N-acetylcysteine Treatment During Acute Stress Prevents Stress-induced Augmentation of Addictive Drug Use and Relapse

Abstract

Converging epidemiological studies show that a life-threatening event increases the incidence of post-traumatic stress disorder (PTSD), which carries 30–50% comorbidity with substance use disorders (SUDs). Such comorbidity results in greater drug use and poorer treatment outcomes. There is overlap between the enduring synaptic neuroadaptations produced in nucleus accumbens core (NAcore) by acute restraint stress and cocaine self-administration. Because of these coincident neuroadaptations, we hypothesized that an odor paired with acute restraint stress would reinstate drug seeking, and chose two mechanistically distinct drugs of abuse to test this hypothesis, alcohol and cocaine. Rats were trained to self-administer either drug beginning three weeks after odor pairing with acute stress or sham, and acute restraint stress increased alcohol consumption. Following context extinction training, the stress-paired odor reinstated both alcohol and cocaine seeking, while an unpaired odor had no effect. N-acetylcysteine (NAC) restores drug and stress-induced reductions in glial glutamate transporter-1, and has proven effective at reducing cue-induced reinstatement of drug seeking. We administered NAC for five days prior to reinstatement testing and abolished the capacity of the stress-paired odor to increase alcohol and cocaine seeking. Importantly, daily NAC given during or just following experiencing acute restraint stress also prevented the capacity of stress-paired odors to reinstate alcohol and cocaine seeking, and prevented stress-induced deficits in behavioral flexibility. These data support using daily NAC treatment during or immediately after experiencing a strong acute stress to prevent subsequent conditioned stress responding, in particular relapse and cognitive deficits induced by stress-conditioned stimuli.

Introduction

Exposure to acute, life-threatening traumatic events, such as occurs in combat or rape, increases the vulnerability to develop post-traumatic stress disorder (PTSD). Furthermore, a diagnosis of PTSD substantially increases the probability of developing substance use disorders (SUDs). For example, veterans diagnosed with PTSD have a 30–50% comorbidity with SUDs (1), which is ~10-fold higher than the 3.5% diagnosis in the general population (2, 3). Moreover, the severity of stress exposure correlates with greater substance use (4) and is a risk factor in triggering relapse (5, 6). Patients with comorbid PTSD/SUDs show poorer treatment outcomes than patients diagnosed with either PTSD or SUDs alone (7, 8). Even though the link between stress and SUDs is evident, effective treatments for comorbid PTSD and SUDs are not available. However, in a recent double-blind, placebo controlled clinical trial with veterans comorbid for PTSD and SUDs, treatment with N-acetylcysteine (NAC) proved effective at reducing symptoms of PTSD and drug craving (9). Meta-analyses of NAC trials support the use of NAC in ameliorating drug craving, although in some trials NAC failed to reduce craving (10, 11).

Rodent models combining stress and substance use reveal that previous exposure to acute or chronic stress predisposes animals to the behavioral effects of psychostimulants and opioids, including the development of behavioral sensitization and drug self-administration (12–14). Acute restraint stress produces long-lasting physiological and morphological adaptations at glutamatergic synapses in the nucleus accumbens core (NAcore) that parallel those produced by drug self-administration and augments the acquisition of cocaine self-administration (13). In particular, 3 weeks after acute restraint stress in rats, the function and expression of the glial glutamate transporter-1 (GLT-1), is reduced, and restoring stress-induced reductions in GLT-1 by daily treatment with ceftriaxone prevents stress-facilitated acquisition of cocaine use (13). Daily NAC treatment also restores levels of GLT-1 and reduces craving for addictive drugs in some clinical trials (11). In parallel with the clinical laboratory, NAC prevents cue- and context-induced drug seeking in rodents withdrawn from cocaine, nicotine, heroin, cannabis or alcohol self-administration, and the efficacy of NAC at reducing drug-seeking is linked to restoring GLT-1 (15).

We established a rat model of conditioned stress-induced alcohol and cocaine relapse. Acute restraint stress was paired with a novel odor 3 weeks prior to beginning drug self-administration. After self-administration and extinction training, the stress-paired odor was used to reinstate drug seeking. Alcohol and cocaine were used as examples of two addictive drugs having distinct molecular sites of action and patterns of use (16, 17), and are among the most abused drugs by veterans comorbid for PTSD and SUDs (9). We show that drug seeking induced by stress-conditioned odor was prevented when daily NAC was administered during extinction training, during stress exposure, or immediately after stress exposure. Further, we show that NAC treatment before and after restraint stress prevented deficits in cognitive flexibility induced by acute stress. These data point to the possible prophylactic use of NAC for reducing comorbidity between PTSD and SUDs in individuals experiencing acute stress, such as rape or combat, which are known to increase the incidence of PTSD.

Methods

Animal housing and stress

Male Sprague-Dawley rats (250 g; Charles River Laboratories, Raleigh, NC, USA) were double-housed with a 12:12 hr dark/light cycle. The animals were approximately 2 months old (± 1 week) and all experiments were conducted during the light cycle. Rats received food and water ad libitum and were allowed at least 1 week to acclimate to the vivarium before any treatment. The acute stress group was restrained for 2 hours (between 10:00 and 14:00 h) in restraining devices paired with an odor cue (lemon or sandalwood oil; Wyndmere Naturals Inc. Minneapolis, MN; wyndmerenaturals.com) in a crossover design where 50% of the rats were paired with one or the other odor. Sham animals were left undisturbed in their home cages and exposed to an odor (1 mL of lemon or sandalwood oil in a petri dish) for 2 hours. The flat-bottomed plexiglas restrainers (3.25 inches in diameter and 8 inches long) were designed so that the rats’ tails emerged from the rear. The animals appeared healthy as shown by their coat texture and no difference in body weight was detected between sham and stress exposed rats at the time animals were used for behavior. Experimental procedures were approved by the Animal Care and Use Committee of the Medical University of South Carolina and performed in accordance with National Institutes of Health guidelines.

Animal surgeries

Two weeks after stress or sham (control) exposure, animals in the cocaine self-administration group underwent surgery to implant intravenous indwelling catheters. Rats were surgically implanted with intravenous silastic catheters in the right jugular vein under anesthesia with ketamine (87.5 mg/kg, i.m.) and xylazine (5 mg/kg, i.m.). Ketorolac (3 mg/kg, i.p.) was administered prior to surgery and as needed postoperatively to provide analgesia. Prophylactic antibiotic (Cefazolin 10 mg/0.1 ml, i.v.) was administered during surgery. The catheter was secured to the vein with silk sutures and was passed subcutaneously to the middle of the back where it terminated in a connector consisting of a modified 22-gauge cannula (Plastics One, Roanoke, VA) embedded in dental cement attached to surgical mesh (Atrium, Hudson, NH). Catheters were flushed daily with heparin (0.1 mL of 100 IU) until the end of cocaine self-administration, and catheter patency was confirmed at the end of each study.

Alcohol and cocaine self-administration and reinstatement

Alcohol self-administration

rats were trained to self-administer ethanol by first exposing them to an intermittent two-bottle choice drinking initiation paradigm for a period of two weeks. Three days per week, two bottles, one containing water and the other containing 20% ethanol, were placed on the cage. The bottles and the rats were weighed and g/kg of ethanol consumed were calculated after each day of the two bottle choice testing. The purpose of this phase was to acclimate the rats to the taste and smell of the alcohol solution. The day following the last two-bottle choice session, the rats were placed in operant chambers and trained to self-administer alcohol on an FR1 schedule of reinforcement based on previously published methods (18). Each active lever press activated the syringe pump to delivered ~ 45 μl of a liquid solution over a 1.5 second period. During alcohol delivery, the stimulus light above the active lever was illuminated and the tone was presented. Following each delivery, a 4-sec timeout period was initiated during which additional active lever presses were recorded but had no programmed consequences. Initially, rats were trained on a 20% ethanol solution in 30 min daily sessions. After stable responding for 20% ethanol (approximately 8–10 sessions) was reached, the concentration of ethanol was reduced to 10% for the remaining sessions (12–16 sessions). After an additional week of daily self-administration sessions, extinction training began in which active lever presses no longer resulted in the delivery of the light/tone stimulus or alcohol.

Cocaine self-administration

Following surgery rats were trained to acquire operant responding for food in two 2-hour sessions prior to beginning 2-hour daily cocaine self-administration paired with light/tone cues. Rats self-administered cocaine (0.2 mg/0.05 ml infusion, ~0.5–0.67 mg/kg/infusion; kindly provided by the National Institute of Drug Abuse) until reaching 10 days with >10 infusions, followed by 10 days of extinction. Training was performed in standard operant chambers containing house and cue lights, tone, and two retractable levers (Med Associates, St. Albans, VT). Subjects were trained on a fixed-ratio 1 schedule of reinforcement paired with light and tone (78 dB, 4.5 kHz) and followed by a 20-s timeout period signaled by absence of the house light. During extinction, active lever presses no longer resulted in delivery of cocaine or cues. Extinction criteria were met when animals averaged <20 active lever presses for 2 days prior to reinstatement testing.

Reinstatement

After extinction training the rats were exposed to the conditioned cue (stress-paired odor) or unconditioned cue (unpaired odor) in a random crossover design where cues (lemon or sandalwood odors) where counterbalanced. The odor containers were placed in the operant box for 5 min before loading the animal and initiating the reinstatement test. During the test, lever presses were recorded but yielded no programmed consequences (e.g. no alcohol or cocaine was provided and no cues associated with the drug were presented).

NAC administration schedules

Several timing variations of NAC were chosen help identify times in the protocol where NAC exposure might reduce conditioned stress reinstatement. Specifically, repeated NAC was given before, during and after stress because this protocol restores GLT-1, while acute pretreatment reduces synaptic glutamate release by indirectly stimulating glutamate autoreceptors (15).

NAC before conditioned odor reinstatement

A group of rats were administered either saline (VEH) or NAC (100 mg/kg, i.p.; Sigma-Aldrich, St. Louis, MO, USA) for 4-days prior to and the day of reinstatement testing. NAC was prepared daily in 27 mg/ml NaOH in saline and adjusted to pH = 7.2. NAC or VEH was administered 2 hr prior to the operant extinction or reinstatement session.

NAC administration before, during and after restraint stress

A group of rats were treated with NAC or VEH for 4-days prior to, the day of and 4 days following acute restraint stress (NAC X 10 days). The treatments were administered at the same time each day and 2 hr prior to acute restraint stress. Two additional groups were examined. One group was treated with NAC or VEH for 4-days prior to and the day of acute restraining stress, and another group was treated with NAC or VEH for 5-days beginning the day after acute restraint stress.

Assessment of behavioral flexibility

Behavioral flexibility was examined using an operant set-shifting procedure (19). In brief, rats were first trained to respond to a light cue in order to receive a reinforcer. Once this rule was acquired, the contingencies of the task changed to a new rule that required the rat to ignore the light cue and respond only on one lever to receive a reinforcer. Behavioral flexibility was then assessed by measuring the number of trials required to learn the new rule. The primary dependent variables in these experiments were the number of completed trials required and errors made to achieve criterion performance of 10 consecutive correct choices.

Statistics

All group sizes were determined using power analyses to ensure sufficient power for statistical evaluation using a significance level of p< 0.05. Data were statistically evaluated using Prism 8.0 (GraphPad, Inc) software to conduct two-tailed Student’s t-tests for single comparisons, and an analysis of variance (ANOVA) for multiple comparisons. If the interaction score in a two- or three-way ANOVA was significant, a Bonferroni post hoc test was performed. For within subject comparisons repeated measures (RM) ANOVAs were conducted.

Results

Stress exposure increases alcohol but not cocaine self-administration

We first examined the long-term effects of acute restraint stress in alcohol and cocaine intake. All rats were exposed during restraint or sham stress to a conditioning odor cue (lemon or sandalwood oil scent). Three weeks following restraint or sham stress, rats were trained to self-administer alcohol or cocaine followed by extinction training (Fig 1a). Exposure to restraint stress increased alcohol self-administration as indicated by increased active lever presses (Fig 1b) compared to sham animals during the last 10-session of alcohol self-administration (2-way RM-ANOVA time F_(5,115)= 69.97, p< 0.001, stress v sham F_(1,22)= 68.54, p< 0.001, interaction F_(9,198)= 5.803, p< 0.001). Stress exposure also increased inactive lever presses (2-way RM-ANOVA time F_(6,129)= 3.241, p= 0.0056, stress v sham F_(1,22)= 28.99, p< 0.001, interaction F_(9,198)= 5.174 p< 0.001), but the effect was irregular, occurring only on self-administration days 5, 6, and 10 and the greatest magnitude of this difference was 6 lever presses (Fig S1c). Similarly, stressed rats took more alcohol reinforcements than sham rats (2-way RM-ANOVA time F_(5,115)= 80.64, p< 0.001, stress v sham F_(1,22)= 58.38, p< 0.001, interaction F_(9,198)= 5.512, p<0.005; Fig 1b). No difference between treatment groups (stress v sham) for alcohol intake during 2-bottle choice or during the first 10 sessions of alcohol self-administration (Fig S1a and b).

Figure 1. Stress pre-exposure leads to increased vulnerability to alcohol but not cocaine self-administration.

(a) Timeline of NAC and vehicle (VEH) treatment before and during extinction and stress-paired odor-induced reinstatement in stressed versus sham animals after alcohol or cocaine self-administration. This group design corresponds to data in Figures 1 and 2. (b) Potentiated alcohol self-administration in stressed compared to sham animals. Left panel, active lever presses over the last 10-days of alcohol self-administration and extinction training. See Figures S1b for the first 10 days of self-administration. Right panel, reinforcers taken during the last 10 days of alcohol self-administration. *p< 0.05, comparing Sham to Stress treatment groups. (c) Lack of effect by stress on cocaine self-administration and extinction active lever pressing. Left panel, active lever presses over cocaine self-administration and extinction training (rats underwent food training prior to beginning cocaine self-administration). Right panel, infusions over 10-days of cocaine self-administration. Note that the high lever pressing on the first day of self-administration resulted from food training rats for two days prior to beginning cocaine (see Methods).

All data are shown as mean ± SEM.

In contrast with alcohol, restraint stress did not alter cocaine self-administration compared to sham (Fig 1c) on either active (2-way RM-ANOVA time F_(17,225)= 0.67, p= 0.667, stress v sham F_(1,25)= 0.50, p= 0.487, interaction F_(17,225)= 1.02, p= 0.425), inactive lever pressing (Fig S1b); regardless of whether or not cocaine intake calculated by body weight (Fig S1e). However, food training prior to beginning cocaine self-administration likely obfuscated the previously reported restraint stress-induced facilitation in the acquisition of cocaine self-administration(13). There was no difference in cocaine infusions between stress and sham (2-way RM-ANOVA time F_(9,225)= 3.49, p= 0.001, stress v sham F_(1,25)= 0.001, p= 0.963, interaction F_(17,225)= 2.06, p= 0.034). Although a significant interaction was identified between treatment and time, a Bonferroni post hoc did not reveal treatment group differences at any time.

Stress exposure increased active lever pressing on multiple days during extinction in alcohol trained rats (2-way RM-ANOVA time F_(5,105)= 161.1, p< 0.001, stress v sham F_(1,22)= 70.83, p< 0.001, interaction F_(9,198)= 8.35, p< 0.001; Fig 1b). Stress did not affect extinction active lever pressing in cocaine trained rats, although there was a significant effect of time due to the extinction of active lever pressing over multiple training sessions (2-way RM-ANOVA: time F_(7,175)= 21.39, p< 0.001, stress v sham F_(1,25)= 0.001, p= 0.983, interaction F_(17,175)= 0.63, p= 0.731). There was no effect of stress on inactive lever pressing during extinction training (Fig S1d).

Stress-conditioned odor-induced alcohol and cocaine seeking is abolished by NAC treatment during extinction and reinstatement

The rats used in Fig 1 were divided into NAC and vehicle (VEH) treatment groups such that rats in the NAC group did not differ in total alcohol (Fig 2a; 2-way ANOVA Stress v Sham F_(1,20)= 347.7, p< 0.001, NAC v Veh F_(1,20)= 0.57, p= 0.460, interaction F_(1,20)= 3.43, p= 0.079) or cocaine (Fig 2b; 1-way ANOVA F_(2,24)= 0.16, p= 0.853) intake from rats in the VEH group. However, as indicated in Figure 1, rats that underwent stress exposure took greater total alcohol reinforcers than sham rats, although the NAC and VEH groups had equivalent intake within the stress and sham groups (Fig 2a). During extinction training and odor-induced reinstatement trials, rats were injected daily with either NAC (100 mg/kg, ip; (20, 21)) or VEH (Fig 1a). Alcohol-trained rats underwent a single reinstatement trial in the presence of either stress/sham-paired or unpaired odor. Only alcohol-treated rats with a history of stress exposure and daily VEH injections reinstated drug seeking in the presence of the stress-paired odor (Fig 2c; 3-way RM-ANOVA Ext v Rst F(1,20)= 56.78, p< 0.001, stress v sham F(1,20)= 55.33, p< 0.001, VEH v NAC F(1,20)= 47.57, p< 0.0001, interaction F(1,20)= 82.07, p<0.001.) NAC treatment abolished drug seeking produced by the stress-paired odor. The stress unpaired odor did not reinstate active lever pressing in any treatment group, regardless of pretreatment with stress, sham, NAC or VEH (Fig 2c; 3-way RM-ANOVA Ext v Rst F(1,20)= 2.32, p=0.144, sham v stress F(1,20)= 0.35, p= 0.559, VEH v NAC F(1,20)= 0.55, p= 0.466, interaction F(1,20)= 0.69, p= 0.416) Inactive lever pressing was unaltered by either odor in any treatment group in alcohol-trained rats (Fig S2a).

Figure 2. Stress-paired odor induced alcohol and cocaine seeking in stressed rats is blocked by NAC treatment during extinction and reinstatement.

The sham and stress alcohol groups were divided after self-administration into VEH and NAC treatment groups. (a) More total alcohol reinforcers were consumed in rats that underwent restraint stress, but the levels were equivalent between the NAC and VEH groups for sham and stress rats. *p< 0.05 comparing stress to sham. (b) The total number of infusions was equivalent between all three cocaine treatment groups. (c) The stress-paired odor reinstated alcohol seeking in VEH rats, and this effect was abolished by NAC treatment. The unpaired odor did not induce alcohol seeking (bottom panel). (d) The stress-paired odor reinstated cocaine seeking in VEH, and this effect was abolished in NAC treated rats. The unpaired odor did not induce cocaine seeking (bottom panel). All data are shown as mean ± SEM. N shown in bars.

*p<0.05, comparing sham to stress treatment, +p<0.05 compared to extinction within each group

After extinction training, cocaine-trained rats underwent two reinstatement trials separated by 2 days of further extinction training using a random crossover design, such that all rats had a reinstatement session with a stress- paired odor and an unpaired odor (Fig 1a). Only stress VEH treated rats significantly reinstated to the stress-paired odor, and NAC abolished this response (Fig 2d; 2-way RM-ANOVA Ext v Rst F_(1,24)= 4.56, p= 0.043, treatment F_(2,24)= 2.07, p= 0.148, interaction F_(2,24)= 4.70, p= 0.019). No treatment group showed increased active lever pressing in response to exposure to the unpaired odor (Fig 2d; Ext v Rst F_(1,24)= 1.84, p= 0.188, treatment F_(2,24)= 0.87, p= 0.433, interaction F_(2,24)= 0.04, p= 0.964). Also, neither odor altered inactive lever presses in any group of cocaine-trained rats (Fig S2b).

NAC treatment during stress prevented stress odor-induced alcohol seeking

Three different NAC treatment experiments were conducted to evaluate whether NAC treatment during or after restraint stress altered the effects of stress-paired odor on alcohol and cocaine seeking. The first experiment was conducted only in alcohol-trained subjects and NAC was administered daily 4-days prior to, during, and for 5-days following restraint stress (NAC 100 mg/kg X 10 days) (Fig 3a). NAC treatment during this time period reduced the stress-induced augmentation of active lever presses and total alcohol reinforcements obtained compared to the VEH stress rats over the last 10 days of alcohol self-administration, to a level of alcohol use equivalent to sham-treated rats (Fig 3b; 2-way RM-ANOVA treatment F_(3,28)= 13.72, p< 0.001, time F_(9,s52)= 248.9, p< 0.001, interaction F_(27,252)= 2.575, p< 0.001; and Fig 3c; 1-way ANOVA F_(3,28)= 226.7, p< 0.001). No statistical difference was observed between groups during 2-bottle choice or the first 10-session of alcohol self-administration (Fig S3a,b). NAC treatment regimen also prevented stress odor-induced reinstatement of alcohol seeking (Fig 3d; 3-way RM-ANOVA Ext v Rst F(1,28)= 37.85, p< 0.0001, treatment F(1,28)= 46.73, p< 0.001, NAC F(1,28)= 69.06, p< 0.001, 3-way interaction F(1,28)= 21.29, p<0.001), without altering inactive lever pressing (Fig S3c).

Figure 3. NAC treatment before, after and during stress prevented stress-paired odor induced alcohol seeking.

(a) Schematic outlines of NAC treatment before and after the 2-hours of acute restraint stress. (b) Potentiated alcohol self-administration in stressed animals compared to all treatments groups. Active lever presses over the last 10-days of alcohol self-administration, *p< 0.05 comparing stress animals to all treatment groups. (c) NAC treatment reduced the stress-potentiated total number of alcohol reinforcers consumed. (d) NAC prevented the capacity of a stress-paired odor to reinstate alcohol seeking. All data are shown as mean ± SEM. N shown in bars.

*p< 0.05 compared all other treatment groups, +p< 0.05 compared to extinction

NAC treatment during or after stress prevented stress odor-induced alcohol and cocaine seeking

In the next two experiments we wanted to determine if NAC was still effective at preventing stress-paired odor-induced reinstatement if given during or after restraint stress in both alcohol- and cocaine-trained rats (Fig 4a). Two different NAC treatment protocols were employed. NAC was administered daily for four days prior to and the day of restraint stress or injected daily over 5 days after restraint stress (Fig 4a). Control VEH injections were administered in parallel only in the group administered NAC after restraint stress.

Figure 4. NAC treatment during or after stress prevented stress-paired odor-induced alcohol and cocaine seeking.

(a) Schematic outlines of NAC treatment before and after the 2-hours of acute restraint stress. (b) Potentiated alcohol self-administration in stressed animals compared to all treatments groups. Active lever presses over the last 10-days of alcohol self-administration, *p< 0.05 comparing stress animals to all treatment groups. (c) Lack of effect by stress on cocaine self-administration. (d) NAC treatment reduced the stress-potentiated total number of alcohol reinforcers consumed. (e) NAC prevented the capacity of a stress-paired odor to reinstate alcohol seeking. (f) NAC treatment did not significantly alter the number of cocaine infusions taken during 10 days of self-administration. (g) NAC prevented the capacity of a stress-paired odor to reinstate cocaine seeking (left panel). The stress-unpaired odor had no effect on cocaine seeking in any treatment group (right panel). All data are shown as mean ± SEM. N shown in bars.

*p< 0.05 compared all other treatment groups, +p< 0.05 compared to extinction

NAC administration prior to and during stress or given following stress exposure decreased the active lever presses and the number of alcohol reinforcers compared to stress-VEH treated rats during the last 10 days of alcohol self-administration (Fig 4b, 2-way RM-ANOVA treatment F_(4,350)= 31.87, p< 0.0001, time F_(9,350)= 179.0, p< 0.001, interaction F_(36,350)= 1.83, p< 0.001, and Fig 4d; 1-way ANOVA F_(4,35)= 231.2, p< 0.001). Moreover, both NAC treatment regimens prevented the reinstatement of alcohol-seeking elicited by stress-conditioned odor (Fig 4e; 2-way RM-ANOVA Ext v Rst F_(1,35)= 57.42, p< 0.0001, treatment F_(4,35)= 28.37, p< 0.0001, interaction F_(4,35)= 29.28, p< 0.0001). No treatment group showed differences in 2-bottle choice, first 10 days of alcohol self-administration and inactive lever pressing between groups or compared to extinction levels of pressing (Fig S4a,b,c).

Neither NAC protocol altered the number of infusions taken during cocaine self-administration or cocaine intake per body weight (Fig 4f; 1-way ANOVA F_(3,17)= 1.38, p= 0.284, and Fig S4d; 1-way ANOVA F_(3,18)= 0.604, p= 0.62). Akin to alcohol-trained rats, both NAC treatment protocols prevented stress odor-induced reinstatement of cocaine seeking (Fig 4e; 2-way RM-ANOVA Ext v Rst F_(1,17)= 14.98, p= 0.001, treatment F_(3,17)= 3.12, p= 0.054, interaction F_(3,17)= 3.70, p= 0.032). The unpaired odor had no effect on lever pressing in cocaine-trained rats compared to extinction levels of pressing or between any treatment group (Fig 4g; 2-way RM-ANOVA Ext v Rst F_(1,17)= 1.40, p= 0.254, treatment F_(3,17)= 0.58, p= 0.637, interaction F_(3,17)= 0.00, p= 0.991). Inactive lever pressing in response to either the stress-paired or -unpaired odor was not different between any treatment group (Fig S4e).

NAC reversed conditioned stress-induced impairment in operant learning

To assess the impact of conditioned stress and NAC treatment on operant-based learning, we utilized an operant set-shifting task that requires the rat to alter its behavior in response to a rule change. NAC or vehicle was administered daily 4-days prior to, during, and for 5-days following restraint stress (NAC 100 mg/kg X 10 days) (Fig 5a). Rats were initially trained on a rule that required choosing the lever indicated by a stimulus light to receive a reinforcer. Rats from all groups required 3–5 daily sessions to acquire the first rule. On the final day of training for the initial rule, the number of trials required to reach criterion using this rule were equal among all treatment groups indicating that there was no effect of NAC treatment or stress exposure on the visual cue discrimination phase of the task (F_(3,4)= 0.038, p= 0.989)(Fig 5b). However, when tested for the ability to shift response strategy, only vehicle-treated, stress-exposed rats required more trials to shift from the visual-cue rule to the location rule (F_(3,20)= 5.77, p= 0.005)(Fig 5b). Further, the increase in errors observed in this group resulted from increased perseverative errors (F_(3,20)= 4.46, p= 0.015)(Fig 5c). These data indicate that NAC did not impair learning and prevented stress cue-induced impairments.

Figure 5. N -Acetylcysteine (NAC) treatment during and after stress prevented stress-induced cognitive deficit.

Rats from both groups required three to five daily sessions to acquire the first rule. A, Schematic outlines of NAC treatment before and after the 2 h of acute restraint stress. B, Final day of training for the initial rule, no difference in the number of trials required to reach criterion on the visual cue discrimination phase of the task between groups. However, when tested their ability to shift their response strategy, only stress-exposed rats that did not receive NAC treatment required significantly more trials to shift from the visual cue rule to the location rule. C, Increased perseverative errors in only stress-exposed rats vehicle treated. No difference in the number of omissions across treatment groups. All data are shown as mean ± SEM. N shown in bars. *P < 0.05, comparing stress-VEH with all other groups, no other significant comparison was found.

DISCUSSION

These experiments provide preclinical evidence that NAC treatment during or after stress may be a successful strategy for treating PTSD and SUDs comorbidity. Thus, NAC reduced stress-potentiated alcohol use and conditioned stress-induced reinstatement of alcohol and cocaine seeking. NAC also inhibited conditioned stress-induced alcohol and cocaine seeking when administered during extinction training and prior to the reinstatement session.

Animal modeling of comorbid PTSD and SUDs

Highly motivated drug-seeking is a cardinal symptom of SUDs, and many clinical and preclinical studies indicate that drug craving in humans and reinstated drug seeking in rodents is potentiated by stress exposure (22–24). Stress is well established to be an instigator of relapse to drug use (22), and the stress imposed on human addicts is often conditioned stress. In contrast, in rodent models most literature uses exposure to a primary stressor to reinstate drug seeking (23). Moreover, persons suffering PTSD avoid re-exposure to the primary stressor (e.g. combat or rape), making the stress that induces relapse to drug use in comorbid PTSD and SUDs more likely to be an environmental cue or context predicting the possible presence of the original stressful experience (25, 26). Thus, it has been argued that conditioned stress may be a useful model of PTSD and comorbidity with SUDs (25, 26).

In addition to the present study, a limited number of other studies have employed conditioned stress to reinstate drug seeking (23). Pairing an odor with repeated social defeat or acute footshock with continuous white noise reinstates alcohol seeking (27, 28). Pairing a context with relatively intense acute footshock leads to persistent enhancements in cue-induced reinstatement of methamphetamine seeking, as well as facilitating cocaine conditioned place preference (29). Furthermore, a history of exposure to a predator odor exposure enhances the reinstatement of methamphetamine seeking and exposure to stress odor also increases responding in a stress-induced reinstatement procedure (30). Stress-conditioned predator odor also potentiates cued reinstatement of cocaine seeking (31). Finally, rats exposed to inescapable foot-shock consume more ethanol compared to controls, suggesting that a single stressful event can have lasting effects on subsequent alcohol consumption (32). In contrast, unpublished data indicate that footshock paired with a tone does not reinstate heroin or cocaine seeking (33). Compared to these other studies, the rat model in the present study was unique in using acute restraint stress. Also, the previous studies introduced the stressor after discontinuing drug use, while in our model the stressor was administered 3 weeks prior to initiating drug self-administration. This was done to more closely model the clinical situation where acute stress exposure induces neurobiological changes that create vulnerability to develop SUDs. Clinically this models the more common sequence of events that occurs when individuals not addicted to drugs undergo a PTSD-initiating acute stressor, such as combat or rape, that then predisposes the individual to developing SUDs and engaging in uncontrollable relapse due to PTSD-associated conditioned stress (34, 35).

Alternate Interpretations and Caveats

Modeling complex disorders such as addiction and PTSD requires salient conditioning of conditioned with unconditioned stimuli that can ultimately shape subsequent instrumental behavior (i.e. Pavlovian-Instrumental Transfer, PIT)(36, 37). This phenomenon could contribute to our findings with cue-induced reinstatement findings. However, in a typical PIT design, the original conditioning stimulus and succeeding operant response are the same (e.g. food reward), which was not the case in our study. Nonetheless, more than one learning mechanism likely contributes to conditioned stress-induced drug seeking, and parsing these mechanisms in future studies is an important requisite towards complete understanding of the neurobiological basis of stress-induced relapse.

The fact that a behavioral response to the stress odor involved a motivated behavior rather than freezing response reflects a growing literature that rodents display both passive and active behaviors when presented with aversive stimuli (37). For example, in one fear conditioning footshock paradigm, rodents actively seek out a platform that allows them to escape when presented with a shock-conditioned stimulus (38). Although the rats in our study likely displayed passive avoidance during the reinstatement test, only the active behavioral strategy of lever pressing was quantified.

In addition to suppressing conditioned stress-induced active coping (lever pressing), it is possible that NAC inhibit the ability of the odor to become an operant conditioned stimulus. However, we found that NAC alone did not alter an operant-based cognitive task (the set shifting paradigm), while still suppressing stress cue-induced cognitive impairments. Also, one interpretation of NAC reductions in alcohol consumption is that NAC treatment changed the value (reinforcing properties) of alcohol. This is a logical explanation that deserves experimental exploration in future studies directly assessing the reinforcing value of alcohol, such as progressive ratio responding.

Neurobiological mechanisms to understand and treat comorbid PTSD and SUDs

A long-standing hypothesis is that monoamine transmission in the brain is altered in PTSD (26), and the only pharmacotherapy currently approved by the US Food and Drug Administration for treating PTSD are the selective serotonin reuptake inhibitors (SSRIs) (39). Unfortunately, SSRIs are only marginally effective at treating comorbid PTSD/SUDs (40). Because of the role of dopamine in drug reward (41), most preclinical studies have focused on stress-induced adaptations in mesocorticolimbic dopamine neurons that might alter the sensitization, self-administration, and reinstatement to addictive drugs. For example, stress-induced increases in glucocorticoid hormones promote vulnerability to drug use via actions on dopamine neurons (42), and blockade of corticotropin releasing factor (CRF) within the ventral tegmental area (VTA) inhibits stress-induced reinstatement of cocaine-seeking (43). While this study show that CRF release into the VTA is critical for stress-induced reinstatement, prefrontal cortical glutamate release into the VTA or nucleus accumbens may be more involved in conditioned stress-induced reinstatement (44). This might explain the failure of oral pexacerfont (a CRF1 antagonist) in a clinical setting to reduce stress-induced alcohol craving, emotional responsiveness or anxiety (45).

Recently it was shown in a pilot double-blind, placebo controlled clinical trial that daily NAC administration for 8 weeks in veterans comorbid for PTSD and SUDs reduced symptoms of PTSD and drug craving (9). In parallel with this clinical study, our experiments involved a rat model of acute restraint stress followed by alcohol and cocaine self-administration to show that NAC prevented conditioned stress-induced reinstatement of drug seeking. Acute stress and self-administration of addictive drugs lower the protein levels and function of GLT-1 in the nucleus accumbens of rodents (13, 15). Restoring GLT-1 with repeated injections of NAC or other compounds such as ceftriaxone or propentofylline prevents cues from reinstating drug seeking, and stress from facilitating the acquisition of cocaine self-administration (13, 15), the augmentation of alcohol intake in Sprague-Dawley (Fig 4b) or alcohol-preferring rats (46). We hypothesize that by increasing GLT-1, NAC restores the capacity of astroglia in the vicinity of accumbens excitatory synapses to eliminate synaptically released glutamate, and thereby inhibit cue-induced reinstatement.

Clinical implications for NAC

The capacity of NAC to simultaneously reduce symptoms in two distinct psychiatric disorders, PTSD and SUDs (9), points to action on a neuropathology and symptomatology that is shared between the disorders. It has been proposed that this shared characteristic is intrusive thinking (47). Intrusive thinking is a symptom that can initiate not only PTSD and drug craving, but also major depression (i.e. rumination) and obsessive compulsive disorder, all of which are disorders where NAC has a positive therapeutic action in double-blind clinical trials (9, 11). The fact that NAC is treating only one symptom in these psychiatric disorders may account for variability in its clinical efficacy. For example, reducing drug craving does not always prevent relapse to drug use, and reducing rumination may not prevent depression. However, in the case of comorbid disorders that share intrusive thinking as a symptom, NAC may be especially effective since weakening PTSD and SUDs comorbidity will increase the efficacy of other therapies designed to treat the individual disorders (8, 48, 49).

Conclusions

Here we provide novel behavioral findings in an animal model of comorbid PTSD and SUDs that parallel the findings shown in a recent clinical trial where NAC reduced symptoms of PTSD/SUDs comorbidity (9). Importantly, we found that NAC could be used prophylactically to prevent or reverse the vulnerability to undergo stress-induced escalation of alcohol use and conditioned stress-induced alcohol or cocaine seeking. Given that daily NAC treatments restore stress- or drug-induced down-regulation of GLT-1, we hypothesize that the efficacy of NAC indicates a novel underlying neurobiological mechanism of PTSD/SUDs comorbidity that parallels the well-studied role of GLT-1 in regulating cue-induced reinstatement of drug seeking. Testing this hypothesis is a topic of ongoing experimentation.