Adult neurogenesis is the process of generating new neurons which integrate into existing circuits after fetal and early postnatal development has ceased. In most mammalian species, adult neurogenesis only appears to occur in the olfactory bulb and the hippocampus. In addition there is a high level of adult neurogenesis in the olfactory epithelium (considered part of the peripheral nervous system) where olfactory receptor neurons are constantly replaced.
The process appears more widespread but still limited in other vertebrate classes, having been described in select brain regions of certain birds, fish and reptiles. Furthermore, many invertebrates and vertebrates have neural regenerative capacities that involve neurogenesis (such as tail regeneration in salamanders).
Neural Stem Cells
New neurons are generated throughout life from a population of dividing cells known as neural stem/progenitor cells (NPCs). Two criteria are typically used to define a cell as a stem cell:1) the potential of self-renewal and 2) the ability to give rise to multiple distinct cell types. NPCs isolated from the adult brain are classified as a ‘’multipotent’’ cell because they can differentiate into the the three main lineage cell types of the nervous system (neurons, astrocytes, and oligodendrocytes) when cultured in vitro. The evidence for multipotency of NPCs in vivo remains scant.
Figure 1: Locations of adult mammalian neurogenesis
There are two neurogenic regions in the adult brain where under physiological conditions NPCs give rise to new neurons: (1) the subventricular zone of the lateral ventricles (SVZ) where NPCs generate cells that migrate into the olfactory bulb, and (2) the subgranular zone (SGZ) of the dentate gyrus (DG) where new granule cells become integrated into the local neuronal network ( Figure 1).
For those two regions several types of dividing progenitors were identified . The “type-1” cells (or ‘B’ cells in the SVZ) are similar to the radial glial cells observed during development, and have a morphology and physiology similar to mature astrocytes. Although they reside in the SGZ, they extend processes up into the molecular layer. Type-1 and B cells are relatively quiescent. In contrast “Type-2” cells (or ‘C’ cells in the SVZ),have a high proliferative activity but have a small roundish morphology. The current hypothesis is that Type-2 cells (or ‘C’ cells) give rise to Type-3 (or A cells) representing neuronally committed neuroblasts.
NPCs are not limited to neurogenic regions of the brain, rather their proliferation can be observed in most CNS regions, especially after injury. However, in these other regions it appears that neurogenesis is actively repressed by the local environment – NPCs from non-neurogenic regions have been observed to give rise to neurons when transplanted into the hippocampus. Some evidence indicates that this effect is mediated by the local astrocyte populations.
NPCs have historically been labeled in the brain by the addition of a proliferation marker, such as 3H-thymadine or bromodeoxyuridine (BrdU; Figure 2, bottom). Immunohistochemistry for BrdU can be combined with the detection of mature markers to identify the phenotype of the newborn cells. Recently, several molecular techniques for labeling adult-born cells have been developed, including transgenic mice with GFP driven by a stem cell gene’s promoter (i.e., Nestin-GFP; Figure 2, top left) and retroviral labeling ( Figure 2, top right). BrdU labeling has been used to definitively show that new neurons are incorporated into the dentate gyrus and olfactory bulb of the adult human brain (Eriksson et al., 1998; Curtis et al., 2007).
Maturation of New Neurons
Adult neurogenesis is unique from developmental neurogenesis because the new neurons must integrate into an established, functioning network. Much of the present knowledge about neuronal development in adult neurogenesis has been reviewed by Kempermann et al.(2004), Ming and Song (2005). Abrous et al. (2005), and Zhao et al.(2008).
Maturation of new neurons in the adult dentate gyrus
Figure 3: Maturation of newborn granule cells
The process of adult hippocampal neurogenesis is entirely confined to the dentate gyrus. Local progenitor cells in the SGZ undergo neuronal differentiation, and may show a limited migration into the GCL. The speed of maturation is likely experience dependent, and varies between neurons. Approximate duration of a number of distinct post-mitotic developmental phases of newborn granule cells are listed here. Figure 3 shows a scematic of the anatomical phases of granule cell growth.
Local GABA: (Less than 1 week old) Immature neurons have neurite outgrowth, but often not polarized towards molecular layer. Few or no synapses, but sensitive to locally diffuse GABA, which is depolarizing.
Synaptic GABA: (1 to 2 weeks old) Dendrites begin to extend into molecular layer (no spines) and axons can be observed in hilus. Synaptic GABA inputs can be observed, which is still excitatory. Glutamatergic inputs are not present. Immature action potentials can be observed when cells are directly stimulated.
Spine formation onset and axon outgrowth: (2 weeks old). By about 16 days neurons begin to develop spines in the molecular layer. GABA transitions to inhibitory around this time. By 17 days, new axons (mossy fibers) can be observed forming functional connections onto downstream hilar neurons and CA3 pyramidal cells.
Functionally immature neurons (~3 weeks to 2 months) Spine formation is gradual, with neurons progressively increasing their dendritic arborization and connections. Mossy fibers continue to mature, with boutons on CA3 neurons growing considerably by 28 days. Neurons still have unique physiological properties, including increased LTP, and different resistance, capacitance and resting potentials.
Fully mature neurons (> 2 months old). Newborn neurons eventually become physiologically indistinguishable from fully mature neurons.
Recent work using immediate early genes such as c-fos, Zif268, and Arc as putative markers of neuronal activity have shown that water maze training (Kee et al., 2007) or exposure to an enriched environment (Tashiro et al., 2007) during this maturation process will cause these neurons to be more responsive upon reexposure to the same condition several weeks later.
Maturation of New Neurons in the Adult Olfactory Bulb
In contrast to adult neurogenesis in the dentate gyrus, cells that were born in the SVZ migrate a long distance into their target area, the olfactory bulb. This long migration gives olfactory neurogenesis a different timescale from DG neurogenesis.
- Migration: (2-6 Days) Newborn cells migrate in chains along the rostral migratory stream (RMS), a structure maintained by specialized astrocytes. After the newborn neurons reach the middle of the OB they detach from the chains and migrate radially.
- Neuronal Differentiation: (15-30 Days) After immature neurons reach the OB, they begin to differentiate into two different types of local interneurons. Over 95% differentiate into GABA-ergic granule neurons whereas the remainder become periglomerular neurons expressing either GABA and/or dopamine as neurotransmitter. Newborn granule cells can be distinguished into cells with dendrites that do not extend beyond the mitral cell layer and other cells that possess non-spiny dendrites reaching into the external plexiform layer.
- Integration into network: (15 -30 Days) Newborn granule cells and periglomerular neurons become integrated into the OB circuitry and respond to olfactory stimuli.
Neuronal Selection and Survival
One critical aspect of adult neurogenesis is the selection process. While large numbers of new neurons are born to the OB and DG, only a fraction of these cells survive. In the dentate gyrus, approximately half of the newborn neurons die within 2 weeks of birth, but this number is heavily regulated by various factors. In contrast to newborn DG neurons the selection process in the OB appears to be later in the development process, when young neurons with extended dendrites already covered with spines are susceptible to cell death.
Regulation of Neurogenesis
The “rediscovery” of neurogenesis in the 1990’s was due in large part to the observation that the levels of new neurons in the adult hippocampus are modulated by a range of factors, including stress (Gould et al., 1990), aging (Kuhn et al., 1996), environment (Kempermann et al., 1998), and activity (van Praag et al., 1999). Numerous drugs and behaviors have since been shown to affect the levels of new neurons in the brain.
Modulation of neurogenesis typically occurs in one of two ways in vivo – either the modulator changes the levels of proliferation of NPCs, or the effect is on the survival of the new neurons. The most studied modulators have been summarized in the following tables. See Ming and Song (2005) and Abrous et al.(2005) for more details.
Several neuro-psychiatric conditions have been associated with altered rates of neurogenesis in animal models, including Alzheimer’s disease, temporal-lobe epilepsy, ischemia, and depression. In each of these cases, it remains unclear whether perturbed neurogenesis is a symptom of the disorder or has a causal role. Aging also has a robust effect on neurogenesis, with levels of new neurons decreasing in later stages of life. The marked decrease occurs fairly early and neurogenesis is maintained at a very low level for most of the life span.
Function of Neurogenesis
While the observation and characterization of neurogenesis has been robust, the role of adding new neurons on a region’s function has remained elusive in most cases. Nonetheless, because neurons are integrating into regions of relatively well described circuitry and function, several behavioral and computational ideas have been explored.
Hippocampus-dependent Behavioral Tasks
Several techniques to reduce adult neurogenesis have been used to look at the process’s effect on hippocampal function. These have included x-ray irradiation, anti-proliferative drugs (MAM) and molecular knock-downs. A range of hippocampus-dependent behaviors have been tested with mixed results (see Deng et al., 2010 for a review). Trace eyeblink conditioning was shown to be affected in MAM experiments, and contextual fear conditioning was impaired following irradiation and genetic ablation of adult neurogenesis.
Morris Water Maze (MWM) testing has shown inconsistent results in several paradigms, with some experimenters seeing deficits in short-term retention, others in long-term retention, and others no discernable differences at all. Furthermore, set of behavioral studies have demonstrated that neurogenesis may have a role in the pattern separation function of the dentate gyrus (Clelland et al., 2009). Finally, a recent study has suggested that new neurons may be important in memory consolidation (Kitamura et al., 2009).
In addition to its presumed role in memory, the correlation of neurogenesis levels to stress has suggested a role in anxiety-related behaviors. For example, fluoxetine (the active compound in Prozac) is not effective as an anti-depressant in mice without adult neurogenesis due to irradiation.
Olfactory Bulb-dependent Behavioral Tasks
The function of the olfactory pathway can be tested with a variety of behavioral tasks that test odor discrimination or odor learning. Using transgenic mice with reduced OB neurogenesis it could be shown that new OB neurons appear to be critically involved in odor discrimination. At the same time odor discrimination learning itself increases the survival of newborn OB neurons. The same effect on survival has been found using odor enrichment resulting in improved odor memory.
Computational Impact of New Neurons
Because the dentate gyrus is the entry structure to the hippocampus, which has a substantial history of neural network modeling, several non-exclusive computational functions have been suggested for neurogenesis. These have arisen from both theoretical and computational modeling ventures. For a more detailed review of the theoretical functions of adult neurogenesis (see Aimone, Deng, and Gage; 2010).
Increase of hippocampal memory capacity – several models predict network capacity will increase with neurogenesis, but Becker’s model (2005) explores the idea in a full hippocampal model. Becker predicts that the increase in possible sparse codes due to new neurons increases the quality of memory formation in downstream hippocampal regions.
Reduction of interference between new and older memories – Wiskott and colleagues (2005) propose that the presence of new neurons helps the dentate gyrus network respond to changing inputs. Specifically, their model suggests that without neurogenesis, the hippocampal network will suffer from “catastrophic interference,” leaving the network unable to effectively encode new memories.
Encoding time in memories – Aimone et al. (2006; 2009) suggest that the different physiological properties of immature neurons will bias the sparse coding function of the dentate gyrus, possibly providing a link between memory associations formed in the recurrent CA3 network.
Olfactory bulb neurogenesis has not been as extensively studied computationally, possibly because the olfactory bulb circuit does not have the history of modeling that the hippocampus has. Cecchi et al.’s (2001) theoretical study of OB neurogenesis suggests functional roles similar to those suggested for newborn neurons in the dentate gyrus. Cecchi’s results suggest that random incorporation of new neurons with activity-dependent survival will maximize the discrimination of odors presented to the network.
Recommended Reading and Links
“Adult Neurogenesis” eds. Fred H Gage, Gerd Kempermann, Hongjun Song. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 2007.
“Adult Neurogenesis” by Gerd Kempermann, Oxford University Press; 2 edition (January 26, 2011)
James B. Aimone’s website
Sebastian Jessberger’s website
Fred H. Gage’s website
Minor Changes Made by Sciencebeta.com: Omitted section on Non-human Neurogenesis, Incorporated Citation References as Hyperlinks