Developmental and Behavioral Phenotypes in a Mouse Model of DDX3X Syndrome

Abstract

Background: Mutations in the X-linked gene DDX3X account for approximately 2% of intellectual disability in females, often comorbid with behavioral problems, motor deficits, and brain malformations. DDX3X encodes an RNA helicase with emerging functions in corticogenesis and synaptogenesis.

Methods: We generated a Ddx3x haploinsufficient mouse (Ddx3x^+/- females) with construct validity for DDX3X loss-of-function mutations. We used standardized batteries to assess developmental milestones and adult behaviors, as well as magnetic resonance imaging and immunostaining of cortical projection neurons to capture early postnatal changes in brain development.

Results: Ddx3x^+/- females showed physical, sensory, and motor delays that evolved into behavioral anomalies in adulthood, including hyperactivity, anxiety-like behaviors, cognitive impairments in specific tasks (e.g., contextual fear memory but not novel object recognition memory), and motor deficits. Motor function declined with age but not if mice were previously exposed to behavioral training. Developmental and behavioral changes were associated with a reduction in brain volume, with some regions (e.g., cortex and amygdala) disproportionally affected. Cortical thinning was accompanied by defective cortical lamination, indicating that Ddx3x regulates the balance of glutamatergic neurons in the developing cortex.

Conclusions: These data shed new light on the developmental mechanisms driving DDX3X syndrome and support construct and face validity of this novel preclinical mouse model.

Keywords: Corticogenesis; DDX3X syndrome; Intellectual disability; Movement disorder; Neurodevelopment.

Figure 1.. Ddx3x^+/− mice have postnatal physical and sensory delays.

A) Ddx3x^+/− pups show delayed growth. The plot shows the body weight from postnatal day (P) 1 to 27 across the three genotypes (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes, Day: F=1055.8, df=26; Genotype: F=4.5, df=1). B) Ddx3x^+/− adults have low body weight. The plot shows the body weight of 4-month old mice (n shown in legend; mean ± SEM; ⊗ indicate outliers; Student’s t test after outliers removal and Shapiro-Wilk test for normality, t=4.3, df = 84). C) Ddx3x^+/− pups have a delay in eye opening. The plot shows the eye opening scores (0, closed; 1, half-opened; 2, open) across the three genotypes (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes; Day: F=593, df=11; Genotype: F=18, df=1; Interaction: F=5, df=11). D) Ddx3x^+/− pups show a delay in developing visual placing skills. The plot shows the visual placing scores (0, response absent; 1, response present) across the three genotypes. Visual response was considered present (score 1) when pups lowered onto a flat surface reached out with the forepaws before the vibrissae touched the surface (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes; Day: F=53.8, df=5; Genotype: F=11.8, df=1; Interaction: F=6.6, df=5). E) Ddx3x^+/− pups have a delay in developing a startle response to an auditory cue. The plot shows the startle response score (0, response absent; 1, response present) across the three genotypes. Startle response was considered present (score 1) when pups turned their heads in response to an 80db click sound (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes; Day: F=285.2, df=12; Genotype: F=12.7, df=1; Interaction: F=4.9, df=12). F) Ddx3x^+/− pups show a delay in developing an ear twitch response to a tactile stimulus. The plot shows the reflex to ear twitch response (0, response absent; 1, response present) across the three genotypes (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes; Day: F=142.6, df=8; Genotype: F=12.2, df=1; Interaction: F=8.5, df=8). All data collected and scored blind to genotype. In all panels, Ddx3x^+/+ (blue), Ddx3x^+/− (red), and Ddx3x^+/y (yellow). *p<0.05, **p<0.01, ***p<0.001; ns, non significant.

Figure 2.. Ddx3x^+/− mice have postnatal motor delays.

A) Ddx3x^+/− pups have a delay in surface righting time. The plot shows the time that it takes for a pup to turn on the four paws from a supine position across the three genotypes (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes; Day: F=131.1, df=11; Genotype: F=24.7, df=1; Interaction: F=4.1, df=11). B) Ddx3x^+/− pups have a delay in acquiring negative geotaxis skills. The plot shows the response of the pup when placed head down on a mesh covered platform at a 45° angle (0, fall; 1, stay/move down/walk down; 2, turn and stay; 3, turn, move up, and stay; 4, turn and move up to the top) across the three genotypes (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes; Day: F=57.5, df=12; Genotype: F=16.1, df=1; Interaction: F=1.8, df=12). C) Ddx3x^+/− pups tend to fall more in a negative geotaxis task. The plot shows the number of falls (score 0) per pup across the three genotypes (n shown in legend; mean ± SEM; ⊗ indicate outliers; Student’s t test after outliers removal and Shapiro-Wilk test for normality; t=6, df=74). D) Ddx3x^+/− pups have reduced motor endurance and performance in a rod suspension test. The plot shows the time that it takes for a pup to fall off a rod across the three genotypes (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes; Day: F=29.4, df=9; Genotype: F=4.4, df=1; Interaction: F=1.3, df=9). E) Ddx3x^+/− pups have reduced grip strength. Grip strength was tested by placing pups on a mesh grid and rotating the grid from a horizontal to vertical position. The plot shows the angle of rotation to which pups fall off the grid across the three genotypes (n shown in legend; mean ± SEM; repeated measure ANOVA between Ddx3x^+/+ and Ddx3x^+/− genotypes; Day: F=94.5, df=9; Genotype: F=17.9, df=1; Interaction: F=5.4, df=9). All data collected and scored blind to genotype. In all panels, Ddx3x^+/+ (blue), Ddx3x^+/− (red), and Ddx3x^+/y (yellow). *p<0.05, **p<0.01, ***p<0.001; ns, non significant.

Figure 3.. Ddx3x^+/− mice have hyperactivity, anxiety-like behaviors, and cognitive deficits.

A-B) Ddx3x^+/− mice show increased locomotor activity in open field test. Panel A shows an example of activity of individual mice in a 10-min interval. The plot in B shows the distance covered in the 30-min open field test (n shown in legend; mean ± SEM; Student’s t test, t=2.1, df = 76). C) Ddx3x^+/− mice display increased latency to enter the center zone. The plot shows the latency to enter the center zone for the two genotypes (n shown in legend; mean ± SEM; Wilcoxon signed-rank test, W=567). D) Ddx3x^+/− mice have increased thigmotaxis. The plot shows the time spent in the center zone for the two genotypes (n shown in legend; mean ± SEM; Wilcoxon signed-rank test, W=1327). E) Ddx3x^+/− mice have increased activity in Y maze spontaneous alternation test. The plot shows the number of arms explored in a 15-min test for the two genotypes (n shown in legend; mean ± SEM; Student’s t test, t=2.9, df = 77). F) Short-term working memory is overall intact in Ddx3x^+/− mice in Y maze spontaneous alternation test. The plot shows the percentage of choices, broken down by choice type. ‘ABC’ indicates correct alternations, while ‘ABA’, ‘ABB’, ‘AAA’ errors (n shown in legend; mean ± SEM; ABA choice: Student’s t test, t=2.7, df=79). G) Ddx3x^+/− mice have no alterations in fear conditioning training. The plot shows the percentage of time spent freezing during training (n shown in legend). H) Ddx3x^+/− mice display weaker initial recall of contextual fear memory. After 24 hours from contextual and cued training (G), mice were exposed to the same environment of training, but without the administration of tones or shocks, for 240 sec. The plot shows the % of time the mice spent freezing during this 240 sec testing trial (n shown in legend; mean ± SEM; repeated measure ANOVA over the first 120 sec, Genotype: F=4.9, df=1). I) Ddx3x^+/− mice show no changes in cued fear memory. After 24 hrs from contextual testing (H), mice were exposed to a new environment, with administration of two tones without subsequent shock exposure. The plot shows the % of time the animals spent freezing during this 400 sec testing trial. J) Ddx3x^+/− mice spend more time with a novel object than a familiar one, as their control littermates. The plot shows the time mice spent interacting a novel object (triangles) or a familiar object (circles) during a 5 min testing session (n shown in legend; mean ± SEM; Ddx3x^+/+: Student’s t test, t=3.9, df=57; Ddx3x^+/−: Wilcoxon signed-rank test, W=436). All data collected and scored blind to genotype. For all tests, statistical testing was conducted after removing outliers (indicated as ⊗) assessing normality with the Shapiro-Wilk test. In all panels, Ddx3x^+/+ (blue) and Ddx3x^+/− (red). *p<0.05, **p<0.01, ***p<0.001; ns, non significant.

Figure 4.. Ddx3x^+/− mice have motor deficits, which attenuate in ageing mice previously exposed to behavioral training.

A) Adult Ddx3x^+/− mice have suboptimal motor performance on a rotarod test. Trials were conducted one hour apart with an acceleration over 5 minutes, and trials 1-3 were performed 24 hrs before trials 4-6, to assess both short-term (1 hour) and long-term (24 hrs) motor learning on 4-month old mice (4mo). The plot shows the latency to fall from the rod (n shown in legend; mean ± SEM; repeated measure ANOVA; Trial: F=52.9, df=5; Genotype: F=4.1, df=1; Interaction: F=0.79, df=5). B) Ageing Ddx3x^+/− mice have altered motor learning on a rotarod test. Trials were conducted one-hour apart with an acceleration over 5 min, with trials 1-4, 5-8 and 9-12 performed 24 hrs apart on one-year old mice (1 yr). The plot shows the latency to fall from the rod (n shown in legend; mean ± SEM; repeated measure ANOVA; Trial: F=125.1, df=11; Genotype: F=5.2, df=1; Interaction: F=4.8, df=11). C-D) Adult and ageing Ddx3x^+/− mice have impaired motor coordination in the balance beam test, but these changes are not observed in ageing mice previously exposed to behavioral training (pre 1yr). Panel C shows the number of frames covered (n shown in legend in panel D; mean ± SEM; 4mo: Welch’s t-test, t=4.04, df=22; 1yr: Welch’s t-test, t=3.24, df=37; pre 1yr: Welch’s t-test, t=0.6, df=29). Panel D shows the number of slips per frame (n shown in legend; mean ± SEM; 4mo: Welch’s t-test, t=4.13, df=11.7; 1yr: Wilcoxon signed-rank test, W=45.5). E) Ageing Ddx3x^+/− mice have age-dependent deficits in turning in the vertical pole test, but these changes are prevented by prior behavioral training (pre 1yr). The plot shows the mean time to turn on the top of the vertical pole (n shown in legend in panel F; mean ± SEM; 1yr: Welch’s t-test, t=3.6, df=22.6). F) Adult and ageing Ddx3x^+/− mice have altered balance during the vertical pole test, but these changes are not observed in ageing mice previously exposed to behavioral training (pre 1yr). The plot shows the mean time to descend the vertical pole (n shown in legend; mean ± SEM; 4mo: Student’s t test, t=3.3, df=20; 1yr: Welch’s t-test, t=3.1, df=19.2). G) Ddx3x^+/− mice have age-dependent changes in endurance in the wire hanging test. The plot shows the latency to fall (n shown in legend in panel H; mean ± SEM; 1yr: Wilcoxon signed-rank test, W=280). H) Ddx3x^+/− mice show changes in motor ability in the wire hanging test. The plot shows the number of frames covered by the mice (n shown in legend; mean ± SEM; 4mo: Wilcoxon signed-rank test, W=112). All data collected and scored blind to genotype. For all tests, statistical testing was conducted after removing outliers (indicated as ⊗) and assessing normality with the Shapiro-Wilk test. In all panels, Ddx3x^+/+ (blue) and Ddx3x^+/− (red). *p<0.05, **p<0.01, ***p<0.001; ns, non significant.

Figure 5.. Ddx3x^+/− mice have voxelwise differences in brain volume.

A) Ddx3x^+/− pups have reduced brain volume. Coronal flythrough from anterior (top row) to posterior (bottom row), showing brain anatomy (left column) and absolute differences in Ddx3x^+/− vs Ddx3x^+/+ mice (right column) at P3. The color coding indicates the false discovery rate (FDR) for regions that are smaller (blue gradient) or larger (red gradient) in Ddx3x^+/− compared to Ddx3x^+/+ mice (n=10/genotype). B) Changes in relative brain volume in Ddx3x^+/− pups. Same as in A, showing voxelwise changes relative to the overall reduction in brain volume (n=10/genotype). C) Specific brain regions are disproportionally reduced in Ddx3x^+/− pups. Same as in A, with the color coding indicating the percentage of difference between Ddx3x^+/− and Ddx3x^+/+ mice for regions that are disproportionally smaller (blue gradient) or larger (red gradient) that the ~10% reduction in overall brain volume (n=10/genotype). All data used to generate these plots, include individual-level data, are reported in Table S4. All data collected and analyzed blind to genotype.

Figure 6.. Ddx3x^+/− mice have altered cortical lamination.

A) Ddx3x^+/− mice show a misplacement of ScPN projection neurons in the developing cortex. Representative confocal images of coronal sections of primary motor cortex (M1) from Ddx3x^+/+ and Ddx3x^+/− mice at P3, immunostained for CTIP2 (red), a marker of ScPN and BRN1 (green), a marker of IT in UL and DL. As expected, CTIP2+ ScPN are restricted to layer V, while BRN1+ IT are predominantly in upper layers in control mice. B) CTIP2+ ScPN extend deeper in Ddx3x^+/− mice in the primary motor cortex. Distribution of the percentage of cells (DAPI+) that are positive to the ScPN marker CTIP2, across ten equally sized bins from the pia (bin 1) to the ventricle (bin 10) in Ddx3x^+/+ and Ddx3x^+/− mice (n shown in legend; 6-8 sections/mouse, with outliers across sections removed; mean ± SEM; Two-way ANOVA: Genotype: F=4, df=1; Bins: F=43.3, df=9; Bin 6: Student’s t test, t=2.3, df=10; Bin 7: Welch Two Sample t-test, t=4.3, df=4.1; Bin 9: Wilcoxon signed-rank test, W=2). C) BRN1+ IT neurons appear grossly normal in the motor cortex of Ddx3x^+/− mice. Distribution of the percentage of cells (DAPI+) that are positive to the IT marker BRN1, across ten equally sized bins from the pia (bin 1) to the ventricle (bin 10) in Ddx3x^+/+ and Ddx3x^+/− mice (n shown in legend; 6-8 sections/mouse, with outliers across sections removed; mean ± SEM; Two-way ANOVA: Bin: F=28.2, df=9). D) CTIP2+ and BRN1+ are mostly mutually exclusive, but not in Ddx3x^+/− mice. The panels show a magnification of the bins 5-7 regions from panel A. As expected, in control mice, a minority of CTIP2+ neurons is positive to BRN1. E) In Ddx3x^+/− mice, there are more CTIP2+BRN1+ neurons in deep layers. Distribution of the percentage of CTIP2+ ScPN that are also positive to BRN1, across bins 5-7, where CTIP2+ ScPN are most abundant (n shown in legend; 6-8 sections/mouse, with outliers across sections removed; mean ± SEM; Two-way ANOVA: Genotype: F=11, df=1; Bin 6: Student’s t test, t=2.5, df=10). F). Same as A, but from the primary somatosensory cortex (S1). G) There is an excess of CTIP2+ ScPN in Ddx3x^+/− mice in the primary somatosensory cortex. Same as B, but on S1. (n shown in legend; 6-8 sections/mouse, with outliers across sections removed; mean ± SEM; Two-way ANOVA: Bin: F=119.5, df=9; Interaction: F=3.9, df=9; Bin 5: Student’s t test, t=3.3, df=9). H) BRN1+ IT neurons appear grossly normal in the somatosensory cortex of Ddx3x^+/− mice. Same as C, but on the somatosensory cortex. (n shown in legend; 6-8 sections/mouse, with outliers across sections removed; mean ± SEM; Two-way ANOVA: Bin: F=108.2, df=9). I) CTIP2+BRN1+ neurons in S1. The panels show a magnification of the bins 5-7 regions from panel I. J) No changes in CTIP2+BRN1+ neurons in S1. Distribution of the percentage of CTIP2+ neurons that are also positive to BRN1, across bins 5-7, where CTIP2+ ScPN are most abundant (n shown in legend; 6-8 sections/mouse, with outliers across sections removed; mean ± SEM; Two-way ANOVA: Bin: F=99.8, df=2). In all panels, scale bar corresponds to 120 μm. All data collected and scored blind to genotype. For all tests, statistical testing was conducted after removing outliers and assessing normality with the Shapiro-Wilk test. In all panels, Ddx3x^+/+ (blue) and Ddx3x^+/− (red). *p<0.05, **p<0.01, ***p<0.001; ns, non significant.