We here report that voluntary wheel running led to a regional increase in the number of newly generated cortical microglia. We asked how adult cortical cell genesis would respond to environmental enrichment and physical activity, both stimuli that robustly induce adult hippocampal neurogenesis. After labeling proliferating cells with bromodeoxyuridine (BrdU) and immunohistochemical detection of BrdU, we found that both experimental paradigms did not result in general effects on cell proliferation and cell genesis in the neocortex. However, there were regionally and layer specific changes in the number of BrdU marked cells, both 1 day and 4 weeks after BrdU. Environmental enrichment led to a significant increase in the number of new astrocytes in layer 1 of the motor cortex. Voluntary wheel running, in contrast, caused an induction in the proliferation of microglia in superficial cortical layers of several brain regions. Under no condition was the number of new oligodendrocytes measurably enhanced. In contrast to the hippocampus, we did not find any new neurons in the cortex. The physiological ‘activation’ of microglia adds a new aspect to the question of microglial function in the healthy brain and of how adult brain cells can plastically react to physiological stimuli.
Cell genesis in the adult mammalian neocortex has received considerable interest, because it was claimed that new neurons could be generated in the adult cerebral cortex (Gould et al., 1999, 2001). The scientific debate that arose (Nowakowski and Hayes, 2000; Grassi Zucconi and Giuditta, 2002; Rakic, 2002) took place in the light of earlier reports on adult neurogenesis in other brain regions (Altman and Das, 1965; Cameron et al., 1993; Lois and Alvarez-Buylla, 1993; Kuhn et al., 1996).
In the dentate gyrus of the adult hippocampus, new neurons are continually generated from a local population of stem or progenitor cells. Adult neurogenesis and its regulation can be linked to hippocampal function (Kempermann, 2002; Kempermann and Gage, 2002), although details of the functional relevance are still unresolved. Similarly, stem cells in the wall of the lateral ventricles sustain adult neurogenesis in the adult olfactory system and here, too, first theories have aimed a linking neurogenesis to specific functional consequences (Petreanu and Alvarez-Buylla, 2002). These findings have sparked interest in other regions of the adult brain that contain neural stem cells (Palmer et al., 1999), but normally do not seem to produce new neurons in situ. Adult cortical neurogenesis had not been observed in mammals until Gould et al. reported a surprising amount of cortical neurogenesis in primates (Gould et al., 1999, 2001). Their claim was disputed by Kornack and Rakic (Kornack and Rakic, 2001) and Koketsu et al. (Koketsu et al., 2003). Magavi et al. found reactive adult cortical neurogenesis in mice only after a highly specific photoablation of thalamo-cortical projection neurons, but not under normal conditions (Magavi et al., 2000).
One possible explanation is that cortical neurogenesis is not generally impossible, but is repressed under normal conditions. This would be similar to neurogenesis in hippocampal region CA1, where no (or perhaps extremely limited) neurogenesis occurs under normal conditions, but can be stimulated in a pathological situation (Nakatomi et al., 2002).
The adult brain responds to the sensory richness of the environment, the complexity of the stimuli and the degree of physical involvement of a subject in experiencing this environment with a broad range of measurable changes. The adult hippocampus seems to be particularly sensitive to such influences (Jones and Smith, 1980), but it has sometimes been suggested that this particular sensitivity reflects a certain hippocampal bias among researchers. But even the earliest research on effects of living in an enriched environment on brain morphology has revealed cellular effects in cortical areas (Altman and Das, 1964; Rosenzweig, 1966; Diamond et al., 1967, 1976). Beaulieau and Colonnier reported experience-dependent changes in the adult visual cortex of the cat, including effects on neuron numbers (Beaulieau and Colonnier, 1987 Beaulieau and Colonnier, 1989). In the early 1980s, M. Kaplan had reported a similar observation in rats (Kaplan, 1981). However, these findings were not taken up by other researchers and remained unconfirmed.
The stimulatory effects of environmental enrichment (Kempermann et al., 1997b, 2002) and voluntary physical activity (Van Praag et al., 1999a,b) on adult hippocampal neurogenesis made us revisit the early findings of experience-dependent effects on cortical cell genesis. With regard to the hippocampus, neural stem cell biology has already fundamentally changed general views on neuroplasticity.
Even in the earliest reports on cell divisions in the adult brain, cell proliferation in cortical regions has been mentioned (Globus and Kuhlenbeck, 1944; Bryans, 1959; Smart and Leblond, 1961), but no studies have addressed this issue systematically and there is only limited knowledge about the exact nature, dynamics and consequences of this proliferative activity. The earlier reports agree that new glial cells are produced in the adult cortex (Korr et al., 1973; Kaplan and Hinds, 1980; Levison and Goldman, 1993) and this is the widely held opinion today. The discussion on adult cortical neurogenesis has now questioned this tacit assumption.
We designed the following experiment to quantify cell genesis in the adult murine cerebral cortex in response to stimuli known to induce neurogenesis in the adult hippocampus. We examined cell proliferation in the cingulate, motor, somatosensory, insular and visual cortex and studied the distinguishable layers in these regions separately. Cell proliferation was assessed immunohistochemically after incorporation of thymidine analogue bromodeoxyuridine (BrdU) into the DNA of cells in s-phase. We compared mice living under three different conditions: (i) in an enriched environment, i.e. a large cage with more opportunities for social and exploratory behavior (but not containing a running wheel); (ii) in a cage equipped with a running wheel; and (iii) in standard laboratory cages.
Our hypothesis was that cell genesis in the adult neocortex would not occur in a random or static distribution, but would show a regionally specific response to activity-dependent regulatory influences.
Materials and Methods
Housing Conditions and BrdU Injections
Forty-eight female C57BL/6 mice, ~2 months old at the beginning of the experiment, were obtained from Charles River (Sulzfeld, Germany). Animals were randomly assigned to the three experimental conditions. One group (ENR) lived in an enriched environment as described previously (Kempermann et al., 2002), except that in the present study the enriched cage did not contain a running wheel. The entire group of 16 mice lived in a large cage of 80 × 80 cm floor area with re-arrangeable plastic tubes, nesting material and mesh wire ladders. They received the same type of food as the other groups. The second group (RUN) inhabited standard-size cages equipped with a running wheel (Tecniplast). The third group (CTR) lived under standard laboratory conditions. ENR stayed in the enriched environment for 40 days and received one daily i.p. injection of BrdU (5-bromo-2-deoxyuridine; daily dose, 50 μg/g body wt; Sigma) during the last 10 days of this period. RUN had access to the running wheel for 10 days, during which they received daily BrdU injections. In parallel, CTR received one daily BrdU injection for 10 days. In each group, eight mice were perfused 1 day after the last injection of BrdU and the remaining mice 4 weeks later, during which time they continued to live under their respective experimental conditions. All mice had access to water and food ad libitum and lived in a 12 h/12 h dark/light cycle. Local and federal regulations regarding animal welfare were followed.
The mice were killed with an overdose of ketamine and perfused transcardially with 4% paraformaldehyde in cold 0.1 M phosphate buffer (pH 7.4). The brains were left in the fixative for 24 h and then transferred into 30% sucrose. Forty micrometer coronal sections were cut from a dry-ice-cooled block on a sliding microtome (Leica). The sections were stored at −20°C in cryoprotectant with 25% ethylene glycol, 25% glycerine and 0.05 M phosphate buffer. Sections were stained with free-floating immunohistochemistry and prepared for BrdU detection by incubation in 2 N HCl for 30 min at 37°C and washing in 0.1 M borate buffer (pH 8.5) for 10 min.
All antibodies were diluted in Tris-buffered saline (TBS, pH 8.5) containing 0.1% Triton X-100 and 3% donkey serum (TBS-plus). Antibodies were tested with the appropriate negative controls (reciprocal omission of primary and secondary antibodies).
The primary antibodies used in this study were: monoclonal rat anti-BrdU (Harlan Seralab) 1:500; monoclonal mouse anti-NeuN (Chemicon) 1:100; polyclonal rabbit-anti S100β (SWant, Bellinzona, Switzerland) 1:2000; monoclonal mouse anti-CNP (Abcam) 1:500; monoclonal mouse anti-APC (Oncogene) 1:200; monoclonal mouse anti-RIP (Developmental Studies Hybridoma Bank, University of Iowa) 1:200; and polyclonal rabbit anti-Iba-1 (a generous gift from Drs Y. Imai and S. Kohsaka, National Institute of Neuroscience, Tokyo, Japan) 1:70. Fluorescent secondary antibodies from donkey conjugated with FITC, rhodamine-X, or CY5 were obtained from Jackson Laboratories (Distributor: Dianova) and all used at 1:250.
For immunofluorescent quantification of both BrdU-labeled cells and of BrdU in combination with phenotypic markers, every sixth section throughout the right cerebral hemisphere was used. After pretreatment (see above) and a blocking step with TBS-plus, sections were incubated in a mixture of the three antibodies of each series for 36 h (45 h for Iba-1) at 4°C. After washing in TBS and TBS-plus, a cocktail of secondary antibodies (rhodamine X to detect BrdU, FITC for NeuN and Iba-1, and CY5 for S100β, CNP or Iba-1) was applied for 2–5 h at room temperature. Sections were mounted in polyvinyl alcohol with diazabicyclo-octane (DABCO) as anti-fading agent.
Nuclei were unambiguously identified by histone staining with TO-PRO3 (Molecular Probes), 1:1000, 15 min in TBS.
Fluorescent signals were detected using a spectral confocal microscope (Leica TCS SP2). For each region, the phenotypes of 40 (S100β) or 50 (Iba-1) BrdU-labeled cells per animal were determined. All analyses were done in sequential scanning mode in order to rule out cross-bleeding between detection channels. Images were processed with Adobe Photoshop 6.0 (Adobe Systems).
Quantification of BrdU-labeled cells in cortical regions and layers was done using a double-stained series with immunoreactions for BrdU (rhodamine-X) and NeuN (FITC) on sections 240 mm apart, covering the entire fronto-occipital extension of the hemisphere. Analyses were performed on a Leica DM-RXE microscope equipped with StereoInvestigator 4 (MicroBrightfield). Cortical regions (cingulate, motor, sensory, insular and visual) and layers (1, 2–3, 4, 5–6, or 1, 2–6, depending on the detectability of an internal granule cell layer, layer 4) were defined by anatomical criteria based on the NeuN staining and following the Mouse Brain Library atlas (Rosen et al., 2000) which is available at http://www.mbl.org. The region boundaries were traced on a video image, acquired with a Hitachi HV-C20A colour video camera, using the StereoInvestigator software. The analysis was performed using filter cube FI/RH (Leica) that allows the simultaneous visualization of rhodamine-X and FITC. The volume of the identified cortical regions was calculated from the traced section areas by multiplying the area with the distance between the sections plus the section thickness (Cavalieri principle; StereoInvestigator software). As BrdU-labeled cells are comparatively rare, the stereological procedure, described elsewhere (Williams and Rakic, 1988), was modified to exclude the uppermost focal plane only. Otherwise, BrdU-labeled cells were counted exhaustively in all chosen regions and in the same session in which the region boundaries were traced. BrdU counts were expressed as absolute numbers per region or per entire hemisphere (excluding frontal cortex). The same procedure was applied to the hippocampus as described previously (Kempermann et al., 1997a).
The relative distribution of phenotypes per 40 or 50 BrdU-labeled cells within each region (see above) was related to the absolute number of BrdU-positive cells in that same region to yield the absolute number of BrdU-labeled cells of the phenotypes of interest (NeuN, S100β, Iba-1 and CNP).
All statistical analyses were performed with Statview 4.5.1 for Macintosh. Factorial analyses of variance (ANOVA) were performed for all comparisons of morphological data followed by Fisher post hoc test, where appropriate.
Activity-dependent Regulation of Adult Neurogenesis in the Hippocampus
As internal control and baseline, we confirmed that the experimental manipulations of environmental enrichment (ENR) and physical activity (RUN) led to the same type of enhancement as in our previous studies (Kempermann et al., 1997b; Van Praag et al., 1999b). Figure 1A shows that this was the case. Under both conditions, increased numbers of BrdU-labeled cells in the hippocampal dentate gyrus could be found at 4 weeks (short for: ‘4 weeks after the last injection of BrdU’). RUN, in addition, had a significant effect on the number of BrdU-marked cells at day 1 (short for: ‘at day 1 after the last injection of BrdU’). As described previously, there is a decrease in BrdU-labeled cell numbers between 1 day and 4 weeks (Kempermann et al., 1997a, 2003). During this period, cells are eliminated by apoptosis (Young et al., 1999; Biebl et al., 2000). In the following paragraphs we concentrate on differences at 1 day and 4 weeks after BrdU, rather than between the two time-points.
Regionally Specific Regulation of Cell Proliferation and Survival of Newborn Cells
In the same tissue sections in which hippocampal cell genesis was examined and applying analogous, yet adapted methods of analysis, BrdU-labeled cells were quantified in the neocortex. BrdU-marked cells could be found in the entire murine cortex without any obvious spatial preference. Combined analysis over all the cortical regions examined revealed no difference in the number of BrdU-labeled cells between RUN, ENR and CTR (Fig. 1B). This was also true if densities (cells per volume) were analysed and none of the two experimental manipulations altered the volume of the cortex (not shown). If regions and layers were analysed individually, differences in absolute numbers of BrdU-labeled cells emerged (Fig. 2). At day 1, RUN had significantly more BrdU-labeled cells in layer 1 of the cingulate cortex (as compared to CTR). This significant difference in layer 1 of the cingulate cortex could also be seen at 4 weeks. Additionally, at 4 weeks RUN had significantly more BrdU-labeled cells in layer 1 of the motor and visual cortex and fewer BrdU-labeled cells in layers 2–6 of the insular cortex. In the comparison ENR/CTR only layers 2–3 of the visual cortex showed a significant difference, with more cells under enriched conditions. Other comparisons were not significant. Taken together, there was no simple picture in the cellular response of cortical proliferating cells and their progeny to the conditions of ENR and RUN. The regions in which a significant increase on the level of BrdU counts was found were selected for further phenotypic analysis.
Regionally Specific Distribution of Phenotypes of the Newly Generated Cells
Multipotent neural stem or progenitor cells can be isolated from the adult rodent cortex that give rise to neurons, astrocytes and oligodendrocytes (Palmer et al., 1997). Thus we investigated whether the BrdU-labeled cells would acquire a neuronal, astrocytic or oligodendrocytic phenotype. At 4 weeks we did not find any signs of cortical neurogenesis. In the cortical regions investigated there were no BrdU/NeuN double-labeled cell as there were in the hippocampus (Fig. 2). As all BrdU-counts were done at the fluorescence microscope with NeuN as neuronal marker to identify the cortical regions, thousands of cells were screened, both at 1 day and 4 weeks. Additionally, identification of astrocytic differentiation was done at the confocal microscope and in sections stained for BrdU, NeuN and S100β at both time-points. In this analysis, too, no new neurons (BrdU/NeuN double-labeled cells) were found in the cortex.
Astrocytes were identified by their expression of S100β. Not surprisingly, astrocytes accounted for a large percentage of BrdU-labeled cells at 4 weeks. Only in layer 1 of the motor cortex, however, did the effects of RUN and ENR on the numbers of BrdU-labeled cells translate into a net increase in astrogenesis (Fig. 2). The post hoc analysis was only significant for the comparison ENR/CTR; but for RUN/CTR, P = 0.0608. An inverse analysis revealed that in all other regions with significant changes in the number of BrdU-positive cells, these changes were due to effects on BrdU+/S100β–cells.
Oligodendrocytes were identified by their expression of CNP (2′3′-cyclic nucleotide 3′-phosphodiesterase) (Trapp et al., 1988), as shown in Fig. 3H. Additionally, we stained with markers APC (Bhat et al., 1996) or RIP (Friedman et al., 1989), which gave more ambiguous staining (not shown). With all three markers, only very few BrdU-labeled oligodendrocytes could be found. Neither RUN nor ENR had an obvious effect on this number. Due to the very low absolute count, a small activity-induced change might have been missed, but new oligodendrocytes certainly did not explain the activity- and experience-dependent effect on S100β-negative cells. Consequently, we considered a cell-type outside the neuroectodermal lineage to be responsible for the observed changes in BrdU-labeled cells.
Voluntary Wheel Running Increases the Number of New Microglia in Specific Cortical Regions
Labeling microglia with antibody Iba-1 revealed that the region-specific increases in the number of BrdU-marked cells in RUN were due to significant increases in microglia. In several cortical regions investigated, BrdU-labeled microglia were significantly more frequent in RUN than in CTR (Fig. 2). This effect was not seen under ENR conditions. Morphologically, the BrdU-labeled microglia appeared ramified and did not show signs of hypertrophy or even macrophagic transformation (Fig. 3A–F). BrdU-labeled chromatin also showed the characteristic intranuclear distribution, with heterochromatin close to the nuclear membrane (Fig. 3C). Co-staining with a fluorescent nuclear dye that binds to histones (TO-PRO3) confirmed that the BrdU-immunoreaction was seen in the nucleus of the microglial cells and that the labeled microglia had only one nucleus (Fig. 3F). Morphologically, microglia in RUN, BrdU-labeled or not, did not look different from microglia in CTR (and ENR). As further clue, whether local microglia were dividing, one mouse was examined at 2 h after a single injection of BrdU. Here, too, we found BrdU-marked microglia in the cortex.
The present data indicate that in the adult brain, voluntary wheel running increased the number of BrdU-marked microglia.
In contrast, neither environmental enrichment nor wheel running resulted in overall differences in the number of cortical BrdU-labeled cells as compared to controls. This suggests that the effect found on microglia was specific with regard to both the cell type and, to some degree, also the localization. Only in the motor cortex did we see an effect of environmental enrichment on macroglia. Due to the huge number of cells to be examined, phenotypic analysis had to be limited to those conditions under which differences on the level of BrdU-labeled cells had been found. It is therefore conceivable that RUN or ENR might have net effects on macro- or microglia in additional regions, for example by means of a relative redistribution between the phenotypes of the newly generated cells. In hippocampal neurogenesis of aged mice living in an enriched environment, we had seen such a phenotypic shift (Kempermann et al., 2002). To further examine potential changes in the number of new microglia in regions without significant change in the number of BrdU-labeled cells, new experiments tailored to the individual regions will be necessary. The same applies for the frontal cortex, which could not be studied quantitatively here, because too few sections covered the region for us to come to valid results.
In vitro, neural stem or progenitor cells isolated from the adult brain have been shown to generate neurons, astrocytes and oligodendrocytes, but not microglia (Palmer et al., 1997). However, the question of whether new microglia can arise from neural stem or progenitor cells resident in the adult brain has not been finally answered. Present evidence indicates that mature microglia can proliferate and undergo a slow turnover in the adult brain (Lawson et al., 1992). In addition, microglia are recruited from circulating monocytes and under normal conditions both processes combined lead to a steady state (Lawson et al., 1992).
The number of BrdU-marked microglia increased between the two time-points of investigation. This might be indicative of a slow continued proliferation, with so few divisions that BrdU was not diluted beyond the threshold of detection. For the hippocampus of C57Bl/6 mice, Hayes and Nowakowski have calculated that BrdU-labeled cells will detectable for about three cell cycles (Hayes and Nowakowski, 2002). A similar approximation for cortical cells would require knowledge of cell cycle parameters, most notably the lengths of the s-phase and of the cell cycle. Such data are not yet available.
Alternatively, or additionally, the increase in BrdU-labeled microglia could reflect a inner-parenchymal redistribution or the invasion of blood-borne microglia (Eglitis and Mezey, 1997; Priller et al., 2001). Although BrdU-labeled pairs of microglial cells and the presence of BrdU-marked microglia immediately after the BrdU-injection are indicative of a proliferation in loco, some of the microglial cells might have incorporated BrdU elsewhere and migrated in. Only complex experiments, for example with bone marrow chimeras (Priller et al., 2001), will allow this problem to be addressed. The main conclusion of our present study, however, is independent of this consideration: wheel running induced the number of new microglia in selected cortical regions. Increased microglial division did not occur uniformly, but preferred the superficial cortical layers. This increase was substantial, particularly in the visual cortex and comparisons reached high levels of significance.
As microglia have phagocytotic properties, it was necessary to establish that BrdU-incorporation into Iba-1 positive cells did not reflect phagocytosis of BrdU-marked other cells, but cell division of microglia. However, BrdU-labeled microglia (i) generally had the classical ramified morphology, which is not the established phagocytotic state (Fig. 3A–D); (ii) were often found in pairs, suggesting a local cell division (Fig. 3B); (iii) could be detected as early as 2 h after BrdU, which is not compatible with the assumed kinetics of phagocytotic microglial transformation (not shown); and (iv) had only one nucleus, which was the one labeled for BrdU (Fig. 3F).
Microglial function in a normal, non-activated state is largely unknown. We found that under conditions that are not generally accepted to qualify as ‘activating’ (as are trauma or ischemia, etc.), an increased number of microglia could be found in several cortical regions. Although there are some previous examples of such ‘pathology without pathology’ involving microglia (Gehrmann et al., 1993), effects of an entirely physiological, voluntary behavior on microglia have not been reported before. This increase in dividing microglia occurred in the absence of macroglial activation in the same area and in the absence of any sign of brain pathology. Also, by morphological criteria most BrdU-labeled cells qualified as ramified microglia, indicative of a resting state (Gehrmann and Kreutzberg, 1995). However, there is evidence that microglia can react to even minor changes in ion homeostasis in vivo (Gehrmann et al., 1993). The Iba-1 antibody is thought to recognize an activated form of microglia, because the microglia-specific calcium-binding protein Iba-1 is involved in the Rac signalling pathway, the key molecule in microglial activation (Imai and Kohsaka, 2002). Iba-1, however, seems to be active up-stream of Rac (Imai and Kohsaka, 2002), suggesting that antibodies against Iba-1 might recognize microglia in their mildest levels of activation.
It is tempting to link these effects on microglia with the beneficial effects of ‘activity’ on neurological health (Laurin et al., 2001; Vaillant and Mukamal, 2001), because microglia are the immunocompetent cells of the brain, involved in numerous protective and responsive immunological effects (Kreutzberg, 1996). Microglia are also considered to support neuronal function by mediating neurotrophic factors in the uninjured brain (Elkabes et al., 1996). Among these neurotrophins, for example, nerve growth factor (NGF) and brain-derived neurotrophic factor (BDNF) are up-regulated by physical activity. Coincidentally, the activity-dependent up-regulation of BDNF was most prominent in layers 1–3 of the caudal neocortex (Neeper et al., 1996). Growth factors might also be involved in mediating the effects of physical activity on cortical microglial proliferation. Running-induced effects on hippocampal cell proliferation and on protection from various types of experimental brain damage seem to depend on insulin-like growth factor 1 (IGF-1) as mediator (Carro et al., 2000, 2001). At least after ischemic brain injury, IGF-1 might act as a mitogen for microglia (O’Donnell et al., 2002).
According to current concepts, microglia are of mesodermal origin. There is no clear evidence of a common progenitor cell for microglia, macroglia and neurons in the adult brain. Thus, our finding of increased microglial proliferation cannot be due to a stimulation of the multipotent cortical neural progenitor cells (Palmer et al., 1999). Consequently, both wheel running and environmental enrichment seemed to have very limited influence on the neural stem or progenitor cells in the neo-cortex, although they robustly affected hippocampal progenitor cells (Kempermann et al., 1997b; Van Praag et al., 1999b).
A possible exception to this conclusion is the increase in astrogenesis found in layer 1 of the motor cortex. This effect was significant and robust, but locally restricted. Activation of astrocytes and induction of gliogenesis in the adult brain is mostly regarded as ‘reactive gliosis’ in response to damage and less as a normal cellular turnover or addition. Earlier researchers on environmental enrichment have discussed changes in glial cell numbers (Altman and Das, 1964; Diamond et al., 1966). From our data, it seems that generalizing claims of experience- or activity-induced gliogenesis cannot be confirmed.
In an apparent contradiction to our finding of only a limited production of new oligodendrocytes, Levison et al. have reported that the cycling cells of the adult rat neocortex would primarily generate oligodendrocytes (Levison et al., 1999). However, in that study progenitor cells were retrovirally labeled at postnatal day 2 or 3 and only the progeny from these progenitors was assessed at later time points, up to the age of 8 months. Thus, besides a potential species difference, the Levison study and our experiment might have investigated different populations of cells.
Importantly, we did not detect adult cortical neurogenesis under our experimental conditions, thus confirming an earlier description by Magavi et al., who did not find new neurons in the cerebral cortex of adult mice (Magavi et al., 2000). This is also in accordance with studies by Kornack and Rakic (Kornack and Rakic, 2001) and Koketsu et al. (Koketsu et al., 2003) in non-human primates. Gould et al., in contrast, had reported large numbers of new neurons in the brain of monkeys (Gould et al., 1999), but later concluded that the majority of these neurons were of transient existence (Gould et al., 2001). Several review papers and comments have discussed the methodological aspects of whether or not adult cortical neurogenesis would occur under physiological conditions in primates (Nowakowski and Hayes, 2000; Grassi Zucconi and Giuditta, 2002; Rakic, 2002). In their 2001 paper, Gould et al. also showed neurogenesis in the anterior cortex of rats, although they did not quantify it (Gould et al., 2001). For the time being, it remains open as to whether rats differ from mice with regard to adult cortical neurogenesis, whether adult cortical neurogenesis might occur only in the rodent frontal cortex (which was not included in our experiment, because not enough tissue was available to cover this region adequately), or whether the discrepancy could indeed be explained on the basis of methodological issues, especially the concentration of BrdU (Cameron and McKay, 1999; Gould et al., 2001).
In general, recent evidence indicates that under influence of a pathological stimulus some adult neurogenesis might be possible in otherwise non-neurogenic regions (Magavi et al., 2000; Arvidsson et al., 2002; Nakatomi et al., 2002). We considered it conceivable that a similar neurogenic stimulus might also be exerted by non-pathogenic yet activating conditions. But in the present study we show that cortical neurogenesis was not induced by two physiological stimuli that at the same time robustly enhanced hippocampal neurogenesis.
Taken together, our study brings microglia into a context that so far has been dominated by the consideration of the other two ‘elements’ of the brain, neurons and astrocytes. Activity-dependent cortical plasticity is far from understood. The finding of regionally specific effects that qualitatively and quantitatively depend on the type of the physiological stimulus and differentially affect different types of brain cells certainly does not simplify the problem.
We would like to thank Christiana Cooper-Kuhn and H. Georg Kuhn for generous support of this study, Y. Imai for the Iba-1 antibody and Josef Priller for comments on the manuscript. The research reported here was financed by grants from Deutsche Forschungsgemeinschaft (DFG) and VolkswagenStiftung.
Address correspondence to Gerd Kempermann, Max-Delbrück-Center for Molecular Medicine (MDC), Robert-Rössle-Str. 10, 13125 Berlin, Germany. Email: firstname.lastname@example.org.