Modeling Microbiome Modulation of Tumor Metabolic Networks to Predict Synergistic Therapies

Area of Science:

• Oncology
• Microbiome Research
• Computational Biology

Background:

• Microbiome composition significantly impacts drug efficacy, but modeling the complex interactions between tumors, microbes, and therapies is challenging.
• Existing methods for predicting combination therapies in colorectal cancer (CRC) often overlook the influence of microbial communities.

Purpose of the Study:

• To develop and validate a generalizable computational framework for predicting synergistic drug combinations in CRC.
• To incorporate the metabolic interplay of microbes, such as Fusobacterium nucleatum (Fn), into therapeutic strategy development.
• To identify context-specific combination therapies tailored to individual microbiome profiles.

Main Methods:

• A machine-learning model was trained on 6,514 drug combinations in microbe-free CRC cell lines.
• Genome-scale metabolic modeling was integrated with machine learning to predict synergistic drug pairs.
• An asymmetric co-culture system simulating the colon's oxygen gradient was used for experimental validation.
• Mechanistic analyses, including pharmacological perturbations, were performed to understand drug synergy determinants.

Main Results:

• The computational framework successfully predicted synergistic drug combinations in both microbe-free and microbe-associated CRC models.
• Predictions generalized to immunotherapy-associated conditions, demonstrating broad applicability.
• Experimental validation confirmed synergistic combinations in HCT116 cells with Fn, including novel therapeutic agents.
• Phospho-inositol metabolism and cysteine transport were identified as key mechanisms driving Fn-dependent drug synergy.

Conclusions:

• The developed framework offers a scalable and microbiome-aware approach to discover context-specific combination therapies for colorectal cancer.
• This work advances the understanding of how microbial metabolites influence drug response and therapeutic outcomes.
• The findings pave the way for more personalized and effective treatment strategies in oncology by integrating microbiome data.