Pten haploinsufficiency causes desynchronized growth of brain areas involved in sensory processing

Abstract

How changes in brain scaling relate to altered behavior is an important question in neurodevelopmental disorder research. Mice with germline Pten haploinsufficiency (Pten ^+/-) closely mirror the abnormal brain scaling and behavioral deficits seen in humans with macrocephaly/autism syndrome, which is caused by PTEN mutations. We explored whether deviation from normal patterns of growth can predict behavioral abnormalities. Brain regions associated with sensory processing (e.g., pons and inferior colliculus) had the biggest deviations from expected volume. While Pten ^+/- mice showed little or no abnormal behavior on most assays, both sexes showed sensory deficits, including impaired sensorimotor gating and hyporeactivity to high-intensity stimuli. Developmental analysis of this phenotype showed sexual dimorphism for hyporeactivity. Mapping behavioral phenotypes of Pten ^+/- mice onto relevant brain regions suggested abnormal behavior is likely when associated with relatively enlarged brain regions, while unchanged or relatively decreased brain regions have little predictive value.

Keywords: Behavioral neuroscience; Biological sciences; Imaging anatomy; Neuroscience; Small animal imaging.

Graphical abstract

Figure 1

Pten^+/- mice have abnormal developmental brain region growth trajectories (A) Correlations between brain region growth indices {[(P60 volume for mouse) – (average P7 volume for genotype)]/(average P7 volume for genotype)} within genotypes are present in both genotypes in some regions (green) but not others (purple). (B–E) Pten^+/- mice have abnormal brain region growth trajectories. We calculated the growth trajectory (B) based on the absolute and relative (percent of total brain) volume for each brain region [(mean_{P60 WT region volume})/(mean_{P7 WT region volume})], and extrapolated the predicted size of each brain region in Pten^+/- mice based on this trajectory [(mean_{P7Pten+/- region volume}) x (WT growth trajectory)]. We then calculated the percent deviation from this predicted size for each mouse {[(region volume) – (predicted region volume)]/(predicted region volume) x 100}. These calculations were performed on absolute (D) and relative (C and E) brain region volumes. (C) Overview of significant Pten^+/- deviations from predicted relative volume. Red regions are relatively increased, blue regions are relatively decreased, and color intensity indicates the degree of deviation. Cortex is translucent to enable visualization of subcortical structures. (D and E) Pten^+/- deviation from absolute (D) and relative (E) predicted brain region volume. P7, postnatal day 7; P60, postnatal day 60. The Scalable Brain Atlas (https://scalablebrainatlas.incf.org/composer/?template=ABA_v3; Bakker et al., 2015) was used to make images in (C). Data are represented as mean ± SEM. Black symbols, one-sample t-tests vs. expected volume. ∗∗∗ p < 0.001, ∗∗ p < 0.01, ∗ p < 0.05, + p < 0.10. See also Figure S1, Tables S1 and S2.

Figure 2

Pten^+/- mice of both sexes are hyporeactive to acoustic stimuli of high intensity; female Pten^+/- mice are hypersensitive to acoustic stimuli of low intensity (A) Expected startle responses to white noise stimuli of different intensities above 70 dB background. (B and C) Female Pten^+/- mice show lower startle thresholds and increased startle amplitude in the low-dB module (B), but lower startle amplitude to stimuli in the high-dB module (C). (D and E) Pten^+/- males also show decreased startle amplitude to stimuli in the high-dB module (E), with limited evidence of lower startle thresholds and increased startle amplitude in the low-dB module (D). (F–N) Fitting sigmoid curves to each mouse and analyzing the 5% startle threshold (5% of saturation), midpoint (50% of saturation), maximum startle amplitude, and slope of the sigmoid function revealed that Pten^+/- females (G) but not males (K) show a lower startle threshold, and both sexes of Pten^+/- mice show lower maximum startle responses (I and M). No genotype differences were found in either sex for midpoint (H and L) or slope (J and N). Data are represented as mean ± SEM. Black symbols, independent-samples t-tests between genotypes. Colored symbols, significant startle (paired-samples t-tests vs. 0 dB above background). ∗∗∗ p < 0.001, ∗∗ p < 0.01, ∗ p < 0.05, + p < 0.10. Main effect of genotype in two-way mixed-model ANOVAs (genotype x stimulus dB): @@ p < 0.01, @ p < 0.05. See also Figure S2, Table S3.

Figure 3

Pten^+/- mice have lower startle amplitude and decreased pre-pulse inhibition (A and D) Expected startle responses to 120-dB white noise stimuli following repeated presentations (A) and within pre-pulse trials (D). (B and C) Both female (B) and male (C) Pten^+/- mice startle less than Pten^+/+ mice to a 120-dB white noise pulse during the pre-pulse inhibition (PPI) assay. Neither Pten^+/+ nor Pten^+/- females show significant habituation [(phase I startle amplitude – phase III startle amplitude)/(phase III startle amplitude) x 100] to the stimulus (B), while both Pten^+/+ and Pten^+/- males do (C). Phase I, average of 6 stimulus-only presentations prior to PPI phase; phase II, average of 12 stimulus-only presentations during PPI phase; phase III, average of 6 stimulus-only presentations following PPI phase. (E–F) Pten^+/- mice of both sexes have impaired PPI [(phase II startle amplitude – pre-pulse startle amplitude)/(phase II startle amplitude) x 100], showing decreased inhibition of the startle response following a pre-pulse acoustic stimulus. Data are represented as mean ± SEM. Black symbols, independent-samples t-tests between genotypes. Colored symbols, significant habituation (one-sample t-tests vs. 0). ∗∗ p < 0.01, ∗ p < 0.05, + p < 0.10. See also Table S3.

Figure 4

Pten^+/- females show altered cognitive behavior (A) Schematic of the novel object recognition task. (B and C) Female Pten^+/+ mice (B), and males of both genotypes (C), prefer the novel object during the test phase, while Pten^+/- females do not (B). Percent time investigating (each object) = {[(time investigating novel or familiar object)]/[(time investigating novel object) + (time investigating familiar object)] x 100}. Discrimination index = {[(time investigating novel object) – (time investigating familiar object)]/[(time investigating novel object) + (time investigating familiar object)] x 100}. (D and E) Pten^+/- female mice investigate the sample objects less than Pten^+/+ females during the first sample phase (D), while male Pten^+/- mice spend less of the trial investigating the sample objects across the three phases (E). Nov, novel object; fam, familiar object; S1, sample phase 1; S2, sample phase 2. (F) Examples of the puzzle box conditions. Conditions 1–4 (C1–C4) were each presented three times, while C0 was used for the first and last trials of the experiment. (G and H) Female Pten^+/- mice performed significantly better on the tissue task (G), but no other genotype differences were observed in either sex (G–H). Data are represented as mean ± SEM. Black symbols, paired-samples t-tests between novel and familiar objects within groups (B–C) or independent-samples t-tests between genotypes (D–E). Black symbols over lines, main effect of genotype in two-way mixed-model ANOVAs (genotype x phase, D-E). Colored symbols (D–E), change in investigation over phases (one-way within-subjects ANOVAs). Dashed lines (G–H), maximum time to complete task. ∗∗∗ p < 0.001, ∗ p < 0.05, + p < 0.10. See also Figure S3, Tables S4 and S9.

Figure 5

Pten^+/- mice were normal on the majority of behavioral assays (A and B) Latency to find platform in Morris water maze improved over acquisition and reversal training in Pten^+/+ and Pten^+/- female (A) and male (B) mice. VP, visual platform test. Dashed line, acquisition or reversal criterion. (C and D) All female (C) and male (D) mice improved their latency to fall across testing days in the rotarod learning assay. (E and F) Female (E) and male (F) mice of both genotypes showed a reduction in the number of reaches to criterion across training in the single-seed reaching task. T1-T7, training days 1–7. (G–L) Normal fear conditioning, including weak trace fear conditioning (G–H), remote memory for trace fear conditioning (I–J), and cued fear conditioning extinction (K–L), was observed in female (G,I, and K) and male (H,J, and L) Pten^+/- mice, except that Pten^+/- males showed impaired remote contextual memory (J). Ext bin, average of 6 extinction trials. Data are represented as mean ± SEM. Black symbols, main effect of genotype in two-way mixed-model ANOVAs (genotype x training day, A–B), independent-samples t-tests between genotypes (C–D, G–L), or Sidak post hoc tests between genotypes from two-way mixed-model ANOVAs (genotype x training day, E-F). Colored symbols, change over time (one-way within-subjects ANOVAs, C–F, or paired-samples t-tests, G–L). ∗∗∗ p < 0.001, ∗ p < 0.05, + p < 0.10. See also Figures S4–S6, Tables S4, S5 and S6.

Figure 6

Male and female juvenile Pten^+/- mice show different aspects of adult acoustic stimuli threshold phenotypes (A–C) Female Pten^+/- mice show hypersensitivity in the low-dB module as juveniles [A(i)) and adults (C(i)], with limited differences as adolescents [postnatal day 45, B(i)]. In the high-dB module, Pten^+/- females only show hyporeactivity in adulthood [C(ii)], not as juveniles [A(ii)] or adolescents [B(ii)]. (D–F) Pten^+/- males express hyporeactivity in the high-dB module throughout development [juveniles, D(ii); adolescents, E(ii); adults, F(ii)], but only show limited evidence of hypersensitivity in the low-dB module as adults [F(i)], with no genotype differences as juveniles [D(i)] or adolescents [E(i)]. (G–L) Fitting sigmoid curves to each mouse and analyzing the 5% startle threshold (5% of saturation), midpoint (50% of saturation), maximum startle amplitude, and slope of the sigmoid revealed that only adult Pten^+/- females showed significantly lower 5% startle threshold [I(i)] and maximum startle amplitude [I(iii)]. No other significant genotype differences were observed in juvenile (G), adolescent (H), or adult (I) female mice, although there was a trend to a lower sigmoid slope in the adolescent Pten^+/- females [H(iv)]. Male Pten^+/- mice showed decreased maximum startle amplitude at all ages [juvenile, J(iii); adolescent, K(iii); adult, L(iii)], but no other genotype differences. Data are represented as mean ± SEM. Black symbols, independent-samples t-tests between genotypes. Colored symbols, significant startle (paired-samples t-tests vs. 0 dB above background). ∗∗∗ p < 0.001, ∗∗ p < 0.01, ∗ p < 0.05, + p < 0.10. Main effect of genotype in two-way mixed-model ANOVAs (genotype x stimulus dB): @@ p < 0.01, @ p < 0.05, # p < 0.10. See also Figure S8, Table S7.

Figure 7

Female Pten^+/- mice show social deficits from P45, while males show normal social approach behavior throughout development (A–H) All groups show significant social preferences as juveniles (females, A; males, E) and at postnatal day 35 (P35; females, B; males, F). Male mice of both genotypes also show significant social preferences at P45 (G) and in adulthood (H). Female Pten^+/-, but not Pten^+/+, mice fail to prefer the social chamber at P45 (C) and in adulthood (D). (I and J) The preference index, which assesses the strength of the preference for the social chamber [(mouse + tube – empty tube)/(mouse + tube + empty tube) x 100], only differed in the females at P45, with a decrease in the Pten^+/- mice (I). (K and L) Other than a trend to increased distance traveled in the juvenile Pten^+/- males (L), no genotype differences were found in locomotion, although all groups showed increased distance traveled across time (females, K; males, L). M, chamber containing mouse + tube; C, center chamber; E, chamber containing empty tube. Heatmaps in A–H are averaged across groups. Data are represented as mean ± SEM. Black symbols, significant preference (paired-samples t-tests between mouse + tube and empty tube chambers, A–H) or independent-samples t-tests between genotypes (I-L). Colored symbols, effect of age (one-way within-subjects ANOVAs). ∗∗∗ p < 0.001, ∗∗ p < 0.01, ∗ p < 0.05, + p < 0.10. See also Table S7.