Electrogene transfer of an Epstein–Barr virus‐based plasmid replicon vector containing the diphtheria toxin A gene suppresses mammary carcinoma growth in SCID mice

Abstract

Experimental mammary cancer therapy in mice was conducted using electrogene transfer of a non‐viral EBV‐based plasmid vector (reduced size of the oriP gene), containing the DT‐A gene. Because the EBV‐based plasmid vector exhibits high transfer efficiency and strong persistent transgene expression due to autonomous replication in human cells, it is particularly suitable as a tool for cancer gene therapy. In vitro, 79% of MDA‐MB231 human mammary carcinoma cells died as a result of the EBV‐based vector containing DT‐A (pEB‐DTA) by 48 h after transfection. DNA synthesis was also significantly decreased as compared to levels with a control vector. In vivo, mammary tumors induced by inoculation of SCID mice with MDA‐MB231 cells were subsequently treated by direct injection of pEB‐DTA vector or pEB‐GFP vector as a control once a week for 5 weeks. After each injection, the tumors were subjected to in vivo electrogene transfer. Significantly reduced tumor volumes were observed for the pEB‐DTA group in experimental week 1 and thereafter throughout the study. At necropsy, strong and extent expression of GFP was still observed in tumors receiving pEB‐GFP 6 days after the last electrogene transfer. The ratio of histological necrotic area to viable area was significantly increased in the pEB‐DTA‐treated tumors, where levels of apoptosis were significantly higher than those observed in the pEB‐GFP group. DNA synthesis showed a tendency to decrease in the pEB‐DTA group but this was not significant. The incidence and multiplicity of lung metastasis were similar between the groups. There was also no difference in the density of microvessels between groups. We therefore conclude that the EBV‐based plasmid vector system combined with in vivo electrogene transfer can result in efficient gene transfection and that the non‐viral replicon vector containing DT‐A suppresses tumor growth due to apoptotic cell death in this model. (Cancer Sci 2005; 96: 434–440)

Abbreviations:

BrdU

5‐bromo‐2′‐deoxyuridine

DT‐A

diphtheria toxin‐A

EBNA1

Epstein–Barr virus nuclear antigen 1

EBV

Epstein–Barr virus

eGFP

enhanced green fluorescence protein

GFP

green fluorescence protein

HRP

horseradish peroxidase

PBS

phosphate‐buffered saline

PI

propidium iodide

SCID

severe combined immune‐deficient

TUNEL

terminal deoxynucleotidyl transferase‐mediated dUTP‐digoxigenin nick end‐labeling

vWF

von Willebrand factor.

In Japan, breast cancer incidence is increasing, and the breast became the leading cancer site for females in 1995 for the first time,^{( 1 )} in line with the rising prevalence worldwide. In the United States, breast cancer is the second leading cause of cancer death in women, with approximately 20 000 American women being diagnosed with breast cancer in 2002, and approximately 40 000 dying from the disease.^{( 2 )} Conventional therapies used in breast cancer have yielded positive results; nevertheless, the rates for cure and survival remain unsatisfactory. There is thus a need for the development of more effective and less toxic therapies that can reduce morbidity and mortality.

The DT‐A fragment is unique in specifically modifying mammalian elongation factor 2; even a single molecule of DT‐A in the cytosol leads to inhibition of protein synthesis and cell death. However, DT‐A alone applied extracellularly is not highly toxic because the binding domain is located in the B fragment. DT‐A therefore cannot enter living cells unless it is associated with the B fragment.^{( 3 )} This characteristic is particularly attractive for application in cancer gene therapy, a rapidly evolving field with many applications and approaches. Viral vectors are generally employed in the majority of current gene therapy studies because of their highly efficient transduction. However, viral vectors also have some disadvantages, such as the potential risk of production of a replication‐competent virus,^{( 4 )} insertional mutagenesis or low expression due to DNA methylation^{( 5 )} with retroviral systems, and the possibility of non‐specific inflammatory responses and antivector cellular immunity with adenoviruses.^{( 6 )} As none of the available vectors used to date satisfies all of the criteria of an ideal gene delivery system, there are still significant hurdles in vector selection and manipulation that must be overcome before large clinical trials for cancer therapy can commence. Therefore, we have tried a different approach for gene transduction.

Gene transfer by in vivo electrogene transfer (electroporation) has been found recently to be effective for introducing DNA into rodent skeletal muscle,^{( 7 )} skin,^{( 8 )} liver^{( 9 )} and urinary bladder tissue.^{( 10 )} Although previously believed to have a very low transfection efficiency, it was reported recently that, under suitable conditions, transduction was approximately equivalent to using an adenoviral vector concentration of 10⁶ transduction units/mL, but it was not as effective when compared to higher viral concentrations.^{( 11 )} The technique offers some attractive advantages over the use of viral vectors: (i) repeat transfections are possible for the accumulation of gene transduction, as vector immunogenicity should not be an issue; (ii) it is very easy to produce plasmids on a large scale that can accommodate vectors with large DNA sizes; and (iii) the procedure is safe in vivo and it does not depend on cell cycling within target tissues.

The EBV‐based plasmid vector (EBV‐based episomal vector) is a non‐viral vector carrying the EBNA1 gene and oriP from the EBV genome.^{( 12 )} EBNA1 binding to the oriP element initiates replication of the plasmid, which persists in cells for long periods. EBNA1 also functions as a nuclear localizer of plasmids and upregulates transcription. Therefore, EBV‐based vectors feature high transfer efficiency, strong transgene expression and long‐term maintenance of expression.^{( 13 )} In addition, EBV‐based vectors cannot produce infectious viral particles.

Experimental cancer gene therapeutic studies have already demonstrated that EBV‐based plasmid vectors containing the suicide gene,^{( 14} , ¹⁵ , ^{16 )} the IL‐12/IL‐18 gene^{( 17 )} or the Fas ligand (FasL) gene^{( 18 )} exert strong antitumor effects when coupled with polyplex or HVJ‐liposomes for gene transfer. Electrogene transfer has also been shown to be effective in experimental animal cancer models.^{( 19} , ²⁰ , ²¹ , ²² , ^{23 )} However, because the EBV‐based plasmid vector is large in size, we previously developed an EBV‐based vector carrying a mini‐oriP gene (a 954‐bp deletion from intact oriP, which is 2200 bp size).^{( 24 )} In the present experiment, EBV‐based plasmid vectors carrying mini‐oriP were applied using in vivo electrogene transfer. When considering specific clinical applications, external neoplasias and organs such as the breast are obvious choices for electrogene transfer because it is easy to conduct electrogene transfer. The present experiment was conducted to evaluate whether electrogene transfer of an EBV‐based replicative plasmid vector containing the DT‐A gene might be used as an effective in vivo therapy in a novel mammary cancer model.

Materials and Methods

Vectors.  The plasmid pEB6cagGFP (10.5 kb),^{( 24 )} which encodes eGFP, was used as a control vector for in vitro and in vivo studies. To produce the pEB6cagDT‐A vector (10.6 kb), the eGFP gene was deleted from pEB6cagGFP via digestion with EcoRI/NotI and then replaced with the DT‐A fragment gene.^{( 25 )} pEB6cagMCS (9.8 kb), an empty control vector, was also used for the in vitro study. These vectors all carry the mini‐oriP gene^{( 24 )} which is driven by the CAG promoter. In this manuscript, the vectors are referred to as pEB‐GFP, pEBmcs and pEB‐DTA, respectively, for simplicity. Plasmids were extracted from Escherichia coli (DH5α strain) and purified by means of a modified alkaline lysis procedure using a Quiagen Plasmid Maxi Kit (Quiagen, Valencia, CA, USA).

Mammary carcinoma cells and animals.  A mammary carcinoma cell line, MDA‐MB231, was established from a human breast cancer and maintained in RPMI 1640 medium containing 10% fetal bovine serum with streptomycin (100 µg/mL)/penicillin (100 unit/mL) in an incubator under 5% CO₂. A total of 20 female 6‐week‐old SCID mice were used in this study (Clea Japan, Osaka, Japan). The animals were housed five to a plastic cage with wood chip bedding and had free access to water and food under conditions of controlled temperature (21 ± 2°C), humidity (50 ± 10%) and lighting (12 : 12 h L : D cycle). All were held for a 1‐week acclimatization period prior to test initiation. Animals were killed using diethyl ether. All manipulations were carried out in accordance with the procedures outlined in the Guide for the Care and Use of Laboratory Animals in Osaka Medical College.

In vitro electrogene transfer.  MDA‐MB231 mammary carcinoma cells were subjected to electrogene transfer. During the transfection process, MDA‐MB231 cells (1 × 10⁶) were suspended in 100 µL of Nucleofector T solution (amaxa, Köln, Germany). Aliquots of the pEB‐GFP, pEBmcs or pEB‐DTA vectors (5 g) were added to the cell suspension and electrogene transfer was conducted using the amaxa Nucleofector with program #A23. The cells were then harvested in a 96‐well optical‐bottom plate for DNA synthesis estimation and a 6‐well plate for cell viability. Because the pEB‐GFP vector encodes eGFP, transfection efficiency was measured by assessing the number of GFP‐positive cells using a flow cytometer (EPICS Elite ESP; Coulter, Miami, FL, USA).

In vitro DNA synthesis and cell viability.  At 24 h after transfection, cells in 96‐well optical‐bottom plates were incubated in medium containing 50 µmol/L BrdU (Sigma, St Louis, MO, USA) for 1 h, rinsed with PBS, incubated with 1 mol/L HCl for 20 min, rinsed with PBS, and exposed to an anti‐BrdU mouse monoclonal antibody (Dakocytomation, Glostrup, Denmark) for 1 h. Following further PBS rinsing and incubation with secondary antibody conjugated with HRP for 1 h, the wells were rinsed and covered with a luminol substrate (Roche Diagnostics, Indianapolis, IN, USA) for 5 min. After sealing with aluminum foil, the light emission of the samples was measured using a luminometer (Luminoskan Ascent; Thermo Electron, Helsinki, Finland). The number of viable cells per well was counted 48 h after transfection using a hemocytometer with trypan blue exclusion.

DT‐A gene therapy using EBV replicon vectors by electrogene transfer in vivo.  MDA‐MB231cells (5 × 10⁶ cells/0.3 mL in PBS) were inoculated subcutaneously into the right inguinal region of 20 female SCID mice. Two weeks later, when tumors had developed to approximately 0.2–0.3 cm in diameter, pEB‐GFP or pEB‐DTA was injected slowly into the implanted tumor using a 27‐gauge needle while the animals were under diethyl ether anesthesia. These vectors (0.5 µg/µL saline) were injected into tumors sized 0.6–0.7 cm in diameter until leakage of the vector solution, while in tumors of larger dimensions, a total volume of 100 µL vector (50 µg) was applied. Electrogene transfer was carried out immediately after the injection by applying a conductive gel (Aquasonic 100, Parker Laboratory, NJ, USA) topically to the unshaved skin over the tumor. Electric pulses were then delivered directly to the tumor via ‘forceps’ plate electrodes (CUY650‐10; Nepa Gene Co., Ichikawa, Japan) using a square‐wave electropulser BTX T820 (BTX, San Diego, CA, USA) monitored by a graphic pulse analyzer (Enhancer 400; BTX). The electrodes consisted of a pair of stainless steel disks 1.0 cm in diameter (Nepa Gene Co.). Optimal conditions for the electrogene transfer were determined previously (eight pulses with a pulse length of 20 ms at 100 V).^{( 22} , ^{23 )} Vector injection and electrogene transfer were carried out once a week. The size of each treated mammary tumor was measured weekly using a caliper and tumor volume was calculated using the following formula: maximum diameter × (minimum diameter)² × 0.4.^{( 26 )} Individual body weights were recorded weekly. During week 6, all surviving mice were killed under ether anathesia and tumors were cut in half and observed 6 days after the fifth electrogene transfer under a fluorescence stereomicroscope (model MZFL III, Leica Microsystems, Bannockburn, IL, USA) for confirmation of extension of GFP expression.

Histopathological analyses.  At necropsy, mammary tumors and lymph nodes (axillary and femoral regions and any appearing abnormal) were removed, immediately fixed in 10% phosphate buffered‐formalin, and embedded in paraffin. Lungs were inflated with the fixative, excised and immersed in the fixative; all lobes were trimmed and processed for histopathology. Sections were cut at a thickness of 4 µm, stained with hematoxylin and eosin (H&E) and examined histopathologically for metastatic foci. All foci >100 µm were counted using an ocular micrometer disk (Olympus Co., Tokyo, Japan) and the data expressed as average number per animal in the group. Areas of viable and central necrotic regions in mammary cancers were determined using a MCID color image processor (Imaging Research, Ontario, Canada) and expressed as the average ratio of viable cells/necrotic cells.

DNA synthesis of mammary cancers given DT‐A gene therapy.  All animals in the groups were injected i.p. with 100 mg/kg body weight BrdU 1 h prior to being killed under ether anesthesia. Within each treatment group, five animals were selected for evaluation of DNA synthesis rates as inferred by BrdU incorporation. DNA in paraffin‐embedded tissue sections was denatured by a 20‐min incubation in 4 mol/L HCl solution for 20 min at 37°C. After incubation with an anti‐BrdU mouse monoclonal antibody (Clone Bu20a; Dakocytomation) and reaction with a secondary antibody conjugated with fluorescein‐isothiocyanate, the tissue sections were further stained with a 50 µg/mL PI solution containing 100 µg/mL RNase A for 30 min at 37°C. Whole sections were scanned automatically with a microscope‐based multiparameter laser scanning cytometer (LSC2; Olympus Optical Co.) under ×200 magnification. The numbers of fluorescein‐isothiocyanate‐positive (S‐phase cells incorporating BrdU) and PI‐positive cells were counted using LSC2 and BrdU labeling indices were expressed as a percentage of the total number of cells counted.

Apoptotic indices of mammary cancers given DT‐A gene therapy.  For quantitative analyses of apoptosis, 4‐µm sections from paraffin‐embedded tumor sections were assayed by the TUNEL method using a fluorescein in situ apoptosis detection kit (ApopTag; Chemicon International, Temecula, CA, USA) according to the manufacturer's protocol with the exception of the proteinase K incubation step (60 µg/mL for 30 min at 37°C). The tissue sections were further stained with a 50 µg/mL PI solution and also scanned automatically with LSC2 under ×200 magnification. The numbers of fluorescein‐isothiocyanate‐positive (TUNEL‐positive cells) and PI‐positive cells were counted using LSC2 and the rates for TUNEL‐positive cells were expressed as a percentage of the total number of cells counted. Sections of testes were used as positive controls.

Density of microvessels in mammary cancers given DT‐A gene therapy.  To quantitatively assess microvessels in mammary carcinomas receiving DT‐A gene therapy, sections were heated in citrate buffer (pH 6.0) using an autoclave (15 min at 121°C) and stained immunohistochemically using an anti‐vWF rabbit polyclonal antibody (blood vessel endothelial marker, diluted 1 : 200; Dakocytomation). The number of vWF‐positive microvessels was assessed according to a published method^{( 27 )}. Briefly, slides were scanned at low‐power magnification to identify the areas with the highest numbers of vessels. The five areas having the highest densities were selected and counted at ×200 to ×400 power magnification. Areas in viable regions of tissues were examined using a color image processor (MCID) and the results were expressed as the average number of microvessels/mm².

Statistical analyses.  The significance of differences between groups in quantitative data were analyzed using the two‐sided Student's t‐test via the method of Welch, which provides for insufficient homogeneity of variance. Differences in incidences were examined by the Fisher's exact probability test.

Results

Effects of DT‐A on cell viability and DNA synthesis in mammary carcinomas in vitro.  The transfection efficiency was 73% in this study using the amaxa electrogene transfer device (Fig. 1a,b). Mammary carcinoma cells transfected with pEB‐DTA (Fig. 1d) showed a significant 79% average reduction in numbers of viable cells as compared to the control pEBmcs case (Fig. 1c,e). DNA synthesis, as assessed by BrdU incorporation, was also decreased markedly in pEB‐DTA‐transfected cells as compared to pEBmcs‐transfected cells (Fig. 1f).

Figure 1.

Human mammary carcinoma cells (MDA‐MB231) were electrotransfected with the pEBV‐based episomal vector containing the eGFP gene (pEB‐GFP). (a) Dark field, (b) light field (original magnification ×200). The transfection efficiency was 73% and MDA‐MB231 cells transfected with pEB‐DTA showed a marked reduction of viable cells (d) as compared to the pEBmcs control vector case (c). (c,d) Phase contrast microscopy (original magnification ×100). (e) Viable cells and (f) DNA synthesis in MDA‐MB231 cells transfected with pEB‐DTA. MDA‐MB231 cells transfected with pEB‐DTA showed a significant reduction of viable cells (e) associated with decreased DNA synthesis (f). Data are mean ± SD, **P < 0.01 as compared with the pEBmcs‐transfected cells.

Body weight and mammary tumor growth in mice given DT‐A gene therapy.  Two mice died in week 4; one in the pEB‐GFP (control) group and the other in the pEB‐DTA group. The body weights of mice receiving vectors are shown in Fig. 2a. Body weight curves were confirmed to be similar between the pEB‐GFP and pEB‐DTA groups. The general condition of all animals given pEB‐based vectors was good throughout the study.

Figure 2.

(a) Body weight and (b) mammary tumor volume in mice receiving the pEBV replicon vector containing the DT‐A gene. (a) The body weights of mice treated with pEB‐DTA were similar to those observed in the pEB‐GFP control group throughout the study. (b) Tumor volume was significantly suppressed in the pEB‐DTA group from week 1 onwards (*P < 0.05; **P < 0.01 as compared with the pEB‐GFP group). Data presented are mean ± SD. (c–f) Expression of GFP within tumors 6 days after the fifth electrogene transfer. (c,d) pEB‐GFP group, (e,f) pEB‐DTA group. Expression of GFP was still strong and extensive in a tumor treated with pEB‐GFP ([c] light field, [d] dark field), while an example treated with pEB‐DTA does not show any GFP fluorescence ([e] light field, [f] dark field). (e) Note the apparent necrotic area after treatment with pEB‐DTA (arrows). Micrographs were taken using a fluorescence stereomicroscope with a 480/40 nm excitation filter. Scale line = 0.5 cm.

Data for tumor volumes in mice receiving vectors are presented in Fig. 2b. Tumor growth was significantly suppressed in the pEB‐DTA group from week 1 as compared to the pEB‐GFP group throughout the study. Fluorescence stereomicroscopy after necropsy (during week 6) revealed strong expression of GFP in mammary tumors treated with pEB‐GFP 6 days after the last (fifth) electrogene transfer (Fig. 2c,d). In contrast, the tumors treated with pEB‐DTA did not show apparent GFP expression (Fig. 2e,f).

Histopathological analyses.  Histopathologically, the implanted mammary carcinomas proved to be moderately differentiated adenocarcinomas (Fig. 3a). Cancer cells had abundant eosinophilic cytoplasm with rounded nuclei and prominent nucleoli, and mitoses were observed frequently (Fig. 3a). Tumors were characterized by areas of central necrosis (Fig. 3b,c) with inflammatory cells, especially neutrophils (Fig. 3d). The ratio of necrotic to viable areas was significantly increased in the tumors treated with pEB‐DTA as compared to the pEB‐GFP group (Fig. 3g).

Figure 3.

Histopathological analysis of tumors in mice treated with EBV replicon vector containing DT‐A. (a) pEB‐GFP group, (b) pEB‐DTA group, (c) pEB‐GFP group. The implanted mammary carcinomas proved to be moderately differentiated adenocarcinomas, characterized by areas of central necrosis (asterisk) with inflammatory cells at the periphery, independent of the treatment. (d) These were mainly neutrophils (pEB‐DTA group). (e) Metastasis to lymph nodes was only observed in mice treated with pEB‐GFP. Note the lymphoid sinuses filled with mammary carcinoma cells (arrow). (f) Metastatic foci in the lungs were observed frequently, independent of the treatment (arrow, pEB‐DTA group). H&E stain. Original magnifications: (a) ×400, (b,c,f) ×40, (d) ×400, and (e) ×100. (g) The ratio of necrotic/viable areas was significantly increased in the pEB‐DTA group (*P < 0.05 as compared with the pEB‐GFP group). (h) Metastasis to the lymph nodes was observed only in mice treated with control pEB‐GFP, while lung metastasis was relatively frequent with no difference between the pEB‐GFP and pEB‐DTA groups. Data presented are mean ± SD.

Metastases to lymph nodes and lungs were also evaluated histopathologically and the results are shown in Fig. 3h,i. The incidence of metastasis to lymph nodes was 20% in the pEB‐GFP group (Fig. 3e), whereas no lymph node metastasis was apparent in the pEB‐DTA group (Fig. 3h). Incidences of lung metastases were 60% in the pEB‐GFP group and 40% in the pEB‐DTA group, the difference not being significant (Fig. 3f,h). Multiplicity of metastases to lymph nodes and lungs also demonstrated similar trends (Fig. 3i).

DNA synthesis and apoptosis.  Levels of DNA synthesis, as evaluated by BrdU labeling, showed a tendency for a decrease in the pEB‐DTA group as compared to the pEB‐GFP group but this did not attain statistical significance because of large variation in the pEB‐GFP group (Fig. 4a). Levels of TUNEL‐positive cells were significantly increased in mammary tumors treated with pEB‐DTA as compared to pEB‐GFP (Fig. 4b).

Figure 4.

(a) DNA synthesis, assessed by BrdU incorporation, showed a tendency for a decrease in the pEB‐DTA group but this was not statistically significant because of large variation in the pEB‐GFP group. (b) The rate for TUNEL‐positive (apoptotic) cells was significantly elevated in the pEB‐DTA group as compared to the pEB‐GFP group (**P < 0.01). (c) The density of microvessels, assessed using anti‐vWF antibody for blood endothelial cell marker, was similar between the pEB‐GFP and pEB‐DTA groups. Data presented are mean ± SD.

Analysis of angiogenesis in tumors receiving DT‐A gene therapy.  Microvessel density, determined by immunohistochemical analysis with the blood vessel endothelial cell marker vWF, did not demonstrate any intergroup difference (Fig. 4c).

Discussion

The present study provided concrete evidence that electrogene transfer of an EBV‐based plasmid replicon vector containing the DT‐A gene suppresses mammary carcinoma growth in SCID mice. It was previously shown that EBV‐based plasmid vectors have high transfection efficiency using HVJ‐liposomes,^{( 28} , ^{29 )} cationic liposomes^{( 29 )} or polyamidoamine dendrimers^{( 30 )} in various cells of human origin in vitro. The efficiency of the transfection in MDA‐MB231 cells (human mammary carcinoma cells) was extremely high (73%) in the present study and a 79% cell death rate was evident after 48 h. DNA synthesis was also significantly decreased in cells transfected with pEB‐DTA as compared to the control vector.

DT‐A exerts toxicity by inhibiting protein synthesis and inducing apoptotic cell death^{( 31} , ^{32 )} and has thus been proposed as a suicide gene for use in cancer gene therapy.^{( 33 )} In fact, histologically necrotic areas were significantly increased in tumors transfected with pEB‐DTA in the present study, and levels of apoptosis were significantly increased. As DT‐A‐mediated apoptotic cell death is cell cycle‐independent, it is particularly attractive for application as a cytoreductive gene therapy. However, in vivo in the present study, although tumor growth was significantly inhibited in the pEB‐DTA group, the inhibitory potential appeared much weaker than earlier reported for EBV‐based vectors containing HSVtk ^{( 14} , ¹⁵ , ^{16 )}, similar to a trend observed in another vector system.^{( 19 )} In this context it should be remembered that DT‐A exhibits no bystander effect, as it is unable to cross cell membranes when lacking the B fragment. This could be a disadvantage in comparison to the HSVtk/GCV system, which is known to be highly effective in rapidly dividing cancer cells due to its pronounced bystander effect.

However, DT‐A‐induced apoptotic cells can induce an immune response specifically against cancer cells.^{( 34 )} In the present study, human cancer cells were used for autonomous replication of the EBV vector and SCID mice were selected for in vivo inoculation because they lack both T cells and B cells. This is presumably another reason why tumor retardation was not as strong as with the HSVtk/GCV system using pEB‐based plasmid vectors.^{( 14} , ¹⁵ , ^{16 )} HSVtk/GCV therapy has a strong bystander effect and it appears that as few as 10% of cells need to express HSVtk to produce this phenomenon.^{( 35 )} While HSVtk/GCV therapy does suppress transplanted murine hepatocellular or mammary carcinomas in immunocompetent syngeneic mice,^{( 22} , ^{36 )} the efficacy is decreased in athymic nude mice that do not have an intact T‐cell system, suggesting that T‐cell‐mediated immune responses may be a critical factor for the bystander phenomenon^{( 36} , ^{37 )}.

In the present study, there were no differences in lung metastasis between pEB‐DTA and control vector groups and microvessel density within tumors also did not vary between groups. Again, DT‐A would not be expected to exert any effect on surrounding non‐transfected cells, including blood vessel or lymphatic endothelial cells, even if released from the dead cells because of the lack of the B fragment.

Tumor growth was suppressed with electrogene transfer conducted once a week in the present study. To completely regress tumor growth future studies should focus on optimal dosage, as it has been reported that electrogene therapy administered three times a week has much stronger effects than with two times a week.^{( 19 )}

Although the mutated DT‐A gene was not strong enough to suppress metastasis, combination of an EBV replicon vector (mini‐oriP, a 954‐bp deletion from intact oriP of 2200 bp size) with electrogene transfer resulted in sustained and expanded GFP expression in mammary carcinomas, so that this combined gene transfer system may be a good tool for gene therapy. Mazda described in a review article that an EBV‐based plasmid vector may not merely be a good ‘expression vector’, but also a ‘gene transfer vector (delivery vehicle)’.^{( 13 )} Other genes incorporated into the EBV replicon vector containing mini‐oriP should now be tested for use in cancer gene therapies.

The EBV is responsible for various human malignancies, as exemplified by Burkitt's lymphoma, Hodgkin's lymphoma, nasopharyngeal carcinoma and gastric cancers, and EBNA1 is expressed constitutively in all malignant cells of EBV‐associated tumors. In contrast, EBNA1 is not expressed in normal cells. With our plasmid vector, very strong expression was found exclusively in EBV‐associated tumors, while transfected normal cells hardly expressed the transgene. For possible clinical application and specific killing of tumor cells, therefore, EBV‐based plasmid vectors appear to be extremely suitable for targeting EBV‐associated tumors.^{( 13 )} However, because EBV‐based plasmid vectors can replicate autonomously in primates, the clinical safety of the EBV vector system must be considered.

So far, the EBV‐based plasmid vector system has also been tried for various non‐neoplastic diseases such as adenosine deaminase deficiency,^{( 38 )} Duchenne muscular dystrophy,^{( 39 )} chronic heart failure^{( 40 )} and diabetes^{( 41 )} using animal models, so that the scope for application appears wide.

There are several incentives for adding in vivo electrogene transfer to the gene therapy arsenal. Problems associated with immunogenicity may be avoided, even with repeated treatments, and the method is relatively easy to apply and is well tolerated. Electroporation‐mediated delivery of anticancer drugs has been carried out in clinical trials and its safety has already been demonstrated in humans.^{( 42} , ^{43 )} When considering specific clinical applications, external neoplasias and organs such as the breast are obvious choices for easiness of electrogene transfer.

In conclusion, the present study demonstrated that an EBV replicon non‐viral vector (carrying mini‐oriP) containing DT‐A suppresses tumor growth in a mouse mammary cancer model after electrogene transfer for in vivo transduction. This growth suppression was particularly associated with increased apoptotic cell death.