Primary social relationships influence the development of the hypothalamic–pituitary–adrenal axis in the rat

Physiology & Behavior 73 (2001) 255 ± 260

Primary social relationships influence the development of the hypothalamic±pituitary±adrenal axis in the rat Seymour Levine* Department of Psychiatry, Center for Neuroscience, University of California-Davis, Davis, CA 95616, USA Received 27 September 2000; accepted 22 February 2001

Abstract In the developing rodent, there is a period from about days 4 to 14 when the adrenal response to stress is either minimal or non-existent. This has been designated as the stress hyporesponsive period (SHRP). Numerous experiments have demonstrated that maternal factors are critical for the regulation of the pup's hypothalamic ± pituitary ± adrenal (HPA) axis and the maintenance of the SHRP. Following 24 h of maternal deprivation, the neonatal rat shows elevated basal levels of corticosterone (CORT) and exhibits a robust CORT and ACTH response to mild stress. Further, c-fos mRNA in the paraventricular nucleus (PVN) is enhanced following stress in deprived pups. At least three aspects of maternal behavior play a role in the regulation of the HPA axis during development. Tactile stimulation appears capable of inhibiting most of the brain-related changes that occur following maternal deprivation. Feeding is essential for maintaining the adrenalunresponsive and reduces the sensitivity of the adrenal to ACTH. Passive contact suppresses the response to stress. In the adult, corticotropin-releasing hormone (CRH) is the major neuropeptide that controls pituitary ACTH secretion. In the maternally deprived pup, CRH gene transcription is downregulated and arginine vasopressin (AVP) appears to assume the major regulatory hormone that modulates ACTH. These data all indicate that maternal factors are responsible for actively inhibiting the endocrine responses to stress postnatally. Thus, during development, most of the peripheral and central stress-responsive systems are capable of being activated. However, under conditions of normal dam ± pup interactions, these responses are mostly suppressed by the dam's behavioral interaction with the pups. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Hypothalamus-pituitary-adrenal development; CRH; Vasopressin; Maternal regulation

1. Introduction In adult organisms, glucocorticoids, in particular corticosterone (CORT), serve a wide variety of regulatory and permissive functions [22]. These functions are aimed basically at the maintenance of basic metabolic processes and regulation of the organism's response to stress. Most of the effects of CORT in the adult are readily reversible. However, during development, administration of glucocorticoids has been demonstrated to have permanent effects on growth and differentiation in a number of systems, including the central nervous system (CNS). In rats, high doses of glucocorticoids administered to the neonate cause decreased mitosis, myelination altered neuromorphogenesis [2]. Animals treated neonatally with glucocorticoids show reduced

* Corresponding author. Tel.: +1-530-752-1887; fax: +1-530-757-8827. E-mail address: [email protected] (S. Levine).

DNA content and brain size, as well as impaired neuroendocrine function [9] and behavior [3] in adulthood. Conversely, the absence of CORT following adrenalectomy also affects CNS development. Thus, relatively constant low levels of CORT are essential for normal maturation. It appears that during development, a delicate and critical balance of CORT is essential. The dynamics of the hypothalamic ± pituitary ± adrenal (HPA) system shows a characteristic developmental pattern. During the final days of gestation and immediately after birth, basal levels of CORT in the rat are declining but are still high [21]. However, during the first two postnatal days, CORT concentrations decrease dramatically and remain at low levels until approximately day 14. In addition, it has also been reported that concentrations of pituitary adrenocorticotropin (ACTH) and hypothalamic corticotropin-releasing hormone (CRH) are also low during the first two postnatal weeks [37]. Further, during this period, stimuli, which normally elicit CORT elevations in adults, are unable to do

0031-9384/01/$ ± see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 0 3 1 - 9 3 8 4 ( 0 1 ) 0 0 4 9 6 - 6

256

S. Levine / Physiology & Behavior 73 (2001) 255±260

so in the infant. This neonatal period, from postnatal days 4 to 14, has been referred to as the ``stress hyporesponsive period'' (SHRP). During the SHRP, there is also a diminution of adrenal sensitivity, evidenced by the fact that the infant fails to show a significant elevation of CORT following administration of high doses of exogenous ACTH [18]. The HPA system does not develop uniformly; different components of the system have different ontogenetic patterns. Thus, although the infant rat is capable of responding to certain stressful stimuli by day 14, circadian rhythms of CORT and negative feedback regulation are not fully developed until much later in development [10,20,36]. The overall picture that emerges, however, is that there are multiple differences in the neonatal neuroendocrine system, which ensure that plasma CORT levels in the developing infant are maintained at relatively low and unperturbable levels. 2. Maternal influences on development There is a growing body of evidence which indicates that the mother relates physiological processes in the developing infant in subtle ways [13,14]. Studies, mostly on developing rodents, have demonstrated that the mother regulates physiological responses in the infant as diverse as heart rate, sleep/wake cycles and growth hormone production. Maternal regulation is manifested by separating mother and infant. Specific physiological changes in the infant, which develop slowly over a relatively protracted period of separation, can be tightly linked to specific features of mother ± infant interaction. For example, when 2-week-old rat pups are separated from their mothers for 24 h, there is a 40% decrease in cardiac rate. This decrease results specifically from the absence of mother's milk, as opposed to lack of maternal contact or other aspects of maternal care. Thus, when milk is infused into the pup directly, in the absence of the mother, normal cardiac rate can be maintained or restored. Another physiological response of the infant rat pup to separation from the mother is a decline in growth hormone secretion and in levels of a tissue enzyme regulating growth and differentiation: omithine decarboxylase (ODC). This decline can be traced precisely to the lack of vigorous tactile stimulation from the mother. Vigorous stroking of a separated rat pup by an experimenter with a moistened artist brush can normalize growth hormone and ODC activity [16,28]. The initial studies, which suggested the possibility that maternal factors may be involved in the regulation of the HPA axis during development, were designed to investigate an entirely different question: whether the infant rat is capable of using consummatory behavior to suppress HPA activity as was demonstrated in the adult. Thus, rats at postnatal days 12, 16 and 20 were maternally deprived for 24 h and then exposed to a novel environment [31]. The two variables that were manipulated were the presence or absence of the mother and the presence or absence of food. The animals were fed

while being exposed to a novel environment through a previously implanted tongue cannula. The results indicated that a brief feeding bout was not capable of suppressing the CORT response following exposure to novelty. However, the presence of an anesthetized mother completely inhibited the CORT response. There was an unexpected finding in this study, namely, 12-day-old pups also showed significant elevation of both basal and stress-induced CORT secretion. This finding was surprising in view of all the evidence, which indicated that at this age, the rat pup is in the midst of the SHRP and therefore should not be exhibiting a CORT response. Although the purpose of the deprivation procedure was to increase the level of food-motivated behavior, so that the infants would consume the infused milk, the results indicated that maternal factors might be critical in the regulation of the neonatal HPA axis. Subsequent experiments revealed that maternal deprivation produces a stress-responsive HPA system during that period in development when non-deprived infants are non-responsive. These data suggested that the HPA axis was yet another physiological system that was influenced by maternal behavior. 3. Maternal influences on the developing adrenal Hofer [13] has designated the process by which maternal factors regulate physiological development as ``hidden regulators.'' One of the criteria for a hidden regulator is that it requires a long period of maternal deprivation to become evident, in contrast to the immediate responses to separation. An examination [19] of the time course of the CORT response following deprivation reveals that the maternal influences on the developmental patterns of this system do indeed meet this criterion established for hidden regulators. Evidence for this conclusion comes from a study in which pups were maternally deprived for varying lengths of time (i.e., 0, 2, 4, 8 and 24 h). At the end of each deprivation period, CORT secretion in response to stress (novelty or novelty plus a saline injection and ACTH) was measured. Basal levels of CORT increased progressively over time in days 7 and 11 (but not day 3) pups. CORT release in response to stress followed a similar pattern. Maternal deprivation at all ages tested resulted in a striking increase of circulating CORT following injection with ACTH. Thus, for animals that are in the midst of the SHRP, prolonged maternal deprivation resulted in a disinhibition of the normal pattern of HPA activity, which characterizes the neonatal rat. In view of the evidence, which convincingly demonstrated that maternal factors are important in the regulation of the neonatal HPA axis, we began to focus on the issue of what specific aspects of the mother ± infant interaction were responsible for the non-responsiveness characteristic of this system. In our early studies, [33] we found that contact with a lactating dam suppressed the CORT response in maternally deprived pups that were either exposed to novelty or injected with saline. A lactating dam was the most effective

S. Levine / Physiology & Behavior 73 (2001) 255±260

stimulus for producing this suppression. The presence of a non-lactating female, a littermate or an adult male was not capable of suppressing the CORT response [32]. Thus, maternal contact has been shown to regulate at least one aspect of the pup's HPA responsiveness during the SHRP. An experiment was designed [4] to investigate whether prolonged 24 h non-nutritional contact with a maternally behaving dam would also inhibit the stress response observed in maternally deprived pups. Seven-, 11- and 15day-old pups were left with a dam that exhibited maternal care but was unable to lactate (i.e., thelectomized, nipples removed) for 24 h, or were maintained with their lactating dams. At the end of this period, basal and stress (novelty) levels of CORT were compared between different treatments. The results provided little evidence that maternal contact in the absence of sucking and/or feeding was capable of downregulating the infant's HPA system. Pups that remained in contact with thelectomized darns showed CORT levels equivalent to deprived infants (with the exception of a small but significant modulation of the CORT response to novelty at day 12). In all cases, however, these levels were still higher than those of the non-deprived pups. Further studies [26] using different techniques substantiated the conclusion that maternal contact in the absence of feeding did not result in a reduction of plasma CORT levels. These studies are inconsistent with the report that CORT levels following 2 h of maternal separation are partially reversed when the pup is allowed to contact a nipplecauterized dam [17]. These findings did suggest that the processes responsible for maintaining the SHRP differ from those that modulate the stress response in an infant that has been rendered responsive by 24 h of maternal deprivation. Thus, we can discriminate two seemingly independent events: (1) an inhibitory effect that maintains hyporesponsiveness and (2) a suppressing effect that prevents the CORT response from being activated once the pup is capable of responding. It is generally assumed that elevations in CORT are under the direct regulation of ACTH. However, there is evidence both in adults [5] and in infants [11] that adrenal sensitivity to ACTH can vary. For example, in adult animals, adrenal sensitivity varies as a function of time of day; the adrenal is more sensitive to ACTH in the morning. In the rat pup, there is evidence that one of the indicators of the SHRP is a reduction in adrenal sensitivity to ACTH. Several studies have shown that following maternal deprivation, CORT levels in the adrenal and in plasma are significantly elevated in response to exogenous ACTH. In contrast, non-deprived pups show little or no CORT response when exposed to high levels of ACTH at all ages tested (days 4, 8, 12 and 16). At day 4, it was demonstrated [26] that maternally deprived pups not only elicit a greater CORT response to equivalent amounts of exogenous ACTH, but also much lower doses are able to evoke a CORT response. Thus, extremely low levels of ACTH (in the order of magnitude of 0.001 IU of ACTH) are capable of inducing a CORT response in these

257

deprived pups, whereas very high concentrations of ACTH are required to elicit any response in non-deprived pups. 4. Pituitary ACTH If, as we have indicated, maternal deprivation produces a highly sensitive adrenal, it could be concluded that the absence of feeding directly affects the adrenal and that other components of the HPA axis still remained hyporesponsive. Thus, experiments were initiated to directly examine the ACTH response to a saline injection and novelty in both maternally deprived and non-deprived pups at days 6, 9 and 12. An initial blood sample was obtained after 24-h deprivation and at 10, 20 and 30 min following the saline injection. The data [34] demonstrate that the ACTH response to the mild stress used in these experiments was robust in deprived pups, although, unlike CORT basal levels, was not elevated. In addition to demonstrating that maternally deprived pups showed an elevation of ACTH following exposure to mild stress, the data also indicated that once the system is activated, there appeared to be a deficiency in the negative feedback mechanism. In adult animals, the ACTH response to a saline injection returns to basal values by 30 min. At days 9 and 12, ACTH remains elevated for at least 2 h following an acute saline injection. At day 15, ACTH is still elevated in maternally deprived pups, but appears to be slowly returning to basal levels. Given this body of data, we returned to the question of the possible inhibitory role of feeding on the pituitary response to stress. In order to accomplish this experiment, the use of a technique for feeding the pup in the absence of the dam was required. Such a technique was developed initially by Hall and Rosenblatt [12] and involved implanting a cannula through which milk could be delivered at regular intervals throughout the 24 h of maternal deprivation. For this study, we used the procedure that was reported by Spear et al. [30]. Milk (0.24 ±0.45 ml) was delivered hourly by an infusion pump attached to the cannula. The amount of milk delivered was calibrated to maintain the normal weight gain that was seen in the non-deprived pups. One of the more pervasive maternal behaviors is anogenital licking. The primary purpose of this behavior is to stimulate the eliminative processes in the pup. In order to prevent accumulation of waste products in the infant, the anogenital area of the pups was stimulated manually by the experimenter on three different occasions separated by 7 ±8 h. This stimulation was also performed on non-fed deprived pups. At the end of 24 h, blood samples were obtained from fed and non-fed pups at days 9, 12 and 15. For half the animals, blood was taken immediately following the 24-h period of maternal separation (non-treated, NT). The remaining infants were injected with saline and blood was taken at 30 min post-injection. The results [25,34] showed that non-fed NT deprived pups had elevated CORT levels when compared to the fed NT deprived animals. Thus,

258

S. Levine / Physiology & Behavior 73 (2001) 255±260

feeding reduced the deprivation-induced elevations in basal levels. This effect was present at days 9 and 12. Although the saline injection produced elevations in CORT above those observed in NT pups, feeding significantly reduced the CORT response at all ages tested. However, the ACTH response did not appear to be affected by feeding. At all ages, the ACTH levels elevated in response to the saline injection and were not significantly different between the fed and non-fed deprived infants. When we examined the results obtained in the experiment described above with those presented in the initial study which investigated the ACTH response in maternally deprived pups, it was evident that the magnitude of the ACTH response was significantly lower than that we had obtained in the previous experiment. The only change in the procedure between the two experiments was that in the feeding experiment, all animals had been stroked three times during the 24-h deprivation period. We therefore specifically investigated [34] the role of stroking in the regulation of the neonatal HPA axis. CORT and ACTH secretion was measured in four groups of pups at two ages (days 9 and 12). Three groups of pups were maternally deprived. One group received no treatment during the deprivation period (DEP), the second group was stroked in the anogenital region using a fine brush (DEP + STROKED), the third group was picked up at the same time as the stroked group but received no stroking (DEP + HANDLED). These animals were compared to a non-deprived group (NDEP). Basal and stress levels of CORT and ACTH were measured. None of the treatments elevated basal levels of ACTH. Stroking reduced the ACTH response to stress to the levels of the non-deprived pups. Handling did not prevent the elevation of ACTH levels induced by deprivation at day 9. At day 12, this manipulation led to a further increase in ACTH secretion. The CORT results, once again, reflected the absence of feeding. Thus, although there were no differences between the non-deprived and stroked infants in ACTH levels, both basal and stress levels of CORT were significantly elevated in the stroked pups, indicating that adrenal sensitivity to same endogenous levels of ACTH was increased. These findings help resolve the paradox created by earlier findings, which seemed to suggest that contact with a non-lactating dam was not effective in preventing activation of the HPA system in non-fed pups. Those findings were based solely on the CORT response. Insofar as different properties of maternal behavior appear to differentially regulate specific components of the developing HPA axis, it is likely that if ACTH was measured during 24 h of contact with the non-lactating dam, we would have observed a suppression of the ACTH response. 5. Hypothalamus Ð CRH and arginine vasopressin (AVP) Insofar it has been demonstrated that the secretion of ACTH was significantly increased following maternal dep-

rivation, it was logical to assume that the central components of the stress response should also reflect the effects of maternal deprivation. One of the features of the SHRP has been the consistent finding that CRH gene expression is difficult to elicit in the neonate [26]. It has been postulated that the cellular regulatory mechanisms for CRH biosynthesis may be deficient early in development. However, from the data presented thus far, it could also be argued that the specific aspects of the dam's behavior actively inhibit CRH biosynthesis. In our initial studies [29], CRH gene expression was examined in maternally deprived pups. There was no evidence of increased basal levels of CRH message. Surprisingly, in 12-day-old deprived pups, basal levels of CRH mRNA were significantly reduced. Further, 2 h following stress, there was no evidence of any changes in CRH message. However, there were other indications that the response of the brain to stress was enhanced as a consequence of maternal deprivation. Increases in the immediate early genes (IEG) c-fos and NGF-1a in the paraventricular nucleus (PVN) were much higher in deprived pups. The downregulation of CRH gene expression and the increase in PVN IEGs could be completely reversed by stroking the anogenital region of the pups on three occasions during the 24-h deprivation period [35]. Although resting levels of CRH mRNA appear to be suppressed following maternal deprivation and there were no changes in CRH gene expression following stress, we once again tested the hypothesis that maternally deprived pups would show an increase in CRH gene transcription. Given that deprived pups show a significant response to mild stress, we hypothesized that a rapid increase in the primary CRH transcript would be detected in deprived pups. Previous work [15] has demonstrated that by using intronic probe technology, stress-induced CRH hnRNA has been detected when no changes in mRNA levels were observed. Examining CRH hnRNA should therefore provide a more direct measure of rapid stimulus-induced transcriptional activation in the neonate. In the following experiment, both CRH hnRNA and CRH mRNA were examined in both non-deprived and deprived pups at 6, 12 and 18 days of age in response to an injection of isotonic saline. As expected, there were significant increases in both ACTH and CORT in deprived pups. However, both CRH hnRNA and CRH mRNA were significantly reduced in deprived pups at all ages. Another unexpected outcome of this study was how rapidly CRH gene transcription occurs in the neonate. Within 15 min of the injection, there was a marked increase in CRH message in non-deprived pups. This rapidity of this response to a very mild perturbation was very surprising. Although CRH hnRNA has been shown to rise within 15 min following stress, CRH mRNA has not been reported to occur earlier than 1 h, and more commonly it takes from 2 to 3 h before CRH gene transcription begins to increase, and only under conditions where the stimulus is more severe. These data indicate that during ontogeny, the cellular mechanisms controlling CRH gene expression are unique to this period in

S. Levine / Physiology & Behavior 73 (2001) 255±260

development and respond rapidly to a mild challenge. In this regard, the hypothalamic level of the HPA axis may in fact be hyperresponsive rather than hyporesponsive in the neonate [7]. Further, maternal deprivation appears to reverse this pattern of rapid CRH gene expression and maternally deprived pups show a reduction in CRH expression. In the experiments described above it, was demonstrated that although non-deprived pups fail to elevate the peripheral hormones ACTH and CORT in response to stress, there is evidence for neural activity (c-fos) and rapid CRH gene expression. In contrast, maternally deprived pups exhibit elevations in the secretion of ACTH and CORT but have a marked reduction in CRH gene expression. This raised the question of whether CRH is the major ACTH secretagogue during development. In the adult rat, AVP exerts a synergistic action on ACTH release [23,24]. Further, under some conditions, there is evidence that AVP may become the predominant ACTH secretagogue. Thus, when exposed to chronic stress, the pattern of secretagogue release is altered, whereas CRH message return to basal levels AVP expression is enhanced [1,6,27]. Based on the assumption that 24 h of maternal deprivation may represent a chronic stress for the pup, it was hypothesized that when a stimulus evokes ACTH secretion during the SHRP, the concerted action of CRH and AVP is essential to result in the peripheral hormone response to stress. To address this question, nondeprived and deprived pups were restrained in small plastic tubes and a time course of CRH and AVP gene expression was examined [8]. The time course of ACTH and CORT secretion was also obtained. Early in development, ACTH and CORT were elevated only in deprived pups. In this experiment, basal levels of CRH and AVP mRNA were not affected by maternal deprivation. Whereas CRH message in non-deprived pups, presumably in the SHRP, was elevated in a manner similar to that observed in response to a saline injection, only in deprived pups, which showed a rise in ACTH, was there an increase in AVP gene expression. In this study, on all occasions when ACTH was elevated, there was a concomitant increase in AVP gene expression. Thus, maternal factors appear to play a major role in determining the pattern of ACTH-releasing factors during development. Therefore, by inhibiting the expression of AVP, the peripheral hormones secreted during stress are also inhibited. It has been demonstrated that many of the neural components of the HPA axis that are affected by maternal deprivation can be reversed primarily by anogenital stroking. It remains to be determined if this aspect of maternal regulation can also reverse the pattern of secretagogue release that appears to be responsible for the increases in ACTH secretion seen in maternally deprived pups. 6. Conclusion The body of evidence that has been presented in this review clearly supports the hypothesis that the HPA axis of

259

the developing rat is, in part, regulated by maternal factors. However, different aspects of the dam's behavior appear to regulate different aspects of the HPA system. Feeding is responsible for the downregulation of the neonate's adrenal and renders the adrenal insensitive to its tropic hormone, ACTH. In contrast, active licking and grooming suppress neural activation and alter the pattern of ACTH-releasing factors. Passive contact and possibly non-nutritive sucking inhibit the stress response to novelty. Although it is possible to experimentally segregate these different components of the dam's behavior, it is more likely that all three of these regulatory processes are acting in concert to limit and prevent the so-called stress hormones in general and more specifically CORT from exceeding some optimal level. It would appear that downregulating the ACTH and CORT responses is sufficiently critical that there are multiple failsafe mechanisms that ensure that these hormones will not respond or respond minimally.

Acknowledgments This research was supported by grants HD-02881 from the National Institute of Child Health and Human Development, MH-45006 from the National Institute of Mental Health (NIMH) and U.S. Public Health Service Research Scientist Award MH-19936 from NIMH to the author.

References [1] Bartanusz V, Jezova D, Bertini L, Tilder F, Aubry J-M, Kiss J. Stressinduced increase in vasopressin and corticotropin-releasing factor expression in hypophysiotropic paraventricular neurons. Endocrinology 1993;132:895 ± 902. [2] Bohn MC. Granule cell genesis in the hippocampus of rats treated neonatally with hydrocortisone. Neuroscience 1980;5:2003 ± 12. [3] Bohn MC. Glucocorticoid-induced teratologies of the nervous system. In: Yanai J, editor. Neurobehavioral teratology. New York: Elsevier, 1984. pp. 365 ± 87. [4] Ciruili F, Gottlieb SL, Rosenfeld P, Levine S. Maternal factors regulate stress responsiveness in the neonatal rat. Psychobiology 1992; 20(2):143 ± 52. [5] Dallman MF, Akana SF, Cascio CS, Darlington DN, Jacobson L, Levin N. Regulation of ACTH secretion: variations on a theme of B. Recent Prog Horm Res 1987;43:113 ± 73. [6] DeGoeji D, Binnekade R, Tilders F. Chronic stress enhances vasopressin but not corticotropin-releasing factor secretion during hypoglycemia. Am J Physiol 1992;263:E394 ± 9. [7] Dent GW, Smith MA, Levine S. Rapid induction of corticotropinreleasing hormone gene transcription in the paraventricular nucleus of the developing rat. Endocrinology 2000;141:1593 ± 8. [8] Dent GW, Okomoto DK, Smith MA, Levine S. Stress-induced alterations in corticotropin-releasing hormone and vasopressin gene expression in the paraventricular nucleus during ontogeny. Neuroendocrinology 2000;71:333 ± 42. [9] Erkine MS, Geller E, Yuwiler A. Effects of neonatal hydrocortisone treatment on pituitary and adrenocortical responses to stress in young rats. Neuroendocrinology 1979;29:191 ± 9. [10] Goldman L, Vinget C, Hollingshead GW, Levine S. Postweaning

260

[11]

[12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24]

S. Levine / Physiology & Behavior 73 (2001) 255±260 development of negative feedback in the pituitary ± adrenal system of the rat. Neuroendocrinology 1973;12:199 ± 211. Grino M, Burgundor JM, Eska RL, Eiden LE. Onset of glucocorticoid responsiveness of anterior pituitary corticotrophs during development is scheduled by corticotropin-releasing factor. Endocrinology 1989;124:2686 ± 92. Hall WG, Rosenblatt JS. Suckling behavior and intake control in the developing rat pup. J Comp Physiol Psychol 1997;91:1232 ± 47. Hofer MA. Hidden regulatory processes in early social relationships. In: Bateson PPG, Klopfer PH, editors. Perspectives in ethology: social behavior. New York: Plenum, 1978. pp. 135 ± 66. Hofer MA. Early stages in the organization of cardiovascular control. Proc Soc Exp Biol Med 1984;175:147 ± 57. Kovaks KJ, Sawchenko PE. Sequences of stress-induced alterations in indices of synaptic and transcriptional activation in parvocellular neurosecretory neurons. J Neurosci 1996;16:262 ± 73. Kuhn CM, Butler SR, Schanberg SM. Selective depression of serum growth hormone during maternal deprivation in rat pups. Science 1978;201:1034 ± 6. Kuhn CM, Pauk J, Schanberg SM. Endocrine responses to mother ± infant separation in developing rats. Dev Psychobiol 1990;23: 395 ± 410. Levine S, Glick D, Nakane PK. Adrenal and plasma corticosterone and vitamin rat adrenal glands during postnatal development. Endocrinology 1967;80:1177 ± 9. Levine S, Huchton DM, Wiener SG, Rosenfeld P. Time course of the effect of maternal deprivation on the hypothalamic ± pituitary ± adrenal axis in the infant rat. Dev Psychobiol 1991;24(8):547 ± 58. Levin R, Levine S. Development of circadian periodicity in base and stress levels of corticosterone. Am J Physiol 1975;229:1397 ± 9. Martin CE, Cake MH, Hartmann PE, Cook IF. Relationship between foetal corticosteroids, maternal progesterone and parturition in the rat. Acta Endocrinol 1977;84:167 ± 76. Munck A, Guyre PM, Holbrook NJ. Physiological functions of glucocorticoids in stress and their relationship to pharmacological actions. Endocr Rev 1984;5(1):25 ± 44. Rivier C, Vale W. Interaction of corticotropin-releasing factor and arginine vasopressin on adrenocorticotropin secretion in vivo. Endocrinology 1983;113:939 ± 42. Rivier C, Rivier J, Mormede P, Vale W. Studies of the nature of the interaction between vasopressin and corticotropin-releasing factor in adrenocorticotropin release in the rat. Endocrinology 1984;115: 882 ± 6.

[25] Rosenfeld P, Ekstrand J, Olson E, Suchecki D, Levine S. Maternal regulation of adrenocortical activity in the infant rat: effects of feeding. Dev Psychobiol 1993;26(5):261 ± 77. [26] Rosenfeld P, Suchecki D, Levine S. Multifactorial regulation of the hypothalamic ± pituitary ± adrenal axis during development. Neurosci Biobehav Rev 1992;16:553 ± 68. [27] Scaccianoce S, Muscolo L, Cigliana G, Navarra D, Nicolai R, Angelucci L. Evidence for the specific role of vasopressin in sustaining pituitary ± adrenocortical stress response in the rat. Endocrinology 1991;128:3138 ± 43. [28] Schanberg SM, Evoniuk G, Kuhn CM. Tactile and nutritional aspects of maternal care: specific regulators of neuroendocrine function and cellular development. Proc Soc Exp Biol Med 1984;175:135 ± 46. [29] Smith MA, Kim S-Y, Van Oers HJJ, Levine S. Maternal deprivation and stress-induced immediate early genes in the infant rat brain. Endocrinology 1997;138:4622 ± 8. [30] Spear LP, Specht SM, Kirstein CL, Kuhn CM. Anterior and posterior, but not cheek, intraoral cannulation procedures elevate serum corticosterone levels in neonatal rat pups. Dev Psychobiol 1989;22:401 ± 11. [31] Stanton ME, Wallstrom J, Levine S. Maternal contact inhibits pituitary ± adrenal stress response in preweaning rats. Dev Psychobiol 1987;20:131 ± 45. [32] Stanton ME, Levine S. Maternal modulation of infant glucocorticoid stress responses: role of age and maternal deprivation. Psychobiology 1988;16(3):223 ± 8. [33] Stanton ME, Levine S. Inhibition of infant glucocorticoid stress response: specific role of maternal cues. Dev Psychobiol 1990;23: 411 ± 26. [34] Suchecki D, Rosenfeld P, Levine S. Maternal regulation of the hypothalamic ± pituitary ± adrenal axis in the infant rat: the role of feeding and stroking. Dev Brain Res 1993;75:192 ± 5. [35] Van Oers HJJ, DeKloet ER, Whelan T, Levine S. Maternal deprivation effect on the infants neural stress markers is reversed by tactile stimulation but not by suppressing corticosterone. J Neurosci 1998;18: 10171 ± 9. [36] Vazquez DM, Akhil H. Pituitary ± adrenal response to ether vapor in the weanling animal: characterization of the inhibitory effect on glucocorticoids on adrenocorticoid secretion. Pediatr Res 1993;34: 646 ± 53. [37] Walker CD, Perrin M, Vale W, Rivier C. Ontogeny of the stress response in the rat: role of the pituitary and the hypothalamus. Endocrinology 1986;118:1445 ± 51.