Hoffman, H-J., Schneider, R. & Crusio, Wim E. (1993) Genetic Analysis of Isolation-Induced Aggression. II. Postnatal Environmental Influences in AB Mice. Behaviour Genetics 23 391-394

Genetic Analysis of Isolation-Induced Aggression. II. Postnatal Environmental Influences in AB Mice

Hans-Jürgen Hoffmann1, Regine Schneider1, and Wim E. Crusio2,3

  1. Institute of Brain Research, Brenneckestraße 6, D-O-3090 Magdeburg, Germany.
  2. Génétique, Neurogénétique et Comportement, URA 1294 CNRS, UFR Biomédicale, Université René Descartes (Paris V), 45 rue des Saints-Pères, 75270 Paris Cedex 06, France.
  3. To whom correspondence should be addressed.

Send correspondence and proofs to: Dr. Wim E. Crusio at the above address.


Recently, we reported on two closely-related inbred mouse strains, ABG and AB//Halle, that display extreme differences in isolation-induced intermale aggression. In the present article we investigated the influence of both maternal and social postnatal environmental influences. No effects were found of the postnatal maternal environment. Likewise, whether animals after weaning were housed together in same-strain or mixed-strain groups did not influence their subsequent aggressive behavior. We conclude that the aggressive behavior of ABG and AB//Halle is rather robust with regard to postnatal environmental modification and that the difference between the two strains is most likely due to only few genetic factors.

Keywords: isolation-induced aggression; genotype; postnatal maternal effects; fostering; inbred strains of mice.


Phenotypic differences observed between inbred strains that have been reared in standardized laboratory environments reflect genetic divergence. However, the physiological pathways leading from genotype to phenotype may vary widely from the usually-assumed direct one from chromosomal locus to phenotype: strain differences may also be influenced directly by non-chromosomal sources (i.e., mitochondrial genes), or indirectly, by pre- and/or postnatal maternal factors (Roubertoux et al., 1990). Prenatal maternal influences and non-chromosomal inheritance may be tested by investigating reciprocal crosses in combination with egg or ovary transplantations (Carlier et al., 1992). The relative contribution of the postnatal maternal environment to some observed strain difference can be detected by using crossfostering techniques (van Abeelen, 1980).

When employing crossfostering techniques, care should be taken to include two necessary control groups in the experimental design. Of course, we have to use unfostered animals but, in addition, infostered animals (that is, animals fostered to dams belonging to the same strain) should be studied to check for possible effects of the fostering procedure per se (Bovet-Nitti et al., 1968; Southwick, 1968; van Abeelen, 1980).

Recently, we reported on extreme differences in spontaneous intermale aggression in two closely related inbred mouse strains: ABG and AB//Halle (Schneider et al., 1992). These differences were even more pronounced after a two-week isolation period, with longer isolation periods having no influence on the behavior of the non-aggressive strain ABG (Schneider et al., 1992). The objective of the present experiments was to assess the importance of genotype-dependent, postnatal environmental influences on isolation-induced aggression in these two strains. With regard to the postnatal maternal environment, we compared unfostered, infostered and reciprocally outfostered groups of animals with each other. To investigate the effects of the postnatal social environment, we examined animals that had been reared together with their littermates from the same strain or in mixed groups together with animals from the opposite strain.



The inbred strains ABG and AB//Halle have been derived separately from the same partially inbred stock in the early 1960s (see Schneider et al., 1992, for more details). Both strains had been maintained by brother-sister mating in our laboratory in Magdeburg for four to six generations at the start of the experiment. The mice were kept under controlled laboratory conditions with a light regimen of LD 12:12 (lights on at 0600) and were provided with food (Altromin 1320 rat and mouse maintenance diet) and water ad libitum. Soiled bedding was changed weekly. For matings, three females and one male were housed together. Females were removed and housed individually immediately after diagnosis of pregnancy in cages measuring 25 x 40 x 18 cm (type I). In the morning of their day of birth, animals were either left with their own dams (unfostered group), or transferred to mothers that had produced offspring on the same day, belonging to either the same or the other strain (infostered and outfostered groups, respectively). The litters were culled to a maximum of eight pups and weaned three weeks after parturition. Males from each litter were then caged in groups of two to six for two weeks (type I cages). However, in a second experiment, a subgroup of the unfostered animals was housed in mixed-strain groups consisting of three males from each strain. Thereafter, subjects were housed individually in smaller plastic cages (type II: 10 x 40 x 10 cm).

Behavioral testing

Experiments were carried out in the breeding room between 0900 and 1300. Following a two-week isolation period, the test male (aged 50 ± 2 days) was put directly into the type I test cage and allowed to habituate for 1 min. An opponent male of the inbred strain C3H/Ola, aged 60­100 days and selected randomly from among the animals available at any given moment, was then introduced. Animals from this strain attack only rarely in this situation and are thus well-suited as standard opponents (Schneider et al., 1992). Each opponent was used once daily and up to five times in total at most, to minimize the influence of previous fighting experience. The test cage was changed after each test session in order to avoid influences of residual odors. Aggressive behavior was assessed following the methods of Hahn and Haber (1982) and Selmanoff et al. (1976) during a 10 min observation period. The frequencies of the following behaviors were registered directly and continuously by either one of two different observers: tail rattling (TR), rapid lateral quivering or thrashing of the tail; attacks (AT), biting of the opponent; and aggressive grooming (AG), vigorous grooming of the opponent mouse from lateral position using teeth and forpaws. Data were expressed as frequencies per minute. In addition, the latency to the first attack was recorded in sec (L1).

Statistical analysis

For statistical analysis we carried out two-way ANOVAs with STRAIN and GROUP as main factors, employing the SPSS and SAS packages for PC (Norusis, 1988; SAS Institute Inc., 1987). Because distributions were not necessarily normal, the results of the ANOVAs were checked by means of nonparametric Kruskal-Wallis (crossfostering experiment) or Mann-Whitney U-tests (socialization experiment; Siegel, 1956) to test for differences between groups within strains. We only present the results obtained with the nonparametric tests if they deviated from those obtained with the ANOVAs.

Results and Discussion

The results of the crossfostering experiment are entered in Table I. The unfostered control groups show large differences between strains for all variables investigated, confirming our previous findings (Schneider et al., 1992). This picture does not change at all when similar comparisons are made between the infostered or outfostered groups. This result is confirmed by ANOVAs, that indicate significant strain differences for all variables measured (TR: F1,149 = 45.1; AT: F1,149 = 52.8; AG: F1,149 = 21.7; L1: F1,149 = 64.7; all P < 0.001).

For none of the variables did the fostering procedure affect the behavior of either strain, as neither the main effect of fostering (TR: F2,149 = 0.6; AT: F2,149 = 0.8; AG: F2,149 = 0.6; L1: F2,149 = 0.7), nor its interaction with the strain effect (TR: F2,149 = 0.7; AT: F2,149 = 0.5; AG: F2,149 = 1.2; L1: F2,149 = 0.1) was significant for any of the variables under consideration. We may therefore conclude that neither fostering per se nor crossfostering has any appreciable influence on the aggressive behavior of these two strains.

Southwick (1968) reported that the aggressive behavior of crossfostered animals tended to resemble that of the opposite strain in one (A/J) of his two strains. This result need not be interpreted as contradicting our findings. Southwick (1968) used the relatively unrelated inbred strains A/J and CFW, whereas we used two closely related strains derived from the same partially-inbred stock. Therefore, as we have argued before (Schneider et al., 1992), only few genotype-dependent differences will be expected between ABG and AB//Halle. It is thus not at all surprising to find that mothers of both strains apparently provide similar postnatal maternal environments to their pups. However, even crossfostering between strains as unrelated as CBA/H and NZB did not have any effects on the intermale aggression of animals from these strains (Roubertoux and Carlier, 1988). Maxson (1992) reviewed crossfostering experiments in which aggression was studied. In 7 out of 9 studies, no effects of fostering were reported. Our results reinforce Maxson's conclusion that, generally speaking, the postnatal maternal environment appears to exert only little influence on the expression of intermale aggression.

The results of the second experiment are presented in Table II. We obtain the expected large strain differences in both the same-strain and the mixed-strain groups: for all variables analyzed, the main effect of strain was significant (TR: F1,110 = 59.8; AT: F1,110 = 60.4; AG: F1,110 = 26.2; L1: F1,110 = 69.7; all P < 0.001). Whether animals were kept together with their littermates of the same strain or together with animals of the other strain, did not appear to influence their subsequent aggressive behavior, as neither a group effect (TR: F1,110 = 0.0; AT: F1,110 = 0.0; L1: F1,110 = 1.2) nor an interaction between housing condition and strain appeared (TR: F1,110 = 0.3; AT: F1,110 = 0.5; AG: F1,110 = 0.2; L1: F1,110 = 0.9). AG was the only exception (group effect: F1,110 = 4.4, P < 0.05), but this appears to be due to distributional problems as nonparametric Mann-Whitney U-tests failed to reveal effects of rearing condition within either strain (ABG: z = 1.15, P = 0.25; AB//Halle: z = 1.82, P = 0.07).

In conclusion, the present study shows that postnatal environmental factors, be they maternal or social, are not or only weakly involved in the genesis of the very large differences in intermale aggressive behavior as shown by the closely related inbred strains ABG and AB//Halle. Although the existence of prenatal influences cannot yet be excluded, it appears highly probable that the behavioral differences are due to genetic divergence. Given the close genetic relationship between these strains, it may be hypothesized that the number of genetic factors involved will be low, possibly only one. A classical Mendelian cross-breeding study will be carried out to test this hypothesis.


We thank Dr. Horst Schicknick for help with the statistical analyses, Antje Leddin (Magdeburg) for expertly conducting a part of the aggression tests, and Profs. Pierre Roubertoux, Michèle Carlier (Paris), and Stephen Maxson (Storrs) for useful discussion.


Bovet-Nitti, F., Oliverio, A., and Bovet, D. (1968). Effects of cross-fostering on avoidance learning and freezing behavior of DBA/2J and C3H/He inbred mice. Life Sci. 7:791-797.

Carlier, M., Nosten-Bertrand, M., and Michard-Vanhée, C. (1992). Separating genetic effects from maternal environmental effects. In Goldowitz, D., Wahlsten, D., and Wimer, R. (eds.), Techniques for the Genetic Analysis of Brain and Behavior: Focus on the Mouse, Techniques in the Behavioral and Neural Sciences, Volume 8, Elsevier, Amsterdam, pp. 111-126.

Hahn, M.E. and Haber, S.B. (1982). The inheritance of agonistic behavior in male mice: A diallel analysis. Aggress. Behav. 8:19-38.

Maxson, S.C. (1992). Methodological issues in genetic analyses of an agonistic behavior (offense) in male mice. In Goldowitz, D., Wahlsten, D., and Wimer, R. (eds.), Techniques for the Genetic Analysis of Brain and Behavior: Focus on the Mouse, Techniques in the Behavioral and Neural Sciences, Volume 8, Elsevier, Amsterdam, pp. 349-373.

Norusis, M.J. (1988). SPSS/PC+ Advanced Statistics V2.0, SPSS Inc.

Roubertoux, P. and Carlier, M. (1988). Differences between CBA/H and NZB mice on intermale aggression. II. Maternal effects. Behav. Genet. 18:175-184.

Roubertoux, P., Nosten-Bertrand, M., and Carlier, M. (1990). Additive and interactive effects of genotype and maternal environment. Adv. Study Behav. 19:205-247.

SAS Institute Inc. (1987) SAS/STAT Guide for Personal Computers, Version 6 Edition. SAS Institute Inc, Cary, NC.

Schneider, R., Hoffmann, J., Schicknick, H., and Moutier, R. (1992). Genetic analysis of isolation-induced aggression. I. Comparison between closely related inbred mouse strains. Behav. Neural Biol. 57:198-204.

Selmanoff, M.K., Maxson, S.C., and Ginsburg, B.E. (1976). Chromosomal determinants of intermale aggressive behavior in inbred mice. Behav. Genet. 6:53-69.

Siegel, S. (1956). Nonparametric statistics for the behavioral sciences, McGraw-Hill, New York.

Southwick, C.H. (1968). Effects of maternal environment on aggressive behavior of inbred mice. Commun. Behav. Biol. 1:129-132.

van Abeelen, J.H.F. (1980). Direct genetic and maternal influences on behavior and growth in two inbred mouse strains. Behav. Genet. 10:545-551.

Table I. Components of Isolation-Induced Aggressive Behavior in Unfostered, Infostered, and Outfostered Male Mice from Two Closely-Related Inbred Strains
Variable Strain Unfostered Infostered Outfostered
Mean ± SEM Mean ± SEM Mean ± SEM
Na ABG 34 11 13
AB//Halle 41 28 28
TR ABG 0.07 ± 0.02 0.05 ± 0.03 0.18 ± 0.09
AB//Halle 0.71 ± 0.10 0.96 ± 0.13 0.81 ± 0.17
AT ABG 0.41 ± 0.13 0.47 ± 0.22 0.35 ± 0.19
AB//Halle 2.42 ± 0.27 3.15 ± 0.50 2.28 ± 0.33
AG ABG 0.44 ± 0.05 0.52 ± 0.06 0.39 ± 0.08
AB//Halle 0.79 ± 0.06 0.66 ± 0.05 0.70 ± 0.06
L1 ABG 499 ± 25 541 ± 29 535 ± 33
AB//Halle 261 ± 28 279 ± 40 300 ± 32

aN, number of subjects; TR, tail rattling; AT, attacks; AG, aggressive grooming; L1, latency to first attack (sec).

Table II. Components of Isolation-Induced Aggressive Behavior in Male Mice from Two Closely-Related Inbred Strains Housed In Same-Strain or Mixed-Strain Groups
Housing Condition Same straina Mixed strain
Variable Strain Mean ± SEM Mean ± SEM
Nb ABG 34 18
AB//Halle 41 21
TR ABG 0.07 ± 0.02 0.02 ± 0.01
AB//Halle 0.71 ± 0.10 0.74 ± 0.10
AT ABG 0.41 ± 0.13 0.23 ± 0.13
AB//Halle 2.42 ± 0.27 2.63 ± 0.43
AG ABG 0.44 ± 0.05 0.34 ± 0.05
AB//Halle 0.79 ± 0.06 0.63 ± 0.06
L1 ABG 499 ± 25 566 ± 19
AB//Halle 261 ± 28 266 ± 44

aSame data as Table I, unfostered group.

bN, number of subjects; TR, tail rattling; AT, attacks; AG, aggressive grooming; L1, latency to first attack (sec).