Figure 3

Confocal microscopy of DCX-positive cells. A, three-dimensional reconstruction of confocal microscopic z-series through sections stained for DCX expression. Two cells with no or short processes (A and B) are shown and two examples of cells in category E. Among all DCX-positive cells, category E is most abundant, accounting for more than 50% (see Fig. 4B) of the cells. Scale bar (in A for all panels), 30 μm. B, Co-localization of DCX and calretinin (CR) expression. CR expression identifies a postmitotic phase of granule cell development [8]. About 70% of all CR-positive cells are DCX-positive. The remaining DCX-positive CR-negative cells can be found in all categories except F (cf. Fig. 6B). Conversely, as depicted here, CR-positive cells can be found in all six categories. This also implies that (as suggested by the morphology) cells in categories A and B become postmitotic. The time span between exit from the cell cycle and the onset of CR expression varies. Single optical plane. Scale bar, 70 μm. C, many DCX-positive cells are in close contact to nestin-positive radial cells, here visualized in a nestin-GFP reporter gene mouse [19, 69]. These "vertical astrocytes" are considered the stem cells of the neurogenic region of the SGZ [1]. They are consistently S100β-negative [5]. 3-D reconstruction. Scale bar, 100 μm. D, Ki67 identifies a DCX-expressing cell with category A morphology as dividing. The panels to the right and on the top depict reconstructions from a confocal z-stack in xz and yz direction to confirm that the Ki67-positive nucleus belongs in fact to the DCX-positive cell. Scale bar, 70 μm. E, three-dimensional reconstruction of a BrdU-labeled cell with B morphology, 4 h after BrdU-injection. Scale bar, 70 μm. F, three-dimensional reconstruction of a BrdU-labeled cell with D morphology, 3 days after BrdU-injection. Scale bar, 70 μm.G, DCX-positive cells of all six morphological categories can be found in contact with astrocytes. In this example, during mitosis a GFAP-GFP-positive process surrounds the dividing DCX-expressing cell in the division plane. The separated sets of chromosomes were visualized with immunohistochemistry against phosphorylated histone H3 (red). Scale bar, 15 μm. H, apoptotic cells in the SGZ and granule cell layer were visualized with the TUNEL method (red). In the adult SGZ, TUNEL-positive cells are rare. Only 32 cells were found in 37 animals. Of these, a substantial number were DCX-positive (arrow in inset; DCX, blue), implying that cell death occurs indeed on the level of DCX expression. Co-localization was further demonstrated by placing virtual slices in yz and xz direction through the cell in question (top right) and by measuring the intensity of the fluorescent signals along a line placed across the cell (bottom right). The curves for the TUNEL signal (top, red) and DCX (bottom, blue) are at the same position. Single optical plane. Scale bar, 50 μm. I, the processes of GFAP-positive cells that engulf DCX-positive cells sometimes form a basket "cradling" the DCX-expressing cell. 3-D reconstruction. Scale bar, 15 μm. K, the vertical astrocytes with their close spatial relationship to DCX-positive cells are negative for S100β (red), whereas processes of horizontal astrocytes (arrow) are positive for GFAP-GFP and DCX. 3-D reconstruction. Scale bar, 50 μm. L, not all DCX-positive cells have close contact to astrocytes as visualized with the GFAP-GFP reporter gene mouse. 3-D reconstruction. Scale bar, 50 μm.