The data indicate that p21 is involved in the fundamental molecular mechanisms underlying the control of neurogenesis in the hippocampus. The present study confirmed previous findings that p21 is expressed exclusively in the SGZ of hippocampus, but not in granular cell layer where the mature neurons reside 
. In the present study we found that p21 is expressed in transient amplifying progenitors, neuroblasts and young developing neurons. In NPC derived from the hippocampi of WT mice, p21 is co-localized with both SOX2 and nestin in early stage neuronal progenitors. After differentiation, GFAP+
astrocytes do not express p21, whereas p21 expression is evident in early stage both DCX+
and immature βIII-tubulin+
) neurons. Co-labeling for p21 and Ki67 show that these two proteins rarely co-localize. Therefore, when p21 is present, proliferation is restrained in these cells.
The in vitro
studies support this hypothesis. Compared to neuropheres derived from WT mice, spheres derived from p21-/- mice were larger in size, the spheres were comprised of more cells and they exhibited higher spherogenic ability and increased proliferation. In addition, the percentage of both SOX2+
cells expressing Ki67 (proliferating cells) was higher in NPC derived from p21-/- mice. The experiments using conditional knock down of p21 in NPC from WT mice showed that when p21 expression is decreased, proliferation of NPC is increased. Taken together, these results strongly support the hypothesis that p21 plays a fundamental role in regulating neuronal lineage cell proliferation in the hippocampus and in keeping these cells in a quiescent state. In vivo
, saline-treated p21-/- mice showed greater cellular proliferation in the SGZ (as detected by increased BrdU labeling) and increased neurogenesis (as indicated by more BrdU+
co-labeled cells) compared to WT mice. These results are in agreement with the finding of increased proliferation in the lateral ventricle wall and in the forebrain stem cells in young p21-/- mice 
Other published studies 
have shown that chronic treatment with nonselective, serotonin- and norepinephrine-selective reuptake inhibitor-type antidepressants stimulate hippocampal neurogenesis. In the present study this was found to be associated with decreased hippocampal p21 protein expression. No significant changes were observed in the expression of other Cdk inhibitors, showing that p21 plays an unique role in antidepressant-induced increases in neurogenesis. These results also indicate that a shared property of reuptake inhibitor-type antidepressant drugs is that they decrease the expression of p21 in the SGZ of the hippocampus. Although DMI appeared to have the greatest suppressive effect on p21 expression, there were no differences among the three antidepressants in terms of the proliferating neuronal progenitors (i.e., stimulation of neurogenesis). It is important to point out that it is not appropriate to try to draw conclusions regarding comparative potencies of the three drugs used because equimolar doses were not administered, nor were dose-response studies conducted.
Tissue specific factors can start or stop the proliferation of mitotic cells. Out of three tasted major Cdk inhibitors (p21, p18 and p27), it is only p21 that changed in response to antidepressant treatment. We can speculate that in the hippocampus, some tissue-specific factors selectively affect p21. Similar tissue specificity has been demonstrated for other brain regions. For example, TGFβ exerts its anti-proliferative properties in the cerebellum, but not in the hippocampus 
. On the other hand, Smad7, one of the transcription factors of the TGFβ signaling pathway, induces proliferation of SGZ and SVZ neuronal progenitors, independent of TGFβ signaling 
. We also examined other Cdk inhibitors using immunofluorescent analysis (not shown here) and found that p27 is most abundantly expressed in all parts of the dentate gyrus. Therefore, it is likely that p27 maintains quiescence of neuronal stem cells and differentiated neurons in the granular cell layer.
An important finding is that neurogenesis was not increased in p21-/- mice chronically treated with imipramine, suggesting that p21 suppression is not simply a correlate of increased neuronal cell proliferation, but that down-regulation of p21 is required for the stimulation of neurogenesis produced by this class of antidepressants. In fact, in p21-/- mice there was a decrease, rather than an increase in the number of both BrdU+
cells in response to chronic imipramine treatment. These observations were supported by the results showing that imipramine decreased protein levels of both DCX and NeuN in the p21-/- mice. The mechanism underlying the decrease is not clear. Cleaved caspase 3 is a protein implicated in cell death signaling pathways 
. In other cell types, chronic treatment with antidepressants can induce apoptosis via caspase 3 activation 
. In the hippocampus, apoptosis occurs mostly in proliferating early neuronal progenitors 
, however it might also occur, although at a slower rate, at the post-mitotic stage 
. In WT mice, chronic imipramine treatment did not induce changes in apoptosis as evident by cleaved caspase 3 levels. Besides its effects on cell cycle regulation, p21 also has antiapoptotic properties 
. It is possible that in the absence of p21, NPC become vulnerable to antidepressant-induced apoptosis. Another possibility is that p21 might directly interact and inactivate caspase 3 and its downstream apoptotic pathway 
, and in the absence of p21, caspase 3 is activated and apoptosis is induced. This is supported by the finding that in p21-/- mice, cleaved caspase 3 protein levels were increased after one week of imipramine treatment, indicating induced apoptosis. This could explain the decreased levels of neurogenesis in p21-/- mice after three weeks of treatment with imipramine. The fact that no changes in the levels of cleaved caspase 3 were found after three weeks might indicate that by this time very few early neuronal progenitors proliferate in the SGZ of p21-/- mice.
Although there is evidence pointing to a mechanistic link between neurogenesis and the mechanism of action of antidepressant drugs, the role of neurogenesis in the pathophysiology of depression remains unclear. Suppression of neurogenesis by hippocampus-directed irradiation 
or by using transgenic animals 
does not produce a depressive or anxious phenotype per se, but does reduce some of the behavioral effects of antidepressants 
. Recent findings also indicate that adult-born hippocampal neurons are required for normal stress response, and support a direct role of neurogenesis in stress-related depressive behavior 
. The link between depression in human and increased immobility in the forced swim test in the mouse must be drawn with great caution 
, but in the present study there was a clear-cut inverse relationship between immobility in the FST and hippocampal neurogenesis. For example, saline-treated p21-/- mice showed less immobility in the FST and increased neurogenesis compared to saline-treated WT mice. However, treatment with imipramine reduced immobility and increased neurogenesis in the WT mice, but increased immobility and decreased neurogenesis in the p21-/- mice. The last imipramine injection was administered to the mice 24 hr before the FST, which rules out acute drug/genotype interactions. These results indicate that imipramine does not show antidepressant activity in the FST or stimulate neurogenesis in p21-/- mice, and strongly suggest that p21 is required for the stimulation of neurogenesis and some of the behavioral effects produced by reuptake inhibitor-type antidepressant drugs.
In summary, p21 regulates the proliferation of neuronal progenitors in the hippocampus, and the stimulation of neurogenesis produced by reuptake inhibitor-type antidepressants might be a consequence of decreased p21 expression and the subsequent release of neuronal progenitor cells from the blockade of proliferation. Because many antidepressants stimulate neurogenesis, it is possible that their shared common mechanism of action is suppression of p21. At this point the mechanisms underlying this effect are not known. As direct effects of the antidepressants on p21 expression in NPC in vitro
were not found (data not shown), it is likely that antidepressants suppress p21 secondary to enhancing neurotransmission. Similar effects was shown for the long term lithium treatment which suppress p53 expression in cerebellum granular cells 
, and p21 is a transcriptional target for p53
. Further studies will be needed to determine the mechanisms and pathways underlying suppressive effects of antidepressants on p21 expression.