Behavioural Brain Research 144 (2003) 199210

Research report

Early social stress in female guinea pigs induces a masculinization of adult behavior and corresponding changes in brain and neuroendocrine function
Sylvia Kaiser a, , Frank P.M. Kruijver b , Dick F. Swaab b , Norbert Sachser a
a b

Department of Behavioural Biology, University of Mnster, Badestr. 9, D-48149 Mnster, Germany Netherlands Institute for Brain Research, Meibergdreef 33, 1105 AZ Amsterdam ZO, The Netherlands

Received 10 January 2003; received in revised form 27 February 2003; accepted 27 February 2003

Abstract This study was undertaken to investigate, in guinea pigs, the effects of pre- and early postnatal social stress on the functioning of hormonal, autonomic, behavioral, and limbicbrain systems. Dams had either lived in groups with a constant composition (i.e. stable social environment) or in groups with changing compositions, that means every 3 days two females were transferred from one group to another (i.e. unstable social environment). The subjects studied were female offspring of dams who had either lived in a stable social environment during pregnancy and lactation (i.e. control daughters, CF) or in an unstable social environment during this period of life (i.e. early stressed daughters, SF). After weaning, each ve groups of CF and SF, consisting of two females each, were established. The spontaneous behavior of the females was recorded, blood samples were taken to determine cortisol, testosterone, dehydroepiandrosterone, dehydroepiandrosterone sulfate and estrogen levels, the adrenals were prepared to determine tyrosinehydroxylase (TH) activities and the brains to investigate the distribution of sex hormone receptors. SF showed not only a behavioral and endocrine masculinization, but also an upregulation of androgen receptor and estrogen receptor- in the medial preoptic area and the nucleus arcuatus of the hypothalamus, the nucleus paraventricularis of the thalamus, and the CA1 region of the hippocampus. These ndings corresponded with distinctly elevated serum-concentrations of testosterone and increased activities of the adrenal TH. In conclusion, early social stress caused by an unstable social environment induces in female guinea pigs a permanent behavioral masculinization that is accompanied by changes in the endocrine and autonomic system as well as by changes in the distribution of sex hormone receptors in the limbic system. 2003 Elsevier Science B.V. All rights reserved.
Keywords: Adrenal tyrosinehydroxylase activity; Androgen receptors; Estrogen receptor- ; Guinea pigs; Hippocampus; Hypothalamus; Prenatal stress; Social behavior

1. Introduction Stressors acting on the organism during the pre- and/or perinatal phase can have long-term effects on behavior, reproductive, and immune function, as well as on the endocrine system [42,43,82]. Brain development can also be distinctly affected by prenatal stress: e.g. the number of hippocampal glucocorticoid receptors is decreased [50], the neurotransmitter content, e.g. acetylcholine, dopamine, adrenaline is changed in specic brain areas [61], and the sexually dimorphic nucleus of the preoptic area is diminished in size
Corresponding author. Tel.: +49-251-8324676; fax: +49-251-8323896. E-mail address: kaisesy@uni-muenster.de (S. Kaiser).

[4]. In humans, prenatal stress is thought to be related to the occurrence of homosexuality [18] and the incidence of schizophrenia [72]. Stressful life events such as bereavement, child abuse, and early maternal separation are risk factors for depression [15,84]. In most experimental studies on the effects of early stress, pregnant females were subjected to nonsocial stressors (i.e. to regimens of heat, bright light, or restraint, e.g. [82]). Often a demasculinization and feminization of prenatally stressed male offspring and a reduced fertility of prenatally stressed female (SF) offspring are described [28,30,63,82]. In our recent work, we showed that the instability of the social environment during pregnancy and lactation represents a powerful social stressor that has important effects on the offsprings behavior and endocrine system: females

0166-4328/$ see front matter 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0166-4328(03)00077-9

200

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

whose mothers had lived in an unstable social environment during this period of life (which is experimentally induced by changing the group composition) displayed signicantly higher amounts of male-typical behaviors than females whose mothers had lived in a stable social environment [68]. The majority of the studies investigating prenatal stress work with rats. However, we have chosen guinea pigs for following reasons. (1) Rats are a species, which gives birth to immature young. Their neuroendocrine development occurs importantly in the postnatal period [16]. In contrast to rats, guinea pigs are precocial: they give birth to mature young and the peak in brain growth occurs during fetal development [70]. (2) Stress during pregnancy may alter the interaction between the female and her offspring. For example, in rats, mothers who were exposed to stressors during pregnancy display changed maternal behavior, which in turn can have distinct effects on the development of the offspring [59]. In the precocial guinea pig, however, the postnatal environment obviously does not play such an important role: the dependence of the offspring on maternal interaction postnatally is signicantly less than in altricial species such as rats. For example, in a recent study we could show that in guinea pigs the behavioral differences between the female offspring of females who had lived in a stable or an unstable social environment are due to the social instability during pregnancy, while the period of lactation does not seem to be of importance for this phenomenon [68]. Thus, we are dealing with a prenatal social stress model. The behavioral masculinization of prenatally stressed female guinea pigs corresponds with increased androgen levels and an elevated activity of the sympathetic adrenomedullary system [36]. In prenatally stressed males, opposite effects are found. When the mothers had lived in an unstable social environment, a behavioral infantilization was found that corresponds with a delayed development of the adrenocortical system and a decreased activity of the sympatheticadrenomedullary system [37]. Sex-specic behavioral patterns are mediated by sex hormones, which act at least partly on specic receptors in certain brain areas [10]. Androgens and/or estrogens are regarded as major candidates for such activational effects [45]. The receptors, namely estrogen receptor (ER) and androgen receptor (AR), have at least partly to be organized during the sensitive phase of sex differentiation [23,51]. Binding of androgens and/or estrogens to androgen or estrogen receptors of the limbic system regulates the display of masculine behavioral patterns [10]. Thus, one main effect of androgens is the activation of masculine behavior in adult male vertebrates. Since our recent studies showed a behavioral and hormonal masculinization of early stressed female guinea pigs we investigated in this study in the same species the effects of early stress not only on the spontaneous behavior, endocrine, and autonomic system, but also on the distribution of AR and ER- in specic areas of the limbic system.

2. Animals and methods 2.1. Subjects The guinea pigs (Cavia aperea f. porcellus) used were descendants of a heterogeneous short-haired and multicolored stock of 40 animals obtained from a breeder in 1975. The animals were bred by chance in closed breeding colonies (1050 animals). Natural markings permit individual recognition of all animals. Commercial guinea pig diet (Hveler Spezialfutter 1070 for guinea pigs, Hveler Spezialfutterwerke GmbH & Co KG, Langenfeld, Germany) and water were available ad libitum. This diet was supplemented regularly with fruits and hay. 2.2. Housing conditions All animals were housed under standard conditions in the same room: 12 h/12 h light/dark cycle, photoperiod 7:0019:00 h, temperature 20 2 C, relative humidity about 60%. The oors were covered with standard bedding material (wood shavings). Cage cleaning took place once a week. 2.3. Procedure The behavior, endocrine parameters as well as the distribution of ARs and ER- in certain brain areas of females whose mothers had lived in two different conditions were compared. These conditions were: (1) a stable social environment during pregnancy and lactation; and (2) an unstable social environment during this period of life. 2.3.1. Housing conditions of the dams All dams were living in groups of one adult male and ve adult females. The groups were established in the following way: at 20 days of age, which is the time around weaning, females were transferred from their natal colonies [67] to 1 m2 enclosures with a resident adult male (618 months of age) until ve nonrelated females were present. The age difference between the females of the same group did not exceed 42 days. At about 30 days of age, the females were mated for the rst time and gave birth to their rst litter after about 67 days of gestation. Thereafter, they were mated postpartum and had their second gestation. The offspring remained with their mothers until weaning (20 days of age) and then were taken out of the enclosures. Part of the groups remained in a constant group composition throughout the whole investigation time, i.e. these mothers had lived in a stable social environment during pregnancy and lactation. In the other groups, the composition of the females was regularly changed beginning with the onset of the rst females second period of pregnancy. The group composition was constant for not more than three successive days: every third day, two females from different groups were exchanged at

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

201

8:30 h. Since each group consisted of ve females, every 15th day the same female was transferred from one group to another. (Example: day 0: female 1 of group A and female 1 of group B were exchanged: 1A B/1B A; day 3: 2A B/2B A; . . . ; day 12: 5A B/5B A; day 15: 1A B/1B B.) In case of lactation, the females were transferred together with their offspring. That is, these females had lived in an unstable social environment. We were careful to catch and handle animals living in the stable and the unstable social environment in the same way at corresponding times. 2.3.2. Housing conditions of the female offspring The experiments were done with 20 females originating from the second litters of mothers who had either lived in a stable or an unstable social environment (litter size: two to six; females/litter: one to four). The following groups were established, consisting of two females each, which were kept in 0.5 m2 enclosures: (a) ve groups of two females each whose mothers had lived in a stable social environment (n = 10); (b) ve groups of two females each whose mothers had lived in an unstable social environment (n = 10). The two females living in the same group came from different natal groups; thus, they were unrelated and unfamiliar with each other. They were placed into the group at 20 days of age. The age difference between the two females of one group did not exceed 8 days. 2.4. Behavioral measures From the 20th to 80th day of age (female guinea pigs become sexually mature around 30 days of age), every females behavior was recorded by videotape. The observation time was divided into three phases: the 20th40th, 41st60th, and 61st80th day of age, respectively. During each of these phases, every females behavior was recorded ve to six times for 1 h between 8:00 and 10:00 h. The behavior was recorded with the focal animal sampling and continuous recording method [53]. The categorization of the behavioral patterns have been described previously [36,37,48]. Courtship behavior: Intensive anogenital snifng (one animal is licking, snifng, or nuzzling another animals anogenital region intensively); rumba (slowly approaching an animal, rhythmically oscillating the hindquarters from side to side, and emitting a burbling vocalization (the rumble); the head stretches forward, held parallel to the ground, and as the displaying animal nears the female its body may assume a curve). Agonistic behavior: Retreat (directed movement away from the opponent at a walking or running pace); face-off (facing the opponent); curved body posture (standing side on in front of the opponent; the body is curved; head and

hindquarters directed to the opponent); attack lunge (a short jump at the opponent). Contact behavior: Naso-nasal snifng (one animal is snifng another animals nasal region). Play behavior: Frisky hops; run off. 2.5. Blood sampling, hormone determination, and histology Each 10 females, whose mothers had lived in a stable social environment during pregnancy and lactation (i.e. control females) and 10 females, whose mothers had lived in an unstable social environment during pregnancy and lactation (i.e. early stressed females) were sacriced at the age of 100 5 days. No female was in estrus at the time of sacrice. Most caviomorph rodents like guinea pigs have a vaginal closure membrane which is perforated only at estrus and parturition [64,80]. Thus, in guinea pigs, the condition of the vaginal membrane can be used as an external indication of estrus. The guinea pigs were weighed, anesthetized with dry ice, and decapitated between 12:00 and 13:00 h. After decapitation, the following procedures were conducted. (a) Blood samples were taken within 3 min after decapitation to determine serum levels of cortisol, testosterone, dehydroepiandrosterone (DHEA), dehydroepiandrosteronesulfate (DHEAS), and estrogen levels. The serum was separated by centrifugation (11, 752 g for 5 min) and deep frozen (20 C) until assayed. Serum levels of cortisol and testosterone were determined by radioimmunoassay without chromatography using specic antibodies against cortisol (Biotrend, Kln, Germany) and testosterone (kindly donated by Dr. Eberhard Nieschlag, Mnster, Germany), respectively. The intraassay and interassay coefcients of variation of cortisol were 4.8 and 8.7%, respectively, and of testosterone 6.2 and 9.1%, respectively (for further details, see [3638]). DHEA, DHEAS, and estrogen levels were determined by an enzyme immunoassay (EIA; [57,77]). The intraassay and interassay coefcients of variation were below 10% in all three EIAs. (b) The adrenals were dissected immediately after decapitation, rapidly cleaned from fat and adherent tissue and frozen in 5 mM TrisHCl buffer (pH 7.2). For analysis of adrenal tyrosinehydroxylase (TH) activity, a radioenzymatic assay has been carried out (for further details, see [36,37]). (c) Immediately after decapitation, the brains were removed and dissected (the cerebellum was removed). The dissected brains included the hypothalamic, thalamic, and hippocampal area that was subsequently xated in 4% formaldehyde at room temperature for 7 days. After the xation, the brains were dehydrated and embedded in parafn. Serial, 6 m frontal sections were cut on a Leitz-microtome (Rockleigh, NJ).

202

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

2.6. Immunohistochemistry Along the rostrocaudal axis, each 25th frontal parafnembedded section of the dissected part of the guinea pig brains was stained for ER- and adjacent sections for AR. Additionally, adjacent sections were stained with thionin to aid in the outlining of cell group boundaries.

The parafn-embedded sections were mounted on SuperFrost/Plus slides and dried overnight on a hot plate at 49 C followed by about 36 h in an oven at 37 C. The staining protocol for ER- was performed as described in [31] with slight modications: deparafnization in xylene and rehydration by a series of decreasing ethanol concentrations, rinsing in distilled water twice for 5 min each time,

Fig. 1. AR and ER- -nucleus-positive neurons in percent in the nucleus arcuatus of the hypothalamus. Values are given as medians and individual values. CF: females whose mothers had lived in a stable social environment during pregnancy and lactation; SF: females whose mothers had lived in an unstable social environment during pregnancy and lactation. Statistic: MannWhitney U-Test (two-tailed); n1 = n2 = 10; P < 0.001.

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

203

rinsing in Tris-containing buffered saline (TBS)-highsalt (pH 7.6) twice for 5 min each time, water bath treatment in 0.05 mol/l TrisHCl buffer (pH 7.6) for 30 min at 90 C, washing in TBS-highsalt for 30 min (3 10 min), incubation in milk-TBS for 1 h at room temperature, washing in TBS-highsalt twice 5 min each time, and incubation with the primary antibody (MC-20, Santa Cruz) in Supermix (0.25% gelatin, 0.5% Triton X-100 in TBS, pH 7.6) for 1 h at room temperature and at 4 C overnight. The next day, the sections were washed thrice in TBS-milk for 10 min each time, washed in TBS-highsalt for 5 min, incubated with the secondary biotinylated antirabbit IgG 1:200 in Supermix for 1 h at room temperature, washed in TBS-milk thrice for 10 min each time, washed in TBS-highsalt for 5 min, incubated with

ABC (Elite kit, Vector Laboratories Inc.) 1:800 in Supermix for 1 h at room temperature, rinsed in 0.05 mol/l TrisHCl (pH 7.6), incubated in TrisHCl containing 0.05 mg/ml 3,3 -diaminobenzidine, 0.01% H2 O2 , and 0.2% nickel ammonium sulfate, washed in TrisHCl twice for 10 min each time, dehydrated in graded ethanols, cleared in xylene, and coverslipped with Entellan mounting medium. The used primary polyclonal rabbit ER- antibody (MC-20) is recognizing the carboxy terminus epitope (i.e. the C-terminal domain) of the ER- . The specicity of this ER- antibody has been characterized and documented by Clarke et al. [13] and Fitzpatrick et al. [21], showing that the antibody MC-20 specically recognizes ER- and does not cross-react with ER- . Furthermore, omission of the

Fig. 2. Androgen receptor-immunoreactivity (AR-ir) in certain brain areas. (A and B) AR-ir in the nucleus arcuatus of an early stressed female (SF; A) and a control female (CF; B); (C and D) AR-ir in the medial preoptic area of an SF (C) and a CF (D); (E and F) AR-ir in the CA1 of an SF (E) and a CF (F). Bar represents 40 m.

204

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

primary antibody revealed no staining in our guinea pig sections. The detection of AR was performed as described earlier with slight modications [20,45]: deparafnization in xylene and rehydration by a series of decreasing ethanol concentrations, rinsing in distilled water twice for 5 min each time, rinsing in TBS (pH 7.6) twice for 5 min each time, microwave treatment in 0.1 mol/l citrate buffer (pH 6.0) at 90 C for 10 min, washing in TBS for 30 min (3 10 min), incubation in milk-TBS for 1 h at room temperature, washing in TBS for 5 min, and incubation with the primary antibody (PG-21, kindly donated by Drs. Gail Prins and Geoffrey Greene, Chicago) in Supermix for 1 h at room temperature and at 4 C for 3 days. The next day, the sec-

tions were washed thrice in TBS-milk for 10 min each time, washed in TBS for 5 min, incubated with the secondary biotinylated antirabbit IgG 1:200 in Supermix for 1 h at room temperature, washed in TBS-milk thrice for 10 min each time, washed in TBS for 5 min, incubated with ABC 1:800 in supermix for 1 h at room temperature, rinsed in 0.05 mol/l TrisHCl (pH 7.6). Subsequently, a signal-amplication method based on the deposition of biotinylated tyramine was performed. This method has been described in detail by Adams [1]. Briey, after washing with TBS, sections were incubated with biotinylated tyramine diluted 1:1000 and 0.01% H2 O2 for 20 min, after which they were rinsed with TBS. The ABC procedure was then repeated, and the slides were subsequently rinsed in 0.05 mol/l TrisHCl

Fig. 3. Estrogen receptor- -immunoreactivity (ER- -ir) in certain brain areas. (A and B) ER- -ir in the nucleus arcuatus of an early stressed female (SF; A) and a control female (CF; B); (C and D) ER- -ir in the medial preoptic area of an SF (C) and a CF (D); (E and F) ER- -ir in the CA1 of an SF (E) and a CF (F). Bar represents 40 m.

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

205

(pH 7.6). Sections were incubated for 10 min in 0.05 M TrisHCl containing 0.05 mg/ml 3,3 -diaminobenzidine, 0.01% H2 O2 , and 0.2% nickel ammonium sulfate. After developing, sections were washed in TrisHCl twice for 10 min each time, dehydrated in graded ethanols, cleared with xylene, and coverslipped with Entellan mounting medium. The primary antibody PG21 is raised in rabbits and directed against a synthetic peptide corresponding to the nal 21 amino acids of the N-terminus of the human AR. Literature data show the specicity of the anti-AR antibody

for brain immunohistochemical studies [58,62,85]. Furthermore, parafn-embedded sections of formalin-xed mouse and human testes show clear AR staining in the peritubalar cells as previously reported [35,71]. In our study, omitting the AR antibody PG21 totally prevented staining. 2.7. Image analysis We analyzed the number of neurons with nucleolus in the medial preoptic area (MPOA) and the nucleus arcuatus (ARC) of the hypothalamus as well as in the nucleus

Fig. 4. AR- and ER- -nucleus-positive neurons in percent in the medial preoptic area of the hypothalamus. Values are given as medians and individual values. CF: females whose mothers had lived in a stable social environment during pregnancy and lactation; SF: females whose mothers had lived in an unstable social environment during pregnancy and lactation. Statistic: MannWhitney U-Test (two-tailed); n1 = n2 = 10; P < 0.001.

Fig. 5. AR- and ER- -nucleus-positive neurons in percent in the pyramidal layer of the CA1-region of the hippocampus. Values are given as medians and individual values. CF: females whose mothers had lived in a stable social environment during pregnancy and lactation; SF: females whose mothers had lived in an unstable social environment during pregnancy and lactation. Statistic: MannWhitney U-Test (two-tailed); n1 = n2 = 5; P < 0.05; P < 0.001.

206

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

paraventricularis of the thalamus (PVT) with nuclear, cytoplasmic, or both forms of staining for ER- and AR. Cytoplasmic staining of both ER- and ARs is described also in other studies [9,20,45,46,85]. We also analyzed the numbers of neurons negative for ER- and AR staining [31]. It should also be noted that the unstained neurons, nuclei, and nucleoli could well be recognized in the sections. Neurons that show only nuclear or nuclear and cytoplasmic staining were designated nucleus-positive cells. We calculated the percentage of nucleus-positive cells. Additionally, we analyzed the number of stained neurons in the pyramidal layer of the CA1, CA2, and CA3 region as well as in the hilus and the dentate gyrus of the hippocampus of ve CF and ve SF. In guinea pigs, a structural sex different region in the hypothalamic MPOA is described which resembles the sexually dimorphic nucleus of the preoptic area (SDN-POA) of the rat [29]. However, in our 6 m thick sections, we could not identify the SDN-POA of the guinea pig. For this reason, we made no distinction between the SDN-POA and the surrounding MPOA area. Instead, we examined the entire MPOA. For the analysis, we used an IBAS image-analysis system (Kontron Instruments Ltd., Munich, Germany). This procedure was previously extensively described and illustrated by Kruijver et al. [44]. Briey, the image analysis system was connected to a scanning stage control box (MCU, Carl Zeiss, Oberkochem, Germany) and had a Sony B/W CCD-camera for image acquisition. Both the scanning stage and the camera were mounted on a microscope (Carl Zeiss) equipped with planabo objectives. To provide optimal contrast and homogenous illumination of the section, the voltage of the light source was set maximally. The light was reduced by neutral grey lters to improve light contrast. For each section, the analysis consisted of the following steps: in each section an area covering one of the regions which should be analyzed (using the 2.5 objective of microscope) was loaded into the IBASimagememory. In this image, the specic areas were outlined manually. Subsequently, the image analyzer covered the outlined area with a grid of rectangular elds, each with the size of the area displayed by the camera when the 63 objective was installed. From this grid, all elds were selected. Taking into account the aberration of the optical axis between the 2.5 and the 63 objective, the pixel positions of the rectangular elds in the 2.5 image were converted into scanning stage coordinates to position the corresponding areas in front of the camera when using the 63 objective. After the 63 objective was installed, the image analyzer moved the scanning stage automatically to the previously dened positions of the high magnication measuring areas. 2.8. Statistics Physiological data are represented as mean S.E.M. Frequencies of behavioral data are given as medians and ranges.

Percentage of sex hormone receptors are represented as medians and individual values. We tested differences between two independent groups with the two-tailed MannWhitney U-Test. The differences were considered signicant if their probability of occurring by chance was less than 0.05.

3. Results 3.1. Behavior Females whose mothers had lived in a stable social environment during pregnancy and lactation (CF) differed signicantly in specic behavioral patterns from females whose mothers had lived in an unstable social environment during this period of life (SF): SF displayed distinctly higher amounts of play behavior (frisky hops and run off), of a contact behavioral pattern (naso-nasal snifng), and of two male-typical courtship behavioral patterns (rumba, intensive anogenital snifng) than CF (Table 1). No differences were found between SF and CF with respect to agonistic behavioral patterns (data not shown). 3.2. Endocrinology The behavioral alterations were accompanied by differences in the endocrine and autonomic systems: SF showed higher plasma concentrations of testosterone and higher activities of the adrenal TH than CF (Table 2). Plasma concentrations of cortisol, DHEA, DHEAS, and estrogen concentrations as well as body weights did not differ between the two categories of females (Table 2). 3.3. ARs and ER- In the ARC, SF showed an upregulation of AR compared to CF (Figs. 1, and 2A and B). The percentage of ER- -nucleus-positive cells in the ARC was also higher in SF than in CF (Figs. 1, and 3A and B). Furthermore, early social stress altered the percentage of AR-nucleus-positive cells in the MPOA: SF showed again an upregulation of AR (Figs. 2C and D, and 4). Additionally, there was a tendency in SF for an upregulation of ER- (Figs. 3C and D, and 4). In the PVT, we found differences between the females, too: ER- as well as AR was upregulated in SF (ERCF median, 83.7%; SF, 92.4%; U = 9; P < 0.01; AR: CF median, 56.1%; SF, 83.7%; U = 8; P < 0.01). Because of these striking results, we decided to investigate additionally the expression of AR and ER- in the hippocampus of 10 animals (each ve CF and SF). In the pyramidal layer of the CA1, we found in SF an upregulation of AR and ER- compared to CF (Figs. 2E and F, 3E and F, and 5). We found also AR- and ER- -expression in the pyramidal layers of CA2, CA3, the dentate gyrus and hilus of the hip-

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210 Table 1 Behavioral parameters in early stressed and control females Days of age 2040 Play behavior Naso-nasal snifng Rumba Anogenital snifng 4160 Play behavior Naso-nasal snifng Rumba Anogenital snifng 6180 Play behavior Naso-nasal snifng Rumba Anogenital snifng 15.67 (3.730.5) 2.42 (1.54.5) 0 (00.67) 0 (00.17) 13.82 (6.327.8) 2.92 (1.56.6) 0.25 (00.72) 0.42 (0.172.0) 46 32 18 1 4.83 (2.715) 1.75 (0.832.3) 0 (00) 0 (00.33) 16.08 (8.324.8) 4.42 (1.85.2) 0.45 (02.0) 0.42 (0.172.0) 4.42 (1.510.7) 1.25 (0.502.2) 0 (00) 0 (00.17) 14.25 (10.327.8) 3.42 (1.06.2) 0 (00) 0 (01.2) 1 9 50 41 Behavioral patterns CF SF U (P)

207

(<0.001) (<0.01) (1.0) (0.44)

8 (<0.01) 8.5 (<0.01) 9 (<0.01) 4 (<0.001) (0.72) (0.28) (<0.05) (<0.001)

Data are given as medians and ranges (in parenthesis) in frequencies per hour. CF (control females): females whose mothers had lived in a stable social environment during pregnancy and lactation; SF (early stressed females): females whose mothers had lived in an unstable social environment during pregnancy and lactation. Statistic: MannWhitney U-Test (two-tailed); n1 = n2 = 10. Table 2 Body weights, adrenal TH activities as well as plasma concentrations of steroid hormones in early stressed and control females Parameter Tyrosine hydroxylase activity (nmol/h/2 adrenals) Cortisol (ng/ml) Testosterone (pg/ml) DHEA (ng/ml) DHEAS (ng/ml) Estrogen (pg/ml) Body weight (g) CF 26.4 352 110 0.32 64 22.3 535 1.5 78.5 8.2 0.013 1.6 2.5 10.5 SF 36.4 373 440 0.29 91.4 21.3 583 1.5 54.9 104 0.015 28.9 1.4 24.4 U (P) 3 (<0.001) 38 (0.39) 10 (<0.01) 30 (0.14) 44.5 (0.67) 48 (0.91) 31 (0.17)

Data are given as means S.E.M. CF (control females): females whose mothers had lived in a stable social environment during pregnancy and lactation; SF (early stressed females): females whose mothers had lived in an unstable social environment during pregnancy and lactation. Statistic: MannWhitney U-Test (two-tailed); n1 = n2 = 10.

pocampus of CF and SF, which were not different (data not shown).

4. Discussion This study shows that early social stress has distinct masculinizing effects on the females behavior, endocrine system, autonomic system, and distribution of sex hormone receptors in various brain areas. When the mothers had lived in an unstable social environment during pregnancy this was resulting in a behavioral activation of the pregnant females [39]then their female offspring (SF) displayed signicantly higher amounts of rumba, intensive naso-anal snifng, naso-nasal snifng as well as frisky hops and run off than female offspring of mothers who had lived in a stable social environment (CF). These results are similar to the results of a former study [68]. The behavioral differences point to a masculinization of SF: intensive naso-anal snifng and rumba are the most conspicuous parts of the courtship

behavior which in mixed-sex groups of guinea pigs is exclusively directed from males to females [33,66]. Further on, frisky hops and run off are the main elements of the guinea pigs play behavior [48] and it is known from different mammalian species that playing is displayed to a distinctly higher amount by males than by females [55]. SF had a higher amount of AR-nucleus-positive cells in certain brain areas than CF. This expression pattern of AR is similar to that of male guinea pigs: male guinea pigs show a higher AR expression compared to females, e.g. in the MPOA ([2]; personal observations). It seems likely that the male-typical patterns of AR expression in the MPOA and ARC are related to the behavioral and endocrine masculinization of early stressed females, since these brain areas are well known to play an essential role in the control of masculine behaviors [26] and the regulation of gonadotropin releasing factors [20]. In addition, we found in the MPOA, ARC, and PVT of SF a higher amount of ER- -nucleus-positive cells compared to CF. Contrary to the early stressed females male-typical

208

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

expression pattern of AR, their expression pattern of ER- is different from those of male guinea pigs. In male guinea pigs, much less ER- -nucleus-positive cells are found than in SF as well as in CF (own observations). Such sex differences, namely that males show less ER in different brain areas than females are also described in other studies (guinea pig [11]; rat [47]). Thus, whether there is a correlation between this upregulation of ER- and the described behavioral and endocrine masculinization is not yet clear. The different expression patterns of AR and ER- in specic brain areas of SF and CF may be explained by organizational effects of sex hormones during the sensitive phase of brain development [51], by an activational effect of sex hormones in development [45] or by a combination of these mechanisms. In order to result in sex-specic behavior patterns, the prenatally organized AR and ER- have to be activated by hormones in adulthood [10]. Androgens and/or estrogens are regarded as major candidates for such activational effects [45]. Indeed, serum testosterone concentrations were higher in early stressed females than in control females. However, the concentrations were still much lower than in males [69]. In female mammals, testosterone can be produced in the adrenal cortex as well as in the ovaries [5,24]. In our study, we cannot determine the source of androgens: the elevated testosterone concentrations could be secreted by the adrenal cortex as well as by the ovaries. Probably the behavioral differences between early stressed and control offspring are due to an interplay of an organizing effect of androgens during embryogenesis and an activational effect of androgens during adulthood. However, confusingly we found in pregnant females living in an unstable social environment decreased dehydroepiandrosterone(sulfate) levels compared to pregnant females living in a stable social environment [39]. Thus, it is of interest to investigate the concentrations of other androgens (e.g. androstenedione) in pregnant females living in different social environments as well as in the fetus. An alternative explanation of our ndings includes the observation that adrenal TH activities differed signicantly between stressed individuals and controls. This enzyme catalyses the rate-limiting step of the catecholamine synthesis and adrenal TH activity represents a reliable indicator of the organisms sympathetic-adrenomedullary activation [19,27,81]. Interestingly, catecholamines are known to activate steroid receptors, e.g. dopamine can mimic the effects of progesterone by stimulating the progesterone receptors [52]. In our study, we found an upregulation of AR and ER- in the ARC, MPOA, PVT, and CA1 in early stressed and later masculinized females as well as much higher TH activity compared to control females. In contrast male guinea pigs, whose mothers had lived in an unstable social environment during pregnancy, show a downregulation of AR in the ARC and MPOA and of ER- the CA1 as well as distinctly lower adrenal TH activities compared to control males [40]. Thus, it seems possible that the sex-specic and sex-reversed effects of early stress on behavior are due

to an activational effect of catecholamines on prenatally organized AR and ER- receptors. Of course, different mechanisms may play a role in different brain areas. Several studies describe effects of prenatal stress on the hippocampus: the hippocampal weight is reduced in prenatal stressed rats [78], neurogenesis can be impaired [6,49], and acetylcholine release in the hippocampus can be inuenced [14]. The present study is the rst to demonstrate an effect of prenatal stress on the expression of steroid, androgen and estrogen, receptors in the hippocampus, with early stress increasing AR and ER- . Also, other studies found AR in the hippocampus of several species [8,12,41,71,86]. Furthermore, ER- expression in the hippocampus is described [7,13,32,73,75]. However, regarding ER- -immunoreactivity in the pyramidal neurons of the rat CA1, there are contrary results: some studies [83] nd no immunoreactivity whereas others [54,73,75,79] report immunoreactive cells in the pyramidal layer of the CA1CA3. The hippocampus plays a central role in the regulation of vegetative functions [17,34] as well as in learning and memory [49,60,76]. In addition, estrogen inuences several processes associated with hippocampal functions like learning behavior [83]. Furthermore, it has been found that androgenic compounds can inuence some hippocampal mediated learning and memory tasks in rats [22,65]. Consequently, expression of sex hormone receptors in the hippocampus may be important for the ability of learning and of memory [73,74]. Interestingly, it is shown in rats and mice that prenatal stress can inuence learning behavior [3,25,49,56,78]). Thus, in a future study, it may be of interest to compare spatial learning and memory between SF and CF also in the guinea pig. In conclusion, the present study shows very clearly that early social stress causes a permanent behavioral masculinization of the females which corresponds with the activity of the endocrine and autonomic system as well as with the distribution of sex hormone receptors in specic brain areas of the limbic system. It would be interesting to investigate also the distribution of AR and ER- in other brain areas like the amygdala and the bed nucleus of the stria terminalis, which are also known to be involved in the regulation of social behavior.

Acknowledgements We would like to acknowledge Dr. M. Heistermann for determining estrogen concentrations, Dr. R. H. Straub for determining the DHEA(S) concentrations, and Bart Fisser for cutting the brains. Research was supported by the Deutsche Forschungsgemeinschaft (KA 1546/1-1), the Ethologische Gesellschaft, and the Gesellschaft zur Frderung der Westflischen Wilhelms-Universitt zu Mnster e.V.. All experiments were announced to the competent local authority and were approved by the Tierschutzbeauftragter of the University of Mnster.

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210

209

References
[1] Adams JC. Biotin amplication of biotin and horseradish peroxidase signals in histochemical stains. J Histochem Cytochem 1992;40:145763. [2] Ahdieh H, Feder HH. Sex differences in nuclear androgen receptors in guinea pig brain and the effects of an alpha2 noradrenergic blocker on androgen receptors. Brain Res 1988;456:2759. [3] Aleksandrov AA, Polyakova ON, Batuev AS. The effects of prenatal stress on learning in rats in a morris maze. Neurosci Behav Physiol 2001;31:714. [4] Anderson DK, Rhees RW, Fleming DE. Effects of prenatal stress on differentiation of the sexually dimorphic nucleus of the preoptic area (SDN-POA) of the rat brain. Brain Res 1985;332:113 8. [5] Askari HA. Sexual differences in the biogenesis of the androgens by the adrenal cortex in rat. Endocrinology 1970;87:137780. [6] Avishai-Eliner S, Brunson K, Sandman C, Baram T. Stressed-out, or in (utero)? Trends Neurosci 2002;25:51824. [7] Bettini E, Pollio G, Santagati S, Maggi A. Estrogen receptor in rat brain: presence in the hippocampal formation. Neuroendocrinology 1992;56:5028. [8] Beyenburg S, Watzka M, Clusmann H, Blumcke I, Bidlingmaier F, Elger CE, et al. Androgen receptor mRNA expression in the human hippocampus. Neurosci Lett 2000;294:258. [9] Blaustein JD, Turcotte JC. Estrogen receptor-immunostaining of neuronal cytoplasmic processes as well as cell nuclei in guinea pig brain. Brain Res 1989;495:7582. [10] Breedlove SM. Sexual differentiation of the brain and behaviour. In: Becker JB, Breedlove SM, Crews D, editors. Behavioral endocrinology. Cambridge: MIT Press; 1992. p. 3968. [11] Brown TJ, Yu J, Gagnon M, Sharma M, MacLusky NJ. Sex differences in estrogen receptor and progestin receptor induction in the guinea pig hypothalamus and preoptic area. Brain Res 1996;725:37 48. [12] Choate JV, Slayden OD, Resko JA. Immunocytochemical localization of androgen receptors in brains of developing and adult male rhesus monkeys. Endocrine 1998;8:5160. [13] Clarke CH, Noreet AM, Clarke MSF, Watson CS, Cunningham KA, Thomas ML. Perimembrane localization of the estrogen receptor protein in neuronal processes of cultured hippocampal neurons. Neuroendocrinology 2000;71:3442. [14] Day JC, Koehl M, Deroche V, Le Moal M, Maccari S. Prenatal stress enhances stress- and corticotropin-releasing factor-induced stimulation of hippocampal acetylcholine release in adult rats. J Neurosci 1998;18:188692. [15] De Bellis MD, Baum AS, Birmaher B, Keshavan MS, Eccard CH, Boring AM, et al. Developmental traumatology. Part I: biological stress systems. Biol Psychiatry 2000;45:125970. [16] Dean F, Yu C, Lingas RI, Matthews SG. Prenatal glucocorticoid modies hypothalamo-pituitary-adrenal regulation in prepubertal guinea pigs. Neuroendocrinology 2001;73:194202. [17] Eichenbaum H, Otto T. The hippocampuswhat does it do? Behav Neural Biol 1992;57:236. [18] Ellis L, Ames MA, Peckham W, Burke D. Sexual orientation of human offspring may be altered by severe maternal stress during pregnancy. J Sex Res 1988;25:1527. [19] Ely DL, Henry JP. Neuroendocrine response patterns in dominant and subordinate mice. Horm Behav 1978;10:15669. [20] Fernndez-Guasti A, Kruijver FPM, Fodor M, Swaab DF. Sex differences in the distribution of androgen receptors in the human hypothalamus. J Comp Neurol 2000;425:42235. [21] Fitzpatrick SL, Funkhouser JM, Sindoni DM, Stevis PS, Deecher DC, Bapat AR, et al. Expression of estrogen receptor- protein in rodent ovary. Endocrinology 1999;140:258191.

[22] Flood JF, Morley JE, Roberts E. Memory-enhancing effects in male mice of pregnenolone and steroids metabolically derived from it. Proc Natl Acad Sci USA 1992;89:156771. [23] Forest MG. Role of androgens in fetal and pubertal development. Horm Res 1983;18:6983. [24] Frank LG, Davidson JM, Smith ER. Androgen levels in the spotted hyeana Crocuta crocuta: the inuence of social factors. J Zool 1985;206:52531. [25] Fujioka T, Fujioka A, Tan N, Chowdhury GMI, Mouri H, Sakata Y, et al. Mild prenatal stress enhances learning performance in the non-adopted rat offspring. Neuroscience 2001;103:3017. [26] Grco B, Edwards DA, Michael RP, Clancy AN. Androgen receptors and estrogen receptors are colocalized in male rat hypothalamic and limbic neurons that express fos immunoreactivity induced by mating. Neuroendocrinology 1998;67:1828. [27] Henry JP, Stephens PM. Stress, health and the social environment. A sociobiological appropach to medicine. New York: Springer; 1977. [28] Herrenkohl LR. Prenatal stress reduces fertility and fecundity in female offspring. Science 1979;206:10979. [29] Hines M, Davis FC, Coquelin A, Goy RW, Gorski RA. Sexually dimorphic regions in the medial preoptic area and the bed nucleus of the stria terminalis of the guinea pig brain: a description and an investigation of their relationship to gonadal steroids in adulthood. J Neuroscience 1985;5:407. [30] Holson RR, Gough B, Sullivan P, Badger T, Sheehan DM. Prenatal dexamethasone or stress but not ACTH or corticosterone alter sexual behavior in male rats. Neurotoxicol Teratol 1995;17:393401. [31] Ishunina TA, Kruijver FPM, Balesar R, Swaab DF. Differential expression of estrogen receptor and immunoreacivity in the human supraoptic nucleus in relation to sex and aging. J Clin Endocrinol Metab 2000;85:328391. [32] Ivanova T, Beyer C. Ontogenic expression and sex differences of aromatase and estrogen receptor- / mRNA in the mouse hippocampus. Cell Tissue Res 2000;300:2317. [33] Jacobs WW. Malefemale associations in the domestic guinea pig. Anim Learn Behav 1976;4:7783. [34] Jacobson L, Sapolsky R. The role of the hippocampus in feedback regulation of the hypothalamicpituitaryadrenocortical axis. Endocr Rev 1991;12:11834. [35] Janssen PJ, Brinkmann AO, Boersma WJ, van der Kwast TH. Immunohistochemical detection of the androgen receptor with monoclonal antibody F39.4 in routinely processed parafn-embedded human tissues after microwave pretreatment. J Histochem Cytochem 1994;42:116975. [36] Kaiser S, Sachser N. The social environment during pregnancy and lactation affects the female offsprings endocrine status and behaviour in guinea pigs. Physiol Behav 1998;63:3616. [37] Kaiser S, Sachser N. Social stress during pregnancy and lactation affects in guinea pigs the male offsprings endocrine status and infantilizes their behaviour. Psychoneuroendocrinology 2001;26:50319. [38] Kaiser S, Brendel H, Sachser N. Effects of ACTH applications during pregnancy on the female offsprings endocrine status and behavior in guinea pigs. Physiol Behav 2000;70:15762. [39] Kaiser S, Heemann K, Straub R, Sachser N. The social environment affects behaviour and androgens, but not cortisol in pregnant female guinea pigs. Psychoneuroendocrinology 2003;28:6783. [40] Kaiser S, Kruijver FPM, Straub R, Sachser N, Swaab DF. Early social stress in male guinea pigs induces permanent changes in adult social behaviour, brain and neuroendocrine functions, submitted for publication. [41] Kerr JE, Allore RJ, Beck SG, Handa R. Distribution and hormonal regulation of androgen receptor (AR) and AR messenger ribonucleid acid in the rat hippocampus. Endocrinology 1995;136:321321. [42] Klein SL, Rager DR. Prenatal stress alters immune function in the offspring of rats. Dev Psychobiol 1995;28:32136.

210

S. Kaiser et al. / Behavioural Brain Research 144 (2003) 199210 [66] Sachser N. Different forms of social organization at high and low population densities in guinea pigs. Behaviour 1986;97:25372. [67] Sachser N. Social organization, social status, behavioural strategies and endocrine responses in male guinea pigs. In: Balthazart J, editor. Hormones, brain and behaviour in vertebrates. Comparative physiology, vol. 9. Basel: Karger; 1990. p. 17687. [68] Sachser N, Kaiser S. Prenatal social stress masculinizes the females behaviour in guinea pigs. Physiol Behav 1996;60:58994. [69] Sachser N, Prve E. Plasma-testosterone development in colony and individually housed male guinea pigs. Ethology 1988;79:6270. [70] Sapolski RM, Meaney MJ. Maturation of the adrenocortical stress response: neuroendocrine control mechanisms and the stress hyporesponsive period. Brain Res Rev 1986;396:6476. [71] Sar M, Lubahn DB, French FS, Wilson EM. Immunohistochemical localization of the androgen receptor in rat and human tissues. Endocrinology 1990;127:31806. [72] Selten J-P, Van der Graaf Y, Van Duursen R, Gispen-de Wied CC, Kahn RS. Psychotic illness after prenatal exposure to the 1953 dutch ood disaster. Schizophr Res 1999;35:2435. [73] Shugrue PJ, Merchenthaler I. Estrogen is more than just a sex hormone: novel sites for estrogen action in the hippocampus and cerebral cortex. Front Neuroendocrinol 2000;21:95101. [74] Shugrue PJ, Merchenthaler I. Evidence for novel estrogen binding sites in the rat hippocampus. Neuroscience 2000;99:60512. [75] Solum DT, Handa RJ. Localization of estrogen receptor alpha (ER alpha) in pyramidal neurons of the developing rat hippocampus. Dev Brain Res 2001;128:16575. [76] Squire LR. Mechanisms of memory. Science 1986;232:16129. [77] Straub RH, Konecna L, Hrach S, Rothe G, Kreutz M, Schlmerich J, et al. Serum dehydroepiandrosterone (DHEA) and DHEA sulfate are negatively correlated with serum interleukin-6 (IL-6), and DHEA inhibits IL-6 secretion from mononuclear cells in man in vitro: possible link between endocrinosenescence and immunosenescence. J Clin Endocrinol Metab 1998;83:20127. [78] Szuran T, Zimmermann E, Welzl H. Water maze performance and hippocampal weight of prenatally stressed rats. Behav Brain Res 1994;65:1535. [79] Tohgi H, Utsugisawa K, Yamagata M, Yoshimura M. Effects of age on messenger RNA expression of glucocorticoid thyroid hormone androgen and estrogen receptors in postmortem human hippocampus. Brain Res 1995;700:24553. [80] Touma C, Palme R, Sachser N. Different types of oestrous cycle in two closely related South American rodents (Cavia aperea and Galea musteloides) with different social and mating systems. Reproduction 2001;121:791801. [81] Von Holst D. The concept of stress and its relevance for animal behavior. In: Lehman DS, Hinde R, Shaw E, editors. Advances of the study of behavior. London: Academic Press; 1998. p. 1131. [82] Ward I. Prenatal stress feminizes and demasculinizes the behavior of males. Science 1972;175:824. [83] Weiland NG, Orikasa C, Hayashi S, McEwen BS. Distribution and hormone regulation of estrogen receptor immunoreactive cells in the hippocampus of male and female rats. J Comp Neurol 1997;388:603 12. [84] Weiss EL, Longhurst JG, Mazure CM. Childhood sexual abuse as a risk factor for depression in women: psychosocial and neurobiological correlates. Am J Psychiatry 1999;156:81628. [85] Wood RI, Newman SW. Androgen receptor immunoreactivity in the male and female Syrian hamster brain. J Neurobiol 1999;39:359 70. [86] Young WJ, Chang C. Ontogeny and autoregulation of androgen receptor mRNA expression in the nervous system. Endocrine 1998;9:79 88.

[43] Koehl M, Darnaudery M, Dulluc J, Van Reeth O, Le Moal M, Maccari S. Prenatal stress alters circadian activity of hypothalamopituitary-adrenal-axis and hippocampal corticosteroid receptors in adult rats of both gender. J Neurobiol 1999;40:30215. [44] Kruijver FPM, Zhou J-N, Pool CW, Hofman MA, Gooren LJG, Swaab DF. Male-to-female transsexuals have female neuron numbers in a limbic nucleus. J Clin Endocrinol Metab 2000;85:203441. [45] Kruijver FPM, Fernndez-Guasti A, Fodor M, Kraan EM, Swaab DF. Sex differences in androgen receptors of the human mamilary bodies are related to endocrine status rather than to sexual orientation or transsexuality. J Clin Endocrinol Metab 2001;86:81827. [46] Kruijver FPM, Swaab DF. Sex hormone receptors are present in the human suprachiasmatic nucleus. Neuroendocrinology 2002;75:296 305. [47] Khnemann S, Brown TJ, Hochberg RB, MacLusky NJ. Sex differences in the development of estrogen receptors in the rat brain. Horm Behav 1994;28:28391. [48] Kunkel P, Kunkel I. Beitrge zur Ethologie des Hausmeerschweinchens. Z Tierpsychol 1964;21:60241. [49] Lemaire V, Koehl M, Le Moal M, Abrous DN. Prenatal stress produces learning decits associated with an inhibition of neurogenesis in the hippocampus. PNAS 2000;97:110327. [50] Maccari S, Piazza PV, Kabbaj M, Barbazanges A, Simon H, Le Moal M. Adoption reverses the long-term impairment in glucocorticoid feedback induced by prenatal stress. J Neurosci 1995;15:1106. [51] MacLusky NJ, Naftolin F. Sexual differentiation of the central nervous system. Science 1981;211:1294303. [52] Mani SK, Allen JM, Clark JH, Blaustein JD, OMalley BW. Convergent pathways for steroid hormone- and neurotransmitter-induced rat sexual behavior. Science 1994;265:12469. [53] Martin P, Bateson P. Measuring behaviour. Cambridge: Cambridge University Press; 1993. [54] McEwen B. Estrogen actions throughout the brain. Recent Prog Horm Res 2002;57:35784. [55] Meaney MJ, Stewart J. Neonatal androgens inuence the social play of prepubescent rats. Horm Behav 1981;15:197213. [56] Meek LR, Burda KM, Paster E. Effects of prenatal stress on development in mice: maturation and learning. Physiol Behav 2000;71:543 9. [57] Meyer HHD, Sauerwein H, Mutayoba BM. Immunoafnity chromatography and a biotin-streptavidin amplied enzymeimmunoassay for sensitive and specic estimation of estradiol-17 . J Steroid Biochem 1990;35:2639. [58] Michael RP, Clancy AN, Zumpe D. Distribution of androgen receptor-like immunoreactivity in the brains of cynomolgus monkeys. J Neuroendocrinol 1995;7:7139. [59] Moore CL, Power KL. Prenatal stress affects motherinfant interaction in norway rats. Dev Psychobiol 1986;19:23545. [60] Morris RGM, Garrud P, Rawlins JNP, OKeefe J. Place navigation impaired in rats with hippocampal lesion. Nature 1982;297:6813. [61] Moyer JA, Herrenkohl LR, Jacobowitz DM. Stress during pregnancy: effect on catecholamines in discrete brain regions of offspring as adults. Brain Res 1978;144:1738. [62] Prins GS, Birsch L, Greene GL. Androgen receptor localization in different cell types of the adult rat prostate. Endocrinology 1991;129:318799. [63] Rhees RW, Fleming DE. Effects of malnutriton, maternal stress, or ACTH injections during pregnancy on sexual behavior of male offspring. Physiol Behav 1981;27:87982. [64] Rood JP, Weir BJ. Reproduction in female wild guinea pigs. J Reprod Fertil 1970;23:393409. [65] Roof RL, Havens MD. Testosterone improves maze performance and development of a male hippocampus in females. Brain Res 1992;572:31031.