A truncating Aspm allele leads to a complex cognitive phenotype and region-specific reductions in parvalbuminergic neurons

Neurodevelopmental disorders are heterogeneous and identifying shared genetic aetiologies and converging signalling pathways affected could improve disease diagnosis and treatment. Truncating mutations of the abnormal spindle-like microcephaly associated (ASPM) gene cause autosomal recessive primary microcephaly (MCPH) in humans. ASPM is a positive regulator of Wnt/β-Catenin signalling and controls symmetric to asymmetric cell division. This process balances neural progenitor proliferation with differentiation during embryogenesis, the malfunction of which could interfere with normal brain development. ASPM mutations may play a role also in other neurodevelopmental disorders, nevertheless, we lack the details of how or to what extent. We therefore assessed neurodevelopmental disease and circuit endophenotypes in mice with a truncating Aspm1–7 mutation. Aspm1–7 mice exhibited impaired short- and long-term object recognition memory and markedly enhanced place learning in the IntelliCage®. This behaviour pattern is reminiscent of a cognitive phenotype seen in mouse models and patients with a rare form of autism spectrum disorder (ASD) as well as in mouse models of altered Wnt signalling. These alterations were accompanied by ventriculomegaly, corpus callosum dysgenesis and decreased parvalbumin (PV)+ interneuron numbers in the hippocampal Cornu Ammonis (CA) region and thalamic reticular nucleus (TRN). PV+ cell number correlated to object recognition (CA and TRN) and place learning (TRN). This opens the possibility that, as well as causing MCPH, mutant ASPM potentially contributes to other neurodevelopmental disorders such as ASD through altered parvalbuminergic interneuron development affecting cognitive behaviour. These findings provide important information for understanding the genetic overlap and improved treatment of neurodevelopmental disorders associated with ASPM.

Associates Inc., VT, and USA). Wheels were not removed prior to behavioral testing. The activity was indexed by the number of wheel revolutions. Total running wheel activity from the first 4 weeks was analysed as well as the activity per day during this time.

Open Field
The Open Field (OF) analysis was carried out as described previously [2]. The arena consisted of a transparent and infra-red light permeable acrylic test arena with a smooth floor (internal measurements: 45.5 x 45.5 x 39.5 cm). Illumination levels were set at approx. 150 lux in the corners and 200 lux in the middle of the test arena. Data were recorded and analysed using the ActiMot system (TSE, Bad Homburg, Germany).

Y Maze
Spontaneous alternations were assessed using the Y-Maze as reported in a foregoing study [3,4]. The Y maze was composed of opaque light grey PVC and had 3 identical arms (30 x 5 x 15 cm) placed at 120° from each other; the illumination in the center of the maze was maintained at 100 lux. For each test, a mouse was placed at the end of one arm and allowed to move freely through the maze during a 5-minute session. Spontaneous alternations (defined as consecutive entries into all three arms without repetitions in overlapping triplet sets) were scored. Total arm entry numbers were collected over the 5 minutes. Spontaneous alternation performance percentage (SPAs %) is further defined as the ratio of actual (total alternations) to possible alternations (total number of triplets) x 100. Alternation behavior is a measure of spatial working memory. The percentage alternate arm returns (AARs %) were also calculated where the animal, after moving to a new arm, returns to the previously visited arm.

Social Discrimination
The Social Discrimination procedure was applied in a manner similar to that previously described [3,5]. The test consists of two 4-min exposures of stimulus animals (ovariectomized 129Sv females) to the test animal in a fresh cage to which the test animal had been moved 2 h prior to testing. During the first exposure, the sample phase, we introduced one stimulus animal into the cage and allowed it to roam freely with the test animal. A trained observer with a hand-held computer recorded the amount of time that the test animal explores (sniffing of the head and body, direct contact) the stimulus. This measure of the sample phase was used as an index of social investigation and social affinity. After a retention interval of 2h, this stimulus animal was re-exposed to the test animal together with an additional, previously not presented stimulus animal. A separate "familiar" and "unfamiliar" stimulus animal was assigned to each test animal. A trained observer with a hand-held computer again recorded the duration of investigatory behavior of the test animal towards the stimulus animals (familiar and unfamiliar) during this test phase. A social recognition index was calculated as time spent investigating the unfamiliar stimulus mouse / time spent investigating both the familiar and unfamiliar stimulus mouse.

Short and long-term object recognition memory
The object recognition procedure implemented has been described in detail [4]. In brief, mice were allowed to habituate to empty cages for 10 minutes once each day during the two days prior to the test. In the sample phase, two identical objects were presented in the cage for the mouse to explore three times for 5 minutes each with a 15-minute inter-trial interval. Three hours later, during the short-term recognition memory test phase, one of these objects with a novel object was presented to the mouse to explore for 5 minutes. Long-term recognition memory was then tested 24 hours later by presenting one of the original objects and another novel object for 5 minutes. The amount of time spent by the mouse investigating the objects in each phase was recorded. A recognition index was then calculated as the percentage time spent investigating the unfamiliar novel object/ (time spent investigating familiar+unfamiliar objects).

Place and reversal learning and working memory in the IntelliCage®
Place and reversal learning and working memory were assessed using the IntelliCage® (detailed procedure in [4]). The IntelliCage® (NewBehavior, TSE Systems GmbH, Bad Homburg, Germany) is an automated behavioral analysis apparatus for assessing individually RFIDtagged (Planet ID "White Label" (Planet ID GmbH, Essen, Germany) mice living in social groups.
The plastic homecage in this case is relatively large (55 x 37.5 x 20.5 cm 3 ) where the four cage corners are operant conditioning chambers (15 x 15 x 21 cm 3 ). Up to 10 mice are housed within one cage and can access drinking water in the corners. Only one mouse can enter each corner at a time ("corner visit") and nosepoke through an opening to access water ("nosepoke"). With this apparatus, the mice were assessed for place learning as an index of spatial reference memory, reversal learning to measure behavioral flexibility and patrolling ability as an index of working memory. Two separate cages were used for this experiment and mice were housed according to genotype. During the "free adaptation phase" (days 1-3), all motorized doors were open and mice had free access to all corners to drink water. During this 3 day period, spontaneous activity and exploratory indices (nosepokes, corner visits) were recorded. The number of corner visits, nosepokes and percentage visits with nosepokes were compared between genotypes during the first hour and 24 hours after introduction into the IntelliCage® as a measure of habituation to a novel environment. On day 4, there was a nosepoke adaptation period where mice learned to nosepoke for water. On days 5-8, the mice learned to access water in one of the IntelliCage® corners ("correct corner"). The corner assigned to each mouse was the least preferred corner during the adaptation phase and attempts to access water in the other three corners registered as an error ("incorrect nosepoke"). The percentage error rate (# incorrect nose pokes/total # nose pokes x 100) was then calculated and a decreased error rate signified enhanced spatial reference memory ability. On days 9-12, after the place-learning task, the assigned corner for each mouse switched to the opposite corner. The percentage error rate (as above for place learning) was calculated for each animal and indexed the behavioral flexibility of the mouse. On days 13-16, working memory of the mice was assessed in the patrolling task. The mice learned that the assigned corner would be the next adjacent corner once water access was successful. This assignment switched to the next corner in a clockwise direction and three yellow LED lights, situated at each door, illuminated when the mouse entered the correct corner. We calculated the percentage error rate for each animal as before during place and reversal learning.

Tissue preparation
Mice were euthanised at the age of 60 weeks and perfused transcardially with a solution of regions, the thalamic reticular nucleus and anterior cingulate cortex (ACC). A preliminary analysis was performed for all ROIs to ascertain the optimal sampling conditions. Using the 10x objective, the ROIs were outlined using the software according to the stereotactic coordinates in the mouse brain atlas as follows: dorsal hippocampal dentate gyrus between -0.94 to -2.92 mm, CA1, 2, 3 between -1.34 to -3.16 mm (sampling grid: 100 µm, counting frame: 100 µm), TRN between -0.46 to -1.94 mm (sampling grid: 300 µm, counting frame: 100 µm) and ACC between 1.18 to -0.10 mm (sampling grid: 100 µm, counting frame: 100 µm) [7]. Total cell number (N) within the ROI was then estimated according to the following formula: In this formula, Qrepresents the counts, ssf is the section-sampling fraction, asf is the areasampling fraction and tsf is the thickness sampling fraction [8]. Volumetric analysis of selected ROIs was performed using the Stereoinvestigator software and the Cavalieri estimator probe as described previously [6]. The following brain regions of interest were analysed: corpus callosum between 1.10 and -0.46 mm (5 sections, sampling grid: 15 µm), lateral ventricles between 1.18 and -1.06 mm (6 sections, sampling grid: 50 µm), CA1, 2, 3 and DG between 0.94 and 3.16 mm (7 sections, sampling grid: 15 µm).

Statistics
Numerical analyses were performed using GraphPad Prism version 7.03 for Windows