Gene therapy provides a novel strategy for cancer treatment, although it is still not capable of eradicating cancer in humans. The biggest challenge in gene therapy for the treatment of cancer is the development of appropriate vectors. Numerous viral and non-viral (synthetic) methods for gene transfer have been developed.(1–3) Generally, viral vectors are more efficient for gene delivery and gene expression than non-viral methods. However, viral vectors pose more risk than non-viral vectors. Moreover, viral vectors are not available for drug delivery, whereas non-viral vectors are capable of delivering anticancer reagents, as well as synthetic oligonucleotides, antibodies and RNA, in addition to therapeutic genes. From this perspective, non-viral vector systems are a favorable means by which to deliver cancer therapy. However, a number of barriers exist, including mechanisms which protect our body from the invasion of exogenous molecules. Non-viral vectors must be capable of overcoming these barriers in order to effectively deliver cancer therapy.
Cancers are diverse and often resistant to therapeutic strategies. Gene therapy has yet to meet the promise of a breakthrough in cancer therapy. There are several barriers to overcome in cancer gene therapy. One of the biggest challenges is the design of appropriate vectors. Numerous viral and non-viral methods for gene transfer have been developed for human gene therapy, but both viral and non-viral vectors have limitations and advantages. In this review article, recent improvements in the development of non-viral vectors for delivering gene therapy for the treatment of cancer will be discussed. (Cancer Sci 2006; 97: 348 –354)
hepatocyte growth factor
herpes simplex virus-thymidine kinase
hemagglutinating virus of Japan
mesenchymal stem cells
short interfering RNA
Although liposomes enable targeted delivery of macromolecules, a low and variable efficiency of gene transfer was observed during the early days of liposome development. The synthesis of cationic lipids produced a revolutionary improvement in gene transfer efficiency.(4) This led to the development of a new model of delivery involving liposome/DNA complexes or lipoplexes. Prior to this, DNA was incorporated into liposomes, however, lipoplexes enabled electrostatic interactions between negatively charged DNA and positively charged cationic liposomes. Numerous cationic lipids with improved transfection efficiency now exist, thus reducing the cytotoxicity of lipoplexes.(3) However, DNA is taken up into cells by endocytosis during lipoplex-mediated transfection.(5) The main problem with endocytosis-mediated delivery is that therapeutic molecules are prone to degradation within endosomes or lysosomes, as shown in Figure 1. Lipids may offer protection against the degradation of therapeutic molecules before they reach the cytoplasm. A neutral lipid, DOPE is capable of facilitating the endosomal release of DNA.(6) This discovery led to the use of a mixture of cationic lipids and DOPE for lipofection. Further analysis of various lipids has revealed that a 1 : 1 mixture of N-[1-(2,3-dimyristyloxy) propyl]-N,N-dimethyl-N-(2-hydroxyethyl) ammonium bromide and cholesterol is capable of destabilizing the endosome membrane more effectively than DOPE.(7) To further protect therapeutic molecules delivered by liposomes, DNA is now conjugated with cationic molecules. For example, protamine sulfate(5) or adenovirus mµ protein(8) are conjugated with DNA, after which the newly formed complexes are incorporated into or mixed with cationic liposomes. It is difficult to evaluate the efficiency of liposome-mediated gene delivery using cultured cells in vitro since the results are not consistent with those observed in vivo.(9) In spite of improvements in the membrane permeability of more complex cationic complexes, simple cationic liposomes remain more popular in clinical trials of cancer therapy. With regard to these trials, cationic liposomes containing HLA-B7 and β-2 microglobulin genes induce antitumor immunity in HLA-B7-negative melanoma patients, thus, a number of institutions have performed clinical trials of a liposomal drug (Allovectin-7) for the treatment of metastatic melanoma.(10) Delivery of the β-interferon gene by cationic liposomes has been evaluated to treat patients suffering from glioblastoma in Japan.(11) Several trials have also evaluated delivery of various anticancer agents using liposomes in humans.(12,13) As a result, liposomes are now considered safe for use in humans.
The requirement for nuclear transport of plasmid DNA poses a significant barrier to effective gene expression following gene therapy using non-viral vectors.(14) A number of trials have evaluated the success of nuclear transport of exogenous DNA using non-viral vectors, such as liposomes.(15,16) Incorporation of the viral machinery capable of mediating nuclear transport of exogenous DNA into non-viral vectors might enhance the migration of exogenous DNA into the nucleus. From this standpoint, more work is needed to reproduce the viral capability of transporting DNA into the nucleus.
Polymers used as non-viral vectors to enhance gene expression can be divided into two categories based on biodegradability. Various cationized non-biodegradable polymers have been evaluated with regard to their success of delivering DNA into cells, resulting in improved gene expression. These include linear cationized polymers of poly(ethyleneimine)(17) and poly-L-lysine.(18) Others are poly(N-ethyl-4-vinylpyridinium bromide), poly(dimethylaminoethyl methacrylate), chitosan, and dimethylaminodextran, or cationic polymers of branched poly(amidoamine) dendrimer and branched poly(ethyleneimine). Generally, because DNA is a large and negatively charged molecule, it has difficulty attaching to the negatively charged cell membrane for internalization. It is well recognized that cationized polymers readily form complexes with negatively charged DNA through electrostatic interactions. This condenses the DNA and creates a positive net electric charge under appropriate conditions. This facilitates cell attachment and subsequent internalization by means of endocytosis. In order to promote the internalization of DNA into cells, several cell receptor ligands have been used to take advantage of receptor-mediated endocytosis. For example, a folate can be covalently attached to a cationized polymer in order to promote DNA transfection. This folate-bound cationized polymer results in selective delivery and internalization of DNA into tumors.(19) Similarly, selective delivery and internalization of DNA into tumors can be achieved with transferrin-bound cationized polymers.(20) Galactose-bound cationized polymers enable direct delivery and internalization of DNA into the liver through the asialoglycoprotein receptor, to which galactose binds.(21) Successful use of the asialoglycoprotein receptor for DNA targeting suggests that polysaccharides might serve as useful non-viral carriers. We have designed cationized derivatives of the polysaccharides pullulan or dextran with spermine for complexation with plasmid DNA. Enhanced gene expression is observed with cationized polysaccharides, compared to commercially available cationic liposomes. Gene expression is reduced by pretreatment of cells with a natural ligand for the asialoglycoprotein receptor, asialofetuin, which clearly supports receptor-mediated endocytosis of cationized polysaccharide-plasmid DNA complexes. Efficient transfection of rat bone marrow-derived MSC with adrenomedullin plasmid DNA can be achieved using a cationized polysaccharide vector. Transplantation of adrenomedullin-transfected MSC into an ischemic site following rat myocardial infarction achieves a superior therapeutic effect to transplantation of MSC alone (Y. Tabata, unpublished data, 2005).
Biodegradable polymers have been used to achieve controlled-release of DNA, thus enhancing and prolonging gene expression. Controlled-release technology increases and prolongs the concentration of DNA around an injection site. Several reports describe the controlled-release of DNA from the matrixes of various biodegradable polymers, including poly(D,L-lactic acid-coglycolic acid), poly(lactic acid)-poly(ethylene glycol), poly(2-aminoethyl propylene phosphate), poloxamer, poly(ethylene-covinyl acetate), silk-elastinlike polymer, atelocollagen, and gelatin. Shea et al. describe the sustained release of plasmid DNA encoding platelet-derived growth factor from a poly(D,L-lactic acid-coglycolic acid) matrix, leading to enhanced deposition of extracellular matrix and blood vessel formation in vivo.(22) Controlled-release of plasmid DNA by an atelocollagen minipellet enhances gene expression and consequent therapeutic effects in animal models of disease. We have prepared cationized gelatin by chemically introducing amine residues to the carboxyl groups of gelatin, subsequently cross-linked by exposure to various concentrations of glutaraldehyde in order to produce several cationized gelatin hydrogels with different propensities toward degradation and release of DNA. Each DNA-containing cationized gelatin hydrogel significantly enhances gene expression, compared to that observed following injection of a solution containing plasmic DNA, and prolongs the duration of gene expression. Using the hydrogel system, the release of plasmid DNA is likely driven by degradation of the vector alone, as opposed to diffusion of plasmid DNA following injection. The fact that plasmid DNA becomes immobilized within the hydrogel through polyionic complexation with the cationized gelatin, makes it all the more likely that plasmid DNA is released as a result of degradation of the hydrogel carrier. Data regarding the release of plasmid DNA over time further support this hydrogel degradation theory. Plasmid DNA is more stable when complexed with a hydrogel, and the controlled-release of DNA from plasmid DNA-cationized gelatin complexes deposits greater concentrations of plasmid DNA around cells, resulting in increased efficacy of gene transfection. As degradation of the hydrogel determines the rate of release of plasmid DNA, it is possible to achieve controlled-release of plasmid DNA using any shape of hydrogel carrier. Controlled-release technology enhances the biological activity of an antitumor DNA plasmid(23) of NK4, which is a protein composed of the NH2-terminal hairpin and subsequent four-kringle domains of HGF. NK4 is a known HGF antagonist, which inhibits the ability of HGF to promote metastasis and angiogenesis. Subcutaneous injection of hydrogel microspheres containing NK4 plasmid DNA into nude mice injected with ascitic AsPC-1 tumor cells significantly prolongs mouse survival, compared with mice injected with NK4 plasmid DNA in the solution form. Thus, the capability to achieve controlled-release of DNA is a promising technology to enhance the in vivo biological effects of plasmid DNA.
Following administration of polymers or liposomes into the body, the generally rapid uptake of vectors by the mononuclear phagocyte system can prevent drugs from reaching their desired site of action if they are not being targeted to mononuclear phagocyte system tissues and organs. One effective way to tackle this problem is to modify the surface of drug carriers with PEG or PEG-like polymers. It is well known that cationized polymers or liposomes modified with PEG interact electrostatically with DNA to form complexes with a core-shell micelle structure. PEGylated vectors containing plasmid DNA circulate in the blood longer than free plasmid DNA. Drug carriers modified with PEG or PEG-like polymers enable DNA to accumulate within tumors or at sites of inflammation due to characteristic changes in the vasculature, including increased vascular permeability and a relative lack of lymph vessels, the so-called enhanced permeability and retention effect.(24) Pronectin F +, an artificial protein with repeated RGDS sequences, has been cationized and modified with PEG. When intravenously injected into mice with subcutaneous masses of Meth-AR-1 fibrosarcoma, PEG-modified cationized Pronectin F + complexed with plasmid DNA demonstrates significantly greater gene expression within tumors than PEG-free cationized Pronectin F + containing plasmid DNA and free plasmid DNA.(25)
Non-viral gene delivery vectors enable delivery of plasmid DNA into target cells. A number of methods to physically force DNA into cells have also been developed. Two such methods are electroporation and ultrasound.
In electroporation, an electric field increases the permeability of the cell membrane to facilitate the introduction of plasmid DNA into cells. Using this method, a 10- to 100-fold increase in gene expression over that obtained with administration of naked DNA alone is achieved.(26) This is a popular method for transfection in vitro. Skin and muscle are the main targets of in vivo transfection and two types of electrodes are used, plate-type and needle-type. The plate-type electrode yields more reproducible results, but the needle-type is more useful for transfection of various tissues. Electroporation poses a problem of tissue damage when high electrical currents are used. To solve this problem, it is now performed at lower voltages without a notable decline in transfection efficiency.(27) For cancer treatment, phase II trials of electro-chemotherapy are underway for head, neck, and skin cancer.(28)
Ultrasound is also available for gene delivery and is more useful in vivo than in vitro.(29) This transfection method depends on the ability of ultrasound to induce cavitation. DNA is first mixed with contrast reagents, such as Optison and Levovist, which are gas-filled particles coated with lipids or albumin. Using ultrasound, microbubbles (1–100 µm in diameter) are then generated and ruptured. This is how ultrasound can be used to induce cavitation. It is thought that membrane permeability might be enhanced when the bubbles are ruptured by the energy of the ultrasound. An alternative theory is that a high-speed jet flow (>600 km/h) is produced which might enhance membrane permeability.(30) Either way, DNA is incorporated into target cells with recovery of cell membrane permeability within 60 s. Transfection efficiency increases with increases in the mechanical index, frequency and exposure time. However, these parameters are also linked to cell damage. Therefore, it is essential to determine optimum conditions for performing ultrasound. We currently use ultrasound with a mechanical index of 0.6, a frequency of 1 MHz, and 1 min of exposure to achieve transfection in skeletal muscle, and 10 s of exposure to achieve transfection in mouse embryo tissue using Ultax UX-301 (Celcom, Fukuoka, Japan). Applications for ultrasound-mediated gene delivery in cancer therapy are similar to the electroporation method.
To avoid degradation prior to reaching the cytoplasm, fusion-mediated delivery systems have been developed, as shown in Figure 1. A fusigenic viral liposome with a fusigenic envelope derived from HVJ (Sendai virus) was first constructed.(31) HVJ is considered a mouse parainfluenza virus and is not a human pathogen. The virus is famous for inducing fusion with the cell membrane at neutral pH, and HN- and F-fusion proteins of the virus contribute to cell fusion.(32) HN binds to acetyl-type sialic acid and degrades the sugar chain with its neuraminidase activity. Then F associates with lipids, such as cholesterol within the cell membrane, to induce cell fusion. The F glycoprotein is first synthesized as inactive F0 in cells infected with HVJ, then cleaved by a host protease to produce the active F1 and F2 forms. The resulting F1 contains hydrophobic peptides of approximately 25 amino acids, which induce cell fusion. To achieve fusion-mediated gene transfer, DNA-loaded liposomes can be fused with UV-inactivated HVJ to form a fusigenic viral-liposome, the HVJ-liposome, which is 400–500 nm in diameter. Primitive HVJ-liposomes are constructed by fusion of liposomes with UV-inactivated HVJ. Reconstituted fusion liposomes can also be constructed. The HVJ virion is first completely lyzed with detergent, after which the lysate is mixed with DNA solution and various lipids are added to the mixture. By removing the detergent with dialysis or column filtration, reconstituted HVJ particles containing DNA can be constructed. Instead of using the entire HVJ virion, fusion proteins (F and HN) isolated from the virion can be mixed with the lipid/DNA mixture in the presence or absence of detergent. Reconstituted fusion liposomes are as effective as conventional HVJ-liposomes using fully intact HVJ virions in terms of delivery of FITC–ODN, as well as the luciferase gene, into cultured cells. The LacZ gene can also be directly transferred to mouse skeletal muscle using reconstituted fusion particles in vivo. Incubation with anti-F protein antibody at least 30 min prior to transfection reduces the efficiency of HVJ F protein-mediated gene delivery. However, a significant reduction in the efficiency of gene delivery is not observed when cells are incubated with wortmannin, an inhibitor of endocytosis, for 15 min prior to transfection.
It is expected that molecules of interest might be protected from degradation within endosomes and lysosomes by fusion-mediated delivery. Fluorescence is detected in nuclei following the introduction into human fibroblast cells of FITC–AS-ODN against the decorin gene using either HVJ-liposomes or lipoplex (Lipofectin) (Fig. 2a–c). However, decorin expression is suppressed following delivery of FITC–AS-ODN by HVJ-liposomes, but not lipoplex (Fig. 2d). Using the fluorescence resonance energy transfer system shown in Fig. 2(e), more than 85% of ODN labeled with two different fluorescent dyes at their 5′ and 3′ ends remain intact within the nucleus following delivery by HVJ-liposomes, compared to 30% following delivery using Lipofectin.(33)
A similar approach to enhance the efficiency of gene transfer uses fusion peptides derived from influenza virus hemagglutinin for receptor-mediated gene delivery. Combining transferrin/poly-L-lysine/DNA complexes with the hemagglutinin peptide increases gene transfer efficiency by more than 1,000-fold in cultured cancer cells, compared to gene transfer in the absence of this peptide.(34)
The use of HVJ-liposomes for delivering cancer treatment has been investigated in animal models. A melanoma-associated antigen gene or RNA injected into skeletal muscle or the spleen successfully evokes tumor-immunity to prevent melanoma growth.(35) A radio-inducible HSV-TK gene driven by an Egr-1 promoter enhances the efficacy of radiotherapy against hepatocellular carcinoma when delivered by HVJ-anionic liposomes.(36)
To achieve sustained gene expression, HVJ-liposomes can be combined with an EBV replicon plasmid containing the cis-acting oriP (the latent viral DNA replication origin) sequence and the trans-acting EBV nuclear antigen-1 gene.(37) The EBV replicon plasmid enhances transcriptional activation and enables stable retention of the plasmid DNA. Using an EBV replicon plasmid containing the HSV-TK gene, suicide gene therapy is more effective against melanoma in tumor-bearing mice.(38)
It is well known that silencing of transgene expression occurs in host cells, despite insertion of transgenes into the host genome.(39) Although limited or silenced transgene expression following transfection is a major problem with human gene therapy, the mechanism(s) by which transgene expression is regulated has yet to be determined. It is thought that histone methylase, acetylase and deacetylase can regulate transcription by modifying chromatin.(40) Histone deacetylase inhibitors can lead to a recovery of transgene expression after silencing.(41) We discovered a novel histone deacetylase, FR901228, which was originally developed as an anticancer drug by Astellas Pharma (Tokyo, Japan), capable of amplifying exogenous DNA expression both in vitro and in vivo. As the drug was first developed as an anticancer compound, synergistic anticancer effects are obtained when suicide gene therapy is combined with administration of FR901228. Co-injection of the HSV-TK gene and FR901228, followed by gancyclovir (GCV) administration, significantly suppresses tumor growth in mice, compared with gene therapy alone. Furthermore, more than 50% of tumor-bearing mice become tumor-free and survive for an extended period of time.(42) Thus, combination therapy is a promising and practical approach for cancer treatment.
HVJ envelope vector
There are some drawbacks to using HVJ-liposomes, even though the vector is widely used for gene transfer both in vitro and in vivo. One disadvantage of using HVJ-liposomes is the difficulty of isolating and producing both inactivated HVJ and DNA-loaded liposomes. Another limitation is that the fusion capacity of HVJ-liposomes decreases to approximately 2% that of native HVJ due to a reduction in the density of fusion proteins on the surface of HVJ-liposomes. To simplify the vector system and to increase the efficiency of gene delivery, we incorporated plasmid DNA into inactivated HVJ particles without liposomes.(43)
As shown in Fig. 3, HVJ can be inactivated by exposure to β-propiolactone (0.0075–0.001%) or UV irradiation (99 mJ/cm2), after which purification by ion-exchange column chromatography and gel filtration can be performed. The HVJ envelope has a diameter of 280 nm, and its zeta potential is approximately −5 mV.
Exogenous plasmid DNA can be incorporated into inactivated HVJ by treatment with mild detergent and centrifugation (10 000 g, 5–10 min). Many detergents, such as Triton X-100, NP-40, and deoxycholate, are available for preparation of the HVJ envelope vector. Without detergent treatment, DNA does not become incorporated into the viral particle. The DNA trapping efficiency of the HVJ envelope vector using this method is approximately 15–20%. Electron microscopy can be used to confirm incorporation of DNA into all inactivated HVJ particles. A 14 kb DNA plasmid is the largest piece of DNA delivered thus far, with a trapping efficiency of approximately 18%.
Similar quantities and molar ratios of F and HN fusion proteins are identified within HVJ envelope vectors as in native HVJ. Therefore, the HVJ envelope vector demonstrates similar fusion capacity as wild-type HVJ. Electron microscopic observations confirm fusion between the vector envelope and cell membrane within 3–5 s after attachment to the cell surface (Fig. 3).
The viral genome is eliminated from the HVJ envelope vector, therefore replication and viral gene expression are lacking in cells transfected with the HVJ envelope vector. Gene transfer to mouse muscle using the HVJ envelope vector is not diminished with repeated injections.(44) Following repeated injections, insufficient anti-HVJ antibody is generated to neutralize the HVJ envelope vector in mice, as unfortunately occurs with HVJ-liposomes.(45) This is probably the result of rapid fusion, such that fusion-mediated drug delivery is completed before sufficient binding of antibody to the vector.
The HVJ envelope vector is also available for drug delivery not involving gene therapy. Using the HVJ envelope vector, Cy3-labeled siRNA can be delivered into the cytoplasm of almost all cultured cells. Much attention has been paid to the use of siRNA for cancer treatment. However, it appears difficult to inhibit tumor growth using siRNA alone, especially in vivo because it is impossible to deliver siRNA into all cells of a tumor mass. A more practical approach to cancer therapy using siRNA might be to use siRNA to enhance the anticancer effects of chemotherapy or radiotherapy. CDDP, one of the most widely used anticancer drugs, inhibits cellular growth by inducing DNA double-strand breaks. However, cells can use DNA repair machinery to respond to DNA damage, thereby inducing resistance to anticancer drugs in human cancer cell lines. Rad51 plays a major role in homologous recombination repair machinery that is involved in the repair of double-strand DNA breaks generated by CDDP.(46) Overexpression of the human Rad51 gene in HeLa cells induces resistance of HeLa cells to CDDP (Fig. 4a). When Rad51 siRNA is delivered to HeLa cells using the HVJ-E, Rad51 expression is completely knocked-out. As shown in Figure 4(b), colony numbers in the presence of 0.02 µg/mL CDDP are less than 10% of those observed in the absence of CDDP, following transfer of Rad51 siRNA.(47) Combined treatment with cisplatin and Rad51siRNA significantly inhibits the growth of HeLa tumors (Fig. 4c). Rad51 siRNA also increases the anticancer effects of another chemotherapeutic drug, bleomycin.
The HVJ envelope vector has enhanced transfection efficiency after conjugation with a biocompatible polymer(48) or magnetic beads.(49) Tissue-specific HVJ envelope vectors have been constructed (Y. Kaneda, unpublished data, 2005). A clinical grade HVJ envelope vector is currently being produced for use in clinical trials. Thus far, the virus has only been produced in chick eggs(32) however, egg-derived HVJ is difficult to use in clinical trials. It is also difficult to produce large amounts of the virus in cultured cells. However, we recently succeeded in producing large amounts of HVJ in human cells using animal product-free medium. Now we can produce more than 10(10) particles/mL of culture medium of human cell-derived HVJ.(44) A pilot plant to commercially produce clinical grade HVJ envelope vector has already been established. Thus, a human cell-derived HVJ envelope vector is now ready for clinical use.