Regulation of MAP Kinase–Directed Mitogenic and Protein Kinase B–Mediated Signaling by Cannabinoid Receptor Type 1 in Skeletal Muscle Cells

Bovine insulin was purchased from Sigma-Aldrich (Munich, Germany). WIN55,212-2 (R-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)-pyrrolo-[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphtalenylmethanone mesylate), AM630 (6-Iodo-2-methyl-1-[2-(4-morpholinyl)ethyl]-1H-indol-3-yl](4-methoxyphenyl)methanone), JWH015 (2-Methyl-1-propyl-1H-indol-3-yl)-1-naphthalenylmethanone), and ACEA (N-(2-Chloroethyl)-5Z,8Z,11Z,14Z-eicosatetraenamide) were all purchased from Tocris Bioscience (Bristol, U.K.). SR141716 (N-(Piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole- 3-carboxamide) was a gift from sanofi-aventis (Frankfurt, Germany). All antibodies were from Cell Signaling Technology (Beverly, MA) unless stated otherwise. Anti-actin was purchased from Sigma-Aldrich (St. Louis, MO). Both IRS-1 and p85 (PI 3-kinase) antibodies were from Millipore (Watford, U.K.). The antibody against the CB1 receptor together with rat CB1 and rat CB2 cDNA were generated by the Mackie laboratory (Indiana University) (28). Horseradish peroxidase–conjugated anti-(rabbit IgG) and anti-(mouse IgG) were obtained from New England Biolabs (Beverley, MA).

L6 cells were cultured to myotubes in α-minimum essential medium containing 2% (vol/vol) FBS (Life Technologies, Paisley, U.K.) and 1% (vol/vol) antibiotic/antimycotic solution (100 units/ml penicillin, 100 μg/ml streptomycin, and 250 ng/ml amphotericin B) at 37°C with 5% CO₂. HEK-293AD and HEK-293T cells were maintained in a humidified 5% CO₂, 37°C incubator in Dulbecco's modified Eagle's medium/F12 with 10% (vol/vol) FBS. Transient transfections in HEK-293 cells were carried out using Lipofectamine 2000 (Invitrogen), according to manufacturer's instructions.

L6 myotubes were lysed after experimental treatment in an appropriate volume of lysis buffer (29). Cell lysates (40 μg) were subjected to SDS/PAGE on a 10% resolving gel and immunoblotted as previously described (29). Immobilon-P membranes (Millipore, Bedford, MA) were probed with different primary antibodies as described in the figure legends. Primary antibody detection was performed with the appropriate horseradish peroxidase–conjugated anti-rabbit or anti-mouse IgG and resulting signals visualized using enhanced chemiluminescence by exposure to Konica Minolta X-ray autoradiographic film.

L6 myotubes were lysed as described above. IRS-1 was immunoprecipitated overnight at 4°C from 400 μg of cell lysate protein using an anti–IRS-1 antibody. The resulting immunocomplex was conjugated to protein-G-Sepharose beads and solubilized in Laemmli sample buffer prior to immunoblotting.

Total RNA was extracted from L6 myoblasts or myotubes using TRI Reagent (Sigma-Aldrich, Dorset, U.K.) according to the manufacturer's instructions. The extracted RNA (1 μg) was used as the template for cDNA synthesis performed using the Improm-II Reverse Transcription System (Promega, Madison, WI) according to manufacturer's instructions. Primers used for PCR amplification were from previously published sequences and were as follows: Forward 5′-3′, GGCATCTCTTTCTCAGTCAC, reverse 5′-3′, ATCAGGTAGGTCTCGTCAAT (30). PCR amplification was performed using the following cycling conditions: 94°C for 2 min followed by (for 28 cycles) 94°C for 1 min, 60°C for 1 min, and 72°C for 2 min. Final extension was carried out at 72°C for 5 min. The amplification reaction was carried out using GoTaq Flexi DNA polymerase according to manufacturer's instructions (Promega). PCR products were then analyzed by agarose gel electrophoresis.

Short hairpin RNA (shRNA) sequences were inserted into the pLKO.1-puro lentiviral vector (Sigma-Aldrich, St. Louis, MO) as follows. Each hairpin consisted of a 21-nucleotide sense sequence, a short hairpin sequence (CTCGAG), a 21-nucleotide antisense sequence, and five thymidines (a stop signal for RNA polymerase). Oligonucleotides (oligos) were constructed for rat CB1 (sense and antisense strands underlined, hairpin loop is indicated by italics): forward CCGGAACAAGTCTCTCTCGTCGTTCAACTCGAGTTGAACGACGAGAGAGACTTGTTTTTTTG, reverse AATTCAAAAAAAGAGCACCGTTAAGATCGCGAACTCGAGTTCGCGATCTTAACGGTGCTCTT. Additional nucleotides were added to the ends of the oligos as shown, such that annealing of the two complementary oligos resulted in overhangs consistent with those generated by EcoRI and AgeI. Oligos were annealed by mixing 2 μg of each oligo, heating to 94°C for 10 min, followed by cooling at a rate 1°C per minute until 21°C is reached. The final double-stranded DNA sequences were then inserted into pLKO.1-puro at the EcoRI and AgeI sites. Correct insertions of shRNA were confirmed by sequencing.

Recombinant lentiviruses were produced by cotransfecting HEK-293T cells with the empty pLKO.1-puro vector, pLKO.1-puro/scramble, or pLKO.1-puro/CB1 plasmids with the envelope vector pCMV-VSVg and the packaging vector pHR-CMV8.2ΔR at ratios of 3:2:2, respectively, according to mass. Infectious lentiviral particles were harvested by collecting the cell culture media 72 h later and filtered through 0.45-μm Minisart cellulose-acetate filters. Infection of L6 myoblasts was carried out by adding 1 ml crude lentivirus preparation to L6 myoblasts seeded in 6-cm dishes in the presence of 8 μg/ml polybrene. Medium was replaced with standard growth medium containing 3 μg/ml puromycin 24 h after transduction and cells were maintained in this selective medium until fully selected.

All statistical analysis was performed using one-way ANOVA. Data analysis was performed using GraphPad Prism software and considered statistically significant at P values < 0.05.

CB1 receptor expression was originally thought to be exclusive to the brain. However, an increasing number of studies suggest that these receptors are also expressed in peripheral tissues, including skeletal muscle (13). However, to date, no specific function(s) has been attributed to these cannabinoid receptors, and their roles in skeletal muscle physiology remain largely unknown. In an attempt to address this, we initially assessed expression of the CB1 receptor in L6 rat skeletal muscle cells. Using RT-PCR and Western blot analysis, respectively, the expression of the CB1 receptor was confirmed in fully differentiated L6 myotubes at both the mRNA (Fig. 1,A) and protein (Fig. 1,B) level. However, CB1 receptor protein was not detected in L6 myoblasts (Fig. 1,B), supporting the view that its expression is differentiation dependent in skeletal muscle (31). Specificity of the CB1 antibody used was confirmed by immunoblot analysis performed using lysates obtained from HEK-293AD cells transiently overexpressing rat CB1 and CB2 receptors (Fig. 1 C).

Treatment of rat skeletal L6 myotubes with 100 nmol/l insulin for 10 min after short-term (2-h) serum starvation leads to robust activation of PKB and ERK1/2 (p42/44 MAPK), as measured by the extent of phosphorylation of PKB at Thr308 and Ser473, and of ERK1/2 at Thr202/Tyr204 (Fig. 2,A). Acute pretreatment (up to 30 min) using the mixed CB1/CB2 agonist WIN55,212-2 (100 nmol/l) has no effect on insulin-stimulated ERK1/2 or PKB activation (Fig. 2,A). However, pretreating L6 myotubes with WIN55,212-2 (100 nmol/l) for 24 h prior to insulin treatment substantially perturbs the activation of ERK but not that of PKB (Fig. 2 B). Therefore, chronic WIN55,212-2 treatment selectively suppresses insulin-induced activation of the ERK1/2 pathway, but has no effect on PKB activation. This is in contrast to a previous study carried out in glioma cells, demonstrating that longer-term (> 12 h) treatment with WIN55,212-2 used at a higher concentration (15 μmol/l) can downregulate both ERK1/2 and PKB activity, although this was done in the absence of any external hormonal stimulus (32). In another study performed in primary human skeletal muscle, acute exposure with the endogenous CB1 agonist anandamide alone was found to transiently activate ERK1/2. However we did not observe this same effect in the L6 myotubes using the synthetic CB1 agonists WIN55,212-2 and ACEA, but it is noteworthy that the effects reported by Eckardt et al. (31) were observed using a high concentration (10 μmol/l) of anandamide, which may potentially modulate a number of cellular responses independently of its effect on cannabinoid receptors.

To determine whether the effect of WIN55,212-2 is mediated through the CB1 and/or CB2 receptors, we tested the CB1 and CB2 selective agonists ACEA and JWH015, respectively. Acute (30 min) treatment with ACEA (10 nmol/l) had no effect on either ERK1/2 or PKB activation in response to insulin (Fig. 2,C). However, similar to the application of WIN55,212-2, a 24-h treatment with ACEA (10 nmol/l) also led to the inhibition of insulin-stimulated ERK1/2 activity but had no effect on PKB activation (Fig. 2,D). In contrast to the effects of WIN55,212-2 and ACEA, treating L6 myotubes with JWH015 either acutely (100 nmol/l, 30 min) or with various concentrations longer term (10 nmol/l to 1 μmol/l, 24 h) does not alter insulin-stimulated activation of ERK1/2 or PKB (Fig. 3), suggesting that the inhibitory effect of WIN55,212-2 on ERK1/2 signaling is mediated selectively through the CB1 receptor. It should be noted that concentrations of JWH015 less than 100 nmol/l have been shown to inhibit cAMP production and those less than 1 μmol/l, to evoke calcium transients in HEK-293 cells expressing either the human or rat forms of the CB2 receptor (33,34).

SR141716 (rimonabant) is a selective CB1 receptor inverse agonist (trade name Acomplia) that has been used to treat obesity (35). Administration of SR141716 causes a transient reduction in food intake and sustained reduction in body weight gain as well as decreases hyperinsulinemia and reduces circulating levels of free fatty acids in diet-induced obese rodents and in humans (17,22). Predictably, these effects of SR141716 are not observed in CB1 receptor–deficient mice fed a high-fat diet (17). To our knowledge, the effects of SR141716 on insulin-induced signaling responses in skeletal muscle have not yet been fully elucidated. Therefore, we investigated whether exposing L6 myotubes to SR141716 would abrogate the effects of CB1 activation on insulin-stimulated PKB and ERK1/2 activation. Our data demonstrate that SR141716 prevents the loss in ERK1/2 activity in response to both WIN55,212-2 and ACEA in a concentration-dependent manner, with the effects of the agonists completely blocked by 100 nmol/l SR141716 (Figs. 4 and 5). In contrast, application of the CB2 selective antagonist AM630 (100 nmol/l) had no effect on WIN55,212-2 and ACEA-mediated loss in ERK1/2 activity, again implying that the modulation of ERK1/2 signaling is selectively mediated through the CB1 receptor (Fig. 4 A).

Consistent with the inhibition of ERK1/2, WIN55,212-2 or ACEA treatment (both 100 nmol/l) for 24 h also leads to suppression of insulin-induced cAMP-responsive element–binding protein (CREB) phosphorylation at Ser133, a downstream target of ERK1/2 as well as activation of MAP kinase kinase (MEK) 1/2, the immediate upstream kinase of ERK (Figs. 4 and 5). Again, similarly to the effects on ERK1/2, the responses on MEK1/2 and CREB phosphorylation are abrogated by coapplication of SR141716 in a concentration-dependent manner. Interestingly, our data also demonstrate that the level of insulin-mediated activation of MEK1/2, ERK1/2, and CREB with coapplication of SR141716 with WIN55,212-2 was higher than that observed with insulin alone (Fig. 4 B). However, this did not appear to be the case in the presence of ACEA, possibly due to the higher affinity of this agonist toward the CB1 receptor thereby possibly competing more effectively with SR141716 with regard to receptor binding.

In addition to the antagonistic effects of SR141716, we also investigated whether application of SR141716 alone would modulate insulin sensitivity in L6 myotubes. Interestingly, pretreatment with SR141716 (100 nmol/l) for 24 h significantly enhances both insulin-stimulated PKB and ERK activity (Fig. 6), with an approximately twofold and threefold greater stimulation observed, respectively. This is despite the lack of any effect of pharmacological CB1 activation on insulin-regulated PKB activity. Consistent with the increased activation of PKB, the insulin-induced phosphorylation of forkhead transcription factor in rhabdomyosarcoma-like 1 (FKHRL1)/Foxo3a (Thr32) and glycogen synthase kinase (GSK) 3α/β (Ser21/Ser9), two important downstream targets of PKB, was also further enhanced after SR141716 pretreatment. It must also be noted that despite mediating the enhancement of insulin-stimulated PKB activity, exposure of L6 myotubes to SR141716 alone does not induce activation of this kinase, contrary to previously published data (27).

To determine whether the insulin-sensitizing effects of SR141716 are mediated through the CB1 receptor, we generated stable CB1 knockdown L6 cells using shRNA expression vectors and confirmed suppression of CB1 expression by ∼70% (Fig. 7). As in normal wild-type L6 cells, those transfected with silencer-negative control small interfering RNA also show enhanced PKB and ERK1/2 activation in response to insulin after SR141716 treatment (Fig. 7). In contrast, genetic silencing of CB1 leads to a loss in SR141716-mediated insulin sensitization in L6 cells with respect to the activation of both ERK1/2 and PKB.

Insulin-stimulated activation of PI 3-kinase is dependent largely on the interaction of its regulatory subunit, p85, with the IRS proteins. Therefore, by immunoprecipitation, we assessed whether the interaction between p85 (PI 3-kinase) and IRS-1 was altered by modulation of CB1 receptor function. Treatment of L6 myotubes with either the CB1 selective agonist ACEA (10 nmol/l) or the CB1 inverse agonist SR141716 (100 nmol/l) for 24 h did not alter IRS-1/p85 association either under basal or insulin-stimulated conditions (Fig. 8,A). This is also consistent with the lack of detectable activation of PKB when SR141716 is applied alone (Fig. 6), and suggests that regulation of CB1 receptor function does not by itself alter PI 3-kinase activity.

Another potential explanation for SR141716-mediated sensitization of insulin-regulated signaling may involve the suppression of phosphatase activity toward target kinases. To explore this possibility, L6 myotubes were treated with or without SR141716 for 24 h prior to insulin stimulation, followed by insulin washout for up to 30 min in the presence of a PI 3-kinase inhibitor, to determine the level of PKB activity under postinsulin washout conditions. In contrast to control (vehicle-treated) cells, those incubated with SR141716 display a significantly delayed loss in PKB phosphorylation at Ser473 during postinsulin washout (Fig. 8 B). This therefore suggests that SR141716 acts through CB1 to suppress phosphatase activity toward this site, thereby maintaining PKB in its active state.

The endocannabinoid system (ECS) has emerged as a key regulator of energy balance, and inhibiting CB1 receptor function is a potential treatment for obesity. Although the effects of the ECS on metabolism are mediated both centrally and peripherally, recent studies have shown that CB1 inverse agonists/antagonists such as SR141716 (Acomplia) do not need to act centrally to improve circulating lipid profile and glucose metabolism in the obese state (10,36). The clinical implementation of SR141716 as a counter-obesity drug has, however, been hindered by its central nervous system side effects, resulting in an unfavorable risk-benefit ratio. Nevertheless, a new generation of peripherally acting CB1 inverse agonists/antagonists is currently under clinical development. Identifying skeletal muscle as a potential site of action for the effects of CB1 antagonists/inverse agonists on metabolic processes supports the idea that peripheral cannabinoid targets may be of therapeutic benefit. Indeed, previous studies have demonstrated that application of the CB1 receptor inverse agonist SR141716 elevates glucose uptake in soleus muscle isolated from leptin-deficient obese mice as well as in cultured L6 skeletal muscle cells (25,27). These studies were the first to implicate a role for the CB1 receptor in skeletal muscle function and suggest that tonic CB1 receptor activity influences glucose metabolism in this tissue. In this study, we have manipulated CB receptor function by pharmacological means in an attempt to evaluate the importance of the ECS upon insulin action in muscle cells.

Our findings indicate that modulating CB1 receptor function by selective agonists/inverse agonists can regulate the responsiveness of skeletal muscle toward insulin through its action on two key signaling pathways, namely PI 3-kinase/PKB and the Raf-MEK1/2-ERK1/2 phosphorylation cascade. This is an important finding that suggests that the upregulation of endocannabinoid/CB1 tone, as observed in obese animal models and humans, may lead to altered insulin sensitivity in peripheral tissues such as skeletal muscle. Crucially, endocannabinoid levels have previously been shown to be elevated in soleus muscle of mice fed a high-fat diet in accordance with hyperglycemia (37). This work also raises the likelihood of cross talk between CB1 receptor function and upstream components of insulin signaling involved in the regulation of PKB and ERK1/2.

Our data demonstrate that longer-term pharmacological activation of the CB1 receptor function leads to the selective inhibition of ERK1/2-mediated signaling, while having little or no effect on PKB activity. Furthermore, we have shown for the first time that treatment with the CB1 inverse agonist SR141716 alone can improve insulin sensitivity in skeletal muscle as determined by its ability to enhance both insulin-stimulated PKB- and ERK1/2-directed signaling. However, although activation and inhibition of CB1 are consistent with the inhibitory and stimulatory effects on ERK1/2 activation, respectively, a discrepancy arises when comparing the effects of CB1 agonism and inverse agonism with regard to modulating PKB activity.

It remains unclear what accounts for the distinct effects of CB1 receptor modulation toward PKB- and ERK1/2-directed signaling, but there are a number of possibilities that may help explain the basis of this differential control. It is plausible, for example, that some level of cross-talk exists between the CB1 receptor and insulin receptor signaling, resulting in the specific downregulation of the Ras/Raf/ERK cascade—for instance, this may be due to interference with the ability of adapter molecules such as growth factor receptor–bound protein 2 and Shc to bind to the insulin receptor, or possibly involving the downregulation in expression or activity of one or more of these components. Furthermore, this might occur without altering IRS protein functionality, thereby maintaining PI 3-kinase/PKB signaling capability. Second, CB1 activation may lead to increased phosphatase activity toward the Ras/Raf/ERK cascade, resulting in its inhibition, while having little or no effect toward components of the PI 3-kinase/PKB pathway.

In contrast to the effects observed with the CB1 agonists, applying SR141716 alone potentiated both insulin-stimulated PKB and ERK1/2 activity. In view of our data showing that applying SR141716 alone does not activate PKB to any appreciable extent, we would discount any additive effect of insulin and SR141716. Instead, our observations support the notion that SR141716 acts to potentiate insulin-stimulated PKB activity and because this enhancement in PKB activation is not observed in cells where CB1 expression has been silenced (Fig. 7), it implicates a role for the CB1 receptor in this sensitization.

The potentiation of PKB and ERK1/2 activation by longer-term SR141716 treatment may arise through several different mechanisms. First, it may be due to an increase in the expression of the upstream components of insulin signaling such as the insulin receptor, IRS-1/2, or the regulatory and catalytic subunits of PI 3-kinase. However, we do not observe any detectable elevation of PKB or ERK1/2 activity when SR141716 is applied alone, as measured by the extent of their phosphorylation as would be expected in such a scenario. In addition, there is no observed enhancement in association between IRS-1 and the p85 regulatory subunit of PI 3-kinase, implying that it is unlikely that SR141716 is acting via the modulation of PI 3-kinase activity. Nevertheless we cannot rule out the possibility that a marginal increase in these upstream components may contribute to the insulin sensitization (27). Another possible explanation may involve SR141716 “priming” the cell for stimulation by insulin by disrupting the inhibitory environment acting on PKB and ERK1/2 under basal conditions. For instance, this may involve relieving the inhibitory action of various phosphatases that can negatively regulate insulin responses (38,,,,–43). We have observed that SR141716 does not alter expression of phosphatase and tensin homolog (PTEN) in L6 myotubes (data not shown), although the possibility that it may change the expression and/or activity of other important protein phosphatase(s) such as protein tyrosine phosphatase 1B (PTP-1B) or Src homology domain 2 (SH2)-containing tyrosine phosphatase-1 (SHP-1) cannot be excluded. Indeed, our data demonstrating that SR141716 delays dephosphorylation of PKB at Ser473 during insulin washout provide evidence in support of such a mechanism. However, further work will be required to elucidate the identity of the phosphatase(s) involved. Another explanation may arise from the localization of CB1 receptors at the plasma membrane. For example, these receptors are found to be present within caveolae, well-characterized lipid rafts that perform a number of signaling functions (44). Caveolae are cholesterol-rich microdomains that act to retain PKB in a repressed state via the action of atypical protein kinase Cs (45). Because insulin receptors are also known to reside within caveolae, it is possible that the close proximity of the cannabinoid receptors causes them to be involved in modulating insulin receptor–mediated signaling. One way to test this idea would be to determine the effects of CB1 activation or antagonism in L6 cells with a diminished number of caveolae.

Consistent with the inhibition of ERK1/2, WIN55,212-2 or ACEA treatment for 24 h also suppresses insulin-induced CREB phosphorylation at Ser133, a downstream target of ERK1/2, as well as the activation of MEK1/2, the immediate upstream kinase of ERK (Figs. 4 and 5). This suggests that the effects of CB1 agonism on ERK1/2-mediated signaling are acting at the level of MEK1/2 or above, possibly affecting the activity of the GTPase Ras or the MAPK kinase Raf, both of which lie upstream of ERK1/2 and MEK1/2 in the signaling cascade. It has also been demonstrated previously that antagonists/inverse agonists of the CB1 receptor can lead to an increase in the levels of cAMP, which in turn promotes activation of protein kinase A (27,46). Protein kinase A has also been shown to phosphorylate Raf-1 thereby preventing it from binding to Ras-GTP through its association with 14-3-3 proteins (47). This would ultimately lead to an inhibition of the Ras-Raf-ERK signaling cascade. However, in our case, SR141716 acts to potentiate the activation of ERK by insulin, thereby implicating a different response that can override any effects of enhanced cAMP production as SR141716 has been shown to cause in L6 cells (27).

This work also raises the question as to whether SR141716 is functioning as an antagonist by inhibiting constitutive/tonic CB1 activity in the L6 cells or by acting as an inverse agonist as reported in other cell types (48,49). Importantly, although complete silencing of CB1 receptor expression using shRNA was not achieved, the remaining levels of receptor (∼30%) are insufficient to convey the insulin-enhancing effects of SR141716. This suggests that some threshold level of CB1 receptor expression is required to be present in order for SR141716 to act as a positive modulator of insulin signaling. In addition, because we did not observe enhancement of PKB and/or ERK1/2 activity in response to insulin in the CB1 knockdown cells compared with the silencer-negative control cells (Fig. 7), we propose that CB1 receptor occupancy by SR141716 is required for mediating these insulin-sensitizing effects, whereas inhibition through suppression of receptor expression does not produce the same response.

Therefore, at least in cultured L6 muscle cells, pharmacological activation of the CB1 receptor acts to selectively inhibit ERK1/2, while largely unaffecting PKB and its associated activity. Intriguingly, blockade of CB1 receptor function acts to abrogate impedance toward signaling and enhance both insulin-stimulated ERK1/2 and PKB activity. In conclusion, our results further contribute to the growing body of evidence that the ECS does have an important role in modulating muscle function and, in accordance with the reported behavioral side effects of SR141716, support the use of peripheral instead of central CB1 inverse agonism/antagonism to treat obesity and metabolic disorders.

We are grateful for financial support from the National Institutes of Health/National Institute on Drug Abuse (DA11322), European Commission (contract LSHM-CT-20004-005272), Diabetes Research & Wellness Foundation, and Diabetes UK.

No potential conflicts of interest relevant to this article were reported.