SEARCH

SEARCH BY CITATION

Keywords:

  • adult neurogenesis;
  • cell death;
  • dentate gyrus;
  • olfactory bulb;
  • postnatal development

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. PCD of neurons in the early postnatal brain (<P30)
  5. PCD in the adult brain
  6. Factors regulating PCD in the adult brain
  7. Significance of PCD in the adult brain
  8. Conclusion and perspectives
  9. Acknowledgments
  10. References

During development, elimination of excess cells through programmed cell death (PCD) is essential for the establishment and maintenance of the nervous system. In many brain regions, development and major histogenesis continue beyond postnatal stages, and therefore, signs of neurogenesis and PCD are frequently observed in these postnatal brain regions. Furthermore, some brain regions maintain neurogenic potential throughout life, and continuous genesis and PCD play critical roles in sculpting these adult neural circuits. Although similar regulatory mechanisms that control PCD during development appear to also control PCD in the adult brain, adult-generated neurons must integrate into mature neural circuits for their survival. This novel requirement appears to result in unique features of PCD in the adult brain. In this article, we summarize recent findings related to PCD in the early postnatal and adult brain in rodents.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. PCD of neurons in the early postnatal brain (<P30)
  5. PCD in the adult brain
  6. Factors regulating PCD in the adult brain
  7. Significance of PCD in the adult brain
  8. Conclusion and perspectives
  9. Acknowledgments
  10. References

During embryonic development, many neural cells undergo programmed cell death (PCD). This active elimination process is believed to be required for proper establishment and maintenance of the nervous system (Buss et al. 2006). PCD of neural cells is found during most stages of development including proliferation, migration, axonal guidance, and synaptogenesis. It has been proposed that PCD has at least three major functions during embryonic development (Buss et al. 2006): (i) limiting the proliferating pool size during neurogenesis; (ii) correcting errors and removing developmentally transient structures; and (iii) quantitatively matching neurons with efferent targets and afferent inputs for optimal neuronal connections during synaptogenesis. The time-course of neuronal development is highly diverse depending on the neuronal population, and accordingly the time-course of PCD in each neural population is different. Because major developmental processes continue postnatally in some brain regions, it is not surprising that PCD occurs in these areas of the postnatal brain (Buss et al. 2006; Oppenheim 1991). For example, neurogenesis of granule-type neurons in the cerebellum, dentate gyrus (DG), and olfactory bulb (OB) is active until 2–3 weeks after birth, and PCD of neural progenitors occurs in these brain regions. Synaptogenesis is also active in many early postnatal brain regions including the cerebral cortex, midbrain, cerebellum, and retina. Therefore, PCD of post-mitotic neurons during synaptogenesis is also abundant in these brain regions.

It has recently become clear that at least some adult brain regions maintain neural stem cells that spontaneously produce new neurons in adulthood. Adult neurogenesis is evident in most species including rodents (Altman 1963; Privat & Leblond 1972), primates (Gould et al. 1998), and humans (Eriksson et al. 1998; Bernier et al. 2000; Curtis et al. 2007). Similar to the situation during embryonic and early postnatal development, substantial numbers of newly generated neurons in adult brain are eliminated by PCD (Biebl et al. 2000; Petreanu & Alvarez-Buylla 2002; Dayer et al. 2003; Sun et al. 2004; Ninkovic et al. 2007; Snyder et al. 2009). Because neurons born in the adult brain must integrate and synaptically connect with other mature neurons for their survival, adult neurogenesis and PCD continuously change neural circuits.

PCD of neurons in the early postnatal brain (<P30)

  1. Top of page
  2. Abstract
  3. Introduction
  4. PCD of neurons in the early postnatal brain (<P30)
  5. PCD in the adult brain
  6. Factors regulating PCD in the adult brain
  7. Significance of PCD in the adult brain
  8. Conclusion and perspectives
  9. Acknowledgments
  10. References

The regulation and utility of PCD in the early postnatal brain are similar to the processes that occur during embryonic development. Therefore, examples of all three major functions of PCD identified during embryonic development are also found during postnatal brain development. The distribution, cell types, and extent of PCD in the early postnatal brain are summarized in Table 1.

Table 1.   Programmed cell death (PCD) during early postnatal development
Brain regionsStem cells/progenitorsTransient cellsMigrating cellsSynaptogenesis
  1. 1(Thomaidou et al. 1997); 2(Derer & Derer 1990); 3(Gohlke et al. 2004); 4(Ferrer et al. 1990); 5(Janowsky and Finlay 1983); 6(Gould et al. 1991); 7(Kim et al. 2009); 8(Ashwell 1990); 9(Lossi et al. 2004); 10(Jung et al. 2008); 11(Yamanaka et al. 2004); 12(Janec and Burke 1993); 13(Oo and Burke 1997); 14(Fiske & Brunjes 2001); 15(Brunjes et al., 1996); 16(Saito et al. 2004); 17(Young 1984); 18(Galli-Resta and Ensini 1996). AC, amacrine cells; BC, basket cells; BP, bipolar cells; CR, Cajal-Retzius cells; DA, dopamine neurons; DG, dentate gyrus; E, embryonic day; GC, granule cells; GMC, glomerular cells; hs, hamster; ms, mouse; P, postnatal day; PC, purkinje cells; RGC, retinal ganglion cells; RMS, rostral migratory stream; SC, stellate cells; SEL, sub-ependymal layer.

CortexE14-P0 (rat)1CR: P0-7 (ms)2 P14-15 (ms)3
Hippocampus[DG] no PCD at P0-10 (rat, hs)4–5  [CA1-3] P0-7 (rat)4 P5-10 (hs)5 [DG] P21 (rat)6 P30 (ms)7
CerebellumGC: E15-P7 (ms)8  PC: E15-P3 (ms)9,10 SC/BC:P5-21 (ms)11
Substantia nigra (pars compacta)E20-4 (rat)12  DA: P14-20 (rat)13
RMS (SEL)  P1-60 (rat)14–15 
Olfactory bulbGC:E15-P10 (rat)14  GC:P1-20 (rat)14,16 GMC:P1-15 (rat)14
RetinaBP:P3-8 (ms)17  BP/AC:P3-48 (ms)17 RGC:P1-5 (ms, rat)17–18

Elimination of excess stem/progenitor populations

Programmed cell death of neural progenitors was thought to be rare or absent in rodents (Ferrer et al. 1992; Reznikov & van der Kooy 1995; Spreafico et al. 1995). However, with improved techniques, considerable numbers of dying cells are detectable in the proliferative zone during embryonic neurogenesis (Blaschke et al. 1996). Computational simulation estimates that PCD of progenitors may reduce the total number of adult neurons by up to 24% in rodents, which would influence the brain size to a greater extent than PCD during synaptogenesis (Gohlke et al. 2004). PCD of progenitor cells is found in the secondary germinal layer of the postnatal cerebellum and subventricular zone (SVZ) (Hatten & Heintz 1995; Krueger et al. 1995; Thomaidou et al. 1997; Levison et al. 2000; Vogel 2002). These types of PCD the in early postnatal brain also function to eliminate unnecessary stem/progenitor cell populations, which limits the population of proliferating cells.

Elimination of erroneous or developmentally transient cell populations for error correction

Recently, we reported that a subset of erroneously migrating Purkinje cells are eliminated from their ectopic position in the postnatal mouse cerebellum (Jung et al. 2008). Similar elimination of neurons that fail to migrate properly also occurs in the developing cerebral cortex (Sun W. unpubl.data, 2007), suggesting that the elimination of cells that fail to migrate may be a common function of PCD for error-correction. PCD in the postnatal brain appear to also serve as a means to remove transient cell populations at the completion of development. During development of the cerebral cortex, Cajal-Retzius (CR) cells are localized in the pia of the cerebral cortex, and they release reelin, a key regulator of neuronal migration. Elimination of CR cells or null mutation of the reelin gene impairs neuronal migration, resulting in disorganization of cortical cell layer formation (Katsuyama & Terashima 2009). By the second week after birth, when major layer formation in the cerebral cortex is complete, most CR cells in mice are eliminated by apoptosis (Derer & Derer 1990). Similarly, a subset of subplate cells that transiently receive thalamic inputs also undergo PCD postnatally (Woo et al. 1991; Soriano et al. 1993).

Elimination of neurons during synaptogenesis

As previously mentioned, the most prominent PCD occurs during synaptogenesis of postmitotic neurons. Because synaptogenesis continues during postnatal stages in many brain regions, many post-mitotic neurons undergo PCD in the postnatal brain (Table 1). However, because precursors and post-mitotic neurons are often intermingled within small anatomical regions, it is often difficult to unambiguously distinguish the cell types undergoing PCD in the developing brain. Recently, we and others have found that Bax-KO mice exhibited virtually normal numbers of postnatally generated granule neuronal populations in the DG and cerebellum (Fan et al. 2001; Jung et al. 2008; Kim et al. 2009). Because Bax-KO mice cannot execute PCD of post-mitotic neurons during development, these results suggest that PCD of these neuronal populations is less than expected. However, it is known that PCD can be mediated by Bax-independent pathway such as autophagy (Yu et al. 2003, 2008), and Bax-independent PCD might remove these populations. Therefore, more comprehensive analyses are needed to clarify this issue.

PCD in the adult brain

  1. Top of page
  2. Abstract
  3. Introduction
  4. PCD of neurons in the early postnatal brain (<P30)
  5. PCD in the adult brain
  6. Factors regulating PCD in the adult brain
  7. Significance of PCD in the adult brain
  8. Conclusion and perspectives
  9. Acknowledgments
  10. References

As we described above, two main populations of cells undergo PCD: neural progenitor/stem cells and postmitotic neurons. Although it is clear that the number of adult neural stem cells is reduced with age (Olariu et al. 2007), it remains unclear whether this spontaneous reduction in neural stem cell number is related to PCD of precursors or differentiation of adult neural stem cells. On the other hand, PCD of post-mitotic neurons in adult brain has been extensively addressed, and here we focus on this type of PCD in adult neurogenic regions.

Dentate gyrus

The production and subsequent PCD of DG neurons continues throughout life (Biebl et al. 2000; Dayer et al. 2003). In the subgranular zone (SGZ) where progenitor cells are found, approximately 9000 new cells are produced per day in 2-month-old rats (Cameron & McKay 2001), and the production of new neurons decreases with age (Seki & Arai 1995; Biebl et al. 2000; Rao & Shetty 2004; Olariu et al. 2007; Snyder et al. 2009). Studies using double-labeling with apoptotic markers and immature neuronal markers have demonstrated that PCD occurs during the maturation of newly generated neurons (Nacher et al. 2001; Brandt et al. 2003; Francis et al. 2004). Consistent with this idea, when PCD is experimentally blocked, immature neurons accumulate significantly in the DG (Biebl et al. 2000; Sun et al. 2004). Approximately half of all newly generated cells are eliminated within the first month after birth. In rats, new neurons that survive beyond this first month live for at least 6–11 additional months (Kuhn et al. 1997; Biebl et al. 2000; Dayer et al. 2003). Due to lower production and greater PCD, fewer new neurons are added to the adult mouse hippocampus, and the overall contribution of adult neurogenesis to hippocampal function is weaker in mice compared with rats (Kempermann et al. 2003; Snyder et al. 2009).

Although the majority of neurons undergoing PCD in the adult hippocampus are adult-generated immature neurons, a subset of mature DG neurons also undergoes spontaneous PCD. Long-term birthdate tracing with BrdU labeling in the rat has revealed that about 25% of DG neurons born during a major developmental neurogenic period (P6) undergo PCD during the first 1–6 months after birth (Dayer et al. 2003). It is unclear whether there is spontaneous PCD of mature DG neurons in mice, but the total number of DG neurons does not change in the adult mouse between 3 and 12 months of age, and there may be compensatory loss of mature neurons (Sun et al. 2004).

The SVZ-RMS-OB system

Compared with the DG, the rate of neurogenesis is higher in the adult SVZ-OB system. It has been estimated that daily 60 000–120 000 cells in 2-month-old rats and 30 000 cells in mice are generated daily (Lois & Alvarez-Buylla 1994; Biebl et al. 2000; Winner et al. 2002). Newly generated neuroblasts that originated from the SVZ migrate through the rostral migratory stream (RMS) into the OB. After reaching the regions of the subependymal layer (SEL), which is the central region of the OB, the neuroblasts migrate radially into the olfactory granule cell layer (OGL), and undergo terminal differentiation into olfactory granule cells (OGCs) and periglomerular cells. PCD occurs prominently in the OB compared with the RMS and the SVZ during neuronal maturation in rats and mice (Biebl et al. 2000; Belvindrah et al. 2002; Petreanu & Alvarez-Buylla 2002). Cells that survive beyond the PCD period (<10% of total GCs) live at least 19–21 months in rats and mice (Kaplan et al. 1985; Winner et al. 2002; Yamaguchi & Mori 2005). Recently, Petreanu & Alvarez-Buylla (2002) demonstrated that there are two waves of cell death in the adult OB. Initially, 50% of newly generated cells are eliminated within 1 month (during synaptogenesis), and then a further 25% of cells undergo PCD gradually between 6 and 19 months after birth. Therefore, during the second wave, surviving and functioning post-mitotic (over 6 months-old) neurons may be gradually removed and replaced by newly born neurons. In this way, newly generated GCs may spontaneously replace earlier born, older cells.

Other brain regions

Although the DG and SVZ are two major adult neurogenic regions, some studies have indicated the production of new neurons in other regions of the adult rodent brain (Gould 2007). For example, it has been reported that a subset of midbrain dopaminergic neurons are spontaneously eliminated, and new neurons replace these degenerated neurons in the rat (Zhao et al. 2003). However, this remains controversial (Frielingsdorf et al. 2004; Shan et al. 2006), and further confirmation is required. In fact, neural stem cells are found in some non-neurogenic regions such as the paraventricular region (PVR), spinal cord, and retina (Takemura 2005; Reh & Fischer 2006; Seri et al. 2006). Although the ability of stem/progenitor cells in non-neurogenic regions to spontaneously produce neurons appears to be limited in vivo, they can produce neuronal cells in vitro or following brain injury, suggesting that they may contribute to brain repair after injury. The extent and regulation of PCD in these non-neurogenic areas needs to be addressed in future studies.

Factors regulating PCD in the adult brain

  1. Top of page
  2. Abstract
  3. Introduction
  4. PCD of neurons in the early postnatal brain (<P30)
  5. PCD in the adult brain
  6. Factors regulating PCD in the adult brain
  7. Significance of PCD in the adult brain
  8. Conclusion and perspectives
  9. Acknowledgments
  10. References

Because PCD in the adult brain is closely related to adult neurogenesis, the factors controlling adult neurogenesis also play an essential role in PCD of adult produced neurons (Fig. 1).

image

Figure 1.  Various factors regulating proliferation and/or long-term survival (>3 weeks) in adult neurogenic regions. (A) Sagittal view of the rodent brain showing the neurogenic regions (red and blue shaded), (B) In the dentate gyrus (DG), neuroblasts located in the subgranular zone (SGZ) and differentiated neurons migrate into the granule cell layer (GCL). Matured neurons extend their dendrites to the molecular layer (ML) and project axons toward CA3. (C) In the subventricular zone (SVZ), stem cells (B-cell) are located adjacent to the ependymal layer (E). Through transit amplifying cells (C-cell), neuroblasts are committed (A-cells). (D) In the olfactory bulb (OB), newly migrated neuroblasts differentiate into granule cells and periglomerular cells. (E) List of factors that increase (columns below upward arrow) or decrease (columns below downward arrow) proliferation and/or survival of DG neurons (black text), SVZ-OB (red), or both (blue). AMPA, α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate; BDNF, brain-derived neurotrophic factor; DHEA, dehydroepiandrosterone; GABA, gamma-Aminobutyric acid; LTD, long-term depression; LTP, long-term potentiation; NE, norepinephrine; NGF, nerve growth factor; NMDA, N-Methyl-D-aspartic acid; PACAP, pituitary adenylate cyclase-activating peptide; TGF, transforming growth factor; TNF, tumor necrosis factor; VEGF, vascular endothelial growth factor.

Download figure to PowerPoint

Environment/experience

It is well known that environmental stimuli influence the extent of neurogenesis. Different stimuli appear to differentially affect the proliferation of neural stem cells vs. the PCD of newly produced neurons. For example, exposure of rodents to an enriched environment increases the survival of newly generated hippocampal neurons, whereas the effect of enrichment on neural stem cell proliferation is strain-specific (Kempermann et al. 1998, 2006). Physical exercise, such as voluntary wheel running, promotes both stem cell proliferation and to a lesser extent the survival of newly born hippocampal neurons in the mouse DG (van Praag et al. 1999; Snyder et al. 2009). Similar to the effect of enrichment on DG neurogenesis, exposure of mice to an enriched odorant environment at the critical period promotes the survival of OB neurons in mice (Rochefort et al. 2002; Rochefort & Lledo 2005; Yamaguchi & Mori 2005). Conversely, olfactory sensory deprivation promotes PCD of olfactory neurons, which may be mediated by gamma-Aminobutyric acid (GABA)-A receptor receptor activation (Yamaguchi & Mori 2005). Repetitive learning, such as eyeblink conditioning, a water maze task, or odor discrimination learning also promotes survival of newly generated neurons (Gould et al. 1998; Leuner et al. 2004; Alonso et al. 2006). Experiencing social dominance enhances neuronal survival in rats (Kozorovitskiy & Gould 2004), whereas social isolation or subordinate social status significantly reduced cell proliferation and hippocampal learning (Gould et al. 1998; Ibi et al. 2008). Thus, social interaction, social status and stress are important factors regulating neurogenesis and neuronal survival.

It appears that prolonged exposure to odors is required for modification of the extent of PCD, and environmental stimuli that are sufficient to induce neuronal plasticity may be required for an effect on PCD of adult-generated neurons. Consistently, the induction of long-term potentiating as a putative cellular mechanism of learning and memory promotes neuronal proliferation and survival in the rat DG (Bruel-Jungerman et al. 2006; Chun et al. 2006). Altered neurogenesis following environmental stimulation contributes to alterations in specific aspects of hippocampal and olfactory function (Saxe et al. 2006; Imayoshi et al. 2008; Moreno et al. 2009; Mouret et al. 2009).

Secretory factors

The behavioral/ethological stimuli-dependent alterations in PCD are likely mediated by extracellular secretory factors such as neurotransmitters, hormones, and growth factors. Neurotransmitters that control neuronal excitation are involved in the control of adult neurogenesis and PCD. Glutamate influences proliferation and survival by acting through N-Methyl-D-aspartic acid (NMDA) and metabotropic receptors. In the DG, NMDA activation promotes neuronal survival (Gould et al. 1997; Yoshimizu & Chaki 2004; Snyder et al. 2009). Conversely, NMDA blockade decreases neuronal survival in the rat OB (Fiske & Brunjes 2001). GABA is a major inhibitory neurotransmitter for mature neurons, but it depolarizes progenitor cells and immature neurons in the adult brain (Imayoshi et al. 2008). GABA released from hippocampal interneurons activates progenitors extrasynaptically, which promotes the maturation of immature neurons. However, the role of tonic GABA in the PCD of immature neurons is currently unclear. Serotonin (or 5-hydroxytryptamine, 5-HT), acting through different subtypes of receptors, upregulates cell proliferation (Brezun & Daszuta 1999; Malberg et al. 2000; Banasr et al. 2004). The adult SVZ and DG also receive dopaminergic input from the substantia nigra and ventral tegmental area (Swanson 1982). Dopaminergic denervation in animal models of Parkinson’s disease causes a dramatic reduction in the number of proliferating cells in mouse SVZ and DG (Baker et al. 2004; Hoglinger et al. 2004).

Various types of stress lead to decreased cell proliferation and/or survival during adult neurogenesis (Gould et al. 1998; Pham et al. 2003; Mirescu et al. 2006). Stress hormones secreted from the hypothalamic-pituitary-adrenal axis regulate adult neurogenesis (McEwen 1999). Thyroid hormones are also important factors that regulate cell proliferation and maturation during brain development, and they also affect adult neurogenesis in rodents (Ambrogini et al. 2005; Desouza et al. 2005). Sex hormones and prolactin also modulate DG neurogenesis, and these factors may be important for reproductive behaviors (Perez-Martin et al. 2003; Shingo et al. 2003; Dalla et al. 2009).

Growth/neurotrophic factors are important regulators of neuronal PCD in the adult brain. Stimulation of brain-derived neurotrophic factor (BDNF) signaling promotes the proliferation and survival of neurons in the DG and/or SVZ in rodents (Zigova et al. 1998; Katoh-Semba et al. 2002; Sairanen et al. 2005; Scharfman et al. 2005; Rossi et al. 2006). Other factors known to influence progenitor cell proliferation include ciliary neurotrophic factor (Emsley & Hagg 2003), basic fibroblast growth factor 2, epidermal growth factor (Kuhn et al. 1997; Enwere et al. 2004), vascular endothelial growth factor (Cao et al. 2004; Kim et al. 2009), and insulin-like growth factor-1 (Aberg et al. 2000; Lichtenwalner et al. 2006). However, their effects on PCD are much less clear.

Molecular pathways for PCD in the postnatal brain

Most neurons undergoing PCD exhibit morphological and biochemical hallmarks of apoptosis, such as nuclear condensation/fragmentation, cell shrinkage, activation of caspases, and DNA fragmentation. Therefore, neuronal PCD is mediated by key players of the apoptosis signal cascade such as pro- and anti-apoptotic members of the Bcl-2 family and pro-apoptotic caspases. Although it is generally believed that Bcl-2 family members are upstream controllers of caspase activation in the ‘mitochondrial pathway’, the consequences of perturbing Bcl-2 family member expression vs. caspase activity are strikingly different. When cell death activators such as caspase-3, -9, or Apaf1 are deleted, PCD of neural progenitors is markedly reduced, which leads to gross brain abnormalities and increased embryonic lethality in mice (Kuida et al. 1998; Yoshida et al. 1998; Honarpour et al. 2000; D’Sa et al. 2003; Song et al. 2005). However, suppression of apoptosome formation or caspase activation does not affect PCD of postmitotic motoneurons (Oppenheim et al. 2008). On the other hand, gene knockout of the pro-apoptotic gene Bax results in accumulation of post-mitotic neurons rescued from PCD (Sun et al. 2003, 2004; Kim et al. 2007), whereas the PCD of neural progenitors appears to be unaffected. These results suggest that these two molecules selectively control different types of PCD: Bcl-2 family members may play a more prominent role in PCD during synaptogenesis, whereas caspases are more important for PCD of neural stem/progenitor cells.

It has been reported that transient blockade of caspase activity by an infusion of caspase inhibitors only transiently protects newly produced neurons from PCD, and there is no significant increase in the number of NeuN-positive mature neurons in rats (Ekdahl et al. 2001). On the other hand, gene knockout of Bax or overexpression of Bcl-2 completely prevents PCD of newly produced DG neurons (Sun et al. 2004; Kuhn et al. 2005). Likewise, PCD of OB neurons appears to be critically modulated by Bax as shown by deletion of this gene (Kim et al. 2007). However, PCD of SVZ neuroblasts also requires Bak, another multidomain pro-apoptotic molecule. Compared with Bax or Bak single knockout mice, Bax/Bak double mutant mice exhibited more severe accumulation of SVZ-derived neuroblasts in the RMS, suggesting that both Bax and Bak play roles in PCD of SVZ neuroblasts (Lindsten et al. 2003, 2000; Kim et al. 2007).

Significance of PCD in the adult brain

  1. Top of page
  2. Abstract
  3. Introduction
  4. PCD of neurons in the early postnatal brain (<P30)
  5. PCD in the adult brain
  6. Factors regulating PCD in the adult brain
  7. Significance of PCD in the adult brain
  8. Conclusion and perspectives
  9. Acknowledgments
  10. References

Elimination of excess cells through PCD is essential for establishing efficient synaptic connections. The neurotrophic hypothesis explains how selection of neuronal survival versus PCD is determined during synaptogenesis (Oppenheim 1991; Buss et al. 2006). Briefly, the amount of target-derived neurotrophic signals required for the survival of neurons is limited, and competition among neurons for access to these trophic signals eventually discriminate the winners (survival) and losers (death). The mechanism of PCD in adult neurogenesis also appears explained by the neurotrophic hypothesis. As discussed earlier, conditions that increase the expression of neurotrophic signals (e.g. learning, physical exercise, brain damage) enhance the survival of adult-produced neurons. By contrast, conditions that suppress the expression of neurotrophic signals (e.g. stress) reduces the survival of new neurons. A single-cell knockout technique has provided direct evidence for the importance of competition among newly-generated neurons for their survival. When the NMDA receptor is genetically ablated in a subset of newly produced neurons, the survival of the affected neurons is significantly reduced. Conversely, the survival of normal neighboring neurons is increased compared with newly produced neurons in control mice. These data suggest that the survival of neurons is regulated by competition among newly produced neurons (Tashiro et al. 2006).

Although competition occurs among homogeneously immature neurons during embryonic/postnatal development, newly born immature neurons in adult neurogenic regions must become incorporated into pre-established mature circuits, and these immature neurons need to compete/interact with mature neurons for new synaptic connections. Using morphological tracing by infecting dividing cells with green fluorescent protein (GFP)-expressing retroviruses, it has been shown that newly born neurons can form synapses with targets (Hastings & Gould 1999; Markakis & Gage 1999; Petreanu & Alvarez-Buylla 2002; Zhao et al. 2006), and they begin to receive glutamatergic input from afferent fibers by 2 weeks after birth (Wang et al. 2000; van Praag et al. 2002; Belluzzi et al. 2003; Carleton et al. 2003). Because the numbers of efferent (e.g. CA3) and afferent (e.g. entorhinal cortex) neurons remain relatively constant in the adult brain, newly generated young DG neurons should be able to find pre-existing synaptic partners. Dendritic filopodia of the newly generated neurons form synaptic contacts with previously formed postsynaptic boutons, resulting in altered connectivity (Toni et al. 2007). Furthermore, by 4–6 weeks after birth, these young immature neurons exhibit stronger synaptic plasticity than mature neurons (Wang et al. 2000; Snyder et al. 2001). This enhanced synaptic plasticity of immature neurons may be due to the lack of GABAergic inhibition, whereas mature DG neurons are strongly inhibited by GABA (Snyder et al. 2001; Esposito et al. 2005; Tozuka et al. 2005; Tashiro et al. 2006). Therefore, newly produced neurons appear to compete with mature neurons for their synaptic connections.

Recently, we addressed the importance of PCD in the adult brain using Bax-KO mice, which cannot eliminate surplus neurons by PCD (Sun et al. 2004). In Bax-KO mice, the number of DG neurons progressively increases with age, due to the survival of all newly produced neurons. Those surviving neurons exhibited mossy fiber (MF) projections and synaptic connections with targets, CA3 postsynaptic boutons. Because of the increased numbers of MFs in the presence of a relatively constant number of CA3 dendritic spines, the MF:CA3 ratio is dramatically increased (Kim et al. 2009), severely impairing synaptic plasticity. These observations imply that newly produced neurons may interact with the pre-existing circuitry, and in the absence of their elimination, these surplus neurons continue to invade and perturb the mature circuits. Therefore, aged Bax-KO mice exhibit a reduction of calbindin, mature DG neuronal marker, expression, and an impairment of associative memory formation (Sun et al. 2004; Perez et al. 2007; Kim et al. 2009). This is in strong contrast to the developmental compensation of motor and sensory behaviors of Bax-KO mice. Although developmental PCD of motor and sensory neurons is also absent in Bax-KO mice, Bax-KO mice show virtually normal motor/sensory behaviors (Sun et al. 2003; Buss et al. 2006). We have determined that the extra motoneurons that survive developmental PCD due to Bax deletion fail to maintain target innervations, are progressively excluded from the circuit, and show atrophied morphology (Sun et al. 2003). Therefore, the numbers of functionally significant (i.e. maintained in the motor circuits) motoneurons are similar in wild-type and Bax-KO mice. Considering that neurotrophic factors promote the growth and axonal elongation of neurons, failure to access sufficient amounts of neurotrophic signals may limit the growth of neurons, resulting in atrophy and exclusion of the excess neurons from functional neural circuits in Bax-KO mice. On the other hand, synaptogenesis and PCD of adult-produced DG neurons occur after initial contacts to the CA3 region in the adult brain, and there is little need for further growth/axonal elongation during PCD of adult-produced neurons. Therefore, failure of PCD in the adult brain may not be efficiently corrected by the subsequent failure of growth and maintenance of synaptic contacts. In this respect, PCD of adult-produced neurons may play an essential and indispensible role in systems-matching of the adult brain.

In contrast to hippocampal function, we failed to observe overt deficits in olfactory functions in Bax-KO mice (Kim et al. 2007). This is mainly due to the compensatory failure of neuronal migration of surplus neuroblasts to the OB. Compensation for the failure of PCD may occur at multiple steps of development, such as before (migration) and after (growth, synaptic maintenance) synaptogenesis. This observation further indicates that PCD is a developmental adaptation for efficient elimination of excess cells, whereas in most cases, other back-up strategies can compensate for the possible failure of PCD.

Conclusion and perspectives

  1. Top of page
  2. Abstract
  3. Introduction
  4. PCD of neurons in the early postnatal brain (<P30)
  5. PCD in the adult brain
  6. Factors regulating PCD in the adult brain
  7. Significance of PCD in the adult brain
  8. Conclusion and perspectives
  9. Acknowledgments
  10. References

The elimination of excess neuronal connections through programmed cell death (PCD) is essential for the establishment and maintenance of efficient synaptic function. Although the distribution and extent of PCD in the adult brain was identified more than 20 years ago, this obvious biological event has been largely ignored until recently, and understanding the significance of PCD in the adult brain is still in its infancy. We now know that the extent of PCD in the adult brain is modified by neuronal activity/experience, and perturbation of PCD significantly impairs normal brain functions. Therefore, PCD in the adult brain is not a mere elimination step removing excess cells, but is an important regulatory step for maintenance of optimal brain function. To underscore the physiological importance of PCD and neuronal replacement in the adult brain, future studies of adult neurogenesis and PCD should use animal models including wild species that are examined in natural or semi-natural conditions, with an emphasis on ethologically relevant behaviors. In addition, modification of adult neurogenesis and PCD appear to be related to several neurodegenerative and psychiatric diseases (Hsieh & Eisch 2000; Santarelli et al. 2003; Kim et al. 2009; Taupin 2009). Therefore, understanding the physiology of PCD in the adult brain may lead to new therapeutic strategies.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. PCD of neurons in the early postnatal brain (<P30)
  5. PCD in the adult brain
  6. Factors regulating PCD in the adult brain
  7. Significance of PCD in the adult brain
  8. Conclusion and perspectives
  9. Acknowledgments
  10. References