Researchers found that even in otherwise healthy adults, fatigue was associated with gut microbial shifts, lower energy-related metabolites, and patterns that overlapped most strongly with ME/CFS and psychiatric disorder datasets.
Study: Fatigue-associated gut bacteria in Japanese healthy adults characterized by metagenomic analysis. Image Credit: Sinhyu Photographer / Shutterstock
In a recent study published in the journal Scientific Reports, researchers examined associations between the gut microbiome, fecal metabolites, and fatigue in Japanese adults.
Dysbiosis, i.e., gut microbiome dysregulation, has been observed in multiple diseases, including schizophrenia, depression, and ME/CFS. Since dysbiosis is frequently observed in people with psychiatric disorders, changes in the gut microbiome may plausibly occur in the pre-disease stage. As such, exploring gut microbial features in otherwise healthy people with fatigue may help develop future prevention and risk-stratification strategies.
The study and findings
In the present study, researchers investigated associations between the gut microbiome and fatigue. They recruited 50 otherwise healthy Japanese adult employees between 2019 and 2021; participants completed the Chalder Fatigue Questionnaire (CFQ), the Epworth Sleepiness Scale (ESS), and the Pittsburgh Sleep Quality Index (PSQI). Subjects self-collected fecal samples for whole-genome shotgun sequencing, organic acid analysis, and metabolomic analysis.
Organic acids in fecal samples were quantified using high-performance liquid chromatography. Species-level taxonomic composition was determined. Taxonomic profiling was performed at species and genus levels. Alpha diversity was evaluated using species richness and the Simpson, Shannon, and Pielou’s evenness indices. In addition, a principal coordinate analysis (PCoA) based on Bray-Curtis dissimilarity was performed to compare the non-fatigue and fatigue groups.
In total, 16 participants had fatigue (CFQ ≥ 17) and showed higher ESS and PSQI scores than the non-fatigue group. Body mass index was significantly different between the fatigue and non-fatigue groups and was treated as a potential confounder in subsequent multivariable analyses. There were no significant differences in alpha diversity or standard PCoA-based beta diversity between groups, although pairwise Bray-Curtis analysis showed greater inter-individual microbiome dissimilarity within the fatigue group. Taxonomic profiling identified 945 species, 405 genera, and 15 phyla across all fecal samples. The fatigue group had significantly greater abundance of six genera than the non-fatigue group.
The non-fatigue group had 11 genera that were significantly more abundant than in the fatigue group. Further, 10 species in the fatigue group and 17 in the non-fatigue group were significantly more abundant. Metabolomic analysis identified 110 metabolites and 11 organic acids in fecal samples. The fatigue group had significantly lower levels of citrate and adenosine and higher levels of tyramine and gamma-aminobutyric acid (GABA).
Next, the researchers investigated whether changes in microbial functional potential accompanied these differences in microbial metabolites and composition. To this end, functional profiling was performed using Kyoto Encyclopedia of Genes and Genomes (KEGG) Orthologs (KOs). These results indicate altered microbial functional potential, not direct evidence of increased oxidative stress, enzyme activity, or metabolic flux. This identified 8,739 KOs across all samples, of which 452 and 197 showed significant enrichment in the fatigue and non-fatigue groups, respectively.
Furthermore, Random Forest (RF) classification was performed using significantly altered KOs, genera, and species to explore whether these microbial characteristics contained discriminatory information about fatigue status. KO-based models had high discriminatory ability across repeated analyses, with a median out-of-bag area under the receiver operating characteristic curve (AUROC) of 0.972. However, held-out test-set performance was lower and variable, so the authors interpreted these findings as exploratory rather than as validation of a predictive classifier.
Enrichment analysis revealed that modules involved in urate production, 2-oxoglutarate synthesis and metabolism, and the urea metabolism pathway were upregulated in the fatigue group. In contrast, modules related to the Wood-Ljungdahl pathway, F-type ATPase, and glycogen and trehalose biosynthesis were downregulated. In addition, multiple KOs involved in the oxidative stress response showed significant upregulation in the fatigue group.
Next, to link fatigue-related functional shifts to microbial genomes, metagenome-assembled genomes (MAGs) were constructed, and their contributions to KO signatures were evaluated. This resulted in 1,351 MAGs; following deduplication, 332 representative MAGs were retained. Comparative analysis indicated that 23 and five MAGs were significantly enriched in non-fatigue and fatigue groups, respectively.
In addition, functional annotation for MAGs was performed using KOs. Overall, the researchers identified 62 fatigue-depleted KOs from MAGs depleted in the fatigue group and 57 fatigue-enriched KOs from those enriched in the fatigue group. Next, they explored correlations between the relative abundance of five selected MAGs (with the most significantly altered KEGG modules) and fecal levels of metabolites that were significantly different between groups.
The relative abundance of Fusicatenibacter saccharivorans and Hominisplanchenecus faecis was significantly positively correlated with citrate levels. Escherichia coli abundance was also positively correlated with tyramine and GABA levels. Meanwhile, the abundance of Blautia_A obeum and Oliverpabstia faecicola was significantly negatively correlated with GABA levels.
Finally, the team examined these fatigue-associated MAGs in external cohorts across conditions in which fatigue is a contributing or primary symptom, including ME/CFS, major depressive disorder (MDD), impaired glucose tolerance (IGT), bipolar disorder (BD), and obesity. The 28 MAGs that were significantly different in the fatigue group were all detected in IGT, BD, MDD, and obesity datasets, while 17 and 22 were detected in two ME/CFS datasets. The five representative MAGs were also detected in the external cohorts, except for E. coli and Blautia_A obeum in one ME/CFS dataset.
However, concordant directional and statistically significant overlap was strongest in ME/CFS cohorts, followed by MDD and BD. No concordant MAGs were identified in obesity or IGT cohorts, and the authors cautioned that cross-cohort comparisons may be affected by differences in participant characteristics and study protocols.
Conclusions
In sum, subjective fatigue in healthy adults was associated with marked shifts in the functional composition of the gut microbiome. The overlap of fatigue-related microbial alterations with those in psychiatric disorders and ME/CFS underscores the potential relevance of gut microbial signatures in fatigue-associated biological states.
Overall, the study results provide a foundation for future studies into whether fatigue-linked gut microbial signatures could support early detection or preventive intervention strategies, although the small, cross-sectional study cannot establish causality or clinical utility.
Download your PDF copy by clicking here.
Researchers urge one health approach to control raccoon-borne water contamination