|Home | About | Journals | Submit | Contact Us | Français|
Autism is diagnosed by three major symptom categories: unusual reciprocal social interactions, impaired communication, and repetitive behaviors with restricted interests. Direct social approach in mice has strong face validity to simple social approach behaviors in humans, which are frequently impaired in autism. This unit presents a basic protocol for a standardized, high-throughput social approach test for assaying mouse sociability. Our automated three-chambered social approach task quantifies direct social approach behaviors when a subject mouse is presented with the choice of spending time with either a novel mouse or a novel object. Sociability is defined as the subject mouse spending more time in the chamber containing the novel target mouse than in the chamber containing the inanimate novel object. The Basic Protocol describes procedures for testing one subject at a time in a single apparatus. A Support Protocol addresses data collection.
Abnormal reciprocal social interactions, which include low interest in peers, difficulty in maintaining social interaction, and failure to communicate effectively using nonverbal facial and gestural expressions, are central to the diagnosis of autism spectrum disorders (American Psychiatric Association, 1994; Lord et al., 2006; Lord and Bishop, 2009). Mouse models with strong face validity to the diagnostic symptoms provide essential tools for studying basic genetic and environmental causes of autism, and for developing treatment strategies. Mus musculus is a social species that exhibits a rich repertoire of social behaviors. Social behaviors in mice are important for forming bonds, foraging, defending territories, avoiding predators, and reproducing successfully (D’Amato and Populin, 1987; Laviola and Terranova, 1998; Moles and D’Amato, 2000; Branchi et al., 2001; Hurst et al., 2001; Miczek et al., 2001; Winslow, 2003; Kavaliers et al., 2006; Crawley, 2007a,b; Panksepp et al., 2007; Arakawa et al., 2008). Mice have a natural tendency to approach and investigate unfamiliar conspecifics, much like the way a person would greet a stranger. A person with autism often avoids interacting with a stranger, either withdrawing when approached or remaining aloof. Therefore, low social approach in mice represents an endophenotype with a reasonable analogy to the types of social deficits that characterize autism (Crawley et al., 2007; Silverman et al., 2010b; Yang et al., 2011b). We have developed equipment and methods that provide a relatively rapid screen for impaired sociability in mouse models of autism (Crawley, 2004; Duncan et al., 2004; Nadler et al., 2004; DeLorey et al., 2008; Jamain et al., 2008; Ryan et al., 2008; Nakatani et al., 2009; Page et al., 2009; Radyushkin et al., 2009; Yang et al., 2009,2011a; Ehninger et al., 2010; Silverman et al., 2010b; Defensor et al., 2011; Hamilton et al., 2011; Silverman et al., 2011). The Basic Protocol in this unit presents methods for testing mouse social approach behaviors using our automated three-chambered social approach task (Nadler et al., 2004; Moy et al., 2007, 2008b; Yang et al., 2009, 2011a; Silverman et al., 2010a,b). The Support Protocol addresses data collection.
The automated three-chambered social approach task was developed in 2004 and has been widely employed as a standard test for assaying sociability in mice. Sociability is defined as the subject mouse spending more time in the chamber containing the novel mouse than in the chamber containing the inanimate novel object. A second corroborative measure is time spent sniffing the novel mouse versus time spent sniffing the novel object, which measures direct social interactions. Number of transitions across chambers offers a built-in control measure of exploratory locomotion (Crawley, 2007b, 2008; McFarlane et al., 2008; Moy et al., 2008b, 2009; Yang et al., 2009, 2011a; Silverman et al., 2010a,b). See the Support Protocol for information about collecting these data.
The three-chambered apparatus was originally designed by Nadler, Moy, Dold, Trang, Simmons, Perez, Young, Barbaro, Piven, Magnuson, and Crawley at the University of North Carolina and National Institute of Mental Health (Nadler et al., 2004). The apparatus and software are commercially manufactured and available from Dold Labs and Engineering. Similar but not identical equipment is commercially available from other sources such as Ugo Basile and Stoelting.
Subject mice: e.g., C57BL/6J (B6) and FVB.129P2-Pde6b+ Tyrc-ch/AntJ (FVB/AntJ) adult mice (The Jackson Laboratory) or C57BL/6N and FVB/NJ (e.g., Charles River Laboratories or Taconic Farms) between 8 weeks and 6 months of age (see Critical Parameters)
Target mice (novel mice): e.g., 129/SvImJ adult mice (The Jackson Laboratory) between 8 weeks and 6 months of age, preferably of the same sex and approximate body weight (within 5 g) as the subjects (see Critical Parameters)
Odorless mild dish soap and long-handled dish sponge
70% (v/v) ethanol in labeled spray bottle
Tap water in labeled spray bottle
Test room, with minimal cues visible to the subject
Two (or more) gooseneck desk lamps with incandescent 75-watt light bulbs
Lux meter (Fisher Scientific)
Automated three-chambered social test apparatus: e.g., Crawley automated three-chambered social approach apparatus for mice, hardware and software [(Fig. 8.26.1; Dold Labs and Engineering; (830)560-1471, email@example.com, firstname.lastname@example.org)]; Ugo Basile, cat. no. 46503; Stoelting, cat. no. 60450
Computer (Dell desktop, or similar PC) with software provided by the manufacturer of the automated three-chambered social test apparatus (Dold Labs and Engineering)
Automated video tracking systems, optional (e.g., see www.noldus.com/animal-behavior-research/solutions/research-small-lab-animals/sociability-test; Page et al., 2009)
Heavy-duty utility paper towels
Cameras (preferably CCTV security cameras, e.g., Panasonic WV-CP280) with mounting bracket
DVD recorder (if recording with CCTV cameras) TV monitor (if recording with CCTV cameras) Video cables
BNC and RCA connectors
Standard mouse group-housing cages
Marking pen (dark), and (if preferred) paw tattoos, ear punches, ear tags, or subcutaneous transponders
Paper tube or small cup (for transporting mouse)
Soft facial tissues
Holding area: dedicated room or quiet area near the testing room
Index cards and broad tip markers
Stopwatches without beepers or with beepers silenced (See Fig. 8.26.2)
Plastic drinking cups filled with small heavy objects (to place on top of the inverted wire pencil cups)
NOTE: All protocols using live animals must first be reviewed and approved by an Institutional Animal Care and Use Committee (IACUC) and must follow officially approved procedures for the care and use of laboratory animals.
The following steps (i.e., acclimation and phases I, II, and III) take place on the same day.
This phase is the first 10-min session of the actual experiment. See the Support Protocol for information about data collection in the experiment.
A fourth 10-min session can be added to measure preference for social novelty, although this is not as relevant to autism-like symptoms as sociability. After completion of the 10-min social approach test for sociability, a second novel mouse is placed inside the previously empty novel object cup, while the first novel mouse remains inside its cup. The subject mouse is then given 10 min to explore all three chambers. Preference for social novelty is defined as more time in the chamber with novel mouse 2 than time in the chamber with novel mouse 1, and more time spent sniffing novel mouse 2 than time spent sniffing novel mouse 1. Most mice prefer to spend more time near the completely unfamiliar novel mouse 2.
Preference for social novelty contains components of social recognition and social memory, which appear to be less relevant to the types of social deficits that define autism. Thus, the preference for social novelty test may be useful to investigate social recognition or social memory. For more information about specific methods for quantifying partner recognition and social memory, and as a control for the sensory ability of the subject mouse to discriminate social odors.
Mice rely on olfactory cues to identify other animals. To accurately assess sociability in the three-chambered task, it is critical to clean the apparatus thoroughly between subjects and wash it at the end of the day. This is a point that cannot be overemphasized.
The time spent in each chamber and the number of entries into the side chambers of the three-chambered box manufactured by Dold Labs and Engineering are automatically measured using the computer software provided by the manufacturer. The chamber has pairs of photobeam emitters and detectors embedded across each doorway. As the animal moves through the doorway, it sequentially breaks and unbreaks two infrared beams. An interface box with an embedded real-time controller monitors beam breaks and calculates transitions in and out of each chamber and time spent in each chamber. At the end of phase II and phase III, chamber time and transition data are automatically tallied by the software. The most recent version of the software saves data to the computer automatically. Video tracking systems have also been used to score this task. Interested readers can find more information in Page et al. (2009), Kaidanovich-Beilin et al. (2011), and at http://www.noldus.com/animal-behavior-research/solutions/research-small-lab-animals/sociability-test.
Time spent sniffing the wire cup containing the novel mouse and time spent sniffing the novel empty wire cup can be scored either live or from recorded experiments, using the same basic methods. When scoring a testing session live, care needs to be taken to avoid making noises that could disrupt ongoing behaviors in the subject mouse. The experimenter should sit at least 2 meters away from the apparatus, facing the center chamber directly, with equal views to both side chambers, and proceed as follows: (1) Hold two silenced stopwatches (see the Basic Protocol, step 25 annotation, and Fig. 8.26.2), one in each hand, for scoring time spent sniffing each cup. (2) Watching the subject carefully, start the timer as soon as the mouse begins sniffing, and stop the timer promptly when the mouse stops sniffing. The subject mouse is considered to be sniffing the cup when its nose is directed to the cup and the distance between the nose tip and wire bars is 1 cm or less. Remember to write down the cumulative time spent sniffing before clearing the timer for the next session.
If test sessions are to be recorded for later scoring, make sure that both cups and the circular region surrounding each cup are clearly visible on the screen. An ideal strategy for achieving this is to videotape from above, with the camera mounted ~1.5 meter above the center chamber. To ensure accuracy, it is advisable for two people to practice with the same videos until inter-rater reliability is 95% or higher.
Because autism is diagnosed purely based on behavioral symptoms, it follows that a mouse model of autism must have strong face validity to clinical symptoms. The three-chambered task was based on similar equipment in the behavioral neuroscience literature, modified to measure simple social approach in mice. While normally developing children prefer to play with friends, children with autism often do not approach other children, but remain alone in a corner, playing with a toy. The three-chambered task simulates the playground scenario by asking the subject mouse, “Do you prefer to spend time with another mouse or with an inanimate object?” Owing to its strong face validity and reliability, the three-chambered task has become one of the most widely used behavioral tests for studying autism-relevant phenotypes in inbred strains and transgenic mouse lines. In recent years, the three-chambered task has been used to investigate environmental and genetic hypotheses relevant to autism (Crawley et al., 2007; Yang et al., 2007a,b, 2008, 2009; Chadman et al., 2008; DeLorey et al., 2008; Jamain et al., 2008; Molina et al., 2008; Matsuo et al., 2009; Nakatani et al., 2009; Page et al., 2009; Zhou et al., 2009; Zhao et al., 2010 Silverman et al., 2011) and to test efficacies of pharmacological and behavioral treatments (Silverman et al., 2010a; Chadman, 2011; Yang et al., 2011a).
Employing multiple units of the three-chambered apparatus simultaneously offers high-throughput, simple, automated, and standardized measures of direct social approach behaviors. Although fewer complex and nuanced differences in reciprocal social interaction are captured, the task is suitable as a first-line screening assay for autism-like phenotypes.
Subject mice are group housed mice between 8 weeks and 6 months of age. They do not have to be experimentally naïve. With the exception of highly stressful tests (e.g., learned helplessness, fear conditioning, Morris water maze), mice having experiences in many other low-stress tests (e.g., developmental milestones, juvenile social interaction, elevated plus-maze, open field exploration, light ↔ dark exploration, general health battery) usually display normal sociability scores in the three-chambered task (Crawley et al., 2007; Chadman et al., 2008; Yang et al., 2009). Retesting the same subjects in the three-chambered task appears to be unproblematic (Moy et al., 2004), provided that there is a break of several weeks between the initial test and the retest.
Sexual experiences and social isolation have strong influences on social behaviors in adult mice (Rawleigh et al., 1993; Koike et al., 2009; Bouet et al., 2011; Kercmar et al., 2011). To avoid these confounding factors, sexually naïve and group housed mice are preferred for this task. Episodic aggressive behaviors in the home cage are not unusual in adult male mice. It is better to avoid testing males that are unusually aggressive or submissive, since their social approach behaviors might be confounded by aggression or social anxiety.
Target mice are group-housed 129/SvImJ adults between 8 weeks and 6 months of age, preferably of the same sex as the subjects. Low activity and low aggression in this strain ensure that the target mouse sits quietly during the test session and that all of the approach is initiated by the subject mouse. Other inactive strains (e.g., AJ, LP/J, 129/SvJae) could be used instead of 129/SvImJ. In our experience (Yang and Silverman, unpub. observ.), sociability scores are usually not affected by the novel mouse’s strain and/or sex. However, we prefer inactive strains because they continue to sit quietly in the apparatus after repeated use, and over a long period of time. A same-sex novel mouse is used to avoid variability related to estrous cycles. The novel mouse could be younger or older than the subject, as long as both animals are within the age range of 2 to 6 months and their body weights are within 5 grams of each other.
The single most important parameter of the three-chambered task is time spent in the two side chambers. Time spent sniffing the novel mouse and the novel object provides a corroborative and more specific measure of social investigation. The comparison between the two sides determines whether a group of animals displays sociability or not. Time spent in the center chamber is shown in graphs for illustrative purpose only, and should NOT be included in the statistical analysis for sociability. Repeated Measures ANOVA or Student’s t test is used to compare time spent in the chamber containing the novel mouse and time spent in the chamber containing the novel object, as well as time spent sniffing the novel mouse and time spent sniffing the novel object. A simple unpaired t test is appropriate for comparing chamber time or sniff time within a single group. Repeated Measures ANOVA is more appropriate for large experiments that have multiple groups, and when sociability needs to be analyzed within each group, e.g., within each drug dose when several doses are tested. This is because Repeated Measure ANOVA avoids the type I error associated with multiple t tests. If the ANOVA or t test result is statistically significant, sociability is present. If the result is not significant, sociability is absent. Sociability is a yes-or-no phenotype, obtained by comparing two sides within each group. Sociability is NOT a graded parameter for quantitatively comparing chamber times across groups. More finegrained reciprocal interaction tests are needed for quantitative comparisons of social behaviors across groups.
The automated three-chambered task was originally designed to investigate simple sociability deficits in mouse models of autism. Positive findings are then followed up with a separate, more complex analysis for reciprocal social interactions (e.g., scoring dyads of freely moving mice on graded, quantitative parameters such as follow, crawl over and under, nose-to-nose sniff, and nose-to-anogenital sniff). An example of a robust mouse model of autism that exhibits deficits in both the three-chambered task and the reciprocal social interaction test is the BTBR T+tf/J inbred strain (Yang et al., 2007b, 2009; McFarlane et al., 2008).
The three-chambered task requires that the subject be able to move around and explore all three chambers. Inactivity is a major confounding factor that could interfere with the experiment and skew results. Number of transitions is an important built-in control parameter that could be used to measure exploratory locomotion. This variable could be compared across different groups to identify differences in activity. Animals with extremely low numbers of transitions during Phase II and Phase III could be excluded from the experiment. For an example of excluding inactive animals from data analysis, see Yang et al. (2009).
In most cases, this is due to photocell malfunctioning, which happens when a photo sensor is blocked, e.g., by mouse hair, dirt, or bedding. To check for problems, turn on the machine and push a toy car through the doorways; monitor the time and transition numbers that occur on the screen as the toy car passes. Locate the problematic photocells, and clean the lens with moistened cotton-tipped swabs. Restart the machine after cleaning. This is usually all it takes to correct the error. If the problem continues, contact the equipment manufacturer.
These two parameters usually correlate positively with each other. However, there are occasions when chamber time is not significant but sniff time is significant. How do you decide whether the animals have displayed sociability or not? There is no simple answer to this quandary, but there are several things to consider: (1) double-check the video to see if sniff time has been scored accurately, and whether the animals exhibited any competing or unusual behaviors, e.g., excessive self-grooming, jumping, or rearing; (2) check the number of entries (in our experience, animals that are hyperactive or inactive tend to have low sniff time); and (3) repeat the experiment with a new cohort of mice. If the same results are found, we suggest drawing conclusions based on chamber time.
Subtle environmental factors could have substantial influences on mouse behaviors. When inexplicable behaviors are observed, the first thing to do is to look into factors that could agitate mice. For example, cage changing could make mice aggressive and hyperactive, an inexperienced animal caretaker or experimenter might handle the mice inappropriately, noises and vibrations could distract the animals, odors that come through a ceiling vent could make the animals rear up and sniff the air frequently, or rats from other nearby laboratories could induce fear reactions in subject mice. This is by no means an exhaustive list. To get reliable results, one needs to be vigilant about potentially interfering factors and meticulous in controlling them.
Mice used as controls in experiments are usually those that are known to display significant sociability. C57BL/6J, FVB/NJ, and FVB/Ant strains have been shown to exhibit high sociability in a number of published studies (Bolivar et al., 2007; Moy et al., 2007, 2008b; Silverman et al., 2010a,b; Yang et al., 2011a). It is therefore expected that these strains should show sociability in every laboratory. If a group of 8 to12 adult mice of one of these highly social strains fails to show sociability, it is almost certain that something is wrong. Procedural errors and environmental stressful factors are the most common sources of unusual results. Experiments should be put on hold until the control strain resumes normal sociability in a consistent manner. To monitor for occurrences of unusual results, always include a few mice from the control strains in your routine experiments. If the control mice behave the way they typically do, you can be more confident about your results.
For researchers who are testing genetic hypotheses of autism by using transgenic and knockout mice, the wildtype littermates are the correct control group. A lack of sociability in the wildtype group makes it difficult to interpret data from the mutant group. A critical factor to consider here is the genetic background of the mice. Because embryonic stem (ES) cells derived from 129 mice colonize germ lines efficiently, they are the most widely used cell lines for producing targeted mutants. Unfortunately, some 129 substrains are not ideal for behavioral experiments, owing to their inactivity, anatomical abnormality, low sociability, and/or high anxiety-like behaviors (Rodgers et al., 2002; Holmes et al., 2003; Moy et al., 2004, 2007). Transgenic mice that are on a 129 background or a mixed background with a substantial portion from a 129 strain are likely to behave differently from mice on a B6 background. Problems in general health, especially neurological reflexes, vision, olfaction, and locomotion can significantly affect behaviors in the social approach task. It is therefore important to evaluate animals for general health as described by Crawley (2007a, 2008) before testing them in the social approach task. Additional tests for anxiety-like phenotypes might be necessary to control for low general exploration and artifactually high amount of time spent in the center chamber. If the wildtypes do not display sociability due to genetic background issues, one solution is to backcross the mutant line with B6 for multiple generations until phenotypes of the wildtype controls are almost identical to those of B6.
Large strain differences in social behaviors exist among inbred mouse strains. This topic has been discussed at length elsewhere (Yang et al., 2011b). Among commonly used strains, multiple laboratories have consistently reported high sociability in B6 and FVB (FVB/NJ and FVB/Ant) strains and low sociability in BTBR and BALB/c strains (Brodkin et al., 2004; Bolivar et al., 2007; Panksepp and Lahvis, 2007; Moy et al., 2007, 2008a; McFarlane et al., 2008; Yang et al., 2007b, 2009, 2011a).
As Figure 8.26.3 illustrates, B6 and FVB/Ant mice spent significantly more time in the chamber containing the novel mouse than in the chamber containing the novel object, and more time sniffing the novel mouse than sniffing the novel object (indicated by asterisks on top of novel mouse bars), whereas BTBR mice spent similar time with both novel stimuli in either chamber. These strains with highly replicable sociability scores could be used as benchmarks in laboratories that conduct the three-chambered task routinely. Unusual behaviors in these strains often indicate methodological problems that require correction before proceeding with experiments.
There is usually a positive and significant correlation between chamber time and sniff time (Fig. 8.26.4), meaning that chamber time alone might be sufficient for high-throughput screening. A lack of significant correlation between chamber time and sniff time might indicate methodological problems to be solved.
The pretest acclimation of novel target mice takes ~1 hr. The social approach task, with its three 10-min sessions (Phase I, II, and III), plus intersession transitions and post-test cleaning, takes ~45 to 50 min to complete. It takes another 5 to 10 min to take notes, organize/transfer data, and reset machines. Altogether, this basic single-unit method takes one person 60 ± 10 min to finish for one mouse. It is reasonable to test six to eight mice a day. With a high-throughput setup using four or more boxes simultaneously, it is possible to finish six rounds, up to 24 subjects a day.
This work was supported by the National Institute of Mental Health Intramural Research Program MH02179-24.