Coping with the experience of frustration throughout life: Sex-and age-speci ﬁ c effects of early life stress on the susceptibility to reward devaluation

Early life stress may lead to lifelong impairments in psychophysiological functions, including emotional and reward systems. Unpredicted decrease in reward magnitude generates a negative emotional state (frustration) that may be involved with susceptibility to psychiatric disorders. We evaluated, in adolescents and adult rats of both sexes, whether maternal separation (MS) alters the ability to cope with an unexpected reduction of reward later in life. Litters of Wistar rats were divided into controls (non handled − NH) or subjected to MS. Animals were trained to ﬁ nd sugary cereal pellets; later the amount was reduced. Increased latency to reach the reward ‐ associated area indicates higher inability to regulate frustration. The dorsal hippocampus (dHC) and basolateral amygdala (BLA) were evaluated for protein levels of NMDA receptor subunits (GluN2A/ GluN2B), synaptophysin, PSD95, SNAP ‐ 25 and CRF1. We found that adult MS males had greater vulnerability to reward reduction, together with decreased GluN2A and increased GluN2B immunocontent in the dHC. MS females and adolescents did not differ from controls. We concluded that MS enhances the response to frustration in adult males. The change in the ratio of GluN2A and GluN2B subunits in dHC could be related to a stronger, more dif ﬁ cult to update memory of the aversive experience.


Introduction
Throughout the course of life, individuals face many challenges, not all of which have the desired outcome.Frustration involves an unfulfilled previous expectation (Amsel, 1962;Papini, 2003).It may occur by the omission or reduction in quantity or quality of an expected reward (Papini et al., 2022;Psyrdellis et al., 2016).In animals, an unpredicted decrease in the magnitude of reward generates a negative emotional state, which leads to a transient performance decrement in learned rewarded responses (Conrad et al., 2022;Manzo et al., 2014;Salinas et al., 1997) and leads the animal to escape the place where the omission occurred, creating aversion (Papini, 2003;Papini et al., 2022).Reward omission leads to frustration which can be demonstrated by the difference in the speed at which rats reach the target previously associated with reward (reviewed by Papini, 2003): evidence shows that rats running at slower speeds on a track using food as reinforcement indicates a measure of frustration behavior in these animals (Amsel, 1962;Salinas et al., 1993Salinas et al., , 1997)).The unexpected loss of reward may lead to lasting aversive effects, while quick recovery from frustration signals adaptive, healthy cognitive functioning; on the other hand, maladaptive coping with frustration is involved with susceptibility to aggressive behaviors, emergence of addictive behaviors and reduced social aptitude in general (Donaire et al., 2022;Dugré and A well-established body of evidence, both in animal models and in human cohorts, suggests that early life stress (ELS) affects brain function, with lifelong effects (J.Alves et al., 2022;Bolton et al., 2017;Malinovskaya et al., 2018;Taylor, 2010).These changes may affect emotional and cognitive functions, enabling adaptation to the environment (Gee et al., 2015); however, they may also increase vulnerability to psychopathologies (Arcego et al., 2024;Kessler et al., 2010;Ohta et al., 2023;Tsotsokou et al., 2021).Both psychiatric disorders (Kessler et al., 2010;Taylor, 2010), as well as changes in cognitive functions, including hindered emotional memory regulation and cognitive flexibility (M.B. Alves et al., 2019;Couto Pereira et al., 2019;Dalmaz et al., 2021;de Lima et al., 2020;Kraaijenvanger et al., 2020;Lazzaretti et al., 2018) may be favored by ELS.
Animal models of ELS have consistently shown that early experiences shape how memories are acquired, maintained and retrieved, particularly those associated with negative emotions (J.Alves et al., 2022;Krugers et al., 2017).The egocentric memory of the frustration event (Papini, 2003), particularly its emotional charge, contributes to maintaining the decreased performance in learned rewarded responses (Salinas et al., 1993(Salinas et al., , 1997)); in fact, interfering with memory consolidation immediately after reward devaluation, results in a rapid return to the preshift search behavior (Ortega et al., 2014;Salinas et al., 1997).The ability to update the memory of the frustrating event facilitates the development of tolerance to the effects of non-reward (Papini, 2003;Papini et al., 2022), improving coping with frustration.Hence, the tendency to form strong negatively-charged memories (J.Alves et al., 2022;Callaghan & Richardson, 2013;Kosten et al., 2007), resistant to update by reconsolidation (Couto-Pereira et al., 2019), may hinder ELS animals ability to adapt to frustration.
The hippocampus and amygdala, amongst other structures, have important roles in learning, memory, and control of emotional behavior (Izquierdo et al., 2016).The dorsal region of the hippocampus (dHc), specifically, is implicated in the cognitive portion of episodic memories (Fanselow and Dong, 2010) and its contextual novelty detection and bidirectional interaction with the amygdala during memory retrieval is implicated in aversive memory update processes (Ferrara et al., 2019;Seidenbecher et al., 2003).Both amygdala and hippocampus are affected by ELS: maternal separation (MS) causes structural and functional synaptic changes in the dHc (Loi et al., 2017;Murthy et al., 2019;Reincke & Hanganu-Opatz, 2017) and the basolateral amygdala (BLA) (Koe et al., 2016), including altered dendritic arborization and spine density (Koe et al., 2016), and impaired synchrony in the dHc-BLA circuit during aversive memory reactivation (Couto Pereira et al., 2019, 2023).The developmental trajectory of both structures is also affected by ELS (Gee et al., 2015;Andersen et al., 2004;Callaghan & Richardson, 2013).
The ability of neurons to modify the number or effectiveness of their synapses is essential for neuroplasticity, allowing adaptation to the environment (Leslie et al., 2011), but may also lead to maladaptive behaviors and have detrimental effects.Changes in both synaptic proteins or receptors (both their amount and/or composition) may underlie the effects of stressful situations on cognition and behavior (Đorović et al., 2024;Karst et al., 2023;Wu et al., 2022).In this regard, variations in the strength of emotional memories may be mediated by the ratio of GluN2A/GluN2B subunits that make up N-methyl-D-aspartate receptors (NMDARs) (S.-H.Wang et al., 2009;Holehonnur et al., 2016).Different subunit composition confer particular properties to these receptors, including ligand affinity, open probability and deactivation kinetics, and mobility, affecting synaptic plasticity (Tian et al., 2021).Indeed, GluN2A and GluN2B ratio adjustment following experiences influences the plasticity of a given memory (for reviews, see Zhang et al., 2018;Ladagu et al., 2023), and early environment may affect this ratio (Pickering et al., 2006).Furthermore, a possible mediator of ELS effects on adult hippocampal and amygdala functions is the persistent corticotropin releasing hormone (CRF) signaling, which is believed to contribute to cognitive and structural impairments leading to maladaptive behaviors (Alcántara-Alonso et al., 2024;Couto-Pereira et al., 2016;Fenoglio et al., 2005;Ivy et al., 2010;Wang et al., 2011), including sensitivity to frustration (Micioni Di Bonaventura et al., 2014).
Another important factor when assessing the long-term consequences of ELS is sex.Preclinical studies suggest that exposure to early adversities differently affects the neurodevelopment of males and females, leading to sex-specific behavioral, endocrinal and neurobiological patterns in adulthood (Lajud & Torner, 2015;van Bodegom et al., 2017;Zanta et al., 2021).These effects are further influenced by the menstrual/estral cycles phase at the time of assessment, as circulating sex hormones modulate the response to stress (Montero-López et al., 2018).Nevertheless, early organizational effects are probably very important in determining how ELS will influence behavior.A second important factor is the age at which outcomes are assessed.Previous studies have reported different behavioral effects of neonatal interventions when the evaluations are made during development (e.g., in the periods of childhood or adolescence) or in adulthood (Leichtweis et al., 2020;Silveira et al., 2006Silveira et al., , 2008)).Furthermore, diverse effects of ELS on neurochemical parameters are also observed when measurements are taken in adolescents or adult animals (Leichtweis et al., 2020;Silveira et al., 2006;Sinani et al., 2022).
We believe that a deeper understanding of how ELS affects longterm behavior and vulnerability to disorders will allow the development of more accurate and individualized treatments for individuals at risk (de Lima et al., 2023).Animal models are valuable when assessing complex psychobiological phenomena, since cultural effects are removed from the scenario, allowing the study of the neurobiological aspects in an isolated manner (Bolton et al., 2017).Here, we evaluated whether MS, a widely used animal model of ELS, alters the animals' ability to cope with an unexpected devaluation of reward later in life.For this, we used different sets of animals, which were evaluated at different stages of life (adolescence and adulthood) in a task of unexpected reward devaluation.We also examined the involvement of the dHc and BLA in the update of the memory of reward magnitude, by verifying N-methyl-D-aspartate (NMDA) receptor subunits, CRF receptor type 1 (CRF1), and pre-and post-synaptic proteins (Synaptosomal-Associated Protein 25 − SNAP-25, synaptophysin and Postsynaptic Density Protein 95 − PSD95).We tested the hypothesis that animals subjected to ELS and exposed to a frustrating experience would show a different adjustment ratio of NMDARs subunits and in proteins related to synaptic plasticity.We also tested the possibility that CRF1 could present increased baseline levels in the MS animals, due to a programming effect of ELS.To our knowledge, this is the first work assessing how an animal model of early adversity affects the susceptibility to frustration throughout development.

Subjects
All procedures were approved by the institutional Research Ethics Committee (CEUA-UFRGS #35364) and followed the Brazilian Law for the use of animals (Federal Law 11.794/2008) and the National Institute of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80-23, 1996).Care was taken to minimize animal suffering during the experiments.
Primiparous pregnant Wistar rats bred at our animal facility were used.Around gestational day 18, they were single-housed in home cages made of Plexiglas (38 × 32 × 17 cm) with sawdust-covered floors and kept in a controlled environment (standard 12 h dark/light cycle, lights on from 7:00 to 19:00, temperature at 22 ± 2 °C, standard rat chow and water provided ad libitum).The day of birth was considered postnatal day 0 (PND 0).All litters were randomly culled to 6-9 pups within 24 h after birth and were randomly assigned to one of the two neonatal interventions (see below).Weaning was performed on PND 21: males and females were housed 3-4 per cage and remained under standard animal facility conditions until the beginning of the experiments.A maximum of 2 animals of each sex and each litter were used for the same group for the behavioral experiment (shifted or unshifted, as described below), and only 1 animal of each sex and each litter was used for the biochemical experiments.

Maternal separation (MS)
From PND 1-10, once a day, pups were gently placed together in a clean box lined with a paper towel, in a warm bath set to 32 °C, where they remained for 3 h.After this period, pups were returned to their respective cages.This procedure was performed during the lights-on cycle, between 12:00 and 17:00.Controls were non-handled (NH) animals: pups and dams were left undisturbed until weaning.From birth to weaning, cleaning of the housing box was carried out by carefully removing dirty sawdust and replacing them with clean sawdust, avoiding the nest area.Each litter was handled with a distinct glove to avoid the spread of odors between nests.During the period of separation, dams remained in their homecage, inside the same room, so they could hear the pups' vocalizations (Diehl et al., 2014).
On PND32 or around PND 77 (3 days prior to the task), the animals' food supply was reduced to and maintained at 80 % of the amount they used to consume.Concurrently, habituation to the consumption of sweet cereals was carried out, each rat receiving 1 unit of the cereal (Kellogg's Froot Loops®, Kellanova, USA) in the housing cage.
The reward devaluation task was adapted from previous studies (Salinas et al., 1993(Salinas et al., , 1997)).Experiments took place from 12:00 to 18:00 h.Before the beginning of each session, animals were taken to the room where they remained for 20 min, for acclimation.The apparatus was a straight-alley maze (12 W × 110 L × 41 H cm), made of MDF with black walls and floor.The designated amount of sweet cereal pellets was placed at the center of a digitally delimited target area (12 × 12 cm), located in one end of the corridor.The animals were always placed at the opposite end of the corridor.
Animals were randomly assigned to either the shifted (receiving 10 pellets of cereal per trial) or unshifted (receiving 1 pellet of cereal per trial) groups, at the beginning of the experiment.One cereal pellet corresponded to 1/4 of a Froot Loops® unit.Training consisted of six trials per day, separated by a 30-s intertrial interval, for 10 days.On Day 11, shifted groups received only 1 pellet per trial.This reduction was maintained for 2 additional days (days 12 and 13).All experiments were recorded and the latency (in secondss) to reach the target, distance covered in the apparatus (in meters) and the number of entries in the target area were determined using the behavioral analysis software ANY-maze (Stoelting Co, IL, USA; RRID:SCR_014289).The number of pellets consumed in each trial was also registered.Rats that did not eat any pellets in 50 % of the trials or more at training days 9 and 10 were excluded (n = 22).
Latency was averaged across 2 blocks of trials, corresponding to the first 3 trials vs. the last 3 trials of each day.The number of entries in the target area was averaged across the 6 trials of each day.

Estrous cycle phase determination
To determine the phase of the estrous cycle (estrus, metestrus, diestrus and proestrus) of the adult females, vaginal lavage was performed using a pipette with a blunt polyethylene tip, containing 40 µl of 0.9 % saline solution.The tip was introduced into the vaginal orifice and the saline solution was injected, aspirated and placed on a previously identified slide and visualized fresh under a microscope, at 10x and 40x magnification (Marcondes et al., 2002).This procedure was performed during all 13 days of behavioral experiments, after the end of the reward devaluation task, and also before euthanasia.

Biochemical analyses
Animals were randomly selected to be used in neurochemical evaluations (totalizing 7 animals per group).Twenty-four hours after the last behavioral session, they were quickly killed by decapitation.Fresh brain tissue was dissected on ice.To dissect the dHc and BLA, coronal brain slices of 2 mm were cut using an acrylic brain matrix (#AL-1160, Alto).Structure boundaries were identified using a rat brain atlas (Paxinos & Watson, 1998).Once dissected, samples from both hemispheres were stored at − 80 °C until further analysis.

Statistical analysis
Data were analyzed using the software SPSS version 23.0 (IBM, NY, USA; RRID:SCR_019096).Mauchly's test was used to assess data sphericity when using repeated measures analysis of variance (ANOVA).If sphericity assumption was violated (p < 0.05), the Greenhouse-Geisser correction was used.Repeated measures ANOVA was used for latency and target area entries, with MS, sex, and group (shifted vs. unshifted) as independent factors; additionally, repeated measures ANOVA was also employed to analyze latency in females using estrous cycle (follicular and luteal phases), MS and group (shifted vs. unshifted) as independent variables.Two-way ANOVA, with MS and sex as factors, were performed for locomotion and food consumption on day 10 of the behavioral experiment.All behavioral analyses were adjusted for batch.Two-way ANOVA with MS and group (shifted vs. unshifted) as factors was used for all biochemical results.When appropriate, post hoc pairwise comparisons were made using the Bonferroni adjustment for multiple comparisons.Data are expressed as the mean ± standard error of the mean (SEM) for latency (which results from averaging trials data from the same animal), boxplots indicating median, interquartile interval and range for integral data (number of entries and pellets consumed), and mean ± standard deviation (SD) for biochemical data.Statistical significance was set at p < 0.05.Plots were built using GraphPad Prism v. 6.01 (GraphPad Software Inc.; RRID:SCR_002798)).

Results
Adult male animals that went through maternal separation are more susceptible to reward devaluation For the reward devaluation task, animals were trained in an apparatus where they were presented with a reward.Training was performed for 10 days (6 trials per day), during which animals received the same amount of reward (unshifted groups received a small reward and shifted groups received a higher reward).On Day 11, an unexpected reward devaluation was introduced for shifted animals (see Fig. 1A and 2A for a schematic representation of the task).
Considering entries in the target area, a virtual square at the center of which the reward was placed, over days 10 to 13, an interaction between day and group (shifted and unshifted) was found [F (3,198) = 13.99,p < 0.001], together with an effect of day [F(3, 198) = 9.29, p < 0.001] (Fig. 1D).An increased number of entries was observed on days 11 to 13 in the shifted groups in the face of an unexpected reduction in reward, when compared to day 10.No effects of MS or sex or interaction between these two factors were observed.On day 10, before the unexpected reward reduction, the consumption of sweet cereals by the groups receiving a higher reward showed only an effect of sex [F(1, 32) = 4.96, p = 0.033], whereby adolescent females consumed a smaller amount of pellets than males (2.9 [2.4-3.8] vs. 3.8 [3.0-4.9]).No effect of MS or interactions were found (p > 0.05; see Fig. 1 E).No effects of MS or sex or interactions were observed in the distance covered in the apparatus on day 10 (p > 0.05, data not shown).
When evaluating the performance of adult animals specifically on day 11, when the shift in reward magnitude was introduced, there was a significant interaction between group, sex and trials block [F (1, 110) = 4.630, p = 0.034], with latency increasing in the second block of attempts for male animals in the shifted group (trials 64-66), when faced with the perception of the unexpected reduction of the reward (Fig. 2B and 2C).Importantly, an interaction between trials block, sex and MS was also observed [F(1, 110) = 4.478, p = 0.037], with MS-shifted male animals presenting higher latencies to reach the target in the second block of trials (Fig. 2B, block 64-66), compared to MS-shifted females (p = 0.008; pairwise comparisons performed with Bonferroni adjustment for multiple comparisons).Using estrous cycle (follicular and luteal phases, corresponding to the proestrus-estrus and metestrus-diestrus phases, respectively (Antunes et al., 2016;Emanuele et al., 2002)) as a covariate in a females-only analysis for day 11, we found a significant interaction trials block x group x estrous cycle phase [F(1, 69) = 6.383; p = 0.014], whereby shifted females in the luteal phases had a higher increase in latency in the second block of trials than those in the follicular phases (luteal: from 6.22 ± 6.64 s to 11.18 ± 12.55 s; follicular: from 5.02 ± 2.36 s to 6.84 ± 2.96 s, see Supplementary Material − Figure S1).No interactions between estrous cycle and MS were present (p > 0.05).
In entries in the target area, over days 10 to 13, an interaction between day, group (shifted and unshifted) and MS [F(3, 330) = 4.343, p < 0.001], an interaction between day and group [F(3, 330) = 13.956,p < 0.001], and an effect of day [F(3, 330) = 4.744, p = 0.003] were found (Fig. 2D).An increased number of entries was observed on days 11 to 13 in the animals in the shifted groups in the face of an unexpected reduction in reward, when compared to day 10.This effect appears to be more relevant for MS animals.
We controlled for the consumption of sweet cereals on day 10, the day before the unexpected reward reduction, by groups receiving a higher reward.No effects of sex or MS were found (p > 0.05 in both cases; see Fig. 2E), suggesting that effects found from Day 11 on were not caused by a previous difference in learning the task or different preference for sweet foods.The distance covered by the animals in the apparatus on day 10 showed an effect of sex, with females presenting higher locomotor activity than males [Females: 3.768 ± 0.152 m; Males: 3.221 ± 0.239 m; F(1, 110) = 62.770, p < 0.001], which is expected (Bishnoi et al., 2021), and an effect of group, as shifted animals explored the apparatus less than unshifted ones [unshifted: 4.416 ± 1.303 m; shifted: 2.810 ± 1.062; F(1, 110) = 6.820, p = 0.010], without significant interactions with MS (p > 0.05).Since shifted animals received 10 pellets, they spent most of the trial time eating the reward, leading to decreased locomotion.This shows that the changes found in latency are not due to effects on motricity or preference for sugar rewards caused by MS.
For detailed statistical results of the behavioral measures, please refer to the Supplementary Material − Tables S1 and S2.
Male MS animals that went through the experience of reward devaluation showed a reduction in GluN2A and an increase in GluN2B subunits immunocontent in the dorsal hippocampus The immunocontent of relevant receptors and synaptic proteins, 24 h after day 13 of the reward devaluation task, was analyzed separately for adult males and females, by Western blot.In the dHc of adult males, we observed an interaction between group (shifted or unshifted) and MS on the immunocontent of the GluN2A subunit of the NMDA receptor [F(1, 24) = 8.710, p = 0.007], since animals from the shifted group that went through MS had a reduction in this protein compared to unshifted MS rats.No difference was observed on the immunocontent of this subunit in NH animals (Fig. 3A).An interaction group x MS was also observed for the GluN2B subunit [F(1, 24) = 9.686, p = 0.005] (Fig. 3B).However, for this subunit, an increase was observed in MS animals in the shifted group.No significant effects were observed for the immunocontent of synaptophysin, SNAP-25, PSD95, or CRF1 receptor (p > 0.05 for all proteins; Fig. 3C-F).
The same proteins were evaluated in the BLA.No effects were observed for any of them, either in males or females (p > 0.05 for all proteins; see Fig. 4).
For detailed statistical results of the neurochemical measures, please refer to the Supplementary Material − Tables S3 and S4.

Discussion
In this study, we observed that ELS induced differential vulnerability to frustration linked to unexpected reduction of a reward, and this effect was age-and sex-specific: adult male rats subjected to MS showed a delay in adapting to a new disadvantageous reward situation compared to control animals and MS females, while as adolescents all groups showed similar response to frustration.We also observed a reduction in the immunocontent of the GluN2A subunit and an increase in the GluN2B subunit of the NMDA receptor in the dHc of male adult rats subjected to ELS, while no effect was observed in females.
Adverse experiences during development can program the organism to respond differently to unfavorable situations in adulthood.To our knowledge, this is the first study evaluating how ELS, more specifically MS, influences the response to unexpected reward reduction.Adversities early in life are known to lead to lasting neurobiological changes, both in humans (Teicher, 2018;Teicher et al., 2016) and in rodents (Gildawie et al., 2020;Honeycutt et al., 2020;Majcher-Maślanka et al., 2019).However, studies on the long-term effects of ELS have focused mainly on male adults, and fewer studies have considered the effects on adolescent behavior, a period of increased susceptibility to the consequences of maladaptive response to frustration, such as risk-taking (Avramescu et al., 2023;Noschang et al., 2021), and aggressive behaviors (Dugré & Potvin, 2023).This prompted us to study the effects of ELS on frustration, comparing both sexes throughout two developmental phases, adolescence and adulthood.
Frustration was induced by reducing the magnitude of an habitual reward; it arises from the violation of a prior expectation (Papini, 2003;Papini et al., 2022).The stronger frustration observed in male animals in the MS group was evidenced by the increased time to reach the target after the animals realize that there had been a reduction in reward (Coleman-Mesches et al., 1996;Leszczuk & Flaherty, 2000;Salinas et al., 1996Salinas et al., , 1997;;Salinas & McGaugh, 1996).The animals that underwent reward reduction had an increased latency to reach the target in the second block of trials on the shift day, showing that their perception of this decrease happens gradually, as expected in this task (Conrad et al., 2022;Salinas et al., 1997).Our results are suggestive of higher vulnerability to frustration in MS male animals, which appears to have emerged in adolescence and remained through adulthood.ELS induces earlier maturation of the prefrontal cortex-amygdala circuit (Gee et al., 2015) and delayed synaptic development of the hippocampus (Andersen et al., 2004), resulting in a profile of altered emotional regulation and emotional learning (Callaghan & Richardson, 2013) which may impact the lifelong maladaptive responses to stressful events observed in MS males, such as the susceptibility to frustration shown here.This is in accordance with the resistance to reconsolidation of aversive memories (Couto Pereira et al., 2019), anxiety-like (de Lima et al., 2020;Kambali et al., 2019), and depressive-like behaviors (Marais et al., 2008), also exhibited by male MS animals.
Differently, adolescent MS females exhibited a delayed response to frustration, i.e., they took longer to respond to reward omission and adapt to the new contingencies.This characteristic dimmed as adults, and these animals presented performance similar to NH females, showing some resilience to frustration.In fact, adult MS females often display attenuated behavioral and physiological consequences of ELS, when compared to MS males (Couto-Pereira et al., 2016;Diehl et al., 2007;Dimatelis et al., 2016;Lee et al., 2020;Roman et al., 2004), which may be related to the recently reported absence of changes in the prefrontal cortex-amygdala circuit connectivity in MS females (Haikonen et al., 2023).The resilient behavioral profile was lower for both female NH and MS rats when reward devaluation was presented in the luteal phase, compared to the follicular phase; this is in accordance with the increased vulnerability to emotional distress when the levels of plasma progesterone rise, observed both in female humans (Steiner, 1997) and rats (Devall et al., 2009;Nin et al., 2012).As a measure of a possible coping strategy in the face of unexpected reward reduction, we analyzed the number of entries into the target area.The increase in the number of entries can be interpreted as noncompliance with the reduction in reward and difficulty in dealing with frustration.In fact, after the reward reduction there was an increased number of entries in the MS group, representing the search for the lost reward (Amsel, 1962;Papini, 2003).The consumption of sweet cereals as well as the locomotion of the animals were also evaluated during the day preceding the reward devaluation (day 10), since both mobility and differences in the value (valence) of the reward could influence the latency after the shift.Taking these variables into account, no significant differences were found between groups, corroborating the interpretation that the results obtained in this task (increased latency and increased number of entries after reward reduction) reflect frustration in the face of a reduced reward, and not previous behavioral characteristics.
Along with the behavioral findings, changes in the immunocontent of NMDARs subunits, namely a decrease in GluN2A and increase in GluN2B levels, were observed in the dHc of MS males 3 days after the experience of frustration, compared to unshifted MS males, but not in the other animals.The update of the egocentric memory of the frustration event may play an important role in the adaptive strategy (Papini, 2003).Memory update can occur after reactivation and depends on processes such as reconsolidation or extinction.Memory reactivation requires a transient increase in synaptic NMDA receptors containing the GluN2B subunit, in detriment of GluN2A (Zhang et al., 2018;Wang et al., 2009;Holehonnur et al., 2016), which is thought to be responsible for the destabilization process.Male MS animals that underwent the experience of frustration maintained the maladaptive behavior throughout days 11 (the shift day) and the following day (day 12, postshift).The lack of molecular shifts in the composition of the dHc NMDA receptor of the NH shifted males (compared to NH unshifted males) possibly indicates that a "stable" GluN2A/ GluN2B ratio was already established by the time samples were collected.For MS animals, on the other hand, memory update processes were possibly delayed compared to male controls and that may explain why we were able to detect them on the third day postshift.The combination of behavioral and molecular results indicate that the frustrating event had a greater negative emotional valence for the MS males, delaying the adaptive coping to a new circumstance in a changing environment (Papini, 2003).In females, no differences were found between groups in the immunocontent of NMDA receptor subunits in dHc, which supports the profile proposed here of resilience of MS females in the face of a situation of unexpected reward reduction.In that regard, females that went through ELS presented a similar response to controls, both in their behavior and in the hippocampal molecular parameters related to the (re)consolidation of an aversive memory in face of a frustrating event (Cuenya et al., 2012;Riemer et al., 2018).
We also evaluated the immunocontent of synaptic plasticity proteins in the dHc.SNAP-25, which is involved in regulating the release of neurotransmitters from synaptic vesicles and stabilizing dendritic spines (Kádková et al., 2019), showed a marginal reduction in MS males, but not in females.Regarding PSD95, which participates in anchoring NMDA receptors in the postsynaptic terminal (Won et al., 2016), there were no differences between the groups.This result reinforces the hypothesis that in MS males that underwent an unexpected reduction in reward there was no reduction in the number of NMDA receptors, but rather an exchange of GluN2A subunits by GluN2B.
No differences were found when analyzing the immunocontent of GluN2A and GluN2B subunits of NMDA receptors in the BLA.The amygdala is one of the structures involved in the response to unexpected reward devaluation (Judice-Daher et al., 2012;Kawasaki et al., 2015;Salinas & McGaugh, 1996), especially in the perception of a decrease in the value of that reward (Kawasaki et al., 2017).Nevertheless, considering that these evaluations were made after the third day of unexpected reward reduction, this result is in line with the suggestion that the amygdala is not a permanent storage site of this memory/perception of the aversive stimulus, but it is involved with the modulation of emotionally significant information that is stored elsewhere (Salinas et al., 1996).Furthermore, it was already observed that molecular differences in memory update between MS male rats and controls arise mainly in the dHc and not in the BLA (Couto Pereira et al., 2019, 2023).The immunocontent of the synaptic plasticity proteins SNAP-25, synaptophysin and PSD95 also showed no differences between groups in the BLA.
One limitation of this study concerning the neurochemical results is the timing of sample collection (3 days postshift), when the other animals (females and NH males) had already adapted to the task, as expected.In this sense, a measure shortly after the reward reduction (1 or 6 h postshift, for example), could add more strength to our claim concerning the delayed memory update in frustrated MS males.The analysis of the ventral hippocampus would also provide relevant information to understand the response to frustration in MS rats.
Our results demonstrated that MS affects the response to unexpected reward reduction, more prominently in male animals.Changes induced by the experience of reward reduction in the immunocontent of the GluN2A and GluN2B subunits of the NMDA receptor in the dHc of male animals that underwent MS in the neonatal period suggest an exchange of these subunits, suggesting a reinforced aversive memory of the reward reduction.Corroborating the resilience profile of adult MS females, no behavioral changes were found, as well as no alterations were detected in the immunocontent of proteins and receptors analysed.
Overall, our results add-on to the existing literature on adverse early life environments and their impacts on brain development, and highlight sensitivity to frustration as one possible mechanism contributing to increased susceptibility to mental disorders observed later in life in these individuals.This work also underlines the importance of considering sex-differences when studying the lasting effects of ELS on behavioral patterns and expands information on the effects of early stress on memories with negative emotional valence.

Fig. 1 .
Fig. 1.Behavioral responses to unexpected reward devaluation for days 10 (last day of training), 11 (reward reduction − shift), 12 and 13 (1 and 2 days postshift, respectively), in adolescent males and females subjected to postnatal maternal separation − MS or non-handled − NS. A. Schematic representation of the experimental timeline.B. Latency to reach the target area (in seconds) grouped in 3-trial blocks."Unshifted" gray line depicts the mean of the 4 unshifted groups, displayed in detail in panel C. C. Latencies for the unshifted groups.D. Number of entries in the target area per day (mean of the 6 trials).E. Number of sweet cereal pellets consumed by shifted animals on day 10.Data is presented as mean ± SEM.Repeated measures ANOVA using MS, sex, group (shifted/unshifted) and trials block as factors, adjusting for batch, was performed; shifted and unshifted groups are shown in separate plots to allow better visualization.n = 5-14 animals per group.

Fig. 2 .
Fig. 2. Behavioral responses to unexpected reward devaluation for days 10 (last day of training), 11 (reward reduction − shift), 12 and 13 (1 and 2 days postshift, respectively), in adult males and females subjected to postnatal maternal separation − MS or non-handled − NS. A. Schematic representation of the experimental timeline.B. Latency to reach the target area (in seconds) grouped in 3-trial blocks."Unshifted" gray line depicts the mean of the 4 unshifted groups, displayed in detail in panel C. C. Latencies for the unshifted groups.D. Number of entries in the target area per day (mean of the 6 trials).E. Number of sweet cereal pellets consumed by shifted animals on day 10.Data is presented as mean ± SD.Repeated measures ANOVA using MS, sex, group (shifted/unshifted) and trials block as factors, adjusting for batch, was performed; shifted and unshifted groups are shown in separate plots to allow better visualization.n = 8-22 animals per group.

Fig. 3 .
Fig. 3. Biochemical analysis of the dorsal hippocampus (dHc) of adult males and females subjected to postnatal maternal separation − MS or non-handled − NS, 1 day after the last postshift session (day 13) of the unexpected reward reduction task. A. Immunocontent of the GluN2A subunit of the NMDA receptor, in males.B. Immunocontent of the GluN2B subunit of the NMDA receptor, in males.C. Immunocontent of the corticotropin releasing factor receptor type 1 − CRF1, in males.D. Immunocontent of the Postsynaptic Density Protein 95 − PSD95, in males.E. Immunocontent of Synaptophysin, in males.F. Immunocontent of Synaptosomal-Associated Protein 25 − SNAP-25, in males.G. Immunocontent of the GluN2A subunit of the NMDA receptor, in females.H. Immunocontent of the GluN2B subunit of the NMDA receptor, in females.I. Immunocontent of the corticotropin releasing factor receptor type 1 − CRF1, in females.J. Immunocontent of the Postsynaptic Density Protein 95 − PSD95, in females.K. Immunocontent of Synaptophysin, in females.L. Immunocontent of Synaptosomal-Associated Protein 25 − SNAP-25, in females.Data is presented as mean ± SD.ANOVA using MS and group (shifted/unshifted) as factors was performed.n = 6-7 animals per group.

Fig. 4 .
Fig. 4. Biochemical analysis of the basolateral amygdala (BLA) of adult males and females subjected to postnatal maternal separation − MS or non-handled − NS, 1 day after the last postshift session (day 13) of the unexpected reward reduction task. A. Immunocontent of the GluN2A subunit of the NMDA receptor, in males.B. Immunocontent of the GluN2B subunit of the NMDA receptor, in males.C. Immunocontent of the corticotropin releasing factor receptor type 1 − CRF1, in males.D. Immunocontent of the Postsynaptic Density Protein 95 − PSD95, in males.E. Immunocontent of Synaptophysin, in males.F. Immunocontent of Synaptosomal-Associated Protein 25 − SNAP-25, in males.G. Immunocontent of the GluN2A subunit of the NMDA receptor, in females.H. Immunocontent of the GluN2B subunit of the NMDA receptor, in females.I. Immunocontent of the corticotropin releasing factor receptor type 1 − CRF1, in females.J. Immunocontent of the Postsynaptic Density Protein 95 − PSD95, in females.K. Immunocontent of Synaptophysin, in females.L. Immunocontent of Synaptosomal-Associated Protein 25 − SNAP-25, in females.Data is presented as mean ± SD.ANOVA using MS and group (shifted/unshifted) as factors was performed.n = 6-7 animals per group.