Neuronal differentiation and cell-cycle programs mediate response to BET-bromodomain inhibition in MYC-driven medulloblastoma

Abstract

BET-bromodomain inhibition (BETi) has shown pre-clinical promise for MYC-amplified medulloblastoma. However, the mechanisms for its action, and ultimately for resistance, have not been fully defined. Here, using a combination of expression profiling, genome-scale CRISPR/Cas9-mediated loss of function and ORF/cDNA driven rescue screens, and cell-based models of spontaneous resistance, we identify bHLH/homeobox transcription factors and cell-cycle regulators as key genes mediating BETi's response and resistance. Cells that acquire drug tolerance exhibit a more neuronally differentiated cell-state and expression of lineage-specific bHLH/homeobox transcription factors. However, they do not terminally differentiate, maintain expression of CCND2, and continue to cycle through S-phase. Moreover, CDK4/CDK6 inhibition delays acquisition of resistance. Therefore, our data provide insights about the mechanisms underlying BETi effects and the appearance of resistance and support the therapeutic use of combined cell-cycle inhibitors with BETi in MYC-amplified medulloblastoma.

Fig. 1

Integrative genomic analysis to identify mediators of BETi response. Schematic depicting overall analysis. Top panel: Treatment of medulloblastoma cells with BETi results in transcriptional suppression, contributing its phenotypic effects. Bottom panel: Integrative genomics approach incorporating genome-scale CRISPR-Cas9 screens to identify cell-essential genes, expression profiling following treatment with BETi to identify genes suppressed by BETi, and near genome-scale ORF rescue screens to identify genes that are required to be suppressed for BETi to exert its functional effect. The combination of all three assays identifies cell-essential genes that are suppressed by BETi, and whose suppression is necessary for decreased viability, thus mediating BETi response

Fig. 2

Rescue genes are enriched for cell-cycle regulators and bHLH/homeobox transcription factors. a Intersection of genes suppressed by 1 μM JQ1 with those identified to be cell essential (green) in D458 (top) and D283 (bottom). p-values indicate significance of overlap as determined with a Chi Square test. Source data: Supplementary Data Files 1 and 2. b Rescue genes identified in D458 (left) and D283 (right) cell lines following treatment with either JQ1 or IBET151. Asterisks indicate genes that scored as statistically significant rescue genes in both cell lines, but only met fold-change thresholds in the cell line shown. p-value indicates significance of overlap as determined by Fisher’s Exact Test. Source Data: Supplementary Data File 4. c STRING protein network analysis to identify direct and functional protein networks that exist between the entire set of candidate rescue ORFs identified across both cell lines. Protein-protein interactions (edges) between ORF rescue genes (nodes) that scored in either D458 of D283 following treatment with JQ1 or IBET151 are shown. p value indicates significance of enrichment of protein-protein interactions. Source Data: Supplementary Data File 4. d Venn diagram depicting overlap of genes that are suppressed by JQ1 (blue), score as dependencies in CRISPR-Cas9 screens (green) and are identified to be rescue genes (red) in D458 (top) or D283 (bottom). *CCND2 met both the q value threshold and the log-fold change threshold in D458, but only the q-value threshold in D283. Source data: Supplementary Data Files 1, 2, and 4

Fig. 3

Expression of cell-cycle regulators, anti-apoptosis genes and bHLH/homeobox transcription factors rescue BETi effects a Low throughput rescue assays in D458 and D283 cells expressing eGFP, CCND2, CCND3, BCL2L1, NEUROD1, NEUROG3, MYOD1, or MYOG that were treated with JQ1 1 μM or DMSO control. Asterisks denote significant differences from eGFP controls (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by Two-tailed unpaired t-tests. Error bars depict mean ± SEM. Source data: Source Data File. b Percentage (%) of apoptotic or dead cells in D458 and D283 cells expressing eGFP, CCND2, CCND3, BCL2L1, NEUROD1, or NEUROG3 treated with JQ1 2 μM for 72 h. Asterisks denote significant differences from eGFP controls (**p < 0.01, ***p < 0.001) as determined by two-tailed unpaired t-tests. Error bars depict mean ± SEM. Source data: Source Data File

Fig. 4

Drug-tolerant cells exhibit attenuated responses to BETi. a Percentage of viable cells among sensitive and drug-tolerant D425 and D458 populations after 72 h of treatment with JQ1. Error bars depict mean across six independent experiments ± SEM. ***p < 0.001 as determined by two-tailed unpaired t-tests. Source data: Source Data File. b Fold change in cell proliferation relative to pre-treatment baseline following 48 h of treatment with JQ1 among sensitive and drug-tolerant D425 and D458 cells that had been maintained in continuous JQ1 treatment (“drug-tolerant continuously treated”) or in which JQ1 had been withdrawn for 30 days (“drug-tolerant treatment withdrawn”). Data depict 20 replicate measurements across 4 independent experiments; error bars depict mean ± SEM. Asterisks (***) depicts p < 0.0001 as determined by two-tailed unpaired t-test. Source data: Source Data File. c Expression of genes in the JQ1 consensus signature in sensitive and drug-tolerant D458 cells following 24 h of treatment with JQ1 (1 μM) or vehicle control. Data from five independent replicates are shown, error bars depict mean ± SEM. Asterisks denote significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by two-tailed unpaired t-tests. Source data: Source Data File. d Expression of JQ1 medulloblastoma target genes in sensitive and drug-tolerant cells treated with JQ1 or vehicle control. Values represent mean of five replicate experiments ± SEM. Asterisks denote significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by two-tailed unpaired t-tests. Source data: Source Data File

Fig. 5

Drug tolerant cells express BETi mediator and rescue genes a Immunoblots probing for BETi-mediator proteins BCL2L1, CCND2, HLX in D458 sensitive and drug-tolerant cells following 24 h of treatment with JQ1 (1 μM) or vehicle control. Vinculin is included as a loading control. Source data: Source Data File. b Immunoblots probing for BETi-rescue proteins NEUROD1 and NEUROG1 in D458 sensitive and drug-tolerant cells following 24 h of treatment with JQ1 (1 μM) or vehicle control. Vinculin is included as a loading control. Source data: Source Data File. c Fold change in proliferation (as measured by luminescence) of D458 sensitive cells (top panel) and D458 drug tolerant medulloblastoma cells (bottom panel) following lentiviral infection with short hairpins targeting eGFP (negative control), SF3B1 (positive control) and BCL2L1. Values represent mean of 15 independent measurements across three replicate experiments ± SEM. Asterisks (***) denotes p values of <0.0001 as determined by two-tailed unpaired T-tests. Source data: Source Data File. d Fold change in proliferation (as measured by luminescence) of D458 sensitive cells (top panel) and D458 drug tolerant medulloblastoma cells (bottom panel) following lentiviral infection with short hairpins targeting eGFP (negative control), SF3B1 (positive control), CCND2 and NEUROD1. Values represent mean of 15 independent measurements across three replicate experiments ± SEM. Asterisks (***) denotes p values of < 0.0001 as determined by two-tailed unpaired t-tests. Source data: Source Data File

Fig. 6

Drug-tolerant cells exhibit changes in chromatin landscape and activating marks preferentially bind to rescue genes. a Heatmap representing abundances of differentially altered chromatin marks as identified by global chromatin profiling of sensitive and drug-tolerant D425 and D458 cells passaged in DMSO or BETi (JQ1 and IBET151) for 24 h. Red and blue respectively represent values above and below the median across each row. Source data: Supplementary Data File 8. b, c ChIP-seq binding scores (Z-transformed) of d H3K4me3 and e H3K27ac marks to rescue genes identified in ORF screens in sensitive cells treated with DMSO (n = 2) or 1 μM JQ1 (n = 2) and drug-tolerant cells passaged in 1 μM JQ1 (n = 4). Error bars depict mean ± SEM. Asterisks denote significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by two-tailed unpaired t-tests. Source data: Source Data File

Fig. 7

Drug tolerant cells exhibit altered cell state. a Immunoblots for neuronal differentiation and stem markers in sensitive and drug-tolerant D458 cells in the presence of JQ1 (1 μM) or DMSO control. Source data: Source Data File. b Percentage of BrdU positive D425 and D458 sensitive and drug-tolerant cells following 72 h of treatment with JQ1 (1 μM) or DMSO control. Data represent three independent experiments per cell line; error bars depict mean ± SEM. Asterisks denote significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by two-tailed unpaired t-tests. Source data: Source Data File

Fig. 8

Combination of BETi and CDK4/CDK6 attenuates resistance in flank models a Cumulative doubling of D458 cells passaged in DMSO, LEE011 (500 nM or 1 μM), JQ1 (1 μM) or the combination of JQ1 (1 μM) with LEE011 at either 500 nM or 1 μM. Data from three independent replicates are shown, error bars depict mean ± SEM. For all panels, asterisks denote statistically significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by two-tailed unpaired t-tests. Source data: Source Data File. b Growth of D458 flank xenografts treated with vehicle, JQ1 (50 mg/kg/day), LEE011 (75 mg/kg/day) or the combination of both compounds (JQ1 50 mg/kg/day and LEE011 75 mg/kg/day). Three vehicle treated mice were euthanized on day 15 due to their tumor size (shown by arrow). Data represent mean tumor volume (n = 10), normalized to day 1, ± SEM. Large circles denote time points with significant (p < 0.0001) improvement in tumor volume relative to DMSO controls. Asterisk denotes significant (p < 0.05) improvement in tumor volume among mice treated with JQ1 and LEE011 relative to those treated with JQ1 alone as determined by two-tailed unpaired t-tests. Source data: Source Data File. c Growth of MB002 flank xenografts treated with vehicle, JQ1 (50 mg/kg/day), LEE011 (75 mg/kg/day or the combination of both compounds (JQ1 50 mg/kg/day and LEE011 75 mg/kg/day). Arrows depict days in which mice were euthanized due to their tumor size. Data represent mean tumor volume (n = 10), normalized to day 1, ± SEM. Medium and large circles denote significant (p < 0.05 and p < 0.0001, respectively) differences between treated tumors at each time point relative to DMSO controls. Asterisks denote statistically significant improvements in tumor volume among mice treated with JQ1 and LEE011 relative to those treated with JQ1 alone (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by two-tailed unpaired t-tests. Source data: Source Data File. d Gene sets enriched in mRNA extracted from in vivo D458 flank xenografts treated with JQ1 relative to xenografts treated JQ1 and LEE011 (and thus suppressed by the combination of JQ1 and LEE011)

Fig. 9

CDK4/6 inhibition attenuates resistance to BETi in intracranial models a Kaplan–Meier survival curves of mice bearing intracranial D458 xenografts treated with vehicle, JQ1 (50 mg/kg/day), LEE011 (75 mg/kg/day) or the combination of both compounds (JQ1 50 mg/kg/day and LEE011 75 mg/kg/day). Asterisks denote statistically significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by log-rank (Mantel-Cox) tests. b Kaplan–Meier survival curves of mice bearing intracranial Med114 xenografts treated with vehicle, JQ1 (50 mg/kg/day), LEE011 (75 mg/kg/day) or the combination of both compounds (JQ1 50 mg/kg/day and LEE011 75 mg/kg/day). Asterisks denote statistically significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by log-rank (Mantel-Cox) tests. c Kaplan–Meier survival curves of mice bearing intracranial Med411 xenografts treated with vehicle, JQ1 (50 mg/kg/day), LEE011 (75 mg/kg/day) or the combination of both compounds (JQ1 50 mg/kg/day and LEE011 75 mg/kg/day). Asterisks denote statistically significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by log-rank (Mantel-Cox) tests

Fig. 10

Drug-tolerant cells exhibit an altered differentiation state that is also present in human medulloblastoma. a Model for altered differentiation in BETi-tolerant medulloblastoma cells. Human medulloblastomas express both stem-cell and neuronal differentiation markers. Following treatment with BETi, drug-tolerant cells exhibit a more neuronal phenotype however do not undergo terminal differentiation, maintaining the ability to progress through the cell-cycle. b Expression levels of TUJ1 (TUBB3) and JQ1 sensitivity of 46 central nervous system cell lines from the CTRP database. AUC denotes area under the curve. Asterisks denote statistically significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by a two-tailed unpaired t-test. Source data: Source Data File. c Representative images of three medulloblastomas stained for TUJ1 (green), SOX2 (red), NESTIN (yellow), OLIG2 (white) and DAPI (blue). Scale bar is 100 μm). Source data: Supplementary Data File 9. d Percentage of D458 sensitive and drug-tolerant cells that are TUJ1 positive and MSI1 negative in the presence of JQ1 (1 μM) or DMSO control. Data from three independent experiments are shown; error bars depict mean ± SEM. Asterisks denote statistically significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) as determined by a two-tailed unpaired t-test. Source data: Supplementary Data File 9. e Schematic depicting barcoding experiments with 600,000 barcode library. A pooled lentiviral library encompassing 600,000 unique DNA barcodes was used to label individual cells. Cells were expanded before being placed in replicate DMSO or JQ1 drug-treatment experiments. Barcodes were PCR amplified prior to treatment (ETP = early time point) and from replicate experiments at the drug of treatment and subjected to next-generation sequencing. f Percentage of shared DNA barcodes across replicates of barcoded D458 cells treated with JQ1 (2 μM). Within each replicate, the number of other replicates with whom each barcode is shared is depicted. Source data: Source Data File