Disentangling the Complexity of HGF Signaling by Combining Qualitative and Quantitative Modeling.

Signaling pathways are characterized by crosstalk, feedback and feedforward mechanisms giving rise to highly complex and cell-context specific signaling networks. Dissecting the underlying relations is crucial to predict the impact of targeted perturbations. However, a major challenge in identifying cell-context specific signaling networks is the enormous number of potentially possible interactions. Here, we report a novel hybrid mathematical modeling strategy to systematically unravel hepatocyte growth factor (HGF) stimulated phosphoinositide-3-kinase (PI3K) and mitogen activated protein kinase (MAPK) signaling, which critically contribute to liver regeneration. By combining time-resolved quantitative experimental data generated in primary mouse hepatocytes with interaction graph and ordinary differential equation modeling, we identify and experimentally validate a network structure that represents the experimental data best and indicates specific crosstalk mechanisms. Whereas the identified network is robust against single perturbations, combinatorial inhibition strategies are predicted that result in strong reduction of Akt and ERK activation. Thus, by capitalizing on the advantages of the two modeling approaches, we reduce the high combinatorial complexity and identify cell-context specific signaling networks.

Detecting and removing inconsistencies between experimental data and signaling network topologies using integer linear programming on interaction graphs.

Cross-referencing experimental data with our current knowledge of signaling network topologies is one central goal of mathematical modeling of cellular signal transduction networks. We present a new methodology for data-driven interrogation and training of signaling networks. While most published methods for signaling network inference operate on Bayesian, Boolean, or ODE models, our approach uses integer linear programming (ILP) on interaction graphs to encode constraints on the qualitative behavior of the nodes. These constraints are posed by the network topology and their formulation as ILP allows us to predict the possible qualitative changes (up, down, no effect) of the activation levels of the nodes for a given stimulus. We provide four basic operations to detect and remove inconsistencies between measurements and predicted behavior: (i) find a topology-consistent explanation for responses of signaling nodes measured in a stimulus-response experiment (if none exists, find the closest explanation); (ii) determine a minimal set of nodes that need to be corrected to make an inconsistent scenario consistent; (iii) determine the optimal subgraph of the given network topology which can best reflect measurements from a set of experimental scenarios; (iv) find possibly missing edges that would improve the consistency of the graph with respect to a set of experimental scenarios the most. We demonstrate the applicability of the proposed approach by interrogating a manually curated interaction graph model of EGFR/ErbB signaling against a library of high-throughput phosphoproteomic data measured in primary hepatocytes. Our methods detect interactions that are likely to be inactive in hepatocytes and provide suggestions for new interactions that, if included, would significantly improve the goodness of fit. Our framework is highly flexible and the underlying model requires only easily accessible biological knowledge. All related algorithms were implemented in a freely available toolbox SigNetTrainer making it an appealing approach for various applications.

Modeling approaches for qualitative and semi-quantitative analysis of cellular signaling networks.

A central goal of systems biology is the construction of predictive models of bio-molecular networks. Cellular networks of moderate size have been modeled successfully in a quantitative way based on differential equations. However, in large-scale networks, knowledge of mechanistic details and kinetic parameters is often too limited to allow for the set-up of predictive quantitative models.Here, we review methodologies for qualitative and semi-quantitative modeling of cellular signal transduction networks. In particular, we focus on three different but related formalisms facilitating modeling of signaling processes with different levels of detail: interaction graphs, logical/Boolean networks, and logic-based ordinary differential equations (ODEs). Albeit the simplest models possible, interaction graphs allow the identification of important network properties such as signaling paths, feedback loops, or global interdependencies. Logical or Boolean models can be derived from interaction graphs by constraining the logical combination of edges. Logical models can be used to study the basic input-output behavior of the system under investigation and to analyze its qualitative dynamic properties by discrete simulations. They also provide a suitable framework to identify proper intervention strategies enforcing or repressing certain behaviors. Finally, as a third formalism, Boolean networks can be transformed into logic-based ODEs enabling studies on essential quantitative and dynamic features of a signaling network, where time and states are continuous.We describe and illustrate key methods and applications of the different modeling formalisms and discuss their relationships. In particular, as one important aspect for model reuse, we will show how these three modeling approaches can be combined to a modeling pipeline (or model hierarchy) allowing one to start with the simplest representation of a signaling network (interaction graph), which can later be refined to logical and eventually to logic-based ODE models. Importantly, systems and network properties determined in the rougher representation are conserved during these transformations.

The Process-Interaction-Model: a common representation of rule-based and logical models allows studying signal transduction on different levels of detail.

BACKGROUND: Signaling systems typically involve large, structured molecules each consisting of a large number of subunits called molecule domains. In modeling such systems these domains can be considered as the main players. In order to handle the resulting combinatorial complexity, rule-based modeling has been established as the tool of choice. In contrast to the detailed quantitative rule-based modeling, qualitative modeling approaches like logical modeling rely solely on the network structure and are particularly useful for analyzing structural and functional properties of signaling systems. RESULTS: We introduce the Process-Interaction-Model (PIM) concept. It defines a common representation (or basis) of rule-based models and site-specific logical models, and, furthermore, includes methods to derive models of both types from a given PIM. A PIM is based on directed graphs with nodes representing processes like post-translational modifications or binding processes and edges representing the interactions among processes. The applicability of the concept has been demonstrated by applying it to a model describing EGF insulin crosstalk. A prototypic implementation of the PIM concept has been integrated in the modeling software ProMoT. CONCLUSIONS: The PIM concept provides a common basis for two modeling formalisms tailored to the study of signaling systems: a quantitative (rule-based) and a qualitative (logical) modeling formalism. Every PIM is a compact specification of a rule-based model and facilitates the systematic set-up of a rule-based model, while at the same time facilitating the automatic generation of a site-specific logical model. Consequently, modifications can be made on the underlying basis and then be propagated into the different model specifications - ensuring consistency of all models, regardless of the modeling formalism. This facilitates the analysis of a system on different levels of detail as it guarantees the application of established simulation and analysis methods to consistent descriptions (rule-based and logical) of a particular signaling system.

Large-scale network models of IL-1 and IL-6 signalling and their hepatocellular specification.

The pro-inflammatory cytokines interleukin 1 (IL-1) and 6 (IL-6) are crucially involved in the regulation of a multitude of physiological processes, in particular coordinating the immune response upon bacterial infection and tissue injury. Both interleukins induce complex signalling cascades and trigger the production of mitogenic, pro-proliferative, anti-apoptotic, chemotactic, and pro-angiogenic factors thereby affecting the delicate balance between regeneration vs. invasive growth, tumourigenesis and metastasis. Moreover, several links to insulin resistance have been found within their associated signalling networks. Focusing on this from a systems biology perspective, we introduce comprehensive large-scale network models of IL-1 and IL-6 signalling which are based on a logical modelling approach and reflect the current biological knowledge. Theoretical network analysis enabled us to uncover general topological features and to make testable predictions on the stimulus-response behaviour of the networks. In this context, non-intuitive network-wide species dependencies as well as structures of regulatory feedback and feed-forward mechanisms could be characterised. By integrating high-throughput phosphoproteomic data from primary human hepatocytes we optimised the model structures to obtain models with high prediction accuracy for hepatocytes. Our model-based data analysis, for instance, suggested model modifications regarding (i) Akt contribution to IL-1-stimulated p38 MAPK activation and (ii) insignificant p38 MAPK activation in response to IL-6. In light of the presented results and in conjunction with the detailed model documentations, both models hold great potential for theoretical studies and practical applications.

Computing combinatorial intervention strategies and failure modes in signaling networks.

The identification of combinatorial intervention strategies and the elucidation of failure modes that may cause aberrant behavior of cellular signaling networks are highly relevant topics in cell biology, medicine, and pharmaceutical industry. We have recently introduced the concept of minimal intervention sets (MISs)--minimal combinations of knock-ins and knock-outs provoking a desired/observed response in certain target nodes--to tackle those problems within a Boolean/logical framework. We first generalize the notion of MISs and then present several techniques for search space reduction facilitating the enumeration of MISs in networks of realistic size. One strategy exploits topological information about network-wide interdependencies between the nodes to discard unfavorable single interventions. A similar technique checks during the algorithm whether all target nodes of an intervention problem can be influenced in appropriate direction (up/down) by the interventions contained in MIS candidates. Another strategy takes lessons from electrical engineering: certain interventions are equivalent with respect to their effect on the target nodes and can therefore be grouped in fault equivalence classes (FECs). FECs resulting from so-called structural equivalence can be easily computed in a preprocessing step, with the advantage that only one representative per class needs to be considered when constructing the MISs in the main algorithm. With intervention problems from realistic networks as benchmarks, we show that these algorithmic improvements may reduce the computation time up to 99%, increasing the applicability of MISs in practice.

Discrete logic modelling as a means to link protein signalling networks with functional analysis of mammalian signal transduction.

Large-scale protein signalling networks are useful for exploring complex biochemical pathways but do not reveal how pathways respond to specific stimuli. Such specificity is critical for understanding disease and designing drugs. Here we describe a computational approach--implemented in the free CNO software--for turning signalling networks into logical models and calibrating the models against experimental data. When a literature-derived network of 82 proteins covering the immediate-early responses of human cells to seven cytokines was modelled, we found that training against experimental data dramatically increased predictive power, despite the crudeness of Boolean approximations, while significantly reducing the number of interactions. Thus, many interactions in literature-derived networks do not appear to be functional in the liver cells from which we collected our data. At the same time, CNO identified several new interactions that improved the match of model to data. Although missing from the starting network, these interactions have literature support. Our approach, therefore, represents a means to generate predictive, cell-type-specific models of mammalian signalling from generic protein signalling networks.

The logic of EGFR/ErbB signaling: theoretical properties and analysis of high-throughput data.

The epidermal growth factor receptor (EGFR) signaling pathway is probably the best-studied receptor system in mammalian cells, and it also has become a popular example for employing mathematical modeling to cellular signaling networks. Dynamic models have the highest explanatory and predictive potential; however, the lack of kinetic information restricts current models of EGFR signaling to smaller sub-networks. This work aims to provide a large-scale qualitative model that comprises the main and also the side routes of EGFR/ErbB signaling and that still enables one to derive important functional properties and predictions. Using a recently introduced logical modeling framework, we first examined general topological properties and the qualitative stimulus-response behavior of the network. With species equivalence classes, we introduce a new technique for logical networks that reveals sets of nodes strongly coupled in their behavior. We also analyzed a model variant which explicitly accounts for uncertainties regarding the logical combination of signals in the model. The predictive power of this model is still high, indicating highly redundant sub-structures in the network. Finally, one key advance of this work is the introduction of new techniques for assessing high-throughput data with logical models (and their underlying interaction graph). By employing these techniques for phospho-proteomic data from primary hepatocytes and the HepG2 cell line, we demonstrate that our approach enables one to uncover inconsistencies between experimental results and our current qualitative knowledge and to generate new hypotheses and conclusions. Our results strongly suggest that the Rac/Cdc42 induced p38 and JNK cascades are independent of PI3K in both primary hepatocytes and HepG2. Furthermore, we detected that the activation of JNK in response to neuregulin follows a PI3K-dependent signaling pathway.