The gut microbiota is crucial for human health, functioning as a complex adaptive system akin to a vital organ. To identify core health-relevant gut microbes, we followed the systems biology tenet that stable relationships signify core components. By analyzing metagenomic datasets from a high-fiber dietary intervention in type 2 diabetes and 26 case-control studies across 15 diseases, we identified a set of stably correlated genome pairs within co-abundance networks perturbed by dietary interventions and diseases. These genomes formed a "two competing guilds" (TCGs) model, with one guild specialized in fiber fermentation and butyrate production and the other characterized by virulence and antibiotic resistance. Our random forest models successfully distinguished cases from controls across multiple diseases and predicted immunotherapy outcomes through the use of these genomes. Our guild-based approach, which is genome specific, database independent, and interaction focused, identifies a core microbiome signature that serves as a holistic health indicator and a potential common target for health enhancement.

The incidence of metabolic syndrome is significantly higher in patients with irritable bowel syndrome (IBS), but the mechanisms involved remain unclear. Gut microbiota is causatively linked with the development of both metabolic dysfunctions and gastrointestinal disorders, thus gut dysbiosis in IBS may contribute to the development of metabolic syndrome. Here, we show that human gut bacterium Ruminococcus gnavus-derived tryptamine and phenethylamine play a pathogenic role in gut dysbiosis-induced insulin resistance in type 2 diabetes (T2D) and IBS. We show levels of R. gnavus, tryptamine, and phenethylamine are positively associated with insulin resistance in T2D patients and IBS patients. Monoassociation of R. gnavus impairs insulin sensitivity and glucose control in germ-free mice. Mechanistically, treatment of R. gnavus-derived metabolites tryptamine and phenethylamine directly impair insulin signaling in major metabolic tissues of healthy mice and monkeys and this effect is mediated by the trace amine-associated receptor 1 (TAAR1)-extracellular signal-regulated kinase (ERK) signaling axis. Our findings suggest a causal role for tryptamine/phenethylamine-producers in the development of insulin resistance, provide molecular mechanisms for the increased prevalence of metabolic syndrome in IBS, and highlight the TAAR1 signaling axis as a potential therapeutic target for the management of metabolic syndrome induced by gut dysbiosis.

Green tea extracts and tea catechins have been shown to prevent or alleviate diabetes. The present study tests the hypothesis that green tea leaves in powder form (GTP), which also contain fiber and other water non-extractable materials, are more effective than the corresponding green tea extracts (GTE) in impeding the development of diabetes in db/db mice. Female db/db mice were treated with a diet containing 1% of GTE, 2% of GTE, 2% of GTP (with the same catechin content as 1% GTE) or 1% GTP. The 1% GTE group had lower food intake, water consumption, body weight and fasting blood glucose levels than the control group, while 2% GTP did not have any significant effect. Dietary 1% GTE also preserved β-cell insulin secretion. However, 1% GTP increased food intake, water consumption and blood glucose levels. Microbiome analysis with 16S rRNA gene V4 sequencing showed that the gut microbiota was modified by GTE and GTP, and a few bacterial guilds were associated with blood glucose levels. In the Random Forest regression model, the leading predictor of metabolic outcome was food consumption, followed by changes in some bacterial guilds. The results illustrate the importance of food consumption and gut microbiota in affecting the progression of diabetes.

Current microbiome signatures for chronic diseases such as diabetic kidney disease (DKD) are mainly based on low-resolution taxa such as genus or phyla and are often inconsistent among studies. In microbial ecosystems, bacterial functions are strain specific, and taxonomically different bacteria tend to form co-abundance functional groups called guilds. Here, we identified guild-level signatures for DKD by performing in-depth metagenomic sequencing and conducting genome-centric and guild-based analysis on fecal samples from 116 DKD patients and 91 healthy subjects. Redundancy analysis on 1,543 high-quality metagenome-assembled genomes (HQMAGs) identified 54 HQMAGs that were differentially distributed among the young healthy control group, elderly healthy control group, early-stage DKD patients (EDG), and late-stage DKD patients (LDG). Co-abundance network analysis classified the 54 HQMAGs into two guilds. Compared to guild 2, guild 1 contained more short-chain fatty acid biosynthesis genes and fewer genes encoding uremic toxin indole biosynthesis, antibiotic resistance, and virulence factors. Guild indices, derived from the total abundance of guild members and their diversity, delineated DKD patients from healthy subjects and between different severities of DKD. Age-adjusted partial Spearman correlation analysis showed that the guild indices were correlated with DKD disease progression and with risk indicators of poor prognosis. We further validated that the random forest classification model established with the 54 HQMAGs was also applicable for classifying patients with end-stage renal disease and healthy subjects in an independent data set. Therefore, this genome-level, guild-based microbial analysis strategy may identify DKD patients with different severity at an earlier stage to guide clinical interventions. IMPORTANCE: Traditionally, microbiome research has been constrained by the reliance on taxonomic classifications that may not reflect the functional dynamics or the ecological interactions within microbial communities. By transcending these limitations with a genome-centric and guild-based analysis, our study sheds light on the intricate and specific interactions between microbial strains and diabetic kidney disease (DKD). We have unveiled two distinct microbial guilds with opposite influences on host health, which may redefine our understanding of microbial contributions to disease progression. The implications of our findings extend beyond mere association, providing potential pathways for intervention and opening new avenues for patient stratification in clinical settings. This work paves the way for a paradigm shift in microbiome research in DKD and potentially other chronic kidney diseases, from a focus on taxonomy to a more nuanced view of microbial ecology and function that is more closely aligned with clinical outcomes.