Rapamycin prevents, but does not reverse, aberrant migration in Pten knockout neurons

Abstract

Phosphatase and tensin homolog (PTEN) is a major negative regulator of the Akt/mammalian target of rapamycin (MTOR) pathway. Mutations in PTEN have been found in a subset of individuals with autism and macrocephaly. Further, focal cortical dysplasia (FCD) has been observed in patients with PTEN mutations prompting us to examine the role of Pten in neuronal migration. The dentate gyrus of Pten(Flox/Flox) mice was injected with Cre- and non-Cre-expressing retroviral particles, which integrate into the dividing genome to birthdate cells. Control and Pten knockout (KO) cell position in the granule cell layer was quantified over time to reveal that Pten KO neurons exhibit an aberrant migratory phenotype beginning at 7.5days-post retroviral injection (DPI). We then assessed whether rapamycin, a mTor inhibitor, could prevent or reverse aberrant migration of granule cells. The preventative group received daily intraperitoneal (IP) injections of rapamycin from 3 to 14 DPI, before discrepancies in cell position have been established, while the reversal group received rapamycin afterward, from 14 to 24 DPI. We found that rapamycin prevented and reversed somal hypertrophy. However, rapamycin prevented, but did not reverse aberrant migration in Pten KO cells. We also find that altered migration occurs through mTorC1 and not mTorC2 activity. Together, these findings suggest a temporal window by which rapamycin can treat aberrant migration, and may have implications for the use of rapamycin to treat PTEN-mutation associated disorders.

Significance statement: Mutations in phosphatase and tensin homolog (PTEN) have been linked to a subset of individuals with autism and macrocephaly, as well as Cowden Syndrome and focal cortical dysplasia. Pten loss leads to neuronal hypertrophy, but the role of Pten in neuronal migration is unclear. Here we have shown that loss of Pten leads to aberrant migration, which can be prevented but not reversed by treatment with rapamycin, a mTor inhibitor. These results are important to consider as clinical trials are developed to examine rapamycin as a therapeutic for autism with PTEN mutations. Our findings show that some abnormalities cannot be reversed, and suggest the potential need for genetic screening and preventative treatment.

Keywords: Autism; Focal cortical dysplasia; Migration; Pten; Rapamycin.

Figure 1

Retrovirus-mediated knockout of Pten in newborn neurons of the neonatal mouse dentate gyrus. Retroviral particles containing mCherry and Cre linked by a T2A-motif (A) was co-injected with a GFP-expressing control retrovirus into the dentate gyrus of Pten^Flox/Flox mice on P7 (B). Pten loss was confirmed through immunohistochemical staining against Pten (C). The channels were merged to aid in the visualization of Cre-positive and Cre-negative neurons (D). Arrows denote Cre-positive cells that lack expression of Pten, while arrowheads denote Cre-negative cells. To examine a potential change in differentiation or mitosis following Pten loss, animals were injected with BrDU either just before viral injection, or 24 or 48 hours after. When BrDU is injected at the time of viral injection we find that ~ 20% of control (F; green) and Pten knockout (E; red) retrovirus labeled cells are co-labeled with BrDU (G; gray) There is no difference in the percentage of BrDU co-labeling between Pten KO and control neurons (H). By 24 and 48 hours post injection less than 4% of retrovirus labeled neurons were co-labeled with BrDU (not shown) indicating specificity of viral labeling for cells dividing on the day of viral injection.

Figure 2

Pten knockout neurons migrate farther into the dentate gyrus granule cell layer. The dentate gyrus of Pten^Flox/Flox mice was injected co-injected with a retrovirus expressing mCherry and Cre (red), to label and KO Pten, and one that expresses GFP only (green), to label controls. Mice were perfused at 7.5, 12.5, 16.5, 21.5, and 24.5 DPI to examine control vs Pten KO granule neurons. Arrows denote Pten KO neurons (red or yellow), asterisks denote control (only green). DAPI staining (turquoise) was used to visualize the granule cell layer (GCL). Newborn granule neurons migrate into the GCL in conjunction with the extension of dendrites through the molecular layer prior to 12.5 DPI (A). Scatter plots of the thickness of the GCL as measured at the same position of each control and Pten KO neuron indicates variability in GCL thickness across time and anatomical region (B; turquoise). Despite this variability, measurement of the raw distance from the hilus/GCL border into the GCL indicate that Pten KO neurons (B; red) migrate farther into the GCL compared to controls (B; white). To control for variability in GCL thickness the position Pten KO neurons (C; red) and control neurons (D; green) were expressed as a percentage of the thickness of the GCL as visualized by the DAPI stain (E; turquoise). Migration was quantified by measuring the distance to the center of the neurons from the hilar border (F) and expressed as a percentage of the thickness of the GCL at the position of each neuron measured (G). Pten KO neurons migrate farther from the hilus compared to controls beginning around 7.5 DPI (p<0.05), with a marked discrepancy in migration around 12.5 DPI (p<0.0001) and at every other time point following 12.5 DPI (p<0.0001; G). See Table I for quantitative values. (Statistical model described in methods; *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001)

Figure 3

Rapamycin-treated animals have decreased weight gain. GFP only and mCherry with Cre retroviruses were co-injected into the dentate gyrus of Pten^flox/flox mice on postnatal day 7 (P7). Rapamycin was injected prior to the establishment of the migratory defect (prevention group) via daily IP injections from 3–14 DPI (P10–21), then perfused on 14 DPI. Rapamycin was also injected after the establishment of the migratory defect (reversal group) via daily IP injections from 14–24 DPI (P21–31), then perfused on 24 DPI (A). Mice were weighed daily and the graphical values represent mean+/−SEM over time to find that rapamycin-treated mice grew less than vehicle-treated controls in the prevention group, although the change in weight was not significant considering all time points, pairwise comparisons reveal a divergence in weight beginning at 12 DPI (treatment p=0.0627, 12 DPI p<0.01, 13 DPI p<0.0001, 14 DPI p<0.0001, B). Rapamycin-treated mice trended towards decreased weight gain in the reversal group but this was not significant (treatment p=0.3852, C). For both growth curves, the number in parenthesis next to rapamycin and vehicle indicate number of animals and the number next to the brackets represent the overall statistical significance of rapamycin when all time-points are considered (two-way ANOVA). Stars represent significance at individual time points after pairwise comparison (Bonferroni’s; *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001).

Figure 4

Rapamycin can prevent and reverse somal hypertrophy, but can only prevent aberrant migration in Pten KO neurons. Representative images of prevention (3–14 DPI) and reversal (14–24 DPI) groups, arrowheads denote KO neurons from vehicle-treated animals, while arrows denote KO neurons from rapamycin-treated animals. Asterisks denote GFP-control cells, while yellow lines represent the GCL border (A). Quantification of soma size and migration in the prevention group reveals that KO cells from vehicle-treated animals were larger and migrated farther from the hilus compared to control cells and Pten KO cells from rapamycin-treated animals (p<0.0001, B; p<0.05, C). In the reversal group, KO cells from vehicle-treated animals were larger but showed no difference in migratory distance compared to KO cells from rapamycin-treated animals (p<0.0001, D; p=0.946, E). Comparisons of soma size and migration between prevention and reversal groups reveal rapamycin treatment could shrink KO cells, as KO cells from rapamycin-treated animals in the reversal group were smaller than KO cells from vehicle-treated animals in the prevention group (p<0.05, F), but it could not shrink KO cells to the size of KO somas from rapamycin-treated animals (p<0.0001, F). Rapamycin-treatment could prevent but not reverse aberrant migration, as KO neurons from rapamycin-treated animals in the reversal group migrate farther from the hilus compared to KO neurons from rapamycin-treated animals in the prevention group (p<0.001, G), but show no difference in migration compared to KO neurons from vehicle-treated animals in the reversal group (p=0.946, G). See Table I for quantitative results. (Statistical model described in methods; *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001)

Figure 5

Rapamycin effects aberrant migration through mTorC1 and not mTorC2. Tissue from vehicle- and rapamycin-treated animals in the prevention group was stained for p-S6 (Ser235/236), an indicator of mTorC1 activity (A), to reveal that Pten KO neurons from vehicle-treated animals have a greater normalized p-S6 intensity compared to control cells from vehicle-treated animals (B; p<0.0001), and KO cells from rapamycin-treated animals (B; p<0.0001). P-Akt (Ser473), an indicator of mTorC2 activity (C), revealed that Pten KO neurons from vehicle-treated animals have a greater normalized p-Akt intensity compared to control cells from vehicle-treated animals (D; p<0.0001), but show no significant difference in intensity as compared to KO cells from rapamycin-treated animals (D; p=0.89). Normalized intensity levels of phosphorylated-GSK3β (Ser9), an indicator of p-Akt activity (E), revealed that Pten KO neurons from vehicle-treated animals have a greater normalized p-GSK3β intensity compared to control cells from vehicle-treated animals (F; p<0.0001). However, KO cells from rapamycin-treated animals showed a modest decrease in p-GSK3β intensity as compared to KO cells from vehicle-treated animals (F; p=0.029; F). In A–C, arrowheads denote Pten KO neurons from vehicle-treated animals, arrows denote Pten KO neurons from rapamycin-treated animals, and asterisks denote control neurons See Table I for quantitative results. (Statistical model described in methods; *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001)