The newborn rat's stress system readily habituates to repeated and prolonged maternal separation, while continuing to respond to stressors in context dependent fashion

Hormones and Behavior 60 (2011) 165–176

Contents lists available at ScienceDirect

Hormones and Behavior

j ourna l homepage: www.e lsev ie r.com/ locate /yhbeh

The newborn rat's stress system readily habituates to repeated and prolongedmaternal separation, while continuing to respond to stressors in contextdependent fashion

Nikolaos P. Daskalakis a,⁎, Sanne E.F. Claessens a, Jasper J.L. Laboyrie a, Leo Enthoven a, Melly S. Oitzl a,Danielle L. Champagne a,b, E. Ronald de Kloet a

a Division of Medical Pharmacology, Leiden/Amsterdam Center for Drug Research, Leiden University Medical Center, Leiden University, Gorlaeus Laboratories, PO Box 9502,Leiden University, 2300 RA Leiden, The Netherlandsb Institute of Biology Leiden (IBL), Department of Integrative Zoology, Sylvius Laboratory, P.O. Box 9505, 2300 RA, Leiden University, Leiden, The Netherlands

⁎ Corresponding author at: Medical Pharmacology, LALeiden, PO Box 9502, 2300 RA, Leiden, The Netherlands

E-mail address: [email protected] (N.

0018-506X/$ – see front matter © 2011 Elsevier Inc. Aldoi:10.1016/j.yhbeh.2011.04.003

a b s t r a c t

Article history:Received 15 February 2011Revised 9 April 2011Accepted 17 April 2011Available online 3 May 2011

Keywords:Maternal separationCorticosteroneACTHSHRPNoveltyEnvironmentGenotypeMaternal careTyrosine hydroxylaseMelanocortin receptor-2

Adrenal corticosterone secretion of newborn mice rapidly desensitizes to repeated maternal absence. Thepresent study investigated the effects of novelty exposure, maternal care and genotype on this phenomenon.Maternal separation (MS) took place on postnatal days (pnd) 3–5. In Wistar rats, the degree of novelty in theMS-environment was varied by exposing pups to: (i) “home separation”: pups remained in the home cage;(ii) “novel separation”: pups were placed individually in a novel cage. Maternal care was recorded on pnd 1 to4. To investigate the effect of genotype, we also examined Long Evans in the “home separation” condition.Basal and stress-induced ACTH and corticosterone levels were measured. Adrenal tyrosine hydroxylase (TH)and melanocortin receptor-2 (MCR-2) proteins served as markers for adrenal function.We show, in both rat strains, that the rise in plasma corticosterone induced by a single 8 h-MS on pnd 5 wasabolished, when this separation procedure had also been performed on pnd 3 and 4. Habituation to maternalabsence occurred irrespective of housing conditions. However, pups in the “home separation” conditionreceived less maternal care upon reunion than those placed in the “novel separation”. These “homeseparation” pups appeared more responsive to a subsequent acute novelty-stressor, and their adrenal TH andMCR-2 were higher. Long Evans rats appeared more stress responsive than the Wistars, in the homeseparation condition.In conclusion, separation environment, maternal care and genotype do not affect adrenal desensitization torepeated 8 h-MS itself, but may modulate the adrenal stress-responsiveness of separated pups.

CDR, Einsteinweg 55, 2333 CC,. Fax: +31 71 527 4715.P. Daskalakis).

l rights reserved.

© 2011 Elsevier Inc. All rights reserved.

Introduction

Aberrant HPA axis activity and corticosterone (CORT) secretioninduced by adverse early life experiences is considered a major riskfactor for the development of psychiatric disorders in humans (Heimet al., 2008; Lupien et al., 2009). Therefore, rodents deprived as pupsfrom maternal care have been widely used as a laboratory model forearly adversity to study the underlyingmechanism of CORT-enhancedvulnerabilities (Plotsky et al., 2005; Pryce et al., 2001a). Severaladversity paradigms are currently used. One common approach is asingle episode of 24 h of maternal absence. Another approach isrepeated daily maternal separations (MS), which include repeatedperiods of 3–8 h absence of the dam during the first two postnatalweeks. It has been proposed that pups experiencing repeated MS

display as adult enhanced stress responsiveness, increased anxiety,helplessness and anhedonia, deficits of sensorimotor gating andincreased propensity for the intake of addictive drugs (Biagini et al.,1998; Brake et al., 2004; Moffett et al., 2007; Plotsky et al., 2005;Zhang et al., 2005). Interestingly, these rodent behaviors, pro-grammed by early adversity, resemble clinical endophenotypes ofdepression and schizophrenia, hence providing face and constructvalidity to these models (Pryce and Seifritz, 2011). However, theoutcome of early adversity depends on strain and gender of theanimals as well as the frequency, duration, age and time point (withinthe light cycle) of MS (Claessens et al., 2011; Lehmann and Feldon,2000). Moreover, the environmental context (housing in groups or inisolation, inside the home cage or in a novel environment) and theambient temperature have an effect (Lehmann and Feldon, 2000;Ruedi-Bettschen et al., 2004).

Cumulatively, these observations have raised the question to whatextent the HPA axis is actually activated during repeated MS. Schmidtand colleagues (Schmidt et al., 2004; Schmidt et al., 2006) found that,

166 N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

in mice, after 8 h of a single MS the basal level of circulating CORT hasslowly reached peak levels, while the stress hyporesponsive period(SHRP) has become disrupted resulting in an enhanced responsive-ness of the adrenocortical secretion of CORT to mild stressors andexogenous ACTH administration (Levine et al., 1991; Rosenfeld et al.,1992b). Enthoven and colleagues hypothesized that 3 consecutivedaily 8 h-MS from postnatal (pnd) 3–5 would amplify neuroendocrineresponses to both separation and novelty exposure. Surprisingly,they reported instead that the increase in HPA axis basal activity thatis first observed after the initial 8 h-MS was rapidly abolished afterrepeated 8 h-MS. This rapid desensitization of the neonate's HPA axisto repeated daily separations from the dam was not due to metabolicfactors such as ghrelin, and also did not occur because of enhancedglucocorticoid negative feedback (Enthoven et al., 2008). In spite ofthis rapid desensitization of the HPA axis to the effect of dailyseparation, the SHRP remained disturbed because a subsequentnovelty stressor triggered an enhanced plasma CORT and c-fosmRNA response in the PVN (Enthoven et al., 2008). This findingraised the question whether the environmental context experiencedby the pup during MS can influence the outcome (Enthoven et al.,2008).

In the present study we have extended these findings to the rat byexamining the immediate effects of 3 daily repeated 8 h-MS from pnd3–5. The objective of these experiments was to investigate further theapparent “desensitization” of HPA axis activity to repeated MS indifferent separation contexts in two rat strains. The degree of noveltywas varied in the separation environment using: (i) “home separa-tion”: the environment was the home cage and pups remainedgrouped together; (ii) “novel separation”: the environment did notcontain any element of the home cage and pups were additionallyisolated from their littermates. The effect of the different MS protocolswas investigated on basal and novelty stress-induced ACTH and CORTlevels on pnd 5. Since in the study of Enthoven et al.(2008), theapparent adrenal sensitivity to ACTH was altered dramatically duringthe repeated separations we also measured two biomarkers foradrenal function: tyrosine hydroxylase (TH) levels as index foradrenal medullary function and the level of melanocortin 2 receptor(MCR-2) as an index of adrenal sensitivity to ACTH. Moreover,maternal care was measured for the first 4 postnatal days to exploreits possible implication in the outcome of the different separationprocedures.

Similar to what we observed in mice, we also found that the MS-induced CORT response is readily abolished in rats if the separationsare repeated daily, but that the animals' ability to respond to a noveltystressor depends on the separation context. Genotype and maternalcare upon reunion did not affect the desensitization phenomenon, butrather appeared to be associatedwith stress responsiveness of the pupto an acute novelty stressor.

Materials and methods

Animals

Wistar rats (originally obtained from Harlan, Horst, The Nether-lands) and Long Evans rats (originally obtained from Elevage Janvier,Le Genest-St-Isle, France) were used in this study and housed in ouranimal facility under a 11:13 h light/dark cycle (lights on at 08.30 h,illumination inside the cage: 20–30 lux, temperature: 20±1 °C,relative humidity: 60±10%) and low volume background noise(40 dB). Food (RM3, Special Diet Services, Witham, Essex, UK) andwater (containing 0.02% HCL) was ad libitum. Upon arrival males andfemales were housed in groups of 2 or 3 in macrolon-polycarbonatetype IV cages with wire lid; 60×38×20 cm; containing sawdustbedding and tissue, and used for breeding at least after a habituationperiod of one week.

Animal experiments were approved by the Local Committee forAnimal Health, Ethics and Research of Leiden University and carriedout in accordance with European Communities Council Directive86/609/EEC.

Breeding

Two or three females of F1 generation, which were group housedfor at least a week, were mated with a male in Type IV macrolon-polycarbonate cages with wire lid. After 10 days, male were removedfrom the cage and pregnant females were transferred individually toclean cages (macrolon-polycarbonate type III cages with wire lid;42.5×26.6×18.5 cm) containing sawdust and two sheets of papertowels for nest material. Pregnant females were checked for littersdaily at 19:30 h starting from 20 days after the start of breeding. Iflitters were present, the day of birth was defined as pnd 0 for thatlitter. On the day after parturition, pnd 1, each litter was culled to 8–10healthy pups (ratio males:females=1:1) and remained undisturbeduntil used in the study.

Maternal behavior

The maternal behavior of each dam was observed and scored forfive-60 min periods per day during the first 4 days post partum usinga procedure originally described before (Champagne et al., 2008;Champagne et al., 2003; Myers et al., 1989). Observations wereperformed at three periods during the light phase (10:00, 13:30, and17:00 h) and two periods during the dark phase (07:30 and 19:30 h;under 2×60WredTLD-light). The behavior of eachmotherwas scoredevery 3 min (20 observations per period, 100 observations per day).

We scored the following maternal behaviors: retrieval: the damretrieves her pups from either a location outside the nest to the nest orfrom inside the nest to a new location, maternal contact: the dam is incontact with the pups but not nursing or licking, licking and grooming(LG): the dam is licking and grooming either the whole body orspecifically the anogenital area of the pup, passive nursing posture:the dam is in a passive posture; she is lying either on her back or sidewhile the pups are nursed, away from nest: there is no maternalcontact, nest building: the dam gathers material to a nest side,redistributes nest material, creates nest or changes shape of nest,burying her pups. Finally, arched-back nursing was also measuredwith a distinction of (i) (passive) low arch: dam in a passive posturewhere she lays over (some or all of) her pups flat (blanket nursing) orwith a low arch. (ii) (active) low arch: the dam is positioned lesspassively over (some or all of) her pups; her possible other activitiesare licking pups, moving nest material, self grooming, repositioningpups in the nest, eating etc. (iii) Middle arch: in this position the damshows a greater arch and her other activities are only pup-oriented(licking pups), (iv) High arch: in this position the limbs of the dam areextended so the pups have full access to her nipples.We considered as(overall) passive nursing the sum of the passive nursing posture andthe (passive) low arch back nursing scores. The other three nursingpostures (active low arch, middle arch, high arch) were considered(overall) active nursing (AN).

Other dam non-maternal care behaviors were also observed likeeating, drinking water, chasing tail, self grooming, digging, andsleeping. Note that some behavioral categories were not mutuallyexclusive. For, example, licking and grooming often occurs while thedam is nursing the pups. Other litter conditions were noticed: splitlitter (pups divided over two positions) and buried pups.We analyzedthe percentage of observations in which: 1) the dams displayed eachbehavior or 2) litters were in a certain condition.

In the result section, we report frequencies of AN (as % ofobservations) and LG (as % of observations). Note that AN and LGfrequencies both include instances where both of the behaviorsoccurred simultaneously and those instances are quite frequent.

167N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

Maternal separation (MS; Fig. 1)

MS occurred at either pnd 3, 4 and 5 or only pnd 5, lasting 8 h each.Litters were randomly distributed over experimental conditions.

Dams' transfer from the litter (“Dam out”)

At 9:00 h, dams selected for MS were removed from their cage(“home” cage), placed in a cage of the same type and transferred to anadjacent room (“dams” room). In the “dams” room, the environmen-tal conditions were the same except the lighting intensity was higher(illumination inside the cage: 50–60 lux).

Separation procedure

After the dam was relocated to a new cage, litters were keptwithout any food or water available for 8 h (9:00 to 17:00 h). Thehome cagewas placed on heating pads (33–38 °C; TM 22, Beurer, Ulm,Germany) to maintain the body temperature of the pups. To acquirethe desired temperature, heating pads were turned on 30 min prioruse.

We used two following separation contexts:

– “Home separation” (HOME SEP; Fig. 1A). The pups remained intheir familiar environment (housing room, home cage) togetherwith their littermates.

– “Novel separation” (NOVEL SEP; Fig. 1B). The pups were moved toan adjacent unfamiliar room, with similar conditions as thehousing room. Pups were put individually in new clean cages

Fig. 1. Sequential steps of early life manipulations: (A) Home separation: (i) Dam istaken out of the nest, but pups (“p”) stay altogether in the home cage on top of theirhome bedding (“H BED”) for 8 h (ii) Dam is reunited with the pups. (B) Novelseparation: (i) Dam is taken out of the nest (ii) Pups (“p”) placed in novel context for8 h (other room, isolated from the littermates, novel bedding: “N BED”) (iii) Pups arebrought back to their home cage (iv) Dam is reunited with the pups.

(macrolon-polycarbonate type II, which were divided in compart-ments of 18×20×14 cm, containing fresh sawdust bedding) andplaced on heating pads. The separated pups housed in thisunfamiliar novel context, experienced the absence of their dam,the home cage environment and proximal contact with theirlittermates (isolation).

Reunion (“Dam back”)

At 17:00 h, the pups were returned to their home cage followed bytheir dams. Dams of separated pups in home and novel contexts werereunited with their litter at the same time.

Control litters

Non-separated (NON SEP) litters remained undisturbed with theirdams in the housing room until the time of testing.

Testing: novelty exposure (pnd 5)

We determined the HPA axis responsiveness to a mild stressor at17:00 h on pnd 5. Pups were removed from their separationenvironment (home or novel) and either sacrificed immediately bydecapitation or placed individually in new clean cages (same type asin the novel separation), containing fresh sawdust. Novelty exposurewas carried out in a separate room, the “novelty exposure” room,under similar environmental conditions as the housing room. Thecages were placed on heating pads (33–38 °C) to maintain the bodytemperature of the pups. After 30 min from the onset of the stressor,the pups were sacrificed.

Note that this manipulation is different for pups of the differentgroups: for the HOME SEP pups (1st HOME SEP, 3rd HOME SEP), it isthe first time they experience a novel and unfamiliar cage, for theNOVEL SEP pups it is a relatively familiar manipulation since they arecoming from a similar environment. Therefore, the perceived degreeof novelty may be different.

Experimental design (Fig. 2)

Litters for the endocrine experiments: in order to minimize inter-litter differences, every treatment group consisted of 4–7 litters and,within each litter; we distributed the pups equally in terms of timepoint of sacrifice and sex. Number of animals per time point in everytreatment group was 8–18.

– Experiment I (Fig. 2A): to determine the effects of repeatedseparations in different separation context on body growth, ACTH& CORT secretion andmaternal care, litters were divided into threetreatment groups: single home separation (1st HOME SEP),repeated home separation (3rd HOME SEP), and repeated novelseparation (3rd NOVEL SEP). We sacrificed rat pups in threedifferent testing conditions: basal levels (basal), 8 h of separation(separated), 8 h of separation+30 min of novelty (novelty). Atotal of 12 litters were used (4 litters for each group with 8 pups inevery time point).

– Experiment II (Fig. 2B): to determine the effects of repeatedseparations in home context on adrenal activity, litters weredivided into two treatment groups: single home separation (1stHOME SEP) and repeated home separation (3rd HOME SEP). Wesacrificed rat pups in two different testing conditions: basal levels(basal) and 8 h of separation (separated). A total of 8 litters wereused (4 litters for each group with 8 pups in every time point).

– Experiment III (Fig. 2C): to determine the effects of repeatedseparation in home context on ACTH&CORT secretion of anotherrat strain (Long Evans), litters were divided into two treatmentgroups: single home separation (1st HOME SEP) and repeated

Fig. 2. Graphical representation of the three experiments for endocrine measures providing an overview of the time lines of different treatment. First is the time bar from postnatalday (pnd) 0 to 5; within each pnd in white is the light period of the light cycle and in black the dark period. Each big white bar under the time bar represents different conditions ofsacrifice. Within every white bar there are boxes indicating the experimental manipulations; gray box represents 8 h of maternal separation in home context, gray box with blackhatched stripes represents 8 h of maternal separation in novel context and black box represents 30 min of novelty exposure. Pups were sacrificed on pnd 5 measured in basalconditions (basal), after 8 h of maternal separation (separated) or 8 h of maternal separation with an additional 30 min of exposure to novelty (novelty). Treatment groups: firstseparation (1st HOME SEP) had no previous history of treatments, third separation animals were exposed to 8 h of maternal separation on pnd 3 and 4 in a home (3rd HOME SEP) ornovel context (3rd NOVEL SEP). Note that novelty exposure is different for pups of the different groups: for the HOME SEP (1st HOME SEP, 3rd HOME SEP) pups, it is the first timethey experience a novel and unfamiliar cage, for the 3rd NOVEL SEP pups it is a relatively familiar manipulation since they are coming from a similar environment.

Table 1Synopsis of the number of litters used.

Exp. group Exp. I Exp. II Exp. III Maternal careobservations

Experiments I–III Maternal careobservations

1st HOME SEP NON SEP 4 4 6 4+6=103rd HOME SEP HOME SEP 4 4 7 4+5=93rd NOVEL SEP NOVEL SEP 4 – – 4+5=9

Note. First separation (1st HOME SEP) or Non separated pups (NON SEP) had noprevious history of treatments pnd 1 to 4, separated pups were exposed to 8 h ofmaternal separation on pnd 3 and 4 in a home (3rd HOME SEP/ HOME SEP) or novelcontext (3rd NOVEL SEP/NOVEL SEP).

168 N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

home separation (3rd HOME SEP). We sacrificed rat pups in threedifferent testing conditions: basal levels (basal), 8 h of separation(separated), 8 h of separation+30-min of novelty (novelty). Atotal of 13 litters were used (6 litters for the 1st HOME SEP groupwith 16 pups in every time point, 7 litters for the 3rd HOME SEPgroup with 18 pups in every time point).

Litters for maternal care observations: maternal care measure-ments were performed for the litters of Experiment I together withsome extra litters that were not used in the endocrine experiments: 6non separated litters, 5 repeatedly separated litters on pnd 3 and 4 in ahome (HOME SEP) or novel context (NOVEL SEP). On pnd 3 & 4, for the3rd HOME SEP/ HOME SEP and 3rd NOVEL SEP/ NOVEL SEP groups,damswere not in contact with the pups at two time points (10:00 and13:30) because they were separated, and, therefore, we could notcollect maternal care data.

See Table 1 for a synopsis of litters used in the differentexperimental groups.

Collection of blood plasma and adrenals

At the designated time point, the pups were sacrificed bydecapitation. Trunk blood from all pups was collected individuallyin 1.5 ml EDTA-coated microcentrifuge tubes. All blood samples werekept on ice and later centrifuged for 15 min at 13,000 rpm at 4 °C.Plasma was transferred to clean 1.5 ml microcentrifuge tubes. Allplasma samples were stored frozen at -20 °C until the determination

of ACTH and CORT. After decapitation, adrenals were dissected andsnap frozen in isopentane on dry ice and stored at −80 °C until usedfor Western blotting.

Measurements

Body weight (gr) was measured just before every experimentalmanipulation with an electronic precision scale (MXX-2001, DenverInstrument, Göttingen Germany; readability 0.1 g, linearity 0.2 g).Since the groups were different in birth weight (pnd 1 weight), wecalculated the ratio (in %) of the body weight measurements to birthweight.

169N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

ACTH was measured by radioimmunoassay (MP Biomedicals, LLC,NY, USA; sensitivity 10 pg/ml, intra-assay variation 4.1%, interassayvariation 4.4%). Samples were determined in a 50% dilution, startingwith 25 μl blood plasma. All samples were analyzed in one assay toexclude inter-assay variability.

CORT was measured by radioimmunoassay (MP Biomedicals, LLC,NY, USA; sensitivity 1.25 ng/ml, intra-assay variation, 4.4%, interassayvariation 6.5%). Concentrations were determined in duplicate from anextended standard curve (0, 6.25, 12.5, 25, 50, 100, 250, 500 and1000 ng corticosterone/ml), since we noted that the lower boundaryprovided by the kit was not sensitive enough to measure neonatalbasal plasma concentrations. All samples were analyzed in one assayto exclude inter-assay variability.

Tyrosine hydroxylase (TH) & melanocortin receptor type 2(MCR-2) protein levels

Adrenals were homogenized in 400 μl lysis buffer (Triethanola-mine, NaCl, DOC, SDS, triton-X-100) and protease inhibitor was addedto inhibit proteins' degradation. This lysate was spun down andsupernatant was kept and stored in−20 °C. Concentration of proteinspresent in the supernatant was determined using a Thermo ScientificPierce BCA Protein Assay. Therefore, a calibration curve (BovineSerum Albumin in 5 dilutions) was done.

Western blotting was performed, according to a previouslydescribed method (Brinks et al., 2007), in duplicate on thesupernatant of the homogenized adrenals to measure TH and MCR-2 protein levels. Each sample supernatant was loaded in a concen-tration of 1 mg/ml (by varying the amount of H2O added to thesample). The samples also included a standard volume of samplebuffer and were denaturized at 95 °C (5 min) and subjected to SDS-PAGE.

After electrophoresis, the proteins were transferred to a mem-brane (blotting) overnight (4 °C, 125 mA). The day after, the blotswere blocked in 10 mM Tris–HCl (pH 8.0), 150 mM NaCl, and 0.05%Tween 20 containing 5% non-fat dried milk powder and, then,incubated with the primary antibody and the secondary antibodyconsecutively. For TH, the primary antibody used was rabbit antiTyrosine Hydroxylase (TH) (AB152) ordered from Millipore in a1:1000 concentration. The secondary antibody used was Goat-anti-rabbit IgG-HRP in a 1:5000 concentration. For MCR-2, the primaryantibody used was mouse anti Melanocortin Receptor-2 (MCR-2)ordered from Chemicon in a 1:1000 concentration. The secondaryantibody used was goat anti mouse IgG-HRP in a 1:5000 concentra-tion. For TH, we used liver tissue as negative control and adult ratadrenal tissue as positive control. For MCR-2, we used human skinsamples as positive control and water as negative control. Sampleswere also tested on their α-tubulin levels, as well, to correct for thetotal amount of protein. The primary antibody used was anti-mouseα-tubulin in a 1:5000 concentration and the secondary antibody usedwas goat anti mouse IgG-HRP in a 1:10,000 concentration.

After washing of the antibodies, blots were incubated withperoxidase-conjugated antibodies (1:10,000; Jackson ImmunoRe-search Laboratories, West Grove, PA). Immunoreactive bands werevisualized by enhanced chemiluminescence and the blots wereexposed to films. The autoradiographs (films) were scanned andoptical density (OD) of the TH, MCR-2 and α-tubulin bands weredetermined using Image J software. The TH and MCR-2 values ofthe samples were corrected for total protein (α-tubulin); therefore,the ratios between the TH and α-tubulin levels and MCR-2 andα-tubulin levels were calculated. In order to compare samples ranin different gels we used also one sample (“standard sample”) thatwas loaded in all gels. Group size n=8 adrenals (from 8 separateanimals) per time point of each treatment group were used for themeasurements.

Statistical analysis

The results were analyzed by analysis of variance (ANOVA) withthe level of significance set at pb0.05. Where appropriate, simple andinteraction main effects were investigated further with subsequentpost-hoc comparisons (by Tukey test). The statistical analysis wasadjusted for non-equivalent groups when needed. The initial analysisof pups' measurements included sex as a factor; once it wasdetermined that sex was not a significant factor, data from malesand females were pooled. Data are presented as mean±SEM. Thelevel of significance was set at p≤0.05.

Results

Experiment I

Wistar pups were exposed to repeated MS under differentseparation contexts with the goal to examine the “desensitization”of the endocrine responses.

Body weight (data not shown)Repeated measures ANOVA, for the ratio of body weight to birth

weight, revealed main effects of time (F1,45=4027.35; p≤0.001) andthe interaction of time and treatment (F1,45=34.36; p≤0.001). Thetreatment effect was significant in all time points (p≤0.001). On pnd5, HOME SEP and NOVEL SEP were not different, but both were lighterthan the NON SEP (p≤0.001).

HOME SEP vs. NOVEL SEP: The two repeatedly separatedgroups were followed for body growth in more time points (pnd 3and pnd 4) than the controls. Therefore, we performed a separatestatistical analysis for their comparison. Repeated measures ANOVArevealed main effects of time (F6,180=1315.76; p≤0.001), but not oftreatment or the interaction of time and treatment. The two groups(HOME SEP and NOVEL SEP) were not different in weight at any timepoint.

ACTH (data not shown)ACTH basal levels for naïve Wistars on pnd 5 were 90.08±11.38

(pg/ml). Two-way ANOVA did not reveal any main effects oftreatment, time or their interaction. ACTH is not different after asingle event of 8 h of MS on pnd 5. Rat pups also did not respond to30 min novelty if it was implemented immediately after the first orthird MS period. There was no effect of treatment on ACTH. However,if repeatedly separated groups were compared separately, ACTHlevels of NOVEL SEP pups, after the thirdMS, were lower than the onesof the HOME SEP pups (p=0.046).

CORT (Fig. 3)Two-way ANOVA revealed effects of treatment (F2,63=51.54;

p≤0.001), time (F2,63=17.50; p≤0.001) and their interaction(F4,63=10.97; p≤0.001).

1st HOME SEPNaïve rat pups responded on pnd 5 with a three-fold increase of

CORT to 8 h-MS (p≤0.001) or to the combination of MS with noveltyexposure (p≤0.001). The novelty exposure did not, however, createan additional increase over the MS levels.

1st HOME SEP vs. 3rd HOME SEP/3rd NOVEL SEPIf on pnd 3 and 4 pups were separated from their mother, on pnd 5,

CORT levels, in response to the third period of MS, was not alteredregardless if the MS happened in a home or a novel context(“separated” vs. “basal” levels in 3rd HOME SEP or in 3rd NOVELSEP). CORT levels after MS or after the combination of MS and novelty(“separated” and “novelty” levels) were lower for the repeatedly

Fig. 3. Corticosterone (ng/ml) blood plasma levels measured at basal conditions (basal;white bars), after 8 h of maternal separation [separated; home separated (HOME SEP):gray bars/novel separated (NOVEL SEP): gray bars with black hatched stripes] or 8 h ofmaternal separation with an additional 30 min of exposure to novelty (novelty; blackbars). First separation (1st HOME SEP) had no previous history of treatments. Thirdseparation animals were exposed to 8 h of maternal separation on pnd 3 and 4 in ahome (3rd HOME SEP) or novel context (3rd NOVEL SEP). Note that novelty exposure isdifferent for pups of the different groups: for the HOME SEP (1st HOME SEP, 3rd HOMESEP) pups, it is the first time they experience a novel and unfamiliar cage, for the 3rdNOVEL SEP pups it is a relatively familiar manipulation since they are coming from asimilar environment. Data represented mean±SEM. Significance level was set atp≤0.05. * vs. basal, # vs. separated levels of the same treatment group, ¥ vs.correspondent value of 1st HOME SEP. n=8 per time point of each treatment group.

170 N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

separated pups in both contexts (p≤0.001) compared to single MSpups.

3rd HOME SEP vs. 3rd NOVEL SEPHowever, when exposed additionally to novelty after the third MS,

the 3rd HOME SEP pups displayed an increase in CORT levels(compared to “basal” levels: p=0.023, 50% increase, but not over“separated” levels: p=0.065). The small increase of CORT whennovelty is combined with MS is absent when MS occurs in the novelcontext.

Maternal Care: in order to investigate whether the rapid de-sensitization of the CORT response to repeated MS was related tomother-pup interaction upon reunion we measured several compo-nents of maternal behavior on pnd 1–4.

Licking & grooming (Fig. 4)Repeated measures ANOVA revealed main effects of time

(F19,475=12.94; p≤0.001) and the interaction of time and treatment(F38,475=3.27; p≤0.001) on LG (Fig. 4A), but overall mean of LGacross the first postnatal days (Fig. 4B) was not significantly dif-ferent between treatment groups (NON SEP: 6.08±0.59, HOME SEP:5.17±0.52, NOVEL SEP: 5.42±0.51).

The time effect was not significant for the dams of the NON SEPpups, but it was significant for the dams of the HOME SEP (p≤0.001)and NOVEL SEP pups (p≤0.001). Post hoc analysis revealed that thedifferences were mainly in the post separation hour observationperiod on pnd 3 and pnd 4.

On pnd 3 (Fig. 4B), the dams of separated pups displayed increasedLG levels upon reunion after the first 8 h-MS from their pups ascompared to before separation levels (for the HOME SEP dams:p=0.012, for the NOVEL SEP dams: p≤0.001). Between the treat-ment groups, before first MS, there was no difference in LG levels.After the separation, only the dams of the NOVEL SEP pups displayhigher levels of maternal care at this time point compared to bothNON SEP (p≤0.001) and HOME SEP groups (p=0.043).

On pnd 4 (Fig. 4C), the dams of separated pups increase their LGlevels upon reunion after 8 h-MS for the second time (for the dams of

HOME SEP: p=0.003, for the dams of NOVEL SEP: p≤0.001), ascompared to before separation levels. Interestingly, the effect ofMS onmaternal care did not habituate. Before the second MS the dams ofNOVEL SEP pups displayed the least LG (p=0.011 vs. NON SEP).When they were reunited with their pups, after the second MS, damsof separated pups in both contexts displayed higher maternal carecompared to the controls (for the dams of HOME SEP: p=0.038, forthe dams of NOVEL SEP: p=0.005).

Active nursing (Fig. 5)Repeated measures ANOVA reveal main effects of time (F19,475=

16.20; p≤0.001) and the interaction of time and treatment (F38,475=7.37; p≤0.001) on AN (Fig. 5A), and of treatment (F2,25=3.62;p=0.041) on the overall mean of AN across the first postnatal days(Fig. 5B). The time effect was significant for all groups (for the dams ofNON SEP pups p=0.008, of HOME SEP pups p≤0.001 and of NOVELSEP pups p≤0.001). Dams of HOME SEP pups, compared to thecontrols (NON SEP), displayed different levels of AN across theindividual time points (p=0.042) and ended-up with less overallmean across pnd 1–4 (p=0.039). We conducted further post hocanalysis for the post separation hour observation period on pnd 3 andpnd 4. On both days, there is no treatment effect before or after thefirst 8 h-MS indicating there is no post-reunion increase of AN in bothdays.

Experiment II

Because the endocrine blood levels imply changes in adrenalsensitivity to subsequent stressors upon repeated MS in the homecontext we have examined two biomarkers that may give anindication how adrenal function is affected.

TH (Fig. 6A)Two-way ANOVA revealed effects of time (F1,31=5.43; p=0.027)

and the interaction of treatment and time (F2,31=4.52; p=0.009). If,on pnd 3 and 4, pups were separated from their mother in a homecontext, on pnd 5, a 65% reduction in TH level is displayed compared tonaïve pups (basal levels 1st HOME SEP vs. 3rd HOME SEP; p=0.014).The reduction is followed by a four-fold increase in response to thethird period of MS (p=0.006) (“separated” vs. “basal” levels in 3rdHOME SEP).

MC2 receptor (Fig. 6B)Two-way ANOVA revealed effect of the interaction of treatment

and time (F2,31=6.09; p=0.020). If, on pnd 3 and 4, pups wereseparated from their mother in a home context, on pnd 5 receptorlevels display a 50% reduction compared to naïve pups (basal levels1st HOME SEP vs. 3rd HOME SEP), but it was not significant(p=0.055). The reduction is followed by an increase (three-fold) inresponse to the third period of MS (p=0.047) (“separated” vs. “basal”levels in 3rd HOME SEP).

Experiment III

In order to assess possible strain differences the effect of repeatedMS in home context, Long Evans rats was used.

ACTH (data not shown)ACTH basal levels for naïve Long Evans on pnd 5 were 35.56±

1.96 (pg/ml). Two-way ANOVA revealed effects of treatment (F1,110=80.51; p≤0.001), time (F2,110=8.25; p=0.002) and their interaction(F2,110=9.25; p≤0.001).

1st HOME SEP. Naïve rat pups on pnd 5 responded with an ACTHincrease to the combination of MS with novelty exposure (p≤0.001;

Fig. 4. Licking & grooming (LG). (A) Time course during pnd 1–4: Non separated pups (NON SEP; empty circles with discontinuous black connecting line) had no previous history oftreatments. Separated pups had experienced 2 times 8 h of maternal separation on pnd 3 and 4 in home (HOME SEP: home separated; gray squares with gray connecting line) ornovel context (NOVEL SEP: novel separated; triangles with black connecting line). On pnd 3 & 4, for the HOME SEP and NOVEL SEP groups, dams were not in contact with the pups atthe time points 10:00 and 13:30. Therefore in the time course graph, they don't have data points and this is highlighted with a symbol;⊗. (B&C): Maternal care at reunion on pnd 3 &4 respectively: LG measured for an hour at the pre-separation time period 7:30–8:30 (white bars) or one hour during the post separation time period at 17:00–18:00 (white barswith black horizontal stripes). Data represented mean±SEM. Significance level was set at pb0.05. τ denotes overall time effect in this group, * vs. pre-separation levels, ¥ vs.correspondent value of NON SEP, £ vs. correspondent value of HOME SEP. n=9–10 per treatment group.

171N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

50% increase). The novelty exposure creates an additional increaseover the MS levels (p=0.030).

3rd HOME SEP. If on pnd 3 and 4 rat pups were separated from theirmother, on pnd 5 ACTH levels, in response to the third period of MS,were decreased (“separated” vs. “basal” levels in 3rd HOME SEP;p=0.038) and they even became lower than basal levels of naïvepups of pnd 5 (p=0.002; 35% decrease).

1st HOME SEP vs. 3rd HOME SEP. ACTH levels, either basal, after MS orafter combination of MS and novelty levels of repeatedly separatedpups are reduced compared to the respective values of the singlyseparated pups (p=0.031 for “basal” values, p≤0.001 for “separated”values, p≤0.001 for “novelty” values).

CORT (Fig. 7)Two-way ANOVA revealed effects of treatment (F1,110=103.63;

p≤0.001), time (F2,110=86.12; p≤0.001) and their interaction(F2,110=37.76; p≤0.001).

1st HOME SEP. Naïve rat pups on pnd 5 responded with a CORTincrease to 8 h-MS (p≤0.001; four-fold increase) or to the combina-tion of MS with novelty exposure (p≤0.001; five-fold increase). Thenovelty exposure resulted in an additional increase over the MS levels(p=0.004).

3rd HOME SEP. If, on pnd 3 and 4, rat pups were separated from theirmother, CORT levels were not altered on pnd 5, in response to thethird period of MS (“separated” vs. “basal” levels in 3rd HOME SEP).The HOME SEP pups displayed an increase in CORT over basal levelson pnd 5 if they were additionally exposed to novelty after MS(p≤0.001; two-fold increase), over separated levels (p=0.015) andover CORT levels of naïve pups (p≤0.001).

1st HOME SEP vs. 3rd HOME SEP. CORT levels after MS or aftercombination of MS and novelty (“separated” and “novelty” levels)were reduced in the repeatedly separated pups compared to therespective values of the singly separated pups (p≤0.001 for bothcomparisons).

Fig. 5. Active nursing (AN). (A) time course during pnd 1–4: Non separated pups (NON SEP; empty circles with discontinuous black connecting line) had no previous history oftreatments. Separated pups had experienced 3 times 8 h of maternal separation on pnd 3 and 4 in home (HOME SEP: home separated; gray squares with gray connecting line) ornovel context (NOVEL SEP: novel separated; triangles with black connecting line). On pnd 3 & 4, for the HOME SEP and NOVEL SEP groups, dams were not in contact with the pups atthe time points 10:00 and 13:30. Therefore in the time course graph, they don't have data points and this is highlighted with a symbol: ⊗. (B) Overall mean across the first fourpostnatal days. NON SEP: white bars, HOME SEP: gray bars, NOVEL SEP: gray bars with black hatched stripes. Data represented mean±SEM. Significance level was set at pb0.05. τdenotes overall time effect in this group, ¥ vs. correspondent value of NON SEP. n=9–10 per treatment group.

172 N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

Discussion

The present study was designed to extend previous data on theimmediate outcome of repeated daily maternal separations on the

Fig. 6. Adrenal TH (A) & MC2 receptor (B) protein levels measured at basal conditions(basal; white bars), after 8 h of maternal separation [separated; home separated(HOME SEP): gray bars]. First separation (1st HOME SEP) had no previous history oftreatments. Third separation animals were exposed to 8 h of maternal separation onpnd 3 and 4 in a home context (3rd HOME SEP). Data represented mean±SEM.Significance level was set at p≤0.05.* vs. basal, ¥ vs. correspondent value of 1st HOMESEP. n=8 adrenals (from separate pups) per time point of each treatment group; theWesterns were performed in duplicate.

HPA axis from the mouse to the rat. Our previous studies had revealedthe remarkable phenomenon that these repeated daily maternalseparations inmice resulted in a desensitization of the CORT-responseto the separation procedure itself, while the pups continued torespond to an acute novelty stressor (Enthoven et al., 2008). In thisstudy we demonstrate that also the rat readily adapts to repeateddaily separations. Specifically, we show that after 8 h of maternalseparation CORT levels increased markedly in the 5 day old Wistarand Long Evans rat pup. However, if the pups had been exposed alsothe two preceding days to 8 h-MS this rise in CORTwas abolished. Theadrenal desensitization induced by the homotypic daily repeated 8 h-MS was not affected by genotype, separation environment ormaternal care upon reunion. However, repeatedly home separatedpups show a subtle enhancement of the adrenal CORT response to aheterotypic acute novelty exposure. Additionally, the adrenal TH andMCR-2 protein content was higher than observed in pups exposed tosolely a single 8 h-MS. This increased stress response to novelty after

Fig. 7. Corticosterone (ng/ml) blood plasma levels measured at basal conditions (basal;white bars), after 8 h of maternal separation [separated; home separated (HOME SEP):gray bars] or 8 h of maternal separation with an additional 30 min of exposure tonovelty (novelty; black bars). First separation (1st HOME SEP) had no previous historyof treatments. Third separation animals were exposed to 8 h of maternal separation onpnd 3 and 4 in a home context (3rd HOME SEP). Data represented mean±SEM.Significance level was set at pb0.05. * vs. basal, # vs. separated levels of the sametreatment group, † vs. basal of 1st HOME SEP, ¥ vs. correspondent value of 1st HOMESEP. n=14–16 per time point of each treatment group.

173N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

repeated episodes of home rather than novel 8 h-MS was observed inboth genotypes, with the Long Evans pups displaying a relativelyhigher stress-response compared to the Wistar pups.

Differential ACTH and CORT time course during maternal absence

In line with expectations, CORT levels following a single 8 h-MS onpnd 5 were significantly increased relative to basal levels. Both in ratsand mice, a single prolonged period of maternal separation of ≥8 hduring the SHRP is needed for a large increase of CORT (Enthoven et al.,2008; Rosenfeld et al., 1992b; Schmidt et al., 2004; Stanton et al., 1988;Walker et al., 1991). As far as ACTH levels are concerned, in both ratstrains, theywere not increased after 8 h ofmaternal absence. However,in mice it is clearly shown by Schmidt et al. that during the course of a24 h-MS, ACTH levels are increased between 8 and 12 h from the onsetof separation (Enthoven et al., 2008; Schmidt et al., 2004). Hence, thelack of ACTH response in rats observed in the present study is probably arat-mouse difference (Schmidt et al., 2002), since our ACTH findings areactually in agreementwith other studies using rats.Walker et al. (1991)showed already at 30 min of maternal absence a rise in ACTH level,which then returned tobaseline after8 hof separationpossibly byeitherdepletion of the pituitary ACTH stores or glucocorticoid feedbackinhibition of ACTH release. That the feedback inhibition alreadyoperatesin newborn rats at pnd 5 was shown by injecting CORT or dexameth-asone to naïve, deprived or adrenalectomized pups (van Oers et al.,1998a; van Oers et al., 1998b; Walker et al., 1990; Walker et al., 1986).Accordingly, a single 8 h-MS period permits a robust CORT increase,while at that time the ACTH response is suppressed. Our findings (inLong Evans) argue against the depletion of ACTH from the pituitarystores sincewe observed increased ACTH levels in response to 30 min ofnovelty following a single 8 h-MS.

Desensitization of CORT response to maternal absence: a robustphenomenon

The interesting novel aspect of our previous studies with mice wasthat upon repeated separations the pup readily adapts to maternalabsence and as a result the separation-induced increase in CORT doesnot occur any longer. We report here that also in the newborn rat thebasal CORT response to daily repeated 8 h-MS is abolished, both inWistar and Long Evans 5 day old rat pups while their adrenals are ableto respond to a single 8 h-MS. This habituation or adaptation of thepup to the experience of repeatedMSwas not due to a shift in the timecourse of theMS-induced CORT response (Enthoven et al., 2008). Also,repeated MS in a home context does not result in a cumulative CORTresponse irrespective of the duration (from 15 min to 8 h) of maternalabsence (D'Amato et al., 1992; Rosenfeld et al., 1992b). It should benoted, however, that these authors did not report desensitization ofthe adrenal response to repeated maternal absence as was shown inEnthoven's study for the mouse, and here in two rat strains. However,the current study differs from the mouse study in one aspect. WhileEnthoven et al. found a persistent decrease in basal CORT release aftertwo daily separations; we did not see this decrease consistently in thetwo rat strains. Possibly this difference could be explained by smallcircadian fluctuation present already in that age (Enthoven et al.,2008). Enthoven measured basal levels at 9:00 h (16 h after previousseparation) and we did at 17:00 h (24 h after the end of the previous8 h-MS). However, the circadian fluctuation of CORT and the potentialinfluence of early life experience have not been reported in neonaterodents so far earlier than the third week of life (Ader, 1969).

In the present study it was shown that if the pups are housed in anovel environment isolated from their littermates and deprived fromall familiar cues during the 8 h-MS, the desensitization of the adrenalstill happens. In the rat, comparable studies have been performed,however with variable outcomes. Several studies using shorterperiods of maternal absence in a novel environment (1 hour x

8 days) showed sensitization of the adrenal response to thisprocedure (Knuth and Etgen, 2005; McCormick et al., 1998). Otherstudies (15 min×8 days, 1 h×3 days) did not (McCormick et al.,1998; Vazquez and Akil, 1992). Thus, taken together, the duration andfrequency of the daily separations might influence the outcome.Overall, the fact that adrenal corticosterone secretion under twowidely different conditions displayed desensitization to maternalabsence demonstrated the robustness of this phenomenon.

Enhanced adrenal sensitivity to stress after repeated maternal absence

While it is now well-established that the adrenal readilydesensitizes to homotypic repeated maternal separations, it is alsowell-established that MS causes increased adrenal sensitivity toheterotypic stressors and exogenous ACTH. This enhanced adrenalsensitivity to a heterotypic stressor needs at least 8 h-MS to developand is profound after 24 h of separation (Levine et al., 1991; Okimotoet al., 2002). Previously, it was reported that despite the rapidadaptation of the HPA axis to daily repeated maternal absence, theCD1 mouse pup stayed on alert and retained the ability to respond tostressors with an increase in ACTH and CORT levels (Enthoven et al.,2008), which indicates a large steroid production capacity andargues against ACTH depletion. Wistar pups, when exposed tonovelty for 30 min immediately after the separation, show no ACTHresponse irrespective of whether the separation was the first (theLong Evans did) or the third. Also it did not matter whether theseparation context was familiar (home cage) or unfamiliar (novelcage). However, a subtle CORT response to the novelty stressor stilloccurred despite desensitization to the homotypic repeated maternalseparation both in Wistar and Long Evans pups.

Interestingly, in Wistars, repeated “home separation” vs. “novelseparation” had a different outcome on the response to the subsequent30 minnovelty stressor. Rat pups separated for 8 hon3 consecutive daysin a novel environment apparently adapted to this condition because theadditional 30 min of novelty stress (which in this group could beconsidered a homotypic stressor) could not trigger a response anymore.Apparently, the previous experience of maternal absence in a novelenvironment not only did prepare the pups for maternal absence butalso for the experience of “novelty” itself. Both Wistar and Long Evanspups, in the “home separation” condition, still show a CORT response.Moreover, Enthoven et al. (2008) reported that a daily repeatedexposure to a combination of 8 h home separation+additional 30 minnovelty was not able to attenuate the response to a subsequent 30 minnovelty exposure itself. This suggests that only if the environment, inwhich the novelty stress is experienced, is intrinsic to the housingconditions during separation, then the novelty stressor can beconsidered “homotypic” and the response to this stressor is abolished.

These observations underscore previous research showing that theneonatal adrenal function is altered after maternal absence (Okimotoet al., 2002). To further examine the altered adrenal function wemeasured ACTH receptors (MC2-R) in the neonate adrenals in anattempt to explain why adrenal sensitivity to the novelty stressor wasincreased in the face of unaltered circulating ACTH levels. We reportthat MCR-2 protein content was reduced 24 h after the second homeseparation, but enhanced after a third separation interval. This wouldpredict enhanced responsiveness to exogenous ACTH as had beenshown before, immediately after prolonged 24 h maternal separation(Okimoto et al., 2002; Rosenfeld et al., 1991). We also measured THprotein level as an indirect measure of medullary-catecholamineresponse to maternal absence (Okimoto et al., 2002). We show thatthe third home separation induced an increase of TH protein levelsover the basal levels (which were reduced 24 h after the secondperiod of MS). Apparently, the increase in adrenal activity developsafter repeated 8 h episodes of maternal absence (in our study) or afterprolonged 24 h of maternal absence (Okimoto study) and not after asingle 8 h episode. These experiments provide some insights into the

174 N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

mechanistic underpinning of enhanced adrenal sensitivity towardsheterotypic stressors, which are in line with the elegant experimentsof singly separated pups subjected to chemical sympathectomy(Walker, 1995).

Maternal care upon reunion

First, we would like to underline a possible methodologicalconstraint in all the above studies (including the present). AN pnd 1–4(alone or together with LG) in our study was 56% (of the observations),which is in the same range of previous reports on pnd 1–4 [49% (Pryce etal., 2001b), 60% (Macri et al., 2004)] andhigher thanpreviously reportedvalues taking into account longer periods (pnd 1–8: 40%; (Pryce et al.,2001b), 47% (Champagne et al., 2003); (Long Evans), 43% (Macri et al.,2004)]. For LGwe reported amean of 6% (of the observations), which isapproximately 2 fold lower than previously reported [14% (Pryce et al.,2001b), 12% (Champagne et al., 2001)]. Strain differences, culling ofpups, litter size and litter's sex ratio could be important factorsinfluencing baseline levels of LG in the various studies (Claessenset al., 2011). However, even the same research group, using the samestrain of rats with small changes in the experimental procedure,reported large differences in the pnd 1–4 LG scores [approximate meanvalues: 11% (Caldji et al., 1998), 12% (Champagne et al., 2001) 14% (Liuet al., 2000), 15% (Champagne et al., 2003)].

The question raised in this study was to investigate to what extentmaternal care the pups received, after reunion with the dam, couldexplain the effect of repeated separations on the HPA axis. It isbelieved that maternal absence can affect adrenocortical celldifferentiation and function (Rosenfeld et al., 1992a) and this effectseems related to feeding rather than tactile stimulation provided bythe dam. Feeding acts as an inhibitory factor to the neonate's basal andstress-induced adrenal activity (Schmidt et al., 2006; Suchecki et al.,1993; van Oers et al., 1998b). Therefore, one other mechanism for theCORT desensitization to maternal absence could be related to thepattern of tactile stimulation (in the form of LG) or food intake(nursing) experienced by the pups after reunion with their dam.

In the current study we observed the following: (i) home andnovel separated groups received overall similar care which, in the caseof AN, was reduced compared to that of the controls; (ii) postseparation bouts of LG were higher for novel separated than for homeseparated pups, (iii) separation did not induce an increase in AN uponreunion. These findings suggest that for some maternal behaviors(LG) the dam compensates upon reunion. However, for others (AN)she does not compensate, resulting in a lack of care. This alteredpattern of care after MSmight have led to a greater suppression of theHPA axis and result in blunted stress activation after the third 8 h-MS.However, differences in maternal care for pups experiencing MS infamiliar or unfamiliar contexts, were not reflected in the CORT-response to repeated absence. However, those differences in maternalcare may explain the differences between separation contexts in theresponse to a heterotypic stressor.

Plotsky and colleagues have argued that maternal care is the majorfactor driving the effects of MS. Their arguments were based onexperiments showing that dam's exposure to foster litters while herpups where in MS, did not lead, in adulthood, to enhanced HPA axis'responsiveness for the separated animals (Huot et al., 2004). Incontrast, other investigators, using a similar design, observed thatdams are actually able to distinguish the separated pups from nonseparated pups, by their altered vocalization behavior, leading to thepost separation bouts of maternal care in favor of the separated pups(Zimmerberg et al., 2003a; Zimmerberg et al., 2003b). Cumulatively,the maternal mediation hypothesis maybe is not the sole mechanismexplaining the effects of early life stress in the HPA axis activity, but itis proposed that environmental adversity and the maternal repertoireboth underlie the lasting alteration on the offspring's HPA response(Macri andWurbel, 2006). Finally, an interesting possibility is that the

response of the dams upon re-unionmight have worked as a cue usedby the pups to predict maternal return after a separation experience.

Mechanism of repeated maternal separations: neonatal learning

The mechanism underlying the effects of repeated separations hasbeen explored previously. It is known that food deprivation leads toincreased adrenocortical output and sensitivity to stressors (Dallmanet al., 1999). It was therefore reasonable to assume that metabolicfactors are involved. However, since in the previous mouse study therise in ghrelin and the decrease in glucose were identical aftereach separation we could rule out involvement of metabolism(Enthoven et al., 2008).

We also could eliminate enhanced glucocorticoid feedback aspotential mechanism for the lack of CORT response to repeatedseparations, since a glucocorticoid antagonist that profoundlyenhanced the CORT response to the 1st separation failed to do soafter the third. Small effects of mineralocorticoid receptor antagonistswere found though suggesting the involvement of higher brainregions in the effect of repeated separations (Enthoven et al., 2008).

The involvement of higher brain regions raises the possibility thatneonatal learning could have a key role since recent reports showedone-trial odor learning in this age (Moriceau and Sullivan, 2006). It isimportant to underline that the odor system is fully developed at thistime.When the dam is present in the nest, adversity towards the pupswill be negligible and attachment to the dam care-giver is expected todevelop irrespective the quality of maternal behavior (Raineki et al.,2010). Until approximately pnd 10, pups exhibit odor preference tonovel odors evenwhen they are pairedwith negative stimuli (Sullivanet al., 2000a). This odor preference is associated with enhanced co-activation of the locus coeruleus–olfactory bulb pathway (Sullivan etal., 2000b). In the post-sensitive period, odor-avoidance behaviorappears and is associated with neural processes in amygdala andpiriform cortex (Sullivan et al., 2000a). Interestingly, during the“sensitive” period when the dam is away, the odor aversion neuronalsystem is activated prematurely and aversive memories can beformed as long as the CORT levels are elevated in blood and amygdala(Moriceau and Sullivan, 2006; Moriceau et al., 2006).

In our experiment, the pups were at an age (pnd 3–5 during theSHRP) that permitted formation of memories only during long-termabsence of the dam. After being separated from their mothers for thefirst time, the pups may have learned to predict the return of themother and thus the reinstatement of maternal care. In other words,the pups previously separated do not respond to the homotypicseparation itself (habituation), but do respond to the heterotypicstressor (30 min of novelty). This notion calls for study of brain areasinvolved in processing of novel information for memory storage. ThePVN c-fos expression data in Enthoven' s experiments support thisline of reasoning (Enthoven et al., 2008), and recent findings from ourlaboratory have demonstrated particularly in the amygdala a rise in c-fos expression if MS rats were exposed to an heterotypic stressor(Daskalakis et al., 2009). As mentioned before also maternal cuesmight have helped in the potential contextual associations. However,we have to be cautious since there is no reported evidence yet ofcontextual fear learning the first week of life apart from the studies ofodor fear learning. The possibility that the odor cues of the familiar ornovel context are more important still remains to be tested.

Conclusion

Taken together, the current study shows that the effect of repeatedseparations on the HPA axis activity previously observed in mice canbe generalized to rats (Enthoven et al., 2008). To explain this we favorthe reasoning that the newborn rats readily learn to predict the returnof the dam after the first experience of 8 h absence irrespective ofwhether the pups are housed in the home or the novel environment.

175N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

This adaptation or habituation of the pup to maternal absencemanifests itself in the desensitization of the adrenocortical CORToutput normally observed after the first separation. It occursirrespective of rat strain and separation context, while metabolicfactors and maternal care upon reunion do not seem to be implicated.Following maternal absence the pups become more sensitive toheterotypic stressors on the adrenal level. We propose that theprotocols employed in MS studies should be standardized because thecurrent data predict a different outcome on stress responsiveness toan acute novel stressor depending on whether the pups wereseparated in home vs novel environments. It would be of interest infuture studies to test the hypothesis that these variations in early lifeexperience have different outcomes for brain function and behavior inadulthood.

Conflict of interest

The authors have no conflict of interest to report.

Acknowledgments

We would like to thank Servane Lachize for help with RIA andWesley L. Fung for help with Western blotting. This work wassupported by the Top Institute Pharmaceutical SciencesT5#209 (NPD,JJLL), EU-Erasmus (NPD), EU-lifespan (SEFC), NWO-NDRF/STIGON(LE), NWO-Aspasia (MSO), NWO-IRTG (MSO), Marie Curie Founda-tion (DLC), the Smart Mix Program of the Netherlands Ministry ofEconomic Affairs and The Netherlands Ministry of Education, Cultureand Science (DLC), and the Royal Netherlands Academy of Arts andSciences (ERdK).

References

Ader, R., 1969. Early experiences accelerate maturation of the 24-hour adrenocorticalrhythm. Science (New York, N.Y.) 163, 1225–1226.

Biagini, G., Pich, E.M., Carani, C., Marrama, P., Agnati, L.F., 1998. Postnatal maternalseparation during the stress hyporesponsive period enhances the adrenocorticalresponse to novelty in adult rats by affecting feedback regulation in the CA1hippocampal field. International Journal of Developmental Neuroscience 16,187–197.

Brake, W.G., Zhang, T.Y., Diorio, J., Meaney, M.J., Gratton, A., 2004. Influence of earlypostnatal rearing conditions on mesocorticolimbic dopamine and behaviouralresponses to psychostimulants and stressors in adult rats. The European Journal ofNeuroscience 19, 1863–1874.

Brinks, V., van der Mark, M., de Kloet, R., Oitzl, M., 2007. Emotion and cognition in highand low stress sensitive mouse strains: a combined neuroendocrine and behavioralstudy in BALB/c and C57BL/6J mice. Frontiers in behavioral neuroscience 1, 8.

Caldji, C., Tannenbaum, B., Sharma, S., Francis, D., Plotsky, P.M., Meaney, M.J., 1998.Maternal care during infancy regulates the development of neural systemsmediating the expression of fearfulness in the rat. Proceedings of the NationalAcademy of Sciences of the United States of America 95, 5335–5340.

Champagne, F., Diorio, J., Sharma, S., Meaney, M.J., 2001. Naturally occurring variationsin maternal behavior in the rat are associated with differences in estrogen-inducible central oxytocin receptors. Proceedings of the National Academy ofSciences of the United States of America 98, 12736–12741.

Champagne, F.A., Francis, D.D., Mar, A., Meaney, M.J., 2003. Variations in maternal carein the rat as a mediating influence for the effects of environment on development.Physiology & Behavior 79, 359–371.

Champagne, D.L., Bagot, R.C., van Hasselt, F., Ramakers, G., Meaney, M.J., de Kloet, E.R.,Joels, M., Krugers, H., 2008. Maternal care and hippocampal plasticity: evidence forexperience-dependent structural plasticity, altered synaptic functioning, anddifferential responsiveness to glucocorticoids and stress. J Neurosci 28, 6037–6045.

Claessens, S.E., Daskalakis, N.P., van der Veen, R., Oitzl, M.S., de Kloet, E.R., Champagne,D.L., 2011. Development of individual differences in stress responsiveness: anoverview of factors mediating the outcome of early life experiences. Psychophar-macology 214, 141–154.

Dallman, M.F., Akana, S.F., Bhatnagar, S., Bell, M.E., Choi, S., Chu, A., Horsley, C., Levin, N.,Meijer, O., Soriano, L.R., Strack, A.M., Viau, V., 1999. Starvation: early signals,sensors, and sequelae. Endocrinology 140, 4015–4023.

D'Amato, F.R., Cabib, S., Puglisi-Allegra, S., Patacchioli, F.R., Cigliana, G., Maccari, S.,Angelucci, L., 1992. Effects of acute and repeated exposure to stress on thehypothalamo-pituitary-adrenocortical activity in mice during postnatal develop-ment. Hormones and Behavior 26, 474–485.

Daskalakis, N., Diamantopoulou, A., Tjälve, M., Meelis, W., Oitzl, M.S., de Kloet, E.R.,2009. Early life adversity load predicts psychosis-like behaviour in the rat.

Neuroscience Meeting Planner. Chicago, IL: Society for Neuroscience, 2009. Online,Program No. 341.349/O318.

Enthoven, L., Oitzl, M.S., Koning, N., van der Mark, M., de Kloet, E.R., 2008.Hypothalamic-pituitary-adrenal axis activity of newborn mice rapidly desensitizesto repeated maternal absence but becomes highly responsive to novelty.Endocrinology 149, 6366–6377.

Heim, C., Newport, D.J., Mletzko, T., Miller, A.H., Nemeroff, C.B., 2008. The link betweenchildhood trauma and depression: insights from HPA axis studies in humans.Psychoneuroendocrinology 33, 693–710.

Huot, R.L., Gonzalez, M.E., Ladd, C.O., Thrivikraman, K.V., Plotsky, P.M., 2004. Fosterlitters prevent hypothalamic-pituitary-adrenal axis sensitization mediated byneonatal maternal separation. Psychoneuroendocrinology 29, 279–289.

Knuth, E.D., Etgen, A.M., 2005. Corticosterone secretion induced by chronic isolation inneonatal rats is sexually dimorphic and accompanied by elevated ACTH. Hormonesand Behavior 47, 65–75.

Lehmann, J., Feldon, J., 2000. Long-term biobehavioral effects of maternal separation inthe rat: consistent or confusing? Reviews in the Neurosciences 11, 383–408.

Levine, S., Huchton, D.M., Wiener, S.G., Rosenfeld, P., 1991. Time course of the effect ofmaternal deprivation on the hypothalamic-pituitary-adrenal axis in the infant rat.Developmental Psychobiology 24, 547–558.

Liu, D., Diorio, J., Day, J.C., Francis, D.D., Meaney, M.J., 2000. Maternal care, hippocampalsynaptogenesis and cognitive development in rats. Nature Neuroscience 3,799–806.

Lupien, S.J., McEwen, B.S., Gunnar, M.R., Heim, C., 2009. Effects of stress throughout thelifespan on the brain, behaviour and cognition. Nature reviews 10, 434–445.

Macri, S., Wurbel, H., 2006. Developmental plasticity of HPA and fear responses in rats:a critical review of the maternal mediation hypothesis. Hormones and Behavior 50,667–680.

Macri, S., Mason, G.J., Wurbel, H., 2004. Dissociation in the effects of neonatal maternalseparations onmaternal care and the offspring's HPA and fear responses in rats. TheEuropean Journal of Neuroscience 20, 1017–1024.

McCormick, C.M., Kehoe, P., Kovacs, S., 1998. Corticosterone release in response torepeated, short episodes of neonatal isolation: evidence of sensitization. Interna-tional Journal of Developmental Neuroscience 16, 175–185.

Moffett, M.C., Vicentic, A., Kozel, M., Plotsky, P., Francis, D.D., Kuhar, M.J., 2007. Maternalseparation alters drug intake patterns in adulthood in rats. BiochemicalPharmacology 73, 321–330.

Moriceau, S., Sullivan, R.M., 2006. Maternal presence serves as a switch betweenlearning fear and attraction in infancy. Nature Neuroscience 9, 1004–1006.

Moriceau, S., Wilson, D.A., Levine, S., Sullivan, R.M., 2006. Dual circuitry for odor-shockconditioning during infancy: corticosterone switches between fear and attractionvia amygdala. The Journal of Neuroscience 26, 6737–6748.

Myers, M.M., Brunelli, S.A., Squire, J.M., Shindeldecker, R.D., Hofer, M.A., 1989. Maternalbehavior of SHR rats and its relationship to offspring blood pressures. Develop-mental Psychobiology 22, 29–53.

Okimoto, D.K., Blaus, A., Schmidt, M.V., Gordon, M.K., Dent, G.W., Levine, S., 2002.Differential expression of c-fos and tyrosine hydroxylase mRNA in the adrenalgland of the infant rat: evidence for an adrenal hyporesponsive period.Endocrinology 143, 1717–1725.

Plotsky, P.M., Thrivikraman, K.V., Nemeroff, C.B., Caldji, C., Sharma, S., Meaney, M.J.,2005. Long-term consequences of neonatal rearing on central corticotropin-releasing factor systems in adult male rat offspring. Neuropsychopharmacology 30,2192–2204.

Pryce, C.R., Seifritz, E., 2011. A translational research framework for enhanced validityof mouse models of psychopathological states in depression. Psychoneuroendo-crinology 36, 308–329.

Pryce, C.R., Bettschen, D., Bahr, N.I., Feldon, J., 2001a. Comparison of the effects of infanthandling, isolation, and nonhandling on acoustic startle, prepulse inhibition,locomotion, and HPA activity in the adult rat. Behavioral Neuroscience 115, 71–83.

Pryce, C.R., Bettschen, D., Feldon, J., 2001b. Comparison of the effects of early handlingand early deprivation onmaternal care in the rat. Developmental Psychobiology 38,239–251.

Raineki, C., Moriceau, S., Sullivan, R.M., 2010. Developing a neurobehavioral animal modelof infant attachment to an abusive caregiver. Biological Psychiatry 67, 1137–1145.

Rosenfeld, P., Gutierrez, Y.A., Martin, A.M., Mallett, H.A., Alleva, E., Levine, S., 1991.Maternal regulation of the adrenocortical response in preweanling rats. Physiology& Behavior 50, 661–671.

Rosenfeld, P., Suchecki, D., Levine, S., 1992a. Multifactorial regulation of thehypothalamic-pituitary-adrenal axis during development. Neuroscience andBiobehavioral Reviews 16, 553–568.

Rosenfeld, P., Wetmore, J.B., Levine, S., 1992b. Effects of repeated maternal separationson the adrenocortical response to stress of preweanling rats. Physiology & Behavior52, 787–791.

Ruedi-Bettschen, D., Feldon, J., Pryce, C.R., 2004. Circadian- and temperature-specificeffects of early deprivation on rat maternal care and pup development: short-termmarkers for long-term effects? Developmental Psychobiology 45, 59–71.

Schmidt, M.V., Oitzl, M.S., Levine, S., de Kloet, E.R., 2002. The HPA system during thepostnatal development of CD1 mice and the effects of maternal deprivation. BrainResearch. Developmental Brain Research 139, 39–49.

Schmidt, M., Enthoven, L., van Woezik, J.H., Levine, S., de Kloet, E.R., Oitzl, M.S., 2004.The dynamics of the hypothalamic-pituitary-adrenal axis during maternaldeprivation. Journal of Neuroendocrinology 16, 52–57.

Schmidt, M.V., Levine, S., Alam, S., Harbich, D., Sterlemann, V., Ganea, K., de Kloet, E.R.,Holsboer, F., Muller, M.B., 2006. Metabolic signals modulate hypothalamic-pituitary-adrenal axis activation during maternal separation of the neonatalmouse. Journal of Neuroendocrinology 18, 865–874.

176 N.P. Daskalakis et al. / Hormones and Behavior 60 (2011) 165–176

Stanton, M.E., Gutierrez, Y.R., Levine, S., 1988. Maternal deprivation potentiates pituitary-adrenal stress responses in infant rats. Behavioral Neuroscience 102, 692–700.

Suchecki, D., Rosenfeld, P., Levine, S., 1993. Maternal regulation of the hypothalamic-pituitary-adrenal axis in the infant rat: the roles of feeding and stroking. BrainResearch. Developmental Brain Research 75, 185–192.

Sullivan, R.M., Landers, M., Yeaman, B., Wilson, D.A., 2000a. Good memories of badevents in infancy. Nature 407, 38–39.

Sullivan, R.M., Stackenwalt, G., Nasr, F., Lemon, C., Wilson, D.A., 2000b. Association of anodor with activation of olfactory bulb noradrenergic beta-receptors or locuscoeruleus stimulation is sufficient to produce learned approach responses to thatodor in neonatal rats. Behavioral Neuroscience 114, 957–962.

van Oers, H.J., de Kloet, E.R., Li, C., Levine, S., 1998a. The ontogeny of glucocorticoidnegative feedback: influence of maternal deprivation. Endocrinology 139,2838–2846.

van Oers, H.J., de Kloet, E.R., Whelan, T., Levine, S., 1998b. Maternal deprivationeffect on the infant's neural stress markers is reversed by tactile stimulation andfeeding but not by suppressing corticosterone. The Journal of Neuroscience 18,10171–10179.

Vazquez, D.M., Akil, H., 1992. Development of pituitary pro-opiomelanocortin gene andpeptide expression: characterization and effect of repeated intermittent maternalisolation. Neuroendocrinology 56, 320–330.

Walker, C.D., 1995. Chemical sympathectomy and maternal separation affect neonatalstress responses and adrenal sensitivity to ACTH. The American Journal ofPhysiology 268, R1281–R1288.

Walker, C.D., Sapolsky, R.M., Meaney, M.J., Vale, W.W., Rivier, C.L., 1986. Increasedpituitary sensitivity to glucocorticoid feedback during the stress nonresponsiveperiod in the neonatal rat. Endocrinology 119, 1816–1821.

Walker, C.D., Akana, S.F., Cascio, C.S., Dallman, M.F., 1990. Adrenalectomy in theneonate: adult-like adrenocortical system responses to both removal andreplacement of corticosterone. Endocrinology 127, 832–842.

Walker, C.D., Scribner, K.A., Cascio, C.S., Dallman, M.F., 1991. The pituitary-adrenocor-tical system of neonatal rats is responsive to stress throughout development in atime-dependent and stressor-specific fashion. Endocrinology 128, 1385–1395.

Zhang, T.Y., Chretien, P., Meaney, M.J., Gratton, A., 2005. Influence of naturally occurringvariations in maternal care on prepulse inhibition of acoustic startle and the medialprefrontal cortical dopamine response to stress in adult rats. The Journal ofNeuroscience 25, 1493–1502.

Zimmerberg, B., Kim, J.H., Davidson, A.N., Rosenthal, A.J., 2003a. Early deprivation altersthe vocalization behavior of neonates directing maternal attention in a rat model ofchild neglect. Annals of the New York Academy of Sciences 1008, 308–313.

Zimmerberg, B., Rosenthal, A.J., Stark, A.C., 2003b. Neonatal social isolation alters bothmaternal and pup behaviors in rats. Developmental Psychobiology 42, 52–63.