Dcdc2 knockout mice display exacerbated developmental disruptions following knockdown of Dcx

Targeted genetic deletion of Dcdc2

In order to generate Dcdc2 mutant mice we sequentially generated mouse lines bearing engineered Dcdc2 alleles; a conditional deletion or “floxed” allele in which exon 2 was flanked by loxp sites, Dcdc2^flox2, and a constitutively deleted allele in which exon2 was deleted, Dcdc2^del2((figure 1 A&B). Deletion of exon 2, an exon present in all annotated splice variants of Dcdc2, is predicted to result in a frame shift and premature stop codon when exon 1 and exon 3 are spliced together. To verify this aberrant splice variant and stop codon in the mutants we used RT-PCR with primers to exon 1 and exon 3, sub-cloned, and then sequenced the resulting fragments. As shown in figure 1C, consistent with loss of the 52 base pair exon 2, the amplified product from homozygous mutant animals (Dcdc2^del2/del2) was approximately 50 bases smaller than that amplified from Dcdc2^wt/wt animals (figure 1C). Furthermore, the sequence of the amplified fragment from homozygous mutants indicated that the exon 1-to-exon 3 spliced sequences contained the expected premature stop codon (figure 1D). Introduction of premature stop codons by mutation often result in transcripts that are degraded by nonsense-mediated mRNA decay. We tested for such decay by quantitative rt PCR ((qPCR) to determine whether Dcdc2 mRNA levels were lower in mice with alleles missing exon 2. In cDNA prepared from RNA isolated from the brains of homozygous mutants we found evidence of approximately 10 fold decrease in Dcdc2 mRNA than in Dcdc2^wt/wt animals. Similarly, in heterozygous animals, with one mutant and one wt allele, we found intermediate levels of Dcdc2 mRNA. These results are consistent with potent nonsense mediated decay and loss of Dcdc2 transcripts in Dcdc2^del2/del2 mutant mice (figure 1E; N=21 p<0.001). We also attempted to confirm decreased expression at the protein level, however five commercially purchased and tested antibodies failed to identify bands of the appropriate MW for Dcdc2 even in wild type brains. Nevertheless, the aberrant splice variant, premature stop codon, and potent nonsense-mediated decay in the mutant provide substantial evidence that the Dcdc2^del2 is a loss of function mutant allele for Dcdc2.

In crosses between mice heterozygous for the del2 and wt alleles of Dcdc2, as well as in crosses between animals homozygous and heterozygous for the del2 alleles, we observed the expected mendelian ratios of 1:2:1 and 1:1 respectively, indicating no evidence of embryonic lethality associated with Dcdc2^del2. Animals homozygous for the del2 deletion are healthy and are fertile. As an initial screening for behavioral changes in Dcdc2 knockout mice we tested Dcdc2^del2/del2, Dcdc2^wt/del2, and Dcdc2^wt/wt mice on open field behavior and on one configuration of the Hebb-Williams maze (maze 1). As shown in figure 1 F&G, there were no significant behavioral differences in either behavioral test. Measures of spontaneous locomotion and exploratory behavior in the open field test showed no significant differences (Figure 1; p >0.05). Similarly, there was no significant difference in performance on Maze 1 of the Hebb-Williams maze, with no significant interaction between trial and genotype for the number of errors committed during testing (p >005.), and no significant main effect of genotype (p> 0.05) in learning the maze. There was, however, a significant main effect of trial across genotypes, confirming that animals learned the task (p< 0.05).

Dcdc2 mutants have structurally normal brains

Dcdc2 knockout mice showed no defects in brain morphology as assessed by comparison of serially sectioned brains from Dcdc2^wt/wt and Dcdc2^del2/del2 mice (figure 2 A,B). Laminated neural structures, including neocortex, hippocampus and cerebellum, all showed typical morphologies. Similarly, the size and organization of major white matter tracts showed no evidence of disruption. There was also no evidence of focal developmental disruptions in neocortex, including neither periventricular heterotopia nor layer 1 ectopia. In addition, the numbers of total neuronal and non-neuronal cells in the cerebral neocortex, as assessed by non-biased stereology of Nissl-stained sections, showed no significant differences between wild type and knockout mice. Furthermore, immunohistochemistry of two neocortical-layer specific markers, Cux1 (layers II–IV) and Tbr1 (layer V and VI) revealed no significant differences in neocortical lamination patterns (figure 2 C&D).

As Dcdc2 has been shown to bind to microtubules, and Dcdc2 expressed in hippocampal neurons, localizes to neuronal cilia and alters cilia signaling (Massinen et al, in press), we performed an assessment of neuronal cillia in hippocampus and cerebral cortex in P54 Dcdc2^wtwt and Dcdc2^del2/del mice. Neuronal cilia can be identified immunocytochemically due to their enriched expression of type III adenylyl cyclase and pericentrin, proteins localized to the axoneme and basal body, respectively (Bishop et al., 2007, Anastas et al., 2011). We therefore performed an immunocytochemical assessment of neuronal cilia in hippocampus and neocortex of 54 day old Dcdc2^del2/del2 and Dcdc2^wt/wt mice. However, the structure, numbers, and lengths of neuronal cilia in neocortex and hippocampus did not differ between Dcdc2^del2/del2 and Dcdc2^wt/wt mice (figure 3). Further studies are needed to determine whether signaling to or from these cilia is altered in Dcdc2 mutants.

Dcdc2 mutants display normal neocortical neurogenesis and neuronal migration

Recent reanalysis of Dcx knockout mice, which were initially shown to have undisturbed neocortical lamination (Corbo JC et al., 2002), revealed significant changes in both neurogenesis and neuronal migration in fetal development in Dcx mutants that were largely resolved and no longer apparent by later postnatal periods (Pramparo T et al., 2010). We therefore investigated both neurogenesis and migration during the fetal period in Dcdc2 knockouts. To assess neurogenesis, we compared the percentage of neocortical progenitors within the VZ that were in M-phase of the cell cycle (phoH3+ cells) at embryonic day 15 (E15). We found no significant differences between wild type and Dcdc2 knockout mice (figure 4A). We also tested whether Dcdc2 loss altered the fraction of neocortical progenitor cells that exit the cell cycle. For this analysis, BrdU injections were made at E15, and brains were harvested and processed for BrdU and Ki67 immunohistochemistry 24 hours later. The percentage of cells that had exited the cell cycle after 1 day, BrdU⁺ and Ki67^- cells, did not significantly differ between Dcdc2^wt/wt and Dcdc2^del2/del2 mice (figure 4A). Together these results indicate that the number of mitotic neural progenitors or the rate at which neocortical neuronal progenitors become postmitotic is not altered by the Dcdc2 deletion mutation.

In order to test whether there were any defects in neuronal migration in Dcdc2 mutants, we performed three different experiments. First, we injected pregnant females at gestational day 15 with BrdU and examined the positions of BrdU positive neurons within neocortex on the day of birth. As shown in Figure 3C, in both wild-type and knockout mice, BrdU labeled cells reached the top of the cortical plate revealing no apparent migration delays or arrest. In an additional assay for migration, we used electroporation of VZ progenitors at the ventricular zone at E15 to label migrating neurons with GFP and then assessed the position of neurons on the day of birth. Similar to the BrdU assay, GFP labeled neurons were present in upper layers in both knockouts and wt mice (figure 4D). Lastly, to test whether cells lacking a functional copy of Dcdc2 would migrate more slowly if migrating within the context of a population of cells with functional Dcdc2 alleles we used animals homozygous for the Dcdc2^flox2 allele and transfected these Dcdc2^flox2/flox2 animals at E15 with plasmids expressing cre, pCAG-Cre, and a conditionally gated GFP, pCALNL-GFP. Cre-transfected cells in wild type and in Dcdc2^flox/flox animals were marked by the expression of GFP and these cells migrated similarly from the VZ to superficial layers of neocortex (figure 4E) in both Dcdc2^flox2/flox2 and Dcdc2^wt/wt. Together, the results of these three neuronal migration assays indicate that genetic deletion of Dcdc2 in mice does not result in impaired neuronal migration of pyramidal neurons in mouse neocortex.

Dcx RNAi impairs neuronal migration more in Dcdc2 knockouts than in wildtype mice

Previous studies of compound Dcx and dclk1 mutations in mice indicated that loss of combinations of these genes results in greater impairments in neuronal migration and differentiation than does loss of any single gene alone (Deuel TAS et al., 2006; Koizumi H et al., 2006). In order to test for evidence of a similar functional relationship between Dcdc2 and Dcx we compared the effects of Dcx RNAi on Dcdc2^del2/del2 and Dcdc2^wt/wt mutant mice. As we previously showed for Dcx RNAi in mice (Ramos RL et al., 2006), we found that Dcx RNAi delivered at E14 to wt type mice causes some cells destined for upper layers to be distributed into deeper layers in mouse neocortex, but does not lead to the formation of subcortical band heterotopia as it does in rat neocortex (Ramos RL et al., 2006). Similarly, in this study, subcortical band heterotopia failed to form in any wild type mice (n=8) transfected with Dcx RNAi (figure 5A). In contrast to the effects of Dcx RNAi in Dcdc2^wt/wt animals, 4 of 9 Dcdc2^del2/del2 mutants transfected with Dcx RNAi developed prominent subcortical band heterotopia in the white matter underlying neocortical lamina (figure 5B). These results indicate that the loss of Dcdc2 function by mutations creates a sensitized condition permissive to the formation of subcortical band heterotopia in mice upon decreased expression of Dcx.

To further confirm that the disruption in migration caused by Dcx RNAi was exacerbated in Dcdc2 knockout mice we quantitatively compared the positions of neurons within neocortex following transfection of a scrambled control and an effective Dcx RNAi in Dcdc2^del2/del2 and Dcdc2^wt/wt mice. As shown in the histogram in figure 5C, there was no significant difference between the distribution of neurons in P14 brains transfected at E15 with the scrambled control RNAi vectors in Dcdc2^del2/del2 and Dcdc2^wt/wt mice; however, there was a significant shift in the proportion of neurons that resided in deeper positions following RNAi against Dcx in the Dcdc2 knockout mice compared to wt controls (N=5, p<0.01). This difference was seen both in a significant decrease in cells residing in superficial layers, and a significant increase in the number of cells in deeper positions. Thus, Dcx RNAi impairs neuronal migration more in Dcdc2 knockouts than in wt animals.

Dcx RNAi impairs dendritic growth and differentiation more in Dcdc2 knockouts than in wt mice

Results from analysis of compound mutant mice for Dcx and Dclk2 indicate a synergistic function for Dclk2 and Dcx in the maturation of dendritic morphologies in hippocampus (Kerjan G et al., 2009). We therefore assessed whether there was a similar synergistic interaction between Dcdc2 and Dcx function in development of dendritic morphologies in the neocortex. For this analysis we measured the basal dendrites of layer III pyramidal neurons in somatosensory cortex in 5 brains within each of four conditions: Dcx shRNA in Dcdc2^wt/wt, Dcx shRNA scramble control in Dcdc2^wt/wt, Dcx RNAi in Dcdc2^del2/del2, and Dcx shRNA scrambled control in Dcdc2^del2/del2. We restricted the analysis to layer III neurons to avoid the possible confound of comparing displaced cells that reside in deeper layers in increased number in the Dcdc2^del2/del Dcx RNAi treated mice. The results of these experiments shown in Figure 6 A-F show that Dcx RNAi compared to control RNAi had no effect on the mean number of primary or secondary basal dendrites in Dcdc2^wt/wt mice (figure 6D). Dcx RNAi in Dcdc2^wt/wt mice, in contrast, significantly decreased the length of basal processes both total dendritic length and length of primary, secondary and tertiary processes (figure 6 E&F; p<0.01). The same Dcx RNAi treatment in Dcdc2^del2/del2 mice, created an even greater decrease in all measures of basal dendritic process number and length (p<0.01). The increased severity of the Dcx RNAi in Dcdc2^del2/del2 mutants consisted of a complete absence of tertiary basal processes following Dcx RNAi (figure 6F). All measures of basal process length and number were most reduced in Dcdc2^del2/del2 mice receiving RNAi targeting Dcx (figure 6 D-F; p<0.01). Thus, similar to neuronal migration, the effects of Dcx RNAi on dendritic elaboration is more severe in Dcdc2 mutants than in wild type mice.

Gene targeting and genotyping

Mice carrying the loxp-exon2-loxP conditional allele of Dcdc2 (Dcdc2^flox2) were made by the University of Connecticut Health Center Gene Targeting and Transgenic Facility by standard methods. Briefly, embryonic stem cells harboring a floxed allele of exon two of Dcdc2 were produced by electroporating mouse ES cells (129S6 (129SvEvTac) with a targeting construct, and subsequently drug selected and screened by PCR for correctly targeted ES cell clones. A single positive colony was expanded and used for embryo re-aggregation to produce 5 chimeric mice. Three of these mice were shown to germline transmit the targeted allele to offspring in a cross with C57BL6 mice. The PGK-Neo cassette in the targeting construct was then removed by crossing these mice with 129S4/SvJaeSor-Gt(ROSA)26Sor^tm1(FLP1)Dym/J mice (JAX labs). These offspring were used to generate a colony of Dcdc2^flox2/flox2 mice. In order to generate Dcdc2^del2/del2 mice with a deletion of exon 2 we crossed Dcdc2^flox2/flox2 mice with Hrpt-Cre mice, C57Bl6-Hprt^tm1(cre)Mnn/J (UCHC). Genotyping was subsequently performed by PCR using two pairs of primers (Loxp F: 5′-agtggatctgcagggttaac, Loxp R: 5′-cttcggttgttacacgagcaat; Exon2 F: 5′-gagtggatctgcagggttaacat; Exon2 R: 5′-aagccgaggcaagcagatcttta).

RT-PCR analysis

Total RNA of the cerebral cortex was extracted from Dcdc2^del2/del2, Dcdc2^del2/wt and Dcdc2^wt/wt knock-out mice and wild-type littermates by RNAqueous (Ambion). Reverse transcription (RT) reactions were performed with 5 μg of total RNA using the SuperScript II reverse transcriptase (200 U per reaction; Invitrogen, Carlsbad, CA). RT-PCR was performed using a forward oligonucleotide primers located in Dcdc2 exon1 (5′-atgaacggtcccagctccag) and reverse primer located in exon5 (5′-cccactccggagagttatctt') to amplify Dcdc2 fragments spanning exon1–exon5. PCR was performed for 35 cycles with a denaturing step at 94°C (1 min), followed by annealing at 58°C (1 min) and extension at 68°C (1 min). PCR products were then purified by agarose gel electrophoresis, ligated into the PCR3.1 vector (invitrogen), and then sequenced. For quantitative real time qRT-PCR of Dcdc2, primers to exon 5 (tat gtg gcc gtc ggc aga g) and exon 7 (ccg atg gtt gac ttg gat tgc) were used with SYBR GREEN (Applied Biosystems) and assayed on an ABI 7500 qPCR machine (Applied Biosystems). Product amplification was validated for linearity in a serial dilution, and the expected single 98 bp amplicon was confirmed by gel electrophoresis. To quantify expression of Dcdc2 mRNA the delta-CT from Gapdh expression (CT GAPDH- CT Dcdc2) was computed in triplicate technical replicates and a mean established for each of 6 brains from animals of each genotype. The primers used for GAPDH were forward (ggcaagttcaacggcacagtc) and reverse (tggtggtgaagacgccagtag).

Behavioral tests

The open field apparatus was composed of a white plastic, square-shaped (45.72 × 48.26 × 60.96 cm) enclosure. The floor was divided into a grid containing sixteen, 11.43 × 12.07 cm squares. The number of squares entered during a five minute testing period was measured by a blind observer. An entry into a field was defined as having all four paws cross the line into a square. The apparatus was cleaned with 70% EtOH after each test. A 60 × 60 cm Hebb-Williams maze (maze 1), with attached start and goal box, was constructed using black acrylic plastic; the removable floor and top were made with clear Plexiglass. One week prior to testing mice were given free access to water and restricted to 2 g of chocolate-flavored food pellets per day (Bio-Serv, Frenchtown, NJ). Mice were weighed daily and maintained at 85% of their normal body weight. During habituation, mice were allowed to explore the maze without interior walls for a maximum of 10 trials of 120 s duration. Once the subject left the start box a black plastic guillotine door was closed to prohibit re-entry. The trial ended once the mouse ate one chocolate-flavored food pellet located in the goal box, or 120s was reached. If the subject did not reach the goal box in 120s he was gently guided to the goal box, the door to the goal box was closed, and the mouse remained until one food pellet was consumed. Habituation ended when all 10 trials were completed, or the subject completed three consecutive trials in less than 30s. The next day mice were tested using the Maze 1 configuration (Meunier, 1986). The mice completed 6 trials in a maximum time of 120s per trial. ANY-maze tracking software (Stoelting) and a web camera (Logitech Quick Cam) were used to record and analyze behavior. The number of errors made during each trial was measured. Errors were calculated when the center of mass of the subject crossed into an error zone.

Serial Section analysis and Stereological Estimate of Neuron Number in the Cerebral Cortex

After perfusion and fixation (4% paraformaldehyde) brains P40-45 were washed in water for 24 hours before being dehydrated in a series of 80%, 95%, 100% ethanol and ethanol/ether. The brains were placed into 3% celloidin for at least a week followed by 12% celloidin for 2–3 days. The celloidin block was trimmed to achieve a stable base and notched on the left side for orientation. The sections were cut coronally on a sliding microtome at 30 μm, segregated into 10 compartments, and stored in 80% ethanol. Every fifth section was stained for Nissl substance with cresyl violet. This procedure entailed washing the sections in distilled water and then placing them in 0.5% cresyl violet acetate solution (which stains the Nissl substance) for 3–5 minutes. Each section was placed in distilled water for 1 minute and then differentiated and dehydrated in 80 and 95% ethanol. A few drops of colophonium were added to the 95% ethanol baths. If differentiation was adequate, the sections were then cleared with terpineol and passed through xylol. Sections were mounted with careful attention to orientation so that left and right were identifiable consistently. The sections were then coverslipped with Permount. All cell estimations were performed under 100× oil-immersion DIC-illuminated objective using the optical fractionator as implemented by Stereo Investigator. Preliminary research has determined the optimal parameters for the optical fractionator. Cells are estimated using a sampling frequency of every 20^th section. Using a sampling grid of 530 × 530 μm, cells that lie within a counting box (15 × 15 × 20 μm) are classified as being either neurons or non-neurons (glia, blood vessel-related endothelial cells, pial, and ependymal cells). All counts were performed using standard stereologic procedures (disector/3D counting).

Histology and immunohistochemistry

For fluorescent immunostaining, brains were dissected and drop fixed for embryonic or neonatal brain, or perfusion fixed for adult with 4% paraformaldehyde/PBS. Brain sections were prepared with vibratome (Leica) at 60∼80 μm and rinsed for 5 minutes in 1× PBS, blocked for 1 hr in blocking solution (1× PBS, 0.3% Triton X-100, and 5% normal goat serum), incubated either for 2 hr at room temperature or overnight at 4°C with the primary antibodies diluted in blocking solution, rinsed three times for 5 min each with 1× PBS, incubated with the appropriate secondary antibodies (Molecular probe, 1:200) diluted in blocking solution for 1 hr at room temperature, rinsed three times for 5 min each with 1× PBS, incubated 10 min with Topro3 (Molecular Probes,1;1,000), rinsed with 1× PBS, and coverslipped with Antifade (Molecular Probe, 1;3,000). Confocal images were captured using a Leica confocal microscope and imported into Adobe Photoshop. The primary antibodies were: Rabbit anti-GFP polyclonal antibody (molecular probe, 1:2,000); rabbit anti-CUX1 polyclonal anitboby, (Santa Cruz Biotechnology, 1:1,000); rabbit polyclonal anti-Tbr1 (Santa Cruz Biotechnology,1:1,000); rat polyclonal anti-Brdu (Accurate Chemical & Scientific,1;100); rabbit polyclonal anti-Ki67 (Novoacastra, 1:200); rabbit anti-phosH-3 polyclonal antibody, (Millipore, 1:200), mouse monoclonal anti-alpha tubulin (Sigma, 1: 2,000). For cilia detection the primary antibodies included a rabbit anti-ACIII (1:1000; Santa Cruz) and mouse anti-pericentrin (1:200; BD Biosciences). Quantifaction of fluorescent images was performed with ImageJ (NIH), and statistical comparisions were made by t-test for comparision of 2 groups and ANOVA for comparision of more than 2.

In Utero Electroporation

Briefly, pregnant mice were euthanized at E14, and uterus was exposed. Lateral ventricles were injected with pulled glass microcapillary needles with plasmids in a 0.01% fast green solution (Sigma). Electrodes were placed on either side of the embryo's head, and 3× 50 ms square pulses at 25 volts were administered at 1s intervals with a BTX830 square-wave pulse generator (Genetronics, Havard Apparatus). Brains were harvested at postnatal 21 days and preceded to immunostaining and imaging analysis.