1. Introduction
Pediatric medulloblastoma and ependymoma represent the first and third most common childhood brain malignancies, respectively [
1]. Current up-front treatment for medulloblastoma involves maximal safe surgical resection, followed by craniospinal irradiation and a combination of chemotherapeutic agents such as tubulin inhibitors and DNA alkylators [
2]. Extensive transcriptomic analyses have categorized medulloblastoma into four molecular subgroups: WNT, SHH, Group 3, and Group 4, that can be further stratified into 13 molecular subtypes [
3]. Each subtype warrants different management regimes to maximize tumor response to treatment due to genetic differences, although these are yet to be defined. Currently, the poorest survival rates are associated with
MYC-amplified Group 3 or
TP53-mutated SHH disease, highlighting the urgency for the development of novel therapies to improve survival rates for these medulloblastoma subtypes.
Ependymoma arises from the ependymal cells lining the ventricular walls of the central nervous system (CNS) and accounts for 6–12% of all intracranial tumors in children, with up to 30% diagnosed in those younger than three years [
4]. Contemporary standard-of-care treatment for ependymoma patients is surgery followed by focal radiotherapy [
5]. Despite this, 40–60% of these tumors will recur at the primary site of disease or with metastatic spread [
4,
6]. The role of chemotherapy in this disease remains uncertain and is currently being investigated in several international clinical trials (clinicaltrials.gov identifiers: NCT01096368 and NCT02265770) [
7,
8]. Although no chemotherapeutic regimen has proven clinically beneficial in the treatment of pediatric ependymoma to date, results from ACNS0831 suggest that some patients may benefit from maintenance chemotherapy [
9]. Complicating the current lack of safe and efficacious therapies, ependymal malignancies with similar histological grades exhibit genetic heterogeneity, leading to a diverse range of patient outcomes despite identical treatment strategies [
10]. On the basis of transcriptomics and DNA methylation patterns, ependymoma has also been delineated into multiple subgroups, of which posterior fossa A ependymoma (EPN_PFA) and
C11orf95 fusion-positive ependymoma (formerly EPN_RELA, and recently renamed due to the identification of other fusion partners for
C11orf95) predominate in children [
11,
12]
While our understanding of the molecular pathogenesis of these diseases has improved dramatically in the last three decades, this knowledge has not translated to increased patient survival. Amongst survivors, current treatments can cause debilitating side effects, including untreatable secondary malignancies [
13], cognitive dysfunction, cardiotoxicity, myelosuppression, renal toxicity, and endocrine problems [
14]. Therefore, improved clinical outcomes are dependent on the identification of efficacious therapies that not only increase survival but reduce treatment-related side effects.
Numerous research studies since the late 1990s have contributed to a growing body of evidence demonstrating that various cannabinoids have anti-cancer effects. These include studies performed in a wide variety of experimental models of cancer, ranging from cancer cell lines in culture to genetically engineered mice, and covering several cancers including lymphomas and adult brain tumors [
15]. Δ9-tetrahydrocannabinol (THC) is one of the primary compounds that can be isolated from the plant
Cannabis sativa L and is known to exert a broad range of biological and psychoactive effects. THC mimics the actions of endocannabinoids, a family of lipid-based signaling molecules, by binding to and activating cannabinoid receptors type 1 (CB
1R) and type 2 (CB
2R) [
16,
17]. These two G-protein coupled receptors are primarily expressed in cells within the CNS and immune system, respectively. Cannabidiol (CBD) is another highly abundant cannabinoid in
Cannabis sativa L extracts that has been demonstrated to also exert biological effects in mammals; however, unlike THC, CBD has low affinity for CB
1R and CB
2R [
18]. Recent studies have shown that CBD instead targets several other G-protein coupled receptors, such as GPR55, GPR18, and 5-HT1A [
19,
20], and transient receptor potential (TRP) channels such as TRPV1 and TRPV2 [
21,
22]; however, its mechanism of action in mammals is yet to be fully elucidated.
Research into the effects of THC and CBD in different types of cancer overwhelmingly show that THC and CBD induce cancer cell death [
23]. The more intriguing brain tumor-related studies have been in mouse models of glioblastoma (the most common adult brain cancer) and showed that both THC and CBD improved animal survival when administered to mice in combination with the standard of care chemotherapy temozolomide [
24,
25]. Furthermore, when glioma cells were pre-treated with THC or CBD, either in vitro or in vivo, this sensitized the cancer cells to radiation-induced death and prolonged survival of mice [
26]. Mechanistically, the effects of cannabinoids on glioblastoma cells are mediated by the inhibition of proliferation and the induction of cell death via autophagy and apoptotic mechanisms [
27,
28,
29].
There is no existing data on the effect of these agents in pediatric brain tumor models in vitro or in vivo. Some anecdotal reports can be found describing the benefits of medicinal cannabis for these patients, but assessment of these is challenging, because the dosages and exact components of the plant extracts used have not been comprehensively documented (as would be done in a conventional clinical trial). As such, any data from these patients is unreliable. With the increasing availability of medicinal cannabis, there is a growing demand from patients, parents, and physicians for better information describing the safety and efficacy of cannabinoids in pediatric brain tumors [
30,
31,
32], although reports describing the proportion of pediatric oncology patients actively seeking or using medical cannabis are scarce. A recent report from the Children’s Hospital of Minnesota stated that hope for an anti-tumor effect was the major reason parents sought medical cannabis for children with brain cancer despite a lack of evidence demonstrating such an effect [
32]. To address these gaps in the literature and to help identify novel anti-cancer agents with better safety profiles for the treatment of these diseases, our study aimed to investigate the anti-cancer efficacy and mechanisms of action for THC and CBD in preclinical models of pediatric medulloblastoma and ependymoma.
3. Discussion
Tumors arising in the CNS are the most common solid cancers of childhood and the major cause of childhood cancer deaths [
53]. This research study has focused on two aggressive pediatric brain cancers that affect young children particularly—medulloblastoma and ependymoma. Despite multimodal treatment protocols including surgery, radiotherapy, and chemotherapy, survival rates for patients with high-risk disease have failed to improve significantly for several decades, and recurrences are common. Extensive molecular characterization of both cancers [
3,
11] has not led to successful translation of many novel therapies to the clinic, even with focused efforts to develop drugs that target specific mutations in these diseases. In addition, survivors frequently encounter significant long-term sequelae including developmental defects, psychosocial deficits, and secondary tumors [
54]. Thus, additional improvements are required for patients with aggressive disease, and new treatment approaches must focus on innovative ways to reduce the long-term toxicity of therapy [
55].
Recent research has demonstrated that cannabinoids, including the phytocannabinoids THC and CBD but also synthetic cannabinoids, exhibit anti-tumor properties in different adult cancer types, including breast cancer, melanoma, pancreatic cancer, lymphoma, and glioblastoma [
23,
56]. The robustness of these studies has resulted in the implementation of new clinical trials, specifically in adult glioblastoma, that are investigating the cannabis-based medicine Sativex (GW Pharmaceuticals) in combination with temozolomide (ClinicalTrials.gov ID: NCT01812603 and NCT01812616). In addition, it has been established that cannabinoids, particularly CBD, exhibit minimal toxicity, may reduce some brain tumor-related symptoms such as seizures, and are well tolerated by patients [
57]. Despite these encouraging data, and the known ability of cannabinoids to penetrate the blood–brain barrier, there is no existing pre-clinical data on the effect of these agents in pediatric brain cancer.
In order to investigate the role of THC and CBD in medulloblastoma and ependymoma, we took advantage of existing in vitro models of each disease [
4,
58,
59,
60,
61]. Specifically, these models represent
MYC-amplified Group 3 medulloblastoma and
C11orf95 fusion-positive ependymoma. In addition, we used multiple cell lines of these two molecular subtypes to confirm the reproducibility of our findings across multiple models. Encouragingly, we observed expression of both canonical and non-canonical cannabinoid receptors within medulloblastoma and ependymoma samples, as well as the cell lines. Given this was the first study of these compounds in pediatric brain cancer, we chose to investigate purified THC and CBD, rather than utilize plant extracts, which contain a complex mixture of many additional compounds with potential therapeutic properties [
62]. While it has been demonstrated that whole-plant cannabis preparations may elicit better therapeutic effects [
63], we aimed to first determine if THC and CBD had any specific effect on pediatric brain cancer cells. Our data demonstrate that both compounds influence medulloblastoma and ependymoma cells, although in different ways.
Consistently, both CBD and THC reduced the viability of these cell lines, although the medulloblastoma cell lines were more sensitive than ependymoma cells. This was particularly encouraging given that the cell lines used represent subgroups of each disease with poor prognosis. Surprisingly, the effects of THC did not appear to be mediated via CB
1R, but instead cannabinoid-induced toxicity could be prevented by α-tocopherol in three out of the four cell lines tested, suggesting that these drugs induce ROS within brain cancer cells. Similar mechanisms of action have been demonstrated in glioblastoma cells where ROS induction was linked to increased apoptosis [
38]. However, neither THC nor CBD induced significant apoptosis in medulloblastoma cells when assessed by flow cytometry. This may be due to the concentrations of THC and CBD used, which were below the concentrations required to completely inhibit viability. The effect of α-tocopherol in ependymoma cells was less clear with it appearing to have a negative effect on DKFZ-EP1NS cell viability.
Given the demonstrated roles for THC and CBD in the modulation of intracellular signaling pathways, we investigated their effects primarily in medulloblastoma cells. In general, an overall decrease in signaling pathway activity was observed; however, these changes were rarely statistically significant when multiple experiments were quantified. In the case of ERK1/2 activity, disparate effects of CBD were observed with one cell line having reduced activity, one with increase phosphorylation, and the other with no change. The variability in response might be due to differences in the underlying genetics of each cell line used; however, given that we have selected three medulloblastoma cell lines with the same molecular classification, these differences have been experimentally minimized. Moreover, it has been shown that cannabinoids have a biphasic effect on ERK1/2 phosphorylation [
40]; thus, it is feasible that timing in response to THC or CBD may differ across cell lines. Both THC and CBD appeared to increase autophagy in medulloblastoma cells, similar to what has been reported in cultured glioblastoma cells [
29]. To date, molecular markers that predict how cancer cells respond to cannabinoids have not been reported but would be useful to distinguish cancers that might be inhibited by these agents.
Studies have shown that combining different cannabinoids, especially THC and CBD, can potentiate the effect of each cannabinoid. It was encouraging that THC and CBD appeared to synergistically reduce medulloblastoma and ependymoma cell viability in vitro; however, when tested in vivo, no survival benefit of either drug, or the combination of THC and CBD, was observed. This contrasts with studies in other brain cancers, where the combination of THC and CBD was superior to the use of each drug as a single agent in vivo [
24]. One major difference is that Torres et al. utilized ectopic models of glioblastoma [
24]. Another study that used an intracranial model of glioblastoma showed that the combination of THC with CBD did not improve animal survival, supporting our data, although they did see a benefit of cannabinoids when they were combined with temozolomide [
64]. This approach of combining multiple different drugs has underpinned much of our past success in the treatment of pediatric cancer; thus, we tested the combination of THC and CBD with the medulloblastoma drug CPA. THC and CBD appeared to exhibit in vitro synergy with CPA; however, these effects were also not replicated in vivo. It is conceivable that the differences between the in vitro and in vivo effects of cannabinoids observed in our and other’s studies are due to their bioavailability and pharmacokinetics. It has been reported that the bioavailability of cannabinoids after oral administration can vary considerably, because they need to be absorbed in the intestine, then metabolized in the liver, before incorporation to the blood stream, thus affecting the cannabinoid concentrations reaching the tumor and consequently impacting any anti-tumor effect. Different studies have compared the pharmacodynamics and pharmacokinetics of cannabinoids in tumor-free mice and rats after different routes of administration, showing that the concentration of CBD in plasma or in brain after oral administration is much lower than intraperitoneal or intravenous administration [
49,
65]. To address the potential poor intra-tumoral drug concentrations that may result from oral administration of cannabinoids, we tested an alternative route of administration of these compounds, but this also did not improve overall survival in the models tested. A limitation of our study is that we did not measure intratumoral cannabinoid concentrations using mass spectrometry, although the inhibition of ERK1/2 phosphorylation we observed within medulloblastomas indicated that sufficient CBD is present within medulloblastoma xenografts to inhibit signaling; however, we did not have a molecular marker demonstrating effects of THC in these cells to validate intra-tumoral drug penetration.
Previously, we reported drug interaction studies between the pan-ErbB inhibitor dacomitinib with CPA using these same medulloblastoma cell lines and found additive or antagonistic in vitro drug interactions were in fact antagonistic in vivo [
66]. Antagonism was not observed when THC or CBD were combined with CPA in vivo, which possibly indicates that although cannabinoids do not improve survival, they do not appear to inhibit the action of conventional chemotherapy. Our flow cytometry data indicated that CBD inhibited medulloblastoma proliferation. Since many anti-cancer chemotherapies, including CPA, require cells to be actively proliferating for cytotoxicity, this CBD-induced cytostasis may be the reason there was no improvement in survival when it was combined with CPA in vivo. A limitation of this study is that we only assessed one of the chemotherapies used in medulloblastoma treatment and have not investigated the influence that THC or CBD may have on other standard agents used in medulloblastoma or ependymoma, such as cisplatin, vincristine, etoposide, or radiation. It has been shown that phytocannabinoids have suppressive effects on immune cells [
67]; however another limitation of our research is that we evaluated the effectiveness of THC and CBD using human brain cancer cells xenografted into immune-deficient mice. While preclinical studies that showed cannabinoids are effective against adult glioblastoma used similar immune-deficient models, it may be useful to determine if the presence of a functioning immune system might alter the effects of cannabinoids on medulloblastoma and ependymoma.
Although no direct anti-tumor effect was observed in our in vivo experiments, cannabinoids have been proven to have numerous other positive effects for cancer patients. For example, modulation of the endocannabinoid system has been shown to prevent cisplatin-induced neuropathic pain in preclinical models [
68], suggesting cannabinoids may potentially improve patient quality of life; however, these data have not yet been validated in a clinical setting. Regardless of the potential positive effects of THC and CBD, the possibility that they may interact with other cancer treatments must be noted. THC and CBD are predominantly metabolised in the liver by cytochrome p450 family enzymes [
51]. Specifically, they are primarily metabolised by CYP3A4 [
50,
51], which also metabolises several brain cancer chemotherapeutics including CPA and vincristine [
69,
70]. Therefore, although we did not observe additional toxicity in our mouse models when THC or CBD were combined with CPA, there is the potential that co-administration of cannabinoids with conventional cancer treatments could alter the bioavailability of chemotherapeutics, prolonging their cytotoxic effects in children.
In summary, we have comprehensively assessed the effects of THC and CBD in mouse models of medulloblastoma and ependymoma and investigated the intracellular effects of these compounds using multiple cell lines that represent two molecular subtypes with poor survival outcomes. While it could be argued that better efficacy may have been observed in models representing less aggressive disease subgroups, those tumors currently respond well to existing therapies, and our goal was to focus on identifying improved therapeutics for subgroups of these diseases most likely to relapse with current gold standard therapies. THC and CBD do not appear to elicit an in vivo survival benefit in mouse models of medulloblastoma or ependymoma. Given the lack of therapeutic efficacy in these cancer models and insufficient data demonstrating other benefits in children, future studies focusing on the potential for these compounds to improve quality of life are required to build more evidence for the use of these drugs in pediatric neuro-oncology.
4. Materials and Methods
4.1. Analysis of Human Medulloblastoma and Ependymoma Expression Data
Boxplots were generated using the R statistical environment to infer the levels of mRNA expression seen in human ependymoma and medulloblastoma samples. Two ependymoma datasets were analyzed: (1) GSE64415 [
11] and (2) RNA-seq from formalin-fixed paraffin embedded (FFPE) tissue in UK ependymomas (Ritzmann et al. in preparation, data available upon reasonable request). One medulloblastoma dataset was analyzed (GSE85217 [
36]). All samples were supported by DNA methylation-based subclassification [
71]. For ependymoma, the predominant pediatric subgroups were included (EPN_PFA, EPN_C11orf95), and for medulloblastoma, the four main subgroups were analyzed (WNT, SHH, Group 3, Group 4). For the two publicly available datasets, the normalized data were downloaded from the Gene Expression Omnibus (GEO) and levels of expression plotted per subgroup. For the RNA-seq expression set, data was generated by 100bp paired-end RNA-seq on total RNA extracted from FFPE tissue following ribodepletion with Ribo-Zero. Libraries were sequenced on an Illumina HiSeq machine targeting 50 million reads per sample and aligned to the human genome (Hg19) and transcriptome using TopHat2 and summarized at gene level using FeatureCounts. DNA Methylation analysis for the RNA-seq dataset was performed as previously described [
6]. Samples were normalized to log2 transcripts per million (log2TPM) for visualization.
4.2. Medulloblastoma and Ependymoma Cell Lines and Culture Conditions
D425 and D283 cells [
58,
59] were a gift from Darell Bigner of Duke University (Durham, NC, USA), and PER547 cells [
60] were a gift from Ursula Kees of Telethon Kids Institute (Perth, Australia). The genetic identity of medulloblastoma cell lines was confirmed by STR analysis and sequencing. The IC-1425EPN cells were gifted by Xiao-Nan Li (Northwestern University, Chicago, IL, USA) [
4], while the DKFZ-EP1NS cells were provided by Till Milde (German Cancer Research Center (DKFZ), Heidelberg, Germany) [
61]. Cells were confirmed to be mycoplasma-free using a MycoAlert™ Mycoplasma Detection Kit (Lonza, Basel, Switzerland). Cells were transduced to express luciferase using the retroviral expression construct MSCV-ires-pacLuc2 (D283 and PER547) or lentiviral expression construct pCL20-MSCV-GFP-ires-Luc2 (D425). Constructs were generously supplied by Drs Suzanne Baker, Richard Williams, and Arthur Nienhuis, of St Jude Children’s Research Hospital (Memphis, TN, USA). Medulloblastoma cell lines were cultured in antibiotic-free media supplemented with Glutamax (Invitrogen, Carlsbad, CA, USA) at 37 °C in 5% CO
2 as follows: D283: MEM-alpha (Gibco, Waltham, MA, USA) with 10% fetal bovine serum (FBS) (CellSera, Rutherford, Australia); PER547: RPMI (Gibco) with 1 mM sodium pyruvate (Invitrogen), non-essential amino acids (Invitrogen), 50 µM 2-mercaptoethanol (Sigma-Aldrich, St Louis, MO, USA), and 10% FBS; D425: modified IMEM (Gibco) with 1 M HEPES (Gibco) and 10% FBS. Ependymoma cell lines were cultured in media supplemented with Glutamax and antibiotics at 37 °C in 5% O
2 and 5% CO
2 as follows: IC-1425EPN (short term cultures of cells isolated from xenografts): DMEM:F12 (Gibco), 10% FBS, 1 µg/mL heparin (Sigma Aldrich, St Louis, MO, USA), 20 ng/mL epidermal growth factor (EGF), and 20 ng/mL basic fibroblast growth factor (FGF) (both from Peprotech, Rocky Hill, CT, USA); DKFZ-EP1NS: Neurobasal medium without vitamin A (Gibco), B27 with vitamin A (Gibco), N2 (Gibco), 20 ng/mL EGF, 20 ng/mL FGF, and 1 µg/mL heparin.
4.3. RNA Isolation and Transcriptome Profiling (RNA-Seq) from Cell Lines
D283, D425, PER547, or IC-1425EPN cells were cultured as described above, and total RNA was extracted using an AllPrep DNA/RNA mini kit (Qiagen, Hilden, Germany). RNA integrity was assessed by bioanalyzer (RIN 10). Total RNA was shipped to the Australian Genome Research Facility for library preparation and sequencing (Illumina NovaSeq 6000, 150bp paired-end (PE) reads). Between 12–59 million sequencing reads were generated per sample. The raw sequencing data are available from the European Genome-Phenome Archive (Accession number: EGAS00001004963).
4.4. Pre-Processing of RNA-Seq Data
FASTQC [
72] was employed for pre-alignment quality control of raw sequence reads. Reads were aligned to the human reference genome (hg38) using HISAT2 [
73] and summarized at the gene-level using featureCounts [
74]. Post-alignment QC was carried out with SAMStat [
75]. The proportion of mapped reads was 93% (range 92.3–5.0%). A gene was deemed expressed with a count ≥ 10 in at least one cell line and visualized showing counts per million (cpm), which normalizes each cell line with respect to sequencing depths.
4.5. Compounds
THC and CBD were purchased from THC Pharm GmbH (Frankfurt, Germany). THC was dissolved in ethanol and stored at −20 °C. Prior to use, ethanol was evaporated in a siliconized tube under a nitrogen stream. CBD was provided as a powder and stored at room temperature. The cannabinoids were either dissolved in DMSO (10 mM) (Sigma Aldrich) for in vitro use, or dissolved directly in Miglyol 812 (IOI Oleochemical GmbH, Hamburg, Germany) for in vivo administration either per os (p.o., oral gavage) or intraperitoneally (i.p.). Dosages are indicated in the text. CPA (Endoxan; Baxter Healthcare, Deerfield, MA, USA) was dissolved in saline and delivered i.p. as described in the text. For in vitro studies, the activated form of CPA was used, and 4HPC (Toronto Research Chemicals, Toronto, ON, Canada) was dissolved in DMSO and stored at −80 °C. The CB1R-selective antagonist SR141716A (SR1) was from MedChemExpress (Monmouth Junction, NJ, USA) and stored at 80 °C, and α-tocopherol (αTOC) was from Sigma-Aldrich. Both compounds were dissolved in DMSO for in vitro experiments.
4.6. Drug Sensitivity and Drug Interaction Assays
Cannabinoids are known to bind serum proteins; therefore, drug sensitivity was assessed in low serum conditions (1.5% FBS). Cells were plated (1500/well) into tissue culture treated 384-well plates (Corning, New York, NY, USA) using a Multidrop Combi (Thermo Scientific, Waltham, MA, USA). Medulloblastoma and ependymoma cells were incubated for 1 and 24 h, respectively, prior to the addition of drug. Drugs (either single drugs or drug combinations) were dispensed using an HP300 digital dispenser (Tecan, Mannedorf, Switzerland) with concentrations indicated in the text. Cells were treated for 72 h and incubated with alamar blue (2.5% methylene blue, 1 mM potassium hexacyanoferrate (III), 1 mM potassium hexacyanoferrate (II) trihydrate, and 0.6 mM resazurin (all from Sigma-Aldrich)) for the final 6 h of treatment. Resorufin fluorescence was detected using a SynergyMX plate reader (Biotek, Winooksi, VT, USA) with 570 nm excitation and 590 nm emission. Data were expressed as a percentage of DMSO-treated controls present on each plate. The ED50 was interpolated from a best-fit dose–response curve determined using Prism v8 (GraphPad Software, San Diego, CA, USA). Drug interactions were analyzed using Combenefit software (Cambridge University, Cambridge, UK) [
45].
4.7. Cannabinoid Receptor Antagonist Assays
Cells were washed and resuspended in 1.5% FBS prior to plating (6000/well) into tissue culture-treated 96-well plates using a Multidrop Combi (Thermo Scientific, Waltham, MA, USA). Drugs were dispensed using an HP300 digital dispenser (Tecan) in a combination array matrix. Cells were treated with 2 µM of CB
1R-selective antagonist SR1 or 10 µM of antioxidant αTOC for 1 h [
41,
76]. THC and CBD were then added (concentrations indicated in the text) for 72 h, and viability assessed using alamar blue as above. The reduction of resazurin to resofurin was measured via absorbance at 600 nm using a SynergyMX plate reader. Raw absorbance data was normalized to the absorbance measured in the corresponding control for each treatment and expressed as a percentage of control.
4.8. Protein Analysis by Western Transfer and Immunoblotting
PER547, D283, and D425 cells were washed and resuspended twice in medium containing 1.5% FBS and incubated for 3 h before being treated with the ED50 and ED80 doses of THC and CBD. Cells were lysed after 24 or 48 h with radioimmunoprecipitation assay (RIPA) buffer containing protease and phosphate inhibitors (Roche, Basel, Switzerland). Protein concentration was quantified using BCA assay (Pearce, Appleton, WI, USA) and 30 µg/lane separated using 4–12% NuPAGE Bis-Tris gels (Invitrogen) followed by transfer onto nitrocellulose membranes. Membranes were immunoblotted with specific primary antibodies, followed by horseradish peroxidase-conjugated secondary antibodies (1:5000) (GE Healthcare, Chicago, IL, USA). Relevant bands were detected using Supersignal West Dura (Pierce) or Clarity Western ECL (BioRad, Hercules, CA, USA) and images collected using a BioRad ChemiDoc. Primary antibodies used were PRAS40 (Cell Signaling Technologies (CST, Danvers, MA, USA) #2691, 1:1000), phosphorylated PRAS40 Thr246 (CST #2997, 1:1000), p42/44 ERK1/2 (CST #9102, 1:1000), phosphorylated ERK1/2 Thr202/Tyr204 (CST #9101, 1:1000), S6 ribosomal protein (CST #2217, 1:1000), phosphorylated S6 ribosomal protein Ser235/236 (CST #2211, 1:1000), 4EBP1 (CST #9452, 1:1000), phosphorylated 4EBP1 Thr37/46 (CST #2855, 1:1000), LC3A/B (CST #12741, 1:1000), cleaved PARP Asp214 (CST #5625, 1:1000), and β-actin (Sigma-Aldrich #A1978, 1:5000).
4.9. Flow Cytometry for Cell Cycle Distribution and Apoptosis
Cell cycle distribution was analyzed using EdU (added 45 min before harvest) to label cells in S phase and DAPI to label DNA content. D283 cells were treated with DMSO (0.1%), 7.5 μM THC or 5.5 μM CBD, in the presence of DMSO or 10 μM 4HPC. Time of harvest is indicated in the figures. Cells were stained using the Click-iT EdU AlexaFluor488 kit (Invitrogen). In addition, cells were stained with AlexaFluor647-conjugated cleaved PARP (CST #68975, 1:50) to identify apoptotic cells, and PE-conjugated phospho-histone H3 Ser10 (CST #5764, 1:50) to mark cells in mitosis. Samples were analyzed using an LSRFortessa X20 (BD, Franklin Lakes, NJ, USA) and results were visualized and quantified using FlowJo software. Data are pooled from two independent experiments and show the mean with standard deviation (SD).
4.10. Orthotopic Implant Models of Medulloblastoma and Ependymoma
Animal experiments were approved by the Animal Ethics Committee of the Telethon Kids Institute and performed in accordance with Australia’s Code for the Care and Use of Animals for Scientific Purposes. For survival studies, cells (100,000 per mouse) were suspended in Matrigel (Corning) and implanted into the right cerebellar hemisphere of 6–10-week-old female athymic mice (Balb/c nude) or NRG mice for medulloblastoma and ependymoma cells, respectively (Animal Resources Centre, Perth, Western Australia, Australia) using a Hamilton syringe as previously described [
77]. Tumor size was monitored by bioluminescence using a LagoX Optical Imager (Spectral Instruments Imaging, Tucson, AZ, USA). Once tumors were established, mice were randomized into groups based on bioluminescence flux to obtain groups of mice with close to equal mean bioluminescence prior to treatment as indicated in the text. For D283, treatment commenced seven days after implantation. For IC-1425EPN, treatment commenced once the average radiance reached 10
6 p/s. Median survival and Kaplan–Meier survival curve comparisons were calculated using GraphPad Prism (v8). An event was recorded when mice were euthanized due to intracranial tumor-related morbidity. Mice requiring euthanasia for non-tumor-related reasons (e.g., weight loss, infection, and physical trauma) were censored. Whole blood was collected weekly, treated with ethylenediaminetetraacetic acid, and parameters measured using a BC-5000VET Hematology Analyzer (MindRay, Shenzhen, China).
4.11. Immunohistochemistry
Mouse brains were embedded in paraffin after fixation in 4% paraformaldehyde in PBS overnight at 4 °C. Tissue sections (5 µm) were processed in a citrate buffer for antigen retrieval before immunostaining with phosphorylated ERK1/2 (pERK1/2) Thr202/Tyr204 (CST #9101, 1:200). An Elite ABC kit with NovaRED substrate was used for antibody detection, and tissue sections were counterstained with Gill’s hematoxylin (all from Vector Laboratories, Burlingame, CA, USA). Positively stained cells from a minimum of four images per tumor were quantified using a Nuance spectral unmixing camera and InForm Tissue Finder software (Perkin Elmer, Waltham, MA, USA).
4.12. Statistical Analyses
Band intensities on immunoblots were quantified using Image J [
78] and treatments compared to DMSO-treated samples using a Kruskal–Wallis test with Dunn’s multiple comparisons test. Flow cytometry and immunohistochemistry results were compared using a one-way ANOVA with Bonferroni’s correction for multiple comparisons. For in vivo experiments, sample size calculations were performed based on the known mean and standard deviation of survival for the orthotopic implant models. These indicated that with four mice per group, we should be able to detect a true difference in the mean response of treated and control mice of −9.76 or 9.76 days with probability (power) 0.80. The Type I error probability associated with this test of the null hypothesis that the population means of the treated and control groups are equal was 0.05. Kaplan–Meier survival curves were compared using the log-rank (Mantel-Cox) test. THC, CBD, or THC:CBD treated groups were compared to vehicle controls, while THC:CPA and CBD:CPA combination-treated groups were compared to mice treated with CPA alone. Values of significance are indicated by asterisks and described in each figure legend where appropriate.