechanical index of 1.0. Graphs represent averaged data from 4 independent experiments. Error bars represent SEM.
MIPS approach may not be preferable for these transfectionpermissive cells. A current drawback of many transfection approaches is their inability to be adapted to transfect populations of cells in vivo. Transfection via microbubble-mediated sonoporation however, does not have this limitation. More specifically, by targeting phagocytic cells expressing phagocytic Fc receptors, the MIPS approach has the potential to preferentially transfect macrophages and FcR-expressing dendritic cells in animal models. We set out to demonstrate the in vivo use of MIPS transfection through pmaxFR-Green transfection of murine peritoneal macrophages in vivo. IgG-microbubbleDNA mixtures were prepared as outlined, suspended in 200 μL PBS and injected into the peritoneal cavity of C57BL/6 mice at a dose of 3–6 × 10 7 microbubbles/mouse and 20 μg DNA/mouse. Ultrasound was applied to the abdomen of mice 20 min post injection using a diagnostic ultrasound transducer at a frequency of 2.5 MHz and a mechanical index of 1.0 for a period of 45 s. Mice were returned to routine husbandry
conditions and sacrificed at 24 or 48 h for harvesting of peritoneal macrophages or in situ imaging respectively. Peritoneal macrophages were harvested by flushing with cold PBS and the cells allowed to adhere overnight to glassbottom fluorodishes before imaging with confocal microscopy (Fig. 3G, I). The percentage of isolated peritoneal macrophages expressing GFP were typically low (less than 1–2%), reflecting either a lower in vivo transfection rate or macrophage turnover in the peritoneum. Aside from the demonstration of MIPS-transfection of macrophages in vivo, this methodology allows for the immediate ex vivo investigation of transfected peritoneal macrophages where the expression of the protein of interest has occurred entirely within an in vivo environment. Further to the isolation of transfected peritoneal macrophages, detection of GFPpositive cells in situ was performed on freshly exposed abdominal organs using an Olympus OV100 microscopy system. Although differentiation of phagocytic and bystander peritoneal cells could not be achieved at this resolution, it is
Fig. 3. Demonstration of in vitro and in vivo MIPS transfection of phagocytes with pmaxFP-Green. (A) Transfection efficiencies and (B–E) representative confocal micrographs of (B) BMMØs, (C) BMDCs, (D) RAW264.7 and (E) DC2.4 cells following MIPS transfection in vitro. (F–I) Demonstration of MIPS transfection of peritoneal macrophages in vivo. Microbubble suspensions with (F–G) or without (H–I) pmaxFP-Green DNA (20 μg) were injected in the peritoneal cavity of mice followed by the application of abdominal ultrasound using a 2.5 MHz diagnostic ultrasound transducer. (F, H) in situ fluorescent micrographs (GFP filter set) of lateral abdominal wall of freshly euthanized mice 48 h after in vivo MIPS transfection. (G, I) Scanning confocal images of ex vivo peritoneal macrophages retrieved by peritoneal lavage of euthanized mice 24 h after in vivo MIPS transfection.
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
157
158
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
apparent that numerous GFP-positive cells, many morphologically consistent with macrophages, are present within the peritoneal space (Fig. 3F, G). 4. Concluding remarks Here we outline a novel methodology that enables the transfection of primary phagocytes known for their high resistance to non-viral gene delivery. While transfection efficiencies achieved through MIPS are modest, this approach is perfectly suited for live-cell, microscopy-based studies which are typically limited to the investigation of transfection permissive phagocyte-like cell lines. The ability to utilize primary phagocytes in these studies would circumvent the biological peculiarities of these immortalized cell lines. The use of MIPS transfection of peritoneal macrophages in vivo will additionally permit the study of phagocyte biology within live animals, or immediately ex vivo, diminishing artefacts arising through extended ex vivo culture. Integration of MIPS transfection of primary phagocytes within existing research programs can be achieved with minimal setup and troubleshooting. The application of in vivo MIPS transfection methodologies to gene therapy and vaccination strategies are under investigation. Acknowledgments We thank Dr. Yan Shi for his critical reading of the manuscript. This work was supported by the Natural Sciences and Engineering Research Council of Canada and Alberta Innovates- Health Solutions.
References Burke, B., Sumner, S., Maitland, N., Lewis, C.E., 2002. Macrophages in gene therapy: cellular delivery vehicles and in vivo targets. J. Leukoc. Biol. 72, 417. Desrosiers, M.D., Cembrola, K.M., Fakir, M.J., Stephens, L.A., Jama, F.M., Shameli, A., Mehal, W.Z., Santamaria, P., Shi, Y., 2007. Adenosine deamination sustains dendritic cell activation in inflammation. J. Immunol. 179, 1884. Lindner, J.R., 2004. Microbubbles in medical imaging: current applications and future directions. Nat. Rev. Drug Discov. 3, 527. Newman, C.M., Bettinger, T., 2007. Gene therapy progress and prospects: ultrasound for gene transfer. Gene Ther. 14, 465. Prentice, P., Cuschieri, A., Dholakia, K., Prausnitz, M., Campbell, P., 2005. Membrane disruption by optically controlled microbubble cavitation. Nat. Phys. 1, 107. Schlicher, R.K., Radhakrishna, H., Tolentino, T.P., Apkarian, R.P., Zarnitsyn, V., Prausnitz, M.R., 2006. Mechanism of intracellular delivery by acoustic cavitation. Ultrasound Med. Biol. 32, 915. Sundaram, J., Mellein, B.R., Mitragotri, S., 2003. An experimental and theoretical analysis of ultrasound-induced permeabilization of cell membranes. Biophys. J. 84, 3087. Takahashi, M., Furukawa, T., Tanaka, I., Nikkuni, K., Aoki, A., Kishi, K., Koike, T., Moriyama, Y., Shibata, A., 1992. Gene introduction into granulocytemacrophage progenitor cells by electroporation: the relationship between introduction efficiency and the proportion of cells in S-phase. Leuk. Res. 16, 761. Van De Parre, T.J., Martinet, W., Schrijvers, D.M., Herman, A.G., De Meyer, G.R., 2005. mRNA but not plasmid DNA is efficiently transfected in murine J774A.1 macrophages. Biochem. Biophys. Res. Commun. 327, 356. Yates, R.M., Hermetter, A., Russell, D.G., 2005. The kinetics of phagosome maturation as a function of phagosome/lysosome fusion and acquisition of hydrolytic activity. Traffic 6, 413. Zhao, Y.Z., Luo, Y.K., Lu, C.T., Xu, J.F., Tang, J., Zhang, M., Zhang, Y., Liang, H.D., 2008. Phospholipids-based microbubbles sonoporation pore size and reseal of cell membrane cultured in vitro. J. Drug Target. 16, 18.
Contents lists available at ScienceDirect
Journal of Immunological Methods j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j i m
Technical Note
In vitro and in vivo transfection of primary phagocytes via microbubble-mediated intraphagosomal sonoporation Jason C.M. Lemmon a, Ryan J. McFarland b, Joanna M. Rybicka a, Dale R. Balce a, Kyle R. McKeown b, Regina M. Krohn a, Terry O. Matsunaga b, Robin M. Yates a, c,⁎ a b c
Department of Comparative Biology and Experimental Medicine, Faculty of Veterinary Medicine, University of Calgary, AB, Canada Department of Radiology Research, University of Arizona, AZ, USA Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Calgary, AB, Canada
a r t i c l e
i n f o
Article history: Received 23 February 2011 Received in revised form 9 May 2011 Accepted 1 June 2011 Available online 14 June 2011 Keywords: Phagocytosis Macrophage Dendritic cell Gene therapy DNA vaccine
a b s t r a c t The professional phagocytes, such as macrophages and dendritic cells, are the subject of numerous research efforts in immunology and cell biology. The use of primary phagocytes in these investigations however, are limited by their inherent resistance to transfection with DNA constructs. As a result, the use of phagocyte-like immortalized cell lines is widespread. While these cell lines are transfection permissive, they are generally regarded as poor biological substitutes for primary phagocytes. By exploiting the phagocytic machinery of primary phagocytes, we developed a non-viral method of DNA transfection of macrophages that employs intraphagosomal sonoporation mediated by internalized lipid-based microbubbles. This approach enables the transfection of primary phagocytes in vitro, with a modest, but reliable efficiency. Furthermore, this methodology was readily adapted to transfect murine peritoneal macrophages in vivo. This technology has immediate application to current research efforts and has potential for use in gene therapy and vaccination strategies. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Professional phagocytes, such as macrophages and dendritic cells, function in a diverse array of physiologic, pathologic and immunologic roles in health and disease. Hence, the cellular biology of these lineages is a focus for numerous research areas. A major barrier to these research efforts is the high resistance of primary macrophages and dendritic cells to DNA transfection, together with the biological insufficiencies of transfectionpermissive macrophage- and DC-like cell lines. Transfection resistance of the primary phagocytes is thought to be due to the innate ability of these cells to degrade foreign nucleic acids within the endolysosomal system (Burke et al., 2002). Methods
⁎ Corresponding author at: Department of Comparative Biology and Experimental Medicine, Faculty of Veterinary Medicine, University of Calgary, 3330 Hospital Dr. NW, Calgary, AB, Canada, T2N 4N1. Tel.: + 1 403 210 6249; fax: + 1 403 210 7882. E-mail address: [email protected] (R.M. Yates). 0022-1759/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2011.06.001
that circumvent DNA degradation within these compartments, such as electroporation, also result in lower transfection efficiencies in phagocytes, indicating that there are additional factors contributing to this resistance (Takahashi et al., 1992; Burke et al., 2002). While some success has been achieved using viral methods of gene transfer, these are typically limited by the size of foreign DNA and their suitability and selectivity for in vivo applications. In addition, the construction of recombinant viruses, particularly for a large number of constructs, is often time-prohibitive. Hence non-viral methods of gene delivery to primary phagocytes, although typically yielding lower transfection efficiencies and cell viabilities, are still preferred in certain scenarios. In particular, non-viral transfection of ex vivo and in vivo macrophages and dendritic cells has generated interest for the potential to use these cells as targets or vehicles in gene therapy (Burke et al., 2002). Electroporation of macrophages can be achieved using systems such as the commercially available Nucleofector™ kit marketed by Lonza. This method, although popular for macrophage-like cell lines, is
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
limited to in vitro applications and is associated with low viability and altered cellular morphology (Van De Parre et al., 2005). Due to the limitations of current methodologies for transfection of primary bone marrow derived macrophages (BMMØs), we sought to develop a novel, non-viral, phagocytespecific method of targeted gene delivery. The method we have developed utilizes intraphagosomal sonoporation mediated by internalized lipid-based microbubbles. Originally developed as contrast agents for use in diagnostic ultrasonography, gas-filled encapsulated microbubbles (herein referred to as microbubbles) have recently received attention for their potential use as gene delivery vectors in genetic therapy (Newman and Bettinger, 2007). Microbubbles are typically 1–3 μm in diameter and consist of heavy-gas cores (perfluorocarbons or sulphur hexafluoride) with stabilizing shells composed of lipid, protein or synthetic polymer (Lindner, 2004). In addition to the acoustic properties, which allow microbubbles to be used as contrast agents, inertial cavitation of microbubbles following insonation results in a large, localized release of mechanical energy (Sundaram et al., 2003; Schlicher et al., 2006). This energy can be released in the form of hydrodynamic shock waves with induced pressures of 4 mPa, microjets with a velocity of 5.5 m s − 1, and pressures of 15 kPa. Both shockwaves and microjets are capable of transiently breaching biological membranes which have an estimated critical pressure threshold of 3 kPa (Prentice et al., 2005; Zhao et al., 2008). This phenomenon has been termed “microbubble-mediated sonoporation”, and has been shown to produce pores up to 100 nm in diameter with half lives of a few seconds (Newman and Bettinger, 2007). Introduction of microbubble-associated plasmid DNA into non-phagocytic eukaryotic cells through microbubble-induced sonoporation has been demonstrated both in vitro and in vivo (Newman and Bettinger, 2007). We hypothesized that we could utilize the phagocytic machinery of primary phagocytes to facilitate their transfection through intraphagosomal sonoporation by the cavitation of internalized microbubbles. Furthermore, we reasoned that sonoporation of perinuclear phagosomes would yield a higher efficiency of transfection due to their proximity to the nucleus and be better tolerated than sonoporation of the plasma membrane. To specifically target phagocytes and to stimulate phagocytic uptake, we developed a cationic-, biotinylated-, lipid-coated microbubble that could be opsonized with an anti-biotin immunoglobulin (IgG) antibody and to which plasmid DNA could be electrostatically bound to the cationic shell. Following phagocytosis, cavitation of the microbubble could be induced by external insonation, which in some cases could result in transfection of the primary phagocyte. This method of transfection we term “microbubble-mediated intra-phagosomal sonoporation (MIPS)”. Although in vitro transfection efficiencies using this novel methodology were modest, transfected phagocytes were morphologically indistinguishable from untransfected controls. Moreover, using this approach we achieved the in vivo transfection of murine peritoneal macrophages. This not only demonstrated the potential of this and similar reagents in future genetic therapies targeting phagocytes, but more immediately provided a research tool that permits transfection and expression of DNA constructs within phagocytes of animal models.
153
2. Materials and methods 2.1. Mice, cells and constructs C57BL/6 (WT) mice were purchased from Charles River Laboratories. All animal experiments were conducted according to protocols approved by the University of Calgary Animal Care and Use Committee. Murine bone marrow-derived macrophages (BMMØs) were derived from 8 to 12 week old mice, as previously described (Yates et al., 2005). Murine peritoneal macrophages (PMØs) were isolated by peritoneal lavage with cold phosphate buffered saline pH 7.2 (PBS) of euthanized 8– 12 week old mice without elicitation. The macrophage and DClike cell lines RAW 264.7 (ATCC) and DC 2.4 cells (kindly provided by Dr. Yan Shi, University of Calgary) were maintained as previously described (Desrosiers et al., 2007). The 4.7-kb plasmid pmaxFP-Green-(C/N) (Lonza) was used in all optimization experiments and contains the fluorescent green protein pmaxFP-Green open reading frame under a CMV IE promoter. Plasmids were prepared using E.N.Z.A. endo-free plasmid maxi kits (Omega bio-tek). 2.2. Microbubble preparation Microbubbles were formed by the dispersion of a mixture of the phospholipids 1,2-dipalmitoyl-sn-glycero3-phosphocholine (DPPC), 1,2-dipalmitoyl-sn-glycero-3phosphoethanolamine-N-[methoxy(polyethylene glycol)2000] (ammonium salt) (DPPE-PEG2000), 1,2-dipalmitoyl3-trimethylammonium-propane (chloride salt) (TAPs), and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[biotinyl(polyethylene glycol)-2000] (ammonium salt) (biotinyl-PEG2000) (60.2:10:28.2:1.5, m:m:m:m) (Avanti Polar Lipids, Alabaster, Ala.), in a diluent comprising normal saline, glycerol, and propylene glycol (8:0.5:1.5, v:v:v). Samples were suspended in the diluent mixtures at a final lipid concentration of 2 mg/mL followed by heating to 55 °C until dissolution occurred. The sample was then allowed to cool to room temperature. The colloidal solution was then added to 2 mL vials (Wheaton, St. Louis, Mo.), which were then capped and sealed. The air was purged from the vials and replaced with N90% by volume perfluorobutane gas (Fluoromed, Round Rock, Tx.) using an in-house manifold. Microbubbles were activated utilizing a vial-mix agitator (Bristol-Myers Squibb Medical Imaging Inc.) at 4000– 4500 rpm for 45 s. Once activated the bubbles were washed with PBS and opsonized with the addition of 1 mL PBS containing 0.8 mg anti-biotin IgG (rabbit) (Rockland Immunochemicals) and incubation for 10 min. Unbound IgG was removed by washing with growth medium. Endotoxinfree plasmid DNA (pmaxFP-Green) was added to the solution followed by a 10 minute incubation period at the concentration indicated. 2.3. In vitro transfection Differentiated cells were transferred to 35 mm tissue culture treated Petri dishes at a concentration of 2 × 106 cells/ dish, and incubated overnight. Following incubation, the growth medium was removed from each dish and replaced with a 200 μL microbubble mixture consisting of growth
154
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
medium containing 7.5–15 × 107 opsonized microbubbles/mL. The dishes were then inverted and incubated at 37 °C for 5 min in 7% CO2 to facilitate the cellular binding of microbubbles. Growth medium (2 mL) was added and cells were further incubated for 5 min to ensure complete phagocytosis. Ultrasound was applied to the bottom of each dish using a fixed 1 MHz transducer (PMT medical) at 0.5 W cm− 1 for 3 s, unless otherwise stated. Insonation with higher frequencies was achieved using a 2.5-8 MHz transducer on a M5 portable diagnostic ultrasound (Mindray). 2.4. In vivo transfection Control and treated mice were injected intraperitoneally with 200 μL prepared microbubble solution in PBS (1.5–3 × 10 8 microbubbles/mL) with or without 20 μg adsorbed DNA. After 20 min, 2.5 MHz ultrasound was applied to the abdomen of each mouse for 45 s using a standard diagnostic ultrasound transducer with mechanical index (MI) of 1.0 to cavitate phagocytosed microbubbles. Mice were sacrificed at 24–48 h post injection and the peritoneal cavity imaged or flushed. 2.5. Flow cytometry and microscopy Transfection efficiencies were determined by flow cytometry using a Cell Lab Quanta™ cytometer with Kaluza software v1.0 (Beckman Coulter). A total of 30,000 cells were analysed per sample with viable cells gated by Coulter volume. Images of transfected in vitro or ex vivo cells were generated by an SP5 scanning laser confocal microscope (Leica). Fluorescent and phase max projections were then constructed using Metamorph software and the projections superimposed. Images of microbubbles or microbubble-containing cells were taken using an Olympus IX-70 phase contrast-equipped microscope. Secondary detection of rabbit IgG on the surface of microbubbles was achieved with incubation of Alexa Fluor 488 goat anti-rabbit IgG (Molecular Probes) (10 μg/mL) in PBS containing 5% bovine serum albumin (m/v). Visualization of endosomes and lysosomes was achieved by pre-incubation of macrophages with the non-toxic, membrane impermeable fluorescent tracer Alexa Fluor 594 hydrazide (Molecular Probes) (10 μg/mL in growth media) for 18 h prior to a 1 hour chase and phagocytosis of microbubbles (Yates et al., 2005). Images of GFP transfected peritoneal cells in situ were taken using an Olympus OV100 microscopy system. Control and transfected images were generated using identical settings. 3. Results and discussion Preliminary “trial and error” experimentation with microbubble formulation, opsonization, handling and delivery to BMMØs (not described) led to the base-line BMMØ transfection protocol using MIPS. In brief, BMMØs were seeded in 35 mm tissue culture treated Petri dishes at a concentration of 2 × 10 6 cells/dish and allowed to adhere overnight. A biotinylated, cationic lipid-based microbubble suspension was synthesized and activated by agitation in perfluorobutane. Microbubbles were washed with PBS, opsonized with rabbit anti-biotin IgG, and incubated with endotoxin-free pmaxFP-Green plasmid DNA (Fig. 1A). IgG opsonization could
be confirmed by detection of the opsonizing IgG with a secondary anti-rabbit IgG fluorescent antibody (Fig. 1A insert). Growth medium was removed from BMMØs and replaced with a 200 μL of growth medium containing approximately 1.5–3 × 10 7 microbubbles (approximately 8–15 microbubbles per BMMØ). To promote BMMØ contact with the buoyant microbubbles, Petri dishes were inverted for 5 min at 37 °C in 7% CO2. The meniscus formed by the microbubble suspension was sufficient to protect the inverted BMMØ monolayer from desiccation over this period. The Petri dishes were subsequently incubated in the upright position with additional growth medium for a further 5 min to ensure complete phagocytosis of adhered microbubbles (Fig. 1B). Microbubble localization within phagolysosomes was demonstrated in live BMMØs pulsed with the lysosomal tracer Alexa fluor 594 hydrazide (Fig. 1C, D). To promote inertial cavitation of the intraphagosomal microbubbles, Petri dishes were insonated for 3 s using a 1 MHz transducer. This was achieved by placing an ultrasound transducer under the sonolucent Petri dish with ultrasound gel to ensure transfer of acoustic energy to the cells. This treatment resulted in the complete cavitation of all intracellular microbubbles (Fig. 1E). The BMMØs were then incubated at 37°C in 7% CO2 to allow expression of introduced constructs. This treatment was generally well tolerated by the BMMØs and resulted in transfection efficiencies between 2 and 8% (Fig. 1F). Insonation with plasmid DNA alone did not result in any transfection, indicating that microbubble-mediation was necessary for membrane sonoporation. Additionally, the absence of IgG opsonization resulted in significantly reduced transfection efficiencies (44.5 ± 10% reduction), suggesting that IgGmediated phagocytosis was necessary to obtain optimal transfection efficiencies. Following the establishment of the basic intraphagosomal microbubble-mediated transfection protocol, we next sought to optimize parameters such as the concentration of plasmid DNA, microbubble:BMMØ ratios and the frequencies of insonation. Transfection efficiencies of BMMØs with pmaxFPGreen were determined after 48 h using flow cytometry with cells gated by Coulter volume. As predicted, the transfection efficiencies correlated to the amount of plasmid DNA loaded onto the microbubbles up to 10 μg/plate (Fig. 2A). The number of phagocytosed microbubbles was also found to correlate with the transfection efficiencies, but was inversely correlated with cell viability, which we reasoned was due to unrecoverable membrane or organellar damage (Fig. 2B). This could have been exacerbated by plasma membrane damage resulting from the cavitation of bound, but not internalized excess microbubbles. Frequency of insonation between 1 and 8 MHz, with a constant mechanical index, generally produced comparable levels of transfection and cell death (Fig. 2C). In general, 10 μg of DNA per 3.5 mm plate, 2–6 microbubbles/cell and insonation at 1–5 MHz produced optimal transfection efficiencies with maximal cellular viability. We next determined whether transfection via MIPS could be used to transfect primary bone marrow derived dendritic cells (BMDCs) and the macrophage- and dendritic cell-like cell lines RAW264.7 and DC-2.4. Using the parameters previously optimized for BMMØs, we successfully transfected these phagocytes with pmaxFP-Green (Fig. 3A–E). Intriguingly, the MIPS approach was less efficient at transfecting the
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
155
Fig. 1. Phase contrast micrographs depicting MIPS transfection of BMMØs in vitro. (A) Micrograph of washed, opsonized microbubbles following loading with plasmid DNA. Bar represents 10 μm. (A insert) Micrographs depicting presence of opsonizing rabbit anti-biotin IgG on the surface of microbubbles as detected by fluorescently conjugated anti-rabbit IgG secondary antibody. (Ai, ii) Microbubble opsonized with anti-biotin IgG. (Aiii, iv) Microbubble not opsonized with antibiotin IgG (negative control). (Ai, iii) Phase contrast and (Aii, iv) fluorescent micrographs. (B) BMMØs following phagocytosis of microbubbles. Arrows indicate internalized microbubbles. (C, D) Phase contrast and fluorescent images of microbubbles within BMMØ phagolysosomes 15 min after internalization. The red fluorescent tracer Alexa fluor 594 hydrazide had been previously pulsed and chased into endolysosomes before microbubble internalization. (C) Phase-fluorescent image overlay. (D) Red fluorescent image depicting lysosomal contents within microbubble-containing phagosomes. Arrows indicate internalized microbubbles. (E) BMMØs directly after insonation showing absence of all internalized microbubbles. (F) Phase-GFP image overlay depicting BMMØ expression of GFP 24 h following MIPS transfection.
macrophage- and DC-like cell lines which are generally permissive to commercially available transfection methodologies. We reasoned that this was most likely due to the lower phagocytic index of these immortalized cell lines, although cell-specific optimization of the protocol could improve trans-
fection efficiencies. Nonetheless, transfection of these phagocyte-like cell lines can be achieved through MIPS. Since the RAW264.7 and DC2.4 cell lines, as opposed to primary BMMØ and BMDCs, can be currently transfected using commercially available transfection approaches (such as electroporation), a
156
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
Fig. 2. Optimization of MIPS transfection parameters for BMMØ transfection. The efficiencies of transfection with pmaxFP-Green are expressed as the percentage of GFP positive cells as detected by flow cytometry with cells gated by Coulter volume. Viability was determined by trypan blue exclusion directly following insonation. (A) BMMØ transfection efficiencies achieved using 1, 2.5, 5, 10 and 20 μg plasmid DNA per 35 mm dish. (B) BMMØs transfection efficiencies and viabilities were determined as a function of the number of microbubbles added to each 35 mm dish. (C) BMMØs transfection efficiencies and viabilities as a function of frequency of insonation with a constant mechanical index of 1.0. Graphs represent averaged data from 4 independent experiments. Error bars represent SEM.
MIPS approach may not be preferable for these transfectionpermissive cells. A current drawback of many transfection approaches is their inability to be adapted to transfect populations of cells in vivo. Transfection via microbubble-mediated sonoporation however, does not have this limitation. More specifically, by targeting phagocytic cells expressing phagocytic Fc receptors, the MIPS approach has the potential to preferentially transfect macrophages and FcR-expressing dendritic cells in animal models. We set out to demonstrate the in vivo use of MIPS transfection through pmaxFR-Green transfection of murine peritoneal macrophages in vivo. IgG-microbubbleDNA mixtures were prepared as outlined, suspended in 200 μL PBS and injected into the peritoneal cavity of C57BL/6 mice at a dose of 3–6 × 10 7 microbubbles/mouse and 20 μg DNA/mouse. Ultrasound was applied to the abdomen of mice 20 min post injection using a diagnostic ultrasound transducer at a frequency of 2.5 MHz and a mechanical index of 1.0 for a period of 45 s. Mice were returned to routine husbandry
conditions and sacrificed at 24 or 48 h for harvesting of peritoneal macrophages or in situ imaging respectively. Peritoneal macrophages were harvested by flushing with cold PBS and the cells allowed to adhere overnight to glassbottom fluorodishes before imaging with confocal microscopy (Fig. 3G, I). The percentage of isolated peritoneal macrophages expressing GFP were typically low (less than 1–2%), reflecting either a lower in vivo transfection rate or macrophage turnover in the peritoneum. Aside from the demonstration of MIPS-transfection of macrophages in vivo, this methodology allows for the immediate ex vivo investigation of transfected peritoneal macrophages where the expression of the protein of interest has occurred entirely within an in vivo environment. Further to the isolation of transfected peritoneal macrophages, detection of GFPpositive cells in situ was performed on freshly exposed abdominal organs using an Olympus OV100 microscopy system. Although differentiation of phagocytic and bystander peritoneal cells could not be achieved at this resolution, it is
Fig. 3. Demonstration of in vitro and in vivo MIPS transfection of phagocytes with pmaxFP-Green. (A) Transfection efficiencies and (B–E) representative confocal micrographs of (B) BMMØs, (C) BMDCs, (D) RAW264.7 and (E) DC2.4 cells following MIPS transfection in vitro. (F–I) Demonstration of MIPS transfection of peritoneal macrophages in vivo. Microbubble suspensions with (F–G) or without (H–I) pmaxFP-Green DNA (20 μg) were injected in the peritoneal cavity of mice followed by the application of abdominal ultrasound using a 2.5 MHz diagnostic ultrasound transducer. (F, H) in situ fluorescent micrographs (GFP filter set) of lateral abdominal wall of freshly euthanized mice 48 h after in vivo MIPS transfection. (G, I) Scanning confocal images of ex vivo peritoneal macrophages retrieved by peritoneal lavage of euthanized mice 24 h after in vivo MIPS transfection.
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
157
158
J.C.M. Lemmon et al. / Journal of Immunological Methods 371 (2011) 152–158
apparent that numerous GFP-positive cells, many morphologically consistent with macrophages, are present within the peritoneal space (Fig. 3F, G). 4. Concluding remarks Here we outline a novel methodology that enables the transfection of primary phagocytes known for their high resistance to non-viral gene delivery. While transfection efficiencies achieved through MIPS are modest, this approach is perfectly suited for live-cell, microscopy-based studies which are typically limited to the investigation of transfection permissive phagocyte-like cell lines. The ability to utilize primary phagocytes in these studies would circumvent the biological peculiarities of these immortalized cell lines. The use of MIPS transfection of peritoneal macrophages in vivo will additionally permit the study of phagocyte biology within live animals, or immediately ex vivo, diminishing artefacts arising through extended ex vivo culture. Integration of MIPS transfection of primary phagocytes within existing research programs can be achieved with minimal setup and troubleshooting. The application of in vivo MIPS transfection methodologies to gene therapy and vaccination strategies are under investigation. Acknowledgments We thank Dr. Yan Shi for his critical reading of the manuscript. This work was supported by the Natural Sciences and Engineering Research Council of Canada and Alberta Innovates- Health Solutions.
References Burke, B., Sumner, S., Maitland, N., Lewis, C.E., 2002. Macrophages in gene therapy: cellular delivery vehicles and in vivo targets. J. Leukoc. Biol. 72, 417. Desrosiers, M.D., Cembrola, K.M., Fakir, M.J., Stephens, L.A., Jama, F.M., Shameli, A., Mehal, W.Z., Santamaria, P., Shi, Y., 2007. Adenosine deamination sustains dendritic cell activation in inflammation. J. Immunol. 179, 1884. Lindner, J.R., 2004. Microbubbles in medical imaging: current applications and future directions. Nat. Rev. Drug Discov. 3, 527. Newman, C.M., Bettinger, T., 2007. Gene therapy progress and prospects: ultrasound for gene transfer. Gene Ther. 14, 465. Prentice, P., Cuschieri, A., Dholakia, K., Prausnitz, M., Campbell, P., 2005. Membrane disruption by optically controlled microbubble cavitation. Nat. Phys. 1, 107. Schlicher, R.K., Radhakrishna, H., Tolentino, T.P., Apkarian, R.P., Zarnitsyn, V., Prausnitz, M.R., 2006. Mechanism of intracellular delivery by acoustic cavitation. Ultrasound Med. Biol. 32, 915. Sundaram, J., Mellein, B.R., Mitragotri, S., 2003. An experimental and theoretical analysis of ultrasound-induced permeabilization of cell membranes. Biophys. J. 84, 3087. Takahashi, M., Furukawa, T., Tanaka, I., Nikkuni, K., Aoki, A., Kishi, K., Koike, T., Moriyama, Y., Shibata, A., 1992. Gene introduction into granulocytemacrophage progenitor cells by electroporation: the relationship between introduction efficiency and the proportion of cells in S-phase. Leuk. Res. 16, 761. Van De Parre, T.J., Martinet, W., Schrijvers, D.M., Herman, A.G., De Meyer, G.R., 2005. mRNA but not plasmid DNA is efficiently transfected in murine J774A.1 macrophages. Biochem. Biophys. Res. Commun. 327, 356. Yates, R.M., Hermetter, A., Russell, D.G., 2005. The kinetics of phagosome maturation as a function of phagosome/lysosome fusion and acquisition of hydrolytic activity. Traffic 6, 413. Zhao, Y.Z., Luo, Y.K., Lu, C.T., Xu, J.F., Tang, J., Zhang, M., Zhang, Y., Liang, H.D., 2008. Phospholipids-based microbubbles sonoporation pore size and reseal of cell membrane cultured in vitro. J. Drug Target. 16, 18.