- Email: [email protected]
Crowding pregnant mice affects attack and threat behavior of male offspring
HORMONES
AND
BEHAVIOR
19, 86-97 (1985)
Crowding Pregnant Mice Affects Attack and Threat Behavior of Male Offspring PHILIP W. HARVEY AND PETER F. D. CHEVINS Department
of Biological Sciences, University of Keele, Staffordshire ST5 5BG, England
Keele,
Attack and threat behavior of adult male offspring of female mice crowded during the final third of pregnancy was investigated. In 5-min test pairings with an anosmic “standard opponent” which had 50 ~1 of male mouse urine applied to its fur, the prenatally stressed group of males showed significantly less attack behavior; attack latency was longer and number of attacks, bites, amount of time spent attacking, and composite aggression scores were all lower, compared with the control group. Similarly, less threat behavior was observed in offspring from crowded dams; there were lower frequencies of tail rattles, rough grooms, and upright threats. Additionally, proportionally fewer males in the prenatally stressed group attacked or displayed threats. A second experiment was designed to investigate the effects of exogenous androgen on the aggressiveness of males from crowded mice: testosterone propionate administration (500 &animal/day, for 5 days prior to testing) abolished differences both in the proportion of males from crowded mice that fought and also apparently abolished differences in intensities of attack and threat behavior between groups. However, trends toward reduced aggression in prenatally crowded males remained. More detailed analysis of these responses, based only on animals that displayed aggression, revealed significantly reduced intensity of aggression in offspring from females crowded during pregnancy, indicating that testosterone propionate therapy did not completely restore this behavior. In order to reduce postnatal effects due to possible differences in mothering, all offspring were fostered to untreated mothers at birth. The results are discussed in terms of in utero exposure of male fetuses of crowded dams to stress-liberated adrenal steroids of maternal origin, and the possible consequences for the endocrine integrity of these offspring. 0 1985 Academic press, hc.
Environmental factors are well known to exert powerful influences on the development of behavior in mammals, especially if they exert their effect during sensitive periods of neural differentiation. Fluctuations in the environment in one generation may therefore add to the phenotypic variation in behavior shown by the next, but little is known of the detailed causation of such developmental alteration. Progress is, however, being made in describing and explaining the effects of adverse environmental conditions during pregnancy on the development of offspring behavior and this study is a contribution to that literature. 86 0018-506X/85 $1.50 Copyright 0 1985 by Academic press, Inc. All rights of reproduction in any form reserved.
PRENATAL
STRESS
AND
AGGRESSION
87
Early experiments on stressful conditions during pregnancy were largely concerned with the effects on offspring exploratory behavior and fear responses (Archer and Blackman, 1971; Joffe, 1978), and often yielded conflicting results. More recently, several authors have described feminization of male sexual behavior resulting both from crowding during pregnancy (Dahlof, Hkd, and Larsson, 1977) and from periodic restraint under hot lights (Ward, 1972; Herrenkohl and Whitney, 1976; Meisel, Dohanich, and Ward, 1979). Impairment of normal masculine sexual behavior (Ward, 1972; Dunlap, Zadina, and Gougis, 1978; Gotz and Domer, 1980; Harvey and Chevins, 1984) has also been reported. These studies suggest that adverse conditions during the last few days of gestation in rodents may disrupt sexual differentiation of the brain. If this is so, other sexually dimorphic patterns of behavior may also be affected: in this study we have examined the effects on intermale aggression in mice. We have chosen to use crowded housing conditions during late pregnancy as our environmental variable; a paradigm which has been previously shown to compromise sexual differentiation of the male, as seen in sexual behavior (Dahlof et al., 1977; Harvey and Chevins, 1984), but which may well prove to be a less severe stressor than restraint, combined with heat and light. Living in high population densities is known to depress activity of the pituitary gonadal axis of male mice (Bronson, 1973; Jean-Faucher, Berger, De Turckheim, Veyessiere, and Jean, 1981) and suppress female reproductive cycles (McKinney, 1972; Nichols, 1980). Activation of the pituitaryadrenal axis has often been invoked as the final common path of such stressors (e.g., Christian, 1964) and may well also be involved in the mediation of the effects of stress during pregnancy. Efforts to demonstrate its involvement in causing defects in male sexual behavior have not, however, yielded consistent results: Chapman and Stern (1978) failed to show demasculinized male offspring after adrenocorticotrophic hormone (ACTH) treatment during pregnancy while other studies show that ACTH during pregnancy impairs masculine sexual responses (Rhees and Fleming, 1981; Stylianopoulou, 1983; Harvey and Chevins, 1984) and aggression (Simon and Gandelman, 1977). For its full expression, intermale aggression in mice, like male sexual behavior in many rodent species, requires the organizing influence of adequate circulating concentrations of androgen on the neural apparatus which will eventually drive the behavior, at the time of its differentiation. Androgen is again required during adulthood for the activation of the behavioral program when the relevant sensory stimuli are present. Absence of normal testosterone titers at either or both of these times could therefore result in failure to display aggression in response to stimuli which would normally elicit it. Having demonstrated loss of normal aggressive behavior in a first experiment we have therefore attempted to examine the nature
88
HARVEY
AND
CHEVINS
of the alteration in the experimental animals by administration of an appropriate dose of a long-acting testosterone preparation before presenting stimuli which normally evoke aggression. METHODS
Females used in these studies were virgin “TO” strain outbred albino mice, obtained from A. Tuck and Son Ltd., Battlesbridge, Essex 2-5 weeks prior to mating or were virgin “TO” females bred in our own laboratories. Prior to mating, females were housed in groups of 10 in large plastic cages (42 x 25 x 11 cm), allowed ad libitum supply of food (Labsure Animal Diet, Christopher Hill Ltd., Dot-set) and water, and maintained on a reverse lighting regime (red lights on 1200-2200 hr) at 18-23°C. At the age of 10 weeks, females were placed individually into small plastic cages (30 x 13 x 11 cm) with a sexually experienced male, and observed daily for the appearance of vaginal plugs, which were deemed to indicate Day 0 of pregnancy. Males were then removed. In Experiment 1, pregnant females were assigned at random to one of two treatment groups: crowding stress (PN, N = 8) or control (CON, N = 8). Females assigned to the stress group were removed from single housing on Day 12 of pregnancy and placed in a large cage of aggressive males. Crowding cages contained 25-28 males and 2-5 females, so that the total housing density was 30 mice per cage. Several males were moved daily between crowding cages to ensure social instability and continued fighting between males within groups. Observation of these cages revealed that this treatment did produce intermale aggression; pregnant females were rarely attacked but were repeatedly pursued and mounted. On Day 17 of pregnancy, the experimental females were removed from the crowding cages and rehoused individually in small cages. Births normally occurred the following day. Throughout the treatment period, control females remained undisturbed and were housed individually in small cages. Pups from litters produced from both treatments were sexed, culled at random to 8 per litter, and fostered within 13 hr of birth to an untreated mother that had given birth with the previous 24 hr. Litters were then left undisturbed until postnatal day 2 1, when offspring were weaned and rehoused 8-10 per large cage according to sex and treatment, so that at least one representative from each litter was contained in each group. At 10-l 1 weeks of age, male offspring were removed from group housing and isolated in small cages for 16 days prior to aggression testing. In Experiment 2 experimental (PN, N = 9) and control (CON, N = 9) litters were identically raised. Male offspring were isolated for 21 days prior to aggression testing during the last 5 days of which each male in both control and prenatally stressed groups received a daily subcutaneous
PRENATAL
STRESS
AND
AGGRESSION
89
injection of 500 pg testosterone propionate (Sigma) in 0.1 ml olive oil (Harvey and Chevins, 1984). Behavioral tests were conducted in a large neutral cage containing clean sawdust and covered in clear Perspex. A standard opponent was placed into the arena for 5 min, after which time the test male was introduced. Standard opponents were group-housed males aged approximately 11 weeks and rendered anosmic by nasal perfusion with 4% zinc sulfate solution. Anosmic males rarely attack other mice (Brain, Benton, Childs, and Parmigiani, 1981). The standard opponents had approximately 50 ~1 of male mouse urine smeared on rump and base of tail. This urine was pooled from five isolated, sexually experienced mice and was collected during the 24-hr period prior to use. Application of urine immediately prior to testing dramatically increases the probability and intensity of aggression directed to these animals-probably due to the presence of an aggression-facilitating pheromone in the urine of male mice (Ingersoll, Bobotas, Ching-Tseu, and Lukton, 1982). Standard opponents were used once or twice only, and where they were used twice these were not in consecutive tests, nor in presentations to test males of the same category (control or experimental). Care was also taken to avoid bias through order effects in second use of standard opponents. The resulting behavioral interaction was observed for 5 min via remote closed-circuit TV monitor and videotaped for later detailed analysis. Behavioral tests were conducted under red light between 1600 and 2100 hr. Measures of aggression were latency to attack, number of discrete biting attacks, number of bites, and cumulative attack time. A composite aggression score (+ 1 point for each sniff, bite, and tail rattle) was also calculated. Sniffs were scored when targeted to the genital area of the standard opponent and tail rattles recorded as each distinct bout of tail rattling. These measures of aggression are largely based on those employed by Brain and Nowell (1970), Brain, Nowell, and Wouters (1971), Brain (1972), and Brain and Poole (1974). Also recorded were rough grooms and upright threats as defined by Simon and Gandelman (1981). All results were analyzed employing nonparametric techniques, and statistical tests used were Mann-Whitney U test and Fisher’s exact probability test. Results are presented as median scores with 95% confidence limits (Campbell, 1979). RESULTS
The results from Experiment 1 show that chronic pregnancy significantly impairs the expression of both responses of adult male offspring in mice (Table 1). crowded females showed longer attack latencies, and
crowding during attack and threat Male offspring of lower numbers of
7 1-21
14 5-39
300* 14-300
CON (N= 15)
PN (N= 15)
35 3-52 0** O-18
0** o-13
(set)
26 3-46
Number of bites
Cumulative attack time
0* o-3
3 l-6
Number of rough grooms
1* o-4
4 1-13
Number of upright threats
28** 24-44
52 43-92
Composite agression score
Significantly different from CON: *P < 0.025, **P < 0.01.
5** o-12
13 8-29
16 14-24 17 13-28
Number of tail rattles
Number of sniffs
TABLE 1 during Pregnancy on Male Offspring Attack and Threat Behavior’
a Medians and 95% confidence limits, CON = control, PN = experimental,
0* o-7
Number of attacks
Attack latency (xc)
The Effects of Chronic Crowding
B s 3
%
2
$ z
PRENATAL
STRESS
AND
AGGRESSION
91
attacks and bites directed toward the standard opponent than did controls. They also spent less time fighting and obtained overall lower values in the composite aggression score compared with controls. Less threat behavior was also apparent in males from crowded dams: these animals displayed fewer tail rattles, rough grooms, and upright threats compared with control offspring. Further analysis of these results based on data only from animals that showed aggression revealed that there were no differences in the intensity of aggression between groups. However, proportional analysis of data (Table 2) revealed that significantly fewer males from crowded mice attacked or threatened. Experiment 2 shows that administration of testosterone propionate prior to tests of aggression results in approximately equal proportions of control males and those from crowded dams attacking or threatening (Table 2) and abolishes significant differences in the frequency of display of these aggressive responses (Table 3). However, an overall deficit in the aggressive responses of males from crowded mice remained. Further analysis of results based on data only from animals that showed aggression (Table 4) revealed that there were significant differences in the intensity of aggression between groups. Aggressive male offspring from crowded mice, having been treated with testosterone propionate prior to aggression testing, displayed longer attack latencies, delivered fewer attacks and bites to the standard opponent, showed fewer tail rattles and rough grooms, spent less time engaged in attacks, and achieved lower values on the composite aggression score than comparable treated control males. Investigation of litter sex ratios at birth showed that there were no significant differences in the proportion of males born in litters from each treatment group. The mean percentage of males in litters from crowded mice was 53.1 + 2.9%, compared with 54.4 + 2.4% from control females. DISCUSSION Chronic crowding during the final third of pregnancy is shown in Experiment 1 to significantly reduce aggressive responses of adult male offspring in mice. Thus the reported effects of stress during pregnancy are extended not merely to another species, but also to another androgendependent sexually dimorphic aspect of behavior. Sexual differentiation of behavior in male rodents does, therefore, seem to be at risk if the pregnant female’s living conditions are adverse during the final stages of fetal development. This lack of masculinization could reflect a failure of development in the brain region concerned with aggressive behavior, or a failure in development of the pituitary-gonadal system. As the latter is ultimately under hypothalamic control, a developmental defect in the central nervous system is implicated in any case, and the question resolves itself to whether the defect is neurobehavioral, neuroendocrine, or both. Our
6 (W 6 (60)
Experiment 2 CON + TP (N = 10) PN + TP (N = 10) 3 (30) 6 (W
12 (80) 6**(40)
Number of males displaying rough grooms
Values in parentheses are percentages. Significantly
6 (60) 6 (6’3
13 (87) 9 (60)
Number of males displaying tail rattles
222.0 36-300
PN (N = 10)
1.5 o-12
9.5 o-25
Number of attacks
1.5 o-22
11.5 O-46
Number of bites
2.5 O-30
21.0 O-72
Cumulative attack time 64
13.5 8-15
15.0 4-18
Number of sniffs
2.0 O-10
8.5 o-17
Number of trail rattles
1.0 O-l
0.0 o-4
Number of rough grooms
’ Analysis based on all animals. Medians and 95% confidence limits. CON = control, PN = experimental.
63.5 21-300
CON (N = 10)
Attack latency (se4
7 (70) 5 m-0
13 (87) 8*(53)
Number of males displaying upright threats
of Male Offspring
0.5 o-7
2.5 O-10
Number of upright threats
16.0 13-36
39.0 14-66
Composite aggression score
Propionate Administration”
different from CON: *P =
upon the Proportion
TABLE 3 Chronic Crowding during Pregnancy and Male Offspring Attack and Threat Behavior: Influence of Testosterone
” Number of animals. CON = control, PN = experimental. 0.047, **P = 0.026.
12 (80) 6**(40)
Experiment I CON (N = 15) PN (N = 15)
Number of males displaying attacks
TABLE 2 The Effects of Chronic Crowding during Pregnancy and Testosterone Propionate (TP) Administration Displaying Attacks or Threats”
G 2 z
5 b
s
T 54
h)
W
36-298
23.5 15-80 7.0* 1-28
31.0 2-47
19.0 2-26
4.0** l-17
Number of bites
Number of attacks
13.0* 2-46
50.0 5-77
Cumulative attack time bed
13.5 8-15
15.0 4-18
Number of sniffs
6.0** 2-10
15.5 3-22
Number of tail rattles
7.0 1-9”
9.0 2-13
4.0 4-5b 1.0** l-4
Number of upright threats
Number of rough grooms
37.5** 13-49
59.0 23-73
Composite agression score
” Analysis based on animals displaying agression. Medians and 95% confidence limits. CON = control, PN = experimental. Significantly different from CON: *P < 0.05, **P -c 0.025. ’ Indicates range of data where N = < 6.
PN
CON
Treatment
Attack latency (set)
TABLE 4 Chronic Crowding during Pregnancy and Male Offspring Attack and Threat Behavior: Influence of Testosterone Propionate Administration”
E g
$
94
HARVEY
AND
CHEVINS
attempt to distinguish between these possibilities by administration of testosterone propionate relied on the rationale that if the damage was purely endocrine, replacement therapy would reinstate normal aggressive behavior. First analysis of the results of Experiment 2 indicates that this is the case, as no difference in the proportions of control and experimental males fighting, and no differences in the latencies or numbers of the component acts are apparent. However, when the analysis is confined only to those males which fight, differences in the intensity of aggression are seen to remain. We interpret this result to mean that all the experimental animals showing aggression have incomplete neurobehavioral masculinization resulting from prenatal stress. In addition, at least a proportion of the males (those which only show fighting after exogenous testosterone administration) have also a neuroendocrine defect. Why some males behave differently from others, in that they fight only when given exogenous testosterone, is a point of some interest. Even among control animals, of course, some fight when others do not. Several studies have drawn attention to the effects of differential exposure to androgens or estrogens in utero which results from fetal position between two male fetuses (2M), two female fetuses (OM), or one of each sex (Vom Saal and Bronson, 1978, 1980; Vom Saal, 1983; Hauser and Gandelman, 1983). These different categories of males have been shown to exhibit different degrees of sexual differentiation of body and behavior, including agonistic behavior (Vom Saal, Grant, McMullen, and Laves, 1983) and one possibility is that those which do not normally fight are the least masculinized “OM” males. If this were so, any alteration in the sex ratio by experimental treatment could alter the proportion of aggressive males, hence we considered it important to record sex ratio at birth. Lane and Hyde (1973) do report selective mortality of fetuses in utero after maternal stress but crowding late during pregnancy did not alter the sex ratio at birth, so this phenomenon cannot explain our results. It remains a possibility that a category of males showing intermediate masculinization (“I,“?) can be induced to fight only when given extra testosterone. Indeed Vom Saal (1983) has recently shown for another sexually dimorphic behavioral trait, infanticide by males, that the effects of the sex of the in utero neighbors interact with those of prenatal stress. No information is as yet available regarding the causation of the aberrant sexual behavior shown by male offspring of rodents stressed during pregnancy (Ward, 1972; Herrenkohl and Whitney, 1976; DahlM et al., 1977; Dunlap et al., 1978; Chapman and Stern, 1978; Meisel, et al., 1979; Gotz and Darner, 1980; Harvey and Chevins, 1984) but Ward and Weisz (1980) report an acceleration of the normal fetal testosterone surge and Darner (1980) showed reduced neonatal testosterone levels resulting from this treatment. The proximate cause may thus be mistiming of the surge of testosterone before birth and its inadequate secretion after birth. The
PRENATAL
STRESS AND AGGRESSION
95
more basic cause both of this effect and of the effects on aggression may yet prove to lie in the maternal pituitary-adrenal stress response: Simon and Gandelman (1977) administered ACTH to mice during late pregnancy, and found a reduction in the number of male offspring showing aggression in adulthood. As in our own study, testosterone propionate treatment of these males was found to increase the proportion of males that fought. More recently, Politch and Herrenkohl (1984a) have investigated the effects of restraint during pregnancy in the mouse: although unable to show that this form of stress affected normal copulatory behavior of adult male offspring, they do report (1984b) reduced masculine sexual behavior following ACTH or corticosterone administration during pregnancy. The weight of evidence at present therefore suggests that the effects of stress during pregnancy on sexual differentiation of behavior of male offspring in rodents are mediated via activation of the maternal pituitary-adrenal system. Perhaps the most important conclusion of this study is that it supports the hypothesis that sexually dimorphic aspects of behavior are particularly vulnerable to environmental insult during pregnancy, at least in the male offspring of altricial rodent species. It will now be of interest to see whether the female offspring are similarly at risk: so far studies in this area have been limited (Herrenkohl and Whitney, 1976; Herrenkohl and Politch, 1978; Herrenkohl and Gala, 1979; Beckhardt and Ward, 1983) but are presently in progress in our laboratory. ACKNOWLEDGMENTS The authors acknowledge the services of D. Bosworth and J. Shaw in caring for the animals and the comments of Dr. P. F. Brain (University College, Swansea). P.W.H. is supported by a University of Keele research studentship.
REFERENCES Archer, J. E., and Blackman, D. C. (1971). Prenatal psychological stress and offspring behavior in rats and mice. Dev. Psychobiol. 4, 193-248. Beckhardt, S., and Ward, I. L. (1983). Reproductive functioning in the prenatally stressed female rat. Dev. Psycho&l. 16, 111-119. Brain, P. F. (1972). Study on the effect of the 4-10 ACTH fraction on isolation-induced intermale fighting behaviour in albino mice. Neuroendocrinology 10, 371-376. Brain, P. F., Benton, D., Childs, Cl., and Parmigiani, S. (1981). The effect of the type of opponent in tests of murine aggression. Behav. Processes 6, 319-327. Brain, P. F., and Nowell, N. W. (1970). Some observations on intermale aggression testing in albino mice. Commun. Behav. Eiol. 5, 7-17. Brain, P. F., Nowell, N. W., and Wouters, A. (1971). Some relationships between adrenal function and the effectiveness of a period of isolation in inducing intermale aggression in albino mice. Physiol. Behav. 6, 27-29. Brain, P. F., and Poole, A. (1974). Some studies on the use of standard opponents in intermale aggression testing in TT albino mice. Behaviour 50, 100-110. Bronson, F. H. (1973). Establishment of social rank among grouped mice: Relative effects on circulating FSH, LH and corticosterone. Physiol. Behav. 10, 947-951.
96
HARVEY
AND CHEVINS
Campbell, R. C. (1979). Statistics for Biologists, 2nd ed. Cambridge Univ. Press, London. Chapman, R. H., and Stem, J. M. (1978). Maternal stress and pituitary-adrenal manipulations during pregnancy in rats: Effects on morphology and sexual behaviour of male offspring. .I. Comp.
Physiol.
Psychol.
92, 1074-1083.
Christian, J. J., Lloyd, J. A., and Davis, D. E. (1964). The role of endocrines in the self regulation of mammalian populations. Rec. Prog. Horm. Res. 21, 507-578. Dahlof, L. G., Hard, E., and Larsson, K. (1977). Influence of maternal stress on offspring sexual behaviour. Anim. Behav. 25, 958-963. Darner, G. (1980). Sexual differentiation of the brain. Vitum. Horm. 38, 325-381. Dunlap, J. L., Zadina, J. E., and Gougis, G. (1978). Prenatal stress interacts with prepuberal social isolation to reduce male copulatory behavior. Physiol. Behav., 21, 873-875. G&z, F., and Darner, G. (1980). Homosexual behavior in prenatally stressed male rats after castration and oestrogen treatment. Endokrinologie 76, 115-I 17. Harvey, P. W., and Chevins, P. F. D. (1984). Crowding or ACTH treatment of pregnant mice affects adult copulatory behavior of male offspring. Horm. Behuv. 18, 101-l 10. Hauser, H., and Gandelman, R. (1983). Contiguity to males in utero affects avoidance responding in adult female mice. Science (Washington, D.C.) 220, 437-438. Herrenkohl, L. R., and Gala, R. R. (1979). Serum prolactin levels and maintenance of progeny by prenatally stressed female offspring. Experientia 35, 702-704. Herrenkohl, L. R., and Politch, J. A. (1978). Effects of prenatal stress on the estrous cycle of female offspring as adults. Experientiu 34, 1240-1241. Herrenkohl, L. R., and Whitney, J. B. (1976). Effects of prepartal stress on postpartal nursing behavior, litter development and adult sexual behavior. Physiol. Behuv. 17, 1019-1021. Ingersoll, D. W., Bobotas, G., Ching-Tse, L., and Lukton, A. (1982). Glucuronidase activation of latent aggression promoting cues in mouse bladder urine. Physiol. Behuv. 29, 789-793. Jean-Faucher, C. Berger, M., De Turckheim, M., Veyessiere, G., and Jean, C. (1981). Effects of dense housing on growth of reproductive organs, plasma testosterone levels and fertility in male mice. J. Endocrinol. 90, 397-402. Joffe, J. M. (1978). Hormonal mediation of the effects of prenatal stress on offspring behaviour. In G. Gottlieb (Ed.), Studies on the Development of Behuviour and the Nervous system. Early InJluences. Academic Press, New York. Lane, E. A., and Hyde, T. S. (1973). The effects of maternal stress on fertility and sex ratio-A pilot study with rats. J. Abnorm. Psycho/. 82, 78-80. McKinney, T. D. (1972). Estrous cycle in house mice: Effects of grouping, preputial gland odors and handling. J. Mammal. 53, 391-393. Meisel, R. L., Dohanich, G. P., and Ward, I. L. (1979). Effects of prenatal stress on avoidance acquisition, open field performance and lordotic behavior in male rats. Physiol.
Behuv.
22, 527-530. Stress in Female Activity and the Oestrous
Nichols, D. J. (1980). Social
Mice:
Effects
of Differentiul
Housing
on
Cycle. Ph.D. thesis, University of Keele. Politch, J. A., and Herrenkohl, L. R. (1984a). Effects of prenatal stress on reproduction in male and female mice. Physiol. Behuv. 32, 95-100. Politch, J. A., and Herrenkohl, L. R. (1984b). Prenatal ACTH and corticosterone: Effects on reproduction in male mice. Physiol. Behav. 32, 135-138. Rhees, R. W., and Fleming, D. E. (1981). Effects of malnutrition, maternal stress or ACTH injections during pregnancy on sexual behavior of male offspring. Physiol. Behav. 27, 879-883. Rose, R. M. (1969). Androgen responses to stress. Psychosom. Med. 31, 405-417. Simon, N. G., and Gandelman, R. (1977). Decreased aggressive behavior in offspring of ACTH treated mice. Behav. Biol. 21, 478-488. Adrenocorticul
PRENATAL
STRESS AND AGGRESSION
97
Simon, N. G., and Gandelman, R. (1981). Threat postures signal impending attack in mice. Behav. Neural Biol. 33, 509-514. Stylianopoulou, F. (1983). Effect of maternal adrenocorticotropin injections on the differentiation of sexual behavior of the offspring. Harm. Behav. 17, 324-331. Vom Saal, F. S. (1983). Variation in infanticide and parental behavior in male mice due to prior intrauterine proximity to female fetuses: Elimination by prenatal stress. Physiol. Behav. 30, 675-681. Vom Saal, F. S., and Bronson, F. H. (1978). In utero proximity of female mouse fetuses to males: Effect on reproductive performance in later life. Biol. Reprod. 19, 842-853. Vom &al, F. S., and Bronson, F. H. (1980). Variation in length of the estrous cycle in mice due to former intrauterine proximity to male fetuses. Biol. Reprod. 22, 777-780. Vom Saal, F. S., Grant, W., McMullen, C., and Laves, K. (1983). High fetal estrogen concentrations: Correlation with increased adult sexual performance and decreased aggression in male mice. Science (Washington, D.C.) 220, 1306-1308. Ward, I. L. (1972). Prenatal stress feminizes and demasculinizes the behavior of males. Science (Washington, DE.) 175, 82-84. Ward, I. L., and Weisz, J. (1980). Maternal stress alters plasma testosterone in fetal males. Science
(Washington,
D.C.)
207, 328-329.
AND
BEHAVIOR
19, 86-97 (1985)
Crowding Pregnant Mice Affects Attack and Threat Behavior of Male Offspring PHILIP W. HARVEY AND PETER F. D. CHEVINS Department
of Biological Sciences, University of Keele, Staffordshire ST5 5BG, England
Keele,
Attack and threat behavior of adult male offspring of female mice crowded during the final third of pregnancy was investigated. In 5-min test pairings with an anosmic “standard opponent” which had 50 ~1 of male mouse urine applied to its fur, the prenatally stressed group of males showed significantly less attack behavior; attack latency was longer and number of attacks, bites, amount of time spent attacking, and composite aggression scores were all lower, compared with the control group. Similarly, less threat behavior was observed in offspring from crowded dams; there were lower frequencies of tail rattles, rough grooms, and upright threats. Additionally, proportionally fewer males in the prenatally stressed group attacked or displayed threats. A second experiment was designed to investigate the effects of exogenous androgen on the aggressiveness of males from crowded mice: testosterone propionate administration (500 &animal/day, for 5 days prior to testing) abolished differences both in the proportion of males from crowded mice that fought and also apparently abolished differences in intensities of attack and threat behavior between groups. However, trends toward reduced aggression in prenatally crowded males remained. More detailed analysis of these responses, based only on animals that displayed aggression, revealed significantly reduced intensity of aggression in offspring from females crowded during pregnancy, indicating that testosterone propionate therapy did not completely restore this behavior. In order to reduce postnatal effects due to possible differences in mothering, all offspring were fostered to untreated mothers at birth. The results are discussed in terms of in utero exposure of male fetuses of crowded dams to stress-liberated adrenal steroids of maternal origin, and the possible consequences for the endocrine integrity of these offspring. 0 1985 Academic press, hc.
Environmental factors are well known to exert powerful influences on the development of behavior in mammals, especially if they exert their effect during sensitive periods of neural differentiation. Fluctuations in the environment in one generation may therefore add to the phenotypic variation in behavior shown by the next, but little is known of the detailed causation of such developmental alteration. Progress is, however, being made in describing and explaining the effects of adverse environmental conditions during pregnancy on the development of offspring behavior and this study is a contribution to that literature. 86 0018-506X/85 $1.50 Copyright 0 1985 by Academic press, Inc. All rights of reproduction in any form reserved.
PRENATAL
STRESS
AND
AGGRESSION
87
Early experiments on stressful conditions during pregnancy were largely concerned with the effects on offspring exploratory behavior and fear responses (Archer and Blackman, 1971; Joffe, 1978), and often yielded conflicting results. More recently, several authors have described feminization of male sexual behavior resulting both from crowding during pregnancy (Dahlof, Hkd, and Larsson, 1977) and from periodic restraint under hot lights (Ward, 1972; Herrenkohl and Whitney, 1976; Meisel, Dohanich, and Ward, 1979). Impairment of normal masculine sexual behavior (Ward, 1972; Dunlap, Zadina, and Gougis, 1978; Gotz and Domer, 1980; Harvey and Chevins, 1984) has also been reported. These studies suggest that adverse conditions during the last few days of gestation in rodents may disrupt sexual differentiation of the brain. If this is so, other sexually dimorphic patterns of behavior may also be affected: in this study we have examined the effects on intermale aggression in mice. We have chosen to use crowded housing conditions during late pregnancy as our environmental variable; a paradigm which has been previously shown to compromise sexual differentiation of the male, as seen in sexual behavior (Dahlof et al., 1977; Harvey and Chevins, 1984), but which may well prove to be a less severe stressor than restraint, combined with heat and light. Living in high population densities is known to depress activity of the pituitary gonadal axis of male mice (Bronson, 1973; Jean-Faucher, Berger, De Turckheim, Veyessiere, and Jean, 1981) and suppress female reproductive cycles (McKinney, 1972; Nichols, 1980). Activation of the pituitaryadrenal axis has often been invoked as the final common path of such stressors (e.g., Christian, 1964) and may well also be involved in the mediation of the effects of stress during pregnancy. Efforts to demonstrate its involvement in causing defects in male sexual behavior have not, however, yielded consistent results: Chapman and Stern (1978) failed to show demasculinized male offspring after adrenocorticotrophic hormone (ACTH) treatment during pregnancy while other studies show that ACTH during pregnancy impairs masculine sexual responses (Rhees and Fleming, 1981; Stylianopoulou, 1983; Harvey and Chevins, 1984) and aggression (Simon and Gandelman, 1977). For its full expression, intermale aggression in mice, like male sexual behavior in many rodent species, requires the organizing influence of adequate circulating concentrations of androgen on the neural apparatus which will eventually drive the behavior, at the time of its differentiation. Androgen is again required during adulthood for the activation of the behavioral program when the relevant sensory stimuli are present. Absence of normal testosterone titers at either or both of these times could therefore result in failure to display aggression in response to stimuli which would normally elicit it. Having demonstrated loss of normal aggressive behavior in a first experiment we have therefore attempted to examine the nature
88
HARVEY
AND
CHEVINS
of the alteration in the experimental animals by administration of an appropriate dose of a long-acting testosterone preparation before presenting stimuli which normally evoke aggression. METHODS
Females used in these studies were virgin “TO” strain outbred albino mice, obtained from A. Tuck and Son Ltd., Battlesbridge, Essex 2-5 weeks prior to mating or were virgin “TO” females bred in our own laboratories. Prior to mating, females were housed in groups of 10 in large plastic cages (42 x 25 x 11 cm), allowed ad libitum supply of food (Labsure Animal Diet, Christopher Hill Ltd., Dot-set) and water, and maintained on a reverse lighting regime (red lights on 1200-2200 hr) at 18-23°C. At the age of 10 weeks, females were placed individually into small plastic cages (30 x 13 x 11 cm) with a sexually experienced male, and observed daily for the appearance of vaginal plugs, which were deemed to indicate Day 0 of pregnancy. Males were then removed. In Experiment 1, pregnant females were assigned at random to one of two treatment groups: crowding stress (PN, N = 8) or control (CON, N = 8). Females assigned to the stress group were removed from single housing on Day 12 of pregnancy and placed in a large cage of aggressive males. Crowding cages contained 25-28 males and 2-5 females, so that the total housing density was 30 mice per cage. Several males were moved daily between crowding cages to ensure social instability and continued fighting between males within groups. Observation of these cages revealed that this treatment did produce intermale aggression; pregnant females were rarely attacked but were repeatedly pursued and mounted. On Day 17 of pregnancy, the experimental females were removed from the crowding cages and rehoused individually in small cages. Births normally occurred the following day. Throughout the treatment period, control females remained undisturbed and were housed individually in small cages. Pups from litters produced from both treatments were sexed, culled at random to 8 per litter, and fostered within 13 hr of birth to an untreated mother that had given birth with the previous 24 hr. Litters were then left undisturbed until postnatal day 2 1, when offspring were weaned and rehoused 8-10 per large cage according to sex and treatment, so that at least one representative from each litter was contained in each group. At 10-l 1 weeks of age, male offspring were removed from group housing and isolated in small cages for 16 days prior to aggression testing. In Experiment 2 experimental (PN, N = 9) and control (CON, N = 9) litters were identically raised. Male offspring were isolated for 21 days prior to aggression testing during the last 5 days of which each male in both control and prenatally stressed groups received a daily subcutaneous
PRENATAL
STRESS
AND
AGGRESSION
89
injection of 500 pg testosterone propionate (Sigma) in 0.1 ml olive oil (Harvey and Chevins, 1984). Behavioral tests were conducted in a large neutral cage containing clean sawdust and covered in clear Perspex. A standard opponent was placed into the arena for 5 min, after which time the test male was introduced. Standard opponents were group-housed males aged approximately 11 weeks and rendered anosmic by nasal perfusion with 4% zinc sulfate solution. Anosmic males rarely attack other mice (Brain, Benton, Childs, and Parmigiani, 1981). The standard opponents had approximately 50 ~1 of male mouse urine smeared on rump and base of tail. This urine was pooled from five isolated, sexually experienced mice and was collected during the 24-hr period prior to use. Application of urine immediately prior to testing dramatically increases the probability and intensity of aggression directed to these animals-probably due to the presence of an aggression-facilitating pheromone in the urine of male mice (Ingersoll, Bobotas, Ching-Tseu, and Lukton, 1982). Standard opponents were used once or twice only, and where they were used twice these were not in consecutive tests, nor in presentations to test males of the same category (control or experimental). Care was also taken to avoid bias through order effects in second use of standard opponents. The resulting behavioral interaction was observed for 5 min via remote closed-circuit TV monitor and videotaped for later detailed analysis. Behavioral tests were conducted under red light between 1600 and 2100 hr. Measures of aggression were latency to attack, number of discrete biting attacks, number of bites, and cumulative attack time. A composite aggression score (+ 1 point for each sniff, bite, and tail rattle) was also calculated. Sniffs were scored when targeted to the genital area of the standard opponent and tail rattles recorded as each distinct bout of tail rattling. These measures of aggression are largely based on those employed by Brain and Nowell (1970), Brain, Nowell, and Wouters (1971), Brain (1972), and Brain and Poole (1974). Also recorded were rough grooms and upright threats as defined by Simon and Gandelman (1981). All results were analyzed employing nonparametric techniques, and statistical tests used were Mann-Whitney U test and Fisher’s exact probability test. Results are presented as median scores with 95% confidence limits (Campbell, 1979). RESULTS
The results from Experiment 1 show that chronic pregnancy significantly impairs the expression of both responses of adult male offspring in mice (Table 1). crowded females showed longer attack latencies, and
crowding during attack and threat Male offspring of lower numbers of
7 1-21
14 5-39
300* 14-300
CON (N= 15)
PN (N= 15)
35 3-52 0** O-18
0** o-13
(set)
26 3-46
Number of bites
Cumulative attack time
0* o-3
3 l-6
Number of rough grooms
1* o-4
4 1-13
Number of upright threats
28** 24-44
52 43-92
Composite agression score
Significantly different from CON: *P < 0.025, **P < 0.01.
5** o-12
13 8-29
16 14-24 17 13-28
Number of tail rattles
Number of sniffs
TABLE 1 during Pregnancy on Male Offspring Attack and Threat Behavior’
a Medians and 95% confidence limits, CON = control, PN = experimental,
0* o-7
Number of attacks
Attack latency (xc)
The Effects of Chronic Crowding
B s 3
%
2
$ z
PRENATAL
STRESS
AND
AGGRESSION
91
attacks and bites directed toward the standard opponent than did controls. They also spent less time fighting and obtained overall lower values in the composite aggression score compared with controls. Less threat behavior was also apparent in males from crowded dams: these animals displayed fewer tail rattles, rough grooms, and upright threats compared with control offspring. Further analysis of these results based on data only from animals that showed aggression revealed that there were no differences in the intensity of aggression between groups. However, proportional analysis of data (Table 2) revealed that significantly fewer males from crowded mice attacked or threatened. Experiment 2 shows that administration of testosterone propionate prior to tests of aggression results in approximately equal proportions of control males and those from crowded dams attacking or threatening (Table 2) and abolishes significant differences in the frequency of display of these aggressive responses (Table 3). However, an overall deficit in the aggressive responses of males from crowded mice remained. Further analysis of results based on data only from animals that showed aggression (Table 4) revealed that there were significant differences in the intensity of aggression between groups. Aggressive male offspring from crowded mice, having been treated with testosterone propionate prior to aggression testing, displayed longer attack latencies, delivered fewer attacks and bites to the standard opponent, showed fewer tail rattles and rough grooms, spent less time engaged in attacks, and achieved lower values on the composite aggression score than comparable treated control males. Investigation of litter sex ratios at birth showed that there were no significant differences in the proportion of males born in litters from each treatment group. The mean percentage of males in litters from crowded mice was 53.1 + 2.9%, compared with 54.4 + 2.4% from control females. DISCUSSION Chronic crowding during the final third of pregnancy is shown in Experiment 1 to significantly reduce aggressive responses of adult male offspring in mice. Thus the reported effects of stress during pregnancy are extended not merely to another species, but also to another androgendependent sexually dimorphic aspect of behavior. Sexual differentiation of behavior in male rodents does, therefore, seem to be at risk if the pregnant female’s living conditions are adverse during the final stages of fetal development. This lack of masculinization could reflect a failure of development in the brain region concerned with aggressive behavior, or a failure in development of the pituitary-gonadal system. As the latter is ultimately under hypothalamic control, a developmental defect in the central nervous system is implicated in any case, and the question resolves itself to whether the defect is neurobehavioral, neuroendocrine, or both. Our
6 (W 6 (60)
Experiment 2 CON + TP (N = 10) PN + TP (N = 10) 3 (30) 6 (W
12 (80) 6**(40)
Number of males displaying rough grooms
Values in parentheses are percentages. Significantly
6 (60) 6 (6’3
13 (87) 9 (60)
Number of males displaying tail rattles
222.0 36-300
PN (N = 10)
1.5 o-12
9.5 o-25
Number of attacks
1.5 o-22
11.5 O-46
Number of bites
2.5 O-30
21.0 O-72
Cumulative attack time 64
13.5 8-15
15.0 4-18
Number of sniffs
2.0 O-10
8.5 o-17
Number of trail rattles
1.0 O-l
0.0 o-4
Number of rough grooms
’ Analysis based on all animals. Medians and 95% confidence limits. CON = control, PN = experimental.
63.5 21-300
CON (N = 10)
Attack latency (se4
7 (70) 5 m-0
13 (87) 8*(53)
Number of males displaying upright threats
of Male Offspring
0.5 o-7
2.5 O-10
Number of upright threats
16.0 13-36
39.0 14-66
Composite aggression score
Propionate Administration”
different from CON: *P =
upon the Proportion
TABLE 3 Chronic Crowding during Pregnancy and Male Offspring Attack and Threat Behavior: Influence of Testosterone
” Number of animals. CON = control, PN = experimental. 0.047, **P = 0.026.
12 (80) 6**(40)
Experiment I CON (N = 15) PN (N = 15)
Number of males displaying attacks
TABLE 2 The Effects of Chronic Crowding during Pregnancy and Testosterone Propionate (TP) Administration Displaying Attacks or Threats”
G 2 z
5 b
s
T 54
h)
W
36-298
23.5 15-80 7.0* 1-28
31.0 2-47
19.0 2-26
4.0** l-17
Number of bites
Number of attacks
13.0* 2-46
50.0 5-77
Cumulative attack time bed
13.5 8-15
15.0 4-18
Number of sniffs
6.0** 2-10
15.5 3-22
Number of tail rattles
7.0 1-9”
9.0 2-13
4.0 4-5b 1.0** l-4
Number of upright threats
Number of rough grooms
37.5** 13-49
59.0 23-73
Composite agression score
” Analysis based on animals displaying agression. Medians and 95% confidence limits. CON = control, PN = experimental. Significantly different from CON: *P < 0.05, **P -c 0.025. ’ Indicates range of data where N = < 6.
PN
CON
Treatment
Attack latency (set)
TABLE 4 Chronic Crowding during Pregnancy and Male Offspring Attack and Threat Behavior: Influence of Testosterone Propionate Administration”
E g
$
94
HARVEY
AND
CHEVINS
attempt to distinguish between these possibilities by administration of testosterone propionate relied on the rationale that if the damage was purely endocrine, replacement therapy would reinstate normal aggressive behavior. First analysis of the results of Experiment 2 indicates that this is the case, as no difference in the proportions of control and experimental males fighting, and no differences in the latencies or numbers of the component acts are apparent. However, when the analysis is confined only to those males which fight, differences in the intensity of aggression are seen to remain. We interpret this result to mean that all the experimental animals showing aggression have incomplete neurobehavioral masculinization resulting from prenatal stress. In addition, at least a proportion of the males (those which only show fighting after exogenous testosterone administration) have also a neuroendocrine defect. Why some males behave differently from others, in that they fight only when given exogenous testosterone, is a point of some interest. Even among control animals, of course, some fight when others do not. Several studies have drawn attention to the effects of differential exposure to androgens or estrogens in utero which results from fetal position between two male fetuses (2M), two female fetuses (OM), or one of each sex (Vom Saal and Bronson, 1978, 1980; Vom Saal, 1983; Hauser and Gandelman, 1983). These different categories of males have been shown to exhibit different degrees of sexual differentiation of body and behavior, including agonistic behavior (Vom Saal, Grant, McMullen, and Laves, 1983) and one possibility is that those which do not normally fight are the least masculinized “OM” males. If this were so, any alteration in the sex ratio by experimental treatment could alter the proportion of aggressive males, hence we considered it important to record sex ratio at birth. Lane and Hyde (1973) do report selective mortality of fetuses in utero after maternal stress but crowding late during pregnancy did not alter the sex ratio at birth, so this phenomenon cannot explain our results. It remains a possibility that a category of males showing intermediate masculinization (“I,“?) can be induced to fight only when given extra testosterone. Indeed Vom Saal (1983) has recently shown for another sexually dimorphic behavioral trait, infanticide by males, that the effects of the sex of the in utero neighbors interact with those of prenatal stress. No information is as yet available regarding the causation of the aberrant sexual behavior shown by male offspring of rodents stressed during pregnancy (Ward, 1972; Herrenkohl and Whitney, 1976; DahlM et al., 1977; Dunlap et al., 1978; Chapman and Stern, 1978; Meisel, et al., 1979; Gotz and Darner, 1980; Harvey and Chevins, 1984) but Ward and Weisz (1980) report an acceleration of the normal fetal testosterone surge and Darner (1980) showed reduced neonatal testosterone levels resulting from this treatment. The proximate cause may thus be mistiming of the surge of testosterone before birth and its inadequate secretion after birth. The
PRENATAL
STRESS AND AGGRESSION
95
more basic cause both of this effect and of the effects on aggression may yet prove to lie in the maternal pituitary-adrenal stress response: Simon and Gandelman (1977) administered ACTH to mice during late pregnancy, and found a reduction in the number of male offspring showing aggression in adulthood. As in our own study, testosterone propionate treatment of these males was found to increase the proportion of males that fought. More recently, Politch and Herrenkohl (1984a) have investigated the effects of restraint during pregnancy in the mouse: although unable to show that this form of stress affected normal copulatory behavior of adult male offspring, they do report (1984b) reduced masculine sexual behavior following ACTH or corticosterone administration during pregnancy. The weight of evidence at present therefore suggests that the effects of stress during pregnancy on sexual differentiation of behavior of male offspring in rodents are mediated via activation of the maternal pituitary-adrenal system. Perhaps the most important conclusion of this study is that it supports the hypothesis that sexually dimorphic aspects of behavior are particularly vulnerable to environmental insult during pregnancy, at least in the male offspring of altricial rodent species. It will now be of interest to see whether the female offspring are similarly at risk: so far studies in this area have been limited (Herrenkohl and Whitney, 1976; Herrenkohl and Politch, 1978; Herrenkohl and Gala, 1979; Beckhardt and Ward, 1983) but are presently in progress in our laboratory. ACKNOWLEDGMENTS The authors acknowledge the services of D. Bosworth and J. Shaw in caring for the animals and the comments of Dr. P. F. Brain (University College, Swansea). P.W.H. is supported by a University of Keele research studentship.
REFERENCES Archer, J. E., and Blackman, D. C. (1971). Prenatal psychological stress and offspring behavior in rats and mice. Dev. Psychobiol. 4, 193-248. Beckhardt, S., and Ward, I. L. (1983). Reproductive functioning in the prenatally stressed female rat. Dev. Psycho&l. 16, 111-119. Brain, P. F. (1972). Study on the effect of the 4-10 ACTH fraction on isolation-induced intermale fighting behaviour in albino mice. Neuroendocrinology 10, 371-376. Brain, P. F., Benton, D., Childs, Cl., and Parmigiani, S. (1981). The effect of the type of opponent in tests of murine aggression. Behav. Processes 6, 319-327. Brain, P. F., and Nowell, N. W. (1970). Some observations on intermale aggression testing in albino mice. Commun. Behav. Eiol. 5, 7-17. Brain, P. F., Nowell, N. W., and Wouters, A. (1971). Some relationships between adrenal function and the effectiveness of a period of isolation in inducing intermale aggression in albino mice. Physiol. Behav. 6, 27-29. Brain, P. F., and Poole, A. (1974). Some studies on the use of standard opponents in intermale aggression testing in TT albino mice. Behaviour 50, 100-110. Bronson, F. H. (1973). Establishment of social rank among grouped mice: Relative effects on circulating FSH, LH and corticosterone. Physiol. Behav. 10, 947-951.
96
HARVEY
AND CHEVINS
Campbell, R. C. (1979). Statistics for Biologists, 2nd ed. Cambridge Univ. Press, London. Chapman, R. H., and Stem, J. M. (1978). Maternal stress and pituitary-adrenal manipulations during pregnancy in rats: Effects on morphology and sexual behaviour of male offspring. .I. Comp.
Physiol.
Psychol.
92, 1074-1083.
Christian, J. J., Lloyd, J. A., and Davis, D. E. (1964). The role of endocrines in the self regulation of mammalian populations. Rec. Prog. Horm. Res. 21, 507-578. Dahlof, L. G., Hard, E., and Larsson, K. (1977). Influence of maternal stress on offspring sexual behaviour. Anim. Behav. 25, 958-963. Darner, G. (1980). Sexual differentiation of the brain. Vitum. Horm. 38, 325-381. Dunlap, J. L., Zadina, J. E., and Gougis, G. (1978). Prenatal stress interacts with prepuberal social isolation to reduce male copulatory behavior. Physiol. Behav., 21, 873-875. G&z, F., and Darner, G. (1980). Homosexual behavior in prenatally stressed male rats after castration and oestrogen treatment. Endokrinologie 76, 115-I 17. Harvey, P. W., and Chevins, P. F. D. (1984). Crowding or ACTH treatment of pregnant mice affects adult copulatory behavior of male offspring. Horm. Behuv. 18, 101-l 10. Hauser, H., and Gandelman, R. (1983). Contiguity to males in utero affects avoidance responding in adult female mice. Science (Washington, D.C.) 220, 437-438. Herrenkohl, L. R., and Gala, R. R. (1979). Serum prolactin levels and maintenance of progeny by prenatally stressed female offspring. Experientia 35, 702-704. Herrenkohl, L. R., and Politch, J. A. (1978). Effects of prenatal stress on the estrous cycle of female offspring as adults. Experientiu 34, 1240-1241. Herrenkohl, L. R., and Whitney, J. B. (1976). Effects of prepartal stress on postpartal nursing behavior, litter development and adult sexual behavior. Physiol. Behuv. 17, 1019-1021. Ingersoll, D. W., Bobotas, G., Ching-Tse, L., and Lukton, A. (1982). Glucuronidase activation of latent aggression promoting cues in mouse bladder urine. Physiol. Behuv. 29, 789-793. Jean-Faucher, C. Berger, M., De Turckheim, M., Veyessiere, G., and Jean, C. (1981). Effects of dense housing on growth of reproductive organs, plasma testosterone levels and fertility in male mice. J. Endocrinol. 90, 397-402. Joffe, J. M. (1978). Hormonal mediation of the effects of prenatal stress on offspring behaviour. In G. Gottlieb (Ed.), Studies on the Development of Behuviour and the Nervous system. Early InJluences. Academic Press, New York. Lane, E. A., and Hyde, T. S. (1973). The effects of maternal stress on fertility and sex ratio-A pilot study with rats. J. Abnorm. Psycho/. 82, 78-80. McKinney, T. D. (1972). Estrous cycle in house mice: Effects of grouping, preputial gland odors and handling. J. Mammal. 53, 391-393. Meisel, R. L., Dohanich, G. P., and Ward, I. L. (1979). Effects of prenatal stress on avoidance acquisition, open field performance and lordotic behavior in male rats. Physiol.
Behuv.
22, 527-530. Stress in Female Activity and the Oestrous
Nichols, D. J. (1980). Social
Mice:
Effects
of Differentiul
Housing
on
Cycle. Ph.D. thesis, University of Keele. Politch, J. A., and Herrenkohl, L. R. (1984a). Effects of prenatal stress on reproduction in male and female mice. Physiol. Behuv. 32, 95-100. Politch, J. A., and Herrenkohl, L. R. (1984b). Prenatal ACTH and corticosterone: Effects on reproduction in male mice. Physiol. Behav. 32, 135-138. Rhees, R. W., and Fleming, D. E. (1981). Effects of malnutrition, maternal stress or ACTH injections during pregnancy on sexual behavior of male offspring. Physiol. Behav. 27, 879-883. Rose, R. M. (1969). Androgen responses to stress. Psychosom. Med. 31, 405-417. Simon, N. G., and Gandelman, R. (1977). Decreased aggressive behavior in offspring of ACTH treated mice. Behav. Biol. 21, 478-488. Adrenocorticul
PRENATAL
STRESS AND AGGRESSION
97
Simon, N. G., and Gandelman, R. (1981). Threat postures signal impending attack in mice. Behav. Neural Biol. 33, 509-514. Stylianopoulou, F. (1983). Effect of maternal adrenocorticotropin injections on the differentiation of sexual behavior of the offspring. Harm. Behav. 17, 324-331. Vom Saal, F. S. (1983). Variation in infanticide and parental behavior in male mice due to prior intrauterine proximity to female fetuses: Elimination by prenatal stress. Physiol. Behav. 30, 675-681. Vom Saal, F. S., and Bronson, F. H. (1978). In utero proximity of female mouse fetuses to males: Effect on reproductive performance in later life. Biol. Reprod. 19, 842-853. Vom &al, F. S., and Bronson, F. H. (1980). Variation in length of the estrous cycle in mice due to former intrauterine proximity to male fetuses. Biol. Reprod. 22, 777-780. Vom Saal, F. S., Grant, W., McMullen, C., and Laves, K. (1983). High fetal estrogen concentrations: Correlation with increased adult sexual performance and decreased aggression in male mice. Science (Washington, D.C.) 220, 1306-1308. Ward, I. L. (1972). Prenatal stress feminizes and demasculinizes the behavior of males. Science (Washington, DE.) 175, 82-84. Ward, I. L., and Weisz, J. (1980). Maternal stress alters plasma testosterone in fetal males. Science
(Washington,
D.C.)
207, 328-329.