Fas activates lipolysis in a Ca2+-CaMKII-dependent manner in 3T3-L1 adipocytes

: Fas (CD95) is a member of the tumor necrosis factor (TNF) receptor superfamily and plays a crucial role in the induction of apoptosis. However, like TNF, Fas can induce non-apoptotic signaling pathways. We previously demonstrated that mice lacking Fas specifically in adipocytes are partly protected from diet-induced insulin resistance, potentially via decreased delivery of fatty acids to the liver as manifested by lower total liver ceramide content. In the present study we aimed to delineate the signaling pathway involved in Fas-mediated adipocyte lipid mobilization. Treatment of differentiated 3T3-L1 adipocytes with membrane-bound Fas ligand (FasL) significantly increased lipolysis after 12 hours without inducing apoptosis. In parallel, Fas activation increased phosphorylation of ERK1/2 and FasL-induced lipolysis was blunted in the presence of the ERK-inhibitor U0126 or in ERK1/2-depleted adipocytes. Furthermore, Fas activation increased phosphorylation of the Ca2+/calmodulin-dependent protein kinases II (CaMKII) and blocking of the CaMKII-pathway (either by the Ca2+ chelator BAPTA or by the CaMKII inhibitor KN62) blunted FasL-induced ERK1/2 phosphorylation and glycerol release. In conclusion, we propose a novel role for CaMKII in promoting lipolysis in adipocytes. Abstract Fas (CD95) is a member of the tumor necrosis factor (TNF) receptor superfamily and plays a crucial role in the induction of apoptosis. However, like TNF, Fas can induce non-apoptotic signaling pathways. We previously demonstrated that mice lacking Fas specifically in adipocytes are partly protected from diet-induced insulin resistance, potentially via decreased delivery of fatty acids to the liver as manifested by lower total liver ceramide content. In the present study we aimed to delineate the signaling pathway involved in Fas-mediated adipocyte lipid mobilization. Treatment of differentiated 3T3-L1 adipocytes with membrane-bound Fas ligand (FasL) significantly increased lipolysis after 12 hours without inducing apoptosis. In parallel, Fas activation increased phosphorylation of ERK1/2 and FasL-induced lipolysis was blunted in the presence of the ERK-inhibitor U0126 or in ERK1/2-depleted adipocytes. Furthermore, Fas activation increased phosphorylation of the Ca 2+ /calmodulin-dependent protein kinases II (CaMKII) and blocking of the CaMKII-pathway (either by the Ca2+ chelator BAPTA or by the CaMKII inhibitor KN62) blunted FasL-induced ERK1/2 phosphorylation and glycerol release. In conclusion, we propose a novel role for CaMKII in promoting lipolysis in adipocytes.


Introduction
White adipose tissue (WAT) has major metabolic and endocrine functions mediated by secretion of different adipokines and fat-derived metabolites such as NEFAs. These molecules regulate food intake, energy expenditure, and glucose homeostasis (1,2). In obesity, excess WAT accumulation is accompanied by local infiltration of macrophages and other inflammatory cells secreting different cytokines such as IL-1α, IL-1β, IL-6, IL-8 (KC) and MCP-1, which in turn alter the expression and secretion pattern of adipokines, cytokines, and stimulate the release of fatty acids by elevating basal lipolysis. All these changes contribute to the development of detrimental complications of obesity such as insulin resistance and diabetes mellitus (2,3).
Fas (FasR, CD95, Apo-1) is a type I transmembrane protein that belongs to the tumor necrosis factor (TNF)/nerve growth factor (NGF) receptor superfamily (7) and is activated by Fas ligand (FasL, CD95L), a type II membrane protein (8). Fas was first described in 1989 as a surface molecule on lymphocytes that can trigger cell death (9). In the adult mouse, Fas and FasL are expressed in several tissues including WAT (10,11). Upon binding of FasL, preformed trimeric Fas complexes undergo a conformational change that results in the formation of a death-inducing signaling complex (DISC) and activation of downstream pathways leading to apoptosis (9,12). However, in addition to this well-established role of Fas in apoptosis, Fas activation contributes to non-apoptotic signaling pathways, including cell proliferation (9,13) and the induction of inflammatory responses in different cell types (14)(15)(16)(17)(18). Moreover, we have recently reported that Fas is increasingly expressed in WAT of obese subjects and that Fas-deficient and adipocyte-specific Fas knockout mice are partly protected from high fat diet-induced insulin resistance (11,19) implicating a role for Fas in the pathogenesis of obesity-associated insulin resistance. While the underlying mechanism remained incompletely understood, livers of adipocyte-specific Fas-KO mice had lower levels of total ceramides, which are potentially metabolized from NEFA delivered to the liver from adipose tissue.
Thus, in the present study we hypothesized that in adipocytes Fas activation can, independently of its pro-apoptotic effects, directly induce the hydrolysis of triglycerides (i.e., increase basal lipolysis), and set-up to investigate the intracellular signaling pathway(s) involved.

Cell Culture
3T3-L1 adipocytes were cultured in Dulbecco's modified Eagle's medium (DMEM, Invitrogen, Basel Switzerland) containing 25 mM glucose (high glucose), supplemented with 10% fetal calf serum (FCS, Socochim SA, Lausanne) and antibiotics (Invitrogen). 48 hours after reaching confluence (day 0, D0), cells were treated with a mixture of 500 μM methylisobutylxanthine, 1 μM dexamethasone, 1.7 μM insulin (all from Sigma, Buchs, Switzerland) and 1 μM rosiglitazone (Alexis Biochemicals) to induce differentiation. Two days later (D2) the medium was changed to high glucose culture medium containing insulin (0.5 μM). Another 2 days later (D4), the medium was replaced by culture medium without insulin. The culture medium was replaced every other day and changed to culture medium containing 5.5 mM glucose (low glucose) after 4 days (D8). Cells were kept at least 2 days on low glucose before experiments were performed. Membrane-bound Fas ligand (Upstate, Lake Placid, NY, USA) was added to low glucose serum-free medium as indicated.
For pretreatment experiments with rosiglitazone (Enzo Life Science, Lausen, Switzerland), mature adipocytes were incubated for 48 hours with 5 μM of the compound in low glucose medium. Thereafter, FasL together with rosiglitazone was applied for another 6 or 12 hours.

Data analysis
Data are presented as means ± SEM and were analyzed by a one sample t test or analysis of variance (ANOVA) with a Newman-Keuls multiple comparison test. *p < 0.05, **p < 0.01, ***p < 0.001.

Fas activation induces lipolysis in 3T3-L1 adipocytes
The Fas receptor (CD95) is expressed in 3T3-L1 preadipocytes, decreases during adipocyte differentiation but is still clearly expressed in mature adipocytes ( Fig. 1). Expression levels of the adipocyte specific proteins PPARγ 2 , C/EBPα and perilipin reflect the respective stage of differentiation. Treatment of differentiated 3T3-L1 adipocytes with 2 ng/ml FasL for 12 hours significantly increased lipolysis (Fig. 2), consistent with our earlier observation (8), and without affecting their viability as assessed by TUNEL assay (supplemental Fig. 1) and MTT determination (11).

Fas activation increases phosphorylation of HSL
Beta adrenergic receptor agonists such as catecholamines stimulate lipolysis via adenylate cyclase-dependent activation of PKA and consecutive activation of HSL, perilipin 1 and ATGL. Inhibiting such effect, insulin activates phosphodiesterase 3, which converts cAMP to 5'-AMP, thereby diminishing cAMP-mediate PKA activity, which results in inhibition of lipolysis. In order to examine whether Fas-mediated lipolysis comprise activation of PKA, phosphorylation of PKA substrates was determined in 3T3-L1 adipocytes. As expected, the β 1,2 -receptor agonist isoproterenol increased phosphorylation of PKA substrates significantly. In contrast, treatment with FasL had no effect on the abundance of phosphorylated PKA substrates (Fig. 3A). However, even though not detected by the PKA substrate antibody, incubation of 3T3-L1 adipocytes with FasL for 6 and 12 hours significantly increased phosphorylation of HSL at Ser563 (Fig. 3B), while it had no effect on total HSL protein levels (supplemental Fig. 4) and phosphorylation of perilipin (Fig. 3B). In addition, ATGL protein levels were slightly but not significantly up regulated upon Fas incubation (Fig. 3C). Thus, Fas-mediated lipolysis may depend on activation of HSL and/or ATGL.

Fas-mediated lipolysis is ERK-dependent
An alternative signaling pathway to activate lipolysis in adipocytes involves the p44/42 MAP kinases (ERK1/2), as was shown for TNFα (21). Since Fas belongs to the tumor necrosis factor receptor superfamily, we postulated that FasL-induced lipolysis is mediated via ERK1/2 activation. Incubation of mature 3T3-L1 adipocytes with 2 ng/ml FasL increased phosphorylation of ERK1/2 significantly after 6 and 12 hours whereas total protein concentration of ERK1/2 was not affected (Fig. 4A). To exclude the possibility that the effects of FasL treatment were due to a FasL-mediated increase in TNFα secretion and, thus, to a paracrine regulatory loop mediated by this cytokine, TNFα concentration was determined in the supernatant.
As depicted (supplemental Fig. 5), incubation with 2 ng/ml FasL for 12 hours did not lead to increased TNFα secretion to the medium. We therefore concluded that Fasmediated ERK1/2 activation was independent of TNFα.
To assess whether Fas-induced lipolysis is dependent on ERK1/2 activation, we incubated 3T3-L1 cells in presence or absence of the MEK1/2 inhibitor U0126.
To further corroborate a Fas-ERK1/2-lipolysis pathway in adipocytes, we tested whether PPARγ agonists such as thiazolidinediones (TZDs), which were demonstrated to inhibit whole body lipolysis in patients with type 2 diabetes (22), can inhibit Fas-induced lipolysis and if so, whether activation of ERK1/2 was also diminished. As shown in Fig. 4D, FasL-stimulated ERK1/2 phosphorylation was reduced in the presence of rosiglitazone without affecting total ERK1/2 protein content. Correspondingly, rosiglitazone also reduced FasL-mediated lipolysis (Fig.   4D).

Fas-mediated lipolysis is CaMKII-dependent
CaMKII was previously shown to mediate magnolol-triggered lipolysis in sterol ester-loaded 3T3-L1 preadipocytes in an ERK1/2 dependent fashion (23). We therefore postulated that FasL-induced lipolysis may be dependent on intracellular changes in calcium levels since it is well accepted that Fas activation can increase intracellular calcium levels (24). Indeed, preincubation of 3T3-L1 adipocytes with the intracellular calcium chelator BAPTA/AM prevented both FasL-induced ERKactivation and lipolysis (Fig. 5A), suggesting that both effects of Fas activation are dependent on an intracellular calcium rise. In response to the latter the developing calcium/calmodulin complex may activate CaMKII leading to intramolecular autophosphorylation at several sites including Thr286, Thr305 and Thr306 (25). As depicted in Fig. 5B, incubation of mature 3T3-L1 adipocytes with 2 ng/ml FasL increased phosphorylation of CaMKII at Thr286 significantly after 6 and 12 hours.
Moreover, pre-treatment with the CaMKII inhibitor KN62 reduced FasL-mediated ERK1/2 activation as well as lipolysis (Fig. 5C). These data strongly suggest that Camediated activation of CaMKII is a proximal signaling response to Fas activation, which is propagated further downstream via ERK1/2 to induce basal lipolysis.

Discussion
In obesity, higher basal lipolysis rate resulting in increased release of NEFAs into the circulation contributes to the development of hepatic and total body insulin resistance (26). Several factors contribute to such increase in lipolysis. First, obesity is associated with a persistent low grade inflammation of adipose tissue as manifested by an increased production and secretion of pro-inflammatory cytokines such as TNFα and IL-6 (27). The latter in turn can directly stimulate adipocyte lipolysis even as isolated factors (21,28,29). Second, hypertrophic adipocytes are characterized by an elevated rate of basal lipolysis (30), which might be at least partly mediated by self-production of inflammatory cytokines acting in an autocrine manner, possibly as a self-protective cellular mechanism against excessive cellular over-growth. Third, since insulin is the major anti-lipolytic hormone, insulin resistance at the adipocyte level results in increased lipolysis, creating a vicious cycle between hypertrophy, inflammation, lipolysis, and insulin resistance. Fas may be a key component of such dys-regulation: Fas is activated by FasL, which can be produced by inflammatory cells infiltrating adipose tissue in obesity. Moreover, expression of Fas is increased in adipose tissue of obese humans and in isolated adipocytes of obese and diabetic mice, and intriguingly, its protein expression correlates with adipocyte size and is therefore increased in hypertrophic adipocytes (11).
In agreement with such notion, we show herein that chronic stimulation of the death receptor Fas induced lipolysis in 3T3-L1 adipocytes. Importantly, this effect occurred under conditions that did not induce apoptosis (11). Like TNFα, FasL stimulates lipolysis through activation of the ERK1/2 MAP kinases since Fas activation lead to increased phosphorylation of ERK1/2 and pre-treatment with the MEK1/2 inhibitor U0126 or siRNA-mediated downregulation of ERK1/2 blocked Fasinduced lipolysis. In the case of TNFα it was proposed that ERK1/2-dependent downregulation of the lipid droplet coating protein perilipin is responsible for the increase in lipolysis (21). We also observed a down-regulation of perilipin in cells treated with FasL for 12 hours. However, in contrast to TNFα such effect was not mediated by ERK1/2, since FasL-induced decrease in perilipin expression was not prevented by ERK-inhibition (supplemental Fig. 6). Moreover, 6 hours of FasL incubation did not decrease perilipin protein content (data not shown) but increased FFA and glycerol release. All these results suggest that Fas-mediated lipolysis is independent of a decrease in perilipin expression.
Herein, we present evidence for calcium-triggered activation of ERK in Fas activation-mediated lipolysis. Fas activation in cells was previously shown to raise intracellular free Ca 2+ levels (24). Similarly, increased intracellular Ca 2+ levels induced by endoplasmic reticulum (ER) stress were shown to induce lipolysis in adipocytes ERK-dependently (31). Hence, Fas-induced ERK activation may be mediated by ER stress-triggered Ca 2+ release. However, as shown in supplemental Moreover, the extracellular Ca 2+ -chelator EDTA blunted Fas-induced ERK activation and lipolysis similar to the intracellular chelator BAPTA (supplemental Fig. 8). Such data suggest that extracellular Ca 2+ -influx rather than Ca 2+ -release from ER is involved in Fas induced lipolysis. Increased intracellular Ca 2+ is bound by the calcium-binding protein calmodulin (CaM) forming a complex. The latter then binds to and thereby activates CaMKII. Activation of CaMKII by the Ca 2+ /CaM complex allows intramolecular autophosphorylation at several sites including Thr286. This generates calcium-independent activity that persists after dissociation of calcium/calmodulin allowing transient calcium elevation to promote prolonged kinase activation (25). We found that FasL treatment increased phosphorylation at Thr286 of CaMKII. Moreover, Fas activation-induced lipolysis was prevented in the presence of the CaMKIIinhibitor as well as of the Ca 2+ -chelator BAPTA. Thus, we postulate that Fas activation in adipocytes increases intracellular Ca 2+ levels leading to activation of CaMKII, which in turn activates ERK1/2 and, thus, lipolysis. Accordingly, it was previously reported that trans-10, cis-12 conjugated linoleic acid-induced ERK1/2 activation in adipocyte is dependent on a rise in intracellular free Ca 2+ levels and consecutive activation of CaMKII (32). Unfortunately, lipolysis was not addressed in this study. The data presented herein may suggest a novel pathway for lipolysis in adipocytes via CaMKII-dependent activation of ERK. Interestingly, such pathway may be very ancient and evolutionarily preserved since CaMKII-mediated release of free fatty acids was recently demonstrated to play a role in pheromone biosynthesis in insects (33).
Interestingly, the TZD rosiglitazone reduced both Fas-mediated ERK1/2 phosphorylation and lipolysis. PPARγ agonists have profound effects on adipocyte metabolism and were shown to improve insulin sensitivity in patients with type 2 diabetes. Moreover, in a recent paper, treatment with another TZD, pioglitazone, was able to reduce whole body lipolysis in type 2 diabetic patients (22). We previously described a potential role for Fas activation in obesity-associated insulin resistance (11) and show herein that Fas activation leads to increased lipolysis. Thus, our experiments may point to a role of TZDs in counteracting Fas-induced metabolic changes in adipocytes and further underscore the importance of ERK-activation for FasL-induced lipolysis. Similarly, TZDs were previously found to antagonize the effects of TNFα on adipocytes (34).
We recently showed that treatment of 3T3-L1 adipocytes with 2ng/ml FasL for 12 hours reduced protein levels of Akt/PKB (protein kinase B) (19). Moreover, a recent publication reported that deletion of rictor, an essential component of the Akt kinase mammalian target of rapamycin complex 2 (mTORC2), increases basal lipolysis as well as phosphorylation of HSL at Ser563 in adipocytes (35). Hence, it is conceivable that Fas mediated reduction in Akt protein levels contributed to FasLinduced lipolysis, potentially via increased phosphorylation of HSL at serine residue 563. Moreover, ERK1/2-mediated phosphorylation of HSL at Ser600 (36) might contribute to Fas-induced lipolysis. However, such corresponding anti-phospho-HSLSer600 antibody is not commercially available and, thus, not available to us for evaluating a potential role of HSLSer600 phosphorylation in this study. Besides activation of HSL, we cannot exclude the involvement of ATGL to Fas activationinduced lipolysis since ATGL protein levels were slight albeit not significantly increased after FasL treatment.
In summary, activation of the Fas receptor alters the metabolism of adipocytes and leads to ERK1/2-mediated lipolysis, which may be triggered by Fas-induced increase in intracellular calcium levels and hence, autophosphorylation of the Ca 2+ /calmodulin-dependent protein kinases II. Thus, our findings suggest an important role of the Fas receptor in the development of adipose tissue dys-function in the context of obesity.