CTD² Centers

Identifying and Targeting Cancer Dependencies With Small Molecules
Principal Investigator: Stuart L. Schreiber, Ph.D.

The Broad CTD² Center aims to accelerate the discovery of patient-matched therapeutics. Cancers, as a result of their specific genetic or cellular features, acquire dependencies required for survival. Some drugs have been developed to target these dependencies and can yield high clinical response rates. However, drugs in this category benefit only a small fraction of cancer patients and, due to drug resistance, the beneficial responses are not always durable. The Broad CTD² Center is discovering cancer dependencies that are targetable with small molecules and identifying combinations of drugs that can avoid or overcome resistance.

As part of the CTD² Network, the Broad Center generated an 'Informer Set' of small-molecule probes, drugs, and select combinations. These compounds selectively target distinct nodes in cell circuitry that collectively modulate a broad array of cell processes. The Center quantitatively measured the sensitivity of deeply characterized cancer-cell lines to the Informer Set compounds, and undertook analyses to connect the sensitivities to cancer-specific alterations, including mutations, gene expression, copy-number variation, and cellular features such as lineage. These analyses and links to the underlying data (e.g., raw or processed data, or fitted curves) are provided openly on the CTD² Data Portal and on an interactive public resource called the Cancer Therapeutics Response Portal (CTRP).

CTRP is a living resource for the biomedical research community, meaning the Center will continue to add data and analyses to updated versions of this interactive resource. The expanded dataset underlying the latest versio of CTRP includes an Informer Set of 545 compounds and select combinations tested for sensitivity in 907 cancer cell lines (specific analyses may be available only for relevant subsets of these totals). It can be mined to develop insights into small-molecule mechanisms of action and novel therapeutic hypotheses, and to support future discovery of drugs matched to patients based on predictive biomarkers. Future versions of the CTRP will include data generated under the following aims:

Aim 1. Integrate small-molecule, RNAi, and overexpression data from genomic cancer cell line profiling within the interactive resource (CTRP) to enable hypothesis generation

Aim 2. Identify and test hypotheses suggested by the interactive resource about cancer genetic dependencies targeted by small molecules

Aim 3. Use the interactive resource to identify combinations of small-molecule agents targeting cancer genetic dependencies that avoid or overcome drug resistance seen with single agents

Computational and Functional Approaches to Validating Cancer Genome Targets
Principal Investigator: Scott Powers, Ph.D.

The Cold Spring Harbor Laboratory (CSHL) CTD² Center examines the vast repertoire of alterations in cancer genomes to discover and functionally validate therapeutic targets. Using sophisticated bioinformatics and high-throughput biological tools, the Center predicts gene sets and networks that are significantly altered in tumors and then uses human and mouse models to annotate drivers, dependencies, and interactions that may be vulnerable to combinatorial targeting. Some tools and approaches used by the Center for identification and in-depth functional validation of targets include:

• Computational technology that uses drug response data to infer network models that predict cellular responses to perturbations
• High throughput cDNA screening to identify oncogenic drivers
• High throughput shRNA screening to discover tumor cell dependencies and combinatorial targets
•  Three dimensional “organoid” cell culture system that more accurately recapitulates in vivo tissue composition and architecture
•  Flexible and rapid “speedy mouse” technology to evaluate the function of single or multiple genes in parallel

To date, the Center has identified and validated over fifty driver genes and/or dependencies; several are compelling therapeutic targets, and two are being tested in the clinic. Fibroblast growth factor 19/fibroblast growth factor receptor 4 (FGF19/FGFR4) is being targeted in a clinical trial for liver cancer and bromodomain containing 4 (BRD4) is in a trial as a target specific to Acute Myeloid Leukemia. The experimental strategies of the CSHL Center provide a blue print that can be adapted by the research community to identify targets across different tumor types.

Systems Biology of Tumor Progression and Drug Resistance
Principal Investigator: Andrea Califano, Ph.D.

At Columbia University, CTD² funding supports efforts to study the systems biology of tumor progression and drug resistance. Researchers led by principal investigator Andrea Califano have developed a pipeline called Cancer Target High-Throughput Optimized Discovery and Evaluation (caTHODE). This pipeline uses both computational and experimental methods to efficiently discover and validate master regulators within the genomic networks that give rise to specific cancer subtypes. Master regulators are key nodes within networks of interacting genes and proteins that act as bottlenecks through which many different cellular signals must pass to initiate downstream activity. For this reason, researchers believe that master regulators may constitute points of vulnerability within a tumor. By computationally predicting and then experimentally validating the roles of master regulators in tumor progression and resistance to chemotherapy, this work is helping to generate a genome-wide list of prioritized targets for further investigation.

The caTHODE pipeline, which is intended to be scalable and effective for any tumor phenotype, utilizes a combination of methods developed at Columbia University. These include:

• Computational algorithms developed in the Califano laboratory that dissect and interrogate networks of transcriptional, post-transcriptional, and post-translational regulatory interactions.
• Genome-wide RNAi screens developed in the laboratory of José Silva (Icahn School of Medicine at Mount Sinai) to validate these master regulators.
• High-throughput chemical screening assays developed in the lab of Brent Stockwell to identify and validate small-molecule inhibitors of targets associated with phenotypes for tumor progression and drug resistance.

To date, the Columbia Center’s contributions to CTD² include the discovery and validation of therapeutic targets, chemical modulators, and biomarkers in three distinct tumor subtypes: glioblastoma multiforme; glucocorticoid resistant T cell acute lymphoblastic leukemia; and an aggressive subtype of diffuse large B cell lymphoma that originates from the progression of follicular lymphoma. In addition, Columbia researchers developed collaborations with other CTD² Centers focusing on additional cancer subtypes. These studies enabled the following findings:

Glioblastoma multiforme: In the mesenchymal phenotype of glioblastoma, the Columbia Center identified four modulators that harbor mutations. Mesenchymal glioblastoma is associated with the worst clinical outcomes for patients with brain cancers. In collaboration with Stuart Schreiber (Broad Institute), they also identified several novel inhibitors of signal transducer and activator of transcription 3 (STAT3), a key transcription factor within the regulatory network that promotes the mesenchymal phenotype.

T cell acute lymphoblastic leukemia:  Researchers at Columbia University identified several candidate master regulators of glucocorticoid (GC) resistance and validated three genes. When these genes were silenced, GC-induced apoptosis increased and GC transcriptional activity was activated. Biochemical and functional assays revealed a mechanism of glucocorticoid resistance, and high-throughput screening uncovered an experimental compound that restores GC sensitivity. In a follow-up study, they inferred and validated additional transcription factors that are master regulators of GC resistance.

Diffuse large B cell lymphoma (DLBCL): NF-κB pathway activation is a hallmark of the most aggressive form of DLBCL, the activated B cell DLBCL (ABC-DLBCL) subtype. The Columbia team identified transcription factors and signaling molecules that are critical to ABC-DLBCL and identified master regulators that contribute to follicular lymphoma transformation.

Ovarian serous cystadenocarcinoma: To identify molecular mechanisms of ovarian cancer pathogenesis, Columbia Center researchers reconstructed the transcriptional, post-transcriptional, and post-translational networks of ovarian serous cystadenocarcinoma. They identified master regulators that indicate poor prognosis, drive tumorigenesis, and promote resistance to cisplatin chemotherapy. Factors within regulatory networks that modulate the activity of the master regulators were also revealed.

Non-small cell lung cancer: To dissect the genome-wide signal transduction network that is regulated by tyrosine kinases, the Califano Lab applied a modified version of the Algorithm for the Reconstruction of Accurate Cellular Networks (ARACNe) algorithm (pARACNe) to a published dataset of phospho-proteomic profiles of non-small cell lung tumors. They also used a modified version of the Master Regulator Inference Algorithm (MARINa) algorithm to compare gene expression patterns and identify master regulators in 50 cell lines.

Functional Annotation of Cancer Genomes
Principal Investigator: William C. Hahn, M.D., Ph.D.

The comprehensive characterization of cancer genomes has and will continue to provide an increasingly complete catalog of genetic alterations in specific cancers. However, most epithelial cancers harbor hundreds of genetic alterations as a consequence of genomic instability. Therefore, the functional consequences of the majority of mutations remain unclear.

The Dana-Farber Cancer Institute (DFCI) CTD² Center has developed tools to perform somatic cell genetics in mammalian cells.  Tumor cell populations are systematically analyzed to identify the function of somatically altered genes identified by genome characterization studies. These approaches also reveal co-dependencies or synthetic lethal partners in tumors.

The Center is using high-throughput genetic perturbation approaches to create genome-wide datasets and applying bioinformatics to identify and credential targets. As part of this effort, the Center is cataloging all essential genes that contribute to proliferation or survival in a large number of ovarian and colon cancer cell lines. Novel bioinformatics approaches are being used to interrogate this functional data and integrate it with other genomic datasets. These approaches in vitro and in silico will identify targets (e.g., candidate oncogenes that are mutated or amplified and essential) in specific genetically defined tumor subtypes. Complementary to this approach, the Center is using multiplexed tumor formation assays and context-specific transformation assays to identify novel oncogenes. They have also developed methods to identify and validate targets in vivo in patient-derived experimental models. Beyond discovery, DFCI CTD² investigators will apply high-throughput gene expression profiling and targeted proteomics to interrogate the signaling networks perturbed by these oncogenes.

All of the outputs of these investigations (data, methodologies and bioinformatics tools) are made readily available through the CTD² Data Portal. Access to this information will enable the cancer research community to identify targets based on both genomic and functional biological evidence. Ultimately, these data will inform the most appropriate genetic context for downstream mechanistic and validation studies and prioritize targets for translation into therapeutics.

High Throughput Protein-Protein Interaction Interrogation in Cancer
Principal Investigator: Haian Fu, Ph.D.

Genomic alterations in various tumor types, as revealed by cancer genomics initiatives such as The Cancer Genome Atlas (TCGA), often lead to re-wired protein-protein interaction networks, which in turn drive tumor initiation and progression. Thus, identifying prominent PPI nodes and networks among oncoproteins and tumor suppressors as enriched by various genomics datasets and the validation of their critical roles in tumorigenesis and progression are expected to reveal an entirely new class of PPI-based cancer targets for therapeutic development.

The Emory Molecular Interaction Center for Functional Genomics (MicFG), with its expertise in high throughput technologies for the study of protein-protein interactions (PPI), productive track record of innovative HTS assay development for chemical lead discovery, and proven cancer genomics mining, database, and data integration capabilities, proposes to utilize high throughput PPI technologies to interrogate cancer genomes through a team science approach (i) to rapidly establish oncogenic PPI networks of selected cancer types based on TCGA and other genomic datasets, (ii) to validate functional roles of key PPI nodes or hubs in tumorigenesis and progression, (iii) to develop HTS assays for critical tumor-associated PPIs to enrich the therapeutic target pipeline of the NCI and the drug discovery field, and (iv) to leverage our informatics capability for genomics data mining for prioritized PPI mapping, functional PPI validation, and trans-network data sharing and collaboration.

We aim to bridge the gap between vast genomic datasets and therapeutic discovery by establishing and interrogating the cancer PPI target space. With our PPI expertise and state-of-the-art high throughput technologies, which are highly adaptable to a variety of outputs, we intend to function as an active, synergistic member in collaborative trans- Network projects. 

An Integrated Functional and Computational Discovery Engine for Preclinically-Validated Cancer Drug Targets
Principal Investigator: Christopher Kemp, Ph.D.

Large-scale molecular analyses have provided an unprecedented global view of the molecular defects in cancers and promise to revolutionize precision cancer medicine by guiding the development of therapies that are matched to genomic alterations in tumors. However, developing and implementing successful targeted therapies remains a daunting challenge. Cancer genomes are complex, so determining which genes to target from the hundreds of possibilities is challenging. Many of the best known oncogenes and tumor suppressor genes have not been successfully targeted directly. In cases where targeted therapies are used clinically, successes have been limited because patients frequently develop drug resistance, underscoring the need for combination therapies. The Fred Hutchison Cancer Research Center (FHCRC) CTD² Center and collaborators at Oregon Health and Science University have developed technology and methods to help address these challenges.

The FHCRC CTD² Center performs genome-scale functional testing of potentially druggable genes, including candidates derived from TCGA and other large genomic datasets, using isogenic cell lines and genomically characterized patient-derived tumor cell cultures. They use cell viability and other assays as readouts for high-throughput well-based siRNA and therapeutic compound screens. Integrating the results of these functional genomic approaches with genotype-specific vulnerabilities, which are inferred through computational analyses of large-scale public datasets, enables the identification of genotype- and patient- specific therapeutic targets that are selectively lethal to human cancer cells carrying defined mutations. 

Currently, the Center uses these strategies to identify and credential novel drug targets for cancers most in need of better therapies, including aggressive subtypes of head and neck squamous cell carcinoma, pancreatic ductal adenocarcinoma, and triple negative breast cancer. Gene targets that are identified and have known pharmacologic inhibitors are tested alone and in combination with existing standard of care drugs to nominate candidate targets for preclinical validation or clinical trials.

Identifying and Validating Tumor-Selective Targets for use in Immunotherapy
Principal Investigators: Martin McIntosh, Ph.D. and Edus H. Warren, M.D.

The CTD² Center at Fred Hutchison Cancer Research Center (FHCRC) aims to identify novel cancer-specific antigenic targets for immunotherapeutic approaches that are specifically toxic to patients’ malignant cells. The Center is identifying numerous promising cancer-specific targets in lung adenocarcinoma and serous ovarian carcinoma. Development of treatments to selectively attack cancer cells requires identification of cell surface proteins or protein complexes that have three essential elements: abundant on malignant cells, absent or rare in somatic tissues, and recognized specifically by a T cell or B cell receptor. The FHCRC Center has developed a high-throughput pipeline to efficiently identify cancer-specific targets that satisfy all three elements and are naturally suited for immunotherapeutic approaches because of their potential to be targeted without adversely affecting normal tissues.

FHCRC identifies potentially immunogenic peptides that arise from cancer-specific changes in a protein’s amino acid composition and are recognized by human T cell or B cell receptors. This recognition activates the body’s immune system to designate the tumor as foreign material and initiate a response to destroy it. Targets for T cell based immunotherapy must be on the cell surface, either as a peptide that is presented by major histocompatibility complex (MHC) class I protein complexes or as a cell surface protein variant. To discover potential candidate targets, the Center applies computational methods to RNA-seq data from The Cancer Genome Atlas and the Genotype Tissue Expression project. These data are combined with FHCRC/University of Washington Comprehensive Cancer Center data that profiles translating ribosomes. Variant coding transcripts that undergo a cancer-specific splicing event and are common in tumors, but are rare or not found in normal tissues, are designated cancer-specific transcripts (CSTs). CSTs that are bound to translating ribosomes or associated with the endoplasmic reticulum (found by another type of RNA-seq assay) are then used to predict the cancer-specific polypeptides (CSPs). These CSPs are experimentally tested to determine if they harbor peptides presented by the MHC complex or if their variant is presented on the cell surface. All discovery and verification efforts use materials from patients who participate in a FHCRC/University of Washington Comprehensive Cancer Center study. 

The CSPs are developed for use in adoptive T cell therapies. To start therapy development, the team identifies from patients or unaffected individuals a human T cell that recognizes the MHC-bound antigen, or a B cell that recognizes the cell surface proteins The T cell or B cell receptor sequence (TCR and BCR, respectively) is then obtained and used to engineer the therapeutic T cell. If necessary, the TCR may be modified to have higher affinity to the antigen. BCR sequences are used to engineer a T cell that adopts the recognition capability of an antibody. This is commonly called Chimeric Antigen Receptor T cell (or CART) therapy. Before they are tested as therapeutics in clinical trials, the active T cells are tested in vitro to ensure the modified receptors recognize the cancer-specific target and in vivo to determine if they elicit the desired immune response. When these methods are used in a patient, his or her T cells are harvested, engineered to express the specific TCR or BCR, and possibly other receptors, and then re-infused into the patient.

The FHCRC Center has active collaborations within and outside the CTD² Network. Other CTD² investigators exploit cancer-specific cell surface proteins identified by FHCRC to deliver toxic materials specifically to malignant cells. To validate targets in biomaterials collected from patients, the Center collaborates with leading clinical researchers who focus on moving therapies from mouse models into human trials.

Functional Analysis of Oncogenic Networks in Primary Organoids
Principal Investigators: Calvin J. Kuo, M.D., Ph.D. and Hanlee Ji, M.D.

Cancer arises from the acquisition and concerted action of multiple mutations and genomic aberrations in discrete combinations of tumor suppressors and oncogenes, known as "drivers". Large cancer genome-scale sequencing studies such as TCGA, are now operative and the Cancer Target Discovery and Development (CTD²) Network seeks "to bridge the gap between the enormous volumes of data generated... and the ability to use these data for the development of human cancer therapeutics". A secondary goal for the CTD² Initiative is that in five years, "the entire CTD² Network is expected to identify and characterize targets for approximately 25 or more (if possible) cancer types," and for applicants "to have or build the capacity for in depth analyses and experimental approaches utilizing datasets for many cancer types." A broad "coverage" is the paradigm for this initiative.

The wealth of TCGA data will be directly coupled to robust in vitro functional validation of candidate cancer driver modules using primary mouse 3D organoid cultures of diverse tissues arrayed in high-throughput format. In Aim 1, the Hanlee Ji and Sylvia Plevritis groups will identify co-segregating mutational modules from TCGA datasets from multiple solid tumor types, using complementary methods of supervised Bayesian analysis and Unsupervised Module Network Analysis for Master Regulators. In Aim 2, these prioritized mutational modules, stratified for clinical significance, will undergo direct functional validation in a broadly applicable, multiplexable, in vitro 3D primary organoid system developed by the Calvin Kuo group, which is amenable to combinatorial gene engineering. In collaboration with Bill Hahn, this will utilize high throughput lentiviral introduction of cDNA or shRNA to systematically interrogate the genes within amplicons and deletions, contextually modeled in the TCGA mutational background in which these copy number variations occur. Additionally, co-segregating mutational modules from diverse tissues will undergo systematic deletion in organoid cultures to define minimal module composition, and we will pursue process development to extend the range of tissues from which organoids can be modeled.

Overall, these studies describe bioinformatic and in vitro modeling approaches that are robustly portable across a variety of organ systems for functional interrogation of diverse TCGA datasets and with attendant implications for cancer biology, diagnosis and therapy.

Systematic Development of Novel Druggable Targets in Glioblastoma
Principal Investigator: Michael E. Berens, Ph.D.

Gene expression patterns across glioblastoma (GBM) cases in The Cancer Genome Atlas (TCGA) were previously queried by a novel computational analysis tool for mining molecular “contexts^”. The analysis uncovered 12 distinct gene-based molecular “contexts of GBM”. Based on molecular profiles from previously established and clinically relevant preclinical models, the Translational Genomics Research Institute (TGen) CTD² Center assembled a collection of 54 patient-derived xenograft GBM models that reside within the same genomic-based contexts. This collection represents diverse GBM tumor phenotypes. Collaborators at TGen, Sanford Burnham, and Thomson-Reuters use these PDX models, in combination with bioinformatic and empiric approaches, to achieve the following goals:

• Identify novel targets associated with subclasses of GBM and discrete features of the “hallmarks of cancer”
• Functionally validate prioritized targets in vitro using high throughput RNAi and chemical based assays
• Validate prioritized targets in vivo using orthotopic human GBM tumor grafts and inducible shRNA and/or hits from chemical based screening assays

Within each distinct context, bioinformatic algorithms (MetaBase^TM) query the patterns of expressed genes and the pathways in which they function. The known networks in which these genes reside are used to discern candidate therapeutic targets specific for certain GBM contexts. Chemical biology libraries and RNAi panels (well characterized for mechanism of action) uncover molecules that specifically and uniquely influence GBM cell behaviors within certain contexts. Compounds identified in empiric studies are also used in high throughput screening against short-term cultures from the GBM PDX models. Each of these strategies can be adapted to assess various hallmarks of cancer, including proliferation, survival, self-renewal, migration, invasion, and apoptosis-resistance.

Integrating in silico and laboratory technologies continuously improves computational approaches to help prioritize targets in the specific tumors. Furthermore, small-molecule screens against identified targets and pathways provide a potential fast-track to develop novel “perturbagens” against the validated targets. Sub-classifying GBM during these experimental processes informs fundamental insight into disease biology and lays the foundation for precision-guided therapy for molecularly-subgrouped human cancers.

By demonstrating an efficient approach for tractable target identification and validation, this project will impact the fields of informatics, cancer biology and drug discovery and accelerate, the translation of genomic discoveries into new treatments.

Transformative Strategies for Dissecting Cancer Pathways
Principal Investigators: Frank P. McCormick, Ph.D., Michael T. McManus, Ph.D. and Jonathan S. Weissman, Ph.D.

The ability to effectively and efficiently perturb endogenous gene expression is essential to uncover tumor-specific vulnerabilities that may be targeted with therapeutics. However, targeting single vulnerabilities in patients often leads to drug resistance and treatment failure. Therefore, the signaling pathways that act synergistically to promote tumor growth must be identified to design effective combination cancer therapies (polytherapies) that target key cancer "driver" pathways. The lack of a systematic method by which to identify tumor-specific vulnerabilities in pathways that functionally cooperate to drive tumor growth creates a major challenge in developing such therapies. Therefore, the search for effective cancer polytherapies has been done largely in an ad hoc manner by exploring limited numbers of potential combinations. The CTD² Center at the University of California San Francisco (UCSF-1) has developed an experimental pipeline with high-throughput technologies to systematically identify pathways that when targeted, lead to specific and synergistic destruction of cancer cells.

The UCSF-1 Center uses the “EXPAND RNAi” high-throughput screening approach based on complex RNAi libraries to identify candidate driver genes. To evaluate the relevance of the potential drivers, the Center quantifies responses to targeted inhibitors using screens in engineered primary cell lines (isogenic cell lines). The quantitative data are integrated with known information for each compound and cell line used in the screen to create chemical-genetic interaction maps. The maps provide critical insights into biological pathways and functional dependencies and can be used to identify and design potential polytherapies.

The UCSF-1 Center also employs a complementary approach to study gene function in cancer cells, the clustered regularly interspaced short palindromic repeats (CRISPR) technique. They modified the CRISPR system to dynamically repress gene transcripts (CRISPRi) as well as established a novel system that results in gene activation (CRISPRa). These methods exhibit low off-target effects and specifically target transcriptional start sites allowing the Center to apply the technology on a whole-genome scale.

Novel approaches developed by the UCSF-1 Center establish systematic methods to uncover gene interaction networks that drive tumor growth. The Center’s experimental pipelines will facilitate the development of cancer polytherapies and create new paradigms for discovering cancer therapeutics that fully capitalize on the genomic profiling of human tumors.

Genetic Network Analysis to Discover Cancer Targets
Principal Investigator: William A. Weiss, M.D., Ph.D.

Cancers acquire genetic changes, or drivers, that are required for tumor survival. Therapies developed to target these cancer drivers often lead to clinical responses. However, the promise of targeted therapies has been tempered in many cancer patients by eventual relapse due to therapy-driven drug resistance that results from rewiring of the tumor’s genetic architecture. Novel approaches are necessary to understand this rewiring, and identify upstream or downstream targets that may yield to small molecule modulation or other therapeutic strategies.

The CTD² Center at University of California San Francisco (UCSF-2) has developed computational network tools to visualize the genetic architecture of tissue samples. These tools correlate gene expression patterns from heterogeneous samples of normal or tumor tissue. This approach identifies genetic networks that control tissue organization or cellular function in complex normal tissues, and details the rewiring of these networks during tumor evolution in vivo.

These methods were initially developed to analyze normal and tumor tissues from mouse skin. Transcriptional profiles of normal skin samples from wild type mice and mice lacking the tumor oncogene, Hras, were compared. This analysis provided insight on changes in genetic architecture that may lead to susceptibility to tumor development in tissues lacking Hras. Recently, these analyses have been extended to neural tissues from mouse models of neuroblastoma. If appropriate mouse models are available to generate expression profiling data on tissues of interest, the computational approach has the potential to identify the genetic architecture of tumor types being studied by other CTD² Centers through trans-Network collaborations.

The UCSF-2 CTD² Center is currently identifying candidate drug targets by mining genomic data from human neuroblastomas and carrying out functional shRNA screens in human models. Results from human data mining are compared with architecture analyses results from the neuroblastoma mouse model to prioritize targets for further investigation. Prioritized targets are analyzed computationally using gene expression and copy number data from multiple developmental stages of carcinoma including benign, malignant, and metastatic samples. This analysis will investigate stage-specific activation/inhibition of candidates and identify changes in their expression architecture during tumor progression. Targets will then be validated using gain- or loss- of function approaches in human derived cell lines and patient derived xenograft models from appropriate human cancers. Potential new targets will be tested for sensitivity to pharmaceuticals. These approaches will identify both new cancer targets and novel roles for existing therapeutics within cancer pathways.

Biological Annotation of Data from Large-Scale Cancer Genomic Initiatives
Principal Investigator: Gordon B. Mills, M.D., Ph.D.

Large-scale tumor profiling efforts by consortia such as The Cancer Genome Atlas (TCGA), Therapeutically Applicable Research to Generate Effective Treatments (TARGET) and the Cancer Genome Characterization Initiative (CGCI) are cataloging genomic aberrations across major cancer lineages. These studies have revealed an extraordinary level of genome complexity. The CTD² Center based at the University of Texas MD Anderson Cancer Center (MDACC) is working to distinguish key “driver” events critical to pathogenesis from the numerous biologically-neutral “passenger” aberrations that accompany unstable tumor genomes. The Center’s goal is to find ways to validate functional driver aberrations, since targeting such events or their activated pathways may improve patient outcomes.

Currently, the CTD² Center at MDACC examines the functionality of thousands of potential driver genes found within breast, pancreas, melanoma and endometrial tumors. MDACC, in collaboration with Baylor College of Medicine, has developed pipelines for the rapid construction of barcoded cDNAs with mutations identified through tumor profiling by the large-scale genomics consortia mentioned above. The Center is actively constructing every somatic event within candidate cancer genes because each change within a given gene may result in a different functional impact or therapeutic response. The cDNAs are suitable for expression in cell and animal cancer models.

To maximize discovery potential, cell lines expressing wild type or mutant cDNA can be used in multiple high-throughput screens in parallel. For example, generalizable sensor cell assays quantify the ability of the mutant genes to induce cell survival and proliferation. Through these assays, driver genes are identified. Cells expressing confirmed driver genes are subjected to compound screening platforms designed to identify drugs that inhibit driver activity. Together, these in vitro approaches permit broad evaluation of driver candidates across multiple cancer lineages. Importantly, these approaches complement RNAi-based gene depletion screening strategies for driver discovery used by other CTD² Centers.

While cell-based screening systems are amenable to high-throughput analyses, in vitro models do not fully recapitulate all hallmarks of tumorigenesis and metastasis. To address this challenge, the Center has developed a Context-Specific Screen (CSS) platform to interrogate the tumorigenic potency of candidate driver genes under the appropriate in vivo genetic and microenvironment contexts. For in vivo screens, non-transformed human primary cells expressing pools of confirmed barcoded driver genes are implanted orthotopically at defined sites in the mouse to ensure the correct microenvironment context. Upon tumor formation, genes driving tumor progression are identified from tissues by barcode amplification and sequencing. Importantly, this approach permits high-throughput discovery of cooperating driver events that are co-selected in output tumors.

In summary, this work will provide the greater research community multi-level functional assessments of oncogenomics data collected by TCGA and other large scale genomic studies. This technology, which facilitates biological annotations of genomic changes will create unique opportunities for transformative cancer research by engaging the research community. Ultimately, these contributions will accelerate drug development and implementation.

A Concerted Attack on Patient-Specific Oncogenic Vulnerabilities in Lung Cancer
Principal Investigators: Michael Roth, Ph.D., Michael A. White, Ph.D., John B. MacMillan, Ph.D., and John D. Minna, M.D.

Lung cancer is a major cause of death in the United States and there is a clear need for additional and better therapeutic approaches. The multiple oncogenotypes that cause lung cancer are known to confer different responses to currently approved therapeutics. However, the molecular basis of these differences is not completely understood and a systematic approach to identify functional differences in cells derived from tumors with different oncogenotypes has not been carried out. The CTD² Center at University of Texas Southwestern (UTSW) Medical Center uses two different high throughput approaches to address this need.

Using siRNA libraries, the Center studies how loss of function of each human gene affects survival of lung cancer cell lines that represent at least 6 different oncogenic subtypes. This high throughput technique measures cancer cell survival. Sensitivity to loss of a gene indicates that the gene (or another protein in its pathway) may be a potential therapeutic target. The targets identified are then tested for therapeutic potential in appropriate tumor models in animals. 

The UTSW Center is also conducting high-throughput screens to identify compounds that may have therapeutic benefit in one or more subsets of genetically distinct lung cancers. In a pilot effort, the Center used a library of 200,000 synthetic drug-like compounds to screen for inhibition of cell survival in the same lung cancer cell lines used in the siRNA screens. The Center is now screening 40,000 natural products derived from marine bacteria. Compounds that inhibit cell survival in culture are tested for the ability to inhibit growth of human tumor explants. These experiments provide a potential fast-track for discovering novel therapies as well as a method for detecting tumor-specific vulnerabilities that are not detected by the RNAi strategy described above.

After identifying tumor specific vulnerabilities through the two independent screening methods, gene knock-down data and compound sensitivity data can be correlated to identify novel drug targets and compounds that inhibit them. A database of the combined functional genomic and compound screening results are available to the public through the CTD² Data Portal and compound structures annotated with biological data are displayed on PubChem.