Modeling microbiome modulation of tumor metabolic networks to predict synergistic therapies

Area of Science:

• Computational biology
• Microbiome research
• Cancer therapeutics

Background:

• Microbiome composition significantly impacts drug response, but modeling microbial metabolic effects on host cells and therapies is challenging.
• Existing methods lack the ability to predict combination therapies considering microbial influence in diseases like colorectal cancer (CRC).

Purpose of the Study:

• To develop and validate a microbiome-aware computational framework for predicting synergistic drug combinations in CRC.
• To investigate the metabolic impact of specific microbes, such as Fusobacterium nucleatum (Fn), on drug synergy.

Main Methods:

• Combined machine learning with genome-scale metabolic models to analyze 6,514 drug combination profiles in CRC cell lines.
• Predicted synergistic drug combinations in both microbe-free and microbe-associated contexts.
• Validated predictions through in vitro assays, co-culture experiments, and metabolic pathway analysis.

Main Results:

• The framework successfully predicted synergistic drug combinations, outperforming existing methods in microbe-associated contexts.
• Identified phosphoinositol metabolism and cysteine transport as critical determinants of Fn-dependent drug synergy.
• Demonstrated the framework's scalability for discovering microbiome-dependent therapies.

Conclusions:

• This work presents a scalable computational strategy for discovering microbiome-dependent combination therapies for colorectal cancer.
• The findings highlight the importance of integrating microbiome data into drug discovery and development pipelines.
• The framework supports the development of personalized cancer treatments considering individual microbiome profiles.