A Novel Combination Approach Targeting an Enhanced Protein Synthesis Pathway in MYC-driven (Group 3) Medulloblastoma

Medulloblastoma is the most common pediatric brain tumor of cerebellar origin (1). Current therapies involve surgical removal of tumor, radiation, and intensive chemotherapy, but fail to cure approximately one-third of patients. Survivors often experience severe long-term side effects including neurocognitive deficits (2, 3). Clearly, development of more effective and less toxic therapies is needed.

Medulloblastoma has four major molecular subgroups: WNT (β-catenin), SHH (Sonic-hedgehog), group 3 (often termed MYC-driven medulloblastoma), and group 4 (heterogeneous genes). The WNT subgroup shows the best clinical prognosis and SHH and group 4 subgroups show intermediate prognosis (4–6). Group 3 medulloblastoma has the worst prognosis, with <50% survival. Aberrant MYC-driven medulloblastoma forms highly malignant tumors, often metastatic with resistance, even to multimodal therapy (7, 8). Thus, targeting MYC-driven oncogenic programs are critical to identifying effective therapeutics for these high-risk patients.

Dysregulation of protein synthesis by aberrant activation of oncogenic signaling pathways has emerged as a key mechanism for cancer progression and therapy resistance (9, 10). Studies show that the oncogenic effects of MYC are due to increased protein synthesis, leading to increased cell proliferation. A major immediate downstream effect of MYC activation is an increase in protein synthesis because it directly regulates and increases protein synthesis rates through the transcriptional control of multiple protein synthesis components, including mRNA translation factors and ribosome biogenesis (11–13). Protein synthesis is also enhanced by activating the mTOR-dependent phosphorylation of 4EBP1 (14). mTOR signaling is another master regulator of protein synthesis frequently deregulated in MYC-dependent cancers including medulloblastoma (15–17). mTOR also controls protein synthesis by phosphorylating the tumor suppressor eukaryotic translation initiation factor 4E (eIF4E) binding protein 1 (4EBP1) and ribosomal protein p70S6 kinase (S6K). mTOR-dependent phosphorylation of 4EBP1 blocks its ability to negatively regulate the translation initiation factor eIF4E, thus promoting eIF4E's ability to initiate protein translation (18). Importantly, MYC stimulates hyperactivation of eIF4E to drive tumorigenesis. mTOR also stabilizes the MYC protein by inducing MYC translation (19, 20). These studies suggest that the potential cross-talk between MYC- and mTOR-dependent mechanisms of translation reprogramming might be a cause of enhanced protein synthesis. Therefore, the MYC/mTOR axis might be an attractive therapeutic target for MYC-driven cancers.

MYC proteins are poor therapeutic targets because of their short half-lives and pleiotropic nature, but alternative strategies have been developed that permit MYC transcription to be regulated epigenetically, and target MYC by inhibiting bromodomain and extraterminal (BET)-containing proteins (21, 22). Preclinical studies have shown BET-bromodomain inhibition is a promising therapeutic strategy to target MYC-driven medulloblastoma and other cancers (22–24). BET protein inhibitors and PI3K–mTOR inhibitors, respectively, can facilitate therapeutic targeting of MYC and mTOR-dependent relevant protein synthesis pathways. Clinical experience of this approach is limited and evidence suggests that individually, such agents have poor antitumor efficacy, due to their insufficient activity on therapeutic targets (25–27). Combination strategies involving inhibitors targeting mTOR signaling and BET proteins may be required to achieve complete blockade of the enhanced protein synthesis pathway.

We evaluated the therapeutic potential of combined inhibition of MYC transcription and mTOR signaling against MYC-driven medulloblastoma. Combined inhibition significantly downregulated the expression of MYC/mTOR–associated key components of the protein synthesis pathway including MYC transcription, and exhibited synergistic anti-medulloblastoma effects in vitro and in vivo.

Medulloblastoma cell lines Daoy, D-283 and D-341 were purchased from ATCC. HD-MB03 medulloblastoma cell line was purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen. ONS-76 and D-425 medulloblastoma cell lines were provided, respectively, by Dr. Sutapa Ray (UNMC, Omaha, NE) and Dr. Sidharth Mahapatra (UNMC, Omaha, NE). Cell lines were authenticated by their respective resources using short tandem repeat profiling. All cell lines were also tested for Mycoplasma contamination using MycoSensor-PCR assay kit (Agilent Technologies). In this study, Daoy and ONS-76 were used as non-MYC–amplified (SHH-MB) cell lines, whereas, D-341, HD-MB03, D-425, and D-283 were used as group 3 medulloblastoma cell lines with MYC amplification (28). Cell lines were cultured and maintained using EMEM or RPMI1640 media supplemented with 10% FBS and 1% penicillin/streptomycin in a humidified incubator at 5% CO₂ and 95% air atmosphere at 37°C. Experiments were performed using less than 10 passages for each cell line. The BET protein inhibitor JQ1, PI3K–mTOR dual inhibitor BEZ235 and mTOR inhibitor temsirolimus were purchased from MedChemExpress LLC. OTX-015, an additional inhibitor of BET proteins, was purchased from Sellekchem LLC.

Cell growth analyses in medulloblastoma cells were performed using the MTT assay as described previously (29).

Twenty thousand medulloblastoma cells resuspended in neural stem cell medium were seeded in 6-well plates and allowed for 72 hours to form spheres. Spheres were treated with inhibitors for an additional 72 hours. Following treatment, aggregates of cells with >50 μm size were counted and imaged using an EVOS-Auto-Imaging System (Life Technologies). Spheres were also subjected to qRT-PCR and Western blot analyses for the expression of neural stem cell markers.

The ability of inhibitors to induce apoptosis in medulloblastoma cells was determined using an Annexin-V:FITC flow-cytometry assay kit (BD Biosciences) following the manufacturer's instructions. Induction of apoptosis by inhibitors was also determined using a Caspase 3/7 Activity Assay Kit (Promega) following the manufacturer's instructions. For cell-cycle analysis, the control and inhibitor-treated medulloblastoma cells were fixed in 75% ethanol and stained using a propidium iodide flow cytometry kit (Abcam).

Cells at 1 × 10⁴ density/well were plated in 96-well plates and treated with inhibitors for 24 hours. After treatment, media were removed and fresh media added containing O-propargyl-puromycin (OPP) for 30 minutes at 37°C. Cells were then stained with the 5-FAMAzide S, and processed for the detection of OPP according to manufacturer's instructions using Protein-Synthesis Assay Kit (Cayman Chemical).

Western blot analysis was performed using a standardized protocol (29). Primary antibodies used in this analysis included c-MYC, BRD2, BRD3, BRD4, and β-Actin (Santa Cruz Biotechnology), p-4EBP1 (Ser65), 4EBP1, p-eIF4E (Ser209), eIF4E, p-mTOR (Ser2448), mTOR, p-S6K (Thr421/Ser424), S6K, Nestin, and SOX2 (Cell Signaling Technology), and Cyclin-D1 and CD133 (BD Biosciences).

RNA was prepared using RNeasy (Qiagen) kit and 2 μg of total RNA was used for cDNA preparation using superscript verso enzyme kit (Promega). cDNA product was amplified in 10 μl reaction using SYBR-Green Super-mix and standard gene-specific primers (Applied Biosystems). All reactions were processed in a QuantStudio-3 PCR System and results analyzed by data analysis software (Applied Biosystems).

All siRNAs used were purchased from Santacruz-Biotechnology. Control (scrambled, sc-37007), c-MYC (sc-29226) and 4EBP1 (sc-29594) siRNAs (a pool of 3 target-specific siRNAs with 50 nmol/L) were transiently transfected into medulloblastoma cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Following 72 hours of transfection, cells were subjected to downstream analyses using Western blotting and MTT assays.

RNA from four medulloblastoma cell lines (DAOY, D-283, D-341, and HD-MB03) and inhibitor-treated HD-MB03 cells, was purified using the Qiagen RNeasy Kit. After confirming sequence grade quality of RNA using an Agilent 2100 Bioanalyzer, a RNA library was prepared using True-Seq RNA Sample Preparation V2 Kit and subjected to RNA sequencing using the Illumina NextSeq550 system in the UNMC Genomics Core Facility. Each sample was processed in triplicate. The original fastq reads were processed by a newly developed standard pipeline utilizing STAR as the aligner and RSEM as the tool for annotation and quantification at both gene and isoform levels. Using these reads, the normalized FPKM and TPM values for all the available genes were calculated.

All animal experiments were performed according to a UNMC Institutional Animal Care and Use Committee–approved protocol. For the subcutaneous xenograft study, 6- to 8-week-old NSG female mice from Jackson Laboratories were injected in the right-flank with 2.5 × 10⁵ HD-MB03 medulloblastoma cells suspended in 100 μL PBS and mixed 1:3 with Matrigel. Ten days post tumor injection, when tumor was palpable, the tumor-bearing mice were divided into six treatment groups (n = 5 each group) and treated twice a week for four weeks. Treatments included vehicle control (1:10 solution of DMSO:10% cyclodextrin) or JQ1 (50 mg/kg, i.p.), BEZ (25 mg/kg, i.p.), and TEM (10 mg/kg, i.p.) alone or combination of JQ1 with BEZ or TEM. Doses for these inhibitors were at ranges of achievable exposures in mice or humans (23, 30, 31). Tumor volume was assessed twice a week using the Biopticon TumorImager. When tumor volume approached 1 cm³, the mice were euthanized using CO₂ and tumor tissues were collected and processed for Western blotting and immunohistologic analyses for the expression of key proteins. Survival of the vehicle or different agent–treated mice was determined by the Kaplan–Meier method.

For the orthotopic xenograft study, 1 × 10⁵ HD-MB03 cells suspended in 8 μL neural stem cell media were injected into cerebella of NSG mice using the Kopf stereotaxic instrument and guidance. Ten days following injection, mice were randomly divided into four treatment (vehicle, JQ1, BEZ, and JQ1 + BEZ) groups (each group n = 7) and treated with inhibitors as indicated above. Following treatments, cerebellar tumors were examined by histologic (H&E) and magnetic resonance imaging (MRI) analyses. The details of H&E, MRI, and IHC analyses are provided in Supplementary Materials and Methods section.

All experiments were repeated at least three times and the mean and SE values calculated. Differences (p value) were calculated using Student t tests or ANOVA. IC₅₀ values of inhibitors for each medulloblastoma line were determined using GraphPad Prism V6 software. To determine inhibitor combinations/interactions, we used the Chou and Talalay CI method using CalcuSyn software (Biosoft). CI < 0.9 indicates synergism, 0.9–1.1 additivity, and >1.1 antagonism (32).

To identify overexpressed genes associated with enhanced translation machinery in MYC-driven medulloblastoma cells, we performed RNA-sequencing analyses to study differential transcriptional changes compared with non-MYC-driven medulloblastoma cells. The RNA-seq results identified over 2,000 genes in MYC-driven medulloblastoma cells compared with non-MYC medulloblastoma cells. Furthermore, gene-set-enrichment-analyses (GSEA) of targeted translation/protein synthesis pathway, using differentially expressed genes, revealed a very significant enrichment of translation associated genes in MYC-driven medulloblastoma cells (Supplementary Fig. S1A). We also observed the overexpression of a significant number of protein synthesis key genes that included ribosome biogenesis components, mRNA translation initiation/elongation factors, and mTOR signaling (Supplementary Fig. S1B). Using publicly available gene expression data for medulloblastoma tumors at R2-Genomics platform, we confirmed that the expression of key components of protein synthesis machinery positively and significantly correlated with MYC expression in group 3 medulloblastoma patient samples (Supplementary Table S1). These results suggest the activation/overexpression of protein synthesis components in MYC-driven medulloblastoma.

To investigate the MYC/mTOR axis as a therapeutic target in MYC-driven medulloblastoma, we used a pharmacologic approach using small-molecule inhibitors. We primarily used a recently developed BET protein inhibitor JQ1 to target MYC transcription. To target mTOR signaling, we used PI3K–mTOR dual inhibitor BEZ235. JQ1 and BEZ235 both have shown promising in vitro and in vivo antitumor activities against various malignancies. Both can cross blood–brain barrier (23, 26, 27, 30). Structurally similar inhibitors of BET protein and PI3K–mTOR are in multiple clinical trials in patients with advanced tumors (26, 31). In addition to JQ1/BEZ235, the clinically relevant BET inhibitor OTX-015 and mTOR inhibitor temsirolimus were included to confirm the therapeutic potential of targeting MYC/mTOR axis.

Using the MTT assay, we first examined the single agent antigrowth effects of JQ1 and BEZ235 against two non-MYC and four MYC-amplified medulloblastoma cell lines. The MTT results (Fig. 1A) showed a dose-dependent cell growth inhibition of all medulloblastoma lines by JQ1 and BEZ235 at low micromolar and nanomolar potencies, respectively. Analysis of IC₅₀ values of inhibitors in cell lines, demonstrated superior efficacy of each inhibitor with lower IC₅₀ values in MYC-amplified lines, compared with non-MYC–amplified lines (Fig. 1B). Superior efficacies of inhibiting MYC/mTOR targets were further confirmed using the alternative and clinically relevant inhibitors (temsirolimus and OTX-015) in medulloblastoma cell lines (Supplementary Fig. S2A and S2B).

We next investigated whether JQ1 and BEZ235 synergistically inhibited MYC-driven medulloblastoma cell growth. Four MYC-amplified and one non-MYC medulloblastoma cell lines were treated with JQ1 and BEZ235 alone or in combination, followed by cell growth analysis using the MTT assay. Combined treatment of JQ1/BEZ235 significantly inhibited growth of all MYC-amplified cell lines in a dose-dependent manner, compared with single-agent treatment (Fig. 1C). Although there was significant effect of this combination in non-MYC medulloblastoma (Daoy) cells treated with a higher dose, growth inhibition was lower than and not as consistent as in MYC-amplified medulloblastoma cells. The combination study was repeated by combining JQ1 with mTOR inhibitor temsirolimus using two MYC-amplified medulloblastoma cell lines. Consistently, a dose-dependent combination of JQ1 and temsirolimus significantly reduced growth of MYC-amplified medulloblastoma cells. The combination index (CI) analyses (Chou–Talalay method) confirmed that combination of JQ1 either with BEZ235 or temsirolimus exerted strong synergistic anticancer effects with the CI value ranges of 0.3 to 0.8 (Fig. 1D). We further confirmed this synergistic potential on growth inhibition of MYC-amplified cells using a combination of BET-protein inhibitor OTX-015 with mTOR inhibitors (BEZ235 or temsirolimus; Supplementary Fig. S2C). These data consistently demonstrated combined antitumor potential of MYC and mTOR inhibition in MYC-driven medulloblastoma cells.

Cisplatin is a standard chemotherapy used in combination with radiation for the treatment of medulloblastoma. We have shown that MYC-driven medulloblastoma cell lines are relatively chemoresistance to cisplatin, compared with non-MYC–amplified medulloblastoma cells (33). Results showed that inhibition of PI3K–mTOR by BEZ235 chemosensitized medulloblastoma cells to cisplatin in vitro and in vivo. Therefore, we wondered whether BET-protein inhibitor JQ1 also chemosensitized medulloblastoma cells. Our results demonstrated that JQ1 also chemosensitized the medulloblastoma cells to cisplatin, in a dose-dependent fashion (Supplementary Fig. S3). These data suggest that inhibition of MYC transcription and mTOR signaling not only have combinatorial antimedulloblastoma potential, but also chemosensitization, particularly in MYC-driven medulloblastoma cells that acquire resistance to chemotherapy in clinic.

To determine whether combination of JQ1 with BEZ235 or temsirolimus induces cell-cycle arrest and apoptosis, medulloblastoma cell lines were treated with a suboptimal dose of each inhibitor alone or combined and subjected to cell-cycle analysis using propidium iodide staining and apoptosis analyses using caspase-3/7 activity assay and Annexin-V staining. We used three MYC-amplified cell lines (D-283, D-341, and HD-MB03) and a non-MYC–amplified Daoy cell line. HD-MB03 and D-341 were used in subsequent studies as these lines have been reported as well-established group 3 medulloblastoma cell lines with MYC amplification status (28). The cell-cycle profile (Fig. 2A; Supplementary S4) across all MYC-amplified lines showed that while JQ1 and BEZ235 alone slightly arrested the cells in G₁, combined treatment with JQ1 and BEZ235 drastically increased the population of cells in G₁ phase compared with individual treatments. Results with apoptosis analyses (Fig. 2B) using a caspase-3/7 assay demonstrated a significantly enhanced activity of caspase-3/7 enzyme, with an increased apoptosis after combined treatment of JQ1 and BEZ235 in all cell lines, compared with single agents. There was significant induced apoptosis by JQ1 and BEZ235 alone compared with the DMSO solvent. Induction of apoptosis by the combination was confirmed using Annexin-V/PI double-positive staining in a representative MYC-amplified HD-MB03 cell line (Fig. 2C; Supplementary S5). We observed minimal and inconsistent combination effects on cell-cycle arrest and apoptosis induction in non-MYC–amplified Daoy cells (Fig. 2A and C), suggesting the MYC-dependent potency of combinations. Results for cell-cycle arrest and apoptosis induction by the combination were consistent when we combined JQ1 with TEM in MYC-amplified HD-MB03 cells (Fig. 2D and E; Supplementary S4 and S5). We further observed this antimedulloblastoma potential on cell-cycle arrest and apoptosis induction using a combination of OTX-015 either with BEZ235 or temsirolimus in HD-MB03 cells (Supplementary Fig. S6). Together, these results suggest that combination of JQ1 or OTX-015 with mTOR inhibitors reduces the cell growth of MYC-driven medulloblastoma cell lines by inducing G₁ cell-cycle arrest and apoptosis.

We next investigated the molecular mechanisms of combined therapy of JQ1/BEZ235 or JQ1/temsirolimus using D-341 and HD-MB03 cell lines. We determined the expression/activation of the key components of the MYC and mTOR signaling by Western blotting. We observed that an individual treatment of JQ1 and BEZ235 or temsirolimus in MYC-amplified cells resulted in significantly downregulated expression of phosphorylated key proteins (p-mTOR, p-S6K, p-4EBP1, and p-eIF4E) of the activated protein synthesis pathway, along with suppressed expression of MYC and BET-proteins (BRD2, BRD3, BRD4; Fig. 3A and B). Because increased protein synthesis correlates with increased cell proliferation and Cyclin D1 overexpression, we confirmed the reduced expression of Cyclin D1 protein upon treatment with inhibitors. Cotreatment of JQ1 with BEZ235 or temsirolimus significantly downregulated the expression of above key components of activated protein synthesis, including MYC expression, compared with individual treatments (Fig. 3A and B; Supplementary S7A). Consistent with previous findings, although we observed some effects of the inhibitors on key targets (BRD4/MYC/p-4EBP1/p-eIF4E) in non-MYC Daoy cells, the combination effect was not consistent as in MYC-driven cells (Supplementary Fig. S7B). These data suggest that inhibition of MYC transcription and mTOR signaling together, act cooperatively and downregulated the protein synthesis pathway, clarifying why this combined inhibition exerts the greatest antitumor effects in MYC-driven medulloblastoma.

We tested whether the inhibition of MYC/mTOR represented a global blockage of protein synthesis utilizing a cellpermeable, puromycin analogue OPP in JQ1 and BEZ235 alone or combined treated HD-MB03 and D-341 cells. Incorporation of OPP to nascent-polypeptide chains can be detected via coppercatalyzed click-chemistry using 5-FAMAzide and total protein synthesis can be detected based on 5FAM-fluorescence intensity. This analysis showed high fluorescence in control DMSOtreated cells, and as expected, low fluorescence signal when protein synthesis was blocked with cycloheximide (Fig. 3C). Although we observed strong inhibition of protein synthesis by JQ1 and BEZ235 alone, the combination further significantly blocked total protein synthesis with lowest fluorescent signal in both medulloblastoma lines, suggesting a cooperative impact of the MYC and mTOR inhibition on global protein synthesis.

We validated the combined therapeutic potential of targeting MYC/mTOR axis using a siRNA knockdown approach. We transiently transfected medulloblastoma cells with MYC-siRNA to target MYC and 4EBP1-siRNA to target mTOR. Our knockdown results in MYC-driven (HD-MB03) medulloblastoma cells showed that combined siRNAs of MYC and 4EBP1 not only significantly reduced the expression of MYC and 4EBP1, but also inhibited cell growth compared with individual siRNAs (Fig. 4A). In parallel, although we observed some effect of these siRNAs individually on their respective targets in non-MYC–amplified Daoy cell line, combined knocking down of MYC and 4EBP1 did not cause significant effects on the expression or cell growth, indicating no cooperation and interaction between MYC and mTOR due to non-MYC or low-MYC expressing nature of this cell line (Fig. 4B). These results further suggest that MYC and mTOR cooperate with each other in a MYC-dependent manner and can be synergistically targeted in medulloblastoma.

We next examined whether genetic inhibition of MYC and 4EBP1 by siRNA enhanced the antitumor potential of pharmacologic inhibitors BEZ235 and JQ1, respectively, in MYC-driven medulloblastoma cells. Downregulation of MYC significantly enhanced the effect of BEZ235 in decreasing expression of both MYC and phosphorylated-4EBP1 and also resulted in significantly reduced cell growth, compared with BEZ235 treatment or MYC-siRNA alone (Fig. 4C). Similarly, knockdown of 4EBP1 resulted in enhanced antimedulloblastoma efficacy of JQ1 on cell growth by further downregulating the expression of MYC and phosphorylated-4EBP1, compared with JQ1 treatment or 4EBP1-siRNA alone (Fig. 4D). These data combined with the targeting of MYC/mTOR using genetic/pharmacologic approaches, support our hypothesis of the effectiveness of targeting MYC and mTOR signaling together, in MYC-driven medulloblastoma.

We next investigated the effect of JQ1 and BEZ235 alone or combined in a spheroid model of MYC-driven medulloblastoma cells. Figure 5A shows a micrograph of sphere formation by JQ1 and BEZ235 alone or combined in HDMB03 cells. Treatment of JQ1 or BEZ235 alone significantly inhibited the sphere formation with the further reduction of spheres when combined (Fig 5B). We further tested the combination potential of JQ1/BEZ235 on the expression of these markers in HD-MB03 spheres by qPCR and Western blot analyses. Both JQ1 and BEZ235 efficiently inhibited the expression of neural stem cell markers (SOX2, CD133, Nestin) with reduced MYC expression at the mRNA (Fig. 5C) and protein (Fig. 5D) levels. Cotreatment of JQ1 and BEZ235 further decreased the expression of CD133 at both the transcript and protein levels. This cotreatment slightly decreased the expression of SOX2 and Nestin compared with single agents. These data suggest that combined inhibition of MYC transcription and mTOR signaling has antitumor effects on medulloblastoma spheres.

MYC plays key role in gene transcription at the global level. Therefore, it was of interest to investigate the effect of our combined therapy on global gene expression in MYC-driven medulloblastoma. We performed RNA-sequencing in HD-MB03 cells after treatment with JQ1 and BEZ235 alone or combined. Using log-2 fold-change cutoff, our gene expression data showed that JQ1 and BEZ upregulated the expression of 75 and 18 genes, respectively, with no sharing of genes with each other (Fig. 5E). However, the combination of JQ1 and BEZ235 upregulated the expression of 144 genes with higher potency, compared with JQ1 and BEZ235 alone. JQ1 and BEZ235 downregulated the expression of 985 and 135 genes, respectively, with significantly higher numbers, compared with upregulated genes. Of these downregulated genes, 50 genes were shared by JQ1 and BEZ235, suggesting that JQ1 and BEZ235 commonly repress target gene expression in MYC-driven medulloblastoma cells. Consistently, a combination of JQ1/BEZ235 downregulated the expression of 1,123 genes with the highest potential, compared with JQ1 or BEZ235-repressed genes alone (Fig. 5E). Together, our data suggest that the combination of JQ1 and BEZ235 synergistically modulated overall gene expression. Our RNA-seq data also showed that JQ1 was approximately 4 to 7 times more potent in modulating gene expression, when compared with BEZ235, suggesting its target specificity at the transcription level. In addition, we have also subjected our expression data to GSEA analyses. GSEA analyses revealed significant enrichment of MYC/mTOR signaling as targets of JQ1/BEZ combination treatment (Fig. 5F).

To evaluate the combined therapeutic potential of JQ1 with mTOR inhibitors against MYC-driven medulloblastoma in vivo, NSG mice were xenografted subcutaneously with HD-MB03 cells and treated with vehicle control or JQ1 and temsirolimus alone or the combination of JQ1 with BEZ235 or temsirolimus. As shown in Fig 6A, treatment with JQ1, BEZ235 and temsirolimus alone significantly suppressed tumor growth by approximately 50%–60%, compared with vehicle control. Combination of JQ1 with BEZ235 or temsirolimus further significantly suppressed tumor growth by approximately an additional 30%–40%, suggesting the antitumor combination potential of JQ1 with mTOR inhibitors against MYC-driven medulloblastoma in vivo. We also observed survival of the mice with a maximum 1,000 mm³ tumor size as endpoint for analyses. The Kaplan–Meier survival analyses showed that combination of JQ1/BEZ235 or JQ1/temsirolimus significantly prolonged the survival of xenografted mice when compared with single agents (Fig. 6B). There was significantly increased survival in the mice treated individually with these inhibitors, compared with vehicle control. In addition, treatment with these inhibitors alone or combined did not cause the significant changes in the total body weights and histopathology of vital organs between control and treatment groups (Supplementary Fig. S8), suggesting the tolerability of these therapies in mice.

We next determined the combined effect of inhibitors on the expression of key components of protein synthesis in xenografted tumors. Western blot analyses (Supplementary Fig. S9) showed that the combination of JQ1 with BEZ235 significantly decreased the expression levels of phosphorylated-4EBP1 and MYC compared with single agents. IHC analyses of MYC/Ki-67 staining further confirmed a significantly reduced proliferation in xenografted tumors treated with combination therapies, compared with single-agent therapy (Fig. 6C and D). These data suggest that combinations not only reduced tumor growth and increased survival, but also targeted relevant associated pathways and downstream molecules in subcutaneous xenografted tumors.

In addition to subcutaneous xenografts, we examined antitumor efficacy of combination therapy using orthotopic xenografts generated by intracerebellar injections of HD-MB03 cells into NSG mice. Ten days post tumor injection, mice (n = 7 per group) were treated with vehicle or JQ1 and BEZ235 alone or a combination of JQ1/BEZ235 similar to subcutaneous xenografts. Following treatments, we assessed tumor occurrence and growth in the cerebellum by MRI and histologic analyses. Histologic analyses showed that approximately 50% of the mice in treatment groups had significant tumor occurrence. Treatment of vehicle, JQ1, BEZ235, and combination of JQ1/BEZ235 showed tumor occurrence in 4, 4, 3, and 3 mice, respectively, out of 7 mice. We measured the tumor area in the cerebellum by Image J and observed that cotreatment with JQ1 and BEZ235 significantly reduced the tumor size in the cerebellum, compared with vehicle or JQ1 and BEZ235 alone (Fig. 7A). There was also a significant reduction of tumor size with the treatment of JQ1 and BEZ235 alone, compared with vehicle control. In addition, tumor shrinkage in treatment groups was further confirmed using MRI in mouse cerebellum (Fig. 7B). The tumor sizes calculated from MRI of the vehicle-, JQ1-, BEZ235-, and JQ1 + BEZ235-treated mice were 60.85, 1.87, 3.45, and 0.39 mm³, respectively. Together, in vivo results suggest a synergistic antimedulloblastoma potential of JQ1/BEZ235 not only against subcutaneous tumors but also orthotopic tumors.

The majority of patients with MYC-driven medulloblastoma face a paucity of effective therapies despite intensive and multimodal treatments (7). Treatment approaches based on targeting a MYC-driven tumor microenvironment are needed. MYC proteins are therapeutically undruggable because of complex protein structures and short half-lives (21). Attenuation of MYC-dependent pathways and MYC-regulatory activities might be an alternative to targeting MYC. A key and immediate downstream effect of MYC activation is to directly regulate the protein synthesis machinery. MYC is known to transcriptionally regulate the transcription of several key components of protein synthesis machinery including eIF4E oncogene and 4EBP1 tumor suppressor (11). Importantly, studies suggest that MYC enhances protein synthesis during tumorigenesis not only through transcriptional control but also by controlling mTOR-dependent translation (14). The translation of MYC protein itself, is further known to be regulated by mTOR/eIF4E signaling (20), suggesting targeting MYC function at both transcription and translation levels is a viable strategy. In this study, we show that combined inhibition of MYC transcription using a BET-protein inhibitor (JQ1), and translation machinery using mTOR signaling inhibitors (BEZ/TEM) displayed broad antitumor activities in MYC-driven medulloblastoma in vitro and in vivo. Our data provide the first preclinical evidence for a potential therapeutic role of inhibition of enhanced protein synthesis by combined targeting of MYC transcription and mTOR signaling against MYC-amplified (group 3) medulloblastoma.

Although the role of MYC-dependent enhanced protein synthesis has been extensively studied, particularly in lymphoid malignant microenvironment, there is no clear evidence of this interaction in medulloblastoma tumor microenvironment. In our studies, we observed overexpression and activation of the key components of protein synthesis pathway including mTOR signaling and MYC targets in MYC-amplified medulloblastoma cell lines, compared with non-MYC medulloblastoma cells, confirming MYC-dependent addiction of enhanced protein synthesis in medulloblastoma.

Recently, BET-bromodomain proteins as a cofactor have been shown to extensively regulate MYC transcription in various cancers. BET protein inhibitors have also shown success as preclinical anticancer agents against cancers including medulloblastoma and have demonstrated potential to selectively inhibit MYC transcription (21–23). Similarly, mTOR inhibitors have been shown to inhibit MYC protein stability at the translation level (33–36). Several small-molecule inhibitors of the BET bromodomain and mTOR signaling are in clinical trials in patients with advanced tumors (26, 27). In this study, we observed superior responsiveness of MYC-amplified medulloblastoma cell lines to BET and mTOR inhibitors, compared with non-MYC medulloblastoma cell lines. We have confirmed that these inhibitors synergistically suppress cell growth and induce apoptosis in MYC-driven medulloblastoma. The therapeutic potential of combined inhibition of MYC and mTOR were further confirmed by a genetic inhibition approach using target-specific siRNAs. Importantly, using subcutaneous/orthotopic mouse models of MYC-driven medulloblastoma, we have confirmed that JQ1 and mTOR inhibitors alone suppress medulloblastoma tumor progression, and the combination of JQ1 with mTOR inhibitors synergistically inhibits medulloblastoma progression. These findings suggest the feasibility of clinical investigation of our combined approach.

Targeting BET proteins has been shown to effectively block cancer cells from eliciting an adaptive signaling response to inhibitors of the PI3K pathway, which at least in some cases, can restore sensitivity to therapy (37). Studies suggest that there can be resistance to BET inhibitors through adaptive kinome reprograming via the activation of receptor tyrosine kinases and downstream signaling by PI3K–AKT–mTOR axis, which are compensatory prosurvival kinase networks (38, 39). This overcomes BET protein inhibition. Therefore, BET inhibitors may be thought of as rational combinatorial partners for reprogrammed oncogenic signaling such as PI3K–mTOR. Recently, combined BET protein and CDK2 inhibition has shown synergistic antitumor efficacy against MYC-driven medulloblastoma in vitro and in vivo. For example, BET protein inhibitor JQ1 and CDK2 inhibitor milciclib were shown to destabilize MYC protein. Both JQ1 and milciclib significantly reduced tumor growth and prolonged survival in medulloblastoma preclinical models (40). Our results suggest that JQ1 not only has antitumor potency in combination with CDK2 inhibitor, but also efficiently inhibits tumor cell growth and prolonged survival in medulloblastoma-bearing animals when combined with mTOR inhibitors.

Resistance to mTOR inhibitors is frequent, due to feedback activation of PI3K upstream receptor tyrosine kinase, providing the rationale for combined inhibition of PI3K/mTOR to efficiently block the mTOR signaling (41, 42). Therefore, we used PI3K–mTOR dual inhibitor BEZ235 that can overcome this feedback activation and efficiently target the mTOR-driven oncogenicity. Because BEZ235 has yet to be incorporated into a clinical setting (43), as an alternative, we used temsirolimus, an FDA-approved mTOR inhibitor alongside BEZ235. Temsirolimus is currently under clinical trials in various pediatric cancers including medulloblastoma (44). We observed similar antimedulloblastoma activities of this inhibitor as of BEZ235, suggesting targeting mTOR by either inhibiting PI3K–mTOR, or selective mTOR, could be beneficial in designing therapies for MYC-driven medulloblastoma.

Translational control of protein synthesis also has a role in the maintenance of self-renewal and pluripotency (45–47). Furthermore, activation of both MYC and mTOR signaling have been shown to play key roles in the biology of cancer “stem” cells, leading to tumor relapse and drug resistance in various malignancies including medulloblastoma (48, 49). Medulloblastoma cells express neural stem cell markers and have ability to form colonies/spheres (50). Our preliminary results with the antimedulloblastoma potential of MYC and mTOR inhibition on spheres indicate that individual or combined inhibition of MYC transcription and mTOR signaling might target cancer stem cells with the potential of reducing recurrence of medulloblastoma.

In summary, MYC-driven medulloblastoma cells displayed increased activation and overexpression of the protein synthesis machinery. Targeting this enhanced protein synthesis pathway with combined inhibition of MYC transcription and mTOR translation signaling by small-molecule inhibitors (Fig. 7C), demonstrated significant preclinical potential in reducing MYC-driven medulloblastoma cell growth, inducing apoptosis with prolongation of survival in xenograft models. The findings from this study warrant further preclinical evaluation in patient-derived xenografts, to translate these approaches to the clinical setting.

No potential conflicts of interest were disclosed.

Conception and design: N.K. Chaturvedi, J.G. Sharp, H. Band, S.S. Joshi, D.W. Coulter

Development of methodology: N.K. Chaturvedi, M.J. Kling, H. Band

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.M. McIntyre, Y. Liu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N.K. Chaturvedi, M.J. Kling, C.N. Griggs, V. Kesherwani, Y. Liu, D.W. Coulter

Writing, review, and/or revision of the manuscript: N.K. Chaturvedi, M.J. Kling, V. Kesherwani, M. Shukla, S. Ray, T.R. McGuire, J.G. Sharp, H. Band, S.S. Joshi, D.W. Coulter

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N.K. Chaturvedi, M.J. Kling

Study supervision: N.K. Chaturvedi, S. Ray, S.S. Joshi

This work was financially supported by the State of Nebraska through the Pediatric Cancer Research Grant Funds (LB905; awarded to D.W. Coulter). The authors thank the Child Health Research Institute and the Fred Pamela Buffet Cancer Center supported Core Facilities at UNMC. The authors also thank the Genomics, Flow-Cytometry, Bio-imaging, and Tissue-Science Core Facilities at UNMC.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.