Social Instability in Adolescence Differentially Alters Dendritic Morphology in the Medial Prefrontal Cortex and its Response to Stress in Adult Male and Female Rats

Abstract

Adolescence is an important period for HPA axis development and synapse maturation and reorganization in the prefrontal cortex (PFC). Thus, stress during adolescence could alter stress-sensitive brain regions such as the PFC and may alter the impact of future stressors on these brain regions. Given that women are more susceptible to many stress-linked psychological disorders in which dysfunction of PFC is implicated, and that this increased vulnerability emerges in adolescence, stress during this time could have sex-dependent effects. Therefore, we investigated the effects of adolescent social instability stress (SIS) on dendritic morphology of Golgi-stained pyramidal cells in the medial PFC (mPFC) of adult male and female rats. We then examined dendritic reorganization following chronic restraint stress (CRS) with and without a rest period in adult rats that had been stressed in adolescence. Adolescent SIS conferred long-term alterations in PL of males and females, whereby females show reduced apical length and basilar thin spine density and males show reduced basilar length. CRS in adulthood failed to produce immediate dendritic remodeling in SIS rats. However, CRS followed by a rest period reduced apical dendritic length and increases mushroom spine density in adolescently-stressed adult males. Conversely, CRS followed by rest produced apical outgrowth and decreased mushroom spine density in adolescently-stressed adult females. These results suggest that stress during adolescence alters development of the prefrontal cortex and modulates stress-induced dendritic changes in adulthood.

INTRODUCTION

Adolescence, a developmental period generally considered to begin near the onset of puberty and end after sexual maturation, is an important period for the development of many forebrain regions that are sensitive to stress (Lupien, McEwen, Gunnar, & Heim, 2009) and are involved in regulating stress responses (Goldman, Winget, Hollingshead, & Levine, 1973). One such brain region is the prefrontal cortex (PFC). For instance, adolescence is a crucial time for synaptic pruning and synaptic stabilization in the PFC of both humans and rodents (Panagiotakopoulos & Neigh, 2014; Selemon, 2013). Both male and female rats exhibit decreases in the density of dendritic spines on apical and basilar dendrites in the medial PFC (mPFC), and this is accompanied by decreased basilar branch length in females (Koss, Belden, Hristov, & Juraska, 2014). Densities of dopaminergic D1-family receptors peak in mid-adolescence, before falling dramatically in late adolescence and early adulthood (Andersen, Thompson, Rutstein, Hostetter, & Teicher, 2000). In addition, numbers of perineuronal nets surrounding parvalbumin-containing neurons—which play an important role in synaptic stabilization—reach adult levels during mid-adolescence (Baker, Gray, & Richardson, 2017). Thus, stress during adolescence could influence the maturation of the PFC, producing lasting consequences not only for PFC-mediated behaviors, but also for the ability of PFC to modulate the HPA axis.

Indeed, in rodents, chronic stress during adolescence has both immediate and lasting effects on corticolimbic brain regions and related behaviors. For instance, chronic variable physical stress during adolescence attenuates hippocampal growth and produces deficits on the Morris water maze in adulthood in males (Isgor, Kabbaj, Akil, & Watson, 2004). Daily restraint stress in early adolescence leads to immediate reductions in dendritic length in the mPFC and hippocampus of males and females and increases in depressive-like behaviors (Eiland, Ramroop, Hill, Manley, & McEwen, 2012). Male rats subjected to social instability in adolescence show reduced social interactions both immediately following the cessation of stress (Hodges et al., 2017) and in adulthood (Green, Barnes, & McCormick, 2013). The same manipulation results in increased ethanol intake in adolescent males (Marcolin, Hodges, Baumbach, & McCormick, 2019), and remodels dendrites in the medial amygdala (Hodges, Louth, Bailey, & McCormick, 2019).

Many of the immediate and remote effects of adolescent stress are sex-dependent. For instance, adolescent social instability increases depression-like behaviors in females but not males immediately post-stress (Mathews, Wilton, Styles, & McCormick, 2008). On the other hand, daily restraint stress in early adolescence impairs temporal order memory in males but not females immediately post-stress (Wei et al., 2014). Adult female rats that were exposed to social instability during adolescence demonstrate increased locomotor activation to nicotine and amphetamine compared to adult males (McCormick, Robarts, Gleason, & Kelsey, 2004; McCormick, Robarts, Kopeikina, & Kelsey, 2005). Chronic variable stress in adolescence results in depressive-like behavior (Bourke & Neigh, 2011; Wulsin, Wick-Carlson, Packard, Morano, & Herman, 2016) and HPA axis hyporeactivity in adult females (Wulsin et al., 2016), whereas males demonstrate HPA axis hyperreactivity (Jankord et al., 2011; Mathews et al., 2008). Together, these studies suggest that adolescent stress alters the development of stress-sensitive brain regions such as the PFC, and that this effect is different in females and males.

Consistent with this notion, stress in adolescence increases risk for many psychological disorders later in adulthood (Lewinsohn, Rohde, Seeley, Klein, & Gotlib, 2000). Notably, women are more likely than men to develop many of these disorders—for instance, depression—as adults (Cover, Maeng, Lebrón-Milad, & Milad, 2014), and this increased risk begins in adolescence—that is, shortly after puberty (Kessler, 2003).

Dysfunction of the PFC has been implicated in the pathophysiology of stress-linked disorders (e.g., Merriam, Thase, Haas, Keshavan, & Sweeney, 1999; Milad et al., 2009; Rajkowska, O’Dwyer, Teleki, Stockmeier, & Miguel-Hidalgo, 2006). PFC modulates several processes, including executive function and emotion regulation, that are impaired in many stress-sensitive disorders (Wellman et al., 2018). In addition, the PFC, along with other corticolimbic brain regions, helps to coordinate the body’s physiological response to stress through feedback to the HPA axis (Diorio, Viau, & Meaney, 1993; McKlveen, Myers, & Herman, 2015).

Sex-specific effects of chronic stress on PFC in adulthood are well documented. For example, chronic stress results in apical dendritic retraction in the mPFC in male rats (e.g., Bloss, Janssen, McEwen, & Morrison, 2010; Cook & Wellman, 2004; Garrett & Wellman, 2009; Moench & Wellman, 2017), but either little change (Moench & Wellman, 2017) or dendritic outgrowth (Garrett & Wellman, 2009) in females. Further, male rats have chronic stress-induced deficits in many mPFC-mediated behaviors such as fear extinction and cognitive flexibility (reviewed in Hurtubise & Howland, 2017; Wellman & Moench, 2019), whereas these deficits are typically absent in chronically stressed females (Baran, Armstrong, Niren, & Conrad, 2010; Snyder, Barry, Plona, et al., 2015). Finally, there are also sex-specific changes in dendritic architecture in the days following the cessation of chronic stress in adults. After a 7-day undisturbed interval after stress, chronically stressed male rats show dendritic outgrowth relative to unstressed rats. This difference from baseline is ameliorated 3 days later, 10 days post-stress. In contrast, females demonstrated minimal morphological changes in the post-stress period (Moench & Wellman, 2017). Dendritic remodeling is also accompanied by sex-dependent changes in morphology of dendritic spines (Moench & Wellman, 2017).

Notably, it is not known whether adolescent stress influences adult dendritic morphology in the mPFC, or whether stress-induced dendritic changes are different in adults with prior stress exposure during adolescence. Therefore, we first investigated the effects of adolescent social instability stress on adult dendritic morphology in the prelimbic subregion (PL) of the mPFC. Subsequently, we investigated chronic stress-induced dendritic remodeling in adults that had experienced social instability stress as adolescents. Previous work from our lab has demonstrated dynamic, sex-dependent dendritic change in the days following the cessation of a chronic stressor, with males showing dendritic outgrowth beyond baseline at 7 days post-stress (Moench & Wellman, 2017), we also characterized dendritric morphology at 7 days post-stress. For simplicity and consistency with previous studies (Moench, Breach, & Wellman, 2019; Ortiz & Conrad, 2018), we will refer to this undisturbed interval as “rest” hereafter.

METHODS

Experiment 1: Effects of adolescent social instability on adult prelimbic dendritic morphology.

Subjects and Stressor.

Male and female Sprague-Dawley rats were obtained from the supplier (Envigo, Indianapolis, IN) on postnatal day (PND) 22. Rats were housed in same-sex groups of three in standard cages (48 cm × 20 cm × 26 cm), with ambient temperature 23–25 °C, free access to food and water, and a 12:12 h light/dark cycle (lights on at 0800 h).

Rats were either unstressed (No SIS; n=6 male and 6 female) or experienced social instability stress (SIS; n=6 male and 6 female) from PND 30–45 using a modification of McCormick’s paradigm (Fig. 1A; see McCormick, Merrick, Secen, & Helmreich, 2007; Schmidt et al., 2007). This time period spans a prepubertal period and includes the average time of onset of puberty for females and males (McCormick, 2010). Unstressed rats were left undisturbed except for standard vivarium care. SIS rats were isolated 1 h per day in an opaque container with a radius of 6.5cm and height of 9.5cm. In addition, every four days during this period rats were placed with new cagemates of the same sex, such that rats were never housed with the same cagemates more than once. This paradigm is ecologically relevant, as social interactions are particularly rewarding for rats during adolescence, and social interactions during adolescence mitigate stress effects (McCormick, 2010). Thus, disrupting these interactions is likely stressful. Indeed, a similar social instability paradigm has been shown to produce anxiety-like behaviors in adult males and diestrus females and changes in adult HPA axis activity (McCormick, 2010). After PND 45, rats were left unhandled except for standard vivarium care until PND 70. All rats were weighed daily from PND 70–79 to assess effects of adolescent stress on adult weight gain. The stress manipulation occurred during the light phase of the light/dark cycle. All procedures were approved by the Bloomington Institutional Animal Care and Use Committee and conducted in accordance with NIH Guidelines.

Figure 1.

(A) Schematic diagram of timeline for Experiment 1. Male and female rats were either unstressed in adolescence (No SIS) or stressed in adolescence using a social instability paradigm (SIS) and dendritic morphology in mPFC was examined in adulthood. Arrows indicate time of euthanasia. N = 6 male and 6 female per stress condition. (B) Digital micrograph of a Golgi-stained pyramidal neuron in layer II-III of PL. Scale bar = 50μm. (C) Schematic representation of coronal sections of the rat mPFC. Shaded areas represent regions of prelimbic cortex (PL) from which neurons were reconstructed. (D) Digital micrograph of dendritic spines on a pyramidal neuron in layer II-III of PL. M, mushroom; S, stubby; T, thin. Scale bar = 5μm.

Estrous Phase Characterization.

Prior to perfusion on PND 79, vaginal lavages were performed and exfoliate cytology was examined immediately under a light microscope to determine estrous phase (Garrett & Wellman, 2009). The majority of females were in diestrus (n=9; proestrus n=1; estrus n=2). Thus, we did not analyze data relative to estrus phase.

Golgi Histology.

On PND 79, rats were overdosed with urethane and transcardially perfused with 0.9% saline before staining using a variation of the Golgi-Cox procedure (Wellman, 2017). Briefly, brains were removed and placed in Golgi solution for 14 days, followed by sucrose for 3–5 days, prior to being sectioned into 220 μm coronal sections on a vibratome (Campden Instruments 5100 mz). Sections were mounted, alkalinized, developed in Dektol (Kodak), fixed in Ilford rapid fixer, dehydrated in a graded series of ethanols, cleared in xylenes, and coverslipped.

Dendritic Analysis.

For each animal, 12 pyramidal neurons from layer II-III of PL (six per hemisphere) were randomly selected and reconstructed at a final magnification of 900x (Fig. 1B, 1C). Neurons were selected based on discriminability, the presence of a non-truncated apical dendrite extending from the apex of the soma towards the pial surface, the presence of dendritic spines, and at least one non-truncated basilar tree. If multiple non-truncated basilar trees were present, the experimenter reconstructed all those that were clearly discriminable. Morphology of apical and basilar dendrites was quantified in three dimensions using a computer-based neuron tracing system (Neurolucida, MBF Biosciences). For basilar dendrites per-tree analyses were performed.

Spine Density Analysis.

Dendritic spines on the distal branches of 12 apical and 12 basilar dendrites per animal were assessed at a final magnification of 1500x and classified into categories corresponding to spine shape (stubby, thin, and mushroom; Fig. 1D). Distal terminal branches were examined as they may be especially responsive to stress manipulations and corticosterone administration (Cook & Wellman, 2004; Wellman, 2001). Whenever possible, we assessed spine density on branches of previously reconstructed neurons.

Statistical analyses.

Weight gain from PND 70–79 and adrenal-to-body-weight ratios were analyzed using two-way ANOVAs. Several aspects of dendritic morphology were examined. For all analyses, values were averaged across neurons within subjects. To assess overall changes in dendritic morphology, mean number and length of apical and basilar dendrites were analyzed using two-way ANOVAs followed by appropriate planned comparisons. To analyze distribution of dendritic material we conducted three-dimensional Sholl analyses. These data were analyzed using a three-way repeated-measures ANOVA (sex × stress × distance from soma). Significant effects were followed up with relevant two-way ANOVAs (e.g., within males only, stress × distance from soma). Overall spine density and the density of each spine type were quantified and compared using two-way ANOVAs (sex × stress), followed by appropriate planned comparisons.

For analyses of overall dendritic morphology and spine densities, planned comparisons consisted of two-group F-tests done within the context of the overall ANOVA, and compared unstressed male and female rats, as well as differences between stress conditions within each sex.

Experiment 2: Effects of chronic stress in adulthood on dendritic morphology in PL in rats that experienced adolescent social instability.

We next examined sex-dependent dendritic responses to chronic restraint stress in adulthood in rats that had experienced SIS in adolescence. Because CRS-induced dendritic retraction in males is well-documented across several labs (reviewed in Wellman & Moench, 2019), we did not include No-SIS groups in this experiment.

Subjects and Stressor.

Male and female Sprague-Dawley rats were obtained and subjected to the SIS paradigm described in Experiment 1. After SIS, they were left undisturbed until PND 70. On PND 70, rats were randomly assigned to three groups (see Fig. 2 for experimental design; n=6/group/sex): unstressed in adulthood (SIS/No CRS; this group comprised the SIS group from Experiment 1), stressed in adulthood and perfused on the day of the last stressor (SIS/CRS), and stressed in adulthood and perfused after a 7 day post-stress rest period (SIS/CRS-Rest).

Figure 2.

Schematic diagram of timeline for Experiment 2. All rats underwent adolescent social instability stress (SIS). Male and female rats were then assigned to one of three groups: unstressed in adulthood (SIS/No CRS), stressed in adulthood (SIS/CRS), or stressed in adulthood followed by a 7-day period during which rats were undisturbed (SIS/CRS-Rest). Arrows indicate time of euthanasia. N = 6 rats per sex per adult stress condition.

Stress in adulthood consisted of chronic restraint stress (3 h/day for 10d) in which rats were placed in a semi-cylindrical Plexiglas tube (male: 16 cm × 6.5 cm × 5 cm; female: 15 cm × 6 cm × 4.5 cm, modified so the tail piece locks into place; Braintree Scientific, Braintree, MA). This procedure increases adrenal weights and attenuates weight gain in males and females (Bollinger, Collins, Patel, & Wellman, 2017; Moench et al., 2019).

Tissue Collection and Analyses.

On the last day of stress or rest, vaginal lavages were performed as described for Experiment 1. The majority of females were in diestrus (n=13) compared to proestrus (n=2) and estrus (n=3). Thus, we did not analyze our data relative to estrous phase. Rats were euthanized and brains were processed for Golgi histology, and dendritic and spine analyses and statistics were conducted as described in Experiment 1. To verify the stress manipulation, weight gain from PND 70–79 and adrenal-to-body-weight ratios were analyzed using two-way ANOVAs.

RESULTS

Experiment 1: Effects of adolescent social instability on adult prelimbic dendritic morphology.

Adolescent stress effects on adult body weight gain and adrenal to body weight ratio.

Overall, females gained less weight compared to males (Fig. 3A; main effect of sex, F_{1, 20} = 85.03, p < 0.001) and SIS reduced adult weight gain (main effect of stress, F_{1, 20} = 5.07, p = 0.04). There was not a significant sex by stress interaction on adult weight gain (F_{1, 20} = 2.75, p = 0.11). However, planned comparisons revealed SIS reduced adult weight gain in males (F_{1, 10} = 6.11, p = 0.03) and not females (F_{1, 10} = 0.24, p = 0.64). Both No SIS (F_{1, 10} = 86.78, p < 0.001) and SIS (F_{1, 10} = 21.70, p = 0.001) females gained less weight compared to their male counterparts.

Figure 3.

(A) Mean weight change in adult rats that were either unstressed (No SIS) or stressed using a social instability paradigm in adolescence (SIS). Social instability reduced adult weight gain in males but not females. (B) Mean adrenal-to-body-weight ratios. Females had significantly higher adrenal weights compared to males. SIS did not alter adult adrenal-to-body-weight ratios. Error bars represent SEM. Comparisons with significant p values are indicated.

Relative adrenal weights were higher in females than males (Fig. 3B; main effect of sex, F_{1, 20} = 120.03, p < 0.001) and SIS did not affect adrenal weights (main effect of stress, F_{1, 20} = 0.03, p = 0.86). Moreover, females had higher adrenal weights compared to males regardless of stress condition (sex by stress interaction, F_{1, 20} = 0.03, p = 0.86).

Adolescent social instability produces sex-dependent dendritic change in apical and basilar dendrites in PL.

Adolescent SIS significantly altered apical branch number in a sex-dependent manner (Fig. 4A, B; main effect of sex, F_{1, 20} = 3.61, p = 0.07; main effect of stress, F_{1, 20} = 2.52, p = 0.13; sex × stress interaction, F_{1, 20} = 6.31, p = 0.02). Planned comparisons demonstrated that apical branch number did not significantly differ between No-SIS males and females (F _{1, 10} = 0.33, p = 0.58). However, branch number was significantly reduced in SIS females relative to SIS males (F _{1, 10} = 6.78, p = 0.03), due to an SIS-induced decrease in apical branch number in females but not in males (F _{1, 10} = 5.77, p = 0.04, and 0.79, p = 0.40, respectively).

Figure 4.

(A) Reconstructions of the apical tree of representative neurons for adult male and female rats that underwent either no stress during adolescence (No SIS) or social instability stress (SIS) during adolescence. For each, mean apical length is near the mean for its group. Scale bar = 50μm. (B and C) Mean apical branch number and length in No SIS and SIS males and females. SIS reduced apical dendritic branch number and length in females but not males. (D) Mean apical length between concentric spheres of 20- μm increments. Dendritic retraction in females is present overall but most pronounced in the proximal-middle portion of the arbor. Error bars represent SEM. Comparisons with significant p values are indicated.

Adolescent SIS also significantly altered apical branch length in a sex-dependent manner (Fig. 4 A, C; main effect of sex, F_{1, 20} = 16.09, p = 0.001; main effect of stress, F_{1, 20} = 8.61, p = 0.01; sex × stress interaction, F_{1, 20} = 3.72, p = 0.07). Planned comparisons revealed that apical branch length did not differ significantly between No-SIS males and females (F _{1, 10} = 2.84, p = 0.12). However, apical branch length was significantly reduced in SIS females relative to males (F _{1, 10} = 14.29, p = 0.004). This was due to an SIS-induced decrease in branch length in females (F _{1, 10} = 11.69, p = 0.007) but not males (F _{1, 10} = 0.51, p = 0.49).

To more closely examine stress-induced changes in the distribution of apical dendritic material, Sholl analyses were performed. Unsurprisingly, apical dendritic length varied across the arbor (Fig. 4D; main effect of distance, F_{15, 300} = 204.49, p < 0.001). In addition, the distribution of dendritic length varied with both sex and stress (main effect of sex, F_{1, 20} = 17.68, p < 0.001; main effect of stress, F_{1, 20} = 7.99, p = 0.01; sex × distance interaction, F_{15, 300} = 2.48, p = 0.002; stress × distance interaction, F_{15, 300} = 2.19, p = 0.007). The interactions of sex and stress (F_{1, 20} = 3.25, p = 0.09) and of sex, stress and distance (F_{15, 300} = 0.42, p = 0.97) were not significant. A follow-up 2-way ANOVA (sex × distance from soma) failed to find a significant difference in apical dendritic distribution in No Stress males and females (main effect of sex, F_{1, 10} = 3.91, p = 0.08; sex × distance interaction, F_{15, 150} = 1.38, p = 0.17). On the other hand, a significant difference was found between SIS males and females (main effect of sex, F_{1, 10} = 14.29, p = 0.004; sex × distance interaction, F_{15, 150} = 1.55, p = 0.09). This was due to an SIS-induced apical dendritic retraction in females (main effect of stress, F_{1, 10} = 11.03, p = 0.008; stress × distance interaction, F_{15, 150} = 1.00, p = 0.46). This significant effect was not present in males (main effect of stress, F_{1, 10} = 0.51, p = 0.49; stress × distance interaction, F_{15, 150} = 1.55, p = 0.09).

For basilar dendrites, overall males and females had similar branch numbers (main effect of sex, F_{1, 20} = 1.51, p = 0.23) and SIS reduced branch number (Fig. 5A, B; main effect of stress, F_{1, 20} = 4.61, p = 0.04). Additionally, there was no significant interaction of stress and sex (F_{1, 20} = 0.37, p = 0.36). However, planned comparisons revealed that whereas there were no sex differences within either No-SIS (F_{1, 10} = 0.07, p = 0.81) or SIS groups (F_{1, 10} = 1.76, p = 0.21), SIS males had significantly reduced basilar branch numbers relative to No SIS males (F_{1, 10} = 4.96, p = 0.05). Similarly, SIS altered basilar branch length (Fig. 5 A, C; main effect of sex, F_{1, 20} = 0.00, p = 0.99; main effect of stress, F_{1, 20} = 15.66, p = .001; sex × stress interaction, F_{1, 20} = 3.13, p = 0.09), and planned comparison indicated that this effect was sex-specific. Males and females did not differ significantly within either No SIS or SIS groups (F_{1, 10} = 1.32 and 1.92, respectively, p = 0.28 and 0.21, respectively). Stress significantly decreased basilar branch length in males (F_{1, 10} = 18.56, p = 0.002) but not females (F_{1, 10} = 2.14, p = 0.17).

Figure 5.

(A) Reconstructions of a basilar tree of representative neurons for adult male and female rats that underwent either no stress during adolescence (No SIS) or social instability stress (SIS) during adolescence. For each, mean apical length is near the mean for its group. Scale bar = 50μm. (B, C) Mean basilar branch number and length in No SIS and SIS males and females. SIS significantly reduced adult basilar dendritic length in males but not females. (D) Mean basilar length between concentric spheres of 20- μm increments. Males demonstrated retraction throughout the basilar arbor. Error bars represent SEM. Comparisons with significant p values are indicated.

Three-way repeated measures ANOVA (sex × stress × distance from soma) revealed that length of basilar dendritic material varied with distance from soma (Fig. 5C; main effect of distance, F_{8, 160} = 395.41, p < 0.001), and that stress reduced basilar branch length across the arbor (main effect of stress, F_{1, 20} = 9.71, p = 0.005; distance × stress interaction, F_{8, 160} = 2.95, p = 0.004). Overall, basilar distribution was not significantly different in males and females (main effect of sex, F_{1, 20} = 0.87, p = 0.36), though this varied across the arbor (interaction of distance from soma and sex, F_{8, 160} = 2.09, p = 0.04). There was no sex by stress interaction (F_{1, 20} = 1.00, p = 0.33), and no interaction of distance, sex, and stress (F_{8, 160} = 1.26, p = 0.27). However, follow-up analyses demonstrated that the effect of SIS on basilar dendritic organization was specific to males. Within the No SIS groups, 2-way repeated measures ANOVA (sex × distance from soma) revealed no overall sex difference (F_{1, 10} = 1.31, p = 0.28), but an effect of sex that varied across the arbor (sex × distance interaction, F_{8, 80} = 2.15, p = 0.04). This reflected subtly shorter basilar dendritic length in No SIS females relative to males, which failed to reach significance at any distance from the soma (all pairwise Fs_{1, 10} < 1.86, 0.20 < p’s < 0.92) except 100 μm (F_{1, 10} = 4.75, p = 0.05). This small sex difference was not present in SIS rats (main effect of sex, F_{1, 10} = 0.00, p = 0.95; sex × distance interaction, F_{8, 80} = 0.92, p = 0.50). Instead, SIS reduced basilar dendritic lengths in males (main effect of stress, F_{1, 10} = 9.82, p = 0.01; stress × distance interaction, F_{8, 80} = 3.79, p = 0.001) but not females (main effect of stress, F_{1, 10} = 1.97, p = 0.19; stress × distance interaction, F_{8, 80} = 0.43, p = 0.90).

Adolescent social instability reduces adult thin spine density on basilar dendrites in PL of females.

Two-way ANOVAS for apical spines revealed that males and females had similar spine densities both overall and across all spine types (Fig. 6A; main effect of sex, Fs_{1, 20} ≤ 1.24, 0.28 < p’s < 0.99). Stress did not alter density either overall or for any spine type (main effect of stress, Fs_{1, 20} ≤ 1.90, 0.18 < p’s < 0.98), and this did not vary by sex (sex × stress interaction, Fs_{1, 20} ≤ 2.95, 0.10 < p’s < 0.89).

Figure 6.

(A) Mean spine densities on apical terminal dendrites of adult male and female rats that underwent either no stress during adolescence (No SIS) or social instability stress (SIS) during adolescence. No significant differences due to stress or sex were found. (B) Mean spine densities on basilar terminal dendrites of adult male and female rats that underwent either no stress during adolescence (No SIS) or social instability stress (SIS) during adolescence. Social instability reduced thin spine density in females but not males. Error bars represent SEM. Comparisons with significant p values are indicated.

Males and females had similar total basilar spine densities (Fig. 6B; main effect of sex, F_{1, 10} = 2.62, p = 0.12), and this did not vary by stress condition (interaction between sex and stress, F_{1, 10} = 1.16, p = 0.29). However, SIS reduced thin spine density (main effect of stress, F_{1, 20} = 9.60, p = 0.006). Planned comparisons revealed this was primarily due to a decrease in thin spine density in the female SIS group relative to the No SIS females (F_{1, 10} = 6.57, p = 0.03), as a similar tendency in males failed to reach significance (F_{1, 10} = 3.08, p = 0.11). Stress did not significantly alter density of stubby or mushroom spines (main effect of stress, Fs_{1, 20} ≤ 2.58, 0.12 < p’s < 0.43), and this did not vary by sex (sex × stress interaction, Fs_{1, 20} ≤ 2.82, 0.11 < p’s < 0.16).

Experiment 2: Chronic stress-induced changes in adult PL in adolescently-stressed rats.

Chronic stress in adulthood attenuates weight gain and increases adrenal weight in adolescently-stressed rats.

Overall, females gained less weight compared to males (Fig. 7A; main effect of sex, F_{1, 30} = 5.18, p = 0.03), and CRS attenuated weight gain (main effect of stress, F_{2, 30} = 38.74, p < 0.001; sex × stress interaction, F_{2, 30} = 2.92, p = 0.07). Planned comparisons showed that SIS/No CRS males gained significantly more weight than SIS/No CRS females (F_{1, 10} = 21.70, p = 0.001). CRS reduced weight gain in males and females. SIS/CRS males (F_1,10 = 40.12, p < 0.001) and SIS/CRS-Rest males (F_1,10 = 36.42, p < 0.001) gained less weight compared to SIS/No CRS males, though they did not differ from each other (F_1,10 = 1.39, p = 0.27). Likewise, SIS/CRS females (F_1,10 = 23.54, p = 0.001) and SIS/CRS-Rest females (F_1,10 = 37.81, p < 0.001) gained less weight compared to SIS/No CRS females, and they also did not differ from each other (F_1,10 = 0.16, p = 0.70).

Figure 7.

Mean weight change (A) and adrenal-to-body-weight ratios (B) in male and female rats that underwent adolescent social instability stress (SIS) and were then either unstressed in adulthood (SIS/No CRS), stressed in adulthood (SIS/CRS), or stressed via chronic restraint in adulthood followed by a 7-day rest period (SIS/CRS-Rest). Adult CRS resulted in persistent attenuation of weight gain for both males and females, and a lasting increase in adrenal weight ratio for males. In females, adrenal weight ratio was initially increased after stress and then significantly decreased after seven days of rest. Error bars represent SEM. Comparisons with significant p values are indicated.

Females had higher relative adrenal weights compared to males (Fig. 7B; main effect of sex, F_{1, 30} = 198.11, p < 0.001), and CRS increased relative adrenal weights (main effect of stress, F_{2, 30} = 6.63, p = 0.004). Planned comparisons found that SIS/No CRS females had higher relative adrenal weights compared to SIS/No CRS males (F_{1, 10} = 56.53, p < 0.001). CRS increased adrenal to body weight ratios, as SIS/CRS males (F_1,10 = 16.00, p = 0.003) and SIS/CRS-Rest males (F_1,10 = 6.43, p = 0.03) had higher adrenal-to-body-weight ratios compared to SIS/No CRS males, though they did not differ from each other (F_1,10 = 0.63, p = 0.45). In females, CRS significantly increased adrenal weight relative to SIS/No CRS rats (F_1,10 = 5.55, p = 0.04). SIS/CRS-Rest females had significantly lower adrenal weight ratios relative to SIS/CRS females (F_{1, 10} = 5.44, p = 0.04) and were not different from SIS/No CRS females (F_{1, 10} = 0.04, p = 0.85).

Chronic stress induces sex-dependent changes in apical dendritic morphology in PL of adolescently-stressed rats following a rest period.

Two-way ANOVA revealed that adult CRS altered apical branch number in a sex-dependent manner (Fig. 8A, B; main effect of sex, F_{1, 30} = 4.36, p = 0.05; main effect of stress, F_{2, 30} = 0.91, p = 0.41; sex × stress interaction, F_{2, 30} = 8.07, p = 0.002). As described in Experiment 1, branch number was significantly reduced in SIS/No CRS females relative to SIS/No CRS males. A similar sex difference was present in SIS/CRS rats (F_{1, 10} = 10.70, p = .008), but not SIS/CRS-Rest rats (F_{1, 10} = 4.11, p = 0.07). This pattern was due to a differential effect of adult CRS between the sexes. Although CRS did not significantly alter apical branch number immediately post-stress in either males or females (F_{1, 10} = 1.06 and 0.71, respectively, p = 0.33 and 0.41, respectively), both males and females showed significant apical remodeling 7 days after the cessation of CRS. SIS/CRS-Rest males had significantly decreased apical branch numbers relative to SIS/No CRS (F_{1, 10} = 15.12, p = .003); this decrease approached significance relative to SIS/CRS rats (F_{1, 10} = 4.60, p = .06). In contrast, SIS/CRS-Rest females demonstrated significantly increased apical branch number relative to SIS/CRS females (F_{1, 10} = 9.91, p = .01) but not SIS/No CRS females (F_{1, 10} = 2.00, p = 0.19).

Figure 8.

(A) Reconstructions of apical trees of representative neurons for male and female rats that underwent adolescent social instability stress (SIS) and were then either unstressed in adulthood (SIS/No CRS), stressed in adulthood (SIS/CRS), or stressed via chronic restraint in adulthood followed by a 7-day rest period (SIS/CRS-Rest). Scale bar = 50 μm. For each, mean apical length is near the mean for its group. (B and C) Mean apical branch number and length in male and female SIS/No CRS, SIS/CRS, and SIS/CRS-Rest rats. SIS/CRS-Rest males had significantly lower apical branch number and length compared to SIS/No CRS males. SIS/CRS-Rest females had significantly longer apical branch length compared to SIS/CRS females. (D) Mean distribution of dendritic material. The dendritic retraction and outgrowth seen after rest in males and females, respectively, is most pronounced in the middle-distal portion of the arbor. Error bars represent SEM. Comparisons with p values approaching or reaching significance are indicated.

Similarly, 2-way ANOVA revealed that females had lower mean apical branch length compared to males (Fig. 8 A, C; main effect of sex, F_1,30 = 7.06, p = 0.01). Overall, CRS did not alter mean apical length (main effect of stress, F_2,30 = 0.91, p = 0.59). However, the effect of stress on apical branch length varied with sex (sex × stress interaction, F_2,30 = 7.33, p = 0.003). As described in Experiment 1, SIS/No CRS males had longer apical dendrites than SIS/No CRS females. Likewise, planned comparisons revealed that SIS/CRS males had longer apical dendrites than SIS/CRS females (F_{1, 10} = 5.49, p = 0.04). Although SIS/CRS males did not differ significantly from SIS/No CRS males, (F_{1, 10} = 1.56, p = 0.24), SIS/CRS-Rest males had significant apical dendritic retraction relative to SIS/No CRS males (F_1,10 = 11.44, p = 0.007). SIS/CRS and SIS/CRS-Rest males did not significantly differ (F_{1, 10} = 2.31, p = 0.16). SIS/CRS females did not differ significantly from SIS/No CRS females (F_{1, 10} = 0.00, p = 0.99). However, SIS/CRS-Rest females tended to have increased dendritic length compared to SIS/No CRS females (F_{1, 10} = 3.83, p = 0.08), and apical dendrites of SIS/CRS-Rest females were significantly longer than those of SIS/CRS females (F_{1, 10} = 6.37, p = 0.03).

To more closely examine stress-induced changes in the distribution of apical dendritic material, Sholl analyses were performed. Three-way repeated measures ANOVA (sex × stress × distance from soma) revealed that the distribution of dendritic material varied across the apical arbor (Fig. 8D; main effect of distance F_{15, 450} = 324.10, p < 0.001). Overall, females had shorter branch length compared to males (main effect of sex, F_{1, 30} = 6.77, p = 0.01). Overall, CRS did not influence apical distribution (main effect of stress, F_2,30 = 0.50, p = 0.61); however, the effect of CRS on apical dendritic distribution varied by sex (sex × stress interaction, F_{2, 30} = 7.41, p = 0.002), and this effect was consistent across the arbor (distance × sex interaction, F_{15, 450} = 0.78, p = 0.70; distance × stress interaction, F_{30, 450} = 0.73, p = 0.86; sex × stress × distance interaction, F_{30, 450} = 0.90, p = 0.63).

Follow-up 2-way repeated-measures ANOVAs (stress × distance from soma) within males revealed that for adolescently-stressed males, SIS/No CRS and SIS/CRS males did not significantly differ at any point along the arbor (main effect of stress, F_1,10 = 1.55, p = 0.24; stress × distance interaction, F_15,150 = 0.35, p = 0.99). However, SIS/CRS-Rest males had significantly reduced apical branch length relative to SIS/No CRS males (main effect of stress, F_1,10 = 11.82, p = .006), and this effect was consistent across the arbor (stress × distance interaction, F_15,150 = 1.08, P = 0.38). SIS/CRS and SIS/CRS-Rest males did not differ significantly (main effect of stress, F_1,10 = 2.43, p = 0.15; stress × distance interaction, F_15,150 = 0.62, p = 0.86).

Likewise, for adolescently-stressed females, CRS alone did not affect apical branch length, as SIS/No CRS rats did not significantly differ from SIS/CRS rats at any point along the apical dendritic arbor (main effect of stress, F_1,10 = 0.04, p = 0.84; stress × distance interaction, F_15,150 = 0.90, p = 0.57). However, SIS/CRS-Rest females showed a tendency towards increased dendritic material relative to SIS/No CRS females (main effect of stress, F_1,10 = 3.81, p = .08; stress × distance interaction, F_15,150 = 1.30, p = 0.21), and this increase was significant relative to SIS/CRS females (main effect of stress, F_1,10 = 5.83, p = .04) and consistent across the apical arbor(stress × distance interaction, F_15,150 = 0.76, p = 0.72).

Overall, basilar branch number and length were similar for males and females (Fig. 9A, B; main effect of sex, F_{1, 30} = 3.38 and 2.80, respectively, p = 0.08 and 0.11, respectively) and CRS did not significantly alter branch number or length (main effect of stress, F_{2, 30} = 0.63 and 1.16, respectively, p = 0.54 and 0.33, respectively), and this did not vary by sex (sex × stress interaction, F_{2, 30} = 1.18 and 0.28, p = 0.32 and 0.76, respectively). Likewise, Sholl analysis revealed that males and females had similar dendritic distributions across the basilar arbor (Fig. 9C; main effect of sex, F_{1, 30} = 0.42, p = 0.52; distance from soma × sex interaction, F_{8, 240} = 0.54, p = 0.83), and adult CRS did not alter basilar distribution (main effect of stress, F_{2, 30} = 0.40, p = 0.67), and this did not vary across the arbor (distance × stress interaction, F_{16, 240} = 0.69, p = 0.80) or between males and females (sex × stress interaction, F_{2, 30} = 0.35, p = 0.71; sex × stress × distance, F_{16, 240} = 0.57, p = 0.90).

Figure 9.

(A and B) Mean branch number and length per basilar tree in male and female rats that underwent adolescent social instability stress (SIS) and were then either unstressed in adulthood (SIS/No CRS), stressed in adulthood (SIS/CRS), or stressed via chronic restraint in adulthood followed by a 7-day rest period (SIS/CRS-Rest). No significant differences were found. (C) Mean distribution of basilar dendritic material for male and female SIS/No CRS, SIS/CRS, and SIS/CRS-Rest rats. Error bars represent SEM.

Chronic stress followed by rest alters spine densities on apical dendrites in PL of adolescently-stressed rats.

Two-way ANOVAs revealed that neither stress nor sex significantly influenced total spine density on apical dendrites (Fig. 10A; main effect of sex, F_{1, 30} = 3.72, p = 0.06; main effects of stress, F_{2, 30} = 0.16, p = 0.86; sex × stress interaction, F_{2, 30} = 0.49, p = 0.62). However, densities of specific spine types varied with both sex and stress. Overall, stubby spine densities were significantly lower in females relative to males (main effect of sex, F_{1, 30} = 4.30, p = 0.05), but densities of thin and mushroom spines did not differ by sex (F_1,30 = 0.62 and 1.43, respectively, p = 0.44 and 0.24, respectively). Overall, stress did not significantly alter stubby, thin or mushroom spine density (main effect of stress, F_{2, 30} = 2.81, 1.18, and 0.99, respectively; 0.08 < p’s < 0.39), and this was consistent across the sexes for stubby and thin spines (sex × stress interaction, F_{2, 30} = 0.71 and 0.10, respectively, p = 0.71 and 0.91, respectively). However, the effect of stress on mushroom spine density was significantly different in males and females (sex × stress interaction, F_{2, 30} = 4.27, p = 0.02). Planned comparisons revealed that CRS alone did not alter mushroom spine density in adolescently-stressed males (F_{1, 10} = 1.98, p = 0.19) or females (F_{1, 10} = 0.01, p = 0.91). However, SIS/CRS-Rest male rats had an increase in mushroom spine density relative to SIS/No CRS males (F_{1, 10} = 7.51, p = 0.02), but not SIS/CRS males (F_{1, 10} = 0.15, p = 0.71). Further, in SIS/CRS-Rest females, mushroom spine density was significantly decreased compared to SIS/No CRS (F_{1, 10} = 7.44, p < 0.02) and SIS/CRS (F_{1, 10} = 5.21, p = 0.05) female rats.

Figure 10.

(A) Mean spine densities on apical terminal dendrites in male and female rats that underwent adolescent social instability stress (SIS) and were then either unstressed in adulthood (SIS/No CRS), stressed in adulthood (SIS/CRS), or stressed via chronic restraint in adulthood followed by a 7-day rest period (SIS/CRS-Rest). Females had lower stubby spine density compared to males. SIS/CRS-Rest males showed increased mushroom spine density on apical dendrites compared to SIS/No CRS males. Conversely, SIS/CRS-Rest females demonstrated decreases in mushroom spine density relative to SIS/No CRS and SIS/CRS females. (B) Mean spine densities on basilar terminal dendrites for male and female SIS/No CRS, SIS/CRS, and SIS/CRS-Rest rats. Females had lower total spine density compared to males. Error bars represent SEM. Comparisons with significant p values are indicated.

Chronic stress induces sex difference in total spine densities on basilar dendrites in PL of adolescently stressed rats.

Two-way ANOVAs revealed that stress did not significantly alter total spine density or the density of specific spine types (Fig. 10B, main effect of stress, F_{2, 30} = 0.22, p = 0.81 for total spines; and 0.65, 1.47, and 0.18, for stubby, thin, and mushroom spines, respectively, 0.25 < p’s < 0.84), and this did not vary with sex (sex-by-stress interaction, all Fs_{2, 30} ≤ 0.74, p = 0.49). However, overall, females had lower total spine density compared to males (main effect of sex, F_{1, 30} = 4.51, p = 0.05), though the tendencies towards lower density of stubby, thin, and mushroom spines in females failed to reach significance (main effect of sex, F_{1, 30} = 3.27, 1.44, and 1.01, respectively, 0.08 < p’s < 0.32).

DISCUSSION

Here we report that adolescent social instability reduces adult weight gain in males but not females, and alters adult dendritic morphology in PL in a sex-dependent manner, with females showing reduced apical dendritic length and basilar thin spine density and males showing reduced basilar length. The lasting effects of adolescent social instability may set the stage for unique responses to adult stress: We found that chronic stress in adulthood results in persistent increases in relative adrenal weights in males. Further, CRS followed by rest reduces apical dendritic length and increases mushroom spine density in adolescently-stressed adult males. Conversely, CRS followed by rest produces apical dendritic outgrowth and decreases mushroom spine density in adolescently-stressed adult females.

Changing cage mates disrupts social bonds and reduces habituation to daily isolation, thereby maintaining the stressful nature of the repeated isolation in this paradigm. Moreover, social instability in adolescence prolongs corticosterone release in response to a novel stressor, reduces weight gain in males, increases CRH mRNA expression in the paraventricular nucleus of the hypothalamus (PVN), and increases PVN activation in response to a new stressor in males (reviewed in McCormick, 2010). These data suggest that the social instability manipulation was stressful.

Male and female rats exhibit adolescent-typical behaviors during PND 28–42 (Spear, 2000). Although our manipulation encompassed the pre- and post- pubertal periods for both sexes (McCormick, 2010), males and females differ in the timing of noticeable pubertal onset (McCormick & Mathews, 2007). Thus, the timing of our manipulation relative to puberty differed slightly for males and females, and this could contribute to the sex differences we report.

Adolescent social instability alters adult body weight and adrenal effects of adult stress in males but not females

Male rats that underwent social instability in adolescence demonstrated significantly decreased weight gain in adulthood compared to their unstressed counterparts, suggesting that adolescent stress may alter long-term feeding behaviors or confer metabolic changes. There is evidence to support this notion. For example, chronic variable stress in male rats throughout adolescence results in increased adrenal size and decreased body weight and fat content in adulthood, and stress in late adolescence also resulted in decreased thymus size and increased HPA axis reactivity (Jankord et al., 2011). Consistent with our data, studies employing daily social isolation in adolescent and adult females have found minimal effect on weight gain (McCormick, 2010). On the other hand, chronic variable stress in late-adolescent females reduced body weight gain, increased thymus and adrenal weights, and blunted stress reactivity in adulthood (Wulsin et al., 2016). These differing results in females may be due to different age of stress onset, stressor type, or intensity of the stressor.

Consistent with previous studies in stress-naïve rats (Baran, Armstrong, Niren, Hanna, & Conrad, 2009; Bollinger et al., 2017; Garrett & Wellman, 2009; Moench et al., 2019), chronic stress attenuated weight gain and increased adrenal-to-body-weight ratios in male and female rats that were stressed in adolescence. However, unlike in our previous study (Moench & Wellman, 2017), in males the increased adrenal weight was maintained after 7 days of rest. As discussed above, chronic variable stress in adolescent females blunted stress reactivity in adulthood (Wulsin et al., 2016), but increased HPA axis reactivity in males (Jankord et al., 2011). Similar, subtle differences in stress reactivity induced by adolescent SIS may be responsible for the sex-dependent, lasting effect of adult stress on adrenal weights seen here. Note also that the rats used in the present study were obtained at PND 22, with shipping of approximately one hour duration from the vendor to the lab. This relatively brief experience may have been stressful, and we cannot rule out the possibility that the stress of shipping may have influenced the effects of subsequent stressors.

Adolescent social instability alters neuronal morphology in adult PL in a sex-dependent manner

In males, adolescent SIS reduced adult basilar dendritic length but did not significantly alter apical dendritic length, whereas in females, the opposite pattern of dendritic remodeling was present, with significant reductions in apical but not basilar length. These patterns of changes contrast with the immediate effects of chronic restraint in adolescence: comparable reductions in apical length in the PL of both males and females, with no significant remodeling of basilar dendrites (Eiland et al., 2012). This suggests that while both sexes show apical dendritic retraction immediately following adolescent stress, further remodeling occurs in males—but not females. This suggests that, as in adults (Moench et al., 2019; Moench & Wellman, 2017), adolescent males and females show differential patterns of prefrontal plasticity in the extended period after stress.

Interestingly, social instability in adolescence resulted in a basal sex difference in mean apical dendritic length of adult rats, whereby females had shorter apical dendrites, a sex difference that was not present in unstressed controls. Previous studies in unstressed adults are limited and mixed, with two studies demonstrating that adult male rats have longer, more complex apical dendritic arbors in medial prefrontal cortex compared to unstressed adult females (Garrett & Wellman, 2009; Markham & Juraska, 2002), and another failing to find this basal sex difference (Moench & Wellman, 2017). The current findings suggest a potential explanation for the mixed results in the literature: differences in early experience.

Likewise, in Experiment 2, adolescently-stressed female rats, regardless of adult stress condition, had significantly lower densities of stubby spines on apical dendrites and reductions in total spine densities on basilar dendrites (the contrasting lack of significant effects in Experiment 1 likely reflects the small effect size combined with the increased power when sex comparisons were collapsed across the adult-CRS groups). To our knowledge, no studies have demonstrated sex differences in adult spine density on basilar dendrites in the PL (but cf Markham & Juraska’s 2002 findings in the anterior cingulate cortex). Thus, as also suggested by the results of Experiment 1, adolescent stress may alter the development of basilar dendrites differently for males versus females, resulting in functional differences in local communication.

In SIS rats, chronic restraint stress in adulthood induces dendritic remodeling in PL only after a rest period

According to previous work, retraction of apical dendrites in PL of adult male rats is evident immediately after chronic stress (e.g., Cook & Wellman, 2004; Garrett & Wellman, 2009; Moench & Wellman, 2017; Radley et al., 2005) and then extend to a length greater than unstressed controls after a 7-day rest period (Moench & Wellman, 2017). Conversely, dendrites of females either extend immediately following stress (Garrett & Wellman, 2009) or demonstrate minimal change (Moench & Wellman, 2017). In the current study, we did not directly compare the effects of adult CRS in SIS rats to adult CRS in No-SIS rats. However, when taken in the context of the literature, it appears that adolescently stressed male and female rats demonstrate delayed retraction and outgrowth, respectively, in response to adult chronic stress. Similarly, we found a sex-dependent stress effect on spine densities, with adolescently stressed males experiencing an increase in mushroom spine density following rest from chronic stress, whereas females experienced a decrease following the rest period. Interestingly, many of the methods used here were nearly identical to methods used in Moench and Wellman (2017), which found stress-induced changes in stubby spines rather than mushroom spines. This suggests that prior adolescent stress alters CRS-induced changes in apical spine density in the PL. These data suggest that history of adolescent stress modulates the onset of chronic stress-induced neuronal remodeling. Given the nonlinear post-stress remodeling previously documented (Moench & Wellman, 2017), it is quite possible that dendritic architecture may change further with more time post-stress.

Prior studies have consistently demonstrated that spine density in adult mPFC is responsive to stress. Chronic restraint stress (6 h a day for 21 d) decreases total (Hains et al., 2009; Radley, Anderson, Hamilton, Alcock, & Romig-Martin, 2013) and mushroom spine density (Radley et al., 2013) on apical dendrites in mPFC of adult male rats. Conversely, our lab previously found decreases in stubby spine density in male rats immediately after 10 days of 3-h CRS and increases in stubby spine density in female rats after seven days of rest following CRS (Hains et al., 2009; Radley et al., 2013). Other chronic stressors appear to affect apical spine density and are associated with behavioral changes. Chronic social isolation decreases levels of synaptic proteins and reduces total spine density in males and females due to decreases in thin and mushroom spines on proximal portions of apical dendrites in PL (Sarkar & Kabbaj, 2016).

Implications of SIS-Induced Dendritic Remodeling for Physiology and Behavior

Although the utility of spine classification based on morphology has recently been questioned (Segal, 2017), spine morphology (e.g. stubby, thin, mushroom) is hypothesized to reflect spine maturity. Mushroom spines are thought to be mature, strong synapses while stubby and thin spines are thought to be unstable and reactive. For instance, AMPA receptors are more abundant in mushroom spines (Matsuzaki et al., 2001), and spine head size correlates with synaptic strength and stability (Chen, Lu, & Zuo, 2014). Additionally, spines are highly dynamic, with changes in spine morphology and number associating with synaptic strength (Matsuzaki et al., 2001) and PFC-dependent learning (Lai, Franke, & Gan, 2012; Moench, Maroun, Kavushansky, & Wellman, 2016). Thus, sex-dependent alterations in spine density and type could produce different alterations of neuronal function for males versus females.

Furthermore, given that dendritic geometry affects neuronal firing patterns (Grudt & Perl, 2002; Lu, Inokuchi, McLachlan, Li, & Higashi, 2001; Mainen & Sejnowksi, 1996), the sex-dependent dendritic reorganization after both adolescent and adult stress likely also has implications for the function of these neurons. For instance, apical and basilar dendrites in pyramidal neurons receive different inputs (Spruston, 2008). Given that basilar dendrites tend to remain within the same layer as the soma, the reduction in basilar length seen in adolescently-stressed adult males may reflect long-term alterations in local communication in the PL, whereas the reduction in apical arbor in females may primarily alter communication with more distal brain regions. Furthermore, the plasticity rules that determine the induction of LTP are different in the apical and basilar arbors of cortical neurons (Gordon, Polsky, & Schiller, 2006). Thus, differential remodeling of these two compartments in males and females could produce sex-dependent patterns of alterations in the function of PL neurons and the behaviors they mediate.

This is consistent with many studies demonstrating that adolescent stress produces changes in a variety of PFC-influenced behaviors in adulthood. For instance, male rats subjected to daily social defeat during early or mid- adolescence demonstrate cognitive deficits on a PFC-dependent operant set-shifting task in adulthood (Snyder, Barry, & Valentino, 2015). The same paradigm resulted in immediate deficits on the task in adolescent females, though unlike in male rats, these deficits did not persist into adulthood (Snyder, Barry, Plona, et al., 2015). Chronic social defeat in adolescent male mice produces delayed deficits on attentional set-shifting that do not manifest until adulthood (Xu et al., 2016). These deficits were accompanied by decreased levels of brain-derived neurotrophic factor (BDNF), which is important for neuronal growth and maintenance in the PFC. Finally, in females, adolescent chronic variable stress resulted in increased depression-like behavior in the forced swim task that did not manifest until adulthood (Wulsin et al., 2016). This “incubation effect” was unique to stress in adolescence: a similar stress protocol in adulthood did not produce long-term behavioral impairments (Cotella et al., 2019). Thus, it is possible that the adolescent stressor in the current study results in altered neuronal communication, and thus delayed morphological responses, which could contribute to long-term and delayed behavioral deficits.

CONCLUSION

Stress in adolescence results in lasting sex-dependent dendritic reorganization that is present in adulthood. These adolescent SIS-induced changes may contribute to the differential patterns of chronic stress-induced dendritic remodeling we saw in adulthood. In turn, such changes may contribute to differential vulnerability to stress-induced behavioral deficits, as several studies have demonstrated that adolescent stress profoundly affects behavior in rodent models of human stress-linked disorders. Understanding how adolescent stress affects stress-induced changes in patterns of prefrontal plasticity may shed light on stress in adolescence as a potential risk factor for the development of stress-linked psychological disorders and lay the groundwork for identification and targeting of mechanisms.