Cellular Signalling 26 (2014) 323–331

Contents lists available at ScienceDirect

Cellular Signalling journal homepage: www.elsevier.com/locate/cellsig

Akt and Rac1 signaling are jointly required for insulin-stimulated glucose uptake in skeletal muscle and downregulated in insulin resistance Lykke Sylow a,⁎, Maximilian Kleinert a, Christian Pehmøller a, Clara Prats b, Tim T. Chiu c, Amira Klip c, Erik A. Richter a, Thomas E. Jensen a a b c

Molecular Physiology Group, Department of Nutrition, Exercise and Sports, August Krogh Centre, University of Copenhagen, Denmark Department of Biomedical Sciences, Center of Healthy Aging, University of Copenhagen, Denmark Program in Cell Biology, The Hospital for Sick Children, Toronto, Canada

a r t i c l e

i n f o

Article history: Received 14 October 2013 Accepted 1 November 2013 Available online 9 November 2013 Keywords: Insulin resistance Type 2 diabetes Glucose metabolism Insulin signaling Actin cytoskeleton GLUT4

a b s t r a c t Skeletal muscle plays a major role in regulating whole body glucose metabolism. Akt and Rac1 are important regulators of insulin-stimulated glucose uptake in skeletal muscle. However the relative role of each pathway and how they interact are not understood. Here we delineate how Akt and Rac1 pathways signal to increase glucose transport independently of each other and are simultaneously downregulated in insulin resistant muscle. Pharmacological inhibition of Rac1 and Akt signaling was used to determine the contribution of each pathway to insulin-stimulated glucose uptake in mouse muscles. The actin filament-depolymerizing agent LatrunculinB was combined with pharmacological inhibition of Rac1 or Akt, to examine whether either pathway mediates its effect via the actin cytoskeleton. Akt and Rac1 signaling were investigated under each condition, as well as upon Akt2 knockout and in ob/ob mice, to uncover whether Akt and Rac1 signaling are independent and whether they are affected by genetically-induced insulin resistance. While individual inhibition of Rac1 or Akt partially decreased insulin-stimulated glucose transport by ~40% and ~ 60%, respectively, their simultaneous inhibition completely blocked insulin-stimulated glucose transport. LatrunculinB plus Akt inhibition blocked insulin-stimulated glucose uptake, while LatrunculinB had no additive effect on Rac1 inhibition. In muscles from severely insulin-resistant ob/ob mice, Rac1 and Akt signaling were severely dysregulated and the increment in response to insulin reduced by 100% and 90%, respectively. These findings suggest that Rac1 and Akt regulate insulin-stimulated glucose uptake via distinct parallel pathways, and that insulin-induced Rac1 and Akt signaling are both dysfunctional in insulin resistant muscle. There may thus be multiple treatment targets for improving insulin sensitivity in muscle. © 2013 Elsevier Inc. All rights reserved.

1. Introduction Abnormalities in the regulation of skeletal muscle glucose uptake lead to insulin resistance at the whole body level, and may ultimately precipitate type 2 diabetes [1]. Glucose uptake into skeletal muscle occurs by translocation of glucose transporter 4 (GLUT4) from intracellular compartments to the plasma membrane and to transverse tubuli (2– 5). The underlying signaling mechanism regulating this process is not well understood but includes activation of Akt2 and the Rho GTPase, Rac1. Understanding these pathways and how they act on glucose transport is essential in order to identify novel molecular targets for Abbreviation: AS160, Akt substrate 160; EDL, extensor digitorum longus; FWHM, Full Width at the Half Maximum; GAP, GTPase activating protein; KO, knockout; PAK1, p21 protein-activated kinase 1; PI3K, phosphatidylinositol 3-kinase; Rac1, Ras-related C3 botulinum toxin substrate 1; Rac1 InhibII, Rac1 Inhibitor II; WT, wild-type; 2DG, 2 deoxy-glucose. ⁎ Corresponding author at: Universitetsparken 13, 2100 Copenhagen O, Denmark. Tel.: +45 20955250; fax: +45 35321600. E-mail address: lshansen@ifi.ku.dk (L. Sylow). 0898-6568/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cellsig.2013.11.007

pharmacological drug development in the treatment of insulin resistant states. Insulin-mediated activation of Akt2 and Rac1 occurs downstream of the lipid kinase phosphatidylinositol 3-kinase (PI3K). Activation of Akt leads to phosphorylation of AS160 (also known as TBC1D4) [2] and stimulation of Rab proteins [3]. Rac1 stimulates actin cytoskeleton reorganization [4] and activates PAK [5] by releasing PAK from its autoinhibitory domain which allows for phosphorylation of Thr423/ Thr402 (PAK1/2) within the activation loop [6,7]. Rac1-dependent actin remodeling is necessary for insulin to induce GLUT4 translocation in L6 myotubes [8,9]. Rac1 seems to exclusively regulate GLUT4 translocation in muscle cells, since GLUT4 translocation in adipocytes is not affected by expression of a dominant negative Rac1 [10]. Interestingly, dominant-negative mutation of Akt decreases insulin-stimulated GLUT4 translocation, without affecting actin remodeling in L6 myotubes [11], pointing to distinct downstream effects of the Rac1 and Akt pathways. Furthermore, knockdown of Rac1 by siRNA [12] or knockout of Rac1 [6] does not affect insulin-stimulated Akt phosphorylation, even though it abolishes insulin-induced GLUT4 translocation [8,13,14].

324

L. Sylow et al. / Cellular Signalling 26 (2014) 323–331

Based on these findings obtained in cell culture, we hypothesized that the insulin signaling pathway bifurcates downstream of PI3K into a ‘Rac1-arm’ and an ‘Akt-arm’, and that these two pathways act on glucose transport by regulating distinct steps in the GLUT4 translocation process. Although the Rac1 and Akt pathways seem to act on GLUT4 translocation through distinct downstream processes, a number of studies point to cross talk or feedback loops between the Rac1 and the Akt signaling pathways. For example, basal, but not insulin-stimulated, Akt activity is required for constitutively active Rac1 to induce GLUT4 translocation in L6 myotubes [13]. Furthermore, Rac1 [15], as well as Rac1's downstream target PAK [16,17], has been shown to act as scaffolds for Akt to facilitate interaction of Akt with its regulators. The actin cytoskeleton is also a critical signaling complex in distinct regions within the cells [18], and in L6 muscle cells, PI3K and Akt redistribute to the remodeled cortical actin cytoskeleton in response to insulin [3,9,19]. Thus, the actin cytoskeleton also contributes to the regulation of the Akt and Rac1 signaling cascades. Moreover, interfering with actin remodeling prevents GLUT4 translocation [20]. Some groups report that insulin resistant skeletal muscle has normal Akt regulation [21–23], while others show a partial reduction in Akt activity [24,25]. We recently documented that insulin-induced Rac1 signaling via PAK was significantly decreased in skeletal muscle of obese type 2 diabetic patients [6]. Insulin-resistant skeletal muscles also display altered actin cytoskeleton [26,27]. Together, these findings suggest that the Akt- and Rac1-signaling cascades both contribute to the impaired glucose uptake observed in insulin-resistant muscles, and underscore the importance of unraveling the individual contributions of these two signaling ‘arms’ in search of novel molecular drug targets in the treatment of insulin resistant states. Therefore our aim was to examine the relative contributions of the Rac1 and the Akt signaling ‘arms’ in the regulation of insulin-stimulated glucose uptake in mature skeletal muscle. Furthermore, we analyzed potential cross talk between these two pathways. We hypothesized that Rac1 and Akt act on glucose transport via distinct parallel signaling pathways which are both dysregulated in insulin resistant mature skeletal muscles. 2. Methods 2.1. Animals Female C57BL/6 mice (Taconic, Denmark), 12–16 weeks old, were used for all inhibitor experiments. Four male and two female Akt2 knockout (KO) mice, plus wild-type mice on a C57BL/6N background (10–12 weeks of age) [28] were used. Akt KO mice were a kind gift from Dr Laurie Goodyear (Joslin Diabetes Center and Department of Medicine, Harvard Medical School, Boston, USA). Ob/ob mice and lean control mice were purchased from Charles River (Germany) and were used for experiments at 14 weeks of age. All animals were maintained on a 12:12-h light-dark cycle, and fed a standard rodent chow diet (Altromin no. 1324; Chr. Pedersen, Denmark), and had access to water ad libitum. All experiments were approved by the Danish Animal Experimental Inspectorate, and complied with the “European Convention for the Protection of Vertebrate Animals Used for Experiments and Other Scientific Purposes.” 2.2. Muscle incubations Soleus and extensor digitorum longus (EDL) muscles were dissected from 2 h fasted, anesthetized (6 mg pentobarbital sodium/100 g body wt) mice and were suspended at resting tension (2–4 mN) in incubation chambers (Multi Myograph system; Danish Myo-Technology, Denmark) in Krebs–Ringer–Henseleit buffer (KRB) with 2 mM pyruvate, and 8 mM mannitol at 30 °C, as described previously [29]. For inhibitor experiments, the muscles were pre-incubated for 1 h in KRB buffer with Wortmannin (WM; 0.5 μM; Sigma Aldrich), Rac1 Inhibitor II

(Rac1 InhibII; 10 μM; Calbiochem), MK2206 (20 μM; SelleckChem), LatrunculinB (5 μM; Cytoskeleton) or a corresponding amount of DMSO as vehicle control. Following the pre-incubation period, muscles were stimulated with insulin (60nM) for 30 min. 2.3. 2-Deoxyglucose (2DG) uptake 2DG uptake was measured with 1 mM 2DG during the last 10 min of the insulin stimulation period using 3H 2DG and 14C mannitol tracers as described [29]. 2.4. Muscle analyses Immediately following insulin stimulation, muscle tissue was quickly frozen in liquid nitrogen and stored at −80 °C. Tissue was homogenized 2 × 1 min at 30 Hz using a Tissuelyser II (Qiagen, USA) in 50 mM HEPES (pH 7.5), 150 mM NaCl, 20 mM sodium pyrophosphate, 20 mM β-glycerophosphate, 10 mM NaF, 2 mM sodium orthovanadate, 2 mM EDTA, 1% NP-40, 10% glycerol, 2 mM PMSF, 1 mM MgCl2, 1 mM CaCl2, 10 μg/ml leupeptin, 10 μg/ml aprotinin, and 3 mM benzamidine. After rotation end-over-end for 1/2 h, lysate supernatants were collected by centrifugation (13,000 ×g) for 20 min at 4 °C. 2.5. Immunoblotting Lysate protein concentrations were measured using the bicinchoninic acid (BCA) method, with BSA standards (Pierce) and BCA assay reagents (Pierce) in triplicates. Total protein and phosphorylation levels of relevant proteins were determined by standard immunoblotting techniques, with equal amounts of protein loaded per well. The primary antibodies used were Akt2, p-AktSer473/Thr308, GLUT4, phospho-Akt Substrate (PAS;p-AS160), actin, PAK1, p-PAK1/2Thr423/402 (Cell Signaling Technology), and Rac1 (Cytoskeleton). Polyvinylidene difluoride membranes (Immobilon Transfer Membrane; Millipore) were blocked for 1/2 h in TBS–Tween 20 plus 2% skim milk or 5% BSA protein, at room temperature. Membranes were incubated with primary antibodies overnight at 4 °C, followed by incubation with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. Bands were visualized using the BIORAD ChemiDoc™ MP Imaging System plus enhanced chemiluminescence (ECL+; Amersham Biosciences). 2.6. Immunohistochemistry on single fibers Basal and insulin-stimulated EDL muscles from lean or ob/ob mice were immersed in KRB buffer containing procaine hydrochloride (35 mg/10 ml) for 5 min. Right after, muscles were fixed by immersion in 2% formaldehyde 0.15% picric acid for 30 min at room temperature, followed by 3.5 h at 4 °C. After isolation of a minimum of 20 single muscle fibers per muscle, immunostaining of Rac1 was performed as described [30]. Briefly, EDL fibers were incubated overnight with an anti-Rac1 antibody (Novus Biologicals) in immunobuffer containing 2% saponin and, after 3 washes with immunobuffer, single EDL fibers were incubated with a secondary antibody conjugated with Alexa Fluor 488 (Invitrogen, UK). The muscle fibers were mounted in Vectashield mounting medium. Confocal images were collected with a Zeiss LSM710 microscope, through a 63 ×/1.4 NA oil DIC PlanApochromat objective at 20 °C. Images were analyzed using Zeiss Zen (2010) software. 2.7. Detection of actin remodeling in L6 myotubes L6 myoblasts were seeded and grown in aMEM + 10% FBS on coverslips until 80% confluency before being switched to 2% FBS differentiation media. The cells were cultured in differentiation media for 7 days with media change every 2 days to allow differentiation into myotubes.

L. Sylow et al. / Cellular Signalling 26 (2014) 323–331

At the end of differentiation, cells were serum starved for 3 h followed by 10 min of 100 nM insulin stimulation. Cells were subsequently washed 2 × with PBS, fixed with 3% PFA for 20 min, quenched with 0.1 M glycine for 10 min, permeabilized with 0.1% TX-100 for 3 min, and blocked with 5% milk for 15 min before being stained with rhodamine-phalloidin to detect F-actin. For experiments with inhibitors, myotubes were pretreated with either DMSO, 10 μM Rac1 inhibitor II, or 15 μM MK2206 for 40 min at the end of 3 h serum starvation before subjecting to 10 min insulin stimulation (100nM) in the presence or absence of the inhibitors. For experiments with FLAG–AS160-4A expression, FLAG–AS160-4A was transfected in myoblasts at day one using Fugene HD (Promega) according to the manufacture's protocol prior to myotube differentiation. Fluorescence images were taken with a Zeiss LSM 510 confocal microscope using an Apochromat 63x/1.4 NA oil objective. 2.8. Statistical analyses Results shown are means ± SEM. Statistical testing was performed using paired t-tests or one- or two-way ANOVA (repeated or non repeated measurements) as appropriate. Tukey's post hoc test was performed when ANOVA revealed significant interaction. Statistical evaluation was performed using Sigmaplot 11.0. The significance level was set at p b 0.05. 3. Results 3.1. Insulin-stimulated glucose uptake relies on both Rac1 and Akt activation To investigate the specific individual contribution of Rac1 and Akt in the regulation of insulin-stimulated glucose uptake, we applied pharmacological inhibitors of PI3K (Wortmannin), Rac1 (Rac1 inhib II), and Akt (MK2206), and measured insulin-stimulated 2DG transport in isolated ex vivo incubated soleus and EDL muscles. As expected, WM completely blocked 2DG transport in response to insulin in both soleus and EDL, while inhibition of Rac1 reduced insulin-stimulated 2DG transport by 40% in both muscles (Fig. 1A and B). In soleus and EDL, the Akt inhibitor MK2206 decreased insulin-stimulated glucose uptake by 70% and 60%, respectively. Interestingly, when both Rac1 and Akt were inhibited, insulin-stimulated 2DG transport was completely blocked in soleus (Fig. 1A) and decreased by 90% in EDL (Fig. 1B), mimicking the effect of PI3K-inhibition by WM. This suggests that the Rac1- and Akt-activated pathways regulate 2DG transport through

325

complementary pathways and are necessary for most if not all insulin-stimulated glucose transport. 3.2. Rac1 and Akt are activated in parallel pathways downstream of PI3K The literature on the involvement of Akt in the regulation of Rac1 is inconclusive and such regulation has not been investigated in differentiated skeletal muscle. To analyze the activity of Rac1, we used a p-PAK1/2Thr423/402 antibody recognizing the phosphorylation sites responsible for activation of PAK, since activated GTP bound Rac1 induces this phosphorylation [7]. We found that p-PAK1/2Thr423/402 was abolished by WM but was unaffected by Akt inhibition, suggesting that Rac1 signaling is activated downstream of PI3K but not of Akt (Fig. 2A–C). Interestingly, we observed only a modest decrease in insulin-stimulated phosphorylation of AS160 by the Akt inhibitor, despite a significant decrease in p-AktSer473/Thr308 (Fig. 2A–C) and 2DG transport (Fig. 1A and B), whereas WM caused a complete suppression of AS160 phosphorylation. This is in agreement with a recent study showing that only activation of a small portion of Akt suffices to induce full activation of downstream targets [31]. Because Rac1 and PAK have been reported to be involved in activation of Akt in various cell types [15–17], we analyzed p-AktSer473/Thr308 when Rac1 was inhibited. Inhibition of Rac1 did not affect insulininduced p-AktSer473 or p-AktThr308 (Fig. 2A–C), suggesting that PAK1 activation is not necessary for insulin to induce phosphorylation of Akt. This is consistent with our previous findings showing no effect of Rac1 knockout on Akt signaling in muscle [6]. 3.3. Rac1, but not Akt, regulates insulin-stimulated 2DG transport via the actin cytoskeleton Rac1 is a regulator of the actin cytoskeleton [4]. To investigate if Rac1 regulates insulin-stimulated 2DG transport by acting on the actin cytoskeleton, we added an actin cytoskeleton-depolymerizing agent, LatrunculinB, on top of the Rac1 InhibII to soleus and EDL muscles. We hypothesized that Rac1 inhibition and depolymerization of the actin cytoskeleton would each partially inhibit insulin-stimulated glucose uptake, with no additive effect of adding the two inhibitors together. On the contrary, adding LatrunculinB on top of MK2206 would have an additive effect because Akt and the actin cytoskeleton act on insulinstimulated glucose uptake via distinct parallel pathways. Consistent with our previous findings, Rac1 InhibII decreased insulin-stimulated glucose uptake by 40% and 50% in soleus and EDL, respectively. MK2206 inhibited glucose uptake in soleus by 70% (Fig. 3A) and by

Fig. 1. A) Insulin-stimulated (60 nM, 30 min) 2DG transport in incubated soleus mouse muscles −/+ 0.5 μM Wortmannin (WM), 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206, 10 μM Rac1 inhibitor II + 15 μM MK2206, or a corresponding amount of DMSO, 1 h pre-incubation. B) Insulin-stimulated (60 nM, 30 min) 2DG transport in incubated soleus mouse muscles −/+ 0.5 μM Wortmannin (WM), 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206, 10 μM Rac1 inhibitor II + 15 μM MK2206, or a corresponding amount of DMSO, 1 h preincubation. Statistical significance between basal and insulin is indicated by */*** (p b 0.05/0.001). Statistical significant effects of the inhibitors compared to DMSO control on insulinstimulated 2DG transport are indicated by ### (p b 0.001). Statistical significant effects of inhibitors are indicated by †/††/††† (p b 0.05/0.01/0.001). Values are mean ± S.E.M. n = 6–8.

326

L. Sylow et al. / Cellular Signalling 26 (2014) 323–331

Fig. 2. A) Bar graphs showing quantifications of p-PAKThr423, p-AktSer473, p-AktThr308 and p-AS160 in soleus and EDL −/+ 0.5 μM Wortmannin (WM), 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206, 10 μM Rac1 inhibitor II + 15 μM MK2206, or a corresponding amount of DMSO, 1 h pre-incubation. B) Representative blots showing insulin-stimulated signaling in soleus −/+ 0.5 μM Wortmannin (WM), 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206, 10 μM Rac1 inhibitor II + 15 μM MK2206, or a corresponding amount of DMSO. C) Representative blots showing insulin-stimulated signaling in soleus −/+ 0.5 μM Wortmannin (WM), 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206, 10 μM Rac1 inhibitor II + 15 μM MK2206, or a corresponding amount of DMSO. Statistical significant effects of the inhibitors on signaling are indicated by #/##/### (p b 0.05/0.01/0.001). Statistical significant difference between basal and insulin is indicated by */**/*** (p b 0.05/0.01/0.001). Values are mean ± S.E.M. n = 6–8.

60% in EDL (Fig. 3B). The effect of LatrunculinB was similar to that of Rac1 InhibII, and decreased insulin-induced glucose transport by 40% in both soleus and EDL. We detected no additive inhibitory effect of

Rac1 InhibII with LatrunculinB, suggesting that Rac1 acts on glucose uptake via its regulatory role of the actin cytoskeleton. In contrast, MK2206 plus LatrunculinB decreased insulin-stimulated glucose uptake by 90%

Fig. 3. A) Insulin-stimulated (60 nM, 30 min) 2DG transport in incubated soleus mouse muscles −/+ 15 μM MK2206, 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206 + 5 μM LatrunculinB (LatB), 10 μM Rac1 inhibitor II + 5 μM LatrunculinB, 5 μM LatrunculinB, or a corresponding amount of DMSO, 1 h pre-incubation. B) Insulin-stimulated (60 nM, 30 min) 2DG transport in isolated incubated EDL mouse muscles −/+ 15 μM MK2206, 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206 + 5 μM LatrunculinB (LatB), 10 μM Rac1 inhibitor II + 5 μM LatrunculinB, 5 μM LatrunculinB, or a corresponding amount of DMSO, 1 h pre-incubation. Statistical significant difference between basal and insulin is indicated by **/*** (p b 0.01/0.001). Statistical significant effects of the inhibitors compared to DMSO control on insulin-stimulated 2DG transport are indicated by ### (p b 0.001). Statistical significant effects of inhibitors are indicated by (†)/†/††† (p b 0.1/0.05/0.001). Values are mean ± S.E.M. n = 6–14.

L. Sylow et al. / Cellular Signalling 26 (2014) 323–331

in soleus (Fig. 3A) and completely abolished this increase in EDL (Fig. 3B). 3.4. Activation of Rac1 and Akt does not rely on an intact actin cytoskeleton in mature skeletal muscle The actin cytoskeleton is an important scaffolding structure and is involved in insulin-stimulated redistribution [3,9,19] and activation [32,33] of Akt and PI3K [9] in cultured muscle cells. However, the involvement of the actin cytoskeleton in insulin signaling on PI3K or the Akt and Rac1 signaling pathways has not been investigated in mature skeletal muscle. Interestingly, LatrunculinB did not affect any of the measured signaling proteins in either soleus or EDL muscles (Fig. 4A–C), despite significantly decreasing insulin-stimulated glucose transport (Fig. 3A and B). These data suggest that the impact of LatrunculinB in mature skeletal muscle differs from that in cultured muscle cells, and/or that in mature muscle, the actin cytoskeleton does not act as a necessary scaffolding structure for insulin-stimulated signaling proteins, but is nonetheless important for insulin-stimulated glucose transport. 3.5. Knockout of Akt2 prevents the insulin-stimulated increment in phosphorylation of PAK To further investigate the role of Akt2 in the regulation of the Rac1 signaling pathway, we analyzed the level of Rac1 and PAK1 protein expression in addition to insulin-stimulated p-PAK1/2Thr423/402 in soleus

327

and EDL muscles from Akt2-KO mice. Akt2 KO did not significantly affect the protein content of either Rac1 or PAK1, suggesting that Akt2 does not regulate the expression of these proteins in skeletal muscle (Fig. 5A and B). However, insulin-stimulated p-PAK1/2Thr423/402 was affected differentially in the two muscles: Akt2 KO had no detectable effect on insulin-stimulated p-PAK1/2Thr423/402 in soleus, but the increment in p-PAK1/2Thr423/402 in response to insulin was prevented in EDL muscle (Fig. 5C). Because the amount of PAK1 trended lower in Akt2 KO mice, we analyzed the level of p-PAK1/2Thr423/402 in relation to the total amount of PAK1 in each individual muscle (Fig. 5D). The increment in insulin-stimulated p-PAK1/2Thr423/402/PAK1 was lower in Akt2 KO muscles due to a higher basal phosphorylation compared to their WT controls. This contrasts with our finding that inhibition of insulin-stimulated Akt phosphorylation by MK2206 does not affect p-PAK1/2Thr423/402 in either soleus or EDL muscles (Figs. 2A, 4A). These findings are interesting, since they suggest that basal activity and/or Akt2 protein per se, rather than its activation in response to insulin, is involved in activation of Rac1 signaling.

3.6. Pharmacological inhibition of Akt and expression of a constitutive active AS160 do not affect insulin-induced actin remodeling It is well established that Rac1 is a regulator of the actin cytoskeleton in the L6 muscle cell line and that Rac1 knockdown by siRNA prevents insulin-stimulated actin reorganization [8]. Hence, actin remodeling is readout of activation of the Rac1 ‘arm’ of insulin signaling. Since we

Fig. 4. A) Bar graphs showing quantifications of p-PAKThr423, p-AktSer473, p-AktThr308 and p-AS160 in soleus and EDL −/+ 15 μM MK2206, 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206 + 5 μM LatrunculinB (LatB), 10 μM Rac1 inhibitor II + 5 μM LatrunculinB, 5 μM LatrunculinB, or a corresponding amount of DMSO, 1 h pre-incubation. B) Representative blots showing insulin-stimulated signaling in soleus −/+ 15 μM MK2206, 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206 + 5 μM LatrunculinB (LatB), 10 μM Rac1 inhibitor II + 5 μM LatrunculinB, 5 μM LatrunculinB, or a corresponding amount of DMSO, 1 h pre-incubation. C) Representative blots showing insulin-stimulated signaling in soleus −/+ 15 μM MK2206, 10 μM Rac1 inhibitor II (Inhib II), 15 μM MK2206 + 5 μM LatrunculinB (LatB), 10 μM Rac1 inhibitor II + 5 μM LatrunculinB, 5 μM LatrunculinB, or a corresponding amount of DMSO, 1 h pre-incubation. Statistical significant effects of the inhibitors on signaling are indicated by #/##/### (p b 0.05/0.01/0.001). Statistical significant difference between basal and insulin is indicated by */**/*** (p b 0.05/0.01/0.001). Values are mean ± S.E.M.

328

L. Sylow et al. / Cellular Signalling 26 (2014) 323–331

Fig. 5. A) Bar graphs showing quantifications of total Rac1 and PAK1 protein expression in soleus and EDL muscles from wildtype (WT) or Akt2 knockout (KO) mice. B) Representative western blots of total Akt2, PAK1, Rac1, Actin, GLUT4, and basal or insulin -stimulated p-AktSer473 and p-PAK1/2Thr423/402 in WT and Akt2 KO mice. C) Bar graphs showing quantifications of insulin-stimulated (60 nM, 20 min) p-PAK1/2Thr423/402 in incubated soleus and EDL muscles from wildtype (WT) or Akt2 knockout (KO) mice. D) Bar graphs showing quantifications of insulin-stimulated (60 nM, 20 min) p-PAK1/2Thr423/402 related to total amount of PAK1 protein in incubated soleus and EDL muscles from wildtype (WT) or Akt2 knockout (KO) mice. Statistical significant effects of Akt2 KO on signaling are indicated by (#)/# (p = 0.1/b0.05). Statistical significant difference between basal and insulin is indicated by */**/*** (p b 0.05/0.01/0.001). Values are mean ± S.E.M.

Fig. 6. A) Representative western blot of basal and insulin-stimulated (100 nM) p-AktThr308 and Rac1 (control) with or without 15 μM MK2206 in L6 myoblasts. B) Representative collapsed (summation of all optical slices) and dorsal (summation of optical slices N2 μm above the stress fibers) images of F-actin in response to 100 nM insulin in L6 myotubes with or without 15 μM MK2206. C) Representative collapsed and dorsal images of F-actin in FLAG–AS160-4A expressing myotubes before and after insulin stimulation. Myotubes are marked by the white dotted lines. Scale bar = 20 μm.

L. Sylow et al. / Cellular Signalling 26 (2014) 323–331

here show evidence to suggest that the Akt and the Rac1 signaling pathways act through distinct distal mechanisms on glucose transport, we enquired what would be the effect of Akt inhibition on insulinstimulated actin remodeling. Due to methodological challenges in detecting cortical actin organization in mature skeletal muscle, we used L6 myotubes for this purpose. When F-actin was visualized by rhodamine–phalloidin in unstimulated myotubes, the staining was predominantly stress fibers. Upon insulin stimulation, F-actin underwent reorganization leading to the well-characterized Rac-dependent cortical actin remodeling at the dorsal section of the cell [8,9]. Similar to incubated skeletal muscles, MK2206 prevented insulin-induced p-AktThr308 in L6 myoblasts (Fig. 6A) but had no effect on the insulin-stimulated reorganization of cortical actin filaments (Fig. 6B). To further confirm that the Akt–AS160 signaling ‘arm’ is not involved in the regulation of actin remodeling, we overexpressed a mutant of AS160 (FLAG–AS160-4A), which is a constitutively active Rab–GAP and hence prevents signaling downstream of Akt [34] (Fig. 6C). AS160 is a GAP–Rab but it is conceivable that it could have additional effects on elements leading to actin dynamics, and in any case it was examined as an element of the Akt signaling pathway. Consistent with the lack of effect of the Akt inhibitor MK2206, the FLAG–AS160-4A mutant did not prevent insulinstimulated actin remodeling (Fig. 6C). These results show that the Akt

329

signaling pathway is not involved in insulin-dependent regulation of the actin cytoskeleton, which instead is downstream of Rac activation [20]. 3.7. Rac1 and Akt signaling are both severely dysfunctional in insulin-resistant ob/ob mice To investigate the extent of the dysfunction in Rac1 and Akt signaling in insulin-resistant muscles, we analyzed 2DG uptake, Rac1 protein expression, signaling, and distribution in soleus muscle from ob/ob mice stimulated with insulin ex vivo. While insulin increased 2DG transport in soleus muscle of lean control mice by 300%, this response was nearly totally prevented in ob/ob mouse muscle, highlighting the severe insulin resistance of muscle in the ob/ob mice (Fig. 7A). Insulin-resistance was accompanied by a ~70% decrease in insulin-stimulated p-AktSer473 and p-AktThr308 in ob/ob muscle (Fig. 7B). In the ob/ob mice, phosphorylation of PAK1/2Thr423/402 and of AS160 in response to insulin was completely prevented. We also analyzed the expression of Akt2, Rac1, and actin protein. Rac1 was lower in soleus muscle (by ~35%) and fat tissue (by ~60%) and tended to decrease in the liver (by ~ 50%) of ob/ob mice compared to lean controls. Likewise, Akt2 protein expression was lower in soleus muscle (by ~ 25%), fat (by ~ 60%) and liver (by ~ 25%).

Fig. 7. A) Insulin-stimulated (60 nM, 30 min) 2DG transport in incubated soleus muscles of Ob/Ob or lean control mice. B) Bar graphs showing quantifications of p-Akt Ser473, p-Akt Thr308 , p-PAK Thr423, and p-AS160 and actin in soleus of Ob/Ob or lean control mice. C) Bar graphs showing quantifications of Rac1, Akt2, and actin protein expression in soleus, gastrocnemius, liver and fat tissue of Ob/Ob and lean control mice. D) Representative images of Rac1 localization in mouse (Ob/Ob vs lean) EDL single fibers +/− 10 min 60 nM ex vivo insulin stimulation and bar graph showing Full Width at the Half Maximum (FWHM) of the striated Rac1 staining. Statistical significant difference between basal and insulin is indicated by */**/*** (p b 0.05/0.01/0.001). Statistical significant differences between Ob/Ob and lean control mice are indicated by †/††/††† (p b 0.05/0.01/0.001). Statistical significant difference between Ob/Ob and lean control mice on insulin-stimulated 2DG transport, signaling or FWHM is indicated by #/##/### (p b 0.05/0.01/0.001). Values are mean ± S.E.M. n = 5, 15 fibers imaged for each condition. Scale bar = 10 μm.

330

L. Sylow et al. / Cellular Signalling 26 (2014) 323–331

Actin expression was unchanged (Fig. 7C). We have previously observed that, like GLUT4 [35], Rac1 relocalizes to the transverse tubuli in response to insulin in mature skeletal muscle [6], suggesting that Rac1 activation specifically at the transverse tubuli is needed for Rac1 to exert its downstream effects. We therefore investigated whether Rac1 redistribution occurred to the same degree in insulin-resistant ob/ob mice as in insulin-sensitive, lean control mice. Insulin induced a small but significant increase (~15%) in Full Width at the Half Maximum in the striated Rac1 pattern (indicating localization to transverse tubuli [35]) in the EDL muscle of lean control mice (Fig. 7D). This redistribution was not detected in ob/ob mice, suggesting that the insulin-induced relocalization of Rac1 to the T-tubuli is also dysregulated in insulin-resistant muscle. 4. Discussion The current study shows that activation of both Rac1 and Akt is necessary for the insulin-dependent regulation of glucose transport in mature skeletal muscle. We also document for the first time, that simultaneous inhibition of these two pathways complementarily and completely blocks insulin-induced glucose uptake, and is a characteristic of insulin-resistant skeletal muscle. Most studies investigating the involvement of Akt, and in particular of Rac1, in insulin signaling and glucose uptake have used cultured muscle cells. Even though these are powerful tools to understand intracellular mechanisms, cultured muscle cells differ from mature skeletal muscle in the expression and reliance of various proteins in the gene expression and structural architecture that may influence the insulin response [35–37]. We found that separate inhibition of either Rac1 or Akt partially inhibits insulin-stimulated glucose uptake, whereas inhibition of both Akt and Rac1 signaling pathways abolished the insulin response. Taken together, this supports the concept that Rac1 and Akt are activated in parallel pathways, and highlights the importance of the simultaneous activation of these two pathways for full insulin action. We also evaluated whether Rac1 acts on glucose uptake via its effect on the actin cytoskeleton. Expression of dominant negative Rac1, or silencing the protein via siRNA, prevents insulin-stimulated actin remodeling [38,39], which suggests an essential role for Rac1 in regulating actin-dynamics in cultured muscle cells. In agreement with these reports, we observed no additive effect of inhibiting Rac1 on top of actin depolymerization, suggesting that Rac1 mediates its effect on glucose transport via the actin cytoskeleton. In contrast, actin depolymerization on top of Akt inhibition mimicked the effect of WM, and completely abolished glucose transport. These findings suggest that Rac1, but not Akt, acts via the actin cytoskeleton in mature skeletal muscle and that both pathways have essential contributions to the regulation of glucose uptake. This is further supported by our findings, that, in contrast to Rac1 knockdown [8], inhibition of the Akt signaling pathway did not affect insulin-induced actin remodeling in L6 myotubes. In cultured muscle cells, the actin cytoskeleton acts as a dynamic scaffolding structure for proteins in the insulin signaling cascade [32,33]. However, we found no alterations in signaling of any of the measured proteins following actin depolymerization, suggesting that skeletal muscle actin cytoskeleton is not essentially involved in regulating insulin-induced signaling, as reported for cultured muscle cells. Yet actin may scaffold insulin signaling proteins following their activation, to induce GLUT4 translocation. Similarly, while signaling measured on lysate remains intact, the signaling molecules may no longer be in the correct membrane compartments to stimulate GLUT4 translocation. Furthermore, in our hands, we cannot directly visualize the action of insulin on the actin cytoskeleton in mature skeletal muscle. The question therefore still remains whether the actin cytoskeleton is as dynamic a structure in the more densely packed mature skeletal muscle, as has been reported in cultured muscle cells. It is likely that the actin cytoskeleton in mature skeletal muscle is more static and therefore may have

different functions in mature skeletal muscle from those in a cultured muscle cell. Future studies should aim at answering these questions in order to understand how the actin cytoskeleton facilitates GLUT4 translocation. This will provide a possibility for developments of novel pharmacological drug targets in the treatment of insulin resistant states, such as in obesity and type 2 diabetes. The existence of cross-talk or feedback loops between the Rac1- and the Akt-signaling ‘arms’ has recently been reported in the literature [13,15,38]. However, we did not observe any effects of pharmacological Akt inhibition on Rac1-signaling in response to insulin or vice versa. In support of this, we previously showed that genetic KO of Rac1 in mice (with decreased insulin-induced muscle glucose transport) does not affect insulin-stimulated p-AktThr308 or p-AktSer473 [6]. Although pharmacological inhibition of Akt did not affect pPAK1/2Thr423/402, KO of Akt2 prevented insulin-stimulated p-PAK1/2Thr423/402 when related to total PAK1 protein in both soleus and EDL muscles, due to a higher basal phosphorylation in the Akt2 KO muscles compared to the WT controls. These findings are interesting, since it suggests that the presence of Akt2 protein is required for normal activation of the Rac1 signaling pathway. However, during normal conditions in which the Akt2 protein is present, our study suggests that there is no cross talk between Akt and Rac1 signaling in response to insulin in mature skeletal muscle. In cultured muscle cells, agents that induce insulin resistance lower Rac1 activation and GLUT4 translocation, with only small reductions in Akt signaling [12]. We recently confirmed this clinically in human models of insulin resistance and T2D in which Rac1 signaling is severely decreased [6], highlighting Rac1 as a major regulator of glucose transport in skeletal muscle. In keeping with those results, it has been proposed that a deficit in the ability to rearrange the cortical actin cytoskeleton in response to insulin stimulation constitutes a central defect in insulin-resistant muscle cells [26,39,40], revealing novel molecular drug targets for the treatment of insulin resistant states.

5. Conclusion In conclusion, we have documented that both Rac1 and Akt signaling are necessary for normal regulation of insulin-stimulated glucose transport in mouse muscle through pathways that are parallel and distinct but which may affect each other. Our findings further suggest that Rac1, but not Akt and AS160, involves the actin cytoskeleton to activate insulin-stimulated glucose uptake in mature skeletal muscle, and that both pathways are severely dysregulated in insulin resistant muscle. The identification of the Rac1 and Akt signaling ‘arms’ as acting via distinct intracellular mechanisms on glucose transport may allow for the possibility of multiple treatment targets for improving insulin sensitivity in muscle.

Authors' contributions L.S. E.A.R., T.E.J. and A.K. designed the study. L.S. conducted the experiments, performed the laboratory analysis, and wrote the manuscript. T.E.J., M.K., C.P., T.C., and E.A.R. took part in conducting the experiments. All authors revised on and approved the final version of the manuscript. T.E.J. is the guarantor of this work and, as such, has full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Grants The study was supported by grants from the Danish Medical Research Council, the Novo Nordisk Foundation, the Lundbeck Foundation, and the UNIK — Food Fitness and Pharma, and the Canadian Institutes of Health Research grant MT3707.

L. Sylow et al. / Cellular Signalling 26 (2014) 323–331

Acknowledgments We acknowledge the skilled technical assistance of Betina Bolmgren (Molecular Physiology Group, DK). Akt2 KO mice were a kind gift from Dr Laurie Goodyear, Joslin Diabetes Center and Department of Medicine, Harvard Medical School, Boston, USA. We thank Dr Jørgen Wojtaszewski for helpful discussions. No potential conflicts of interest relevant to this article were reported. Reference [1] C.M. Taniguchi, B. Emanuelli, C.R. Kahn, Nat. Rev. Mol. Cell Biol. 7 (2006) 85–96. [2] H.F. Kramer, C.A. Witczak, E.B. Taylor, N. Fujii, M.F. Hirshman, L.J. Goodyear, J. Biol. Chem. 281 (2006) 31478–31485. [3] Y. Sun, P.J. Bilan, Z. Liu, A. Klip, Proc. Natl. Acad. Sci. U. S. A. 107 (2010) 19909–19914. [4] N. Tapon, A. Hall, Curr. Opin. Cell Biol. 9 (1997) 86–92. [5] E. Manser, T. Leung, H. Salihuddin, Z.S. Zhao, L. Lim, Nature 367 (1994) 40–46. [6] L. Sylow, T.E. Jensen, M. Kleinert, K. Hojlund, B. Kiens, J.F. Wojtaszewski, C. Prats, P. Schjerling, E.A. Richter, Diabetes (6) (2013) 1865–1875. [7] Z.S. Zhao, E. Manser, X.Q. Chen, C. Chong, T. Leung, L. Lim, Mol. Cell. Biol. 18 (1998) 2153–2163. [8] L. JeBailey, A. Rudich, X. Huang, C. Di Ciano-Oliveira, A. Kapus, A. Klip, Mol. Endocrinol. 18 (2004) 359–372. [9] Z.A. Khayat, P. Tong, K. Yaworsky, R.J. Bloch, A. Klip, J. Cell Sci. 113 (Pt 2) (2000) 279–290. [10] J. Marcusohn, S.J. Isakoff, E. Rose, M. Symons, E.Y. Skolnik, Curr. Biol. 5 (1995) 1296–1302. [11] Q. Wang, R. Somwar, P.J. Bilan, Z. Liu, J. Jin, J.R. Woodgett, A. Klip, Mol. Cell. Biol. 19 (1999) 4008–4018. [12] L. JeBailey, O. Wanono, W. Niu, J. Roessler, A. Rudich, A. Klip, Diabetes 56 (2007) 394–403. [13] S. Ueda, T. Kataoka, T. Satoh, Biol. Cell. 100 (2008) 645–657. [14] S. Ueda, S. Kitazawa, K. Ishida, Y. Nishikawa, M. Matsui, H. Matsumoto, T. Aoki, S. Nozaki, T. Takeda, Y. Tamori, A. Aiba, C.R. Kahn, T. Kataoka, T. Satoh, FASEB J. 24 (2010) 2254–2261. [15] M.K. Delaney, J. Liu, Y. Zheng, M.C. Berndt, X. Du, Arterioscler. Thromb. Vasc. Biol. 32 (2012) 2761–2768. [16] T. Ijuin, T. Takenawa, Mol. Cell. Biol. 32 (2012) 3570–3584. [17] M. Higuchi, K. Onishi, C. Kikuchi, Y. Gotoh, Nat. Cell Biol. 10 (2008) 1356–1364. [18] A. Viola, N. Gupta, Nat. Rev. Immunol. 7 (2007) 889–896.

331

[19] N. Patel, A. Rudich, Z.A. Khayat, R. Garg, A. Klip, Mol. Cell. Biol. 23 (2003) 4611–4626. [20] T.T. Chiu, T.E. Jensen, L. Sylow, E.A. Richter, A. Klip, Cell. Signal. 23 (2011) 1546–1554. [21] H. Storgaard, C.B. Jensen, M. Bjornholm, X.M. Song, S. Madsbad, J.R. Zierath, A.A. Vaag, J. Clin. Endocrinol. Metab. 89 (2004) 1301–1311. [22] Y.T. Kruszynska, D.S. Worrall, J. Ofrecio, J.P. Frias, G. Macaraeg, J.M. Olefsky, J. Clin. Endocrinol. Metab. 87 (2002) 226–234. [23] K. Hojlund, P. Staehr, B.F. Hansen, K.A. Green, D.G. Hardie, E.A. Richter, H. Beck-Nielsen, J.F. Wojtaszewski, Diabetes 52 (2003) 1393–1402. [24] J.T. Brozinick Jr., B.R. Roberts, G.L. Dohm, Diabetes 52 (2003) 935–941. [25] K. Morino, K.F. Petersen, S. Dufour, D. Befroy, J. Frattini, N. Shatzkes, S. Neschen, M.F. White, S. Bilz, S. Sono, M. Pypaert, G.I. Shulman, J. Clin. Invest. 115 (2005) 3587–3593. [26] K.M. Habegger, B.A. Penque, W. Sealls, L. Tackett, L.N. Bell, E.K. Blue, P.J. Gallagher, M. Sturek, M.A. Alloosh, H.O. Steinberg, R.V. Considine, J.S. Elmendorf, Diabetologia 55 (2012) 457–467. [27] P. Tong, Z.A. Khayat, C. Huang, N. Patel, A. Ueyama, A. Klip, J. Clin. Invest. 108 (2001) 371–381. [28] H. Cho, J. Mu, J.K. Kim, J.L. Thorvaldsen, Q. Chu, E.B. Crenshaw III, K.H. Kaestner, M.S. Bartolomei, G.I. Shulman, M.J. Birnbaum, Science 292 (2001) 1728–1731. [29] T.E. Jensen, A.J. Rose, S.B. Jorgensen, N. Brandt, P. Schjerling, J.F. Wojtaszewski, E.A. Richter, Am. J. Physiol. Endocrinol. Metab. 292 (2007) E1308–E1317. [30] T. Ploug, D.B. van, H. Ai, S.W. Cushman, E. Ralston, J. Cell Biol. 142 (1998) 1429–1446. [31] K.L. Hoehn, C. Hohnen-Behrens, A. Cederberg, L.E. Wu, N. Turner, T. Yuasa, Y. Ebina, D.E. James, Cell Metab. 7 (2008) 421–433. [32] K. Peyrollier, E. Hajduch, A. Gray, G.J. Litherland, A.R. Prescott, N.R. Leslie, H.S. Hundal, Biochem. J. 352 (Pt 3) (2000) 617–622. [33] C.A. Eyster, Q.S. Duggins, A.L. Olson, J. Biol. Chem. 280 (2005) 17978–17985. [34] H. Sano, S. Kane, E. Sano, C.P. Miinea, J.M. Asara, W.S. Lane, C.W. Garner, G.E. Lienhard, J. Biol. Chem. 278 (2003) 14599–14602. [35] H.P. Lauritzen, T. Ploug, C. Prats, J.M. Tavare, H. Galbo, Diabetes 55 (2006) 1300–1306. [36] D. Tondeleir, D. Vandamme, J. Vandekerckhove, C. Ampe, A. Lambrechts, Cell Motil. Cytoskeleton 66 (2009) 798–815. [37] H.P. Lauritzen, J.D. Schertzer, Am. J. Physiol. Endocrinol. Metab. 299 (2010) E169–E179. [38] T.T. Chiu, Y. Sun, A. Koshkina, A. Klip, J. Biol. Chem. 288 (2013) 17520–17531. [39] E.M. Horvath, L. Tackett, A.M. McCarthy, P. Raman, J.T. Brozinick, J.S. Elmendorf, Mol. Endocrinol. 22 (2008) 937–950. [40] K.M. Habegger, N.J. Hoffman, C.M. Ridenour, J.T. Brozinick, J.S. Elmendorf, Endocrinology 153 (2012) 2130–2141.