|
|
||||||||
Ernest Gallo Clinic and Research Center, Department of Neurology, University of California, San Francisco, Emeryville, California
Submitted 24 July 2007; accepted in final form 13 August 2007
| ABSTRACT |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
From a physiological perspective, midbrain DA neurons present two patterns of activity—regular, pacemaker firing and burst firing (reviewed in Cooper 2002
; Komendantov et al. 2004
)—that are considered important for control of different physiological responses. Tonic, pacemaker-like firing may be altered under conditions ranging from withdrawal from addictive drugs to schizophrenia (Grace 2000
; Weiss et al. 1992
). In addition, burstlike firing in VTA DA neurons, which may produce phasic increases in DA in the different target areas of the VTA (reviewed in Overton and Clark 1997
), is thought to encode salience of important stimuli (Robinson and Berridge 2001
) and may represent a teaching signal (Schultz 2002
).
A number of channels, including small conductance, calciumactivated potassium channels (SK) and inwardly rectifying hyperpolarization-activated cation channels (HCN, responsible for generating the Ih current), can regulate pacemaker and burst firing (Johnson and Seutin 1997
; Neuhoff et al. 2002
; Seutin et al. 1993
, 2001
). However, little is known about the long-term effects of chronic ethanol treatment on the firing properties of VTA DA neurons. Furthermore, intermittent ethanol exposure can enhance the locomotor-activating effects of cocaine (Itzhak and Martin 1999
), but the cellular mechanisms underlying hyperlocomotion following this pattern of ethanol exposure are poorly understood. In this study, we evaluated the function of different channels involved in control of firing in VTA DA neurons 7 days after the end of a 5-day repeated, intermittent ethanol treatment, and also the ability of repeated ethanol exposure and withdrawal to enhance locomotor activation by cocaine.
| METHODS |
|---|
|
|
|---|
Male Sprague–Dawley rats were singly housed in a temperature-controlled (21 ± 1°C) room on a 12-h light/dark cycle (lights on at 07:00). Food and water were available without restriction. Rats were acclimated to handling procedures starting 3 days before experiments began. All ethanol injections and behavioral testing were performed during the light cycle. Rats were injected with ethanol [2 g/kg, administered intraperitoneally (ip), twice a day at 10:00 and 18:00, from a 20% vol/vol solution] or an equivalent volume of saline for 5 days, with injections starting at postnatal day (P) 21.
Slice preparation
All electrophysiological recordings were performed in rats 7 days after the end of repeated treatment with ethanol or saline (described earlier). To prepare VTA slices, animals were anesthetized with halothane and perfused transcardially with about 10 ml of nearly frozen (
0°C) perfusion solution at a rate of about 10 ml/min. This perfusion solution was saturated with 95% O2-5% CO2 and contained (in mM): 119 NaCl, 2.5 KCl, 1.0 NaH2PO4, 6.1 MgCl2, 0.1 CaCl2, 26.2 NaHCO3, 1.25 glucose, 50 sucrose, 1 ascorbic acid, and 3 kynurenic acid. Horizontal slices of 230 µm containing the VTA were prepared with a VT1000S vibratome (Leica, Nussloch, Germany) in the same perfusion solution at 4°C. Slices were then placed in a holding chamber at 31–32°C containing artificial cerebrospinal fluid (aCSF) saturated with 95% O2-5% CO2 and composed of (in mM): 119 NaCl, 2.5 KCl, 1.0 NaH2PO4, 1.3 MgCl2, 2.5 CaCl2, 26.2 NaHCO3, and 11 glucose. Ascorbic acid (1 mM) was added about 15 min before slices were placed in the recovery chamber. Slices were allowed to recover for
45 min before being placed in the recording chamber and superfused with oxygenated aCSF at 31–32°C with picrotoxin (100 µM) added to block
-aminobutyric acid type A (GABAA)–receptor-mediated inhibitory postsynaptic currents.
Electrophysiology
Cells were visualized using infrared differential interference contrast video microscopy. Whole cell current- and voltage-clamp recordings were made using a Multiclamp 700A or 700B amplifier and Clampex 9.2 (Axon Instruments, Foster City, CA). Electrodes (2.8–4.0 M
) generally contained: 130 mM KOH, 105 mM methanesulfonic acid, 17 mM hydrochloric acid, 20 mM HEPES, 0.2 mM EGTA, 2.8 mM NaCl, 2.5 mg/ml MgATP, and 0.25 mg/ml GTP (pH 7.2–7.4, 280–290 mOsm). A low level of the calcium buffer EGTA was included in the pipette solution to preserve calcium-dependent potassium currents during whole cell current- and voltage-clamp recordings (Wolfart et al. 2001
). For experiments examining brief (30 s) bath application of N-methyl-D-aspartate (NMDA) under voltage clamp (Vholding = +40 mV), the internal solution was (in mM): 117 Cs-methanesulfonic acid, 20 HEPES, 0.4 EGTA, 2.8 NaCl, 5 TEA-Cl, 2.5 Mg2ATP, and 0.25 Mg2GTP (pH 7.2–7.4, 280–290 mOsm). The reversal potential of NMDA currents was not determined.
Current-clamp data were acquired at 20 kHz and filtered at 2 kHz and voltage-clamp data were acquired at 10 kHz and filtered at 2 kHz. VTA DA neurons were identified by the presence of a large Ih current (Johnson and North 1992
). Because Ih is present in both principal and tertiary VTA neurons (Margolis et al. 2006
), we recognize that its presence does not unequivocally identify DA neurons in midbrain slices. However, in previous work (Ungless et al. 2001
) and in the present study, this criterion was enough to obtain differences between control and experimental treatments. In addition, prior results from our laboratory (Borgland et al. 2006
and F. Sarti, unpublished observations) suggest a higher (
75%) correlation between the presence of Ih and tyrosine hydroxylase, indicative of DA neurons, in neurons just medial to the medial terminal nucleus of the accessory optic tract (MT), where nearly all the recordings in the present study were performed. Also, the decrease in Ih amplitude in ethanol animals was only about 40% (![]()
![]()
![]()
![]()
![]()
Fig. 7) and thus was not sufficient to prevent the use of Ih to identify putative DA neurons relative to GABA neurons, which do not have an Ih current (Johnson and North 1992
; Margolis et al. 2006
).
|
|
|
|
|
|
|
10 sequential APs with a coefficient of variation of the interspike interval (ISI) of <20%. The average basal instantaneous firing frequency and ISI during periods of pacemaker firing were then determined across 30 successive APs just after breaking into a neuron, before any dialysis of the neuron with the intracellular pipette solution could occur. However, prolonged dialysis with the internal solution did not alter the firing pattern or AP waveform of neurons (data not shown). In addition, due to variability among neurons in the basal pacemaker firing rate, neurons were brought to about 1 Hz by injecting DC current through the patch amplifier before experiments examining the effects of NMDA, 4-(N-ethyl-N-phenylamino)-1,2-dimethyl-6-(methylamino) pyridinium chloride (ZD7288), or recovery from hyperpolarization.
All voltage values were corrected for the liquid junction potential, estimated to be 10 mV using the Junction Null Calculator in Clampex 9.2, and also by direct measurement of the potential difference between internal and external solutions present after zeroing the pipette current. Bridge balance was used to compensate 60–80% of the series resistance. Also, because the average series resistance and instantaneous currents evoked by hyperpolarization (taken as a measure of input resistance) were the same between ethanol and saline animals (see RESULTS), any errors in compensation would be similar in both groups and thus would not explain the differences in channel function observed here. Further, to estimate whether repeated ethanol exposure and withdrawal might have altered the size of VTA neurons, we estimated the whole cell capacitance using a method adapted from Gentet et al. (2000)
. In Ih experiments, the voltage step from –60 to –70 mV was used for analysis of capacitance, where the series resistance was determined from the peak of the capacitive transient; the input resistance was determined 40 ms into plateau of the voltage response; the capacitive current transient just after application of the voltage step was best fit with two exponentials; and the fast exponential was used in the determination of cell capacitance. For SK experiments, the voltage step from –70 to –60 mV was used for capacitance analysis. Both methods gave equivalent measures of capacitance (data not shown).
To investigate SK tail currents, neurons were held at –70 mV, and depolarizing current steps (400 ms, from –60 to –20 mV in 10-mV steps) were applied, with 1 s between successive steps. On returning to –70 mV, a tail current was evident (see Fig. 3A). The peak magnitude of the SK tail current was determined for the steps to –20, –30, and –40 mV. In addition, the charge transfer for SK (in picocoulombs) was calculated by integrating (nanoamperes x milliseconds) the tail current evoked after the depolarizing pulse, beginning from the initiation of the tail current to 550 ms into the tail current because the apamin-sensitive component of the tail current should be <5% by 550 ms into the tail current (Abel et al. 2004
). The area under the curve was determined using GraphPad Prism (San Diego, CA). Finally, the rate of inactivation of the SK current was estimated in the tail current evoked after a –20-mV depolarization by fitting a single exponential from about 55 ms after the peak of the tail current to 550 ms after the peak of the tail current (Abel et al. 2004
).
To examine Ih, neurons were held at –60 mV and a series of voltage steps (500 ms, from –40 to –150 mV in 10-mV steps, with 4 s between steps) was applied. In some Ih voltage-clamp experiments, to reduce contamination from other currents and to better isolate Ih, we performed experiments under conditions where many currents other than Ih were inhibited using a modified internal solution (3 mM BAPTA added) and external solution (aCSF containing 2 mM tetraethylammonium, 20 mM MgCl2, 0.5 µM tetrodotoxin, and 2 mM 4-aminopyridine, with NaCl reduced in an equimolar fashion) (Liu et al. 2003
).
Ih was determined from the difference between the peak current response at the end of the pulse and the instantaneous current response about 30 ms after the onset of the hyperpolarizing pulse (see Fig. 7A; Watts et al. 1996
). In addition, we determined the Ih charge transfer using methods similar to those described earlier for SK, except that Ih charge transfer was determined by integrating the slowly activating inward current beginning from 30 ms after the initiation of the –140-mV hyperpolarizing voltage pulse (Neuhoff et al. 2002
). We also determined the magnitude and charge transfer of the instantaneous current, which could represent inwardly rectifying potassium channels (IRK) and/or potassium leak channels (Watts et al. 1996
; Wilson 2005
). We did not perform experiments to determine the relative contribution of these currents to the instantaneous current. The voltage dependence of activation of Ih was determined as described (Liu et al. 2003
) using the equation G = I/ (E – Erev), where –55 mV was used for the Erev for Ih because it was not different between saline and ethanol animals (Fig. 8F). Data were fit with a Boltzmann equation (Liu et al. 2003
) using GraphPad Prism to determine the half-activation voltage (V1/2) and slope factor. The Ih reversal potential was determined using described methods (Cathala and Paupardin-Tritsch 1997
; Mayer and Westbrook 1983
), where the reversal potential of Ih was estimated by comparing the instantaneous current evoked from a voltage of –60 mV, where Ih was not active, and the instantaneous current evoked from a voltage of –80 mV, where Ih was activated (see Fig. 8E). Because instantaneous currents evoked from –60 mV likely represent IRK/leak channels (Watts et al. 1996
; Wilson 2005
), whereas currents evoked from –80 mV represent Ih in addition to IRK/leak, one can estimate the reversal potential of Ih by determining the point where the two current–voltage (I–V) curves intersect, which essentially represents the voltage where the Ih current is zero after the IRK/leak current is subtracted out.
|
Motor activity assay
Locomotor activity was measured in a 17 x 17-in. open-field chamber lined with three 16-beam infrared arrays. Distance traveled (in centimeters) was measured using Open Field Activity software (MED Associates, St. Albans, VT). Animals were habituated to the locomotor chamber during the 4th through 6th days after the last ethanol injection. Each day, the animals were allowed 1 h of free exploration, then were given an ip injection of saline (1 ml/kg) and then returned to the chamber for 30 min. This was followed by a second ip injection of saline and a final 30 min in the locomotor chamber. On the 7th day after ethanol injection, animals were allowed 1 h of free exploration, followed by an injection of saline and then 30 min in the locomotor chamber. The animals then received an ip injection of cocaine (15 mg/kg) and were returned to the chamber for a final 30 min.
One concern with repeated injections is the possibility of stress-induced changes. However, as subsequently described, saline and naïve neurons exhibited a similar afterhyperpolarization (AHP) and voltage-clamp measures of SK and Ih, and saline animals exhibited a locomotor response to cocaine very similar to that in naïve animals. These results suggest that any stress related to multiple injections or handling procedures did not alter channel function or enhance the locomotor response to cocaine. As an additional indicator of possible stress due to injections, we examined the weight of the animals across the period of injection. Saline and ethanol animals showed a similar weight before injections (saline: 51.6 ± 1.4 g; ethanol: 52.2 ± 1.2 g, n = 25 both for saline and ethanol), but by the 5th day the weight of ethanol animals was slightly but significantly reduced (
10%) relative to saline animals (saline: 74.9 ± 1.6 g; ethanol: 68.4 ± 1.7 g, P < 0.05). The cause of this difference is unclear, but may relate to differences in activity of the animals, with two periods of intoxication and lethargy each day in ethanol but not saline animals. However, the weight of saline and ethanol animals was not significantly different after the 7-day withdrawal period (saline: 129.6 ± 3.4 g; ethanol: 121.0 ± 5.2 g), suggesting that repeated ethanol injection did not result in persisting weight loss due to stress associated with injection.
Reagents
All reagents were purchased from Sigma (St. Louis, MO), except for ZD7288, which was purchased from Tocris Cookson (Ellisville, MO). Reagents were prepared in a 1:1,000 stock that was frozen in aliquots at –20°C, except for cocaine, which was made fresh on the day of experimentation. Apamin was prepared in distilled water and ZD7288 in DMSO.
Statistical analysis
All data are expressed as means ± SE. Statistical analysis was performed with a two-tailed, unpaired t-test, except, where indicated, a two-way, repeated-measures ANOVA was used. Significance was determined using GraphPad Prism with confidence intervals of
95%. All exponentials were fit using Clampfit. During experiments aimed at examining changes in firing during NMDA application, the coefficient of variance of the interspike interval (CV-ISI) was determined from 100 successive action potentials. CV-ISI was calculated as the SD divided by the mean and expressed as a percentage (Zhang et al. 1994
).
| RESULTS |
|---|
|
|
|---|
Pacemaker activity was not altered after repeated ethanol treatment followed by 7 days withdrawal
We first investigated the basal pacemaker activity in VTA DA neurons. Here, neurons were considered able to fire in pacemaker mode if they were able to fire
10 sequential APs (see METHODS). We recorded neurons meeting this criterion from 18 saline and 19 ethanol animals. The average basal instantaneous firing frequency and ISI during periods of pacemaker firing were determined across 30 successive action potentials just after breaking into a neuron. When pacemaker firing occurred, there were no significant differences in the average instantaneous firing frequency (Fig. 1B; saline, 1.19 ± 0.07 Hz; ethanol, 1.27 ± 0.12 Hz; not significant, P > 0.05 (n.s.)) or the average ISI (Fig. 1C; saline, 958 ± 69 ms; ethanol, 967 ± 109 ms; n.s.) of neurons from saline and ethanol rats, with firing rates similar to those in several other studies using whole cell recording (Margolis et al. 2006
; Marinelli et al. 2005
). In addition, although a resting membrane potential was difficult to determine because neurons were undergoing pacemaker firing, the average membrane potential during pacemaker firing was not different between saline and ethanol animals (saline, –59.1 ± 1.2 mV; ethanol, –59.3 ± 0.6 mV, n.s.), in agreement with no change in many aspects of the action potential waveform (see following text).
SK function was reduced in ethanol animals
To further characterize the firing properties of VTA DA neurons, we examined the action potential waveform of neurons from saline and ethanol animals (n = 21 and 20, respectively). A fundamental feature of neuronal firing is the generation of action potentials, and analysis of specific components of the action potential waveform can often provide evidence that the function of a particular channel has been altered. For example, decreased Na+ channel function could be reflected by an increase in action potential threshold and a decrease in the action potential peak (Zhang et al. 1998
). However, neither the threshold nor the peak of the action potential were different between saline and ethanol animals (Fig. 2, A and B; threshold: saline, –43.5 ± 0.5 mV; ethanol, –44.1 ± 0.7 mV; peak: saline, 17.6 ± 1.4 mV; ethanol, 16.3 ± 1.5 mV; n.s.), suggesting no changes in Na+ channel function. Further, a reduction in the peak value of the AHP is often observed during inhibition of fast-repolarizing potassium channels (such as BK and Kv3 delayed rectifier channels; Appel et al. 2003
; Grillner and Mercuri 2002
). However, the peak value of the AHP was not altered in ethanol animals (Fig. 2, A and B; saline, –69.0 ± 1.7 mV; ethanol, –69.8 ± 0.8 mV; n.s.), suggesting no change in function of the fast-repolarizing channels. In addition, the action potential width, determined halfway between the action potential threshold and peak, was not different between saline and ethanol animals (saline, 1.27 ± 0.05 ms; ethanol, 1.29 ± 0.05 ms; n.s.).
These data suggest that some of the basic electrophysiological properties of VTA DA neurons were not modified by withdrawal after repeated ethanol treatment. However, the time-to-peak of the AHP (TTP-AHP), determined relative to the action potential threshold, was significantly reduced in ethanol animals (Fig. 2C shows results from neurons treated with apamin: saline, 70.9 ± 8.7 ms; ethanol, 46.6 ± 7.0 ms; n = 5 for saline and 10 for ethanol, P < 0.05; significant differences were also apparent in an additional set of neurons from 17 saline and 14 ethanol animals in which apamin was not tested: saline, 56.0 ± 4.8 ms; ethanol, 37.5 ± 4.6 ms; P < 0.05). Although a number of potassium channels contribute to the AHP, the small-conductance, calcium-dependent potassium current (SK) is of particular interest because it is hypothesized to regulate the transition from pacemaker to burst firing (Komendantov et al. 2004
). Neurons from saline animals exhibited a very broad AHP with a greatly delayed TTP-AHP, typical of naïve neurons with a strong SK (age-matched naïve, 61.2 ± 11.1 ms, n = 7; and see Ping and Shepard 1996
; Shepard and Bunney 1988
). Further, the SK-selective antagonist apamin (200 nM) significantly reduced the TTP-AHP in both saline and ethanol animals (Fig. 2, D and F; P < 0.05; n = 5 for saline and n = 10 for ethanol), with a significantly greater change in the TTP-AHP in saline animals (Fig. 2, E and F; saline, 60.6 ± 9.9 ms; ethanol, 35.8 ± 6.3 ms; P < 0.05), and resulted in a similar TTP-AHP in both groups (Fig. 2D; saline, 10.4 ± 1.5 ms; ethanol, 11.8 ± 1.1 ms; n.s.). Because SK was a dominant contributor to the TTP-AHP (see also Ping and Shepard 1996
; Shepard and Bunney 1991
), the significantly reduced TTP-AHP in ethanol animals suggests that withdrawal after repeated ethanol exposure reduced SK function.
To more directly investigate the effects of ethanol withdrawal on SK function, we performed voltage-clamp experiments to better isolate the apamin-sensitive SK current. Neurons were held at –70 mV, then depolarized for 400 ms with pulses ranging from –20 to –60 mV in 10-mV steps. On returning to –70 mV, a tail current due to slow deactivation of a channel was evident (see Fig. 3A; see also Paul et al. 2003
). Both the peak amplitudes of the tail currents evoked after depolarizing steps to –20, –30, and –40 mV and the tail current charge transfer, calculated by integrating the tail current evoked after the depolarizing pulse (see METHODS), were significantly reduced in ethanol animals (Fig. 3, A–D), with saline similar to age-matched naïve animals (naïve, 240 ± 40 pA for step to –20 mV, n = 7; compared with 239 ± 12 pA for saline and 178 ± 13 pA for ethanol). In addition, the tail current was inhibited by >90% by apamin in both saline and ethanol animals (analyzed for the depolarizing step to –20 mV; saline, 92.9 ± 2.5% block; ethanol, 92.8 ± 2.5% block; n = 6 and 9 for saline and ethanol, respectively), indicating that SK is the primary channel active during the tail current following depolarization (see also Paul et al. 2003
). In this regard, the apamin-sensitive component of the peak tail current was significantly reduced in ethanol relative to saline animals (Fig. 3E; saline, 60.6 ± 9.9 pA; ethanol, 34.8 ± 5.8 pA, P < 0.05), strongly suggesting that SK function was reduced in ethanol animals.
Differences in SK function could result from changes in basic membrane properties. However, neither the series resistance (saline, 14.7 ± 0.9 M
; ethanol, 16.1 ± 0.8 M
; n.s.) nor the instantaneous currents evoked by hyperpolarization (see Fig. 7, E and F, taken as a measure of input resistance) were altered in ethanol animals. In addition, we estimated whether cell capacitance might be different in neurons from saline and ethanol animals, perhaps indicating differences in cell size (see Neuhoff et al. 2002
), following a method adapted from Gentet et al. (2000)
. Although there could be considerable error using this method due to space-clamp issues associated with low-pass filtering and voltage attenuation during patch clamp of neurons in brain slice with intact dendrites (Major 1993
; Spruston et al. 1993
), as well as a lack of direct measurement of the cell surface area, we observed no differences in the fast exponential capacitive transient after a voltage step from –70 to –60 mV (saline, 0.99 ± 0.06 ms; ethanol, 1.08 ± 0.05 ms; n.s.) or the cell capacitance (saline, 76.7 ± 3.4 pF; ethanol, 77.6 ± 4.0 pF; n.s.). Further, differences between saline and ethanol animals in integrated currents of SK (Fig. 3C) and Ih (see following text) persisted when normalized to the cell capacitance estimated in this way (data not shown). Taken together with a lack of differences in series resistance or the instantaneous currents evoked by hyperpolarization, these results suggest that apparent changes in SK function did not result from differences in basic membrane parameters.
Reduced SK function in ethanol animals could be due to several factors, including altered kinetics or reversal potential. However, the time constant of inactivation (fit from 55 to 550 ms into the tail current; Abel et al. 2004
) was not different between ethanol and saline animals (saline, 139 ± 6 ms; ethanol, 140 ± 7 ms; n.s.). Also, the reversal potential of SK (determined by depolarizing to –20 mV for 400 ms, then stepping the cell to voltages ranging from –70 to –120 mV in 5-mV increments, in the presence of the Ih blocker ZD7288, 30 µM) was not different between groups (saline, –89.0 ± 3.2 mV; ethanol, –87.7 ± 1.9 mV; n = 4 for each group; n.s.). The deviation of the SK reversal potential relative to the EK+ derived from the Nerntz equation (
–100 mV) could be due to small space-clamp errors while patch-clamping neurons in a brain slice and/or localization of SK channels in the dendrites (Abel et al. 2004
). Additionally, smaller SK amplitudes in ethanol animals might impair accurate determination of the reversal potential in those neurons. However, because the whole cell capacitance, input resistance, and series resistance were similar between saline and ethanol animals, any such errors should be equivalent in both groups.
NMDA burst firing was enhanced in ethanol animals
In addition to shaping the AHP, SK can regulate the firing pattern of VTA DA neurons. In particular, studies have demonstrated that SK inhibition facilitates the transition to bursting after application of NMDA (Johnson and Seutin 1997
; Seutin et al. 1993
). Entry of Na+ and Ca2+ after activation of NMDA receptors produces a depolarization and an increase in intracellular Ca2+ that, in addition to the inactivation of SK (which reduces the AHP), may facilitate the transition from pacemaker firing to burst firing (Johnson and Seutin 1997
; Komendantov et al. 2004
; Seutin et al. 1993
). Because our action potential waveform analysis and voltage-clamp data suggest that SK function was reduced in VTA DA neurons from ethanol animals, we studied whether reduced SK function in ethanol animals would facilitate the ability of NMDA to induce bursting activity. To more precisely quantify bursting, we determined the coefficient of variation of the interspike interval (CV-ISI), a measure of spike train irregularity, before and after NMDA application. CV-ISI was defined as the SD divided by the mean of the ISI distribution (Zhang et al. 1994
) and was calculated across 100 action potentials. The change from pacemaker to bursting activity increases the CV-ISI, where the CV-ISI is small for very regular spike trains (like pacemaker) and larger in irregular or bursting spike trains.
Bath application of NMDA (20 µM) for 10 min increased the firing frequency in neurons from saline animals, but generally did not induce bursting activity, indicated by no difference in the CV-ISI before and after application of 20 µM NMDA (Fig. 4, A and D; pre-NMDA, 23.8 ± 4.8%; post-NMDA, 31.1 ± 9.6%; n = 10; n.s.). In contrast, the same dose of NMDA (20 µM; n = 11) induced bursting activity in ethanol animals, indicated by a significant increase in CV-ISI (Fig. 4, A and D; pre-NMDA, 27.5 ± 5.5%; post-NMDA, 79.4 ± 16.1%; P < 0.05), with a significant difference in the CV-ISI between saline and ethanol after 20 µM NMDA (P < 0.05). Example histograms of the ISI distribution before and after 20 µM NMDA are shown in Fig. 4, B and C, and demonstrate that NMDA enhanced firing rate but maintained a regular firing pattern in saline animals, whereas NMDA produced an ISI distribution with both short ISIs (during a burst) and longer ISIs (between bursts) in ethanol animals.
A lower dose of NMDA (10 µM) increased firing rate but did not produce bursting or alter the CV-ISI in either group (Fig. 4D; saline, 20.2 ± 7.5% post-NMDA; ethanol, 33.0 ± 10.7% post-NMDA; n = 7 and 6 for saline and ethanol, respectively). A lack of bursting in saline animals was not due to insufficient NMDA receptor activation because 50 µM NMDA also did not significantly increase the CV-ISI in saline animals (Fig. 4D; 34.1 ± 14.1% post-NMDA; n = 5). However, neurons from saline animals exhibited bursting and a significantly elevated CV-ISI when 20 µM NMDA was combined with SK inhibition by apamin (Figs. 4D and 5B; 78.8 ± 6.2% post-NMDA; n = 5; P < 0.05), in agreement with studies showing bursting after NMDA and apamin in combination (Johnson and Seutin 1997
; Seutin et al. 1993
).
Although NMDA alone generally did not elicit bursting in saline neurons, bursting was occasionally observed (an example is shown in Fig. 5B). Thus to further define whether bursting was induced within a particular neuron, a criterion was defined whereby individual neurons would be considered bursting if they had a CV-ISI greater than the mean + 2 SDs of the CV-ISI values taken from neurons from saline animals before NMDA addition, yielding a threshold of a 68.5% CV-ISI (Fig. 5A). With this criterion, bursting was observed in saline animals in 2 of 10 neurons exposed to 20 µM NMDA, 1 of 5 neurons exposed to 50 µM NMDA, but 4 of 5 neurons exposed to 20 µM NMDA combined with apamin (Fig. 5A). Further, bursting was observed in 8 of 11 neurons from ethanol animals exposed to 20 µM NMDA (Fig. 5A). Thus NMDA alone can induce bursting in some neurons from saline animals, but the incidence of bursting in 20 µM NMDA was significantly greater in ethanol animals (P < 0.05,
2 test). Further, inhibition of SK at the same time as depolarization with NMDA strongly facilitated the transition to bursting in saline animals and generated bursting that was very similar to that observed in ethanol animals exposed only to NMDA (Figs. 4D and 5B).
Because there was variability among neurons regarding the pre-NMDA baseline (Fig. 5A), we also calculated a CV-ISI ratio, defined as the ratio of the CV-ISI after NMDA to the CV-ISI before NMDA. The CV-ISI ratio was significantly greater in ethanol relative to saline neurons (Fig. 5C; saline, 1.3 ± 0.3; ethanol, 3.0 ± 0.6; P < 0.05), in agreement with a greater increase in variability in ISIs that typifies burst firing (see Fig. 4, A–D). In addition, we observed that the CV ratio was significantly negatively correlated with the SK tail current (Fig. 5D; r2 = 0.387; P < 0.05). Because neurons from ethanol animals that exhibited smaller SK function showed a significantly larger increase in the CV ratio, this suggests that reduced SK function might be responsible for the enhanced NMDA-induced transition to bursting in ethanol animals.
Finally, differences in NMDA-induced bursting could result from different NMDA-dependent receptor activation. To examine this possibility, we determined the current response to a 30-s exposure to NMDA (Vholding = +40 mV) under voltage clamp with a Cs+-based internal solution (Borgland et al. 2006
). Bath application of 10 or 20 µM NMDA produced a dose-dependent increase in evoked current that was similar in neurons from saline and ethanol animals (Fig. 6; n = 6 for all groups and doses), suggesting that NMDA-receptor activation was not different between the two groups. Thus bursting induced by NMDA was significantly facilitated in ethanol animals, with no change in NMDA-receptor activity, whereas neurons from saline animals could exhibit bursting with 20 µM NMDA combined with apamin, but generally did not exhibit bursting after NMDA alone. Taken together, these results strongly suggest that the transition to burst firing was facilitated in ethanol animals and that the reduced SK function in ethanol animals might play a key role in this enhanced transition to bursting.
Ih function was reduced and recovery from hyperpolarization impaired in ethanol animals, but reduced Ih did not account for enhanced transition to bursting
The Ih current can also regulate the activity of VTA DA neurons (Neuhoff et al. 2002
; Okamoto et al. 2006
; Seutin et al. 2001
). To measure Ih, we performed voltage-clamp experiments where a series of hyperpolarizing voltage steps (500 ms, ranging from –40 to –150 mV in 10- mV steps) was applied to VTA neurons from a holding potential of –60 mV, and the magnitude of Ih determined from the slowly developing current sag (Cathala and Paupardin-Tritsch 1997
; Liu et al. 2003
; Neuhoff et al. 2002
). We also determined the Ih charge transfer by integrating the area under the slowly activating inward current in the –140-mV hyperpolarizing voltage pulse (see METHODS). Finally, to better isolate Ih, we performed some voltage-clamp experiments under conditions where many currents other than Ih were inhibited using modified internal and external solutions (see METHODS and Liu et al. 2003
).
Figure 7, A and B shows examples of the current response to a step from –60 to –140 mV in both experimental groups. The Ih current was significantly decreased in ethanol compared with saline animals, measured both in terms of peak current (Fig. 7, A–D; P < 0.01) and Ih charge transfer (Fig. 7G, measured for step to –140 mV; P < 0.05), and using either standard intracellular and extracellular solutions (Fig. 7, A, C, and E; saline, n = 11; ethanol, n = 9) or solutions that provide better isolation of Ih (Fig. 7, B, D, and F; saline, n = 6; ethanol, n = 8). Also, the Ih current in standard solutions in saline neurons was similar to that in neurons from age-matched naïve rats (measured for step to –140 mV, naïve, 625 ± 74 pA, n = 7; compared with 670 ± 59 pA from saline and 476 ± 61 pA from ethanol). In addition, Ih was determined in neurons from the additional 18 saline and 16 ethanol animals using a current step from –80 to –140 mV. Although the Ih measured in this manner was significantly reduced in ethanol relative to saline animals (saline, 382 ± 20 pA; ethanol, 321 ± 18 pA; P < 0.05), Ih is partially active at –80 mV (Cathala and Paupardin-Tritsch 1997
; Liu et al. 2003
; Neuhoff et al. 2002
), and thus these experiments were not used for kinetic analyses. Finally, instantaneous currents were not altered in ethanol relative to saline animals, measured either as peak currents or charge transfer during a hyperpolarizing step from –60 mV (Fig. 7, E, F, and H), perhaps suggesting no changes in IRK and/or potassium leak channels (Watts et al. 1996
; Wilson 2005
).
Although total Ih currents were reduced in ethanol animals, several kinetic parameters of Ih were not altered. First, the voltage dependence of activation of Ih, fit with a Boltzmann equation (see METHODS), was not different between saline and ethanol animals (Fig. 8, A and B), with no differences in V1/2 (Fig. 8C) or slope factor (Fig. 8D). Thus ethanol reduced the total Ih current without altering the voltage dependence of activation of Ih (see also Okamoto et al. 2006
). Repeated ethanol exposure also did not alter the reversal potential of Ih (Fig. 8, E and F; n = 6 for saline and n = 9 for ethanol), which was estimated, as described (Cathala and Paupardin-Tritsch 1997
; Mayer and Westbrook 1983
), by comparing the instantaneous currents evoked from a holding voltage of –60 mV, where Ih is not active, and the instantaneous currents evoked from a holding voltage of –80 mV, where Ih is activated (Cathala and Paupardin-Tritsch 1997
; Liu et al. 2003
; Neuhoff et al. 2002
). In particular, one can estimate the reversal potential of Ih from the point where the two I–V curves intersect; because currents from a –60-mV holding potential contain IRK/leak but little Ih, whereas currents from a –80-mV holding potential contain both IRK/leak and Ih due to activation of Ih at –80 mV, the intersection of the I–V curves from –60 and –80 mV essentially represents the voltage where the Ih current is zero after the IRK/leak current is subtracted out.
Ih has been reported to alter burst firing in other types of neurons (Wilson 2005
). Thus we examined the effects of the Ih antagonist ZD7288 (30 µM) on firing or NMDA modulation of bursting after 10-min exposure to ZD7288, after which time >90% of Ih should be inhibited (Fig. 9C; Gasparini and DiFrancesco 1997
; Gu et al. 2007; Okamoto et al. 2006
). In saline animals, ZD7288 did not facilitate a transition to bursting when NMDA was applied subsequent to ZD7288, indicated by no significant increase in the CV-ISI (Fig. 9A; baseline: 31.1 ± 6.6%; after ZD7288 addition: 25.1 ± 5.5%; after ZD7288 and NMDA: 32.3 ± 10.8%; n.s.; n = 5). Thus the reduced Ih in ethanol animals cannot account for the facilitated transition to bursting by NMDA in ethanol animals. In addition, ZD7288 did not alter pacemaker firing, measured by determining the instantaneous firing frequency, in either saline animals (Fig. 9B; before ZD7288: 0.75 ± 0.06 Hz; after ZD7288: 0.72 ± 0.14 Hz; n = 8; n.s.) or ethanol animals (before ZD7288: 0.86 ± 0.05 Hz; after ZD7288: 0.98 ± 0.10 Hz; n = 8; n.s.), in agreement with no differences in pacemaker firing (Fig. 1).
|
Withdrawal from repeated ethanol treatment enhanced the locomotor response to cocaine
Finally, we wanted to establish whether the repeated ethanol treatment used here could be sufficient to produce physiological changes in ethanol animals. In this regard, drug-induced locomotor sensitization has been associated with some aspects of drug reward, dependence, and relapse (Robinson and Berridge 2001
), and repeated ethanol exposure has been shown to produce a cross-sensitization to the locomotor-activating effects of cocaine (Itzhak and Martin 1999
). Thus we examined whether the ethanol exposure protocol used here enhanced the motor response to cocaine, examined after 7 days of withdrawal.
Animals withdrawn after repeated ethanol or saline were habituated to the open-field locomotor chamber for 3 days (see METHODS) before administration of cocaine on the 7th day of withdrawal. Ethanol animals exhibited a significantly greater locomotor response to an acute injection of cocaine (15 mg/kg, ip) relative to saline animals or naïve animals (Fig. 10; n = 16, 13, and 11 for saline, ethanol, and naïve animals, respectively; P < 0.05). No differences were observed between groups in the locomotor activity during a 60-min period of habituation or in response to an acute injection of saline (Fig. 10). Further, saline animals exhibited a locomotor response to cocaine very similar to that in naïve animals, suggesting that any stress related to multiple injections or handling procedures did not enhance the locomotor response to cocaine. Taken together, these results indicate that the repeated ethanol exposure and withdrawal paradigm used here altered channel function and also produced a cross-sensitization to the acute locomotor-activating effects of cocaine.
|
| DISCUSSION |
|---|
|
|
|---|
Pacemaker activity was not altered in VTA DA neurons from ethanol animals
Pacemaker firing is under strong control of a number of currents including Ca2+ channels and A-type potassium channels (see Marinelli et al. 2006
; Puopolo et al. 2007
) and the Na+/K+-ATPase (Johnson et al. 1992
). However, studies of SK and Ih regulation of pacemaker activity in midbrain DA neurons have yielded mixed results. When pacemaker frequencies are more similar to those observed here, SK inhibition with apamin did not alter pacemaker firing (Brodie et al. 1999
; Johnson and Wu 2004
; Seutin et al. 1993
; Waroux et al. 2005
), although, with more rapid firing, apamin can increase firing rates (Ping and Shepard 1996
; Shepard and Bunney 1988
; Waroux et al. 2005
; Wolfart et al. 2001
). Further, Ih-dependent depolarization enhanced firing in some studies (Okamoto et al. 2006
; Seutin et al. 2001
), but other results suggest a less-consistent relationship between Ih and pacemaker activity (Liu et al. 2003
; Neuhoff et al. 2002
), and Ih can reduce firing by decreasing the input resistance (Fan et al. 2005
). Thus the currents responsible for the strong oscillatory drive in VTA neurons could override the contribution of Ih and SK during pacemaker firing, perhaps also explaining the lack of difference in pacemaker activity between saline and ethanol animals despite functional changes in several currents.
Previous in vivo and in vitro observations during withdrawal after chronic ethanol exposure found reduced firing rates in adult midbrain DA neurons after 1–3 days (Bailey et al. 1998
; Diana et al. 1993
, 1996
; Shen 2003
; but see Brodie 1999) or 6 days of withdrawal (Bailey et al. 2001
), with no changes in firing rates of active DA neurons
3 wk after repeated ethanol exposure (Bailey et al. 2001
; Shen et al. 2007
), although the number of neurons firing was reduced (Shen et al. 2007
). Here, we observed no changes in pacemaker activity 7 days after repeated ethanol treatment. Taken together, these observations suggest that withdrawal from chronic ethanol treatment decreases pacemaker activity for several days (but see Brodie 1999), but that this effect is absent after
7 days of withdrawal. However, it should be noted that our studies used juvenile animals, and thus our results might not fully reflect those collected from adult animals, although differences in species (mice vs. rat) and route of ethanol administration (liquid diet vs. intermittent ip injections) could also explain apparent discrepancies between our studies and those of Bailey and colleagues (2001)
.
SK function was reduced and burst firing enhanced in ethanol animals
Several lines of evidence suggest that repeated ethanol exposure decreased SK function in VTA neurons. The time-to-peak of the AHP (TTP-AHP) was significantly reduced in ethanol animals, with no apparent changes in several other parameters of the action potential waveform. The TTP-AHP was also greatly reduced by the SK antagonist apamin (Ping and Shepard 1996
; Shepard and Bunney 1988
), suggesting a primary contribution of SK to the TTP-AHP. Further, SK-mediated tail currents (Paul et al. 2003
) were significantly reduced in ethanol animals, with a significant correlation between the TTP-AHP and the tail current magnitude in ethanol animals. Taken together, these results strongly suggest that SK function was reduced in ethanol animals. Further experiments will be required to determine whether differences in SK function in ethanol animals reflect changes in number or regulation of SK or of the calcium channels required for SK activation (Bond et al. 2005
).
In addition to regulation of the AHP, SK inhibition facilitates NMDA-induced burst firing (Johnson and Seutin 1997
; Seutin et al. 1993
; Waroux et al. 2005
). Here, NMDA (20 µM) increased pacemaker frequency in saline animals, but induced burst firing in ethanol animals, which could be explained by decreased SK activity facilitating NMDA induction of bursting in ethanol animals. Differential NMDA induction of burst firing was not due to differences in NMDA receptor activation because the current response to NMDA, measured under voltage clamp, was similar in saline and ethanol animals. In addition, control neurons were capable of bursting because SK inhibition in combination with NMDA elicited bursting in saline animals, as is widely observed in control neurons (Johnson and Seutin 1997
; Seutin et al. 1993
). Thus neurons from saline animals could exhibit bursting, especially when SK was inhibited during NMDA application, but NMDA induction of bursting was greatly facilitated in ethanol animals, perhaps due to decreased SK function. In addition, NMDA induction of bursting in ethanol neurons was significantly negatively correlated with the SK tail current magnitude, strongly supporting the possibility that reduced SK function might be responsible for the enhanced NMDA-induced transition to bursting in ethanol animals.
An increased ability of the glutamate system in the VTA to induce bursting activity after repeated drug exposure may contribute to drug seeking and enhance the vulnerability to relapse during ethanol withdrawal. Glutamatergic afferents into the VTA can regulate the shift from pacemaker to burst firing (Chergui et al. 1994
; Floresco et al. 2001
; Grillner and Mercuri 2002
), which increases DA release in midbrain target regions and may contribute to some aspects of drug-induced reinstatement (Grace 2000
; Schmidt and Pierce 2006
; Tupala and Tiihonen 2004
).
Ih function was reduced and recovery from hyperpolarization impaired in ethanol animals
Repeated ethanol exposure and 7 days' withdrawal also significantly reduced hyperpolarization-evoked Ih currents, which could significantly influence firing properties. For example, Ih drives certain aspects of oscillatory activity in striatal cholinergic interneurons (Wilson 2005
), and thus decreased Ih might retard bursting by reducing the depolarizing drive (Wilson 2005
). However, NMDA-induced bursting was enhanced in ethanol animals and Ih inhibition in saline animals did not enhance NMDA-induced bursting, suggesting that differences in Ih did not account for the increased NMDA-induced transition to bursting in ethanol animals. In addition, Ih can facilitate recovery from hyperpolarization (Neuhoff et al. 2002
; Satoh and Yamada 2000
; Wilson 2005
). Firing is delayed after strong hyperpolarization (see Fig. 9D) due to an A-type potassium current (Koyama and Appel 2006
) and Ih activation at hyperpolarized potentials can provide depolarization to drive neurons to voltages permissive for