Antitumor Activity and Induction of TP53-Dependent Apoptosis toward Ovarian Clear Cell Adenocarcinoma by the Dual PI3K/mTOR Inhibitor DS-7423

DS-7423, a novel, small-molecule dual inhibitor of phosphatidylinositol-3-kinase (PI3K) and mammalian target of rapamycin (mTOR), is currently in phase I clinical trials for solid tumors. Although DS-7423 potently inhibits PI3Kα (IC50 = 15.6 nM) and mTOR (IC50 = 34.9 nM), it also inhibits other isoforms of class I PI3K (IC50 values: PI3Kβ = 1,143 nM; PI3Kγ = 249 nM; PI3Kδ = 262 nM). The PI3K/mTOR pathway is frequently activated in ovarian clear cell adenocarcinomas (OCCA) through various mutations that activate PI3K-AKT signaling. Here, we describe the anti-tumor effect of DS-7423 on a panel of nine OCCA cell lines. IC50 values for DS-7423 were <75 nM in all the lines, regardless of the mutational status of PIK3CA. In mouse xenograft models, DS-7423 suppressed the tumor growth of OCCA in a dose-dependent manner. Flow cytometry analysis revealed a decrease in S-phase cell populations in all the cell lines and an increase in sub-G1 cell populations following treatment with DS-7423 in six of the nine OCCA cell lines tested. DS-7423-mediated apoptosis was induced more effectively in the six cell lines without TP53 mutations than in the three cell lines with TP53 mutations. Concomitantly with the decreased phosphorylation level of MDM2 (mouse double minute 2 homolog), the level of phosphorylation of TP53 at Ser46 was increased by DS-7423 in the six cell lines with wild-type TP53, with induction of genes that mediate TP53-dependent apoptosis, including p53AIP1 and PUMA at 39 nM or higher doses. Our data suggest that the dual PI3K/mTOR inhibitor DS-7423 may constitute a promising molecular targeted therapy for OCCA, and that its antitumor effect might be partly obtained by induction of TP53-dependent apoptosis in TP53 wild-type OCCAs.


Introduction
The phosphatidylinositol 3-kinase (PI3K)-AKT signaling pathway is frequently activated in various types of cancers, and several inhibitors that target this pathway have been developed as potential cancer therapeutics. The constitutive activation of the PI3K-AKT pathway results from various types of alterations, including the overexpression of receptor tyrosine kinases (RTKs), as well as mutations of Ras, the catalytic subunit p110a of phosphoinositide-3-kinase (PIK3CA), and PTEN [1]. Class I PI3Ks include four isoforms of the catalytic subunit (p110a, p110b, p110c, and p110d). Among these four isoforms, p110a is broadly mutated and predominantly activated in various types of human cancers, although p110b and p110d might be selectively activated in certain tumors such as those with loss of PTEN function [2,3]. Mammalian target of rapamycin (mTOR) is the catalytic subunit found in two distinct complexes: the raptor-containing complex mTORC1 and the rictor-containing complex mTORC2 [4]. AKT activates mTORC1 signaling and also phosphorylates other downstream proteins, including GSK3b, forkhead box-O tran-scription factors (FOXOs), and mouse double minute 2 homolog (MDM2) [5]. mTORC1 controls protein synthesis and cell proliferation via the phosphorylation of its downstream targets, 4E-BP1 and S6 kinase 1 (S6K1) [6]. Rapamycin and its analogs (rapalogs) block mTORC1 activity, but not mTORC2 activity [7]. One of the AKT downstream targets, MDM2, is a negative regulator of TP53 that induces its ubiquitination and subsequent degradation [8]. Although the cytostatic effects of PI3K pathway inhibitors have been reported in various types of cancers [9][10][11][12], targeting the PI3K pathway might induce cytotoxic effects by suppressing anti-apoptotic signals through the dephosphorylation of FOXOs and stabilization of TP53. It seems reasonable to suspect that targeting the PI3K-mTOR axis might be a promising therapeutic strategy to selectively induce apoptosis of cancer cells, especially those without mutations in TP53.
Epithelial ovarian cancer is a leading cause of death resulting from gynecological malignancies. Ovarian clear cell adenocarcinoma (OCCA) is the second most common cause of death from ovarian cancer, with a higher incidence in Asia, especially in Japan (.25%), than in other continents [13]. OCCA is derived primarily from ovarian endometriosis, and the clinical outcome is generally poor, owing to low response rates to conventional platinum-based chemotherapy [14]. Thus, novel therapeutic strategies are warranted to improve the clinical outcome of OCCA. In histological terms, ovarian serous adenocarcinoma (OSA) is the most common variant of ovarian carcinoma [15]. It is highly sensitive to platinum-based chemotherapy, with a primary clinical response rate of .70%. The mutational spectrum differs between OCCA and OSA, with TP53 mutations observed in almost all (96%) OSA tumors, but in only 10% of OCCA tumors [15,16]. In particular, mutations of RB1 and BRCA1/2 are much more common in OSA than in OCCA. However, PIK3CA mutations are more frequent in OCCA (.40%) than in OSA (,10%) [17]. Although mutations of KRAS and PTEN are rare (,10%), the overexpression of several RTKs has been reported in OCCA, including human epidermal growth factor receptor 2 (HER2) with a frequency of approximately 40% and cMET with a frequency of approximately 30% [18][19][20][21]. Taken together, these observations suggest that the RTK-PI3K/mTOR signaling axis might be broadly activated in OCCA.
DS-7423 is a novel, small-molecule compound that inhibits both PI3K and mTOR (mTORC1/2). It inhibits all class I PI3K isoforms with greater potency against p110a than against the other p110 isoforms. Relevant activity (IC 50 ,200 nM) was not observed in any of the 227 kinases tested, except for Mixed Lineage Kinase 1 (MLK1) and Never-In-Mitosis Gene A (NIMA)related kinase 2 (NEK2). The compound is currently in phase I clinical trials for solid tumors. In this study, we evaluated its antitumor efficacy in a panel of OCCA cell lines. We focused in particular on the ability of DS-7423 to induce apoptosis, and on whether the apoptosis might be mediated by TP53.

Small-molecule compounds
The small molecule compound DS-7423 was provided by the Daiichi-Sankyo Company, Ltd (Tokyo, Japan). The drug information about DS-7423 is available on the ClinicalTrials.gov website (NCT01364844). The mTOR inhibitor rapamycin was purchased from Cayman Chemical (Michigan, USA).

Polymerase chain reaction (PCR) and sequencing
The mutational status of all nine OCCA cell lines was analyzed by PCR and direct sequencing. The PCR conditions and primers for the analysis of PTEN (exons 1-9) and K-Ras (exon 1 and 2) sequences were described previously [24][25][26]. The entire coding region of PIK3CA was analyzed by reverse transcription (RT)-PCR with LA-Taq according to the manufacturer's protocol (Takara BIO, Madison, WI) [27]. The PCR primers for analysis of TP53 (exons 4-8) were described previously [28].

Cell cycle analysis
Cells (5610 5 ) were seeded in 60-mm dishes and treated with DS-7423 for 48 h. Floating and adherent cells were collected by trypsinization and washed twice with phosphate buffer saline (PBS). Cells were resuspended in cold 70% ethanol and maintained at 4uC overnight. After being washed twice with PBS, cells were incubated in RNase A (0.25 mg/mL) (Sigma) for 30 min at 37uC, followed by staining with propidium iodide (PI; 50 mg/ mL) (Sigma) at 4uC for 30 min in the dark. Cells were then analyzed using flow cytometry (BD FACS Calibur HG, Franklin Lakes, NJ). Cell cycle distribution was analyzed using CELL Quest pro ver. 3.1. (Beckman Coulter Epics XL, Brea, CA). All experiments were repeated three times.
Detection of apoptosis by staining with annexin-V FITC Cells (5610 5 ) were cultured in 60-mm plates for 24 h before treatment with either DMSO (control), and 156 nM DS-7423, or 2,500 nM DS-7423 for 48 h. Cells were trypsinized, washed twice with PBS, and then analyzed after double staining with annexin-V fluorescein isothiocyanate (FITC) (Abcam, Cambridge, MA) and PI. The apoptotic cell population was analyzed using flow cytometry. All experiments were performed three times.

Ethics statement for animal experiments
This study was approved by Animal Care and Use Committee, Daiichi-Sankyo Pharmaceutical Co. Ltd. Athymic mice were maintained in an SPF (Specific Pathogen Free) facility according to our institutional guidelines, and experiments were conducted under an approved animal protocol.

Tumor xenografts in nude mice
Specific pathogen-free female nude mice (BALB/cAJcl-nu/nu), 6 weeks old, were purchased from CLEA Japan, Inc (Tokyo, Japan). Subcutaneous xenograft tumors in the mice were established by the injection of a 100-mL suspension containing 5610 6 cells of the TOV-21, RMG-I, or ES-2 lines in PBS. Tumors were removed after exponential growth, cut into 3-mm pieces, and transplanted subcutaneously into other mice for RMG-I cells. DS-7423 was suspended in 0.5 w/v% Methyl Cellulose 400 solution (Wako Pure Chemical Industries, Ltd.) Oral daily administration of DS-7423 started 8-22 days later, following the injection of the cells (5-6610 6 cells/0.1 mL). One week after tumor transplantation, mice were assigned randomly to one of the three treatment regimens: (1) non-treated control, (2) DS-7423 (1.5 mg/kg), (3) DS-7423 (3 mg/kg), and (4) DS-7423 (6 mg/kg). Each treatment group consisted of five mice. DS-7423 was injected orally (p.o.) once a day. Tumor volumes (in mm 3 ) were calculated by the formula: ([major axis] 6[minor axis] 2 /2). After the treatment, the tumors were removed and analyzed by western blotting. Tumor weight (wet weight) was measured, and the average weight was calculated for each group.

Semi-quantitative RT-PCR analysis
OCCA cells were treated with either DMSO or the indicated concentration of either DS-7423 or rapamycin for 24 h. Total RNAs of these cells were extracted with the RNeasy Mini Kit according to the manufacturer's instructions (QIAGEN, Valencia, CA). cDNAs were synthesized from total RNAs by using the Super Script III First-strand Synthesis SuperMix (Invitrogen, Carlsbad, CA). The exponential phase of the RT-PCR occurred between 15-30 cycles, and these cycles were monitored to allow semiquantitative comparisons among the cDNAs developed from identical reactions. The primers and conditions for the amplification of p53AIP1, p21, and GAPDH sequences were described previously [29]. The PCR primers for PUMA were 59-TGAGA-CAAGAGGAGCAGCAG- 39 (forward) and 59-AC-CTAATTGGGCTCCAT CTC-39 (reverse). The primers for p53R2, TIGAR, GLS2, GADD45, 14-3-3 sigma and PAI-1 were described previously [30][31][32][33][34][35]. Each PCR regimen involved a 2min initial denaturation step (94uC), which was followed by 15-30 cycles at 94uC for 30 s, then at 55uC for 30 s, and finally, at 72uC for 30 s using a Thermal Cycler Gene Atlas instrument (ASTEC, Fukuoka, Japan).

Gene silencing
Cells were plated at approximately 30% confluence in 100-mm plates and incubated for 24 h before transfection with small interfering RNA (siRNA) duplexes at the concentrations indicated, using Lipofectamine 2000 RNAiMAX (Invitrogen, Carlsbad, CA) and Opti-MEM medium (Life Technologies, Grand Island, NY). The siRNAs specific for TP53 were purchased from Invitrogen. A negative control kit was used as a control (Invitrogen, Carlsbad, CA).

Luciferase assay
Transfection was performed using Effectene reagent (QIAGEN, Valencia, CA) according to the manufacturer's recommendation. The TP53 expression plasmid (0.1 mg/mL) was cotransfected with pp53 TA Luc (0.25 mg/mL). The phRL CMV-Renilla plasmid (Promega, Madison, WI) was also transfected in all experiments as the internal control to normalize the transfection efficiency. The assays, each involving triplicate wells, were repeated three times.

Statistical analysis
The data were expressed as means 6 standard deviations of three independent determinations. The significance of the difference between two samples was analyzed using the Student's t-test, and a p-value of ,0.05 was considered to denote a statistically significant difference.

Results
Genetic alterations and activation of the PI3K-AKT signaling pathway in OCCA cell lines We evaluated the phosphorylation (p-) levels of the proteins in the PI3K-AKT pathway by using an immortalized epithelial cell line from an ovarian endometrial cyst as a control. AKT was phosphorylated at Thr308 in seven of the nine OCCA cell lines tested ( Figure 1). The cell lines OVMANA and ES-2 had low levels of p-AKT (Thr308) (Figure 1). The phosphorylation levels of S6, 4E-BP1 and/or FOXO1/3a, the downstream targets of AKT, were upregulated in the OCCA cells, including OVMANA and ES-2.
Four of the nine cell lines possessed PIK3CA mutations (44%) (Figure 1 and Table S1), and one of these four, TOV-21G, also possessed mutations in PTEN and K-Ras (11%). TP53 mutations were detected in three cell lines (33%) ( Table S1). The mutational status of PIK3CA was not associated with the phosphorylation of AKT or proteins that act downstream of AKT. Next, the expression and phosphorylation levels of three RTKs (HER2, HER3, and MET), which have been reported to be overexpressed in OCCA, were evaluated. The levels of phosphorylation of both HER2 (Tyr1221/1222) and HER3 (Tyr1289) were correlated with the abundances of these two proteins ( Figure 1). p-HER2 and p-HER3 levels were elevated in four (44%: OVISE, SKOV3, JHOC7 and RMG-I) and six (67%: TOV-21G, OVISE, OVMANA, OVTOKO, JHOC-7 and RMG-I) cell lines, respectively ( Figure 1). The expression of MET was higher in all nine OCCA cell lines than in the control, although the level of p-MET was increased in only two cell lines (22%: JHOC-7 and RMG-I).
Taken together, all the OCCA cell lines, except for ES-2 and JHOC-9, possessed one or more activating alterations in the RTK-PI3K genes examined ( Figure 1 and Table S1). Each of the four cell lines with PIK3CA mutations showed concomitant activation of RTKs, defined as high levels of phosphorylation of HER2 and/or HER3.

Anti-proliferative effect of DS-7423 in OCCA cell lines
We tested the anti-proliferative effects of the dual PI3K/ mTOR inhibitor, DS-7423, and the mTOR (mTORC1) inhibitor, rapamycin, in each of the nine OCCA cell lines. Exposure to 156 nM DS-7423 inhibited cell growth by 70%-97%, and the IC 50 values for cell proliferation were 20-75 nM ( Figure 2A). Dose-dependent growth suppression was more clearly induced by DS-7423 than by rapamycin in each of the nine cell lines (Figure 2A). The IC 50 value was not reached with rapamycin at any of the concentrations tested (2.45-2,560 nM) in five (OVMANA, SKOV3, OVTOKO, JHOC-7 and RMG-I) of the nine OCCA cell lines. We also examined the effect of DS-7423 in seven OSA lines. The IC 50 values with DS-7423 were .100 nM in four of these seven OSAs ( Figure 2B). The ratio of resistant cell lines (IC 50 .100 nM) was significantly higher in OSA cell lines (57%) than in OCCA cell lines (0%) (p = 0.019 by Fisher's exact test).
We performed immunoblotting on the lysates prepared from the cells treated with DS-7423 or rapamycin. DS-7423 suppressed the phosphorylation of AKT (Thr308 and Ser473) and S6 (Ser235/ 236 and Ser240/244) at doses of 39-156 nM and higher ( Figure 3A and Figure S1). DS-7423 suppressed the phosphorylation levels of the targeted proteins at comparable doses in the AKT pathway (AKT, FOXO1/3a, and MDM2) and mTORC1 pathway (S6). Rapamycin did not suppress p-Akt at any dose, and suppressed p-S6 at 2.45 nM or higher doses ( Figure 3B). On the   dose-dependent manner ( Figure 4B). No significant adverse effects, including body weight loss of more than 10%, were observed in the mice examined ( Figure 4B). Treatment with DS-7423 suppressed the levels of p-AKT (Thr308) and p-S6 (Ser240/244) in the TOV-21G and RMG-I xenografts ( Figure S2A). Compared with TOV-21G and RMG-I xenografts, the anti-tumor effect of DS-7423 was weaker in xenografts with ES-2, for which the basal level of p-Akt (Thr-308) was low ( Figure S2B).

Induction of apoptosis by DS-7423 in TP53 wild-type cell lines
The data collected from FACS analysis suggested that DS-7423 has a cytotoxic and cytostatic effect in certain OCCA cell lines. We combined the DS-7423 treatment (156 nM or 2,500 nM) with double staining with annexin-V FITC and PI to evaluate the proportion of cells that underwent apoptosis. DS-7423 at 156 nM induced apoptosis at 4-12% in five of the six cell lines that lacked mutations in TP53 ( Figure 5A and 5B). In these five cell lines, 2,500 nM DS-7423 induced apoptosis in 10-16% of the cells. In three cell lines with TP53 mutations, DS-7423 did not induce apoptosis in .5% of the cells at any of the doses tested ( Figure 5A). The size of the population of apoptotic cells was significantly higher in cells that lacked mutations in TP3 when compared with cells with mutated TP3 at either 156 nM (p = 0.0352) or 2,500 nM (p = 0.0368) DS-7423 according to the Student t-test ( Figure 5C). Rapamycin did not induce apoptotic cell death in .5% of the OCCA cells, even at 2,500 nM. The percentage of apoptotic cells was significantly higher in OVISE cells treated with DS-7423 than that in those treated with rapamycin ( Figure S3). This result indicates that mTORC1 inhibition alone is insufficient to induce apoptosis in OCCA cell lines. Immunoblotting analysis revealed that DS-7423 induced the cleavage of PARP within 2 h in OVMANA cells without mutations in TP53 ( Figure 5D). The induction of cleaved-PARP was observed at 39 nM, and the effect increased in a dose-dependent manner up to a concentration of 2,500 nM ( Figure 5D).

Induction of p-TP53 at Ser46 and expression of p53AIP1 by DS-7423
The phosphorylation of MDM2 is associated with the activation of MDM2 and degradation of TP53, with the phosphorylation of TP53 at Ser46 playing a key event in the TP53-dependent apoptosis (28). Treatment with DS-7423 reduced the level of p-MDM2 in a dose-dependent manner ( Figure 3A and 6A). Inversely, DS-7423 increased TP53 level even at lower doses, resulting in increased expression of p-TP53 (Ser15 and Ser46) ( Figure 6A). However, only p-TP53 (Ser46), not p-TP53 (Ser15), was clearly induced by high doses of DS-7423 (156-2,500 nM). We then used semi-quantitative RT-PCR to evaluate the regulation of genes that are directly regulated by TP53 in OVMANA and OVISE cells. DS-7423 induced the expression of the pro-apoptotic genes p53AIP1 and PUMA at 39 nM or higher doses, but did not induce the expression of p21 at any of the three doses tested (39,156, and 2,500 nM) ( Figure. 6B and 6C). We also performed semi-quantitative RT-PCR of other TP53 target genes involved in DNA repair (p53R2), metabolism (TIGAR and GLS2), G2/M arrest (GADD45), and cell cycle arrest/senescence (14-3-3 sigma and PAI-1) to test whether other TP53 target genes are induced by DS-7423. GADD45 was significantly induced by DS-7423 in OVISE cells ( Figure 6B and 6C), in which G2/M arrest was enhanced by DS-7423 according to the MTT assay ( Figure 4A). The other TP53-downstream genes tested were not induced by DS-7423 in both OVISE and OVMANA cells, and expression of TIGAR was rather decreased in OVMANA cells ( Figure S4).

TP53 activation is responsible for DS-7423-mediated apoptosis
We used siRNAs specific to TP53 to knockdown TP53 expression in OVISE cells, and treated the cells with DS-7423 at either 156 or 2,500 nM. The size of the population of apoptotic cells was calculated by annexin-V FITC-PI double staining 48 h after treatment of DS-7423. Knockdown of TP53 levels rescued cells from apoptotic cell death induced by treatment with both DS-7423 doses ( Figure 6D). Immunoblotting indicated that two independent siRNAs (siRNA1 and siRNA2) specific to TP53 suppressed the expression of TP53 by .80% ( Figure 6E). Next, we performed the MTT assay by applying both DS-7423 and siRNA to TP53 in OVISE cells (wild-type TP53). The anti-proliferative effect of DS-7423 was significantly reduced when combined with the knockdown of TP53 ( Figure 6F). The effect of DS-7423 on the transcriptional activity of TP53 was also examined by luciferase assays in ES-2 cells with mutations in TP53. The cells were treated with DS-7423 for 24 h at the indicated doses, and then cotransfected with both pp53-TA-luc plasmid (containing TP53 binding sites) and a plasmid that encodes TP53. The relative luciferase activity of TP53 was significantly enhanced by DS-7423 in a dose-dependent manner ( Figure 6G).

Discussion
The effects of the PI3K/mTOR inhibitor, DS-7423, on OCCA cell lines were examined with a particular focus on (i) the antiproliferative effect of DS-7423, (ii) the induction of apoptosis by DS-7423, and (iii) the identification of predictive biomarkers for (i) and (ii).
MTT assays revealed a clear dose-dependent effect of DS-7423 on cell proliferation, with all nine OCCA cell lines displaying sensitivity to DS-7423 (IC 50 at 75 nM or lower), regardless of mutations on PIK3CA. The sensitivity to DS-7423 was significantly higher in OCCA than in OSA cell lines. The prevalence in OCCA cell lines of activating mutations in genes that encode components of the RTK-PI3K-AKT signaling pathway might account, at least in part, for their broad sensitivity to DS-7423. Differences in the dose-dependence of the anti-proliferative effects of DS-7423 and rapamycin suggest differences in the modes of action of these two drugs. Whereas DS-7423 showed a more robust anti-proliferative  these cells. Thus, the action mechanism of DS-7423 might be distinct in each type of cells, regardless of the TP53 status. Resistance to mTOR (mTORC1) inhibitors might be induced by several mechanisms, including increased activity of another mTOR complex, mTORC2, or upregulation of receptor tyrosine kinases such as insulin-like growth factor-1 receptor (IGF-R1) [37,38]. The use of mTORC1 inhibitors to treat OCCAs is currently being investigated in phase 2 clinical trials. The currently ongoing GOG (Gynecologic Oncology Group)-0268 (NCT01196429) trial recruits OCCA patients and treats the subjects with carboplatin and paclitaxel, followed by temsirolimus (CCI-779). A report on six cases with weekly administration of temsirolimus in recurrent OCCA patients showed partial response in one patient and stable disease in another patient [39]. However, given that our data suggest that dual PI3K/mTOR inhibitors, such as DS-7423, might be more promising than single mTORC1 inhibitors, clinical trials that involve a dual PI3K/mTOR inhibitor, such as DS-7423, seem warranted for OCCA.
DS-7423 induced significantly higher levels of apoptotic cell death in OCCA cells without mutations in TP53 than in OCCA cells with TP53 mutations. This result suggests both that the mutational status of TP53 might be a good biomarker to predict apoptosis induction by DS-7423, and that apoptosis depends on TP53 function. TP53 is degraded by MDM2, a ubiquitin ligase for TP53, and the MDM2 function is augmented by the kinase activity of Akt. Akt-mediated phosphorylation of MDM2 blocks its binding to p19ARF, increasing the degradation of TP53 [40,41]. DS-7423 increased the level of p-TP53 at Ser46, which results in induction of p53AIP1 and PUMA (genes involved in TP53mediated apoptosis) [29,[42][43][44]. This data suggests that the apoptotic effect of DS-7423 depends, at least in part, on TP53 activity. The reasons for p-TP53 (Ser46), not p-TP53 (Ser15), being clearly induced and for apoptosis being preferentially knockdown on apoptosis induction by DS-7423. TP53 was knocked down by two independent siRNAs specific to TP53 (siRNA1 and 2) in OVISE cells, which do not carry any mutation in TP53. The apoptotic cell population was evaluated using annexin-V staining, as described in Figure 5. The experiments were repeated 3 times, and each value is shown as the mean of 3 experiments 6 SD. *p,0.05 (E) Suppression of TP53 expression by siRNAs was confirmed by immunoblotting. (F) Effect of TP53 knockdown on cell proliferation by DS-7423 in MTT assay of OVISE cells. TP53 was knocked down by a siRNA1 specific to TP53 and MTT assay was subsequently performed as in Figure 2. Knockdown of TP53 diminished the antiproliferative effect caused by DS-7423 on OVISE cells. The experiments were repeated 3 times, and each value is shown as the mean of 3 experiments 6 SD. *p,0.05 (G) TP53 expression plasmid (0.1 mg/mL) was cotransfected with pp53 TA Luc (0.25 mg/mL) plasmid into ES-2 cells mutated in TP53. The addition of DS-7423 increased the relative luciferase activity of TP53 in a dose-dependent manner. The experiments were repeated 3 times, and each value is shown as the mean of 3 experiments 6 SD. *p,0.05. doi:10.1371/journal.pone.0087220.g006 PI3K/mTOR Inhibitor in Ovarian Clear Cell Cancer PLOS ONE | www.plosone.org induced by high doses of DS-7423 should be further clarified. In addition, other non-apoptotic genes were not significantly induced by DS-7423, except for GADD45 in OVISE cells. Further analyses are warranted whether TP53 function is more involved in apoptosis rather than in cell cycle arrest and/or DNA repair process by DS-7423. Another possibility is that other proteins (such as FOXOs) which act downstream of Akt might also play a role in the induction [45]. Dephosphorylation of FOXOs at their Akt sites induces their nuclear translocation and triggers apoptosis by induction of prosurvival genes of the BCL2 family [46,47]. The observation that the phosphorylation of FOXO1/3a was suppressed by DS-7423, regardless of TP53 status, suggests that the pro-apoptotic effect of DS-7423 cannot be explained exclusively by the phosphorylation of FOXOs. The use of siRNA to knockdown TP53 rescued OCCA cells from apoptosis caused by DS-7423. We also confirmed by MTT assay that the antiproliferative effect of DS-7423 was significantly diminished by knocking down TP53, suggesting that intact TP53 function might enhance the anti-tumor effect of DS-7423. Recently, it was reported that cell death caused by a PI3K inhibitor, BKM-120, was associated with TP53 status in glioma cells [48], and that PI3K/AKT inhibition was suggested to induce TP53-dependent apoptosis in HTLV-1-transformed cells [49]. These data also support the importance of wild-type TP53 in the induction of the cytotoxic effect of PI3K pathway inhibitors.
The involvement of multiple molecules in the activation of the PI3K/mTOR pathway underscores the critical need to develop predictive biomarkers that might also serve as therapeutic targets. Mutations of PIK3CA and amplification of HER2 have been proposed to be useful biomarkers in breast cancer [50,51], whereas mutant Ras has been suggested to be a biomarker of resistance in several solid tumor cells [52]. All these biomarkers (PIK3CA, HER2 and Ras) are focused on the RTK/Ras/PI3K pathway itself, and not on the cytotoxic effects associated with PI3K/ mTOR inhibitors. Our data suggest that the presence of PIK3CA mutation and any other PI3K-activating alteration alone might not predict the sensitivity of OCCA cells to DS-7423. ES-2 cells, with no mutations in the RTK/Ras/PI3K pathway genes examined, showed low level of p-Akt, and the effect of DS-7423 in ES-2 xenografts was less robust, suggesting that the level of PI3K pathway activation would still be important for the sensitivity. However, the mutational status of TP53 might represent a better biomarker for the selection of tumors that could be killed by DS-7423 treatment. The frequency of mutations in TP53 in OCCA was much less frequent than for ovarian cancers with other histology types [15,53]. These results indicate that OCCAs would be good candidates for clinical studies on the dual PI3K/mTOR inhibitor, DS-7423.
Our study has several limitations. First, cytostatic effect is still essential to suppress cell proliferation, regardless of TP53 status. Second, the ratio of apoptotic cells is low (less than 20%) even at high concentrations of DS-7423. Third, the mechanism of cytostatic effect by DS-7423 in OCCA is cell type dependent (i.e. G1 arrest was not induced in OVISE and ES-2 cells). Thus, careful consideration is required to evaluate the TP53-dependent cytotoxic effect of DS-7423. Further studies are warranted to elucidate the mechanism of action of DS-7423, and more efficient induction of apoptosis might be needed for clinical application of this drug in OCCA. Figure S1 Immunoblotting of OCCA cells (ES-2 and JHOC-9), treated with DS-7423 at concentrations ranging from 0 to 2,500 nmol/L. As shown in Figure 3, phosphorylation of AKT and its target proteins were downregulated by DS-7423. In ES-2 cells, basal level of p-AKT at Thr 308 was very low (as shown in Fig. 1), but p-AKT at Ser473 was clearly suppressed by DS-7423. (PPTX) Figure S2 In vivo effect of DS-7423 in nude mice. (A) Western blot of total lysates from the TOV-21G and RMG-1 xenografts. total lysates were harvested 2 and 6 h after the last drug administration of DS-7423. The levels of p-Akt (Thr-308) and p-S6 (Ser-240/244) were assessed. (B) Subcutaneous xenograft tumors in athymic BALB/c mice were established after injection of ES-2 cells. Mice were treated daily at the indicated doses (1.5, 3 or 6 mg/kg/day, totally 8 times) of DS-7423 or non-treated control. Estimated tumor volumes were smaller in mice treated daily with 6 mg/kg of DS-7423, compared to the control. Western blot of total lysates from the ES-2 xenografts (treated with 6 mg/kg of DS-7423) was also shown below. (PPTX) Figure S3 The size of apoptotic cell population was compared between DS-7423 and rapamycin in OVISE cells, using annexin-V FITC and PI double staining (as shown in Fig. 5A-5B). The percentage of apoptotic cells was significantly higher in cells treated with DS-7423, compared with those with rapamycin. (PPTX) Figure S4 Semi-quantitative RT-PCR in OVMANA and OVISE cells treated with DS-7423 at the indicated doses. Each expression level of p53R2, TIGAR, GLS2, GADD45, 14-3-3 sigma and PAI-1 was not enhanced by DS-7423. Each experiment was repeated 3 times, and each value is shown as the mean of 3 experiments 6 SD. (PPTX)

Supporting Information
Table S1 Phosphorylation and mutational status in 9 OCCA cell lines. Elevated phosphorylation of cMET, HER2 and HER3, and mutations of PIK3CA, PTEN, KRAS and TP53 were listed in each cell line. (XLSX)