Francesco Iorio

Systematic identification of genomic markers of drug sensitivity in cancer cells

Clinical responses to anticancer therapies are often restricted to a subset of patients. In some cases, mutated cancer genes are potent biomarkers for responses to targeted agents. Here, to uncover new biomarkers of sensitivity and resistance to cancer therapeutics, we screened a panel of several hundred cancer cell lines—which represent much of the tissue-type and genetic diversity of human cancers—with 130 drugs under clinical and preclinical investigation. In aggregate, we found that mutated cancer genes were associated with cellular response to most currently available cancer drugs. Classic oncogene addiction paradigms were modified by additional tissue-specific or expression biomarkers, and some frequently mutated genes were associated with sensitivity to a broad range of therapeutic agents. Unexpected relationships were revealed, including the marked sensitivity of Ewing’s sarcoma cells harbouring the EWS (also known as EWSR1)-FLI1 gene translocation to poly(ADP-ribose) polymerase (PARP) inhibitors. By linking drug activity to the functional complexity of cancer genomes, systematic pharmacogenomic profiling in cancer cell lines provides a powerful biomarker discovery platform to guide rational cancer therapeutic strategies.

Discovery of drug mode of action and drug repositioning from transcriptional responses

A bottleneck in drug discovery is the identification of the molecular targets of a compound (mode of action, MoA) and of its off-target effects. Previous approaches to elucidate drug MoA include analysis of chemical structures, transcriptional responses following treatment, and text mining. Methods based on transcriptional responses require the least amount of information and can be quickly applied to new compounds. Available methods are inefficient and are not able to support network pharmacology. We developed an automatic and robust approach that exploits similarity in gene expression profiles following drug treatment, across multiple cell lines and dosages, to predict similarities in drug effect and MoA. We constructed a “drug network” of 1,302 nodes (drugs) and 41,047 edges (indicating similarities between pair of drugs). We applied network theory, partitioning drugs into groups of densely interconnected nodes (i.e., communities). These communities are significantly enriched for compounds with similar MoA, or acting on the same pathway, and can be used to identify the compound-targeted biological pathways. New compounds can be integrated into the network to predict their therapeutic and off-target effects. Using this network, we correctly predicted the MoA for nine anticancer compounds, and we were able to discover an unreported effect for a well-known drug. We verified an unexpected similarity between cyclin-dependent kinase 2 inhibitors and Topoisomerase inhibitors. We discovered that Fasudil (a Rho-kinase inhibitor) might be “repositioned” as an enhancer of cellular autophagy, potentially applicable to several neurodegenerative disorders. Our approach was implemented in a tool (Mode of Action by NeTwoRk Analysis, MANTRA, http://mantra.tigem.it).

Heterogeneity of genomic evolution and mutational profiles in multiple myeloma

Multiple myeloma is an incurable plasma cell malignancy with a complex and incompletely understood molecular pathogenesis. Here we use whole-exome sequencing, copy-number profiling and cytogenetics to analyse 84 myeloma samples. Most cases have a complex subclonal structure and show clusters of subclonal variants, including subclonal driver mutations. Serial sampling reveals diverse patterns of clonal evolution, including linear evolution, differential clonal response and branching evolution. Diverse processes contribute to the mutational repertoire, including kataegis and somatic hypermutation, and their relative contribution changes over time. We find heterogeneity of mutational spectrum across samples, with few recurrent genes. We identify new candidate genes, including truncations of SP140, LTB, ROBO1 and clustered missense mutations in EGR1. The myeloma genome is heterogeneous across the cohort, and exhibits diversity in clonal admixture and in dynamics of evolution, which may impact prognostic stratification, therapeutic approaches and assessment of disease response to treatment.

A Landscape of Pharmacogenomic Interactions in Cancer

Summary Systematic studies of cancer genomes have provided unprecedented insights into the molecular nature of cancer. Using this information to guide the development and application of therapies in the clinic is challenging. Here, we report how cancer-driven alterations identified in 11,289 tumors from 29 tissues (integrating somatic mutations, copy number alterations, DNA methylation, and gene expression) can be mapped onto 1,001 molecularly annotated human cancer cell lines and correlated with sensitivity to 265 drugs. We find that cell lines faithfully recapitulate oncogenic alterations identified in tumors, find that many of these associate with drug sensitivity/resistance, and highlight the importance of tissue lineage in mediating drug response. Logic-based modeling uncovers combinations of alterations that sensitize to drugs, while machine learning demonstrates the relative importance of different data types in predicting drug response. Our analysis and datasets are rich resources to link genotypes with cellular phenotypes and to identify therapeutic options for selected cancer sub-populations.

Machine Learning Prediction of Cancer Cell Sensitivity to Drugs Based on Genomic and Chemical Properties

Predicting the response of a specific cancer to a therapy is a major goal in modern oncology that should ultimately lead to a personalised treatment. High-throughput screenings of potentially active compounds against a panel of genomically heterogeneous cancer cell lines have unveiled multiple relationships between genomic alterations and drug responses. Various computational approaches have been proposed to predict sensitivity based on genomic features, while others have used the chemical properties of the drugs to ascertain their effect. In an effort to integrate these complementary approaches, we developed machine learning models to predict the response of cancer cell lines to drug treatment, quantified through IC50 values, based on both the genomic features of the cell lines and the chemical properties of the considered drugs. Models predicted IC50 values in a 8-fold cross-validation and an independent blind test with coefficient of determination R2 of 0.72 and 0.64 respectively. Furthermore, models were able to predict with comparable accuracy (R2 of 0.61) IC50s of cell lines from a tissue not used in the training stage. Our in silico models can be used to optimise the experimental design of drug-cell screenings by estimating a large proportion of missing IC50 values rather than experimentally measuring them. The implications of our results go beyond virtual drug screening design: potentially thousands of drugs could be probed in silico to systematically test their potential efficacy as anti-tumour agents based on their structure, thus providing a computational framework to identify new drug repositioning opportunities as well as ultimately be useful for personalized medicine by linking the genomic traits of patients to drug sensitivity.

Transcriptional data: a new gateway to drug repositioning?

Highlights ► Can every biological state be represented by a given gene expression signature? ► Recently, signatures have been used as proxies of clinicopathological phenotypes. ► Drug–drug/drug–disease ‘connections’ have been inferred by signature matching. ► This allowed the prediction of new application for already approved drugs. ► A massive amount of public transcriptional data is ready to be exploited in this way.

Transcriptional gene network inference from a massive dataset elucidates transcriptome organization and gene function

We collected a massive and heterogeneous dataset of 20 255 gene expression profiles (GEPs) from a variety of human samples and experimental conditions, as well as 8895 GEPs from mouse samples. We developed a mutual information (MI) reverse-engineering approach to quantify the extent to which the mRNA levels of two genes are related to each other across the dataset. The resulting networks consist of 4 817 629 connections among 20 255 transcripts in human and 14 461 095 connections among 45 101 transcripts in mouse, with a inter-species conservation of 12%. The inferred connections were compared against known interactions to assess their biological significance. We experimentally validated a subset of not previously described protein–protein interactions. We discovered co-expressed modules within the networks, consisting of genes strongly connected to each other, which carry out specific biological functions, and tend to be in physical proximity at the chromatin level in the nucleus. We show that the network can be used to predict the biological function and subcellular localization of a protein, and to elucidate the function of a disease gene. We experimentally verified that granulin precursor (GRN) gene, whose mutations cause frontotemporal lobar degeneration, is involved in lysosome function. We have developed an online tool to explore the human and mouse gene networks.

A CRISPR Dropout Screen Identifies Genetic Vulnerabilities and Therapeutic Targets in Acute Myeloid Leukemia

Summary Acute myeloid leukemia (AML) is an aggressive cancer with a poor prognosis, for which mainstream treatments have not changed for decades. To identify additional therapeutic targets in AML, we optimize a genome-wide clustered regularly interspaced short palindromic repeats (CRISPR) screening platform and use it to identify genetic vulnerabilities in AML cells. We identify 492 AML-specific cell-essential genes, including several established therapeutic targets such as DOT1L, BCL2, and MEN1, and many other genes including clinically actionable candidates. We validate selected genes using genetic and pharmacological inhibition, and chose KAT2A as a candidate for downstream study. KAT2A inhibition demonstrated anti-AML activity by inducing myeloid differentiation and apoptosis, and suppressed the growth of primary human AMLs of diverse genotypes while sparing normal hemopoietic stem-progenitor cells. Our results propose that KAT2A inhibition should be investigated as a therapeutic strategy in AML and provide a large number of genetic vulnerabilities of this leukemia that can be pursued in downstream studies.

BRAF inhibitor resistance mediated by the AKT pathway in an oncogenic BRAF mouse melanoma model.

Significance Using Sleeping Beauty transposon mutagenesis in a melanoma model driven by oncogenic BRAF (B-Raf proto-oncogene, serine/threonine kinase), we identified both known and novel candidate genes that mediate resistance to the BRAF inhibitor PLX4720. We validate ES-cell expressed Ras as a novel promoter of BRAF inhibitor resistance and propose that AKT (v-akt murine thymoma viral oncogene homolog 1)-mediated inactivation of BAD (BCL2-associated agonist of cell death) constitutes a pathway that may contribute to hepatocyte growth factor-mediated therapy resistance. Our work establishes Sleeping Beauty mutagenesis as a powerful tool for the identification of novel resistance genes and mechanisms in genetically modified mouse models. BRAF (v-raf murine sarcoma viral oncogene homolog B) inhibitors elicit a transient anti-tumor response in ∼80% of BRAFV600-mutant melanoma patients that almost uniformly precedes the emergence of resistance. Here we used a mouse model of melanoma in which melanocyte-specific expression of BrafV618E (analogous to the human BRAFV600E mutation) led to the development of skin hyperpigmentation and nevi, as well as melanoma formation with incomplete penetrance. Sleeping Beauty insertional mutagenesis in this model led to accelerated and fully penetrant melanomagenesis and synchronous tumor formation. Treatment of BrafV618E transposon mice with the BRAF inhibitor PLX4720 resulted in tumor regression followed by relapse. Analysis of transposon insertions identified eight genes including Braf, Mitf, and ERas (ES-cell expressed Ras) as candidate resistance genes. Expression of ERAS in human melanoma cell lines conferred resistance to PLX4720 and induced hyperphosphorylation of AKT (v-akt murine thymoma viral oncogene homolog 1), a phenotype reverted by combinatorial treatment with PLX4720 and the AKT inhibitor MK2206. We show that ERAS expression elicits a prosurvival signal associated with phosphorylation/inactivation of BAD, and that the resistance of hepatocyte growth factor-treated human melanoma cells to PLX4720 can be reverted by treatment with the BAD-like BH3 mimetic ABT-737. Thus, we define a role for the AKT/BAD pathway in resistance to BRAF inhibition and illustrate an in vivo approach for finding drug resistance genes.

Identifying Network of Drug Mode of Action by Gene Expression Profiling

Drug mode of action (MOA) of novel compounds has been predicted using phenotypic features or, more recently, comparing side effect similarities. Attempts to use gene expression data in mammalian systems have so far met limited success. Here, we built a drug similarity network starting from a public reference dataset containing genome-wide gene expression profiles (GEPs) following treatments with more than a thousand compounds. In this network, drugs sharing a subset of molecular targets are connected by an edge or lie in the same community. Our approach is based on a novel similarity distance between two compounds. The distance is computed by combining GEPs via an original rank-aggregation method, followed by a gene set enrichment analysis (GSEA) to compute similarity between pair of drugs. The network is obtained by considering each compound as a node, and adding an edge between two compounds if their similarity distance is below a given significance threshold. We show that, despite the complexity and the variety of the experimental conditions, our approach is able to identify similarities in drug mode of action from GEPs. Our approach can also be used for the identification of the MOA of new compounds.