RESEARCH Integrative Cardiovascular Physiology and Pathophysiology Detrimental effects of 2-arachidonoylglycerol on whole blood platelet aggregation and on cerebral blood ﬂow after a focal ischemic insult in rats

Detrimental effects of 2-arachidonoylglycerol on whole blood platelet aggregation and on cerebral blood ﬂow after a focal ischemic insult in rats. Am J Physiol (2-AG) is a major modulator of blood ﬂow and platelet aggregation and a potential neuroprotectant. The present study investigated, for the ﬁrst time, the effects of 2-AG on cerebral blood ﬂow (CBF) in the ﬁrst critical hours during middle cerebral artery occlusion (MCAO) and on platelet aggregation in rats. Adult male Sprague-Dawley rats ( n (cid:2) 30) underwent permanent MCAO under isoﬂurane anesthesia and were randomly assigned to receive either 2-AG (6 mg/kg iv), monoacylglycerol lipase inhibitor JZL-184 (10 mg/kg iv), or vehicle ( n (cid:2) 6 rats/group) treatment. CBF and cardiovascular responses were mea- sured, by a blinded investigator, for up to 4 h. In separate experiments, platelet aggregation by 2-AG (19–300 (cid:3) M) was assessed by whole blood aggregometry ( n (cid:2) 40). 2-AG and JZL-184 signiﬁcantly increased the severity of the CBF deﬁcit versus vehicle (20.2 (cid:4) 8.8% and 22.7 (cid:4) 6.4% vs. 56.4 (cid:4) 12.1% of pre-MCAO baseline, respec-tively, P (cid:5) 0.05) but had no effect on blood pressure or heart rate. While JZL-184 signiﬁcantly increased the number of thrombi after MCAO, this did not reach signiﬁcance by 2-AG. 2-AG induced platelet aggregation in rat whole blood in a similar manner to arachi- donic acid and was signiﬁcantly reduced by the cyclooxygenase inhibitors indomethacin and ﬂurbiprofen and the thromboxane recep- tor antagonist ICI 192,605 ( P (cid:5) 0.05). This is the ﬁrst study showing that 2-AG increases the severity of the CBF deﬁcit during MCAO, and further interrogation conﬁrmed 2-AG-induced platelet aggregation in rats. These ﬁndings are important because 2-AG had previously been shown to exert neuroprotective actions and therefore force us to reevaluate the circumstances under which 2-AG is beneﬁcial. NEW & NOTEWORTHY 2-Arachidonoylglycerol (2-AG) has neuroprotective properties; however, the present study revealed that 2-AG increases the severity of the cerebral blood ﬂow deﬁcit during middle cerebral artery occlusion in rats. Further interrogation showed that 2-AG induces platelet aggregation in rats. These ﬁndings force us to reevaluate the circumstances under which 2-AG is beneﬁcial.


INTRODUCTION
Ischemic stroke is caused by a cerebral blood flow (CBF) deficit due to vessel occlusion by thrombi or emboli and is a major cause of mortality and morbidity in older age. With a paucity of treatments, new information on the role of potential neuroprotective therapies in ischemic stroke is critical.
The endocannabinoid 2-arachidonoylglycerol (2-AG) induces neuronal survival after brain injury such as traumatic brain injury (TBI) (48), spinal cord injury (1), and cerebral ischemic insult (59). 2-AG can act as a full agonist at both cannabinoid (CB)1 and CB2 receptors and can be rapidly metabolized in vivo to arachidonic acid and prostaglandin glycerol ester through enzyme pathways including monoacylglycerol lipase (MAGL) (16) and cyclooxygenase (34). Cyclooxygenase can further metabolize arachidonic acid to PGG 2 and PGH 2 , which in turn may be converted to thromboxane, prostacyclin, and other prostaglandins, all of which have important physiological and pathological actions. 2-AG can be produced by many cells, including endothelial cells and platelets (47). Taken together, endogenous 2-AG carries tremendous implications for disease prevention and progression.
With regard to effects of 2-AG on endothelial cells, 2-AG has potent vasodilatory effects on the rat middle cerebral artery (MCA) (24) and acts as a potent vasorelaxant against endothelin-1 (9), which is a potent vasoconstrictor linked to CBF deficit after ischemic stroke (42). Therefore, 2-AG-induced vasodilatory effects may alleviate the CBF deficit in ischemic stroke, which in turn, at least in part, may be responsible for the neuroprotective effect of 2-AG after MCA occlusion (MCAO) observed by Wang et al. (59). However, no data are available concerning the effects of 2-AG on CBF during MCAO.
Regarding the effects of 2-AG on platelets, 2-AG is the primary endocannabinoid released from activated platelets (41) with potent proaggregatory effects evident in human platelet studies (30,41). Platelet aggregation is known to exacerbate the severity of the CBF deficit after MCAO; however, despite evidence that drugs that modify platelet aggregation influence the outcome in rat models of cerebral ischemia (27), no studies, to our knowledge, have assessed 2-AG effects on rat platelets. Therefore, to establish whether 2-AG alleviates or exacerbates the severity of the CBF deficit during a cerebral ischemic insult, we monitored the effect of 2-AG on CBF in the critical hours after MCAO onset and measured the evolution of pathology. JZL-184 is a potent and selective inhibitor of MAGL and has been shown to increase endogenous 2-AG in the brain at 30 min after administration (38). Therefore, we have addressed our objectives using two strategies, namely, 2-AG treatment to examine exogenous 2-AG effects and JZL-184 treatment to examine endogenous 2-AG effects. We report that exogenous and endogenous 2-AG aggravated rather than alle-viated the CBF deficit with a subsequent increase in the evolution of neuronal injury but with no effects on blood pressure (BP) or heart rate (HR). Interrogation of rat platelet effects confirmed, for the first time, potent proaggregatory effects of 2-AG in rats as a potential explanation for the increased CBF deficit caused by 2-AG.

Ethical Statement
All animal experiments were carried out in strict accordance with the United Kingdom Home Office Guidelines on the Operation of the Animals (Scientific Procedures) Act of 1986 (Project License Nos. PPL 60/3775 and 60/3665; Personal License No. PIL 60/11356) and were performed under ethical approval of University of Strathclyde Animal Welfare Ethical Review Body. All experimental design, data analysis, and statistical procedures were performed in accordance with Animal Research Reporting In Vivo Experiments guidelines for reporting in vivo experiments (31).

Animals
Male adult Sprague-Dawley rats (stroke study: n ϭ 30 and platelet aggregation study: n ϭ 40) were obtained from in-house breeding in the Biological Procedures Unit, University of Strathclyde. Male animals, used to exclude potential estrous cycle effects (13), were group housed under standard conditions at 21°C on a 12:12-h light-dark cycle with environmental enrichment and ad libitum access to water and food. All efforts were made to minimize animal suffering, and the number of animals used, n ϭ 6 rats/group for the MCAO study and n ϭ 5-7 rats/group for the platelet study, was based on a priori power calculations at 80% power and a significance level of 5% (52) and on previous publications (30,43,54). In all experiments, terminal anesthesia was induced and maintained by inhalation of isoflurane. Suitable depth of anesthesia was assessed by lack of a pedal withdrawal reflex or corneal reflex, and this depth of anesthesia was maintained throughout the procedure. Body temperature was monitored continuously with a rectal probe throughout the procedure and maintained within 36.5-37.5°C using a homeothermic blanket (Harvard Apparatus, Kent, Edenbridge, UK).

Permanent Cerebral Ischemia: MCAO Model
Induction of ischemia and measurement of CBF. Anesthesia was induced in 30 male Sprague-Dawley rats (275-350 g, 8 -10 wk old) by inhalation of isoflurane (5% for induction, 2.25-2.5% for maintenance via artificial ventilation) in 30% O 2-70% N2O at 1 l/min. Physiological measurements were monitored wth a pulse oximeter/capnograph (Medair Life-Sense), and a femoral artery was cannulated for continuous BP monitoring and arterial blood gas measurement before and at 2 and 4 h after occlusion. A venous cannula was inserted for drug administration. CBF was monitored throughout the procedure using a laser-Doppler probe (Moor Instruments, Axminster, UK) in the cortical region distal to the MCA at approximately anterior/posterior Ϫ3 mm, medial/lateral 4 mm. Readings were recorded beginning 60 min before the induction of ischemia, at 15-min intervals, and data were expressed as a percentage of the baseline value. Permanent MCAO was induced by a 5-0 silicon-coated monofilament (50-3033, Doccol, Sharon, MA), which was inserted into the left external carotid artery and advanced along the internal carotid artery to occlude the origin of the MCA (39). Sham-operated animals underwent MCAO surgery without insertion of the filament. Animals were maintained for up to 4 h after the onset of MCAO, as neuroprotective effects of 2-AG have been shown to be mediated by a reduction in the early (1-4 h) inflammatory response (48) or early endogenous activation of CB receptors (2). Therefore, the 4 h time point not only allows cardiovascular and cerebrovascular responses to be monitored continuously during MCAO but also allows insights into the effects of 2-AG on the evolution of neuronal injury and the early microglial response.
A priori exclusion criteria were any rat that exhibited no ischemic injury on histological examination, showed signs of rupture of the internal carotid artery, whose laser-Doppler signal was not reduced to Յ60% of baseline following filament insertion, whose physiological measurements were outside normal parameters based on Flecknell (18), or whose oxygen saturation remained below 90%. On the basis of these criteria, no animals were excluded from the present study.
Treatment protocol. Rats were randomly assigned to five groups (n ϭ 6 rats/group). Work was done in six batches. For each batch, the names of the five groups were written on one piece of paper each, crumpled up, and selected out of a beaker by a third party. This process was repeated for each batch of experiments (or n) with a different third party each time. In the treatment groups, animals received either 2-AG (1.2 mg/kg iv bolus over 2 min followed by 4-h infusion to give a total dose of 6 mg/kg), the MAGL inhibitor JZL-184 (5 mg/kg iv bolus over 5 min followed by 4-h infusion to give a total dose of 10 mg/kg), or vehicle (PEG-300 iv) starting 15 min before MCAO. In two groups, rats underwent either MCAO surgery without treatment or sham surgery. The dose of 2-AG used in this study was based on Panikashvili et al. (49) and Wang et al. (59), who both used a single dose of 2-AG at 5 mg/kg (ip bolus) to test effects on neuronal damage after traumatic brain injury and cerebral ischemia, respectively. Given that the total dose in the present study was not administered until the end of the occlusion period, an additional 20% of 2-AG and JZL-184 was used.
Termination and histological processing. The experimenter was blinded to the experimental group to which the animal belonged (treatment received, stroke, or sham surgery) in all analyses by recoding of animals by an independent investigator. At the end of each experiment, animals were euthanized quickly by exsanguination and decapitation, and the brains were immediately and carefully removed and frozen in isopentane (Ϫ42°C) in an attempt to maintain structural integrity and morphology, avoid autolysis, and avoid pressure on the unfixed brain, which might have led to morphological artifacts (29,61). Tissue was stored at Ϫ20°C for histological analysis. For the measurement of injury volume, eight coronal sections (20 m thick) were taken from distinct neuroanatomic regions within the forebrain (45) and stained with hematoxylin and eosin. Hematoxylin and eosinstained sections were viewed under a light microscope to identify damaged tissue based on neuronal morphology. Irreversibly damaged neurons were pyknotic (shrunken or triangular in shape), and the surrounding neuropil was disrupted and displayed pallor. The injury was transcribed onto line diagrams of the eight coronal levels for quantification as previously described (60). The injury boundary was transcribed based on its relation to anatomic landmarks shown on the line diagrams. The use of line diagrams prevented quantification of injury volume from being influenced by tissue edema. Each line diagram was calibrated before use by confirming that the area of each coronal section was measured accurately. The injury area in each line diagram was then quantified using MCID image analysis software. The injury volume was then calculated from the area under the curve measured at each of the eight sections against the interaural (IA) distance, where y intersected x-axes at 12.72 mm IA and 0 mm IA, respectively, and expressed in cubic millimeters.
Immunofluorescence was performed to identify and quantify microglia numbers and activation, based on cell morphology, within the MCA territory at the coronal level of the globus pallidus (2.92 mm IA) and anterior hippocampus (5.88 mm IA). Microglia were labeled using a goat polyclonal antibody for ionized Ca 2ϩ -binding adaptor molecule 1 (Iba-1; 1:250) and FITC-conjugated rabbit anti-goat secondary antibody (1:200). Cell nuclei were stained with Vectashield mounting medium containing DAPI. Thrombi occluding vessels in the ischemic hemisphere were identified by double labeling with rabbit antibody for CD41 (integrin-␣II, 1:250) and mouse antibody for CD62P (P-selectin, 1:100) and appropriate secondary antibodies (FITC-conjugated goat anti-rabbit and TRITC-conjugated goat antimouse antibodies).
Multilayer z-stack images (1 m) were taken at the cortical injury boundary in the ipsilateral hemisphere and corresponding contralateral regions using a laser scanning confocal microscope (Leica Microsystems, Milton Keynes, UK) and analyzed using Volocity imageanalysis software (Perkin-Elmer, Cambridge, UK). Activated microglia cells were clearly identified by a skilled observer on the basis of their morphology, including an enlarged nucleus and shorter, thicker processes. For each image, the number of microglia present was quantified and the number of activated microglia was expressed as a percentage of the total number of microglia present.

Whole Blood Aggregometry
Whole blood aggregometry measures changes in impedance between two electrodes as platelets adhere without the removal of other blood constituents that can affect platelet function (21). Forty naïve adult male Sprague-Dawley rats (300 -500 g, aged 10 -16 wk) were anesthetized with 5% isoflurane in 100% O 2 at 1 l/min and maintained with 2-2.75% isolflurane in O2 delivered via a nose cone. A carotid artery was dissected free of connective tissue and cannulated with polythene tubing containing heparinized saline (10 U/ml). A 9-to 10-ml volume of arterial blood was withdrawn into heparinized syringes (20 U/ml of whole blood), diluted 1:1 with 0.9% (wt/vol) NaCl solution (saline), and stored at room temperature before aggregometry experiments. After blood withdrawal, rats were euthanized with an overdose of 1 ml pentobarbital sodium (Euthatal, 200 mg/ml iv). The diluted blood from each rat was divided into 1-ml aliquots in plastic cuvettes and warmed to 37°C for 5-10 min before any aggregating agents were added. Platelet aggregation was measured using a two-channel Chrono-log 590 whole blood aggregometer (Labmedics, Stockport, UK), which measured changes in impedence across electrodes to which platelets and platelet aggregates adhered (7). After the addition of an aggregatory agent, activated platelets will adhere and impede conductance of the electrical current between the electrodes.
Treatment protocol. After an equilibrium period, platelet aggregation was induced with ADP (0.1-30 M, n ϭ 5-6), 2-AG (n ϭ 6 -13), and arachidonic acid (n ϭ 4 -11) at 19 -300 M, and responses were measured at the peak or after 10 min, respectively. In further experiments, aggregation to ADP ( . Interactions of 2-AG and arachidonic acid with 1 M ADP were studied using concentrations that produced a small aggregation response when added alone (75 M 2-AG and 38 M arachidonic acid), added 1 min before ADP, and the response was measured at 10 min (n ϭ 12). Interactions of 2-AG and arachidonic acid with ADP were also studied in the presence of cyclooxygenase inhibitors indomethacin and flurbiprofen and the TP antagonist ICI-192,605 added 10 min before the first aggregating agent (n ϭ 6).

Data and Statistical Analysis
Values are expressed as means Ϯ SE of n rats. In MCAO experiments, multiple treatment groups were compared using one-way ANOVA with Bonferroni correction or a paired t-test for ipsilateral versus contralateral hemisphere comparisons. Physiological parame-ters and CBF were recorded at baseline and at 1-h intervals throughout the experimental period and compared using mixed-model repeatedmeasures ANOVA. In the whole blood aggregometry experiments, aggregation responses in the presence and absence of antagonists or inhibitors and agonist interactions were compared in aliquots of blood taken from a single animal. Since some responses were not normally distributed, as determined by the Shapiro-Wilk test, aggregation responses in the absence and presence of antagonists or inhibitors were compared using Friedman tests. For agonist interactions, area under the curve values were calculated. A probability of P Ͻ 0.05 was considered to be statistically significant for all comparisons.

2-AG Increased the CBF Deficit During MCAO With No Effects on BP or HR
To assess the effects of endogenous and exogenous 2-AG on local cortical blood flow during MCAO, we monitored CBF by laser-Doppler flowmetry throughout the experiment. A significant increase in the severity of the CBF deficit was observed during MCAO in 2-AG-and JZL-184-treated animals compared with vehicle treatment (Fig. 1). After placement of the microvascular clip on the left internal carotid artery, blood flow dropped to~80% of baseline levels as seen in sham animals [as previously described (43)]. Perfusion was further decreased upon insertion of the filament in vehicle-treated and untreated MCAO groups, to~45% of baseline flow, where it remained stable throughout the occlusion period. However, there was a further significant reduction in CBF to~20% of baseline levels in 2-AG-and JZL-184-treated animals versus vehicle treatment (Fig. 1).
Given that the potential hypotensive effects of 2-AG (for a review, see Ref. 46) might explain the observed decreased CBF, we monitored BP and HR throughout the experiment. However, we found no significant difference in 2-AG-and JZL-184-treated groups compared with vehicle treatment for BP or HR (Fig. 2, A and B). In addition, blood gases and body temperature were tightly controlled, with no significant differences among the groups (Table 1).
To determine the effect of this reduction in CBF on the evolution of pathology, we measured early neuronal perikaryal damage and early microglial activation at the termination of the experiment. Neuronal damage in the cortex was significantly increased by~40% and 50% by 2-AG and JZL-184, respectively, versus vehicle at 4 h after MCAO (Fig. 3, A and C, and  Fig. 4). Increased neuronal damage was not evident in subcortical regions, consistent with the MCA branches supplying the striatum, being end arteries with no collateral flow, unlike cortical branches (19). Microglia activation occurs as early as 3-4 h after stroke and is known as a key initial event in the evolution of damage after stroke (50); however, the efficacy of 2-AG to modulate microglia after stroke has not been studied. Therefore, the microglial response was assessed in the present study in brains at 4 h after MCAO. Whereas there was a significant increase in activation in the ipsilateral hemisphere versus contralateral, there were no significant effects of 2-AG or JZL-184 compared with vehicle control (Fig. 3, B and D). 2-AG has previously been shown to induce platelet aggregation in human blood. As such, we also assessed thrombus formation in the ischemic hemisphere in the present study to investigate the potential role of platelet aggregation in mediating the effect of 2-AG. There was a significant increase in the number of thrombi in the ipsilateral hemisphere after treatment with JZL-184 (Fig. 3, E and F). However, this did not reach significance in 2-AG-treated animals. Collectively, these data show that exogenous and endogenous 2-AG aggravated the deficit in CBF with potential pathological consequences, most likely not mediated via changes in BP or HR.

2-AG Induced Platelet Aggregation Via the Cyclooxygenase Pathway and Thromboxane A 2
To further interrogate mechanisms of 2-AG-induced exacerbation of ischemia, proaggregatory effects of 2-AG, previously shown in human platelets, were studied for the first time in rats. 2-AG induced platelet aggregation over a concentration range of 19 -300 M with a concentration response curve comparable to that of arachidonic acid, which were both less potent than ADP with maximum responses of 12.0 Ϯ 2.1 ⍀ (n ϭ 6) and 11.7 Ϯ 0.5 ⍀ (n ϭ 11) for 2-AG and arachidonic acid, respectively, and 18.7 Ϯ 1.5 ⍀ (n ϭ 6) for ADP (Fig.   5A). 2-AG (150 M) produced slowly developing platelet aggregation with the maximum response occurring at the end of the 10-min observation period (Fig. 5B). A similar aggregation response was observed with arachidonic acid (150 M; Fig. 5B). In contrast, ADP (1 M) produced a rapid and reversible response, which peaked at 2 min (Fig. 5B).
2-AG can be metabolized to arachidonic acid with the further production of proaggregatory mediators. Given the similar profiles of 2-AG-and arachidonic acid-induced aggregation, the role of arachidonic acid metabolism and production of proaggregatory mediators in the response to 2-AG were assessed. The cyclooxygenase inhibitors indomethacin (10 M) and flurbiprofen (10 M) and the TP antagonist ICI-192,605 (1 M) significantly abolished aggregation induced by 2-AG and arachidonic acid (Fig. 6, A and B). 2-AG-induced aggregation was also significantly reduced by indomethacin at 3 M (Fig. 6A); however, this concentration did not affect the response to arachidonic acid (Fig. 6B).
To establish whether MAGL is responsible for the release of arachidonic acid in 2-AG-induced aggregation, we used a selective inhibitor of MAGL (JZL-184) as well as FAAH (URB597). 2-AG-and arachidonic acid-induced aggregation were not altered by URB597 (0.3 M) or JZL-184 (0.1 M) (Fig. 7). In addition, attenuation of the 2-AG response by the TP antagonist ICI-192,605 was not prevented by either URB597 or JZL-184. To confirm the role of MAGL in rat platelets, an additional set of experiments was performed in which a higher concentration of JZL-184 (1 M) was preincubated with blood for a longer period (30 min at 37°C) before the addition of the aggregating agent. In these experiments, JZL-184 (1 M) significantly inhibited the aggregation response to both 2-AG (n ϭ 5) and arachidonic acid (n ϭ 7) (Fig. 7).
2-AG is known to interact with other physiological agonists to enhance aggregation in human platelet-rich plasma (40). We investigated whether 2-AG/ADP-enhanced aggregation involved arachidonic acid metabolism and production of thromboxane A 2 , as previously described by Baldassarri et al. (3), in human platelets. Suboptimal concentrations of 2-AG (75 M) and arachidonic acid (38 M) prolonged and potentiated ADPinduced aggregation, which was significantly reduced by flurbiprofen (Fig. 8, A and B) and partially reduced by indomethacin at 10 M and ICI-192,605 (Fig. 8A). In contrast, indo- methacin at 3 M significantly increased 2-AG potentiation of ADP (n ϭ 6; Fig. 8A).

DISCUSSION
The major finding in the present study is that 2-AG and JZL-184 aggravated the severity of the CBF deficit in the first few pivotal hours during MCAO compared with vehicle, which led to a different evolution of brain injury.
To our knowledge, this is the first study to investigate the effects of 2-AG and JZL-184 on CBF during stroke. The first few hours after permanent MCAO are critical to the development of neuronal damage, as the fate of vulnerable, but potentially viable, tissue surrounding the core is dependent on residual blood flow through collaterals, and the duration and severity of ischemia (23). The laser-Doppler probe was placed above the region perfused by the MCA to measure collateral blood flow in numerous small vessels supplied by the anterior communicating and anterior cerebral arteries during the first 4 h. After intraluminal filament insertion, there was a sharp reduction in CBF, which stabilized at 45% of the baseline signal in the vehicle-treated and untreated groups, indicating successful occlusion of the MCA (28), as confirmed by the presence of neuronal perikaryal damage. There was a further significant reduction in CBF to~20% of baseline levels in 2-AG-treated animals versus vehicle treatment. This result was reproduced by pharmacological enhancement of endogenous 2-AG using JZL-184. In addition, the enhanced CBF deficit resulted in a significant increase in early neuronal damage in 2-AG-and JZL-184-treated groups versus the vehicle-treated group. 2-AG and JZL-184 administration were started 15 min before MCAO, but CBF baseline monitoring was started 60 min before MCAO. Therefore, it can be seen from our data that there was no influence of either agent before (independently of) the induction of cerebral ischemia. Taken together, we are confident that the laser-Doppler flow differences observed in the present study were indeed due to 2-AG-induced aggravation of the CBF deficit during MCAO, probably due to reduced blood flow through collaterals, given the placement of the laser-Doppler probe. Although the present results are perhaps surprising, given that 2-AG is a known vasodilator (24), they provide new information about the role of 2-AG in ischemic stroke in rats and provide some insights into the published results from recent clinical studies that have shown that use of cannabis increased stroke risk (56).
If CBF drops below a certain threshold, tissue damage will be extended (25). To assess whether the drop in CBF by 2-AG and JZL-184 affects the evolution of ischemic injury, we assessed early pathology by identifying pyknotic neurons and activated microglia histologically postmortem at 4 h post-MCAO. 2-AG or JZL-184 did not significantly increase microglial activation as detected by Iba-1 immunostaining. Iba-1 is highly and specifically expressed in monocytic cell lines, including peripheral macrophages as well as both resting and activated microglia (27). However, at this early time point, macrophages are unlikely to have infiltrated the brain parenchyma (51). A significant increase in microglial activation in the ipsilateral hemisphere versus contralateral was observed, which is consistent with the literature at 4 h after MCAO (50), indicating that the lack of effect of 2-AG and JZL-184 on microglial activation is likely to be a valid result and not a type  (36). Therefore, further studies at a later time point postocclusion (Ն24 h) are needed to conclude whether changes in microglial activation across different experimental groups exist. On the other hand, 2-AG and JZL-184 increased the development of neuronal perikaryal damage at 4 h after MCAO in the present study. In the present study, we used fresh-frozen brains for analysis rather than perfused brains to allow examination of intravascular cellular responses (platelets and thrombi). This approach has the potential to generate histological artifacts; therefore, due care was taken to exsanguinate quickly and remove the brains carefully into Ϫ42°C to maintain structural integrity and morphology, avoid autolysis, avoid pressure on the unfixed brain, and avoid dark neuron, vacuole, or pigment artifacts (29,61). Although it is recognized that early ischemic neuronal changes are histologically detectable at this early 4 h time point (58), the infarction is not fully evolved, and the effects of intravenous 2-AG on ischemic damage needs to be further investigated in future studies with a longer (24 -48 h) survival time point to draw conclusions on final infarct size. Wang et al. (59) showed that 2-AG induced neuroprotection 24 h after MCAO, implying that the early increase in injury in the present study was likely to be an acceleration in evolution of damage rather than a definitive exacerbation of infarct. Whether a different dose of Values are means Ϯ SE; n ϭ 6 for all groups for injury, n ϭ 4 -6 for microglia activation, and n ϭ 4 -5 for thrombus formation. For injury volume, *P Ͻ 0.05 vs. the vehicle-treated group; for microglial activation and thrombus formation, *P Ͻ 0.05 vs. the contralateral hemisphere by paired t-test. x P Ͻ 0.05 vs. the sham group by one-way ANOVA followed by multiple comparisons with Bonferroni correction for treatment group comparisons.
2-AG in the present study may be beneficial despite the deleterious effects of platelet aggregation is worthy of note; however, Wang et al. used a single dose similar to that in the present study. Also worthy of note is that Wang et al. administered 2-AG by an intraperitoneal bolus rather than intravenous infusion, which was chosen in the present study to increase the likelihood of exogenous 2-AG reaching the ischemic tissue, given its high labile nature in rat plasma (33) and which may lead to greater platelet effects than an intraperitoneal bolus. The mechanisms behind the aggravation of the CBF deficit during MCAO in the present study are currently unclear. Parameters such as age, body temperature, BP, HR, and arterial blood gases were similar in 2-AG-and JZL-184-treated animals versus the vehicle control and are unlikely to account for the aggravation of the CBF deficit, implying some other underlying cerebrovascular effects of 2-AG. Vasodilatory action on uncompromised vessels is known to cause redistribution of blood away from compromised regions if autoregulation is failing, known as the "steal" phenomenon (20). However, the steal phenomenon after cerebrovasodilators still remains controversial (for a review, see Ref. 35), and whether this explains the increased severity of CBF deficit by 2-AG after stroke needs further investigation. After MCAO, the cerebral vasculature becomes prothrombotic (10,14), and platelet aggregation contributes significantly to MCAO-induced focal cerebral ischemia (11). We showed that, whereas JZL-184 induced a significant increase in the number of thrombi in the ipsilateral hemisphere, this significant increase was not apparent in 2-AG-treated animals. Whether this reflects a slightly more potent or local effect of endogenous 2-AG on thrombi is unclear. 2-AG is a known proaggregant in human platelets (30) but has not yet been studied in rat platelets. Therefore, to establish whether proaggregation is a plausible mechanism of exacerbation of the ischemic deficit in our rat model, we confirmed that 2-AG does indeed induce aggregation of rat platelets. Following this line of thought, increased neuronal perikaryal damage was not evident in subcortical regions, consistent with the MCA branches supplying the striatum, being end arteries with no collateral flow (37,53), which would be expected to be more resistant to the detrimental influence of proaggregatory effects of 2-AG in permanent MCAO models.
In the present study, 2-AG induced aggregation of rat platelets at similar concentrations used in human platelet studies (5,30,40), indicating that any species differences in the stability of 2-AG in plasma between rat and humans (33) is unlikely to have influenced the response to 2-AG in the present study. Washed human platelet studies showed a maximal effect of 2-AG at lower concentrations (10 M) (3,55), probably due to lack of other blood components (blood cells, enzymes, or plasma proteins) that affect the labile nature of 2-AG and/or platelet function (21).
2-AG and arachidonic acid induced similar aggregation responses, which were reduced markedly by cyclooxygenase inhibitors and a TP antagonist, indicating that 2-AG induces aggregation of rat platelets through metabolism of arachidonic acid via the cyclooxygenase pathway and production of the proaggregatory mediator thromboxane A 2 . This mechanism is in agreement with work in human whole blood (5,30) and washed platelets (3), where 2-AG stimulated the synthesis of thromboxane A 2 in human platelets (3). However, an earlier study (40) in platelet-rich plasma did not show any sensitivity of the 2-AG response to cyclooxygenase inhibition. This discrepancy may again be explained by the different methodologies of platelet analysis (22). In the present study, whole blood aggregometry was used to allow platelet function to be evaluated in its physiological milieu, in the presence of red blood cells and leukocytes, which can modulate platelet aggregation (17), more akin to the in vivo microenvironment. Our findings that neither the FAAH inhibitor nor the MAGL inhibitor altered the aggregation response to 2-AG or arachidonic acid are in agreement with some human platelet studies (3,40,55) but not others (5,30). The reasons for this discrepancy are unclear. We used incubation periods (10 min Taken together, our data suggest that 2-AG is unlikely to induce platelet aggregation in rat whole blood through metabolism by FAAH or MAGL to form arachidonic acid. Rather, it is possible that 2-AG may induce aggregation via phospholipase A 2 -mediated release of arachidonic acid or via metabolism of 2-AG to arachidonic acid by other enzymes, such as cyclooxygenase-2, or plasma esterases (33), which are present in whole rat blood. In addition, given the reduction in poten- tiation by ICI-192,605 and flurbiprofen (Fig. 6A), the interaction between 2-AG and ADP in the present study appears also to involve metabolism of arachidonic acid and enhanced secretion of thromboxane A 2 .
In summary, whether 2-AG-induced platelet aggregation is responsible for the increased CBF deficit in our study needs further analysis, given that a significant increase in thrombi was evident only after JZL-184 and not after 2-AG treatment in the present study, and especially given that preventing platelet aggregation had beneficial effect in only some (26) but not all stroke studies (32). However, with millions of people worldwide taking cannabis, and with endocannabinoids recognized as potential therapeutic agents in cerebral and myocardial ischemia (for reviews, see Refs. 4,6,and 14), it is essential to identify the mechanisms responsible for the detrimental effects of 2-AG in stroke uncovered in this study, so that new endocannabinoid-related therapies can be developed that will selectively mediate the beneficial effects but not the detrimental effects in ischemia. Therefore, further, more indepth studies, outside the scope of the present work, are required to determine whether 2-AG, by inducing platelet aggregation, contributes to the formation of thrombi in the brain vessels that reduce the blood flow after stroke and whether the effect of 2-AG on CBF deficit is reversed by treatment with cyclooxygenase inhibitors or thromboxane receptor antagonists. The present findings force us to reevaluate the circumstances under which 2-AG is beneficial and whether systemic 2-AG treatment may represent an increased risk of thromboembolic-related disorders.