Effects of arachidonic acid on unitary calcium currents in rat sympathetic neurons

Abstract

1. We have characterized the actions of arachidonic acid (AA) on whole cell and unitary calcium (Ca²⁺) currents in rat neonatal superior cervical ganglion (SCG) neurons using barium (Ba²⁺) as the charge carrier.

2. Whole cell currents were elicited by stepping the membrane potential from −90 mV to +10 mV. Arachidonic acid (5 μm) was introduced into the bath in the continued presence of 1 μm (+)-202-791, an L-type Ca²⁺ channel agonist. Under these conditions, the peak current, comprised mainly of N-type current, and a slow, (+)-202-791-induced component of the tail current were inhibited by 67 ± 6 and 60 ± 10%, respectively, indicating that AA inhibits both N- and L-type currents.

3. At a test potential of +30 mV, AA (5 μm) decreased unitary L- and N-type Ca²⁺ channel open probability (P_o) in cell-attached patches that contained a single channel. For both channels, the underlying causes of the decrease in P_o were similar. Arachidonic acid caused an increase in the percentage of null sweeps and in the number of null sweeps that clustered together. In sweeps with activity, the average number of openings per sweep decreased, while first latency and mean closed time increased. Arachidonic acid had no significant effect on unitary current amplitude or mean open time.

4. Our findings are the first description of the inhibition of unitary L- and N-type Ca²⁺ channel activity by AA and are consistent with both channels spending more time in their null mode and with increased dwell time in one or more closed states.

Arachidonic acid (AA, C20:4 n-6), a polyunsaturated fatty acid, appears to play an important role in normal intracellular and/or intercellular signal transduction in the nervous system. It may also play a critical role in certain pathological conditions (see review by Bazan, 1989). A number of neurotransmitters stimulate the release of AA from nerve cell membranes (see review by Katsuki & Okuda, 1995). Once liberated, a major effect of free AA appears to be the modulation of ion channels, including voltage-gated ion channels (sodium, potassium, calcium and chloride channels), ligand-gated ion channels (GABA, NMDA and nicotinic receptors), intracellular calcium (Ca²⁺) release channels and gap junctions (Miller et al. 1992; Sakai et al. 1992; Bouzat & Barrantes, 1993; Striggow & Ehrlich, 1997; Weingart & Bukauskas, 1998; see reviews by Ordway et al. 1991; Meves, 1994; Katsuki & Okuda, 1995). Arachidonic acid is also a candidate retrograde messenger in long-term depression (Bolshakov & Siegelbaum, 1995) and potentiation (Williams et al. 1989; O'Dell et al. 1991). Thus, AA may play a functional role in coordinating ion channel modulation with nerve cell excitability.

We are particularly interested in the effects of AA on Ca²⁺ channel activity in neurons because Ca²⁺ influx through a variety of types of Ca²⁺ channels coordinates electrical activity with many cellular processes, such as enzyme activation, neurotransmitter release and gene expression (see review by Berridge, 1998). Bath application of micromolar amounts of AA inhibits whole cell Ca²⁺ currents in nerve cell lines (Schmitt & Meves, 1995), carotid body nerve cells (Hatton & Peers, 1998), sensory (Piomelli et al. 1987; Khurana & Bennett, 1993), sympathetic (Bug et al. 1989) and central neurons (Keyser & Alger, 1990), indicating that AA has the capacity to modulate neuronal Ca²⁺ channel activity. However, the types of Ca²⁺ currents inhibited by AA have not been determined. Furthermore, direct evidence of the effects of AA on Ca²⁺ channel activity at the single channel level has been lacking.

To begin to determine which Ca²⁺ channel types are sensitive to AA and to characterize in detail its effects on neuronal Ca²⁺ currents, we investigated whether AA could modulate L- and N-type currents in neonatal rat superior cervical ganglion (SCG) neurons. These neurons possess both L- and N-type Ca²⁺ channels and at least 80 % of the whole cell peak current in these cells is carried through N-type Ca²⁺ channels (Plummer et al. 1989; Regan et al. 1991). Moreover, SCG neurons have been used widely as a model system for studying the modulation of nerve cell excitability (Adams et al. 1986; Plummer et al. 1991; see review by Hille, 1994). We have found that application of AA to the bath inhibits both L- and N-type whole cell currents. To determine how AA alters L- and N-type Ca²⁺ channel activity, we characterized the actions of AA on unitary L- and N-type Ca²⁺ channel activity recorded in the cell-attached patch configuration.

METHODS

Cell preparation

Sympathetic neurons were obtained from SCG of neonatal Sprague-Dawley rats (1-3 days old) following decapitation. Ganglia were removed and mechanically dissociated (Hawrot & Patterson, 1979), producing cells that were free of processes. Neurons were plated on poly-L-lysine coated glass coverslips and cultured in Dulbecco's modified Eagle's medium that was supplemented with 7.5 % calf serum, 7.5 % fetal bovine serum, 4 mM glutamine, 100 IU ml⁻¹ penicillin, 100 μg ml⁻¹ streptomycin (all purchased from Sigma, St Louis, MO, USA) and 0.2 μg ml⁻¹ nerve growth factor (Bioproducts for Science, Indianapolis, IN, USA). Cells were maintained at 37.5°C in a 5 % CO₂ incubator. Cells were used for whole cell experiments within 24 h to avoid recording from cells with processes. For cell-attached patch experiments, SCG neurons were used within 24–48 h. Animals were cared for and killed according to a University of Massachusetts Medical School approved protocol in compliance with the scientist-related provisions of the Federal Animal Welfare Act, and the US Public Health Service ‘Policy on Human Care and Use of Animals’.

Current recordings

Whole cell membrane currents were obtained with standard patch clamp techniques (Hamill et al. 1981). The electrode current was zeroed prior to sealing. Currents were recorded at room temperature (20-24°C) with a Dagan 3900 (Dagan Corp., Minneapolis, MN, USA) or Axon 200A (Axon Instruments, Foster City, CA, USA) patch clamp amplifier. Capacitive transients were compensated by at least 70 %. Sweeps were low-pass filtered at 5 kHz using the four-pole Bessel filter of the patch clamp amplifier and digitized at 20 kHz unless stated otherwise. Current traces were stored on the hard drive of a PDP-11 computer or Pentium PC. Data were acquired either using custom software written for the PDP-11 or the Patch software suite (Cambridge Electronic Design, Cambridge, UK). Electrodes were pulled from borosilicate glass capillaries (Drummond Scientific Company, Broomall, PA, USA), coated with Sylgard (DOW Corning Corporation, Midland, MI, USA) to reduce capacitance and electrical noise, and fire-polished to a tip diameter of ∼1 μm. Pipette resistance varied from 2–2.5 MΩ. Following the method of Plummer et al. (1991), the membrane potential was held at −90 mV and stepped to +10 mV for 40 ms, then stepped back to an intermediate negative potential of −50 mV for 50 ms, except where noted. Test pulses were delivered at 4 s intervals. Drug introduction and changes in the bath solution were accomplished by superfusion.

Currents from cell-attached patches were recorded at room temperature with an Axopatch 200A patch-clamp amplifier by holding the patch at −90 mV and stepping to +30 mV for 700 ms, every 4 s. Data were acquired using the Patch software suite. Currents were low-pass filtered at 1 kHz (four-pole Bessel filter) and sampled at 5 kHz. Electrodes were similar to those for whole cell experiments except that the tip of each electrode was fire-polished to give a pipette resistance that ranged from 3–5 MΩ. The electrode current was zeroed prior to sealing. Cells were exposed to AA by lowering a puffer pipette containing 5 μM AA into the bath solution within 20 μm of the cell. Due to the large diameter of the puffer pipette and the positive pressure (100 kPa) applied to it with a Picospritzer II (General Valve Corp., Fairfield, NJ, USA) prior to entering the bath, the AA solution freely flowed out of the pipette upon immersion; no additional pressure was necessary. Bovine serum albumin (BSA) was introduced into the chamber by superfusion.

Solutions and chemicals

For measuring whole cell currents, the external solution contained (mM): 20 barium acetate, 125 N-methyl-D-glucamine, 10 Hepes and 0.001 tetrodotoxin (293 mosmol l⁻¹). The pipette solution was composed of (mM): 123 cesium aspartate, 10 EGTA, 10 Hepes, 5 MgCl₂, 4 ATP (296 mosmol l⁻¹). The pH of both solutions was adjusted to 7.5 with CsOH. Where noted, 0.4 mM guanosine triphosphate (GTP) was included in the pipette solution.

For cell-attached patch recordings, the pipette solution contained (mM): 110 BaCl₂ and 10 Hepes. The pH was adjusted to 7.5 with tetraethylammonium hydroxide. The membrane potential of the cell outside the patch was zeroed by bathing the cells in a solution of (mM) 140 potassium aspartate, 10 Hepes, 5 EGTA and 500 nM (+)-202-791. The pH was adjusted to 7.5 with KOH.

Arachidonic acid (NuCheck, Minneapolis, MN, USA) and (+)-202-791 (a gift from Dr Hoz, Sandoz, Switzerland) were prepared from stock solutions made up in 100 % ethanol and diluted with bath solution to a final ethanol concentration of less than 0.17 %. This maximal final concentration of ethanol had no significant effect on Ca²⁺ currents (data not shown). Stock solutions of AA were kept in sealed glass vials under nitrogen at −90°C. Bovine serum albumin (BSA, essentially fatty acid free, Sigma) was added directly to the bath solution. Tetrodotoxin (Peninsula Laboratories, Inc., Belmont, CA, USA or Sigma) was made up in double distilled water.

Data analysis

Prior to analysis, whole cell currents were corrected for leak currents by adding the average current from 10 current traces that were generated with a test pulse to −110 mV and scaled accordingly to the whole cell test pulse amplitude. The amplitude of whole cell currents was measured as the average of 10 sampling points around a cursor set at 15 or 39 ms after the onset of the test pulse (peak current), unless otherwise indicated. In some experiments, currents were measured approximately 12 ms after a repolarization step to an intermediate potential, defined as the slow component of the tail current.

For cell-attached patch recordings, an average current, created from 5–15 sweeps that contained no openings (null sweeps), was subtracted from all sweeps to remove leak current and any remaining capacitive current. Unitary transitions between open and closed states were detected by setting the threshold at 50 % of the open channel level. Changes in currents due to AA were calculated by starting the analysis at the time the puffer pipette had reached its final position near the cell. Analysis of unitary currents was obtained by characterizing the channel activity in at least 40 consecutive, leak-subtracted sweeps. The number of sweeps was matched for control and AA. Cumulative first latency plots were generated using Excel (Microsoft, Seattle, WA, USA) and Origin (Microcal Software, Inc., Northampton, MA, USA). All-points amplitude histograms, mean ensemble currents, open probability (P_o), and arithmetic means of first latencies, open and closed times were generated with the Patch software suite. Changes in P_oversus time were calculated by determining the average P_o during each 700 ms test pulse to +30 mV. To calculate the time to observe an effect of AA, the time between the positioning of the puffer pipette and the first sweep with a decreased P_o that also showed an obvious change in gating properties was calculated.

Statistical significance was determined in Origin using Student's two-tailed, paired t test unless stated otherwise, to determine whether the difference between two mean values was significant, defined as equal to, or less than 0.05. Data are expressed as means ± standard error of the mean.

RESULTS

Arachidonic acid inhibits whole cell Ba²⁺ currents in neonatal SCG neurons

To determine whether AA inhibits both L- and N-type Ca²⁺ channel activity in SCG neurons, whole cell Ba²⁺ currents were studied in the presence of the dihydropyridine agonist (+)-202-791, an L-type Ca²⁺ channel agonist that elicits long-lasting channel openings (Plummer et al. 1989). Application of 1 μM (+)-202-791 to the bath increased the peak current only marginally (Fig. 1), consistent with the relatively low contribution of L-type Ca²⁺ channel activity to the whole cell peak current (Plummer et al. 1991). When the membrane potential was stepped from the test potential of +10 mV to an intermediate tail potential of −50 mV in the presence of (+)-202-791, a slow component of the tail current followed the typical fast deactivating component. An example of this is shown in Fig. 1A. These recording conditions allowed us to examine the effects of AA on the peak current, which was composed primarily of N-type current, and on the slow component of the tail current, made up entirely of L-type current.

Figure 1. Bath application of arachidonic acid (AA) inhibits both whole cell L- and N-type Ba²⁺ currents in SCG neurons.

A, peak (upper) and slow component of the tail current (lower) amplitude vs. time. In the continued presence of the dihydropyridine agonist (+)-202-791 ((+)-202; 1 μM), application of 5 μM AA to the bath decreased both the peak current, primarily made up of N-type current, and the slow component of the tail current, made up entirely of L-type current. Continuous bars indicate periods of exposure of the cell to different compounds. Inset: sweeps taken at the times indicated by the lower case letters. Dashed line indicates zero current level. B, summary of the effects of AA on the peak and the slow component of the tail current. (+)-202-791 increased the peak current 1.23 ± 0.54-fold and the slow component of the tail current 17.6 ± 5.7-fold. Arachidonic acid (5 μM) in the continued presence of (+)-202-791 significantly (*P < 0.01) reduced both the peak and slow component of the tail current by 67 ± 6 and 60 ± 10 %, respectively (n = 5 cells). C, mean current-voltage relationships in the presence (•) and absence (▪) of 5 μM AA (n = 5 cells). (+)-202-791 (1 μM) was present in the bath throughout each experiment. D, upper traces, AA (5 μM) inhibits whole cell currents generated with 700 ms test pulses in the absence of (+)-202-791 (Control). Lower trace, subtraction of the inhibited current from the control current revealed the current inhibited by AA was non-inactivating.

In the continued presence of 1 μM (+)-202-791, superfusion of 5 μM AA significantly inhibited both the peak and the slow component of the tail current, when measured approximately 15 min after the application of AA to the bath (Fig. 1A and B). From the current-voltage relationship in the presence of (+)-202-791 (Fig. 1C), it appears that at negative voltages, current inhibition by AA was minimal whereas significant inhibition of inward and outward current occurred at a number of test potentials greater than 0 mV (P < 0.05). These results suggest that the inhibition of the whole cell peak current by 5 μM AA is voltage dependent. To examine the kinetics of the inhibited component, currents in the presence of AA were subtracted from control currents (Fig. 1D). The inhibited component of current was non-inactivating in 4 out of 5 recordings.

We performed two experiments to rule out certain potential non-specific effects of AA. One concern was that AA inhibited whole cell currents simply by disrupting the binding of (+)-202-791 to L-type Ca²⁺ channels. To verify that the actions of AA resulted in changes in the channels themselves, rather than by displacing (+)-202-791, the effect of AA on whole cell currents was tested in the absence of (+)-202-791 (Fig. 2A). Under these conditions, application of 5 μM AA to the bath was still able to inhibit the peak current by 48 ± 6 % (n = 5). Similar inhibition was also observed with longer test pulses (Fig. 1D). A second concern was that AA might somehow irreversibly disrupt the membrane, leading to a non-specific decrease in whole cell Ba²⁺ currents. However, this does not appear to be the case. Though difficult to wash out normally (data not shown), the inhibition of both the peak and the slow component of the tail current by AA was reversible in 5 out of 5 cells tested when BSA (essentially fatty acid free) was included in the wash solution (Fig. 2B). BSA binds free AA in the bath, effectively decreasing the external concentration and maintaining a steep concentration gradient for AA to diffuse from the cell into the external solution (Spector, 1975; Schmitt & Meves, 1994). Taken together the results from these whole cell patch clamp experiments suggest that AA inhibits both whole cell L- and N-type currents in neonatal SCG neurons.

Figure 2. The inhibitory effects of AA appear to be independent of (+)-202-791 and reversible.

A, whole cell currents generated with 20 ms test pulses were inhibited by 48 ± 6 % (n = 5) by AA, in the absence of the dihydropyridine agonist. (Sampling rate, 5 kHz.) B, in the continued presence of 1 μM (+)-202-791 ((+)-202), the effects of AA could be reversed by including bovine serum albumin (BSA, 1.0 mg ml⁻¹) in the bath solution.

Arachidonic acid inhibits unitary L-type Ca²⁺ channel activity

We used the cell-attached patch configuration to confirm the whole cell results and to examine how AA decreased unitary currents in patches containing a single L- or N-type Ca²⁺ channel. A patch was considered to have only one channel if the all-points amplitude histogram showed no current greater than −1.2 pA. At this amplitude simultaneous channel openings should become apparent. Also no superimposed openings were observed in at least 120 consecutive sweeps. Differences in their properties were used to distinguish single L-type from N-type Ca²⁺ channel activity. In the presence of the dihydropyridine agonist (+)-202-791, two characteristics were especially useful for identifying unitary L-type Ca²⁺ channel activity. First, L-type Ca²⁺ channels displayed long-lasting channel openings approximately −0.74 pA in amplitude, many of which were at least 8 ms in duration (Plummer et al. 1989). Second, unitary tail currents were observed in some sweeps due to these long openings remaining during the repolarizing step back to the holding potential (Plummer et al. 1989). Examples of dihydropyridine-induced unitary tail currents can be viewed in the first sweep of both the left and right panels of Fig. 3A. In contrast, N-type Ca²⁺ channel activity was characterized by bursts of smaller amplitude openings of shorter duration (0.4-3.0 ms; Plummer et al. 1989).

Figure 3. Unitary L-type Ca²⁺ channel activity is inhibited by AA.

A, cell-attached recordings from a patch containing one L-type Ca²⁺ channel. The left panel (Control) shows six representative consecutive sweeps of L-type Ca²⁺ channel activity in the presence of (+)-202-791 (500 nM). Under these conditions, long-lasting channel openings with an amplitude of −0.8 pA were prominent. The right panel (AA) shows that after 5 μM AA was puffed into the bath in the continued presence of (+)-202-791, channel activity decreased. Dashed lines indicate zero current level. Channel closings that do not return to the zero current level appear to be due to small changes in the leak. Under control conditions (left) and in the presence of AA (right) mean ensemble currents (B), and all-points amplitude histograms (C) were each constructed from 88 consecutive sweeps from the same recording as in A.

Unitary currents in patches containing only one L-type Ca²⁺ channel were analysed under control conditions and after the introduction of AA. An example of the effects of AA on consecutive sweeps of single L-type Ca²⁺ channel activity is illustrated in Fig. 3. Under control conditions, channel openings at +30 mV were fairly evenly distributed throughout the duration of the 700 ms long test pulse. When 5 μM AA was puffed onto the cell in the continued presence of 500 nM (+)-202-791, unitary channel activity decreased. The presence of AA decreased the amplitude of mean ensemble currents 75.6 ± 7.8 % (n = 6), an example of which can be observed in Fig. 3B. This also can be seen in the all-points amplitude histogram which shows that AA decreased the time the channel spent in the open state (Fig. 3C). In addition, the amplitude histogram and direct measurement of unitary currents indicated that AA did not significantly affect the unitary L-type Ca²⁺ channel amplitude (Table 1), compared with the control group (P > 0.05).

Table 1. Biophysical characteristics of unitary L-type Ca²⁺ channel currents.

Indexes	Control	Arachidonic acid
Null sweeps/recording (%)	16.2 ± 2.3	50.7 ± 10.2†
Null sweeps/cluster	1.6 ± 0.1	4.6 ± 0.8*
Mean number of openings/sweep in sweeps with activity	56.4 ± 15.0	10.8 ± 2.6*
Mean unitary current amplitude (pA)	0.74 ± 0.02	0.73 ± 0.02
Mean open time (ms)	4.77 ± 1.34	4.63 ± 1.42
Mean closed time without last closing (ms)	3.7 ± 1.40	13.1 ± 4.9
Mean closed time with last closing (ms)	11.8 ± 4.5	35.5 ± 7.1†
Mean first latency (ms)	30.5 ± 5.5	91.3 ± 8.1†

Results from 6 patches.

^*P < 0.05

^†P < 0.01.

Since no change in unitary current amplitude was observed, the decrease in the ensemble currents due to AA was most likely due to a decrease in P_o. Indeed, after application of AA, there was a delay of approximately 90 ± 18 s (n = 6 patches) before the P_o precipitously decreased and then remained low in the continued presence of AA (Fig. 4A). The average P_o from six single L-type Ca²⁺ channel patches significantly decreased (Fig. 4B; P < 0.05) in the presence of AA (0.24 ± 0.06), compared with control (0.53 ± 0.06). A large part of this decrease could be accounted for by an increase in the percentage of null sweeps per recording (Table 1). In addition, these null sweeps clustered together. Clustering of sweeps with no activity has been described previously as mode 0 behaviour (Hess et al. 1984). When null sweeps were observed in control recordings, on average 1–2 null sweeps would cluster together before returning to sweeps with activity. After the addition of AA to the bath, the average number of null sweeps found consecutively increased approximately 3-fold (Table 1). These data suggest that AA stabilizes L-type Ca²⁺ channel activity in mode 0.

Figure 4. Arachidonic acid decreases unitary L-type Ca²⁺ channel open probability.

A, open probability (P_o) versus time from the same L-type Ca²⁺ channel shown in Fig. 3. The horizontal bar indicates the recording conditions: Control, 500 nM (+)-202-791; AA, 5 μM AA puffed onto the cell in the continued presence of (+)-202-791. P_o was calculated for each sweep and plotted against time. B, summary of the effects of 5 μM AA on L-type Ca²⁺ channel P_o (n = 6 patches). Arachidonic acid significantly decreased the P_o of single L-type Ca²⁺ channels compared with control P_o (*P < 0.05).

In addition to the increase in null sweeps, one of the prominent aspects of the inhibition of unitary L-type Ca²⁺ channel activity by AA was the large decrease in channel activity in many of the sweeps with activity. This can be seen by comparing the consecutive sweeps, shown in the left and right panels of Fig. 3A, in the plot of P_oversus time shown in Fig. 4A and quantified as a decrease in the number of channel openings per sweep in the sweeps that displayed activity (Table 1). Despite the decrease in the number of openings, the mean open time did not change significantly (Table 1). It was possible that the lack of change in the mean open time was due to a larger number of openings occurring in sweeps at the beginning versus the end of the AA period, thus biasing the mean open time towards control values. To control for this, mean open time was recalculated with the first minute of data collected under control conditions and after application of AA removed (n = 6 patches). Under these conditions, mean open time in the presence of AA (4.51 ± 1.50) was again not significantly different from control (4.95 ± 1.26), consistent with the initial finding that AA does not affect the open state. In addition, this finding suggests that AA does not decrease L-type Ca²⁺ channel activity by displacing (+)-202-791.

To determine whether other changes in unitary channel activity occurred, further kinetic analyses were performed. As the number of openings per sweep decreased, yet the mean open time did not change significantly, we examined whether three other parameters changed: mean closed time, mean closed time when truncated last closings were included, and first latency. To determine whether an increase in closed time contributed to the low activity observed in the presence of AA, mean closed time was calculated by excluding the first latency and truncated last closings. Following the application of AA, the mean closed time increased 3.3 ± 0.9-fold. Surprisingly, this increase in the presence of AA was not significantly different from control (Table 1). Since so little activity occurred in the presence of AA, often only a few openings per sweep, the exclusion of truncated last closings from the analysis greatly underestimated the time the channel spent in closed or inactivated states. Therefore the mean closed time was recalculated with truncated last closings included. The mean closed time of the AA group now was significantly longer than that of the control group (Table 1), reflecting the large decrease in channel activity with AA compared with controls.

The increase in mean closed time when last closings were included, could be due to an increase in the dwell time in one or more closed states or to an increase in dwell time in one or more inactivated states. If a channel has been stabilized in a closed state, and therefore dwelling there longer, an increase in the first latency would be expected because the channel takes longer to transition to the open state. However, if a channel inactivates due to AA, it will recover from inactivation primarily at negative potentials. If a channel is inactivated, stepping to +30 mV will not facilitate recovery and therefore it should not open. On the other hand, if a channel has recovered from inactivation during the repolarization period, it should open normally. Thus, an increase in null sweeps, but no obvious change in first latency in sweeps with activity, would be expected if long truncated closings in the presence of AA were due to channel inactivation.

When, first latencies were calculated, AA was found to increase the mean first latency significantly compared with controls (Table 1), in support of AA stabilizing a closed state. Plots of cumulative first latency verified the decreased rate of opening following a voltage step to +30 mV. In six recordings, the control cumulative first latency plots could be described by a fast time constant (τ₁) of 7.3 ± 2.1 ms and a slow time constant (τ₂) of 97.4 ± 40.3 ms. Both time constants increased in the presence of AA (τ₁ =19.1 ± 6.0 ms, P < 0.05, two-way, paired t test; τ₂ =346.7 ± 141.3 ms, P < 0.05, one-way, paired t test) with no obvious shift in the appearance of short versus long first latencies. Together, these results indicate that in addition to increasing the percentage of null sweeps, AA reduced the number of openings in sweeps with activity by increasing the first latency and mean closed time, when last closings were included.

Arachidonic acid inhibits unitary N-type Ca²⁺ channel activity

Recordings of unitary N-type Ca²⁺ channel activity were made with 500 nM (+)-202-791 in the bath solution to ensure that N-type rather than L-type Ca²⁺ channel activity was recorded. If channel activity showed no dihydropyridine agonist-induced, long-lasting openings or tail currents, and had a unitary current amplitude of approximately −0.69 pA, it was identified as N-type. In addition, these channels almost always displayed modal gating patterns composed of clusters of null, inactivating or non-inactivating sweeps, which have been described previously for unitary N-type Ca²⁺ channel activity in SCG neurons (Plummer & Hess, 1991; Rittenhouse & Hess, 1994).

To determine whether AA inhibits N-type Ca²⁺ channel activity, unitary currents in patches containing a single N-type Ca²⁺ channel were analysed under control conditions and after 5 μM AA was puffed onto the cell. An example of six consecutive sweeps of control activity (Fig. 5A, left panel) shows that in sweeps where the N-type Ca²⁺ channel opened, it did so in bursts, primarily at the onset of the test pulse. Bursts could be short lived, as seen in Fig. 5A, or last the length of the 700 ms test pulse. After the application of 5 μM AA, channel activity decreased (Fig. 5A, right panel). This decrease was also obvious in mean ensemble currents (Fig. 5B) from the same patch as in Fig. 5A. The mean ensemble current amplitude from five single channel recordings was reduced by 58.9 ± 11.4 % after puffing AA onto the cell. The decrease in unitary N-type Ca²⁺ channel activity was also apparent when comparing all-points amplitude histograms, constructed from equal numbers of control sweeps and sweeps recorded in the presence of 5 μM AA. The examples shown in Fig. 5C indicate that the time the N-type Ca²⁺ channel spent in the open state decreased in the presence of 5 μM AA. The mean unitary current amplitude was not significantly different (n = 7 patches, P > 0.05) between the control group (-0.69 ± 0.03 pA) and the AA group (-0.67 ± 0.03 pA; Table 2).

Figure 5. Unitary N-type Ca²⁺ channel activity is inhibited by AA.

A, each panel shows six consecutive leak-subtracted sweeps from the same patch containing one N-type Ca²⁺ channel before (Control) and after 5 μM AA. (+)-202-791 (500 nM) was present in the bath throughout the experiment. Dashed lines indicate zero current level. Under control conditions (left) and in the presence of AA (right) mean ensemble currents (B), and all-points amplitude histograms (C) were each constructed from 103 consecutive sweeps from the same recording as in A.

Table 2. Biophysical characteristics of unitary N-type Ca²⁺ channel currents.

Indexes	Control	Arachidonic acid
Null sweeps/recording (%)	38.0 ± 7.6	59.0 ± 7.0†
Null sweeps/cluster	2.8 ± 0.7	8.3 ± 2.1*
Mean number of openings/sweep in sweeps with activity	52.5 ± 17.5	14.0 ± 6.3*
Mean unitary current amplitude (pA)	0.69 ± 0.03	0.67 ± 0.03
Mean open time (ms)	0.8 ± 0.1	0.7 ± 0.1
Mean closed time without last closing (ms)	2.5 ± 0.3	3.6 ± 0.6
Mean closed time with last closing (ms)	9.4 ± 2.9	25.5 ± 6.9*
Mean first latency (ms)	26.4 ± 14.1	55.9 ± 13.1

Results from 7 patches.

^*P < 0.05

^†P < 0.01.

Consistent with the decrease in mean ensemble current amplitude, P_o was reduced following the application of AA. Plots of P_oversus time were used to examine the time course of the current inhibition. An example of P_oversus time (Fig. 6A) taken from the same single N-type Ca²⁺ channel recording shown in Fig. 5 illustrates the effects of AA on unitary N-type Ca²⁺ channel activity over time. 5 μM AA decreased the P_o of single N-type Ca²⁺ channels after a delay of approximately 83 ± 33 s (n = 7). The average P_o of single N-type Ca²⁺ channel activity decreased significantly (P < 0.05) in the presence of AA (0.12 ± 0.02) compared with control activity (0.23 ± 0.04; Fig. 6B; n = 7 patches). A major part of the decrease in unitary N-type Ca²⁺ channel P_o could be accounted for by an increase in the percentage of null sweeps per recording (Table 2). As with L-type Ca²⁺ channel activity, null sweeps cluster together giving rise to the null mode (Plummer & Hess, 1991). We investigated whether AA stabilized the null mode by determining if the number of null sweeps that formed a cluster increased after puffing AA onto the cell. Similar to the increase observed with L-type Ca²⁺ channels, the number of null sweeps per cluster increased 3.3 ± 0.6-fold (n = 7 patches; Table 2). These data suggest that AA stabilizes the null mode of N-type Ca²⁺ channels.

Figure 6. Arachidonic acid decreases N-type Ca²⁺ channel open probability.

A, P_oversus time obtained from the single N-type Ca²⁺ channel shown in Fig. 5. The bar indicates different recording conditions: Control, 500 nM (+)-202-791; AA, 5 μM AA was puffed onto the cell in the continued presence of (+)-202-791. B, summary bar graph (n = 7 patches) demonstrates that 5 μM AA in the continued presence of (+)-202-791 also significantly decreased the P_o of single N-type Ca²⁺ channels compared with control P_o (*P < 0.05).

To determine whether additional gating changes in N-type Ca²⁺ channel activity occurred in the presence of AA, further kinetic analyses were performed. The presence of AA significantly decreased the number of N-type Ca²⁺ channel openings in sweeps with activity, but had no effect on the mean open time (Table 2). The mean closed time also did not significantly change with AA (Table 2), reflecting that when the N-type Ca²⁺ channel opened, it appeared to do so in bursts that looked similar to control activity. These data suggest that AA does not affect closed states that can transition directly to the open state, otherwise the bursting pattern would change. However, as with unitary L-type Ca²⁺ channel activity, exclusion of truncated-last closings resulted in underestimating the time the channel spent closed or inactivated in the presence of AA. When the mean closed time was calculated with last closed events included, the mean closed time following the application of AA was now significantly longer than control (Table 2).

The increase in mean closed times, when the last closings were included, could be due either to changes in other closed states or an increase in inactivation. To probe whether AA stabilized the N-type Ca²⁺ channel in a closed state, we examined whether AA affected first latency by employing the same strategy as was used with the L-type Ca²⁺ channel. In the presence of AA, the first latency approximately doubled (Table 2). Cumulative first latency plots also indicated an increase in the first latency with AA. In seven control recordings, the cumulative first latency plots could be described by a fast time constant (τ₁) of 3.0 ± 0.6 ms and a slow time constant (τ₂) of 120.5 ± 47.6 ms. Both increased in the presence of AA (τ₁ =9.4 ± 4.5 ms; τ₂ =266.7 ± 75.3 ms), though only the change in the slower time constant was significant (P < 0.05, one-way, paired t test).

In addition to increases in closed time and first latency, we tested whether the rate of fast inactivation changes with the inhibition of unitary N-type Ca²⁺ channel activity by AA, since there was some indication of this in Fig. 5B. We fitted mean ensemble currents in the presence and absence of AA with time constants for fast inactivation. As mentioned previously, unitary N-type Ca²⁺ channel activity is modal with variability among individual channels in the incidence of inactivating versus non-inactivating modes (Plummer & Hess, 1991; Rittenhouse & Hess, 1994). In some recordings the percentage of inactivating sweeps was sufficiently low that it was difficult to fit accurately a time constant to the control or AA mean ensemble current. Nevertheless in three recordings it was possible, and we found that AA did not significantly alter the rate of inactivation (τ_i =100.9 ± 34.5 ms) compared with control (τ_i =138.0 ± 54.2 ms). Moreover, in the presence of AA, currents generated with long test pulses in the whole cell mode (Fig. 1D) showed no change in their inactivation kinetics compared with control sweeps; in 4 out of 5 cells the component of current inhibited by AA was non-inactivating. Taken together, these findings are consistent with the possibility that AA enhances the incidence of the null mode, decreases the non-inactivating mode, but appears to have no effect on the inactivating mode.

Bovine serum albumin reverses the effects of AA on multi-channel, cell-attached patch recordings

Arachidonic acid decreased L- and N-type Ca²⁺ channel activity in single channel patches to very low levels. One concern with such low activity is that the AA may have irreversibly disrupted the membrane rather than acting directly or indirectly through a signalling cascade to modify channel activity. A second concern is that puffing the AA onto the cell might have disrupted the patch mechanically, also resulting in a loss of channel activity. To address these possibilities, we tested whether the inhibition of Ca²⁺ channel activity in multi-channel, cell-attached patch recordings by AA was reversible. In patches that contained 1–3 channels (n = 6 patches), superfusion of cells with 0.5 mg ml⁻¹ BSA in the continued presence of 500 nM (+)-202-791 had no obvious effects on NP_o (0.47 ± 0.07) compared with NP_o (0.50 ± 0.05) in the presence of (+)-202-791 alone. However, superfusion of the external solution with 0.5 mg ml⁻¹ BSA in the continued presence of (+)-202-791 reversed the inhibition of Ca²⁺ channel activity by 5 μM AA in 4 out of 6 multi-channel cell-attached patches. Figure 7 shows an example of the reversibility of the inhibitory effects of AA on channel activity (Fig. 7A), mean ensemble currents (Fig. 7B) and NP_ovs. time (Fig. 7C).

Figure 7. Arachidonic acid reversibly inhibits Ca²⁺ channel activity in multi-channel patches.

An example of Ca²⁺ channel activity in the presence of 0.5 mg ml⁻¹ BSA in a cell-attached patch containing one L-type channel with high activity and one N-type Ca²⁺ channel with very low activity. A, six consecutive leak-subtracted sweeps recorded under different conditions illustrate the recovery of channel activity when AA was washed away with a solution containing BSA. (+)-202-791 (500 nM) was present throughout the recording. Dashed lines indicate zero current level. B, mean currents, in the presence of BSA (left), in the presence of 5 μM AA (centre) and after washing with BSA (right), each were generated from 95 consecutive sweeps. C, plot of N (number of channels) multiplied by P_o (open probability) versus time from the same patch as in A and B. Horizontal bar indicates the change in conditions.

Lastly, to determine whether the effects of AA were concentration dependent, we lowered the concentration of AA from 5 to 1 μM. Under these conditions, 1 μM AA puffed onto cells (n = 2) had no effect on currents recorded from multi-channel, cell-attached patches (data not shown), indicating that the decrease in channel activity was not due to some artifact caused by the puffer. The lack of effect of AA at 1 μM is consistent with our whole cell studies. We found the IC₅₀ for inhibition of the slow component of the tail current to be 5.4 μM (data not shown). Taken together, these results suggest that the inhibition of channel activity is due to concentration-dependent, specific actions of AA and not due to disruption of the patch or other artifacts.

DISCUSSION

Arachidonic acid inhibits L- and N-type Ba²⁺ currents in SCG neurons

To determine whether L- and N-type currents are sensitive to AA, the effects of AA on Ba²⁺ currents in SCG neurons were examined. We found that bath application of 5 μM AA significantly inhibited the whole cell peak current, mostly made up of N-type current, and the slow component of the tail current, made up entirely of L-type current. We confirmed these findings in the cell-attached patch configuration and examined how AA inhibits unitary Ba²⁺ currents at the single channel level. Arachidonic acid, puffed onto SCG neurons, decreased mean ensemble currents that were created from patches containing a single L- or N-type Ca²⁺ channel. For both channel types, AA had no significant effect on unitary current amplitude, but did decrease channel P_o, suggesting that channel gating, but not permeation, is affected.

Analysis of single channel kinetics provided additional information as to how AA reduced channel P_o. Unitary current kinetics of both L- and N-type Ca²⁺ channel activity appeared to be affected similarly by AA. The most prominent effect of AA was to increase the percentage of null sweeps and the number of null sweeps that clustered together, indicating that AA enhanced the null mode. When activity did occur, AA decreased the number of openings per sweep, but had no significant effect on mean open time. In addition, mean closed time, when last closings were included in the analysis, and mean first latency increased. Taken together, these changes are most consistent with a mechanism of action where AA stabilizes one or more closed states that can transition directly to an inactivated state.

It remains possible that AA does affect inactivation and the longer component of the first latency reflects slower recovery from inactivation. At the single channel level, if a particular inactivated state is not fully absorbing at +30 mV, a channel that happens to be inactivated at this test potential will transiently recover if given enough time. No changes in the kinetics of, or the incidence of the inactivating mode appear to occur in the presence of AA. However, changes in slower forms of inactivation also could account for the longer truncated closings, increases in the mean first latency and increases in the null mode. Whether AA induces changes in closed times that promote inactivation and/or causes changes in inactivation, the finding of increases in the null mode are consistent with our whole cell data. In these studies, bath application of AA had no significant effect on the voltage dependence of activation but did enhance holding potential-dependent inactivation (Liu et al. 2000). Further investigation is necessary to resolve whether changes in closed state dwell times can account for the effects of AA and/or whether slow forms of inactivation are affected by the presence of AA.

Non-specific effects of AA cannot account for decreased L- and N-type Ca²⁺ channel activity

The decreases in whole cell currents and unitary L- and N-type Ca²⁺ channel activity appear to be due to specific actions of AA. Arachidonic acid inhibited whole cell currents either in the presence or absence of the L-type Ca²⁺ channel agonist (+)-202-791, indicating that it does not inhibit L-type current simply by displacing the agonist (Shimasue et al. 1996). Furthermore, N-type Ca²⁺ channels do not bind dihydropyridines; therefore inhibition by AA must occur independently of any competition with dihydropyridines. At the single channel level, the lack of change in unitary amplitude or mean open time suggests, but does not rule out the possibility, that application of AA to the bath does not act by some non-specific plugging of the channel pore. At higher concentrations (>10 μM), AA can disrupt membrane integrity due to various non-specific detergent effects on membranes and ion channels (Meves, 1994). However, this seems unlikely in our recording conditions since inhibition of whole cell and unitary currents by 5 μM AA could be reversed with BSA. Though the in vivo concentrations of released AA are unknown, it has been suggested that the actions of low micromolar concentrations of AA may have physiological significance (Anderson & Welsh, 1990) since the oxygenases that metabolize AA have K_m values for AA in the range of 3–28 μM (Needleman et al. 1986). Whether AA confers the changes in N- or L-type Ca²⁺ channel gating directly, or by activating other signalling molecules, or indirectly via one of its metabolites, remains to be elucidated.

Arachidonic acid modulates the activity of other ion channels

The biophysical effects of AA, when examined in other nerve cell types, are consistent with our findings (Piomelli et al. 1987; Bug et al. 1989; Keyser & Alger, 1990; Khurana & Bennett, 1993; Hatton & Peers, 1998). In other mammalian neurons, inhibition of whole cell Ca²⁺ currents by AA caused no obvious changes in whole cell current kinetics (Keyser & Alger, 1990; Khurana & Bennett, 1993; Schmitt & Meves, 1995; Hatton & Peers, 1998). The significant increase in the percentage of null sweeps for both channels is consistent with whole cell data from cardiac and intestinal smooth muscle cells which show that AA shifts the holding potential-dependent inactivation curve for L-type current in a negative direction (Shimada & Somlyo, 1992; Petit-Jacques & Hartzell, 1996).

Arachidonic acid appears to cause a variety of changes in the activity of other voltage-gated ion channels including chloride, sodium and potassium channels that result in the inhibition of some currents and the enhancement of others. Where examined, gating and not permeation appears to be altered (Anderson & Welsh, 1990; Hwang et al. 1990; Devor & Frizzell, 1998; Bringmann et al. 1998), as is the case with L- and N-type Ca²⁺ channel activity. In one case, AA was found to increase the activity of a large conductance potassium channel by decreasing the dwell time constants for two intermediate closed states, but without affecting dwell time in the open state or in a closed state that leads to the open state (Devor & Frizzell 1998). These findings suggest that AA may affect a potassium current in a similar manner to Ca²⁺ currents, except that AA enhanced rather than inhibited the current. In cases where AA inhibits other currents, such as an outwardly rectifying chloride current (Anderson & Welsh, 1990), a transient A current (Bringmann et al. 1998) and a sodium current (Fraser et al. 1993), AA enhances inactivation. Whether AA acts at homologous sites among these different ion channels is not known, though it need not do so. Indeed, if AA can act directly and/or indirectly, the site of action on each channel type may vary widely, as could changes in channel behaviour (Duerson et al. 1996; Petit-Jacques & Hartzell, 1996; Skinner et al. 1997).

Physiological implications of AA inhibition of Ca²⁺ channel activity

Arachidonic acid and its metabolites are thought to participate in pathophysiological processes during stroke, convulsions, epilepsy, and neurodegenerative diseases that result in impaired brain function (see review by Bazan, 1989). In particular, ischaemia is accompanied by an accumulation of free fatty acids, with the largest relative change occurring in the free AA concentration (Bazan, 1976; Siesjö, 1981; Yasuda et al. 1985). The generation of free radicals produced during AA metabolism is thought to contribute significantly to cell damage in the brain (Bazan, 1989). It also has been reported that ischaemic injury in neurons is mediated by Ca²⁺ influx through receptor- and voltage-gated ion channels (see reviews by Choi, 1988; Siesjö & Bengtsson, 1989). Arachidonic acid can enhance NMDA receptor-mediated currents (Miller et al. 1992), while our data have shown that AA inhibits L- and N-type currents in neurons. These findings suggest that the release of AA from the membrane and the modulation of ion channel activity by AA may potentially play antagonist roles during ischaemia-reperfusion-induced cytotoxicity. Under more physiological conditions, a number of neurotransmitters stimulate the release of AA from nerve cell membranes where it appears to participate in modulating membrane excitability (Gammon et al. 1989; Felder et al. 1990; Tence et al. 1994; Rodriguez-Alvarez et al. 1997).

In conclusion, we have found that AA inhibits both L- and N-type Ba²⁺ currents in whole cell and cell-attached patch recording configurations. Analysis of unitary currents indicates that AA inhibits L- and N-type currents similarly by increasing the percentage of null sweeps and the number of null sweeps in a cluster. In sweeps with activity, mean first latency and closed time (when including last closings) increased, while the number of openings per sweep decreased. Mean open time showed no significant change. These changes in kinetics are consistent with the possibility that AA enhances the dwell time in the null mode and stabilizes one or more closed states in sweeps with activity. Changes in slow forms of inactivation may also occur. Our findings add to the growing body of evidence that AA can modulate the activity of multiple types of ion channels in the nervous system.