Targeted retroviral gene delivery using ultrasound - Wiley Online Library

THE JOURNAL OF GENE MEDICINE RESEARCH J Gene Med 2007; 9: 77–87. Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/jgm.1003

ARTICLE

Targeted retroviral gene delivery using ultrasound

Sarah L. Taylor1 Ahad A. Rahim1 Nigel L. Bush2 Jeffrey C. Bamber2 Colin D. Porter1 * 1

The Institute of Cancer Research, Section of Cell and Molecular Biology, 237 Fulham Rd., London SW3 6JB, UK 2

The Institute of Cancer Research and The Royal Marsden NHS Trust, Joint Department of Physics, Downs Road, Sutton, SM2 5PT, UK *Correspondence to: Colin D. Porter, The Institute of Cancer Research, 237 Fulham Rd., London SW3 6JB, UK. E-mail: [email protected]

Abstract Background Achieving specificity of delivery represents a major problem limiting the clinical application of retroviral vectors for gene therapy, whilst lack of efficiency and longevity of gene expression limit non-viral techniques. Ultrasound and microbubble contrast agents can be used to effect plasmid DNA delivery. We therefore sought to evaluate the potential for ultrasound/microbubble-mediated retroviral gene delivery. Methods An envelope-deficient retroviral vector, inherently incapable of target cell entry, was combined with cationic microbubbles and added to target cells. The cells were exposed to pulsed 1 MHz ultrasound for 5 s and subsequently analysed for marker gene expression. The acoustic pressure profile of the ultrasound field, to which transduction efficiency was related, was determined using a needle hydrophone. Results Ultrasound-targeted gene delivery to a restricted area of cells was achieved using virus-loaded microbubbles. Gene delivery efficiency was up to 2% near the beam focus. Significant transduction was restricted to areas exposed to ≥ 0.4 MPa peak-negative acoustic pressure, despite uniform application of the vector. An acoustic pressure-dependence was demonstrated that can be exploited for targeted retroviral transduction. The mechanism of entry likely involves membrane perturbation in the vicinity of oscillating microbubbles, facilitating fusion of the viral and cell membranes. Conclusions We have established the basis of a novel retroviral vector technology incorporating favourable aspects of existing viral and non-viral gene delivery vectors. In particular, transduction can be controlled by means of ultrasound exposure. The technology is ideally suited to targeted delivery following systemic vector administration. Copyright  2007 John Wiley & Sons, Ltd. Keywords

ultrasound; microbubble contrast agent; retrovirus; targeted delivery

Introduction

Received: 20 June 2006 Revised: 20 November 2006 Accepted: 5 December 2006

Copyright  2007 John Wiley & Sons, Ltd.

Neither viral nor non-viral approaches have satisfactorily addressed the requirement for efficient, targeted transgene delivery into cells [1]. Viral vectors are capable of efficient gene delivery, of which retro- and lentiviral vectors can achieve sustained gene expression following genome integration. However, widespread use of such vectors is compromised by the lack of control over gene delivery to target cells. Specificity of delivery is a critical issue for situations best suited to a systemic route of administration, for which attempts have focused on modification of the vector itself [2]. In contrast, several ‘physical’ means have been adopted for non-viral vectors [3], some of

78

which are amenable to targeted delivery; e.g. cells become more receptive to DNA uptake when exposed to ultrasound, due to temporary modification of cell membrane permeability [4]. However, low efficiency and transient gene expression are issues that limit non-viral vectors. In this study, we seek to combine the respective strengths of the viral and non-viral approaches. Ultrasound exposure can enhance delivery of naked DNA into cells [5,6]. Although the exact mechanism is unclear, gas bubble oscillation (and/or collapse) in the ultrasound field [7] generates shear forces [8–10] or jetting phenomena [11] in the cell growth medium which are the likely cause of transient cell membrane pores, through which DNA may enter the cytoplasm [12–14]. Efficiency can be increased significantly by the addition of microbubbles – stabilised gas bubbles used for image contrast – to the transfection medium [15–17], and the technique can be further developed by attachment of the genetic material to the microbubble [18,19]. In vivo, ultrasound has been used to deliver therapeutic genes to muscle [20] or tumour [21] following direct injection of microbubbles and DNA, or to the heart following intravenous administration [22,23]. The further development of microbubbles as molecular imaging probes by incorporating targeting ligands [24] promises to improve accumulation of systemically administered vector at the target site [25], prior to insonation. However, while this technique enables specific gene delivery by means of controlled ultrasound exposure, the transgene is lost when cell division occurs, and gene expression remains short-term; this represents a major limitation for some therapeutic applications. Murine leukaemia virus (MLV)-based vectors enter the cell by receptor-mediated virus-cell membrane fusion, resulting in the cytoplasmic release of the uncoated core. With few exceptions, envelope engineering for retargeted binding and entry has not been successful [2]. MLV particles lacking the envelope protein are incapable of infection, although the cores are still competent if the block to entry can be overcome by other means [26]. Since, in respect of their systemic utility and capacity for being targeted, microbubbles possess the properties that have been sought for retroviral vectors, and since they can additionally achieve DNA entry dependent on ultrasound exposure, we investigated whether microbubbles could be used to effect targeted, ultrasound-dependent entry of the core of an envelopedeficient retroviral vector.

S. L. Taylor et al.

10% foetal calf serum (FCS). For gene delivery experiments, TE671 cells were seeded in OptiCell cell culture units (BioCrystal Inc., Westerville, OH, USA) and incubated in DMEM (10% FCS) to reach 60% confluence. TELCeB6 cells, carrying the MFGnlsLacZ retroviral vector and expressing Moloney-MLV gag-pol proteins, produce viral particles with a vector genome encoding nuclearlocalised β-galactosidase that are non-infectious because they lack envelope proteins; derivative TELCeB6/AF7 cells produce infectious amphotropic particles [27]. Envelopedeficient and infectious MLV were harvested in serum-free OptiMEM I (Invitrogen Ltd., Paisley, UK) from confluent cultures of TELCeB6 and TELCeB6/AF7 cells, respectively, and the supernatant passed through a 0.45 µm filter before use. Mature capsid (p30) protein content was equivalent by Western blot analysis (see below) with/without envelope protein expression, indicating equivalent virus particle production and maturation; the particle concentration in such preparations has previously been determined to be 7 × 108 /ml by enumeration of individual particles visualised by immunofluorescence [28].

Preparation of virus-loaded microbubble vector

Materials and methods

Cationic lipid-shelled, perfluorocarbon-filled microbubbles [19] were kindly supplied by Alexander Klibanov, University of Virginia. Excess lipid was removed by centrifugation (50 g, 5 min) of the preparation in a syringe barrel, to which a tap was attached; this allowed for subsequent removal of the liquid fraction whilst retaining the microbubbles that had concentrated above due to their lower density. The tap was closed and the microbubble fraction resuspended in phosphate-buffered saline (PBS). Microbubbles were washed by two further rounds of centrifugation/flotation and their concentration was determined using a haemocytometer. To attach envelope-deficient virus to microbubbles sufficient for three experimental replicates, 6 × 107 microbubbles were added to 6 ml virus (4.2 × 109 particles) and the mixture incubated at room temperature for 30 min to allow an electrostatic attachment. The microbubbles were washed twice, as above, to remove unbound virus and recover microbubble-associated virus. The virus-loaded microbubble fraction was resuspended in 30 ml OptiMEM I before application to cells (10 ml per OptiCell ). Plasmid DNA-loaded microbubbles were similarly obtained following incubation of 50 µg pNGVL1β-gal [29,30] per OptiCell with 2 × 107 microbubbles.

Cell culture and envelope-deficient retrovirus production

Western blot analysis of microbubble and retrovirus association

Human rhabdomyosarcoma TE671 cells (and virus producer cells derived therefrom) were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with

Samples taken from each step of the process of producing virus-loaded microbubbles were diluted in SDSloading buffer, heat-denatured and loaded onto a 12%


J Gene Med 2007; 9: 77–87. DOI: 10.1002/jgm

79

Ultrasound-Targeted Retroviral Transduction

acrylamide gel for sodium dodecyl sulfate polyacrylamide electrophoresis (SDS-PAGE) analysis of viral capsid (p30) protein content. The separated viral proteins were then transferred to a nitrocellulose membrane. The membrane was ‘blocked’ in 5% milk in Tris-buffered saline overnight, followed by incubation with goat anti-Rauscher leukaemia virus p30 antibody (Quality Biotech Inc., Camden, NJ, USA). The membrane was washed and incubated with a horseradish peroxidase-conjugated rabbit anti-goat secondary antibody (Dako Ltd., Ely, UK) in 5% milk in Tris-buffered saline. Finally, the membrane was washed and developed using enhanced chemiluminescence (Amersham Biosciences, Buckinghamshire, UK) and autoradiography.

Ultrasound exposure tank and apparatus A water tank was built for reproducible cell exposure to ultrasound at the focal distance of a 1 MHz transducer (Figure 1A). The transducer was held at the base of the tank against a Mylar window, using contact gel for efficient ultrasound transmission. A positioning device enabled vertical alignment of the beam for a reproducible acoustic field. To avoid ultrasound reflection at the surface of the plastic tissue culture vessel and avoid the presence of a medium/air interface above the cells, which could attenuate the ultrasound and/or create standing waves and thus undermine efforts to control ultrasound exposure, we used the OptiCell culture system for insonation of cells. The OptiCell consists of two 100-µm-thick membranes, treated for cell growth, which are acoustically transparent [15,30]; the membranes create a sterile, 2-mm-deep 10 ml chamber, accessible through self-sealing ports. The OptiCell was supported orthogonal to the beam axis, such that the membrane holding the cell monolayer was at the height of the transducer focus. An anechoic absorber was placed at the top of the tank to prevent reflection of the ultrasound. Pulsed 1 MHz ultrasound signals were produced using a computer-controlled sinusoidal waveform signal generator, and fed to the transducer via a power amplifier [30]. The transducer was a single piezoelectric element of 20 mm diameter, spherically focused with a radius of curvature of 67 mm (Imasonic, Besan¸con, France).

Measurement of the acoustic field Acoustic pressure in the ultrasound field was measured using a calibrated 0.2 mm diameter needle hydrophone (Precision Acoustics Ltd., Dorchester, UK) and digital oscilloscope [30]. The transducer was fixed at the base of the tank, and the hydrophone tip placed at the focal distance. The hydrophone was moved under computer Copyright  2007 John Wiley & Sons, Ltd.

control along orthogonal axes in the focal plane; peakto-peak and peak-negative pressures were measured in 0.5 mm increments.

Ultrasound-mediated transduction protocol The growth medium was replaced with 10 ml OptiMEM I containing 1 ml virus alone, 1 ml virus mixed with 1 × 107 microbubbles, or virus-loaded microbubbles (prepared as described above). Note that to make the combined vector and account for loss of approximately half of each component during the association and washing steps, 2 ml virus (1.4 × 109 particles) per OptiCell was initially incubated with 2 × 107 microbubbles. After injection of the microbubbles into the OptiCell , each unit was allowed to stand for 10 min to allow the microbubbles to rise up against the cells. Insonations were performed by placing the OptiCell at the focal distance of the transducer in the custom-built water tank (see above). The tank was filled with water at a temperature of 37 ◦ C, in which the cells were allowed to acclimatise for 60 s prior to ultrasound exposure. Exposure to 1 MHz pulsed ultrasound was performed at 1 kHz pulse repetition frequency, with ten identical sinusoidal cycles per pulse (pulse length) and at an amplifier input voltage amplitude of 60 or 80 mV, respectively producing a focal peak-negative pressure of 1.0 or 1.2 MPa. Overall exposure time was 5 s. After ultrasound exposure the cells were allowed to recover for 4 h at 37 ◦ C before replacing the medium with DMEM (10% FCS) and incubating cells for a further 48 h. Controls were sham-exposed to provide a ‘no ultrasound’ comparison; additionally, because the experimental setup resulted in a distribution of acoustic pressure across the cell monolayer, all experiments were internally controlled for ultrasound-independent background events because the outermost regions of the OptiCell were essentially unexposed.

Gene expression assay Cells were fixed with 0.5% gluteraldehyde (15 min, room temperature) and histochemically stained with 1 mg/ml X-Gal at 37 ◦ C for 4 h, as described previously [31]. Numbers of cells with blue nuclei were counted by observation with a microscope within three boundaries with radii at 2.5, 7.5 and 13 mm from the beam axis. The counts were then normalised, to account for varying cell number between experiments, by dividing the number of stained cells by the total number of cells. Total cell counts were used to assess viability of cells exposed to ultrasound relative to sham-exposed controls since non-viable cells did not remain attached to the OptiCell membrane. Additionally, in order to evaluate the spatial distribution of transduction, efficiency was determined for each 2 mm diameter microscopic field in a 13 × 13 grid centred on the beam axis. J Gene Med 2007; 9: 77–87. DOI: 10.1002/jgm

80

S. L. Taylor et al.

(a)

(b)

(c) Figure 1. Equipment used for ultrasound exposure, and measurement of the acoustic field in the tank. (A) Schematic representation of the exposure tank used for insonating a cell monolayer in the focal plane of the transducer. The cells are adherent to the upper TM TM of the two acoustically transparent OptiCell membranes; virus and microbubbles are added to the OptiCell prior to insonation. An acoustic absorber prevents unwanted reflections. (B) Acoustic pressure within the ultrasound field of the 1 MHz transducer was determined by tracking a 0.2 mm diameter needle hydrophone in the focal plane of the transducer, which is also the plane in which the cell monolayer is positioned in the tank during exposure. Plots are shown of the peak-negative pressure amplitude as a function of radial distance from the acoustic beam axis along orthogonal axes (black and grey) using an amplifier input amplitude of 80 mV (peak 1.2 MPa). (C) Hydrophone trace at the transducer focus, demonstrating the smooth leading/trailing-edged pulse of constant acoustic pressure amplitude

Results Ultrasound/microbubble-mediated retroviral transduction system Target TE671 cells adherent to the acoustically transparent membrane of an OptiCell cell culture unit were exposed to ultrasound in the presence of cationic Copyright  2007 John Wiley & Sons, Ltd.

lipid microbubbles and envelope-deficient retrovirus. The OptiCell was held at the focal distance of a 1 MHz singleelement transducer, orthogonal to the beam axis, in a purpose-built ultrasound exposure apparatus; an acoustic absorber was placed behind to prevent reflections from the water surface, which would result in additional exposure or the establishment of standing wave conditions that would complicate measurement of the J Gene Med 2007; 9: 77–87. DOI: 10.1002/jgm

81


acoustic field to which the cells and microbubbles were exposed (Figure 1A). The transducer was driven by a pulsed sinusoidal signal with an amplitude determined in preliminary experiments to be capable of achieving microubble destruction; similarly, the exposure duration was limited to 5 s to satisfy the constraint that a high level (95%) of cell viability should be maintained. The acoustic pressure varied equivalently along orthogonal axes through the focus (Figure 1B): peak-negative pressure ≥1 MPa was reached over an area with 5 mm diameter, with a maximum value of 1.2 MPa at the centre of the beam; pressures of ≥0.4 and ≥0.2 MPa corresponded to diameters of 15 and 26 mm, respectively. The acoustic pressure range was thus more extensive than the range of values that we have used previously for DNA delivery [29]. Envelope-deficient retrovirus carrying the nuclearlocalised β-galactosidase marker gene was added to TE671 cells with cationic lipid microbubbles for exposure to ultrasound. Cationic microbubbles were chosen in anticipation of their electrostatic association with retroviral particles, comparable to the use of the former for binding plasmid DNA [19] and the association of retroviruses with cationic liposomes [26]. Ultrasound pulses that consisted of ten identical cycles of a sinusoidal wave of 1 MHz frequency (Figure 1C) were generated with a repetition frequency of 1 kHz, amounting to 50 ms of ultrasound ‘on time’ during the 5 s exposure. In an initial experiment, there was no transduction with virus alone, or when the microbubbles or ultrasound were omitted; in contrast, transduction was detected near the centre of the area exposed to ultrasound (albeit at a mean efficiency of only 0.01%) when virus and microbubbles were added together and cells were exposed to ultrasound, indicating that ultrasound-induced microbubble behaviour is capable of effecting entry of the viral core.

Association of envelope-deficient retrovirus and cationic microbubbles To enhance their association – with a view to improving the efficiency of transduction – microbubbles and virus were incubated before addition to the cells. Because of the geometry of the experiment (Figure 1A), association with the microbubbles would be expected to bring virus into close proximity to the cells during exposure, mainly via buoyancy of the microbubbles, but also from the effect of acoustic radiation force in the direction of sound propagation [32]. Microbubbles were incubated with virus and then washed by flotation under low-speed centrifugation, to separate microbubble-associated virus from free virus. Association was assessed by Western blot analysis of viral capsid protein in both the microbubble and wash fractions: approximately 50% of virus was retained by the microbubbles (Figure 2A). Virus-loaded microbubbles were observed microscopically to have attached homogeneously to the cell monolayer during brief (10 min) incubation with cells prior to Copyright  2007 John Wiley & Sons, Ltd.

(a)

(b) Figure 2. Gene delivery using a virus-loaded microbubble vector, as a function of the peak-negative acoustic pressure to which the cells were exposed. (A) Western blot analysis of viral capsid protein content of the input virus, two successive wash fractions obtained during preparation of the virus-loaded microbubble vector, and of the final vector preparation used in transduction experiments. Envelope-deficient virus was incubated with microbubbles, following which the microbubble fraction was recovered by flotation whilst being gently centrifuged; unconjugated virus remained in the underlying medium (wash fraction 1, w1). The microbubble fraction was resuspended in PBS and recovered in the same way, resulting in wash fraction 2 (w2) and the preparation of virus-loaded microbubbles (‘final vector’). Samples from each step were normalised to the original volume of virus used (‘input virus’) to assess the efficiency of association of virus with the microbubbles. (B) Dependence of ultrasound-mediated transduction by virus-loaded microbubbles on peak-negative acoustic pressure amplitude, determined by counting the number of β-galactosidase-expressing cells in regions of the cell monolayer defined by boundaries centred on the acoustic axis. With an amplifier input amplitude of 80 mV (peak 1.2 MPa) radii at 2.5, 7.5 and 13 mm defined regions corresponding to peak-negative pressure >1, 0.4–1, 0.2–0.4 and 0.3 MPa. Transduction was greatly (>100-fold) enhanced using virus-loaded microbubbles with the ultrasound exposure conditions used above; β-galactosidase-positive cells were scored within annuli bounded by circles of diameter 5, 15 and 26 mm centred on the focus, from which an estimation of acoustic pressure-dependence was obtained (Figure 2B). An acoustic pressure threshold existed for entry of the envelope-deficient virus, with little transduction occurring for peak-negative pressure 0.4 MPa. (Being microbubbleand incubation time-dependent, the low background was attributed to the cationicity of the microbubbles directly aiding virus entry [26].)

Spatial localisation of transduction The data in Figure 2B provide averages and standard deviations (SDs) of mean transduction efficiency within acoustic pressure bands defined by radial position of cells in the focal plane of the ultrasound source. The SDs help in assessing the differences between means in different bands; however, the averaging of cell counts within bands presupposes spatial homogeneity of transduction around each annulus. To evaluate the spatial correlation of transduction with acoustic pressure more carefully, and particularly to confirm the expected symmetry of transduction about the beam axis, the transduction efficiency was determined for each 2 mm diameter microscopic field in a 13 × 13 grid centred on the axis (Figure 3). Transduction was symmetrical, with significant efficiencies restricted to areas exposed to ≥ 0.4 MPa peak-negative pressure. Gene delivery was reduced at the very centre of the beam, which was also associated with a reduction (15%) in cell viability.

Acoustic pressure-dependence of transduction The transducer had been designed such that the circularly symmetrical fall in acoustic pressure within the ultrasound beam with increasing distance from the beam axis would be sufficiently gentle to provide for spatial registration of transduced cells with the local pressure: i.e. to enable the relationship between acoustic pressure and transduction to be determined by spatially registering the counts of transduced cells with the local pressure values. Whilst the data in Figure 3 verified acoustic pressure-dependence, the resolution of analysis (which was limited by the transduction efficiency) was insufficient to derive such Copyright  2007 John Wiley & Sons, Ltd.

S. L. Taylor et al.

a relationship. Transduction efficiency could, however, be related to the spatial average acoustic pressure for each of the annuli. Moreover, by varying the transducer input amplitude, and therefore the spatial averages, it was possible to determine intermediate values (Figure 4). Above a minimum threshold of 0.2 MPa there was a monotonic increase in transduction efficiency with increasing spatial average peak-negative pressure. The SDs associated with the efficiency measurements were larger for the higher values of pressure/efficiency due to the greater variation intrinsic to counting fewer cells within the smaller areas; i.e. the datapoints for higher efficiency nevertheless reflect fewer positive cells because the region area (and hence the total number of cells exposed to the respective pressure) is much less. For this reason it is not clear whether the relationship has reached saturation by 1 MPa or whether a further increase in efficiency would result from greater acoustic pressures. Cell viability was unaffected in this experiment even for the band with the highest acoustic pressure.

Virus concentration-dependence of transduction To determine the dependence of transduction efficiency on the virus concentration, 2-fold serially diluted preparations were incubated with microbubbles, which were subsequently washed and used for ultrasoundmediated transduction, as described above. Whilst there was a trend for greater levels of transduction for more concentrated virus, it was evident that the dependence approximated linearity only for the more diluted virus and that the above experiments were conducted with a concentration that was saturating (Figure 5). This indicates that factors other than the virus concentration place the limit on transduction efficiency.

Stability of gene expression Stability of gene expression following ultrasound/microbubble-mediated transduction with envelope-deficient virus was explored both by serial passage of transduced cells and by clonal expansion of such cells following limiting dilution. Control OptiCells were exposed to envelope-deficient virus that was not microbubbleassociated, or to microbubble-complexed plasmid DNA encoding nuclear-localised β-galactosidase; infectious retrovirus (without microbubbles or ultrasound) was used as a positive control. The whole experiment was performed in quadruplicate, of which one replicate was stained with the cells in situ, whilst the cells from the centre of the monolayer were recovered and processed from the other three: a portion of the OptiCell membrane (approximately 2 cm × 2 cm) was excised, washed and immersed in trypsin solution. Cells were subsequently washed and plated in 24-well plates for serial passage over the ensuing 3 weeks; duplicate plates at each J Gene Med 2007; 9: 77–87. DOI: 10.1002/jgm


83

TM

Figure 3. Spatial distribution of retroviral gene delivery relative to the acoustic beam axis. (A) Photograph of an OptiCell following transduction and assay for β-galactosidase expression; marks represent the position of one or more cells with blue nuclei as a crude demonstration that transduction was localised to the area closest to the transducer focus. (B) Detailed analysis of percentage of transduced cells in fields of view within an area exposed to ultrasound of peak-negative pressure >0.2 MPa for the three experiments described in Figure 2. The data represent the mean efficiency for each field of view. Circles indicate the boundaries, centred on the beam axis, for peak-negative pressures of 1, 0.4 and 0.2 MPa (radii of 2.5, 7.5 and 13 mm, respectively). The density of viable cells was equivalent throughout the regions exposed to 100fold enhancement of transduction over their simultaneous addition to the cells. Commercial neutral lipid microbubbles (SonoVue , Bracco) failed both to associate with virus and to achieve transduction; albumin-shelled microbubbles (Optison , Amersham International) were similarly ineffective. One reason for this is that attaching particles to the surface of a buoyant microbubble aids their mass transfer to the cell surface. Additionally, close proximity of viral particles and microbubbles may be required for ultrasound-induced fusion; certainly, contact of the virus-loaded microbubbles with the target cell was essential, consistent with a mechanism of entry involving localised membrane perturbation: exposure of cells growing on the lower membrane of the OptiCell (by inversion of the unit), 2 mm below the accumulation of microbubbles at the upper membrane due to their buoyancy, abolished gene delivery (data not shown). The transduction efficiency of ultrasound/microbubblemediated envelope-deficient virus gene delivery can be compared with that of enveloped virus by considering the number of transduction events versus the frequency of virus particle encounter with the cell monolayer. For the amphotropic control virus used in Figure 6, 1.4 × 108 particles in an OptiCell resulted in transduction of 3% of the monolayer of 1.7 × 106 cells for 1 h exposure. In this case, particle encounter with the cell monolayer was limited by Brownian motion in the medium, with a diffusion limit of ∼200 µm for 1 h [34]; since the medium depth was 2 mm, the theoretical number of particles encountering the cell monolayer was 1.4 × 107 (i.e. ∼8 particles/cell), resulting in a transduction efficiency of ∼0.4% per particle. For virus-loaded microbubbles, a particle/microbubble ratio of 70 : 1 and microbubble/cell ratio of ∼6 : 1 (consistent with microscopic observations) led to ∼400 particles/cell and a peak transduction efficiency of 2% (the maximum local efficiency recorded in Figure 3B); this corresponds to a particle transduction efficiency of ∼0.005% (although, since the level of virus particle loading on the microbubble was saturating with respect to efficiency, this is an under-estimate). The calculated relative particle transduction efficiency is thus J Gene Med 2007; 9: 77–87. DOI: 10.1002/jgm

86

very similar to the value of 1–2% for entry of this virus mediated by the cationic liposome DOTAP [26]. However, in the present case, delivery followed ultrasound exposure for 5 s only and was spatially localised. Whilst rapid entry is likely to be important for clinical application, and vector circulation will enable cumulative transduction at the site of insonation due to continual replenishment, there is much scope for improving the efficiency of delivery. There are a number of ways in which to seek to optimise ultrasound/microbubble-mediated transduction with envelope-deficient retrovirus. Although there is sufficient virus loading of microbubbles (based on the virus concentration-dependence of Figure 5) it may be of interest to modify the nature of the association, either by altering the amount of charged lipid in the microbubble formulation or by using a ligand-based approach (e.g. avidin/biotin), since avidity of binding may be an issue affecting entry. Modification of the lipid composition of the microbubble shell or addition of substances to improve membrane permeability/fluidity [35] may increase transduction by facilitating fusion. Varying the ultrasound exposure parameters will also likely affect efficiency, since the optimal peak-negative acoustic pressure range of 0.4–1 MPa relates specifically to the single parameter values chosen for pulse length, repetition rate and exposure time; indeed, from the relationship of Figure 4B, it is not clear that the maximum effective pressure has been reached, although further increase may be offset by loss in cell viability. Finally, attaching targeting ligands to the microbubble shell will induce close apposition of membranes to enhance fusion. Indeed, such molecular targeting may be necessary for in vivo application, since microbubbles exhibit similar behaviour to red blood cells in the circulation, moving at high speeds through the vasculature [36]; e.g. microbubbles carrying echistatin accumulate in angiogenic vessels of gliomas via attachment to the receptor αV β3 [37]. Combined with the targeting due to positioning of the ultrasound beam the potential for ligand-modified microbubbles to achieve proximity specifically to the desired target cells (which could be a molecularly distinguishable subset of cells within the insonation zone) offers the means for highly targeted transduction in vivo. Targeting microbubbles to vascular endothelial markers would be an important step in the development of this vector towards clinical application in cancer and cardiovascular disease. The novelty of our approach is that spatial control at a distance is exerted upon retroviral delivery by exposure to ultrasound. For eventual clinical application, this process is non-invasive and the necessary acoustic pressure can be applied specifically to the target site. A further advantage for clinical application is that microbubble accumulation and destruction at the target site can be monitored using ultrasound imaging to assess vector delivery in real-time. By successfully combining the desirable attributes of viral and nonviral technologies, our data both provide the means of achieving stable ultrasound/microbubble-mediated gene delivery and delineate a solution to the critical issue of Copyright  2007 John Wiley & Sons, Ltd.

S. L. Taylor et al.

specificity following systemic administration of retroviral vectors.

Acknowledgements This work was funded by The Institute of Cancer Research and the Biotechnology and Biological Sciences Research Council. We are grateful to Alexander Klibanov and Jonathan Lindner for providing the microbubbles used in these experiments. We thank Gail ter Haar for constructive advice.

References 1. Verma IM, Weitzman MD. Gene therapy: twenty-first century medicine. Annu Rev Biochem 2005; 74: 711–738. 2. Lavillette D, Russell SJ, Cosset FL. Retargeting gene delivery using surface-engineered retroviral vector particles. Curr Opin Biotechnol 2001; 12: 461–466. 3. Wells DJ. Gene therapy progress and prospects: electroporation and other physical methods. Gene Ther 2004; 11: 1363–1369. 4. Newman CM, Lawrie A, Brisken AF, et al. Ultrasound gene therapy: on the road from concept to reality. Echocardiography 2001; 18: 339–347. 5. Fechheimer M, Boylan JF, Parker S, et al. Transfection of mammalian cells with plasmid DNA by scrape loading and sonication loading. Proc Natl Acad Sci U S A 1987; 84: 8463–8467. 6. Kim HJ, Greenleaf JF, Kinnick RR, et al. Ultrasound-mediated transfection of mammalian cells. Hum Gene Ther 1996; 7: 1339–1346. 7. Ayme-Bellegarda EJ. Collapse and rebound of a gas-filled spherical bubble immersed in a diagnostic ultrasonic field. J Acoust Soc Am 1990; 88: 1054–1060. 8. van Wamel A, Bouakaz A, Versluis M, et al. Micromanipulation of endothelial cells: ultrasound-microbubble-cell interaction. Ultrasound Med Biol 2004; 30: 1255–1258. 9. Marmottant P, Hilgenfeldt S. Controlled vesicle deformation and lysis by single oscillating bubbles. Nature 2003; 423: 153–156. 10. Wu J. Theoretical study on shear stress generated by microstreaming surrounding contrast agents attached to living cells. Ultrasound Med Biol 2002; 28: 125–129. 11. Postema M, van Wamel A, Lancee CT, et al. Ultrasound-induced encapsulated microbubble phenomena. Ultrasound Med Biol 2004; 30: 827–840. 12. Bao S, Thrall BD, Miller DL. Transfection of a reporter plasmid into cultured cells by sonoporation in vitro. Ultrasound Med Biol 1997; 23: 953–959. 13. Miller DL, Quddus J. Lysis and sonoporation of epidermoid and phagocytic monolayer cells by diagnostic ultrasound activation of contrast agent gas bodies. Ultrasound Med Biol 2001; 27: 1107–1113. 14. Deng CX, Sieling F, Pan H, et al. Ultrasound-induced cell membrane porosity. Ultrasound Med Biol 2004; 30: 519–526. 15. Miller DL, Dou C, Song J. DNA transfer and cell killing in epidermoid cells by diagnostic ultrasound activation of contrast agent gas bodies in vitro. Ultrasound Med Biol 2003; 29: 601–607. 16. Greenleaf WJ, Bolander ME, Sarkar G, et al. Artificial cavitation nuclei significantly enhance acoustically induced cell transfection. Ultrasound Med Biol 1998; 24: 587–595. 17. Lawrie A, Brisken AF, Francis SE, et al. Microbubble-enhanced ultrasound for vascular gene delivery. Gene Ther 2000; 7: 2023–2027. 18. Frenkel PA, Chen S, Thai T, et al. DNA-loaded albumin microbubbles enhance ultrasound-mediated transfection in vitro. Ultrasound Med Biol 2002; 28: 817–822. 19. Christiansen JP, French BA, Klibanov AL, et al. Targeted tissue transfection with ultrasound destruction of plasmid-bearing cationic microbubbles. Ultrasound Med Biol 2003; 29: 1759–1767. 20. Lu QL, Liang HD, Partridge T, et al. Microbubble ultrasound improves the efficiency of gene transduction in skeletal muscle J Gene Med 2007; 9: 77–87. DOI: 10.1002/jgm


21. 22.

23. 24. 25. 26. 27. 28. 29.

in vivo with reduced tissue damage. Gene Ther 2003; 10: 396–405. Miller DL, Song J. Tumor growth reduction and DNA transfer by cavitation-enhanced high-intensity focused ultrasound in vivo. Ultrasound Med Biol 2003; 29: 887–893. Bekeredjian R, Chen S, Frenkel PA, et al. Ultrasound-targeted microbubble destruction can repeatedly direct highly specific plasmid expression to the heart. Circulation 2003; 108: 1022–1026. Korpanty G, Chen S, Shohet RV, et al. Targeting of VEGFmediated angiogenesis to rat myocardium using ultrasonic destruction of microbubbles. Gene Ther 2005; 12: 1305–1312. Klibanov AL. Ligand-carrying gas-filled microbubbles: ultrasound contrast agents for targeted molecular imaging. Bioconjugate Chem 2005; 16: 9–17. Lindner JR. Molecular imaging with contrast ultrasound and targeted microbubbles. J Nucl Cardiol 2004; 11: 215–221. Porter CD. Rescue of retroviral envelope fusion deficiencies by cationic liposomes. J Gene Med 2002; 4: 622–633. Cosset FL, Takeuchi Y, Battini JL, et al. High-titer packaging cells producing recombinant retroviruses resistant to human serum. J Virol 1995; 69: 7430–7436. Pizzato M, Marlow SA, Blair ED, et al. Initial binding of murine leukemia virus particles to cells does not require specific Envreceptor interaction. J Virol 1999; 73: 8599–8611. Rahim AA, Taylor SL, Bush NL, et al. Spatial and acoustic pressure dependence of microbubble-mediated gene delivery targeted using focused ultrasound. J Gene Med 2006; 8: 1347–1357.


87 30. Rahim A, Taylor SL, Bush NL, et al. Physical parameters affecting ultrasound/microbubble-mediated gene delivery efficiency in vitro. Ultrasound Med Biol 2006; 32: 1269–1279. 31. Porter CD, Lukacs KV, Box G, et al. Cationic liposomes enhance the rate of transduction by a recombinant retroviral vector in vitro and in vivo. J Virol 1998; 72: 4832–4840. 32. Rychak JJ, Klibanov AL, Hossack JA. Acoustic radiation force enhances targeted delivery of ultrasound contrast microbubbles: in vitro verification. IEEE Trans Ultrason Ferroelectr Freq Control 2005; 52: 421–433. 33. Taniyama Y, Tachibana K, Hiraoka K, et al. Development of safe and efficient novel nonviral gene transfer using ultrasound: enhancement of transfection efficiency of naked plasmid DNA in skeletal muscle. Gene Ther 2002; 9: 372–380. 34. Chuck AS, Clarke MF, Palsson BO. Retroviral infection is limited by Brownian motion. Hum Gene Ther 1996; 7: 1527–1534. 35. Nozaki T, Ogawa R, Feril LB Jr, et al. Enhancement of ultrasound-mediated gene transfection by membrane modification. J Gene Med 2003; 5: 1046–1055. 36. Lindner JR, Song J, Jayaweera AR, et al. Microvascular rheology of Definity microbubbles after intra-arterial and intravenous administration. J Am Soc Echocardiogr 2002; 15: 396–403. 37. Ellegala DB, Leong-Poi H, Carpenter JE, et al. Imaging tumor angiogenesis with contrast ultrasound and microbubbles targeted to alpha(v)beta3. Circulation 2003; 108: 336–341.

J Gene Med 2007; 9: 77–87. DOI: 10.1002/jgm