The therapeutic value of XL388 in human glioma cells

XL388 is a highly efficient and orally-available ATP-competitive PI3K-mTOR dual inhibitor. Its activity against glioma cells was studied here. In established and primary human glioma cells, XL388 potently inhibited cell survival and proliferation as well as cell migration, invasion and cell cycle progression. The dual inhibitor induced significant apoptosis activation in glioma cells. In A172 cells and primary human glioma cells, XL388 inhibited Akt-mTORC1/2 activation by blocking phosphorylation of Akt and S6K1. XL388-induced glioma cell death was only partially attenuated by a constitutively-active mutant Akt1. Furthermore, it was cytotoxic against Akt1-knockout A172 glioma cells. XL388 downregulated MAF bZIP transcription factor G (MAFG) and inhibited Nrf2 signaling, causing oxidative injury in glioma cells. Conversely, antioxidants, n-acetylcysteine, pyrrolidine dithiocarbamate and AGI-106, alleviated XL388-induced cytotoxicity and apoptosis in glioma cells. Oral administration of XL388 inhibited subcutaneous A172 xenograft growth in severe combined immunodeficient mice. Akt-S6K1 inhibition and MAFG downregulation were detected in XL388-treated A172 xenograft tissues. Collectively, XL388 efficiently inhibits human glioma cell growth, through Akt-mTOR-dependent and -independent mechanisms.


INTRODUCTION
Glioma is the most common primary brain tumor and among the most aggressive of human cancers [1,2]. Over 10, 000 people are diagnosed with glioma each year in the United States, mostly with high-grade tumors [3]. The average survival of glioma patients is less than a year after initial diagnosis [1,2]. Significant progress has been made in glioma treatments, including neurosurgical resection, radiation and chemotherapy [1,2]. However, the five-year survival rate remains very disappointing [1,2]. It is therefore urgent to explore more efficient targeted therapies [4].
By applying "Transwell" and "Matrigel Transwell" assays, we show that XL388 (250 nM) inhibited A172 cell migration ( Figure 1F) and invasion ( Figure 1G) in vitro. The quantitative analysis demonstrated that it significantly reduced the number of migrated ( Figure  1F) and invaded ( Figure 1G) A172 cells. Analyzing cell cycle, by the propidium Iodide (PI)-FACS assay, show that XL388 (250 nM) treatment in A172 cells led to an increase in G1 phase cells, but decreases in S-/G2-phase cells ( Figure 1H), indicating G1-S arrest in XL388treated cells. For the cell migration/invasion and cell cycle analyses, cells were treated with XL388 (250 nM) for 24h or less, when no significant cytotoxicity or proliferation inhibition were detected ( Figure 1A). Therefore, XL388 potently inhibited A172 cell viability, proliferation, migration, invasion and cell cycle progression. The potential effect of XL388 in other human glioma cells was studied as well. As described early, the primary human glioma cells, Pri-1/Pri-2 (derived from two different patients [12][13][14]), as well as the established U251 glioma cells, were cultured and treated with XL388 (250 nM). XL388 treatment resulted in significant inhibition of cell viability (CCK-8 OD, Figure 1I), proliferation (EdU incorporation, Figure 1J) and migration ("Transwell" assays, Figure  1K).

XL388 induces significant apoptosis activation in glioma cells
In human cancer cells, proliferation inhibition and cell cycle arrest could induce cell apoptosis. We therefore studied the potential effect of XL388 on glioma cell apoptosis. In A172 cells, the caspase-3 activity ( Figure  2A) and the caspase-9 activity ( Figure 2B) increased over 4-6 folds after XL388 treatment (250 nM, 24h). Furthermore, cleavages of caspase-3, caspase-9 and PARP were detected in XL388-treated A172 cells ( Figure 2C). Following XL388 treatment, mitochondrial depolarization was detected in A172 cells, which was evidenced by accumulation of JC-1 green monomers ( Figure 2D). Further studies demonstrated that TUNELpositive cell nuclei ratio was significantly increased following XL388 treatment in A172 cells ( Figure 2E). Cell apoptosis was further supported by a significant increase of Annexin V-positive staining in XL388treated cells ( Figure 2F). In U251 cells and Pri-1/Pri-2 primary glioma cells, XL388 similarly increased AGING caspase-3 activity ( Figure 2G) and induced apoptosis activation (TUNEL assays, Figure 2H). Significantly, XL388, at 250 nM, failed to inhibit cell viability ( Figure  2I) and induce cell apoptosis ( Figure 2J) in primary human astrocytes ("Astrocytes") and HCN-1a neuronal cells. Both are non-cancerous cells [12,13]. These results demonstrated that XL388 induced significant apoptosis activation in glioma cells.
Notably, the three were utilized at higher concentrations than XL388.

Figure 2. XL388 induces significant apoptosis activation in glioma cells. A172 cells (A-F), U251MG ("U251") (G and H) and primary
human glioma cells ("Pri-1/Pri-2") (G and H) as well as the primary human astrocytes ("Astrocytes") and HCN-1a neuronal cells (I and J) were treated with XL388 (250 nM), and cultured for applied time periods, then cell apoptosis was analyzed by the mentioned assays (A-H and J), with cell viability tested by CCK-8 assay (I). Data were presented as mean ± SD (n=5). * p <0.05 vs. "C" cells. Experiments in this figure were repeated three times, and similar results were obtained. Bar= 100 μm (D and E).
In A172 cells, Akt1 KO (see Figure 3) failed to downregulate MAFG expression ( Figure 4I) or inducing ROS production (CellROX intensity, Figure 4J). However, XL388 induced MAFG downregulation ( Figure 4I) and significant ROS production ( Figure 4J) in the Akt1-KO A172 cells. These results further supported that XL388-induced MAFG downregulation and ROS production are independent of Akt inhibition in glioma cells. In the primary human astrocytes, MAGF expression is low ( Figure 4K). XL388 treatment had little effect on ROS in astrocytes ( Figure 4L). This could also explain why XL388 is ineffective against astrocytes ( Figure 2I and 2J).

XL388 oral administration inhibits A172 xenograft growth in severe combined immunodeficient (SCID) mice
As described in our previous studies [12,13], A172 glioma cells were injected s.c. to the SCID mice. The A172 tumor xenografts were established within two weeks (volume of each tumor around 100 mm 3 , "Day-0"). The tumor-bearing SCID mice were treated with XL388 or the vehicle control. As shown, oral administration of XL388 (5 mg/kg body weight, daily, × 14d) [16,24] potently inhibited A172 xenograft growth in SCID mice ( Figure 5A). Volumes of XL388-treated A172 xenografts were lower than those of vehicle control tumors ( Figure 5A). The estimated daily tumor growth, calculated by the formula: (Tumor volume at Day-35-Tumor volume at Day-0)/35, again demonstrated that XL388 significantly inhibited A172 xenograft growth ( Figure 5B). At Day-35, all tumors were isolated and weighted individually. XL388-treated A172 xenografts were significantly lighter than the vehicle tumors ( Figure 5C). Thus, XL388 oral administration inhibited A172 xenograft growth in SCID mice. The mice body weights were not significantly different between the vehicle group and XL388 treatment group ( Figure 5D).
At treatment Day-7, two hours afterXL388 or vehicle administration, two xenografts of each group were isolated and tissue lysates were achieved. Results in Figure 5E and 5F confirmed that phosphorylation of Akt and S6K1 was largely inhibited in XL388-treated tumors, indicating Akt-mTORC1/2 inactivation. Furthermore, MAFG downregulation was detected ( Figure 5E and 5F).

DISCUSSION
Glioma is the most common primary tumor in central never system and is among the most aggressive of all human malignancies [1,2]. Overactivation of Akt-mTOR cascade is frequently detected in human glioma, promoting tumor cell survival, growth, proliferation, motility, angiogenesis and apoptosis-resistance [40][41][42]. Small molecule inhibitors of Akt-mTOR cascade have exhibited favorable preclinical results and entered clinical trials for human glioma [40,41]. However, the limited single-agent activity of rapamycin analogs in several glioma trials [43,44] provides a rationale for further testing other Akt-mTOR inhibitors against human glioma [40,41].

AGING
Here we show that XL388 blocked Akt-mTORC1/2 activation in established and primary human glioma cells. XL388 potently inhibited glioma cell viability and proliferation as well as cell migration, invasion and cell cycle progression. The PI3K-mTOR dual inhibitor induced significant apoptosis activation in glioma cells. Significantly, oral administration of XL388 potently inhibited A172 xenograft growth in SCID mice. These results indicated that XL388 might have important therapeutic value for human glioma.
Although XL388 blocked Akt-mTORC1/2 activation, XL388-induced cytotoxicity in glioma cells is not solely dependent on Akt-mTORC1/2 inhibition. First, XL388 is significantly more potent than other known Akt-mTOR inhibitors (LY294002, perifosine and rapamycin) in killing glioma cells. Second, restoring Akt-mTOR activation, by caAkt1, only partially attenuated XL388-induced glioma cell death. Third, XL388 is yet still cytotoxic and pro-apoptotic in Akt1-KO A172 cells. These results confirmed the coexistence of Akt-mTOR-independent mechanisms responsible for XL388-induced anti-glioma cell activity.
Indeed, we show that MAFG-Nrf2 inhibition could be another mechanism for XL388-induced actions in glioma cells.
Emerging studies have proposed that MAFG could be an important oncogenic gene for tumorigenesis and progression [45]. Liu et al., showed that MAGF is overexpressed in hepatocellular carcinoma (HCC), associated with tumor progression and reduced survival time [45]. Vera-Puente et al., proposed that MAFG is a potential therapeutic target of non-small-cell lung cancer (NSCLC). MAFG silencing increased ROS production to sensitize cancer cells cisplatin [32]. Fang and colleagues demonstrated that BRAF V600E -stabilized MAFG initiated recruitment of a co-repressor complex to CpG island methylator phenotype (CIMP) gene promoters. It will then promote tumorigenesis in colorectal cancer [46]. Conversely, MAFG silencing inhibited CRC cell growth [46].
We showed that XL388 downregulated MAFG and inhibited Nrf2 signaling, causing significant ROS production and oxidative injury in glioma cells. Several Figure 5. XL388 oral administration inhibits A172 xenograft growth in SCID mice. The SCID mice bearing A172 xenografts (n=10 per group) were administrated with vehicle (saline, "Veh") or XL388(5 mg/kg body weight, daily, × 14d), then tumor volumes (in mm 3 ) (A) and mice body weights (in grams) (D) was recorded every seven days for a total of 35 days; The estimated daily tumor growth (in mm 3 per day) was calculated as described (B). At treatment Day-35, all tumors were isolated and individually weighted (C). At treatment Day-7, two hours after initial XL388administration, the xenograft tumors were isolated. Tissue lysates were subjected to Western blotting assays of listed proteins (E and F). *p < 0.05 vs. "Veh" group (A-C).
AGING antioxidants, including NAC, PDTC and AGI-1067, alleviated XL388-induced glioma cell death and apoptosis. Importantly, MAFG expression was unchanged in Akt-KO A172 cells. These results indicated that MAGF downregulation could be an unique action of XL388, which is responsible for the superior anti-glioma cell activity by this compound.
The current in vitro results and subcutaneous xenograft studies could not be directly translated to humans, and thus the efficacy and safety of XL388 against human glioma will definitely need further characterizations. Testing this compound at lower concentrations in an insitu glioma xenograft model is certainly needed in the following studies.
The underlying signaling mechanisms of XL388-induced MAFG downregulation and Nrf2 inhibition warrant additional studies as well.

Cell culture
Cultures of established glioma cell lines, A172 and U251MG, human neuronal HCN-1a cells, as well as the primary human astrocytes and primary human glioma cells ("Pri-1/Pri-2", derived from two primary glioma patients) were described in detail in our previous studies early [12][13][14]47]. The protocols of this study were approved by the Ethics Review Board (ERB) of Shanghai Jiao-Tong University School of Medicine, according to the principles of Declaration of Helsinki.

Quantitative real-time reverse transcriptase polymerase chain reaction (qPCR)
The detailed protocols for qPCR, using SYBR Master Mix and the ABI Prism 7500H Fast Real-Time PCR system, were described early [12,13]. Quantization of target mRNA was through the 2 -ΔCt method. The mRNA primers of humanNrf2, HO1, NQO1 and GAPDH were described previously [48]. The mRNA primers of human MAFG were purchased from Origene (Beijing, China).

Constitutively-active mutant Akt1
The recombinant adenovirus expressing constitutivelyactive Akt1(caAkt1, S473D, with GFP tag) construct was a gift from Dr. Fang at Shanghai Jiao Tong University [30]. caAkt1 adenovirus or the empty vector adenovirus (Ad-GFP) was added to A172 cells. The infected cells expressing GFP were sorted by FACS and stable cells established. Expression of caAkt1 was verified by Western blotting.

Akt1 knockout
A lenti-CRISPR-GFP Akt1-knockout (KO) construct was from Dr. Zhang at Soochow University [29]. A172 cells were cultured into six well plates at 60% confluence, transfected with the Akt1-KO construct. The transfected cells with GFP were sorted by FACS, and stable single cells were established. Akt1 KO was verified by Western blotting.

Colony formation assay
A172 glioma cells (5, 000 cells for each treatment) were resuspended in agar (0.5%, Sigma)-containing complete medium (with 10% FBS), added on the top of 10-cm culture dishes. XL388-contianing medium or the vehicle control medium was renewed every two days for 10 days. Afterwards, A172 colonies were stained and manually counted.

EdU staining assay of cell proliferation
Cells were seeded into six-well plates at 1 × 10 5 cells per well. Following the applied treatments, an EdU (5ethynyl-20-deoxyuridine) Apollo-567 Kit (RiboBio, Guangzhou, China) was utilized to quantify cell proliferation. EdU and DAPI were both added to the cultured cells, and visualized under a fluorescent microscope. EdU ratio (% vs. DAPI) was calculated.

Apoptosis and cell cycle assays
Cell apoptosis was tested by Annexin V FACS, nuclear TUNEL staining and caspase-3/caspase-9 activity AGING assays. The detailed protocols were described in previous studies [49,50]. Propidium Iodide (PI)-FACS assay of cell cycle progression was described early [51].

Cell death detection by trypan blue staining
Following the applied treatment, trypan blue was added to stain the "dead" glioma cells. Cell death percentage was calculated by the automated cell counter (Merck Millipore, Soochow, China), and the Trypan blue ratio was recorded.

Mitochondrial depolarization assay
JC-1, a mitochondrial fluorescence dye, will aggregate in the mitochondrial inner membrane of stressed cells with mitochondrial depolarization, forming green monomers [52]. Glioma cells were seeded into 24-well plates at 50-60% of confluence, and treated with XL388. Afterwards, cells were stained with JC-1 (10 μg/mL, Sigma), washed and tested under a fluorescence spectrofluorometer (F-7000, Hitachi, Japan) at test-wavelength of 488 nm (green). The representative JC-1 images, integrating both green fluorescence (at 488 nm) and red fluorescence (at 625 nm), were presented as well.
In vitro migration and invasion assays As described early [12], for each treatment 3 × 10 4 glioma cells were seeded onto the upper surface of the "Transwell" chambers (12-μm pore size, BD Biosciences, Shanghai, China). The lower compartments were filled with complete medium (with 10% FBS). After incubation for 16h, the non-migrated cells on the upper surfaces were removed, and on the lower surfaces the migrated cells were fixed, stained and counted. To test cell invasion, "Transwell" chambers were coated with Matrigel (Sigma, Shanghai, China).

Western blotting
Western blotting protocol was described in our previous studies [12,53,54]. Note that the same set of lysates were run in sister gels to test different proteins.

Lipid peroxidation
Using a previously-described protocol [55] cellular lipid peroxidation level was analyzed. In brief, A172 cells were seeded (1 × 10 5 cells per well into six-well plates). Following XL388 treatment, a lipid peroxidation assay kit (Abcam, Shanghai, China) was applied to quantitatively measure cellular lipid peroxidation intensity, tested by the thiobarbituric acid reactive (TBAR) concentration through the described protocols [55,56].

ROS detection
Cells were seeded into six-well plates at 1 × 10 5 cells per well. Following the applied treatments, cells were stained with CellROX dye for 30 min under the dark. The red fluorescence (at 625 nm) was detected and representative CellROX images were shown.

Xenograft assay
As previously reported [12,13], the female severe combined immunodeficient (SCID) mice were purchased from The Animal Center of Soochow University (Suzhou, China) and housed under the standard procedures. A172 cells (5×10 6 cells of each mice in 200 µl of Matrigel gel, no serum) were subcutaneously (s.c.) injected to the flanks of the SCID. In two weeks with the volume reaching approximately 100 mm 3 for each tumor ("Day-0"), mice were randomly assigned into two groups with 10 mice per group. Tumor volumes were calculated as described [12,13]. All animal procedures were approved by IACUC of Shanghai Jiao-Tong University School of Medicine.

Statistical analyses
In this study, statistics were calculated by using SPSS 23.0 software (SPSS Co., Chicago, IL). Descriptive statistics including mean and standard deviation (SD) along with one-way ANOVAs were applied to determine significant differences. A Two-tailed unpaired T test (Excel 2013) was utilized to test significance between two treatment groups. P values < 0.05 were considered significant.

AUTHORS CONTRIBUTIONS
All listed authors designed the study, performed the experiments and the statistical analysis, and wrote the manuscript. All authors have read the manuscript and approved the final version.

CONFLICTS OF INTEREST
The authors declare no conflicts of interest.