Glioblastoma is a universally lethal cancer with a median survival time of approximately 15 months1. Despite substantial efforts to define druggable targets, there are no therapeutic options that notably extend the lifespan of patients with glioblastoma. While previous work has largely focused on in vitro cellular models, here we demonstrate a more physiologically relevant approach to target discovery in glioblastoma. We adapted pooled RNA interference (RNAi) screening technology2,3,4 for use in orthotopic patient-derived xenograft models, creating a high-throughput negative-selection screening platform in a functional in vivo tumour microenvironment. Using this approach, we performed parallel in vivo and in vitro screens and discovered that the chromatin and transcriptional regulators needed for cell survival in vivo are non-overlapping with those required in vitro. We identified transcription pause–release and elongation factors as one set of in vivo-specific cancer dependencies, and determined that these factors are necessary for enhancer-mediated transcriptional adaptations that enable cells to survive the tumour microenvironment. Our lead hit, JMJD6, mediates the upregulation of in vivo stress and stimulus response pathways through enhancer-mediated transcriptional pause–release, promoting cell survival specifically in vivo. Targeting JMJD6 or other identified elongation factors extends survival in orthotopic xenograft mouse models, suggesting that targeting transcription elongation machinery may be an effective therapeutic strategy for glioblastoma. More broadly, this study demonstrates the power of in vivo phenotypic screening to identify new classes of ‘cancer dependencies’ not identified by previous in vitro approaches, and could supply new opportunities for therapeutic intervention.
Access optionsAccess options
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Gene Expression Omnibus
The authors thank the Hemann, Paddison, and Zuber laboratories, along with K. Alazem for technical help with RNAi screening. We thank B. Kristensen and A. M. Rosager for input on histological analysis and A. Saiakhova, E. R. Chan, M. Squatrito and W. Liu for bioinformatic support. Additional support was provided by the Cytometry & Imaging Microscopy and Genomics core facilities of the Case Comprehensive Cancer Center at Cleveland Clinic and Case Western (P30CA043703), respectively. We also thank members of the Rich, Tesar and Scacheri laboratories for input on the project and discussion about the manuscript. This work was supported by Velosano (J.N.R.); New York Stem Cell Foundation-Robertson Investigator Award and philanthropic support from the Goodman family (P.J.T.); CIHR Banting Fellowship (S.C.M.); National Institutes of Health grants CA183510 (T.E.M.); GM007250 (T.E.M., A.R.M., L.J.Y.K., J.J.M.); CA189647 (C.G.H.); CA154130, CA169117, CA197718, CA171652, NS087913 and NS089272 (J.N.R.).
Extended data figures
Primary Screen Results. Worksheet 1 contains the normalized RPM (reads per million) values for each hairpin for each replicate from the in vivo screen (13-14 individual animals per replicate). Median RPM value for each replicate was used for analysis. Signal to Noise of replicates was used to calculate individual shRNA score based on their ability to deplete cells in the induced arm compared to the control arm. Worksheet 2 contains the ranked list of genes from in vivo screen. 2nd-Best shRNA score from each expressed gene was used to rank genes for hits. 2nd-Best shRNA score of 2 or greater was considered a hit. Worksheet 3 and 4 contain the same information for the in vitro screen.
RNAseq Results. RNA sequencing results for GBM528 and GBM3565 from in vivo and in vitro conditions. All expressed genes listed. FPKM (Fragments Per Kilobase of transcript per Million mapped reads) values for known genes were calculated using Cufflinks. FPKM values were quantile normalized. To generate an expressed genes list, an average of replicates for each condition was calculated and genes with FPKM value >0.25 in either intracranial sample or culture sample were considered expressed. Genes that did not meet this expression cut-off (Replicate average FPKM <0.25 in Intracranial and in Culture conditions) were removed as not expressed. Expressed genes were tabled by converting FPKMs <0.25 to 0.25.
Gene Ontology Enrichment Values. Raw outputs for gene ontology enrichments. Each worksheet contains program, settings, and files used to generate output. Last worksheet contains the GSEA genesets analyzed when using GSEA.
Super-Enhancer List for GBM528. Lists of all Super-Enhancers detected in in vivo intracranial condition (IC) (worksheet 1) and in the in vitro culture condition (worksheet 2). In vivo and in vitro SEs were identified using the dynamic Enhancer software (retrieved from https://github.com/BradnerLab/pipeline). H3K27Ac peaks identified as enhancers and separated by less than 12.5 kb were stitched together. All stitched peaks were then ranked by the density of H3K27Ac minus input. Peaks higher than the inflection point on the density curve were designated SEs. Worksheet 3 contains called condition-specific enhancers. To call condition-specific SEs, SEs from in vivo and in vitro conditions were merged and H3K27Ac signals for all merged SEs in each cell line were calculated. To be considered specific to in vivo or in vitro, SEs had to be only called a SE in a single condition and had to have at least 1.5-fold change in H3K27Ac signal between conditions.
JMJD6 correlation with TCGA genes. List of all gene expression correlations with JMJD6 in TCGA primary tumours. To test correlation of all genes with JMJD6 across glioblastoma tumours, TCGA RNAseqV2 data were downloaded and analysed. All glioblastoma tumours in the TCGA with RNA-seq data available were used.
About this article
Nature Reviews Molecular Cell Biology (2018)