Animals and keeping

One week before the experiment set off, male Wistar rats (Harlan-Italy, Udine, Italy), weighing 210 ± 20 g (mean ± SD) were housed under controlled temperature (21 ± 2 °C), relative humidity (60 ± 5%) and artificial 12 h light/dark cycle, avoiding all stressful stimuli. Animal handling and care throughout the experimental procedures met with national (Legislative Decree n. 26, 04/04/2014) and international (Directive 2010/63/EU in Europe) laws and policies. The experimental protocols were carried out in compliance with the guidelines of the Animal Ethics Committee of the University of Cagliari. Standard laboratory food (A04, Safe, Augy, France) and water were freely available ad libitum.

According to the optimum standard for the evaluation of lipids in tissue and plasma [41, 42], animals received no food for 12 h before surgery.

Rats (n = 112) were randomly assigned to 2 groups that received a pre-treatment 6 h before the surgery: one group (vehicle-treated; n = 60) was pre-treated with the vehicle, i.e. 0.3 ml of sunflower oil, while the other group (BCP-treated; n = 52) was previously gavage-fed (with the help of feeding needle) with 40 mg of BCP (Sigma-Aldrich, St Louis, Mo, USA) (corresponding to 180 mg/kg), dissolved in 0.3 ml of sunflower oil. Each group was further subdivided into sham-operated rats- represented by animals that underwent surgery without common carotid arteries (CCA) occlusion- and rats submitted to BCCAO/R and processed for lipid analysis in brain tissue and plasma (vehicle-treated n = 24; BCP-treated n = 20), western blot (vehicle-treated n = 24; BCP-treated n = 20) in brain tissue homogenates, and immunohistochemistry (vehicle-treated n = 12; BCP-treated n = 12) in brain tissue sections.

Surgery

Surgical procedure for induction of BCCAO/R was adapted from the method of Iwasaki et al. [43] and performed in all cases between 13:00 and 16:30 p.m.. Rats were anesthetized with intraperitoneal administration of Equitesin (4.2% w/v chloral hydrate, 2.12% w/v MgSO4, 16.2% w/w pentobarbital, 39.6% w/w propylene glycol, and 10% w/w ethanol in sterile distilled H₂O) (5 ml/100 g bodyweight). After a midline cervical incision and blunt dissection of muscles, the CCA were exposed caring to leave the vagus nerve intact. Cerebral blood flow reduction was produced by placement of two atraumatic microvascular clips for 30 min on CCA. The reperfusion period was achieved by removing the clips and restoring blood flow through the stenosed vessels for 60 min. The control animals, used to determine the effects of anaesthesia and surgical manipulation on the results, were represented by sham-operated rats.

Sampling

At the end of the procedure, brain samples were collected either as fresh tissue for lipid analysis and Western blot or after transcardial perfusion fixation with ice cold 0.1 M phosphate buffer (PB), pH 7.4 for immunohistochemistry. The frontal cortex was rapidly dissected out by a transverse cut made at the level of the optic chiasm, at the approximate bregma level of − 1.0 mm [44], and frozen at − 80 °C until HPLC or western blot analysis. Temporal-occipital cortex, dissected out by a transverse cut at the approximate bregma level of − 4.5 mm, was also sampled as a control cortical area not irrorated by the internal carotid artery branches. Blood was quickly collected from the trunk of killed animals into heparinised tubes and centrifuged at 1500 g for 10 min at 8 °C. The resulting plasma was frozen at − 20 C° until assayed for lipids. Perfused brains were dissected out, fixed by overnight immersion in 4% formaldehyde in 0.1 M phosphate buffer (PB), pH 7.4 and then rinsed in 0.1 M PB, pH 7.3, containing 20% sucrose. After sucrose infiltration, samples were embedded in Optimal Cutting Temperature (OCT) medium for cryostat sectioning. For each assay, the investigator was blind with respect to the experimental condition of rats.

Endocannabinoid and congener quantification

Frozen tissues were homogenized and extracted with 50 mM chloroform/methanol/Tris-HCl, pH 7.5 (2:1:1, v/v), containing internal deuterated standards for AEA, 2-AG, PEA and OEA quantification by isotope dilution ([2H]⁸ AEA, [2H]⁵ 2-AG, [2H]⁴ PEA, [2H]⁴ OEA; Cayman Chemical, Ann Arbor, MI, USA). AEA, 2-AG, PEA, and OEA were quantified by liquid chromatography–atmospheric pressure chemical ionization–mass spectrometry [1100 HPLC system (Agilent Technologies, Santa Clara, CA, USA) equipped with MS Detector 6110 single quadrupole] and using selected ion monitoring at M1 values for the four compounds and their deuterated homologs, as described previously [41]. Concentrations (nmoles/g; nmoles/ml) are shown as histograms in Figs 1, 2.

Measurement of fatty acid composition of tissue phospholipids

Total lipids were extracted from different brain areas using chloroform/methanol 2:1 (v/v). Aliquots were mildly saponified as previously described [45] in order to obtain free fatty acids for high-performance liquid chromatography (HPLC) analysis. Separation of fatty acids was carried out with an Agilent 1100 HPLC system (Agilent Technologies) equipped with a diode array detector as previously reported [42]. Concentrations (nmoles/g; nmoles/ml) are shown as histograms in Figs 1, 2.

Western blot

Tissue homogenate were prepared in a 2% solution of sodium dodecyl sulfate (SDS) containing a cocktail of protease inhibitors (cOmplete, Mini Protease Inhibitor Cocktail Tablets, Roche, Basel, Switzerland). Protein concentrations were determined using the Lowry method of protein assay [46] with bovine serum albumin as standard. Proteins for each tissue homogenate (40 μg), diluted 3:1 in 4X loading buffer (NuPAGE LDS Sample Buffer 4X, Novex by Life Technologies, Carlsbad, CA, USA), were heated to 95 °C for 7 min and separated by SDS-polyacrilamide gel electrophoresis (SDS-PAGE) using precast polyacrylamide gradient gel (NuPAGE 4–12% Bis-Tris Gel Midi, Novex by Life Technologies) in the XCell4 Sure Lock™ Midi-Cell chamber (Life Technologies). Internal mw standards (Precision Plus Protein™ WesternC™ Standards, Bio-Rad, Hercules, CA, USA) were run in parallel. Two gels at a time were run for Coomassie staining and immunoblotting, respectively. Proteins for immunoblotting were electrophoretically transferred on a polyvinylidene fluoride membrane (Amersham Hybond™-P, GE Healthcare, Little Chalfont, United Kingdom) using the Criterion™ Blotter (Bio-Rad). Blots were blocked by immersion in 20 mM Tris base and 137 mM sodium chloride (TBS) containing 5% milk powder and 0.1% Tween 20 (TBS-T), for 60 min at room temperature and incubated overnight at 4 °C with rabbit polyclonal antisera directed against CB1 receptor (Synaptic System, Göttingen, Germany), diluted 1:500, CB2 receptor (Cayman Chemical, Ann Arbor, Mi, USA), diluted 1:1000, PPAR-α (Thermo Scientific, Waltham, MA, USA), diluted 1:1000, and COX-2 (residues 570–598) (Cayman Chemical), diluted 1:200 in TBS containing 5% milk powder and 0.02% sodium azide, used as primary antisera. After TBS-T rinse, blots were incubated for 60 min, at room temperature, with peroxidase-conjugated goat anti-rabbit serum (Sigma Aldrich), diluted 1:10,000 in TBS-T. Loading controls were obtained by stripping and immunostaining the membranes with a mouse monoclonal antibody against the housekeeping protein glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (EMD Millipore, Darmstadt, Germany), diluted 1:1000, as primary antiserum, and a peroxidase-conjugated goat anti-mouse serum (EMD Millipore), diluted 1:5000, as secondary antiserum. In order to control for non specific staining, blots were stripped and incubated with the relevant secondary antiserum. After TBS-T rinse, protein bands were visualized using the ECL chemiluminescent system according to the protocol provided by the company (GE Healthcare), under ImageQuant LAS 4000. Approximate molecular weight (mw) and relative optical density (O.D.) of immunolabelled protein bands were evaluated by a “blind” examiner, and were quantified by comparing the position of relevant bands on the digital images with those of the GAPDH bands, respectively. The ratio of the intensity of CB1-, CB2-, PPAR-α- and COX-2-positive bands to the intensity of GAPDH-positive ones was used to compare relative expression levels of these proteins following BCCAO/R procedure. The O.D. was quantified by Image Studio Lite Software (Li-Cor, http: //www.licor.com/bio/products/software/image_studio_lite/) and is shown as histograms in Fig. 3.

Immunohistochemistry

The avidin–biotin–peroxidase complex (ABC) technique was used to process cryostat semiconsecutive sections (16 μm thick), collected on chrome alum-gelatin coated slides. Coronal serial sections at + 4.70 to - 1.70 mm and at − 3.5 to - 8.0 mm Bregma levels, respectively, were used to focus the observations to frontal and temporal-occipital cortex (Paxinos and Watson, 2007). The endogenous peroxidase activity was blocked with 0.1% phenylhydrazine in phosphate buffered saline (PBS) containing 0.2% Triton X-100 (PBS/T) followed by incubation with 20% of either normal goat or normal horse serum (Vector, Burlingame, CA, USA) for 1 h at RT and then incubated with rabbit polyclonal antibody against CB1 (Synaptic System, Germany), diluted 1:1000, and against COX-2 (Cayman Chem., USA), diluted 1:300, as primary antibodies. Incubations with primary antiserum were carried out overnight at 4 °C. Biotin-conjugated goat anti-rabbit serum (Vector, Burlingame, CA, USA), diluted 1:400, was used as secondary antiserum. The ABC (BioSpa Div. Milan, Italy), diluted 1:250, followed by a solution of 0.1 M PB, pH 7.3, containing 0.05% 3,3′-diaminobenzidine (Sigma, Milan, Italy), 0.01% hydrogen peroxide and 0.04% nickel ammonium sulfate was used to reveal the reaction product. Incubations with secondary antiserum and ABC lasted 60 min and were performed at RT. Negative control preparations were obtained by incubating tissue sections in parallel with either PBS-T alone or with the relevant primary antiserum preabsorbed with an excess of the corresponding peptide antigen. Alternate sections were stained with modified Mayer’s hematoxylin. Slides were examined by the same examiner blinded to animals’ treatment with an Olympus BX61 microscope and digital images were captured with a Leica DF 450C camera.

Statistical analysis

HPLC and Western blot data from the four experimental subgroups, i.e. vehicle- and BCP-treated sham-operated animals, and vehicle- and BCP-treated BCCAO/R rats, have been analyzed by two way analysis of variance (ANOVA) [main factors: a) BCP-treatment (i.e. vehicle- vs BCP-treatment) and b) BCCAO/R (i.e. sham-operation vs BCCAO/R] by using GraphPad Prism 7.03 for Windows (GraphPad Software, La Jolla California USA). Wherever appropriate (i.e. p for the main factors and their interaction < 0.05), multiple pair-wise contrasts were made and multiplicity adjusted p value for each comparison was calculated using the Bonferroni’s post hoc test.

Analysis of eCB and fatty acid profiles in brain tissue

eCB and fatty acid profiles in plasma

BCP- vs vehicle-treated differences were detected in the control animals for the AEA (− 64%; post hoc p <  0.0013), and in pair-wise contrasts of BCCAO/R-BCP-treated vs BCCAO/R-vehicle-treated rats, that showed a highly significant decrease of the AEA (− 56%; post hoc adjusted p <  0.0001) and lipoperoxides (− 38%; post hoc adjusted p = 0.0219) concentrations after BCP-treatment (Table 1; Fig. 2).

Western blot

The effects of the administration of BCP on the concentrations of CB1, CB2, PPAR-alpha and COX-2 are reported in Table 1 and graphically shown in Fig. 3. As in the case of lipids, WB analysis of the tissue homogenates demonstrated that molecular changes were observed only in the frontal cortex. The effect of BCP-pre-treatment was evident in the pair-wise comparison of the sham+BCP vs sham+vehicle rats, with increase in the relative protein levels of CB1 (+ 53%; post hoc adjusted p 0.0481), and in the pair-wise contrast between the BCCAO/R rats, where PPAR-alpha levels significantly increased in BCCAO/R-BCP- vs BCCAO/R-vehicle-treated animals (+ 130%; post hoc adjusted p <  0.0001) (Table 1; Fig. 3). The post hoc analysis also showed an effect of the BCCAO/R per se; thus, relative protein levels increased in BCCAO/R- vs sham-vehicle-treated rats in the case of CB1 (+ 101%; post hoc adjusted p = 0.0279) and increased in BCCAO/R + BCP vs sham+BCP animals in the case of CB1 (+ 100%; post hoc adjusted p = 0.0433), CB2 (− 50%; post hoc adjusted p = 0.0010), and PPAR-alpha (+ 64%; post hoc adjusted p < 0.0001) (Table 1; Fig. 3). Consistently, BCP treatment x BCCAO/R interaction was observed for CB2 (p = 0.0401) and PPAR-alpha (p = 0.0163) (Table 1). As for COX-2, pair-wise comparisons showed no significant differences, however an interaction between the two ANOVA main factors was observed (p = 0.0299) (Table 1). No statistically significant changes were observed in the temporal-occipital cortex (data not shown).

The antibodies against CB1 and COX-2 were the only ones to produce a reliable immunostaining in tissue sections of rat cerebral cortex. For this reason, the following immunohistochemical data are based exclusively on the immunoreactivity obtained with them.

Immunohistochemistry

In order to associate the molecular changes observed by HPLC and western blot analyses and the tissue morphology, immunoreactivities to CB1 and COX-2 were also examined in the cerebral cortex (Figs. 4, 5). All markers labeled neuronal structures distributed throughout the rostro-caudal extension of the frontal cortex (Figs. 4, 5) and the temporal-occipital cortex (data not shown).

The CB1 receptor-antibody labelled a dense to moderate plexus of beaded fibers and some neuronal cell bodies distributed throughout the cortical layers in both vehicle- and BCP-treated rats (Fig. 4). As a general rule the density of labelled nerve terminals and fibers appeared higher in BCCAO/R animals (Fig. 4b, d) than in the sham-operated ones (Fig. 4a, c).

The antibody against COX-2 labelled mainly neuronal perikarya, that showed different density of intracytoplasmic immunostaining, and proximal neuronal processes (Fig. 5). Immunolabelled neurons were frequent in cortical layers II/III and V. As a general rule, vehicle-pre-treated rat brains showed that the staining intensity and density of labelled structures were higher in BCCAO/R (Fig. 5c, d) than in sham-operated rat brains (Fig. 5 a, b), whereas this difference was not evident in the BCP-pre-treated rats (Fig. 5 e-h).

BCP effects on the eCB system following hypoperfusion/reperfusion

Our results, demonstrating that a single dose of BCP may modulate lipid metabolism and related receptor proteins, are in agreement with pharmacokinetic data [51] and with data on the effectiveness of BCP-pretreatment and extend our previous observations on the beneficial effects of the Pistacia lentiscus L. essential oil in slightly different experimental conditions [3].

Several BCP properties that have been recently reported can explain the molecular changes we observed in the rat BCCAO/R model [20, 30, 31, 34, 35, 37, 38, 52–54]. Indeed, BCP is a selective agonist for CB2 and PPAR-alpha receptors [30–32, 55] and, in addition, has been shown to inhibit fatty acid amide hydrolase (FAAH) [20, 31, 54] and COX-2 [52].

It is known that brain ischemia and reperfusion activate the ECS, by driving a combination of biochemical adaptations of the NAE degrading and synthesizing enzymes that, collectively, lead to the accumulation of eCBs [12, 56]. Similarly, the BCCAO/R model of transient global hypoperfusion/reperfusion activates the ECS [present data; 4]. Among the eCB congeners, PEA represents an important player of the endogenous defense against neuroinflammation [11, 57–59]. Importantly, beyond the modulation of the immune cells [59], PEA exerts its neuroprotective effects by activating PPAR-alpha [23, 24] and blunting COX-2 activity [11]. Thus, exogenous administration of PEA in an acute stroke model is effective in reducing the infarct size [7]. Moreover, in stroke patients eCBs and congeners, including PEA, have been shown to increase during the acute ischemic phase during which they may play a role through multiple potential mechanisms [17]. Our data indicate that, after a single dose of BCP, the tissue levels of PEA, as well as levels of AEA and OEA, increase in the sham-operated, while, as a general rule, they do not change after induction of BCCAO/R.

Though it is difficult to speculate on the possible biological significance of the increase of AEA, PEA and OEA levels in basal conditions, it can be suggested a scenario where BCP, known to elicit a full agonist action on CB2 receptors and PPAR-alpha and gamma isoforms [32, 55], interferes with the endogenous signaling of eCB and congeners by adjusting their relative levels in both basal conditions and after BCCAO/R. Interestingly, it has been shown that PPAR-alpha induction elicits an increase of PEA and OEA [60]. Thus, in the presence of exogenous BCP, levels of AEA, 2-AG and PEA in the BCCAO/R rat frontal cortex are either unchanged or reduced; these findings, together with the increases of CB receptors and PPAR-alpha and the marked decrease of lipoperoxides, and the interaction between BCCAO/R and BCP-treatment observed for the COX-2 expression lead to suggest that a modulation of the endogenous anti-inflammatory milieu is occurring in the tissue challenged by the BCCAO/R-induced oxidative stress [see 23]. In our basic experimental setting it is not feasible to speculate whether the PPAR-alpha receptor is induced by its NAE ligands, including PEA, or viceversa. However, it has been suggested that enhancing the tissue availability of PEA through inhibition or modulation of its enzymatic breakdown may represent a complementary therapeutic approach to counteract neuroinflammation [20, 56, 60]. Thus, it is tempting to speculate that BCP may increase the eCB and congeners by activating PPAR-alpha [32, 55, 61], and initiate a preventive action on some aspects of the tissue physiological response to the BCCAO/R-induced oxidative stress.

The present data are also consistent with previous results obtained in serum from patients undergoing carotid endoarterectomy, where increased levels of lipoperoxides and isoprostanes and a concomitant increase of their catabolism in peroxisomes have been shown to be directly correlated to the hypoperfusion/reperfusion-induced oxidative challenge [62]. In particular, it is interesting that peroxisomal beta-oxidation increased during the first 30 min of reperfusion only in patients having contralateral carotid stenosis higher than 50% [62]. On the other hand, recent polypharmacological studies have pointed out that, in a therapeutic setting, a critical role is played by the synergistic modulatory effects due to both the modulation of CB2 receptors and the increase of AEA levels as a consequence of the inhibition of its main degrading enzyme [31]. Thus, more specific experiments should be carried out to evaluate whether, upon specific inhibition of FAAH and/or combined inhibition of COX-2, the BCP effects are confined to brain.

Interestingly, our data further show that BCP decreases AEA levels in plasma. Though it is difficult to speculate whether the plasmatic decrease of AEA is related to the eCB modifications in brain tissue, it can be hypothesized that the general decrease of peripheral AEA observed in both the sham+BCP and the BCCAO/R + BCP animals may reflect the integrity of brain tissue. Clinical studies support the notion that neurological and neuropsychiatric disorders are characterized by detectable changes in eCB plasma levels and that the AEA and PEA plasma levels are correlated with neurological disability, so that patients with higher AEA and PEA levels had greater neurological impairment [17 and refs therein].

In this study, we have also demonstrated that, after BCCAO/R, higher concentration of eCBs are associated with an increase of the relative levels of CB1 and CB2 receptor proteins. Both these receptors may contribute to the physiological tissue response to the BCCAO/R challenge. As already proposed [63–65], a parallel raise in eCBs and CB1 receptors indicates a sensitization of the cannabinoidergic system that may contribute to regulate cellular functions that depend upon CB1 receptor activation. Thus, after BCCAO/R, this sensitization would be able to modulate events such as neurotransmitter release, calcium cellular influx, oxidative stress damage, and vascular tone that appear to be crucial in the response of cerebral tissue to the hypoperfusion/reperfusion insult. Interestingly, it has been reported that CB2 remains inactive under physiological conditions [39, 53] and that its activation is crucial in injury models [66, 67], such as middle cerebral artery occlusion (MCAO) and reperfusion [39, 53]. Thus the BCP pre-treatment enhances the biological activities that have been associated to CB2 ischemia/reperfusion-induced activation, such as amelioration of microcirculation dysfunction and anti-inflammation [39, 53]. The parallel increase of cortical CB2 relative protein levels may also contribute to the physiological response by regulating the production of pro-inflammatory molecules by glial cells, through which CB2 may either prevent the detrimental effects of neuroinflammatory reaction or participate in adaptive changes to the brain insult [53, 68, 69]. Further, in light of the evidence that eCBs may act as ligands for receptors other than CBs [70], it is interesting that the endocannabinoid and endovanilloid pathways have been found to antagonistically interact to adjust synaptic strength of inhibitory synapses [71].

As for OEA, BCP pre-treatment causes its positive modulation in sham animals suggesting a complex role for OEA in the BCAO/R-induced insult perhaps linked, as suggested above, to the reported BCP-induced activation of PPAR-alpha [32, 55]. In agreement with earlier experimental evidence arguing against the utility of OEA as neuroprotectant prior to ischemic stroke [72], it can be suggested that in our experimental setting the PPAR-alpha receptor is playing a dual role, both anti-inflammatory, likely via repression of NF-kB signalling [73], and anti-oxidative stress by enhancing lipoperoxide degradation [74].

BCP effects on the hypoperfusion/reperfusion-induced oxidative stress

Consistently with our previous findings in absence of any medicinal treatment [3, 4], one of the unfavorable biological effects through which BCCAO/R can affect the tissue homeostasis is the decrease of DHA tissue content, a polyunsaturated fatty acid that is naturally abundant and avidly retained in the brain [75, 76]. Interestingly, the pre-treatment with BCP totally prevented the BCCAO/R-induced decrease of DHA. The disruption of DHA levels is a key factor that affects the efficiency of membrane-depending molecular mechanisms [77]. In fact, DHA is particularly predisposed to lipid peroxidation [76] and, therefore, potentially apt to contribute to the hypoperfusion/reperfusion-induced oxidative stress. The observation that BCP pre-treatment also induces an increase of PPAR-alpha relative levels, an inverse correlation between COX-2 relative levels in BCCAO/R-vehicle- vs BCCAO/R-BCP-pre-treated rats (as shown by the statistically significant BCP-pre-treatment*BCCAO/R interaction), and a decrease of lipoperoxide concentration, leads to suggest that BCP activates multiple and concurrent factors to counteract the BCCAO/R-induced tissue reaction, hence helping to preserve the brain structure. Interestingly, it has been recently reported that the exogenous administration of BCP in combination with DHA has a unique effect in inducing analgesia after formalin injection [78]. Since the synergic activity of BCP and DHA took place in an experiment carried out in vitro, in which the results could not be explained by systemic effects, Fiorenzani et al. [78] proposed that it is likely that both BCP and DHA show competitive binding/interaction for the same receptor. Indeed, there is evidence that DHA activates PPAR-alpha [79, 80], and also inhibits COX-2 and prostaglandins formation during neuroinflammation [81]. In this context, it is relevant that recent prophylactic and therapeutic approaches for cerebrovascular disease take into account the pathways of brain accretion and delivery of DHA [77, 82].

As a matter of fact, in the present study, BCP treatment triggers a marked decrease of tissue and plasmatic concentrations of lipoperoxides. Fatty acid hydroperoxides are quite unstable compounds capable of extending the free radical oxidative damage and form pro-inflammatory substances [83, 84]. Several physiopathological conditions, not necessarily associated to early obvious neurological signs [85], share the occurrence of cerebral hypoperfusion episodes for which the detection of molecular indicators in the early hours may be useful in clinical settings to prevent irreversible cerebral damage. Whether plasmatic changes of AEA and lipoperoxides could represent additional specific markers in humans and if BCP holds promise as an effective nutraceutical compound should be further investigated.