JCI-20679 was synthesized as described previously (11) and dissolved in dimethyl sulfoxide (DMSO; 13445-45, Nacalai Tesque, Kyoto, Japan). Compound C (044-33751, Wako Pure Chemical Industries, Osaka, Japan), inosine (I4125, Sigma-Aldrich, St. Louis, MO, USA), CCCP (M34152, Thermo Fisher Scientific, Waltham, MN, USA), rotenone (R8875, Sigma-Aldrich) and cyclosporin A (031-24931, Wako Pure Chemical Industries) were dissolved in DMSO.

Wild-type C57BL/6 and BALB/c mice were obtained from Oriental BioService (Kyoto, Japan). The mice were handled in the animal facility of the Bioscience Research Center at Kyoto Pharmaceutical University according to a protocol approved by the Institution Animal Care and Use Committee.

The procedure was performed as described previously (16,17). Briefly, to generate Sleeping-Beauty transposon-mediated de novo glioblastoma, neonatal mice under hypothermia anesthesia were placed in a stereotaxic instrument (51730D, Stoelting Co., Wood Dale, IL, USA). DNA complexed with polyethyleneimine was injected into the lateral cerebral ventricle at 1 µl/min for 2 min using a 10 µl Hamilton syringe with a 30 gauges needle and an automated infusion system (Legato 130, KD Scientific, Holliston, MA, USA). The coordinates for DNA injection were +1.5 AP, 0.7 ML, and −1.5 DV from λ. In vivo-JetPEI (101000040, Polyplus Transfection, New York, NY, USA), which is an in vivo-compatible DNA transfection reagent, was used. The following DNA plasmids were used for glioblastoma induction: PT2/C-Luc//PGK-SB13 (0.2 µg, 20207, Addgene, Watertown, MA, USA), PT/Caggs-NRASV12 (0.4 µg, 20205, Addgene), PT3.5/CMV-EGFRvIII (0.4 µg, 20280, Addgene), and PT2-shP53 (0.4 µg, 124261, Addgene). Tumor formation was confirmed 10 min after injection of D-luciferin (126-05116, Wako) into the abdominal cavity of the 1–3 months old mice using an IVIS Lumina XR imaging system (Summit Pharmaceuticals International, Tokyo, Japan). Mice were sacrificed by using cervical dislocation for glioblastoma tissue collection.

The procedure used to establish primary GSCs was performed as described previously (17). Briefly, tumor tissues were diced with scalpels and digested with accutase (AT-104, Innovative Cell Technologies, San Diego, CA, USA) at 37°C for 20–30 min, and then washed and incubated with serum-free neural stem-cell culture medium consisting of neurobasal medium supplemented with B27, N2 (17504044, 17502001, Gibco/Thermo Fisher Scientific), and 10 ng/ml EGF and bFGF (236-EG-200, 233-FB-025, R&D Systems, Minneapolis, MN, USA) to maintain GSCs as neurospheres. The neurospheres were split into single cells using accutase and maintained in neural stem-cell culture medium. The procedure to establish differentiated glioblastoma cell lines in an adherent culture system was performed as described previously (18). A172 and U251 human adherent glioblastoma cells obtained from Riken BRC (Tsukuba, Japan) were maintained in DMEM supplemented with 10% FBS (175012, HyClone, South Logan, UT, USA) and penicillin-streptomycin (15140-148, Gibco/Thermo Fisher Scientific). All cells were incubated at 37°C in a 5% CO₂ atmosphere.

GSCs were treated with JCI-20679 at the indicated concentrations for 72 h. The cells were harvested, and the proportion of cells in the DNA synthesis phase was detected by using an APC BrdU flow kit (552598, BD Biosciences, San Diego, CA, USA) following the manufacturer's instructions. At least 3,000 cells per specimen were measured and assessed by flow cytometry using a BD LSRFortessa X-20 cell analyzer (BD Biosciences).

Mitochondrial membrane potentials were detected using JC-1 MitoMP Detection kit (MT09, Dojindo Laboratories, Kumamoto, Japan) according to the manufacturer's instructions and assessed by flow cytometry using a BD LSRFortessa X-20 cell analyzer (BD Biosciences). For monitoring the mitochondrial membrane potential, we used the PE and FITC fluorescence as the JC-1 red and JC-1 green fluorescence. The mitochondrial membrane potential was analyzed by using the ratio of the mean of the JC-1 red/green fluorescence signal intensities.

The oxygen consumption rate was measured using an extracellular oxygen consumption assay (ab197243, Abcam, Cambridge, UK) according to the manufacturer's instructions. At least 700,000 cells per specimen were analyzed every minute for 1 h. These data were normalized by cell numbers.

Mitochondrial reactive oxygen species (ROS) were measured using MitoRos 580 dye optimized for detecting ROS in mitochondria (16052, AAT Bioquest, Sunnyvale, CA, USA) according to the manufacturer's instructions, and were assessed by flow cytometry using a BD LSRFortessa X-20 cell analyzer (BD Biosciences). At least 10,000 cells per specimen were measured.

The procedure was carried out as described previously (19). Briefly, cellular ATP was measured using a CellTiter-Glo luminescent cell viability assay kit, and cellular AMP was measured using an AMP-Glo assay kit (G7570, V5011, Promega, Madison, WI, USA), and then the AMP/ATP ratio was calculated.

The procedure was performed as described previously (17). Briefly, RNAi clones targeting AMPKβ were obtained from Sigma-Aldrich. Transduction using lentiviruses with non-targeting shRNA (SHC016, Sigma-Aldrich, Sequence: CCGGGCGCGATAGCGCTAATAATTTCTCGAGAAATTATTAGCGCTATCGCGCTTTTT) or shRNA targeting AMPKβ was performed at a multiplicity of infection of 10. The following shRNAs were used: sh-AMPKβ #1 (TRCN0000025105, Sequence: CCGGCCCTCCTCTACAAGCCGATATCTCGAGATATCGGCTTGTAGAGGAGGGTTTTT), sh-AMPKβ #2 (TRCN0000274638, Sequence: CCGGCCATGATCCTTCTGAGCCAATCTCGAGATTGGCTCAGAAGGATCATGGTTTTT). The cells were maintained with 0.75 µg/ml puromycin (164-23154, Wako Pure Chemical Industries).

To extract proteins, cells were lysed by 1% SDS buffer supplemented with a protease inhibitor cocktail for use with Mammalian Cell and Tissue Extracts (25955-11, Nacalai Tesque) and PhosSTOP EASYpack (04 906 845 001, Roche Diagnostics, Indianapolis, IN, USA). Total proteins were separated by SDS-PAGE and transferred to PVDF membranes (IPVH00010, Millipore, Billerica, MA, USA). The membranes were blocked with 3–5% fat-free dried milk or 5% BSA in Tris-buffered saline (TBS; 50 mM Tris, 2.68 mM KCl, 137 mM NaCl, PH 7.4) with 0.05% Tween20 (TBST) or Blocking One-P (05999-84, Nacalai Tesque), and then incubated with primary and secondary antibodies. The chemiluminescence of proteins was detected using a ChemiDoc XRS Plus system (Bio-Rad) through the reaction with Clarity Western ECL substrate (170-5060, Bio-Rad Laboratories, Hercules, CA, USA) or ChemiLumi One Super substrate (02230-30, Nacalai Tesque). The following antibodies were used; p-AMPKα (1:1,000; 2535, Cell Signaling Technology, Danvers, MA, USA), AMPKα (1:1,000; 5832, Cell Signaling Technology), GAPDH (1:1,000; 016-25523, Wako Pure Chemical Industries), p-CAMKII (1:1,000; 12716, Cell Signaling Technology), CAMKII (1:1,000; 4436, Cell Signaling Technology), NFATc2 (1:1,000; 5861, Cell Signaling Technology), Vinculin (1:2,000; 66305-1-Ig, Proteintech), Lamin A/C (1:1,000; 2032, Cell Signaling Technology), horseradish peroxidase-conjugated horse anti-mouse IgG (1:2,000; PI-2000, Vector Laboratories, Burlingame, CA, USA), and HRP-conjugated goat anti-rabbit IgG (1:2,000; 7074, Cell Signaling Technology).

GSCs were treated with JCI-20679 at the indicated concentrations for the indicated durations, and then cellular proteins were fractionated into nuclear and cytoplasmic fractions using a LysoPure nuclear and cytoplasmic extractor kit (295-73901, Wako) according to the manufacturer's protocol.

The differentiated glioblastoma cells were treated with JCI-20679 at the indicated concentrations for 72 h. The cells were harvested, and calcineurin phosphatase activity in the same number of the cells was measured using a calcineurin cellular activity assay kit (BML-AK816, Enzo Life Sciences, Farmingdale, NY, USA) following the manufacturer's instructions. At least 10,000 cells per specimen were measured.

Cells were lysed with Trizol (15596026, Thermo Fisher Scientific) to extract total RNA. The RNAs were purified with an RNeasy mini kit (74106, Qiagen, Hilden, Germany). Quantitative PCR analysis was performed with THUNDERBIRD SYBR qPCR mix (QPS-201, TOYOBO, Osaka, Japan) using cDNA synthesized from the extracted RNAs using ReverTra Ace qPCR RT master mix with gDNA (genome DNA) remover (FSQ-301, TOYOBO) and a Light Cycler 96 system (Roche). A standard ΔCt method which had been reported previously (20) was used for data analysis. Gene expression levels were normalized to mGAPDH. Primers were obtained from Eurofins Genomics (Tokyo, Japan). The following primers were used: mNFATc2 primer #1 (F: GTGCAGCTCCACGGCTACAT, R: GCGGCTTAAGGATCCTCTCA), mNFATc2 primer #2 (F: GAGAAGACTACAGATGGGCAG, R: ACTGGGTGGTAGGTAAAGTG), mGAPDH (F: GGTGTGAACGGATTTGGCCGTATTG, R: CCGTTGAATTTGCCGTGAGTGGAGT).

NFATc2 overexpression was performed using a mouse neural stem-cell nucleofector kit (VPG-1004; Lonza, Tokyo, Japan) and a nucleofector 2b device (AAB-1001, Lonza) with an A-033 program optimized for mouse GSCs. The NFATc2 gene (MR209524, Origene Technology, Rockville, USA) or empty vector (PS100001, Origene) was introduced into mouse GSCs, and cells containing the NFATc2 gene or empty vector were selected using 50–250 µg/ml G418 (09380-86, Nacalai Tesque) for 2 weeks. The procedure was performed following the manufacturer's instructions. The nucleofection efficacy was confirmed by western blot analysis.

GSCs were injected into 20 sex-matched mice (6 weeks old; n=10 in each treatment group) for the analyzing the event-free survival rate, and GSCs of a deferent line were injected into 14 sex-matched mice (6 weeks old; n=7 in each treatment group) for the assessing the glioblastoma size. Cells (1×10³) were suspended in 2 µl PBS (166-23555, Wako Pure Chemical Industries). Mice were anchored to the stereotaxic instrument. Cells were injected at a speed of 1 µl/min with an infusion system (Legato 130). The syringe was withdrawn from the brain 2 min after the injection. JCI-20679 was injected intraperitoneally at 20 mg/kg three times a week. JCI-20679 was dissolved at 10 mg/ml in DMSO with Kolliphor EL (C5135, Sigma-Aldrich) (final 10%) and then diluted in saline (total 200 µl/20 g body weight) to the required concentration. DMSO was used as a control. The occurrence of death, obvious neurological symptoms, or a decrease in body weight of more than 2 g was used to generate event-free survival curves. The JCI-20679 treatment was started after 3 days from the GSCs transplantation to analyze the event-free survival rate. Tumor formation was analyzed every week 10 min after injection of D-luciferin (126-05116, Wako) into the abdominal cavity of mice using an IVIS Lumina XR imaging system. For comparison of bioluminescence intensity, tumor initiation (1×10⁴−5×10⁵ photons/sec) was confirmed and then randomized before starting treatment. Mice with neurological symptoms by the tumor formation, such as slow moving or abnormal posture, were sacrificed by using cervical dislocation.

Two-tailed unpaired Student's t-test, two-tailed Welch's t-test, one-way ANOVA with Dunnett's multiple comparison test or 2-way ANOVA with Bonferroni's multiple comparison test were used to determine the statistical significance for in vitro studies. Data are displayed as the mean ± SD unless otherwise indicated. All analyzed data were obtained from at least 3 independent experiments. For the in vivo study, log-rank tests were used to determine significant differences in the survival curves by the Kaplan-Meier method. Bioluminescence intensities were compared by Wilcoxon rank sum test. P-values were calculated using Microsoft Excel software. P<0.05 was considered statistically significant.

First, we analyzed the effects of JCI-20679 on the proliferation of GSCs maintained as neurospheres and differentiation-induced cells in adherent culture; both cell types were derived from identical murine glioblastoma tissues (Fig. 1A). The properties of GSCs and differentiation-induced cells have been established and described previously (17). We found that the proliferation of GSCs was more effectively suppressed by JCI-20679 than the proliferation of differentiated cells (Fig. 1B). JCI-20679 suppressed the proliferation of GSCs in a concentration-dependent manner (Fig. 1C). JCI-20679 treatment did not increase the number of trypan blue-positive dead cells (data not shown). In addition, the BrdU incorporation assay showed that JCI-20679 significantly reduced the number of cells entering the DNA synthesis phase of the cell cycle (Fig. 1D, E). However, we observed no significant induction of apoptosis by the Annexin-V assay and sub-G1 population by the cell cycle analysis (data not shown). These results suggest that JCI-20679 suppresses GSCs proliferation by inducing not apoptosis but cell cycle arrest.

A previous study showed that JCI-20679 inhibits the function of the bovine mitochondria complex I in vitro (15). We next investigated the effects of JCI-20679 on mitochondrial function in cancer cells. It was difficult to detect the mitochondrial membrane potential, which directly reflects mitochondrial function, in GSCs so we used the differentiated glioblastoma cells and A172 human glioblastoma cells as models. JCI-20679 significantly reduced the mitochondrial membrane potential, as did the established mitochondrial inhibitor CCCP, which was used as a positive control (Fig. S1A, B). We next measured the oxygen consumption rate of GSCs. JCI-20679 inhibited the oxygen consumption rate, which is an index of oxidative phosphorylation. Rotenone, an established mitochondrial complex I inhibitor, was used as a positive control (Fig. 2A). Moreover, we found that JCI-20679 increased mitochondrial ROS generation in the differentiated glioblastoma cells in a concentration-dependent manner (Fig. 2B). These results indicate that JCI-20679 inhibits mitochondrial function in cancer cells.

Inhibition of mitochondrial function activates AMPK through its phosphorylation (21), so we evaluated p-AMPKα protein levels in GSCs treated with JCI-20679. Our results showed that JCI-20679 concentration-dependently increased p-AMPKα protein levels (Fig. 3A). Consistent with these results, treatment of GSCs with JCI-20679 decreased intracellular ATP levels and increased intracellular AMP levels, resulting in a significant increase in the AMP/ATP ratio (Fig. 3B-D). Co-treatment with compound C or inosine, which are known AMPK inhibitors, significantly reversed the JCI-20679-mediated suppression of cell proliferation (Fig. 3E, F). These results suggest that JCI-20679 activates AMPK to suppress GSC proliferation.

Inhibition of mitochondrial function modulates the calcium signaling pathway (22), so we investigated the effect of JCI-20679 on factors involved in the calcium signaling pathway. Our results showed that JCI-20679 significantly reduced the protein levels of p-CaMKII and NFATc2 in GSCs (Fig. 4A, B). NFATc2 protein levels decreased in both nuclear and cytoplasmic protein fractions (Fig. 4C). Adequate quality of cellular fractionation was confirmed by western blot analysis (Fig. S1C). Moreover, JCI-20679 reduced calcineurin phosphatase activity (Fig. 4D). These results suggest that JCI-20679 suppresses the calcium signaling pathway.

We next investigated the implication of the JCI-20679-mediated reduction in NFATc2 expression. JCI-20679 decreased NFATc2 mRNA levels in GSCs (Fig. 5A). It has been reported that AMPKα phosphorylation is suppressed by knockdown of AMPKβ, which is crucial for AMPKα phosphorylation (21). Thus, we used AMPKβ shRNA for suppressing AMPK activation. We confirmed suppressing AMPK activation successfully by western blot analysis. In addition, depleted AMPKβ has no effect on expression levels of NFATc2 protein (Fig. S1D). The decrease of NFATc2 mRNA levels mediated by JCI-20679 treatment was recovered when phosphorylation of AMPKα was suppressed (Fig. 5B). Furthermore, the experiment used another shRNA sequence and another NFATc2 primer gave similar result (Fig. S1E). In addition, cell viability of GSCs treated by JCI-20679 were recovered by knockdown of AMPKβ (Fig. 5C). Moreover, we tested the effects of NFATc2 overexpression in GSCs and differentiated glioblastoma cells. We confirmed NFATc2 overexpression by western blot analysis (Fig. 5D). NFATc2 overexpression significantly reversed the JCI-2067-mediated suppression of proliferation (Fig. 5E). These results indicate that the anti-proliferative effects of JCI-20679 are mediated by decreased NFATc2 expression levels that are caused by AMPK activation.

To examine the antitumor effects of JCI-20679 in vivo, JCI-20679 was administered intraperitoneally (20 mg/kg three times a week) to a mouse model orthotopically transplanted with GSCs. JCI-20679 significantly improved event-free survival (Fig. 6A). We confirmed the efficacy of JCI-20679 using orthotopically transplanted with a different GSC line. JCI-20679 significantly decreased the luminescent signal, which reflects glioblastoma size, 3 weeks after JCI-20679 treatment was initiated (Fig. 6B, C). These results indicate that JCI-20679 suppresses glioblastoma proliferation in vivo.