Behavioural and Neurochemical Consequences of Early Weaning in Rodents
Takefumi Kikusui, Companion Animal Research, School of Veterinary Medicine, Azabu University, Laboratory of Veterinary Ethology, Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-17-71 Fuchinobe, Sagamihara, Kanagawa 229-8501, Japan (e-mail: email@example.com).
Among all mammalian species, pups are highly dependent on their mother not only for nutrition, but also for physical interaction. Therefore, disruption of the mother–pup interaction changes the physiology and behaviour of pups. We review how maternal separation in the early developmental period brings about changes in the behaviour and neuronal systems of the offspring of rats and mice. Early weaning in mice results in adulthood a persistent increase in anxiety-like and aggressive behaviour. The early-weaned mice also show higher hypothalamic-pituitary-adrenal activity in response to novelty stress. Neurochemically, the early-weaned male mice, but not female mice, show precocious myelination in the amygdala, decreased brain-derived neurotrophic factor protein levels in the hippocampus and prefrontal cortex, and reduced bromodeoxyuridine immunoreactivity in the dentate gyrus. Because higher corticosterone levels are persistently observed up to 48 h when the mice are weaned on postnatal day 14, the exposure of the developing brain to higher corticosterone levels may be one of the effects of early weaning. These results suggest that deprivation of the mother–infant interaction during the late lactating period results in behavioural and neurochemical changes in adulthood and that these stress responses are sexually dimorphic (i.e. the male is more vulnerable to early weaning stress).
The importance of mother–infant relationships in developing mammalian infants, including the role of nutrition and physical contact, has been characterised from behavioural, endocrine and neurochemical perspectives. In domestic rats (Rattus norvegicus), complete maternal deprivation of intensive grooming and licking behaviours during the first 10 days of the neonatal period has been shown to have serious consequences for the pups later in life, including increased anxiety-like behaviour (1, 2) and enhanced neuroendocrine response to stressors (3, 4), particularly among the male offspring. These behavioural and neuroendocrine changes persist throughout life, and associated epigenetic changes are also found in the central nervous system, indicating that maternal presence during the neonatal period has significant effects on the development of neural networks (5).
In a natural setting, the domestic rat mother stays with the pups for more than 12 h every day and shows intense nursing behaviour until weaning between 25–33 days postpartum (6). In a laboratory setting, female rats generally show postpartum oestrous and become pregnant again, which results in another delivery every 21 or 22 days. Therefore, to increase productivity, laboratory rats are generally weaned on postnatal day 21. Rat pups are able to eat, maintain body temperature and evacuate after approximately 2 weeks of birth (7), and weaning within this period has received scientific attention in terms of its neurobehavioural consequences. Similarly, in domestic mice (Mus musculus), a study based on semi-natural observations described separation from the dam occurred when the offspring reached puberty at the age of 4–5 weeks (8). Furthermore, the natural weaning time depends on environmental resources available during the nursing period (i.e. if mothers rear their offspring under severe conditions, there is a trade-off between weaning earlier, reducing offspring survival and increasing maternal survival) (9). Here, weaning is defined as breakage of the bond between the dam and her offspring, which includes the cessation of suckling and physical separation from the dam with the cessation of social protection by the dam. Such phenomena are generally observed in various kinds of mammalian species, including humans.
Development of behaviour
Play behaviour is typically observed in juvenile animals and occurs before weaning, when siblings are often readily available. The expression of play behaviour appears to be affected by the mother–pup relationship because pups that received less anogenital licking from the mother (10), and pups raised by naturally observed low-licking mothers (11), showed more play behaviour during development. Furthermore, early weaning reduced the rate of play-fighting and exploratory behaviour in adulthood (12). Our recent study found similar results (i.e. early weaning decreased play behaviour in both male and female rats), and these behavioural changes were specific to play behaviour because there were no changes in the other behaviours, such as social sniffing and self-grooming (13). The social development of rats and mice is different in that mice, unlike rats, do not display similar social play behaviour at any developmental stage. However, Terranova and Laviola (14) demonstrated that early weaning increases social interactions, such as jumping to other siblings, among juvenile mice.
A few reports have demonstrated the effects of early weaning on the anxiety-like behaviour of adult rodents. In rats, the effect of early weaning on anxiety-like behaviour was assessed by the elevated plus-maze test, hole-board test and open-field test (15). Rat pups were weaned at 16 or 30 days of age and tested at 7–9 weeks of age. In the elevated plus-maze test, the classical parameters of low anxiety-like behaviour, namely time spent in and frequency of entry into open arms, were lower in the early-weaned rats, which spent more time in closed arms, indicating that early weaning increased their anxiety levels (16). In addition, the hole-board test yielded the clearest intra group difference in anxiety-like behaviour. In early-weaned rats, the frequency and duration of head-dipping behaviour were lower compared to normally-weaned rats, indicating that early weaning increased the anxiety-like behaviour in the hole-board test as well. Data on mice weaned at 14 versus 21 days postpartum and subjected to anxiolytic tests at 8 and 22 weeks of age suggested that the increased fearfulness in the early-weaned animals persisted at least until the middle of their lifespan. Consequently, early weaning appears to have long-lasting effects on the fearfulness of adult laboratory rodents, leading to increased anxiety-like behaviour in response to challenging events presented in novel environments (15–19).
Monogamous reproductive pairs with offspring exist among mice close to each other in an area, and dominant male mice showed high territorial aggression to protect the pair’s nest (8). Once male mice experience sexual intercourse, their territorial aggression increases in comparison with the sexually naïve mice (20). This type of aggression is expressed as biting, lateral posture and tail rattling, which have all been interpreted as threatening behaviour (21). The territorial aggression toward an unfamiliar, sexually mature male mouse introduced in the cage of the resident male was compared between early- (14 days postpartum) and normally-weaned (21 days postpartum) mice; no significant differences were found in the physical attacking/biting frequency. However, the individual variations in the biting frequency increased significantly in the early-weaned group (16, 22). These behavioural changes induced by early weaning might reflect anxiety toward the unfamiliar intruder.
Ecological and social stress-enhanced aggression
Early stress experience probably augments stress responses rather than changing the baseline behaviour or neuroendocrine status in adulthood (23). If so, the behavioural changes observed in early-weaned animals might be more evident under stress. To evaluate this hypothesis, we selected two stressors (i.e. a social stressor and an ecological stressor) aiming to assess the effects of early weaning on stress-related aggression. Social instigation was used as the social stressor because it has been demonstrated to increase aggression (24). In addition, food restriction was used as the ecological stressor because a relationship exists between food deprivation and elevated aggression. We also examined the effect of early weaning on the mRNA expression of 5-hydoroxytryptamine (5-HT) receptors in the hippocampus, particularly 5-HT1A and 5-HT1B, because these are reported to be involved in aggressive behaviour (25). Normally-weaned mice showed twice the baseline level of attack bites after social instigation for 5 min, whereas early-weaned animals were less aggressive after social instigation. However, the early-weaned mice were more aggressive after food-restriction stress than the normally-weaned mice, suggesting a lower threshold for aggressive behaviour after food shortage. Moreover, early-weaned mice had lower 5-HT1B expression levels in the hippocampus than the normally-weaned mice; no effect was found with regard to 5-HT1A expression. These results suggest that manipulation of the weaning time modulates adult aggressive behaviour depending on the stressors imposed and that this change may involve the 5-HT1B receptor system in the hippocampus (17).
In adult female rats, maternal behaviour is characterised by the amount of maternal care that the rat received as an infant (4, 26). Rat pups that have received more maternal care display more consistent maternal behaviour (e.g. grooming), and this phenomenon is nongenetically inherited by the next generation (non-genomic transmission) (4, 26). Considering these lines of evidence, it is very important to understand how early weaning affects the subsequent maternal behaviour of the female offspring in terms of animal welfare and commercial productivity.
Novakova (27) first reported the effects of early weaning on subsequent maternal behaviour in rats, and found that early weaning decreased nest building and the acceptance of fostered pups. Because maternal licking and grooming of the pups has a critical role in non-genomic transmission, we measured these behaviours and the nursing behaviour from postpartum day 15–21 to describe the type of maternal care of which the early-weaned mice were deprived (28). It was shown that the mothers spent 3% of their time licking/grooming or in arched-back nursing of their pups on postpartum day 15, and the duration of these behaviours gradually decreased till postpartum day 21. Taken together, in this period, mothers spent 50% of their time attending to pups (i.e. they adopted a passive nursing posture by lying down next to the pups). These observations suggest that early-weaned female mice receive less maternal licking and grooming behaviour than the normally-weaned mice. Subsequently, the maternal behaviour of early- and normally-weaned females was analysed. Early-weaned female mice performed less licking/grooming and arched-back nursing of their pups than the normally-weaned mice, although the time spent away from or attending to pups did not differ between the groups (28). These findings suggest that early weaning deprives the offspring of maternal care and, consequently, the offspring shows less maternal behaviour in adulthood.
Neuroendocrine response to stress
Neonatal maternal separation is an intense stressor that induces long-lasting hyperactivity of the hypothalamic-pituitary-adrenal (HPA) axis (29, 30). In adulthood, this is characterised by augmented levels of circulating glucocorticoids at rest and in response to mild stressors (29, 30). The maternal environment, particularly the intensity of maternal licking and grooming, alters the expression of glucocorticoid receptors (GR) in the hippocampus, which is the key target of the glucocorticoid feedback loop (31). Rat offspring deprived of maternal care or receiving less maternal grooming had reduced GR expression in the hippocampus, indicating decreased negative feedback by glucocorticoids, followed by increased glucocorticoid secretion in response to stress episodes.
In mice, early weaning has been shown to affect the neuroendocrine response to stress (18). Baseline levels of corticosterone, the corticosterone response to a novel cage, and hippocampal GR mRNA expression levels were assayed at the age of 3, 5 and 8 weeks (18). Basal corticosterone levels in early-weaned male mice, but not in early-weaned female mice, were higher than those in the normally-weaned mice. When the mice were subjected to the elevated plus-maze test, the early-weaned mice showed greater anxiety-like behaviour and secreted more corticosterone than the normally-weaned mice. In addition, GR expression in the early-weaned mice was higher at the age of 3 weeks, but lower at the age of 8 weeks, compared to normally-weaned mice (18). Decreased GR expression in the hippocampus lowers feedback inhibition by circulating glucocorticoids, resulting in higher corticosterone secretion after stress exposure, which is consistent with our endocrine data.
We also investigated the impact of early weaning on HPA activity immediately after weaning, which would be the key factor inducing behavioural changes in the early-weaned mice. When the mice were weaned on postnatal day (PD)14, both male and female mice showed higher corticosterone levels up to 48 h after weaning (32). By contrast, after normal weaning on PD21, corticosterone levels returned to the baseline within 1 h. The prolonged and significant increase in corticosterone levels is likely to involve behavioural as well as neurochemical changes induced by early weaning, which have been described below. Further studies are needed to clarify this issue.
Neural changes induced by early weaning
Early-life events affect not only the neuroendocrine system but also cognition (33), learning, memory (34) and vulnerability to drug abuse (35); for example, maternally-deprived mice showed robust and persistent responses to cocaine behavioural sensitisation (35). Therefore, we speculate that the whole brain structure is influenced by the social environment in which the neonate is reared. In abused human children, the hippocampus and corpus callosum are shrunken and the cerebral ventricle is enlarged (36), suggesting that early life events have a significant impact on the developmental structure of the brain, including the prefrontal cortex and the limbic system. Additionally, endogenous glucocorticoids have a critical role in signalling the initiation and enhancing the rate of myelin formation (37). We measured the levels of myelin basic proteins (MBPs), the major components of the myelin sheath, as an index of brain development. Expression of these MBPs increased gradually in normally-weaned mice. By contrast, in the early-weaned male mice, but not the early-weaned female mice, it increased robustly up to the third postnatal week and declined by the fifth postnatal week, compared to normally-weaned mice. These results suggest that early weaning influences not only anxiety-like behaviour, but also myelin formation in the brain during the developmental period, particularly between 3–5 weeks of age (16). Subsequently, we investigated the influence of early weaning on the structural development of the limbic system in mice by assessing the glycolipid, galactosylceramide, a specific myelin constituent in the brain. Different regions of the brain undergo different myelin formation processes (38), and anxiety-like behaviour is thought to be controlled by the limbic system, which includes the amygdala, prefrontal cortex and hippocampus. In the limbic system, the amygdala is strongly associated with the expression of anxiety-like behaviour. Lipid analysis of mice amygdala showed the early accumulation of galactosylceramide at 5 weeks in early-weaned male mice, but not in early-weaned female mice. The precocious accumulation of galactosylceramide in early-weaned male mice was observed only in the amygdala and not in the prefrontal cortex or hippocampus (39). Electron microscopy showed an increase in the number and decrease in the diameter of myelinated axons in the anterior part of the basolateral amygdala in early-weaned male mice at 5 weeks. These results confirmed that the higher anxiety-like behaviour in early-weaned male mice could be related to precocious myelin formation, particularly in the anterior part of the basolateral amygdala (39).
Brain derived neurotrophic factor and neurogenesis in the hippocampus
The candidate molecules that modulate brain development in the early-weaned mice are trophic factors that also regulate neurogenesis in the brain. Both brain-derived neurotrophic factor (BDNF) and neurogenesis are reported to be decreased by higher level of glucocorticoids. Thus far, mother–pup interactions have been known to affect neurotrophic factor expression; for example, maternally-deprived rats exhibit suppressed adult neurogenesis (40) and BDNF levels in the hippocampus (33). These neurochemical and morphological changes are correlated with psychopathogical symptoms, such as post-traumatic stress disorders and depression (41). With regard to neurogenesis, antidepressant treatments increase the proliferation of granule cells in the dentate gyrus with a delay paralleling the improvement of clinical symptoms (42). Chronic social stress, a key factor for mental illness, decreases adult neurogenesis and neural survival (43). Recently, Tsankova et al. (44) demonstrated that chronic social defeat decreases BDNF expression in the hippocampus through DNA modification and can be recovered with antidepressant treatment.
Early-weaned males, but not females, had less BDNF protein in the hippocampus at 3 weeks of age than the normally-weaned mice (32). Newly-grown cells were labelled with bromodeoxyuridine (BrdU) injections at 2, 3 or 5 weeks of age and were assayed at 3, 5 and 8 weeks of age, respectively. Early-weaned males had fewer BrdU-immunoreactive cells in the dentate gyrus at 3, 5 and 8 weeks of age. Contrastingly, in early-weaned females, fewer BrdU-positive cells were observed only at 5 weeks. Double-staining with BrdU and the neurone markers, NeuN and Tuj1, demonstrated that neurogenesis was lower in the early-weaned mice at 5 weeks of age. These results suggest that deprivation of the mother–infant interaction during the late lactating period increases corticosterone for 2 days and decreases BDNF synthesis and neurogenesis in the hippocampus; these stress responses are somewhat sexually dimorphic (i.e. males are more vulnerable to early weaning stress).
As mentioned above, early weaning changes glucocorticoid receptor expression in the hippocampus; this indicates that early weaning decreases neurogenesis and BDNF expression in the hippocampus via higher sensitivity to glucocorticoids.
Gender differences in the response to early weaning
There are several observations of gender differences in the anxiety-like behaviour and stress responses among maternally-deprived rodents (45, 46). Male rats and mice were more vulnerable to early weaning stress than females. Specifically, high anxiety-like behaviour in the elevated plus-maze test and augmented autonomic and HPA-responses in novelty tests of early-weaned animals were clearly observed in male rats, but not in females (16, 18, 19). Furthermore, early weaning has shown to significantly affect precocious myelin formation (38), BDNF expression and neurogenesis (32) in males. These data suggest that, among laboratory rodents, males are more vulnerable to early weaning stress than females. The effects of gonadal steroids on the sexual differences in response to early weaning are now under investigation.
The effect of the early maternal environment on the mammalian offspring is critical. Natural early weaning depends on the availability of environmental resources during the nursing period (i.e. if mothers rear their offspring under severe conditions, there is a trade-off between weaning earlier, reducing offspring survival, and increasing maternal survival) (9, 47). As a result of early weaning, the offspring become more sensitive to stressors and more anxious in novel environments. They may show marked territorial aggression if food resources are limited, have unstable social relationships, and obtain more resources in the group when they are adults. These behavioural changes are natural adaptations to the environment because the early-weaned animals lived in severe conditions, which their mother experienced during nursing. This review focuses on the early weaning of laboratory rodents and concludes that early weaning: (i) leads to persistent increased fearfulness and autonomic responses toward novelty stress; (ii) increases the neuroendocrine response to stress possibly due to a reduction in the number of hippocampal glucocorticoid receptors; and (iii) decreases the amount of maternal behaviour exhibited by the females, which is critical because of its non-genomic transmission to subsequent generations.
This study was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (#14760187 to T. K. and #15GS0306 to Y. M.), and The Promotion and Mutual Aid Corporation for Private Schools of Japan, Grant-in-Aid for Matching Fund Subsidy for Private Universities.