Acute inhibition of diacylglycerol lipase blocks endocannabinoid-mediated retrograde signalling: evidence for on-demand biosynthesis of 2-arachidonoylglycerol

Abstract

The endocannabinoid (eCB) 2-arachidonoylglycerol (2-AG) produced by diacylglycerol lipase α (DGLα) is one of the best-characterized retrograde messengers at central synapses. It has been thought that 2-AG is produced ‘on demand’ upon activation of postsynaptic neurons. However, recent studies propose that 2-AG is pre-synthesized by DGLα and stored in neurons, and that 2-AG is released from such ‘pre-formed pools’ without the participation of DGLα. To address whether the 2-AG source for retrograde signalling is the on-demand biosynthesis by DGLα or the mobilization from pre-formed pools, we examined the effects of acute pharmacological inhibition of DGL by a novel potent DGL inhibitor, OMDM-188, on retrograde eCB signalling triggered by Ca²⁺ elevation, G_q/11 protein-coupled receptor activation or synergy of these two stimuli in postsynaptic neurons. We found that pretreatment for 1 h with OMDM-188 effectively blocked depolarization-induced suppression of inhibition (DSI), a purely Ca²⁺-dependent form of eCB signalling, in slices from the hippocampus, striatum and cerebellum. We also found that at parallel fibre–Purkinje cell synapses in the cerebellum OMDM-188 abolished synaptically induced retrograde eCB signalling, which is known to be caused by the synergy of postsynaptic Ca²⁺ elevation and group I metabotropic glutamate receptor (I-mGluR) activation. Moreover, brief OMDM-188 treatments for several minutes were sufficient to suppress both DSI and the I-mGluR-induced retrograde eCB signalling in cultured hippocampal neurons. These results are consistent with the hypothesis that 2-AG for synaptic retrograde signalling is supplied as a result of on-demand biosynthesis by DGLα rather than mobilization from presumptive pre-formed pools.

Key points

• 2-Arachidonoylglycerol (2-AG), one of the best-characterized retrograde messengers at central synapses, has been thought to be produced ‘on demand’ through a diacylglycerol lipase α (DGLα)-dependent pathway upon activation of postsynaptic neurons (on-demand synthesis hypothesis).

• However, recent studies propose an alternative hypothesis that 2-AG is pre-synthesized by DGLα, stored in neurons, and released from such ‘pre-formed pools’ without the participation of DGLα (pre-formed pool hypothesis).

• To test these hypotheses, we examined the effects of acute pharmacological inhibition of DGL by a novel potent DGL inhibitor, OMDM-188, on retrograde 2-AG signalling.

• We found that 2-AG-mediated retrograde signalling was blocked after 1 h treatment with OMDM-188 in acute slices from the hippocampus, striatum and cerebellum, and was blocked several minutes after OMDM-188 application in cultured hippocampal neurons.

• These results fit well with the on-demand synthesis hypothesis, rather than the pre-formed pool hypothesis.

Introduction

Endocannabinoids (eCBs) are released from postsynaptic neurons and negatively regulate synaptic transmission through presynaptic cannabinoid type 1 (CB₁) receptors (Kano et al. 2009; Regehr et al. 2009; Castillo et al. 2012; Katona & Freund, 2012; Ohno-Shosaku et al. 2012). While anandamide and 2-arachidonoylglycerol (2-AG) have been identified as two major eCBs (Piomelli, 2003), recent studies have revealed that 2-AG but not anandamide mediates retrograde signalling at synapses (Gao et al. 2010; Tanimura et al. 2010; Yoshino et al. 2011). This conclusion is based on results from mice deficient in the 2-AG-synthesizing enzyme diacylglycerol lipase α (DGLα) and β (DGLβ). The mobilization of eCB from postsynaptic neurons is triggered by strong depolarization of postsynaptic neurons and resultant elevation of intracellular Ca²⁺ concentration (Ca²⁺-driven eCB release (ER); Kreitzer & Regehr, 2001; Ohno-Shosaku et al. 2001; Wilson & Nicoll, 2001), strong activation of postsynaptic G_q/11 protein-coupled receptors at basal Ca²⁺ level (basal receptor-driven eCB release (RER); Maejima et al. 2001; Varma et al. 2001), or simultaneous Ca²⁺ elevation and G_q/11 protein-coupled receptor activation (Ca²⁺-assisted RER; Varma et al. 2001; Kim et al. 2002; Ohno-Shosaku et al. 2002). In DGLα knockout mice but not in DGLβ knockout mice, all of the three forms of eCB-mediated retrograde signalling were absent (Tanimura et al. 2010), indicating that the 2-AG produced by DGLα mediates retrograde signalling.

As for the production of eCBs, it has long been thought that eCB is produced ‘on demand’ in activated neurons (Piomelli, 2003). Two recent studies challenged this ‘dogma’ of eCB production. A novel potent DGL inhibitor, OMDM-188 (Ortar et al. 2008; Di Marzo, 2011), does not block either depolarization-induced suppression of inhibition (DSI), a representative form of Ca²⁺-driven ER (Min et al. 2010b), or RER by activation of group I metabotropic glutamate receptors (I-mGluRs; Zhang et al. 2011). To reconcile the discrepancy between the results from DGLα knockout mice and OMDM-188, the authors proposed that 2-AG is pre-synthesized by DGLα and pooled in neurons, and is mobilized from these hypothetical ‘pre-formed 2-AG pools’ upon stimulation without the contribution of DGLα (Min et al. 2010a; Alger & Kim, 2011; Fig. 6A, ‘pre-formed’). However, this hypothesis depends largely on the results from pharmacological experiments which are inconsistent. Min et al. showed that 2 μm OMDM-188 did not block DSI in hippocampal slices (Min et al. 2010b). Contrary to this report, Zhang et al. showed that 5 μm OMDM-188 blocked DSI but did not suppress RER by activation of I-mGluR in hippocampal slices (Zhang et al. 2011). The two differing effects of the same DGL inhibitor in the same preparation are not explained by the pre-formed 2-AG hypothesis. Furthermore the absence of RER suppression by OMDM-188 is also at variance with the previous studies that a broad spectrum DGL inhibitor, tetrahydrolipstatin (THL) or RHC-80287, blocked RER and Ca²⁺-assisted RER (Melis et al. 2004; Haj-Dahmane & Shen, 2005; Safo & Regehr, 2005; Hashimotodani et al. 2007b; Uchigashima et al. 2007; but see Edwards et al. 2006, 2008). Therefore, it is necessary to systematically evaluate the effects of OMDM-188 on Ca²⁺-driven ER, RER and Ca²⁺-assisted RER, and to determine which of the three forms of eCB release require the immediate activity of DGLα for 2-AG mobilization.

Figure 6. Summary diagrams of the models for 2-AG-mediated retrograde signalling.

A, schematic diagram illustrating two models that explain how 2-AG is released upon activation of postsynaptic neurons. In the on-demand biosynthesis model (‘on demand’), postsynaptic activation induces production of diacylglycerol (DG). DG is then converted to 2-arachidonoylglycerol (2-AG) by DG lipase α (DGLα) and 2-AG is released as retrograde messenger. In this model, acute pharmacological inhibition of DGLα is expected to block the retrograde signalling. In the pre-formed pool model (‘pre-formed’), 2-AG is synthesized constitutively by DGLα and stored in neurons. Postsynaptic activation triggers 2-AG release from the 2-AG pool to initiate retrograde signalling. In this model, the retrograde signalling is expected to remain intact under pharmacological blockade of DGLα until the presumptive 2-AG pool is depleted. B, molecular mechanisms of 2-AG-mediated retrograde signalling based on the on-demand synthesis model. Activation of G_q/11 protein-coupled receptors (G_q/11PCRs) stimulates PLCβ, which produces the 2-AG precursor DG. Activation of voltage-gated Ca²⁺ channel (VGCC) causes Ca²⁺ elevation, which induces DG production through an unknown pathway and also facilitates the G_q/11PCR–PLCβ signalling pathway. 2-AG is then produced from DG by DGLα and released. The released 2-AG activates presynaptic CB₁ receptors, and suppresses the transmitter release.

Methods

All experiments were performed according to the guidelines laid down by the animal welfare committees of the University of Tokyo and Kanazawa University.

Recordings from acute slices

Transverse hippocampal slices (400 μm thick), coronal striatum slices (300 μm thick) and parasagittal cerebellar slices (250 μm thick) were prepared from C57BL/6 mice at postnatal days 20–33, 17–20 and 9–27, respectively. Mice were decapitated under CO₂ anaesthesia, and the brains were cooled in ice-cold cutting solution with the following compositions (in mm): 120 choline chloride, 2 KCl, 8 MgCl₂, 28 NaHCO₃, 1.25 NaH₂PO₄, and 20 glucose, bubbled with 95% O₂ and 5% CO₂, for the hippocampus and striatum; or normal artificial cerebrospinal fluid (ACSF) (in mm): 125 NaCl, 2.5 KCl, 2 CaCl₂, 1 MgSO₄, 1.25 NaH₂PO₄, 26 NaHCO₃, and 20 glucose, bubbled with 95% O₂ and 5% CO₂, for the cerebellum. Slices were cut with a vibroslicer (Leica VT-1200) and incubated at room temperature for at least 1 h before experiments.

The recording chamber was perfused with ACSF supplemented with 10 μm 2,3-dioxo-6-nitro-1,2,3,4-tet-rahydrobenzo[f]quinoxaline-7-sulfonamide and 10 μm 3-((R)-2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid for recording IPSCs, or with 100 μm picrotoxin for recording EPSCs at a flow rate of 1.5–2.0 ml min⁻¹. For recording IPSCs from hippocampus, concentrations of Ca²⁺ and Mg²⁺ were increased to 2.5 mm and 1.3 mm, respectively. In all experiments, we used an upright microscope (Axioskop; Zeiss, Germany) equipped with an infrared CCD camera system (Hamamatsu Photonics, Hamamatsu, Japan).

For the DSI experiments in the hippocampus, whole-cell recordings were made from visually identified pyramidal neurons in the CA1 region. IPSCs were evoked by monopolar stimulation with a glass pipette filled with the bath solution and placed at the inner one-third of the stratum radiatum. Patch pipettes (3–5 MΩ) were filled with an internal solution with the following composition (in mm): 90 CsCH₃SO₃, 45 CsCl, 10 Hepes, 0.2 EGTA, 5 MgCl₂, 5 Na₂ATP, 0.2 Na₂GTP and 5 QX314 (pH 7.3).

For the DSI experiments in the striatum, whole-cell recordings were performed from visually identified medium spiny (MS) neurons (Narushima et al. 2007) and IPSCs were evoked by bipolar stimulation with a pair of glass pipettes filled with the bath solution and placed near the neurons being recorded. Patch pipettes (3–5 MΩ) were filled with the same internal solution as used for the experiments in the hippocampus.

For the experiments of DSI and depolarization-induced suppression of excitation (DSE) in the cerebellum, whole-cell recordings were made from visually identified Purkinje cells (PCs), and inhibitory inputs and excitatory parallel fibre (PF) inputs were stimulated by a glass pipette filled with the bath solution and placed above the PC layer and in the middle molecular layer, respectively. Patch pipettes (2–4 MΩ) were filled with the same internal solution as used in hippocampal slices for DSI experiments or with solutions of the following compositions (in mm): 140 CsCl, 1 EGTA, 10 Hepes, 4.6 MgCl₂, 0.1 CaCl₂, 4 Na₂ATP, 0.4 Na₂GTP (pH 7.3) for DSE experiments; 130 potassium d-gluconate, 6 KCl, 10 NaCl, 10 Hepes, 0.16 CaCl₂, 2 MgCl₂, 0.5 EGTA, 4 Na-ATP and 0.4 Na-GTP (pH 7.3) for experiments of synaptically induced retrograde suppression.

In all experiments, membrane currents were recorded with a patch-clamp amplifier (Axopatch 1D, Molecular Devices, Sunnyvale, CA, USA). The signals were filtered at 3 kHz and digitized at 20 kHz and the pipette access resistance was compensated by 80%. All experiments were performed at 32°C, except for those using hippocampus which were performed at room temperature.

Recordings from cultured hippocampal neurons

Cultured hippocampal neurons were prepared from newborn Sprague–Dawley rats as described previously (Ohno-Shosaku et al. 2001). Briefly, rats were decapitated under isoflurane anaesthesia, and brains were rapidly isolated. Then, cells were mechanically dissociated from the hippocampi and plated onto poly-l-ornithine-coated plastic dishes. The cultures were kept at 36°C in 5% CO₂ for 12–15 days before use.

Double whole-cell recordings were performed using a patch-clamp amplifier (EPC10/2; HEKA Electronik, Lambrecht/Pfalz, Germany). Each neuron of a pair was voltage-clamped at −80 mV using a patch pipette (3–5 MΩ) filled with an internal solution of the following composition (in mm): 130 potassium gluconate, 15 KCl, 10 Hepes, 0.2 EGTA, 6 MgCl₂, 5 Na₂ATP and 0.2 Na₂GTP (pH 7.3). One neuron was stimulated by applying positive voltage pulses (to 0 mV, 2 ms) at 0.5 Hz, and IPSCs were measured from the other neuron. For the experiments with the I-mGluR agonist dihydroxyphenylglycine (DHPG), an internal solution with the following composition was used for postsynaptic neurons (in mm): 120 potassium gluconate, 15 KCl, 10 Hepes, 5 EGTA, 6 MgCl₂, 5 Na₂ATP and 0.2 Na₂GTP (pH 7.3). The external solution contained (in mm): 140 NaCl, 2.5 KCl, 2 CaCl₂, 1 MgCl₂, 10 Hepes, 10 glucose and 1 kynurenic acid (pH 7.3). The recording chamber was perfused with the external solution with or without drugs at a flow rate of 1–3 ml min⁻¹.

Application of drugs

A potent DGL inhibitor, OMDM-188, was dissolved in DMSO to yield a 10 mm stock solution and stored at −30°C. OMDM-188 was diluted in the bath solution to the final concentrations before use. To effectively deliver lipophilic compounds such as OMDM-188 to neurons in brain slices, we added 0.2 mg ml⁻¹ bovine serum albumin (BSA) into the ACSF (Makriyannis et al. 2005; Tanimura et al. 2009) during the treatment of slices with OMDM-188. Slices were incubated with BSA (control) or with OMDM-188 plus BSA for 1 h. Recordings were performed without OMDM-188 or BSA and each experiment was terminated within 40 min after transferring slices to the recording chamber.

Quantification of 2-AG

Hippocampal slices were prepared and incubated with OMDM-188 (2 μm) or vehicle for 1.5–2.0 h. Slices were frozen quickly in liquid nitrogen, and stored at −80°C. Quantification of 2-AG was performed as described previously (Tanimura et al. 2010). The samples were crushed to powder without thawing using AutoMill (Tokken, Chiba, Japan) and lipids were extracted with 1 ml of acetonitrile containing deuterated 2-AG (2-AG-d8; Cayman, Ann Arbor, MI, USA) as an internal standard. The extracts were further cleaned up and concentrated using Oasis HLB 1 cc Vac solid phase extraction cartridge (10 mg sorbent, Waters, Milford, MA, USA) as follows: samples were diluted 3-fold with water containing 0.03% formic acid and loaded to the preconditioned Oasis HLB cartridge. The cartridges were serially washed with water and 15% ethanol, and then eluted with 150 μl of acetonitrile, 10 μl of which was analysed with a liquid chromatography-tandem mass spectrometry system.

An ultra-performance liquid chromatograph (UPLC, Waters) coupled to a TSQ Quantum Ultra triple-stage quadrupole mass spectrometer (Thermo Scientific, Waltham, MA, USA) was used. The following UPLC conditions were used: column, ACQUITY UPLC BEH C18 (1.0 mm × 100 mm); flow rate, 100 μl min⁻¹; mobile phase, A: 20 mm NH₄HCO₃ in water, B: acetonitrile; gradient program, 20–95% linear gradient of mobile phase B (0–20 min) and hold at 95% of B (20–27 min); column temperature, 45°C. With this condition, 2-AG was baseline resolved from its structural isomer 1-AG. The TSQ Quantum Ultra mass spectrometer with electrospray ionization ion source was operated in selected-reaction monitoring (SRM) mode under the following conditions: spray voltage, +3.5 kV, ion transfer tube temperature, 350°C; sheath gas, 65 (arbitrary units); auxiliary gas, 5 (arbitrary units); collision gas (Ar, 99.999%), 1.2 mTorr; Q1 and Q3 resolutions, both at 0.7 Da (full width at half-maximum); SRM transitions, m/z 379→287 (2-AG) and m/z 387→294 (2-AG-d8). Quantification was performed with an internal standard method using known amounts of 2-AG as the primary standard and 2-AG-d8 as an internal standard. Data were analysed with Xcalibur 2.0.6 software package (Thermo Scientific).

Data analysis and statistics

For cultured hippocampal neurons, magnitudes of DSI, 2-AG-induced suppression and I-mGluR-mediated suppression were measured as a percentage of the mean amplitudes of five consecutive IPSCs after depolarization or drug application relative to those acquired for 30 s before conditioning. For brain slices, magnitudes of DSI for the hippocampus, striatum and cerebellum and those of DSE and synaptically induced retrograde suppression for the cerebellum were calculated as the percentage of reduction in the mean of the three or four (for DSI in the cerebellum) consecutive IPSC/EPSC amplitudes following depolarization or train of PF stimulation relative to the baseline IPSC/EPSC amplitudes before conditioning.

Results are presented as means ± SEM. Normality of the distribution was tested by the Kolmogorov–Smirnov (KS) test. Statistical significance was assessed by paired or unpaired Student's t test. When sample data did not pass the KS test, the non-parametric Mann–Whitney U test was used. One-way ANOVA and the post hoc Bonferroni test was used for comparison between more than two groups.

Results

OMDM-188 blocks DSI in hippocampal slices

Min et al. reported recently that DSI in hippocampal slices was not blocked by OMDM-188 (Min et al. 2010b). In their study, hippocampal slices prepared from C57BL/6 mice were incubated with 2 μm OMDM-188 for at least 30 min, and IPSC recordings were performed in the presence of OMDM-188. Using the same concentration, we first re-examined the effect of OMDM-188 on DSI in hippocampal slices of C57BL/6 mice. Hippocampal slices were pretreated with 2 μm OMDM-188 for 1 h, and recordings were performed in the absence of OMDM-188. After stable recording of basal IPSCs at 0.33 Hz, DSI was induced by depolarizing pyramidal neurons from −70 mV to 0 mV for 5 s. In contrast to the previous report (Min et al. 2010b), 2 μm OMDM-188 blocked DSI completely (Fig. 1A–C). Then, we measured the 2-AG content in hippocampal slices to check whether OMDM-188 depleted basal 2-AG contents in hippocampal slices. In these experiments, we used a longer incubation time (1.5–2.0 h) to reduce the risk of insufficient OMDM-188 penetration into cells in the depths of slices. The 2-AG content of the OMDM-188-treated slices was reduced to about 70% of the control slices (Fig. 1D). These results indicate that pharmacological inhibition of DGL by OMDM-188 blocks DSI in hippocampal slices without depleting basal 2-AG substantially. A rational explanation for the reduction in 2-AG content is provided in the Discussion.

Figure 1. Effects of OMDM-188 on DSI and basal 2-AG contents in acute hippocampal slices.

A–C, representative experiments (A and B) and summary bar graph (C) showing the blockade of DSI by treatment of hippocampal slices with 2 μm OMDM-188 for 1 h. Time courses of DSI in control (A) and OMDM-188-treated slices (B) are shown. Pyramidal neurons were depolarized (from −70 mV to 0 mV for 5 s) at time 0. Insets: IPSC traces before (black) and after depolarization (red). D, summary bar graph showing basal 2-AG contents in hippocampal slices. Numbers of tested cells are indicated in parentheses in C and D, and also in similar bar graphs shown in subsequent figures. *P < 0.05, ***P < 0.001, t test

OMDM-188 blocks eCB-mediated retrograde synaptic suppression in the striatum and cerebellum

Genetic deletion of DGLα abolishes eCB signalling in the cerebellum, hippocampus and striatum (Tanimura et al. 2010). To address whether acute pharmacological inhibition of DGL by OMDM-188 blocks eCB signalling in brain regions other than the hippocampus, we examined DSI in striatal and cerebellar slices after pretreatment with OMDM-188 for 1 h. After stable recording of basal IPSCs (at 0.2 Hz for MS neurons and at 0.33 Hz for PCs), DSI was elicited by single depolarizing pulses (from −70 mV to 0 mV for 5 s) to MS neurons or by applying a series of depolarizing pulses (10 pulses of 100 ms duration from −70 mV to 0 mV repeated at 1 Hz) to PCs. Notably, DSI was completely blocked by 2 μm OMDM-188 in both striatal and cerebellar slices (Fig. 2A–D). We also examined the effect of OMDM-188 on DSE at PF–PC synapses in cerebellar slices. After stable recording of basal EPSCs at 0.33 Hz, DSE was induced by single depolarizing pulses (from −70 mV to 0 mV for 1 s) to PCs. DSE at PF–PC synapses was markedly reduced by OMDM-188 at 2 μm, and was completely blocked at 5 μm (Fig. 2E and F). Furthermore, we examined the effect of OMDM-188 on retrograde suppression of PF-EPCSs induced by PF burst stimulation (Brown et al. 2003; Maejima et al. 2005), which is known to be caused by Ca²⁺-assisted RER (Maejima et al. 2005). After stable recording of basal EPSCs at 0.2 Hz, PFs were stimulated by 10 pulses at 100 Hz with an electrode placed at the edge of the molecular layer. We found that 2 μm OMDM-188 completely blocked the synaptically induced retrograde suppression of PF EPSCs (Fig. 3). Taken together, these results clearly demonstrate that acute inhibition of DGL by OMDM-188 effectively blocks eCB-mediated synaptic suppression in the striatum and cerebellum as well as in the hippocampus.

Figure 2. Effects of OMDM-188 on eCB signalling in acute slices from the striatum and cerebellum.

A and B, representative data (A) and summary bar graph (B) showing the blockade of DSI by OMDM-188 (2 μm) in striatal MS neurons. Time courses of DSI in control (open circles) and OMDM-188-treated slices (filled circles) are shown. MS neurons were depolarized (from −70 mV to 0 mV for 5 s) at time 0. C and D, representative data (C) and summary bar graph (D) showing the blockade of DSI by OMDM-188 (2 μm) in cerebellar PCs. Time courses of DSI in control (open circles) and OMDM-188-treated slices (filled circles) are shown. A series of depolarizing pulses (10 pulses of 100 ms duration from −70 mV to 0 mV repeated at 1 Hz) were applied to PCs at time 0. Insets in A and C: IPSC traces before (black) and after depolarization (red). E, representative data showing the blockade of DSE by OMDM-188 (5 μm) at PF–PC synapses in cerebellar slices. Time courses of DSE in control (open circles) and OMDM-188-treated slices (filled circles) are shown. PCs were depolarized (from −70 mV to 0 mV for 1 s) at time 0. Insets: EPSC traces before (black) and after depolarization (red). F, summary bar graph showing dose-dependent blockade of DSE by OMDM-188. **P < 0.01, ***P < 0.001, t test.

Figure 3. Effects of OMDM-188 on synaptically induced eCB signalling in acute cerebellar slices.

Representative experiments (A and B) and summary bar graph (C) showing the blockade of transient suppression induced by PF burst stimulation by OMDM-188 (2 μm). PFs were stimulated by 10 pulses at 100 Hz at time 0. In B, EPSC traces before (black) and after PF burst stimulation (red) are shown. ***P < 0.001, one-way ANOVA with Bonferroni post hoc test.

OMDM-188 blocks eCB-mediated retrograde synaptic suppression in cultured hippocampal neurons

Then, we examined the effect of OMDM-188 on retrograde eCB signalling in cultured hippocampal neurons from rat. This reduced preparation provides agonists and blockers of eCB signalling with much better access to target cells compared to acute slice preparations (Hashimotodani et al. 2007b, 2008). The improved accessibility is expected to allow rapid inactivation of DGL by OMDM-188. We performed paired whole-cell recordings from hippocampal neurons and isolated cannabinoid-sensitive unitary IPSCs. We determined the cannabinoid sensitivity of IPSCs by checking their responses to 2-AG (0.1 μm) (Hashimotodani et al. 2007b). We adopted neuron pairs that exhibited more than 50% suppression of IPSCs by 2-AG for the following experiments.

We tested the effect of OMDM-188 at 0.2 μm, a concentration 10 times lower than that used for acute slice experiments. While postsynaptic depolarization for 5 s elicited prominent DSI in control neurons, pretreatment with 0.2 μm OMDM-188 for 20–60 min completely blocked DSI (Fig. 4A–C). Then we examined the effect of OMDM-188 on RER by I-mGluR activation. We found that the same OMDM-188 pretreatment significantly blocked the IPSC suppression induced by the I-mGluR agonist DHPG (Ohno-Shosaku et al. 2002; Fig. 4D and E). We confirmed that presynaptic CB₁ receptor function was unchanged by OMDM-188, since the amplitude of 2-AG-induced suppression of IPSCs was 127 ± 21% of control (n= 3) after OMDM-188 treatment.

Figure 4. Effects of pretreatment with OMDM-188 on eCB signalling in cultured hippocampal neurons.

A–C, representative experiments (A and B) and summary bar graph (C) showing the blockade of DSI by treatment with 0.2 μm OMDM-188 for 20–60 min. Open and filled circles represent data from experiments in control and OMDM-188-containing solutions, respectively. Postsynaptic neurons were depolarized (from −80 mV to 0 mV for 5 s) at time 0. IPSC traces (B) before (black) and after depolarization (red) are shown. D and E, representative experiments (D) and summary bar graph (E) showing the blockade of DHPG (50 μm, 1 min)-induced IPSC suppression by treatment with 0.2 μm OMDM-188 for 20–60 min. Time courses of DHPG-induced IPSC suppression in control (open circles) and OMDM-188-containing (filled circles) solutions are shown. Insets are IPSC traces before (black) and during application of DHPG (red). ***P < 0.001, Mann–Whitney U test (C) or t test (E).

Next, we tested whether brief application of OMDM-188 to cultured hippocampal neurons blocks DSI. After observation of DSI in the control solution, 2 μm OMDM-188 was bath applied. After 2 min of OMDM-188 application, the same postsynaptic depolarization induced significantly smaller DSI (Fig. 5A and B). We induced DSI repeatedly with 2 min intervals and found that DSI was blocked almost completely after 8 min of OMDM-188 application (Fig. 5A and B). To test whether repeated DSI induction is required for almost complete blockade of DSI, we performed another series of experiments in which DSI was induced only twice, before and 8 min after OMDM-188 application. We found that the magnitude of DSI was reduced from 87.3 ± 4.9% to 11.5 ± 1.8% after 8 min treatment with OMDM-188 (n= 6, P < 0.001, paired t test), indicating that the OMDM-induced blockade of DGL is not activity dependent. The same OMDM-188 treatment was also tested for RER. In each neuron pair, DHPG was applied twice. Without OMDM-188 treatment, repeated application of DHPG induced similar extents of IPSC suppression (Fig. 5C, upper panel; Fig. 5D, Control). When DHPG was applied before and after 8 min treatment with 2 μm OMDM-188, the IPSC suppression induced by the second DHPG application was markedly reduced (Fig. 5C, lower panel; Fig. 5D, OMDM). These results demonstrate that acute pharmacological inhibition of DGL for several minutes is sufficient for suppressing retrograde eCB signalling.

Figure 5. Effects of brief treatment with OMDM-188 on eCB signalling in cultured hippocampal neurons.

A and B, in each pair, the postsynaptic neuron was depolarized (from −80 mV to 0 mV for 5 s) before and 2 min, 4 min and 8 min after application of 2 μm OMDM-188. Representative IPSC traces obtained before (black) and after (red) depolarization (A) and summary graph (B) show that OMDM-188 treatment for 2 min effectively blocked DSI. C and D, representative experiments (C) and summary graph (D) showing the blockade of DHPG (50 μm, 1 min)-induced IPSC suppression by 8 min treatment with 2 μm OMDM-188. IPSC traces acquired before (black) and during DHPG application (red) are shown in inset. *P < 0.05, **P < 0.01, one-way ANOVA with Bonferroni post hoc test (B) or paired t test (D).

Discussion

In the present study, we used a potent DGL inhibitor, OMDM-188, to test whether 2-AG is produced on demand or is mobilized from pre-formed pools upon activation of postsynaptic neurons (Fig. 6A). We demonstrated that OMDM-188 effectively blocked DSI in hippocampal and striatal slices, DSI, DSE and Ca²⁺-assisted RER in cerebellar slices, and DSI and RER in cultured hippocampal neurons. These results support the idea that on-demand 2-AG biosynthesis is required for retrograde eCB signalling (Fig. 6B), although we cannot entirely rule out the hypothesis that 2-AG is mobilized from pre-formed 2-AG pools. We assume it very unlikely, but a possibility remains, that readily depleting 2-AG pools may act as 2-AG sources. Even if such 2-AG pools may exist, our data obtained from cultured hippocampal neurons clearly indicate that they must be very labile and easily depleted within 2 min of DGL blockade even in silent neurons that are voltage-clamped to −80 mV. Another possible explanation of the present data with the pre-formed 2-AG pool hypothesis is that OMDM-188 might inhibit the presumptive 2-AG transporter. This explanation has yet to be tested because it remains unclear whether 2-AG is released from neuronal membrane by passive diffusion or is actively carried by a specific transporter. Moreover, the molecular identity of such a presumptive 2-AG transporter remains undetermined, even if it is present. Therefore, it is very difficult to test the possibility that OMDM-188 might inhibit the presumptive 2-AG transporter.

We found that OMDM-188 treatment for 1.5–2.0 h reduced the 2-AG content in hippocampal slices by about 30%. These data per se do not support any particular hypothesis. Although the data might support the pre-formed 2-AG pool hypothesis such that inhibition of 2-AG synthesis by OMDM-188 depleted the 2-AG pools, there are alternative interpretations. There are several studies suggesting the presence of constitutive eCB tone, at least under certain experimental conditions, depending on the balance between constitutive synthesis and degradation (Wilson & Nicoll, 2001; Hentges et al. 2005; Slanina & Schweitzer, 2005; Hashimotodani et al. 2007b; Neu et al. 2007; Oliet et al. 2007; Pan et al. 2011; Zhong et al. 2011). We previously examined the roles of the 2-AG degradation enzyme monoacylglycerol lipase (MGL) in the retrograde eCB signalling, by using an MGL inhibitor. Our data suggest that 2-AG is constitutively synthesized, predominantly through a DGL-dependent pathway, released and hydrolysed by presynaptic MGL (Hashimotodani et al. 2007b). It is therefore likely that the inhibition of constitutive 2-AG synthesis by OMDM-188 reduces the 2-AG content and, therefore, the data of about 30% reduction of 2-AG content by OMDM-188 treatment support the on-demand synthesis hypothesis.

Effects of DGL inhibitors on DSI/DSE have been investigated in several previous reports, but the data are not necessarily consistent (Alger & Kim, 2011). For example, THL is reported to block DSI in the hippocampus in some studies (Hashimotodani et al. 2007b, 2008; Edwards et al. 2008; Zhang et al. 2011) but not in others (Edwards et al. 2006; Szabo et al. 2006). Surprisingly, as for the effect of OMDM-188, even the two previous studies that are the bases for the pre-formed 2-AG hypothesis are mutually contradictory. Min et al. showed that 2 μm OMDM-188 did not block DSI in hippocampal slices but blocked eCB-mediated slow self-inhibition (SSI) in the cortex (Min et al. 2010b). In marked contrast, Zhang et al. showed that 5 μm OMDM-188 blocked DSI but did not block I-mGluR-mediated RER by the first DHPG application in hippocampal slices (Zhang et al. 2011). How can these apparently contradictory results be reconciled? One possibility would be that different results may have been attributable to the difference in the depth of the neuronal structures that are the targets of OMDM-188 from the surface of the slice. Since THL and OMDM-188 are lipophilic, they are generally difficult to penetrate into brain tissues. Therefore, these DGL inhibitors have easier access to somas that are located near the surface of the slice than to the dendrites that are ramified and distributed in the depths of the slice. SSI, DSI and I-mGluR-mediated RER are considered to depend on 2-AG produced locally from subcellular neuronal compartments, around the soma for SSI, along the somatodendritic domain for DSI, and around the dendrites with glutamatergic synapses for I-mGluR-mediated RER. It is therefore likely that the difference in the accessibility of OMDM-188 might be the main reason for the apparently differential effects of OMDM-188 on SSI and DSI in the study of Min et al. (2010b). Likewise, the finding that I-mGluR-mediated RER is more resistant to OMDM-188 than DSI (Zhang et al. 2011) could be explained by the different degree of OMDM-188 penetration into the somatodendritic domain (for DSI) and the dendrites with glutamatergic synapses (for I-mGluR-mediated RER). Moreover, whether whole-cell recordings were made from visually identified cell bodies or by using the blind patch-clamp technique might be an important factor, since the cell bodies recorded with the blind-patch method are generally located in the depth of the slice. In contrast to slice preparations, the dissociated culture system has the clear advantage that target neurons are directly exposed to extracellular fluid and lipophilic compounds including OMDM-188 and THL can easily get access to their targets.

Other possible factors that may cause different results for the effect of OMDM-188 include the temperature, the presence or absence of BSA in the extracellular solution, whether or not DGL activity is rate limiting in the 2-AG synthesis for the biological phenomenon under examination, and whether the modulatory effect of 2-AG on synaptic transmission is maximal or submaximal. Because temperature influences the diffusion rate of molecules, the fluidity of membranes and enzyme activities, the difference in temperature may produce different results. As for BSA, it remains uncertain whether the addition of BSA makes any difference. In the present study, we regularly used BSA for OMDM-188 treatment. However, we found that BSA might not be essential for the delivery of OMDM-188 because we observed that OMDM-188 treatment in the absence of BSA was also effective in blocking DSI in hippocampal slices (data not shown). Even though BSA is not necessary, it may help the penetration of drugs into brain slices to some extent. Whether DGL activity is rate limiting in the 2-AG synthesis can also influence the effect of DGL inhibitors. If DGL activity is rate limiting in a certain form of eCB-mediated cellular response but not in others, the former may be more sensitive to partial inhibition of DGL than the latter. Finally, whether 2-AG signalling maximally suppresses synaptic transmission can also affect the results. For example, if synaptic transmission is maximally suppressed by saturating concentrations of 2-AG, 50% reduction in the 2-AG synthesis may not cause a detectable change in the magnitude of 2-AG-mediated synaptic suppression.

The signalling cascades for RER and Ca²⁺-assisted RER are well studied (Hashimotodani et al. 2007a; Kano et al. 2009; Ohno-Shosaku et al. 2012). Strong activation of G_q/11 protein-coupled receptors at basal [Ca²⁺]_i or synergistic G_q/11 protein-coupled receptor activation and elevation of [Ca²⁺]_i drives phospholipase Cβ (PLCβ), leading to the production of diacylglycerol (DG; Hashimotodani et al. 2005; Maejima et al. 2005; Fig. 6B). DG is then hydrolysed to 2-AG by DGLα. It is thought that PLCβ is activated on demand when G_q/11 protein-coupled receptors are activated. In contrast to RER and Ca²⁺-assisted RER, the molecules that provide DG during Ca²⁺-driven ER have not been identified. Nevertheless, real-time imaging studies have demonstrated that DG, or its parallel byproduct inositol trisphosphate, is generated from phosphatidylinositol 4,5-bisphosphate by Ca²⁺ elevation alone in neurons (Okubo et al. 2001; Codazzi et al. 2006). Moreover, both DG and 2-AG are generated by Ca²⁺ elevation in synaptosomes (Oka et al. 2007). All these results suggest that neurons are capable of producing 2-AG on demand upon G_q/11 protein-coupled receptor activation and/or Ca²⁺ elevation, causing RER, Ca²⁺-assisted RER and Ca²⁺-driven ER (Fig. 6B).

Glossary

2-AG

2-arachidonoylglycerol

BSA

bovine serum albumin

CB₁

cannabinoid receptor type 1

DG

diacylglycerol

DGL

diacylglycerol lipase

DHPG

dihydroxyphenylglycine

DSE

depolarization-induced suppression of excitation

DSI

depolarization-induced suppression of inhibition

eCB

endocannabinoid

ER

endocannabinoid release

I-mGluR

group I metabotropic glutamate receptor

MS

medium spiny

PC

Purkinje cell

PF

parallel fibre

PLCβ

phospholipase Cβ

RER

receptor-driven endocannabinoid release

SSI

slow self-inhibition

THL

tetrahydrolipstatin

Additional information

Competing interests

None declared.

Author contributions

Y.H., T.O.-S. and M.K. designed the research; Y.H., A.T., Y.K., Y.S and T.O.-S. performed the research and analysed the data; T.S. contributed unpublished reagents/analytical tools; V.DiM. contributed a new compound and critically reviewed the paper; Y.H., T.O.-S. and M.K. wrote the paper. The experiments were carried out at the University of Tokyo and Kanazawa University. All authors approved the final version of the paper.

Funding

This work was supported by Grants-in Aid for Scientific Research (21220006 and 25000015 to M.K.), the Strategic Research Program for Brain Sciences (Development of Biomarker Candidates for Social Behavior), and Global COE program (Integrative Life Science Based on the Study of Biosignaling Mechanisms) from MEXT, Japan.

Author's present address

Y. Hashimotodani: Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY 10461, USA.