Selective inhibition of 2-AG hydrolysis enhances endocannabinoid signaling in hippocampus

The functions of 2-arachidonoylglycerol (2-AG), the most abundant endocannabinoid found in the brain, remain largely unknown. Here we show that two previously unknown inhibitors of monoacylglycerol lipase, a presynaptic enzyme that hydrolyzes 2-AG, increase 2-AG levels and enhance retrograde signaling from pyramidal neurons to GABAergic terminals in the hippocampus. These results establish a role for 2-AG in synaptic plasticity and point to monoacylglycerol lipase as a possible drug target.

The endocannabinoids modulate brain neurotransmission by activating CB 1 cannabinoid receptors mostly localized on axon terminals of GABAergic interneurons 1 .In the hippocampus, an endocannabinoid released from depolarized pyramidal cells transiently depresses GABA release from basket cell terminals in a retrograde signaling process called depolarization-induced suppression of inhibition (DSI) [2][3][4] .Although the role of CB 1 receptors in DSI is well documented, the identity of its endocannabinoid mediator remains elusive.Two molecules, anandamide and 2-AG, meet the defining criteria of an endocannabinoid.They are produced by neurons in an activitydependent manner, they engage CB 1 receptors with high affinity and they are eliminated through regulated transport and intracellular hydrolysis 5 .In neurons, the hydrolysis of anandamide and 2-AG is catalyzed by two distinct serine hydrolases: fatty-acid amide hydrolase (FAAH), which cleaves anandamide and other lipid amides 6 , and monoacylglycerol lipase (MGL), which hydrolyzes 2-AG and other 2-monoacylglycerols 7 .This catabolic segregation offers the opportunity to investigate the functions of each endocannabinoid by blocking its deactivation and thereby amplifying its actions at CB 1 receptors.Using this approach, it has been shown that inhibition of FAAH activity does not affect hippocampal DSI 8 , which suggests that neither anandamide nor other FAAH substrates with cannabinoid-like activity (for example, virodhamine and N-arachidonoyl-dopamine) 9,10 contribute to this process.
of CA1 pyramidal neurons elicited DSI in B50% of the cells tested.
The functions of 2-AG in synaptic plasticity have been previously inferred from experiments with non-specific biosynthesis inhibitors 5 or Cox-2 inhibitors 8 , whose effects on brain endocannabinoid metabolism are unknown.Our results, showing that blockade of intracellular MGL activity selectively increases brain 2-AG levels and prolongs hippocampal DSI, provide unambiguous evidence that 2-AG mediates this form of retrograde signaling.Potent and selective MGL inhibitors designed on the scaffolds of URB602 and URB754 will provide a valuable tool to explore the physiological roles of 2-AG and validate MGL as a drug target.
Note: Supplementary information is available on the Nature Neuroscience website.

ACKNOWLEDGMENTS
We thank N. Ha ´jos for valuable comments and J. Kim for help with cultures.This research was supported by grants from the US National Institute on Drug

Figure 1
Figure 1 Effects of MAFP on DSI in hippocampal CA1 pyramidal cells.Top, traces from a representative experiment showing the effects of vehicle (ethanol, 0.00003%) or MAFP (Tocris, 45 nM) on the transient reduction of spontaneous inhibitory postsynaptic potentials (IPSCs) elicited by a depolarizing stimulus (arrow).Scale bars, 100 pA, 5 s.Bottom left, averaged time-course of DSI after administration of vehicle (solid squares) or MAFP (open squares).Bottom right, DSI area in the first 30 s after stimulus application was significantly larger in MAFP-treated than in control slices.

Figure 2 VOLUME 8 [
Figure 2 Inhibition of MGL, but not FAAH, activity prolongs DSI in CA1 pyramidal cells.(a-c) Time course of DSI measured in the same cell before (filled squares) and after (open squares) treatment with (a) URB597 (0.1 mM; n ¼ 7); (b) URB602 (100 mM; n ¼ 8) or (c) URB754 (0.5 mM; n ¼ 7).Panels on the right show representative DSI traces from individual experiments before and after drug application.Arrows indicate the depolarizing stimuli used to elicit DSI (from -60 mV to 0 mV, 1 s).Scale bars, 100 pA, 10 s.(d) Application of URB754 (0.5 mM) and URB602 (100 mM), but not URB597 (0.1 mM), led to greater DSI area than in control (Ctr).(e) The drugs did not influence basal charge transfer.*, P o 0.05, Student's t-test.
All experimental procedures were in accordance with Society for Neuroscience and European Union guidelines and were approved by the institutional animal care and use committees.