Pten regulates neuronal arborization and social interaction in mice

Abstract

CNS deletion of Pten in the mouse has revealed its roles in controlling cell size and number, thus providing compelling etiology for macrocephaly and Lhermitte-Duclos disease. PTEN mutations in individuals with autism spectrum disorders (ASD) have also been reported, although a causal link between PTEN and ASD remains unclear. In the present study, we deleted Pten in limited differentiated neuronal populations in the cerebral cortex and hippocampus of mice. Resulting mutant mice showed abnormal social interaction and exaggerated responses to sensory stimuli. We observed macrocephaly and neuronal hypertrophy, including hypertrophic and ectopic dendrites and axonal tracts with increased synapses. This abnormal morphology was associated with activation of the Akt/mTor/S6k pathway and inactivation of Gsk3beta. Thus, our data suggest that abnormal activation of the PI3K/AKT pathway in specific neuronal populations can underlie macrocephaly and behavioral abnormalities reminiscent of certain features of human ASD.

Figure 1. Pten Deletion Occurred in Differentiated Neurons

(A) In 2-week-old Nse-cre; Rosa26R mice, β-galactosidase signal was detected in MAP2-expressing neurons in the polymorphic layer (PML) and the outer granular layer (GL), close to the molecular layer (ML), in the dentate gyrus, but not in the inner GL, subgranular zone (SGZ), or in cells expressing nestin. Although a similar level and pattern of MAP2 or nestin immunoreactivity appeared in control (either without cre or Rosa26R), β-galactosidase was not detected (data not shown). Scale bar, 200 µm. (B) P15 Nse-cre; Rosa26R mice injected with BrdU at P14 retained the BrdU signal mainly in the SGZ, which was absent in β-galactosidase-expressing cells in the outer GL and PML. Four weeks after the BrdU injection, BrdU signal was detected in the GL, and some of them colocalized with β-galactosidase-expressing granule neurons (arrow, for example). Scale bar, 100 µm. (C) In mutant tissue at 2 weeks of age, similar to β-galactosidase activity in Nse-cre; Rosa26R brain, Pten-negative (blue), and P-Akt-positive signals (brown) were detected in the PML and the outer GL. At 4 weeks of age, the number of Pten-negative, P-Akt-positive cells increased in mutant dentate gyrus. Scale bar, 200 µm. (D) In the sensory cortex layers III to VI, Pten (brown) was detected in most neurons in control. In mutant cortex, Pten-negative cells were mainly detected at layers III to V. Scale bar, 100 µm.

Figure 2. The Pten Mutant Mice Were Abnormal in Behavioral Tests for Social Interaction and Social Learning

(A) At day 1, mutants spent significantly less time interacting with a conspecific juvenile compared to controls (n = 12). Control mice spent significantly less time interacting with the same juvenile 3 days hence (p < 0.05), yet mutants did not decrease their interaction time (p = 0.8). Legend in this panel applies to all bar graphs. (B) Mutant mice showed significant deficits in nest formation (n = 12). ANOVA revealed a significant effect of genotype (F_1,22 = 7.97, p = 0.01), time (F_3,66 = 11.37, p < 0.00001), and an interaction between genotype and time (F^3,66 = 6.26, p < 0.001). (C) Time spent interacting with a novel inanimate object under the same conditions as in (A) was not significantly affected by genotype (n = 12). (D) Mutant mice did not show significant difference from control in latency to find a buried treat following overnight food deprivation (n = 12). (E) When exposed to caged social and inanimate targets in an open field, controls showed normal preference for a social target over an inanimate target, while mutants spent similar time interacting with both targets (n = 12). Furthermore, mutant mice spent significantly less time interacting with a social target compared to controls (p < 0.00001). In this task, there was also a significant decrease in inanimate object interaction time between genotypes (p < 0.01), unlike in (C). (F) In a social preference task, mutants spent less time with a social target compared to controls (n = 12). Time spent with an inanimate object was not significantly different in both groups. (G) In a preference for social novelty task, controls showed a preference for social novelty, while mutants showed no preference between the social targets (n = 12). Mutants spent significantly less time interacting with a novel social target compared to controls. (H) Mutant female mice delivered normal-sized pups. Mortality of pups between P0 and P5 was significantly higher in mutants (n = 6) compared to controls (n = 8). *p < 0.05 to controls and < 0.005 to P0.

Figure 3. The Pten Mutant Mice Showed Abnormalities in Responses to Sensory Stimuli, Anxiety, and Learning

(A) Mutants exhibited increased locomotor activity in an open field test (n = 16 mutants, 17 controls). Average speed was 11.30 ± 0.84 cm/s for mutants and 8.51 ± 0.47 cm/s for controls (p = 0.006). Legend in this panel applies to all bar graphs. (B) Initial startle response was significantly increased in mutants (n = 11 mutants, 14 controls). Data represent the average startle response to the first six presentations of a 40 ms, 120 dB white noise stimulus. (C) In a prepulse inhibition test, mutants showed significantly impaired sensorimotor gating (n = 11 mutants, 14 controls). (D) Mutants exhibited anxiety-like behavior as they spent significantly less time in the center zone of the open field apparatus (n = 16 mutants, 17 controls). (E) The latency to enter the light side of the dark/light boxes was significantly elevated in mutants (n = 16 mutants, 17 controls). (F) In elevated plus maze test, mutants exhibited significantly increased duration in open arm (n = 16 mutants, 17 controls). (G) Mutants exhibited a significantly decreased learning curve in the submerged platform version of the water maze when latency to reach the platform was measured (n = 9 mutants, 12 controls). ANOVA revealed a main effect of genotype (F_1,17 = 11.17, p < 0.01) and day number (F_10,170 = 2.21, p < 0.05). Similar effect was seen in distance traveled to reach the platform (data not shown). Three mutant mice died during the water maze task due to seizure activity during training. (H) Controls showed clear preference for target quadrant versus opposite whereas mutants showed no preference (n = 9 mutants, 12 controls). (I) Mutant mice spent significantly increased time along the edge of the water maze (thigmotaxis, n = 9 mutants, 12 controls). ANOVA revealed a main effect of genotype (F_1,17 = 39.76, p < 0.00001).

Figure 4. Progressive Macrocephaly and Regional Hypertrophy in the Pten Mutant Mice

(A) Relative sizes of mutant brain to control at different ages indicate progressive macrocephaly in mutant mice (n = 4 or more per group). *p < 0.005 versus control. (B) A representative mutant brain at 10 months of age (right) is bigger than that from littermate control. Scale bar, 4 mm. (C) H/E staining on coronal sections shows that the thickness of the cerebral cortex (arrows), the length between the pial surface and the corpus callosum (CC), increased in adult mutant brain (upper panels). In the hippocampus, progressively enlarged dentate gyri (DG) and compressed or absence of CA1 were seen in mutant brains (lower panels). Scale bars, 200 µm.

Figure 5. Hypertrophic and Ectopic Axonal Tract with Increased Synapses in Pten-Deleted Dentate Gyrus

(A) Horizontal floating sections from 10-month-old brains were stained for synapsin I (red) and calbindin (green). Confocal images showed that elongated and dispersed mossy fiber tract from the granular layer (GL) of mutant dentate gyrus and an ectopic layer of axonal signals (arrows) in the molecular layer (ML), compared to those of control (upper panels). Scale bar, 500 µm. High-magnification images revealed that the mossy fiber synapses spanned a larger area in mutant versus control animals (lower panels; from the boxes in upper panels). (B) The inner ML of 7-month-old mutant dentate gyrus was positive for Timm staining (arrows), while such signals were absent in control ML. Scale bar, 100 µm. (C) Electron microscopic analysis of the inner ML (IML) of dentate gyri revealed that the increased axonal staining in mutant was due to enlarged presynaptic varicosities (red highlight). The varicosities of mutant contained a large number of densely packed synaptic vesicles. Scale bar, 0.5 µm.

Figure 6. Golgi Stain Revealed Dendritic Hypertrophy, Ectopy, and Increased Spine Density in Pten-Deleted Brain

(A) Thickened or elongate neuronal processes (arrow and arrow heads, respectively) were present in mutant cerebral cortex compare to control at 3 months of age. Scale bar, 100 µm. (B) At 8 months of age, increased length of the dendritic arbors in the molecular layer (ML) and ectopic neuronal processes (arrow) in the polymorphic layer (PML) were observed in mutant compare to control (upper panels). Reconstructions of single neurons made from image stacks emphasize the dendritic hypertrophy of the mutant neurons compare to control (lower panels). An axon can be seen emanating from both mutant and control neurons (arrowheads). Scale bar, 100 µm. (C) Mutant ML was significantly thicker than that of control at all adult ages tested (p < 0.05). (D) Higher-magnification images of dendrites in the ML revealed increased thickness and spine density in mutant (1.434 ± 0.064 spines/µm) versus control (1.077 ± 0.033 spines/µm) brains (p < 0.000005, n = 23 and 26 dendritic branches from three brains, respectively). Scale bar, 10 µm.

Figure 7. Ectopic Dendrites in Pten-Deleted Granule Neurons

(A) High-magnification images on Golgi-stained granule neurons in 8-month-old dentate gyri showed the presence of ectopic, spiny neuronal processes (arrows) in mutant, but not in control. Scale bar, 10 µm. (B) Coronal floating sections were stained for MAP2 (green) and P-Akt (red). Increased P-Akt was apparent in the granular (GL) and molecular layers (ML) of mutant dentate gyri at all ages, but not in controls. Mutant dentate gyrus at 10 months of age had an ectopic layer of dendrites (arrow heads) between the GL and polymorphic layer (PML), which was absent at 4 weeks of age. Scale bar, 200 µm. (C) Higher-magnification confocal images revealed that the ectopic dendrites were from P-Akt-positive granule neurons (light blue arrows, for example), but not from P-Akt-negative neurons (white arrows, for example). Scale bar, 50 µm.

Figure 8. Molecular Signaling Downstream of Pten-Deleted Neurons

(A) At 2 months of age, mutant dentate gyri exhibited increased signal (brown) for P-Akt, P-S6 and P-Gsk3β, compared to those in control. The increased staining in mutants was apparent in the granular layer (GL) and inner molecular layer (ML) for P-S6 and in the GL for PGsk3β, whereas both the GL and ML displayed increased P-Akt. All sections were counterstained with hematoxylin, except for P-Gsk3β panels that were stained with methyl green. Scale bar, 200 µm. (B) Western blot analysis showed decreased Pten and increased P-Akt, P-S6, P-Gsk3β, and P-Tuberin in mutant versus control at all ages tested. p < 0.005 for Pten at 2 months and P-Akt; p < 0.1 for P-S6, P-Gsk3β and P-Tuberin.