In 1962, a peptide purified from the salivary gland of a mouse was shown to accelerate incisor eruption and eyelid opening in newborn mice. The peptide was named the epidermal growth factor (EGF) (Cohen, 1962). Soon thereafter, EGF and members of the family of peptide growth factors had been identified in countless physiological and pathological contexts. EGF binds to a cell surface receptor (EGFR, epidermal growth factor receptor), inducing its dimerization and phosphorylation of several tyrosines residues within its cytoplasmic tail. The phosphorylated tyrosines provide binding sites for cytoplasmic proteins; this couples the activated receptor to the signal transduction cascades and, eventually, to gene expression and processes such as cell division, differentiation, or migration. Abnormal EGFR signaling due to overactive receptors or overexpressed ligands leads to developmental defects and is also associated with many types of cancers. This much was understood about the EGF system when the 1986 Nobel Prize in Medicine was awarded to Stanley Cohen and Rita Levi-Montalcini for their discovery of EGF and NGF (nerve growth factor).
Today, the EGFR is the subject of some 30 000 research papers. Many individual molecules mediating the EGFR-induced responses have been identified and are drug targets in oncology and other areas of medicine (Yarden and Sliwkowski, 2001). Genomics and proteomics approaches will soon make it possible to follow all genes and protein/protein interactions affected by EGFR/ligand binding on the cell surface. The structural details of EGFR interaction with its ligands are becoming progressively understood at the atomistic level and can be followed in real time with modern imaging tools (Burgess et al., 2003). However, neither the contribution of EGFR to tissue morphogenesis in development nor the exact role of deregulated EGFR signaling in disease processes is understood at this time. For example, the mechanistic explanation of the EGF-induced eyelid opening is only beginning to emerge and the attempts to use EGFR expression as prognostic marker in cancer have met with only limited success (Arteaga, 2003).
Quantitative models are necessary in order to deal with the immense complexity associated with the large number of signaling components and multiple feedback loops in the EGFR network (Wiley etal., 2003). Synthesis of the existing information into quantitative models is becoming possible due to the highly conserved nature of EGFR signaling and the ability to test the constructed models in well-developed experimental systems. Only a few dozen out of ~30 000 EGFR-related Medline entries are dedicated to modeling and computational analysis. The existing models can be classified into three groups. Models of binding and trafficking describe the levels and cellular location of free and ligand-bound receptors. Explicit models of signal transduction describe the biochemical transformations induced by extracellular ligands at the level of a single cell. Finally, molecular modeling has provided insight into the structural biology of the EGFR system and has guided the design of receptor tyrosine kinase (RTK) inhibitors.
Quantitative models of ligand-receptor binding and trafficking have been indispensable in dissecting the individual steps in the cycle of receptor (and ligand) endocytosis (Wiley, 2003). Using these models to fit data obtained for the kinetics of ligand-induced receptor internalization and degradation, it has been possible to extract the rate constants for different steps within the endocytic cycle. The extracted constants have explained the experimentally observed differences in the biological effects of different EGFR ligands (notably EGF and TGF-a). At the same time, the pharmacological effects of drugs can now be interpreted and predicted in terms of their effects on different steps in receptor-mediated endocytosis. Importantly, the models developed for the EGFR have been successfully used in the quantitative analysis of other ligand-receptor systems.
Several computational models have been used to interpret and plan EGFR signal transduction experiments in cell culture assays. These models accounted for ligand-receptor binding, receptor dimerization, endocytosis, and signal transduc-tion through the PLCy and MAPK cascades (Schoeberl et al., 2002; Kim et al., 1999; Haugh et al., 2000). The models were used to describe the signals induced by a steplike change in the concentration of exogenous ligand. The predictive capability of these models was demonstrated in two separate cell culture experiments, even though the parameters used for the computational models had been assembled from measurements in multiple cell types and from several in vitro studies.
Almost exclusively, current models of EGFR signaling focus on events at the level of a single cell (or at the molecular level). However, many of the most important in vivo effects of EGFR signaling involve interactions between multiple cells. There are no existing models that describe EGFR signaling at this level, and there is relatively little quantitative understanding that can be translated into directed manipulation of tissue regulation and development. Systematic analysis of EGFR-mediated cell communication requires computational models that simultaneously account for intracellular events and intercellular communication by locally produced EGFR ligands. Only with approaches that consider all of these elements simultaneously can predictions be made that are of value on the organismal scale. Given this complexity, integrated models are nontrivial to test experimentally. Appropriate experimental paradigms that can be used to directly test predictions of complex models are critically needed. Cultured epithelial layers and model organisms of developmental genetics, such as the fruit flies, show great promise for achieving this goal (Shvartsman et al., 2002; Vermeer et al., 2003; Casci and Freeman, 1999).
Even though much of the quantitative studies of EGFR network have assumed that it acts as an independent signaling module, it is highly unlikely that EGFR acts alone in mediating a specific cell or tissue response. Fortunately, an increasing amount of information on how the EGFR system interacts with other pathways, such as TGF^ and cytokine receptor pathways, promises to help us understand how cells integrate signals from multiple pathways to produce a final response.
