Activity-based anorexia (ABA) is a syndrome that can occur in humans and other animals, characterized by suppressed food intake, below-normal body weight, and hyperactivity (113, 114). Detailed descriptions of the ABA model can be found in Chaps. 16 and Chaps. 17 of this volume. In brief, in rats, ABA develops when animals are provided with a limited (1–2 h) daily period of food access and otherwise unlimited (22–23 h) access to a running wheel. Under these conditions, rats fail to consume enough calories during the once-daily meal to compensate for energy expended during wheel running. Body weight drops and running increases progressively across days, whereby reduced eating results in a vicious cycle that can lead to death by starvation (115, 116).

Few studies have examined early life environmental regulation of stress reactivity and later susceptibility to ABA. Recently, Carrera et al. (117) examined the effects of early postnatal handling of male and female rats on wheel-running rates, food intake, and weight loss in the ABA paradigm. Handling blunts stress reactivity (118) and should, therefore, make rats less susceptible to ABA-induced effects. Handling did not, however, affect any of these measures in young adolescent males. In adult females, handled runners required more days to reach the study removal criterion (i.e., a dangerously low body weight) than did animal facility-reared runners, but no differences were observed in weight loss and food intake. Hancock and Grant (119) examined the effects of prolonged periods of free wheel running, in combination with a restricted feeding schedule, on food consumption, weight loss, and running rates in male and female rats that experienced handling or MS repeatedly during PND 1–14. In comparison to handled rats, MS rats with 22-h daily access to a running wheel, in combination with a 1-h daily restricted feeding schedule, exhibited a faster rate of body weight loss, lower levels of food intake, greater daily increases in wheel running, and faster attainment of the removal criterion (119). In order to examine the differential effects of postnatal treatment on food intake and weight loss in the ABA paradigm depending on gender and developmental age, Hancock and Grant (120) used a milder version of the ABA paradigm, comprising 2-h daily running wheel access followed by 1-h food access. Handled and MS rats in both genders were tested either in adolescence or adulthood. Compared to handled females, MS females demonstrated greater increases in wheel running and a more pronounced running-induced suppression of food intake during adolescence, but not in adulthood. In contrast, it was only in adulthood that wheel running produced prolonged anorexic effects in MS more so than in handled males. These findings highlight the interplay between early postnatal treatment, gender, and developmental age on running, food intake, and rate of body weight loss in a milder version of the ABA paradigm.

Hyperactivity of the HPA axis is a marked feature of anorexia nervosa (121, 122). Hancock and Grant (119) hypothesized that the differential effects of postnatal treatment on food intake and weight loss during the ABA paradigm might result from differences in the HPA axis reactivity. Running-induced suppression of food intake and weight gain in rats is concurrent with increased expression of the hypothalamic corticotropin-releasing factor (123), and circulating adrenocorticotrophin (124) and corticosterone (124, 125). In MS rats, the HPA axis activity is already increased (1) and likely to be exacerbated by wheel-running stress. As such, increased weight loss and decreased food intake in MS rats subjected to the ABA paradigm may result from heightened release of stress hormones during running, which may account for the increased anorexic effect in these rats.

Interactions with peers during adolescence are thought to be of principal importance for social development in human adolescents, since individuals spend more time interacting with peers during adolescence than at any other developmental period (30, 31). In rats, early disruption of social interactions, such as isolated rearing, may affect brain development and produce profound, long-term neurochemical, endocrinological, and behavioral effects (38–40, 43, 126). Isolation-reared rats showed alterations in hippocampal neurotransmission (59), disregulation of the HPA axis activity responding to stress (55, 66), and anxiety-like behaviors (53, 55, 56). As described above, food intake and weight gain were significantly increased in our MS males, but not in NH, by isolation rearing (54). Increased serum cortisol levels are implicated in anxiety (70), depression (13), and binge eating disorders (19, 21). We have found that the basal plasma level of corticosterone was elevated in isolation-reared MS males compared with their counterpart NH (102), although it did not differ between group-housed MS and NH males (90). Thus, it is likely that increased food intake and/or body weight gain in our MS pups by isolated rearing may be related to a tonic increase in the plasma corticosterone level, which may worsen their psychoemotional behaviors.

Dysfunction of the HPA axis has been implicated in the pathogenesis of eating disorders (19–21). Symptoms of anxiety and depression are associated with the pathophysiology of eating disorders (71), especially with binge-like eating disorders (72, 73). The behavioral scores of our MS rats, such as ambulatory counts, rearings, defecation scores, immobility duration during the swim test, and the arm stays and entries of elevated plus maze test, did not seem to be further worsened by isolation rearing, i.e., those scores of single-housed MS rats did not differ from group-housed MS rats per se (54). However, further analyses revealed an interaction between MS and postweaning isolation in the time spent in both open and closed arms of the elevated plus maze test, suggesting an impact of postweaning isolation on anxiety-like behaviors in MS pups. In addition, they showed an interaction in defecation scores, representing emotional status (127–129), between MS and isolation rearing conditions. Thus, isolation-induced increases in food intake and weight gain observed in our MS males are likely to be related to their impacts on the psychoemotional behaviors representing anxiety.

However, isolation rearing did not affect food intake and weight gain in our MS females (unpublished observation), contrary to its significant effect observed in MS males. Plasma corticosterone level of isolated MS females showed a trend toward being increased compared to group-housed MS females. The HPA axis response to stress is known to be affected by gender, and the effects of gender or sex steroid hormones on the HPA axis vary with species and stressors. Currently, we do not have a clear explanation of the mechanisms underlying the differential gender effect of isolation rearing on food intake and weight gain in MS rats.

Glucocorticoids, stress hormones released by activation of the HPA axis, are known to be involved in the regulation of energy balance (130, 131). It has been demonstrated that food deprivation markedly elevates the plasma levels of corticosterone, representative of glucocorticoids in rodents (132–135). In adulthood of our MS model, food deprivation significantly elevated the plasma corticosterone levels of NH males, but not MS males, and the basal plasma levels of corticosterone did not differ between NH and MS rats (89). This result suggests that the responsiveness of the HPA axis to fasting, a stressful episode, is altered by the experience of neonatal MS in our model. It was reported that 3-h daily MS during PND 2–14 blunts the release of adrenocorticotrophic hormone (ACTH) responding to restraint stress in adult rats (10). Together with our result, this supports the conclusion that the experience of neonatal MS may blunt the HPA axis activation, stress response, of the offspring in adulthood. However, the patterns of plasma ACTH (5) and corticosterone (1) responses after restraint stress were not changed by the experience of 3-h daily of MS during the first 2 weeks of life, while foot-shock stress increased the plasma corticosterone level only in the MS group, but not in the NH group (136). MS effects on the characteristics of the HPA axis activation in offspring appear to vary depending on stressors. Also, response characteristics of the HPA axis to stressful stimuli at later ages are reported to vary depending on the timing of separation during the postnatal period (4).

Some of the central effects of glucocorticoids are believed to be mediated by hypothalamic neuropeptides (137, 138), and the hypothalamic mRNA expression (139–141) and release (142) of neuropeptide Y (NPY), a potent orexigenic peptide, increases during food deprivation. MS experience blunted the fasting-induced increase not only of plasma corticosterone, but also of the arcuate NPY mRNA expression in our MS males (89). There are some reports suggesting that glucocorticoids may control the hypothalamic NPY expression, and that the arcuate NPY neurons contain glucocorticoid receptors (143). NPY mRNA expression was increased by glucocorticoids in vitro (144, 145), and adrenalectomy down-regulated the hypothalamic NPY expression (146–148). Furthermore, elevated plasma corticosterone was necessary for fasting-induced increase of NPY mRNA expression in mice (133, 149). However, other studies have suggested that the hypothalamic NPY may influence the level of plasma glucocorticoids, i.e., NPY increases the mRNA expression (150), content and release (151) of corticotropin-releasing hormone in the hypothalamus, and the pituitary release of ACTH (152). Also, it was reported that fasting-induced increase of NPY expression requires neither an elevation of plasma corticosterone (153) nor the existence of endogenous glucocorticoids (148). Therefore, although the causal relation between plasma corticosterone and the hypothalamic NPY expression is controversial, reports support that the blunted NPY expression in our MS males during food deprivation may be related to the blunted response of plasma corticosterone involved with fasting.

In our female MS model, the plasma corticosterone increases responding to 48 h of food deprivation tended to be greater in MS than in NH females, although a significant difference was not found in statistics (112). Desbonnet and colleagues (154) have suggested that MS stress may result in a more reactive neuroendocrinological stress system in females than in males, i.e., acute swim stress increased the plasma corticosterone level in MS females but not in MS males, and the corticotropin-releasing factor response to stress was enhanced in MS females relative to males. This supports our findings that the HPA axis response to food deprivation is blunted in MS males (89), but not in MS females (112). Whereas most of the studies investigating long-lasting effects of MS have used male subjects, those that studied both males and females have revealed significant gender differences in the MS effects on psychoemotional behaviors (110, 155) and HPA axis status (156).

In the arcuate nucleus of our MS females, increased NPY expression and decreased pro-opiomelanocortin and cocaine- and amphetamine-regulated transcript expressions responding to 48 h of food deprivation were exaggerated compared to NH females (112), while the arcuate NPY expression responding to food deprivation was reduced in MS males compared with NH males (89). These results support the idea that MS stress may result in a more reactive neuroendocrinological stress system in females than in males (154). Male and female rats differ in numerous neuroendocrine and behavioral parameters, and vulnerability to stress is gender dependent (75–78). Thus, it is likely that MS experience may increase stress vulnerability in female rats and exaggerate the feeding peptides expression in the arcuate nucleus responding to a metabolic stress, food deprivation, but currently molecular mechanisms underlying the exaggerated response of the hypothalamic feeding peptides expression in MS females to food deprivation are not clear.

Hypothalamic NPY expression responding to food deprivation was blunted in our MS males in young adulthood, which was accompanied by a blunted response of plasma corticosterone (89). NPY potently stimulates food intake (157–159), and increased NPY expression in the hypothalamus appears to be implicated in the induction of hyperphagia (160, 161). Rats display compensatory hyperphagia when food is returned with ad libitum access following food deprivation. Thus, it is likely that compensatory hyperphagia following food deprivation may be diminished, or at least reduced, due to blunted increases of the hypothalamic NPY and/or the plasma corticosterone in our MS males.

In order to determine if MS rats show different feeding responses to fasting trials, we subjected the NH and MS male pups to 24 h of fasting and 24 h of refeeding ad libitum repeatedly during the adolescent period (90), especially because the symptoms of eating disorders mostly start with dieting and the incidence of eating disorders is higher among adolescents and youth (162). Not only NH pups, but also MS pups showed compensatory hyperphagia on PND 30 following the first fasting trial, with no difference in the amount of foods consumed (90). The arcuate NPY expression responding to fasting on PND 29 was blunted in our MS pups; however, the plasma corticosterone levels increased after the first fasting trial not only in the NH but also in the MS group, in accordance with previous reports that food deprivation elevates the plasma level of corticosterone (132–135). Adrenal glucocorticoids, corticosterone in rodents, have been implicated in the regulation of energy homeostasis (130, 131), and centrally administered glucocorticoids increase food intake and weight gain in rodents (138). Thus, it is concluded that elevated plasma corticosterone during the first fasting trial on PND 29 might have contributed to compensatory hyperphagia in both NH and MS pups. The findings also suggest that increased expression of the arcuate NPY is not necessary to induce compensatory hyperphagia following food deprivation, nor for the orexigenic action of corticosterone.

Compensatory hyperphagia responding to 24 h of food deprivation diminished after the second fasting/refeeding cycle in NH males; however, it persisted in MS males throughout the entire experimental period (90). It has been reported that repeated weight cycling (or fasting/refeeding cycling) may reduce metabolic rate and, consequently, reduce the need for energy intake (163–165). Thus, it seems that diminished hyperphagia in male NH pups on repeated fasting/refeeding cycles may be due to a reduced metabolic rate, and experience of neonatal maternal separation may affect the metabolic adaptation to repeated weight cycling.

Interestingly, both the arcuate NPY expression and the plasma corticosterone level were increased in satiated MS pups that still showed compensatory hyperphagia following the six sets of fasting/refeeding cycles, contrary to the fact that neither NPY nor plasma corticosterone increased in NH pups that did not show hyperphagia in the same condition (90). Plasma corticosterone is implicated in the hypothalamic NPY expression (133, 143, 146, 149). Thus, it is likely that a chronic increase of plasma corticosterone during the repeated fasting/refeeding cycles might have contributed to a tonic increase of NPY expression in MS rats, and the tonic increases in the plasma corticosterone and NPY expression in our MS rats on repeated fasting/refeeding cycles may partly be in charge of the sustained compensatory hyperphagia.

Repeated fasting/refeeding cycles resulted in a marked weight loss both in NH and MS males (90), in accordance with previous reports that 24 h of fasting and 24 h of refeeding cycles lead to a significant suppression in weight gain (163, 164, 166, 167). Interestingly, MS pups appeared to lose more weight during food deprivation, and gain more during refeeding, than NH pups (90), suggesting that experience of repeated MS during the preweaning period may lead to an exaggerated response to caloric challenges in weight gain of the offspring later in life. Notably, the increased weight loss of MS pups on each fasting day was less significant than the increased weight gain on each refeeding day; consequently, total weight loss after the six sets of fasting/refeeding cycle was significantly reduced in MS rats compared to NH rats (90). This appeared to be related to the increased feeding response (sustained hyperphagia) in MS rats during repeated fasting/refeeding cycles. Compared to NH males, the plasma corticosterone levels responding to either acute fasting or repeated fasting/refeeding cycles were further elevated in MS males (90). Thus, it is concluded that in our male MS model, experience of neonatal MS may lead to an exaggerated feeding response to repeated fasting/refeeding challenges at adolescence, possibly, due to increased responsiveness of the HPA axis.

Altered emotional and mood states, including depression and anxiety, affect eating behavior and food choice. Depression and anxiety can be linked to compulsive behaviors such as drug use and craving for palatable food, which induce feelings of pleasure (168, 169). Studies of humans showed that most subjects reported a preference for palatable food rich in fat and sugar during negative emotions (170). It is hypothesized that sustained hyperphagia during repeated fasting/refeeding cycles observed in our MS males may, at least partly, be related with their psychoemotional status. That is, depression- and anxiety-like behaviors were increased in MS males compared with NH males (54, 91), and repeated fasting/refeeding cycles, metabolic stress challenges, not only resulted in sustained hyperphagia but also improved depression-like behaviors of our MS males (unpublished observation). It was reported that rats consuming palatable foods after exposure to a stressor displayed reduced signs of stress (171, 172). Stress-stimulated consumption of palatable food is proposed as reward-based eating, which indirectly blunts the stress response (173).

Rearing rodents in persistent social isolation from weaning, to deprive them of social play, is a relevant paradigm for studying early life stress and produces a large array of consistent long-lasting neuroendocrinological and behavioral alterations compared with group-housed controls (41, 43, 126). The reported behavioral and neuroendocrinological effects of postweaning isolation in rats, which have been mostly studied in males, have strongly suggested its tentative impact on feeding behaviors; however, isolation rearing in male rats did not cause consistent alteration in body weight and food intake from age-matched controls (see Sect. 1.2). In particular, the effect of social isolation in adolescence on body weight and food intake of female rats has rarely been reported. We have demonstrated that adolescence social isolation may promote food intake and weight gain of female rats (74). The stress-induced elevation of the plasma corticosterone was blunted in isolated females, and furthermore, the basal plasma level of corticosterone was elevated in the isolates compared with group-caged ones. Increased serum glucocorticoids have been implicated in binge-like eating disorders (19, 21, 90), and central administration of glucocorticoids increased food intake and weight gain in rodents (138). Notably, isolated females showed a selective increase in cookie intake when they had access to cookies in addition to standard chow (74). This result is in accordance with previous studies showing that binge eating in animal models is evident in the selective increase in palatable food intake, resulting in an overall increase in caloric intake (213–215).

In our study, isolated young females showed hyperactivity in accordance with previous reports (75, 216, 217). Hyperactivity observed in isolation-reared female Hooded Lister rats was accompanied by strong and stable preferences for their most preferred food (216). Increased food intake, especially palatable food intake, in chronically stressed rats has been suggested to correlate with anxiety-like behaviors (218). Although the behavioral scores of isolated females during elevated plus maze test did not differ from group-housed ones, number of rears and repetitive standing with two forepaws up were increased in isolated females during the activity test (74). Increased rearing activity in rats has been reported to reveal an anxiety-related behavior responding to stress, as a proactive emotional coping behavior (219, 220). It has been suggested that increased food intake responding to stress is a stress coping behavior and consumption of palatable food dampens psychological and physiological responses to stress (221, 222). The stress-induced elevation of plasma corticosterone was blunted in our isolated females that ate more food than group-housed ones. Therefore, increased consumption of food in isolated females may be a stress coping behavior, likely in relation to anxiety-related behaviors, dampening psychological and physiological responses to chronic social isolation stress.

Loss of social contact and behavioral withdrawal are associated with the etiology of depression in men. However, little change was observed in immobility or struggling time during the Porsolt forced swim test in isolation-reared male rats (54, 223, 224). Women are almost twice as likely to suffer from depressive disorders as men (225). In several putative animal models of depression, female rather than male rats show a greater response to stressors (226). Teenage girls, in particular, report more stressful experiences than do teenage boys or older women and men (227, 228). Immobility duration in the Porsolt swim test was significantly increased in isolated females compared with group-housed controls (74), revealing that adolescence social isolation may increase depression-like behaviors in female rats. Association of binge-like eating disorders with symptoms of anxiety and depression has been reported (72, 73). Previous studies have suggested that increased serum cortisol levels are implicated in depression (13) and binge eating disorders (19, 21). Taken all together, it is concluded that social isolation in adolescence increases food intake and depression-like behaviors in female rats, and a tonic increase of the HPA axis activity responding to chronic isolation stress may play a role in its pathophysiology.

Male and female rats differ in numerous neuroendocrine and behavioral parameters, and vulnerability to stress is gender dependent (75–78). For example, compared to males, female rats are more active in the open arms of the elevated plus maze (229, 230) and the open field (75, 231), indicative of higher levels of general arousal and exploration, as well as decreased anxiety among females (76). The anxiogenic property of postweaning social isolation has been found in male rats with elevated plus maze test (54, 55), but not in female rats (55, 74). In male rats, isolation rearing did not increase depression-like behaviors (54, 223, 224), nor cause any consistent alterations in food intake (41, 53–55).

Women with anorexia and bulimia nervosa have been found to display high levels of social insecurity and social isolation (232). The consistent relationship between disordered eating behaviors and negative friendship qualities (friend alienation, friend conflict) was reported in adolescent girls (233), and depressed mood was associated with poor body image and disordered eating symptoms (233–235). Obese girls were more likely to report more serious emotional problems, hopelessness, and a suicide attempt, when compared to their normal weight peers (236). Depression and obesity were positively associated in females and the entire sample, but not in males (237). Here, we propose that social isolation in adolescence of female rats can be used as an animal model to study the pathophysiology of binge-like eating behaviors associated with symptoms of depression in young females.