|Home | About | Journals | Submit | Contact Us | Français|
It is now clear that neurogenesis occurs in the brain of adult mammals. Many studies have attempted to establish relationships among neurogenesis, depression and the mechanism of action of antidepressant drugs. Therapeutic effects of antidepressants appear to be linked to increased neurogenesis in the hippocampus. Cdk inhibitors are expressed in multiple brain regions, presumably maintaining quiescence in differentiated neurons. Recently, the abundant expression of p21Cip1 was found in neuroblasts and in newly developing neurons in the subgranular zone of the hippocampus, a region where adult neurogenesis occurs. Chronic treatment with the tricyclic antidepressant imipramine markedly decreased p21Cip1 mRNA and protein levels and stimulated neurogenesis in this region. These results suggest that p21Cip1 restrains neurogenesis in the hippocampus, and antidepressant-induced stimulation of neurogenesis might be a consequence of decreased p21Cip1 expression, with the subsequent release of neuronal progenitor cells from the blockade of proliferation. These findings suggest the potential for new therapeutic strategies for the treatment of depression that target cell cycle proteins. However, there is a possibility that long-term stimulation of neurogenesis might exhaust the proliferation potentials of neuronal progenitors.
For many years the production of new neurons in the brains of mammals had been considered confined to development. This implied that any loss of neurons was irreversible and inevitable because damaged or dying neurons could not be replaced in adult brain. The non-renewability of neurons was a fundamental premise underlying the pathophysiology of some neurological and neurodegenerative disorders and the responses to brain injury. The “no new neurons in adult brain” doctrine was based upon the lack of observable mitotic divisions and the absence of neurons showing a transition from an immature to a mature state in adult brain. With the development and implementation of new methodologies, such as 3H-thymidine labeling in the1960s, the presence of neurogenesis was first observed in adult mammals.1 Later, using bromodeoxyuridine (BrdU) labeling, several thousand dividing cells could be detected in the hippocampus of young adult mice.2 Of these cells, half express neuron-specific markers. The rate of generation of new neuron in young adult mice and rats has been estimated to be from 1.5% to 6% of the total hippocampal granule cell population per month. Although differences have been found between and among species, adult neurogenesis has been found in all mammals studied, including various species of rodents, nonhuman primates and humans (reviewed in refs. 3–6). Adult neurogenesis has decreased over evolution. Compared to nonmamalian vertebrates such as birds and reptiles, the rate and extent of neurogenesis are much lower in mammals.7 Furthermore, the rate of neurogenesis is age-dependent; it decreases from adolescence to adulthood, and is even lower in aged animals.7–10
Neurons are involved in information processing, whereas glia (astrocytes and oligodendrocytes) provide an essential supportive role for the neurons. Adult neural stem cells (NSC) are cells that can self-renew and differentiate into all types of neural cells, including neurons, astrocytes and oligodendrocytes.11 These properties of adult NSC have been shown in vitro, but have not been convincingly demonstrated in vivo. Therefore, these dividing cells are frequently referred to as “neural progenitors”.4 Neuronal progenitors can be isolated from many areas of the adult nervous system, but adult neurogenesis is not a global phenomenon throughout the brain, but is restricted to specific regions. Neurogenesis has been most clearly demonstrated in two brain locations: the subventricular zone (SVZ), located next to the ependyma, a thin cell layer that lines the lateral ventricles; and the subgranular zone (SGZ) of the dentate gyrus of the hippocampus. Neurons born in the adult SVZ migrate over a great distance through the rostral migratory system, eventually turning towards the granule and periglomerular cell layers in the olfactory bulb (OB). Neurons born in the adult SGZ migrate into the granular cell layer of the dentate gyrus and become granule cells.3–5 While neurogenesis can be stimulated in other brain regions by various conditions and insults such as injury, it remains unclear how much neurogenesis takes place in brain areas other then the SGZ and SVZ under normal conditions.4,6,12 In some species adult neurogenesis has been reported to occur in the neocortex, hypothalamus, striatum, amygdala, substantia nigra and brainstem, however, some of these findings have been disputed.6 In humans, it is generally accepted that neurogenesis occurs in the hippocampus,13 and recent evidence suggests that it also takes place in the olfactory bulb.14 The greatest controversy is whether neurogenesis occurs in nonhuman primate and human neocortex.6,15–17
NSC and progenitor cells within the SGZ represent a spectrum of several distinct progenitors cell types. Using a neural progenitorspecific, nestin-GFP-expressing transgenic mouse, these cells have been defined as types 1, 2a, 2b and 3b.3,4,18–21 The early neural progenitors, type 1 and 2a cells, lack committed neuronal markers and express nestin, a primitive neuroepithelial marker. In contrast, nestin expression is downregulated in late, committed neuronal progenitors or type 2b and type 3b cells; they begin to express markers of developing young neurons such as microtubule protein doublecortin (DCX) and polysialyted neural cell adhesion molecule (PS-NCAM).20–22 Normally quiescent astrocyte-like type 1 cells express nestin and contain a radial glial fibrillary acidic protein (GFAP)-expressing process and are believed to represent self-renewing stem-like cells.23 Type 2 cells, which arise from asymmetric division of type 1 cells, represent fully committed neuroblasts. They express nestin but lack GFAP, and are distinguished by the absence (type 2a) or presence (type 2b) of DCX.23,24
The survival and the integration of newborn neurons in the SGZ are largely determined during a critical time window (1–3 weeks) when these neurons are immature and display unique physiological properties. Approximately 60% of the newborn cells that don’t terminally differentiate die within one week of their generation.25 The neurons that survive and mature are thought to be integrated into existing circuitry as functional neurons.26,27 Neurons that do not die within this initial period survive for at least 19 months in rodents.28 Survival of new neurons can be increased by some interventions, such as learning new tasks and environmental enrichment.29,30
A number of factors are involved in the regulation of the process of the birth of new neurons and their subsequent migration, maturation and survival. Neurotransmitters have complex effects on hippocampal neuronal proliferation. Serotoninergic, cholinergic and noradrenergic activation all increase neurogenesis in the dentate gyrus.31–33 GABA, acting through GABAA receptors, decreases proliferation in both the dentate gyrus and the SVZ,34,35 and dopaminergic signaling has a similar inhibitory effect.36–38 Growth factors also can affect neurogenesis. For example, brain-derived neurotrophic factor (BDNF) enhances hippocampal proliferation, an effect that is blocked by BDNF-specific antibodies.39 Insulin-like growth factor 1 (IGF-1),40 transforming growth factor (TGF), vascular endothelial growth factor (VGEF) and others growth factors have been shown to induce proliferation in the dentate gyrus.41–45 The Wnt signaling pathway was identified to induce neurogenesis in the dentate gyrus.46 Corticosteroids were the first hormones found to have an affect on hippocampal neurogenesis.41 Activation of hypothalamic-pituitary-adrenal axis and sustained elevation of glucocorticoids can lead to chronic inhibition of neurogenesis,47 whereas adrenalectomy leads to an increase in the number of neurons in the adult rat hippocampus.48 Thyroid hormones also stimulate neurogenesis.47 Inflammation, which gives rise to microglial activation in the area where the new neurons are born, strongly impairs hippocampal neurogenesis in rodents.49,50 Inflammation does not affect precursor cell proliferation, but it reduces neuronal differentiation and survival, effects likely due to microglia-derived interleukin-6 (IL-6). In contrast, leukemia inhibitory factor (LIF), a cytokine that belongs to the same IL-6 family of cytokines, is induced in the brain in response to brain injury and hypoxia, and increases neuronal proliferation and survival.51 Thus, neurotransmitters, multiple growth factors, hormonal status and pro-inflammatory molecules all can modulate neurogenesis.
The generation of neurons by progenitor cells involves the tight coordination of multiple cellular activities, including exit from the cell cycle, migration and differentiation. During development of the central nervous system, cycling progenitor cells generate post-mitotic neuronal precursors that rapidly migrate out of germinal zones and initiate neuronal differentiation (reviewed in ref. 52). The mechanisms involved in the maintenance of quiescence or the initiation of neuronal cell proliferation remain a central question in developmental biology. In adult mammalian brain, most neuronal cells stay in a quiescent, differentiated state. In post-mitotic, terminally differentiated neurons, it is likely that cell cycle activity is arrested by the increased expression of Cdk inhibitors. Numerous cell cycle inhibitors are expressed in the brain either in a ubiquitous fashion (e.g., p19Ink4d) or in specific regions: p15Ink4d in the forebrain; p21Cip1 in the cerebellum53 and p27Kip1 in the cerebellum and cortical post-mitotic neurons.53,54 In animals derived from backcrosses between p27Kip1-null and p19Ink4d-null mice, a subpopulation of neurons showed increased proliferation in the brain, indicating that these two Cdk inhibitors cooperate to maintain differentiated neurons in a quiescent but potentially reversible state.55
Cell cycle exit and differentiation also has been shown to involve negative cell cycle regulators. In developing olfactory epithelium, there is a temporal progression of the expression of p27Kip1, p18Ink4c, p16Inc4d and p21Cip1 during the transition from neural cell cycle exit to olfactory neuron maturation. Thus, different Cdk inhibitors act sequentially,56 and this non-overlapping pattern suggests that these proteins have independent roles in the brain. In primary adult hippocampal-derived stem cells, exposure to cell cycle exit factors increases p21Cip1 expression.57 In mammal brain, the role of cell cycle related proteins often goes beyond regulation of cell cycle progression. They can act as neuronal differentiation factors, especially during development. 52 Increased expression of p21Cip1 is associated with terminal differentiation in many cell types.58 In PC12 pheochromocytoma cells, a model system to study neuronal differentiation, induction of differentiation leads to p21Cip1 accumulation.59,60 Another Cdk inhibitor p57Kip2 is predominantly expressed in fully differentiated neurons in the SVZ.61 p27Kip1 plays an important role in neurogenesis in the developing mouse cerebral cortex by promoting differentiation and radial migration of cortical projection neurons.62 p27Kip1, paired with Cdk5, controls cytoskeletal changes essential for neuronal migration.63 The Xenopus Cdk inhibitor Xic1 shows homology with mammalian p21Cip1, p27Kip1 and p57Kip2. When stimulated by Notch1, Xic1 promotes neuronal differentiation, and this function is distinct from its ability to inhibit the cell cycle.64
It has been hypothesized that the microenvironments in the SVZ and SGZ, known as the neurogenic niche, might have specific factors necessary for the proliferation and differentiation of neural progenitors and their integration into existing neural networks.65 Cdk inhibitors might control neural progenitor proliferation within neurogenic niches. p27Kip1 regulates the division of transit amplifying progenitors (type 2a cells) in the adult SVZ.66 Loss of p27Kip1 has no effect on the number of NSC, but selectively increases the number of transit-amplifying progenitors concomitantly with a reduction in the number of neuroblasts in the SVZ. Recently, we demonstrated the abundant expression of intranuclear p21Cip1 in the SGZ, where it is colocalized with NeuN, a marker for neurons67 (Fig. 1). DCX is a marker of young neurons. By using FACS analysis we found that among DCX-positive cells, 42.8% stained positive for p21Cip1, indicating that p21Cip1 is expressed in newly developing neurons (type 2b cells). p21Cip1-null mice were examined, and the rate of cellular proliferation was increased in the SGZ of p21Cip1-deficient mice compared with wild-type animals. In addition, the levels of both DCX and NeuN protein were increased in adult p21Cip1-null mice, further demonstrating increased hippocampal neuronal proliferation. Surprisingly, p21Cip1 was not expressed in terminally differentiated granular neurons in the dentate gyrus (Fig. 1). Others have found that brain ischemia produces a greater activation of neuronal proliferation in the SGZ and SVZ of p21Cip1-null mice compared to wild-type mice,68 indicating that cell cycle activation of neuronal precursors is intrinsically inhibited by p21Cip1. Based upon these data, we speculate that p21Cip1 expressed in the SGZ is a part of the hippocampal neurogenic niche, and has an important role in regulating the number of newly developing neurons in this region. Cell-specific properties of cell cycle regulators have been described. For example, RB induces the p21Cip1 promoter in epithelial cells, but not in fibroblasts.69 Pituitary tumor-transforming gene (PTTG) deletion resulted in p21Cip1 induction in the pituitary, but not in the thyroid, thymus or spleen.70 Thus, p21Cip1 protein might play specific role in hippocampal neurogenesis in the SGZ.
Neurogenesis declines with aging in both SVZ and SGZ.71 In rodents, the decline in SGZ neurogenesis mainly occurs between 2 and 12 months of age,3 and control of cell division and death is critical in this process. Progenitor proliferation in the SVZ, neurogenesis in OB as well as multipotent progenitor frequency and self-renewal potentials all decline with age.72 This is associated with increased expression of the Cdk inhibitor p16Ik4a, a protein that has been linked to cellular senescence. Aging p16Ink4a-null mice have smaller declines in SVZ proliferation and multipotent progenitor frequency and self-renewal potentials. Interestingly, p16Ink4a deficiency does not affect progenitor function in the SGZ, indicating regional differences during aging.72 In mice overexpressing the short (truncated) form of p53 (p44 mice),73 the interaction between the full-length and short isoforms of p53 stabilizes p53 function and increases transcription of p21Cip1. This increase in p21Cip1 expression is associated with a diminished capacity for neurogenesis. As these mice age they undergo a more pronounced decrease in neuronal proliferation in the SVZ, a reduced number of neurons in the OB, and exhibit features of premature aging.74 Thus, the progressive loss of neurogenesis might involve age-related changes in the regulation of cell cycle inhibitors.
The functional significance of adult neurogenesis is not known. The hippocampus is involved in a number of important functions, including memory formation and retrival, learning, and neuroendocrine and mood regulation. Many studies have attempted to link changes in hippocampal neurogenesis to alterations or deficits in these endpoints. The relationships among hippocampal neurogenesis, depression and the mechanism of action of antidepressant drugs have generated a considerable amount of interest and controversy. This has led to a neurogenesis theory of depression (or antidepressant action). In rodents, chronic stress (or glucocorticoid administration) induces “depression-like” behavior, and this is accompanied by decreased hippocampal neurogenesis.75,76 In humans, stress has been found to play one of the leading roles in the pathogenesis of depression.77,78 Chronic stress (or glucocorticoid administration) leads to hippocampal atrophy in rodents, and in humans, depression is associated with hippocampal atrophy.75,76 However, it is improbable that depression-related decreases in hippocampal neurogenesis could produce a sufficient reduction in hippocampal volume that it could be detected using imaging. Major depression is associated with neurocognitive deficits and neuroendocrine dysregulation, both of which could involve impairment of hippocampal function.
If depression is due to decreased hippocampal neurogenesis, then reducing neurogenesis should produce depression. This has not been found to be the case because reducing hippocampal neurogenesis does not produce depression-like behavior in animals models, although it has been argued that this is due to the lack of appropriate and valid animal models of human depression. A seminal study reported that radiation-induced inhibition of hippocampal neurogenesis blocked the behavioral effects of a tricyclic and a SSRI antidepressant.79 However, this study was criticized because the behavioral test used might have been sensitive to the anxiolytic properties of the drugs and was not appropriate for the detection of antidepressant activity, and the selectivity of the radiation-induced lesion was problematic.
One of the most compelling lines of evidence linking depression and neurogenesis are the findings that antidepressant drugs, and procedures and tasks that reduce depression, such as electroconvulsive shock78,80–82 and exercise,83 increase neurogenesis. Most antidepressant drugs produce a rapid increase in extracellular levels of norepinephrine and/or serotonin; however, the onset of appreciable clinical improvement usually takes 3–4 weeks.84 This delay suggests that long-term adaptation in neurotransmitter systems and/or their downstream targets might be necessary for their therapeutic effects.
Antidepressants increase neurogenesis in the adult hippocampus.80,85 They appear to increase the proliferation rate of neuronal progenitors without affecting the rate of survival or the rate of differentiation into neurons.86 The induction of neurogenesis requires repeated treatment with the antidepressant, and the time course of neurogenesis is consistent with time course of its therapeutic effects.48 Upregulation of neurogenesis occurs after the chronic administration of different classes of antidepressants, including tricyclic antidepressants, selective serotonin (SSRI) and norepinephrine (SNRI) reuptake inhibitors and monoamine oxidase inhibitors.80,85 Some drugs that do not directly affect biogenic amine reuptake also stimulate neurogenesis and have antidepressant activity. For example, agomelatine, a new antidepressant that acts as a melatonin agonist, also stimulates neurogenesis in rat hippocampus,87 as do thyroid hormones, which have been used in the treatment of depression.47 Psychotropic drugs that do not have antidepressant activity, such as haloperidol, have no effect on neurogenesis.85 Thus, the induction of adult neurogenesis might represent a common final target for different classes of antidepressants. Determining how antidepressant drugs stimulate neurogenesis remains an interesting and thought-provoking area of research.
Although it is well established that antidepressant therapies stimulate neuronal proliferation, very limited data are available on the fundamental mechanisms underlying this effect. It recently was shown that VEGF signaling through a high affinity receptor tyrosine kinase (Flk-1) is necessary and sufficient for the both the neurogenic and behavioral actions of multiple classes of antidepressants88 The signaling pathway by which antidepressants induce VEGF expression could involve the cAMP response element binding protein (CREB).80,89 There is evidence that antidepressant treatment might affect neuronal progenitor cells. Thus, the selective serotonin re-uptake inhibitor fluoxetine does not affect the division of stem-cell like cells, but increases the division of early neuronal progenitors.23
In our experiments, chronic treatment with the tricyclic antidepressant imipramine stimulated neurogenesis in the hippocampus. This imipramine-induced increase in the number of SGZ neurons was associated with markedly decreased hippocampal p21Cip1 mRNA and protein levels, concordant with an increased number of mature neurons expressing NeuN.67 These results suggest that p21Cip1 restrains neurogenesis in the SGZ, and the imipramine-induced stimulation of neurogenesis might be a consequence of decreased p21Cip1 expression and the subsequent release of neuronal progenitor cells from the blockade of proliferation. These data were the first to show that proteins involved in the regulation of the cell cycle are a possible mechanistic link between neurogenesis and action of antidepressants. The signaling cascade underlying imipramine-induced suppression of p21Cip1 is not known. Imipramine inhibits the reuptake of both norepinephrine and serotonin, and it is likely that imipramine suppresses p21Cip1 promoter activity secondary to its effects on one or both neurotransmitters. Local hippocampal factors also could be involved in the regulation of p21Cip1 expression. For example, the interplay between transforming growth factor beta-2 (TGFβ2) and brain-derived neurotrophic factor in the cerebellum has been suggested to account for anti-proliferative and pro-differentiating activities observed in post-mitotic cerebellar neurons.90 Because many antidepressants stimulate neurogenesis, it is possible that their shared common mechanism of action is suppression of p21Cip1. Importantly, these studies indicate that processes regulating cell cycle progression are operable in the adult brain and therefore may potentially be a target for therapy.
The discovery of adult neurogenesis raised expectations for improved treatments for patients with central nervous system disorders. Despite all of the controversy surrounding the role of neurogenesis in etiology of depression, the consensus opinion is that increased neurogenesis triggered by antidepressants in some manner contributes to their therapeutic effects. It is possible that new therapeutic strategies for the treatment of depression could involve stimulating neurogenesis by targeting cell cycle related proteins. The new findings with regards to antidepressants and neurogenesis have a number of interesting ramifications. For example, it has been suggested that more neurogenesis is not necessarily beneficial.30 Increased mitotic expansion without corresponding increases in the rate of differentiation could lead to a large pool of undifferentiated cells, which could have adverse consequences. Excessive neurogenesis also could result in inappropriate migration, differentiation and integration into existing neural networks, and could underlie such pathological conditions as epilepsy.30
The studies of the effects of antidepressants on neurogenesis have focused on relatively short-term antidepressant administration (i.e., 28 days or less). Constant and/or long-term stimulation of neurogenesis by antidepressant treatment might have adverse consequences. Relative quiescence might extend stem cell longevity (long-term maintenance of self-renewal ability) by preventing the exhaustion of the capacity to proliferate.91,92 It was shown that NSC have finite proliferation potentials and p21Cip1 negatively regulates adult NSC proliferation.93,94 We found that treatment with antidepressant for three weeks suppresses murine SGZ p21Cip1 expression with a subsequent increase in hippocampal neurogenesis.67 Therefore it is conceivable that suppression of p21Cip1 by longer treatment could result in eventual and premature exhaustion of neuronal precursors as it was observed in aging p21Cip1-null mice.93 If this is the case, then in some instances it could lead to the loss of therapeutic efficacy, or the development of treatment refractory depression, and this appears to be true. A relatively common and troubling phenomenon is that antidepressant drugs, especially SSRIs, loose their effectiveness after long-term use.95 This has been called relapse during maintenance treatment (in common terms, “Prozac poop-out”). A possible mechanism underling this effect is exhaustion of neurogenesis due to excessive stimulation.
The role of neurogenesis in normal and diseased states is not known, but it has been hypothesized to be involved in neuronal plasticity underlying emotional states, memory and cognition. It has been reported that neurocognitive deficits can be observed in patients undergoing long-term antidepressant treatment even though their mood disorder is well controlled by the drug.96 One study found moderate changes in memory and cognitive function in patients with major depression after an average of 2 years of treatment with different classes of antidepressants.97 The extent to which these deficits in cognitive function are due to the underlying depressive disorder or antidepressant treatment are not known, but it is possible that long-term antidepressant therapy might affect cognitive function by disrupting the normal regulation of neurogensis. This might be more significant in elderly patients, who also have age-related deficits in cognitive function. It is interesting to note that depression is very common in the elderly, who conceivably have a reduced rate of hippocampal neurogenesis. It is possible that neurogenesis-dependent effects of antidepressants might be decreased in this population.
Antidepressants are commonly prescribed drugs that undoubtedly are effective in treating mood disorders. They are widely used not only in adults, but also in children and adolescents. The rate of antidepressant drug treatment has increased more than four times between early 1990s and early 2000s.98 We are beginning to learn more and more about the effects of antidepressants on neuronal function. The finding that antidepressants affect cell cycle proteins opens up new avenues for research and drug development. However, with this new information comes the possibility that some the adverse effects produced by antidepressants and the limitations of drug therapy might involve interactions with cell cycle regulatory proteins.
This work was partially support by a NARSAD Young Investigator Award (V.C), National Institutes of Health Grants MH079988 (V.C), MH078037 (R.N.P.) and MH079370 (R.N.P.), and the Levine Family Fund Research Endowment (R.N.P.).