Photodynamic Therapy Sensitivity Is Not Altered in Human Tumor Cells after Abrogation of p53 Function¹

The tumor suppressor gene p53 encodes a protein that functions as a transcription factor involved in regulating genes that control the cell cycle and apoptosis (1). An increasing number of studies describe altered cell sensitivity to chemotherapy or radiation therapy when p53 is mutated. A positive correlation is observed between wt³ p53 expression and apoptotic cell death induced by chemotherapy or radiation (2, 3, 4). However, wt p53 expression cannot universally be considered a positive predictor of radiation or chemosensitivity because negative correlations also exist. For example, human breast and colon carcinoma cell lines have equivalent sensitivity to ionizing radiation, Adriamycin, and etoposide but have increased cisplatin sensitivity after the disruption of wt p53 expression (5). Abrogation of wt p53 with transfected papillomavirus type 16 oncoprotein is associated with increased radiation resistance in human diploid fibroblasts, whereas transfected human lymphoblasts show delayed apoptosis but no alteration in radiation sensitivity (6, 7). Differences in cell type seem to be one determinant affecting whether p53 plays a role in modulating treatment response. A number of clinical studies continue to suggest that mutated p53 is a prognostic factor associated with treatment resistance and poor prognosis in patients with hematological, breast, and colorectal cancers (8, 9).

PDT is a treatment for solid malignancies using tissue-penetrating visible light after the administration of a tumor-localizing photosensitizer (10). This procedure shows considerable promise for tumors of the bronchus, esophagus, breast, bladder, colon, and brain, and many of these malignancies present with mutated and/or nonfunctioning p53 (10, 11, 12). Minimal information is available defining PDT responsiveness as a function of p53 status. Photosensitivity of HL60 human promyelocytic leukemia cells (which are p53 null) increased in a transfectant cell line expressing wt p53 (13). These p53-expressing transfectant cells have a concomitant down-regulated BCL-2 expression, and this agrees with a report that elevated BCL-2 expression induces PDT resistance (14). Recently, we demonstrated that a human colon carcinoma cell line with a wt p53 phenotype (LS513) exhibits increased PDT sensitivity when compared with a mutated p53 colon carcinoma cell line (LS1034; Ref. 15). Cellular differences unrelated to p53 expression could certainly play a role in the differential photosensitivity observed in these nonisogeneic colon carcinoma cell lines.

The aim of the current study was to determine whether selective abrogation of p53 function in cell lines derived from solid tumors affects cellular sensitivity to PDT-mediated oxidative stress. We used established human colon (LS513) and breast (MCF-7) carcinoma cell lines with normal p53 function and produced stable clones from these parental cell lines in which p53 function was disrupted by transfection with the HPV-16 E6 gene. The viral oncoprotein E6 specifically targets p53 for ubiquitin degradation and disrupts the transactivation function of this transcription factor (16). Using these isogeneic cells, we were able to investigate the effect of E6 expression on photosensitivity as well as on PDT-induced apoptosis, cell cycle alterations, and p53-associated gene expression.

Human colorectal carcinoma cells (LS513) were purchased from American Type Culture Collection (Rockville, MD; CRL 2134). The cells were grown as a monolayer in RPMI 1640 supplemented with 10% FCS and 1% penicillin/streptomycin and had a plating efficiency of 55%. Human breast carcinoma cells (MCF-7) were a gift from Dr. A. Masumder (University of Southern California, Los Angeles, CA). They were grown as a monolayer in RPMI 1640 supplemented with 10% FCS and 1% penicillin/streptomycin and had a 51% plating efficiency. Both cell lines have a wt p53 phenotype. Transfectants (which are described below) were maintained in growth medium containing 400 μg/ml G418.

Parental LS513 and MCF-7 cells were cotransfected by calcium phosphate-DNA precipitation with expression plasmid p1322 encoding full-length HPV-16 E6 under the control of a human B-actin promoter (a gift from Dr. T. Fung, Children’s Hospital Los Angeles, Los Angeles, CA; Ref. 17) and pMC1neo polyadenylic acid encoding neomycin resistance (Stratagene, La Jolla, CA). Cells were selected in growth medium containing 800 μg/ml G418, and clones were isolated and screened for E6 function by Western blot analysis of p53 expression. The viral oncoprotein HPV-16 E6 marks p53 for ubiquitin degradation, and clones expressing high levels of E6 had abrogated basal and inducible p53 protein levels. The stable transfectant clones were designated LS513/E6 and MCF-7/E6. Parental cells were also transfected with pMC1neo polyadenylic acid alone to provide vector control cells, and these clones were designated LS513/neo and MCF-7/neo.

p53 function was measured using a transactivation assay in which cells were transiently transfected by calcium phosphate-DNA precipitation with a p53-HBS reporter plasmid, as described previously (15). This plasmid contains two copies of a 20-bp motif encoding a p53-HBS ligated immediately upstream from a minimal thymidine kinase promoter linked to CAT. Reporter gene activation occurs only when p53 binds to the HBS motif of the promoter. Cells were collected 48 h after transfection and assayed for CAT activity by standard thin-layer chromatography analysis. Assays were repeated in four separate experiments.

PH was a gift from QLT Phototherapeutics, Inc. (Vancouver, British Columbia, Canada), and this photosensitizer was used in all PDT experiments. A stock solution of PH was diluted to 25 μg/ml in RPMI 1640 containing 5% FCS. Cells were incubated in PH-containing medium for 16 h, rinsed for 30 min in complete growth medium, and exposed to red light (15). Control treatments included incubation with PH alone and exposure to light alone. X-irradiation was carried out using a ¹³⁷Cs gamma irradiator (Atomic Energy of Canada, Limited, Ontario, Canada) with a dose rate of 7.7 Gy/min. Cells were treated with Adriamycin (Cetus, Emeryville, CA) at a final concentration of 1.0 μg/ml in growth medium for time periods ranging from 2–48 h. Cytotoxicity was determined using a standard clonogenic assay. Cells were plated in triplicate in 60-mm culture dishes 24 h before the various procedures. After treatments, the dishes were incubated for 14 days (MCF-7) or 21 days (LS513), and the resulting colonies (>50 cells) were stained and counted. Clonogenic survival data were calculated from three separate experiments.

Cells were collected at various times after treatments in a SDS lysing buffer [4% SDS, 0.125 m Tris base, 10% glycerol; 4% 2-mercaptoethanol, and 0.02% bromphenol blue (pH 6.8)]. Protein samples were size-separated on discontinuous polyacrylamide gels (7.5–12.5%) and transferred overnight to nitrocellulose membranes. Filters were blocked for 1 h in 5% nonfat milk before incubation with antibodies, which included mouse monoclonal antihuman p53 (Clone DO-1; Lab Vision Corp., Fremont, CA), rabbit polyclonal antihuman p21 (C-19; Santa Cruz Biotechnology, Inc., Santa Cruz, CA), mouse monoclonal antihuman retinoblastoma protein (Clone G3–245; PharMingen, San Diego, CA), and mouse monoclonal anti-actin (C4, ICN Biomedicals, Inc., Aurora, OH). Filters were then incubated with either an antimouse or antirabbit peroxidase conjugate (Sigma, St. Louis, MO), and the resulting complexes were visualized by enhanced chemiluminescence autoradiography (Amersham Life Science, Arlington Heights, IL).

Treated and control cells were collected at various time intervals (0–48 h) by trypsinization and stored frozen at −80°C in citrate buffer (1.18% sodium citrate, 8.55% sucrose, and 5% DMSO) until analyzed. Cells were prepared using a detergent/trypsin method of nuclei preparation and stained with a propidium iodide/RNase solution (18). Cell cycle distributions were obtained from an analysis of 10,000 events using a Becton Dickinson FACScalibur flow cytometer and ModFit 2.0 software.

Treated and control cells were collected by trypsinization for quantitative measurement of DNA fragmentation using an Apo-Direct kit (PharMingen). Cells were stored and labeled as specified by the manufacturer. Briefly, samples were fixed in 1% paraformaldehyde, washed and resuspended in 70% ethanol, and stored at −20°C until assayed. Cells were sequentially labeled with terminal deoxynucleotidyltransferase and FITC-dUTP and analyzed by flow cytometry. Cell samples were also collected for DNA electrophoresis as described previously (13). This procedure demonstrates nucleosomal ladders as an indicator of apoptosis. Cell samples were incubated overnight in lysis buffer [10 mm Tris, 100 mm EDTA, and 0.5% SDS (pH 8.0)], and DNA was phenol extracted, ethanol precipitated, and separated by electrophoresis on a 1% agarose gel.

The significance of p53 expression as a predictor of treatment outcome continues to be examined by many investigators (8, 9, 12). It is generally thought that p53 regulates treatment responsiveness via apoptosis and cell cycle arrest (1). PDT functions by generating reactive oxygen species, and this treatment can induce both a rapid form of apoptosis and cell cycle arrest (11). In addition, many of the types of clinical tumors treated with PDT have a high frequency of p53 mutation (12). It is therefore of both basic mechanistic and clinical interest to evaluate the significance of the p53 phenotype in tumor cells exposed to PDT-mediated oxidative stress. The results obtained in the current study demonstrate that p53 expression does not directly modulate human tumor cell photosensitivity in either apoptosis-responsive (LS513) or nonresponsive (MCF-7) cells.

We used a strategy of selective abrogation of p53 to examine the role of this tumor suppressor gene in modulating the photosensitivity of cancer cells. Human colorectal (LS513) and breast (MCF-7) carcinoma cell lines were chosen for study because they have wt p53 phenotypes and represent malignancies amenable to PDT (5, 19). Disruption of p53 was accomplished after transfection and stable integration of the HPV-16 E6 gene into parental LS513 and MCF-7 cells. The E6 gene product is specifically involved in p53 binding and degradation with concomitant disruption of p53 transactivation function (16). This experimental procedure has been used effectively by other groups investigating p53 mechanisms of action (5, 6, 7). We successfully isolated G418-resistant MCF-7 and LS513 clones with stable integration of E6 for our study. The effectiveness of E6 transfection in attenuating p53 function was first examined using a transactivation reporter gene assay. A CAT reporter expression plasmid driven by a minimal thymidine kinase promoter containing a p53-HBS element was transiently transfected into MCF-7 and LS513 parental cells and their respective E6 and neo clones. Expression of CAT occurs only when wt p53 binds to the HBS motif within the promoter (20). Fig. 1 provides relative CAT activity (and, by inference, functional p53 levels) for parental and E6-containing tumor cells. High CAT activity was found in both parental (wt p53) tumor cell lines and cells containing only the transfected neo resistance gene. Greatly reduced CAT activity (and, by inference, p53 transactivation) was observed in the E6-containing clones. These results confirm that E6 expression abrogated p53 function in both the MCF-7 and LS513 cells.

The effect of E6 on p53 protein levels was also documented in parental and E6 cells using Western blot analysis. Fig. 2 shows that the basal and inducible levels of p53 expression were greatly attenuated in control and Adriamycin-treated LS513/E6 and MCF-7/E6 cells compared to parental cell lines. These observations agree with our p53 transactivation results described above and confirm the ability of E6 to abrogate p53.

Cytotoxicity of PDT-treated cells was determined using a clonogenic assay that measures reproductive integrity and the ability to proliferate. Fig. 3,A shows that LS513 and LS513/E6 cells exhibited equivalent photosensitivity. This indicates that the selective abrogation of p53 function did not alter PDT sensitivity in these cells. Interestingly, we previously compared PDT sensitivity in colon carcinoma cells from two patients (LS513 with wt p53 and LS1034 with a mutated/nonfunctioning p53 phenotype) and found the LS513 cells to be markedly more sensitive to PDT (15). This differential photosensitivity was not due to modifications in photosensitizer uptake. Our current data would suggest that cellular properties besides p53 activity influenced photosensitivity. Further evidence that p53 status does not correlate with treatment responsiveness is seen in Fig. 3 B, which shows that MCF-7 and MCF-7/E6 breast carcinoma cells also exhibited equivalent PDT sensitivity. Nevertheless, an earlier study performed in our laboratory demonstrated increased PDT sensitivity when p53 null HL60 promyelocytic leukemia cells were transfected with wt p53 (13). This finding concurs with the observation that cell types programmed for apoptosis often show altered sensitivity to therapeutic treatments, whereas cells derived from solid tumors do not show a clear relationship between p53 status and sensitivity (5). Additional studies using immunodeficient mice are required to confirm the in vivo PDT responsiveness of tumors derived from parental and E6 cell lines.

Fig. 2 shows that PDT-mediated oxidative stress induced high p53 expression in parental MCF-7 cells 24–48 h after treatment. In LS513 cells, PDT induced only slight p53 expression, which was seen at 48 h. Treatment with extended Adriamycin incubation was associated with a prolonged expression of p53 in both parental cell lines. Western analysis also shows the induction of p21 and pRb dephosphorylation in parental LS513 and MCF-7 cells exposed to PDT or Adriamycin. Incubation of LS513 and MCF-7 cells with PH alone resulted in protein expression levels for p53, p21, and pRb that were identical to those of nontreated controls. Recently, phthalocyanine-mediated PDT was found to induce p21 in human A431 epidermoid carcinoma cells that harbor a nonfunctional p53 mutation (21). PDT-mediated p21 induction in A431 cells was followed by the inhibition of cyclin D1/E, cyclin-dependent kinase 2 and cyclin-dependent kinase 6; G₀-G₁ cell cycle arrest; and apoptosis. Oxidative stress associated with diethyl maleate-mediated glutathione depletion also activates p21 expression and arrests cell cycle progression via a pathway that is independent of p53 (22). In our study, pRb dephosphorylation occurred in both parental and E6 cells treated with PDT but only in parental cells exposed to Adriamycin. This indicates that Adriamycin induced p53 and p21 expression and pRb hypophosphorylation in a sequential pathway, whereas PDT did not require p53 or p21 induction as an initiating event for pRb hypophosphorylation. The relevance of the reversal in pRb hypophosphorylation seen at 48 h in PDT-treated LS513 cells is currently unknown. In preliminary experiments, Western blot analysis showed a reversal of PDT-induced hypophosphorylation when cells were incubated with the phosphatase 1 and 2A inhibitor okadaic acid immediately after light exposure (data not shown). This suggests that PDT-associated pRb dephosphorylation may involve the activation of serine/threonine phosphatases, and studies are in progress to directly address this possibility. Interestingly, an earlier study indicates that inhibition of serine/threonine phosphatases 1 and 2A by calyculin inhibits PDT-induced apoptosis (23).

Inducible expression of p53 and p21 is often directly associated with G₁ cell cycle arrest and apoptosis (24). Downstream dephosphorylation of pRb regulates the G₁ arrest at the G₁-S-phase boundary (25). Recent work has demonstrated that PDT causes a significant G₀-G₁-phase arrest in A431 cells after p21 induction (21). Table 1 describes the cell cycle distribution in exponentially growing parental cells exposed to PH-mediated PDT at a dose that induces p21 expression as well as in E6 clones in which p21 was attenuated. All cell lines were analyzed over a 48-h time period, and only minimal cell cycle alterations were observed. A modest G₁ arrest (defined as an increase in G₁-phase cells and/or a decrease in S-phase cells) was detected in both LS513 and LS513/E6 cells. A prolonged G₁ arrest was observed in parental MCF-7 cells, whereas MCF-7/E6 cells showed a greatly reduced G₁ arrest. Our data suggest that the difference in G₁ arrest observed between parental and E6 cells could involve modifications in p53 function.

The role of HPV-16 E6-mediated abrogation of p53 function in PDT-induced apoptosis was also examined. Ionizing radiation is an efficient inducer of apoptosis and was used as a positive control in DNA fragmentation experiments. Radiation-induced apoptosis is characteristically measured 24 h after treatment, whereas PDT leads to a more rapid form of apoptosis detectable within a few hours of treatment (1, 23, 26). DNA nucleosomal laddering was first observed 2 h after PDT in LS513 cells and continued to increase at 6 and 24 h after light exposure (data not shown). Table 2 shows that parental LS513 cells underwent apoptosis, whereas LS513/E6 cells, with their greatly diminished constitutive and inducible p53 levels, exhibited reduced apoptosis after PDT and radiation exposures. These results indicate that p53 may play a role in the apoptotic response induced by both PDT and radiation in LS513 cells. However, PDT also induces rapid apoptosis in p53 null HL60 cells (13). MCF-7 cells exhibited only minimal PDT- or radiation-inducible apoptosis, and levels were further decreased in MCF-7/E6 cells. This agrees with previous reports indicating that MCF-7 cells do not readily undergo apoptosis because of the absence of a functional caspase-3 (5, 27). Interestingly, the differences in PDT-induced apoptosis observed in cells with wt and abrogated p53 did not translate into altered photosensitivity when measured by clonogenic survival. PDT-mediated oxidative stress produces both cell necrosis and apoptosis (1, 23, 26), and our results indicate that blocking one of the pathways may not alter the overall effectiveness of the treatment.

In conclusion, selective abrogation of p53 expression with HPV-16 E6 did not modulate the clonogenic survival of colon or breast carcinoma cells exposed to PDT-mediated oxidative stress. Likewise, p53-associated changes in apoptosis did not translate into alterations in PDT-induced cell killing. These results suggest that p53 status may not be useful as a predictive factor of PDT sensitivity in colon or breast carcinomas. Procedures such as PDT, which are equally cytotoxic irrespective of a tumor’s p53 status, are essential in the armamentarium of cancer therapies because alterations in p53 are found in approximately half of all human malignancies (12). The lack of correlation between p53 expression and PDT efficacy could explain the responsiveness of neoplastic cells to oxidative stress regardless of patterns of resistance to drugs and ionizing radiation.

We thank Lora Barsky for assistance with flow cytometric analysis.