Dose-dependent effect of retrieval-extinction on preventing reinstatement of cocaine-associated memory in mice

Heng-Ai Chang; Wen Dai; Sherry Shu-Jung Hu

doi:10.4103/0304-4920.354804

ORIGINAL ARTICLE

Year : 2022 | Volume : 65 | Issue : 4 | Page : 159-170

Dose-dependent effect of retrieval-extinction on preventing reinstatement of cocaine-associated memory in mice

Heng-Ai Chang¹, Wen Dai², Sherry Shu-Jung Hu²
¹ Department of Psychology; Institute of Basic Medical Sciences, National Cheng Kung University, Tainan, Taiwan
² Department of Psychology, National Cheng Kung University, Tainan, Taiwan

Date of Submission	07-Apr-2022
Date of Decision	06-Jun-2022
Date of Acceptance	09-Jun-2022
Date of Web Publication	26-Aug-2022

Correspondence Address:
Dr. Sherry Shu-Jung Hu
Cannabinoid Signaling Laboratory, Department of Psychology, National Cheng Kung University, 1 University Road, Tainan 70101
Taiwan

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/0304-4920.354804

Abstract

As a noninvasive behavioral intervention, the retrieval-extinction (R-E) procedure has drawn much research attention for its capacity to target the reconsolidation of maladaptive memories. However, later research findings suggest that the cause and consequence of R-E may be more complicated than previously suggested. For example, the R-E procedure could increase an animal's motivation for drug-seeking under certain circumstances, and the reversed extinction-retrieval (E-R) procedure could also suppress the drug memory. Two possible mechanisms underlying the R-E procedure have been proposed: the reconsolidation-update and extinction-facilitation hypotheses. To elucidate the paradoxical prior findings and examine these two hypotheses, we systematically examined the efficacy of the extinction (E), R-E, and E-R procedures in mice's low-dose versus high-dose cocaine-induced conditioned place preference (CPP) memory. We showed that the dose of cocaine is a crucial determinant of the efficacy of the three behavioral interventions. The E procedure exerted a long-lasting suppression of the low-dose cocaine CPP memory, while the R-E procedure induced more memory defects than the E and E-R procedures in its long-term suppression of the high-dose cocaine CPP memory. It warrants further investigation of whether the R-E procedure's underlying neurochemical and molecular mechanisms differ from the E and E-R procedures.

Keywords: Addiction, cocaine dose, retrieval-extinction, reconsolidation-update, extinction-facilitation, relapse

How to cite this article:
Chang HA, Dai W, Hu SS. Dose-dependent effect of retrieval-extinction on preventing reinstatement of cocaine-associated memory in mice. Chin J Physiol 2022;65:159-70

How to cite this URL:
Chang HA, Dai W, Hu SS. Dose-dependent effect of retrieval-extinction on preventing reinstatement of cocaine-associated memory in mice. Chin J Physiol [serial online] 2022 [cited 2023 Sep 26];65:159-70. Available from: https://www.cjphysiology.org/text.asp?2022/65/4/159/354804

#Heng-Ai Chang and Wen Dai contributed equally to this work.

Introduction

Why emotion can boost our memory is traced back to the long evolutionary history, as remembering the emotionally arousing stimuli well promotes the survival of a species. When an emotional stimulus (unconditioned stimulus, US; e.g., footshock) is paired with an initially neutral cue (conditioned stimulus, CS; e.g., tone, light, or context), the association is consolidated into long-term memory within hours after learning. It was originally thought that memory, once consolidated, is difficult to disrupt and becomes a permanently encoded memory trace.^[1],[2] Extinction or exposure-based therapy, a noninvasive and well-established procedure, is used to reduce fear and anxiety in animals and humans.^[3] By repeatedly presenting a previously trained CS in the absence of US, extinction leads to a progressive decrease in the CS-elicited fear.^[4] It is argued that the extinction (E) procedure does not modify the original memory trace but instead engages a new-learning process of the CS-noUS association that inhibits responding to the original CS-US association.^[5] The E procedure involves a competition between the original and extinguished memories, thus resulting in a short-term suppression of the original memory. The extinguished fear responses can recover with the passage of time (spontaneous recovery [SR]);^[6],[7],[8] when the CS is presented in a new context different from the extinction context (renewal);^[9],[10] when exposing the animal to the unsignaled US or stressor (reinstatement).^[4],[11]

Due to the difficulties in disrupting the well-consolidated memory,^[12] recent research has focused on reactivating the previously consolidated memory, rendering it into a labile period. Once the original memory is retrieved and made transiently labile, it undergoes destabilization and restabilization to be updated and restored. This restabilization process is termed “reconsolidation” in the associative learning paradigms.^[13],[14] Targeting reconsolidation by the pharmacological intervention (e.g., protein synthesis inhibitors, β-blockers, etc.) during the reconsolidation window results in memory deficits at later times, suggesting that memories are either abolished (i.e., due to storage deficits) or persistently inhibited (i.e., due to retrieval deficits).^{[13],[15],[16],[17],[18]}

Monfils et al. are the first to combine the strengths of reconsolidation and extinction to combat the recurring fear.^[19] By presenting a retrieval trial to reactivate memory shortly before extinction, Monfils et al. demonstrated that the retrieval-extinction (R-E) procedure caused an enduring attenuation in renewal, reinstatement, and SR of learned fear in adult rats.^[19] They also showed that the control group that received the standard E procedure had a recurrence of the fear-conditioned responses. Flavell et al. replicated Monfils et al.'s findings by showing that the R-E procedure prevented the return of contextual fear memories, but not the auditory cued fear conditioning.^[19],[20] These authors further reconditioned all rats with a weaker footshock in the same context and found that those who had received E alone reacquired contextual fear memory. The rats that had received the R-E procedure were immune to reconditioning and thus retained diminished fear-conditioned responses.^[20] Schiller et al. soon applied the non-invasive R-E procedure to humans.^[21] These authors showed that when human subjects were given a single retrieval trial 10 min before extinction, they no longer expressed fear memories even when tested one year later. In addition to fear memories, drug-associated memories can also be modified. Xue et al. found that the R-E procedure completely blocked SR and drug-primed reinstatement in rat cocaine-and morphine-induced conditioned place preference (CPP) tasks.^[22] These authors also showed that the R-E procedure attenuated the SR, renewal, and reinstatement of drug-seeking behaviors in rat cocaine and heroin self-administration (SA) models.^[22] Finally, they demonstrated that the R-E procedure caused a long-lasting attenuation of the cue-induced increase in drug craving and blood pressure in abstinent heroin addicts.^[22]

The mechanisms underlying the efficacy of the R-E procedure have been proposed in two hypotheses: reconsolidation-update versus extinction-facilitation.^{[19],[21],[22],[23],[24],[25]} The reconsolidation-update hypothesis argues that when extinction is presented after a retrieval trial, the original memory enters a new phase of stabilization, thereby allowing the extinction to modify and restabilize the original memory.^[19],[26] The original memory, after its reactivation (by retrieval) and modification (by extinction), becomes a renewed memory that no longer supports reinstatement of drug-seeking^[22] or return of fear.^[19] Despite the fact that the existence of a reconsolidation process could only be inferred from its absence, the reconsolidation-update hypothesis yield two predictions that can be empirically verified.^[27],[28] The first is that extinction must occur within the reconsolidation window for the R-E procedure to be effective. When the retrieval is omitted, the memory should remain unmodified by standard E. The second is that the disruption of reconsolidation should protect animals against reinstatement, rendering them back to a state similar to the originally naïve animals. The evidence for and against these two predictions has been accrued. For example, Monfils et al.^[19] and Xue et al.^[22] supported the first prediction by showing that extinction conducted outside the reconsolidation window failed to prevent later reinstatement, and the standard E procedure caused a recurrence of the fear-conditioned and drug-primed responses. However, by employing a reversed sequence (i.e., extinction then retrieval), Millan et al.^[29] and Baker et al.^[30] opposed the first prediction by showing that the reversed extinction-retrieval (E-R) procedure, like the R-E procedure, was able to attenuate the reinstatement of alcoholic beer-seeking and fear-conditioned responses. The authors deduced that the reversed E-R procedure should have been invalid since a retrieval trial was unlikely to act retrospectively on the encoding of the extinguished memory. Moreover, other studies failed to show the efficacy of E-R,^{[20],[31],[32],[33],[34],[35]} which suggested that critical boundary conditions such as memory type, memory age, memory strength, or retrieval duration affect whether memory could be reactivated or modified after a retrieval trial.^{[36],[37],[38],[39],[40],[41]} As for the second prediction, Millan et al.^[29] included an originally naïve (Ori-Naïve) group that had never been trained or extinguished before and examined whether the R-E procedure could resume those animals back to their originally naïve state.^[29] Surprisingly, the authors found that the E and R-E groups, when tested under a progressive ratio schedule, both displayed an enhanced motivation for drug-seeking than the Ori-Naïve group. The increase in original breakpoints was even more remarkable in the R-E group than in the E group, implying that R-E could promote relapse under certain circumstances.^[29] Their results suggest that the R-E procedure is not always protective, and its underlying mechanism is more complex than we initially thought. On the other hand, the extinction-facilitation hypothesis suggests that the R-E procedure strengthens the extinction training, rendering the extinguished memory more resistant to SR, renewal, and reinstatement.^[24],[25] While the findings of Millan et al.^[29] and Baker et al.^[30] described above better support the extinction-facilitation hypothesis, but they did not discern which hypothesis (reconsolidation-update or extinction-facilitation or both) underlies the efficacy of the R-E procedure.^{[19],[22],[42]}

This study aimed to examine the reconsolidation-update and extinction-facilitation hypotheses in the mouse model of cocaine-induced CPP. Thus far, no study has compared the efficacy of the R-E procedure with the reversed E-R procedure and the Ori-Naïve group in the cocaine-induced CPP memory. Only Millan et al.^[29] and Baker et al.^[30] conducted the reversed E-R experiments in the alcoholic beer SA and cued fear conditioning tasks, respectively. Hence, we systematically examined the efficacy of the E, R-E, and E-R procedures on cocaine-induced CPP memory in mice. First, we examined the effects of the E and R-E procedures on reinstatement of the 10 mg/kg cocaine-induced CPP memory while adding the reversed E-R procedure for comparison. Second, we added an Ori-Naïve group and a cocaine CPP retraining procedure to examine whether the above three behavioral procedures could return the mice back to an originally naïve state or protect them from the reacquisition of the 10 mg/kg cocaine-induced CPP. Finally, we investigated the role of cocaine dose in the efficacy of the three behavioral procedures by using the 40 mg/kg of cocaine to establish CPP. We previously showed that the modulatory effect of the CB₁ antagonist rimonabant depended on the strength of CPP memory.^[43] We observed a graded dose-response curve of CPP induced by various doses of cocaine, with the 40 mg/kg cocaine inducing the strongest CPP memory. Rimonabant differentially modulated the low-dose (10 mg/kg) versus high-dose (40 mg/kg) cocaine-induced CPP memory, implying that different mechanisms underlay various doses of cocaine-associated memories.^[43] Therefore, we examined whether the E, R-E, and E-R procedures would exert the same efficacy in the low-dose versus high-dose cocaine-induced CPP memory.

Materials and Methods

Animals

All animal experiments conformed to the Guide for the Care and Use of Laboratory Animals in the National Institutes of Health Guideline for Animal Research and were reviewed and approved by the Institutional Animal Care and Use Committee of the National Cheng Kung University College of Medicine (approval number: 104116). Eight- to twelve-week-old male C57BL/6J mice, which weighed 22–28 g, were obtained from the National Laboratory Animal Center (NLAC, Taipei, Taiwan). They were housed in groups (four to five mice per plastic cage) under controlled temperature (22°C ± 1°C), humidity (70%), and lighting conditions (12-h light/dark cycle with lights on 0700). Food and tap water were accessible ad libitum.

Drug treatment

Cocaine hydrochloride was purchased from Sigma (St. Louis, MO, USA) and dissolved in sterile saline for intraperitoneal (i.p.) injection before use. The doses of 10 and 40 mg/kg of cocaine were chosen because we previously showed that rimonabant differentially modulated the low-dose (10 mg/kg) versus the high-dose (40 mg/kg) cocaine-induced CPP memory.^[43] These results imply that different mechanisms underlie various doses of cocaine-associated memories.

Cocaine-induced conditioned place preference

The CPP apparatus is composed of a neutral compartment (7.2 cm × 7.2 cm × 13 cm length × width × height) in the middle, and two main compartments (13 cm × 13 cm × 13 cm length × width × height) on each side. These two side compartments (with white or black walls and ceilings, steel bar grid or wire mesh floors, and strong or dim roof light) are separated from the middle compartment (with gray walls, ceiling, and platform floor) by movable sliding doors. Chambers were thoroughly wiped and deodorized with a 70% isopropyl alcohol-rinsed paper towel and dried before each training and test bout.

Establishment of conditioned place preference

Pretest

On day 1, the unconditioned preference of each subject for a particular environment was assessed by placing the mice in the center chamber and allowing them to explore the three compartments for 15 min (pretest). After the pretest, two behavioral criteria were applied to avoid apparent unconditioned preference toward any compartment. First, the mouse that spent longer than 180 s in the middle compartment was excluded. Second, the mouse was excluded if the time difference between the two side compartments was greater than 300 s.

Conditioning

This study used a biased design for cocaine-induced CPP. During days 2-7, all mice underwent 6 days of cocaine-CPP conditioning (CPP training). The mice received an i.p. injection of saline and were immediately confined in their unconditioned preferred compartment for 30 min (days 2, 4, and 6); the same mice received a cocaine injection (10 or 40 mg/kg, i.p.) and were confined in the unconditioned least-preferred compartment for 30 min (days 3, 5, and 7).

Posttest

On days 8 and 15, the mice, free of cocaine, were placed in the center chamber and allowed free exploration for 15 min in the CPP test (test 1 and test 2). After each test, the mouse that spent longer than 180 s in the middle compartment or 600 s in each side compartment was excluded. This standard excluded the probability that mice's substantial preference toward the neutral middle compartment would confound their CPP magnitude. The 600-s standard was adopted in all posttests because the 40 mg/kg cocaine could induce high CPP scores. The cocaine-induced CPP score (in s) was presented as “time spent difference,” which was calculated by subtracting the time spent in the saline-paired compartment from the time spent in the cocaine-paired compartment. A significantly greater cocaine-CPP score in the test than in the pretest reflected the establishment of CPP.

Extinction, retrieval-extinction, and extinction-retrieval procedures

During days 9–14, all mice, free of cocaine, underwent 6 days of extinction training with different procedures. The mice were randomly assigned into three groups: E, R-E, and E-R. The E group underwent a standard extinction procedure, in which they were alternately confined in the saline-paired compartment (days 9, 11, 13) and the cocaine paired-compartment (days 10, 12, 14) for 40 min. The R-E group first allowed free exploration in the CPP apparatus for 10 min as a retrieval trial and returned to the home cage for a 10-min break. They were then alternately confined in the saline-paired compartment (days 9, 11, 13) and the cocaine-paired compartment (days 10, 12, 14) for 30 min as an extinction trial. The E-R group had their extinction training in reverse order. The mice first underwent a 30-min extinction procedure where they were alternately confined in the saline-paired compartment (days 9, 11, 13) and the cocaine-paired compartment (days 10, 12, 14). After a 10-min break in their home cage, they were allowed free exploration in the CPP apparatus for 10 min as a retrieval trial.

Experimental procedures and groupings

For all of the following tests, the same behavioral criteria were applied. In other words, the mouse that spent longer than 180 s in the middle compartment or 600 s in each side compartment was excluded.

Experiment 1

At first, fifty-two mice underwent cocaine (10 mg/kg)-induced CPP. After the mice required cocaine CPP memory (test 1), they were randomly divided into three groups as described above: E (initial n = 18), R-E (initial n = 18), and E-R (initial n = 16). After the CPP test on day 15 (test 2), the mice were returned to the home cage. Since SR is defined as the extinguished responses that have recovered with the passage of time, drug-free mice's CPP preference scores were assessed for 15 min as the SR 1 and SR 2 on days 22 and 36, respectively. On day 37, all mice received a cocaine administration (5 mg/kg) right before the cocaine priming test (cocaine priming) [Figure 1]a. At the end of experiment 1, five mice were excluded using the behavioral criteria. Thus, each group had the final subject number as follow: E (n = 17), R-E (n = 18), and E-R (n = 12).

Figure 1: Effects of the extinction, retrieval-extinction, and extinction-retrieval procedures on the low-dose (10 mg/kg) cocaine-induced CPP memory. (a) Timeline of the experiments. (b) Effects of the three behavioral procedures on the spontaneous recovery and reinstatement of cocaine-associated memory. E, extinction; R, retrieval; E group, extinction group; R-E group, retrieval-extinction group; E-R group, extinction-retrieval group. Data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001, which show differences between groups were statistically significant.

Click here to view

Experiment 2

At the beginning, sixty-three mice underwent low-dose cocaine (10 mg/kg)-induced CPP. They were then randomly assigned into three groups: E (initial n = 16), R-E (initial n = 16), and E-R (initial n = 16). Another group of mice, which we called the original-naïve group (Ori-Naïve: initial n = 15), only underwent the CPP pretest on day 1 (pretest) and were kept in their home cage through days 2–19. On days 20–21, all four groups were confined in the saline-paired compartment (day 20) and the cocaine paired-compartment (day 21) for 30 min. This cocaine-CPP conditioning served as the retraining procedure for the first three groups; as the one pair of new training for the Ori-Naïve group. On days 22, 29, and 30, all mice underwent the CPP test (test 3), the SR, and the cocaine (5 mg/kg) priming test (cocaine priming), respectively [Figure 2]a. According to the behavioral criteria, seven mice were excluded. The final subject number for each group was: E (n = 14), R-E (n = 14), E-R (n = 16), and Ori-Naïve (n = 12).

Figure 2: Effects of the extinction, retrieval-extinction, and extinction-retrieval procedures, compared with the original-naïve mice after cocaine CPP retraining on the low-dose (10 mg/kg) cocaine-induced CPP memory. (a) Timeline of the experiments. (b) Effects of the three behavioral procedures, when compared with original-naïve mice after cocaine reacquisition, on spontaneous recovery and reinstatement of the low-dose cocaine-associated memory. E, extinction; R, retrieval; E group, extinction group; R-E group, retrieval-extinction group; E-R group, extinction-retrieval group; Ori-Naïve group: original-naïve group. Data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001, which show differences between groups were statistically significant.

Click here to view

Experiment 3

Originally, sixty mice in Experiment 3 were randomly assigned to the four experimental groups: E (initial n = 15), R-E (initial n = 15), E-R (initial n = 15), and Ori-Naïve (initial n = 15). They underwent the same experimental procedures as those in Experiment 2. The only difference is that the mice in Experiment 3 went through high-dose cocaine (40 mg/kg)-induced CPP conditioning [Figure 3]a. At the end, five mice were excluded, which left the final subject number of each group as follow: E (n = 14), R-E (n = 15), E-R (n = 15), and Ori-Naïve (n = 11).

Figure 3: Effects of the extinction, retrieval-extinction, and extinction-retrieval procedures, compared with the original-naïve mice after cocaine CPP retraining on the high-dose (40 mg/kg) cocaine-induced CPP memory. (a) Timeline of the experiments. (b) Effects of the three behavioral procedures, when compared with original-naïve mice after cocaine reacquisition, on spontaneous recovery and reinstatement of the high-dose cocaine-associated memory. E, extinction; R, retrieval; E group, extinction group; R-E group, retrieval-extinction group; E-R group, extinction-retrieval group; Ori-Naïve group: original-naïve group. Data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001, which show differences between groups were statistically significant.

Click here to view

Statistical analysis

SPSS Statistics 17.0 (IBM, Armonk, NY, USA) was used for statistical analysis. The cocaine-CPP score (in s) was presented as the mean ± standard error of the mean. In three experiments, paired t-tests (pretest vs. test 1) were first used to ensure the reliable acquisition of cocaine-induced CPP. Various sets of paired t-tests (test 1 vs. test 2; test 1 vs. test 3; test 1 vs. SR 1; test 1 vs. SR 2; test 1 vs. priming) were employed to assess whether each behavioral intervention (E, R-E, E-R) affected the cocaine CPP memory. For the Ori-Naïve group, a paired t-test (pretest vs. test 3) was used to examine whether they reliably acquired cocaine-CPP memory after the one pair of new CPP training. Paired t-tests (test 3 vs. SR; test 3 vs. priming) were then used to assess the SR and cocaine priming effects in the Ori-Naïve group. To compare the memory-impairing effects among different behavioral interventions (and their effects compared with the Ori-Naïve group), one-way analysis of variances (ANOVAs) were used for multiple-group comparisons at each time point, followed by the post hoc Fisher's least significant difference (LSD) tests. The levels of statistical significance were set at P < 0.05.

Results

Only the E, but not the R-E or E-R procedure, exerted long-term suppression of the low-dose cocaine conditioned place preference memory

The effects of the E, R-E, and E-R procedures on the low-dose (10 mg/kg) cocaine-CPP memory were examined [Figure 1]a. The mice were first pooled together, and the CPP scores (in s) in the pretest (-102.34 ± 12.14) and test (135.36 ± 21.39) were compared, which showed that all mice reliably acquired cocaine CPP memory [paired t-test (pretest vs. test 1): t₄₆ = 11.677, P = 0.0001] [Figure 1]b: #a]. These mice were then randomly assigned to the E, R-E, and E-R groups. A one-way ANOVA revealed that there was no difference among the three groups before the three extinction procedures began [test 1: F (2, 44) =0.232, P = 0.794]. In the test 2, the E and E-R procedures (paired t-test [test 1 vs. test 2]: E: t₁₆ = 4.062, P = 0.001; E-R: t₁₁ = 2.466, P = 0.031), but not the R-E procedure (P = 0.154), diminished cocaine CPP memory [[Figure 1]b: #b, #c]. In the SR 1, only the E procedure impaired mice's memory (paired t-test [test 1 vs. SR 1]: t₁₆ = 2.907, P = 0.01); not the R-E or E-R procedure (P = 0.357; 0.271). None of the E, R-E, and E-R groups, however, damaged memory in the SR 2 (P = 0.221; 0.194; 0.176). In the priming test, only the E group (paired t-test [test 1 vs. priming]: t₁₆ = 2.584, P = 0.02), but not the R-E or E-R group (P = 0.595; 0.693), impaired cocaine memory [[Figure 1]b: #b, #c]. To compare the memory-impairing efficacy between the E, R-E, and E-R groups, one-way ANOVAs followed by post hoc LSD were employed. In test 2, the difference among three groups were marginally significant [test 2: F (2, 44) =2.999, P = 0.06], and the E and E-R groups showed more pronounced memory deficits when compared with the R-E group: (E vs. R-E: P = 0.039; E-R vs. R-E: P = 0.048). The differences among the three groups were not significant in the SR1, SR2, and priming tests (all P > 0.05). However, the E group, when compared with the R-E group, showed diminished memory in the SR 1: (E vs. R-E: P = 0.035) and the priming test: (E vs. R-E: P = 0.05) [Figure 1]b. Together, these results indicated that the R-E procedure was not as effective as the E and E-R procedures. Only the E procedure displayed a long-term suppression effect on the low-dose cocaine CPP memory.

The E, R-E, and E-R procedures did not return the low-dose cocaine-conditioned mice back to the original-naïve state

To investigate whether the E, R-E, and E-R procedures could return the low-dose (10 mg/kg) cocaine-conditioned mice back to their original-naïve state, we added the CPP retraining procedure and the Ori-Naïve mice group. The Ori-Naïve mice did not receive any drug treatment until the one pair of CPP new training on days 20 and 21, when the E, R-E, and E-R groups underwent cocaine CPP retraining [Figure 2]a. Thus, the CPP scores of the three groups were compared with the Ori-Naïve group in test 3, SR, and the priming test. First, the mice were pooled together. Their CPP scores (in s) in the pretest (-85.25 ± 19.03) and test (163.2 ± 22.53) were compared, and the results showed that all mice reliably acquired cocaine CPP memory (paired t-test [pretest vs. test 1]: t₄₃ = 12.074, P = 0.0001) [Figure 2]b: #a]. These mice were randomly divided into the E, R-E, and E-R groups. A one-way ANOVA revealed that the CPP scores of three groups were equivalent before each extinction procedure began [test 1: F (2, 41) = 0.274, P = 0.762]. In test 2, the E and E-R procedures impaired cocaine CPP memory (paired t-test [test 1 vs. test 2]: E: t₁₃ = 3.568, P = 0.003; E-R: t₁₅ = 3.462, P = 0.003), while the R-E procedure barely affected the cocaine CPP memory (t₁₃ = 2.156, P = 0.05). In the test 3, the mice in the E group had diminished cocaine CPP memory (paired t-test [test 1 vs. test 3]: t₁₃ = 2.38, P = 0.033), but not those in the R-E or E-R group (P = 0.681; 0.454). The Ori-Naïve mice, after their one pair of new training, obtained cocaine CPP memory in test 3 (paired t-test [pretest vs. test 3]: t₁₁ = 5.723, P = 0.0001). The Ori-Naïve mice showed diminished cocaine CPP memory in the SR [paired t-test (test 3 vs. SR): t₁₁ = 2.716, P = 0.02], and the priming test [paired t-test (test 3 vs. priming): t₁₁ = 2.556, P = 0.027]. Nevertheless, none of the E, R-E, or E-R procedures impaired mice's cocaine CPP memory in the SR [P = 0.431; 0.19; 0.619]. While the E and R-E procedures did not affect the cocaine-primed CPP memory (P = 0.685; 0.999), the E-R procedure paradoxically boosted cocaine-primed memory in the priming test (paired t-test [test 1 vs. priming]: t₁₅ = 2.942, P = 0.01) [[Figure 2]b: #b, #c, #d, #e]. In test 2, one-way ANOVA revealed a significant difference among three groups [test 2: F (2, 41) =6.489, P = 0.004]. The post hoc LSD tests showed that the E and E-R groups had more pronounced memory deficits when compared with the R-E group (E vs. R-E: P = 0.001; E-R vs. R-E: P = 0.019). In test 3, one-way ANOVA indicated no difference among groups (P > 0.05), but post hoc LSD revealed that the E group showed more serious cocaine memory deficits than the R-E group (E vs. R-E: P = 0.034). In order to examine whether the E, R-E, and E-R procedures could bring the mice back to their originally naïve status, we compared the memory-impairing efficacy of the E, R-E, and E-R groups with the Ori-Naïve group. In the SR, one-way ANOVA revealed a marginally significant difference among groups [SR: F (3, 52) = 2.601, P = 0.062], and the post hoc LSD tests indicated that the Ori-Naïve group had memory deficits (Ori-Naïve vs. E: P = 0.022; Ori-Naïve vs. R-E: P = 0.031; Ori-Naïve vs. E-R: P = 0.02). In the priming test, one-way ANOVA showed a significant group difference [priming: F (3, 52) =2.794, P = 0.049], and the post hoc LSD tests indicated that the Ori-Naïve group had memory deficits [Ori-Naïve vs. E: P = 0.037; Ori-Naïve vs. R-E: P = 0.044; Ori-Naïve vs. E-R: P = 0.008] [Figure 2]b. Our findings indicated that only the E procedure, but not the R-E or E-R procedure, temporarily suppressed the low-dose cocaine CPP memory after CPP retraining. However, none of the three behavioral manipulations (E, R-E, and E-R) had returned the low-dose cocaine-conditioned mice to their original-naïve state.

The R-E procedure showed a greater magnitude than the E and E-R procedures in the long-term suppression of the high-dose cocaine conditioned place preference memory

We used the high-dose (40 mg/kg) cocaine to examine the effects of the E, R-E, and E-R procedures, when compared with Ori-Naïve mice, on memory retention after one pair of CPP retraining [Figure 3]a. We first showed that all mice reliably acquired cocaine CPP memory by comparing the pooled CPP scores (in sec) in the pretest (-106.45 ± 16.51) and test (230.93 ± 18.99) [paired t-test (pretest vs. test 1): t₄₃ = 17.816, P = 0.0001] [[Figure 3]b: #a]. The mice were then randomly assigned to the E, R-E, and E-R groups, and one-way ANOVA revealed that these three groups were equivalent before each extinction procedure began [test 1: F (2, 41) =1.062, P = 0.355]. In the test 2, the E, R-E, and E-R procedures all significantly impaired cocaine CPP memory (paired t-test [test 1 vs. test 2]: E: t₁₃ = 8.558, P = 0.0001; R-E: t₁₄ = 7.15, P = 0.0001; E-R: t₁₄ = 6.239, P = 0.0001). After cocaine CPP retraining, the E, R-E, and E-R procedures reduced cocaine CPP memory in the test 3 (paired t-test [test 1 vs. test 3]: E: t₁₃ = 3.705, P = 0.003; R-E: t₁₄ = 4.409, P = 0.001; E-R: t₁₄ = 3.934, P = 0.001). The Ori-Naïve mice, after the one pair of new training, acquired the high-dose cocaine CPP memory in the test 3 (paired t-test [pretest vs. test 3] t₁₀ = 5.877, P = 0.0001). The cocaine CPP memory of the Ori-Naïve mice, however, remained undamaged in the SR (paired t-test [test 3 vs. SR]: t₁₀ = 0.746, P = 0.473), and the priming test (paired t-test [test 3 vs. priming]: t₁₀ = 0.217, P = 0.833). The memory-impairing effects of the E, R-E, and E-R procedures lasted in the SR (paired t-test [test 1 vs. SR]: E: t₁₃ = 3.032, P = 0.01; R-E: t₁₄ = 3.849, P = 0.002; E-R: t₁₄ = 4.613, P = 0.0001]), and the priming test (paired t-test [test 1 vs. priming]: E: t₁₃ = 3.261, P = 0.006; R-E: t₁₄ = 4.663, P = 0.0001; E-R: t₁₄ = 2.323, P = 0.036) [[Figure 3]b: #b, #c, #d]. In the test 2, one-way ANOVA showed a significant group difference [test 2: F (2, 41) = 3.671, P = 0.034], and the post hoc LSD tests indicated that the R-E group, when compared with the E and E-R groups, had severer memory deficits (R-E vs. E: P = 0.045; R-E vs. E-R: P = 0.015) [Figure 3]b. We also compared the E, R-E, and E-R groups with the Ori-Naïve mice, for their memory-impairing effects after the cocaine CPP retraining. One-way ANOVAs revealed significant group differences in the test 3 [test 3: F (3, 51) =3.627, P = 0.019] and the SR [SR: F (3, 51) = 3.859, P = 0.015], as well as a marginally significance among groups in the priming test [priming: F (3, 51) =2.613, P = 0.061]. The post hoc LSD tests further indicated that it is the R-E procedure, when compared with the E and Ori-Naïve mice, significantly suppressed cocaine memory in the test 3 (R-E vs. E: P = 0.044; R-E vs. Ori-Naïve: P = 0.002), the SR [R-E vs. E: P = 0.003; R-E vs. Ori-Naïve: P = 0.018], and the priming test (R-E vs. E: P = 0.014; R-E vs. Ori-Naïve: P = 0.034) [Figure 3]b. In summary, while all three behavioral manipulations induced long-term suppression of the high-dose cocaine CPP memory, the R-E procedure produced more pronounced memory deficits than the E and E-R procedures. Because the Ori-Naïve mice maintained good cocaine memory throughout the test 3, SR, and priming test, only the R-E procedure chronically protected mice from the reacquisition of the high-dose cocaine CPP memory.

Discussion

The main objective of this study is to examine whether the reconsolidation-update hypothesis or the extinction-facilitation hypothesis is the mechanism underlying the R-E procedure in the mouse model of cocaine-induced CPP. We systematically examined the efficacy of the E, R-E, and E-R procedures, which were then compared with the Ori-Naïve mice in the low-dose vs. the high-dose cocaine CPP memory. The most significant finding was that the efficacy of the three behavioral manipulations depended on the dose of cocaine used. In other words, while the E procedure exerted a long-lasting suppression of the low-dose cocaine CPP memory, the R-E procedure induced a more pronounced memory deficit than the E and E-R procedures in the long-term suppression of the high-dose cocaine CPP memory. However, none of the E, R-E, and E-R procedures rendered the mice back to their originally naïve state.

First, we observed that R-E was inefficient in the low-dose cocaine CPP paradigm, which is opposite to the long-term impairing effect of E on the SR and reinstatement of cocaine memory, and to the acute memory-impairing effect of E-R [Figure 1]b. It is suggested that a typical extinction training involves at least two competitive memory traces: the original CS-US trace and the new CS-noUS trace.^[5],[41] With our low-dose cocaine-induced CPP protocol, the E and E-R procedures appeared to induce the dominance of the inhibitory CS-noUS trace over the excitatory CS-US trace, which was displayed by blockade of the cocaine CPP memory. The R-E procedure, however, seemed to cause an opposite process such that the extinction training following the retrieval trial paradoxically strengthened the excitatory CS-US trace, which was manifested by the inefficacy of R-E in preventing mice's low-dose cocaine memory. The null effect of R-E adds to a growing literature indicating the variable nature of the R-E procedures, which can be specified by the boundary conditions. We will discuss these complex boundary conditions in detail below.^{[20],[31],[32],[33],[34],[35]}

Our null finding of R-E differs from Xue et al.'s^[22] finding that R-E prevented SR and reinstatement of the 10 mg/kg cocaine-induced CPP memory. These contrasting results further emphasize that R-E is not always practical and the apparent discrepancy between this study and Xue et al.^[22] can be attributed to potential differences such as species, duration of extinction, and other methodological issues. As for the species, we used male C57BL/6J mice as subjects, while Xue et al.^[22] used Sprague-Dawley rats. The species-dependent differences in pharmacokinetics and metabolism of cocaine between mice and rats could account for the apparent discrepancy. For example, following a cocaine (10 mg/kg, i.p.) injection, it took 5 min for mice^[44] but 30 min for rats^[45] to have their brain cocaine concentrations reach peak values. The half-life of cocaine (10 mg/kg, i.p.) to be eliminated from the plasma and brains was 16 min for mice, shorter than that of 25.4 min for rats.^[44],[46] We thus postulated that the same dose of cocaine stayed longer in the plasma and brains of rats than mice, which could lead to cocaine's varying rewarding effects on rats and mice. As for the duration of extinction, Xue et al. employed the eight daily 55-min extinction sessions for extinction training and used the eight daily sessions of a 10-min retrieval followed 10 min later by a 45-min extinction for the R-E procedure.^[22] In our study, the E group received the six daily 40-min extinction sessions; the R-E group received the six daily sessions of a 10-min retrieval followed 10 min later by a 30-min extinction; the E-R group received the six daily reversed sessions of a 30-min extinction followed 10 min later by a 10-min retrieval. Despite the differences, the magnitudes of cocaine CPP memory established in our prior^[43] and current studies are comparable to Xue et al.'s study.^[22] As for the methodological issues, we and Xue et al. both used the unforced choice procedure in a three-compartment apparatus for cocaine-induced CPP.^[22] To assess each subject's initial preference for a particular compartment, we adopted a more rigorous behavioral criterion than Xue et al.^[22] in the pretest. We excluded subjects with a staying time longer than 180 s in the middle and a time difference greater than 300 s between the two side compartments. On the other hand, Xue et al.^[22] excluded subjects with a staying time longer than 540 s in any of the three compartments. The behavioral criteria used in both studies ensured that the three-compartment apparatus was an unbiased apparatus for all of the subjects included, as Cunningham et al. suggested.^[47] Moreover, we used a biased subject assignment procedure, while Xue et al. used an unbiased subject assignment procedure.^[22] We assigned the initially nonpreferred compartment for each subject to be the cocaine-paired compartment. In contrast, Xue et al. randomly assigned a particular compartment for pairing with cocaine in a counterbalanced manner, regardless of each subject's initial preference. The use of an unbiased apparatus, however, is a more important choice than the choice between the biased vs. unbiased subject assignment procedure. Cunningham et al.^[48] showed that when mice were trained in the unbiased apparatus, ethanol-induced CPP was successfully established regardless of whether the drug was paired with the mice's initially preferred or non-preferred compartment. However, when the apparatus was biased, the CPP could only be established when ethanol was paired with the initially non-preferred compartment (i.e., the biased subject assignment procedure), and this asymmetrical outcome was likely due to the ceiling effect of measurement.^[48] Since this study and Xue et al.^[22] both using the unbiased apparatus, we agreed with Cunningham et al.'s^[48] argument that “In an unbiased apparatus, the type of subject assignment procedure (biased or unbiased) may not matter.” Our viewpoint was further supported by Blander et al.'s findings that there were no differences in the magnitudes of preference between the biased and unbiased subject assignment procedure in morphine-induced CPP in rats.^[49] Although we speculated that cocaine's varying rewarding effects on mice and rats were a significant factor for the apparent discrepancy (i.e., lack of efficacy of R-E on preventing mice's low-dose cocaine CPP memory), some minor methodological issues (e. g., how many sessions of conditioning, when to test SR and reinstatement of the cocaine memory) could also jointly contribute to the observed discrepancy between our and Xue et al.'s studies.^[22]

Next, we compared the effects of the E, R-E, and E-R procedures with the Ori-Naïve mice after the low-dose (10 mg/kg) cocaine CPP retraining. None of the behavioral procedures rendered the mice back to their originally naïve state. We first replicated our findings that both E and E-R suppressed the low-dose cocaine CPP memory when tested 24 h later [Figure 1]b and [Figure 2]b. The Ori-Naïve mice acquired cocaine CPP memory after one pair of the CPP new training, and their memory became diminished in the SR and cocaine priming tests. After one pair of the CPP retraining, only E, but not R-E or E-R, acutely suppressed cocaine CPP memory in the 24-h test. Our results imply that the qualitative differences in the neural circuits underlie the E, R-E, and E-R procedures. While R-E was ineffective in impairing the low-dose cocaine CPP memory, at least E and E-R successfully suppressed the initial training memory (i.e., excitatory CS-US trace), which was likely due to the dominance of the inhibitory CS-noUS trace. E even temporarily protected mice from the reacquisition of cocaine memory, suggesting that the dominance of the inhibitory CS-noUS trace could last for a while after the CPP retraining. Previous studies showed that inhibition of the nucleus accumbens (NAc) shell induced cocaine-seeking^[50] and alcoholic beer-seeking^[51] in the extinguished rats that underwent the standard E in a SA paradigm. In contrast, Millan et al. showed that inactivation of NAc shell did not promote alcoholic beer-seeking in the extinguished rats receiving the R-E training.^[29] Their findings supported the notion that R-E was qualitatively different from E. After the reacquisition of the low-dose cocaine CPP, we observed that none of the E, R-E, and E-R procedures suppressed the SR and cocaine-primed reinstatement. The reversed E-R procedure even paradoxically facilitated the cocaine-primed memory. Our CPP results echoed Millan et al.'s SA results that R-E facilitated rats' reacquisition of alcohol-seeking under the progressive ratio condition.^[29] Therefore, our findings supported that the reversed E-R procedure acted differently from the R-E procedure under certain circumstances. Whether or not the R-E and E-R procedures share the same neural circuits warrants future investigation.

Furthermore, we compared the effects of the E, R-E, and E-R procedures with the Ori-Naïve mice after the high-dose (40 mg/kg) cocaine CPP retraining. We found that all three behavioral manipulations induced a long-term suppression of the high-dose cocaine CPP memory after memory reacquisition. When compared with E and E-R, the R-E procedure induced more pronounced memory deficits in the initial training memory and the 24-h test, SR, and cocaine priming test after memory reacquisition [Figure 3]b. After one pair of the high-dose cocaine new learning, the Ori-Naïve mice's memory remained undamaged throughout the 24-h test, SR, and cocaine priming test. Only the R-E procedure induced memory deficits when compared with the Ori-Naïve mice.

Our most significant finding was that the dose of cocaine determined the efficacy of the E, R-E, and E-R procedures in preventing the SR and reinstatement of cocaine-associated memories. In other words, the E procedure induced a long-term suppression of the low-dose cocaine CPP memory, while the R-E procedure was better at inducing a long-term suppression of the high-dose cocaine CPP memory. Cocaine dose was the key independent variable in this study, and our prior and other studies have delineated a monotonic dose-response curve for cocaine-induced CPP memories.^{[43],[52],[53],[54]} That is, the higher the dose of cocaine used, the stronger the memory trace established. Therefore, the efficacy of the E and R-E procedures seemed to depend on the strength of cocaine CPP memory. Our findings that E inhibited the SR and cocaine-primed reinstatement of the low-dose cocaine CPP suggested that the E procedure was more efficient in suppressing weak cocaine-associated memory, usually in the beginning stage of drug abuse. In contrast, the R-E procedure specifically protected the mice from the SR and reinstatement during reacquisition of the high-dose cocaine CPP memory. Hence, after the reacquisition of the high-dose cocaine memory, the R-E procedure was the most efficient intervention to provide the mice with long-term protection against cocaine reinstatement. These results suggest that the R-E procedure is more efficient in suppressing strong cocaine-associated memory. We tried to explain the above results using the two competitive memory traces.^[5],[41] With our high-dose cocaine-induced CPP protocol, the original excitatory CS-US trace seemed to be stronger. It competed with the inhibitory CS-noUS trace in the E and E-R procedures. In contrast, the R-E procedure triggered reconsolidation that allows the extinction to modify the original memory into the new CS-noUS memory or facilitated the extinction training that allows the extinguished memory to be more resistant to SR and reinstatement. At first glance, our findings and explanations seem to contradict the statement “the ability to induce reconsolidation is inversely related to the strength of the memory trace,"^[35] which was supported by several studies.^{[15],[36],[41]} For example, Suzuki et al.^[36] systematically examined the boundary conditions such as memory strength, memory age, and retrieval duration (i.e., reexposure duration to the CS) and how these conditions interacted to affect the likelihood of memory reconsolidation occurring. In particular, they showed that weak memories, compared with the strong ones, were more prone to anisomycin's disruptive effects during reconsolidation.^[36] Nevertheless, our findings that the R-E procedure was better at suppressing the strong cocaine memory suggested two possibilities. First, the memory-suppressing effect of R-E might not be due to its disruption of the reconsolidation process (the reconsolidation-update hypothesis) but to its enhancement of the extinction process (the extinction-facilitation hypothesis).^[28] Second, some other factors interacted with memory strength to determine the efficacy of R-E. For example, prediction error (PE), which is defined as a discrepancy between actual and expected events, is determined by the interaction between original learning and retrieval.^[55],[56] Research showed that PE is necessary but not sufficient to induce memory reconsolidation.^[56] In our cocaine-induced CPP paradigm, the original learning was stronger in the high-dose cocaine conditioning (CPP magnitude: 337.39 ± 18.94 s) compared with the low-dose one (CPP magnitude: 248.46 ± 20.58 s). We speculated that the PE, the discrepancy between our retrieval trial and high-dose cocaine-induced strong original memory was big enough to trigger memory reconsolidation. This significant PE might be the reason why the R-E procedure was more efficient in suppressing the high-dose cocaine-induced CPP memory. On the other hand, after the reacquisition of the low-dose cocaine memory, the memory-impairing effect of E was limited to the 24-h test. The E procedure is considered as new learning in which the animals acquire the new association of CS-no US; thus, the animals inhibit their responding to the original CS-US association.^[57] In this study, it would be easier for E-induced new learning to suppress the weak original learning induced by the low-dose cocaine rather than to inhibit the strong associated learning in the reacquisition of low-dose cocaine CPP and in the original acquisition of high-dose cocaine CPP. Therefore, our results suggest that the mechanism underlying the R-E procedure is qualitatively different from the E procedure, as Millan et al. suggested.^[29]

Moreover, it is essential to note that the efficacy of the reversed E-R procedure was not equivalent to that of R-E in this study. The reversed E-R procedure, when compared with R-E, was more effective for its acute impairing effect on the low-dose cocaine memory and less effective for its long-term suppression effect on the high-dose cocaine memory. Paradoxically, the reversed E-R procedure enhanced the cocaine-primed reinstatement during the low-dose cocaine memory reacquisition. Despite being inferred from its absence,^[27] the reconsolidation-update hypothesis suggests that the capability of R-E to prevent the return of fear or drug memories is due to its disruption of the reconsolidation of original memory.^{[19],[21],[22],[23]} The observed differences between the R-E and E-R procedures support the reconsolidation-update hypothesis. That is, a retrieval trial must occur before extinction; thus, the original memory can be first reactivated by the retrieval and then modified by the extinction.^{[15],[26],[28],[58]} When extinction precedes a retrieval trial, the retrieval cannot act retrospectively on the encoding of the extinguished memory, so the reversed E-R procedure should yield different results. Our CPP findings agreed with the above ratiocination, which was different from Millan et al.'s findings that the R-E procedure and the reversed E-R procedure both impaired the reinstatement of alcoholic beer-seeking in a SA paradigm (the extinction–facilitation hypothesis).^[29] The contrasting results were likely due to the qualitative differences inherited in the CPP and SA paradigms. While drug-induced CPP is viewed as Pavlovian conditioning that measures the association between a drug's interoceptive reward cues and a drug-paired context, the drug SA measures the operant responding as drug motivation. This dichotomy, however, is oversimplified when one considers that CPP involves an operant-like response allowing an animal to move from a neutral compartment to one previously paired with a drug and that SA involves the Pavlovian cues such as lights or tones pairing with each drug infusion.^[54] Incongruous findings between CPP and SA have been documented. For example, LSD and buspirone elicited CPP but not SA, while pentobarbital and phencyclidine induced SA but not CPP.^[59] These results echoed human research findings that a drug's subjective feelings did not always parallel with human SA behaviors.^[60] It is argued that the difference between drug reward and drug motivation accounts for the discrepancies between CPP and SA. CPP measures pure drug reward (i.e., the duration spent in the drug-paired compartment), which is qualitatively different from the drug motivation measured in SA (i.e., operant responses such as lever press or nose poke). If a CPP paradigm wished to measure the motivation for a drug, it should have measured an increase in the number of entries into the drug-paired compartment.^[61] Some other non-motivational differences could also explain the discrepancies between CPP and SA. CPP typically produces a monotonic dose-response curve without a plateau,^[62] which differs from the biphasic inverted U-shaped curves in the fixed ratio of SA and from the monotonic curve with an apparent plateau at the high doses in the progressive ratio of SA.^[63] Hence, the divergent parameters measured in the CPP and SA paradigms could account for the discrepancy between our and Millan et al.'s findings.^[29]

Cahill and Milton proposed that deciphering the neurochemical and molecular mechanisms underlying the standard E and R-E (or its reversed sequence) procedures may help distinguish the reconsolidation-update hypothesis from the extinction-facilitation hypothesis.^[64] Prior findings that inactivation of NAc shell only prevented the expression of extinction in animals that received the standard E but not the R-E procedures suggested that these two behavioral manipulations recruited different neural substrates.^[29],[51] On the one hand, Flavell et al. showed that nimodipine, an L-type voltage-gated calcium channel blocker, prevented the detrimental effect of the R-E procedure on contextual fear memory owing to nimodipine's ability to block memory destabilization after their retrieval. These findings support the reconsolidation-update hypothesis that the efficacy of R-E depends on the destabilization of the initial memory after its retrieval.^[20] Based on a thorough literature review, Cahill and Milton indicated that the memory destabilization process recruited the GluN2B-NMDAR-mediated signaling in the amygdala. The authors thus suggested that if the reconsolidation-update hypothesis is correct, the R-E procedure should require GluN2B-NMDAR activity to induce memory destabilization.^[64] On the other hand, the standard E required an increase in calcineurin expression in the amygdala.^[64] Hence, if the extinction-facilitation hypothesis is correct, the R-E procedure should increase calcineurin expression in the amygdala.

In sum, our study emphasizes the role of cocaine dose in determining the efficacy of the E, R-E, and E-R procedures. For the potential clinical application, the E procedure is better at conquering weak cocaine memory in the early phase of cocaine addiction, while the R-E procedure is better at conquering strong cocaine memory in the late phase of cocaine addiction. However, several limitations of this study need to be addressed. First, the sex of the subjects was limited to males only; thus, we did not know whether the efficacy of the E, R-E, and E-R procedures in suppressing the low-dose versus the high-dose cocaine memory would be different in females. Indeed, a prior study showed that the R-E procedure was more effective in reducing fear reinstatement in females than in males.^[65] Second, unlike human research,^[55],[56] this study did not employ an independent non-invasive measure for memory destabilization other than the occurrence of the reconsolidation process itself. Hence, we could only infer the underlying mechanism from the behavioral outcome. Finally, future studies that examine the molecular mechanisms underlying the efficacy of the E, R-E, and E-R procedures, especially in the low-dose versus the high-dose cocaine-associated memory, should be conducted to further clarify the reconsolidation-update and extinction-facilitation hypotheses.

Conclusion

We are the first to highlight the role of cocaine dose in the efficacy of the E, R-E, and E-R procedures in cocaine-induced CPP memory in mice. Our results showed that standard E was more efficient in its long-lasting suppression of the low-dose cocaine CPP memory. At the same time, the R-E procedure was more efficient in its long-term suppression of the high-dose cocaine CPP memory. By adding cocaine dose as one of the boundary conditions for the R-E intervention, our results emphasize that R-E is far more complicated than we previously thought. Therefore, we urge future studies to focus on deciphering the neurochemical and molecular mechanisms underlying the effects of the E, R-E, and E-R procedures on various doses of cocaine-induced CPP memories.

Ethical approval

The animal protocol (NCKU IACUC approval number: 104116) for the experimental procedures and housing conditions was reviewed and approved by the Institutional Animal Care and Use Committee of the National Cheng Kung University College of Medicine. All experiments conformed to the Guide for the Care and Use of Laboratory Animals in the National Institutes of Health Guideline for Animal Research.

Acknowledgment

We thank all of the Hu lab members for their technical assistance.

Financial support and sponsorship

This research was supported by the Ministry of Science and Technology (MOST) in Taiwan (Grants 104-2410-H-006-025-MY3, 105-2410-H-006-019-MY2 to SSH). There was no role of the funding source in the study design, the collection, analysis, and interpretation of the data, the writing, and submission of the paper.

Conflicts of interest

There are no conflicts of interest.

References

1.	Glickman SE. Perseverative neural processes and consolidation of the memory trace. Psychol Bull 1961;58:218-33.
2.	McGaugh JL. Time-dependent processes in memory storage. Science 1966;153:1351-8.
3.	Myers KM, Davis M. Behavioral and neural analysis of extinction. Neuron 2002;36:567-84.
4.	Rescorla RA, Heth CD. Reinstatement of fear to an extinguished conditioned stimulus. J Exp Psychol Anim Behav Process 1975;1:88-96.
5.	Bouton ME. Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychol Bull 1993;114:80-99.
6.	Rescorla RA. Spontaneous recovery. Learn Mem 2004;11:501-9.
7.	Quirk GJ. Memory for extinction of conditioned fear is long-lasting and persists following spontaneous recovery. Learn Mem 2002;9:402-7.
8.	Burdick CK, James JP. Spontaneous recovery of conditioned suppression of licking by rats. J Comp Physiol Psychol 1970;72:467-70.
9.	Bouton ME, King DA. Contextual control of the extinction of conditioned fear: Tests for the associative value of the context. J Exp Psychol Anim Behav Process 1983;9:248-65.
10.	Bouton ME, Bolles RC. Contextual control of the extinction of conditioned fear. Learn Motiv 1979;10:445-66.
11.	Bouton ME, Bolles RC. Role of conditioned contextual stimuli in reinstatement of extinguished fear. J Exp Psychol Anim Behav Process 1979;5:368-78.
12.	Schafe GE, Nadel NV, Sullivan GM, Harris A, LeDoux JE. Memory consolidation for contextual and auditory fear conditioning is dependent on protein synthesis, PKA, and MAP kinase. Learn Mem 1999;6:97-110.
13.	Nader K, Schafe GE, Le Doux JE. Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature 2000;406:722-6.
14.	Nader K, Schafe GE, LeDoux JE. The labile nature of consolidation theory. Nat Rev Neurosci 2000;1:216-9.
15.	Tronson NC, Taylor JR. Molecular mechanisms of memory reconsolidation. Nat Rev Neurosci 2007;8:262-75.
16.	Debiec J, Ledoux JE. Disruption of reconsolidation but not consolidation of auditory fear conditioning by noradrenergic blockade in the amygdala. Neuroscience 2004;129:267-72.
17.	Duvarci S, Nader K, LeDoux JE. Activation of extracellular signal-regulated kinase- mitogen-activated protein kinase cascade in the amygdala is required for memory reconsolidation of auditory fear conditioning. Eur J Neurosci 2005;21:283-9.
18.	Frankland PW, Ding HK, Takahashi E, Suzuki A, Kida S, Silva AJ. Stability of recent and remote contextual fear memory. Learn Mem 2006;13:451-7.
19.	Monfils MH, Cowansage KK, Klann E, LeDoux JE. Extinction-reconsolidation boundaries: Key to persistent attenuation of fear memories. Science 2009;324:951-5.
20.	Flavell CR, Barber DJ, Lee JL. Behavioural memory reconsolidation of food and fear memories. Nat Commun 2011;2:504.
21.	Schiller D, Monfils MH, Raio CM, Johnson DC, Ledoux JE, Phelps EA. Preventing the return of fear in humans using reconsolidation update mechanisms. Nature 2010;463:49-53.
22.	Xue YX, Luo YX, Wu P, Shi HS, Xue LF, Chen C, et al. A memory retrieval-extinction procedure to prevent drug craving and relapse. Science 2012;336:241-5.
23.	Milton AL, Everitt BJ. Neuroscience. Wiping drug memories. Science 2012;336:167-8.
24.	Rescorla RA. Deepened extinction from compound stimulus presentation. J Exp Psychol Anim Behav Process 2006;32:135-44.
25.	Janak PH, Corbit LH. Deepened extinction following compound stimulus presentation: Noradrenergic modulation. Learn Mem 2011;18:1-10.
26.	Torregrossa MM, Taylor JR. Learning to forget: Manipulating extinction and reconsolidation processes to treat addiction. Psychopharmacology (Berl) 2013;226:659-72.
27.	Lattal KM, Wood MA. Epigenetics and persistent memory: Implications for reconsolidation and silent extinction beyond the zero. Nat Neurosci 2013;16:124-9.
28.	Hutton-Bedbrook K, McNally GP. The promises and pitfalls of retrieval-extinction procedures in preventing relapse to drug seeking. Front Psychiatry 2013;4:14.
29.	Millan EZ, Milligan-Saville J, McNally GP. Memory retrieval, extinction, and reinstatement of alcohol seeking. Neurobiol Learn Mem 2013;101:26-32.
30.	Baker KD, McNally GP, Richardson R. Memory retrieval before or after extinction reduces recovery of fear in adolescent rats. Learn Mem 2013;20:467-73.
31.	Chan WY, Leung HT, Westbrook RF, McNally GP. Effects of recent exposure to a conditioned stimulus on extinction of Pavlovian fear conditioning. Learn Mem 2010;17:512-21.
32.	Costanzi M, Cannas S, Saraulli D, Rossi-Arnaud C, Cestari V. Extinction after retrieval: Effects on the associative and nonassociative components of remote contextual fear memory. Learn Mem 2011;18:508-18.
33.	Soeter M, Kindt M. Disrupting reconsolidation: Pharmacological and behavioral manipulations. Learn Mem 2011;18:357-66.
34.	Auber A, Tedesco V, Jones CE, Monfils MH, Chiamulera C. Post-retrieval extinction as reconsolidation interference: Methodological issues or boundary conditions? Psychopharmacology (Berl) 2013;226:631-47.
35.	Struik RF, De Vries TJ, Peters J. Detrimental effects of a retrieval-extinction procedure on nicotine seeking, but not cocaine seeking. Front Behav Neurosci 2019;13:243.
36.	Suzuki A, Josselyn SA, Frankland PW, Masushige S, Silva AJ, Kida S. Memory reconsolidation and extinction have distinct temporal and biochemical signatures. J Neurosci 2004;24:4787-95.
37.	Milekic MH, Alberini CM. Temporally graded requirement for protein synthesis following memory reactivation. Neuron 2002;36:521-5.
38.	Cammarota M, Bevilaqua LR, Medina JH, Izquierdo I. Retrieval does not induce reconsolidation of inhibitory avoidance memory. Learn Mem 2004;11:572-8.
39.	Muravieva EV, Alberini CM. Limited efficacy of propranolol on the reconsolidation of fear memories. Learn Mem 2010;17:306-13.
40.	Wang SH, de Oliveira Alvares L, Nader K. Cellular and systems mechanisms of memory strength as a constraint on auditory fear reconsolidation. Nat Neurosci 2009;12:905-12.
41.	Eisenberg M, Kobilo T, Berman DE, Dudai Y. Stability of retrieved memory: Inverse correlation with trace dominance. Science 2003;301:1102-4.
42.	Ma X, Zhang JJ, Yu LC. Post-retrieval extinction training enhances or hinders the extinction of morphine-induced conditioned place preference in rats dependent on the retrieval-extinction interval. Psychopharmacology (Berl) 2012;221:19-26.
43.	Hu SS, Liu YW, Yu L. Medial prefrontal cannabinoid CB1 receptors modulate consolidation and extinction of cocaine-associated memory in mice. Psychopharmacology (Berl) 2015;232:1803-15.
44.	Benuck M, Lajtha A, Reith ME. Pharmacokinetics of systemically administered cocaine and locomotor stimulation in mice. J Pharmacol Exp Ther 1987;243:144-9.
45.	Pan HT, Menacherry S, Justice JB Jr. Differences in the pharmacokinetics of cocaine in naive and cocaine-experienced rats. J Neurochem 1991;56:1299-306.
46.	Ma F, Falk JL, Lau CE. Within-subject variability in cocaine pharmacokinetics and pharmacodynamics after intraperitoneal compared with intravenous cocaine administration. Exp Clin Psychopharmacol 1999;7:3-12.
47.	Cunningham CL, Gremel CM, Groblewski PA. Drug-induced conditioned place preference and aversion in mice. Nat Protoc 2006;1:1662-70.
48.	Cunningham CL, Ferree NK, Howard MA. Apparatus bias and place conditioning with ethanol in mice. Psychopharmacology (Berl) 2003;170:409-22.
49.	Blander A, Hunt T, Blair R, Amit Z. Conditioned place preference: An evaluation of morphine's positive reinforcing properties. Psychopharmacology (Berl) 1984;84:124-7.
50.	Peters J, LaLumiere RT, Kalivas PW. Infralimbic prefrontal cortex is responsible for inhibiting cocaine seeking in extinguished rats. J Neurosci 2008;28:6046-53.
51.	Millan EZ, Furlong TM, McNally GP. Accumbens shell-hypothalamus interactions mediate extinction of alcohol seeking. J Neurosci 2010;30:4626-35.
52.	Raybuck JD, McCleery EJ, Cunningham CL, Wood MA, Lattal KM. The histone deacetylase inhibitor sodium butyrate modulates acquisition and extinction of cocaine-induced conditioned place preference. Pharmacol Biochem Behav 2013;106:109-16.
53.	Tzeng WY, Chuang JY, Lin LC, Cherng CG, Lin KY, Chen LH, et al. Companions reverse stressor-induced decreases in neurogenesis and cocaine conditioning possibly by restoring BDNF and NGF levels in dentate gyrus. Psychoneuroendocrinology 2013;38:425-37.
54.	Green TA, Bardo MT. Opposite regulation of conditioned place preference and intravenous drug self-administration in rodent models: Motivational and non-motivational examples. Neurosci Biobehav Rev 2020;116:89-98.
55.	Sevenster D, Beckers T, Kindt M. Prediction error governs pharmacologically induced amnesia for learned fear. Science 2013;339:830-3.
56.	Sevenster D, Beckers T, Kindt M. Prediction error demarcates the transition from retrieval, to reconsolidation, to new learning. Learn Mem 2014;21:580-4.
57.	Bouton ME. Context, ambiguity, and unlearning: Sources of relapse after behavioral extinction. Biol Psychiatry 2002;52:976-86.
58.	Nader K, Hardt O. A single standard for memory: The case for reconsolidation. Nat Rev Neurosci 2009;10:224-34.
59.	Bardo MT, Bevins RA. Conditioned place preference: What does it add to our preclinical understanding of drug reward? Psychopharmacology (Berl) 2000;153:31-43.
60.	Fischman MW. Relationship between self-reported drug effects and their reinforcing effects: Studies with stimulant drugs. NIDA Res Monogr 1989;92:211-30.
61.	Bardo MT, Miller JS, Neisewander JL. Conditioned place preference with morphine: The effect of extinction training on the reinforcing CR. Pharmacol Biochem Behav 1984;21:545-9.
62.	Brown RW, Kirby SL, Denton AR, Dose JM, Cummins ED, Drew Gill W, et al. An analysis of the rewarding and aversive associative properties of nicotine in the neonatal quinpirole model: Effects on glial cell line-derived neurotrophic factor (GDNF). Schizophr Res 2018;194:107-14.
63.	Green TA, Gehrke BJ, Bardo MT. Environmental enrichment decreases intravenous amphetamine self-administration in rats: Dose-response functions for fixed- and progressive-ratio schedules. Psychopharmacology (Berl) 2002;162:373-8.
64.	Cahill EN, Milton AL. Neurochemical and molecular mechanisms underlying the retrieval-extinction effect. Psychopharmacology (Berl) 2019;236:111-32.
65.	Chen W, Lin XY, Li JJ, Zhang WX, Sun N, Zheng XF. Gender difference in retrieval-extinction of conditioned fear memory. Acta Psychol Sin 2021;53:1082-93.