Intermittent and extended cocaine access attenuate cocaine-vs-food choice in male and female rats

Subjects

A total of 35 (18 Male / 17 Female) Long Evans rats weighing approximately 225–275g for females and 250–300g for males were purchased from a commercial supplier (Inotiv, Frederick, MD, USA). All subjects were single-housed in a temperature-controlled and AAALAC-accredited vivarium under a 12-h light:dark cycle with lights on from 6 AM to 6 PM. Subjects were given ad libitum access to both food (Tekland Rat Diet, Envigo) and water in their home cage and were weighed weekly, except during the supplemental access sessions when the rats were weighed daily. Animal research and maintenance were conducted following the 2011 NIH Guide for the Care and Use of Laboratory Animals, Eighth Edition. All enrichment and experimental protocols were approved by the Virginia Commonwealth University Institutional Animal Care and Use Committee (IACUC).

Catheter Implantation and Maintenance

Subjects were aseptically implanted with an indwelling intravenous (IV) femoral catheter and vascular access port as previously described [26]. Catheters were flushed daily with a 0.1 mL of cefazolin (50 mg/mL) and 0.1 mL heparin (250 U/mL) sequentially. Catheter patency was confirmed at the end of each experiment by IV methohexital (1.6 mg/0.1 mL) administration and patency was confirmed by observing instantaneous loss of muscle tone and ambulation.

Apparatus

Modular operant chambers (Med Associates, St. Albans, VT) housed in sound-attenuating chambers (Med Associates) were used as described previously [26]. Briefly, each retractable lever had tricolor LED lights (red, yellow, green) located directly above the lever. The syringe pump (PHM-100, Med Associates) was connected to a fluid swivel (375/22PS, Instech Laboratories), and the IV line was protected by a stainless steel, magnetic tether (Instech Laboratories). The operant chamber also contained a retractable dipper located between two levers with a 0.1 mL cup for liquid food delivery. (vanilla-flavored Ensure^® diluted in tap water, Abbott Laboratories, Chicago, IL). All behavioral experiments were conducted using custom programs written in Med-State Notation (Med Associates).

Cocaine-vs-Food Choice Training

Initial operant training and cocaine-vs-food choice sessions were conducted between ~11AM and ~2PM, Monday through Friday, unless otherwise noted. Subjects were initially trained to respond for 0.32 mg/kg/infusion cocaine on the right lever under a fixed-ratio (FR)1, 15-s time out schedule of reinforcement up to the terminal FR5 schedule as previously described [26]. Following three consecutive days of ≥ 20 cocaine infusions under the FR5 schedule with no upward or downward trends, rats then advanced to food-maintained responding training. Subjects were initially trained to respond for a 5-s presentation of 0.1 mL of 32% Ensure on the left lever under an FR1 schedule up to the terminal FR5 schedule. Acquisition of food-maintained responding was defined as ≥ 100 food deliveries during a 2-h behavioral session.

Subsequently, subjects were trained on the terminal cocaine-vs-food choice procedure as previously described [26]. Briefly, the choice session consisted of five sequential 20-min components with a 5-min time out preceding each component. During the timeout, subjects received a non-contingent infusion of the cocaine dose available followed by a non-contingent food presentation, both of which served as discriminative stimuli for the reinforcer magnitudes available during that component. The response period of each component then began with insertion of both levers into the chamber and illumination of the stimulus lights above each lever. Increasing cocaine doses (0, 0.032, 0.1, 0.32, 1.0 mg/kg/infusion) were available as the alternative to food during successive components under a concurrent FR5:FR5 schedule of reinforcement. Responding on any one lever reset the ratio requirement of the other lever. During each component, subjects could complete up to 10 total ratio requirements across both levers. At the end of each component, the levers retracted and stimulus lights were turned off. Cocaine choice was considered stable when the smallest cocaine dose that maintained ≥80% cocaine choice did not vary more than 0.5 log units over three consecutive days.

Supplemental Reinforcer Access Manipulations

Subjects were randomized into one of three supplemental cocaine access groups: no cocaine access (NA =11, 6M/5F), intermittent cocaine access (IntA =12, 6M/6F), and extended cocaine access (EA =12, 6M/6F). The four-week experimental procedure consisted of one baseline week with choice sessions only followed by two supplemental access weeks that included both choice sessions + the designated supplemental access conditions from 6PM to 6AM daily and finally one post-access week of choice sessions only (Figure 1). During the two supplemental access weeks, NA subjects were left in the vivarium during the supplemental cocaine access period. Subjects in the IntA condition could self-administer 0.32 mg/kg/infusion cocaine under an FR5/no timeout schedule for 5 min at the beginning of each 30 min interval (5 min: ON/ 25 min: OFF) during the 12h supplemental access period (Figure 1B). Subjects in the EA condition could self-administer 0.32 mg/kg/infusion cocaine under an FR5/10s timeout schedule during the 12h supplemental access period (Figure 1C). Choice sessions were instituted from ~11 AM – 1 PM Monday through Friday throughout the four-week experiment (Figure 1). Subjects were placed back into their home cage and returned to the vivarium following both supplemental cocaine access and cocaine choice sessions.

Following the supplemental cocaine access studies, subjects were quasi-randomized into one of three supplemental food access experimental groups (NA: 4M/ 4F, IntA: 4M/ 4F, and EA: 5M/ 3F) and counterbalanced based on cocaine-access history. The four-week experimental timeline and daily food access conditions were the same as described above and in Figure 1 with the exception that 0.1 mL of 32% Ensure was available rather than cocaine during the two-week supplemental access period. To conduct the supplemental food access experiments, we created a custom Med-PC program to automatically refill the liquid reservoir with 0.1 mL diluted Ensure after each response requirement completion using the syringe pump and a 50 mL syringe containing 32% Ensure and tubing. We calculated the liquid food reservoir capacity was ~80 mL and combined with the 50 mL syringe provided a total Ensure volume of 130 mL. However, we also determined the bottom 10 mL of the liquid reservoir was too shallow to deliver a full 0.1 mL, meaning that the total functional volume was ~120 mL or ~1200 food reinforcers during the supplemental food access session. During these experiments, no subject earned more than 1200 food reinforcers during the 12-h supplemental food access session.

Data Analysis

The primary dependent measure for the supplemental cocaine- or food-access sessions was the number of cocaine or food reinforcers earned during each 12-h supplemental access session. These measures were plotted as a function of access condition day. Because the IntA condition provided a total of 2h of cocaine availability each day (i.e. each 12-h supplemental IntA session contained 24 intervals of 5 min cocaine availability during each 30-min interval), cocaine intake during the first and last 2h of EA supplemental access were also determined for comparison. The primary dependent measure for the cocaine choice sessions was percent cocaine choice, defined as {(cocaine reinforcers earned ÷ total number of reinforcers earned) * 100}. This measure was plotted as a function of the unit cocaine dose. Additional dependent measures were the numbers of reinforcers (food, cocaine, and total) earned across all components during the choice session. Cocaine choice data represents the average from the last two days of each experimental week.

Data were analyzed using either a one-way, two-way, or Brown-Forsythe repeated-measures ANOVA, mixed-effect analysis, or Welch’s t-test as appropriate using GraphPad Prism (v11.0.2 for macOS, La Jolla, CA). The main factors were unit cocaine dose and access condition. Post-hoc comparisons following a significant treatment effect and/or interaction were performed using a Tukey or Šidák post-hoc test. The Geisser-Greenhouse correction was used to correct for any sphericity violations. Statistical significance for all experiments was established a priori at the 95% confidence level (p < 0.05).

Drugs

Cocaine HCl was provided by the National Institute on Drug Abuse Drug Supply Program (Bethesda, MD). Methohexital HCl was purchased from a commercial vendor (PAR Pharmaceutical, Chestnut Ridge, NY). All solutions were dissolved in bacteriostatic saline, and IV solutions were passed through a 0.22 μm sterile filter before use. Cocaine doses were expressed as the salt form listed above.

Supplemental Cocaine and Food Access

Figure 1E–H shows the temporal response patterns during IntA and EA cocaine and food supplemental access sessions. For supplemental cocaine access, there was a significant effect of day and access condition, but no significant day × access condition interaction (Day: F_{2.4, 49.9} = 4.40, p = 0.01; Access Condition: F_{1, 22} = 14.48, p = 0.001; Interaction: F_{2.4, 49.9} = 1.13, p = 0.34); Because there was a main effect of access condition but no significant interaction, we performed a t-test to compare access conditions, and a greater number of cocaine infusions were earned each day under EA compared to IntA (t= 7.42, df = 13.13, p < 0.0001; Figure 1E). Total cocaine intake over the 10 supplemental access sessions was greater under EA compared to IntA (t= 3.74, df = 18.49, p = 0.001; Supplemental Figure 1A). However, IntA resulted in significantly more cocaine infusions compared to the first or last 2h of EA during week 1 (F_{2.0, 12.5} = 31.04, p <0.0001) and week 2 (F_{2.0, 27} = 24.82, p < 0.0001) (Figure 1G). Therefore, IntA resulted in less overall cocaine intake than EA, but higher rates of cocaine self-administration when cocaine was available. There were no significant sex differences across supplemental cocaine access conditions (Supplemental Figure 2).

Unlike supplemental cocaine access, there was no significant effect of supplemental food access condition on food reinforcers per day (Day: F_{2.9, 20.3} = 9.6, p = 0.0004; Access Condition: F_{1, 14} = 1.13, p = 0.31; Interaction: F_{1.5, 19.8} = 0.38, p = 0.63; Figure 1F). There was also no significant difference between IntA and EA conditions in total food intake over the 10 supplemental food access sessions (t= 1.07, df = 14.0, p = 0.30; Supplemental Figure 1B). IntA food resulted in more food reinforcers compared to either the first or last 2h of EA food per day during week 1 (F_{2, 21} = 5.10, p < 0.0001) and week 2 (F_{2, 21} = 1.31, p < 0.0001) of supplemental food access (Figure 1H). Therefore, IntA and EA access conditions resulted in similar overall food intake, but rates of food-maintained responding during periods of food availability were higher under IntA conditions. There were sex differences during supplemental food access, such that food-maintained responding was higher in females compared to males for both IntA food (t = 4.40 df = 18, p = 0.0003, Supplemental Figure 3A) and EA food (t = 5.52 df = 18, p < 0.0001, Supplemental Figure 3B). Additionally, females also earned more food reinforcers than males during the first week of IntA food (Sex: F_{5, 18} = 30.3, p < 0.0001; Supplemental Figure 3C).

Cocaine Choice During Differential Supplemental Cocaine Access

Figure 2 shows the effect of supplemental cocaine access on cocaine choice and session total, food, and cocaine reinforcers. During week 1 of supplemental cocaine access, there was a significant cocaine dose × access condition interaction on cocaine choice (Cocaine Dose: F_{1.5, 45.4} = 64.7, p < 0.0001; Access Condition: F_{2, 31} = 25.63, p <0.0001; Interaction: F_{3.1, 45.4} = 7.9, p = 0.0001; Figure 2A); post-hoc analysis showed that both IntA and EA cocaine attenuated 0.32 mg/kg/infusion choice (p < 0.0001 for both access conditions) and 1.0 mg/kg/infusion (IntA: p = 0.04; EA: p = 0.03) choice compared to the NA group. There was also a significant access condition effect on choice session reinforcers (Reinforcer: F_{1.2, 36.6} = 190.6, p < 0.0001; Access Condition: F_{2, 31} = 0.61, p = 0.55; Interaction: F_{2.4, 36.6} = 10.65, p = 0.0001; Figure 2B); post-hoc analysis showed that IntA cocaine increased session food reinforcers (p = 0.03) and both IntA (p = 0.0004) and EA (p = 0.0002) cocaine decreased session cocaine reinforcers compared to NA.

During week 2 of supplemental cocaine access, there was a significant cocaine dose × access condition interaction on cocaine choice (Cocaine Dose: F_{1.6, 47.8} = 83.76, p < 0.0001; Access Condition: F_{2, 31} = 49.56, p < 0.0001; Interaction: F_{3.3, 47.8} = 19.00, p < 0.0001; Figure 2C); post-hoc analysis showed that both IntA and EA cocaine attenuated 0.32 mg/kg/infusion cocaine choice (p < 0.0001 and EA cocaine also attenuated 1.0 mg/kg/infusion (p < 0.0001) cocaine choice compared to NA. Additionally, 1.0 mg/kg/infusion cocaine choice was significantly lower during EA cocaine compared to IntA cocaine (p = 0.004). There was also a significant effect of supplemental cocaine access condition on choice session reinforcers (Reinforcer: F_{1.3, 38.7} = 289.2, p < 0.0001; Access Condition: F_{2, 31} = 0.41, p = 0.66; Interaction: F_{2.5, 38.7} = 12.93, p < 0.0001; Figure 2D); post-hoc analysis showed that EA cocaine increased session food reinforcers (p = 0.02) and both IntA (p = 0.0007) and EA (p < 0.0001) cocaine reduced session cocaine reinforcers compared to NA. In addition, session cocaine reinforcers were smaller for EA compared to IntA cocaine (p = 0.004).

There were no significant effects of cocaine-access condition on cocaine choice (Figure 2E) or session total, food, and cocaine reinforcers (Figure 2F) up to seven days after termination of supplemental cocaine access. Moreover, there were no significant sex differences on cocaine choice up to seven days post supplemental cocaine access (Supplemental Figure 4).

Cocaine Choice During Differential Supplemental Food Access

Figure 3 shows supplemental food access effects on cocaine choice and session total, food, and cocaine reinforcers. During week 1, there was a significant supplemental food access effect on cocaine choice (Cocaine Dose: F_{2.2, 45.7} = 77.9, p < 0.0001; Access Condition: F_{2, 21} = 6.5, p = 0.006; Interaction: F_{4.4, 45.7} = 1.6, p = 0.19; Figure 3A); however, due to the absence of a significant cocaine dose × access condition interaction no post-hoc analyses were performed. There was also a significant supplemental food access effect on choice session reinforcers (Reinforcer: F_{1.2, 25.6} = 78.3, p <0.0001; Access Condition: F_{2, 21} = 1.1, p = 0.35; Interaction: F_{2.4, 25.6} = 5.5, p = 0.007; Figure 3B); post-hoc analysis showed that IntA food decreased session food reinforcers (p = 0.04) and increased session cocaine reinforcers (p = 0.03) compared to NA food. There was no significant sex difference in cocaine choice during week 1 of supplemental food access (Supplemental Figure 5A–C)

Supplemental food access had no significant effect on cocaine choice during week 2 (Figure 3C). However, 0.1 mg/kg/infusion cocaine choice was higher in males compared to females in the IntA group (Cocaine Dose: F_{1.8, 11.1} = 36.03, p < 0.0001; Sex: F_1,6 = 4.56, p = 0.02; Interaction: F_{1.8, 11.1} = 4.96, p = 0.03; Supplemental Figure 5D). Although there was a significant access condition × reinforcer interaction for choice session reinforcers during week 2 (Reinforcer: F_{1.3, 26.4} = 55.75, p<0.0001; Access Condition: F_{2, 21} = 3.25, p = 0.06; Interaction: F_{2.5, 26.4} = 3.59, p = 0.03; Figure 3D), post hoc analyses correcting for multiple comparisons did not detect any significant differences.

Lastly, there were no significant effects of supplemental food access on either cocaine choice (Figure 3E) or session reinforcers (Figure 3F) up to seven days after termination of supplemental food access. There were also no significant sex differences in cocaine choice up to seven days post supplemental food access (Supplemental Figure 5G–I).

Cocaine- and food-maintained responding under IntA and EA conditions

Consistent with the extant literature, IntA cocaine resulted in less overall cocaine intake compared to EA cocaine [17,19,27,28]. Additionally, rates of cocaine self-administration were higher during IntA cocaine compared to EA cocaine, which was also consistent with the literature [19,27,29–32]. We did not observe “escalation” defined as a progressive increase in cocaine self-administration rates over time during either IntA or EA cocaine supplemental sessions in the present study, which was inconsistent with a subset of previous IntA [33,34] and EA [16,35–37] rat cocaine self-administration studies. However, this “escalation” behavioral pattern has been inconsistently reported in the literature for both IntA [32,38,39] and EA [40–43] cocaine self-administration in rats and rhesus monkeys [14,44,45]. Regardless of whether “escalation” of cocaine self-administration rates is observed, supplemental cocaine access conditions can significantly impact behavioral measures [34,40,42,46], including cocaine-vs-food choice (present study). Therefore, “escalation” of drug intake may be unnecessary for investigating the behavioral consequences of differential drug access conditions effects.

Similar to IntA cocaine, IntA food also resulted in higher-rates of food-maintained responding compared to EA food. The present IntA food results were consistent with previous IntA saccharin [47] and IntA sucrose [48] studies reporting that IntA resulted in higher response rates than EA. The general congruence of the present cocaine and food self-administration patterns during both IntA and EA with the literature provides an empirical foundation to interpret how different cocaine access conditions and associated temporal patterns of cocaine intake impact cocaine-vs-food choice.

IntA and EA cocaine attenuated cocaine choice

Consistent with previous cocaine choice studies in rats [49–51] and nonhuman primates [52,53], cocaine-maintained a dose-dependent increase in percent cocaine choice under baseline conditions. This cocaine dose-dependent increase in cocaine choice was primarily due to cocaine-associated reinforcement effects rather than reinforcement-independent cocaine effects on food-maintained responding [49–51]. Unexpectedly, both IntA and EA cocaine attenuated cocaine choice when assessed approximately 5h after the IntA/EA cocaine session. Specifically, the present results were inconsistent with previously published IntA and EA cocaine effects demonstrating increased cocaine PR breakpoints and BE demand (Supplemental Tables 1 and 2). However, the present cocaine results were consistent with IntA and EA methamphetamine effects on methamphetamine-vs-social choice [21] and with IntA and EA cocaine effects on cocaine-vs-food choice [41,54,55] such that neither IntA nor EA drug increased drug choice in any experiment. Moreover, the present results are consistent with IntA ethanol effects on ethanol-vs-sucrose choice [56]. Thus, the present results were consistent with the extant literature highlighting the conclusion that preclinical drug-choice metrics of reinforcing efficacy provide distinct information on addictive drug reinforcement processes that would not be predicted based on single-operant drug self-administration procedures [57–59]. There are at least three potential reasons for the differential IntA cocaine effects on cocaine choice compared to previous IntA cocaine effects on PR breakpoints and BE demand.

First, the differential IntA and EA cocaine effects between the present cocaine choice and previous PR and within-session BE demand procedures may be related to the self-administration procedure used to assess reinforcing efficacy or strength. All preclinical drug self-administration procedures are influenced by both the positive reinforcing effects and the intake-limiting effects (e.g. motor or cognitive impairment) of the self-administered drug [23,60,61]. The present study used a cocaine-vs-food choice procedure to minimize the influence of intake-limiting effects by self-administered cocaine on the primary dependent measure of cocaine reinforcement [25,60,62]. In contrast, metrics of cocaine reinforcement in PR and BE procedures are more sensitive to intake-limiting effects and could increase because supplemental access produced tolerance to intake-limiting effects rather than sensitization to reinforcing effects [24]. Additionally, within-session BE demand procedures are influenced by discriminative stimulus effects produced by large cocaine doses earned during early components of the within-session demand procedure [24]. Because IntA cocaine sensitizes discriminative stimuli associated with cocaine self-administration [29] and EA cocaine produces tolerance to discriminative stimuli associated with cocaine self-administration [63], the effects of IntA and EA cocaine on discriminative stimulus versus reinforcing effects would be difficult to disentangle under the within-session BE demand procedure. Overall, the present results do not support the conclusion that IntA or EA cocaine increased measures of cocaine reinforcing efficacy in the context of a cocaine-vs-food choice procedure that minimized the influence of these reinforcement-independent intake-altering effects.

Second, the differential results may be related to the length of the supplemental cocaine-access session. Most IntA studies using the 5-min ON; 25-min OFF procedure had IntA session durations of 2–6 h, resulting in a total of 20–60 min of cocaine availability (Supplemental Table 1). The present study used a 12h IntA session resulting in 120 min of cocaine availability over 12h. We elected to use a 12h IntA session to match the total time of the EA sessions and allow for 2h of cocaine self-administration, which also matched the cocaine-vs-food choice session length. Although shorter IntA session durations may sensitize cocaine PR breakpoints and BE demand, longer IntA sessions may not, and future studies could manipulate IntA session length as an independent variable as has been done with EA cocaine session duration [35]. Previous studies perhaps used shorter sessions conducted entirely during the light cycle, whereas the sessions used here were longer and allowed SA during both segments of the light/dark cycle which may have produced more robust changes in diurnal rhythm as previously reported in both rats [64–66] and monkeys [13,14]. Additionally, IntA cocaine studies that have used longer session durations, but instituted other procedural variables such as discrete trials or time outs to program intermittency, have generally reported similar results to the present study. For example, a 24h discrete-trial IntA procedure, which allowed rats to self-administer cocaine every 12 min for a maximum of five injections per hour (a so-called DT 5 schedule), decreased cocaine PR breakpoints [67]. Other 21–24h discrete-trial IntA procedures reported in Supplemental Table 1 did not significantly alter cocaine PR breakpoints in rats [68,69] and either had no effect [44,70], increased [71], or decreased [70,72] cocaine-vs-food choice in rhesus monkeys.

Lastly, the interval between supplemental cocaine access and assessment of cocaine reinforcing efficacy may have influenced the present findings. In contrast to the present study, most IntA and EA cocaine self-administration studies referenced in Supplemental Tables 1 and 2 assessed cocaine breakpoints and BE demand at least 24h after the last IntA/EA cocaine self-administration session. The interval between supplemental cocaine access and subsequent assessment of cocaine reinforcement imposes a period of drug abstinence, and previous studies suggest that abstinence duration may influence both drug [73–75] and nondrug self-administration [76,77], although see [44,78,79]. Accordingly, future studies could explicitly manipulate this interval and the associated degree of cocaine abstinence when cocaine and nondrug reinforcement is evaluated.

IntA food increased cocaine choice

Similar to IntA cocaine attenuation of cocaine choice, IntA food attenuated the relative reinforcing efficacy of food resulting in increased cocaine choice. Consistent with the present study, 3h supplemental saccharin access increased cocaine-vs-saccharin choice in rats [80]. Moreover, previous IntA saccharin and sucrose studies in rats reported increased PR breakpoints following IntA sessions similar to what has been reported for cocaine [47,48]. Thus, just like IntA cocaine, effects of IntA food on metrics of food reinforcing efficacy under single-operant self-administration procedures are not predictive of effects under a concurrent “choice” procedure, such that IntA food effects on PR breakpoints were not predictive of effects on choice. Overall, the present results and the literature suggest that both IntA and EA effects on measures of reinforcer efficacy are not selective for addictive drugs and generalize across different positive reinforcers [36,81].

Role of reinforcer economy in cocaine-vs-food choice

The similar effects of IntA and EA cocaine on cocaine choice and reciprocal effects of IntA food on cocaine choice support the hypothesis that the openness of the reinforcer economy and the associated daily amounts of cocaine or food consumed influenced cocaine-vs-food choice more so than the temporal patterns of operant responding during the supplemental sessions. In behavioral economic terms, the economy for a given reinforcer is considered to be “open” if that reinforcer is available outside an experimental session or “closed” if the reinforcer is available only during the experimental session [82–84]. Supplemental IntA or EA sessions open the economy relative to the choice session used to assess relatively reinforcing efficacy, and the present results suggest that opening the economy for a reinforcer reduces the relative reinforcing efficacy of that reinforcer during a choice session. Fully opening the cocaine economy resulted in deleterious health effects in both nonhuman primates [13–15] and rats [85]. Moreover, the present IntA food effects on cocaine choice in rats were consistent with food prefeeding effects on cocaine-vs-food choice in nonhuman primates [52,53]. Overall, the present results are consistent with contemporary addiction theories and provide further evidence that behavioral allocation between an addictive drug and nondrug reinforcers are environmental and context dependent [3,59,86].