Multi-omic definition of metabolic obesity through adipose tissue-microbiome interactions.

A metabolome-informed obesity metric (metBMI) was developed from deep multi-omics phenotyping of 1,408 individuals that captures adipose tissue-related dysfunction across organ systems.

• The metric was derived from a discovery cohort of 1,408 individuals.
• metBMI was designed to capture metabolic heterogeneity not fully captured by body mass index (BMI).
• The metric reflects adipose tissue-related dysfunction across multiple organ systems.
• metBMI was validated in an external cohort of n=466 individuals.

In an external validation cohort, metBMI explained 52% of BMI variance and more accurately reflected adiposity than other omics models.

• The external cohort comprised n=466 individuals.
• metBMI explained 52% of BMI variance in the external cohort.
• metBMI outperformed other omics-based models in reflecting adiposity.
• This validation demonstrated generalizability of the metabolome-based metric across cohorts.

Individuals with higher-than-expected metBMI had 2–5-fold higher odds of multiple cardiometabolic conditions.

• Elevated metBMI was associated with 2–5-fold higher odds of fatty liver disease, diabetes, severe visceral fat accumulation and attenuation, insulin resistance, hyperinsulinemia, and inflammation.
• These associations were observed relative to individuals with lower-than-expected metBMI for a given BMI.
• The findings suggest metBMI captures metabolic risk beyond what BMI alone indicates.
• Multiple cardiometabolic endpoints were assessed, covering hepatic, endocrine, and inflammatory domains.

Individuals with higher-than-expected metBMI undergoing bariatric surgery achieved 30% less weight loss.

• The bariatric surgery subcohort comprised n=75 individuals.
• Those with an elevated obesogenic metBMI signature achieved 30% less weight loss following bariatric surgery.
• This finding suggests metBMI may have clinical utility in predicting surgical outcomes.
• The result highlights metabolic heterogeneity in treatment response among individuals with obesity.

The obesogenic metBMI signature was associated with reduced gut microbiome richness, altered ecology, and altered functional potential.

• Higher metBMI aligned with reduced microbiome richness.
• Altered microbial ecology and functional potential were observed in individuals with higher-than-expected metBMI.
• These microbiome associations suggest a gut microbial component to the metabolic obesity phenotype.
• Microbiome characteristics were part of the multi-omic framework linking adipose dysfunction to systemic metabolic changes.

A 66-metabolite panel retained 38.6% explanatory power for the metBMI signature, with 90% of these metabolites covarying with the microbiome.

• A condensed panel of 66 metabolites was identified as sufficient to capture a substantial portion of the metBMI signal.
• This panel retained 38.6% explanatory power relative to the full metabolome-based metric.
• 90% of the 66 metabolites covaried with the gut microbiome.
• The covariation suggests the metabolite panel captures a microbiome-linked metabolic signal.

Mediation analysis revealed a bidirectional, metabolite-centered host-microbiome axis mediated by lipids, amino acids, and diet-derived metabolites.

• Mediation analysis was used to characterize directionality and mediators of host-microbiome interactions.
• The axis was found to be bidirectional, with metabolites serving as central mediators.
• Key metabolite classes mediating this axis included lipids, amino acids, and diet-derived metabolites.
• These findings position circulating metabolites as functional intermediaries between gut microbial ecology and host adipose-related dysfunction.

BMI does not fully capture obesity's metabolic heterogeneity, motivating the development of alternative metrics.

• The paper states that 'obesity's metabolic heterogeneity is not fully captured by body mass index (BMI).'
• This limitation of BMI in clinical and research settings provided the rationale for the multi-omics approach.
• Metabolic obesity phenotypes can differ substantially among individuals with similar BMI values.
• The study was designed to address this gap by developing a metabolome-informed alternative.