Joshua Gould

GenePattern 2.0

Stephen J Chapman1,2, Chiea C Khor1, Fredrik O Vannberg1, Nicholas A Maskell2, Christopher WH Davies3, Emma L Hedley2, Shelley Segal4, Catrin E Moore4, Kyle Knox5, Nicholas P Day6, Stephen H Gillespie7, Derrick W Crook5, Robert JO Davies2 & Adrian VS Hill1 1The Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford OX3 7BN, UK. 2Oxford Centre for Respiratory Medicine, Churchill Hospital Site, Oxford Radcliffe Hospital, Oxford OX3 7LJ, UK. 3Department of Respiratory Medicine, Royal Berkshire Hospital, Reading RG1 5AN, UK. 4Department of Paediatrics, John Radcliffe Hospital, Oxford OX3 9DU, UK. 5Department of Microbiology, John Radcliffe Hospital, Oxford OX3 9DU, UK. 6Centre for Clinical Vaccinology and Tropical Medicine, Oxford OX3 9DU, UK. 7Centre for Medical Microbiology, Department of Infection, University College London, London NW1 2BU, UK. e-mail: adrian.hill@well.ox.ac.uk

Tumour micro-environment elicits innate resistance to RAF inhibitors through HGF secretion

Drug resistance presents a challenge to the treatment of cancer patients. Many studies have focused on cell-autonomous mechanisms of drug resistance. By contrast, we proposed that the tumour micro-environment confers innate resistance to therapy. Here we developed a co-culture system to systematically assay the ability of 23 stromal cell types to influence the innate resistance of 45 cancer cell lines to 35 anticancer drugs. We found that stroma-mediated resistance is common, particularly to targeted agents. We characterized further the stroma-mediated resistance of BRAF-mutant melanoma to RAF inhibitors because most patients with this type of cancer show some degree of innate resistance. Proteomic analysis showed that stromal cell secretion of hepatocyte growth factor (HGF) resulted in activation of the HGF receptor MET, reactivation of the mitogen-activated protein kinase (MAPK) and phosphatidylinositol-3-OH kinase (PI(3)K)–AKT signalling pathways, and immediate resistance to RAF inhibition. Immunohistochemistry experiments confirmed stromal cell expression of HGF in patients with BRAF-mutant melanoma and showed a significant correlation between HGF expression by stromal cells and innate resistance to RAF inhibitor treatment. Dual inhibition of RAF and either HGF or MET resulted in reversal of drug resistance, suggesting RAF plus HGF or MET inhibitory combination therapy as a potential therapeutic strategy for BRAF-mutant melanoma. A similar resistance mechanism was uncovered in a subset of BRAF-mutant colorectal and glioblastoma cell lines. More generally, this study indicates that the systematic dissection of interactions between tumours and their micro-environment can uncover important mechanisms underlying drug resistance.Drug resistance presents a challenge to the treatment of cancer patients. Many studies have focused on cell-autonomous mechanisms of drug resistance. By contrast, we proposed that the tumour micro-environment confers innate resistance to therapy. Here we developed a co-culture system to systematically assay the ability of 23 stromal cell types to influence the innate resistance of 45 cancer cell lines to 35 anticancer drugs. We found that stroma-mediated resistance is common, particularly to targeted agents. We characterized further the stroma-mediated resistance of BRAF-mutant melanoma to RAF inhibitors because most patients with this type of cancer show some degree of innate resistance. Proteomic analysis showed that stromal cell secretion of hepatocyte growth factor (HGF) resulted in activation of the HGF receptor MET, reactivation of the mitogen-activated protein kinase (MAPK) and phosphatidylinositol-3-OH kinase (PI(3)K)-AKT signalling pathways, and immediate resistance to RAF inhibition. Immunohistochemistry experiments confirmed stromal cell expression of HGF in patients with BRAF-mutant melanoma and showed a significant correlation between HGF expression by stromal cells and innate resistance to RAF inhibitor treatment. Dual inhibition of RAF and either HGF or MET resulted in reversal of drug resistance, suggesting RAF plus HGF or MET inhibitory combination therapy as a potential therapeutic strategy for BRAF-mutant melanoma. A similar resistance mechanism was uncovered in a subset of BRAF-mutant colorectal and glioblastoma cell lines. More generally, this study indicates that the systematic dissection of interactions between tumours and their micro-environment can uncover important mechanisms underlying drug resistance.

Discovery and prioritization of somatic mutations in diffuse large B-cell lymphoma (DLBCL) by whole-exome sequencing

To gain insight into the genomic basis of diffuse large B-cell lymphoma (DLBCL), we performed massively parallel whole-exome sequencing of 55 primary tumor samples from patients with DLBCL and matched normal tissue. We identified recurrent mutations in genes that are well known to be functionally relevant in DLBCL, including MYD88, CARD11, EZH2, and CREBBP. We also identified somatic mutations in genes for which a functional role in DLBCL has not been previously suspected. These genes include MEF2B, MLL2, BTG1, GNA13, ACTB, P2RY8, PCLO, and TNFRSF14. Further, we show that BCL2 mutations commonly occur in patients with BCL2/IgH rearrangements as a result of somatic hypermutation normally occurring at the IgH locus. The BCL2 point mutations are primarily synonymous, and likely caused by activation-induced cytidine deaminase–mediated somatic hypermutation, as shown by comprehensive analysis of enrichment of mutations in WRCY target motifs. Those nonsynonymous mutations that are observed tend to be found outside of the functionally important BH domains of the protein, suggesting that strong negative selection against BCL2 loss-of-function mutations is at play. Last, by using an algorithm designed to identify likely functionally relevant but infrequent mutations, we identify KRAS, BRAF, and NOTCH1 as likely drivers of DLBCL pathogenesis in some patients. Our data provide an unbiased view of the landscape of mutations in DLBCL, and this in turn may point toward new therapeutic strategies for the disease.

GSEA-P

UNLABELLED Gene Set Enrichment Analysis (GSEA) is a computational method that assesses whether an a priori defined set of genes shows statistically significant, concordant differences between two biological states. We report the availability of a new version of the Java based software (GSEA-P 2.0) that represents a major improvement on the previous release through the addition of a leading edge analysis component, seamless integration with the Molecular Signature Database (MSigDB) and an embedded browser that allows users to search for gene sets and map them to a variety of microarray platform formats. This functionality makes it possible for users to directly import gene sets from MSigDB for analysis with GSEA. We have also improved the visualizations in GSEA-P 2.0 and added links to a new form of concise gene set annotations called Gene Set Cards. These additions, as well as other improvements suggested by over 3500 users who have downloaded the software over the past year have been incorporated into this new release of the GSEA-P Java desktop program. AVAILABILITY GSEA-P 2.0 is freely available for academic and commercial users and can be downloaded from http://www.broad.mit.edu/GSEA

Widespread Genetic Heterogeneity in Multiple Myeloma: Implications for Targeted Therapy

We performed massively parallel sequencing of paired tumor/normal samples from 203 multiple myeloma (MM) patients and identified significantly mutated genes and copy number alterations and discovered putative tumor suppressor genes by determining homozygous deletions and loss of heterozygosity. We observed frequent mutations in KRAS (particularly in previously treated patients), NRAS, BRAF, FAM46C, TP53, and DIS3 (particularly in nonhyperdiploid MM). Mutations were often present in subclonal populations, and multiple mutations within the same pathway (e.g., KRAS, NRAS, and BRAF) were observed in the same patient. In vitro modeling predicts only partial treatment efficacy of targeting subclonal mutations, and even growth promotion of nonmutated subclones in some cases. These results emphasize the importance of heterogeneity analysis for treatment decisions.

Systematic investigation of genetic vulnerabilities across cancer cell lines reveals lineage-specific dependencies in ovarian cancer

A comprehensive understanding of the molecular vulnerabilities of every type of cancer will provide a powerful roadmap to guide therapeutic approaches. Efforts such as The Cancer Genome Atlas Project will identify genes with aberrant copy number, sequence, or expression in various cancer types, providing a survey of the genes that may have a causal role in cancer. A complementary approach is to perform systematic loss-of-function studies to identify essential genes in particular cancer cell types. We have begun a systematic effort, termed Project Achilles, aimed at identifying genetic vulnerabilities across large numbers of cancer cell lines. Here, we report the assessment of the essentiality of 11,194 genes in 102 human cancer cell lines. We show that the integration of these functional data with information derived from surveying cancer genomes pinpoints known and previously undescribed lineage-specific dependencies across a wide spectrum of cancers. In particular, we found 54 genes that are specifically essential for the proliferation and viability of ovarian cancer cells and also amplified in primary tumors or differentially overexpressed in ovarian cancer cell lines. One such gene, PAX8, is focally amplified in 16% of high-grade serous ovarian cancers and expressed at higher levels in ovarian tumors. Suppression of PAX8 selectively induces apoptotic cell death of ovarian cancer cells. These results identify PAX8 as an ovarian lineage-specific dependency. More generally, these observations demonstrate that the integration of genome-scale functional and structural studies provides an efficient path to identify dependencies of specific cancer types on particular genes and pathways.

Harnessing Connectivity in a Large-Scale Small-Molecule Sensitivity Dataset

UNLABELLED Identifying genetic alterations that prime a cancer cell to respond to a particular therapeutic agent can facilitate the development of precision cancer medicines. Cancer cell-line (CCL) profiling of small-molecule sensitivity has emerged as an unbiased method to assess the relationships between genetic or cellular features of CCLs and small-molecule response. Here, we developed annotated cluster multidimensional enrichment analysis to explore the associations between groups of small molecules and groups of CCLs in a new, quantitative sensitivity dataset. This analysis reveals insights into small-molecule mechanisms of action, and genomic features that associate with CCL response to small-molecule treatment. We are able to recapitulate known relationships between FDA-approved therapies and cancer dependencies and to uncover new relationships, including for KRAS-mutant cancers and neuroblastoma. To enable the cancer community to explore these data, and to generate novel hypotheses, we created an updated version of the Cancer Therapeutic Response Portal (CTRP v2). SIGNIFICANCE We present the largest CCL sensitivity dataset yet available, and an analysis method integrating information from multiple CCLs and multiple small molecules to identify CCL response predictors robustly. We updated the CTRP to enable the cancer research community to leverage these data and analyses.

Comparative gene marker selection suite

MOTIVATION An important step in analyzing expression profiles from microarray data is to identify genes that can discriminate between distinct classes of samples. Many statistical approaches for assigning significance values to genes have been developed. The Comparative Marker Selection suite consists of three modules that allow users to apply and compare different methods of computing significance for each marker gene, a viewer to assess the results, and a tool to create derivative datasets and marker lists based on user-defined significance criteria. AVAILABILITY The Comparative Marker Selection application suite is freely available as a GenePattern module. The GenePattern analysis environment is freely available at http://www.broad.mit.edu/genepattern.

Prognostic Gene Expression Signature for Patients With Hepatitis C–Related Early-Stage Cirrhosis

BACKGROUND & AIMS Cirrhosis affects 1% to 2% of the world population and is the major risk factor for hepatocellular carcinoma (HCC). Hepatitis C cirrhosis-related HCC is the most rapidly increasing cause of cancer death in the United States. Noninvasive methods have been developed to identify patients with asymptomatic early-stage cirrhosis, increasing the burden of HCC surveillance, but biomarkers are needed to identify patients with cirrhosis who are most in need of surveillance. We investigated whether a liver-derived 186-gene signature previously associated with outcomes of patients with HCC is prognostic for patients with newly diagnosed cirrhosis but without HCC. METHODS We performed gene expression profile analysis of formalin-fixed needle biopsy specimens from the livers of 216 patients with hepatitis C-related early-stage (Child-Pugh class A) cirrhosis who were prospectively followed up for a median of 10 years at an Italian center. We evaluated whether the 186-gene signature was associated with death, progression of cirrhosis, and development of HCC. RESULTS Fifty-five (25%), 101 (47%), and 60 (28%) patients were classified as having poor-, intermediate-, and good-prognosis signatures, respectively. In multivariable Cox regression modeling, the poor-prognosis signature was significantly associated with death (P = .004), progression to advanced cirrhosis (P < .001), and development of HCC (P = .009). The 10-year rates of survival were 63%, 74%, and 85% and the annual incidence of HCC was 5.8%, 2.2%, and 1.5% for patients with poor-, intermediate-, and good-prognosis signatures, respectively. CONCLUSIONS A 186-gene signature used to predict outcomes of patients with HCC is also associated with outcomes of patients with hepatitis C-related early-stage cirrhosis. This signature might be used to identify patients with cirrhosis in most need of surveillance and strategies to prevent the development of HCC.

The Drug Repurposing Hub: a next-generation drug library and information resource

To the Editor: Drug repurposing, the application of an existing therapeutic to a new disease indication, holds promise of rapid clinical impact at a lower cost than de novo drug development. So far, there has not been a systematic effort to identify such opportunities, limited in part by the lack of a comprehensive library of clinical compounds suitable for testing. To address this challenge, we hand-curated a collection of 4,707 compounds, experimentally confirmed their identities, and annotated them with literature-reported targets. The collection includes 3,422 drugs that are marketed around the world or that have been tested in human clinical trials. Compounds were obtained from more than 50 chemical vendors, and the purity of each sample was established. We have thus established a blueprint for others to easily assemble such a repurposing library, and we have created an online Drug Repurposing Hub (http:// www.broadinstitute.org/repurposing) that contains detailed annotation for each of the compounds. Repurposing is attractive and pragmatic, given the substantial cost and time requirements—on average, a decade or more—for drug development1. In addition, a large number of potential drugs never reach clinical testing. Moreover, fewer than 15% of compounds that enter clinical development ultimately receive approval, despite the majority of them being deemed safe2. For either approved or failed drugs for which safety has already been established, finding new indications can rapidly bring benefits to patients. Prior drug-repurposing successes span disease areas; examples include the cyclooxygenase inhibitor aspirin to treat coronary-artery disease, the phosphodiesterase inhibitor sildenafil to treat erectile dysfunction, and the antibiotic erythromycin for impaired gastric motility (Supplementary Table 1)3. Even drugs associated with troubling side effects merit reconsideration, as evidenced by the successful repurposing of the antiemetic thalidomide to treat multiple myeloma4. Risk-mediating measures for avoiding the potential teratogenicity of thalidomide and its derivatives are reasonable in patients with life-threatening cancer, whereas the use of these drugs to treat nausea remains unacceptable. Although the benefits of repurposing are clear, successes thus far have been mostly serendipitous. Systematic, large-scale repurposing efforts have not been possible owing to the lack of a definitive physical drug collection, the low quality of drug annotations, and insufficient readouts of drug activity from which new indications can be predicted. Recent technological advances have enabled a step change in our ability to assess drug activities comprehensively. For example, perturbational gene expression profiles can now be obtained at high throughput across multiple cell types5. Gene expression profiling has enabled recent repurposing discoveries, including sirolimus for glucocorticoid-resistant acute lymphocytic leukemia, topiramate for inflammatory-bowel disease, and imipramine for small-cell lung cancer. For cancer therapeutics, a recently developed assay known as PRISM, which uses barcoded cell lines, enables rapid testing of many drugs against a large number of cancer cell lines in pools6. Molecular features of the cell lines (for example, gene expression, mutation, or copy-number variation) can then be used to identify predictive biomarkers of drug sensitivity (Supplementary Table 2). Finally, morphologic changes in cells can be assessed using high-throughput microscopy and machine-learning approaches. Such imaging-based screening unexpectedly identified the cholesterol drug lovastatin as a potent inhibitor of leukemia stem cells. To take advantage of these advances in experimental methods, we sought to assemble a comprehensive library of drugs that have reached the clinic. Surprisingly, we found that no such chemical library of approved and clinical trial drugs is available for purchase. In particular, drugs that have been tested in clinical trials but did not reach approval are not readily accessible. Even obtaining a complete list of such drugs and their annotations is challenging. A prior effort led by the US National Institutes of Health (NIH) focused on drugs approved by the US Food and Drug Administration (FDA), but the library has few compounds that have yet to achieve FDA approval7. Some chemical vendors offer a subset of approved drugs, but most of these commercial libraries overlap in their content and include only a small fraction of the approximately 10,000 drugs that have reached the clinic in the United States and Europe. Given that no complete collection exists, we launched a three-step effort to create the Repurposing Library by (i) identifying and purchasing compounds; (ii) comprehensively annotating their known activities and clinical indications; and (iii) experimentally confirming drug identity and purity. We employed two approaches to identify clinical-drug structures for the Repurposing Library. First, we searched existing databases, both publicly accessible and proprietary, for clinically tested drugs and then manually integrated them to ensure sufficient drug coverage and chemical-structure reliability (Supplementary Table 3). Sources included DrugBank, the NCATS NCGC Pharmaceutical Collection (NPC), Thomson Reuters Integrity, Thomson Reuters Cortellis, and Citeline Pharmaprojects7–9. Second, we located marketed or approved ingredient lists from regulatory agencies worldwide, including the FDA. After structure standardization and the removal of duplicates, approximately 10,000 small-molecule drugs with disclosed structures were found to have reached clinical development. Most of these drugs are not widely available in commercial screening libraries. Through structure-matching (as opposed to relying on compound names), chemical suppliers were identified for 5,691 compounds (Fig. 1). Controlled substances, nonpharmaceutical substances, and redundant elemental formulations were not pursued further. To assemble the collection, we ultimately purchased 8,584 samples (representing 5,691 unique compounds) from 75 chemical vendors, at an average cost of