- The CSBC U54 Research Center at Vanderbilt
Overall Project Title
- Phenotype Transitions in Small Cell Lung Cancer
- Under construction
Our Center addresses a fundamental issue in oncology: the pervasive and ubiquitous occurrence of phenotypic heterogeneity in any cancer, at all stages of progression. The overarching goal of the Center is to produce a quantitative understanding of plasticity and dynamics of cancer cell phenotypes in a tumor, as we believe this knowledge holds the promise of major advances in treatment. Both genetic and epigenetic factors contribute to tumor heterogeneity. Furthermore, multidirectional interactions of distinct tumor cell phenotypes amongst themselves and with host cells shape the evolutionary trajectory of a tumor, including its metastatic properties. Thus, heterogeneity is a complex, multiscale problem (from genes to molecules to cells to tissues), intrinsically unsuitable to reductionist approaches. Rather, we consider a systems-level approach to current challenges, which include: 1) identifying useful quantitative metrics of a tumor phenotypic space; 2) defining deterministic and stochastic components of heterogeneity at molecular and cellular levels; 3) deriving emergent tumor phenotype dynamics from single-cell behavior; and 4) designing effective treatment strategies based on this system-level knowledge.
To tackle these challenges, we frame tumors as complex adaptive systems and apply concepts from dynamical systems theory. Furthermore, we resolved to focus a multitude of theoretical and experimental approaches on one single cancer type, Small Cell Lung Cancer (SCLC), which at the moment is a disease with dismal outcomes and no improvement in treatment approaches for over half a century.
SCLC is an ideal model system to study the impact of phenotypic heterogeneity on progression and treatment resistance. Virtually all tumors exhibit degrees of plasticity, but SCLC is on the extreme side, facilitating data interpretation and possibly reducing the influence of confounding factors. For instance, SCLC tumors exhibits rapid relapse after initial high levels of response to chemoradiation treatment. Relapse is not associated with recurrent acquired resistance mutations, pointing to a major role for phenotypic state transitions. SCLC initiation and promotion involve obligatory biallelic inactivation of tumor suppressors p53 and Rb, while driver oncogenes appear to be absent. Nonetheless, diverse intra-tumor SCLC cellular phenotypes emerge in patients and mouse model tumors, supporting the idea that SCLC aggressiveness and treatment escape depend of phenotype diversity. Our working hypothesis is that heterogeneous phenotypes and their interaction dynamics form a robust SCLC tumor ecosystem adaptable to perturbations and treatment. In our Center, the overarching goal is to test this hypothesis with a combination of experimental and modeling approaches and if correct, at least in part, demonstrate its general value for other cancer types.
Vito Quaranta, MD
Vito Quaranta, MD, trained in cell biology and immunology, is the Director of the Quantitative Systems Biology Center, and Professor of Biochemistry and Pharmacology at Vanderbilt University School of Medicine. Having studied cancer and the tumor microenvironment for most of his career, for over a decade he has been implementing cutting-edge interdisciplinary efforts melding mathematics, engineering, computation and biology to solve the problem of cancer invasion and metastasis. He has co-developed multiscale mathematical models that predict tumor aggressiveness based on the physical properties of extracellular matrix and adhesive properties of cancer cells. Dr. Quaranta has established single-cell techniques to quantify the rate of proliferation of single cells in response to perturbations (e.g., Fractional Proliferation), which can be applied in high-throughput fashion to measure the dynamics of cancer cell response to drugs. He is also expert in single-cell methodologies to evaluate mechanism of action of cancer targeted therapy, based on the merging of automated time-lapse microscopy with image analysis and computational modeling. To explore the roots of differential cell behavior and tumor heterogeneity, Dr. Quaranta studies the dynamics of transcription factor and signaling networks that define and maintain cell identity, and ultimately contribute to forming the phenotypic landscape of the tumor microenvironment.
Leonard A. Harris, Ph.D
Leonard A. Harris, Ph.D., is a post-doctoral fellow in the Quaranta and Lopez laboratories, with extensive experience in mathematical modeling. He is interested in understanding and uncovering the molecular basis for cancer cell heterogeneity and its connection to drug treatment resistance using a systems biology approach combining computational modeling and in vitro and in vivo experimentation. Formally trained in chemical engineering, Leonard is a leader in the development of BioNetGen (BNGL) rule-based modeling, and an expert in stochastic simulations with the Gillespie method.
Jonathan M. Lehman, M.D., Ph.D.
Jonathan M. Lehman, M.D., Ph.D., is a physician scientist with a research focus on SCLC. One of his original contributions in this field is the emphasis on the underappreciated role of metabolism in SCLC and SSTR2 signaling with its influence on AMPKa phosphorylation in metabolic flux and survival. Jon’s role in the Center will encompass the characterization of SCLC phenotypes both in cell lines and PDXs that he initiates and curates. He works in close collaboration with Pierre Massioni, MD, Director of Cancer Early Detection and Prevention Initiative and an expert in lung cancer.
Carlos F. Lopez, Ph.D
Carlos F. Lopez, Ph.D., is an Assistant Professor of Biochemistry, Biomedical Informatics and Pharmacology at VUSM. He is formally trained in the fields of chemistry, physics, biophysics, biochemistry, and molecular cell biology with a unique emphasis on the development and application of computational modeling techniques. Carlos and his group build cancer models across multiple spatiotemporal time-scales, using techniques from statistical mechanics, molecular simulation, mesoscale modeling, reaction kinetics, bridging the gap between single cells and cell populations. He has a deep commitment to interface experimental and theoretical groups in order to test and expand hypotheses, with the ultimate goal of achieving a fundamental understanding of the rules that govern cancer as a complex system.
Christine M. Lovly, M.D., Ph.D.
Christine M. Lovly, M.D., Ph.D., is an Assistant Professor of Medicine at Vanderbilt University. Research in the Lovly laboratory is directed at understanding and developing improved therapeutic strategies for specific clinically relevant molecular subsets of lung cancer. Her laboratory has a particular focus on mechanisms of sensitivity and resistance to agents used in clinical practice. Christine has positioned herself at the forefront of translational SCLC research with emphasis on liquid biopsies.
Julien Sage, Ph.D.
Julien Sage, Ph.D., is a Professor of Pediatrics and Genetics at Stanford University and the co-Director of the Stanford Cancer Biology PhD program. His laboratory focuses on mouse models of cancer, especially for tumor types mutant for the retinoblastoma (RB) tumor suppressor gene such as small cell lung cancer (SCLC). His laboratory has developed novel mouse models for SCLC and has used these models to investigate key signaling pathways in SCLC, including the role of developmental signaling pathways such as Notch and Hedgehog, as well as paracrine and autocrine signals that promote SCLC growth.
Alissa M. Weaver, M.D., Ph.D.
Alissa M. Weaver, M.D., Ph.D., is the Cornelius Vanderbilt Professor of Cell and Developmental Biology and Pathology, Microbiology, and Immunology. She was appointed as AAAS fellow in 2016. The overall goal of her research is to understand cellular and molecular mechanisms of tumor progression. Her laboratory has a particular focus on the role of membrane trafficking and cytoskeletal rearrangements in promoting tumor invasion and metastasis, especially the role of exosomes and other extracellular vesicles in driving those and other tumor-relevant processes. In this project, her laboratory will focus on the role of secretion in driving SCLC cell-cell interactions, phenotype dynamics, and aggressiveness.
Core Directors and Participating Faculty
Yu Shyr, Ph.D.
Yu Shyr, Ph.D., is the Chair of the Department of Biostatistics at Vanderbilt University School of Medicine. He is expert in biostatistical methodologies for clinical trial design, high-dimensional data preprocessing, and statistical or bioinformatics approaches to big data. He is on the advisory board of several Cancer Centers and SPOREs, and is Principal Investigator of a UO1 grant for the Barrett’s esophagus translational research network coordinating center (BETRNetCC), and PI for data coordination of the NCI SCLC Consortium. Dr. Shyr’s current research interests focus on developing statistical bioinformatics methods for analyzing next-generation sequencing data including sample size requirements for studies based on cell population and single-cell RNA sequencing.
Jonathan M. Irish, Ph.D.
Jonathan M. Irish, Ph.D., is Assistant Professor in the Department of Cell & Developmental Biology, and Pathology, Microbiology & Immunology at VUSM. He is Scientific Director of the Cancer & Immunology Core and the Mass Cytometry Center of Excellence. Research areas in Jonathan’s laboratory include neural stem cell and immune cell signaling interactions in brain tumors; machine learning algorithms to identify healthy and malignant cell types; and, immune mechanisms of disease and therapy response. The central goal of his research is a sytems-level understanding of changes at the single-cell level that alter signaling and lead to therapy-resistant populations in cancer and immunological diseases.
Ken Lau, Ph.D.
Ken Lau, Ph.D., is Assistant Professor of Cell and Developmental Biology at Vanderbilt University. His background is in computational modeling combined with experimentation to study complex biological phenomena. Ken is interested in developing a systems-level understanding of the molecular basis for cell decisions, by utilizing high-dimensional, high-resolution data generated by technologies such as mass cytometry on single cells from fixed tissue, multiplex tissue imaging, and single-cell RNA-seq. To interpret such data, he recently developed p-Creode, an unsupervised algorithm that produces multi-branching graphs and infers a statistically robust hierarchy of cell-state transitions that define developmental trajectories. Ken’s overarching goal is to unravel heterogeneity and plasticity of epithelial and cancer cell populations, and translate such knowledge into therapeutic approaches.
Qi Liu, Ph.D.
Qi Liu, Ph.D., is Assistant Professor of Biomedical Informatics at VUSM. She is expert in bioinformatics methodology development and next generation sequencing data analysis. Her research interests include integrative system biology approaches to the biological basis of complex diseases, and transcriptional and post-transcriptional regulation networks.
Project 1: Modeling the SCLC Phenotypic Space
Leaders: Vito Quaranta, Carlos F. Lopez
Co IS: Christine M. Lovly, Jonathan M. Lehman
Figure1: Potential Sources and Dynamics of Tumor Heterogeneity in the SCLC Phenotypic Landscape. Diversity of tumor cell phenotypic states can occur based on genetic or epigenetic changes that allow cells to adapt to various circumstances, including the microenvironment. Additional features, such as stochasticity of molecular states and cellular interactions (e.g. through secretion) will also contribute to transitions between phenotypic states and differential biological functions of tumors, such as altered ratios of drug resistant, sensitive, and quiescent cell populations
The overarching goal of this project is to develop a global blueprint of SCLC phenotypic space, clarifying bias imposed by genomic alterations, and epigenetic basis for phenotype transition or selection in response to drugs. We leverage our expertise in modeling the origin of heterogeneous phenotypes from transcription factor (TF) networks inferred from transcriptomics data. Simulations of TF network dynamics via logic-based models enable the identification of attractors, which roughly correspond to SCLC differentiation states defined by profiles of activated or silenced TFs. We define a gene ontology metric to identify biological similarities and differences between phenotypes across a variety of experimental systems including human cell lines, patient-derived xenografts and primary tumors, as well as tumors from SCLC genetically engineered mouse models (GEMMs). The resulting phenotype map informs studies aimed at connecting model systems to patients. A limitation of SCLC is the scarcity of patient specimens, since biopsies or surgery are rarely performed beyond initial diagnosis. We circumvent this barrier by using instead liquid biopsies of circulating, cell-free DNA as a clinical proxy for the primary tumor, allowing a connection between genomic alterations and phenotypic space states of SCLC tumors. To evaluate the relative role of state transitions vs. selection underneath SCLC plasticity and drug treatment evasion, we use DNA barcoding and information theory techniques to quantify rates of diversification of SCLC phenotypes in response perturbations. Specifically, we map trajectories of cells within the SCLC phenotype space as cells adapt and evade treatment. In summary, we propose to develop a comprehensive view of SCLC phenotypic heterogeneity, linking transcriptomic, genomic, and functional features of SCLC cells across diverse experimental model systems and patient primary tumor specimens. We will link these observations to clinically measurable variables, develop a unified map of phenotypic response dynamics in response to therapy, and analyze the role of plasticity in phenotype shifting and evolution of drug resistance., providing possible novel avenues to SCLC treatment strategies.
Figure 2: Mutational analysis of plasma cfDNA from 27 patients with SCLC. Summary of mutations identified by patient at any time point. Patients are separated by stage of disease at diagnosis (LS-limited stage; ES- extensive stage). Alterations are color coded per the figure legend below the image. The mutation frequencies for each gene are plotted on the right panel.
Project 2: Phenotype interactions and dynamics in SCLC tumors
Leaders: Alissa M. Weaver, Julien Sage, Carlos F. Lopez
Key Personnel: Leonard A. Harris
Figure 3: Schematic of the preliminary SCLC population dynamics (PopD) model. Black arrows represent cell fate (division, death, differentiation); red arrows represent positive feedback; red circles represent negative feedback; ∅ represents cell death.
In this project, we assess interactions of distinct SCLC cell phenotypes with each other, directed at forming an ecosystem that supports progression and treatment escape. Using a combination of mathematical modeling and experimentation, we aim to quantitatively appraise the relative role of each phenotype in driving the dynamics of tumor aggressiveness as a whole under a variety of perturbation. Cell population models will be simulated both deterministically, to predict the average behavior of a tumor, and stochastically, to evaluate the role of intrinsic and extrinsic noise. The overarching goal is to predict dependencies that are targetable, both at the molecular and cellular level, and then test these predictions in vitro and in experimental SCLC GEMMs.
Tumor-propagating cells (TPCs) in SCLC GEMM tumors are neuroendocrine and strongly tumorigenic. We also characterized cell populations derived from these TPCs with distinct phenotypes, including non-neuroendocrine subpopulations that can promote the growth and the spread of neuroendocrine TPCs. With a combination of experimental and mathematical approaches we will predict key interactions between SCLC phenotypic subpopulations to uncover fragility/intervention points that could be used for treatment. We will also determine whether specific molecular cargoes carried by purified extracellular vesicles and/or secreted as soluble factors affect SCLC phenotype maintenance, growth, and survival dynamics in vitro and in vivo. The identification of secreted factors involved in SCLC communication will allow disruption of feedback loops for cell-population level mathematical model validation and eventual therapeutic intervention.
Single Cell Biology and Data Analysis (SCB-DA) Core
Leaders: Jonathan Irish, Yu Shyr
Co-Is: Ken S. Lau, Qi Liu
This Center shared resource provides two critical services: 1) state-of-the-art single-cell analytic experimental procedures, including single-cell RNA-sequencing on several platforms, and mass cytometry analyses; 2) bioinformatics and statistical data analysis services, including dimensionality reduction for interpretation and visualization of high-dimensional data. These services support in a cost-efficient manner analysis and integration of data related to phenotypic heterogeneity in SCLC. The SCB-DA Core will drive development and/or application of cutting edge single-cell and bioinformatics technologies, such as the MEM score metric for separation of cell subpopulations with, e.g., mass cytometry, or the p-Creode algorithm to infer phenotypic trajectories.
Leaders: Alissa M. Weaver
Co-Lead: Carlos F. Lopez
In addition to cutting edge cancer systems biology research, we are committed to reaching out to the community to disseminate our techniques, modeling approaches and accomplishments. Activities include outreach through: 1) our website; 2) hands-on workshops focused on single-cell technologies and modeling approaches; and 3) seminars and symposia. Our Center investigators are leaders in the area of tumor heterogeneity, network biology, single-cell biology, analysis and modeling of complex systems. They will continue to innovate and drive forward these fields as part of the Center. As new tools and approaches are developed, we will disseminate them through hands-on workshops and online modules.