Dependence on the 26S proteasome is an Achilles’ heel for triple-negative breast cancer (TNBC) and multiple myeloma (MM). The therapeutic proteasome inhibitor, bortezomib, successfully targets MM but often leads to drug-resistant disease relapse and fails in breast cancer. Here we show that a 26S proteasome-regulating kinase, DYRK2, is a therapeutic target for both MM and TNBC. Genome editing or small-molecule mediated inhibition of DYRK2 significantly reduces 26S proteasome activity, bypasses bortezomib resistance, and dramatically delays in vivo tumor growth in MM and TNBC thereby promoting survival. We further characterized the ability of LDN192960, a potent and selective DYRK2-inhibitor, to alleviate tumor burden in vivo. The drug docks into the active site of DYRK2 and partially inhibits all 3 core peptidase activities of the proteasome. Our results suggest that targeting 26S proteasome regulators will pave the way for therapeutic strategies in MM and TNBC.

SUMMARYExtrachromosomal circular DNA (ecDNA) is an important driver of aggressive tumor growth, promoting high oncogene copy number, intratumoral heterogeneity, accelerated evolution of drug resistance, enhancer rewiring, and poor outcome. ecDNA has been reported in medulloblastoma (MB), the most common malignant pediatric brain tumor, but the ecDNA landscape and its association with specific MB subgroups, its impact on enhancer rewiring, and its potential clinical implications, are not known. We assembled a retrospective cohort of 468 MB patient samples with available whole genome sequencing (WGS) data covering the four major MB subgroups WNT, SHH, Group 3 and Group 4. Using computational methods for the detection and reconstruction of ecDNA1, we find ecDNA in 82 patients (18%) and observe that ecDNA+ MB patients are more than twice as likely to relapse and three times as likely to die of disease. In addition, we find that individual medulloblastoma tumors often harbor multiple ecDNAs, each containing different amplified oncogenes along with co-amplified non-coding regulatory enhancers. ecDNA was substantially more prevalent among 31 analyzed patient-derived xenograft (PDX) models and cell lines than in our patient cohort. By mapping the accessible chromatin and 3D conformation landscapes of MB tumors that harbor ecDNA, we observe frequent candidate “enhancer rewiring” events that spatially link oncogenes with co-amplified enhancers. Our study reveals the frequency and diversity of ecDNA in a subset of highly aggressive tumors and suggests enhancer rewiring as a frequent oncogenic mechanism of ecDNAs in MB. Further, these results demonstrate that ecDNA is a frequent and potent driver of poor outcome in MB patients.

Extrachromosomal circular DNA (ecDNA) is an important driver of particularly aggressive human cancers. However, the prevalence of ecDNA, and its role in tumor development and progression in the different molecular subgroups of medulloblastoma (MB), remain unknown. To answer these questions, we have assembled a multi-institutional retrospective cohort of 472 MB patients with available whole genome sequencing (WGS) data, drawing from three cancer genomic data repositories and covering all MB subgroups (WNT, SHH, Group 3 and Group 4). Using recent computational methods to detect and reconstruct ecDNA, we find ecDNA in 66 patients (14%) and observe that the presence of ecDNA is associated with significantly poorer outcomes. By subgroup, ecDNA was found in 0/24 WNT (0%), 22/109 SHH (20%), 15/107 Group 3 (14%) and 20/181 Group 4 (11%) patients. Affected genomic loci harbor up to hundredfold amplification of oncogenes including MYC, MYCN, TERT, and other novel putative oncogenes. We further analyzed 24 patient-derived xenograft (PDX) and four cell line models of MB tumors. ecDNA was substantially more frequent in patient-derived models (17 of 29, 59%) than in our patient cohort. To elucidate the functional regulatory landscapes of ecDNAs in MB, we generated transcriptional (RNA-seq), accessible chromatin (ATAC-seq), and chromatin interaction (Hi-C) profiles of 6 MB tumor samples. In each case, we identify regulatory interactions that cross fusion breakpoints on the ecDNA, representing potential “enhancer rewiring” events which may contribute to transcriptional activation of co-amplified oncogenes. To test this hypothesis, we are currently conducting in-vitro CRISPRi screens targeting regulatory regions on the ecDNA of a MB cell line to determine whether these enhancers promote proliferation. In summary, our study analyzes the frequency, diversity and functional relevance of ecDNA across MB subgroups and provides strong justification for continued mechanistic studies of ecDNA in MB with the potential to uncover new therapeutic approaches.

Ependymoma is a tumor of the brain or spinal cord. The two most common and aggressive molecular groups of ependymoma are the supratentorial ZFTA-fusion associated and the posterior fossa ependymoma group A. In both groups, tumors occur mainly in young children and frequently recur after treatment. Although molecular mechanisms underlying these diseases have recently been uncovered, they remain difficult to target and innovative therapeutic approaches are urgently needed. Here, we use genome-wide chromosome conformation capture (Hi-C), complemented with CTCF and H3K27ac ChIP-seq, as well as gene expression and DNA methylation analysis in primary and relapsed ependymoma tumors, to identify chromosomal conformations and regulatory mechanisms associated with aberrant gene expression. In particular, we observe the formation of new topologically associating domains (‘neo-TADs’) caused by structural variants, group-specific 3D chromatin loops, and the replacement of CTCF insulators by DNA hyper-methylation. Through inhibition experiments, we validate that genes implicated by these 3D genome conformations are essential for the survival of patient-derived ependymoma models in a group-specific manner. Thus, this study extends our ability to reveal tumor-dependency genes by 3D genome conformations even in tumors that lack targetable genetic alterations.

By profiling enhancers in primary ependymoma tumors, we have recently identified putative oncogenes, molecular targets, and functional pathways. Inhibition of selected targets diminished the proliferation of patient-derived neurospheres and increased survival in mouse models of ependymoma. While enhancers frequently regulate the nearest gene, unambiguous identification of enhancer target genes remains to be a challenge in the absence of chromosome conformation information. Consequently, we have now used HiC to map the 3-dimensional organization of tumor chromatin in the two most common and aggressive ependymoma subgroups: posterior fossa group A (PF-EPN-A) and supratentorial ependymomas with gene fusions involving the NF-κB subunit gene RELA (ST-EPN-RELA). By an integrative analysis of enhancer and gene expression in the context of the newly derived HiC data, we find that a large amount of the previously predicted enhancer target genes can be confirmed by physical interactions. Importantly, we also identify many new putative tumor-dependency genes activated by long-range promoter-enhancer interactions. Complementary to the analysis of gene-enhancer interactions, we have also leveraged the HiC data for resolving structural rearrangements underlying copy number alterations frequently observed in PF-EPN-A tumors. Especially copy number gains of the 1q arm of chromosome 1 are associated with poor survival. Our preliminary results reveal complex structural variants that underlie 1q gains, which lead to inter-chromosomal rearrangements and affect several genes that potentially contribute to poor survival. We now aim to test the relevance of the novel candidate tumor-dependency genes for tumor cell growth and proliferation in patient derived ependymoma models.

Medulloblastoma is a malignant brain tumor that occurs in the cerebellum, most frequently in children. Medulloblastoma is molecularly classified into four major groups, and therapies are now being developed according to the nature of these groups and subgroups. However, there are currently no effective molecularly targeted drugs for most of these groups. In recent years, we have been analyzing the genomes of medulloblastomas to identify genetic mutations involved in tumorigenesis. Among them, mutations in chromatin modifiers are frequently detected in medulloblastoma, suggesting the importance of alterations in the epigenome in tumor development. However, it remains unclear how epigenomic changes are involved in tumorigenesis. Here, we have used the SHH-group medulloblastoma (MB-SHH) mouse model to understand the epigenomic changes during tumor development and identify new therapeutic targets for medulloblastoma. To this end, we utilized an Atoh1-EGFP:Ptch1+/- mice that specifically label cerebellar granule cell progenitors (GNPs) that are known to be the cellular origin of MB-SHH, preneoplastic (PNCs) and tumor cells with EGFP during MB-SHH formation. Following FACS-based collection of EGFP-positive cells, comparative ATAC-seq analysis revealed that the open chromatin regions dynamically changed during transformation of GNPs into PNCs with enrichment of nuclear factor I (NFI) binding motifs. Cut & Tag analysis on these cells showed that NFI proteins bound chromatin regions that become more accessible during tumorigenesis, suggesting that NFI could play an important role in tumorigenesis after the epigenomic changes. Indeed, CRISPR-engineered in vivo somatic deletion of Nfia and/or Nfib prevented hyperplasia formation, confirming their essential role in tumor initiation. Knockdown of either NFIA or NFIB in patient-derived xenograft models also attenuated tumor growth. Thus, our study has uncovered a novel oncogenic mechanism that contributes to the development of MB-SHH tumors via alterations of accessible chromatin and aberrant DNA binding of NFI transcription factors.

Ependymoma is a tumor of the brain or spinal cord. The two most common and aggressive molecular groups of ependymoma are the supratentorial RELA-fusion associated group and the posterior fossa ependymoma group A. In both groups, tumors occur mainly in young children and frequently recur after treatment1. Although the molecular mechanisms underlying these diseases have recently been uncovered, they remain difficult to target and innovative therapeutic approaches are urgently needed. Here, we use genome-wide chromosome conformation capture (Hi-C), complemented with CTCF (insulators) and H3K27ac (active enhancers) ChIP-seq as well as gene expression and whole-genome DNA methylation profiling in primary and relapsed ependymoma tumors and cell lines to identify chromosomal rearrangements and regulatory mechanisms underlying aberrant expression of genes that are essential for ependymoma tumorigenesis. In particular, we observe the formation of new topologically associating domains (‘neo-TADs’) by intra- and inter-chromosomal structural variants, tumor-specific 3D chromatin complexes of regulatory elements, and the replacement of CTCF insulators by DNA hyper-methylation as novel oncogenic mechanisms in ependymoma. Through inhibition experiments we validated that the newly identified genes, including RCOR2, ITGA6, LAMC1, and ARL4C, are highly essential for the survival of patient-derived ependymoma models in a disease subtype-specific manner. Thus, this study identifies potential novel therapeutic vulnerabilities in ependymoma and extends our ability to reveal tumor-dependency genes and pathways by oncogenic 3D genome conformations even in tumors that lack known genetic alterations.

By profiling enhancers in primary ependymoma tumors, we have recently identified putative oncogenes, molecular targets, and functional pathways. Inhibition of selected targets diminished the proliferation of patient-derived neurospheres and increased survival in mouse models of ependymoma. While enhancers frequently regulate the nearest gene, identification of enhancer target genes remains to be a challenge in the absence of chromosome conformation information. Consequently, we have now used HiC to map the 3-dimensional organization of tumor chromatin in the two most common and aggressive ependymoma subgroups: posterior fossa group A (PF-EPN-A) and supratentorial ependymomas with gene fusions involving the NF-κB subunit gene RELA (ST-EPN-RELA). By an integrative analysis of enhancer and gene expression in the context of the newly derived HiC data, we find that a large number of the predicted enhancer target genes are enriched for strong physical interactions. Importantly, we also identify many new putative tumor-dependency genes activated by long-range promoter-enhancer interactions and complex tumor-specific chromatin clusters of regulatory elements. Complementary to the analysis of gene-enhancer interactions, we have also leveraged the HiC data for resolving structural rearrangements underlying copy number alterations. Copy number gains of the 1q arm of chromosome 1 are especially associated with poor survival. Our preliminary results in PFA relapse samples show complex structural variants underlying 1q gain that lead to inter-chromosomal rearrangements and affect several genes that potentially contribute to poor survival. In ongoing work we are testing the relevance of the novel candidate genes for tumor cell growth and proliferation in-patient derived ependymoma models.