Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Activation of the cannabinoid receptor in cultured neurons inhibits adenylyl cyclase (Little and Martin, 1989; Bidaut-Russell et al., 1990) and enhances an outward potassium current in cultured hippocampal neurons (Deadwyler et al., 1993, 1995a; Hampson et al., 1995). The mechanism of cannabinoid receptor modulation of this potassium current has been previously described (Deadwyler et al., 1995b). This current has been identified as similar to A current (I_A) in other reports on the basis of activation and inactivation time constants (5 and 50 ms, respectively), steady-state inactivation and activation voltage ranges (V_1/2 = 70 to 75 and 20 to 15, respectively), refractory period (~200 ms), insensitivity to tetraethylammonium (TEA; 25 mM), and sensitivity to high concentrations of 4-aminopyridine (4-AP; 5 mM; Storm, 1990). Several G_i/o protein-linked receptor systems, such as -aminobutyric acid_B (Gage, 1992), serotonin_1A (Segal, 1980), and adenosine A₁ (Mei et al., 1995), enhance I_A by producing a positive shift in steady-state inactivation of the channel, resulting in fewer inactivated channels and hence increased current at membrane potentials between 80 and 50 mV. The cannabinoid receptor agonist WIN 55,212-2 produces a 15- to 25-mV positive shift in I_A through a similar second messenger cascade (Hampson et al., 1995; Mu et al., 1996) and can be blocked by SR141716A (Mu et al., 1995), a competitive antagonist of the CB₁ cannabinoid receptor (Rinaldi-Carmona et al., 1994).

Initial descriptions of the cannabinoid receptor modulation of I_A in hippocampal neurons focused on the cannabinoid produced shift in voltage dependence of steady-state inactivation of I_A. However, an additional effect on this current was observed in which the inactivation time constant () was also markedly reduced from 50 to 25 ms after cannabinoid exposure (Mu et al., 1997). This effect was observed only with higher concentrations (>300 µM) of 4-AP. Although voltage-dependent outward potassium currents with both 25 and 50 ms  values have been previously observed (Storm, 1990; Sheng et al., 1993), it is likely that a composite of two different potassium currents (only one of which was I_A) with overlapping voltage dependencies really comprised this TEA-insensitive "control" current reported in those studies.

A likely candidate for the second current was potassium D or delay current (I_D) on the basis of the inactivation , insensitivity to TEA, and high sensitivity to 4-AP (100-500 µM; Storm, 1990; Wu and Barish, 1992; Locke and Nerbonne, 1997a). I_D or I_D-like currents have been characterized in several different mammalian neurons, including cultured hippocampal neurons (Wu and Barish, 1992; Luthi et al., 1996). It differs from I_A in that I_D is steady-state inactivated and activated at approximately 30 mV more positive membrane potentials than I_A (Storm, 1990; Wu and Barish, 1992). I_D is also slower to activate ( = 20 ms) and inactivate ( = 100 ms) with a refractory period of up to 20 s (Storm, 1990). In addition, I_D is blocked by low concentrations of 4-AP (<1 mM) and by dendrotoxin (DTX; 2 µM), whereas I_A is insensitive to DTX and requires much higher concentrations of 4-AP (5 mM) for blockade.

There have been no prior reports of the sensitivity of I_D to cannabinoid receptor modulation in any type of central nervous system neuron. In the present experiments, we therefore investigated two main issues: 1) whether the second current contributing to the composite outward potassium current was I_D, and 2) if so, whether that current also was modulated by cannabinoid receptor activation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture. The preparation of hippocampal neurons in culture was similar to that described in several previous reports (Deadwyler et al., 1993, 1995a; Hampson et al., 1995). Hippocampi from fetal (E-18) rats (Zivic-Miller) were incubated with neutral protease (2 U/ml Dispase 1; Boehringer Mannheim Biochemica, Mannheim, Germany) for 40 to 50 min at 37°C. After stopping the enzymatic reaction with 1.0 mM NaEDTA, cells were dissociated by gentle trituration via two flame-polished Pasteur glass pipettes and plated at a density of 3 to 4 × 10⁵ cells/35-mm dish. The plating medium consisted of 59% Dulbecco's modified Eagle's medium (1×), 19.5% Ham's F-12 Nutrient Mixture (1×), 10% FBS, 10% horse serum, and 1% L-glutamine (200 mM) (all from GIBCO BRL, Gaithersburg, MD). Cultures were grown at 37°C in a humidified 5% CO₂ incubator. After 48 h, half of the medium was replaced by "feeding" medium, which consisted of 98% neurobasal medium, 2% B-27 supplement, 0.25% L-glutamine (200 mM), 0.1% 2-mercaptoethanol (all purchased from GIBCO BRL), and 25 mM KCl. At 72 h after plating, half the medium was again replaced with feeding medium, and cultures were treated with 0.75 µM cytosine--D-arabinofuranoside (Sigma Chemical Co., St. Louis, MO) to prevent proliferation of glia. The culture medium was then changed every 3 days for the remainder of the experiment. Experiments were performed on cultured cells between days 7 and 15.

Recording Methods. The procedure for whole-cell recording was similar to that reported previously (Deadwyler et al., 1993). Briefly, patch electrodes were prepared from 1.5-mm o.d./1.1-mm i.d. borosilicate glass capillaries to produce 1- to 2-µm (2-5 M) tip openings. Electrodes were filled by suction and backfilling with a standard intracellular solution of 140 mM KCl, 11 mM EGTA, 1 mM CaCl₂, 2 mM MgCl₂, 2 mM ATP, 200 µM GTP, and 20 mM HEPES buffer (Sigma Chemical Co.). Hippocampal cells in primary culture (7-15 days) were washed and constantly perfused with extracellular medium consisting of 140 mM NaCl, 5 mM KCl, 2.5 mM CaCl₂, 2 mM MgCl₂, 10 mM glucose, and 20 mM HEPES, with 1 µM tetrodotoxin (TTX; Sigma Chemical Co.) added to block voltage-gated sodium channels. Slowly activating/noninactivating potassium currents were blocked by 35 mM TEA, leaving only the transient I_A and I_D (Storm, 1990; Wu and Barish, 1992). Cultured cells were heated (37°C) and perfused with oxygenated (95% O₂/5% CO₂) bathing medium throughout the experiment. The osmolality of the bath was 320 ± 10 mOsm, and the pipette solution was 280 ± 10 mOsm. Osmolality was adjusted to prevent cell shrinkage or swelling resulting from dialysis with the recording pipette solution. Cells were covered to a depth of 2 ml with bathing medium and placed in a Leiden microincubator (Medical Systems, Inc., Greenvale, NY) on a Nikon Diaphot inverted microscope. Positive pressure was applied to the recording pipette as it was lowered into the medium and approached the cell membrane. Constant negative pressure was applied to form the seal. A sharp pulse of negative pressure opened the cell membrane for whole-cell recording (Hamill et al., 1981).

Drug Preparation. The cannabinoid drug was applied via a pressure pipette (10- to 50-µm tip opening) controlled by a solenoid valve (Picospritzer II; General Valve Co., Fairfield, NJ) modified to eject a steady stream of drug-containing media over the surface of the cell. WIN 55,212-2 (Sterling Drug Co., Malvern, PA) was prepared daily from a 10 mM stock solution in ethanol, diluted with extracellular bathing medium, and the ethanol evaporated under a constant stream of nitrogen (Deadwyler et al., 1993). Equivalent bath concentrations corresponding to the pressure pipette concentrations of WIN 55,212-2 are reported in the text. The drug solution was titrated to the same osmolality as the extracellular bathing medium. Due to the lipophilicity of the drug, a 30-s ejection via the application pipette was followed by a washout period of at least 2 min. Previous studies demonstrated that the effects of pressure pipette applications of WIN55,212-2 were rapid (~10-s onset) and were fully reversed after 2-min perfusion of bathing medium after the application; therefore, all current traces were obtained during this 30-s application period (Deadwyler et al., 1993). No effect was observed on whole-cell currents with the application of vehicle-only solution via the same procedure. Controls consisted of vehicle-only applications to separate neurons, as well as pre- and post-drug measurements of I_A and I_D within the same WIN 55,212-2-treated neurons. DTX (Sigma Chemical Co.) used to selectively block I_D (Storm, 1990; Wu and Barish, 1992) was dissolved directly in bathing medium to a concentration of 2 µM. Controls consisted of recordings in normal medium before perfusion with DTX-containing medium. In experiments designed to inhibit cannabinoid receptor coupling to G proteins, cultured neurons (6-15 days) were pretreated with pertussis toxin (PTX; islet-activating protein; Sigma Chemical Co.; Deadwyler et al., 1993). PTX (10 µg/ml) was added to the culture medium 18 h before recording and to the pipette solution for dialyzation into the cell during recording. Control cells were from the same batch of culture plates with no PTX added. To irreversibly activate G_i/o proteins, 600 µM guanosine-5'-O-(3-thio)triphosphate (GTPS), a nonhydrolyzable GTP analog, was added to the pipette solution. Controls were conducted with neurons in the same culture plate not exposed to GTPS. All results were from cells exposed to only one of the above conditions unless otherwise specified. The water-soluble cAMP analog 8-bromo-cAMP (8-br-cAMP; 10 µM) was applied to the cells via the bathing medium. The protein kinase A inhibitors IP-20 ("Walsh" inhibitory peptide, Sigma Chemical Co.) and Rp-cAMPS [(Rp)-diastereomer of cAMP; Sigma Chemical Co.] were dialyzed into the cell through the recording solution. The cannabinoid receptor (CB₁-specific) antagonist SR141716A (provided by Sanofi Reserche, Montpelier, France) was prepared as a 1 mM stock solution in ethanol and diluted in bathing medium to 300 to 500 nM, and the ethanol was removed by evaporation under nitrogen.

Analysis. Time constants () for inactivation of potassium currents were calculated by curve-fitting to the inactivating portion of the outward current trace evoked by a depolarizing voltage step using the CLAMPEX and CLAMPAN patch-clamp acquisition and analysis programs (Axon Instruments). The separation of currents from the composite outward potassium current was also performed using the CLAMPAN program. Measurement of observed changes in voltage dependence of the A and D currents was accomplished by fitting Boltzmann functions to the isolated I_A or I_D generated by the inactivation and activation protocols (see Deadwyler et al., 1993). Peak current amplitude (I_max) in the inactivation protocol was calculated using a 120-mV hyperpolarizing prepulse, whereas peak activation amplitude was calculated using a depolarizing step to +50 mV. I_max was the same or very near the same for all treatment conditions in a given cell. For each combination of prepulse and step voltage, I_A or I_D (I/I_max) was converted to relative conductance [G/G_max; because G_A = I_A/(V_m  E_K), with E_K determined to be approximately 100 mV in cultured hippocampal neurons] and plotted to compare changes in voltage dependence across different drug conditions and different cells (Saint et al., 1990). Boltzmann functions were fitted to mean data points of G/G_max and conditioning prepulse voltages (V_pp) via the following formula: G/G_max = 1/[1 + exp((V_pp  V_1/2)/k)], where V_pp is the conditioning prepulse potential for inactivation; V_1/2 is the voltage at which G/G_max = half-maximum conductance, and  is a slope factor. The coefficients (V_1/2 and k) from the above Boltzmann equations were calculated via nonlinear regression (PC-SAS NLIN).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of Composite Outward Current in Cultured Hippocampal Neurons. Figure 1 (top) shows the nondifferentiated whole-cell current elicited by the voltage step protocol for outward potassium currents in hippocampal cells (Saint et al., 1990; Storm, 1990; Wu and Barish, 1992; Deadwyler et al., 1993). The "composite" current (I_comp) was recorded after the addition of TEA (25 mM) to the external medium, which abolished the noninactivating delayed rectifier, I_K, and other noninactivating outward potassium currents elicited by this protocol (Storm, 1990). The residual I_comp displays fast activation time (<5 ms) and an inactivation time constant ("") of 50 ms. The current has a steady-state voltage-dependent activation between 50 and 0 mV and steady-state inactivation between 90 and 50 mV, with a refractory period of approximately 200 ms. I_comp is blocked by high concentrations (>1 mM) of 4-AP (cf. Saint et al., 1990; Wu and Barish, 1992; Deadwyler et al., 1993) and has been extensively characterized in prior studies of cannabinoid receptor modulation (Deadwyler et al., 1993, 1995a,b; Hampson et al., 1995).

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Separation and identification of voltage-dependent IA and ID. Total current, whole-cell voltage-gated potassium current elicited by the inactivation protocol shown at the top. The holding potential was 50 mV; prepulses were stepped from 90 to 20 mV in 15-mV increments and then followed by a 375-ms, +50-mV pulse (protocol inset). TTX (0.5 µM) was present in the bath solution for block of voltage-gated sodium current and presynaptic activity. Composite (Icomp), after blocking the noninactivating IK with TEA (25 mM), TEA-insensitive outward potassium current shows time-dependent inactivation. IA, residual inactivating current after the addition of 4-AP (500 µM) or DTX (2 µM) to block ID. The remaining current fits the criteria for IA. ID was calculated by subtraction of the 4-AP- and DTX-insensitive IA from Icomp. Amplitude and inactivation time constant () of subtracted or "recovered" ID are within the same range as determined by direct activation (see Fig. 3). Note differences in inactivation  and peak current amplitude for IA and ID.

Effects of Cannabinoids on I_comp. Figure 2 shows the effect of exposure to the potent cannabinoid WIN 55,212 (40 nM) on I_comp elicited by the above steady-state inactivation protocol (see Materials and Methods). These effects were compared with those of DTX (2 µM) on I_comp using the same inactivation protocol. In both cases, mean  was significantly reduced from control I_comp levels (Table 1; F_1,95 = 17.75, p < .001). However, an effect not mimicked by DTX was for WIN 55,212-2 to increase the amplitude of I_comp (now I_A) at all voltage steps (see intermediate traces shown in Fig. 2, bottom left), resulting in a positive shift in the voltage dependence of steady-state inactivation. This is shown graphically by the Boltzmann curves at the right in Fig. 2. As reported elsewhere (Deadwyler et al., 1993, 1995a; Hampson et al., 1995), both steady-state inactivation (circles) and activation (squares) of I_comp were positively shifted by exposure to WIN 55,212-2 (mean I_comp inactivation V_1/2 = 71.3 ± 3.1 mV, mean I_comp + WIN 55,212-2 inactivation V_1/2 = 54.3 ± 4.4 mV, F_1,95 = 12.57, p < .001; mean I_comp activation V_1/2 = 20.1 ± 2.6 mV, mean I_comp + WIN 55,212-2 activation V_1/2 = 3.5 ± 2.4 mV, F_1,95 = 14.2, p < .001). The positive shift in the Boltzmann curves for I_comp produced by WIN 55,212-2 was blocked by simultaneous exposure to SR141716A, the CB₁ receptor antagonist (Table 1). As stated above, this shift in the voltage dependence of I_comp steady-state inactivation did not occur after exposure to DTX (cf. Fig. 1; mean I_comp + DTX inactivation V_1/2 = 72.1 ± 4.8; DTX versus WIN 55,212-2, F_1,95 = 0.56, p = .45) or to WIN 55,212-2 + DTX (Fig. 2; mean I_comp + WIN 55,212-2 + DTX inactivation V_1/2 = 53.2 ± 2.4, F_1,95 = .38, p = .54).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Comparison of the effects of DTX and cannabinoid receptor agonist WIN 55,212-2 on Icomp with cannabinoid enhancement of IA. Left, Icomp elicited with steady-state inactivation protocol under control conditions (top) and after exposure to DTX (2 µM, middle) or WIN 55,212-2 (40 nM, bottom). Both DTX and WIN 55,212-2 produced a similar reductions in inactivation  values (compare with dashed trace for Icomp), leaving a residual current characteristic in both cases of IA. WIN 55,212-2 also produced an increase in peak amplitudes at intermediate voltage steps, indicating a shift in the voltage dependence of steady-state inactivation of IA. Scale bars: 50 ms, 0.5 nA. Right, Boltzmann curves for steady-state inactivation and activation of IA. IA inactivation ( and ). Peak IA amplitude (Imax) was measured for 120 mV prepulse with depolarizing step to +50 mV. Hyperpolarizing prepulses were then varied in 15-mV steps (positive), and the ratio of IA to Imax was computed. The resulting relative current (IA/Imax) was converted to relative conductance (G/Gmax; see text) and fitted by nonlinear regression to the Boltzmann function: G/Gmax = 1/[1 + exp((Vpp  V1/2)/k)]. Symbols indicate mean ± S.E.M. conductance (G/Gmax) across different cells (DTX: n = 8; WIN 55,212-2, n = 12). , Boltzmann curve for IA (DTX treated). , Boltzmann curve and conductance for WIN 55,212-2-treated IA and reflect a 15-mV positive shift in the steady-state inactivation of IA from the DTX-treated condition. IA activation ( and ). Peak IA amplitude (Imax) was measured for 120-mV prepulse and depolarizing step to +50 mV. Depolarizing prepulse was then varied in 10-mV steps (negative); IA/Imax was computed, converted to conductance, and fitted to the Boltzmann equation: G/Gmax = 1/[1 + exp(+(Vpp  V1/2)/k)].

Effects of Cannabinoids on I_D. Figure 3 shows the mean Boltzmann curves for steady-state inactivation and activation of I_D recorded in cultured hippocampal neurons (n = 12). The insets show individual currents developed by the respective steady-state inactivation and activation protocols (see Materials and Methods). Because there is no pharmacological means of blocking I_A independent of I_D and the inactivation protocol includes voltages within a range that will also affect inactivation of I_A, the steady-state inactivation curve for I_D (Fig. 3, circles) was constructed by subtracting I_A from I_comp. To reconstruct I_D, I_comp was recorded at each voltage step in the presence and absence of 4-AP (500 µM) to isolate I_A. The latter current (I_A) was then subtracted from the untreated I_comp to provide the curve for I_D. The V_1/2 for I_D steady-state inactivation was determined to be 39.9 ± 3.5 as shown in Table 1, 30 mV more positive than I_A (Deadwyler et al., 1993). The curve for steady-state activation of I_D was constructed with an activation protocol and membrane holding potential at 40 mV (Fig. 3, squares). The V_1/2 for steady-state activation of I_D was +2.6 ± 2.7 mV (Table 1). Either method (subtraction or direct activation) produced comparable current profiles (insets, Fig. 3) respect to both  and amplitude of I_D. The voltage dependence for both steady-state inactivation and activation were consistent with several previous reports of I_D characteristics (Storm, 1987, 1990; Wu and Barish, 1992).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Steady-state inactivation and activation of ID. Peak ID amplitude (Imax) was derived and measured using the same inactivation and activation protocols as in Figs. 1 and 2. IA was pharmacologically separated from Icomp in the steady-state inactivation protocol and then subtracted from Icomp to produce ID. IA (40-mV prepulse) was eliminated in the activation protocol (50-mV prepulse, depolarizing step to +50 mV). For each combination of prepulse (inactivation) or step voltage (activation), Boltzmann curves for steady-state inactivation and activation were fitted as in Fig. 2. Points are mean ± S.E. conductance elicited for each voltage (n = 21 cells). Inset traces, isolated ID currents elicited by the inactivation and activation protocols. ×, on the Boltzmann curves, single-cell conductance measurements obtained from the traces shown. Scale bars: 50 ms, 0.5 nA.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Cannabinoid effects on ID. Left, WIN 55,212-2 effects on ID activation (steady-state activation protocol). TEA, activation of TEA-insensitive ID in normal media with 25 mM TEA added. Steady-state inactivation of IA is effected by a step to 40 mV before depolarizing steps to 0 to +50 mV (inset, ID activation protocol). WIN 55,212-2, exposure to 10 and 40 nM WIN 55,212-2 (in TEA) produces a dose-dependent decrease in ID amplitude with no change in . DTX, exposure to DTX (2 µM) produced a slightly greater reduction in ID. Scale bar: 50 ms, 0.5 nA. Right, Boltzmann curves for steady-state inactivation and activation of ID. Exposure to WIN 55-212-2 (solid lines) produced an apparent shift in voltage dependence of inactivation and activation compared with control (dashed line); however, the shifts were accompanied by symmetric decreases in slope for activation and inactivation (see text).

Relative Potency and Efficacy of Cannabinoids on I_A and I_D. Figure 5 shows the concentration-effect (2.5-40 nM) curves for the WIN 55,212-2 modulation of I_D amplitude (filled circles) and shift in V_1/2 for steady-state inactivation of I_A (open squares). A comparison of the EC₅₀ values for each effect revealed that the cannabinoid reduction in mean I_D amplitude occurred at lower concentrations (13.9 ± 1.6 nM) than the mean V_1/2 shift in I_A voltage dependence (20.6 ± 1.9 nM; T₁₄ = 2.69, p < .01). Thus, I_D was 34% more sensitive to the influence of cannabinoids than I_A, a further indication that cannabinoid alteration of I_comp was produced by the above demonstrated independent effects on I_A and I_D.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Concentration-dependent effects of WIN 55,212-2 on IA and ID (range, 2.5-100 nM). Dose-effect curves for WIN 55,212-2 effect on reduction of ID amplitude (). Half-maximal effect (EC50) was 13.9 nM, and the maximal effect was obtained at 70 nM (not shown). The dashed line shows the effects of the same concentrations of WIN 55,212-2 on the shift in voltage dependence (V1/2) for steady-state inactivation of IA (). The EC50 value for cannabinoid effects on IA was 20.6 nM, and the maximal effect was obtained at 80 nM (at least 4 cells/point; range, 4-21).

Cannabinoid Effects on I_A and I_D Are Mediated by G_i/o Proteins, cAMP, and Protein Kinase A (PKA). It has been firmly established that cannabinoid receptor inhibition of adenylyl cyclase is mediated through PTX-sensitive G proteins (Howlett et al., 1986; Bidaut-Russell et al., 1990; Pacheco et al., 1991). Consistent with this effect of cannabinoid receptor activation, previous reports demonstrated that the effects of cannabinoids on I_A were G protein (Deadwyler et al., 1993) and cAMP dependent (Deadwyler et al., 1995a). In those studies, increased levels of cAMP (8-br-cAMP), PKA, protein kinase inhibitors (IP-20 and Rp-cAMPS), and WIN 55,212-2 were all used (Deadwyler et al., 1993, 1995a; Hampson et al., 1995). The results of similar treatments with respect to effects on I_A were replicated in the present study and are summarized in Table 1 and Fig. 6, A and B. In general, manipulations that increased cAMP levels produced a negative shift in the voltage dependence of I_A but not the slope of the steady-state activation and inactivation curves (Deadwyler et al., 1995a; Hampson et al., 1995). Treatments that decreased cAMP levels or blocked PKA produced a positive shift in I_A activation and inactivation with no change in slope of the Boltzmann curves (Fig. 6, A and B). Because the effects on PKA presumably altered phosphorylation of membrane channel proteins, the catalytic subunit of PKA (Cat.S.; Table 1) and the phosphatase inhibitor okadaic acid (OA; Table 1) were also tested. Both Cat.S. and OA induced a negative shift in I_A voltage dependence that was similar to increased cAMP levels (8-br-cAMP). Table 1 shows that the effects of WIN 55,212-2 on I_A and I_comp were similar in direction and magnitude to decreased cAMP levels or PKA inhibition (especially IP-20) and opposite in effect to increased cAMP levels and enhanced PKA-dependent phosphorylation by other agents. These results indicate that the differential shifts in I_A voltage dependence were regulated by the phosphorylation status of the I_A channel protein. Furthermore, because exposure to 8-br-cAMP, Cat.S., or OA blocked cannabinoid effects on I_A (Table 1) and IP-20 or Rp-cAMPS plus WIN 55,212-2 were not additive, it is therefore likely that the same cAMP/PKA cascade was involved for all of the above influences on I_A.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   cAMP/PKA modulation of IA and ID steady-state inactivation and activation. A, Boltzmann curves (see Fig. 2) for steady-state inactivation of IA (see Protocol inset) calculated for currents recorded in 8-br-cAMP (10 µM), WIN 55,212-2 (WIN 55,212-2; 40 nM), or DTX (2 µM) in the bathing solution or after dialyzing the cell, via the recording pipette solution, with the PKA inhibitory peptide IP-20 (6 µM) or partial PKA inhibitor Rp-cAMPS (2 µM). Symbols reflect mean ± S.E. conductance across all cells for each prepulse voltage (see Table 1 for V1/2 and n values). B, activation of IA. Treatment conditions were the same as above for Inactivation. Symbols reflect mean ± S.E. conductance across cells for each depolarizing step voltage shown in Protocol inset. Inset, current traces show change in peak IA (activation) after exposure to IP-20 and 8-br-cAMP and under control conditions. C, ID inactivation. Boltzmann curves for steady-state inactivation of ID (by subtraction of IA from Icomp) were calculated by subtraction of IA from Icomp (Fig. 2, see also Table 1 for V1/2 and n values). D, Boltzmann curves for activation of ID. Treatment conditions and symbols are the same as in B. Inset, current traces elicited by the ID steady-state activation protocol for 8-br-cAMP, IP-20, and control. Scale bar: 50 ms, 0.5 nA

View larger version (39K): [in this window] [in a new window]   Fig. 7.   Maximum ID activation. ID was recorded using three combinations of prepulse (holding) voltage (120, 80, and 40 mV) with a +80-mV depolarizing voltage in the protocol indicated. Voltages were selected to exceed the range used to elicit maximum currents in Fig. 6, C and D, to test whether the slope changes resulted from a shift in voltage dependence or decreased maximal conductance. Peak current was not significantly different for each protocol (Control) but was greatly reduced after exposure to PKA inhibitors, WIN 55,212-2 and DTX. Rp-cAMPS, thio-cAMP stereoisomer, WIN 55,212-2, cannabinoid agonist WIN 55,212-2; and IP-20, PKA inhibitory peptide (20 amino acid).

Contributions of I_A and I_D to I_comp. Figure 8 shows the net changes in I_comp with different manipulations of cAMP. The solid traces show I_comp recorded from a single neuron under control conditions (i.e., "resting" levels of cAMP) and after exposure to 8-br-cAMP (10 µM) or the cannabinoid agonist WIN 55,212-2 (40 nM). The horizontal lines at the top in Fig. 8 show that increased cAMP levels (8-br-cAMP) reduced the peak amplitude of I_comp by 16% relative to control (long vertical arrow); in contrast, cannabinoid exposure increased peak I_comp amplitude by only 4% (short vertical arrow). However, at the same time, increased cAMP levels (8-br-AMP) caused a 27% increase in I_comp  relative to control, whereas cannabinoids decreased the I_comp  by 58%, almost twice the change produced by cAMP. The dashed line in Fig. 8 (8-br-cAMP and Cannabinoid) indicates the degree of change in time course of I_comp under both conditions as indicated by the direction of the arrow. Thus, the maximum range in I_comp  from increased cAMP levels (Fig. 8, 8-br-cAMP) to decreased cAMP levels produced by cannabinoid receptor activation (Fig. 8, Cannabinoid) was 68%. The dotted traces in Fig. 8 depict I_A (8-br-cAMP and Control) and show that I_comp amplitude was reduced in conjunction with I_A amplitude after exposure to increased levels of cAMP. Conversely, the dash-dotted traces (Control and Cannabinoid) show the reciprocal change in amplitude and  of I_D as a function of decreased cAMP levels produced by cannabinoid exposure. In each case, it is clear that an increase in cAMP produced an increase in the duration of I_comp but at the expense of a marked reduction in amplitude. Decreasing cAMP levels via cannabinoid receptor activation resulted in a marked decrease in the duration of I_comp, with a relatively insignificant increase in amplitude. This dual modulation of I_comp shown in Fig. 8 is cAMP dependent; however, the differential modulation of I_A and I_D must occur "downstream" at the level of PKA-dependent phosphorylation, as indicated by the differential effects of IP-20 and Cat.S. on both currents (see Table 1).

View larger version (12K): [in this window] [in a new window]   Fig. 8.   Effects of differential modulation of IA and ID by cAMP on Icomp. The three sets of single-cell voltage-clamp current traces show Icomp under control (center), 8-br-cAMP (10 µm, left), and WIN 55,212-2 (40 nM, right) recording conditions. Solid current traces indicate Icomp recorded under each of the three conditions. Dashed current traces at left (8-br-cAMP) and right (WIN 55,212-2 ) are superimposed control Icomp (center) traces. Bold arrows show differential change in  from control Icomp. Horizontal lines and arrow at top depict change in Icomp amplitude (right to left) produced by increasing levels of cAMP. Dotted traces (middle and left) depict decrease in IA as cAMP levels increase (8-br-cAMP). Dash-dotted traces (middle and right) show decrease from maximum ID (8-br-cAMP), through control (middle), to minimum ID after exposure to WIN 55,212-2 (right).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The above results clearly demonstrate that a major source of voltage-dependent potassium current (I_comp) in hippocampal cells is selectively modulated by cannabinoid receptor occupancy. I_A and I_D, which contribute to I_comp in cultured hippocampal neurons, are voltage-dependent outward currents with overlapping activation ranges; both are differentially sensitive to 4-AP, and both are TEA resistant. Because the two currents, I_A and I_D, inactivate at different  values (I_A  25 ms; I_D  100 ms), have different voltage dependencies for steady-state inactivation (V_1/2: I_A  70 mV; I_D  40 mV), and can be totally differentiated pharmacologically with DTX (2 µM) or low-to-moderate concentrations of 4-AP (500 µM), the cannabinoid receptor modulation of each current was examined independent of the other.

It is clear from the above data that the previously reported effects of cannabinoids (Deadwyler et al., 1993, 1995a; Hampson et al., 1995) on I_A in cultured hippocampal neurons can now be identified with respect to the contributions of I_A and I_D to I_comp (Figs. 2, 6, and 8). Hence the positive shift in voltage dependence of I_comp resulted from cannabinoid-mediated enhancement of I_A, whereas the reduction in  of I_comp was derived from decreased I_D amplitude. The addition of the selective channel blocker DTX produced a reduction in the I_comp  similar to WIN 55,212-2 (Figs. 1 and 2) and blocked any further actions of WIN 55,212-2 (Table 1). Exposure to WIN 55,212-2 also produced a reduction in I_D amplitude in both the steady-state inactivation and activation protocols (Fig. 4). DTX, however, did not provoke a change in the voltage dependence of I_A (Fig. 2). Further evidence that both the positive shift in I_A voltage dependence and the reduction in I_D amplitude were cannabinoid receptor mediated was provided by blockade by the CB₁ receptor antagonist SR141716A (Table 1). Although a potential alternative interpretation for the dual effects of cannabinoids on I_comp is that the shift in voltage dependence of I_A (Fig. 2) could have resulted directly from the elimination or marked reduction of I_D (Fig. 4), the removal of a current with a more positive Boltzmann curve (V_1/2) than I_A (as was the case for I_D; see Fig. 6) would have produced a negative (not a positive) shift in the voltage dependence of I_comp (Fig. 2). The predicted negative shift would result only if both currents were activated with the same time constant (cf. Fig. 6). The slower activation time constant for I_D (20 ms) requires that the initial peak of I_comp (5-ms duration) consists almost entirely of I_A; thus, the positive shift in I_comp voltage dependence was strictly the result of the cannabinoid receptor activation on I_A, not I_D. Similarly, the reduction in I_comp  resulted from direct cannabinoid effects on I_D amplitude because I_A inactivation  was not altered by cannabinoids (Table 1).

Prior reports have established that cannabinoid receptor modulation of I_A voltage dependence can be attributed ultimately to PKA-dependent phosphorylation (Table 1; Hampson et al., 1995; Mu et al., 1996). Several manipulations of the cAMP/PKA cascade, including direct application of Cat.S (Table 1), confirmed that the reduction in I_D amplitude produced by cannabinoid receptor activation was consistent with a decrease in PKA dependent phosphorylation (presumably of I_D channel proteins) in these cells (Fig. 6, Table 1). Figure 8 further demonstrates that the effect of increased cAMP on I_D steady-state inactivation was reciprocal to modulation by the cannabinoid receptor of I_A. The fact that the control (untreated) measures of I_D amplitude and I_A V_1/2 were between these two extremes shows that a tonic level of cAMP is normally present and active, allowing both I_A and I_D to be only partially "expressed" under normal, control conditions (Control trace in Fig. 1). Because there was a change in slope in the Boltzmann curves for I_D, it was not possible to determine whether the voltage dependence of either activation or inactivation of I_D was altered by cannabinoids and other agents that modulate cAMP levels (Fig. 6). However, there was a cAMP-dependent decrease in I_D amplitude, presumably due to total blockade of the channel as occurs with DTX (Storm, 1990; Wu and Barish, 1992). The fact that the EC₅₀ value was lower for cannabinoid reduction in I_D amplitude than for shifting the voltage dependence of I_A suggests a slight bias toward decreasing the  of I_comp.

Mechanism of Cannabinoid Modulation of I_A and I_D. The above results provide evidence that a reciprocal relationship exists between I_A and I_D with respect to cannabinoid receptor modulation of levels of cAMP. Modulation of voltage-gated potassium channel currents (I_comp) can play a major role in altering the temporal and amplitude characteristics of action potentials in hippocampal neurons (Segal et al., 1984). Such reciprocity is likely explained by differences in the subtypes of K⁺ channels producing I_A and I_D. If the two currents were produced by different subtypes (i.e., shaker-type Kv1 versus shal-type Kv4 channels or even Kv4.2 versus Kv4.3 channels), then the common factor of cannabinoid receptor-mediated inhibition of adenylyl cyclase, PKA, and subsequent protein phosphorylation could be translated into opposite actions on I_A and I_D. The I_A channel described in these studies fits the description of the Kv4.2 or Kv4.3 subtypes, which have been shown to be present in hippocampal pyramidal cells and interneurons (Villarroel and Schwarz, 1996; Serodio and Rudy, 1998). Although the profile of I_A is also satisfied by Kv1.4, this can be ruled out on the basis of a lack of effect of H₂O₂ on either I_A or I_comp (Mu et al., 1997). Different inactivation mechanisms have been demonstrated for different types of potassium channels (Levitan, 1994). When phosphorylated, the Kv4.2- and Kv4.3-type homomultimers have low conductances that change drastically toward maximum under conditions of dephosphorylation (Aiello et al., 1995). Thus, for this putative I_A channel, cannabinoid receptor-mediated decreases in PKA phosphorylation would lead to a positive shift in the inactivation voltage of the I_A channel (Deadwyler et al., 1993; Hampson et al., 1995).