Neocarzinostatin Induces an Effective p53-Dependent Response in Human Papillomavirus-Positive Cervical Cancer Cells

Carcinoma of the uterine cervix is one of the most common neoplasia in women, and more than 90% of all cervical carcinomas are epidemiologically associated with infection by high-risk human papillomaviruses (HPVs) such as types 16, 18, 31, and 33. These viruses encode two oncoproteins, E6 and E7, capable of altering proliferation and apoptosis, as well as inducing immortalization of primary human keratinocytes (zur Hausen, 2000). E6 specifically binds to wild-type p53 protein inducing its cytoplasmic sequestration and subsequently enhancing its degradation in an ATP/ubiquitin-dependent manner, resulting in deregulated cell growth and enhanced genomic instability (Scheffner et al., 1990; Liang et al., 1993).

p53 is a transcription factor with sequence-specific DNA-binding activity, and there is evidence that this tumor suppressor plays an important role in the maintenance of genomic stability after DNA damage (Kastan et al., 1991). Intracellular p53 protein levels increase following DNA damage, inducing cell cycle arrest in G₁, which allows the cell to undergo efficient DNA repair. In this way, wild-type p53 counteracts the emergence of cell clones with acquired DNA mutations. In contrast, mutant p53 is unable to induce G₁ arrest in response to DNA damage, resulting in inefficient DNA repair and the emergence of genetically unstable cells. If damage is not repaired, the cells are eliminated by p53-induced apoptosis. Therefore, the ability of p53 to induce apoptosis is of central importance to its tumor suppressor activity (Symonds et al., 1994).

Since p53 protein is potentially functional in cervical cancer cell lines, blocking E6-mediated p53 inactivation is a main therapeutic goal (Butz et al., 1995). Previous studies have shown that inhibition of E6-mediated degradation of p53 in cervical cancer frequently results in increased levels of p53 expression (Butz et al., 1995). In many cases, however, it is necessary to induce DNA damage so that p53 becomes stable (Mantovani and Banks, 1999). Thus, therapeutic strategies that selectively rescue the p53 pathway may help sensitize HPV-positive transformed cells to undergo growth arrest or apoptosis.

The enediynes are a class of DNA strand-breaking agents that show promise as anticancer drugs because they are extremely potent cytotoxic and antimitotic agents, and some of them exhibit selectivity for neoplastic cells relative to normal human bone marrow (Nicolaou et al., 1992). Neocarzinostatin (NCS), a product naturally found as a complex of an enediyne chromophore and a 10-kDa protein, was the first member of this family to be described. It has been previously shown to induce both single- and double-strand breaks in DNA. The oligonucleosomal cleavage of DNA seen after NCS treatment is associated with apoptosis rather than being the direct result of the strand-cleaving effects of the drug itself (Hartsell et al., 1995).

NCS is considered a prodrug that requires reduction for activation, and its cytotoxicity has been demonstrated to vary directly with the sulfhydryl content of the cell. Interestingly, higher Bcl-2 protein expression, frequently observed after p53 inactivation, enhances NCS effects (Cortazzo and Schor, 1996; Schor et al., 1999). Enediynes that require reductive activation might be clinically useful agents in the therapy of tumors related to up-regulation of the proto-oncogene bcl-2 such as neuroblastomas, estrogen-responsive breast cancers (Cortazzo and Schor, 1996), and cervical cancer (Liang et al., 1995).

In this study, we have determined the effect of NCS on stabilization and activation of p53, as well as on cellular proliferation and apoptosis induction in HPV-positive and -negative cervical cancer cell lines. NCS was capable of inhibiting cellular proliferation by inducing G₂ cell cycle arrest and apoptosis in both HPV-positive and -negative cell lines. NCS lethality was higher for tumor cells than for normal human keratinocytes under the same experimental conditions. In addition, we observed a significant increase in nuclear p53 levels and transcriptional function in HPV-positive and also in ViBo (HPV-negative) cells. Higher levels of p21 transcripts accompanied this effect in all cell types tested. In contrast, a decline in Bcl-2 protein levels and after NCS treatment was observed only in CaSki, C33A (24 h after exposition), and ViBo cells (48 h after exposition).

These results indicate that HPV-positive cervical cancer cell lines have an effective NCS-induced DNA damage response despite E6 expression, suggesting an anticancer potential for this drug against cervical cancer, independently of HPV infection.

Materials and Methods

Reagents. A 0.6 mg/ml stock solution of neocarzinostatin (10,700 molecular weight; Kayaku Pharmaceuticals Ltd., Tokyo, Japan) in 10 mM ammonium acetate buffer, pH 5.0, kindly provided by Dr. Irving Goldberg (Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA) and Dr. Nina Schor (Department of Neurology, University of Pittsburgh, Pittsburgh, PA), was kept in the dark at 4°C and diluted with Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Carlsbad, CA) immediately before each experiment.

Plasmids. The p53-responsive reporter plasmid p53CON, which contains a single 19-base pair p53-recognition sequence and is under minimal human hsp70 promoter control, as well as the basic vector pGUP.PA.8, were kindly donated by Dr. Hoppe-Seyler (Projektgruppe Angewandte Tumorvirologie, Heildelberg, Germany).

Cells and Cell Cultures. Cervical cancer cell lines HeLa (HPV18), CaSki (HPV16), and C33A (HPV-negative, p53 mutant) were purchased from American Type Culture Collection (ATCC, Rockville, MD). INBL (HPV18) and ViBo (HPV-negative, wild-type p53) cervical cancer cell lines were established previously from surgical tumors of Mexican patients diagnosed with epidermoid cervical carcinoma (Monroy et al., 1992) and analyzed in our laboratory for HPV. Cells were grown in DMEM supplemented with 10% fetal bovine serum and 2 mM glutamine (Invitrogen). Normal human keratinocytes were prepared from neonatal foreskins using keratinocyte-serum free medium (Keratinocytes-SFM; Invitrogen) according to manufacturer's instructions and grown in this medium supplemented with 20 mg/ml bovine pituitary extract and 0.1 ng/ml epidermal growth factor (Invitrogen). One day before the treatment, cells were plated on either 35, 60, 100-mm diameter, 96-well plates or on glass coverslips on 35-mm plates at a density of 7800 cells/cm² and maintained overnight at 37°C in a humidified atmosphere with 5% CO₂. Cells were then treated with NCS at the concentration indicated in each case. NCS-exposed and control cells (treated with the same amount of vehicle alone diluted with DMEM) were incubated at 37°C for 1 h. Medium was removed, and the cells were washed free of NCS. Fresh medium was added to each plate, and cells were incubated at 37°C for the rest of the experiment.

Viable Cell Count. Viable cell number was determined 0, 24, 48, and 72 h after NCS exposure with the standard trypan blue exclusion method by counting blue and clear cells. Daily cell counts were reported as the average of triplicate cell counts expressed as a percentage of the average number of cells on day 0. Counts obtained under each experimental condition were compared with those obtained for controls.

Cell Cycle Analysis. Cells untreated and treated for 1 h with 40 nM NCS were harvested at 0, 24, 48 and 72 h, adjusted to 1 × 10⁶ cells/tube, fixed in 70% ethanol (1 h on ice), and stored at –20°C until the assay. Immediately before analysis, cells were washed with phosphate-buffered saline (PBS) and incubated at 4°C in a staining solution containing 50 μg/ml propidium iodide (PI) and 100 K units RNase A in PBS. After 30 min of staining, DNA fluorescence was measured with a FACSort flow cytometer (BD Biosciences, San Jose, CA). For each sample, 10,000 events were acquired on a linear scale, and percentages of cells in G₀/G₁, S, and G₂/M phases of the cell cycle were determined using the FACSort ModFitLT V2.0 software.

Apoptosis Analysis. Cells growing in coverslips were treated for 1 h with 40 nM NCS. Seventy-two hours after NCS removal, cells were fixed in methanol/acetone 1:1 (v/v) for 2 min and then stained with 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI; 2 μg/ml in 100 mM Tris-HCl, pH 7.5) for 30 min in the dark. Cells were mounted in PBS/70% glycerol mounting medium and examined for morphological signs of apoptosis under a fluorescence microscope. Apoptosis was also determined by terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) technique by flow cytometry. The TUNEL assay was modified slightly from the manufacturer's recommended protocol (Roche, Mannheim, Germany). Cells were harvested and fixed in freshly prepared 1% para-formaldehyde in PBS for 30 min at 4°C and then in 70% ethanol for 1 h at 4°C. Fixed cells were permeabilized using 0.2% Triton X-100 in 0.1% sodium citrate. The DNA labeling mixture containing terminal deoxynucleotidyl transferase was then added. Cells were incubated overnight at room temperature and washed twice with PBS. Controls were resuspended in the TUNEL reaction mixture containing fluorescent dUTP without terminal deoxynucleotidyl transferase. To correlate apoptosis with cell cycle status, cells were stained with PI, as described previously and analyzed by flow cytometry. A dot plot of PI-red fluorescence (index of DNA content) versus green fluorescein isothiocyanate-fluorescence (dUTP incorporation) was constructed.

Detection of p53 and Bcl-2 by Flow Cytometry and Immunocytochemistry Analysis. For flow cytometric analysis, cells were treated for 1 h with 40 nM NCS, washed twice with PBS, and harvested using versene (0.02% EDTA) after 0, 24, and 48 h of NCS exposure. Cells were adjusted to 1 × 10⁶ cells/ml and fixed with 0.25% para-formaldehyde for 15 min in the dark followed by 70% methanol for 1 h at 4°C. Fixed cells were washed twice in PBS and resuspended in 50 μl of a solution containing 0.25% Tween, 1% bovine serum albumin, and 10 μl of fluorescein isothiocyanate-conjugated antibody against either p53, Bcl-2, or an isotype control (Dako, Santa Barbara, CA) in PBS. Following 45 min of incubation in the dark, cells were washed in 0.25% Tween/PBS. After centrifugation, cells were resuspended in 600 μl of PBS and analyzed on a FACSort flow cytometer. At least 10,000 events were acquired per sample. Data analysis and Kolmogorov-Smirnov statistics (Conover, 1980) were done using CellQuest V2 software. Each determination was made twice on independent cultures.

Simultaneously, cells were grown on coverslips and treated with NCS, as described above. Cells were washed with PBS and fixed for 2 min with methanol/acetone 1:1 (v/v) at –20°C. After PBS washings, cells were incubated for 5 min with 3% H₂O₂, permeabilized with PBS containing 0.25% Triton X-100, and incubated for 30 min with 1% bovine serum albumin. Cells were then incubated (1 h at 37°C) with antibody against p53 (Dako) diluted 1:50 in PBS. Cells were washed with PBS-0.5% Tween-20 and incubated (45 min) with biotinylated anti-mouse or anti-rabbit IgG antibodies (Dako). After washing twice with PBS, samples were incubated (20 min) with peroxidase-conjugated streptavidin and washed several times with PBS. Cells were incubated with 3-amino-9-ethyl carbazole (Zymed, Carlton Court, South San Francisco, CA) or diamino benzidine (Dako) chromogen solution. Finally, samples were mounted. To determine changes in cellular localization of p53, samples were taken 0, 3, 6, 12, 24, and 48 h after 40 nM NCS treatment and processed by immunocytochemical analysis as above.

E6/E7 and p21 Expression. To investigate the possible effect of NCS on E6/E7 expression and on the main p53-checkpoint gene p21, cells were analyzed by RT-PCR at 0, 3, 6, 12, 24, and 48 h after 40 nM NCS treatment. Total cellular RNA was prepared using TRIzol (Invitrogen), and 1.5 μg of total RNA was retrotranscribed with Superscript II RNase H^– reverse transcriptase (Invitrogen) (final volume, 20 μl). The cDNA (1 μl) was then amplified in a DNA GeneAmp 2400 PerkinElmer Thermal Cycler (PerkinElmer Instruments, Norwalk, CT) with 2.5 U TaqDNA polymerase (Invitrogen), 100 nM upstream, and downstream primers for E6/E7 (Fujinaga et al., 1991), p21 (Saegusa et al., 2001), or human β2-microglobulin (5′-ACC CCC ACT GAA AAA GAT GAG TAT-3′ forward and 5′-ATG ATG CTG CTT ACA TGT CTC GAT-3′ reverse primers) in a volume of 50 μl. Amplification conditions for β2-microglobulin and E6/E7 were as follows: 94°C, 15 s; 60°C, 1 min and 72°C, 1 min, for a total of 25 cycles. Amplification conditions for p21 were: 94°C, 30 s; 59°C, 1 min and 72°C, 1 min for a total of 33 cycles. The PCR products were analyzed on a 3.0% agarose gel containing 0.5 μg/ml ethidium bromide and visualized by ultraviolet irradiation.

Transfection and Reporter Gene Analysis. To determine whether p53 protein is functionally impaired in HPV-positive cervical cancer cell lines and whether NCS is able to induce or increase the transcriptional function of endogenous p53 protein, transient transfections were performed using Lipofectamine 2000 reagent (Invitrogen), 4 μg of p53 CON or pGUP.PA.8, and 1 μg of pGFP (green fluorescent protein, as an internal standard) plasmids were transfected into 1 × 10⁶ cells seeded on 60-mm plates 1 day before transfection. After 24 h, transfected cells were treated with 40 nM NCS for 1 h. Twenty-four hours later, cells were lysed directly on the plate in 1× reporter lysis buffer (Promega, Madison, WI). Lysates were centrifugated at 3000g for 2 min at 4°C. Supernatants were analyzed for luciferase activity as recommended in the luciferase assay system protocol (Promega) using a Fluoroskan Ascent FL2.1. Variations in transfection efficiencies were corrected by determination of green fluorescent protein positive cells by flow cytometry and expressed as the -fold increase with respect to untreated control cells. All data shown represent the mean values from multiple independent experiments.

NCS Toxicity Determination in Normal Human Keratinocytes. To test whether NCS affects cell proliferation or survival of normal cells, primary cultures of normal human keratinocytes were compared with HeLa cells. Cellular proliferation was measured by incorporation of 5-bromo-2′-deoxyuridine (BrdU; Roche). Cells were treated with 0, 10, 20, and 40 nM NCS in a 96-well plate, as described above. Seventy-two hours after NCS removal, BrdU was added to a final concentration of 10 mM, and the incubation was continued for 4 h. Cells were processed according to the manufacturer's instructions. Absorbance was measured in an ELISA reader at 370 nm (reference wavelength 492 nm). In parallel, the percentage of apoptotic cells was measured by TUNEL, as described previously.

Results

Inhibition of Cellular Growth by NCS. To determine the optimal concentration of neocarzinostatin that affects cell growth, adherent cell number was measured by counting the number of adherent cells at various times (0, 24, 48, and 72 h) after 1 h exposure to 10, 20, 40, and 80 nM NCS (Fig. 1). Although viable cell number started to decrease 24 h after 20 nM treatment, the maximal inhibitory effect on cellular growth was observed at 48 to 72 h using a concentration of 40 to 80 nM NCS. Similar results were observed when cell proliferation was determined using the BrdU incorporation assay 72 h after NCS exposure. For each cell line, DNA synthesis began to decrease at a concentration of 20 nM NCS (data not shown). Based on these results, we used 40 nM NCS in the following experiments. It was also observed that some cells tend to round up and detach from the dish surface within 48 h after drug treatment (data not shown), leading to fragmentation into smaller bodies that retained membrane integrity characteristics of apoptosis.

Induction of G₂-Cell Cycle Arrest by NCS. To examine a possible block of the cell cycle in HPV-positive or -negative cervical cancer cell lines, the distribution of cells in the different phases of the cycle was studied by flow cytometry following NCS treatment. Most cells remained in the G₁ phase at time 0. After 24 h, the number of cells in G₂ phase increased 2- to 6-fold (Fig. 2), with a corresponding decrease of cells in G₁ phase. For ViBo and C33A, however, a modest change was observed in G₂ phase. Intensity and duration of the G₂ arrest varied between cell lines, being faster and irreversible in both HPV18-positive HeLa and INBL cells, slow in HPV-negative ViBo and C33A cells, and transitory in the CaSki cell line (Fig. 2). The high proportion of cells in G₂ strongly suggests a cellular growth arrest in this phase.

Induction of Apoptosis by NCS. Most cell lines cannot recover from G₂ arrest produced by NCS treatment and eventually die, with the exception of ViBo, which only stops growing. Since it has been observed that the toxicity of DNA damaging agents in several tumor systems appears to be mediated by p53-induced apoptosis, we investigated if these cervical cancer cell lines are eliminated through this process after being exposed to NCS. Untreated and NCS-treated cells were analyzed for apoptosis by DAPI staining and TUNEL assay. NCS treatment resulted in apoptosis of both HPV-positive and C33A cell lines, showing typical features such as chromatin condensation and fragmentation. These changes were observed after approximately 48 h of NCS treatment (data not shown), and the fraction of apoptotic cells, as determined by DAPI staining, increased further 72 h after exposure (Fig. 3A). These results were confirmed by TUNEL assay. After 72 h of NCS treatment, 55 to 83% of the cells underwent apoptosis, except for the cell line ViBo, which was resistant to the induction of this process but finally became senescent (Fig. 3B).

To correlate apoptosis with cell cycle status, cells incubated with TUNEL mixture were counterstained with PI, and flow cytometric analysis was performed. These results indicate that the apoptotic process occurs predominantly at the G₂ phase (Fig. 3C), correlating with the G₂ blocking previously found.

NCS Increases p53 Levels. Expression of p53 and Bcl-2 in cell lines exposed to NCS was investigated by flow cytometry. A significant increase in the percentage of p53-positive cells was observed in the first 24 h of NCS treatment in HPV-positive (HeLa, INBL, and CaSki), as well as ViBo cell lines. In the case of C33A, there was virtually no change in the number of p53 positive cells (Fig. 4A). Additionally, a significant increase in p53 levels was detected in the former cell lines (D > 0.5, Kolmogorov-Smirnov statistics; D, the greatest difference between the two curves), while again for C33A there was no evident change in the levels of this protein (Fig. 4B). No change in Bcl-2 protein levels or the percentage of Bcl-2 positive cells was observed in either INBL (Fig. 4C) or HeLa cell lines (data not shown). In CaSki and C33A cells, however, Bcl-2 protein levels decreased 24 h after NCS treatment but were restored after 48 h, while in ViBo, cells this antiapoptotic protein decreased until 48 h after treatment. These changes were statistically significant (D >0.05, Kolmogorov-Smirnov Statistics).

To confirm these results and to study p53 subcellular localization, cells were grown on coverslips and analyzed by immunocytochemistry (Fig. 5). At time 0, p53 was found in the cytoplasm with predominant perinuclear localization in HPV-positive cells. However, 24 h after NCS treatment, p53 levels increased and the protein accumulated in the nucleus, its normal functional location. Nuclear localization of p53 was observed from time 0 in the HPV-negative ViBo and C33A cell lines, albeit at a very low level in ViBo. Nuclear p53 levels increased further in ViBo but not in C33A cells 24 h after NCS treatment (Fig. 5).

To establish the exact moment of nuclear accumulation, samples from both HeLa and INBL cell lines were taken at 0, 3, 6, 12, 24, and 48h after NCS exposure. In both cell lines, p53 accumulation in the nucleus was evident 6 h after drug treatment, with a subsequent gradual increase (Fig. 6).

Effect of NCS on E6/E7 and p21 Gene Expression. Since NCS treatment resulted in an intranuclear increase of p53 protein levels, kinetic studies were performed to determine whether the effect of NCS includes modification of E6/E7 and p21 mRNA expression. Treatment with NCS resulted in a significant reduction of E6/E7 expression only in the HeLa cell line (HPV18), while in INBL cells, the levels of these transcripts showed no change (Fig. 7). In contrast, RT-PCR kinetic study showed an increase in p21 mRNA levels in HeLa, INBL, CaSki, and ViBo cells, starting at earlier times of NCS treatment (3 h). Down-regulation of E6/E7 expression in HeLa cells was not due to a nonspecific toxic effect since β2-microglobulin transcript levels (internal control) were not altered (Fig. 7), and p21 gene expression was induced under the same experimental conditions. In agreement with previously reported observations (Kastan et al., 1991; Amundson et al., 1998; Sionov and Haupt, 1999; Hietanen et al., 2000), indicating that stress regulation of p53 occurs primarily at the protein level, we observed that p53 mRNA levels were not changed after NCS treatment (data not shown).

NCS Increases p53 Function as Transcriptional Factor. In addition to the increase in both the percentage of p53-positive cells and the nuclear p53 protein levels, NCS also promoted the trans-activation function of this tumor suppressor protein (Fig. 8). After NCS treatment of transfected cells, activation of a p53-responsive promoter (p53CON reporter gene) increased significantly (*p < 0.05) in all cell lines containing wild-type p53 but not in C33A, which contains mutant p53. This result suggests that p53 may trans-activate its target genes to inhibit cell proliferation and induce apoptosis in response to NCS treatment. It also confirms the observed increase in p21 mRNA levels detected after NCS treatment in all HPV-positive cells and HPV-negative ViBo cells and the absence of this effect in C33A cells (Fig. 7).

NCS Does Not Affect Growth of Normal Cells. We compared the effect of NCS on the growth rate of normal human keratinocytes and HeLa cells. No significant inhibitory effect on cellular growth was observed on human keratinocytes even at 40 nM NCS, a concentration that caused significant cell growth decrease in neoplastic cells (Fig. 9A). On the other hand, treatment with 40 nM NCS induces minimal (6.9%) apoptotic death in normal primary keratinocytes compared with HeLa cells (85%), as determined by TUNEL assays (Fig. 9B). Thus, NCS appears to be a good chemotherapeutic candidate specific for cervical cancer cells.

Discussion

In this study, we show that NCS treatment of both HPV-positive and -negative cervical cancer cell lines results in a decrease in the rate of cellular growth and DNA synthesis. In addition, we have determined that, in HPV-positive cells, NCS causes significant cell cycle arrest in G₂ and apoptosis.

Proliferating cells normally respond to DNA damage by delaying the G₁-S and G₂-M cell cycle transition phases or by reducing the rate of replicon initiation in S phase (Kaufmann and Wilson, 1994). In this context, p53 has been involved in G₁ and G₂ checkpoint control through the transactivation of mediator genes like p21, GADD45, and 14–3-3σ (for review, see Amundson et al., 1998; Sionov and Haupt, 1999).

Since p53 plays an important role in the maintenance of genomic stability after DNA damage, we believe that the effects observed after NCS treatment may be attributable to activation of this tumor suppressor gene. It has been observed that in experimentally transformed cells expressing exogenously high risk HPV-E6, the G₁ checkpoint is lost very early, and G₂ checkpoint sometimes is disrupted (Amundson et al., 1998).

With respect to endogenously expressed E6 and its effect on p53, genotoxic agents like mitomycin C, cisplatin, or UV radiation have been shown to cause an increase in nuclear p53 protein levels, followed by an increase in p21 mRNA and G₁ arrest (Butz et al., 1995, 1996). Different cell cycle arrest phenotypes were found among the HPV-positive cell lines studied, even in cells associated with the same type of HPV, which suggests that some HPV-positive cancer cells can efficiently respond to genotoxic stress. In this article, we observed a strong G₂ arrest in HeLa and INBL cell lines, while this effect was reversible in CaSki cells. In contrast, the HPV-negative C33A (mutated p53) and ViBo (wild-type p53) cell lines showed a minor G₂ arrest.

Although previous evidence indicates that E6 protein expressed from strong heterologous promoters can inhibit p53-mediated transcription in vivo, more recent observations show that in HeLa and CaSki cells endogenous p53 is able to activate its target promoters in spite of E6 expression (Butz et al., 1995). Consistent with these observations, we found a basal level of p53 activity and a strong increase in transcriptionally active p53 in both HPV-positive and ViBo cell lines following NCS treatment. In addition, p53 was accumulated in the nucleus and p21 transcript levels increased in spite of minor changes in E6/E7 transcript levels, suggesting an important role for p53 in the anticancer mechanism of NCS. In CaSki cells, however, the high p53-transcriptinal activity (Fig. 8) does not totally correlate with p21 expression (Fig. 7), suggesting a p53-independent induction.

Despite C33A cells underwent apoptotic death, a change either in p53 levels, which were already elevated, or in its transcriptional function after NCS treatment, were not observed. This was expected since this cell line contains a mutated p53 protein with a longer half-life, which induces its nuclear accumulation (Scheffner et al., 1991). Butz et al. (1995) also found that p53 levels in C33A cells were not disturbed after treatment with other genotoxic agents. In accordance with our results in C33A, which indicate a moderate G₂ block after NCS treatment, Paules et al. (1995) observed a G₂ blockage by ionizing radiation in low-passage primary cultures from Li-Fraumeni patients, as well as in human fibroblasts expressing HPV16-E6 and embryonic p53 –/– cells, indicating that the lack of wild type p53 does not compromise the G₂ arrest.

Although wild-type p53 usually sensitizes cells to apoptosis, some p53 mutations preserves sensitivity to radiation and drugs. In fact, C33A cells have shown significant apoptosis in response to γ-irradiation, mitomycin C, and staurosporine (Kilic et al., 1999).

Furthermore, it is important to consider that p53 protein contains several functional domains for inducing cell cycle arrest and apoptosis. While N-terminal activation domain 1 (residues 1–42), activation domain 2 (residues 43–63), and the proline-rich domain are required for inducing cell cycle arrest, the C-terminal basic domain (residues 364–393) and the proline-rich domain (residues 64–91) are required for apoptotic activity (Zhu et al., 2000). Interestingly, C33A has a mutation in residue 273, which is not included within the apoptotic domains. In this respect, it has been shown that, p53 can induce apoptosis independently of its transactivation function, either through a wild-type proline-rich domain (for review see, Sionov and Haupt, 1999) or by direct interaction with mitochondria (Mihara et al., 2003). Besides, participation of alternative mechanisms like p53-independent caspase activation can be considered. Thus, NCS could induce apoptosis in C33A cells through p53-dependent or p53-independent pathways.

Nevertheless, despite the activation of transcriptionally active p53 protein observed in ViBo cells, causing a partial block in G₂, apoptosis was not induced, suggesting that a p53-dependent apoptosis pathway is defective in these cells. There are many tumors that retain wild-type p53 and carry alterations in the downstream mediators of p53 function like defects in Bax induction (Rampino et al., 1997; Vousden, 2002) or the combined loss of p63 and p73, which results in the failure of cells containing functional p53 to undergo apoptosis in response to DNA damage (Flores et al., 2002). On the other hand, it is known that the viral oncoprotein E7 binds to Rb and inactivates its growth suppressive function releasing proapoptotic E2F proteins. In this respect, it has been found that human fibroblasts overexpressing E7 (like HeLa, INBL, and CaSki cases) undergo apoptosis in response to ionizing radiation, while the same cells without E7 (as in ViBo case) would normally undergo growth arrest in response to similar treatment. Thus, ViBo cells lack oncogenic stress signals present in HPV-infected cells (White et al., 1994). These signals could play a major role in NCS-induced apoptosis in HPV-positive cells.

Contrary to our expectations, no significant decrease in anti-apoptotic Bcl-2 protein levels was observed after NCS treatment of INBL or HeLa cells, in spite of the great increase in p53 levels and apoptosis observed. Interestingly, in accordance with this results, Bax and Bcl-2 protein levels did not change after Ad5-p53 transfection of HeLa and SiHa, even though they underwent apoptosis at 48 h post-transfection (Huang et al., 2000).

In contrast to INBL and HeLa cells findings, a transitory decrease in Bcl-2 levels occurred in CaSki and C33A cell lines (24 h after NCS treatment). In agreement with this observation, Liang et al. (2002) demonstrated that NCS-induced apoptosis in cells overexpressing Bcl-2 was associated to the cleavage of this protein promoted by caspase-3. In addition, they found that the 23 kDa cleavage product of Bcl-2 was detected between 4 and 24 h after NCS treatment. This could be the case for CaSki and C33A cells, opening an interesting research possibility in these cell lines.

Several groups have extensively studied NCS, with an emphasis on chemotherapy of hepatic, gastric, and urinary cancers. In this work, we observed that NCS inhibits cell growth and induces apoptosis in cervical cancer cell lines and that these cells are much more sensitive than normal human keratinocytes. In accordance with these results, normal human keratinocytes under genotoxic stress are able to develop a fast adaptive response, absent in transformed keratinocytes, that enhance their resistance to apoptosis (Chouinard et al., 2002). On the contrary, high proliferating cells, like cervical cancer cells (Fig. 9A), are more susceptible to apoptosis by DNA damage agents. Moreover, we have also found a higher level of Bcl-2 protein in HeLa (fluorescence intensity = 349) than in normal keratinocytes (fluorescence intensity = 100) as determined by flow cytometry (data not shown), which should synergize the apoptotic affect of NCS in HeLa cells (Cortazzo and Schor, 1996). On the other hand, in normal keratinocytes, where the Rb pathway remains intact and the proapoptotic E2F-1 factor is consequently down-regulated, the induction of p53 by genotoxic insult is most likely to result in cell-cycle arrest. In cervical tumoral cells, by contrast, in which E2F-1 activity is deregulated and the oncogenic stress is higher than in normal keratinocytes, reactivation wild-type p53 increases E2F-1-induced apoptosis (Bates and Vousden, 1999). Furthermore, HeLa cells are highly sensitive to apoptosis by the Fas pathway, which can be mediated through wild-type p53 (Sionov and Haupt, 1999; Aguilar et al., 2001). Additionally, the difference in apoptosis induction between HeLa and normal keratinocytes could reflect the genetic background of the former, since c-Myc, jun-B, Bcl-2, and c-Ha-ras gene expression is higher in HeLa, and a more efficient DNA-repair mechanism has been reported in normal keratinocytes (Pelisson et al., 1992; Choo et al., 1995; Liang et al., 1995; Rey et al., 1999). Our data are also supported by results found in other cell systems including skin, lung and embryonal fibroblasts, as well as lymphocytes, where normal cells are less sensitive to NCS than their tumoral counterpart (Maeda, 1981). In fact, the Health Ministry and Social Security of Japan approved its use in hepatoma patients in 1993 (Maeda, 2001).

In summary, we have documented NCS as a potent inducer of p53 activity, showing that reactivation of this tumor suppressor in HPV-positive cell lines can be an effective tumor therapy strategy.