Papillary serous endometrial carcinoma is an aggressive tumor characterized by late-stage presentation, i.p. spread, and poor prognosis. It is histologically similar to serous papillary carcinoma of the ovary. Preclinical studies have shown that adenovirus-mediated expression of p53 in ovarian cancer cell lines causes growth inhibition and apoptosis in vitro and in vivo. Such studies provide the rationale for Phase I Adp53 gene therapy clinical trials in ovarian cancer. In the present study, we compared the efficacy of adenoviral vectors containing p53 (Adp53) or p21 (Adp21) in a papillary serous endometrial tumor cell line (SPEC-2) that contains mutated p53. Growth assays revealed that both Adp53 and Adp21 were efficacious in decreasing cell proliferation as assessed by anchorage-dependent and anchorage-independent growth assays. However,as compared with Adp53, the effects of Adp21 tended to be more transient and less marked. Strikingly, Adp21, but not Adp53, induced a G1 arrest in SPEC-2 endometrial adenocarcinoma cells. In contrast, as assessed by induction of hypodiploid peaks, free DNA ends detected by a terminal deoxynucleotidyl transferase-based assay, and annexin V positivity, p53 was more effective than p21 in inducing cell death by apoptosis. Compatible with the more efficient induction of apoptosis, Adp53, but not Adp21, induced a marked increase in expression of the preapoptotic molecule BAX without a concomitant change in expression of the antiapoptotic mediator Bcl-2. The differential effects of Adp53 and Adp21 on cell cycle progression and apoptosis may be related to the reversibility of p21-induced cell cycle arrest and the irreversibility of p53-induced apoptosis. Thus, at least in the papillary serous endometrial carcinoma cell line SPEC-2, Adp53 may be more effective than Adp21 as a gene therapeutic. Nevertheless,these preclinical studies suggest that papillary serous endometrial carcinoma is a potential target for p53- or p21-mediated gene therapy.
UPSC2 is a rare aggressive endometrial tumor characterized by late stage at presentation, i.p. spread, and poor prognosis. UPSC is similar to serous papillary ovarian cancer in presentation and outcome. The reported 5-year survival rates for UPSC are only 45–62% for stage I–II and 11% for stage III–IV (1, 2, 3). Effective treatment for advanced or recurrent UPSC does not exist because available surgical, chemotherapeutic, or radiotherapeutic techniques do not result in significant cure rates. Initial treatment by hysterectomy and extended surgical staging is usually recommended. Postoperative treatment may consist of adjuvant pelvic and whole abdominal radiation therapy or chemotherapy (3, 4, 5). Regardless of the type of treatment, more than 50% of cases recur (1, 2, 3). Stage-specific survival rates may be worse for UPSC than for endometrioid endometrial cancer due to inaccurate staging or to an altered natural history of the disease. For example, as many as 44% of cases of UPSC are upstaged from clinical staging after surgical evaluation (3). The poor prognosis in UPSC may also be related to the 50% lymph-vascular space invasion found at the time of UPSC diagnosis compared to the 14% lymph-vascular space invasion in typical endometrioid adenocarcinoma, reflecting a different pathophysiology (6).
Preclinical studies of i.p. delivery of adenovirus carrying p53 (Adp53)have led to Phase I studies of Adp53 in women with ovarian cancer. In addition to the potential efficacy of Adp53 gene therapy in a number of tumor models, gene therapy offers the theoretical advantage of locoregional therapy of ovarian cancer and potentially of UPSC when delivered into the peritoneal cavity (7, 8, 9). The disease distribution and histological similarities of UPSC and ovarian cancer suggest that UPSC patients may also be good candidates for novel i.p. gene therapy delivery systems if used in combination with hysterectomy.
Immunohistochemical studies suggest that p53 is aberrant in 50–90% of UPSC tumors in comparison to 10–30% of typical endometrioid adenocarcinomas (10, 11, 12, 13). This high level of abnormality is similar to or exceeds the 50% positivity found in advanced epithelial ovarian cancers (11). Presumably, these findings correlate with the presence of mutated p53 or simply abnormal overexpression of p53 protein (11). Thus, replacement of a normal functioning p53 could contribute to the eradication of these tumor cells. Studies evaluating the effect of wtp53 transfection into colorectal, glioblastoma, and hepatocellular carcinomas have supported the hypothesis that introduction of normal p53 into cells with aberrant p53 leads to decreased growth and increased cell death (14).
Although a downstream effector of p53 mediated G1 arrest,the role of p21 in the apoptotic pathway remains controversial. Indeed,studies of p21 in knockout mice failed to demonstrate an obligatory role for p21 in p53-mediated apoptosis (15). In certain cell lines, infection with Adp21 does not lead to apoptosis but rather results in growth arrest and protection from p53-dependent apoptosis (16, 17). Infection with Adp21 has been demonstrated to cause permanent in vitro growth arrest but not apoptosis and to decrease tumor growth, thus increasing the survival of treated mice carrying prostate tumors (18, 19).
We demonstrate that replacement of both p53 and p21 in the UPSC cell line SPEC-2 induces significant growth suppression, alterations in population distribution throughout the cell cycle, and apoptotic cell death. However, infection with Adp53 is much more effective than Adp21 in inducing apoptosis, whereas Adp21 is more effective than Adp53 in inducing cell cycle arrest.
MATERIALS AND METHODS
The SPEC-2 UPSC cell line was developed by Drs. J. Boyd and D. G. Kaufman at the University of North Carolina (Chapel Hill, NC) and was kindly provided by Dr. Janet Price (M. D. Anderson Cancer Center,Houston, TX; Ref. 20). Cells were grown in complete MEM with 10% heat-inactivated fetal bovine serum and P/S.
SPEC-2 cells were transduced with a recombinant adenovirus containing human cDNA for either the p53 or p21 tumor suppressor gene driven by the cytomegalovirus promoter with a polyadenylation signal and a minigene cassette inserted into the E1 deleted region of modified adenovirus (Adp53 and Adp21). Viral stocks were propagated in 293 cells as described previously (21). Monolayer SPEC-2 cells in 10% MEM were infected with the chosen MOI and incubated with agitation at 37°C for 60 min. At the completion of the incubation period,additional medium was added without washing the virus from the cells,except where indicated.
Adenoviral expression of Escherichia coli β-gal gene(Ad5CMVβ-gal) and GFP (AdGFP) was used to determine transduction efficiency. Cells were plated at 2.5 × 105cells/35-mm2 dish and analyzed as described previously (22). Transduction efficiency was further evaluated using an adenovirus harboring a natural GFP (AdGFP). Cells were infected and incubated as with the Ad5CMVβ-gal at varying MOIs; however, at the completion of the incubation period, they were trypsinized, washed in PBS, and evaluated for expression of AdGFP by flow cytometry without further preparation.
Growth Suppression Assays.
Cell counting assays were used to evaluate the effect of infection on SPEC-2 cell growth. Cells were plated in triplicate at a density of 2 × 104 cells/well in 6-well plates. Cells were either uninfected, infected with Ad5CMVβ-gal as a vector control, or infected with Ad5CMVp53 or Ad5CMVp21. Cells were harvested and counted with a Coulter counter (Model ZM, Coulter Corp., Miami, FL) on days 1,3, 5, and 7 after infection. Growth effect was also monitored using dye conversion in a microculture MTT assay as described previously (23).
Protein expression after infection was assessed via Western blotting. Total cell lysates were prepared by sonicating the cells 24 and 72 h after infection in radioimmunoprecipitation assay buffer [150 mm NaCl, NP40, 0.5% sodium deoxycholate, 0.1% SDS, and 50 mm Tris (pH 8.0)] for 5 s. Fifty μg of protein were loaded onto a 13% SDS-polyacrylamide gel for evaluation of p53, or 25μg of protein were loaded onto a 10% gel for evaluation of p21, BAX,or Bcl-2 expression. Protein was electrophoresed and transferred to a Hybond-enhanced chemiluminescence membrane (Amersham, Arlington Heights, IL). Membranes were blocked with 5% nonfat dry milk and 0.1%Tween-20 and then probed with mouse antihuman p53 (1:1000) antibody,primary rabbit antihuman p21 antibody (1:1500), rabbit anti-Bax (Santa Cruz Biotechnology, Santa Cruz, CA), or mouse anti-Bcl-2 antibody(DAKO, Carpinteria, CA) antibody. Appropriate IgG horseradish peroxidase-conjugated goat antimouse or rabbit secondary antibodies were used, and blots were processed as per the manufacturer’s suggestions.
Cell Cycle Analysis.
Cells were incubated with virus for 72 h at a density of 1 ×106 in a 10-ml plate. Cells were fixed in ethanol and stored at 4°C. At the time of cell cycle analysis, cells were washed and resuspended in PI staining buffer (50 mg/ml PI and 15 mg/ml Rnase)in PBS. DNA content was evaluated by flow cytometry (Coulter Epics XL-MCL).
Evidence for apoptosis was evaluated by the presence of a hypodiploid peak during cell cycle analysis and via a TUNEL-based assay (Apo-BRDU kit; Phoenix Flow Systems, San Diego, CA). Cells were plated at 1 × 106; mock-infected or infected with p53, p21, or β-gal adenoviral vectors at a MOI of 50; and then harvested 3 and 5 days after infection. This time period correlated with the appearance of increased numbers of detached cells. Cells were initially fixed in 1%formaldehyde and then in 70% EtOH. Fixed cells were incubated with Tdt enzyme and bromodeoxyuridine triphosphate, rinsed, and then incubated with fluorescent mAb and PI/RNase staining buffer. The percentage of positive cells were determined by flow cytometry (Coulter Epics XL-MCL).
To independently assess necrosis and apoptosis, annexin binding in early apoptotic cells was evaluated (24). Cells were plated at 106 in 10-ml plates and infected, as described in the TUNEL-based assay, at a MOI of 50 the following day. Three and 5 days after infection, cells were harvested and incubated with annexin V-FITC according to the manufacturer’s recommendations (Annexin V-FITC, Trevigen). Cells were stained with PI as described above. Annexin V-positive and PI-positive cells were determined to be apoptotic.
Anchorage-independent Colony Formation.
Cells were plated and infected 24 h later at a MOI of 50 in 6-well plates. Twenty-four h after infection, cells (1 ×103) were washed and plated in 0.6% agar in triplicate in three separate experiments. Plates were incubated at 37°C, and colonies consisting of at least 30 cells were counted on an inverted microscope at 100× magnification 7–13 days after agar plating (25).
SPEC-2 Cell Line Harbors a Mutated p53.
p53 sequencing in our laboratory revealed a p53 frameshift mutation in exon 6 as a result of a 3-bp deletion at amino acid 218 (data not shown). Evaluation by the cytogenetics department at the M. D. Anderson Cancer Center revealed a complex karyotype characterized by multiple chromosomal deletions, duplications, and translocations (data not shown).
Transduction Efficiency of Adp53.
Transduction efficiency of adenoviral constructs has typically been defined by the percentage of cells staining positive for a transducedβ-gal gene. Because this is not only tedious but subjective, relying on the ability to visibly discern “positive blue” from a nonstained or a lightly stained cell, we evaluated transduction efficiency using an adenoviral vector including a GFP reporter. The β-gal assay suggested that 50% transduction efficiency occurred at a MOI of 200(Fig. 1). In comparison, studies with AdGFP indicated that a 50% transduction efficiency occurred at only a MOI of 5. Therefore, evaluation of transduction efficiency via β-gal expression may significantly underestimate the efficiency of the adenovirus to infect SPEC-2 cells. Objective determination using the AdGFP vector and flow cytometry may provide a superior quantification technique.
p53 Induces p21 Expression.
To determine the levels of expression of p53 and p21 after infection,cell lysates at 24 and 72 h after infection were collected. As expected, Western blotting confirmed increased p53 and p21 expression after transduction with Adp53. p21 protein levels after infection with Adp21 were significantly higher than the level achieved after activation of p21 transcription with Adp53 infection (Fig. 2).
Adp21 and Adp53 Cause Significant Growth Suppression.
To identify the MOI with the least amount of vector toxicity and the greatest amount of p53-induced growth inhibition, we evaluated growth suppression by dye conversion using a MTT dye. Cells were mock-infected or infected with multiple dilutions of Adβ-gal or Adp53. Using this method, a MOI of 50 appeared to produce the greatest p53-specific effect with the least viral-dependent effect. This suggests that as compared with AdGFP staining, p53 growth inhibition may require a higher MOI.
Adp53 was much more effective than Adp21, inducing a 92% decrease in cell counts as compared to an 80% decrease induced by Adp21 on day 7. Strikingly, in contrast to the Adp53-infected cells, the Adp21-infected cells began to proliferate rapidly 7 days after infection (Fig. 3). Furthermore, 72 h after Adp53 infection, most of the cells were shrunken and detached, a morphological change that was not observed with Adp21-infected cells at the same time point (data not shown).
Anchorage-independent Growth Suppressed.
We evaluated SPEC-2 colony formation ability in agar after infection with Adp53 or Adp21 and compared the results to mock-infected andβ-gal-infected cells. p53 infection resulted in a 52% reduction in colony formation, whereas p21 infection resulted in a 20% reduction(the average of three separate experiments). β-Gal infection did not decrease colony formation ability (Fig. 4). Thus, this reduction in the ability to produce colonies in an anchorage-independent environment supports the potential for this type of anticancer therapy in vivo.
Cell Cycle Analysis.
Either cell cycle arrest or cell death through apoptosis could account for the inhibition of anchorage-dependent and anchorage-independent growth after infection with Adp53 and Adp21. To assess the effect on cell cycle progression, population distribution throughout the cell cycle was examined via flow cytometry. Of the nonhypodiploid viable cells, Adp53 induced a minimal effect on G1 but was associated with a marked decrease in the S-phase population (38%),with a concommitant decrease in G2-M phase compared to controls (Table 1). In contrast,infection with Adp21, at the same MOI, resulted in a marked increase in the G1 population (28%) in addition to decreases in S-phase and G2-M-phase populations of the cell cycle. Adp53 and Adp21 induced a marked increase in hypodiploid cell populations(compatible with apoptotic cell death) compared to controls (20% and 7.4%, respectively).
Adp53 Is More Effective in Inducing Programmed Cell Death than Adp21: TUNEL-based Assay.
As indicated in Table 1, growth suppression after infection with Adp53 maybe primarily apoptotic, whereas that after Adp21 may be due primarily to cell cycle arrest with a lesser component of apoptosis. A TdT TUNEL-based assay was used to determine the degree of apoptotic cell death at 3 and 5 days after infection with either vector. Three days after infection with a MOI of 50, Adp53 infection induced 87%apoptotic cells as indicated by TdT positivity in four similar experiments. With the same MOI, 70.5% of cells were apoptotic after Adp21 infection. Five days after infection, the numbers of apoptotic cells increased with both infections (Table 1). This is compatible with the presence of hypodiploid peaks. Thus, although both Adp53 and Adp21 induce cell death, Adp53 may be more efficient.
TUNEL-based assays may not differentiate between necrotic and programmed cell death (26). The annexin V assay evaluates membrane changes specific for apoptosis by binding to phosphotidylserine in the membranes of cells undergoing apoptosis (24, 27). Average results from three similar experiments 72 h after infection showed cells infected with p53 to be 54%positive for annexin binding compared with a 35% positivity after p21 infection. Annexin positivity increased 5 days after infection with both agents (Table 1). These results again indicate that Adp53 is more efficient than Adp21 at producing apoptotic cell death.
BAX Expression Is Increased after Infection with Adp53.
p53 has been demonstrated to induce expression of BAX, potentially accounting for its effect on cellular apoptosis (28). We therefore assessed the effect of Adp53 and Adp21 on the expression of the BAX and Bcl-2 mediators of apoptosis. Strikingly, Adp53, which induced high levels of apoptosis in SPEC-2 cells, induced a marked increase (22-fold) in BAX expression. In contrast, Adp21, which was much less efficient in inducing cell death, did not alter BAX expression (Fig. 5). For unknown reasons,the adenovirus alone induced a modest increase in BAX expression (Fig. 5). None of the constructs altered the level of Bcl-2 expression. The altered ratio of BAX:Bcl-2 may thus contribute to Adp53-induced programmed cell death.
The histology and gross presentation of UPSC is often similar to serous papillary carcinoma of the ovary. Indeed, epidemiological studies and preliminary molecular genetic analysis suggest that serous papillary carcinoma of the ovary can arise in endometrial implants,suggesting a common cell of origin (29). Similar to epithelial ovarian carcinomas, i.p. involvement at the time of diagnosis and the presence of psammoma bodies are common in UPSC (2). As with ovarian cancer, CA-125 has been shown to be predictive of disease status for some women with UPSC (30).
In this series of experiments, we evaluated the effects of Adp53 and Adp21 gene therapy on the UPSC cell line SPEC-2. This is the first study evaluating the effects of adenoviral gene therapy on an endometrial cancer cell line. The importance of this study lies in the clinical and molecular similarities between papillary serous endometrial cancer and epithelial ovarian cancer and the potential application of this type of therapy for UPSC patients. Our results have shown that the SPEC-2 cell line is easily transduced by adenoviral vectors containing p21 or p53 tumor suppressors. Interestingly, we were able to more objectively quantitate the degree of transfection using flow cytometric evaluation of a transduced GFP than the more conventional β-gal vector. However, the actual transduction efficiency may not directly correlate with the needed intracellular functional threshold to successfully convert to a p53 repair/apoptosis pathway.
Transfection of p53 or p21 by an adenoviral vector resulted in growth suppression and cell death by apoptosis in the SPEC-2 cell line. In general, p53 appears to initiate an accelerated or more efficient programmed cell death pathway involving increased expression of the BAX protein. The kinetics of p53-induced cell death did not appear to involve a G1 arrest, in spite of the fact that infection with Adp53 resulted in increased p21 expression. In contrast, infection with Adp21 activated a kinetically slower and milder apoptotic pathway. Adp21 infection caused a significant G1 arrest but did not appear to induce changes in the expression of the Bcl-2 family of proteins.
The minimal G1 arrest after p53 infection was an unexpected finding because p53 induces p21 expression. The reason for this unexpected finding may also be due to the kinetics of p53-dependent cell death. p53 infection may activate programmed cell death so efficiently that G1 arrest is not observed. An alternative possibility is that SPEC-2 may harbor a mutated p21 that is unable to interact with the cyclin-dependent kinase-cyclin complex and therefore unable to initiate G1 arrest (31). Alternatively, cell cycle arrest may be dependent on achieving a threshold level of functional p21 that infection with Adp53 did not achieve. Studies that demonstrate that cell lines with nonfunctional p21 undergo apoptosis after expression of wtp53, in contrast to cell lines with functional p21, which arrest in G1 after p53 activation, have been interpreted to indicate that p21 protects against p53-mediated apoptosis (16). This may actually be due to the ratio of expression of p21:p53. In the referenced studies, at low doses, infection with Adp21 reduced cell death in cells expressing a p53 temperature-sensitive mutant. However, at a higher MOI, p21 alone induced cell death (16). This appears to support the hypothesis that a p21:p53 protein expression ratio may be integral in determining phenotypic outcome.
p21 has been shown to be involved in reversible and irreversible growth arrest, apoptosis, terminal differentiation, and possibly cell death due to necrosis (18, 19, 32, 33). Our results suggest that p21 produces apoptosis in SPEC-2 cells but is less efficient than p53. Furthermore, apoptosis after infection with p21 appears to be a p53-independent mechanism indicated by a lack of alteration in p53 protein expression on Western blot after p21 infection and the presence of mutant p53 in SPEC-2 (data not shown). Concordant with the differential ability of Adp53 and Adp21 to induce apoptosis, p53 markedly induced expression of BAX, whereas Adp21 did not. Nevertheless, Adp21 did induce BAX-independent apoptotic cell death.
p53-dependent programmed cell death has been proposed to be related to changes in the expression of Bcl-2 family members (28),FAS (34), or redox-related genes (35) rather than through its downstream effector, p21 (15). Replacement of wtp53 in SPEC-2 causes significant and rapid cell death associated with elevations in BAX along with a stable level of Bcl-2 expression. To confirm that the observed cell death caused by Adp53 and Adp21 infection was programmed rather than necrotic, three methods were used to detect apoptosis: (a) the presence of hypodiploid cells on flow cytometry; (b) TUNEL-based assay; and(c) annexin V staining. All three methods showed evidence of apoptotic cell death in the SPEC-2 cell line after infection with Adp53 or Adp21. However, Adp53 was more efficient than Adp21 in producing an apoptotic cell death. Taken together, our results indicate that Adp53 and Adp21 result in a marked decrease in cell proliferation, albeit through different primary mechanisms. This supports additional studies of the potential of in vivo use of Adp53 and Adp21 gene therapy in UPSC patients.
The abbreviations used are: UPSC, uterine papillary serous carcinoma; MOI, multiplicity of infection; PI,propidium iodide; β-gal, β-galactosidase; wtp53, wild-type p53;GFP, green fluorescent protein; mAb, monoclonal antibody; P/S,penicillin/streptomycin; Tdt, terminal deoxynucleotidyl transferase;TUNEL, Tdt-mediated nick end labeling; MTT,3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide.
|.||Control .||β-Gal .||p53 .||p21 .|
|Cell cycle (72 h)|
|% G2-M phase||8.8||9.9||16.9||6.5|
|% S phase||32.9||15.4||23.5||8.1|
|TdT % (72 h)||1.9||5.0||87.0||70.5|
|Annexin % (120 h)||7.5||8.4||54.0||35.9|
|.||Control .||β-Gal .||p53 .||p21 .|
|Cell cycle (72 h)|
|% G2-M phase||8.8||9.9||16.9||6.5|
|% S phase||32.9||15.4||23.5||8.1|
|TdT % (72 h)||1.9||5.0||87.0||70.5|
|Annexin % (120 h)||7.5||8.4||54.0||35.9|
Cells were plated in 10-ml plates at 1 × 106cells/plate in MEM supplemented with 10% serum, P/S, nonessential amino acids, sodium pyruvate, and l-glutamine (complete MEM). Twenty-four h after seeding plates, infection (see “Materials and Methods”) with a MOI of 50 was performed. Cells were harvested 72 h after infection via trypsinization and fixed in 70% EtoH. DNA content was determined by flow cytometry after staining with PI(see “Materials and Methods”). Hd refers to hypodiploid population. Infection with Adp53 resulted in a significant reduction in S phase and an increased percentage of cells in G0-G1. Infection with Adp21 shifted cells to a G1 arrest and also decreased the S-phase population. Results represents one of three similar experiments. Three days after infection, cells were harvested via trypsinization and fixed in 1% paraformaldehyde, rinsed, and then fixed in 70% EtoH. Fixed cells were incubated with TdT enzyme and Br-dUTP, followed by incubation with fluorescent mAb (see“Materials and Methods”). Percentage of apoptotic cells was recorded by flow cytometry. Infection with Adp53 or Adp21 resulted in increasing percentages of apoptosis cells. Results represents one of three similar experiments. Five days after infection, cells were harvested via trypsinization, counted, and washed in PBS. Five ×105 cells were incubated with annexin V-FITC and PI (see“Materials and Methods”). Percent of apoptotic cells was recorded by flow cytometry. Infection with Adp53 or Adp21 resulted in increasing percentages of apoptotic cells by 5 days. Results from one experiment 5 days after infection are presented; note the percentage of apoptotic cells is less than that seen with TUNEL-based assay. Results represent one of three similar experiments.
We thank Ying Henderson for sequencing p53 in the SPEC-2 cell line, Karen Rameriz for technical assistance with flow cytometry,Dianne Fightmaster for technical assistance with the MTT assay, and the M. D. Anderson Cancer Center Cytogenetics Department for evaluating SPEC-2 chromosomes.