Antidepressants stimulate hippocampal neurogenesis by inhibiting p21 expression in the subgranular zone of the hipppocampus

Abstract

The relationships among hippocampal neurogenesis, depression and the mechanism of action of antidepressant drugs have generated a considerable amount of controversy. The cyclin-dependent kinase (Cdk) inhibitor p21(Cip1) (p21) plays a crucial role in restraining cellular proliferation and maintaining cellular quiescence. Using in vivo and in vitro approaches the present study shows that p21 is expressed in the subgranular zone of the dentate gyrus of the hippocampus in early neuronal progenitors and in immature neurons, but not in mature neurons or astroglia. In vitro, proliferation is higher in neuronal progenitor cells derived from p21-/- mice compared to cells derived from wild-type mice. Proliferation is increased in neuronal progenitor cells after suppression of p21 using lentivirus expressing short hairpin RNA against p21. In vivo, chronic treatment with the non-selective antidepressant imipramine as well as the norepinephrine-selective reuptake inhibitor desipramine or the serotonin-selective reuptake inhibitor fluoxetine all decrease p21 expression, and this was associated with increased neurogenesis. Chronic antidepressant treatment did not affect the expression of other Cdk inhibitors. Untreated p21-/- mice exhibit a higher degree of baseline neurogenesis and decreased immobility in the forced swim test. Although chronic imipramine treatment increased neurogenesis and reduced immobility in the forced swim test in wild-type mice, it reduced neurogenesis and increased immobility in p21-/- mice. These results demonstrate the unique role of p21 in the control of neurogenesis, and support the hypothesis that different classes of reuptake inhibitor-type antidepressant drugs all stimulate hippocampal neurogenesis by inhibiting p21 expression.

Figure 1. p21 is expressed in neuronal progenitors in the SGZ of the hippocampus.

A) The confocal image shows that p21 (red) is expressed in the SGZ of the dentate gyrus and is not expressed in mature neurons (NeuN, green) in the GCL. Here, and in other confocal images, nuclei are stained with the DNA specific dye DAPI (blue). p21-positive cells appear pink. p21 is expressed mostly in nuclei; however, some cytoplasmic staining also is observed. SGZ, subgranular zone; GCL, granular cell layer; ML, molecular layer; B–E) Confocal images depicting p21 co-localization with neuronal markers in the SGZ of the hippocampus. p21-positive nuclei line the SGZ and appear light blue. Upper panel-low magnification (x20), lower panel–higher magnification (x63). B) p21 (green) is co-localized with SOX2 (intranuclear, red). Cells that express both proteins appear yellow; cells that expressed SOX2 only appear pink; C) p21 (green) is co-localized with nestin (cytoplasmic, red); D) p21 (green) is co-localized with DCX (cytoplasmic, red); E) p21 (green) is not co-localized with GFAP (cytoplasmic, pink). Some of p21 positive cells are marked with arrows.

Figure 2. p21 is expressed in hippocampal NPC of a neuronal lineage.

Cultured neuronal progenitors were stained for p21 (intranucelar, red) and various neuronal markers. A) co-localization of p21 and SOX2 (green, intranuclear, double-positive nuclei appear yellow; B) co-localization of p21 and nestin (green, cytoplasmic); In differentiating NPC, p21 co-localizes with C) DCX (green, cytoplasmic); and with D) Tuj-1 (green, cytoplasmic), but not with E) GFAP (green, cytoplasmic). Numbers indicate the percentage of p21-positive cells among the total number of cells expressing a respective marker. Counting was performed in triplicate (>500 cells per sample). Data are presented as mean ± SEM.

Figure 3. Increased proliferation of NPC derived from p21-/- mice.

A) The spheres derived from WT and p21-/- hippocampi 7 days after plating. Note bigger size and more cellularity in p21-/- spheres; B) Seven days after seeding, the NPC spheres were collected, dispersed and the number of cells was counted; C) Sphere self-renewal analysis. NPC spheres were dispersed and cells were seeded at 0.5 cell/well into 96-well plates. The percent of cells forming new spheres was determined after 10 days; D) The percent of Ki67-positive cells co-expressing nestin or SOX2. Counting was performed in triplicate (>500 cells per sample). Data are presented as a mean ± SEM, *, p<0.05, **, p<0.01; E) Representative confocal images of proliferating NPC derived from WT and p21-/- hippocampi and stained for Ki67 (intranucelar, red) and nestin (cytoplasmic, green) or SOX2 (intranuclear, green). p21-/- NPC exhibit more Ki67-positive cells; F) Proliferating neuronal precursors derived from WT and p21-/- hippocampi and stained for Ki67 (intranucelar, red) and p21 (intranuclear, green). Note that WT NPC cells positive for Ki67 are negative for p21; G) Double-labeling analysis of WT NPC with a color-coded fluorogram. The confocal image shows the absence of co-localization between p21 (green) and Ki67 (red).

Figure 4. Conditional suppression of p21 enhances proliferation of WT NPC.

A) One thousand WT NPC cells were kept in proliferating conditions for 4 days. Cells were cultured in 24-well plates in triplicates. The mean number of cells at the time of collection; B) Cells were fixed and stained for Ki67 or p21 . The graph depicts the percentage of p21- or Ki67-positive cells in each culture. Counting was performed in triplicates (>500 cells per sample). Data are presented as a mean ± SEM, *, p<0.05. Two independent experiments were conducted, and representative results are shown; C) Representative confocal images of NPC treated with either control lentivirus (control) or lentivirus expressing p21 shRNA. The cells were stained for Ki67 (intranucelar, red) and p21 (intranuclear, green). Note the increased number of Ki67-positive and the decreased number of p21-positive cells after p21 shRNa treatment.

Figure 5. Chronic treatment with antidepressants suppressed p21 expression and increased neuronal proliferation in the SGZ of hippocampus.

A) Western blot analysis of p21 and other Cdk inhibitors in hippocampus of mice chronically treated with normal saline (NS), imipramine (IP), desipramine (DMI) or fluoxetine (FL). For each sample hippocampi from 5 mice/group were pooled; B) Quantitative analysis of three independent experiments. The intensity of p21 bands was measured, normalized to the loading control (β actin) and the ratios were corrected to NS control to quantify relative changes; C) Number of BrdU⁺ cells. Non-parametric Kruskal-Wallis test [(H3,N = 20) = 11.75, p<0.008]. D) Number of BrdU⁺/DCX⁺ cells. One way ANOVA followed by Dunnett's post-hoc test [(F3,16) = 3.88, p = 0.029 for all groups] in the SGZ. For each sample hippocampi from 5 mice/group were analyzed. Data are presented as a mean ± SEM.

Figure 6. SGZ neurogenesis and immobility in the FST in p21-/- mice chronically treated with normal saline (NS) or imipramine (IP).

A) Number of BrdU⁺ cells. The effects of treatment x genotype were assessed by two-way ANOVA followed by Neuman-Keuls post-hoc test [(F1,16,) = 20.18, p = 0.004]; B) Number of BrdU⁺/DCX⁺ cells in the SGZ. Two-way ANOVA [(F1,16) = 23.483, p = 0.00018]. Data are presented as a mean ± SEM, *, p<0.05 vs WT NS, #, p<0.05 vs p21-/- NS; Western blot analysis of C) Neuronal markers and Cdk inhibitors; and D) Cleaved caspase 3. For each sample hippocampi from 3 mice/group were pooled; E, F) Quantitative analysis of three independent experiments. The intensity of E) DCX and NeuN, and F) cleaved caspase 3 bands were measured, normalized to the loading control (β actin) and ratios were corrected to WT NS control to quantify relative changes. Data are presented as a mean ± SEM, *, p<0.05 vs WT NS; G) Immobility in the FST. The values are the total time spent immobile in sec. Data are presented as a mean ± SEM, n = 9–10/group *, p<0.05 vs WT NS or p21-/- NS, #, p<0.05 vs p21-/- NS. Imipramine decreased immobility in WT treated mice [(M = 143.08 SD = 39.02) compared to WT untreated (M = 176.99 SD = 20.39) mice, t(17) = 2.33, P  = 0.032]. The effects of treatment and genotype were assessed by two-way ANOVA followed by Neuman-Keuls post-hoc test[ F(1, 33) = 12.551, p = .001]. The p21-/- NS group had significantly less immobility time compared to WT NS whereas the p21-/- IP treated group had significantly more immobility than the p21-/- NS group. This indicates that antidepressant treatment produces a paradoxical effect in the p21-/- mice.