Cells are constantly bombarded with molecular signals that they must interpret in order to respond appropriately to environmental cues. Much of the cellular decision making is carried out via signaling cascades that translate the engagement of surface receptors first into altered protein phosphorylation patterns and then into gene regulation. A critical issue in this field of signal transduction is to understand how cells accurately respond to environmental cues. To address this issue, two types of cells (PC12 pheochromocytoma and T lymphocytes) will be studied in this project. For these cells, the activation of a specific signaling cascade (e.g. mitogen-associated protein kinase pathway or MAPK) is triggered upon sensing of external ligands, and this cascade results in either cellular quiescence, differentiation and/or proliferation. PC12 cells can discriminate between nerve growth factor and epidermal growth factor ligands by differentially regulating the dynamics of their signaling response. T cells detect the presence of few foreign-derived ligands while avoiding spurious activation by large quantities of self-derived ligands. In both types of cell, this ligand discrimination has been shown to be highly tunable (e.g. during neurocytic or thymic development) and at the same time reliable (despite stochastic variability among cells).
In this project, the PI will combine computational and experimental approaches to reconcile this robustness and variability in cell signaling. First, computer models of cell signaling will be developed to probe the theoretical robustness of ligand discrimination in the context of stochastic fluctuations in chemical reactions, and phenotypic variation (i.e. expression of different levels of signaling components). Second, the variability of signaling responses due to phenotypic variability will be monitored experimentally: alteration in cell responsiveness will be mapped onto the differential expression of key signaling components at the single cell level. The PI will also use perturbation (through chemical inhibition of kinase/phosphatase enzymes or up/down regulation of signaling components) to modulate cell responsiveness and test theoretical predictions.
This project is based on the PI's development of computer models of signal transduction, as well as the development of multiplexed detection of phosphorylated states in individual cells. Thus, it is a unique opportunity to merge theoretical and experimental approaches to study the systems biology of cell signaling.
Broader impact
This project addresses a fundamental paradox in signal transduction whereby PC12 cells and T lymphocytes utilize feedback regulation of MAPK activation to enforce ligand detection with functional variability yet with reliability. Understanding such feedback regulation will allow the derivation of general principles at the systems biology level to explain how signaling networks can reconcile robustness and variability in their input/output relationships. The project will include the training of postdoctoral, graduate and undergraduate students in systems biology, ranging from experimentation to computer modeling. Tutorials, journal clubs and discussion groups will be used to introduce students and postdoctoral fellows to systems biology. Moreover, new implementations of biochemical computer models will be broadcast to the scientific community such that other researchers can use them, edit them, and test their own hypotheses in signal transduction. Experimental data will also be deposited in repositories (e.g. DREAM3 database) to help the systems biology community test different modeling approaches.
Research activities Using funding from the NSF, we have successfully launched a sustainable research effort to characterize quantitatively the immune system. We pursued our interest in documenting the phenotypic variability of cells, in varied systems (from T lymphocytes of the mouse immune system, to peripheral blood mononuclear B cells of humans, to cancer cell lines). We established a new methodology (termed "Single-cell Heterogeneity Analysis" or SCHA) to systematically correlate the variability in signal transduction with the endogenous variability in the expression levels of receptors and signaling regulators (kinase, phosphatases, scaffolds etc.). This methodology generates quantitative observations that computational modeling must fit and that drives further understanding of the system. We applied SCHA to dissect the variability in cytokine signaling in T lymphocytes. As part of the SCHA methodology, we created a software package, ScatterSlice, to quantify the functional heterogeneity of cellular signaling responses stemming from variation in expression of key signaling molecules. As a proof of principle, we apply this new methodology to probe the signaling responses for receptors in the common gamma chain (gc) family, comprised of the multi-subunit receptors for interleukins- (IL-), 2, 4, 7, 9, 15 and 21. This receptor family, by definition, shares the gc chain (CD132). IL-2 and IL-15 also share an additional chain, the β chain (CD122). Each cytokine receptor has a proprietary α chain that endows it with specificity through enhanced affinity for its respective cytokine. Quantitative understanding of the differential responsiveness to gc cytokines at the level of individual cells remains an elusive issue. As such, we investigated whether the degree of activation influences the signaling properties of other cytokine. We first quantified, with high resolution in a population of isogenic T cells, the influence of IL-2Rα on IL-2 sensitivity. In agreement with previous results, we report that large variation in IL-2Rα accounts for much of the variability of cells’ sensitivity to IL-2. More surprisingly, we identified IL-2Rα as a potent regulator of IL-7 signaling. Modulation of IL-7Rα expression or signaling capacity could be ruled out as a mechanism explaining the dependency of IL-7 sensitivity on IL-2Rα. Consequently, we introduced a simple thermodynamic model to demonstrate that high levels of IL-2Rα could inhibit IL-7 sensitivity through preformation of heterotrimeric IL-2 receptors in the absence of IL-2. This modeling effort relies on Bayesian statistics to integrate prior thermodynamic knowledge (e.g. affinity constants for all chemical reactions) into parameter optimization and model inference. Extending our model to other cytokines of the common gc family, we then recapitulate different measured effects of IL-2Rα on cytokine sensitivities. The significance of our findings for the common γ chain cytokine pathways relate to the sharp transition between effector and memory phenotypes in the T cell compartment. We further applied SCHA to document the intricacies of signaling crosstalks in lymphocytes. We systematically measured the dynamics of accumulation of the IL-2 cytokine in in vitro cultures of primary mouse T lymphocytes. In parallel, we monitored the expression of IL-2 receptor chains and the downstream signaling response. A simple thermodynamic model of the IL-2 pathway was shown to be insufficient to fit our experimental data, in particular because of the large amount of variability within the isogenic population of T cells under consideration. Using SCHA, we parsed our data to quantify the responsiveness of the IL-2 receptor at the individual cell level, upon activation and differentiation. We found that there exists a critical inhibition of the IL-2R signaling response while antigen signaling response lasts: despite upregulation of IL-2Rα and accumulation of IL-2 in the extracellular medium, T cells were found to be hyporesponsive in an antigen-dependent manner. This enforces a time separation between TCR signaling and IL-2R signaling. The dynamic implications of this cross-talk between antigen and cytokine receptors were probed experimentally and modeled quantitatively. Functionally, our study highlights how variability in signaling response can be leveraged by a population of T cells to regulate and synchronize its collective response to external perturbations. We also used NSF funding to apply our SCHA methodology into human immunology. In collaboration with clinicians at MSKCC, we probed the response of human B and T lymphocytes. The major findings in this project quantifies how dysregulation of the BCR signaling pathway discriminates between normal B cells (collected from blood samples of healthy volunteers) and B cells harvested from chronic lymphocytic leukemia patients (CLL): specifically, we found an hypo-responsiveness in the PLCg2/BLNK pathway compared to SYK activation that correlates stringently with the severity of the disease. In the context of human T cell biology, our SCHA methodology demonstrated how variable the antigen response of cytotoxic T cells was, specially based on varied response to the IL-15 cytokines. These studies are direct application of our SCHA methodology that we developed based on NSF funding.
