|Start Date / Time||September 23, 2015 at 9:00 AM|
|End Date / Time||September 23, 2015 at 10:00 AM|
|Location||9455 MRB IV|
|Presenter Name||Vito Quaranta, M.D.|
|Presentation Title||Emergence of cellular phenotypes from transcription factor networks: A systems-level approach|
|Status||This meeting has already occurred|
Short Summary: Our systems biology pipeline aims to identify cell reprogramming strategies by an iterative process of data-driven modeling, model-driven hypothesis generation, experimental testing or validation, and mathematical model refinement. Hopefully, this systems-level approach will open avenues to non-invasive treatment of Small Cell Lung Cancer and diabetes.
Long Summary: We have applied a systems biology approach to Small Cell Lung Cancer (SCLC), in order to test the hypothesis that SCLC phenotypic heterogeneity is a main driver of dismal treatment outcomes in this lethal disease (<5% survival at 1 year from diagnosis). Previously, SCLC has been reported to adopt one of two phenotypic states, neuroendocrine (NE) or mesenchymal-like (ML). In contrast, our systems-level analyses uncover an unsuspected degree of phenotypic heterogeneity. Thus, by co-expressed gene analysis of SCLC tumors and cell lines datasets (WGCNA), coupled to ARACNE transcription factor (TF) identification, we derive a SCLC-specific network of 33 TFs. Logic-based model simulations predict that the SCLC TF network settles into ~40 equilibrium states (technically referred to as attractors), which can be broadly classified into two clusters, NE or ML, by knowledge-based algorithms. This predicted vast phenotypic heterogeneity was validated experimentally by profiling TF expression in cell lines and tumor specimens, as well as by multidimensional single-cell biomarker plots. Within an SCLC cell line (presumably genetically homogeneous), both NE and ML phenotypes are observed, sometimes within a single cell, suggesting that the origin of SCLC phenotypic heterogeneity is at least in part non-genetic and is rooted in the topology itself of the SCLC TF network. Our current challenge is to determine whether state transitions within this phenotypic heterogeneity landscape are linked to treatment failure in SCLC.
Our results in SCLC have broad implications for normal and malignant cell reprogramming. More specifically, they pose the question as to whether the heuristics developed in modeling SCLC are generalizable to describing TF network dynamics in stable physiological systems, such as differentiation of exocrine and endocrine cells in pancreas development. We intend to use a literature derived Boolean model of pancreatic TF networks to generate hypotheses of reprogramming strategies that would facilitate transitions between differentiated endocrine cells. Preliminary results show our pancreatic TF network is able to recapitulate the development of stable alpha and beta cell attractor states from an endocrine precursor state, after optimization of interaction strengths using particle swarm non-linear optimization algorithms.
In summary, our systems biology pipeline allows for an iterative process of hypothesis generation, experimental testing, and model refinement of cell reprogramming strategies that hopefully will open up avenues to non-invasive treatment of diabetes and SCLC.