A conserved transcriptomic model defines metabolic resilience and vulnerability in obesity

Abstract

BackgroundObesity arises from a complex interplay of genetic and environmental factors, with alterations of transcriptional networks that integrate metabolic, immune, and regulatory pathways. Conventional measures such as body mass index (BMI) quantify body size but fail to capture the molecular heterogeneity underlying divergent metabolic outcomes. We therefore sought to construct a gene expression-based transcriptomic representation of obesity, using BMI as a practical training anchor, and to use this framework to delineate transcriptional programs associated with metabolically healthy and pathogenic obesity, with subsequent projection to mouse transcriptomic data for cross-species validation. MethodsTranscriptome data of human visceral adipose tissue (N= 1,298) were used to derive the transcriptomic BMI model, and genes contributing to the model were functionally annotated by gene set enrichment analysis. The human-trained model was subsequently applied to mouse selection lines (N = 18) with divergent obesity phenotypes. In the human cohort, post hoc stratification into metabolically healthy obesity (MHO) and metabolically unhealthy obesity (MUO) groups was performed using a downstream classification framework incorporating observed BMI together with predicted BMI, to assess whether model-derived predicted BMI reflected obesity-related pathophysiologic status. ResultsModel-selected genes were involved in coordinated regulation of lipid metabolism, immune activation, and growth signaling, extending to mitochondrial and translational pathways. Cross-species analyses uncovered conserved metabolic polarization: DU6 mice exhibited lipid-anabolic and inflammatory remodeling, whereas DU6P mice displayed oxidative, mitochondrial, and GH-axis-enriched transcriptional states. In human cohorts, MHO individuals showed upregulation of mitochondrial energetics and protein synthesis, while MUO individuals were characterized by increased autophagy, lipid catabolism, and stress-adaptive signaling on the transcriptional level. Together, these findings define a conserved molecular continuum linking oxidative efficiency to metabolic health and inflammation to metabolic vulnerability. ConclusionsThis integrative transcriptomic framework bridges human and mouse adipose biology to uncover conserved mechanisms underlying obesity phenotypes. By contrasting mitochondrial and translational programs with inflammatory and catabolic pathways, it provides mechanistic insight into metabolic resilience and a foundation for precision approaches to obesity management.