European Journal of Pharmacology, 216 (1992) 191-198
191
© 1992 Elsevier Science Publishers B.V. All rights reserved 0014-2999/92/$05.00
EJP 52472
Blockade by trifluoperazine of a CaZ+-activated K + channel in rat hippocampai pyramidal neurons Y o s h i m i I k e m o t o a, A t s u y a Y o s h i d a b a n d M a s a o O d a c a Department of DentalAnesthesiology, Faculty of Dentistry, Kyushu University, Fukuoka 812, Japan, h Second Department of Oral Surgery, Faculty of Dentistry, Kyushu University, Fukuoka 812, Japan and c Department of Anesthesiology, Saga Medical School, Saga 840-01, Japan Received 14 November 1991, revised MS received 11 February 1992, accepted 10 March 1992
The effects of trifluoperazine, a phenothiazine derivative, on the large-conductance Ca2+-activated K + channel (BKca) in dissociated rat hippocampal pyramidal neurons were examined using the inside-out configuration of the patch-clamp technique. The BKc~ was activated by 12.6 tzM Ca 2÷ on the internal surface of the membrane patch. The single channel conductance of the BKca was 244 + 17.5 pS (n = 10) in symmetrical solutions of 150 mM K +. Trifluoperazine, applied on the internal surface of the membrane, decreased the open probability of the channel without changing the single channel conductance. The reduction in the open probability was well described by a block of the open state of the channel in a simple sequential model. The apparent dissociation constant (K o) for the reduction was calculated to be 1.4/zM and the Hill coefficient 0.69 at + 20 mV. The inhibition was voltage dependent, being more pronounced at depolarized voltages. The voltage dependence enabled us to estimate that the binding site for the agent in the channel lies about half way across the membrane electrical field. It is concluded that trifluoperazine blocks the open state of the BKca, which is known to provide an outward current for repolarization and afterhyperpolarization of the neuronal action potential. This may result in a decrease in spike intervals during burst firing of neurons. Ca2+-activated K + currents; Pyramidal neurons (dissociated); Hippocampus (rat); Single channel recording; Trifluoperazine
1. Introduction
Phenothiazines are used in the treatment of patients with psychosis, particularly schizophrenia, and have been reported to exert their therapeutic effects by antagonizing dopamine binding to its receptors (Seeman, 1980). Using intracellular recordings of neurons in slice preparations of nucleus accumbens septi of the guinea pig, Uchimura et al. (1986) showed that activation of D~ receptors induced a hyperpolarization as a result of an increase in K ÷ conductance and that activation of D 2 receptors resulted in depolarization accompanied by a decrease in K + conductance. It is also known that the binding of dopamine to D~ receptors activates adenylate cyclase and that D 2 receptors are coupled to G-protein and inhibit adenylate cyclase activity (Civelli et al., 1989). Phenothiazines are reported to suppress the binding of dopamine to D e receptors more potently than the binding to D~-recep-
Correspondence to: Y. Ikemoto, Department of Dental Anesthesiology, Faculty of Dentistry, Kyushu University, Fukuoka 812, Japan. Tel. 81.92.641 1151 ext. 5291, fax 81.92.641 8027.
tors (Billard et al., 1984). Besides antipsychotic effects, phenothiazines have sedative, anxiolytic, antiemetic and antihistaminic effects. In addition, they also potentiate the effects of analgesics and general anesthetics (Baldessarini, 1990) and may produce seizure patterns in the electroencephalogram and induce convulsions in patients who have a history of epilepsy or a condition that predisposes them to seizures (Itil, 1978). In addition to dopamine receptor blockade, investigators have shown that these drugs also suppress other ion currents, such as the Na + current (Ogata and Tatebayashi, 1989), a Ca 2+ current (Ogata et al., 1990), and K ÷ currents including the one that underlies the slow afterhyperpolarization (Ogata et al., 1989; Dinan et ai., 1987). This suppression may contribute to the actions of these agents in the central nervous system. At least three classes of CaZ+-activated K ÷ channels have been described in a variety of excitable cells including central neurons, namely, (1) large-conductance Kca channels (BKca) , (2) intermediate-conductance Kca channels (IKca) and (3) small-conductance Kca channels (SKca) (Halliwell, 1990). The influx of Ca 2+ through Ca z+ channels during the action potential and the release of Ca 2+ from the endoplasmic
192 reticulum increase the intracellular Ca 2+ concentration ([Ca2+]i) , which in turn may activate the three classes of K + channels. The BKca provides an outward current for repolarization and afterhypolarization of the action potential (Storm, 1987; Lancaster and Nicoll, 1987; Yoshida et al., 1991). In the central nervous system, this outward current may play an important role in controlling the interval between action potentials in the early phase of spike bursts, and is thus probably involved in the suppression of epileptogenesis. We now demonstrate that trifluoperazine, a phenothiazine derivative, blocks the open state of the BKca in dissociated pyramidal neurons of the rat hippocampus.
2. Materials and methods
2.1. Dissociation of the neuron Single hippocampal pyramidal neurons were dissociated from the brain of Wistar rats (7-10 days old) of both sexes, using a technique similar to that developed by Kaneda et al. (1988). Briefly, under diethyl ether anesthesia, the rats were decapitated and the brain was rapidly dissected out and cut into slices of 400-600 ~m. These slices were then treated for 20 min at 37°C with collagenase (0.04%, Type I, Sigma, USA) and actinase (0.056%, Kaken Chemical Co., Japan) dissolved in the control external solution. The enzyme was washed out with external solution containing 20% fetal calf serum. The treated brain slices of hippocampal tissue were stored in the external solution. The CA-1 region of the slice was punched out and the pyramidal neurons were mechanically dissociated in culture dishes just before use.
descn0ed by Ikemoto et al. (1989). Trifluoperazine was applied in the 12.6 ~ M CaZ+-internal solution. All data were digitized with a PCM converter (PCM 501-ES, Sony, Japan: altered for DC to 20 kHz) and stored on video tapes for later analysis, which was performed with a microcomputer (PC-9801XL, NEC, Japan). The data were replayed and filtered at a cut-off frequency of 2 kHz with a Bessel-type low-pass filter (48 d B / o c tave, FV-625A, NF Electronic Instruments, Japan) before sampling at an appropriate frequency. 2.3. Solutions The composition of the control external solution was (in mM): NaC1 150.0; KCI 5.0; MgC12 1.0; CaCI 2 1.0; glucose 10.0; H E P E S 10.0. The solution was titrated to pH 7.4 with Tris base. The composition of the internal solution used for the inside-out patch recordings was (in mM): KC1 150.0; H E P E S 5.0; E G T A 1.0 (pH 7.2 with KOH). The concentration of free Ca 2+ was adjusted to 12.6 p.M by adding an appropriate amount of CaC12 according to Fabiato and Fabiato (1979) and Tsien and Rink (1980). The Ca2+-free internal solution contained 1 mM E G T A but no Ca 2+. The pipette solution contained (in raM): KC1 150.0; CaC12 1.0; H E P E S 5.0 (pH 7.4 with KOH). 2.4. Drugs The drugs used were purchased from the following companies: collagenase (type I, Sigma, USA), actinase (Kaken Chemical Co., Japan), fetal calf serum (Dainippon Chemical Co., Japan) and trifluoperazine (Sigma, USA). All experiments were carried out at room temperature (20-25°C). Statistical values are given as means _+ S.D.
2.2. Single channel recordings 3. Results
After the cells had settled onto the bottom of the dish, the experiments were started. The glass electrode was usually filled with a 150 mM K + pipette solution and had a tip resistance of 5-10 MO. With a micromanipulator (MO-203, Narishige, Japan), the electrode was slowly moved clown onto the surface of a neuron in the external solution. After a gigaohm seal had been made with a small negative pressure applied in the electrode, the electrode was drawn back to excise an inside-out membrane patch (Hamill et al., 1981). Under this inside-out patch condition, Ca 2+ (1 raM) in the external solution activated single K + currents in more than 90% of the membrane patches. These currents were recorded with a patch-clamp amplifier (EPC-7, List, FRG). The internal solution (Ca2+-free or 12.6 p~M Ca 2+) was applied on the internal surface of the membrane patch with the rapid application technique
3.1. Reduction in the open probability in response to trifluoperazine We have previously described the kinetic properties of the CaZ+-activated K + channel now studied (Yoshida et al., 1991). The channel is activated by Ca 2+, Sr 2+ and Ba 2+ with increasing sensitivity at depolarized potentials. The kinetics of opening and closing of the channel are well explained by a simple sequential model. The open time histogram can be fitted with a single exponential function and the closed time histogram shows at least two exponential components at varius Ca 2+ concentrations. The channel has a large single-channel conductance of 200-250 pS and is blocked by Ba 2+ or tetraethylammonium. These properties indicate that the channel can be classified as the
193
BKca channel type reported in other tissues (Latorre et al., 1989). Representative records of the K current activated by 12.6/zM [Ca2+]~ in symmetrical 150 mM K ÷ solutions are shown in fig. 1A. The channel activity increased with increasing depolarization, as mentioned above, and at +40 mV the channel showed almost sustained bursting activity. The channel closing events of relatively long duration observed at negative potentials became less frequent with depolarization, leaving the repetitive brief closing events in the burst. After recording the control current, we applied 10/zM trifluoperazine to the internal surface of the membrane.
A
a
b
Control
.= ~2 v
C
TFP 10-8M
p=0.85
p=0.48
TFP 2XtOeM
p=0.26
[3.
A
L~
0
~ 10
0
0
0
0
10
10
Amplitude (pA)
A
Ca2* 1.26X1(~eM
b
a Control
B
TFP 1()5M
1.0
mv 40 2omv
~
2
0.5
-20 , ~ - ~ ~ ~
_,o
-
1.26XI0-5M
Ktw- ~4Xtd6M ~k
,
# __J
lOpA
lOOms 0
B
BKca lO 0 Control z~ TFP 1135M
i
-40
! ~
i
-20/[0
/
of° i
20
J
40my
/ O~
~
J
~
J
10-6
10-5
10-4
TFP (M)
pA
5
i
10-7
-5
-10 Fig. 1. (A) Single BKca currents and effects of trifluoperazine (TFP), recorded in the inside-out patch configuration. The channel was activated by 12.6 /zM [Ca2+ ]i. The pipette solution contained 150 mM K +. The standard internal solution (150 mM K + ) containing 10 IzM trifluoperazine was perfused after the control current had been recorded. The m e m b r a n e potentials are given on the left. T h e upward deflections represent the outward current, and the dashed lines indicate closed channel current level. (B) The voltage-current relationship of the BK¢~ current in the control condition (©) and in the presence of 10 IzM trifluoperazine (,x). The amplitude of the current was measured as described in fig. 2. The single channel conductance in this patch was 220 pS and was not affected by the drug. The straight line was drawn by eye. All data were taken from the same patch. The effects were reversible by washing out for several minutes.
Fig. 2. (A) The amplitude histograms of the BKc, currents recorded at + 20 mV in the absence (a) and presence of internal trifluoperazine (TFP: b 1 /zM and c 2/zM). The open probability (p) was 0.85 in control conditions and decreased to 0.48 and 0.26 in the presence of 1 and 2 /.tM of the drug, respectively. Ordinate: percentage of the total number of measured points. Abscissa: amplitude of the current in pA. (B) The dose-response curve for the inhibition of the open probability by trifluoperazine at + 20 inV. The relative value of the open probability in the presence of the drug (PTFP) to that in control (Pco.t) is plotted against the drug concentration. Each point represents the mean of five to seven measurements. Vertical bars indicate one S.D. A least-squares fit yielded an apparent dissociation constant (K D) of 1.4/zM and a Hill coefficient of 0.69.
The open probability of the channel was markedly reduced but the amplitude of the current was not changed by the drug (fig. lab). The reduction was greater at depolarized potentials. The amplitude of the single channel current was obtained as described below (see fig. 2A) and plotted against the membrane potential in fig. 1B (open circles). The K channel had an ohmic conductance of 220 pS with a reversal potential of 0 mV. The average conductance of 10 patches from 10 neurons was 244 +
194
17.5 pS. The current amplitude measured in the presence of trifluoperazine is also plotted in fig. 1B (open triangles), which shows that the single channel conductance was not altered by the drug. Amplitude histograms were constructed from records obtained at + 2 0 mV (fig. 2A). The open channel current showed a symmetrical distribution, and the very brief closing events caused a small continuous distribution between the full open level and the closed level under each experimental condition. The amplitude of the single channel current was measured
A
a
b
e
d
c
Control
TFP 10-6M
5XlO6M
2Xl()S M
,o[ •to o •
tau = 30.2ms
9.6ms
c
[I 1 2ms 30
'tL_"
1(?
10
10
I°Ik~ol
i,
0
100
0
25
50
0
11~Oo
where I stands for mean patch current, i for the unit current amplitude, and N for channel number within the patch. In the present study N was 1 in every experiment analyzed; p was 0.85 in the control condition at + 20 mV and was reduced to 0.48 and 0.26 by 1 and 2/~M trifluoperazine, respectively (fig. 2A). Figure 2B illustrates the dose-inhibition relationship of the drug at + 20 mV. The continuous line is a least-squares fit according to the following equation: P = K~/(D n + K~)
20
2O
(1)
I(~5M
n
3O
30
p = I/(i'N)
(2)
40~
L2rns
5.2ms
)
0
I
from the peak position, and was not changed by 1 and 2 /xM trifluoperazine (fig. 2Ab,c). The amplitude histogram also indicates the open probability (p) of the channel, which was calculated as the time during which the channel was in a open state:
t5
where P denotes the fig. 2), D the drug dissociation constant fitting gave a K D of 0.69.
relative value of p (PTFP/Pco,t in concentration, K D the apparent and n the Hill coefficient. The 1.4 txM and a Hill coefficient of
3.2. Open channel blockade by trifluoperazine 0
5
open Time (ms)
B ° 80O
Inorder to elucidate how trifluoperazine reduces the open probability, we proposed that trifluoperazine blocked open channels (Neher, 1983), since this mode of action has been proposed for neuroleptics in airway smooth muscle cells (McCann and Welsh, 1987). The model is described below with one closed state for the sake of simplicity: kt
ft'[D]
C ~ O ~
O'B
(3)
k_ m
/
400
/
where C denotes the closed state, O the open state and O • B the blocked state of the channel, kl, k_~, a and /3 are rate constants and [D] the concentration of the drug. The following three equations are derived from equation (3):
o
o
TFP (XlO-e M)
Fig. 3. (A) The open time histograms obtained in a patch u n d e r control conditions (a) and in the presence of trifluoperazine (TFP: 1, 2, 5 and 10 p~M in b, c, d and e, respectively). All histograms were fitted with single exponential functions. The time constant (tau) of the histogram, which reflects the m e a n open time, was 30.2 ms in control. Trifluoperazine decreased the m e a n open time in a dose-dependent manner. Superimposed curves are the best fits of single exponential functions. Ordinate: percentage of the total n u m b e r of the events. Abscissa: time in ms. (B) The rate constant ( 1 / t a u ) is plotted against the concentration of trifluoperazine. The slope gave /3 of 82 s -1 /zM - I in equation (4) in the text.
1 / ( m e a n open time) = k i + / 3 . [ D ]
(4)
1 / ( m e a n block time) = a
(5)
KD=a//3
(6)
Figure 3A illustrates the open time histograms, which were well fitted with a single exponential curve in the absence and presence of trifluoperazine. The drug decreased the time constant (tau), which describes the mean open time, in a dose-dependent manner. 1 / ( m e a n open time) is plotted against trifluoperazine concentration in fig. 3B. The data points fell on a straight line and the slope gave /3 a value of 82 s -1 /zM -~ in equation (4). The average of /3 with nine patches was 93.2 + 7.6 s -] /xM -~
195
A a
tau~--0.54rns
B
btau2--4m3mS
a
C tau =8.6ms
I
+,g4° t
el
P 30
20
10
oL
10
5
~
16~M
rFP
5.6ms
2C.
tN
C
30-
20
O-
b 2x IO"SM
20
0
50
25
Closed Time (ms)
25
k
0
Closed Time (ms)
25
Fig. 4. (A) The closed time histogram under control conditions. The histogram was fitted with three exponential functions with a time constant of 0.54, 4.3 and 8.6 ms, respectively. (B) In the presence of higher concentrations of trifluoperazine (TFP), the closed time histrogram was fitted with a single exponential function. The time constant was 6.3, 5.4 and 5.6 ms at (a) 2, (b) 5 and (c) 10/~M, respectively, giving the mean block time ( I / a ) in equation (5). Ordinate: percentage of the total number of the events. Abscissa: time in ms.
The closed time histogram was fitted with the sum of three exponential functions in the control condition (fig. 4A). In the presence of high concentrations of trifluoperazine, the histogram was fitted with a single exponential function, the time constant of which reflects the transition rate constant from O . B to O,
A
since most channels were blocked by the agent. As predicted by the model, the time constant of blockade (mean block time) did not vary over the concentration range of 2 - 1 0 / x M trifluoperazine (fig. 4B). The mean block time with 10 /zM of the drug was 5.6 + 0.3 ms (eight patches from eight neurons), and gave a a value
B-20mY
20mV
acontrol
bTFP 1()5 M
C -40mY %
a control
4
t~SM
acontrol
4
b F p Io-sM
4
t-
o(D CL p=0.89,
p=0.67
p=0.56
2
2
1
0.20
0.20
o,
0 0
10
0
10
, 0
, 10 Amplitude
10
0
10
10
(pA)
Fig. 5. The voltage-dependent reduction in the open probability by trifluoperazine (TFP). The drug decreased the open probability from 0.89 to 0.19 at +20 mV (A), from 0.67 to 0.20 at - 2 0 mV (B) and from 0.56 to 0.20 at - 4 0 mV (C). Ordinate: percentage of the total number of measured points. Abscissa: amplitude of the current in pA.
196 of 178/s. a and /3, being based on the sequential model in equation (6), yielded a K D value of 1.9 ~ M , which is close to the K D obtained from the reduction in the mean patch current (fig. 2B; 1.4 /~M). This finding indicates that trifluoperazine acts on the BKca by blocking open channels.
B
A 0
2z
KD(pM)
3.3. Voltage-dependent effects of trifluoperazine
/ o / ~ /
The voltage dependence of the effect of trifluoperazine, which was already seen in fig. 1A, was studied in more detail. Figure 5 depicts the voltage-dependent effects of trifluoperazine on the open probability. The channel was activated by 12.6 /~M Ca 2+ in a voltagedependent manner, namely, p = 0.89 at + 20 mV, 0.67 at - 2 0 mV and 0.56 at - 4 0 inV. Trifluoperazine (10 ~ M ) decreased the probability to 21, 30 and 36% of the control value at + 20, - 2 0 and - 4 0 mV, respectively. According to the sequential model, the K D can be determined as following (Benham et al., 1985; McCann and Welsh, 1987): KD = [D]'P'Pd/(P
-- P a )
(7)
where [D] denotes the drug concentration, and p and Pd the open probabilities in the absence and presence of the drug, respectively. To get a linear relation, equation (7) is modified as: (P-Pd)/P'Pd = [D]/KD
(8)
( P - P d ) / P ' P d iS plotted against the drug concentration in fig. 6A. The linear relationships indicate that equation (8) is valid for the data of this study and that the block occurs on a basis of a one-to-one binding reaction, which is consistent with the Hill coefficient of 0.69 in fig. 2B. The values of K D obtained from the slopes of the linear relationships are plotted against the voltage in fig. 6B, where the interpolation to 0 m V yielded the apparent dissociation constant at 0 mV (KD(0)). The binding site in the m e m b r a n e electrical field of a drug is determined by the following equation (Woodhull, 1973), KD
=
KD(0 ) .exp( - 6zFV~/RT)
(9)
where 6 is a partition parameter, Vm the m e m b r a n e potential, and z, F, R and T have their usual meanings in thermodynamics. To get a linear relationship, equation (9) is modified as: RT/zF.In(KD(0)/KD) = 6 'Vm
(10)
Figure 6C shows a plot of the left term of equation (10) against the voltage. Trifluoperazine dihydrochloride has two dissociation constants: pK 1 = 3.9 and p K 2 = 8.1. At physiological pH, namely 7.2 in the present experiments, the drug dissociated according to p K 2, with about 95% being monovalent. The valence of unity
,
lo TFP
20
-40
,
20
?" 2_0 mY
(pM)
C
mV
20[
0
-10
-20
o/
t L
Fig. 6. Estimation of the binding site for trifluoperazine (TFP) in the channel• (A) Determination of the K D for the blocking effect of the agent at +20 (©), -20 (o) and -40 mV (z~). (P--Pd)/P'Pd is plotted against the drug concentration. The reciprocal of the slope gave the K o at each voltage. (B) Interpolation of the KD-voltage plot yielded the KD(0). (C) RT/zF.In(KD(0)/K D) is plotted against the membrane potential. The slope gave a partition parameter, 6, of 0,47, which suggests that the binding site for trifluoperazine lies about half way across the electrical field from the internal pore of the channel.
(z = 1) was therefore used in equation (9). The value of 6 was 0.47 _+ 0.03 (n = 4), suggesting that the blocking site of trifluoperazine lies about half way from the internal pore across the m e m b r a n e electrical field.
4. Discussion Single channel recordings have been used in the study of drug effects on various types of ionic channels. We have reported that the charged form of local anesthetics inside a neuron produce an open state blockade of the BKca (Oda et al., 1992). Neuroleptics, including haloperidol and trifluoperazine, have been shown to block the open state of the BKca in dissociated smooth muscle cells, and the blocking effect was shown not to be related to the inhibition of calmodulin (McCann and Welsh, 1987; Kihara et al., 1990). Benishin et al. (1988) studied the effects of neuroleptics on CaZ+-de pendent 86Rb+ efflux in rat brain synaptosomes stimulated with 100 m M K +. At concentrations which did not depress 45Ca2+ uptake, trifluoperazine inhibited the 86Rb+ effiux with a biphasic d o s e - r e s p o n s e curve,
197
thereby suggesting the existence of two populations of Ca2+-activated K + channels with different sensitivities. The high sensitivity component (30-40% of the flux) had an ICs0 of 20-30 nM and the low sensitivity component an ICs0 of 1-3 tzM. The K D value of 1.4 /zM determined in our experiments is consistent with the value reported by Benishin et al. (1988), suggesting that the low sensitivity component corresponds to the BKc~ in hippocampal pyramidal neurons. The open probability was calculated as the time during which the channel was in an open state. Trifluoperazine reduced the open probability without changing the single channel conductance (fig. 2). The Kt~ for the inhibition was 1.4/zM and the Hill coefficient was 0.69 at +20 mV, values which are similar to those obtained by McCann and Welsh (1987) in cultured airway smooth muscle cells. We assumed in the present study that the drug simply blocked the open state of the channel as described in equation (3). On the basis of this assumption, the K D was calculated from the opening and closing rate constants in the presence and absence of the drug. We obtained a K o value of 1.9 p,M, which was close to that obtained from the mean patch current (1.4 p~M), a result supporting the assumption that trifluoperazine acts on the BKc, by blocking open channels. This finding differs from the results obtained with pentobarbitone: simple openchannel block failed to explain the open-close kinetics of acetylcholine-activated channels (Gage and McKinnon, 1985). The blocking action was greater at depolarized potentials, reflecting that the charged form of the drug molecule affected the channel. An analysis of the voltage dependence provided an estimation that the drug binds about half way across the membrane electrical field from the internal pore to block the channel (fig. 6). The voltage-dependent effect was also noticed in airway smooth muscle cells by McCann and Welsh (1987), who reported that the blocking site for haloperidol was about one third of the way across the electrical field. To our knowledge there have been no reports concerning the plasma concentration of trifluoperazine in clinical practice. The plasma concentration of chlorpromazine, another phenothiazine derivative, is reported to be about 1/zM under steady-state conditions (Cooper et al., 1976). The concentration of trifluoperazine may be lower since it is administered at smaller doses. Shortly after administration, however, the trifluoperazine concentration may well be high enough to block the BKc~. The drug may not affect the channel from the outside of the cell membrane, as was reported by McCann and Welsh (1987). Trifluoperazine, which has a pK z of 8.1, will dissociate according to the pH of the extracellular space (pH 7.4). About 15% of the molecules may stay uncharged and will rapidly pass through the membrane into the neuron until an equi-
librium is reached. The drug dissociates again according to the intracellular pH (about 7.0), where about 96% of the molecules may become charged. The charged form enters the internal pore of the channel and binds about half way from the internal surface with a one-to-one stoichiometry. Thus, the BKc~ becomes blocked, which results in a reduction in the outward current for repolarization and spike afterhyperpolarization, and then in a decrease in the intervals between the action potentials of neurons. This even may help the dopaminergic block produce some antipsychotic effects a n d / o r may induce seizure activities by facilitating burst firing of the neurons, since the hippocampus is sometimes a focus of seizures (Alger, 1984). In summary, we have shown that trifluoperazine blocks the open state of the BKc~ of dissociated rat hippocampal pyramidal neurons. Together with the suppression of dopaminergic transmission and the reduction in voltage-dependent currents, this blockade may contribute to the various actions of phenothiazines in the mammalian central nervous system.
References Alger, B.E., 1984, Hippocampus-electrophysiological study of epileptiform activity in vitro, in: Brain Slices, ed. R. Dingledine (Plenum Press, New York) p. 155. Baldessarini, R.J., 1990, Drugs and the treatment of psychiatric disorders, in: The Pharmacological Basis of Therapeutics, 8th edn., eds. A.G. Gilman, T.W. Rail, A.S. Nies and P. Tayler (Pergamon Press, New York) p. 383. Benham, C.D., T.B. Bolton, R.J. Lang and T. Takewaki, 1985, The mechanism of action of Ba 2+ and TEA on single Ca2+-activated K+-channels in arterial and intestinal smooth muscle cell membranes, Plfiigers Arch. 403, 120. Benishin, C.G., B.K. Krueger and M.P. Blaustein, 1988, Phenothiazines and haloperidol block Ca-activated K channels in rat fore brain synaptosomes, Mol. Pharmacol. 33, 195. Billard, W., V. Ruperto, G. Crosby, L.C. lorio and A. Barnett, 1984, Characterization of the binding of 3H-SCH 23390, a selective D-I receptor antagonist ligand, in rat striatum, Life Sci. 35, 1885. Civelli, O., J. Bunzow, H. Van Tol, D. Grandy, P. Albert, J. Salon, C. Machida and K. Neve, 1989, Cloning of a rat Dz-receptor cDNA, in: Molecular Biology of Neuroreceptors and Ion Channels, NATO ASI Series, Vol. 32, ed. A. Maelicke (Springer Verlag, Berlin) p, 259. Cooper, T.B., G.M. Simpson and J.H. Lee, 1976, Thymoleptic and neuroleptic drugs plasma levels in psychiatry: current status, Int. Rev. Neurobiol. 19, 269. Dinah, T.G., V. Crunelli and J.S. Lelly, 1987, Neuroleptics decrease calcium-activated potassium conductance in hippocampal pyramidal cells, Brain Res. 407, 159. Fabiato, A. and F. Fabiato, 1979, Calculator programs for computing the composition of the solutions containing multiple metals and ligands used for experiments in skinned muscle cells, J. Physiol. (Paris) 75, 463. Gage, P.W. and D. McKinnon, 1985, Effects of pentobarbitone on acetylcholine-activated channels in mammalian muscle, Br. J. Pharmacol. 85, 229. Halliwell, J.V., 1990, K + channels in the central nervous system, in: Potassium Channels: Structure, Classification, Function and
198 Therapeutic Potentials, ed. N.S. Cook (Ellis Horwood Limited, Chichester) p. 348. Hamill, O.P., A. Marty, A.E. Neher, B. Sakmann and F.J. Sigworth, 1981, Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches, PfliJgers Arch. 391, 85. Ikemoto, Y., K. Ono, A. Yoshida and N. Akaike, 1989, Delayed activation of large-conductance Ca2+-activated K channels in hippocampal neurons of the rat, Biophys. J. 56, 207. ltil, T.M., 1978, Effects of psychotropic drugs in qualitatively and quantitatively analyzed human EEG, in: Principles of Psychopharmacology, 2nd edn., eds. W.G. Clark and J. Del Giudice (Academic Press, New York) p. 261. Kaneda, M., H. Nakamura and N. Akaike, 1988, Mechanical and enzymatic isolation of mammalian CNS neurons, Neurosci. Res. 5,299. Kihara, M., K. Matsuzawa, H. Tokuno and T. Tomita, 1990, Effects of calmodulin antagonists on calcium-activated potassium channels in pregnant rat myometrium, Br. J. Pharmacol. 100, 353. Lancaster, B. and R.A. Nicoll, 1987, Properties of two calciumactivated hyperpolarizations in rat hippocampal neurons, J. Physiol. (London) 389, 187. Latorre, R., A. Oberhauser, P. Labarca and O. Alvarez, 1989, Varieties of calcium-activated potassium channels, Ann. Rev. Physiol. 51,385. McCann, J.D. and M.J. Welsh, 1987, Neuroleptics antagonize a calcium-activated potassium channel in airway smooth muscle, J. Gen. Physiol. 89, 339. Neher, E., 1983, The charge carried by single-channel currents of rat cultured muscle cells in the presence of local anaesthetics, J. Physiol. (London) 339, 663.
Oda, M., A. Yoshida and Y. Ikemoto, 1992, Blockade by local anesthetics of the single Ca2+-activated K + channel in rat hippocampal neurones, Br. J. Pharmacol. 105, 63. Ogata, N. and H. Tatebayashi, 1989, Modulation of sodium current kinetics by chlorpromazine in freshly-isolated striatal neurons of the adult guinea-pig, Br. J. Pharmacol. 98, 1173. Ogata, N., M. Yoshii and T. Narahashi, 1989, Psychotropic drugs block voltage-gated ion channels in neuroblastoma cells, Brain Res. 476, 140. Ogata, N., M. Yoshii and T. Narahashi, 1990, Differential block of sodium and calcium channels by chlorpromazine in mouse neuroblastoma cells, J. Physiol. (London) 420, 165. Seeman, P., 1980, Brain dopamine receptors, Pharmacol. Rev. 32. 229. Storm, J.F., 1987, Action potential repolarization and a fast after-hyperpolarization in rat hippocampal pyramidal cells, J. Physiol. (London) 385, 733. Tsien, R.Y. and T.J. Rink, 1980, Neutral carrier ion-selective microelectrodes for measurement of intracellular free calcium, Biochim. Biophys. Acta 599, 623. Uchimura, N., H. Higashi and S. Nishi, 1986, Hyperpolarizing and depolarizing actions of dopamine via D-1 and D-2 receptors on nucleus accumbens neurons, Brain Res. 375, 368. Woodhull, A.M., 1973, Ionic blockade of sodium channels in nerve, J. Gen. Physiol. 61,687. Yoshida, A., M. Oda and Y. Ikemoto, 199i, Kinetics of the Ca 2÷activated K + channel in rat hippocampal neuron, Jap. J. Physiol. 41,297.
191
© 1992 Elsevier Science Publishers B.V. All rights reserved 0014-2999/92/$05.00
EJP 52472
Blockade by trifluoperazine of a CaZ+-activated K + channel in rat hippocampai pyramidal neurons Y o s h i m i I k e m o t o a, A t s u y a Y o s h i d a b a n d M a s a o O d a c a Department of DentalAnesthesiology, Faculty of Dentistry, Kyushu University, Fukuoka 812, Japan, h Second Department of Oral Surgery, Faculty of Dentistry, Kyushu University, Fukuoka 812, Japan and c Department of Anesthesiology, Saga Medical School, Saga 840-01, Japan Received 14 November 1991, revised MS received 11 February 1992, accepted 10 March 1992
The effects of trifluoperazine, a phenothiazine derivative, on the large-conductance Ca2+-activated K + channel (BKca) in dissociated rat hippocampal pyramidal neurons were examined using the inside-out configuration of the patch-clamp technique. The BKc~ was activated by 12.6 tzM Ca 2÷ on the internal surface of the membrane patch. The single channel conductance of the BKca was 244 + 17.5 pS (n = 10) in symmetrical solutions of 150 mM K +. Trifluoperazine, applied on the internal surface of the membrane, decreased the open probability of the channel without changing the single channel conductance. The reduction in the open probability was well described by a block of the open state of the channel in a simple sequential model. The apparent dissociation constant (K o) for the reduction was calculated to be 1.4/zM and the Hill coefficient 0.69 at + 20 mV. The inhibition was voltage dependent, being more pronounced at depolarized voltages. The voltage dependence enabled us to estimate that the binding site for the agent in the channel lies about half way across the membrane electrical field. It is concluded that trifluoperazine blocks the open state of the BKca, which is known to provide an outward current for repolarization and afterhyperpolarization of the neuronal action potential. This may result in a decrease in spike intervals during burst firing of neurons. Ca2+-activated K + currents; Pyramidal neurons (dissociated); Hippocampus (rat); Single channel recording; Trifluoperazine
1. Introduction
Phenothiazines are used in the treatment of patients with psychosis, particularly schizophrenia, and have been reported to exert their therapeutic effects by antagonizing dopamine binding to its receptors (Seeman, 1980). Using intracellular recordings of neurons in slice preparations of nucleus accumbens septi of the guinea pig, Uchimura et al. (1986) showed that activation of D~ receptors induced a hyperpolarization as a result of an increase in K ÷ conductance and that activation of D 2 receptors resulted in depolarization accompanied by a decrease in K + conductance. It is also known that the binding of dopamine to D~ receptors activates adenylate cyclase and that D 2 receptors are coupled to G-protein and inhibit adenylate cyclase activity (Civelli et al., 1989). Phenothiazines are reported to suppress the binding of dopamine to D e receptors more potently than the binding to D~-recep-
Correspondence to: Y. Ikemoto, Department of Dental Anesthesiology, Faculty of Dentistry, Kyushu University, Fukuoka 812, Japan. Tel. 81.92.641 1151 ext. 5291, fax 81.92.641 8027.
tors (Billard et al., 1984). Besides antipsychotic effects, phenothiazines have sedative, anxiolytic, antiemetic and antihistaminic effects. In addition, they also potentiate the effects of analgesics and general anesthetics (Baldessarini, 1990) and may produce seizure patterns in the electroencephalogram and induce convulsions in patients who have a history of epilepsy or a condition that predisposes them to seizures (Itil, 1978). In addition to dopamine receptor blockade, investigators have shown that these drugs also suppress other ion currents, such as the Na + current (Ogata and Tatebayashi, 1989), a Ca 2+ current (Ogata et al., 1990), and K ÷ currents including the one that underlies the slow afterhyperpolarization (Ogata et al., 1989; Dinan et ai., 1987). This suppression may contribute to the actions of these agents in the central nervous system. At least three classes of CaZ+-activated K ÷ channels have been described in a variety of excitable cells including central neurons, namely, (1) large-conductance Kca channels (BKca) , (2) intermediate-conductance Kca channels (IKca) and (3) small-conductance Kca channels (SKca) (Halliwell, 1990). The influx of Ca 2+ through Ca z+ channels during the action potential and the release of Ca 2+ from the endoplasmic
192 reticulum increase the intracellular Ca 2+ concentration ([Ca2+]i) , which in turn may activate the three classes of K + channels. The BKca provides an outward current for repolarization and afterhypolarization of the action potential (Storm, 1987; Lancaster and Nicoll, 1987; Yoshida et al., 1991). In the central nervous system, this outward current may play an important role in controlling the interval between action potentials in the early phase of spike bursts, and is thus probably involved in the suppression of epileptogenesis. We now demonstrate that trifluoperazine, a phenothiazine derivative, blocks the open state of the BKca in dissociated pyramidal neurons of the rat hippocampus.
2. Materials and methods
2.1. Dissociation of the neuron Single hippocampal pyramidal neurons were dissociated from the brain of Wistar rats (7-10 days old) of both sexes, using a technique similar to that developed by Kaneda et al. (1988). Briefly, under diethyl ether anesthesia, the rats were decapitated and the brain was rapidly dissected out and cut into slices of 400-600 ~m. These slices were then treated for 20 min at 37°C with collagenase (0.04%, Type I, Sigma, USA) and actinase (0.056%, Kaken Chemical Co., Japan) dissolved in the control external solution. The enzyme was washed out with external solution containing 20% fetal calf serum. The treated brain slices of hippocampal tissue were stored in the external solution. The CA-1 region of the slice was punched out and the pyramidal neurons were mechanically dissociated in culture dishes just before use.
descn0ed by Ikemoto et al. (1989). Trifluoperazine was applied in the 12.6 ~ M CaZ+-internal solution. All data were digitized with a PCM converter (PCM 501-ES, Sony, Japan: altered for DC to 20 kHz) and stored on video tapes for later analysis, which was performed with a microcomputer (PC-9801XL, NEC, Japan). The data were replayed and filtered at a cut-off frequency of 2 kHz with a Bessel-type low-pass filter (48 d B / o c tave, FV-625A, NF Electronic Instruments, Japan) before sampling at an appropriate frequency. 2.3. Solutions The composition of the control external solution was (in mM): NaC1 150.0; KCI 5.0; MgC12 1.0; CaCI 2 1.0; glucose 10.0; H E P E S 10.0. The solution was titrated to pH 7.4 with Tris base. The composition of the internal solution used for the inside-out patch recordings was (in mM): KC1 150.0; H E P E S 5.0; E G T A 1.0 (pH 7.2 with KOH). The concentration of free Ca 2+ was adjusted to 12.6 p.M by adding an appropriate amount of CaC12 according to Fabiato and Fabiato (1979) and Tsien and Rink (1980). The Ca2+-free internal solution contained 1 mM E G T A but no Ca 2+. The pipette solution contained (in raM): KC1 150.0; CaC12 1.0; H E P E S 5.0 (pH 7.4 with KOH). 2.4. Drugs The drugs used were purchased from the following companies: collagenase (type I, Sigma, USA), actinase (Kaken Chemical Co., Japan), fetal calf serum (Dainippon Chemical Co., Japan) and trifluoperazine (Sigma, USA). All experiments were carried out at room temperature (20-25°C). Statistical values are given as means _+ S.D.
2.2. Single channel recordings 3. Results
After the cells had settled onto the bottom of the dish, the experiments were started. The glass electrode was usually filled with a 150 mM K + pipette solution and had a tip resistance of 5-10 MO. With a micromanipulator (MO-203, Narishige, Japan), the electrode was slowly moved clown onto the surface of a neuron in the external solution. After a gigaohm seal had been made with a small negative pressure applied in the electrode, the electrode was drawn back to excise an inside-out membrane patch (Hamill et al., 1981). Under this inside-out patch condition, Ca 2+ (1 raM) in the external solution activated single K + currents in more than 90% of the membrane patches. These currents were recorded with a patch-clamp amplifier (EPC-7, List, FRG). The internal solution (Ca2+-free or 12.6 p~M Ca 2+) was applied on the internal surface of the membrane patch with the rapid application technique
3.1. Reduction in the open probability in response to trifluoperazine We have previously described the kinetic properties of the CaZ+-activated K + channel now studied (Yoshida et al., 1991). The channel is activated by Ca 2+, Sr 2+ and Ba 2+ with increasing sensitivity at depolarized potentials. The kinetics of opening and closing of the channel are well explained by a simple sequential model. The open time histogram can be fitted with a single exponential function and the closed time histogram shows at least two exponential components at varius Ca 2+ concentrations. The channel has a large single-channel conductance of 200-250 pS and is blocked by Ba 2+ or tetraethylammonium. These properties indicate that the channel can be classified as the
193
BKca channel type reported in other tissues (Latorre et al., 1989). Representative records of the K current activated by 12.6/zM [Ca2+]~ in symmetrical 150 mM K ÷ solutions are shown in fig. 1A. The channel activity increased with increasing depolarization, as mentioned above, and at +40 mV the channel showed almost sustained bursting activity. The channel closing events of relatively long duration observed at negative potentials became less frequent with depolarization, leaving the repetitive brief closing events in the burst. After recording the control current, we applied 10/zM trifluoperazine to the internal surface of the membrane.
A
a
b
Control
.= ~2 v
C
TFP 10-8M
p=0.85
p=0.48
TFP 2XtOeM
p=0.26
[3.
A
L~
0
~ 10
0
0
0
0
10
10
Amplitude (pA)
A
Ca2* 1.26X1(~eM
b
a Control
B
TFP 1()5M
1.0
mv 40 2omv
~
2
0.5
-20 , ~ - ~ ~ ~
_,o
-
1.26XI0-5M
Ktw- ~4Xtd6M ~k
,
# __J
lOpA
lOOms 0
B
BKca lO 0 Control z~ TFP 1135M
i
-40
! ~
i
-20/[0
/
of° i
20
J
40my
/ O~
~
J
~
J
10-6
10-5
10-4
TFP (M)
pA
5
i
10-7
-5
-10 Fig. 1. (A) Single BKca currents and effects of trifluoperazine (TFP), recorded in the inside-out patch configuration. The channel was activated by 12.6 /zM [Ca2+ ]i. The pipette solution contained 150 mM K +. The standard internal solution (150 mM K + ) containing 10 IzM trifluoperazine was perfused after the control current had been recorded. The m e m b r a n e potentials are given on the left. T h e upward deflections represent the outward current, and the dashed lines indicate closed channel current level. (B) The voltage-current relationship of the BK¢~ current in the control condition (©) and in the presence of 10 IzM trifluoperazine (,x). The amplitude of the current was measured as described in fig. 2. The single channel conductance in this patch was 220 pS and was not affected by the drug. The straight line was drawn by eye. All data were taken from the same patch. The effects were reversible by washing out for several minutes.
Fig. 2. (A) The amplitude histograms of the BKc, currents recorded at + 20 mV in the absence (a) and presence of internal trifluoperazine (TFP: b 1 /zM and c 2/zM). The open probability (p) was 0.85 in control conditions and decreased to 0.48 and 0.26 in the presence of 1 and 2 /.tM of the drug, respectively. Ordinate: percentage of the total number of measured points. Abscissa: amplitude of the current in pA. (B) The dose-response curve for the inhibition of the open probability by trifluoperazine at + 20 inV. The relative value of the open probability in the presence of the drug (PTFP) to that in control (Pco.t) is plotted against the drug concentration. Each point represents the mean of five to seven measurements. Vertical bars indicate one S.D. A least-squares fit yielded an apparent dissociation constant (K D) of 1.4/zM and a Hill coefficient of 0.69.
The open probability of the channel was markedly reduced but the amplitude of the current was not changed by the drug (fig. lab). The reduction was greater at depolarized potentials. The amplitude of the single channel current was obtained as described below (see fig. 2A) and plotted against the membrane potential in fig. 1B (open circles). The K channel had an ohmic conductance of 220 pS with a reversal potential of 0 mV. The average conductance of 10 patches from 10 neurons was 244 +
194
17.5 pS. The current amplitude measured in the presence of trifluoperazine is also plotted in fig. 1B (open triangles), which shows that the single channel conductance was not altered by the drug. Amplitude histograms were constructed from records obtained at + 2 0 mV (fig. 2A). The open channel current showed a symmetrical distribution, and the very brief closing events caused a small continuous distribution between the full open level and the closed level under each experimental condition. The amplitude of the single channel current was measured
A
a
b
e
d
c
Control
TFP 10-6M
5XlO6M
2Xl()S M
,o[ •to o •
tau = 30.2ms
9.6ms
c
[I 1 2ms 30
'tL_"
1(?
10
10
I°Ik~ol
i,
0
100
0
25
50
0
11~Oo
where I stands for mean patch current, i for the unit current amplitude, and N for channel number within the patch. In the present study N was 1 in every experiment analyzed; p was 0.85 in the control condition at + 20 mV and was reduced to 0.48 and 0.26 by 1 and 2/~M trifluoperazine, respectively (fig. 2A). Figure 2B illustrates the dose-inhibition relationship of the drug at + 20 mV. The continuous line is a least-squares fit according to the following equation: P = K~/(D n + K~)
20
2O
(1)
I(~5M
n
3O
30
p = I/(i'N)
(2)
40~
L2rns
5.2ms
)
0
I
from the peak position, and was not changed by 1 and 2 /xM trifluoperazine (fig. 2Ab,c). The amplitude histogram also indicates the open probability (p) of the channel, which was calculated as the time during which the channel was in a open state:
t5
where P denotes the fig. 2), D the drug dissociation constant fitting gave a K D of 0.69.
relative value of p (PTFP/Pco,t in concentration, K D the apparent and n the Hill coefficient. The 1.4 txM and a Hill coefficient of
3.2. Open channel blockade by trifluoperazine 0
5
open Time (ms)
B ° 80O
Inorder to elucidate how trifluoperazine reduces the open probability, we proposed that trifluoperazine blocked open channels (Neher, 1983), since this mode of action has been proposed for neuroleptics in airway smooth muscle cells (McCann and Welsh, 1987). The model is described below with one closed state for the sake of simplicity: kt
ft'[D]
C ~ O ~
O'B
(3)
k_ m
/
400
/
where C denotes the closed state, O the open state and O • B the blocked state of the channel, kl, k_~, a and /3 are rate constants and [D] the concentration of the drug. The following three equations are derived from equation (3):
o
o
TFP (XlO-e M)
Fig. 3. (A) The open time histograms obtained in a patch u n d e r control conditions (a) and in the presence of trifluoperazine (TFP: 1, 2, 5 and 10 p~M in b, c, d and e, respectively). All histograms were fitted with single exponential functions. The time constant (tau) of the histogram, which reflects the m e a n open time, was 30.2 ms in control. Trifluoperazine decreased the m e a n open time in a dose-dependent manner. Superimposed curves are the best fits of single exponential functions. Ordinate: percentage of the total n u m b e r of the events. Abscissa: time in ms. (B) The rate constant ( 1 / t a u ) is plotted against the concentration of trifluoperazine. The slope gave /3 of 82 s -1 /zM - I in equation (4) in the text.
1 / ( m e a n open time) = k i + / 3 . [ D ]
(4)
1 / ( m e a n block time) = a
(5)
KD=a//3
(6)
Figure 3A illustrates the open time histograms, which were well fitted with a single exponential curve in the absence and presence of trifluoperazine. The drug decreased the time constant (tau), which describes the mean open time, in a dose-dependent manner. 1 / ( m e a n open time) is plotted against trifluoperazine concentration in fig. 3B. The data points fell on a straight line and the slope gave /3 a value of 82 s -1 /zM -~ in equation (4). The average of /3 with nine patches was 93.2 + 7.6 s -] /xM -~
195
A a
tau~--0.54rns
B
btau2--4m3mS
a
C tau =8.6ms
I
+,g4° t
el
P 30
20
10
oL
10
5
~
16~M
rFP
5.6ms
2C.
tN
C
30-
20
O-
b 2x IO"SM
20
0
50
25
Closed Time (ms)
25
k
0
Closed Time (ms)
25
Fig. 4. (A) The closed time histogram under control conditions. The histogram was fitted with three exponential functions with a time constant of 0.54, 4.3 and 8.6 ms, respectively. (B) In the presence of higher concentrations of trifluoperazine (TFP), the closed time histrogram was fitted with a single exponential function. The time constant was 6.3, 5.4 and 5.6 ms at (a) 2, (b) 5 and (c) 10/~M, respectively, giving the mean block time ( I / a ) in equation (5). Ordinate: percentage of the total number of the events. Abscissa: time in ms.
The closed time histogram was fitted with the sum of three exponential functions in the control condition (fig. 4A). In the presence of high concentrations of trifluoperazine, the histogram was fitted with a single exponential function, the time constant of which reflects the transition rate constant from O . B to O,
A
since most channels were blocked by the agent. As predicted by the model, the time constant of blockade (mean block time) did not vary over the concentration range of 2 - 1 0 / x M trifluoperazine (fig. 4B). The mean block time with 10 /zM of the drug was 5.6 + 0.3 ms (eight patches from eight neurons), and gave a a value
B-20mY
20mV
acontrol
bTFP 1()5 M
C -40mY %
a control
4
t~SM
acontrol
4
b F p Io-sM
4
t-
o(D CL p=0.89,
p=0.67
p=0.56
2
2
1
0.20
0.20
o,
0 0
10
0
10
, 0
, 10 Amplitude
10
0
10
10
(pA)
Fig. 5. The voltage-dependent reduction in the open probability by trifluoperazine (TFP). The drug decreased the open probability from 0.89 to 0.19 at +20 mV (A), from 0.67 to 0.20 at - 2 0 mV (B) and from 0.56 to 0.20 at - 4 0 mV (C). Ordinate: percentage of the total number of measured points. Abscissa: amplitude of the current in pA.
196 of 178/s. a and /3, being based on the sequential model in equation (6), yielded a K D value of 1.9 ~ M , which is close to the K D obtained from the reduction in the mean patch current (fig. 2B; 1.4 /~M). This finding indicates that trifluoperazine acts on the BKca by blocking open channels.
B
A 0
2z
KD(pM)
3.3. Voltage-dependent effects of trifluoperazine
/ o / ~ /
The voltage dependence of the effect of trifluoperazine, which was already seen in fig. 1A, was studied in more detail. Figure 5 depicts the voltage-dependent effects of trifluoperazine on the open probability. The channel was activated by 12.6 /~M Ca 2+ in a voltagedependent manner, namely, p = 0.89 at + 20 mV, 0.67 at - 2 0 mV and 0.56 at - 4 0 inV. Trifluoperazine (10 ~ M ) decreased the probability to 21, 30 and 36% of the control value at + 20, - 2 0 and - 4 0 mV, respectively. According to the sequential model, the K D can be determined as following (Benham et al., 1985; McCann and Welsh, 1987): KD = [D]'P'Pd/(P
-- P a )
(7)
where [D] denotes the drug concentration, and p and Pd the open probabilities in the absence and presence of the drug, respectively. To get a linear relation, equation (7) is modified as: (P-Pd)/P'Pd = [D]/KD
(8)
( P - P d ) / P ' P d iS plotted against the drug concentration in fig. 6A. The linear relationships indicate that equation (8) is valid for the data of this study and that the block occurs on a basis of a one-to-one binding reaction, which is consistent with the Hill coefficient of 0.69 in fig. 2B. The values of K D obtained from the slopes of the linear relationships are plotted against the voltage in fig. 6B, where the interpolation to 0 m V yielded the apparent dissociation constant at 0 mV (KD(0)). The binding site in the m e m b r a n e electrical field of a drug is determined by the following equation (Woodhull, 1973), KD
=
KD(0 ) .exp( - 6zFV~/RT)
(9)
where 6 is a partition parameter, Vm the m e m b r a n e potential, and z, F, R and T have their usual meanings in thermodynamics. To get a linear relationship, equation (9) is modified as: RT/zF.In(KD(0)/KD) = 6 'Vm
(10)
Figure 6C shows a plot of the left term of equation (10) against the voltage. Trifluoperazine dihydrochloride has two dissociation constants: pK 1 = 3.9 and p K 2 = 8.1. At physiological pH, namely 7.2 in the present experiments, the drug dissociated according to p K 2, with about 95% being monovalent. The valence of unity
,
lo TFP
20
-40
,
20
?" 2_0 mY
(pM)
C
mV
20[
0
-10
-20
o/
t L
Fig. 6. Estimation of the binding site for trifluoperazine (TFP) in the channel• (A) Determination of the K D for the blocking effect of the agent at +20 (©), -20 (o) and -40 mV (z~). (P--Pd)/P'Pd is plotted against the drug concentration. The reciprocal of the slope gave the K o at each voltage. (B) Interpolation of the KD-voltage plot yielded the KD(0). (C) RT/zF.In(KD(0)/K D) is plotted against the membrane potential. The slope gave a partition parameter, 6, of 0,47, which suggests that the binding site for trifluoperazine lies about half way across the electrical field from the internal pore of the channel.
(z = 1) was therefore used in equation (9). The value of 6 was 0.47 _+ 0.03 (n = 4), suggesting that the blocking site of trifluoperazine lies about half way from the internal pore across the m e m b r a n e electrical field.
4. Discussion Single channel recordings have been used in the study of drug effects on various types of ionic channels. We have reported that the charged form of local anesthetics inside a neuron produce an open state blockade of the BKca (Oda et al., 1992). Neuroleptics, including haloperidol and trifluoperazine, have been shown to block the open state of the BKca in dissociated smooth muscle cells, and the blocking effect was shown not to be related to the inhibition of calmodulin (McCann and Welsh, 1987; Kihara et al., 1990). Benishin et al. (1988) studied the effects of neuroleptics on CaZ+-de pendent 86Rb+ efflux in rat brain synaptosomes stimulated with 100 m M K +. At concentrations which did not depress 45Ca2+ uptake, trifluoperazine inhibited the 86Rb+ effiux with a biphasic d o s e - r e s p o n s e curve,
197
thereby suggesting the existence of two populations of Ca2+-activated K + channels with different sensitivities. The high sensitivity component (30-40% of the flux) had an ICs0 of 20-30 nM and the low sensitivity component an ICs0 of 1-3 tzM. The K D value of 1.4 /zM determined in our experiments is consistent with the value reported by Benishin et al. (1988), suggesting that the low sensitivity component corresponds to the BKc~ in hippocampal pyramidal neurons. The open probability was calculated as the time during which the channel was in an open state. Trifluoperazine reduced the open probability without changing the single channel conductance (fig. 2). The Kt~ for the inhibition was 1.4/zM and the Hill coefficient was 0.69 at +20 mV, values which are similar to those obtained by McCann and Welsh (1987) in cultured airway smooth muscle cells. We assumed in the present study that the drug simply blocked the open state of the channel as described in equation (3). On the basis of this assumption, the K D was calculated from the opening and closing rate constants in the presence and absence of the drug. We obtained a K o value of 1.9 p,M, which was close to that obtained from the mean patch current (1.4 p~M), a result supporting the assumption that trifluoperazine acts on the BKc, by blocking open channels. This finding differs from the results obtained with pentobarbitone: simple openchannel block failed to explain the open-close kinetics of acetylcholine-activated channels (Gage and McKinnon, 1985). The blocking action was greater at depolarized potentials, reflecting that the charged form of the drug molecule affected the channel. An analysis of the voltage dependence provided an estimation that the drug binds about half way across the membrane electrical field from the internal pore to block the channel (fig. 6). The voltage-dependent effect was also noticed in airway smooth muscle cells by McCann and Welsh (1987), who reported that the blocking site for haloperidol was about one third of the way across the electrical field. To our knowledge there have been no reports concerning the plasma concentration of trifluoperazine in clinical practice. The plasma concentration of chlorpromazine, another phenothiazine derivative, is reported to be about 1/zM under steady-state conditions (Cooper et al., 1976). The concentration of trifluoperazine may be lower since it is administered at smaller doses. Shortly after administration, however, the trifluoperazine concentration may well be high enough to block the BKc~. The drug may not affect the channel from the outside of the cell membrane, as was reported by McCann and Welsh (1987). Trifluoperazine, which has a pK z of 8.1, will dissociate according to the pH of the extracellular space (pH 7.4). About 15% of the molecules may stay uncharged and will rapidly pass through the membrane into the neuron until an equi-
librium is reached. The drug dissociates again according to the intracellular pH (about 7.0), where about 96% of the molecules may become charged. The charged form enters the internal pore of the channel and binds about half way from the internal surface with a one-to-one stoichiometry. Thus, the BKc~ becomes blocked, which results in a reduction in the outward current for repolarization and spike afterhyperpolarization, and then in a decrease in the intervals between the action potentials of neurons. This even may help the dopaminergic block produce some antipsychotic effects a n d / o r may induce seizure activities by facilitating burst firing of the neurons, since the hippocampus is sometimes a focus of seizures (Alger, 1984). In summary, we have shown that trifluoperazine blocks the open state of the BKc~ of dissociated rat hippocampal pyramidal neurons. Together with the suppression of dopaminergic transmission and the reduction in voltage-dependent currents, this blockade may contribute to the various actions of phenothiazines in the mammalian central nervous system.
References Alger, B.E., 1984, Hippocampus-electrophysiological study of epileptiform activity in vitro, in: Brain Slices, ed. R. Dingledine (Plenum Press, New York) p. 155. Baldessarini, R.J., 1990, Drugs and the treatment of psychiatric disorders, in: The Pharmacological Basis of Therapeutics, 8th edn., eds. A.G. Gilman, T.W. Rail, A.S. Nies and P. Tayler (Pergamon Press, New York) p. 383. Benham, C.D., T.B. Bolton, R.J. Lang and T. Takewaki, 1985, The mechanism of action of Ba 2+ and TEA on single Ca2+-activated K+-channels in arterial and intestinal smooth muscle cell membranes, Plfiigers Arch. 403, 120. Benishin, C.G., B.K. Krueger and M.P. Blaustein, 1988, Phenothiazines and haloperidol block Ca-activated K channels in rat fore brain synaptosomes, Mol. Pharmacol. 33, 195. Billard, W., V. Ruperto, G. Crosby, L.C. lorio and A. Barnett, 1984, Characterization of the binding of 3H-SCH 23390, a selective D-I receptor antagonist ligand, in rat striatum, Life Sci. 35, 1885. Civelli, O., J. Bunzow, H. Van Tol, D. Grandy, P. Albert, J. Salon, C. Machida and K. Neve, 1989, Cloning of a rat Dz-receptor cDNA, in: Molecular Biology of Neuroreceptors and Ion Channels, NATO ASI Series, Vol. 32, ed. A. Maelicke (Springer Verlag, Berlin) p, 259. Cooper, T.B., G.M. Simpson and J.H. Lee, 1976, Thymoleptic and neuroleptic drugs plasma levels in psychiatry: current status, Int. Rev. Neurobiol. 19, 269. Dinah, T.G., V. Crunelli and J.S. Lelly, 1987, Neuroleptics decrease calcium-activated potassium conductance in hippocampal pyramidal cells, Brain Res. 407, 159. Fabiato, A. and F. Fabiato, 1979, Calculator programs for computing the composition of the solutions containing multiple metals and ligands used for experiments in skinned muscle cells, J. Physiol. (Paris) 75, 463. Gage, P.W. and D. McKinnon, 1985, Effects of pentobarbitone on acetylcholine-activated channels in mammalian muscle, Br. J. Pharmacol. 85, 229. Halliwell, J.V., 1990, K + channels in the central nervous system, in: Potassium Channels: Structure, Classification, Function and
198 Therapeutic Potentials, ed. N.S. Cook (Ellis Horwood Limited, Chichester) p. 348. Hamill, O.P., A. Marty, A.E. Neher, B. Sakmann and F.J. Sigworth, 1981, Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches, PfliJgers Arch. 391, 85. Ikemoto, Y., K. Ono, A. Yoshida and N. Akaike, 1989, Delayed activation of large-conductance Ca2+-activated K channels in hippocampal neurons of the rat, Biophys. J. 56, 207. ltil, T.M., 1978, Effects of psychotropic drugs in qualitatively and quantitatively analyzed human EEG, in: Principles of Psychopharmacology, 2nd edn., eds. W.G. Clark and J. Del Giudice (Academic Press, New York) p. 261. Kaneda, M., H. Nakamura and N. Akaike, 1988, Mechanical and enzymatic isolation of mammalian CNS neurons, Neurosci. Res. 5,299. Kihara, M., K. Matsuzawa, H. Tokuno and T. Tomita, 1990, Effects of calmodulin antagonists on calcium-activated potassium channels in pregnant rat myometrium, Br. J. Pharmacol. 100, 353. Lancaster, B. and R.A. Nicoll, 1987, Properties of two calciumactivated hyperpolarizations in rat hippocampal neurons, J. Physiol. (London) 389, 187. Latorre, R., A. Oberhauser, P. Labarca and O. Alvarez, 1989, Varieties of calcium-activated potassium channels, Ann. Rev. Physiol. 51,385. McCann, J.D. and M.J. Welsh, 1987, Neuroleptics antagonize a calcium-activated potassium channel in airway smooth muscle, J. Gen. Physiol. 89, 339. Neher, E., 1983, The charge carried by single-channel currents of rat cultured muscle cells in the presence of local anaesthetics, J. Physiol. (London) 339, 663.
Oda, M., A. Yoshida and Y. Ikemoto, 1992, Blockade by local anesthetics of the single Ca2+-activated K + channel in rat hippocampal neurones, Br. J. Pharmacol. 105, 63. Ogata, N. and H. Tatebayashi, 1989, Modulation of sodium current kinetics by chlorpromazine in freshly-isolated striatal neurons of the adult guinea-pig, Br. J. Pharmacol. 98, 1173. Ogata, N., M. Yoshii and T. Narahashi, 1989, Psychotropic drugs block voltage-gated ion channels in neuroblastoma cells, Brain Res. 476, 140. Ogata, N., M. Yoshii and T. Narahashi, 1990, Differential block of sodium and calcium channels by chlorpromazine in mouse neuroblastoma cells, J. Physiol. (London) 420, 165. Seeman, P., 1980, Brain dopamine receptors, Pharmacol. Rev. 32. 229. Storm, J.F., 1987, Action potential repolarization and a fast after-hyperpolarization in rat hippocampal pyramidal cells, J. Physiol. (London) 385, 733. Tsien, R.Y. and T.J. Rink, 1980, Neutral carrier ion-selective microelectrodes for measurement of intracellular free calcium, Biochim. Biophys. Acta 599, 623. Uchimura, N., H. Higashi and S. Nishi, 1986, Hyperpolarizing and depolarizing actions of dopamine via D-1 and D-2 receptors on nucleus accumbens neurons, Brain Res. 375, 368. Woodhull, A.M., 1973, Ionic blockade of sodium channels in nerve, J. Gen. Physiol. 61,687. Yoshida, A., M. Oda and Y. Ikemoto, 199i, Kinetics of the Ca 2÷activated K + channel in rat hippocampal neuron, Jap. J. Physiol. 41,297.
