Merrimack Pharmaceuticals (www.merrimackpharma.com) has developed a computational model of the ErbB receptor network. "This presentation will show how network biology can help you to understand these complex systems and perhaps help you to either develop better drugs or target a specific patient population," says Birgit Schoeberl, Ph.D., associate director, computational biology.
The ErbB (epidermal growth factor receptor family) signal transduction pathway comprises four receptors with 12 different ligands known to bind to those receptors. Because of its complexity, it is difficult to develop drugs high in specificity.
"The idea is, if we better understand how these four different receptors interact and how these downstream signal transduction pathways influence each other, it will help us develop better drugs," Dr. Schoeberl explains. She also plans to present the response patterns for ten different tumor cell lines to anti-EGFR drugs comparing simulation results with wet-lab experimental data.
Utilizing high-density antibody microarrays with proprietary surface chemistry, the company is able to quantify receptors on different tumor cell lines via a high throughput method that also makes data generation more efficient.
The computational model, which is based on up to 500 differential equations, is trained with experimental data (from A431 cells) using protein expression levels and time course data of protein phosphorylation, specifically ERK and AKT phosphorylation. These are two kinases downstream of the ErbB receptor and essential to inducing transcription factors, which leads to cell proliferation.
"We use ERK and AKT phosphorylation as a read out for the inhibitor efficacy. Then we validate the model, making predictions of how an inhibitor would inhibit, for example, ErbB1 phosphorylation in the cell lines. Then we compare the experiments with the model prediction.
"We tested this on eight different types of breast and ovarian cancer cell lines, and the model predicted those responses quite well," summarizes Dr. Schoeberl. The company says its approach is unique because of the tight interaction between experiments, modeling, and antibody engineering.
GeneGo's (www.genego.com) MetaCore is a software platform that can link high throughput data to signaling pathways. "The idea is to present several case studies to show how we can use our system to cross-validate different types of data," states Tatiana Nikolskaya, Ph.D., CSO of GeneGo.
Each high throughput method provides researchers with an idea of the real cellular processes. Different methods provide different data. MetaCore can align that information and compare it. "Such cross-validation of different HT data allows us to reconstruct cellular processes more accurately and provides better drug targets," explains Dr. Nikolskaya.
She says the firm's biggest challenge now is to gain access to different types of data created on the same data set, especially in proteomics and metabolomics where there is not a lot of data available in the public domain. However, she adds that the company does have metabolomics data available in a separate network, which will soon be merged with their signaling network.
Many signaling cascades have an effect on metabolic levels when they turn off and on certain metabolic subsystems. Dr. Nikolskaya says they are hoping to be able to show cascades from the receptor through kinases to transcriptional factors and then to particular metabolic pathways.
Currently, the software's main application is mining of microarray data. "It's nice to have the same system across different departments because it's not enough to have integrated data and store and access them; it's very important to be able to mine them at the same time."