Inhibition of Angiogenesis and Induction of Apoptosis Are Involved in E1A-mediated Bystander Effect and Tumor Suppression

Introduction

The bystander effect was originally identified through a study in which cancer cells modified by the herpes simplex virus thymidine kinase gene caused the death of nearby unmodified cancer cells after administration of the drug ganciclovir (1) . In this case, gap junctional intercellular communication was the mechanism responsible for bystander killing (2) . However, recent studies have shown that tumor cells transfected with genes such as IFN-β, interleukin 4, and p53 exhibit bystander effects with alternative mechanisms, including enhancement of the immune response and suppression of angiogenesis (3, 4, 5) , which has been shown to play an important role in tumor growth and metastasis (6 , 7) .

Ad5 3 E1A has recently been demonstrated to exhibit antioncogenic activity by suppressing transformation, metastasis, and tumorigenicity via multiple mechanisms (8, 9, 10, 11, 12, 13) . The molecular mechanisms that may be associated with this antioncogenic activity include down-regulation of transforming and metastasis genes such as HER-2/neu, Axl, stromelysin, plasminogen activator, urokinase, and type 4 collagenase (8 , 14) .

E1A has also been shown to increase the susceptibility of cells to tumor necrosis factor and γ-radiation (15 , 16) , enhance their susceptibility to host cellular immune responses, and sensitize them to apoptosis induced by various stimuli such as high cell density. E1A can also induce apoptosis in a cell-free system and in both p53-dependent and p53-independent manners (17, 18, 19) . Moreover, local and systemic liposome- and adenovirus-mediated delivery of E1A results in significant tumor suppression in cell lines and mouse models (10 , 11) . Current methodology for E1A transfection is not 100% successful; hence, we raised the question as to whether E1A-mediated tumor suppression might be due in part to a bystander effect. To test this possibility, we investigated the bystander effect on tumor development in nude mice inoculated with Ad5 E1A transfectant cells.

Materials and Methods

Cell Lines and Culture.

Cell lines used were parental human ovarian cancer line SKOV3.ip1, stable Ad5 E1A transfectants of SKOV3.ip1 (ip1-E1A2), E1A frameshift mutant transfectants of SKOV3.ip1 (ip1-Efs), and the human breast cancer cell line MDA-MB-435 (American Type Culture Collection, Manassas, VA). All cultures were established and grown as described by Shao et al. (16) .

Tumor Cell Inoculation.

Female 4–5-week-old nude mice were purchased from Harlan Bioproducts for Science, Inc. (Indianapolis, IN) and maintained in the animal facility of The University of Texas M. D. Anderson Cancer Center. Mice were divided into six experimental groups containing five mice each. One group was inoculated s.c. with a cell mixture of 2× 10⁶ ip1-E1A2 cells and an equal amount of parental SKOV3.ip1 cells to induce tumor. A second group was inoculated with 1 × 10⁶ ip1-E1A2 cells and an equal amount of MDA-MB-435 cells. The remaining groups were inoculated with equal numbers of SKOV3.ip1, MDA-MB-435, or ip1-E1A2 cells alone. Cells were injected into the flanks of the mice, and each mouse was inoculated with tumors at two sites. Tumor volumes were measured and recorded once a week for 4–7 weeks. Each experiment was repeated at least one time.

Immunohistochemistry.

Histological sections of ip1-E1A2/SKOV3.ip1-, ip1-E1A2-, and SKOV3.ip1-induced tumors were taken, fixed with formalin, and embedded in paraffin. The sections were incubated with rabbit antibody against factor VIII (1:200; Dako Corp., Carpinteria, CA), followed by incubation with biotinylated goat antirabbit IgG (1:200; Vector Laboratories, Inc., Burlingame, CA). Sections were then incubated with an avidin-biotin complex-horseradish peroxidase conjugate and developed using aminoethylcarbazole chromogen (Sigma, St. Louis, MO) as a substrate. Positive signals were visualized by light microscopy at high power (×400), and the number of microvessels was counted in six ×200 fields for each tumor section.

Tumor Apoptosis Assay.

The sections from ip1-E1A2/SKOV3.ip1-, ip1-E1A2-, and SKOV3.ip1-induced tumors were taken and analyzed for apoptosis by TUNEL assay as described by Shao et al. (15) .

[³H]Thymidine Incorporation.

Cells were cocultured in a 24-well Transwell unit containing an upper and a lower chamber (Costar Corporate, Cambridge, MA). The chambers were separated by a membrane that allowed the medium, but not the cells, to penetrate freely. The upper chamber was filled with 2 × 10⁴ ip1-E1A2 cells or ip1-Efs, SKOV3.ip1, or MDA-MB-435 control cells in 100 μl of medium. The lower chamber was filled with 1.5 × 10⁴ SKOV3.ip1 or MDA-MB-435 cells in 600 μl of medium. The supernatant from the upper chamber, which was replaced with 100 μl of fresh medium, was filtered and used to replace 100 μl of the medium from the lower chamber each day. After 72 h of culture, 3 μCi of [³H]thymidine in 20 μl of medium were added into the lower chamber, and incubation was continued for another 16 h. Thymidine incorporation was determined with a scintillation counter.

In Vitro Apoptosis Assay.

To analyze apoptosis in bystander cells induced by E1A transfectants in culture, we cultured 3 × 10⁵ ip1-E1A2, ip1-Efs, or SKOV3.ip1 cells in the upper chamber and 1.5 × 10⁴ SKOV3.ip1 cells in the lower chamber of the Transwell unit under serum starvation conditions. After 96 h of culture, cells were trypsinized and harvested for TUNEL assay as described by Shao et al. (15) .

Results and Discussion

E1A has exhibited multiple tumor-suppressive properties (8) . To test whether E1A could mediate a bystander effect on tumor development, a human ovarian cancer cell line derivative (SKOV3.ip1), the wild-type Ad5 E1A stable transfectant of SKOV3.ip1 (ip1-E1A2), the human breast cancer cell line MDA-MB-435, and ip1-E1A2/SKOV3.ip1 and ip1-E1A2/MDA-MB-435 mixtures were used to inject tumors ectopically in nude mice. Tumors induced by the cell mixtures or ip1-E1A2 cells alone grew more slowly than those induced by either SKOV3.ip1 or MDA-MB-435 cells alone (Fig. 1, A and B ⇓ ). These results, consistent with previous results from our laboratory (8 , 10 , 11) , indicated that E1A mediates tumor suppression. In addition, data showed that ip1-E1A2 cells mediated a bystander effect by suppressing tumor growth induced by the SKOV.3 ip1 or MDA-MB-435 cells.

Recent studies have shown that angiogenesis is one of the important factors affecting tumor growth. Factor VIII, which presents on the blood vessel endothelial cells, has been used as a marker for angiogenesis (6) . To examine whether E1A-mediated tumor suppression involves a change in angiogenesis and whether the change is involved in the bystander effect, we performed immunohistochemical staining for factor VIII in tumor tissue. Discrete microvessel counts and neovascularization in tumors induced by ip1-E1A2/SKOV3.ip1 cell mixtures (Fig. 2A ⇓ ) and ip1-E1A2 cells alone (Fig. 2B ⇓ ) were dramatically lower than those in tumors induced by SKOV3.ip1 cells alone (Fig. 2C ⇓ ). These results suggest that E1A mediates the antiangiogenic suppression process of blood vessel formation and that this action is involved in the bystander effect.

In addition to angiogenesis, apoptosis has been implicated in cancer pathogenesis and treatment (20) and as a mechanism for bystander cell death (2) . E1A has been shown to induce apoptosis in vitro; therefore, to determine whether E1A could mediate apoptotic changes in vivo, we analyzed tumor sections by TUNEL assay. Tumors induced by ip1-E1A2/SKOV3.ip1 cell mixtures and by ip1-E1A2 cells alone exhibited apoptosis in 7% and 9% of the cells examined, respectively (Fig. 3, A and B ⇓ ), whereas tumors induced by SKOV3.ip1 cells showed apoptosis in <1% of cells examined (Fig. 3C ⇓ ). These results indicate that E1A mediates apoptosis not only in tumors induced by the E1A transfectants but also in tumors induced by the cell mixtures.

It has been shown that cytokines or other factors secreted by transfected cells can induce bystander effects (3, 4, 5) . To further investigate whether our bystander effect was mediated by a factor secreted from ip1-E1A2 cells, we cocultured SKOV3.ip1 or MDA-MB-435 cells with ip1-E1A2 or control ip1-Efs cells in a Transwell unit. Their mitogenic activity was measured by[ ³H]thymidine incorporation. The mitogenic activity of SKOV3.ip1 or MDA-MB-435 cells cocultured with ip1-E1A2 cells was lower than that of SKOV3.ip1 or MDA-MB-435 cells cocultured with ip1-Efs cells, which do not produce functional E1A proteins (Fig. 4, A and B ⇓ ). This indicated that ip1-E1A2 cells could mediate inhibition of mitogenic activity of SKOV3.ip1 and MDA-MB-435 cells and that the effect was caused by a secretable factor(s) present in the ip1-E1A2 cell supernatant. To further examine whether ip1-E1A2 cells could secrete this factor(s) to induce apoptosis in vitro, we cultured SKOV3.ip1 cells alone or with either ip1-E1A2 or control ip1-Efs cells in Transwell units under serum starvation conditions to induce apoptosis. Twenty percent of SKOV3.ip1 cells underwent apoptosis when cultured with ip1-E1A2 cells as compared with 5% and 3% of SKOV3.ip1 cells cultured with ip1-Efs cells or alone, respectively (Fig. 4C ⇓ ). These results indicate that cells transfected with E1A mediated a bystander effect by releasing a secretable factor to inhibit mitosis and enhance apoptosis through the culture supernatant under serum withdrawal conditions. Taken together, our results suggest that E1A-mediated inhibition of angiogenesis, induction of apoptosis, and suppression of mitosis may contribute to the bystander effect on E1A-mediated tumor suppression.