Hao-Ran Wang, Xue-Hui Lv, Yue Sun, Zhi-Chao Gu, Meng Lin, Le Guan, Qing-Feng Wang, Shi-Meng Wang
Extracellular vesicles (EVs)-lipid bilayer-enclosed nanoparticles secreted by virtually all cell types-have shifted from being viewed as passive cellular byproducts to active signaling nodes orchestrating inter-cellular and inter-organ communication. Yet the field has accumulated faster than it has integrated: hundreds of EV-cargo-phenotype associations exist as isolated edges of a network whose system-level architecture remains poorly defined. Here we propose the vesiculome as an operational framework for that network, resting on three falsifiable axioms: (i) the EV complement of an organism constitutes a network addressable by donor cell × target tissue × cargo class; (ii) cargo composition tracks donor-cell metabolic state in a quantitatively predictable way; and (iii) the integrated balance between pro-inflammatory and pro-resolving vesicle outputs - rather than any single edge - determines the organismal metabolic phenotype. Organized along the chain of EV generation, immune-metabolic interaction, disease mechanism, and translational application, the review synthesizes how this network sustains metabolic homeostasis and how its dysregulation drives metaflammation, the chronic low-grade inflammation underlying obesity, type 2 diabetes mellitus, metabolic dysfunction-associated steatotic liver disease, and atherosclerosis. We dissect two reciprocal arcs of this vesiculome. Polarization-specific EVs from M1/M2 macrophages, Th17 and regulatory T cells, dendritic cells, NK cells, and neutrophils deliver inflammatory or protective cargo to adipose tissue, liver, and pancreas. EVs from adipocytes, hepatocytes, skeletal myocytes, pancreatic β-cells, and intestinal epithelial cells reciprocally reshape the immune microenvironment. Disease arises as a network-level configuration of these arcs rather than as isolated edge failures. We apply a four-tier causality framework to every claim, distinguishing correlational evidence from cargo-depletion and physiological-dose validation. Only miR-155, miR-122, miR-690, and miR-33 currently satisfy the highest tier; most cargoes require systematic experimental escalation before therapeutic translation. We further separate preclinical from clinically validated evidence, and downgrade plant-derived nanovesicles to a methodological cautionary note given non-reproducible cross-kingdom claims. The translational arm evaluates EV-based liquid biopsy biomarkers, native and engineered therapeutic EVs, and mesenchymal stem cell-derived EVs against this framework. We close with a structured catalogue of foundational, causal, and translational knowledge gaps and a priority research agenda built on single-EV multi-omics, in vivo tracking, multi-organ organoid-on-chip platforms, and longitudinal human cohorts.