Somatostatin interneuron fate-mapping and structure in a Pten knockout model of epilepsy

Abstract

Disruption of inhibitory interneurons is common in the epileptic brain and is hypothesized to play a pivotal role in epileptogenesis. Abrupt disruption and loss of interneurons is well-characterized in status epilepticus models of epilepsy, however, status epilepticus is a relatively rare cause of epilepsy in humans. How interneuron disruption evolves in other forms of epilepsy is less clear. Here, we explored how somatostatin (SST) interneuron disruption evolves in quadruple transgenic Gli1-CreER^T2, Pten^fl/fl, SST-FlpO, and frt-eGFP mice. In these animals, epilepsy develops following deletion of the mammalian target of rapamycin (mTOR) negative regulator phosphatase and tensin homolog (Pten) from a subset of dentate granule cells, while downstream Pten-expressing SST neurons are fate-mapped with green fluorescent protein (GFP). The model captures the genetic complexity of human mTORopathies, in which mutations can be restricted to excitatory neuron lineages, implying that interneuron involvement is later developing and secondary. In dentate granule cell (DGC)-Pten knockouts (KOs), the density of fate-mapped SST neurons was reduced in the hippocampus, but their molecular phenotype was unchanged, with similar percentages of GFP+ cells immunoreactive for SST and parvalbumin (PV). Surviving SST neurons in the dentate gyrus had larger somas, and the density of GFP+ processes in the dentate molecular layer was unchanged despite SST cell loss and expansion of the molecular layer, implying compensatory sprouting of surviving cells. The density of Znt3-immunolabeled puncta, a marker of granule cell presynaptic terminals, apposed to GFP+ processes in the hilus was increased, suggesting enhanced granule cell input to SST neurons. Finally, the percentage of GFP+ cells that were FosB positive was significantly increased, implying that surviving SST neurons are more active. Together, findings suggest that somatostatin-expressing interneurons exhibit a combination of pathological (cell loss) and adaptive (growth) responses to hyperexcitability and seizures driven by upstream Pten KO excitatory granule cells.

Keywords: epilepsy; mTORopathies; mammalian target of rapamycin (mTOR); morphology; phosphatase and tensin homolog (Pten); somatostatin interneurons; transgenic mice.

FIGURE 1

Pten knockout from dentate granule cells. (A) Confocal images showing the dentate granule cell layer and hilus in control and DGC-Pten KO mice. Knockout cells are identified by lack of Pten immunostaining, evident as a dark blue band of Nuclear Blue counterstained neurons concentrated predominantly along the hilar-granule cell layer border. Examples of KO cells are indicated by the arrowheads (white). Scale, 50 μm. (B) Percentage of KO dentate granule cells for each DGC-Pten KO mouse. Bars represent animal means ± SEM.

FIGURE 2

Reduced SST interneuron density in the hippocampus. (A) Confocal images of hippocampus showing the distribution of eGFP+ cells in Gli1-CreER^{T 2–/–}, Pten^{fl /fl}, SST-FlpO^+/–, frt-eGFP^+/– (control) and Gli1-CreER^{T 2+/–}, Pten^{fl /fl}, SST-FlpO^+/–, frt-eGFP^+/– (DGC-Pten KO) mice. Scale, 200 μm. (B) eGFP+ cells within the hilus illustrating reduced cellular density in DGC-Pten KO mice. Scale, 100 μm. (C–F) Density of eGFP-SST interneurons in the hippocampus, hilus, CA1, and CA3, respectively. (G–J) Areas of the hippocampus, hilus, CA1, and CA3, respectively. *p < 0.05. Bars represent animal means ± SEM.

FIGURE 3

Expression of 3-nitrotyrosine and cleaved caspase-3 is similar between control and DGC-Pten KO mice. (A) Representative images of 3-NT immunostaining of hilar eGFP-SST interneurons in control and DGC-Pten KO groups. Scale, 50 μm. (B) The percentage of 3-NT+ hilar eGFP-SST interneurons did not differ significantly between groups. (C) Representative images of cleaved caspase-3 (CC-3) immunostaining of the dentate gyrus in control and DGC-Pten KO groups. The arrowhead denotes a CC-3+, GFP- cell in the dentate gyrus granule cell layer. Scale, 50 μm. Bars represent animal means ± SEM.

FIGURE 4

Preserved molecular phenotype of SST neurons in DGC-Pten KO mice. (A) Representative images of the hilus showing overlap between eGFP-labeled cells and SST immunostaining. Scale, 50 μm. (B) The percentage of hilar SST-immunoreactive cells that co-express eGFP. (C) Graph shows the percentage of hilar eGFP+ cells that were immunopositive for SST. (D) Representative images of the hilus showing overlap between eGFP-labeled cells and PV immunostaining. Scale, 50 μm. (E) The percentage of hilar PV+ cells that co-express eGFP. (F) The percentage of hilar eGFP+ cells that also express PV. Bars represent animal means ± SEM.

FIGURE 5

Hilar SST interneurons undergo morphological changes in DGC-Pten KO mice. (A) Representative images showing hilar eGFP-SST interneuron (HIPP) cell bodies and dendrites in control and DGC-Pten KO mice. Scale, 10 μm. (B) Graph illustrating the significant increase in eGFP-SST soma area in DGC-Pten KO mice. (C) Scatterplot showing the soma area for each reconstructed eGFP-SST interneuron in control (118 cells) and DGC-Pten KO (153 cells) mice. The red lines depict the mean soma areas. (D) Average soma roundness did not differ significantly between groups. (E) The number of primary dendrites per eGFP-SST interneuron did not differ significantly between groups. (F) Average soma area was positively correlated with the percentage of Pten KO DGCs in each mouse (n = 6 DGC-Pten KO mice). (G) Average hilar SST cell density was negatively correlated with the percentage of Pten KO DGCs in each mouse (n = 6 DGC-Pten KO mice). *p < 0.05. Bars represent animal means ± SEM.

FIGURE 6

Increased apposition between Znt3 puncta and eGFP-SST interneuron processes. (A) Depth coded images of the hilus showing the area coverage of eGFP-SST processes. Scale, 20 μm. (B) Bar graph showing similar area coverage of eGFP-SST processes. (C) Representative images showing Znt3 puncta (purple) and eGFP-SST neurons in the dentate hilus. The lower panel shows magnified images from the regions in the upper panel indicated by white boxes. Arrowheads (white) indicate examples of apposition between Znt3 puncta and eGFP-SST processes. Scale, 10 μm. (D) Bar graph shows the significant increase in apposition of Znt3 puncta to eGFP-SST interneuron processes and cell bodies in DGC-Pten KO mice. **p < 0.01. Bars represent animal means ± SEM.

FIGURE 7

eGFP-SST interneuron processes density is maintained in DGC-Pten KO mice. (A) Representative images of the hilus (H), dentate granule cell body layer (DGCL), and dentate inner (IML), middle (MML), and outer molecular layers (OML) in control and DGC-Pten KO (PKO) mice. Scale, 20 μm. (B) Bar graph showing the area occupied by eGFP-SST interneuron processes in the IML, MML, and OML, respectively. (C) Bar graph depicting the significant increase in molecular layer area among DGC-Pten KO mice relative to controls. *p < 0.05, **p < 0.01, ****p < 0.0001. Bars represent animal means ± SEM.

FIGURE 8

Altered Fos labeling of eGFP-SST interneurons and dentate granule cells in DGC-Pten KO mice. (A) c-Fos and eGFP-SST labeling in the dentate gyrus. Scale, 100 μm. (B) Percentage of hilar eGFP-SST interneurons with c-Fos immunostaining. (C) Density of c-Fos+ dentate granule cells. (D) FosB and eGFP-SST labeling in the dentate gyrus. Scale, 100 μm. (E) Percentage of hilar eGFP-SST interneurons immunoreactive for FosB. (F) Density of FosB+ dentate granule cells. *p < 0.05, **p < 0.01. Bars represent animal means ± SEM.