Atypical teratoid/rhabdoid tumors (AT/RT) are aggressive infantile brain tumors with poor survival. Recent advancements have highlighted significant molecular heterogeneity in AT/RT with an aggressive subgroup featuring overexpression of the MYC proto-oncogene. We perform the first comprehensive metabolic profiling of patient-derived AT/RT cell lines to identify therapeutic susceptibilities in high MYC-expressing AT/RT.
Metabolites were extracted from AT/RT cell lines and separated in ultra-high performance liquid chromatography mass spectrometry. Glutamine metabolic inhibition with 6-diazo-5-oxo-L-norleucine (DON) was tested with growth and cell death assays and survival studies in orthotopic mouse models of AT/RT. Metabolic flux analysis was completed to identify combination therapies to act synergistically to improve survival in high MYC AT/RT.
Unbiased metabolic profiling of AT/RT cell models identified a unique dependence of high MYC AT/RT on glutamine for survival. The glutamine analogue, DON, selectively targeted high MYC cell lines, slowing cell growth, inducing apoptosis, and extending survival in orthotopic mouse models of AT/RT. Metabolic flux experiments with isotopically labeled glutamine revealed DON inhibition of glutathione (GSH) synthesis. DON combined with carboplatin further slowed cell growth, induced apoptosis, and extended survival in orthotopic mouse models of high MYC AT/RT.
Unbiased metabolic profiling of AT/RT identified susceptibility of high MYC AT/RT to glutamine metabolic inhibition with DON therapy. DON inhibited glutamine-dependent synthesis of GSH and synergized with carboplatin to extend survival in high MYC AT/RT. These findings can rapidly translate into new clinical trials to improve survival in high MYC AT/RT.
Atypical teratoid/rhabdoid tumors (AT/RT) are the most common malignant brain tumors of infancy. Despite intensive chemotherapy and radiation, median survival remains less than 1 year. Recent advancements have identified significant molecular heterogeneity in AT/RT and an especially aggressive subgroup with high expression of the MYC proto-oncogene. We performed the first comprehensive metabolic profiling of patient-derived AT/RT cell lines and identified increased glutamine metabolism in high MYC-expressing AT/RT. We show that the glutamine metabolic inhibitor 6-diazo-5-oxo-L-norleucine (DON) selectively targets high MYC AT/RT and combines synergistically with carboplatin to extend survival. We also identify potential biomarkers to predict AT/RT sensitivity to DON therapy. These findings can rapidly translate to new clinical trials aimed at improving survival in AT/RT.
Atypical teratoid/rhabdoid tumors (AT/RT) are the most common malignant brain tumors of infancy (1). Standard treatment involves intensive radiation and chemotherapy that causes severe morbidity but leads to a median survival of just 6–11 months (2). New precision therapies are needed to reduce side effects and improve survival.
Preclinical studies have largely developed therapies directed at AT/RT as a molecularly homogenous tumor (3–6) given exome studies that identified just one recurring genetic mutation in the SMARCB1 gene (7). Recent advancements in our understanding of AT/RT have emphasized intertumor molecular heterogeneity. Large scale epigenetic analyses have defined three molecularly distinct subgroups of AT/RT (8, 9). Each subgroup has a unique pattern of proto-oncogene expression that distinctly contributes to AT/RT growth and tumorigenicity. This molecular heterogeneity must be addressed to successfully develop new treatment strategies for AT/RT.
These recent bioinformatics studies identified overexpression of the MYC proto-oncogene in a subgroup of AT/RT. Johann and colleagues described a MYC subgroup of AT/RT based on the proto-oncogene's central role in driving this subgroup's tumorigenicity (8). Torchia and colleagues identified a more heterogeneous pattern of MYC expression largely centered on the subgroup 2B (9). Preliminary clinical studies suggested that high MYC AT/RT have an especially poor prognosis (10). This aggressive phenotype matches other high MYC- or MCYN-expressing tumors such as neuroblastoma and medulloblastoma (11–13).
We develop a targeted therapeutic strategy directed at high MYC AT/RT. Previous studies have demonstrated the dependence of MYC-expressing AT/RT on MYC for survival (14). Direct inhibition of MYC is challenging given its large, flat, featureless protein–protein binding sites (15). Preclinical studies using bromodomain inhibitors to target high MYC AT/RT and medulloblastoma have revealed a modest improvement in survival (14, 16). We aim to identify MYC-dependent adaptations in AT/RT that can be targeted downstream of MYC to further improve survival in AT/RT.
MYC is a known regulator of cancer cell metabolism (17, 18). Rapidly dividing, aggressive cancer cells have unique metabolic needs to support high energy demands, protect from excessive reactive oxygen species, and meet the demand for increased utilization of nucleic acids and amino acids (19). Increased expression of the MYC proto-oncogene regulates adaptations in cancer cell metabolism to better meet these demands (18). We hypothesized that MYC-specific adaptations in cancer cell metabolism can be targeted to improve survival in high MYC AT/RT.
We explore the efficacy of 6-diazo-5-oxo-L-norleucine (DON) to precisely target high MYC AT/RT. DON is a glutamine analogue that selectively inactivates glutamine-utilizing reactions by binding competitively to glutamine active sites and forming a covalent adduct to irreversibly inhibit enzymes (20, 21). DON has previously been tested in a pediatric phase I clinical trial and was well tolerated without reaching a MTD (22). Other glutamine inhibitors like CB-839 are not likely to be effective in treating brain tumors due to poor central nervous system penetration (23). DON has never been tested specifically in MYC-expressing tumors.
In this study we perform the first comprehensive metabolic profiling of human-derived AT/RT cell lines to identify MYC-altered metabolic pathways that can be targeted to improve survival in AT/RT. These studies explore the efficacy for a novel targeted therapeutic strategy that addresses the recently identified molecular heterogeneity in AT/RT.
Materials and Methods
Cell lines and cell culture
AT/RT cell lines, BT-12 (RRID:CVCL_M155), BT-16 (RRID:CVCL_M156), BT-37 (RRID:CVCL_JL57), and CHLA-06-ATRT (ATCC catalog no. CRL-3038, RRID:CVCL_AQ42) are described previously (6, 24, 25). BT-12 and BT-16 cells were derived by Dr. Peter J. Houghton (St. Jude Children's Research Hospital, Memphis, TN) and Dr. Jackie Biegel (The Children's Hospital of Philadelphia, Philadelphia, PA), and obtained from the Children's Oncology Group cell repository (25). BT-37 cells were derived from a human xenograft at St. Jude Children's Research Hospital (Memphis, TN; ref. 26). CHLA-02-ATRT (ATCC catalog no. CRL-3020, RRID:CVCL_B045), CHLA-04-ATRT (ATCC catalog no. CRL-3036, RRID:CVCL_0F38), and CHLA-05-ATRT (RRID:CVCL_AQ41) were obtained from Children's Hospital of Los Angeles (24, 27). CHLA-266 (RRID:CVCL_M149) cells were obtained from the Children's Oncology Group cell repository (28, 29). Cell lines were confirmed through STR testing by the Johns Hopkins Genetics Resources Core Facility. CHLA-02-ATRT, CHLA-04-ATRT, and CHLA-266 cell lines were cultured in the same media as CHLA-06-ATRT as described previously (26–28). All cell lines were verified as Mycoplasma free through frequent PCR testing, last performed on May 1, 2019. All cells were grown in a humidified 37°C chamber with 5% CO2. Cell lines were used between passage 1 and 20 from collection to use in experiments.
DON (Bachem, catalog no. 4006579.0500) was dissolved in purified water for in vitro experiments as was carboplatin (Selleckchem, catalog no. S1215). DON and carboplatin were dissolved in PBS for in vivo experiments.
Cells were plated on a 96-well plate, with 2,000 cells/well for cell lines BT-12, BT-16, BT-37, CHLA-05-ATRT, and CHLA-06-ATRT. CHLA-02-ATRT and CHLA-04-ATRT were grown at a density of 200,000 cells/well in a 6-well plate then plated in a 96-well plate. All cells were plated in triplicate in all conditions. Mitochondrial activity was determined with the MTS 3-(4, 5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) from Promega. MTS reagent (40 μL) was added to 200 μL of media containing cells and/or drugs per well and incubated for 1 hour at 37°C with 5% CO2 then measured with Epoch Micro-Volume Spectrophotometer Plate Reader System (BioTek) at 490 nm absorbance. Measurements were taken at day 0, 3, and 5.
Cell viability assays were completed with MUSE Count and Viability Assay Kit (Millipore MCH100102). Cells were prepared in a uniform cell suspension for counting, and cells were incubated with Muse Count and Viability Reagent for 5 minutes. Cell viability was analyzed on Muse Cell Analyzer.
Western blotting and antibodies
Western blots were performed by extracting cell pellets or tumor samples using RIPA buffer. Protein concentrations were quantified using Bradford assay. Antibodies were used according to the manufacturer's instructions: MYC (#5605), C-PARP (#9541), p-H2A.X (Ser139; #9718; from Cell Signaling Technology), and β-Actin (#47778; Santa Cruz Biotechnology, Inc.). Densitometry was performed using ImageJ v1.440 software as described previously (6).
Metabolomics and stable isotope analysis
Stable isotope experiments were completed after cells were treated for 24 hours with DON and incubated for 2 hours with 4 mmol/L glutamine (13C5, 15N2, 99% purity) from Cambridge Isotope (catalog no. CNLM-1275-H-0.5). Cell samples were washed with ice-cold PBS, pelleted, and metabolites extracted with chilled 80% HPLC grade methanol as described previously (30). Analyses were completed on an Agilent 1290 liquid chromatography system coupled to an Agilent 6520 quadrupole time of flight mass spectrometer. The mass spectrometer, equipped with a dual electrospray ionization source, was run in negative ion and then positive ion mode. The scan range was 50–1,600 m/z. The source settings consisted of: drying gas flow rate: 11 L/min; nebulizer: 40 pounds per square inch gauge; gas temp: 350°C; capillary voltage: 3,000 V (neg), 2,500 V (pos). Metabolite identification was determined using MS-MS, with fragments compared against the Agilent Metlin Metabolomics Database and Library.
LC/MS data were analyzed using Agilent Qualitative Analysis B.07.00, El-MAVEN (Elucidata), and Metabolomic Analysis and Visualization ENgine (MAVEN) as described previously (31).
Viral infections and induction of MYC
Lentivirus with MYC plasmid (Addgene plasmid 17758) was subcloned in pWPI (Addgene 12254) as described previously (32). To produce the required lentiviral particles, 293T cells were transfected with vesicular stomatitis virus G envelope plasmid, delta 8.9 gag/pol plasmid, and the plasmid of interest using Fugene Transfection Reagent (Roche) per the manufacturer's instructions as described previously (24). Supernatants were collected at 48, 72, and 96 hours and stored at 4°C. For infection of AT/RT cell lines, adherent cells and neurospheres were dissociated into single cells with gentle titration and Accutase, respectively, and incubated with the lentivirus-containing MYC plasmid, or empty pWPI vector. All experiments were performed within 14 days of lentiviral infection.
For glutathione (GSH) rescue experiments, 4 mmol/L cell permeable GSH-reduced ethyl ester (Sigma-Aldrich G1404) or 5 mmol/L N-acetyl-cysteine (Sigma-Aldrich A7250) were titrated to a pH of 7. Cells were treated just before DON and carboplatin treatments. Similarly l-Buthionine-sulfoximine (BSO; Sigma-Aldrich B2515) was used at 10 μmol/L and titrated to pH of 7 as described previously (30).
Intracranial xenograft tumors
For animal care and anesthesia, “Principles of Laboratory Animal Care” (NIH publication No. 86-223, revised 1985) was followed, using a protocol approved by the Johns Hopkins Animal Care and Use Committee, in compliance with United States Animal Welfare Act regulations and Public Health Service Policy. Intracranial xenografts were produced in anesthetized animals as described previously (6). Four-to-6-week old female Nu/Nu mice were obtained from Charles River Laboratories. Intracranial injection Burr holes were created from an 18-gauge bevel tip needle. A total of 2.5 × 105 (BT12) or 1 × 104 (CHLA-06-ATRT) of viable cells were suspended in 5 μL of media and injected in the right striatum, 3 mm anteroposterior, 2 mm mediolateral, 3 mm dorsoventral. Injections were made with a Hamilton syringe with a needle guard of 3 mm. Incisions were closed with an Autoclip System (Fine Science Tools, catalog no. 12020-00).
The day after injection, mice were treated with vehicle (PBS) or 30 mg/kg of DON. Treatments were administered via intraperitoneal injections once weekly. The combination treatment consisted of four groups: vehicle control, 30 mg/kg DON, 50 mg/kg carboplatin i.p. injection, and combination of DON and carboplatin. This dose of DON in mice converts to a human dose well below the projected MTD in the pediatric phase I trial of DON, while the carboplatin dose in mice converts to an equivalent human dose that has been well tolerated in human clinical trials (22, 33, 34). Animals were sacrificed upon distress, poor grooming, or loss of 20% body mass.
Statistical analysis was conducted with GraphPad Prism (GraphPad Software) or Excel (Microsoft). Single group comparisons were performed with Student t test, and multiple group comparisons were performed with one-way ANOVA. P < 0.05 was considered significant. Synergy calculations were conducted per guidelines described by Chou and Talalay method (35). Data were processed with CompuSyn for drug combinations and for general dose-effect analysis (http://www.combosyn.com/).
Primary brain tumor samples
Brain tumor samples were obtained by the Johns Hopkins University Department of Pathology (Baltimore, MD) and St. Jude Children's Research Hospital (Memphis, TN) with institutional review board approval. All samples were deidentified.
IHC was performed on tissue microarrays as described previously (6). For each tumor, 2 cores measuring 0.6 mm diameter were used per array. Antigen unmasking was performed with hot sodium citrate pH 6.0 × 30 minutes followed by 3% hydrogen peroxide × 30 minutes and 0.4% TBS × 30 minutes. Slides were blocked with 4% normal goat serum × 1 hour and were incubated overnight at 4°C with primary antibodies. The following primary antibody was used: 1:50 c-MYC (Abcam #33072). AT/RT tissue cores were scored by a neuropathologist (C.G. Eberhart) using H-scores (H) (0–255), which were obtained by multiplying the intensity of stain (0, no stain; 1, weak stain; 2, strong stain) by percentage (0–100) of neoplastic cells showing the staining intensity. Tumors were scored for MYC as being negative (score of 0), low (<40), intermediate (40–100), or high (>100).
IHC of primary AT/RT identified a subgroup of tumors with high MYC expression
Large scale genetic and epigenetic analyses of primary AT/RT identified a subgroup of tumors with high MYC expression (8, 9). To investigate whether this genetic heterogeneity translates to a similar pattern of MYC protein expression, we performed IHC on 22 primary human AT/RT. Quantification of protein expression was completed by H-score analysis evaluated by a neuropathologist (C.G. Eberhart). Similar to previously reported genetic analyses, 32% of human AT/RT exhibited strong staining for MYC, 27% exhibited moderate staining, 32% mild staining, and 9% did not stain for MYC (Fig. 1A).
We next evaluated eight patient-derived AT/RT cell lines and identified similar patterns of MYC protein expression with 50% of tumors expressing moderate to high levels of MYC (Fig. 1B). Of note, CHLA-266 expressed low levels of MYC protein; however it has previously been categorized in the AT/RT subgroup 2B that generally expresses high levels of MYC protein (9). This discrepancy may be due to growing these cells in low-serum media compared with high-serum media in the prior publication. MYC expression had no significant impact on cell proliferation (MTS assay after 5 days of cell growth, P = 0.53; Supplementary Fig. S1). These studies confirm that the genetic and epigenetic subgrouping correlates with MYC protein expression in primary human tumors. Human-derived cell lines express similar heterogeneity in MYC expression making them good models to study directed therapy against high MYC AT/RT.
Metabolic profiling of AT/RT cell lines identified an increased dependence on glutamine for growth and survival in high MYC AT/RT
To identify unique pathways that can be targeted in high MYC AT/RT, we performed unbiased metabolic profiling of each of 8 patient-derived AT/RT cell lines. Metabolites were extracted from each cell line and separated in an ultra-high performance liquid chromatography mass spectrometry. Partial Least Squares-Discriminant Analysis (PLS-DA) identified a unique metabolic profile in high MYC cell lines compared with low MYC cell lines (Fig. 2A). This analysis indicates that there are global metabolic variations between high and low MYC-expressing AT/RT.
Further evaluation of these metabolic pathways identified amino acid metabolism, tricarboxylic acid (TCA) cycle intermediaries, and GSH metabolism as the most altered metabolic pathways between low and high MYC AT/RT cell lines (Fig. 2B). In particular, glutamine intensity is greater in high MYC cell lines while glutamate intensity is lower (Fig. 2C). The glutamine:glutamate ratio is significantly elevated in high MYC AT/RT compared with low MYC cell lines (P < 0.001; Fig. 2D). The glutamine:UDP-glucose ratio is also significantly increased in high MYC AT/RT (P < 0.001; Fig. 2E). The change in glutamine:UDP-glucose suggests that high MYC cell lines increase the utilization of glutamine greater than glucose to meet metabolic demands. These studies support that AT/RT cell metabolism is altered in high MYC AT/RT leading to unique patterns of glutamine utilization.
We next titrated glutamine from 0 mmol/L to normal growth media concentrations to determine how altered metabolism in high MYC AT/RT affected dependence on glutamine for cellular proliferation and survival. Low MYC AT/RT achieved normal cellular density growing for 5 days in 0.5 mmol/L glutamine while high MYC AT/RT had reduced cellular density after growing in glutamine concentrations as high as 2 mmol/L (Fig. 2F and G; Supplementary Fig. S2). These studies demonstrate that high MYC AT/RT is more dependent on glutamine for cell growth and survival.
DON selectively targets high MYC AT/RT
Because of this relative dependence on glutamine in high MYC AT/RT, we hypothesized that pharmacologic inhibition of glutamine metabolism is an effective strategy to selectively target high MYC AT/RT. To evaluate this hypothesis, we tested the in vitro efficacy of treating AT/RT with the glutamine analogue, DON. DON 15 μmol/L significantly slowed AT/RT cell growth in high MYC cell lines (15 μmol/L DON vs. vehicle control; P < 0.001; Fig. 3A) while low MYC AT/RT was relatively resistant. The median concentration to decrease 50% of cell growth (IC50) was 7.84 μmol/L in high MYC AT/RT compared with 33.5 μmol/L in low MYC AT/RT (P < 0.001; Supplementary Fig. S3A). Similarly, DON induced high levels of apoptosis as determined by C-PARP expression in high MYC AT/RT while having little effect on low MYC cell lines (Fig. 3B; Supplementary Fig. S3B). Lentiviral-mediated forced expression of MYC in the low MYC AT/RT cell lines BT-37 and CHLA-05-ATRT led to increased apoptosis in response to DON treatment (Fig. 3C and D). These studies demonstrate that high MYC expression sensitizes AT/RT to the inhibition of glutamine metabolism with DON therapy.
We next established orthotopic mouse models of AT/RT using the high MYC cell lines BT-12 and CHLA-06-ATRT. We treated mice with 30 mg/kg DON by intraperitoneal injection once weekly and compared survival with vehicle control–treated mice. DON was well tolerated (Supplementary Fig. S4) and nearly doubled median survival in BT-12 orthotopic mouse models from 21 to 36 days (P = 0.0027; Fig. 3E). In the aggressive CHLA-06-ATRT–derived orthotopic mouse model, DON treatment also significantly improved survival (P = 0.0189; Fig. 3F).
DON depletes AT/RT cells of intracellular GSH
While DON treatment extended survival in high MYC orthotopic mouse models of AT/RT, mice still died from progressive disease. To design combination therapy that will complement and extend these survival benefits, we performed metabolic flux analysis of cells incubated with isotopically labeled glutamine (13C5, 15N2) to determine how DON affects glutamine metabolic pathways. We treated the high MYC AT/RT cell lines BT-12 and CHLA-06-ATRT with DON versus vehicle control for 24 hours. We then spiked media with isotopically labeled glutamine for 2 hours and determined the fate of labeled carbons and nitrogen. DON inhibited the conversion of glutamine to glutamate, which in turn decreased the production of intracellular GSH (Fig. 4A).
DON treatment led to a significant decrease in glutamate (M+6; P = 0.0002; Fig. 4B), as well as a decrease in GSH (M+6; P < 0.0001; Fig. 4B) in BT-12. Similarly, we found decreases in glutamate (M+6) production (P < 0.0001; Fig. 4B) and GSH (M+6; P = 0.0004; Fig. 4B) in CHLA-06-ATRT. Both glutamate (M+6) and GSH (M+6) are more than 50% reduced in these cell lines indicating that glutaminase activity is reduced with DON treatment.
GSH is known to protect cells from platinum chemotherapy–induced DNA damage and apoptosis (36, 37). GSH binds platinum chemotherapies and expels the GSH–platinum conjugate from the cell before it reaches the nucleus and causes DNA damage (Fig. 4C). Therefore, we hypothesized that the selective depletion of intracellular GSH in DON-treated high MYC AT/RT will block this natural defense and sensitize cancer cells to platinum therapies.
DON combines synergistically with carboplatin to target high MYC AT/RT
Platinum-based agents are an important component of the backbone of AT/RT therapy (2). We next tested whether DON-induced depletion of GSH sensitized high MYC AT/RT to carboplatin. DON effectively combined with carboplatin to slow AT/RT cell growth compared with each medication alone and vehicle control (CHLA-06-ATRT, P < 0.001; BT12 P < 0.001; Fig. 5A). This combination also decreased the number of viable cells more than each medication alone and vehicle control (CHLA-06-ATRT, P < 0.001; BT12, P < 0.01; CHLA-02-ATRT, P < 0.01; Fig. 5B). Combination therapy significantly increased tumor cell killing compared with each medication alone and vehicle control (CHLA-06-ATRT, P < 0.001; BT12, P < 0.01; CHLA-02-ATRT, P < 0.01; Fig. 5C) and induced higher rates of apoptosis (CHLA-06-ATRT, P < 0.001; BT12, P < 0.01; CHLA-02-ATRT, P < 0.01; Supplementary Fig. S5). Formal synergy testing demonstrated that DON and carboplatin act synergistically to induce apoptosis in high MYC AT/RT (Supplementary Fig. S6).
To test our hypothesis that depletion of GSH is responsible for the increased sensitivity to carboplatin, we added either GSH (reduced ethyl ester) or the GSH precursor (N-acetyl-cysteine) to combination DON and carboplatin. Both agents rescued the effect of DON combined with carboplatin on DNA damage (pH2AX) and apoptosis (cPARP) in BT-12 and CHLA-06-ATRT cell lines (Fig. 5D). We then combined carboplatin with BSO, which is also known to deplete intracellular GSH (38), and replicated the impact of DON combined with carboplatin on DNA damage and cell death (Fig. 5D).
DON combines with carboplatin to improve survival in orthotopic models of high MYC AT/RT
To determine the survival benefit of combining DON with carboplatin therapy, we established orthotopic mouse models of AT/RT with the high MYC cell lines BT-12 and CHLA-06-ATRT. We treated mice in four groups, vehicle control, carboplatin 50 mg/kg i.p. injection administered weekly, DON 30 mg/kg i.p. weekly, and carboplatin 50 mg/kg i.p. weekly combined with DON 30 mg/kg i.p. weekly. Carboplatin had little effect on survival while DON significantly extended survival compared with vehicle control. Combination of carboplatin with DON treatment of BT-12 orthotopic mouse models was well tolerated (Supplementary Fig. S4A) and significantly extended median survival from 33 days in the DON-treated mice to 45 days in the DON combined with carboplatin–treated group (combination therapy vs. control, P = 0.0021; combination vs. carboplatin, P = 0.0033; combination vs. DON, P = 0.0408 by log-rank test adjusted using the Bonferroni–Sidak method of multiple comparisons; Fig. 6A). Combination treatment in CHLA-06-ATRT orthotopic mouse models was also well tolerated (Supplementary Fig. S4B) and extended median survival (combination therapy vs. control, P = 0.0003; combination vs. carboplatin, P = 0.0003; combination vs. DON P = 0.0253 by log-rank test adjusted using the Bonferroni–Sidak method of multiple comparisons; Fig. 6B). Combination of DON with carboplatin was also measured in BT-12 flank tumors for accurate measurements of growth rate and change in volume over the course of therapy. DON combined with carboplatin slowed cell growth over the course of treatments and led to a significantly smaller final tumor mass (Supplementary Fig. S7).
The effect of DON and carboplatin in combination was additionally evaluated in vivo by Western blot for C-PARP expression (Fig. 6C). Mice were treated with a single dose of vehicle control, DON 30 mg/kg i.p., carboplatin 50 mg/kg i.p., or combination of DON 30 mg/kg i.p. with carboplatin 50 mg/kg i.p. Brains were harvested 72 hours after treatment. DON treatment induced apoptosis in the tumor while normal cerebral cortex was unaffected (Fig. 6C). Combination of DON with carboplatin induced higher levels of apoptosis compared with DON, carboplatin, or vehicle control. Combination therapy did not induce apoptosis in normal cerebral cortex (Fig. 6C).
AT/RTs are aggressive cancers with very poor survival (2). Tremendous advancements have been made in our understanding of the molecular heterogeneity of AT/RT (8, 9), but our therapeutic tools are still in the infancy of development. Past directed therapies have failed to meaningfully improve AT/RT survival as median survival remains less than 1 year from diagnosis (1). We developed a novel combination therapy that addresses this newly identified molecular heterogeneity to improve survival in high MYC AT/RT.
We previously demonstrated that the cell reprogramming factors LIN28A and LIN28B are highly expressed in AT/RT primary tumors and are key drivers of AT/RT tumorigenesis (24). LIN28 influences proto-oncogene expression through its primary role in regulating the Let-7 family of miRNAs (39, 40). We have shown that high levels of LIN28 expression lead to high KRAS and mTOR expression in AT/RT (6, 24). Pharmacologic inhibition of these LIN28-regulated oncogenic pathways improves survival in preclinical models of AT/RT. Similarly, through this regulation of Let-7 miRNAs, LIN28 is a known regulator of MYC expression (41, 42).
MYC is involved in the maintenance of cancer stem cells and drives tumor cell growth, proliferation, and tumorigenesis (43, 44). High MYC-expressing cancer cells have increased energy demands and require high biomass accumulation to support this rapid cellular growth (18). MYC regulates adaptations in cancer cell metabolism to meet these high demands (18). We hypothesized that MYC-dependent adaptations in cellular metabolism may be targeted to selectively kill high MYC-expressing AT/RT.
Little is known about the impact of MYC expression on AT/RT cell metabolism. We report here the first comprehensive metabolic profiling of patient-derived AT/RT cell lines. Unbiased mapping of metabolite intensities identified a unique metabolic profile in high MYC cell lines. Amino acid metabolism is one of the pathways most affected by MYC protein expression. Significantly altered ratios of glutamine:glutamate and glutamine:UDP-glucose suggests that glutamine metabolism is altered in high MYC AT/RT. In fact, high MYC cell lines are unable to grow in low glutamine concentrations while low MYC cell lines grow relatively normal. These data suggest that MYC expression drives a reliance on glutamine to meet AT/RT cancer cell's metabolic demands.
These results are in line with previous studies demonstrating that MYC induces transcription of genes involved in glutamine metabolism (45, 46). MYC increases the expression of the ASCT2 glutamine transporter that carries glutamine into the cell. MYC also increases glutaminase protein levels leading to more rapid glutamine catabolism. Glutamine metabolism can produce ATP or function as a precursor for the synthesis of proteins, other amino acids, nucleic acids, and GSH (18). The rapidly dividing cancer cell may preferentially use glutamine to more efficiently meet high energy demands while protecting from resulting reactive oxygen species (47).
We exploit this dependence on glutamine by targeting high MYC AT/RT with the glutamine analogue, DON. DON has previously been used in phase I clinical trials in pediatrics and has been proven to be well tolerated without reaching a MTD (22). While DON is not currently commercially produced, there is significant interest in developing targeted prodrugs of DON (48). These prodrugs were designed to circulate intact and inert in plasma and be selectively converted into DON in target tissues, such as brain and tumor (21). Preclinical studies using DON therapy provide support for testing of these novel prodrugs in human subjects.
High MYC AT/RTs are relatively sensitive to DON therapy, leading to a slowing of AT/RT cell growth and inducing high rates of apoptosis. DON crosses the blood–brain barrier in sufficient concentrations to significantly extend survival in orthotopic mouse models of AT/RT when administered to mice at a dose comparable with well-tolerated doses in the pediatric phase I study (22, 33). MYC expression is responsible for this sensitivity to DON therapy because forcing MYC expression in low MYC-expressing AT/RT sensitizes cancer cells to DON therapy.
Our work suggests clinical biomarkers could be established to predict tumor sensitivity to DON therapy. MYC expression in AT/RT cells and tumors is correlated with increased reliance on glutamine and sensitivity to DON. MYC expression in our AT/RT cells is heterogeneous with some cell lines robustly positive and others negative. IHC for MYC on AT/RT biopsy samples may predict sensitivity to DON treatment in humans. We also found that the glutamine:glutamate ratio may serve as a biomarker to predict sensitivity to DON therapy in AT/RT. This ratio can be especially useful because glutamine and glutamate peaks can be identified on MR spectroscopy (48). Clinicians may be able to identify tumors sensitive to DON therapy from MRI even before tumor biopsy.
While the survival benefit of DON alone is exciting in this aggressive subgroup of AT/RT, orthotopic mouse models of AT/RT still die from tumor progression. On the basis of data from our metabolic flux experiments where we found that DON treatment led to depletion of the detoxifying peptide GSH, we combined DON therapy with carboplatin. We found that DON sensitizes AT/RT to platinum therapies slowing cell growth and inducing high rates of apoptosis. Exogenous GSH and the GSH precursor NAC independently rescued this combinatorial effect on apoptosis. Combining DON with carboplatin in orthotopic mouse models of high MYC AT/RT was well tolerated, induced high levels of apoptosis in the tumor while sparing normal cerebral cortex, and further improved AT/RT survival, suggesting that this combination therapy would have a broad therapeutic index.
DON combined with carboplatin is a novel drug combination that can translate into a new clinical trial. The well-established safety profile of DON in pediatric patients and the mechanistic synergy with platinum-containing agents presented here, suggest that DON may be incorporated into the backbone of standard AT/RT therapy for high-MYC tumors. DON may further sensitize tumors to standard therapies, reduce AT/RT drug resistance, and improve survival in this poor-prognosis subgroup of AT/RT.
Disclosure of Potential Conflicts of Interest
J. Alt is a consultant/advisory board member for Dracen Pharmaceuticals. B. Slusher has ownership interests (including patents) at and reports receiving commercial research grants from Dracen Pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.
Conception and design: S.Z. Wang, A.R. Hanaford, C.G. Eberhart, E.H. Raabe, J.A. Rubens
Development of methodology: S.Z. Wang, B. Poore, J. Alt, A. Price, E.H. Raabe, J.A. Rubens
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.Z. Wang, J. Alt, S.J. Allen, H. Kaur, B.S. Slusher, E.H. Raabe, J.A. Rubens
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.Z. Wang, B. Poore, H. Kaur, C.G. Eberhart, E.H. Raabe, J.A. Rubens
Writing, review, and/or revision of the manuscript: S.Z. Wang, B.S. Slusher, C.G. Eberhart, E.H. Raabe, J.A. Rubens
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B.A. Orr, E.H. Raabe
Study supervision: E.H. Raabe, J.A. Rubens
Other (provided research materials): B.S. Slusher
This work was supported by grants from Alex's Lemonade Stand (to J.A. Rubens and E.H. Raabe); Department of Defense (to J.A. Rubens, CA171021), and Giant Food Pediatric Cancer Research Fund. This work was also supported by NCI Core Grant to Johns Hopkins (SKCCC P30CA006973) and 1R01NS103927-01A1 (to B.S. Slusher and E.H. Raabe).
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.