Final Report Summary - AMPKIN (Systems biology of the AMP-activated protein kinase pathway)
The ultimate aim of the AMPKIN project was to achieve an advanced understanding of the dynamic operation of the AMP-activated protein kinase (AMPK) signaling pathway. This pathway played a central role in monitoring the cellular energy status of cells and organisms. The main objective of the project was to generate predictive kinetic mathematical descriptions of pathway activation/deactivation in yeast and mammalian cells and, thereby, to identify potential drug targets to treat human metabolic diseases. This overall project objective led to the following specific scientific, technical and innovation objectives:
- to establish and critically compare the network structures of the AMPK pathway from yeast and from mammalian cells using existing data and knowledge from literature and databases: the objective was achieved as a reconstruction of the cellular Snf1 network in yeast was generated, which showed that yeast and mammalian AMPK have similar targets and physiological roles;
- to generate, optimise and verify assay systems for as many different steps as possible in the AMPK pathway of yeast and mammalian cells in order to generate quantitative data and maximise the use of real data in modelling: the objective was achieved for mammalian cells and more than fully achieved for yeast cells, where an impressive repertoire of tools, many suitable beyond the organism, have been generated;
- to generate reference quantitative dynamic datasets following activation and deactivation of the AMPK pathway in yeast and mammalian cells. This reference data set would be used for generating dynamic models of the pathways and to optimise parameters that could not be determined experimentally: the datasets generated allowed building mathematical models of the yeast Snf1 pathway. The available data, however, limited to some extent the modelling of the mammalian AMPK. The objective was fully achieved with respect to yeast;
- to generate and critically compare dynamic models for the yeast and mammalian AMPK pathway and to use information from the yeast model to complement gaps in the mathematical description of the mammalian model: the objective was only fully achieved for the yeast Snf1 pathway;
- generation of tools for system perturbation, which would be used to generate data for model testing and iterative model improvement and potentially for development of drug screening approaches: the success of expressing an entire heterologous complex in yeast was a major breakthrough. At that point it was not known why the mammalian complex expressed in yeast did not respond to compounds that are effective in mammalian cells. There were different hypotheses that are tested by subsequent work;
- to provide 'dynamic' datasets from experiments employing a range of defined system perturbations in both yeast and mammalian cells with the aim to test and iteratively improve the models and to optimise the underlying parameters: a wealth of quantitative and dynamic data has been generated both for the yeast and the mammalian AMPK pathways;
- to generate iteratively improved mathematical models in order to determine system properties and to provide an assessment of similarities and dissimilarities of the models in yeast and mammalian cells and hence of the significance and the limitations of the approach of comparative modelling from experimental and theoretical perspectives: the objective was achieved for the yeast system. Modelling combined with experimentation allowed addressing different relevant research questions, such as the role of feedback and feed-forward loops. In this way, modelling was closely integrated with experimentation to address and solve research problems.
- to establish and critically compare the network structures of the AMPK pathway from yeast and from mammalian cells using existing data and knowledge from literature and databases: the objective was achieved as a reconstruction of the cellular Snf1 network in yeast was generated, which showed that yeast and mammalian AMPK have similar targets and physiological roles;
- to generate, optimise and verify assay systems for as many different steps as possible in the AMPK pathway of yeast and mammalian cells in order to generate quantitative data and maximise the use of real data in modelling: the objective was achieved for mammalian cells and more than fully achieved for yeast cells, where an impressive repertoire of tools, many suitable beyond the organism, have been generated;
- to generate reference quantitative dynamic datasets following activation and deactivation of the AMPK pathway in yeast and mammalian cells. This reference data set would be used for generating dynamic models of the pathways and to optimise parameters that could not be determined experimentally: the datasets generated allowed building mathematical models of the yeast Snf1 pathway. The available data, however, limited to some extent the modelling of the mammalian AMPK. The objective was fully achieved with respect to yeast;
- to generate and critically compare dynamic models for the yeast and mammalian AMPK pathway and to use information from the yeast model to complement gaps in the mathematical description of the mammalian model: the objective was only fully achieved for the yeast Snf1 pathway;
- generation of tools for system perturbation, which would be used to generate data for model testing and iterative model improvement and potentially for development of drug screening approaches: the success of expressing an entire heterologous complex in yeast was a major breakthrough. At that point it was not known why the mammalian complex expressed in yeast did not respond to compounds that are effective in mammalian cells. There were different hypotheses that are tested by subsequent work;
- to provide 'dynamic' datasets from experiments employing a range of defined system perturbations in both yeast and mammalian cells with the aim to test and iteratively improve the models and to optimise the underlying parameters: a wealth of quantitative and dynamic data has been generated both for the yeast and the mammalian AMPK pathways;
- to generate iteratively improved mathematical models in order to determine system properties and to provide an assessment of similarities and dissimilarities of the models in yeast and mammalian cells and hence of the significance and the limitations of the approach of comparative modelling from experimental and theoretical perspectives: the objective was achieved for the yeast system. Modelling combined with experimentation allowed addressing different relevant research questions, such as the role of feedback and feed-forward loops. In this way, modelling was closely integrated with experimentation to address and solve research problems.
