Int. J. Cancer: 73, 592–599 (1997) r 1997 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer LACK OF CORRELATION BETWEEN CISPLATIN-INDUCED APOPTOSIS, p53 STATUS AND EXPRESSION OF Bcl-2 FAMILY PROTEINS IN TESTICULAR GERM CELL TUMOUR CELL LINES Herman BURGER, Kees NOOTER*, Antonius W.M. BOERSMA, Christine J. KORTLAND and Gerrit STOTER Department of Medical Oncology, University Hospital Rotterdam and Rotterdam Cancer Institute (Daniel den Hoed Kliniek), Rotterdam, The Netherlands We investigated the role of p53 and of the Bcl-2 family proteins in the apoptotic response of a panel of testicular tumour cell lines (NT2, NCCIT, S2 and 2102 EP). The p53 gene status and the capacity of the p53 protein to transactivate the p21/WAF/CIP gene were determined, and we examined the correlation between p53 status and the susceptibility to cisplatin-induced apoptosis. In contrast to wild-type p53containing NT2 and 2102 EP cells, NCCIT (mutant p53) and S2 (no p53 protein) cells were shown to be p53-transactivation defective. However, NCCIT and S2 cells with nonfunctional p53 were readily triggered into apoptosis by cisplatin, whereas p53-transactivation competent 2102 EP cells failed to undergo cisplatin-induced apoptosis. The defective apoptotic pathway in 2102 EP cells was reflected by a 4-fold decreased sensitivity to cisplatin in the MTT assay. We further demonstrated that the p53-independent differential cisplatin sensitivity among the testicular germ cell tumour (TGCT) cell lines was not due to differences in cellular cisplatin accumulation or DNA platination. The pattern of endogenous expression levels of Bax, Bcl-2, Bcl-x and Bak, which was not modulated by cisplatin treatment, demonstrated that these Bcl-2 family proteins are not involved in drug-induced apoptosis in the TGCT cell lines. Our results suggest a lack of correlation between cisplatin-induced apoptosis, p53 status and expression of Bcl-2 family proteins in our panel of TGCT cell lines. We conclude that the cisplatininduced apoptotic pathway in TGCT cell lines might be p53-independent and is probably not associated with differences in the Bcl-2/Bax rheostat. Int. J. Cancer 73:592–599, 1997. r 1997 Wiley-Liss, Inc. Testicular germ cell tumours (TGCTs) comprise a versatile group of neoplasms often showing complex histological features. They are classified into seminomas (SE), which are composed of the neoplastic counterparts of primordial germ cells, and nonseminomatous (NS) TGCTs, which are neoplastic caricatures of early embryonic development. TGCTs represent one of the few tumour types that are curable by chemotherapy and radiotherapy, with an overall cure rate of approximately 80% (Einhorn, 1990). As yet, the nature of the exceptional sensitivity of testicular tumours to cytoreductive therapy has not been defined. Because it has now been established that ionizing radiation and a large variety of anti-cancer drugs exert their cytotoxic action through the induction of apoptosis, and that inhibition of the apoptotic pathway may lead to cytotoxic drug resistance (reviewed by Fisher, 1994), one might speculate that TGCTs are hypersensitive to treatment-induced apoptosis. Several lines of evidence show that the product of the p53 tumour suppressor gene plays a pivotal role in the sensitivity of tumour cells to apoptosis induced by chemotherapy or radiation therapy (Lowe et al., 1993; Clark et al., 1993; Fisher, 1994; Harris, 1996). In some circumstances, functional inactivation of p53 by mutations or interactions with cellular or viral proteins can lead to resistance to genotoxic agents commonly used in anti-cancer therapies. Indeed, cells from transgenic mice homozygous for p53 null alleles were shown to be more resistant to induction of apoptosis by chemotherapeutic agents than cells from their normal wild-type p53 (wtp53) littermates (Lowe et al., 1993). Another line of evidence for a role of wtp53 in apoptosis comes from in vitro studies which show that mutations in the p53 gene may render cells resistant to induction of apoptosis by ionizing radiation and chemotherapeutics (reviewed by Fisher, 1994). Members of the Bcl-2 family of proteins are involved in the control of apoptosis in a range of different cell types (reviewed by Reed et al., 1996). The bcl-2 gene was first identified at the t(14;18) translocation commonly found in human follicular lymphoma (Korsmeyer, 1992). Several cellular and viral homologs of Bcl-2 have been identified that belong to this rapidly expanding Bcl-2 protein family (Reed et al., 1996). Bcl-2 family proteins can function either as inhibitors (e.g., Bcl-2, Bcl-xL, Mcl-1, A1 and Bag) or promoters (e.g., Bax, Bcl-xS, Bad and Bak) of cell death and can physically interact with each other and subsequently modulate apoptosis (Reed et al., 1996). For instance, Bcl-2 and Bax form homodimers and heterodimers, and the balance between the respective dimers (i.e., Bcl-2/Bcl-2; Bcl-2/Bax; Bax/Bax) determines the extent to which apoptosis is induced or suppressed. It has been proposed that Bax homodimers promote apoptosis and that the Bax-mediated cell death is counteracted by Bcl-2/Bax heterodimerization. Thus, the ratio of Bcl-2/Bax represents one cell-autonomous rheostat that determines the cell’s fate; bcl-2 and bax are immediate early response genes of the p53 tumour suppressor gene (Miyashita et al., 1994). The promoter of the bax gene contains several p53 consensus binding sites, and wtp53, but not mutant p53 (mtp53), can transactivate the expression of the bax gene. In contrast, the bcl-2 gene contains a p53-dependent negative response element through which the p53 protein can function as a repressor of bcl-2 expression. Thus, Bcl-2 and Bax are independently regulated by p53, and its effects on bcl-2 and bax gene expression may determine the vulnerability of cells to apoptotic stimuli. Accordingly, in a panel of human cells with defined p53 status, it has been shown that DNA damage-induced upregulation of Bax and downregulation of Bcl-2 was dependent on the presence of wtp53 (Zhan et al., 1994). Although it is generally accepted that the presence of wtp53 increases the susceptibility of cells to induction of apoptosis by a wide variety of genotoxic agents, probably by modulation of the Bcl-2/Bax rheostat, an increasing amount of evidence indicates that genotoxic insults can trigger apoptosis also via p53-independent mechanisms (Clarke et al., 1993). Some of these p53-independent apoptotic routes, but not all, are also regulated by members of the Bcl-2 protein family. The existence of p53-independent apoptotic pathways likely implies that the regulatory role of p53 in apoptosis is influenced by the particular cellular context in which the product of the p53 tumour suppressor gene is expressed. Hence, p53 is apparently not the only regulator of DNA damage-associated checkpoint(s) of apoptosis. Contract grant sponsor: Dutch Cancer Society; Contract grant number: DDHK 94-846. *Correspondence to: Department of Medical Oncology, University Hospital Rotterdam, Room D337, Dr. Molewaterplein 40, 3015 GD, Rotterdam, The Netherlands. Fax: (31) 10 463 4627. E-mail: [email protected] Received 21 May 1997; Revised 8 July 1997 CISPLATIN-INDUCED APOPTOSIS IN TGCT CELL LINES Clinical data revealed that the vast majority of TGCTs has no mutations in their p53 alleles (Peng et al., 1993); this lack of p53 mutations has been implicated in the high response rate of this neoplasm to combination chemotherapy. Several human NS TGCT cell lines have been established that retain their relative sensitivity to cytotoxic agents, demonstrating that these cell lines are representative models of chemo-sensitive tumours. In the present study, we investigated whether the presence of functional wtp53 is a prerequisite for the onset of apoptosis in TGCTs. We studied the role of p53 and the Bcl-2 family proteins in drug-induced apoptosis in a panel of well-defined TGCT cell lines. Our data suggest that the cisplatin-induced apoptotic pathway in TGCT cell lines is p53independent and most likely not correlated with inherent or drug-induced differences in the Bcl-2/Bax rheostat. MATERIAL AND METHODS TGCT cell lines and culture conditions Four established TGCT cell lines were analyzed. The TGCT cell line NT2 (ATCC CRL-1973) and the 2102 EP cell line (Wang et al., 1981) have been derived from NS, whereas S2 (a gift from A. von Keitz, Marburg, Germany) and NCCIT are 2 cell lines that exhibit some SE-like characteristics (Damjanov et al., 1993). The cell lines were grown as monolayer and maintained at 37°C in a humidified cell culture incubator with 8.5% CO2 in HEPES-buffered RPMI 1640 supplemented with 10% FCS (GIBCO BRL, Paisley, UK), 100 IU/ml penicillin (Sigma-Aldrich, Zwijndrecht, the Netherlands), 100 µg/ml streptomycin (Sigma) and 2 mM L-glutamine (GIBCO BRL). Drug sensitivity assay The MTT colorimetric assay, which measures the number of viable cells capable of reducing the tetrazolium compound (SigmaAldrich, Zwijndvecht, The Netherlands) to a blue formazan product, was used to quantitate the chemosensitivity of the cell lines to cisplatin (cis-diamminedichloroplatinum II). Briefly, cells were harvested during the exponential growth phase and seeded into 96-well (3000 cells/well) tissue culture plates (Microtest III, Falcon 3072, Beckton Dickinson, Lincoln Park, NJ). After overnight pre-incubation at 37°C, serial dilutions of cisplatin (Platosin, Pharmachemie, Haarlem, The Netherlands) were added to quadruplicate wells, and the cells were exposed to the drug for an additional 4 days. Processing of the 96-well microtiterplates and absorbance measurements were performed according to standard procedures. The IC50 and IC90 values, defined as the cisplatin concentration that reduced the absorbance with 50% or 90%, respectively, were estimated graphically from the concentration response curves. Intracellular platinum (Pt) accumulation Triplicate 75-cm2 tissue culture flasks with exponentially growing cells were exposed to 33 µM (10 µg/ml) and 100 µM (30 µg/ml) of cisplatin for 2 hr. Following exposure to drugs, the cells were immediately washed to remove free cisplatin, harvested by trypsinization, washed with ice-cold PBS (3 3 10 ml) and lysed on ice in 500 µl of 0.2% (w/v) Triton-X-100 (Sigma)/H2O. Protein concentration in the resulting lysates was determined using a Bio-Rad (Veenendaal, The Netherlands) protein assay kit. Total Pt content was determined in duplicate by atomic absorption spectrometry (AAS) using a flameless Perkin-Elmer (Foster City, CA) 4110 ZL spectrometer. Intracellular Pt levels were expressed as µg of Pt per mg of protein (µg Pt mg21 protein). Determination of Pt bound to DNA Following cisplatin incubation, triplicate monolayer cultures were washed with PBS (3 3 20 ml), and the cells were lysed in the tissue culture flasks at 37°C for 16 hr with 10 ml of DNA lysis buffer containing 0.5% SDS, 10 mM Tris-HCl (pH 8.2), 400 mM NaCl, 2 mM EDTA and 0.5 mg/ml proteinase K (Boehringer Mannheim, Germany). The salting out procedure for extracting DNA (Miller et al., 1988) was used to prepare genomic DNA. Subsequently, the DNA samples were sonicated at 4°C for 1 hr, and 593 DNA content (absorbance at 260 nm) and Pt-DNA adducts (AAS) were determined. The DNA platination levels were expressed as pg of Pt per µg of DNA (pg Pt µg21 DNA). Induction of apoptosis in TGCT cell lines by cisplatin Cells from exponential phase cultures were used for the induction of apoptosis by cisplatin. The cells were seeded at a density of 106 cells per 75 cm2 in tissue culture flasks and 24 hr later incubated with various concentrations (3.1 µM, 6.3 µM, 12.5 µM and 25 µM) of cisplatin. After incubation with drug (2 hr at 37°C in culture medium), the cells were washed with culture medium and further cultured in drug-free medium for 24, 48 or 72 hr. Microscopical detection of apoptotic cells Apoptotic cells were recognized by the appearance of condensed nuclear chromatin and fragmented nuclei. For the visualization of these features in apoptotic cells, 2 DNA stains Hoechst 33342 and PI were used (both obtained from Calbiochem, La Jolla, CA). Hoechst 33342 is a fluorescent dye that is used to stain DNA structures in viable cells. Because PI can only enter cells with a disrupted cell membrane, this dye can be used both as a DNA stain and concomitantly to determine the integrity of the cell membrane. Cells were simultaneously incubated with Hoechst 33342 (0.5 µg/ml) and PI (2.5 µg/ml) for 20 min and viewed under a fluorescence microscope (Carl Zeiss, Weesp, The Netherlands); micrographs were taken according to standard procedures. Quantification of apoptotic cells by annexin V labeling Single cell suspensions for annexin V labeling were obtained by trypsinization, washed with culture medium containing 10% FCS and thereafter washed twice with culture medium without phenol red. Annexin V labeling of cells was performed as described previously (Boersma et al., 1996). Briefly, floating and adherent cells were incubated for 1 hr at 37°C in HEPES (10 mM HEPES/NaOH, pH 7.4) buffered RPMI (GIBCO BRL) culture medium (without phenol red) supplemented with 0.5 µg/ml FITCconjugated annexin V (BioWhittaker, Verviers, Belgium), washed twice in culture medium and kept on ice until further processing. Flow cytometry was performed on a FACScan flow cytometer (Beckton Dickinson, San Jose, CA) tuned at 488 nm. The FITC fluorescence (515–545 nm) was measured in logarithmic mode, whereas forward light scatter (FLS) and perpendicular light scatter (PLS) were measured in linear mode. Cell debris was excluded from analysis by appropriate FLS threshold setting. Irradiation To investigate the effect of irradiation on p53 and p21/CIP/WAF expression, TGCT cell lines were g-irradiated (10 Gy) in 75 cm2 tissue culture flasks positioned between 2 opposing 137Cs sources (Gamma Cell 40, Atomic Energy of Canada, Ottawa, Canada) at a dose rate of 1.06–1.08 Gy/min. Cells were seeded at a density of 2 3 106 cells/tissue culture flask and 24 hr later used for the irradiation experiment. Irradiated cells were incubated for an additional 6–24 hr and thereafter further processed for Western and Northern blot analysis. Immunoblotting Expression of p53, Bcl-2, Bcl-x, Bak and Bax proteins was assessed in the TGCT cell lines by Western blot analysis. After exposure to cisplatin, both detached and adherent cells were combined and washed with ice-cold PBS (3 3 10 ml). Cell pellets were resuspended in 10 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM EDTA, 1% Triton-X-100 and the protease inhibitors phenylmethylsulfonyl fluoride (1 mM PMSF; Boehringer Mannheim), aprotinin (0.23 U/ml; Sigma) and leupeptin (10 µM; Boehringer Mannheim). Samples containing 20 µg of protein in loading buffer (50 mM Tris-HCl; pH 6.8, 0.1% bromophenol blue, 2% SDS, 10% glycerol, 100 mM dithioerythritol) were boiled for 3 min, subjected to SDS-PAGE (12%) and transferred to Immobilon-P transfer membrane (Millipore, Bedford, MA) using a semi-dry blotting system (Hoefer, San Francisco, CA) with a continuous buffer BURGER ET AL. 594 system (39 mM glycine, 48 mM Tris, 0.0375% SDS and 20% methanol) at 0.8 mA/cm2 for 90 min. Membranes were preincubated in 5% (w/v) non-fat dry milk (Protifar, Nutricia, Zoetermeer, the Netherlands) in TBS (50 mM Tris, pH 7.5; 150 mM NaCl) for 1 hr, washed and subsequently incubated overnight at 4°C in TBST (2% Tween 20 in TBS) supplemented with specific monoclonal antibody (MAb) or polyclonal antibody (PAb). Bcl-2 specific mouse IgG1 MAb (100), p53 specific mouse IgG2a MAb (DO-1), Bax specific rabbit IgG PAb (N-20), Bcl-xS/L specific rabbit IgG PAb (S-18) and Bak specific rabbit IgG PAb (G-23) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and used at a dilution of 1:5,000 (Bax) or 1:1,000 (p53, Bcl-2, Bcl-xS/L, and Bak). Poly(ADP-ribose) polymerase (PARP) specific mouse MAb (C-2-10) and Bax specific mouse IgG2b MAb were purchased respectively from BIOMOL (Plymouth Meeting, PA) and Immunotech (Coulter, Mijdrecht, The Netherlands). MAb were both used at a dilution of 1:5,000. Immunological complexes were visualized by enhanced chemiluminescence (Pierce, Cambridge, UK) using horseradish peroxidase-conjugated goat anti-mouse or goat antirabbit IgG (1:4,000; Santa Cruz Biotechnology). Isolation and sequencing of p53 cDNAs To establish the p53 DNA sequences of the 4 studied TGCT cell lines, total cellular RNA was isolated according to standard procedures and reverse-transcribed using antisense p53 primers. The single stranded p53 cDNA was subsequently amplified by the polymerase chain reaction using 4 different sets of p53-specific primers. The 4 p53 cDNA fragments (fragment 1: nt 137–486 corresponding to aa 1–117; fragment 2: nt 411–762 corresponding to aa 93–202; fragment 3: nt 694–1029 corresponding to aa 188–291; fragment 4: nt 993–1317 corresponding to aa 287–387) were completely sequenced according to the chain-termination method, and p53 nt and aa positions are as described by ZakutHouri et al. (1985). Northern blot analysis Total cytoplasmic RNA was isolated and subsequently equal amounts of RNA (20 µg/lane) were loaded on 1% agarose gels, size-fractionated by gel electrophoresis and blotted onto Hybond N1 nylon membranes (Amersham, Aylesbury, UK) according to standard procedures. Blots were hybridized with [a-32P] dATP (Amersham) radiolabeled cDNA probes according to the recommendation of the Random Primed DNA Labeling Kit (Boehringer Mannheim). The used cDNA probes were the NotI fragment purified from plasmid pCEP-WAF-1 and the b-actin probe. RESULTS Sensitivity to cisplatin and growth characteristics The sensitivity of the 4 TGCT cell lines for cisplatin-induced growth inhibition was estimated by the MTT assay. The sensitivity of these cell lines for cisplatin expressed in IC50 values ranged between 1.17 6 0.59 µM (mean 6 standard deviation [SD]) of cisplatin for the NT2 cell line and 4.01 6 0.75 µM for the 2102 EP cell line (Table I). Based on these IC50 values, 2102 EP cells were approximately 4-fold less sensitive to cisplatin compared with NT2 cells. The IC90 values for these TGCT cell lines show an identical pattern of cisplatin sensitivity. We further investigated whether differences in growth characteristics could account for the differential chemosensitivity of these cell lines. The in vitro doubling times of the TGCT cell lines were in the same order of magnitude (Table I), and cell cycle analysis of exponentially growing cells showed no differences in the distribution over the different cell cycle phases (G1, S, G2/M) between the TGCT cell lines (data not shown). Apparently, the differential cisplatin sensitivity of the TGCT cell lines is not due to differences in their growth characteristics. Cisplatin accumulation and DNA platination To investigate whether differences in cisplatin accumulation could account for the differential chemosensitivity among the TGCT cell lines, we determined the intracellular Pt accumulation for the different TGCT cell lines (Table I). Cells were incubated for 2 hr with 10 µg/ml (33.3 µM) or 30 µg/ml (100 µM) of cisplatin, and the intracellular Pt concentration was quantitated by AAS. All cell lines display an approximately linear correlation between incubation dose and accumulation. NT2 and 2102 EP cells accumulated about equal amounts of cisplatin, while these cell lines showed a 4-fold difference in cisplatin sensitivity as determined by the MTT assay. The intracellular Pt concentration of cisplatin-treated S2 and NCCIT cells were, respectively, approximately 1.5- and 2-fold higher than in NT2 or 2102 EP cells. Thus, the intracellular Pt levels were not correlated with the cisplatin sensitivity of these TGCT cell lines. We also determined the total DNA platination in the TGCT cell lines (Table I). Similar to the accumulation experiments, cells were incubated with 10 or 30 µg/ml of cisplatin and analyzed for Pt-DNA adduct formation. The DNA platination data show that the small differences in accumulation are completely paralleled by similar differences in Pt-DNA adduct formation. As for the intracellular Pt accumulation, no clear correlation between total DNA lesions in the overall genome and cisplatin sensitivity of these TGCT cell lines was noted. Moreover, equimolar doses of cisplatin in NT2 and 2102 EP cells were found to be equi-damaging with respect to Pt-DNA adduct formation but showed a 4-fold difference in drug sensitivity as measured by the MTT assay. These results indicate that the differential cisplatin sensitivity of the TGCT cell lines is not attributed to differences in DNA platination. Induction of apoptosis To evaluate whether observed differences in cisplatin sensitivity of these TGCT cell lines are related to their proneness to drug-induced apoptosis, we determined the threshold of susceptibility for induction of apoptosis by cisplatin. Apoptosis was examined morphologically by Hoechst/PI staining, assessed by immunoblot analysis of proteolytic cleavage of PARP and quantitated flow cytometrically by annexin V labeling. TABLE I – CHARACTERISTICS OF TGCT CELL LINES Cell line DT (h) NT2 NCCIT S2 2102 EP 33.5 6 9.8 32.4 6 9.5 30.7 6 6.0 24.1 6 1.4 IC50 (µM) 1 1.17 6 0.59 1.99 6 0.64 2.75 6 0.62 4.01 6 0.75 IC90 (µM) 1 Cisplatin accumulation (ng Pt mg21 protein) 2 DNA platination (pg Pt µg21 DNA) 3 p534 8.49 6 4.43 7.04 6 0.28 15.18 6 2.94 27.24 6 3.63 26.0 6 3.1 (85.5 6 8.5) 39.6 6 5.1 (117.3 6 1.8) 52.0 6 6.3 (193.6 6 9.8) 23.6 6 2.8 (84.8 6 3.7) 33.1 6 3.1 (76.3 6 8.2) 30.8 6 6.7 (98.4 6 9.7) 37.2 6 4.9 (132.9 6 8.9) 24.6 6 8.3 (79.2 6 6.7) wt/wt mt/2 wt/wt wt/wt 1IC and IC are those drug concentrations effective in inhibiting 50% and 90%, respectively, of the cell growth measured by MTT assay after 50 90 4 days of continuous exposure to the drug. The MTT data were derived from 3 independent experiments.–2Total intracellular Pt concentration (mean 6 SD) after 2-hr exposure of the cells to 10 µg/ml of cisplatin. Values in parentheses are the accumulation data of 30 µg/ml of cisplatin exposure. The Pt accumulation data were derived from 3 independent experiments.–3Total amount of Pt bound to DNA measured by AAS after 2-hr exposure of the cells to 10 µg/ml of cisplatin. Values in parentheses are the accumulation data of 30 µg/ml of cisplatin exposure. The Pt/DNA adducts data were derived from 3 independent experiments.–4The p53 gene status of the TGCT cell lines was determined by sequencing their complete p53 cDNA. Sequence data revealed that NCCIT cells are hemizygous for p53 containing 1 mutated allele carrying 1 bp deletion.–DT, In vitro doubling time. CISPLATIN-INDUCED APOPTOSIS IN TGCT CELL LINES The induction of apoptosis was recognized by morphological changes such as chromatin condensation and nuclear fragmentation. These typical nuclear changes, visualized by staining the cells with Hoechst 33342 (blue) and PI (pink), were monitored with a fluorescence microscope and were apparent 48–72 hr after cisplatin treatment. Spontaneous apoptosis in TGCT cell cultures was low (,3%). Micrographs of representative samples illustrate that the vast majority of cisplatin-treated NT2, NCCIT and S2 cells have an apoptotic appearance (Fig. 1a). In contrast, cisplatin-treated 2102 EP cells did not differ from untreated cells; even at a dose of 25 µM, only approximately 8% of cells exhibited apoptotic nuclei. A family of cysteine proteases showing structural homology to interleukin-1b-converting enzyme (ICE) have been implicated in the initiation of the active phase of apoptosis through specific proteolytic cleavage of essential substrates including PARP, a nuclear enzyme involved in DNA repair (Martin and Green, 1995). Therefore, we studied the proteolytic cleavage of PARP using a PARP specific MAb (C-2-10) that recognizes both 116 kDa PARP and the 85 kDa apoptosis-related cleavage fragment. The presence of 85 kDa proteolytic PARP fragment was demonstrated by Western blot analysis in cell lysates of cisplatin-treated NT2, NCCIT and S2 cultures (Fig. 1b). Although the intact 116 kDa PARP protein was readily detected, the apoptosis-specific 85 kDa PARP fragment was not present in drug-treated 2102 EP cells. The apparent absence of proteolytic cleavage of PARP confirmed the morphological data and indicates that the 2102 EP TGCT cell line is relatively resistant to chemotherapy-induced apoptosis. An established approach for the quantification of apoptotic cells is based on the fact that the surface of apoptotic cells differs from normal cells in that phosphatidylserine (PS) is aberrantly exposed on the external face of the cell membrane. Annexin V binds preferentially to PS, which can be used to quantitate the early phases of the apoptotic process (Boersma et al., 1996). Representative FITC frequency histograms of the 4 TGCT cell lines (Fig. 1c) show that untreated control cells had a mean fluorescence intensity of approximately 20–30 arbitrary units (a.u.) and illustrate that the apoptotic cells are represented by an additional distinct subpopulation with a fluorescence intensity $100 a.u. (Boersma et al., 1996). Nearly all NT2, NCCIT and S2 cells that were still present in the drug-treated culture were annexin-V–positive, suggesting massive induction of apoptosis in these TGCT cell lines. In contrast, hardly any of the treated 2102 EP cells showed increased ($100 a.u.) annexin V staining, suggesting that the vast majority of 2102 EP cells had failed to undergo drug-induced apoptosis. At 72 hr after cisplatin treatment (compared with 48 hr), the percentage of annexin-V–positive 2102 EP did not increase, suggesting that the apoptotic response in 2102 EP was not simply delayed. In summary, morphological studies, analysis of PARP cleavage and annexin V labeling results support the concept that the 2102 EP cell line, compared with the other TGCT cell lines, is significantly less sensitive to cisplatin-induced apoptosis. p53 status (wt vs. mt) of the TGCT cell lines The p53 gene status of the TGCT cell lines was determined by sequence analysis of the complete p53 cDNA. The sequence data revealed that NT2, S2 and 2102 EP cells are homozygous for wtp53 (Table I). In contrast, NCCIT cells were shown to be hemizygous for p53 containing 1 mutated allele carrying a 1 bp deletion. This deletion (G at nt position 949; codon 272) results in a frameshift mutation leading to altered downstream codons and eventually to a stopcodon at position 1114. This mutated p53 allele encodes a truncated p53 protein of 347 amino acids. Concomitantly, NCCIT p53 protein migrates faster than wtp53 protein as detected by Western blot analysis using the DO-1 MAb (Fig. 2a). NT2 and 2102 EP cells express similar basal levels of wtp53 protein. Surprisingly, no endogenous expression of p53 protein was detected in S2 cells despite the presence of wtp53 alleles. 595 Effects of cisplatin on p53 protein accumulation and p21/WAF/CIP expression The p53 response to genotoxic insults was investigated by monitoring p53 protein levels in cisplatin-treated TGCT cells using Western blot analysis (Fig. 2b). In contrast to NT2 and 2102 EP cells that show a normal p53 response, NCCIT and S2 cells did not show a significant increase in the level of p53 protein after cisplatin treatment. The presence of transactivation-competent p53 protein was determined by its ability to activate the p21/WAF/CIP gene, whose expression is regulated at the transcriptional level by the wtp53 but not the mtp53 protein (Harris, 1996). Indeed, p21/WAF/CIP mRNA was found to be significantly upregulated 6–12 hr following g-irradiation (10 Gy) in wtp53-expressing NT2 and 2102 EP cells (Fig. 3). No increased p21/WAF/CIP mRNA levels were detected in p53 non-expressing S2 cells, and a complete absence of p21/WAF/ CIP mRNA expression was found in mtp53-containing NCCIT cells. A study on the time course of p21/WAF/CIP expression demonstrated that the absence of detectable p21/WAF/CIP mRNA levels at 6 hr after irradiation was not simply the result of an inappropriate sampling time. Apparently, NCCIT and S2 cells are p53-transactivation deficient. Expression of different members of the Bcl-2 protein family Several studies have implicated members of the rapidly expanding Bcl-2 protein family as important components of the apoptosis pathway in a wide range of different cell types (Reed et al., 1996). As an initial approach to determine the possible involvement of the Bcl-2 family proteins in the apoptotic pathway of TGCT cells, we examined the constitutive protein expression of Bax, Bcl-2, Bcl-x and Bak by Western blot analysis (Fig. 4). Rather high but comparable endogenous levels of Bax protein were found in all TGCT cell lines. Identical results were found with a MAb directed against the Bax protein that became available only recently. No Bcl-2 protein expression was found in NCCIT cells, whereas comparable Bcl-2 protein levels were found in the other TGCT cell lines. The endogenous protein levels of Bcl-xL, the long form of Bcl-x that can heterodimerize with Bax and consequently influences the Bcl-2/Bax rheostat, were also shown to be comparable. In addition, the short form of Bcl-x (Bcl-xS ) was never detected. Bak expression was found to be somewhat more heterogeneous among the studied TGCT cell lines but showed no apparent association with induction of apoptosis. These studies on the intrinsic expression of Bcl-2 family proteins revealed no correlation between the Bcl-2/Bax ratio and the induction of apoptosis. The constitutive expression levels of death-suppressing and death-promoting genes are inherent to the cell. Because genotoxic stress-induced DNA damage may modulate these expression levels, we examined the effect of cisplatin treatment on the expression of the Bcl-2 family proteins. The TGCT cells were incubated for 2 hr with different doses of cisplatin (3.1–12.5 µM); at 6, 12, 24 and 48 hr after treatment, the expression level of these death-related proteins was determined by Western blot analysis (Fig. 4). Notably, the endogenous expression of Bax, Bcl-2, Bcl-x and Bak was not affected by cisplatin at all time points studied. DISCUSSION TGCTs represent one of the few types of cancer that are curable by chemotherapy and radiotherapy (Einhorn, 1990). Consistently, most TGCT cell lines display an unusually high sensitivity to cytotoxic agents. Analysis of potentially relevant parameters, including cellular detoxification mechanisms (e.g., the glutathione and the metallothionein system), Pt accumulation, DNA platination and repair as well as topoisomerases activity, did not elucidate the nature of the exceptional sensitivity of TGCTs to cyto-reductive therapy (Masters et al., 1993; Sark et al., 1995). Most TGCT cell lines are prone to drug-induced apoptosis that may be an important determinant in the chemotherapeutic response of this type of cancer (Cresta et al., 1996). However, the cellular components that 596 BURGER ET AL. FIGURE 1 CISPLATIN-INDUCED APOPTOSIS IN TGCT CELL LINES determine the low threshold for apoptosis induction have not been defined, although it has been suggested that p53 status plays a pivotal role (Peng et al., 1993; Cresta et al., 1996). In the present study, we investigated the role of p53 and members of the Bcl-2 protein family in drug-induced apoptosis in a panel of TGCT cell lines. The induction of apoptosis, recognized morphologically by chromatin condensation and nuclear fragmentation, was apparent 48–72 hr after cisplatin treatment in cultures of NT2, NCCIT and S2 cells, but not in 2102 EP cells (Fig. 1a). Accordingly, the apoptosis-associated proteolytic cleavage of PARP was readily detected in drug-treated NT2, NCCIT and S2 cells, although it was absent in 2102 EP cells (Fig. 1b). Quantification of the apoptotic process by annexin V labeling demonstrated that in contrast to the other TGCT cell lines, the vast majority of 2102 EP cells are resistant to the induction of apoptosis (Fig. 1c). The insensitivity of 2102 EP cells to drug-induced apoptosis correlates with the MTT data (Table I). Cisplatin resistance has been associated with reduced intracellular accumulation. However, chemosensitive NT2 and chemoresistant 2102 EP cells accumulated approximately equal amount of cisplatin and showed no difference in DNA platination. The observed intracellular Pt levels in these TGCT cell lines excluded the possibility that increased efflux as a result of functional overexpression of an ATP-dependent glutathione S-conjugate pump (Ishikawa et al., 1994) accounted for the differential chemosensitivity among these TGCT cell lines. The hypersensitivity of TGCTs may be related to a defective capacity to remove (repair) Pt-DNA adducts (Hill et al., 1994). However, the identical DNA platination in chemosensitive NT2 and chemoresistant 2102 EP cells was accompanied by an identical capacity to remove the Pt lesions from the genome (data not shown), indicating that differential DNA repair is presumably not involved. In line with our observations, the absence of a correlation between DNA repair and chemosensitivity in TGCT cell lines has been reported by others as well (Köberle et al., 1996). From these data we concluded that the observed differential cisplatin sensitivity of the TGCT cell lines was not associated with differences in growth characteristics, Pt accumulation, Pt-DNA adduct formation or DNA repair, but probably has to be attributed merely to different thresholds to drug-induced apoptosis. The product of the p53 tumour suppressor gene is involved in multiple cellular processes, including gene transcription, DNA repair, genomic stability, cell cycle control and apoptosis (reviewed by Harris, 1996). The status of p53 is thought to be an important mediator in the cellular response to chemotherapy. With respect to testicular tumours, it has been suggested that the absence of mutations in the p53 gene accounts for the hypersensitivity of FIGURE 1 – Induction of apoptosis in TGCT cell lines by cisplatin. (a) Typical morphological changes associated with apoptosis were examined by Hoechst/PI staining and visualized by fluorescence microscopy analysis. Representative micrographs of untreated (top) and cisplatin-treated (12.5 µM cisplatin, 2 hr at 37°C) TGCT cell cultures (bottom) at 72 hr after treatment are shown. The different TGCT cell lines are indicated. Scale bar 5 30 µm. (b) Western blot analysis of cisplatin-induced cleavage of PARP in TGCT cell lines. Cells were incubated in the absence (2) and presence (1) of drug (12.5 µM cisplatin, 2 hr at 37°C). At 48 hr after treatment, the integrity of PARP was monitored by immunoblotting. The different TGCT cell lines are indicated. A Western blotting control (C) to detect specific PARP cleavage was included and represents cell extracts of untreated (2) and etoposide-treated (1) human HL60 leukemia cells. The results of 1 experiment are typical of 3 independent replicates. (c) Annexin V binding histograms of cisplatin-treated TGCT cells. At 48 hr after drug treatment (12.5 µM cisplatin, 2 hr at 37°C), the cells were stained with FITC-conjugated annexin V. The binding of annexin V was quantitated by flow cytometry, and the fluorescence intensity was expressed in arbitrary units (a.u.). Solid lines represent untreated (control) cells, and dotted lines represent drug-treated cells. Apoptotic cells with a high fluorescence intensity ($100 a.u.) are indicated by the shaded areas. The different TGCT cell lines are indicated. The data shown are representative univariate histograms. 597 FIGURE 2 – Western blot analysis of p53 protein levels in TGCT cell lines. Basal levels of p53 protein (a) and the effects of cisplatin treatment (12.5 µM cisplatin, 2 hr at 37°C) on endogenous p53 protein levels (b) are shown. Cells were incubated in the absence (2) and presence (1) of cisplatin; at 48 hr after drug treatment, the p53 protein levels were monitored by immunoblotting. The different TGCT cell lines are indicated. FIGURE 3 – Northern blot analysis of endogenous (2) and radiationinduced p21/WAF/CIP mRNA levels (1) in TGCT cell lines. Total RNA was isolated from the cells at 6 hr after g-irradiation (10 Gy). Equal amounts of RNA (10 µg/lane) were size fractionated on 1% agarose gels, transferred onto a nylon membrane and probed with the NotI fragment of the human p21/WAF/CIP cDNA. The different TGCT cell lines are indicated. Equal intensity of ethidium bromide staining of ribosomal RNA bands and the hybridizing results of b-actin confirmed the presence of equal amounts of RNA within each lane. Autoradiographs were exposed for 16 hr with intensifying screens. TGCT cells to genotoxic agents (Cresta et al., 1996). Therefore, we evaluated the role of p53 in drug-induced apoptosis in our panel of TGCT cell lines. In contrast to wtp53-containing NT2, S2 and 2102 EP cells, NCCIT cells were found to be hemizygous for p53 containing 1 mutated allele carrying a 1 bp deletion at position 949 of codon 272. Consequently, NCCIT cells express a truncated p53 protein of 347 amino acids that lacks the tetramerization, nuclear localization and DNA damage recognition site (Harris, 1996). The carboxy-terminus encoded by exons 9–11 of the p53 protein, which is absent in the truncated p53 protein of NCCIT cells, has been implicated in the induction of apoptosis (Harris, 1996). However, 598 BURGER ET AL. FIGURE 4 – Western blot analysis of endogenous and cisplatin-induced Bax, Bcl-2, Bcl-x and Bak protein levels in TGCT cell lines. Cells were incubated in the absence (c) and presence of cisplatin (3.1, 6.3, 12.5 µM); at 48 hr after drug treatment, the Bcl-2 family protein levels were monitored by immunoblotting. The Bak PAb reacted, apart from the Bak specific band (Mr: approx. 24,000), with an Mr: approx. 21,000 band that possibly represents a proteolytic degradation product or an alternatively spliced variant of Bak. Coomassie staining of duplicate blots showed that equivalent amounts of protein were present in all samples analyzed. The different TGCT cell lines are indicated. Molecular weight markers are indicated in kDa at the left, and the arrows indicate the respective position of the Bcl-2 family protein. we clearly demonstrated that mtp53 NCCIT and wtp53 NT2 cells are equally sensitive to drug-induced apoptosis. Although no basal level of p53 protein was detected in S2 cells, despite the presence of p53 mRNA, drug-induced apoptosis was evident. Whether S2 cells have an impaired translational machinery or whether the p53 protein is specifically inactivated by cellular or viral proteins remains to be elucidated. In contrast to wtp53 NT2 and 2102 EP cells, the endogenous p53 protein levels of NCCIT and S2 cells were not affected by cisplatin treatment or g-irradiation, indicating that NCCIT and S2 cells have a general defect in their p53 response to genotoxic insults. Consistently, NT2 and 2102 EP cells normally transactivated the expression of p21/WAF/CIP after DNAdamaging treatment, although no increase was observed in NCCIT and S2 cells (Fig. 3). These results indicate that NT2 and 2102 EP cells express functional wtp53, whereas NCCIT and S2 cells are deficient in p53-mediated transactivation. Obviously, no correlation between p53 status and induction of apoptosis was detected, suggesting the presence of a p53-independent apoptotic pathway in TGCTs. Previously, p53-independent apoptosis has been observed in a variety of tumour cell types. Although it is clear that chemotherapy- and radiation-induced apoptosis can proceed despite the absence of functional p53, little is known about the activation and signaling pathway involved in p53-independent apoptosis. Our data might imply that the onset of apoptosis in TGCTs is p53-independent. In contrast to this hypothesis, Cresta et al. (1996) concluded that the hypersensitivity of TGCT cell lines to etoposide-induced apoptosis was associated with functional p53. However, this conclusion was primarily based on differential apoptosis induction between chemo-sensitive wtp53 TGCT cell lines and chemo-resistant bladder cell lines expressing mtp53, rather than differential sensitivity among TGCT cell lines with different p53 status. Alternatively, although less likely, the onset of cisplatin- and etoposide-induced apoptosis may proceed through different apoptotic pathways. Notably, preliminary results showed that the p53-independent induction of apoptosis in TGCT cell lines is not cisplatin-specific. Both the NCCIT cell line expressing transactivation-deficient mtp53 and the NT2 cell line endowed with wtp53 alleles were shown to be highly susceptible to radiationinduced apoptosis, whereas the p53-transactivation competent 2102 EP cell line appeared to be resistant to induction of apoptosis by g-radiation. Members of the Bcl-2 protein family have been implicated as important components of the apoptosis pathway in a wide range of different cell types (reviewed by Reed et al., 1996). Based on the capacity of these proteins to form dimers, a simple competitive binding model has been established in which the Bcl-2/Bax ratio determines the extent to which apoptosis is induced or suppressed (Reed et al., 1996). Our data, as presented in Figure 4, reveal that the endogenous protein levels of the studied members of the Bcl-2 protein family (Bcl-2, Bax, Bcl-x and Bak) were not associated with the differential susceptibility to drug-induced apoptosis. Moreover, the expression of the Bcl-2 family proteins was not affected by genotoxic stress, indicating that the Bcl-2/Bax rheostat is not involved in the p53-independent induction of apoptosis in TGCT cell lines. Although no Bax upregulation was apparent by immunoblotting, we previously demonstrated heterogeneous Bax expression and upregulation within the apoptotic NT2 population CISPLATIN-INDUCED APOPTOSIS IN TGCT CELL LINES by flow cytometry (Boersma et al., 1997). However, no Bax upregulation could be detected by flow cytometry in the other TGCT cell lines (data not shown). In line with our results, Bax protein was not increased after etoposide treatment in any of the 6 TGCT cell lines studied by Cresta et al. (1996). In addition, high levels of ectopically expressed Bcl-2 protein in S2 cells transfected with the complete bcl-2 cDNA were not sufficient to inhibit cisplatin-induced apoptosis in this TGCT cell line (data not shown). Our findings in TGCT cell lines are in accordance with a previous report demonstrating that differential sensitivity of Burkitt’s lymphoma cell lines to g-irradiation is independent of the status of p53 and the expression of Bax, Bcl-2 and Bcl-x (Khanna et al., 1996). In contrast, it has been suggested that inherent protein levels of Bax and Bcl-2 in TGCT cell lines determine the threshold of susceptibility to apoptosis induction by etoposide (Cresta et al., 1996). However, this premise was primarily based on differences in Bcl-2 expression between chemosensitive TGCT cell lines and chemoresistant bladder cell lines. From our present results on the cisplatin-induced apoptotic pathway in TGCT cell lines, we concluded that the susceptibility to apoptosis induction does not appear to be related to p53 status, to endogenous protein levels of the members of the Bcl-2 protein family and to drug-induced modulation of the expression of Bcl-2 family proteins. This report demonstrates that drug-induced apoptosis in TGCT cell lines is not regulated by Bcl-2 family proteins and can proceed in the absence of functional wtp53. The results of our study and the finding that alterations of the p53 gene in carcinoma 599 in situ of the testis are more common than previously thought (Kuczyk et al., 1996) may indicate that the presence of functional wtp53 is not required for the successful treatment of TGCTs. Furthermore, p53 inactivation in both experimental and human tumorigenesis is believed to be a late event and superimposed on a series of progressive genetic abnormalities that may all influence the outcome of anti-cancer therapies (Fisher, 1994). The existence of many Bcl-2 family proteins and the fact that new members are still being identified may suggest that a unified model of apoptosis is more complex than the rather simple Bcl-2/Bax competitive binding model or that multiple models of apoptosis exist. Apparently, distinct cellular thresholds for induction of apoptosis exist among different tumour types. Thus, further unraveling of the apoptotic pathway and identification of key determinants of chemo-sensitivity in TGCTs may lead to more appropriate and successful anti-cancer treatment modalities for tumour types that are now considered to be drug resistant. ACKNOWLEDGMENTS We thank Dr. B. Vogelstein (The Johns Hopkins University School of Medicine, Baltimore, MD) for providing the pCEPWAF-1 plasmid. We also thank Mr. E.C.E. Brouwer for assistance with Pt determinations by AAS and Mr. R.H.A.M. Vossen (Laboratory of Anthropogenetics, Sylvius Laboratories, Leiden University, Leiden, The Netherlands) for his help with p53 sequence determinations. REFERENCES BOERSMA, A.W.M., NOOTER, K., BURGER, H., KORTLAND, C.J. and STOTER, G., Bax upregulation is an early event in cisplatin-induced apoptosis in human testicular germ cell tumour cell line NT2, as quantitated by flow cytometry. Cytometry, 27, 275–282 (1997). BOERSMA, A.W.M., NOOTER, K., OOSTRUM, R.G. and STOTER, G., Quantification of apoptotic cells with FITC-labeled Annexin V in CHO cells treated with cisplatin. Cytometry, 24, 123–130 (1996). CLARKE, A.R., PURDIE, C.A., HARRISON, D.J., MORRIS, R.G., BIRD, C.C., HOOPER, M.L. and WYLLIE, A.H., Thymocyte apoptosis induced by p53-dependent and independent pathways. Nature (Lond.), 362, 849–852 (1993). CRESTA, C.M., MASTERS, J.R.W. and HICKMAN, J.A., Hypersensitivity of human testicular tumours to etoposide-induced apoptosis is associated with functional p53 and a high Bax:Bcl-2 ratio. Cancer Res., 56, 1834–1841 (1996). DAMJANOV, I., HORVAT, B. and GIBAS, Z., Retinoic acid-induced differentiation of the developmentally pluripotent human germ cell tumour-derived cell line, NCCIT. Lab. Invest., 68, 220–232 (1993). EINHORN, L.H., Treatment of testicular cancer: a new and improved model. J. clin. Oncol., 8, 1777–1781 (1990). FISHER, D.E., Apoptosis in cancer therapy: crossing the threshold. Cell, 78, 539–542 (1994). HARRIS, C.C., Structure and function of the p53 tumour suppressor gene: clues for rational cancer therapeutic strategies. J. nat. Cancer Inst., 88, 1442–1452 (1996). HILL, B.T., SCANLON, K.J., HANSSON, J., HARSTRICK, A., PERA, M., FICHTINGER-SCHEPMAN, A.M.J. and SHELLARD, S.A., Deficient repair of cisplatin-DNA adducts identified in human testicular teratoma cell lines established from tumours from untreated patients. Europ. J. Cancer, 30A, 832–837 (1994). ISHIKAWA, T., WRIGHT, C.D. and ISHIZUKA, H., GS-X pump is functionally overexpressed in cis-diamminnedichloroplatinum(II)-resistant human leukemia HL-60 cells and down-regulated by cell differentiation. J. biol. Chem., 269, 29085–29093 (1994). KHANNA, K.K., WIE, T., SONG, Q., BURROWS, S.R., MOSS, D.J., KRAJEWSKI, S., REED, J.C. and LAVIN, M.F., Expression of p53, bcl-2, bax, bcl-x2 and c-myc in radiation-induced apoptosis in Burkitt’s lymphoma cells. Cell Death Differentiation, 3, 315–322 (1996). KÖBERLE, B., PAYNE, J., GRIMALDI, K.A., HARTLEY, J.A. and MASTERS, J.R.W., DNA repair in cisplatin-sensitive and resistant human cell lines measured in specific genes by quantitative polymerase chain reaction. Biochem. Pharmacol., 52, 1729–1734 (1996). KORSMEYER, S.J., Bcl-2 initiates a new category of oncogenes: regulators of cell death. Blood, 80, 879–886 (1992). KUCZYK, M.A., SERTH, J., BOKEMEYER, C., JONASSEN, J., MACHTENS, S., WERNER, M. and JONAS, U., Alterations of the p53 tumour suppressor gene in carcinoma in situ of the testis. Cancer, 78, 1958–1966 (1996). LOWE, S.W., RULEY, H.E., JACKS, T. and HOUSMAN, D.E., p53-Dependent apoptosis modulates the cytotoxicity of anticancer agents. Cell, 74, 957–967 (1993). MARTIN, S.J. and GREEN, D.R., Protease activation during apoptosis: death by a thousand cuts? Cell, 82, 349–352 (1995). MASTERS, J.R.W., OSBORNE, E.J., WALKER, M.C. and PARRIS, C.N., Hypersensitivity of human testis tumour cell lines to chemotherapeutic drugs. Int. J. Cancer, 53, 340–346 (1993). MILLER, S.A., DYKES, D.D. and POLESKY, H.F., A simple salting out procedure for extracting DNA from human nucleated cells. Nucl. Acids Res., 16, 1215 (1988). MIYASHITA, T., KRAJEWSKI, S., KRAJEWSKI, M., WANG, H.G., LIN, H.K., LIEBERMANN, D.A., HOFMANN, B. and REED, J.C., Tumour suppressor p53 is a regulator of bcl-2 and bax gene expression in vitro and in vivo. Oncogene, 9, 1799–1805 (1994). PENG, H.Q., HOGG, D., MALKIN, D., BAILEY, D., GALLIE, B.L., BULBUL, M., JEWETT, M., BUCHANAN, J. and GROSS, P.E., Mutations of the p53 gene do not occur in testis cancer. Cancer Res., 53, 3574–3578 (1993). REED, J.C., MIYASHITA, T., TAKAYAMA, S., WANG, H.-G., SATO, T., KRAJEWSKI, S., AIMÉ-SEMPÉ, C., BODRUG, S., KITADA, S. and HANADA, M., Bcl-2 family proteins: regulators of cell death involved in the pathogenesis of cancer and resistance therapy. J. cell. Biochem., 60, 23–32 (1996). SARK, M.W.J., TIMMER-BOSSCHA, H., MEIJER, C., UGES, D.R.A., SLUITER, W.J., PETERS, W.H.M., MULDER, N.H. and DE VRIES, E.G.E., Cellular basis for differential sensitivity to cisplatin in human germ cell tumour and colon carcinoma cell lines. Brit. J. Cancer, 71, 684–690 (1995). WANG, N., PERKINS, K.L., BRONSON, D.L. and FRALEY, E.E., Cytogenetic evidence for premeiotic transformation of human testicular cancers. Cancer Res., 41, 2135–2140 (1981). ZAKUT-HOURI, R., BIENZ-TADMOR, B., GIVOL, D. and OHREN, M., Human p53 cellular tumour antigen: cDNA sequence and expression in COS cells. EMBO J., 4, 1251–1255 (1985). ZHAN, Q., FAN, S., BAE, I., GUILLOUF, C., LIEBERMANN, D.A., O’CONNOR, P.M. and FORNACE, A.J., Induction of bax by genotoxic stress in human cells correlates with normal p53 status and apoptosis. Oncogene, 9, 3743–3751 (1994).