Induction of p53-Dependent Apoptosis by Prostaglandin A2.

Prostaglandin (PG) A2, one of cyclopentenone PGs, is known to induce activation of apoptosis in various cancer cells. Although PGA2 has been reported to cause activation of apoptosis by altering the expression of apoptosis-related genes, the role of p53, one of the most critical pro-apoptotic genes, on PGA2-induced apoptosis has not been clarified yet. To address this issue, we compared the apoptosis in HCT116 p53 null cells (HCT116 p53-/-) to that in HCT116 cells containing the wild type p53 gene. Cell death induced by PGA2 was associated with phosphorylation of histone H2A variant H2AX (H2AX), activation of caspase-3 and cleavage of poly(ADP-ribose) polymerase 1 in HCT116 cells. Induction of apoptosis in PGA2-treated cells was almost completely prevented by pretreatment with a pan-caspase inhibitor, z-VAD-Fmk, or an inhibitor of protein synthesis, cycloheximide. While PGA2 induced apoptosis in HCT116 cells, phosphorylation of p53 and transcriptional induction of p53-target genes such as p21WAF1, PUMA, BAX, NOXA, and DR5 occurred. Besides, pretreatment of pifithrin-α (PFT-α), a chemical inhibitor of p53’s transcriptional activity, interfered with the induction of apoptosis in PGA2-treated HCT116 cells. Pretreatment of NU7441, a small molecule inhibitor of DNA-activated protein kinase (DNA-PK) suppressed PGA2-induced phosphorylation of p53 and apoptosis as well. Moreover, among target genes of p53, knockdown of DR5 expression by RNA interference, suppressed PGA2-induced apoptosis. In the meanwhile, in HCT116 p53-/- cells, PGA2 induced apoptosis in delayed time points and with less potency. Delayed apoptosis by PGA2 in HCT116 p53-/- cells was also associated with phosphorylation of H2AX but was not inhibited by either PFT-α or NU7441. Collectively, these results suggest the following. PGA2 may induce p53-dependent apoptosis in which DNA-PK activates p53, and DR5, a transcriptional target of p53, plays a pivotal role in HCT116 cells. In contrast to apoptosis in HCT116 cells, PGA2 may induce apoptosis in a fashion of less potency, which is independent of p53 and DNA-PK in HCT116 p53-/- cells


Introduction
Tumor suppressor gene, p53, which is one of the most crucial tumor suppressor genes, exerts its anti-cancer effect by activating cell death, including apoptosis and autophagic cell death, as well as cell cycle arrest in cancer cells [1]. Apoptosis induced by p53 is carried out by apoptosis-relating proteins whose expression is modulated by p53. Besides, p53 can induce apoptosis by directly stimulating cytochrome c release from mitochondria through its mitochondrial translocation [2][3][4]. The intracellular (Hyclone, Logan, UT, USA), 100 units/mL penicillin (Hyclone) and glutamate (Invitrogen, Carlsbad, CA, USA) at 37 • C under 5% CO 2 . The subculture of cells or replacement of media was done every three days.

Cell Death Assay
Cell death induced by PGA 2 was assessed by annexin V assay. After PGA 2 treatment, cells were stained with fluorescein isothiocyanate (FITC)-labeled annexin V to measure the translocation of phosphatidylserine from inner leaflets to outer leaflets in the cytoplasmic membranes and propidium iodide to detect necrotic nuclei using Annexin V apoptosis assay kit (BD Biosciences, San Jose, CA, USA). Then, fluorescence of stained cells was measured on FACSCanto II (BD Biosciences) and analyzed using BD FACSDiva program.

Quantitative Real Time Polymerase-Chain Reaction
For the analysis of the mRNA level of target genes of p53 in PGA 2 -treated cells, quantitative real time polymerase-chain reaction (qPCR) was performed as follows. First-strand cDNA was synthesized from total RNA using PrimeScript TM RT reagent Kit (Takara Korea Biomedical Inc., Seoul, Korea). First-strand cDNA was then amplified by specific primers against target genes of p53 using SYBR FAST qPCR Kit (KAPAbiosystems, Woburn, MA, USA) on ABI 7300 Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA). GAPDH mRNA level normalized each mRNA level of p53-target genes in the same sample and their relative changes among samples were calculated by the ∆∆Ct method [22].

Transfection of Small Interfering RNA (siRNA)
For the knockdown of DR5 expression, siRNA against DR5 was transfected into HCT116 cells using Lipofectamine TM RNAiMAX Transfection reagent (Thermo Fisher Scientific, Waltham, MA, USA). The final concentration of DR5 siRNA was 1 nM, and the volume of transfection reagent was 3 µL.

Statistical Analysis
All data in this study are expressed as the means ± standard error of the mean, which were obtained from three independent experiments performed in triplicate. Statistical analysis was performed using a paired Student's t-test. P-values of data were indicated in each figure.

Results
3.1. PGA 2 Induces Caspase-Dependent Apoptosis in HCT116 Cells But Not in HCT116 p53-/-Cells First, we attempted to observe the difference in the level of apoptosis between HCT116 cells and HCT116 p53-/-cells treated with PGA 2 . As shown in Figure 1A,B, and Figure S1A-C, annexin V-positive HCT116 cells were increased according to concentrations and incubation times of PGA 2 , but in HCT116 p53-/-cells treated with PGA 2 , annexin V-positive cells were hardly increased. To be consistent with the result of annexin V assay, PGA 2 induced cleavage of poly(ADP-ribose) polymerase 1 (PARP1) and caspase-3 only in HCT116 cells ( Figure 1C). Moreover, when HCT116 cells were pretreated by z-VAD-Fmk, a pan-caspase inhibitor, PGA 2 -induced apoptosis was almost wholly prevented ( Figure 1D,E). Collectively, these data suggested that PGA 2 induces caspase-dependent apoptosis in a p53-dependent manner in HCT116 cells.

PGA 2 Activates p53 via DNA-PK during the Induction of Apoptosis in HCT116 Cells
Then, we analyzed whether and how p53 was activated in HCT116 cells during the PGA 2 -induced apoptosis. Whereas HCT116 p53-/-cells showed no expression of p53, p53 was phosphorylated at Ser-15 and at Ser-46 by PGA 2 treatment in HCT116 cells, and the extent of p53 phosphorylation was increased in parallel with concentrations of PGA 2 (Figure 2A, Supplementary S2). Notably, the protein level of p53 was also increased in the same pattern with that of p53 phosphorylation, implying that phosphorylation of p53 protein may result in its stabilization. Transcripts of p53's target genes such as PUMA, BAX, p21 WAF1 , and NOXA were augmented by PGA 2 treatment in HCT116 cells but not in HCT116 p53-/-cells, suggesting that PGA 2 should increase transcriptional activity of p53 by phosphorylating it ( Figure 2B, Figure S3A). Messenger RNAs of p21 WAF1 and NOXA showed an increasing tendency in PGA 2 -treated HCT116 p53 -/-cells, implying p53-independent gene expression by PGA 2 . However, pifithrin (PFT)-α, a chemical inhibitor of p53's transcriptional activity, prevented the increase of p21 WAF1 and NOXA mRNAs in HCT116 cells, but not in HCT116 p53 -/-cells ( Figure S3B). It was reported that PGA 2 can increase the level of p21 WAF1 mRNA via HuR-mediated stabilization of p21 WAF1 mRNA in the absence of p53 [23]. In addition, it was shown that transcription factors such as SP1, p73, ATF3 and ATF4 can activate transcription of NOXA without involvement of p53 [24,25]. So, it can be speculated that PGA 2 activates p53-dependent transcription in HCT116 cells, and PGA 2 also increases of p21 WAF1 and NOXA mRNAs in HCT116 p53-/-cells through mRNA stabilization and activation of various transcription factors. The increase of p21 WAF1 , PUMA, and NOXA expression was also observed at the level of protein ( Figure 2C, Figure S4). While multiple protein kinases including HIPK-2, p38MAPK, PKC-δ, and DYRK2 phosphorylate p53 at Ser-46, enzymes involved in DNA damage response such as DNA-PK, ATM, and ATR, phosphorylate p53 at Ser-15 [26,27]. Since phosphorylation of p53 Ser-46 is very subtle and PGA 2 was reported to activate DNA-PK [21], DNA-PK was suspected to be a principal enzyme to phosphorylate p53 Ser-15 and induce apoptosis. And as expected, NU7441, an inhibitor of DNA-activated protein kinase catalytic subunit (DNA-PKcs) prevented PGA 2 -induced phosphorylation of p53 and increase of p21 WAF1 protein ( Figure 2D), demonstrating a causative role of DNA-PK in activation of p53. To be consistent with its effect on PGA 2 -induced transcriptional activity of p53, NU7441 suppressed PGA 2 -induced apoptotic findings such as cleavage of PARP1 and increase of annexin V-positive cells as well ( Figure 2D,E). NU7441 showed no effect on survival of HCT116 p53-/-cells ( Figure S5). Accordingly, these data suggested that PGA 2 may activate p53 via inducing DNA-PKcs activity, which plays a critical role in PGA 2 -induced apoptosis.

PGA 2 -Induced Apoptosis Was Dependent on the Transcriptional Activity of p53
Activation of p53 induces apoptosis in a manner of both dependent on and independent of its transcriptional activity in which these two pathways are not mutually exclusive [1,28]. Then, we attempted to determine whether the effect of p53 was exerted via transcriptional regulation of apoptotic genes or via its effect on the mitochondrial apoptotic pathway in PGA 2 -induced apoptosis. As shown in Figure 3A,B, and Figure S1D, PGA 2 -induced apoptosis was gradually decreased according to concentrations of PFT-α. Consistently with these findings, cleavage of PARP1 and caspase-3 was reduced, but phosphorylation of p53 was marginally affected, indicating that the caspase activation cascade was downstream of p53 activation ( Figure 3C). Considering that pretreatment of PFT-α reduced expression of PGA 2 -induced p21 WAF1 , the inhibitory effect of PFT-α on PGA 2 -induced apoptosis must be due to its inhibitory effect against the transcriptional activity of p53 ( Figure 3C). Besides, p53 did not move to mitochondria and, instead, was accumulated in cytosol ( Figure S6A). Furthermore, phosphorylated p53, which has transcriptional activity, was increased in the nuclear fraction ( Figure S6B). Therefore, these results suggested that PGA 2 -induced apoptosis may be dependent on the transcriptional activity of p53 but not mitochondrial p53. The result of three independent annexin V assay was presented as mean ± SEM. (C) Whole cell lysates (WCLs) were prepared and subjected to immunoblot analysis against indicated proteins using GAPDH as an internal reference protein.

PGA 2 -Induced Apoptosis Is Dependent on de novo Protein Synthesis of p53 Target Genes
Then, we speculated that proteins of p53 target genes might play a critical role in PGA 2 -induced apoptosis. To prove this speculation, we analyzed the effect of cycloheximide (CHX), an inhibitor of translation, on PGA 2 -induced apoptosis. Pretreatment of CHX prevented the induction of apoptosis by PGA 2 (Figure S7A,B), and cleavage of both PARP1 and caspase-3 as well without an effect on phosphorylation of p53 ( Figure S7C), implying the critical role of de novo proteins synthesized by p53 in this apoptosis model. Although PUMA and NOXA proteins were synthesized by p53, knockdown of PUMA and NOXA using siRNAs did not suppress PGA 2 -induced apoptosis ( Figure S8). Not only pro-apoptotic BCL-2 family proteins but also death receptor (DR) proteins such as DR4 (tumor necrosis factor receptor superfamily member 10a, TNFRSF10A), DR5 (tumor necrosis factor receptor superfamily member 10b, TNFRSF10B), and FAS (Fas cell surface death receptor) are involved in the induction of p53-induced apoptosis [29,30]. Among death receptors, DR5 but neither DR4 nor FAS was significantly increased at the level of protein by PGA 2 , which was accompanied by phosphorylation of p53, an increase of p21 WAF1 protein, and cleavage of both PARP1 and caspase-3 ( Figure 4A). In the reporter-luciferase gene assay, the promoter activity of DR5 gene was upregulated by PGA 2 treatment in HCT116 cells but not in HCT116 p53-/-cells, which was suppressed by pretreatment with PFT-α, indicating an increase of DR5 expression at the level of transcription by PGA 2 -induced transcriptional activity of p53 ( Figure 4E).
Supporting the result of reporter gene assay, expression of DR5 mRNA was increased by PGA 2 treatment, and this increase was alleviated by PFT-α pretreatment ( Figure S9). Notably, knockdown of DR5 expression using siRNA suppressed both an increase of annexin V-positive cells ( Figure 4B) and cleavage of PARP1 ( Figure 4C) in PGA 2 -treated cells. Moreover, both NU7441 and PFT-α, which suppressed PGA 2 -induced apoptosis, prevented the expression of DR5 as well as cleavage of PARP1 ( Figure 4D). Collectively, these data suggested that DR5 upregulated by p53 plays a pivotal role in PGA 2 -induced apoptosis.

Discussion
The results of this study can be, collectively, summarized that PGA 2 induces the activation of p53 in HCT116 cells via DNA-PK, and p53, in turn, upregulates the expression of DR5 at the level of transcription, which finally leads to caspase-dependent apoptosis.
In this study, how PGA 2 increases the activity of DNA-PKcs was not clarified. The activation of DNA-PKcs occurs through multiple pathways [31]. DNA-PKcs is primarily activated in response to the damaged DNA via interaction with Ku70 and Ku80 proteins, which are recruited to the end of broken DNA. DNA-PKcs is also activated through protein kinases such as AKT and casein kinase II even in the absence of DNA damage. In the previous report, PGA 2 -induced DNA-PKcs activity was owing to reactive oxygen species (ROS)-induced DNA damage [21]. However, 15d-PGJ 2 , another cyclopentenone PG, induced activation of ATM, a protein kinase responsible for breakage of double stranded DNA, by direct interaction with ATM at cysteine residues, resulting in activation of ATM [32]. So, it may be possible that PGA 2 induces activation of DNA-PKcs through damaging DNA via ROS accumulation and physical interaction with DNA-PKcs as well. To identify the involvement of DNA damage in PGA 2 -induced apoptosis, phosphorylation of histone H2A variant H2AX at Ser-139 (γ-H2AX), which is a sensitive marker of DNA damage [33], was detected by immunoblot analysis. As shown in Figure S10A, PGA 2 treatment increased γ-H2AX remarkably, indicating that DNA damage occurred in PGA 2 -treated cells. Besides, pretreatment of PFT-α did not affect the level of γ-H2AX, while it reduced the level of cleaved PARP1 ( Figure S10A, Figure 4D). Moreover, PGA 2 treatment increased γ-H2AX in HCT116 p53-/-cells as well, which was not suppressed by either PFT-α or NU7441 ( Figure S10B). Therefore, it can be speculated that activation of DNA-PK may be due to damaged DNA, and both DNA damage and hence activation of DNA-PK occur upstream of p53 activation and induction of apoptosis.
Many target genes of p53 are involved in apoptosis [29][30][31]34]. As shown in Figure S11, mRNA levels of p53 target genes that were reported to induce apoptosis including FAS and FAS ligand (Fas-L) in PGA 2 -treated cells were not significantly increased by PGA 2 treatment. So, it seems that in HCT116 cells, PGA 2 -activated p53 induces expression of limited numbers of pro-apoptotic genes such as BAX, PUMA, NOXA, and DR5. Therefore, DNA-PK-p53-DR5 pathway might be a sole apoptosis-activating mechanism for PGA 2 -induced apoptosis in HCT116, although transcriptional target genes of p53 involved in apoptosis were not investigated at the level of transcriptome in this study, and thus the possibility of involvement of (a) novel gene(s) still remains. It is not clarified in this study how DNA-PK/p53 activation leads to specific transcriptional induction of DR5. Based on the report of Woo et al, it can be speculated that DNA-PK activation is necessary but not sufficient for p53-mediated DR5 transcription [35]. When HCT116 cells were treated with nutlin-3, an inhibitor of MDM2, which does not activate DNA-PK, p53 target genes were induced ( Figure S12) and moreover, transcription of DR5 was turned out to increase through stabilized p53 [36]. Thus, DNA-PK activity seems to be necessary for stabilization of p53 but not binding of p53 protein to p53 response element in promoter region of target genes.
Active p53 protein can affect both intrinsic and extrinsic apoptosis by upregulating the expression of several genes of each pathway or physical interaction with mitochondria [28]. Although it is not clear if p53 affect the intrinsic apoptosis or the extrinsic apoptosis at this point in PGA 2 -induced apoptosis in HCT116 cells, considering no attenuating effect of siRNAs against PUMA and NOXA on PGA 2 -induced apoptosis ( Figure S8) and no release of cytochrome c from mitochondria into cytosol ( Figure S13) leads to the speculation that PGA 2 -induced apoptosis in HCT116 cells may occur through the extrinsic apoptotic pathway which is activated by upregulated DR5.
Several reports have demonstrated that PGA 2 inhibits the growth and induce cell death in cancer cells [8,37]. Although the growth inhibitory pathway and apoptotic pathway induced by PGA 2 are all dependent on de novo protein synthesis, their relationship is not clearly established. In this study, PGA 2 induced apoptosis in HCT116 cells containing wild-type p53 gene to a much higher extent than that in HCT116 p53-/-cells. However, surprisingly, PGA 2 inhibited the growth of both HCT116 cells and HCT116 p53-/-cells to similar extent in a dose-dependent manner by CCK-8 assay which detects live cells ( Figure S14). In cell cycle distribution analysis, PGA 2 induced G2M arrest in both cell lines, but accumulated sub-G1 apoptotic cells, only in HCT116 cells ( Figure S15). Therefore, the growth inhibitory effect of PGA 2 may be exerted through both p53-dependent and -independent manners in HCT116 cells, although PGA 2 -induced apoptosis was dependent on p53 in HCT116 cells. Based on the result that PGA 2 induced G2M arrest at higher level in HCT116 p53 -/-cells than in HCT116 cell, it can be assumed that PGA 2 might activate stronger cell cycle arrest in HCT116 p53 -/-cells than that in HCT116 cells or PGA 2 might activate p53-independent growth inhibitory mechanism in HCT116 p53 -/-cells, finally resulting in similar CCK-8 results.
Besides, PGA 2 has also been reported to induce apoptosis in HL-60 cells and Hep3B cells whose p53 is deleted or mutated [11,15]. So, the dependency of PGA 2 -induced apoptosis on p53 may not be applied for all cell types. In HL-60 cells, PGA 2 induced intrinsic apoptosis by direct interaction with mitochondria without the involvement of de novo protein synthesis. In Hep3B cells, PGA 2 induced apoptosis via up-regulation of SOX-4, suggesting that PGA 2 can induce apoptosis via multiple pathways according to cellular contexts. Interestingly, PGA 2 induced apoptosis in HepG2 and Hep3B cells in a caspase-independent manner. Therefore, these reports suggested that PGA 2 induces apoptosis via the combinatorial pathways composed of caspase and new protein synthesis. The involvement of caspase activity may be determined by the proteins induced by PGA 2 .
Since cyclooxygenase-2 (COX-2) which synthesizes PGE 2 is highly expressed in many types of cancer, PGE 2 is elevated in cancer tissues. PGE 2 secreted from cancer tissues increases cancer cell survival, angiogenesis, invasion, and metastasis, playing a promoting role in carcinogenesis and immunosuppression in microenvironment around cancers [38,39]. Thus, COX-2/PGE 2 axis has been proposed as a therapeutic target in cancer treatment [40]. In the meanwhile, PGA 2 which is produced from PGE 2 by non-enzymatic dehydration, shows opposite effects to those of PGE 2 against cancers, proposing PGA 2 as a therapeutic molecule for cancer treatment [41]. So, it can be speculated that the relative amount between PGE 2 and PGA 2 in tumor microenvironment may contribute to the determination of the efficacy of both anti-cancer chemotherapy and radiotherapy, the effect of which is dependent on p53 activity. Therefore, the genetic information of cancer tissues should be integrated with profiles of inflammatory cytokines, including PGE 2 and PGA 2 in tumor microenvironment to maximize the effects of anti-cancer therapeutics in the future.

Conclusions
Treatment of PGA 2 induces caspase-dependent apoptosis in HCT116 cells containing wild type p53 gene, but not in HCT116 p53-/-cells which are deficient in functional p53. Activation of p53 by PGA 2 in HCT116 cells was dependent on the activity of DNA-PKcs. Among transcriptional target genes of p53, DR5 was responsible for the induction of apoptosis in PGA 2 -treated cells. PGA 2 increases the expression of DR5 at the level of transcription via p53 activity. Therefore, PGA 2 induces p53-dependent apoptosis by activating DNA-PKcs-p53-DR5 pathway in HCT116 colorectal cancer cells.
Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/2218-273X/10/3/492/s1, Figure S1: Induction of apoptosis by PGA 2 in HCT116 cells. Figure S2: Phosphorylation of p53 at Ser-46 by PGA 2 . Figure S3: Measurement of p53 target genes. Figure S4: Densitometric analysis of p53 target genes. Figure S5: The effect of NU7441 on survival of HCT116 p53-/-cells. Figure S6: The subcellular localization of p53 expression in PGA 2 -treated cells. Figure S7: The effect of CHX on the PGA 2 -induced apoptosis in HCT116 cells. Figure S8: The effect of knockdown of PUMA and NOXA expression on PGA 2 -induced apoptosis. Figure S9: The effect of PFT-α on the PGA 2 -induced increase of DR5 mRNA. Figure S10: Phosphorylation of histone H2AX by PGA 2 . Figure S11: Measurement of pro-apoptotic p53 target genes in PGA 2 -treated cells. Figure S12: Induction of p53 target genes by nutlin-3. Figure S13: Analysis of cytochrome c release in PGA 2 -treated cells. Figure S14: The effect of PGA 2 on the growth of HCT116 cells and HCT116 p53-/-cells. Figure S15: The effect of PGA 2 on the distribution of cell cycle of HCT116 cells and HCT116 p53-/-cells.