Sequential Injection of Cationic Liposome and Plasmid DNA ...

doi:10.1006/mthe.2001.0311, available online at http://www.idealibrary.com on IDEAL

ARTICLE

Sequential Injection of Cationic Liposome and Plasmid DNA Effectively Transfects the Lung with Minimal Inflammatory Toxicity Yadi Tan,*,† Feng Liu,* Zhiyu Li,*,1 Song Li,* and Leaf Huang*,†,2 *Center for Pharmacogenetics, School of Pharmacy, and †Department of Pharmacology, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania 15213 Received for publication December 6, 2000; accepted in revised form March 15, 2001; published online April 16, 2001

A major hurdle to lipoplex-based systemic gene delivery is acute inflammatory toxicity. In this study, a safe, simple, and effective alternative to lipoplex administration, specifically, sequential injection of cationic liposome and plasmid DNA, was evaluated. When plasmid DNA was injected into the tail vein of mice 2–5 min after the injection of cationic liposomes, 50 – 80% lower levels of proinflammatory cytokines, including TNF-␣, IL-12, and IFN-␥, were observed compared to lipoplex injection. The sequential injection technique yielded a two- to fivefold higher level of transgene expression in the lung and was more effective in repeated dosing than lipoplex. Other types of lipoplex-associated toxicities, such as neutropenia, lymphopenia, thrombocytopenia, and complement depletion, were also significantly reduced with sequential injection. The reduction in cytokine release was observed with several different liposome formulations and appeared to be a general phenomenon. Key Words: sequential injection; cationic liposome; gene transfer; lipoplex; systemic; inflammatory toxicity; cytokine induction; plasmid DNA.

INTRODUCTION Cationic liposome-mediated gene transfer efficiently delivers genes to pulmonary endothelium and holds promise for treating pulmonary metastasis and other diseases (1–3). However, its application is hampered by associated side effects. The major toxicity is the rapid induction of large quantities of proinflammatory cytokines, such as TNF-␣,3 IL-6, IL-12, and IFN-␥ (4 –7), which is due to the stimulation of immune cells by the unmethylated CpG motifs in the plasmid DNA (pDNA) (6 –12). The systemic administration of cationic liposome–pDNA complex also causes other side effects, including transient lymphopenia, thrombocytopenia, complement activation/depletion, and the elevation of hepatic alanine aminotransferase (ALT) and aspartate aminotransferase (AST) (13). The proinflammatory cytokines induced by cationic lipo-

1 Current address: Department of Biological Chemistry, Johns Hopkins University School of Medicine, Baltimore, MD 21204. 2 To whom correspondence and reprint requests should be addressed at 633 Salk Hall, Center for Pharmacogenetics, School of Pharmacy, University of Pittsburgh, Pittsburgh, PA 15213. Fax: (412) 648-1664. E-mail: [email protected]. 3 Abbreviations used: LPD, liposome–protamine–DNA; pDNA, plasmid DNA; TNF, tumor necrosis factor; IFN, interferon; CX, complex; SQ, sequential; DEX, dexamethasone.

MOLECULAR THERAPY Vol. 3, No. 5, May 2001, Part 1 of 2 Parts Copyright © The American Society of Gene Therapy 1525-0016/01 $35.00

some–pDNA complexes may prove to be beneficial in genetic immunization (14, 15) and in cancer therapy (16, 17). However, the side effects of cationic liposome-based vectors need to be well controlled when these vectors are used for the treatment of nonmalignant diseases. Various approaches have been taken to reduce this inflammatory response. These include the elimination of CpG motifs in the pDNA (18), targeting of the DNA to endothelium (19), the use of PCR-amplified DNA fragments instead of full-length pDNA (20), and the use of immunosuppressant (12) or endocytosis blocking agents (18). These approaches are somewhat complicated and have had limited success. In the search for an improved means to overcome the inflammatory toxicity of cationic liposome-mediated gene transfer, we have revisited a previously reported gene transfer method. This method involves injecting the cationic liposome and pDNA sequentially, rather than as a complex (21). This method was initially designed to investigate the mechanism of pulmonary transfection by pDNA. Here we present data showing that this protocol is more efficient in gene transfer than lipoplex and has an excellent safety profile. It is our hope to bring this safe and efficient gene transfer method into broad use and to accelerate the progress of gene therapy studies utilizing cationic liposome-based vectors.

673

ARTICLE MATERIALS

AND

METHODS

Chemicals. 1,3-Dioleoyl-3-trimethylammonium propane (DOTAP) and dioleoyl phosphatidylethanolamine (DOPE) were purchased from Avanti Lipids, Inc. (Alabaster, AL). N-(2,3-Dioleoyloxy) propyl)-N,N,N-trimethyl ammonium chloride (DOTMA) was a gift from Dr. Dexi Liu (University of Pittsburgh, Pittsburgh, PA). 3-␤[N-(N⬘,N⬘-Dimethylaminoethane)carbamoyl]cholesterol (DC-Chol) was synthesized as described (22). Cholesterol and dexamethasone 21-phosphate disodium salt was obtained from Sigma (St. Louis, MO). Protamine sulfate-USP was purchased from Eli Lilly (Indianapolis, IN). Linear polyethylenimine (PEI) of 22 kDa was obtained from Euromedex (Souffleweyersheim, France). SuperFect was purchased from Qiagen (Valencia, CA). Animals. CD-1 mice (4-week, female, 16 –18 g) were from Charles River Laboratories (Wilmington, MA). The animals were kept at the University of Pittsburgh Central Animal Facility. All experiments were conducted under protocols approved by the Institutional Animal Care and Use Committee. Plasmid. Plasmid pNGVL3-Luc (pDNA) contains the cDNA of firefly luciferase inserted into vector pNGVL3 obtained from the National Gene Vector Laboratory (http://huffamoose.im.med.umich.edu/vcore/). The plasmid was amplified in the DH5␣ strain of Escherichia coli, isolated by alkaline lysis, and purified with a Qiagen Endo-Free Giga-Prep kit (Qiagen). The plasmid was suspended in water and stored at ⫺80°C until use. Preparation of liposomes. Liposomes containing DOTAP and cholesterol in a 1:1 molar ratio were prepared as previously described (6). The lipid mixture in chloroform was placed under a stream of nitrogen to evaporate the solvent until a thin lipid film formed in the flask. It was further vacuum-desiccated for 2 h and then hydrated in water to a final concentration of 10 mg DOTAP/ml. The lipid suspension was briefly sonicated and then sequentially extruded through polycarbonate membranes of pore size of 1.0, 0.4, and 0.1 ␮m while the temperature of the solution was kept at 50°C. The resulting liposomes were primarily unilamellar vesicles. Liposomes containing DOTMA and cholesterol in a 1:1 molar ratio were prepared the same way as DOTAP/Chol liposomes. DC-Chol/DOPE liposomes were produced by microfluidization at a 3:2 molar ratio of DC-Chol to DOPE at a concentration of 2 ␮mol/ml (1.2 mg/ml total lipid) as described previously (23). Preparation of lipo(poly)plex and polyplex. For lipoplex and polyplex, cationic liposomes/polymer and pNGVL3-Luc were separately diluted with PBS to equal volumes. Lipoplex/polyplex was prepared by adding DNA solution dropwise into the liposome/polymer solution in a 50-ml conical tube while the tube was being gently swirled. The resulting lipoplex suspension was translucent and the polyplex suspension was clear. The dilution was calculated such that a final volume of 200 ␮l lipoplex/ polyplex was to be injected into one animal. For lipopolyplex liposome–protamine–DNA (LPD), all the components were diluted with dextrose solution to reach a final concentration of 5% dextrose. Plasmid DNA solution was added dropwise to an equal volume of LP (DOTAP– cholesterol liposomes and protamine) mixture at a ratio of 1200 nmol DOTAP:60 ␮g protamine:100 ␮g DNA (24). The volume of injection was also 200 ␮l per animal. All complexes were incubated at room temperature for at least 10 min before injection. Lipoplex vs sequential injection of DOTAP/cholesterol liposomes and pDNA. If not otherwise indicated, the amount of DOTAP/cholesterol liposomes and pDNA was 900 nmol and 25 ␮g, respectively. All the dilutions were made in PBS (pH 7.4). For lipoplex, the mice received an injection of a 200-␮l suspension of lipoplex via the tail vein. For sequential injection, if not indicated otherwise, the animals first received an injection of 100 ␮l liposome and then 100 ␮l plasmid DNA into the tail vein. There were three mice per group. Blood collection and cytokine assay. At indicated time points following injection, mice were bled via their tail artery after a brief warming. The blood was allowed to clot on ice for at least 4 h and then centrifuged at 3000g for 20 min at 4°C. Serum was collected and the concentrations of cytokines (TNF-␣, IL-1␤, IL-12, IFN-␥) were determined with a mouse cytokine immunoassay kit (R&D Systems, Minneapolis, MN).

674

Evaluation of transgene expression in mouse lung. The plasmid pNGVL3Luc was used as a reporter gene in all experiments. Unless otherwise specified, the mice were sacrificed 8 h after vector injection. Each mouse lung was collected and placed in 1 ml of ice-cold lysis buffer (0.05% Triton X-100, 2 mM EDTA, and 0.1 M Tris, pH 7.8) and homogenized with a tissue tearer (BioSpec Products, Bartlesville, OK) for 30 s at high speed. The homogenates were then centrifuged at 14,000g for 10 min at 4°C. Ten microliters of the supernatant was analyzed with the luciferase assay system (Promega, Madison, WI) using an automated LB953 luminometer (Berthod, Bad Wildbad, Germany). The protein content of the supernatant was measured with the Bio-Rad Protein Assay System (Bio-Rad, Hercules, CA). Luciferase activity was expressed as relative light units (RLU) per milligram protein. Hematological and serological analysis. Mouse blood samples were collected at indicated time points and sent to Antech Diagnostics (Farmingdale, NY) for analysis of hematological and serological parameters. These samples were collected in EDTA KE microtubes (Sarstedt, Newton, NC) for cell and platelet counts. Samples were also collected in serum gel microtubes (Sarstedt) for enzyme ALT and AST measurements. Complement assay. Serum complement C3 activity was measured with the complement assay system from Sigma. The assay was modified to a 96-well format according to the method described previously (25, 26). Antibody-sensitized sheep erythrocytes were washed and diluted with Gelatin Veronal Buffer to a concentration of 1 ⫻ 108 cells/ml. Serum samples from mice were serially (2n) diluted with Gelatin Veronal Buffer to a range of 1:40 and 1:640. One hundred microliters of erythrocyte suspension, 50 ␮l of each serum dilution, and 5 ␮l of fivefold-diluted C3-deficient serum were added to each well of a round-bottomed 96-well plate. The plate was tape sealed and incubated at 37°C for 1 h with vigorous shaking. The plate was centrifuged at 2000g for 5 min to pellet unlysed erythrocytes. One hundred microliters of the supernatant was transferred to a flat-bottomed 96-well plate and the released hemoglobin in the supernatant was measured at 415 nm with a plate reader. The H50 (serum dilution necessary to lyse 50% of the erythrocytes in a given assay) was calculated from the data. Then, H50 values were converted to percentage values relative to the H50 of untreated mice to indicate the degree of complement depletion. Comparison of complex vs sequential injection with different formulations. All solutions were made in PBS except for LPD nanoparticles for which the dilution was made with dextrose solution. The injection volume was 200 ␮l for complex and 100 ␮l ⫹ 100 ␮l for sequential injections. All sequential injections were performed with 2-min intervals. The dose of pDNA was 25 ␮g in all groups with the exception of the DOTMA liposome group due to the toxicity of the formulation. The dose per animal for each vector was as follows: DOTAP/cholesterol liposome, 900 nmol DOTAP; DOTMA/cholesterol liposome, 600 nmol DOTMA with 20 ␮g pDNA; LPD, 300 nmol DOTAP/cholesterol liposome and 15 ␮g protamine; DC-Chol/DOPE liposome, 200 nmol total lipids; PEI, 25 ␮g PEI; SuperFect, 300 ␮g SuperFect (commercially available at a concentration of 3 mg/ml). There were three mice per group. Statistical analysis. Data were analyzed by the two-tailed unpaired Student’s t test using the PRISM software program (GraphPad Software, Inc.). Data were considered statistically significant if P ⬍ 0.05 (* in the figures). ** indicates P ⬍ 0.01 and *** indicates P ⬍ 0.001 in the figures.

RESULTS Sequential Injection of Cationic Liposomes and pDNA Induced Significantly Lower Levels of Proinflammatory Cytokines and Yielded Higher Transgene Expression DOTAP/cholesterol cationic liposomes (900 nmol) and pDNA (25 ␮g) were injected into mice via the tail vein sequentially (2 min apart) or as a lipoplex. The doses of liposome and pDNA were chosen as described by Song et MOLECULAR THERAPY Vol. 3, No. 5, May 2001, Part 1 of 2 Parts Copyright © The American Society of Gene Therapy

ARTICLE FIG. 1. The effect of sequential injection on cytokine induction (A) and transgene expression (B). Each group of mice (n ⫽ 3) received tail vein injection of 900 nmol DOTAP/cholesterol cationic liposomes (L) and/or 25 ␮g pDNA (D) carrying a luciferase gene. The L and D components were administered either in the form of an LD complex or sequentially injected 2 min apart with L first (L 3 D) or D first (D 3 L). Blood was collected 2 h after injection and serum TNF-␣ and IL-12 were measured by ELISA. Luciferase expression in mouse lung was assayed 8 h after injection. *P ⬍ 0.05 vs LD group.

al. (21) for optimal transgene expression. Cytokine assay was performed with blood collected at 2 h, when the concentration of TNF-␣, the major source of toxicity, reached its peak (6, 12). The levels of both cytokines TNF-␣ and IL-12 were significantly lower in L 3 D sequential injection than in lipoplex (LD) injection (81 ⫾ 11% reduction for TNF-␣ and 93 ⫾ 1% reduction for IL-12; P ⬍ 0.05) (Fig. 1A). Luciferase activity in the lung was fivefold higher in the L 3 D group than in the LD group (Fig. 1B). In contrast, in D 3 L sequential injection, both cytokines’ levels and transgene expression were at similar levels in comparison with lipoplex injection. Naked DNA (D) produced background levels of transfection. Since decreased levels of proinflammatory cytokines and improvement of transgene expression were observed only in L 3 D and not in D 3 L injection, “sequential injection” in subsequent studies shall refer to L 3 D injection if not otherwise indicated.

Time Window for an Effective Gene Transfer with Sequential Injection The interval between the injections of cationic liposomes and pDNA critically affects the level of transgene expression (21). This may be due to the kinetics of the cationic liposomes in the blood, which limits their capacity to interact with subsequently injected pDNA. In our MOLECULAR THERAPY Vol. 3, No. 5, May 2001, Part 1 of 2 Parts Copyright © The American Society of Gene Therapy

experiments, when pDNA was injected 2 to 5 min after DOTAP/cholesterol liposome administration, significantly less TNF-␣ (P ⬍ 0.01) and IL-12 (P ⬍ 0.05) production was observed, as well as a severalfold increase in transgene expression compared to LD lipoplex (Fig. 2). As the interval of injection increased, the cytokine levels were still significantly lower, but the transgene expression in the lung gradually decreased with time (Fig. 2). Nevertheless, transgene expression at 10- or 20-min sequential injection intervals was at a level comparable to that of lipoplex injection. In subsequent experiments, all sequential injections were performed at 2-min intervals.

Dose–Response Curve for Sequential vs Lipoplex Systemic Gene Transfer Next, we examined the effect of sequential injection over a range of pDNA doses from 5 to 50 ␮g (Fig. 3). The cationic liposome dose was proportionally adjusted to keep the ratio of liposome to pDNA constant. The dose– response curves for TNF-␣ and IL-12 levels in the sequential injection group were essentially unchanged. In contrast, LD complex induced much higher TNF-␣ and IL-12 levels at 25- and 50-␮g doses. For transgene expression in the lung, the sequential injection group had either similar

675

ARTICLE lated as the indicator of total exposure to TNF-␣ by the mice, SQ injection had an overall 75– 80% lower TNF-␣ level. For IL-12 and IFN-␥, which peaked at 6 h, SQ injection induced much lower levels of cytokines at each time point (range from 33 to 75% reduction). Overall reduction of the two cytokines, as calculated by AUC, was approximately 51% for IL-12 and 60% for IFN-␥.

Kinetics of Lung Transgene Expression in Sequential vs Lipoplex Gene Transfer Figure 5 shows the kinetics of luciferase gene expression in the mouse lung over a 72-h period with sequential and lipoplex injection. Transgene expression from sequential

FIG. 2. The effect of injection interval on TNF-␣ and IL-12 induction and transgene expression. 25 ␮g pDNA was injected via the mouse tail vein at different time points following injection of 900 nmol DOTAP/cholesterol liposomes, or pDNA and liposomes were injected as a complex (LD). Blood was collected at 2 h after injection and serum TNF-␣ and IL-12 were measured by ELISA. Luciferase activity in mouse lung was assayed 8 h after. n ⫽ 3 per group. *P ⬍ 0.05; **P ⬍ 0.01 vs LD group.

or higher transgene activity than the LD group at each dose.

Kinetics of Cytokine Induction in Sequential vs Lipoplex Systemic Gene Transfer In order to fully characterize the effect of sequential injection, the kinetics of three important proinflammatory cytokines (TNF-␣, IL-12, and IFN-␥) in mice blood were examined over a period of 24 h following intravenous injection of 900 nmol DOTAP/cholesterol liposomes and 25 ␮g pDNA. These injections were either LD complex (CX) or sequential (SQ) (Fig. 4). Compared to CX injection, SQ injection had a dramatically lower TNF-␣ level during the 0- to 4-h time period. When the area under the curve (AUC) (time ⫻ concentration) was calcu-

676

FIG. 3. Dose–response curve for cytokine TNF-␣ and IL-12 induction and transgene expression following complex (LD) or sequential (L 3 D) injection. Different doses of pDNA and DOTAP/cholesterol liposomes were injected into groups of mice (n ⫽ 3) via the tail vein either as an LD complex or 2 min apart sequentially. The liposome dose was 180, 450, 900, or 900 nmol and DNA dose was 5, 12.5, 25, or 50 ␮g, respectively. Blood was collected 2 h after injection and the serum TNF-␣ and IL-12 levels were measured by ELISA. Luciferase activity in mouse lung was assayed 8 h after. *P ⬍ 0.05; ***P ⬍ 0.001 SQ vs LD group. MOLECULAR THERAPY Vol. 3, No. 5, May 2001, Part 1 of 2 Parts Copyright © The American Society of Gene Therapy

ARTICLE FIG. 4. Kinetics of cytokines TNF-␣, IL-12, and IFN-␥ induced by the sequential injection (SQ) of DOTAP/cholesterol liposomes and pDNA vs complex (CX) injection. The groups of mice (n ⫽ 3) received injection of 900 nmol liposomes and 25 ␮g pDNA. At each indicated time point, blood was collected from two groups of mice and cytokines were measured with ELISA. *P ⬍ 0.05; **P ⬍ 0.01 SQ vs CX group.

injection was consistently higher (at least three- to fivefold) than that of lipoplex injection at all points examined.

Sequential Injection Had a Shorter Refractory Period The high level of proinflammatory cytokines induced by lipoplex not only reduces the level of transgene expression but also correlates with a refractory period during which repeated dosing yields a significantly lower transfection (6, 12). Figure 6 demonstrates our examination of MOLECULAR THERAPY Vol. 3, No. 5, May 2001, Part 1 of 2 Parts Copyright © The American Society of Gene Therapy

the refractory period of sequential vs lipoplex transfection. Mice were given injections at days 0 and 3, 5, or 7 and luciferase expression in the lung was assayed 8 h after the final injection. When both injections were sequential (SQ, SQ), the second injection at day 3 yielded approximately 40% of the level of the first SQ transfection. The transfection level was then fully restored when a second injection was performed at day 5. In comparison, when both injections were lipoplex (CX, CX), the second injection at day 3 and day 5 yielded only 13–15% of the level

677

ARTICLE

FIG. 5. Kinetics of transgene expression in mouse lung following injection of liposome and pDNA in a complex (CX) or sequentially 2 min apart (SQ). Groups of mice (n ⫽ 3) received injection of 900 nmol liposomes and 25 ␮g pDNA. Luciferase activity in the lung was measured at indicated time points. *P ⬍ 0.05; **P ⬍ 0.01 SQ vs CX group.

of the first CX transfection (5% of the level of the first SQ transfection) and the level of transgene expression recovered only at day 7. When the first transfection was lipoplex and second was sequential (CX, SQ), the transgene expression from the second injection was consistently higher than that of the (CX, CX) group, but lower than that of the (SQ, SQ) group until the refractory period ended.

Dexamethasone (DEX) Further Reduced the Levels of Proinflammatory Cytokines in Sequential Injection of Cationic Liposomes and pDNA DEX, a potent immunosuppressant, was shown to be a safe and effective agent for inhibiting the induction of proinflammatory cytokines by lipo(poly)plex (12). As shown in Fig. 7, DEX decreased the cytokine levels induced by sequential injection of cationic liposomes and pDNA. For TNF-␣, a single dose of DEX (100 ␮g/mouse, ip) prior to sequential injection decreased the 2-h peak level by 75% to 90 ⫾ 53 pg/ml. This is more than 90% reduction compared to the TNF-␣ level in lipoplex injection (⬃2000 pg/ml) (Fig. 1). For IL-12, DEX continually reduced levels at each time point from 4 to 8 h by 70%. It reduced AUC by 97.2% in comparison with sequential injection and 98.6% compared to the lipoplex curve (Fig. 4). Interestingly, IFN-␥ level in sequential injection was not changed by DEX. The kinetics for IFN-␥ in SQ and SQ ⫹ DEX were essentially identical (data not shown).

Other Toxicities Associated with Lipoplex Were Less Severe in Sequential Injection Although the major concern was the induction of high levels of proinflammatory cytokines, systemic administration of lipoplex also causes hematological and serological changes, including lymphopenia, thrombocytopenia, complement depletion, and elevations in serum enzyme

678

activity such as ALT and AST (13). These toxicities are part of the inflammatory response and may be related to the proinflammatory cytokines. Cytokines such as TNF-␣ can induce changes in hepatic enzyme activities and induce the expression of adhesion molecules which could be responsible for the transit reduction in lymphocytes. Therefore the hematological and serological toxicities associated with sequential injection should be reduced in comparison to lipoplex injection, and the results in Table 1 support this concept. Platelet count, complement activity, hepatic enzyme activity, and WBC count were examined after injection of gene vectors or various controls (DOTAP/cholesterol liposome, PBS, or no injection) (Table 1). For platelet count, complement activity, and hepatic enzyme activity, SQ caused significantly less change than CX compared to naı¨ve mice, while liposome alone or PBS had no effect. The platelet count was 45% of the control level in the CX and 70% in the SQ group (P ⬍ 0.01 CX vs SQ). Complement activity was 12.8 ⫾ 10.1% in the CX group and 42.7 ⫾ 12.6% in the SQ group in comparison to the level of 100 ⫾ 6.8% in naı¨ve mice (P ⬍ 0.05 CX vs SQ). The elevation of ALT activity was more than 10-fold in CX and approximately 3-fold for SQ injection. The AST level was increased 3-fold in CX but only 2-fold in SQ injection (P ⬍ 0.05 CX vs SQ for ALT and AST). The white blood cell (WBC) count demonstrated a different toxicity. There were similar degrees of reduction in the numbers of lymphocytes and monocytes, in the CX, the SQ, and the liposome groups, compared to PBS and untreated groups. This result indicates that the reduction in WBC counts was primarily the effect of the cationic liposome.

Efficient Gene Transfer and Reduction of Inflammatory Toxicity Was Achieved with Sequential Injection of the Components of Lipoplex and Lipopolyplex, but Not of Polyplex For Figs. 1 through 7, DOTAP/cholesterol liposome was used for the sequential and lipoplex transfection. We reason that the advantage of sequential transfection may not be limited to this formulation alone. Sequential injection of the components of several different formulations, including two lipoplexes, one lipopolyplex, and two polyplex formulations, were performed and the cytokine induction and transgene expression were measured (Table 2). The doses for the four in vivo formulations were chosen from the observations of others (19, 21, 24) or determined in our experiments for DOTMA. For the two formulations DC-Chol and SuperFect, which were usually employed only for in vitro transfections, we fixed the amount of pDNA at 25 ␮g and used the maximal dose of the cationic lipids or polymers at the injection volume of 100 ␮l. DOTMA is an analog of DOTAP but with higher in vivo transfection activity as well as higher toxicity (21, 27). When tested in the sequential injection protocol, the dose of the DOTMA liposome had to be lowered to 600 nmol MOLECULAR THERAPY Vol. 3, No. 5, May 2001, Part 1 of 2 Parts Copyright © The American Society of Gene Therapy

ARTICLE FIG. 6. The effect of different injection regimes on transgene expression following repeat dosing. Groups of mice (n ⫽ 3) received either CX or SQ injection of 900 nmol liposomes and 25 ␮g pDNA at day 0 and then CX or SQ injection at indicated later time points (day 3, 5, or 7). Luciferase activity in the lung was assayed at 8 h after either the first or the second injection. *P ⬍ 0.05 between indicated groups.

(with pDNA of 20 ␮g) to ensure survival of the mice for the transgene assay. At this dose, the proinflammatory cytokines TNF-␣ and IL-1␤ at 2 h were both 65–75% lower in sequential injection than those in lipoplex; transgene expression in the lung was efficient although without much improvement. The lack of improvement in transgene expression appeared to be toxicity related since the mice appeared unhealthy following injection despite the lowered cytokine levels. Similarly, at the highest dose (Fig. 3C, 50 ␮g of DNA) in the DOTAP formulation, the mice appeared sick even with reduced cytokine levels, and no advantage was observed in transgene expression following sequential injection. When a lower dose for the DOTMA formulation (360 nmol liposome and 10 ␮g pDNA) was used, the improvement in transgene expression in the lung by sequential injection was 5- to 10-fold (data not shown). These observations indicate that in addition to the proinflammatory cytokine response, the

cationic lipids at high doses also contribute to the in vivo toxicity in cationic liposome-mediated systemic gene transfer. Different formulations differ greatly in this regard; DOTAP appeared to have a lower dose-limiting toxicity than DOTMA. DOTAP/Chol liposome–protamine–DNA lipopolyplex when injected sequentially (LP 3 D) showed a substantially (54%) lower level of TNF-␣ and a more than fivefold higher transgene expression than LPD lipopolyplex injection. DC-Chol is a cholesterol-based cationic lipid and is inefficient in systemic gene delivery. No improvement in gene transfer was found for this lipid using the sequential protocol. Polyplex formulations, however, did not benefit from sequential injection. PEI/pDNA complex was efficient in pulmonary gene transfer with relatively low levels of TNF-␣ and IL-1␤ at the 25-␮g DNA dose and N/P ratio of 6. Sequential injection of PEI and pDNA gave significantly

FIG. 7. The effect of dexamethasone on cytokine kinetics induced by sequential injection of cationic liposomes and pDNA. Groups of mice (n ⫽ 3) received intraperitoneal injection of 100 ␮g DEX (1 mg/ml in PBS) or PBS within a half-hour before tail vein injection of SQ 900 nmol liposomes and 25 ␮g pDNA. Blood was collected at indicated time points and serum levels of TNF-␣ and IL-12 were assayed by ELISA. **P ⬍ 0.01; ***P ⬍ 0.001 SQ vs SQ ⫹ DEX group.

MOLECULAR THERAPY Vol. 3, No. 5, May 2001, Part 1 of 2 Parts Copyright © The American Society of Gene Therapy

679

ARTICLE TABLE 1 Hematological and Serological Parameters Affected by Injection of DOTAP/Cholestrol Liposome and pDNA Sequentially or as a Complex Naı¨ve Platelets at 24 h (⫻1000/␮l) Complement activity at 48 h (100% of naı¨ve) Liver enzyme at 24 h ALT (U/L) AST (U/L) WBC counts at 24 h (100% of naı¨ve) Lymphocyte Monocyte Neutrophil

CX

SQ

DOTAP

PBS

914 ⫾ 120

454 ⫾ 50**1

700 ⫾ 21*1**2

914 ⫾ 78

999 ⫾ 0

100 ⫾ 6.8

12.8 ⫾ 10.1***1

42.7 ⫾ 12.6**1*2

91.0 ⫾ 2.6

96.3 ⫾ 14.8

35 ⫾ 7 179 ⫾ 20

415 ⫾ 88**1 581 ⫾ 53***1

112 ⫾ 84*2 351 ⫾ 85*1*2

26 ⫾ 9 135 ⫾ 18

24 ⫾ 11 146 ⫾ 18

100 ⫾ 12.6 100 ⫾ 38.5 100 ⫾ 29.4

32.5 ⫾ 4.2***1 40.1 ⫾ 36.8 84.2 ⫾ 47.5

42.8 ⫾ 10.5**1 35.7 ⫾ 14.3 111.8 ⫾ 34.7

47.7 ⫾ 25.4*1 39.0 ⫾ 13.7 77.1 ⫾ 7.6

69.7 ⫾ 14.11 72.0 ⫾ 31.3 93.1 ⫾ 22.7

Note. Injections for CX and SQ groups were performed as described under Materials and Methods. Each animal in the DOTAP group received tail vein injection of 200 ␮l solution containing 900 nmol DOTAP/cholesterol liposome. Mice in the PBS group received 200 ␮l of PBS buffer. Naı¨ve mice had no injections. n ⫽ 4 mice/group in the complement assay and n ⫽ 3 in all other assays. *1, **1, and ***1: significantly different compared to naı¨ve group. *2 and **2: significantly different compared to CX group.

lower transgene expression in the lung. SuperFect, an activated dendrimer formulation commercially available for in vitro transfection, was inefficient in transfection and cytokine induction in vivo. Sequential injection improved transfection more than 10-fold but the expression level was still low.

DISCUSSION Several groups have recently reported efficient pulmonary gene transfer via systemic administration of lipoplex or lipopolyplex (24, 28 –35). Gene transfer via these vectors is largely due to a passive mechanism of vector aggregation (36 –38), in which serum proteins induce aggregation of the lipidic vectors. Aggregates are efficiently entrapped in pulmonary vasculature due to their relatively large size. While aggregation allows for a sufficient interaction of the vector or released DNA with the target cells (endothe-

lial cells), the nonspecific interactions of the vector with immune cells also trigger the induction of large amounts of proinflammatory cytokines (6, 12). The essential components responsible for cytokine induction are the unmethylated CpG motifs in the DNA (6 –12). Cationic lipids play a synergistic role, helping to protect the DNA from degradation and present the CpG motifs to immune cells. These proinflammatory cytokines not only are toxic to animals but also inhibit transgene expression (6, 12, 18). Systemic administration of naked plasmid DNA under physiological conditions is not effective in gene transfer. Injection of naked DNA via the vascular route also appears to be inert with respect to cytokine induction. A minimal cytokine response with naked DNA might be due to inefficient interactions of the naked DNA with immune cells and the rapid degradation of free DNA in the blood circulation. Some recent studies in our lab suggest that

TABLE 2 Cytokine and Transgene Expression Levels Following Sequential vs Complex Administration of the Components of Several Lipo(poly)plex and Polyplex Formulations Formulation TNF-␣ at 2 h (pg/ml) IL-1␤ at 2 h (pg/ml) Luciferase activity in the lung at 8 h (⫻103 RLU/mg protein)

CX SQ CX SQ CX SQ

DOTAP

DOTMA

LPD

DC-Cho1

PEI

SuperFect

3684 ⫾ 840 729 ⫾ 226** 264 ⫾ 41 102 ⫾ 21** 108 ⫾ 33 4840 ⫾ 1482**

1618 ⫾ 750 432 ⫾ 101 302 ⫾ 58 103 ⫾ 57* 896 ⫾ 349 1078 ⫾ 216

1604 ⫾ 508 738 ⫾ 642 102 ⫾ 21 95 ⫾ 58 320 ⫾ 349 2067 ⫾ 125**

140 ⫾ 125 69 ⫾ 5 31 ⫾ 8 24 ⫾ 2 5.3 ⫾ 4.8 5.2 ⫾ 3.6

290 ⫾ 26 172 ⫾ 73 27 ⫾ 5 30 ⫾ 4 186 ⫾ 35 31 ⫾ 20**

295 ⫾ 62 395 ⫾ 204 32 ⫾ 16 20 ⫾ 10 2.2 ⫾ 1.0 29 ⫾ 16*

Note. CD-1 mice received complex (CX) or sequential (SQ) injection of each formulation as described under Materials and Methods. Blood was withdrawn at 2 h after injection for cytokine assay. Mice were sacrificed at 8 h after injection for measurement of luciferase activity in the lung. n ⫽ 3 mice/group. ** P ⬍ 0.01 and * P ⬍ 0.05, SQ vs CX group.

680


ARTICLE plasmid DNA can be active in gene transfer provided a sufficient interaction with the microvascular endothelium in a given organ. For example, temporary clamping of the hepatic artery, portal vein, and vena cava following tail vein injection of plasmid DNA in 0.1 ml solution led to efficient liver transfection in mice (Liu and Huang, submitted for publication). The mouse diaphragm can also be efficiently transfected in a similar manner (Liu and Huang, submitted for publication). The ability of plasmid DNA to transfect lung via the vascular route has been demonstrated in a recent study from D. Liu et al. They have shown that preinjection of free cationic liposomes followed by injection of free plasmid DNA leads to efficient pulmonary gene transfer (21). In vitro studies of the interaction of cationic liposomes or lipoplex with mouse serum suggest that it is unlikely that the aggregated liposomes can interact with the subsequently injected plasmid DNA to form new complexes in the blood circulation. They speculate that the naked DNA is responsible for lung transfection. Liposome aggregates appear to facilitate the interaction of plasmid DNA with vascular endothelium by slowing down the pulmonary microcirculation. The observation that the administration of plasmid DNA into an isolated lung led to efficient lung transfection supports this concept (21). While the detailed mechanism responsible for lung transfection by complex injection or sequential injection remains to be unveiled, the sequential injection protocol should be associated with a reduced cytokine response. The DNA administered in sequential protocol is probably largely free, or weakly associated with partially neutralized cationic liposomes, and therefore the DNA interacts poorly with immune cells. This hypothesis was tested and confirmed in the current study. As shown in Figs. 1– 4, sequential injection leads to a 50 – 80% lower level of several key proinflammatory cytokines, including TNF-␣, IL-12, and IFN-␥, than lipoplex injection. Sequential injection is also associated with fewer hematological and serological side effects than lipoplex injection (Table 1). In addition to a reduced cytokine response, sequential injection also yields more efficient pulmonary gene transfer compared to complex injection. The improved pulmonary transfection might partly be due to the inherently different efficiency of the two transfection mechanisms (free DNA vs complex transfection). Sequential injection may have the advantage of an even distribution of DNA and the uptake of DNA by a larger number of pulmonary endothelial cells, compared to an aggregation-mediated uptake process. The improved lung transfection may also be due to a reduced inhibitory effect of cytokines on transgene expression. A TNF-␣ neutralizing antibody, a circulating TNF-␣ receptor, and the immunosuppressant DEX significantly improved transgene expression (6, 12, 39). In addition, high levels of cytokines caused a prolonged refractory period for repeated dosing (6), which could be lessened with DEX (12). Sequential injection not only yields higher transfection following single dosing, it also has a much shorter refractory period (Figs. 5 and 6). All of the data suggest that the reduction of proinflamMOLECULAR THERAPY Vol. 3, No. 5, May 2001, Part 1 of 2 Parts Copyright © The American Society of Gene Therapy

matory cytokines in sequential injection contributes to the improved pulmonary transfection. The low level of cytokine induction and improved pulmonary gene transfection with sequential injection protocol appears to be a general phenomenon for several cationic lipids examined in this study. Cationic polymers, such as PEI, appear to be an exception (Table 2). Transgene expression in the lung using the sequential injection protocol was much lower than that using PEI/DNA complex. The mechanism for this difference remains unknown. Some studies suggest that the PEI/DNA complex transfects the lung via a different mechanism compared to lipoplex (40, 41). While sequential injection brings about a reduction of serum levels of several cytokines, the response is different for each cytokine. The most significant effect is seen with TNF-␣, and a less dramatic reduction is observed with IL-12 and IFN-␥. While the mechanism for the difference is not clear and requires further studies, the unique cytokine profile may suggest an application in cancer gene therapy. IFN-␥ and IL-12 are generally believed to beneficial for the treatment of tumors (42, 43). While TNF-␣ possesses anti-tumor activity, at high dose it is involved in systemic toxicity. The sequential injection protocol which retains favorable cytokines and eliminates the unfavorable ones could have an improved therapeutic index in cancer gene therapy. Recently, several approaches have been employed to reduce the proinflammatory response associated with cationic lipid vectors. These include the deletion of CpG motifs (18), targeting the complex to endothelium (19), the use of PCR-amplified gene fragments (20), and the use of immunosuppressive (12) or endocytosis blocking agents (18). In comparison, sequential injection is the simplest and most effective approach and leads to significantly improved transfection and reduced toxicity without the use of an additional agent. Sequential injection also can be combined with other approaches such as DEX (Fig. 7) and CpG motif-depleted plasmid to further reduce the levels of the proinflammatory cytokines. To summarize, we have shown that the sequential protocol is highly efficient in pulmonary gene transfer with significantly reduced toxicity. This new protocol shall empower the use of cationic liposome-based gene vectors for the treatment of pulmonary metastasis. When combined with other approaches such as CpG motif depletion and the use of DEX to reduce cytokines, this gene transfer protocol could be useful for treating other diseases such as pulmonary hypertension, arteriosclerosis, and pulmonary thrombosis. ACKNOWLEDGMENTS This work was supported by NIH Grants AI 48851, DK 54225, CA 74918, DK 44935, and AR 45925 (Dr. Leaf Huang) and HL63080 (Dr. Song Li). We thank Dr. Gilbert Burckart and Ms. Jennifer Roach for critically reading the manuscript.

REFERENCES 1 Nikitin, A. Y., Juarez-Perez, M. I., Li, S., Huang, L., and Lee, W. H. (1999). RB-mediated suppression of spontaneous multiple neuroendocrine neoplasia and lung metastases in Rb⫹/⫺ mice. Proc. Natl. Acad. Sci. USA 96: 3916 –3921.

681

ARTICLE 2 Conary, J. T., et al. (1994). Protection of rabbit lungs from endotoxin injury by in vivo hyperexpression of the prostaglandin G/H synthase gene. J. Clin. Invest. 93: 1834 –1840. 3 Zhou, M. Y., et al. (1998). In vivo expression of neutrophil inhibitory factor via gene transfer prevents lipopolysaccharide-induced lung neutrophil infiltration and injury by a beta2 integrin-dependent mechanism. J. Clin. Invest. 101: 2427–2437. 4 Freimark, B. D., et al. (1998). Cationic lipids enhance cytokine and cell influx levels in the lung following administration of plasmid: Cationic lipid complexes. J. Immunol. 160: 4580 – 4586. 5 Scheule, R. K., et al. (1997). Basis of pulmonary toxicity associated with cationic lipid-mediated gene transfer to the mammalian lung. Hum. Gene Ther. 8: 689 –707. 6 Li, S., et al. (1999). Effect of immune response on gene transfer to the lung via systemic administration of cationic lipidic vectors. Am. J. Physiol. 276: L796 – 804. 7 Yew, N. S., et al. (1999). Contribution of plasmid DNA to inflammation in the lung after administration of cationic lipid:pDNA complexes. Hum. Gene Ther. 10: 223–234. 8 Krieg, A. M., et al. (1995). CpG motifs in bacterial DNA trigger direct B-cell activation. Nature 374: 546 –549. 9 Klinman, D. M., Yi, A. K., Beaucage, S. L., Conover, J., and Krieg, A. M. (1996). CpG motifs present in bacteria DNA rapidly induce lymphocytes to secrete interleukin 6, interleukin 12, and interferon-␥. Proc. Natl. Acad. Sci. USA 93: 2879 –2883. 10 Ballas, Z. K., Rasmussen, W. L., and Krieg, A. M. (1996). Induction of NK activity in murine and human cells by CpG motifs in oligodeoxynucleotides and bacterial DNA. J. Immunol. 57: 1840 –1845. 11 Sparwasser, T., et al. (1998). Bacterial DNA and immunostimulatory CpG oligonucleotides trigger maturation and activation of murine dendritic cells. Eur. J. Immunol. 28: 2045–2054. 12 Tan, Y., Li, S., Pitt, B. R., and Huang, L. (1999). The inhibitory role of CpG immunostimulatory motifs in cationic lipid vector-mediated transgene expression in vivo. Hum. Gene Ther. 10: 2153–2161. 13 Tousignant, J. D., et al. (2000). Comprehensive analysis of the toxicities induced by systemic administration of cationic lipid:plasmid DNA complexes in mice. Hum. Gene Ther. 11: 2493–2513. 14 Sato, Y., et al. (1996). Immunostimulatory DNA sequences necessary for effective intradermal gene immunization. Science 273: 352–354. 15 Weiner, G. J., Liu, H. M., Wooldridge, J. E., Dahle, C. E., and Krieg, A. M. (1997). Immunostimulatory oligodeoxynucleotides containing the CpG motif are effective as immune adjuvants in tumor antigen immunization. Proc. Natl. Acad. Sci. USA 94: 10833– 10837. 16 Hofland, H., and Huang, L. (1995). Inhibition of human ovarian carcinoma cell proliferation by liposome–plasmid DNA complex. Biochem. Biophys. Res. Commun. 207: 492– 496. 17 Whitmore, M., Li, S., and Huang, L. (1999). LPD lipopolyplex initiates a potent cytokine response and inhibits tumor growth. Gene Ther. 6: 1867–1875. 18 Yew, N. S., et al. (2000). Reduced inflammatory response to plasmid DNA vectors by elimination and inhibition of immunostimulatory CpG motifs. Mol. Ther. 1: 255–262. 19 Li, S., Tan, Y., Viroonchatapan, E., Pitt, B. R., and Huang, L. (2000). Targeted gene delivery to pulmonary endothelium by anti-PECAM antibody. Am. J. Physiol. Lung Cell Mol. Physiol. 278: L504 –511. 20 Hofman, C. R., Dileo, J. P., Li, Z., Li, S., and Huang, L. (2001). Efficient in vivo gene transfer by PCR fragment with reduced inflammatory activity. Gene Ther. 8: 71–74. 21 Song, Y. K., Liu, F., and Liu, D. (1998). Enhanced gene expression in mouse lung by prolonging the retention time of intravenously injected plasmid DNA. Gene Ther. 5: 1531– 1537. 22 Gao, X., and Huang, L. (1991). A novel cationic liposome reagent for efficient transfection of mammalian cells. Biochem. Biophys. Res. Commun. 179: 280 –285.

682

23 Sorgi, F. L., and Huang, L. (1996). Large scale production of DC-Chol liposomes by microfluidization. Int. J. Pharm. 144: 131–139. 24 Li, S., and Huang, L. (1997). In vivo gene transfer via intravenous administration of cationic lipid/protamine/DNA (LPD) complexes. Gene Ther. 4: 891–900. 25 Klerx, J. P., Beukelman, C. J., Van Dijk, H., and Willers, J. M. (1983). Microassay for colorimetric estimation of complement activity in guinea pig, human and mouse serum. J. Immunol. Methods 63: 215–220. 26 Barron, L. G., Meyer, K. B., and Szoka, F. C., Jr. (1998). Effects of complement depletion on the pharmacokinetics and gene delivery mediated by cationic lipid–DNA complexes. Hum. Gene Ther. 9: 315–323. 27 Ren, T., Song, Y. K., Zhang, G., and Liu, D. (2000). Structural basis of DOTMA for its high intravenous transfection activity in mouse. Gene Ther. 7: 764 –768. 28 Zhu, N., Liggitt, D., Liu, Y., and Debs, R. (1993). Systemic gene expression after intravenous DNA delivery into adult mice. Science 261: 209 –211. 29 Song, Y. K., Liu, F., Chu, S., and Liu, D. (1997). Characterization of cationic liposomemediated gene transfer in vivo by intravenous administration. Hum. Gene Ther. 8: 1585– 1594. 30 Hofland, H. E., et al. (1997). In vivo gene transfer by intravenous administration of stable cationic lipid/DNA complex. Pharm. Res. 14: 742–749. 31 Huang, L., and Li, S. (1997). Liposomal gene delivery: A complex package. Nat. Biotechnol. 15: 620 – 621. 32 Templeton, N. S., Lasic, D. D., Frederik, P. M., Strey, H. H., Roberts, D. D., and Pavlakis, G. N. (1997). Improved DNA:liposome complexes for increased systemic delivery and gene expression. Nat. Biotechnol. 15: 647– 652. 33 Hong, K., Zheng, W., Baker, A., and Papahadjopoulos, D. (1997). Stabilization of cationic liposome–plasmid DNA complexes by polyamines and poly(ethylene glycol)– phospholipid conjugates for efficient in vivo gene delivery. FEBS Lett. 400: 233–237. 34 Liu, F., Qi, H., Huang, L., and Liu, D. (1997). Factors controlling the efficiency of cationic lipid-mediated transfection in vivo via intravenous administration. Gene Ther. 4:517–523. 35 Liu, Y., et al. (1997) Factors influencing the efficiency of cationic liposome-mediated intravenous gene delivery. Nat. Biotechnol. 15: 167–173. 36 Li, S., Rizzo, M. A., Bhattacharya, S., and Huang, L. (1998). Characterization of cationic lipid–protamine–DNA (LPD) complexes for intravenous gene delivery. Gene Ther. 5: 930 – 937. 37 Li, S., Tseng, W. C., Stolz, D. B., Wu, S. P., Watkins, S. C., and Huang, L. (1999). Dynamic changes in the characteristics of cationic lipidic vectors after exposure to mouse serum: Implications for intravenous lipofection. Gene Ther. 6: 585–594. 38 McLean, J. W., et al. (1997). Organ-specific endothelial cell uptake of cationic liposome–DNA complexes in mice. Am. J. Physiol. 273: H387– 404. 39 Zhang, H. G., Zhou, T., Yang, P., Edwards, C. K., 3rd, Curiel, D. T., and Mountz, J. D.(1998). Inhibition of tumor necrosis factor ␣ decreases inflammation and prolongs adenovirus gene expression in lung and liver. Hum. Gene Ther. 9: 1875–1884. 40 Coll, J. L., et al. (1999). In vivo delivery to tumors of DNA complexed with linear polyethylenimine. Hum. Gene Ther. 10: 1659 –1666. 41 Zou, S. M., Erbacher, P., Remy, J. S., and Behr, J. P. (2000). Systemic linear polyethylenimine (L-PEI)-mediated gene delivery in the mouse. J. Gene Med. 2:128 –134. 42 Nemunaitis, J., et al. (1998). Phase I trial of retroviral vector-mediated interferon gamma (IFN-␥) gene transfer into autologous tumor cells in patients with metastatic melanoma. Cancer Gene Ther. 5: 292–300. 43 Tahara, H., and Lotze, M. T. (1995). Antitumor effects of interleukin-12 (IL-12): Applications for the immunotherapy and gene therapy of cancer. Gene Ther. 2: 96 –106.
