Glioblastoma (GBM) constitutes the most common and aggressive primary brain tumor. To better understand how GBM evolves we analyzed longitudinal genomic and transcriptomic data of 114 patients. The analysis reveals a highly branched evolutionary pattern in which 63% of patients experience expression-based subtype changes. The branching pattern together with estimates of evolutionary rates suggest that the relapse associated clone typically preexisted years before diagnosis. 15% of tumors present hypermutations at relapse in highly expressed genes with a clear mutational signature. We find that 11% of recurrent tumors harbor mutations in LTBP4, a protein binding to TGF-β. Silencing LTBP4 in GBM cells leads to TGF-β activity suppression and decreased proliferation. In IDH1-wild-type recurrent GBM, high LTBP4 expression is associated with worse prognosis, highlighting the TGF-β pathway as a potential therapeutic target in GBM.

No abstract

Activating mutations in PIK3CA are frequent in human breast cancer, and phosphoinositide 3-kinase alpha (PI3Kα) inhibitors have been approved for therapy. To characterize determinants of sensitivity to these agents, we analyzed PIK3CA-mutant cancer genomes and observed the presence of multiple PIK3CA mutations in 12 to 15% of breast cancers and other tumor types, most of which (95%) are double mutations. Double PIK3CA mutations are in cis on the same allele and result in increased PI3K activity, enhanced downstream signaling, increased cell proliferation, and tumor growth. The biochemical mechanisms of dual mutations include increased disruption of p110α binding to the inhibitory subunit p85α, which relieves its catalytic inhibition, and increased p110α membrane lipid binding. Double PIK3CA mutations predict increased sensitivity to PI3Kα inhibitors compared with single-hotspot mutations.

Dissecting the pathogenesis of classical Hodgkin lymphoma (cHL), a common cancer in young adults, remains challenging because of the rarity of tumor cells in involved tissues (usually <5%). Here, we analyzed the coding genome of cHL by microdissecting tumor and normal cells from 34 patient biopsies for a total of ∼50 000 singly isolated lymphoma cells. We uncovered several recurrently mutated genes, namely, (32% of cases), (24%), (18%), and (16%), and document the functional role of mutant STAT6 in sustaining tumor cell viability. Mutations of genetically and functionally cooperated with disruption of, a JAK-STAT pathway inhibitor, to promote cHL growth. Overall, 87% of cases showed dysregulation of the JAK-STAT pathway by genetic alterations in multiple genes (also including ,, ,, and ), attesting to the pivotal role of this pathway in cHL pathogenesis and highlighting its potential as a new therapeutic target in this disease.

Precision medicine in cancer proposes that genomic characterization of tumors can inform personalized targeted therapies1–5. This proposition, however, is complicated by spatial and temporal heterogeneity6–14. Here we study genomic and expression profiles across 127 multi-sector or longitudinal specimens from 52 glioblastoma (GBM) patients. Using bulk and single-cell data, we find that samples from the same tumor mass share genomic and expression signatures, while geographically separated multifocal tumors and/or long-term recurrent tumors are seeded from different clones. Chemical screening of patient-derived glioma cells (PDCs) shows that therapeutic response is associated to genetic similarity, and multifocal tumors enriched with PIK3CA mutations have a heterogeneous drug response pattern. Importantly, we show that targeting truncal events is more efficacious in reducing tumor burden. In summary, this work demonstrates that evolutionary inference from integrated genomic analysis in multi-sector biopsies can inform targeted therapeutic interventions for GBM patients.

Atypical miRNA substrates do not fit criteria often used to annotate canonical miRNAs, and can escape the notice of miRNA genefinders. Recent analyses expanded the catalogs of invertebrate splicing-derived miRNAs (''mirtrons''), but only a few tens of mammalian mirtrons have been recognized to date. We performed meta-analysis of 737 mouse and human small RNA data sets comprising 2.83 billion raw reads. Using strict and conservative criteria, we provide confident annotation for 237 mouse and 240 human splicing-derived miRNAs, the vast majority of which are novel genes. These comprise three classes of splicing-derived miRNAs in mammals: conventional mirtrons, 59-tailed mirtrons, and 39-tailed mirtrons. In addition, we segregated several hundred additional human and mouse loci with candidate (and often compelling) evidence. Most of these loci arose relatively recently in their respective lineages. Nevertheless, some members in each of the three mirtron classes are conserved, indicating their incorporation into beneficial regulatory networks. We also provide the first Northern validation for mammalian mirtrons, and demonstrate Dicer-dependent association of mature miRNAs from all three classes of mirtrons with Ago2. The recognition of hundreds of mammalian mirtrons provides a new foundation for understanding the scope and evolutionary dynamics of Dicer substrates in mammals.[Supplemental material is available for this article.]Diverse pathways of conserved post-transcriptional gene regulation are mediated by Argonaute proteins and their guide, short RNAs. Among Argonaute-mediated small RNA pathways, the beststudied are the microRNAs (miRNAs). Generally speaking, miRNAs are ;21 to 24-nucleotide (nt) RNAs whose termini are precisely defined, and derive from precursor transcripts bearing one or more inverted repeats or hairpins (Axtell et al. 2011). The first miRNAs emerged from genetic studies of Caenorhabditis elegans developmental mutants (Lee et al. 1993;Reinhart et al. 2000), and were only recognized as noncoding loci upon their cloning. This set the stage for the directed identification of miRNA genes from cloned short RNAs (Lagos-Quintana et al. 2001;Lau et al. 2001;Lee and Ambros 2001). In animals, most miRNAs are generated by stepwise cleavage of primary miRNA transcripts ). These are processed in the nucleus by the Drosha RNase III enzyme to release an ;50-to 80-nt pre-miRNA hairpin, and again in the cytoplasm by a Dicer-class RNase III enzyme to yield a small RNA duplex. One of the strands is preferentially stably incorporated as a singlestranded RNA in an Argonaute (Ago) complex, and guides it to target transcripts (Czech and Hannon 2010).Although bioinformatic strategies have been used to identify miRNA genes Lim et al. 2003;Huang et al. 2007; van der Burgt et al. 2009), these have mostly been superseded by deep sequencing. This is in large part due to the fact that effective computational methods rely on comparative genomics and are ill-suited to identify species-specific miRNAs with reasonable specificit...

assisted with the initial computational analyses. H.Z. helped with the mouse in vivo experiment. K.J. assisted with patient selection. P.R. performed the nested control study, and assisted with patient sample procurement and survival analyses. P.R. and C.S. also performed the patient clinical annotation. J.S.R. viewed the FFPE slides, performed the laser microdissection and provided intellectual support. C.K. supervised the SWI/SNF complex ChIP-seq, helped with the SWI/SNF complex ChIP-seq data interpretation and provided intellectual insights.

Mirtrons are microRNA (miRNA) substrates that utilize the splicing machinery to bypass the necessity of Drosha cleavage for their biogenesis. Expanding our recent efforts for mammalian mirtron annotation, we use meta-analysis of aggregate datasets to identify ~500 novel mouse and human introns that confidently generate diced small RNA duplexes. These comprise nearly 1000 total loci distributed in four splicing-mediated biogenesis subclasses, with 5'-tailed mirtrons as, by far, the dominant subtype. Thus, mirtrons surprisingly comprise a substantial fraction of endogenous Dicer substrates in mammalian genomes. Although mirtron-derived small RNAs exhibit overall expression correlation with their host mRNAs, we observe a subset with substantial differences that suggest regulated processing or accumulation. We identify characteristic sequence, length, and structural features of mirtron loci that distinguish them from bulk introns, and find that mirtrons preferentially emerge from genes with larger numbers of introns. While mirtrons generate miRNA-class regulatory RNAs, we also find that mirtrons exhibit many features that distinguish them from canonical miRNAs. We observe that conventional mirtron hairpins are substantially longer than Drosha-generated pre-miRNAs, indicating that the characteristic length of canonical pre-miRNAs is not a general feature of Dicer substrate hairpins. In addition, mammalian mirtrons exhibit unique patterns of ordered 5' and 3' heterogeneity, which reveal hidden complexity in miRNA processing pathways. These include broad 3'-uridylation of mirtron hairpins, atypically heterogeneous 5' termini that may result from exonucleolytic processing, and occasionally robust decapitation of the 5' guanine (G) of mirtron-5p species defined by splicing. Altogether, this study reveals that this extensive class of non-canonical miRNA bears a multitude of characteristic properties, many of which raise general mechanistic questions regarding the processing of endogenous hairpin transcripts.