15 NSFBIO: Causal modeling of T cell signaling in time and space

A great challenge in biomedical research is to understand how the regulation of cellular activation occurs in the interaction of dozens of signalling components. As most current research only addresses single components of signalling systems, new strategies to address entire signalling systems are required. With the dual objective to further develop methods for the investigation of complex signalling systems and to gain biological insight we study signal amplification in T cell activation by CD28. To explain what that means: T cells or T lymphocytes are central regulatory cells in the immune system. Their activation is critical in immune responses to pathogens, in cancer, and in autoimmune disease. For their activation they require two inputs. The first signal directly communicates the presence of a pathogen; the second signal, costimulation by CD28, communicates that other components of the immune system have recognised the same pathogen. Costimulation then amplifies the intracellular signalling processes triggered by the first signal. However, it remains unknown how this amplification is accomplished.

Signal amplification, similar to many other signalling processes, is of great complexity, as many proteins need to collaborate. A critical component of such complexity is that proteins are rarely evenly distributed throughout cells but enrich at particular subcellular locations at particular times, thus generating complex spatiotemporal distributions. Co-enrichment of two proteins enhances their interaction efficiency. At the scale of many signalling proteins of a cell, spatiotemporal distributions thus determine how information flows through signalling networks in time and space thus regulating cellular function. Microscopy can determine the subcellular distributions of signalling proteins in live cells over time, a process referred to as imaging. Only when applied at a large scale as we uniquely do, imaging can capture the information flow across complex signalling systems as an efficient and unique means to understand the regulation of cellular function. To understand how CD28 amplifies T cell signalling, we will image signalling in T cells lacking CD28 engagement. As the imaging data thus acquired contains very large amounts of data, computational image analysis approaches are required. We develop such approaches collaboratively. Our work is part of a NSF/BBSRC US/UK binational pilot programme. Our US partner, the Murphy laboratory at Carnegie Mellon University, is supported by the NSF to develop advanced computational image analysis approaches for the imaging data we acquire. In combination large-scale imaging and computational image analysis are expected to reveal the mechanisms used by T cells to amplify signalling. Importantly, our strategy will be generally applicable to the analysis of complex signalling systems and thus can be transferred to the analysis of cellular activation in many other physiologically important settings. In addition, as signal amplification is wide spread, we also expect that data gained in T cells will inform mechanisms of signal amplification in other cell types.

Understanding T cell signal amplification is also of medical interest. T cells are of great medical importance, particularly in autoimmune diseases and the immune response to cancer. In collaboration with groups at the University of Bristol and outside in academia and industry, we have begun to explore the role of signalling organisation in the autoimmune disease multiple sclerosis and its therapy and in primary immunodeficiency. Methods and data generated here will be transferred to these projects in the future.

We address how signaling intermediates function as an integrated ensemble in the regulation of cellular activation using mainly large-scale live cell imaging of signal transduction. In particular, we ask how the spatial distribution of a given protein depends upon the spatial distributions of other proteins at that and earlier time points, as these dependencies encode system control of cell function. We will develop general open source computational tools and collect experimental data to solve this challenge for costimulatory signaling during T cell activation. As part of the NSF/BBSRC US/UK binational pilot, this proposal is driven by an exceptionally close collaboration between experimental (UK) and computational (US) scientists. The two principle outcomes will be an integrated general strategy for the quantitative investigation of complex signaling processes as they occur inside living cells and generalizable insight into mechanisms of signal modulation by costimulatory receptors.

Develop open-source tools and collect movies to generate and interpret spatiotemporally-aligned maps of a representative set of signaling molecules. a) Create maps for 35 key molecules in T cell signalling under full stimulus and costimulation-block conditions; b-d) Identify putative complexes by clustering whole cell maps and individual voxels and incorporate diverse sources of information using factor graphs.

Develop open-source tools to infer causal relationships between spatiotemporal patterns. a-c) Estimate spatiotemporal causal relationships among sensors using three computational approaches; d) Test causal relationships using manipulation of spatial distributions.

Develop open-source tools and collect phosphoproteome data to improve models of signaling interactions by including phosphorylation states. a) Collect phosphoproteome kinetics data; b) Compare with putative complex formation from imaging; and c) Use simulations to estimate apparent affinities between proteins.

Data and methods for the academic and industrial scientific community
Complex and spatiotemporally diverse signalling networks are a general feature of cellular regulation. Our strategy to harness spatiotemporal signalling distributions to gain functional insight into such complex systems and the resulting data on organisational mechanisms of signal amplification are therefore of general interest and will be easily transferable to other research projects. They will be made freely available.

Therapeutic potential and its exploitation
Signal amplification in complex spatiotemporally constrained networks underlies many diseases, in particular multigenic diseases such as autoimmune disease, type II diabetes, and cancer. Costimulation specifically plays an important role in autoimmune disease. The application of our functional system imaging strategy to understand T cell signal amplification is likely to substantially promote scientific insight suitable for the development of new therapeutic strategies. While the exploitation of such strategies is beyond the BBSRC remit, the development of methods for their efficient discovery in academia and industry adds to the impact of this application.

Illustrating the comparatively straightforward road from imaging strategy development to future therapeutic application, in a collaborative project our expertise and tools are applied to understand the function of T cells that emerge as the consequence of treatment with immune regulatory peptides (Prof. Wraith, U. Bristol, as commercially pursued by the biotech company Apitope). A contract-based collaboration with UCB Celltech uses our system imaging strategy to characterize a lead compound. Our functional imaging strategy is also being made widely available through incorporation into the portfolio of the contract research company KWS. Academic collaborative access to our strategy is detailed in the academic beneficiaries section. Thus our strategy is likely to enable future applications, in our own research, in the wider academic community and in industry. Its further development thus entails benefits for the wider population in the medium to longer term.

Communications and Engagement
This research has important implications for our understanding of signalling complexity in the regulation of cellular function that commonly underlies multigenic disease. It is therefore essential that our findings be disseminated appropriately to the general public. The Department and University websites publicize grants and important research papers, with key achievements given press releases. The PI maintains his own website for data dissemination (www.bristol.ac.uk/cellmolmed/infect-immune/wuelfing/).

Data on spatiotemporal distributions in T cell activation can be visually arresting. We will use this visual appeal to entice high school students in Bristol to become engaged in science through the Centre for Public Engagement at the University of Bristol, by packaging the full Wuelfing lab website into teacher support materials, and through potential participation in the biennial "Science Alive" fair organized by the University of Bristol.

The proposed studies will present a unique educational opportunity for the two postdoctoral fellows to be employed, one in Bristol supported by the BBSRC and the other in the laboratory of Dr. Robert Murphy at the Carnegie Mellon University in Pittsburgh, our partner in this NSF/BBSRC US/UK binational pilot project supported by the NSF. Coming from a computational and biological background, respectively, both will receive cross-disciplinary training by working on the project together. Such combined training in imaging and computational image analysis will enable them to apply quantitative system scale imaging to other model systems upon establishment of their own independent groups.