The Cancer Target Discovery and Development initiative (CTD 2) is a multi-institutional, collaborative program of the National Cancer Institute whose goal is to leverage the huge amount of data that is now available about cancer to improve treatments for people with disease. This project is now possible by using large data sets such as The Cancer Genome Atlas (TCGA), Therapeutically Applicable Research to Generate Effective Treatments (TARGET), and the Cancer Genomic Characterization Initiative (CGCI), which contain pairs of molecular profiles of both cancer samples and normal tissues from which those samples developed. Using experimental techniques, high-throughput approaches, and bioinformatics methods, participating institutes are analyzing this information to define the molecular networks that underlie specific cancer subtypes, identify genes and pathways within those networks that hold potential as targets for therapies, and find small molecules able to perturb these targets in ways that can be used to prevent or treat disease.

Master regulators may constitute points of vulnerability within a tumor.

At Columbia University, CTD2 funding is supporting a project to study the systems biology of tumor progression and drug resistance. We have developed a pipeline called Cancer Target High-Throughput Optimized Discovery and Evaluation (caTHODE), which 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 function as bottlenecks through which many different cellular functions must pass. For this reason, researchers believe that master regulators may constitute points of vulnerability within a tumor. By predicting and then validating the roles of master regulators of tumor progression and resistance to chemotherapy, this work is helping to generate a genome-wide, prioritized list of targets for further investigation.

caTHODE: a Pipeline for Cancer Target Discovery

Our research pipeline, called caTHODE, utilizes a combination of methods developed at Columbia University to predict and then validate master regulators of disease. The approach involves:

  • using computational algorithms developed in principal investigator Andrea Califano’s laboratory to dissect transcriptional (ARACNe), post-transcriptional (Cupid, Hermes), and post-translational (MINDy) regulatory interactions; and to interrogate these networks, also called interactomes, to identify master regulators of disease (MARINa).
  • experimentally validating a prioritized list of highly probable master regulators using genome-wide RNAi screens developed in the laboratory of José Silva
  • performing high-throughput chemical screening assays developed in the lab of Brent Stockwell to identify and validate small-molecule inhibitors of molecular targets associated with phenotypes for tumor progression and drug resistance.

The caTHODE pipeline is intended to be scalable and to allow for processing of a novel tumor phenotype every 18 to 24 months. It is making it possible to validate individual targets and synergistic molecular relationships that constitute either oncogene or non-oncogene dependencies of the tumor, or that increase sensitivity to existing FDA-approved drugs or compounds in the late stages of development.

Our projects

The original goal of Columbia’s contribution to CTD 2 was to discover and validate therapeutic targets, chemical modulators, and biomarkers in three distinct tumor subtypes: glioblastoma multiforme (GBM); glucocorticoid resistance in T cell acute lymphoblastic leukemia (T-ALL); and an aggressive subtype of diffuse large B cell lymphoma (DLBCL) that originates from the progression of follicular lymphoma. In addition, Columbia researchers developed collaborations focusing on additional cancer subtypes, including ovarian serous cystadenocarcinoma, non-small cell lung cancer, and a second aggressive subtype of DLBCL, the activated B cell subtype (ABC).

Glioblastoma multiforme (GBM): The mesenchymal phenotype of glioblastoma is associated with the worst clinical outcomes for patients with brain cancers. It is driven by oncogene addiction in a small regulatory module that includes the transcription factors C/EBPβ, C/EBPδ, and STAT3. Using a genomic analysis that integrates gene expression and copy number variation profiling, deep sequencing data, and reverse-engineering to define the interactomes of molecular subtypes of GBM, we identified four modulators that harbor specific  mutations in mesenchymal phenotype samples. In a collaboration with Stuart Schreiber (Broad Institute), we also screened compounds previously shown to bind STAT3, and found a number that decrease STAT3 activity. We identified several additional STAT3 inhibitors using a luciferase-based screen for compounds that inhibit the transcriptional activity of C/EBPβ,.

T cell acute lymphoblastic leukemia (T-ALL): Using the MARINa algorithm, we identified several candidate master regulators of glucocorticoid (GC) resistance. RNAi-mediated validation of the top candidates identified three genes whose silencing increased GC-induced apoptosis and activated GC transcriptional activity. Biochemical and functional assays revealed a mechanism of glucocorticoid resistance, and high-throughput screening revealed that an experimental compound can restore GC sensitivity. This activity was confirmed both in resistant T-ALL cell lines and in primary patient transplants into irradiated mice. In a follow-up study that integrated a pooled RNAi screen with computational predictions of master regulators, we inferred and validated additional candidate transcription factors that are master regulators of GC resistance.

Diffuse large B cell lymphoma (DLBCL): NF-kB pathway activation is a hallmark of the most aggressive DLBCL, the activated B cell DLBCL  (ABC-DLBCL) subtype. We attempted to identify therapeutic targets in ABC-DLBCL, comparing cells from this lineage with cells of the germinal center B-cell like DLBCL (GCB). By integrating RNAi screened candidates with genomics-inferred drivers of ABC and GCB, we identified transcription factors and signaling molecules that are critical to ABC-DLBCL. We also used a previously identified gene expression profile signature for follicular lymphoma (FL) to identify master regulators of FL transformation using the MARINa algorithm and the Human B Cell Interactome.

Ovarian serous cystadenocarcinoma: In an effort to identify molecular mechanisms of ovarian cancer pathogenesis, we used data from TCGA, as well as the MARINa and MINDy algorithms, respectively, to: (1) reconstruct the transcriptional, post-transcriptional, and post-translational networks of ovarian serous cystadenocarcinoma, and (2) identify master regulators that drive tumorigenesis, poor prognosis, and resistance to cisplatin chemotherapy, as well as factors within regulatory networks that modulate their activity.

Non-small cell lung cancer: We applied a modified version of the ARACNe algorithm (pARACNe) to a published dataset of phospho-proteomic profiles of non-small cell lung tumors in order to dissect the genome-wide signal transduction network that is regulated by tyrosine kinases. We also used a modified version of the MARINa algorithm to perform master regulator analysis in 50 cell lines.