Social context in renewal of cocaine and fentanyl seeking: role of peer presence versus peer absence

Abstract

Rationale:

Drug relapse among individuals with substance use disorders can be triggered by social cues that were present in the original drug-taking environment. Previous preclinical work demonstrated this process using the ABA renewal design in which rats were first trained to lever press for contingent cocaine with a peer (context A), then the response was extinguished without the peer (context B), and finally responding was shown to return or renew with re-presentation of the peer (context A).

Objective:

This study sought to determine if the presence and/or absence of a peer would differentially elicit cocaine and fentanyl seeking.

Methods:

In 4 separate experiments, rats were tested for either ABA or BAB renewal of either cocaine seeking (Experiments 1A and 1B) or fentanyl seeking (Experiments 2A and 2B). In addition, in the fentanyl experiments, naltrexone (0.1–1 mg/kg) was administered prior to both ABA and BAB renewal tests to determine if renewal of fentanyl seeking was altered. An additional experiment (Experiment 2C) tested the effect of naltrexone (0.1–1 mg/kg) on locomotor activity.

Results:

Results showed that cocaine seeking was obtained using either ABA or BAB renewal designs, whereas fentanyl seeking was obtained only using the ABA renewal design. In addition, ABA renewal of fentanyl seeking was decreased by naltrexone at a dose (1 mg/kg) that had no effect on locomotion.

Conclusions:

These results indicate that peer presence can trigger both cocaine and fentanyl seeking, whereas peer absence specifically triggers cocaine seeking. ABA renewal of fentanyl seeking was decreased by naltrexone, indicating that opiate receptors are involved.

INTRODUCTION

Cocaine and opioid use disorders are major public health concerns, with both disorders having high relapse rates following treatment. For cocaine, the relapse rate is high even among patients admitted to long-term residential treatment programs, with approximately one-half of patients returning to cocaine use within one year after treatment (Simpson et al. 2002; Sinha 2013). For fentanyl, 90% of patients relapse to opioid use, even with intensive non-pharmacological treatment. With medication, about 40–50% of patients relapse to opioid use, and rates revert up to 90% when medication is discontinued (Rhodes 2017). Although medication can be important for the treatment of substance use disorders, psychosocial factors play a role in recovery success as well, including marital status and employment history (Kassani et al. 2015). In addition, patients who stay drug free tend to have larger and more reciprocal prosocial networks compared to patients prone to relapse (Panebianco et al. 2016) and when an individual relapses to drug use, it can cause others within the peer network to relapse as well (Karakos 2014). Moreover, reassociation with drug-using peers after a period of abstinence in an in-patient rehabilitation facility is a self-reported trigger for relapse (Brewer et al. 1998; Sun 2007).

Preclinical research has also shown that drug-associated social peers can serve as a potent trigger for relapse (Meng et al. 2021; Weiss et al. 2018). In a recent study conducted in our laboratory, we used the ABA renewal paradigm to show that a peer can serve as a trigger for cocaine seeking in rats (Humburg and Bardo 2025). In ABA renewal, also known as context-induced reinstatement (Crombag et al. 2008), acquisition of an operant response is first established in one distinct context (context A) and then is extinguished in a different context (context B). When placed back into context A, the operant response is increased or renewed (Sweeney and Shahan 2015). In the case of Humburg and Bardo (2025), context A consisted of a compound stimulus (peer and house light illumination) and context B consisted of no peer or house light illumination. When tested with each element alone (peer or house light) and as a compound (peer + house light), rats renewed responding to the peer alone and peer + house light compound, but not to the house light alone. A follow-up experiment demonstrated that ABA renewal was obtained when the house light alone (no peer) served as context A, demonstrating that the house light was sufficiently salient to control responding. These results indicate that a social peer is a potent stimulus that can completely overshadow a house light to trigger ABA renewal of cocaine seeking. However, it remains to be determined if this preclinical finding is specific to cocaine or whether it generalizes to other drugs such as fentanyl.

While the results of Humburg and Bardo (2025) demonstrate that a drug-associated peer can trigger cocaine seeking, another translationally relevant question is to determine if a peer that is associated with drug abstinence (extinction) protects against relapse and thus, may also serve as a trigger for drug seeking when it is removed. To do this in a preclinical model, one can use a BAB renewal paradigm in which drug self-administration is acquired in the absence of a peer (context B), extinguished in the presence of a peer (context A) and then tested by removing the peer (context B). Like ABA renewal, Browning and Shahan (2018) obtained BAB renewal of sucrose seeking using a peer stimulus as the context. However, as noted in that report, a social peer is not merely a neutral stimulus, as it also has incentive motivational value in the form of social reward (Calcagnetti and Schechter 1992; Evans et al. 1994; Humphreys and Einon 1981; Venniro and Shaham 2020). This has important implications for interpreting results from the BAB design because the results could also be interpreted as resurgence rather than renewal; i.e., resurgence is the return of a response following removal of an alternative reinforcer (Winterbauer and Bouton 2010). Hence, to better understand the stimulus versus rewarding effects of a social context, the present study tested both ABA and BAB renewal of drug seeking (cocaine or fentanyl) using a social peer (presence or absence) as the context.

A secondary purpose of these experiments was to determine if a clinically relevant medication would block peer-induced renewal of drug seeking. Since cocaine use disorder does not have any FDA-approved medication treatment options, we tested naltrexone against peer-induced renewal of fentanyl seeking, as naltrexone is effective for treatment of opioid use disorder (Bart 2012). Varying doses of naltrexone were administered prior to tests sessions using either ABA or BAB renewal of fentanyl seeking.

METHODS

Animals

A total of 69 Sprague Dawley rats at 50–55 days of age were purchased from Inotiv, Inc (West Lafayette, IN). Forty-five rats (18 male, 27 female) were used for drug self-administration testing and 24 were used as passive (non-self-administering) peers. Peers were randomly selected upon arrival in the laboratory and were naïve to any experimental treatment but were handled daily throughout the study. Rats were housed individually in a temperature-controlled colony room on a 12:12 h light schedule (lights on at 0800 h) with testing occurring during the light phase. Rats had ad libitum access to food and water in their home cage, except where noted. All procedures were in accordance with the “Guide for the Care and Use of Laboratory Animals” [National Research Council (U.S.), 2011] and were approved by the Institutional Animal Care and Use Committee at the University of Kentucky.

Apparatus

Pre-training with food reinforcement occurred in a single operant conditioning chamber (28 × 24 × 21 cm; ENV-008CT; Med Associates, S. Albans, VT, USA) containing two retractable levers on the same wall. There was a food tray centered between the two levers into which food pellets could be dispensed from a food hopper. A cue light was located above each lever and a house light was mounted over the center of the opposite wall.

Acquisition of cocaine and fentanyl self-administration occurred in a custom-built apparatus consisting of two operant chambers joined together and separated by a wire screen partition (1.27 cm, 19 gauge) which allowed limited interaction with a peer (see Humburg and Bardo, 2025). One side of the chamber was used as the self-administration chamber and the adjacent chamber was used for presentation of a social peer. Each chamber contained two retractable levers, with a 28-V white cue light located 6 cm above each lever, that were placed on opposite walls of the chamber closest to the wire screen partition. A white house light was present at the top of each chamber. In the social peer chamber, the levers were always retracted. The apparatus was located within a sound attenuating cabinet.

Locomotor activity was recorded automatically using an animal activity monitoring system with Versamax System software (AccuScan Instruments Inc., Columbus, OH). Rats were placed in monitoring chambers (42 × 42 × 30 cm) made of clear acrylic walls and floor. Each chamber incorporated a horizontal 16 × 16 grid of photo beam sensors, with each beam 2.5 cm apart and 7 cm above the chamber floor. Horizontal activity was expressed as distance traveled (cm).

Food reinforcement pre-training

Rats were handled for 3–5 days after arrival to the laboratory and, on the last day of handling, rats were moved to food restriction (5–15 g/day). The next day, rats began food-reinforced pretraining to press a lever in the single operant conditioning chamber that was distinctly different from the dual-compartment chamber. Presses on the active lever resulted in a food pellet (Bio-Serve Dustless Precision pellet 45mg, Product # F0021) delivery on a fixed-ratio (FR) 1 schedule of reinforcement. There was no house light, cue light, or time out period present during these sessions. Rats were trained to lever press for food pellets until they earned at least 50 pellets during a 1-hr session, a criterion that was obtained within 3–5 sessions for all rats. Following food-reinforced pretraining to press a lever, rats were taken off food restriction and allowed ad libitum access to food in the home cage for the remainder of the experiment.

Surgery

After food reinforcement pretraining, rats underwent surgery to insert an indwelling jugular catheter. Rats were anesthetized with a ketamine (75 mg/kg) + xylazine (7.5 mg/kg) mixture (males also received acepromazine; 0.75 mg/kg) and received a chronically indwelling jugular catheter that exited from an acrylic head mount attached to the skull with jeweler screws. One day prior to surgery, and two days after surgery, rats were treated with carprofen (5.56 mg/kg) to manage pain. Following catheter surgery, rats were allowed to recover for 5–7 days and then began acquisition of drug self-administration.

General Methods

For each daily drug self-administration session (1-hr), indwelling catheters were flushed with saline, then the rat was hooked up to an infusion pump tether and was placed into the chamber where they were trained to self-administer either cocaine or fentanyl in separate experiments. When the active lever was pressed (lever position counterbalanced across rats), an infusion of cocaine (0.75 mg/kg/infusion; Experiments 1A and 1B) or fentanyl (2.5 μg/kg/infusion; Experiments 2A and 2B) was delivered in a volume of 0.1 ml over 5.9 sec. Concomitant with the infusion, illumination of the cue light above the active lever signaled a time out (TO) period in which pressing the active lever had no programmed consequence. Pressing on the inactive lever at any point during the session was recorded but did not result in any programmed consequence. House lights on both sides of the chamber were illuminated to signal the start of the session and they stayed illuminated continuously until the end of the 1-hr session.

For acquisition, rats learned to self-administer cocaine or fentanyl on an FR1 schedule. The response requirement was then increased to FR3 for 7 sessions, followed by terminal FR5 for 7 more sessions. The program that controlled each session was started immediately after the self-administering rat was hooked to the infusion tether and the doors of the sound attenuating cabinet were closed. Each self-administration session was signaled by both house lights illuminating and extension of both the active and inactive levers into the self-administration chamber. The presence or absence of the peer depended on the renewal design. In the ABA experiments, rats had a peer in the adjacent chamber; in the BAB experiments, rats did not have a peer in the adjacent chamber. In all cases, the peer was the same sex and age as the self-administering rat. At the end of the session, the house lights turned off and the levers were retracted. Rats then had their catheters flushed with a mixture of saline, heparin, and gentamicin and were put back into their home cage.

For extinction, after the last self-administration session (session 22), rats were placed into extinction training. Daily extinction sessions were like FR5 acquisition sessions, except that the drug was removed. In addition, the presence or absence of the peer depended on the renewal design. In the ABA experiments, rats did not have a peer in the adjacent chamber during extinction; in the BAB experiments, rats had a peer in the adjacent chamber during extinction.

Renewal testing began the day after the last extinction session. For ABA groups, rats were returned to context A; for BAB groups, rats were returned to context B; and for the AAA group in Experiment 1A (see below), rats simply received continued exposure to context A. As during extinction, the signaled TO period was present on an FR5 schedule without drug delivery.

Specific Experiments

Experiment 1A: ABA renewal of cocaine seeking

Acquisition:

On the first 1-hr acquisition session, rats did not have a peer present; this procedure helped ensure that the rats would learn to self-administer cocaine without distraction from the peer on the first day. The next 7 daily sessions were like the first session; however, a peer was placed in the adjacent chamber before the self-administering rat was placed in the chamber (context A) and illumination of both house lights signaled the session. The same peer was used throughout acquisition training.

Extinction:

Following acquisition with the peer (context A), rats were randomly assigned into 2 equal-sized groups: (1) AAA (control) and (2) ABA (renewal). During extinction, the AAA group had the same context as acquisition (i.e., same peer). In contrast, the ABA group was extinguished without the peer (context B). In both groups, illumination of both house lights continued to signal the session. Rats received 10 daily extinction sessions.

Renewal Tests:

Rats in the ABA group were tested with peer present (context A) and peer absent (context B), in counterbalanced order; intervening between the 2 tests, rats were given 3 additional extinction sessions in context B. Across these same 5 sessions, rats in the AAA group were simply tested each day with the peer present (context A). As during acquisition and extinction, illumination of both house lights signaled the session.

Experiment 1B: BAB renewal of cocaine seeking

Acquisition:

To match the timeline of Experiment 1A, rats self-administered cocaine for 22 daily sessions without a peer present (context B) in the adjacent chamber.

Extinction:

All rats were assigned to a BAB group in which a peer was placed into the adjacent chamber before the self-administering rat was placed into the chamber (context A); there was no BBB group. The self-administering rat had the same peer present across all 10 extinction sessions.

Renewal Tests:

All rats were tested under both renewal (context B) and extinction (context A) conditions. Intervening between the 2 tests, rats were given three additional extinction sessions in context A.

Experiment 2A: ABA renewal of fentanyl seeking and effect of naltrexone

Acquisition:

The procedures were like Experiment 1A, except rats self-administered fentanyl instead of cocaine and there was only an ABA group.

Extinction:

Procedures were like Experiment 1A, except all rats had the peer removed during extinction and the number of extinction sessions increased from 10 to 14.

Renewal Tests:

All rats were tested for renewal with the peer (context A) 5 min following an injection of naltrexone (0.1, 0.3, or 1 mg/kg, s.c.) or saline, with dose order counterbalanced across rats. The same peer from acquisition was placed in the adjacent chamber prior to the self-administering rat for each of the four tests. Intervening between the 4 tests, rats were given three additional extinction sessions in context B.

Experiment 2B: BAB renewal of fentanyl seeking and effect of naltrexone

Acquisition:

Procedures were like Experiment 1B, except rats self-administered fentanyl instead of cocaine.

Extinction:

Procedures were like Experiment 1B, except the number of extinction sessions increased from 10 to 14.

Renewal Tests:

All rats were tested for renewal with the peer absent (context B) 5 min following an injection of naltrexone (0.1, 0.3, or 1 mg/kg, s.c.) or saline, with dose order counterbalanced across rats. Intervening between the 4 tests, rats were given 3 additional extinction sessions in context A.

Experiment 2C: Locomotor effect of naltrexone

To determine if the doses of naltrexone used in Experiments 2A and 2B had any nonspecific suppressant effect on behavior, a follow-up experiment was conducted using the 6 passive peers (3 male and 3 female) from Experiment 2A. Rats were first allowed to habituate to the locomotor chamber for one 1-hr session where no injections were given. Then, on four subsequent days, rats were injected with naltrexone (0.1, 0.3, or 1 mg/kg, s.c.) or saline and, 5 min later, were placed into the locomotor apparatus for 1 hr; doses were given in counterbalanced order across rats. In between each locomotor test session, rats were rested in the home cage without any injection or exposure to the locomotor apparatus.

Drugs

Cocaine HCl and fentanyl HCl were acquired from the National Institute on Drug Abuse (NIDA Drug Supply Program, Rockville, MD) and dissolved in 0.9% bacteriostatic saline to make the doses used for self-administration (cocaine unit dose of 0.75 mg/kg and fentanyl unit dose of 2.5 μg/kg; 0.1 ml infusion volume for both cocaine and fentanyl). Naltrexone HCl (RTI International, Durham, NC) was dissolved in 0.9% bacteriostatic saline to prepare the three treatment doses (0.1, 0.3, and 1.0 mg/kg, s.c.). All drug doses were expressed as salt weight.

RESULTS

Experiment 1A: ABA renewal of cocaine seeking

Results from the acquisition, extinction, and renewal phases for AAA and ABA groups are illustrated in Figure 1. For acquisition, with the data collapsed across group and sessions within each FR requirement, a 2-way ANOVA revealed significant main effects of lever (F(1,42) =89.09, p < 0.0001) and FR (2,42) =10.59, p = 0.0002), as well as a lever x FR interaction (F(2,42) =17.21, p < 0.0001). A Šídák’s multiple comparisons test revealed that active lever presses were significantly greater than inactive lever presses on the FR3 (95% CI [−53.21, −0.2197], p < 0.05) and FR5 (95% CI [−105.8, −52.82], p < 0.0001), but not on the FR1 (Figure 1A). A follow-up analysis that included group assignment based on subsequent extinction in either context A or B (groups AAA vs ABA) showed no significant difference between groups during acquisition.

Figure 1.

Results from Experiment 1A: ABA renewal of cocaine seeking. A) Timeline for Experiment 1A. B) Acquisition of cocaine self-administration in context A collapsed across AAA and ABA groups, showing active and inactive lever presses. *Indicates significant difference between active and inactive lever presses, p < 0.05. C) Extinction of active lever pressing in the AAA group in context A and the ABA group in context B. ****Indicates significant decrease in active lever pressing collapsed across groups, p < 0.0001. D) Renewal of active lever pressing in AAA and ABA groups under the extinction and peer conditions. **Indicates a significant difference in active lever pressing between the AAA and ABA groups when the peer is present, p < 0.01. ****Indicates a significant increase in active lever pressing in the ABA group between the extinction and peer test, p < 0.0001. All data expressed as mean ± SEM based on n= 6 in AAA group and n = 16 in ABA group.

For extinction, a 2-way ANOVA revealed significant main effects of session (F(9,180) = 25.37, p < 0.0001) and group (F(1,20) = 4.59, p = 0.045), with active lever presses decreasing across sessions and, regardless of session, rats in the ABA group having overall less responding than rats in the AAA group (Figure 1B).

For renewal testing, a two-way ANOVA revealed a significant group x test condition interaction (F(1,20) = 13.03, p = 0.002). An uncorrected Fischer’s LSD multiple comparisons test revealed that the number of active lever presses on the extinction test was not different between the AAA and ABA groups. However, when tested in context A, rats in the ABA group showed significantly increased lever pressing compared to extinction (95% CI [−40.35, −17.78], p < 0.0001), indicating that peer-induced renewal was obtained. In contrast, no significant effect of the peer was observed in the AAA group (Figure 1C).

Experiment 1B: BAB renewal of cocaine seeking

Results from the acquisition, extinction, and renewal phases are illustrated in Figure 2. For acquisition, with the data collapsed across sessions within each FR requirement, a mixed-effects ANOVA revealed significant main effects of lever (F(1,12) = 35.56, p < 0.0001) and FR (F(2.471,29.53) = 27.27, p < 0.0001), as well as a lever x FR interaction (F(20,239) = 22.39, p < 0.0001). A Šídák’s multiple comparisons test revealed that active lever presses were significantly greater than inactive lever presses on the FR3 (95% CI [6.126, 35.30], p = 0.0053) and FR5 (95% CI [24.13, 59.01], p < 0.0001), but not on the FR1 (Figure 2B).

Figure 2.

Results from Experiment 1B: BAB renewal of cocaine seeking. A) Timeline for Experiment 1B. B) Acquisition of cocaine self-administration in context B, showing active and inactive lever presses. **Indicates significant difference between active and inactive lever presses on an FR3 schedule, p < 0.01, and ****indicates significant difference between active and inactive lever presses on an FR5 schedule, p < 0.0001. C) Extinction of active lever pressing in context A (not significant, p = 0.0943). D) Renewal of active lever pressing under the extinction and no peer conditions. *Indicates a significant increase in active lever pressing between the extinction and no peer condition, p < 0.05. All data expressed as mean ± SEM based on n= 7.

For extinction, a one-way repeated measures ANOVA revealed no significant effect of session (95% CI [−5.179, 49.75], p = 0.094; Figure 2C).

For renewal testing, a paired t-test revealed that omission of the peer (context B) increased active lever pressing compared to the last day of extinction with the peer (context A; t(6) = 3.36, p = 0.015; Figure 2D).

Experiment 2A: ABA renewal of fentanyl seeking and effect of naltrexone

Results from the acquisition, extinction, and renewal phases are illustrated in Figure 3. For acquisition, with the data collapsed across sessions within each FR requirement, a mixed factor ANOVA revealed significant main effects of lever (F(1,8) = 56.08, p < 0.0001) and FR (F(20,160) = 13.76, p < 0.0001), as well as a lever x FR interaction (F(20,160) = 12.79, p < 0.0001). A Šídák’s multiple comparisons test revealed that active lever presses were significantly greater than inactive lever presses on the FR3 (95% CI [9.898, 51.44], p = 0.002) and FR5 (95% CI [48.34, 89.88], p < 0.0001), but not on the FR1 (Figure 3B).

Figure 3.

Results from Experiment 2A: ABA renewal of fentanyl seeking and effect of naltrexone. A) Timeline for Experiment 2A. B) Acquisition of fentanyl self-administration in context A, showing active and inactive lever presses. **Indicates significant difference between active and inactive lever presses on an FR3 schedule, p < 0.01, and ****indicates significant difference between active and inactive lever presses on an FR5 schedule, p < 0.0001. C) Extinction of active lever pressing in context B. ****Indicates significant decrease in active lever pressing, p < 0.0001. D) Renewal of active lever pressing under the extinction and peer condition, using various doses of naltrexone. **Indicates a significant increase in active lever pressing between the extinction and saline condition, p < 0.01. *Indicates a significant decrease in active lever pressing between the saline condition and the highest dose of naltrexone (1.0 mg/kg), p <= 0.05. All data expressed as mean ± SEM based on n = 9.

For extinction, a one-way repeated measures ANOVA revealed a significant main effect of session (F(3.697,29.58) = 11.83, p < 0.0001), with number of active lever presses decreasing across sessions (Figure 3C).

For renewal testing, a paired t-test revealed a significant increase in active lever pressing in context A compared to the last day of extinction in context B for the saline control condition (t(8) = 3.94, p = 0.004). For the different naltrexone dose conditions in context A, a one-way repeated measures ANOVA revealed a significant main effect of dose (F(2.68,21.44) = 4.83, p = 0.012). A paired t-test revealed a significant decrease in active lever pressing at 1.0 mg/kg naltrexone compared to saline (t(8) = 2.67, p = 0.029; Figure 3D).

Experiment 2B: BAB renewal of fentanyl seeking and effect of naltrexone

Results from the acquisition, extinction, and renewal phases are illustrated in Figure 4. For acquisition, with the data collapsed across sessions within each FR requirement, a mixed factor ANOVA revealed significant main effects of lever (F(1,6) = 33.03, p = 0.0012) and FR (F(20,120) = 22.76, p < 0.0001), as well as a lever x FR interaction (F(20,120) = 14.43, p < 0.0001). A Šídák’s multiple comparisons test revealed that active lever presses were significantly greater than inactive lever presses on the FR3 (95% CI [13.59, 44.41], p < 0.0001) and FR5 (95% CI [29.16, 59.98], p < 0.0001), but not on the FR1 (Figure 4B).

Figure 4.

Results from Experiment 2B: BAB renewal of fentanyl seeking and effect of naltrexone. A) Timeline for Experiment 2B. B) Acquisition of fentanyl self-administration in context B, showing active and inactive lever presses. ****Indicates significant difference between active and inactive lever presses on the FR3 and FR5 schedules, p < 0.0001. C) Extinction of active lever pressing in context A. *Indicates significant decrease in active lever pressing, p < 0.05. D) Renewal of active lever pressing under the extinction and peer condition, using various doses of naltrexone (not significant). All data expressed as mean ± SEM based on n = 7.

For extinction, a one-way repeated measures ANOVA revealed a significant main effect of session (F(4.105,24.63) = 5.57, p = 0.023), with number of active lever presses decreasing across sessions (Figure 4C).

For renewal testing, a paired t-test revealed no significant difference in active lever presses on the renewal test in Context B compared to the last extinction session in Context A for the saline control condition. For the different naltrexone dose conditions in Context A, a one-way repeated measures ANOVA also revealed no significant effect of dose (Figure 4D).

Experiment 2C: Locomotor effects of naltrexone

Results from locomotor experiment are illustrated in Figure 5. When the data were analyzed in 5-min intervals, a two-way repeated measures ANOVA revealed a significant main effect of time (F(2.966, 14.83) = 88.44, p < 0.0001. A Tukey’s multiple comparisons test revealed that distance traveled decreased between the first and last five-minute time bins for saline (95% CI[650.1, 2263], p < 0.01), 0.1 mg/kg naltrexone (95% CI[458, 2446], p = 0.01), 0.3 mg/kg naltrexone (95% CI[503.8, 1874], p < 0.01), and 1 mg/kg naltrexone (95% CI[704.8, 1650], p < 0.001). When the same data were collapsed over the session, a one-way repeated measures ANOVA showed no significant effect of dose.

Figure 5.

Results from Experiment 2C. Main figure) Time course of horizontal distance traveled expressed as distance traveled in cm following naltrexone (0.1, 0.3 or 1 mg/kg) or saline. **Indicates a significant difference from the final time interval (p < 0.01). Insert figure) Total distance traveled across each naltrexone dose condition expressed in m (not significant). All data expressed as mean ± SEM based on n = 6.

DISCUSSION

Regardless of the drug (cocaine vs. fentanyl) or peer presence (ABA vs BAB), rates of acquisition on the incremental FR5 schedule of reinforcement were similarly robust across the four self-administration experiments. This outcome in the cocaine experiments is similar to other studies showing that the presence or absence of a passive peer does not significantly change the rate of cocaine self-administration, regardless whether rats are housed in isolate or social conditions (Robinson et al. 2017; Smith 2012). In contrast, however, the finding in the fentanyl experiments is dissimilar to previous work showing that rats trained to self-administer remifentanil with a peer (whether the peer self-administered remifentanil or saline) show greater rates of intake than those trained without a peer (Hofford et al. 2020). One important difference between Hofford et al. (2020) and the current experiments is the half-life of the opioid used in each. Remifentanil has a shorter half-life than fentanyl (Haidar et al. 1997) and it engenders a higher rate of responding on an FR schedule, thus perhaps making it more sensitive to the presence of a peer. Regardless, the fact that we observed similar rates of acquisition across all four current experiments indicates that any differences noted in either extinction or renewal phases were not due to differential rates of acquisition.

As expected, regardless of the drug (cocaine vs. fentanyl) or peer presence/absence (ABA vs BAB), there was also a decrease in responding on the active lever when the drug was omitted, although it was not statistically significant in Experiment 1B (cocaine BAB group). Examination of the extinction results across experiments revealed that the error variance (SEM) in Experiment 1B was larger than in Experiments 1A, 2A and 2B (cf. Fig 2C vs Figs 1C, 3C and 4C) and thus, the lack of a statistically significant decrease in responding across extinction sessions may simply reflect a type II error. Alternatively, however, it could be that the presence of the novel peer in BAB rats could have suppressed lever pressing at the start of the extinction phase due to having incentive value as an alternative non-drug reward (Calcagnetti and Schechter 1992; Sweeney and Shahan 2015), thus blunting the normal rate of decline in responding across extinction sessions. However, it is difficult to explain why this would have occurred with extinction of cocaine self-administration (Experiment 1B), but not with extinction of fentanyl self-administration (Experiment 2B), and thus we favor a type II error interpretation.

One of the key findings from the renewal tests across experiments was that peer-induced ABA renewal generalized from cocaine to fentanyl, confirming that a peer stimulus previously associated with drug self-administration is a potent trigger for drug seeking across different drug classes. However, while BAB renewal was also obtained with cocaine, it was not obtained with fentanyl. The failure to demonstrate BAB renewal of fentanyl seeking was an unexpected outcome, especially given that it was obtained using either cocaine (Experiment 1B) or food (Browning and Shahan 2018) as the reinforcer. One potential explanation for this unique outcome with the fentanyl self-administration experiments may relate to evidence indicating the rewarding effect of social interaction depends on activation of opioid receptors (Manduca et al. 2016; Trezza et al. 2011). Thus, in the case of ABA renewal of fentanyl seeking, acquisition was in the presence of two different opioid receptor activators (fentanyl and peer), followed by extinction with neither activator (no fentanyl and no peer) and then testing with reintroduction of one of the activators (peer). In contrast, in the case of failing to show BAB renewal of fentanyl seeking, acquisition was in presence of one opioid receptor activator (fentanyl), extinction was in the presence of a different activator (peer) and renewal was tested with neither activator (no fentanyl and no peer). These results suggest that renewal of fentanyl seeking may depend on opioid receptor activation during the test session, either by fentanyl or peer presence. To test the validity of this interpretation, future work could use a BBA renewal paradigm in which acquisition and extinction occur without a peer (context B) and renewal testing occurs with a peer (context A). Demonstrating BBA renewal of fentanyl seeking would support the potential opioid receptor activation explanation outlined above.

An alternative explanation for the differential outcome between cocaine and fentanyl for yielding BAB renewal may relate to their dissociable neurobehavioral mechanisms of action. In terms of their similarity, both cocaine and fentanyl reinforcement and seeking are known to involve either direct or indirect activation of the mesolimbic reward system (Chaudun et al. 2024; Pierce and Kumaresan 2006). However, cocaine and fentanyl also have dissociable neurobehavioral effects. For example, while cocaine targets primarily monoamine transporter systems (Ritz et al. 1990), fentanyl is a relatively selective mu opioid receptor agonist that is thought to produce analgesia and reward through G-protein signaling, while also activating the β-arrestin pathway to produce muscle rigidity and respiratory depression (Comer and Cahill 2019). In addition, at the behavioral level, cocaine dose-dependently increases locomotor activity (Simmons et al. 2013), whereas fentanyl has a biphasic effect, with low doses increasing activity and high doses decreasing activity (Varshneya et al. 2023). Further work is needed to determine what role these dissociable neurobehavioral mechanisms play, if any, in the differential ability of cocaine and fentanyl to support BAB renewal.

Another important finding was that the highest dose of naltrexone (1.0 mg/kg) attenuated ABA renewal with fentanyl. This is consistent with clinical literature showing that extended-release, injectable naltrexone significantly decreases relapse rates and cravings in human subjects with opioid use disorder (Krupitsky et al. 2011). Along with this, preclinical behavioral economic studies show that naltrexone pretreatment significantly increases elasticity and decreases demand for fentanyl (Hammerslag et al. 2020). Importantly, the effect of naltrexone was not due to a general locomotor suppressant effect because the doses tested had no effect on locomotor activity (Experiment 2C). It would be interesting to determine if agonist replacement pharmacotherapies for OUD (buprenorphine and/or methadone) would also reduce peer-induced ABA renewal of fentanyl seeking.