Introduction
Mutations in
isocitrate dehydrogenase 1/2 (IDH
mut) occurs early in gliomagenesis (Sanson et al.
2009) and lead to neomorphic production of the oncometabolite D-2-hydroxyglutaraste (D2HG). High intracellular levels of D2HG result in inhibition of DNA and histone demethylase enzymes, thereby leading to epigenetic dysregulation through increased DNA and histone methylation. DNA and histone methylation is associated with epigenetic repression of genes through formation of heterochromatin. Of interest is the stability of this DNA methylation observed in IDH
mut glioma tumors, with much of the CpG island methylation that defines IDH
mut glioma being retained in patients post-resection and chemotherapy (Mazor et al.
2017). Based on this, we became interested in investigating epigenetic inhibitors as potential therapeutic targets for IDH
mut glioma.
The bromodomain and extraterminal (BET) family of proteins represents a class of epigenetic “readers” that bind to acetylated chromatin residues to facilitate transcriptional activation of nearby genes (Jain et al.
2017), hence these proteins are also termed transcriptional coactivators. These BET proteins have proven roles in various physiological and pathophysiological processes such as maintenance of stem-cell populations (Horne et al.
2015; Wu et al.
2015; Bolden et al.
2014) and promotion of tumorigenesis (Xu et al.
2018). The BET family of proteins included BRD2, BRD3, BRD4, and the testis-specific isoform BRDT (Taniguchi
2016). Most research has focused on the function of the BRD4 isoform for its major role in transcriptional pause-release via recruitment of the elongation factor complex PTEF-b to transcriptionally paused genes (Miller et al.
2017). With the observation that BET proteins are upregulated in cancer (Pastori et al.
2014) and facilitate transcription of oncogenes such as c-MYC (Delmore et al.
2011), this led to the concept that BET proteins, particularly BRD4, facilitate tumorigenesis via transcriptional coactivation of oncogenes. Inhibition of receptor tyrosine kinase (RTK) pathways, such as PDGFR signaling (Wang et al.
2015), has also been observed via BET inhibition. These observations support the investigation of small molecule BET inhibitors for utilization as potential cancer therapeutics.
In the 1990s, Mitsubishi Pharmaceuticals identified the first known bromodomain inhibitors using the benzodiazepine pharmacophore thienodiazepam. This chemical scaffold was eventually further improved upon as a lead compound by Filippakopoulos et al. in 2011 with successful development of the selective bromodomain inhibitor JQ1 (Filippakopoulos et al.
2010). This drug exhibited pan-BET inhibition activity, with similar binding affinity for BRD2/3/4 but low binding to BRDT. JQ1 exhibited very promising pharmacodynamic properties when investigated as an anticancer therapeutic by targeting oncogene expression, such as c-MYC. This drug was also shown to readily penetrate the blood–brain barrier, but the pharmacokinetics proved troublesome with a plasma half-life of just 1–1.5 h (Filippakopoulos et al.
2010). Nevertheless, this compound has been used extensively in the literature as a tool compound to investigate BET-related functions in various ailments such as cancer (Cheng et al.
2013) and inflammatory diseases (Belkina et al.
2013).
Of interest is the role of MYC proteins in the progression and malignancy of IDH
mut gliomas, with MYC activation associated with progression from low grade glioma to high-grade GBM (Korshunov et al.
2019). Additionally, it has been shown that other IDH
mut cancers, such as cholangiocarcinomas and colorectal cancers, have also exhibited enhanced sensitivity to BET inhibitors (Fujiwara et al.
2018; McCleland et al.
2016). Therefore, we became interested in whether IDH
mut gliomas more heavily rely upon the oncogenic functions of BET transcriptional coactivators for cellular survival and proliferation.
Here, we investigate the role of BET protein function in IDH wildtype (IDHwt) vs IDHmut glioma using The Cancer Genome Atlas (TCGA) clinical data and in vitro cell assays. We utilized a panel of IDHwt/mut glioma cell lines, representing Grade 3 and Grade 4 gliomas, along with an IDHwt cell line bearing a stable transgenic R132H IDH1 mutation to investigate the hypothesis that IDHmut gliomas are more sensitive to the antiproliferative and/or cytotoxic effects of BET inhibition.
Materials and methods
Glioma cell culture
0827, 0923, and 0905 GBM stem-like cells were previously isolated by Dr. Howard Fine’s lab and have been previously characterized for GBM stem-like properties and genomic alterations (Toledo et al.
2015; Ene et al.
2012; Son et al.
2009). BT142 (Luchman et al.
2011) is a grade III anaplastic astrocytoma and was purchased from ATCC, while TB096 (Moure et al.
2019), another grade III anaplastic astrocytoma, was a gift from Dr. Hai Yan at Duke University. Validation of
IDH1 mutations and 2-HG production has been previously conducted by our lab for all cell lines used in this study (Sears et al.
2021). All cell lines were grown in defined neurobasal cell culture medium that includes B-27 and N2 supplements along with EGF and FGF growth factors (NBE). 0905 and BT142 required 100 ng/mL PDGF-AB in the medium, while TB096 required culture conditions containing half NBE and half DMEM with 10% FBS. All cell cultures were cultured at 37C and 5% CO
2, included penicillin–streptomycin, and frequently monitored for mycoplasma contamination. Additionally, all cells were cultured in suspension as neurospheres except for TB096 which exhibited adherent growth. Cell culture medium was replenished 3 ×/week. JQ1 (Millipore Sigma; SML1524) was dissolved in 100% ethanol, sterile filtered, and diluted to 1000 × in sterile ethanol for drug treatments. Vehicle concentration was 0.1% or less for all treatments.
TCGA survival analysis
The UCSC Xena platform (Goldman et al.
2020) was utilized for analysis of patient overall survival in IDH
wt/mut glioma based on BET isoform (
BRD2/3/4) gene expression. Low grade glioma and glioblastoma patients were analyzed together via the TCGA GBMLGG cohort dataset. IDH
mut astroctyomas and oligodendrogliomas were distinguished based on 1p/19q codeletion status. This dataset includes information on copy number alterations, DNA methylation, gene expression, somatic mutations, and survival for each patient.
Cell viability and apoptosis assays
A Luminex Guava Muse Cell Analyzer was utilized to conduct cell viability and apoptosis assays using a Muse Cell Count and Viability Kit or a Muse Annexin V and Dead Cell Kit. Cells prepared as single-cell suspensions were analyzed on the Muse Cell Analyzer according to the manufacturer’s protocols. Briefly, cells were trypsinized (0.05%), mixed with Cell Count and Viability reagent or 1:1 NBE to Annexin V and Dead Cell reagent, and then incubated according to the manufacturer’s protocol. Stained samples were then run on the Muse Cell Analyzer, and gates were set based on a positive control sample.
Immunoblotting
Cell culture samples were washed with 1 × DPBS, pelleted via centrifugation, flash-frozen in liquid nitrogen, and stored at − 80 °C for later cell lysis. For preparation of whole cell lysates, cell pellets were resuspended in 1 × RIPA buffer supplemented with 1 × Halt protease and phosphatase inhibitor and then pellets to homogenized via a probe sonicator for 3 × 20 s at 30% amplitude. Insoluble material was pelleted via centrifugation, and the supernatant protein concentration quantified via Bradford assay. SDS-PAGE and transfer steps were conducted using the Novex immunoblot system. All blots used NuPage 4–12% Bis–Tris gradient gels. Blots were visualized via a ProteinSimple Fluorchem E Imager. Antibody information can be found in Table S1.
BrdU incorporation assays
BrdU incorporation assays were conducted using a BD Pharmingen BrdU Flow kit according to the manufacturer’s protocol. Briefly, cells were treated for 48 h with drug and then pulsed with 10 µM BrdU, except for our background stain control, for 2.5 h before washing, fixing, and storing at − 80 °C in solution of 1:9 FBS to DMSO. For analysis on flow cytometer, cells were thawed, permeabilized, and incubated with anti-BrdU antibody according to manufacturer’s protocol. Stained cell samples and unstained controls were run on a Beckman Coulter Cytomics FC500 flow cytometer with FITC filter applied. Forward- and side scatter voltages and gates were set using the No BrdU control, whereas FITC voltages were set using a BrdU positive control sample to determine BrdU + and BrdU − cells.
Plasmids
pSLIK-IDH1-FLAG (Addgene plasmid # 66802;
http://n2t.net/addgene:66802; RRID: Addgene_66802) and pSLIK-GFP (Addgene plasmid # 66844;
http://n2t.net/addgene:66844; RRID: Addgene_66844) were gifts from Christian Metallo37. These plasmids were provided as bacterial stabs and once acquired were grown in LB and clones screened for fidelity via restriction digestion and Sanger sequencing. Acceptable clones were then further grown in LB and plasmids isolated via Qiagen Maxiprep kit. Plasmid concentration was then quantified prior to development of lentivirus.
Lentivirus production
Lentiviral vectors were generated by transfecting either pSLIK-GFP or pSLIK-IDH1-R132H-FLAG transfer plasmids with an envelope plasmid (pCMV-VSV-G) and packaging plasmid (pCMV-dR8.2 dvpr). These three plasmids were transfected into 293 T cells using Polyplus Jetprime transfection reagent according to the manufacturer’s protocol and then cell culture media was collected every 24 h for 3 days and stored at − 80 °C. Lentivirus-containing cell culture medium was centrifuged on tabletop centrifuge to remove cell debris, and then further centrifuged on a Beckman Coulter Optima XL-100 K Ultracentrifuge at 4C and 25,000 g for 1 h. Lentivirus-containing pellets were resuspended in 1 × DPBS and then assayed via Takara Bio’s Lenti-X GoStix for presence of adequate lentiviral titer.
Lentiviral transduction of cell lines
0923 GBM cells were plated in 6-well plates and then lentivirus was titrated over 5 wells. After 2 days inoculation, lentivirus-containing NBE was swapped out with hygromycin-containing (150 μg/mL) NBE. After 2 weeks in hygromycin, transduced cells were grown in normal NBE and then validated for R132H IDH1 gene expression via immunoblot and 2HG production via ELISA.
Statistical analysis
2-way ANOVA and Student’s t test statistical analyses were performed via GraphPad Prism 8.0 software package. Logrank statistical analysis of Kaplan–Meier survival curves was conducted via the UCSC Xena Browser platform. All replicate analyses are presented as mean ± standard deviation of the mean for three independent replicates at a significance level (α) of 0.05 unless otherwise indicated in the figure legends. Statistical significance was denoted in figures using the following system: *p ≤ 0.05; **p ≤ 0.001; ***p ≤ 0.0001.
Discussion
In this study, we utilized TCGA clinical data combined with cell health assays and showed that IDH
mut gliomas are more reliant on the BET transcriptional coactivators for cell survival and proliferation. Analysis of clinical TCGA low grade glioma and glioblastoma data shows BRD4 gene expression correlates with decreased survival only in IDH
mut glioma, while BRD3 gene expression correlates with reduced patient survival only in IDH
wt glioma. This supports the notion that BET isoform activity promotes malignancy preferentially in IDH
mut glioma and suggests that pan-BET inhibitors would be more useful as cancer therapeutics in IDH
mut glioma, though our TCGA data also points to the promise of BRD4-selective inhibitors in the treatment of IDH
mut glioma. Furthermore, this suggests the clinical benefit of BET inhibitors in the treatment of IDH
wt glioma may be limited. Accordingly, a recent Phase II clinical trial with the pan-BET inhibitor OTX015 was withdrawn due to lack of clinical activity (Hottinger et al.
2016).
Previous studies with JQ1 and other BET inhibitors in glioma and other cancers shows that the effects of BET inhibition are primarily antiproliferative while having little effect on cytotoxicity (Pastori et al.
2014; Filippakopoulos et al.
2010; Cheng et al.
2013). Indeed, in cell health studies using JQ1, we observed very little cytotoxicity but moderate antiproliferative effects in IDH
wt glioma cells. In contrast, we observed a JQ1-mediated selective cytotoxic response associated with
R132H IDH1 in our endogenous IDH
mut glioma cells (0905, BT142) and dox-inducible
R132H IDH1 model that is little observed in the IDH
wt counterpart (0827, 0923). Immunoblot and flow cytometric analysis supported an apoptotic mode of cell death in our IDH
mut cells. Interestingly, this cytotoxic response is delayed to 10 days incubation with BET inhibitor, with early (3 days) effects of BET inhibition being primarily antiproliferative. This is significant since most in vitro cytotoxicity assays using BET inhibitors utilize endpoints of 3 days or less (Cheng et al.
2013; Filippakopoulos et al.
2010; Pastori et al.
2014), whereas treatment with BET inhibitors in vivo can occur over a period of weeks to months (Pastori et al.
2014; Cheng et al.
2013; Hottinger et al.
2016; Lewin et al.
2018). In addition to the assessments of cytotoxicity, we observed a modest-to-major decrease in cellular proliferation in all cell lines treated with JQ1. We also observed a statistically significant, though modest, preferential decrease in cellular proliferation in our IDH
mut glioma cells compared to the IDH
wt counterpart. Therefore, our data indicates that IDH
mut glioma is preferentially reliant on BET protein function for maintenance of cellular survival and proliferation, and suggests moving onto in vivo models of glioma for further validation of this hypothesis.
Though our molecular data is not conclusive, it presents a potential rationale for the sensitivity of IDH
mut glioma to BET inhibition. BRD4 inhibition has been shown to inhibit transcriptional activation of the oncogene c-MYC (Delmore et al.
2011), which has been poised as particularly important for progression from low grade to high grade IDH
mut glioma (Bai et al.
2016; Odia et al.
2013). Also interesting are the effects of BET inhibition on PDGFRα activity. There is very little scientific literature regarding BET-mediated regulation of PDGFRα (Wang et al.
2015), but our data suggests that JQ1-mediated PDGFRα downregulation is not uniformly seen at early incubation timepoints (3 days), but becomes more pronounced and accompanied by decreased activation of p-PDGFRα at late timepoints (10 days). This suggests that though total PDGFRα downregulation may occur early in the time course of JQ1 treatment, the functional effects of PDGFRα downregulation are not observed until later during JQ1 treatment. Additionally, IDH
mut has been shown to activate PDGFRα signaling via epigenetic silencing of CTCF activity (Flavahan et al.
2016), and in our cell cultures PDGF-AB is supplemented in culture medium only in IDH
mut cells (0905, BT142). However, we were not able to ascertain the specific role decreased PDGFRα activation played in the cytotoxic and antiproliferative responses of our glioma cells to JQ1.
Additional molecular analysis revealed a selective downregulation of various active chromatin marks, notably H3K14/27Ac (Karmodiya et al.
2012; Creyghton et al.
2010) and H3S10Ph (Zippo et al.
2009; Karam et al.
2010), only in our IDH
mut glioma cells. This is particularly interesting as it suggests selective epigenetic, and therefore transcriptional, effects in our IDH
mut glioma cells. Through our limited molecular analysis via immunoblotting, we were unable to find any selective upregulation or downregulation of pathways that would corroborate selective transcriptional alterations in our IDH
mut glioma cells. This supports further analysis with more comprehensive, unbiased molecular techniques such as ATAC-Seq or ChIP-Seq, coupled with RNA-Seq, to identify unique changes in transcriptional programs in IDH
wt vs IDH
mut glioma in response to BET inhibition.
In this study, we provide clinical and in vitro evidence that IDHmut glioma exhibits a preferential reliance on BET protein activity to maintain cellular survival, proliferation, and tumor malignancy. Considering BET inhibitors have had limited success in the clinic, this study suggests repurposing these BET inhibitors for the treatment of IDHmut glioma may yield a promising therapeutic strategy.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.