Abstract
Common neurobiological substrates contribute to the progressively increased behavioral effects (i.e., sensitization) that occur with repeated intermittent treatments of cocaine and morphine. Consequently, repeated exposure to cocaine can augment responding to morphine (termed cross-sensitization). Drug-induced sensitization in rats may model aspects of the dysfunction in motivation that are imposed by addiction. The ventral pallidum (VP) is involved in motivated behaviors and its function is altered by acute administration of cocaine and morphine, but the effects of repeated drug exposure remain unknown. Targeting this paucity, the present study evaluated electrophysiological changes in the VP of rats exposed to five once-daily cocaine treatments (15 mg/kg i.p.). This regimen also induced behavioral-sensitization that was expressed 3 days later when the rats received either an acute injection of cocaine (15 mg/kg i.p.) or morphine (10 mg/kg i.p.). VP neurons recorded in vivo 3 days after the repeated cocaine treatment regimen demonstrated increased excitatory responding to microiontophoretic applications of morphine and glutamate. The maximal effect (Emax) was increased without altering potency, suggesting a change in the functional efficacy of the respective receptor systems. This did not represent a potentiation in transmission in general, for the effects of GABA were diminished. The results provide the first evidence for cellular adaptation in the VP after a sensitizing drug treatment paradigm and reveal that cross-sensitization of drug-induced behaviors temporally correlates with changes in VP neuronal responding. These findings advance an emerging theme that alterations in the VP may contribute to the increased motivation for drug seeking that occurs in drug-withdrawn addicts.
The repeated, intermittent administration of a variety of potentially addictive drugs produces persistent increases in both their psychomotor activating effects (Post and Rose, 1976; Stripling and Ellinwood, 1977) and their incentive motivational properties (Lett, 1989; Piazza et al., 1989; Shippenberg and Heidbreder, 1995). This phenomenon is known as behavioral sensitization. The neural changes that underlie behavioral sensitization are thought to contribute to the development of the compulsive patterns of drug seeking and drug craving that characterizes addiction (Robinson and Berridge, 1993; for review, see Stewart and Badiani, 1993). Most research on the neurobiology of sensitization has focused on mesocorticolimbic brain regions because these regions share the ability to influence both the motor and reinforcement properties of abused drugs. It is well established that the ventral tegmental area (VTA), the nucleus accumbens (NAc), and the prefrontal cortex are critical contributors to drug-induced behavioral sensitization (Kalivas and Stewart, 1991; for review, see Pierce and Kalivas, 1997; Wolf, 1998; Vanderschuren and Kalivas, 2000). The ventral pallidum (VP) serves as a major output for the NAc (Groenewegen and Russchen, 1984; Zahm and Heimer, 1988; Heimer et al., 1991; Chrobak and Napier, 1993) and is a common target site for projections from the VTA (for review, see Napier et al., 1991a), prefrontal cortex (Sesack et al., 1989), and amygdala (Fuller et al., 1987; Mitrovic and Napier, 1998). As these anatomical connections suggest, behavioral studies have revealed that the VP is involved in both the motor (Austin and Kalivas, 1990, 1991; Hoffman et al., 1991; Napier, 1992; Napier and Chrobak, 1992; Fletcher et al., 1998) and the incentive (Hiroi and White, 1993; Gong et al., 1996; Gong et al., 1997; Fletcher et al., 1998; McFarland and Kalivas, 2001) properties of psychomotor stimulants and opiates. Moreover, injection of the μ-opioid receptor antagonist D-Phe-Cys-Tyr-D-Trp-Orn-Thr-Pen-Thr-NH2 (CTOP) into the VP prevents the behavioral sensitization produced by systemic injections of morphine (Johnson and Napier, 2000), suggesting that this region may also regulate sensitization processes.
Electrophysiological evaluations have greatly aided in defining the functional correlates at the cellular level for drug-induced sensitization observed behaviorally. For example, the NAc of cocaine-sensitized rodents exhibits neuronal activity changes, which, depending on the post-treatment withdrawal period, can include increases in dopamine receptor function (Henry and White, 1991; Beurrier and Malenka, 2002), reductions in the excitatory effects of glutamatergic ligands (White et al., 1995; Thomas et al., 2001), and enhancements in the inhibitory effects of GABA (Henry and White, 1995). We sought to show behavioral sensitization would also correlate with changes in the function of VP neurons. To do so, we electrophysiologically investigated VP neuronal responding in chloral hydrate-anesthetized rats at a time period when cocaine-induced behavioral sensitization was shown (in a separate group of rats) to be maintained. Evaluated were transmitter systems contained in the projections from several limbic brain regions, the VTA (dopamine), amygdala/cortex (glutamate), and the NAc (GABA and enkephalin). Cocaine-sensitized rats exhibit enhanced motor responding to an acute challenge of the μ-opioid receptor agonist morphine (Shippenberg and Heidbreder, 1995), a phenomenon referred to as cross-sensitization. Thus, we opted to focus on the μ-opioid receptor of the enkephalinergic system. To ascertain whether VP μ-opioid receptors also were capable of cross-sensitization at a withdrawal time when cross-sensitization occurred behaviorally, we compared the functional consequences of an acute morphine challenge both at the behavioral and cellular level in cocaine-sensitized rats.
Materials and Methods
Animals
Male Sprague-Dawley rats weighing 280 to 340 g (Harlan, Indianapolis, IN) were housed in pairs in environmentally controlled conditions (7:00 AM light/7:00 PM dark cycle, with temperature maintained at 23–25°C) with continuous access to rat chow and water. They were acclimated to colony conditions for at least 1 week before experimentation. All animals were handled in accordance with the procedures established in the Guide for the Care and Use of Laboratory Animals (National Research Council, Washington, DC). The specific protocols were approved by the Loyola University Medical Center, Institutional Animal Care and Use Committee.
Behavioral Sensitization Protocol
Drug Treatments and Behavioral Assessments. The treatment protocol used was one previously shown to result in behavioral sensitization (White et al., 1998). It consisted of three phases, conducted over a 15-day period (with the exception of the control study presented in Fig. 4). Phase 1. For the first 5 days, the rats were handled using procedures that emulated the subsequent treatment injection and behavioral assessments (including acclimation to motor test box). Motor activity was quantified on the last day, and this score was considered the rat's baseline. Phase 2. One day later, the rats were randomly assigned to one of two treatment groups; one group received saline (1 ml/kg i.p.) once daily for 5 days, and the other group was similarly injected with 15 mg/ml/kg cocaine (as the HCl salt in saline; Sigma-Aldrich, St. Louis, MO). This phase is referred to as the repeated treatment (RT). On all days, rats were transferred from home cage into clear test boxes (of the same dimensions as the home cage). After a 30-min preinjection habituation, the rats received their respective i.p. injections and were replaced into the test box for 60 min. Motor activity was quantified for both the pre- and postinjection periods via five sets of infrared photocell beams along the longitudinal axis and 3 cm up from the floor (Applied Concepts, Ann Arbor, MI), and the number of photobeam breaks was tallied by a computer in 5-min bins. Behavioral observations also were conducted by two observers who were unaware of the rats' treatment history. Scores were assigned after 1 min of observation every 10 min according to the following motor category score: 1) asleep: resting, eyes closed; 2) inactive: resting with eyes open, little or no movement; 3) slow active: periodic sniffing, infrequent locomotion and rearing/wall climbing, intermittent grooming; 4) normal active: occasional locomotion and rearing/wall climbing, investigational sniffing without a repetitive pattern; 5) hyperactive: faster, nonstereotyped sniffing, locomotion, with frequent rearing/wall climbing, head bobbing and sniffing; 6) intermittent stereotypy: a repetitive motor pattern that often included rearing/wall climbing, head-bobbing, and sniffing in one corner and then rapidly trans-locomoting to next corner and repeating the behavior; and 7) stereotypy: fast, perseverative behaviors with prominent sniffing and head bobbing, with little or no locomotion. Eighty-six percent of the time, the two observers agreed in the assignment of scores; otherwise, the average of the two was used for subsequent analysis. Phase 3. Three days later, half of each of the treatment groups received an acute challenge (AC) injection of cocaine (15 mg/ml/kg i.p.) and the other half received morphine (10 mg/ml/kg i.p. as the SO4 salt in saline; Mallinckrodt, Hazelwood, MO). The AC treatment designations were conducted by ranking the RT5 beam break scores and alternately assigning the rats to the two AC treatment groups, so that the average RT5 score for those receiving AC cocaine were not different from those injected with morphine. The test paradigm was the same as that used for RT days 1 through 5 (RT1–5), with the exception that motor activity was assessed for 90 min postinjection. This longer observation period was required to accommodate the longer half-life of morphine. The morphine-induced behaviors were characterized by comparing the number of times rats demonstrated one of two mutually exclusive “cataleptic-like” behavioral patterns. The cataleptic-like aspect of the patterns included a rather rigid, hypomotoric (“frozen”) state accompanied by a Straub tail, exophthalmos, pilo-erection, and a hyper-reactivity to sensory stimuli. The distinguishing features of the pattern involved two vastly different motor behaviors that intermittently interrupted this frozen state. One type of interruption involved bursts of rapid locomotor activity where the animal often ran to the opposite side of the cage. This pattern is abbreviated CL. The other pattern was distinguished by the lack of locomotion. Here, the frozen state was intermittently interrupted by confined motor behaviors largely comprised of rapid, jerky side-to-side movements of the front paws and/or head, abbreviated CPH.
For the control study presented in Fig. 4, phases 1 and 2 were conducted as stated above. Phase 3, however, involved three groups of rats: untreated rats, and those who received five repeated injections (RT1–5) of cocaine or RT1–5 of saline. The last two groups received an acute challenge of cocaine (15 mg/kg i.p.) 3 days after RT5 (AC1) and again 24 h later (AC2). The next day, all rats received a morphine injection (10 mg/kg i.p.; AC3).
Behavioral Data Summary and Analysis. To evaluate responding during phase 2, a two-way repeated measures ANOVA (rmANOVA) was used to compare responding on RT1 versus RT5, ascertaining the effects of time versus treatment and their interaction; pairwise contrasts between pretreatment groups were analyzed at specific time points with Newman-Keuls. Because a repeated measures evaluation of responding during RT1 versus RT5 was conducted in the same rats for both dependent variables (time and treatments), the chance for a type II error is increased. To reduce this chance, for these evaluations (both the ANOVA and the Newman-Keuls) the level for a significant α was set at 0.025. For evaluations of scores totaled across time, within pretreatment group scores were analyzed for the RT1 and RT5 using paired t tests. Between-pretreatment group effects of an acute challenge (AC; phase 3) were evaluated with ANOVA or Student's t test. A Mantel-Haenszel chi square was used for distribution statistics. p < 0.05 was used to indicate significance. Data are presented as the mean ± S.E.M.
Electrophysiological Experiments
Three days after the last repeated injection of cocaine or saline as described above, a set of rats separate from those used to assess behavior was anesthetized with chloral hydrate (400 mg/kg i.p.; Sigma-Aldrich) and prepared for in vivo electrophysiological recordings of VP neurons. A lateral tail vein was cannulated for i.v. supplements of chloral hydrate that were given as necessary to maintain a surgical level of anesthesia for the duration of the experiment. (No anesthesia was administered during microiontophoretic applications of test ligands). The animals were placed into a stereotaxic apparatus (David Kopf Instruments, Tujunga, CA) with the nosepiece set at 3.3 mm below the horizontal. A midline scalp incision was made and the skull was exposed. To allow sampling throughout the VP, a 3-mm burr hole was drilled in the skull, centered at 0.04 mm posterior to bregma and 2.3 mm lateral to the midline. Core temperature of the rat was maintained at 35–37°C with a thermostatically controlled heating pad (Fintronics Inc., Orange, CT). Electrophysiological recordings were made during the animals' light cycle. Frequently, more than one cell per rat was evaluated.
A multibarrel pipette-microelectrode assembly was used for microiontophoretic drug administration and extracellular monitoring of individual action potentials, as used previously (Mitrovic and Napier, 1995; Turner et al., 2001). In brief, single glass pipettes and multibarrel pipettes purchased prefilled with glass fibers (A-M Systems Inc., Everett, WA) were heat-pulled to form tips that were broken back to 2 to 3 and 10 to 15 μm, respectively. The single pipette served as the recording microelectrode, and it was glued in parallel with the multibarrel iontophoretic pipette so that the microelectrode extended 20 to 22 μm beyond the iontophoretic pipette tip. The microelectrode and the center barrel of the iontophoretic pipette (used for current balancing) were filled with a 0.5 M sodium acetate/2% pontamine sky blue solution. The four outside barrels of the microiontophoretic pipette were individually filled with GABA, glutamate or morphine (100 mM as the base in sterile water). GABA and morphine were ejected with a cationic current; glutamate was ejected with an anionic current. Pipette impedances, measured in vitro at 165 Hz (microelectrode tester; Winston Electronics, San Francisco, CA), ranged from 4 to 8 MΩ for recording microelectrodes, 10 to 22 MΩ for current-balancing barrels, and 25 to 85 MΩ for drug-containing barrels.
The pipette assembly was stereotaxically lowered through the burr hole and advanced through the VP with a hydraulic microdrive (Trent Wells, Inc., Southgate, CA). Action potentials were amplified, filtered (200 Hz and 2 kHz, half-amplitude frequency cut-off), and isolated from background using an amplifier/voltage discriminator (Fintronics Inc.) with the discriminator output fed into a computer fitted with an interface card and using in-house software for data collection and histogram generation. Action potentials and the discriminator output were continuously monitored on a storage oscilloscope (Tektronix Inc., Southgate CA). A stable baseline firing was obtained (i.e., spike frequency varying <15% for at least 5 min) during which the action potential characteristics were documented (for details, see Turner et al., 2002). Each ligand then was ejected serially, and spiking rate changes were monitored. Ejection-response curves were obtained as follows. For GABA and glutamate, two epochs of 9-s ejection/18-s retention in 10-nA increments from 10 to120 nA were used, with the second epoch typically used for data analysis. For morphine, continual incremental current application of 1, 2, 4, 8, 16, 32, 64, and 128 nA/60 s was used. Based on our prior evaluations with morphine iontophoresis (Johnson and Napier, 1997b), a 20% change from baseline firing rate by 32 nA obtained by two trials was considered the minimal indication of responsiveness to an iontophoresed ligand.
At the end of the experiment, an anionic current was passed through the recording electrode to deposit pontamine sky blue dye at the electrode tip and mark the last recording site. Rats were administered an overdose of chloral hydrate. Brains were removed, blocked, and freeze mounted onto microtome/cryostat chucks. Fifty-micrometer-thick slices of the forebrain were obtained, mounted on gel-coated slides, and then stained with cresyl violet. The anatomical location of the pontamine sky blue dye deposit was determined on agreement of at least two persons. This site was used as a reference to locate the remaining recording sites, based on their stereotaxic coordinates. Only recording sites contained within the VP and rostral aspect of the sublenticular substantia innominata (Fig. 5) were included in the study.
Electrophysiological Data Summary and Analysis. Ligand-induced changes in spiking were summarized as percentage of basal activity. For morphine, basal spiking was averaged from ten 3-s time bins (30 s) that occurred immediately before initiating the continuous incremental ejection currents. For glutamate and GABA, basal activity was obtained for each retention/ejection epoch using the three bins (9 s) immediately preceding each ejection period. In 7 of 30 glutamate recordings analyzed, basal spiking was not consistently reobtained after using higher ejection currents (due to residual actions of higher synaptic glutamate concentrations). Thus, for these, pretreatment basal activity was quantified, as for morphine. The microiontophoretic ejection current-response relationship (analogous to dose or concentration-response comparisons) was used to ascertain the following questions: 1) Does the magnitude of the response correlate with the ejection current magnitude (addressed using repeated measures ANOVA); and if so, do responses along the ejection current curve differ between saline- and cocaine-pretreated rats (addressed with a post hoc Newman-Keuls). To allow this analysis, VP-located neurons that responded to a ligand and were tested by at least four incremental ejection current levels were averaged, and those recordings containing a response that fell outside 2 S.D. of the mean were excluded from further analysis (the latter criterion eliminated four of 81 recordings. 2) Did cocaine pretreatment alter the relative efficacy/maximal effect (Emax) and/or ligand potency, i.e., the ejection current required to produce half Emax (ECUR50)? To address this question, nonlinear regression analysis was carried out for the ejection current-response curves with more than four levels of ejection current tested (n = 73). Those curves with an R2 > 0.7 (this criterion omitted 1 of 73) were used to calculate single point assessments of Emax and ECUR50 (Turner et al., 2002). Student's t test comparisons were made of the data that fell inside 2 S.D. of the mean from saline- and cocaine-pretreated rats (4 of 72 values were >2 S.D. and thus were not analyzed).
For glutamate and morphine, the higher ejection currents frequently induced a decrease in action potential amplitude, often accompanied by an increase in spike duration that rapidly heralded a cessation of spiking. Full recovery occurred immediately after terminating the ejection current. As previously characterized for VP neurons (Turner et al., 2002), this profile likely reflects excessive activation such that a depolarization-induced inactivation of the spike-generating processes ensues. Thus, to aid in establishing an Emax asymptote for the rate-increase phase of the bimodal curves, the “cut-off” ejection current for statistical evaluations was standardized to the level that preceded a decrease in peak responding by 20% for morphine and 100% for glutamate. To ascertain whether cocaine pretreatment could alter the propensity of VP neurons to enter into this “overexcited state”, apparent depolarization-induced inactivation was considered to have been initiated at the ejection current where an approximate decrease by 25% of the spike amplitude was first observed. A Pearson or Mantel-Haenszel chi square test was employed for distribution statistics. p ≤ 0.05 was used to indicate significance, and data are presented as the mean ± S.E.M.
Results
Behavioral Evaluations
Behavioral assessments were conducted to verify that the cocaine treatment protocol used 1) was sufficient to induce sensitization, 2) that after the cocaine treatment was terminated, this effect persisted to the time when the electrophysiological studies were conducted, and 3) that cross-sensitization to morphine occurred at this withdrawal time period. Figures 1, 2, 3 revealed that all three objectives were met.
A qualitative comparison between RT1 saline in Fig. 1, A to C, with RT1 cocaine in Fig. 1, D–F, shows that a single injection of cocaine (15 mg/kg i.p.) increases motor activity in rats. This was demonstrated by both the automated assessments (Fig. 1, D and E) as well as observationally scored behavior (Fig. 1F). The motoric effects were enhanced by five once-daily injections of cocaine (RT1 versus RT5; Fig. 1, D and E, statistical analyses are in figure legends). Observational determinations revealed that the sensitized response included enhanced locomotion and rearing/wall climbing that were intermittently interrupted with stereotypic head-bobbing and air sniffing (Fig. 1F). The peak motor response occurred within 10 min of the i.p. injection for both the first (RT1) and fifth (RT5) injection, with the magnitude of the RT5 peak being double that obtained for RT1 by automated evaluations. The enhanced responding was relatively brief as RT5 scores approached those obtained on RT1 by 40 min postinjection.
We next determined whether the RT history of the rat influenced responding to an AC after three drug-free days. To do so, the rats from the saline and cocaine RT groups were randomly assigned to two AC treatments: cocaine (15 mg/kg i.p.; Fig. 2) or morphine (10 mg/kg i.p.; Fig. 3). An AC with cocaine produced enhanced responding in rats with a cocaine history in all three motor assessments. Thus, cocaine-induced behavioral sensitization is expressed after a 3-day withdrawal from five daily cocaine treatments. Additional relevant information can be implied by comparing the data presented in Figs. 1 and 2. First, the profile seen after the first cocaine injection (i.e., RT1, closed diamonds, and the first bar in the bar graphs in Fig. 1, D–F) is recapitulated by the AC in rats with a saline pretreatment history (open squares and bar graphs in Fig. 2). This qualitative conclusion is also obtained if the data are statistically “reanalyzed” so that the group comparisons are RT1 cocaine versus responding after AC cocaine in rats with RT saline, i.e., p > 0.05 with repeated measures ANOVA. These observations suggest that repeated exposure to the test paradigm imposed by five daily injections of saline does not contribute to the response obtained with a single cocaine injection. Further studies are required to validate this possibility. Second, the peak cocaine effect seems to be maintained after drug withdrawal; compare RT5 in Fig. 1, D–F (the closed squares and right bar in the bar graphs), with AC cocaine after RT cocaine in Fig. 2 (the closed symbols and bars), making note of the different graph scales. This conclusion is suggested when the data were reanalyzed statistically, for no difference was obtained between RT5 cocaine and AC cocaine in cocaine-pretreated rats (p > 0.05).
An AC with morphine (10 mg/kg i.p.) evoked a behavioral profile that was distinct from that obtained with AC cocaine. One striking difference is that the magnitude of the automated assessments is about one-third of that obtained with AC cocaine (note the scale differences between Figs. 2, A and B, and 3, A and B). Another is the biphasic nature of the response profile to morphine. In drug-naive rats (i.e., those in the RT saline group, shown in Fig. 3 as open symbols and bars), morphine AC caused an initial cataleptic-like state that was intermittently interrupted with bouts of front paw/head jerks or locomotion bursts. During the last 30 min of the 90-min observation period, the rats' motor behavior gradually returned to, and slightly increased over, preinjection baseline levels. As illustrated by the beam breaks and crossings assessments, this latter motor increase was enhanced in rats with a cocaine RT history. The observational scoring method that successfully described the particular behaviors that were sensitized after cocaine AC was ineffectual in characterizing pretreatment group differences in the morphine AC. However, observational assessments that focused on motor behaviors that seemed to “break through” a morphine-induced cataleptic-like state revealed that the increased automated scores shown in Fig. 3, A and B, related to more frequent intermittent locomotor “bouts” in rats with a cocaine history (abbreviated CL). It is noteworthy that these bursts of locomotion most typically involved complete translocations of the activity chamber, as also was reflected in the crossings scores shown in Fig. 3B. The CL profile replaced the paw and head jerks (CPH) seen with morphine in saline-pretreated rats (Fig. 3C). Thus, the observational results show that morphine AC induced a cataleptic-like state that was interrupted with localized paw and head-jerking behaviors and that this profile was expanded to include locomotor outbursts if the rats had a prior experience of repeated cocaine injections.
The three motoric assessments concur in showing that the locomotion phase of motor responses to 10 mg/kg morphine demonstrated sensitization in rats withdrawn from five repeated cocaine treatments. Sensitization did not occur when the morphine AC is given 24 h after only two once-daily injections of cocaine (Fig. 4). This lack of responding reflects the reduced number of cocaine pretreatments, not the 24 h postcocaine withdrawal time, for morphine did enhance responding 24 h after cocaine if the rats had been subjected to a three once-daily cocaine treatment regimen (Fig. 4).
Insofar as behavior is a useful index of the state of the brain, we determine to replicate the 5-day repeated cocaine plus a 3-day withdrawal treatment protocol in a separate set of rats to ascertain whether VP neuronal responding to various transmitter ligands is changed in a cocaine-sensitized brain.
Electrophysiological Evaluations
In total, 128 neurons were evaluated and as shown in Fig. 5, the entire rostral/caudal extend of the VP was sampled for each of the ligands tested. In common with prior studies (Turner et al., 2001; Mitrovic and Napier, 2002), 92% of the recordings in saline-pretreated controls were of action potentials with a biphasic waveform, 8% were triphasic, and 26% of all action potentials exhibited a positive-going initial deflection. The average action potential amplitude was 520 ± 40 μV, and the duration was 1.5 ± 0.06 ms. In general, similar action potential characteristics were obtained from VP recordings taken from rats pretreated with cocaine; with p > 0.05 in t tests for amplitude and duration. Chi square evaluations for action potential wave form descriptors (e.g., bi- or triphasic waveforms) revealed no distribution differences between the two pretreatment groups, except that a higher portion of cells recorded from saline-pretreated group exhibited action potentials with an initial negative deflection [χ2(1) = 6.20, p = 0.013]. Thus, to the extent that as a collective, these electrophysiological characteristics can indicate, it seems that similar populations of neurons were spontaneously active, and sampled, in the saline- and cocaine-pretreated groups. In contrast, the basal/spontaneous firing rate of neurons was altered; VP recordings from saline-pretreated rats equaled 16.0 ± 1.6 spikes/s (n = 78), whereas the rate was 11.4 ± 1.1 spikes/s (n = 50) in cocaine-pretreated rats [t(68) = 2.26, p = 0.027]. To control for this difference in spontaneous firing between the two RT groups, responses of individual neurons to microiontophoretically applied ligands were standardized to percentage of their spontaneous firing rate (baseline) as used previously (White and Wang, 1984; Henry et al., 1989; Pitts and Marwah, 1989; Pitts et al., 1990; Napier et al., 1991b; Simson et al., 1992; Heidenreich et al., 1995; Mitrovic and Napier, 1995; Johnson and Napier, 1997a,b).
Locally applied morphine is known to increase spiking in VP neurons (Napier et al., 1992; Chrobak and Napier, 1993; Mitrovic and Napier, 1995), and this effect may involve an inhibition of GABAergic terminals on projections from the nucleus accumbens (Chrobak and Napier, 1993). Others have shown that with higher microiontophoretic ejection currents, morphine-induced rate increases often progress to where the action potentials displayed hallmarks of depolarization inactivation (Matthews and German, 1984; Henry et al., 1992), i.e., increased duration and decreased amplitude of the action potential until the spike can no longer be discriminated from background “noise” (Grace and Bunney, 1986; Turner et al., 2002). Thus, in the present study, we evaluated the effect of cocaine pretreatments on the morphine response profile of both indices: rate increase and apparent depolarization inactivation. Of the VP neurons tested in saline-pretreated rats, 71% showed rate increases during morphine applications (Table 1; Fig. 6). Similarly, 78% of the tested neurons in rats withdrawn from repeated injections of cocaine showed morphine-induced rate increases. Although the proportion of responding neurons did not change, the ejection current profile did, with enhanced responding occurring at 32 and 64 nA (Fig. 6B) in recordings from cocaine-pretreated rats. This was not reflected statistically in the single-point assessments of Emax or ECUR50 (Fig. 6B, bar graphs), likely owing to the fact that these indices could not be reliably calculated from many of the individual ejection current-response curves because they did not exhibit asymptotes within the range of ejection currents used. So, although it remains unclear whether the shift of the morphine ejection current-response curve reflects changes in efficacy, it is apparent that the capacity of morphine to cause increased spiking is enhanced in rats with a history of cocaine treatments. As shown in Fig. 6C, this increase in the rate-enhancing properties of morphine occurred without altering the sensitivity of the neurons to the apparent depolarization inactivation properties of the drug.
Three days after the same cocaine treatment regimen used in the present experiments, the rate-enhancing effects of glutamate are attenuated in the NAc (White et al., 1995), but the ability of the excitatory amino acid to induce depolarization inactivation is enhanced in dopaminergic neurons recorded from the VTA (White et al., 1995; Zhang et al., 1997). We determined that responding of the VP was similar to that seen for dopaminergic neurons, for the rate-enhancing effects of glutamate were augmented in the cocaine-pretreated group. This response was distinguished by an increase in the apparent efficacy of the excitatory effects of glutamate (Fig. 7B, left bar graph). Moreover, although depolarization inactivation to local applications of glutamate readily occurred in the VP (Fig. 7A), it took considerably larger ejection currents of glutamate to induce inactivation in the cocaine-pre-exposed VP (Fig. 7C). These response profile changes occurred in the absence of changes in the proportion of responding neurons (Tables 1 and 2).
GABA transmission is altered in striatal brain regions of cocaine-sensitized rats (Henry and White, 1995; Jung et al., 1999). Because the effects of morphine in the VP may be mediated by GABA (Chrobak and Napier, 1993), we considered the possibility that GABAergic transmission may also be altered in the VP of cocaine-pretreated rats. As shown in Fig. 8, a rightward shift in the ejection current-response curve for the rate suppressant effects of GABA was obtained from the cocaine group. This shift was reflected by a decrease in relative potency for GABA (i.e., increased ECUR50; Fig. 8B, bar graphs). Complete rate suppression (Emax) was obtained in both pretreatment conditions (Fig. 8B, bar graphs). The number of neurons that responded to GABA also was not altered by the cocaine pretreatments (Tables 1 and 2).
Discussion
Aside from the general paucity of electrophysiological evaluations of neuronal changes induced by repeated cocaine treatments, several features of the present study are uniquely relevant to the understanding of brain adaptations that occur in the cocaine-sensitized animal. First is the demonstration that spontaneous neuronal activity in the VP (of a chloral hydrate-anesthetized rat) is reduced in cocaine-sensitized rats. Second is the discovery that neurotransmitter function within the VP is different in cocaine-sensitized rats than in controls. Third is the demonstration that behavioral cross-sensitization between morphine and cocaine occurs at the same time period after the repeated cocaine treatment as do changes in VP neuronal responding to morphine. These findings extend our view of the neuronal substrates that underlie drug-induced sensitization.
Spontaneous spiking reflects both in the intrinsic excitability of neurons as well as receptor activation after the tonic release of endogenous transmitters. Our data suggest that one or both of these functions are altered in the VP of drug-withdrawn cocaine-sensitized rats. Although not yet studied for the VP, there are numerous reports of other Limbic brain regions that a variety of biochemical and electrophysiological events that may be involved in such adaptations (for review, see White and Kalivas, 1998). Particularly compelling is a demonstration with the same cocaine pretreatment/withdrawal period protocol used in the present study, of a clear reduction in voltage-dependent sodium currents and a hyper-polarization of the resting membrane potential of medium spiny neurons in the NAc (Zhang et al., 1998). It is plausible that the observed reduction in VP spontaneous activity reflects a similar adaptation. Moreover, because medium spiny neurons project to the VP (Walaas and Ouimet, 1989) and their activation evokes both excitatory and inhibitory VP responses (Chrobak and Napier, 1993), the finding that these accumbal neurons are less likely to fire after short-term withdrawal from repeated cocaine also implies that there should be a decrease in the tonic endogenous release of accumbal-pallidal transmitters. Pharmacological inactivation of the NAc readily decreases spontaneous firing of VP neurons (Napier, 1992). This rate decrease likely reflects a reduction in the release of transmitters contained in this pathway that can increase VP spiking (either directly or indirectly), including substance P (Mitrovic and Napier, 1998) and opioids (Chrobak and Napier, 1993). The enhancement of pallidal responses to exogenous (microiontophoretic) applications of the opiate morphine in cocaine-sensitized rats may be an adaptive consequence of a decrease in endogenous opioid release.
The ability of repeated treatments of psychomotor stimulants to enhance responding to a subsequent challenge of an opiate is contingent upon the route and timing of administration of the two classes of drugs. When the stimulant amphetamine is repeatedly administered directly into the VTA (e.g., five to eight infusions, given less frequently than once daily), locomotor cross-sensitization is seen to low doses of systemically administered morphine (1–2.5 mg/kg) 2 to 11 days later (Vezina and Stewart, 1990; Cador et al., 1995). However, when the acute challenge is given 3 weeks after 14 daily, intravenous cocaine self-administration sessions, locomotor cross-sensitization does not occur with heroin but is observed with amphetamine (De Vries et al., 1998). Moreover, rats pretreated with amphetamine (five daily treatments of 2.5 mg/kg i.p.) fail to demonstrate locomotor cross-sensitization to morphine (2 or 5 mg/kg s.c.) when tested after either 3 days or 3 weeks of amphetamine withdrawal (Vanderschuren et al., 1999). Clearly, the drug and dosing regimen used, as well as the withdrawal period imposed before the acute challenge greatly impacts on the ability of repeated psychomotor stimulant treatments to enhance responding to opiate administration. The results shown here reveal that five once-daily i.p. injections of cocaine (15 mg/kg) are sufficient to induce cross-sensitization to the motor effects of a systemically administered, moderate dose (10 mg/kg i.p.) of morphine after a 3-day withdrawal period.
The phenomenon of cross-sensitization implies a common underlying mechanism for the tested ligands, in this case, cocaine and morphine. Concurring with this interpretation is the large body of literature showing that repeated treatments with cocaine or opiates enhance dopamine transmission in the NAc, and such a mechanism is thought to commonly regulate the hyperactivity seen with both drugs (for review, see White and Kalivas, 1998). We demonstrate here that the locomotor-activating effects of cocaine and morphine are indeed commonly enhanced with a cocaine history; however, several unique motor features of each of the drugs also were augmented. For cocaine, the rearing/wall climbing behaviors seen with acute treatments are exaggerated, and the rats become more stereotypic if they had a cocaine treatment history. In contrast, morphine-induced cataleptic-like states seen with acute administration become interrupted with intermittent, localized paw and head-jerking behaviors in rats with a cocaine treatment history. These observations suggest that adaptations in distinct neuroanatomical substrates may overlay adaptations in common substrates. Neurons in the VP may serve this role, for behaviorally sensitizing dosing regimens of cocaine (present study) and morphine (our unpublished results) beget very different adaptations in VP transmission: VP neurons recorded from rats with a cocaine history show enhancements in the rate altering effects of morphine and glutamate, with a lessening of GABAergic inhibition, whereas in rats repeatedly pretreated with morphine, VP cell responding to glutamate and GABA are not altered, and the rate-enhancing effects of morphine are reduced. Because the withdrawal period for the morphine study was longer than the present one (i.e., 14 versus 3 days), additional studies with similar withdrawal periods are needed to validate this intriguing possibility.
The behavioral role of the VP in opiate cross-sensitization to cocaine also includes reward function, for cocaine-induced place conditioning is attenuated if during conditioning the cocaine is preceded by intra-VP injections of the μ-opioid receptor antagonist naloxone (Skoubis and Maidment, 2003). Regardless of the specific role, it is noteworthy that when behavioral cross-sensitization to morphine occurs, neurons in the VP become hypersensitive to morphine. This effect may involve VP neuronal adaptations in the expression or turnover regulation of μ-opioid receptors for repeated cocaine administration increases μ-opioid receptor binding in several brain regions, including the VP (albeit with longer treatments of higher doses than those used here) (Hammer, 1989; Unterwald et al., 1994).
Throughout the brain, an inhibition of GABAergic transmission is involved in the excitatory effects of morphine (Antonelli et al., 1986; Jiang and North, 1992; Johnson and North, 1992), including in the VP (Chrobak and Napier, 1993). After repeated cocaine in the present study, morphine-induced rate enhancement was up-regulated along with a down-regulation of the ability of GABA to decrease VP cell firing. This apparently occurs without changes in the presynaptic function of GABA, for repeated exposure to cocaine does not alter extracellular fluid levels of GABA in the VP (determined when cocaine was present) (Sizemore et al., 2000), and the number of GABAergic boutons is not changed in the VP after a 2- or 14-day withdrawal from five twice-daily injections of cocaine (De Leon et al., 2000). Thus, a postsynaptic mechanism may underlie the decrease in GABA potency seen in the present study for VP neurons in cocaine-pretreated rats. It is noteworthy that the VP GABAergic effects do not parallel that reported for other forebrain regions. For example, in striatal regions of cocaine-sensitized rats subjected to a short-term withdrawal (1–3 days), the inhibitory effects of microiontophoretically applied GABA are augmented (Henry and White, 1995) and GABA release is decreased (Jung et al., 1999). These regional differences illustrate the value of further investigations regarding GABAergic involvement in the neuronal adaptations that occur with repeated cocaine exposure. Underscoring the need to include the VP in these evaluations are the recent demonstrations that VP GABA influences behaviors associated with self-administration (McFarland and Kalivas, 2001) and repeated noncontingent injections (Fletcher et al., 1998) of stimulants.
An alteration in responsiveness to excitatory amino acid transmission also occurs after withdrawal from cocaine, and, like GABA, the profile of the alteration differs among various brain regions. In the VTA of rats withdrawn for 3 days from a 5-day cocaine treatment protocol, the rate-enhancing effects of lower ejection currents for glutamate are not altered, but the ability of higher ejection currents to induce an apparent depolarization-induced inactivation of firing is enhanced (White et al., 1995). Similar treatment and withdrawal protocols potentiate the rate-enhancing effects of low ejection currents of glutamate in the prefrontal cortex (Peterson et al., 2000). Within the NAc, the cocaine treatment protocol used in the present study attenuates responding to glutamate (White et al., 1995; Zhang et al., 1998). Continuing the theme of regional differences in responding to chronic cocaine treatments, we observed a robust enhancement of glutamatergic influences in the VP, whereas the portion of neurons responding to local application of glutamate was not altered by prior exposure to cocaine. Thus, it seems that additional neuronal populations that are sensitive to glutamate did not contribute to the observed augmented responding, and the enhanced sensitivity to glutamate within the VP may have occurred within a particular subset of neurons. These functional adaptations to noncontingently administered cocaine in the present study may occur without an alteration in glutamatergic inputs. Supporting this possibility, Smith et al. (2003) revealed that rats yoked to animals self-administering cocaine do not show changes in VP glutamate turnover (even though increases were seen in the self-administering rats).
Conclusions
In summary, this study demonstrated that repeated administration of cocaine profoundly alters neurotransmission in the VP, an effect that is correlated positively with motor activity. Glutamatergic inputs from, e.g., the cortex and amygdala, converge in the VP with accumbal opioid and GABAergic inputs. The behavioral sequelae that occurs during chronic exposure to cocaine is associated with particular adaptations within each of these regions of the brain's limbic system. Given that the VP is a primary output for the limbic system, an alteration in afferent influences within the VP would be anticipated to have a high impact on the behavioral consequences of repeated cocaine exposure. Recent reports describing the role of the VP in the drug-environment associations that occur with repeated psychostimulant treatments in a place preference paradigm (Gong et al., 1996) and cocaine reinstatement after extinction from cocaine self-administration (McFarland and Kalivas, 2001) provide a compelling evidence for this emerging theme.
Footnotes
-
This research was supported by the Ralph and Marian Falk Medical Research Trust (to J.M. and T.C.N.), the Irene Whitney Foundation through the Neuroscience and Aging Institute Loyola University Chicago (to T.C.N.), and U.S. Public Health Service Grants DA05255 and DA015760 (to T.C.N.).
-
doi:10.1124/jpet.105.084038.
-
ABBREVIATIONS: VTA, ventral tegmental area; NAc, nucleus accumbens; VP, ventral pallidum; RT, repeated treatment; AC, acute challenge; rmANOVA, repeated measures analysis of variance; ANOVA, analysis of variance.
- Received January 19, 2005.
- Accepted February 17, 2005.
- The American Society for Pharmacology and Experimental Therapeutics