Research Article

Cell Proliferation Without Neurogenesis in Adult Primate Neocortex

See allHide authors and affiliations

Science  07 Dec 2001:
Vol. 294, Issue 5549, pp. 2127-2130
DOI: 10.1126/science.1065467


A recent assertion that new neurons are continually added to the neocortex of adult macaque monkeys has profound implications for understanding the cellular mechanisms of higher cognitive functions. Here we searched for neurogenesis in adult macaques by using immunofluorescent triple labeling for the DNA-replication indicator, bromodeoxyuridine (BrdU), and neuronal and glial cell markers. Although numerous BrdU-labeled cells were distributed throughout the cerebral wall, including the neocortex, these were identified as nonneuronal cells; evidence for newly generated neurons was limited to the hippocampus and olfactory bulb. Thus, our results do not substantiate the claim of neurogenesis in normal adult primate neocortex.

Higher cognitive functions in primates, including humans, depend on the appropriate number, organization, and connectivity of neurons in the association areas of the neocortex (1, 2). Studies with 3H-thymidine (3H-TdR) autoradiography, which labels DNA in dividing cells and their progeny, have indicated that neocortical neurons are generated before or shortly after birth in all species examined (3–10). In macaque monkeys, cortical neurogenesis occurs in a strict inside-to-outside (deep to superficial) laminar order before birth (7), whereas gliogenesis—in macaques as in other species—continues postnatally (11–13).

More recently, the generation of cells in the nervous system has been studied with the thymidine analog, 5-bromodeoxyuridine (BrdU), which also labels DNA during cell division (14). This method has been combined with cell-type–specific immunomarkers and confocal microscopic imaging to detect newly generated neurons in the adult hippocampus and olfactory bulb of rodents (15–17) and primates (18–21). With this approach, it was recently reported that considerable numbers of new neurons are continuously added to neocortical association areas during adulthood in macaque monkeys (22). The report further indicated that these neurons are produced in the subventricular zone (SVZ) of the lateral cerebral ventricles and then migrate in streams through the subcortical white matter to the prefrontal, posterior parietal and inferior temporal neocortex. Moreover, after reaching the neocortex within 2 weeks after their generation, the new neurons were reported to extend axons to other cortical areas (22) and then die within the next 7 weeks (23,24).

Because of the considerable conceptual and biomedical implications of this claim, it is essential to validate the reliability and robustness of this putative phenomenon. Accordingly, we examined the proliferation and phenotypic differentiation of cells in the cerebrum of adult macaque monkeys at various times after BrdU injections by using triple-label immunofluorescence for BrdU, glial fibrillary acidic protein (GFAP, a marker for astrocytes), and either NeuN (a marker for adult neurons) or class III β-tubulin (TuJ1, a marker for immature neurons) (25). We surveyed the prefrontal cortex—including the principal sulcus, where the majority of adult-generated neurons were reported—by examining every third section in a sequential series of coronal sections. The sections were first scanned with conventional epifluorescence microscopy with a dual-band filter set, which allowed simultaneous viewing of the fluorescence signals of BrdU and neuronal markers. All cells suspected of being double-labeled for BrdU and a neuronal marker were examined in detail with three-dimensional confocal analysis.

BrdU-labeled nuclei were observed in the neocortex of each monkey, irrespective of the particular injection schedule. The animals perfused 10 and 23 days after the final BrdU injection were most informative because these survival times correspond to the time point when adult-generated neurons are reportedly present in greatest numbers in the cortex before their demise. We found that many BrdU-labeled cells within the neocortex were distributed as “doublets” and were the likely daughter cells of a mitotic event within the neocortex (Fig. 1). The phenotype of BrdU-labeled cells in the cortex was examined with triple-label immunofluorescence histochemistry for BrdU, NeuN, and GFAP. Although we surveyed more than 1000 BrdU-labeled cells in the prefrontal neocortex in each monkey, in no instance did we find a BrdU-labeled cell that was colabeled with NeuN. This result differs markedly from the previous study (22), which claimed that 1 to 2 weeks after BrdU injections, 38 to 52% of the BrdU-labeled cells in the principal sulcus colabeled for NeuN. Given the number of BrdU-labeled cells that we surveyed in the principal sulcus alone, the probability of missing 38 to 52% [or even the revised, lower estimate of 25% (24)] of the total BrdU-labeled cell population as double-labeled neurons is infinitesimally small. This discrepancy may be due to differences in histological analysis [e.g., (26)] or other artifacts (27,28). For example, upon initial inspection of our material, some BrdU-labeled nuclei appeared to belong to cortical neurons, which would be indicative of adult-generated neurons. However, in such cases, detailed confocal z-series analyses (i.e., examining sequential optical sections at intervals of 0.8 μm in thez axis) revealed that these BrdU-labeled nuclei actually belonged to cells that were closely apposed to neurons but were themselves immunonegative for NeuN (Fig. 2). The nuclei of NeuN-positive neurons were invariably immunonegative for BrdU (Fig. 2). Thus, the BrdU-labeled nuclei in these initially ambiguous cases appear to belong to newly generated satellite glial cells, which have been previously described in the neocortex of adult monkeys with both3H-TdR and BrdU methods (29–31) as well as in the cortex of other mammals (11, 12,26, 32).

Figure 1

Confocal microscopic images of BrdU-labeled cells in the frontal cortex of adult macaque monkeys. BrdU-labeled cells in cortex were often distributed as “doublets” (arrows), as illustrated by these examples in cingulate cortex (A toC) and prefrontal cortex (D) 10 days [(A) to (C)] and 32 days (D) after BrdU injections. BrdU-labeled nuclei (green) are closely associated with the perikarya of NeuN-positive neurons (red) but are not immunopositive for NeuN, indicating a nonneuronal phenotype. GFAP immunoreactivity (blue) indicates the presence of astrocytes. None of the neurons were labeled with BrdU. Scale bar, 20 μm for (A) to (D).

Figure 2

Images of cells in the adult macaque monkey prefrontal cortex that appear to be double-labeled for BrdU and NeuN are instead resolved to be two closely apposed single-labeled cells. (A) A NeuN-positive neuron (red) appeared to be colabeled with BrdU (green) in a merged image. However, a z-series analysis (B to E) revealed that the BrdU-labeled nucleus (arrow) was located in a different focal plane than the NeuN-positive neuronal cell body. The BrdU-positive nucleus is visible in (B) and (C), and the NeuN-positive nucleus and nucleolus are visible in (D) and (E) (arrowhead), demonstrating that the neuron is not BrdU-labeled. (F to J) An example of a pyramidal neuron that appears to be BrdU-labeled in sections [(F) to (H), arrow] is shown to have its nucleus and nucleolus in a different focal plane [(I) and (J), arrowhead]. These BrdU-labeled nuclei apparently belong to adult-generated satellite glia, which typically associate with neuronal perikarya (see Fig. 1). Astrocytes are indicated by GFAP immunoreactivity (blue). Scale bars, 20 μm.

Gould et al. have proposed that most of the putative new neurons in the monkey neocortex die within 9 weeks after being generated and that the previous 3H-TdR autoradiography studies failed to detect them because of the longer survival periods used in these studies (22–24). However, in the present analysis, new cortical neurons were not detected even in the animals with short survival intervals. Even if, because of the sampling procedure or technical limitations, we have missed some BrdU-labeled neurons, their number would be exceedingly small. It would also be essential to exclude BrdU labeling resulting from DNA synthesis in response to cell damage, abortive mitoses, or initial steps in cell death, all of which could occur without mitotic division (33–35), before concluding that the labeled cells are newly generated neurons. Our results are in harmony with previous3H-TdR autoradiographic studies in both primate and nonprimate species, indicating that neurogenesis of the neocortex is normally confined to developmental periods (3–10,12, 13). Consistent with our findings, a recent study with immunofluorescent double-labeling of cortical cells with BrdU and NeuN in adult mice also failed to detect neurogenesis under normal conditions (36).

Finally, the negative findings in the present study cannot be attributed to inadequate histological processing because we could positively identify adult-generated neurons colabeled for BrdU and NeuN in the hippocampal dentate gyrus and olfactory bulb in the very animals examined for cortical neurogenesis (Fig. 3) (20, 21). Moreover, we could identify immature BrdU/TuJ1 double-stained neuroblasts in the SVZ and olfactory tract that migrate to the olfactory bulb (see Fig. 3) (21) as they do in other mammalian species (17, 37, 38). However, we did not detect any migrating neuroblasts entering the overlying subcortical white matter, by using double-staining for BrdU and TuJ1 (39). We did observe BrdU-labeled cells in the subcortical white matter, but these were not arrayed in the migratory stream from the SVZ to the prefrontal principal sulcus as diagrammed in Fig. 4 in Gould et al. (22). Rather, in our material, these cells were distributed throughout the white matter along blood vessels or along myelinated fiber tracts, typical of newly generated endothelial cells and oligodendrocytes, respectively, which have also been observed in numerous previous studies (11–13, 28–32, 40–46). The aligned rows of cells identified as chains of migratory neurons in Fig. 5D of Gould et al. (22) and Fig. 4C of their subsequent paper (24) appear to be endothelial cells lining a longitudinally cut capillary. The cytological features used previously to identify BrdU-labeled cells as migrating neurons in the adult macaque [i.e., bipolar morphology and colabeling for CRMP-4 (collapsin response-mediated protein 4; also named TUC-4/Ulip1, formerly “TOAD-64”)] (22) are also exhibited by cells in the oligodendrocyte lineage (43, 44). Thus, these criteria cannot be used to identify BrdU-labeled cells in the white matter as migrating neurons. A more detailed discussion of the caveats of the methods used, as well as a critique of the specific aspects of the previous studies, is provided elsewhere (27,28).

Figure 3

Newly generated cells with neuronal characteristics in the adult macaque monkey brain. (A and B) Immature neurons expressing TuJ1 (red) are distributed in the SVZ along the lateral and ventrobasal walls of the lateral cerebral ventricle. BrdU-labeled nuclei (greenish-yellow; arrows) indicate that these are newly generated cells. (A) Immature morphology and the colocalization of BrdU within the nuclei of TuJ1-labeled neuroblasts. (B) Neuroblasts in the SVZ coalesce into migratory chains that extend via the rostral migratory stream to the olfactory bulb. Migrating neuroblasts were not observed in the subcortical white matter. Ependymal and astrocytic cells are indicated by GFAP (blue). LV, lateral ventricle. (C and D) An example of a newly generated granule cell in the olfactory bulb that is colabeled for NeuN [arrow, red in (C)] and BrdU [arrow, yellow-green in (D)]. A z-stack series through this cell (56) revealed that the BrdU signal is confined within and coextensive with the NeuN-labeled nucleus of the cell. Scale bars, 20 μm.

The present findings lead to the conclusion that neocortical neurons are not normally renewed during the life-span of macaque monkeys, which can last three decades (13); similar limits may exist in the human forebrain (47). Although preservation of neurons during one's life-span is considered important for the storage of memory and life-long experience (13), the decreased capacity for neurogenesis in adulthood may be an impediment when the need arises for replacement of lost neurons in trauma and neurodegenerative disorders (48). The challenge in the coming years will be to determine how to replace lost neurons in brain regions where they are normally not renewed and how to incorporate them into an adult brain environment to restore lost function (36, 37, 49–54).

  • * To whom correspondence should be addressed. E-mail: pasko.rakic{at}


View Abstract

Navigate This Article