Effect of dietary krill oil supplementation on the endocannabinoidome of metabolically relevant tissues from high-fat-fed mice

The endocannabinoid system, and in particular the cannabinoid receptor type 1 (CB₁), are profoundly involved in the regulation of energy balance, at both the central and peripheral levels [1, 2]. Apart from CB₁, this system includes the cannabinoid receptor type 2 (CB₂) [3, 4], the endogenous agonists, N-arachidonoylethanolamine (anandamide, AEA) and 2-arachidonoyl glycerol (2-AG), known as endocannabinoids (ECs) [5‐7], and enzymes for EC biosynthesis and degradation [8‐11]. ECs are nominally synthesized from arachidonic acid (AA), but are phospholipid (PL)-derived mediators. AEA and 2-AG are synthesized, respectively, through the N-acylphosphatidylethanolamine-selective "phospholipase D" (NAPE-PLD), among others, and the diacylglycerol lipases (DAGL). The main enzymes for EC degradation are fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL), for AEA and 2-AG, respectively. NAPE-PLD can also produce other N-acylethanolamines (NAE), like N-palmitoylethanolamine (PEA) and N-oleoylethanolamine (OEA), from the corresponding N-acylphosphatidylethanolamines, and these two compounds are also good substrates for FAAH [12]. PEA and OEA do not bind CB₁ and CB₂ receptors, but act via other receptor types, such as the peroxisome proliferator-activated receptor-α (PPAR-α) [13] and the transient receptor potential vanilloid type-1 (TRPV1) channel [14].

The beneficial effects of omega-3 polyunsaturated fatty acids (ω-3 PUFAs) on the risk of developing cardiovascular diseases (CVD) [15] and metabolic disorders, such as insulin resistance, fatty liver and hypertension [16], have been extensively studied. The most studied ω-3 PUFAs are eicosapentaenoic acid (EPA, 20:5) and docosahexaenoic acid (DHA, 22:6), present in the diet as such or biosynthesized from α-linolenic acid (ALA, 18:3) [17]. In contrast, derivatives of ω-6 PUFAs, originated from linoleic acid (LA, 18:2), and of which the major exponent is AA (ω-6 20:4), are known for their pro-inflammatory and other disease-propagating effects [18]. Both ω-3 and ω-6 PUFAs precursors are essential fatty acids because they cannot be synthesized de novo in mammals and have to be obtained through the diet or from dietary precursors. Major sources of ω-3 PUFAs are fish oils, whereas other types of oil, such as flaxseed oil, are a good source of ALA [19]. Krill oil (KO) is a relatively new source of ω-3 PUFAs, extracted from Antarctic krill (Euphasia superba) [20], a marine crustacean with the most abundant biomass on earth [21]. The chemical composition of KO is unique. It is rich in ω-3 PUFAs in the form of PL rather than triglycerides (TG) and this may influence the bioavailability and tissue incorporation of ω-3 PUFAs [22]. Moreover, KO contains a powerful antioxidant, astaxanthin, a lipid-soluble pigment that preserves KO from oxidation. This special composition possibly underlies in part the health-promoting effects of KO, such as its anti-inflammatory and hypolipidemic properties in humans [23, 24].

The levels of the ECs and their lipid congeners, in peripheral tissues controlling energy homeostasis, are altered in obese animals and humans, particularly when increased adiposity is the consequence of a prolonged high fat (HF) diet, and this alteration contributes to obesity-related metabolic disorders [25]. The dysregulation of EC concentrations can be due in part to changes in the activity and/or expression of EC biosynthetic enzymes, but also to alterations in EC ultimate PL biosynthetic precursors, which in turn are sensitive to quantitative and qualitative changes in dietary fats. It was reported that dietary ω-3 PUFAs in rats can alter NAE levels in several tissues [26], whereas, in vitro, AA, DHA and EPA influence the AA content of PL in isolated mouse adipocytes, thereby modulating the levels of EC biosynthetic precursors and of ECs in these cells [27]. On the other hand, increased dietary fat in mice is well known to lead not only to obesity and increased adiposity, but also to several metabolic sequelae, such as insulin resistance, dyslipidemia and increased expression in hepatocytes of sterol regulatory element binding protein 1c (SREBP-1c), and of its target proteins acetyl-CoA-carboxylase-1 (ACC1) and fatty acid synthase (FAS), with subsequent enhanced hepatic de novo lipid biosynthesis [28]. These effects are due in part to elevation of peripheral EC levels and subsequent over-activation of CB₁, since CB₁ knock-out mice, or mice chronically treated with CB₁ antagonists, exhibit little or no response to high fat diets in terms of body weight gain and increased adiposity, fatty liver, insulin resistance and dyslipidemia [28]. Importantly, dietary KO was recently shown to reduce fatty liver and systemic inflammation in obese Zucker rats through the negative modulation of EC and EC PL precursor levels [29].

We have recently observed that a shorter duration (8 instead of 14 weeks) of a certain type of HF diet with reduced levels of ω-3 and, particularly, ω-6 PUFA fatty acid precursors, and higher cholesterol levels, instead decreases hepatic FAS and ACC1 expression in mice, with no change in SREBP1 expression nor in plasma glucose, non-esterified fatty acids (NEFA) and adiponectin levels [30]. This regimen, however, still causes fatty liver, increased adiposity and hypercholesterolemia, all of which are still attenuated by increasing levels of dietary KO. In the present study, we aimed at verifying whether this HF diet with a different fat composition and a shorter duration with respect to that previously used [28], may have affected EC tissue levels in a different way, and, subsequently, if the beneficial effects caused by KO in this case [30] were still due to normalization of EC dysregulation. Moreover, by using sensitive lipidomic approaches based on either single-quadrupole MS or high resolution mass spectrometric (MS) and MS-MS detection [31], both combined with liquid chromatography, we analyzed, under the different dietary regimens, the levels of EC biosynthetic precursors, as well as of anandamide congeners, OEA and PEA.

The dietary unbalance among macronutrients leads to metabolic derangement of glucose and lipid disposal characterised by a marked dyslipidemia, increased insulin resistance and fatty liver, which are some of the characteristic features of the metabolic syndrome. Whether not only the amount but also the composition of dietary fat may influence the development of the metabolic syndrome, mediated by the EC system, has been recently investigated [29]. In the present study, 8 week administration of a HF diet low in the n-6 PUFA precursor LA, led to a significant elevation of body weight and liver weight, as well as of hepatic TG, cholesterol and TNFα levels, and serum TG, cholesterol and PL levels. On the other hand, we did not observe induction of high fasting glycemia and hypoadiponectinemia, contrary to the findings previously obtained using a different type of HF diet, which had significantly higher LA content [28]. At the same time, the HF diet used here reduced, instead of increasing, EC levels in the liver, in contrast to what previously described using the other HF diet [28], which led to significantly enhanced hepatic AEA levels. The only explanation that we have for this discrepancy lies in the different n-6 PUFA precursor content of the HF diets used by us and Osei-Hyiaman and colleagues, who observed the strongest effect on AEA levels after 3 weeks of the dietary regimen. While these authors used partially hydrogenated oils (PHO), we used butter. These two fats have a dramatic different LA content (about 4% with PHO vs about 0.4% with butter). It has been pointed out before that the effect of HF diet on peripheral EC levels greatly depend on the fatty acid composition of the diet [33]. Chronic administration of a diet such as the one used in the current study, with a lower amount of LA than in a normal diet, may have caused a net decrease in hepatic AA, and hence in the levels of the ultimate PL biosynthetic precursors of ECs. However, as assessed by the "endocannabinoidomics" methodology used here, we could not see any decrease in the levels of AEA and 2-AG direct biosynthetic precursors following the HF diet, thus suggesting that strongly different fatty acid compositions of prolonged dietary regimens might affect EC levels also at other levels (see below). The higher amount of cholesterol in the HF diet used here is, instead, unlikely to have caused any effect on EC levels, since we have previously shown that an even more hypercholesterolemic diet affects tissue EC concentrations in mice only when it also provokes accumulation of atherosclerotic plaques [34].

Interestingly, a further change of dietary fat composition imposed with KO supplementation during the HF diet resulted in the elevation of both AEA (at both the lowest and highest dose) and 2-AG (only at the highest dose) hepatic levels, thus clearly dissociating the hepatic disorders accompanying HF diet-induced obesity from EC overactivity, at least with the type of HF diet used here. Previously, we had shown that in the liver of obese Zucker rats, in which elevated hepatic levels of both AEA and 2-AG occur [35], amelioration of hepatic steatosis following dietary KO was accompanied by a strong reduction of AEA levels, but also by an increase of the tissue concentration of 2-AG, which, in both species, is significantly more abundant than AEA. These data, taken together, might suggest that EC overactivity in the liver is not necessarily a determining factor for obesity-associated liver steatosis and underlying inflammation in rodents. On the other hand, contrary to the HF diet used by Osei-Hyiaman et al. [28], the dietary regimen employed here caused a reduction, rather than an increase, of two de novo lipogenesis-associated genes, FAS and ACC1, and this effect was potentiated by KO. These data can now be partly accounted for by the present finding of decreased or increased EC levels after the HF diet or the HF diet plus KO, respectively, and support the hypothesis that hepatic de novo lipogenesis is instead strongly sustained by ECs via CB₁ receptor activation [28] but also counteracted by KO independently from the EC system. Clearly, obesity- and HF diet-associated-hepatosteatosis in mice is only part due to de novo lipogenesis in hepatocytes, and might be caused principally by other factors (such as fatty acid flux from the adipose tissue).

We found here that, contrary to EC levels, KO increases the levels of the PPARα activator, PEA, albeit in a non-dose-dependent manner. Given the beneficial effects of PPARα agonists on liver steatosis and inflammation [36], it is possible that the beneficial effects of KO in this case are due to its stimulatory action on this compound. However, KO reduced the levels of the other PPARα activator, OEA, thus possibly resulting in an overall null effect on PPARα activity. Importantly, like with the HF diet, also the effects of KO on hepatic EC, PEA and OEA levels did not appear to be largely accounted for by effects on the direct biosynthetic precursors of these compounds. Instead, the HF diet-induced increase of OEA levels and decrease in 2-AG levels might be explained by our previous finding [28] that this dietary regimen: 1) on the one hand, increases the expression of both CD36 and stearoyl-CoA desaturase, two proteins involved in the uptake and biosynthesis, respectively, of oleic acid, the ultimate precursor of OEA; and 2) on the other hand, increases also the expression of the monoacylglycerol lipase, the major enzyme involved in 2-AG degradation. These effects were significantly counteracted by KO [28], which, as mentioned above, also reduced and enhanced hepatic OEA and 2-AG, respectively, thus supporting the potential role of these proteins and enzymes in determining the levels of these two lipid mediators in the liver. KO at the highest dose, however, also increased the levels of 18:0-20:4-DAG, which corresponded to increased 2-AG levels.

Another major difference between the results of the present study and that of others [37], possibly attributable to the difference in dietary fat composition, was the fact that here the 8-week HF diet did not cause elevation of 2-AG levels in the epididymal AT (a type of adipose depot similar to intra-abdominal adipose tissue in humans), although it did provoke a reduction of EC levels in the inguinal AT similar to that observed in other subcutaneous adipose depots of both obese rodents [35, 38] and humans [39, 40]. The HF diet used here caused instead a significant reduction of AEA levels in the epididymal AT. Interestingly, unlike other types of HF diet, the one used here also did not induce high fasting glycemia or reduce serum adiponectin levels [30], thus strengthening the link between excessive EC tone in the visceral AT, or lack thereof, and insulin resistance or sensitivity, respectively. In this case, the relatively low levels of LA present in our HF diet are unlikely to have caused an increase of AA in PL biosynthetic precursors of ECs, since both NArPE and sn-2-arachidonoyl-containing DAGs were increased, rather than decreased, by the HF diet. Despite the lack of stimulatory effect of the HF diet on epididymal AT EC levels, dietary KO strongly reduced 2-AG levels in the epididymal AT, and also down-regulated both AEA and 2-AG levels in the inguinal AT, whilst reducing fasting glycemia and elevating adiponectin levels. These data, taken together with the little effect of KO on PEA or (with the exception of the epididymal AT) OEA levels, suggest that EC tone in the white AT, rather than in the liver, might be an important determinant of both insulin resistance and ectopic (hepatic) fat formation. The lowering of AT EC levels by dietary KO might, therefore, represent an interesting alternative to CB₁ receptor antagonists to counteract the development of pre-diabetes and hepatosteatosis. Importantly, unlike the effects exerted by the HF diet per se, the alterations (or lack thereof) caused by KO on EC, PEA and OEA levels in both AT depots analysed here could be mostly accounted for by similar effects on the direct biosynthetic precursors of these mediators.

Excessive EC levels in the skeletal muscle was suggested as another potentially important cause of high fasting glycemia [33]. In fact, activation of CB₁ receptors in myotubes modifies insulin signalling, reduces sensitivity to insulin and impairs glucose uptake [41]. Furthermore, elevation of EC levels triggers an inhibition of AMPK1α and, subsequently, of glucose uptake and oxidation from the skeletal muscle [42]. Here we found, in the gastrocnemius skeletal muscle of HF diet-fed mice, a significant elevation of 2-AG levels. Likewise, in a previous study, Matias and collaborators [33] found an increase of 2-AG levels after a 3-week regimen with another type of HF diet and, although this effect was not significant at 8 weeks of regimen, it was observed still after 14 weeks. More importantly, we observed here that dietary KO supplementation decreased 2-AG levels in a dose-dependent manner and also reduced the levels of one of the DAGs analysed, suggesting a beneficial role of KO at enhancing energy expenditure and reducing fasting glucose levels. Furthermore, we observed that KO supplementation was accompanied also by elevation of PEA levels. Given the role of PEA as anti-inflammatory agent [43], and the previous observation that omega-3 PUFAs exert a protective effect against muscle damage induced by the pro-inflammatory cytokine TNF-α [44], we speculate that increased PEA levels might protect skeletal muscle from the damaging effect of TNF-α, and contribute, together with KO-induced (an possibly AT EC decrease-mediated) elevation of adiponectin levels, to the anti-inflammatory effects of KO. Our "endocannabinoidomic" profiling data also indicate that the changes caused by HF diet or KO on EC levels in the skeletal muscle are mostly accounted for by similar effects on the direct biosynthetic precursors of these mediators.

Another novel observation of this study was the finding that dietary KO during a HF diet can counteract the effect of the latter on heart EC levels. HF-induced elevation of cardiac EC levels was observed also previously using a different HF diet [33], and we have shown here that KO dose-dependently reduces AEA levels, and is effective at re-establishing normal 2-AG levels already at the lowest dose tested, whilst producing little, if any, effect on PEA and OEA levels. ECs directly affect cardiac function by acting at CB₁ and CB₂ receptors and, in the case of AEA and OEA, potentially also via TRPV1 receptors [45]. Although the subsequent overactivity of CB₁ receptors might enhance lipogenesis in the heart, thus contributing to formation of pericardial fat, which is a risk factor for cardiovascular events [33], the endocannabinoid system may also be activated as a compensatory mechanism in various forms of hypertension to limit pathologically increased blood pressure and myocardial contractility [46]. In obese Zucker rats, KO was previously shown to reduce cardiac AEA levels, whilst elevating 2-AG levels, and concomitantly reducing TG accumulation in the heart. Although we did not measure here fat accumulation in the heart of HF-treated mice, nor the amount of pericardial AT, we can speculate that KO-induced reduction of cardiac EC levels might contribute to reducing ectopic fat formation in this organ, and hence cardiovascular risk. Our "endocannabinoidomics" results also suggest that the decrease of EC levels carried about by KO can be due to lesser availability of EC biosynthetic precursors.

In 2007, Janiak and collaborators showed that chronic blockade of CB₁ receptors could reduce renal failure caused by obesity and type 2 diabetes in Zucker rats in a way partly independent from inhibition of food-intake [47]. Furthermore, in an another study Matias et al. demonstrated that 8-week HF fed mice caused a strong elevation of AEA levels, possibly contributing to decreased glomerular filtration rate and increased renal blood flow typical of obesity [32]. In the present study, both AEA and 2-AG levels increased after the HF diet, although only 2-AG in a statistically significant manner. Importantly, the kidney levels of both compounds were strongly decreased after KO supplementation, which instead did not so strongly affect OEA and PEA levels. Since renal EC overactivity could contribute to kidney malfunctioning and cardiovascular disease, in part due to renal AT formation [48], KO could be a promising alternative to pharmacologic treatment for the renal complications of obesity and the metabolic syndrome. Importantly, whilst the effects of KO on kidney 2-AG levels were reflected by its effects on the DAG precursors of this compound, KO supplementation did not significantly affect the levels of the direct AEA biosynthetic precursor, NArPE, although a trend towards its decrease was observed.

In conclusion, our data have shown that also in mice with HF-induced obesity and dysmetabolism, n-6 PUFAs dietary content influences EC precursors and biosynthesis, and that the addition of KO to the diet can ameliorate several metabolic disturbances and reduce EC levels in most of those peripheral organs the malfunctioning of which is responsible for such disturbances. Furthermore, by using a novel lipid profiling methodology, we have confirmed the hypothesis according to which KO-induced reduction of the levels of AEA and/or 2-AG is at least in part due to reduction of the tissue levels of the corresponding direct precursors of these mediators. Finally, again by carrying out the simultaneous quantification of N-acylethanolamine activators of PPARα, such as PEA and OEA, we have shown that KO affects the levels of these mediators, although only in some of the analysed tissues. Since these two mediators are dysregulated in several tissues of obese Zucker rats [49], these data might suggest that KO can potentially produce beneficial metabolic effects against dysmetabolism and inflammation in obesity also by re-equilibrating the activity of PPARα.