Segregation of Human Neural Stem Cells in the Developing Primate Forebrain

See allHide authors and affiliations

Science  07 Sep 2001:
Vol. 293, Issue 5536, pp. 1820-1824
DOI: 10.1126/science.1060580


Many central nervous system regions at all stages of life contain neural stem cells (NSCs). We explored how these disparate NSC pools might emerge. A traceable clone of human NSCs was implanted intraventricularly to allow its integration into cerebral germinal zones of Old World monkey fetuses. The NSCs distributed into two subpopulations: One contributed to corticogenesis by migrating along radial glia to temporally appropriate layers of the cortical plate and differentiating into lamina-appropriate neurons or glia; the other remained undifferentiated and contributed to a secondary germinal zone (the subventricular zone) with occasional members interspersed throughout brain parenchyma. An early neurogenetic program allocates the progeny of NSCs either immediately for organogenesis or to undifferentiated pools for later use in the “postdevelopmental” brain.

As cells with stemlike qualities have come to be identified within a widening range of organs [e.g., (1, 2)], new questions have arisen about their relevance to normal development. The central nervous system (CNS) may serve as a bellwether for insights in this field. NSCs have been identified in the mammalian CNS, including humans (3–9), at stages from fetus to adult in a surprisingly wide range of regions (10–13). NSCs, defined as self-renewing, propagatable primordial cells each with the capacity to give rise to differentiated progeny within all neural lineages in all regions of the neuraxis, are posited to exist in the embryonic and fetal ventricular germinal zone (VZ) where they participate in CNS organogenesis (5, 14, 15). Cells equally “stemlike” in their potential have been identified at later stages (including old age) from a variety of regions: subventricular (SVZ) (13–17) and ependymal (18) zones of the forebrain, subgranular zone of the hippocampus (6–10, 19), retina (20) and optic nerve (10, 11), cerebellum (12), spinal cord (21), and even cortical parenchyma (10, 15, 22). How might these observations be reconciled? Are such stemlike pools, particularly those isolated from various parenchymal regions at “postdevelopmental” periods, of physiological relevance or artifacts of experimental manipulation (10, 11)? Do these populations represent the same lineage or unique pools (17)? Of what relevance are these cells to normal human CNS development and repair?

We hypothesized that multiple stem cell pools, descendants of a common NSC, emerge during early cerebrogenesis as cells are used in organogenesis and concurrently also set aside to establish a reservoir for subsequent use in homeostasis and repair. This could represent a developmental strategy in which plasticity is programmed into the CNS at the single-cell level from early stages of embryogenesis.

We sought to determine how progeny of a single traceable clone of NSCs get segregated during development by using a system that might also lend insight into human development. We grafted a clone of NSCs of human derivation (5, 23) into the developing brains of fetal bonnet monkeys (Macaca radiata), an Old World species (Web note 1) (24). We asked what the fate would be of human cells transplanted at a time when neocortical cell genesis, migration, and differentiation are intensive (25–27). The primate neocortex, at the appropriate developmental stage, allows a distinction between layers of active neuron birth and layers where neurogenesis has been completed and glial cells are instead acquired (27) (Web note 2) (24) (Fig. 1, schematics I and II). One can discern experimentally the responses to local developmental cues simply by assaying the spatial segregation and patterns of differentiation of NSCs of a single clone in a given animal's brain after a single transplantation procedure. [A summary of simian cortical development is provided in an expanded legend to Fig. 1 in Web note 2 (24)]. Under transabdominal ultrasonic guidance, bonnet monkey fetuses at 12 to 13 weeks gestation received a single in utero injection of ∼2 × 107 clonally related undifferentiated NSCs [prelabeled with the nuclear marker 5-bromo-2′-deoxyuridine (BrdU)] into the left lateral cerebral ventricle, allowing the cells access to the VZ from which the cerebral cortex is derived (23). [At 12 to 13 weeks, VZ cells normally cease giving rise to the neurons in layers IV to VI and begin contributing to neurogenesis in layers II and III (27) ( Fig. 1, schematic I).] Pregnancy was allowed to continue to the completion of most cortical neurogenesis at ∼16 to 17 weeks gestation (Fig. 1, schematic II), when the fetuses were delivered by Cesarean section and their brains were processed for histological analysis (28) (Fig. 2). Distribution of donor human NSCs (hNSCs) in the monkey brains was monitored by immunocytochemical staining for the BrdU marker (Figs. 1 and3) (28). To provide further independent confirmation of the cells' origin, we used, in parallel, antibodies against additional donor-specific markers, including the human-specific nuclear mitotic antigen (NuMA) as well as other species-specific tags (28). The phenotypes of these cells were characterized by immunocytochemistry (28) (Fig. 3).

Figure 1

Clonal hNSCs migrate from ventricular germinal zone (VZ) into developing neocortex. On the left are schematics of the developing monkey neopallium at the time of transplantation (I) [12 to 13 weeks postcoitum (pc)] and at the time of death (II) (16 to 17 weeks pc). See (27) and Web note 2 for brief description of primate neocortical development. Transplantation is described in (23). (A toC) Photomicrographs from selected locations spanning the neopallium. (Their location relative to the schematic is indicated by brackets.) (A) Injected into the left lateral ventricle and having integrated throughout the VZ, the hNSC-derived cells (d), identified by their BrdU immunoreactivity (black nuclei), migrated along the monkey's radial glial processes [visualized with an antibody to vimentin (brown)] through the neopallial wall to reach their temporally appropriate destination in the nascent superficial layers II and III (A), where they detached from the radial glia and took up residence as neurons (see Fig. 3 for closeups and characterization). Arrows indicate climbing cells (both donor- and host-derived) positioned along the processes of the vimentin-positive host radial glia. Some cells (inset) are pictured still attached to these fibers and in the process of migration. The photomicrographs in (B) and (C) show examples of immature, donor hNSC-derived (BrdU-positive black nuclei, d) astrocytes (brown vimentin-positive immunostain) intermixed with host-derived astrocytes in deeper cortical lamina, having differentiated as expected for that site and time. Abbreviations: MZ, marginal zone; CP, cortical plate; SP, subplate; WM, white matter; II to VI, cortical layers. Scale bar, 35 μm.

Figure 2

Evaluation of engrafted monkey brains at the conclusion of neocortical development. Brains were retrieved as described in (23) and analyzed in three anterior-posterior segments (I to III), as illustrated. For histological methods, see (28).

Figure 3

Segregation of the fates of hNSCs and their progeny into two subpopulations in the brains of developing Old World monkeys. Schematics (left) and photomicrographs (right) illustrating the distribution and properties of clonal hNSC-derived cells. [Each coronal section in the schematic (I to III) corresponds to a coronal level (I to III) in Fig. 2.] hNSCs [labeled with BrdU and implanted as per (23)] dispersed throughout and integrated into the VZ. From there, clonally related hNSC-derived cells pursued one of two fates, as shown by immunocytochemical analysis (A to I) (28). Those donor cells that migrated outward from the VZ along radial glial fibers (as per Fig. 1) into the developing neocortex constituted one pool or subpopulation. The differentiated phenotypes of cells in this subpopulation 1 (red stars in the schematic) (particularly in layers II and III) are pictured in panels (A) to (G). (A) An hNSC-derived BrdU-positive cell (black nucleus, arrow)—likely a neuron according to its size, morphology, large nucleus, and location—is visualized (under Nomarski optics) intermingled with the monkey's own similar neurons (arrowheads) in neocortical layers II and III. The neuronal identity of such donor-derived cells is confirmed by immunocytochemical analysis in (B) to (D). (B, C, and E to G) High-power photomicrographs of human donor-derived cells integrated into the monkey cortex double-stained with antibodies against BrdU and cell type–specific markers: (B) NeuN and (C) calbindin for neurons (arrows, donor-derived cells; arrowheads, host-derived cells). (E) CNPase for oligodendroglia (arrow, BrdU-positive black nucleus in CNPase-positive brown cell; arrowhead indicates long process emanating from the soma). (F and G) GFAP for astroglia [antibody to Brd U revealed via fluorescein in (F); antibody to GFAP revealed via Texas Red in (G)]. The human origin of the cortical neurons is further independently confirmed in (D) where the human-specific nuclear marker NuMA (black nucleus) is colocalized in the same cell with neurofilament (NF) immunoreactivity (brown). Progeny from this same hNSC clone were also allocated to a second cellular pool—subpopulation 2 [blue dots in the schematic and pictured in (H) and (I) (arrows)]—that remained mainly confined to the SVZ and stained only for an immature neural marker [vimentin (brown) colocalized with BrdU (black nucleus) better visualized in inset (arrows); arrowhead indicates host vimentin-positive cell]. Some members of subpopulation 2 were identified within the developing neocortex (blue dots) intermixed with differentiated cells. (F) and (G) use immunofluorescence; the other immunostains use a DAB-based color reaction. The photomicrographs were taken from different animals as representative of all animals. ve, lateral cerebral ventricle; arrow, BrdU-positive donor-derived cell; arrowhead, BrdU-negative, host-derived cell except in (E). Scale bars, 30 μm [(A) to (C)]; 20 μm [(D) to (I)].

Unilaterally injected hNSCs distributed themselves throughout both cerebral hemispheres symmetrically and at most levels of the neuraxis, settling in diverse widespread regions of the telencephalon, principally at the frontal and frontoparietal levels (Fig. 3). Although the individual hNSCs were clonally related, they appeared to segregate into two subpopulations (Fig. 3), as follows.

Cells in subpopulation 1 (red stars in Fig. 3) appeared to traverse great distances (∼1.6 cm or ∼1600 times a migrating cell body diameter) from the periventricular germinal zones along host radial glial processes (Fig. 1A) to terminate at developmentally and temporally appropriate cortical laminae and differentiate into several neuronal (Fig. 3, A to D) and glial (Fig. 3, E to G) cell types. Those hNSCs that migrated to the superficial neurogenic cortical layers II and III (Fig. 1A, schematic II) appropriately became neurons (Fig. 3A, arrow), identified by dual immunoreactivity to antibodies to NeuN, calbindin, and neurofilament (Fig. 3, B to D, arrows), intermixed with the monkey's own neurons (arrowheads). The majority of the hNSC-derived neurons were found in cortical layers II and III [which, at the time of transplant, profited from an intensive supply of newly formed neurons (27, 29)]. Those hNSC-derived cells that stopped and integrated within the deeper cortical layers IV to VI differentiated appropriately into glial cells (Fig. 1, B and C, schematic II), identified by immunoreactivity to glial fibrillary acidic protein (GFAP) (for astrocytes) or to 2′,3′-cyclic nucleotide 3′-phosphohydrolase (CNPase) (for oligodendrocytes) (Fig. 3, E to G). [Glial cells of donor origin were also appropriately observed in the marginal zone (MZ, layer I) (Fig. 1, schematic II) and in subcortical regions. Some donor cells contributed also to the radial glial cell population.]

Cells in subpopulation 2 (blue dots in Fig. 3) were small, undifferentiated BrdU-positive cells lacking neuronal processes and were dispersed throughout the SVZ as single cells or small clusters intermingled with the germinal cells of the host (Fig. 3, H and I). When double-stained for cell type–specific antigens, these cells expressed vimentin (an immature progenitor/stem cell marker) (Fig. 3I and inset) but were negative for all other markers of differentiation. The majority of such undifferentiated hNSC-derived cells remained within the SVZ [none in the ependyma (18)]. The SVZ has been implicated in postnatal and adult homeostatic mechanisms (16, 17, 30, 31) and as an ongoing source of cortical neurons after overt cortical development has ceased (32–34). A small number of subpopulation 2 cells, however, were present within the striatum and cortex, intermixed with the differentiated cells (Fig. 3). These cells may provide a local resident pool for self-repair and plasticity and may represent the stemlike cells extracted by several investigators (10, 13, 15, 22,35). [This observation favors the interpretation that such reported cells are not simply the result of dedifferentiation of committed progenitors, an artifact of experimental manipulation, as has occasionally been speculated (10, 11)].

Our data provide a plausible dynamic for how multiple, disparate stem cell populations are generated as part of a single strategy of NSC allocation. The clonal progeny of a given NSC segregate to yield some differentiated cells for organogenesis (e.g., subpopulation 1) and other cells (e.g., subpopulation 2) for deposition in secondary germinal zones (e.g., the SVZ) as a reservoir. The NSCs that have been isolated from adults are likely descendants of the same NSCs that contributed to embryonic and fetal CNS development and thus do not represent a unique pool. In this view, ongoing lifelong self-repair and plasticity are a fundamental developmental program set in place during early stages of brain organogenesis. Grafted hNSCs appear to become integrated into the morphogenetic program of the developing primate host brain (Figs. 1 and 3) (36). Although it was not technically possible in these monkeys to quantify rigorously the percentage of grafted cells that survived, the histological images show that a large number of donor-derived cells were present bilaterally in all recipients (37, 38). That hNSCs can migrate through the large expanse of the primate cerebrum, not merely through the much smaller rodent brain (5–7), suggests that migration may be a fundamental stem cell property limited only by available terrain (large or small). In rodents, NSCs have been shown to be well-suited for transplant-based approaches to gene therapy and/or cell replacement in diseases characterized by extensive or global abnormalities (39). Our results suggest that this approach may similarly be feasible in large primates and possibly humans.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed at present address: Beth Israel Deaconess Medical Center, Department of Neurology, 855 Harvard Institute of Medicine, 77 Avenue Louis Pasteur, Boston, MA 02115, USA. E-mail: esnyder1{at} or vouredni{at}

  • Co-senior authors.


View Abstract

Stay Connected to Science

Navigate This Article