Sex-dependent effects of acute stress in adolescence or adulthood on appetitive motivation

Abstract

Rationale:

Intensely stressful experiences can lead to long-lasting changes in appetitive and aversive behaviors. In humans, post-traumatic stress disorder increases the risk of comorbid appetitive disorders including addiction and obesity. We have previously shown that an acute stressful experience in adult male rats suppresses motivation for natural reward.

Objectives:

We examine the impact of sex and age on the effects of intense stress on action-based (instrumental) and stimulus-based (Pavlovian) motivation for natural reward (food).

Methods:

Rats received 15 unsignaled footshocks (stress) followed by appetitive training and testing in a distinct context. In Experiment 1, stress occurred in either adolescence (PN28) or adulthood (PN70) with appetitive training and testing beginning on PN 70 for all rats. In Experiment 2, stress and appetitive training/testing occurred in adolescence.

Results:

Acute stress in adolescent females suppressed instrumental motivation assessed with progressive ratio testing when testing occurred in late adolescence or in adulthood, whereas in males stress in adolescence did not suppress instrumental motivation. Acute stress in adulthood did not alter instrumental motivation. In contrast, Pavlovian motivation assessed with single-outcome Pavlovian-to-instrumental-transfer (SO-PIT) was consistently enhanced in females following adolescent or adult stress. In males, however, stress in adolescence had no effect, whereas stress in adulthood attenuated SO-PIT.

Conclusions:

Acute stress in adolescence or adulthood altered instrumental motivation and stimulus-triggered Pavlovian motivation in a sex and developmentally specific manner. These findings suggest that the persistent effects of acute stress on Pavlovian and instrumental motivational processes differ in females and males, and that males may be less vulnerable to the deleterious effects of intense stress during adolescence.

Introduction

Stress is a normal part of life, but intensely stressful events can lead to long-term aberrations in learning, memory, and motivation (Giovanniello et al. 2023). In humans, post-traumatic stress disorder (PTSD) is comorbid with several appetitive motivational disorders including addictions, obesity, and eating disorders (Kessler et al. 2005; Pagoto et al., 2012; Swinbourne and Touyz 2007). Age and sex may be important determinants in the development of PTSD with comorbid disorders, with adolescents and females exhibiting increased risk (Gordon, 2002; Haskell et al. 2010; Shansky 2015; Perkonigg et al. 2009; Torchalla et al. 2014; Wang et al., 2010). A key aspect of PTSD and its comorbid conditions is its persistent nature. While there is a substantial literature on the immediate effects of acute and chronic stress on appetitive behaviors (see Mantsch, et al. 2016 for review) much of the work on the persistent effects of stress on reward processing has focused on the effects of chronic stress (see Lisieski et al. 2018 for review). Very little is known about the persistent effects of an acutely stressful experience on appetitive behaviors.

A simple acute procedure for inducing lasting stress effects is based on the stress-enhanced fear learning (SEFL) approach. In SEFL, a battery of shocks in one context leads to a sensitized fear response following a single shock in a novel context (Rau et al, 2005; Nishimura et al, 2022). We have adapted this approach to study persistent effects of an intensely stressful experience on reward-based behaviors (Pizzimenti et al., 2017; Derman and Lattal 2022; 2023). Subjects first undergo fear conditioning in one context followed by appetitive training/testing in a novel context. This approach has several important advantages. First, because this procedure is an isolated event, it allows for studying the effects of intense stress at specific developmental ages. Second, it is an automated and highly replicable procedure, minimizing potential experimenter bias. Third, it entails a well-characterized and easily measured behavioral response, freezing, which not only provides a readout of the stress treatment in real-time but also allows assessment of its carryover effects in new contexts.

Using this SEFL-based approach, we have demonstrated that in adult male rats, this single stressful event attenuates cue-triggered food seeking (Pavlovian-to-instrumental transfer, PIT) and instrumental motivation (progressive ratio, PR) long after stress-induction (Derman and Lattal, 2022; 2023). These data suggest that in adult males both stimulus-based (Pavlovian) and action-based (instrumental) motivation for natural reward are suppressed long after stress-induction. In the following experiments, we evaluate sex differences in these effects when acute stress occurs during adolescence or adulthood.

General Methods

Subjects:

Long Evan rats were used in all experiments (N=245); group sizes, ages, and sexes are outlined below in individual methods sections. All rats were offspring of rats purchased from Charles River and bred at OHSU. Housing density ranged from 2–4 rats per cage, depending on age and sex. Experimental procedures were conducted during the dark phase of the cycle. For all experiments, rats were food restricted during treatments (see each experiment for details).

Apparatus:

Three distinct contexts housed in separate rooms were used, one for stress induction (Ctx A), one for appetitive training and testing (Ctx B), and one for stress-enhanced fear learning (SEFL; Ctx C). Chambers differed in dimensions, patterned backdrops, olfactory cues, floor textures, and other internal features. All chambers were standard Med Associates operant chambers housed in sound attenuating cabinets. Ctx A was 29.53 × 24.84 × 18.67 cm (LxWxH) with metal rod grid floors (19, 4.7mm rods) and a single house light. The visual and olfactory distinguishing features of Ctx A were (1) a horizontal zig zagged pattern backdrop and (2) clover leaf essential oil (Crafter’s Choice^™ Clove Leaf EO- Certified 100% Pure 1050) infused gauze pads placed under the grid at the base of the chambers. Ctx B was 29.53 × 23.5 × 27.31 cm with metal rod grid floors, two retractable levers flanking a recessed food magazine, a single speaker above the magazine, two flat lights above the levers and a single hooded houselight on the opposing wall situated at the top and center. The tactile, visual, and olfactory distinguishing features of Ctx B were a wire mesh insert (19-Gauge Wire Mesh Fence; 0.5” mesh size) placed over the grid floors, a backdrop with randomly distributed differently sized squares or circles, and citrus cilantro fragrance oil infused gauze pads (Crafter’s Choice^™ Citrus Cilantro Fragrance Oil 548). Other internal features of Ctx B chambers included two retractable levers flanking a recessed food magazine, a single speaker above the magazine, and two flat lights above the levers on the front wall of the chamber. On the opposing wall, a single hooded houselight was situated at the top and center. Food hoppers were externally housed and connected to the magazine for delivery of pellets. Ctx C was 29.53 × 23.5 × 27.31 cm with metal rod grid floors to provide visual, dimensional, and olfactory distinctiveness, a 0.24 cm thick black neoprene sheet (29.5 × 61 cm) was inserted to vertically hug two of the chamber walls. These inserts altered the inner shape creating a curved appearance to two of the chamber walls and had a strong rubber smell providing a distinct olfactory feature.

Stress Induction:

Naive rats were placed into a novel operant chamber (Ctx A) for a 90-minute session, in which 15, 1 second, 1mA footshocks were delivered through the metal rods of chamber’s grid floor. Shocks were presented on a unsignaled variable time schedule of 360 seconds (VT 360”). For control treatment, rats were left in Ctx A for a 90-minute session without any shocks. All sessions were video recorded for subsequent quantification of freezing behavior. Freezing was quantified by sampling behavior every 10 seconds in the 3-minutes leading up to every shock; freezing was scored binarily with a ~2 second evaluation window. For control treatment, scoring was conducted using yoked timepoints.

Magazine Training:

All appetitive training and testing took place in Ctx B. Each magazine training session lasted 20 minutes during which 20 pellets (Bioserv Dustless Precision Pellets, 45mg; product# F0021) were delivered on a VT60” schedule into the chambers magazine. Rats received at least two sessions but were given additional training if more than 2 pellets were left unconsumed. The magazines contained an infrared beam across the entrance of the magazine that automatically detected entries which were recorded for analysis. The first magazine training session was video recorded to evaluate generalized freezing, which was assessed every 10 seconds for the session duration.

Continuous Reinforcement Training:

Next rats were trained to press a lever to earn these pellets. During these sessions an active and inactive lever were inserted into the chambers flanking the magazine (positions counterbalanced). Every press on the active lever earned a single pellet, whereas presses on the inactive lever had no programmed consequences. Sessions were terminated when either 50 presses were made, or 40 minutes had lapsed from session start. An acquisition criterion of 50 active lever presses in a single session was set for this training phase. Daily sessions were given till meeting this criterion, with a maximum of 6 total sessions allowed per rat. Lever and magazine responses were automatically recorded for analysis.

PR Testing:

During PR testing the active and inactive levers were inserted into the chambers and presses on the active lever were reinforced on a progressively increasing ratio schedule of reinforcement. Each session lasted 120 minutes and the number of responses required to earn a single pellet increased as follows: 1, 2, 4, 6, 9, 12, 15, 20, 25, 32, 40, 50, 62, 77, 95, 118, 145, 178, 219, 268, 328, 402, 492, 603, 737, 901, 1102 (as in Derman and Lattal, 2023; Richardson and Roberts, 1996).

Variable Interval Training:

Instrumental training then continued using variable interval (VI) schedules of reinforcement which were made progressively leaner across 8 sessions. Each session lasted 20 minutes and the daily VI schedules were as follows: 2xVI10”, 2xVI30” and 4xVI60”. Lever and magazine responses were automatically recorded for analysis.

Appetitive Pavlovian Conditioning:

During 8 sessions of conditioning, rats were presented with auditory conditioned stimuli (CSs) that were either paired with pellets (CS+) or not (CS−). These CSs were white noise and a tone (80dB) which were presented for 2 minutes per trial (CS+/− assignments counterbalanced). On CS+ trials, 4 pellets were delivered on a VT30” schedule. Each session lasted 55m and trials were separated by an inter trial interval (ITI) averaging 5 minutes (range 3–7 minutes). During Pavlovian conditioning, the levers were always retracted and unavailable to the rats. Magazine responses were automatically recorded for analysis.

PIT Testing:

During testing the active and inactive levers were inserted into the chambers and after the first 10 minutes of the session, CSs were presented every 4 minutes (i.e, with a fixed 2-minute ITI). Active lever presses and CS+ presentations did not deliver pellets. Each CS was presented 4 times in a semi-random order. The session ended 1 minute after the last CS offset, with a session duration of 40 minutes. Lever and magazine responses were automatically recorded for analysis.

SEFL Testing:

To evaluate SEFL, after appetitive training and testing in Ctx B, rats were brought into the novel Ctx C and given a single 1 second, 1 mA footshock at 3.2 minutes into a 3.8-minute session. The following day rats were brought back into Ctx C for a 12-minute test session during which no shocks occurred. All sessions were video recorded for subsequent quantification of freezing behavior.

Ctx A Fear Retention Testing:

After SEFL testing in Experiment 2, rats were brought back into Ctx A for a 12-minute test session during which no shocks occurred. All sessions were video recorded for subsequent quantification of freezing behavior.

Experiment 1: Effects of Adolescent or Adult Stress on PR and SO-PIT during Adulthood

Experiment 1a Effects of Adolescent vs Adult Stress on PR:

Figure 1A depicts the timeline of Experiment 1a. Rats were weaned at PN21 and then on PN27 food was removed from their homecages. The next day (PN28) rats designated for adolescent treatment received stress induction treatment (Group Adolescent Shock: n=16; f=10; m=6) or control treatment (Group Adolescent Ctrl: n=15; f=9; m=6) in Ctx A. The rats designated for adult treatment were left undisturbed in their homecages. All rats were then returned to ad libitum food access until PN69, when food was removed, and weights were maintained at 85–90% of ad libitum weights. On PN70 all rats designated for the adult pre-training treatment group received stress induction treatment (Group Adult Shock: n=13; f=8; m=5) or control treatment (Group Adult Ctrl: n=13; f=9; m=4) in Ctx A; all other groups remained in their homecages. Rats received magazine training in Ctx B over the next two days, followed by CRF training to criterion, followed by two PR test sessions.

Fig. 1.

Experiments 1a/b initial training. A) Timeline for Experiments 1a and 1b. On PN28 subjects in the adolescent treatment group received shock or control treatment in Ctx and all other animals remained in their homecages. On PN70 rats in the adult treatment group received shock or control treatment in Ctx and all other animals were left in their homecages. On PN71 magazine training began in Ctx B, followed by CRF training, and then PR testing for Experiment 1a and PIT training and testing for Experiment 1b. (*: for experiment 1b we included a group that received Ctx a shock or control treatment after appetitive training, but before PIT testing.) B) Following adolescent shock, acute recovery of weight-loss from food restriction was blunted by shock in females (n=6) and males (n=6) compared to non-stressed rats (females, n=18 [homecage controls, n=12; n=6]; males=15 [homecage controls, n=9; n=6]). In addition, weight recovery was overall lower in females compared to males. C) Responding during magazine training. Magazine responding was suppressed in adolescent shocked females (n=18) compared to female controls (n=46), but in adult shocked females (n=24) and females in the post-training group (n=12) there was no difference in responding compared to controls. In males, both adolescent (n=18) and adult shocked (n=21) rats exhibited suppressed magazine responding compared to controls (n=35). Unexpectedly the post-training group (n=20) also showed reduced magazine responding compared to controls. D) Summary of freezing across magazine training. In females, adolescent shocked females (n=11) showed greater freezing than adolescent controls (n=37), whereas adult shocked females (n=15) did not show significant freezing above controls. In males, both adult (n=18) and adolescent (n=9) shocked rats showed significantly greater freezing than male controls (n=47). In addition, adolescent shocked males showed greater freezing than adult shocked females. E) Timecourse of freezing across the first magazine training session. In females, adolescent shocked rats showed robust freezing that peaked in the first 5 minutes of the session and persisted variably at low levels for the remainder of the session. Whereas the adult shocked females exhibited minimal freezing that peaked in minute 4 and then disappeared for the remainder of the session. In males, adolescent shocked males, freezing peaked in the first 5 minutes and remained evident throughout the entire session. Whereas in adult shocked males, freezing was low, but evident in the first 10 minutes and then tapered off for the remainder of the session. F) Time to acquire CRF training. In females, adolescent shocked rats (n=18) the time to acquisition was significantly longer than controls (n=58), but the time to acquire did not differ between adult shocked (n=24) and control rats. In males, time to acquire was significantly longer in both adolescent (n=18) and adult (n=21) shocked rats compared to controls (n=55). (For all panels, data shown as means with SEM; #: different from Ctrl, p<0.05; @ sex difference, p<0.05)

A subset of rats received fear extinction and SEFL testing after PR testing. Some rats were re-exposed to Ctx A for a 180 minute extinction session (Adolescent Ctrl: n=6; f=3; m=3; Adolescent Shock n=5; f=3; m=2; Adult Ctrl: n=5; f=3; m=2; Adult Shock n=6; f=3; m=3) or left undisturbed in their homecages (Adolescent Ctrl: n=6; f=3; m=3; Adolescent Shock n=7; f=3; m=4; Adult Ctrl: n=5; f=3; m=2; Adult Shock n=6; f=3; m=3). The next day, rats were brought to context C for SEFL.

Experiment 1b Effects of Adolescent vs Adult Stress on SO-PIT:

As detailed above adolescent and adult rats were given Ctx A treatment on PN28 and PN70 respectively followed by appetitive training in Ctx B. An additional cohort of rats was given Ctx A treatment after appetitive training (Post-training groups). All rats remained in their homecages when not receiving Ctx A treatment. The groups sizes were as follows: Adolescent Pre Ctrl: n=16; f=8; m=8; Adolescent Pre Shock n=16; f=8; m=8; Adult Pre Ctrl: n=33; f=19; m=14; Adult Pre Shock n=29; f=15; m=14; Adult Post Ctrl: n=15; f=5; m=10; Adult Post Shock n=17; f=7; m=10). Appetitive training in Ctx B began on PN71 for all groups and consisted of 2 sessions of magazine training, CRF training to criterion, 8 sessions of VI training, and 8 sessions of Pavlovian conditioning. Ctx A treatment (shock or control) was given to the Post groups the day before PIT testing and all other rats were left undisturbed in their homecages. This was followed by two PIT test sessions.

Experiment 2: Effects of Adolescent Stress on PR and SO-PIT during Adolescence

Experiment 2a Effects of Adolescent Stress on PR in Adolescence:

Rats were weaned on PN21 and food was removed on PN27. Rats were food restricted for the duration of the experiment but allowed to grow ~7g daily; this was achieved by titrating chow rations with the following formula: 0.1 × bodyweight (g) = chow (g). This target daily weight gain was chosen to ensure typical growth (in accordance with published Charles River Long Evans growth curve). On PN28 rats received stress induction (Shock: n=7; f=3; m=4) or control treatment (Ctrl n=9; f=6; m=3) in Ctx A followed the next day by appetitive training in Ctx B, which consisted of 2 sessions of magazine training and CRF to criterion, followed by two sessions of PR testing. Rats were given a 2 week break and then given a third PR test. In the days following, rats were tested for SEFL in Ctx C and then finally all rats were returned to Ctx A to evaluate retention of contextual fear learning.

Experiment 2b Effects of Adolescent Stress on PIT in Adolescence:

Rats were weaned at PN21 and food restricted as above starting on PN27, followed by Ctx A treatment as described for Experiment 2a on PN28. Appetitive training began in Ctx B on PN29 (Group sizes: Ctrl n=16; f=8; m=8; Shock: n=16; f=8; m=8;), consisting of 2 magazine training sessions, CRF training to criterion, 8 VI training sessions, and 8 Pavlovian conditioning sessions. Following training, PIT was tested in 2 sessions, which occurred between PN51 and PN57 in late adolescence. In the days following PIT testing, rats received SEFL conditioning and testing in Ctx C and the day after testing for SEFL rats were tested for retention of fear in Ctx A (details in Supplemental Text).

Statistical Analysis:

Data were analyzed using two-way ANOVAs and repeated measures ANOVAs (RM-ANOVAs). Planned comparisons and post-hoc comparisons were performed using Holm-Sidak’s multiple comparison tests. One of our goals was to examine sex as a biological variable thus we established a priori that comparisons would be made between treatment conditions within each sex separately when groups were sufficiently powered. These comparisons were made even in the absence of main effects of sex or interactions with sex. When adolescent and adult control subgroups did not differ, these groups were collapsed into one large control group for analysis of the effects of shock treatment. For Experiment 1b, the post-training group treatments (control or shock) were given after training; thus, the training data from these groups were collapsed with controls when behavior was statistically indistinguishable from pre-training control groups. Data were analyzed in Graphpad Prism and IBM SPSS.

Results

Experiment 1: Effects of Adolescent or Adult Stress on PR and SO-PIT during adulthood

Experiments 1a and 1b:

Training in the first stages (pre-training Ctx-A treatment, magazine and CRF training) of Experiments 1a and 1b was identical and therefore the data from these phases were collapsed across the experiments.

Stress induction in adolescence or adulthood:

Shocked rats acquired high levels of freezing during acquisition in adolescence and adulthood (see Supplemental Text and Supplemental Fig. 1A).

Adolescent stress suppresses weight gain transiently:

Fig. 1B shows the acute effects of shock on weight recovery on the day following relief from food restriction (PN29). Acute weight recovery following food restriction was blunted by shock in females and males, and males overall gained more weight than females (main effects of Ctx-A-treatment, F_(1,41)=7.99, p<0.01, and sex, F_(1,41)=8.38, p<0.01; no sex × Ctx-A-treatment interaction, p=0.97). This effect of shock did not persist; over the next five weeks, adolescent control and shock groups did not differ in weight gain (Supplemental Fig. 1B).

Stress prior to appetitive training alters magazine responding and generalized freezing in adulthood:

Fig. 1C shows the mean magazine responses during magazine training (averaged across 2 training sessions). Magazine responding was significantly lower in the post-training groups (i.e., rats destined to receive Ctx A treatment after appetitive training) compared to controls, so these data were kept separate. Stress suppressed magazine responding, with a stronger effect in males compared to females (Fig. 1C: reliable main effects of Ctx-A-treatment, F_(3,186)=12.31, p<0.01, sex, F_(1,186)=8.6, p<0.01, and a non-significant Ctx-A-treatment × sex interaction, F_(3,186)=2.34, p=0.08). In females, shocked adolescents responded less than controls (t₍₁₈₆₎=2.70, p=0.02), but responding was similar between adult shocked rats and controls (p=0.45). In males, responses were lower in rats shocked in adolescence or adulthood compared to controls (Ctrl vs Adolescent Shock t₍₁₈₆₎=4.99, p<0.01; Ctrl vs Adult Shock t₍₁₈₆₎=3.98, p<0.01). In both females and males, magazine responding was significantly lower in the post-training group (post-hoc comparisons: Females: Ctrl vs Post, t₍₁₈₆₎=2.33, p=0.04; Males: Ctrl vs Post, t₍₁₈₆₎=3.43, p<0.01). Between-sex post-hoc comparisons revealed that adult shocked males showed suppressed responding compared to adult shocked females (t₍₁₈₆₎=2.99, p<0.01).

As can be seen in Fig. 1D, stress in adolescence led to increased generalized freezing during magazine training, especially in males (main effects of Ctx-A-treatment, F_(2,131)=81.38, p<0.01; sex, F_(2,131)=12.30, p<0.01; and a significant Ctx-A-treatment × sex interaction, F_(2,131)=5.49, p<0.01). Freezing was significantly greater in adolescent shocked females than in controls (t₍₁₃₁₎=6.88, p<0.01) but freezing in adult shocked females did not differ from controls (p=0.42). In contrast, freezing was significantly greater than controls in both adolescent shocked males, t₍₁₃₁₎=11.02, p<0.01, and adult shocked males, t₍₁₃₁₎=2.55, p=0.01. Between-sex comparisons in each Ctx-A-treatment group revealed that adolescent shocked males showed stronger freezing than females (t₍₁₃₁₎=3.69, p<0.01), but freezing did not differ by sex in the adult shocked or control groups (ps>0.3). As can be seen in Fig. 1E, closer temporal analysis of freezing revealed that freezing peaked within the initial 5 minutes across all shocked groups and was more prolonged in males compared to females. Furthermore, adolescent shocked rats exhibited more pronounced freezing compared to adult shocked rats.

Acquisition of continuous reinforcement training in adulthood is affected by prior stress in a sex-specific pattern.

Fig. 1F shows the acquisition time for CRF training (main effect of Ctx-A-treatment, F_(2,187)=12.29, p<0.01; no effect of sex, p=0.34; and no interaction, p=0.17). In females, rats shocked in adolescence showed delayed acquisition compared to controls (t₍₁₈₇₎=2.59, p=0.02), whereas adult shocked rats acquired at the same speed as controls (p=0.94). In males, acquisition was delayed in rats shocked in adolescence or adulthood compared to controls (ts₍₁₈₇₎>2.40, ps<0.02). From this point on, rats received either PR testing (Experiment 1a) or SO-PIT (Experiment 1b).

Experiment 1a: PR Testing:

Fig. 2A shows lever responding during PR testing in 20-minute bins. Control groups were not collapsed for analyses as responding differed as function of sex, treatment age, and time. A RM-ANOVA (sex × Ctx-A-treatment × Ctx-A-age × time bin) revealed a significant effect of time, F_(5,240)=30.64, p<0.01 and a significant 4-way interaction, F_(5,240)=4.08, p<0.01. Given this interaction, we followed up with Ctx-A-treatment by bin RM-ANOVAs for each age group and sex separately. Adolescent shocked females showed suppressed responding relative to control females (main effect of Ctx-A-treatment, F_(1,17)=4.67, p<0.05, and Ctx-A-treatment × time interaction, F_(5,85)=2.16, p=0.06). Responding did not differ between adult shocked and control females (Ctx-A-treatment, p=0.20; Ctx-A-treatment × time interaction F_(5,75)=2.15, p=0.07). In males, shock had no effect on responding whether it occurred in adolescence or adulthood (ps>0.30). Figure 2B shows the mean breakpoint from PR testing. A Two-way ANOVA found a main effect of treatment condition, F_(3,58)=2.94, p=0.04, a non-significant effect of sex, F_(1,58)=3.54, p=0.07, and no interaction, p=0.14. Post-hoc analyses found that adolescent control females reached significantly higher breakpoints than adolescent shocked females t₍₅₈₎=3.57, p<0.01. Breakpoints were similar between control and shocked groups for adult-treated females, adolescent-treated males, and adult-treated males (ps>0.86).

Fig. 2.

Experiment 1a PR and SEFL. A) Timecourse of active lever responding across PR testing. In females, adolescent and adult control groups differed, so analysis was conducted separately for adult and adolesce Ctx-A treatment groups. In adolescent treated females, active lever responding was significantly lower in shocked (n=9) rats compared to controls (n=10). In adult treated females, responding was similar between control (n=9) and shocked rats (n=8). In males, responding did not differ between adolescent control (n=10) and shocked rats (n=9) or between adult control (n=4) and shocked (n=7). B) Breakpoints from PR testing. In females, breakpoints were highest in rats given control treatment in adolescence compared to all other groups. In males, breakpoints were similar across all groups. C) Ctx-C SEFL testing following either Ctx-A extinction or no extinction. In non-extinguished rats, adolescent (n=7; f=3; m=4) and adult (n=6; f=3; m=3) shocked rats exhibited significantly greater freezing compared to controls (n=11; f=5; m=6). In contrast, in the extinguished group, freezing was significantly greater in adult shocked (n=6; f=3; m=3) rats compare to controls (n=10; f=5; m=5) but freezing did not differ between controls and adolescent shocked rats (n=5; f=3; m=2). (Data shown as means with SEM; *: different from controls, p<0.05; f: females; m: males).

Experiment 1a: SEFL Testing: Re-exposure to Ctx A weakens Stress Enhanced Fear Learning:

In one cohort, SEFL was conducted in a novel context, Ctx C. This dataset was not sufficiently powered for evaluation of sex differences, so sex was not examined as an independent variable here. On the first day of SEFL prior to the single shock in Ctx C, no freezing was observed in any animals when exposed to this new context (data not shown). On the following day, freezing was assessed in Ctx C during a SEFL test. Fig. 2C shows the average percent of freezing across SEFL testing. These data show that re-exposure to Ctx A eliminated SEFL in adolescent shocked rats, but not in adult shocked rats. A Ctx-A-pre-treatment (shock or no shock) × Ctx-A-post-treatment (extinction or no extinction) ANOVA revealed a significant main effect of pre-treatment, F_(2,39)=12.88, p<0.01, as well as a significant main effect of post-treatment F_(1,39)=5.93, p=0.02. Planned comparisons indicated that in rats not re-exposed to Ctx A, freezing was greater in both adolescent shocked rats, t₍₃₉₎=2.57, p=0.01, and adult shocked rats, t₍₃₉₎=4.87, p<0.01, compared to controls. However, in rats re-exposed to Ctx A, the adolescent shocked rats did not show greater freezing than controls (p=0.81), whereas adult shocked rats did, t₍₃₉₎=2.33, p<0.05.

Experiment 1b: Variable Interval Training and Pavlovian Conditioning:

Training data are described in full detail in Supplemental Text. VI training and Pavlovian conditioning are shown in Supplemental Fig. 2. All groups developed reliable instrumental and Pavlovian responding without apparent systematic effects of pre-training shock or sex on performance.

Experiment 1b: PIT Testing:

Fig. 3 shows Pavlovian and instrumental responding during PIT testing. Males and females showed discriminated Pavlovian magazine responding to CS+ and CS− (main effect of phase (ITI, CS+, CS−), F_(2,232)=120.86, p<0.01; phase × Ctx-A-treatment × sex interaction, F_(6,232)=2.62, p=0.02). Females in the post-training shock group exhibited greater CS+ magazine responding compared to female controls (Fig. 3A; main effect Ctx-A-treatment, F_(3,58)=5.82, p<0.01; phase × Ctx-A-treatment interaction, F_(6,114)=2.33, p=0.04; post-training shock vs. controls, t₍₁₇₁₎=5.13, p<0.01). In contrast, males in the adolescent shock and adult pre-training shock groups displayed lower CS+ responding than male controls (Fig. 3B; controls vs. adolescent pre-training shock, t₍₁₇₄₎=3.10, p<0.01; controls vs. adult pre-training shock, t₍₁₇₄₎=3.29 p<0.01), but ITI and CS− responding did not differ in any of the groups. Separate sex × phase RM-ANOVAs for each Ctx-A-treatment group revealed sex differences in the adult pre-training shock group, F_(1,24)=5.94, p=0.02, and the adult post-training shock group (F_(1,15)=11.86, p<0.01) and post-hoc comparisons determined that CS+ magazine responding was greater in females than males in the adult pre-training shock (t₍₇₂₎=3.07 p<0.01) and adult post-training shock (t₍₄₅₎=3.53 p<0.01) groups, whereas responding in the ITI and CS− did not differ between sexes in these groups (ps>0.43). For the control and adolescent pre-training shock groups no sex differences or phase × sex interactions were found (ps>0.13).

Fig. 3.

Experiment 1b PIT Testing data. A) Conditioned responding in females (Ctrl, n=32; Adolescent Pre-training Shock, n=8; Adult Pre-training Shock, n=12; Adult Post-training Shock, n=10). Conditioned discrimination was evident in all groups; magazine responding was significantly greater during the CS+ than CS− and ITI responding. Between group analyses revealed that in females given post-training shock, CS+ responding was greater than controls. B) Conditioned responding in Males (Ctrl, n=32; Adolescent Pre-training Shock, n=8; Adult Pre-training Shock, n=14; Adult Post-training Shock, n=7). Discrimination in conditioned responding was evident in all groups; CS+ magazine responding was significantly greater than CS− and ITI responding. Between groups comparisons found that CS+ responding was significantly lower in adolescent and adult rats shocked prior to appetitive training. C) PIT effect in females. All female groups showed PIT; in each groups active lever responding was significantly greater during CS+ presentations than during CS− presentations. Active lever responding was also greater during the CS+ than during the ITI. Between group comparisons showed that in all shocked groups, active lever responding was significantly greater than in the control group. The shaded region in the backdrop depicts the mean PIT effect size for controls. D) PIT effect in males. PIT was evident in control and adolescent shock males with CS+ presentations eliciting significantly greater responding than CS− presentations. In contrast, in the adult pre-training shock and post-training shock groups, PIT was absent with no difference in active lever responding between CS+ and CS− presentations. In controls, adolescent shock, and adult pretraining shock groups, CS+ presentations significantly increased responding over the ITI response rate. Between group comparisons showed that responding on the CS− was greater in the post-training group than controls. The shaded region in the backdrop depicts the mean PIT effect size for controls. E) Within trial timecourse of PIT in females. In the female control, adolescent shock, and pre-training adult shocked groups, PIT was evident during the entire CS window, whereas in the post-training shock group PIT was reliable in the second half of the CS. E) Withing trial timecourse of PIT in males. In control males PIT was reliable for the entire CS durations, whereas in adolescent the adolescent shock group the effect was reliable from 30–90s within the CS. In contrast, in the adult pre-training shock and post-training shock groups, PIT was absent during the entire CS window. (Data shown as means with SEM; #: different from ITI, p<0.05; $: different from CS−, p<0.05; @: different from controls, p<0.05)

We next evaluated PIT by analyzing active lever responding, where PIT was defined a priori as higher responding during the CS+ compared to the CS−. Figs 3C and 3D show mean active lever responding in females and males, respectively (main effect of phase, F_(2,232)=54.67, p<0.01 and a significant phase × Ctx-A-treatment × sex interaction, F_(6,232)=2.90, p=0.01). In females, within group planned comparisons revealed that PIT was observed across all groups (F_(2,114)=32.19; CS+ vs CS: ts₍₁₁₄₎>2.73, ps<0.02). Moreover, shock treated females showed greater CS+ active lever responding than female controls (F_(3,57)=5.00, p<0.01; ts₍₁₇₁₎>2.43, ps<0.02). In males, however, reliable PIT was only seen in the control (t₍₁₁₆₎=6.56, p<0.01) and adolescent pre-training shock groups (t₍₁₁₆₎=2.38, p=0.04), while it was absent in the adult pre-training shock and adult post-training shock groups (ps>0.19). To determine the source of the sex differences identified by the 3-way interaction between Ctx-A-treatment, sex, and phase identified by our original RM-ANOVA we also followed up with sex × phase RM-ANOVAs within each CtxA-treatment group. In controls, we found no main effect of sex (p=0.13) but a significant phase × sex interaction (F_(2,126)=3.15, p<0.05) however, post-hocs did not identify any reliable sex difference in responding during the ITI, CS+, or CS− (ps>0.09). In the adolescent pre-training shocked groups, we found a main effect of sex (F_(1,14)=9.27, p<0.01) and post-hoc comparisons identified that this sex difference was reliable in CS+ responding (t₍₄₂₎=3.00, p<0.01) but not during the ITI or CS− (ps>0.39). In the adult pre-training and adult-post-training shocked groups, there was no main effect of sex nor a sex × phase interaction (ps>0.27).

Further analysis of the temporal pattern of responding during the PIT test confirmed the deficit in adolescent or adult shocked males. Fig. 3 shows active lever responding in 30s bins from 60s prior to CS onset to 60s post CS offset. In females (Fig. 3E), PIT was evident across the entire CS period in all Ctx-A-treatment groups, with some continuation of this effect 30s following CS offset in controls (CS × time interactions: control: F_{(7, 224)}=4.56, p<0.01; adolescent shock: F_{(7, 49)}=3.78, p<0.01; adult pre-shock: F_{(7, 91)}=5.00, p<0.01; adult post-training shock: F_{(7, 42)}=2.34, p=0.04). In males (Fig. 3F) PIT was apparent within the entire CS window in controls, whereas in males shocked in adolescence PIT was more variable during the CS window and was only reliable later in the CS window (30–90s). In the adult pre-training shocked and adult post-training shocked groups, PIT was entirely absent during the CS window (CS × time interactions: control: F_{(7, 217)}=11.27, p<0.01; adolescent shock: F_{(7, 49)}=2.35, p=0.04; adult pre-shock: p=0.58; adult post-training shock: p=0.74).

Experiment 2: Effects of Adolescent Stress on PR and SO-PIT during Adolescence

Experiment 2a and 2b:

Training in the first stages of Experiments 2a and 2b were identical and therefore the data from these phases were collapsed across the experiments.

Adolescent weight gain under food restriction:

All rats were food restricted as adolescents at the beginning of the experiment but were allowed to gain weight throughout the experiment while under food restriction. Fig. 4B shows weight gain across the experiment. Males were heavier than females at the start and grew heavier with time (main effect of sex, F_(1,22)=20.42, p<0.01, and time, F_(22,528)=1486.99, p<0.01; sex × time interaction, F_(22,528)=41.24, p<0.01; no sex × time × Ctx-A treatment interaction, p=0.58). Shock did not affect weight gain (no effect of Ctx-A-treatment, p=0.39).

Fig. 4.

Timeline and initial training for Experiments 2a and 2b. A) Experimental timeline. At PN28 rats were shocked or given control treatment in Ctx A and appetitive began the following day in Ctx B. Following appetitive training and testing all rats were tested for SEFL in Ctx C and then finally tested for fear retention in Ctx A. B) Weight gain under food restriction across the duration of the experiment. Males (Ctrl, n=13; Shock, n=12) gained significantly more weight than females (Ctrl, n=12; Shock, n=14) and there were no differences in weight gain between shock and control rats. C) Responding during magazine training. Both shocked females (n=14) and males (n=12) showed suppressed magazine training compared to controls (females, n=12; males, n=13). D) Timecourse of freezing during magazine training. Both females and males show moderate freezing levels in the first fifteen minutes of training. There were no apparent sex differences. E) Time to acquire CRF training. In females, the time to acquire was significantly longer in shocked rats (n=14), compared to controls (n=12). In males, there was no difference in the time to acquire between control (n=13) and shocked (n=12) rats. (Data shown as means with SEM; *: different from controls, p<0.05)

Stress induction in adolescence:

Shocked males and females rapidly acquired freezing in Ctx A. These data are described in Supplemental Text and shown in Supplemental Fig. 3A.

Adolescent shock suppresses magazine training in adolescence:

Fig. 4C shows that magazine responding was suppressed by adolescent shock (main effect of Ctx-A-treatment, F_(1,47)=20.50, p<0.01; no effect of sex, p=0.24; no interaction, p=0.47). Post-hoc comparisons confirmed that compared to controls responding was lower in shocked females and males (ts₍₄₇₎>2.65, ps<0.01). Fig. 4D shows that freezing in the first magazine training session was higher in shocked females and males compared to controls with all groups extinguishing by the end of the session (main effects of treatment, F_(1,47)=14.45, p<0.01, and time, F_(19,893)=2.60, p<0.01; no effect of sex, p=0.75; significant Ctx-A-treatment × time interaction, F_(19,893)=3.14, p<0.01).

Adolescent stress delays acquisition of continuous reinforcement training in adolescent females, but not males:

Fig. 4E shows that shock delayed acquisition of instrumental responding in females (main effect of Ctx-A-treatment (F_(1,47)=4.20, p=0.05), no effect of sex (p=0.46), and no interaction (p=0.23)). Planned comparisons in each sex revealed that acquisition times were longer in shocked females compared to control females (t₍₄₇₎=2.33, p<0.05) but there was no difference in acquisition times between control and shocked males (p=0.56).

Experiment 2a: Adolescent stress and PR testing in adolescence.

PR testing is shown in 20-min bins across 3 tests in Fig. 5. A test × bin × Ctx-A-treatment × sex RM-ANOVA found that responding systematically differed across time (main effect of test, F_(2,24)=5.35, p=0.01; main effect of time bin, F_(5,60)=29.67, p<0.01) and sex (F_(1,12)=6.75, p=0.02). Analysis of the time course of responding in females found a significant time × Ctx-A treatment interaction (F_(17,130)=1.94, p=0.02). Post-hoc analyses were unable to identify the source of this interaction, however ordinally responding appeared to increase between tests in controls, but less so in shocked rats. In males responding did not differ between shock or control rats across tests, statistically or ordinally (p=0.68).

Fig. 5.

Experiment 2a PR testing in adolescence. Timecourse of active lever testing during 3 PR tests. A) In females, responding did not differ between shocked (n=6) and control (n=4) rats across testing. B) In males responding did not differ between shocked (n=3) or control (n=3) rats. (Data shown as means with SEM; @: different from controls)

Adolescent stress does not alter PIT training:

Training data are described in full detail in Supplemental Text and VI training and Pavlovian conditioning are shown in Supplemental Figs 4B and 4C, respectively. Control and shocked groups developed reliable instrumental responding and reliable Pavlovian conditioning with no effects of shock on performance. During instrumental training females showed lower response rates independent of shock.

Adolescent stress potentiates SO-PIT expression in late adolescence in females, but not males:

Fig. 6 shows Pavlovian (Fig. 6A) and instrumental (Fig. 6B) responding from the PIT test. All groups showed reliable Pavlovian CS+ vs. CS− discrimination (F_(2,56)=53.35, p<0.01) with no reliable main effects or interactions involving sex or Ctx-A-treatment (ps>0.2). PIT was observed in all rats, but in females CS+ active lever responding was stronger in shocked rats than in controls (Fig. 6B; main effect of phase (ITI/CS+/CS−), F_(2,56)=35.65, p<0.01; no effect of sex or Ctx-A treatments, ps>0.32; marginal sex × Ctx-A-treatment interaction, F_(1,28)=3.60, p=0.07). In females, we found main effects of phase (F_(2,28)=18.19, p<0.01) and Ctx-A-treatment (F_(1,14)=7.07, p=0.02) with reliable PIT in control (t₍₂₈₎=2.79, p=0.03) and shocked females (t₍₂₈₎=4.95, p<0.01), but the magnitude of PIT was greater in shocked females compared to control females (t₍₄₂₎=3.12, p<0.01). In males, we found a main effect phase (F_(2,28)=18.38, p<0.01) and no effect of treatment (p=0.61). Planned within group comparisons between the CS+ and CS− confirmed that PIT was reliably expressed in control and shocked males (ts₍₂₈₎>3.50, ps<0.01). Finally, to assess the sex differences we compared control females to males and separately shocked females to males using a phase × sex RM-ANOVA. In controls, we found a marginal effect of sex, F_(1,14)=3.99, p=0.07 and post-hoc comparisons revealed that CS+ active lever responding was significantly greater in male controls compared to females, t₍₄₂₎=2.65, p=0.03, but did not differ during the ITI or CS−. No sex differences were found between shocked females and males, ps>0.51.

Fig. 6.

Experiment 2b PIT testing in adolescence. A) Conditioned responding during testing. All groups showed reliable conditioned discrimination; CS+ magazine responding was significantly greater than CS− and ITI magazine responding (n=8 per group). B) Active lever responding. PIT was evident in all groups; active lever responding was significantly greater during CS+ presentation than CS− presentations in all groups. Responding was also significantly greater during the CS+ than the ITI. Between group comparisons within sex revealed that in females, shocked rats showed significantly greater CS+ active lever responding compared to controls. Responding in males did not differ between controls and shocked rats. C) Timecourse of within trial active lever responding. In females, PIT was evident in the entire CS window, and in the shocked rats this effect carried over into the first 30 seconds post CS in shocked females, but not controls. In males, the PIT effect was reliable in the first minute of the CS window, becoming unreliable in the second half of the CS window, whereas in shocked males PIT was evident in the entire CS window and similar to shocked females, this carried over into the 30s post CS bin. (Data shown as means with SEM; #: different from ITI, p<0.05; $: different from CS−, p<0.05; @: different from controls, p<0.05)

Figs. 6C and 6D shows the time course of PIT expression during the CS. In both males and females, the PIT effect was temporally organized, with augmentation of the active lever starting after CS onset and terminating after CS offset. The pattern of responding in this timecourse analysis is consistent with the session summary data.

Experiment 2a and 2b SEFL and fear retention testing following appetitive testing in adolescents:

At the end of Experiments 2a and 2b rats were tested for SEFL in a novel context (Ctx C) and then tested for retention of fear to the original shock context (Ctx A). In the case of SO-PIT this test occurred following PIT testing, and in the case of PR, testing occurred after the final PR test. In both cases the tests were conducted around PN 50–55. The data from these tests were collapsed across the two experiments as no differences were found between performance within the shock or control groups in these separate experiments.

Experiment 2a and 2b SEFL and Ctx A Fear Retention Test:

SEFL was conducted in novel Ctx C. On day 1, freezing prior to the first shock did not differ between control or shocked rats or between females and males (data not shown: Sex × Ctx-A treatment RM-ANOVA: no effect of Ctx-A treatment, p=16; no effect of sex 0.30; no interaction, p=29), Fig. 7A shows a summary of freezing during SEFL testing. Freezing was significantly greater in shocked rats compared to controls and this effect was stronger in males. During the first session in Ctx C freezing did not differ between groups or sexes prior to shock, ps>0.16. A sex × Ctx-A treatment RM-ANOVA revealed main effects of Ctx-A treatment, F_(1,40)=19.54, p<0.01 and sex, F_(1,40)=5.51, p=0.02. Comparisons between shocked and control rats confirmed that freezing was significantly higher in shocked rats compared to controls in both females, t₍₄₀₎=1.98, p=0.05 and males, t₍₄₀₎=4.17, p<0.01. Comparisons between sexes revealed that freezing was lower in shocked females compared to shocked males t₍₄₀₎=3.04, p<0.01, but that freezing did not differ between sex in controls, p=0.69.

Fig. 7.

Experiment 2a/b SEFL and Ctx A fear retention test. A) Summary of freezing during SEFL testing in Ctx B. During SEFL testing both female (n=13) and male (n=11) shocked rats showed significantly greater freezing than controls (females, n=11; males, n=9). Shocked males exhibited higher freezing than shocked females. B) Summary of freezing during a test for fear retention in Ctx A. Shocked rats showed more freezing than controls. This effect was weak in females and shocked males showed significantly greater freezing than shocked females. (Data shown as means with SEM; @: different from controls, p<0.05; &: sex difference, p<0.05)

In a final test, rats were returned to Ctx A to evaluate retention of fear. A summary of freezing during this retention test is shown in Fig. 7B, where shocked rats showed evidence for retained fear evident by greater freezing in shocked vs control rats. This was supported by a sex × Ctx-A treatment ANOVA which found a main effect of treatment, F_(1,46)=18.22, p=0.12 and no effect of sex, p=0.12.

Discussion

An acute stressor in adolescence or adulthood altered instrumental motivation (progressive ratio; PR) and stimulus-triggered Pavlovian motivation (single-outcome Pavlovian-to-instrumental transfer; SO-PIT) for natural reward. The effects on Pavlovian motivation were particularly striking in revealing sex differences. Shock during adolescence or adulthood promoted SO-PIT in females, relative to controls. In males, shock during adulthood impaired SO-PIT. Males shocked during adolescence, however, did not show this effect, showing similar SO-PIT to controls. Instrumental motivation, on the other hand, revealed a different pattern, with males generally being insensitive to the effects of shock, and female adolescents showing deficits. These findings suggest that the persistent effects of stress on Pavlovian and instrumental motivational processes differ in females and males, and that males may be less vulnerable to the effects of stress during adolescence.

Sex differences in the effects of stress on cue-triggered motivation

A growing literature suggests that females are particularly sensitive to the enhancing effects of stress on cue-mediated behaviors (Barker and Taylor, 2021; Feltenstein et al., 2011; Moran-Santa Maria et al., 2016; Mathews et al., 2008). Our findings extend this work by showing that stress enhances cue-triggered food-seeking in females whether stress occurred in adolescence or in adulthood. Epidemiological evidence suggests that PTSD is a stronger risk factor for obesity in females than males (Haskell et al. 2010; Perkonigg et al. 2009) and obesity vulnerability is linked to enhanced SO-PIT and general PIT expression in rats (Derman and Ferrario, 2018; 2020; Gladding et al. 2023). Thus, our current data support the idea that intense stress may render females more susceptible to obesity by augmenting the ability of food cues to drive food-seeking behavior.

In contrast to females, stress in adult males impaired SO-PIT, but stress in adolescence had no effect. These findings in adults are consistent with the literature on stress prior to PIT training, although stress after training but before PIT testing may have weaker effects (Derman and Lattal, 2022; Pielock, et al., 2013). It is interesting to note that qualitatively the absence of PIT in adult shocked males appears to be driven in part by modest augmentation in lever responding by CS− presentations. Though speculative, this failure to express PIT here might reflect some generalization of the excitatory properties of the CS+ to the CS−. We have previously shown that sensory-specific PIT is unaffected by shock; however, in that study the CSs were from distinct modalities (visual vs auditory), whereas with SO-PIT in the present experiments the CSs were both auditory. The shared modality of the CSs here may have encouraged a tendency to generalize in adult shocked males.

The absence of an effect in adolescence suggests that in males, adolescence may present a time of insensitivity to the long-lasting effects of stress on motivation for natural reward. Given that adolescence is associated with increased risk-taking, particularly in males, it makes adaptive sense that this age may also be accompanied by increased resilience to stress (Burke and Miczek 2015). Indeed, there is some evidence that adolescent males may be resilient particularly to the persistent effects of stress on appetitive behaviors (Bourke and Neigh, 2011; Toth et al., 2008b; Pohl et al., 2007). Rising testosterone, in combination with relatively high levels of corticosterone in adolescence may be responsible for this resiliency (Burke and Miczek 2014; Cueva et al. 2015; Russo et al. 2012).

It is also possible that male adolescents may be particularly sensitive to the motivational control of cues over instrumental responding. Indeed, Marshal et al. (2020) found that adolescent males will preferentially exhibit PIT, even under conditions that are optimized to promote magazine responding at the expense of PIT. Females were not included in their study, so it is not clear whether adolescent females show a similar propensity for stimulus-based instrumental motivation, but in our current study control females given adolescent PIT training exhibited lower PIT than males. Thus, female adolescents may not possess this same baseline amplification in stimulus-based motivation. Our finding here that stress in adult males blocks PIT, but that adolescent stress does not, may reflect the imperturbability of stimulus-based motivational control unique to adolescent males.

There is one more final consideration worth addressing with regard to the differential effects between adolescent and adulthood stress in males. In Experiment 1b comparisons were made between littermates shocked in adolescence versus adulthood. A complication in interpreting these data is that the time between shock and appetitive training is longer in the adolescent group compared to the adult group. This is an inherent complication in many developmental studies that aim to identify long-term effects of stress in youth. In an effort to address this limitation, we conducted Experiment 2b in which adolescent rats were given appetitive training immediately after shock. Here our effects were consistent with what we observed in the adolescent group in Experiment 1b. This increases our confidence that developmental age rather than time between shock and training accounted for the null effects of shock in males. Nevertheless, it remains to be seen whether introduction of an interval between adulthood stress and appetitive training produce different effects than the suppression of SO-PIT we have shown in males here and in Derman and Lattal (2022).

Shock in adolescence suppresses instrumental motivation in adult females

In Experiment 1 we found that females shocked in adolescence (but not in adulthood) exhibited reduced instrumental motivation (responding on a progressive ratio schedule) compared to controls when tested in adulthood. There is limited data on the effects of stress on instrumental motivation for females; however, Williams et al. (2022) found that chronic early life stress from PN2–9 reduced instrumental motivation for sucrose in females in adulthood, consistent with our effects in adolescent shocked females tested in adulthood. In males, we found no effect of shock on progressive ratio responding whether shock occurred in adolescence or adulthood. We have previously found that in adult males, shock leads to suppressed motivation for natural reward when testing occurred long after shock and after extensive appetitive training and experience with the outcome (Derman and Lattal, 2023). Others have found that early life stress in males suppresses PR for sucrose in adulthood (Campbel et al. 2017), and yet others have found the opposite (Williams et al., 2022). We found here that shock in adulthood that closely precedes testing has no effect on PR, whereas in our previous study PR was tested 45+ days after shock and we observed a strong suppression that persisted across repeated testing.

Taken together these studies suggest that time between shock and testing may be an important determinant in observing stress-associated suppression in instrumental motivation in adult males, as has been shown with other learning processes (Knox et al. 2012). The absence of an impact on instrumental motivation in males in the case of relatively recent stress is consistent with the finding that ongoing chronic mild stress does not impact PR responding for sucrose in males (Barr and Phillips,1998). A challenge to the interpretation that time is a critical factor in observing stress mediated suppression in instrumental motivation is that our adolescent shocked males tested in adulthood showed no effect of stress on PR. However, adolescence in males might signify a period of comparatively reduced sensitivity to the impacts of stress on motivation for natural rewards. Supporting this interpretation is our observation that shock during adolescence also does not affect SO-PIT, whereas shock during adulthood disrupts it, consistent with findings that adolescent stress in males does not suppress sucrose intake, whereas stress in adulthood does (Bourke and Neigh, 2011; Toth et al., 2008b; Pohl et al., 2007).

Males exhibit stronger fear generalization than females

In our experiments, rats received shocks in Context A followed by appetitive training in Context B. When shocked in adolescence, both sexes showed similarly low levels of freezing in Context B in adolescence, but when tested in Context B in adulthood, both sexes showed substantial freezing with increased freezing in males. Similarly, when shocked in adulthood, generalized freezing in Context B occurred in males but not in females. These findings suggest that males show increased generalized fear in adulthood, whether the shock occurred in adolescence or adulthood. This is consistent with other studies showing lower freezing levels in females (Inslicht et al.,2013; Gruene et al.,2015; Colon et al., 2018; Trott et al., 2022; but see Keiser et al., 2017; Lynch et al., 2013) and these sex differences could reflect any number of factors, including how contexts are processed (Asok et al. 2019; Colon and Poulos, 2020) or what behaviors are evoked by those contexts (Shanazz et al. 2022; Kirk et al., 1985; Steenbergen et al., 1990; Heinsbroek et al., 1991).

Our observed sex difference in freezing in adulthood, but not in adolescence, is consistent with findings that sex differences in emerge with age (Colon et al., 2018; Odynocki, et al., 2022). Colon et al. (2018) found a linear increase in fear generalization with age in males and linear decrease in females, consistent with our finding that adolescent females showed some fear generalization the day after shock, but adult females tested the day after shock showed no reliable fear generalization. They also found that this sex difference emerged with longer intervals between fear conditioning and testing, consistent with our findings in Experiment 2 of no sex difference in fear generalization in adolescence one day after shock, but an apparent sex difference in SEFL in these same rats when tested in late adolescence.

Finally, in Experiment 1 we observed a weaker SEFL effect when the Context A shocks occurred in adolescence compared to adulthood. Furthermore, this effect in adolescence was completely reversed by extinction of Context A. Although underpowered to detect sex differences, this general finding suggests that the effects of stress during adolescence may be more readily reversible by extinction. More work is clearly needed on developmental effects of SEFL and on the conditions in which extinction weakens SEFL (Hassien, et al. 2020; Rau et al. 2005; Williams, et al. 2019).

Stress transiently suppresses responding early in appetitive training in males more than females

Sex differences in freezing generally corresponded to effects on appetitive responding during the magazine training and continuous reinforcement phases of training. Although it is likely that the generalized freezing seen in shocked rats contributed to the magazine response suppression effect, it cannot fully account for it. In Experiment 1, we saw minimal Ctx B freezing in both females and males shocked in adulthood, but we found robust suppression in magazine responding in males, and no suppression in females. Whereas in Experiment 2, males and females showed similarly moderate levels of freezing, but only females showed suppressed magazine responding. Both of these observations highlight incongruity between the Ctx B generalized freezing and the suppression of magazine responding. Thus, although it is likely that some of the magazine suppression resulted from response competition with freezing, appetitive behavior was suppressed even when freezing did not occur. Given that these behaviors appear somewhat dissociable, they may present ideal endpoints for distinguishing factors that contribute to persistent effects of intense stress on aversive versus appetitive processes.

Neurobiological Considerations

Our approach to induce acute stress here is based on the SEFL approach developed by Fanselow and colleagues (Rau et al., 2005). This approach has been shown to cause long duration changes in behavior that last at least 90 days (Blouin et al., 2016). SEFL in early life (P19) is associated with changes in glucocorticoid and neuropeptide Y receptor expression in the amygdala in adulthood (Poulos et al., 2014). Chronic stress in adolescence alters developmental changes that would normally occur within prefrontal cortex-basolateral amygdala (BLA-PFC) circuit and these changes impact fear expression and reward-seeking in adulthood (Sillivan et al., 2020; Uliana et al., 2021). In addition, the PFC and BLA have both been implicated in expression of PIT (Keistler et al., 2015; Sias et al., 2021). Thus, it is possible that the persistent effects we observe here are mediated by stress-induced alterations in these circuits.

With respect to sex differences, the Nucleus Accumbens (NAc) may play a role in our observed effects of stress on PIT. The NAc is critical for PIT (Hall et al., 2001), and it exhibits several sexually dimorphic characteristics including differences in glutamatergic inputs and cell morphology (Cao et al. 2016; Forlano and Woolley 2010). Moreover, stress produces sex-specific alterations to the transcriptomic profiles in the NAc (Hodes et al. 2015; Ordoñes Sanchez et al. 2021). This suggests that the NAc is a good candidate nucleus for future examination of sex-specific mechanisms of the persistent effects of stress on reward-seeking.

Conclusions

We found striking sex differences in the effects of shock on fear generalization and cue-triggered motivation. In humans there are well established differences in the rates of PTSD diagnoses between females and males and the greater prevalence of comorbidities of appetitive motivational disorders in females. In females the effect of stress on stimulus-based reward-seeking is consistent between natural reward and drug reward. In males, stress has been shown to persistently augment drug-seeking (Pizzimenti et al., 2017), but our data here and in Derman and Lattal (2022, 2023) show that stress compromises reward seeking using natural reward. This discrepancy between drug and natural reward seeking in males may reflect differences in the approaches taken (cue-induced reinstatement vs PIT) or they may indicate a genuine disparity in the effect of stress on seeking of natural and drug rewards (Hanson, et al. 2021).

Highlights.

Acute stress during adolescence or adulthood enhanced Pavlovian motivation in females.

Acute stress during adulthood impaired Pavlovian motivation in males, but adolescent stress had no effect.

Data availability statement:

The data that support the findings of this study are available from the corresponding author, KML, upon request.