Metabolic pathway and network analysis integration for discovering the biomarkers in pig feces after a controlled fruit-vegetable dietary intervention

Authors: LIU, ZHIHAO - University Of Maryland; Solano-Aguilar, Gloria; Lakshman, Sukla; Urban, Joseph; ZHANG, MENGLIANG - Middle Tennessee State University; Chen, Pei; YU, LIANGLI - University Of Maryland; Sun, Jianghao

Submitted to: Food Chemistry
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/9/2024
Publication Date: 8/10/2024
Citation: Liu, Z., Solano Aguilar, G., Lakshman, S., Urban, J.F., Zhang, M., Chen, P., Yu, L., Sun, J. 2024. Metabolic pathway and network analysis integration for discovering the biomarkers in pig feces after a controlled fruit-vegetable dietary intervention. Food Chemistry. 461:140836. https://doi.org/10.1016/j.foodchem.2024.140836.
DOI: https://doi.org/10.1016/j.foodchem.2024.140836

Interpretive Summary: Exploring the link between diet and health has obtained significant attention. Among dietary components, the consumption of fruits and vegetables has emerged as a focal point due to their enrich and diverse phytochemical composition. Dietary assessments are traditionally measured by dietary recalls, food records, self-reported intake questionnaires, which have various limitations. Recent advances in nutritional metabolomics allow for more objective assessment of dietary exposures, which have been proposed as an alternative or complement to traditional methods such as self-reporting of food intake. The purpose of this study was to identify potential dietary biomarkers from pig fecal samples with fruits and vegetables (FV) intervention . The mixed FV diet selected for this investigation included spinach, apple, grape, green bean, celery, strawberry, kale, broccoli, blackberry, and blueberry, which represent a diverse range of flavors, colors, and nutritional profiles. However, it is a challenge to provide intuitive information on thousands of metabolites in a single metabolomics approach, especially with the overwhelming complexity of the current FV intervention study. To overcome the challenge, we integrated liquid chromatography-high resolution mass spectrometry(LC-HRMS), multivariate statistical analysis, metabolic pathway prediction and network exploration for the discovery of biomarkers in pigs with FV intervention. The integration of data-driven and knowledge-based analytical methods enable the exploration of interconnected metabolic pathways and potential interactions between identified biomarkers. Based on the strategy, we successfully monitored a significant number of potential biomarkers, which are related to the intake of the diet enriched in flavanols, flavones, flavan-3-ols, and anthocyanins, etc. The biomarkers identified were generally not detected as isolated entities, most of which are as part of larger sets of metabolites of the same chemical classes, thus further increasing the reliability of the identification. Overall, the identified biomarker in our study could be applied to evaluate the efficiency of FV interventions for maintaining and improving health at the individual level.

Technical Abstract: This study aimed to establish a strategy to identify dietary intake biomarkers using a non-targeted metabolomic approach, including metabolic pathway and network analysis. The strategy was successfully applied to identify dietary intake biomarkers in fecal samples derived from pigs fed two doses of a polyphenol-rich fruit and vegetable (FV) diet following the Dietary Guidelines for Americans (DGA) recommendations. Potential biomarkers were differentially detected among dietary treatment groups using liquid chromatography-high resolution mass spectrometry (LC-HRMS) based on a non-targeted metabolomic approach, and metabolic pathway and network analysis were performed for final metabolite identification. The principal component analysis (PCA) results showed significant differences between the fecal metabolite profiles in the control and two FV intervention groups, indicating a diet-induced differential fecal metabolite profile after FV intervention. Metabolites from common flavonoids, e.g., (epi)catechin and protocatechuic acid, or unique flavonoids, e.g., 5,3',4'-trihydroxy-3-methoxy-6,7-methylenedioxyflavone and 3,5,3',4'-tetrahydroxy-6,7-methylenedioxyflavone, were identified as highly discriminating factors and confirmed their potential as fecal markers for the FV diet intervention. The microbiota pathway prediction using targeted flavonoids provided invaluable and reliable biomarker exploration with high confidence. A correlation network analysis between these discriminatory ion features was applied to find connections to possible dietary biomarkers, further validating these biomarkers with biochemical insights. This study demonstrates that integrating metabolic pathways and network analysis with a non-targeted metabolomic approach is highly effective for rapid and accurate identification and prediction of fecal biomarkers under controlled dietary conditions in animal studies. The approach can also be utilized for microbial metabolisms and human clinical research.