We studied, in behaving mice, the contribution of CB1 receptors to the activity-dependent changes induced at the hippocampal CA3–CA1 synapse by associative learning and following experimentally evoked long-term potentiation (LTP). Mice were classically conditioned to evoke eyelid responses with a trace paradigm using a tone as conditioned stimulus (CS) and an electric shock as unconditioned stimulus (US). Field excitatory postsynaptic potentials (fEPSPs) were evoked at the CA3–CA1 synapse during the CS–US interval across training. Conditioning was performed in presence of an agonist (WIN55,212-2) alone or with an antagonist (AM251) of the CB1 receptor, injected either systemically or locally. Conditioned responses (CRs) and fEPSP potentiation were depressed by WIN55,212-2. LTP was evoked by high-frequency stimulation of Schaffer collaterals after systemic or local WIN55,212-2 and AM251 injections. WIN55,212-2 affected the induction phase of LTP, mainly when injected locally. The addition of AM251 canceled out the effects of WIN55,212-2. Similar experiments were carried out in animals lacking the CB1 receptor (CB1−/− mice) and following silencing of hippocampal CB1 receptors (CB1R-siRNA–injected animals). In this case, CRs (CB1−/− mice) and LTP (CB1−/− and CB1R-siRNA–injected mice) reached lower values than their respective controls. Results offer new insights for understanding CB1 receptor contribution to associative learning and to CA3–CA1 synaptic plasticity.
Mammalian tissues contain at least 2 types of cannabinoid receptor: CB1 and CB2 (Matsuda et al. 1990; Munro et al. 1993). Both are classic G-protein–coupled receptors containing 7 transmembrane domains (Howlett 1995). The cannabinoid CB1 receptor is present in the central nervous system (CNS) and also in certain peripheral tissues, whereas CB2 has thus far been located primarily in peripheral tissues. However, recent investigations have shown CB2 receptor expression in neuronal populations (Morgan et al. 2009). CB1 receptors are expressed at nerve terminals, and an important function of these receptors is to suppress the release of a range of excitatory and inhibitory neurotransmitters (Gessa et al. 1997; Shen and Thayer 1998).
Learning and memory impairments are among the most commonly reported behavioral effects of cannabinoids (Lichtman et al. 1995; Pamplona and Takahashi 2006). These effects are thought to be associated with the hippocampus, since CB1 receptors are expressed at especially high density in the dentate gyrus, CA1, and CA3 regions (Herkenham et al. 1990, 1991; Matsuda et al. 1990; Pertwee 1997, 2001; Tsou et al. 1998; Julián et al. 2003; Martín et al. 2008). Both systemic injections (Iwasaki et al. 1992; Lichtman et al. 1995; Ferrari et al. 1999; Varvel et al. 2001; Da Silva and Takahashi 2002) and the direct administration of cannabinoid receptor agonists into the hippocampus (Lichtman et al. 1995; Egashira et al. 2002; Suenaga et al. 2008) induce deficits in various hippocampal-dependent tasks, such as the radial and water mazes. Furthermore, cannabinoid receptor activation in hippocampal slices impairs synaptic transmission and long-term potentiation (LTP) (Nowicky et al. 1987; Collins et al. 1994; Terranova et al. 1995, 1996; Serpa et al. 2009), whereas CB1 blockade enhances them (Terranova et al. 1995, 1996; Hoffman et al. 2007). Taken together, considerable evidence demonstrates that cannabinoid agonists impair, whereas cannabinoid antagonists improve memory and synaptic plasticity (Sullivan 2000; Davies et al. 2002). However, recent studies suggest that the cannabinoid system cannot be categorized into these previously described simple patterns. For example, the neural processes underlying memory formation are differentially sensitive to cannabinoid receptor activation or deactivation depending on the type of memory under examination (Abush and Akirav 2010). In addition, although initial studies using CB1-deficient mice reported increased memory performance in hippocampal-dependent tasks (Terranova et al. 1996) and enhancement of hippocampal LTP (Böhme et al. 2000), it has been reported latterly that the absence of CB1 receptors leads to an accelerated decline in cognitive functions (Bilkei-Gorzo et al. 2005).
In order to obtain a more complete perspective of cannabinoid CB1 receptor actions in alert behaving animals, we attempted to study here how the cannabinoid system contributes to several forms of in vivo hippocampal synaptic transmission processes (input/output curves, paired-pulse (PP) facilitation, and LTP induction in behaving animals), as well as to the acquisition of a well-known associative learning task—namely, the classical conditioning of eyelid responses using a trace paradigm. For this purpose, experimental manipulations at different levels were carried out: from the systemic injection of CB1 agonists and antagonists to the local silencing of hippocampal CB1 receptors (Ramiro-Fuentes et al. 2010), including the direct infusion of a cannabinoid agonist into the hippocampus and experiments carried out with mice lacking the CB1 receptor (Ledent et al. 1999; Aso et al. 2008). Results indicate a definite and complex role of cannabinoid CB1 receptors in hippocampal synaptic plasticity during experimentally evoked LTP and associative learning tasks.
Materials and Methods
Experiments were carried out in C57BL/6 mature (3/5 months old; 25–30 g) male mice obtained from an official supplier (University of Granada Animal House, Granada, Spain). Additional experiments were carried out with mature (3/5 months old) male CB1 knockout (CB1−/−) mice on a CD1 background and their wild-type littermates (CB1+/+) provided by the Animal House of the Pompeu Fabra University (Barcelona, Spain). In order to homogenize the genetic background of the mice, the first-generation heterozygotes were bred for 15 generations on a CD1 background (Charles River, L'Arbresle, France) with selection for the mutant CB1 gene at each generation. The CB1 receptor knockout mice derived from the backcrossing of chimeric CD1-CB1 receptor knockout mice developed by Ledent et al. (1999) with wild-type CD1 females (Charles River). All CB1−/− and CB1+/+ mice used in the experiment originated from the same breeding and were matched for age and weight.
Upon arrival, animals were housed in shared cages (n = 10 per cage), but they were switched to individual cages after surgery. Mice were kept on a 12 h light:dark cycle with constant ambient temperature (21.5 1 °C) and humidity (55 8%). Food and water were available ad libitum. Experiments were carried out in accordance with the guidelines of the European Union (2003/65/CE) and recent Spanish regulations (BOE 252/34367-91, 2005) for the use of laboratory animals in chronic studies. All experimental protocols were also approved by the local Ethics Committee of the Pablo de Olavide University (Seville, Spain). Unless otherwise indicated, a total of 10 successful animals were used per experimental group. We considered successful animals those that finished the selected experimental protocol presenting field excitatory postsynaptic potentials (EPSPs) (fEPSPs) and electromyographic (EMG) recordings that did not deteriorate over time because of disconnection or polarization of recording and/or stimulating electrodes. Additional animals were used in preliminary studies to select the appropriate doses of the drugs used in this study, lentivirus diffusion in the hippocampus, as well as the stability of the recording and stimulating systems.
Animals were anesthetized with 0.8–1.5% isoflurane, supplied from a calibrated Fluotec 5 (Fluotec-Ohmeda, Tewksbury, MA) vaporizer, at a flow rate of 1–2 L/min oxygen (AstraZeneca, Madrid, Spain) and delivered by a mouse anesthesia mask (David Kopf Instruments, Tujunga, CA). Animals were implanted with stimulating electrodes on the left supraorbital nerve and with recording electrodes in the ipsilateral orbicularis oculi muscle (Fig. 1A). These electrodes were made from 50 μm, Teflon-coated, annealed stainless steel wire (A-M Systems, Carlsborg, WA). Mice were also implanted with stimulating electrodes in the contralateral (right) Schaffer collateral/commissural pathway of the dorsal hippocampus (2 mm lateral and 1.5 mm posterior to bregma and 1–1.5 mm from the brain surface; Paxinos and Franklin 2001) and with a recording electrode placed in the right CA1 stratum radiatum (1.2 mm lateral and 2.2 mm posterior to bregma and 1–1.5 mm from the brain surface). These hippocampal electrodes were made from 50 μm, Teflon-coated tungsten wire (Advent Research, Eynsham, UK). A 26G stainless steel cannula (Plastic One, Reanoke, VA) was implanted close to the recording hippocampal electrode (1.6 mm lateral and 1.8 mm posterior to bregma and 1 mm from the brain surface, i.e., 0.5 mm above the infusion target) and a bare silver wire affixed to the bone as ground. All the implanted wires were soldered to two 4-pin sockets (RS Amidata, Madrid, Spain) and fixed to the skull with dental cement (Fig. 1A,B; for details, see Gruart et al. 2006).
Whenever it was required (either for drug administration or lentivirus injection), animals were also implanted chronically with a blunted, stainless steel, 26G cannula (Plastic One) in the CA1 pyramidal layer contralateral to the implanted eyelid (1.6 mm lateral and 1.8 mm posterior to bregma and 1 mm from the brain surface; Paxinos and Franklin 2001). The tip of the cannula was aimed to be located 0.5 mm above the infusion target (Fig. 1B).
Recording and Stimulation Procedures
Recording sessions were carried out with 4 animals at a time. Animals were placed in separate small (5 × 5 × 10 cm) plastic chambers located inside a larger Faraday box (30 × 30 × 20 cm). The EMG activity of the orbicularis oculi muscle was recorded with Grass P511 differential amplifiers (Grass-Telefactor, West Warwick, RI) at a bandwidth of 0.1 Hz–10 kHz. Field EPSP recordings were also made with Grass P511 differential amplifiers through a high-impedance probe (2 × 1012 Ω, 10 pF).
Input/Output Curves and PP Facilitation
For input/output curves (Fig. 2), animals were stimulated at the Schaffer collaterals with PPs (40 ms of interstimulus interval) at increasing intensities (0.02–0.3 mA). Pulses consisted of 100-μs, square, biphasic (negative–positive) waves. We also checked the effects of PPs at different (10, 20, 40, 100, 200, and 500 ms) interstimulus intervals and intensities corresponding to 30–40% of the amount necessary to evoke a saturating response. In all cases, the pair of pulses of a given intensity was repeated ≥5 times with time intervals ≥30 s to avoid as much as possible interferences with slower short-term potentiation (augmentation) or depression processes (Zucker and Regehr 2002). Moreover, to avoid any cumulative effects, intensities and intervals were presented at random. At the range of intensities used here, no population spikes were observed in the collected recordings. For baseline recordings and to determine the stability of fEPSPs evoked by the electrical stimulation of Schaffer collaterals, single pulses or PP (40 ms of interval) were presented at a rate of 1/20 s (Fig. 1C).
LTP Evoked by High-Frequency Stimulation
Field EPSP baseline values (Figs 3 and 4) were collected 15 min prior to LTP induction using paired (40 ms interstimulus interval) 100-μs, square, biphasic pulses. Pulse intensity was set at 30–40% of the amount necessary to evoke a maximum fEPSP response (0.05–0.15 mA)—that is, well below the threshold for evoking a population spike (Gureviciene et al. 2004; Gruart et al. 2006). An additional criterion for selecting stimulus intensity was that the second stimulus evoked a larger (>20%) fEPSP than the first (Bliss and Gardner-Medwin 1973). For LTP induction, animals were presented with a high-frequency stimulation (HFS) session consisting of five 200 Hz, 100 ms trains of pulses at a rate of 1/s. This protocol was presented 6 times at intervals of 1 min. Thus, a total of 600 pulses were presented during an HFS session. In order to avoid evoking large population spikes and/or the appearance of electroencephalography (EEG) seizures, the stimulus intensity during HFS was set at the same as that used for generating baseline recordings. Animals presenting after discharges or motor seizures following the HFS protocol (as checked by online EEG recordings and visual observation of the stimulated animal) were excluded from the study. After each HFS session, the same PP stimuli (40 ms interstimulus interval) were presented every 20 s for 30 min and for 15 min the following day (Figs 3 and 4; for details, see Gruart et al. 2006; Madroñal et al. 2007).
Classical Eyeblink Conditioning
Classical conditioning was achieved using a trace paradigm (Fig. 5). For this, a tone (20 ms, 2.4 kHz, 85 dB) was presented as a conditioned stimulus (CS). The unconditioned stimulus (US) consisted of a 500-μs, 3× threshold, square, cathodal pulse applied to the supraorbital nerve. The US started 500 ms after the end of the CS. A total of 2 habituation and up to 12 conditioning sessions were carried out (Fig. 6). A conditioning session consisted of 60 CS–US presentations and lasted ≈30 min. For a proper analysis of the conditioned response (CR), the CS was presented alone in 10% of the cases. CS–US presentations were separated at random by 305 s. Animals received just 1 training session per day. For habituation sessions, only the CS was presented, also for 60 times per session, at intervals of 305 s. As criteria, we considered a “CR” the presence, during the CS–US interval, of EMG activity lasting >10 ms and initiated >50 ms after CS onset. In addition, the integrated EMG activity recorded during the CS–US interval had to be at least 2.5 times greater than the averaged activity recorded immediately before CS presentation (Porras-García et al. 2005).
Field EPSPs were evoked at the CA3–CA1 synapse during habituation and conditioning sessions by a single 100-μs, square, biphasic (negative–positive) pulse applied to Schaffer collaterals 300 ms after CS presentation (Fig. 5A). Stimulus intensities ranged from 0.05 to 0.35 mA.
The CB1 receptor agonist WIN55,212-2 and the selective antagonist AM251 (Sigma-Aldrich, Madrid, Spain) were used in this study. Doses of WIN55,212-2 (1.5 mg/kg) and AM251 (3 mg/kg) were determined following preliminary tests. Drugs were dissolved in 0.9% saline and Tween 80 (Sigma-Aldrich), at a 9.75:0.25 ratio, with the help of a shaker (≈10 h). Drugs were injected subcutaneously (s.c.) dissolved in 0.4 mL of vehicle. For intrahippocampal injections (Figs 4 and 6), 1.5 μg of WIN55,212-2 were dissolved in 2 μL of vehicle and injected through the cannula at a rate of 0.2 μL/min. Unless otherwise indicated, AM251 was injected 15 min before WIN55,212-2, and the latter was injected 15 min before the selected experiment.
In Vivo Silencing of the Hippocampal CB1 Receptor
Silencing of CB1 receptor expression in vivo was carried out in accordance with recent reports from our laboratory (Martín et al., 2008; Ortiz et al. 2010; Ramiro-Fuentes et al. 2010). Three sequences were designed, targeted to different regions of the CB1 receptor messenger RNA (mRNA) sequence. Based on Hannon's design criterion (Dreyer 2010; Ortiz et al. 2010; Ramiro-Fuentes et al. 2010), the following targets within the mRNA sequence were selected: First sequence, AAA AAA GTA CCT GCA AGG CCG CCT AAG ATC AAC CTC AAG CTT CAA GTC GAT CTT AGA CGG CCT TGC AGA TAC GGT GTT TCG TCC TTT CCA CAA; second sequence, AAA AAA GGC AGA CGT ATC CGT AGA CAC AGA CAT AGC AAG CTT CCC ATG TCT GTG TCC ACA GAC ACG TCT GCC GGT GTT TCG TCC TTT CCA CAA; and third sequence, AAA AAA GGC CTG ATG ACG ATC CTC CTA CAG ACC AGC AAG CTT CCT GGC CTA TAA GAG GAT CGT CAC CAG GCC GGT GTT TCG TCC TTT CCA CAA.
To each oligo, a NheI restriction site was added at 3′ and a U6-3′–specific 10mer at 5′. Using the pSilencer 1.0-U6 (Ambion, Huntingdon, UK) as a template and a U6 promoter-specific forward primer containing XhoI restriction site (GCT CGA GCC GCT CTA GAA CTA GTG C), each siRNA target was added to the mouse U6 promoter by polymerase chain reaction (PCR). The PCR conditions were highly drastic to prevent mutations within the targets. The following PCR program was performed: 120 s at 94 °C (initial denaturation), followed by 94 °C (45 s), 64 °C (45 s), and 72 °C (45 s), repeated over 35 cycles. The PCR reaction contained 4% dimethyl sulfoxide (Sigma-Aldrich, Switzerland). The PCR products were digested with NheI and XhoI, cloned into similar sites into pTRIPΔU3[PGK-B2-IRES2-eGFP-WPR] (gift of Dr Uwe Maskos, Pasteur Institute, Paris, France), after removing B2 sequence using the same restriction sites, and sequenced to verify the integrity of each construct. These vectors are nonregulatable. To sum up, 3 nonregulatable lentiviruses, Lenti-CB1R-siRNA1, Lenti-CB1R-siRNA2, and Lenti-CB1R-siRNA3, were constructed. They expressed small hairpin RNAs, aimed at silencing CB1R expression, and specifically targeted against different regions of the CB1R mRNA.
Lentivirus was produced as follows: The vector plasmids, together with the packaging construct plasmid and the envelope plasmid, were co-transfected into HEK293T cells to produce the viral particles (Bahi, Boyer, Gumy, et al. 2004; Bahi, Boyer, Kafri, and Dreyer 2004; Ortiz et al. 2010). The viral titers were determined by p24 antigen measurements (KPL, Gaithersburg, MD). For in vivo experiments, the different viral stocks were matched for viral particle content, mixed, and used at 2 × 109 particles/μL.
Mice were injected with 2 μL of the mixed concentrated lentiviral stock (CB1R-siRNAs), slowly infused at a rate of 0.5 μL/min. Sham animals were injected in the same area with viral particles obtained with the same lentiviral vector not containing the CB1R-siRNA sequence, named Lv-green fluorescent protein (GFP) (Figs 7–9).
At the end of the experiments, mice were deeply reanesthetized (4% chloral hydrate solution, 10 mL/kg) and perfused transcardially with saline and 4% phosphate-buffered paraformaldehyde. Their brains were removed, postfixed overnight at 4°C, and cryoprotected in 30% sucrose in phosphate-buffered saline. To determine the location of implanted electrodes and cannula, dorsal hippocampus sections were obtained in a microtome (Leica, Wetzlar, Germany) at 50 μm, mounted on glass slides, and stained with 0.1% toluidine blue (Fig. 1B). For lentiviral diffusion, free-floating sections were stained with a polyclonal rabbit anti-green fluorescent protein (anti-GFP) antiserum (diluted 1:2000; Molecular Probes/Invitrogen, Eugene, OR) using the standard immunocytochemical fluorescence protocols (Darmopil et al. 2009; Granado et al. 2011) and using an Alexa fluor 488 green as secondary antibody (1:400; Molecular Probes/Invitrogen). Sections were then mounted on gelatin-coated slides in Mowiol solution (Calbiochem, San Diego, CA) and stored at 4 °C in the dark until visualized.
Data Storage and Analysis
EMG and hippocampal extracellular activity and 1-V rectangular pulses corresponding to CS and US presentations were stored digitally on a computer through an analog/digital converter (CED 1401 Plus, Cambridge, England). Data were analyzed off-line for quantification of CRs and fEPSPs with the Spike 2 (CED) program. The slope of evoked fEPSPs was computed as the first derivative (volts/second) of fEPSP recordings (volts). Three to 5 successive fEPSPs were averaged, and the mean value of the slope during the rise-time period (i.e., the period of the slope between the initial 10% and the final 10% of the fEPSP) was determined. Computed results were processed for statistical analysis using the Sigma Stat for Windows package. Data are always represented as the mean standard error of the mean. Acquired data were analyzed using a 2-way analysis of variance, with days as repeated measure and with a contrast analysis for a further study of significant differences.
Input–Output Curves Evoked at the Hippocampal CA3–CA1 Synapse Were Modified with the Subcutaneous Administration of the CB1 Receptor Agonist WIN55,212-2
In order to characterize the functional properties of the CA3–CA1 synapse under CB1 receptor activation conditions, the facilitation evoked by the presentation of a pair of pulses to Schaffer collaterals was tested in the presence of both the cannabinoid agonist (WIN55,212-2; 1.5 mg/kg, s.c.) and the antagonist (AM251; 3 mg/kg, s.c.). Drugs were injected 30 min before the test. We recorded the changes in fEPSP slopes evoked in the pyramidal CA1 area by PP (40 ms interval) stimulation in vehicle-, WIN-, and WIN + AM–injected mice (Fig. 2). As shown in Figure 2A–C for controls, the slope of fEPSPs evoked in the CA1 area by the first pulse (black triangles) increased steadily with current strength (range 0.02–0.3 mA). In contrast, fEPSPs evoked by the second pulse (white triangles) increased initially in parallel with (but with larger values than) fEPSPs evoked by the first pulse until stimuli >0.20 mA, from which the fEPSP slopes evoked by the second stimulus were smaller than those evoked by the first (* in Fig. 2B; F14,126 = 4.988, P < 0.001). As previously described for behaving mice (Madroñal et al. 2009), the addition of fEPSP slopes evoked by the 2 stimuli (first + second, black circles, Fig. 2B) presented a sigmoid-like shape. Furthermore, the second/first PP ratio (Fig. 2C, black squares) decreased progressively from facilitation to depression, with an inflexion point at ≈0.2 mA.
WIN-injected mice presented input/output curves different from controls (Fig. 2D–F). First, the fEPSP evoked by the first pulse was smaller than the fEPSP evoked by the second stimulus for the whole range of intensities tested (0.02–0.3 mA), with significant differences in the slopes of fEPSPs evoked by the first and the second pulse for intensities >0.1 mA (F14,126 = 1.871, P < 0.001; Fig. 2E). As a consequence (Fig. 2F), PP facilitation was not changed into PP depression as in control animals, even at high intensities. Nevertheless, the addition of fEPSP slopes evoked by the first and the second stimuli (black circles, Fig. 2E) continued presenting a sigmoid-like shape, as previously described for vehicle-injected animals. Finally, fEPSP slopes evoked by the first pulse in WIN-injected animals were significantly (P < 0.001) lower than those evoked by the first pulse in controls for intensities >0.16 mA.
The effects of WIN55,212-2 on input/output curve were antagonized by a simultaneous injection of AM251 (Fig. 2G–I), since the evolution of the first and the second fEPSP evoked by the same pair of pulses at the same range of intensities (0.02–0.3 mA) was similar to those evoked in vehicle-injected mice. This group (WIN + AM) presented significant PP facilitation of fEPSPs evoked by the second pulse with respect to the first (F14,126) = 42,052, P < 0.001; Fig. 2H) for both low intensities. Indeed, the PP facilitation was changed into PP depression at high intensities as previously described for controls (>0.18 mA; Fig. 2I, black squares).
We next checked the effects of PP stimulation at a fixed intensity (≈40% of the amount needed for evoking a maximum fEPSP response) of the CA3–CA1 synapse at different time intervals in controls and WIN- and WIN + AM–injected mice (not illustrated). In this test, no significant differences were found between data collected from the 3 experimental groups (F10,90 = 0.436, P = 0.925). However, the 3 groups of animals presented significant PP facilitation of fEPSPs evoked by the second pulse with respect to the first at 40 ms interpulse intervals (F5,45 = 7.643, P < 0.001).
LTP Evoked at the Hippocampal CA3–CA1 Synapse Was Depressed by the Parenteral Administration of CB1 Receptor Agonist WIN55,212-2
One week after the 2 tests mentioned above, the 3 groups of animals were also used for an LTP study (Fig. 3). Animals were injected (s.c.) with vehicle, WIN55,212-2 (1.5 mg/kg), or WIN55,212-2 + AM251 (1.5 mg/kg of WIN and 3 mg/kg of AM), and 30 min after the injection, they were stimulated with pairs of pulses (40 ms interval; first, black triangles; second, white triangles; Fig. 3A,C,E) during 15 min (at a rate of 3/min) at Schaffer collaterals in order to obtain a baseline for evoked fEPSPs (Fig. 3B,D,F). For LTP induction, each animal received an HFS session (arrows, Fig. 3B,D,F), and its evolution was checked presenting the same pairs of pulses at Schaffer collaterals. With this protocol, the vehicle-injected group presented a significant LTP lasting at least 24 h for both the first (≈160% of the baseline) and the second (≈200%, compared with baseline values for fEPSP evoked by the first pulse) evoked fEPSPs (F11,99 = 5.955, P < 0.001; Fig. 3B). Differences between the first and the second fEPSPs for control mice were significant during the baseline and immediately after the HFS session (F11,99 = 5.575, P < 0.05).
Compared with controls, WIN-injected mice presented a significantly (P < 0.01) smaller fEPSP (≈150% for the first evoked fEPSP and ≈250% for the second fEPSP; Fig. 3B,D). In this case, significant differences between the first and the second evoked fEPSPs were detected not only during baseline recordings but also after the HFS session and even 24 h later (F11,99 = 1.139, P < 0.05; Fig. 3D), indicating that the PP ratio [(second/first) × 100] presented larger values in WIN-injected mice than in controls (see asterisks in Fig. 3B,D).
Mice injected with both the cannabinoid agonist and the antagonist (WIN + AM group, Fig. 3E,F) presented LTP values similar to those reached by the control group—namely, the first evoked fEPSP reached values of ≈200% (over baseline) after the HFS session (F11,99 = 25.481, P < 0.001). In contrast to what was found in the WIN group, significant differences between the first and the second evoked fEPSP were present only during baseline recordings (F11,99 = 4.857, P < 0.05; Fig. 3F). Indeed, the PP ratio between the slopes of the second and the first fEPSP after LTP induction was similar in WIN + AM mice and in controls (not illustrated).
LTP Evoked at the Hippocampal CA3–CA1 Synapse Was Depressed by the Intrahippocampal Administration of CB1 Receptor Agonist WIN55,212-2
LTP induction was also studied after local activation of hippocampal CB1 receptors (Fig. 4). In this experiment, pairs of stimuli (40 ms interval) were presented at Schaffer collaterals during 15 min in order to obtain baseline values. Three groups of animals were studied: controls (2 μL of vehicle injected locally), WIN (1.5 μg dissolved in 2 μL of vehicle and injected locally), and WIN + AM (1.5 μg of WIN injected locally and 3 mg/kg AM, s.c.). Fifteen minutes after injections, the same pair of stimuli was presented for another 15 min in order to obtain a second baseline recording. The aim was to detect any possible change in basal synaptic transmission after drug infusions (Figs 1C and 4B,D,F). A nonsignificant decrease of evoked fEPSPs was detected in the WIN-injected group (Lees and Dougalis 2004).
For LTP induction, animals were presented with the HFS protocol described in the previous section. After the HFS session, fEPSP slope evolution was checked during the following 30 min and, in addition, for 15 min 24 h after the HFS session. In this situation, LTP was induced in vehicle- and in WIN + AM–injected animals, but not in the WIN group. Specifically, and compared with fEPSP slope values recorded during the baseline after the injection, control mice presented a potentiation of ≈250% for the first evoked fEPSP (Fig. 4B, black triangles) and of ≈330% for the second (Fig. 4B, white triangles) during the 30 min following HFS (F14,126 = 5.392, P < 0.001). LTP induced in control animals remained significantly (P < 0.001) larger than baseline values up to 24 h after the HFS session. Significant differences between the first and the second fEPSPs evoked at the CA3–CA1 synapse by PP stimulation were present during the whole test, both before and after LTP induction (asterisks in Fig. 4B, F14,126 = 0.262, P < 0.05).
Regarding the effects of hippocampal CB1 receptor activation on LTP induction, animals locally injected with WIN55,212-2 did not present any significant increase in fEPSP slopes after HFS, and both the first and the second evoked fEPSPs remained around—even below—baseline values (Fig. 4D; F14,126 = 0.258, P = 0.997). Significant differences between the 2 evoked fEPSPs were detected during baseline recordings and also during the 30 min following the HFS session (asterisks in Fig. 4D; F14,126 = 0.546, P < 0.05).
As expected, LTP induced in WIN + AM mice reached values similar to those recorded in controls, indicating that the parenteral injection of the cannabinoid antagonist AM251 reversed the effects on LTP of WIN55,212-2 administered intrahippocampally. Field EPSP slopes recorded in the WIN + AM group (compared with the baseline values after drug injection) were ≈250% for the first evoked fEPSP and ≈310% for the second (F14,126 = 6.651, P < 0.001; Fig. 4F). Differences between the 2 evoked fEPSPs were significant during the whole LTP test (asterisks in Fig. 4F; F14,126 = 0.379, P < 0.05).
Parenteral Injections of WIN55,212-2 Prevent Classical Eyeblink Conditioning and Activity-Dependent Potentiation of the Hippocampal CA3–CA1 Synapse
In another group of experiments, we studied the effects of WIN55,212-2 and AM251 on classical eyeblink conditioning and on activity-dependent potentiation of the CA3–CA1 synapse (Gruart et al. 2006). First, we evaluated whether these 2 drugs have any detectable effect on the CA3–CA1 synapse. For this, animals (n = 4) were stimulated for 15 min at Schaffer collaterals (Fig. 1C). The stimulus consisted of a single 100 μs, square, biphasic pulse presented 3 times/min. In this, as well as in the following experiments, pulse intensity was set at ≈40% of the amount necessary to evoke a maximum fEPSP response (Gruart et al. 2006, Sahún et al. 2007). WIN55,212-2 was then systemically injected (1.5 mg/kg, s.c.), and the same stimulus was presented again to Schaffer collaterals for 60 min. Although there were no significant differences between fEPSP slope values before and after the injection (P = 0.42), it can be seen in Figure 1C that WIN administration had a depressing effect on the CA3–CA1 synapse, which started 20 min after the drug injection and reached the maximum effect 30 min after injection. The same experimental procedure was carried out with AM251 (3 mg/kg); in this case, no detectable change in hippocampal basal synaptic transmission was noticed (P = 0.532, not illustrated).
For classical eyeblink conditioning (Fig. 5), 3 groups of mice (control, WIN, and WIN + AM) were conditioned using a trace paradigm. The experimental design also included the presentation of a single stimulus to Schaffer collaterals 300 ms after CS presentation (Fig. 5A), which allowed us to follow the learning-dependent changes taking place at the hippocampal CA3–CA1 synapse during the acquisition process. WIN55,212-2 (1.5 mg/kg, s.c.) or WIN55,212-2 + AM251 (1.5 mg/kg of WIN, s.c. and 3 mg/kg of AM, s.c.) was administered 30 min before each conditioning session—that is, for 10 consecutive days. Control animals received the same volume of vehicle (0.4 mL) also 30 min before each learning session.
As shown in Figure 5D (black circles), the percentage of CRs in controls increased progressively across conditioning from 27.9% during the first session to 63.3% during the 10th. In contrast, mice injected daily with WIN55,212-2 were unable to present a normal conditioning curve (Fig. 5D, white circles). The percentage of CRs presented by the WIN group stayed around 20% for all the conditioning sessions—that is, well below values observed for the control group. However, the effect of WIN55,212-2 on classical eyeblink conditioning was antagonized by the pretreatment with AM251 (Fig. 5D, gray circles), since animals included in the WIN + AM group were able to learn the trace paradigm as well as vehicle-injected mice did. The percentages of CRs presented by control and WIN + AM groups were statistically different from values presented during the 2 habituation sessions from the 4th to the 10th conditioning sessions (F11,99 = 9.316, P < 0.001). Furthermore, the percentage of CRs observed in WIN-injected mice was significantly smaller than values collected from both control and WIN + AM groups from the 3rd to the 10th training session (asterisks in Fig. 5D, F22,198 = 4.483, P < 0.001).
As shown in Figure 5C, fEPSP slopes evoked in controls by the stimulation of Schaffer collaterals increased progressively across conditioning sessions (Fig. 5C, black triangles), reaching a maximum value of ≈140% during the 10th session (fEPSP slopes recorded during habituation were considered 100%). In contrast and in correlation with the low percentage of CRs acquired by these animals, fEPSPs collected from the WIN-injected group did not present any significant increase across conditioning sessions (Fig. 5C, white triangles). Interestingly, from the first to the seventh conditioning sessions, WIN-injected mice presented fEPSP slopes even smaller than those collected during the 2 habituation sessions, indicating that, as well as previously described for hippocampal slices, the in vivo application of WIN55,212-2 inhibits basal excitatory transmission (Lees and Dougalis 2004).
As expected, AM251 injections reversed the effects of WIN on hippocampal synaptic function (Fig. 5C, gray triangles), since the WIN + AM–injected group presented the normal concomitant learning-dependent synaptic changes at the hippocampal CA3–CA1 synapse (Gruart et al. 2006). Vehicle- and WIN + AM–injected mice presented a similar evolution of their evoked fEPSPs across conditioning sessions, these fEPSP slopes being different from habituation values from the 7th to the 10th conditioning sessions (F11,99 = 6.782, P < 0.001). Statistically significant differences were found between fEPSPs recorded from WIN-injected mice and those collected from vehicle- and WIN + AM–injected animals from the 7th to the 10th conditioning session (asterisks in Fig. 5C; F22,198 = 1.101, P < 0.05). In summary, in vivo CB1 receptor activation by the cannabinoid agonist WIN55,212-2 inhibits both the acquisition of a classical eyeblink conditioning test and the synaptic changes underlying associative learning. In addition, these effects are reversed by the pretreatment with the CB1-specific antagonist AM251.
Local Injections of WIN55,212-2 Prevent Classical Eyeblink Conditioning and Activity-Dependent Potentiation of the Hippocampal CA3–CA1 Synapse
Results set out above clearly demonstrate that CB1 receptor activation impairs the acquisition of a trace eyeblink conditioning and blocks certain forms of both short- and long-term hippocampal synaptic transmissions. However, the effects on learning of WIN55,212-2 administered s.c. could be mediated by nonhippocampal CB1 receptors, since—besides its expression in the hippocampus—this cannabinoid receptor is expressed throughout the CNS (Sullivan 2000; Julián et al. 2003; Martín et al. 2008). To study the specific role of hippocampal CB1 receptors on classical eyeblink conditioning, we performed a new series of experiments in which WIN55,212-2 was injected directly into the CA1 area of the hippocampus through a chronically implanted cannula (Figs 1A and 6A). Intracranial drug infusion took place 10 min before the corresponding session, since a preliminary study showed that the slope of fEPSPs evoked by single pulses presented to Schaffer collaterals started to decrease ≈10 min after the intrahippocampal injection of WIN55,212-2 (data not shown).
Three new experimental groups (control, WIN, and WIN + AM) were classically conditioned as described above. Animals were trained in the same conditions as before, until the seventh conditioning session, where they presented ≈55% of CRs (Fig. 6C). At this point, the percentage of CRs presented by the 3 groups was significantly different from values collected during the 2 habituation sessions (F13,117 = 27.170, P < 0.001). Ten minutes before the eighth conditioning session, control mice were locally injected with 2 μL of vehicle into the CA1 area, the WIN group was injected with 1.5 μg of WIN55,212-2 dissolved in 2 μL, and the WIN + AM group was administered 3 mg/kg of AM251 s.c. 15 min prior to hippocampal WIN55,212-2 injection (1.5 μg).
The local activation of hippocampal CB1 receptors in WIN-injected mice led to a significant (P < 0.001) decrease in the percentage of CRs evoked during the eighth session (white circles, Fig. 6C), which dropped from 54.2% during the seventh conditioning session to 33.3% during the eighth. These results indicate that hippocampal CB1 receptors play a specific role in the acquisition of classical eyeblink conditioning. In contrast, the control and WIN + AM groups did not present any disturbance of their learning curves (Fig. 6C; vehicle, black circles; WIN + AM, gray circles), and both groups reached values of ≈55% of CRs during the eighth conditioning session. Thus, the pretreatment with AM251 reverses the effects of WIN55,212-2 injected into the hippocampus. A comparison of the percentage of CRs presented by the 3 experimental groups after their corresponding injections indicated that differences were statistically significant (asterisks in Fig. 6C; F26,234 = 4.336, P < 0.001).
As expected, during the first 7 conditioning sessions, fEPSP slopes for the 3 experimental groups (Fig. 6B; vehicle, black triangles; WIN, white triangles; WIN + AM, gray triangles) increased progressively until reaching values of ≈110–120% as compared with habituation sessions. In concordance with the evolution in the percentage of CRs, the local injection of WIN55,212-2 caused a marked decrease in the slope of evoked fEPSPs (Fig. 6B, white triangles)>—specifically, from 122.2% during the seventh session to 64.4% during the eighth. In contrast, the local injection of vehicle into the CA1 10 min before the eighth conditioning session caused no detectable changes in fEPSP slopes (Fig. 6B, black triangles). Furthermore, the inhibition of the hippocampal synaptic transmission caused by the local activation of CB1 receptors was prevented by the prior subcutaneous injection of AM251 (Fig. 6B, gray triangles). The slopes of fEPSPs evoked in control and WIN + AM groups versus those evoked in WIN-injected mice were significantly different (asterisks in Fig. 6B; F26,234 = 1.990, P < 0.005).
In order to compare the effects of a single intrahippocampal infusion of WIN55,212-2 with the effects of its systemic administration, we extended the associative learning test for 2 additional conditioning sessions (Fig. 6, 11th and 12th sessions). Thirty minutes before the 11th session, vehicle, WIN (1.5 mg/kg), or WIN + AM (1.5 mg/kg of WIN and 3 mg/kg of AM) was administered (s.c.). The systemic injection of WIN55,212-2 caused a noticeable decrease in the percentage of CRs (from 48.9% reached during the 10th conditioning session to 8.7% during the 11th session; Fig. 6C, white circles). This pronounced decrease in the percentage of CRs suggests that part of the learning impairment produced by the activation of cannabinoid receptors is mediated by nonhippocampal CB1 receptors, since the learning deficit produced by a parenteral injection of WIN55,212-2 was more noticeable than that produced by the intrahippocampal infusion of the same cannabinoid agonist. As expected, the percentage of CRs in control and WIN + AM groups was not modified by vehicle or WIN + AM (s.c.) injections.
Furthermore, recorded fEPSPs evoked by the stimulation of Schaffer collaterals presented a decreased slope in the WIN-injected group (from 109.6% during the 10th conditioning session to 80% during the 11th session). No significant changes in fEPSP slopes were detected in controls and in the WIN + AM group during the 11th conditioning session.
Effects of In Vivo Hippocampal CB1 Receptor Silencing on Synaptic Transmission
We next studied the functional properties of the hippocampal CA3–CA1 synapse after the in vivo gene silencing of hippocampal CB1 cannabinoid receptors with the lentiviral particles mix. CB1R-siRNA or sham CB1R-siRNA was injected into the hippocampal CA1 area through a chronically implanted cannula. Several functional properties of the CA3–CA1 synapse were analyzed 1 and 21 days after lentiviral injections (Figs 7–9). We first determined the spread of the virus within the hippocampus in the CB1R-siRNA–injected mice as well as in the sham-injected mice using immunohistochemistry for the GFP protein and found that particles infected approximately 2 mm2 along the rostrocaudal axis, infecting most of the dorsal hippocampus, including the pyramidal cell layer and dentate gyrus in both CB1R-siRNA and sham CB1R-siRNA mice (Fig. 7).
Figure 8 shows the evolution of fEPSPs evoked in the CA1 area by single stimuli presented at Schaffer collaterals at increasing intensities (0.02–0.3 mA) 24 h (black triangles) and 21 days (white triangles) after CB1R-siRNA injection. In both cases, fEPSP slopes increased progressively for higher intensities. fEPSP slope values reached 21 days after the injection were lower than those 1 day after it (1300% 1 day after injection; 917% 21 days after injection; Fig. 8C), although no significant difference was detected. When the sham CB1R-siRNA group was injected, the evolution of fEPSPs was similar 1 and 21 days after the intrahippocampal infusion, reaching values of ≈1400% of the baseline (Fig. 8D, black and white triangles). As a control of the efficiency of the CB1R-siRNA, we also injected, through the same implanted cannula, 1.5 μL of WIN55,212-2. Interestingly, the cannabinoid agonist had no effect when infused 21 days after siRNA injection (Fig. 8C, gray triangles) but still had the expected depressing effect in sham CB1R-siRNA animals (Fig. 8D, gray triangles). This inhibition of hippocampal synaptic strength in CB1R-siRNA sham-injected mice was more obvious at high intensities (asterisks in Fig. 8D, F42,228 = 1.834, P < 0.005).
The effects of PP stimulation at the CA3–CA1 synapse were also studied in these 2 groups of animals (Fig. 8E,F). Six different time intervals (10, 20, 40, 100, 200, and 500 ms), at a fixed intensity, were tested 1 and 21 days after siRNA or sham siRNA injections. The PP ratio [(second/first) × 100] was not affected by the in vivo silencing of hippocampal CB1 receptors, since no significant difference was detected between CB1R-siRNA–injected mice and the sham group, either 1 or 21 days after intrahippocampal infusions (Fig. 8E,F; black and white squares).
LTP evoked at the CA3–CA1 synapse as described above was also affected by CB1R-siRNA injections. As shown in Figure 9, the LTP evoked in the CB1R-siRNA group by the first pulse was significantly smaller than that evoked in sham CB1R-siRNA animals (F11,99) = 1.857, P < 0.05). However, both groups of animals presented a significant increase in fEPSP slopes following HFS, as compared with baseline values (siRNA, F11,99 = 6.239, P < 0.001; sham, F11,99 = 8.901, P < 0.001). Moreover, fEPSP slopes evoked by the second pulse were always larger than those evoked by the first in both groups of animals (asterisks in Fig. 9B,D; siRNA, F11,99 = 1.739, P < 0.05; sham, F11,99 = 0.497, P < 0.05). Although these results suggest that the effects of the HFS reached a larger hippocampal area than that covered by the siRNA injection, the LTP evoked in the CB1R-siRNA group was significantly smaller than that evoked in sham CB1R-siRNA animals.
Classical Eyeblink Conditioning and Hippocampal Synaptic Plasticity in Mice Lacking the CB1 Receptor
Mice lacking the cannabinoid CB1 receptor (CB1−/−) and their corresponding wild-type littermates (CB1+/+) were classically conditioned with the trace paradigm described above. CB1+/+ mice learnt the paradigm (black circles, Fig. 10B) as already described in other control groups (Fig. 5D), reaching asymptotic values of CRs from the 7th to the 10th conditioning sessions (≈67% of CRs). In contrast, CB1−/− mice were unable to learn the paradigm as well as CB1+/+ mice did. The maximum percentage of CRs reached by the CB1−/− group was 49% during the 10th session. Differences in the percentage of CRs between CB1+/+ and CB1−/− mice were statistically significant from the 7th to the 10th conditioning sessions (asterisks in Fig. 10B; F11,99 = 2.151, P < 0.05). As expected, the corresponding hippocampal synaptic changes underlying associative learning were seen in the CB1+/+ group (Fig. 10A, black triangles), whose fEPSP slopes increased across conditioning sessions until reaching ≈130% of control values by the 10th conditioning session. In contrast, CB1−/− mice presented an increase of only 16% (from 100% to 116%) in the fEPSP slopes from the 1st to the 10th conditioning sessions. The differences between CB1+/+ and CB1−/− were statistically different for the 9th and the 10th conditioning sessions (asterisks in Fig. 10A; F11,99 = 0.842, P ≤ 0.05).
We next characterized the functional properties of the CA3–CA1 synapse in CB1−/− mice and in the corresponding CB1+/+ group (Fig. 11). Figure 11A displays the evolution of fEPSPs evoked by single stimuli at Schaffer collaterals at increasing intensities (0.02–0.3 mA). At low intensities (0.02–0.16 mA), values of the fEPSP slopes for CB1+/+ and CB1−/− animals were similar (Fig. 11A; CB1+/+, black circles; CB1−/−, white circles). However, for higher intensities (>0.16 mA), fEPSP slopes for the CB1+/+ group continued increasing, reaching values up to 2400% of baseline records (at 0.3 mA of stimulating intensity), while fEPSPs evoked in CB1−/− animals reached values well below those collected from the CB1+/+ group (≈1300% of baseline at 0.3 mA; Fig. 11A, white circles). These differences between groups were significant in the range 0.24–0.3 mA (asterisks in Fig. 11A; F14,126 = 2.750, P = 0.001). When pairs of pulses were presented at different interstimulus intervals to Schaffer collaterals of CB1+/+ and CB1−/− mice (Fig. 11B), the PP ratio between the second and the first evoked fEPSP was decreased in CB1−/− mice at short intervals (10, 20, and 40 ms). At an interval of 40 ms, significant differences were found between the 2 groups (asterisks in Fig. 11B; F5,45 = 1.790, P < 0.05). These results indicated that PP facilitation was depressed in CB1−/− mice as compared with their littermate controls.
Finally, an LTP study was carried out in these 2 groups (CB1+/+ and CB1−/−) of animals (Fig. 11C,D). Pairs of stimuli (40 ms interval) were presented at Schaffer collaterals during 15 min in order to obtain a baseline for both the first and the second evoked fEPSPs. Then, mice received an HFS session, which induced a noticeable increase in the slopes of both the first (black triangles) and the second (black squares) fEPSPs in CB1+/+ mice (Fig. 11C; F11,99 = 16.650, P < 0.001). This increase remained significant 24 h after HFS. In contrast, CB1−/− mice (Fig. 11D) presented a small, but significant, increase (F11,99 = 1.947, P = 0.042) in fEPSP slopes (first pulse, white triangles; second pulse, white squares) after the HFS session, although during the 30 min following HFS presentation and 24 h later it was significantly smaller (F11,99 = 5.210, P < 0.05) than the potentiation induced in wild-type mice. In summary, input–output curves, PP facilitation, and LTP were significantly depressed in CB1−/− animals as compared with their littermate controls.
According to the present results, the administration of a CB1 agonist decreased the acquisition of an associative learning task and the concomitant increase in strength of the CA3–CA1 synapse, as well as LTP evoked at the same hippocampal synapses, in alert behaving mice. Unexpectedly, CB1−/− mice also showed a decrease in the percentage of CRs, a diminished potentiation of the CA3–CA1 synapse across training, and a lower LTP. As discussed below, those effects can be ascribed only partially to compensatory mechanisms because similar results were collected from CB1R-siRNA–injected animals (see Murray et al. 2007). As proposed previously (Abush and Akirav 2010), our findings suggest diverse effects of the cannabinoid system on hippocampal CA1 pyramidal cells with regard to activity-dependent synaptic plasticity that cannot be categorized as just impairing or enhancing effects of cannabinoid receptors. These points are discussed in detail below.
The hippocampus is strongly involved in learning and memory (Woody 1986; Sunayashiki-Kusuzaki et al. 1993; Zola-Morgan and Squire 1993; Bramham 2007; Neves et al. 2008; Sacktor 2008) and contains one of the highest CB1 receptor densities (Herkenham et al. 1990, 1991). Although the cannabinoid system has been widely studied, its role in hippocampal-dependent functions is still a matter of discussion. For example, numerous studies have shown that the administration of CB1 agonists impairs both memory and hippocampal synaptic plasticity (Lichtman et al. 1995; Hampson and Deadwyler 1999; Hernández-Tristán et al. 2000; Davies et al. 2002; Puighermanal et al. 2009), whereas antagonists have been reported to improve them (Terranova et al. 1996; Lichtman 2000; Takahashi et al. 2005; Wolff and Leander 2003), to have no effect (Da Silva and Takahashi 2002; Davies et al. 2002) or even to cause memory deficit and block LTP induction (de Oliveira Alvares et al. 2005, 2006). In an interesting series of experiments, it has been shown that cannabinoids play an important role in spike-timing coordination (Robbe et al. 2006), fire/bursting of hippocampal principal cells (Goonawardena et al. 2011), and in theta regulation of pyramidal firing (Robbe and Buzsáki 2009), 3 processes directly involved in spatial navigation and in hippocampal-mediated memory functions. Nevertheless, those studies did not address 2 important questions. First, most of them were focused on spatial learning tasks, such as the radial and water mazes, but little is known about the effects of cannabinoids on other complex hippocampal-dependent learning forms such as associative learning tasks. Second, synaptic plasticity mediated by the cannabinoid system has been widely studied in hippocampal slices, and it has been assumed that the in vivo system will simply follow the same pattern. In accordance with results previously described for other learning tests (Lichtman et al. 1995), our results indicate that the systemic activation of CB1 receptors by the cannabinoid agonist WIN55,212-2 impairs the acquisition of a trace eyeblink conditioning paradigm and that this effect is antagonized by pretreatment with the cannabinoid antagonist AM251. Importantly, the chronic electrodes used here allowed us to observe the correlation between cannabinoid receptor activation, impaired learning, and blocked hippocampal synaptic plasticity in alert behaving animals.
However, experiments in which the agonist is systemically injected do not specifically show that cannabinoids impair learning and memory via a direct action on hippocampal circuits. It has been shown previously that the spatial memory–impairing effects of systemic cannabinoid administration can be reproduced by intrahippocampal administration of cannabinoid receptor agonist (Lichtman et al. 1995). Our results indicate that the local hippocampal activation of CB1 receptors is able to decrease both the percentage of CRs and the concomitant changes in synaptic strength taking place at the hippocampal CA3–CA1 synapse. Taken together, these findings suggest that a selective disruption of hippocampal CB1 receptors is contributing significantly to cannabinoid-mediated impairment of performance of these associative learning tasks. Importantly, our results also suggest that at least part of the learning impairment produced by cannabinoids is mediated by nonhippocampal CB1 receptors. Since the systemic activation of CB1 cannabinoid receptors caused a larger decrease in the percentage of CRs than the local activation did, the deleterious effects of cannabinoids on associative learning are not mediated solely by hippocampal CB1 receptors. In fact, systemically administered cannabinoid receptor agonists produce additional sensorimotor, physiological, and subjective effects (Järbe and McMillan 1980; Little et al. 1988) that could be involved in the decrease in the percentage of CRs. In this regard, it has been reported that intrahippocampal injections of CB1 agonists failed to produce these effects (Lichtman et al. 1995).
As previously described for hippocampal slices (Terranova et al. 1995; Davies et al. 2002), results presented here demonstrate that cannabinoid agonists applied s.c. also inhibit the in vivo induction of LTP in the CA1 area of the hippocampus. Although this inhibition is detected after both the systemic and the local activation of cannabinoid CB1 receptors (Figs 3 and 4), the inhibition is more pronounced when the drug is injected directly into the hippocampus. This difference can be explained considering that cannabinoid agonists are highly lipophilic and have a poor penetration into brain tissue (Bajo et al. 2009). It is widely accepted that the mechanism by which cannabinoids impair synaptic transmission is not through any direct effect on the molecular machinery underlying this process but by reducing neurotransmitter release from presynaptic terminals. Because hippocampal LTP requires depolarization of the postsynaptic membrane to relieve magnesium blockade of N-methyl-d-aspartate (NMDA) receptors and allow entry of calcium, a reduction in neurotransmitter release could impair long-term synaptic plasticity by failing to depolarize the postsynaptic CA1 membrane to a level that relieves magnesium blockage (Misner and Sullivan 1999). The specific mechanism by which the cannabinoid agonist WIN55,212-2 reduces glutamate release in the hippocampus remains unknown, although 2 processes have been suggested: the first involves a G-protein–mediated inhibition of presynaptic N- and P/Q-type calcium channels and is probably the primary molecular mechanism, but modulation of proteins involved in vesicle release may also have a role (Sullivan 2000; Puighermanal et al. 2009). The variation in the PP ratio evoked by WIN55,212-2 administration in input/output curves and LTP suggests a definite involvement of CB1 receptors in presynaptic processes underlying learning and memory. In this regard, it has been recently proposed that presynaptic mechanisms play an important role in both LTP and learning-dependent synaptic changes in strength (Madroñal et al. 2009). As a further support of this contention, it has been shown that CB1 receptors are located on presynaptic terminals of forebrain principal neurons (Domenici et al. 2006) and that regulate glutamate release from Schaffer collaterals (Németh et al. 2008), probably by interaction with adenosine A1 receptors (Hoffman et al. 2010; Sousa et al. 2011).
It has been suggested that the mechanism by which cannabinoids inhibit the induction of LTP is also the mechanism by which they impair learning and memory (see Sullivan 2000). In this regard, we have demonstrated here that CB1 receptor activation also inhibits another form of long-lasting synaptic plasticity—the activity-dependent synaptic potentiation that takes place at the CA3–CA1 synapse during associative learning. The present results further support the contention that these 2 forms of long-term synaptic potentiation share a cellular mechanism mediated by cannabinoid CB1 receptors.
Although it is clear that CB1 receptors are expressed at an especially high density in the dentate gyrus, CA1, and CA3 regions of the hippocampus (Matsuda et al. 1990; Tsou et al. 1998), there is currently controversy on whether CB1 receptors can be found in excitatory pyramidal neurons (Pettit et al. 1998) or are instead located exclusively in the presynaptic terminals of inhibitory interneurons (Tsou et al. 1999). Although only a few in vitro studies have provided strong evidence for the presence of CB1 receptors in presynaptic glutamatergic terminals (Pettit et al. 1998; Katona et al. 2006), the fact that local hippocampal injection of the cannabinoid agonist WIN55,212-2 suddenly blocks synaptic transmission at the CA3–CA1 synapse—that is, at a glutamatergic synapse—provides in vivo evidence for the existence of CB1 receptors at excitatory terminals.
Initial studies using mice lacking the cannabinoid CB1 receptor reported enhanced memory retention in the object-recognition task (Reibaud et al. 1999). However, later studies have shown that the deletion of the CB1 receptor gene causes a deficit in reversal learning in the water maze test (Varvel and Lichtman 2002) and a delayed extinction learning in a fear-conditioning paradigm (Marsicano et al. 2002; Martin et al. 2002). Interestingly, our results indicate that CB1-deficient mice present a clear learning impairment for acquiring a classical eyeblink conditioning using a trace paradigm. One plausible explanation for the impaired performance of this associative learning task is the acceleration in the decline of cognitive functions described in the absence of CB1 receptors (Bilkei-Gorzo et al. 2005). In this regard, young (6/8 weeks old) CB1-deficient mice show enhanced learning compared with their corresponding wild-type littermates, but mature (3/5 months old) mice present important learning deficits and perform similarly to old (14/17 months old) mice in various learning tasks. Since the mice used in our experiments were around 4 months old at the beginning of the conditioning sessions, we can assume that their learning abilities were already deteriorated. Furthermore, this accelerated decline in learning performance in mature CB1−/− mice seems to be related with a significant reduction of the neural density in both CA1 and CA3 areas (Bilkei-Gorzo et al. 2005). Although it has been reported that LTP is facilitated in slices obtained from young CB1 knockout mice (Böhme et al. 2000), our results indicate that in the absence of CB1 receptors, the in vivo LTP induction in mature mice is blocked, as well as the learning-dependent synaptic changes observed at the CA3–CA1 synapse. This blocked synaptic transmission is presumably due to the loss of hippocampal neurons, a phenomenon not detected in young mice (Bilkei-Gorzo et al. 2005).
Finally, results obtained with CB1−/− mice prompted us to fine-tune this study, silencing exclusively CB1 receptors in the CA1 region of the hippocampus. In accordance with results previously obtained in mice lacking the endogenous receptor, the local silencing of hippocampal CB1 receptors also blocked the induction of LTP and the acquisition of conditioned eyeblink responses, indicating that a local alteration in the hippocampal endocannabinoid system is sufficient to impair both learning and activity-dependent synaptic plasticity. As a whole, the present results further confirm the involvement of hippocampal CB1 receptors in learning and memory processes and reinforces the putative role of presynaptic mechanisms (Madroñal et al. 2009) in activity-dependent changes in synaptic strength.
(MICINN-BFU2008-0899 to J.M.D.G., MICINN-BFU2008-03390 to A.G., MICINN-BFU2010-20664); Plan Nacional sobre Drogas from the Spanish Ministerio de Sanidad y Política Social to R.M.
We would like to thank Mr Roger Churchill for his editorial help. Conflict of Interest: None declared.