An operant social self-administration and choice model in rats

Abstract

It is difficult to translate results from animal research on addiction to understanding the behavior of human drug addicts. Despite decades of basic research on neurobiological mechanisms of drug addiction, treatment options remain largely unchanged. A potential reason for this state-of-affairs is that mechanistic studies using rodent models do not incorporate a critical facet of human addiction: volitional choices between drug use and non-drug social rewards (such as employment and family). Recently, we developed an operant model in which rats lever press for rewarding social interaction with a peer and then choose between the addictive drugs heroin or methamphetamine versus social interaction. Using this model, we showed that rewarding social interaction suppresses drug self-administration, relapse to drug seeking, and brain responses to drug-associated cues. Here, we describe a standardized protocol of operant social interaction self-administration, discrete-trial choice between drugs versus social interaction that causes voluntary abstinence from the drug, and tests for incubation of drug craving (the time-dependent increase in drug seeking during abstinence). This protocol is flexible, but most commonly requires 7–8 weeks for completion. We also provide a detailed description of the technical requirements and procedures for building the social self-administration and choice apparatus. Our protocol provides a reliable way to study the role of operant social reward in addiction and addiction vulnerability in the context of choices. We also propose that this protocol can be used to study brain mechanisms of operant social reward, and potentially impairments in social reward in animal models of psychiatric disorders and pain.

Editorial summary:

In this operant model rats lever press to obtain rewarding social interaction with a peer or addictive drugs. The model can thus be used to study the role of operant social reward in addiction and addiction vulnerability in the context of choices.

Introduction

Decades of preclinical research on the pharmacological, circuit, and molecular mechanisms of opioid and psychostimulant addiction^1–3 has yet to be translated to successful clinical treatments^4,5. Heilig et al.⁵ suggested that incorporating social factors, which play a critical role in human drug addiction⁶, into mechanistic studies using animal addiction models will improve their predictive validity. Based on these considerations, we recently introduced a rat model⁷ that mimics features of a clinical behavioral treatment known as the Community Reinforcement Approach, or CRA^8,9. The CRA employs principles of operant conditioning by substituting drug use with positive social reinforcers (e.g., family support, employment adherence), contingent in part on cessation of drug use^10–12.

Our rat “CRA model” is based on studies of choice between food versus drug reward in rats and monkeys^13–18. This model is also an extension of our studies on relapse to drug seeking after voluntary abstinence induced by providing rats mutually exclusive choices between the self-administered drug and palatable food^19–22. Using the rat CRA model, we showed that rewarding social interaction suppresses methamphetamine and heroin self-administration in established addiction models^23–25, incubation of methamphetamine craving (the time-dependent increases in drug seeking that occurs during homecage forced abstinence or food choice-induced voluntary abstinence²⁶), and brain responses to methamphetamine-associated cues⁷. We also showed that social choice-induced voluntary abstinence decreases incubation of heroin craving²⁷.

Here, we describe how to use our rat “CRA model”, providing a standardized protocol of operant social interaction self-administration, discrete-trial choice between drugs versus social interaction that causes voluntary abstinence from the drug, and tests for incubation of drug craving. We also demonstrate potential applications of the protocol to study individual variability in the context of the choice between drug and social reward, as well as the motivation to seek operant social reward. We also provide a detailed description of the technical requirements, procedures, and troubleshooting for building a fully automatic social self-administration and social versus drug choice apparatus (Fig. 1). Our protocol provides a reliable way to study the role of social reward in rat addiction models and addiction vulnerability. We also propose that the protocol can be used to investigate the brain mechanisms of volitional social reward, and potentially impairments in social reward in animal models of pain, autism, depression, anxiety, PTSD, and schizophrenia.

Figure 1. Social-choice self-administration apparatus.

(A) Steps to build the apparatus (Stage VII). The base of the apparatus is a standard modular operant test chamber with modified top for rat (ENV – 008CT). (B) Automatic social-choice self-administration apparatus. (Top left panel) The picture represents the apparatus with the configuration described in this protocol and detailed measurements. Food and water are available during the self-administration and choice sessions and are attached in the front door. Legend: 1. Retractable lever 1; 2. Retractable lever 2; 3. Inactive lever; 4. Tone; 5. House light white; 6. House light red; 7. Cue light white; 8. Food magazine (optional); 9. Food receptacle (optional); 10. Pump; 11. Door; 12. Barrier; 13. Fan. 14. SmartCtrl Connection Panel (8 in/16 out). All measurements are in cm. (Bottom left panel) Social apparatus enclosed in the Med-Associates cubicle. Magnification of the modified cubicle for an easy access to the left retractable active lever. (Right panel) Different barrier types. Top panel: barrier for female rats; bottom: barrier for male rats (all measurements are in cm).

Overview of the social choice self-administration protocol

The timeline of the protocol is depicted in Fig. 2. After receiving male and female rats (Charles River Laboratories), we house them two per cage by sex for 2–3 weeks and then individually, starting 1 week prior to social self-administration. We randomly assign the rats to resident (drug user) and social partner (drug naïve) conditions (Stage I). In our protocol, we use male and female rats, with male partners for male rats and female partners for female rats. After an acclimation period, we train the rats to self-administer for access to their social partner during daily sessions using a discrete trial design (see below for details). During this phase, each resident rat lever presses for its previously paired partner (Stage II, social self-administration). After stable social self-administration, we insert a Silastic catheter into the resident rats’ jugular vein. We first allow the rats to recover from surgery for several days and then train them to lever press for intravenous drug infusions (Stage III & IV, drug self-administration). The social- and drug-paired levers are on different sides of the apparatus, and we use different sets of discriminative and discrete cues for the two different rewards (Fig. 3 – see Equipment section). Subsequently, we allow the rats to choose between the drug-paired lever and social interaction, using a discrete-trial procedure (Stage V, voluntary abstinence). We take advantage of the rats’ strong preference for social interaction over drugs to achieve voluntarily drug abstinence^7,27. Finally, we test the rats for drug seeking on abstinence day 1 (the day after the last day of drug self-administration) and on abstinence day 15 (Stage VI, relapse or incubation tests); only the drug-paired lever is available during testing. The two tests at different timepoints allow us to determine the effect of social choice-induced voluntary abstinence on incubation of drug craving (the time-dependent increase in non-reinforced drug seeking abstinence²⁸). Active lever presses on the previously drug-paired lever during testing, the operational measure of drug seeking in incubation of craving and relapse studies^26,29, cause contingent presentations of the light cue previously paired with drug infusions but no drug is delivered.

Figure 2. Protocol timeline.

Stage I – Steps 1–2, Social housing and rats’ separation: ~1 week acclimation to new colony, ~2 weeks social housing and 1 week single housing; Stage II – Step 3, Social self-administration: 6 days social self-administration; Stage III & IV – Step 4, Surgery 1 day plus ~3–4 days recovery; Step 5, Drug self-administration 12 days; Stage V – Step 6, Discrete choice procedure: 10 sessions over 14 days; Stage VI – Step 7, Relapse or Incubation tests: 2 days, on abstinence day 1 and day 15. This protocol is flexible, but most commonly requires 7–8 weeks for completion.

Figure 3. Using the protocol to study incubation of drug craving.

(A) Effect of social choice-induced voluntary abstinence on incubation of methamphetamine craving. (Left panel – Self-administration training). Number of social (60 trials) rewards and methamphetamine infusions (6 h), in male and female rats. (Center panel – Voluntary abstinence). Number of social rewards and methamphetamine infusions earned during the 10 discrete-choice sessions. (Right panel – Relapse or Incubation tests). Active-lever presses during the 30-min test sessions; left: forced abstinence, right: social choice abstinence. During testing, active-lever presses led to contingent presentation of the light cue previously paired with methamphetamine infusions during training, but not methamphetamine delivery (extinction conditions). Adapted from ref.⁷. Data are mean±SEM; * different from test day 1. (B) Effect of social choice-induced voluntary abstinence on incubation of heroin craving. (Left panel – Self-administration training). Number of social (60 or 20 trials) rewards and heroin infusions (6 h), in male and female rats. (Center panel – Voluntary abstinence). Number of social rewards and methamphetamine infusions earned during the 10 discrete-choice sessions. (Right panel – Relapse or Incubation tests). Active-lever presses during the 30-min test sessions; left: forced abstinence, right: social choice abstinence. Adapted from ref.²⁷.. Data are mean±SEM; * different from test day 1; # different from social choice abstinence. This study has been approved by the NIDA IRP Animal Care and Use Committee.

Below we provide additional details of the experimental procedures, including a description of the steps required and a list of materials to build our fully automatic social self-administration and choice apparatus (Stage VII), a “Troubleshooting” section, and “Anticipated results” section.

Comparison with other models

The protective effects of social interaction on addiction-related behaviors were previously demonstrated in rodent models³⁰. Experimenter-imposed social interaction either outside or inside the testing cage decreases drug conditioned place preference (CPP), self-administration, and reinstatement (relapse) of drug seeking^31–34. Group-housed rats exposed to an enriched environment show decreased drug CPP, drug self-administration, and reinstatement^31,33,34. Similarly, pairing a peer rat with a nondrug context decreases both cocaine CPP and drug priming-induced reinstatement of CPP^35,36. Other studies focused on social facilitation and social inhibition of drug intake in rats and showed that drug self-administration is facilitated in socially housed rats if both members of the pair had access to drug and, conversely, drug self-administration is inhibited if only one rat of the pair has access to drug^32,37–40.

However, these previous models do not incorporate the volitional choice between social interaction and drug use that occurs in human drug users. Additionally, in monkeys and rodents, drug self-administration is reliably decreased by operant availability of other nondrug rewards such as palatable food¹³. And most rats choose sucrose or saccharin over heroin or cocaine, even after a long history of drug self-administration¹⁵.

Advantages

Our protocol allows researchers to study whether drug self-administration is reduced by providing rats volitional or subject-controlled operant choice between drug and social reward, a setup that more closely models the human condition⁵. The fully automatic social choice apparatus²⁷ allows researchers to simultaneously train many rats, while eliminating the intense experimenter workload and confounds related to rat-human interactions of the semi-automatic model we introduced in our original study⁷. The fully automatic procedure is generalizable to different rat strains (Fig. 4). The fully automatic model also makes it feasible to incorporate both classical and cutting-edge techniques in behaving rats (e.g. in vivo electrophysiology, calcium imaging, etc.) to study mechanisms underlying complex and ethologically relevant volitional social behaviors. Additionally, our social choice procedure has long-lasting effects on drug relapse and craving, even when the social reward is discontinued. We showed that social choice-induced voluntary abstinence prevents incubation of methamphetamine craving (Fig. 3A)⁷ and decreases incubation of heroin craving (Fig. 3B)²⁷.

Figure 4. Generalization of the protocol to Long-Evans rats and remifentanil.

(A) Timeline of the experiment. (B) Self-administration training. Number of social rewards (20 trials) and remifentanil infusions (2 h), n=8 Long-Evans male rats (plus n=8 social partners). (C) Choice tests. Number of social rewards and remifentanil infusions earned during the 5 discrete-choice sessions. Data are mean±SEM. This study was approved by the NIDA-IRP Animal Care and Use Committee.

From a translational perspective, our choice procedure models the human treatment in which abstinence is rewarded with alternative social-based nondrug incentives^41,42. While almost 100% of our rats prefer to interact with a social partner rather than self-administer a drug, some human drug users will continue to use drugs despite the availability of social support⁴³. The difference between the clinical scenario and our rats is related to the time of social reward delivery: in our rats it is immediate upon lever pressing while in humans social reward is often delayed⁴⁴. However, when we introduced a delay to get access to a social partner the abstinence rate decreased to ~50% (Exp. 2; Fig. 5), consistent with findings in humans⁴⁵. Thus, our model can be adapted to study addiction vulnerability and potentially model addiction of people who appear to benefit less from the protective effects of social support.

Figure 5. Using the protocol to study addiction vulnerability.

(A) Timeline of the experiment. (B) Self-administration training. Number of social rewards (20 trials) and heroin infusions (6 h), n=12 Sprague- Dawley female rats (plus n=12 social partners). (C) Choice tests. Number of social rewards and heroin infusions earned during the 12 discrete-choice sessions. After the first 3 choice sessions (yellow dash line), we introduced a 60-s delay for the social reward. (D) Individual data. Heat map of the preference score (number of social rewards/[number of social rewards + number of drug infusions]) for each rat during the 12 discrete-choice sessions. 1.0 (dark blue) indicates preference for social and 0.0 (white) indicates preference for heroin. Data are mean±SEM. This study was approved by the NIDA IRP Animal Care and Use Committee.

More broadly, the role of social interaction is not only critical for drug addiction but also plays a major role in the etiology of several human psychiatric disorders. Therefore, here we show that our protocol can be adapted to study motivational aspects of social reward (Exp. 3; Fig. 6), using the progressive ratio reinforcement schedule, commonly used to study the motivation to seek drug and food rewards⁴⁶, and a social seeking test analogous to the one routinely used in drug studies²⁶ (Exp. 4; Fig. 7). These procedures provide a platform for future studies on behavioral, pharmacological, and circuit mechanisms of social reward and social seeking. We propose that progressive ratio responding for social reward and social seeking tests at different times after cessation of operant social self-administration can be used to study disruption of social reward in animal models of pain, autism, depression, anxiety, PTSD, and schizophrenia.

Figure 6. Using the protocol to study the motivation to seek operant social reward.

(A) Timeline of the experiment. (B) Self-administration training. Number of palatable food pellets (20 trials) and social rewards (20 trials); n=3 male Sprague Dawley rats (plus n=3 male social partners). (C) Choice individual data. Heat map of the preference score (number of social rewards/[number of social rewards + number of palatable pellets]) for each rat during the 5 discrete-choice sessions. 1.0 (dark blue) indicates preference for social and 0.0 (white) indicates preference for food. (C) Progressive ratio test. Number of food or social rewards earned during progressive-ratio tests. Data are mean±SEM. This study was approved by the NIDA IRP Animal Care and Use Committee.

Figure 7. Using the protocol to study social seeking.

(A) Timeline of the experiment. (B) Self-administration training. Number of social rewards (60 trials); n=12 Sprague Dawley – 6 males and 6 females – rats (plus n=12 – 6 males and 6 females – social partners). (C) Social-seeking test. Active and inactive lever presses during the 30-min test sessions performed 1 d after the last social self-administration session. During testing, active-lever presses led to contingent presentation of the light cue previously paired with social reward during training, but not social interaction (extinction conditions). Data are mean±SEM. This study was approved by the NIDA IRP Animal Care and Use Committee.

Finally, although it is now possible to buy the entire apparatus directly from Med-Associates (note that the authors do not have any commercial agreement with this company), it is significantly cheaper to build our custom-made automatic apparatus (Fig. 1). As described below, it requires principal components purchased from Med-Associates and then additional inexpensive materials (primarily plastic and Plexiglas). This represents a cheaper solution, and therefore our custom-made apparatus is potentially affordable to any laboratory.

Limitations

Our protocol requires double the number of rats for any given experiment (as resident and social partners are required). This increases the overall cost of the experiment and requires more animal facility space (and therefore the associated per diem costs). Another limitation is that the protocol is sensitive to the sizes of the barrier holes that separate the two sides of the chamber (Fig. 1). It is critical to allow the rats to at least touch each other’s faces during social interactions. We recommend using the barrier with the dimensions described below; if the holes are too small the rats do not maintain stable social self-administration and social preference (Table 1 & 2).

Table 1.

Troubleshooting table

Step	Problem	Possible reason	Solution
	Apparatus not working	Electrical or Med-PC interface	Check that all the cables and wires are properly connected and the SmartCtrl card pins interface with the chamber
	Social partners escaping from the chamber	Open space between the back wall and the stainless-steel grid floor	Add a polycarbonate panel to the back of the chamber
10	Apparatus does not fit inside the Med-Associates cubicle	Not enough space for the left retractable lever	Enlarge the already-existing hole on the left side of the cubicle for easy passage of the left active retractable lever
10	Difficult to access to the SmartCtrl Connection Panel (8 in/16 out) for plugging in all cables	Not enough space for the panel	Attach the panel (using Velcro) to the back of either the IKEA cabinet or the sound-attenuating cubicle.
13	Low number of rewards earned during the session	Rats cross the barrier that separates the two sides of the social-choice self-administration chamber	Recreate the barrier with smaller holes and increase the number of social self-administration sessions
		The holes of the barrier are too small to allow enough social interaction	Recreate a barrier with larger holes and increase the number of social self-administration sessions
16	Rats lose weight during recovery	Infection close to the catheter site	Provide antibiotics and contact the facility veterinary staff
	Rats in discomfort, clean fluids coming out of the rats back, or tension in the catheter during flushing	Malfunction or blocked catheter	Flush the catheter with saline and heparin. If the problem persists, implant the catheter in the other vein. If this strategy does not work, remove the rat form the experiment
17	Rats do not acquire stable self-administration	Wrong dose of drug or catheter malfunction	Prepare new drug and check catheter patency
	Low number of rewards but high number of active lever presses	Drug syringe is empty, or rats are disconnected from the drug lines	Change the syringes every time they are below ∼ 9 ml. Flush the catheter and re-connect the rats
22	Missing trials during the session	Rats crossing the two sides of the chamber or not enough space for social interaction	Recreate the barrier (as described above)
		Levers, cues, or door malfunction	Change levers, cues, or door

Table 2.

Scenario of protocol issues

Scenario	Consequences
While we were developing the automatic social choice procedure, we tested several different screens. One of them had circular holes of 3 cm of diameter. Some of the rats (especially males) were stuck into the holes. We had to manually free the rats by breaking the plastic screen	Rats did not reliably perform the operant task. Therefore, during social self-administration (in one case) and choice (in another case) included missed sessions/trials. This is because the experience of being stuck into the holes was extremely stressful for the rats and they stopped lever pressing
In one occasion the guillotine door was not properly working. The gears inside were spinning and the door was not opening making an extremely loud noise for the duration of the experiment	The rat was too scared to perform the task for the next three days.

Experimental Design

Our procedures follow the guidelines outlined in the Guide for the Care and Use of Laboratory Animals (8th edition; http://grants.nih.gov/grants/olaw/Guide-for-the-Care-and-Use-of-Laboratory-Animals.pdf). The studies from which we show results here were all approved by the NIDA-IRP Animal Care and Use Committee. Obtaining appropriate permissions and conforming to regulations is important.

Experimental timing and organization of the protocol.

This protocol is flexible, but most commonly requires 7–8 weeks for completion. Typically, we run 6 days of social self-administration, 12 days of drug self-administration (which varies based on the drug procedure used for each experiment – see ref.⁷), 10 sessions of social choice-induced voluntary abstinence over 14 days, and two sessions of relapse/incubation tests (on abstinence days 1 and 15). We often run experiments in cohorts ranging from 8 to 32 rats (plus social partners), with runs occurring once daily for each rat in the different cohorts. We run our experiments throughout the entire day, and we have not observed any differences related to the time of day of testing. We suggest that experimenters new to this model start with cohorts of no more than 8 rats (plus 8 social partners). We also suggest that social self-administration sessions include ~60 trials (60 s interaction) per session (~ 2 h) to give the rats enough exposure to their social partners. Using the automatic social choice procedure described here, this phase does not require the presence of the experimenter to separate the two rats after each trial. However, we recommend checking on the rats periodically to make sure that none cross the barrier (this can occur with small female rats). After surgery, we recommend flushing the catheters daily with sterile saline containing gentamicin (4.25 mg/mL) to maintain the catheter patency.

Effects of modifications to the procedure.

In our published papers and Fig. 3^7,27, we used male and female Sprague Dawley rats. We also used our protocol with Long-Evans rats (Exp.1; Fig. 4). Thus, we speculate that our protocol can be used also to train a different strain of rats. We observed no sex differences and the protocol is adaptable for both male and female rats.

Typically, we start the protocol with rats weighing 150 to 175 g (~ 40–60 days old) on arrival. In our experience, a rat’s age (within 2–6 months⁷) is not a critical feature for maintaining stable social self-administration or social preference. However, we recommend keeping the residents and their social partners at the same body weight range or age range in the different experiments. Additionally, while partner rats can be socially housed two per cage, resident rats should be single housed to prevent catheter damage.

We have also showed that the rats’ strong social preference over drug persists even if rats are socially housed or if they are lever pressing to gain access to an unfamiliar rat (see ref.⁷). In a modified version of this protocol, we trained the rats first to self-administer methamphetamine and then for social interaction, followed by a series of choice tests. We investigated whether the most ‘addicted’ rats (selected using an established rat model of individual differences in addiction vulnerability²⁵) would prefer drug over social interaction. We found that independent of the addiction score and independent of the sequence of the training procedure, the rats preferred social interaction over drug⁷. This set of data shows that the preference for social interaction over a drug is independent of the training sequence for the two rewards.

Materials

Animals

Male and female Sprague Dawley or Long Evans rats (Charles River Laboratories), weighing 150 to 175 g on arrival. ∧ CRITICAL We house the rats 2 per cage by sex for 2–3 weeks and then individually, starting 1 week prior to social self-administration. We randomly assign rats to resident (drug user) and social partner (drug naïve) conditions. ! CAUTION Experiments must follow all governmental and institutional guidelines for care and use of laboratory animals. Moreover, it is critical to report any excluded rats (due to catheter failure or lack of reliable self-administration, etc.). The studies described below were approved by the NIDA-IRP Animal Care and Use Committee.

REAGENTS

Drugs [We used drugs provided by NIDA Pharmacy].∧ CRITICAL We have used different class of drugs (methamphetamine (Fig. 3A)⁷, heroin (Fig. 3B)^7,27, and remifentanil (Exp. 1, Fig. 4)) in our protocol. Thus, it is highly likely that our protocol can be used to study the effect of social reward on self-administration and relapse to other drugs. ∧ CRITICAL For any new drugs, we recommend using standard doses reported in the literature and validating the selected drug dose with proper dose-effect curve (see ref.⁷). ∧ CRITICAL For any drug, report in the drug laboratory book the exact amount of drug removed from the original container. ! CAUTION We recommend using all laboratory safety precautions to handle drugs to avoid potential contact with the syringe needles or contact of the solutions with the experimenter’s eyes. Dispose all the sharp materials to the designed sharp container to avoid hazard materials to the other users of the lab space.

Equipment

Parts for custom-made social choice apparatus: ∧ CRITICAL All parts can be found at Med-Associates Inc

• ENV – 008CT Standard modular operant test chamber with modified top for rat;

• DIG-716B – SmartCtrl interface module (8 input / 16 output);

• SG-716B – SmartCtrl connection panel (8 input / 16 output);

• SG-210CB – DB25 SmartCtrl cable, M/F 25’ (7.6m);

• SG-210CP-25 – Power Cable, 25’ (7.6 m);

• ENV-010BS – Auto guillotine door;

• FAB-ENV-008–32 – Social chamber aluminum mesh (plus Thumbscrew HAR-Thumb-4–40X5/16-LowPro);

• ENV-307A-GFW – Stainless steel grid floor (white front and back) for ENV-307A;

• ENV-307–07 – Waste pan for ENV-307A chamber;

• ENV-025F28 – Replacement 28 VDC fan with cable;

• ENV-112CM – Retractable lever (n=2);

• ENV-110M – Standard fixed lever (inactive);

• ENV-223AM – Sonalert module with volume control for rat chamber, 2,900 Hz;

• ENV-215M – House light hooded, 100 mA, 28 V DC;

• FAB-ENV-008–07 – Right/Left/Center front support (n=1 each);

• ENV-221M – Stimulus light, 1″ (1 white and 1 red lens);

• ENV-203M-45 – Modular pellet dispenser, magazine type 45 mg (optional) ∧ CRITICAL our apparatus can be used also for food versus social choice studies);

• ENV-200xxx – Pellet receptacle + plastic tube (optional);

• PHM-100 – Infusion pump;

Other equipment:

• Clear impact-resistant polycarbonate (12” x 12” x 1/4” Sheet – purchased from McMaster Carr). ! CAUTION For any cutting and assembling, we recommend applying the appropriate safety measures (goggles, cut resistant gloves, coats, etc.)

! CAUTION We recommend building the apparatus far from any behavioral/facility rooms because the noise of a Dremel or drill can affect rats’ behavior in nearby rooms.

• Clear rectangular cage for double-housed rats (34 cm (w) x 40 cm (d) x 20 cm (h)) with plastic cover top and lid for food and water (Fig. 2);

• Clear rectangular cage for single-housed rats (23 cm (w) x 35 cm (d) x 20 cm (h)) with plastic cover top and lid for food and water (Fig. 2);

• Hard woodchip bedding;

• Cleaning solution and wipes.

• Med-Associates programs, written using Med-PC code to automatically run social and drug self-administration, social versus drug choice, and relapse tests (available upon request from the authors)

Equipment setup

Surgery

We anesthetize the rats with isoflurane (5% induction; 2–3% maintenance). We then insert Silastic (Silastic Silicone Laboratory Tubing, 0.020 3 0.037 “wall 0.009”; VWR, Radnor, PA) catheters into the jugular vein, which we pass subcutaneously to the midscapular region and attach to a modified 22-gauge cannula cemented to polypropylene mesh (Small Parts, Amazon.com). ∧ CRITICAL We inject ketoprofen (2.5 mg/kg, s.c., Butler Schein, Dublin, OH) after surgery to relieve pain and decrease inflammation. We allow the rats to recover from surgery for 3–4 d. ∧ CRITICAL We flush the catheters daily with sterile saline containing gentamicin (4.25 mg/mL, Fresenius Kabi, Lake Zurich, IL) during the recovery, training, and voluntary abstinence phases. ! CAUTION We recommend using all the safety laboratory precautions to handle gentamicin and ketoprofen to avoid potential contact with the syringe needles or contact of the solutions to the experimenter’s eyes. Dispose all the sharp materials to the designed sharp container to avoid hazard materials to the other users of the lab space.

Rats social- and single-housed

Once we receive the rats, we group-house them 2 per cage by sex. Fig. 2 depicts a standard cage for double-housed rats. We provide free access to water and food during the entire duration of the experiments. ∧ CRITICAL We suggest allowing the rats to acclimate to the facility for 1 week and then handling them twice a week for the following 2 weeks before separation. After this period, we randomly assign rats to resident (drug user) and social partner (drug naïve) conditions. Fig. 2 also depicts a standard cage for single-housed rats. ∧ CRITICAL Both double- and single-housed cages should be cleaned 2–3 times per week.

Custom-made social-choice self-administration apparatus

The standard modular operant test chamber for rat (ENV-008CT) is the base for the custom-made social-choice self-administration. The apparatus can be enclosed in regular Med-Associates sound attenuating cubicle (ENV-017M) with a small modification in the lateral side (Fig. 1B; we enlarged the already-existing hole on the left side of the cubicle for easy passage of the left retractable lever). Alternatively, the apparatus can be housed within inexpensive IKEA cabinet frames (catalog #: 802.653.98). All the parts listed above are necessary to build 1 apparatus (Fig. 1; all parts can be found at Med-Associates Inc).

A standard Med Associates self- administration chamber is combined with a custom-made social-partner chamber that is separated by a guillotine door (ENV-010BS). Each chamber should have a discriminative cue on the right panel (white house-light; ENV-215M) to signal the insertion and subsequent availability of the social reward-paired active (retractable) lever located near the guillotine door and a discriminative cue on the left panel (ENV-221M, red lens) to signal the insertion and subsequent availability of the drug-paired active (retractable) lever located on the left side. Locate the levers 6 cm above the grid floors and a white discrete light cue (ENV-221M, white lens) above the drug-paired lever and a discrete tone cue (ENV-223AM) above the social-paired lever. The left side can also be equipped with a pellet dispenser, pellet receptacle (our apparatus can be used also for food versus social choice study), and an inactive (stationary) lever. Attach a fan on the back of the cubicle for background noise. Include a bottle of water and food hopper to provide food and water during the self-administration and choice sessions (Fig. 1B). ∧ CRITICAL Check rats’ body weight daily before the beginning of each session. ∧ CRITICAL After surgery, flush the rats’ catheter daily with sterile saline containing gentamicin (4.25 mg/mL). ∧ CRITICAL Before the beginning of each drug self-administration session,check whether each line and swivel are properly working (we manually move the infusion pump and flush the Tygon tube that delivers drug from the syringe to the rats’ catheter). Prepare 20 ml syringes of drug and change them once they are below ~ 9 ml to prevent the rats from emptying the syringes during the self-administration sessions. ! CAUTION We recommend periodically checking the rats and making sure that the programs are working properly, that the rats are not crossing the barrier during social self-administration, and that they remain connected to the drug lines for the entire duration of the drug self-administration session. ∧ CRITICAL At the end of each session, we change bedding from both the resident and the social partner side of each chamber. Also, we wipe the stainless-steel grid floors with water once a week and clean the entire apparatus before each relapse test to avoid any confounds due to the previous presence of the social partners.

Data collection

We collect behavioral data via a computer using the Med-PC program. Subsequently, we transfer the data to an Excel file and analyze the data using SPSS (IBM, version 25, GLM procedure). See below for “Anticipated results” section.

Procedure

Stage I: Building the social-choice self-administration apparatus • TIMING 3–6 days per chamber

∧ CRITICAL This phase can vary on timing depending on the experimenter’s experience – Each step is descripted in Fig. 1A. ∧ CRITICAL As an alternative to self-building this apparatus, it can be obtained from Med-Associates Inc.

• 1Remove the second center front support from a standard modular operant chamber (ENV – 008CT);
• 2Make three holes on the white plastic base 20 cm from the support removed above;
• 3Add the right/center/left front supports;
• 4Cut the right metal flap. One screw is attached to the center support, but this still leaves space for the door; this will allow easy access to levers and cues in that side;
• 5Attach the auto guillotine door (ENV-010BS) to the top of the modular chamber;
• 6Cut two rectangular pieces (20 cm x 15 cm) from the clear impact-resistant polycarbonate sheet; attach them to the front and back of the partner side of the chamber (between the auto guillotine door and the metal supports);
• 7Cut a third piece of polycarbonate to create the top of the partner chamber (18 cm x 18 cm). Prepare a metal flap (18 cm) and attach it to the back of the partner chamber and the polycarbonate piece to allow the top part to open and close easily. Complete the top part with a latch to keep the chamber close when the social partner is inside.
• 8Refine the plastic white back part of the stainless-steel grid floor (ENV-307A-GFW) to be ~ 5 cm;
• 9Add all the other components (levers, cues, etc.).
• 10Remove the SmartCtrl Connection Panel (8 in/16 out) (SG-716B) from the white base and attach it to the back of either the IKEA cabinet or the sound attenuating cubicle.

  ? TROUBLESHOOTING

Stage II: Social housing and separation of rats • TIMING 2–3 weeks

• 11Purchase male and female Sprague Dawley or Long Evans rats (Charles River Laboratories), weighing 150 to 175 g. House the rats 2 per cage by sex, with free access to food and water, and allow rats to habituate to their new colony facility for a minimum of 1 week prior to handling. ∧ CRITICAL Our protocol is likely to work with other strain of rats.
• 12Two weeks later, randomly assign rats to resident (drug user) and social partner (drug naïve) conditions. ∧ CRITICAL STEP Mark the rat tails to keep track of the different social pairs.

  ? TROUBLESHOOTING

Stage III: Social self-administration • TIMING 1 week

• 13Train rats to self-administer for access to their social partner during daily 120-min sessions (60 trials/session, 60 s social interaction) using a discrete trial design. Each resident rat lever presses for its previously paired partner. First bring the rats from the facility and move them from their home cage to their assigned side of the operant chamber (residents and partners). Start the session by uploading the Med-PC social self-administration program that, once issued, will automatically start the session with illumination of the social-paired houselight followed 10 s later by insertion of the social-paired lever; we allow resident rats, via the Med-PC social self-administration program, 60 s to press the active lever (fixed-ratio-1 reinforcement schedule) before lever retraction and houselight turning off. CRITICAL STEP Successful lever presses from the resident rats cause the retraction of the active lever, a discrete 20-s tone cue, and opening of the guillotine-style sliding door. Resident rats are subsequently allowed to interact with their social partner for 60 s until the houselight turns off, at which point the guillotine door closed. Record the number of successful trials and inactive lever presses. ∧ CRITICAL STEP Check whether the rats are crossing the barrier.

  ? TROUBLESHOOTING

• 14At the end of the session, remove both resident and partner rats and bring them back to the facility in their respectively home cages.
• 15During this training phase, run the social self-administration procedure for 6 sessions by repeating steps 13 and 14.

Stage IV & V: Surgery and drug self-administration • TIMING 3 weeks

• 16The day after the last day of social self-administration, anesthetize the rats with isoflurane (5% induction; 2–3% maintenance), insert a Silastic catheters attached to a modified 22- gauge cannula cemented to polypropylene mesh into the jugular vein and pass it subcutaneously to the midscapular region. Inject ketoprofen (2.5 mg/kg, s.c.) after surgery to relieve pain and decrease inflammation. Allow rats to recover from surgery for several days. ∧ CRITICAL STEP Flush the catheters daily with sterile saline containing gentamicin during the entire duration of the experiment.

  ? TROUBLESHOOTING

• 17After 5–7 days of recovery from surgery, train the resident rats to self-administer drugs. First bring the resident rats (e.g. drug users) from the facility and move them from their home cage to the operant chamber. Start the self-administration sessions at the onset of the dark cycle, and start each session by uploading the Med-PC drug self-administration program that, once issued, will automatically start the session with the presentation of the discriminative cue (red houselight) and 10 s later the insertion of the drug-paired lever; the red houselight remains on for the duration of the session. CRITICAL STEP Successful lever presses (fixed-ratio-1 20-s time-out reinforcement schedule) from the drug user rats cause the retraction of the active lever, a discrete 20-s light cue, and the automatic delivery of a drug infusion via the infusion pump (active for 3.5 seconds). The Med-PC drug self-administration program will run for six 1-h sessions separated by 10-min off periods. At the end of each 1-h session, the red houselight is automatically turned off and the active lever is retracted. The program will also automatic limit the number of drug infusions to 15 per hour to avoid overdose. ∧ CRITICAL Check whether the rats are connected to the drug lines. ∧ CRITICAL STEP Record the number of drug infusions and active and inactive lever presses. ∧ CRITICAL
• 18At the end of the session, remove the drug user rats and bring them back to the facility in home cages.

  ? TROUBLESHOOTING

• 19Run rats for drug self-administration for 12 sessions, by repeating steps 17 and 18 for 12 consecutive days. Note that this phase can vary on timing depending on the self-administration procedure used up to 3 months or more; in this scenario, repeat steps 17 and 18 for 5–6 days a week for 3 months (provide some off days over the weekends).

Stage VI: Relapse or Incubation test • TIMING 1 day

∧ CRITICAL The relapse test in the presence of drug cues usually consists of 30-, 60-, or 90-min (or longer) sessions on abstinence day 1 (the day after the last day of drug self-administration) and day 15 (the day after the last day of voluntary abstinence – see stage VIII below).

• 20Bring the resident rats from the facility and move them from their home cage to the operant chamber. Start the session by uploading the Med-PC relapse program that, once issued, will automatically start the session with the presentation of the red discriminative houselight, followed 10 s later by the insertion of the drug-paired lever; the red house light remains on for the duration of the session. Active lever presses during testing, the operational measure of drug seeking in incubation of drug craving and relapse^26,29, will automatically result in contingent presentations of the light cue previously paired with drug infusions, but not drug delivery. At the end of the session, the active lever is automatically retracted, and the red houselight is turned off. ∧ CRITICAL Before the beginning of the session, remove the syringes from the infusion pumps. ∧ CRITICAL STEP Record the number of active and inactive lever presses. During this phase, we do not present the lever or cues previously associated with social interaction.
• 21At the end of the session, remove the drug user rats and bring them back to the facility in home cages.

Stage VII: Discrete Choice Procedure • TIMING 2 weeks

• 22The day after the Day 1 relapse test, bring both resident and partner rats from the facility and move them from their home cage to their assigned side of the operant chamber (residents and partners). Start the session by uploading the Med-PC choice program that, once issued, will automatically start the session with the presentation of the discriminative cues for social interaction (houselight) and drug (red light), followed 10 s later by the insertion of the levers paired with both rewards. Rats can then select one of the two levers. If the rats respond within 6 min, they only receive the reward that corresponds with the selected lever (60 s social interaction for the social-paired lever and one drug infusion for the drug-paired lever). Thus, on a given trial, the rat can earn either the social reward or the drug reward, but not both. Each reward delivery is signaled by the social- (20-s tone cue and opening of the guillotine-style sliding door) or drug-associated (20-s light cue and activation of infusion pump) discrete cue, the retraction of both levers, and the extinguishing of both discriminative cues. If the rat fails to respond on either active lever within 6 min, both levers are retracted, and their related discriminative cues are extinguished with no reward delivery. The Med-PC choice program allows rats to choose between the social- and drug-paired levers in a discrete-trial choice procedure; it divides each 120-min choice session into 15 discrete trials separated by 8 min. ∧ CRITICAL STEP Record the number of social and drug rewards, and inactive lever presses.
• 23At the end of the session, remove the drug user rats and bring them back to the facility in home cages.
• 24During this voluntary abstinence phase, run the choice procedure for 10 sessions over 14 days (provide some off days over the weekends) by repeating steps 20 and 21.

  ? TROUBLESHOOTING

Stage VIII: Relapse or Incubation test • TIMING 1 day

• 25The day after the last day of voluntary abstinence repeat steps 20 and 21 to test the rats’ drug seeking at late abstinence phase.

• TIMING

The duration of an entire social-choice self-administration protocol, including acclimation and separation of drug users and social partners, is ~7–8 weeks.

Stage I – Steps 1–10, Social-choice self-administration apparatus building: 3–6 days depending on experience.

Stage II – Steps 11–12, Social housing and rats’ separation: ~1-week acclimation to the new colony, ~2 weeks social housing, and 1-week single housing;

Stage III – Step 13–15, Social self-administration 6 days;

Stage IV & V – Step 16, Surgery 1 day plus several days of recovery; Step 17–19, Drug self-administration 12 days;

Stage VI – Step 20–21, Relapse or Incubation test: 1 session on abstinence day 1;

Stage VII – Step 22–24, Voluntary abstinence: 10 sessions in 14 days;

Stage VII – Step 25, Relapse or Incubation test: 1 session on abstinence day 15.

Troubleshooting

Our protocol has been used and tested in more than ~600 rats (including published and unpublished data) with consistent and reliable data across experiments. Therefore, we do not anticipate critical issues at any of the Stages reported above. However, based on our experience, we provide below (Table 1) a list of solutions to potential problems that experimenters may encounter while running the protocol as described here. Additionally, we provide (Table 2) examples of scenarios in which the experiment went wrong to facilitate future troubleshooting. For any unanticipated issues, experimenters can contact the corresponding authors for suggestions.

Anticipated results

We report the social self-administration data as number of rewards (equal to the successful trials) that rats earn during the 120-min daily sessions. We then report the number of drug infusions earned by the rats during the 6-h self-administration sessions. For the voluntary abstinence phase, we report the number of social rewards and drug infusions earned during the 10 discrete-choice sessions. Finally, for the relapse or incubation tests we report the number of previously active and inactive lever presses during the sessions. We use factorial ANOVA and t tests using SPSS (IBM, version 25, GLM procedure) for statistical analysis of the behavioral data. When we obtain significant main effects and interaction effects (P < 0.05, two-tailed), we follow them with post- hoc tests (Fisher PLSD). For choice data, the statistical analyses are performed on a social preference ratio score (number of social rewards/[number of social rewards + number of drug infusions])²⁷. Usually, we do not present the inactive lever data in figures, because responding on this lever during the any stage of the protocol is very low (though we typically report the range of inactive lever presses for each experiment). Moreover, in our studies we did not observe sex differences during either social or drug self-administration, social preference over drug during the voluntary abstinence phase, and for the relapse/incubation tests^7,27. We do not use statistical methods to pre-determine sample sizes, and our sample sizes are similar to those reported in previous publications^7,27.

Using this protocol, we have shown that social choice-induced voluntary abstinence prevents incubation of methamphetamine craving⁷ (Fig. 3A) and reduces incubation of heroin craving²⁷ (Fig. 3B).

Here, we also provide unpublished data to show generalization of our protocol to Long Evans rats (Exp. 1 – n=8 male drug users and n=8 male social partners, weighing 150–175 g upon arrival) and the opioid drug remifentanil. A timeline of the experiment is reported in Fig. 4A. The rats reliable lever press for social interaction (Fig. 4B). The analysis of number of operant social interactions showed a significant main effect of session (F_7,49=5.1, p<0.001; partial Eta²=0.4). The rats also reliably lever pressed for remifentanil infusions (Fig. 4B). The analysis of number of infusions showed a significant effect of session (F_7,49=15.8, p<0.001; partial Eta²=0.7). We then tested the rats during 10 discrete choice sessions, and they showed strong preference for social interaction over remifentanil (Fig. 4C). The analysis of the social preference score showed a significant effect of session (F_4,28=0.4, p=0.8; partial Eta²=0.05).

Our protocol can be adapted to study addiction vulnerability. Based on our original findings with methamphetamine⁷, in Exp. 2, we tested whether rats would reverse their social preference over heroin by introducing a delay of the social reward during the choice session. A timeline of the experiment is shown in Fig. 5A. The rats (n=12 female drug users and n=12 female social partners, weighing 150–175 g upon arrival), increased the number of social rewards over time (Fig. 5B). The analysis of number of operant social interactions and heroin infusions showed a significant effect of session (F_5,55=16.6, p<0.001; partial Eta²=0.6; F_5,55=2.8, p=0.03; partial Eta² = 0.2, respectively). We then tested the rats during 12 discrete choice sessions. After the first 3 choice sessions, we introduced a 60-s delay for the social reward but not for heroin. This manipulation decreased rats’ preference to ~50% (Fig. 5C). The analysis of the social preference score showed a significant effect of session (F_11,121=13.2, p<0.001; partial Eta² = 0.6). However, this reversal occurred with different magnitude to different rats, showing individual variability in delayed social choice (Fig. 5D).

Our protocol can also be adapted to study motivational aspects social reward. In Exp. 3, we directly compare the rats’ motivation to seek two natural rewards (e.g. palatable food and social interaction) using our choice procedure and progressive ratio reinforcement schedule. A timeline of the experiment is shown in Fig. 6A. Nonfood deprived rats (n=3 male rewards users and n=3 male social partners, weighing 250–275 g upon arrival), increased both the number of food and social rewards over time (Fig. 6B; Session: F_5,10=3.9, p=0.03; partial Eta²=0.7; F_5,10=50.2, p<0.001; partial Eta² = 0.9, respectively). We then tested the rats during 5 discrete choice sessions showing that 2 out 3 rats preferred social interaction over palatable food pellets (Fig. 6C). Then we tested the rats’ motivation to seek the two rewards using a progressive ratio reinforcement schedule and we found no differences between the two reward types (p values>0.05).

Our protocol can also be adapted to study more broadly social reward and non-reinforced social seeking. In Exp. 5, we tested whether rats would perform stable social self-administration for longer periods (12 days) than the usual 6–8 sessions and if they would show social seeking after a short abstinence period (1 day). A timeline of the experiment is reported in Fig. 7A. The rats (n=12 drug users – 6 females and 6 males – and n=12 social partners – 6 females and 6 males – weighing 150–175 g upon arrival), increased the number of social rewards over time (Fig. 7B; Session: F_11,110=9.0, p<0.001; partial Eta²=0.5). One day after social self-administration training, the rats showed robust non-reinforced social seeking (Fig. 7C). The analysis showed a main effect of Lever (inactive, active): F_1,10 = 53.7, p < 0.001; partial Eta² = 0.9). During this phase, we presented the cues and the lever previously paired with social self-administration, but there was no social interaction and the door remained closed during the test. For both social self-administration and non-reinforced social seeking, we did not observe sex differences (p values > 0.05).

Taken together, these data show that our protocol is flexible and can be extended to different strains of rats and different drugs of abuse. Our data also show that the protocol can also be used to study addiction vulnerability, and the mechanisms of social reward and social reward seeking.