20.109(F13): Mod 2 Day 1 System Design

Introduction

Today we will begin our journey through Systems Biology and the careful consideration of experimental and computational design that is required for large-scale biological engineering studies. At the end of lab today, you will choose two chemical inhibitors that target the Epidermal Growth Factor Receptor (EGFR) signaling network with the goal of overcoming a growing common problem in the treatment of disease -- drug resistance.

Part 1: Simulation of the EGFR network

The first mathematical model published in the journal Cell described the binding of EGF to EGFR. That paper ushered in an era of multidisciplinary research which brought together scientists and engineers to build mathematical models of intracellular signaling pathways. In fact, the field of Systems Biology, often thought of as the use of computational models to study multiple signaling pathways across variable time and length scales, grew directly from these first collaborations.

There are many approaches to computational modeling, each aiming to understand how information travels through a signaling network to ultimately produce a cell phenotype. We will discuss some of the various modeling techniques in lecture. In lab you will design and perform a high-throughput experiment that alters cell proliferation, a cell phenotype that is often dysregulated in cancer. There are several distinct signaling pathways downstream of EGFR activation that influence cell proliferation and viability. The best understood pathways are the MEK/Erk pathway, PI3K/Akt pathway, and the STAT transcriptional regulation pathway. One important tool that we can take advantage of as biological engineers is the wealth of mathematical models describing the dynamics of these pathways upon EGF stimulation. It is good practice to gather as much data as possible before designing an experiment -- therefore, we will use computational simulation to better understand how chemical inhibition will alter information flow through the MEK/Erk, PI3K/Akt, or STAT pathways.

In 2012, Bidkhori et al. employed an ordinary differential equation (ODE)-based mathematical model to study the effect of a common mutant EGFR protein on the pathways that control cell proliferation/viability. Today we will utilize two web-based programs, CellDesigner and COPASI, that facilitate building and visualizing cell signaling models to interogate the Bidkhori model. In fact, the computational study of cell signaling has become such an useful tool that a standardized programing language SBML has been developed. Concomitant with the development of SBML, a number of user friendly GUIs have popped up to make simulation of models accessible to everyone -- including the two that we use today.

Figure 1 from Bidkhori et al. graphically illustrating the signaling cascades used in their modeling scheme. Bidkhori G, Moeini A, Madoudi-Nejad A (2012) Modeling of Tumor Progression in NSCLC and Intrinsic Resistance to TKI in Loss of PTEN Expression. PLoS ONE 7(10)

First, you will find the Bidkhori model on the BioModels database. Next, you will visualize the network structure using CellDesigner. Finally, you'll simulate the network to get a feel for the information flux through the MEK/Erk, PI3K/Akt, and STAT3 pathways using COPASI. After your brain is full of EGFR signaling, you'll choose a combination of two inhibitors in an attempt to curb signaling and kill the CHO-EGFR cells, a model of ovarian cancer.

Import the SBML code

The BioModels database is a very large compendium of biochemical models that are for public use. The curators of the database accept SBML code from model builders and then validate that the original figures published with the model can be recapitulated using the submitted code. The Systems Biology community has tried very hard to make their computational models available to everyone, and you'll find that many engineers will start with a published model and then build from there. This is not only an acceptable practice, but an encouraged one -- with two important rules: 1. You must understand the output and design of the model that you start with and 2. The authors of the original model are always cited in your work. Note: copying models to use in an undergraduate class is generally not acceptable.

1. Go to the BioModels database.
2. In the search box enter 'EGFR'.
   • You should find a list of 22 'curated' models.
   • Scroll all the way down to the bottom of that list and choose BIOMD0000000452 Bidkhori2012 - normal EGFR signaling
     • Note: take care not to scroll through the models that have not been curated (or validated).
3. Click the disk icon to save the models SBML code to your computer.
4. Now click on the BioModel database number associated with the model.
   • In this window you will find information about the paper the model was described in, other models that the current one is derived from, and its creation date.
5. Navigate to the Curation tab.
   • Here is the information provided by the BioModels curators as to how the model was validated. You'll see that the author of this tab also used COPASI to run the model.
   • If you maximize the image of the plots, you can get an idea for what type of output your can expect from your simulation.
6. Double check that you've downloaded the SBML file to your computer. If you have, you're good to move on to the next section.

Open the model in CellDesigner

1. Open CellDesigner
2. Use the Open command under File to open the SBML code that you downloaded form the BioModels database.
3. Next, clean up the CellDesigner workspace a bit to make things easier to view:
   1. Maximize the CellDesigner window.
   2. Under View:
      • unselect Show Reaction Id
      • under Change Toolbar Visible, select Hide all
4. Click the Proteins near the bottom of the window. How many proteins and protein-complexes are described by this model?
5. Click Reactions. How many different ODEs are used the describe this biochemical system?
6. Now, minimize the bottom panel using the small, gray arrows in the upper left hand corner.
7. The components of the EGFR network and their complexed states (for example EGF + EGFR --> EGF-EGFR) are illustrated in the large window in the middle of the screen. The number of model components makes it very difficult to view the structure of the model. To see a quick view of the system, under View choose Zoom Fit.
8. The 'hairball' view of the model includes arrows (edges) and boxes (nodes) that represent the different biochemical reactions of the system. To better organize the model for ease of visualizing the information flow, go to Layout and choose Hierarchic Layout.