Depletion of central BDNF in mice impedes terminal differentiation of new granule neurons in the adult hippocampus

Abstract

Granule neurons generated in the adult mammalian hippocampus synaptically integrate to facilitate cognitive function and antidepressant efficacy. Here, we investigated the role of BDNF in facilitating their maturation in vivo. We found that depletion of central BDNF in mice elicited an increase in hippocampal cell proliferation without affecting cell survival or fate specification. However, new mutant neurons failed to fully mature as indicated by their lack of calbindin, reduced dendritic differentiation and an accumulation of calretinin(+) immature neurons in the BDNF mutant dentate gyrus. Furthermore, the facilitating effects of GABA(A) receptor stimulation on neurogenesis were absent in the mutants, suggesting that defects might be due to alterations in GABA signaling. Transcriptional analysis of the mutant hippocampal neurogenic region revealed increases in markers for immature neurons and decreases in neuronal differentiation facilitators. These findings demonstrate that BDNF is required for the terminal differentiation of new neurons in the adult hippocampus.

Figure 1. TrkB receptor expression in cycling progenitor cells and immature neurons in the adult dentate gyrus

A–C, Representative confocal images showing TrkB expression (red) in BrdU⁺ precursors (green) (arrow) in adult wild type dentate gyrus twenty four hours after BrdU administration. D–F, Representative confocal images showing TrkB receptor expression (green) in BrdU (blue) and NeuN (red)-containing immature neurons (arrow) in wild type dentate gyrus 5 days following BrdU administration. Scale bars =10µm

Figure 2. Organization and cytoarchitecture of the BDNF^2L/2LCk-cre mutant hippocampus is grossly normal

Volumetric measurements of the granule cell layer (A), hilus (B) and CA3 region (C) of wild type (WT) and BDNF^2L/2LCk-cre conditional mutant (CM) mice. Representative coronal sections containing hippocampus from wild type (D and F) and BDNF conditional mutant (E and G) mice showing comparable levels and patterns of NeuN (D and E) and synaptophysin (F and G) immunolabeling.

Figure 3. Proliferation of hippocampal progenitor cells in wild type (WT) and BDNF conditional mutant (CM) mice

A, Representative images of dentate gyrus from animals that received BrdU 24 hours earlier show robust increase of BrdU⁺ cells (dark brown staining) in the BDNF mutant (bottom) compared to the wild type control (top). Sections were conterstained with cresyl violet. B, BDNF mutants (n = 9) exhibited a significant 43% increase in BrdU⁺ cells compared to wild type animals (n = 8); *, P < 0.05.

Figure 4. Survival of newly generated cells is normal in BDNF conditional mutant mice

A, The calculated proportion of surviving new cells is similar in wild type (36%) and BDNF mutant (41%) mice (n = 12). B, Cell death, quantified by the total amount of TUNEL⁺ cells, was similar between wild type (n = 6) and BDNF mutant (n = 7) mice (P= n/s).

Figure 5. BDNF regulates later phases of neuronal differentiation in the adult hippocampus

Wild types (WT) and BDNF conditional mutants (CM) (n = 4) had comparable proportions of new cells that expressed the neuronal marker NeuN (A) or the glial marker GFAP (B) in the adult DG. C, BrdU and calbindin double immunolabeling of hippocampal sections from wild types and BDNF conditional mutants that received BrdU 4 weeks earlier. A lower proportion of new cells in the DG of BDNF mutants expressed calbindin, a terminal differentiation marker (n = 4); *, P = 0.008. D, Percentage of the total number of BrdU⁺ cells that also contained calbindin 8 weeks after BrdU administration (n = 4); *, P < 0.0001. E, Representative images showing accumulation of immature calretinin⁺ neurons (red signal, arrows) in the SGZ of BDNF conditional mutants (bottom panel) compared to wild types (top panel). F, Quantification of number of calretinin-containing cells in the wild type and BDNF mutant SGZ (n = 4); *, P = 0.02. G, Percentage of the total number of BrdU⁺ cells that also contained calretinin 4 weeks after BrdU administration (n = 4); *, P = 0.01. H, Representative images showing accumulation of immature DCX⁺ neurons (green signal) in the SGZ of BDNF conditional mutants compared to wild types.

Figure 6. Dendritic differentiation of adult-generated granule neurons is compromised in BDNF conditional mutants

Representative confocal images (A and C) and traces (B and D) of new granule neurons from wild type (A and B) and BDNF conditional mutant (C and D) mice that were marked in vivo with a GFP-encoding retroviral vector twenty days earlier; Scale bars: 25 µm. Newly generated neurons in BDNF conditional mutant mice (CM; n = 40 cells from 4 animals) displayed significantly decreased (E) dendrite branching and total length (F) when compared to wild type control cells (WT; n = 30 cells from 3 animals); * P < 0.05. Sholl Analysis for dendrite complexity (G) also indicated a significant decrease in dendritic complexity among BDNF mutant cells (closed circles) when compared to wild type control cells (open circles) at a distance of 50 µm from the soma; *, P = 0.02; #, P = 0.068.

Figure 7. Newly generated neurons in BDNF conditional mutants exhibit deficient migration to outer segments of the dentate gyrus

A, To analyze granule cell migration, a retrovirus encoding GFP was delivered to the hippocampus of wild type (WT) and BDNF conditional mutant (CM) mice, and GFP⁺/NeuN⁺ cells analyzed 28 days later; Scale bar = 50µm. B, Quantification of the proportion of GFP⁺/NeuN⁺ cells in the inner (white bar), middle (grey bar) or outer (black bar) segments of the granule cell layer. Mutants had a significantly lower proportion of cells that reached the outermost segment (closest to molecular layer) of the granule cell layer when compared to wild type animals; n = 6; *, P < 0.05). A concomitant increase of new neurons was observed in the middle segment of the granule cell layer of BDNF mutants when compared to wild types; *, P < 0.05.

Figure 8. GABA_AR-mediated changes in proliferation and differentiation of neural precursors in the adult hippocampus are absent in BDNF mutants

A, Blockade of GABA_A receptors with picrotoxin (PT) induced a significant increase in proliferation in the DG of wild types compared to their saline (Sal) treated counterparts (n = 4); F = 12; *, P = 0.01 in Fishers PLSD test. Picrotoxin treatment had no effect on the BDNF mutants (n = 4). B, Quantification of new calbindin-containing cells in the adult DG of mice (n = 4) that received saline or pentobarbital treatment 4 weeks earlier. Whereas pentobarbital stimulation of GABA_A receptors significantly increased the number of mature new granule neurons in the wild types (F = 10.8), it had no effect on the BDNF mutants (n = 4); *, P = 0.02.

Figure 9. Laser capture microdissection (LCM) of cells in the hippocampal neurogenic region of adult mice

Representative images of hippocampal sections before (left panel) and after (right panel) LCM. Arrows indicate areas of dissected tissue.