Merging Metabolic Modeling and Imaging for Screening Therapeutic Targets in Colorectal Cancer

Table of Contents

• Introduction
• Authors
• Prerequisites
• Files

Introduction

Cancer-associated fibroblasts (CAFs) play a key role in metabolic reprogramming and are well-established contributors to drug resistance in colorectal cancer (CRC). To exploit this metabolic crosstalk, we integrated a systems biology approach that identified key metabolic targets in a data-driven method and validated them experimentally. This process involved high-throughput computational screening to investigate the effects of enzyme perturbations predicted by a computational model of CRC metabolism to understand system-wide effects efficiently. Our results highlighted hexokinase (HK) as one of the crucial targets, which subsequently became our focus for experimental validation using patient-derived tumor organoids (PDTOs). Through metabolic imaging and viability assays, we found that PDTOs cultured in CAF conditioned media exhibited increased sensitivity to HK inhibition. Our approach emphasizes the critical role of integrating computational and experimental techniques in exploring and exploiting CRC-CAF crosstalk.

Authors

Niki Tavakoli¹⁺, Emma J. Fong²⁺, Abigail Coleman², Yu-Kai Huang², Mathias Bigger^2,3, Michael E. Doche², Seungil Kim², Heinz-Josef Lenz⁶, Nicholas A. Graham³, Paul Macklin⁴, Stacey D. Finley^1,3,5,*, Shannon M. Mumenthaler^1,2,6,*

¹Alfred E. Mann Department of Biomedical Engineering, University of Southern California, Los Angeles, CA, 90089, USA
²Ellison Institute of Technology, Los Angeles, CA, 90064, USA
³Mork Family Department of Chemical Engineering and Materials Science, University of Southern California, Los Angeles, CA, 90089, USA
⁴Department of Intelligent Systems Engineering, Indiana University, Bloomington, IN, 46202, USA
⁵Department of Quantitative and Computational Biology, University of Southern California, Los Angeles, CA, 90089, USA
⁶Division of Medical Oncology, Norris Comprehensive Cancer Center, University of Southern California, Los Angeles, CA, 90033, USA
⁺Denotes co-first authors
^*Co-correspondence: [email protected] (S.D.F.), [email protected] (S.M.)

Prerequisites

Before you begin, ensure you have met the following requirements.

• MATLAB: You need to have MATLAB installed on your machine. For installation details, see the official MATLAB website.
• COBRA Toolbox: You also need to have the COBRA Toolbox installed in MATLAB. Installation instructions can be found on the COBRA Toolbox GitHub page.
• Python: You need to have Python installed on your machine. You can download it from the official Python website.
• TensorFlow: After installing Python, you need to install TensorFlow. You can install TensorFlow via pip with the following command:

  pip install tensorflow
  

Files

Enzyme Knockdowns

• 📁 enzyme_knockdowns : contains all of the code for performing partial enzyme knockdowns for all cell conditions

  • 📁 baselineCode : contains the code developed by Wang et al. (2022), which employs upFBA to predict the network's flux distributions based on the constraints outlined in their study.
  • 📁 baselineCode : contains the outputted data from run_upFBA_model.m in baselineCode.

• 📜 gene_knockdowns.m : a script that sets up the partial enzyme knockdowns for all cell conditions.

• 📜 main_gene_knockdowns.m : a script that executes gene_knockdowns.m.

---

Representation Learning

• 📁 representation_learning : contains all of the code for performing representation learning on the the data from 📜 gene_knockdowns.m.
  • 📜 trainNN.py : a script that loads in the data, defines the neural network architecture, specifies the loss function and optimizer, and runs the training loop where the model learns from the training data.
  • 📜 genNN.py : a script that generates outputs using the trained neural network.
  • 📜 data_generator.py : a script designed to generate / preprocess data.