Abstract
Previous reports of membrane hyperpolarizations, associated with acute application of cocaine, have been recorded from brain slice preparations containing aminergic nuclei and have always been attributed to cocaine’s ability to elevate levels of local biogenic amines followed by activation of their receptors. The majority of these studies were conducted with brain slices obtained from rats that had not received prior chronic in vivo treatment with cocaine. We observed that cocaine alone, at 3 μM, could induce a membrane hyperpolarization (COC-HYP) in 100% of rat dorsolateral septal nucleus (DLSN) neurons from brain slices of rats treated chronically with cocaine for either 14 or 28 days in vivo. The DLSN is a nucleus absent of biogenic amine cell bodies, but does contain biogenic amine terminals with GABAergic cell bodies and terminals. Cocaine applied to brain slices from rats not previously administered cocaine or administered cocaine for up to seven days in vivo yielded a maximum incidence of COC-HYPs at only 50%. COC-HYPs recorded from DLSN neurons were not blocked by previous treatment with amine receptor antagonists or by a TTX and zero calcium medium. Based on these results, the ability of DLSN neurons to respond to a cocaine challenge with a COC-HYP did not involve inhibition of amine reuptake/uptake or action potential release of neuroactive substances. Rather, the COC-HYP, with an apparent reversal potential of -80 mV, was reduced by the GABA receptor antagonists-bicuculline and CGP-55845A. Lowering extracellular Na+ or Cl− , lowering of temperature, or previous superfusion with the GABA uptake blocker NO-711 could block the COC-HYP. In summary, our data suggest that COC-HYPs, after application of a cocaine challenge to brain slices from rats treated chronically (14 - 28 days, but not acutely, 7 days) with cocaine are due to cocaine-induced changes in GABA release and/or transporter function. The latter changes in transporter function may involve the reversal of the GABA transporter with release of GABA and subsequent activation of postsynaptic GABAA and GABAB receptors.
The majority of studies to determine the cellular actions of cocainein vitro has used CNS synapses of drug naive rats. These synapses have included both aminergic cell body regions and their terminal fields. There has been a remarkable consistency in terms of the responses obtained after administration of cocaine to aminergic cell body regions. That is, when cocaine alone is applied and recordings are made from cell body areas, a significant inhibitory effect, e.g., a membrane hyperpolarization is recorded from biogenic amine-containing neurons. This membrane hyperpolarization associated with acute cocaine application results from a potentiation of the typical actions of the transmitter released endogenously (Suprenant and Williams, 1987; Pan and Williams, 1989;Lacey et al., 1990; Bonci and Williams, 1996). In general, the ability of cocaine to potentiate biogenic amine responses has been attributed to its well-known action to bind to the respective amine transporters (Ross and Renyi, 1969; Ritz et al., 1987). Subsequently, inhibition of the transporter would inhibit uptake of biogenic amines applied exogenously to the slice or block reuptake of endogenous biogenic amines released within an aminergic synapse.
However, in brain areas lacking biogenic amine-containing somata,i.e., terminal field areas, responses to cocaine administration have varied from alterations of membrane potential, modification of synaptic activity or no effect (Jahromi et al., 1993; Simms and Gallagher, 1996). Although the effects of cocaine when administered alone have been equivocal, cocaine would potentiate the actions of exogenously applied biogenic amines or prolong evoked aminergic synaptic potentials (Uchimura and North, 1990;Bobker and Williams, 1991; Jahromi et al., 1993; Simms and Gallagher, 1996).
The DLSN is a biogenic amine terminal area and one that contains a very high density of cell bodies and terminals for inhibitory (GABA) and excitatory (glutamate) amino acids (Jakab and Leranth, 1994). Because its synaptic pathways contain a diversity of receptors (Gallagheret al., 1995), and its implication in emotion and anxiety (Gray, 1982), we chose the DLSN to investigate the multiple actions of cocaine (Woolverton and Johnson, 1992).
Furthermore, because cocaine dependence is a process that is essentially chronic in nature, studies of the effects of repeated cocaine administration in animals are required to suggest a possible interpretation of the clinical effects of chronic cocaine administration in humans. For instance, although cocaine initially produces euphoria and mood elevation, continued abuse can lead to psychiatric problems such as anxiety, depression and psychosis (Fischman, 1987). Accordingly, it has become especially important to study the cellular mechanism(s) by which chronic cocaine in vivo affects neural function both prior to and after an acute “cocaine challenge” or reexposure to cocaine in vitro.
We had initially examined the acute in vitro actions of cocaine (Simms and Gallagher, 1993; Simms et al., 1994) applied to brain slice preparations from rats pretreated with saline or not pretreated with cocaine in vivo. Data from these experiments were similar to those reported by Jahromi et al.(1993). These investigators recorded from multiple biogenic amine terminal areas and observed a variety of effects after an acute application of cocaine. However, like us, none of their effects were significant statistically. No consistent changes were reported in any of the following parameters: resting membrane potential, input resistance or spontaneous and evoked inhibitory and excitatory synaptic activity.
However, after a 14-day, but not 7-day, in vivo exposure to cocaine (15 mg/kg, twice daily) we observed that all of the above properties were altered significantly and consistently (Shoji et al., 1997: figs. 1 and 2). Furthermore, we also demonstrated (Simms and Gallagher, 1997) that after the same 14-day, but not 7-day, chronic treatment with cocaine, the distribution of cell types within the DLSN (Gallagher et al., 1995) differed with chronic cocaine exposure. All of the changes we have reported before occurred in the absence of any additional acute drug treatment, and thus represented the chronic effects of cocaine to alter the intrinsic electrical and synaptic properties of DLSN neurons (Simms and Gallagher, 1996; 1997; Shoji et al., 1997).
Our study was undertaken to characterize the cellular response to a “challenge” dose of cocaine in vitro 16 hr after chronic periods of intermittent cocaine exposure in vivo.
Materials and Methods
Cocaine treatment regimen.
Male Sprague-Dawley rats (Harlan, 75–250 g) were housed three to four per cage with free access to food and water. Each rat received injections with either saline (0.9%) or cocaine HCl [Sigma Chemical Co., St. Louis, MO or National Institute on Drug Abuse (NIDA), Rockville, MD] [15 mg/kg, i.p., twice daily (9:00 a.m. and 4:00 p.m.)] for 7 (COC-7), 14 (COC-14) or 28 (COC-28) consecutive days. It is well established that behavioral sensitization can develop to locomotor activity and stereotyped behavior, specifically, rearing, fast repetitive head and/or foreleg movement, induced by cocaine when it is administered intermittently (Post, 1977). We used the development of behavioral sensitization to cocaine as an indicator of the effectiveness of our cocaine injections. Behavioral sensitization was measured as enhanced exploratory locomotor activity and induced stereotypic behavior in all animals 15 min after twice daily treatment with cocaine for periods of either 7, 14 or 28 days (see Simms and Gallagher, 1996).
Because we did not find any appreciable differences in the electrophysiological responses obtained from brain slices derived from saline-injected rats and rats not exposed to cocaine or saline, data from these groups were pooled. All chronic results only represent data collected from rats treated chronically with cocaine and killed 1 hr before their next (and final) scheduled cocaine injection,i.e., at 0800 on days 15 or 29. As a result, this paradigm yielded brain slices having a period (16 hr) of early withdrawal from cocaine.
Preparation of brain slice.
Rat forebrain coronal slices (500-μm thick) containing the DLSN were prepared using standard techniques (Stevens et al., 1984). Briefly, the rat was decapitated and the brain rapidly removed and immersed in a modified cold ACSF solution. The ACSF solution was maintained at 6°C and bubbled continuously with 95% O2 and 5% CO2 to maintain proper pH (7.3–7.4). The composition of the ACSF solution was as follows: NaCl, 117 mM; KCl, 4.7 mM; NaH2PO4, 1.2 mM; MgCl2, 1.2 mM; CaCl2, 2.5 mM; NaHCO3, 25 mM and glucose, 11.5 mM. In the cold (6°C) solution, the brain was quickly blocked to transverse sections 2-mm thick with the caudal edge at the level of the optic chiasm. Diagonal cuts were then made lateral to the anterior commissure to remove most of the cortex and striatum. The resulting block of tissue was glued (Duro Super Glue, Loctite Corp., Rocky Hill, CT) to a chuck and placed in the bath of a Vibroslice (752 M, Campden Instruments, Ltd., London, England, UK) in similarly treated cold ACSF solution. Serial slices were made rostral to caudal until a section containing medial and lateral septal nuclei was produced. The slice was then placed in a superfusion chamber maintained at 32 ± 2°C and superfused at a flow rate of 1 to 1.5 ml/min with ACSF solution bubbled continuously with 95% O2 and 5% CO2. We routinely use the following two criteria as indices of viable slices. First, stable MP of at least -50 mV must be maintained for at least 10 min. Second, the neurons must respond to direct positive current stimulation with a rapid and overshooting sodium spike.
Recording from brain slice.
Sharp intracellular recordings were obtained using Frederick-Haer standard wall 1.0-mm fiber filled glass microelectrodes pulled to final tip resistances of 70 to 100 MΩ and filled with 2 M potassium acetate. Ri was routinely measured by passing hyperpolarizing current pulses of known intensities through the recording electrode using a bridge-type circuit. Voltage signals and applied current were recorded with an Axoclamp 2A amplifier (Axon Instruments, Inc., Foster City, CA). The output of the amplifier was D.C. coupled to a storage oscilloscope (Model 5111, Tektronix, Portland, OR) and a dual channel Gould (Cleveland, OH) (Model 220) chart recorder. A Model 4208 Panasonic VCR/Recorder (A.R. Vetter Co., Rebersburg, PA) was used to capture all tracings for storage. The stored signal can be played back and analyzed using a pClamp Version 6.0 Software with a DigiData 1200 interface to a Gateway 2000 4DX2–66V computer. Paper copies of the waveforms were generated with a Hewlett Packard Laserjet 4 printer.
Electrical stimulation of brain slice.
In some experiments the brain slice was stimulated electrically to yield low frequency induced orthodromic responses via outputs from a Grass (Quincy, MA) S-88 stimulator with isolation units. Focal stimulation was applied through a low resistance concentric bipolar electrode (Frederick-Haer) inserted into the dorsolateral aspect of the DLSN nucleus of the septum. Stimulus parameters were adjusted to yield consistent responses, e.g., 100-μsec duration and 1 to 10 V intensity, at a frequency of 0.17 Hz. We have demonstrated previously that neurons in the DLSN display a series of synaptic potentials in the slice preparation after electrical stimulation of fimbrial afferents and/or local interneurons. These include: 1) an excitatory amino acid- (probably glutamate) mediated EPSP that results from activation of both non-NMDA and NMDA (N-methyl-d-aspartic acid) receptors (Gallagher and Hasuo, 1989); 2) a f-IPSP-mediated by GABA acting at GABAA receptors (Stevens et al., 1984) and 3) a s-IPSP-mediated, at least in part, by GABA acting at GABAB receptors (Hasuo and Gallagher, 1988).
Drug application.
Pharmacological sensitivity and drug testing were carried out by superfusion of known concentrations of substances. Substances were dissolved in the ACSF and entered the recording chamber through a gravity feed inlet of the superfusion system. All drug stock solutions were made up in distilled H2O. The drugs used in the these experiments were as follows: (-)-bicuculline methiodide and TTX from Sigma; (±) sulpiride, idazoxan HCl, CPT; atropine sulfate; p-MPPF dihydrochloride; NO-711-HCl, and d-AP5 from Research Biochemicals Incorporated (RBI, Natick, MA); CGP-55845A from Ciba-Geigy (Basel, Switzerland); CNQX from Tocris-Cookson (Essex, UK); (-)-cocaine HCl from Sigma and the National Institute on Drug Abuse (NIDA).
Data analysis.
Cocaine was applied by superfusion to the slice and a change in membrane potential was recorded intracellularly. We chose to superfuse routinely and primarily with 3 μM cocaine because our preliminary results showed this concentration of cocaine produced consistent hyperpolarizations and was incapable of local anesthetic effects such as increasing the threshold or width of a sodium spike induced by positive current injection. Furthermore,in vivo microdialysis studies have shown that dialysate cocaine concentrations obtained from brain tissue of chronic cocaine treated rats approaches 3 μM after a cocaine challenge injection (Pettit et al., 1990). Moreover, this concentration of cocaine approximates closely the brain concentrations of cocaine found in users of the drug (Javaid et al., 1978; Van Dyke et al., 1978). All cellular data are expressed as mean ± S.E.M. Statistical analyses used in these studies were the unpaired one-tailed Student’s t test (SigmaPlot, Windows, Ver. 1.0). Statistical significance was determined at the level of P ≤ .05. Graphs and histograms were generated using SigmaPlot (Windows, Ver.1.0) software (Jandel Scientific Corp., San Rafael, CA). In comparing groups of small sample size (fig. 1, bottom) a Fisher exact test (Sigmastat, Ver. 1.0) was used with statistical significance determined at the level of P ≤ .05.
Results
Expression of the COC-HYP is Dependent upon Duration of in Vivo Treatment
Cocaine (1–10 μM) applied by superfusion to brain slices (n = 63) in vitro obtained from rats treated chronically with cocaine in vivo [15 mg/kg, i.p., twice daily (BID) × 14 or 28 days] resulted in a slow onset hyperpolarization (COC-HYP) of the MP of DLSN neurons (at 3 μ M, ΔMP = -4.9 ± 0.4 mV, n = 10, fig.1, top). We conclude that COC-HYPs are an all-or-none event with a change of –2 mV from MP being the minimum recorded threshold of occurrence. A change of –2 mV is also biologically relevant, because at –60 mV a DLSN neuron could exhibit spontaneous firing activity, whereas, at –62 mV the same DLSN neuron would exhibit no spontaneous activity (Gallagher et al., 1995). All COC-HYPs were recorded in the presence of biogenic amine antagonists in the superfusion medium.
While monitoring the incidence and concentration effect of an acute challenge with cocaine, we noted that COC-HYPs were observed in all rats having had a 14-day treatment regimen (fig. 1, bottom). Furthermore, the concentration of cocaine needed to record a consistent COC-HYP became less as the in vivo exposure was increased from 7 to 28 days of chronic cocaine treatment, i.e., the concentration-effect was shifted leftward (fig. 1, bottom). We conclude that a 14-day chronic cocaine regimen resulted in a “cellular sensitization” to an acute cocaine challenge.
We define the state of “cellular sensitization” to an acute cocaine challenge as an increased incidence of observing a COC-HYP after a challenge dose of cocaine (fig. 1, bottom). We observed that following a 14- or 28-day chronic treatment regimen, a cocaine challenge of ≥3 μM would result in a hyperpolarization from every cell. Thus, “cellular sensitization” represents an increase in the proportion of DLSN neurons exhibiting a hyperpolarization to a cocaine challenge. This conclusion is based on the application of the Fisher exact probability test to the data depicted in figure 1, bottom. This analysis demonstrates that there is no difference between the incidence of observing a hyperpolarization after a cocaine challenge with the three different concentrations of cocaine to neurons treated with saline (control) or cocaine for 7 days. However, after a 14- or 28-day treatment with cocaine there was a significant difference in the incidence of observing a hyperpolarization when recording DLSN neurons from rats receiving a saline (control) or treatment with cocaine for only 7 days.
Non-Amine Mechanism of COC-HYP
Cocaine is well known for its ability to inhibit the transporters responsible for the uptake of biogenic amines within the nervous system (Goeders and Smith, 1986; Ritz et al., 1987). The DLSN contains postsynaptic receptors for each of the biogenic amines: an alpha-2-adrenotropic receptor, a D-2-dopamine receptor and a 5HT-1A-serotonin receptor (for review see Gallagher et al., 1995). Cocaine and other biogenic amine uptake blockers potentiate the hyperpolarizations recorded from DLSN neurons and induced by exogenous application of these biogenic amines (Joëlset al., 1987; Simms and Gallagher, 1996).
In an effort to minimize a potential endogenous biogenic amine-mediated effect, we have included selective dopamine (DA) and norepinephrine (NE) specific receptor antagonists (sulpiride, 1 μM, and idazoxan, 10 μM, respectively, Gallagher et al., 1995) with all cocaine challenges. These biogenic amine antagonists could also block the activation of nerve terminal catecholamine receptors present on glutamate, GABA or other terminals within the slice. Moreover, we demonstrate that the COC-HYP (fig. 2A) persists when a brain slice is superfused with TTX (1 μM, fig. 2B), and, in addition, persists in the presence of zero calcium medium with TTX (fig. 2C). This combination of a voltage-dependent sodium channel antagonist and lack of extracellular calcium will effectively block synaptic transmission within the slice and eliminate action potential dependent, but not spontaneous release of endogenous biogenic amines and other endogenous neuroactive substances. The presence of TTX and zero calcium medium would not alter non-vesicular release or release mediated by reversal of the uptake carrier for GABA or other transmitters, e.g., glutamate, on neurons or glia.
We have also conducted experiments with six rats, not previously administered cocaine in vivo, but pretreated with reserpine (5 mg/kg, 48 hr before death; Calabresi et al., 1988). When cocaine (3 μM) was applied to brain slices from these reserpinized rats a 50% incidence of membrane hyperpolarizations was observed. This 50% incidence is identical to that obtained with nonreserpinized rats (fig. 1, bottom). These results add additional support to the concept that a COC-HYP recorded from the DLSN is not mediated by biogenic amines.
GABA Dependence of the COC-HYP
Antagonists were applied along with idazoxan (10 μM) and sulpiride (1 μM) to brain slices from rats treated chronically with cocaine. In conjunction with these two antagonists, the coapplication of the adenosine-1 receptor antagonist CPT (10 μM; n= 3), muscarinic antagonist atropine (1 μM; n = 3) or the 5-HT1A antagonist p-MPPF (1 μM; n = 3) did not alter the COC-HYP.
However, coapplication of bicuculline, a GABAAreceptor antagonist, which initially resulted in a depolarization of the MP (Shoji et al., 1994, 1997), subsequently depressed a typical COC-HYP. Similarly, application of a GABAB receptor antagonist, CGP-55845A, also initially resulted in a depolarization of the MP, but subsequently depressed a COC-HYP. Coapplication of both bicuculline and CGP-55845A suppressed but did not block completely the COC-HYP (fig.3). A residual hyperpolarization (1.3 ± 0.3 mV, n = 5) persisted, even in the presence of both GABAA and GABAB receptor antagonists.
Ionic mechanism of COC-HYP.
Figure 1, top, 2A and 4A demonstrate the time required for onset and typical decrease of cellular input resistance associated with a COC-HYP. Note that a relatively long (5–10 min) continuous superfusion with cocaine is required to observe a COC-HYP. Shorter superfusion times were without effect. Figure 5A plots an estimation of the reversal potential of the COC-HYP. The apparent reversal potential for the COC-HYP obtained in this manner is -80 ± 1.4 mV (n = 5), a value not commonly attributed to a single ion. A combination of ions that could explain such a reversal potential is potassium and chloride, with EK ∼-90 mV and ECl ∼-70 mV in this preparation and extracellular media.
Although we attempted to obtain a direct reversal of the COC-HYP, repeated applications of cocaine to the same brain slice often resulted in serially smaller COC-HYPs, even when spaced at intervals of 45 min. This apparent tachyphylaxis to repeated applications of cocaine and inability consistently to reproduce comparable COC-HYPs in the same brain slice is also evident with a single prolonged exposure to cocaine (fig. 1, top and 4A). As previously mentioned, the COC-HYP does not persist during the continued presence of cocaine, but rather “fades” back to the original MP. This fade is similar to membrane potential changes resulting from activation of GABAA receptors after prolonged application of GABAA agonists (Gallagher et al., 1983). Also apparent is the inability of cocaine to maintain a reduced membrane input resistance and block of spontaneous action potential activity (figs. 1 and 4A). In contrast, when biogenic amines are superfused continuously or applied serially by drop application, the hyperpolarization they each induce does not exhibit a fade (Gallagheret al., 1995).
Figure 5B also depicts the results obtained when extracellular potassium was shifted from its normal value of 4.7 mM to either 10 or 0.5 mM. The reversal potential of the COC-HYP shifted in a linear Nernstian fashion with the altered potassium concentration. The slope of this linear shift was 29 mV/10 mM suggesting that potassium, although contributing significantly to the COC-HYP, was not the only ion responsible.
Blockade of COC-HYPs.
When extracellular chloride was substituted with isethionate (117 mM) the COC-HYP was not observed (fig. 4B; n = 4); a result suggesting that chloride ion or a chloride-dependent process contributed significantly to the COC-HYP. To investigate a possible role for chloride, we determined the reversal potential of fast GABAA-receptor mediated f-IPSPs. F-IPSPs could be isolated by blocking excitatory transmission with a combination of CNQX (20 μ M) and D-AP5 (50 μ M); later, slower GABAB-receptor mediated IPSPs were blocked with CGP-55845A (1 μM). Under these conditions, the reversal potential of f-IPSPs in DLSN neurons from control brain slices was -70.1 ± 0.5 mV, n = 14, although in brain slices from rats administered cocaine chronically for 14 or 28 days, the reversal potential of the f-IPSP was shifted in a statistically significant manner (P ≤.05) to -67.4 ± 0.3 mV, n= 6. The reversal potential of the f-IPSP is equivalent to the reversal or equilibrium potential for chloride ion (ECl-), because the GABAA-ionotropic receptor is coupled to a chloride (Cl− ) channel. This shift in the reversal potential for Cl− suggests that the Cl− -pump, which maintains Cl− concentrations across the cell membrane, has been inhibited by the chronic cocaine treatment. An inhibition of this outwardly directed pump would result in an elevation of both intracellular chloride and sodium.
Another membrane process possibly affected, as a result of chronic cocaine, is an electrogenic Na+/Cl− exchanger that is essential for the GABA-transporter. Reduction of extracellular sodium by replacement with glucosamine (117 mM) prevented the expression of the COC-HYP (fig. 4C). This latter result, which is similar to that observed with chloride replacement, suggests that, like chloride, sodium is required—not necessarily as a charge carrier—but primarily as a metabolic cofactor along with Cl− , for the COC-HYP. Thus, at least, potassium, chloride and sodium are necessary to observe the COC-HYP.
Temperature is an additional factor that is essential for the generation of COC-HYPs in brains slices obtained from rats treated chronically with cocaine. Lowering the bath temperature from 34°C to room temperature (22°C) prevented the appearance of COC-HYPs (fig.4D). This result supports the hypothesis that the COC-HYP involves one or more metabolically dependent phenomena.
GABA transport inhibitor, NO-711, blocks COC-HYP.
A process that is metabolically dependent, sensitive to a lowering of Na+ or Cl− , highly active at GABA synapses, and persists in the presence of TTX and zero calcium is GABA transport. Because cocaine inhibits biogenic amine transporters, we considered the possibility that as a genetically related protein (Uhl and Johnson, 1994) the GABA transporter may also be affected after chronic cocaine exposure.
Application of NO-711 (3 μ M, Suzdak et al., 1992) to brain slices from rats treated chronically with cocaine for 14 or 28 days initially produced a membrane hyperpolarization (2.4 ± 0.4 mV, n = 5) that faded back to control values in the continued presence of the drug (fig. 6). Subsequently, application of cocaine (3 μM) in the continued presence of NO-711 resulted in an almost complete block of the COC-HYP (0.3 ± 0.05 mV; n = 5). These data suggest that NO-711 may mimic a novel action of chronic cocaine, which is to inhibit and reverse the GABA transporter.
Discussion
Previous intracellular electrophysiological results of cocaine challenge in vitro after chronic cocaine in vivo.
In only one other laboratory have intracellular recording techniques been used to examine the actions of cocainein vitro with a brain slice preparation following chronic cocaine in vivo (Harris and Williams, 1992). In that study, the effects of repeated cocaine (20 mg/kg/daily i.p., × 14 days) and a 1-wk withdrawal period were examined on norepinephrine neurons of rat brain slices containing the locus ceruleus. No baseline changes of synaptic transmission or electrical properties were noted before reexposure to cocaine. However, neurons from animals administered chronic cocaine exhibited a significant increase in sensitivity to the effects of cocaine which prolonged the time course of the NE-mediated inhibitory postsynaptic potential.
In a more recent study, Bonci and Williams (1996) concluded that presynaptic adenosine (A1) receptors are involved in the reversal of the typical augmentation of GABAB mediated IPSPs observed in brain slices from naı̈ve rats to an inhibition of this IPSP recorded from dopamine neurons in ventral tegmental area brain slices of chronic cocaine-treated rats.
We tested this possibility at the DLSN, which contains A1-receptors (Hasuo et al., 1992). The A1 antagonist, CPT, did not alter the expression of COC-HYPs and did not alter the typical IPSPs recorded from DLSN neurons. Moreover, our preliminary results with application of DA to slices from naive rats or rats treated chronically with cocaine demonstrate that DA only depresses rather than augments an evoked slow-IPSP (S. Shoji and J. P. Gallagher, unpublished observations). This depressant action of DA also persists in the presence of the A1 antagonist CPT.
Williams’ data were collected at a longer withdrawal period (1-wk) than we used (16 hr). Our data may represent an early withdrawal effect, although Williams’ data may represent the effect of a longer withdrawal period. In this regard, a report by Brandon and White (1997)demonstrates that membrane property changes, recorded from neurons in brain slices of the nucleus accumbens and observed at 3 days after chronic cocaine treatment, are lost when animals receiving the same chronic cocaine regimen are tested at 14 days following withdrawal of cocaine. Thus, it appears that different mechanisms may be responsible for the phenomena associated with early withdrawal from cocaine as compared to a period of late withdrawal.
Superfusion of cocaine results in a GABA-dependent hyperpolarization within the DLSN.
We have demonstrated in vitro with brain slices from rats treated chronically with cocainein vivo, for 14 or 28 days that superfusion with cocaine (3 μM) results in a GABA-receptor dependent, biogenic amine independent, MP hyperpolarization of DLSN neurons. Support for the GABA dependence of this hyperpolarization stems from experiments that effectively reduced the COC-HYP; each experiment was conducted in the presence of biogenic amine antagonists. Each experiment points to a number of different mechanisms that may all contribute to the GABA-dependent COC-HYPs.
A combination of the GABAA-receptor antagonist, bicuculline (10 μM), with the GABAB-receptor antagonist, CGP-55845A (1 μM) depresses the COC-HYP (fig. 3). Neither antagonist alone is completely effective, although both antagonists, together are still not completely effective. The remaining residual hyperpolarization may be due to an electrogenic component of the GABA response that is resistant to receptor antagonism (Haugh-Scheidtet al., 1995).
Two additional experiments, involving transporter mechanisms, could also contribute to the COC-HYP; these include the ionic dependence of the transporters for chloride and/or GABA. As noted earlier, the COC-HYP is not an immediate response after application of cocaine to brain slices of rats treated chronically with cocaine. Unlike a typical agonist-receptor-effector sequence, which generates a response within msec or seconds, the COC-HYP exhibits a delayed onset of several minutes (fig. 1, top). Lowering of extracellular sodium, or, lowering of extracellular chloride each prevent a COC-HYP (fig. 4). This COC-HYP is not mediated by the ability of cocaine to inhibit biogenic amine transporters, but instead may result from the redistribution of ions (Na+ and Cl− ) across the neuronal cell membrane. Both these ions are critical for the normal functioning of the transporters for chloride and GABA. In this regard, we have considered the possibility that chronic cocaine may affect the Cl− - or GABA-transporters independently or concomitantly, but in a manner analogous to its typical ability to affect biogenic amine transporters, acutely. Indeed, the GABA-transporters belong to one of two subfamilies of Na+-and Cl−-dependent transporters whose genes encode glycoproteins with 12 putative transmembrane regions; the other subfamily includes the biogenic monoamine transporters. These encoded glycoproteins act as neurotransmitter/ion symporters that use transmembrane Na+ gradients to drive cellular accumulation of their respective transmitters (Uhl and Johnson, 1994). We suggest that cocaine, in addition to biogenic amine transport inhibition, may also alter the function of chloride and GABA transport via alterations of intracellular sodium. Furthermore, following chronic cocaine treatment, these chloride and GABA transport systems become “sensitized” to cocaine’s effect such that application of cocaine to brains treated chronically with cocaine enhances their altered activity. These effects of chronic cocaine upon GABA and chloride transport are opposite to the down-regulation of serotonin transport we observed with chronic cocaine (Simms and Gallagher, 1996).
A general feature of the neuronal and glial transporters for transmitters is their reversibility (Levi and Raiteri, 1993; Bonanno and Raiteri, 1994). The direction of the transport process can be manipulated by altering ionic gradients across a cell membrane, particularly the gradients for Na+ and Cl−. For instance, GABA accumulation (uptake and reuptake) occurs with a Na+ gradient of (out > in) and/or a chloride gradient (out > in). As these ionic gradients are shifted, they produce a net increase of intracellular sodium, e.g., by inhibition of the chloride pump and excessive activation of ionotropic glutamate receptors. The enhanced activation of postsynaptic glutamate receptors could also include activation of metabotropic receptors; these receptors are activated by excessive glutamate release due to disinhibition of the nerve terminal GABAB receptors (Shoji et al., 1997). As a result, the transporters could function in the opposite direction,i.e., to pump (release) GABA from the intracellular compartment back into the synapse (fig. 7). Such a process may include the release of cytosolic (nonvesicular) GABA, which persists in the presence of TTX and zero calcium extracellularly (fig. 2). Support for a “carrier-mediated release” process, i.e., release of GABA by reversed uptake stems from work done in a variety of slice and synaptosomal preparations (for review see Adam-Vizi, 1992; Bernath, 1992). Our previous data (Shoji et al., 1997) demonstrating an increase in spontaneous IPSP frequency and prolongation of evoked IPSPs (Gallagher and Shoji, 1995) also provide evidence for such a process (Haycock et al., 1978; Pin and Bockaert, 1989) at DLSN neurons of brain slices from rats treated chronically with cocaine. Our data (fig. 6) also support the concept that GABA transport is altered by chronic cocaine, since the GABA transport inhibitor, NO-711, blocked COC-HYPs.
We have already demonstrated (Shoji et al., 1997) that chronic cocaine reduces the negative feedback control of both GABA and glutamate release within the DLSN synapse. Schwartz (1987) had originally demonstrated that depolarization due to excessive activation of retinal neurons by glutamate resulted in elevation of intracellular sodium and possibly calcium. This phenomenon has been demonstrated in tissues in addition to the retina with a similar result,i.e., an elevation of intracellular sodium/calcium and the transformation of GABA transporters. The typical inwardly directed neurotransmitter transporters are now reversed and function in an “inside-out” manner, a phenomenon termed “carrier-mediated release” (Levi and Raiteri, 1993: Bonanno and Raiteri, 1994). If a similar process occurs within the DLSN of rats treated chronically with cocaine, then the concentration of carrier-mediated released GABA is now sufficient to activate postsynaptic GABA receptors and induce a COC-HYP; these COC-HYPs are blocked by blocking the reversed transporter with the GABA uptake blocker NO-711.
We propose that chronic cocaine could also, by altering the typical negative feedback inhibitory function of the nerve terminal GABAB receptors on both GABA and glutamate neurons within the DLSN, although not affecting the post-synaptic GABA receptors (Shoji et al., 1997), prevent these transmitters’ transporters from acting in their normal uptake/reuptake manner and catalyze their conversion to inside-out transporters with release of their respective transmitters.
Another possible contributing factor to the phenomenon of “cellular sensitization” could be a shift in the distribution of DLSN cell types that occurs following chronic cocaine administration (Simms and Gallagher, 1997).
A combination of mechanisms contribute to the expression of COC-HYPs within the DLSN.
Our data suggest the following series of events as a complex mechanism (fig. 7), which underlies the COC-HYP recorded from DLSN neurons. An elevation of intracellular chloride results from inhibition of an outwardly directed chloride pump (Cl− -transporter). We provide data that demonstrate a shift in the equilibrium potential of the GABAA-receptor mediated f-ipsp. Elevation of [Cl-I] is associated with movement of Na+ into the cell to neutralize the increased intracellular negativity brought about by Cl− . Furthermore, an additional increase of [Na+i] and possibly [Ca++ i] results from excessive activation of ionotropic glutamate receptors on DLSN neurons. An excessive release of glutamate is due to a decreased efficacy (“down-regulated”) of the GABA-heteroreceptor (fig. 7) on glutamate releasing neurons (Shoji et al., 1997). Our ability to block the COC-HYP by lowering extracellular Na+ or Cl− , and by lowering temperature both support a cellular mechanism whereby chronic cocaine has resulted in stimulation of the GABA-carrier (reversed transporter), such that when a cocaine challenge is applied, GABA is secreted (released) from neurons and/or glia resulting in a postsynaptic hyperpolarization mediated by GABAA and GABABreceptors.
These results do not speak against an action of cocaine to hyperpolarize CNS neurons at non-GABA synapses via their classical ability to enhance synaptic levels of biogenic amines, which subsequently activate their respective receptors. Rather, these results suggest that complex mechanisms associated with cocaine usage may also involve transmitters and pathways in addition to the biogenic amines, particularly, the GABAergic and glutamatergic systems. Our data are especially relevant after the recent suggestion by Wise (1995) that a “hope for pharmacotherapy of addiction (stimulant abuse, including cocaine) lies in the development of drugs that may act at other stages (non-dopamine) of the brain’s reward circuitry—perhaps in the anatomical cascade of GABAergic efferents… ”. Thus, based on our data and the comments made by Wise, v.s., we propose that GABA transport inhibitors, by inhibiting the reversed transporters (GABA-carriers) activated by chronic cocaine, should be considered as possible therapy for cocaine addiction.
Acknowledgments
The authors thank G. R. Hillman, Ph.D. for his expertise regarding statistical evaluation of the data.
Footnotes
-
Send reprint requests to: Dr. Joel P. Gallagher, Department of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, TX 77555-1031.
-
↵1 This work was supported by National Institutes of Health, National Institutes of Drug Abuse Grant DA-07190 and Training Grant T32-DA07287
- Abbreviations:
- DLSN
- dorsolateral septal nucleus
- COC-HYP
- membrane hyperpolarization associated with cocaine challenge
- TTX
- tetrodotoxin
- GABA
- γ-aminobutyric acid
- GABAA
- GABAA receptor
- GABAB
- GABABreceptor
- COC-7
- chronic in vivo 7-day, twice daily, cocaine treatment
- COC-14
- chronic in vivo 14-day, twice daily, cocaine treatment
- COC-28
- chronic in vivo28-day, twice daily, cocaine treatment
- Ri
- membrane input resistance
- f-ipsp
- fast inhibitory synaptic potential
- s-ipsp
- slow inhibitory synaptic potential
- EPSP
- excitatory postsynaptic potential
- CPT
- 8-cyclopentyl-1, 3-dimethylxanthine
- ACSF
- artificial cerebrospinal fluid
- d-AP5
- (d)-2-amino-5-phosphonovaleric acid
- CNQX
- cyano-7-nitroquinoxaline-2,3-dione
- MP
- membrane potential
- DA
- dopamine
- NE
- norepinephrine
- 5-HT
- serotonin
- NIDA
- National Institute of Drug Abuse
- Received August 11, 1997.
- Accepted March 27, 1998.
- The American Society for Pharmacology and Experimental Therapeutics