Review article
Adolescent development, hypothalamic-pituitary-adrenal function, and programming of adult learning and memory

https://doi.org/10.1016/j.pnpbp.2009.09.019Get rights and content

Abstract

Chronic exposure to stress is known to affect learning and memory in adults through the release of glucocorticoid hormones by the hypothalamic-pituitary-adrenal (HPA) axis. In adults, glucocorticoids alter synaptic structure and function in brain regions that express high levels of glucocorticoid receptors and that mediate goal-directed behaviour and learning and memory. In contrast to relatively transient effects of stress on cognitive function in adulthood, exposure to high levels of glucocorticoids in early life can produce enduring changes through substantial remodeling of the developing nervous system. Adolescence is another time of significant brain development and maturation of the HPA axis, thereby providing another opportunity for glucocorticoids to exert programming effects on neurocircuitry involved in learning and memory. These topics are reviewed, as is the emerging research evidence in rodent models highlighting that adolescence may be a period of increased vulnerability compared to adulthood in which exposure to high levels of glucocorticoids results in enduring changes in adult cognitive function.

Introduction

Animal models provide a valuable approach for the investigation of stress-related developmental pathology due to experimental controls that would not be possible in investigations of people. Although preclinical research has focused on earlier and later periods of development, there is increasing interest in investigating stress-related plasticity that may be unique to adolescence. In the same way that adolescence in humans is a period of development that, despite a gradual onset and offset, is qualitatively different than childhood and adulthood, the behavioural repertoire of adolescent rodents is substantially different from that of neonates and pre-weanlings and from that of adults. For example adult-typical social behaviour and risk-taking behaviours emerge in adolescence, although adolescents and adults differ in some of the expression of these behaviours (Sisk and Foster, 2004, Spear, 2000). Rough-and-tumble play begins at about postnatal day 18 and begins to increase thereafter, especially in males, until puberty, and declines thereafter (Meaney, 1989, Pellis et al., 1997). The behavioural differences across ontogeny are a manifestation of the restructuring and ongoing maturation of nervous system function over adolescence in rats, just as there is in people, with prefrontal cortical regions of the brain having significant ongoing development over adolescence (Crews et al., 2007, Giedd, 2004). The behavioural and neural changes are also necessary adaptations to the demands required of animals after weaning, such as exploration at greater distances from the nest and eventual dispersal, broader social interactions, and preparation for sexual and reproductive function. Even when adolescent and adult rats are behaviourally similar, such as the transient expression of parental behaviour found in early adolescence (reviewed in Nephew et al., 2008), the behaviours of adolescents often involve different neural and endocrine mechanisms than those of adults.

Biological transitions are times of increased vulnerability that may be related to change in the reactivity of the stress systems, notably the hypothalamic-pituitary-adrenal (HPA) axis (Dorn and Chrousos, 1997). In early prenatal and neonatal ontogeny, prolonged activation of the HPA axis by stressors influences the development of the central nervous system (CNS), thereby programming adult function. The transitions from neonate to adolescence to adulthood also involve significant changes in HPA function and ongoing maturation of the CNS. Thus, adolescence too may be an opportunity for further environmental tuning of CNS development and its learning and memory functions. Here, we review research in rodent models investigating the HPA axis and its role in learning and memory-associated neuroplasticity in adults to then address general adolescent development and HPA function, brain development and learning and memory during adolescence, and the extent to which the experience of chronic stressors in adolescence results in relatively permanent alteration of learning and memory function in adulthood.

The physiological response to stressors occurs in two temporally distinct waves. The immediate, ‘first wave’ response is initiated within seconds through activation of the sympathetic nervous system (SNS) and results in rapid accumulation of circulating catecholamines that prepare the body for survival. The second wave, or the endocrine response to stress involving the HPA axis, is slower and results in the release of glucocorticoids (primarily cortisol in people, corticosterone in rodents), which play a critical role in long-lasting adaptations to stressors. The SNS consists of sympatho-neural and sympatho-adrenomedullary components, which release catecholamines from sympathetic nerves and the adrenal medulla, respectively (reviewed in Kvetnansky et al., 2009). Catecholamine release in both systems is regulated by cholinergic projections from the intermediolateral cell column of the spinal cord (T1-L2) and release of catecholamines is triggered by binding of nicotinic and muscarinic receptors on sympathetic ganglia and chromaffin cells (reviewed in Kvetnansky et al., 2009). In turn, the firing of cholinergic cells in the spinal cord is regulated by descending projections from integrative centres in the limbic system, hypothalamus, cortex, and the brainstem (see reviews by Kvetnansky et al., 2009, Ulrich-Lai and Herman, 2009). The primary role of peripherally released catecholamines is to redistribute resources toward processes that promote survival, resulting in increased glucose utilization, increased heart rate, pupil dilation, bronchodilation, decreased gut motility and vasoconstriction.

In contrast to the rapid effects exerted by SNS catecholamines, glucocorticoid effects are slower due to the time required to initiate their release and act on the genome, although there is increasing evidence for rapid actions of glucocorticoids at membrane receptors (Karst et al., 2005, Mikics et al., 2004, Sandi et al., 1996). Concentrations in circulation peak approximately 10–20 min after stress exposure and transcriptional effects become evident beginning one hour after stress exposure (Sapolsky et al., 2000). Glucocorticoid release is the endpoint of HPA axis activation, which is initiated in the paraventricular nucleus of the hypothalamus (PVN). The PVN releases corticotrophin releasing hormone (CRH) and arginine vasopressin (AVP) in response to stressor-specific inputs from sensory, brainstem, limbic and cortical regions. In turn, CRH and AVP stimulate the release of ACTH (adrenocorticotropic hormone) from the anterior pituitary, which initiates the release of glucocorticoids from the adrenal cortex. The release of glucocorticoids is moderated also by factors other than pituitary ACTH, such as paracrine factors within the adrenal and a multi-synaptic pathway from the suprachiasmatic nucleus of the hypothalamus to the adrenal (rev. in Bornstein et al., 2008). Glucocorticoids are transported in the circulatory system via corticosteroid binding globulin (CBG), which regulates their access to cells (Breuner and Orchinik, 2002, Hammond et al., 1991). The HPA axis is involved in “allostasis”, which means that the axis is a mechanism by which adaptive changes are initiated to maintain physiological stability, influencing diverse actions such as energy metabolism, immune and cardiovascular function through actions of glucocorticoids on almost every cell type of the body (Korte et al., 2005, McEwen, 2000, Munck et al., 1984).

The genomic actions of glucocorticoids are mediated by mineralocorticoid (MR) and glucocorticoid (GR) receptors. MR and GR are found primarily in the cytosol and nucleus of target cells, which glucocorticoids access by passive diffusion (for evidence of GR in synaptosomes, see Wang and Wang, 2009). Upon binding and translocation to the nucleus, the ligand–receptor complexes act as transcription factors via transactivation (actions of hetero- or homo-dimers at response elements in promoter regions or introns of responsive genes) and transrepression [actions of monomers at transcription factors such as activator protein 1 (AP-1) and cyclic AMP response element-binding (CREB)] (rev. in Datson et al., 2008, De Bosscher et al., 2008). The genomic actions also involve the recruitment of coactivators and corepressors. MR has a 10-fold higher affinity for corticosterone than does GR, and most MR are occupied even under low basal concentrations of hormone (> 80% MR occupied compared to ~ 15% of GR) (reviewed in de Kloet et al., 2008). GR are thus more sensitive to changes in glucocorticoid concentrations, and are involved in negative feedback (reviewed in Datson et al., 2008).

GR are more widely distributed in the brain than are MR, with the highest densities found in limbic regions of the brain, particularly the hippocampus, medial prefrontal cortex (mPFC), and amygdala, which provide important inputs (largely indirectly via the bed nucleus of the stria terminalis) to the PVN (Ulrich-Lai and Herman, 2009) (see Fig. 1). The hippocampus and infralimbic mPFC are implicated in the negative feedback regulation of the HPA axis, whereas nuclei of the amygdala and the prelimbic mPFC are involved in activating the HPA axis (reviewed in Ulrich-Lai and Herman, 2009). Whereas much of the role of glucocorticoids throughout the body is allostatic and involves meeting the immediate energy demands initiated by a stressor, it is through their actions in the CNS that glucocorticoids have more lasting effects. By altering synaptic structure and function in pathways mediating goal-directed behaviour (dopamine pathway from the ventral tegmental area to the nucleus accumbens) and learning and memory (mPFC, amygdala, and hippocampus), glucocorticoids influence the behavioural and physiological responses to subsequent encounters (described more in later sections).

For rodents and for people, adolescence is a transitional time between childhood and adulthood that lacks clear markers of onset and offset. Whereas some aspects of development are continuous and quantitative, other aspects of development involve discontinuities and qualitative change. An example of the former is skeletal maturation in the rat, which proceeds along an exponential trajectory from birth continuously through adolescence and begins to slow toward adulthood at approximately postnatal day 65 (P65) (Horton et al., 2008). An example of the latter is the onset of puberty, a hallmark of adolescence that involves the maturation of the hypothalamic-pituitary-gonadal (HPG) axis and other systems for reproductive function. Behavioural factors also are implicated in puberty. For example, in males, genital self-grooming, which is highest at puberty, is linked to the growth of the prostate and seminal vesicles (reviewed in Hernandez-Gonzalez, 2000).

Adrenarche precedes gonadarche in rats, and the adrenal gland is thought to influence development of the HPG axis (reviewed in Pignatelli et al., 2006) (for a review of prenatal and postnatal development of the HPA axis, see Walker and McCormick, 2009). Basal corticosterone concentrations rise gradually from about P14 to approximately P40 in both sexes, after which the increase is steeper in females than in males resulting in higher concentrations in females than in males at all ages thereafter (Pignatelli et al., 2006). Adult-like concentrations in both sexes are found by about P45 (Pignatelli et al., 2006). Serum luteinizing hormone (LH) and estradiol concentrations increase around P21 (weaning) to peak in female rats at about P35 (Zapatero-Caballero et al., 2004), with the first ovulation occurring at the time of vaginal opening (Ojeda and Urbanski, 1994). Regular estrous cycles begin approximately a week after vaginal opening (reviewed in Lohmiller and Sonya, 2006). Sperm production begins around the day of preputial separation, but is not optimal until several weeks later (reviewed in Lohmiller and Sonya, 2006, Ojeda and Urbanski, 1994). In males, there is a slight increase in serum testosterone concentrations from P21 to P45, and a steeper increase to adult concentrations at P60 (Pignatelli et al., 2006, Zapatero-Caballero et al., 2003). Sexual maturity as evidenced by mature reproductive behaviour is not found until several weeks after puberty in some strains of male rats, although the reproductive behaviour of other strains, such as Long-Evans rats, may mature earlier (reviewed in Hernandez-Gonzalez, 2000).

Because the timing of some of the neuroendocrine changes of puberty are associated with readily measureable discrete physical markers (e.g., day of vaginal opening and first ovulation, P35 ± 2; day of preputial separation and sperm in the epididymis, P42 ± 2), these markers are often used in defining adolescence (McCormick and Mathews, 2007). Thus, day of vaginal opening and of preputial separation are widely used as readily accessible indicators for developmental status. However, the changes in the nervous system over adolescence involve more than the HPG axis and reproductive function (Sisk and Zehr, 2005), and involve a timeframe that extends before and after the overt physical markers of puberty. As a result, a widely used and conservative age range for classification as adolescent in rodents is postnatal days 28–42 (P28–P42), with weaning most commonly occurring within P21–P25 and designated as a juvenile period, and classification as adult at P60 (Spear, 2000) (see Fig. 2). For male rats, this classification places adolescence in a pre-pubertal period, whereas classification of adolescence in people places it in the years after puberty, with a relatively shorter time included before puberty. A less conservative classification system used for rats spans P21 to P59 and includes three subgroups; a prepubescence/early adolescence period from P21 to P34, a mid-adolescence period from P34 to P46, and a late adolescence period from P46 to P59 (Tirelli et al., 2003). The same ages have been used for defining adolescence in males and females, although some aspects of development mature earlier in adolescence in males (e.g., volume of the locus coeruleus attained at P45 in males and P60 in females, Pinos et al., 2001) and others mature earlier in females than in males (e.g., gonadal function, Pignatelli et al., 2006).

Some of the earliest studies of HPA function in response to stressors in adolescence were those of S. Schapiro and colleagues. They proposed that before P30, the negative feedback system of the HPA axis was immature based on evidence of higher plasma corticosterone concentrations 60 min after footshock in rats at P21 and P25 than at P30 (Schapiro et al., 1962) and no difference in post-weaning juveniles and adults in the half-life of corticosterone (Schapiro et al., 1971). Goldman et al. (1973) found that the time course of corticosterone release after novelty stress was the same in P25 rats as in adults (P65), but that P25 rats had higher and more prolonged release of corticosterone than adults when exposed to ether or to footshock, which suggests that the nature of the stressor or stressor intensity is a key factor in age differences in HPA function. They also reported that the corticosterone response of P45 rats to footshock did not differ from that of P65 rats, suggesting that HPA function may be mature by P45. Some evidence for differences in HPA negative feedback was the finding of less suppression of corticosterone responses to stress after peripheral injection of dexamethasone in P27–28 than in P81–86 rats, although the two groups did not differ in suppression when tested after hypothalamic implants of dexamethasone (Goldman et al., 1973).

In the years since the study by Goldman and colleagues, the investigations of the development of the HPA axis over adolescence have been relatively few, although the available studies also report higher or prolonged ACTH and corticosterone release in pre-pubertal adolescent compared to adult males (Cruz et al., 2008, Gomez et al., 2002, Romeo et al., 2006a, Romeo et al., 2006b, Romeo et al., 2004a, Romeo et al., 2004b, Vazquez and Akil, 1993). However, this pattern of age difference is less consistently observed in comparisons of pre-pubertal and adult females (Doremus-Fitzwater et al., 2009, Hodes and Shors, 2005, Viau et al., 2005). As found for post-pubertal adolescents, one study found that P46 females did not differ from adult females at any point after 15 min confinement to an elevated platform, and that the corticosterone concentrations of P46 males were less than that of adult males initially after the stress exposure, but were higher than adults 45 and 90 min after return to the home cage (McCormick et al., 2008). In another study, the pattern of corticosterone release after a forced swim did not differ between P46 and adult rats for either males or females (Mathews et al., 2008). In sum, although stage of adolescence, sex of the animals, and type of stressor are important factors in the ontogeny of HPA responses to stress, there is considerable evidence that the response of pre-pubertal males tends to be prolonged relative to that of adult males.

The basis for the prolonged HPA responses to a stressor in adolescence compared to in adults remains unknown, although it cannot be explained by differences in clearance rates of ACTH and corticosterone or by differential sensitivity of the adrenal to ACTH (Goldman et al., 1973, Vazquez and Akil, 1993). Further, there is no difference between adolescents and adults in the binding capacity of MR and GR (Vazquez and Akil, 1993) or in stress-induced regulation of GR mRNA (Romeo et al., 2008). Growth and metabolism appear to be protected from the greater HPA response of young adolescents than adults to an acute stressor: Adolescent and adult male rats did not differ after an acute stressor for several metabolic parameters (e.g., plasma glucose, insulin, thyroid hormones) (Romeo et al., 2007). Another age difference in HPA function is that pre-pubertal adolescents of either sex appear unresponsive to sex steroids, whereas in adulthood, androgens tend to inhibit and estradiol tends to potentiate corticosterone release in response to stressors (reviewed in McCormick and Mathews, 2007).

There are fewer investigations of HPA function under conditions of chronic stress in adolescence than under conditions of acute stress. One study of HPA function after repeated restraint exposures (30 min exposures for seven days beginning at P22 or P70) found that adults maintained elevated corticosterone levels for longer than did pre-pubertal males (Romeo et al., 2006a). Repeated social defeat over P35 to P40 was found to sensitize the corticosterone response to defeat in males (Watt et al., 2009), which was not found after repeated defeat in adulthood (Covington and Miczek, 2005). Another study using five days of 90 min of restraint beginning at either P30–32 or ~ P70 found adolescent males had higher plasma corticosterone concentrations than did adult males during and 30 min after the fifth restraint exposure, and that adult males had a greater reduction in corticosterone release from the first to fifth restraint exposure compared to adolescent males (Doremus-Fitzwater et al., 2009). Adolescent and adult females did not differ significantly in corticosterone release during the fifth restraint exposure. We have found that the extent of habituation of HPA responses to repeated stressors in adolescence depends on the type of stressor, with less habituation when the stressor involved social instability (16 daily changes of cage partners and 1 h isolation, a confinement similar to restraint) than when it involved only repeated isolation (McCormick et al., 2007). In addition, habituation of corticosterone release to social instability in adolescence may be offset by reductions in plasma corticosteroid binding globulins, thereby maintaining high levels of free corticosterone (unbound corticosterone, the fraction considered biologically-active) (McCormick et al., 2007).

Investigations using measures other than the hormones of the HPA axis also suggest that stress responses differ in adolescence than in adulthood in rodents. Studies of pharmacological control of cardiac responses to aversive stimuli found that whereas pre-weanlings had parasympathetic nervous system (PNS) activation and adult rats had PNS withdrawal and SNS activation to noise stress, adolescent rats showed only PNS withdrawal (Kurtz and Campbell, 1994). Another research approach highlighting differences in the stress response of adolescents compared to adults is the investigation of changes in expression of enzymes and proteins in the PVN and other neural regions. In male grass rats, 2 h in a novel environment resulted in higher corticosterone concentrations and greater Fos (a marker of neuronal activation) expression in P28 than in P60 rats in several brain regions (medial amygdala, PVN, cingulate gyrus, lateral septum, paraventricular nucleus of the thalamus and centromedial nucleus of the thalamus) (Novak et al., 2007). In laboratory rat strains, 30 min of restraint elicited Fos expression in the PVN more rapidly in pre-pubertal adolescent than in adult males (Romeo et al., 2006a, Viau et al., 2005), and five days of daily 30 min of treadmill running produced greater Fos expression in the hippocampus at P28 than at P56 (Kim et al., 2004). Another study found no difference between adolescent and adult rats in Fos expression in the PVN after either 15 min or 2 h of restraint, although Fos expression was limited to the PVN in adolescents and was more widespread in adults (Kellogg et al., 1998).

The expression of ACTH secretagogues after a stressor also is age-dependent. After 30 min of restraint, AVP heteronuclear mRNA expression was higher and CRH mRNA expression was lower in the PVN of P28–P32 adolescents compared to adults (Romeo et al., 2006a, Viau et al., 2005). These age differences were not found in females (Viau et al., 2005). Cocaine-and-amphetamine regulated transcript (CART) is a neuropeptide that regulates HPA function and, in turn, its expression is regulated by glucocorticoids in a sex-specific manner (Koylu et al., 2006). CART also has been implicated as a factor in the onset of puberty based on its regulation of GnRH (Lebrethon et al., 2000). Expression of CART mRNA in the dentate gyrus and the central nucleus of the amygdala was higher in P28 than in P77 rats, and expression of CART mRNA in the central nucleus was increased 2 h after 30 min of restraint in P77 and unchanged in P28 rats, perhaps reflecting a ceiling effect (Hunter et al., 2007).

Adolescent males (P38) had increased glutathione peroxidase (GPx; protects against free radicals) activity in the prefrontal cortex to both a low and a higher intensity of footshock, whereas in adult males an increase had previously been found only after higher intensity of acute footshock (Uysal et al., 2005). GPx activity did not increase after either intensity of footshock in adolescent females, but the antioxidant superoxide dismutase activity did increase in the hippocampus and prefrontal cortex of adolescent females and not in adolescent males. These findings suggest that adolescent males may be more sensitive to a stressor than adult males, and that there is sex-specificity in adolescence as to the enzymes activated by a stressor.

Another important consideration is that behavioural responses to stressors change over ontogeny. For example, defensive burying is considered to be part of the rat's natural behavioural repertoire, and refers to the rat's tendency to cover unfamiliar or aversive objects with material using its forepaws and head (see review by De Boer and Koolhaas, 2003). Defensive burying in response to a shock-prod has been used extensively as a measure since first developed by Pinel and Treit (1978). Defensive burying is rare before P21, and increases steadily from P21 to P77, after which it declines (Arakawa, 2007, Lopez-Rubalcaba et al., 1996).

In sum, the available evidence indicates that stress-related behavioural and physiological function of adolescent rodents differs both quantitatively and qualitatively from that of adults, which may reflect the ongoing maturation of the limbic structures that regulate HPA function in adulthood. Thus, the consequences of exposure to stressors in adolescence are likely very different than in adulthood, and the ongoing development over the brain during adolescence may render the animal more vulnerable such that the impact of the stress exposure may be evident long after the stress exposure. Learning and memory processes may be particularly sensitive to stress exposures in adolescence given the well-established effects of glucocorticoids and stressors on learning and memory and given the overlap in the neural structures regulating HPA function and subserving learning and memory. The next sections describe briefly the relationships between HPA function and learning and memory from investigations in adults, learning and memory in adolescence, and then review the literature as to immediate and lasting consequences of stressors in adolescence on learning and memory.

The extra-hypothalamic limbic structures that are important for regulating HPA function, notably the hippocampus, prefrontal cortex, and amygdala, have long-established roles in learning and memory. Their roles also involve the dopaminergic projections from the ventral tegmental area to the nucleus accumbens, a pathway that is important for motivation, and the function of which is regulated by glucocorticoids (Marinelli and Piazza, 2002). Although there are functional interactions among these structures in learning and memory processes (see Fig. 3), there are also distinct contributions among and within the structures. For example, whereas the ventral medial prefrontal cortex is important for working memory and attentional processing, the dorsal medial prefrontal cortex has a greater involvement in memory for motor processes (reviewed in Dalley et al., 2004). Although both regions are involved in spatial learning, the ventral subregion of the hippocampus has a greater involvement in emotional processing than the dorsal subregion, which has a greater involvement in spatial learning and memory and contextual fear conditioning (reviewed in Bannerman et al., 2004, Jay et al., 2004). The differential involvement of the neural structures in memory is often more apparent under conditions of stress (e.g., Chauveau et al., 2008, Chauveau et al., 2009).

Intrinsic stress (i.e. stress related to the cognitive task) generally enhances the consolidation of memory through actions of norepinephrine and glucocorticoids on the neural circuits activated by the learning experience (see review by Bisaz et al., 2009). When the stress derives from conditions other than the cognitive task (i.e. extrinsic stress), the effects are more varied and more specific to the type of learning involved (Bisaz et al., 2009). Acute extrinsic stress enhances aversive hippocampal-dependent tasks such as contextual fear conditioning and trace eye-blink conditioning (e.g., Sandi et al., 2005, Shors, 2001), and impairs hippocampal-dependent spatial memory retrieval without impairing the ability to learn a spatial memory task (see review by Howland and Wang, 2008). Long-term potentiation (LTP) and long-term depression (LTD) are considered to be the neurophysiological correlates of the plasticity associated with learning and memory, and are highly sensitive to extrinsic stressors. For example, acute stressors or glucocorticoids impede LTP in the hippocampus (Foy et al., 1987), in the basolateral amygdala–prefrontal cortical pathway (Maroun and Richter-Levin, 2003), and in the hippocampal–prefrontal cortical pathway (Rocher et al., 2004). Consistent with the role of the basolateral amygdala in modulating stress-related memory consolidation in other brain regions (reviewed in Roozendaal et al., 2009), the amygdala also regulates LTP in the hippocampus (e.g., Kim et al., 2001).

Exposure to recurring or chronic stress impairs medial prefrontal cortical-dependent tasks such as the recall of extinction of conditioned fear (Miracle et al., 2006) and attentional set-shifting (Cerqueira et al., 2007, Liston et al., 2006) and impairs hippocampal-dependent performance in spatial tasks such as the spatial Y-maze (Conrad et al., 1996) and radial arm maze (Luine et al., 1994). Numerous markers of neuroplasticity are altered by chronic stress and may contribute to the changes in learning and memory (see reviews by Conrad, 2008, Czeh et al., 2008, Fuchs et al., 2006, Lupien et al., 2009). Chronic stress and chronic elevation of glucocorticoids result in retraction of dendrites in the hippocampus (McLaughlin et al., 2007, Watanabe et al., 1992), the medial prefrontal cortex (Brown et al., 2005b, Radley et al., 2004, Shansky et al., 2009, Wellman, 2001), and, depending on the type of stressor, enhancement or retraction of dendrites in the basolateral amygdala (Vyas et al., 2002). Chronic stress and/or chronic elevation of glucocorticoids decrease neurogenesis in the dentate gyrus (Cameron and Gould, 1994, Dagyte et al., 2009, Pham et al., 2003).

Most of the research described above involved male rats only. When females have been included, effects of acute and chronic stress on behavioural and neural plasticity are often in the opposite direction than for males (e.g., Bowman et al., 2001, Shors, 2006, Shors et al., 2001, Westenbroek et al., 2004), which likely reflects sex differences in neural structures, in HPA function, and in HPG–HPA interactions.

In sum, chronic stress dramatically alters brain structure and function. However, the effects of chronic stress in adulthood on learning and memory typically are not lasting (Luine et al., 1994, McEwen and Sapolsky, 1995, Sousa et al., 2000), and the changes in neural structure also are reversed with time (e.g., Sousa et al., 2000), at least in the hippocampus if not in the amygdala (Vyas et al., 2004). Whether the effects of chronic stress in adolescence results in greater and or more enduring changes for learning and memory is considered in the next sections by first reviewing the development of the relevant brain structures in adolescence, learning and memory function in adolescence, and then the consequences of stressors in adolescence for learning and memory in adulthood.

The brain regions that are critically involved in learning and memory, including the medial prefrontal cortex (mPFC), hippocampus, and amygdala, have high expression of corticosteroid receptors, are involved in regulation of the HPA axis, and in turn, their function is influenced by glucocorticoids. These regions of the rodent brain undergo extensive morphological and functional remodeling in adolescence, with maturation occurring earlier in subcortical limbic regions than in frontal cortical regions (rev. in Andersen, 2003, Casey et al., 2008, Crews et al., 2007, Spear, 2000).

Prominent changes are observed in synaptic complexity and reorganization of projections that confer greater control over emotional and cognitive processes to the mPFC (Andersen, 2003, Chambers et al., 2003). In the ventral mPFC, frontal white matter volume increases between P30 and P90 in both sexes, whereas neuron number decreases in females more extensively than in males during the same period (Markham et al., 2007). Glutamatergic projections from the basolateral amygdala and dopaminergic projections from the ventral tegmental area to the mPFC continue to increase throughout adolescence and into young adulthood (> P60) (Benes et al., 2000, Berger et al., 1985, Cunningham et al., 2002, Cunningham et al., 2008, Kalsbeek et al., 1988), whereas cortical GABA and forebrain NMDA receptors in the mPFC decrease after P30 (Insel et al., 1990, Yu et al., 2006). Glutamatergic projections from the mPFC to the core of the nucleus accumbens also increase progressively throughout adolescence (Brenhouse et al., 2008) and maturation of electrophysiological properties of striatal medium spiny neurons parallels the development of cortical inputs to this region (Kasanetz and Manzoni, 2009, Tepper et al., 1998). In addition, dopaminergic regulation of electrophysiological properties of mPFC neurons and of corticoaccumbens synapses matures between adolescence and adulthood (Benoit-Marand and O'Donnell, 2008, Tseng and O'Donnell, 2007, Yu et al., 2006).

The volume of the amygdala (Koshibu et al., 2004, Rubinow et al., 2009a) and the hippocampus (Koshibu et al., 2004, Yildirim et al., 2008) increases during adolescence. There is also an increase in cholinergic projections from the basal forebrain to the basolateral amygdala that continues into adulthood (P60) (Berdel et al., 1996), whereas the number of neurons and glia in this brain region decreases from P35 to P90 (Rubinow et al., 2009a). In the hippocampus, neurogenesis and dendritic spine density are higher just before puberty (< P35) and decrease to adult levels thereafter (He and Crews, 2007, Yildirim et al., 2008). The composition of hippocampal signalling proteins also changes between weaning and adulthood, indicating that signalling cascades in adolescence may be different than in adulthood (Weitzdorfer et al., 2008).

Catecholaminergic systems that play an important role in attention, arousal, goal-directed behaviour, and stress are also remodeled in adolescence. In the ventral tegmental area, dopaminergic cell bodies transiently increase in size between P14 and P30 before decreasing to adult levels by P35 (Park et al., 2000). In contrast, the number of dopaminergic cells in the ventral tegmental area and the substantia nigra does not change in adolescence (Arbogast and Voogt, 1991). New noradrenegic neurons are also formed in the locus coeruleus (LC) during adolescence, reaching stable levels at P45 in males and P60 in females (Pinos et al., 2001). Changes are also observed at the synaptic level, with dopamine transporter (DAT) density in the midbrain reaching adult levels by P42 (Coulter et al., 1996, Galineau et al., 2004, Tarazi and Baldessarini, 2000).

Dopamine receptors are also transiently elevated in brain regions that are innervated by midbrain dopamine projections, including the nucleus accumbens, dorsal striatum, and the medial prefrontal cortex. Dopamines D1 and D2 receptor density peaks between P28 and P40 in the nucleus accumbens and the striatum, decreasing thereafter to adult levels (Andersen et al., 2000, Gelbard et al., 1989, Giorgi et al., 1987, Tarazi and Baldessarini, 2000, Teicher et al., 1995). In contrast, accumbal D3 receptors continue to increase from P20 to P60 (Stanwood et al., 1997). Receptor overexpression is higher in the striatum than in the nucleus accumbens, is more prominent in male than in female rats (e.g., Andersen et al., 2000, Teicher et al., 1995), and is not attenuated by pre-pubertal gonadectomy (Andersen et al., 2002). In contrast, D1, D2 and D4 receptor density in the hippocampus and the entorhinal cortex increases gradually until adult levels are reached by P35 (Tarazi and Baldessarini, 2000).

A similar but protracted pattern of receptor overexpression and pruning is found also in the PFC, where D2 receptors reach stable levels by P80 and D1 receptors continue to decrease until P100 (Andersen et al., 2000). In addition, serotonin transporter development is also protracted in the PFC, as evidenced by steady increases in receptor density until P90, in contrast to stable levels in the striatum being reached by P50 (Moll et al., 2000). In contrast, noradrenergic transporter density in the frontal cortex reached stable levels by P50, whereas dopamine transporter (DAT) density in the striatum was pruned dramatically between P50 and P90 (Moll et al., 2000). However, others reported that DAT density in the striatum and the NAc reached adult levels by P28, with no evidence of subsequent pruning (Galineau et al., 2004, Tarazi and Baldessarini, 2000).

In sum, the above highlights some of the maturation that occurs over adolescence in key neural structures for learning and memory and that underlie the differences in cognitive performance over ontogeny.

Although there is significant change in the neural structures subserving learning and memory over adolescence, some aspects of learning and memory are mature in early adolescence. Acquisition and performance on associative learning tasks such as eye-blink conditioning and odour-aversion learning is adult-like by P24 (Raineki et al., 2009, Stanton and Freeman, 2000). However, there is evidence that the threshold for aversive motivational learning such as conditioned taste aversions is higher in adolescents than in adults (reviewed in Pautassi et al., 2008). Fear conditioning over ontogeny also has been investigated extensively in rodents (e.g., Barnet and Hunt, 2005, Rudy, 1994). Freezing responses to aversive conditioned stimuli emerge before potentiated startle responses, with potentiated startle responses first appearing around the age of weaning (Hunt et al., 2007), although responses were found to be lower at P28 and P35 than at P56 (Gewirtz et al., 2008). There also is evidence for enhanced contextual fear conditioning in adolescence (Esmoris-Arranz et al., 2008, Ito et al., 2009). Differences between adolescent and adult rodents in contextual fear conditioning suggest that ongoing development of both the amygdala and hippocampus may underlie the age differences (Ito et al., 2009).

Performance on spatial memory tasks that involve long delay intervals is lower in P30 rats than in adults (reviewed in Brown et al., 2005a), although performance on several hippocampal-dependent tasks may appear adult-like in young, pre-pubertal rats (Vorhees et al., 2005). Further, the well-established sex differences in spatial performance emerge only at puberty (Kanit et al., 2000). In addition, efficient search strategies in the Morris water maze task of spatial learning appear relatively late in adolescence (P42) (Schenk, 1985), and adolescents do not show the benefit on retention of spatial learning that adults do from spaced training in the water maze (Spreng et al., 2002). Differences between adolescents and adults in learning and memory cannot be attributed solely to age differences in stress systems impacting performance on stressful tasks, because adolescents showed impaired learning relative to adults on a food-conditioned place preference task that does not involve a stress component (Rubinow et al., 2009b). However, the age difference in the latter study may be due to the stress of, or motivational factors related to, the food deprivation required for the task rather than an age difference in learning and memory.

The experience of stress appears to impact learning and memory differently in adolescence than in adulthood (reviewed in Shors, 2006). In early adolescent rats of either sex (P25–P29), there was no effect of footshock stress on trace eye-blink conditioning, whereas stress enhanced conditioning in male and female rats in mid-adolescence (P35–P40) and in adult males (> P60) and impaired conditioning in adult females (Hodes and Shors, 2005). Thus, evidence of ongoing maturation during adolescence of learning and memory systems based on brain measures and in behavioural performance suggests that these processes may be disrupted or moulded by the experience of stressors in adolescence such that the effect of the stress history is evident in adulthood.

Environmental experiences in adolescence influence later performance on spatial cognition. For example, exposure to enriched environments over adolescence, but not later in life, improved Morris water maze performance in mice (Williams et al., 2001) and Morris water maze performance (Paylor et al., 1992) and radial arm maze performance in rats (Lores-Arnaiz et al., 2007). In contrast, exposure to stressors in adolescence has a negative impact on spatial cognition. Exposure to variable physical stressors (forced swim, restraint, cold, noise, ether) and not variable social stressors (litter shifting, subordination, crowding, isolation, novel environment) daily from P28 to P56 in male rats led to decreased performance in a Morris water maze test and decreased hippocampal volume several weeks later (tested at P77) (Isgor et al., 2004). The effect of variable physical stress on hippocampal structure and spatial performance was not evident when tested at P56, which suggests that the stress exposures altered ongoing maturation. A milder, briefer stress procedure (30 min on elevated platform) administered daily from P28 to P30 in male rats also altered performance in a Morris water maze when tested in adulthood, particularly when the adolescent stress was followed by a new stress exposure in adulthood (Avital and Richter-Levin, 2005). The performance difference among the groups was interpreted as more efficient spatial memory in the control groups than in the stress groups. In another study, aged male mice tested 12 weeks after a chronic social stress procedure (twice weekly change of cage mates) administered for seven weeks over adolescence and early adulthood had decreased spatial recognition memory in a Y maze and decreased performance in a Morris water maze, and did not differ in social recognition or object recognition memory, compared to controls (Sterlemann et al., 2009). The chronic-stressed mice also had reduced long-term potentiation in the CA1 pyramidal layer and reduced expression of several hippocampal proteins associated with learning and memory compared to controls.

Lasting effects of stress in adolescence have also been observed for fear conditioning. Male rats that were housed in isolation for three weeks beginning at P21 and then returned to social housing for two weeks did not differ from controls in unconditioned fear but had enhanced auditory fear conditioning compared to controls (Lukkes et al., 2009). A briefer stress exposure (exposure to cat odour P28–P30) had no effect on fear conditioning (contextual and auditory) in either males or females when tested beginning at P83; however, males were impaired in the extinction of the conditioning (Toledo-Rodriguez and Sandi, 2007). Stress at an earlier period post-weaning (footshocks P21 to P25) also impaired extinction of contextual fear conditioning and impaired the reversal of decreased synaptic transmission in the medial PFC during the extinction trials when tested 10–12 weeks later, whereas the same stress administered pre-weaning was without effect (Koseki et al., 2009). Three days of varied stress exposures (forced swim, elevated platform, and footshock) in pre-pubertal male rats was found to impair shuttle avoidance when tested in adulthood (Tsoory et al., 2007, Tsoory and Richter-Levin, 2005).

Evidence that the impact of a manipulation in adolescence on adult learning and memory is greater than when administered in adulthood comes from a recent study in which the demyelinating drug cuprizone was administered to mice either from P29 to P56 or from P57 to P84 (Makinodan et al., 2009). When tested immediately after treatment, both age groups showed impairments compared to controls in a spatial memory Y-maze test. However, when tested at P126 after a remyelination period, whereas the adult-treated rats had improved performance, the adolescent-treated rats continued to be impaired despite their lengthier time in recovery. Neither group was impaired on a non-hippocampal-dependent object recognition task (a perirhinal cortex-dependent task), which, together with the results above involving stressors, suggests that the later developing hippocampal–prefrontal cortical regions may be particularly sensitive to manipulations in adolescence.

Another indication that developing learning and memory systems may be altered by stressors in adolescence is evidence from investigations of drug-related plasticity. Drugs of abuse produce structural changes in the nucleus accumbens, hippocampus, and PFC, and the molecular mechanisms underlying the plasticity are thought to be the same as those underlying learning and memory (Robinson and Kolb, 2004). Thus, any adolescent stress-induced alteration in learning and memory may be accompanied by altered behavioural and structural responses to drugs of abuse. Although there are few of such studies, and the studies differ as to stress procedures, age at testing, and drug (amphetamine, nicotine, cocaine, alcohol), the general finding is that stress exposures in adolescence alter sensitization to drugs of abuse (reviewed in McCormick, 2009).

The investigation of lasting consequences of stress exposures in adolescence is in its infancy, and many challenges remain. For example, many of the learning and memory tests that have been used are inherently stressful (e.g., Morris water maze, fear conditioning). Thus, performance may be affected by the extent to which an adolescent stress procedure alters HPA responses to a new stressor, which many adolescent stress procedures do (e.g., Mathews et al., 2008, Uys et al., 2006). Similarly, adolescent stress-induced changes in emotional or motivational systems (reviewed in McCormick et al., 2010) may influence performance on learning and memory tests, such that performance may reflect, for example, increased anxiety rather than deficits in the working memory per se. Use of a broader range of cognitive tests, particularly ones that are not inherently stressful or that do not involve food deprivation (given the possibility of stress alterations of metabolic or motivational systems) will help elucidate the specific neuropsychological processing underlying performance differences related to stressors in adolescence.

Another important question to address is the extent to which the enduring effects of stressors in adolescence are mediated or moderated by changes in the HPG axis. First, there are widespread activational effects of gonadal hormones on learning and memory, and like glucocorticoids, gonadal hormones influence the underlying neuroplasticity (reviewed in Hajszan et al., 2007, Luine, 2008). There is increasing evidence that gonadal hormones exert relatively permanent organizational effects on ongoing brain development over adolescence, and exposure to gonadal hormones in adolescence influences gonadal actions in the brain in adulthood (e.g., Ahmed et al., 2008, Cooke and Woolley, 2009, Hebbard et al., 2003, Sanz et al., 2008). Given the role of glucocorticoids in the timing of HPG development (Pignatelli et al., 2006), stress exposures may thus influence the effects of the HPG axis on ongoing brain development. For example, isolation-housing is used by many labs as a stressor in rats, and female rats singly-housed since P28 reached reproductive maturity earlier as indicated by regular estrous cycles that were shorter in length than those of group-housed females (Hermes and McClintock, 2008). Stress of injection alone over adolescence (from P21 to P60) also was found to alter estrous cyclicity in female rats (Raap et al., 2000). In mice, stress in adolescence altered behavioural and physiological effects of gonadal hormones in adulthood (Laroche et al., 2009). Given the role of gonadal hormones in modulating HPA responses to stressors (McCormick and Mathews, 2007), the effects of stress in adolescence on adult HPA function may be mediated by the HPG axis. Further investigation of HPA–HPG interactions will better our understanding of some of the sex-specific effects of stressors in adolescence.

In sum, there is ample evidence that glucocorticoid hormones and the exposure to chronic stressors influence the neuroplasticity of learning and memory systems. The available evidence suggests that the transitions of adolescence in HPA and HPG function and ongoing maturation of the CNS render the animal uniquely vulnerable to long-lasting changes in learning and memory after chronic stress. Continued investigation of risk and resilience in the adolescent period in rodents will serve to understand the bases for greater vulnerability in adolescent humans for stress-related pathologies.

Section snippets

Acknowledgements

CMM is funded by a Natural Sciences and Engineering Research Council of Canada Discovery Grant and the Canadian Foundation for Innovation. IZM holds a Natural Sciences and Engineering Research Council of Canada Graduate Fellowship.

References (184)

  • B.J. Casey et al.

    The adolescent brain

    Dev Rev

    (2008)
  • F. Chauveau et al.

    Prefrontal cortex or basolateral amygdala lesions blocked the stress-induced inversion of serial memory retrieval pattern in mice

    Neurobiol Learn Mem

    (2008)
  • F. Chauveau et al.

    The hippocampus and prefrontal cortex are differentially involved in serial memory retrieval in non-stress and stress conditions

    Neurobiol Learn Mem

    (2009)
  • C.L. Coulter et al.

    Postnatal development of the dopamine transporter: a quantitative autoradiographic study

    Dev Brain Res

    (1996)
  • F. Crews et al.

    Adolescent cortical development: a critical period for vulnerability for addiction

    Pharmacol Biochem Behav

    (2007)
  • B. Czeh et al.

    Chronic stress-induced cellular changes in the medial prefrontal cortex and their potential clinical implications: does hemisphere location matter?

    Behav Brain Res

    (2008)
  • J.W. Dalley et al.

    Prefrontal executive and cognitive functions in rodents: neural and neurochemical substrates

    Neurosci Biobehav Rev

    (2004)
  • N.A. Datson et al.

    Central corticosteroid actions: search for gene targets

    Eur J Pharmacol

    (2008)
  • S.F. De Boer et al.

    Defensive burying in rodents: ethology, neurobiology and psychopharmacology

    Eur J Pharmacol

    (2003)
  • E.R. de Kloet et al.

    Neuropharmacology of glucocorticoids: focus on emotion, cognition and cocaine

    Eur J Pharmacol

    (2008)
  • T.L. Doremus-Fitzwater et al.

    Social and non-social anxiety in adolescent and adult rats after repeated restraint

    Physiol Behav

    (2009)
  • F.J. Esmoris-Arranz et al.

    Contextual fear conditioning differs for infant, adolescent, and adult rats

    Behav Processes

    (2008)
  • M.R. Foy et al.

    Behavioral stress impairs long-term potentiation in rodent hippocampus

    Behav Neural Biol

    (1987)
  • L. Galineau et al.

    Ontogeny of the dopamine and serotonin transporters in the rat brain: an autoradiographic study

    Neurosci Lett

    (2004)
  • H.A. Gelbard et al.

    Postnatal development of dopamine D1 and D2 receptor sites in rat striatum

    Brain Res

    (1989)
  • J.C. Gewirtz et al.

    Effects of gestational iron deficiency on fear conditioning in juvenile and adult rats

    Brain Res

    (2008)
  • O. Giorgi et al.

    Developmental and age-related changes in D1-dopamine receptors and dopamine content in the rat striatum

    Dev Brain Res

    (1987)
  • G.L. Hammond et al.

    Molecular studies of corticosteroid binding globulin structure, biosynthesis and function

    J Steroid Biochem Mol Biol

    (1991)
  • J. He et al.

    Neurogenesis decreases during brain maturation from adolescence to adulthood

    Pharmacol Biochem Behav

    (2007)
  • P.C. Hebbard et al.

    Two organizational effects of pubertal testosterone in male rats: transient social memory and a shift away from long-term potentiation following a tetanus in hippocampal CA1

    Exp Neurol

    (2003)
  • M. Hernandez-Gonzalez

    Prepubertal genital grooming and penile erections in relation to sexual behavior of rats

    Physiol Behav

    (2000)
  • G.E. Hodes et al.

    Distinctive stress effects on learning during puberty

    Horm Behav

    (2005)
  • J.G. Howland et al.

    Synaptic plasticity in learning and memory: stress effects in the hippocampus

    Prog Brain Res

    (2008)
  • R.G. Hunter et al.

    Regulation of CART mRNA by stress and corticosteroids in the hippocampus and amygdala

    Brain Res

    (2007)
  • T.R. Insel et al.

    The ontogeny of excitatory amino-acid receptors in rat forebrain.1. N-methyl-d-aspartate and quisqualate receptors

    Neuroscience

    (1990)
  • L. Kanit et al.

    Sexually dimorphic cognitive style in rats emerges after puberty

    Brain Res Bull

    (2000)
  • C.K. Kellogg et al.

    Adolescent development alters stressor-induced fos immunoreactivity in rat brain

    Neuroscience

    (1998)
  • S.M. Korte et al.

    The darwinian concept of stress: benefits of allostasis and costs of allostatic load and the trade-offs in health and disease

    Neurosci Biobehav Rev

    (2005)
  • K. Koshibu et al.

    Sex-specific, postpuberty changes in mouse brain structures revealed by three-dimensional magnetic resonance microscopy

    Neuroimage

    (2004)
  • E.O. Koylu et al.

    Cocaine and amphetamine regulated transcript (CART) and the stress response

    Peptides

    (2006)
  • S. Lores-Arnaiz et al.

    Exposure to enriched environments increases brain nitric oxide synthase and improves cognitive performance in prepubertal but not in young rats

    Behav Brain Res

    (2007)
  • V. Luine et al.

    Repeated stress causes reversible impairments of spatial memory performance

    Brain Res

    (1994)
  • J.L. Lukkes et al.

    Adult rats exposed to early-life social isolation exhibit increased anxiety and conditioned fear behavior, and altered hormonal stress responses

    Horm Behav

    (2009)
  • J.A. Markham et al.

    Neuron number decreases in the rat ventral, but not dorsal, medial prefrontal cortex between adolescence and adulthood

    Neuroscience

    (2007)
  • I.Z. Mathews et al.

    Heightened neuroendocrine function in males to a heterotypic stressor and increased depressive behaviour in females after adolescent social stress in rats

    Behav Brain Res

    (2008)
  • C.M. McCormick

    An animal model of social instability stress in adolescence and risk for drugs of abuse

    Physiol Behav

    (2010)
  • C.M. McCormick et al.

    HPA function in adolescence: role of sex hormones in its regulation and the enduring consequences of exposure to stressors

    Pharmacol Biochem Behav

    (2007)
  • E.I. Ahmed et al.

    Pubertal hormones modulate the addition of new cells to sexually dimorphic brain regions

    Nat Neurosci

    (2008)
  • S.L. Andersen et al.

    Dopamine receptor pruning in prefrontal cortex during the periadolescent period in rats

    Synapse

    (2000)
  • H. Arakawa

    Ontogeny of sex differences in defensive burying behavior in rats: effect of social isolation

    Aggress Behav

    (2007)
  • Cited by (173)

    • Adolescent neurocognitive development

      2023, Encyclopedia of Child and Adolescent Health, First Edition
    View all citing articles on Scopus
    View full text