1In recent decades evidence has accumulated demonstrating the birth and functional integration of new neurons in specific regions of the adult mammalian brain, including the dentate gyrus of the hippocampus and the subventricular zone.
2Studies in a variety of models have revealed genetic, environmental and pharmacological factors that regulate adult neurogenesis. The present review examines some of the molecular and cellular mechanisms that could be mediating these regulatory effects in both the normal and dysfunctional brain.
3The dysregulation of adult neurogenesis may contribute to the pathogenesis of neurodegenerative disorders, such as Huntington's, Alzheimer's and Parkinson's disease, as well as psychiatric disorders such as depresssion. Recent evidence supports this idea and, furthermore, also indicates that factors promoting neurogenesis can modify the onset and progression of specific brain disorders, including Huntington's disease and depression.
Neurogenesis is a critical process in the development of the embryonic brain; neuronal precursors located in the primitive ectoderm of the neural tube and the neural crest have the capacity to differentiate in response to neural-inducing signals, migrate and form proscribed functional circuits, thereby creating the central and peripheral nervous systems, respectively. In the adult, however, the capacity for neurogenesis is considerably limited. Destruction of the cells of the nervous system by injury or disease produces minimal regeneration in comparison to the renewal of skin and hepatic tissues, an observation that, for many years, led scientists to believe that the adult brain of higher vertebrates was ‘hard-wired’ and consequently incapable of producing new neurons. This was further reinforced by classical ideologies, which dictated that the continued production of new neurons would alter the characteristics and personality of an individual.
These beliefs were first challenged by studies of neurogenesis in adult rats by Joseph Altman in 1962. Using [3H] thymidine autoradiography, he identified new neurons in the brains of adult rats. In particular, the olfactory bulb1 and hippocampus2 were identified as key areas in the adult brain containing pluripotent stem cells, which have the potential to proliferate, differentiate and even incorporate into neural circuits and function.3
Since those early studies, neurogenesis has been consistently reported to occur in two regions of the adult brain (Fig. 1): (i) the subventricular zone (SVZ), which lines the lateral ventricles and gives rise to new interneurons that reach the olfactory bulb via the rostral migratory stream; and (ii) the subgranular layer (SGL) of the dentate gyrus, at the interface of the granule cell layer and the hilus.4 The precursor cells in the SGL divide and produce daughter cells that differentiate into granule neurons.5 The granule neurons become incorporated into the granule cell layer and express markers of mature neurons such as neuron-specific enolase (NSE), the calcium-binding protein calbindin, the N-methyl-d-aspartate (NMDA) receptor subunit NR1 and neuronal nuclei (NeuN) protein.4,6–8 Within a few days, these neurons project to the CA3 hippocampal region, receive afferents and exhibit electrophysiological properties very similar to those of mature dentate granule neurons.3,9–11 Hippocampal neurogenesis has also been reported to occur in the adult human and monkey brain12–14 and, importantly, can be correlated with changes in hippocampal plasticity.15 Furthermore, adult neurogenesis appears to be an evolutionary ancient phenomenon, having also been described in non-mammalians as diverse as birds and insects.16,17
REGULATORS OF NEUROGENESIS IN THE HEALTHY ADULT BRAIN
The extent of basal hippocampal neurogenesis is limited, but can be upregulated in response to seizures and environmental manipulation,18 and administration of drugs such as selective serotonin re-uptake inhibitors.19,20 Wheel-running in adult mice and rats has been consistently shown to increase hippocampal progenitor proliferation and neurogenesis;21,22 an effect that appears to be specific to the hippocampus and has not been reported to occur in the olfactory bulb.23 Several studies have also shown that environmental enrichment (involving exposure to novel objects and environmental complexity relative to standard housing) can also increase hippocampal progenitor proliferation23–25 and enhance the survival of newly generated cells in the hippocampus.26–28
Even more interesting are studies showing that these increases in neurogenesis can be correlated with improvements on behavioural tasks in rodents. A study by Gage and colleagues showed that 129/SvJ mice, which are known to do poorly on learning tasks, produce more astrocytes and fewer neurons than other strains, but that exposure to an enriched environment stimulated neurogenesis and improved learning in these mice.26 Similarly, in a study of aged mice at 10 and 20 months of age, an enrichment-induced increase in hippocampal neurogenesis correlated with improvements in learning parameters, exploratory behaviour and locomotor activity.25 These studies show that enrichment has a sustained, positive effect on hippocampal neurogenesis, implying that the ageing brain has a capacity for cellular regeneration, when driven by sensory stimulation and activity. Furthermore, adult neurogenesis is not only a transient phenomenon but leads to the lasting presence of newly generated neurons,27 which can differentiate and migrate to form appropriate functional connections in the brain.3,29
Taken together, these observations showing that increased neurogenesis can improve measures of cognition and behaviour, even in aged and diseased rodents, have led to interest in the molecular mediators regulating this beneficial effect. The cumulative research efforts of the past decade have shown that, rather than acting through a single mechanism, environmental enrichment induces a host of complex biochemical changes that promote the generation of new cells in the hippocampus (Fig. 2).
The dentate gyrus is enriched with 5-HT1A receptors,30 and receives serotonergic innervation from the brainstem.31 Serotonergic axons from the dorsal raphe nucleus terminate in the molecular layer and those from the median raphe nucleus terminate in the hilus. Granule cells continue to be produced in the innermost part of the hilus, become integrated into the layer and extend axons along the mossy fibre pathway throughout the hilus.
A considerable amount of evidence suggests that 5-hydroxytryptamine (5-HT) may stimulate the production of neurons in the dentate gyrus; pharmacological manipulations, such as administration of selective serotonin re-uptake inhibitor (SSRI) antidepressants,19,20 and 5-HT1A agonists, such as 8-hydroxy-2-(di-n-propylamino)-tetralin (8-OH-DPAT),32 stimulate adult hippocampal neurogenesis. Inhibition of serotonin (5-HT) synthesis or lesions of 5-HT neurons are associated with decreases in the number of newly generated cells in the dentate gyrus and SVZ.33 Lesions and inhibition of 5-HT synthesis using parachlorophenylalanine (PCPA) administration both show that it is the loss of 5-HT that underlies decreases in cell numbers. 5-HT can also promote the survival of neurons in the adult brain, as demonstrated by the abilities of a 5-HT receptor agonist and SSRI to protect neurons against excitotoxic and ischaemic injury in animal models.34,35
Recently there has been some debate as to the extent to which serotonin dysregulation affects hippocampal neurogenesis. In a recent study by Jha and colleagues, systemic administration of the serotonergic neurotoxin 5,7-dihydroxytryptamine, to male Wistar rats, did not alter the proliferation, survival or differentiation of adult hippocampal progenitors.36 By contrast, treatment with PCPA, a serotonin synthesis inhibitor, which also reduces hippocampal noradrenaline levels, resulted in a significant decline in both the proliferation and survival of adult hippocampal progenitors, without affecting differentiation. These results support the view that a number of factors are involved in regulating neurogenesis. Furthermore, the substances used in this experiment were both acute pharmacological treatments. Chronic alterations in serotonin levels appear to have more robust and striking effects on hippocampal neurogenesis, perhaps mediated by the effects of serotonin on specific serotonergic receptors. Given that chronic antidepressant treatment regulates the expression of specific serotonin receptor subtypes,37 it is possible that the neurogenic response to elevated serotonin levels following an acute versus chronic SSRI treatment may be mediated by a different complement of serotonin receptors.
Current evidence already indicates that different receptor subtypes may be responsible for mediating neurogenic activity in discrete areas of the brain.38 Administration of 8-OH-DPAT to rats followed by a single injection of 5-bromodeoxyuridine (BrdU) before killing showed that activation of 5-HT1A receptors produces similar increases in the numbers of BrdU+ cells in the SGL and SVZ. By contrast, activation of 5-HT2C receptors by 1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane (DOI) mediates increased cell proliferation in the SVZ, whereas blockade of 5-HT2A receptors (with ketanserin) produces a 65% decrease in the number of proliferating cells in the SGL, showing that 5-HT2C and 5-HT2A receptors are involved in the regulation of cell proliferation in the SVZ and SGL, respectively. Double labelling for NeuN showed that administration of selective 5-HT1A and 5-HT2C receptor agonists produce consistent increases in the number of newly formed neurons in the dentate gyrus.39 Taken together, these results show that serotonin is capable of promoting neurogenesis, but that different receptor subtypes may promote neurogenesis in specific areas. Further studies are required to determine whether specific serotonin receptors mediate acute versus chronic effects on neurogenesis.
Growth factors and neurotrophins
Enhanced physical exercise via a running wheel in the home-cage environment of rodents also increases levels of key growth factors in the periphery, including fibroblast growth factor (FGF-2),40,41 insulin-like growth factor 1 (IGF-1)42 and vascular endothelial growth factor (VEGF).43,44 Exogenous administration of each of these factors is known to also stimulate neurogenesis in the absence of exercise.45–47
Fibroblast growth factor appears to promote proliferation and differentiation of neural stem cells.48–50 When FGF-2 is overexpressed in mice using viral vector delivery, neurogenesis is upregulated in the dentate gyrus.51 Intercerebral administration of FGF-2 or heparin-binding epidermal growth factor (HB-EGF) to mice at 20 months of age also rescues an age-related decline in neurogenesis.52 Fibroblast growth factor 2 was subsequently shown to be capable of promoting the differentiation of neural stem cells into striatal-like neurons and protecting striatal neurons in toxin-induced models of Huntington's disease (HD).53 Following on from this work, a recent preclinical study has demonstrated that FGF-2 promotes neurogenesis in a mouse model of HD.54 Subcutaneous injection of FGF-2 into R6/2 HD mice was followed by labelling for BrdU. The results were striking; the number of new cells in the SVZ of HD mice treated with FGF-2 was 150% greater than those counted in the control HD mice and five times those observed in the wild-type mice treated with FGF-2. The increase in neurogenesis was accompanied by migration of the new neurons into the striatum where they assumed properties of striatal cells as evidenced by dopamine- and cAMP-regulated phosphoprotein (32 kDa) (DARPP-32) expression, indicating they were capable of differentiating into appropriate neuronal phenotypes. Furthermore, a cohort of R6/2 mice treated with FGF-2 showed increased survival and behavioural improvements on the rotarod. Whether the properties of FGF-2 in this respect are neurogenic or neuroprotective is still contentious,53 but the findings do raise the exciting possibility that FGF-2 could be one future therapeutic target for treatment of brain injury or disease.
Insulin-like growth factor 1 is a polypeptide hormone that induces neurogenesis in the adult mammalian brain both in vivo and in vitro. Subcutaneous infusion of IGF-1 into hypophysectomized rats (with low levels of circulating IGF) for 6 days significantly increased proliferation of neural progenitors in the hippocampal dentate gyrus compared with control animals.45 After 20 days of IGF-1 treatment, a significantly greater proportion of the newborn cells expressed neuronal-specific proteins compared with control animals, suggesting that IGF-1 promotes neurogenesis in the adult rat hippocampus. These findings have been confirmed recently in normal (non-hypophysectomized) rats42 and are consistent with a previous study that demonstrated IGF-1 induced neurogenesis in the adult songbird.55 However, the effects of IGF-1 on proliferation, differentiation and survival of new neurons appears to depend on whether delivery is systemic or central.56 In contrast with these studies, infusion of IGF-1 directly into ventricles of adult brown Norway rats increased hippocampal neurogenesis threefold through a dramatic effect on cell proliferation, but had no effect on neuronal differentiation.56
The blockade of circulating FGF-2 or IGF-1 by injections of a neutralizing antibody has been shown to suppress basal neurogenesis.47,57 Administration of an antibody that blocked the passage of systemic IGF-1 to the brain during running also has been shown to prevent exercise-induced neurogenesis, demonstrating IGF-1 is a necessary component of physical activity-induced neurogenesis.42
Unlike IGF-1, FGF-2 does not increase hippocampal neurogenesis when infused into the adult rat.47,58 However, FGF-2 induces upregulation of IGF-1 receptors and binding proteins59 and may serve to potentiate the mitogenic effects of IGF-1 on neural progenitors.
The molecular basis of growth factor-induced neurogenesis is poorly understood. The observation that the proliferative actions of FGF-2 may be ascribed to MAP/ERK signalling, and the fact that this pathway is also implicated in neurogenesis promoted by neurotrophin 3 (NT-3) and brain-derived neurotrophic factor (BDNF), suggests that common downstream pathways are involved.54 A recent study also implicates the neuronal PAS domain protein 3 (Npas3) transcription factor in FGF-mediated adult hippocampal neurogenesis in mice. Mice homozygous for a null mutation in the Npas3 gene are deficient in the expression of the hippocampal FGF receptor subtype 1 mRNA and have reduced levels of basal neurogenesis compared with wild-type littermates.60
Another growth factor implicated in neurogenesis is VEGF, a secreted angiogenic and neurogenic protein. This exists in two isoforms (i.e. VEGFA and VEGFB), both of which have been shown to have effects on neurogenesis.61,62 When the gene for VEGFB is knocked out in mice, neurogenesis is impaired. Intraventricular administration of VEGFB to these mice restored neurogenesis to levels observed in wild-type mice, thus implicating VEGF in the regulation of neurogenesis.63 Levels of VEGF are also increased by exercise,44 and in wheel-running mice.63 Peripheral blockade of the angiogenic effects of VEGF using a viral vector encoding the antagonist sFlt has been shown to abolish running-induced neurogenesis in C57/BL6 wild-type mice,64 suggesting that VEGF could be acting both indirectly through modulation of the neurovasculature where hippocampal neurogenesis occurs,65 thus promoting transport of other endogenous neurotrophins, and directly on neural precursor cells located within the vascular niche.61 Infusion of VEGF into the lateral ventricle of adult rats resulted in reduced apoptosis but unaltered proliferation, suggesting that VEGF has a survival promoting effect on neural progenitor cells rather than directly upregulating neurogenesis per se.
Recent evidence indicates that VEGF also may act in a biphasic dose-dependent manner. Administration of a high dose of VEGF (500 ng/mL) to adult mice was shown to downregulate VEGF receptors 1 and 2 in the SVZ, concomitant with an increase in the proliferation and differentiation of neural progenitors. By contrast, a low dose of VEGF upregulated the endogenous VEGF receptors, but did not increase proliferation and differentiation.66 It is possible that a similar form of feedback regulation may be employed in other growth factor pathways that promote neurogenesis, thus enabling increasing neurogenesis in response to demand (e.g. after ischaemic or focal trauma).
There is strong evidence to suggest that VEGF may be a key component of exercise-induced neurogenesis. Hippocampal VEGF levels are increased by environmental enrichment and learning in a spatial maze, both paradigms being associated with increased neurogenesis.67 Overexpression of VEGF using gene transfer in adult rats resulted in increased neurogenesis, together with improved performance of the rats in cognitive tests. By contrast, inhibition of VEGF expression by RNA interference completely blocks the environmental induction of neurogenesis.67
A number of neurotrophic substances, such as BDNF, also have been shown to promote neurogenesis and cell survival68–71 both during development and after ischaemic trauma to the adult brain.72 Brain-derived neurotrophic factor appears to work in feedback loops with a number of other neurogenic factors, including serotonin,73,74 FGF-275 and nitric oxide (NO),69 to promote the differentiation and survival of new neurons. Treating explants of adult human SVZ cells with FGF-2 (for 1 week) followed by BDNF (for 8 weeks), resulted in a substantial increase in the number of surviving neurons after 9 weeks compared with those cells treated with unsupplemented medium or BDNF or FGF-2 alone.75 Serotonin, by its action on 5-HT receptors coupled to adenosine 3",5"-cyclic monophosphate (cAMP) production and cAMP response element binding protein (CREB) activation, influences BDNF levels through increased transcription of the BDNF gene, and NO has been reported to act in a positive feedback loop with BDNF to promote proliferation and differentiation.69 Paracrine NO signalling appears to contribute to the BDNF-induced transition of neuronal precursor cells from proliferation to differentiation, thus promoting neurogenesis.
Lesion studies on rodents have shown that key neurotransmitters such as acetylcholine and noradrenaline are required for the proliferaton and survival of new neurons. Lesions of the forebrain cholinergic input in Sprague-Dawley rats, using stereotaxically targeted injections of 192 IgG-saporin, reduced dentate gyrus neurogenesis with a concurrent impairment in spatial memory. By contrast, systemic administration of the cholinergic agonist physostigmine increases neurogenesis in the dentate gyrus. Furthermore, mice lacking the β-2 subunit of the neural nicotinic acetylcholine receptor have significantly decreased cell proliferation in the hippocampus compared with age-matched controls.76 Considered together, these results indicate that impaired cholinergic function may contribute to deficits in learning and memory through a decrease in the production of new hippocampal neurons.77
Noradrenergic denervation by the noradrenaline neurotoxin DSP-4 also has been shown to reduce the proliferation, but not the survival or differentiation, of hippocampal progenitor cells.78 The authors suggest that this could be a direct result of the loss of noradrenergic input to α1-adrenoceptors or could be occurring indirectly through loss of other trophic factors, including FGF-2, expressed by noradrenergic neurons.79
The adult hippocampus receives excitatory γ-aminobutyric acid (GABA)-ergic inputs, which have been shown to promote activity-dependent differentiation of progenitor cells in culture.80γ-Aminobutyric acid also promotes the synaptic integration of newly generated neurons in the adult brain.81 The release of GABA from hippocampal GABA-ergic interneurons is, in turn, regulated by a number of other neurotransmitter substances, including serotonin,82 and cannabinoids,83 which have been implicated also in the regulation of adult neurogenesis.
The N-methyl-d-aspartate (NMDA) receptors subtype of glutamate receptors plays a key role in the regulation of adult hippocampal neurogenesis. In the dentate gyrus of the adult rat, activation of NMDA receptors rapidly decreases cell proliferation, and blockade of NMDA receptors rapidly increases precursor proliferation.84 Acute treatment with the NMDA receptor antagonists MK-801 and CGP 37849 increased the birth of neurons and increased the overall density of neurons in the granule cell layer. However, there does not appear to be any significant functional expression of NMDA-Rs on adult progenitor cells80,85 in the adult dentate gyrus, so this effect may be mediated by GABA-ergic inputs onto hippocampal progenitor cells, as GABA release from hippocampal interneurons requires NMDA-R activation,82 and can be blocked by in vivo administration of the NMDA-R antagonists MK-801.86
Recent work has shown that embryonic and adult hippocampal neural precursor cells are both immunoreactive for CB1 cannabinoid receptors, thereby indicating that cannabinoids could also be involved in the regulation of neurogenesis. Indeed, chronic adminstration of HU210, a synthetic cannabinoid, has been shown to promote neurogenesis in the dentate gyrus of adult rats. The increased neurogenesis also produced anxiolytic and antidepressant-like effects, as determined by improved performance of the rats on the novelty-suppressed feeding test and the forced swim test.87 These findings are in agreement with other studies showing that mice lacking the CB1 receptor have significantly decreased neurogenesis,88 whereas those deficient in the endocannabinoid-inactivating enzyme fatty acid amide hydrolase (FAAH) have increased neurogenesis.89
Studies on rats have shown that neuronal proliferation is inhibited by treatment with exogenous corticosterone and the glucocorticoid receptor agonist dexamethosone,90 whereas adrenalectomy dramatically increases neurogenesis.91–93 Adrenal hormones appear to mediate deleterious effects of stressful experiences, which suppress neurogenesis in the adult dentate gyrus.92 Corticosterone, the main adrenal steroid in rodents, regulates its own secretion through negative feedback by interacting with two receptors present in the dentate gyrus that differ in their affinity. The type I receptors, mineralocorticoid receptors, have a high affinity for corticosterone, whereas the type II receptors, glucocorticoid receptors, have a low affinity for the hormone.94 However, very few precursor cells in the dentate gyrus express either of these receptors,91 so it is possible that adrenal steroids regulate cell proliferation indirectly by increasing glutamate release in the dentate gyrus which, in turn, inhibits neurogenesis through NMDA receptors. These increases in glutamate within physiological parameters may not be excessive and thus are likely to be below neurotoxic levels.
It is interesting to note that during the first two postnatal weeks, the period during which the majority of granule cells are generated, levels of adrenal steroids are low, whereas older animals experience more sustained stress-induced increases in adrenal steroids than younger animals do.95 However, the provision of enriched environments to adult mice and rats decreases stress responses and reduces levels of circulating corticosterone and adrenocorticotrophic hormones.96–98 Therefore, it appears that as well as having direct effects on neurogenesis through growth factors, enrichment may be acting indirectly to increase neurogenesis via a reduction in the suppressive effects of adrenal steroids on neurogenesis.
A number of toxins and drugs have been reported to also decrease neurogenesis in the adult hippocampus, including opiates,99 alcohol,100 nicotine,101 cocaine102 and methylazoxymethanol (MAM).103 The effect of ethanol appears to depend on the dose administered, with lower doses of ethanol being reported to increase neurogenesis,57 whereas progressively higher doses appear to inhibit neurogenesis.100,104 One potential confound in all such in vivo studies of adult hippocampal neurogenesis is that any treatment that alters the home-cage activity of the animals could have indirect effects on neurogenesis, since we know that increased physical exercise alone can enhance neurogenesis.21 Therefore, it is crucial that such studies on hippocampal neurogenesis include appropriate behavioural tests, such as measures of spontaneous activity.
Genetic constraints on neurogenesis
Given the number of factors affecting neurogenesis, it is hardly surprising that a number of genetic loci have been implicated in the regulation of neurogenesis in the adult brain. Differences in neurogenic potential have been reported in various mouse strains,26 with concomitant differences in performance on hippocampal-dependent learning tasks.105 This suggests that neurogenesis may contribute to hippocampal-dependent memory and that genotype regulates the extent of neurogenesis in the adult. For example, C57/BL6 mice have higher levels of baseline neurogenesis than DBA2 mice. When these mice were bred to generate 10 different strains of mice, those with higher levels of neurogenesis also demonstrated improved learning in the water maze task.105 Although genetic background may indeed influence hippocampal neurogenesis, it does not follow that the relevant genes are directly involved in regulating neurogenesis. Some genes could be promoting indirect effects such as the upregulation of hormones or growth factors, whereas others may render a particular strain of mouse more susceptible to processes that have deleterious effects on neurogenesis, such as excitotoxic cell death. Furthermore, other genes might be involved in regulating the spontaneous activity levels of the mice, thereby producing an indirect impact on levels of hippocampal neurogenesis.
REGULATORS OF NEUROGENESIS IN THE DISEASED ADULT BRAIN
It has been proposed that neurodegenerative diseases, such as Huntington's, may be partly attributed to a failure of neuroregenerative processes.106 Indeed, recent studies have shown that neurogenesis is altered in a number neurodegenerative diseases, including Alzheimer's,107,108 Huntington's,109 and also with ageing110 and epilepsy.111 What follows is a discussion of the manner in which neurogenesis is altered in each of these pathological states, and the treatments that could potentially reverse neurogenic aberrations and thus mitigate disease symptoms.
Post-mortem studies on human brains show that neurogenesis appears to be increased in the SVZ of patients with HD.109,112 Studies in the quinolinic acid (QA) model of HD are in agreement with these findings, demonstrating that the selective loss of GABA-ergic medium spiny neurons in the QA-lesioned striatum results in increased progenitor cell proliferation and compensatory neurogenesis.113,114 However, the combination of increased progenitor cell proliferation and compensatory striatal neurogenesis has been observed in a number of models of neurodegenerative diseases,115–119 and is probably related to cell death levels and inflammation. These findings contrast with those from the R6/1120,121 and R6/2122 mouse models of Huntington's disease, which show little difference in basal SVZ neurogenesis from their wild-type littermates and a striking decrease in neurogenesis in the dentate gyrus of the hippocampus.120,122 However, there is no difference in the in vitro growth of the adult neuronal precursor cells between the R6/1 and R6/2 genotypes. This could be owing to the in vivo microenvironment in which the cells are located rather than reflecting a dysfunction of the neuronal precursor cells per se.123 Although these models do show cell death in the striatum,124,125 the excitotoxic stimulus promoting progenitor cell proliferation may not be as great as that produced by the QA lesion model or the human HD brain at late stages of the disease.
Further support for the notion that environmental cues from dying cells in the striatum may promote progenitor cell proliferation comes from recent work on the role of activating transcription factor 2 (ATF-2) in neurogenesis.126 This protein is highly expressed in the brain stem, substantia nigra and also the granule cells of the hippocampus. Although levels of ATF-2 are decreased in the hippocampus in a number of neurological diseases, including Huntington's, Alzheimer's and Parkinson's, one study has demonstrated that ATF-2 levels are actually increased in the subependymal layer in HD,126 a region reported to contain increased numbers of proliferating progenitor cells. It is possible that ATF-2 is activated by environmental cues related to cell death in the HD striatum and that increased AFT-2 reflects the proliferative response of SVZ progenitor cells to striatal cell death. Such a mechanism could, in part, account for the increased neurogenesis observed in the SVZ of human HD brains at post-mortem.
Alternatively, it may be that ATF-2 acts as an effector of progenitor cell proliferation in the SVZ, and that the reduction of ATF-2 observed in the HD hippocampus may prevent progenitor cell proliferation, as seen in the transgenic R6/1 and R6/2 lines. Although further work is needed to confirm whether aberrations in the levels of ATF-2 occur in mouse models of HD, these findings do suggest that ATF-2 and other microenvironmental cues, possibly from the striatum, influence adult neurogenesis in an, as yet, unknown manner.
Transgenic HD mice have been shown to have decreased levels of a key neurotrophin, BDNF, in the striatum and hippocampus.127 There is evidence supporting disruption of BDNF levels with transcriptional dysregulation,128 as well as protein trafficking.127,129,130 The fact that environmental enrichment delays disease onset in HD mice28,131,132 and ameliorates a hippocampal BDNF deficit127 is consistent with the effects of enrichment on hippocampal neurogenesis.133 Whether enrichment-induced neurogenesis has clear functional benefits in wild-type rodents is still unclear,134 and a recent study indicates that the behavioural effects of enrichment are not dependent upon enhanced hippocampal cell proliferation.135
Alzheimer's disease (AD) is a progressive neurodegenerative disorder caused by the presence of senile plaques and neurofibrillary tangles in areas of the brain, including the hippocampus and neocortex. Affected individuals undergo a general cognitive decline, behavioural and personality changes and, as the disease advances, motor complications and severe dementia develop, rendering them incapacitated and dependent on caregivers.
In parallel with research on neurogenesis, several mouse models of AD also have been found to have decreased hippocampal neurogenesis.136–141 Furthermore, neurogenesis in the post-mortem AD brain142 and in one AD mouse model107 appears to be increased, not decreased. It is interesting that in both AD and HD, hippocampal neurogenesis appears to be decreased in most mouse models, raising the possibility that, in both disorders, similar mechanisms might underlie the pathological aberration in neurogenesis, despite the differing genetic mutations that are present.
A recent study to quantify and characterize adult-generated neurons in the PDAPP mouse model of AD showed that old, but not young, PDAPP mice have a 50% reduction in neurogenesis and that new SGZ neurons exhibit abnormal maturation. Furthermore, the investigators also reported that ectopic neurogenesis occurred in the outer granule cell layer (oGCL) in PDAPP mice. However, these cells do not survive to maturity, thus providing a potential explanation of the differing findings in human brain and mouse models of neurodegenerative disease. It would be interesting to see whether mouse models of HD also exhibit similar ectopic neurogenesis in the oGCL.
As hippocampal learning can stimulate neurogenesis,143 it has been postulated that the decreased SGZ neurogenesis reported in the PDAPP mouse is secondary to the diminished learning capacity present. Dissociating whether decreased SGZ neurogenesis precedes or follows degeneration of cortical afferents to the GCL will help distinguish which is affected first – neurogenesis or learning. Does decreased learning and activity in disease decrease neurogenesis or vice versa? This is a question that has yet to be addressed specifically, although the observation that cell proliferation is decreased even in very young HD mice,122,133 prior to the onset of cognitive deficits and the decreased activity that accompanies disease progression, indicates that any learning deficits are a consequence and not a cause of aberrations in neurogenesis.
It appears that several factors may be affecting neurogenesis in AD. The mutant amyloid precursor protein (APP) may interfere directly with cell proliferation and maturation144 or indirectly through the production of Aβ42, a peptide that has been shown to have negative effects on cell survival in vivo and in vitro.145 Equally, the loss of other growth factors reviewed here might also contribute to decreased SGZ neurogenesis and enhanced GCL neurogenesis. Given that acetylcholine is required for proliferation of neuronal precursors, it is unsurprising that the loss of ascending basal forebrain cholinergic inputs in AD 146 has been shown to reduce dentate gyrus neurogenesis in rats, together with impaired spatial memory performance of these animals in the Morris water maze.77 Dietary restriction, shown to be beneficial in models of AD,147 upregulates neurogenesis,148,149 which is associated with an increase in BDNF.149 Levels of IGF-1 are altered in AD150 and levels of serotonin are decreased.151 Enhanced immunoreactivity for VEGF in the brains of patients with AD has been observed post-mortem152 and the serum levels of VEGF are also markedly increased in patients with AD.153 As vascular abnormalities are present in individuals with AD, the increased VEGF could reflect a compensatory repair mechanism, as well as contributing to changes in neurogenesis seen in AD. Clearly, however, further study is needed to dissect the potential contribution of each of these factors to the hippocampal pathology reported in AD.
The pathological effects of stress and depression also affect neurogenesis. Preclinical and clinical studies demonstrate that stress and depression are associated with reductions of the total volume of this structure, and atrophy and loss of neurons in the adult hippocampus. Several studies have implicated altered regulation of neurogenesis in mood disorders.19,20,154 Stress increases the levels of circulating corticosteroids, which are known to influence hippocampal plasticity during adulthood in a complex manner.155 Experimental increases in the levels of adrenal steroids result in significant decreases in the rate of granule cell production in adulthood.156 Furthermore, the removal of adrenal steroids stimulates the proliferation of granule cell precursors,92 the vast majority of which differentiate into neurons.156
Transgenic R6/1 HD mice also show a depressive phenotype on the Porsolt forced swim test, modelling the depression seen in many patients with HD. Concomitant with the depressive phenotype in this mouse model is a reduction in hippocampal neurogenesis. Furthermore, both the depressive phenotype and reduced hippocampal neurogenesis are rescued by the administration of fluoxetine, an SSRI.120
Adult neurogenesis may also be altered in other psychiatric disorders. For example, there is evidence for decreased hippocampal volume in schizophrenia, a disorder involving cognitive symptoms such as spatial working memory and olfactory deficits, as well as psychotic symptoms. Therefore, we propose that hippocampal and SVZ neurogenesis should be investigated in post-mortem tissues from patients with schizophrenia.
Given that fluoxetine ameliorates disorders of mood and neurogenesis in a transgenic mouse model of HD, it appears that serotonin dysregulation contributes to the pathogenesis of these deficits. Several other studies support this notion. Levels of 5-HT are diminished in R6/2 HD mice157 and decreased 5-HT receptor binding sites have also been reported in human studies.158–160 Serotonin is also involved in the regulation of learning and memory via its action on autoreceptors located on 5-HT neurons (raphe nuclei) or as a heteroreceptor on non-5-HT neurons, mainly in the hippocampus. 5-HT1A knockout mice show impaired spatial learning on the Morris water maze,161 as do mice treated with serotonin re-uptake inhibitors.162
There is conflicting evidence concerning the effect of epilepsy and seizures on neurogenesis. A number of studies indicate that seizures promote neurogenesis in the dentate gyrus and SVZ,163–169 a phenomenon thought to represent a compensatory response to the loss and damage of existing cells.170,171
Experimental paradigms of seizure-induced activity differ between laboratories, thereby limiting direct comparison between studies. However, the available evidence indicates that the effect of isolated seizures on neurogenesis may be different to the effect of chronic seizure activity seen in conditions such as temporal lobe epilepsy (TLE). A study by Kralic and colleagues demonstrated that although acute status epilepticus, induced by kainic acid injection in a mouse model, increased cell proliferation in the subgranular zone and dentate gyrus, models of temporal lobe epilepsy showed a marked reduction in neurogenic potential.172 This was attributed to the dispersion of dentate granule cells seen in the brains of these mice – a feature also observed in many patients with TLE.173
Hippocampal gene expression profiles obtained using microarrays performed on tissue from Sprague-Dawley rats after induction of electroconvulsive seizures (ECS) show upregulation of growth factors, transcription factors, neurotrophic and angiogenic factors.174 Many of the same molecular factors also have been implicated in the regulation of seizure-induced neurogenesis, including FGF,175 BDNF,176 NO,177 FGF-2178 and galanin.179
Studies on animal models using a middle cerebral artery occlusion (MCAO) protocol have shown that ischaemic trauma promotes neurogenesis in the subgranular zone/GCL and SVZ, and accumulating evidence indicates that these new neurons act to replace those lost at sites of degeneration.119,180–184 Recent work indicates that ischaemia-induced neurogenesis can persist for up to 4 months after the initial insult.185
Environmental enrichment and exercise have been shown to improve motor and cognitive outcomes after transient MCAO in rodents.186–190 The increased neurogenesis postischaemia and the beneficial effects of enrichment and exercise could well be accounted for by the action of a number of neurotrophins and growth factors that were upregulated following ischaemia.191 Factors reported to be increased secondary to ischaemia include bFGF,192,193 BDNF,194 glial cell line-derived neurotrophic factor (GDNF),195 epidermal growth factor, FGF-2,196 tumour necrosis factor,197 erythropoietin198 and bone morphogenic protein.199
Inflammation also accompanies ischaemic insults, although there is conflicting evidence on the impact of inflammatory mediators on adult neurogenesis. Administration of pharmacological agents such as indomethacin200,201 and tissue kallikrein, both of which decrease inflammation, increases progenitor cell proliferation following stroke. However, mice deficient in C3, a powerful inflammatory mediator, show decreased neurogenesis after transient MCAO.202 Further work is needed to clarify the effect of inflammation and intracerebral pressure on neurogenesis, and functional outcomes following stroke if effective pharmacological treatments for brain repair are to be developed.
TREATMENT OF BRAIN INJURY AND DISEASE VIA UPREGULATION OF NEUROGENESIS: A THERAPY FOR THE FUTURE?
Given the evidence reviewed earlier, would stimulating neurogenesis by targeting of the postulated pathways involved produce a functional improvement in individuals with neurodegenerative diseases? The potential therapeutic utility of various manipulations known to enhance neurogenesis is discussed below.
In a recent study, we treated R6/1 HD mice and wild-type controls (aged 10–20 weeks) with a course of either fluoxetine (an antidepressant with SSRI activity) or saline. Extensive phenotypic testing to determine the effect of fluoxetine on disease progression and behaviour was followed by analysis of the brains with a marker of cell division (i.e. BrdU) and a neuronal marker (i.e. NeuN) to determine whether promoting neurogenesis via administration of fluoxetine would mitigate symptoms of HD in a transgenic mouse model.
The results demonstrated that the R6/1 mouse model of HD has decreased neurogenesis in the dentate gyrus concomitant with affective deficits on the Porsolt forced swim test and cognitive deficits as measured by decreased alternation on the T-maze.120 Furthermore, fluoxetine administration mitigated all of these symptoms. In fluoxetine-treated HD mice, there was an increase in neurogenesis, with a concurrent improvement in performance on behavioural tests. Fluoxetine-treated mice showed a reduction in immobility time on the forced swim test, indicating an improved affective state compared with their saline-treated counterparts. The HD mice treated with fluoxetine also had a significantly higher percentage of alternation on the T-maze test, which measures the ability of the mouse to remember the arm entered previously, and scores behavioural performance according to whether the mice spontaneously alternate between trials.120
One implication of these findings is that SSRI such as fluoxetine, which has been found to rescue hippocampal neurogenesis and cognitive deficits in HD mice, may also have therapeutic benefit in AD, and we are currently exploring this possibility. Furthermore, such antidepressants, and other pharmacological modulators of neurogenesis, may have therapeutic potential for a wider range of brain disorders.
Behavioural and dietary modifications
As well as pharmaceutical treatments, environmental enrichment, enhanced exercise and dietary modifications may provide additional means of increasing neurogenesis and mitigating symptoms of brain injury or disease, which could be combined with pharmacotherapy. The effects of environmental enrichment and increased physical exercise on hippocampal neurogenesis have been described earlier. Dietary restriction (DR) in rodent models of stroke, AD, HD and Parkinson's disease (PD) has been shown to improve functional outcomes, possibly via a neuroprotective effect. Dietary restriction also has been shown to increase neurogenesis in the hippocampus in mice and also increases levels of neurotrophic factors such as BDNF and heat shock proteins.148 When heterozygous BDNF knockout (BDNF +/–) mice and wild-type mice were maintained for 3 months on either restricted or ad libitum diets, levels of BDNF protein in hippocampal neurons were decreased in the BDNF +/– mice but were increased by DR in the wild-type mice and, to a lesser extent, in the BDNF +/– mice. The reduced neurogenesis in BDNF +/– mice was associated with a significant reduction in the volume of the dentate gyrus, thus suggesting that BDNF is important in regulating basal levels of neurogenesis and, furthermore, that DR is important in enhancing neurogenesis.148 However, most studies do not examine whether DR induces increased physical activity (i.e. via weight loss and the motivation of hunger), which could be the reason why some of its effects are similar to those of environmental enrichment and wheel running. It is also possible that DR, environmental enrichment and exercise all induce a mild stress response which, in turn, may act as the trigger for the neurotrophic changes, such as increased BDNF promoting increased cell proliferation and differentiation. A recent study showing that folate deficiency and subsequent hyperhomocysteinaemia may adversely affect brain function and plasticity through deranged one-cell metabolic processes203 lends further support to the argument that dietary components may also affect neurogenesis.
Electroconvulsive therapy, a clinically proven treatment for depression and refractory mania, has been shown to increase hippocampal neurogenesis. This could be because of the increased neural activity concomitant with an increased production of growth factors and neurotrophins.174 Whether this therapy would be of benefit to patients with neurodegenerative disorders is, as yet, unknown.
Although these findings are encouraging, further work is still needed to ascertain whether a sustained increase in neurogenesis as a result of activity-driven or pharmaceutical means would produce a long-term functional benefit to patients with neurodegenerative diseases such as HD and AD, as well as other brain disorders. Ultimately, it is hoped that a greater understanding of the factors involved in regulating adult neurogenesis may pave the way for developing suitable treatments and preventative strategies to delay the onset and mitigate the symptoms of a number of devastating brain disorders.
We thank Ms CM Hannan and Dr J Nithianantharajah for comments and editing assistance with the manuscript. AJH is supported by an RD Wright Award and project grants from the Australian National Health and Medical Research Council and by the Lord Mayor's Charitable Fund (Eldon and Anne Foote Trust).