Hypoxia induces p53 accumulation in the S-phase and accumulation of hypophosphorylated retinoblastoma protein in all cell cycle phases of human melanoma cells.

Hypoxia has been shown to induce accumulation of p53 and of hypophosphorylated retinoblastoma protein (pRb) in tumour cells. In this study, the cell cycle dependence of p53 accumulation and pRb hypophosphorylation in four human melanoma cell lines that are wild type for p53 was investigated using two-parameter flow cytometry measurements of p53 or pRb protein content and DNA content. The hypoxia-induced increase in p53 protein was higher in S-phase than in G1 and G2 phases in all cell lines. The accumulation of p53 in S-phase during hypoxia was not related to hypoxia-induced apoptosis or substantial cell cycle specific cell inactivation during the first 24 h of reoxygenation. pRb was hypophosphorylated in all cell cycle phases by hypoxia treatment. The results did not support a direct link between p53 and pRb during hypoxia because p53 was induced in a cell cycle-specific manner, whereas no cell cycle-dependent differences in pRb hypophosphorylation were detected. Only a fraction of the cell populations (0.60+/-0.10) showed hypophosphorylated pRb. Thus, pRb is probably not the only mediator of the hypoxia-induced cell cycle block seen in all cells and all cell cycle phases. Moreover, the cell cycle-dependent induction of p53 by hypoxia suggests that the primary function of p53 accumulation during hypoxia is other than to arrest the cells.

genomic integrity (Lane. 1992). If the repair fails. the accumulation of p53 induced by DNA damage may trigger apoptosis. Hypoxia has been shown to induce accumulation of p53 protein in the nucleus of most cell lines that are wild type for p53. but not in cell lines that possess mutant p53 (Graeber et al. 1994: Amellem et al. 1997. Cells are arrested in the cell cycle by hypoxia treatment (Graeber et al. 1994: Amellem and Pettersen. 1991: Ludlow et al. 1993). However. this arrest seems to be independent of the p53 status (Graeber et al. 1994). and might. therefore. not be caused by the p53 accumulation. Hypoxia treatment can induce apoptosis in some cell lines (Muschel et al. 1995: Yao et al. 1995: Graeber et al. 1996: Shimizu et al. 1996: Amellem et al. 1997. The role of p53 in hypoxia-induced apoptosis is not completely clarified (Amellem et al. 1997).
The retinoblastoma protein. pRb. is a nuclear phosphoprotein which regulates the cell cycle. pRb is expressed in all cell types and exists in an active hypophosphorylated and an inactive hyperphosphorylated state (Weinberg. 1995). Under normal conditions. pRb is hypophosphorylated and bound to the nucleus only in early G, phase (Mittnacht and Weinberg. 1991: Templeton. 1992: Stokke et al. 1993. It is proposed that pRb serves as a brake on the progression of cells from GI to S-phase of the cell cycle when the protein is in its active state. Hypoxia treatment has been shown to induce reversible hypophosphorylation of pRb in all phases of the cell cycle (Ludlow et al. 1993: Amellem et al. 1996. This hypoxia-induced hypophosphorylation is probably independent of p53 status. Moreover. pRb hypophosphorylation does not seem to cause the hypoxia-induced arrest in GI (Amellem et al. 1996).
Studies of the cell cycle dependence of hypoxia-induced p53 accumulation have not been reported. nor have measurements of hypoxia-induced changes in p53 and pRb in the same cell line. 1547 However, knowledge from such studies might elucidate the roles of pRb and p53 in the regulation of cell proliferation and cell viability under hypoxic conditions. In the present work, hypoxiainduced cell cycle-dependent changes in pRb and p53 were investigated in the human melanoma cell lines A-07, D-12, R-18 and U-25 by two-parameter flow cytometry measurements. The cell cycle variability of these gene products was, thereafter, related to proliferation and viability data to reveal information about the functions of p53 and pRb during hypoxia.

MATERIALS AND METHODS Cell lines
Four established human melanoma cell lines (A-07. D-12, R-18. U-25) were used (Rofstad, 1994). Mutations in p53 were analysed by CDGE (constant denaturant gel electrophoresis) with primers covering exon 5-8 (B0rresen, 1996), as well as by complete sequencing of the p53 cDNA using primers covering the open reading frame (Smith-S0rensen et al, 1998). The analysis showed that all four cell lines express wild-type p53. Cell lines were maintained in monolayer culture in RPMI-1640 medium (25 mm HEPES and L-glutamine) supplemented with 13% fetal calf serum, 250 mg 1-1 penicillin and 50 mg 1-1 streptomycin. Cultures were incubated at 37°C in a humidified atmosphere of 5% carbon dioxide in air and subcultured by trypsinization (0.05% trypsin/0.02% EDTA solution).

Exposure to hypoxia
Cells from cultures in exponential growth were plated in glass dishes, incubated at 37°C in a humidified atmosphere of 5% carbon dioxide in air for 24 h, and then exposed to hypoxia for 16 h. The culture medium was removed and replaced by fresh medium supplemented with 2.2 g 1-' sodium bicarbonate immediately before hypoxia treatment. The glass dishes were placed in air-tight steel chambers during the hypoxia tratment. The steel chambers were flushed with a humidified, highly purified gas mixture consisting of 95% nitrogen and 5% carbon dioxide at a flow rate of 51 min-'. Measurements showed that the concentration of oxygen in the medium was < 10 p.p.m. after 1 h of flushing. The pH in the medium at the end of the exposure was within the range of 7.3-7.5. Cell proliferaton and survival during hypoxia Cell proliferation during hypoxia was determined by counting aerobic control cells or hypoxia-treated cells in a haemocytometer and measuring DNA histograms by flow cytometry as described previously (Sanna and Rofstad, 1994). Cell survival was measured by using a colony assay (Rofstad, 1992). Briefly, 1-ml aliquots of cell suspension were seeded in 25-cm2 tissue culture flasks, containing 10lethally irradiated feeder cells in 4 ml of medium. After 8-21 days of incubation at 37°C in a humidified atmosphere of 5% carbon dioxide in air, cells were fixed in ethanol, stained with methylene blue, and colonies containing more than 50 cells were counted.
Ap tosis measuremes Apoptotic cells were detected by immunofluorescence using the ApopTag (Oncor) in situ assay (Yao et al, 1995). Cells washed in Ca'+-and Mg'+-free Hank's buffered salt solution (HBSS) were resuspended in 4% neutral-buffered paraformaldehyde for 10 min at room temperature. Aliquots of approximately 50 gl of cell suspension were dropped on microscope slides. Two drops of equilibration buffer were added before the slides were incubated at room temperature in a humidified atmosphere for 5 mnn. Approximately 50 jtl of a terminal deoxynucleotide transferase and digoxigenin-ll-deoxyuridine triphosphate solution was applied to the cell preparations to end-label DNA frgments. The slides were continuously incubated at 37°C for 1 h and washed in prewarmed stop/wash buffer at 37°C for 30 min. Approximately 50 .tL of an antidigoxigenin-fluorescein solution was pipetted onto the slides for detection of the end-labelled DNA fragments before the slides were incubated at room temperature for 30 min. Finally, the slides were mounted under a glass coverslip with a drop of propidium iodide/Antifade staining buffer. The cell preparations were washed three times in phosphate-buffered saline (PBS) between each step in the procedure. The cells were visualized by epifluorescence using standard fluorescein excitation and emission filters. The fraction of cells in apoptosis was determined by scoring 500 cells in each sample.

Energy charge measurements
The relative concentraions of adenylate phosphates in control cells and cells exposed to hypoxia were measured by high performance liquid chromatography at 254 nm. The medium of the cell cultures was replaced with acetonitrile and the lysed cells were scraped off the glass dishes. The cell lysates were dried and low-strength buffer (100 mm potassium dihydrogen phosphate. 1.5% acetonitrile. 0.08% C16HH6NBr. pH 5.0) was added before the samples were centrifuged. The supernatant was used in the further analysis. The elution conditions that provided the best resolution were 80% lowstrength buffer and 20% high-strength buffer (150 mm potassium dihydrogen phosphate, 10% acetonitrile. 0.08% tetrabutylammuonium bromide, pH 5.0) for 10 min. followed by a linear gradient to 100% high-strength buffer during the next 10 min. Elution with the high-strength buffer was continued for anothr 10 min. The separation was carried out with a Supelcosil LC-18-T 5-pm cartridge (Supelco, USA). The flow rate was 1.0 ml min-'. The adenylate energy charge was calculated by the equation:  Science) diluted 1:50 in staining buffer. After incubation for 30 minu the nuclei were washed once with staining buffer and resuspended in 500 g1 of staining buffer containing 2 jg ml-' Hoechst 33258. The samples were filtered through a 30-jm nylon mesh and analysed in a flow cytometer.

Staining for BrdUrd and DNA content
To investigate the progression of cells in the different phases of the cell cycle after hypoxia treatment, cells were incubated with 50 jg ml' BrdUrd (5-bromo-2'-deoxyuridine) for 30 min immediately before hypoxia treatment. The culture medium containing BrdUrd was removed, the cells were washed severat times with PBS and fresh medium supplemented with sodium bicarbonate was added before hypoxia treatment. The cells were incubated at 37°C in a humidified atmosphere of 5% carbon dioxide in air for various times after reoxygenation. Cells were harvested and fixed in 100% methanol and stored at -20°C. Immunostaining for BrdUrd was performed according to the protocol of Gerlyng et al (1992). Briefly, the cells were digested in a suspension of 0.2% pepsin in 2 M hydrochloric acid and stained for BrdUrd by a three-layer procedure. The anti-BrdUrd antibody (Becton Dickinson). the secondary antibody (biotinylated anti-mouse Ig from sheep. Amersham Life Science) and streptavidin-FITC (Amersham Life Science) were diluted in PBS containing 0.5% Tween 20 and 0.5% bovine serum albumin. PBS was employed for washing between incubations. After incubation with streptavidin-FITC, the pellets were washed once with PBS and resuspended in 500 jl of PBS with 2.5 jg ml-' propidium iodide (Calbiochem) and 100 jg ml-RNAase A (Pharmacia). The samples were filtered through a 30-jm nylon mesh and analysed in a flow cytometer.

Flow cytomew
The stained cells or nuclei were analysed in a FACStarP'Ls flow cytometer (Becton Dickinson) equipped with two argon lasers (SpectraPhysics) that were tuned to 488 nm and UV respectively. FITC fluorescence, forward light scatter and side scatter pulse amplitudes, as well as propidium iodide fluorescence pulse height. pulse width and pulse area were measured upon excitation by the 488-nm laser. Hoechst 33258 fluorescence pulse height pulse width and pulse area were measured upon excitation by the UV laser. Hoechst 33258 fluorescence pulse area or propidium iodide fluorescence pulse area were used as a measure of DNA content. The nuclei were gated for doublet discrimination in a diagram of integrated Hoechst 33258 fluorescence against Hoechst 33258 pulse width. The green fluorescence intensities were calibrated with fluorescent beads in each experiment such that the FITC fluorescence intensities measured in different expeiments could be compared.

Data t nt and analysis
The median FITC fluorescence intensity (FLI-H) of the anti-p53stained sample subtracted by the median fluorescence intensity of the corresponding control sample (which had received no primary antibody) was used as a measure of relative level of p53 protein in the nuclei. Measurements of relative levels of p53 protein in each phase of the cell cycle were obtained by generating FLI -H histograms from three narrow gates in the DNA histograms. T-47D British Journal of Cancer (1998) 78 (12) fluorescence; F32-A) and p53 protein content (FITC fluorescence; FL1-H). Histograms A-D represent aerobic cells, whereas histograms E-K represent cells exposed to hypoxia for 16 h. The dual parameter histogam A and E represent control cels which received no primary antbody against p53. The dual parameter hi s B and F show the distribuon of stained p53 prtein in the nuclei throughout the cel cycle. The DNA content of these nucei is also shown in one-parameter F32-A histograms (C and G). The gates Rl, R2 and R3 in these histograns were applied to generate FL1-H histograms for cells in G1 phase, S-phase and G2 phase of the cell cyCe respeely. p53 protein Ctent of different cell cycle phases is shown for hypoxia-treated cells in histograms I-K Histograms for total p53 protein lvel are shown in D and H cells, which are mutant for p53, were used as a positive control.
They showed a relative level of p53 protein that was about 20 times the value of the untreated melanoma cells, which are wild type for p53.
T'he extraction procedure provides samples where only the detergent-resistant bound pRb is present, i.e. hypophosphorylated pRb (Mittnacht and Weinberg, 1991;Templeton, 1992;Stokke et al, 1993). At normal conditions, pRb is only bound in the nucleus in G . Two distinct peaks with G, DNA content were observed in the FL1-H histograms (FITC fluorescence) of aerobic nuclei. Previous studies have shown that the PMG3-245 anti-pRb monoclonal antibody has a high degree of non-specific binding (Stokke et al, 1993;Jonassen et al, 1994). The nuclei with low FL1-H fluorescence intensities were, therefore, assumed to have only nonspecific binding of the pRb antibody, and were termed pRbnuclei. High FL 1-H fluorescence intensities were assumed to and U-25 cells from aerobic control cels (C) and hypoxitreated cels (H).
Proteins from 5 x 105 cells were baded in each lane. The band below p53 was due to non-specific bincing, as shown in blots stained without the p53 antibody represent nuclei with bound hypophosphorylated pRb. which were called pRb+ nuclei. The fraction of pRb+ nuclei was obtained by measuring the relative number of nuclei with high FL 1-H. In addition to omission of the primary antibody. the B-cell lymphoma cell line U698 was used as a negative control.

Westem bloffing
Cells harvested from cultures in exponential growth were boiled in Laemmli's lysis buffer (Laemmli, 1970) for 5 mnm. Proteins were separated by 9% sodium dodecyl sulphate polyacrylamide gel electrophoresis and transferred to a polyvinylidene fluoride (PVDF) membrane. The mouse monoclonal antibodies PAbl8Ol (Oncogene Science) and PMG3-245 (Pharmingen) were used for specific staining of p53 and pRb. Staining was performed using a three-layer procedure, using biotinylated anti-mouse antibody (Amersham Life Science) as second layer, streptavidin-alkalin phosphatase (Amersham Life Science) as third layer and BCIP/NBT (Sigma) as substrate.

Statistical analysis
Statistical comparsons of data were performed by parametric analysis using Student's t-tests and one-way analysis of vanrance. Multivanrate statistical analysis was applied to compare measured quantities in the phases of the cell cycle for each of the four cell lines (Johnsen and Wichern 1992). A significance criterion of P < 0.05 was used.

RESULTS
The melanoma cells are arrested in all cell cycle phases during hypoxia treatment The cell number of the D-12 line remained constant dunrng hypoxia treatment ( Figure 1A). p53 protein is induced specifically in S-phase under hypoxic condifions An example of results from flow cytometric analysis of aerobic control cells and hypoxia-treated D-12 cells which were stained for p53 protein and DNA content is shown in Figure 2. The histograms in Figure 2B and F show the distributions of p53 protein throughout the cell cycle for aerobic control cells and hypoxia-treated cells respectively. Hypoxic conditions induced accumulation of p53 protein in a cell cycle-dependent manner. To investigate this cell cycle-dependent accumulation of p53 protein more precisely, histograms of p53 protein content were generated for each phase of the cell cycle using narrow gates set in the DNA histograms ( Figure 2C and G). Measurements of relatixe lexve1 of p5 3 protein in each phase of the cell cxcle under aerobic and hx-poxic condition, are shovx-n in Ficure 3. The relatix e lex el of p53 protein tended to increase in all cell c\cle phases after h\poxia treatment. Hoxvever. the hxpoxiainduced accumulation of p53 protein in S-phase xxas the dominatine effect. and a _icnificant increase in p5 protein xwas seen in all cell lines A-O-. D-l1. R-18. P < 0.0(X)5: t-25. P = 0.01I. Moreover. the increase in p53 protein in S-phase xvas significantl\ higher than in G pha;se in all cell lines iP < 0.05 and sigoniticantlv higher than in G phas>e in all cell lines i P < GS0i5) except in R-IS iP = 0.054i. The total cell populations Figure 2D  The hypoxia-treated cells were reoxygenated and f time points after reoxygenation. Cells that are reow hypoxia treatment will re-enter the cell cycle be metabolically viable immediately after hypoxia tre and Sutherland. 1989). However. a certain number clonogenically dead. and will at some time point al tion stop cycling and disintegrate. These dead cells to the surface of the glass dishes. and will. therefor of the fixed cell population. Thus. analysis of BrdU content over a broad time period after reoxygenatioi whether the majority of the inactivated cells we durinc hypoxia.
An example of results from flovw cytometric anal control cells and hypoxia-treated A-07 cells stain and DNA content is shown in Figure 8. The BrdUrd-negative (Rl) and BrdUrd-positive (R2) z cells through the cell cycle immediately after BrdUrd incorporation is shown in Figure 8A. The cell cycle distribution of BrdUrdnegative (RI) and BrdUrd-positive (R2) hypoxia-treated cells is shown at O h (B). 4 h (C) and 13 h (D) after hypoxia treatment. As pRbj<ts expected from Figure lB. there were no significant differences between the aerobic control cells ( Figure 8A) and the hypoxia-07, D-12, R-18 treated cells ( Figure 8B) immediately after BrdUrd incorporation ratnd Cells (H). or hypoxia treatment. With increasing reoxygenation times (up to mid above pRb C vithout the pRb 24 h). the BrdUrd-positive hypoxia-treated cells became gradually distributed throughout all phases of the cell cycle. Four hours after reoxygenation. BrdUrd-labelled cells (R2) from hypoxia-treated populations had moved on to GJM and the next G1 ( Figure 8C). iase of the cell although the cell cycle time of the hypoxia-treated cells seemed to -action of pRb+ be increased compared with aerobic control cells (data not shown).

Ily significant
Thirteen hours after reoxygenation ( Figure 8D). cells that were in cell lines. For S-phase during hypoxia were distributed among all cell cycle i did not vary phases. Thus. at least some of the cells which were in S-phase ell populations during the hypoxia treatment were still cycling. Cells in G and os. Comparison G/M phase during hypoxia treatment (BrdUrd-negative cells) ells resulted in were also cycling ( Figure 8D). Moreover. the ratio of BrdUrdthe cell lines negative and BrdUrd-positive hypoxia-treated cells did not varv re pRb+ under substantially during the first 24 h of reoxygenation. The ratios of BrdUrd-negative and BrdUrd-positive ceHls at time points later ccumulation of than 24 h were not analysed. as they were disturbed bv cell divint (Figure 7). sion. Thus. the BrdUrd analysis did not reveal anv substantial ie nucleus after hypoxia-induced cell cycle-specific inactivation during the first 24 h of reoxygenation.
I hypoxia-DISCUSSION ivatlonl Cells expressing wild-type p53 accumulate p53 protein when at the start of exposed to stresses such as ionizing radiation. UV light. heat thal effects of shock. starvation and hypoxia (Ko and Prives. 1996). We found a . 1991). On the significant hypoxia-induced increase in p53 protein level in the c p53 induction S-phase of four melanoma cell lines that are wild type for p53.

stigate whether
Moreover, the increase in p53 level in S-phase was significantly activation. and higher than in GI phase in all cell lines. and significantly higher re sensitive to than in G, phase in three out of four cell lines. Cell cycle-depens were labelled dent accumulation of p53 by hypoxia treatment has not been It to follow the reported previously. In a recent study of three human wild-type hat are synthe-p53 cell lines. y-irradiation was shown to induce p53 selectivelv in nod will incor-G, phase and early S-phase (Komarova et al. 1997). Moreover. UV compared with radiation of wild-type p53 NIH3T3 cells induced p53 selectivelỹ l phase cells). in the S-phase of cells that were synchronized by serum stars ation fixed at several (Haapajarvi et al. 1995). In contrast. p53 induction by heat shock tygenated after and UV has been shown to be independent of cell cycle phase in cause they are human fibroblasts (Yamaizumi and Sugano. 1994: Sugano et al. matment (Kwok 1995). The p53 induction by hypoxia. however. cannot be of the cells are compared with that of DNA-damaging treatments because fterreoxygenahypoxia probably induces little or no DNA damage. Whereas will not attach DNA lesions in cells may trigger p53 accumulation to maintain -e. not be a part genomic integrity (Lane. 1992). neither the trigger for hypoxiard versus DNA induced p53 accumulation nor all cellular effects of p53 during n should reveal hypoxia are known. ere in S-phase In the present study. p53 induction by hypoxia was only studied in melanoma cell lines that are wild type for p53. There are rather lysis of aerobic few reports in which cell lines with mutations in the p53 gene have led for BrdUrd been exposed to hypoxia. Graeber et al ( 1994)  hypoxia. Moreover, no induction of p53 protein was observed when the mutant p53-expressing breast cancer cell line T47-D was treated with hypoxia (Amellem et al. 1997). These observations are in agreement with the lack of increase in p53 levels seen in most cells with mutant p53 when exposed to DNA-damaging agents (Kastan et al, 1991: O'Connor et al. 1993. Hypoxia treatment induced accumulation of hypophosphorylated pRb in all the melanoma cell lines. However. only a fraction of the cells showed hypophosphorylated pRb bound to the nucleus. This cell fraction did not seem to differ between the cell cycle phases. Our observations are consistent with those of Amellem et al (1996). who found that pRb was hypophosphorylated in only a certain fraction of the cell populations when the cells were exposed to less than 4 p.p.m. oxygen. In their work. however. the fraction of the cells with hypophosphorylated pRb differed between cell lines, whereas no significant differences could be detected between the four melanoma lines studied here.
p53 accumulation induced by DNA-damaging agents can mediate G, phase arrest (Ko and Prives. 1996). G, arrest by hypoxia. however. seems to be independent of p53 accumulation (Graeber et al. 1994). The arrest of the melanoma cells in all phases of the cell cycle did not correlate with the cell cycle-dependent induction of p53. The primary function of p53 induction by hypoxia treatment is. therefore. probably not to arrest the cells.
Consequently. we suggest that factors other than pS3 are essential for the hypoxia-induced cell cycle arrest of the melanoma cells.
Several physiological growth inhibitory signals as well as DNAdamaging agents can block pRb phosphorylation and, thereby.
arrest the cells in G, phase (Weinberg, 1995). However. because only a fraction of the melanoma cells showed pRb hypophosphorylation by hypoxia. pRb activation is probably not the only cause of the hypoxia-induced cell cycle arrest as all cells were arrested. Thus. pRb hypophosphorylation does not seem to be essential for the hypoxia-induced cell cycle arrest of the melanoma cells. in accordance with the observations of Amellem et al ( 1996). The arrest of cells in the cycle has been observed in different cell lines when exposed to a sufficiently low oxygen concentration. However. although all cell lines are arrested on the GI/S border by hypoxia. some cell lines progress from mitosis to G, (Spiro et al. 1984: Amellem andPettersen. 1991). whereas others are also arrested in mitosis (Shrieve et al, 1983. the present study). The differences in cell cycle arrest between cell lines are probably not caused by different levels of oxygenation. as all investigators used an oxygen concentration of less than 10 p.p.m. Thus, there seems to be cell line-specific differences in the regulation of the hypoxia-induced cell cycle arrest. As most of the cell lines used in these experiments are tumour lines. such differences could be due to mutations. causing changes in the expression of so far unknown proteins regulating the cell cycle progression during hypoxia.
Several studies have suggested a link between p53 and pRb in the control of cell growth (Demers et al. 1994: Hickman et al. 1994: Slebos et al. 1994: Haupt et al. 1995. When cells are exposed to DNA-damaging agents. the p53 induction leads to an increase in p21. At high protein concentrations. p21 inhibits the functions of CDKs (cyclin-dependent kinases). allowing the accumulation of hypophosphorylated pRb. which might result in radiation-induced GI arrest. Hypoxia. however. has not been shown to activate this G, block pathway. The present data do not support a link between p53 and pRb during hypoxia because p53 was induced in a cell cyclespecific manner. whereas no cell cycle-dependent differences in the fraction of cells with hypophosphorylated pRb were detected. Our observations are consistent with those of Graeber and co-workers (Graeber et al. 1994). who suggested that hypoxia induces p53 protein by a different pathway to DNA damaging agents.
Hypoxic conditions of 10 p.p.m. oxygen for some hours inevitably led to cell inactivation. despite the ability of cells to adapt to an oxygen-poor atmosphere. In the melanoma lines. approximately one-half of the cells lost clonogenicity after 16 h of hypoxia treatment. This cell inactivation is probably not due to energy depletion. as adenylate energy charge values after 16 h of hypoxia treatment were rather high [between 0.68 ± 0.02 (U-25) and 0.84 ± 0.03 (D-12 and R-18)]. Because the adenylate energy charge of most normal cells are in the range 0.80-0.95 (Stryer. 1988). it is unlikely that the hypoxia-induced decrease in energy charge will lead to substantial cell inactivation. Results from studies of rodent cell lines support the present observation. as they suggest that other mechanisms were involved in the hypoxiainduced cell inactivation (Shrieve et al. 1983: Rotin et al. 1986). However, the mechanism by which hypoxia induces cell inactivation has not been fully elucidated. In the present work. we attempted to clarify the roles of p53 and pRb in hypoxia-indtuced cell inactivation.
Induction of apoptosis has been shown to contribute to hypoxiainduced cell death (Muschel et al. 1995: Yao et al. 1995: Graeber et al. 1996: Rofstad et al. 1996b: Shimizu et al. 1996. Furthermore. a relationship between the accumulation of p53 during hypoxia and hypoxia-induced apoptosis has been suggested (Graeber et al. 1996). as a parallel to the p53-dependent apoptotic pathway induced by DNA damage. The hypoxia-induced p53 accumulation in S-phase melanoma cells did not induce apoptosis. Whereas all the melanoma lines showed a significant increase in p53 level after 16 h of hypoxia treatment. none of the melanoma lines showed a significant increase in the apoptotic fraction. The lack of apoptosis induction despite the significant pS3 accumulation in the melanoma lines suggests that induction of p53 by hypoxia does not necessarily lead to apoptosis in tumour cells that are wild type for p53. A possible overexpression of the proteins Bcl-2 and Bcl-xL might explain the lack of hypoxia-induced apoptosis in the melanoma cell lines, as these proteins have been shown to prevent hypoxia-induced apoptosis (Graeber et al. 1996: Shimizu et al. 1996. Some studies have shown. by measuring clonogenic survival. that cells in S-phase at the start of hypoxia treatment are most sensitive to the lethal effects of hypoxia (Spiro et al. 1984: Amellem andPettersen. 1991). However, others have found little difference in the survival of cells to hypoxia during the cell cycle (Kwok and Sutherland. 1989). The BrdUrd analysis of melanoma cells did not reveal any substantial cell cycle phase-specific inactivation during the first 24 h after hypoxia treatment. However. as the applied method is rather approximate. we cannot exclude the possibility that a larger fraction of S-phase cells than of G, and G./M cells are inactivated after hypoxia treatment. Nevertheless. our results suggest that many cells residing in S-phase during hypoxia treatment survive. in contrast to a previous study on human cells (Amellem and Pettersen. 1991).
In conclusion, four human melanoma cell lines exposed to hypoxia showed accumulation of p53 pnrmarily in the S-phase and induction of hypophosphorylated pRb in all phases of the cell cycle. The hypoxia-induced p53 accumulation was not associated with apoptosis or cell cycle-specific cell inactivation. Moreover. (1998) 78(12) Hypoxia-induced changes in p53 arnd pRb 1557 neither p53 accumulation nor pRb hypophosphorylation seemed to be essential for the hypoxia-induced cell cycle block seen in all phases of the cell cycle.