Address correspondence to Dr Thomas N. Seyfried, Biology Department, Boston College, Chestnut Hill, MA 02467, U.S.A. E-mail: firstname.lastname@example.org
Purpose: Autism is a multifactorial disorder that involves impairments in social interactions and communication, as well as restricted and repetitive behaviors. About 30% of individuals with autism develop epilepsy by adulthood. The EL mouse has long been studied as a natural model of multifactorial idiopathic generalized epilepsy with complex partial seizures. Because epilepsy is a comorbid trait of autism, we evaluated the EL mouse for behaviors associated with autism.
Methods: We compared the behavior of EL mice to age-matched control DDY mice, a genetically related nonepileptic strain. The mice were compared in the open field and in the light–dark compartment tests to measure activity, exploratory behavior, and restricted and repetitive behaviors. The social transmission of food preference test was employed to evaluate social communication. Home-cage behavior was also evaluated in EL and DDY mice as a measure of repetitive activity.
Key Findings: We found that EL mice displayed several behavioral abnormalities characteristic of autism. Impairments in social interaction and restricted patterns of interest were evident in EL mice. Activity, exploratory behavior, and restricted behavior were significantly greater in EL mice than in DDY mice. EL mice exhibited impairment in the social transmission of food preference assay. In addition, a stereotypic myoclonic jumping behavior was observed in EL mice, but was not seen in DDY mice. It is of interest to note that seizure activity within 24 h of testing exacerbated the autistic behavioral abnormalities found in EL mice.
Significance: These findings suggest that the EL mouse expresses behavioral abnormalities similar to those seen in persons with autism. We propose that the EL mouse can be utilized as a natural model of autism and epilepsy.
Autism involves a spectrum of multifactorial neurodevelopmental disorders affecting up to 1 in 110 children in the United States (Rutter, 2005; Centers for Disease Control and Prevention, 2009). Autism occurs in early childhood with symptoms that include impairments in social interactions and communication, as well as restricted and repetitive behaviors (Rapin & Tuchman, 2008). A strong interplay of genetics and environmental risk factors influence the autistic phenotype, as the concordance rate among monozygotic and dizygotic twins is 88% and 31%, respectively (Rosenberg et al., 2009).
Several neurochemical abnormalities have been characterized in EL mice. EL mice exhibit an increased expression of glial fibrillary acidic protein (GFAP) in the hippocampus compared to nonepileptic control C57BL/6J mice, although there is no increase in hippocampal GFAP-positive astrocytes prior to seizure onset (Brigande et al., 1992). Similar to individuals with autism (Vargas et al., 2005), EL mice exhibit increased microglial activation in the absence of obvious neuronal loss (Brigande et al., 1992; Drage et al., 2002). Ramified microglia appear in the cerebral cortex, hippocampus, cerebellum, as well as throughout the brain in EL mice. Brain gangliosides, sialic acid–containing glycosphingolipids, are also altered in EL mice. Ganglioside GD1a concentration is elevated in the cerebral cortex and hippocampus of EL mice (Brigande et al., 1992). EL mice have lowered levels of GT1b and GQ1b gangliosides and elevated levels of ganglioside GD3 in the cerebellum (Brigande et al., 1992). Because these gangliosides are enriched in neural membranes, these findings suggest that EL mice express altered neural membrane structures.
Along with a robust epileptic phenotype and defined neurochemical abnormalities, an abnormal behavioral profile of EL mice has also been well characterized. Motor abnormalities have been observed in EL mice as early as postnatal day 5 (McFadyen-Leussis & Heinrichs, 2005). EL mice have an increased latency in their surface-righting reflex (time to return from a supine to a pronated position) and their negative geotaxis reflex (time to orient head from downslope to upslope) at postnatal day 5 that persists through postnatal day 9. In addition to early motor abnormalities, EL mice also show an inability to habituate to a novel environment at postnatal days 25–30 (McFadyen-Leussis & Heinrichs, 2005). Therefore, EL mice express abnormal behavior by postnatal day 5.
The abnormal motor behavior of EL mice also extends to irregularities in social behavior. EL mice exhibit short-term memory deficits in social investigation with juvenile conspecific mice (Lim et al., 2007). EL mice also exhibit a deficit in interacting with and investigating an unfamiliar adult mouse in a novel cage environment (Turner et al., 2007). These findings indicate that EL mice exhibit impaired social interactions. The abnormal social behavior of EL mice is also connected to maternal deficiencies. For example, EL dams display nursing deficiencies and have an increased latency in initiating pup retrieval (Bond et al., 2003). EL dams also display a preponderance of nonmaternal activity between postpartum days 2 and 5. During this time, self-grooming and exploration is significantly elevated in EL dams. The maternal deficiencies support an impaired social interaction phenotype.
The epilepsy-prone EL mouse strain and the nonepileptic EL background strain, DDY, were originally obtained from J. Suzuki (Tokyo Institute of Psychiatry, Tokyo, Japan) and from Clea Japan, Inc. (Tokyo, Japan), respectively. The mice were maintained in the Boston College Animal Care Facility as inbred strains. The mice were group-housed (prior to initiation of study) and kept on a 12-h light–dark cycle, with lights on at 6 a.m. and lights off at 6 p.m. All mice were tested between noon and 5 p.m. The procedures for animal use were in strict accordance with the NIH Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee.
Open field test
Mice were placed in the center of a 40 × 20 × 20-cm open field apparatus, and behavior was measured for 30 min. Mouse movement within the apparatus was tracked by the SmartFrame Cage Rack System using infrared beams, and data were analyzed using MotorMonitor software (Kinder Scientific, San Diego, CA, U.S.A.). The infrared beams register mouse location, distance traveled, and rearing capabilities. Locomotor activity was measured as the number of basic movements and rearing events the mice performed, along with time spent rearing and resting in the periphery of the open field. Rearing events were measured as the number of times the mouse stood on its hind legs. The number of entries into the center, distance traveled, and total time spent in the center of the open field measured exploratory behavior and the ability of mice to express fear-related responses.
Light–dark compartment test
The testing apparatus consisted of two compartments: a dark compartment and a light compartment. The dark compartment served as the control environment and the light compartment served as the novel environment. Mice were initially placed in the dark compartment and were allowed to move freely between the light and dark compartments. The latency for mice to completely enter the illuminated compartment and total time that the mice spent in the illuminated compartment were considered for statistical analysis. Each test lasted 10 min and was performed once for each mouse.
Myoclonic jumping assessment
Mice were singly housed for 24 h and provided food and water ad libitum. Each mouse was observed in its home cage for 1 min at 10-min intervals according to standard procedures for assessing myoclonic jumping behavior (Weiner et al., 1979; Nausieda et al., 1982; Carvey et al., 1989). The observation period lasted for a total of 70 min. Myoclonic jumping was defined as repetitive, hind-limb jumping in the corner of a cage with the muzzle oriented upward. Mice were scored as either positive or negative for exhibiting the myoclonic jumping behavior.
Social transmission of food preference
Flavored, powdered chows were used for testing according to standard procedure (Wrenn, 2004). The flavored chows consisted of either 1% ground cinnamon or 2% baking cocoa mixed by weight with rodent chow. Food was placed into a 25-ml beaker filled with baseplate wax up to the 25-ml mark. This permitted up to 5.0 g of powdered chow to be placed in the beaker and allowed mice easy access to the food. Mice were separated into “demonstrator” and “observer” groups. Demonstrator mice were fasted and then singly placed into a mouse cage with the hopper and bedding removed. Juvenile mice were fasted for 12 h, whereas adult mice were fasted for 24 h. A weighed beaker, containing either cinnamon or cocoa chow (cued food), was randomly assigned and placed in the cage. Mice were fed for 1 h and the amount of food consumed was measured. A demonstrator mouse was then placed into a cage containing three conspecific, age-matched observer mice. Demonstrator and observer mice interacted for 30 min. Each observer mouse was then singly housed and fasted for 12 or 24 h, depending on age. After the fast, the observer mouse was then placed into a test cage (without bedding and hopper) containing a weighed beaker filled with cinnamon chow randomly placed on one side of the cage and a weighed beaker of cocoa chow placed opposite the cinnamon chow (cued vs. novel food). Position-bias was prevented by placing cages onto a rack in the middle of a testing room with light sources on both sides of the rack. Mice were allowed to feed for 1 h and the food consumed was measured. Only mice that had consumed at least 0.2 g of food were considered for analysis. The percentage of cued food consumed to total food consumed was analyzed.
Analysis of variance (ANOVA) and the two-tailed Student’s t-test were used to evaluate the significance of differences for each behavioral parameter between strain and age (SPSS Software version 10, SPSS Inc, Chicago, IL, U.S.A.). Female and male mice were pooled for analysis, as statistical differences were not observed between sexes. Juvenile and adult mice were pooled across age for analysis in the social transmission of food preference test, since no statistical difference was observed between ages.
Two separate mouse cohorts were tested. The first cohort consisted of juvenile female and male EL and DDY mice (30–40 days of age). Testing of these mice was completed at or before 40 days. In addition, adult female (>180 days of age) EL and DDY mice were tested to assess the effects of seizures on the autistic phenotype. Of the 12 adult EL mice used in the first cohort, 7 experienced a seizure within 24 h prior to testing and 5 had not experienced a seizure within 1 week prior to testing. To control for litter effects, we used juvenile EL and DDY mice from five and four litters, respectively. Adult EL and DDY mice were derived from five and four litters, respectively. These mice were used in the open field test, light–dark compartment test, and the myoclonic jumping assessment. Myoclonic jumping was evaluated first followed randomly on the same day by evaluation in the open field test or the light–dark compartment test. The remaining test (open field or light–dark compartment) was performed on the following day. The second cohort consisted of juvenile and adult female EL and DDY mice. The EL and DDY mice used were derived from seven litters. The second mouse cohort was tested for social transmission of food preference. Adult EL male mice were excluded from the study, as adult EL males die sporadically from acute uremia poisoning due to urinary retention (Todorova et al., 2003).
Open field test
The open field test examines abnormal behavior in locomotor/exploratory activity and assesses the level of fear-related response in mice (File, 1980; Crawley & Paylor, 1997; Paylor et al., 1998). Restricted and repetitive behavior was measured by fine movement in the open field. Fine movement specifically measures self-grooming and head weaving, which are both considered forms of restricted and repetitive behavior when performed excessively (Devaud, 2003; Lewis et al., 2007). Significant differences in locomotor activity in the open field were observed between EL and DDY mice (Fig. 1). Basic movement (whole body position changes) in the open field was greater in EL mice than in DDY mice (Fig. 1A). EL mice spent significantly less time resting in the periphery of the open field than DDY mice, indicating a more active phenotype (Fig. 1B). EL mice exhibited a profile of increased and sustained rearing compared to DDY mice (Fig. 1C,D).
Exploratory behavior, along with the level of fear-related response, was measured by the activity of the mice in the center of the open field. Mice normally fear and avoid the center of open fields (Parks et al., 1998). The center of the open field was defined as a 20 × 10 cm area within the 40 × 20 cm open field. Exploratory behavior was greater and the level of fear-related response in the open field was less in EL mice than in DDY mice (Fig. 2). EL mice entered (Fig. 2A), perambulated (Fig. 2B), and remained in the center of the open field significantly more than DDY mice (Fig. 2C), as illustrated by the open field hot spots (Fig. 3). DDY mice displayed normal thigmotaxic behavior in the open field, as evidenced by their preponderance of activity near the corners and sides of the cage, whereas EL mice displayed indiscriminate activity throughout the open field (Fig. 3). Repetitive and restricted behavior, as measured by fine movement, was also significantly greater in EL mice than in DDY mice (Fig. 4A). Therefore, the open field test indicated that EL mice express abnormally high locomotor activity, have high levels of exploratory behavior and decreased fear-related responses, and engage in restricted and repetitive behaviors.
Light–dark compartment test
The light–dark compartment test examines the propensity of mice to explore a novel environment (Crawley et al., 1997; Crawley, 1999; Bourin & Hascoet, 2003). Exploratory behavior was significantly greater in EL mice than in DDY mice (Fig. 5). Latency for emergence from the dark compartment was significantly less in EL mice than in DDY mice (Fig. 5A). EL mice also spent a significantly greater amount of time in the light compartment than DDY mice (Fig. 5B). The light–dark compartment test confirmed findings in the open field test that EL mice engage in high levels of exploratory activity.
The social transmission of food preference test examines the ability of mice to communicate through social interaction (Galef & Wigmore, 1983; Wrenn, 2004; McFarlane et al., 2008; Ryan et al., 2010). EL mice showed significantly less preference for the cued food compared to DDY mice (Fig. 6). Cued food consumption relative to total food consumption was significantly less in EL mice (61 ± 6%) than in DDY mice (79 ± 3%).
Seizure activity influence on autistic behavior
Seizures occur in EL mice due to fear and stress, and routine handling induces seizure activity in EL mice (Todorova et al., 1999). Seven EL mice had seizures induced within 24 h prior to testing (seizing EL mice) and were compared to five EL mice that had not seized within 7 days prior to or during testing. Significant differences in behavior were observed between nonseizing and seizing EL mice at adult ages (Table 1). In the open field, EL mice that had seized exhibited greater basic movements and spent less time resting in the periphery than nonseizing EL mice. In addition, entrance and perambulation in the open field center was significantly greater in EL mice that seized than in EL mice that did not seize. EL mice that seized also had a tendency to spend more time in the center of the open field than EL mice that did not seize, although this difference in behavior did not reach statistical significance (p=0.081). A greater number of fine movements were recorded from the EL mice that seized compared to nonseizing EL mice. In the light–dark compartment test, EL mice that seized spent significantly more time in the light compartment than nonseizing EL mice. EL mice that seized also had a tendency to emerge from the dark compartment quicker than nonseizing EL mice, although this difference was not statistically significant (p =0.546). The myoclonic jumping behavior was observed significantly less in the seizing EL mice compared to seizure-free EL mice. These results suggest that seizure activity exacerbated the hyperactivity and exploratory behavior in EL mice.
Table 1. Comparative analysis of adult DDY and EL mice and influence of seizures on behavioral tasksa
aData are presented as mean ± SEM.
bControl adult DDY mice (n = 12).
cEL mice that had not seized within 24 h of testing (n = 5).
dEL mice that had seized within 24 h of testing (n = 7).
*Level of significance between seizing and nonseizing EL mice (p < 0.05).
Results from the open field and the light–dark compartment tests showed that the EL mice were hyperactive. Hyperactivity is a symptom prevalent in children with autism (Ando & Yoshimura, 1979). In addition, attention-deficit hyperactivity disorder has been suggested as a comorbid trait of autism (Spencer et al., 1999; Frazier et al., 2001; Leyfer et al., 2006). EL mice also expressed abnormal ambulatory behavior in the open field, and decreased fear-related response in the open field and in light–dark compartment tests. Mice are normally fearful of and avoid lighted areas and open space, as this exposes mice to predators (Archer, 1973; Walsh & Cummins, 1976; Powell & Banks, 2004). It is known that individuals with autism have a deficit in perception and attention to environmental stimuli (Courchesne et al., 1994). Similar deficits in EL mice may prevent focus on environmental cues that would normally evoke a fear-related response. It is also possible that EL mice have inhibitory deficits in their locomotor activity and exploratory behavior, even in the presence of a threatening situation. A lack of inhibition leading to hyperexcitability has been well-established with regard to seizure generation (Balcar et al., 1978; Ribak et al., 1979; Ribak & Reiffenstein, 1982; Ribak et al., 1989). Inhibition deficits may exist in neural circuits responsible for controlling locomotion and exploratory behavior in EL mice. This, coupled with their hyperactivity, may lead to the observed high exploratory behavior in EL mice. Further behavioral tests will be needed to discern if an inability to perceive environmental stimuli or inhibit behaviors actually exists in the EL mouse.
Restricted and repetitive behavioral abnormalities, part of the diagnostic criteria for autism (American Psychiatric Association, 2000), were also observed in EL mice. This was seen in the open field as fine movements in the form of excessive self-grooming and head weaving. Repetitive and restricted behavior was also exhibited as a whole-body movement in the form of myoclonic jumping. The myoclonic activity observed in EL mice is reflective of an autistic phenotype and not a seizure phenotype, since juvenile mice exhibited the behavior prior to seizure onset. Hence, a robust phenotype consisting of restricted and repetitive behaviors is present in EL mice.
We used the social transmission of food preference assay, as it is considered the standard test for measuring communication in mice (Boylan et al., 2007; McFarlane et al., 2008; Ryan et al., 2010). It is important to note that EL mice do not have a deficit in investigating novel odors (Pascual & Heinrichs, 2007). The results of the social transmission of food preference assay show that EL mice exhibit a deficit in olfactory communication, an important mode of social communication in mice (Hurst, 1989). This deficit validates the autism diagnostic criteria of impaired communication (Ryan et al., 2008).
Our results of stereotypic behaviors and impaired communication, together with previous reports of social deficits in EL mice (Lim et al., 2007; Turner et al., 2007), are the first to demonstrate salient abnormalities relevant to all three diagnostic components of autism in EL mice. This, along with other abnormalities relevant to autism (Table 2), highlights the value of the EL mouse as a model of autism and epilepsy.
Table 2. Summary of the fidelity of the EL mouse as a natural model of autism
Characteristics of persons with autism and EL mice
Increased locomotor activity and exploratory behavior (Figs. 1 and 5)
Epilepsy (Malow, 2006)
Model of epilepsy
The behavioral abnormalities associated with autism generally occur prior to the onset of seizures in children with autism and epilepsy (Volkmar & Nelson, 1990; Tuchman & Rapin, 2002). This developmental relationship between autism and epilepsy was also observed in EL mice, as seizures do not occur in juvenile mice, but rather at sexual maturity (Todorova et al., 1999). Interestingly, we found that seizures exacerbated the autistic behavior in EL mice. The EL mice that had not seized within 24 h prior to testing were confirmed to be seizure-free for this period, since seizures do not occur spontaneously in EL mice, and must be induced through handling (Todorova et al., 1999). Seizures increased hyperactivity, exploratory behavior, the stereotypic behaviors of self-grooming, and head weaving, but reduced the incidence of myoclonic jumping. To our knowledge, the influence of clinical seizures on autistic behavior has not been well-characterized in children (Di Martino & Tuchman, 2001; Canitano, 2007). Previous reviews suggest that autistic-like behaviors could be improved following treatment of epileptiform discharges (Binnie, 1994; Tuchman, 2000). A previous study showed that the behavior in autistic children with EEG abnormalities, but who never experienced clinical seizures, was improved following the administration of valproic acid (Plioplys, 1994). We suggest that seizure management might reduce some autistic behaviors in individuals with autism and epilepsy.
The utility of the EL mouse as a model of autism and epilepsy is seen in the robustness of the model’s phenotype, along with the ability of the model to replicate several behavioral abnormalities relevant to the major diagnostic criteria of autism. Previous models of autism have been generated through targeted gene disruption or through surgical brain lesions (Murcia et al., 2005; Klauck & Poustka, 2006; Moy et al., 2006). Other models of autistic behaviors have been developed through a forward genetics approach (Gilby, 2008; Ryan et al., 2010). The EL mouse is the first model in which all animals consistently replicate multiple behavioral abnormalities relevant to autism while sharing the comorbid disorder of epilepsy. Given the multifactorial nature of autism and the lack of consistent neurologic abnormalities found in autism, research has trended toward identifying autistic subpopulations that share neurologic abnormalities. The EL mouse can be an important model to parse the underlying etiology responsible for the codevelopment of autism and epilepsy. The EL mouse will also be integral in studying the unique neurologic challenges that affect individuals with autism and epilepsy.
We thank Dr. Xandra Breakefield and her staff (Molecular Neurogenetics Unit, Neuroscience Center, Massachusetts General Hospital, and Neurology Department, Harvard Medical School, Boston, MA, U.S.A.) for kindly supplying the open field apparatus and Motor Monitor software for our use during the study. We also thank Dr. Stephen C. Heinrichs (VA Medical Center, Research 151–Neuropharmacology, Boston, MA, U.S.A.) whose helpful comments improved the scope and focus of the manuscript. This research was supported by the Boston College Research Fund, and NIH grant (NS055195).
None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.