Apoptosis and autophagic cell death are programmed cell deaths that are involved in cell survival, growth, development and carcinogenesis. p53, the most extensively studied tumor suppressor, regulates apoptosis and autophagy by transactivating its downstream genes. It also stimulates the mitochondrial apoptotic pathway and inhibits autophagy in a transactivation-independent manner. However, the contribution of apoptosis and autophagic cell death to p53-dependent cell death is unclear. Using wild-type (WT) and mutant (MT) p53 inducible cell lines in TP53-null SF126 glioblastoma cells, we examined the apoptosis and autophagic cell death induced by p53. WT p53 expression in SF126 cells induced apoptosis and autophagy, and reduced the cell number. An autophagy inhibitor reduced autophagy, increased the S-phase fraction, and attenuated the inhibition of cell proliferation induced by WT p53. Pan-caspase inhibitor reduced apoptosis but showed weaker inhibition of cell proliferation than the autophagy inhibitor. We concluded that p53-dependent cell death in SF126 cells comprises caspase-dependent and caspase-independent apoptosis and autophagic cell death, and the induction of autophagy as well as apoptosis could be a new strategy to treat some type of WT p53-retaining tumors. (Cancer Sci 2011; 102: 799–807)
Both apoptosis and autophagic cell death are included in programmed cell death. They are involved in not only cell homeostasis processes such as cell survival, growth and development, but also cancer progression and proliferation. Classical chemotherapeutic drugs target DNA to eliminate cancer cells by inducing DNA damage and subsequent apoptosis. Current molecular target drugs also activate the apoptotic pathway, but recent study shows some of these classical drugs and molecular target drugs are also involved in the autophagic pathway.(1,2)
Apoptosis is a well-known programmed cell death. Morphological features of apoptosis include membrane changes, chromatin condensation, nuclear fragmentation and apoptotic body formation.(3) On the other hand, autophagy is a bulk degradation mechanism that plays an essential role in the removal and recycling of cytoplasmic proteins and organelles as well as the maintenance of cellular nutrient homeostasis during nutrient deprivation. Morphological features of autophagy are the formation of a closed double membrane vacuole called an autophagosome, which matures by fusion with a lysosome to create an autolysosome. Although autophagy is necessary for cell survival, recent studies have demonstrated that it induces cell death under specific circumstances such as cancer.(4–6) Cell death with the morphological feature of autophagy is defined as autophagic cell death. In cell death due to autophagy, the fate of the cell should be altered and result in long-term cellular survival when an autophagy inhibitor (such as 3-methyladenine) or siRNA of an autophagy-related gene (such as Atg5) inhibits autophagy.(7) Autophagy removes harmful proteins and organelles that induce DNA damage in a cell; therefore, it works against carcinogenesis. Indeed, beclin-1, a phylogenetically conserved autophagy-related gene, is often inactivated monoallelically in human cancers.(8) Furthermore, beclin-1+/− mutant mice suffer from a high incidence of spontaneous tumors.(9) These indicate that beclin-1 is a haploinsufficient tumor suppressor gene.
Tumor suppressor p53 regulates both apoptosis and autophagy.(10) p53 is activated by a variety of cellular genotoxic stressors(11) and subsequently, p53 transactivates apoptosis- and autophagy-inducing genes (apoptosis: BAX,(12)p53AIP1(13) and PUMA;(14) and autophagy: DRAM,(15)Sestrin 1 and Sestrin 2 [SESN 2](16,17)). Moreover, p53 also regulates apoptosis and autophagy in a transactivation-independent manner. Cytoplasmic p53 induces mitochondrial outer membrane permeabilization, followed by the release of cytochrome c and subsequent apoptosis,(18,19) and attenuated the induction of autophagy.(20,21)
Although p53 regulates both apoptosis and autophagy by diverse mechanisms, it remains unclear how much they contribute to actual p53-dependent cell death and inhibition of cell proliferation. Furthermore, whether the apoptosis and autophagy induced by p53 are mutually independent is unclear. To answer these questions, we constructed a wild-type (WT) p53-inducible cell line in SF126 glioblastoma cells and analyzed the contribution of apoptosis and autophagy to WT p53-dependent cell death using apoptosis and autophagy inhibitors. We also established S121F and R306G mutant (MT) p53-inducible cell lines and analyzed the contribution of apoptosis and autophagy the same as WT p53. These two mutants have not been reported in human tumor. But the cell biological behaviors of these mutants are interesting. S121F has a stronger apoptosis-inducing ability than WT p53.(22) R306G localizes in the cytoplasm due to lack of association to importin alpha, and the truncated form of importin alpha is identified in breast cancer,(23) and might have a strong autophagy-inhibiting ability because of its cytoplasmic localization.(20,21) We considered that comparing the phenotypes of MT p53 to WT p53 might be useful for assessing if apoptosis and autophagy are functionally independent from each other.
Materials and Methods
Cell lines and cell culture. The stable SF126 cell line expressing doxycycline (Dox)-inducible WT p53 (SF126-tet-p53: WT clone) was previously constructed.(24) SF126-tet-TON (Mock clone), SF126-tet-S121F (S121F clone) and SF126-tet-R306G (R306G clone) were constructed according to the protocol described in a previous report(24) using the three plasmids, pcDNA5/TON, pcDNA5/TON-p53S121F and pcDNA5/TON-p53R306G, respectively.(25) Wild-type or MT p53 was induced with 10 ng/mL Dox (Sigma-Aldrich, St. Louis, MO, USA). For inhibition of autophagy and apoptosis, cells were cultured with 6 mM 3-methyladenine (3MA; Sigma-Aldrich) and 10 μM benzyloxycarbonyl-valyl-alanyl-aspartic acid (O-methyl)-fluoro-methylketone (VAD; R&D systems, Minneapolis, MN, USA), respectively.
Subcellular fractionation. Before trypsinization, cells were washed once in cold phosphate-buffered saline (PBS), lysed and separated into subcellular fractions using the CelLytic NuCLEAR Extraction kit (Sigma-Aldrich) following the manufacturer’s recommendations.
Western blotting analysis. Cells were harvested and resuspended in lysis buffer containing 50 mM Tris–HCl (pH 8.0), 150 mM NaCl, 5 mM EDTA and 1% protease inhibitor cocktail (Sigma-Aldrich). The lysate was analyzed by western blotting as described previously(26) using anti-p53 (sc-6243; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-actin (A2066; Sigma-Aldrich), anti-LC3 (PM036; Medical & Biological Laboratories, Nagoya, Japan), anti-BAX (M010-3; Medical & Biological Laboratories), anti-DRAM (ab72171; Abcam, Cambridge, MA, USA), anti-14-3-3σ (ab14123; Abcam), anti-SESN2 (ab57810; Abcam), anti-MDM2 (sc-965; Santa Cruz Biotechnology) and anti-p21 (sc-397; Santa Cruz Biotechnology) antibodies.
Immunofluorescence analysis. The strategy for p53 immunofluorescence was described in a previous report.(24) For LC3 immunofluorescence, cells were harvested at 18 h after treatment and fixed with 4% paraformaldehyde and incubated for 10 min at room temperature. After washing with PBS, they were then permeabilized with 50 μg/mL digitonin (Sigma-Aldrich) for 15 min at room temperature. After washing with PBS, these cells were incubated with anti-LC3 (PM036; Medical & Biological Laboratories) diluted 1:250 for 1 h. After washing with PBS, they were incubated with Anti-Rabbit IgG(H + L)FITC (Beckman Coulter, Brea, CA, USA) for 1 h and visualized using LSM5 PASCAL (Carl Zeiss, Oberkochen, Germany). We observed cells with ×400 magnification and presented the mean percentage LC3 puncta-positive or puncta >10 cells ±SEM in five different fields of vision. In each field, approximately 30 cells were analyzed.
Cell proliferation assay. A total of 5 × 103 cells per well were seeded and incubated in a 96-well plate for 24 h. They were then treated with drugs and further cultured until 48 h at 37°C. Cell proliferation assay were performed with a Cell Counting Kit-8 (Dojin Laboratories, Kumamoto, Japan) as described previously.(24)
Cell cycle analysis by FACS. A total of 1.5 × 104 cells per plate were seeded and incubated in a 6-cm culture plate for 24 h. They were further incubated in the presence of drugs for 48 h. These cells were collected and FACS was performed as described previously.(26)
Characterization of established cell lines. As shown in Figure 1, the established cell lines expressed p53 in a Dox-dependent manner, but with different expression levels, and the expression level of WT p53 was lower than those of S121F and R306G. To determine the transcriptional activities of the expressed p53, we also performed western blotting of well-known p53 downstream gene products. Although the WT p53 expression level was lower than that of MT p53s, all the examined downstream gene products were highly induced by WT p53. DRAM and SESN2 are p53 downstream genes that induce autophagy and they also upregulated in our system.
Immunofluorescence analysis revealed that WT p53 and S121F mutants were mainly localized in the nucleus, and R306G mutants were mainly localized in the cytosol (Fig. 2a). To quantify the distribution of p53 protein, we performed subcellular fractionation to divide the cells into nuclear and cytosolic fractions for immunoblotting. As shown in Figure 2(b,c), more than 80% of WT p53 and S121F mutants were localized in the nucleus, while approximately 80% of R306G was localized in the cytosol.
We then performed a cell proliferation assay (Fig. 3a). At 48 h after Dox treatments, cell proliferation was inhibited by WT p53 expression and the total cell number was equal to that at the start. Cell proliferation was also inhibited by S121F expression, but the total cell number decreased markedly compared with that seen with WT p53 expression. R306G expression inhibited cell proliferation, but its effect was weak in comparison to that exerted by WT p53 and S121F.
Finally, we examined the effect of WT and MT p53 expression on the cell cycle by FACS (Fig. 3b). The sub-G1 fraction was prominently increasing in S121F, followed by WT p53 and R306G. On the other hand, the proportion of the S-phase fraction was decreased by WT p53 and S121F expression, but not by R306G expression. These results showed that WT p53 had cell cycle arrest-inducing as well as apoptosis-inducing activity, while S121F tended to induce apoptosis rather than cell cycle arrest.
Analysis of autophagy inducible abilities of WT and MT p53. Autophagy is induced under nutrient deprivation, and stimulated autophagy was observed by increasing LC3 puncta-positive cells and LC3-I to LC3-II conversion.(27)
Thus, we examined whether autophagy was stimulated in established cell lines under serum-starved conditions. As shown in Figure S1, LC3 puncta-positive cells were increased under serum-starved conditions. LC3-I to LC3-II conversion was also accelerated in all cell lines (Fig. S2); thus, autophagy was detected in SF126 cells under serum-starved conditions.
We then examined whether WT and MT p53 expression stimulates autophagy even under serum-rich conditions. LC3 puncta-positive and puncta >10 cells accumulated in all Dox-treated WT clones, and even MT clones, but not in mock clones (Fig. 4). In the same way, LC3-I to LC3-II conversion was also accelerated when WT and MT p53 were expressed (Fig. 5 lanes 1 and 5).
These data showed that WT and MT p53 expression stimulates autophagy in established cells.
Contribution of autophagy to apoptosis, cell cycle arrest and the inhibition of cell proliferation induced by WT and MT p53. To examine the p53-induced autophagy influence on apoptosis, cell cycle arrest and inhibition of cell proliferation we analyzed them in the established cell lines under the p53 expression conditions using the autophagy inhibitor 3MA or pan-caspase inhibitor VAD.
Adding 3MA, LC3-I to II conversions were decreased in the WT and MT p53 clones (Fig. 5 lanes 5 and 6), but it was rather slightly increased in the mock clones. The number of LC3 puncta-positive cells in WT and MT p53 expressed cells also decreased by adding 3MA (Fig. S3). In contrast to 3MA, VAD did not change LC3-I to II conversions and the number of LC3 puncta-positive cells (Fig. 5 lanes 5 and 7, Fig. S3). These data showed that 3MA block the WT and MT p53-induced autophagy. We then analyzed if the characters of these clones induced by WT and MT p53 change under the blocking autophagy by 3MA.
In mock clones with Dox, adding 3MA or VAD did not increase the sub-G1 fraction, but the S-phase fraction was decreased and cell proliferation was blocked by the 3MA treatments (Fig. 6a,b lanes 5–7). Similar results were obtained from other clones without p53 expression (Fig. 6c–h lanes 1–3). These data showed that under the serum-rich and without p53 condition, 3MA did not influence the autophagy status, but induced cell cycle arrest and resulted in inhibiting cell proliferation, without apoptosis.
In contrast to the mock clone, adding 3MA slightly decreased the sub-G1 fraction and increased the S-fraction on the WT p53 clone (Fig. 6c lanes 5 and 6), thus attenuating the inhibition of cell proliferation (Fig. 6d lanes 5 and 6). Thus, blocking autophagy by 3MA under WT p53 expression conditions results in antagonizing p53 functions such as increasing sub-G1 fraction, cell cycle arrest (S phase) and inhibition of cell proliferation. VAD treatment slightly decreased the sub-G1 fraction, and the proportion of the S-phase fraction did not change (Fig. 6c lanes 5 and 7); however, the inhibition of cell proliferation was attenuated (Fig. 6d lanes 5 and 7). To reveal that increased sub-G1 fraction induced by WT p53 represents the increasing apoptosis fraction, we measured the activity of caspase-3/7 activities (Fig. S4). VAD treatment blocks the pro-caspase activities completely, but 3MA treatments did not influence caspase activities in WT, but rather increased it. On the other side, 3MA treatment increased the total cell number (Fig. 6d lanes 5 and 6). We also found that DRAM knockdown, one of the key molecules of p53-induced autophagy, also increased the total cell number (data not shown). These data indicate that p53-induced autophagy is one of the reasons for the amount of cell death.
On the other hand, apoptosis induced by S121F expression was not rescued by 3MA and the S-phase fraction was slightly decreased (Fig. 6e lanes 5 and 6). The inhibition of cell proliferation was slightly attenuated, but the total cell number was decreased compared with that seen before S121F expression by the 3MA treatment (Fig. 6f lanes 5 and 6). VAD treatment also had a minimal effect on apotosis and inhibition of cell proliferation induced S121F (Fig. 6e,f lanes 5 and 7). VAD treatment strongly blocked the caspase-3/7 activities in S121F cell lines the same as WT, while the subG1 fraction did not decrease (Fig. 6e). We consider that WT p53 and especially S121F induce caspase-independent apoptosis(28) and the subG1 fraction was composed of caspase-independent apoptosis and caspase-dependent apoptosis.
Finally, we examined whether 3MA and VAD treatments influence the R306G phenotype. Treatment with 3MA did not change the sub-G1 fraction or the proportion of the fraction in each cell cycle phase (Fig. 6g lanes 5 and 6). Furthermore, inhibition of cell proliferation was slightly attenuated (Fig. 6h lanes 5 and 6). VAD also did not change the sub-G1 fraction, the proportion of the fraction in each cell cycle phase and the inhibition of cell proliferation (Fig. 6g,h lanes 5 and 7).
These data indicated that 3MA also inhibited R306G-induced autophagy as well as other WT p53 and S121F, but blocking autophagy did not influence its phenotype.
Table 1. The summary of character of established cell lines treating with pan-caspase and autophagy inhibitors
Dox (p53) + 3MA
Dox (p53) + VAD
†Values shown are mean ± SD (95% CI).
Cell proliferation (%)
215 ± 11.4†
189 ± 38.4†
132 ± 5.8†
209 ± 6.9†
1.1 ± 0.62†
0.5 ± 0.00†
1.4 ± 0.11†
0.5 ± 0.17†
S phase (%)
15.5 ± 1.19†
20.6 ± 0.20†
9.3 ± 0.66†
16.3 ± 1.09†
Puncta-positive cells (%) (>10)
27.9 ± 3.44† (0.0)
29.0 ± 5.57† (0.0)
29.1 ± 14.53† (0.0)
32.4 ± 9.50† (0.0)
Cell proliferation (%)
137 ± 22.8†
3 ± 17.7†
100 ± 4.7†
48 ± 11.1†
0.7 ± 0.00†
12.8 ± 1.54†
9.5 ± 0.82†
9.7 ± 0.53†
S phase (%)
19.4 ± 1.41†
9.4 ± 1.31†
17.9 ± 0.91†
7.6 ± 0.24†
Puncta-positive cells (%) (>10)
34.8 ± 7.07† (0.0)
90.0 ± 3.37† (25.2 ± 1.90†)
75.2 ± 3.49† (7.6 ± 2.69†)
93.4 ± 4.06† (31.5 ± 6.55†)
Cell proliferation (%)
161 ± 19.6†
−79 ± 1.4†
−44 ± 9.1†
−77 ± 0.9†
0.8 ± 0.13†
65.6 ± 2.94†
68.1 ± 1.05†
59.2 ± 2.66†
S phase (%)
25.3 ± 3.99†
17.1 ± 1.79†
11.6 ± 0.58†
14.1 ± 0.92†
Puncta-positive cells (%) (>10)
16.8 ± 5.67† (0.0)
91.7 ± 2.33† (37.7 ± 8.68†)
82.9 ± 4.26† (19.1 ± 2.29†)
96.8 ± 2.25† (52.3 ± 9.17†)
Cell proliferation (%)
212 ± 32.1†
94 ± 7.9†
109 ± 4.3†
122 ± 13.4†
0.7 ± 0.11†
1.9 ± 0.41†
2.5 ± 0.53†
1.3 ± 0.17†
S phase (%)
22.9 ± 0.62†
23.8 ± 0.59†
22.5 ± 1.14†
22.9 ± 0.46†
Puncta-positive cells (%) (>10)
27.4 ± 4.76† (0.0)
77.1 ± 4.42† (18.1 ± 2.29†)
32.3 ± 8.06† (2.2 ± 3.53†)
82.3 ± 6.86† (18.0 ± 3.92†)
Dual contribution of apoptosis and autophagic cell death to p53-dependent cell death. Previous studies have reported that overexpression of WT p53 induces both apoptosis and autophagy.(15,29–31) However, it is still unclear whether the autophagy induced by p53 contributed to p53-dependent cell death. In this study, we demonstrated that conditional overexpression of WT p53 in the human glioblastoma cell line SF126 induced both apoptosis and cell death by autophagy, and consequently inhibited cell proliferation. Inhibition of p53-induced autophagy by 3MA released cell cycle arrest and inhibition of cell proliferation but slightly inhibited apoptosis. Inhibition of p53-induced apoptosis by VAD released inhibition of cell proliferation without affecting autophagy. The degree of recovery of cell proliferation by 3MA was higher than that of VAD, indicating that contribution of autophagic cell death was important to p53-dependent cell death as well as apoptosis. We also examined whether the autophagy was induced by p53 expression using TP53-null SaOS2 cells, but under the p53 expressed conditions in SaOS2 cells, the autophagy was not induced (unpublished data, Y.S.). The reason for this remains unclear, but the pathway of p53-induced autophagy may be uncontrollable in SaOS2 cells.
These results suggest that activation of p53-induced autophagy, as well as activation of apoptosis, is a potential therapeutic target for cancer treatment in some types of tumors with WT p53. But it remained to dissolve that precise molecular mechanism of p53 inducing autophagic cell death.
Separation of p53-induced autophagy and apoptosis pathways. Compared with WT p53, S121F demonstrated a stronger ability to induce cell death but retained a similar ability to induce autophagy. When S121F-induced autophagy was inhibited by 3MA, it did not affect caspase-dependent apoptosis and did not release inhibition of cell proliferation. These results suggest that a p53-induced autophagy pathway is separable from a p53-induced caspase-dependent apoptosis pathway.
One possible explanation is that the induction of apoptosis and autophagy is independently regulated by distinct p53-downstream genes, and the genes are differentially transactivated by WT p53 and by S121F. For example, S121F might preferentially transactivate pro-apoptotic genes. However, we have previously shown a lack of correlation between p53-dependent transcriptional activity of downstream genes for apoptosis and cell cycle arrest and the ability to induce apoptosis among 179 mutant p53 including S121F.(26) In that study, there was no correlation between the ability to induce apoptosis and p53-downstream pro-apoptotic genes among WT p53 and S121F. Because the induction of autophagy was similar among WT p53, S121F and R306G, induction of p53-dependent autophagy may not be correlated with p53-dependent transactivation of known downstream genes for apoptosis and cell cycle arrest.
Whether p53-induced autophagy depends on the transactivation function of p53 needs to be elucidated by further studies including identification of new p53-downstream genes related to autophagy and analysis of the transactivation-independent mechanism of p53-induced autophagy.
Cytoplasmic p53 and induction of apoptosis and autophagy. Among common p53 polymorphisms R72 and P72, the R72 variant has a stronger ability to induce apoptosis and a greater ability to localize to the cytoplasm (especially to mitochondria) of cells than the P72 variant.(19) Because R306G mainly locates cytoplasm, we speculate that it may possess a stronger ability to induce apoptosis than WT p53. Unexpectedly, induction of apoptosis by R306G was only slight and was much weaker than those of WT p53 and S121F. It has been shown that apoptosis by cytoplasmic p53 needs collaboration of PUMA and BAX, p53-downstream proteins, in mitochondria.(10) We speculate that R306G may not induce efficient apoptosis because it does not efficiently transactivate downstream genes.
Recently, cytoplasmic p53 has been shown to inhibit autophagy.(20) However, in this study, cytoplasmic mutant R306G failed to inhibit autophagy, but rather slightly induced it. There are at least two speculations explaining the distinct observations. First, cytoplasmic R306G partially inhibits autophagy induced by nuclear R306G, and the observation is the sum total. Second, as in the case of apoptosis, autophagy by cytoplasmic p53 needs collaboration of some p53-downstream proteins, and R306G may not induce the downstream genes.
In conclusion, we demonstrated that in the human glioblastoma cell line SF126, inhibition of cell proliferation by WT p53 consisted of apoptosis and autophagic cell death, and that the contribution of autophagic cell death to p53-dependent cell death was stronger than that of apoptosis. We also demonstrated that the p53-induced autophagy pathway was independent from the p53-induced apoptosis pathway.
The authors thank Eri Yokota for her technical assistance. This study was supported by grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology (12217010 and 17015002), and the Gonryo Medical Foundation to C.I. and Sapporo Bioscience Foundation to S.K.