Age‐related decline in neurogenesis: Old cells or old environment?

As the baby boom generation ages, the numbers of patients afflicted with various age-related pathologies of the brain will increase, creating a tremendous burden on the health care segment of the economy as well as great human suffering. Observations that neurogenesis continues in the adult brain (Kaplan and Hinds, 1977; Nottebohm, 1985; Eriksson et al., 1998) and may be part of a normal repair process has led to ideas for therapeutic interventions, such as: transplantation of healthy neural stem/precursor cells and induction of neurogenesis in situ (reviewed by Magavi and Macklis, 2001). Unfortunately, age-related changes in the brain may limit therapeutic success. Transplantation of young neural progenitors or stem cells may not succeed in repairing damaged brains if the environment of the recipient brain is hostile to survival and differentiation due to aging (Sykova, 2001). Induction of neurogenesis in situ may be thwarted by the inability of the resident neural stem/progenitor cells to respond due to aging (Armstrong and Barker, 2001). If we are to efficiently exploit adult neurogenesis for therapeutic ends, we must understand the effect of aging on the component parts of neurogenesis. Do age-related changes occur in the stem/progenitor cells present in the brain or do age-related changes occur in the local and systemic environments that allow these cells to proliferate and differentiate when needed? Neural stem cells are defined as cells that are capable of producing additional undifferentiated stem cells as well as progeny that go on to differentiate into several cellular phenotypes: neurons, glial cells and oligodendrocytes. Neural progenitor cells are defined as cells with a limited replicative potential that are committed to the neuronal lineage. Neural progenitors are derived from neural stem cells either during fetal development or in the adult. Neurogenesis, the production of new neurons, depends upon neural stem cells and neural progenitors being present and responsive to induction. Neurogenesis can be assessed in vitro (Fig. 1) or in vivo. Mouse cell positive for neuron-specific marker MAP2. MAP2 is a microtubule-associated protein that promotes tubulin polymerization and is involved in neurite outgrowth. In the mature neuron, MAP2 is present in the soma and dendrites of the neuron. Using a transgenic line of mice, C57BL/6-TgN(ACTbEGFP)1Osb (The Jackson Laboratory, Bar Harbor, ME), that constitutively expresses green fluorescent protein (GFP) in every cell of the body, newborn mouse brains were dissociated into single cells. In serum-free medium containing growth factors, these cells were grown as unattached neurospheres for 19 days. At that time the growth factors were removed and the cells were allowed to attach to polyethylenimine coated sterile coverslips. After an additional 38 days in which the cells migrated out from the neurospheres and differentiated into a variety of cell phenotypes, the cells were fixed with 4% paraformaldehyde and permeabilized with Triton X-100. MAP2 was visualized using a mouse monoclonal antibody (clone AP-20; Sigma-Aldrich Inc., St. Louis, MO) followed by counterstaining (red color) with Alexa Fluor 568 signal-amplification kit for mouse antibodies (A-11066; Molecular Probes, Eugene, OR). The nucleus was stained with DAPI (purple). Evidence that cell proliferation occurs in the adult brain is often based upon labeling with 5-bromo-2′-deoxyuridine (BrdU), a thymidine analog. BrdU is incorporated into DNA during replication and repair, and can subsequently be detected by immunohistology using an antibody against BrdU. In response to criticism that BrdU incorporation might be detecting DNA repair in postmitotic neurons instead of cell division in progenitors or stem cells, Cooper-Kuhn and Kuhn (2002) investigated this question in 8-week-old Wistar rats. After labeling with BrdU, brains were fixed at defined time intervals. Immunohistological analysis of the granule cell layer of the dentate gyrus of the hippocampus revealed that BrdU positive cells co-expressed markers for mitosis, progenitors, immature and mature neurons in that sequence with increasing time after BrdU labeling. BrdU positive cells were not found to co-localize with cells undergoing apoptosis, and irradiation, which stimulates DNA repair, caused a decrease in BrdU labeling. The authors concluded that incorporation of BrdU is a reliable method for detecting cells undergoing mitosis, and can be used to detect both neural stem and progenitor cells. BrdU labeling has indicated that cell division takes place in at least two regions of the adult brain: the dentate gyrus of the hippocampus and the subependyma surrounding the anterior lateral ventricle. Using adult Wistar rats (8–10 weeks old) and adult CD1 mice (6–8 weeks old), Seaberg and van der Kooy (2002) carefully dissected out tissue from each region, extracted cells and tested these cells for self-renew in tissue culture. In both species of rodent the cells derived from the dentate gyrus exhibited a limited ability for self-renewal whereas the cells derived from the subependyma exhibited more proliferative potential, suggesting that the dentate gyrus contains only progenitors and the subependyma contains both progenitors and stem cells. In the dentate gyrus of the hippocampus age-related changes have been observed. Using highly polysialylated neural cell adhesion molecule (NCAM-H) as a marker for newly generated granule cells and BrdU as a marker for newly synthesized DNA, Seki and Arai (1995) compared neurogenesis in the dentate gyrus of Wistar rats 35 days to 18 months of age. Immunohistology showed that the numbers of newly generated granule cells decreased with increasing age. Using a double labeling protocol consisting of BrdU and either anti-NeuN (neuron-specific nuclear protein) or anti-calbindin-D28k (an intracellular protein with high affinity for calcium), Kuhn et al. (1996) compared 6-month-old Fischer 344 rats with 12- and 27-month-old rats. Immunohistology revealed a reduction in the differentiating progeny of granule cells with increasing age. That this phenomenon was not limited to rats was shown by Kempermann et al. (1998) who compared 8-month and 20-month-old C57Bl/6 mice and found that BrdU labeling in the dentate gyrus of the hippocampus decreased with increasing age. In the forebrain subependma age-related changes have been observed by Tropepe et al. (1997). When 2–4 month and 23–25 month-old SW/COBS mice were compared, the older mice exhibited a 50–60% decrease of BrdU positive progenitor cells. Sequential labeling with BrdU and tritiated thymidine revealed lengthening of the cell cycle in the older mice. When cells were isolated from the subependyma and proliferative potential was assessed in tissue culture, however, no age-related decrease was observed, suggesting a lack of aging in the neural stem cells. Additional insights were gained from transplantation studies that introduced fetal or embryonic neural stem cells into aged brains. Hodges et al. (2000) injected conditionally immortalized hippocampal cell lines derived from mouse embryos into the ventricles of 22-month-old Sprague-Dawley rats. Qu et al. (2001) injected primary neural stem cells derived from human fetuses into the ventricles of 24-month-old Fischer 344 rats. Both kinds of cells migrated widely throughout the recipient brains, differentiating into neurons and astrocytes, and both procedures resulted in significant improvements in the ability of the aged rats to remember the location of a submerged platform in the water maze tests. The possible mechanism by which memory was enhanced, however, remained ambiguous: did the donor cells augment deficient resident neural stem/progenitor cells or did neurotrophic factors released by the donor cells improve the aged environment? Specifically addressing possible age-related changes in the host environment, Zaman and Shetty (2002) freshly isolated rat fetal hippocampal cells (prelabeled in utero with BrdU), then injected these cells into the ventricles of Fischer 344 rats of three ages: young adult, middle-aged adults (12–14 months) and old adults (11–24 months). Thirty days after transplantation, absolute graft cell survival was determined by immunostaining for BrdU and stereological cell counting methods. When the hippocampi were undamaged approximately 18–25% of donor cells survived at all ages of recipient. When hippocampi were first damaged with kainic acid (KA), however, then given fetal hippocampal cells, an age-related change in absolute graft cell survival was revealed: an average of 72% survived in young brains vs. 30% in middle-aged and old brains. The failure to observe a further decline in the brains of old as compared to middle-aged host brains raises the question that the decrements in middle-aged brains could be related to maturation rather than aging. Additional experiments, however, did support continued decrements into old age. Graft cell survival could be improved by incubation of the freshly isolated fetal hippocampal donor cells for 3 hr in the presence of brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3) and caspase inhibitor before transplantation to KA damaged brains. Survival increased in middle-aged rats to 51% and in old rats to 63%. The authors speculated that damage to brain tissue stimulates the release of neurotrophic factors. As an animal ages this response to damage becomes impaired, so that transplanted cells experience a less supportive environment in the middle-aged and old brains. This age-related deficiency could be overcome to an extent by artificially supplying the neurotrophic factors. Another damaging treatment that reveals age-related changes in the brain is transient forebrain ischemia, a model for stroke (Yagita et al., 2001). Wistar rats were first subjected to ischemia, then injected with BrdU to label proliferating cells. When hippocampal tissue was fixed 1 day after BrdU labeling, ischemic animals exhibited an increase in labeled cells in the subgranular zone of the hippocampus when compared to control animals: a 5.7-fold increase for young adults (3–4 months old) and a 10.6-fold increase for old adults (24 months old). When hippocampal tissue was fixed 28 days after BrdU injection, the number of nuclei still labeling with BrdU could be compared to the number at Day 1, assessing the survival of the labeled cells during that period of time. In young adult tissue 65.5% of the labeled cells survived, whereas in old adult tissue only 15.3% remained. Double staining for BrdU and NeuN showed that 71.6% of the surviving cells had differentiated into neurons. The authors suggested that although both young and old tissues responded to ischemia with increased incorporation of BrdU, the younger brains retained a higher number of such cells, reflecting a higher survival rate of such cells and suggesting that the young brains provided a more supportive environment for survival. These authors postulated that an age-related increase in glucocorticoids might be inhibiting survival of the transplanted cells. Global approaches to improving memory or stimulating neurogenesis in old animals have provided a number of interesting results. Nichols et al. (2000) review the evidence that increases in glucocorticoids with increasing age may cause synaptic loss, and inhibit production of newly born granule neurons. Lichtenwalner et al. (2001) and Aberg et al. (2000) suggest that decreases in levels of IGF-I in the aged brain may be responsible for decreased neurogenesis. Mayo et al. (2001) report that infusions of the neurosteroid, pregnenolone sulfate (PREG-S), will reverse memory impairments in aged rats. Nicolle et al. (2001) and Manev et al. (2000) discuss the possible role of oxidative stress in increasing neuronal vulnerability in aged hippocampal neurons. Lee et al. (2000), Mattson (2000), and Duan et al. (2001) describe the effects of dietary restriction (the only intervention known to increase lifespan) on neural progenitor/stem cells: an increase in newly generated neural cells in the adult brain, an increase in expression of neurotrophic factors such as BDNF, and an increase in the resistance of neurons to dysfunction and apoptosis. None of these studies directly address our original question, however, because both neural progenitor/stem cells and support cells may be changed by any of these experimental manipulations. Nearly half the total cells in the brain are astrocytes, star-shaped glial cells that support the proliferation, survival and maturation of neurons. Studying differentiation in vitro, Song et al. (2002) have shown that co-culture of astrocytes derived from the adult hippocampus with adult neural stem cells promotes neuronal fate commitment, whereas co-culture of astrocytes derived from the adult spinal cord does not promote neurogenesis from the same stem cells. These data suggest that astrocytes play an active role in controlling neurogenesis, and that astrocytes in different areas of the central nervous system have distinct characteristics in this regard. With increasing age the ratio of glial cells to neurons increases, resulting in gliosis (reviewed in Cotrina and Nedergaard, 2001). With increasing age, might astrocytes also lose the ability to stimulate neurogenesis? In summary, the results of many experiments are consistent with the hypothesis that old brains are less able to support proliferation and differentiation of neural progenitor cells, but no data have been gathered as to whether intrinsic age changes occur in the neural progenitors themselves. The picture is even less clear concerning neural stem cells. What is the relationship between age-changes in neurogenesis and dementia of the Alzheimer type (DAT)? Do the disease processes involved in neurodegeneration cause defects in neurogenesis or do defects in neural stem/progenitor cells contribute to DAT? Mutations in the Presenilin-1 (PS1; PSEN-1) gene is a significant cause of early-onset familial Alzheimer disease (Selkoe, 1998). Presenilin protein is thought to be part of a complex of proteins involved in the cleavage of amyloid precursor protein-C terminal fragments (APP-CTF) to produce the beta-amyloid peptides that accumulate during the pathogenesis of DAT (reviewed in Esler and Wolfe, 2001). Knock-out mice lacking a functioning presenilin-1 gene exhibit greatly reduced numbers of neural progenitor cells at 9.5 days of gestation, and later in gestation the brain exhibits hemorrhages and cerebral cavitation (Shen et al., 1997). These mice also have gross skeletal malformations and die within minutes of being born. Handler et al. (2000) subsequently determined that the reduced number of neural progenitor cells was caused by premature differentiation. Normally during early neurogenesis, most progenitor cells would continue to cycle producing an expansion of the progenitor population. When PS1 function is missing the premature differentiation of neural progenitor cells severely depletes this pool of progenitors, which are not available when needed later in development. These data suggest that PS1 may be necessary for the maintenance of undifferentiated neural progenitor cells during development. A clever technique that allows the study of PS1 knockout mice while avoiding the perinatal lethality is the production of conditional knockout (cKO) mice. Tsien et al. (1996) inserted loxP sequences into the genome of embryonic stem (ES) cells so that they flanked part of the PS1 gene, but did not interfere with the normal expression of the gene. The ES cells were used to generate homozygous mice, which were then crossed with a second mouse line transgenic for Cre recombinase under the control of the α-CaMKII (Ca2+/calmodulin-dependent protein kinase II) promoter. The activity of this promoter is restricted to the forebrain and starts about 4 weeks after birth. Therefore, in the hybrid animals the Cre recombinase is expressed after birth, catalyzing recombination between the loxP recognition sequences, and deleting part of the PS1 gene only in some cells of the forebrain. This efficient deletion occurs in postmitotic neurons. Comparing control and cKO mice, Feng et al. (2001) found that basal levels of neurogenesis as assessed by BrdU incorporation were not significantly different; however, when the mice were exposed to an enriched environment known to stimulate neurogenesis (Kempermann et al., 1997) a difference was revealed. The cKO mice increased neurogenesis 37% less than the control mice. To test the correlation between memory and neurogenesis, Feng et al. (2001) subjected both control and cKO mice (10–17 months old) to various behavioral experiments. A significant difference between the two groups was found only with the cued fear conditioning test. In this test, mice are given a one-time shock paired with a tone, then 15 days later the mice are exposed to the tone without shock. The cKO mice exhibited more freezing responses than the controls, leading the authors to suggest that neurogenesis in the hippocampus is correlated with the stability of memory in the hippocampus. Newly generated neurons in the dentate gyrus make new synapses, but the cells live only 3 weeks (Cameron et al., 1993; Hastings and Gould, 1999). Three weeks is about the time that memory is stored in the hippocampus before being transferred elsewhere in the brain. Production of new neurons during this time period actually degrades the resident memory traces in the hippocampus. In the cKO mice, the absence of neurogenesis allows the hippocampal memory of the conditioned fear response to be retained, whereas in the control mice normal neurogenesis allows new memories to take the place of old. Additional evidence that PS1 is important in adult neural progenitor cells comes from Wen et al. (2002). Two-month-old C57B16/DBA F1 hybrid mice were prelabeled for 4 hr with BrdU and 24 hr later the brains were fixed and prepared for histology. Co-staining for BrdU and a marker of early neuronal differentiation was observed in only 11–14% of BrdU labeled cells. Co-staining for BrdU and a marker of mature neurons or astrocytes was not observed. But co-staining for BrdU and nestin (a marker for immature neural progenitors) was observed in 70% of BrdU labeled cells. These data indicate that not enough time had passed at fixation for the proliferating cells of the hippocampus to differentiate into neurons or glia. Double immunofluorescent staining for BrdU and PS1 revealed an average of 64% of all BrdU labeled cells were also PS1 positive. These data suggest that PS1 expression is common in undifferentiated neural progenitor cells of the adult brain. In summary dominant mutations in the gene for PS1 are associated with early-onset familial DAT. PS1 is expressed in neural progenitor cells and may be necessary for maintenance of the undifferentiated state. Might mutations in PS1 lead to a deficit of or premature differentiation of neural progenitor cells in DAT? The most thoroughly studied of all adult stem cells are hematopoietic stem cells (HSCs). Insights gained from the study of these adult stem cells can be used to guide experiments with neural stem cells. Globerson (2001) and Schlessinger and Van Zant (2001) have reviewed the evidence for age-related changes in HSCs. Extensive transplantation studies from young into old and old into young have established a consensus: basal hematopoiesis is maintained throughout life, but the ability to respond to stress decreases with age and cells derived from aged donors exhibit some alterations in their ability to differentiate. Similar experiments should be conducted with adult neural stem cells; i.e., the transplantation of stem cells from elderly animals into young animals and vice versa with suitable controls. Such experiments would provide data concerning: 1) numbers of stem cells that can be isolated from old tissue vs. young tissue, 2) the ability of the stem cells from elderly animals to proliferate and migrate in the environment provided by the young brain, and 3) the extent to which stem cells from elderly animals retain the ability to differentiate into various cell phenotypes. Variations on this theme would include damaging the brains of the young hosts to elicit neurotrophic factor release and assessing the ability of the stem cells derived from elderly animals to repair such damage. If deficits are identified in the old neural stem cells, pretreatment with growth factors or hormones could be attempted to determine if such deficits can be reversed. Are similar experiments feasible using neural stem cells with mutations for familial DAT? Palmer et al. (2001) reported recently the successful extraction of neural stem cells from postmortem human brains 2–20 hr after death. Embryonic and fetal human neural progenitors have been successfully transplanted into rat brain where the human cells migrate and differentiate (Aleksandrova et al., 2001; Qu et al., 2001; Englund et al., 2002). Adult human neural precursors have been transplanted to adult rat spinal cord where remyelination was elicited (Akiyama et al., 2001). Therefore, it might be possible to extract neural stem cells from postmortem samples of human dentate gyrus, from patients with familial DAT, age-matched controls and young controls. After characterization of proliferative and differentiation potential in vitro, the cells could be characterized further by transplantation into young rat brains with and without damage, to test for proliferative and differentiation potential in vivo. If deficits are identified in neural stem cells from patients with DAT, experimental analysis to identify the metabolic bases for such deficits would follow with the eventual goal of generating rational protocols for intervention.