JOURNAL OF CELLULAR PHYSIOLOGY 209:413–421 (2006)
Ultrasound Guided Site Specific Gene Delivery System Using Adenoviral Vectors and Commercial Ultrasound Contrast Agents CANDACE M. HOWARD,1 FLEMMING FORSBERG,1 CORRADO MINIMO,2 JI-BIN LIU,1 DANIEL A. MERTON,1 AND PIER PAOLO CLAUDIO3* 1 Department of Radiology, Thomas Jefferson University, Philadelphia, Pennsylvania 2 Department of Pathology, Main Line Clinical Laboratories, Wynnewood, Pennsylvania 3 Sbarro Institute for Cancer Research and Molecular Medicine, College of Science and Technology, Center for Biotechnology, Temple University, Philadelphia, Pennsylvania We have evaluated if ultrasound imaging (US) and various commercially available contrast microbubbles can serve as a non-invasive systemically administered delivery vehicle for site-specific adenoviral-mediated gene transfer in vitro and in vivo. The contrast agents were tested for their ability to enclose and to protect an adenoviral vector carrying the GFP marker gene (Ad-GFP) into the microbubbles. We have also evaluated the ability of the innate immune system to inactivate free adenoviruses as well as unenclosed viruses adsorbed on the surface of the contrast agents and in turn the ability of the microbubbles to enclose and to protect the viral vectors from such agents. In vitro as well as in vivo, innate components of the immune system were able to serve as inactivating agents to clear free viral particles and unenclosed adenoviruses adsorbed on the microbubbles’ surface. Systemic delivery of Ad-GFP enclosed into microbubbles in the tail vein of nude mice resulted in specific targeting of the GFP transgene. Both fluorescence microscopy and GFP immunohistochemistry demonstrated US guided specific transduction in the targeted cells only, with no uptake in either heart, lungs or liver using complement-pretreated Ad-GFP microbubbles. This approach enhances target specificity of US microbubble destruction as a delivery vehicle for viral-mediated gene transfer. J. Cell. Physiol. 209: 413–421, 2006. ß 2006 Wiley-Liss, Inc.
Progress in gene therapy has been hindered by concerns over the safety and practicality of viral vectors and the inefficiency of currently available non-viral transfection techniques (Bekeredjian and Shohet, 2004). One of the major challenges for gene therapy is systemic delivery of nucleic acids directly into an affected tissue. This requires developing a vehicle that is able to protect the nucleic acid from degradation, while delivering the gene of interest to the specific tissue and specific subcellular compartment. Viruses are attractive delivery vectors because of their ability to efficiently transfer genes with sustained expression. However, this very same attribute also serves as one of the greatest drawbacks due to the potential for insertional mutagenesis. Residual viral elements can be immunogenic, cytopathic, or recombinogenic. Recombinant adenoviruses are one of the most common gene transfer vectors utilized in human clinical trials, but systemic administration of this virus will also be met by host innate and adaptive antiviral immune responses which can limit and/or preclude repetitive regiments (Jiang et al., 2004). The complement system, a group of proteins that has evolved to rapidly recognize foreign microbes and viruses and to clear them from the circulatory system, is one such innate host defense. The quest for novel, safe and efficient systemic gene delivery systems has raised ultrasound (US) contrast agents as one of the putative candidates (Unger et al., 1997, 2001; Voss and Kruskal, 1998; Lawrie et al., 1999, 2000; Miller et al., 1999; Newman et al., 2001; Ng and Liu, 2002; Bekeredjian et al., 2003; Hosseinkhani et al., 2003; Dijkmans et al., 2004; Lavon and Kost, 2004; Shimamura et al., 2004; Wei et al., 2004; Bekeredjian ß 2006 WILEY-LISS, INC.
et al., 2005; Larina et al., 2005). Ultrasound has been used diagnostically and therapeutically for decades, and its safety is well established (Jakobsen et al., 2005). Even high-energy US is used clinically in hyperthermia, however, transdermal drug delivery is accomplished with ultrasound, at low frequencies because of more efficient cavitation (Tsutsui et al., 2004). Second generation US contrast agents (microbubbles) contain high-molecular weight gasses with less solubility and diffusivity, which improves microbubble persistence and allows transpulmonary passage. This allows peripheral vein injection, because the more robust bubbles can recirculate through the systemic circulation numerous times, surviving for several minutes within the bloodstream (Goldberg et al., 1994). The ideal microbubble diameter most likely is 4 mm. This is small enough to prevent entrapment within the pulmonary capillary bed (ranging from 5 to 8 mm in diameter), but big enough to entrap and protect viral vectors such as adenoviruses (Ad) from the environment. Microbubble
Contract grant sponsor: W. W. Smith Charitable Trust; Contract grant sponsor: Toshiba America Medical Systems, Tustin, CA. *Correspondence to: Pier Paolo Claudio, College of Science and Technology, Center for Biotechnology, Bio Life Sciences Building, Suite 333, 1900 North 12th Street, Philadelphia, PA 19122-6099. E-mail:
[email protected] Received 24 May 2006; Accepted 12 June 2006 DOI: 10.1002/jcp.20736
414
HOWARD ET AL.
shells range from 10 to 200 nm in thickness and help to prevent destruction and diffusion of the gas core. This, in turn, improves the echogenicity and longevity of the microbubble. Shells are composed of phospholipids, proteins, biocompatible polymers, or surfactants. The nature of the shell determines the flexibility of the microbubble, the effect of US, and the binding properties to specific cell types (Porter et al., 1997; Dayton et al., 1999, 2001; Lindner et al., 2000). The gas-filled microspheres effectively lower the energy threshold for nonthermal cavitation. This allows diagnostic transducers operating within the energy levels mandated by the FDA to be used for drug/gene delivery. In the sonification zone the microbubbles are destroyed releasing their contents. Cavitation also creates small shock waves that increase cell permeability by disruption of the endothelial barrier (Pitt et al., 2004). Gene-based drugs are highly active and therefore are applicable to microbubble vectors where the amount of gene injected is on the order of micrograms; therefore, a large volume of bubbles is not required for delivery. US contrast agents offer the advantage of target-specific release by imaging guidance, thereby improving gene delivery specificity (Dijkmans et al., 2004; Howard, 2004; Pitt et al., 2004). The endothelial barrier precludes hematogenous delivery of viral agents such as adenoviruses. However, ultrasound-targeted microbubble destruction (UTMD) circumvents this considerable obstacle by taking advantage of the physical properties of microbubbles to enable focal release of entrapped materials as well as the creation of small shock waves that increase cellular permeability. In addition, the microbubbles protect the viruses from rapid degradation by the immune system, thus allowing for intravenous injection rather than direct target organ delivery by catheter-based approaches or operative bed injection (Howard, 2004). This is particularly important in cardiovascular as well as cancer gene therapy of inaccessible tumors. The microbubbles may also limit the amount of inflammatory response to the viruses and may allow repeated injections. Thus, the purpose of our study was to evaluate the feasibility of site-specific gene delivery mediated by diagnostic US using Ad-GFP encapsulated in commercially available US contrast agents in vitro and in vivo. The goal of the study was also to determine if incubation of the microbubbles with complement could improve specificity of viral transgene transduction to the target tissue/organ allowing a simplified approach to encapsulation of the viral vectors with commercially available contrast agents. MATERIALS AND METHODS Cell lines, cell culture and adenoviral production
The DU-145 (human prostate adenocarcinoma), H23 (human lung adenocarcinoma), DB-1 (human melanoma), and 293 (primary human embryonic kidney) cell lines were obtained from the American Type Culture Collection (ATCC, Rockville, MD) and were grown at 378C, in a 5% CO2/95% atmosphere, in Dulbecco’s modified Eagle’s medium (Mediatech Inc., Herndon, VA) supplemented with 10% fetal bovine serum (FBS) from Hyclone, Inc., (Logan, UT). Ad-GFP, which expresses the green fluorescence protein marker gene under the strong cytomegalovirus (CMV) constitutive promoter, was generated using the AdEasy system (Carlsbad, CA), amplified as previously described (Claudio et al., 1999), and purified with the BD Adeno-X virus purification kit (BD Biosciences, Mountain View, CA) following the manufacturer directions. A viral titer of 1.2 1012 plaque-forming units (pfu)/ml was determined with a plaque assay for the Ad-GFP viruses. Journal of Cellular Physiology DOI 10.1002/jcp
Preparation of microbubbles and inactivation of unenclosed adenoviruses
Optison (GE Healthcare, Princeton, NJ), SonoVue (Bracco, Princeton, NJ), Levovist (Schering AG, Berlin, Germany), Sonazoid (GE Healthcare, Oslo, Norway) and Imagent (IMCOR Pharmaceuticals, Inc., San Diego, CA) contrast agents were prepared following manufacturers’ instructions. Microbbubbles were reconstituted in the presence or absence of 2 ml of 1.2 1012 pfu of Ad-GFP viruses. To inactivate unenclosed and free adenoviruses, 1 volume of microbubbles formed in the presence of Ad-GFP were incubated with 10 volumes of a solution containing 60 mg/ml of human complement (Sigma Aldrich, Saint Louis, MS) for 30 min at 378C. Microbubbles were then washed with 10 ml of phosphate buffer saline solution (PBS). The milky white suspension floating on the top of PBS was then collected and used in the in vitro and in vivo experiments. For the in vitro experiments US exposure was achieved with a 2.5 MHz phased array and a PowerVision 7000 scanner (Toshiba America Medical Systems, Tustin, CA). The acoustic output of the transducer was measured in water at 1 cm depth using a 0.5 mm broadband acoustic hydrophone (Precision Acoustics Ltd., Dorchester, UK). This hydrophone has an excellent sensitivity over 1–20 MHz. Equal number of cells were plated in six-well dishes in triplicates and insonified at 535 or 207 kPa peak negative pressure (corresponding to the 0 and-5 dB output settings, respectively) for 1 min after administration of 100 mL of bubbles reconstituted with the viral vector. Experiments were repeated with the delivery vehicle incubated with complement for 30 min at 378C to inactivate unenclosed Ad-GFP and with controls (contrast bubbles only, in one well per agent). After 24 h transduction efficiency was demonstrated by fluorescent microscopy. Localization of viruses within the microbubble preparation
Imagent (IMCOR Pharmaceuticals, Inc., San Diego, CA) contrast agents were prepared following manufacturers instructions. Microbbubbles were reconstituted in the presence or absence of 2 ml of 1.2 1012 pfu of Ad-GFP viruses. Unenclosed adenoviruses were inactivated as described above. Microbubbles were then collected and reacted with a rabbit anti-GFP polyclonal antibody for 10 min at room temperature and then with an anti rabbit-TRITC conjugated antibody for another 10 min at room temperature. Microbubbles were placed between a glass slide and cover slip for immunohistochemistry and were sealed with nail polish. A deconvolution fluorescence microscope (Olympus IX81, Melville, NY) was used to document fluorescence on/of the microbubbles. Microbubbles reconstituted in the absence of Ad-GFP constituted the negative control. Animal preparation and ultrasonic bubble destruction
Animal studies were performed in accordance with NIH recommendations and the approval of the institutional animal research committee. Animal care and humane use and treatment of mice were in strict compliance with (1) institutional guidelines, (2) the Guide for the Care and Use of Laboratory Animals (National Academy of Sciences, Washington, DC, 1996), and (3) the Association for Assessment and Accreditation of Laboratory Animal Care International (Rockville, MD, 1997). Twenty nude mice (female NU/NU-nuBR outbred, isolatormaintained mice, 4–5 weeks old, from Charles Rivers Laboratories, Wilmington, MA) were implanted with the human melanoma cell line DB-1 as a xenograft model (injecting 2.5 106 DB-1 cells per flank). After 20 days, when the tumors reached a volume of approximately 50 mm3, mice were sedated with 0.0133 ml/g of a mixture of Xylazine hydrochloride (10 mg/kg; Gemini, Rugby Laboratory, Rockville Centre, NY) and Ketamine hydrochloride (20 mg/kg; Ketaset, Aveco, Fort Dodge, IA) administered intraperitoneally. The mice were placed on a mat warmed with 378C circulating water for the entire procedure. A 27-gauge needle with a heparin lock was placed within a lateral tail vein for administration of
415
ULTRASOUND GUIDED SITE SPECIFIC GENE DELIVERY
TABLE 1. Summary of the experimental design Group #1
Group #2
Group #3
Imagent þ Ad-GFP Complement treated 657 kPa
Imagent þ Ad-GFP Untreated
Imagent þ Ad-GFP Complement treated 506 kPa
657 kPa
Group #4 Imagent þ Ad-GFP Untreated 506 kPa
contrast material. The nude mice received an injection of 100 mL of Imagent with/without Ad-GFP through the tail vein. The mice were split into two control groups (one mouse receiving 100 mL of PBS and US, and another control mouse receiving both Imagent and US) and six active groups of three mice each (all receiving Imagent and Ad-GFP viruses and US) but with/without pretreatment with the inactivating agent) (see Table 1). Grayscale US imaging was performed with an Aplio scanner (Toshiba America Medical Systems, Tustin, CA) using a broad bandwidth (4–10 MHz) linear array operating at 4.0 MHz. On the Aplio scanner three different acoustic output settings (100, 64, and 20%) were employed for 4 min of exposure. These output power settings correspond to peak negative pressures of 657, 506, and 304 kPa (respectively determined as described previously). Ultrasound images were recorded as digital clips. The mice were sacrificed 48 h after completion of the experiments, by placing them in a CO2 gas jar placed in a ventilated fume hood. The tumors, heart, lungs and liver were harvested. Tissues to be sectioned for fluorescence analysis were placed in OCT (Sakura Finetek USA, Inc., Torrance, CA), frozen in liquid nitrogen, and stored at 808C. Tissues to be sectioned for immunohistochemical analysis were preserved in neutral buffered formalin at 48C prior to embedding in paraffin. Determination of GFP fluorescence and immunohistochemical staining
For GFP imaging, frozen specimens were sectioned 3–4 mm thick using a cryostat. The sections were mounted with Vecta Shield (Vector Laboratories, Burlingame, CA) and expression of GFP was examined under fluorescence microscopy (Olympus, Melville, NY). Sections were also processed for hematoxylin and eosin (H&E) staining. For immunohistochemical (IHC) analysis, formalin fixed and paraffin embedded specimens were sectioned 3–4 mm thick. Sections were deparaffinized, re-hydrated and then quenched in 3% H2O2 for 20 min. Sections were washed with PBS and blocked in PBS containing 1% BSA for 20 min at 378C. Monoclonal anti-GFP (Invitrogen, Carlsbad, CA) 1:2,000 was incubated 1 h at 378C and then washed three times in PBS. Sections were incubated with an avidin-biotin-peroxydase complex (Vectastain Elite ABC kit, Vector Laboratories) and then washed two more times in PBS. The immunoreactivity was determined using diaminobenzidine (DAB) as the final chromogen. Finally, sections were counterstained with Meyer’s Hematoxylin, dehydrated through a sequence of increasing concentration alcoholic solutions, cleared in xylene and mounted with epoxydic medium. During the immunohistochemical assay proof slides were coupled with negative control slides on which the primary antibody was omitted.
RESULTS Test of various ultrasound contrast agents in vitro for auto-fluorescence
Because the purpose of our study was to evaluate the feasibility of gene delivery mediated by diagnostic ultrasound and adenovirus-GFP encapsulated, pretreated or not with complement, in a rodent tumor xenograft model, we tested various ultrasound contrast agents and cell lines for auto-fluorescence. Three cell lines of different origin (H23, human lung adenocarcinoma; DU-145, human prostate adenocarcinoma; and DB-1, human melanoma) were chosen for their ability to Journal of Cellular Physiology DOI 10.1002/jcp
Group #5 Imagent þ Ad-GFP Complement treated 304 kPa
Group #6
Group #7
Group #8
Imagent þ Ad-GFP Untreated
CTRL 1
CTRL 2
PBS þ US
Imagent þ US
304 kPa
657 kPa
657 kPa
grow in nude mice. Cells were plated on glass coverslips, which were placed into 60 mm dishes, and 100 mL of the contrast agents Optison, SonoVue, Levovist, Sonazoid, and Imagent were added to the culturing medium. High or low intensities of ultrasound were applied for 2 min to each dish. Twenty-four hours later cells were fixed in 4% paraformaldheyde and observed under a fluorescence microscope. Figure 1 demonstrates that the DU-145 cells exposed to the various ultrasound contrast agents did not show any sign of fluorescence following exposure to US at 657 or 304 kPa. Comparable results were obtained at the same pressure levels with the H23 and DB-1 cells (data not shown). Complement inactivates adenoviruses adsorbed on the microbubbles’ surface in vitro
Reconstitution of microbubbles in the presence of adenoviruses may result in different scenarios. Viruses may attach to the microbubbles’ membrane or may be imbedded in the membrane, or may be enclosed within the bubbles. If viruses are attached to the surface of the bubbles, non-targeted viral release within the capillary bloodstream may occur (Howard, 2004). Therefore, the hypothesis that complement may inactivate not only free (not enclosed) adenoviruses but also adenoviruses adsorbed on the surface of the reconstituted ultrasound contrast agent was tested. If the hypothesis is confirmed the target specificity of gene therapy via ultrasound microbubble destruction would be improved. First, the ability of complement to inactivate adenoviruses was tested. Du-145 and H23 cells were plated on glass coverslips, which were placed into 60 mm dishes, and 10 mL of Ad-GFP (1 1012/ml) were used to transduce the cells (Fig. 2A,D). Another set of dishes was also transduced with 100 mL of a solution of Ad-GFP incubated with 10 volumes of complement for 30 min at 378C (Fig. 2B,E). Clearly, 30 min of incubation of the adenovirus with complement was enough to inactivate the adenovirus as demonstrated by the lack of fluorescence of Figure 2(parts B and E compared to parts A and D). Second, the viruses with respect to the microbubbles were localized following complement inactivation of adenoviruses adsorbed on the surface of freshly formed microbubbles. Imagent AF0150 microbubbles were prepared in the presence (Fig. 3A) or absence of AdGFP and the ability of complement to inactivate the adenovirus not encapsulated into the bubbles by treating the bubbles for 30 min at 378C with complement was tested. The bubbles were then reacted with a rabbit anti-GFP polyclonal antibody and with an anti rabbit-TRITC conjugated antibody. Figure 3B shows that Ad-GFP microbubbles not treated with complement but reacted with rabbit anti-GFP and anti-rabbit TRITC appeared as a yellow color indicating that the anti-rabbit TRITC conjugated antibody was reacting with the anti-GFP antibody on the bubble’s surface (see red arrowheads). On the other hand, Ad-GFP microbubbles treated with complement to inactivate the
416
HOWARD ET AL.
Fig. 1. Fluorescence microscopy image of DU-145 cells exposed to 100 mL of different contrast agents: (A) Imagent, (B) Optison, (C) SonoVue, (D) Sonazoid, and (E) Levovist. Lack of fluorescence following exposure to US at 657 or 304 kPa is demonstrated. F: Negative controls were treated with ultrasound alone.
adenovirus and reacted with rabbit anti-GFP and antirabbit TRITC appeared as a green color indicating that the anti-rabbit TRITC conjugated antibody was not reacting with the anti-GFP antibody on the bubble’s surface (Fig. 3C, see yellow arrowheads). This data indicates that the treatment with complement of the Ad-GFP bubbles inactivated the adenoviruses present on the bubble’s surface leaving instead intact and viable
the adenoviruses encapsulated by the microbubbles without destroying the integrity of the Imagent bubble shell. A second set of experiments was performed to evaluate the ability of complement to inactivate adenoviruses adsorbed on the microbubbles’ membrane. For this experiment, DU-145 cells were plated on glass coverslips, which were placed into 60 mm dishes.
Fig. 2. Fluorescence microscopy image of Du-145 and H23 cells transduced with Ad-GFP. A: DU-145 cells transduced with 10 mL of Ad-GFP (1 1012/ml); (B) DU-145 cells transduced with 100 mL of a solution of Ad-GFP incubated with 10 volumes of complement for 30 min at 378C; (C) DU-145 cells incubated with 100 mL of PBS (negative control); (D) H23 cells transduced with 10 mL of Ad-GFP (1 1012/ml); (E) H23 cells transduced with 100 mL of a solution of Ad-GFP incubated with 10 volumes of complement for 30 min at 378C; (F) H23 cells incubated with 100 mL of PBS (negative control).
Journal of Cellular Physiology DOI 10.1002/jcp
ULTRASOUND GUIDED SITE SPECIFIC GENE DELIVERY
417
Fig. 3. Fluorescence microscopy images of Ad-GFP/bubbles. A: Imagent bubbles reconstituted in the presence of Ad-GFP and reacted with anti-rabbit TRITC conjugated antibody. B: Imagent bubbles reconstituted in the presence of Ad-GFP and reacted with rabbit anti-GFP antibody and anti-rabbit TRITC conjugated antibody. C: Imagent bubbles reconstituted in the presence of Ad-GFP, treated with complement for 30 min at 378C and reacted with rabbit anti-GFP antibody and anti-rabbit TRITC conjugated antibody. D: Phase contrast image of Imagent bubbles.
Fig. 4. Fluorescence microscopy images of Du-145 and H23 cells transduced with the aid of Ad-GFP/bubbles. A: Cells insonified at 535 kPa peak negative pressure after administration of Ad-GFP Sonazoid bubbles; (B) Cells insonified at 207 kPa peak negative pressure after administration of Ad-GFP Sonazoid bubbles; (C) Cells insonified at 535 kPa peak negative pressure after administration of complement treated Ad-GFP Sonazoid bubbles. D: Cells insonified at 207 kPa peak negative pressure after administration of complement treated AdGFP Sonazoid bubbles. E: Cells insonified at 535 kPa peak negative
Journal of Cellular Physiology DOI 10.1002/jcp
pressure after administration of Ad-GFP Imagent bubbles; (F) cells insonified at 207 kPa peak negative pressure after administration of Ad-GFP Imagent bubbles; (G) cells insonified at 535 kPa peak negative pressure after administration of complement treated AdGFP Imagent bubbles. H: Cells insonified at 207 kPa peak negative pressure after administration of complement treated Ad-GFP Imagent bubbles. I: Cells insonified at 207 kPa peak negative pressure after administration of Imagent bubbles (negative control).
418
HOWARD ET AL.
Fig. 5. Fluorescence imaging of mice tissues and DB-1 melanoma xenografts obtained from mice injected in their tail vein with untreated Ad-GFP/microbbubbles. A: Nude mice with melanoma xenografts (DB-1 cells) implanted on its flank; (B) dissected tumor showing the vascular axis; (C) US enhancement of tumor vasculature by US-targeted microbubble destruction (see red arrowhead). D–F: H&E staining of tumor, hearth, and lung, respectively; (G–I) light microscopy of tumor, hearth, and lung, respectively; (J–L) fluorescence microscopy of tumor, hearth, and lung, respectively.
Optison, SonoVue, Levovist, Sonazoid, and Imagent contrast agents were tested for this purpose. Best results were obtained using Sonazoid (Fig. 4A–D) and Imagent (Fig. 4E–H) and are shown. Cells were insonified at 535 (Fig. 4A, C, E, and G) or 207 kPa (Fig. 4B, D, F, and H) peak negative pressure for 1 min after administration of complement treated or untreated bubbles prepared in the presence of the viral vector for gene delivery. Complement treated Sonazoid bubbles were insonified at 535 (Fig. 4C) or 207 kPa (Fig. 4D) peak negative pressure. Complement treated Imagent bubbles were insonified at 535 (Fig. 4G) or 207 kPa (Fig. 4H) peak negative pressure. Microbubbles alone were used as negative control (Fig. 4I). Twentyfour hours later the marker gene (GFP) transduction efficiency was demonstrated by fluorescence microscopy. Imagent more efficiently enclosed and protected Ad-GFP viruses demonstrating that this particular simplified approach could be a viable gene delivery strategy with Imagent. In fact, 207 kPa US peak negative pressure efficiently delivered the GFP adenovector enclosed in Imagent bubbles and treated with complement to inactivate the unenclosed viral particles to 93% of the cells (Fig. 4H). Five hundred thirty-five kilo Pascals peak negative pressure did not restore either Sonazoid (Fig. 4C, 27% of the cells) or Imagent (Fig. 4G, 28% of the cells) GFP signal intensity. Sonazoid also Journal of Cellular Physiology DOI 10.1002/jcp
partially restored the GFP signal intensity, but to a lower extent (56% of the cells) using 207 kPa ultrasound waves (Fig. 4D); therefore, it was decided to proceed with Imagent alone for the in vivo experiments. Complement inactivates adenoviruses adsorbed on the microbubbles’ surface and ultrasound allows site-specific gene delivery system in vivo
Nude mice with melanoma xenografts (DB-1 cells) implanted on their flank, received injections of either PBS or Imagent Ad-GFP (dose: 0.1 ml) through their tail vein. Mice were split into two control groups and six active groups, see Table 1. US was performed for 4 min at 4.0 MHz using an Aplio scanner (Toshiba America Medical Systems, Tustin, CA) operating at one of three power levels (Fig. 5C). Animals were sacrificed 48 h after imaging, and the tumors, heart, lungs, and liver were harvested and either frozen in OCT (Sakura Finetek USA) or formalin fixed and paraffin embedded. Specimens underwent phase contrast light, fluorescent microscopy, and H&E staining as well as immunohistochemistry using a monoclonal antibody against GFP. Tumors and their vascular axis are shown in Figures 5A–C; 6A–C; and 7A,B, respectively. In vivo, systemic delivery of Ad-GFP microbubbles not treated with complement and injected in the tail vein of
ULTRASOUND GUIDED SITE SPECIFIC GENE DELIVERY
419
Fig. 6. Fluorescence imaging of mice tissues and DB-1 melanoma xenografts obtained from mice injected in their tail vein with complement treated Ad-GFP/microbbubbles. A: Nude mice with melanoma xenografts (DB-1 cells) implanted on its flank; (B) dissected tumor showing the vascular axis; (C) US enhancement of tumor vasculature by US-targeted microbubble destruction (see red arrowhead). D–F: H&E staining of tumor, hearth, and lung, respectively; (G–I) light microscopy of tumor, hearth, and lung, respectively; (J–L) fluorescence microscopy of tumor, hearth, and lung, respectively.
nude mice resulted in non-specific targeting of the GFP transgene following 4 min of US application (Fig. 5J–L, respectively). In fact, not only the tumors but also nontargeted organs such as the heart and lungs were positive to 488 nm fluorescence light excitation. Figure 5A,B shows the gross anatomy of a tumor xenograft and its vascular axis. Figure 5C shows the tumor vascular bed imaged with ultrasound and in the presence of the contrast agent (see red arrowhead). Parts (D, G), (E, H), and (F, I) show the hematoxylin and eosin staining and the transmission light microscope image of tumor, heart and lung tissues harvested from the same mouse, respectively. On the other hand, in vivo, systemic delivery of AdGFP microbubbles pretreated with complement resulted in specific targeting of the GFP transgene (compare Fig. 6 parts J–L). In this set of mice, tumors targeted by 304 kPa of US were positive to the GFP marker, whereas heart and lung harvested from the same mice were negative to the 488 nm fluorescence light excitation. Figure 6A,B shows the gross anatomy of a tumor xenograft and its vascular axis. Figure 6C shows the tumor vascular bed imaged with US and contrast agent (see red arrowhead). Figure 6 parts Journal of Cellular Physiology DOI 10.1002/jcp
(D, G), (E, H), and (F, I) shows the H&E staining and the transmission light microscope image of sections from the tumor, heart and lung harvested from the same mouse, respectively. Systemic delivery of Ad-GFP microbubbles pretreated with complement and injected in the tail vein of nude mice resulted in image guided site specific targeting of the GFP transgene within the tumor alone as shown in Figure 6J. Immunohistochemical analysis using a monoclonal antibody against GFP was also performed. Figure 7 parts A–D shows the H&E staining of tumor, lung, heart and liver sections from the same mouse, respectively. Untreated Ad-GFP microbubbles resulted in non-targeted GFP expression evidenced by the DAB positive staining in all tested tissues, see red arrowheads (Fig. 7 parts E–H, respectively). Complement treated Ad-GFP microbubbles resulted in specific GFP expression targeted solely to the tumor (Fig. 7I, see red arrowhead) and not the other tested tissues, as evidenced by immunohistochemical analysis (compare Fig. 7 parts I–L). Negative control mice were treated with ultrasound contrast agents alone and ultrasound and did not show any immunohistochemical staining (Fig. 7M–P, respectively).
420
HOWARD ET AL.
Fig. 7. Immunohistochemical analysis of mice tissues and DB-1 melanoma xenografts prepared from mice injected in their tail vein with either untreated or with complement treated Ad-GFP/microbbubbles using an antibody against GFP. A–D: H&E staining of tumor, lung, hearth, and liver, respectively; (E–H) immunohistochemical analysis of nude mice injected with untreated Ad-GFP/ microbubbles into the tail vein and resulting in untargeted GFP expression in tumor, lung, hearth and liver, respectively (see red
arrowheads); (I–L) immunohistochemical analysis of nude mice injected with complement treated Ad-GFP/microbubbles into the tail vein and resulting in specific GFP expression only in tumor (see red arrowhead), but not in lung, hearth and liver, respectively; (M–P) immunohistochemical analysis of mice injected with Imagent alone into the tail vein resulting in negative staining of all the examined tissues (tumor, lung, hearth and liver).
DISCUSSION
microbubbles (Fig. 3). This opens up novel therapeutic avenues for patients in need of a less invasive and more specific systemic approach. Moreover, we found that US imaging of microbubbles containing Ad-GFP at diagnostic acoustic pressures (
