Ewing sarcoma is an aggressive paediatric cancer of the bone and soft tissue. It results from a chromosomal translocation, predominantly t(11;22)(q24:q12), that fuses the N-terminal transactivation domain of the constitutively expressed EWSR1 protein with the C-terminal DNA binding domain of the rarely expressed FLI1 protein. Ewing sarcoma is highly sensitive to genotoxic agents such as etoposide, but the underlying molecular basis of this sensitivity is unclear. Here we show that Ewing sarcoma cells display alterations in regulation of damage-induced transcription, accumulation of R-loops and increased replication stress. In addition, homologous recombination is impaired in Ewing sarcoma owing to an enriched interaction between BRCA1 and the elongating transcription machinery. Finally, we uncover a role for EWSR1 in the transcriptional response to damage, suppressing R-loops and promoting homologous recombination. Our findings improve the current understanding of EWSR1 function, elucidate the mechanistic basis of the sensitivity of Ewing sarcoma to chemotherapy (including PARP1 inhibitors) and highlight a class of BRCA-deficient-like tumours.

Proteins are manufactured by ribosomes-macromolecular complexes of protein and RNA molecules that are assembled within major nuclear compartments called nucleoli 1,2. Existing models suggest that RNA polymerases I and III (Pol I and Pol III) are the only enzymes that directly mediate the expression of the ribosomal RNA (rRNA) components of ribosomes. Here we show, however, that RNA polymerase II (Pol II) inside human nucleoli operates near genes encoding rRNAs to drive their expression. Pol II, assisted by the neurodegeneration-associated enzyme senataxin, generates a shield comprising triplex nucleic acid structures known as R-loops at intergenic spacers flanking nucleolar rRNA genes. The shield prevents Pol I from producing sense intergenic noncoding RNAs (sincRNAs) that can disrupt nucleolar organization and rRNA expression. These disruptive sincRNAs can be unleashed by Pol II inhibition, senataxin loss, Ewing sarcoma or locus-associated R-loop repression through an experimental system involving the proteins RNaseH1, eGFP and dCas9 (which we refer to as 'red laser'). We reveal a nucleolar Pol-II-dependent mechanism that drives ribosome biogenesis, identify disease-associated disruption of nucleoli by noncoding RNAs, and establish locus-targeted R-loop modulation. Our findings revise theories of labour division between the major RNA polymerases, and identify nucleolar Pol II as a major factor in protein synthesis and nuclear organization, with potential implications for health and disease. Various proteins self-organize via liquid-liquid phase separation (LLPS) into nucleolar subdomains, which are needed for highly stereotyped ribosome assembly 1,2. At fibrillar centres in the heart of mammalian nucleoli, the major rRNA molecules needed to assemble ribosomes are generated by Pol-I-dependent transcription of rRNA genes within ribosomal DNA (rDNA) repeats 1,3. Within rDNA, rRNA genes are separated by large intergenic spacers (IGSs) (Extended Data Fig. 1a). At nucleolar rRNA genes, Pol I synthesizes precursor rRNAs (pre-rRNAs) that are processed into mature 28S, 18S and 5.8S rRNA molecules as they migrate to the granular component at the nucleolar periphery. Outside nucleoli, Pol III synthesizes 5S rRNA molecules that are targeted to nucleoli for processing. Mature rRNAs are packaged into 40S and 60S ribosomal subunits for export to the cytoplasm. Traditionally, the nucleolar Pol I and nucleoplasmic Pol III are viewed as the sole mammalian RNA polymerases that directly mediate housekeeping ribosome biogenesis. Interestingly, in the budding yeast Saccharomyces cerevisiae, Pol II is physically enriched at rDNA IGSs, but this phenomenon is deleterious because it drives ageing without affecting rRNA expression 3-5. It is unclear whether nucleolar Pol II exists in higher organisms or directly promotes ribosome biogenesis in any species. Active Pol II at rDNA IGSs To investigate whether Pol II exists within human nucleoli, we first used immunofluorescence coupled to super-resolution microscopy. Within nucleoli, which we...

BackgroundThe study of high-throughput genomic profiles from a pharmacogenomics viewpoint has provided unprecedented insights into the oncogenic features modulating drug response. A recent study screened for the response of a thousand human cancer cell lines to a wide collection of anti-cancer drugs and illuminated the link between cellular genotypes and vulnerability. However, due to essential differences between cell lines and tumors, to date the translation into predicting drug response in tumors remains challenging. Recently, advances in deep learning have revolutionized bioinformatics and introduced new techniques to the integration of genomic data. Its application on pharmacogenomics may fill the gap between genomics and drug response and improve the prediction of drug response in tumors.ResultsWe proposed a deep learning model to predict drug response (DeepDR) based on mutation and expression profiles of a cancer cell or a tumor. The model contains three deep neural networks (DNNs), i) a mutation encoder pre-trained using a large pan-cancer dataset (The Cancer Genome Atlas; TCGA) to abstract core representations of high-dimension mutation data, ii) a pre-trained expression encoder, and iii) a drug response predictor network integrating the first two subnetworks. Given a pair of mutation and expression profiles, the model predicts IC50 values of 265 drugs. We trained and tested the model on a dataset of 622 cancer cell lines and achieved an overall prediction performance of mean squared error at 1.96 (log-scale IC50 values). The performance was superior in prediction error or stability than two classical methods (linear regression and support vector machine) and four analog DNN models of DeepDR, including DNNs built without TCGA pre-training, partly replaced by principal components, and built on individual types of input data. We then applied the model to predict drug response of 9059 tumors of 33 cancer types. Using per-cancer and pan-cancer settings, the model predicted both known, including EGFR inhibitors in non-small cell lung cancer and tamoxifen in ER+ breast cancer, and novel drug targets, such as vinorelbine for TTN-mutated tumors. The comprehensive analysis further revealed the molecular mechanisms underlying the resistance to a chemotherapeutic drug docetaxel in a pan-cancer setting and the anti-cancer potential of a novel agent, CX-5461, in treating gliomas and hematopoietic malignancies.ConclusionsHere we present, as far as we know, the first DNN model to translate pharmacogenomics features identified from in vitro drug screening to predict the response of tumors. The results covered both well-studied and novel mechanisms of drug resistance and drug targets. Our model and findings improve the prediction of drug response and the identification of novel therapeutic options.

ETMRs are aggressive pediatric embryonal brain tumors with universally dismal outcome1. We collected 193 primary ETMRs and 23 matched relapses to investigate the genomic landscape of this distinct entity. We found that patients having tumors without C19MC amplification, the proposed driver [3][4][5] , frequently harbor DICER1 germline mutations or other miRNA-related aberrations including somatic miR-17-92 amplifications. Whole-genome sequencing revealed an overall low recurrence of SNVs, but prevalent R-loop-associated chromosomal instability, of which we show that this can be induced by loss of DICER1 function. Comparing primary tumors and matched relapses revealed a strong conservation of SVs but low conservation of SNVs. Moreover, many newly acquired SNVs are associated to a new cisplatin treatment related mutational signature. Finally, we show that targeting R-loops with topoisomerase and PARP inhibitors might be an effective treatment strategy for this deadly disease.Users may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research, subject always to the full Conditions of use:

Cohesin SA1 (STAG1) and SA2 (STAG2) are key components of the cohesin complex. Previous studies have highlighted the unique contributions by SA1 and SA2 to 3D chromatin organization, DNA replication fork progression, and DNA double-strand break (DSB) repair. Recently, we discovered that cohesin SA1 and SA2 are DNA binding proteins. Given the recently discovered link between SA2 and RNA-mediated biological pathways, we investigated whether or not SA1 and SA2 directly bind to RNA using a combination of bulk biochemical assays and single-molecule techniques, including atomic force microscopy (AFM) and the DNA tightrope assay. We discovered that both SA1 and SA2 bind to various RNA containing substrates, including ssRNA, dsRNA, RNA:DNA hybrids, and R-loops. Importantly, both SA1 and SA2 localize to regions on dsDNA that contain RNA. We directly compared the SA1/SA2 binding and R-loops sites extracted from Chromatin Immunoprecipitation sequencing (ChIP-seq) and DNA-RNA Immunoprecipitation sequencing (DRIP-Seq) data sets, respectively. This analysis revealed that SA1 and SA2 binding sites overlap significantly with R-loops. The majority of R-loop-localized SA1 and SA2 are also sites where other subunits of the cohesin complex bind. These results provide a new direction for future investigation of the diverse biological functions of SA1 and SA2.