Normal radial migration and lamination are maintained in dyslexia-susceptibility candidate gene homolog Kiaa0319 knockout mice

Developmental dyslexia is a common disorder with a strong genetic component, but the underlying molecular mechanisms are still unknown. Several candidate dyslexia-susceptibility genes, including KIAA0319, DYX1C1, and DCDC2, have been identified in humans. RNA interference experiments targeting these genes in rat embryos have shown impairments in neuronal migration, suggesting that defects in radial cortical migration could be involved in the disease mechanism of dyslexia. Here we present the first characterisation of a Kiaa0319 knockout mouse line. Animals lacking KIAA0319 protein do not show anatomical abnormalities in any of the layered structures of the brain. Neurogenesis and radial migration of cortical projection neurons are not altered, and the intrinsic electrophysiological properties of Kiaa0319-deficient neurons do not differ from those of wild-type neurons. Kiaa0319 overexpression in cortex delays radial migration, but does not affect final neuronal position. However, knockout animals show subtle differences suggesting possible alterations in anxiety-related behaviour and in sensorimotor gating. Our results do not reveal a migration disorder in the mouse model, adding to the body of evidence available for Dcdc2 and Dyx1c1 that, unlike in the rat in utero knockdown models, the dyslexia-susceptibility candidate mouse homolog genes do not play an evident role in neuronal migration. However, KIAA0319 protein expression seems to be restricted to the brain, not only in early developmental stages but also in adult mice, indicative of a role of this protein in brain function. The constitutive and conditional knockout lines reported here will be useful tools for further functional analyses of Kiaa0319. Electronic supplementary material The online version of this article (doi:10.1007/s00429-016-1282-1) contains supplementary material, which is available to authorized users.

The expected effect of the trapping cassette in KO1 and NZ alleles is to prevent normal splicing after exon 5, which will be joined to the En2 exon present in the cassette, and resulting in a short chimeric protein (p.D374GfsX102) lacking the PKD, C6, transmembrane and cytoplasmic domains of KIAA0319 protein (Velayos-Baeza et al. 2007, 2008. The predicted result at the protein level from the Null allele is p.D374VfsX14 although the resulting transcript, unlike with the KO1 and NZ alleles, would be expected to undergo degradation by nonsense-mediated mRNA decay (NMD) and no protein would be produced.
C57BL/6J mice were obtained from Harlan Laboratories UK. Both FLPe and Sox2cre mouse lines were already available at the WTCHG animal facility, kept by backcrossing into C57BL/6J.

Custom anti-KIAA0319 antibodies
Specific antiserum R7 against the C-terminal domain of mouse KIAA0319 protein has been previously described (Velayos-Baeza et al. 2010). Custom polyclonal rabbit antiserum R5 against the ectodomain of KIAA0319 was obtained from Eurogentec Ltd after immunisation with peptides [E+F] (residues 150-164, C+ PEETTEYSDEYKDLE, and 208-222, MEKLQDPTPHPLDQE+C, respectively, of the mouse KIAA0319 protein). Characterisation of reactivity in Western blotting and Immunofluorescence applications, including affinitypurified antibodies against each of the immunisation peptides and performed using overexpression of mouse and human proteins in mammalian cells, showed that only epitope F in the mouse protein was recognised and that only non-glycosylated protein seemed to be detected (results not shown). Therefore this antiserum is not able to recognise the endogenous, glycosylated, mouse KIAA0319 protein. However, transient overexpression of the KIAA0319 protein allows accumulation of non-modified protein that can be detected by the R5 antiserum.

Behavioural tests
Elevated Plus Maze (EPM). The test was essentially carried out as previously described (Ufartes et al. 2013). The maze consisted of 2 opposing open and 2 opposing closed arms.
All arms were 30 cm long and 5 cm wide. Half the mice were placed in the EPM facing left and the other half facing right in the open arm and allowed to freely explore the apparatus over a 10-min trial. Animals were tracked using the AnyMaze System (Stoelting, USA). Entry into an arm of the EPM was defined using the entire animal with at least 70% of the animal Open field. Locomotor activity was measured using the PAS Home Cage system (San Diego Instruments, San Diego, CA, US). Mice were placed into a plexiglass cage with 4x8 photobeam configuration and tested over a period of 60 min. The total beam breaks were measured and divided into 10 minute time bins for analysis.
Locomotor habituation. Locomotor habituation to the increasingly familiar environment was assessed using a computer controlled system (TSE-Systems, Bad Homburg, Germany). Mice were exposed 3 times for 10 minutes to the same rectangular Plexiglas cage [35 cm (w) × 20 cm (d) × 20 cm (h)] inserted into the TSE-System with 1 hour interval between exposures. Cages were cleaned with water and alcohol between animals. Animals' locomotor activity was automatically scored by the photobeam system (15x15 beams); decrease in distance travelled was used as a measure of habituation.
Rotarod. The acceleration rotarod was carried out as described (Ufartes et al. 2013). Mice were given 3 trials per session and 3 sessions per day with a 60-min intersession interval.
Mean speed at fall from three consecutive trials was used as a measure of motor coordination. If a mouse failed to grip in three times in the initial 10 s/4rpm period it was excluded.
Inverted screen, Grip strength, Spatial novelty (Y maze) and Social behaviour / social memory. Tests were performed as described (Ufartes et al. 2013).
Spontaneous alternations (T maze). The apparatus and method have been described (Ufartes et al. 2013). Each mouse received 6 trials over 2 sessions. The inter-session interval was 80 minutes. The percentage of passes over 6 trials was calculated to indicate the degree to which the mouse explored a different goal arm to the previous run (alternation).
Object recognition. Med Associates activity chambers (ENV-510) were used containing a black Perspex box (27 x 27 cm) insert. The objects were metal brackets glued to 5 x 5 cm metal bases and novel and familiar objects were similar in size, approximately 6.5-8.5 cm x 3.5-4 cm. During testing, the objects were placed 4 cm from the middle part of the walls to allow investigation around the object. Tests were recorded and analyzed by the AnyMaze System (Stoelting, USA). The test consisted of three sessions. During session 1 (habituation phase) the mouse was allowed to freely explore in an empty open field arena for 10 minutes and locomotor activity was assessed. During session 2 (exposure phase), two identical objects were placed in the arena and the mouse freely allowed to investigate the objects.
Response to object novelty was examined during session 3 (test phase) by replacing one of the familiar with a novel object and the mouse was allowed to explore. Session 2 and 3 were 5 min in duration and were separated by a 3-min inter-trial interval (ITI) where the mouse was placed in a holding cage next to the testing box. During this period, the box and objects were wiped with water and 70% alcohol. The duration and frequency of object exploration during each session was recorded and response to object novelty was assessed by comparing the time spent in contact with the novel object versus the time spent in contact with the familiar object. Object exploration was scored only when the mouse's head was in contact with the object. The position of the novel and familiar objects were counterbalanced.
A performance ratio was calculated (novel/novel+familiar) where 0.5 denotes equal object preference whilst higher or lower denotes a preference for novel or familiar respectively.
Light/dark box. The test was performed as previously described (Scheneider et al. 2012) in Med Associates activity chambers (ENV-510). The dark box insert was made of black Perspex designed to cover half the area of the activity chamber (27 x 13.9 x 21.5 cm) with a 4 x 4 cm hole placed in the middle of the wall at floor level. Time spent in and latency to enter light and dark zones as well as the number of full-body transitions between the light (300 lux) and dark (2 lux) compartments were automatically scored by Med Associates activity software. Animals were started in the light compartment; the session lasted 10 minutes.
Pre-pulse inhibition of acoustic startle. Startle response and pre-pulse inhibition of acoustic startle responses were measured by the SR-Lab System (San Diego Instruments, San Diego, CA, USA). The mouse was placed in a Plexiglas cylinder and left to acclimatise for 5 min with constant 65dB sound (background noise). A test session contained ten periods in which the trial stimuli were included in pseudorandom order so that they appeared once within each period of 12 trial stimuli: startle stimulus (40 ms, 120 dB sound burst), pre-pulse stimulus (20 ms, 81 dB), baseline stimulus (20 ms, 65 dB, to measure baseline movement in the cylinders), and 9 combinations of pre-pulse and startle stimuli ("pre-pulse-plus-pulse") spaced by either 10,20,30,40,50,100,200,400, 800 ms delays, starting at the end of the pre-pulse stimulus. The test session started and finished with five startle stimuli. The average inter-trial interval was 15 s (ranged from 10 to 20 s). The startle response was recorded every 1 ms during a 65 ms sampling window starting with the onset of the startle stimulus or the 81dB pre-pulse alone stimulus. The following formula was used to calculate the percentage of pre-pulse inhibition of a startle response: 100-("pre-pulse-plus-pulse" ×100/"pulse-alone").
Unistat 6.5 (Unistat Ltd, London, UK) software was used to perform statistical data analysis using one way one or two-way analysis of variance (ANOVA). To follow up, post-hoc analysis using Bonferroni-modified least significant difference was used. When animals were tested multiple times in the same task or within session periods were used, ANOVA with repeated measures was applied. replaced by a trapping cassette (containing two FRT sites (F1, F2) and two loxP sites (P1, P2)) and a loxP-cassette (containing a third loxP site (P3)), respectively, to obtain a "knockout first" (tm1a or KO1) allele. After recombination with Cre, the region between sites P1 and P3 is deleted to obtain a "Null-lacZ" (tm1b or NZ) allele. Deletion of region between sites F1 and F2 by Flp recombination generates a "floxed" (tm1c or Flx) allele with conditional knockout potential: a "Null" (tm1d or Null) allele where exon 6 is deleted can be generated after Cre recombination between sites P2 and P3. Long-range PCR fragments (KO1-1 to KO1-4) used for target confirmation and sequencing, and the PCRs used for identification of the different alleles (genotyping) are shown. Details about elements in the targeting cassette and the different PCRs are shown in Fig. S3 and S4. b Results of long-range PCRs from KO1 homozygous mice obtained from two different male chimeras (80% and 90%). DNA ladder sizes are shown on the right. c Results obtained with genotyping PCRs from mice homozygous for wild type (ww), KO1 (aa), NZ (bb), Flx (cc) or Null (dd) alleles; heterozygous Flx (cw) also included to show double band with KWF. Size of each fragment (bp) is shown on the right. d Western blotting analysis from different organ lysates from 6-week old mice with specific antiserum R7 shows a heterogeneous band pattern in wild type (wt) samples (left panel). The same pattern is maintained in samples from homozygous NZ mice (middle and right panels) except for a band of ~170kDa (white asterisk), the expected size for glycosylated full-length KIAA0319 protein, in brain and cerebellum (and possibly testis) samples. 30 µg (left panel) or 50 µg (middle and right panels) total protein loaded per lane.