NEW YORK – An international team led by scientists at China's Shanghai Jiao Tong University has identified two "competing guilds" of gut bacteria — groups of microbes that appear to work together — that may comprise a new biomarker of human health.
The study's results, published in Cell on Monday, also suggest that this "two competing guilds" (TCG) model might be used as an indicator of health and to predict response to therapy.
The study analyzed approximately 4,000 metagenomic samples from 38 studies, covering 15 diseases.
To identify the kinds of stable relationships that might comprise core microbiome components, the researchers analyzed data from an interventional study, in which participants with type 2 diabetes (T2D) were randomized to receive either a high-fiber diet or standard of care over the course of three months, followed by one year of follow-up.
Individuals on the high-fiber diet showed "significant" alterations in gut microbe structure over the three months of intervention, as measured by shotgun metagenomic sequencing from fecal samples. These alterations reverted to baseline by the end of the follow-up period, suggesting highly resilient microbial community structures.
The investigators searched their sequencing data for stably correlated bacterial genomes and identified 635 such correlations involving 184 high-quality metagenome-assembled genomes. These formed two microbial guilds that were linked through negative correlations, indicating a potentially competitive relationship.
One guild appeared to be comprised entirely of species within the phylum Firmicutes, whereas the other was composed of five different phyla, namely Firmicutes, Bacteroidota, Proteobacteria, Actinobacteriota, and Fusobacteriota.
Among patients placed on the high-fiber diet, the single-phylum guild (guild 1) increased in abundance over the study period, while the five-phylum guild (guild 2) decreased. Abundances of both groups remained unchanged in the control arm of patients.
The changes in abundance appeared to correlate tightly with clinical parameters such as adiponectin and leptin levels, body weight, and HbA1c levels.
The researchers also found pairs of competing microbial groups within stable clusters of genomes across 10 independent metagenomic case-control datasets spanning six distinct disorders. These comprised atherosclerotic cardiovascular disease, ankylosing spondylitis, colorectal cancer, inflammatory bowel disease, liver cirrhosis, and schizophrenia.
Since previous studies have linked gut microbiota composition to therapy response, the researchers evaluated microbiome differences in responders and non-responders from 11 pretreatment metagenomic datasets related to inflammatory bowel disease, rheumatoid arthritis, advanced melanoma, and B cell melanoma. In all cases, they found pairs of competing guilds that predicted outcomes such as symptom improvement, overall response rate, and progression-free survival.
The authors likened the microbial communities of the TCG model to housekeeping genes, writing that in the same way that housekeeping genes are ubiquitously expressed and are essential for maintaining basic cellular functions, the core members identified in the TCG model represent foundational components of the gut microbiome and are critical for maintaining its overall stability and functionality.
"These results underscore the potential of stably correlated [high-quality metagenome-assembled genomes] as a more precise biomarker in disease classification and prediction, setting a new standard for microbiome research," the authors wrote. "Ultimately, the consistent core structure of TCGs that our research highlights may propel forward the fields of proactive and preventive healthcare, marking a significant stride in microbiome studies and their clinical ramifications."