The endocannabinoid system (ECS) is a lipid signaling system involved in the regulation of appetite, fatty acid oxidation, glucose metabolism, and inflammation [1,2,3]. Energy homeostasis regulation via the ECS works through both central and peripheral mechanisms [4]. There is evidence from both in vitro and in vivo studies implicating the ECS in a series of pathological conditions [5]. Antagonism of cannabinoid receptor 1 (CB₁), which is upregulated in obesity, promotes weight loss and decreases obesity-associated co-morbidities [6]. The use of the global CB₁ antagonist Rimonabant in the treatment of obesity has been hampered by its adverse effects on the central nervous system (CNS) resulting in adverse psychological side effects [7].

Skeletal muscle plays an important role in whole-body energy metabolism including fatty acid oxidation and glucose homeostasis [5]. An increase in intramyocellular fatty acids due to decreased oxidative capacity in skeletal muscle is thought to result in cellular stress and contribute to the low-grade chronic inflammation observed in obesity [6]. Antagonism of CB₁ in skeletal muscle upregulates the expression of genes involved in fatty acid oxidation and glucose control [6], thereby having a beneficial effect on skeletal muscle function in obesity. Additionally, the adipokine adiponectin can act on skeletal muscle adiponectin receptors (AdipoR1, and AdipoR2) [8] to positively affect insulin-stimulated glucose transporter expression and glucose uptake [8]. CB₁ antagonism increases plasma adiponectin levels independent of weight loss [9], whether this increase in adiponectin correlates with improved adiponectin signaling in skeletal muscle is unknown.

In addition to the role of the ECS in skeletal muscle, elucidating the mechanisms of adipocyte dysfunction in obesity and its links to the ECS may be important in identifying additional novel anti-obesity treatment targets. The global CB₁ inverse agonist/antagonist AM251 has been shown to reduce weight gain in animal models of obesity by contributing to a reduction in food intake [10] and an increase in energy expenditure [11].

The effect of AM251 to reduce white adipose tissue (WAT) weight in high fat diet (HFD) fed models is thought to be mainly a result of increased sympathetic tone to WAT depots, while AM251 can increase uncoupling protein 1 (UCP1) mRNA and protein levels in brown adipose tissue (BAT) [12]. It is unclear whether there is any “browning” of WAT depots with AM251; white adipocytes from CB1 knockout mice have increased UCP1 mRNA levels associated with increased cellular respiration [12], which could indicate that CB1 antagonism could directly result in WAT browning.

Therefore, we investigated in a diet induced obesity (DIO) rat model (1) the physiological effects of the CB₁ ligand, AM251, (2) the inflammatory profile of AM251 treatment (3) the effect of AM251 treatment on specific adipocyte genes in both WAT and BAT depots and (4) the effects of AM251 treatment and acute adiponectin exposure on fatty acid oxidative pathways in skeletal muscle. The hypothesis of this study is that AM251 will influence adiponectin signaling pathways in skeletal muscle and contribute to “browning” of WAT.

Despite advances in research, limited pharmacological treatments are available for obesity and its co-morbidity. Antagonism of CB₁ induces weight loss leading to a reduction in various co-morbidities [1]. The current study demonstrated that 6 weeks of treatment with the global CB₁ inverse agonist/antagonist, AM251, in DIO rats clearly influences whole body physiological parameters including decreased fat pad weight, and, as we have previously demonstrated decreased food intake and weight gain [13]. At the tissue level, specific changes were observed in the expression of several genes in adipose tissue depots, but no changes were observed in skeletal muscle. These results demonstrate the complex interplay between the endocannabinoid system and metabolic tissues in obesity and weight loss. However, the administration of AM251 increased systemic inflammation with increased levels of a wide range of cytokines.

The browning of WAT depots is thought to be an avenue for therapeutic approaches to obesity [14]. Our hypothesis was that AM251 may increase browning of WAT that contributes to its actions on lowering body weight and reduced fat mass. However, AM251 treatment in HFD rats did not support this hypothesis in eWAT and pWAT, as there was a decrease in PRDM16 and CPT1b mRNA expression and no alteration in UCP1 mRNA expression. This is despite evidence that CB₁ knockout mice have WAT displaying increased mitochondrial function and elevated UCP1 levels [12], and chronic treatment of dogs with rimonabant results in the increased expression of brown fat markers (including UCP1) in WAT [12]. An upregulation of β₃-ARs in WAT is thought to be partly responsible for the browning of WAT by rimonabant [15]. Our results show that AM251 treatment was accompanied by a downregulation in β₃-AR mRNA expression. Since β₃-ARs are important in lipolysis and lowering body fat levels [16], this may suggest that the lowering of body fat by AM251 may be independent of β₃-ARs in WAT.

BAT is an important regulator of whole-body energy homoeostasis and involved in thermoregulation of rodents [16], and its presence/function in humans has led BAT to be a new target for human obesity treatments [16]. In DIO mice treated with rimonabant, improvements in glucose utilisation, increased energy expenditure and activation of BAT independent on weight loss were reported [17]. In our DIO rat model, AM251 treatment caused a decrease in Adrb3, UCP1, CPT1b, GLUT4, PRDM16, and leptin mRNA levels in BAT. It’s important to note that changes in UCP1 mRNA levels do not necessarily reflect changes in UCP1 protein levels [16] and that UCP1 still needs to be activated by a stimuli such as sympathetic stimulation of β3-ARs. In our AM251 treated DIO rats, there was however a decrease in Adrb3 which was coupled with a decrease in UCP1 as well as CPT1b and PRDM16 mRNA levels, which may suggest a blunting of thermogenic responses in these animals. However, functional readouts in the future would enable a better conclusion to be made.

The decrease in plasma leptin in our DIO model following AM251 treatment agrees with previous research in mice following 15-day treatment with rimonabant [18]. Ghrelin which is implicated to stimulate appetite [13] was decreased following six weeks of treatment with AM251, and supports our previous observations in this cohort of a decrease in food intake, albeit only transiently [13]. In further agreement, glucagon which has been demonstrated to be reduced following weight loss [13] also decreased following the treatment period. In contrast to these observed changes in leptin, ghrelin, glucagon, and PAI-1 we did not observe any change in plasma adiponectin levels. This disagrees with previous research in animals and humans that have demonstrated that antagonists/inverse agonists of CB₁ increase plasma adiponectin levels [1].

While no changes in plasma adiponectin were observed in the current study, we were interested in whether there was any alteration in some key genes involved in adiponectin signaling and muscle oxidative capacity, following treatment with AM251 in HFD fed rats and whether these genes were altered following acute exposure to adiponectin. It has previously been demonstrated in human primary skeletal muscle myotubes derived from both lean and obese individuals that AM251 treatment increases AMPKα1 and decreases PDK4 mRNA expression, while no changes were observed in the expression of AMPKα2 [6]. In the current study, despite the reduction in weight gain and alteration in adipose tissue mass, we did not observe any changes in mRNA expression following AM251 treatment in key genes involved in adiponectin signaling or alteration in PDK4. However, in agreement with the human primary skeletal muscle culture work [6,19] we did not observe any changes in a key gene involved in mitochondrial biogenesis, PGC-1α, which is supported by the lack of change in citrate synthase activity in the current study.

We have previously demonstrated in human primary skeletal muscle myotubes that 6 h of treatment with 0.1 µg/mL of globular adiponectin (gAd) increased AMPK activity in myotubes derived from lean individuals but only the higher dose of 0.5 µg/mL of gAd increased AMPK activity derived from individuals with obesity and obesity and diabetes [20]. The current study demonstrates in rodents that acute, 30-min exposure to adiponectin following six weeks of an HFD did not alter AMPKα2 mRNA in the soleus or EDL. We have also previously demonstrated that 6 h of adiponectin treatment in human primary myotubes derived from lean individuals increases AdipoR1 but not AdipoR2 mRNA expression, whereas myotubes derived from individuals who had obesity or obesity and diabetes did not alter AdipoR1 or AdipoR2 [21]. In agreement with this work, whether the HFD animals were treated with AM251 or not, acute exposure to adiponectin did not alter AdipoR1 or AdipoR2 in the soleus or EDL. While overexpression of Appl1 has been demonstrated to increase adiponectin signaling [22] we also did not see any compensatory increase in the mRNA expression of Appl1 in the animals following adiponectin treatment. While the lack of change in PGC-1α supports our previous work in human primary myotubes following gAd treatment [21,23], it is possible that the acute ex vivo exposure to adiponectin was not long enough to enable the detection of any subsequent mRNA changes and future work should increase the ex vivo exposure time.

AM251 treatment of HFD fed rats caused an increase in several plasma pro-inflammatory cytokines: namely IFN-γ, IL-1α, IL-1β, IL-2, IL-4, IL-6, IL-12p7, IL-17α, IL-18, and RANTES. Adipose tissue remodelling during weight loss can initiate an inflammatory response [24], and may at least partially explain the increased levels of circulating cytokines observed in the current study. IL-6 has also been observed to increase following weight reduction [25]. However, in contrast to these studies, other circulating cytokines have been shown to decrease when reducing adiposity [26] and a higher dose of AM251 (10 mg/kg) for 4 weeks reduced pro-inflammatory M1 adipose tissue macrophages in the presence of a reduction in fat pad mass [27]. It is known that some cytokines such as IL-18 may not be affected by weight loss but with alterations in insulin resistance [28]. IL-18 however, increased in the current study after AM251 treatment. Interestingly plasma adiponectin levels have been demonstrated to be negatively correlated with IL-6 and IL-18 [29]. Thus, it is possible that the increases in these cytokines because of AM251 treatment negated the expected increase in plasma adiponectin as a result of decreased adipose tissue mass. Rimonabant treatment for 12 weeks in obese women with polycystic ovarian syndrome resulted in increased circulating levels of VEGF however, the treatment did not alter TNF-α, IL-1β, IL-2, and IL-6 [30], which is in contrast to our results following AM251 treatment. IL-10, which has been previously demonstrated to correlate with circulating adiponectin levels [31], did not change in the current study. In view of these collective results, we measured the mRNA expression of three of these cytokines (IL-1β, IL-6 and TNF-α) in the different fat depots to determine if the source of the inflammation was adipose tissue (we were not able to detect sufficient amounts of TNF-α in pWAT and BAT for quantification). Taken together our observed decrease in IL-1β in BAT and lack of alterations of this and IL-6 and TNF-α in other adipose tissue depots and previously observed reductions in IL-6 and TNF-α in adipose tissue macrophages following treatment with AM251 [27] suggest that the source of increased circulating cytokines in this study does not appear to be derived from adipose tissue. Indeed a recent paper [32] has indicated that CB1 deficiency in non-immune cells promotes DIO resistance, while CB1 deficiency in immune cells further worsens DIO induced inflammation. This may indicate that blockade of the CB1 receptor pharmacologically has both beneficial and detrimental effects on DIO phenotypes. Both rimonabant and AM251 are effective in reducing/preventing obesity in a number of rat nd mouse models as reviewed previously [33]. Structurally, they are similar, differing only in a single halogen substituent. Both are potent and selective CB1 receptor antagonists, and can both act as inverse agonists at higher concentrations. However, there are differences in their selectivity profiles, with AM251 being ~1000 fold more selective for the mouse CB1 vs. the mouse CB2, as opposed to rimonabant being only 25 fold more selective [34]. There are also reported differences in their off-target activities, with Soethoudt et al. [34] showing that AM251 had activity at 16 off-target receptors including GPR55, whereas rimonabant only had 8 off-target activities. An additional level of complexity may arise from ligand biased signaling, which is the ability of different ligands to differentially activate different receptor signaling effectors/responses. Ligand biased signaling at the CB1 receptor has been reported [35] although to our knowledge there are no studies comparing bias of AM251 vs. rimonabant. Given that research into AM251 and other CB1 antagonists/inverse agonists is still continuing, it will be important for future studies to determine if there are critical differences between these ligands and rimonabant, especially its effect on adverse psychiatric events.

The data from the current study demonstrate the changes in vivo and ex vivo following treatment with the global CB₁ inverse agonist/antagonist, AM251. This study has shown reduced epididymal, peri-renal and brown fat pad mass, and circulating levels of leptin, glucagon, ghrelin, and GLP-1. In skeletal muscle, AM251 treatment for six weeks, or when the muscle from these animals were exposed to ex vivo acute treatment with adiponectin, did not alter mRNA expression of genes associated with adiponectin signaling or oxidative capacity. We also saw no evidence for the browning of WAT in chronically treated AM251 rats on an HFD, which suggests that browning of WAT is not a mechanism contributing to the anti-obesity effects of AM251. We also observed that AM251 treatment resulted in an increase in pro-inflammatory cytokines, representing whole body systemic inflammation. Thus, despite AM251 having a positive impact on obesity with respect to food intake, weight loss and fat pad weight, it results in systemic inflammation bringing into question its continued use in research.