In this project, we identified early microbiome and metabolites markers for NAFLD. Using this information and integrating it with clinical metadata, we generated a model that can assess the risk of a patient developing NAFLD. Our model, which is based on machine learning and taking into consideration microbiome features, outperforms existing clinical screening tools. These results serve as a proof-of-concept for developing highly accurate and explanatory AI-tools for NAFLD diagnosis.
We also conducted the first study characterizing metabolic changes in NAFLD patients due to physical exercise in various sample matrices. We demonstrated that a 12-week high-intensity interval training exercise program improves the condition of NAFLD patients on a whole-body level, and this is achieved even without requiring dietary changes and without significant weight loss. Exercise decreased blood glucose levels and increased exercise parameters that protect against hyperglycemia. We observed several metabolomic signatures indicating that NAFLD patients doing exercise have better use and metabolism of glucose, but also different metabolism of amino acids, lipids, and bile acids.
In a separate study, we followed NAFLD patients who consumed a normal diet or a resistant starch (a prebiotic nondigestible fiber that feeds beneficial bacteria in the gut) for 4 months. We found that the dietary switch to resistant starch consumption alters gut microbiota and metabolites and improves clinical parameters. In NAFLD patients we found that one species, Bacteroides stercoris, is associated with poor liver status and performance, and the change in the diet improves all these parameters and decreases the presence of Bacteroides stercoris in the gut. Moreover, we could prove that Bacteroides stercoris promotes NAFLD progression in a mouse model of the disease.
One of the aims of BestTreat was to screen and select beneficial bacterial solutions for the treatment of NAFLD. We screened a broad group of bacterial strains for traits with potential benefits in NAFLD in vitro and selected two strains for further analysis in animal models.If the experiments in animals prove positive, we plan to scale up production for use in dietary supplements, infant formulae, or pharmaceutical products.
We also explored engineered microbial strains for therapeutic use. We engineered the probiotic Escherichia coli Nissle 1917 (EcN) to express Aldafermin, a potential treatment option for patients with NAFLD. We tested the beneficial role of these strains in a comprehensive study in mice, analyzing anatomical, histological, and clinical features, as well as transcriptomics and metabolomic profiles. We found that the administration of these engineered strains has numerous beneficial effects in a mouse model of NAFLD. Our results suggest the potential efficacy of engineered EcN strains in combination with dietary changes as a novel non-invasive strategy to treat NAFLD condition, with clear benefits in restoring lipid metabolism and AA dysfunctions commonly observed in NAFLD patients.
To ensure the dissemination of our action, all publications arising from the BestTreat program are open-access and can be obtained from our Website (www.besttreat.eu).