Regions from the dentate molecular layer of juvenile rats were examined to investigate the cellular and subcellular localization of the CB1 receptor (Fig. 1). Astrocytes and their processes were identified by DAB immunodeposits of GFAP and the CB1 receptor was detected by immunogold labeling. As expected, presynaptic inhibitory terminals that formed symmetric synapses highly expressed CB1 receptor immunoparticles. In contrast, labeling was less abundant at excitatory terminals that formed asymmetric synapses with dendritic spines, as well as on astrocytic and mitochondrial membranes.

In the initial experiments we sought to determine if administering a LFS train of 6000 pulses at 10 Hz could reliably induce LTD in the juvenile rat DG. We found that the 10 Hz LFS induced a significant and immediate short-term depression (STD; −78.25 ± 3.96%, P = 3.12 × 10⁻⁵; n = 16 slices from seven animals) that transitioned into a longer-lasting (60 min) LTD of MPP evoked synaptic responses (−18.74 ± 3.50%; P = 0.00020; n = 16 slices from seven animals; Fig. 2). The induction of LTD with the 10 Hz-LFS had no effect on paired-pulse ratios (P = 0.22), indicating that this LTD does not alter presynaptic neurotransmitter release probability. In agreement with results obtained in mice (Peñasco et al. 2019), the LTD induced by the 10 Hz LFS could be blocked by the CB1 inverse agonist, N-(Piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide (AM 251; 4 µM; −0.55 ± 4.87%; n = 6 slices, 2 animals; P = 0.012) and by bath application of the selective metabotropic glutamate receptor 5 (mGluR₅) antagonist 2-Methyl-6-(phenylethynyl)pyridine (MPEP; 10 µM; 2.07 ± 3.71%; n = 4 slices, two animals; P = 0.0026). The competitive NMDAR antagonist DL-2-Amino-5-phosphonopentanoic acid (DL-APV; 50 µM), however failed to attenuate 10 Hz-LTD (−21.45 ± 10.96%; n = 6 slices, four animals; P = 0.82). The failure of the NMDAR antagonist to even attenuate the 10 Hz LFS induced LTD was in stark contrast to the large effect size observed with AM 251 (d = 1.31) and MPEP (d = 1.54).

The delivery of 900 pulses at 1 Hz induced significant de novo STD (−46.30 ± 4.67% LTD; n = 11 slices, eight animals; P = 1.79 × 10⁻⁶) and LTD (−24.85 ± 4.30% LTD; n = 11 slices, 8 animals; P = 0.000216; Fig. 3) in the DG. While 10 Hz and 1 Hz-induced similar magnitudes of LTD (P = 0.288), the 1 Hz LFS induced significantly less STD than was observed with the 10 Hz stimuli (P = 3.12 × 10⁻⁵; d = 0.52), likely reflecting the difference in the number of stimuli administered (10,000 vs. 900). As with the LTD induced by the 10 Hz stimuli, the LTD induced by 1 Hz LFS did not alter paired-pulse ratios (P = 0.52), suggesting that this LTD also did not involve a change in presynaptic release probability. In contrast, while GABA_A antagonist picrotoxin (PTX; 100 µM) was required for the induction of LTD with the 10 Hz stimuli, it did not have any effect on the magnitude of LTD induced by the 1 Hz LFS (−27.59 ± 6.60%; n = 9 slices, four animals; P = 0.509; data not shown).

To determine if the 10 Hz stimuli and the 1 Hz stimuli induced LTD that required the activation of similar receptors, we examined the capacity for the 1 Hz LFS to induce LTD in the presence of MPEP (−20.15 ± 6.40%; n = 6 slices, four animals), AM 251 (−15.83 ± 6.24%; n = 17 slices, eight animals) or DL-APV (−26.68 ± 5.11%; n = 6 slices, three animals). In each case, blocking either mGluR, CB1, or NMDA receptors alone had no effect on the magnitude of 1 Hz-LTD (P = 0.789; P = 0.556; P = 0.17). Similarly, inhibiting L-type calcium channels with nimodipine (20 µM) also failed to block 1 Hz-induced LTD (−26.80 ± 2.88%; n = 2 slices, two animals; P = 0.731). Surprisingly, preventing the phosphorylation of tyrosine residues on the GluA2 subunit (10 µM Tat-GluA2_3Y peptide), thought to be critical for AMPAR endocytosis (Man et al. 2000; Ahmadian et al. 2004) failed to block 1 Hz-LTD in the DG with the LFS stimulation (−22.79 ± 1.63%; n = 6 slices, four animals; P = 0.235). A complete blockade of the LTD induced by the 1 Hz LFS was only reliably observed in the presence of both AM 251 and MPEP simultaneously (−0.82 ± 4.14%; n = 10 slices, six animals; P = 0.000548; effect size d = 1.75). These data indicate that both mGluR and CB1 receptors are involved in LTD induced by both stimulation paradigms, however inhibition of both receptors are required to block 1 Hz LTD whereas inhibition of either receptor alone is sufficient to block 10 Hz-induced LTD.

Since both CB1 and mGluR₅ receptors appear to be necessary for the induction of LTD in the DG, we sought to determine if the induction of LTD with 10 Hz stimulation (6000 pulses) would occlude, or reduce, the induction of LTD with the 1 Hz LFS (900 pulses). The application of the 10 Hz stimuli induced a robust LTD (−21.23 ± 3.67%), however the subsequent application of the 1 Hz stimuli resulted in significantly more LTD being induced (∼−17%) for a combined LTD magnitude of −37.99 ± 5.02% (n = 11 slices, six animals; P = 0.00176; Fig. 4). There was a large effect size for the amount of LTD induced by the 1 Hz LFS following 10 Hz-LTD (d = 1.15) induction, indicating that both of these two stimulus paradigms may be recruiting mGluR and CB1 receptors.

To better determine the role of mGluR receptors in LTD in the DG, we first sought to pharmacologically induce mGluR-dependent LTD by bath applying (RS)-3,5-Dihydroxyphenylglycine (DHPG; 30 µM) (Camodeca et al. 1999). The bath application of DHPG reliably induced a persistent de novo LTD of synaptic efficacy (−21.96 ± 3.68%; P = 0.00055; n = 8 slices, five animals; Fig. 5). When DHPG was coapplied with AM 251 the LTD was abolished (−0.44 ± 3.20%; n = 5 slices, four animals; P = 0.001; d = 2.30). Similarly, coapplication with MPEP led to a significant attenuation of the synaptic depression MPEP (−10.18 ± 3.50%; n = 9 slices, five animals; P = 0.035; d = 1.13). The induction of the mGluR-dependent LTD did not occlude further LTD induced by 1 Hz LFS (−41.86 ± 4.97%; n = 8 slices, five animals; P = 0.00506; d = 1.61), indicating that while these forms of synaptic depression recruit overlapping mechanisms they are distinct.

To determine whether LTD can be induced solely by activation of CB receptors, we pharmacologically activated these receptors with Win-55, 212-2 (Win, 5 µM), a high affinity agonist of CB1 but with agonist properties for the lesser-expressed CB2 receptors in the DG (Peñasco et al. 2019). Bath application of Win alone caused a significant depression in synaptic efficacy (29.6 ± 7.49%; n = 4 slices, one animal; P = 0.024; Fig. 6). These data provide evidence that CB1 receptor activity alone is sufficient in inducing synaptic depression in the juvenile rat DG.

All animal procedures were approved by, and performed in accordance with, the University of Victoria animal care committee's regulations. Sprague Dawley rats were housed in pairs or triplets under standard, unenriched conditions prior to experimental use with ad libitum access to solid rodent chow and drinking water. The facility was maintained on a 12-h light:dark cycle with lights on at 5:00 a.m.

The method used was described previously (Puente et al. 2011). Coronal hippocampal vibratome sections were cut at 50 µm and collected in 0.1 M PBS (pH 7.4) at room temperature (RT), preincubated in 10% bovine serum albumin, 0.1% sodium azide, and 0.02% saponin prepared in Tris-HCl buffered saline, pH 7.4, for 30 min (RT) and incubated with the primary goat polyclonal anti-CB1 receptor antibody (1:100; Frontier Institute Co., ltd; goat polyclonal; CB1-Go-Af450; AB_2571592) together with the mouse monoclonal anti-GFAP antibody (1:1000; Sigma-Aldrich; mouse monoclonal; G3893; AB_257130) on a shaker for 2 d at 4°C. Tissue was incubated in a secondary 1.4-nm gold-labeled rabbit anti-goat immunoglobulin G (Fab′ fragment; 1: 100; Nanoprobes Inc., cat#2004) and in a biotinylated anti-mouse secondary antibody (1: 200; Vector Labs; BA-2000; AB_2313581) for 4 h at RT. Then, sections were incubated in avidin-biotin peroxidase complex (ABC) (1:50), subsequently washed in 1% BSA/TBS overnight at 4°C and postfixed in 1% glutaraldehyde in TBS for 10 min at RT. Gold particles were silver intensified with an HQ Silver kit (Nanoprobes Inc.) for ∼12 min (in dark). Tissue was then incubated in 0.05% diaminobenzidine (DAB) and 0.01% hydrogen peroxide prepared in 0.1 M PB for 3 min. Labeled sections were osmicated (1% osmium tetroxide in 0.1 M PB, pH 7.4, 20 min), dehydrated in graded alcohols to propylene oxide, and plastic-embedded in Epon resin 812. Ultrathin sections (50 nm thick) were collected on mesh nickel grids, counterstained with 2.5% lead citrate for 20 min and examined with a Philips EM208S electron microscope. Tissue sections were imaged using a Morada digital camera (Olympus) coupled to the electron microscope. Figures were created with Adobe Photoshop (CS3, Adobe Systems; RRID: SCR_014199).

As described previously, juvenile male (n = 41) and female (n = 27) rats were used to examine LTD in the adolescent hippocampal DG (PND 21–28) (Fontaine et al. 2019). Animals were anesthetized with isoflurane until unresponsive and then rapidly decapitated. Brains were quickly removed in cold (1°C–2°C) artificial cerebrospinal fluid (aCSF; 125 mM NaCl, 2.5 mM KCl, 1.25 mM NaH₂PO₄, 25 mM NaHCO₃, 2 mM CaCl₂, 1.3 mM MgCl₂ and 1.4 mM Dextrose) equilibrated with carbogen (95% oxygen, 5% carbon dioxide). Transverse slices (400 µm) were then cut using a vibratome (Pelco 100, Ted Pella Inc.) in the cold aCSF, before being allowed to recover in warmed aCSF (32 ± 0.5°C) for at least 1 h prior to being used for experimentation.

For all recordings, slices were submerged in, and continuously perfused with, carbogenated aCSF at 30 ± 0.5°C. An upright microscope (Olympus BX51WI, Olympus) was used to visually position a concentric bipolar stimulating electrode (FHC, Bowdoinham, ME) into the middle third of the dentate molecular layer (DML) to activate medial perforant path (MPP) fibres. Extracellular field excitatory postsynaptic potentials (fEPSPs) were recorded using a glass micropipette (1 MΩ), filled with aCSF, and visually positioned in the middle third of the DML to align with the stimulating electrode. fEPSP data were collected using an Axon Multiclamp 700B amplifier and recorded using Clampex 10.5 software (Axon Instruments, Molecular Devices; RRID: SCR_011323).

Electrode placement was optimized on a slice-by-slice basis. All slices included in this study had maximal fEPSP amplitudes of at least 0.7 mV. The stimulation intensity was reduced to 70% of the maximal response for all experiments. Single pulse stimulation (0.067 Hz) was delivered until a stable fEPSP slope (<10% variability) was maintained for at least 20 min prior to application of one of the following conditioning stimulus (CS) protocols.

Prior to the induction of LTD and at the end of the postconditioning recording, the presynaptic neurotransmitter release probability was evaluated using a paired-pulse stimulation protocol. To this end, a pair of pulses (intensity unchanged) were delivered 50 msec apart and a ratio was constructed comparing the slope of the second fEPSP relative to the slope of the first fEPSP.

The antagonists N-(Piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide (AM251; 4 µM, Tocris Bioscience; dissolved in 15% DMSO) 2-Methyl-6- (phenylethynyl)pyridine (MPEP; 10 µM; Tocris Bioscience) and DL-2-Amino-5-phosphonopentanoic acid (DL-APV; 50 µM; Tocris Bioscience) were used to inhibit CB1, mGluR₅, and NMDA receptors, respectively (Table 1). The antagonist nimodipine (NIMO; 20 µM; Tocris Bioscience) was used to block L-type calcium channels and a synthetic peptide, Tat-GluA2_3Y (Tat-GluA2; 10 µM; donated by Yu Tian Wang, University of British Columbia) was used to block phosphorylation of tyrosine residues on the GluA2 classically associated with AMPAR endocytosis (Man et al. 2000; Ahmadian et al. 2004). Unless otherwise specified, receptor antagonists were dissolved in deionized water, aliquoted and stored at −20°C until use when they were thawed and pipetted directly into aCSF to yield the appropriate concentration. All antagonists were applied for the last 10 min of the preconditioning recording period, as well as during the administration of any of the conditioning protocols with the exception of Tat-GluA2 which was applied for the complete preconditioning period, throughout the conditioning protocol and for 30 min of the postconditioning recording.

For all recordings, the average fEPSP slope is presented as the mean ± SEM, and graphically is depicted as the average of four responses (i.e., 1 min). Student's t-tests were used to statistically compare magnitudes of LTD. Statistical significance was set at P < 0.05. Effect sizes were calculated as Cohen's d and are provided when statistical significance was achieved.