Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

One of the most promising nonviral methods of in vivo gene delivery involves the use of cationic liposomes complexed to plasmid expression vectors (Felgner et al., 1997). The resulting aggregates typically have a net positive charge and, therefore, have the propensity to interact with negatively charged surfaces including cell membranes (Smith et al., 1993, Wong et al., 1998). Although cationic liposome/plasmid DNA aggregates have been shown to mediate transfection of selected tissues after i.v. administration in mice (Hyde et al., 1993; Zhu et al., 1993; Thierry et al., 1995), difficulties in defining systemically viable lipid-based DNA formulations have hampered development of systemic applications (Bally et al., 1998). For this reason many applications have focused on development of cationic liposomes as carriers of DNA to be administered within defined regions. Examples would include lung (Brigham et al., 1989; Stribling et al., 1992; Alton et al., 1993; Wheeler et al., 1996), nasal epithelium (Caplen et al., 1995), arterial endothelium (Nabel et al., 1989, 1992; Stephan et al., 1996), spleen (Philip et al., 1993), brain (Ono et al., 1990; Jiao et al., 1992), as well as tumors (Nabel et al., 1990, 1993; Plautz et al., 1993; Parker et al., 1996). Importantly, regional administration has proven to be a viable option for the treatment of several diseases including cancer and cystic fibrosis, and is already in clinical trials (Caplen et al., 1995; Silver et al., 1996; Chang et al., 1997; Rubin et al., 1997; Stopeck et al., 1997). These preclinical and clinical studies, where liposome/DNA formulations are given regionally, demonstrate the feasibility, safety, and therapeutic potential of such gene therapy approaches.

It is evident that cationic liposome/DNA aggregates (referred to hereinafter as DNA lipoplexes; Felgner et al., 1997) are suitable and effective where regional dosing is a viable route of administration. However, the factors that govern optimal transgene expression after regional administration are not well understood. There is no evidence suggesting, for example, that transgene expression levels correlate with DNA delivery. In a previous study evaluating DNA delivery to melanoma cells in vitro, we obtained data that suggested that gene expression levels do not correlate well with DNA delivery (Reimer et al., 1997). These in vitro data suggested that delivery of intact DNA may not be a barrier to transfection and that the processing of the liposome/DNA complex is crucial to the expression of the DNA.

A careful assessment of parameters affecting chloramphenicol acetyltransferase (CAT) expression in mouse lung after intratracheal administration of DNA lipoplexes has also been completed (Meyer et al., 1995). These investigators demonstrated that in comparison to free plasmid DNA, lipid-based carriers improved DNA delivery within the lung airways, both in terms of enhanced DNA stability and DNA retention in lung after intratracheal administration. Transfection efficiencies in the lung were, however, comparable for the cationic liposome/DNA aggregates and free plasmid DNA. The dilemma that such data create concerns distinguishing the role of plasmid expression delivery from that of transgene expression. The latter can not occur without the former, but successful delivery does not guarantee efficient transgene expression.

The aim of this study was to assess the role of lipid-mediated DNA delivery to tumors after regional administration of free and cationic DNA lipoplexes. For this purpose, we used a mouse tumor model consisting of B16/BL6 melanoma cells grown i.p.. Yang and Huang (1996) have demonstrated gene transfer into B16/BL6 tumors after intratumoral injection; however, we believe that it is important to test transfection activity in a solid tumor model in the absence of direct injection. Such an approach allows us to investigate selectivity of the lipoplex formulations, penetration into the tumor, and tumor cell delivery. In this initial study, cationic DNA lipoplexes were injected i.p. into mice bearing i.p. solid tumors. Subsequently, the tumors were assayed for: 1) expression of the CAT reporter gene, 2) liposomal lipid and DNA delivery to the tumor, and 3) the integrity tumor-associated plasmid DNA. To our surprise, DNA lipoplexes were efficiently bound to B16/BL6 tumors 2 h after injection with as much as 18% of the injected DNA dose associated with tumors.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Dioleoyldimethylammonium chloride (DODAC) was synthesized and supplied by Steven Ansell of Inex Pharmaceuticals, Inc. (Vancouver, BC, Canada). Dioleoylphosphatidylethanolamine (DOPE) was purchased from Avanti Polar Lipids (Alabaster, AL). ¹⁴C-Labeled cholesterylhexadecyl ether (CHE) and ¹⁴C-DOPE were obtained from Amersham (Oakville, ON, Canada). CAT and lactose were obtained from Sigma (St. Louis, MO). All other chemicals used in this study were reagent grade.

The plasmid pInex CAT v2.0, containing the Escherichia coli CAT gene under the control of the CMV promoter was constructed and provided by Roger Graham of Inex Pharmaceuticals. Briefly, pInex CAT was constructed by transferring the 0.8kb HindIII fragment of pCMV4CAT plasmid (generously provided by K. Brigham, Vanderbilt University, Nashville, TN) into the NotI site of pCMV (obtained from Clontech, Palo Alto, CA) after addition of NotI linkers. The resulting plasmid contains the Immediate Early Cytomegalovirus promoter/enhancer, SV40 intron A, the CAT gene, and the SV40 polyadenylation signal. Large scale plasmid DNA preparations were made by propagating plasmid in DH5 E. coli and purified according to the methods outlined in the Qiagen Plasmid Purification Kit (Qiagen, Chatsworth, CA). The nucleic acid concentration was measured by UV absorption at 260 nm and verified by electrophoresis on 0.8% agarose gels.

Radiolabeled plasmid was purified from JM101 species-harboring pInex CAT v2.0 that were metabolically labeled with 1.0 mCi of [³H]-thymidine-5'-triphosphate (DuPont-NEN, Boston, MA) in 100-ml supplemented M9 minimal media. The plasmid was purified using standard techniques as described above. The specific activity of [³H]-pInex CAT v2.0 was approximately 200,000 dpm/µg.

The murine B16/BL6 melanoma cell line was obtained from the National Cancer Institute Tumor Repository 12-105-54 (Bethesda, MD) and was maintained in Eagle's minimal essential medium supplemented with 5% fetal bovine serum at 37°C in 5% CO₂ with no antibiotics.

Preparation of Cationic Liposome/DNA Aggregates. DODAC/DOPE liposomes (50:50 mol %) were prepared according the method of Hope et al. (1985). Lipids were dissolved in chloroform (20 mg/ml) and radiolabeled at a specific activity of 1 to 2 µCi/50 mg with ¹⁴C-CHE as a nonmetabolizable and nonexchangeable liposomal marker (Scherphof et al., 1987). For tracking radiolabeled lipid after injection, ¹⁴C-DOPE was used as the liposomal marker. The lipids were dried to a thin film under a stream of nitrogen gas and vacuum dried at >76 cm Hg for at least 4 h. The films were hydrated in filter-sterilized 300 mM lactose and passed 10 times at room temperature through an extruder (Lipex Biomembranes, Vancouver, BC, Canada) containing three stacked 80-nm polycarbonate membranes (Poretics Corp., Livermore, CA). The lipid concentration of the resulting liposomes was determined by liquid scintillation (Packard TR 1900 Scintillation Counter) using ¹⁴C-CHE or ¹⁴C-DOPE as a marker. The size of the liposomes was measured by quasielastic light scattering using a Nicomp Submicron Particle Sizer (model 270, Pacific Scientific, Santa Barbara, CA) operating at a wavelength of 632.8 nm. All liposomes had a mean diameter of 100 to 140 nm by Gaussian analysis and were stored at 4°C until use.

Transfection. Adult female C57BL/6J mice (7-8 weeks old) were used for all experiments. All procedures were performed in accordance with Canadian Council of Animal Care Guidelines for the Care and Use of Laboratory Animals. Mice (four per group) were injected with B16/BL6 murine tumor cells i.p. (1 × 10⁵ cells) in Hanks' balanced salt solution in a volume of 0.5 ml. The tumors were allowed to grow for 7, 10, or 13 days. The mice were injected i.p. with DNA lipoplexes at the indicated liposome and DNA concentrations. After 24, 48, or 72 h, the tumors were harvested, weighed, and stored at 70°C until assayed for CAT activity. All samples were analyzed within 1 week after isolation.

Assay of CAT Activity. Tumors were thawed in the presence of buffer [15 mM Tris-HCl (pH = 8.0), 60 mM KCl, 15 mM NaCl, 5 mM EDTA (pH = 8.0), 0.15 mM spermine, 1.0 mM dithiothreitol, 35 µg/ml phenylmethylsulfonyl fluoride, 0.5 µg/ml leupeptin, 0.5 µg/ml aprotinin, 5 µM paraoxon] to make a final concentration of 10% (w/v). Tumors were homogenized on ice using a Polytron homogenizer (Brinkman Instruments Canada, Mississauga, ON). Samples (100 µl) were transferred to 1.5-ml microcentrifuge tubes and subjected to three cycles of freeze/thaw consisting of immersion in liquid nitrogen followed by thawing in a 37°C water bath. Samples were centrifuged at 10,000g in an Eppendorf microcentrifuge for 10 min at room temperature; the supernatants were recovered and heat-inactivated for 15 min at 65°C. Samples were centrifuged for 10 min at 10,000g, and 55 µl of the supernatant from each sample was evaluated for CAT activity. To each sample, 50 µl (250,000 dpm) of ¹⁴C-chloramphenicol (NEN-DuPont, Boston, MA) and 25 µl N-butyryl CoA (5 mg/ml) was added and incubated at 37°C for 2 h. Mixed xylenes (Aldrich Chemical Co., Milwaukee, WI; 300 µl) were added to each tube and vortexed vigorously for 30 s, followed by centrifugation for 3 min at 10,000 rpm in an Eppendorf microcentrifuge at room temperature. The upper phase was transferred to a fresh microcentrifuge tube and 750 µl buffer [15 mM TRIS-HCl (pH = 8.0), 60 mM KCl, 15 mM NaCl, 5 mM EDTA (pH 8.0)] were added to each sample, which were then vortexed and recentrifuged. For each sample, 100 µl of the resulting upper phase was sampled, 5 ml of Picofluor scintillant (Packard Instrument Co., Meriden, CT) was added, and radioactivity (¹⁴C) was determined in a Canberra-Packard scintillation counter (1900 TR Tri Carb). CAT units were determined by comparison to a standard curve generated for each experiment. Values were converted to and expressed as mU CAT/g wet weight. Each CAT assay was performed in triplicate and expressed ± S.E.

Quantification of Plasmid DNA after i.p. Administration of the Lipoplex Formulation. Mice bearing 7-day B16/BL6 i.p. tumors were injected with DODAC/DOPE liposome/pInex CAT v2.0 lipoplexes or free plasmid and at 30, 60, and 120 min after injection mice were sacrificed and the peritoneal cavity was lavaged with 3 ml of Hanks' balanced salt solution. Blood was obtained by cardiac puncture and the lavage fluid and blood were immediately analyzed for the presence of plasmid DNA. Tumor, spleen, pancreas, and liver were excised, weighed, and stored at 20°C until further analysis. Plasmid DNA associated with tumor tissue was quantified by two methods. First, tumors (and spleen, liver, pancreas, blood, and lavage fluid) were evaluated for the presence of ³H after administration of ³H-pInex CAT v2.0. Briefly, blood (100 µl), lavage fluid (1 ml), whole tumor, spleen, and pancreas were incubated with 0.5 ml Solvable (NEN-DuPont) at 50°C for 18 h. Liver was homogenized in water to make a 25% homogenate and 200 µl was added to 0.5 ml Solvable and incubated as described. The samples were subsequently decolorized by the addition of H₂O₂ and HCl, scintillation fluid was added, and the samples were counted for ³H radioactivity. The second method involved evaluating the tumors for the presence of plasmid DNA using dot blot analysis. Freshly collected tumor tissue was homogenized for 20 s on ice in the buffer used for CAT assay (100 mg tumor/ml buffer). One hundred microliters homogenate was removed from each sample and dissolved in DNAzol (Gibco-BRL, Burlington, ON, Canada) at room temperature for 30 min. Cold 95% ethanol was added to each tube (1 ml) and the DNA precipitated for 1 h at room temperature. DNA pellets were recovered by centrifugation at 10,000 rpm for 10 min at room temperature, rinsed with 70% ethanol, and dissolved in 100 µl TE buffer [10 mM Tris-HCl (pH = 8.0), 1 mM EDTA (pH = 8.0)]. Purified DNA was applied to a nitrocellulose membrane using a dot blot apparatus and the blots were hybridized using ³²P random prime-labeled pInex CAT v2.0 as described by Sambrook et al. (1989). Plasmid DNA associated with tumors 2 and 24 h after administration was quantified using a PhosphoImager (Molecular Dynamics, Sunnyvale, CA). pInex CAT v2.0-specific DNA values were standardized using known pInex CAT v2.0 standards. Four animals were evaluated for the 2- and 24-h time points with three replications per assay. Data are expressed as means ± S.E.

Assessment of Intact DNA. Plasmid DNA associated with tumors 2 and 24 h after injection of lipoplexes was isolated along with genomic DNA using standard SDS/proteinase K techniques (Sambrook et al., 1989). DNA was extracted by phenol/chloroform and precipitated with 2.5 volumes 95% ethanol. The DNA was resuspended in TE buffer [10 mM Tris-HCl (pH = 8.0), 1 mM EDTA] and evaluated for concentration using spectrophotometric readings at A₂₆₀. DNA samples (5 µg) were loaded onto a 1% agarose gel and subjected to electrophoresis at 5V in TBE buffer (89 mM TRIS-Borate, 2 mM EDTA) for 18 h. The DNA was transferred to nitrocellulose membrane and hybridized with ³²P random prime-labeled pInex CAT v2.0 following the hybridization method of Sambrook et al. (1989). The hybridized blot was exposed and the image digitized using a PhosphoImager.

Statistical Analysis. Quantitative data generated for CAT activity were statistically evaluated using ANOVA (Statistical Software Inc., Tulsa, OK).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

CAT Expression in Melanoma Tumors. C57BL/6J mice were injected i.p. with B16/BL6 melanoma cells and 7, 10, or 13 days after cell inoculation 25 µg (500 µl) of free pInex CAT v2.0 plasmid or DODAC/DOPE DNA lipoplexes (10 nmol lipid/1 µg DNA) were injected i.p. Two days after administration of the plasmid DNA, the tumors were removed and the level of gene expression was determined by measuring CAT activity (Fig. 1). After administration of free plasmid DNA, the level of CAT activity measured in homogenized tumors ranged from 13.2 to 30.4 mU/g wet weight. In contrast, the level of CAT activity in tumors that were grown for 7 days and treated with DODAC/DOPE DNA lipoplexes were approximately 500 mU/g wet weight. Even though the time point selected for assessing CAT gene expression was comparable (48 h), the level of expression decreased when the tumors progressed for 10 and 13 days before plasmid administration. The B16/BL6 tumors progressed rapidly and the tumors isolated 10 and 13 days after cell inoculation were much larger (>200 mg) then those isolated 7 days after cell injection (<100 mg). These data have also been plotted as a function of CAT activity per tumor (Fig. 1, inset) and suggest that transfection is optimal when the plasmid is administered in animals where tumor growth has just initiated (7 days after cell administration). It is important to note that measurable levels of CAT expression were observed after administration of free plasmid and that the increased levels of transfection activity observed after administration of the lipid-based plasmid delivery system are only significant for tumors recovered 7 and 10 days after cell inoculation.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   CAT activity expressed 48 h after transfection of B16/BL6 melanoma tumors grown i.p. for 7, 10, or 13 days with DODAC/DOPE/pInex CAT v2.0 (filled columns) or free pInex CAT v2.0 DNA (unfilled columns). Animals were administered 25 µg plasmid DNA in a volume of 500 µl of 300 mM lactose. The ratio of lipid to DNA was 10 nmol lipid/1 µg DNA. The levels of CAT activity were determined based on known standards and expressed as mU/g wet weight (see Materials and Methods). Each column represents the mean ± S.E. of at least n = 4. Inset shows data expressed as CAT activity (mU) per tumor.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   CAT activity expressed in B16/BL6 i.p tumors grown for 7 days and evaluated 24, 48, or 72 h after i.p. injection of DODAC/DOPE/pInex CAT v2.0 (filled columns) or free pInex CAT v2.0 DNA (unfilled columns). Animals were administered 25 µg plasmid DNA in a volume of 500 µl 300 mM lactose. The ratio of lipid-to-DNA was 10 nmol lipid/1 µg DNA. The levels of CAT activity were determined based on known standards and expressed as mU/g wet weight (see Materials and Methods). Each column represents the mean ± S.E. of at least n = 3.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Compilation of all results obtained after i.p. administration of 25 µg plasmid DNA complexed to DODAC/DOPE liposomes (10 nmol lipid/1 µg DNA). CAT expression was measured 24 h after administration and the individual data points (milliunit CAT activity per gram tumor, see Materials and Methods) were plotted as function of tumor size.

pInex CAT v2.0 Plasmid DNA Dose Response. Figure 4 illustrates data obtained when increasing amounts of DNA were injected i.p. into animals bearing 7-day tumors and CAT expression was determined 24 h later. In this experiment, 25, 50, 75, and 100 µg of DNA were complexed with DODAC/DOPE liposomes such that the lipid/DNA ratio remained at 10 nmol lipid/1 µg DNA. When 50 and 75 µg plasmid DNA were administered, CAT activities of 1334 ± 286 and 1674 ± 124 mU/g wet weight were obtained, respectively. These CAT activities were 2- to 3-fold higher than those obtained using 25 µg plasmid DNA. Increasing the amount of injected DNA to 100 µg yielded CAT activities of 1082 ± 335 mU/g wet weight, thus, the activity observed appears to saturate at doses above 75 µg DNA/mouse. Injection of free plasmid alone yielded CAT expression levels ranging from 39 ± 11 mU/g wet weight at the lowest DNA dose to 72 ± 9 mU/g wet weight at the highest dose (data not shown), again demonstrating that under the conditions used, transfection is enhanced through the use of a lipid-based delivery system.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Effect of DNA dose on CAT activity after transfection of B16/BL6 i.p. tumors with DODAC/DOPE/pInex CAT v2.0. Animals were injected with various amounts of lipid and DNA (ratio 10 nmol lipid/µg DNA) in 500 µl 300 mM lactose. The level of CAT activity was determined as in Fig. 1. Each column represents mean ± S.E. of at least n = 3.

Plasmid DNA and Liposomal Lipid Biodistribution after i.p. Administration. Having established conditions where cationic DNA lipoplexes appear to enhance transgene expression (relative to free plasmid) in B16/BL6 tumors growing in the peritoneal cavity, an evaluation of the biodistribution of plasmid DNA and associated lipid after i.p. administration was completed. Radiolabeled [³H]-plasmid DNA was used to measure DNA delivery and ¹⁴C-DOPE was used as a liposomal lipid marker for detection of the DODAC/DOPE liposomes. The level of DNA and liposomal lipid was evaluated in lavage, blood, and tumors 30, 60, and 120 min after i.p. administration of DODAC/DOPE/pInex CAT v2.0 aggregates. For comparison, DNA and lipid levels in these tissues were also measured after administration of free plasmid DNA and DODAC/DOPE liposomes without bound DNA. The data, presented in Table 1, were derived after analysis of the indicated samples as described in Materials and Methods. After subtraction of a background correction factor (<50 dpm), measurements of <100 dpm were assumed to be below the detection limits and values were assigned as not detectable.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Quantification of DNA (µg) and lipid (nmol) 2 h after i.p. administration of mice with B16/BL6 i.p. tumors with DODAC/DOPE/pInex CAT v2.0 (filled columns) and free DNA (unfilled columns, top) or free liposomes (unfilled columns, bottom) in different tissues. DNA and lipid associated with tumor (tum), spleen (spl), pancreas (pan), liver (liv), and blood (bl) were expressed per gram tissue ± S.E. of at least n = 4. DNA and lipid associated with lavage fluid (lav) was expressed per total lavage (3 ml). Plasmid was measured using [3H]-plasmid and liposomal lipid was measured using [14C]-DOPE as a marker (see Materials and Methods).

DNA Quantification by Dot Blot Analysis and DNA Integrity. It is important to recognize that DNA levels in association with the isolated tumors and other tissues/fluids have been quantified using [³H]-labeled plasmid, and it is possible that the label may not be associated with intact/functional DNA or may represent [³H]-metabolites that have be released after metabolism/degradation of the plasmid. This problem is likely to be a more significant issue when time points beyond 2 h are evaluated. To address this issue, we used a second technique (dot blot analysis) to quantify the amount of DNA, and Southern analysis to assess DNA integrity. The dot blot assay is designed to detect the presence of plasmid DNA sequences and was applied to samples containing genomic and plasmid DNA that was isolated and blotted onto nitrocellulose membranes as described in Materials and Methods. DNA was quantified from dot blots using ³²P random prime-labeled plasmid hybridization and the results analyzed using a phosphoimager. Tumor-bearing animals were injected i.p. with free DNA and DNA lipoplexes and the level of plasmid DNA in isolated tumors was determined 2 and 24 h after administration. The results, shown in Fig. 6, indicated that after injection of free plasmid, DNA levels of 5.86 ± 1.42 µg and 6.33 ± 2.26 µg per g tumor were obtained at 2 and 24 h, respectively. The values obtained at 2 h are consistent with results observed using ³H-plasmid DNA as the label (see Table 1). The amount of DNA associated with the tumors 2 and 24 h after injection of DNA lipoplexes was 88.72 ± 50.58 and 61.29 ± 53.10 µg/g tissue, respectively. The error associated with the dot blot analysis was greater than that observed using [³H]-plasmid; however, levels measured in association with isolated tumors 2 h after i.p. administration were comparable. Data obtained using [³H]-plasmid DNA as a label were 46.01 ± 6.57 µg/g tissue compared with 88.72 ± 50.58 µg/g tissue determined using dot blot analysis. These data substantiate the fact that a large amount of DNA was associated with B16/BL6 tumors after i.p. injection of DNA lipoplexes and, in addition, suggest that the DNA remains associated with the tumors for at least 24 h after administration.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Quantification by dot blot analysis of DNA associated with 7-day B16/BL6 i.p. tumors 2 and 24 h after administration of DODAC/DOPE/pInex CAT v2.0 (filled columns) and free DNA (unfilled columns). The dot blot procedure is described in Materials and Methods. Each column represents mean ± S.E., n = 3.

View larger version (63K): [in this window] [in a new window]   Fig. 7.   Southern blot of plasmid DNA extracted from tumors after i.p. administration of DODAC/DOPE/pInex CAT v2.0 and free DNA to animals with 7-day B16/BL6 tumors. DNA was extracted from tumors and evaluated by Southern analysis 2 (lanes 1 and 2) and 24 h (lanes 3 and 4) after transfection with free DNA (lanes 1 and 3) and DNA lipoplexes (lanes 2 and 4). P = control plasmid DNA, C = tumor DNA without transfection, SC = supercoiled plasmid, LIN = linear plasmid, and OC = open circled plasmid.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Many investigators believe that lipid-based DNA transfer systems are well suited for use in gene therapy for the treatment of cancer. To date, such treatment strategies have focused on either ex vivo use of DNA transfer systems or regional administration. Focusing on the latter, intratumoral injection of cationic DNA lipoplexes has proven to be effective in augmenting gene transfer, such that transgene expression levels are sufficient to elicit a therapeutic response. One of the first and most advanced clinical studies involves transfer of the HLA-B7 gene to human melanoma cells through the use of cationic liposomes, complexed to a plasmid expression vector, that are directly injected into accessible lesions (Nabel et al., 1993; Silver et al., 1996; Rubin et al., 1997; Stopeck et al., 1997). Characterization of gene expression after direct intratumoral injection has, therefore, been studied extensively. Results obtained using the i.p. B16/BL6 tumor as a model to assess regional administration of DNA lipoplex formulations should be considered in comparison with data obtained after direct intratumor injection as well as the factors that may influence gene delivery and gene expression. This discussion will focus specifically on these issues.

It is of critical importance when considering the development of drug carrier technology that the carrier imparts some beneficial effects, typically expressed as an improvement in the therapeutic properties of the associated drug. As greater emphasis is placed on plasmid expression vector delivery in vivo it is becoming apparent that free plasmid, in the absence of a delivery system, is capable of accessing certain cells. Studies such as those reported by Meyer et al. (1995) and Yang and Huang (1996) have demonstrated (in lung and tumor tissue, respectively) higher levels of gene expression after administration of free DNA. Yang and Huang (1996) showed that DNA formulated with cationic liposomes and injected intratumorally actually inhibited gene expression in vivo. In contrast, Egilmez et al. (1996) demonstrated that the human interleukin-2 gene could be successfully delivered intratumorally into established human tumor xenografts in SCID mice through the use of cationic liposome-mediated DNA delivery, and that expression levels were enhanced through use of lipid-based DNA delivery systems. It is interesting to note that the results presented in Fig. 1 suggest that, depending on the progression of the i.p. tumor, cationic DNA lipoplexes can enhance gene expression at least 10-fold in comparison with free plasmid after i.p. injection. These data also suggest, however, that conditions can be defined where the transgene expression is greater for free plasmid in comparison with lipoplex formulations.

This is the first report, to our knowledge, that quantifies the delivery of DNA in tumors as well as other surrounding tissues after i.p. administration of cationic DNA lipoplexes. Quite remarkably, 2 h after administration, approximately 18% of the injected dose of DNA was associated with tumors (Table 1). Moreover, this plasmid DNA was associated with the tumors for at least 24 h after administration (Fig. 7). DNA association was confirmed by Southern analysis (Fig. 7) and these data suggest that the tumor-associated DNA is intact 2 h as well as 24 h after i.p. administration. Whether this DNA has been taken up by tumor cells has yet to be evaluated. In addition, the mechanism of internalization of cationic-liposome/DNA aggregrates by cells in vivo is not clearly understood. We have shown using the B16/BL6 i.p. tumor model that complexing DNA to liposomes can enhance delivery to the tumor.

Although i.p. injection of free DNA does result in measurable levels of transgene expression in the B16/BL6 tumors, the levels are significantly lower than those that can be obtained after injection of the DNA lipoplex formulation (Fig. 1). The results shown in Fig. 7 clearly suggest that reduced levels of expression must be partly due to degradation of the free plasmid. We have shown in this report that when cationic DNA lipoplexes are administered i.p., a proportion of the DNA that is specifically associated with tumors is intact for up to 24 h. This likely contributes significantly to the improved transfection levels observed when using the DNA lipoplexes in comparison to free DNA. It should be noted that we have no direct evidence that transfection of B16/BL6 tumors involves gene transfer to the tumor cell or whether transfection is restricted to some host-derived cell population.

It is unclear whether the difference in Results obtained after intratumoral injection compared to data reported here are a consequence of different expression vectors, different formulation attributes, or differences in the route of injection. The approach developed in this study, although also relying on regional administration, does not use direct injection of tumors. The model is based on transfecting a small tumor, attached to organs or the peritonium, after injection of cationic DNA lipoplexes into the peritoneal cavity. This model has allowed us to measure gene transfection of the tumor as well as the efficiency of plasmid DNA delivery. For this reason, we believe that this model should help address whether selected formulation attributes (surface charge, size, lipid composition, DNA/lipid ratio) modulate gene delivery and whether such changes in delivery influence transgene expression. This information is fundamental if we are to gain a better understanding of the factors that govern efficient transfection in vivo and design better formulations to achieve more efficient gene transfer.

Regardless, it is important to recognize that the i.p. tumor model described here is likely to be of value only in the development of delivery systems to be administered via the i.p. route. In particular, formulation attributes that are required for optimal delivery and expression after i.p. administration will be significantly different then those developed for i.v., i.m., intratracheal, or intratumor routes of administration. The DODAC/DOPE DNA lipoplex formulation used in the studies reported here, for example, do not transfect B16/BL6 tumors (grown i.p. or s.c.) after i.v. inoculation. These formulations are not systemically viable because of residual surface charge and large size (>300 nm; Bally et al., 1998). For these reasons the lipoplex formulations are rapidly eliminated after i.v. administration and tumor delivery of the associated plasmid expression vector is below detectable limits.

We believe that the mechanisms of gene delivery must be elucidated if effective in vivo liposome-mediated DNA delivery systems are to be developed. In particular, it is important to understand the phenomenon of how the lipid and DNA dissociate from one another and how this may affect the ability of the DNA to be processed after binding and cell internalization as well as access to transcriptional and translational machinery. It has been suggested that the strength of the interactions between the lipid and the DNA will affect DNA stability as well as the process by which the DNA is transported to the nucleus (Zabner et al., 1995; Szoka et al., 1996; Reimer et al., 1997; Harvie et al., 1998). For this reason we also must consider the possibility that much of the tumor-associated DNA observed after i.p. administration of cationic DNA lipoplexes may not be in a form that can facilitate transfection. Analogous to our research with liposomal encapsulated drugs (Bally et al., 1998), if the lipid-based delivery system does not incorporate features that encourage release (dissociation) of the therapeutically active agent, then the carrier will effectively inhibit the agent's biological activity. We believe that DNA dissociation will be one of the most critical attributes to consider when designing effective cationic liposome DNA delivery systems.