Δ⁹-Tetrahydrocannabinol (THC), the active agent in the medication dronabinol and the principal psychoactive ingredient in marijuana, exerts its effects on the central nervous system via cannabinoid type 1 receptors (CB1Rs) (Chaperon and Thiebot, 1999). CB1Rs are G protein-coupled receptors (GPCRs) that activate K⁺ channels and mitogen-activated protein kinases and inhibit adenylyl cyclase and voltage-gated Ca²⁺ channels (VGCCs) (Howlett, 2005). The short-term inhibition of synaptic transmission by cannabinoids is primarily mediated by inhibition of presynaptic VGCCs (Brown et al., 2004). Endocannabinoids (eCBs) produced by postsynaptic neurons diffuse in the retrograde direction where they, too, act on presynaptic CB1Rs to inhibit excitatory and inhibitory synaptic transmission (Lovinger, 2008). The effects of THC on synaptic transmission and eCB signaling are not entirely explained by it simply mimicking eCBs.

THC is a partial agonist of CB1Rs, as indicated by weak stimulation of guanosine 5′-O-(3-thio)triphosphate binding in rodent brain (Breivogel and Childers, 2000) and modest inhibition of excitatory postsynaptic currents (EPSCs) (Shen and Thayer, 1999). The CB1R antagonist rimonabant completely blocked THC-mediated inhibition of glutamatergic synaptic transmission. THC produces a submaximal response even at concentrations that saturate CB1Rs; thus, THC will attenuate the actions of full agonists, including eCBs (Kelley and Thayer, 2004). In addition to its low intrinsic activity, THC is highly lipophilic, which imbues it with properties incompatible with many experimental techniques, including poor washout and insufficient penetration of brain slices (Lundberg et al., 2005; Lovinger, 2008). Thus, more water-soluble full CB1 agonists such as Win55212-2 are widely used for in vitro studies (Eissenstat et al., 1990). However, the recreational and clinical use of THC provides a compelling reason to study this compound specifically. In the few studies to examine the short-term effects of THC on synaptic transmission, the magnitude of the responses varied widely (Shen and Thayer, 1999; Straiker and Mackie, 2005). The low intrinsic activity of THC confers a high sensitivity to CB1R density (Selley et al., 2001). However, there is another important aspect to cannabinoid action that might explain the varied effects reported for THC.

Increases in firing rate will overcome the inhibition of neurotransmitter release produced by activation of presynaptic GPCRs (Brenowitz et al., 1998; Frerking and Ohliger-Frerking, 2006). Indeed, the presynaptic inhibition produced by Win55212-2 is attenuated at high firing rates (>20 Hz) (Foldy et al., 2006). Because THC is widely used in humans, has weak and variable affects on synaptic transmission, and has potentially complex interactions with the eCB signaling system, we examined the effects of firing rate on the presynaptic inhibition produced by THC.

Here, we tested the hypothesis that increases in presynaptic firing rate would attenuate THC-mediated inhibition of excitatory synaptic transmission and that at high firing rates, THC would antagonize eCB signaling. Our data indicate that modest increases in stimulus frequency had a profound effect on THC-mediated effects. At low firing rates, THC exhibited classic agonist properties. In contrast, at elevated firing rates, THC occupied CB1Rs but did not effectively couple to VGCCs, and thus, THC acted as an antagonist. This state-dependent switching from agonist to antagonist may account for the complex actions of THC in vivo.

Materials and Methods

Materials. Dulbecco's modified Eagle's medium (DMEM) and sera were purchased from Invitrogen (Carlsbad, CA). THC was obtained from the National Institute on Drug Abuse (Research Triangle Institute, Research Triangle Park, NC). Bicuculline methochloride and 2-arachidonylglycerol were purchased from Tocris Cookson (Ellisville, MO). All other reagents were purchased from Sigma (St. Louis, MO).

Cell Culture. Rat hippocampal neurons were grown in primary culture as described previously (Pottorf et al., 2006) with minor modifications. Fetuses were removed on embryonic day 17 from maternal rats, anesthetized with CO₂, and killed by decapitation under a protocol approved by the University of Minnesota Institutional Animal Care and Use Committee in accordance with the National Institutes of Health guide for the care and use of laboratory animals. Hippocampi were dissected and placed in Ca²⁺- and Mg²⁺-free HEPES-buffered Hanks' salt solution, pH 7.45. HEPES-buffered Hanks' salt solution was composed of the following: 20 mM HEPES, 137 mM NaCl, 1.3 mM CaCl₂, 0.4 mM MgSO₄, 0.5 mM MgCl₂, 5.0 mM KCl, 0.4 mM KH₂PO₄, 0.6 mM Na₂HPO₄, 3.0 mM NaHCO₃, and 5.6 mM glucose. Cells were dissociated by trituration through a series of flame-narrowed Pasteur pipettes, pelleted, and resuspended in DMEM without glutamine, supplemented with 10% fetal bovine serum and penicillin/streptomycin (100 U/ml and 100 μg/ml, respectively). Dissociated cells then were plated at a density of 10,000 to 15,000 cells/dish onto a 25-mm round cover glass precoated with matrigel (250 μl, 0.1 mg/ml). Neurons were grown in a humidified atmosphere of 10% CO₂ and 90% air at 37°C and fed on days 1 and 6 by exchange of 75% of the media with DMEM supplemented with 10% horse serum and penicillin/streptomycin. Cells used in these experiments were cultured without mitotic inhibitors for a minimum of 12 days.

EPSC Recordings. 6-Cyano-2,3-dihydroxy-7-nitroquinoxaline-sensitive EPSCs were recorded using the whole-cell configuration of the patch-clamp technique (Kouznetsova et al., 2002). Pipettes (Narishige, Greenvale, NY) with open resistances of 3 to 5 MΩ were filled with solution that contained 120 mM potassium gluconate, 15 mM KCl, 6 mM MgCl₂, 0.2 mM EGTA, 10 mM HEPES, and 5 mM Na₂ATP, adjusted to pH 7.3 with KOH and to 290 mOsm/kg with sucrose. Recordings were performed at room temperature (22°C) in an extracellular solution that contained 140 mM NaCl, 5 mM KCl, 9 mM CaCl₂, 6 mM MgCl₂, 5 mM glucose, 10 mM HEPES, and 0.01 mM bicuculline methochloride, adjusted to pH 7.4 with NaOH and to 325 mOsm/kg with sucrose. Solutions were applied by a gravity-fed superfusion system. Membrane potential was held at -70 mV, and monosynaptic EPSCs were evoked with a bipolar platinum electrode (FHC Inc., Bowdoinham, ME) placed near a presynaptic neuron. Voltage pulses (0.1 ms) were applied at a fixed rate of either 0.1 or 0.5 Hz using a Grass S44 stimulator with a SIU-5 stimulus isolation unit (Astro-Med Inc., West Warwick, RI). For depolarization-induced suppression of excitation (DSE) recordings, EPSCs were evoked at 0.5 Hz. DSE was elicited by depolarizing the postsynaptic cell to 0 mV for 15 s followed by continued recording at 0.5 Hz. A second DSE was elicited 5 min after the first response.

VGCC Recordings. Whole-cell I_Ca recordings were performed using pipettes filled with 145 mM CsMeSO₄, 10 mM HEPES, 10 mM BAPTA, 5 mM MgATP, and 1 mM Na₂GTP, adjusted to pH 7.35 with CsOH and to 315 mOsm/kg with sucrose. Seals were formed in 140 mM NaCl, 5 mM KCl, 9 mM CaCl₂, 6 mM MgCl₂, 5 mM glucose, and 10 mM HEPES, adjusted to pH 7.4 with NaOH and to 325 mOsm/kg with sucrose and then switched to buffer to isolate I_Ca that contained 143 mM tetraethylammonium chloride, 5 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, 10 mM glucose, and 0.1% bovine serum albumin, adjusted to pH 7.4 with tetraethylammonium OH and to 325 mOsm/kg with sucrose. Series resistance was compensated by a minimum of 75%. Membrane potential was held at -80 mV, and currents were evoked by stepping to 0 mV for 40 ms every 2 or 10 s (0.5 or 0.1 Hz) as noted. Recordings were not corrected for leak because there was essentially no current at the 0-mV test potential in the presence of 200 μM Cd²⁺. For tail current recordings, membrane potential was stepped in 10-mV intervals from -40 to +70 mV for 20 ms followed by repolarization to -40 mV for 20 ms with the resultant tail current peak amplitude analyzed after subtracting leak current measured in 200 μM Cd²⁺. Voltage-dependent reversal of cannabinoid inhibition of I_Ca was tested by comparing a control current evoked by stepping from -80 to 0 mV for 40 ms to a current evoked 500 ms later after a 40-ms prepulse to either 0 or +80 mV. Voltage-sensitivity of cannabinoid inhibition was also assayed using paired 20-ms steps from -80 mV to 0 separated by a 10-ms return to -80 mV.

Data Acquisition and Analysis. Currents were amplified using an Axopatch 200A; for EPSC and I_Ca recordings, respectively, data were filtered at 2 and 1 kHz and digitized at 11 and 5 kHz with a Digidata interface controlled by pClamp software (MDS Analytical Technologies, Toronto, ON, Canada). Tail current and voltage-sensitivity recordings were digitized at 50 kHz and filtered at 10 kHz. Leak currents were digitally subtracted from corresponding tail currents using Clampfit 9.0 and curves fit using Origin 6.0 (OriginLab Corp., Northampton, MA). Access resistance and leak currents were monitored continuously, and the recording was excluded if either changed significantly. In EPSC experiments, sweeps preceded or followed by spontaneous synaptic currents were excluded from analysis. To calculate the percentage of inhibition of EPSC and I_Ca amplitudes, the mean peak current from the sweeps collected during the 1-min preceding drug application was compared with the average peak current during the final minute of drug treatment, the time at which inhibition was maximal. I_Ca elicited at 0.5 Hz underwent significant rundown that was well described by a monoexponential decay function (r² = 0.99). Thus, to calculate changes in current amplitude, rundown was determined by fitting an exponential equation to the 3-min preceding drug application; control current amplitude (0% inhibition) was defined as the value extrapolated to the time at which drug inhibition was calculated.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Increased stimulus rate blocks THC inhibition of EPSC amplitude. A bipolar concentric electrode placed near a presynaptic neuron was used to evoke EPSCs in a postsynaptic cell held at -70 mV in whole-cell voltage clamp. Representative plots show peak EPSC amplitude versus time. EPSCs were evoked every 10 s (0.1 Hz) (A) or every 2 s (0.5 Hz) (B). 1 μM THC was applied during the time indicated by the black horizontal bars. Gray bars indicate the time and amplitude of the corresponding averaged EPSCs displayed in the insets, denoted as a or b. Insets, vertical bar represents 100 pA, and horizontal bar represents 20 ms.

All data are presented as mean ± S.E. Significance was determined using Student's t test or analysis of variance with Bonferroni post test for multiple comparisons.

Results

THC Inhibition of EPSCs Is Modulated by Stimulus Frequency. The effects of CB1R agonists were studied on excitatory synaptic transmission between rat hippocampal neurons in culture. Synaptic currents were recorded from a postsynaptic cell held at -70 mV in the whole-cell configuration of the patch clamp. Stimulation of the presynaptic neuron with an extracellular electrode at a rate of 0.1 Hz evoked reproducible EPSCs. Application of 1 μM THC inhibited EPSC amplitude by 43 ± 10% (n = 7) (Fig. 1A). This level of inhibition by a maximally effective concentration of THC is consistent with the partial agonist properties of THC acting on CB1Rs (Shen and Thayer, 1999). Five-minute pretreatment with 300 nM AM251, a CB1 antagonist, reduced THC-induced inhibition of EPSC amplitude to 5 ± 4% (n = 3, p < 0.01), consistent with THC acting on CB1 receptors. It is noteworthy that increasing the stimulation rate to 0.5 Hz reduced the inhibition produced by THC to only 15 ± 2% (n = 6), which was significantly different from that seen at the 0.1-Hz stimulus rate (Fig. 1B, p < 0.01). Because of the inherent lipophilicity of THC, effective washout from the preparation was not possible, and thus THC was irreversible over the time course studied here. We next examined the frequency-dependence of EPSC inhibition by the full agonists Win55212-2 (100 nM) and 2-arachidonylglycerol (2-AG) (1 μM) (Fig. 2). In contrast to THC, Win55212-2 inhibition of EPSC peak amplitude was not different at 0.1 Hz versus 0.5 Hz (Fig. 2, A and B), as indicated by a 69 ± 4 and 58 ± 7% inhibition of peak current, respectively (n = 6). The inhibition produced by the eCB agonist 2-AG was also similar at 0.1 and 0.5 Hz, producing 36 ± 9 and 40 ± 15% inhibition, respectively (n = 6). If THC were occupying all of the CB1Rs as would be expected for a maximally effective concentration of a partial agonist, then THC might be expected to antagonize the effects of a full agonist (Shen and Thayer, 1999; Kelley and Thayer, 2004; Straiker and Mackie, 2005). Accordingly, when applied in the continued presence of THC, Win55212-2 failed to affect EPSC amplitude (8 ± 6%, n = 6, p < 0.001; Fig. 3, A and B). Thus, THC blocked the inhibition normally produced by 100 nM Win55212-2, even under conditions in which THC had little effect by itself. These data are consistent with the hypothesis that agonists with low intrinsic activity are particularly sensitive to attenuated presynaptic inhibition by increases in the stimulation rate.

THC Inhibition of I_Ca Depends on Depolarization Rate. Activation of presynaptic CB1Rs inhibits N and P/Q type VGCCs (Lovinger, 2008), resulting in the attenuation of the evoked release of neurotransmitter. Because THC prevented the inhibition of evoked EPSCs by full agonists, we postulated that the reduced inhibition of EPSC amplitude by THC at 0.5 Hz was due to impaired inhibitory coupling to VGCCs. To test this hypothesis we elicited whole-cell I_Ca from hippocampal neurons at frequencies of 0.1 and 0.5 Hz. I_Ca was evoked from a holding potential of -80 mV by a 40-ms depolarizing step to 0 mV (Fig. 4). Win55212-2 (100 nM) inhibited whole-cell I_Ca evoked at 0.1 Hz by 51 ± 6% (n = 5). Stimulation at 0.5 Hz produced significant Ca²⁺ current rundown that in control recordings was well described by an exponential function (r² = 0.99, n = 4). Thus, to correct for Ca²⁺ current rundown in drug studies, 3 min of baseline preceding the addition of drug was fit to a single exponential equation and extrapolated to the time at which drug inhibition was calculated (Fig. 4, B and D). Win55212-2 (100 nM) inhibited I_Ca evoked at 0.5 Hz by 46 ± 4% (n = 5) (Fig. 4, B and E). THC (1 μM) inhibited whole-cell I_Ca elicited at 0.1 Hz by 26 ± 3% (n = 5). (Fig. 4, C and E). Similar results have been demonstrated in CB1R-expressing neurons using common eCBs, including 2-AG (Guo and Ikeda, 2004). However, when the currents were evoked at 0.5 Hz, THC had no effect (n = 4) (Fig. 4, D and E). Thus, THC inhibition of I_Ca was highly sensitive to stimulus frequency (p < 0.01).

THC Inhibition of VGCCs Is More Sensitive to Voltage than Inhibition Produced by Win55212-2. Previous studies have demonstrated that inhibition of VGCCs by Gβγ is voltage-dependent (Bean, 1989; Ikeda, 1996; Agler et al., 2003). We hypothesized that inhibition of I_Ca by THC would be more easily reversed by depolarization than by inhibition mediated by Win55212-2. The voltage-dependent activation of I_Ca was determined in the absence (control) or presence of Win55212-2 (100 nM) or THC (1 μM; Fig. 5, A and B). Tail currents were recorded at a holding potential of -40 mV after depolarizing test pulses from -40 to +70 mV. Activation curves were well described by a single Boltzmann function: I/I_max = 1/[1 + exp{(V_0.5 - V_m)/k}]. V_m is defined as the activating potential, k is the slope factor, and V_0.5 is the half-maximum activation potential. Under control conditions, V_0.5 was -3 ± 1 mV (n = 6). Win55212-2 (100 nM) caused a positive shift in the voltage-dependence of activation (V_0.5 = 8 ± 1 mV; n = 3), whereas activation kinetics in the presence of THC (1 μM) were comparable with control (V_0.5 = -3 ± 1 mV; n = 5). To determine whether this difference resulted from an increased sensitivity of THC inhibition of I_Ca to depolarization relative to that produced by Win55212-2, we studied the effects of depolarizing prepulses on drug-induced inhibition of I_Ca (Fig. 5, C-E). A conditioning prepulse to 0 mV facilitated I_Ca in the presence of THC, whereas the facilitation ratio in the presence of Win55212-2 was comparable with that of control (Fig. 5, C-D). The ratio of the peak amplitude of the first I_Ca to the second I_Ca (after prepulse) equaled 0.73 ± 0.01 (n = 27) under control conditions, 0.75 ± 0.03 in the presence of Win55212-2 (100 nM; n = 12), and 0.81 ± 0.02 in the presence of THC (1 μM; n = 6). A conditioning prepulse to +80 mV relieved both THC and Win55212-2-mediated suppression of I_Ca (Fig. 5, C and D). The facilitation ratio for control experiments equaled 1.11 ± 0.03 (n = 35) versus 1.67 ± 0.08 in the presence of Win55212-2 (100 nM; n = 12) and 1.35 ± 0.10 in the presence of THC (1 μM; n = 6). A second protocol using paired depolarizations to 0 mV verified that THC-induced inhibition of I_Ca is more sensitive to mild depolarization than Win55212-2 (Fig. 5E). The facilitation ratio for control (n = 41) and Win55212-2 (100 nM, n = 12)-treated currents were 0.83 ± 0.01 and 0.83 ± 0.02, respectively, and the ratio for THC (1 μM; n = 10) was 0.96 ± 0.02. Thus, THC-mediated inhibition of I_Ca is more sensitive to positive shifts in membrane potential than inhibition produced by Win55212-2.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Win55212-2 (Win-2) and 2-arachidonylglycerol (2-AG) inhibit EPSC amplitude similarly at 0.1 and 0.5 Hz stimulus rates. Representative plots show peak EPSC amplitude versus time. EPSCs were evoked every 10 s (0.1 Hz) (A and C) or every 2 s (0.5 Hz) (B and D). Win-2 and 2-AG were applied during the times indicated by the black horizontal bars. Gray bars indicate the time and amplitude of the corresponding averaged EPSCs displayed in the insets, denoted as a, b, or c. Insets, vertical bar represents 100 pA, and horizontal bar represents 20 ms.

THC Inhibits DSE. The frequency-dependence of THC-induced inhibition of I_Ca and EPSCs suggests that THC may act as an agonist or antagonist depending on the firing rate of the synapse. This frequency-dependent shift in efficacy could have profound effects on eCB signaling. To examine the effects of THC on CB1R activation by the retrograde actions of eCBs, we used the DSE protocol described previously for hippocampal neurons in culture (Ohno-Shosaku et al., 2002; Straiker and Mackie, 2005). In these experiments, EPSCs were evoked at 0.5 Hz, and endocannabinoid production was induced by depolarizing the postsynaptic neuron to 0 mV for 15 s. Immediately after depolarization, a transient DSE (DSE1) lasting 60 to 90 s was observed, with inhibition of EPSCs reaching 46 ± 8% (Fig. 6, A and D; n = 8). After 5 min of continued stimulation (0.5 Hz), a second depolarization to 0 mV was applied (DSE2) (Fig. 6, B and D). Under control conditions, DSE2 produced a 47 ± 10% inhibition of EPSC amplitude (n = 4), similar to DSE1. DSE was completely blocked by pretreatment (5 min) with 300 nM AM251 (data not shown). Similar to the actions of the antagonist, in synaptic pairs treated for 5 min with 1 μM THC, DSE2 was completely blocked (6 ± 6%; n = 4) (Fig. 6, C and D). THC significantly reduced DSE relative to control (p < 0.001).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

THC antagonizes Win55212-2 inhibition of EPSCs evoked at 0.5 Hz. A, representative plot shows peak EPSC amplitude versus time. THC (1 μM) and 100 nM Win55212-2 (Win-2) were applied during the times indicated by the black horizontal bars. Gray bars indicate the time and amplitude of the corresponding averaged EPSCs displayed in the inset, denoted as a, b, or c. Inset, vertical bar represents 100 pA, and horizontal bar represents 20 ms. B, bar graph shows the percentage of inhibition of EPSCs evoked at 0.1 (□) or 0.5 Hz (▪) treated with 100 nM Win55212-2 (Win-2), 1 μM 2-arachidonylglycerol (2-AG), 1 μM THC, or 300 nM AM251 as indicated. ^**, p < 0.01, 0.1 Hz versus 0.5 Hz; ††, p < 0.01, compared with THC at 0.1 Hz; ###, p < 0.001, compared with Win-2 at 0.5 Hz.

Discussion

The effects of THC on synaptic networks are complex and are not fully explained by agonist actions. In the current study, we show that the effects of THC on glutamatergic synapses between hippocampal neurons switch between agonist and antagonist actions on CB1Rs, depending on stimulus rate. When EPSCs were evoked at 0.1 Hz, THC acted as a weak agonist of CB1Rs to inhibit synaptic transmission. In contrast, when stimulated at 0.5 Hz, THC had little direct effect and actually acted as an antagonist capable of blocking the actions of exogenously applied and endogenously produced cannabinoids. Voltage-sensitive coupling of CB1Rs to VGCCs caused these state-dependent actions of THC. The extreme sensitivity of the actions of THC to firing frequency was not previously appreciated, and the concept that agonists of low intrinsic activity might act as antagonists during certain patterns of activity has broad implications for the many drugs that act on presynaptic GPCRs to inhibit neurotransmitter release.

Frequency-Dependent Coupling of CB1Rs to Presynaptic VGCCs. Activation of CB1Rs inhibits EPSCs and sensitive inhibitory postsynaptic currents (IPSCs) between hippocampal neurons (Shen et al., 1996; Hajos et al., 2000). In general, presynaptic inhibition mediated by GPCRs can be overcome by increases in firing rate (Brenowitz et al., 1998; Frerking and Ohliger-Frerking, 2006). Indeed, increasing the presynaptic firing rate to >20 Hz reversed Win55212-2 inhibition of IPSCs (Foldy et al., 2006). In the current study, inhibition of EPSCs by THC was completely blocked by a modest increase in presynaptic stimulus rate to 0.5 Hz. Thus, the inhibition mediated by THC is approximately 40 times more sensitive to the firing rate than that mediated by full CB1R agonists. We speculate that the high sensitivity of THC-mediated inhibition of synaptic transmission to firing frequency contributes to the variable effects reported for this drug in in vitro models. The only experiments to describe THC-mediated inhibition of synaptic transmission were performed at a stimulus rate of 0.1 Hz or less (Pertwee et al., 1996; Shen and Thayer, 1999; Azad et al., 2008) or used intermittent burst-type stimulus protocols with prolonged interstimulus intervals (Pertwee et al., 1992). It is noteworthy that a partial agonist at 5-hydroxytryptamine-3 receptors was shown previously to display a frequency-dependent reduction in the efficacy for inhibition of neurotransmitter release (Van der Vliet et al., 1988), suggesting that a general property of presynaptic inhibition produced by agonists of low intrinsic activity may be an especially high sensitivity to firing rate.

The short-term effects of cannabinoid agonists on synaptic transmission are primarily mediated by the inhibition of presynaptic VGCCs (Lovinger, 2008). Because the relationship of presynaptic [Ca²⁺]_i to vesicular release is a power function, small changes in Ca²⁺ influx have large effects on neurotransmission. CB1Rs couple to VGCCs via the βγ subunits of inhibitory G-proteins (Agler et al., 2003). We found that an increased firing rate reduced THC inhibition of whole-cell I_Ca through VGCCs. Frequency-dependent relief of VGCC channel inhibition has been attributed to the accumulation of residual Ca²⁺ in the presynaptic terminal at high frequencies (Kreitzer and Regehr, 2000) and might contribute to the frequency-dependence of THC inhibition of EPSCs but would not account for the frequency-dependent modulation of VGCCs. The most likely explanation for the reduction in CB1R coupling to VGCCs is the relief of G-protein-mediated inhibition that occurs during repetitive physiological stimuli (Brody et al., 1997). This attenuation of GPCR-mediated inhibition of VGCCs results from voltage-dependent reversal of Gβγ coupling to N- and P/Q-type Ca²⁺ channels (Bean, 1989; Ikeda, 1996; Agler et al., 2003). We observed a positive shift in the activation curve for I_Ca in the presence of Win55212-2, consistent with a voltage-sensitive coupling between CB1R and VGCC. In the presence of THC, I_Ca was facilitated by a small depolarizing prepulse, indicating that THC inhibition of VGCC was even more sensitive to voltage than that produced by Win55212-2. Thus, the inefficient release of βγ subunits evoked by THC (Breivogel and Childers, 2000) relative to full agonists seems to make THC-mediated synaptic inhibition extremely sensitive to voltage.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

THC inhibition of VGCC is dependent on stimulus rate. A to D, representative plots show peak I_Ca versus time. I_Ca were evoked by a depolarizing step from -80 to 0 mV for 40 ms at either 0.1 Hz (A and C) or 0.5 Hz (B and D). Win55212-2 (Win-2; 100 nM) or 1 μM THC was applied during the times indicated by the black horizontal bars. For recordings in which the depolarizing stimulus was applied at 0.5 Hz (B and D), current rundown was determined by fitting an exponential equation to the 3 min preceding drug application; control current amplitude (0% inhibition) was defined as the value extrapolated to the time at which drug inhibition was calculated. Insets, mean current trace from the sweeps collected during the 1 min preceding drug application superimposed on the averaged current traces during the final minute of drug treatment. Inset, vertical bars represent 500 pA, and horizontal bars represent 20 ms. The small horizontal lines in the insets to B and D represent the extrapolated curve fit during drug treatment. E, bar graph shows the inhibition of I_Ca evoked at 0.1 Hz (□) or 0.5 Hz (▪) by Win55212-2 (Win-2) and THC. ^**, p < 0.01 THC effects at 0.1 versus 0.5 Hz.

THC and the Endocannabinoid System. If the attenuation of THC-mediated synaptic inhibition results from a frequency-dependent reduction in βγ coupling to VGCCs, then the binding of THC to CB1Rs would not necessarily be affected by stimulus rate. Indeed, at the high (0.5 Hz) stimulus rate, THC clearly occupied the receptor, even though it failed to affect synaptic transmission, as indicated by block of Win55212-2-mediated inhibition of EPSC amplitude. This observation has significant implications for how THC interacts with the eCB system. We found that THC acted as an antagonist in DSE experiments, presumably because the DSE protocol requires fast stimulus rates to resolve the inhibition of EPSCs produced by the transient production of eCBs. Our results agree with those of Straiker and Mackie (2005) in that THC antagonized the eCB system. However, they did not observe an effect of THC at low stimulus frequency; presumably the combination of low receptor density and low agonist efficacy prevents effective CB1R coupling to VGCC. Long-term exposure to THC-desensitized CB1Rs, consistent with studies that found prolonged treatment with THC in vivo, produced a functional desensitization that impaired synaptic plasticity and prevented Win55212-2-mediated inhibition of IPSCs (Mato et al., 2004; Hoffman et al., 2007). The desensitization of CB1Rs, mediated by βγ activation of G-protein receptor kinases (Jin et al., 1999; Kouznetsova et al., 2002), would not be expected to display the frequency (voltage)-dependence described for coupling to VGCCs.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

THC inhibition of I_Ca is more sensitive to voltage than inhibition by Win55212-2. A, plots display I_Ca activation curves in the absence (control; open box) or presence of 1 μM THC (gray circle) or 100 nM Win55212-2 (Win-2; closed triangle). Voltage protocol is displayed in the inset. Normalized tail current amplitudes are plotted versus the voltage of a test prepulse, and the solid lines describe curves fit to the data with a single Boltzmann function: I/I_max = 1/[1 + exp{(V_0.5 - V_m)/k}]. Calculated values for V_0.5 and k were, respectively, -3 and 10 mV for control, -3 and 10 mV in THC, and 8 and 17 mV in Win55212-2. B, representative tail currents evoked from +10- and +70-mV prepulses are overlaid. Vertical bars represent 1 nA, and horizontal bars represent 1 ms. C, representative I_Ca resulting from prepulse voltage protocol (top). A prepulse to 0 mV facilitated I_Ca in the presence of THC. A prepulse to +80 mV facilitated I_Ca in the presence of both THC and Win-2. Vertical bars represent 200 pA, and horizontal bars represent 20 ms. D, bar graphs summarize the facilitation of I_Ca by a prepulse to either 0 or +80 mV as described in C. ^*, p < 0.05 THC versus control at 0 mV; ^***, p < 0.001 Win-2 versus control at +80 mV; †, p < 0.05 THC versus control at +80 mV. E, bar graph displays the ratio of I_Ca amplitudes evoked by paired depolarizations to 0 mV. Representative traces show currents for each respective condition. Vertical bars represent 200 pA, and horizontal bars represent 10 ms. ^***, p < 0.001, THC versus control; †††, p < 0.001, THC versus Win-2.

In contrast to presynaptic inhibition mediated by autoreceptors that respond to released neurotransmitter, THC modulates a receptor in which the endogenous ligand is distinct from that which mediates synaptic transmission. Thus, we envision at least four states under which THC will exert different effects on synaptic transmission. At low firing frequencies, THC will mimic the actions of eCBs to inhibit synaptic transmission. If eCBs are present, their actions would be occluded. In contrast, at high firing rates, THC will have little direct effect. When eCB production is stimulated and the synapse is firing at a high frequency, THC will block the actions of eCBs.

THC Affects Firing Patterns and Behavior. The inhibition of low-frequency EPSCs by THC (as shown in Fig. 1A) is predicted to remove the slow component of a complex firing pattern. The idea that the frequency-dependence of drug-induced presynaptic inhibition applies a high-pass filter to affected synapses has been suggested previously (Frerking and Ohliger-Frerking, 2006). For example, baclofen (a GABA_B receptor agonist) acted to enhance the contrast between low- and high-frequency field excitatory postsynaptic potentials during application of physiological spike trains taken from animals in a delayed nonmatched-to-sample behavioral paradigm. The drug preferentially inhibited field excitatory postsynaptic potentials that followed long inter spike intervals (ISI) at CA3 to CA1 synapses. The average firing rate of CA3 neurons was near 1 Hz, and the change in ISI near 0.1 Hz, close to the frequencies over which THC displayed dramatic changes in efficacy in the current study. It is noteworthy that cells exhibiting longer ISIs (near 0.1 Hz) tended to fire in response to specific cues and were inhibited by baclofen, whereas neurons that integrated multiple inputs had shorter ISIs (near 1 Hz) and were less sensitive to baclofen. Our data suggest that THC might also enhance the importance of high-frequency integrative inputs relative to more specific low-frequency inputs.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Inhibition of DSE by THC. A to C, EPSCs were evoked at 0.5 Hz. Plots show mean EPSC amplitudes, normalized to the 15 responses immediately before depolarization, plotted versus time. The postsynaptic cell was depolarized to 0 mV for 15 s at the times indicated by arrows. A, an initial control DSE (DSE1) was elicited in each recording. A second postsynaptic depolarization (DSE2) was evoked after 5-min treatment with vehicle (B) or 1 μM THC (C). D, bar graph shows the mean magnitude of DSE1 and DSE2 in the absence (□) and presence of 1 μM THC (▪). The percentage of DSE was calculated according to the following equation: % DSE = 100(EPSC_control - EPSC_DSE)/EPSC_control, where EPSC_control is the average amplitude of the 15 EPSCs immediately before depolarization and EPSC_DSE is the average amplitude of the two EPSCs immediately after depolarization. ^***, p < 0.001 DSE2 in the absence relative to the presence of THC.

THC has broad actions in the central nervous system, most of which are reversed by CB1R antagonists (Rinaldi-Carmona et al., 1994; Chaperon and Thiebot, 1999). However, some behavioral responses to THC suggest actions more complex than simple agonist effects. THC suppresses certain operant behaviors in a manner not effectively reversed by rimonabant, a CB1R antagonist (De Vry and Jentzsch, 2004; McMahon et al., 2005). The eCB system also participates in complex hippocampal functions, including memory extinction (Zhuang et al., 2005). THC retarded the extinction of an adverse associative memory (Ashton et al., 2008) in contrast to acceleration induced by Win55212-2 (Pamplona et al., 2006) and similar to the slowed extinction found in CB1R knockout animals (Marsicano et al., 2002). A human study found that THC actually enhanced a spatial working memory task for females and increased intrusion errors in spatial span tasks in both males and females (Makela et al., 2006). THC antagonism of eCB-mediated plasticity might underlie these effects, although linking synaptic transmission experiments to behavior is highly speculative.

Conclusions

We have shown that the effects of THC on neurotransmission depend on the firing rate of the synapse and the presence of eCBs. Because the actions of THC are more sensitive to the firing rate than highly efficacious cannabinoid agonists such as Win55212-2, we caution that results obtained in vitro with full agonists may not accurately reflect the more complex actions of THC in vivo. Going forward, it will be important to resolve the electrophysiological and behavioral consequences for presynaptic inhibition produced by agonists of differing intrinsic activities.