- Email: [email protected]
Integration and Long-Term Expression in Xenografted Human Glioblastoma Cells Using a Plasmid-Based Transposon System
ARTICLE
doi:10.1016/j.ymthe.2004.05.005
Integration and Long-Term Expression in Xenografted Human Glioblastoma Cells Using a Plasmid-Based Transposon System John R. Ohlfest, Paul D. Lobitz, Scott G. Perkinson, and David A. Largaespada* Department of Genetics, Cell Biology, and Development, University of Minnesota, and University of Minnesota Cancer Center, Arnold and Mabel Beckman Center for Transposon Research, Minneapolis, MN 55455, USA
Available online *To whom correspondence and reprint requests should be addressed at the Department of Genetics, Cell Biology, and Development, University of Minnesota, 6-160 Jackson Hall, 321 Church Street SE, Minneapolis, MN 55455, USA. E-mail: [email protected].
Available online 11 June 2004
Gene therapy has the potential to become an effective component of cancer treatment by transferring genes that cause immunomodulation or tumor cell death or that inhibit angiogenesis into tumor cells or tumor-associated stroma. Viral vectors have been the primary gene transfer vehicles used for intratumoral gene transfer to date. Plasmid-based vectors may be safer and more scalable than viral vectors. However, attempts at plasmid-based intratumoral gene transfer have been met with transient expression and poor gene transfer efficiency. Here we report integration and long-term expression of reporter genes in human glial tumors, growing in nude mice, using the Sleeping Beauty (SB) transposon system. A two-plasmid system was used, in which linear polyethylenimine was complexed with a GFP, NEO, or luciferase transposon plasmid and a SB transposase-expressing plasmid. SB-mediated transposition led to chromosomal integration of the NEO transgene in roughly 8% of tumor cells. SB-mediated insertions were cloned from the genomes of glial tumor cells to provide molecular proof of transposase-mediated integration. Luciferase studies showed that SB facilitated long-term expression of the transgene in glial tumors. SB-mediated intratumoral gene transfer is a novel, nonviral technique that could be used to augment conventional therapy for glioblastoma or other cancers. Key Words: glioblastoma, Sleeping Beauty, cancer gene therapy, transposon
INTRODUCTION Cancer gene therapy studies have demonstrated that tumors can be killed directly, starved of blood supply, or targeted for immune-mediated destruction by the expression of exogenous gene products within the tumor [1 – 10]. Viral vectors have proven to be an effective and widely used gene delivery system for delivering anti-tumor vectors into solid tumors [1,2,11 – 14]. However, immune reactions to viral vectors can inhibit gene transfer efficacy [15] and such reactions raise safety concerns, as demonstrated by an unfortunate clinical trial death [16]. In addition, although most viral vectors are designed to be replication incompetent, several instances of recombination events leading to replication-competent virus have been reported [17,18]. There is little doubt that viral vectors can be made less immunogenic [19] and that production challenges can be overcome. Nonetheless, there is
260
clearly considerable benefit in developing nonviral methods for gene therapy. Attempts at intratumoral gene transfer with plasmid DNA have yielded transient or steadily declining expression ([20,21], Fig. 5). One possible explanation for transient expression when using plasmids is the eventual degradation of episomal plasmid or a decreasing percentage of cells harboring the plasmid in highly mitotic tumors. Long-term expression of genes that inhibit angiogenesis, invoke tumor-specific immune responses, or cause tumor cell death is needed to enhance the therapeutic effect of nonviral cancer gene therapy vectors [reviewed in [22]]. Integrating vectors could lead to clonal expansion of transduced tumor cells, thereby increasing the anti-tumor effect. Brain tumors, such has glioblastoma multiforme (GBM), represent an ideal target for gene therapy because the primary tumor causes death and conventional therapies yield a median survival of only
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy 1525-0016/$30.00
ARTICLE
doi:10.1016/j.ymthe.2004.05.005
1 year [13]. Moreover, preclinical work in animal models has shown considerable efficacy for gene therapy on glioblastoma [6,14,23 – 25]. If antiangiogenic gene therapy is to halt tumor growth, expression of the inhibitor must be sustained to prevent angiogenesis and subsequent tumor regrowth [2,26]. Inhibitors of angiogenesis such as endostatin exert an anti-tumor affect, at least in part, by causing tumors to downregulate proangiogenic factors such as vascular endothelial growth factor [27,28]. Therefore, to be effective the inhibitor must be present long-term because the proangiogenic stimulus may always be present [2,26]. Likewise, suicide gene therapy should also benefit from long-term expression of the transgene. Herpes simplex virus type-1 thymidine kinase (HSV-tk) has become a widely used suicide gene because its expression is toxic only to dividing cells when the patient or animal is given nucleotide analogues, such as ganciclovir [6,12,22]. However, glioma cells are highly migratory and can become quiescent during migration, making them temporarily resistant to ganciclovir/HSV-tk therapy [22,29]. Thus, it is important to achieve long-term expression of HSV-tk to maximize the anti-tumor effects on migratory tumor cells that resume mitosis. Clinical trials for GBM that utilized HSV-tk demonstrated good safety but poor efficacy [12,13]. This was largely due to poor gene transfer rates [12,13] and highlights the need for improved gene delivery technology that is also scalable. In sum, cancer gene therapy is at a crossroads where experiments in animal models have shown great promise [1,5 – 7,14], but achieving high clinical efficacy has been difficult. Therefore there is a need to develop vector systems that are more effective, safer, and easier to produce and that support long-term expression of the transgene. Here we report transposition and long-term expression from plasmid DNA into the chromosomes of human glioma cells growing in nude mice. Sleeping Beauty (SB) is a ‘‘cut and paste’’ transposable element of the Tc1/ mariner superfamily and is active in both murine and human cells [30]. The SB transposase gene can be provided in trans (on a separate DNA molecule) or in cis (on the same DNA molecule) via plasmid. SB binds the inverted terminal repeat/direct repeat sequence of the transposon and catalyzes transposition into genomic DNA [30,31]. Preliminary studies on SB insertion site preference indicate that SB integrates at random into a TA dinucleotide [30,32,33]. To assess SB’s utility as an intratumoral gene transfer vector for GBM we studied the effects of different DNA doses, injection techniques, and doses of the SB transposase when delivered using polyethylenimine (PEI) as gene transfer reagent. PEI was used because it has been shown to transfect both tumors and murine glial cells effectively with no apparent morbidity or toxicity [20,34,35]. In sum, we show that SB transposon vectors can be delivered to xenografted GBM cells and support long-term expression in a significant percentage of cells.
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
RESULTS AND DISCUSSION Distribution of Injected DNA in Glioma Xenografts To determine the distribution of gene transfer within human glioma xenografts, we inoculated nude mice with U373 cells in the flank and subjected them to intratumoral injections of 10 Ag of a GFP plasmid (Fig. 1) complexed in PEI (N/P = 10). We injected tumors that were 6 – 8 mm in diameter. Rapid injection (100 Al in >5 s) of GFP/PEI complexes with a syringe resulted in poor gene transfer, in which most of the expression was observed on the periphery of the tumor (data not shown) and this has been reported before [20]. Slow injections with a syringe into all four tumor quadrants (100 Al in 60 s) was able to provide more evenly distributed GFP expression within the tumor. We analyzed GFP expression by fluorescence microscopy performed on wholemount tumors and cryosections and by immunohistochemistry done on cryosections (Fig. 2). After examining GFP expression distribution in cryosections (N = 7 tumors) we estimated that a range of 7 – 25% (average of 12.5%) of the tumor cells expressed GFP transiently. Tumors were taken at 48 h post injection. SB-Mediated Transposition Leads to Chromosomal Integration To quantitate the number of SB-mediated transposition events that occur intratumorally, we co-injected U373 tumors with 10 Ag of a NEO-resistance transposon plasmid and 500 ng of a CMV-driven SB transposase plasmid (Fig. 1). After a series of injections (twice over 1 week or
FIG. 1. Plasmid vectors. (A) Transposon vectors used. All vectors have an identical ColE1/AMPr backbone that is flanked by two IR/DR elements that serve as transposase binding sites and mark the end of the integrating unit. Simian virus 40 (SV40), cytomegalovirus (CMV), or chimeric chicken h-actin/ CMV enhancer (Caggs) promoter elements were used to drive expression of the transgenes. (B) Sleeping Beauty (SB) transposase vectors used. All vectors have an identical ColE1/AMPr backbone and express SB under the control of the CMV, ubiquitin (UB), and phosphoglycerate kinase (PGK) promoter elements. DDE-SB is a catalytically inactive transposase due to deletion of the catalytic domain [42] and has been used in SB gene transfer studies as a negative control [37,46].
261
ARTICLE
doi:10.1016/j.ymthe.2004.05.005
FIG. 2. GFP expression analysis on glioma tumor xenografts. Tumors were injected with PEI alone (A, B, C) or PEI/GFP plasmid complexes (N/P = 10) (D – I). (A – I) 48 – 72 h postinjection the tumors were removed with a scalpel. (A, D, G) Freshly removed tumor was placed in 1 PBS in 6-cm petri dishes and then visualized at 10 using a GFP filter set on a Zeiss microscope. (B, E, H) Tumors were removed with a scalpel, placed in OCT compound (Tissue-Tek), and flash frozen in liquid nitrogen. 10-Am frozen sections were thawed and visualized at 10 using a GFP filter on a Zeiss microscope. (C, F, I) Frozen sections were thawed and immunohistochemistry was performed using anti-GFP antibody. E and F are adjacent sections from the same tumor; note that GFP expression (E) colocalizes (arrow) with GFP immunostaining (F). The percentage of GFP-positive cells within a tumor was estimated by dividing GFP-positive cell area by total cell area for each section scored (two sections/tumor for seven tumors). These numbers were then averaged to yield an estimate of 12.5% GFP-positive cells within a tumor that express GFP transiently.
six times over 3 weeks) we sacrificed the mice, removed the tumors with a scalpel, and prepared a single-cell suspension of tumor cells. We then plated tumor cells in tissue culture dishes and selected them in 400 Ag/ml G418 or no drug to determine plating efficiency. Two slow injections of the NEO/SB/PEI complexes yielded an
262
average of 2% of the viable tumor cells that were able to form colonies under G418 selection (data not shown). Whereas six injections over 3 weeks boosted the transposition rate, an average of 8.3% of viable tumor cells formed colonies under G418 selection (Fig. 3A). We believe that colony formation is due to SB-mediated
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
doi:10.1016/j.ymthe.2004.05.005
ARTICLE
to integrate into genes but integrates randomly. In contrast, murine leukemia virus and HIV-1-based vectors have a strong preference to land near promoter/enhancer regions or within genes, respectively [36]. In the case of cancer gene therapy, one could reasonably argue that insertional mutagenesis is a not an issue because the patient already has cancer. Nevertheless, it must be taken into account that any vector delivered directly into a tumor could transfect normal cells surrounding the tumor or escape into the blood. Moreover, additional mutations in the tumor clone could accelerate disease. Therefore it is desirable to develop the safest and most effective vector systems to increase clinical efficacy and minimize nontarget effects for cancer gene therapy.
FIG. 3. Sleeping Beauty-mediated transposition in glioma tumor xenografts. (A) Tumors were co-injected with a NEO-resistance transposon plasmid and CMV-driven SB transposase (CMV-SB10, n = 6) or inactive SB transposase (CMV-DDE, n = 5) plasmid complexed in PEI (N/P = 10). After six injections (two/week for 3 weeks) the tumors were removed with a scalpel, minced, trypsinized, and vortexed to generate a single-cell suspension. Total cells were counted and plated in 400 Ag/ml G418 selection or without drug to determine viability/plating efficiency. After 3 weeks of selection colonies were counted and this number was used to determine plating efficiency (no drug) or percentage of viable cells that had stable NEO expression (400 Ag/ml G418). Error bar represents standard deviation. (B) Ten G418-resistant colonies were isolated and expanded from plates described in (A). Genomic DNA was isolated by phenol/chloroform extraction and used as template for splinkerette-mediated cloning PCR. Insertions that were cloned by splinkerette PCR were Blasted against the human genome using the ENSEMBL database.
transposition and not to random integration for two reasons: (1) when NEO is injected with a catalytically inactive SB transposase no colonies form under G418 selection (Fig. 3A) and (2) 10 insertions were cloned by linker-mediated PCR [32] and all of them appeared to be transposon integrations into a TA dinucleotide because no flanking plasmid sequence was recovered and instead insertion of the NEO transposon had occurred in glioma cell genomic DNA (Fig. 3B). SB Insertion Site Analysis Of the 10 insertions cloned, 3 landed in known or predicted genes with no apparent preference for chromosomal destination (Fig. 3B). Vigdal et al. cloned over 100 SB-mediated insertions in human cells and found that roughly 1/3 of the insertions were in genes [33], a percentage similar to that for randomly chosen TA dinucleotides [32,33]. Taken together, the preliminary data on SB insertion site preference indicate that SB does not prefer
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
Direct Injection with a Syringe Yielded Higher Gene Transfer Than a Micropump Coll et al. have previously demonstrated that a micropump is more effective at delivering DNA/PEI complexes to lung tumor xenografts than a syringe [20]. To optimize gene transfer to glioma xenografts we compared two different delivery methods and DNA doses. To determine early expression rates, we first injected U373 tumors with 10 Ag of a luciferase plasmid using either a micropump or a syringe. At 24 h post injection, we observed no significant difference in luciferase expression between micropump- or syringe-injected tumors (n = 3/group; P = 0.76). In both groups we observed roughly 1 104 RLU/mg protein/15 s (data not shown). We repeated this experiment by injecting 10 or 30 Ag of plasmid with a syringe or micropump and examined tumors for luciferase expression 72 h postinjection. At this later time point a syringe provided significantly higher gene transfer than a micropump at either DNA dose (10 Ag, P = 0.043; 30 Ag, P = 0.049, Fig. 4). These data are in clear contrast with those reported by Coll et al. in lung tumor xenografts and there are several possible explanations for this. Optimal transfection of glioma cells may differ from lung cell lines due to altered membrane permeability or other inherent cellular characteristics. In addition, the tumor stroma, microenvironment, and mouse strain could also affect transfection dynamics. As additional xenograft models are studied, including intracranial, it will become clear what method of DNA/PEI complex delivery is suitable for each individual cancer. Optimal intratumoral transfection conditions should be tested empirically by investigators when untested animal cancer models are being studied. To determine if gene transfer occurred in other tissues, we assayed several tissues for luciferase, including liver, kidney, and spleen. Generally all of these tissues had no luciferase expression, but one animal of six assayed had low-level (100 – 1000 RLU/mg protein) luciferase expression in the liver and kidney (data not shown). Thus intratumoral injection of DNA/PEI complexes primarily targets the tumor and this was verified by luciferase in vivo imaging (Fig. 5).
263
ARTICLE
doi:10.1016/j.ymthe.2004.05.005
FIG. 4. Dose and delivery optimization in glioma tumor xenografts. Tumors were injected with 10 or 30 Ag of Caggs-driven luciferase plasmid/PEI complexes (N/P = 10) with either a syringe or a micropump. 72 h after injection tumor lysates were analyzed for luciferase expression. Syringe-injected tumors had significantly higher expression at 10- (P = 0.043) or 30-Ag (P = 0.049) doses. Error bars represent standard deviation; P values derived from Student’s t test.
Long-Term Gene Transfer by Transposition in Glioma Xenografts We hypothesized that tumor cells that underwent transposition would clonally expand and lead to increasing luciferase signal over time, whereas tumors injected with a catalytically inactive form of SB transposase and luciferase transposon should eventually lose any detectable luciferase signal due to plasmid degradation or loss upon cell division. In addition, we wanted to rule out the possibility that the gene transfer we had observed in U373 xenografts was not an artifact of that particular cell line, but instead a general and reproducible result. Therefore we inoculated nude mice with another GBM-derived cell line, U87, in four places (Fig. 5). When the tumors were 6 – 8 mm in diameter we administered a single injection of luciferase transposon plasmid with or without active SB plasmid. We then monitored luciferase expression by in vivo luciferase imaging for the following 4 weeks. At 72 h postinjection, we observed relatively similar luciferase expression in all tumors as expected (Fig. 5). However, 2 weeks after injection tumors (N = 4 tumors/ group) that did not receive active SB had lost detectable luciferase expression. At 4 weeks, we observed a noticeable increase in luciferase expression in tumors that received CMV-driven SB at a 1:20 ratio of SB to luciferase transposon plasmid, respectively (Fig. 5). One explanation for increasing expression over time would be clonal expansion of cells that harbor transposon insertions. Regardless of whether clonal expansion is occurring, these data suggest that transposition facilitates long-term expression of the luciferase transgene. Taken together, these data demonstrate that integrating vectors such as SB are well suited to cancer gene
264
therapy because the anti-tumor effect of the transgene can be sustained and may even increase as the tumor cells divide. In the case of antiangiogenic vectors, longterm expression is highly desirable because decreasing endostatin expression has been reported to correlate with tumor regrowth when endostatin was delivered with a nonintegrating adenoviral vector [2]. A similar argument could be made for immunomodulation, in which it is desirable to maintain or increase expression of the transgene until the immune system kills the tumor or metastasis. Likewise, suicide gene therapy would also benefit because tumor cells that are not mitotic, and thus not susceptible to HSV-tk-mediated killing, could be killed by repeated ganciclovir injections if the cells resume mitosis. Optimizing SB-Mediated Transposition by Varying Promoter Strength and Plasmid Ratios Yant et al. have shown that too much transposase expression leads to less stable gene transfer in liver-directed studies in C57/Bl6 mice; they found that when using a strong promoter such as CMV, a ratio of 1:25 facilitates more efficient transposition than higher ratios [37]. This phenomenon is known as overexpression inhibition and has been observed in related transposable elements [38]. Moreover, a similar phenomenon was observed in cultured cell lines [39]. To improve the SB gene transfer system, it is desirable to have SB transposase expressed in every cell that has transposon substrate. Therefore two weaker promoters [40, Paul Score, unpublished], ubiquitin (UB) and phosphoglycerate kinase (PGK), were used to drive SB in trans at a 1:2.5 or 1:1 ratio of SB to transposon substrate, respectively (Fig. 5). Tumors that received these
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
doi:10.1016/j.ymthe.2004.05.005
ARTICLE
FIG. 5. SB facilitates long-term luciferase expression compared to plasmid alone. (A) Nude mice (n = 4) were given four tumors each by sc inoculation with 3.5 106 U87 cells. Tumors were then injected with 15 Ag of pT2C-Luciferase and 0.75 Ag CMV-DDE (1:20), 0.75 Ag CMV-SB (1:20), 6 Ag UB-SB (1:2.5), or 15 Ag PGK-SB (1:1) complexed with PEI (N/P = 10). The mice were then imaged for in vivo luciferase expression using the Xenogen imaging system at 72 h, 2 weeks, and 4 weeks. One representative mouse is shown at 72 h, 2 weeks, and 4 weeks post injection. (B) Graph of imaging data showing the average luciferase expression of each SB promoter group (n = 4 tumors/group). At 4 weeks post injection, CMVdriven SB facilitated significantly higher luciferase expression compared to CMV-DDE-SB (plasmid without active transposase; P < 0.001). UB- and PGKdriven SB also had significantly higher expression (P = 0.009) compared to CMV-DDE, albeit less than CMVdriven SB. Error bars represent standard deviation; P values derived from Student’s t test.
higher ratios of SB had persistent gene expression at 4 weeks, but CMV-driven SB at a 1:20 ratio resulted in the highest luciferase signal at 4 weeks (Fig. 5). We and others have also tested cis vectors in which UB or CMV is used to drive SB on the same plasmid with transposon [40]. Yet UB-driven SB in cis with a NEO transposon was not as effective in the tumor colony forming assay compared to trans delivery at a 1:1 molar ratio (data not shown) and CMV-driven cis vectors have also been reported to be inferior to trans delivery [37,40].
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
In contrast, a recent report indicates that cis vectors are superior to trans delivery when extremely weak promoters, such as uninduced metallothione (MT) or PGK, are used to drive SB expression [40]. PGK and MT are both very small promoters at 510 and 262 bp, respectively. It is possible that these promoters are ideal for cis plasmids because they result in smaller plasmid size, yet they also produce a low but optimal amount of SB protein [40]. However, these cis vectors were optimized in mouse liver using hydrodynamic gene delivery [40] and the studies
265
ARTICLE
described herein were conducted in human glioma xenografts using PEI gene delivery. Further study of cis and trans delivery of SB will facilitate optimal gene transfer in a variety of tissue types and gene delivery methods. Many factors that influence SB-mediated transposition have been identified, including the ratio of SB gene to transposon substrate in trans, orientation of SB expression cassette in cis vectors, SB promoter strength, cellular cofactors, and size of the transposon [39 – 41]. Nevertheless, CMV-driven SB at a 1:20 ratio of transposase to transposon was the most efficient combination we tested for delivering transposons to tumors with PEI. Implications for Cancer Treatment and Gene Therapy Here we described the development of a nonviral, yet integrative, intratumoral gene delivery technology for human glioblastoma cells. SB has the ability to mediate chromosomal integration, which in turn leads to longterm expression and possibly clonal expansion of mitotic tumor cells harboring transposon insertions (Figs. 3 and 5). Integration and long-term expression of HSV-tk transgenes should improve efficacy because not all cells within a tumor are mitotic at one time and are therefore temporarily resistant to HSV-tk/ganciclovir-mediated killing [22], but could be killed with ganciclovir if mitosis resumed. Accordingly, antiangiogenic gene therapy should benefit from long-term expression because the tumor’s proangiogenic stimulus may always be present and will require continuous suppression [2,26]. SB is a relatively safe integrating vector because there is no apparent preference to land in genes (Fig. 3B; [32,33]). In addition, because the SB system can be delivered as plasmid DNA, it has the potential to be more scalable for clinical use and may have a better biosafety profile than viral vectors. Although high clinical efficacy for gene therapy on glioblastoma has been difficult to achieve with other gene delivery methods [12,13], there is a large body of literature that shows good efficacy for gene therapy on glioblastoma in animal models [reviewed in 22]. SB has been shown to mediate transposition in a wide variety of human cell lines [30,42] and thus should be tested in other animal cancer models for safety and efficacy. Clinically, direct intratumoral injection of DNA/PEI complexes could be most useful when the glioma is inoperable due to anatomical location in the brain. However, the level of gene transfer that can be achieved by intratumoral injection in tumors larger than 6 – 8 mm in diameter (as described herein) remains to be ascertained. The use of improved transposon vectors [39 – 41] and hyperactive SB transposase mutants [39,41] will likely increase the percentage of tumor cells that can be stably integrated by intratumoral injection of DNA/PEI complexes. Alternatively, PEI/DNA complexes could potentially be delivered into the tumor resection cavity to prevent relapse caused by migratory glioma cells that commonly escape resection [22]. Nonviral gene transfer
266
doi:10.1016/j.ymthe.2004.05.005
to the mouse brain has been achieved by intraventricular injection of DNA/PEI complexes with no morbidity [34]. The idea of using transgenes that encode secreted proteins that act ‘‘cell nonautonomously’’ is appealing for cancer gene therapy because the gene transfer rate does not need to be high. De Bouard et al. have shown that when only 1% of an intracranial glioma xenograft expresses a secreted antiangiogenic molecule, animal survival is significantly increased [43]. We have conducted pilot experiments using PEI gene delivery, in larger glioma xenografts (up to 15 mm in diameter), and seen significant anti-tumor effects with secreted molecules, including endostatin (data not shown). HSV-tk has become widely studied due to the bystander effect; as little as 10% of the tumor transduced with HSV-tk has been sufficient to regress the tumor with ganciclovir [5]. Although 8% was the average amount of tumor cells that integrated the NEO transposon in this study, we observed one tumor with a 15% rate after six injections (Fig. 3A). Further optimization of DNA dose, injection technique, and injection frequency will likely boost the average transposition rate above the 10% minimum HSV-tk threshold that has been observed [5]. Moreover, several HSV-tk fusion proteins have been made that are secreted, significantly enhance bystander effect, and require less gene transfer than intracellular HSV-tk to ablate cells [44,45]. Therefore it is likely that the tumor transduction rate of 8% reported herein is high enough to achieve anti-tumor effects with secreted HSV-tk proteins. In addition to SB’s potential use for cancer therapy, SB also has been used to correct safely the phenotype of mouse models for hemophilia B and tyrosinemia in a transposition-dependent fashion [37,46]. Further study of SB in a variety of disease models will determine its potential for clinical use. Indeed, as viral and nonviral vector systems are improved and gene therapy strategies are refined, an increase in clinical efficacy will likely follow.
MATERIALS AND METHODS Plasmid DNA, PEI, and cell lines. Human glioblastoma-derived cell lines U373 and U87 were kindly provided by Dr. Daniel Vallera (University of Minnesota, Minneapolis, MN). All cell lines were grown in RPMI 1640 (Gibco) supplemented with 10% heat-inactivated FBS, 1% penicillin/ streptomycin, and 1% NEAA. Reporter plasmids pT2SVNEO, pT2-CMVGFP, and pT2C-Luc were generated in our laboratory using standard cloning methods. Linear polyethylenimine (aka jet PEI) was purchased from Polyplus Transfection (France). Tumor inoculations and intratumoral injections. Female BALB/c homozygous (nu/nu) mice, 6 – 8 weeks of age, were maintained in a pathogenfree environment throughout the experiments. Animals were inoculated by sc injection with GBM-derived cell lines U373 (5 106) and U87 (3.5 106). Three to four weeks after inoculation the tumors were 6 – 8 mm in diameter and were injected with 100 Al of DNA/PEI complexes in 5% glucose in all four tumor quadrants with a syringe (100 Al in 60 s) or micropump (100 Al at 0.6 ml/h). A kdScientific 100 micropump was used for all micropump injections.
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
doi:10.1016/j.ymthe.2004.05.005
Tumor GFP expression analysis. Mice were euthanized by CO2 and tumors were removed with a scalpel. For whole-mount GFP microscopy, large (4 – 6 mm) pieces of fresh tumor were imaged using a GFP filter set at 10 with a Zeiss microscope. Frozen sections were generated by snapfreezing tumor tissue (suspended in OCT) in liquid nitrogen and sectioning (10-Am sections). Frozen sections were imaged for GFP using the same equipment as for whole mount. The percentage of GFP-positive cells within a frozen section was estimated by dividing GFP-positive cell area by total cell area. GFP immunohistochemistry was conducted by following the manufacturer’s protocol. In brief, sections were probed with antiGFP primary antibody (Clontech No. 8367-2), biotin-conjugated secondary antibody (Antibodies, Inc. No. 48-156-B), and avidin-conjugated-HRP (Antibodies, Inc. No. 49-801H) and then developed using a DAB substrate kit (Zymed No. 00-2014) and visualized using the same Zeiss microscope. NEO colony-forming assay and insertion cloning. Mice were euthanized with CO2 and tumors that received pT2SVNEO plasmid were removed with a scalpel. A single-cell suspension was generated by mincing the tumors with a razor blade into small pieces, followed by incubation in 0.25% trypsin, 0.1% EDTA (Gibco) at 37jC for 20 min with regular vortexing. The cells were then plated in tissue culture dishes and grown with (to determine stable NEO expression) or without (to determine plating efficiency) 400 Ag/ml G418. Stable NEO-resistant clones were isolated and used as a source of genomic DNA for PCR-mediated insertion cloning, which has already been described [32]. In vivo luciferase imaging and assays. Nude mice were anesthetized with avertin and injected with 150 Al of luciferin substrate (28.5 mg/ml; Xenogen). Five minutes after luciferin injection, mice were imaged for luciferase expression according to the manufacturer’s protocol using the Xenogen imaging system. Luciferase assays done on tumor cell lysates were preformed using a luciferase kit according to the manufacturer’s instructions (Promega).
ACKNOWLEDGMENTS We thank the Arnold and Mabel Beckman Foundation and all members of the Beckman Center for Transposon Research. This work was supported by NIHsponsored Cancer Biology Training Grant CA09138. We thank Paul Score for his promoter strength studies. RECEIVED FOR PUBLICATION MARCH 2, 2004; ACCEPTED MAY 2, 2004.
REFERENCES 1. Pulkkanen, K. J., et al. (2002). The combination of HSV-tk and endostatin gene therapy eradicates orthotopic human renal cell carcinomas in nude mice. Cancer Gene Ther. 9: 908 – 916. 2. Kuo, C. J., et al. (2001). Comparative evaluation of the antitumor activity of antiangiogenic proteins delivered by gene transfer. Proc. Natl. Acad. Sci. USA 98: 4605 – 4610. 3. Mullen, C. A., Anderson, L., Woods, K., Nishino, M., and Petropoulos, D. (1998). Ganciclovir chemoablation of herpes thymidine kinase suicide gene-modified tumors produces tumor necrosis and induces systemic immune responses. Hum. Gene Ther. 9: 2019 – 2030. 4. Dranoff, G. (2003). GM-CSF-secreting melanoma vaccines. Oncogene 22: 3188 – 3192. 5. Freeman, S. M., et al. (1993). The ‘‘bystander effect’’: tumor regression when a fraction of the tumor mass is genetically modified. Cancer Res. 53: 5274 – 5283. 6. Culver, K. W., Ram, Z., Wallbridge, S., Ishii, H., Oldfield, E. H., and Blaese, R. M. (1992). In vivo gene transfer with retroviral vector-producer cells for treatment of experimental brain tumors. Science 256: 1550 – 1552. 7. Dolivet, G., et al. (2002). In vivo growth inhibitory effect of iterative wild-type p53 gene transfer in human head and neck carcinoma xenografts using glucosylated polyethylenimine nonviral vector. Cancer Gene Ther. 9: 708 – 714. 8. Scappaticci, F. A., et al. (2001). Statin-AE: a novel angiostatin – endostatin fusion protein with enhanced antiangiogenic and antitumor activity. Angiogenesis 4: 263 – 268. 9. Schultz, J., Pavlovic, J., and Moelling, K. (2000). Immune modulation in cancer using DNA inoculation—antitumour effect of interleukin-12. Dev. Biol. 104: 109 – 114. 10. Li, D., et al. (2002). Interleukin 2 gene transfer prevents NKG2D suppression and enhances antitumor efficacy in combination with cisplatin for head and neck squamous cell cancer. Cancer Res. 62: 4023 – 4028. 11. Ebara, S., and Nasu, Y. (2003). [Suicide gene therapy for prostate cancer—update of current development]. Nippon Rinsho 61: 1257 – 1265.
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
ARTICLE
12. Rainov, N. G. (2000). A phase III clinical evaluation of herpes simplex virus type 1 thymidine kinase and ganciclovir gene therapy as an adjuvant to surgical resection and radiation in adults with previously untreated glioblastoma multiforme. Hum. Gene Ther. 11: 2389 – 2401. 13. Shand, N., et al. (1999). A phase 1 – 2 clinical trial of gene therapy for recurrent glioblastoma multiforme by tumor transduction with the herpes simplex thymidine kinase gene followed by ganciclovir. GLI328 European – Canadian Study Group. Hum. Gene Ther. 10: 2325 – 2335. 14. Lyons, R. M., et al. (1995). An improved retroviral vector encoding the herpes simplex virus thymidine kinase gene increases antitumor efficacy in vivo. Cancer Gene Ther. 2: 273 – 280. 15. Yang, Y., Li, Q., Ertl, H. C., and Wilson, J. M. (1995). Cellular and humoral immune responses to viral antigens create barriers to lung-directed gene therapy with recombinant adenoviruses. J. Virol. 69: 2004 – 2015. 16. Marshall, E. (1999). Gene therapy death prompts review of adenovirus vector. Science 286: 2244 – 2245. 17. Zhu, J., et al. (1999). Characterization of replication-competent adenovirus isolates from large-scale production of a recombinant adenoviral vector. Hum. Gene Ther. 10: 113 – 121. 18. Broeke, A. V., and Burny, A. (2003). Retroviral vector biosafety: lessons from sheep. J. Biomed. Biotechnol. 2003: 9 – 12. 19. Ritter, T., Lehmann, M., and Volk, H. D. (2002). Improvements in gene therapy: averting the immune response to adenoviral vectors. BioDrugs 16: 3 – 10. 20. Coll, J. L., Chollet, P., Brambilla, E., Desplanques, D., Behr, J. P., and Favrot, M. (1999). In vivo delivery to tumors of DNA complexed with linear polyethylenimine. Hum. Gene Ther. 10: 1659 – 1666. 21. Egilmez, N. K., Iwanuma, Y., and Bankert, R. B. (1996). Evaluation and optimization of different cationic liposome formulations for in vivo gene transfer. Biochem. Biophys. Res. Commun. 221: 169 – 173. 22. Lam, P. Y., and Breakefield, X. O. (2001). Potential of gene therapy for brain tumors. Hum. Mol. Genet. 10: 777 – 787. 23. Niranjan, A., et al. (2003). Treatment of rat gliosarcoma brain tumors by HSV-based multigene therapy combined with radiosurgery. Mol. Ther. 8: 530 – 542. 24. Kuriyama, N., Kuriyama, H., Julin, C. M., Lamborn, K. R., and Israel, M. A. (2001). Protease pretreatment increases the efficacy of adenovirus-mediated gene therapy for the treatment of an experimental glioblastoma model. Cancer Res. 61: 1805 – 1809. 25. Gondi, C. S., et al. (2003). Expression of antisense uPAR and antisense uPA from a bicistronic adenoviral construct inhibits glioma cell invasion, tumor growth, and angiogenesis. Oncogene 22: 5967 – 5975. 26. Kisker, O., et al. (2001). Continuous administration of endostatin by intraperitoneally implanted osmotic pump improves the efficacy and potency of therapy in a mouse xenograft tumor model. Cancer Res. 61: 7669 – 7674. 27. Shichiri, M., and Hirata, Y. (2001). Antiangiogenesis signals by endostatin. FASEB J. 15: 1044 – 1053. 28. Hajitou, A., et al. (2002). The antitumoral effect of endostatin and angiostatin is associated with a down-regulation of vascular endothelial growth factor expression in tumor cells. FASEB J. 16: 1802 – 1804. 29. Goldbrunner, R. H., Bernstein, J. J., and Tonn, J. C. (1999). Cell – extracellular matrix interaction in glioma invasion. Acta Neurochir. (Wien) 141: 295 – 305; discussion 304 – 295. 30. Izsvak, Z., Ivics, Z., and Plasterk, R. H. (2000). Sleeping Beauty, a wide host-range transposon vector for genetic transformation in vertebrates. J. Mol. Biol. 302: 93 – 102. 31. Cui, Z., Geurts, A. M., Liu, G., Kaufman, C. D., and Hackett, P. B. (2002). Structure – function analysis of the inverted terminal repeats of the sleeping beauty transposon. J. Mol. Biol. 318: 1221 – 1235. 32. Carlson, C. M., Dupuy, A. J., Fritz, S., Roberg-Perez, K. J., Fletcher, C. F., and Largaespada, D. A. (2003). Transposon mutagenesis of the mouse germline. Genetics 165: 243 – 256. 33. Vigdal, T. J., Kaufman, C. D., Izsvak, Z., Voytas, D. F., and Ivics, Z. (2002). Common physical properties of DNA affecting target site selection of sleeping beauty and other Tc1/mariner transposable elements. J. Mol. Biol. 323: 441 – 452. 34. Abdallah, B., Hassan, A., Benoist, C., Goula, D., Behr, J. P., and Demeneix, B. A. (1996). A powerful nonviral vector for in vivo gene transfer into the adult mammalian brain: polyethylenimine. Hum. Gene Ther. 7: 1947 – 1954. 35. Hirko, A. C., Buethe, D. D., Meyer, E. M., and Hughes, J. A. (2002). Plasmid delivery in the rat brain. Biosci. Rep. 22: 297 – 308. 36. Wu, X., Li, Y., Crise, B., and Burgess, S. M. (2003). Transcription start regions in the human genome are favored targets for MLV integration. Science 300: 1749 – 1751. 37. Yant, S. R., Meuse, L., Chiu, W., Ivics, Z., Izsvak, Z., and Kay, M. A. (2000). Somatic integration and long-term transgene expression in normal and haemophilic mice using a DNA transposon system. Nat. Genet. 25: 35 – 41. 38. Lampe, D. J., Grant, T. E., and Robertson, H. M. (1998). Factors affecting transposition of the Himar1 mariner transposon in vitro. Genetics 149: 179 – 187. 39. Geurts, A. M., et al. (2003). Gene transfer into genomes of human cells by the sleeping beauty transposon system. Mol. Ther. 8: 108 – 117. 40. Mikkelsen, J. G., Yant, S. R., Meuse, L., Huang, Z., Xu, H., and Kay, M. A. (2003). Helper-independent Sleeping Beauty transposon – transposase vectors for
267
ARTICLE
efficient nonviral gene delivery and persistent gene expression in vivo. Mol. Ther. 8: 654 – 665. 41. Zayed, H., Izsvak, Z., Walisko, O., and Ivics, Z. (2004). Development of hyperactive sleeping beauty transposon vectors by mutational analysis. Mol. Ther. 9: 292 – 304. 42. Ivics, Z., Hackett, P. B., Plasterk, R. H., and Izsvak, Z. (1997). Molecular reconstruction of Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell 91: 501 – 510. 43. De Bouard, S., et al. (2003). Antiangiogenic therapy against experimental glioblastoma
268
doi:10.1016/j.ymthe.2004.05.005
using genetically engineered cells producing interferon-alpha, angiostatin, or endostatin. Hum. Gene Ther. 14: 883 – 895. 44. Dilber, M. S., et al. (1999). Intercellular delivery of thymidine kinase prodrug activating enzyme by the herpes simplex virus protein, VP22. Gene Ther. 6: 12 – 21. 45. Tasciotti, E., Zoppe, M., and Giacca, M. (2003). Transcellular transfer of active HSV-1 thymidine kinase mediated by an 11-amino-acid peptide from HIV-1 Tat. Cancer Gene Ther. 10: 64 – 74. 46. Montini, E., et al. (2002). In vivo correction of murine tyrosinemia type I by DNAmediated transposition. Mol. Ther. 6: 759 – 769.
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
doi:10.1016/j.ymthe.2004.05.005
Integration and Long-Term Expression in Xenografted Human Glioblastoma Cells Using a Plasmid-Based Transposon System John R. Ohlfest, Paul D. Lobitz, Scott G. Perkinson, and David A. Largaespada* Department of Genetics, Cell Biology, and Development, University of Minnesota, and University of Minnesota Cancer Center, Arnold and Mabel Beckman Center for Transposon Research, Minneapolis, MN 55455, USA
Available online *To whom correspondence and reprint requests should be addressed at the Department of Genetics, Cell Biology, and Development, University of Minnesota, 6-160 Jackson Hall, 321 Church Street SE, Minneapolis, MN 55455, USA. E-mail: [email protected].
Available online 11 June 2004
Gene therapy has the potential to become an effective component of cancer treatment by transferring genes that cause immunomodulation or tumor cell death or that inhibit angiogenesis into tumor cells or tumor-associated stroma. Viral vectors have been the primary gene transfer vehicles used for intratumoral gene transfer to date. Plasmid-based vectors may be safer and more scalable than viral vectors. However, attempts at plasmid-based intratumoral gene transfer have been met with transient expression and poor gene transfer efficiency. Here we report integration and long-term expression of reporter genes in human glial tumors, growing in nude mice, using the Sleeping Beauty (SB) transposon system. A two-plasmid system was used, in which linear polyethylenimine was complexed with a GFP, NEO, or luciferase transposon plasmid and a SB transposase-expressing plasmid. SB-mediated transposition led to chromosomal integration of the NEO transgene in roughly 8% of tumor cells. SB-mediated insertions were cloned from the genomes of glial tumor cells to provide molecular proof of transposase-mediated integration. Luciferase studies showed that SB facilitated long-term expression of the transgene in glial tumors. SB-mediated intratumoral gene transfer is a novel, nonviral technique that could be used to augment conventional therapy for glioblastoma or other cancers. Key Words: glioblastoma, Sleeping Beauty, cancer gene therapy, transposon
INTRODUCTION Cancer gene therapy studies have demonstrated that tumors can be killed directly, starved of blood supply, or targeted for immune-mediated destruction by the expression of exogenous gene products within the tumor [1 – 10]. Viral vectors have proven to be an effective and widely used gene delivery system for delivering anti-tumor vectors into solid tumors [1,2,11 – 14]. However, immune reactions to viral vectors can inhibit gene transfer efficacy [15] and such reactions raise safety concerns, as demonstrated by an unfortunate clinical trial death [16]. In addition, although most viral vectors are designed to be replication incompetent, several instances of recombination events leading to replication-competent virus have been reported [17,18]. There is little doubt that viral vectors can be made less immunogenic [19] and that production challenges can be overcome. Nonetheless, there is
260
clearly considerable benefit in developing nonviral methods for gene therapy. Attempts at intratumoral gene transfer with plasmid DNA have yielded transient or steadily declining expression ([20,21], Fig. 5). One possible explanation for transient expression when using plasmids is the eventual degradation of episomal plasmid or a decreasing percentage of cells harboring the plasmid in highly mitotic tumors. Long-term expression of genes that inhibit angiogenesis, invoke tumor-specific immune responses, or cause tumor cell death is needed to enhance the therapeutic effect of nonviral cancer gene therapy vectors [reviewed in [22]]. Integrating vectors could lead to clonal expansion of transduced tumor cells, thereby increasing the anti-tumor effect. Brain tumors, such has glioblastoma multiforme (GBM), represent an ideal target for gene therapy because the primary tumor causes death and conventional therapies yield a median survival of only
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy 1525-0016/$30.00
ARTICLE
doi:10.1016/j.ymthe.2004.05.005
1 year [13]. Moreover, preclinical work in animal models has shown considerable efficacy for gene therapy on glioblastoma [6,14,23 – 25]. If antiangiogenic gene therapy is to halt tumor growth, expression of the inhibitor must be sustained to prevent angiogenesis and subsequent tumor regrowth [2,26]. Inhibitors of angiogenesis such as endostatin exert an anti-tumor affect, at least in part, by causing tumors to downregulate proangiogenic factors such as vascular endothelial growth factor [27,28]. Therefore, to be effective the inhibitor must be present long-term because the proangiogenic stimulus may always be present [2,26]. Likewise, suicide gene therapy should also benefit from long-term expression of the transgene. Herpes simplex virus type-1 thymidine kinase (HSV-tk) has become a widely used suicide gene because its expression is toxic only to dividing cells when the patient or animal is given nucleotide analogues, such as ganciclovir [6,12,22]. However, glioma cells are highly migratory and can become quiescent during migration, making them temporarily resistant to ganciclovir/HSV-tk therapy [22,29]. Thus, it is important to achieve long-term expression of HSV-tk to maximize the anti-tumor effects on migratory tumor cells that resume mitosis. Clinical trials for GBM that utilized HSV-tk demonstrated good safety but poor efficacy [12,13]. This was largely due to poor gene transfer rates [12,13] and highlights the need for improved gene delivery technology that is also scalable. In sum, cancer gene therapy is at a crossroads where experiments in animal models have shown great promise [1,5 – 7,14], but achieving high clinical efficacy has been difficult. Therefore there is a need to develop vector systems that are more effective, safer, and easier to produce and that support long-term expression of the transgene. Here we report transposition and long-term expression from plasmid DNA into the chromosomes of human glioma cells growing in nude mice. Sleeping Beauty (SB) is a ‘‘cut and paste’’ transposable element of the Tc1/ mariner superfamily and is active in both murine and human cells [30]. The SB transposase gene can be provided in trans (on a separate DNA molecule) or in cis (on the same DNA molecule) via plasmid. SB binds the inverted terminal repeat/direct repeat sequence of the transposon and catalyzes transposition into genomic DNA [30,31]. Preliminary studies on SB insertion site preference indicate that SB integrates at random into a TA dinucleotide [30,32,33]. To assess SB’s utility as an intratumoral gene transfer vector for GBM we studied the effects of different DNA doses, injection techniques, and doses of the SB transposase when delivered using polyethylenimine (PEI) as gene transfer reagent. PEI was used because it has been shown to transfect both tumors and murine glial cells effectively with no apparent morbidity or toxicity [20,34,35]. In sum, we show that SB transposon vectors can be delivered to xenografted GBM cells and support long-term expression in a significant percentage of cells.
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
RESULTS AND DISCUSSION Distribution of Injected DNA in Glioma Xenografts To determine the distribution of gene transfer within human glioma xenografts, we inoculated nude mice with U373 cells in the flank and subjected them to intratumoral injections of 10 Ag of a GFP plasmid (Fig. 1) complexed in PEI (N/P = 10). We injected tumors that were 6 – 8 mm in diameter. Rapid injection (100 Al in >5 s) of GFP/PEI complexes with a syringe resulted in poor gene transfer, in which most of the expression was observed on the periphery of the tumor (data not shown) and this has been reported before [20]. Slow injections with a syringe into all four tumor quadrants (100 Al in 60 s) was able to provide more evenly distributed GFP expression within the tumor. We analyzed GFP expression by fluorescence microscopy performed on wholemount tumors and cryosections and by immunohistochemistry done on cryosections (Fig. 2). After examining GFP expression distribution in cryosections (N = 7 tumors) we estimated that a range of 7 – 25% (average of 12.5%) of the tumor cells expressed GFP transiently. Tumors were taken at 48 h post injection. SB-Mediated Transposition Leads to Chromosomal Integration To quantitate the number of SB-mediated transposition events that occur intratumorally, we co-injected U373 tumors with 10 Ag of a NEO-resistance transposon plasmid and 500 ng of a CMV-driven SB transposase plasmid (Fig. 1). After a series of injections (twice over 1 week or
FIG. 1. Plasmid vectors. (A) Transposon vectors used. All vectors have an identical ColE1/AMPr backbone that is flanked by two IR/DR elements that serve as transposase binding sites and mark the end of the integrating unit. Simian virus 40 (SV40), cytomegalovirus (CMV), or chimeric chicken h-actin/ CMV enhancer (Caggs) promoter elements were used to drive expression of the transgenes. (B) Sleeping Beauty (SB) transposase vectors used. All vectors have an identical ColE1/AMPr backbone and express SB under the control of the CMV, ubiquitin (UB), and phosphoglycerate kinase (PGK) promoter elements. DDE-SB is a catalytically inactive transposase due to deletion of the catalytic domain [42] and has been used in SB gene transfer studies as a negative control [37,46].
261
ARTICLE
doi:10.1016/j.ymthe.2004.05.005
FIG. 2. GFP expression analysis on glioma tumor xenografts. Tumors were injected with PEI alone (A, B, C) or PEI/GFP plasmid complexes (N/P = 10) (D – I). (A – I) 48 – 72 h postinjection the tumors were removed with a scalpel. (A, D, G) Freshly removed tumor was placed in 1 PBS in 6-cm petri dishes and then visualized at 10 using a GFP filter set on a Zeiss microscope. (B, E, H) Tumors were removed with a scalpel, placed in OCT compound (Tissue-Tek), and flash frozen in liquid nitrogen. 10-Am frozen sections were thawed and visualized at 10 using a GFP filter on a Zeiss microscope. (C, F, I) Frozen sections were thawed and immunohistochemistry was performed using anti-GFP antibody. E and F are adjacent sections from the same tumor; note that GFP expression (E) colocalizes (arrow) with GFP immunostaining (F). The percentage of GFP-positive cells within a tumor was estimated by dividing GFP-positive cell area by total cell area for each section scored (two sections/tumor for seven tumors). These numbers were then averaged to yield an estimate of 12.5% GFP-positive cells within a tumor that express GFP transiently.
six times over 3 weeks) we sacrificed the mice, removed the tumors with a scalpel, and prepared a single-cell suspension of tumor cells. We then plated tumor cells in tissue culture dishes and selected them in 400 Ag/ml G418 or no drug to determine plating efficiency. Two slow injections of the NEO/SB/PEI complexes yielded an
262
average of 2% of the viable tumor cells that were able to form colonies under G418 selection (data not shown). Whereas six injections over 3 weeks boosted the transposition rate, an average of 8.3% of viable tumor cells formed colonies under G418 selection (Fig. 3A). We believe that colony formation is due to SB-mediated
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
doi:10.1016/j.ymthe.2004.05.005
ARTICLE
to integrate into genes but integrates randomly. In contrast, murine leukemia virus and HIV-1-based vectors have a strong preference to land near promoter/enhancer regions or within genes, respectively [36]. In the case of cancer gene therapy, one could reasonably argue that insertional mutagenesis is a not an issue because the patient already has cancer. Nevertheless, it must be taken into account that any vector delivered directly into a tumor could transfect normal cells surrounding the tumor or escape into the blood. Moreover, additional mutations in the tumor clone could accelerate disease. Therefore it is desirable to develop the safest and most effective vector systems to increase clinical efficacy and minimize nontarget effects for cancer gene therapy.
FIG. 3. Sleeping Beauty-mediated transposition in glioma tumor xenografts. (A) Tumors were co-injected with a NEO-resistance transposon plasmid and CMV-driven SB transposase (CMV-SB10, n = 6) or inactive SB transposase (CMV-DDE, n = 5) plasmid complexed in PEI (N/P = 10). After six injections (two/week for 3 weeks) the tumors were removed with a scalpel, minced, trypsinized, and vortexed to generate a single-cell suspension. Total cells were counted and plated in 400 Ag/ml G418 selection or without drug to determine viability/plating efficiency. After 3 weeks of selection colonies were counted and this number was used to determine plating efficiency (no drug) or percentage of viable cells that had stable NEO expression (400 Ag/ml G418). Error bar represents standard deviation. (B) Ten G418-resistant colonies were isolated and expanded from plates described in (A). Genomic DNA was isolated by phenol/chloroform extraction and used as template for splinkerette-mediated cloning PCR. Insertions that were cloned by splinkerette PCR were Blasted against the human genome using the ENSEMBL database.
transposition and not to random integration for two reasons: (1) when NEO is injected with a catalytically inactive SB transposase no colonies form under G418 selection (Fig. 3A) and (2) 10 insertions were cloned by linker-mediated PCR [32] and all of them appeared to be transposon integrations into a TA dinucleotide because no flanking plasmid sequence was recovered and instead insertion of the NEO transposon had occurred in glioma cell genomic DNA (Fig. 3B). SB Insertion Site Analysis Of the 10 insertions cloned, 3 landed in known or predicted genes with no apparent preference for chromosomal destination (Fig. 3B). Vigdal et al. cloned over 100 SB-mediated insertions in human cells and found that roughly 1/3 of the insertions were in genes [33], a percentage similar to that for randomly chosen TA dinucleotides [32,33]. Taken together, the preliminary data on SB insertion site preference indicate that SB does not prefer
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
Direct Injection with a Syringe Yielded Higher Gene Transfer Than a Micropump Coll et al. have previously demonstrated that a micropump is more effective at delivering DNA/PEI complexes to lung tumor xenografts than a syringe [20]. To optimize gene transfer to glioma xenografts we compared two different delivery methods and DNA doses. To determine early expression rates, we first injected U373 tumors with 10 Ag of a luciferase plasmid using either a micropump or a syringe. At 24 h post injection, we observed no significant difference in luciferase expression between micropump- or syringe-injected tumors (n = 3/group; P = 0.76). In both groups we observed roughly 1 104 RLU/mg protein/15 s (data not shown). We repeated this experiment by injecting 10 or 30 Ag of plasmid with a syringe or micropump and examined tumors for luciferase expression 72 h postinjection. At this later time point a syringe provided significantly higher gene transfer than a micropump at either DNA dose (10 Ag, P = 0.043; 30 Ag, P = 0.049, Fig. 4). These data are in clear contrast with those reported by Coll et al. in lung tumor xenografts and there are several possible explanations for this. Optimal transfection of glioma cells may differ from lung cell lines due to altered membrane permeability or other inherent cellular characteristics. In addition, the tumor stroma, microenvironment, and mouse strain could also affect transfection dynamics. As additional xenograft models are studied, including intracranial, it will become clear what method of DNA/PEI complex delivery is suitable for each individual cancer. Optimal intratumoral transfection conditions should be tested empirically by investigators when untested animal cancer models are being studied. To determine if gene transfer occurred in other tissues, we assayed several tissues for luciferase, including liver, kidney, and spleen. Generally all of these tissues had no luciferase expression, but one animal of six assayed had low-level (100 – 1000 RLU/mg protein) luciferase expression in the liver and kidney (data not shown). Thus intratumoral injection of DNA/PEI complexes primarily targets the tumor and this was verified by luciferase in vivo imaging (Fig. 5).
263
ARTICLE
doi:10.1016/j.ymthe.2004.05.005
FIG. 4. Dose and delivery optimization in glioma tumor xenografts. Tumors were injected with 10 or 30 Ag of Caggs-driven luciferase plasmid/PEI complexes (N/P = 10) with either a syringe or a micropump. 72 h after injection tumor lysates were analyzed for luciferase expression. Syringe-injected tumors had significantly higher expression at 10- (P = 0.043) or 30-Ag (P = 0.049) doses. Error bars represent standard deviation; P values derived from Student’s t test.
Long-Term Gene Transfer by Transposition in Glioma Xenografts We hypothesized that tumor cells that underwent transposition would clonally expand and lead to increasing luciferase signal over time, whereas tumors injected with a catalytically inactive form of SB transposase and luciferase transposon should eventually lose any detectable luciferase signal due to plasmid degradation or loss upon cell division. In addition, we wanted to rule out the possibility that the gene transfer we had observed in U373 xenografts was not an artifact of that particular cell line, but instead a general and reproducible result. Therefore we inoculated nude mice with another GBM-derived cell line, U87, in four places (Fig. 5). When the tumors were 6 – 8 mm in diameter we administered a single injection of luciferase transposon plasmid with or without active SB plasmid. We then monitored luciferase expression by in vivo luciferase imaging for the following 4 weeks. At 72 h postinjection, we observed relatively similar luciferase expression in all tumors as expected (Fig. 5). However, 2 weeks after injection tumors (N = 4 tumors/ group) that did not receive active SB had lost detectable luciferase expression. At 4 weeks, we observed a noticeable increase in luciferase expression in tumors that received CMV-driven SB at a 1:20 ratio of SB to luciferase transposon plasmid, respectively (Fig. 5). One explanation for increasing expression over time would be clonal expansion of cells that harbor transposon insertions. Regardless of whether clonal expansion is occurring, these data suggest that transposition facilitates long-term expression of the luciferase transgene. Taken together, these data demonstrate that integrating vectors such as SB are well suited to cancer gene
264
therapy because the anti-tumor effect of the transgene can be sustained and may even increase as the tumor cells divide. In the case of antiangiogenic vectors, longterm expression is highly desirable because decreasing endostatin expression has been reported to correlate with tumor regrowth when endostatin was delivered with a nonintegrating adenoviral vector [2]. A similar argument could be made for immunomodulation, in which it is desirable to maintain or increase expression of the transgene until the immune system kills the tumor or metastasis. Likewise, suicide gene therapy would also benefit because tumor cells that are not mitotic, and thus not susceptible to HSV-tk-mediated killing, could be killed by repeated ganciclovir injections if the cells resume mitosis. Optimizing SB-Mediated Transposition by Varying Promoter Strength and Plasmid Ratios Yant et al. have shown that too much transposase expression leads to less stable gene transfer in liver-directed studies in C57/Bl6 mice; they found that when using a strong promoter such as CMV, a ratio of 1:25 facilitates more efficient transposition than higher ratios [37]. This phenomenon is known as overexpression inhibition and has been observed in related transposable elements [38]. Moreover, a similar phenomenon was observed in cultured cell lines [39]. To improve the SB gene transfer system, it is desirable to have SB transposase expressed in every cell that has transposon substrate. Therefore two weaker promoters [40, Paul Score, unpublished], ubiquitin (UB) and phosphoglycerate kinase (PGK), were used to drive SB in trans at a 1:2.5 or 1:1 ratio of SB to transposon substrate, respectively (Fig. 5). Tumors that received these
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
doi:10.1016/j.ymthe.2004.05.005
ARTICLE
FIG. 5. SB facilitates long-term luciferase expression compared to plasmid alone. (A) Nude mice (n = 4) were given four tumors each by sc inoculation with 3.5 106 U87 cells. Tumors were then injected with 15 Ag of pT2C-Luciferase and 0.75 Ag CMV-DDE (1:20), 0.75 Ag CMV-SB (1:20), 6 Ag UB-SB (1:2.5), or 15 Ag PGK-SB (1:1) complexed with PEI (N/P = 10). The mice were then imaged for in vivo luciferase expression using the Xenogen imaging system at 72 h, 2 weeks, and 4 weeks. One representative mouse is shown at 72 h, 2 weeks, and 4 weeks post injection. (B) Graph of imaging data showing the average luciferase expression of each SB promoter group (n = 4 tumors/group). At 4 weeks post injection, CMVdriven SB facilitated significantly higher luciferase expression compared to CMV-DDE-SB (plasmid without active transposase; P < 0.001). UB- and PGKdriven SB also had significantly higher expression (P = 0.009) compared to CMV-DDE, albeit less than CMVdriven SB. Error bars represent standard deviation; P values derived from Student’s t test.
higher ratios of SB had persistent gene expression at 4 weeks, but CMV-driven SB at a 1:20 ratio resulted in the highest luciferase signal at 4 weeks (Fig. 5). We and others have also tested cis vectors in which UB or CMV is used to drive SB on the same plasmid with transposon [40]. Yet UB-driven SB in cis with a NEO transposon was not as effective in the tumor colony forming assay compared to trans delivery at a 1:1 molar ratio (data not shown) and CMV-driven cis vectors have also been reported to be inferior to trans delivery [37,40].
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
In contrast, a recent report indicates that cis vectors are superior to trans delivery when extremely weak promoters, such as uninduced metallothione (MT) or PGK, are used to drive SB expression [40]. PGK and MT are both very small promoters at 510 and 262 bp, respectively. It is possible that these promoters are ideal for cis plasmids because they result in smaller plasmid size, yet they also produce a low but optimal amount of SB protein [40]. However, these cis vectors were optimized in mouse liver using hydrodynamic gene delivery [40] and the studies
265
ARTICLE
described herein were conducted in human glioma xenografts using PEI gene delivery. Further study of cis and trans delivery of SB will facilitate optimal gene transfer in a variety of tissue types and gene delivery methods. Many factors that influence SB-mediated transposition have been identified, including the ratio of SB gene to transposon substrate in trans, orientation of SB expression cassette in cis vectors, SB promoter strength, cellular cofactors, and size of the transposon [39 – 41]. Nevertheless, CMV-driven SB at a 1:20 ratio of transposase to transposon was the most efficient combination we tested for delivering transposons to tumors with PEI. Implications for Cancer Treatment and Gene Therapy Here we described the development of a nonviral, yet integrative, intratumoral gene delivery technology for human glioblastoma cells. SB has the ability to mediate chromosomal integration, which in turn leads to longterm expression and possibly clonal expansion of mitotic tumor cells harboring transposon insertions (Figs. 3 and 5). Integration and long-term expression of HSV-tk transgenes should improve efficacy because not all cells within a tumor are mitotic at one time and are therefore temporarily resistant to HSV-tk/ganciclovir-mediated killing [22], but could be killed with ganciclovir if mitosis resumed. Accordingly, antiangiogenic gene therapy should benefit from long-term expression because the tumor’s proangiogenic stimulus may always be present and will require continuous suppression [2,26]. SB is a relatively safe integrating vector because there is no apparent preference to land in genes (Fig. 3B; [32,33]). In addition, because the SB system can be delivered as plasmid DNA, it has the potential to be more scalable for clinical use and may have a better biosafety profile than viral vectors. Although high clinical efficacy for gene therapy on glioblastoma has been difficult to achieve with other gene delivery methods [12,13], there is a large body of literature that shows good efficacy for gene therapy on glioblastoma in animal models [reviewed in 22]. SB has been shown to mediate transposition in a wide variety of human cell lines [30,42] and thus should be tested in other animal cancer models for safety and efficacy. Clinically, direct intratumoral injection of DNA/PEI complexes could be most useful when the glioma is inoperable due to anatomical location in the brain. However, the level of gene transfer that can be achieved by intratumoral injection in tumors larger than 6 – 8 mm in diameter (as described herein) remains to be ascertained. The use of improved transposon vectors [39 – 41] and hyperactive SB transposase mutants [39,41] will likely increase the percentage of tumor cells that can be stably integrated by intratumoral injection of DNA/PEI complexes. Alternatively, PEI/DNA complexes could potentially be delivered into the tumor resection cavity to prevent relapse caused by migratory glioma cells that commonly escape resection [22]. Nonviral gene transfer
266
doi:10.1016/j.ymthe.2004.05.005
to the mouse brain has been achieved by intraventricular injection of DNA/PEI complexes with no morbidity [34]. The idea of using transgenes that encode secreted proteins that act ‘‘cell nonautonomously’’ is appealing for cancer gene therapy because the gene transfer rate does not need to be high. De Bouard et al. have shown that when only 1% of an intracranial glioma xenograft expresses a secreted antiangiogenic molecule, animal survival is significantly increased [43]. We have conducted pilot experiments using PEI gene delivery, in larger glioma xenografts (up to 15 mm in diameter), and seen significant anti-tumor effects with secreted molecules, including endostatin (data not shown). HSV-tk has become widely studied due to the bystander effect; as little as 10% of the tumor transduced with HSV-tk has been sufficient to regress the tumor with ganciclovir [5]. Although 8% was the average amount of tumor cells that integrated the NEO transposon in this study, we observed one tumor with a 15% rate after six injections (Fig. 3A). Further optimization of DNA dose, injection technique, and injection frequency will likely boost the average transposition rate above the 10% minimum HSV-tk threshold that has been observed [5]. Moreover, several HSV-tk fusion proteins have been made that are secreted, significantly enhance bystander effect, and require less gene transfer than intracellular HSV-tk to ablate cells [44,45]. Therefore it is likely that the tumor transduction rate of 8% reported herein is high enough to achieve anti-tumor effects with secreted HSV-tk proteins. In addition to SB’s potential use for cancer therapy, SB also has been used to correct safely the phenotype of mouse models for hemophilia B and tyrosinemia in a transposition-dependent fashion [37,46]. Further study of SB in a variety of disease models will determine its potential for clinical use. Indeed, as viral and nonviral vector systems are improved and gene therapy strategies are refined, an increase in clinical efficacy will likely follow.
MATERIALS AND METHODS Plasmid DNA, PEI, and cell lines. Human glioblastoma-derived cell lines U373 and U87 were kindly provided by Dr. Daniel Vallera (University of Minnesota, Minneapolis, MN). All cell lines were grown in RPMI 1640 (Gibco) supplemented with 10% heat-inactivated FBS, 1% penicillin/ streptomycin, and 1% NEAA. Reporter plasmids pT2SVNEO, pT2-CMVGFP, and pT2C-Luc were generated in our laboratory using standard cloning methods. Linear polyethylenimine (aka jet PEI) was purchased from Polyplus Transfection (France). Tumor inoculations and intratumoral injections. Female BALB/c homozygous (nu/nu) mice, 6 – 8 weeks of age, were maintained in a pathogenfree environment throughout the experiments. Animals were inoculated by sc injection with GBM-derived cell lines U373 (5 106) and U87 (3.5 106). Three to four weeks after inoculation the tumors were 6 – 8 mm in diameter and were injected with 100 Al of DNA/PEI complexes in 5% glucose in all four tumor quadrants with a syringe (100 Al in 60 s) or micropump (100 Al at 0.6 ml/h). A kdScientific 100 micropump was used for all micropump injections.
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
doi:10.1016/j.ymthe.2004.05.005
Tumor GFP expression analysis. Mice were euthanized by CO2 and tumors were removed with a scalpel. For whole-mount GFP microscopy, large (4 – 6 mm) pieces of fresh tumor were imaged using a GFP filter set at 10 with a Zeiss microscope. Frozen sections were generated by snapfreezing tumor tissue (suspended in OCT) in liquid nitrogen and sectioning (10-Am sections). Frozen sections were imaged for GFP using the same equipment as for whole mount. The percentage of GFP-positive cells within a frozen section was estimated by dividing GFP-positive cell area by total cell area. GFP immunohistochemistry was conducted by following the manufacturer’s protocol. In brief, sections were probed with antiGFP primary antibody (Clontech No. 8367-2), biotin-conjugated secondary antibody (Antibodies, Inc. No. 48-156-B), and avidin-conjugated-HRP (Antibodies, Inc. No. 49-801H) and then developed using a DAB substrate kit (Zymed No. 00-2014) and visualized using the same Zeiss microscope. NEO colony-forming assay and insertion cloning. Mice were euthanized with CO2 and tumors that received pT2SVNEO plasmid were removed with a scalpel. A single-cell suspension was generated by mincing the tumors with a razor blade into small pieces, followed by incubation in 0.25% trypsin, 0.1% EDTA (Gibco) at 37jC for 20 min with regular vortexing. The cells were then plated in tissue culture dishes and grown with (to determine stable NEO expression) or without (to determine plating efficiency) 400 Ag/ml G418. Stable NEO-resistant clones were isolated and used as a source of genomic DNA for PCR-mediated insertion cloning, which has already been described [32]. In vivo luciferase imaging and assays. Nude mice were anesthetized with avertin and injected with 150 Al of luciferin substrate (28.5 mg/ml; Xenogen). Five minutes after luciferin injection, mice were imaged for luciferase expression according to the manufacturer’s protocol using the Xenogen imaging system. Luciferase assays done on tumor cell lysates were preformed using a luciferase kit according to the manufacturer’s instructions (Promega).
ACKNOWLEDGMENTS We thank the Arnold and Mabel Beckman Foundation and all members of the Beckman Center for Transposon Research. This work was supported by NIHsponsored Cancer Biology Training Grant CA09138. We thank Paul Score for his promoter strength studies. RECEIVED FOR PUBLICATION MARCH 2, 2004; ACCEPTED MAY 2, 2004.
REFERENCES 1. Pulkkanen, K. J., et al. (2002). The combination of HSV-tk and endostatin gene therapy eradicates orthotopic human renal cell carcinomas in nude mice. Cancer Gene Ther. 9: 908 – 916. 2. Kuo, C. J., et al. (2001). Comparative evaluation of the antitumor activity of antiangiogenic proteins delivered by gene transfer. Proc. Natl. Acad. Sci. USA 98: 4605 – 4610. 3. Mullen, C. A., Anderson, L., Woods, K., Nishino, M., and Petropoulos, D. (1998). Ganciclovir chemoablation of herpes thymidine kinase suicide gene-modified tumors produces tumor necrosis and induces systemic immune responses. Hum. Gene Ther. 9: 2019 – 2030. 4. Dranoff, G. (2003). GM-CSF-secreting melanoma vaccines. Oncogene 22: 3188 – 3192. 5. Freeman, S. M., et al. (1993). The ‘‘bystander effect’’: tumor regression when a fraction of the tumor mass is genetically modified. Cancer Res. 53: 5274 – 5283. 6. Culver, K. W., Ram, Z., Wallbridge, S., Ishii, H., Oldfield, E. H., and Blaese, R. M. (1992). In vivo gene transfer with retroviral vector-producer cells for treatment of experimental brain tumors. Science 256: 1550 – 1552. 7. Dolivet, G., et al. (2002). In vivo growth inhibitory effect of iterative wild-type p53 gene transfer in human head and neck carcinoma xenografts using glucosylated polyethylenimine nonviral vector. Cancer Gene Ther. 9: 708 – 714. 8. Scappaticci, F. A., et al. (2001). Statin-AE: a novel angiostatin – endostatin fusion protein with enhanced antiangiogenic and antitumor activity. Angiogenesis 4: 263 – 268. 9. Schultz, J., Pavlovic, J., and Moelling, K. (2000). Immune modulation in cancer using DNA inoculation—antitumour effect of interleukin-12. Dev. Biol. 104: 109 – 114. 10. Li, D., et al. (2002). Interleukin 2 gene transfer prevents NKG2D suppression and enhances antitumor efficacy in combination with cisplatin for head and neck squamous cell cancer. Cancer Res. 62: 4023 – 4028. 11. Ebara, S., and Nasu, Y. (2003). [Suicide gene therapy for prostate cancer—update of current development]. Nippon Rinsho 61: 1257 – 1265.
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy
ARTICLE
12. Rainov, N. G. (2000). A phase III clinical evaluation of herpes simplex virus type 1 thymidine kinase and ganciclovir gene therapy as an adjuvant to surgical resection and radiation in adults with previously untreated glioblastoma multiforme. Hum. Gene Ther. 11: 2389 – 2401. 13. Shand, N., et al. (1999). A phase 1 – 2 clinical trial of gene therapy for recurrent glioblastoma multiforme by tumor transduction with the herpes simplex thymidine kinase gene followed by ganciclovir. GLI328 European – Canadian Study Group. Hum. Gene Ther. 10: 2325 – 2335. 14. Lyons, R. M., et al. (1995). An improved retroviral vector encoding the herpes simplex virus thymidine kinase gene increases antitumor efficacy in vivo. Cancer Gene Ther. 2: 273 – 280. 15. Yang, Y., Li, Q., Ertl, H. C., and Wilson, J. M. (1995). Cellular and humoral immune responses to viral antigens create barriers to lung-directed gene therapy with recombinant adenoviruses. J. Virol. 69: 2004 – 2015. 16. Marshall, E. (1999). Gene therapy death prompts review of adenovirus vector. Science 286: 2244 – 2245. 17. Zhu, J., et al. (1999). Characterization of replication-competent adenovirus isolates from large-scale production of a recombinant adenoviral vector. Hum. Gene Ther. 10: 113 – 121. 18. Broeke, A. V., and Burny, A. (2003). Retroviral vector biosafety: lessons from sheep. J. Biomed. Biotechnol. 2003: 9 – 12. 19. Ritter, T., Lehmann, M., and Volk, H. D. (2002). Improvements in gene therapy: averting the immune response to adenoviral vectors. BioDrugs 16: 3 – 10. 20. Coll, J. L., Chollet, P., Brambilla, E., Desplanques, D., Behr, J. P., and Favrot, M. (1999). In vivo delivery to tumors of DNA complexed with linear polyethylenimine. Hum. Gene Ther. 10: 1659 – 1666. 21. Egilmez, N. K., Iwanuma, Y., and Bankert, R. B. (1996). Evaluation and optimization of different cationic liposome formulations for in vivo gene transfer. Biochem. Biophys. Res. Commun. 221: 169 – 173. 22. Lam, P. Y., and Breakefield, X. O. (2001). Potential of gene therapy for brain tumors. Hum. Mol. Genet. 10: 777 – 787. 23. Niranjan, A., et al. (2003). Treatment of rat gliosarcoma brain tumors by HSV-based multigene therapy combined with radiosurgery. Mol. Ther. 8: 530 – 542. 24. Kuriyama, N., Kuriyama, H., Julin, C. M., Lamborn, K. R., and Israel, M. A. (2001). Protease pretreatment increases the efficacy of adenovirus-mediated gene therapy for the treatment of an experimental glioblastoma model. Cancer Res. 61: 1805 – 1809. 25. Gondi, C. S., et al. (2003). Expression of antisense uPAR and antisense uPA from a bicistronic adenoviral construct inhibits glioma cell invasion, tumor growth, and angiogenesis. Oncogene 22: 5967 – 5975. 26. Kisker, O., et al. (2001). Continuous administration of endostatin by intraperitoneally implanted osmotic pump improves the efficacy and potency of therapy in a mouse xenograft tumor model. Cancer Res. 61: 7669 – 7674. 27. Shichiri, M., and Hirata, Y. (2001). Antiangiogenesis signals by endostatin. FASEB J. 15: 1044 – 1053. 28. Hajitou, A., et al. (2002). The antitumoral effect of endostatin and angiostatin is associated with a down-regulation of vascular endothelial growth factor expression in tumor cells. FASEB J. 16: 1802 – 1804. 29. Goldbrunner, R. H., Bernstein, J. J., and Tonn, J. C. (1999). Cell – extracellular matrix interaction in glioma invasion. Acta Neurochir. (Wien) 141: 295 – 305; discussion 304 – 295. 30. Izsvak, Z., Ivics, Z., and Plasterk, R. H. (2000). Sleeping Beauty, a wide host-range transposon vector for genetic transformation in vertebrates. J. Mol. Biol. 302: 93 – 102. 31. Cui, Z., Geurts, A. M., Liu, G., Kaufman, C. D., and Hackett, P. B. (2002). Structure – function analysis of the inverted terminal repeats of the sleeping beauty transposon. J. Mol. Biol. 318: 1221 – 1235. 32. Carlson, C. M., Dupuy, A. J., Fritz, S., Roberg-Perez, K. J., Fletcher, C. F., and Largaespada, D. A. (2003). Transposon mutagenesis of the mouse germline. Genetics 165: 243 – 256. 33. Vigdal, T. J., Kaufman, C. D., Izsvak, Z., Voytas, D. F., and Ivics, Z. (2002). Common physical properties of DNA affecting target site selection of sleeping beauty and other Tc1/mariner transposable elements. J. Mol. Biol. 323: 441 – 452. 34. Abdallah, B., Hassan, A., Benoist, C., Goula, D., Behr, J. P., and Demeneix, B. A. (1996). A powerful nonviral vector for in vivo gene transfer into the adult mammalian brain: polyethylenimine. Hum. Gene Ther. 7: 1947 – 1954. 35. Hirko, A. C., Buethe, D. D., Meyer, E. M., and Hughes, J. A. (2002). Plasmid delivery in the rat brain. Biosci. Rep. 22: 297 – 308. 36. Wu, X., Li, Y., Crise, B., and Burgess, S. M. (2003). Transcription start regions in the human genome are favored targets for MLV integration. Science 300: 1749 – 1751. 37. Yant, S. R., Meuse, L., Chiu, W., Ivics, Z., Izsvak, Z., and Kay, M. A. (2000). Somatic integration and long-term transgene expression in normal and haemophilic mice using a DNA transposon system. Nat. Genet. 25: 35 – 41. 38. Lampe, D. J., Grant, T. E., and Robertson, H. M. (1998). Factors affecting transposition of the Himar1 mariner transposon in vitro. Genetics 149: 179 – 187. 39. Geurts, A. M., et al. (2003). Gene transfer into genomes of human cells by the sleeping beauty transposon system. Mol. Ther. 8: 108 – 117. 40. Mikkelsen, J. G., Yant, S. R., Meuse, L., Huang, Z., Xu, H., and Kay, M. A. (2003). Helper-independent Sleeping Beauty transposon – transposase vectors for
267
ARTICLE
efficient nonviral gene delivery and persistent gene expression in vivo. Mol. Ther. 8: 654 – 665. 41. Zayed, H., Izsvak, Z., Walisko, O., and Ivics, Z. (2004). Development of hyperactive sleeping beauty transposon vectors by mutational analysis. Mol. Ther. 9: 292 – 304. 42. Ivics, Z., Hackett, P. B., Plasterk, R. H., and Izsvak, Z. (1997). Molecular reconstruction of Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell 91: 501 – 510. 43. De Bouard, S., et al. (2003). Antiangiogenic therapy against experimental glioblastoma
268
doi:10.1016/j.ymthe.2004.05.005
using genetically engineered cells producing interferon-alpha, angiostatin, or endostatin. Hum. Gene Ther. 14: 883 – 895. 44. Dilber, M. S., et al. (1999). Intercellular delivery of thymidine kinase prodrug activating enzyme by the herpes simplex virus protein, VP22. Gene Ther. 6: 12 – 21. 45. Tasciotti, E., Zoppe, M., and Giacca, M. (2003). Transcellular transfer of active HSV-1 thymidine kinase mediated by an 11-amino-acid peptide from HIV-1 Tat. Cancer Gene Ther. 10: 64 – 74. 46. Montini, E., et al. (2002). In vivo correction of murine tyrosinemia type I by DNAmediated transposition. Mol. Ther. 6: 759 – 769.
MOLECULAR THERAPY Vol. 10, No. 2, August 2004 Copyright B The American Society of Gene Therapy