SEARCH

SEARCH BY CITATION

Keywords:

  • Affiliation;
  • knockout mice;
  • conditional knockout mice;
  • cre recombinase;
  • forebrain;
  • memory

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Acknowledgments

We studied three lines of oxytocin (Oxt) and oxytocin receptor (Oxtr) knockout (KO) male mice [Oxt−/−, total Oxtr−/− and partial forebrain Oxtr (OxtrFB/FB)] with established deficits in social recognition to further refine our understanding of their deficits with regard to stimulus female's strain. We used a modified social discrimination paradigm in which subjects are singly housed only for the duration of the test. Additionally, stimulus females are singly housed throughout testing and are presented within corrals for rapid comparison of investigation by subject males. Wild-type (WT) males from all three lines discriminated between familiar and novel females of three different strains (C57BL/6, BALB/c and Swiss-Webster). No KO males discriminated between familiar and novel BALB/c or C57BL/6 females. Male Oxt−/− and Oxtr−/− mice, but not OxtrFB/FB mice, discriminated between familiar and novel Swiss-Webster females. As this might indicate a global deficit in individual recognition for OxtrFB/FB males, we examined their ability to discriminate between females from different strains and compared performance with Oxtr−/− males. WT and KO males from both lines were able to distinguish between familiar and novel females from different strains, indicating the social recognition deficit is not universal. Instead, we hypothesize that the Oxtr is involved in ‘fine’ intrastrain recognition, but is less important in ‘broad’ interstrain recognition. We also present the novel finding of decreased investigation across tests, which is likely an artifact of repeated testing and not because of stimulus female's strain or age of subject males.

Social recognition, the ability to distinguish familiar from novel conspecifics, is imperative for the display of appropriate social behaviors (Ferguson et al. 2002; Winslow & Insel 2002). Oxytocin (Oxt) administration facilitates social recognition (Benelli et al. 1995; Dluzen et al. 1998; Popik & van Ree 1991; Popik et al. 1996) and the development of knockout (KO) mice provided further evidence of Oxt's role in social recognition. Oxt KO mice (Oxt−/−) fail to distinguish between familiar and novel conspecifics (Ferguson et al. 2000, 2001). Two independently derived lines of Oxt receptor KO (Oxtr−/−) mice also have social recognition deficits (Lee et al. 2008; Takayanagi et al. 2005), as does a relatively forebrain-specific Oxtr KO (OxtrFB/FB) (Lee et al. 2008). However, the social recognition deficit differs between Oxtr−/− and OxtrFB/FB males when given the same social recognition task (two trial), whereas OxtrFB/FB males show no social recognition deficit on the five-trial habituation/dishabituation task (Lee et al. 2008). These results indicate a need for further examination of the social recognition deficits of these two Oxtr KO mouse lines.

To do so, we have developed a modified version of the two-trial social discrimination task previously used by Engelmann et al. (1995). However, here we use group-housed subject males, which show better social memory than singly housed subject males (Kogan et al. 2000). We also use singly housed stimulus females to prevent transfer of major urinary proteins between females, which could reduce recognition of individuals by a confusion of odors (see Cheetham et al. 2007). Stimulus females are presented within corrals, which reduces direct contact between the subject and stimulus animals while still eliciting high interest and investigation by the subject (Choleris et al. 2003; Kudryavtseva et al. 2002). As males cannot directly contact corralled females there is no need to extinguish sex behavior before the test, as with other social recognition paradigms (see Ferguson et al. 2001; Lee et al. 2008). Finally, presenting females in corrals allows the use of gonadally intact females, which elicit higher interest from males than ovariectomized females when contact is prevented (Muroi et al. 2006).

The modified two-trial social discrimination task was used to examine social recognition in male Oxt, Oxtr and OxtrFB/FB KOs and their male wild-type (WT) siblings. The ability of WT and KO males from each line to discriminate between females of the same strain was examined to determine whether using stimulus females from inbred (C57BL/6 or BALB/c) or outbred (Swiss-Webster: SW) strains affects recognition abilities. WT males from all three lines distinguish between individual females of the same strain. Oxt−/− and Oxtr−/− males distinguish between SW females, but not females from the inbred strains. However, OxtrFB/FB males have a complete inability to discriminate between females of the same strain. Therefore, the ability of OxtrFB/FB males to distinguish between females from different strains is assessed and compared with Oxtr−/− males. We find that the discrimination deficit is not universal, but is instead specific to intrastrain recognition.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Acknowledgments

Animals and housing

The development and genotyping of the Oxt−/−, Oxtr−/− and OxtrFB/FB mice were described previously (Lee et al. 2008; Young et al. 1996). Oxt KO subjects were originally an equal mix of the C57BL/6J and 129X1/SvJ strains (RW4 embryonic stem cell line; Shipley et al. 1996) but have been repetitively backcrossed for greater than 10 generations with C57BL/6J mice (from Jackson Labs, Bar Harbor, ME, USA). The Oxtr gene for recombination was amplified by polymerase chain reaction (PCR) from 129/Sv DNA and targeted into the R1 (from 129X1/SvJ x 129S1 cross; Nagy et al. 1993) embryonic stem cell line (Lee et al. 2008). Germline transmission then resulted in an equal mix of the C57BL/6J and 129 strains. To generate forebrain-specific Oxtr KOs, we crossed the L7ag13 transgenic line (C57BL/6J genetic background) that expresses Cre recombinase under the control of the Camk2a promoter (Dragatsis & Zeitlin 2000; Zakharenko et al. 2003) with Oxtr+/flox or Oxtrflox/flox mice that had been generated by crossing with Flp recombinase-expressing mice (Jackson Labs, stock no. 003800; C57BL/6J background) to eliminate the neomycin resistance cassette. Oxtrflox/flox male mice were crossed with female Oxtr+/flox mice that contained one transgenic allele expressing Cre recombinase (Oxtr+/flox,cre or Oxtr+/FB). The offspring thus had the following genotypes: (1) Oxtr+/flox, (2) Oxtrflox/flox, (3) Oxtr+/flox,cre and (4) Oxtrflox/flox,cre. The former two are considered WT, the third heterozygous forebrain inactivation and the fourth one forebrain-specific Oxtr KO (OxtrFB/FB). To generate whole body Oxtr KO mice, we bred male Oxtr+/flox,cre mice, which have germ cell expression of Cre recombinase, with female Oxtrflox/flox mice (Bastia et al. 2005). This led to heterozygous progeny with one Oxtr allele inactivated (Oxtr+/−). We crossed these mice to get homozygous total Oxtr KO (Oxtr−/−) mice. Further details are available in the original description (Lee et al. 2008). As both the transgenic lines expressing Flp or Cre recombinase were on a C57BL/6J background, the resulting Oxtr−/− and OxtrFB/FB mice studied here were approximately 88% and 81% C57BL/6J (rest being 129/S), respectively. All OxtrFB/FB mice are heterozygous for the Cre recombinase transgene.

All subjects used were littermates from crosses of heterozygote mice. Within each strain, subjects were taken from six to seven breeders, with equivalent numbers of WT and KO subjects tested from each breeder (see Crusio et al. 2009). Subject males were group housed (two to five animals per cage) upon weaning at 21–28 days old in single-sex cages throughout testing. Gonadally intact BALB/c, C57BL/6 and SW female mice were purchased from NCI-Frederick (Frederick, MD, USA) for use as stimulus females. BALB/c females were chosen to compare with our previous social recognition results using this strain (Lee et al. 2008). C57BL/6 mice were chosen because this strain is the primary background strain of our KO lines, and SW females were chosen because they are a readily obtained outbred strain. Stimulus females were group housed (five per cage) upon receipt into the colony, and singly housed 1 week before onset of all experiments.

All animals were maintained under a 12:12 dark cycle (lights on at 0300 h) with food and water available ad libitum. All tests were conducted during the light phase of the light cycle beginning between 930 and 1100 h under dim lighting (approximately 30–40 lux inside the testing cages). All animal procedures were approved by the National Institute of Mental Health Animal Care and Use Committee and were in accordance with the National Institutes of Health guidelines on the care and use of animals.

Experiment 1

Adult males (90–120 days old) from the Oxt KO, Oxtr KO and partial forebrain-specific Oxtr KO lines were tested three times on the modified social discrimination paradigm (Fig. 1). Group sizes differ across tests as data from any male that investigated during either trial 1 or trial 2 for < 30 seconds (<10% of time allotted) was discarded. In the first test, BALB/c females (mean age: 110 days old) were used as stimulus animals [males: (Oxt+/+, n = 10; Oxt−/−, n = 10), (Oxtr+/+, n = 7; Oxtr−/−, n = 9) and (Oxtr+/+, n = 8; OxtrFB/FB, n = 9)]. Two weeks later, the same males were again administered the social discrimination task, with C57BL/6 females (mean age: 80 days old) used as stimulus animals [males: (Oxt+/+, n = 10; Oxt−/−, n = 7), (Oxtr+/+, n = 9; Oxtr−/−, n = 9) and (Oxtr+/+, n = 9; OxtrFB/FB, n = 8)]. In the third test (2 weeks after the second), SW females (mean age: 80 days old) were used as stimulus animals [males: (Oxt+/+, n = 7; Oxt−/−, n = 6), (Oxtr+/+, n = 9; Oxtr−/−, n = 9) and (Oxtr+/+, n = 9; OxtrFB/FB, n = 7)].

image

Figure 1. Schematic representation of the final social discrimination paradigm used in Experiments 1, 2 and 3.

Download figure to PowerPoint

On each test day, group-housed subject males were placed singly into new, clean cages with < 1/8′′ bedding on the cage floor 1 h before start of testing. Males were allowed to acclimate to the new cage for 30 min, at which point two wire corrals (www.kitchen-plus.com, Item #31570) were introduced into the cage (4 1/4′′ high ×4′′ diameter). The space between the wire bars of the corrals permitted passage of volatile olfactory cues from the stimulus female to the subject male without allowing the male to deposit his scent on the stimulus females (Insel & Fernald 2004), but still maintain high interest and investigation times (see Choleris et al. 2003, 2006). Males were given 30 min to habituate to the presence of the corrals, after which a stimulus female was placed into either the left or right corral (placement was counterbalanced across genotype); the second empty corral was removed from the cage. Males were given 5 min of exposure to the stimulus female (trial 1); investigation was scored whenever the male did the following: (1) made direct contact with the female with either the nose or the forepaw; (2) sniffed or reached inside of the lower half of the corral or (3) sniffed any portion of the female external to the corral (e.g. her tail). The female was removed after 5 min and returned to her home cage. The unused empty corral was returned to the test cage during the 30-min intertrial delay, although not necessarily in the same location as the corrals' left-right placements within the cage were counterbalanced across genotype to avoid development of a side preference.

Upon re-exposure (trial 2), the female from trial 1 (familiar) was placed into the same corral as in trial 1 to avoid cross-contamination of female odors. A novel female (with which the male had no prior experience) was placed into the previously empty corral. Exploration of both familiar and novel corralled females was recorded for 5 min using the same criteria for exploration as during trial 1. Stimulus females were returned to their home cages; subject males were weighed and returned to their group-housed home cage.

Data from each line (Oxt, Oxtr and OxtrFB/FB) were analyzed using spss version 13.0 (Chicago, IL, USA). Time spent investigating (1) familiar and (2) novel females was examined via repeated-measures multivariant analysis of variance (manova) (genotype × test). Post hoc analysis to determine differences across the three tests was carried out using Bonferroni correction. Depending upon significance from the manova, one-way analysis of variance (anova)s compared (1) differences in investigation times within each genotype across the three tests and (2) differences across each genotype within each test. For all three lines, paired samples t-tests were run within each genotype and test to assess ability to discriminate between familiar and novel females. For all statistical tests, significance was set at P < 0.05.

Experiment 2

A new cohort of males from the OxtrFB/FB line was examined once on the social discrimination task to verify their inability to distinguish between SW females. The social discrimination task was administered as described in Experiment 1 (see Fig. 1). Subject males averaged 100 days old, and consisted of three genotypes: Oxtr+/+ (n = 8), OxtrFB/FB (n = 7) and Oxtr+/flox,cre (n = 8). The latter group was tested to determine whether expression of Cre recombinase in the transgenic line used to generate the OxtrFB/FB mice might affect the social recognition phenotype. An anova was run to assess differences across genotypes for time spent investigating (1) familiar and (2) novel females. Paired samples t-tests were run within each genotype to assess ability to discriminate between familiar and novel females. For all statistical tests, significance was set at P < 0.05.

Experiment 3

New cohorts of males from the Oxtr−/− and OxtrFB/FB lines were examined in the social discrimination task to determine ability to discriminate between females of two different strains. The social discrimination task was administered as described in Experiment 1 (see Fig. 1). Adult males from both the Oxtr−/− and OxtrFB/FB lines were tested three times, with 4 weeks between tests. On the first test [(Oxtr line: average 95 days old; Oxtr+/+, n = 9; Oxtr−/−, n = 10) and (OxtrFB/FB line: average 102 days old; Oxtr+/+, n = 12; OxtrFB/FB, n = 11)], a C57BL/6 and a BALB/c female were presented. On the second test [(Oxtr line: average 128 days old; Oxtr+/+, n = 9; Oxtr−/−, n = 10) and (OxtrFB/FB line: average 136 days old; Oxtr+/+, n = 12; OxtrFB/FB, n = 11)], a C57BL/6 and a SW female were presented. On the final test [(Oxtr line: average 155 days old; Oxtr+/+, n = 7; Oxtr−/−, n = 10) and (OxtrFB/FB line: average 166 days old; Oxtr+/+, n = 11; OxtrFB/FB, n = 10)], a BALB/c and a SW female were presented. Strain and left-right position of female presented as ‘familiar’ were counterbalanced across genotype and test. Females were approximately 60 days old on test 1, 85 days old on test 2 and 110 days old on test 3. Statistical analysis was identical to that described for Experiment 1.

Experiment 4

The results from Experiments 1 and 2 indicated that age could be a factor in the decreased exploration observed from tests 1 to 3. Therefore, comparisons were made between old (160–190 days; Oxtr+/+, n = 7; OxtrFB/FB, n = 5) and young (80–95 days; Oxtr+/+, n = 9; OxtrFB/FB, n = 8) males. Subjects were tested once using BALB/c females (approximately 60 days old) as stimulus animals. A two-way manova (genotype × age) was run to assess differences across genotypes and ages for time spent investigating (1) familiar and (2) novel females. Paired samples t-tests were run within each genotype for each age to assess ability to discriminate between familiar and novel females. For all statistical tests, significance was set at P < 0.05.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Acknowledgments

Experiment 1

Social discrimination in Oxt KO males

Oxt+/+ and Oxt−/− males were assessed three times on the social discrimination task (Fig. 2). A repeated-measures manova showed significant main effects of genotype (WT, KO; F1,44 = 3.86, P < 0.05), female (familiar, novel; F1,44 = 39.55, P < 0.001) and test (1, 2, 3; F2,44 = 9.07, P < 0.001). Generally, Oxt+/+ males investigated all females more than Oxt−/− males, and novel females were investigated more than familiar females by all subjects. Post hoc analysis via Bonferroni correction showed decreased investigation times for familiar and novel females on test 3 (SW females) compared with test 1 (BALB/c females; P < 0.001) and test 2 (C57BL/6 females; P < 0.001; Fig. 2). There was a trend toward a genotype × female interaction (F1,44 = 3.39, P = 0.07), but no other significant interactions. Accordingly, anovas comparing within-genotype differences across the three tests and between-genotype differences within each test could not be carried out.

image

Figure 2. Social recognition memory in Oxt WT and KO mice. Data are investigation times in seconds (mean ± SEM) and were analyzed via repeated-measures manova (genotype × test). (a) Oxt+/+ (WT) males significantly discriminated between familiar and novel BALB/c stimulus females (left; **P < 0.01), C57BL/6 stimulus females (middle; *P < 0.05) and SW females (right; **P < 0.01). (b) Oxt−/− (KO) males were unable to discriminate between familiar and novel BALB/c stimulus females (left) or C57BL/6 stimulus females (middle), but discriminated between familiar and novel SW females (right; **P < 0.01).

Download figure to PowerPoint

Within each genotype for each of the three tests, investigation times of familiar and novel females were assessed using paired samples t-tests. Oxt+/+ males (Fig. 2a) discriminated between familiar and novel BALB/c females (P < 0.05; left), C57BL/6 females (P < 0.05; middle) and SW females (P < 0.01; right), as measured by a significant difference in investigation times. In contrast, Oxt−/− males (Fig. 2b) did not discriminate between familiar and novel BALB/c females (P > 0.08; left) or C57BL/6 females (P > 0.10; middle), but did discriminate between familiar and novel SW females (P < 0.01; right).

Social discrimination in Oxtr KO males

Oxtr+/+ and Oxtr−/− males were tested three times on the social discrimination task (Fig. 3). A repeated-measures manova showed a significant main effect of female (familiar, novel; F1,46 = 72.81, P < 0.001) and test (1, 2, 3; F2,46 = 11.32, P < 0.001). Generally, all subjects investigated the novel females more than the familiar females. Post hoc analysis via Bonferroni correction showed decreased investigation times on test 3 (SW females) compared with test 1 (BALB/c females; P < 0.05) and test 2 (C57BL/6 females; P < 0.001; Fig. 3).

image

Figure 3. Social recognition memory in Oxtr WT and KO mice. Data are investigation times in seconds (mean ± SEM) and were analyzed via repeated-measures manova (genotype × test). (a) Oxtr+/+ (WT) males significantly discriminated between familiar and novel BALB/c stimulus females (left; ***P < 0.001), C57BL/6 stimulus females (middle; *P < 0.05) and SW females (right; *P < 0.05). Oxtr+/+ males investigated novel BALB/c females significantly more than novel C57BL/6 or SW females (##P < 0.01). (b) Oxtr−/− (KO) males were unable to discriminate between familiar and novel BALB/c stimulus females (left) or C57BL/6 stimulus females (middle), but discriminated between familiar and novel SW females (right; *P < 0.05). Oxtr−/− males investigated familiar SW females significantly less than familiar BALB/c or C57BL/6 females (#P < 0.05).

Download figure to PowerPoint

There was no significant main effect of genotype (WT, KO) (F1,46 = 2.88, P > 0.05), but there were significant female × genotype (F1,46 = 8.73, P < 0.01) and female × genotype × test (F2,46 = 5.19, P < 0.01) interactions. Therefore, data from Oxtr+/+ and Oxtr−/− males were analyzed separately via one-way anovas comparing performances within genotype across the three tests. Oxtr+/+ males differed across the three tests in investigation of the novel females (F2,24 = 11.93, P < 0.001). Post hoc analysis via Bonferroni correction showed that Oxtr+/+ males spent more time investigating the novel females during test 1 (BALB/c females) than during test 2 (C57BL/6 females) and test 3 (SW females); P < 0.01 for both (Fig. 3a). Oxtr−/− males differed across the three tests in investigation of the familiar females (F2,24 = 6.10, P < 0.01). Post hoc analysis via Bonferonni correction showed that Oxtr−/− males spent significantly less time investigating the familiar females during test 3 than test 1 or test 2 (P < 0.05; Fig. 3b).

Within each genotype for each of the three tests, investigation times of familiar and novel females were assessed using paired samples t-tests. Oxtr+/+ males (Fig. 3a) discriminated between familiar and novel BALB/c females (P < 0.001; left), C57BL/6 females (P < 0.05; middle) and SW females (P < 0.05; right), as measured by a significant difference in investigation times. In contrast, Oxtr−/− males (Fig. 3b) did not discriminate between familiar and novel BALB/c females (P > 0.08; left) or C57BL/6 females (P > 0.10; middle), but did discriminate between familiar and novel SW females (P < 0.01; right).

Social discrimination in OxtrFB/FB males

Oxtr+/+ and OxtrFB/FB males were tested three times on the social discrimination task (Fig. 4). A repeated-measures manova showed significant main effects of female (familiar, novel; F1,44 = 41.00, P < 0.001) and test (1, 2, 3; F2,44 = 9.58, P < 0.001). Generally, Oxtr+/+ males investigated all females more than OxtrFB/FB males, and all subjects investigated the novel females more than the familiar females. Post hoc analysis via Bonferroni correction showed decreased investigation times for familiar and novel females on test 3 (SW females) compared with test 1 (BALB/c females; P < 0.01) and test 2 (C57BL/6 females; P < 0.001; Fig. 4).

image

Figure 4. Social recognition memory in OxtrFB/FB WT and KO mice. Data are investigation times in seconds (mean ± SEM) and were analyzed via repeated-measures manova (genotype × test). (a) Oxtr+/+ (WT) males significantly discriminated between familiar and novel BALB/c stimulus females (left; **P < 0.01), C57BL/6 stimulus females (middle; **P < 0.01) and SW females (right; *P < 0.05). Oxtr+/+ males investigated novel SW females significantly less than novel BALB/c or C57BL/6 females (#P < 0.05). (b) OxtrFB/FB (KO) males were unable to discriminate between familiar and novel BALB/c stimulus females (left), C57BL/6 stimulus females (middle) and SW females (right). OxtrFB/FB males investigated familiar SW females significantly less than familiar BALB/c or C57BL/6 females (###P < 0.001).

Download figure to PowerPoint

There was no significant main effect of genotype (WT, KO; F1,44 = 3.79, P = 0.06), but there was a significant female × genotype interaction (F1,44 = 6.17, P < 0.05). No other interactions were significant. To examine the interaction, data from Oxtr+/+ and OxtrFB/FB males were analyzed separately via one-way anova comparing performances within genotype across the three tests. Oxtr+/+ males differed across the three tests in investigation of the novel females (F2,25 = 3.98, P < 0.05). Post hoc analysis via Bonferroni correction showed that Oxtr+/+ males spent less time investigating the novel females during test 3 than during test 1 or test 2 (P < 0.05; Fig. 4a). OxtrFB/FB males differed across the three tests in investigation of the familiar females (F2,23 = 5.38, P < 0.05). Post hoc analysis via Bonferroni correction showed that OxtrFB/FB males spent significantly less time investigating the familiar females during test 3 than test 1 (P < 0.001; Fig. 4b).

Within each genotype for each of the three tests, investigation times of familiar and novel females were assessed using paired samples t-tests. Oxtr+/+ males (Fig. 4a) discriminated between familiar and novel BALB/c females (P < 0.01; left), C57BL/6 females (P < 0.01; middle) and SW females (P < 0.05; right), as measured by a significant difference in investigation times. In contrast, OxtrFB/FB males (Fig. 4b) did not discriminate between familiar and novel BALB/c females (P > 0.10; left), C57BL/6 females (P > 0.08; middle) or SW females (P > 0.10; right).

Experiment 2

A new cohort of males from the OxtrFB/FB line was tested once on the social discrimination task using SW females as stimulus animals (Fig. 5). An anova showed a significant effect of genotype (F4,40 = 3.22, P < 0.05) and a significant genotype × familiar female interaction (F2,20 = 4.35, P < 0.05). Post hoc analysis via Bonferroni correction showed that OxtrFB/FB males spent significantly more time investigating familiar females than did Oxtr+/+ or Oxtr+/flox,cre males. Within each genotype, differences in investigation times of familiar and novel females were assessed using paired samples t-tests. Oxtr+/+ males (Figure 5, left) and Oxtr+/flox,cre males (Fig. 5, middle) both discriminated between familiar and novel SW females (P < 0.01 and P < 0.001, respectively), whereas OxtrFB/FB males (which are also heterozygous for the Cre recombinase transgene) spent equivalent times with familiar and novel SW females (Fig. 5, right; P > 0.10).

image

Figure 5. Presence of the Cre transgene does not affect social recognition memory. Social recognition memory for SW females in Oxtr WT, Oxtr+/flox,cre (heterozygous for both forebrain Oxtr inactivation and the Cre recombinase transgene) and OxtrFB/FB (homozygous for forebrain Oxtr inactivation and heterozygous for the Cre recombinase transgene) mice. Data are investigation times in sections (mean ± SEM) and were analyzed via paired samples t-tests. WT (Oxtr+/+, left) and heterozygote (Oxtr+/flox,cre, middle) males both discriminated between familiar and novel SW stimulus females (*P < 0.05, **P < 0.01). KO males (OxtrFB/FB) did not.

Download figure to PowerPoint

As there were no differences across genotypes in investigation of the novel females (F2,20 = 2.05, P > 0.10), these results indicate that the inability of OxtrFB/FB males to discriminate between familiar and novel SW females is because of increased investigation of the familiar females. These results are different from those observed in Experiment 1, in which OxtrFB/FB males seemed to spend slightly more time investigating the novel SW females, although not significantly so (Fig. 4, right). These differences in investigation of the familiar females are likely an artifact of the repeated testing design used in Experiment 1, as SW females were presented on the third test in Experiment 1, but on the first test in this experiment.

Experiment 3

New cohorts of males from the Oxtr+/+ and OxtrFB/FB lines were tested three times each on the social discrimination task, using females from two different strains as stimulus animals at each test (test 1: C57BL/6 vs. BALB/c; test 2: C57BL/6 vs. SW and test 3: BALB/c vs. SW) (Fig. 6).

image

Figure 6. Social recognition memory in two lines of Oxtr mice for females of different strains. Data are investigation times in seconds (mean ± SEM) and were analyzed via repeated-measures manova (genotype × test). (a) WT (Oxtr+/+) and (b) KO (Oxtr−/−) males from the total Oxtr line. (c) WT (Oxtr+/+) and (d) KO (OxtrFB/FB) males from the partial forebrain-specific Oxtr line all successfully discriminated between familiar and novel stimulus females (*P < 0.05, **P < 0.01, ***P < 0.001). OxtrFB/FB males investigated the novel females significantly more on the first test (d, left) than the second (d, middle) or third (d, right) tests (#P < 0.05).

Download figure to PowerPoint

Oxtr−/− males

A repeated-measures manova showed significant effects of female (familiar, novel; F1,49 = 195.42, P < 0.001) and test (1, 2, 3; F2,49 = 6.77, P < 0.01). Generally, all subjects investigated the novel females more than the familiar females. Post hoc analysis via Bonferonni correction showed decreased investigation times for test 3 (P < 0.01) compared with test 1. There were no other significant main effects of genotype and no significant interactions. Accordingly, anovas comparing within-genotype differences across the three tests and between-genotype differences within each test could not be carried out.

Within each genotype for each of the three tests, investigation times of familiar and novel females were assessed using paired samples t-tests. Oxtr+/+ males (Fig. 6a) investigated the novel females more than the familiar females on test 1 (P < 0.001; left), test 2 (P < 0.01; middle) and test 3 (P < 0.001; right). Oxtr−/− males (Fig. 6b) also discriminated between novel and familiar females, spending significantly more time investigating the novel females on test 1 (P < 0.001; left), test 2 (P < 0.001; middle) and test 3 (P < 0.05; right).

OxtrFB/FB males

A repeated-measures manova showed significant effects of genotype (WT, KO; F1,61 = 5.23, P < 0.05), female (familiar, novel; F1,61 = 82.18, P < 0.001) and test (1, 2, 3; F1,61 = 21.33, P < 0.001). Generally, OxtrFB/FB males investigated all females more than Oxt+/+ males, and novel females were investigated more than familiar females by all subjects. Post hoc analysis via Bonferonni correction showed decreased investigation times by all subjects for test 2 (P < 0.01) and test 3 (P < 0.001) compared with test 1.

There was a significant female × test interaction (F2,31 = 8.31, P < 0.001), but no other significant interactions. To examine the interaction, data from Oxtr+/+ and OxtrFB/FB males were analyzed separately with a one-way anova comparing performance within genotype across the three tests. Oxtr+/+ males did not differ in investigation of familiar or novel females across the three tests (Fig. 6c), whereas OxtrFB/FB males differed in investigation of the novel females across the three tests (F2,31 = 8.31, P < 0.001). Post hoc analysis via Bonferroni correction showed that OxtrFB/FB males spent significantly more time investigating the novel females during test 1 than test 2 or test 3 (P < 0.05; Fig. 6d).

Within each genotype for each of the three tests, investigation times of familiar and novel females were assessed using paired samples t-tests. Oxtr+/+ males (Fig. 6c) investigated the novel females more than the familiar females on test 1 (P < 0.01; left), test 2 (P < 0.001; middle) and test 3 (P < 0.01; right). OxtrFB/FB males (Fig. 6d) also discriminated between novel and familiar females, spending significantly more time investigating the novel females on test 1 (P < 0.01; left), test 2 (P < 0.05; middle) and test 3 (P < 0. 01; right).

Experiment 4

To determine whether age affected performance on the social discrimination task, old (160–190 days) and young (80–95 days) Oxtr+/+ and OxtrFB/FB males were tested using BALB/c females as stimulus animals. A two-way (genotype × age) manova showed no significant main effects of genotype (WT, KO) or age (old, young) and no significant interaction (P > 0.05 for all). Therefore, age of subject males did not affect investigation of stimulus females. Complementary with results from Experiments 1 and 2, Paired samples t-tests showed that all Oxtr+/+ males (regardless of age) investigated novel females significantly more than familiar females (P < 0.01; Fig. 7, left), whereas OxtrFB/FB males did not (P > 0.08 for both old and young; Fig. 7, right).

image

Figure 7. Social recognition memory in young and old OxtrFB/FB males. Data are investigation times (mean ± SEM) for Oxtr+/+ (WT, left) and OxtrFB/FB (KO, right) males and were analyzed via manova (genotype × age). No differences existed in investigation due to age; only Oxtr+/+ males discriminated between familiar and novel females (**P < 0.01).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Acknowledgments

We used a modified version of a social discrimination task to assess social recognition memory in WT and KO males from Oxt, Oxtr and partial forebrain-specific Oxtr mouse lines. In agreement with previous studies using other social recognition tasks (Ferguson et al. 2000; Lee et al. 2008; Takayanagi et al. 2005), WT males from all three lines discriminated between familiar and novel stimulus females of the same strains (Figs 2–7). This study is the first to show intact social recognition in WT males independent of stimulus female's strain, but impaired social recognition in KO males dependent upon stimulus female's strain. Oxt−/− and Oxtr−/− males did not discriminate between familiar and novel BALB/c or C57BL/6 females (as shown previously in Oxtr lines: Lee et al. 2008; Takayanagi et al. 2005), but did discriminate between familiar and novel outbred SW females (contrary to Choleris et al. 2003 in a different paradigm; Figs 2 and 3). The ability to discriminate between SW females may be specific to the strain and not generalizable to all outbred strains, as Oxt KO mice cannot distinguish in another paradigm between CD-1 mice (Ferguson et al. 2000, 2001).

Feral female mice are more likely to mate within their own strain than with a male from another strain. Oxt−/− females are unable to successfully discriminate between odors from healthy males and those infected with a parasite (Kavaliers et al. 2003). Oxt is therefore implicated in the ability to choose a mate who is parasite-free, more fit and is less likely to infect the female during mating (see Kavaliers et al. 2005), consistent with an increased emphasis on the role of Oxt in intrastrain recognition.

Unlike the Oxt and Oxtr KO males, OxtrFB/FB males did not discriminate between familiar and novel females regardless of stimulus strain, including SW (Fig. 4). OxtrFB/FB males spent only 60% of their time investigating the novel SW females compared with 69% in Oxtr−/− males and 80% in Oxt−/− males. We have previously reported a general decrease in investigation of all stimulus animals (whether familiar or novel) in OxtrFB/FB males, but a maintained recognition memory on the habituation-dishabituation social recognition task (Lee et al. 2008). The current results refine those obtained with the two-trial social recognition task (Lee et al. 2008), clarifying the deficit in social memory in the OxtrFB/FB line.

The targeted mutation is the same in the Oxtr−/− and OxtrFB/FB mice and the three strains differ slightly in genetic background (see Methods). However, as performances on the social discrimination task by WT males from the Oxt and Oxtr lines are identical, it is unlikely that background strain plays a large role. Both total Oxt−/− and Oxtr−/− males show the same social discrimination deficit, indicating that 129Sv flanking DNA is unlikely to be responsible. The presence of Cre recombinase allele in the OxtrFB/FB males could account for the inability of OxtrFB/FB males to discriminate between SW females. However, in Experiment 2 heterozygous mice (for both Cre recombinase and forebrain Oxtr inactivation) had nearly identical performances to the WT males, and discriminated between familiar and novel SW females. Therefore, Experiment 2 indicates that the Cre recombinase transgene does not account for the social recognition defect in OxtrFB/FB males.

Instead, the results from Experiments 1 and 2 indicated that OxtrFB/FB males may have a general inability to recognize individuals. To test this hypothesis, we examined the ability of OxtrFB/FB and Oxtr−/− males to discriminate between females from different strains. We hypothesized that OxtrFB/FB males would not make this discrimination, but WT males and Oxtr−/− males would. Contrary to the hypothesis, WT and KO males from both lines were equally able to discriminate between familiar and novel females of different strains (Fig. 6). Why should OxtrFB/FB males be unable to distinguish between females of the same strain, but able to distinguish between females of different strains? One possibility may be that the postnatal loss of the Oxtr in the OxtrFB/FB in contrast to never expressing Oxtr in the total Oxtr KO leads to the difference. Another possibility may lie with the distribution of the Oxtr within the brain. Deletion of the Oxtr in the OxtrFB/FB conditional line is not complete, with Oxtr expression remaining within the medial amygdala, olfactory bulb and neocortex, but lacking in the hippocampal formation (Lee et al. 2008). As both the amygdala and the hippocampal formation (including the entorhinal cortex) are important for individual recognition (Ferguson et al. 2001; Petrulis 2009, respectively), the altered Oxtr expression pattern in the OxtrFB/FB conditional line may prevent OxtrFB/FB males from discriminating between females of the same strain, even if outbred, but allow discrimination between the greatly different odors of females from different strains. In other words, lack of and/or presence of Oxtr in the above-mentioned regions interferes with ability to make ‘difficult’ discriminations (i.e., between females of the same strain), but is not involved in making ‘easy’ discriminations (i.e., between females of different strains). Results from Oxtr KO males indicate that in the complete absence of Oxtr mice are able to distinguish between different strains. Combined, our data indicates that Oxt acting at the Oxtr may be involved in more detailed individual recognition, but some other peptide or hormone may underlie broad recognition of individuals. For example, follow-up studies using this social discrimination paradigm in vasopressin 1a and 1b receptor KO males are planned to determine whether the closely related peptide may be involved in more general individual recognition.

WT males from all three lines were very consistent in their investigation times, with all spending approximately 75% of the time with the novel females on test 1; 63–68% of the time investigating the novel females on test 2 and 65–70% of the time investigating the novel females on test 3. In contrast, the two-trial social recognition tasks may result in greatly different investigation times of familiar females (compared to novel) (e.g., only ∼ 40% in Lee et al. 2008, but ∼ 75% in Ferguson et al. 2000). Inconsistent performance across KO lines when utilizing the same task makes it difficult to compare results directly. We hope that the current task, which uses group-housed subjects and singly housed gonadally intact stimulus animals contained within corrals, will result in consistent performance across labs using C57BL/6 and C57BL/129 background lines. We have found highly correlated scores from two different intralab investigators (r = 0.991; data not shown), indicating that with appropriate training and exposure to the task, scoring is consistent across investigators.

There is a possible confound of estrous state influencing male investigation with use of gonadally intact females, which were chosen as they elicit higher interest from males than ovariectomized females. However, the day of estrous cycle has previously not been shown to greatly influence male investigation (Muroi et al. 2006). Females in all states of estrous elicit higher investigation from males than do ovariectomized females (Ingersoll & Weinhold 1987). Sexually naïve males (as are the subjects in the current experiments) indicate no preference between receptive and non-receptive females' odors (Carr et al. 1965). Furthermore, subjects of both genotypes would experience any effect of estrous state, thereby eliminating possible genotype × estrous condition interaction.

We believe this study is the first to repeatedly test males' social recognition memory. We find that repeated testing does not interfere with discrimination ability, but does generally result in decreased investigation times by test 3. A previous study indicates that mice from C57BL/6 and 129S2/SvHad strains decrease exploration and activity with repeated testing and handling (Voikar et al. 2004). Males were 2–3 months older at the third test compared with the first. However, we found no overall effect of age on investigation times for both Oxtr+/+ and OxtrFB/FB males (Fig. 7). It is therefore likely that the decrease in investigation in the current study is an artifact of repeated testing. Furthermore, as in the third test in Experiment 1, Experiment 2 (Fig. 5) shows that WT males discriminated in a single test between familiar and novel SW females, while OxtrFB/FB males did not. These results indicate that the social recognition deficit observed in Experiment 1 is true, and not an artifact of repeated testing. Thus, Experiment 2 suggests that order of presentation of stimulus female's strain does not affect social recognition ability in OxtrFB/FB males. These results do underscore the importance of limiting testing history for animals used in social recognition tests, as previous testing can impact willingness to investigate.

The current study expands upon earlier findings that Oxt and the Oxtr are important in social recognition. Our results indicate that Oxt and the Oxtr are less important for ‘broad’ social recognition (interstrain), but are seemingly more important for ‘fine’ social recognition (intrastrain). These results were obtained utilizing a modified social discrimination paradigm that generates very similar investigation times by WT males of stimulus females across all lines tested.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Acknowledgments

Acknowledgments

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References
  7. Acknowledgments

We thank the NEI/NIMH Animal Facilities staff in Building 49 for their exceptional animal care and daily assistance, Particularly Ana Carderas and Darwin Romero. Drs Scott Wersinger, Heather Caldwell and Jerome Pagani provided excellent comments on early versions of this manuscript. We also thank Jim Heath and Emily Shepard for their technical assistance. This research was supported by NIMH (Z01-MH-002498-20) from the Division of Intramural Research, National Institute of Mental Health, National Institutes of Health, DHHS.