Inhibition of human recombinant T-type calcium channels by the endocannabinoid N-arachidonoyl dopamine

Abstract

Background and purpose:

N-arachidonoyl dopamine (NADA) has complex effects on nociception mediated via cannabinoid CB₁ receptors and the transient receptor potential vanilloid receptor 1 (TRPV1). Anandamide, the prototypic CB₁/TRPV1 agonist, also inhibits T-type voltage-gated calcium channel currents (I_Ca). These channels are expressed by many excitable cells, including neurons involved in pain detection and processing. We sought to determine whether NADA and the prototypic arachidonoyl amino acid, N-arachidonoyl glycine (NAGly) modulate T-type I_Ca

Experimental approach:

Human recombinant T-type I_Ca (Ca_V3 channels) expressed in HEK 293 cells and native mouse T-type I_Ca were examined using standard whole-cell voltage clamp electrophysiology techniques.

Key results:

N-arachidonoyl dopamine completely inhibited Ca_V3 channels with a rank order of potency (pEC₅₀) of Ca_V3.3 (6.45) ≥ Ca_V3.1 (6.29) > Ca_V3.2 (5.95). NAGly (10 µmol·L⁻¹) inhibited Ca_V3 I_Ca by approximately 50% or less. The effects of NADA and NAGly were voltage- but not use-dependent, and both compounds produced significant hyperpolarizing shifts in Ca_V3 channel steady-state inactivation relationships. By contrast with anandamide, NADA and NAGly had modest effects on Ca_V3 channel kinetics. Both NAGly and NADA inhibited native T-type I_Ca in mouse sensory neurons.

Conclusions and implications:

N-arachidonoyl dopamine and NAGly increase the steady-state inactivation of Ca_V3 channels, reducing the number of channels available to open during depolarization. These effects occur at NADA concentrations at or below to those affecting CB₁ and TRPV1 receptors. Together with anandamide, the arachidonoyl neurotransmitter amides, NADA and NAGly, represent a new family of endogenous T-type I_Ca modulators.

Introduction

Endogenous compounds consisting of arachidonic acid conjugated with amino acids or neurotransmitter amines are a large class of molecules found in many tissues including brain and spinal cord (Huang et al., 2001; 2002; Milman et al., 2006; Saghatelian et al., 2006; Rimmerman et al., 2008). While details of their synthesis and metabolism are only beginning to emerge, several of these compounds have been shown to have profound effects on nociception (Burstein et al., 2000; Huang etal., 2001; 2002), inflammation (Burstein et al., 2007) and vascular function (O'Sullivan et al., 2004; Milman et al., 2006) following systemic administration or incubation with tissue in vitro. The molecular targets of these compounds are also beginning to be defined, and in some respects they resemble those of the endogenous cannabinoid agonist, anandamide (AEA) (N-arachidonoyl ethanolamide). Thus, N-arachidonoyl dopamine (NADA) is an agonist at cannabinoid CB₁ receptors (Bisogno et al., 2000), transient receptor potential vanilloid receptor 1 (TRPV1) receptors (Huang et al., 2002) and a weak inhibitor of a major AEA hydrolysing enzyme, fatty acid amide hydrolase (FAAH, Bisogno et al., 2000). Intriguingly, the N-arachidonoyl amide most closely related to AEA, N-arachidonoyl glycine (NAGly), has very low affinity for CB₁ receptors (Sheskin et al., 1997), modest inhibitory potency at FAAH (Huang et al., 2001) and negligible activity at TRPV1 receptors (De Petrocellis et al., 2000). It is, however, the prototypic agonist for the orphan G protein coupled receptor, GPR 18 (Kohno et al., 2006), and an agonist of another orphan receptor, GPR 92 (Oh et al., 2008).

The mechanisms through which NADA and NAGly modulate nociception remain incompletely understood. NADA has pro- and anti-nociceptive effects, depending on the route and site of administration (Bisogno et al., 2000; Huang et al., 2002). In different experiments, both pro- and anti-nociceptive effects have been explained on the basis of TRPV1 activation (Sagar et al., 2004; Huang and Walker, 2006), while some anti-nociceptive actions involve CB₁ receptors (Sagar et al., 2004). There is also evidence for CB₁/TRPV1-independent actions of NADA in sensory neurons (Price et al., 2004). The antinociceptive effects of NAGly are not sensitive to CB₁ antagonists (Succar et al., 2007; Vuong et al., 2008), indicating that its main mechanism of action is unlikely to be elevation of endocannabinoids following inhibition of FAAH. Recent studies have described selective NAGly inhibition of the glycine transporter GlyT2A and complex effects on glycine receptors (Wiles et al., 2006; Yang et al., 2008), both of which could conceivably contribute to anti-nociception. As part of a search for possible sites of action relevant to the anti-nociceptive effects of NADA and NAGly, we examined the effects of these compounds on T-type calcium channels, low-voltage-activated channels that have an important role in sensory processing and are an emerging target for analgesics (Shin et al., 2008). We report that NADA and NAGly robustly inhibit recombinant human and native mouse T-type calcium channels. The effects are broadly similar to those previously reported for AEA (Chemin et al., 2001; 2007), but quite distinct from the effects of NADA and NAGly on high-voltage-activated, N-type calcium channels (Guo et al., 2008).

Methods

Cell culture

HEK 293 cells stably transfected with plasmids containing cDNA for the human Ca_V3.1, Ca_V3.2 or Ca_V3.3 (Cribbs etal., 1998; 2000; Gomora et al., 2002, Ross et al., 2008) were cultivated in Dulbecco's modified Eagle's Medium supplemented with 100 U penicillin, 100 µg streptomycin, 10% fetal bovine serum or donor bovine serum and 1 mg·mL⁻¹ G418 (Invitrogen, Mt. Waverly, Australia).

Isolation of sensory neurons

All animal procedures were approved by the Royal North Hospital Animal Care and Ethics Committee. Male C57Bl6 mice at least 8 weeks old were anaesthetized with isofluorane, decapitated and the trigeminal ganglia removed. Adult mouse trigeminal ganglion neurons were isolated as described in Ross et al. (2008). Briefly, ganglia were placed in a modified HEPES-buffered saline (mHBS) containing (in mM): 130 NaCl, 2.5 KCl, 1.8 CaCl₂, 10 MgCl₂. 10 HEPES, 10 glucose (pH to 7.3 with NaOH, osmolarity 330 ± 5 mosmol). The ganglia were cut into pieces with iridectomy scissors and incubated in mHBS containing 20 U·mL⁻¹ papain for 25 min at 37°C. The reaction was stopped with mHBS containing 1 mg·mL⁻¹ BSA and 1 mg·mL⁻¹ trypsin inhibitor (Type II-O). The tissue was then washed with mHBS and cells released by gentle trituration through fire-polished Pasteur pipettes. Cells were plated onto tissue culture dished and used within 8 h of isolation.

Electrophysiology

Voltage-gated calcium channel currents (I_Ca) in HEK 293 cells were recorded in the whole-cell configuration of the patch-clamp method (Hamill et al., 1981) at room temperature, unless otherwise noted (Ross et al., 2008). Dishes were perfused with HBS containing (in mmol·L⁻¹): 140 NaCl, 2.5 KCl, 2.5 CaCl₂, 1 MgCl₂, 10 HEPES, 10 glucose (pH to 7.3, osmolarity = 330 ± 5 mosmol). For recording Ca_V3.1 and 3.2 currents, cells were bathed in an external solution containing (in mM): 140 tetraethylammonium chloride, 2.5 CsCl, 10 HEPES, 10 glucose, 1 MgCl₂, 5 CaCl₂ (pH to 7.3, osmolarity = 330 ± 5 mosmol). For recording Ca_V3.3 currents, 5 mmol·L⁻¹ CaCl₂ was replaced by 5 mmol·L⁻¹ BaCl₂ in the external solution (see Ross et al., 2008). Recordings were made with fire-polished borosilicate glass pipettes with resistance ranging from 2 to 3 MΩ. For recording Ca_V3.1 and 3.2 currents, the internal solution contained (in mmol·L⁻¹): 130 CsCl, 10 HEPES, 2 CaCl₂, 10 EGTA, 5 MgATP (pH to 7.3, osmolarity = 285 ± 5 mosmol). For recording of Ca_V3.3 currents, 10 mmol·L⁻¹ EGTA was replaced by 10 mmol·L⁻¹ BAPTA, and the concentration of MgATP was reduced to 1 mmol·L⁻¹. Recordings were made with a HEKA EPC 10 amplifier with Patchmaster acquisition software (HEKA Elektronik, Lambrecht/Pfalz, Germany), an Axopatch 1D amplifier using pClamp 5 software (Molecular Devices, Sunnyvale, CA, USA) and an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA, USA) using AxoGraph X software (http://axographx.com/). Data were sampled at 5–20 kHz, filtered at 2 kHz and recorded on a hard disk for later analysis. Series resistance ranged from three to 10 MΩ, and was compensated by at least 80% in all experiments. Leak subtraction using a P over 4 protocol (with 10 mV test steps) was used for some experiments where cells were being stepped to a single potential, but it was not employed for experiments where more complex waveforms were applied to the cells (e.g. inactivation). Uncompensated leak in these latter experiments did not exceed −30 pA at −106 mV, and cells with a leak current of greater than −30 pA were not used for any experiment. Cells were exposed to drugs via flow pipes positioned approximately 200 µm from the cell.

Recordings from trigeminal ganglion sensory neurons were made as outlined in Borgland et al. (2001). The solutions were the same as those outlined above, except that T-type I_Ca were recorded in an external TEA solution containing 2.5 mmol·L⁻¹ Ca²⁺ and 1 mmol·L⁻¹ Mg²⁺. Other membrane currents were recorded using HBS as the external solution. Recordings were made from small to medium sized cells (<25 µm diameter) which were identified as Type 1 or Type 2 cells from their I_Ca signatures derived from a current–voltage protocol, as outlined in Borgland et al. (2001). Type 1 neurons express little or no T-type I_Ca (measured at a test potential of −40 mV) and typically express TRPV1 and µ-opioid receptors. Type 2 neurons have significant I_Ca at a test potential of −40 mV, do not express TRPV1 and are not sensitive to µ-opioid agonists.

Effects of drug vehicle (ethanol)

Drugs were kept in concentrated stock solutions in ethanol and stored at −20°C. Daily dilutions from these stocks were made; the final ethanol concentration in all solutions was 0.1%. Ethanol at this concentration did not significantly affect the properties of the Ca_V3 channels. After 10 min recording in control conditions or ethanol (0.1%), the V₅₀ for activation for the Ca_V3 channels was not different: Ca_V3.1: 43 ± 1 mV, 43 ± 1 mV; Ca_V3.2: 40 ± 2 mV, 39 ± 2 mV; Ca_V3.3: −37 ± 1 mV, −35 ± 2 mV (n = 5 to 9 cells for each condition). The amplitudes of the currents elicited by stepping from −106 mV to −26 mV were also similar after 10 min in control conditions and 10 min in ethanol (0.1%): Ca_V3.1: −642 ± 115 pA, −742 ± 150 pA; Ca_V3.2: −331 ± 74 pA, −258 ± 77 pA; Ca_V3.3: −1.82 ± 0.4 nA, −2.02 ± 0.4 nA (n = 5 to 9 cells for each condition). Acute application of ethanol (0.1%) inhibited Ca_V3.1 channels activated by stepping from −86 mV to −26 mV by 1 ± 1% (n = 12).

Data analysis

Concentration–response (I_Ca) curves were generated by fitting data to a sigmoidal dose–response function in GraphPad Prism 4. Steady-state activation curves were generated from current–voltage relationships, while steady-state inactivation curves were generated by measuring the peak current from a 50 ms step to −26 mV following a series of 5 s steps ranging from potentials of −126 mV to −46 mV. Reported potentials are corrected for a junction potential of −6 mV. Activation curves were generated by fitting data to a Boltzmann sigmoidal function Y = 1/(1 + e^{((V0.5−Vm)/Slope)}). Inactivation curves were generated by fitting data to a Boltzmann sigmoidal function Y = 1 − 1/(1 + e^{((V0.5−Vm)/Slope)}).

Statistical significance for comparing the V_0.5 values of activation and inactivation was determined using unpaired t-tests comparing values of V_0.5 calculated for individual experiments. In order to compare the changes in the time constant of inactivation and deactivation, two-way anova was used with a Bonferroni post-test to compare values at different potentials.

Materials

N-arachidonoyl dopamine, NAGly and AEA were obtained from either Alexis Biochemicals (Lausen, Switzerland), Biomol (Plymouth Meeting, PA, USA) or Cayman Chemical (Ann Arbor, MI, USA). Where possible, the same drug was purchased from several sources. In all cases, results were similar with drugs purchased from different suppliers. Papain was from Worthington, and all other drugs and chemicals were from Sigma Australia.

Drug and molecular target nomenclature conforms to the BJP Guide to Receptors and Channels (Alexander et al., 2008).

Results

N-arachidonoyl glycine is structurally very similar to the endocannabinoid AEA, differing only by having an additional oxygen molecule, that is, having a terminal carboxylic acid instead of alcohol. AEA is a reasonably potent inhibitor of T-type calcium channels (Chemin et al., 2001; 2007), so we examined the effects NAGly on T-type calcium channels in mouse trigeminal sensory neurons (Borgland et al., 2001). Superfusion of NAGly onto Type 2 trigeminal ganglion neurons inhibited the I_Ca evoked by a step from −86 mV to −46 mV (Figure 1). The highest concentration of NAGly testable (30 µmol·L⁻¹) inhibited the currents at −46 mV by 80 ± 8% (n = 6), 10 µmol·L⁻¹ NAGly inhibited the currents by 52 ± 7%. As the low-voltage-activated calcium currents in native sensory neurons are likely to represent a mixed population of channels, we examined the effects of NAGly and the CB₁/TRPV1 agonist NADA on recombinant human Ca_V3 channels stably expressed in HEK293 cells.

Figure 1.

N-arachidonoyl glycine (NAGly) inhibits native T-type calcium channels. Whole-cell patch clamp recordings were made from acutely isolated mouse trigeminal ganglion neurons. Currents were evoked by stepping from −86 mV to −46 mV. (A) Time plot of the peak amplitude of the I_Ca at −46 mV, illustrating the effects of 10 µmol·L⁻¹ NAGly (perfused for the duration of the bars). (B) Example traces from the above cell in the absence and presence of NAGly. The dotted line represents zero current. I_Ca, voltage-gated calcium channel current.

N-arachidonoyl glycine and NADA inhibited each of the human Ca_V3 subtypes (Figures 2 and 3). The inhibitory effects of lower concentrations of drug on Ca_V3 channels could be at least partially reversed by washing. At a concentration of 10 µmol·L⁻¹ or 30 µmol·L⁻¹, NADA completely inhibited each of the Ca_V3 channels, while NAGly was considerably less effective, with about 50% or less inhibition of each channel at a concentration of 10 µmol·L⁻¹ (Figure 4). The potency of NADA and NAGly inhibition of Ca_V3 channels was determined by superfusing single concentrations of drug onto cells repetitively stepped from −86 mV to −26 mV. NADA inhibited Ca_V3.1 with pEC₅₀ of 6.29 ± 0.03, Ca_V3.2 with a pEC₅₀ of 5.95 ± 0.02 and Ca_V3.3 with a pEC₅₀ of 6.45 ± 0.02. The endogenous compound N-oleoyl dopamine (18 : 1ω9, 10 µmol·L⁻¹, Chu et al., 2003) inhibited Ca_V3 channels to similar degree as NADA (10 µmol·L⁻¹), but the unsaturated N-palmitoyl dopamine (C16, 10 µmol·L⁻¹) was much less effective (Figure 3). NAGly inhibited Ca_V3.1 channels with a notional EC₅₀ of 16 µmol·L⁻¹, but the highest concentration tested (30 µmol·L⁻¹) did not completely inhibit the channels (Figure 4).

Figure 2.

N-arachidonoyl glycine (NAGly) inhibits recombinant human T-type calcium channels. Whole-cell patch clamp recordings were made from human Ca_V3.1, Ca_V3.2 and Ca_V3.3 channels stably expressed in HEK 293 cells. Currents were evoked by stepping from −86 mV to −26 mV. The effect of NAGly on each of Ca_V3.1, Ca_V3.2 and Ca_V3.3. are illustrated, with a representative time plot and example traces. Each trace is an example of at least six similar experiments. The dotted line represents zero current.

Figure 3.

N-acyl dopamine compounds inhibit recombinant human T-type calcium channels. Whole-cell patch clamp recordings were made from human Ca_V3.1, Ca_V3.2 and Ca_V3.3 channels stably expressed in HEK 293 cells. Currents were evoked by stepping from −86 mV to −26 mV. (A) The effect of N-arachidonoyl dopamine (NADA) N-oleoyl dopamine (OLDA) and N-palmitoyl dopamine (PALDA) on each of Ca_V3.1, Ca_V3.2 and Ca_V3.3. are illustrated, with a time plot for the effects of NADA. (B) A summary of the effects of 10 µmol·L⁻¹ of each of the N-acyl dopamines on each of the recombinant human Ca_V3 channels. The bars are the mean ± SEM of 6–9 cells for each drug on each channel. The dotted line represents zero current.

Figure 4.

Concentration–response plots for N-arachidonoyl dopamine (NADA) and N-arachidonoyl glycine (NAGly) inhibition of Ca_V3 channels. Whole-cell patch clamp recordings were made from human Ca_V3.1, Ca_V3.2 and Ca_V3.3 channels stably expressed in HEK 293 cells. Currents were evoked by stepping from −86 mV to −26 mV. A single concentration of drug was superfused over each cell, each point represents the mean ± SEM of at least six cells. (A) NADA inhibits Ca_V3 channels with a rank order of Ca_V3.3 (350 nmol·L⁻¹) ≥ Ca_V3.1 (500 nmol·L⁻¹) > Ca_V3.2 (1.13 µmol·L⁻¹). (B) Concentration-response plot for NAGly inhibition of Ca_V3.1. The highest concentration of NAGly tested was 30 µmol·L⁻¹, assuming complete inhibition the notional EC₅₀ was 16 µmol·L⁻¹. The inhibition of Ca_V3.2 and Ca_V3.3 by 10 µmol·L⁻¹ NAGly is plotted for comparison.

We further examined the interactions of NADA and NAGly on Ca_V3.1 by examining whether inhibition of the channel was use-dependent or influenced by the membrane potential at which the cell was voltage clamped during the experiment. The amount and macroscopic time-course of inhibition of Ca_V3.1 were the same for NADA (300 nmol·L⁻¹) and NAGly (10 µmol·L⁻¹) whether the currents were evoked at 1 Hz or 0.05 Hz (Figure 5). However, the amount of inhibition by NADA (300 nmol·L⁻¹) and NAGly (10 µmol·L⁻¹) was strongly influenced by the holding potential of the cell, with significantly greater inhibition at −86 mV than −126 mV (Figire 5). Inclusion of the competitive inhibitor of G protein activation, GDPβS (1.2 mmol·L⁻¹), failed to affect the inhibition of Ca_V3.1 by NADA (300 nmol·L⁻¹; 50 ± 5% in GTP vs. 49 ± 6% in GDPβS, n = 6 each) or NAGly (10 µmol·L⁻¹; 52 ± 2% in GTP vs. 53 ± 8% in GDPβS, n = 6 each) applied 10 min after breaking into the cells.

Figure 5.

Inhibition of Ca_V3.1 by N-arachidonoyl dopamine (NADA) and N-arachidonoyl glycine (NAGly) was voltage- but not use-dependent. Whole-cell patch clamp recordings were made from human Ca_V3.1 channels stably expressed in HEK 293 cells. (B) The time course of inhibition of Ca_V3.1 currents by NADA (300 nmol·L⁻¹) and NAGly (10 µmol·L⁻¹) were similar when currents were evoked by stepping from −86 mV to −26 mV at 1 Hz or 0.05 Hz. (B) The inhibition of Ca_V3.1 by NADA (300 nmol·L⁻¹) and NAGly (10 µmol·L⁻¹) was significantly less when cells were voltage clamped at −126 mV and stepped to −26 mV than when cells were stepped to −26 mV from a holding potential of −86 mV. **P < 0.01 versus −86 mV, Student's t-test.

We compared the effects of NADA (500 nmol·L⁻¹), NAGly (10 µmol·L⁻¹) and AEA (300 nmol·L⁻¹) on Ca_V3.1 channel kinetics by comparing the effects of 5 min applications of compound with time-matched controls. These concentrations of drug inhibited the peak current currents by about 50% (Figure 6). Currents were elicited from a holding potential of −106 mV and we measured the time to peak and time constant of channel inactivation from an open state. Data from test potentials of −46 mV (approximately 50% of channels active) and −26 mV (all channels open) are illustrated in (Figure 6). NAGly did not affect time to peak or inactivation from an open state at either potential, while NADA modestly accelerated inactivation from an open state at −26 mV (P < 0.05, Figure 6). AEA significantly accelerated open-state inactivation kinetics of Ca_V3.1 channels at both potentials (P < 0.01), and also decreased the time to peak at −46 mV, consistent with the previously reported effects of AEA on Ca_V3.1 (Chemin et al., 2001). NADA did not affect time to peak of Ca_V3.2 or Ca_V3.3 currents evoked from −106 mV. NADA or NAGly did not affect the time to peak or deactivation time constant of currents evoked from a holding potential of −86 to a test potential of −26 mV in any Ca_V3 channel (Table 1).

Figure 6.

N-arachidonoyl dopamine (NADA) and N-arachidonoyl glycine (NAGly) do not strongly affect Ca_V3.1 channel kinetics. Whole-cell patch clamp recordings were made from human Ca_V3 channels stably expressed in HEK 293 cells. Channel activation from a holding potential of −106 mV was measured 5 min after breaking into the cell and then again after 5 min in NADA (500 nmol·L⁻¹), NAGly (10 µmol·L⁻¹) or anandamide (AEA, 300 nmol·L⁻¹). The graphs illustrate: (A) current amplitude at −26 mV, the time to peak and time constant of inactivation at (B) −26 mV and (C) −46 mV. The values in drug are expressed as a percentage of the values at 5 min, control cells were continuously superfused with vehicle alone. Each bar represents the mean ± SEM of at least six cells. Statistical comparisons were made with control cells recorded on the same day (Student's t-test, *P < 0.05, **P < 0.01). Panel (D) illustrates typical currents elicited by a step from −106 mV to −26 mV recorded in control conditions and after 5 min in NADA, AEA and NAGly. Currents have been normalized to the peak inward current to allow ready comparison of inactivation kinetics.

Table 1.

The effects of NADA and NAGly on the kinetics of Ca_V3 channels

CaV channel	Time to peak (ms)			Deactivation τ (ms)
	Control	NADA	NAGly	Control	NADA	NAGly
3.1	7.2 ± 0.3	6.2 ± 0.3	7.3 ± 0.3	3.2 ± 0.1	3.1 ± 0.1	3.2 ± 0.2
3.2	11.2 ± 0.4	11.5 ± 0.6	11.6 ± 0.4	3.1 ± 0.2	3.1 ± 0.2	3.1 ± 0.2
3.3	43 ± 1	44 ± 2	43 ± 2	2.4 ± 0.1	2.4 ± 0.2	2.3 ± 0.1

Cells expressing recombinant Ca_V3 channels were voltage clamped at −86 mV and then stepped to −26 mV. For examples of these experiments see Figures 2 and 3. The time to peak was measured directly and the decay of the current following repolarization of the membrane to −86 mV fit with a single exponential function. The concentration of NADA was 300 nmol·L⁻¹ for Ca_V3.1 and Ca_V3.3 and 1 µmol·L⁻¹ for Ca_V3.2. 10 µmol·L⁻¹ NAGly was used for each channel. There were no differences in time to peak or deactivation for any current with either drug (paired t-test vs. predrug values for each cell). n = 6–8 for each condition.

NADA, N-arachidonoyl dopamine; NAGly, N-arachidonoyl glycine.

Effects of NADA and NAGly on channel activation and inactivation

AEA inhibits Ca_V3 channels in part by increasing steady-state inactivation and thus reducing the numbers of channels available to open during a depolarization. We examined whether the inhibition of Ca_V3 channels by NADA or NAGly could be due to effects on channel availability or activation. Activation curves were constructed by stepping cells from −106 mV to potentials between −86 and +59 mV, and then were repeated after 5 min perfusion of approximately EC₅₀ concentrations of NADA (Figure 7), NAGly (Figure 8) and AEA (not shown). In the presence of the compounds there were small (2–3 mV) shifts in the potential at which half the channels were activated, these shifts were not different from those seen with time-matched vehicle controls (Table 2).

Figure 7.

N-arachidonoyl dopamine (NADA) affects steady-state inactivation but not activation of Ca_V3 channels. Whole-cell patch clamp recordings were made from human Ca_V3 channels stably expressed in HEK 293 cells, 5 min after breaking into the cell and then again after 5 min in NADA. To measure channel activation, cells were voltage clamped at −106 mV and stepped to potentials above −86 mV in 5 mV increments. To measure steady-state inactivation, cells were voltage clamped for 5 s at potentials between −126 mV and −46 mV, and then stepped to a test potential of 26 mV. The peak current at each test potential is plotted for activation curves, the current at −26 mV following 5 s at the indicated holding potential is plotted for inactivation. Curves are a Boltzmann fit of the data (see Methods). NADA did not affect activation, but produced a significant hyperpolarizing shift in the membrane at which 50% of the channels are inactivated for each Ca_V3 subtype (Table 2).

Figure 8.

N-arachidonoyl glycine (NAGly) affects steady-state inactivation but not activation of Ca_V3.1. Whole-cell patch clamp recordings were made from human Ca_V3.1 channels stably expressed in HEK 293 cells, 5 min after breaking into the cell and then again after 5 min in NAGly (10 µmol·L⁻¹). To measure channel activation, cells were voltage clamped at −106 mV and stepped to potentials above −86 mV in 5 mV increments. To measure steady-state inactivation, cells were voltage clamped for 5 s at potentials between −126 mV and −46 mV, and then stepped to a test potential of −26 mV. The peak current at each test potential is plotted for the activation curve, the current at −26 mV following 5 s at the indicated holding potential is plotted for inactivation. Curves are a Boltzmann fit of the data (see Methods). NAGly did not affect activation, but produced a significant hyperpolarizing shift in the membrane at which 50% of the Ca_V3.1 channels are inactivated (Table 2).

Table 2.

The effects of NADA, NAGly and anandamide (AEA) on the parameters of steady-state activation and inactivation of Ca_V3 channels

Drug	CaV channel	Channel V0.5 (mV)
		Activation	Inactivation
300 nmol·L−1 NADA	3.1	−3 ± 0.3	−11 ± 1**
1 µmol·L−1 NADA	3.2	−2.5 ± 2	−10 ± 3**
300 nmol·L−1 NADA	3.3	−1 ± 0.3	−10 ± 1**
10 µmol·L−1 NAGly	3.1	1 ± 1	−8.5 ± 1**
300 nmol·L−1 AEA	3.1	−2.7 ± 0.5	ND
No drug	3.1	−2 ± 1	−2 ± 3
No drug	3.2	1 ± 2	−2 ± 2
No drug	3.3	−2 ± 2	−2 ± 2

Cells expressing recombinant Ca_V3 channels were voltage clamped at −106 mV and then stepped to potentials above −86 mV (activation) or stepped for 5 s to potentials between −126 and −36 mV before stepping to the test potential of −26 mV. The resulting peak currents were fitted to a Boltzmann equation. Changes in the voltage for half activation/inactivation (V_0.5) of the curves are reported. The “No drug” values represent time-dependent changes under our recording conditions. Curves for NADA and NAGly are illustrated in Figures 7 and 8. NADA, N-arachidonoyl dopamine; NAGly, N-arachidonoyl glycine; ND, not determined.

*P < 0.05,

^**P < 0.01 from control.

Steady-state inactivation was determined by holding cells at −106 mV and then stepping them for 5 s to test potentials between −126 mV and −51 mV before measuring the current following a step to −26 mV. This was repeated after 5 min in approximately EC₅₀ concentrations of drug. NADA produced a significant hyperpolarizing shift in the membrane potential at which 50% of the channels were available for activation for all channels examined (Figure 7, Table 2). A similar effect was observed for NAGly (10 µmol·L⁻¹) on Ca_V3.1 (Figure 8, Table 2). The shifts in steady-state inactivation in cells exposed to vehicle alone for 5 min were less than 2 mV (Table 2). The increase in steady-state inactivation is likely to make a major contribution to the inhibition of Ca_V3 channel currents by NADA and NAGly.

In small to medium sized mouse trigeminal ganglion neurons, T-type I_Ca and capsaicin responses are largely segregated into two different populations of neurons (Borgland et al., 2001; Roberts et al., 2002). Type 1 cells are usually sensitive to capsaicin but do not express T-type I_Ca while Type 2 cells express prominent T-type I_Ca but are insensitive to capsaicin. Superfusion of NADA (300 nmol·L⁻¹) inhibited the I_Ca evoked by stepping Type 2 neurons from −86 mV to −46 mV by 30 ± 7% (Figure 9). NADA alone did not produce an inward current in Type 2 cells. NADA is also an agonist at TRPV1 receptors, and superfusion of NADA (300 nmol·L⁻¹) produced modest inward currents in capsaicin-sensitive Type 1 cells, but not Type 1 cells insensitive to capsaicin (n = 6) (Figure 9). The NADA currents were 10 ± 6% of the size of the current produced by a subsequent application of high concentration of capsaicin (10 µmol·L⁻¹, mean current 1.5 ± 0.5 nA, n = 7). NAGly (10 µmol·L⁻¹) did not activate currents in capsaicin-sensitive Type 1 cells at room temperature (n = 6) and only produced a small inward current at 33°C in capsaicin-sensitive Type 1 cells (16 ± 8 pA, n = 6).

Figure 9.

N-arachidonoyl dopamine (NADA) inhibits native T-type calcium channels and activates TRPV1. Whole-cell patch clamp recordings were made from acutely isolated mouse trigeminal ganglion neurons. Currents were evoked by stepping from −86 mV to −46 mV. (A) Time plot of the peak amplitude of the I_Ca at −46 mV, illustrating the effect of NADA (perfused for the duration of the bar). (B) Typical traces from the same cell as above in (A), in the absence and presence of NADA. (C) Example trace from a Type 1 sensory neuron voltage clamped at −60 mV and superfused with NADA and a maximally effective concentration of the TRPV1 agonist capsaicin. NADA (300 nmol·L⁻¹) produced an inward current of, on average, 10 ± 6% of the capsaicin (10 µmol·L⁻¹) current. The dotted line represents zero current. These traces are not leak subtracted. I_Ca, voltage-gated calcium channel current; TRPV1, transient receptor potential vanilloid receptor 1.

Discussion

The major finding of this study is that NADA inhibited T-type calcium channels with a similar potency to its agonist actions at native CB₁ or TRPV1 receptors and with a similar or greater potency to the inhibitory effects of the prototypic endocannabinoid AEA at Ca_V3 channels (Chemin et al., 2001). At concentrations of 300 nmol·L⁻¹ to 1 µmol·L⁻¹, NADA strongly inhibited native and recombinant T-type calcium channels, while other studies have reported EC₅₀ values of approximately 1 µmol·L⁻¹ for NADA activation of native TRPV1 (Huang et al., 2002, Price et al., 2004) and CB₁ receptors (Bisogno, et al., 2000). By contrast, NAGly, the prototypic arachidonoyl amino acid and close structural analogue of AEA, was a much weaker inhibitor of T-type calcium channels than AEA or NADA. Interestingly, unlike AEA or NADA, NAGly is also not a CB₁ receptor ligand (Sheskin et al., 1997), has little activity at TRPV1 (De Petrocellis et al., 2000) and does not activate a recently described AEA-stimulated current in mouse trigeminal neurons (Roberts et al., 2008).

T-type calcium channels are involved in a wide range of physiological processes (Perez-Reyes, 2003, Shin et al., 2008), including many that are also modulated by activation of CB or TRPV1 receptors. For several reasons, it is difficult to assess the contribution of NADA modulation of T-type channels to the known effects of NADA administered to animals. First, the role of T-type channels cannot be readily dissected pharmacologically because of the lack of inhibitors selective for T-type channels over other voltage-gated calcium channels (Heady et al., 2001). Conversely, while there are neurobehavioural assays of nociception, sleep and seizure activity sensitive to known T-type channel inhibitors (Perez-Reyes, 2003; Shin et al., 2008), none of these assays are specific assays for T-type channel blockers. Second, many of the drugs used to antagonize the effects of NADA at CB₁ receptors or TRPV1 also inhibit T-channels at pharmacologically relevant concentrations. These include the CB₁ antagonists SR 141716A (Chemin et al., 2001) and AM251 (Ross et al., 2008) and the TRPV1 antagonist capsazepine, (Docherty et al., 1997; Connor, unpublished observations). Thus, while the reversal of a NADA effect by any of these agents provides good evidence for the involvement of the cognate protein, pre-treatment with these ligands will occlude any effects of subsequently administered NADA mediated by actions on T-type channels.

The effects of NADA and NAGly on Ca_V3 channel activity shared some but not all of the hallmarks of AEA modulation of the channel. All three ligands produce strong hyperpolarizing shifts in the membrane potential at which Ca_V3 channels inactivate, and this would have the effect of reducing the number of channels available to open from all but the most negative membrane potentials. This is a common mechanism for modulation of Ca_V3 channels, and is seen with arachidonic acid (Zhang et al., 2000; Talavera et al., 2004; Chemin et al., 2007) and cannabinoid ligands such as Δ⁹-tetrahydrocannabinol and cannabidiol (Ross et al., 2008). However, AEA (and arachidonic acid) both have effects on the kinetics of Ca_V3 channels, manifested as an acceleration of channel opening and open-state inactivation, with this latter effect further limiting calcium entry through Ca_V3 channels. Neither NADA nor NAGly strongly affected channel opening or open-state inactivation, which is similar to the effect of cannabidiol on Ca_V3 channels (Ross et al., 2008). NADA inhibition of Ca_V3 channels, particularly Ca_V3.3, was more potent than cannabidiol (EC₅₀ approximately 800 nmol·L⁻¹ for Ca_V3.1 and Ca_V3.2, 4 µmol·L⁻¹ for Ca_V3.3) and Δ⁹-tetrahydrocannabinol (EC₅₀ approximately 1 µmol·L⁻¹ for Ca_V3.1 and Ca_V3.2, 4 µmol·L⁻¹ for Ca_V3.3) recorded in identical conditions, (Ross et al., 2008) while that of NAGly was similar to that of arachidonic acid (Chemin et al., 2007). The rank order of potency for NADA effects on Ca_V3 channels (3.3 ≥ 3.1 > 3.2) is also distinct from that reported for AEA (3.2 > 3.3 > 3.1) (Chemin et al., 2001) and cannabidiol/Δ⁹-tetrahydrocannabinol (3.1 = 3.2 > 3.3) (Ross et al., 2008).

It is likely, but not proven, that NADA and NAGly were acting directly on the Ca_V3 channels. NADA has not been reported to be a ligand for any G protein coupled receptor other than the CB₁ receptor and, while HEK 293 cells express mRNA for the NAGly-activated GPR 18 (Kohno et al., 2006, Johnson and Connor, unpublished observations), the inhibition of Ca_V3.1 by either compound was not sensitive to the non-specific inhibitor of G protein activation, GDPβS. Receptor- or second messenger-dependent modulation of T-type calcium channels is complex (Perez-Reyes, 2003) but we are not aware of any mechanisms described that produce inhibitory effects on all three isoforms of Ca_V3 channels that are similar to those observed in this study (Hildebrand et al., 2007; Iftinca et al., 2007; Tao et al., 2008).

The site(s) of action for fatty acids and their derivatives modulating Ca_V3 channels is not known and there is limited information about where drugs that modulate Ca_V3 channels could bind to affect channel function. Unsaturated fatty acids seem to be relatively non-selective inhibitors of Ca_V3 channels, and a hyperpolarizing shift in channel inactivation potential is commonly observed with lipid-soluble modulators of Ca_V3 (Heady et al., 2001; Ross et al., 2008) and other voltage-dependent channels (Lundbaek 2008). Amphiphiles which act in the plasma membrane to increase bilayer elasticity produce hyperpolarizing shifts in the inactivation of N-type calcium channels and voltage-gated sodium channels (Lundbaek et al., 1996; 2004), reminiscent of the effects of AEA, NADA and NAGly. The partial reversibility of NADA, NAGly and AEA (Chemin et al., 2001) are also consistent with an interaction mediated through the plasma membrane. However, the effects of AEA on bilayer lipid dynamics have been suggested to be minimal (Tian et al., 2005) and those of NADA and NAGly are unknown. Further evidence inconsistent with a major effect of NADA and NAGly on lipid elasticity, at the concentrations that inhibit Ca_V3 channels, comes from a study showing that NADA (10 µmol·L⁻¹) had no effect on native N-type calcium channels while NAGly (10 µmol·L⁻¹) enhanced N-channel activation (Guo et al., 2008). The differences in the absolute potency (up to 30-fold), differences in rank order of potency at Ca_V3 subtypes and the distinct effects on Ca_V3.1 kinetics of iso-inhibitory concentrations of AEA, NADA and NAGly suggest that a specific site mediates the effects of these compounds. This site seems to have a sufficiently well defined structure to be sensitive to the nature of the head group of the compounds and we suggest is likely to be part of the Ca_V3 channel itself.

In this study, we have shown that endogenous acyl amides are powerful inhibitors of T-type calcium channels, with potencies similar to that previously reported for their major sites of action, CB₁ receptors and TRPV1 ion channels. T-type calcium channels, CB₁ receptors and TRPV1 are often co-expressed, and thus at concentrations previously shown to modulate peripheral nociception (Chu et al., 2003; Sagar et al., 2004; Huang and Walker, 2006) and synaptic transmission (Marinelli et al., 2007), NADA is also likely to be strongly affecting T-type calcium channels in these tissues. However, confirmation that T-type calcium channel modulation contributes to the physiological or pharmacological effects of NADA and NAGly awaits the development of more selective T-type calcium channel modulators, or studies in Ca_V3 knockout animals.

Glossary

Abbreviations:

FAAH

fatty acid amide hydrolase

I_Ca

voltage-gated calcium channel current

NADA

N-arachidonoyl dopamine

NAGly

N-arachidonoyl glycine

TRPV1

transient receptor potential vanilloid receptor 1

Conflict of interest

None.