Millions of people with chronic fatigue syndrome (CFS), a debilitating condition often overlooked due to a lack of diagnostic tools, may be closer to personalized care, according to new research showing how the disease disrupts interactions between microbiota, the immune system and metabolism. CFS is characterized by severe symptoms that significantly impair physical and mental wellbeing, including persistent fatigue, sleep disturbances, dizziness, and chronic pain. Experts often compare CFS to COVID-19, as both conditions often result from viral infections, such as the EBV. In the United States, CFS affects between 800,000 and 3 million individuals (many undiagnosed) and costs the economy between $18 and $51 billion annually due to healthcare costs and lost productivity, according to the CDCs.
Previous studies have found immune system alterations in ME/CFS. This new research builds on these findings, investigating the interaction between the gut microbiome, its metabolites and immune responses. The team linked these connections to 12 patient-reported symptom classes, aggregated from hundreds of data points generated by patient health and lifestyle surveys. These include sleep disturbances, headache, fatigue, dizziness and other symptoms, which the researchers mapped comprehensively, from microbiome alterations to metabolites, immune responses and clinical symptoms. The findings come from data on 249 individuals analyzed using a novel artificial intelligence (AI) platform that identifies disease biomarkers from stool, blood and other routine laboratory tests.
To conduct the study, researchers analyzed comprehensive data collected by the Bateman Horne Center, a leading research center for CFS, long-term COVID-19, and fibromyalgia in Salt Lake City, Utah. Scientists developed a deep neural network model called BioMapAI. The tool integrates gut metagenomics, plasma metabolomics, immune cell profiles, blood test data, and clinical symptoms from 153 patients and 96 healthy individuals over four years. Immune cell analysis proved most accurate in predicting symptom severity, while microbiome data best predicted gastrointestinal, emotional, and sleep disorders. The model linked thousands of patient data points, reconstructing symptoms such as pain and gastrointestinal problems, among many others.
It also revealed that patients with the disease for less than four years had fewer disrupted networks than those with the disease for more than ten years. The study included 96 healthy controls matched for age and sex, demonstrating balanced interactions between the microbiome, metabolites and the immune system, in contrast to the significant disruptions in CFS patients related to fatigue, pain, emotional regulation and sleep disturbances. CFS patients also had lower levels of butyrate along with other essential nutrients, inflammation control and cellular energy. Patients with elevated levels of tryptophan, benzoate and other markers indicated a microbial imbalance. Increased inflammatory responses were also observed, particularly affecting MAIT cells, which are sensitive to gut microbiota health.
MAIT cells link gut health to broader immune functions, and their disruption, along with the butyrate and tryptophan pathways (normally anti-inflammatory), suggests a deep imbalance. Although the findings require further validation, they significantly advance scientists’ understanding of CFS and provide clearer hypotheses for future research. Furthermore, because animal models cannot fully reflect the complex neurological, physiological, immune and other alterations observed in CFS, it will be crucial to directly study humans to identify modifiable factors and develop targeted treatments. Scientists know that the microbiome and metabolome are dynamic. This means it may be possible to intervene in ways that genomic data alone cannot. These include lifestyle, dieta and targeted drugs.
BioMapAI also achieved approximately 80% accuracy in external datasets, confirming the key biomarkers identified in the original cohort. This consistency across the different datasets was surprising, the authors stated. The researchers intend to share their dataset widely with BioMapAI, which supports analyses of various symptoms and diseases by effectively integrating multi-omics data that are difficult to replicate in animal models. The researchers’ stated goal is to build a detailed map of how the immune system interacts with gut bacteria and the chemicals they produce. By connecting these dots, we can begin to understand what’s causing disease and pave the way for truly precision medicine, which has long been out of reach.
- Edited by Dr. Gianfrancesco Cormaci, PhD, specialist in Clinical Biochemistry.
Scientific references
Xiong R et al. Nature Med. Jul 25; in press.
Berkis U et al. Front Immunol. 2023; 14:1294758.
Maksoud R et al. J Transl Med. 2021; 19(1):81.
Stoll SV et al. BMJ Open. 2017; 7(9):e015481.
