SEARCH

SEARCH BY CITATION

Keywords:

  • aging;
  • BrdU;
  • hippocampus;
  • memory;
  • spatial learning

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. References

Age-related decrements in hippocampal neurogenesis have been suggested as a basis for learning impairment during aging. In the current study, a rodent model of age-related cognitive decline was used to evaluate neurogenesis in relation to hippocampal function. New hippocampal cell survival was assessed approximately 1 month after a series of intraperitoneal injections of 5-bromo-2′-deoxyuridine (BrdU). Correlational analyses between individual measures of BrdU-positive cells and performance on the Morris water maze task provided no indication that this measure of neurogenesis was more preserved in aged rats with intact cognitive abilities. On the contrary, among aged rats, higher numbers of BrdU-positive cells in the granule cell layer were associated with a greater degree of impairment on the learning task. Double-labelling studies confirmed that the majority of the BrdU+ cells were of the neuronal phenotype; the proportion of differentiated neurons was not different across a broad range of cognitive abilities. These data demonstrate that aged rats that maintain cognitive function do so despite pronounced reductions in hippocampal neurogenesis. In addition, these findings suggest the interesting possibility that impaired hippocampal function is associated with greater survival of newly generated hippocampal neurons at advanced ages.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. References

Neurogenesis within the hippocampal formation occurs throughout life in species ranging from rodent to human (Altman & Das, 1965; Caviness, 1973; Eriksson et al., 1998). Newly generated cells within the adult dentate gyrus can differentiate into neurons and become functionally integrated into existing hippocampal circuitry (Kaplan & Hinds, 1977; Hastings & Gould, 1999; Markais & Gage, 1999; van Praag et al., 2002). Adult neurogenesis has been implicated in mnemonic function, and defining a role for the new neurons in memory processes is a topic of great current interest (Lemaire et al., 2000; Shors et al., 2001). Experiments investigating the link between neurogenesis and hippocampal function in young animals have yielded some supportive evidence. Kemperman & Gage (2002) reported a relationship between higher levels of hippocampal neurogenesis and better acquisition, but not probe trial performance, in young mice trained on the Morris water maze task. In addition, Gould et al. (1999) reported an increase in the survival of new hippocampal neurons after different types of hippocampal-dependent learning that included the spatial version of the Morris water maze, supporting a role for these new cells in cognitive processes. Using the toxin methylazoxymethanol acetate, which disrupts cell proliferation, these investigators found impairment in trace eyeblink conditioning but did not find impairment in their water maze task (Shors et al., 2001, 2002).

With respect to aging, mechanisms that regulate hippocampal neurogenesis are also reportedly altered with advanced age (Stenvers et al., 1996; Cameron & McKay, 1999; Bizon et al., 2001; Lichtenwalner et al., 2001). Moreover, studies of aged rodents have demonstrated marked decreases in neurogenesis (Seki & Arai, 1995; Kuhn et al., 1996; Lichtenwalner et al., 2001), suggesting that such reductions might contribute to the well-documented age-related deficits in cognitive capacities supported by the hippocampus. We recently completed a study, however, that failed to find a relationship between the degree of age-related reduction in new cell proliferation in the hippocampus and hippocampal function as assessed by the Morris water maze in a naturally occurring rat model of cognitive aging (Bizon & Gallagher, 2003). In this model, the status of learning and memory supported by the hippocampus varies considerably across aged individuals (Gallagher et al., 1993). These individual differences provide a useful context for evaluating the functional significance of neural alterations associated with aging and identifying age-related changes that contribute to cognitive decline. Indeed, many neurobiological changes in the hippocampal formation have been correlated with cognitive status in this aged study population (Colombo et al., 1997; Nicolle et al., 1999, 2001; Rapp et al., 1999; Smith et al., 2000; Bizon et al., 2001; Colombo & Gallagher, 2002). For example, in the dentate gyrus, volumetric changes and attenuated synaptophysin staining indicating a loss of integrity in the input from entorhinal cortex are associated with worse cognitive performance in the aged animals (Rapp et al., 1999; Smith et al., 2000).

Our previous neurogenesis study in this model focused on the birth of new cells within the hippocampus and did not address the long-term survival and differentiation of these cells with respect to hippocampal function during aging. Evaluating survival and differentiation is important because only a fraction of hippocampal cells born in the adult survive for long periods of time and ultimately differentiate into neurons that might contribute to cognitive function (Dayer et al., 2003). Moreover, certain trophic factors and other cell survival mechanisms that are affected by age (Gallagher et al., 1993; Sugaya et al., 1998) might preferentially influence the survival, rather than the proliferation, of new hippocampal neurons (Gould et al., 1999; Lee et al., 2002). Finally, recent evidence indicates that some markers of neurogenesis, including those associated with survival and differentiation, are actually elevated in Alzheimer's disease, a pathological aging condition that is associated with profound memory impairment (Jin et al., 2004).

The current study investigated the relationship between age-related changes in the survival and differentiation of adult generated hippocampal cells and spatial learning performance. The current data provide no evidence that preserved neurogenesis in aging is associated with preserved cognitive abilities in the Long-Evans model of aging. By contrast, an inverse relationship between the survival of putative neurons born in the aged hippocampus and spatial learning was observed, suggesting that greater neurogenesis in aging brain might be associated with impaired cognitive abilities.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. References

Spatial learning performance

The results of the spatial learning assessment in young and aged rats were similar to those of previous studies using this protocol (Gallagher et al., 1993), with both the young and the aged rats improving their performance during the course of training (F3,57 = 46.2, P < 0.0001). Although the average search error of the young and aged rats did not differ on the first training trial (F1,19 = 2.1, n.s.), aged rats were subsequently impaired in the acquisition of the spatial learning task (F1,19 = 13.0, P < 0.01; Fig. 1a). Figure 1(b) shows individual learning index scores for young and aged rats. The spatial learning index is derived from search error measures on probe trials interpolated throughout the training protocol. Lower learning indices indicate better performance. Analysis of this measure confirmed an impairment of the aged group relative to the young group (F1,19 = 17.3, P < 0.0005); however, note the wide range of individual performance within the aged group. Some aged rats had learning indices similar to young rats whereas others fell outside the range of the young group. Young and aged rats had comparable escape latencies in locating a visible platform during cue training (young: 11.9 ± 5.7; aged: 12.4 ± 2.1).

image

Figure 1. Performance in hippocampal-dependent spatial learning. (a) Cumulative search error ± SE averaged over five trial blocks is shown for young (•) and aged (○) rats over the course of training in the spatial water maze task. Aged rats were impaired relative to young rats (see text for statistical analysis). (b) Distribution of individual spatial learning indices for young and aged rats. Note the variability in performance among aged rats, with some aged rats performing on par with young cohorts and others demonstrating impairment on the task.

Download figure to PowerPoint

BrdU+ cell survival in hippocampus: relation to cognitive performance

Representative hippocampal 5-bromo-2′-deoxyuridine (BrdU)-positive (BrdU+) labelling for young and aged rats is shown in Fig. 2. In both age groups, labelled cells were situated within the granule cell layer and the hilus of the dentate gyrus 3 weeks after the final BrdU injection. As shown in Fig. 3(a), stereological cell counts of BrdU+ profiles revealed a pronounced (−85%) age-related decline in the number of BrdU+ cells in the granule cell layer (F1,19 = 88.1, P < 0.0001). A much more modest (−24%), but significant, age-related decline was also observed in the hilus (F1,19 = 5.0, P < 0.05). To assess the relationship between the BrdU+ hippocampal cells and spatial learning ability in aged rats, a Pearson's correlation was performed on individual subjects within this age group (Fig. 3b). The results of these analyses yielded a significant positive correlation among the aged subjects between the number of BrdU+ cells in the granule cell layer and the spatial learning index, such that higher numbers of BrdU+ cells were associated with poorer spatial learning abilities (r = 0.57, P < 0.05). At the same time, no significant relationship was found between the number of BrdU+ cells in the hilus and spatial learning ability.

image

Figure 2. Labelling of BrdU+ cells in young and aged rats. Bright-field photomicrographs (20×) show BrdU-positive profiles, labelled with black-coloured reaction product, located in the dentate gyrus of representative young (a) and aged (b) rats with a 3-week survival after BrdU administration. Note the pronounced decrease in BrdU+ labelling in the granule cell layer (gcl) of the aged rat.

Download figure to PowerPoint

image

Figure 3. New cell survival in the granule cell layer with age: relationship to spatial learning performance. (a) Bar graphs showing the averaged number of BrdU+ cells ± SE in young (▪) and aged (□) rats with a 3-week survival following BrdU administration. Note the large decrease of BrdU+ cells in the granule cell layer of aged rats. See text for statistical analysis. (b) The relationship between the number of BrdU+ cells in the granule cell layer and spatial learning performance in aged rats. Lower learning indices indicate better performance. A significant negative correlation was observed among aged subjects; greater numbers of BrdU+ cells were associated with poorer performance on the spatial task. See text for statistical analysis.

Download figure to PowerPoint

Differentiation of new hippocampal neurons: relation to cognitive performance

Double-labelling studies were performed in order to assess the percentages of BrdU+ cells in the granule cell layer that were of the neuronal phenotype. Figure 4 shows representative BrdU+ cells from young and aged subjects that are also immunoreactive for the neuronal marker, NeuN. Confocal analyses of double-labelled cells revealed that the majority of BrdU+ cells in both age groups coexpressed NeuN; however, there was a small (−19%), but significant, age-related decline in the percentage of BrdU+ cells in the granule cell layer that were double-labelled with NeuN (F1,19 = 5.9, P < 0.05). Only a small fraction (5%) of BrdU+ cells in the aged rats and no BrdU+ cells in young rats were found to coexpress the marker of mature astrocytes, glial fibrillary acidic protein (GFAP).

image

Figure 4. Neuronal differentiation of BrdU+ cells in young and aged rats. High power (100×) photomicrographs showing BrdU (red) and NeuN (green) double-labelling in the granule cell layer. Panels (a) and (b) show representative examples of BrdU+/NeuN+ cells in a young and aged rat, respectively. Panel (c) shows an example of a single-labelled BrdU+ cell (red) located in the apex of the granule cell layer. Scale bar = 10 µm.

Download figure to PowerPoint

To evaluate whether the percentage of newly generated cells that differentiated into neurons differed with respect to the cognitive status of the aged subjects, the aged rats were subgrouped on the basis of their water maze performance. The ‘best’ and ‘worst’ performers among aged rats (n = 7 in each group) had an average spatial learning index score of 217 ± 5.9 and 278 ± 7.5, respectively. This difference in spatial learning ability was significant between these groups as assessed using a one-way anova (F1,12 = 40.9, P < 0.0001). In agreement with the significant correlation among aged subjects, higher numbers of BrdU+ cells were observed in the worst compared with the best aged rats, although this difference fell just sort of significance (Fig. 5; F1,12 = 4.45, P = 0.056). However, the percentage of BrdU+ cells coexpressing NeuN (‘best aged’ = 65 ± 6%, ‘worst aged’ = 69 ± 5%) did not significantly differ between the best and worst performers on the spatial task (Fig. 5; F1,12 = 0.31, n.s.). These analyses demonstrate that despite a significant age-related reduction in the differentiation of new hippocampal cells into neurons, similar percentages of adult generated hippocampal cells differentiate into neurons across a broad spectrum of cognitive outcomes in aging.

image

Figure 5. Bar graph showing the total number of BrdU+ cells (open bar) and the proportion of BrdU+ cells double-labelled with NeuN (shaded bar) in aged rats subgrouped with respect to spatial learning performance. The difference between groups in the total number of BrdU+ cells was very nearly significant (P = 0.056) but there was no difference in the proportion of BrdU+ cells in each group double-labelled with NeuN (see text for statistical analysis). An additional comparison was conducted on these combined datasets to provide a rough estimate of the total number of new neurons in each group. This analysis revealed a trend toward higher numbers of new neurons in the worst performing aged compared with the best performing aged groups (F1,12 = 3.1, P = 0.1).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. References

The use of a model in which individual differences occur in cognitive aging provides a setting that is relevant to the potential role of neurogenesis in normal hippocampal function. Indeed, many neurobiological alterations in hippocampus have been found to be predictive of cognitive impairment in the study population used in the current investigation (Colombo et al., 1997; Nicolle et al., 1999, 2001; Rapp et al., 1999; Smith et al., 2000; Bizon et al., 2001; Colombo & Gallagher, 2002). As demonstrated by the current data, however, a pronounced loss of neurogenesis is detected during the aging process even in animals that maintain spatial learning capacity. As such, these data are consistent with evidence that interference with neurogenesis using antimitotic agents in young rats spares certain cognitive abilities that depend on hippocampal function (Shors et al., 2002). The large magnitude of reduced neurogenesis across the aged spectrum of cognitive ability suggests that other neurobiological alterations in this system play a more determining role in cognitive outcome.

Our initial work examining new cell birth in the dentate gyrus of young and aged rats in the Long-Evans rat model failed to detect any relationship between new cells born in the dentate gyrus and spatial learning performance of rats at any age (Bizon & Gallagher, 2003). In support of this finding, other investigators also failed to find a relationship between individual differences in water maze performance and proliferation in the granule cell layer in aged female Fisher 344 rats (Merril et al., 2003). By contrast, it should be noted that a relationship has been reported between greater rates of proliferation and better water maze performance in aged male Sprague–Dawley rats (Drapeau et al., 2003). The discrepancy between this result and the findings of Bizon & Gallagher (2003) and Merril et al. (2003) is at present unclear. The use of different rat strains offers one plausible explanation. Consideration of rat strain can be particularly important in aging research given the genetic influences that might control processes such as cell death, hormonal changes and inflammation, all of which might be very relevant to the regulation of neurogenesis. It will be important to characterize each of these models carefully across such modalities because unique and overlapping findings across animal models might provide valuable insight into human aging processes.

Another consideration when comparing the existing data on the relationship between spatial learning and new cell production in the hippocampus relates to the age of the rats included in each study. Both studies that reported an absence of a relationship between spatial learning and hippocampal cell genesis used young rats that were 6 months of age and aged subjects that were 24–27 months of age, whereas the rats used in both age groups in Drapeau et al. (2003) were substantially younger (young: 3 months; aged: 20 months). Bizon & Gallagher (2003) have shown that hippocampal cell genesis is decreased progressively across the lifespan. Moreover, the current results suggest that the relationship between recently generated hippocampal cells and spatial learning might radically change across the lifespan, but exactly when this occurs and the mechanisms responsible for the change remain uncertain. A redundancy in neurobiological mechanisms involved in learning is likely and, indeed, there is evidence that learning mechanisms used at advanced ages can be equally effective but very different from those recruited in young adulthood (e.g. Hazlett et al., 1998; Cabeza et al., 2002). The accumulating data clearly indicate that neurogenesis is not necessary for spatial learning in either young (Shors et al., 2002) or aged rats (Bizon & Gallagher, 2003; Merril et al., 2003); however, recently generated neurons might be engaged in new spatial learning at some points during the lifespan whereas other neurobiological mechanisms might be preferentially recruited early and very late in life.

There is accumulating evidence that the role of new neurons in mnemonic processes might differ between young and aged subjects. Kempermann & Gage (2002) have previously described a positive correlation among young mice, with higher numbers of newly generated neurons in the granule cell layer associated with better acquisition of the water maze task. Although the number of young subjects in the current study was limited (n = 7), there was a trend in the same direction in the current material (r = −0.4, n.s.). By contrast, we also report here an unexpected significant correlation among aged rats, in which greater survival of new cells was associated with worse cognitive abilities. These latter findings are particularly interesting in light of a new study reporting that markers of hippocampal neurogenesis are elevated in humans with the age-related memory disorder Alzheimer's disease (Jin et al., 2004). It appears that endogenous mechanisms may be in place in both normal and pathological aging that can stimulate different components of hippocampal neurogenesis. Indeed, as we have shown in the Long-Evans model, the greater survival of new hippocampal cells born in aged rats is linked to impaired cognitive abilities.

Obviously, the existing data indicate that any endogenous up-regulation of new neurons that occurs during aging is insufficient to restore cognitive abilities. Perhaps the lack of functional benefit is because the number of new neurons in the aged subjects remains insufficient. Indeed, even the highest numbers of hippocampal BrdU-positive cells observed in aged cognitively impaired rats in our study were far below values for young rats. Several studies have shown that aged rats do have the capacity to regain young levels of hippocampal neurogenesis; neurochemical, endocrine and environmental manipulations can dramatically up-regulate neurogenesis in aged rodents (Cameron & McKay, 1999; Lichtenwalner et al., 2001; Kempermann et al., 2002). Thus, it remains an intriguing question as to whether boosting neurogenesis to a greater degree in the aged rats, for example to a level on a par with younger animals, would have significant functional benefit and result in reversal of cognitive deficits.

An alternative explanation for the observed correlation between BrdU-positive cell number and impaired cognition in the aged group is that neurobiological changes that are associated with hippocampal dysfunction and cognitive impairment produce a secondary effect on the survival of newly born neurons. For example, recent evidence from the aging model used in the present study indicates that oxidative stress is more pronounced in aged rats with cognitive impairment relative to aged cohorts with preserved cognitive abilities (Nicolle et al., 2001). Expression of certain survival factors might be endogenously up-regulated to combat cellular stress and in turn promote a trophic environment that favours the survival of newly generated granule cells. For example, growth factors such as brain-derived neurotrophic factor (BDNF) have been implicated in the survival of hippocampal granule cells both during development and in the adult brain (e.g. Lowenstein & Arsenault, 1996; Lee et al., 2002). Previous work in this aging model found that hippocampal BDNF expression was related to spatial learning performance and that BDNF mRNA was significantly elevated in the hippocampus of aged rats (Sugaya et al., 1998). Thus, BDNF expression during aging might contribute to the enhanced survival observed in the aged rats with impaired function. It will be of substantial interest to elucidate further the molecular mechanisms whereby the survival of hippocampal cells is preferentially sustained in cognitively impaired aged rats.

Another possibility for the relationship between elevated markers of neurogenesis and worse memory in aging relates to the possibility that new neurons produced in the aged brain are not integrated into hippocampal functions. Perhaps, the inability to make use of newly formed neurons is responsible for their failure to improve cognitive function. If neurons born in the aged brain are not functional or long-lived, then rather than conferring functional benefits, these cells might be detrimental to mnemonic function in aged cognitively impaired subjects. Certainly, the cellular environment in the aged brain is significantly different than in the young adult and, as such, the integration of new neurons into hippocampal functions might be compromised. Data obtained using unbiased stereology indicate that the total number of granule cells in the dentate gyrus is maintained during aging in the Long-Evans study population, irrespective of cognitive outcome (Rapp & Gallagher, 1996). These data, together with the current findings, suggest that any increase in neurogenesis in the dentate gyrus of cognitively impaired aged rats occurs at the expense of losing more mature granule cells. Indeed, future work directed at determining the functional status of recently generated and existing granule cell neurons born in the aged brain is an important first step in targeting hippocampal neurogenesis for therapeutic treatments of memory dysfunction associated with aging.

In agreement with Lichtenwalner et al. (2001), an age-related decrease in the proportion of BrdU/NeuN double-labelled cells was observed in the present study. Nevertheless, a high proportion of newly generated cells located in the granule cell layer of aged rats still expressed the neuronal marker NeuN in the current material. High levels of BrdU/NeuN coexpression (> 50%) were similarly observed by Drapeau et al. (2003), although other groups have reported much more pronounced effects of age on differentiation of new hippocampal cells into the neuronal phenotype (Lichtenwalner et al., 2001; Kempermann et al., 2002; Heine et al., 2004). Beyond possible species and strain differences, methodological variations in the sample size and location of BrdU+ cells selected for phenotypic analysis could contribute to these variable findings. In addition, it is interesting to consider that even slight variations in the age of BrdU+ cells employed for phenotype studies might significantly impact differentiation estimates because both rate of differentiation and the longevity of fully differentiated neurons in the aged brain are not well characterized. Finally, it is notable that some groups report difficulty with penetration of the NeuN antibody (Cameron & McKay, 1999), a fact that could contribute to an underestimation of double-labelled BrdU/NeuN cells. In the current material, NeuN labelling was present throughout the section thickness. Most relevant to the current study is that despite a relatively large proportion of BrdU cells in the aged brain coexpressing NeuN, the percentage of new cells that differentiated into neurons was comparable across aged rats that differed significantly with respect to cognitive function. These data indicate that phenotype differentiation of newly formed hippocampal cells among aged rats does not appear to be a significant factor in the maintenance or failure of cognitive abilities.

The current analysis does not support the notion that the decrease in neuronal differentiation with age was due to an increase in the number of new cells that differentiate into astrocytes, as only 5% of the BrdU+ cells in aged rats were found to coexpress GFAP. We cannot exclude the possibility that this percentage is an underestimate because, we did not observe full penetration of the GFAP antibody in the current material, contrasting the results with NeuN. Nevertheless, the phenotype analysis was not able to account for a sizable population (about 35%) of adult generated cells in aged rats that survive many weeks after birth. The single-labelled BrdU+ cells may belong to a yet undetermined phenotype or might reflect new cells that have not yet fully differentiated 3 weeks after birth. It will be of substantial interest in the future to define each of the cell types in the aged rat brain to which new cells are being added throughout life and, ultimately, examine their contribution to cognitive processes.

The current study reports the novel finding that the survival of recently generated hippocampal cells is inversely related to cognitive performance during aging. These data strongly support the notion that age-related decreases in hippocampal neurogenesis are not responsible for the emergence of cognitive deficits but leave open the possibility that aged rats with impaired cognitive abilities are attempting to recruit new neurons to compensate for the impairment.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. References

Subjects

Pathogen-free young (n = 7, 7 months old) and aged (n = 15, 25 months old) male Long Evans rats (Charles River Laboratories) were maintained on a 12-h light/dark cycle (lights on at 08:00 h) and climate controlled at 25 °C. All animal procedures were in accordance with approved institutional animal care procedures and NIH guidelines.

Water-maze testing

The maze consisted of a circular tank (diameter 183 cm; wall height 58 cm) filled with water (27 °C) made opaque with powdered milk (0.9 kg). An escape platform (height 34.5 cm) was located 1 cm beneath the water surface. White curtains affixed with large black geometric designs provided extramaze cues. Data were analysed using an HVS Image Analysing VP-116 video tracking system and an IBM PC with software developed by HVS Imaging (Hampton, UK).

As detailed previously (Gallagher et al., 1993), rats received three trials per day for eight consecutive days using a 60-s intertrial interval. For spatial learning assessment, the location of the platform remained constant and the starting position for each trial was varied. Every sixth trial was a probe, during which the platform was retracted for the first 30 s of the trial. After place training, rats received one session (six trials) of cue training using a visible platform located 2 cm above the water surface that varied in position from trial to trial.

Accuracy of performance in spatial learning was assessed using search error computed from training trials and a learning index score computed from probe trials. The distance of the rat from the platform was sampled ten times per second and these distances were averaged into 1-s bins. Search error is the sum of these 1-s averages on training trials corrected for start and platform location. The learning index is derived from average proximity (search error divided by the probe trial length) on the second, third and fourth probe trials. Scores from these trials are weighted and summed to provide a specific measure of hippocampal-dependent learning (Gallagher et al., 1993).

BrdU administration

One week after testing, BrdU (50 mg kg−1) dissolved 0.9% saline was administered (i.p.), twice daily (09:00 h and 17:00 h) for 5 days. Three weeks following the last injection, rats were given an overdose of sodium pentobarbital (100 mg kg−1) and perfused with 4% paraformaldehyde in 0.1 m phosphate buffer, pH 7.4. Brains were removed, post-fixed, cryoprotected in 0.1 m phosphate-buffered saline (PBS) containing 20% sucrose, frozen on powdered dry ice and stored at −80 °C. Brains were sectioned on a freezing microtome and a one in six series of sections through the hippocampus (40 µm, coronal) was collected into cold cryoprotectant solution (25% ethylene glycol, 25% glycerol, 50% 0.1 m phosphate buffer, pH 7.4) and stored at −20 °C until further processing. One complete series of sections was processed for single BrdU immunohistochemistry and another adjacent series was processed for triple immunofluorescence for BrdU, NeuN and GFAP.

BrdU immunohistochemistry

For DNA denaturation, free-floating tissue sections were pretreated (2 h, 60 °C) in 50% formamide/2× SSC, rinsed in 2× SSC, incubated in 2 n HCl (30 min, 37 °C) and rinsed in 0.1 m sodium borate buffer, pH 8.5 (10 min). After PBS rinses, sections were washed in 0.6% H2O2 (30 min) and pre-incubated in 3% normal goat serum (NGS) and 0.3% Triton-X in PBS (1 h, 4 °C). Sections were then incubated for 72 h in the same solution including an anti-BrdU primary antibody (Accurate Chemicals, 1 : 100). After PBS washes, sections were incubated in a biotinylated secondary antibody (1 : 200) and 3% NGS in PBS (1 h) and then avidin/biotin complex (Vector Laboratories, 2 h). Finally, sections were reacted using a Vector SG substrate kit for peroxidase. Sections were mounted onto Superfrost++ slides, air-dried, counterstained with Nuclear Fast Red (Vector Laboratories), dehydrated through washes in ascending concentrations of alcohol, defatted in Xylene and coverslipped with Permount.

Immunofluorescence

Free-floating tissue sections were pretreated as described above. Sections were then rinsed in Tris-buffered saline (TBS; 100 mm Tris-HCl, 150 mm NaCl, pH 7.5), and pre-incubated in blocking solution containing 3% normal donkey serum (NDS), 0.3% Triton X-100 in TBS for 1 h. Sections were then incubated in the blocking solution with the addition of an anti-BrdU primary antibody (Accurate Chemicals, 1 : 100), an anti-NeuN primary antibody (Chemicon, 1 : 200) and an anti-GFAP primary antibody (Chemicon, 1 : 150) for 72 h. After the primary incubation, sections were rinsed in TBS and incubated in TBS containing 2% NDS, goat anti-mouse IgG–Alexa 488 (Molecular Probes, 1 : 300), donkey anti-rat IgG–Cy3 (Jackson Immunochemicals, 1 : 300) and donkey anti-rabbit IgG–Cy5 (Jackson Immunochemicals, 1 : 300) for 2 h in the dark. Sections were then mounted on to Superfrost++ slides, coverslipped with Gel/Mount (Biomeda), sealed with Cytoseal 60 and stored in the dark at −20 °C until analysis.

Quantification of BrdU-positive profiles

Stereological principals and analyses were conducted as described in Rapp & Gallagher (1996) and Bizon & Gallagher (2003). All analyses were conducted blind with respect to age and cognitive status. Labelled hippocampal cells were exhaustively counted in both hemispheres of every sixth section through the entire rostrocaudal extent of the hippocampus (240 µm apart) using a combination of 400× and 1000× magnification. The NeuroLucida system consisted of a colour video camera interfaced to a Leitz Medilux microscope. A computer-driven motorized stage allowed sections to be surveyed at evenly spaced x–y intervals and standardized on the z-axis. Total section thickness was determined for each section using a microcator (resolution = 0.5 µm). Using a 3-µm guard (in which cells were not counted), the BrdU+ nuclei were counted as they came into focus while scanning through the next 15 µm of the tissue section thickness (z-axis). The total number of BrdU-positive cells was estimated by multiplying the sum of the neurons counted by the reciprocal of the volume of the hippocampus that was sampled (i.e. the fraction of histological sections examined and the fraction of the total section thickness examined).

Analysis of immunofluorescence

Sections labelled by immunofluorescence for BrdU, NeuN and GFAP were evaluated by confocal microscopy (Carl Zeiss Laser Scanning System LSM 510). Immunoreactive cells were visualized using a 63× oil-immersion objective. Each BrdU+ cell analysed was sectioned in the z-plane (z-stacks consisted of 1-µm slices). Zeiss LSM ImageBrowser software was used for analysis. This software allows the observer to step systematically through the 1-µm z-slices in all dimensions and examine each label independently and in combination. Between 30 and 50 BrdU+ cells from each of the rats included in this study were z-plane sectioned and analysed for the coexpression of NeuN or GFAP. NeuN and BrdU immunoreactivity were observed throughout the entire thickness of the section; by contrast, GFAP immunoreactivity did not penetrate the entire section in every case. Every attempt was made to confine the confocal analysis to regions in which all antibodies were fully present. Nevertheless, the number of BrdU+ cells expressing GFAP is probably an underestimate.

Statistical analyses

Age comparisons on the spatial task were made using a two-way anova (age × trial block) to compare cumulative search error, and one-way anovas to compare learning index scores from spatial probe trials and mean path lengths from cue training trials. Numbers of BrdU+ profiles in the granule cell layer and hilus were averaged separately for each age group and the significance of age differences was determined using a one-way anova with a Fisher's PLSD used for post-hoc analyses. In order to investigate the relationship between the total number of BrdU+ cells and cognitive status, separate correlation analyses (Pearson's r) were conducted on values for individual animals in the aged group. The percentage of BrdU+ cells that expressed NeuN was assessed using a one-way anova to compare data obtained from young (n = 7) and aged rats subgrouped with respect to cognitive status (n = 7 best performers and n = 7 worst performers). In total, the numbers of BrdU+ cells per group analysed were as follows: young = 281 cells, best-aged n = 261 and worst-aged n = 241.

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. References

We wish to thank Joe Cantalini and Derek Young for technical assistance. Grant sponsor: NIA; Grant number: PO1-AG09973 (to M.G.) and F32-AG19601 (to J.L.B.).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. References