|Home | About | Journals | Submit | Contact Us | Français|
Leptin regulates energy balance and glucose homeostasis. Shortly after leptin was identified, it was established that obesity is commonly associated with leptin resistance, though the molecular mechanisms remain to be identified. To explore potential mechanisms of leptin resistance, we employed organotypic brain slices to identify candidate signaling pathways that negatively regulate leptin sensitivity. We found that elevation of adenosine 3′, 5′-monophosphate (cAMP) levels impairs multiple signaling cascades activated by leptin within the hypothalamus. Notably, this effect is independent of protein kinase A activation. In contrast, activation of Epac, a cAMP-regulated guanine nucleotide exchange factor for the small G protein Rap1, was sufficient to impair leptin signaling with concomitant induction of SOCS-3 expression. Epac activation also blunted leptin-induced depolarization of hypothalamic POMC neurons. Finally, central infusion of an Epac activator blunted the anorexigenic actions of leptin. Thus, activation of hypothalamic cAMP-Epac pathway is sufficient to induce multiple indices of leptin resistance.
Obesity arises when energy intake chronically exceeds energy expenditure. Obesity is associated with several comorbidities, including type 2 diabetes mellitus, several types of cancer, and cardiovascular disease (Gaede et al., 2008; Stamler et al., 1993; Van Gaal et al., 2006). Reduction in body weight has a beneficial impact on a number of metabolic and cardiovascular risk factors. Thus, development of effective strategies to fight obesity will reduce the incidence of a myriad of diseases. Leptin is a hormone that plays a central role in the regulation of energy balance and glucose homeostasis via activation of leptin receptors, particularly within the central nervous system (Bjørbaek and Kahn, 2004; Flier, 2006; Friedman, 2004; Zhang et al., 1994). Leptin administration decreases food intake, reduces body weight, and increases systemic insulin sensitivity when administrated to lean humans and animals. However, nearly all forms of obesity are associated with higher levels of circulating leptin (Considine et al., 1996; Frederich et al., 1995; Maffei et al., 1996). Thus, obese humans and animals display a decreased response to endogenous and exogenous leptin. This has been demonstrated by several measures, including attenuated anorectic responses, reduction of both signal transducers and activators of transcription (STAT) 3 phosphorylation, and neuro-peptide release after leptin administration (Bjørbaek and Kahn, 2004; Enriori et al., 2007; Flier, 2006; Friedman, 2004; Münzberg et al., 2004; Zhang et al., 1994). Thus, understanding the molecular mechanisms underlying and developing strategies to combat “leptin resistance” is a major goal of obesity research.
Cellular leptin resistance may be mediated by negative regulatory pathways of leptin receptor signaling. Indeed, recent studies have identified molecules that act as negative regulators of leptin signaling. These include suppressor of cytokine signaling-3 (SOCS-3) (Bjørbaek et al., 1999; Howard et al., 2004; Mori et al., 2004), protein tyrosine phosphatase 1B (PTP1B) (Banno et al., 2010; Bence et al., 2006; Cook and Unger, 2002; Zabolotny et al., 2002), and inflammatory signals such as IKKβ/NFκB and ER stress (Ozcan et al., 2009; Zhang et al., 2008). However, the signaling networks that confer central leptin resistance remain to be fully established. Therefore, delineation of cellular signaling networks responsible for leptin sensitivity is a high priority.
Accumulating evidence suggests that adenosine 3′, 5′-monophosphate (cAMP) generally suppresses proinflammatory cytokine signaling (Serezani et al., 2008). Numerous studies have demonstrated that cAMP plays a role as a downregulator of IL-6 signaling which uses JAK-STAT3 signaling (Bousquet et al., 2001; Delgado and Ganea, 2000; Fasshauer et al., 2002; Gasperini et al., 2002; Sands et al., 2006). Previous studies showed that leptin reduces cAMP levels through activation of phosphodiesterase 3B in pancreatic beta cells (Zhao et al., 1998) and in the brain (Sahu and Metlakunta, 2005; Zhao, 2005; Zhao et al., 2002). Leptin and its receptor are structurally and functionally related to the proinflammatory cytokine IL-6 cytokine family (Tartaglia, 1997). Thus, we hypothesized that cAMP-related signal might interfere with leptin signaling pathways and could be involved in central leptin resistance.
Classically, cAMP acts through cAMP-dependent protein kinase (PKA) (Daniel et al., 1998). However, it is unsettled whether or not PKA mediates the inhibitory effects of cAMP on the JAK-STAT3 pathway (Bousquet et al., 2001; Gasperini et al., 2002; Sands et al., 2006). cAMP also acts through Epac (exchange protein directly activated by cAMP), the guanine nucleotide-exchange factor for the small GTPase Rap1(de Rooij et al., 1998; Gloerich and Bos, 2010). Notably, Epac-mediated activation of Rap1 induces SOCS-3 expression in endothelial cells (Sands et al., 2006). However, the molecular pathway linking cAMP to the JAK-STAT3 pathway has not yet been established.
To facilitate the identification of possible signaling pathways that contribute to leptin resistance, we established an in vitro system of leptin action in the hypothalamus. We used an organotypic slice culture system, which enabled a direct assessment and manipulation of candidate molecular pathways inducing leptin resistance within hypothalamic neurons. By employing this system, we investigated whether activation of cAMP-dependent pathways induced leptin resistance in our hypothalamic slice model.
Organotypic slices of mouse hypothalamus were prepared and then maintained at an air-media interface on Millicell-CM filters in MEM base medium for 10 days. We first validated the model system by assessing the effects of leptin treatment of the slices. We found that leptin induced robust phosphorylation of STAT3 (Figures 1A and 1B). In contrast, no STAT3 phosphorylation was observed in the saline-treated slices (Figures 1 A and 1B), as shown by both immunohistochemistry and western blot analysis with an anti-phospho-STAT3 antibody. Notably, leptin-induced phosphorylation of STAT3 was only seen in hypothalamic nuclei such as arcuate nucleus (Arc) and the ventromedial hypothalamus (VMH), known to express leptin receptors (Elmquist et al., 2005; Scott et al., 2009) (Figure 1A). Treatment of organotypic slices with leptin also led to phosphorylation of S6K (Figure 1B), another mediator of leptin action (Cota et al., 2006). We confirmed that leptin-induced STAT3 phosphorylation is mediated via the leptin receptor, by using slices prepared from the leptin receptor null (Leprneo/neo mice) mice (Figures 1D and 1E) (Coppari et al., 2005). Collectively, these results demonstrate that the in vitro system is a model of leptin-induced STAT 3 and S6K activation. We also tested whether this model system was able to recapitulate leptin-induced leptin resistance (Myers et al., 2010). We found that pretreatment of slices with higher levels of leptin (120 nM for 6 hr) strongly impaired leptin-induced pSTAT3 (Figure S1 available online). This indicates that an in vitro system can mimic leptin-induced leptin resistance.
To investigate whether cAMP impairs the leptin signaling pathways in the hypothalamus, we utilized pharmacological reagents to modulate cellular cAMP levels in slices. Interestingly, we found that treatment of the slices with both forskolin (an adenylate cyclase activator; Fsk [20 μM]) plus low level of leptin (0.5 nM)) (Fsk/Lep) had a potent inhibitory effect on leptin-induced STAT3 phosphorylation in the whole hypothalamus (Figures 2A–2D) and in the arcuate nucleus (Figure 2G). This was in contrast to the robust STAT3 phosphorylation observed in control slices (Figure 2A). Forskolin alone clearly blunted leptin-induced STAT3 phosphorylation (Figure 2B), and this effect was enhanced by the presence of low levels of leptin (0.5 nM) (Figure 2B). Treatment of the slices with the low dose of leptin (0.5 nM) alone had little inhibitory effect on leptin-STAT3 signaling as this dose did not elicit STAT3 phosphorylation at either 30 min or 6 hr (Figure 2B).
To further investigate potential signaling pathways involved, we used two different types of phosphodiesterase (PDE) inhibitors known to elevate intracellular cAMP levels: IBMX, a nonspecific inhibitor of PDE, and cilostamide, a selective inhibitor for PDE3. Both PDE inhibitors enhanced the inhibitory effects of forskolin on leptin-induced STAT3 phosphorylation (Figure 2C). We also found that Fsk/Lep also dampened the leptin-induced phosphorylation of S6 Kinase (Figure 2D), a response that is also impaired in the high-fat diet-induced obese rodents (Cota et al., 2008). Notably, Fsk/Lep treatment increased hypothalamic SOCS-3 and PTP1B at the levels of protein and messenger RNA (mRNA) (Figures 2E and 2F), both of which have been demonstrated to contribute to leptin resistance (Bence et al., 2006; Bjørbaek et al., 1999; Cook and Unger, 2002; Howard et al., 2004; Mori et al., 2004; Zabolotny et al., 2002).
We next dissected potential downstream pathways mediating the inhibitory effects of Fsk/Lep on signaling from the leptin receptor. Classically, cAMP exerts many of its effects through protein kinase A (PKA) (Daniel et al., 1998). Thus, we tested whether the inhibitory action of cAMP is mediated in a PKA-dependent manner. Treatment of slices with an inhibitor of PKA, H89, had no effect on cAMP inhibition of leptin-induced phosphorylation of STAT3 and S6K (Figure 2H). Fsk/Lep also induced hypothalamic SOCS-3 in the presence of H 89, a PKA inhibitor (Figure 2I). These results suggest that elevations of cAMP negatively regulate leptin signaling in a PKA-independent manner.
In addition to the PKA pathway, increased cAMP also activates an alternative pathway via Epac, a guanine nucleotide exchange factor, which activates a small G protein Rap1 (de Rooij et al., 1998). We found that Fsk/Lep treatment activated endogenous Rap1 in the hypothalamus (Figure S3). Thus, we next examined whether activation of the Epac-Rap1 pathway is sufficient to evoke an inhibitory effect on leptin receptor signaling. We used a hydrolysis-resistant Epac activator [8-(4-chlorophenylthio)-2′-O-methyladenosine-3′, 5′-cyclic monophosphorothioate, 8-pCPT-2′-O-Me-cAMP] that selectively binds and activates Epac (Enserink et al., 2002). Treatment of the slices with 8-pCPT-2′-O-Me-cAMP plus low level of leptin (8-pCPT-2′-O-Me-cAMP/Lep) impaired leptin-induced phosphorylation of STAT3 (Figures 3A and 3B). Either leptin or 8-pCPT-2′-O-Me-cAMP alone had little effect on leptin-dependent pSTAT3 phosphorylation (Figure 3C). Epac activation also blunted the ability of leptin to modulate other cellular signaling of leptin: phosphorylation of S6K (Figures 3D and 3E) within the hypothalamus. Further, we found that 8-pCPT-2′-O-Me-cAMP induced SOCS-3 and PTP1B (Figures 3F and 3G). This induction occurred in a dose-dependent manner in the presence of low level of leptin, which alone had no effect on induction of either protein (Figure 3H). Collectively, these data suggest that activation of cAMP-Epac/Rap1 impairs hypothalamic leptin receptor signaling.
Recent studies have shown that ciliary neurotrophic factor (CNTF) and leptin have similar anorectic effects (Gloaguen et al., 1997) by modulating similar intracellular signaling cascades. In addition, CNTF can activate leptin-like signaling pathways and can reduce body weight in leptin-resistant obesity (Gloaguen et al., 1997; Lambert et al., 2001). To determine whether the cAMP-Epac pathway is specifically involved in leptin resistance, we assessed whether CNTF induction of STAT3 phosphorylation was blunted by activation of Epac. We found that CNTF caused STAT3 phosphorylation in hypothalamic slices even after Fsk/Lep pretreatment of the slices (Figure S4A). cAMP signaling has been reported to interfere with interleukin 6 (IL-6)-STAT3 signaling in different cell lines (Bousquet et al., 2001; Delgado and Ganea, 2000; Fasshauer et al., 2002; Gasperini et al., 2002; Sands et al., 2006). Thus, we tested whether elevation of cAMP impaired IL6-induced STAT3 phosphorylation in our system. As expected, Fsk/Lep pretreatment blocked IL-6 dependent STAT3 phosphorylation (Figure S4B).
We also employed electrophysiological approach to assess the potential inhibitory effects of cAMP-Epac signaling on leptin’s cellular actions. We assessed the ability of leptin to directly activate pro-opiomelanocortin (POMC) neurons, which are identified targets of leptin. In order to identify POMC neurons for whole-cell patch-clamp recordings, we used POMC-GFP mice (Parton et al., 2007; Ramadori et al., 2008). Whole-cell patch-clamp recordings were performed to assess the effects of leptin on membrane potential. In agreement with previous reports (Cowley et al., 2001; Hill et al., 2008; Williams et al., 2010), leptin caused rapid depolarization from rest in 8 of 12 POMC neurons in organotypic slices (6.4 ± 0.6 mV; resting membrane potential, −51.7 ± 1.3 mV; n = 8; Figures 4A and 4B). We next used organotypic slices from POMC-GFP mice that were pretreated for 6 hr with either Fsk/Lep or 8-pCPT-2′-O-Me-cAMP/Lep. We found that pretreatment with Fsk/Lep prevented the leptin induced depolarization in all POMC neurons examined (0.8 ± 0.3 mV, n = 13; resting membrane potential, −51.3 ± 2.3 mV; n = 13; Figure 4A and 4B). Similarly, leptin failed to depolarize POMC neurons of slices pretreated with 8-pCPT-2′-O-Me-cAMP/Lep (0.6 ± 0.6 mV; resting membrane potential, −50.3 ± 1.0 mV; n = 12; Figures 3A and 3B). Notably, the low dose of leptin alone failed to inhibit the leptin-induced depolarization of POMC neurons (5.8 ± 0.6mV; resting membrane potential, −52.2 ± 1.5mV; Figures 4A and 4B). Additionally, in the presence of tetrodotoxin, which blocks action potential-mediated synaptic transmission, Fsk/Lep still prevented leptin-induced depolarization of POMC neurons (0.5 ± 0.5mV; resting membrane potential, −49.9 ± 1.3mV; Figures 4A and 4B). Collectively, these data suggest that activation of cAMP-Epac signaling directly desensitizes POMC neurons to leptin, which is independent of action potential-mediated synaptic transmission.
To provide further insights to the functional significance of the Epac pathway as it relates to leptin signaling in vivo, we tested whether activation of the Epac pathway impairs the ability of exogenous leptin to inhibit food intake. We performed intracerebroventricular (ICV) infusions of a selective Epac activator (8-pCPT-2′-O-Me-cAMP) in chow fed lean mice. We first identified a dose of 8-pCPT-2′-O-Me-cAMP (5 μg) that did not alter food intake. We next evaluated the ability of ICV leptin injections to suppress food intake following an ICV pretreatment with this dose of the Epac activator. As expected, leptin treatment markedly reduced food intake in mice pretreated with vehicle (Figure 4C). In contrast, in the mice pretreated with the Epac activator leptin did not significantly reduce food intake 4 hr after leptin injection (Figure 4C). These results suggest that the acute anorexic responses to exogenous leptin are blunted by prior central infusions of an activator of the Epac-Rap1 pathway. Finally, we assessed Rap1 activity in the hypothalamus of obese mice after chronic exposure to a high-fat diet. We found that hypothalamic Rap1 activity is increased in the mice fed on high fat diet for 4 weeks (Figure 4D), suggesting that the hypothalamic Epac-Rap1 pathway is activated in the mice exposed to high fat.
Collectively, our in vitro and in vivo results support the model that activation of the cAMP-Epac pathway induces cellular leptin resistance within hypothalamic neurons, a critical site of leptin action. Activation of the cAMP-Epac pathway blunted signaling pathways downstream from the leptin receptor, including JAK-STAT3, mTOR-S6K, ERK, and AMPK. Thus, signaling systems downstream of the leptin receptor are negatively regulated by the cAMP-Epac pathway. Notably, activation of Epac induced SOCS-3 and PTP1B, two negative regulators of leptin signaling in vivo (Banno et al., 2010; Bence et al., 2006; Bjørbaek et al., 1999; Cook and Unger, 2002; Howard et al., 2004; Mori et al., 2004; Zabolotny et al., 2002). Our study is also consistent with a previous report that cilostamide, an inhibitor for phosphodiesterase 3B, inhibited leptin-induced suppression of food intake and STAT3 phosphorylation (Zhao et al., 2002). Taken together, our findings provide evidence linking the cAMP-Epac pathway to central leptin resistance, as seen in obesity.
One implication of our results is that leptin resistance may be induced by a broad range of extracellular stimuli that can activate the cAMP-Epac-Rap1 pathway. For example, any G protein-coupled receptors that act through either Gs or Gi would be a potential target to perturb intracellular cAMP levels (Neves et al., 2002). Thus, we tested whether α-MSH, an agonist of melanocortin receptors MC3R and MC4R impairs leptin-induced pSTAT3, since MC3R and MC4R activate Gs and are expressed in POMC and NPY neurons (Cone, 2005). However, we did not detect an inhibitory effect of α-MSH pretreatment on subsequent leptin signaling in our slices (data not shown). Thus, to date we have not found a “physiological” regulator that drives both activation of the pathway and leptin resistance. However, our findings support the notion that cAMP-Epac signaling may contribute leptin resistance in obesity, as we found elevations of Epac signaling in the hypothalamus of the mice fed with high-fat diet. Clearly, further studies are required to determine whether the Epac-Rap1 pathway is required to maintain leptin sensitivity in vivo. Nonetheless, perturbation of Epac-Rap1 signaling may contribute to the pathophysiology of leptin resistance.
The authors gratefully acknowledge Michelle Choi and Danielle Lauzon for technical assistance, Claudia Vianna and Yong Xu for comments on the manuscript, and Philipp E. Scherer for interpretive assistance. This work was supported by National Institutes of Health grants PL1 DK081182 and UL1 RR024923; by R01DK53301, R01MH61583, RL1DK081185, (to J.K.E.), and by 1F32DK077487 and K01 DK087780 (to K.W.W.); by the American Diabetes Association and a Smith Family Foundation Pinnacle Program Project Award to J.K.E.; and by Scientist Development Grant from the American Heart Association to M.F.
Supplemental Information includes Supplemental Experimental Procedures and four figures and can be found with this article online at doi:10.1016/j. cmet.2011.01.016.