Background
Autism spectrum disorder (ASD) is a common neurodevelopmental disorder marked by lifetime social functional impairment [
1,
2]. The essential features of ASD include impairment in reciprocal social communication; a deficit in communication ability; and restricted, repetitive behavior and interests [
1]. ASD occurs at a higher incidence in males than females, with a common consensus ratio of 4:1 [
3]. The prevalence of ASD has been estimated at approximately 1.5% and has increased dramatically over the past few decades [
4,
5]. Various genetic studies have provided convincing evidence that ASD is a complex and highly polygenic disease [
6]. However, the genetic underpinnings of ASD remain unclear, which impedes understanding of the disease pathology and the search for treatments.
Our previous study identified two novel chromosomal deletions in two unrelated ASD patients [
7]. One of the deletions, which spans 8p23.3-pter, contains three genes––
DLGAP2,
CLN8, and
ARHGEF10––that may be relevant to neurological functions. The functional loss of these genes might contribute to the clinical symptoms of ASD [
8,
9]. In this study, we target the gene
ARHGEF10 to further investigate the impacts of its functional loss. ARHGEF10, as a rho guanine nucleotide exchange factor (GEF), regulates rho GTPases by catalyzing the exchange of G-protein-bound GDP for GTP. There are over 60 rho GEFs identified in the human genome. However, few have been functionally evaluated in animal models. The Thr109lle mutation of the rho GEF 10 (
ARHGEF10) gene was found in a family of patients with slow nerve conduction and thinly myelinated peripheral nerves [
10]. In addition to its possible role in the myelination process, the single nucleotide polymorphism of ARHGEF10 has also been reported to be associated with schizophrenia [
11]. GEFs are the main regulators that facilitate the activation of rho GTPases by converting them from the GDP-bound state to an active GTP-bound state. Rho GTPases have been widely studied in neuronal development and neuronal diseases [
12]. Given the importance of rho GTPases in the nervous system, dysregulation of rho GEFs is believed to be involved in neurodevelopmental diseases. For example, a mutation of
ARHGEF6 has been associated with intellectual disability [
13]. These findings encouraged us to further study the possible role of
ARHGEF10 in ASD.
To explore the role of
ARHGEF10 in ASD and to further understand the impacts of
ARHGEF10 deletion on the molecular mechanisms of neurological function, we generated an
Arhgef10 knockout mouse model by deleting exons 4 and 5 of the mouse
Arhgef10 gene. The mice without the
Arhgef10 gene have normal fertility and body weight gain. The
Arhgef10 knockout mice were then subjected to tests measuring their startle responses, motor behavior, spatial learning, and social behavior. Behavioral changes may reflect a disturbance of neurotransmission. Previous studies have shown that the dysregulation of biosynthesis, transportation, and degradation of neurotransmitters could be associated with ASD [
14]. In addition, genetic association studies have identified some genes that encode the transporters or degradation enzymes of neurotransmitters as contributing to the risk of ASD [
15]. For instance, monoamine oxidase A (MAO-A), an enzyme that is important for the metabolism of serotonin and norepinephrine, is associated with ASD in a population-based study [
15]. Another population-based association study also indicates that changes in monoamine oxidase B (MAO-B) activity increase ASD risk in males [
16]. Based on those findings, we evaluated the levels of monoamines in
Arhgef10 knockout mice to understand the possible roles of
Arhgef10 in the serotonergic system.
In this study, we found that Arhgef10 knockout mice exhibited impaired social interaction and social recognition, representing the key characteristics of ASD. In addition, Arhgef10 knockout mice also displayed reduced anxiety-like and depression-like behaviors and increased locomotor activity. Additionally, serotonin, norepinephrine, and dopamine were elevated in the frontal cortex, hippocampus, and amygdala in Arhgef10 knockout mice. Moreover, MAO-A, a molecule that regulates the levels of certain neurotransmitters, was also reduced in Arhgef10 knockout mice. Behavioral studies and biochemical examination showed that ARHGEF10 not only plays a critical role in social behavior but also participates in the regulation of neurotransmitters.
Methods
Animals and experimental design
Arhgef10 knockout mice were generated by the deletion of exon 4 and exon 5 using the Cre-loxP site-specific knockout according to a method described previously [
8]. The strategy for generating
Arhgef10 knockout mice is shown in Additional file
1: Figure S1. In brief, exons 4 and 5 of ARHGEF10 in embryonic stem cells from the 129S1/Sv mouse strain were replaced with a construct containing ARHGEF10 exons 4 and 5 interposed between two loxP sites and a NEO cassette to produce Cre-induced homologous recombination.
Arhgef10 knockout mice were then backcrossed for at least ten generations with C57BL/6J mice. The
Arhgef10 knockout mice used in this study were produced by heterozygous breeding pairs in a trio breeding format. All mice were kept under standard temperature, humidity, and timed lighting conditions and provided with mouse chow and water ad libitum.
Arhgef10 knockout mice and their control littermates were housed in groups (3–5 mice per cage).
Behavioral tests were conducted during the light cycle (07:00–19:00) in a testing room next to the mouse housing room. Eight- to twelve-week-old male Arhgef10 KO mice and their WT littermates were used in this study. All behavioral tests were carried out with male mice. Animals in the same littermates were used in the same behavioral test, except for the plus maze, open field, and water maze tests. In these three tests, the intervals between tests were approximately 7 days each. The open field test was conducted first, followed by the plus maze and then the water maze. Mice were transported to the testing room and habituated for 30 min before behavioral testing. All animal experiments were approved by the Ethical Committee for the Animal Research of National Taiwan University.
Sample preparation
Mice were anesthetized with isoflurane and then decapitated. For Nissl staining, the brains of 10-week-old mice were removed after saline perfusion and post-fixed with 4% paraformaldehyde (PFA) overnight. Tissue samples for Western blots and high-performance liquid chromatography (HPLC) were prepared by dissecting four parts from the fresh brains of 10- to 12-week-old mice: frontal cortex, striatum, hippocampus, and amygdala. These tissues were then weighed and homogenized.
Nissl staining
Nissl staining was performed as previously described [
17]. In brief, 40-μm-thick frozen brain sections from WT and
Arhgef10 knockout mice were mounted on gelatin-coated slides and air-dried overnight. The slices were then placed directly into xylene for 3 min and then rehydrated with 100, 95, and 85% alcohol, followed by distilled water. The slices were stained with 0.02% crystal violet (Sigma-Aldrich, St. Louis, MO, USA) solution for 25–35 min and then rinsed quickly in distilled water. Finally, the slices were soaked in xylene and mounted with permanent mounting medium.
Open field test
Mice were placed in an open plastic chamber (40 × 40 × 38 cm), and locomotor activity was monitored for 1 h. The central region of the open field was defined as a region of 20 × 20 cm2, and all activity was measured by a 16 × 16 photobeam sensor connected to an automated tracking program (PAS-Open Field, San Diego Instruments). The number of beam breaks were used as an indicator of locomotor activity.
Elevated plus maze test
The elevated plus maze (EPM) apparatus (made of white Plexiglas) was elevated 40 cm above the floor and consisted of two open arms alternating with two closed arms. The open arms and closed arms were made of plastic and measured 25 cm long × 5 cm wide. The two closed arms were protected by 15-cm-high walls. Mice were placed in the EPM for 5 min, and the time spent in each arm and frequency of entry into each arm were recorded by Noldus EthoVision 3.0 (EthoVision®, Noldus Information Technology).
Tests for sociability and social novelty preference in a three-chamber apparatus
The tests for sociability and social novelty preference were performed in a three-chamber apparatus as previously described with minor modification [
18,
19]. The animals used here were all age- and sex-matched littermates; C57BL/6J mice were used as the stranger mice. All animals were habituated to the test chamber for 30 min 1 day before the behavioral test. The three-chamber apparatus was a 21 cm (height) × 25 cm (width) × 48 cm (length) black plastic box. Before the sociability test, the animal was free to explore the apparatus for 5 min. In the sociability test session, an unfamiliar same-sex mouse, designated as stranger 1, was placed in the plastic cylinder; an empty cylinder was placed in the opposite chamber. The cylinders (13 cm in height, 10 cm in diameter) were transparent, and their walls contained holes that allowed the mice to sniff each other. The test mouse was placed in the central chamber and allowed to freely explore these three chambers for 10 min. In the social novelty preference test, a second novel mouse, stranger 2, was placed in the opposite chamber, which was empty in the previous session. The test mouse was allowed to freely explore the chambers for 10 min in this session. The time spent in each chamber and the time spent sniffing or interacting with the stranger mouse were recorded. In addition, the number of entries into all chambers were analyzed using EthoVision.
Tail suspension test and forced swim test
The tail suspension test (TST) [
20] and the forced swim test (FST) are the most commonly used tests to evaluate depression-like behavior in rodents. In the TST, the animals were suspended by the tail with adhesive tape for 6 min. In the FST, the animals were placed in a plastic cylinder tank (30 cm in height × 15 cm in diameter) filled with 25 ± 2 °C water for 6 min. The complete TST and FST sessions were all videotaped for analysis. Immobility was defined as a lack of motion of the whole body at 2–6 min, when the mice ceased struggling and began passively floating, making only the movements necessary to keep the head above water.
Pre-pulse inhibition test
Pre-pulse inhibition (PPI) was tested in an acoustic startle chamber (SR Lab, San Diego Instruments, San Diego, CA). The chamber contained a Plexiglas cylinder (12.6 cm in length × 4.0 cm in internal diameter) fixed on a platform under which a piezoelectric accelerometer recorded and transduced the motion of the tube. The sensor transmitted the digitized signals to a computer interfaced with the startle apparatus. To provide a consistent acoustic environment and mask external noise, we maintained a continuous background of 65-dB white noise within each chamber throughout the PPI tests. The inter-trial interval was between 10 and 20 s. Each startle trial consisted of a single 120-dB white noise burst lasting 40 ms. Each PPI trial consisted of a pre-pulse (20 ms burst of white noise with an intensity of 70, 74, or 82 dB) followed by the startle stimulus 100 ms later (120 dB, 40 ms of white noise). Each of the three types of pre-pulse trials (70, 74, and 82 dB) was presented 10 times. The percentage of PPI was calculated according to the following formula: %PPI = (S − (P + S))/S × 100%, where P + S is the recorded response amplitude for pre-pulse plus startle pulse trials, and S is the recorded response amplitude for startle pulse-only trials.
Morris water maze test
Spatial learning was evaluated behaviorally using the Morris water maze. The maze consisted of a circular pool of water, 105 cm in diameter and 21 cm in height. The pool was divided into four equal quadrants: the target zone, two adjacent zones, and the opposite zone. The pool was filled with water at 22 ± 0.5 °C. A circular platform of 10 cm diameter was placed at the center of the target quadrant, and approximately 1.5 cm below the surface of the water. The maze was surrounded by four simple visual cues external to the maze. The movements of the mice were recorded with a video camera placed on the ceiling over the center of the maze, and the paths of the mice were analyzed using EthoVision. The training was started by acclimating the mice to the task environment with 1 day of free swimming in the pool without the platform. Each mouse underwent four consecutive training trials per day for 4 days. In each trial, the mice were placed along the edge of one quadrant of the maze; all quadrants except the target quadrant were used as starting locations. The duration from the time when the mouse entered the water until it climbed onto the platform was recorded as escape latency, and the mean latency to find the hidden platform was calculated for each individual mouse on each day. The maximum duration of each trial was 1 min, and the mice were removed from the pool by the experimenter after each trial. If the mice failed to find the platform within 1 min, they would be placed on the platform, where they were allowed to remain for 15 s. For the probe trial, 24 h after the final trial, the platform was removed. Each mouse was placed in the water maze at the edge of the former platform location and allowed to swim freely for 60 s. The total time that each mouse spent in each quadrant of the tank was recorded.
The levels of serotonin, norepinephrine, and dopamine in the frontal cortex, striatum, hippocampus, and amygdala were measured by HPLC analysis. Four different mouse brain areas were dissected from the cerebral hemisphere, weighed, and homogenized in 0.1 N perchloric acid by sonication. The homogenates were then centrifuged at 14,500×g for 30 min at 4 °C, and the supernatants were collected for HPLC analysis. HPLC–ECD (Sykam, Gilching, Germany) was used to evaluate levels of monoamines, including dopamine, 5-HT, and NE and their metabolites, in the samples (20 μl). The chromatogram peaks corresponding to the monoamines were identified by their retention times compared with the elution times of monoamine standards (Sigma, St. Louis, MO). The levels of dopamine, 5-HT and NE, were then estimated by the ratios of the peak heights to those of the internal standards and expressed as nanograms of neurotransmitter per gram of tissue.
Western blotting
Western blot analysis was performed to characterize ARHGEF10 (Proteintech Group, IL, USA; 11112-1-AP), GAPDH (Santa Cruz, Dallas, TX, USA; sc25778), β-actin (Sigma-Aldrich, St. Louis, MO, USA; A5316), MAO-A (Abcam, Cambridge, UK; ab126751), monoamine oxidase B (MAO-B) (Abcam, Cambridgeshire, UK; ab137778), dopamine ß-hydroxylase (DβH) (Abcam, Cambridge, UK; ab108384), and tryptophan hydroxylase (TPH) (Abcam, Cambridge, UK; ab52954) expression in the brains of WT and ARHGEF10 knockout mice. Mouse brain tissues were lysed in RIPA buffer (150 mM NaCl, 50 mM Tris–HCl, 1 mM EGTA, 1% Nonidet P-40, 0.25% deoxycholate, 1 mM sodium fluoride, 50 mM sodium orthovanadate) supplemented with Halt protease inhibitor cocktail (Thermo, IL, USA). Proteins were separated by SDS-PAGE, transferred to PVDF membranes, and blocked with 5% non-fat milk for 1 h. The membranes were then incubated with primary antibodies at 4 °C overnight followed by the appropriate secondary antibodies at room temperature for 1 h. The protein bands were visualized using enhanced chemiluminescence (ECL) (Millipore, MA, USA) reagent, and blot images were captured using a UVP imaging system with LabWorks Software (Upland, CA, USA).
Data analysis
Throughout the study, parametric analysis was performed by one- or two-way ANOVA, using a between-group factor of genotype and a within-group factor of each behavioral parameter. If the overall ANOVA showed a significant difference, Tukey’s test was used for post hoc comparisons. Statistical analysis was conducted using SPSS (IBM Inc., Somers, NY, USA) and GraphPad Prism (San Diego, CA, USA) software. Graphs of the data were also created using GraphPad Prism. The data were represented as the mean ± standard error.
Discussion
In a previous clinical study, we found a terminal deletion containing the gene
ARHGEF10 on chromosome 8 in autistic patients. The current study demonstrates that loss of function of
Arhgef10 leads to impairment of social behavior and altered mood-related behavior in mice. Importantly,
Arhgef10 knockout mice showed indifference toward the novel objects and stranger mice in the free social interaction assay, as well as a lack of preference between the novel mouse and the familiar mouse during the social recognition test. Overall,
Arhgef10 knockout mice displayed no preference between the empty chamber and stranger mouse 1 and showed no preference for social interaction with the novel mouse, stranger 2. These results indicate that ARHGEF10 deficiency might cause a deficit in social interaction and social recognition. Interestingly, although
Arhgef10 knockout mice stayed longer in the chamber with the first stranger mouse than the chamber with the second stranger mouse (Fig.
2d), there was no difference in the time of social interaction between the two chambers containing stranger mice (Fig.
2e). These results suggest that
Arhgef10 knockout mice lacked interest in interaction with stranger mice. Moreover,
Arhgef10 knockout mice exhibited a decreased tendency to explore the area with the second mouse. This decreased tendency to investigate the area with a novel stranger mouse may correlate with an aloof character typical of autism. The reduced social interaction and lack of preference for social novelty observed in the three-chamber social behavior tests support our speculation that the deletion of ARHGEF10 might partly be correlated with clinical manifestations in individuals with ASD.
Several published observations have shown that children with ASD are at a high risk of anxiety disorders [
21]. In this study, a reduced anxiety level was found in knockout mice, compared with WT mice, in the EPM test.
Arhgef10 knockout mice spent a longer duration in the open arms than WT mice, implying reduced anxiety-like behavior in
Arhgef10 knockout mice. Moreover, the reduced anxiety-like behavior in EPM could result from the decreased innate fear responses to elevated and open areas. Additionally,
Arhgef10 knockout mice entered the open arm more frequently than WT, which suggests
Arhgef10 deficiency did not affect the natural tendency to explore novel places. However, in the three-chamber test, the tendency to investigate the place with a novel stranger was reduced in
Arhgef10 knockout mice, indicating possible social withdrawal behavior. Although the finding of a lower level of anxiety in
Arhgef10 knockout mice is inconsistent with another reported autism-like mouse strains [
22,
23], the reduced anxiety in
Arhgef10 knockout mice indicates that social interaction defects were not caused by anxiety-like behavior. The reduction of anxiety-like behavior strengthened the correlation between
Arhgef10 and social behavior and suggests the involvement of
Arhgef10 in anxiety-related behavior.
We further evaluated specific depression-like behavior using the FST and the TST, to verify that social deficits in
Arhgef10 knockout mice are not caused by depression-related effects. Compared with WT mice,
Arhgef10 knockout mice exhibited lower immobility time in the FST and the TST than WT mice, suggesting that
Arhgef10 KO reduces depression-like behavior. These results revealed that ARHGEF10 is also involved in the regulation of depression-like behavior. The lower levels of anxiety-like and depression-like behavior further demonstrated that the social traits in
Arhgef10 knockout mice were not the results of mood-related behavioral abnormalities. In addition,
Arhgef10 knockout mice exhibited enhanced locomotor activity in a novel open field, which might be a reflection of the hyperactivity that is often observed clinically in autistic patients. The features of hyperactivity have also been reported in other animal models of autism [
24‐
26]. Although
Arhgef10 knockout mice displayed characteristic hyperactivity, the number of entries into each compartment was not increased in the three-chamber tests, indicating that the reduced social behavior was not a result of hyperactivity. When the results of social impairment and hyperactivity are taken together, the ASD-associated social inability observed in the
Arhgef10 deficiency mice appears not to result from low levels of spontaneous activity. The reduced anxiety-like behavior and hyperactivity phenotype observed in
Arhgef10 knockout mice may be a sign of increased impulsive behavior. We thus further evaluated impulsive behavior using the electro-foot shock aversive water drinking test (EFSDT) [
27].
Arhgef10 knockout mice displayed no obvious impulsive behaviors in comparison with WT mice (Additional file
2: Figure S2-2).
Although some studies suggest that autism features sensorimotor deficits [
28], we did not find any PPI deficit in
Arhgef10 knockout mice. No abnormalities were found in the startle response or sensory gating, indicating that the
Arhgef10 knockout does not affect the sensory gating system in mice. Moreover, escape latencies were comparable between WT and
Arhgef10 knockout mice in the Morris water maze, suggesting that the deletion of ARHGEF10 does not impair spatial-learning-based behavior.
Social behavior can be affected by many behavioral factors, including anxiety levels, spontaneous activity, depression, sensory perception, memory, and cognition. Arhgef10 knockout mice displayed normal sensorimotor responses in the PPI test and normal spatial learning behavior in the water maze. Moreover, Arhgef10 knockout mice did not exhibit depression-like or anxiety-like behavior, suggesting that the social deficit was not confounded by these factors. Our study demonstrated that the loss of function of ARHGEF10 mainly caused social impairment in mice.
In this study, changes in neurotransmitter levels were observed as well. Serotonin and norepinephrine levels were significantly increased in the frontal cortex and amygdala of
Arhgef10 knockout mice. The amygdala has been implicated in various functions concerning emotion and social behavior. Several studies in primates and other mammalian species have implicated the amygdala as an important contributor to social behavior [
29,
30]. Neonatal lesions of the amygdala in rats results in a social behavior deficit, and these rats exhibit autistic-like symptoms [
31]. Furthermore, an early bilateral lesion of the amygdala in rhesus monkeys leads to reduced social interactions with other monkeys [
32]. On the basis of the biological function of the amygdala, it is proposed to be one of the main regulatory brain regions driving the pathology of ASD [
33]. Furthermore, the frontal cortex is the region responsible for emotion control and decision making and has been reported to be associated with the cognitive and anatomical abnormalities in patients with autism [
34]. Based on the distinctive roles of the frontal cortex and amygdala in autism, the neurochemical changes in these brain regions may be linked to the observed phenomena in
Arhgef10 knockout mice. The neurochemical study of autism began with a report of changes in serotonin [
35]. An elevated serotonin level in whole blood is a well-characterized biomarker in autism research [
6]. Additionally, increased serotonin synthesis capacity has been reported in autistic children, suggesting that serotonin may play an important role in the development of autism [
36]. The elevated serotonin levels found in the brains of
Arhgef10 knockout mice correlate with attenuated depression-like behaviors observed in the FST and the TST. These findings also support the role of the serotonin system in some autism patient symptoms. Apart from serotonin, norepinephrine is also an important monoamine associated with neuronal functions and levels of anxiety, arousal, and stress sensitivity, and many of these functions are also impaired in ASD. Evidence for involvement of norepinephrine in ASD comes from a report indicating that patients with autism and their families showed elevated plasma levels of NE and low enzyme activity [
37]. The Angelman syndrome mouse model, which exhibits autistic-like behavior, shows increased norepinephrine and serotonin levels in different brain regions [
23]. In addition, the elevated serotonin level in brain found in
Arhgef10 knockout mice can also explain their attenuated depression-like behaviors in the FST and the TST. The changes of serotonin and norepinephrine in the frontal cortex and amygdala indicate that ARHGEF10 might play a role in linking the phenotype of social impairment to the balance of monoamines. Therefore, the loss of homeostasis of serotonin and norepinephrine in the frontal cortex and amygdala of the
Arhgef10 knockout mice might be the basis of their autistic-like behaviors.
We sought to further understand the underlying mechanisms that contribute to the increased level of amine neurotransmitters. Therefore, we examined MAO-A and MOA-B, which are the key enzymes responsible for catalyzing the metabolism of monoamines, including serotonin and norepinephrine. In addition, the enzymes dopamine β-hydroxylase (DBH), and tryptophan hydroxylase (TPH), which are responsible for the synthesis of norepinephrine and serotonin, respectively, were also examined using Western blot. MAO-A was decreased in the frontal cortex and amygdala in
Arhgef10 knockout mice. However, the levels of DBH and TPH were comparable to those of WT littermates, indicating that the synthesis of the corresponding neurotransmitters was not affected by the
Arhgef10 knockout. The lower level of MAO-A may reduce the capacity for serotonin and NE degradation. The decreased MAO-A in
Arhgef10 knockout mice provides a possible explanation for the higher levels of serotonin and norepinephrine in the frontal cortex and amygdala. Recently, a study demonstrated that patients with autism had decreased MAO-A activity in the frontal cortex and cerebellum [
38]. Moreover, mice lacking MAO-A displayed increased serotonin and norepinephrine levels and aggressive behavior [
39]. Another knockout mouse lacking MAO-A/B also exhibited elevated serotonin levels and autistic-like features [
40]. These previous reports support our findings in
Arhgef10 knockout mice, which implied that the reduced MAO-A levels and elevated amine levels might be the causes of social impairment. Although the phenotypes might result from changes in amine levels in the frontal cortex and amygdala, further study of the neurotransmitters and related molecules that are regulated by ARHGEF10 is required to clarify the underlying mechanism.