Genome-scale modeling of the influence of microbiota-derived butyrate on the regulation of human metabolism by the histone deacetylase sirtuin 1.

BACKGROUND

Genome-scale metabolic models predict metabolic flux distributions but typically lack explicit transcriptional regulation, limiting their ability to simulate graded effects of epigenetic modulators such as Sirtuin1.

OBJECTIVES

This study aims to develop and validate a continuous regulatory-metabolic framework integrating sirtuin 1-dependent transcriptional control into human genome-scale metabolism and to quantify the metabolic impact of microbiome-derived butyrate in intestinal epithelial cells.

METHODS

A curated Sirtuin1-centered regulatory network comprising 8 transcriptional regulators, 487 metabolic genes, and 2296 reactions (∼22% of Recon3D) was integrated into the Recon3D reconstruction to generate iSirtuin1_HumanMet. Continuous regulatory logic was implemented within steady-state regulatory flux balance analysis. Tissue-specific models were derived from genotype-tissue expression transcriptomic data using FASTCORE. Human Caco-2 intestinal epithelial cells were treated with 0 to 9 mM sodium butyrate for 72 h. Sirtuin1 protein expression was quantified by Western blot and modeled using an inverse exponential regression (R2 = 0.669). Predicted maximal intracellular production capacities were compared with independent metabolomics data using Spearman correlation.

RESULTS

Simulated Sirtuin1 activation (0.0-1.0) modulated 2296 reactions, with 34.2% of upregulated reactions belonging to fatty acid oxidation. Increasing Sirtuin1 promoted gluconeogenesis and lipid utilization while repressing glycolysis and nucleotide interconversion. Tissue-specific simulations across 54 tissues revealed distinct clustering of metabolic responses. Incorporation of experimentally derived butyrate-Sirtuin1 inhibition resulted in concordant monotonic trends between predicted and measured intracellular metabolites for 10 of 12 metabolites (83%), with Spearman ρ ranging from -0.64 to 0.85 (median ρ ≈ 0.69). Integration of microbiome-predicted butyrate fluxes showed strong host metabolic associations, including correlations as strong as ρ = -0.92 (P = 8.77 × 10-22).

CONCLUSIONS

In Caco-2 intestinal epithelial cells and tissue-specific human metabolic models, continuous integration of Sirtuin1 regulation enables quantitative simulation of graded transcriptional control and microbiome-derived metabolic modulation, providing a systems-level framework to study diet-microbiome-host metabolic interactions.