(2008) 15, 257–266 & 2008 Nature Publishing Group All rights reserved 0969-7128/08 $30.00 www.nature.com/gt ORIGINAL ARTICLE Ultrasound-based nonviral induces bone formation in vivo

D Sheyn1, N Kimelman-Bleich1, G Pelled1, Y Zilberman1, D Gazit1,2 and Z Gazit1 1Skeletal Biotech Laboratory, Hebrew University of Jerusalem, Jerusalem, Israel and 2Department of Surgery, International Stem Cell Institute, Cedars Sinai Medical Center, Los Angeles, CA, USA

Nonviral gene delivery is a promising, safe, therapeutic tool transgenic mice that express the Luc gene under the control in regenerative medicine. This study is the first to achieve of a human osteocalcin promoter. Following rhBMP-9 nonviral, ultrasound-based, osteogenic gene delivery that sonoporation, osteocalcin-dependent Luc expression lasted leads to bone tissue formation, in vivo. We hypothesized that for 24 days and peaked on day 10. Bone tissue was formed direct in vivo sonoporation of naked DNA encoding for the in the site of rhBMP-9 delivery, as was shown by micro- osteogenic gene, recombinant human bone morphogenetic computerized tomography and histology. The sonoporation protein-9 (rhBMP-9) would induce bone formation. A method was also compared with previously developed luciferase plasmid (Luc), encoding rhBMP-9 or empty electrotransfer-based gene delivery and was found signifi- pcDNA3 vector mixed with microbubbles, was injected into cantly inferior in its efficiency of gene delivery. We conclude the thigh muscles of mice. After injection, noninvasive that ultrasound-mediated osteogenic gene delivery could sonoporation was applied. Luc activity was monitored serve as a therapeutic solution in conditions requiring bone noninvasively, and quantitatively using bioluminescence tissue regeneration after further development that will imaging in vivo, and found for 14 days with a peak increase the efficiency. expression on day 7. To examine osteogenesis in vivo, Gene Therapy (2008) 15, 257–266; doi:10.1038/sj.gt.3303070; rhBMP-9 plasmid was sonoporated into the thigh muscles of published online 22 November 2007

Keywords: direct gene delivery; sonoporation; bone morphogenetic proteins; bone formation

Introduction in this pilot study can be related to the same group of physical gene delivery as electrotransfer. Nonviral gene delivery approaches hold great promise for Sonoporation (transient ultrasound-induced increase the development of efficient and safe gene therapy for in cell membrane permeability) has been shown in clinical orthopedic applications. Nonviral approaches are several studies to enhance gene transfer both in vitro6–8 considered relatively safe. Unlike their viral counterparts, and in vivo.9–18 The increase in cell membrane perme- very little inflammation or disease is associated with ability is probably caused by oscillation of the membrane nonviral vectors.1 Viral vectors that integrate into the in response to acoustic cavitation, which results from a genome carry the risks of oncogenesis and mutagenesis. nonthermal interaction between a propagating pressure These risks are significantly reduced in nonviral gene wave induced by ultrasound and a gas included in delivery, which results in episomal insertion. Although aqueous medium.19 This oscillation may be amplified by virus-mediated transfer generally produces higher trans- the presence of microbubbles in the medium. The fection efficiencies and a longer period of transgene application of ultrasound causes vibration and collapse expression,2 the transient expression of the transgene of the microbubbles. Destruction of the microbubbles associated with nonviral methods represents a potential causes high-energy microstreams, or microjets, that cause advantage when only a short expression period is desired. shear stress on the membrane of a cell and increases its Common transfection methods, mainly used for in vitro permeability. That is one of the proposed mechanisms by gene delivery, have been shown to be up to 70% effective which microbubbles can lower the threshold for cavita- in well-established cell lines.3 Electrical field-mediated tion.20 In this study we used a suspension of perfluor- gene transfer, better known as DNA electrotransfer, is one opropane-filled albumin microspheres (Optison, of the most efficient methods for in vivo direct gene Amersham Health, Princeton, NJ, USA), which is transfer.4,5 The ultrasound-enhanced gene transfer utilized approved for use in echocardiography by the US Food and Drug Administration as a contrast agent for diagnostic ultrasonography. In numerous recent studies, Correspondence: Dr Z Gazit, Skeletal Biotech Laboratory, Hebrew Optison has been found to noticeably enhance the University of Jerusalem, Hadassah medical campus, Jerusalem efficiency of ultrasound-based gene delivery into skeletal 91120, Israel. 17,21,22 9,13 15 E-mail: [email protected] muscle, tumors, and the liver. Sonoporation 18 Received 3 April 2007; revised 16 October 2007; accepted 16 October has been shown to promote drug delivery and increase 2007; published online 22 November 2007 viral vector efficiency in vivo,10,11,13,14 in addition to its Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 258 major contribution to gene therapy: having the capability Results to transfer naked plasmid DNA locally to target cells. In previous studies, sonoporation-based gene delivery Sonoporation of luciferase reporter gene in vivo was mainly applied in the setting of cardiovascular Delivery of the reporter luciferase (Luc) gene and tissues10–13,17,18,23 and cancer.9,13 quantitative in vivo imaging were used to evaluate the DNA electrotransfer or is the use of efficiency of transfection and to calibrate the sonopora- short high-voltage pulses to overcome the barrier of the tion parameters. Based on preliminary experiments (data cell membrane. The electrical pulse destabilizes the cell not shown) and previous publications,16,21,33 an initial membrane for a short time, and then small molecules can protocol was established. The initial acoustic parameters enter the cell using simple diffusion. Bigger molecules were based on those outlined in previous studies in (such as DNA or RNA) are being ‘pushed’ by the which similar sonoporation systems had been used.11,16,22 electrical current, and so can enter the cell. Electro- The results of in vivo imaging indicated that the poration is now in use for delivery of ions, drugs, dyes, expression of the transgene was higher when the power tracers, antibodies and RNA and DNA into cells.24 The input of the sonoporation device was of high intensity exact entrance mechanism of electroporated DNA into (5 W cmÀ2). We tested two frequencies enabled by the the nucleus is not clear, but it seems that the DNA Sonitron2000 device: 1 and 3 MHz. The most conven- migrates electrophoretically through the membrane and tional frequency used in ultrasound-based gene delivery, then diffuses toward the nucleus.24,25 1 Mhz,11 was more efficient in the induction of In the present study, sonoporation was used to Luc expression than 3 MHz. The effect of exposure transfer osteogenic gene and induce de novo formation duration was tested as well and, as expected, transgene of bone tissue in vivo. It was compared to electroporation expression increased with a longer time of exposure. The as the gold standard of nonviral direct gene delivery injected plasmid amount and the Optison microbubbles method. concentration were also optimized. The results indicated The increasing disparity between the demand and that the 5% concentration of Optison was more effective availability of human donor organs for the replacement than the 10 or 25% concentration, corresponding with of dysfunctional organs has prompted an intensive previous findings.16 The increase in plasmid DNA search for alternative sources of organs. In the field of amounts did not further improve the efficiency of orthopedics, massive bone defects caused by trauma, sonoporation (data not shown). cancer, primary joint replacement, revision joint surgery Based on preliminary findings by us and others,16,21,33 and congenital anomalies continue to pose a great the following experimental protocol was established: challenge to reconstructive surgery. The present solu- the ultrasound frequency was set at 1 MHz and the tions for these skeletal disorders include artificial intensity of the power output for high transfection implants or allograft and autograft transplants, which efficiency at 5 W cmÀ2. The exposure duration we used often fail in the long term, are not always available, are was 10 min, and the duty cycle (ratio of run time to total costly and are associated with the risk of immune system cycle time) was set at 50%. The plasmid DNA solution rejection. Bone morphogenetic proteins (BMPs) are that we administered contained 50 mg DNA and 5% members of the transforming growth factor-b super- Optison. A time-dependent trend line of Luc expression family and are known for their ability to induce bone induced by sonoporation was constructed (Figure 1). formation; they frequently are the focus of therapeutic The expression profile indicated that sonoporated gene applications for bone repair and regeneration.26,27 It was translation has reached detectable level on day 4 post recently shown that the BMP-2, -6, -7 and -9 genes are the sonoporation. The peak expression was detected on day most potent inducers of osteogenic differentiation among 7(Luc expression doubled between days 4 and 7 post the BMP family.28–30 In this study we used the BMP-9 sonoporation), and then becomes undetectable by day 19. gene, since our previous observations indicated it could This trend line was found similar to previously reported lead to more significant osteogenesis than the human transient Luc expression using adenoviral vector.34 BMP-2 plasmid.28 Assuming that the expression profile of the sonoporated We hypothesized that, in optimized acoustic condi- human rhBMP-9 gene follows the same expression tions, the direct in vivo sonoporation of naked DNA profile as the Luc gene, transient expression of rhBMP-9 encoding an osteogenic gene recombinant human BMP-9 will induce ectopic bone formation in vivo in the (rhBMP-9) would induce bone formation. As previously sonoporated area.28,35 shown, viral delivery of the rhBMP-9 gene has thera- peutic potential in bone regeneration.31 However, viral gene delivery carries risks and nonviral gene delivery Electroporation of Luc reporter gene in vivo techniques are considered to have greater potential Using the bioluminescence imaging system, we noticed clinically because they are safer and more available. an expression pattern consistent with nonviral gene Significant advances have been made by sonoporation in delivery systems. The cooled charge-coupled device the field of gene delivery during the last decade.32 (CCCD) signal depicted in representative images in However, no report has been made of in vivo tissue Figure 2a were quantitatively analyzed and gene de novo formation using this delivery method. This is expression profile Luc expression peaked on day 6 the first report that demonstrates that the sonoporation- (which was the first time point checked), and was and electroporation-mediated delivery of plasmid DNA decreased in later time points (Figure 2). Compairing encoding the osteogenic gene rhBMP-9 induces bone these results to those achieved by sonoporation, we can tissue formation, in vivo. Our data demonstrate, though, conclude that the reporter Luc gene was about one log that sonoporation was significantly less efficient in bone more efficiently delivered by electroporation, than by formation than electroporation. means of sonoporation.

Gene Therapy Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 259

Figure 1 Reporter gene Luc expression profile in mice injected with Luc with or without sonoporation, as quantified by in vivo bioluminescence. (a) Representative CCCD images showing Luc expression in the left hindlimb, which received an injection of Luc followed by sonoporation but no expression in the right hindlimb, which received an injection of Luc but no sonoporation. (b) Quantitative analysis of the bioluminescent signal. The results are expressed in relative light units (RLU). Bars indicate s.e., n ¼ 5.

Figure 2 Reporter gene Luc expression profile in mice injected with Luc or with non-Luc plasmid (rhBMP-9) after electroporation, as quantified by in vivo bioluminescence. (a) Representative CCCD images showing Luc expression in the left hindlimb, which received an injection of Luc followed by electroporation but no expression in the right hindlimb, which received an injection of rhBMP-9. (b) Quantitative analysis of the bioluminescent signal. The results are expressed in relative light units (RLU). Bars indicate s.e., n ¼ 5.

Luc expression: immunohistochemical analysis revealed widespread expression of Luc in the sonopo- To demonstrate Luc expression in a sonoporated skeletal rated skeletal muscle on day 10 after sonoporation muscle, thigh tissue was embedded in paraffin and (Figure 3a). No staining was seen in control tissues that stained immunohistochemically to detect Luc expression. had received injections of the Luc gene without applica- The red stain, which indicates positive Luc expression, tion of sonoporation (Figure 3b).

Gene Therapy Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 260 In mice whose thigh muscles were injected with the osteogenic inducer gene rhBMP-9 followed by sonopora- tion, the expression of osteocalcin increased during the first 2 weeks and peaked on day 10, which is the expected profile in the bone formation process.36 Indeed, as depicted in Figures 4a and b, osteocalcin-triggered Luc expression increased and peaked on day 10 post sonoporation, after which it declined gradually and was no longer detectable by day 24 post sonoporation. The expression of osteocalcin in the mouse pelvis on day 17 (Figure 4a) is probably an artifact originated in the fact that it is transgenic mice that react in Luc expression to every bone formation or remodeling processes. In the contralateral thigh muscles, which were injected with empty pcDNA3 plasmid vector (the same plasmid backbone as rhBMP-9 plasmid used) and sonoporated using the same conditions serving as control group, low expression of Luc was observed. The osteocalcin-induced Luc expression in the contralateral thigh muscles was detected during the first days after sonoporation and then declined. We believe that the short osteocalcin expression was caused by minor tissue injury. Contrary to Luc expression in wild-type C3H mice, which originated in the sonoporated DNA, as described in the ‘Electroporation of Luc reporter gene in vivo’ section (Figure 1), in this instance osteocalcin expression origi- nated in the transgenic mice genome and was triggered by the sonoporated osteogenic gene. The values presented in Figure 3 were normalized to constitutive tail expression of Luc, as previously described,36 so that we could compare different mice that express the transgene with different intensities.

Ectopic bone formation induced by sonoporation and electroporation of rhBMP-9 Figure 3 In situ Luc expression induced by sonoporation of Luc The osteogenic process resulted in ectopic bone forma- in vivo and detected by immunohistochemical staining. (a) Thigh tion in the thigh muscles of six mice out of eight muscle of a mouse subjected to sonoporation after an injection of sonoporated mice and in five out five electroporated Luc DNA. (b) Thigh muscle of a mouse given an injection of Luc mice. This was validated by performing qualitative and DNA without application of sonoporation. Both tissue samples were quantitative micro-computerized tomography (micro- harvested 10 days after the initiation of sonoporation. CT) scanning (Figure 5a). The quantitative analysis of bone tissue demonstrated that the volume of newly formed bone after sonoporation reached 0.9±0.17 mm3 (mean±s.e., n ¼ 6; Figure 5c). The ectopic bone formation Osteogenic gene expression in vivo induced by after electroporation of the same amount of rhBMP-9 sonoporation resulted in bone volume of 3.12±0.8 mm3 (mean±s.e., We applied a direct gene delivery paradigm to deliver n ¼ 5; Figure 5c). The bone volume density of ectopic the rhBMP-9 gene to an ectopic thigh muscle site via bone formation generated using sonoporation and sonoporation. Our experimental model included trans- electroporation was calculated by dividing total tissue genic mice (Oc-Luc), which harbor the Luc gene under volume to bone volume (BV) and was found to be the control of the osteogenic tissue-specific promoter, 0.7±0.09 (mean±s.e., n ¼ 6) and 0.27±0.02 (mean±s.e., osteocalcin, as described elsewhere,36 and use of a n ¼ 5), respectively (Figure 5d). The bone mineral density quantitative, noninvasive CCCD camera, which is used was found to be 793.51±68.44 mg HA per cm3 (HA, as a real-time monitoring system. In this model, hydroxyapatite; mean±s.e., n ¼ 6) in sonoporated mus- osteogenic activity in the transgenic mice is reflected in cles and 604.47±14.28 mg HA per cm3 (mean±s.e., Luc expression, which is activated by the osteocalcin n ¼ 6) in electroporated muscles (Figure 5d). Entire promoter and is detected by the CCCD as an emitted sonoporated hindlimbs were dissected and subjected to light following an injection of luciferin.36 In these a histological analysis. Standard hematoxylin and eosin experiments, we used the contralateral thigh muscles of (H&E) and bone matrix-specific (Masson trichrome, treated mice as control tissue; muscles of the contral- MTC) staining of these tissues are shown in Figure 6. ateral thigh were injected with empty plasmid pcDNA3 Morphology typical of bone marrow was found inside vector in a phosphate-buffered saline (PBS) and micro- the bone formation, resembling to other approaches bubbles solution, and the same sonoporation protocol inducing ectopic bone formation.37,38 Micro-CT images was applied. and histological slides clearly demonstrated osteogenic

Gene Therapy Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 261

Figure 4 Bioluminescence induced by osteocalcin expression in transgenic Oc-luc mice as a result of sonoporation of either plasmid recombinant human bone morphogenetic protein-9 (rhBMP-9) or the control empty plasmid vector pcDNA3. (a) Representative CCCD images of mice injected with the empty vector in the left thigh and rhBMP-9 in the right thigh, and subjected to sonoporation. (b) Quantitative analysis of the bioluminescent signal. Bars indicate s.e., n ¼ 5.

Figure 5 Ectopic bone formation induced by sonoporation and electroporation of rhBMP-9 in vivo. Ectopic bone formation was detected and quantified by high-resolution micro-CT. (a and b, respectively). Three-dimensional image of representative ectopic bone formation in vivo induced by sonoporation of plasmid DNA encoding rhBMP-9. Enlarged two-dimensional image of representative ectopic bone formation is depicted on the left side of panel a. (c–e) Quantitative analysis of ectopic bone formation: bone volume (BV) (c), bone volume density (BVD) (d), and bone mineral density (BMD) (e). Bars indicate s.e., n ¼ 9. tissue formation in mice subjected to sonoporation with (injected with empty plasmid vector, pcDNA3 and rhBMP-9 (Figures 5 and 6). There was no indication of sonoporated or electroporated under the same condi- osteogenic tissue formation in the control hindlimbs tions used in the experimental group) in the limbs of all

Gene Therapy Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 262

M BM

BM

BF BF

X10 X40

BF BM BM BM

BF M M X10 X40

Figure 6 Ectopic bone formation induced by sonoporation of rhBMP-9 in vivo. Histological analysis was carried out 6 weeks after sonoporation. (a and b) H & E staining. (c and d) Masson trichrome staining. BF, bone formation; BM, bone marrow; M, skeletal muscle; yellow arrow indicates active osteocytes.

mice, as was already indicated by the bioluminescence orthopedic gene therapy. However, it is important to imaging. As a reference to the gene delivery method we note that sonoporation-mediated bone formation was used electroporation of the same plasmid DNA enco- less efficient than electroporation-driven one. Further ding to rhBMP-9, using the same DNA amounts. The research and development of the gene delivery system is electroporation in vivo is a very well-known method and needed in order to significantly increase the efficiency of highly efficient. The ectopic bone mass formed in the the ultrasound-based transfection. Direct gene delivery electroporated thigh muscle is evident in Figure 5b and involving a safe and noninvasive injectable technique its volume is significantly higher than the bone that was has enormous clinical significance because it enables generated by means of sonoporation (Figure 5c). How- incremental and controlled tissue formation using ever, bone volume density of the newly formed tissue repeated treatments. generated by sonoporation was found significantly This study is the first to use sonoporation in vivo for higher than by the electroporation method. The bone skeletal tissue formation de novo in combination with mineral density indicates the maturation state of the noninvasive quantitative imaging of the osteogenic bone tissue. Both types of bone tissue originated in process. The sonoporation technique has previously been nonviral gene delivery of rhBMP-9 gene and formed applied to skeletal muscle, mostly to deliver reporter during the same period (5 weeks), therefore the bone genes.14,17,22 Comparing the proposed gene delivery mineral density was found not significantly different technique with other direct gene delivery systems, one (Figure 5e). can conclude that chemical gene delivery agents possess rather low efficiency in vivo, and the most efficient physical gene delivery system described to date is 39 Discussion electroporation. Electroporation is very well known and widely studied gene delivery method. Our com- Nonviral gene delivery systems offer a safe and versatile parison between sonoporation and electroporation alternative to their viral counterparts. However, a major methods showed that electroporation is significantly limiting factor for efficient nonviral gene therapy more efficient and potent method in bone tissue remains the lack of a suitable and efficient vector for formation de novo. Moreover, the sonoporation technique gene delivery. In this study we propose a direct, should be further developed and studied in order to noninvasive, nonviral, safe and efficient delivery system, reach the efficiency of electroporation. However, applica- which is based on therapeutic ultrasound. As we tion of electroporation is clinically inferior because tissue hypothesized, naked plasmid DNA encoding for can be damaged.40,41 Sonoporation is considered as a safe rhBMP-9 that is delivered by sonoporation into the technique. It can lead to overheating of tissue, which skeletal muscle induced the osteogenic process and may in turn cause tissue damage, but the overheating resulted in bone formation in vivo. These results open a can be avoided without decreasing the efficiency of the new avenue in osteogenic tissue engineering and transfection, however, by lowering the ultrasound duty

Gene Therapy Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 263 cycle to 50% or lower and extending the exposure utilizing this technique in tissue formation de novo duration. A low duty cycle previously was shown to approach. Compared with the stem cell-mediated increase cell survival after sonoporation.33 approach, the sonoporation-based system offers direct An essential tool in tissue engineering in vivo lies in nonviral gene delivery, which is easier to administer, imaging systems, which have been used to monitor gene does not require tissue culture, saves time and money. expression and tissue formation. In this study, marker This proposed sonoporation-based bone-tissue engi- gene expression was monitored noninvasively by using neering approach should be further developed to an in vivo bioluminescence system. Although Luc gene increase gene delivery efficiency, and would be applied delivery was carried out by sonoporation and the to bone fracture repair models and other tissue regenera- expected peak expression was in the first days, and tion models. Although the candidate cell population in because the nonviral delivery occurs during the sono- nonunion bone defect site differs from skeletal myocytes poration and the likelihood uptake of the plasmid DNA that were targeted in this study, the direct gene transfer later on is rather low, our results showed that the peak into bone defect has been shown to be successful expression was detected on day 7. This was probably due before.43,44 to the protein accumulation effect in the nondividing sonoporated myocytes that occurred in the first days till the Luc reached the detectable level on day 3 and the Materials and methods peak expression on day 7. One of the most notable advantages of the present study over previous investiga- Plasmid DNA constructs and microbubbles tions of sonoporation is the quantitative in vivo analysis Two constructs of plasmid DNA were used: plasmid Luc of gene expression and protein translation in a long- under the control of the cytomegalovirus (CMV) promo- itudinal profile. This feature enables us to study bone ter and plasmid recombinant human BMP-9 (rhBMP-9) formation and gene expression in the same animal over based on the pcDNA3 vector and driven by CMV time, generating a more reliable insight into those promoter as well. Both constructs were amplified using processes. standard procedures and purified using an EndoFree Kit The exact mechanism of gene delivery by sonopora- (Qiagen, Valancia, CA, USA). After purification, both tion is not clearly understood. Recent reports have constructs were eluted in a PBS solution for in vivo revealed that a type of ultrasonography contrast agent, injection. The plasmid DNA was prepared in aliquots known as a microbubble, generates cavitation when containing 50-mg DNA in 100 ml PBS solution with 5% exposed to ultrasound. This cavitation results in the Optison, unless otherwise indicated. For mock transfec- bursting of microbubbles, leading to distribution of tions we used aliquots of 50-mg pcDNA3 empty vector in material over a specific area of interest. Sonoporation PBS solution with 5% Optison. and cavitation are thought to produce synergistic effects, yielding increased transfection efficiency.7,17 Animal models In comparison with other tissue-engineering ap- Female wild-type C3H/HeN mice, each between 6 and 9 proaches such as cell-based therapy, direct gene delivery months old, were used in the experiments for optimiza- is less complicated, less expensive and less time tion of the sonoporation and electroporation protocols, consuming. We believe that therapeutic agents (nucleic which was carried out using the reporter gene Luc. For acids and contrast agents) have the potential to be the osteogenesis studies, in which we delivered rhBMP-9 developed into products easily stored on the shelf. We in vivo, we used FVB/N transgenic mice, which had been also believe that the procedure can be repeated because it generated to harbor the Luc gene under the control of is noninvasive, no tissue damage is observed in the the human osteocalcin promoter (hOc), as previously histological analysis and no immune response to naked described.45 nucleic acids is expected when compared with methods involving the viral vectors currently in use in direct Sonoporation technique tissue engineering. Direct gene delivery usually gener- The mice were anesthetized with a mixture of xylazine ates less bone than the cell-based therapy that was also and ketamine (0.15% xylazine and 0.85% ketamine), developed in our laboratory,28 however, its advantage which was injected intraperitoneally (i.p.) at 1 mlgÀ1 lies in the simplicity and cost effectiveness of the body weight. Before the DNA solution was injected, the therapeutic approach. We do not have the tools to animals’ back limbs were prepared by shaving the skin indicate what cell types expressed the delivered DNA or hair and swabbing it with 70% isopropanol and 0.5% what cells formed the bone tissue. Nevertheless, we can chlorhexidine. The solutions of plasmid DNA with speculate that the gene is delivered to myocytes, muscle microbubbles in PBS were injected into the thigh muscle progenitors and satellite cells, and is expressed by them. close to the femur. The needle penetrated a point in the We believe that the rhBMP-9 protein recruits both bone thigh muscle and was pushed farther in until the femur marrow-derived osteoprogenitor cells circulating in the was reached. Once the needle reached its destination, we blood stream and skeletal muscle-derived osteoprogeni- released the plasmid–Optison mix slowly into the thigh tor cells within the muscle, which have been shown to muscle. Right after the injection, the needle was removed respond to secreted BMP-2 in a similar model to produce and the skin was covered with Aloe-sound gel (Rich-Mar intramuscular bone.42 Corp., Inola, OK, USA) at designated sonoporation sites. Finally, we have shown that rhBMP-9 gene delivery Immediately thereafter, transcutaneous, noninvasive using sonoporation is sufficient to elicit bone formation sonoporation was applied using the 6-mm probe of a in an ectopic site. However, its efficiency is still Sonitron2000 device kindly provided by Rich-Mar Corp. significantly inferior comparing to the electroporation- at a power output intensity of 5 W cmÀ2 and various mediated gene delivery. This is the first pilot study acoustic parameters. According to the manufacturer, the

Gene Therapy Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 264 beam of the ultrasound evenly distributes the energy to in 1:100 in PBS. Slides were counterstained with H&E, at least 4 mm of tissue depth covering most of the mouse washed and mounted. thigh muscle. In the experiment inducing ectopic bone formation, the mice received infiltration injections into Evaluation of bone formation in vivo the thigh muscle followed by sonoporation, three times, For a detailed qualitative and quantitative three-dimen- separated by a day of rest, on days 0, 2 and 4. sional evaluation of the bone that formed, we used a scanning protocol that yielded a high resolution.48 The Electroporation technique dissected specimen was mounted on a turntable and 500 The mice were anesthetized as described above. Desig- projections were obtained. The long axis of the femur nated electroporation sites were prepared in the same was placed perpendicular to the X-ray beam axis. The manner as for the sonoporation technique, and 50 mg X-ray tube was operated at 60 KeV and 150 mA, and the of plasmid DNA encoding for Luc or rhBMP-9 in 100 ml integration time was 200 ms. Scans were obtained of 1 Â PBS was injected into the left or right thigh muscle at a resolution of 20 mm in all three spatial dimensions. of the mice, respectively. Immediately post injection, the Two-dimensional CT images were reconstructed in skin was covered with 40% glycerol in 1 Â PBS and 1024 Â 1024 pixel matrices by using a standard convolu- transcutaneous electric pulses were applied using two tion back-projection procedure with a Shepp and Logan 1-cm cube electrodes placed 2–4 mm apart at both sides filter. The bone tissue was segmented from bone marrow of the limb. To generate pulses an ECM 830 electro- and soft tissue by using a global thresholding proce- porator (BTX, San Diego, CA, USA) was used. We dure.49 In addition to the visual assessment of structural applied 4 Â 4 pulses of 100 V for 20 ms per pulse, 1 Hz images, morphometric indices were determined on the frequency, rotating the electrodes around the limb basis of microtomographic data sets by performing direct between the sets of pulses. three-dimensional morphometry.50,51

Noninvasive bioluminescence imaging Histological analysis The imaging unit consisted of an intensified Peltier Tissue samples from an intramuscular site were har- CCCD, Model LN/CCD-1300EB, which was equipped vested 5 weeks after the application of sonoporation, with an ST-133 controller and a 50-mm Nikon lens fixed in 4% formalin overnight, decalcified using Calci- (Roper Scientific, Princeton Instruments, Trenton, NJ, Clear Rapid (National Diagnostics, East Riding of USA). In this system, a pseudocolor image represents Yorkshire, England) for 48 h, passed through increasing light intensity (with blue representing least intense and concentrations of ethanol and embedded in paraffin. red most intense). In all cases, the integrated light is the Five-micrometer-thick sections were cut from each result of a 2-min exposure and acquisition. The exposure paraffin block using a motorized microtome (Leica conditions (including time, f/stop, position of the stage, Microsystems, Nussloch, Germany). The slides were binding ratio and time after the injection of luciferin) heated at 65 1C for 45 min and was followed by were maintained at identical levels so that all measure- deparaffinization. H&E and MTC staining procedures ments would be comparable.36 were carried out in a manner previously reported.38 In all experiments, the mice were anesthetized with a mixture of xylazine and ketamine (0.15 ml xylazine and À 0.85 ml ketamine), which was injected i.p. at 1 mlg 1 Acknowledgements body weight. Ten minutes before the light emissions were monitored, each mouse was given an i.p. injection We thank Rich-Mar Corp. (Inola) for kindly providing of beetle luciferin (Promega Corp., Madison, WI, USA) in the sonoporation Sonitron2000 device for the experi- PBS at 126 mg kgÀ1 body weight. The mice were placed ments described in this study. in a light-tight chamber (a dark box) and a gray-scale image was first recorded using dimmed light. Photon emission was then integrated over a period of 2 min and References recorded as pseudocolor images.34,36,46 On the basis of our published data, an internal control (at the tail level) 1 Parker SE, Vahlsing HL, Serfilippi LM, Franklin CL, Doh SG, was needed for quantification of the results due to Gromkowski SH et al. Cancer gene therapy using plasmid DNA: differences in basal expression levels of Luc.36 safety evaluation in rodents and non-human primates. Hum Gene Ther 1995; 6: 575–590. Immunohistochemical analysis of Luc expression 2 Burke B, Sumner S, Maitland N, Lewis CE. Macrophages in gene The immunohistochemical analysis was carried out as therapy: cellular delivery vehicles and in vivo targets. J Leukoc 72 previously described.47 Briefly, thigh muscles sonopo- Biol 2002; : 417–428. 3 Hamm A, Krott N, Breibach I, Blindt R, Bosserhoff AK. Efficient rated with the Luc gene were harvested 10 days after transfection method for primary cells. Tissue Eng 2002; 8: sonoporation. Treated and control organs were fixed in 235–245. 4% paraformaldehyde for 3 days. Then the samples were 4 Faurie C, Phez E, Golzio M, Vossen C, Lesbordes JC, Delteil C decalcified in EDTA (0.5 M, pH 8) for 14 days, embedded et al. Effect of electric field vectoriality on electrically mediated in paraffin, and cut into 5-mm-thick sections. Immuno- gene delivery in mammalian cells. Biochim Biophys Acta 2004; histochemical staining was carried out using a Histostein 1665: 92–100. SP kit (Zymed Laboratories, San Francisco, CA, USA). 5 Cemazar M, Wilson I, Dachs GU, Tozer GM, Sersa G. Direct Primary Luc antibody (polyclonal rabbit anti-mouse) and visualization of electroporation-assisted in vivo gene delivery to a secondary goat anti-rabbit immunoglobulin G biotin- tumors using intravital microscopy—spatial and time depen- conjugated antibody (Zymed Laboratories) were diluted dent distribution. BMC Cancer 2004; 4: 81.

Gene Therapy Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 265 6 Ward M, Wu J, Chiu JF. Experimental study of the effects of 25 Golzio M, Teissie J, Rols MP. Cell synchronization effect on Optison concentration on sonoporation in vitro. Ultrasound Med mammalian cell permeabilization and gene delivery by electric Biol 2000; 26: 1169–1175. field. Biochim Biophys Acta 2002; 1563: 23–28. 7 Lawrie A, Brisken AF, Francis SE, Cumberland DC, Crossman 26 Reddi AH. Role of morphogenetic proteins in skeletal DC, Newman CM. Microbubble-enhanced ultrasound for tissue engineering and regeneration. Nat Biotechnol 1998; 16: vascular gene delivery. Gene Therapy 2000; 7: 2023–2027. 247–252. 8 Duvshani-Eshet M, Adam D, Machluf M. The effects of albumin- 27 Einhorn TA. Clinical applications of recombinant human BMPs: coated microbubbles in DNA delivery mediated by therapeutic early experience and future development. J Bone Joint Surg Am ultrasound. J Control Release 2006; 112: 156–166. 2003; 85-A (Suppl 3): 82–88. 9 Duvshani-Eshet M, Machluf M. Efficient transfection of tumors 28 Aslan H, Zilberman Y, Arbeli V, Sheyn D, Matan Y, Liebergall M facilitated by long-term therapeutic ultrasound in combination et al. Nucleofection-based ex vivo nonviral gene delivery to with contrast agent: from in vitro to in vivo setting. Cancer Gene human stem cells as a platform for tissue regeneration. Tissue Ther 2007; 14: 306–315. Eng, accepted for publication (MS: TEN-05-0208.R1) 2006; 12: 10 Amabile PG, Waugh JM, Lewis TN, Elkins CJ, Janas W, Dake 877–889. MD. High-efficiency endovascular gene delivery via therapeutic 29 Cheng H, Jiang W, Phillips FM, Haydon RC, Peng Y, Zhou L et al. ultrasound. J Am Coll Cardiol 2001; 37: 1975–1980. Osteogenic activity of the fourteen types of human bone 11 Chen S, Shohet RV, Bekeredjian R, Frenkel P, Grayburn PA. morphogenetic proteins (BMPs). J Bone Joint Surg Am 2003; 85- Optimization of ultrasound parameters for cardiac gene delivery A: 1544–1552. of adenoviral or plasmid deoxyribonucleic acid by ultrasound- 30 Brown MA, Zhao Q, Baker KA, Naik C, Chen C, Pukac L targeted microbubble destruction. J Am Coll Cardiol 2003; 42: et al. Crystal structure of BMP-9 and functional interactions 301–308. with pro-region and receptors. J Biol Chem 2005; 280: 12 Hashiya N, Aoki M, Tachibana K, Taniyama Y, Yamasaki K, 25111–25118. Hiraoka K et al. Local delivery of E2F decoy oligodeoxynucleo- 31 Li JZ, Li H, Dunford B, Holman D, Beres B, Pittman DD et al. Rat tides using ultrasound with microbubble agent (Optison) strain differences in the ectopic osteogenic potential of recombi- inhibits intimal hyperplasia after balloon injury in rat nant human BMP adenoviruses. Mol Ther 2003; 8: 822–829. carotid artery model. Biochem Biophys Res Commun 2004; 317: 32 Bekeredjian R, Grayburn PA, Shohet RV. Use of ultrasound 508–514. contrast agents for gene or drug delivery in cardiovascular 13 Hauff P, Seemann S, Reszka R, Schultze-Mosgau M, Reinhardt medicine. J Am Coll Cardiol 2005; 45: 329–335. M, Buzasi T et al. Evaluation of gas-filled microparticles and 33 Pan H, Zhou Y, Izadnegahdar O, Cui J, Deng CX. Study of sonoporation as gene delivery system: feasibility study in rodent sonoporation dynamics affected by ultrasound duty cycle. tumor models. Radiology 2005; 236: 572–578. Ultrasound Med Biol 2005; 31: 849–856. 14 Lu QL, Liang HD, Partridge T, Blomley MJ. Microbubble 34 Honigman A, Zeira E, Ohana P, Abramovitz R, Tavor E, Bar I ultrasound improves the efficiency of gene transduction in et al. Imaging transgene expression in live animals. Mol Ther skeletal muscle in vivo with reduced tissue damage. Gene Therapy 2001; 4: 239–249. 2003; 10: 396–405. 35 Dayoub H, Dumont RJ, Li JZ, Dumont AS, Hankins GR, Kallmes 15 Miao CH, Brayman AA, Loeb KR, Ye P, Zhou L, Mourad P et al. DF et al. Human mesenchymal stem cells transduced with Ultrasound enhances gene delivery of human factor IX plasmid. recombinant bone morphogenetic protein-9 adenovirus promote Hum Gene Ther 2005; 16: 893–905. osteogenesis in rodents. Tissue Eng 2003; 9: 347–356. 16 Nakashima M, Tachibana K, Iohara K, Ito M, Ishikawa M, 36 Iris B, Zilberman Y, Zeira E, Galun E, Honigman A, Turgeman G Akamine A. Induction of reparative dentin formation by et al. Molecular imaging of the skeleton: quantitative real-time ultrasound-mediated gene delivery of growth/differentiation bioluminescence monitoring gene expression in bone repair and factor 11. Hum Gene Ther 2003; 14: 591–597. development. J Bone Miner Res 2003; 18: 570–578. 17 Taniyama Y, Tachibana K, Hiraoka K, Namba T, Yamasaki K, 37 Moutsatsos IK, Turgeman G, Zhou S, Kurkalli BG, Pelled G, Tzur Hashiya N et al. Local delivery of plasmid DNA into rat carotid L et al. Exogenously regulated stem cell-mediated gene therapy artery using ultrasound. Circulation 2002; 105: 1233–1239. for bone regeneration. Mol Ther 2001; 3: 449–461. 18 Tsutsui JM, Xie F, Porter RT. The use of microbubbles to target 38 Turgeman G, Pittman DD, Muller R, Kurkalli BG, Zhou S, Pelled drug delivery. Cardiovasc Ultrasound 2004; 2: 23. G et al. Engineered human mesenchymal stem cells: a novel 19 Miller DL, Quddus J. Sonoporation of monolayer cells by platform for skeletal cell mediated gene therapy. J Gene Med diagnostic ultrasound activation of contrast-agent gas bodies. 2001; 3: 240–251. Ultrasound Med Biol 2000; 26: 661–667. 39 Andre F, Mir LM. DNA electrotransfer: its principles and an 20 Dijkmans PA, Juffermans LJ, Musters RJ, van Wamel A, ten Cate updated review of its therapeutic applications. Gene Therapy FJ, van Gilst W et al. Microbubbles and ultrasound: from 2004; 11 (Suppl 1): S33–S42. diagnosis to therapy. Eur J Echocardiogr 2004; 5: 245–256. 40 Lefesvre P, Attema J, van Bekkum D. A comparison of efficacy 21 Pislaru SV, Pislaru C, Kinnick RR, Singh R, Gulati R, Greenleaf JF and toxicity between electroporation and adenoviral gene et al. Optimization of ultrasound-mediated gene transfer: transfer. BMC Mol Biol 2002; 3: 12. comparison of contrast agents and ultrasound modalities. Eur 41 Kishimoto KN, Watanabe Y, Nakamura H, Kokubun S. Ectopic Heart J 2003; 24: 1690–1698. bone formation by electroporatic transfer of bone morphogenetic 22 Liang HD, Lu QL, Xue SA, Halliwell M, Kodama T, Cosgrove protein-4 gene. Bone 2002; 31: 340–347. DO et al. Optimisation of ultrasound-mediated gene transfer 42 Musgrave DS, Bosch P, Ghivizzani S, Robbins PD, Evans CH, (sonoporation) in skeletal muscle cells. Ultrasound Med Biol 2004; Huard J. Adenovirus-mediated direct gene therapy with bone 30: 1523–1529. morphogenetic protein-2 produces bone. Bone 1999; 24: 541–547. 23 Kodama T, Tan PH, Offiah I, Partridge T, Cook T, George AJ et al. 43 Betz OB, Betz VM, Nazarian A, Pilapil CG, Vrahas MS, Bouxsein Delivery of oligodeoxynucleotides into human saphenous veins ML et al. Direct percutaneous gene delivery to enhance healing and the adjunct effect of ultrasound and microbubbles. Ultra- of segmental bone defects. J Bone Joint Surg Am 2006; 88: 355–365. sound Med Biol 2005; 31: 1683–1691. 44 Chakkalakal DA, Strates BS, Mashoof AA, Garvin KL, Novak JR, 24 Gehl J. Electroporation: theory and methods, perspectives for Fritz ED et al. Repair of segmental bone defects in the rat: an drug delivery, gene therapy and research. Acta Physiol Scand experimental model of human fracture healing. Bone 1999; 25: 2003; 177: 437–447. 321–332.

Gene Therapy Sonoporation of BMP-9 induces bone formation in vivo D Sheyn et al 266 45 Clemens TL, Tang H, Maeda S, Kesterson RA, Demayo F, 48 Ruegsegger P, Koller B, Muller R. A microtomographic system Pike JW et al. Analysis of osteocalcin expression in transgenic for the nondestructive evaluation of bone architecture. Calcif mice reveals a species difference in vitamin D regulation of Tissue Int 1996; 58: 24–29. mouse and human osteocalcin genes. J Bone Miner Res 1997; 12: 49 Muller R, Ruegsegger P. Micro-tomographic imaging for the 1570–1576. nondestructive evaluation of trabecular bone architecture. Stud 46 Contag CH, Spilman SD, Contag PR, Oshiro M, Eames B, Health Technol Inform 1997; 40: 61–79. Dennery P et al. Visualizing gene expression in living mammals 50 Balto K, Muller R, Carrington DC, Dobeck J, Stashenko P. using a bioluminescent reporter. Photochem Photobiol 1997; 66: Quantification of periapical bone destruction in mice by micro- 523–531. computed tomography. J Dent Res 2000; 79: 35–40. 47 Lavon I, Goldberg I, Amit S, Landsman L, Jung S, Tsuberi BZ et 51 Muller R, Van Campenhout H, Van Damme B, Van Der Perre G, al. High susceptibility to bacterial infection, but no liver Dequeker J, Hildebrand T et al. Morphometric analysis of human dysfunction, in mice compromised for hepatocyte NF-kappaB bone biopsies: a quantitative structural comparison of histological activation. Nat Med 2000; 6: 573–577. sections and micro-computed tomography. Bone 1998; 23: 59–66.

Gene Therapy