GNE-493 inhibits prostate cancer cell growth via Akt-mTOR-dependent and -independent mechanisms

GNE-493 is a novel PI3K/mTOR dual inhibitor with improved metabolic stability, oral bioavailability, and excellent pharmacokinetic parameters. Here GNE-493 potently inhibited viability, proliferation, and migration in different primary and established (LNCaP and PC-3 lines) prostate cancer cells, and provoking apoptosis. GNE-493 blocked Akt-mTOR activation in primary human prostate cancer cells. A constitutively-active mutant Akt1 restored Akt-mTOR activation but only partially ameliorated GNE-493-induced prostate cancer cell death. Moreover, GNE-493 was still cytotoxic in Akt1/2-silenced primary prostate cancer cells. Significant oxidative stress and programmed necrosis cascade activation were detected in GNE-493-treated prostate cancer cells. Moreover, GNE-493 downregulated Sphingosine Kinase 1 (SphK1), causing ceramide accumulation in primary prostate cancer cells. Daily single dose GNE-493 oral administration robustly inhibited the growth of the prostate cancer xenograft in the nude mice. Akt-mTOR inactivation, SphK1 downregulation, ceramide level increase, and oxidative injury were detected in GNE-493-treated prostate cancer xenograft tissues. Together, GNE-493 inhibited prostate cancer cell growth possibly through the Akt-mTOR-dependent and -independent mechanisms.


INTRODUCTION
Prostate cancer is the prostate epithelial malignancy affecting male populations [1,2]. Even with the latest development of prostate-specific antigen (PSA) screening, it still accounts for a considerable proportion of global cancer-related mortalities [1,2]. Radiotherapy and surgery remain to be the primary curative therapies in the management of localized prostate cancer [3][4][5][6].
Both mTORC1 and mTORC2 are required for the development and progression of prostate cancer [5,7,8]. Guertin et al. found that mTORC2 activation is important for PTEN loss-induced development of prostate cancer in mice [12]. Jiang et al. reported that a mTORC1/2 dual inhibitor INK-128 impeded prostate cancer cell growth [13]. Rapamycin (and its analogs) could partly suppress the activation of mTORC1, showing no direct effect on mTORC2 [14,15]. Moreover, mTORC1 inhibition could induce feedback and sustained activation of Akt and Erk, both are pro-cancerous cascades [16,17]. The ATP-competitive mTOR inhibitors were developed. These mTOR kinase inhibitors can block mTORC1 and mTORC2 simultaneously, often resulting better and more dramatic anti-cancer results [18].
Other prostate cancer cells, including the primary human prostate cancer cells of two other patients, priCa-2 and priCa-3, as well as the immortalized prostate cancer cell lines, LNCaP and PC-3, were also treated with GNE-493 (250 nM). The PI3K-mTOR dual inhibitor robustly inhibited CCK-8 cell viability (Fig. 1G) and induced significant cell death (Fig. 1H) in the prostate cancer cells. In addition, GNE-493 (250 nM) robustly inhibited proliferation (tested by the decreased ratio of EdU-incorporated nuclei, Fig. 1I) and migration (Fig. 1J) of the primary and immortalized prostate cancer cells.

GNE-493 induces oxidative stress and programmed necrosis in prostate cancer cells
Although GNE-493 provoked apoptosis activation in prostate cancer cells, the caspase inhibitors failed to completely reverse GNE-493-induced cell death (see Fig. 2). Recent studies have proposed an active and "programmed" mitochondrial necrosis cascade, or programmed necrosis. It is essential in mediating cancer cell death by a number of anticancer agents and various other stimuli [20][21][22][23][24][25][26][27][28][29]. Different cancer-killing agents could induce robust ROS production and significant oxidative injury, leading to p53 protein translocation to mitochondrion and association with CyPD and ANT1. The formation of this mitochondrial p53-CyPD-ANT1 complex will then induce the opening of mPTP channel (mitochondrial permeability transition pore) and depolarization of mitochondria, and eventually causing cell necrosis [22,24,26,30,31]. By employing a CellROX staining assay, we found that levels of ROS (by measuring the CellROX intensity) were significantly increased in GNE-493-treated priCa-1 and priCa-2 primary cancer cells (Fig. 4A). Significant lipid peroxidation, reflected by the increase of the TBAR activity, was observed as well ( Fig. 4B). Moreover, ssDNA accumulation was detected after GNE-493 treatment, reflecting increased DNA damage in priCa-1 and priCa-2 primary cancer cells (Fig. 4C).
"Inputs"). Significantly, the medium LDH contents were increased in GNE-493-treated primary cancer cells (Fig. 4F), supporting cell necrosis induction. Therefore, GNE-493 induced significant oxidative injury and programmed necrosis cascade in prostate cancer cells.

DISCUSSION
Due to PTEN depletion, PIK3CA amplification, and other genetic changes, the PI3K-Akt-mTOR pathway is most frequently overactivated in prostate cancer [5,7,8], which is associated with cancer progression, cancer metastasis, as well as development of drug resistance [5,7,8]. Preclinical studies examining specific PI3K-Akt-mTOR inhibitors have yielded promising results, yet the results of the clinical trials are less convincing [5,7,8]. Emerging studies have suggested that dual PI3K-mTOR inhibitors, which can block the entire PI3K-Akt-mTOR cascade, could achieve better and more robust anti-cancer efficiency [5,7,8].
Here we showed that GNE-493 efficiently inhibited prostate cancer cell growth. In multiple primary prostate cancer cells and established cell lines (LNCaP and PC-3), GNE-493 robustly inhibited cell viability, proliferation, and in vitro migration, while inducing cell cycle arrest and provoking apoptosis. GNE-493 blocked Akt-mTOR activation in prostate cancer cells. In vivo studies showed that a daily single dose of GNE-493 administration significantly inhibited prostate cancer xenograft growth in nude mice. In GNE-493-treated prostate cancer xenografts, Akt-mTOR inactivation was also detected.  1 μM), the control cells were treated with the control shRNA lentiviral particles ("shC") plus DMSO (0.1%); Cells were cultivated for additional 12 h, levels of SphK1 protein (H) and ceramide contents (I) were tested. *P < 0.05 versus "Veh" group. # P < 0.05 versus "GNE-493" only treatment. "n.s." stands for non-statistical difference.
Akt-mTOR blockage however is not the sole mechanism responsible for prostate cancer cell death by GNE-493. caAkt1 reserved Akt-S6K1 phosphorylations yet only partially attenuated GNE-493-induced prostate cancer cell death. GNE-493 was still cytotoxic and induced apoptosis in Akt1/2-silenced cancer cells. Indeed we discovered that Akt-mTOR-independent mechanisms by GNE-493, including ROS production and oxidative injury, programmed necrosis, SphK1 downregulation, and ceramide accumulation, were also important in mediating prostate cancer cell death.
A number of anti-cancer agents could provoke programmed necrosis cascade. Different stimuli will induce p53 mitochondrial translocation and association with CyPD and ANT-1 to open mPTP. This will lead to MMP reduction and finally cause cell necrosis [22,24,26,30,31]. PF-543, the known SphK1 inhibitor, activated programmed necrosis pathway in human colorectal cancer (CRC) cells [21]. AICAR activated AMPK-independent programmed necrosis cascade to cause death of prostate cancer cells [22]. Qin et al. found that glioma cell death by salinomycin was due to programmed necrosis, which can be inhibited by CyPD silencing or inhibition [23].
In addition, a number of oxidative stimuli could also provoke the programmed necrosis cascade [27,[34][35][36]. Here we discovered that GNE-493 provoked oxidative injury and programmed necrosis cascade in prostate cancer cells. Significant ROS production, increased lipid peroxidation, and accumulated DNA damage were detected in GNE-493-treated prostate cancer cells. Moreover, GNE-493 induced p53 translocation to mitochondria and p53-CyPD-ANT1 complex formation as well as mitochondrial depolarization and necrosis. Importantly, the antioxidant NAC, the CyPD inhibitor CsA as well as CyPD shRNA ameliorated GNE-493induced cytotoxicity in prostate cancer cells. Therefore, besides apoptosis, oxidative stress and programmed necrosis are contributors of GNE-493-induced prostate cancer cell death.
SphK1 is overexpressed and/or overactivated in prostate cancer, serving as an important diagnosis marker and therapeutic target [37,38]. It has been shown that circulating S1P contents and erythrocyte SphK1 activity could be novel and efficient biomarkers for the early detection of prostate cancer [39]. Sauer et al. reported that pharmacological inhibition or siRNA-mediated silencing of SphK1 sensitized hormone-resistant prostate cancer cells to docetaxel-induced apoptosis [40]. Pchejetski et al. reported that FTY720, a SphK1 inhibitor, induced prostate cancer cell apoptosis [41]. Dayon et al. showed that SphK1 activation upon chronic androgen deprivation was essential for prostate cancer cell growth and survival [42]. We here discovered that SphK1 protein downregulation and ceramide accumulation, independent of Akt-mTOR inactivation, were also important for prostate cancer cell death by GNE-493. the xenografts were all carefully isolated and weighted at Day-42 (C). At experimental Day-3 and experimental Day-9, one tumor from each group was isolated 6 h after initial GNE-493 or vehicle administration. In the tumor tissues listed proteins were examined (E and F). Total ceramide contents (G) and the relative TBAR activity (H) were examined as well. *P < 0.05 versus "Veh" group.

Cell lines
The immortalized prostate cancer cell lines, LNCaP and PC-3, as well as RWPE1 immortalized prostate epithelial cells, were from Dr. Wang [13]. Cells were cultivated using the described protocols [13].
Primary culture of patient-derived primary human prostate cancer cells Three primary prostate cancer patients undergoing prostate resection, with the written-informed consent, were enrolled in this study. The prostate cancer tissues and adjacent surrounding normal prostate tissues were carefully separated. Tissues were minced and digested as described [13]. Primary prostate cancer cells or prostate epithelial cells were cultivated in medium descried [13]. Primary prostate cancer cells were derived from three patients, "priCa-1", "priCa-2", and "priCa-3". The primary prostate epithelial cells were from one patient and were named as "priEpi" cells. This study and the protocols were approved by the Ethics Committee of The Second Affiliated Hospital of Soochow University and were in accordance with Declaration of Helsinki principles.

Colony formation assay
The prostate cancer cells were treated as descried and re-suspended in complete medium with 0.5% agar (Sigma) and plated on the top of a 10cm diameter culture dish. Every two days GNE-493-containing medium was renewed. Afterwards, the number of viable prostate cancer cell colonies were manually counted.

FACS
Briefly, cells were fixed and stained with propidium iodide (PI, at 10 μg/mL) for 30 min. The FACS assays were carried out to examine cell cycle distribution.

Constitutively-active mutant Akt1
Prostate cancer cells were cultivated in the polybrene-containing complete medium (with FBS) and treated with the recombinant adenoviral constitutively-active mutant S473D Akt1 (caAkt1, from Dr. Li [46,47]). Expression of caAkt1 in the infected cells was verified by Western blotting.

Akt1/2 shRNA
Prostate cancer cells were cultivated in polybrene-containing medium (with FBS) and infected with the Akt1/2 shRNA lentiviral particles (catalog number sc-37030-V, Santa Cruz Biotech) for a total of 24 h. Afterwards, puromycin (3.5 μg/mL) was included the medium and stable cells were established.
Lipid peroxidation assay. Prostate cancer cells were initially seeded into the six-well plates at 8 × 10 4 cells per well and subjected to the applied GNE-493 treatment. The lipid peroxidation levels in total cellular lysates and tissue lysates were measured and quantified using thiobarbituric acid reactive substances (TBAR) method according to the detailed protocols described [48,49].

Reactive oxygen species (ROS) detection
Prostate cancer cells were initially seeded onto the six-well plates at 8 × 10 4 cells per well and subjected to applied GNE-493 treatment. Afterwards, cells were washed with cold PBS and stained with CellROX (3.5 μg/mL). CellROX fluorescence intensity, reflecting ROS contents, was detected by the fluorescence spectrofluorometer at 625 nm.

Ceramide assay
Prostate cancer cells were subjected to applied GNE-493 treatment. Ceramide contents in total cellular lysates and tissue lysates were tested via the protocol as described previously [50].

CyPD shRNA
In brief, the primary human prostate cancer cells were treated with CyPD shRNA lentiviral particles (Santa Cruz Biotech) for 48 h. Afterwards, to select stable cells puromycin (2.5 μg/mL) was added for another 4-5 days. CyPD silencing, with over 90-95% knockdown efficiency, was tested by Western blotting assays.

SphK1 expression
A SphK1 over-expressing lentiviral construct was from Dr. Yao [53]. The SphK1-expressing construct, together with the lentivirus Helper plasmids, were co-transfected to HEK-293T cells. The SphK1 over-expressing lentivirus, LV-SphK1, was thereafter added to primary prostate cancer cells. SphK1 overexpression was always confirmed by Western blotting assays.

Xenograft assay
The male nude mice, at 6-8 weeks old, 18.5-19.0 g, were purchased from the animal facility of The Second Affiliated Hospital of Soochow University, and maintained under the Institutional Animal Care Use Committee guidelines. priCa1 primary prostate cancer cells were mixed at 1:1 ratio with Matrigel. To the right flanks of the mice, six millions cells (in 100 μL suspension) were subcutaneously (s.c.) injected to each nude mice. Treatment began four weeks post pri-Ca1 cells implantation, and the volume of each tumor reached around 100 mm 3 . The xenograft bearing nude mice were orally administrated with GNE-493 (20 mg/kg) or the vehicle control daily for 12 consecutive days [54]. Measuring the tumor volumes, the mice body weights, and the tumor weights were determined as described [55]. To test signaling changes, tumor tissues were minced, homogenized, and dissolved in tissue lysis buffer. Protocols of handling the nude mice were reviewed and approved by the Institute Animal Ethics Review Board of The Second Affiliated Hospital of Soochow University.

DATA AVAILABILITY
All data are available upon request.