Center for Cancer Systems Therapeutics (CaST)
Project 1 - Elucidating the regulatory logic that is responsible for maintaining cancer cell state, independent of specific initiating events and endogenous/exogenous perturbations.
Among the methods CaST is using are approaches for annotating structural and functional features using a template library. The structure of a query protein (A) is used to scan a library of templates with known function (B). Templates can be proteins with various binding partners including other proteins (green), peptides (teal), RNA/DNA (brown) or small molecules (red star). For each complex, the template and query are superposed based on global or local similarity (C, dotted line). An interaction model is created and used to determine whether the query has functional properties similar to the template; e.g., physical interaction energy derived from residues interactions (D, yellow lines). Features such as these can then be used as imput to machine learning approaches.
The aim of this project is to systematically elucidate and experimentally validate the mechanisms that are responsible for tumor homeostasis. We will study tumor cell states in the full context of their heterogeneity, with respect to both tumor composition and microenvironment.
Additionally, we will develop novel methodologies and algorithms to 1) elucidate the logic governing tumor state transition events that result in progression or drug resistance and 2) dissect the full complement of tumor states that co-exist within a patient.
These algorithms will be instrumental in furthering our understanding of the complex cell-cell molecular interactions between tumor and stroma, including those relevant to immunotherapy. Addressing these challenges will require the use of novel methodological approaches that can account for both cellular heterogeneity and the time-dependent nature of tumor homeostasis, including:
- Integration of genetic, epigenetic, structural, and biochemical approaches. For instance, we are integrating structural and functional biology approaches pioneered in the Honig lab to functionally annotate the key proteins that regulate tumor homeostasis. We plan to create models of transcriptional, post-transcriptional, and signal transduction regulation to study the time-dependent regulatory feedback mechanisms contributing to homeostatic control of cell state.
- Probabilistic methodologies developed previously in the Califano and Vitkup labs will be further refined, leading to fully mechanistic kinetic models.
- Attractor metagenes, which have been identified by the Anastassiou lab as highly effective predictors of tumor outcome, are driven by the homeostatic logic of the cancer cell. Similarly, we will study homeostatic tumor cell control (a) across the heterogeneity of tumor-microenvironment interactions, and (b) across distinct tumor cell states that coexist within the same tumor.
These studies and the development of novel methodologies for the study of time-dependent behavior and cell–cell communication processes will require entirely novel data not currently available in the public domain. As a result, we will leverage the pioneering technology work by the Sims lab, available via the Molecular Profiling Core (MPC), as well as advanced CRISPR-based methodologies developed in the Wang lab, to generate time-series molecular profiles of both bulk tumors and individual tumor cells following genetic and small-molecule perturbations. Availability of such data will be critical to implementing novel methodologies for the dissection of feedback loops and more predictive kinetic models of the homeostatic control machinery of the cancer cell.
Finally, using a pan-cancer approach, we will apply our methodologies to each patient in the Cancer Genome Atlas (TCGA), to assess tumor homeostasis and heterogeneity on a per-patient basis. We expect to identify and functionally characterize a small subset of proteins (estimated to be around 400) that fully recapitulate tumor homeostatic control across 20 distinct cancer organ sites and ~100 molecularly distinct subtypes. We will fully annotate these proteins — both functionally and structurally — and will deposit the results in a comprehensive database, CHoPD (Cancer Homeostasis Protein Database). We will also generate a matched and equally comprehensive set of reagents and libraries for systematic, CRISPR-based perturbation of these proteins and protein pairs that are relevant across the full spectrum of TCGA samples.