Neurolucida reconstruction of a BrdU labeled, GFP-expressing adult-generated hippocampal granule cell exposed to status epilepticus when the cell was one-week-old. The general locations of the dentate granule cell layer, inner molecular layer (IML), middle molecular layer (MML) and outer molecular layer (OML) are shown. Confocal reconstructions of portions of the dendritic tree from each of the three molecular layers regions are shown on the left. Afferent input to the three layers is denoted on the right. Scale bar = 20 μm for the reconstruction and 10 μm for the confocal images.

Neurolucida reconstructions of newborn granule cells from Thy1-GFP expressing control animals (top row) and animals treated with pilocarpine to induce status epilepticus. All cells are 12-weeks-old, and cells from epileptic animals were exposed to status when these cells were one-week-old. Apical dendrites are shown in black, basal dendrites in blue, and axons in green. Yellow lines denote the borders of the granule cell body layer and the hippocampal fissure. While many granule cells from epileptic animals exhibited the normal “fanlike” dendrite trees (boxed cells) typical of cells from control animals, a subset of cells exhibited trees that were heavily lateralized, giving them a “windswept” appearance or trees that failed to spread within the molecular layer (“closed parasol deformation”). Finally, several cells exhibited irregular dendritic trees (a.) that were difficult to characterize, but included abnormalities such as a recurrent basal dendrites (arrowheads), crossing dendrites (blue arrow) and dendrites that projected outside of the molecular layer (red arrow). Red asterisks denote granule cells with high spine densities. Scale bar = 100 μm.

Neurolucida reconstructions of molecular layer border (MLB) granule cells from Thy1-GFP expressing control animals (top row) and animals treated with pilocarpine to induce status epilepticus (bottom row). Apical dendrites are shown in black. Yellow lines denote the molecular layer/granule cell body layer border and the hippocampal fissure. Cells from epileptic animals are grossly normal, lacking the overt structural abnormalities evident among newborn cells exposed to status. Scale bar = 100 μm.

Confocal reconstructions of dentate granule cell layer (DGCL) and inner (IML), middle (MML) and outer (OML) molecular layer dendrites of 12-week-old BrdU-labeled, GFP-expressing granule cells from control and epileptic (SE) animals. Cells from epileptic animals exhibited lower spine densities in all regions except the DGCL. Scale bar = 5 μm.

Estimated spine numbers (spine density X estimated AD length) for BrdU-labeled, GFP-expressing 12-week-old granule cells from control animals (gray boxes) and epileptic animals (SE, blue boxes). Granule cells from epileptic animals exhibited significant reductions in the spine number in the inner (IML) and middle (MML) molecular layers relative to controls (A). Total spine number was also significantly reduced (B). Box plotes shows 5^th and 95^th percentiles, median lines and outliers. *P<0.05 vs. control.

A & B: Three dimensional scatter plots showing relationships between estimated (EST.) inner molecular layer (IML) spine number, soma area and basal dendrite length. For 12-week-old granule cells from control animals (A) regression analyses indicated that soma area was not predictive of spine number (and basal dendrites were absent, so all length values =0). By contrast, both soma area and basal dendrite length predicted IML spine density in epileptic animals, although notably, the former two variables did not always co-segregate. C & D: Two-dimensional scatter plots showing the relationships between estimated total spine number and soma area indicate that the two variables are not correlated in control animals (P=0.548; Spearman rank order) and that the cells tend to cluster together. Age-matched granule cells from epileptic animals, on the other hand, show a significant positive correlation between soma area and total spine number (P=0.029). The resulting pattern strongly suggests the existence of two distinct groups: the majority with low spine densities and smaller somas and the minority (approximately 10%) with higher spine densities and larger somas. F: Confocal reconstruction showing a middle molecular layer segment from the highlighted cell in D. This cell, from an epileptic animal, had the greatest number of spines of any cell in the study. Scale bar = 5 μm.

A–F: Confocal reconstructions of GFP-expressing newborn granule cells from Gli1-CreER^T2 X GFP reporter mice. A: Confocal reconstruction of a newly-generated granule cell from a control animal with typical apical dendritic structure. B: Reconstruction of a “windswept” GFP-expressing granule cell (orange) from an epileptic animal. Adjacent GFP expressing granule cells are shown in green, while Nissl staining is shown in blue. C: Newly-generated granule cell from an epileptic animal exhibiting the “closed parasol” deformation. D: Hypertrophied granule cell (arrow) from an epileptic animal co-immunostained with GFP (orange) and the granule cell specific marker Prox1 (pink). Pink asterisks denote granule cells labeled with Prox1 in a focal plane not included in the image composite. Granule cells single-labeled for Prox1 are shown in green. E: Co-immunostaining with GFP and ZnT3 (cyan) showing a hypertrophied granule cell (arrow) and two typically-sized neighbors. F: example of numerous typically-sized granule cells adjacent to a hypertrophied granule cell (arrow). Scale bars = 30 μm.

A: Hippocampus from a control Thy1-GFP expressing mouse co-immunostained for ZnT-3 (A), which reveals mossy fiber terminals, and GFP (A.1 shows merged images). In control animals, ZnT-3 staining is most prominent in the mossy fiber pathway (hilus and stratum lucidum). Three months after status epilepticus, staining is also present in the inner molecular layer (B, B.1), indicative of mossy fiber sprouting (arrowheads in B). C: Control Gli1-CreER^T2 X GFP reporter mouse co-immunostained with GFP and ZnT-3. Newly-generated granule cells exhibit typical morphology and positioning, and mossy fiber sprouting is absent. Layer denotations in C (IML, inner molecular layer; MML, middle molecular layer; DGC-L, dentate granule cell layer) apply to D as well. D: Epileptic Gli1-CreER^T2 X GFP reporter mouse showing GFP-expressing granule cells generated after status. Note the appearance of cells with basal dendrites and ectopically located somata. ZnT-3 labeling further reveals mossy fiber sprouting in the inner molecular layer. E, F: Examples of sprouted mossy fiber axons with terminals in the IML. Insets show double-labeling of the terminals with ZnT-3. Scale bars = 200 μm (A,B), 60 μm (C,D) and 10 μm (E,F).

Confocal images of GFP-expressing apical dendrite segments from the inner molecular layer. Sections are co-stained for ZnT-3, which reveals mossy fiber sprouting in these epileptic animals. Dendrites pictured belong to granule cells with basal dendrites. Examples of dendritic spines apposed to ZnT-3 immunoreactive puncta are shown (arrows). Scale bar = 5 μm.

A: Estimated (EST.) number of inner molecular layer (IML) spines possessed by GFP-expressing newborn granule cells from control and epileptic Gli1-CreER^T2 X GFP expressing mice treated with tamoxifen three months earlier. For epileptic animals (SE), cells were further subdivided into those with (BD+) and without (no BD) basal dendrites. Spine density was significantly reduced for cells lacking basal dendrites relative to all other groups, while cells with long basal dendrites were statistically equivalent to controls. **, P<0.01 relative to all other groups. B: Cells lacking basal dendrites possessed significantly more ZnT-3 apposed spines relative to control cells, and cells with basal dendrites had significantly more spines relative to all other groups. ***, P<0.001 relative to all other groups.