Obesity: stressing about unfolded proteins

Two reports in this issue identify a link between insulin action and the unfolded protein response—a pathway that helps the endoplasmic reticulum cope with cellular stress (pages 429–437 and 438–445). The results expand our understanding of the pathogenesis of insulin resistance and provide new targets for prevention and treatment of obesity.

Weight gain often results in the development of insulin resistance—a metabolic state that reduces the ability of cells to take up glucose from the blood in response to normal circulating levels of insulin, and is a risk factor for diabetes. At the same time, weight gain also enhances insulin-stimulated lipogenesis, leading to excess fat storage in tissues other than adipose tissue, which can create a state of chronic low-level inflammation¹.

Why obesity is associated with insulin resistance and inflammation is not fully understood, but two reports in this issue of Nature Medicine bring us one step closer to reaching an answer at the molecular level^2,3. A chronic imbalance between energy supply and demand, as occurs in obesity, exposes cells to toxic fatty acids, which activate cellular stress pathways^4,5. The epicenter for these stress responses is the endoplasmic reticulum (ER), a membranous network that functions in the synthesis and assembly of secretory and membrane proteins in addition to the metabolism of carbohydrates, lipids, steroids and xenobiotics. Any perturbations that disrupt ER function elicit a tripartite ER-to-nucleus signaling pathway known as the unfolded protein response (UPR)^5,6,7. The new studies, by Park et al.² and Winnay et al.³, highlight how the ability of the UPR to help maintain glucose tolerance and prevent activation of inflammatory pathways requires a unique partnership between a transcription factor induced by one arm of the UPR and the p85 regulatory subunit of phosphoinositide 3-kinase (PI3K), an enzyme regulated by insulin.

The UPR regulates the expression of genes that encode chaperones—proteins that help other proteins fold properly—and other proteins involved in alleviating ER stress. The UPR is activated after feeding as a way to expand the capacity of the ER to help manage an increased energy load. Activation of the UPR is governed by three ER transmembrane proteins: inositol-requiring kinase-1 (IRE1), protein kinase–like ER kinase (PERK) and activating transcription factor-6 (ATF6) (Fig. 1)^5,6.

Figure 1: A link between insulin signaling and a stress response pathway of the endoplasmic reticulum.

IRE1, PERK and ATF6 are ER transmembrane proteins that monitor the accumulation of misfolded protein inside the ER. Upon ER stress, IRE1 endonuclease facilitates the splicing of XBP-1 mRNA, leading to translation of an activated version of the transcription factor XBP-1s. XBP-1s binding of one of the PI3K regulatory subunits, p85α or p85β, enhances nuclear entry of this transcription factor and induction of UPR target genes. The p85 subunit does not directly modulate other arms of the UPR, including the association of IRE1 with TRAF2 and the subsequent activation of JNK and PERK phosphorylation of eIF2 (p-eIF2), which represses protein synthesis coincident with preferential translation of the UPR transcriptional activator ATF4. However, the ATF6 arm of the UPR, which involves proteolytic cleavage and release of the transcription factor into the nucleus, may be activated through association with p85 (ref. ⁴). GADD34, growth arrest and DNA damage–inducible protein.

Exactly how obesity activates these three regulatory proteins is not well understood, but studies suggest that ER stress releases the ER chaperone BiP from binding them and inhibiting their functions. Once released from BiP, each regulatory protein realizes its potential⁴.

Activation of IRE1 facilitates cytoplasmic splicing of the mRNA for X-box–binding protein-1 (XBP-1), leading to the synthesis of a highly active transcription factor called XBP-1s that turns on UPR gene expression. PERK phosphorylates the translation initiation factor eIF2, averting the build-up of misfolded proteins by lowering the rate of mRNA translation in the ER⁸ (Fig. 1). At the same time, phosphorylation of eIF2 leads to preferential synthesis of ATF4, another activator of UPR gene expression. Finally, BiP release of ATF6 in the ER allows ATF6 to migrate to the Golgi, where its cytosolic domain is proteolytically cleaved, releasing this transcriptional activator of UPR target genes⁴.

The importance of the IRE1–XBP-1 arm of the UPR in the management of obesity and diabetes prevention is illustrated by the observation that mice deficient in XBP-1 develop diet-induced insulin resistance⁷. In addition, in association with tumor necrosis factor receptor–associated factor-2 (TRAF-2), IRE1 activates the c-Jun N-terminal kinase (JNK) pathway, which interferes with insulin signaling via serine phosphorylation of the insulin receptor substrate-1 (IRS-1)^5,6,7.

The studies by Park et al.² and Winnay et al.³ now add to this body of evidence, demonstrating that the IRE1–XBP-1 arm of the UPR is linked to insulin action and development of insulin resistance and inflammation: they show that obesity reduces a key binding event between XBP1s and the p85 subunit of PI3K.

The class 1A PI3K signaling pathway is activated by receptor tyrosine kinases and is implicated in a wide range of cellular functions, including insulin-stimulated glucose uptake and signal transduction^9,10. PI3K is a cytosolic heterodimer consisting of a p85 regulatory subunit (one of three isoforms α, β or γ) and a p110 catalytic subunit. Park et al.² and Winnay et al.³ suggest that p85, in addition to being a regulatory subunit for p110, also regulates XBP-1. They show that insulin promotes the interaction between p85 and XBP-1s, which, in turn, leads to XBP-1s's nuclear translocation and possibly stabilization. Notably, the interaction between p85 and XBP-1s is not regulated by PI3K catalytic activity, indicating that this interaction is independent from the role of insulin in glucose uptake by cells.

Collectively, these studies find that depletion of p85 in the liver of lean mice inhibits XBP-1s nuclear translocation and decreases the induction of XBP-1s–regulated genes following ER stress. In contrast, overexpression of p85 in mutant ob/ob mice, a model for obesity and diabetes, restores XBP-1s nuclear translocation and UPR expression in response to feeding, resulting in improved glucose tolerance².

Some of these findings are difficult to reconcile with studies reporting that genetic deletion of p85 in mice fed a high-fat diet results in increased insulin sensitivity^9,10. A deeper understanding of the roles of the various p85 isoforms among tissues and metabolic states may help resolve some of these discrepancies.

Short-term ER stress induced by feeding selectively accentuates XBP-1s activation of UPR gene expression to expand the capacity of the ER to fold proteins that function in the metabolism of a meal. As long as the energy load delivered to cells matches their energy storage capacity, this limited activation of the UPR is sufficient to restore ER homeostasis (Fig. 1). However, in the ob/ob mouse model, the chronic mismatch between energy delivery and storage capacity leads to a reversal of these adaptive features of the UPR. Upon feeding, ob/ob mice show reduced association between p85 with XBP-1s in the liver, and consequently lower nuclear targeting of XBP-1s to induce key UPR genes². As a result, ER stress continues to build, amplifying activation of the PERK arm of the UPR along with JNK kinase activity toward IRS-1, promoting insulin resistance. Thus, the extent to which p85 is available to guide XBP-1s to the nucleus affects the level of ER stress produced by a meal. Obesity makes it harder for XBP-1s to be activated, contributing to insulin resistance. In support of this conclusion, overexpression of the p85α or p85β isoforms in the liver of the ob/ob mice restores nuclear XBP-1s and markedly improves glucose disposal².

The current studies emphasize that insulin mediates metabolic control in part by activating XBP-1s–regulated genes. Failure of the IRE1–XBP-1s–p85 pathway to regain ER homeostasis then signals for full UPR activation, promoting a proinflammatory and proapoptotic state. Many questions, however, remain. How is the interaction between p85 and XBP-1s regulated by insulin and fatty acids in the lean and obese states? What tissue-specific differences exist for this partnership? And how do chemical chaperones, which improve glucose tolerance and insulin action in diabetic ob/ob mice through a postulated enhancement of protein folding, influence this new pathway¹¹?

The new link between the insulin signaling and the ER stress pathways suggests that p85 offers a new therapeutic target for obesity and type 2 diabetes. In future studies, it will be important to delineate the role of the XBP-1s–p85 partnership in metabolism, as well as to elucidate the full mechanistic details for p85 control of the UPR.