Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cocaine addiction is a serious problem in the United States today for which there is no effective treatment (Kosten, 2001). Both genetic and environmental risk factors are involved in cocaine addiction with the genetic contribution explaining about one-third of the variance (Pickens and Svikis, 1988; Tsuang et al., 1998). We and others have examined the role of genotype in responsiveness to drugs of abuse through the study of two inbred rat strains, Lewis and Fischer 344, because they provide a valuable tool to help elucidate the neuropharmacological bases of cocaine addiction (Kosten and Ambrosio, 2001).

Lewis and Fischer 344 rats show marked differences in behavioral responses to a variety of drugs across a range of procedures thought to reflect drug reward. For example, Lewis rats exhibit greater conditioned place preference for cocaine (Kosten et al., 1994) and morphine (Guitart et al., 1992), but not for amphetamine (Stohr et al., 1998) compared with Fischer 344 rats and more readily acquire oral drug self-administration of opiates (Suzuki et al., 1992) and ethanol (Suzuki et al., 1988). Acquisition of intravenous morphine (Ambrosio et al., 1995) and cocaine (Kosten et al., 1997) self-administration also occurs more readily in Lewis compared with Fischer 344 rats. Once drug self-administration is acquired, both Lewis and Fischer 344 rats maintain the behavior in a dose-related manner (Kosten et al., 1997) and extinguish the behavior when the drug is no longer available (Ambrosio et al., 1995; Kosten et al., 1997).

The behavioral effects of cocaine are linked to activation of the mesocorticolimbic dopamine (DA) system (Ritz et al., 1987), a system that consists of cell bodies in the ventral tegmental area (VTA) that project to forebrain structures, including nucleus accumbens (NAc), medial prefrontal cortex, and others (Fuxe et al., 1985). Cocaine increases synaptic DA levels indirectly in NAc (Pettit and Justice, 1989) by inhibiting the DA transporter at presynaptic terminals (Harris and Baldessarini, 1973). The increased synaptic DA interacts with both D1- and D2-like receptors postsynaptically, and both types of receptors play a role in the behavioral effects of cocaine (Spealman et al., 1992).

Protein levels within the mesolimbic DA system differ between Lewis and Fischer 344 rats. In VTA, Lewis rats exhibit lower neurofilament and higher tyrosine hydroxylase and glial fibrillary acidic protein levels compared with Fischer 344 rats (Guitart et al., 1992; Haile et al., 2001). In NAc, Lewis rats have lower levels of D2-like DA receptors in NAc (Flores et al., 1998), Gi_1/2 (Guitart et al., 1993; Haile et al., 2001) with correspondingly higher levels of adenylyl cyclase and protein kinase A compared with Fischer 344 rats (Guitart et al., 1993). There are no strain differences in D1-like receptors (Flores et al., 1998) or in levels of its intracellular second messenger G_s (Guitart et al., 1993; Haile et al., 2001) in this area. In conjunction with human neuroimaging data demonstrating that lower striatal D2 levels are associated with cocaine addiction (Volkow et al., 1997), we have speculated that inherent strain differences in these mesolimbic protein level characteristics, particularly in the D2 receptor/intracellular system, may relate to the behavioral differences (Haile et al., 2001). Thus far, this supposition is based on correlative data.

The purpose of this study was to test the hypothesis that pharmacologically altering D2 receptor levels in rats that show inherent differences in populations of this receptor in NAc would lead to differential effects on the reinforcing and discriminative stimulus effects of cocaine. Specifically, we hypothesized that D2-like agents would have greater effects in Lewis rats because they have relatively lower D2 levels compared with Fischer 344 rats. The effects of D1-like agents may also vary due to the differences in D1-D2 interactions (Walters et al., 1987). Previous studies demonstrate that D1- and D2-like agents alter cocaine discrimination and self-administration behaviors (Spealman et al., 1992). However, this study tests whether there are differential effects based on inherent neuropharmacological characteristics.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals and Housing

Male Lewis and Fischer 344 rats (Harlan, Indianapolis, IN) weighing 280 to 380 g were used in these studies. Food (Purina Chow; Purina, St. Louis, MO) and tap water were available ad libitum except in experiments with food reinforcement (experiments 1 and 2). In these cases, rats were given a daily ration of food in an amount that allowed them to be maintained at 85% of free-feeding body weight. Food rations were presented at random times after training and testing sessions. Rats were individually housed in hanging, wire-mesh cages in a temperature-controlled colony room maintained on a 12-h light/dark cycle (lights on at 7:00 AM). All procedures were approved by the Institutional Animal Care and Use Committee in strict accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Apparatus

All experiments were conducted using standard operant chambers (Coulbourn Instruments, Allentown, PA). The chambers were housed in ventilated sound-attenuating cubicles (Coulbourn Instruments) equipped with fans to mask outside noise. Two response levers were located on one wall of the chamber with a food trough located between the levers. Each lever was located 1 inch from the side wall with three "cue" lights positioned directly above each lever. Downward pressure (about 25 g) on a lever could result in a programmed consequence and was tabulated. Chambers were interfaced to a PC computer and a software program (L2T2 or SAD8V5; Coulbourn Instruments) was used to program the session parameters and tabulate data automatically.

Drugs

Cocaine hydrochloride and dl amphetamine were provided by the National Institute on Drug Abuse (Research Triangle Institute, Research Triangle Park, NC). (±)-SKF-38393 hydrochloride, R-(+)-SCH-23390 hydrochloride, (±)-quinpirole dihydrochloride, S-()-eticlopride hydrochloride were purchased from Sigma/RBI (Natick, MA). Pentobarbital sodium was purchased from Sigma (St. Louis, MO). All drugs included the weight of the salt and were dissolved in sterile isotonic saline and administered in a volume of 1 ml/kg.

Experiment 1: Food-Maintained Responding

This study assessed the effects of various doses of the D1- and D2-like compounds on schedule-controlled responding for food in Lewis and Fischer 344 rats. One purpose was to obtain information on the appropriate dose to choose of each drug for experiment 3, the cocaine self-administration study. Doses that did not disrupt schedule-controlled responding for food (<75% of baseline response rates) in either strain were used in experiment 3, the cocaine self-administration study. Choices for doses of the drugs were also based on the literature.

Training. Lewis and Fischer 344 rats (n = 8/strain) were initially trained to lever-press for food pellets (45 mg; Bio-Serv, Frenchtown, NJ) under a fixed ratio (FR) 1 schedule of reinforcement. A total of 50 reinforcers could be obtained by depression of either lever during daily sessions. Sessions began with the illumination of the houselight and both sets of cue lights and terminated after 50 reinforcers were earned or 30 min elapsed, whichever occurred first. The FR requirement was gradually increased until an FR15 was achieved. Test sessions began once the rat showed consistent response rates (<20% variance over three consecutive days).

Testing. Test sessions were performed twice weekly (Tuesday and Friday) with training sessions continuing on Monday, Wednesday, and Thursday. The effects of pretreatment (30 min) with various doses of the DAergic compounds on operant response rates were examined. These drugs included the D2 antagonist eticlopride (0.01-0.1 mg/kg), the D1 antagonist SCH 23390 (0.005-0.05 mg/kg), the D2/D3 agonist quinpirole (0.001-0.1 mg/kg), and the partial D1 agonist SKF 38393 (0.1-10 mg/kg). Drugs and doses were tested in a nonsystematic manner across rats. The effects of cocaine (1-10 mg/kg i.p.) on operant response rates were also assessed.

Data Analysis. The dependent measure was response rate during the test session as a percentage of rates during the previous baseline training day. Data obtained from the effects of DAergic agents on responding for food were analyzed using analysis of variance (ANOVA) with between-group factor of strain and drug dose as a repeated measure. Separate analyses were performed for each drug. Baseline response rates were analyzed using Student's t test. Significance was set at P < 0.05.

Experiment 2: Cocaine Discrimination

This study assessed the effects of various doses of the D1- and D2-like compounds on cocaine discrimination in Lewis and Fischer 344 rats. This was the first study to examine cocaine discrimination in Lewis and Fischer 344 rats although a previous study examined morphine discrimination (Morgan et al., 1999). Thus, the present study included tests to examine dose-related discrimination behavior and the pharmacological specificity of the behavior in these two strains.

Training. Lewis and Fischer 344 rats (n = 8 per strain) were food-deprived and initially trained to lever press for food pellets (45 mg; Bio-Serv) under an FR1 schedule of reinforcement. The onset of the training sessions was signaled by illumination of the houselight and of the cue lights above both levers. Lever pressing for food reinforcement was shaped successively until completion of an FR10 requirement on either lever resulted in obtaining one reinforcer. When responding on either lever was stable, cocaine discrimination began. On cocaine training days, cocaine (10 mg/kg i.p.) was administered and the rat placed in the operant chamber. The session began 15-min later when the house and cue lights were illuminated. Every 10th response on the cocaine-designated lever produced a food pellet. On vehicle training days, saline was administered. In these sessions, every 10th response on the other lever (vehicle-designated) produced a food pellet. The designation of cocaine- and vehicle-appropriate levers was nonsystematic across rats. Sessions terminated after 50 reinforcers were obtained or 15 min had elapsed, whichever occurred first. A double alternation sequence of training (cocaine, cocaine, saline, saline, cocaine, etc.) was used. Training sessions were performed 5 days per week. The criterion for acquiring cocaine discrimination was when 90% or greater of the responses were on the drug-appropriate lever for six consecutive days.

Testing. Once discrimination criterion was met, substitution and antagonism tests were performed. Test sessions were conducted twice a week (Tuesday and Friday), with discrimination training sessions continuing on the other days (Monday, Wednesday, and Thursday). If discrimination performance fell below 90% correct responding on an intervening training session, test sessions were not conducted until three consecutive days of criterion performance was demonstrated. Test sessions were similar to training sessions, except food reinforcers were available when either lever was depressed 15 times.

Data Analysis. A rat must have pressed at least 45 responses on one lever to be included in the data analysis. Most tests were conducted on all rats. However, in some instances not all rats from a strain were tested in all dose combinations or data could not be included due to the exclusion criterion. The percentage of cocaine-appropriate lever responding (total cocaine lever responses/total lever responses × 100) and response rates (total lever responses/time of session) were determined for each test session. The criterion for full substitution was defined at 80% or more of total responses on the cocaine-appropriate lever. Discrimination data and response rates were analyzed with ANOVA with the between group factor of strain and repeated measure of dose. Separate analyses were performed for each drug test. Significance was set at P < 0.05.

Experiment 3: Intravenous Cocaine Self-Administration

This study assessed the effects of the D1- and D2-like compounds on cocaine self-administration in Lewis and Fischer 344 rats. This was the first study that examined the effects of pharmacological manipulations on the maintenance of cocaine self-administration in Lewis and Fischer 344 rats. A previous study investigated these effects in Fischer 344 rats compared with two other inbred rat strains (Ward et al., 1996).

Catheter Implant Surgery. Rats were anesthetized with pentobarbital (Nembutol, 50 mg/kg; Abbott Diagnostics, Chicago, IL) and implanted with indwelling catheters made of Silastic tubing as described previously (Kosten et al., 1997). Catheters were implanted into the right jugular vein and passed under the skin to exit at the nape of the neck through a cannula (22-gauge, curved metal tube) that was attached to the skull with jeweler's screws and dental acrylic cement. Rats were allowed to recover for a minimum of 4 days before experiments were initiated. Catheters were flushed once daily with a heparin-saline solution and catheter patency was tested at the end of the experiment by the ability of Brevital (Eli Lilly, Indianapolis, IN) injection (0.2 ml of a 10-mg/ml solution) to cause loss of consciousness within 1 to 2 s.

Training. Prior to surgery, rats were trained to lever press for food reinforcement under a continuous reinforcement schedule until consistent lever press responding was achieved (e.g., obtaining the maximum number of 50 food reinforcers on two occasions). After recovery from surgery, the cocaine self-administration training sessions were initiated. Rats were placed in the operant chambers and their catheter/cannula system was attached to a syringe pump system, which consisted of an infusion pump (Razel model A) with a 20-ml glass syringe, connected by a swivel (Stoelting #1) with Teflon tubing. The swivel was connected to the animals' cannula assembly with Tygon tubing protected by a metal spring and secured to a screw mounted on the animal's head. Initially, one depression of the active lever resulted in an intravenous infusion of 100-µl injection of cocaine (FR1) in which a dose of 1.0 mg/kg was delivered per infusion. The cocaine infusion was delivered over a 10-s period, followed by a 5-s time-out. Active lever presses emitted during the infusion time or during the time-out period were tabulated but did not result in any further cocaine delivery. During the infusion, the cue lights were illuminated. The house lights were turned off during the entire infusion and time-out periods. When cocaine showed control over behavior under the FR1 schedule of reinforcement, the FR requirement was raised to FR2 and then to FR3, the schedule used for the remainder of the experiment. We find that rats will reliably maintain cocaine self-administration under an FR3 schedule of reinforcement, and there is a decreased chance of tabulating spurious behavior with this schedule. Experimental parameters (i.e., schedules of reinforcements, time periods, etc.) were programmed using a software package (Coulbourn Instruments) installed on a PC computer, and data were tabulated using a Coulbourn-designed system. Once consistent self-administration responding was shown (number of self-injections <20% for three consecutive days), tests for dose-related responding to cocaine were initiated.

Testing. We used a within-session procedure that consisted of four, 1-h components in which the training dose (1 mg/kg) was available the 1st h and cocaine doses of 1.0, 0.5, and 0.25 mg/kg were presented, one dose per hour, during the subsequent 3 h. Doses were altered by varying the infusion times (20, 10, and 5 s for 1.0, 0.5, and 0.25 mg/kg/injection). This procedure was validated previously in our laboratory (Kosten et al., 1997) and in other laboratories (Caine and Koob, 1994; Sizemore et al., 1997). Furthermore, we have found no difference in rates of acquisition of cocaine self-administration between rats trained using either 10- or 20-s infusion times (T. A. Kosten and C. N. Haile, unpublished observations). The effects of pretreatment with eticlopride (0.03 mg/kg), SCH 23390 (0.01 mg/kg), quinpirole (0.01 mg/kg), and SKF 38393 (1 mg/kg) on cocaine self-administration were then assessed. Doses of the pretreatment drugs were based on data obtained from experiments 1 and 2. Additional tests were conducted because the preliminary findings suggested that eticlopride had no effect in Fischer 344 rats and SCH 23390 had no effect in Lewis rats. Thus, tests with higher doses of eticlopride (0.1 mg/kg) in Fischer 344 rats and of SCH 23390 (0.02 mg/kg) in Lewis rats were performed. Test sessions were run twice weekly (Tuesday and Friday) with training sessions continuing on Monday, Wednesday, and Thursday. Test sessions were run only if the rat showed consistent responding on the prior baseline, training day. Drug tests were run in a nonsystematic manner across rats.

Data Analysis. The effects of the drug treatments on numbers of active and inactive lever presses and self-injections were compared with those obtained under baseline (saline injections) conditions. Baseline data were only included in a drug comparison if the rat had been tested with that drug. Most drug tests included six to seven rats per strain, except for the quinpirole (n = 4) and SKF 38393 (n = 5) tests with Fischer 344 rats. The two additional tests, 0.1 mg/kg eticlopride and 0.02 mg/kg SCH 23390, used three Fischer 344 rats and five Lewis rats, respectively, but these tests were not included in the overall analyses. Data from each drug test were analyzed separately with a three-way mixed ANOVA with the between-group factor of strain and repeated measures of drug condition (baseline versus drug) and dose. If significant drug or strain effects were found, additional analyses were performed separately by strain. Significance was set at P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Experiment 1: Food-Maintained Responding

Baseline rates of responding do not differ significantly between strains (P > 0.10). The mean (±S.E.M.) response rate is 0.98 ± 0.06 lever presses/s for Fischer 344 rats and 1.16 ± 0.11 for Lewis rats. The effects of pretreatment with the various compounds on response rates (as percentage of baseline rates) are shown in Table 1. The D2-like antagonist eticlopride alters response rates significantly by dose, F(3,42) = 13.29; P < 0.0001, in both strains. As seen in Table 1, response rates increase slightly above baseline at lower doses (e.g., 0.03 mg/kg) and then decrease at the highest dose (0.1 mg/kg). The strain and strain by dose interaction effects are not significant (P values > 0.10). The D1-like antagonist SCH 23390 decreases response rates significantly by dose, F(4,56) = 41.52; P < 0.0001, in both strains. Response rates decrease to minimal levels at the highest doses (0.03 and 0.05 mg/kg) as seen in Table 1. Again, the strain and interaction effects are not significant (P values > 0.10). The D2/D3 agonist quinpirole decreases response rates significantly by dose, F(2,43) = 89.27; P < 0.0001. As seen in Table 1, response rates decrease with increasing dose, but this effect occurs more gradually in Fischer 344 rats compared with Lewis rats as supported by the significant interaction of strain by dose, F(2,43) = 10.57; P < 0.001. At the highest dose of quinpirole (0.1 mg/kg) response rates are disrupted to a much greater extent in Lewis rats compared with Fischer 344 rats. The main effect of strain is not significant (P > 0.10). Finally, the partial D1 agonist SKF 38393 also decreases response rates significantly by dose, F(2,40) = 15.05; P < 0.0001, as seen in Table 1. The strain and strain by dose interaction effects are not significant (P values > 0.10). The effects of cocaine on response rates do not differ by dose or by strain as seen in Table 1 (P values > 0.10).

The results of this study suggest that the following doses be used for experiment 3, the cocaine self-administration study: 0.03 mg/kg eticlopride, 0.01 SCH 23390, 0.01 quinpirole, and 1 mg/kg SKF 38393. These are the doses of each drug that do not disrupt responding (<75% of baseline response rates) in either strain.

Experiment 2: Drug Discrimination

There are no strain differences in the ability of cocaine (10 mg/kg) to gain stimulus control over behavior (P > 0.10). The number of days to meet acquisition criterion is 24 ± 3.0 days for Lewis rats and 24 ± 2.4 days for Fischer 344 rats (P > 0.10). Cocaine produces significant dose-related increases in cocaine-appropriate responding, as supported by the significant dose effect, F(4,44) = 24.59; P < 0.0001. Specifically, the lowest cocaine dose occasions vehicle-appropriate responding and the highest cocaine dose occasions cocaine-appropriate responding, as seen in Fig. 1A. Full substitution (80% cocaine-appropriate responding) occurs for a majority of Fischer 344 rats at 3.0- and at the 10-mg/kg training dose and for a majority of Lewis rats at 10 mg/kg as seen in Table 2. There are no strain differences in dose-related cocaine discrimination behavior (P values > 0.1). Response rates during discrimination sessions differ by strain as supported by the significant strain effect, F(1,13) = 9.00; P < 0.005. As seen in Fig. 1B, rates are lower in Fischer 344 rats compared with Lewis rats. The dose effect and interaction term are not significant (P values > 0.10).

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Mean (±S.E.M.) percentage of responding on the cocaine-appropriate lever (A) and mean (±S.E.M.) response rates (number of lever presses/s; B), after administration of various doses of cocaine are presented for Lewis (closed symbols) and Fischer 344 (open symbols) rats.

Substitution tests with amphetamine show that cocaine-appropriate responding increases with increasing amphetamine doses as seen in Fig. 2A. This is supported by the significant dose effect, F(3,36) = 13.09; P < 0.0001. Full substitution is seen in a majority of Fischer 344 rats at the 0.3- and 1.0-mg/kg doses, whereas amphetamine substitutes fully in only three to four of seven Lewis rats at these doses as seen in Table 2. The effects of amphetamine on response rate do not differ between strains nor does it show a dose effect (P values > 0.10; Fig. 2B). Pentobarbital does not substitute for cocaine in either strain as seen in Table 2. Pentobarbital causes a dose-related disruption in response rates and this effect is greater in Fischer 344 rats compared with Lewis rats as seen in Fig. 2B. This statement is supported by the significant effects of strain, F(1,11) = 17.25; P < 0.001, and dose, F(2,11) = 15.20; P < 0.001. 

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Mean (±S.E.M.) percentage of responding on the cocaine-appropriate lever (A) and mean (±S.E.M.) response rates (number of lever presses/s; B), after administration of various doses of amphetamine (squares in A; circles in B) and pentobarbital (triangles in A; diamonds in B) are presented for Lewis (closed symbols) and Fischer 344 (open symbols) rats.

The results of the substitution tests with the partial D1 agonist SKF 38393 and the D2/D3 agonist quinpirole are shown in Fig. 3 and in Table 2. SKF 38393 does not engender significant cocaine-appropriate responding at any dose tested and does not substitute for cocaine in any rat tested. There are no strain differences in discriminative behavior with SKF 38393 as seen in Fig. 3A (P values > 0.10). SKF 38393 decreases response rates as supported by the significant effect of dose, F(2,11) = 11.12; P < 0.001. As seen in Fig. 3B, response rates are lower in Fischer 344 rats compared with Lewis rats as supported by the significant effect of strain, F(1,11) = 8.47; P < 0.01. Quinpirole has dose-related effects on cocaine-appropriate responding as supported by the significant effect of dose, F(3,42) = 59.09; P < 0.001. As seen in Table 2, quinpirole substitutes fully for cocaine for a majority of Fischer 344 and Lewis rats at the highest dose tested (1.0 mg/kg). However, response rates are disrupted at this dose. The effect of quinpirole on response rates is dose-related, as supported by the significant dose effect, F(2,13) = 15.20; P < 0.001, and differs by strain, F(1,13) = 14.58; P < 0.001. As seen in Fig. 3B, response rates are lower in Fischer 344 rats compared with Lewis rats.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Mean (±S.E.M.) percentage of responding on the cocaine-appropriate lever (A) and mean (±S.E.M.) response rates (number of lever presses/s; B), after administration of various doses of quinpirole (squares in A; circles in B) and SKF 38393 (triangles in A; diamonds in B) are presented for Lewis (closed symbols) and Fischer 344 (open symbols) rats.

The effects of the D1 antagonist SCH 23390 and the D2 antagonist eticlopride, each in combination with the training dose of cocaine (10 mg/kg), are shown in Fig. 4. Increasing doses of SCH 23390 attenuate cocaine-appropriate responding to a low of 40 ± 17% in Lewis rats and to 32 ± 16% in Fischer 344 rats, as seen in Fig. 4A. There is a significant effect of dose, F(3,30) = 7.26; P < 0.001, but no significant effect of strain (P > 0.10). Response rates are disrupted by SCH 23390 differently by strain as supported by the significant strain effect, F(1,4) = 11.19; P < 0.005, and by the trend toward significance of the dose by strain interaction, F(2,6) =2.76; P = 0.08. As seen in Fig. 4B, response rates are disrupted to a greater extent in Fischer 344 versus Lewis rats particularly at the highest dose. Eticlopride does not alter the discriminative stimulus effects of cocaine in either strain as seen in Fig. 4A (P values > 0.10). Eticlopride dose dependently disrupts response rates as supported by the significant effect of dose, F(1,12) = 11.59; P < 0.005, but this effect does not differ by strain (P values > 0.10).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Mean (±S.E.M.) percentage of responding on the cocaine-appropriate lever (A) and mean (±S.E.M.) response rates (number of lever presses/s B), after administration of SCH 23390 (squares in A; circles in B) and eticlopride (triangles in A; diamonds in B) in combination with the training dose of cocaine (10 mg/kg) are presented for Lewis (closed symbols) and Fischer 344 (open symbols) rats.

Experiment 3: Intravenous Cocaine Self-Administration

Baseline responding. The data from a total of 12 Lewis rats and 10 Fischer 344 rats are used in this study of the effects of D1- and D2-like compounds on cocaine self-administration. The baseline numbers of self-administered cocaine injections emitted by these rats are presented in Fig. 5. Increasing cocaine doses are associated with decreasing numbers of self-administered cocaine infusions in both strains as supported by the significant effects of dose, F(2,40) = 44.10; P < 0.0001. Fischer 344 rats self-administer greater numbers of cocaine infusions as supported by the significant effect of strain, F(1,20) = 10.02; P < 0.005, and seen in Fig. 5. Similar effects are seen for numbers of active lever presses (data not shown). These data also show significant effects of dose, F(2,40) = 30.68; P < 0.0001, and strain, F(1,20) = 7.51; P < 0.05. There are no significant effects of dose or strain on numbers of inactive lever presses under baseline conditions (data not shown; P values > 0.10).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Mean (±S.E.M.) numbers of self-administered cocaine infusions under baseline conditions by cocaine dose (mg/kg) are presented for Lewis (closed squares) and Fischer 344 (open squares) rats.

Eticlopride. The effects of the D2-like antagonist eticlopride (0.03 mg/kg) on numbers of self-administered cocaine injections in Lewis and Fischer 344 rats are shown in Fig. 6. Active and inactive lever press data under baseline and drug conditions for both strains are presented in Table 3. Increasing doses of cocaine are associated with decreasing numbers of self-injections and active lever presses as supported by the significant effects of dose, F(1,10) = 81.10 and 103.90; P values < 0.0001 (injections and active lever presses, respectively). Number of cocaine injections are lower in Lewis rats (Fig. 6A) compared with Fischer 344 rats (Fig. 6B) as supported by the significant strain effect, F(1,10) = 19.79; P < 0.005. Yet, no significant strain differences are seen in active or in inactive lever presses (P values > 0.10).

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Mean (±S.E.M.) numbers of self-administered cocaine infusions under baseline (open symbols) and after administration of eticlopride (0.03 mg/kg; closed symbols) by cocaine dose (mg/kg) are presented for Lewis (A) and Fischer 344 (B) rats. A second dose of eticlopride (0.1 mg/kg) was tested in Fischer 344 rats (circles) only.

SCH 23390. The effects of the D1-like antagonist SCH 23390 (0.01 mg/kg) on numbers of self-administered cocaine injections for Lewis and Fischer 344 rats are shown in Fig. 7. Active and inactive lever press data under baseline and drug conditions for both strains are presented in Table 3. Increasing doses of cocaine are associated with decreasing numbers of self-injections and active lever presses as supported by the significant effects of dose, F(2,20) = 64.91 and 49.66; P values < 0.0001 (injections and active lever presses, respectively). Numbers of cocaine injections and active lever presses tend to be lower in Lewis rats (Fig. 7A) compared with Fischer 344 rats (Fig. 7B), F(1,10) = 3.77; P = 0.08 and 3.20; P = 0.10 (injections and active lever presses, respectively). SCH 23390 shifts the cocaine dose-response curve for self-injections to the right or upwards in Fischer 344 rats (Fig. 7B) but not in Lewis rats (Fig. 7A). This is supported by the significant drug effect, F(1,10) = 13.76; P < 0.01, and the dose by drug interaction, F(2,20) = 4.84; P < 0.05. Similar effects are seen for numbers of active lever presses (Table 3), a significant drug effect, F(1,10) = 8.86; P < 0.05, and dose by drug interaction, F(2,20) = 3.40; P < 0.05. Separate analyses in Fischer 344 rats reveal significant effects of drug, F(1,10) = 11.39; P < 0.01, dose, F(2,20) = 30.45; P < 0.0001, and their interaction, F(2,20) = 3.92; P < 0.05, on number of self-injections. In Lewis rats, only the dose effects are significant, F(2,20) = 19.03 and 9.54; P values < 0.01 (injections and active lever presses, respectively). The test of the higher SCH 23390 dose (0.02 mg/kg), which was conducted with Lewis rats only, leads to similar effects as is seen with the lower SCH dose. Numbers of inactive lever presses are minimal under both baseline and drug conditions for both Lewis and Fischer 344 rats and there are no significant effects for this measure (P values > 0.10).

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Mean (±S.E.M.) numbers of self-administered cocaine infusions under baseline (open symbols) and after administration of SCH 23390 (0.01 mg/kg; closed symbols) by cocaine dose (mg/kg) are presented for Lewis (A) and Fischer 344 (B) rats. A second dose of SCH 23390 (0.02 mg/kg) was tested in Lewis rats (circles) only.

Quinpirole. The effects of the D2-like agonist quinpirole (0.01 mg/kg) on numbers of self-administered cocaine injections for Lewis and Fischer 344 rats are shown in Fig. 8. Active and inactive lever press data under baseline and drug conditions for both strains are presented in Table 3. Increasing doses of cocaine are associated with decreasing numbers of self-injections and active lever presses as supported by the significant effects of dose, F(2,18) = 195.16 and 165.06; P values < 0.0001 (injections and active lever presses, respectively). Numbers of active lever presses are lower in Lewis rats compared with Fischer 344 rats, F(1,9) = 6.10; P < 0.05. Quinpirole has differential effects on cocaine self-administration behavior in Lewis and Fischer 344 rats as seen by comparing Fig. 8, A (Lewis rats) and B (Fischer 344 rats). Quinpirole decreases responding at the lowest cocaine dose in Lewis rats with little or no effect seen at the other two doses. In Fischer 344 rats, quinpirole increases responding slightly at the lowest cocaine dose and decreases responding slightly at the middle cocaine dose. These statements are supported by the significant effect of strain (active lever presses), F(1,9) = 6.10; P < 0.05, and the significant interactions of cocaine dose by strain by drug, F(2,18) = 4.8 and 7.01 (injections and active lever presses, respectively); P values <0.05. Numbers of inactive lever presses are minimal under both baseline and drug conditions for both Lewis and Fischer 344 rats and there are no significant effects for this measure (P values > 0.10).

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Mean (±S.E.M.) numbers of self-administered cocaine infusions under baseline (open symbols) and after administration of quinpirole (0.001 mg/kg; closed symbols) by cocaine dose (mg/kg) are presented for Lewis (A) and Fischer 344 (B) rats.

SKF 38393. The effects of the D1-like agonist SKF 38393 (1 mg/kg) on numbers of self-administered cocaine injections for Lewis and Fischer 344 rats are shown in Fig. 9. Active and inactive lever press data under baseline and drug conditions for both strains are presented in Table 3. Increasing doses of cocaine are associated with decreasing numbers of self-injections and active lever presses as supported by the significant effects of dose, F(2,18) = 52.46 and 33.3; P values < 0.005 (injections and active lever presses, respectively). Numbers of self-injections and active lever presses are lower in Lewis rats compared with Fischer 344 rats, F(1,9) = 14.96 and 14.32; P values < 0.005. SKF 38393 has differential effects on cocaine self-administration behavior in Lewis and Fischer 344 rats as seen by comparing Fig. 9, A (Lewis rats) and B (Fischer 344 rats). SKF 38393 decreases responding in Lewis rats and increases responding in Fischer 344 rats particularly at the lowest cocaine dose. These statements are supported by the significant interaction of cocaine dose by strain by drug, F(2,18) = 3.33; P < 0.05, for active lever presses and the trend for significance of this interaction, F(2,18) = 3.16; P < 0.10, for number of self-injections. Numbers of inactive lever presses are minimal under both baseline and drug conditions for both Lewis and Fischer 344 rats and there are no significant effects for this measure (P values > 0.10).

View larger version (18K): [in this window] [in a new window]   Fig. 9.   Mean (±S.E.M.) numbers of self-administered cocaine infusions under baseline (open symbols) and after administration of SKF 38393 (1 mg/kg; closed symbols) by cocaine dose (mg/kg) are presented for Lewis (A) and Fischer 344 (B) rats.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

D1- and D2-like agents have differential effects on cocaine self-administration in Lewis and Fischer 344 rats. Yet, no strain differences are seen on how these agents affect cocaine discrimination and minimal differences are seen on rates of operant responding for food. The lack of strain differences in these other procedures may be due to differences in reinforcer (cocaine versus food) and/or the schedule of reinforcement. Or, the conditions under which these procedures were run may have not been optimal to see strain differences. Lower cocaine-training doses or different reinforcement schedules may uncover strain differences in future studies. Nonetheless, the results of the cocaine self-administration study are consistent with inherent differences in NAc DA receptor populations. Lewis rats have lower levels of D2 receptor (Flores et al., 1998) and of G_i in NAc compared with Fischer 344 rats (Guitart et al., 1993; Haile et al., 2001). Although no strain differences exist in NAc levels of D1 receptors (Flores et al., 1998) or in G_s (Guitart et al., 1993; Haile et al., 2001)_, D1-D2 interactions differ between strains and both of these receptor populations contribute to cocaine-induced behaviors (Walters et al., 1987; Spealman et al., 1992).

Cocaine self-administration was maintained in a dose-related manner in both strains as shown previously (Kosten et al., 1997). Similarly, cocaine exerted discriminative control in a pharmacologically specific manner in both strains. Cocaine generalized to amphetamine but not to pentobarbital as seen for outbred rats (Barrett and Appel, 1989; Witkin et al., 1991; Terry et al., 1994). There were no strain differences in acquisition of cocaine discrimination behavior in contrast to findings that Lewis rats acquired cocaine self-administration more readily than Fischer 344 rats (Kosten et al., 1997). However, this lack of strain difference may reflect limited experimental conditions (i.e., one training dose). Indeed, Fischer 344 rats showed more rapid acquisition of morphine discrimination to a low training dose (Morgan et al., 1999).

Compared with Lewis rats, Fischer 344 rats showed lower response rates in the discrimination study, slightly lower response rates in the operant responding for food study, but emitted greater numbers of lever presses during cocaine self-administration. Although increases in self-administration under the conditions of the present study may reflect decreased efficacy of the reinforcing effects of the drug, our preliminary data showed no strain differences in how threshold or peak cocaine doses maintain behavior. Finally, strain differences in cocaine pharmacokinetics can be ruled out; no differences in cocaine plasma levels occurred after i.p. (Guitart et al., 1992) or i.v. (Kosten et al., 1997) cocaine administration.

Strain differences in NAc levels of D2 receptors and G_i led us to hypothesize that D2 agents would have strain-selective effects on behaviors maintained by cocaine. Indeed, the D2 antagonist eticlopride caused a rightward, or upward, shift in the cocaine self-administration dose response function in Lewis rats in a manner consistent with pharmacological antagonism. Numbers of self-administered cocaine infusions increased under eticlopride similar to what is seen by decreasing the unit cocaine dose as shown previously in Sprague-Dawley and Wistar rats with other D2 antagonists (deWit and Wise, 1977; Roberts and Vickers, 1984; Britton et al., 1991; Caine and Koob, 1994). This eticlopride-induced shift in the cocaine dose-response function was not seen in Fischer 344 rats. The higher eticlopride dose, given to Fischer 344 rats only, suppressed responding to minimal levels. Yet, under a progressive ratio schedule, a small effect of eticlopride was reported for Fischer 344 rats (Ward et al., 1996). Similar drug doses were used in both studies, suggesting the schedule difference accounts for the discrepancy. Indeed, Caine and Koob (1994) showed that the schedule of reinforcement was an important factor in determining whether a DAergic agent altered cocaine self-administration.

In contrast to the strain-selective effects on cocaine self-administration, there were no strain differences of eticlopride on cocaine discrimination or on operant responding for food. Eticlopride did not alter cocaine-appropriate responding in either strain although a previous study reported partial blockade (Woolfolk and Holtzman, 1997). Finally, eticlopride had an inverted U-shaped dose-response function on operant responding for food in both strains, increasing rates at low doses and decreasing rates at higher doses. Decreased rates to higher doses were reported previously (Hemby et al., 1996).

The D2/D3 agonist quinpirole decreased cocaine self-administration at the lowest cocaine dose in Lewis rats and at the middle cocaine dose in Fischer 344 rats perhaps reflecting the slight difference in baseline response rates. Quinpirole-induced reductions in cocaine self-administration were reported previously (Caine and Koob, 1993). Quinpirole substituted for the cocaine cue in both strains at a dose that disrupts responding. Previous studies showed quinpirole substitutes for cocaine in monkeys (Katz and Witkin, 1992) and rats (Barrett and Appel, 1989; Terry et al., 1994) but not in all studies (Kleven et al., 1990; Witkin et al., 1991). Unlike the other compounds, quinpirole showed strain-selective effects on rates of operant responding for food. Lewis rats were more sensitive to its rate-decreasing effect.

Although there were no strain differences in NAc levels of D1 receptors or G_s, the D1-D2 dynamics is altered perhaps leading to strain-selective differences in D1 receptor stimulation. Indeed, the D1 antagonist SCH 23390 caused a rightward, or upward, shift in the cocaine self-administration dose-response function in both strains as reported previously in outbred rats (Corrigall and Coen, 1990; Caine and Koob, 1994). However, this effect was greater in Fischer 344 rats. Although this difference was striking and a higher dose of SCH 23390 was not more effective for Lewis rats, this may reflect baseline rate differences. SCH 23390 attenuated cocaine discrimination in both strains as shown previously in rats (Barrett and Appel, 1989) and monkeys (Kleven et al., 1990) at a dose that disrupted responding in Fischer 344, but not Lewis rats. For both strains, SCH 23390 had an inverted U-shaped dose on operant responding for food and decreased rates with high doses were reported previously (Corrigall and Coen, 1990).

The partial D1 agonist SKF 38393 had opposing effects on cocaine self-administration in these strains. Whereas SKF 38393 caused a leftward shift in the cocaine dose-response function in Lewis rats, consistent with an agonist-like effect, it caused a rightward shift in Fischer 344 rats, an effect associated with an antagonist. Partial agonists, such as SKF 38393, have reduced efficacy compared with full agonists acting as antagonists or agonists depending on various factors. SKF 38393 is classified as a partial agonist based, in part, on its lower capacity to stimulate adenylyl cyclase activity compared with other D1 agonists (Izenwasser and Katz, 1993). Baseline activity along this same neurochemical pathway is up-regulated in Lewis versus Fischer 344 rats (Guitart et al., 1993). In addition to intrinsic activation, conventional receptor theory suggests that agonist efficacy reflects the degree of receptor reserve and availability of exogenous ligands (Ariens, 1983). In the presence of cocaine, SKF 38393 may act as antagonist as it did on cocaine self-administration in Fischer 344 rats. Yet, SKF 38393 acted like an agonist on cocaine self-administration in Lewis rats, the strain with lower D2 receptor and G_i levels, increased adenylyl cyclase, and functionally increased D1 receptor availability in NAc. SKF 38393 does not substitute for cocaine in either strain, consistent with previous studies in monkeys (Kleven et al., 1990; Katz and Witkin, 1992) and rats (Barrett and Appel, 1989), although other studies find partial (Witkin et al., 1991) or full (Terry et al., 1994) substitution. SKF 38393 disrupted response rates in the discrimination and operant-responding procedures in both strains.

Chronic cocaine exposure may have had strain-selective effects on protein levels in neural areas subserving cocaine behaviors. However, chronic experimenter-delivered cocaine did not alter NAc G protein levels in either strain. Tyrosine hydroxylase and glial fibrillary acidic protein immunoreactivity in NAc and VTA were altered by chronic cocaine in Fischer 344 rats as were FosB levels for both strains (Haile et al., 2001). Cessation of response-dependent cocaine decreased NAc adenyl cyclase levels and D1, but not D2, receptor density in limbic areas (DeMontis et al., 1998).

The D2-like properties of cocaine are important for maintaining behavior in both Lewis and Fischer 344 rats. Quinpirole substitutes for the cocaine cue and decreases cocaine self-administration. Yet, the degree to which D1 and D2 antagonists alter behavior differs between strains. Eticlopride causes greater increases in cocaine self-administration and more readily decreases operant responding in Lewis rats. Conversely, SCH 23390 causes greater increases in cocaine self-administration and more readily decreases operant responding in Fischer 344 rats. Perhaps, D1 receptor stimulation is more important in Fischer 344 rats, whereas D2 receptor stimulation is more important in Lewis rats. Inherent differences in D1-D2 receptor dynamics in populations under study could lead to different results, even opposing effects like that seen for the D1 partial agonist in Lewis and Fischer 344 rats. Such differences have implications for treatment because like Lewis rats, many cocaine addicts have low striatal D2 levels (Volkow et al., 1997).