Fig. 1
Aggressive behavior, self-injurious behavior, and disruptive behavior (mean scores and SEM) for 3 rating scales of the Autism Spectrum Disorders-Behavior Problems for Adults (ASD-BPA) in 3 groups of patients with Autism Spectrum Disorder (ASD), Pervasive Developmental Disorder—Not Otherwise Specified (PDD-NOS) and intellectual disabilities (ID). Drawn from Table 2 [20]
The early detection of clinical signs of autism is one of the criteria that must treat any ASD model in mice. ASD is observable during the first months of life and the possibility of a diagnosis within the first year has been considered as confirming ASD diagnosis for a long time [21–30]. The onset of social behavior in mice (Sect. 2.1.1) will be examined in the prospect of defining a tool capable to detect early signs of sociability impairment. Social behavior in newborn mice and then social behavior in juveniles will be considered first. A deficit in communication remains the core feature of ASD. Several authors have considered the vocalizations emitted by the pups as “distress calls.” Could the deficit in vocalizations emitted by newborn mice serve to model the impoverished communication of autistic infants? Social behavior in adults will focus on the measures of sociability and interest for social novelty (Sect. 2.5). An attempt to model disruptive behavior is made in Sect. 2.6.
2.1.1 Onset of Social Behavior in Mice
Social behavior is one of the main issues in modeling autism or pervasive disorders in mice and special attention must be paid to the identification as well as the measurement of social relationships. The first step deals with the identification of social behavior during the days following birth and the second with the onset of social behavior in juvenile mice.
The first signs of social behavior appear with suckling behavior. It begins thus as soon as the first days of life with the competition for the best nipple. Jay Rosenblatt demonstrated that newborn rats compete for the lower nipples that are the easiest to reach and that provide the most abundant milk. A nipple once selected by a newborn is recognized and searched [7, 31, 32]. Competition for the best nipples can be investigated in the mouse 2 days after birth with a very simple method. The mother is removed gently from the nest and every pup is identified by recognizable markings. We use nontoxic markers or India ink or natural dye (see the list published by the European Food Safety Authority, EFSA, among others). The pups are put back in the nest with the mother. There are five pairs of nipples, from the rostrum to the caudal region: pars cervicalis, pars thoracica cranialis, pars thoracica caudalis, pars abdominalis, and pars inguinalis [33]. Later, it is possible to observe the position of the pups in relation with the nipples by lifting up the cage and looking through the floor if a transparent cage has been used. It is also possible to push gently the mother and to observe the position of the pups. Marking on the back makes the observation easier and more rapid. The observation can be made 2 days after birth and reiterated 2 days later to check for the persistence of the preferences. The marking must be renewed until the age when an unchangeable marking can be used. The pups occupying the pars cervicalis are scored 0, pars thoracica cranialis scored 1, pars thoracica caudalis scored 2, pars abdominalis scored 3, and pars inguinalis scored 4, according to the nutritive quality associated with each location.
2.1.2 Social Behavior in Juveniles
Categories of interactions have been reported in adult mice under several conditions. It is thus possible to date the onset of these behaviors in juvenile mice.
Sniffing the ground and digging or pushing the litter in the presence of partners are part of social behavior because they consist of marking the territory.
Self-grooming is considered as an indicator of stress in young mice but it may be different in adults as shown in sect. 3.1.3. The face and the other parts of the body are scored separately. The number of episodes and the length of self-grooming are scored.
Sniffing is characterized by the part of the body of the partner towards which the snout is directed. We distinguish here towards the face, the urogenital region, and the flank of the partner. The number of rummaging episodes and the length of the episodes are recorded.
Nesting is noted when the juveniles rest together either in the nest or in another area. The duration of the episode is scored. The position of the juvenile, below or on the top, is indicative of dominance. The below position is more protective than the top against predators and the dominant mouse is therefore below.
Rummaging through the fur of a partner (grooming the partner) is accompanied by placing one or the two fore paws on the body of the partner. The quantity (number) of sniffing and the length of the episode are recorded.
Rummaging can degenerate into bites that are identified by vocalizations or escape of the partner. Running away follows heavy grooming. These behaviors are scored.
Boxing is seen when two partners stand face to face, the fore paws agitated in a boxing manner or a scratching manner.
Wrestling appears when the two partners roll on the ground.
The conditions under which the social behaviors can be observed are highly variable.
(1) The between-sex social behavior is not of interest before 3 weeks of age because the young age excludes apparently a sexual component of social behavior. In this condition, males and females can be observed during independent sequences. (2) The pups of interest can belong to the litter when juvenile mutants are identified in the litter or when some juveniles from the litter have been subjected to appropriate treatments. In another situation the pups of interest can be introduced in the litter before the observation. The second situation adds attraction and rejection for a new partner and is more difficult to analyze.
2.1.3 Practical Considerations for the Observation of Social Behavior
Social observation requires video equipment to film the sequence. A snapshot is required to analyze the behavioral sequences. Male or female mice are placed in a round grey area with a 50 cm diameter, containing a mixture of clean and cage sawdust and 60 lux lighting on the ground. The sequence lasts for 40 min. As it is difficult or impossible to score the behavior of several pups during one viewing, we suggest viewing one video to score one pup. In the following illustration we viewed four times the video.
Table 1 indicates the number and duration of the previously defined social behavior observed in a litter of seven pups; four of them were males (observed here) and one of them (pup 2) carried an extra chromosome region from HSA21 chromosome (Roubertoux unpublished). The table indicates the number of behaviors directed towards one of the other partners, and summarises the results of nondirected behaviors.
Table 1
Social behavior in 18-day-old males of the same litter. Pups 1, 3, and 4 were wild-type mice and pup 2 carried an extra fragment of HSA21
(a) Social interactions with the sibs | (b) Non-directed social behaviors | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Identification of the pup | Average distance (cm) | Sniffing the face (nb) | Duration sniffing the face (s) | Sniffing urogenital region (nb) | Duration sniffing the urogenital region (s) | Sniffing the flanks (nb) | Duration sniffing the flanks (s) | Rummaging episodes (nb) | Duration of rummaging episodes (s) | Rummaging followed by bites (nb) | Boxing episodes (nb) | Wrestling episodes (nb) | Sniffing the ground duration (s) | Digging the litter duration (s) | Pushing the litter duration (s) | Self-groomings (nb) | Duration of Self-grooming (s) | |
Pup 1 towards | 2 | 17 | 11 | 44 | 7 | 5 | 8 | 51 | 20 | 65 | 2 | 2 | 0 | |||||
3 | 27 | 19 | 51 | 5 | 11 | 5 | 47 | 12 | 52 | 1 | 2 | 1 | ||||||
4 | 32 | 15 | 42 | 9 | 17 | 6 | 38 | 19 | 60 | 0 | 2 | 0 | ||||||
Σ | 76 | 45 | 137 | 21 | 33 | 19 | 136 | 51 | 177 | 3 | 6 | 1 | 21 | 54 | 38 | 8 | 35 | |
Pup 2 towards | 1 | 32 | 2 | 7 | 1 | 0 | 1 | 2 | 2 | 13 | 0 | 0 | 0 | |||||
3 | 49 | 3 | 10 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||||||
4 | 81 | 5 | 23 | 2 | 3 | 0 | 0 | 5 | 7 | 0 | 0 | 0 | ||||||
Σ | 162 | 10 | 40 | 3 | 3 | 1 | 2 | 7 | 20 | 0 | 0 | 0 | 121 | 68 | 87 | 42 | 89 | |
Pup 3 towards | 1 | 23 | 21 | 61 | 8 | 14 | 3 | 14 | 19 | 57 | 1 | 2 | 0 | |||||
2 | 35 | 19 | 81 | 11 | 51 | 1 | 4 | 9 | 28 | 3 | 0 | 1 | ||||||
4 | 39 | 25 | 45 | 21 | 33 | 0 | 0 | 11 | 35 | 0 | 0 | 0 | ||||||
Σ | 97 | 65 | 187 | 40 | 98 | 4 | 18 | 29 | 120 | 3 | 2 | 1 | 11 | 25 | 11 | 4 | 46 | |
Pup 4 towards | 1 | 43 | 16 | 42 | 12 | 28 | 6 | 23 | 5 | 32 | 1 | 0 | 0 | |||||
2 | 38 | 22 | 60 | 15 | 32 | 2 | 16 | 3 | 65 | 0 | 1 | 1 | ||||||
3 | 52 | 30 | 72 | 9 | 54 | 8 | 21 | 3 | 47 | 0 | 2 | 0 | ||||||
Σ | 133 | 68 | 174 | 36 | 114 | 16 | 60 | 11 | 144 | 1 | 3 | 1 | 10 | 34 | 27 | 12 | 29 | |
Figure 2 offers a visualization of the same results. It shows that pup 2 initiates few contacts with its sibs but that it performs more nondirected behaviors. The intensive territory marking or self-protection observed in pups may be analyzed as territory marking and social withdrawal. Table 2 shows the behaviors initiated by the four observed pups and directed to them. Pup 2 not only does not take the initiative of social contact but is also, in addition, more intensely targeted by the other partners.
Fig. 2
Activities in a group of 4 pups; (a) activities directed towards others; (b) spontaneous activities non-directed towards another pup. Black area: data of pup 2
Table 2
Social behaviors initiated by the 4 observed pups and directed to them
Pup 1 | Pup 2 | Pup 3 | Pup 4 | |
|---|---|---|---|---|
Initiated social behaviors | 139 | 21 | 162 | 135 |
Target of social behaviors | 101 | 143 | 97 | 118 |
2.2 Are Newborn Mice Vocalizations a Model of Human Communication?
Thirty years ago, one of us (PLR) said as a joke that New Zealand (NZB/BlNJ) pups could be considered as models of autistic disorders because they have learning deficits and because they do not emit vocalizations, or produce few vocalizations assimilated to a language deficit. Was the joke taken seriously? In any case, several papers modeling autism and pervasive disorders used vocalizations in mice as an index of the ability to communicate [34–53]. The question is “Are vocalizations in rodents communication processes?”
Mouse vocalizations are heterogeneous productions that include audible productions (below 20,000 Hz considered as the auditory human limit) and ultrasounds. Mice emit audible vocalizations in reaction to pain or discomfort. Ultrasounds include whistles (10–60 ms, between 30,000 and 130,000 Hz) and clicks (shorter than 5 ms) that are composite frequencies—between 10,000 and 30,000 Hz [54, 55] (see Fig. 3). The word “vocalization” includes the audible and the ultrasound production and the word “ultrasounds” includes whistles and clicks. A complete disappearance of ultrasounds after 11 days of age is mentioned [56–59]. They can reappear later under particular conditions (copulation, threat, or aggression) but there is no evidence for common mechanisms between the pup and the adult vocalization production.
Fig. 3
Whistle emitted by a 6-day C3H/He pup. Vocalizations were recorded on an Ampex 707 tape recorder, at a tape speed of 154 cm/s, enabling the recording of frequencies from 4 Hz to 100 kHz. The filter was a Belin-type F260 with a high-pass filter set at 20 kHz. A Bruel and Kjaer microphone-cathode follower 2614 with capsule 4130 and bandwidth from 20 kHz to 100 kHz was placed at 0 ~ incidence and 2.5 cm above the head of the pup. The signal presented here results from the transcription onto a sonagraph. The frequency and the length are shown on Y and X axes, respectively [54, 55, 183]
The typical internal temperature of a pup is 38.5 °C and the nest with the mother inside offers a similar temperature. The internal temperature drops rapidly when the pup is out of the nest until 7–9 days of age as the pup is poikiloterm during the period. The drop of temperature accelerates the respiratory rhythm producing a hyperventilation. The ultrasounds are probably not produced by the larynx but by the lungs [60]. The respiratory system is not mature and the pup is still not able to regulate the pulmonary fluids. Whistles could be caused by bronchial mucus and clicks by phlegm. Ultrasound production stops when the pup controls its own temperature. Vocalizations could be considered as an indirect indicator of internal thermoregulation processes. The capacity to decode the vocalizations in mice models of ASD makes sense if we evaluate the communicative value of the vocalizations.
2.3 Social Behavior in Adult Mice
Persons with ASD show deficits in social communication and deficits in social interactions (see Chap. 2). Social communication is defined by exchanges of signals between partners. Social interaction includes sociability and social novelty seeking. Sociability in mice is defined as the number of contacts displayed by the test male with a partner. Social novelty seeking is the number of contacts with a new partner. Excellent reviews have described situations that provide the possibility for evaluating both of them [61–64]. The Brodkin review [63] is particularly relevant because he hypothesizes that the BALB/cJ inbred strain presents several features that are observed in ASD including abnormal social behavior and he compares subsequently different measures of social behavior in which BALB/cJ are tested.
2.4 Modeling Communication Impairment in Mice
A communication process in two partners includes an emitter and a receptor and communication impairment in ASD can result from the inability to produce adequate social signals in the autistic person and/or from the inability to perceive the social meaning of the signals emitted by the non-autistic partner. The two processes are difficult to disentangle even in mice. Let us consider a dyadic encounter. The emitter produces a sensory cue. The receptor perceives the cue but it must be able to detect its meaning. Vocalizations are currently recorded in adult mice to model the difficulty of communication observed in ASD [3, 36, 37, 44–48, 50, 52, 53, 62, 65, 66]. The relevance of ultrasound production for studies of communication is different in newborn and in adult. The newborn ultrasound production results from a respiratory phenomenon excluding the larynx whereas the larynx contributes to ultrasounds in adult as shown by bilateral sections of the superior laryngeal nerve that reduce the number of ultrasounds and delay the emission [67]. This difference between the two processes is not sufficient to justify the exclusive use of vocalization counts to model communication deficits. The relationships between the number of vocalizations and social behaviors are weak and depend on internal and environmental factors. Aggression as well as copulation can occur in the absence of vocalization [67, 68]. We lack information on vocalization production and vocalization perception and this lack should incite to be cautious. Do vocalization patterns and their physiological correlates correspond with different social situations and different physiological states in the emitter? Is the poor production of vocalization, observed in a mutant, a general inability to produce or an inability to react to a social condition? What is the feedback from the receptor? Do its reactions reflect a non-understanding of the signals? Some results are encouraging but much more information is required before assimilating mouse vocalization and ASD communication defects. A common pathway between the murine homolog of SRPX2 gene (associated with human language) and mouse vocalization has been suggested [69]. Cortex regions specialized in the recognition of auditory patterns in mice have been identified [70]. Vocalizations could carry different meanings for the mice, suggesting that the variety of vocalization patterns is more interesting than the number of vocalizations [71].
Briefly, we suggest to use the reception of the vocalizations rather than vocalization production to model ASD communication impairment. In any case, the physical structure of the productions does not provide workable information when it is separated from physiological information about emission and reception processes.
The first solution consists of analyzing ultrasound production including the shape of the whistles, their harmonics, and their development [54]. Such an approach requires expensive material and collaboration with experts in acoustics. An alternative solution consists of counting ultrasounds. The equipment requires a bat detector. The S25 Bat Detector (Ultra Sound Advice, http://www.ultrasoundadvice.co.uk) has a wide frequency range because it covers the whole range of audible and ultrasonic range of mouse vocalizations (10–180 kHz). The bat detector is paired with a microphone. Due to recording high frequencies, an ultrasonic microphone must be employed such as the SM2 Microphone (Ultra Sound Advice). It is possible to count the whistles because the bat detector transforms the high frequencies into frequencies that are audible for the experimenters. More sophisticated analysis can be performed by filtering. According to the filter it is possible to select several bandwidths (20–40 kHz, 40.1–60 kHz, 60.1–80 kHz, and so on) and to count only the signals within the bandwidth. It is possible to collect rough information on the number of ultrasounds and their frequency. More sophisticated analysis can be performed, the difficulty being the accumulation of the data and the behavioral interpretation.
The recording protocol varies according to the age of the mouse. The pup is dropped off on a humid blotting paper covering a plate (avoid closed containers such as a cup for acoustic reasons; sawdust is not recommended because pup movements create acoustic artifacts). The temperature should be controlled because the number of ultrasounds depends on the temperature. The microphone is positioned at 4 cm from the pup, the distance being a crucial issue because the intensity of the ultrasounds varies as the inverse of the log value of the distance. The record starts 15 s after putting the pup on the blotting paper because the contact generates vocalizations. Recording ultrasounds in adult dyadic encounters requires a previous identification of the frequencies corresponding to each emitter. The microphone is oriented towards each of the two partners for this purpose. Two solutions are available to record vocalizations during the session. A complete recording and a graphic transcription offer the possibility to measure the parameters and the rhythm of the production during the encounter. Another solution consists of using two detectors, each being calibrated on the bandwidth of a partner.
2.5 Social Interactions in Mice
Very simply, observations during dyadic encounters are available. They have been used for a long time. The limit is the rapidity of the interactions that are often impossible to score. Video recording with several cameras is needed.
The situations proposed for the juvenile can be transferred to adulthood. The identification of the body part can be of interest as shown in Caroline and Robert Blanchards’ works [72, 73]. This will be discussed in the next section (disruptive behavior). It could be observed in a test male with an anesthetized partner [74]. Several authors [75] observed social barbering in dual encounters. The dominant rummages the whisker of the dominated.
2.5.1 The Two-Chamber Test
The test was considered as an anxiety test [76]. It presents all the characteristics of the most recent tests used to measure social interaction as previously observed [63]. We have slightly modified the procedure for our own observations. The housing cage was divided into two equal parts (see Fig. 4). A wire mesh partition separated equally the two parts of the cage. The level of light was 50 lux on the ground. The test male was housed for 10 min in area 2 and the partner was placed in area 1 at the end of the 10 min. At the end of the new period, partner 1 was removed and another partner (partner 2) was placed in area 1. We measured the time spent by the test male in the virtual area (subarea of area 2) during the three periods:
Fig. 4
Testing social relationships in mice. (a) The two-chamber test device. The squared cage is divided in two equal parts by a wire mesh. Area 1 is successively empty, occupied by partner 1 and then by partner 2. The dotted zone is the virtual zone. (b) Time spent in the virtual zone by the euploid and by the 152F7 segmental trisomic mice. (c) The three-chamber test as used in our laboratory. (d) Time spent in interactions by euploid and by the 152F7 segmental trisomic mice under the three conditions. (e) Adaptation of the three-chamber test
Empty area 1 (10 min)
Area 1 with partner 1 (10 min)
Area 2 with partner 2
We used the Videotrack system from ViewPoint (Lyon, France). We finalized the procedure with a mouse model of segmental trisomy 21. The 152F7 mouse which carries an extra copy four genes from HAS 21 was used [77, 78]. They were faced with independent C57BL/6ByJ as first partners and FVB as second partners. Six test males and six controls were used. As the extra segment is on the blind FVB background, the transgenic males resulted from C57BL/6ByJ x transgenic FVB crosses. The controls were C57BL/6ByJ x FVB wild type. Device and results are reported in Fig. 4 and the discriminating capacity of the two chambers and three-chamber test will be compared at the end of the last section.
2.5.2 The Three-Chamber Test
It was proposed by Crawley [79] and is widely used now. Several homemade three-chamber tests exist as well as variations on the general protocol. We used a 39 cm × 13 cm × 30 cm transparent Plexiglas box that was divided into three chambers of equal surfaces (13 cm × 13 cm) (see Fig. 4c). The three chambers communicate by 6 cm wide doors. The doors can be closed by removable sheets of transparent Plexiglas. The light level is 50 lux on the ground. We observed that brighter light inhibits movement. The first part is a progressive habituation. The test mouse is gently dropped off in the central chamber. The mouse is left in the central chamber—the doors being closed—for 10 min. At the end of minute 10, the mouse is directed to the central chamber and the doors are closed. A partner is placed in a wire pencil box in chamber 1. Beakers are used sometimes. They must be in glass, transparent, and perforated with several holes: in glass to be cleaned easily, transparent to see the partner inside, and perforated to allow odor, and vocalizations to pass [10]. The partner being in chamber 1, another wire pencil box (or a beaker), is placed in chamber 3. It includes an odorless objet, having a volume approaching the volume of the mouse. The doors are open. Time to explore the wire pencil boxes is measured during 10 min. The difference between exploration of the wire pencil box with the partner and the wire pencil box with a novel object is a measure of sociability.
There are three difficulties with the test. First, the measure is not an absolute measure because we have no standardization in the psychometric sense, for the test in the mouse. The socialization score has to be compared to the score obtained by a control. The test mouse is more or less socialized than another. Second, exploration has to be operationally defined. Exploration can be defined by the time spent within a virtual circle around the wire pencil box. The time can be recorded by a videotracking system (Videotrack, Viewpoint-Behavior technologies: http://www.viewpoint.fr/news.php). The number of contact (paw or nose) on the wire pencil box and duration of the contacts can also be recorded. Third, it is not possible to disentangle social curiosity and aggressive preliminaries. A solution consisted of removing the wire pencil box containing the partner, at the end of the 10 min of sociability observation. The test male was observed during the next 10 min to detect attack behavior [64].
Social novelty seeking is measured at the end of the test of sociability as in the two-chamber test. In the three-chamber test, the following procedure followed the one proposed by [79]. At the end of the sociability session, the test male is gently pushed towards chamber 2. The doors are closed. A male mouse replaces the object in wire pencil box 2 in chamber 3. The male has the age of the male in chamber 1 but it belongs to another strain. The doors being open, the time spent by the test male in interactions with partner 1 in wire pencil box 1 and in interaction with partner 2 in wire pencil box 2 is measured. The interactions are measured as in the sociability test. We draw a virtual circle with a diameter of 2 cm larger than the wire pencil box. The Videotrack system detected the time spent in this area. We tested 6 naïve 152F7 mice, and 6 C57BL/6ByJ x FVB wild-type controls with C57BL/6ByJ partner 1 and FVB partner 2. The results are shown in Fig. 4.
The performances of the euploid males follow the same trend for sociability and for social novelty seeking in the two devices (Fig. 4a–c). The 152F7 males interacted more with partner 1 than with the object in the two-chamber test and in the three-chamber test but the difference (empty area vs. partner 1) is smaller in 152F7 than in euploid mice. The two tests indicate nonsignificant social novelty seeking in 152F7. Briefly, the two procedures provide similar results.
We recently developed a modification of the three-chamber test that includes a new scoring method to measure the interactions. We measured the number of contacts towards a partner housed in a compartment by counting the number of nose pokes on the compartment and not the time spent in the neighboring pencil boxes. A 51 × 51 cm PVC grey box was divided into two equal parts by an opaque wall. A semicircular (5 cm diameter) in the lower part of the wall ensured the communication between the two parts: one being empty, and the other containing two compartments (wire round pencil box, 7 cm diameter, 10 cm high) separated by 20 cm (Fig. 4e). The tested mouse was introduced in the empty part, the door being closed, for 5 min. The door was then open and the mouse was allowed to explore the whole device for 30 additional minutes. The mouse was gently oriented then to the empty part and the door was closed. A C57BL/6J male was placed in one of the compartments, the other being empty. The door was open and the number of nose pokes on each of the two compartments was counted during 30 min after the tested mouse came into the two-compartment part of the box. The mouse was then gently oriented towards the empty part of the box and the door was closed. A SWR male mouse was placed in the second compartment. The door being open, the observations started again for 30 min and the number of nose pokes on each compartment was counted. A camera was located 2.60 m above the setup and all the observations were made on a screen. We measured the number of contacts towards a partner housed in a compartment by counting the number of nose pokes on the compartment. The succession C57BL/6J—SWR an SWR- C57BL/6J was alternated. The box and compartments were washed with household soap, hot water and dried between the tested males.
Deficit in sociability and seeking for social novelty can be modeled in mice. Two different tests gave similar results. Modeling the ASD deficit in communication seems more difficult. More advances in the study of mouse communication are required to develop test with acceptable face validity.
2.6 Disruptive Behavior
There is a consensus to use the tendency to initiate attack behavior for estimating the tendency to generate disruptive behavior in mice. Female mice exhibit agonistic behavior under specific conditions. Non-lactating females attack lactating intruders [80]. Neonatal injection of testosterone induces aggressive behavior in female mice. Males display spontaneous agonistic actions and are preferred in aggression studies for this reason. Different dyadic encounters have been imagined but the principle is always the same [81–85]. The tested male is brought face to face with an opponent and different indicators of aggressiveness are counted. The application however varies, creating different situations in which the indicators of aggression have neither the same meaning nor the same biological correlates [82].
2.6.1 Dyadic Encounter Designs Enter into a Three-Dimensional Structure
The social status of the test male corresponds to the first dimension [82]. This status is highly variable across the experiments (see Fig. 5). The male can be isolated, caged with a female, caged with other males or maintained in one of the conditions, and subjected to treatments.
Fig. 5
Three-dimensional representation of the most widely used tests for measuring attack behavior against conspecific male mice (modified from [82]
The males reared in different conditions have experienced different social relationships. Males are very often isolated as soon as weaning. Under these conditions, an isolated male does not know that some postures or actions are threatening or dissuasive and that they are emitted to discourage the partner. A male reared with other males has a rank in the hierarchy and it does not react as a naïve male. Previously defeated males attack less frequently and present neurochemical modifications [76]. The endocrine state of the male is also associated with its social status. Rodents are gregarious species and isolation generates stress and an increase of corticosteroids [86]. Is the rise of agonistic behavior the emergence of aggressiveness, an ability to recruit corticosteroids or a strong reactivity to isolation? The endocrine status of a male caged with several other males depends on its social status [87]. Two males with the same genotype and the same mother will differ because of their social status in the group.
The second dimension is defined by the place where the encounter occurs: territory of the test male, territory of the opponent or neutral area. Agonistic behavior can be considered as a protection of the territory in the first case, as a conquest in the second case and as spontaneous violence in the neutral area. Only the last condition may correspond to disruptive behavior.
The third dimension considers that the opponent modulates the agonistic behavior of the test male. Ginsburg and Allee pointed out strain-related differences in their pioneering study [88]. The NZB/BlNJ males display different agonistic behavior against opponents from different strains [89]. The opponent contributes to the score of aggressiveness that is attributed to the test male. The selection of the opponent is always difficult for these reasons. The males initiate attack behavior in several inbred strains of mice and the reply of the test male is rather defensive than offensive. Figure 6 indicates the ability of 11 strains to develop offensive behavior. The male can be also irritating. We preferred generally the very passive A/J male in most of our studies on aggression. The frequency of agonistic behavior is low in A/J and does not exceed 1/1,500.
Fig. 6
(a) Percentage of attacking males (same strains, different mice) in two similar conditions (1 and 1′). (b) Percentage of attacking males in independent groups of mice from the same strains. Mean score obtained in the same condition (1 + 1′)/2 vs. score obtained under condition 5 as described in the legend of Fig. 8
Each study of aggression can be assigned to a compartment in Fig. 5 according to the categories that are selected on the three dimensions. Although two compartments refer to different aspects of aggression, they are often considered as referring to “aggressiveness” as a single concept. Guillot et al. show that identical strains with two different aggression tests resulted in diametrically opposed conclusions. Different combinations are not equivalent although several authors refer to “aggressiveness” as a unique concept [90].
2.6.2 Measuring Aggressiveness
The dyadic encounter is generally employed. A standardized design has been proposed (see Fig. 7) [91, 92].
The test male is introduced into the arena where the observation is performed. Thirty seconds after, the opponent is presented. The observation starts with the first contact. The observation lasts 2 min after the first attack but it stops when the test male does not attack within 360 s after first contact. During the 2 min the number of attacks is counted, an attack being defined as a physical assault by the test male. Fight and chase are sometimes observed in the opponent. Tail rattling can be described as a fast waving movement of the tail. It precedes attack and it is considered as a component of aggression [93, 94]. It was considered as a warning signal to the intruder or as an indicator of anxiety [95]. It could be considered as an event that occurs during a dyadic encounter because the correlations that we observed between the number of rattling and the number of attacks are not correlated [85]. Brain adrenocorticotrophic hormone (ACTH) and beta-endorphin levels are significantly and negatively correlated with the amount of rattling, which is consistent with the hypothesis that rattling is a stress-related behavior or a maladaptive response to an intruder or to an unexpected change of the situation [83]. In this perspective, it could be of interest to evaluate the magnitude of the disruption caused by the intrusion. The latency of attack, rattling, chase or flight is measured. Limiting the assessment to 2 min takes into consideration the bias introduces by the negative correlation between latency of an event and number of events: It is obvious that male mice that attack late present (in probability) less attacks than those who attack earlier. Limitation to 2 min of the fight improves moreover animal welfare.
2.6.3 Frequently Used Protocols
A neutral area with standard opponent was described in 1986 [92], modified in 1988 [91], and selected for most of our studies on aggression. The dyadic encounter is performed in a transparent cage (L 42x W 26x H 18 cm) with a transparent lid as described previously. The floor of the cage is covered with sawdust from different housing cages of the tested males. The light level is 60 lux on the floor. The test male is introduced and 30 s after the A/J opponent is presented and the dyadic encounter takes place as described previously. The opponents were never used twice. The test males were not isolated but they were housed at weaning with a sister from the same litter. Several variants have been introduced: clean sawdust, round Robin test (the males of a strain are tested with the opponents belonging to several strains), the “dangler” procedure that accelerates the onset of attack behavior (in the dangler procedure, the opponent is picked up by the tail and brought close to the head of the test male) and the age of tested males. Their effects on attack behavior have been analyzed by a pioneering paper [88]. Figure 8 shows that the different procedures are more or less statistically independent.
Fig. 8
Factor analysis of attack behavior scores obtained in different tests. Loadings for five testing conditions: (1) non-isolated males, neutral area, tested once, opponent A/J ; (2) non-isolated males, neutral area, tested once, opponent A/J (duplicates condition 1); (3) isolated for 24 h, the day before testing, neutral area, tested once, opponent A/J; (4) isolated for 13 days, tested male cage (resident intruder procedure), tested once, opponent A/J; (5) isolated for 13 days, tested male cage (resident intruder procedure), tested once, opponent belongs to the strain of the opponent (graphical representation from previous results (Table 2—[85]))
The aggression score depends on the previous experiences in the field of social behavior. Defeated mice have a lower propensity to attack and winners seem to be boosted by a first victory. Mice from defeated strains can be aggressive when previously placed under conditions that ensure their dominance [96–100]. Using a mouse for several successive aggression tests is not a commendable strategy. This reactivity to previous experiences is a limit when the reliability of the aggression scores is sought. We overcame the difficulty by using different mice but the same strains. We computed the correlations between the mean scores of the strains. Figure 6a shows that the aggression mean scores did not differ in two experiences that have been separated by several months [82].
The resident intruder test owes its popularity to its rapidity and to the belief that the procedure induces attack behavior. The dyadic encounter is performed in the housing cage of the test male. Generally the test male is isolated, with isolation lasting from 2 weeks to several months. The resident intruder plus the isolation of the test male modulated agonistic behavior as shown in Fig. 6b, and the condition interacts with the strain because 45 % of the strains are stable or decreasing. Several variants are introduced: the size of the cage, the length of isolation, or the opponent status.
Social dominance is quite easy to perform but much more difficult to analyze. The task was first described by a group of pioneers in the field of behavior-genetic analysis [101]. A centimeter-graduated glass tube (4 cm in diameter and 40 cm length) was scotch taped on the table. The test male and an opponent selected for identical age and weight were placed at the opposite ends of the tube, the head into the tube. The retreat latency and the distance that the mice had moved inside the tube when the heads met are measured. The test was used to test Mecp2308/Y model of Rett syndrome [102] Dhcr7-HET model of Smith–Lemli–Opitz syndrome [103], SynI−/− , SynII−/−and SynIII−/−models of synapsin [104], and serotonin transporter Ala56 knock-in [105, 106]. A rapid retreat is expected as an indicator of reject of forced social contact.
It is not possible to select the best strategy to measure aggressiveness because the proneness to initiate attack behavior varies according to the testing and rearing conditions and to the genotype. Non-induced aggression (observed in a neutral area without isolation) is associated with steroid sulfatase that has a key role in neurosteroid modifications, plasma testosterone, and is negatively related with brain 5-HT, beta-endorphin, and ACTH. Aggressive behavior associated to isolation is not correlated with met-enkephalin, endorphin or dynorphin [82, 83].
The different tests measuring aggression in mice are available but they are not equivalent. Modeling disruptive behavior requires several tests. Naïve mice must be used in every test because the life events and particularly previous agonistic experiences modulate the results of dyadic encounter.
2.7 Technical Recommendations on Modeling Social Interests
It would be an error to consider communication or social behavior as independent from other characteristics in ASD and in mouse models of ASD. This remark leads to the need to control a large set of traits that are not specific to ASD but could help to analyze the experimental results. Certain of these traits are presented below.
Sensory abilities are well documented in inbred strains but they must be screened when the mice are born from crosses between non-inbred strains as in conditioned targeting [107–109]. Detection of sensory impairment is useful because it helps to select the tests and to interpret the results. For example, is it wise to subject albino mice to task such as open field or mice with retinal degeneration to elevated plus maze? Identifying sensory impairments in putative models of ASD is of interest per se because sensory and perceptive deficits are frequently reported in ASD [110] as well as impaired multisensory temporal integration [111, 112].Related posts:
Serotonin Disturbance in Mouse Models of Autism Spectrum Disorders
Zebrafish Social Behavior Testing in Developmental Brain Disorders
The Autistic Spectrum Disorders (ASD): From the Clinics to the Molecular Analysis
Invertebrate Models of Synaptic Transmission in Autism Spectrum Disorders
In Vivo Imaging in Mice
Modeling Autism Spectrum Disorders Motor Deficits in Mice
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
