Enhanced antitumor effect of EGF R–targeted p21WAF-1 and GM-CSF gene transfer in the established murine hepatoma by peritumoral injection


One of the major obstacles in current cancer gene therapy is the lack of a gene delivery system with high efficiency and targetability. In this paper, a nonviral gene delivery system GE7, which was designed to target EGF receptor (EGF R) overexpressed on the surface of cancer cells through an EGF R–binding oligopeptide (GE7), was used for in vivo gene therapy in a murine subcutaneous hepatoma model. It was demonstrated that the GE7 system could target the reporter gene β-gal to EGF R–expressing hepatoma cells with high efficiency after in vitro transfection and in vivo peritumoral injection. To improve the therapeutic effect elicited by single gene transfer, human cyclin-dependent kinase inhibitor gene p21WAF-1 and murine cytokine gene GM-CSF were used simultaneously in peritumoral injection of the GE7/DNA polyplex. The results showed that combined gene transfer of p21WAF-1 and GM-CSF could inhibit the growth of pre-established tumor more effectively and prolong the survival time of hepatoma-bearing mice more significantly than the transfer of a single gene. Apoptosis in the tumor tissues were found when injected with the p21WAF-1 DNA polyplex. Prominent inflammatory infiltration was observed in the tumor tissue transfected with the GM-CSF DNA polyplex. Our data demonstrate that the GE7 system–mediated, EGF R–targeted cotransfer of p21WAF-1 and GM-CSF genes exhibit more potent antitumor effect by inducing tumor cell apoptosis and inflammatory responses.


Gene therapy represents an important strategy in the treatment of cancer. Although significant progress in cancer gene therapy has been made within the past decade, two major obstacles must be addressed if cancer gene therapy is to be successful. First, an efficient vector needs to be designed that can cause prolonged high expression of the transduced gene. Second, the therapeutic gene can be delivered to a particular tumor cell, that is to say, a targeted gene delivery system with high transduction efficiency needs to be established.1,2,3,4 However, most of the current gene delivery systems are able to transfer genes into cancer cells in vivo, but are nonselective in their targeting. To enhance the efficiency and targetability of gene transfer, a combination of a ligand oligopeptide for IGF I receptor, EGF receptor (EGF R) or VEGF receptor covalently linked to polylysine combined with an endosomolytic peptide (HA20)-polylysine, has been developed in our laboratory.5,6 One of these effective gene delivery systems — an EGF R ligand (16 amino acids) oligopeptide-polylysine combined with an HA20 (20 amino acids) oligopeptide-polylysine, designated as the GE7 system, has shown that it could not only efficiently transfer exogenous gene into tumor in vivo, but also transduce genes into tumor at a very low dose of 0.2 μg DNA plus 0.2 μg polypeptide per mice. To the best of our knowledge, no such formulation of targeting vector systems has been described and reported previously. Based on the overexpression of EGF R in hepatocellular carcinoma cells,7 in this report, we used the GE7 system for the in vivo delivery of therapeutic genes in a murine subcutaneous hepatoma model.

It has been reported that various kinds of therapeutic genes have been transferred to tumor tissues that have resulted in tumor suppression.8,9 Recent reports focus on combined gene administration aiming to obtain cumulative or synergistic antitumor effects. Reports indicated that combinations of suicide genes, cytokine genes, and tumor suppression genes could induce the synergistic therapeutic effects.10,11,12,13,14,15

p21WAF-1, a wide-spectrum cyclin-dependent kinase inhibitor (CKI), was confirmed to participate in G1/S point regulation downstream from p53, and was shown as a tumor suppressor gene by evidences of preclinical administration in many tumor models,16,17 We have also reported that p21WAF-1 gene transfer could remarkably inhibit the growth of hepatoma cells in vitro and in vivo.18

It is well known that defects in immunosurveillance system are involved in most human malignancies. Therefore, different approaches have been designed to induce potent immune response for the treatment of cancer. Granulocyte–macrophage colony-stimulating factor (GM-CSF) is a well-characterized cytokine to stimulate an effective antitumor immunity by promoting the activation and maturation of antigen-presenting cells (APCs), including dendritic cells, and increasing the immunogenicity of tumor cells.19,20 Our previous work on GM-CSF by using it as an immune-enhancing agent revealed that GM-CSF gene transfer has the ability to stimulate potent antitumor response when used alone or in combination with other genes.21,22,23

It was known that p21WAF-1 gene transfer could achieve tumor growth inhibition, but that it does not possess the potent capacity of inducing antitumor immunity. For cytokine gene transfer alone, it usually induced the antitumor immune response, but seems not strong enough to eradicate established tumors. Therefore, in the present study, we employed a new strategy of combined gene therapy inwhich the CKI gene p21WAF-1 and the murine cytokine gene GM-CSF were delivered simultaneously into murine subcutaneous hepatoma with high EGF R expression by our EGF R–targeting nonviral gene delivery system GE7 in the form of a polyplex of plasmid DNA and oligopeptides. By use of this experimental system, we investigated whether the in vivo antitumor effect could be enhanced by cotransfer of p21WAF-1 and GM-CSF genes. Additionally, peritumoral injection was used in this study to deliver the therapeutic agent, aiming to achieve increased drug concentration in the target tissues and to explore an experimental model for regional interventional gene therapy. Our results demonstrate that combined transfer of p21WAF-1 and GM-CSF genes generates a more potent antitumor effect than the single gene transfer of p21WAF-1 or GM-CSF alone.

Materials and methods

Cell line and mice

Mouse hepatoma cell line Hepa (CRL-1830, ATCC, Manassas, VA), originated from C57BL/6 mice, was grown and maintained in DMEM supplemented with 10% heat-inactivated fetal bovine serum, penicillin (100 U/mL), streptomycin (100 mg/mL), and amphotericin (4 μg/mL). All culture media were obtained from Gibco-BRL (Grand Island, NY). Female C57BL/6 mice, 6 weeks of age, were provided by Joint Ventures SIPPR BK Experimental Animals (Shanghai, China) and housed in a specific pathogen-free condition for all experiments.

Preparation of GE7-poly-L-lysine/DNA/HA20-poly-L-lysine polyplex

The human p21WAF-1 (provided by Dr. Bert Vogelstein, Johns Hopkins University, Baltimore, MD), murine GM-CSF cDNA sequences obtained by RT-PCR, or E. coli β-gal cDNA sequences purchased from Promega (Madison, WI) was cloned into the multiple cloning region of pCEP vector (gift from Dr. Bert Vogelstein), respectively, yielding the recombinant clones of pCEP-p21WAF-1, pCEP-mGM-CSF, and pCEP-β-gal.

An oligopeptide of 16 amino acids (NPVVGYIGERPQYRDL) was designed from the EGF binding domain by the assistance of a computer. It was named GE7 for EGF R recognition. HA20 from influenza hemagglutinin was synthesized for endosomolysis24 and the poly-L-lysine (Sigma, St. Louis, MO) was used for DNA binding as a backbone. The preparations of GE7-poly-L-lysine (GE7-PL) and HA20-poly-L-lysine (HA20-PL) were performed as previously described by us.4

A mixture of GE7-PL and HA20-PL at a molar ratio of 1:1 was added dropwise to different DNA solutions followed by reaction at 25°C for 30 minutes. The molar rate of polypeptide to DNA was also 1:1. To prepare the complex GE7-PL/p21WAF-1+pGM-CSF/HA20-PL, the two plasmid DNA were mixed well before reaction with the GE7-PL/HA20-PL mixture. Agarose gel electrophoresis was used to examine the formation of GE7-PL/DNA/HA20-PL polyplex. All of the GE7-PL/DNA/HA20-PL polyplex used in transfection in the following experiments were named briefly as GE7/p21WAF-1, GE7/pGM-CSF, GE7/β-gal, GE7/pCEP, and GE7/p21WAF-1/pGM-CSF individually.

EGF R evaluation by immunohistochemistry assay

Hepa hepatoma cells were collected by trypsin digestion and washed three times by PBS. After fixing in 10% formaldehyde solution for 30 minutes, the cells were washed again and dripped onto the holder plates that had been coated with polylysine. Standard immunohistochemistry assay was accomplished after the cells were fixed on plates by heating at 55°C. Anti–EGF R mAb and HRP-labeled IgG (obtained from Dako, Carpinteria, CA) were used as the first and the secondary antibody, respectively.

Determination of transduction efficiency by β-gal expression assay

The in vivo β-gal assay after peritumoral injection with GE7/β-gal polyplex was carried out as previously described.4,5 GE7/β-gal polyplex was injected peritumorally at a dose equivalent to 0.2 μg DNA per mouse in 100 μL normal saline when Hepa-derived subcutaneous hepatoma were established at a size of 5 mm in diameter. Forty-eight hours later, 10 organs and tissues including tumor were excised, fixed at 4°C for 1.5 hours, and stained with X-gal solution [1 mg/mL 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside, 5 mM K3 (Fe(CN)6), 5 mM K4 (Fe(CN)6), and 2 mM MgCl2]. For histological examination, stained tumors and tissues were paraffin embedded for sections and counterstained with Nuclear Fast Red. The rate of gene transduction was determined by averaging the number of blue cells and of total cells for 10 randomly chosen fields.

Preparation of hepatoma-bearing mice and treatment procedures

Hepa hepatoma were obtained and cut into pieces. Tumor cubes (1 mm3) were implanted subcutaneously into the right flank of female C57BL/6 mice. Three days later, the animals received peritumoral injection once a week of the GE7/p21WAF-1 polyplex, GE7/pGM-CSF polyplex, or GE7/p21WAF-1/pGM-CSF polyplex containing plasmid DNA at a dose equivalent to 0.2 μg in 0.1 mL normal saline for each injection, respectively. Mice treated with normal saline, GE7-PL/HA20-PL polyplex, GE7/pCEP polyplex, or plasmid DNA only were used as controls. Tumors were measured with a caliper in two perpendicular diameters once a week after inoculation and the volume was expressed according to the formula V=πab2/6, where a is the largest tumor diameter and b the smallest diameter. Life spans of the mice among different groups were expressed by the percentage of surviving mice. Tumor size measurements and survival observations were continued for 3 months.

Assay for the expression of exogenous genes in transfected tumor tissues

Immunohistochemistry evaluation

Immunohistochemical a- nalysis of exogenous gene expression in tumor sections was carried out as described previously.25 Tumor tissue samples obtained from mice injected with GE7-PL/DNA/HA20-PL polyplex were fixed, paraffin embedded, and sectioned at 5 μm. The sample sections were subsequently deparaffinized in xylem, dehydrated with a graded series of ethanol/water washes, then processed with a standard immunohistochemistry procedure. After incubation with anti–p21WAF-1 mAb (Oncogen, Cambridge, MA) as primary antibody at 37°C for 1 hour, the sections were incubated with HRP-labeled secondary Ab (Dako) for another hour.

Western blot analysis

Tumor tissues were prepared from mice after peritumoral injection of GE7-PL/DNA/HA20-PL polyplex once a week for 1 month, and lysed by a buffer solution [50 mM Tris–Cl (pH 8.0), 2% SDS, 2 mM PMSF]. For each sample, 20 μg of proteins were subjected to Western blot analysis. Protein concentration, transfer, and blocking were performed in turn followed by being probed, respectively, with mouse anti–hu-p21WAF-1 mAb (Oncogen) and goat anti–mouse-GM-CSF mAb (Oncogen). The HRP-labeled anti–mouse IgG and anti–goat IgG were used as secondary antibodies. Tumor tissues from mice treated with normal saline and GE7/pCEP polyplex were used as control.26

Detection of DNA fragmentation after GE7/p21WAF-1 transduction

Hepatoma-bearing mice were peritumorally injected with GE7/pCEP polyplex, GE7/p21WAF-1 polyplex, GE7/pGM-CSF polyplex, or the GE7/p21WAF-1/pGM-CSF polyplex once a week for 1 month. Genomic DNA of tumor tissues was isolated, respectively, and analyzed in 1.5% agarose gel electrophoresis for examination of the DNA-ladder formation.

Histological analysis of tumor tissues

After mice were injected with GE7-PL/DNA/HA20-PL polyplex for about 1 month, tumor tissues were carefully excised when mice were sacrificed, fixed in 10% solution of formalin in PBS for 24 hours, and processed for paraffin embedding. Serial 5-μm sections were prepared and stained with hematoxylin and eosin using standard protocols. Histological sections were examined by microscope to determine the presence of inflammatory reaction, immune cell infiltration, and other pathological changes in tumor tissues after treatment.

Statistical analysis

All experiments were carried out in triplicate. One-way analysis of variance (ANOVA) followed by two-tailed Student's t test was used for statistical evaluation of differences. Results were represented as mean±SE, and the statistical significance was analyzed as described in each legend of the figures and table.


EGF R expression in Hepa-derived hepatoma cells

The EGF R expression in Hepa hepatoma cells was examined by immunohistochemical assay. The result in Figure 1 demonstrated that high EGF R expression was detected on the surface of Hepa hepatoma cells.

Figure 1

Immunohistochemical analysis of EGF R expression on Hepa hepatoma cells. Hepa cells were collected after trypsin digestion and fixed in 10% formaldehyde solution for 30 minutes. After washing with PBS, Hepa cells were dripped onto the holder plates. Standard immunohistochemistry assay was then performed after the cells were fixed on plates by heating at 55°C.

Transduction efficiency and targetability of GE7/β-gal polyplex

To observe the in vivo gene transfer efficiency and targetabilility of GE7 gene delivery system, the expression of a reporter gene, β-gal, was detected in hepatoma-bearing mice after peritumoral injection of GE7/β-gal polyplex. It was revealed that tumor tissues possessed a high level of transgene expression at 48 hours after injection and the transduction rate could reach over 60%, whereas tumor tissues from control animals injected with GE7/pCEP polyplex showed no β-gal expression after X-gal staining (Fig 2). We also examined the β-gal expression in different tissues and organs to evaluate the distribution of transgene in mice bodies after peritumoral injection of GE7/β-gal polyplex; the result showed that no reporter gene expression was detected in all 10 organs and tissues including heart, liver, spleen, lung, kidney, stomach, intestine, womb, muscle, as well as brain. All of these results indicated that peritumoral administration of the GE7 gene delivery system could target exogenous gene to EGF R positive tumor cells with high efficiency and targetability.

Figure 2

β-Gal expression in tumor tissues after the GE7 system–mediated in vivo gene transfer. Hepa hepatoma–bearing mice were injected peritumorally with the GE7/β-gal polyplex. Forty-eight hours later, animals were sacrificed and different organs and tissues were dissected, fixed, and stained with X-gal. Paraffin-embedded sections were counterstained with Nuclear Fast Red. Photos demonstrate the β-gal expression in tumor tissues. A: Tumor tissue from mice injected with the GE7/pCEP polyplex as control. B: Tumor tissue from mice injected with the GE7/β-gal polyplex.

Efficient expression of therapeutic genes in tumor tissues after in vivo targeted delivery by the GE7 system

Immunohistochemical and Western blot were carried out to investigate the expression status of p21WAF-1 and GM-CSF proteins in tumor tissues of hepatoma-bearing mice after peritumoral injection of GE7-PL/DNA/HA20-PL polyplex. The immunohistochemical examination in tumor tissues of the mice injected with GE7/p21WAF-1 polyplex and GE7/p21WAF-1/pGM-CSF polyplex revealed a characteristic staining of p21WAF-1 protein, which was not found in tumors of mice injected with GE7/pGM-CSF polyplex and in control vehicle (Fig 3). As shown in Figure 4, Western blot analysis demonstrated that specific bands corresponding to the molecular mass of 21 and 27 kDa were detected in tumors of mice injected peritumorally with GE7/p21WAF-1 polyplex and GE7/pGM-CSF polyplex, respectively, and both of these two bands could be seen in tumor tissues of mice injected with GE7/p21WAF-1/pGM-CSF polyplex. Although a light band related to 27 kDa was also found in tumor tissues of animals that received GE7/pCEP polyplex, GE7/p21WAF-1 polyplex, and normal saline injection, the levels of GM-CSF expression in these tissues were much less than that in GE7/pGM-CSF polyplex– and GE7/p21WAF-1/pGM-CSF polyplex–transduced tumor tissues. It was considered as the endogenous expression of GM-CSF in Hepa hepatoma tissues. Thus, therapeutic genes, including p21WAF-1 and pGM-CSF, could be transferred into EGF R–positive tumor tissues and be expressed efficiently in the target tissues after in vivo delivery by GE7 system.

Figure 3

Immunohistochemical assay for p21WAF-1 expression in tumor tissues transfected with the GE7-PL/DNA/HA20-PL polyplex. One month after injection with the GE7-PL/DNA/HA20-PL polyplex peritumorally, animals were sacrificed and tumor tissues were dissected for paraffin-embedded sections. Mouse anti–human p21WAF-1 monoclonal antibody was used to determine the expression of p21WAF-1 protein. Tumor tissue from mice injected with the GE7/pCEP polyplex (A), GE7/p21WAF-1 polyplex (B), GE7/pGM-CSF polyplex (C), and GE7/p21WAF-1/pGM-CSF polyplex (D).

Figure 4

Western blot analysis of p21WAF-1 and murine GM-CSF expression in tumor tissues transfected with the GE7-PL/DNA/HA20-PL polyplex. Hepatoma-bearing mice received peritumoral injection once a week for 1 month and tumors of the mice from different groups were isolated for Western blot analysis to examine the expression of p21WAF-1 and GM-CSF. Lanes 1 to 5 represent tumor tissue from mice injected with the GE7/pCEP polyplex, GE7/p21WAF-1 polyplex, GE7/p21WAF-1/pGM-CSF polyplex, GE7/pGM-CSF polyplex, and normal saline, respectively.

Enhanced antitumor effect of in vivo cotransfer of p21WAF-1 and GM-CSF genes delivered by the GE7 system

To determine the antitumor effect induced by the combination of p21WAF-1 and GM-CSF gene therapy, hepatoma-bearing mice were prepared by subcutaneous implantation with tumor cubes and injected with GE7-PL/DNA/HA20-PL polyplex peritumorally. Average tumor volumes on day 30 after tumor inoculation are illustrated in Table 1. It was shown that GE7 system–mediated gene transfer of either p21WAF-1 or GM-CSF alone could suppress tumor growth obviously when compared with mice in control groups (P<.05). The average tumor volume in mice injected with GE7/p21WAF-1 polyplex or GE7/pGM-CSF polyplex was 0.30 and 0.29 cm3, respectively, versus 0.64 cm3 in mice injected with GE7/pCEP polyplex as control. However, the tumor growth in the hepatoma-bearing mice injected with GE7/p21WAF-1/pGM-CSF polyplex was inhibited most significantly and the average tumor volume was only 0.10 cm3. Statistical differences in the tumor volume were achieved among the combined gene therapy and single gene utilization (P<.05). The observation of tumor volume continued to 3 months and we concluded that tumor growth was arrested completely in mice that received cotransfer of p21WAF-1 and GM-CSF genes (Fig 5A).

Table 1 Tumor Growth of Hepatoma-Bearing Mice After Different Treatments
Figure 5

Inhibitory effect of combined gene transfer of p21WAF-1 and GM-CSF on the tumor growth of hepatoma-bearing mice after peritumoral injection of the GE7/p21WAF-1/pGM-CSF polyplex. Tumor cubes (1 mm3) were inoculated subcutaneously in female C57BL/6 mice. Three days later, hepatoma-bearing mice were injected peritumorally with different polyplex in which the amount of DNA used was at a dose equivalent to 0.2 μg. Treatments were performed in separate groups of animals once a week for 3 months. Tumor volume was measured and expressed as mentioned in Materials and Methods. A: Average tumor volumes of mice in different groups on day 30 after inoculation (n=6 or 7). B: Survival period of hepatoma-bearing mice. *P<.05, **P<.01 indicates statistical significance as compared with GE7/pCEP polyplex control; +P<.05, ++P<.01 indicates statistical significance when the GE7/p21WAF-1/pGM-CSF polyplex group was compared with the other six groups.

Moreover, we observed the effect of peritumoral injection of GE7/p21WAF-1/pGM-CSF polyplex on the survival of hepatoma-bearing mice. As shown in Figure 5B, all mice in control groups injected with normal saline, GE7-PL/HA20-PL polyplex, GE7/pCEP polyplex, or plasmid DNA died within 3 months, whereas hepatoma-bearing mice that received cotransfer of p21WAF-1 and GM-CSF genes were able to survive much longer than the mice injected with GE7/p21WAF-1 and GE7/pGM-CSF polyplex alone. Seventy-one percent of hepatoma-bearing mice that received combined gene therapy, 33% of mice that received p21WAF-1 gene transfer, and 28% of mice that received GM-CSF gene transfer were found to be tumor free 90 days after treatment. Therefore, in vivo combined transfer of p21WAF-1 and GM-CSF genes exhibited more potent antitumor effect in the established hepatoma-bearing mice.

Induction of tumor cell apoptosis by in vivo p21WAF-1 gene transfer

To determine the major tumor-suppression mechanism by p21WAF-1 overexpression, DNA fragmentation pattern in genomic DNA in tumor tissues characteristic of apoptosis was examined by agarose gel electrophoresis analysis (Fig 6). A very low amount of ladder-type DNA could be seen in tumors of mice injected with GE7/pGM-CSF polyplex, whereas the GE7/pCEP polyplex control did not result in any DNA ladder typical for apoptotic cell death. However, tumors of mice injected with GE7/p21WAF-1 polyplex or GE7/p21WAF-1/pGM-CSF polyplex did show a considerable amount of DNA fragmentation along with DNA degradation. Therefore, it was suggested that gene transfer of p21WAF-1 by GE7 system might induce tumor cell death by apoptosis.

Figure 6

Detection of apoptosis in tumor tissues by DNA fragmentation analysis in hepatoma-bearing mice injected with the GE7-PL/DNA/HA20-PL polyplex. Hepa hepatoma–bearing mice were injected peritumorally with the GE7 complex once a week for a month. Genomic DNA of tumor tissues in mice from different groups were isolated and analyzed for the presence of nucleosomal DNA laddering. Lane 1 represents λ DNA/HindIII marker; lane 2, tumor tissues from mice injected with the GE7/pCEP polyplex; lane 3, tumor tissues from mice injected with the GE7/p21WAF-1 polyplex; lane 4, tumor tissues from mice injected with the GE7/pGM-CSF polyplex; lane 5, tumor tissues from mice injected with the GE7/p21WAF-1/pGM-CSF polyplex; and lane 6 represents pGZM-3f/HaeIII marker.

More significant necrosis and inflammatory infiltration in tumor microenvironment after in vivo cotransfer of p21WAF-1 and GM-CSF genes

In an effort to detect the histopathological changes of tumor mass after GE7-PL/DNA/HA20-PL treatment, the area of inflammatory or liquefactive necrosis was evaluated through microscopic examination of biopsy samples. Histopathological assay demonstrated that hemorrhage and necrosis were generally observed in tumor samples of mice peritumorally injected with GE7/p21WAF-1 polyplex, GE7/pGM-CSF polyplex, and GE7/p21WAF-1/pGM-CSF polyplex, especially in mice that received p21WAF-1 or combined gene treatment. In addition, more massive inflammatory infiltration was detected by pathological examination in the tumors of mice injected with GE7/pGM-CSF polyplex or GE7/p21WAF-1/pGM-CSF polyplex, but no statistical difference in inflammatory infiltration was found between the two groups. In contrast, the control tumor samples from the mice injected with GE7/pCEP polyplex revealed minimal extent of necrosis and inflammatory infiltration in their biopsies (Fig 7).

Figure 7

Histopathological analysis of Hepa hepatoma after peritumoral injection of the GE7-PL/DNA/HA20-PL polyplex. After hepatoma-bearing mice were injected with GE7/pCEP polyplex (A), GE7/p21WAF-1 polyplex (B), GE7/pGM-CSF polyplex (C), and GE7/p21WAF-1/pGM-CSF polyplex (D) once a week for 1 month, tumor tissues were prepared for paraffin-embedded sections and stained with hematoxylin/eosin.


It has been shown that regional administration of vectors encoding therapeutic proteins or local transfer of therapeutic genes could achieve potent antitumor effects.27 For the gene delivery system, virus vectors were commonly used to introduce exogenous genes into tumor cells for cancer gene therapy; however, the disadvantages were transient expression of transduced genes, hepatocyte injury, and immune rejection.28 For seeking an effective gene delivery system with high targetability and security, we developed a tumor-targeted gene transfer system composing of a synthetic oligopeptide GE7 used as a ligand for EGF R binding and oligopeptide HA20 for endosomolysis. Our previous work has proved that the GE7 system could transfer exogenous genes into SMMC-7721 and BEL-7402 human hepatoma cell lines in vitro and also into subcutaneously transplanted human hepatoma in nude mice after peritumoral injection. The mechanism of GE7-mediated gene delivery into target cells was studied by means of the transmission and scanning electron microscopes, fluorescence microscope, etc. It was shown that the membrane entrance process was mediated by binding with the EGF R. Various amounts of EGF at the molar ratio of 5-, 10-, 15-, and 20-fold excess over GE7 was used in the competitive binding experiment, and the result indicated that the GE7 has high affinity with its receptor. In addition, the biodistribution of the GE7 DNA polyplex, when injected through the tail vein, was studied in another part of our studies. Results suggested that the GE7 gene delivery system could target exogenous genes into EGF R–expressing tumor tissues with high efficiency within 12 hours, except for transient and low expression in a few organs, such as liver, spleen, and lung (data not shown). Thus, our previous work has proven that the GE7 gene delivery system targets EGF R.2

In this study, an immune-competent mouse model was required to explore the immune response induced by GM-CSF gene transfer. Murine Hepa hepatoma cell line was confirmed with high EGF R expression and strong tumorigenicity in vivo. First, we demonstrated the successful gene delivery into subcutaneously transplanted tumors by peritumoral injection of GE7/β-gal polyplex. The result indicated that high β-gal expression could be detected in tumor tissues and the expression persisted for more than 15 days. No transgene products were found in 10 organs and tissues including heart, liver, spleen, lung, kidney, brain, stomach, intestine, womb, and muscle tissue, which implied that the GE7 system could mediate a high level of targetable expression of exogenous genes in target tumor tissues after peritumoral injection.

The purpose of our study was to find a therapeutic agent that could arrest the growth of cancer cells but have no apparent cytotoxic effect on normal nonreplicating cells. Recently, some experiments have been conducted in many tumor models to test the suitability of CKIs for gene therapy.29,30,31 Our study was performed to compare the effectiveness of different CKIs in the suppression of tumor growth. We proved that p21WAF-1 gene transfer was as effective as p53 gene transfer and more potent than p15 or p16 (data not shown). Alternatively, GM-CSF is a hematopoietic growth factor that has been reported to have significant effects on the induction of antitumor immune response by increasing MHC molecule expression, enhancing dendritic cell maturation, inducing localized inflammation, and exerting immune modulatory effects by its systemic actions on the cytokine network.32 Here, we present a new protocol for cancer gene therapy based on the combined transfer of the p21WAF-1 gene, which was expected to reduce the growth of tumor, and the GM-CSF gene, which was proposed to improve the antitumor immunity to achieve an enhanced therapeutic effect.

We once demonstrated that when Hepa cells were transfected with GE7/pGM-CSF polyplex or GE7/p21WAF-1/pGM-CSF polyplex, GM-CSF production by transfected Hepa cells could reach up to 20 ng/mL per 106 cells (data not shown). In this study, Western blot assays were performed to evaluate the in vivo expression status of the therapeutic proteins in tumor tissues after peritumoral injection of therapeutic genes. The results indicated that exogenous genes p21WAF-1 and GM-CSF delivered by GE7 system could be transferred into the target tissues and be expressed.

The transfer of p21WAF-1 and GM-CSF genes in hepatoma-bearing mice by peritumoral injection of the GE7-PL/DNA/HA20-PL polyplex showed that the antitumor effect on pre-established tumor was enhanced more remarkably by the combined use of the two therapeutic genes than the transfer of a single gene. A much higher percentage of hepatoma-bearing mice receiving combined gene therapy was tumor free when compared with hepatoma-bearing mice injected with either GE7/p21WAF-1 or GE7/pGM-CSF polyplex alone. It was clearly shown that the GE7 nonviral delivery system–mediated combined gene therapy of p21WAF-1 and GM-CSF generated an enhanced antitumor effect.

The mechanism of cell death due to a high level of p21WAF-1 protein expression has not been reported to be apoptosis. p21WAF-1 gene is known as a general inhibitor of G1/S CDKs and can suppress tumor growth by regulating cell cycle progression33. However, in this study, a typical pattern of DNA ladder characteristic of apoptosis was discovered in both GE7/p21WAF-1– and GE7/p21WAF-1/pGM-CSF–transduced tumor DNA. Similar results were obtained in in vitro experiments that Hepa cells transfected with the p21WAF-1 gene or cotransfer of p21WAF-1 and GM-CSF genes exhibited apoptosis when examined by Hoechst 33 258 staining and TUNEL assay (data not shown). Further study of DNA content by cell cycle assay found S phase arrest in transfected cells in vitro. It was known that the uncoupling between mitosis and S phase might induce cells culminating in polyploid and thereby undergoing apoptosis,34 which might be the interpretation of the subsequent observation of DNA ladder related to DNA degradation in p21WAF-1–transduced tumor tissues. For explanation, we postulate that the GE7 oligopeptide gene delivery system might contribute to the apoptosis induced by p21WAF-1 gene transfer because GE7 oligopeptide is the binding domain of the EGF macromolecule and could bind and possibly activate EGF R, and thus could provide a message of cell cycle progression, whereas p21WAF-1 gene transfer results in cell cycle arrest. The clash of signals from opposite biological responses would probably trigger apoptosis.35,36,37 However, the postulated mechanisms of apoptosis induced by p21WAF-1 gene transfer remains to be further studied.

As an immunomodulator, GM-CSF can activate macrophages, neutrophils, and eosinophils to lyse tumor cells directly or to mediate ADCC response.35 Our data showed that necrosis or hemorrhage was much more frequently seen in tumors of mice injected peritumorally with GE7/p21WAF-1 polyplex, GE7/pGM-CSF polyplex, or GE7/p21WAF-1/pGM-CSF polyplex than in tumors of mice injected with GE7/pCEP control. Moreover, the obvious difference in the inflammatory response was observed among tumors of mice injected with or without GE7/pGM-CSF polyplex. Results here are consistent with the observations of several other groups using GM-CSF gene transfection.38,39,40,41,42 Although the antitumor immune response induced by GM-CSF was not examined in this study, in another part of our studies, we showed the transfection of tumor cells with the GM-CSF gene could enhance the expression of MHC I and B7-1 molecules on the tumor cells, and when the two genes were cotransferred via tail vein into hepatoma-bearing mice, cytotoxic activities of NK and CTL were increased, suggesting that combined gene therapy of p21WAF-1 and GM-CSF might enhance the specific and nonspecific antitumor immunity (data not shown).

In conclusion, a new strategy was designed for in vivo tumor-directed gene therapy mediated by an EGF R–targeted nonviral gene delivery system GE7, which has the ability to transfer therapeutic genes to EGF R–expressing tumor tissues. The peritumoral injection of GE7/DNA polyplex results in the transduction of exogenous genes with high efficiency and targetability. It was encouraging that the GE7 gene delivery system–mediated cotransfer of p21WAF-1 and GM-CSF genes could elicit more potent antitumor effect in the pre-established hepatoma-bearing mice than transfer of two therapeutic genes alone.


  1. 1

    Cristiano RJ, Roth JA . Epidermal growth factor mediated DNA delivery into lung cancer cells via the epidermal growth factor receptor Cancer Gene Ther 1996 3: 4–10

  2. 2

    Foster BJ, Kern JA . HER2-targeted gene transfer Hum Gene Ther 1997 8: 719–727

  3. 3

    Frederiksen KS, Abrahamsen N, Cristiano RJ et al. Gene delivery by an epidermal growth factor/DNA polyplex to small cell lung cancer cell lines expressing low levels of epidermal growth factor receptor Cancer Gene Ther 2000 7: 262–268

  4. 4

    Mohr L, Yoon SK, Eastman SJ et al. Cationic liposome-mediated gene delivery to the liver and to hepatocellular carcinomas in mice Hum Gene Ther 2001 12: 799–809

  5. 5

    Tian PK, Ren SJ, Ren CC et al. A novel receptor-targeted gene delivery system for cancer gene therapy Sci China Ser C Life Sci 1999 42: 216–224

  6. 6

    Li JM, Han JS, Huang Y et al. A novel gene delivery system targeting cells expressing VEGF receptors Cell Res 1999 9: 11–25

  7. 7

    Hamazaki K, Yunoki Y, Tagashira H et al. Epidermal growth factor receptor in human hepatocellular carcinoma Cancer Detect Prev 1997 21: 355–360

  8. 8

    Dilloo D, Bacon K, Holden W et al. Combined chemokine and cytokine gene transfer enhances antitumor immunity Nat Med 1996 2: 1090–1095

  9. 9

    Roth JA, Swisher SG, Meyn RE . p53 tumor suppressor gene therapy for cancer Oncology 1999 13: 148–154

  10. 10

    Ju DW, Tao Q, Cheng DS et al. Adenovirus-mediated lymphotactin gene transfer improves therapeutic efficacy of cytosine deaminase suicide gene therapy in established murine colon carcinoma Gene Ther 2000 7: 329–338

  11. 11

    Rogulski KR, Zhang K, Kolozsvary A et al. Pronounced antitumor effects and tumor radiosensitization of double suicide gene therapy Clin Cancer Res 1997 3: 2081–2088

  12. 12

    Aruga A, Tanigawa K, Aruga E et al. Enhanced adjuvant effect of granulocyte–macrophage colony-stimulating factor plus interleukin-12 compared with either alone in vaccine-induced tumor immunity Cancer Gene Ther 1999 6: 89–95

  13. 13

    Sandig V, Brand K, Herwig S et al. Adenovirally transferred p16INK4/CDKN2 and p53 genes cooperate to induce apoptotic tumor cell death Nat Med 1997 3: 313–319

  14. 14

    Putzer BM, Bramson JL, Addison CL et al. Combination therapy with interleukin-2 and wild-type p53 expressed by adenoviral vectors potentiates tumor regression in a murine model of breast cancer Hum Gene Ther 1998 9: 707–718

  15. 15

    Cao X, Huang X, Ju DW et al. Enhanced antitumoral effect of adenovirus-mediated cytosine deaminase gene therapy by induction of antigen-presenting cells through stem cell factor/granulocyte–macrophage colony-stimulating factor gene transfer Cancer Gene Ther 2000 7: 177–186

  16. 16

    McKenzie PP, Guichard SM, Middlemas DS et al. Wild-type p53 can induce p21 and apoptosis in neuroblastoma cells but the DNA damage-induced G1 checkpoint function is attenuated Clin Cancer Res 1999 5: 4199–4207

  17. 17

    Joshi US, Chen YQ, Kalemkerian GP et al. Inhibition of tumor cell growth by p21WAF1 adenoviral gene transfer in lung cancer Cancer Gene Ther 1998 5: 183–191

  18. 18

    Ren CC, Tian PK, Qu SM et al. Expression of cyclin-dependent kinase inhibitor genes induces apoptosis in human hepatoma cell line Chin Sci Bull 1997 42: 2000–2005

  19. 19

    Dranoff G, Jaffee E, Lazenby A et al. Vaccination with irradiated tumor cells engineered to secrete murine granulocyte–macrophage colony-stimulating factor stimulates potent, specific, and long-lasting anti-tumor immunity Proc Natl Acad Sci USA 1993 90: 3539–3543

  20. 20

    Burger JA, Baird SM, Powell HC et al. Local and systemic effects after adenoviral transfer of the murine granulocyte–macrophage colony-stimulating factor gene into mice Br J Haematol 2000 108: 641–652

  21. 21

    Cao X, Ju DW, Tao Q et al. Adenovirus-mediated GM-CSF gene and cytosine deaminase gene transfer followed by 5-fluorocytosine administration elicit more potent antitumor response in tumor-bearing mice Gene Ther 1998 5: 1130–1136

  22. 22

    Cao X, Zhang W, Wang J et al. Therapy of established tumour with a hybrid cellular vaccine generated by using granulocyte–macrophage colony-stimulating factor genetically modified dendritic cells Immunology 1999 97: 616–625

  23. 23

    Ju DW, Cao X, Acres B . Intratumoral injection of GM-CSF gene encoded recombinant vaccinia virus elicits potent antitumor response in a mixture melanoma model Cancer Gene Ther 1997 4: 139–144

  24. 24

    Murata M, Kagiwada S, Takahashi S et al. Membrane fusion induced by mutual interaction of the two charge-reversed amphiphilic peptides at neutral pH J Biol Chem 1991 266: 14353–14358

  25. 25

    Su W, Yeong KF, Spencer J . Immunohistochemical analysis of human CD5 positive B cells: mantle cells and mantle cell lymphoma are not equivalent in terms of CD5 expression J Clin Pathol 2000 53: 395–397

  26. 26

    Burnette WN . “Western blotting”: electrophoretic transfer of proteins from sodium dodecyl sulfate — polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A Anal Biochem 1981 112: 195–203

  27. 27

    Johnson DH . Locally advanced, unresectable non-small cell lung cancer: new treatment strategies Chest 2000 117: 123S–126S

  28. 28

    Ghosh SS, Takahashi M, Thummala NR et al. Liver-directed gene therapy: promises, problems and prospects at the turn of the century J Hepatol 2000 32: 238–252

  29. 29

    Chow NH, Tzai TS, Cheng HL et al. The clinical value of p21WAF1/CIP1 expression in superficial bladder cancer Anticancer Res 2000 20: 1173–1176

  30. 30

    Rich JN, Zhang M, Datto MB et al. Transforming growth factor-beta–mediated p15(INK4B) induction and growth inhibition in astrocytes is SMAD3-dependent and a pathway prominently altered in human glioma cell lines J Biol Chem 1999 274: 35053–35058

  31. 31

    Kobayashi S, Shirasawa H, Sashiyama H et al. P16INK4a expression adenovirus vector to suppress pancreas cancer cell proliferation Clin Cancer Res 1999 5: 4182–4185

  32. 32

    Warren TL, Weiner GJ . Uses of granulocyte–macrophage colony-stimulating factor in vaccine development Curr Opin Hematol 2000 7: 168–173

  33. 33

    Hunter T, Pines J . Cyclins and cancer. II: cyclin D and CDK inhibitors come of age Cell 1994 79: 573–582

  34. 34

    Waldman T, Zhang Y, Dillehay L et al. Cell-cycle arrest versus cell death in cancer therapy Nat Med 1997 3: 1034–1036

  35. 35

    Lowe SW, Ruley HE, Jacks T, Housman DE . p53-dependent apoptosis modulates the cytotoxicity of anticancer agents Cell 1993 74: 957–967

  36. 36

    Symonds H, Krall L, Remington L et al. p53-dependent apoptosis suppresses tumor growth and progression in vivo Cell 1994 78: 703–711

  37. 37

    Fisher DE . Apoptosis in cancer therapy: crossing the threshold Cell 1994 78: 539–542

  38. 38

    Hogge GS, Burkholder JK, Culp J et al. Preclinical development of human granulocyte–macrophage colony-stimulating factor–transfected melanoma cell vaccine using established canine cell lines and normal dogs Cancer Gene Ther 1999 6: 26–36

  39. 39

    Mach N, Gillessen S, Wilson SB et al. Differences in dendritic cells stimulated in vivo by tumors engineered to secrete granulocyte–macrophage colony-stimulating factor or Flt3-ligand Cancer Res 2000 60: 3239–3246

  40. 40

    Chang AE, Li Q, Bishop DK et al. Immunogenetic therapy of human melanoma utilizing autologous tumor cells transduced to secrete granulocyte–macrophage colony-stimulating factor Hum Gene Ther 2000 11: 839–850

  41. 41

    Hiltbold EM, Alter MD, Ciborowski P et al. Presentation of MUC1 tumor antigen by class I MHC and CTL function correlate with the glycossylation state of the protein taken Up by dendritic cells Cell Immunol 1999 194: 143–149

  42. 42

    Tamada K, Shimozaki K, Chapoval AI et al. Modulation of T-cell–mediated immunity in tumor and graft-versus-host disease models through the LIGHT co-stimulatory pathway Nat Med 2000 6: 283–289

Download references


We thank Dr. Jingjun Li for pathological section preparation and immunohistochemistry analysis and Dr. Dian-Wen Ju for his helpful discussion. This work was supported by a grant of Biotechnology Project, National High Technology Program of China (Project No. Z-20-01-01).

Author information

Correspondence to Xuetao Cao or Jianren Gu.

Rights and permissions

Reprints and Permissions

About this article


  • nonviral vector
  • p21WAF-1
  • GM-CSF
  • gene therapy
  • hepatoma

Further reading