Inhibition of autophagy and induction of glioblastoma cell death by NEO214, a perillyl alcohol-rolipram conjugate

ABSTRACT

Glioblastoma (GBM) is the most aggressive primary brain tumor, exhibiting a high rate of recurrence and poor prognosis. Surgery and chemoradiation with temozolomide (TMZ) represent the standard of care, but, in most cases, the tumor develops resistance to further treatment and the patients succumb to disease. Therefore, there is a great need for the development of well-tolerated, effective drugs that specifically target chemoresistant gliomas. NEO214 was generated by covalently conjugating rolipram, a PDE4 (phosphodiesterase 4) inhibitor, to perillyl alcohol, a naturally occurring monoterpene related to limonene. Our previous studies in preclinical models showed that NEO214 harbors anticancer activity, is able to cross the blood-brain barrier (BBB), and is remarkably well tolerated. In the present study, we investigated its mechanism of action and discovered inhibition of macroautophagy/autophagy as a key component of its anticancer effect in glioblastoma cells. We show that NEO214 prevents autophagy-lysosome fusion, thereby blocking autophagic flux and triggering glioma cell death. This process involves activation of MTOR (mechanistic target of rapamycin kinase) activity, which leads to cytoplasmic accumulation of TFEB (transcription factor EB), a critical regulator of genes involved in the autophagy-lysosomal pathway, and consequently reduced expression of autophagy-lysosome genes. When combined with chloroquine and TMZ, the anticancer impact of NEO214 is further potentiated and unfolds against TMZ-resistant cells as well. Taken together, our findings characterize NEO214 as a novel autophagy inhibitor that could become useful for overcoming chemoresistance in glioblastoma.

Abbreviations: ATG: autophagy related; BAFA1: bafilomycin A₁; BBB: blood brain barrier; CQ: chloroquine; GBM: glioblastoma; LAMP1: lysosomal associated membrane protein 1; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; MGMT: O-6-methylguanine-DNA methyltransferase; MTOR: mechanistic target of rapamycin kinase; MTORC: MTOR complex; POH: perillyl alcohol; SQSTM1/p62: sequestosome 1; TFEB: transcription factor EB; TMZ: temozolomide

Introduction

Glioblastoma (GBM) is one of the most common and highest mortality intracranial primary tumors, with an incidence of about 45.6% of primary malignant brain tumors [1]. According to the histopathological, morphological, and molecular biological characteristics of gliomas, the World Health Organization divides gliomas into four levels I-IV according to their malignant degree and Isocitrate dehydrogenase (IDH) mutation status. GBM is an IDH wildtype, grade IV tumor according to the new WHO classification [2]. Despite the clinical use of surgical resection, combined with radiation and chemotherapy, the patients’ median survival time is still only about 12–15 months [3]. Treatment with the standard of care chemotherapeutic agent, temozolomide (TMZ), usually results in drug resistance and ultimately tumor recurrence [4,5]. Therefore, there is a need for the development of nontoxic, effective drugs that specifically target chemoresistant gliomas.

It has been shown that autophagy is one of the main mechanisms that causes drug resistance of gliomas [6,7]. Autophagy is an evolutionarily conserved process stimulated in response to cellular starvation and stress that degrades damaged or unnecessary organelles and proteins in a lysosome-dependent manner [8]. In mammalian cells, there are three types of autophagy: macroautophagy, microautophagy, and chaperone-mediated autophagy [9]; macroautophagy is referred to as autophagy hereafter. Autophagy is a dynamic process consisting of initiation, elongation, fusion, maturation, and degradation, which are accurately directed by numerous autophagy-related (ATG) proteins and multiple signaling pathways [10]. Under stimulation, the phagophore membranes generate in the cytoplasm, expand to form autophagosomes, and encircle the degradative cargos through two ubiquitination-like systems, namely, the phosphatidylethanolamine-modified MAP1LC3/LC3 (microtubule associated protein 1 light chain 3) system, and the ATG12–ATG5-ATG16L1 system [10,11]. Then these two ubiquitinated conjugation systems participate in the expansion of the phagophore membrane and the maturation of autophagosomes [12,13]. The autophagosomes then fuse with lysosomes to form autolysosomes and subsequently degrade their cargos, which are dependent on the lysosomal functions [14]. Autophagy is a dynamic process involving multiple steps. Therefore, the detection of all autophagy processes from the formation of the initial autophagosome to autolysosome degradation is currently the recommended detection method.

Accumulating evidence suggests that autophagy plays an important role in different diseases, including cancer, neurodegenerative diseases, as well as metabolic disorders [9]. The role of autophagy in cancer is rather complex and one important understanding is that autophagy is a double-edged sword in cancer. In most cases, autophagy is a survival mechanism and protects against apoptosis [15,16]. Autophagy prevents tumor initiation in healthy tissues, but favors cancer progression once the tumor is formed. Whereas, inhibition of autophagy sensitizes cancer cells to therapy and enhances the cytotoxic effects induced by chemotherapeutic agents [17–19]. Therefore, inhibition of autophagy has been exploited as a possible therapeutic strategy in different tumor models, including glioma, glioblastoma, neuroblastoma, and breast cancers [20–22]. Inhibition of autophagy may sensitize tumor cells to common drugs or may overcome the resistance acquired by those cells to chemotherapeutic agents. More studies have shown autophagy inhibitors can reduce the resistance of gliomas to chemotherapy drugs. As classic autophagy inhibitors, chloroquine (CQ) and hydroxychloroquine have been used as important adjuvant chemotherapeutic drugs for phase I and phase II clinical trials [23]. CQ and hydroxychloroquine inhibit the process of autophagic flux primarily by blocking the formation of autophagic lysosomes and enhancing the response of cells to drug treatment [24]. However, in the clinical trial of Rosenfeld et al [25], long-term treatment with CQ showed toxic effects, such as skin damage and nephrotoxic damage in patients. Therefore, there is an urgent need for developing novel autophagy inhibitors that are more effective and safer to long-term use.

NEO214, the conjugate of rolipram and perillyl alcohol (POH), was synthesized based on the premise that both rolipram and POH have anti-cancer properties and are able cross the BBB [26–28]. Rolipram is a selective inhibitor of PDE4 (phosphodiesterase 4), and the mechanisms associated with the anti-tumor activity of PDE4 inhibitors have been shown to involve decreased angiogenesis and inhibition of inflammatory cell migration [29,30]. POH is a naturally occurring monoterpene [31]. Several clinical trials have shown that orally administered high doses of POH were moderately effective as anti-cancer agents, but also highly toxic, whereas intranasal administration of POH (and in particular its pharmaceutical-grade version, NEO100 was much better tolerated, along with promising activity against GBM [26,32,33]. NEO100 has been approved for fast-track and orphan drug status by the FDA and is currently undergoing a phase I-phase II trial for patients with recurrent GBM after failure of the Stupp protocol [4]. Based on the anti-cancer properties of rolipram and POH, the novel conjugate NEO214 was synthesized and tested for its anti-cancer efficacy. Previous studies from our laboratory have shown that: (1) NEO214 can cross the BBB and induces apoptosis in various human cancer cell types [34,35]; (2) NEO214 induces cell cycle arrest and growth inhibition of tumor cells in vitro and in vivo [34,35]; (3) NEO214 May trigger tumor cell death by activation of endoplasmic reticulum (ER) stress and the death receptor pathway [35]. Taken together, it appears that NEO214 is a compound with great potential as a cancer therapeutic agent.

At present, there is no report regarding the effect of NEO214 on autophagy. However, a recent report has provided some clues that NEO214 can affect ER stress by initiating the unfolded protein response (UPR) in human glioma cells [24], and in view of the conjunctional relationship of ER stress/UPR and autophagy [36], we hypothesized that NEO214 might interfere with autophagy.

In the current report, we present our findings to demonstrate that NEO214 acts as an autophagy inhibitor that interferes with the late stage of the autophagy pathway to induce apoptotic glioma cell death. We found that NEO214 inhibits glioma growth via the MTOR (mechanistic target of rapamycin kinase) complex 1 (MTORC1)-TFEB-autophagy-lysosomal signaling pathway, and significantly reduces the metastatic invasion ability of TMZ-resistant glioma cells through targeting autophagy. The finding of this study highlights the possibility of using NEO214 as a novel autophagy inhibitor that might become useful for overcoming autophagy-induced chemoresistance in glioma.

Results

Effect of NEO214 on apoptosis and autophagy in glioma cells

We used Alamar blue assay to determine the concentration-dependent cytotoxic activity of NEO214 in TMZ-sensitive U251 cells, TMZ-resistant U251 cells (U251TR), and TMZ-resistant (based on high expression of MGMT [O-6-methylguanine-DNA methyltransferase]) T98G glioma cells. The data showed that NEO214 was cytotoxic to both TMZ-sensitive and TMZ-resistant glioma cells and the half-maximal inhibitory concentrations (IC 50) were approximately 100 μmol/L at 48 h (Figure 1A). To examine the effect of NEO214 on autophagy, we first tested the changes of GFP-LC3, a well-known marker for autophagosome formation [37]. As shown in (Figure 1B), NEO214 enhanced the GFP-LC3 puncta formation in cells cultured in full-medium, as well as in cells under starvation. This effect of NEO214 was similar to that of CQ, although to a much lesser extent. Next, we checked the changes of LC3 conversion and SQSTM1/p62 protein levels by western blot. The protein levels of LC3-II and SQSTM1 were markedly enhanced in a time- and dose-dependent manner in two human glioma cancer cell lines treated with NEO214 (Figure 1C–J). Since the increase of GFP-LC3 puncta or LC3-II level may represent either the increased generation of autophagosomes or a blockage in autophagosomal maturation and degradation [38,39], we therefore examined the changes of autophagic flux by NEO214 by tracking the SQSTM1 protein level. SQSTM1 is selectively incorporated into the autophagosome through direct binding to LC3 and is efficiently degraded during autophagy flux. When autophagy is suppressed, SQSTM1 level is increased [40]. As shown in Figure S1A-B, autophagy was inhibited significantly more when cells were treated with NEO214 as compared to rolipram alone, NEO100 alone, or their combination in U251 and T98G cell lines, as evidenced from the increases in the amount of LC3-II and SQSTM1. We have shown in our previous publication that NEO214 is a uniquely conjugated drug that shows significantly greater potency as compared to the potency of its individual components alone, or compared to a mere mixture of its two individual compounds. Conjugation makes NEO214 a unique and functionally stable drug. We surmise that the conjugation increases the efficiency of cellular uptake and thus results in greater bioavailability. Rolipram inhibits PDE4 and as a result, increases cyclic AMP (cAMP) levels in the cells. The cAMP is considered to be an important second messenger that activates MTORC1 levels by increasing the interaction between RHEB and MTOR [41]. As shown in Figure S1E, rolipram inhibits autophagy at higher concentrations and NEO214 retains PDE4 enzyme inhibition characteristics [34]. Taken all these together, NEO214 has greater effect in inhibition of autophagy.

Figure 1.

Effect of NEO214 on the autophagy and apoptosis in glioma cells. (A) Cytotoxic effect of NEO214 on U251, U251TR, and T98G glioma cells using Alamar Blue assay. Cells were treated with different concentrations of NEO214 for 24 h and 48 h and survival of the cells was measured. (B) GFP-LC3-expressing transient transfection with GFP-LC3 plasmid in U251 cells were treated with NEO214 (100 μmol/L) or CQ (25 μmol/L) in full-medium 24 h or HBSS for 6 h. The GFP-LC3 punctation was observed under a confocal microscope (X60) and representative images are shown. The number of GFP-LC3 punctations per cell was quantified and presented as mean ± SD from 50 randomly selected cells based on one batch of three independent experiments. U251 (C-E) and T98G (F-H) were treated with different concentrations of NEO214 (0, 25, 50, 75, 100, 150, 200 μmol/L) for 24 h. Western blots for LC3, SQSTM1 and ACTB were performed. NEO214 promotes LC3-II conversion and SQSTM1 accumulation in dose-dependent manners in U251 and T98G. (I-J) U251 and T98G were treated with 100 μmol/L of NEO214 for 1, 3, 6, 9, 15, and 24 h. Western blots for LC3, SQSTM1, and ACTB were performed. LC3-II:LC3-I ratio and SQSTM1 were quantified and the folds of increase were presented based on three independent experiments (n = 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001,scale bars: 10 μm).

Moreover, the cytotoxic effect of NEO214 was not dependent on the activity of MGMT, because it also occurred in the presence of O^6-benzylguanine (O⁶-BG; an MGMT-blocking agent known to eliminate the protective effect of MGMT in glioma cells) (Figure S1C-D, F). Western blot analysis was performed after cells were treated with NEO214 (100 μmol/L) for 24 h. This result showed that NEO214 is cytotoxic to both TMZ-sensitive and TMZ-resistant glioma cells and its cytotoxic activity was independent of MGMT status.

NEO214 suppresses autophagy at its late stage by affecting lysosomal functions and interfering with the fusion of the autophagosome with the lysosome

To examine potential changes of autophagic flux of glioma cells treated with NEO214, we added well-known late stage-autophagy inhibitors CQ (25 μmol/L) and BAFA1 (100 nmol/L) together with NEO214 (100 μmol/L). The western blot results are shown in (Figure 2A,C) and the ratio of LC3-II:LC3-I and SQSTM1 levels are shown in (Figure 2B,D). As shown in (Figures 2A,B) the ratio of LC3-II:LC3-I by NEO214 alone was a little higher than CQ alone, but NEO214 failed to further enhance the LC3-II:LC3-I levels in U251 cells treated with CQ. In T98G cells, the ratio of LC3-II:LC3-I by CQ alone was higher than NEO214 alone, but again NEO214 could not enhance the LC3-II:LC3-I level treated with CQ. Similar results are shown in BAFA1 treatment with NEO214 (Figures 2C,D). Thus, these results suggest that the increased autophagic markers (GFP-LC3 puncta and LC3II level) in NEO214-treated cells are unlikely due to enhanced autophagic flux, but rather by suppression of the late maturation and degradation stage, an effect similar to that of CQ and BAFA1. Therefore, we next attempted to examine the effect of NEO214 on lysosomal functions. The intra-lysosomal pH is a critical factor in determining the lysosomal functions [42]. We thus examined the effect of NEO214 on intra-lysosomal pH by using the LysoTracker, which exhibits a pH-dependent increase in fluorescence intensity upon acidification [43]. As shown in (Figures 2E,F) NEO214 treatment decreased the fluorescence intensity, especially in T98G cells, indicating NEO214 affected intral-lysosomal pH. CQ treatment effectively abolished the fluorescence in U251 and T98G cells, indicating the neutralization of the intra-lysosomal pH.

Figure 2.

NEO214 enhances the level of autophagy markers by suppressing autophagic flux at its late-stage. (A-B) NEO214 does not promote autophagic flux. U251 and T98G cells were treated with NEO214 alone, CQ (25 μmol/L), or a combination of NEO214 and CQ. Cell lysates were prepared and western blot was performed with LC3, SQSTM1, and ACTB antibodies. ACTB was used as a loading control. LC3-II:LC3-I ratio and SQSTM1 were quantified. (*, P < 0.05; **, P < 0.01; ***, P < 0.001). (C-D) U251 and T98G cells were treated with NEO214 alone, BAFA1 (100 nmol/L), or a combination of NEO214 and BAFA1. Cell lysates were prepared and western blots were performed with LC3, SQSTM1, and ACTB antibodies. ACTB was used as a loading control. LC3-II:LC3-I ratio and SQSTM1 were quantified. (*, P < 0.05; **, P < 0.01; ***, P < 0.001). (E-F) NEO214 affects intra-lysosomal pH. U251 and T98G cells treated with 100 μmol/L NEO214 and 25 μmol/L CQ for 24 h, cells were stained with 5 mmol/L LysoTracker®Red DND99 30 min to evaluate lysosomal pH change. The fluorescence intensity was observed under a confocal microscope 570/590 nmand measured in 10 randomly selected fields. The number of LysoTracker puncta formation per cell was quantified and presented as mean ± SD from 50 randomly selected cells based on one batch of three independent experiments. (G-H) NEO214 interferes with a fusion of the autophagosome with the lysosome. GFP-LC3 expressing stable U251 cells were treated with 100 μmol/L NEO214, 100 nmol/L Baf A1 or 25 μmol/L CQ, respectively, in full-medium 24 h or HBSS for 6 h, then processed for LAMP1 immunostaining and observed under a confocal microscope (*600). The enlarged areas were presented to show the colocalization of GFP-LC3 (in green) with LAMP1 (in red). Quantifying with a colocalization tool from Olympus Fluoview software (Olympus). (I-J) U251 and T98G cells transfected with Mrfp-GFP-LC3 were treated with 100 μmol/L NEO214 and the number of GFP⁺ mRFP⁺ (autophagosome-yellow) and GFP⁻ mRFP⁺ (autolysosome-red) puncta per cell in different group was quantified. 50 cells were randomly selected and counted, and data (mean ± SD) are representative of three independent experiments (n = 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.S., not significant; scale bars: 10 μm).

To further understand the inhibitory effect of NEO214 on autophagy, we next examined the autophagosome-lysosome fusion process by tracking the late endosome/lysosome marker LAMP1 to detect its colocalization with the autophagosomal marker, GFP-LC3 in U251 cells stably expressing GFP-LC3. As a positive control, we used BAFA A1, which is known as a vacuolar H⁺-ATPase inhibitor and able to block the fusion between autophagosome and lysosome [11]. As shown in (Figure 2G), NEO214 treatment induced the GFP-LC3 punctation in cells, but most of the GFP-LC3 puncta (green) did not colocalized with LAMP1 (red) in cells cultured in either full medium or EBSS. These changes were similar to those in cells treated with BAFA1. In contrast, there was extensive colocalization of GFP-LC3 and LAMP1 in CQ-treated and Torin-treated cells (indicated by yellow color). Torin is an MTOR inhibitor and it induces autophagy. Our observations were confirmed by a quantification analysis for the colocalization coefficient as presented in (Figure 2H). Moreover, we transfected a tandem mRFP-GFP-LC3 construct into U251 and T98G cells and examined autophagosome maturation and autolysosome formation. Essentially, this assay is based on the fact that the low pH inside the lysosome quenches the fluorescent signal of GFP, whereas mRFP exhibits more stable fluorescence in acidic compartments. Therefore, if the autolysosome maturation proceeds normally, it would give rise to more red-only puncta. Conversely, if the autophagosome does not fuse with lysosome or the lysosome function is impaired, most of the puncta should exhibit both red and green signals and appear to be yellow. As shown in (Figure 2I,J) in control condition, about 70% of autophagic vacuoles had only red fluorescence signal while the 30% had yellow signal in U251 cells, and which is half of each in T98G cells (Figure 2J). After treatment of the cells with NEO214, yellow punctate fluorescence significantly increased while only-red puncta markedly decreased, indicating a blockade of autophagosome maturation/autophagosome-lysosome fusion. Such findings thus raise the evidence that the NEO214 May inhibit autolysosome formation by interfering fusion of autophagosome with lysosome. Thus, data from these this part of our study present clear evidence that NEO214 inhibits autophagosome maturation by affecting the lysosomal function and impairing autophagosome-lysosome fusion, thus further strengthening our earlier observations that the enhanced autophagic marker in cells treated with NEO214 is due to the suppression of autophagy at the late stage.

NEO214 suppresses autophagic flux dependent on MTOR signaling

To further investigate how NEO214 regulates the autophagy-lysosomal pathway in gliomas, we performed quantitative PCR (qPCR) using mRNA extracted from NEO214-treated U251 and T98G cells. NEO214 decreased the expression of most autophagy-lysosome-related genes (Figure 3A,B) as well as the protein levels (Figure 3C). Moreover, we examined the effect of NEO214 on the autophagy-related proteins that were involved in the formation and extension of autophagosomes. NEO214 seemed to have no obvious effects on the expression of BECN1 (beclin 1), ATG3, ATG5 and ATG9A which are involved in the formation of autophagosomes (Figure 3D). According to our later experiments, we found that NEO214 regulates the transcription factor of TFEB, which regulates the transcription of genes involved in all steps of lysosome biogenesis, thereby affecting the level of transcription and translation of autophagy-lysosome-related genes and proteins.

Figure 3.

NEO214 has an adverse effect on lysosomal function and blocked autophagic flux in glioma cells dependent on MTOR signaling. (A-B) Expression analysis of the autophagy-lysosome relevant genes in 100 μmol/L NEO214-treated U251 and T98G cells. (C) WB analysis of ATP6V1G1, ATP6V1D, ATP6V1H and ATP6V1C1 in U251 and T98G cells were treated with NEO214(100 μmol/L) in full-medium 24 h. (D) NEO214 does not affect the autophagosome formation process in glioma cells. Western blot analysis of the protein expression of BECN1, ATG3, ATG5, ATG9A, LC3, and SQSTM1 in U251 and T98G cells treated by 100 μmol/L NEO214 for 24 h. (E) Western blot analysis of MTOR signaling proteins expressions in U251 and T98G cells treated by 100 μmol/L NEO214 24 h. (n = 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001).

MTOR is the key modulator of autophagy, which stimulates protein translation and cell growth while suppressing amino acid-dependent autophagy at the same time via regulating downstream effectors such as TFEB (transcription factor EB) [44]. We therefore investigated whether blockage of autophagic flux by NEO214 was dependent on MTOR signaling. We found that NEO214 exerted effects on phosphorylation of MTOR itself and its specific substrates RPS6KB/p70S6 kinase and EIF4EBP1 in U251 and T98G cells (Figure 3E), indicating that the inhibitory effect of NEO214 is MTOR dependent, but not associated with the AKT pathway, as indicated by Ser473 phosphorylation. Moreover, we tested whether NEO214 can affect the enzymatic activity of CTSB (cathepsin B) and CTSD (cathepsin D), the most important group of proteases inside the lysosome via western blot [45]. As shown in Figure S5A, treatment with NEO214 was highly effective in suppressing cathepsin B in U251 and T98G cells, but only suppressing CTSD expression in T98G cells. Enzymatic activity assays of CTSB and CTSD in two cell lines showed similar results with western blot in Figure S5B. Collectively, these results demonstrate that NEO214 has adverse effects on lysosomal function and its blockage of autophagic flux was dependent on MTOR signaling.

NEO214 suppresses autophagy-lysosome biogenesis through MTORC1

In mammals, MTOR is composed of two multi-protein complexes, MTORC1 and MTORC2. MTORC1 is constituted of MTOR, MLST8, and RPTOR/raptor. NEO214 treatment resulted in increased phosphorylation of RPS6KB/p70S6K and EIF4EBP1, two known MTOR substrates, suggesting alteration of MTORC1 signaling (Figure 3E). To find out whether inhibition of autophagy by NEO214 is through MTORC1, we knockdown (KD) the MTOR signaling pathway components, e.g., MTOR, RPTOR, and RICTOR by transfection of independent short hairpin RNAs (shRNA). The phosphorylation of EIF4EBP1 have no change of RPTOR-KD in T98G cells means the MTORC1 complex activation have been abolished, whereas KD of RICTOR abolished the MTORC2 complex activation in T98G cells. As shown in western blot results in (Figure 4A–D), phosphorylation of EIF4EBP1 increased in T98G cells with NEO214 treatment and allowed cells to start translation, but phosphorylation of 4EBP1 did not change after NEO214 treatment of T98G cells with KD of MTOR (Figure 4A) or KD of RPTOR (Figure 4C). When we investigated the autophagy markers in T98G cells, NEO214 treatment increased the LC3-II and SQSTM1 levels, which suggested inhibition of autophagic flux. In MTOR-KD and RPTOR-KD T98G cells, NEO214 treatment increased LC3-II level but prevented SQSTM1 induction. This result demonstrated that the autophagy is induced upon MTORC1 inactivation. However, KD of RICTOR had minimal effects on autophagy in T98G cells (Figures 4E,F). These findings were further confirmed by gene expression analysis as shown in (Figure 4G). NEO214-associated downregulation of the autophagy-lysosomal gene expression signature was increased upon RPTOR-KD (Figure 4G). Taken all together, these results suggest that NEO214 activated the MTORC1 pathway, and autophagic flux was inhibited via the MTORC1 pathway.

Figure 4.

NEO214 suppresses autophagy-lysosome biogenesis through MTORC1. (A-B) the effect of MTOR-KD on NEO214’s autophagic flux inhibition in gliomas cells. T98G cells were transfected with specific shRNA targeting MTOR genes, treated by 100 μmol/L NEO214 for 24 h, and cell lysates were prepared. Western blot analysis shows endogenous protein expressions in MTOR-KD-T98G cells after NEO214 treatment. ACTB was used as an internal control. Scramble: cells were transfected with scrambled shRNA as a negative control (Ctrl sh). LC3-II:LC3-I ratio and SQSTM1 were quantified at right. (C-D) the effect of RPTOR-KD on NEO214’s autophagic flux inhibition in gliomas cells. T98G cells were transfected with specific shRNA targeting RPTOR genes, treated by 100 μmol/L NEO214 for 24 h, and cell lysates were prepared. Western blot analysis shows endogenous protein expressions in RPTOR-KD-T98G cells after NEO214 treatment. ACTB was used as an internal control. Scramble: cells were transfected with scrambled shRNA as a negative control (Ctrl sh). LC3-II:LC3-I ratio and SQSTM1 were quantified at right. (E-F) the effect ofRICTOR-KD on NEO214’s autophagic flux inhibition in gliomas. T98G cells were transfected with specific shRNA targeting RICTOR genes, treated by 100 μmol/L NEO214 for 24 h, and cell lysates were prepared. Western blot analysis shows endogenous protein expressions in RICTOR-KD-T98G cells after NEO214 treatment. ACTB was used as an internal control. Scramble: cells were transfected with scrambled shRNA as a negative control (Ctrl sh). LC3-II:LC3-I ratio and SQSTM1 were quantified (F). (G) Expression analysis of the autophagy-lysosome relevant genes in 100 μmol/L NEO214-treated sh-Rptor-T98G cells and sh-Rictor-T98G cells. (Downregulation is indicated in blue and purple, and a similar expression is indicated in yellow and green) (n = 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001).

NEO214 regulates the subcellular distribution of TFEB via MTORC1

Previous studies have reported that TFEB positively regulates the transcription of genes involved in all steps of lysosome biogenesis, thereby promoting lysosomal proliferation, acidification, and exocytosis [45]. Since the initial characterization of TFEB as a transcriptional modulator of the lysosomal system, cytoplasm-to-nucleus translocation of TFEB has been identified as the main mechanism underlying the regulation of TFEB activity [46]. Therefore, we studied the effect of NEO214 on TFEB nuclear translocation although it failed to alter TFEB expression in the whole-cell lysates (Figure 5A). TFEB expression was detected in cytoplasm and nucleus via cell fractionation analysis in U251 and T98G cells. As shown in (Figure 5B), NEO214 induced up-regulation of TFEB in the cytoplasm but down-regulation in the nucleus, indicating its capability to block the nuclear translocation of TFEB. Similar results were shown by anti-TFEB immunostaining fluorescence in U251 and T98G cells, where NEO214 triggered cytoplasm translocation of TFEB (Figure 5C). We performed the experiment with the mRFP-GFP-LC3 puncta assay to confirm autophagic and TFEB activity assays (Figure S6). We hypothesized that MTORC1 might regulate the distribution and activation of TFEB under NEO214 treatment. To test this possibility, we infected T98G cells with sh-Rptor or sh-Rictor and treated the cells with HBSS for 6 h, then treated with 100 μmol/L NEO214 for 24 h. As predicted, inhibition of MTORC1 rapidly induced nuclear translocation of TFEB (Figure 5D). In contrast, the inactivation of MTORC2 by depletion of RICTOR did not change the distribution of TFEB (Figure 5E). Altogether, our results reveal a clear correlation between the activity of MTORC1 and the subcellular distribution of TFEB.

Figure 5.

NEO214 regulates the subcellular distribution of TFEB via MTORC1. NEO214 induces cytoplasmic enrichment of TFEB but does not affect TFEB expression in WCL. (A) Immunoblots for TFEB in whole-cell lysates of U251 and T98G cells treated with DMSO or NEO214 (100 μmol/L, 24 h). (B) Immunoblots for TFEB in cytoplasmic and nuclear fractions of U251 and T98G cells treated with DMSO or NEO214 (100 μmol/L, 24 h). LMNB1 serves as the control for the nuclear fractions, while TUBA1B serves as the control of the cytoplasmic fractions. TFEB was quantified and the folds of increase were presented. The statistical results were from three independent nucleo-cytoplasmic separation and western blot experiments. (C) the expression of TFEB on NEO214 treated U251 cells under HBSS starvation. U251 cells were treated with 100 μmol/L NEO214 or not and stained with anti-TFEB antibody and imaged by confocal microscopy. The nucleus was stained by DAPI (blue). The number of TFEB punctations per cell was quantified and presented as mean ± SD from 50 randomly selected cells based on one batch of three independent experiments. (D) Immunofluorescence confocal microscopy showing TFEB (green) and MTOR (red) localization in T98G cells depleted of RPTOR. (E) Immunofluorescence confocal microscopy showing TFEB (green) and MTOR (red) localization in T98G cells depleted of RICTOR (n = 150 cells per time point, pooled from three independent experiments; *, P < 0.05; **, P < 0.01; ***, P < 0.001; scale bars: 10 μm).

TFEB is an important mediator of NEO214 regulated V-ATPases gene expression by MTORC1

Lysosomal biogenesis is known to be regulated by TFEB [47], and among the genes reported to be regulated by TFEB are several V-ATPases. We therefore set out to examine whether TFEB may be involved in MTORC1-dependent regulation of V-ATPases in response to NEO214 treatment. To determine how broadly TFEB regulated gene expression downstream of MTORC1, we performed gene expression studies in T98G stably depleted of TFEB, and TFEB S211A mutants, which cause default TFEB nuclear translocation and therefore NEO214 cannot inhibit TFEB function. We measured relative TFEB mRNA levels and TFEB protein levels in each construct after NEO214 treatment (Figures 6A,C). As shown in (Figure 6C), NEO214 increased phospho-TFEB levels in T98G control cells, while mutations of serine 211 decreased phospho-TFEB, which increased TFEB nuclear localization. As shown in (Figure 6B), the eight V-ATPase mRNA levels were down-regulated by MTORC1 activation through NEO214 treatment. And, as expected, we observed down-regulation of V-ATPase gene expression in TFEB-depleted T98G cells. On the contrary, the V-ATPase mRNA levels significantly increased with the expression of the S211A mutant, despite MTORC1 remaining active through NEO214. Furthermore, the results were confirmed for two V-ATPase genes by western blot and V-ATPase activity assay (Figure 6C–E). The result showed the protein level and activity of V-ATPase decreased in TFEB-depleted T98G cells, whereas increased with the expression of the S211A mutant. These results showed a consistent correlation of V-ATPase mRNA and protein levels. These results suggest that TFEB is an important effector downstream of MTORC1 involved in NEO214-controlled regulation of a substantial number of V-ATPase genes.

Figure 6.

TFEB is an important mediator of NEO214 regulates V-ATPases gene expression by MTORC1. V-ATPase regulation by NEO214 mediating MTORC1 requires TFEB. (A) Qrt-PCR to test the mRNA expression of TFEB is stably transduced with sh-TFEB vectors and TFEB-S211A in T98G cells. (B) Qrt-PCR analysis shows V-ATPase endogenous expressions in TFEB-KD-T98G cells and TFEB-S211A. (C-D) Western blot analysis shows V-ATPase endogenous protein expressions and phospho-TFEB protein expression in TFEB-KD-T98G cells and TFEB-S211A. ATP6V1A and ATP6V1B2 were quantified and the folds of increase were presented based on one batch of three independent experiments as shown on the right. (E) V-ATPase activity assay in T98G cells with 100 μmol/L of NEO214 treatment for 24 h. (F) Luciferase-positive human glioma cells (U251TR: 2 × 10⁵ cells) were intracranially injected into the brain parenchyma of the athymic female nude mice (8–10 weeks old). Treatment started 9 days post-implantation. Vehicle group (n = 3) and NEO214 (50 mg/kg) group (n = 4) were treated subcutaneously for 30 days with 5 days on and 2 days off schedule. Tumor growth was monitored by imaging the mice and survival were monitored. Animals treated subcutaneously with NEO214 (50 mg/kg) demonstrated a significant decrease in tumor growth as compared to the vehicle control group (p < 0.0032). The median survival for the vehicle group was 36 days, while the median survival for the NEO214-treated group was 87 days, i.e, there was a 2.4-fold increase in median survival after NEO214 treatment. (G) H&E-stained section and immunohistochemical (IHC) staining in indicated T98G xenograft tumor genotypes. SQSTM1, p-MTOR, and TFEB staining in the T98G implanted mouse brain tumor 20 days after treatment of 100 mg/kg NEO214. The black arrow shows nuclear TFEB positive staining. (H) Western blot analysis shows SQSTM1, LC3, MTOR, p-MTOR, TFEB and phospho-TFEB protein levels in tumor tissues (*, P < 0.05; **, P < 0.01; ***, P < 0.001; scale bars: 50 μm).

NEO214 is effective in both subcutaneous and intracranial mouse model

To determine whether NEO214 is effective in an in vivo mouse model, we used U251TR luciferase-positive human glioma cells that were intracranially injected into the brain of athymic nude mice. NEO214 treatment started nine days post-implantation. Mouse survival was monitored by noninvasive imaging (Figure S2A), and survival of the animals was monitored (Figure 6F). As shown in Figure 6F, animals treated subcutaneously with NEO214 (50 mg/kg) demonstrated a significant decrease in tumor growth as compared to the vehicle control group (p < 0.0032). The median survival for the vehicle group was 36 days, while the median survival for the NEO214-treated group was 87 days, i.e, there was a 2.4-fold increase in median survival after NEO214 treatment. The tumor progression imaging data is shown in Figure S2A. The U251 subcutaneous xenografts (Figure S2B) showed that NEO214 at 25 mg/kg (administered subcutaneously for 30 days) significantly delayed tumor growth compared to vehicle treatment. These data indicate that NEO214 is active against TMZ-sensitive and TMZ-resistant glioblastoma in both subcutaneous and intracranial models.

To assess the impact of NEO214 on autophagic progression in vivo, we xenografted NOD/SCID mice with T98G tumor cells and found that NEO214 treatment resulted in increased SQSTM1 and MTOR expression in tumor tissue. Moreover, phospho-TFEB increased by NEO214, meaning NEO214 enhanced cytoplasmic accumulation of TFEB, as compared to tumor tissue from control animals (Figures 6G). These results revealed that the effects of NEO214 on autophagy and TFEB, as presented above with in vitro models, also take place in the in vivo setting.

NEO214 enhances the cytotoxic effects of Chloroquine and Temozolomide

We have shown previously that treatments with combinations of CQ and TMZ lead to efficient glioma cell death in vivo and in vitro, even in glioma cell lines that develop chemo-resistance to TMZ [7]. CQ was previously reported to combine with conventional therapy, which significantly increased the median survival of GBM patients from 11.4 to 25 months [48]. However, the clinical trial of Rosenfeld et al. [25] found that long-term treatment with CQ resulted in toxicity that included skin damage and nephrotoxic damage. We now investigated whether NEO214 as an autophagy inhibitor would be able to enhance the effects of CQ, which might allow the use of lower CQ dosages, potentially minimizing unwanted side effects. Furthermore, we investigated whether this approach might be able to overcome resistance to TMZ. We used the same TMZ-sensitive and TMZ-resistant cell lines as above, and exposed them to single, dual, and triple combination drug treatments for 48 h.

Based on the many possible combinations at different concentrations, a large data set was obtained and representative results from MTT assays are presented in (Figure 7) and Figure S3. Consistent with our previous report [7,35], the results showed that TMZ was only cytotoxic for U251 (TMZ-sensitive) cells and minimally affected the TMZ-resistant cells (Figure S3A). CQ was shown to inhibit cell proliferation in all cells either alone or in combination with TMZ (Figures S3B and S3D), with IC50 of 20 μmol/L for U251, 50 μmol/L for U251TR and T98G. We chose 100 μmol/L TMZ and 10 μmol/L CQ to combine with NEO214 for subsequent treatments. As shown in (Figures 7A,B) the addition of NEO214 significantly further increased the cytotoxic effects of the dual treatment in TMZ-resistance glioma cells (U251TR, T98G), 50 μmol/L NEO214 combined with 100 μmol/L TMZ and 10 μmol/L CQ reduced the percent viability of U251TR and T98G by 50%, although 100 μmol/L TMZ and 10 μmol/L CQ combination reduced the percent viability of U251TR and T98G by only 10%.

Figure 7.

NEO214 enhances the cytotoxic effects of Chloroquine and Temozolomide by blocking autophagy. (A-B) Cell growth and survival of U251TR (A) and T98G (B) cell lines were determined by MTT assays after 48 h of culture in the presence of increasing concentrations of NEO214 with CQ and TMZ. (C-D) TMZ-resistant cell lines: U251TR and T98G were exposed to 100 μmol/L TMZ, 10 μmol/L CQ, 100 μmol/L NEO214, and a combination of three drugs for 48 h, then cultured for another 7–10 days in fresh medium in the colony formation assay (CFA). The number of colonies was quantified and presented as mean ± SD from three independent experiments. (E-F) Effects of TMZ-resistant cell lines U251TR and T98G were exposed to 100 μmol/L TMZ, 10 μmol/L CQ, 100 μmol/L NEO214, and a combination of three drugs for 48 h. Cell flow cytometry analyzes the effects of NEO214 and other drugs on the cell cycle of U251TR cells and T98G cells. Statistical analysis of the G₀/G₁ phase and G₂/M phase of U251TR and T98G cells in each drug treatment group. (G-H) Expressions of cyclin-related proteins and apoptosis-related proteins in TMZ-resistant cell lines U251TR and T98G were exposed to 100 μmol/L TMZ, 10 μmol/L CQ, 100 μmol/L NEO214 and a combination of three drugs for 48 h (n = 3; *, P < 0.05).

To confirm the results obtained with short-term MTT assays, we performed a series of longer-term colony formation assays (CFAs). Two of the TMZ-resistant glioma cell lines (U251TR and T98G) were exposed to different concentrations of drugs, singly and in combination for 48 h, and the ability of cells to survive and spawn a colony of descendants during the following 2 weeks was determined. As shown in Figure 7C, single TMZ and CQ or TMZ-CQ-combination treatments were relatively non-effective, and CQ in particular exerted no detectable effect on long-term cellular viability at the chosen concentration. In comparison, single NEO214 treatment reduced cell viability to about 40%, and the combination of NEO214 with TMZ or CQ further enhanced this effect. The addition of NEO214 to the TMZ and CQ dual combination resulted in greatly increased toxicity, where survival dropped to only 10%. These results demonstrate that NEO214 combined with TMZ and CQ significantly enhanced tumor cell death (Figure 7D).

Furthermore, we showed that the effect of NEO214 on the TMZ-resistant glioma cells was due to a growth arrest in cell proliferation in the G₀/G₁ phase, whereas TMZ caused U251 cell growth arrest in the G2/M phase (Figure 7E,F and Figure S4A-B). The results showed that NEO214 combined with TMZ and CQ exhibited a significant additive effect for the TMZ-resistant glioma cell proliferation in the G₀/G₁ phase, compared to each drug alone (Figure 7E,F and Figure S4A-B). Based on the information that NEO214 enhanced TMZ+CQ cytotoxicity in TMZ-resistant glioma cells and acted as an autophagy inhibitor, we next investigated the contribution of autophagy inhibition to the cytotoxic properties of our three-drug cocktail by western blot analysis (Figures 7G,H). We first analyzed the expression levels of p-TP53/p53, PARP, and CDKN1B/p27, three indicators of apoptosis promotion, and CCND1 (cyclin D1), CCND2 (cyclin D2), and CCNE1 (cyclin E1), three general cell cyclin markers in the G1 phase. As shown in Figure S4C, NEO214 treatment resulted in increased expression levels of all three apoptosis markers, but decreased CCND1 and CCND2 protein levels. In comparison, the addition of NEO214 to TMZ or CQ or three-drug treatment caused stronger expression of these markers (Figure 7G,H). Furthermore, tumor metastasis correlated with suppressed autophagy-lysosomal function, as indicated by decreased expression of EMT markers after NEO214 treatment (Figure S4D). Altogether, our results demonstrate that the addition of NEO214 greatly increases the cytotoxic efficacy of low concentrations of CQ and combinations of TMZ+CQ in TMZ-resistant glioma cell lines.

Discussion

It has been shown that inhibition of autophagy may facilitate apoptosis or necrosis [49]. Inhibition of autophagy for cancer treatment has been demonstrated to be a promising therapeutic strategy [50,51] in different tumor models, including gliomas [6,7]. Previous studies indicated that the accumulation of autophagosomes could increase the therapeutic effects of anti-cancer drugs. For example, the accumulation of autophagosomes via treatment with the autophagy inhibitor chloroquine remarkably increased the anti-cancer effect of vorinostat, a histone deacetylase inhibitor, in advanced solid tumors [52]. Inhibition of autophagy at early or late stages leads to different consequences; prevention of autophagosome formation may neutralize the protective role of autophagy, sensitizing cells to chemotherapeutic agents, and enabling a good strategy for combination regimens. Conversely, lysosomotropic agents produce the accumulation of autophagy vacuoles, leading to cellular stress and a consequent cytotoxic effect, being able to kill the tumor in a single therapy [53].

In our current study, we discovered that NEO214 acts as a new type of autophagy inhibitor, resulting in glioma cell death. Specifically, we found that NEO214 inhibits the autophagy-lysosome stage of glioma cells through the MTORC1-TFEB-autophagy-lysosomal axis (Figure 8). This molecular mechanism of NEO214 also applies to TMZ-resistant tumor cells and therefore presents NEO214 as an agent that might become useful for targeting autophagy as an Achilles’ heel in TMZ-resistant gliomas.

Figure 8.

Schematic diagram of NEO214 inhibiting glioma progression through a MTORC1-TFEB signaling pathway. NEO214 inhibits TFEB nuclear translocation by activating MTORC1 activity and as a result, reducing the expression of autophagy-lysosomal related genes (e.g., V-ATPases) and blocking the late stage of autophagy flow in glioma cells. NEO214, as a new type of autophagy target small molecule drug, has a significant inhibitory effect on the cell proliferation and cell cycles of glioma cells.

In the first part of our study, we present evidence that NEO214 is a potent autophagy inhibitor via disrupting the autophagic flux of glioma cells, based on the following observations: (1) NEO214 increases autophagic markers by inhibiting autophagic flux; (2) NEO214 has an adverse effect on lysosomal pH and cathepsin enzyme activity, although it affected lysosome functions differently in different glioma cell lines; and (3) NEO214 prevents the colocalization of GFP-LC3 (a marker for the autophagosome) and LAMP1 (a marker for the lysosome). These effects of NEO214 are similar to those of bafilomycin A1, which is known to block the autophagosome-lysosome fusion process. Data from this study thus expand our current knowledge about the biological mechanism of action of NEO214.

By examining how autophagy is regulated by NEO214, we found NEO214 has obvious effects on MTOR phosphorylation, indicating activation of this kinase. In mammals, MTOR is a key modulator of autophagy [43] and is composed of two multi-protein complexes, MTORC1 (harboring RPTOR protein) and MTORC2 (harboring RICTOR protein). NEO214 treatment up-regulated the MTORC1 pathway and SQSTM1 protein levels. However, in RPTOR-deficient glioma cells, NEO214 could no longer trigger MTORC1 activation, nor increase SQSTM1 levels (Figure 4A–D). These findings were further confirmed by gene expression analysis, whereby NEO214-associated down-regulation of the autophagy-lysosomal signature was prevented upon RPTOR-KD (Figure 4G). To understand the functional interaction of autophagy inhibition with MTORC1 signaling, we found that NEO214 suppressed TFEB through MTORC1-mediated TFEB phosphorylation, leading to its cytoplasmic retention and thereby inactivation. Particularly, TFEB is involved in the regulation of a substantial number of V-ATPase genes (Figure 6C–E), and NEO214 abolished this regulation as well.

Rapamycin and its derivatives (i.e., rapalogs) are a class of allosteric inhibitors of MTORC1 which were extensively evaluated in clinical trials as anticancer agents. Despite some promising results in preclinical models of cancer, these molecules have demonstrated only limited successes in clinical trials, especially when used as monotherapies. Based on their ability to decrease protein translation and attenuate cell cycle progression, rapamycin and its derivatives were shown to have primarily cytostatic effects, which makes them mainly useful as disease stabilizers. Because of their ability to inhibit MTORC1, rapamycin and its derivative are considered autophagy inducers. For this reason, this class is able to mimic a nutrient deprivation scenario. However, this effect appears to be only short-lived because the MTORC1 inhibition by rapalogs ultimately fails to repress a negative feedback loop which results in phosphorylation and activation of AKT. Nonetheless, the effect of rapamycin on autophagy induction is very well documented. In fact, most if not all autophagy induction conditions such as nutrient or growth factor deprivation and low cellular energy levels have been shown to act via inhibition of MTORC1 activity. This establishes a tight, inverse coupling between autophagy induction and MTORC1 activation. Unlike rapamycin, NEO214 appears to do the opposite: i.e., as an MTORC1 activator, it acts as an autophagy inhibitor via cytoplasmic sequestration of TFEB. This mechanism of action by NEO214 might prove to be extremely useful in a clinical situation in which the cancer patient is able tolerate a calorie restriction regimen (i.e., intermittent fasting) while undergoing chemotherapy. We speculate that in this scenario, NEO214 might be able significantly augment the effects of fasting by blocking the ability of cancer cells to turn on the autophagic process.

Another important finding from our study is that NEO214 enhances the cytotoxic effects of chloroquine and temozolomide by blocking autophagy. Previously, we have demonstrated that the combination of CQ and TMZ results in decreased glioma cell proliferation [7]. Because it has been proposed that autophagy may protect glioma cells from chemotherapeutics treatment [54], we investigated whether blockage of autophagy would restore TMZ sensitivity of TMZ-resistant glioma cells. Towards this goal, we added NEO214 to CQ and TMZ treatment and found that NEO214 greatly increased the cytotoxic efficacy of CQ+TMZ treatment in TMZ-resistant glioma cells. CQ mainly affects the acidity activity [24] and NEO214 mainly inhibits autophagy by impairing autophagosome fusion with lysosome and endo-lysosomal systems, which might explain the synergistic impact of NEO214 and CQ when combined. However, in this report, we primarily show results obtained with lower CQ concentrations, because the mutually enhancing effects emerged most impressively under these conditions. Although we did not demonstrate in this paper the potential role of autophagy inhibition in stimulating immunotherapy, Jiang et al. recently reviewed how autophagy may have an additive or inhibitory role in the immunotherapy response for tumors by affecting both innate and adaptative immune responses [55]. Our own preliminary data with NEO214 suggests that this mechanism may indeed be important, and will be the subject of future work.

In conclusion, our study demonstrates that NEO214 is a novel autophagy inhibitor that blocks the late stage of the autophagy pathway, resulting in glioma cell death. This compound inhibits the autophagy-lysosome stage of glioma cells through the MTORC1-TFEB signaling pathway. In combination with CQ and TMZ, NEO214 enhances the cytotoxicity of TMZ in otherwise TMZ-resistant glioma cells. In view of the commonly emerging treatment resistance of GBM in clinical practice, which leaves no effective treatment options and spells dismal prognosis for the affected patients [56,57]. the development of novel agents with the potential to overcome drug resistance is of great interest and urgency [58].

Our previous results show that NEO214 triggers ER stress in tumor cells mainly through Ca²⁺ leak mediated by the translocon (TLC). Blocking intracellular Ca²⁺ leak through TLC by anisomycin blocks the ER stress induction by NEO214 [35]. The induction of autophagy may be the initial cellular protective mechanism against ER stress. Our current finding consolidates our understanding of NEO214, not only as an inducer of ER stress and apoptosis, but also as an inhibitor of autophagy, the protective mechanism against apoptosis. The concurrence between ER stress and autophagy is common in human diseases. Therefore, it is important to understand how autophagy can be manipulated via ER stress regulators to favor prosurvival or prodeath signaling. Development of new drugs that target the linkages between the UPR branches and autophagy will have significant therapeutic benefits in the treatment of such diseases. Within this context, further development of NEO214 might hold promise as a potentially effective second-line treatment with an autophagy-centered mechanism of action that is different from that of common alkylating agents.

Materials and methods

Reagents

NEO214 was provided by NeOnc Technologies, Inc. (Los Angeles, CA) as a crystalline powder. It was prepared as 100 mmol/L stock solution in DMSO and stored in aliquots at −20°C. Chloroquine (CQ; Sigma Aldrich, C6628-25 G) was prepared as a 100 mmol/L stock solution in phosphate-buffered saline (PBS; Genesee Scientific, 25–508) and stored at −20°C. Bafilomycin A₁ (BAFA1; Sigma Aldrich, 19–148) was prepared as 100 mmol/L stock solution in DMSO and stored at −20°C. Torin 1 (Sigma Aldrich, 475991-10 MG) was prepared as a 100 mmol/L stock solution in DMSO and stored at −20°C. All drugs were added to the culture medium in a manner so as to keep the final concentration of DMSO below 0.4%. There were no differences in media control cultures in the presence or absence of DMSO. Capthepsin D Activity Assay (Fluorometric; Abcam, ab65302) and Capthepsin B Activity Assay (Fluorometric; Abcam, ab65300) were used according to the manufacturer’s instructions.

Cells and cell culture

The human glioma cell lines U251(Sigma Aldrich, 09063001) and T98G (ATCC, CRL-1690) were purchased. U251TR cells represent a TMZ-resistant subline of parental U251 cells that was developed in our laboratory as described previously [28,59]. All cells were cultured in DMEM (Genesee Scientific, 25–500) supplemented with 10% fetal calf serum (FCS), 100 U/ml penicillin, and 0.1 mg/mL streptomycin, in a humidified incubator at 37°C and 5% CO₂.

Cell viability assay

Cells (5000 cells/well) were seeded in 96-well assay black plates with a flat bottom (Genesee Scientific, 25–109). After 24 h, different concentrations of drugs were added to the cells and incubated for 48 h. The Alamar blue assay was performed according to the manufacturer’s protocol (Life Technology, DAL1025). Fluorescence was measured using an excitation wavelength of 540–570 nm (peak excitation is 570 nm) and emission at 580–610 nm (peak emission is 585 nm). The average fluorescence values of the cell culture medium alone (background) were subtracted from the fluorescence values of experimental wells. Percent survival was calculated relative to untreated control cells. All experiments were performed in triplicate.

Colony-forming assay

Cells were seeded in 6-well plates at 200 cells/well and allowed to adhere overnight. Subsequently, cells were treated with drugs for 48 h; the medium was then removed and a fresh medium (without drug) was added. Cells were incubated for an additional 7–10 days. At the termination of the assay, colonies were visualized by staining with 1% methylene blue (ThermoFisher Scientific, 414240250) in methanol for 4 h. Dyes were washed out with water and plates were air-dried. The stained colonies were counted. Percent colonies were calculated relative to untreated control cells. All experiments were performed in triplicate.

Western blot analysis

Cells grown in 6-well or 24-well plates were washed with PBS and harvested. Total cell lysates were prepared by disrupting cells with RIPA buffer (Sigma Aldrich, R0278) containing “protease and phosphatase” (ThermoFisher Scientific, 78442) inhibitors. Protein concentrations were determined using the BCA protein assay reagent (ThermoFisher Scientific, 23227). 50 µg of total cell lysate was added to each lane; 10%, 12.5%, or 15% of SDS-PAGE gels were used according to the size of the protein of interest. Trans-blot (Bio-Rad, Hercules, CA, USA) was used for the semi-dry transfer. Immunoblotting was performed with PARP (Cell Signaling Technology, CS-9542), LC3B (Cell Signaling Technology, CS-2775), SQSTM1/p62 (Cell Signaling Technology, CS-5114), BECN1/beclin 1 (Cell Signaling Technology, CS-3738), RPTOR/raptor (Cell Signaling Technology, CS-48648), RICTOR (Cell Signaling Technology, CS-2140), p-AKT (Cell Signaling Technology, CS-4058), p-RPS6KB/p70S6K (Cell Signaling Technology, CS-9206), RPS6KB/p70S6K (Cell Signaling Technology, CS-9202), p-MTOR (Cell Signaling Technology, CS-2971), p-EIF4EBP1 (Cell Signaling Technology, CS-9456), EIF4EBP1 (Cell Signaling Technology, CS-9452), MGMT (Cell Signaling Technology, CS-2739S), CTSD/cathepsin D (Cell Signaling Technology, CS-2284), ATG3 (Cell Signaling Technology, CS-3415), ATG5 (Cell Signaling Technology, CS-12994), TP53 (Cell Signaling Technology, CS-9282), cleaved CASP7/caspase 7 (Cell Signaling Technology, CS-9491), TUBA1B/α-Tubulin (Cell Signaling Technology, CS-2125), LMNB1/Lamin B1 (Cell Signaling Technology, CS-13435), ATP6V1A (Abcam, ab -199,326), ATP6V1B2 (Abcam, ab -73,404), MTOR (Abcam, ab-2732), and CTSB/cathepsin B (Abcam, ab58802), TFEB (ThermoFisher Scientific, PAS-34360; 1:100 for IP, 1:200 for IF, 1:1000 for WB), TSC2 (Proteintech, 24601–1-AP), TSC1 (Proteintech, 20988–1-AP), ATP6V1D (Proteintech, 14920–1-AP), ATP6V1G1 (Proteintech, 16143–1-AP), ATP6V1H (Proteintech, 26683–1-AP), ATP6V1C1 (Proteintech, 16054–1-AP), AKT (GeneTex, GTX-121937), ATG9A (GeneTex, EPR-5973), CCND1/cyclin D1 (Santa Cruz Biotechnology, SC-753), CCND2/cyclin D2 (Santa Cruz Biotechnology, SC-754), CCNDE1/cyclin E1 (Santa Cruz Biotechnology, SC-377100), ACTB/actin (Santa Cruz Biotechnology, SC-8432), CTSB/capthepsin D (Santa Cruz Biotechnology, sc -377,124). Horseradish peroxidases (HRP)-conjugated secondary antibodies were purchased from Vector Laboratories (BA-1000). The protein bands were visualized with a SuperSignal West Pico Chemiluminescence Substrate (ThermoFisher Scientific, 34080). LAS-4000 imaging system (Fujifilm Life Science, Cambridge, MA, USA) was used to detect the specific protein signals. Densitometric values were determined and quantified using the ImageJ software and normalized against the ACTB loading control.

RNA extraction, cDNA synthesis, and qPCR analysis

RNeasy Mini kit (QIAGEN, 74104) was used to extract the total RNA of U251 and T98G cells. The RNA products isolated using this kit were loaded on a 1% agarose gel and separated by electrophoresis. Subsequently, bands for RNA28S, RNA18S, and RNA5S RNAs were inspected to ensure that RNA was extracted correctly. Next, the RNA quality and concentration were assayed using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, Inc.). A total of 1 µg RNA was used for cDNA synthesis using the iScript™ cDNA Synthesis Kit (Bio-Rad Laboratories, Inc., 170–8890). The reverse transcription (RT) conditions were as follows: 25°C for 5 min, 46°C for 20 min, 95°C for 60 s, holding at 4°C. Primers used for qPCR are listed in Table SI. PerfeCTa® SYBR® Green SuperMix (Quantabio, 95054–100) was used for qPCR and the conditions were as follows: 95°C for 3 min, 95°C for 15 s, 55°C for 45 s, 72°C for 30 s, a total of 35 cycles. A CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.) was used to obtain and analyze the data. ACTB/actin was used to normalize gene expression levels. For quantification, the 2-ΔΔCq method was used for this experiment [60].

Transient transfection with GFP-LC3 and Mrfp-GFP-LC3

EGFP-LC3 (11546; deposited by Karla Kirkegaard) and pcDNA3-GFP-LC3-RFP-LC3ΔG (168997; deposited by Noboru Mizushima) were purchased from Addgene. U251 cells or T98G cells were transiently transfected with GFP-LC3 plasmid or mRFP-GFP-LC3 plasmid for 24 h using Lipofectamine 2000 transfection reagent (Life Technology, 11668030) according to manufacturer’s instructions. Cells grown on coverslips were then treated with NEO214 or CQ in complete medium 24 h or Hanks’ Balanced Salt Solution (HBSS; VWR, VWRL0121–0500) for 6 h. The GFP-LC3 puncta formation was observed under a confocal microscope (x60).

Immunofluorescence staining assay

U251 cells were transiently transfected with GFP-LC3 plasmid for 24 h using Lipofectamine 2000 transfection reagent. Cells grown on coverslips were treated with drugs for 24 h then processed for LAMP1 immunostaining. To detect the colocalization of GFP-LC3 puncta with LAMP1, the cells were fixed with 4% PFA in PBS for 30 min at room temperature, washed with PBS, permeabilized with 0.3% Triton X-100 (Sigma Aldrich, T8787) in PBS for 30 min at 37°C. Subsequently, slides were rinsed with PBS for an additional 5 min and blocked with SEA blocking buffer (ThermoFisher Scientific, 37527) for 1 h at room temperature. Primary antibody for LAMP1 (BD Biosciences, 555798) was added to the coverslips for overnight incubation at 4°C. The next day, slides were washed with PBS and incubated with the appropriate Alexa Fluor-conjugated secondary antibody (Life Technologies, A32723) at the recommended concentration for 1 h. The images were obtained using confocal microscopy.

Immunohistochemistry

Tissue sections were fixed in 10% neutral buffered formalin at room temperature for 48 h and then embedded in paraffin. The thickness of the tissue sections was 4 µm. For immunohistochemical staining, slides were placed in a 65°C incubator for 30 min. Tissue slides were deparaffinized in xylene for 5 min, then in fresh xylene for another 5 min, 100% ethanol for 5 min, 95% ethanol for 5 min, and 75% ethanol for 5 min, and then washed in water for 5 min. Endogenous peroxidase was blocked with 3% hydrogen peroxide. Antigen retrieval was achieved using a hot water bath (100°C) and 10 mM citric sodium buffer (pH 6.0) for 15 min. Sections were then blocked with 5% BSA (Fisher Scientific, SH3057402) for 1 h at room temperature and incubated overnight at 4°C with the indicated primary antibodies in 5% BSA: SQSTM1/p62 (Cell Signaling Technology, CS-5114; dilution, 1:500), p-MTOR (Cell Signaling Technology, CS-5114;dilution, 1:100) and TFEB (ThermoFisher, PAS-34360; dilution, 1:100). Antibody binding was detected using the EnVision™ Dual Link System-HRP DAB kit (Dako, K4010; undiluted). Anti-rabbit HRP labeled polymer (Dako, K4010; undiluted) was used to cover tissues, followed by incubation for 30 min at room temperature. Subsequently, 20 µl DAB (~1 drop) in 1 ml DAB substrate buffer was applied to each slide and incubated for 5 min. Sections were then counterstained with hematoxylin for 5 min at room temperature, followed by washing of the slides at room temperature. After staining, the slides were washed with running tap water for 5 min. Subsequently, tissue slides were placed in 75% ethanol for 5 min, 95% ethanol for 5 min, 100% ethanol for 5 min, xylene for 5 min, and then in fresh xylene for another 5 min. For negative controls, primary antibodies were excluded. The mitotic index was quantified by viewing and capturing images of 10 random high-power fields for each tissue section on a Keyence All-In-One Fluorescence Microscope (Keyence Corporation), using a 40× or 20× objective. For evaluation and quantification of immunohistochemical data, 10 fields within the tumor area under high power magnification (40×) were randomly selected for evaluation using ImageJ version 1.49 software (National Institutes of Health). The investigators performed blind counting for all quantifications.

Lentiviral gene knock-down by short-hairpin RNA

Lentiviral-compatible shRNAs against MTOR (sh1: sense: GCAGTTTGCCAGTGGCCTAAA; sh2:sense:GCAAAGATCTCATGGGCTTCG), RPTOR (sh1:sense:GGATGAAGGATCGGATGAAGA; sh2:sense:GCTTGATCGTCAAGTCCTTCA), RICTOR (sh1:sense:CAGACCTCATGGATAATTA; sh2:sense: CAGGCCAGACCTCATGGAT) were purchased from Open Biosystem (Huntsville, AL). For lentivirus production, HEK293T cells were transfected with the transfer vector (e.g., pCDH-CMV-MCS-EF1-Puro or pGIPZ; pCMV-dR8.91 packaging plasmid, and pCMV-VSV-G envelope plasmid (Addgene, 8455; deposited by Bob Weinberg) in a 5:1:4 ratio using the Calcium Phosphate Transfection Kit (Clontech, 631312). The medium was replaced 12 h later. Viral particles were collected 48 h post-transfection, filtered through a 0.45 μm sterile filter, and concentrated overnight by Lenti-X concentrator (Takara, 631312) at a ratio of 3:1, followed by centrifugation at 4°C (28,800 × g, 2 h, ThermoFisher Sorva RC 6+). Viral particles were resuspended in a fresh medium with 8 μM/mL polybrene (Santa Cruz Biotechnology, sc -134,220) and were plated with target cells for 24 h. Lentiviral-transduced cells were selected in 2 μg/mL puromycin for 7 days with the medium changed daily [19].

Image capture and analysis

Images were acquired using Olympus Stream software (version 2.4; Olympus Corporation) under a fluorescence microscope (Olympus Corporation), and ImageJ (version 1.49; National Institutes of Health) was used to analyze the images. Band densitometry and data quantification were also conducted using ImageJ. All images in the present study were grouped using Adobe Illustrator software (version 24.0; Adobe Systems, Inc.).

LysoTracker Red DND-99

Cells were seeded in 12-well plates at 200 cells/well and allowed to adhere overnight. Subsequently, cells were treated with drugs for 24 h, the medium was then removed and LysoTracker Red DND-99 dye (ThermoFisher Scientific, L7528) was added, then fixed with 4% PFA at room temperature for 15 min, and blocked with a blocking solution. After DAPI staining, the images were obtained using confocal microscopy.

V-ATPase enzyme activity measurements

V-ATPase activity of 10 µg microsomal membranes was colorimetrically determined as described previously [61,62] after an incubation period of 40 min at 28°C. Reactions were terminated by adding 40 mM citric acid, pH X. For the blank value 10 μg BSA was used instead of microsomal protein. The V-ATPase assay medium contained 25 mM Tris-Mes, pH 7.0, 4 mM MgSO₄ *7 H₂O (Sigma, 63138), 50 mM KCl (Sigma, 58221), 1 mM NaN₃ (Sigma, RTC000068), 0.1 mM Na₂MoO₄ (Sigma, 243655), 0.1% Brij 35 (Sigma, 1.01894), 500 μM NaVO₄ (Sigma, 450243), and 2 mM Mg-ATP (Sigma, SRE0045). Activity was expressed as the difference between the measurements in the absence and the presence of 100 nM concanamycin A (Millipore Sigma, 27689).

Cell cycle analysis

Cells were collected by trypsinization and pipetting to form a single-cell suspension, followed by washing with PBS and fixation in 70% ethanol at 4°C overnight. After gentle centrifugation and washing with PBS to remove ethanol, propidium iodide (PI) was added to the cells. Incubation with PI was for 30 min at 37°C in the dark. Excitation wavelength was 488 nm, and flow cytometry software Modifit was used for cell DNA content analysis.

In vivo intracranial glioma rodent model

All animal protocols were approved by the University of Southern California Institutional Animal Care and Use Committee (IACUC). Intracranial implantation of tumor cells was performed as previously described [62]. Briefly, 8–10-week-old athymic nude mice were anesthetized and fixed onto a stereotactic head frame. A 0.7-mm diameter hole was drilled at the coordinates, 1.0 mm posterior, 1.0 mm lateral (right) with respect to bregma, and a 25 G Hamilton syringe was inserted at a depth of 2.5 mm ventral. 2 × 10⁵ luciferase-positive temozolomide-resistant human glioma cells were then injected into the subcortical brain parenchyma. The implanted mice were imaged by bioluminescent imaging 10 days after implantation. Once tumors were confirmed by imaging, the animals were randomly placed into different groups, and treatment was started. NEO214 at 50 mg/kg was diluted in the vehicle (50% glycerol:50% ethanol) and administered subcutaneously into the neck scruff region at a volume of 100 µL. Tumor growth was monitored by imaging; survival was documented. For pathology analysis and drug toxicity analysis, mice were treated with NEO214 at 100 mg/kg diluted in the vehicle (50% glycerol: 50% ethanol) and administered subcutaneously into the neck scruff region for 20 days (5 days on and 2 days off). After the treatment period, organs (i.e., spleen, liver, kidney, intestines, heart, and lung), bone marrow and blood were harvested. Organ samples were fixed in 10% formalin, paraffin-embedded, sectioned, and stained with hematoxylin and eosin.

Subcutaneous glioma mouse model

All animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of USC, and all rules and regulations were followed during experimentation on animals. Athymic mice were implanted with 2 × 10⁶ tumor cells into the right flank. About 10 days later, once palpable tumors had developed, animals were assigned to different treatment groups (n = 4 per group). The control group received vehicles only (50% glycerol:50% ethanol), whereas the treatment groups received drugs dissolved in the vehicle. Tumor volume was measured with calipers every 2–3 days. In parallel, body weight was recorded.

Statistical analysis

All values were expressed as mean ± S.E.M. Statistical significance was evaluated using the Student’s two-tailed t-test for all in vitro experiments. The log-rank test was used to evaluate significance for the survival curve; p values < 0.05 were considered significant. NS: Non-significant, *p < 0.05, **p < 0.01, ***p < 0.001. In all cases, at least three independent experiments were conducted to warrant that the results were representative.

Funding Statement

This work was supported by NeOnc Technologies (Los Angeles, CA; TCC), the Wright Foundation grant (USC to HYC), the Garza Foundation (TCC), the Sounder Foundation (TCC), the National Natural Science Foundation of China (31670952 to LLT) and the China Scholarship Council (201606050090 to MO).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2023.2242696