Adipose Modulation of ABCG1 Uncovers an Intimate Link Between Sphingomyelin and Triglyceride Storage

In the early 1980s, it was established that adipose tissue not only plays an important role in triglyceride storage but also accounts for approximately 25% of total body cholesterol in humans (1). However, in obese individuals this proportion of cholesterol can increase to well over half and could contribute to adipocyte dysfunction and obesity-mediated metabolic syndrome, including low levels of plasma HDL cholesterol (1–4). Consistent with the view that adipose tissue behaves as a cholesterol sink, and its accumulation is proportional to triglycerides in response to nutritional changes to maintain cellular integrity and to regulate cellular hypertrophy, it is not surprising that adipocytes have developed a unique ability to deal with cholesterol (1). It may be for this reason that adipocytes have evolved with an extremely limited capacity to perform de novo cholesterol synthesis (1) but possess multiple strategies to extract cholesterol from circulating lipoproteins (5,6). Despite cholesterol being delivered as cholesteryl ester in adipocytes, approximately 95% of cholesterol paradoxically exists as free cholesterol (FC) (7) and resides in the plasma membrane or the cytosolic interface of lipid droplets where it is readily available for mobilization (8). Cholesterol mobilization involves adipose tissue ATP-binding cassette transporter A1 (ABCA1), which has recently been shown to contribute to HDL biogenesis in vivo (9,10). These findings illustrate an exquisite balance between adipose tissue cholesterol storage and plasma HDL cholesterol concentration and raise the question about the role that HDL cholesterol plays in obesity and associated metabolic complications.

ATP-binding cassette transporters ABCA1 and ABCG1 primarily mediate the efflux of cholesterol from peripheral cells. These transporters promote unidirectional cholesterol efflux to lipid-poor apolipoprotein A-I (apoA-I), apoE, or HDL particles, respectively, and are under the transcriptional control of liver X receptors (LXRs), master intracellular sensors that are activated in response to cholesterol biosynthetic intermediates and oxysterols (11,12). This contrasts to scavenger receptor class B type I (SR-BI), which is not an LXR target and facilitates the bidirectional flux of cholesterol depending on the cholesterol status of the cells (11). Remarkably, HDL cholesterol transport machinery is highly represented in adipocytes (5,6,13) but the relevance to obesity and metabolic complications remains poorly understood. While LXR-deficient mice are leaner than littermate controls and are protected against diet-induced obesity, the mechanisms that control fat loss in these mice are still a matter of debate and could involve either reduced lipogenesis, enhanced lipid oxidation and lipolysis, or reduced cholesterol efflux pathways (14). Consistent with apoE being an LXR target gene in adipocyte (13), apoE-deficient mice also exhibit a lean phenotype and are resistant to diet-induced obesity (15–17). However, the interpretation of this phenotype can be confounded by the fact that apoE plays a dual role promoting cholesterol efflux and mediating lipoprotein clearance (15–17) similar to SR-BI (5,6,11). LXRs also transcriptionally regulate the cholesterol efflux transporters ABCA1 and ABCG1, and deficiency of ABCA1 is associated with decreased body weight (18), while ABCG1 deficiency protects against diet-induced obesity (19). Altogether these studies point to an overall consensus in which a defective cholesterol efflux pathway is associated with reduced fat storage, an effect that could be attributed to adipose cachexia (20). However, recent studies suggest that myeloid-specific ABCA1 deletion does not alter diet-induced obesity (21), and paradoxically, specific adipocyte ABCA1 deficiency leads to increased body weight gain and fat storage in response to diet-induced obesity due to adipose cholesterol accumulation altering lipolysis (22) (Fig. 1). These complex findings emphasize the need for a more nuanced understanding of how cholesterol efflux pathways affect adipose function.

In this issue, Frisdal et al. (23) use a number of in vitro and in vivo approaches along with data from people carrying single nucleotide polymorphisms associated with increased ABCG1 expression to put forward the idea that deletion of adipose ABCG1 is responsible for the protection against adiposity as seen in the global Abcg1^−/− mice (19). To test their hypothesis in vivo, the authors developed an innovative approach of a local injection of Abcg1 short hairpin RNA lentiviral particles into the epididymal fat pads of mice. While one of the caveats of this approach could be the targeting of adipose macrophages, the authors were keen to show that deletion of ABCG1 in adipose tissue led to reduced inflammation, inconsistent with a major targeting of Abcg1 in macrophages promoting an inflammatory response (12) but consistent with reduced adipocyte fat storage (23). Future studies will be needed to elucidate the molecular mechanisms as reduced adipocyte inflammation has also been recently observed in apoA-I transgenic mice (24). Several mechanisms could be considered. First, reduced cellular cholesterol was observed in both studies and could reflect reduced proinflammatory sterols in adipocytes or a shunt into production of anti-inflammatory sterols such as desmosterol (12). Second, modulation of fatty acid metabolism−related genes observed in ABCG1-deficient adipocytes could be linked to lower inflammation and/or decreased transcriptional responses of nuclear factor-κβ target genes as a result of altered histone acetylation in their enhancer/promoter regions (12). Third, modulation of fatty acid oxidation (especially enhanced activity of the uncoupling protein 1 [UPC1]) as observed in LXR knockout mice (14) could lead to adipose tissue browning. Finally, the impact of enhanced cellular sphingomyelin (SM) content on adipocyte inflammation in ABCG1-deficient adipocytes still needs to be elucidated.

In contrast to a previous report (9), Frisdal et al. claim that ABCG1 contributes to adipose cholesterol efflux even though this mechanism does not appear to drive their phenotype, as adipose ABCG1 deficiency was associated with lower cellular cholesterol content independent of any change in ABCA1 expression or apoE secretion in adipocytes. This contrasted with the adipose cholesterol accumulation observed in adipose ABCA1 deficiency driving an upregulation of LXR target genes such as ABCG1 (21). So, this raises the question: Where did the cholesterol go in ABCG1-deficient adipocytes? As cholesterol synthesis is extremely low in adipocytes, it is most likely that alternative cholesterol efflux pathways are taking place in these cells. As ABCG1 appears to promote movement of cholesterol from the endoplasmic reticulum to the plasma membrane to release sterols more readily to extracellular lipoprotein acceptors (12), it is tempting to speculate that cholesterol trafficking is diverted to lysosomal pathways at the expense of the endoplasmic reticulum–like structure surrounding lipid droplets in ABCG1-deficient adipocytes (8). This could explain the higher lipid raft visualization in these cells despite reduced cellular cholesterol and triglyceride content (23). Further studies are required to better comprehend cholesterol trafficking in adipocytes and the interplay between ABCA1, ABCG1, and SR-BI in adipocyte cholesterol efflux.

Perhaps the most intriguing finding in this study is the role of ABCG1 mediating SM efflux in adipocytes. One previous model based on transfected cells suggested that the ability of ABCG1 to promote SM efflux could lead to such an extensive change in the equilibrium of membrane components that the outer leaflet becomes more attractive to sterol efflux (25). Frisdal et al. (23) translate this concept in a more physiological context by showing that defective ABCG1-mediated SM efflux could be a primary function of this transporter in adipocytes with pathological consequences on cholesterol efflux and adipose functions. Because SM avidly binds cholesterol, accumulation of intracellular SM in ABCG1-deficient adipocytes could also prevent cholesterol transport from lysosomes to the plasma membrane further contributing to reduce cholesterol efflux (26). However, reduced cellular cholesterol content was observed in ABCG1-deficient adipocytes, suggesting that accumulation of SM occurred at the level of the plasma membrane in these cells (23). Jiang and colleagues (27) have previously reported that reduced plasma membrane SM prevented fat storage by enhancing adipose tissue insulin sensitivity. The current study rather proposes that part of the reduced triglyceride storage in ABCG1-deficient adipocytes could result from enhanced plasma membrane SM content preventing lipoprotein lipase (LPL) activity (Fig. 1). This was illustrated by showing that SM depletion by sphingomyelinase rescued not only LPL activity but also the fat loss of ABCG1-deficient adipocytes, while specific removal of cholesterol by cyclodextrin had no impact.

A possible involvement of ABCG1 in preventing human adipose fat regulation is illustrated in the current study through human genetics (23). A beneficial association between body fatness and triglycerides was also previously reported for both ABCA1 and apoE loss of function (28,29). However, because both LXRα gain of function and LXRβ loss of function could be associated with obesity (for review, see 14), more studies are required to elucidate the complex interactions between cholesterol efflux pathways and adipocyte plasticity and functions in humans. Perhaps one intervention that is of more immediate relevance is the contribution of cholesteryl ester transfer protein (CEPT) inhibitors, which raise endogenous HDL, as results from phase III clinical trials are expected soon. For instance, it will be important to determine the relevance of the pathway identified by Frisdal et al. (23) in the presence of CETP as mice do not express it and transgenic mice overexpressing CETP exhibit reduced adipose fat storage (30).