Effects of early life trauma are dependent on genetic predisposition: a rat study

There is increasing evidence that environmental factors, particularly stressful events experienced early in life, increase the risk of developing a psychiatric illness and/or a behavioural disorder [1‐4]. Following birth, the brain continues to develop, rendering it vulnerable to adverse external influences. These may lead to persistent behavioural, endocrine, and neurochemical changes [5‐8]. It is unclear, however, how the physiological and psychological consequences of early life stress are influenced by an inherited predisposition to a developmental disorder, such as attention-deficit/hyperactivity disorder (ADHD).

ADHD is a heterogeneous behavioural disorder characterised by inattention, impulsivity, and hyperactivity [9, 10]. It is one of the most common child-psychiatric disorders [11]. Although there is evidence for ADHD having a genetic component, the exact aetiology of the disorder is unknown [10, 12]. It is possible that early life experiences interact with the development and expression of an inherited ADHD genotype [10, 13].

Maternal separation is a valid model used in research to mimic early life trauma during childhood in humans [7, 14‐18]. Applied to rats, it involves separating the pups from their mother for a set period of time each day during the first 2 weeks of life. Maternal separation has been found to permanently modify characteristics of the stress response system in offspring, leading to elevated and prolonged stress-induced secretion of corticosterone (the rat equivalent of human cortisol) and adrenocorticotropic hormone [18, 19]. These findings of a prolonged endocrine stress response suggest that the stress response system adapts to early life trauma by diminishing negative feedback regulation [18, 19]. Most maternal separation paradigms also result in behavioural changes as these rats show increased anxiety-like and/or depressive-like behaviours [16, 20, 21]. However limited attention has been given to the effects of maternal separation on genetic animal models of human disorders.

While the mechanisms underlying the long-term effects of environmental stress early in life are not known, they are likely to involve activation of intracellular signalling pathways, leading to modifications of the genome, resulting in changes in gene expression and neural function (see [22]). In this way the stressful early-life environment programmes gene expression of the offspring, so that development of a defensive physiological response to stress occurs - a response that would possibly confer protection should the stressful environment persist [22]. Given the influence of early life environmental conditions on gene expression, it is of interest to investigate how the long-term effects of early life trauma are influenced by the individual's genotype. In this view, the long-term effects of maternal separation may differ between different genetic animal models of human disorders. Furthermore, interactions between early life experience and genetic predisposition may account for the variation of the behavioural characteristics of human disorders [23‐28].

Another factor influencing the outcome of maternal separation stress may be the sex of the rat. Although many studies have investigated effects of maternal separation in both male and female rats [29‐32], there is little consensus on whether one sex is more vulnerable to the effects of postnatal environmental manipulations than the other. It was shown that Sprague-Dawley (SD) females (P28-30) were more resistant to the effects of maternal separation (6 h/day) on measures of impulsive behaviour and cerebral cortical thickness compared to males [31]. Using Wistar rats, an increase in corticotropin-releasing factor-positive cells was found in the paraventricular nucleus of the hypothalamus after an acute swim-stress of maternally separated females, but not males, suggesting that maternal separation increased stress responsivity in females only [29]. Differences in the strain of rat used when investigating the factor of sex on maternal separation effects may account for contradictory findings in the literature.

The present study therefore investigated effects of maternal separation in a rat model of ADHD, the spontaneously hypertensive rat (SHR). The SHR is a well established animal model of ADHD, displaying the three main behavioural characteristics of the disorder: inattention, impulsivity, and hyperactivity, when compared to their normotensive control strain, the Wistar-Kyoto rats (WKY) [33‐35], from which they were originally bred [36]. The ADHD-like behaviour of SHR is genetically determined and not a consequence of the SHR dam's behaviour towards the pups, since cross-fostering SHR pups onto WKY or SD dams did not alter their neurochemistry or their behaviour [37]. The use of SHR to investigate how a genetic predisposition to developing ADHD-like behaviour alters the consequences of early life stress is therefore relevant and appropriate. The aim of the present study was to investigate the effects of chronic, mild, developmental stress on behaviour and basal levels of corticosterone in prepubescent male and female rats of different genotypes - the SHR and WKY strains.

Behaviour of the experimental rats was assessed during the 5^th week of life (prepuberty) using the elevated plus maze (EPM), a valid test for assessing levels of anxiety-like behaviour [48], and the forced-swim test (FST), a putative assessment of depressive-like behaviour [42], originally derived to screen the efficacy of antidepressant drugs [49, 50]. At the end of the 5^th week (P35), 4 days after the last behavioural testing day, the rats were killed by decapitation. Their trunk blood was collected in order to measure basal plasma corticosterone levels to determine hypothalamic-pituitary-adrenal (HPA) axis activity in SHR and their control WKY strain. Results show that SHR do not develop increased anxiety-like behaviour following early life trauma as WKY do, but instead respond to the trauma by developing increased behavioural activation, behaviour that correlates positively with high basal plasma corticosterone levels.

Results from the current study show that maternal separation affects anxiety-like behaviour and locomotor activity of SHR and WKY in different and possibly opposing ways. Maternal separation increased the distance travelled in the EPM and number of entries into the open arms made by SHR, particularly in the females, while it had the opposite effect on WKY, suggesting that SHR developed behaviour that was more active and devoid of anxiety following early life trauma, while WKY developed increased anxiety-like behaviour that was less active following early life trauma. The increased anxiety-like behaviour and decreased behavioural activity observed in WKY following maternal separation agrees with reported effects of maternal separation on other rat strains [16, 51].

The distance travelled in the EPM reflects activity within a novel environment, and therefore cannot be used as a measure of activity on its own, due to the confounding effect of anxiety. Therefore, changes in distance travelled in the EPM due to maternal separation may be due to altered levels of anxiety and/or altered levels of activity per se. In SHR, maternal separation appeared to increase activity in the EPM, decreasing anxiety-like behaviour. This effect was most notable in the female SHR. This is an important finding considering evidence in the literature showing that the long-term effects of childhood trauma are variable, and highly dependent on the individual [13, 24‐26, 52, 53]. Studies have shown that early life trauma is associated with increased risk of developing a number of different disorders later in life, including anxiety disorders and/or depression, conduct disorder, hyperactivity, psychotic symptoms and ADHD [2, 13, 24‐26, 52‐57]. The disorders which arise in response to environmental adversity during development are thus highly dependent on an individual's genotype [23‐27]. Furthermore, it has been shown that genetic predisposition may confer resilience to the development of a disorder following childhood trauma [24, 26]. A study by Polanczyk et al. [25] found that a haplotype in the corticotrophin releasing hormone receptor gene 1 was associated with protective effects against adult depression in individuals who reported maltreatment during childhood (assessed by the Childhood Trauma Questionnaire). Caspi et al. [26] showed that children with a functional polymorphism in the promoter region of the monoamine oxidase A (MAOA) gene, leading to high levels of MAOA expression, were less likely to develop antisocial behaviour following childhood maltreatment. It has also been shown, by Stevens et al. [24], that the long term effects of early life deprivation (reared in deprived conditions from infancy for more than 6 months before being adopted) on ADHD-like behaviour in childhood to mid-adolescence is moderated by polymorphisms of the dopamine transporter (DAT1) gene.

Our findings contribute to the literature highlighting the importance of genetic predisposition on the outcome of early life adversity. SHR may provide a genetic animal model for the instances when early life trauma leads to the development of abnormal behaviours that are not marked by anxiety and/or depression, but rather by increased activity in a novel environment and decreased anxiety.

Although maternal separation altered the number of open arm entries in the EPM made by SHR and WKY, suggesting altered anxiety-like behaviour, there was no effect of maternal separation on time spent in the open arms of the EPM, a second measure of anxiety-like behaviour. Maternal separation also had no effect on depressive-like behaviour, measured by time spent immobile in the FST, in SHR or WKY. This is contrary to studies showing decreased time spent in open arms [16, 51] and diminished activity in the FST [17, 58] in other rat strains due to maternal separation. There have, however, also been studies in the literature reporting no effect of maternal separation on open arm entries and time spent in the open arms of the EPM [32, 58]. The mentioned studies all focused on the behavioural consequences of maternal separation in adulthood, and did not use SHR or WKY. It is possible that anxiety-like and depressive-like behavioural characteristics that result from maternal separation manifest fully in adulthood, and are only partially expressed in prepubescent rats, or that maternal separation effects are highly strain-dependent.

Consistent with their hyperactive phenotype, SHR travelled a greater distance in the EPM and spent less time immobile in the FST than WKY, as reported in previous studies [37, 59, 60]. Since the environment of the EPM was novel to the animals in the present study, distance travelled in the EPM provided a measure of exploratory behaviour [61]. Decreased distance travelled by WKY due to maternal separation implies increased passive, inhibited behaviour. WKY typically show high levels of behavioural inhibition [62, 63], characterised by a reserved response or inactivity in the face of novelty [64]. In the literature they are reported to show decreased ambulation and high behavioural passivity in a novel environment compared to other strains [65, 66]. SHR, on the other hand, show high behavioural activation, characterised by approach motivation and spontaneity [67], and respond actively in a novel situation [37]. Maternal separation tended to increase the active response of SHR to a novel environment, while it accentuated the inhibited behaviour of WKY, thereby amplifying the genetically predisposed behavioural temperaments of the strains. This finding suggests that the early life experience of separation from the dam influenced development of the behavioural inhibition and/or behavioural activation systems.

Studies have shown that children exposed to an unstable and/or traumatic environment very early in life are at a greater risk of developing chronic behavioural inhibition in childhood [68, 69]. Contrary to this, however, it has also been shown that children exposed to early life stress are at a higher risk of developing behavioural disinhibition [70]. The long term effects of early life trauma on the behavioural inhibition/disinhibition profile of a child may be dependent on the child's genetic predisposition [68, 70]. In one study it was shown that children are more likely to develop behavioural inhibition during middle childhood if they are raised in a low social support environment, and have a specific polymorphism in the promoter region of the gene for the serotonin transporter [68]. Another study found that the MAOA functional promoter polymorphism interacts with family adversity and stressful events experienced early in life to influence behavioural disinhibition in children, marked by hyperactivity and conduct disturbances [70]. In the present study the rat strain predisposed to developing ADHD-like behaviour responded to early life trauma with an increase in disinhibited/active behaviour, while their control strain developed increased behavioural inhibition. These results suggest that genetic predisposition interacts with the environment giving rise to either increased behavioural inhibition or increased behavioural disinhibition/activation.

It has been hypothesised that behaviour in the FST reveals the animal's coping strategy when confronted with an inescapable stressor [65, 71]. While SHR adopted a more active strategy (higher climbing and swimming behaviour), displaying behavioural activation, WKY responded in a more passive manner, showing typical behavioural inhibition. The most dramatic difference in FST behaviour between SHR and WKY was apparent in the first minute, when SHR spent a greater percentage of the time climbing than WKY. The strain effect in climbing behaviour apparent for the total 5-minute period was due solely to the significant difference between the strains in the first minute. Climbing behaviour did not differ significantly between the two strains from the second minute onwards. This finding demonstrates that the initial response to an acute, inescapable stress may be determined by genetic predisposition.

It has been observed that, while having similar effects on time spent immobile in the FST, different antidepressant drugs produce characteristic differential effects on the active behaviours in the FST, depending on the neurotransmitter systems which they target [41, 49, 72]. While selective serotonin reuptake inhibitors (SSRIs), such as fluoxetine, sertaline, and paroxetine, increase levels of swimming in the FST, drugs that block the reuptake of norepinephrine, such as desipramine, and maprotiline, increase climbing behaviour in the FST [41, 49, 71, 73]. Supporting these findings is a study by Brenes et al. [74], in which hippocampal concentrations of norepinephrine and serotonin were measured in SDs using high-performance liquid chromatography coupled with electrochemical detection, and then correlated with behaviour in the FST. They found that hippocampal norepinephrine concentrations correlated positively with climbing behaviour and negatively with immobility, while hippocampal serotonin concentration correlated positively with swimming behaviour and negatively with immobility [74].

Attaching neurochemical profiles to active behaviours in the FST has allowed researchers to infer which neurotransmitter systems are involved in the antidepressant-like effects of novel treatments [41]. Although time spent immobile differed between SHR and WKY in each minute of the 5-minute FST, the difference in the first minute was due to SHR spending significantly more time climbing. From the second minute onwards, differences in immobility between the strains were due to SHR spending more time swimming. It is possible that in response to the acute stress, SHR had a greater and more immediate release of norepinephrine into various parts of the brain, possibly due to decreased autoreceptor-mediated inhibition of norepinephrine release [75], resulting in increased climbing behaviour relative to WKY. This agrees with the physiological role of norepinephrine and its release in response to stress [76, 77], and with in vitro superfusion experiments showing that release of norepinephrine in the prefrontal cortex and hippocampus in response to a pulse of glutamate is significantly greater in SHR compared to WKY [37, 75, 78, 79]. Strain differences in behaviour from the second minute onwards could reflect lower basal levels of serotonin in WKY brain relative to SHR [41]. These suggestions are, however, speculative, and require further investigation.

The climbing behaviour of non-maternally separated SHR increased dramatically from the 3^rd to the 4^th minute of the FST, while that of maternally separated SHR did not, resulting in a significant treatment effect in SHR during this minute. The sudden increase in climbing of non-maternally separated SHR could reflect a second attempt to escape, an attempt that was absent in those SHR that had experienced mild, chronic, developmental stress. This result could indicate a change in the coping strategy of SHR due to the maternal separation stress. However, further research is required. Maternal separation had the opposite effect on WKY climbing behaviour in the 4^th minute, tending to increase climbing behaviour in this minute. While the neurochemical basis of this effect of maternal separation is not known, the result indicates opposing effects of chronic, mild, developmental stress on the same behaviour of the two different rat strains.

The high levels of immobility, also referred to as learned helplessness, exhibited by WKY in the FST, is in agreement with the literature [59, 60, 71], and with their passive behaviour in the EPM. Based on results such as these, WKY has been proposed as an animal model for depression [65, 80]. Other studies suggest that the passive response of WKY in behavioural assays indicates behavioural inhibition, and propose WKY as an animal model of anxiety vulnerability rather than depression [62, 63]. Since maternal separation had no effect on immobility time of WKY in the FST, but did enhance WKY's passivity in the EPM, it appears that maternal separation increased behavioural inhibition of WKY in the face of a novel environment (EPM), but not acute stress (FST).

Antidepressant drugs that block re-uptake of norepinephrine in the brain, but not those that block the re-uptake of serotonin, are effective in decreasing immobility in the FST in WKY [71, 81] and increasing locomotor activity of WKY in the open field test [81], thereby increasing their behaviour towards that of SHR. Since SHR were originally derived from WKY [36], the two strains may still maintain some similarities in their neurochemistry. The differences in the neurochemistry between the two strains that have resulted in SHR being more active and displaying symptoms of ADHD may be related to the noradrenergic system [37, 75, 78, 79]. Furthermore, the effect of maternal separation on levels of behavioural inhibition and activation in SHR and WKY may be related to alterations in noradrenergic transmission.

Sex effects in the EPM showed that females were more active and less anxious than males, in agreement with previous studies [32, 37, 82, 83]. Females spent more time swimming than males in the FST, and the SHR females spent less time immobile than SHR males. These results indicate that females respond in a more active way to an acute, inescapable stress.

High locomotor activity in the novel environment of the EPM correlated positively with active behaviour in the FST across both strains, and negatively with immobility in the FST. The positive correlation suggests that the ability to respond actively in the two different behavioural assays may depend on similar brain circuitry. Furthermore, basal plasma corticosterone correlated positively with distance travelled in the EPM, swimming in the FST, and climbing in the first minute of the FST, and correlated negatively with immobility in the FST. This suggests that basal plasma corticosterone correlates positively with an active response to a novel environment and to an inescapable stress, and correlates negatively with a passive, ambivalent response.

While SHR were found to have elevated basal plasma corticosterone compared to WKY, this finding was likely due to the increased basal plasma corticosterone of SHR females (see figure 3). Maternal separation increased basal levels of plasma corticosterone in SHR females only, demonstrating that effects of chronic mild developmental stress are not only strain-dependent, but also sex-dependent. In a study using Wistar rats, Desbonnet et al. [29] found an increase in CRF cells in the PVN due to maternal separation, in females only. Renard et al. [30] found, also using Wistar rats, reduced HPA axis response to an acute stress in males following maternal separation. While not significant, the HPA axis response to an acute stress appeared heightened due to maternal separation in females [30]. The present study, therefore, contributes to the evidence in the literature for sexually divergent effects of maternal separation on regulation of plasma stress hormones. Since the active response by SHR females in both the EPM and FST in the present study was not influenced negatively by maternal separation, but tended to increase, our results suggest that SHR females adapted to the early life trauma by increasing their behavioural activation, maintaining an active response to the stress of a novel situation or inescapable swim stress. Developmental changes in the locus-coeruleus-norepinephrine system and the HPA axis may have occurred, which is reflected in their persistently high behavioural activation [77]. This adaptive mechanism may have required, or given rise to, the increase in basal corticosterone levels present in maternally separated SHR females.

Since SHR did weigh significantly less than WKY at the time of testing, it is arguable that the strain differences in behaviour are related to differences in weight rather than strain. It is also possible that the difference in weight between strains was due to differences in behaviour - i.e. SHR weighed less than WKY because they were far more active. The strain effects evident in the results do, however, agree with the literature. Since maternal separation and sex had no effect on weight, treatment effects and sex effects present in the results cannot be attributed to differences in weight.

In conclusion, our results suggest that SHR and WKY may be appropriate genetic models for the opposing effects of early life trauma on the behavioural inhibition and behavioural activation systems. It also appears that basal levels of corticosterone are associated with the variation in behavioural response to early life trauma, and may be useful as a marker for susceptibility to a certain behavioural temperament. Further research is required, however, to establish the effects of maternal separation on behavioural inhibition versus behavioural disinhibition in these strains, and to establish the role of basal plasma corticosterone. Investigating these strains with this novel hypothesis in mind may allow establishment of risk factors for the development of various disorders characterised by increased behavioural inhibition or disinhibition that arise following traumatic experiences early in life.