Extensive characterization of a Williams syndrome murine model shows Gtf2ird1‐mediated rescue of select sensorimotor tasks, but no effect on enhanced social behavior

Abstract Williams syndrome is a rare neurodevelopmental disorder exhibiting cognitive and behavioral abnormalities, including increased social motivation, risk of anxiety and specific phobias along with perturbed motor function. Williams syndrome is caused by a microdeletion of 26–28 genes on chromosome 7, including GTF2IRD1, which encodes a transcription factor suggested to play a role in the behavioral profile of Williams syndrome. Duplications of the full region also lead to frequent autism diagnosis, social phobias and language delay. Thus, genes in the region appear to regulate social motivation in a dose‐sensitive manner. A “complete deletion” mouse, heterozygously eliminating the syntenic Williams syndrome region, has been deeply characterized for cardiac phenotypes, but direct measures of social motivation have not been assessed. Furthermore, the role of Gtf2ird1 in these behaviors has not been addressed in a relevant genetic context. Here, we have generated a mouse overexpressing Gtf2ird1, which can be used both to model duplication of this gene alone and to rescue Gtf2ird1 expression in the complete deletion mice. Using a comprehensive behavioral pipeline and direct measures of social motivation, we provide evidence that the Williams syndrome critical region regulates social motivation along with motor and anxiety phenotypes, but that Gtf2ird1 complementation is not sufficient to rescue most of these traits, and duplication does not decrease social motivation. However, Gtf2ird1 complementation does rescue light‐aversive behavior and performance on select sensorimotor tasks, perhaps indicating a role for this gene in sensory processing or integration.

motivation. However, Gtf2ird1 complementation does rescue light-aversive behavior and performance on select sensorimotor tasks, perhaps indicating a role for this gene in sensory processing or integration.  7), showing the importance of gene dosage in the pathophysiology of these disorders. 1,30,34 Both syndromes result in altered craniofacial features, cardiac issues, motor coordination deficits and behavioral challenges. 13,20 Many people with WS exhibit hypersociability and tend to approach strangers with little apprehension, although the lack of social anxiety does not preclude a more generalized anxiety and occasional extreme phobias, both of which are more prevalent in individuals with WS than the general population. 2,6,15 Unfortunately, the underlying mechanisms of these behavioral differences are not well understood and thus no targeted treatments exist to help individuals with WS navigate the expectations of society, similar to the struggle autistic individuals face. However, unlike the complex etiology of idiopathic autism, the discrete genetic foundation of WS provides a unique opportunity to uncover these mechanisms, as a relatively small deletion leads to such a recognizable behavioral profile.
As WS and Dup7 are rare, understanding the complex etiology and circuit pathology underlying behavioral phenotypes in humans, or with human brain samples, is challenging. While cellular phenotypes can be investigated in iPSC models, 1 animal models are still required to uncover the link between gene dosage and behavioral phenotypes.
Fortunately, the WSCR is syntenic in mice, and a complete deletion (CD) mouse model eliminating most of the region has been developed that recapitulates many features of WS. 29 An initial survey of various features in the CD mouse line discovered mild cardiac deficits, craniofacial anomalies and some alterations in behavior. Specifically, the authors reported deficits in motor performance in a rotarod task and a decreased habituation to social stimuli in an open field social interaction test of sociability, or the tendency to initiate social contact. 21,29 However, the previous measures on a mostly C57BL/6J background ($F5) showing increased social interest were only conducted in males and they did not utilize the classic 3-chamber social approach task. In our hands, the 3-chamber social approach task showed no difference in CD mice (albeit on a FVB/AntJ Â C57BL/6J F1 hybrid background) 11 ; the FVB strain generally shows less social approach than C57BL/6J, 24 suggesting this background may be less sensitive for CD social phenotyping.
Furthermore, social motivation, the amount of work an animal is willing to do to engage with a conspecific, has not been directly measured. Likewise, while motor learning has been assessed on the rotarod apparatus, 29 less work has characterized motor strength or coordination generally and the results were not consistent on the hybrid background with slightly different test parameters. 11 Finally, anxiety has also been a difficult domain to assess consistently in mice, as transient emotionality can affect the results. 26 For example, similar mouse models deleting a single gene in the WSCR, Gtf2ird1, show opposite results in anxiety-like behavior. 28,32 Overall, a deeper phenotyping of these domains would be of use, especially to provide a foundation for studies that address the effects of WSCR copy number variation at the level of mechanisms or circuits.
Prior to the development of the CD line modeling the full deletion, single gene deletions were the most common approach in trying to elucidate function. Gtf2ird1 is one such gene that has been implicated in a variety of hallmark WS phenotypes, from craniofacial to cognitive and behavioral differences. Gtf2ird1 is often implicated alongside its neighbor and family member, Gtf2i. As these genes occur in tandem in the WSCR and are rarely found separately affected by atypical deletions, it is difficult to isolate their effects using human studies alone. While both genes are conserved in the mouse genome, there has been more difficulty with reliably producing a Gtf2ird1 knockout animal. Alternative and frame-shifted start codons allow truncated versions of the protein to be expressed, even preserving much of its function outside of its negative autoregulation. 12

Verifying
Gtf2ird1 expression, or lack thereof, was also unreliable prior to the relatively recent development of effective antibodies.
To avoid the trouble of deleting this elusive gene, we adopted a different strategy to gauge the influence of Gtf2ird1 on relevant phenotypes; we designed a study to assess the impact of Gtf2ird1 while also providing an extensive characterization of the CD model, as both of these contributions would benefit our understanding of the WSCR.
Thus, we present a novel Gtf2ird1 transgenic expression line, which we use to thoroughly assess the role of Gtf2ird1. We test the hypothesis that Gtf2ird1 plays a dose-dependent role in the cognitive and behavioral symptoms of WS and concurrently examined the effects of Gtf2ird1 overexpression on a mostly C57BL/6J background and in the presence of WSCR deletion. We used a comprehensive battery of tasks designed to elucidate the contributions of Gtf2ird1 to WS-relevant phenotypes. Simultaneously, using the same extensive suite of behavioral measures, we provide a detailed assessment of the CD mouse, providing key information on additional phenotypes related to motor, anxiety, fear and social behaviors, broadening the initial characterization.
In these studies, we replicate and extend the previously reported social differences in the CD mice, 29 showing enhanced social approach and motivation, in addition to sensorimotor differences and greater avoidance behavior in some anxiety-related tasks. Finally, we rule out Gtf2ird1 as being the sole mediator of the social changes, as duplication of this gene did not decrease these behaviors, nor did its complementation of the CD rescue any notable social phenotypes.
However, it does appear to mediate aspects of light-induced anxietyrelated behaviors and sensorimotor coordination, as complementation can ameliorate the deficits observed in the CD mice, suggesting a role for Gtf2ird1 in sensorimotor processing.

| Gtf2ird1 transgenic mouse creation
We selected a bacterial artificial chromosome (BAC) clone (RP24-508D22) which contained the entirety of the 100 kb Gtf2ird1 gene, and 89 kb (60 kb at the 5 0 end and 28 kb at the 3 0 ) of flanking regulatory sequence (e.g., the Gtf2ird1 promoter, etc.), but no additional intact genes or their promoters. This was then recombineered using standard methods to insert an HA tag in-frame directly before the stop codon of the beta isoforms. 7 (TG-Gtf2ird1-HA or simply TG) were created by injecting this modified BAC into C57BL/6NTac mouse oocytes and transplanting these eggs into pseudopregnant surrogates to carry them to term. Transgene-specific primers (BgenoF3-CAACATTCCCAAGCGCAAGAG and Bge-noR3-GATAACTGATCGCGGCCAGC, which produce a 440 bp product in TG animals and no product in WT animals) were used for identification of TG founder animals. BAC copy number was determined to be 2-4. Multiple founders were evaluated to confirm transgenic RNA production by RT-PCR, and a single line was taken forward for evaluation.
Lines were backcrossed to C57BL6/J for over four generations prior to commencing experiments.

| Husbandry
All mice used in this study were maintained and bred in the vivarium at Washington University in St. Louis on a 12/12 h light/dark cycle with food and water provided freely. Three distinct mouse lines were used: C57BL/6J (WT; RRID:IMSR_JAX:000664), the CD mouse modeling deletion of the WSCR 29 and a novel transgenic line (TG) overexpressing Gtf2ird1 with an HA tag. CD and TG lines were maintained as heterozygotes by crossing to WT animals. Heterozygous CD and TG mice were crossed to produce behavioral cohorts containing WT, TG, CD and TG/CD littermates to best compare across genotypes. Animals were group housed by genotype and sex at weaning. Tissue was collected from pups for initial genotyping and again after death to verify genotype via PCR amplification.

| Molecular validation
Molecular analysis to assess RNA and protein levels via RT-qPCR and Western blotting was performed as previously described. 12 Briefly, whole brains were collected from pups $E13.5 for initial characterization of the novel line and just prior to weaning at postnatal days 19 or 20 (P19-P20) for validation of the crosses. Brains were immediately homogenized in Pierce RIPA buffer (ThermoFisher Scientific) with 1Â protease inhibitor (Roche) and RNase inhibitors RNasin (Promega) and SUPERaseIn (ThermoFisher Scientific). Two hundred and fifty microliters of the homogenate was transferred to 750 μl of TRIzol LS, then frozen at À80 C until RNA extraction and subsequent qPCR. Please see Kopp et al. 12 for detailed methods of the complete RT-qPCR protocol used to quantify Gtf2ird1, Gtf2i and Gapdh. The rest of the whole brain lysate was spun down to separate the protein fraction (supernatant) from the rest of the cell contents (pellet). The supernatant was kept in 2-3 aliquots at À80 C until protein quantification via Pierce BCA assay (ThermoFisher Scientific) and Western blotting.
RNA expression and protein levels were assessed relative to Gapdh using primers and antibodies described previously. 12 For Western blotting, 50 μg of protein were loaded onto 4%-15% Mini-PROTEAN TGX protean gels (Bio-Rad). Transfer was carried out using the Trans-Blot Turbo Transfer System (Bio-Rad). The membrane was cut $75 kD to simultaneously probe for GTF2IRD1 (Rabbit anti-GTF2IRD1, 1:500, Novus, NBP1-91973) and a GAPDH antibody as a loading control (Mouse anti-GAPDH, 1:10,000, Sigma Aldrich, G8795). All further incubation and imaging steps are detailed in Kopp et al. 12 Expression of the transgenic allele was also verified by a Western blot using an antibody to the HA tag (Mouse anti-HA, 1:1000, Thermo Scientific, 26183, not shown). As the HA tag is included only on some transcripts of Gtf2ird1 because of alternative splicing of the last exon, quantification focused on the pan-GTF2IRD1 antibody validated in Kopp et al. 12

| Behavioral testing
For behavioral analysis, three separate cohorts of mice were used to assess a variety of characteristics (Table 1). All tasks were run by female experimenters during the light phase. Tasks within each behavioral battery were ordered from least to most stressful to minimize the effect one task had on subsequent tasks. Adolescent and adult mice were handled for 5 days prior to starting the first behavioral task and the tails of mice in Cohorts 1 and 2 were marked with a non-toxic, permanent marker during weight collection and regularly thereafter to easily distinguish mice during testing. Males were run before female animals to avoid olfactory cue influence on behavior. Testing orders were randomly counterbalanced for group across apparatuses and trials.
Unless otherwise indicated, all equipment was cleaned between animals with a 0.02% chlorhexidine diacetate solution (Nolvasan, Zoetis).

| Open field
To assess activity levels and approach-avoidance behavior, we used the open field task adapted from our previously published methods. 3 Briefly, mice were placed in a 50 Â 50 Â 45 cm clear acrylic enclosure under red light at 9 lux, within a sound-and scent-attenuated white opaque box (70.5 Â 50.5 Â 60 cm) to minimize external stimuli, and allowed to freely explore for 60 min. Any-Maze software (Stoelting, Co) tracked the movement via the body center, beginning when the doors to the chamber were closed via video captured with an overhead CCTV camera. A center zone was designated as the middle 50% of the total chamber area. Movement of the animal throughout the arena was quantified as distance traveled, and as time spent in and entries into the center and perimeter zones. The center zone was designated as the inner 50% of the open field area, and the perimeter was the outer 50%.

| Open field social approach
Replicating previously published methods, 27

| Elevated plus maze
Anxiety-like behaviors were tested using the elevated plus maze (EPM) as previously described. 11 Briefly, mice were placed in the center of the apparatus, which contained two open and two closed arms, and allowed to explore for 5 min in the dark. This was repeated for two more days. Outcomes were averaged across days for analysis. Trials were recorded under infrared illumination with an overhead camera using Ethovision software (Noldus Information Technology) to track movement of the animal in the apparatus.

| Marble burying
The marble burying task was used to assess compulsive digging behavior, adapted from previous methods. 16 As in prior work, the mice were introduced to a novel, transparent enclosure (47.6 Â 25.4 Â 20.6 cm, contained within a sound-and scentattenuated white opaque box (70.5 Â 50.5 Â 60 cm) to minimize external stimuli) with 20 evenly spaced, clear marbles on clean, novel, autoclaved aspen bedding. Animals were allowed to explore freely for 30 min. After the animals were removed, two independent scorers recorded the number of marbles buried (defined as at least two-thirds covered with bedding). These scores were averaged for analysis.
T A B L E 1 Behavioral cohort sample size and task order.  Between mice, the marbles were cleaned with 70% ethanol. For this study, we also tracked the animals' movement in the apparatus via Any-maze tracking software and quantified distance traveled for overall activity levels and time spent in the center 50% of the arena.

| Open field novel object exploration
Adapting previously published methods, 29  Mice were placed in the apparatus to explore freely for 20 min while movement was recorded and tracked using Any-Maze software. A 2 cm investigation zone was defined around the object, in addition to center and perimeter areas of the arena.

| Social motivation operant conditioning
Social motivation, or how hard an animal will work for access to a social partner, was assessed using our social motivation operant assay.
Our 16-day paradigm allows for assessment of both social reward seeking and social orienting, two components of social motivation. 4 We followed the procedure outlined in our previous work. 3,18 Briefly, an operant conditioning chamber was modified to include a door that raised in response to a nosepoke in the active hole to provide 12 s of access to a novel sex-and age-matched partner stimulus mouse. To assess social reward seeking, the number of active (i.e., elicits a reward) versus inactive nosepokes were quantified. To assess social orienting, the behavior of the animal was tracked using Ethovision Mice that had at least 40 active nosepokes, 75% accuracy (active: inactive) and 65% successful rewards (interactions during rewards) were considered to have met conditioning criteria and progressed to a fixed ratio of 3 (FR3), where 3 nosepokes were required to receive the reward. After 3 days of FR3 (or 10 days of FR1 for mice who failed to reach criteria), mice were tested in a progressive ratio of 3 (PR3), where the first reward was provided after 3 active nosepokes and each subsequent reward required 3 additional nosepokes to obtain. The breakpoint was measured as the number of rewards a mouse was able to acquire before 30 min of nosepoke inactivity.

| Conditioned fear
To assess associative and anxiety-related memory, mice were tested in a conditioned fear task over 3 days following previously published methods 3,12,19 using the Actimetrics conditioning system (Wilmette, IL, United States). Briefly, following pairing of a tone and context with a 1.0 mA footshock on Day 1, all mice were tested for contextual fear memory on Day 2 and cued fear memory in response to the tone only on Day 3. Shock sensitivity was evaluated as we previously described. 12 4 | COHORT 2

| Sensorimotor battery
Adult mice were evaluated with a battery of sensorimotor measures to assess motor initiation, balance, strength and coordination using previously published methods. 3,12 The battery included evaluation of walk initiation, balance (Ledge and Platform tests), fine motor coordi-

nation (Pole test) and strength with coordination (Inclined and
Inverted Screen tests).

| Rotarod
Motor coordination was assessed using the Rotarod (Rotamex-5; Columbus Instruments, Columbus, OH) following our previously published methods. 19 Briefly, latency to fall was measured for each mouse in three different situations: a stationary rod (for up to 60 s), a continuously rotating rod (3.0 rpm; for up to 60 s) and an accelerating rod

| Light/dark box
The light/dark box was used to assess anxiety-related approachavoidance behavior leveraging the mouse's innate preference for dark spaces. Mice were placed in the dark side of a chamber (47.6 Â 25.4 Â 20.6 cm) and were allowed to explore freely. The light side was illuminated at 65 lux with incandescent desk lamps. Beam brakes were used to measure time spent in each chamber during the 6-min task. For the first 2 min, mice were confined to the dark side of the apparatus, then mice had 4 min to explore the entire apparatus.
Time spent in and latency to move to the light side during the latter 4 min were used as a proxy for anxiety-like behavior, with more anxious-like mice avoiding the brightly lit open space.

| 3-Chamber social approach
Sociability and preference for social novelty were examined in the social approach task, following our previously published methods. 3 The mice received two, 10-min habituation trials: first to the center chamber of the apparatus and then to the entire chamber including the empty social investigation cups. Next, sociability was assessed for 10 min during which a novel age-and sex-matched conspecific was placed under one cup (the side used was counterbalanced across groups). During the fourth 10-min trial, a second, age-and sexmatched novel conspecific was placed under the other cup to assess preference for social novelty. The time spent in and number of entries into the investigation zone (defined as a 2 cm zone surrounding each cup), as well as time in and entries into each chamber and total distance traveled, was quantified using Any-Maze video tracking software.

| Tube test of social dominance
As social creatures, mice create social hierarchies within their social groups. Thus, laboratory mice acquire social hierarchical rank behaviors within their cage environments between 6 and 8 weeks of age, which can be leveraged to examine normal social dominance behavior.
We tested for this normal hierarchical behavior in our mice using the tube test for social dominance following our previously described methods. 3,16 4.6 | Acoustic startle/pre-pulse inhibition task Sensorimotor gating and startle reactivity were assessed using the acoustic startle/pre-pulse inhibition (PPI) task following our previously published methods. 3,12 Briefly, acoustic startle to a 120 dB auditory stimulus pulse (40 ms broadband burst) and PPI (response to a prepulse plus the startle pulse) were measured concurrently using computerized instrumentation (StartleMonitor, Kinder Scientific) over 65 randomized trials. A percent PPI score for each trial was calculated using the following equation: % PPI = (startle pulse alone À (prepulse + startle pulse))/startle pulse alone Â 100.

| Resident intruder
To assess agonistic, and thus aggression-related, behaviors, we used the resident intruder paradigm that leverages the propensity of male mice to defend their territory from an unfamiliar male as previously described. 11,33 Briefly, male mice were single housed for 6 days followed by 4 more days co-housed with a female to establish a terri- In addition to USVs, weight and temperature were recorded for each mouse at each time point. A non-contact HDE Infrared Thermometer was used to take the temperature of each mouse before they were removed from the nest for USV recording. Mice were weighed after recording. Pinnae detachment was also assessed at P5, and eye opening was documented at P14. At P14, the righting reflex was evaluated for each mouse by measuring the time for pup to right itself after being held on its back for 5 s as described previously. 3 Three trials, limited to 1 min, were performed for each mouse, averaged for analysis and direction of righting was noted. To determine the role Gtf2ird1 plays in the WS behavioral repertoire, we first generated and validated a novel mouse overexpressing the general transcription factor GTF2IRD1 (TG-Gtf2ird1-HA) via its endogenous regulatory elements, engineered using a BAC with an HA tag that was inserted just prior to a stop codon of Gtf2ird1
Next, we showed the ability of the TG-Gtf2ird1-HA mouse to  (Table 1), enabling a study of main effects of each allele, as well as detection of interactions. We likewise included sex in all subsequent analyses, and report sex effects when significant.

| Gtf2ird1 restoration ameliorates select sensorimotor coordination deficits in CD mice
Both WS and Dup7 are associated with strength deficits and motor delays. While Gtf2ird1 has been connected to the WS craniofacial phenotype and is suspected to play a role in the unique cognitive profile (which includes visuospatial processing deficits) and behavioral features of WS, its role in sensorimotor features of WS has not been thoroughly defined. To address both the impact of Gtf2ird1 on these features and the complete WS deletion in CD mice, we devised a comprehensive assessment of sensorimotor abilities, which also provided information necessary to properly interpret tasks relying on adequate motor performance. The wide-ranging compilation of tasks addressed a variety of basic motor abilities and more complex tasks requiring integration of sensory information (in mice, coordinated movement often is informed by their whiskers, rather than their eyes). 9 We split relevant tasks across two cohorts; in the first cohort ( Figure 2A, above midline), we tested activity over 1 h in an open field apparatus and natural digging behaviors as observed in the marble burying task. Animals in the second cohort (Figure 2A, below midline) were tested using the rotarod task, acoustic startle response/PPI assay and a sensorimotor battery, which included walk, inverted screen, pole, platform and ledge tasks to assess a variety of movement To examine motor coordination more directly, we used the rotarod task ( Figure 2M), which showed another partial rescue ( Figure 2N; CD Â TG interaction: F(1,69) = 6.977, p = 0.01). While all mice learned the task and generally improved over subsequent trials, CD and CD/TG animals had a shorter latency to fall relative to WT and TG animals ( Figure 2N; F(1,69) = 35.227, p = 1.06 Â 10 À7 ). The interaction between CD and TG alleles (e.g., rescue) was most apparent in females ( Figure 2O Sensory sensitivity is another feature of WS that warrants investigation, as WS individuals are more reactive to sounds. 8,15 In the acoustic startle/PPI task ( Figure 2S

| Restoring Gtf2ird1 expression in CD mice rescues light-avoidant but not center-avoidant anxiety-like behaviors
Anxiety is another feature common to WS and Dup7, although the specific forms differ. Non-social anxiety and increased prevalence of phobias are over-represented in the WS population, while Dup7 is characterized by greater social anxiety and separation anxiety, with no clear phenotype related to fear. As there are no specialized treatments for these symptoms among patients, having a well characterized model for preclinical screening of therapeutics may eventually lead to better care. Thus, we thoroughly assessed non-social anxiety-like features in the CD mouse model to identify tasks sensitive to this mutation and evaluate the potential impact of Gtf2ird1.
Anxiety-like behavior is measured in rodents by quantifying approach-avoidance behavior in low-threat situations in which perceived danger is diffuse and uncertain. 14 Table S2. and the light/dark box task was performed utilizing the second cohort of animals ( Figure 3A, below midline). differences in freezing were observed during the training day of the conditioned fear task, although in general females froze more than males (inset). (I) Day 2 of the conditioned fear task measuring context-based fear recall also showed no significant differences, except between the sexes (inset). (J) While differences between WT and CD animals were not significant, the TG allele increases percent freezing during cued recall in the third day of conditioned fear relative to those animals without that allele. (K) The TG effect on percent time freezing during cued recall is greater in female mice. T + S = tone + shock. All statistical details including sample sizes are reported in Table S3.
Interestingly, during the light/dark box task ( Figure 3F), we observed a significant interaction of CD and TG alleles on the percent time spent in the light ( Figure 3G; F(1,72) = 5.250, p = 0.025). CD animals spent significantly less time in the light relative to their WT ( p = 0.024) and TG/CD ( p = 0.040) counterparts, while TG/CD animals were not significantly different from the WT group, reflective of the TG allele rescuing CD deficits in this task. Thus, Gtf2ird1 complements the CD mutation for this phenotype.
In contrast to the approach-avoidance anxiety-like measures, fear responses, which have a component of anxiety, are quantified in situations where a threat is imminent and well-defined. 14 We used the fear conditioning task in the first cohort of animals to further evaluate fear responses and anxiety-related associative memory by quantifying freezing behavior in response to a shock paired with a novel auditory cue and spatial context. No differences in freezing response to the pairing of the shock and tone + context were observed between genotypes on Day 1 ( Figure 3H). However, females froze more than males overall (Day 1, min 3-5; F(1,85) = 5.606, p = 0.02). This sex effect was also observed during contextual fear recall on Day 2 when mice were re-exposed to the spatial context to test hippocampaldependent spatial conditioning ( Figure 3I; F(1,85) = 5.650, p = 0.02).
The CD mice also showed reduced freezing, similar to our previous reports ( Figure 3I,J), 25 but did not pass the significance threshold.
Mice with only the CD allele showed reduced freezing behavior compared with all other groups, replicating our previous effect, 25 although the comparison to WT mice did not pass the significance threshold (p = 0.078). Shock sensitivity was comparable across groups (Table S3).

| Enhanced social approach and motivation is independent of Gtf2ird1
Finally, given the interesting contrasting social motivation phenotypes in WS and Dup7 patients, 2,6 we conducted a comprehensive phenotyping of social behavior in our cohorts (Figures 4A and 5A). To identify early signs of social behavior changes, we assessed social communication in pups via the maternal isolation-induced USV paradigm in independent cohort 3 ( Figure 4A,B). Given elevated aggression in Dup7 patients, 10 we included standard measures of social dominance (tube test) and aggression (resident intruder) measured in cohort 2. Sociability differences in the CD model was originally identified in a modified single chamber version of social approach (OFSA), rather than the typical 3-chamber social approach task widely used in ASD models. 5,29 Previous work in our lab failed to identify differences in the 3-chamber task alone, although on a C57BL/6J Â FVB hybrid background that showed lower social approach in general. 11,24 Thus, our comprehensive battery here included a deliberate precise replication of the OFSA conditions as a baseline control in cohort 1, 27,29 the standard 3-chamber social approach assay in cohort 2 5,21 and finally a 14-day social operant task we recently designed to be a direct measure of social motivation in rodents in cohort 1. 3,18 Early differences in communication were evident across the 3 days of the USV task. All groups exhibited a decreased number of calls overall compared with WTs (F(3,83) = 7.635, p = 0.00014; Figure 4C). Examination of the spectrotemporal features of the calls showed TG mice produced overall shorter calls compared with all other groups (F(3,91) = 3.411, p = 0.020; Figure 4D) with increased pause time at P7 (F(3,204) = 3.332, p = 0.021; Figure 4E). Call pitch and sound pressure levels were examined to identify possible laryngeal muscle abnormalities. Mean pitch frequency was lower in the TG/CD pups on P5, and both TG/CD and CD pups on P7 (F(8,129) = 4.998, p = 0.00002; Figure 4F), while the frequency range was narrower on P7 for TG/CD and TG pups (F(8,144) = 6.211, p = 6.9 Â 10 À7 ; Figure 4G). Finally, the sound pressure level, or volume, of calls was lower for CD pups on P5 compared with WTs (F (8,131) = 2.742, p = 0.008; Figure 4H). The differences in USV number do not appear to be because of gross developmental delays as they do not follow the same pattern as the weight data. Specifically, weights were comparable between CD and TG/CD animals from P5 to P9, despite both groups weighing significantly less than WT and TG mice (Geno: F(3,79) = 13.619, p = 2.8 Â 10 À7 ; Figure 4I). No significant differences were observed across groups for pup body temperature during call recordings ( Figure 4J). Acquisition of the surface righting reflex and eye opening was also assessed at P14. All pups had the ability to flip themselves upright, with no difference in time to exhibit the righting reflex ( Figure 4K), and no differences in the number of pups with eyes open across groups. Altogether, these data suggest there is an early disruption to social communicative behavior with haploinsufficiency for the WSCR, as well as overexpression of Gtf2ird1, which may be driven in part by musculature issues possibly because of loss of the Eln (Elastin) gene as suggested by the spectrotemporal call features.
In previous research, adult CD mice on the FVB Â C57BL/6J hybrid background showed decreased dominance in the tube test and reduced aggression in the resident intruder paradigms. 11 In our current study, CD mice on the C57BL/6J background seem to win less although no statistically significant difference was observed in the 1 day tube test paradigm ( Figure 5B; H(7) = 11.102, p = 0.134). In addition, there were no significant differences in the number of attacks exhibited across groups in the resident intruder task Similar to Segura-Puimedon et al., 29 in the OFSA task ( Figure 5D) we found CD animals spent a greater portion of time investigating the social stimulus mouse relative to WT animals ( Figure 5E; p = 0.013).
In fact, we observed this increased social approach behavior in all groups relative to WT levels ( Figure 5E, H(3 To control for the potential impact of novelty on the OFSA task ( Figure 5D), we next tested the reaction to a non-social novel object,   Table S3.  (Table S4), which was calculated as above, but utilizing "time in novel zonetime in familiar zone" as the numerator.  Table S4.
platform and rotarod tasks) and potentially sensory processing more generally (sound sensitivity and potentially light sensitivity). While Gtf2ird1 affected a few features clearly, most features were not significantly impacted by its rescue or overexpression, suggesting either  Table S4. extensively alternatively spliced gene, with numerous uncharacterized isoforms. Our TG-Gtf2ird1-HA mouse tags less than half of the isoforms (only the beta variants that contain the full exon 30) 31 ; leveraging this fact, we can investigate how these two groups of isoforms differ. Utilizing an HA antibody in a pull-down assay would effectively separate these isoform groups for downstream analysis to compare their functions, particularly in regard to genomic binding.
Synthesizing the results of this study beyond the contributions Kozel et al. 13 The hypersocial phenotype that has been documented in people with WS is also recapitulated in the CD mice in multiple behavioral tasks (OFSA, 3-chamber social approach and social motivation operant conditioning). Additionally, while an avoidance of the center space was consistently seen across many tasks, it was not observed in the OFSA task, showing how the presence of a conspecific can potentially overcome the center anxiety seen when no social stimulus is present.
Whether this means social stimuli can be more salient than anxiety cues or whether this indicates a separate circuit for social anxiety, the CD mouse model is appropriate for further teasing apart the underpinnings of this hypersocial phenotype as well as the anxietylike features mentioned above. Tasks with no significant data were not shown, including elevated plus maze, tube test for social dominance and resident intruder. Measures with an asterisk (*) have additional statistical findings, such as an interaction between CD and TG alleles or between sex and allele; sex effects alone are not shown. Arrow size reflects p value range: smallest arrow = p < 0.05, medium arrow = p < 0.01, largest arrow = p < 0.001. A lack of an arrow reflects no difference from WT, while an equal sign (=) indicates a rescue phenotype (i.e., a significant difference between CD and TG/CD groups although the TG/CD comparison to WT was not significantly different). The original figures and relevant statistical tables are included in the last column for reference.