Lipid Microdomains and Insulin Resistance: Is There a Connection?

See allHide authors and affiliations

Science's STKE  25 Jan 2005:
Vol. 2005, Issue 268, pp. pe3
DOI: 10.1126/stke.2682005pe3


The potential contribution of lipids to insulin signaling has excited interest because of the notion that cholesterol and sphingolipids form functional microdomains—lipid rafts—in cell membranes and that these domains may affect signal transduction. In this Perspective, we discuss the evidence suggesting that cholesterol-sphingolipid rafts play a role in the pathogenesis of insulin resistance. The data relating insulin signaling to lipid rafts in the main insulin target tissues are briefly summarized, including partially controversial findings on the role of caveolae versus other types of rafts. In addition, recent results pointing toward the importance of raft perturbations in the pathogenesis of insulin resistance are discussed. Notably, several studies suggest a correlation between membrane lipid composition and insulin sensitivity. We put forward the idea that the dyslipidemic changes typically associated with insulin resistance and metabolic syndrome may impair the functionality of rafts in insulin target cells, thereby promoting insulin resistance.

Insulin resistance, abdominal obesity, and dyslipidemia coupled with high blood pressure and a proinflammatory state are increasingly common disorders associated with metabolic syndrome, type 2 diabetes, and cardiovascular diseases (1). An important question is how these interrelated phenomena may be mechanistically coupled. Insulin resistance—the unresponsiveness of tissues that are insulin targets to physiological levels of the hormone—is considered to be one of the primary defects in the development of type 2 diabetes (2). This has led several laboratories to study the mechanisms that lead to the desensitization of cells to insulin.

Insulin acts through the insulin receptor (IR), a plasma membrane receptor tyrosine kinase. Upon insulin binding, the receptor is autophosphorylated on tyrosine residues and can thereby phosphorylate other proteins, transducing the signal to metabolic and mitogenic signaling networks in a cell type–specific and regulated manner (3). Many of the signaling molecules are, however, shared by several pathways. One way to achieve signaling specificity in the face of this molecular promiscuity is the spatial segregation of interactions. The ligand-dependent localization of the IR in distinct membrane regions is well understood. The ligand/receptor complexes move from microvilli to nonvillous segments of the membrane where endocytosis occurred (4) in Chinese hamster ovary cells. The rate of receptor endocytosis appears to be influenced by the lateral mobility of the receptor (5). In a simple scenario, lipid imbalance could affect insulin signaling at the plasma membrane level. By hitting receptor activation or endocytic uptake or receptor routing, lipid imbalance could target multiple downstream effectors. Alternatively, the downstream components themselves may be differentially sensitive to perturbation of the lipid environment.

The potential contribution of lipids to insulin signaling has excited interest because of the notion that cholesterol and sphingolipids may form functional domains (lipid rafts) in cellular membranes, and these lipid rafts may affect several processes, including signal transduction (6, 7). These domains are considered to be small (beyond the resolution of light microscopy), dynamic, and based on multiple weak affinities that may become transiently stabilized by various cues. The specific protein and lipid assemblies functioning as domains and the underlying biophysical principles operating in living cells remain incompletely understood. Caveolae that form as oligomers of the structural protein caveolin are a subset of rafts particularly abundant in adipocytes (8).

At present, a wealth of partially controversial information relates insulin signaling to rafts [recently reviewed in (9, 10)]. Most of the data derive from adipocytes, but evidence is also accumulating from other important insulin target tissues, including muscle, liver, and pancreas (Fig. 1). Among the main reasons for the differential results appear to be the different methodologies and criteria used to study raft dependence. This applies both to the use of biochemical techniques, such as detergent solubility (11), and to methods used to morphologically visualize domains. In adipocytes, caveolae form large domains that are typically resolved by light microscopy as caveolin-positive rosettes. However, these structures contain all the elements of the plasma membrane, including nonraft markers (12) (Fig. 2A). In spite of several debatable elements, the general notion that raft-dependent interactions may help to segregate signaling components is clearly acknowledged. This is demonstrated, for example, by the differential sensitivity of two key IR-mediated signaling pathways to raft perturbation: activation of the small guanosine triphosphatase TC10 and phosphoinositide 3-kinase (PI3K) (13) (Fig. 2B).

Fig. 1.

Role of lipid rafts or caveolae in insulin signaling in its target tissues. The main target tissues for insulin action are striated muscle, adipose tissue, the liver, and pancreatic β cells, where insulin regulates its own secretion. The differential coloring of the membrane indicates differential cell-specific lipid composition in both raft (dark) and nonraft (light) areas. Caveolar invaginations are abundant in adipocytes and myocytes. The numbering refers to the experimental manipulations performed to address the contribution of rafts, and the parameters analyzed in different tissues are indicated. Cholesterol manipulation includes cholesterol depletion, loading, or chelation in the membrane; sphingolipid manipulation includes antibody cross-linking or modulation of sphingolipid synthesis; and mouse models include caveolin-1–null mice, caveolin-3–null mice, or GM3-deficient mice. DRMs, detergent-resistant membranes. EM, electron microscopy.

Fig. 2.

(A) Electron micrograph of caveolae in an adipocyte. The adipocyte plasma membrane contains large ring-shaped "caves," which, in addition to caveolae, include all the elements of the plasma membrane, such as clathrin-coated pits and noncaveolar markers. At the light microscopy level, structures in caves may appear to colocalize with caveolin. However, higher resolution is needed to assess true caveolar localization. This image shows a 3T3-L1 adipocyte surface-labeled with a cholera toxin conjugate at 4°C and then processed for Epon embedding. Semi-thick sections(120 to 150 nm) were cut parallel to the substratum. An electron-dense (black) reaction product, indicating surface-connected structures, is associated with caves (asterisks), which are studded with caveolae (some indicated by arrowheads). The caves appear within the cell interior, in this case close to a lipid body (LB), but are actually connected to the cell surface out of the plane of the section. Scale bar, 500 nm. [Electron micrograph courtesy of Robert G. Parton, University of Queensland] (B) Insulin signaling and lipid rafts or caveolae in adipocytes. The localization of the IR into lipid rafts or caveolae is controversial. However, caveolar integrity is required for the activation of the CAP-Cbl-TC10 signaling pathway in response to insulin (13). IRS-1 to PI3K signaling is at least partly activated in endosomes, and alteration of GM3 levels in lipid rafts interferes with this event (15). P, phosphotyrosine; CrkII, Crk family adaptor protein II; APS, adaptor protein containing PH (pleckstrin homology) and SH2 (Src homology 2) domains; FLT, flotillin; PIP3, phosphoinositol(3,4,5)-triphosphate; CAP, Cbl-associated protein.

The question of whether lipid rafts play a role in the pathogenesis of insulin resistance is receiving increasing attention. A link between insulin resistance induced by the proinflammatory cytokine tumor necrosis factor–α (TNF-α) and rafts was uncovered in adipocytes. TNF-α stimulated expression of the gene encoding GM3 synthase (14), resulting in the doubling of GM3 ganglioside levels in rafts and a parallel reduction in IR raft association (15). This is in line with the observed sensitivity of IR raft association and insulin signaling on plasma membrane ganglioside distribution (16), as well as the enhancement of insulin signaling in mice lacking GM3 synthase (17). Furthermore, mitogen-activated protein kinase activation by insulin was not impaired, whereas insulin-dependent IR internalization and IR substrate 1 (IRS-1) phosphorylation were decreased, which suggests GM3-dependent segregation of signaling pathways and the involvement of endocytic circuits (15) (Fig. 2B).

Several observations from both experimental animals and humans suggest a correlation between insulin sensitivity and membrane lipid composition. In aging Wistar rats that became insulin-resistant, increased plasma membrane viscosity was associated with decreased IR activity in the liver (18). In humans, a relation between the fatty acid composition of skeletal muscle membranes and insulin sensitivity has been demonstrated: The greater the percentage of polyunsaturated fatty acids, the better the insulin action (19, 20). Moreover, fasting plasma insulin levels were positively correlated with adipocyte membrane sphingomyelin and cholesterol content in obese women (21). Whether the changes observed in membrane lipids in insulin-resistant states are associated with an altered assortment of raft proteins and lipids and whether this affects IR signaling through rafts have not been investigated.

Perhaps the most interesting question relates to the consequences that the dyslipidemia typically associated with insulin resistance and metabolic syndrome (hypertriglyceridemia and low high-density lipoprotein cholesterol) may have on membrane lipid domains. Primary defects in energy balance that produce visceral adiposity are thought to be of key importance in the dyslipidemia pathogenesis. The increased availability of fatty acids, primarily from visceral fat, leads to accelerated hepatic synthesis of very-low-density lipoprotein, and hypertriglyceridemia in turn promotes the modification of other lipoproteins, contributing to their atherogenic profile [reviewed in (1, 22)]. Serum lipoproteins are actively exchanging lipids with cells. It would therefore be interesting to analyze how such a dyslipidemic population of lipoproteins contributes to the composition and functionality of lipid domains in insulin target cell membranes. It is worth remembering that rather than the lipids per se, their metabolites (such as fatty acyl coenzyme A, diacylglycerol, or ceramide) may also modulate membranes and serve as potential mediators of insulin resistance.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
View Abstract

Navigate This Article