SEARCH

SEARCH BY CITATION

Keywords:

  • Autism;
  • mirrored chamber;
  • partition test;
  • social anxiety;
  • social interaction;
  • tube test

Abstract

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

The loss of fragile X mental retardation (FMR1) gene function causes fragile X syndrome (FXS), a common mental retardation syndrome. Anxiety and abnormal social behaviors are prominent features of FXS in humans. To better understand the role of FMR1 in these behaviors, we analyzed anxiety-related and social behaviors in Fmr1 knockout (KO) mice. In the mirrored chamber test, Fmr1 KO mice showed greater aversion to the central mirrored chamber than wild-type (WT) littermates, suggesting increased anxiety-like responses to reflected images of mice. Fmr1 KO mice exhibited abnormal social interactions in a tube test of social dominance, winning fewer matches than WT littermates. In a partition test, Fmr1 KO mice had normal levels of social interest and social recognition. However, during direct interaction tests, Fmr1 KO mice showed significant increases in sniffing behaviors. We further tested the influence of environmental familiarity on the social responses of Fmr1 KO mice to unfamiliar partners. In unfamiliar partitioned cages, Fmr1 KO mice did not differ from WT mice in investigation of unfamiliar partners. However, in familiar partitioned cages, Fmr1 KO mice showed less investigation of a newly introduced partner during the first 5 min and more investigation during the last 5 min of a 20-min partition test, behaviors consistent with initial social anxiety followed by enhanced social investigation. Our findings indicate that the loss of Fmr1 gene function results in altered anxiety and social behavior in mice and demonstrate that the Fmr1 KO mouse is a relevant animal model for the abnormal social responses seen in FXS.

Fragile X syndrome (FXS) is the most common inherited cause of human mental retardation with an estimated prevalence of one in 4000 males (Turner et al. 1996). Physical features of the disorder include mild facial anomalies (long narrow face, prominent jaw and protruding ears), hyperextensible joints, flat feet and, in postpubescent males, macroorchidism. Behavioral features of the disorder include cognitive impairment, hyperactivity, attention deficits, sensory hypersensitivity and hyperarousal, social isolation and anxiety and autistic-like behaviors such as gaze avoidance, perseverative language and hand stereotypies (Hagerman 2002).

Fragile X syndrome results from the loss of expression of the fragile X mental retardation (FMR1) gene located on the X chromosome. In most cases, expansion of a CGG repeat region in the 5′ untranslated region of the FMR1 gene to over 230 copies leads to hypermethylation and subsequent transcriptional inactivation (O'Donnell & Warren 2002). FMR1 mRNA is expressed at high levels during early development and in the brain and testis in adults (Devys et al. 1993; Hinds et al. 1993). In brain, the protein product of FMR1 (FMRP) is highly expressed in the cytoplasm and dendrites of neurons (Feng et al. 1997). FMRP is an RNA-binding protein that is thought to regulate the intracellular transportation and translation of mRNAs important for synaptic function (Jin & Warren 2000). Several RNA targets related to synaptic function have been recently identified (Brown et al. 2001; Chen et al. 2003; Darnell et al. 2001; Miyashiro et al. 2003), supporting the hypothesis that FMRP plays an important role in synaptic plasticity.

Fmr1 knockout (KO) mice, which have undetectable levels of Fmr1 mRNA and FMRP (Bakker et al. 1994; but see Yan et al. 2004), exhibit several of the physical and behavioral characteristics of the human syndrome. In addition to having macroorchidism, these mice show hyperactivity and have a mild spatial learning impairment in the Morris water maze (Bakker et al. 1994). Both macroorchidism and hyperactivity were over-corrected in FMR1 yeast artificial chromosome (YAC) transgenic mice carrying the human FMR1 gene, indicating a role for FMR1 in these processes (Peier et al. 2000). Contrary to expectations based on the human phenotype, Fmr1 KO mice showed decreased anxiety-like responses in open-field and light/dark tests. Nonetheless, this phenotype was over-corrected in the FMR1 YAC transgenic mice, indicating a role for the Fmr1 gene in anxiety-related behaviors (Peier et al. 2000). The unexpected nature of the results, however, suggests the need to assess anxiety-like behavior in additional rodent tests for anxiety. In addition, it is important to note that the anxiety responses observed in FXS are predominant in social settings.

Abnormal social behaviors, especially shyness and social withdrawal, can be a prominent feature of FXS. Social anxiety is often observed in high-functioning males with only mild or moderate deficits in FMRP and is frequently the presenting problem in females with FXS (Hagerman 2002). A negative correlation between FMRP production and severity of behavioral problems such as social withdrawal or anxious/depressed behavior in girls with FXS was recently reported (Hessl et al. 2001). The association between FMRP levels and social behavior suggests that the FMR1 gene plays a role in mediating some aspects of social behavior.

Fmr1 KO and transgenic mouse models of FXS (Bakker et al. 1994; Peier et al. 2000) would be useful to determine more specifically whether the FMR1 gene plays a role in social behavior. However, despite extensive characterization of behavior in these mice, very little has been reported about their social behavior (Mineur et al. 2002). Therefore, we designed a series of experiments to further characterize the anxiety-related responses in male Fmr1 KO mice and to begin to determine whether Fmr1 KO mice display abnormal social interactions. Our findings indicate that Fmr1 KO mice display (a) a moderate increase in anxiety-related response in a mirrored chamber test and (b) abnormal social behavior that is influenced by experience.

Materials and methods

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

Animals

Behavioral testing was performed on two different colonies of male Fmr1 KO mice and their wild-type (WT) littermates. The initial tube test (Fig. 2) was performed on animals bred at Baylor College of Medicine from stock provided by S. Warren, which originated from 129/OLA ES cells from B. Oostra. These mice had been backcrossed to C57BL/6 for five generations (Peier et al. 2000). That colony of Fmr1 mice (Colony 1) was destroyed by a flood in June 2001. With the exception of the initial tube test, all other experiments were performed with mice from a second colony (Colony 2) of Fmr1 KO mice generously provided by The Jackson Laboratory (Bar Harbor, ME). Colony 2 mice were backcrossed to C57BL/6J for 11 generations. All mice for the current study were generated by mating Fmr1 heterozygous female mice with Fmr1 WT male mice. Only male Fmr1 KO (Fmr1–/y) and WT littermates were tested in this study. Additional male C57BL/6J mice from The Jackson Laboratory were used for the partition test. Mice were housed (two to five per cage) in a room with a 12-h light/dark cycle (lights on at 0600 h, off at 1800 h) with access to food and water ad libitum. In general, behavioral testing was performed between 0900 h and 1500 h. At the start of testing, the mice were 3–4 months of age. An experimenter who was blind to the genotypes of the mice conducted the experiments. All behavioral testing procedures were approved by the Baylor College of Medicine Animal Care and Use Committee and followed the National Institutes of Health Guidelines.

image

Figure 2. Tube test of social dominance. The percentage of wins by Fmr1 knockout (KO) is presented in black bars and wild-type (WT) littermates in white bars. The data represent results from matches between unfamiliar KO and WT mice (i.e. non-cagemates) from Colony 1 and Colony 2. The asterisks denote results significantly different from a chance 50:50 outcome (P ≤ 0.05).

Download figure to PowerPoint

Open-field test

We tested 15 Fmr1 KO and 15 WT mice from our new colony to confirm basic findings from an earlier study in which Fmr1 KO mice exhibited increased exploratory activity and decreased anxiety-like responses in the open field (Peier et al. 2000). The open-field activity data were analyzed by one-way analysis of variance (anova). Because changes in one direction only were expected based upon previous results, one-tailed P values were calculated.

Mirrored chamber

The mirrored chamber was similar to the apparatus described by Toubas et al. (1990). It consisted of an open cube (30 × 30 × 30 cm), constructed with five-mirrored interior walls and black Plexiglas exterior walls, placed on one side in the center of a 45 × 45 × 32 cm high black Plexiglas box such that the opening faced one of the walls of the box. The black exterior walls of the mirrored chamber in conjunction with the walls of the black box formed three 7.5-cm corridors around the exterior of the mirrored chamber. An additional panel of mirror glass was placed on the interior of the black box opposite the opening of the mirrored chamber, forming an open mirrored alleyway.

Naïve, group-housed mice (15 KO and 15 WT) were placed into a dark corner of the chamber and allowed to explore for 10 min. Using a hand-held computer (Psion Workabout mx, Psion Teklogix, Erlanger, KY) with the observer® program (Noldus Information Technologies, Leesburg, VA), mice were scored for the number and latency of entries and time spent in each of three zones: dark alleys, the mirrored chamber and the open mirrored alley directly in front of the mirrored chamber. An entry was defined as the mouse placing all four feet into the zone. A center mirror ratio (i.e. the time spent in the mirrored chamber divided by the time spent in both mirrored areas – mirrored chamber and mirrored alleyway) was calculated to increase the sensitivity of the test by eliminating the time spent in the dark areas of the test apparatus and to reduce potential confounds because of differences in basal exploration. Data were analyzed by one-way anova.

Tube test of social dominance

As an initial screen for normal social behavior, mice were tested in a tube test for social dominance (Lijam et al. 1997; Lindzey et al. 1961; Messeri et al. 1975; Shahbazian et al. 2002). In this test, one Fmr1 KO mouse and one WT mouse were placed head first at opposite ends of a white plastic tube (3.7 cm inner diameter, 30.5 cm in length) and released simultaneously. The match ended when one mouse completely retreated from the tube. The mouse remaining in the tube was designated the winner (score = 1), and the retreating mouse was the loser (score = 0). Matches lasting more than 2 min or in which the mice crossed over each other were not scored. The scores of each genotype were not independent in matches of Fmr1 KO vs. Fmr1 WT mice; hence, a χ2 one-sample analysis was used to determine whether the scores of the mutant mice were significantly different from an outcome expected by chance (i.e. a 50:50 win–lose outcome).

Colony 1 (KO vs. WT)

Nine Fmr1 KO and nine WT mice from Colony 1 were tested against three different mice of the opposite genotype (total 27 matches). For this part of the test, only mice from different homecages were used. An additional nine matches were performed between KO and WT mice from the same homecage.

Colony 2 (KO vs. WT)

We sought to replicate results obtained with Colony 1 in our new colony and also extend the study to examine the effect of increased test familiarity on responses in the tube test. Fifteen Fmr1 KO mice and 15 WT littermates from Colony 2 were matched in tube tests organized in a round-robin scheme 1 week following testing in the open field. Each mouse performed five matches against mice of the opposite genotype each day over 3 days for a total of 225 matches. Matches between mice from the same homecage (familiar) were interspersed randomly among matches between mice from different homecages (unfamiliar), but results from these two types of matches were analyzed separately. For the purpose of replicating the Colony 1 tube test result, matches between unfamiliar mice on day 1 of testing were analyzed. For the purpose of examining the issue of test familiarity, results from matches between unfamiliar mice on day 3 of testing were also analyzed. For matches between familiar cagemate mice, matches from all 3 days of testing were included in the analysis. This experiment was a replication and extension of experiment 1; hence, the appropriate one-tailed P values were calculated.

Partition test

A partition test was used to evaluate social behaviors in mice during a non-contact version of a social interaction test. The method used was adapted from Kudryavtseva (1994) to include an assessment for basic social recognition in addition to assessment of social interest.

Experimental animals were individually housed for 3–4 days by placing a mouse into one side of standard cage divided in half by a clear perforated (0.6 cm-diameter holes) partition. At approximately 1400 h, the day before the test, a partner C57BL/6J mouse (gender-, age- and weight-matched) was placed in the side opposite the experimental mouse.

Test validation with C57BL/6 mice

In the validation experiment, 30 C57BL/6J male mice were individually housed for 3 days before being placed into partitioned cages opposite the age- and weight-matched group-housed male C57BL/6J partner mice. On the day of the test, approaches and time spent at the partition by the experimental mice were measured (Psion computer with the Noldus observer® program) in a series of tests in which the familiar overnight partner and several new unfamiliar partners were presented in different orders. The experimental mice were split into three groups according to partner order (10 mice per group): group 1 – overnight familiar partner, unfamiliar partner 1, familiar partner, unfamiliar partner 2 and unfamiliar partner 3; group 2 – overnight familiar partner, unfamiliar partner 1, unfamiliar partner 2, familiar partner and unfamiliar partner 3; group 3 – overnight familiar partner, unfamiliar partner 1, unfamiliar partner 2, unfamiliar partner 3 and familiar partner. There were no delays between the various partner tests. Each partner test lasted for 5 min resulting in a total 25-min session for each experimental mouse.

Partition test in Fmr1 KO and WT mice

A total of 30 Fmr1 KO and 30 WT mice were housed individually for 4 days before the partition test. On the day before the test, C57BL/6J partner mice were placed on the opposite side of the partition. For this partition test, there were three phases: First, approaches and time spent at the partition by the experimental mice were measured with the original partner. Second, the behavior of the experimental animal was scored with a new partner. Third, the behavior of the experimental animal was scored with the original partner.

Data were analyzed by two-way (genotype X partner) anova with repeated measures. Paired t-tests were used for post hoc comparisons.

Social interaction test

KO vs. WT interactions

Four days after the standard partition test, Fmr1 KO and WT mice were rehoused into new partitioned cages such that an Fmr1 KO mouse occupied one compartment and a weight-matched WT mouse (non-littermate) occupied the other compartment. Forty-eight hours later a 10-min direct social interaction test was conducted. A total of 18 KO–WT pairs were tested.

To conduct the direct social interaction test, the following procedures were used. Before testing, the standard filtered top of the cage was replaced with a clear lid. After a 5-min acclimation period, the partition was removed and social interaction between the mice was videotaped for 10 min. Behavioral responses were scored from the videotapes by an observer blind to the genotypes of the mice by using the Psion hand-held computer with the Noldus observer® program. Scored behaviors were divided into three groups. Active social behavior, which is initiated by the experimental mouse toward the partner, includes: (1) anogenital sniffing; (2) sniffing of any part of partner's body excluding anogenital area; (3) direct aggressive attacks accompanied by bites toward the back of the partner and kicking movements of the rear legs; (4) lateral threats; (5) tail rattling; (6) chasing; (7) aggressive grooming – intensive nibbling/licking of partner's scruff area and (8) wrestling/boxing – fighting between mice when they are equal and it is impossible to tell the direction of the behavior. Passive social behavior, which occurs as a reaction to the active/aggressive behavior of the partner mouse toward an experimental mouse, includes: (9) receptive – when the experimental mouse receives the sniffing from the partner but does not show signs of defensive or submissive behavior; (10) escape – reaction to partner's aggressive act, which might be expressed from leap to full flight; (11) freezing – immobilization as a reaction to aggressive action; (12) defeat postures (passive defense) – reaction to frequent or strong attacks, a posture with limp forelimbs and upward angled head; (13) upright defense (active defense) – experimental mouse rears up on its hindlegs, rotating the upper torso toward the aggressive opponent, sometimes with kicking movements of forelimbs. Non-social behaviors scored include: (14) exploration – walking around and sniffing the bedding and walls of the cage; (15) rearing in any part of the cage; (16) rest – sitting, sleeping or eating; (17) digging up and scattering the bedding and (18) self-groom – body-care activities. In addition, the amount of time spent in the mouse's own territory and in the partner's territory during the 10-min test was measured. Data were analyzed by one-way anova both within the broad categories and for each parameter.

KO vs. KO and WT vs. WT interactions

Because the sniffing/anogenital sniffing and receptive parameters were related to each other in the KO vs. WT experimental set up, an independent experiment was performed to assess direct social interactions in pairs of mice of the same genotype. For this experiment, 18 KO (i.e. nine KO : KO pairs) and 14 WT (i.e. seven WT : WT pairs) mice were individually housed for 3 days, followed by 2 days of paired housing on opposite sides of a partition. The social interaction test was performed as described above.

Cage familiarity experiment

Naïve, group-housed Fmr1 mice (17 KO and 18 WT) were housed individually in standard cages for 4 days. On day 5, the experimental mice were placed into a partitioned cage (novel environment) opposite an unfamiliar male C57BL/6J mouse. Thus, in contrast to the standard partition test described above, the experimental and partner mice did not have any time in the partition cage prior to the first test. The number of approaches and time spent at the partition were then recorded for 20 min. Following the test, the partner C57BL/6J mouse was removed and the Fmr1 KO and WT mice continued to be housed individually in the partitioned cages in order to familiarize them with the cage. After 24 h, a new unfamiliar male C57BL/6J mouse was introduced into the side opposite the Fmr1 KO or WT mouse, and the number of approaches and time spent at the partition were recorded for 20 min. Data were analyzed by repeated measures anova with follow-up comparisons at each interval by t-tests.

Statistical analysis

All data were analyzed by using the spss for Mac® statistical software package (SPSS, Chicago, IL). Unless noted otherwise, two-tailed P values were determined. For all comparisons, the level of significance was set at P ≤ 0.05. Non-significant trends are reported for P ≤ 0.1.

Results

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

Anxiety-related responses in Fmr1 KO mice

Open-field test in Fmr1 mice from Colony 2

Results (not shown) obtained with male Fmr1 mice from Colony 2 agree with our earlier findings that Fmr1 KO mice exhibit increased exploratory activity and decreased anxiety-like responses in the open field (Peier et al. 2000). Fmr1 KO mice from Colony 2 were significantly more active than WT littermates in total distance traveled (F1,28 = 9.829, P = 0.001) and time spent moving (F1,28 = 10.852, P = 0.002). There was no significant difference in average speed of movement (F1,28 = 1.990, P = 0.08) or vertical activity (F1,28 = 0.760, P = 0.2). Fmr1 KO mice explored the center of the open field, as a proportion of total distance traveled, significantly more than WT littermates (F1,28 = 4.096, P = 0.03), indicating less anxiety-like behavior. Fmr1 KO mice produced fewer boli than WT mice (F1,28 = 4.139, P = 0.03), providing another indication of less anxiety-like behavior.

Mirrored chamber test

As shown in Fig. 1, Fmr1 KO mice had a lower center mirror ratio, spending significantly less time in the mirrored chamber as a proportion of time in both mirrored areas (F1,28 = 4.015, P = 0.05), indicating that in this assay Fmr1 KO mice displayed more anxiety-related responses. There was no difference between Fmr1 KO and WT mice in the latency to enter the mirrored chamber (F1,28 = 1.765, P = 0.2), number of entries to the mirrored chamber (F1,28 = 0.000, P = 1.0) or time spent in the mirrored chamber (F1,28 = 0.234, P = 0.6) (data not shown).

image

Figure 1. Anxiety-related response in the mirrored chamber test. The center mirror ratio is the time spent in the mirrored chamber divided by time spent in both mirrored areas (mirrored chamber and mirrored alley). Values represent mean ± SEM. Black bar indicates Fmr1 knockout mice and white bar indicates respective wild-type (WT) littermates. Asterisk denotes significant difference from WT littermates (P ≤ 0.05).

Download figure to PowerPoint

Tests of social behavior in Fmr1 mice

Tube test of social dominance (KO vs. WT)

Fmr1 KO mice from Colony 1 won significantly fewer matches (7/26, 27%) against WT non-cagemates than expected by chance (χ2 = 5.538, P = 0.02) (Fig. 2, left); however, when matched against WT cagemates, Fmr1 KO mice won five of nine matches (χ2 = 0.111, P = 0.7) (data not shown).

The tube test experiment was repeated with Fmr1 mice from Colony 2 and extended to examine the effect of test familiarity. On the first day of testing, Fmr1 KO mice won significantly fewer matches (25/64, 39%) against unfamiliar WT mice (χ2 = 3.063, P = 0.04 against chance outcome). On the third day of testing, Fmr1 KO mice won 45% (28/62) of matches against unfamiliar WT mice (χ2 = 0.581, P = 0.2). Against familiar WT cagemates, Fmr1 KO mice won 10 of 20 matches (χ2 = 0.000, P = 1.000).

Partition test validation experiment

Figure 3 shows the time spent at the partition, the main parameter reflecting social interest in the partner, for each test-order group in response to the partners presented in different orders. There was a main within-subject effect in time spent at the partition (F4,24 = 9.178, P < 0.001). There was also a repeated testing × partner-order interaction (F8,48 = 2.787, P = 0.01), indicating that mice can distinguish the familiar partner presented in different orders. Post hoc analysis revealed that mice show significantly increased interest in new unfamiliar partners presented during the second 5-min interval (group 1, P = 0.02; group 2, P = 0.04 and group 3, P = 0.001) and third 5-min interval (group 2, P = 0.005; group 3, P = 0.05) and partially during the fourth session (group 1, P = 0.07; group 3, P > 0.05), whereas they drop their interest to the baseline level when faced with the returning familiar partner (P's > 0.05). The reaction of the mice to familiar and unfamiliar partners during the fifth 5-min interval did not differ from the initial level (P's > 0.05), suggesting habituation to the experimental conditions. The results show that C57BL/6J mice are able to recognize at least three different mice.

image

Figure 3. Validation of social recognition in partition test. Time spent at the partition is shown in seconds for groups of C57BL/6J mice presented with their familiar overnight partner (white bars) and three different unfamiliar partners (black, striped and stippled bars) during five sequential 5-min tests. The familiar partner was presented in the third (test order 1), fourth (test order 2) or fifth session (test order 3) to test recognition. Values represent mean ± SEM. Asterisks denote significant differences from the first baseline familiar test session as determined by paired t-tests *(P ≤ 0.05, **P < 0.01). P values are shown for results in which P > 0.05 but P ≤ 0.1.

Download figure to PowerPoint

Partition test in Fmr1 Mice

Figure 4 shows the time spent at the partition during each 5-min test session for both Fmr1 KO mice and WT littermates. Overall, there was a main within-subjects effect of test session in the amount of time spent at the partition (F2,116 = 165.897, P < 0.001), a trend toward a significant test session × genotype interaction (F2,116 = 2.709, P = 0.07) and no effect of genotype (F1,58 = 1.059, P = 0.3). Post hoc analysis revealed that both Fmr1 KO and WT mice showed increased interest in the new unfamiliar mouse relative to the familiar partner (P's ≤ 0.05) and recognition of the returning familiar partner (evidenced by time spent at the partition back to near baseline levels, P's > 0.05). There were no significant differences between WT and KO mice in time spent at the partition during the first, second or third test sessions (P's > 0.05). These results indicate that Fmr1 KO mice showed normal levels of social interest and social recognition in the partition test.

image

Figure 4. Partition test with Fmr1 mice. Time spent at the partition is shown in seconds for wild-type (WT) or Fmr1 knockout (KO) mice during three sequential 5-min tests with their familiar overnight partner (white bars) and an unfamiliar partner (gray bars). Values represent mean ± SEM.

Download figure to PowerPoint

Social interaction test (KO vs. WT)

In direct social interactions with WT mice, Fmr1 KO mice showed enhanced active social behavior (F1,34 = 14.983, P < 0.001) and decreased passive social behavior (F1,34 = 10.083, P = 0.003). As seen in Fig. 5(a), the enhanced active social behavior observed in Fmr1 KO mice was primarily because of significantly higher sniffing duration (F1,34 = 21.739, P < 0.001) and anogenital sniffing duration (F1,34 = 13.704, P = 0.001) compared with WT mice. Fmr1 KO mice also exhibited higher frequency of sniffing (F1,34 = 19.942, P < 0.001) and of anogenital sniffing (F1,34 = 21.983, P < 0.001) than WT mice (data not shown). Among parameters within the passive social category (Fig. 5b), the main difference between Fmr1 KO and WT mice was in ‘receptive’ behavior, when the mice simply allow themselves to be sniffed without signs of submissive or defensive behavior (F1,34 = 19.060, P < 0.001). There were no differences in non-social behaviors or time spent in the home or partner's compartment (P's > 0.05) (data not shown).

image

Figure 5. Fmr1 knockout (KO) vs. wild-type (WT) social interaction test. Percentage of time spent in various (a) active and (b) passive social behaviors by Fmr1 KO (black bars) and WT (white bars) mice during a 10-min interaction test. In this experiment, Fmr1 KO mice were paired with non-cagemate WT littermates. Values represent mean ± SEM. Asterisks denote significant differences between KO and WT mice for a particular behavior (***P ≤ 0.001).

Download figure to PowerPoint

Social interaction test (KO vs. KO and WT vs. WT)

When compared to interactions with the same genotype partner, there was no overall difference between KO and WT mice in any of the active, passive or non-social broad categories of behavior (P's > 0.05). However, analysis of the various individual parameters revealed significant increases in sniffing and anogenital sniffing duration (F1,30 = 10.809, P = 0.003; F1,30 = 6.909, P = 0.01) and sniffing and anogenital sniffing frequency (F1,30 = 10.967, P = 0.002; F1,30 = 6.184, P = 0.02) in Fmr1 KO compared with WT mice (Fig. 6a). In contrast to the data from the KO vs. WT interactions, the duration and frequency of receptive behavior were significantly higher in Fmr1 KO mice compared with WT (duration: F1,30 = 7.899, P = 0.009; frequency: F1,30 = 11.775, P = 0.002) (Fig. 6b). As in the previous experiment there were no differences in non-social behaviors between the two genotypes (data not shown); however, Fmr1 KO mice spent more time than WT animals in their home compartment (F1,30 = 4.358, P = 0.05) (data not shown).

image

Figure 6. Fmr1 knockout (KO) vs. KO and wild-type (WT) vs. WT social interaction tests. Percentage of time spent in various (a) active and (b) passive social behaviors by Fmr1 KO (black bars) and WT (white bars) mice during a 10-min interaction test. In this experiment, Fmr1 KO mice were paired with non-cagemate Fmr1 KO littermates, and WT mice were paired with non-cagemate WT littermates. Values represent mean ± SEM. Asterisks denote significant differences between KO and WT mice for a particular behavior (**P ≤ 0.01).

Download figure to PowerPoint

Influence of cage familiarity on response of Fmr1 mice to unfamiliar social partners

Results from experiments described thus far suggested that the social responses of Fmr1 KO mice might be sensitive to environmental and/or partner familiarity (Table 1). To address this possibility, we tested the social responses of Fmr1 mice to unfamiliar partners in both novel and familiar cage environments. Figure 7 shows the time spent at the partition by Fmr1 KO and WT mice during 5-min intervals in both the unfamiliar and familiar cage situations. On the first day, during the unfamiliar cage test, there was a main effect of time interval for time spent at the partition (F3,99 = 43.234, P < 0.001) but no interval × genotype interaction (F3,99 = 0.874, P = 0.5) and no between-subjects effect of genotype (F1,33 = 0.249, P = 0.6). There was also no difference between WT and Fmr1 KO mice in latency to first approach the partition (F1,33 = 1.611, P = 0.2) (data not shown).

Table 1.  Summary of Fmr1 social responses
TestEnvironmentOpponent/partnerResult
  1. KO, knockout; WT, wild type.

Tube test
 Colony 1 – day 1NovelNovelKO had fewer wins
 Colony 1 – cagematesFamiliarfamiliarKO = WT
 Colony 2 – day 1NovelNovelKO had fewer wins
 Colony 2 – day 3FamiliarNovelKO = WT
 Colony 2 – cagematesNovel or familiarFamiliarKO = WT
Partition test (5 min each)
 Baseline testFamiliarFamiliarKO = WT
 Social interest – unfamiliarFamiliarNovelKO = WT
 Social recognitionFamiliarFamiliar-returnKO = WT
Social interaction (10 min)
 KO vs. WTFamiliarFamiliarKO sniffed partner more
 KO vs. KO, WT vs. WTFamiliarFamiliarKO sniffed partner more
Cage familiarity experimentNovelNovelKO = WT
 FamiliarNovelKO < WT first 5 min KO = WT second, third 5 min KO > WT last 5 min
image

Figure 7. Cage familiarity experiment. Time spent at the partition is shown in seconds during 5-min intervals for 20-min tests in which Fmr1 knockout (KO) mice (closed circles) or their respective wild-type (WT) littermates (open circles) were presented with unfamiliar C57 partners. In the first test, KO and WT mice were tested in unfamiliar partitioned cages. The partners were removed after the test and the KO and WT mice were left in the cages. After 24 h, KO and WT mice were presented with new unfamiliar partners in their now-familiar partitioned cages. Values represent mean ± SEM. Asterisks denote significant differences between KO and WT mice for a particular time interval (*P ≤ 0.05).

Download figure to PowerPoint

On the second day, in the now-familiar cage environment, there was a main effect of time interval (F3,93 = 7.075, P < 0.001) and a significant interval × genotype interaction for time spent at the partition (F3,93 = 4.964, P = 0.003) (Fig. 7 right). There was a non-significant trend toward an overall effect of genotype in time spent at the partition (F1,31 = 2.182, P = 0.1). Post hoc analysis of the data at each interval revealed that Fmr1 KO mice spent significantly less time at the partition during the first 5 min of the test (P = 0.02) and significantly more time at the partition during the last 5 min (P = 0.04). Fmr1 KO mice also showed a tendency to take longer time to first approach the partition (31.1 ± 9.8 seconds vs. 13.0 ± 2.2 seconds; F1,31 = 3.419, P = 0.07). Thus, there were significant differences between Fmr1 KO and WT mice in their investigation of an unfamiliar social partner depending upon degree of familiarity with environmental surroundings.

Discussion

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

Anxiety is a prominent behavioral characteristic of FXS. In particular, many people with FXS exhibit shyness or social anxiety and are easily upset by novel people or environments (Reiss & Freund 1992). Previous studies have found, contrary to expectations, that Fmr1 KO mice showed decreased anxiety-like behavior both in the open-field and in the light/dark tests (Bakker et al. 1994; Peier et al. 2000). A study examining responses in the elevated plus maze found no difference between Fmr1 KO and WT mice (Mineur et al. 2002). However, the over-expression of FMRP in FMR1 YAC transgenic mice resulted in a phenotype of increased anxiety in the open-field and light–dark tests, indicating that FMRP is involved in modulating anxiety-provoking responses (Peier et al. 2000).

Increased anxiety-related responses in the mirrored chamber

Our current findings are the first to show an increase in anxiety-related responses in Fmr1 KO mice. Individuals with FXS appear to have a particular kind of anxiety involving social and novel situations or environments; therefore, the type of test chosen to evaluate anxiety-related responses in FXS mouse models is likely to be an important factor. Both the open-field and light–dark tests are tests for exploratory-based anxiety that use retreat from brightly lit, open spaces as an indicator of anxiety in rodents. Given that the Fmr1 KO mice are hyperactive, increased locomotor activity could confound assessment of anxiety in these tests. In the open-field test, the center distance ratio is a useful measure of anxiety that corrects for differences in locomotor activity, making it possible to distinguish locomotor activity from anxiety-related responses (Peier et al. 2000).

The Fmr1 KO mice displayed an increased anxiety-related response in the mirrored chamber. The mirrored chamber test is also an exploration-based test of anxiety, but it has a potential social component in that mirrors are used to provoke approach-avoidance conflict behavior in the mice (Toubas et al. 1990). Because most animals respond to their mirror image as if it were another animal (Gallup 1968), the appearance of multiple animals in the chamber of mirrors is thought to produce an artificial social stimulus. The central mirrored chamber is very aversive to naïve, untreated mice, as evidenced by the low percentage of time spent in the chamber and the latency to enter the chamber (Reddy & Kulkarni 1997; Seale et al. 1996; Toubas et al. 1990). Treatment with diazepam (Toubas et al. 1990) or other anxiolytic drugs (Reddy & Kulkarni 1997; Seale et al. 1996) reduces the latency to enter the mirrored chamber, providing pharmacological validation for the test as appropriate for assessment of anxiety-related behavior. The mirrored chamber test is potentially confounded by two factors: the low sensitivity of the assay because of the highly aversive nature of the mirrored chamber and differences in locomotor activity (Kliethermes et al. 2003). The use of the center mirror ratio helps both to increase the sensitivity of the assay for detecting an increase in anxiety-like behavior and to normalize the data for differences in locomotor activity.

It is important to point out that the current study is the first to show that a phenotypic difference in the mirrored chamber is also present in animals that have abnormal social behaviors in other tests of social interaction. As discussed below, the results of the tests of social behavior in Fmr1 KO mice are consistent with the hypothesis that Fmr1 KO mice have increased social anxiety. Therefore, we suggest that the mirrored chamber test may be appropriate in assessing social anxiety in mice, although additional studies will be important to validate the use of this test for that purpose. Currently, our laboratory is studying the responses of Fmr1 KO mice in non-exploratory tests of anxiety, such as stress-induced hyperthermia (Borsini et al. 1989; Bouwknecht & Paylor 2002; Van der Heyden et al. 1997) or light-enhanced startle (Hironaka et al. 2002; Walker & Davis 1997), which will be important to better understand the relationship between exploration and anxiety-related responses in these mice.

Abnormal social interactions in Fmr1 KO mice

The social responses of Fmr1 KO mice are complex and may even appear to be conflicting. In the partition test, which is a good test to evaluate social responses in a ‘non-contact’ situation, Fmr1 KO and WT mice displayed comparable levels of interest with familiar and unfamiliar partners as well as recognition of their familiar partner. Importantly, this test indicated that there are conditions under which Fmr1 KO mice exhibit normal social behavior. We find it interesting that the KO mice had a normal response to unfamiliar mice in this 5-min test, but not during the first 5 min of the familiar cage experiment. The major difference in the set up of these experiments is that the mice were housed with overnight partners in the first partition test, but alone in the cage familiarity experiment. Thus, the experience of becoming familiar with the overnight partner may have attenuated the KO response to the next unfamiliar partner.

In the tube test, Fmr1 KO mice retreated more frequently than WT mice. The tube test of social dominance was originally developed to test dominance hierarchies in mice (Lindzey et al. 1961). Dominance in the tube test correlates with whisker trimming (Strozik & Festing 1981), a behavioral trait closely associated with social dominance (Long 1972; Sarna et al. 2000). While mice typically use aggression to establish dominance, it is unclear whether tube test performance is directly related to aggression. A recent report by Rodriguiz et al. (2004) looked at the interactions between cagemates inside a clear tube during a tube test and found that the mice were typically engaged not only in pushing behavior but also in more aggressive agonistic behaviors such as clawing, tail rattling, biting and chasing. Although we were unable to directly observe the interactions in our tube tests, we did not observe tail rattling or chasing by either Fmr1 KO or WT mice. Thus, we suggest that the tendency of Fmr1 KO mice to retreat from interactions involving unfamiliar opponents in the tube test is not likely because of differences in aggressive behavior but is more consistent with a possible increase in social anxiety. An important observation for the tube test indicates that although Fmr1 KO mice are in general more likely to back out of the tube, this effect is determined by the amount of experience with the partner and experience with the test. During tube tests with cagemate opponents (data not shown), Fmr1 KO mice did not show an increased tendency to retreat first, suggesting that they respond differently to familiar vs. unfamiliar opponents. In addition, the tendency of Fmr1 KO mice to retreat from the tube was attenuated when tested on subsequent days. These findings suggest that social responses of Fmr1 KO mice are influenced by factors such as test experience and familiarity with the test partner.

The experiment to directly assess the effect of cage familiarity showed that the social responses of Fmr1 KO mice to unfamiliar partners were very sensitive to environmental familiarity. Whereas social investigation of unfamiliar partners was similar for both Fmr1 KO and WT mice in an unfamiliar partitioned cage, Fmr1 KO mice reacted differently than WT mice to a new unfamiliar partner after familiarization with the cage environment. The findings from this ‘cage familiarity experiment’ are important for several reasons. First, they clearly indicate that social behaviors of Fmr1 KO mice are dependent on experience. The findings from the tube test also support this contention. Second, the data from the first 5 min of the test when the experimental mice were faced with an unfamiliar partner in a familiar cage indicate that the Fmr1 KO mice were initially less likely to approach the new mouse, a finding that we suggest may reflect social anxiety. Lastly, the Fmr1 KO mice adapt to this social anxiety and actually display more social responses/interest compared with WT mice, which is consistent with the direct social interaction data. The fact that social responses of Fmr1 KO mice are differentially influenced by experience compared to WT mice may turn out to be a fundamental observation in other behavioral test settings. Preliminary data from other studies in our lab indicate that the impact of drug treatment on hyperactivity in the open field is dependent on test experience. Thus, future investigations of Fmr1 KO mice will need to incorporate experimental designs to clearly assess the influence of test experience on the behavior of Fmr1 KO mice.

Social interaction in FXS and Fmr1 mice

Abnormal social behavior is a significant problem for individuals with FXS and their families. In addition to shyness or social anxiety, some of the behavioral traits contributing to this impairment include hyperarousal, tactile defensiveness, gaze aversion and repetitive and perseverative behaviors such as hand-flapping or hand-biting (Reiss & Freund 1992). Many of these characteristics are similar to those found in individuals with autism. Indeed, it has been estimated that 15–25% of people with FXS meet the diagnostic criteria for autism (Dykens & Volkmar 1997). However, the behavior of people with FXS is somewhat different from people with classic autism (Reiss & Freund 1992). Most FXS males show more social interest and are more socially attuned to others than males with autism (Bailey et al. 2000). They are able to recognize unfamiliar faces and understand emotional facial expressions (Simon & Finucane 1996; Turk & Cornish 1998). However, despite their interest in interacting with people, direct gaze is highly aversive to individuals with FXS (Cohen et al. 1988; Cohen et al. 1991). Many males with FXS show a unique pattern of gaze aversion, illustrated by the so-called ‘fragile X handshake’ in which males with FXS readily take the hand of others, but turn their head as they do so, mumble a greeting and then return their gaze after the other person looks away (Wolff et al. 1989). This greeting behavior has been linked to excessive anxiety in social situations.

The findings of this study suggest that the complex social phenotype of the Fmr1 KO mouse has some consistencies with the human fragile X phenotype. We suggest that findings from three of our experiments support the notion that the KO mice display abnormal social responses that may be considered consistent with ‘social anxiety’: (a) decreased center mirror ratio in the mirrored chamber test, (b) earlier retreat in the tube test and (c) decreased investigation of an unfamiliar partner during the first 5 min of the familiar cage experiment. Similar to people with FXS, Fmr1 KO mice did not show defects in social interest and social recognition. Indeed, the KO mice showed more active social behavior with mice familiar to them. Although the social responses of Fmr1 KO mice are complex and influenced by experience with partners and test environment, the present findings demonstrate that the Fmr1 KO mouse is a relevant animal model for the abnormal social responses seen in FXS. These mice will be very useful in future studies to elucidate mechanisms by which loss of FMR1 gene function alters social responses and to test potential therapies to ameliorate these responses. The discovery that the FMR1 gene plays a role in social behavior in mice has important implications not only for fragile X but also for other disorders with impaired social behavior, such as autism.

Fmr1 KO mouse model: molecular considerations

A recent molecular characterization of the Fmr1 KO mouse model indicates that Fmr1 mRNA is expressed in these mutant animals; therefore, they are not a ‘molecular null’ (Yan et al. 2004). Apparently, there are splicing mutations in the Fmr1 KO mice that result in transcripts that are aberrantly spliced. Yan et al. (2004) suggest that there is no reason for proteins that are produced from these transcripts to be inactive; however, no protein data were included. At this point, it is unclear whether and how these abnormal splice variants contribute to the behavioral phenotypes of the Fmr1 KO mice, because it is not known whether there is variable protein expression in different brain regions that correlates with different behavioral responses including the present anxiety and social phenotypes. More research is necessary to fully understand the implications of the findings from Yan et al. (2004).

References

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

Acknowledgments

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

This research was supported by FRAXA Research Foundation, the Baylor Fragile X Research Center (HD24064-S1) and the Administrative and Neurobehavioral Cores of the Baylor Mental Retardation and Developmental Disabilities Research Center (HD24064).