| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Molecular Imaging |
and
Michael Schirner, MD
1 From Schering, Corporate Research Business Area Diagnostics and Radiopharmaceuticals, Ultrasound & New Modalities Research, Müllerstrasse 178, D-13342 Berlin, Germany (P.H., S.S., M. Schultze-Mosgau, M.R., T.B., M. Schirner); Group Drug Targeting, Max-Delbrück-Center for Molecular Medicine, Berlin, Germany (R.R.); and Department of Hepatology, Gastroenterology, Endocrinology and Metabolism, Charité, Campus Virchow-Klinikum, Humboldt-University, Berlin, Germany (T.P., S.R.). Received May 13, 2004; revision requested July 30; revision received September 16; accepted October 20. Address correspondence to P.H. (e-mail: peter.hauff{at}schering.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: This study was performed according to a protocol approved by the regional animal research committee. The model plasmid UT651 (pUT651) that contained the Escherichia coli LacZ gene for ß-galactosidase was used to demonstrate the feasibility of ultrasound-mediated gene delivery in CC531 liver tumors in rats. In preliminary experiments, a single injection of pUT651-containing GFMP was administered intraarterially (n = 4) or intravenously (n = 6) with simultaneous sonication (color Doppler mode, maximum mechanical index) of the GFMP passing through the capillaries of the tumors. All animals were sacrificed 2–5 days later, and liver tumors were examined for ß-galactosidase expression histochemically. Subsequently, potential medical usefulness of this delivery system was tested in nude mice bearing Capan-1 tumors (adenocarcinoma of the human pancreas) by using the plasmid RC/CMV-p16 (pRC/CMV-p16), which contains tumor suppressor gene p16. The tumor suppressor gene p16 is deleted in Capan-1 cells. Twenty-five tumor-bearing mice were classified into five groups (four to six mice per group, one treatment group, four control groups) at random. All mice were treated once weekly for 5 weeks with intravenous infusion of p16-containing GFMP or control substances with simultaneous tumor sonication with color Doppler mode ultrasound and maximum mechanical index or without ultrasound treatment. The therapeutic effect of p16 was measured as an increase in tumor volume doubling time. Data were analyzed with analysis of variance. Results were considered significant at the 5% critical level (P < .05).
RESULTS: A clear expression of pDNA was found in tumors in rats treated with a combination of pUT651-containing GFMP and ultrasound; relevant controls showed a significantly lower expression of marker gene. The controlled ultrasound-triggered release of pRC/CMV-p16 from GFMP leads to a strong tumor growth inhibition, which is significant (P < .002), compared with that in controls.
CONCLUSION: A combination of GFMP and ultrasound provides an effective approach for nonviral gene therapy–based cancer treatment.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Several viral and nonviral delivery systems are being investigated for transporting and delivering therapeutic genes in vivo, and both types of systems have advantages and disadvantages. Viral or virus-associated delivery systems are usually able to achieve a relatively high gene transfection of tumor cells. They are disadvantaged, however, by their immunogenicity (which can be boosted after repeated administrations) and a certain risk of insertional mutagenesis, which could lead to an activation of cellular oncogenes or inactivation of tumor suppressor genes (4,5). Nonviral delivery systems such as cationic lipids, liposomes, or polymeric microspheres have other potential advantages, such as ease of preparation and scale-up, as well as immune tolerance and safety profile. They are limited, however, by a relatively low transfection efficiency (6). Hence, the currently available viral and nonviral delivery systems are limited by their lack of organ and cell specificity.
To overcome this inherent limitation, we developed a nonviral gene delivery system with gas-filled microparticles (GFMP) and diagnostic ultrasound (7). Ultrasonographic contrast media research has shown that diagnostic ultrasound at suitable frequencies and intensities can destroy GFMP, and destruction results in the release of the encapsulated gas. Provided that the additional inclusion of suitable substances is feasible, these GFMP can be used as carriers for genetic material, and this role permits their potential use as a nonviral gene delivery system. Moreover, DNA encapsulated in this new nonviral carrier may be protected from harmful influences (eg, nucleases) during its circulation in blood. Ultrasound-triggered release of DNA would allow spatially and temporally controlled and precisely targeted cancer gene therapy. In addition, diagnostic ultrasound can promote and enhance released DNA uptake by target cells, and this capability is called sonoporation (8).
The advantages of this approach include the low level of invasiveness of intravenous administration of microencapsulated DNA, as well as the possibility to visualize the breakup of the GFMP with concomitant release of contents on the screen of the ultrasound device through the color pixels that are produced when the particles are destroyed (stimulated acoustic emission effects) by using the color Doppler mode (9). Thus, the purpose of our study was to evaluate the feasibility of gene delivery mediated with diagnostic ultrasound and plasmid DNA (pDNA) encapsulated in GFMP in rodent tumor models.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Chemicals
Poly(D,L-lactide-co-glycolide) (Resomer RG503; Boehringer Ingelheim, Ingelheim, Germany), with a monomer ratio of 50:50 and inherent viscosity of 0.4 dL/g, was purchased. Gelatin from porcine skin (Fluka, Neu-Ulm, Germany), with a medium gel strength of 180 g Bloom, was purchased, and the DNA-complexing poly-L-lysine (Sigma, Deisenhofen, Germany), with molecular weight of 9.5 kDa, was obtained. Dulbecco modified Eagle medium, RPMI 1640 medium, fetal calf serum, and a mixture of penicillin and streptomycin (Life Technologies, Karlsruhe, Germany) were obtained. All the other chemicals and reagents were of analytic quality.
Plasmids
The plasmid UT651 (pUT651) containing the marker gene Escherichia coli LacZ for ß-galactosidase (Cayla, Toulouse, France), flanked upstream by the human cytomegalovirus (CMV) promoter and downstream by the transcriptional termination and polyadenylation signals from simian virus 40, was obtained. The plasmid was propagated in an E coli strain DH5
culture (Invitrogen, Carlsbad, Calif) and purified with a large-scale plasmid DNA purification kit (Plasmid Maxi Kit; Qiagen, Hilden, Germany). The eukaryotic expression plasmid RC/CMV (pRC/CMV) (Invitrogen) contains neomycin resistance with control of the CMV promoter. (Hereafter, neomycin resistance will be designated as "neo.") Full-length human p16 complementary DNA was subcloned into pRC/CMV, generating the pRC/CMV-p16 construct. Orientation of inserts was confirmed with sequencing (10).
Preparation of GFMP
GFMP of poly(D,L-lactide-co-glycolide) that contained pDNA were prepared as described in detail previously (7). The pUT651 containing the E coli LacZ gene for ß-galactosidase for marker gene studies, the pRC/CMV-p16 containing the tumor suppressor gene p16 for the treatment study, and the pRC/CMV-neo used as an empty vector control were encapsulated with a water-in-oil-in-water-emulsion solvent evaporation technique. The pDNA was stabilized with poly-L-lysine prior to encapsulation with a pDNA-polymer ratio of 2:1. The integrity of the encapsulated pDNA was checked by means of agarose gel electrophoresis (data not shown) by two investigators (S.S., M. Schultze-Mosgau).
Animals
Ten male WAG/Rij-rats (Harlan Winkelmann, Borchen, Germany) that weighed 250–300 g were purchased for the marker gene studies, and 25 female NMRI/nu mice (TZH Schönwalde, Schönwalde, Germany) that weighed 18–20 g were purchased for the tumor suppressor gene study. The studies were performed in accordance with a protocol approved by the regional animal research committee.
Tumor Cells
The CC531 rat colon carcinoma cell line was grown as a subconfluent monolayer in Dulbecco modified Eagle medium containing 10% fetal calf serum and growth medium (GlutaMax; Invitrogen). The CC531 cell line, which is syngeneic with the WAG/Rij rat, was used for the marker gene studies in two preliminary experiments. The Capan-1 human pancreatic adenocarcinoma cell line was obtained from a supplier (American Type Culture Collection, Rockville, Md). The Capan-1/p16 and Capan-1/neo clones were prepared as described previously (10). Both Capan-1 cell lines were grown as subconfluent monolayers in RPMI 1640 medium containing 15% fetal calf serum. Culture media were supplemented with 2 mmol/L glutamine, 100 IU/mL penicillin, and 100 µg/mL streptomycin. All cell lines were cultured in a humidified atmosphere of 95% air and 5% CO2 at 37°C by one researcher (T.P.).
Tumor Induction
For the marker gene studies, anesthetized rats underwent a midline laparotomy, and 3 x 105 CC531 colon carcinoma cells were implanted in the liver subcapsularly. For the treatment study, the prepared Capan-1/neo and Capan-1/p16 tumor cell suspensions were implanted subcutaneously in the right region of the flank in anesthetized nude mice in a concentration of 1.5 x 106 cells per mouse. One investigator (S.S.) implanted Capan-1/neo tumor cells in 20 nude mice and Capan-1/p16 tumor cells in five nude mice. The rats were anesthetized with an intraperitoneal injection of 60 mg sodium pentobarbital per kilogram of body weight (Nembutal; Wirtschaftsgenossenschaft deutscher Tierärzte eG, Hannover, Germany), and the mice, with an intraperitoneal injection of 100 mg/kg ketamine hydrochloride (Ketavet; Pharmacia & Upjohn, Erlangen, Germany) and 10 mg/kg xylazine (Rompune; Bayer, Leverkusen, Germany).
Experimental Design
For a first preliminary in vivo marker gene experiment, midline laparotomy was performed in four anesthetized rats 10 days after CC531 tumor implantation. Subsequently, of four rats, two rats received an intraarterial infusion (1 mL in 5 minutes) via the arteria hepatica of 20 µg of naked pUT651-DNA, and the other two rats received encapsulated pUT651-DNA. Both forms were suspended in 1 mL 0.9% NaCl solution. One rat of each group was treated with ultrasound (ATL-UM9; ATL, Bothell, Wash) with a linear-array transducer (L10-5; ATL), and the other rat of each group was not treated with ultrasound. The following settings were used: color Doppler mode, penetration depth of 3.0 cm, and mechanical index of 1.0. Ultrasound treatment was started 1 minute prior to substance administration and was continued during the infusion and stopped at the end of the infusion by means of the continuous movement of the transducer over the tumor surface. Five days later, all animals were sacrificed, and the liver tumors were removed and frozen in 2-methylbutane cooled in liquid nitrogen and stored at –80°C by two investigators (P.H., M.R.).
The second preliminary in vivo marker gene experiment was performed in six anesthetized rats 10 days after CC531 tumor implantation. The abdominal region was shaved, and the residual hair was removed with a depilatory cream (Pilca; GlaxoSmithKline Consumer Healthcare, Brühl, Germany). The rats then were placed on their backside and received an infusion (1 mL in 5 minutes) via the tail vein of 50 µg encapsulated pUT651-DNA suspended in 2 mL 0.9% NaCl solution. Transcutaneous ultrasound treatment of the liver tumors was started 1 minute prior to substance administration and was continued during the infusion and stopped at the end of the infusion by using the ultrasound device (HDI 5000; ATL) and a transducer (L12-5; ATL) with the color Doppler mode, a penetration depth of 2.5 cm, and a mechanical index of 0.9. Twenty-four hours prior to the administration of the encapsulated pUT651-DNA, three of these rats were pretreated with intravenous administration of 0.75 mg GdCl3 dissolved in 0.9% NaCl solution (Sigma-Aldrich, Taufkirchen, Germany) per kilogram body weight; this substance effectively depresses the reticuloendothelial activity (11). This pretreatment was performed to determine whether macrophages have a significant influence on the pDNA release at the target site because of their physiologic uptake and degradation of, for example, abnormal particulate matter such as the injected circulating GFMP. The six rats were sacrificed at three time points (2, 3, and 5 days after pUT651-DNA release), two rats per time point, one with and the second without GdCl3 pretreatment. Different time points were chosen to identify the time with the strongest ß-galactosidase expression. Two investigators (P.H., M.R.) removed the liver tumors, froze them in 2-methylbutane cooled in liquid nitrogen, and stored them at –80°C.
For a treatment study, 25 mice were classified into five groups (four to six mice per group) at random. Capan-1/neo tumor cells were implanted in mice of groups 1–4, and Capan-1/p16 tumor cells were implanted in mice of group 5. All mice were visited three times per week to check for the initiation of tumor development; after initiation of tumor growth was observed, the tumor growth (length and width in millimeters) was measured at each visit until the end of the whole experiment (1 week after the sixth treatment). Treatment studies were started in mice of groups 1–4 with a tumor volume of 150–250 mm3. All mice were treated once weekly for 5 weeks as follows: group 1 (treatment group, n = 5) received encapsulated pDNA-p16 with ultrasound treatment of the tumor, group 2 (control group 1, n = 6) received encapsulated pDNA-p16 without ultrasound treatment of the tumor, group 3 (control group 2, n = 5) received encapsulated pDNA-neo with ultrasound treatment of the tumor, and group 4 (control group 3, n = 4) received only 0.9% NaCl solution with ultrasound treatment of the tumor. The Capan-1/p16 tumor cells in mice of group 5 (control group 4, n = 5) were used as a control group to show the functionality of the p16 suppressor gene.
Before the treatment was started, the mice were anesthetized and then placed on their left side. Each mouse received a total dose of 120 µl (1.2 µg encapsulated pDNA in 120 µL 0.9% NaCl solution) per treatment, which was administered intravenously in three fractions (40 µL per fraction). Each fraction was infused within 45 seconds, with an infusion break interval of 4 minutes. Sonication of the tumors was performed without any interruptions from the start of the infusion until no further stimulated acoustic emission effects signals were visible on the screen of the ultrasound device, which indicated that no further intact pDNA-containing GFMP were available. The total duration of sonication per treatment was about 12 minutes. The ultrasound treatment was performed by using an ultrasound device (ATL-UM9; ATL) with a linear-array transducer (L 10-5; ATL) on which an ultrasound gel pad (Aquaflex; Parker Laboratories, Fairfield, NJ) of 2.0 cm in thickness was mounted. The following settings were used: color Doppler mode, penetration depth of 2.5 cm, and mechanical index of 1.5. Three investigators (P.H., S.S., M.R.) performed the procedure. All sonographic treatments were documented with a videocassette tape–format video system (S-VHS, Panasonic model AG-8700E; Matsushita Electronic Industrial, Osaka, Japan).
Histochemical Analysis
Cryosections of CC531 liver tumor samples of 12-µm thickness were mounted on aminopropyltriethoxysilane-coated slides. After 1–2 minutes of air drying, the sections were fixed for 10 minutes at 4°C in 2% paraformaldehyde solution (0.1 mol/L Na2HPO4/NaH2PO4, 0.2% glutardialdehyde, and 2% paraformaldehyde), washed twice in phosphate-buffered saline (0.1 mol/L Na2HPO4/NaH2PO4), with pH 7.3, and stained for 16 hours at 37°C with a solution containing the reagent 5-bromo-4-chloro-3-indolyl-ß-D-galactoside (X-Gal; Roth, Karlsruhe, Germany). The solution contained 0.4 mg/mL of the reagent, 0.9 mmol/L MgCl2, 3 mmol/L K3(Fe[CN]6), and 3 mmol/L K4Fe(CN)6. The sections were then rinsed in phosphate-buffered saline twice, counterstained with eosin, dehydrated, and mounted in a quick-hardening mounting medium (Eukitt; Sigma-Aldrich). The prepared liver tumor sections of both preliminary rat experiments were visually scanned for ß-galactosidase expression (blue-staining products) by using a light microscope (Axioskop 40; Zeiss, Göttingen, Germany) equipped with a second observation module that allowed the parallel assessment of each slide by two investigators (P.H., S.S.) in consensus. The investigators evaluated the degree of ß-galactosidase expression subjectively in regard to the amount of blue-staining ß-galactosidase products in the whole section, according to the following scale: no ß-galactosidase expression; poor ß-galactosidase expression, less than 1%; moderate ß-galactosidase expression, greater than 1% but less than 10%; or strong ß-galactosidase expression, greater than 10%.
Data and Statistical Analysis
The treatment period was classified into three evaluation intervals: the first interval was from the first treatment until the beginning of the third treatment (the first 14 days); the second interval, from the third treatment until the beginning of the fifth treatment (the next 14 days); and the third interval, from the fifth treatment until the end of 1 week after the sixth treatment (a further 14 days). The tumor volumes were calculated with the equation: V = (a · b2)/2, where a is the length of the tumor in millimeters, b is the width of the tumor in millimeters, and V is the tumor volume in cubic millimeters (12). The slope and intercept of the linear regression of the logarithm of the volume versus time since the first treatment were estimated for each animal and each interval.
Tumor volume doubling times were calculated for each animal and each interval with the following equation: TVDT = ln(2)/
, where TVDT is the tumor volume doubling time, and is the slope (13). In all three intervals, tumor volume doubling times were analyzed by using analysis of variance. Analysis of variance was performed by using generalized linear models (GLM method of the statistical software [SAS/STAT, version 8.2 of the SAS System for Windows, 2001; SAS Institute, Cary, NC]). To compare values of the treatment group with the values of the controls, linear contrasts were used that are a part of the GLM method. All computations were performed by using statistical software. Results were considered significant at the 5% critical level (P < .05).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
In this study, we demonstrated that GFMP are suitable for the transport and delivery of genes in vivo. Moreover, we were able to demonstrate that widely used diagnostic ultrasound can also be applied for the delivery of DNA. In contrast, therapeutic and high-intensity focused ultrasound have been used for most of the described ultrasound-facilitated gene delivery and gene transfection experiments (15–21). Microbubbles, which are clinically established as contrast agents for diagnosis with ultrasound-based imaging, were shown by some of the researchers in these studies to lead to a higher DNA transfection of cells both in vitro and in vivo (22–27). The DNA, however, was either directly coupled to the surface of the microbubbles or mixed with microbubbles immediately prior to their administration. In contrast to those approaches, we developed a technique that allowed the incorporation of the DNA into the GFMP with the objective of longer DNA protection from potential neutralization caused by nucleases in the blood until immediately prior to their delivery. In addition, this technique allowed us to produce a larger amount of DNA-containing GFMP, to store this formulation for several months (data not shown), and to use it if required.
Further differences between our experiments with ultrasound-facilitated gene delivery and those in published reports include the administration procedure and duration of ultrasound treatment. Although most investigators used bolus injections and a relatively short ultrasound treatment, we decided to infuse the DNA-containing GFMP and to apply ultrasound to the tumors for as long as 12 minutes. The reason for this procedure was to ensure that as much DNA as possible could be released at the target site and that the effect of sonoporation (18,28–30), that is, the ultrasound-enhanced transfection of cells with, for example, DNA, could last as long as possible.
By using the described procedure, we were able to achieve a significant inhibition of the tumor growth in the treatment experiment. A measurable treatment success, however, was observed only from the second evaluation interval to the end of this experiment. Hence, at least two to three treatments were necessary. On the other hand, we could show that multiple injections of DNA-containing GFMP and the DNA release at the target with ultrasound are feasible for several weeks with this nonviral gene delivery approach. This approach has a clear advantage over viral or virus-associated delivery systems in regard to their immunogenicity, which can be boosted after repeated administration (4).
In most of the published reports about ultrasound-facilitated gene delivery experiments, researchers focused on finding an effective method for the genetic treatment of ischemic diseases, mainly with intramuscular injection of DNA and/or microbubbles (8,19,26,27,31). In addition, very few of the researchers investigated this approach for tumor gene therapy by using marker genes that do not exert a therapeutic effect, such as ß-galactosidase and luciferase, mainly by intratumoral substance administration (16,18). To the best of our knowledge, for the first time, we determined a tumor treatment effect after intravenous injection of p16-containing GFMP and the target- and time-controlled DNA release with a common diagnostic ultrasound device. A further improvement of this approach can be expected with the capability of targeting microbubbles to different tissues or regions of disease, as was delineated in previously published articles (32–35). This improvement would allow a site-specific delivery of therapeutic genes, for example, to tumor tissue or other diseased areas.
In summary, we demonstrated that GFMP and widely used diagnostic ultrasound provide a new approach for nonviral delivery of therapeutic genes for the treatment of tumors. Furthermore, the treatment procedure could be repeated several times in the same mouse. Additional advantages lie in monitoring of the DNA release through the visible stimulated acoustic emission effects by means of disruption of GFMP, cell transfection enhancement by using sonoporation, less invasiveness, and use in each facility that has access to diagnostic ultrasound systems.
Several limitations of the study must be addressed. First, because of the small number of animals included in the preliminary rat studies, the results should be interpreted with caution (interpretation should be limited). Second, the dosages of the pDNA-containing GFMP and the sonication procedures that were used have not been optimized for rat and mouse experiments. Finally, it is likely that treatment results will vary with other types of tumors because of differences, for example, in their blood supply or in their responses to a certain therapeutic gene.
Successful gene therapy requires not only gene constructs that can interfere with or replace the function of specific genes but also an effective delivery system to bring the gene constructs to the target site. In this study, results of the delivery approach tested show that even common diagnostic ultrasound systems are suitable for ultrasound-mediated gene delivery and that multiple injections of encapsulated DNA are practicable. Our findings suggest that this approach has future potential as a safe and effective nonviral gene delivery system for the treatment of tumors or other diseases.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: CMV = cytomegalovirus • GFMP = gas-filled microparticles
See Materials and Methods for pertinent disclosures.
Deceased
Author contributions: Guarantors of integrity of entire study, P.H., S.S., R.R., M.R., M. Schirner; study concepts, P.H., S.S., R.R., M. Schultze-Mosgau, T.P., S.R., M. Schirner; study design, P.H., S.S., R.R., M.R., T.B.; literature research, P.H., S.S., R.R.; experimental studies, P.H., S.S., M. Schultze-Mosgau, M.R., T.P.; data acquisition, P.H., S.S., M. Schultze-Mosgau, M.R., T.P.; data analysis/interpretation, P.H., S.S., R.R., M.R., T.B., M. Schirner; statistical analysis, P.H., T.B.; manuscript preparation, P.H., S.S., M.R., T.B.; manuscript definition of intellectual content, P.H., R.R., S.R., M. Schirner; manuscript editing, P.H., M.R., M. Schirner; manuscript revision/review and final version approval, all authors
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Duvshani-Eshet, O. Benny, A. Morgenstern, and M. Machluf Therapeutic ultrasound facilitates antiangiogenic gene delivery and inhibits prostate tumor growth Mol. Cancer Ther., August 1, 2007; 6(8): 2371 - 2382. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
