Broad Institute/Dana-Farber Integrative Cancer Biology Program
Cancer biology is in the midst of major transition — from a primary focus on isolated components (e.g., oncogenes) to an integrated focus on systems (e.g., pathways). This transition has become feasible only recently, following two decades of intensive work on the molecular biology of cancer (leading to the identification of many oncogenes) and on the human genome (providing the ability to take comprehensive global views of biological systems).
A "systems biology" approach is clearly essential to the understanding and treatment of cancer. We need an integrated view expressed in terms of pathways — consisting of comprehensive knowledge of those pathways that are active and those that are essential in each cancer. Such a systems-level understanding of cancer will have major implications for (i) understanding the development and progression of cancer and (ii) creating therapeutic agents and combinations and matching them to patients.
Recognizing the need for integrated systems approaches to cancer, the National Cancer Institute’s Integrated Cancer Biology Program (ICBP) has created a network of Centers for Systems Cancer Biology (CCSB). This network aims to work collaboratively to provide a systems-level understanding of cancer via the generation of multidimensional data, sophisticated yet accessible computational tools, experimental techniques and interdisciplinary training for the next generation of cancer systems biologists.
Phase I (2004-2009)
In 2004, we launched the Broad Institute/Dana-Farber Cancer Institute Center for Integrative Cancer Biology, funded as part of the NCI’s ICBP network. This center created a community of scientists working at the interface of computational cancer biology and experimental cancer biology. The focus of the Center from 2004-2009 has been on establishing patterns of kinase essentiality in cancer, with the notion that what was developed and learned through that kinase-focused effort could later expand to a broader approach to cancer systems biology. The Center was successful beyond all expectations — as measured by productivity (>400 publications by Center-funded investigators during the funding period, including 49 multidisciplinary, collaborative papers between Center-funded PIs); by the development of new systems biology-oriented laboratory methods (e.g., RNAi screening and phospho-proteomic profiling); by computational method development (e.g., Gene Set Enrichment Analysis, Non-Negative Matrix Factorization, and the Connectivity Map project), and by direct uptake of Center methods by the scientific community (e.g., >8000 users of the Connectivity Map, >12,000 users of GSEA software, and >10,000 users of GenePattern software).
Phase II (2010-2015)
In 2010, the Broad Institute ICBP/Center for Cancer Systems Biology (CCSB) received sustaining funding as part of the second round of NCI ICBP funding. We are now expanding the scope of our previous studies to genome-scale and have assembled an integrated team with the following scientific focus:
The Broad CCSB has the overall goal of developing computational models that predict genome-wide essentiality based on the molecular characteristics of the tumor. A critical goal in cancer research is to accurately predict essential genes/proteins across a diversity of tumor subtypes. Such a capability would allow for i) the elucidation of the molecular targets against which therapeutics should be developed, and ii) the identification of specific patient populations likely to respond to such targeted interventions. Accomplishing this goal will require developing a deep understanding of the cellular circuitry of tumor cells — the details of which are increasingly recognized to depend on the genetic subtype of the tumor. A single set of network dependencies is unlikely to explain the diversity of even a single tumor type (e.g., whereas EGFR mutation in lung adenocarcinoma signals EGFR essentiality in some patients, those also harboring KRAS mutation are EGFR-independent. Similarly, EGFR mutations do not appear to confer EGFR essentiality in glioblastoma, where more complex oncogenic and feedback mechanisms appear to be at play). A way to tie functional dependencies to genetic diversity is therefore needed.
A number of technology developments over the past several years are now making it possible to seriously take on this goal. For example, the ability to perform genome-wide RNAi screens now makes it possible to systematically perturb the cancer genome, thereby identifying those genes that are essential in a given experimental system. In addition, it is becoming increasingly possible to perform extensive molecular characterization of tumor samples and model systems — including the patterns of gene copy number, gene expression, somatic mutation, tyrosine phosphorylation, etc., which might serve as predictive features of essentiality.
Major obstacles currently prevent rapid progress toward the goal of predicting essentiality: First, there is a lack of data of sufficient scope (i.e., genome-wide RNAi screening data and extensive molecular characterization) and scale (i.e., data spanning large numbers of cell lines that adequately represent the true genetic diversity of cancer). Second, there is a lack of a computational framework proven to be successful for the integrative challenge of generating predictive models of cellular phenotypes based on molecular features of the cell.
Our Center aims to overcome these obstacles, with the ultimate goal of developing predictive models that accurately identify the “Achilles’ heels” of tumors of different genotypes. We are focusing on lung cancer and melanoma because i) there is great unmet medical need, ii) there exist appropriate experimental systems, iii) there is commitment of the investigators to these diseases, and iv) the focus on two cancer types will allow us to determine to what extent the predictive “rules” that govern the behavior of one tumor type are applicable to the other.