Adenovirus type 5 E1A-induced apoptosis in COX-2-overexpressing breast cancer cells

Abstract

Introduction: Suppression of Bcl-2 expression can overcome cellular resistance to apoptosis induced by the adenovirus type 5 gene E1A in models of ovarian and breast cancer. Celecoxib, a cyclooxygenase-2 (COX-2) inhibitor, is known to downregulate Bcl-2 expression. We hypothesized that celecoxib would enhance E1A-induced apoptosis by suppressing Bcl-2 through suppressing COX-2 expression. If successful, this strategy could represent a means of overcoming resistance to E1A gene therapy.

Methods: We first established the cytotoxicity of celecoxib in two COX-2-overexpressing E1A-transfected breast cancer cell lines (MDA-MB-231 and MDA-MB-435) and in two low-COX-2-expressing E1A-transfected cell lines (MCF-7 (breast cancer) and SKOV3.ip1 (ovarian cancer)). We next tested whether higher sensitivity to celecoxib among these cell lines resulted from increased apoptosis by flow cytometry and western blotting. We further investigated whether suppression of Bcl-2 by celecoxib was involved in the apoptosis resulting from celecoxib treatment, and we explored whether the celecoxib-induced apoptosis in these cells depends on a COX-2 downstream pathway.

Results: The two COX-2-overexpressing cell lines MDA-MB-231-E1A and MDA-MB-435-E1A were more sensitive to celecoxib than the corresponding control cells, but the two low-COX-2-expressing cell lines MCF-7-E1A and SKOV3.ip1-E1A were no more sensitive than control cells to celecoxib. Therefore, we used the MDA-MB-231-E1A and MDA-MB-435-E1A cells for all further experiments. In both cell lines, sub-G1 fraction was increased, or cleavage of PARP and caspase-9 were increased after 5 days of exposure to 40 microM celecoxib. However, Bcl-2 was suppressed only in the MDA-MB-435-E1A cells and not in the MDA-MB-231-E1A cells. Restoring Bcl-2 expression in the MDA-MB-435-E1A stable transfectants did not affect their sensitivity to celecoxib. However, adding prostaglandin E2 (PGE2) or PGF2alpha blunted the sensitivity to celecoxib of both E1A stable transfectants.

Conclusion: We speculate that one mechanism by which celecoxib enhances E1A-induced apoptosis in cells that express high levels of COX-2 is through blocking PGE2 or PGF2alpha.

Figure 1

COX-2 protein expression in breast and ovarian cancer cell lines stably transfected with E1A. (a) MDA-MB-231-E1A stable transfectants produced the greatest amounts of COX-2; MDA-MB-435-E1A cells produced 78% of that amount, but the SKOV3.ip1-E1A and MCF-7-E1A cells produced only 6% of that amount. MDA-MB-231-E1A cells expressed slightly more E1A than the other three cell lines. (a) COX-2 protein expression level between the E1A stable transfectants and their corresponding vector control cells or parent cells. If the COX-2 expression levels of each E1A transfectant is defined as 100%, the corresponding COX-2 expression levels of the vector controls were as follows: 65% for MDA-MB-231, 144% for MDA-MB-435, 71% for SKOV3.ip1 and 67% for MCF-7.

Figure 2

Celecoxib decreases the viability of MDA-MB-231-E1A and MDA-MB-435-E1A stable transfectants. (b) MTT assays after a 5-day exposure to 0–60 μM celecoxib indicate substantial reductions in cell viability in all variants of the MDA-MB-231 and MDA-MB-435 cell lines (E1A stable transfectants, vector control, and parental control cells). Each point represents means from tests performed in quadruplicate; the bars are standard deviations. In both cell lines, the E1A transfectants were more sensitive than the vector control or parental control cells. (b) MTT assays in all variants of the SKOV3.ip1 and MCF-7 cell lines. The E1A transfectants have no difference to sensitivity for celecoxib in both cell lines. (c) Trypan blue assays after a 5-day exposure to 40 μM celecoxib show substantial reductions in viability of MDA-MB-231 and MDA-MB-435 cells. Values shown are normalized to the viability of the control (untreated) cells. Each bar represents means from tests performed in quadruplicate; bars are standard deviations. P values are from two-tailed paired t tests.

Figure 3

Celecoxib enhances apoptosis of MDA-MB-231-E1A and MDA-MB-435-E1A stable transfectants. (a) Cell cycle distribution of MDA-MB-231-E1A and MDA-MB-435-E1A cells was detected by fluorescence-activated cell sorting after a 5-day exposure to 0 or 40 μM celecoxib. The percentage of cells in sub-G₁ (apoptosis) appears at the upper right of each graph. (b) Western blots of MDA-MB-231 and MDA-MB-435 cells treated with 0 or 40 μM celecoxib for 5 days and tested for cleaved caspase-9 (cl-cas-9), uncleaved and cleaved caspase-8 (cl-cas-8), PARP (uncleaved and cleaved), E1A, and actin. Cleaved PARP and cleaved caspase-9 levels were higher after celecoxib treatment in the MDA-MB-231-E1A and MDA-MB-435-E1A stable transfectants, but expression of cleaved caspase-8 (cl-cas-8) did not change.

Figure 4

Celecoxib downregulated COX-2 protein expression in all MDA-MB-231 and MDA-MB-435 variants, but celecoxib downregulated Bcl-2 expression in only the MDA-MB-435-E1A stable transfectants. (a) Western blots of MDA-MB-231 and MDA-MB-435 cells treated with 0 or 40 μM celecoxib for 5 days and tested for COX-2 and Bcl-2. Percentages indicate differences relative to the 0 μM control samples. Protein expression was considered to be downregulated if the treated condition was at least 20% less than the control (untreated) condition. (b) Time course of Bcl-2 expression after treatment with 0 or 40 μM celecoxib in MDA-MB-435-E1A and MDA-MB-231-E1A stable transfectants. Bcl-2 was suppressed at both 72 and 96 h in the MDA-MB-435-E1A stable transfectants but was not suppressed in the MDA-MB-231-E1A stable transfectants. (c) Transfection of MDA-MB-435 cells with Bcl-2 DNA (+) or a control DNA (-) led to overexpression of Bcl-2 in all variants. (d) MDA-MB-435-E1A cells made to overexpress Bcl-2 and non-Bcl-2-overexpressing cells were treated with 0 or 40 μM celecoxib for 5 days, and cell viability was determined with a trypan-blue assay. Bcl-2 overexpression did not restore sensitivity to celecoxib (P = 0.11).

Figure 5

Celecoxib-induced apoptosis of MDA-MB-231-E1A and MDA-MB-435-E1A cells depends on PGE₂ or PGF_2α. Treatment of MDA-MB-231-E1A and MDA-MB-435-E1A stable transfectants with 0 or 40 μM celecoxib (CLX) plus 10 μM of either PGE₂or PGF_2α for 5 days blunted sensitivity to celecoxib in both cell lines.