Experiment 1

Effects of social interaction with relapsed partner. Rats were pair-housed and subjected to 10 days cocaine self-administration training, followed by extinction. 24 h after extinction test, one of the two rats (relapsed partner) was primed with cocaine (10 mg/kg, i.p.), and subjected to 1 h reinstatement test. Next, the relapsed partner was put back to their homecage, and interacted with the other rat (targeted rat) for 30 min. Then the targeted rat was immediately subjected to 1 h reinstatement test.

Experiment 2

Effects of social interaction with unrelapsed partner. Rats were pair-housed and subjected to cocaine self-administration training and extinction. 24 h after extinction test, one of the two rats (unrelapsed partner) was injected with saline (1 ml/kg, i.p.), and subjected to reinstatement test. Next, the unrelapsed partner was put back to their homecage, and interacted with the other rat (targeted rat) for 30 min. Then the targeted rat was subjected to reinstatement test immediately.

Experiment 3

Effects of social interaction with relapsed partner for different time. Rats were pair-housed and subjected to cocaine self-administration training and extinction. 24 h after extinction test, one of the two rats (relapsed partner) was primed with cocaine (10 mg/kg, i.p.), and subjected to reinstatement test. Next, the other rat (targeted rat) interacted with the relapsed partner in their homecage for 0, 10, 30, or 60 min, followed by reinstatement test immediately.

Experiment 4

The maintenance of facilitation effect after 30 min social interaction with the relapsed partner. Rats were pair-housed and subjected to cocaine self-administration training and extinction. 24 h after extinction test, one of the two rats (relapsed partner) was primed with cocaine (10 mg/kg, i.p.), and subjected to reinstatement test. Next, the other rat (targeted rat) interacted with the relapsed partner in their homecage for 30 min, and subjected to reinstatement test immediately (0 min), or 1–7 days later.

Experiment 5

Effects of social interaction with relapsed stranger. Rats were pair-housed and subjected to cocaine self-administration training and extinction. 24 h after extinction test, a rat kept in another cage (relapsed stranger) was injected with a single dose of cocaine (10 mg/kg, i.p.) and subjected to reinstatement test. Then the relapsed stranger was put into the homecage of the targeted rat (its partner was removed from the cage) for 30 min social interaction, and the targeted rat was subjected to reinstatement test immediately.

Experiment 6

Effects of social interaction on relapse to cocaine-induced CPP. Rats were pair-housed and subjected to the baseline test. After CPP training and extinction, the rats were subjected to the extinction test. 24 h later, the targeted rats interacted with relapsed partners, unrelapsed partners, or relapsed strangers for 30 min in homecages, and then were subjected to reinstatement test immediately.

We used adult male Sprague–Dawley (SD) rats (280–300 g upon arrival) purchased from Vital River Laboratories. The rats were pair-housed in an animal facility with appropriate temperature (22 ± 2°C) and humidity (50 ± 10%), as well as freely accessible water and food. The lighting time was controlled, under a 12 h light/dark circle. All behavioral experiments were performed under the dark circle and in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and approved by the Biomedical Ethics Committee for Animal Use and Protection of Peking University.

Rats were under anesthetic through pentobarbital sodium (20 mg/ml, 60 mg/kg, i.p.) and received catheters insertion into the right jugular vein with the tip reaching the opening of right atrium. After the surgery, rats were recovered for 7 days with penicillin administration (resolved in 0.2% heparin sodium) every day, preventing infection and cannula blocking.

Cocaine-HCl was purchased from the Qinghai Pharmaceutical Factory and resolved in 0.9% saline (5 mg/ml).

The training procedure was based on our previous studies (Xue et al., 2012; Luo et al., 2015). The chambers (AniLab Software and Instruments) were equipped with two nosepoke holes 9 cm above the underside, one was active-nosepoke hole and the other was inactive-nosepoke hole. The two rats in each cage were trained at the same time in different chambers for fixed ratio 1 (FR1) cocaine self-administration training for 10 days during three 1 h sessions per day with 5 min intervals. Every session started from the illumination of a house light. Poking to the active-nosepoke hole led to a cocaine infusion (0.75 mg/kg), accompanied with a 5 s tone-light cue, while poking to the inactive-nosepoke hole did not result in cocaine infusion or tone-light cue. There was a 40 s time-out phase between each infusion, after which the house light would turn on again. The number and time of active nosepokes, inactive nosepokes, and infusions were recorded. To prevent rats administrating overdose of cocaine, the number of infusions was limited to 20 times in each session. After training, the rats were returned to their homecages.

The two rats in each cage were subjected to extinction at the same time in different chambers. During the extinction sessions, the conditions were the same as those during the self-administration training. But there was no cocaine infusion after rats poked the active-nosepoke hole. At the end of extinction every day, they were put back in their homecages. The extinction was performed until the number of active-nosepokes decreased to below 20% of the mean nosepokes during the last 3 days of self-administration training for at least two consecutive days. 24 h later, the rats were subjected to test for drug seeking (extinction test).

The procedure was based on our previous studies (Xue et al., 2012). Three-chamber apparatuses were used, and time the rats spent in each chamber was recorded. The two rats in each cage were trained at the same time in different apparatuses. For the baseline test (day 1), there were no partitions among the chambers. Rats were put into the middle chamber and allowed to move freely for 15 min. Rats that presented a preference for one of the boxes (resistance time >540 s) were ruled out. Then the rats were trained for cocaine-induced CPP for 8 days. Rats received intraperitoneal cocaine (10 mg/kg, day 2/4/6/8) or saline (1 ml/kg, day 3/5/7/9) injections alternatively and were confined to the conditioning chambers for 45 min after injection. The rats were returned to their homecages after training every day. 24 h after the last-day training (day 10), drug seeking test (training test) was performed.

The two rats in each cage were subjected to extinction at the same time in different chambers. The conditions for extinction were the same as training except that no injections were given. At the end of extinction every day, they were put back in their homecages. After 8 days of extinction, rats were subjected to drug seeking test (extinction test).

For cocaine priming or saline injection, the rats were intraperitoneally injected with 10 mg/kg cocaine or 1 ml/kg saline, and then they were delivered to drug seeking test (reinstatement test).

To investigate the effect of social interaction with a relapsed partner, one rat in each cage was intraperitoneally injected with cocaine (10 mg/kg), and tested for drug seeking. Next, it was put back to its homecage as a relapsed partner, and interacted with the other rat (targeted rat) for 0, 10, 30, or 60 min.

To test the effect of social interaction with an unrelapsed partner, saline (1 ml/kg) was intraperitoneally injected to one rat in each cage, and then drug seeking was tested. Then it was put back to its homecage as an unrelapsed partner, and interacted with the other rat (targeted rat) for 30 min.

To investigate the effect of social interaction with a relapsed stranger, one rat in each cage was intraperitoneally injected with cocaine (10 mg/kg) and tested for drug seeking. Then it was put back to another cage as a relapsed stranger, and interacted with the rat (targeted rat) in this cage for 30 min. The partner of the targeted rat was removed from their homecage in advance.

After social interaction, the targeted rats were subjected to drug seeking test (reinstatement test).

The conditions during the drug seeking tests in the self-administration model were the same as those during the extinction sessions, and the tests lasted for 1 h.

The conditions during the drug seeking tests in the CPP model were the same as those during the baseline test, and the tests lasted for 15 min. The time spent in the cocaine-paired chamber minus the time spent in the saline-paired chamber was calculated as the CPP score.

All of the statistical analyses were performed using SPSS 20.0 software (SPSS, Chicago, IL, United States). The data were expressed as mean ± SEM, and analyzed by repeated measures analysis of variance (ANOVA) with appropriate within-group factors for each experiment (see Results), followed by least significant difference (LSD) post hoc tests in Experiment 6. Values of p < 0.05 were considered statistically significant.

First, we investigated the effect of social interaction on cocaine seeking behavior in the self-administration model. We first explored whether social interaction with a relapsed partner would induce relapse (Figures 1A–C). The repeated measures ANOVA with the within-subjects factors (extinction test and reinstatement test) showed that cocaine injection increased the number of active-nosepokes (F_1,10 = 6.670, p < 0.05), and had no effect on the number of inactive-nosepokes (p > 0.05) in the reinstatement test, indicating that cocaine priming induced reinstatement of drug seeking (Figure 1D). In the targeted rats group, the repeated measures ANOVA of nosepokes with the within-subjects factors (extinction test and reinstatement test) revealed that the number of active-nosepokes was increased after social interaction with relapsed partner (F_1,10 = 17.862, p < 0.01), and inactive-nosepokes had no change (p > 0.05) in the reinstatement test, suggesting that social interaction with relapsed partner resulted in relapse of cocaine seeking (Figure 1E).

We also assessed the effect of social interaction with an unrelapsed partner on cocaine seeking (Figures 1A,F,G). The repeated measures ANOVA showed that the number of active-nosepokes had no significant change (p > 0.05) after saline injection in the reinstatement test, indicating that saline injection did not induce cocaine relapse (Figure 1H). The repeated measures ANOVA of the number of active-nosepokes of the targeted rats showed no significant change (both p > 0.05) after social interaction with unrelapsed partner, indicating that social interaction with unrelapsed partner did not induce relapse of cocaine seeking.

We next explored the effect of different interaction time on the relapse to cocaine seeking, and targeted rats interacted with relapsed partner for 0, 10, 30, or 60 min (Figure 2A). The repeated measures ANOVA showed that the numbers of active-nosepokes of the targeted rats were elevated after 30 min (F_1,7 = 9.171, p < 0.05) and 60 min (F_1,6 = 7.088, p < 0.05) social interaction with the relapsed partner, but not after 0 min (no social interaction) or 10 min social interaction (Figure 2B, both p > 0.05). The above findings indicated that social interaction required a certain amount of time (no less than 30 min) to produce the facilitation effect on relapse.

Then we further tested how long the facilitation effect of social interaction with relapsed partner on relapse could last (Figure 2A). The repeated measures ANOVA showed that the targeted rats had a higher number of active-nosepokes when tested immediately (F_1,7 = 9.171, p < 0.05), or 1 day (F_1,5 = 13.164, p < 0.05), but not 7 days (p > 0.05), after 30 min social interaction with relapsed partners, while no significant change was found when the targeted rats which did not interact with relapsed partners were tested immediately or 1 day later (both p > 0.05). Interestingly, when tested 7 days later, the number of active-nosepokes of the targeted rats that did not experience social interaction was increased during the drug seeking test (F_1,6 = 6.161, p < 0.05), possibly resulting from spontaneous recovery of drug seeking (Figure 2D). These results demonstrated that the effect of social interaction on cocaine seeking lasted for at least 1 day.

Next, we investigated whether social interaction with relapsed strangers could also induce cocaine relapse (Figures 3A–C). The repeated measures ANOVA with within-subject factors (extinction test and reinstatement test) showed that the numbers of active-nosepokes of the strangers were elevated after cocaine injection (Figure 3D, F_1,8 = 13.857, p < 0.05). No significant change was found in the numbers of active-nosepokes of the targeted rats after social interaction with relapsed strangers (Figure 3E, p > 0.05). The above findings indicated that transfer of cocaine relapse occurred during social interaction with relapsed partner rather than relapsed stranger.

Finally, we validated the facilitation effect of social interaction on cocaine relapse through the cocaine-induced conditioned place preference (CPP) model, which is also a widely used animal model to study drug addiction (Figure 4A). The repeated measures ANOVA of CPP scores of relapsed partners with within-subjects factors (baseline test, training test, extinction test, and reinstatement test), showed a significant main effect (Figure 4B, F_3,15 = 4.665, p < 0.05). Post hoc tests revealed significant differences between baseline test and training test (p < 0.05), training test and extinction test (p < 0.01), or extinction test and reinstatement test (p < 0.05). Meanwhile, the repeated measures ANOVA of CPP scores of the targeted rats with within-subjects factors (baseline test, training test, extinction test, and reinstatement test), showed a significant main effect (Figure 4B, F_3,15 = 14.166, p < 0.01). Post hoc tests revealed significant differences between baseline test and training test (p < 0.01), training test and extinction test (p < 0.01), or extinction test and reinstatement test (p < 0.05). The above results suggested that social interaction with relapsed partners promoted the transfer of cocaine relapse. But if the targeted rats did not interact with relapsed partners, no difference was found between the extinction test and reinstatement test (Figure 4C, p > 0.05).

For the unrelapsed partners, the repeated measures ANOVA of CPP scores showed a significant main effect (Figure 4D, F_3,15 = 3.785, p < 0.05). Post hoc tests revealed significant differences between the baseline test and training test (p < 0.01), or training test and extinction test (p < 0.05), but no difference between the extinction test and reinstatement test (p > 0.05). The repeated measures ANOVA of the CPP scores of targeted rats which interacted with unrelapsed partners revealed a significant main effect (Figure 4D, F_3,15 = 9.560, p < 0.01). Post hoc tests revealed significant differences between baseline test and training test (p < 0.05), or training test and extinction test (p < 0.01). No change was found between the extinction test and reinstatement test (p > 0.05), demonstrating that social interaction with unrelapsed partners had no effect on the transfer of cocaine relapse.

A significant main effect was shown for the CPP scores of the relapsed strangers by repeated measures ANOVA (Figure 4E, F_3,18 = 14.296, p < 0.01), and post hoc tests revealed significant differences between baseline test and training test (p < 0.01), training test and extinction test (p < 0.01), or extinction test and reinstatement test (p < 0.05). Meanwhile, the repeated measures ANOVA of CPP scores of the targeted rats interacting with relapsed strangers showed a significant main effect (Figure 4E, F_3,18 = 4.050, p < 0.05). Post hoc revealed significant differences between baseline test and training test (p < 0.01), or training test and extinction test (p < 0.05), while no change was found between the extinction test and reinstatement test (p > 0.05), indicating that social interaction with relapsed strangers did not affect cocaine seeking.