The JNK signaling pathway plays a key role in methuosis (non-apoptotic cell death) induced by MOMIPP in glioblastoma
- 294 Downloads
Synthetic indolyl- pyridinyl- propenones (IPPs) induce methuosis, a form of non-apoptotic cell death, in glioblastoma and other cancer cell lines. Methuosis is characterized by accumulation of cytoplasmic vacuoles derived from macropinosomes and late endosomes, followed by metabolic failure and rupture of the plasma membrane. However, not all IPPs that cause vacuolization are cytotoxic. The main goals of the present study were to identify key signaling pathways that contribute to methuosis induced by cytotoxic IPPs and to evaluate the anti-tumor potential of a prototype IPP in vivo.
We utilized metabolic flux analysis, glucose uptake, immunoblotting, and selective pharmacological inhibitors to compare the effects of closely related cytotoxic and non-cytotoxic IPPs in cultured glioblastoma cells. To determine whether the use of methuosis-inducing IPPs might be feasible in a therapeutic context, we quantified the distribution of our lead IPP compound, MOMIPP, in mouse plasma and brain, and tested its ability to inhibit tumor growth in an intracerebral glioblastoma xenograft model.
The cytotoxic IPP compound, MOMIPP, causes early disruptions of glucose uptake and glycolytic metabolism. Coincident with these metabolic changes, MOMIPP selectively activates the JNK1/2 stress kinase pathway, resulting in phosphorylation of c-Jun, Bcl-2 and Bcl-xL. At the same concentration, the non-cytotoxic analog, MOPIPP, does not activate these pathways. Pharmacologic inhibition of JNK activity promotes survival, even when cells are extensively vacuolated, but suppression of c-Jun transcriptional activity offers no protection. MOMIPP readily penetrates the blood-brain barrier and is moderately effective in suppressing progression of intracerebral glioblastoma xenografts.
The results suggest that interference with glucose uptake and induction of JNK-mediated phosphorylation of pro-survival members of the Bcl-2 family represent key events in the methuosis death process. In addition to providing new insights into the underlying molecular mechanism of methuosis, the results indicate that compounds of the cytotoxic IPP class may have potential for further development as therapeutic agents for brain tumors.
KeywordsGlioblastoma Methuosis Cell death Chalcones C-Jun N-terminal kinase Vacuoles Endosomes
Dulbecco’s modified Eagle medium, high glucose formulation
Fetal bovine serum
Indolyl- pyridinyl- propenone
c-Jun N-terminal kinase
Polyacrylamide gel electrophoresis
Phosphate-buffered normal saline
Sodium dodecyl sulfate
Cancer cells often harbor genetic mutations that diminish the expression or function of proteins that regulate apoptosis (e.g., PTEN, p53, Rb) [1, 2]. Consequently, tumors may exhibit reduced sensitivity to therapeutic agents that trigger this classical form of programmed cell death. This limitation has spurred interest in identifying alternative modes of cell death that can be exploited for cancer therapy. Methuosis is one of the most recent additions to the growing array of non-apoptotic cell death mechanisms [3, 4]. Cells undergoing methuosis initially exhibit massive vacuolization of macropinosomes and late endosomes. Our recent studies indicate that this is due to defective trafficking of late endosomes to lysosomes, with concomitant homotypic fusion of the affected vesicular compartments . The defect in lysosome-directed trafficking also affects autophagic flux, with resultant accumulation of autophagosomes . Ultimately, the integrity of the cell membrane is compromised and the vacuolated cells rupture in a manner reminiscent of necrosis. These morphological features are distinct from apoptosis and other non-apoptotic cell death processes, and they cannot be prevented by caspase inhibitors or agents that block necroptosis or autophagy .
Methuosis was first described in glioblastoma cells induced to over-express activated forms of the Ras GTPase [6, 7]. Subsequently, our group discovered a series of synthetic chalcones (indolyl-pyridinyl-propenones; IPPs) that can trigger methuosis, independent of constitutive Ras activation [8, 9]. The prototype compound, 3-(5-methoxy-2-methyl-1H-indol-3-yl)-1-(4-pyridinyl)-2-propene-1-one (abbreviated MOMIPP) induces methuosis in glioblastoma (GBM) and other cancer cell lines at low- micromolar concentrations. A recent study provided the first identification of a specific molecular target for MOMIPP; i.e., the phosphatidylinositol-3-phosphate 5-kinase (PIKfyve) . The product of PIKfyve, PI(3,5)P2, is known to play a critical role in late endosome trafficking [11, 12]. Since our initial description of methuosis, a number of other reports have noted similar cell death phenotypes promoted by a variety of chemical agents and natural products [13, 14, 15]. Features of methuosis have also been described in cells responding to overexpression of miR-199a-3p , co-expression of mutant EGFR and K-Ras , immunotargeting of CD99 , treatment with an oligonucleotide aptamer , or NGF-stimulation of TrkA .
Despite the growing recognition of the morphological hallmarks of methuosis, the specific molecular mechanisms that link vacuolization of endocytic compartments to loss of cell viability remain poorly understood. Our structure-activity studies of MOMIPP and numerous analogs in GBM cells have provided valuable chemical tools to address this question. Specifically, we found that minor structural modifications of the indole ring yielded a functionally distinct sub-group of IPPs that retained the ability to induce robust morphological vacuolization, with greatly reduced cytotoxicity [21, 22]. By comparing the effects of MOMIPP with one of the non-lethal analogs (MOPIPP; with propyl substituted for methyl at the 2-position of the indole ring), we noted that cells treated with MOMIPP had more severe inhibition of endolysosomal degradation pathways for EGF and LDL receptors . Coincidentally, MOMIPP shows stronger binding affinity (lower Kd) for PIKfyve than the non-lethal analogs , despite the fact that the cells treated with these compounds have similar vacuolated morphologies.
In the present study, the objective was to expand the comparative analysis of cytotoxic versus non-cytotoxic vacuole-inducing IPPs in GBM cells, with the goal of defining pathways essential for triggering cell death. The results indicate that early impairment of glucose uptake and glycolytic metabolism, with attendant activation of JNK signaling and Bcl-2 phosphorylation, are key elements in the methuosis death program.
Human glioblastoma cell lines, U251 (deposited by Darrell Bigner), SF295 (deposited by Paul Kornblith), and SNB19 and SNB75 (deposited by M.L. Rosenblum), were obtained from the Developmental Therapeutics Program (DTP) Tumor Repository, NCI Division of Cancer Treatment and Diagnosis (DCTD) (operated by Charles River Laboratories for the National Cancer Institute, Frederick, MD). The A172 (Cat. No. CRL-1620), LN229 (Cat No. CRL-2611), T98G (Cat No.CRL-1690), and U87MG (Cat No. HTB-14) cell lines were purchased from the American Type Culture Collection (Manassas, VA). Normal human skin fibroblasts were originally derived from a skin biopsy as described previously . All cell lines were maintained in Dulbecco’s modified Eagle medium (DMEM; ThermoFisher, Walthham, MA), supplemented with 10% (v/v) fetal bovine serum (FBS) (JR Scientific, Woodland, CA) at 37 °C with 5% CO2/95% air. Cell lines were confirmed negative for Mycoplasma contamination by periodic staining with DAPI or use of the PlasmoTest assay (InvivoGen, San Diego, CA).
Chemicals and antibodies
Indolyl chalcone compounds; MOMIPP , MOPIPP , 2a and 2q  were synthesized as described previously. SP600125 and YM201636 were obtained from Cayman Chemical, Ann Arbor, MI. Bafilomycin-A1, n-methylpyrrolidone (NMP), 5-(N-Ethyl-N-isopropyl)amiloride (EIPA), oligomycin, phloretin, cytochalasin B, carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP) and Solutol-HS15 were from Sigma-Aldrich, St. Louis, MO. XenoLight D-Luciferin-K+ salt was obtained from Perkin Elmer, Waltham, MA.
Immunoblots and antibodies
U251 cells were typically seeded at 1 × 106 cells per 10 cm dish and maintained for 24 h in DMEM containing 10% FBS. On the day after plating, test compounds (or an equivalent volume of DMSO) were added in fresh medium and cells were harvested after the indicated times. Cells were lysed in SDS sample buffer (65 mM Tris-HCl, 10% glycerol, 0.5% β-mercaptoethanol, 2% SDS, pH 6.8) and the protein concentration was determined by colorimetric assay (Bio-Rad, Hercules, CA). Equal amounts of protein (80 μg) were subjected to SDS-PAGE, transferred to PVDF membrane. Immunoblot procedures were described previously . Chemiluminescent signals were quantified using an Alpha Innotech FluorChem HD2 imaging system with Alpha View software.
Antibodies against the following proteins were obtained from Cell Signaling Technology (Danvers, MA): phospho JNK 1/2 (Thr183/Tyr185) (Cat. No. 9251), JNK 1/2 (Cat. No. 9252), phospho-p38 MAPK (Thr180/Tyr182) (Cat. No. 9215), p38 MAPK (Cat. No. 9212S), phospho-Bcl-2 (Ser70) (Cat. No. 2827), Bcl-xL (Cat. No. 2764), phospho-c-Jun (Ser63) (Cat. No. 9261), c-Jun (Cat. No. 9165), phospho-SEK1/MKK4 (Cat. No. 9156), SEK1/MKK4 (Cat. No. 9152), and BAX (Cat. No. 5023). Antibodies against Bcl-2 (Cat. No. sc-7382) and the HA-epitope tag (Cat. No. sc-57,592) were from Santa Cruz Biotechnology, Dallas, TX. The antibody against phospho-Bcl-xL (Ser62) (Cat. No. PA535496) was from ThermoFisher, and the antibody against α-tubulin (Cat. No. T5168) was from Sigma. HRP- coupled goat anti-mouse IgG (Cat. No. 554002) and goat anti-rabbit IgG (Cat. No. 554021) were obtained from BD Biosciences (San Jose, CA).
Phase contrast microscopy
For live cell imaging studies, 1 × 105 U251 cells were seeded in 35-mm dishes and incubated for 24 h before commencing treatment with compounds. Cells were examined by phase-contrast microscopy using an Olympus IX70 inverted microscope equipped with a heated stage, a DP80 digital camera, and cellSens™ software (Olympus America, Center Valley, PA).
For immunofluorescence localization of Glut1, cells were grown on coverslips, treated with MOMIPP or DMSO for 4 h, washed with PBS, and then fixed in ice-cold methanol. After pre-blocking with 10% goat serum, cells were incubated with anti-Glut1 polyclonal antibody (Millipore, Cat. No. 07–1401) at 1/50 dilution for 1 h, followed by goat-anti-rabbit IgG, conjugated to AlexaFluor-488 (ThermoFisher) at 1/600 dilution. Fluorescence images were captured with an Olympus IX70 inverted fluorescence microscope, using a DP80 digital camera, and cellSens™ software (Olympus America).
Viability of cells in vitro was assessed by measuring cellular ATP using the CellTiter Glo® luminescence assay according to the manufacturer’s protocol (Promega Corp., Madison, WI). U251 cells were seeded in white-walled opaque 96-well plates (2000 cells/well) with four replicate wells for each culture condition. After addition of compounds at the indicated concentrations, cell viability was assayed at a 48 h endpoint. Luminescence was quantified with a Berthold Tech Centro XS3 LB 960 luminometer, using the preinstalled MikroWin software.
Metabolic flux analysis
Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were determined using a Seahorse XFP analyzer, following standard protocols provided by the manufacturer (Agilent Technologies, Santa Clara, CA). Cell plating densities and concentrations of oligomycin and FCCP were optimized according to the recommended protocols. U251 cells were seeded at 15,000 cells per well in DMEM + 10% FBS in Seahorse miniplates. On the day after seeding, cells in parallel wells were changed over to fresh medium containing either 10 μM MOMIPP or 10 μM MOPIPP, with controls containing equivalent amounts of DMSO. Five hours after addition of the compounds, (or DMSO for controls), the cells were washed and switched to the recommended serum-free base medium for each type of metabolic flux assay. After insertion of the pre-hydrated sensor cartridges, the mini-plates were placed in the Seahorse instrument. OCR was measured under basal conditions (+glucose), and after sequential addition of oligomycin and FCCP. In separate assays, ECAR was measure under basal (glucose-free) conditions, and after sequential addition of glucose and oligomycin.
2-deoxy-D-glucose (2-DG) uptake
2-deoxy-D-glucose [1,2-3H(N)] (Cat. No. NET549A) was purchased from Perkin Elmer (Waltham, MA). For cells incubated with IPPs for 24 h, cellular uptake of 2-DG was determined essentially as described by Wood et al. . Briefly, U251 cells were plated in 6 cm dishes at 2.5 × 105 cells per dish in standard medium. On the next day, fresh medium was added with the indicated compounds and cells were incubated for 24 h. After removal of medium and washing the cells, the cultures were incubated for 5 min at 37o C in HEPES-buffered saline solution (20 mM HEPES, 140 mM NaCl, 2.5 mM MgSO4, 1 mM CaCl2, 5 mM KCl, pH 7.4) containing 1 μCi [3H]2-DG and 10 μM unlabeled 2-DG. Parallel dishes were incubated under the same conditions, except that 10 μM cytochalasin B was added to block glucose transporter activity. Reactions were stopped by addition of ice-cold 0.9% NaCl, and after three washes, the cells were harvested in 1 ml lysis buffer: 20 mM Tris, 135 mM NaCl, 1 mM MgCl2, 1% Triton-X100, pH 8.0. Tritium in the lysates was quantified in a liquid scintillation counter and values for treated vs. control cultures were compared after subtraction of background from the cytochalasin B-treated cultures. To correct for possible differences in cell density after 24 h drug exposure, a portion of each cell lysate was utilized for quantification of α-tubulin by western blot analysis, and the tritium counts were normalized to the tubulin values before being compared to the controls. For short-term drug exposures (2–4 h), differences in cell density were not a factor, so the net tritium uptake values in control and treated cultures were compared directly without normalization to tubulin.
[3H]2-DG uptake in freshly isolated mouse erythrocytes was assayed essentially as described by Ohmori et al. . Briefly, erythrocytes were pretreated for 10 min at 37o C in HEPES-buffered saline solution containing 10 μM MOMIPP, 200 μM phloretin + 1 mM HgCl2, 100 μM cytochalasin B or DMSO equivalent to the amount added with MOMIPP. Then labeling solution containing 1 μCi [3H]2-DG was added and incubation was continued for 5 min. Ice-cold stop solution was added to arrest the reaction and the erythrocytes were pelleted by centrifugation at 4o C. The final pellets were washed twice with stop solution, dissolved in Solvable® (Perkin Elmer), treated with H2O2 and then subjected to liquid scintillation counting.
86Rb (Cat. No. NEZ072001) was purchased from Perkin Elmer. Ouabain-sensitive 86Rb uptake was measured as described by Galuska et al. . U251 cells were plated in parallel wells of a 12-well dish at 50,000 cells/well. On the next day, fresh medium (DMEM + 10% FBS) was added, with 10 μM MOMIPP or an equivalent volume of DMSO. After 4 h the cells were washed and pre-incubated for 10 min in serum-free DMEM, with or without 1 mM ouabain. Finally, 1 μCi of 86Rb was added to each culture and incubation was continued at 37 °C for an additional 10 min. Reactions were terminated by washing the cells with ice-cold 100 mM MgCl2 and cell lysates were prepared for liquid scintillation counting. The 86Rb counts in parallel cultures with ouabain were subtracted from those without ouabain to obtain the net ouabain-sensitive 86Rb uptake.
C-Jun dominant negative construct
The c-Jun dominant-negative (DN) construct, pMIEG3-Jun DN , was provided by Alexander Dent (plasmid # 40350; Addgene, Cambridge, MA). The c-JunDN cDNA, which lacks the sequence encoding the first 122 amino acids of c-Jun, was amplified by PCR, using a forward primer: 5′- AAA AAA GGT ACC ATG ACT AGC CAG AAC ACG CTG CCC AGC GTC-3′ to add a KpnI site and a reverse primer: 5‘-AAA AAA-TCT AGA-TCA GCT GGC ATA GTC AGG CAC GTC ATA AGG ATA GCT AAA TGT TTG CAA C-3’ to add an XbaI site. The resulting PCR product was ligated into the pTRE-Tight vector (Clontech, Mountain View, CA), to generate pTRE-Tight (c-JunDN). The latter was combined with pTK-Hyg (Clontech) and introduced into the U251-Tet-On cell line, which stably expresses the reverse tetracycline-controlled transactivator (rtTA) , using a Nucleofector™ device and kit T reagents from Lonza (Walkersville, MD). The cells were selected in DMEM containing 10% FBS and 200 μg/ml of both G418 and hygromycin. Single colonies were picked, expanded and tested by western blot analysis for expression of HA-tagged c-JunDN in response to addition of 1 μg/ml doxycycline to the medium.
RT-PCR assessment of gene expression
Total RNA was extracted and purified from cultured cells using RNeasy Mini kit, following the manufacturer’s protocol (SA Biosciences/Qiagen, Germantown, MD). cDNA was generated by reverse transcription of 100 ng of total RNA with the RT2 first strand kit (Qiagen). RNA and cDNA were quantified and checked for purity (OD 260/280) using a Nano-Drop-1000 spectrophotometer (Thermo Fisher). RT-PCR reactions were carried out using primers for SERPINA3 and GFAP from ThermoFisher and GAPDH primers from Qiagen, using an Applied Biosystems StepOne Plus™ system. After 40 cycles, the ΔCT values for SERPINA3 and GFAP versus the GAPDH standard were calculated. ΔΔCT values indicate the difference between the average ΔCT values for the MOMIPP-treated cells and the controls. The expression fold change was determined as 2-ΔΔCt.
Turnover and brain uptake of MOMIPP in vivo
All animal studies were performed in compliance with United States Public Health Service Policy on Humane Care and Use of Laboratory Animals, under protocols approved by University of Toledo Institutional Animal Care and Use Committee (Reference Number 107491).
Swiss Webster mice (8–10 weeks, female) were obtained from Charles River Laboratories and housed in ventilated cages on a 12 h light-dark cycle. MOMIPP was dissolved in NSP (10% NMP, 15% Solutol-HS15 in PBS) and administered to mice via intraperitoneal (IP) injection at a dose of 80 mg/kg. For pharmacokinetic studies, 3 mice were used for each time point. Mice were euthanized by CO2 asphyxiation followed by cervical dislocation at the indicated times after drug injection. Blood was collected by cardiac puncture, held on ice in lithium heparin-coated tubes, and centrifuged at 10,000 x g for 20 min. Plasma samples were stored at − 80 °C until analysis. Brains were placed in vials and frozen in liquid N2. Prior to quantification of MOMIPP, the frozen brains were weighed and homogenized 1:9 (w/v) in RIPA buffer (150 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, 1% SDS, 50 mM Tris, pH 7.5).
To determine the concentration of MOMIPP, samples of plasma (200 μl) or brain homogenate (500 μl) were extracted with 1 ml ethyl acetate at 37 °C for 20 min, followed by centrifugation for 2 min at 16,000×g. 800 μl of the resulting extract was vacuum centrifuged at 30 °C for 1 h and the residue was suspended in 100 μl of chromatography elution solution: 30% (v/v) acetonitrile and 0.1% (v/v) formic acid in deionized water. Samples (10 μl) were injected onto a Waters Ascentis Express C18 column (75 × 21 mm, 2.7 μm) with matching guard column on a Waters 2795 HT-Alliance LC Separations Module, and isocratic elution was performed at a flow rate of 0.3 ml/min. MOMIPP was detected via multiple reaction monitoring on a Micromass Quattro Micro Mass Spectrometer (MS) in ESI+ mode with capillary voltage 3.0 kV, source temperature 100 °C, desolvation temperature 400 °C, desolvation gas flow 650 l/h, cone gas flow 40 l/hr., and dwell time 0.2 s. MOMIPP was identified as 293.1 > 95.9 at cone voltage 40 V, collision energy 23 V, and a column retention time of 2.1 min. Nitrogen was used as the carrier gas through the mass spectrometer, while argon was used in the collision chamber.
For plasma standards and quality control calculations, solutions of known amounts of MOMIPP were prepared in methanol and the solution was evaporated at 30 °C for 1 h in a vacuum centrifuge. 200 μl of mouse plasma (Pel-Freez Biologicals, Rogers, AZ) was added to each prepared tube to give six total plasma standard concentrations ranging from 10 to 100,000 nM, and three quality control samples from 30 to 30,000 nM. Brain standards were prepared in the same way, except that 500 μl of brain homogenate from untreated mice was added, and the measured concentrations for brain tissue were corrected for the 10-fold dilution of brain tissue in RIPA buffer during homogenization. Levels of MOMIPP in plasma were based on volume and expressed as nM concentration. Brain homogenate levels were measured based on tissue mass prior to homogenization, so that these concentrations represent nmol/kg tissue. For ease of comparison, tissue densities were approximated as 1.0 g/ml, so that brain levels could also be expressed as nM concentration.
Anti-tumor efficacy study in an orthotopic xenograft model
Athymic CrTac:NCR-Foxn1<nu> mice (female, 7–8 weeks) were purchased from Taconic Biosciences (Rensselaer, NY) and housed under SPF conditions (three per cage) on a 12 h light-dark cycle in ventilated cages with barrier filters. For initiation of xenografts that could be detected by bioluminescence imaging (BLI), we nucleofected U251 cells with pCMV5neo-Luc, which encodes the firefly luciferase gene excised from pGL3 (Promega, Madison, WI). Clonal selection was performed in medium containing G418. The resulting stable cell line (U251-LUC), was maintained in DMEM supplemented with 10% (v/v) FBS and 200 μg/ml G418 and was periodically tested for uniform luciferase by immunofluorescence analysis using a luciferase antibody (Sigma, Cat. No. L2164).
The intracerebral xenografts were established based on the method described by Ozawa and James . Mice were anesthetized by IP injection with a mixture of ketamine (100 mg/kg) and xylazine (7.5 mg/kg) and placed on a warming pad. A 1 cm incision was made down the midline of the scalp. A small hole was drilled in the skull using a 25 g needle at a point 1 mm anterior and 2 mm lateral to the bregma. 4 × 105 cells (suspended in 3 μl of DMEM) were injected into the brain using a 10 μl Hamilton syringe equipped with a 27 g needle, with a cuff placed 3 mm from the tip of the needle to control the depth of injection. The hole in the skull was sealed with sterile bone wax and the skin incision was closed with surgical glue. Post-surgical analgesia was provided with buprenorphine (0.1 mg/kg) administered twice a day for the first two days. Tumor growth was monitored by BLI. All mice had focal BLI in the range of 107–108 photons by the fourth day after tumor cell implantation. Mice were then grouped randomly into control and MOMIPP treatment groups (10 mice for the control and 11 mice for the treatment group; one mouse did not survive after surgery). Sample size was determined using Biostat Power and Precision® software, which indicated that 10 mice per group was sufficient to detect a 25% difference in the mean tumor size between control and treated mice (80% confidence), assuming a standard deviation of 20% in each group. MOMIPP (80 mg/kg, in NSP) or vehicle was administered by IP injection every 24 h for 15 consecutive days. Tumor progression was monitored by BLI on the 7th, 11th, and 15th days after commencement of drug treatment. On the 15th day, all mice were euthanized by CO2 asphyxiation, followed by cervical dislocation. Blood was collected by cardiac puncture and utilized for blood chemistry analysis with a comprehensive diagnostic profile kit and a VetScan VS2 Analyzer (Abaxis, Union City, CA), following protocols recommended by the manufacturer.
GraphPad Prism software was used for statistical analyses. Student’s unpaired t-test was applied to cell culture studies and Mann-Whitney unpaired test was used for in vivo tumor studies. P values ≤ 0.05 were regarded as significant.
Differential effects of cytotoxic versus non-cytotoxic IPP’s on glycolytic metabolism
Cytotoxic IPPs selectively activate the JNK stress kinase pathway
Two MAP kinase kinases, MKK4 and MKK7, mediate phosphorylation of JNK on Tyr and Thr . These kinases can, in turn, be activated by several different upstream kinases (e.g., ASK1, MLK, TAK1). As shown by the western blots in Fig. 4d, MOMIPP and 2q triggered a robust increase in phosphorylation of MKK4. Treatment with MOPIPP and 2a caused a more modest increase. By comparison, neither MOMIPP nor MOPIPP dramatically altered phosphorylation of MKK7 (Fig. 4d). These observations are consistent with the concept that MKK4 responds predominantly to environmental stress, whereas MKK7 is associated with cytokine-mediated pathways [35, 36].
Cytotoxic IPPs trigger increased phosphorylation of Bcl-2 and Bcl-xL
In addition to the c-Jun transcription factor, other known substrates for JNK include pro-survival members of the Bcl-2 family, which play key roles in protecting cells from apoptosis and necrosis [37, 38] and regulating autophagy. Indeed, phosphorylation of Bcl-2 and Bcl-xL by JNK1 is postulated to interfere with their interactions with pro-apoptotic members of the Bcl-2 family (e.g., BAX) and the autophagy regulator, Beclin-1, thereby promoting cell death or autophagy [39, 40, 41]. The results depicted in Fig. 4e indicate that increased phosphorylation of Bcl-2 and Bcl-xL accompanied the activation of JNK in cells treated with the methuosis-inducing compounds, MOMIPP and 2q.
Activation of JNK is an early step in methuosis
To determine if activation of JNK represents an initial step in the evolution of the methuosis phenotype, we examined the phosphorylation of JNK after only 4 h exposure of cells to MOMIPP or MOPIPP. The results show an early increase in phosphorylation of JNK upon treatment with MOMIPP, but not MOPIPP (Fig. 4f). Interestingly, the western blots revealed a time-dependent change in the phosphorylation pattern of the JNK splice variants, with the p54 isoform predominating at 4 h (Fig. 4f), and the p46 isoform predominating at 24 h (Fig. 4c&f). Consistent with the activation of JNK, phosphorylation of c-Jun also increased 4 h after addition of MOMIPP (Additional file 1: Figure S1A). In contrast to JNK, there was no detectable change in phosphorylation of the p38 MAP kinase after 4 h of MOMIPP treatment (Additional file 1: Figure S1B). As in the studies conducted at 24 h, we observed substantial increases in phosphorylation of Bcl-2 and Bcl-xL at 4 h in cells treated with MOMIPP (Additional file 1: Figure S1C). However, we did not detect any change in expression or phosphorylation of the pro-apoptotic Bcl-2 family member, BAD (Additional file 1: Figure S1C).
Relationship between vacuole formation and JNK activation
The observation that cells treated with MOMIPP or MOPIPP have indistinguishable vacuolated morphologies at early time points, while only the cells exposed to MOMIPP show increased JNK phosphorylation, prompted us to further explore the relationship between JNK activation and endosomal vacuolization.
Activation of c-Jun-mediated transcription is not required for methuosis
To determine if c-Jun activation plays a key role in methuosis, U251-c-JunDN cells were treated with MOMIPP, with or without prior induction of c-JunDN. The expression of c-JunDN had no effect on the morphological hallmarks of methuosis (e.g., extreme vacuolization, cell rounding, detachment and lysis) (Fig. 8c), and it did not prevent the loss of cell viability, assessed by the ATP assay (Fig. 8d). Western blot analysis confirmed that expression of c-JunDN did not interfere with stress signaling upstream from c-Jun, since JNK phosphorylation was still increased by MOMIPP (Fig. 8e). Consistent with the preservation of JNK activation in cells expressing c-JunDN, the MOMIPP-induced phosphorylation of Bcl-2 and Bcl-xL was also sustained (Fig. 8e).
Evaluation of MOMIPP pharmacokinetics
The ultimate goal of developing chemical inducers of non-apoptotic cell death in glioblastoma is to employ these compounds as novel therapeutic agents. Therefore, we extended our studies to assess the pharmacokinetic properties of MOMIPP, evaluate its ability to penetrate the blood-brain barrier (BBB), and obtain a preliminary indication of its potential anti-tumor efficacy in a xenograft model.
Effect of MOMIPP on tumor progression
Based on the foregoing studies, we elected to initiate a study to evaluate the effects of MOMIPP on the growth of orthotopic GBM xenografts in immunocompromised mice. The U251 human glioblastoma cell line has been widely used to establish aggressive intracerebral tumors with many of the pathological features of human glioblastoma . To permit monitoring of tumor progression by bioluminescence imaging (BLI), we developed a stable U251 cell line engineered to express firefly luciferase (U251-LUC) (Additional file 2: Figure S2). Four days after intracerebral inoculation of U251-LUC cells, mice were sorted randomly into control and treatment groups (Day-0). At this point BLI indicated no significant difference between the groups (control, 1.14 ± 0.67 × 108 photons; treatment group 1.11 ± 0.45 × 108 photons) (Fig. 9b). Treatment of the mice was then started, with mice in the treatment group receiving daily injections of MOMIPP (80 mg/kg in NSP), while the control group received an equivalent volume of the NSP vehicle. By day-7 the BLI signals in the control group (3.27 ± 0.45 × 108 photons) had tripled compared to day-0, indicating substantial tumor growth. By comparison, tumor progression in the MOMIPP treatment group (1.79 ± 1.72 × 108 photons) was significantly inhibited (Fig. 9b). Subsequent BLI images acquired on day-11 and day-15 (Fig. 9b) confirmed a continued increase in the size of the tumors in the control group, with significant suppression of tumor progression in the MOMIPP-treated group.
The behavior of the MOMIPP-treated mice was not altered during the course of the study. Their weight remained comparable to the controls until day-15, when the controls began to lose weight, prompting termination of the study (Additional file 3: Figure S3). Blood chemistry profiles obtained for all of the control and treated mice at the end of experiment showed no significant differences in key enzymes and metabolites (Additional file 4: Table S1), suggesting that 15-day treatment with high-dose MOMIPP did not cause systemic toxicity or organ failure.
In the present studies, we used MOMIPP and structurally-related indolyl chalcones (IPPs) to explore the mechanism of cytotoxicity underlying the form of non-apoptotic cell death termed ‘methuosis’. The results indicate that although these compounds initially induce morphological phenotypes that appear identical (i.e., massive vacuolization of macropinosomes and endosomes), the cytotoxic compounds differ from the non-lethal compounds in having distinct effects on glycolytic metabolism and the stress-induced JNK signaling pathway. Within the first several hours after vacuoles appear, the prototype methuosis-inducing compound, MOMIPP, caused a substantial reduction in glucose uptake and a decline in glycolytic function. In contrast, the non-cytotoxic analog, MOPIPP did not have major metabolic effects, despite the fact that in previous studies the number of vacuoles per cell was not statistically different between cells treated with MOMIPP or MOPIPP . One possible explanation for these differential effects emerges from the recent identification of the endosomal phosphoinositide kinase, PIKfyve, as a MOMIPP target .
The phosphoinositide products of PIKfyve play key roles in the biogenesis of multivesicular endosomes and the assembly of protein complexes required for vesicular trafficking within the endocytic pathway [11, 48]. Indeed, perturbation of the enzyme is known to cause vacuolization of endocytic compartments , enlargement and fusion of lysosomes , and disruption of autophagic flux . PIKfyve also appears to affect endosome recycling under some circumstances, although this aspect of its function is still not completely understood. For instance, inhibition of PIKfyve disrupts recycling of some junction proteins (claudins) but not others (occludin, E-cadherein) . Of particular relevance for the present study, perturbation of PIKfyve can impair insulin-stimulated glucose uptake and translocation of GLUT4 and GLUT1 from endosomes to the cell surface . It is interesting to note that when Cho et al.  compared several IPPs from our series , they found that the cytotoxic MOMIPP had a significantly greater affinity for PIKfyve than the non-lethal vacuole inducer, compound, 2a. This raises the possibility that despite similar morphological appearances (i.e., number and size of phase-lucent vacuoles), the cells treated with the more potent PIKfyve inhibitors like MOMIPP may experience a more stringent block in some of the vesicular trafficking functions regulated by this kinase. This concept is supported by our recent study showing that MOMIPP causes a more severe disruption of EGF-receptor and LDL-receptor lysosomal degradation than the non-lethal MOPIPP . Similarly, it is conceivable that compared to MOPIPP and other non-cytotoxic IPPs, MOMIPP may cause greater retention of glucose transporters in non-recycling endosomal vacuoles, ultimately causing a decrease in glucose uptake. While this mechanism merits further investigation, it is not clear whether it alone could account for the differential cytotoxicity of the IPP series. For instance, our finding that the known PIKfyve inhibitor, YM201636, had a much smaller effect on [3H]2-DG uptake than MOMIPP (Fig. 6) leaves open the alternative possibility that MOMIPP causes reduced glycolysis and activation of JNK through interactions with targets other than PIKfyve. In this regard, intriguing possibilities include PFKFB3 and PFKFB4 (6-phosphofructo-2-kinase/fructose-2,6-biphosphatases 3 and 4), which play important roles in regulating glycolytic flux, especially in cancer cells [54, 55]. Recently developed antineoplastic inhibitors of PFKFB3 (3PO and PFK15) have chalcone-like structures reminiscent of MOMIPP, with pyridinyl groups as key elements [56, 57]. Therefore, in the future it will be interesting to determine if MOMIPP can directly inhibit PFKFB3.
The results from this study demonstrate that the JNK signaling axis plays a critical role in methuosis, a unique form of cell death induced by MOMIPP and other cytotoxic IPPs. Stress-induced activation of the JNK signaling pathway is widely recognized as a stimulus that can lead to apoptosis [35, 58]. However, previous studies from our group and others have established that methuosis is distinct from apoptosis, based on absence of classical apoptotic hallmarks (e.g., chromatin condensation, DNA laddering, cell shrinkage) and insensitivity to caspase inhibitors [3, 8]. JNK has also has been implicated in regulation of necroptosis, a form of programmed necrosis . But, unlike necroptosis, methuosis is not attenuated by the RIP1 kinase inhibitor, necrostatin-1 (Additional file 5: Figure S4). Thus, methuosis can be added to the growing list of non-apoptotic cell death pathways in which JNK plays a pivotal role .
Phosphorylation by JNK increases the activity of c-Jun and other nuclear transcription factors. However, the results of our studies utilizing the c-JunDN construct suggest that MOMIPP-induced cell death does not depend on changes in gene expression mediated by activation of c-Jun. Rather, it appears that post-translational modification of cytoplasmic JNK substrates represents the most likely trigger for cell death. Among the many known JNK targets in the cytoplasm, we focused on Bcl-2 and Bcl-xL, because multi-site phosphorylation of Bcl-2 (T56, S70,T74,S87) and phosphorylation of Bcl-xL (S62) by JNK are thought to disrupt the pro-survival functions of these proteins [39, 59, 60]. The precise mechanism through which Bcl-2 phosphorylation promotes cell death is still unclear, and it may vary depending on the cell type and the stimulus . Possibilities include: 1) disruption of interactions with pro-apoptotic proteins like Bax, facilitating outer mitochondrial membrane pore formation and release of cytochrome c ; 2) alteration of Bcl-2 function in the endoplasmic reticulum, causing discharge of Ca2+, with consequential mitochondrial calcium overload ; 3) release of Beclin-1 from inhibitory Beclin-1/Bcl-2 complexes, causing increased autophagy ; or 4) loss of mitochondrial membrane potential and release of proteins like AIF, which are capable of inducing caspase-independent cell death [64, 65]. The latter mechanism merits serious consideration, since methuosis is not blocked by caspase inhibitors [8, 15]. Finally, it should be noted that activation of JNK by MOMIPP could have pleiotropic effects beyond phosphorylation of Bcl-2/Bcl-xL. For example, JNK activation is known to amplify production of reactive oxygen species , which has the potential to compromise metabolic functions and trigger necrotic cell death .
The molecular events linking endosomal vacuolization and changes in glycolytic metabolism to the activation of JNK remain to be defined. Two upstream kinases, MKK4 (SEK1) and MKK7, are known to phosphorylate JNK, with MKK7 mediating signals from proinflammatory cytokines (e.g., TNF) and MKK4 serving as a sensor for environmental stress [35, 36]. Previous studies have suggested that glucose deprivation may modulate MKK4 by altering the crosstalk between two JNK-interacting scaffold proteins, JIP1 and JIP3, resulting in activation of ASK1, a kinase that targets MKK4 . Therefore, it is conceivable that the JIP1/3 → ASK1 → MKK4 signaling module may serve as a conduit to activate JNK when endosomal trafficking is disrupted and glucose levels fall during the early stages of methuosis. Our finding that MKK4 was robustly activated by the methuosis-inducing IPPs, MOMIPP and 2q, and less so by the non-cytotoxic vacuole-inducers, MOPIPP and 2a, supports this concept.
Thus far, evaluation of the anticancer activity of methuosis-inducing compounds has been confined mainly to cell culture systems. However, two recent studies have highlighted the potential for exploiting this novel form of cell death for treating cancers in vivo. In one study Huang et al.  identified a unique 4′6’-disubstituted aza-indole that selectively induced methuosis in a broad panel of cancer cell lines in vitro and suppressed the growth of subcutaneous MDA-MB-231 breast cancer xenografts in immunocompromised mice. In a separate study, Ahlstedt et al.  found that a quinolone-based methuosis inducer, Vacquinol-1, reduced the size of brain tumors in syngeneic rat models, although no survival advantage was noted. In the present study, we found that MOMIPP readily penetrates the BBB and significantly suppresses the progression of intracerebral GBM xenografts without overt toxicity in nude mice. However, overall growth suppression was modest, and high doses of MOMIPP were required to compensate for the rapid clearance of the compound from the circulation. Even with relatively high daily doses, our pharmacokinetic observations suggest that by the end of each 24 h period the brain levels of MOMIPP would likely fall below the concentrations found to be therapeutically effective when maintained in vitro. For these reasons, we did not carry out survival studies or immunohistochemical analyses of JNK activity in tumor tissues at this stage. The initial findings reported here, coupled with the fact that MOMIPP can kill GBM cells that are resistant to the standard drug, temozolomide , suggest that further development of IPPs as possible therapeutic agents for brain tumors is warranted. Improvements in efficacy may be realized by structural modifications that increase potency or prolong drug half-life in vivo. In addition, the incorporation of these compounds into sustained-release formulations or targeted delivery vehicles may prove advantageous. Finally, the new mechanistic insights pointing to decreased glycolytic function and induction of the JNK stress pathway as important precipitating events in methuosis may suggest opportunities for synergistic combinations with other therapeutic agents that alter these pathways .
These studies provide new insights into the molecular mechanisms underlying methuosis, a non-apoptotic form of cell death that can be induced in glioblastoma and other types of cancer by small molecules. Herein we found that massive vacuolization of endosomal compartments induced by MOMIPP causes an early suppression of glucose uptake and glycolytic metabolism, accompanied by induction of the JNK stress-signaling pathway. These events appear to be pivotal for cell death, since they are not triggered by closely related non-cytotoxic IPPs and pharmacological inhibition of JNK offers substantial protection. Among the consequences of JNK activation, posttranslational events, such as phosphorylation of pro-survival members of the Bcl-2 family, are probably more important than transcriptional events mediated by c-Jun, since a dominant-negative form of c-Jun had no effect on methuosis. Finally, the present studies provide preliminary support for development of IPPs as potential therapeutic agents for brain tumors by showing that the prototype compound, MOMIPP, can readily penetrate the BBB and can inhibit the growth of orthotopic glioblastoma xenografts in mice.
Funding for these studies was provided by the National Institutes of Health (R01 CA 115495), the Helen and Harold McMaster Endowment for Biochemistry and Molecular Biology, and the University of Toledo Foundation (James Laur and William and Leah Molle Brain Tumor Research Fund). The funders provided financial support for research supplies, equipment and personnel salaries, but they did not participate directly in the design or execution of the studies.
Availability of data and materials
All of the data supporting the conclusions of this article are included within the article and its additional supporting files.
ZL, NM, WM, and JO participated in the overall research design. ZL, NM, JO, SG and JS conducted experiments. CT and PE designed and synthesized the chemical compounds used for all studies. ZL, NM, JO, JS, SG and WAM performed data analysis. ZL, NM and WAM wrote the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
The reported studies did not involve human subjects. All animal studies were performed in compliance with United States Public Health Service Policy on Humane Care and Use of Laboratory Animals, under protocols approved by University of Toledo Institutional Animal Care and Use Committee (Reference Number 107491).
Consent for publication
Not applicable. The reported studies did not involve human subjects.
The chemical compounds used in this study are included in a patent licensed to Systems Oncology, Inc. As co-inventors, WAM, PWE, JO and CJT receive a portion of licensing fees and royalties resulting from the agreement. The other authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 29.Ozawa T, James CD. Establishing intracranial brain tumor xenografts with subsequent analysis of tumor growth and response to therapy using bioluminescence imaging. J Vis Exp. 2010;(41):e1986. https://doi.org/10.3791/1986.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.