Adipocyte-Derived Extracellular Vesicles: Caveolin Matters

Caveolin proteins that form the coat of small flask-shaped invaginations of the cell plasma membrane, called caveolae, were codiscovered in the early 1990s by several groups from different fields (1,2), including P.E. Scherer in the H.F. Lodish laboratory, who was working on adipocyte insulin-responsive glucose transport (3). Since then, the caveolin protein family has been the subject of active research in the context of protein–lipid interactions, membrane dynamics, vesicular trafficking, mechanotransduction, shear stress, and receptor signaling, which has unraveled multiple unrelated functions and many unsolved issues (4). In the search for physiological roles, mouse models with caveolin gene deletion were generated (reviewed in Le Lay and Kurzchalia (5)) that defined a muscle-specific function of caveolin-3 and prominent roles of caveolin-1 (Cav1) linked to vascular permeability, pulmonary functions, and glucose/lipid metabolism.

Among the metabolic defects in Cav1-ablated mice, lipoatrophy develops over time and is highly exacerbated when animals are fed an obesogenic high-fat diet (6). Indeed, the absence of caveolin expression reveals the inability of adipose tissue (AT) to respond to insulin stimulation and to expand despite preserved adipocyte differentiation capacity (7). Such a lipoatrophic phenotype is also found in the very few patients with Cav1-null mutation described thus far (8).

In the present issue of Diabetes, Crewe et al. (9) address the question of the specific role of adipocyte Cav1. The authors report that the sole ablation of Cav1 in mature adipocytes (white or brown) by targeting Cre recombinase with an adiponectin promoter-based construct (Ad-Cav1KO, for adipocyte-specific Cav1 knockout [Cav1KO]) can reproduce most of the AT defects found in whole-body Cav1-null mice. Thus, Ad-Cav1KO mice display blunted insulin-stimulated AKT activation, partial lipodystrophy, and dysregulated lipid metabolism, as well as significant AT inflammation, fibrosis, and mitochondrial dysfunction, when fed a high-fat diet. This was quite an expected finding, as previous studies that used a strategy of tissue-specific reexpression of Cav1 in mice with a caveolin-null background (10) reported normalization of all vascular defects with specific endothelial Cav1 rescue but persistence of lipoatrophy and high-fat-feeding–related metabolic failure (11). This study by Crewe et al. (9) therefore confirms the central role played by adipocyte Cav1 in the control of fat cell expandability and insulin response in mice. However, no clear Cav1 expression changes are seen in humans when lean subjects are compared with patients with different degrees of obesity or with graded severity of obesity-related complications. In human obesity studies, whole AT is often evaluated in which the caveolin content results from the combined expression of Cav1 expressed by adipocytes and endothelial cells, both caveolin-rich cell types, thereby creating an important confounding factor. Conversely, when adipocyte Cav1 expression alone was examined, a positive correlation was found with fat cell size expansion in a group of human volunteers enrolled in a controlled overfeeding trial to promote fattening (12), arguing favorably for a critical role of Cav1 in the healthy adaptation of AT to overnutrition.

The present report by Crewe et al. (9) highlights more surprising findings regarding the metabolic phenotype of Ad-Cav1KO animals (Fig. 1). Unexpectedly, these mice displayed ameliorated systemic clearance of triglycerides (TG) following an oral TG gavage, a phenotype that is opposite the one described for whole-body Cav1KO mice that shows severely impaired blood TG clearance. Furthermore, even though adipocyte Cav1 ablation resulted in dysfunctional AT, which is recognized as a major contributor to whole-body glucose and lipid metabolism alterations, Ad-Cav1KO mice unpredictably exhibited improved blood glucose clearance compared with controls or whole-body Cav1KO mice when fed a short-term high-fat diet. Detailed metabolic phenotyping indicated that this was due to increased insulin-independent glucose uptake by white AT and reduced hepatic gluconeogenesis.

In search of an explanation for systemic metabolic amelioration in the Ad-Cav1KO model, the authors focused on AT-derived extracellular vesicles (AT-EVs) by detecting increased serum small EVs (sEVs) in Ad-Cav1KO mice, having previously demonstrated their role as mediators of interorgan communication (13). The authors demonstrated that intravenous injection of AT-EVs from Ad-Cav1KO mice into control mice phenocopied the endogenous glucose phenotype of the Ad-Cav1KO animals, i.e., improved glucose clearance compared with high-fat-diet mice injected with PBS. Conversely, intravenous injection of control mice with AT-EVs from high-fat-diet–fed control mice led to blunted glucose clearance compared with mice receiving PBS. By fluorescent labeling of AT-EVs prior to their injection, the authors determined that AT-EVs preferentially targeted liver, heart, pancreas, and AT (but not muscle). These data confirm the predominant hepatic targeting of EVs regardless of their origin, size, or method of administration (14) and are in line with another recent report from the team that found in vivo AT-EVs control AT–heart metabolic crosstalk (13).

It was also observed that AT-EVs from Ad-Cav1KO mice could propagate a negative proinflammatory and profibrotic phenotype to the liver. How Cav1 depletion influences AT-EV lipid and protein composition as well as cell/tissue anchorage or targeting has not been investigated. Moreover, the question remains whether AT-EVs impact metabolic pathways directly or indirectly by changing the immunophenotyping of resident macrophages, as this cell population is known to efficiently take up EVs. Overall, how Cav-1 influences Ad-EV composition and function requires further investigation at the molecular level.

Of note, isolated adipocytes from Ad-Cav1KO mice display no visible adipocyte caveolae and strongly depressed Cav1 mRNA expression but showed noticeable Cav1 protein levels (only a 50% reduction compared with control adipocytes), which might be confusing for the interpretation of the results. Indeed, the authors previously reported that adipocytes could acquire Cav1 from EVs produced by surrounding endothelial cells (15), and such a transfer was likely the cause of the residual Cav1 adipocyte expression in Ad-Cav1KO. Thus, Ad-Cav1KO fat, and probably AT in general, appears to be a hub of the complex exchange of EVs between cells and to and from mature adipocytes. Nonetheless, how caveolin traffic from endogenous or exogenous sources can modulate EV production via adipocytes remains unknown. Another interesting issue relates to how the present observation is linked to adipocyte mitochondrial dysfunction observed in Cav1KO mice and to the process of EV-mediated mitochondrial transfer from adipocytes (13,16,17).

There remains much to learn in this emerging field, with great potential for adipocyte capabilities in the orchestration of metabolic health.