The voltage-gated Cav Ca2+ channel subunit α2δ-4 is required for locomotor behavior and sensorimotor gating in mice

Voltage-gated Cav Ca2+ channels are critical for the development and mature function of the nervous system. Variants in the CACNA2D4 gene encoding the α2δ-4 auxiliary subunit of these channels are associated with neuropsychiatric and neurodevelopmental disorders. α2δ-4 is prominently expressed in the retina and is crucial for vision, but extra-retinal functions of α2δ-4 have not been investigated. Here, we sought to fill this gap by analyzing the behavioral phenotypes of α2δ-4 knockout (KO) mice. α2δ-4 KO mice (both males and females) exhibited significant impairments in prepulse inhibition that were unlikely to result from the modestly elevated auditory brainstem response thresholds. Whereas α2δ-4 KO mice of both sexes were hyperactive in various assays, only females showed impaired motor learning/coordination in the rotarod assay. Female but not male α2δ-4 KO mice exhibited anxiolytic and anti-depressive behaviors in the elevated plus maze and tail suspension tests, respectively. Our results reveal an unexpected role for α2δ-4 in cognitive and motor function and identify α2δ-4 KO mice as a novel model for studying the pathophysiology associated with CACNA2D4 variants.


Introduction
Voltage-gated Ca v Ca 2+ channels mediate Ca 2+ signals that initiate a vast array of signaling events including gene transcription, protein phosphorylation, and neurotransmitter release. The main properties of these channels are determined by a pore-forming  1 subunit, while auxiliary  and  2  subunits regulate the trafficking and some functional aspects of these channels (1). These subunits are encoded by four genes each (2), with additional functional diversity conferred by extensive alternative splicing (3). The physiological importance of Ca v channels is reflected in the numerous diseases that linked to mutations in the genes encoding the Ca v subunits which include migraine, ataxia, and disorders of vision and hearing (4,5).
In recent years, variants in Ca v encoding genes have been consistently identified in genome-wide association studies of neuropsychiatric disorders. One of the most prominent of such studies analyzed single-nucleotide polymorphisms (SNPs) in ~60,000 individuals and uncovered CACNA1C, the gene encoding Ca v 1.2, as a major risk gene for schizophrenia, bipolar disorder, major depressive disorder, autism spectrum disorder, and attention deficit hyperactivity disorder (ADHD). Pathway analysis further revealed an association of other Ca v -encoding genes with these disorders, including CACNA2D4 that encodes the  2 -4 subunit (6). This result was rather unexpected given that  2 -4 was thought to be expressed primarily in the retina, where it associates with the Ca v 1.4 channel and regulates the structure and function of photoreceptor synapses (7)(8)(9).  2  is an extracellular protein that regulates the cell-surface trafficking of Ca v channels (10), but may have additional roles. For example,  2 -1 binding to thromobospondins promotes synapse formation in a manner that is inhibited by the analgesic and anti-convulsant drug, gabapentin (11). At the Drosophila neuromuscular junction,  2 -3 is required for proper synapse morphogenesis-a process that does not involve its association with the Ca v 2.1 channel (12). In the retina, the formation of photoreceptor synapses involves the role of  2 -4 as a Ca v 1.4 subunit and as a mediator of trans-synaptic interactions of the cell adhesion molecule, ELFN-1, with postsynaptic glutamate receptors (9).
Despite the association of  2 -4 with neuropsychiatric disease, how  2 -4 contributes to cognitive and affective functions is unknown. To address this question, we examined the behavioral phenotypes of  2 -4 knockout (KO) mice (8).

Animals
All procedures using animals were approved by the University of Iowa Institutional Animal Care and Use Committee. The  2 -4 KO mouse line was bred on a C57/Bl6 background and characterized previously (8). Separate cohorts of males (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25) week old, n= 10 wild-type (WT), n= 11 KO) and females (11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22) week old, n= 11 WT, n= 11 KO) were analyzed. Before beginning handling and testing, mice were ear punched for identification. All mice were housed in groups of 2-3 animals per cage for the duration of the handling and testing periods with food and water ad libitum (describe details of the cages). The room in which the mice were housed was maintained on a consistent light cycle with lights on at 0900 and lights off at 2100 and testing took place between 0800 to 1300. Males and females were tested in separate cohorts at different time points to prevent pheromones on the testing apparatus from impacting results. Mice were generally acclimatized for 30 min in the room in which the assay was conducted prior to initiating the test.
A full week was taken between every test to reduce the impact of stress from previous tests on the next result. The order of testing was designed to minimize the impact of preceding assays by performing those with the least stressful tasks first.
Recordings were conducted on both ears of all animals on a heating pad using electrodes placed subcutaneously in the vertex and underneath the left or right ear. Clicks were square pulses 100 ms in duration, and tone bursts were 3 ms in length at distinct 8-, 16-, and 32 kHz frequencies.
ABRs were measured using BioSigRZ software (Tucker-Davis Technologies), with stimulus levels adjusted in 5-dB increments between 25 and 100 dB SPLs in both ears. Electrical signals were averaged over 512 repetitions and ABR threshold was defined as the lowest sound level at which a reproducible waveform was measured.

Elevated Plus Maze
The testing apparatus consisted of a plus-shaped maze elevated 40 cm above the floor. Two

Light Dark Box
The testing apparatus consisted of a chamber divided into a light and dark compartment equipped with infrared beam tracking (Med Associates). The apparatus was divided into 2 chambers with a gap in the wall between them. Mice were tested using a very bright light in the light chamber (27,000 lux). Mice were moved from the home cage to the light side of the apparatus facing away from the dark chamber. The animals were allowed to freely explore and move between the chambers for 30 min and the animals' movements were documented in sequential 5 min intervals via infra-red tracking. Time spent in either compartment was analyzed by Activity Monitor software.

Open Field Test
The testing apparatus consisted of an open square chamber with walls of 40 cm height and width.
Illumination intensity in the central square was approximately 500 lux. Mice were moved from the home cage to the center of the open chamber. The animals were allowed to freely explore the chamber for 10 minutes. Animal behavior was evaluated using video recording and Anymaze software. Relative time spent in the inner and outer portion of the box were taken as a measure of the animals' anxiety-like behavior. Total distance traveled over the 10 minutes was taken as a measure of the animals' basal activity level.

Prepulse Inhibition
The testing apparatus consisted of a startle response box (SR-LAB from San Diego Instruments).
A restraint chamber consisted of a clear plastic tube from which the tremble response of the animal could be measured via an accelerometer underneath the chamber. Animals were placed in the restraint chamber and allowed to acclimate to the chamber for 10 min with a consistent background white noise level of 65 dB which was present for the entire experiment. The 25-min testing period was divided into 3 blocks each consisting of 6 or 60 trials. All trials were presented with a randomly spaced intertrial interval ranging from 7 to 15 seconds. The first block consisted of 6 pulse trials at 120 dB. The second block contained 12 of each of the following trial types: standard pulse at 120 dB, no stimulation, prepulse of +4 dB above background, prepulse of +8 dB, and prepulse of +16 dB. The third block consisted of 6 pulse trials at 120 dB. Startle response amplitudes (in mV) were measured in SR-LAB software and %PPI measured as (startle response for pulse alone -startle response for pulse with pre-pulse) / startle response for pulse alone) X 100.

Rotarod Test
The testing apparatus consisted of a rotating spindle 3.0 cm in diameter that will increase in speed over the course of the trial (Rotamex 5). Mice were trained for 2 consecutive days with 3 testing trials per mouse each day separated by at least 30 min. For the testing trial, the speed of rotation was increased by 1.2 rpm every 20 s and the latency to fall was recorded. The 6 testing trials were averaged for each mouse.

Forced Swim Test
The testing apparatus consisted of a 2-liter beaker filled with 1200 ml of water at room temperature. Mice were placed in the water and monitored for 6 min, then were dried and placed in a recovery cage with a cage warmer. Time spent immobile was recorded, with immobility defined as lack of motion in the hind legs except necessary movement to balance and keep the head above the water.

Tail Suspension Test
The testing apparatus consisted of a metal bar suspended 30-40 cm above the table. Tails of the mice were wrapped in adhesive tape within the last 1 cm of the tail. A clear plastic tube was placed around the animal's tail to prevent climbing up the tail and onto the bar. Time spent immobile was recorded, with immobility defined as lack of attempting to move their limbs as described previously (14). (denoted in all graphs as follows: *p < 0.05; **p < 0.01; ***p < 0.001).

Results and Discussion
 2 -4 KO mice were born at normal Mendelian ratios and did not exhibit any overt behavioral phenotypes other than hyperactivity. The control wild-type (WT) strain corresponded to C57Bl6 strain on which the  2 -4 KO mice were bred for at least 10 generations. Cohorts of male and female mice were analyzed separately, and there were no differences in body weight of the WT and  2 -4 KO mice used in this study (Table 1). Prepulse inhibition is impaired in  2 -4 KO mice. Sensorimotor gating is a form of pre-attentive processing that is commonly studied in humans and animals using prepulse inhibition (PPI). In this test, a response to a strong acoustic stimulus is generally diminished when it is preceded by a subthreshold stimulus (15). Reductions in PPI are thought to reflect impairments in working memory in individuals diagnosed with schizophrenia, bipolar disorder, and post-traumatic stress disorder and in animal models of these conditions (16,17). Because of the association of Ca vencoding genes with these disorders (6), we tested whether  2 -4 KO mice exhibit deficits in PPI.
WT and  2 -4 KO were tested for startle responses to a 120 dB acoustic stimulus that was administered alone or after a prepulse stimulus of 4, 8, or 16 dB, and PPI was expressed as the % change in the response amplitude due to the prepulse (%PPI, Fig.1A,B). In this assay, there was a significant main effect of both sex (F 1, 39 = 7.876, p < 0.01) and genotype (F 1, 39 = 10.26, p < 0.01), but no interaction between these variables (F 1,39 = 0.0028, p = 0.958; Fig.1B). PPI was significantly lower for  2 -4 KO than for WT mice in the cohort of females (p < 0.05 for both 8 and 16 dB prepulse) and males (p < 0.05 for 8 dB, p < 0.01 for 16 dB prepulse). In some mouse strains, relatively low levels of PPI correlate with low basal startle amplitudes (17). However, basal startle amplitudes were significantly higher in  2 -4 KO mice than in WT mice (F 1,39 = 55.50, p < 0.001; Fig.1C). Some studies have shown that patients with schizophrenia have an impaired habituation to the startle pulse (18), which would manifest as a difference in startle response to the 120 dB-stimulus administered without the prepulse (blocks 1-3, Fig.1A). There was no effect of genotype on this parameter (F 2, 78 = 1.580, p = 0.213). Collectively, these results show that  2 -4 KO mice exhibit impaired PPI without alterations in habituation.   Fig.2). Nevertheless, all  2 -4 KO mice displayed functional hearing above 60 dB, the range used in the PPI assays, which argues against the possibility that the reduced PPI of the  2 -4 KO mice resulted from hearing loss. respond to the aversive stimuli of the OFT and EPM since these mice have normal vision in daylight but not dim light conditions (8). As a further test, we performed the light dark box assay in which avoidance of a chamber with a bright light stimulus is taken as a measure of anxiety-like behavior. The  2 -4 KO spent more time in the lighted chamber than WT mice (η 2 = 0.100, p < 0.05 by Kruskal-Wallis; Fig.3E,F). The light intensity used in the lighted chamber was 27,000 lux, which is well above the visual threshold for  2 -4 KO mice (8). Taken together, results from the EPM and LD assays support an anxiolytic phenotype in  2 -4 KO mice.   Abnormal motor behaviors are a common feature of neurodevelopmental disorders including ASD and ADHD (28)(29)(30). Thereofre, we tested motor function of  2 -4 KO mice in the rotarod assay. In this assay, the mice are placed on a rotating cylinder that is gradually accelerated and the length of time the animal can stay on the cylinder is taken as a measure of balance, coordination, and motor planning (31). The latency to fall was shorter for  2 -4 KO than for WT mice (F 1, 39 = 6.457, p < 0.05; Fig.5A). To further assess motor phenotypes in the  2 -4 KO mice, we analyzed data in the OFT, EPM, and LD assays for aberrant locomotion. In each case, the total distance traveled by  2 -4 KO (both males and females) mice was significantly greater than for WT mice (OFT η 2 = 0.266, p < 0.001; EPM: F 1, 34 = 16.09, p < 0.001; LD: η 2 = 0.490, p < 0.001; Fig.5B-D). These results show that  2 -4 KO mice exhibit signs of hyperactivity and sex-dependent impairment in motor coordination. Our results show that  2 -4 KO mice exhibit a pattern of cognitive, affective, and motor behaviors that resemble those in neuropsychiatric disorders that are linked to variants in Ca vencoding genes (6). Because  2  proteins enhance the cell-surface trafficking of Ca v channels (5), the phenotypes of  2 -4 KO mice could result from loss-of function of Ca v channels in key brain regions such as the hippocampus, cerebral cortex, and cerebellum. A caveat is that compared to other  2  subunits,  2 -4 is nearly undetectable in these brain regions (32).
Moreover, mice lacking the expression of the major Ca v subtype implicated in neuropsychiatric/neurodevelopmental disorders (i.e., Ca v 1.2) exhibit increased anxiety-like behavior  2 -4 KO mice (33) whereas  2 -4 KO mice present with an anxiolytic phenotype in the EPM and LD assays (Fig.3C-F). However, deletion of Ca v 1.2 in the prefrontal cortex causes antidepressant behavior in the TST (34), similar to that of  2 -4 KO mice in our study (Fig.4A-C).
 2 -4 could be expressed in a small subset of neurons implicated in these behaviors, thus escaping detection by quantitative PCR in homogenates of specific brain regions (32).
Alternatively,  2 -4 could undergo pathological upregulation. For example,  2 -4 expression is increased in hippocampus of humans and mice following epileptic seizures (35). Given that  2  proteins regulate synapse formation in part through trans-synaptic interactions with proteins other than Ca v channels (9,11,12,36), aberrant expression of  2 -4 could cause defects in neuronal connectivity that modify cognitive/affective behaviors.
Although blind under dim-light conditions,  2 -4 KO mice are expected to have normal vision under the lighting conditions used in our study (8,9). To date, alterations in cognitive and/or affective function in individuals diagnosed with CACNA2D4-related vision impairment have not been reported. However, the etiology of most neuropsychiatric disorders is complex and likely involves hundreds to thousands of risk alleles distributed across the genome (37). Our findings that  2 -4 KO mice exhibit defects in PPI, motor function, and anxiety/depression-related behaviors validate the importance of CACNA2D4 as one such risk allele and that studies of the extra-retinal functions of  2 -4 warrant further study.