Progressive development of the dentate neurogenic zones is revealed by Nestin-GFP transgene and Tbr2. (A) At E14.5, Nestin-GFP is upregulated in the dentate primordium abutting the cortical hem, which is shown in higher magnification in (A1), with a few Nestin-GFP+ cells leaving the dentate VZ. (A2) Tbr2+ cells, which represent neurogenic precursors, have migrated away from the SVZ (arrowheads). (A3) The fimbria has started to form in the relatively cell-free region of the cortical hem (shown by nuclei staining with DAPI). (A4) The organization of the medial wall with all the markers. Scale bar: 200 μm in A; 50 μm in A1-A4. (B) At E15.5, Nestin-GFP+ cells (green) formed a new neurogenic zone in the subpial region of the fimbriodentate junction (FDJ). (B1-B4) The distribution of Nestin-GFP+ and Tbr2+ cells in the subpial region of the FDJ. Some of the Nestin-GFP+ cells (arrowheads in B1 and B2) were Tbr2- (arrowheads in B3 and B4). Scale bar: 100 μm in B; 20 μm in B1-B4. (C) At E17.5, Nestin-GFP+ cells were distributed across the hilus. (C1-C3) The distribution of Nestin-GFP+ cells (C1,C3), Tbr2+ cells (red, C2,C3) and DAPI (blue, C3) at higher magnification. Scale bar: 200 μm in C; 50 μm in C1-C3. (D) At P0, the subpial neurogenic zone was established around the edge of the dentate pole. (D1-D3) Both Nestin-GFP+ and Tbr2+ cells were largely subpially localized in the forming dentate pole (arrowheads). Scale bar: 200 μm in D; 100 μm in D1-D3. (E) At P2, the subgranular zone (SGZ) started to take shape. (E1-E3) Nestin-GFP+ cells began to seed the nascent SGZ with the largest population being adjacent to the pia, whereas most Tbr2+ cells were distributed in the ML. Scale bar: 200 μm in E; 50 μm in E1-E3. (F) By P7, the permanent neurogenic niche was formed in the subgranular zone. (F1-F3) Most Nestin-GFP+ and some Tbr2+ cells were present in the SGZ. Scale bar: 200 μm in F; 50 μm in F1-F3. (G1-G4) A schematic representation of the two stages involved in neurogenic zone relocation during the development of the dentate gyrus. Green dots indicate the location of stem/progenitor cells as labeled by Nestin-GFP transgene. D, dentate; DN, dentate notch; F, fimbria; FDJ, fimbriodentate junction; GCL, granule cell layer; Hem, cortical hem; Hip, hippocampus; HF, hippocampal fissure; ML, molecular layer; SGZ, subgranular zone; SVZ, subventricular zone; VZ, ventricular zone.

Development of the transhilar glial scaffolding proceeds without evidence of VZ connection. (A) By E18.5, a distinct upper blade was formed as labeled by granule cell marker Prox1. The GFAP+ glial scaffolding was highly concentrated at the border of the fimbria and extended into FDJ. (B,C) GFAP+ glial orientation was correlated with the granule cell arrangement in the forming dentate plate. At the entrance of the hilus (red arrow in B), GFAP+ radial glia radiated out into HF, whereas most Prox1+ granule cells were gathered towards HF. Scale bar: 50 μm in B and C. (D) Labeling radial glia in the ventricular zone by DiO injected into the ventricle at E18.5 revealed that no radial glial fibers directly projected from the VZ into the forming dentate. (E-H) The dynamic changes of radial glial projection in the developing dentate by in utero electroporation (EP) of GFP expression vector targeting the medial wall. (E) Examined at E14.5 after EP at E13.5, GFP labeled radial glia at the dentate VZ showed the association of the endfeet with pia (arrowheads). (F) Examined at E15.5 after EP at E14.5, GFP labeled radial glia no longer extended into the field of the dentate gyrus, in contrast to the GFP labeled glial fibers radially oriented elsewhere in the hippocampal fields. A few cells (arrows) migrated toward the dentate gyrus. (G) Forty-eight hours after EP at E17.5, GFP marked the VZ of the whole medial wall. Radial glial processes in the hippocampal field were clearly labeled but not in the dentate gyrus. GFP labeled a stream of cells along the edge of the fimbria (arrows). Some of them already reached the entrance of hilus, which are shown at higher power in H. (I-K) The ongoing relocation of BLBP+ glia at E18.5. BLBP+ glial processes were revealed by thin section confocal imaging (I). BLBP+ cell bodies (arrow) were seen in the FDJ (inset in I). z-projection of image stacks revealed the long processes spanning from FDJ to HF. The location of individual cell bodies (red and yellow arrows in J) was identified through tracing serial sections. A few glial cell bodies already arrived at the HF (white arrows in J). (K) The two long processes and corresponding cell bodies (arrows), in addition to the glial cell bodies (in green) localized in the FDJ and HF. FDJ, fimbriodentate junction; HF, hippocampal fissure; Hip, hippocampus.

Sustained subpial neurogenic zone in the reeler mice. (A-C,A′-C′) The formation of subpial neurogenic zone proceeded in reeler mutants. By P0, the Nestin-GFP+ (A,A′) and Tbr2+ (B,B′) cells were localized in the subpial region in both control and reeler mutant (arrowheads in A,B,A′,B′). Merged images for Nestin-GFP and Tbr2 in control and mutant are shown in C and C′, respectively. (D,D′) Granule cells migration relied on reelin signaling. At P0, Prox1+ granule cells were collected as a well-defined upper blade in the control (D), whereas they were evenly scattered in the dentate plate of the Reeler mutant (D′). (E-F,E′-F′) The prolonged presence of neurogenic precursors (both Nestin-GFP+ and Tbr2+ cells) in the subpial neurogenic zone in the reeler mutant at P4. By P4, as most of the Nestin-GFP+ cells left the subpial region in the control (E), many cells still persisted subpially in the reeler mutant (E′). Tbr2+ cells were loosely distributed in the molecular layer in control (F and inset). By contrast, they formed a compact subpial layer in the reeler mutant (F′ and inset). (G) The regional schema for quantification of Tbr2+ cells in the dorsal half of the dentate gyrus. The distribution of Tbr2+ cells at P4 was quantified in the MZ (in red) and the area (in blue) including granule cell layer and hilus. (H) The percentage of Tbr2+ cells in the control marginal zone (28.6±1%) was significantly lower than reeler mutant (75.4±0.9%, n=6, ^*P<0.001,χ ²-test). Ctr, control; Rlr, reeler; MZ, marginal zone; GCL, granule cell layer; SGZ, subgranular zone. Scale bar: 100 μm in A-D and A′-D′; 200 μm in E-F,E′-F′.

Defective subpial neurogenic zone, abnormal radial glial scaffolding and premature granule cell production in Cxcr4 null. (A-D) Defective subpial neurogenic zone in Cxcr4 mutant. By E18.5, Nestin-GFP+ and Tbr2+ cells were scarce in the subpial region in the Cxcr4 mutant compared with the control (white arrows in A-D). Tbr2+ cells were ectopically packed in the hilus in the mutant (yellow arrow in D). (E-H) Aberrant development of the transhilar glial fibers in the Cxcr4 mutant. At E18.5, GFAP+ processes were more enriched in the HF in the controls compared with the mutants (arrows in E,F). A subset of transhilar glial fibers labeled by BLBP at E17.5 in the controls (G) were almost lost in the mutants (H). BLBP+ somas seen in the HF and hilus of controls were absent in the mutants (arrows in G,H). (I-O) Premature production of granule cells in the Cxcr4 mutant. By E18.5, acute BrdU labeling was decreased in the forming dentate plate in the mutant (J) when compared with the control (I). However, Prox1+ granule cells were not overwhelmingly affected in the mutant despite the abnormal distribution on the migratory stream (J). By E15.5, Nestin-GFP+ precursors were able to leave the dentate primordium and form a migratory stream in the mutant (arrow in L), although it was not as robust as in the control (arrow in K). At this stage, there were more Prox1+ granule cells in the mutant (N) compared with the control (M). Boxed areas in M and N are shown at higher power in M′ and N′. (O) Birthdating analysis with BrdU pulses (at E15.5 and E16.5) and quantification at E18.5 showed that the density of BrdU+/Prox1+ cells produced at E15.5 was much higher in the mutants (227±21%), but much lower at E16.5 (51.7±7.5%) compared with the controls (n=4, ^*P<0.01, Student's t-test).

Compromised organization of the subgranular niche in Emx1-PTX mice. (A-H) Abnormal SGZ organization at early postnatal ages in Emx1-PTX mice. By P4, Nestin-GFP+ cells started to populate the SGZ in the controls (A), whereas they showed patchy distribution in the Emx1-PTX animals (red outlines in B). Compared with the control (C and E), the Ki67+ proliferating cells tended to cluster in the hilus and Tbr2+ cells were more widely spread in the Emx1-PTX animal (D,F). In contrast to the controls (G), the granule cell layer revealed by Prox1 staining was very poorly organized in the Emx1-PTX animal, with ectopic cells in the migratory stream (arrow in H). (I-R) Persistent defects in the organization of subgranular zone in Emx1-PTX mice at P10. Compared with the controls (I,K,M,O), the transgranular radial glial scaffolding labeled by BLBP failed to form properly (J) and the SGZ (shown by BrdU, Ki67 and Tbr2 staining in L, N and P) was dramatically disorganized in the Emx1-PTX animals. In sharp contrast to the control (Q), the Prox1+ GCL did not form a distinct boundary with the hilar region and there were numerous ectopic granule cells in the hilus and in the migratory route in the Emx1-PTX animals (R). Scale bar: 100 μm in A,B; 200 μm in C-H; 200 μm in I,J; 200 μm in K-R.

Contribution of subpial precursors to the SGZ. (A) Tamoxifen (TM) was administered at E17.5 into Gli1^CreERT2 line to mark the Shh-responding cells with Rosa-lacZ reporter. Twenty-four hours later, the labeled cells were found in the marginal zone (inset 1) and the outer edge of the upper blade (inset 2). (B) Forty-eight hours later, the labeled cells spread into the granule cell layer and appeared in the hilus (inset). (C) When tamoxifen was administered at E18.5, most labeled cells appeared 24 hours later in the upper blade (inset) and the future lower blade. (D-F) RosaYFP reporter was used to mark the Shh-responding cells when tamoxifen was administered at E17.5 into Gli1^CreERT2 line. Only very few cells were GFP+. One GFP+ cell in the upper blade did not show Prox1 expression (E), whereas one of the GFP+ doublet started to have weak Prox1 expression (arrows; F,F′). (G-I) The Shh-responding cells marked at E17.5 gave rise to Prox1+ granule cells at P14 (G and inset). Some of them were BLBP+ (H,H′) or GFAP+ (I,I′). Boxed areas in H and I are shown at higher power in H′,I′, respectively. GCL, granule cell layer; MZ, marginal zone; SGZ, subgranular zone. Scale bar: 100 μm in A-D,G-I; 10 μm in E,F,F′; 20 μm in H′-I′.

Schematic representation of the progression of dentate stem cells to form the SGZ. This figure shows our conception of the events leading to the formation of the SGZ. (A) Canonical Wnts are embryonically expressed in the fimbria/cortical hem and support expansion of dentate progenitors. (B) During their migration, the dentate progenitors are located in the transient neurogenic zone, where their position is maintained by Cxcl12 signaling. (C) By the first postnatal week they relocate to the SGZ where they are regulated by Wnts from the hilus and Shh (not shown) from unknown sources. DN, dentate notch; F, fimbria; FDJ, fimbriodentate junction; HF, hippocampal fissure.