Abstract
In this study we investigated whether the pattern of expression of the cyclin-dependent kinase inhibitor p19INK4dby the unique progenitor cells of the neonatal anterior subventricular zone (SVZa) can account for their ability to divide even though they express phenotypic characteristics of differentiated neurons. p19INK4d was chosen for analysis because it usually acts to block permanently the cell cycle at the G1phase. p19INK4d immunoreactivity and the incorporation of bromodeoxyuridine (BrdU) by SVZa cells were compared with that of the more typical progenitor cells of the prenatal telencephalic ventricular zone. In the developing telencephalon, p19INK4d is expressed by postmitotic cells and has a characteristic perinuclear distribution depending on the laminar position and state of differentiation of a cell. Moreover, the laminar-specific staining of the developing cerebral cortex revealed that the ventricular zone (VZ) is divided into p19INK4d(+) and p19INK4d(−) sublaminae, indicating that the VZ has a previously unrecognized level of functional organization. Furthermore, the rostral migratory stream, traversed by the SVZa-derived cells, exhibits an anteriorhigh–posteriorlowgradient of p19INK4d expression. On the basis of the p19INK4d immunoreactivity and BrdU incorporation, SVZa-derived cells appear to exit and reenter the cell cycle successively. Thus, in contrast to telencephalic VZ cells, SVZa cells continue to undergo multiple rounds of division and differentiation before becoming postmitotic.
- cyclin-dependent kinase inhibitors
- p19INK4d
- progenitor cells
- proliferation
- subventricular zone
- ventricular zone
During development the proliferating cells within the ventricular and subventricular zones that line the lateral ventricles give rise to the neurons of the mammalian forebrain. From tritiated thymidine and bromodeoxyuridine (BrdU) birth-dating studies and the staining patterns of cell type-specific markers (Brand and Rakic, 1979; Menezes and Luskin, 1994; Kornack and Rakic, 1998) current thinking suggests that the progenitor cells of the telencephalic ventricular zone (VZ) give rise to cells that exit the mitotic cycle at the ventricular surface and then undergo differentiation and migration (Rakic, 1974; Bayer et al., 1991). The specialized neuronal progenitor cells located within the anterior part of the subventricular zone of the postnatal forebrain (SVZa) (Luskin, 1993) constitute the major exception to this orderly process, by which proliferation precedes differentiation and migration (Menezes et al., 1995). In particular, as the progeny of the SVZa cells migrate to the olfactory bulb along a restricted pathway, the rostral migratory stream (RMS), they concurrently undergo division, despite expressing a neuronal phenotype. Thus, unlike the cells of the VZ, SVZa-derived cells initiate differentiation before becoming postmitotic.
An obligatory step for a cell to become postmitotic is considered to be the withdrawal from the cell cycle at the G1phase, before reentry into the S phase (Sherr, 1994). Progression from G1 to S phase is negatively regulated by two families of cyclin-dependent kinase inhibitors (CDKIs), the CIP/KIP and INK4 families (Hirai et al., 1995; Weinberg, 1995; Sherr and Roberts, 1999). An analysis of the INK4 family consisting of p15INK4b, p16INK4a, p18INK4c, and p19INK4d has revealed that they exhibit differential expression patterns. In particular, p18INK4c and p19INK4d are expressed during neurogenesis, whereas p15INK4b and p16INK4a are absent during development (Zindy et al., 1997a,b). Moreover, p27KIP1, a member of the CIP/KIP family, was shown to be essential for cell cycle exit mechanisms as well as the differentiation of oligodendrocytes (Raff et al., 1998; Tikoo et al., 1998; Casaccia-Bonnefil et al., 1999). The involvement of CIP/KIP proteins in the genesis of neurons has not been elucidated. The INK4 family members, specifically p18INK4c and p19INK4d, however, have been suggested to play a role in the development of both the cerebral cortex (Zindy et al., 1997b, 1999) and cerebellum (Watanabe et al., 1998). Thus, INK4 proteins might couple proliferation arrest to terminal differentiation by cell cycle arrest at G1.
An understanding of how proliferation is inversely coupled to differentiation during cortical development must take into account the cellular kinetics that progenitor cells undergo. As the VZ cells proliferate, their nuclei oscillate in a process referred to as “interkinetic nuclear migration,” such that there is a relationship between the position of the nucleus and the phase of the cell cycle (Takahashi et al., 1996). During G1, the nuclei ascend from the ventricular surface to the basal border of the VZ, whereas in S phase they undergo DNA synthesis (see Fig. 1). In G2, nuclei return to the ventricular surface before entering the M (mitotic) phase. An individual progenitor cell can divide either symmetrically (both daughter cells either become postmitotic or proliferate) or asymmetrically (one daughter cell becomes postmitotic and the other continues to proliferate).
In this study, we have investigated whether the unusual proliferative capacity of SVZa progenitor cells can be explained by a differential regulation of the G1–S progression. Therefore, we have compared the p19INK4d expression pattern of the cells of the postnatal SVZa and RMS with that of telencephalic VZ cells. We demonstrated an anteriorhigh–posteriorlowgradient of p19INK4d expression along the RMS, indicating that few cells withdraw from the cell cycle in the SVZa and increasingly more as the olfactory bulb is approached. In addition, we showed that the VZ is comprised of a predominantly p19INK4d(+) sublamina (apposed to the ventricular surface) and a predominantly p19INK4d(−) sublamina, indicating that VZ can be divided on the basis of its pattern of p19INK4d expression. Furthermore, the intracellular localization of p19INK4dvaries as a function of the laminar position of a cell in the developing cerebral cortex, suggesting that p19INK4d is dynamically regulated.
MATERIALS AND METHODS
BrdU injections and perfusion. To obtain embryos of particular gestational ages as well as neonatal pups for the analysis of p19INK4d expression in the forebrain, Sprague Dawley rats were mated overnight in the colony, which we maintain. The day of conception, detected by the presence of a vaginal plug, was considered to be embryonic day 0 (E0). Birth usually occurred on the 21st or 22nd day of gestation. To normalize the ages of the postnatal animals examined, E22 was considered equivalent to postnatal day 0 (P0).
To label replicating cells in the embryonic telencephalon between E14 and E20, as well as in the P0–P2 SVZa and RMS thymidine analog used as a cell proliferation marker BrdU was administered intraperitoneally (200 mg/kg) to pregnant dams. The BrdU was administered 45 min or 3, 4, or 9 hr before perfusion to visualize labeled cells in different positions of the VZ as they undergo interkinetic nuclear migration. Timed embryos were removed from the uterus of pregnant dams that were deeply anesthetized with chloral hydrate (350 mg/kg) and perfused transcardially with cold 4% paraformaldehyde in 0.1 mPBS, pH 7.4. The P0–P2 pups were also administered BrdU after anesthetization by ether inhalation, 3 or 9 hr before they were again anesthetized and perfused with 4% paraformaldehyde. The perfused brains of both the embryonic and neonatal rats were removed from the skull, post-fixed overnight at 4°C in the same fixative, and cryoprotected with 20% sucrose for 24 hr. All brains were embedded in O.C.T. (Miles, Elkhart, IN), frozen with liquid nitrogen, cut in the sagittal plane on a cryostat at 10 μm, and then mounted onto Superfrost(plus) slides (Fisher Scientific, Houston, TX).
Immunohistochemical procedures. The immunohistochemical detection of p19INK4d, neuron-specific type III β-tubulin (detected by the antibody referred to as TuJ1), and BrdU was performed on the brains of at least three rats examined at E14, E15, E16, E18, E20, P0, and P2. For some analyses, we modified a previously described procedure from Menezes et al. (1995) to double-label sections with anti-p19INK4dand TuJ1 or either antibody alone. Sections were rinsed in 0.1m PBS, pH 7.4, and kept in blocking serum (1.5% normal goat serum and 0.01% Triton X-100 in 0.1 m PBS, pH 7.4) for 1 hr. Subsequently, the sections were incubated overnight at 4°C in anti-p19INK4d (Santa Cruz Biotechnology, Santa Cruz, CA) or TuJ1 (Promega, Madison, WI) at 1:100 and 1:400 dilutions, respectively, or a combination of anti-p19INK4d and TuJ1 with the same final dilutions. The following day the sections were rinsed with PBS and incubated, at room temperature, in the appropriate secondary antibodies or antibody (Jackson ImmunoResearch, West Grove, PA) added to blocking serum at a 1:200 final dilution. For consistency we always used fluorescein-conjugated goat anti-rabbit IgG for detection of the anti-p19INK4d and rhodamine-conjugated goat anti-mouse for visualization of the TuJ1. For double-labeling, cocktails of secondary antibodies were added to blocking serum at 1:200 final dilution and applied to the sections. After a 1 hr incubation, the sections were washed with PBS and coverslipped with Vectashield (Jackson ImmunoResearch).
An adjacent set of sections to those used for the p19INK4d/TuJ1 labeling was used to reveal BrdU immunoreactivity in sections also stained with anti-p19INK4d. For the detection of BrdU-labeled cells, the sections were rinsed in 0.1 m PBS, immersed in 2N hydrochloric acid (HCl) at 40°C for 20 min to denature the DNA, and then rinsed twice with 0.1 m borate buffer, pH 8.4, for 15 min to neutralize the HCl. After several PBS washes, the sections were immersed in blocking serum for 1 hr and then incubated overnight at 4°C in a combination of rat monoclonal anti-BrdU (Accurate Chemicals, Westbury, NY) and anti-p19INK4d added to blocking serum at a 1:400 and 1:100 final dilution, respectively. The next day, the sections were rinsed with PBS and incubated for 1 hr at room temperature in a cocktail of the appropriate secondary antibodies (fluorescein-conjugated goat anti-rabbit IgG for anti-p19INK4d and rhodamine-conjugated goat anti-rat for anti-BrdU). Both secondary antibodies were added to blocking serum at a 1:200 final dilution. After the secondary antibody incubation, sections were washed with PBS and coverslipped with Vectashield. To demonstrate the specificity of labeling, the primary antibodies were omitted in the control sections in all stainings.
To distinguish cytoarchitectural features, chosen sections neighboring those stained for p19INK4d were stained with cresyl violet (CV). Briefly, after incubation in a series of descending alcohol concentrations and then xylene, the sections were stained with CV (0.5%) for 30 min. Sections were examined using aZeiss Axioscope fluorescence microscope or a confocal Zeiss Axioplan equipped with LSM 510. Confocal images were obtained from a single optical section with a thickness of 1 μm. The laser scanning at 488 nm (for fluorescein) and 543 nm (for rhodamine) was performed sequentially to avoid bleed-through of the separate fluorescent signals. The captured images were merged and processed using Adobe Photoshop.
RESULTS
We studied the regulation of the proliferation and differentiation of cells in the embryonic telencephalon and neonatal RMS by analyzing the pattern of expression of the CDKI p19INK4d by the cells of the prenatal and postnatal forebrain. To determine whether and which cells have initiated differentiation, on the basis of their expression of p19INK4d and neuron-specific tubulin, rat embryos and pups were administered the cell proliferation marker BrdU at various times before their perfusion, and subsequently the brains of these animals were immunostained with p19INK4d and anti-BrdU, as well as the neuron-specific antibody TuJ1, which recognizes neuron-specific type III β-tubulin. p19INK4d was chosen for analysis because of its governing role in determining whether a cell in the developing CNS continues to divide or withdraw from the cell cycle. Recent data suggest that p19INK4d not only regulates cell cycle exit at the G1 phase but also plays a role in maintaining cells in the postmitotic state (Zindy et al., 1999). The state and phenotype of differentiation of the cells in the embryonic telencephalon and RMS were determined by double-labeling sections with an antibody to p19INK4d and with the antibody TuJ1 (Lee et al., 1990; Easter et al., 1993).
The spatiotemporal expression pattern of p19INK4d in the developing cerebral cortex of rats was examined between E14 and E20, during which time the majority of the neurons of the cerebral cortex are generated. During embryonic development, the cerebral cortex expands from the VZ, a pseudostratified neuroepithelium surrounding the lateral ventricles, to a multilayered laminar structure. The progenitors of the VZ, which undergo interkinetic nuclear migration, divide at the ventricular surface and generate immature neurons after withdrawing from the cell cycle. In contrast to the VZ cells, the progenitor cells of the SVZ essentially divide at the same place where they synthesize their DNA. In an orderly manner, the newly generated postmitotic neurons migrate away from the VZ (or overlying SVZ), usually along radial glia, to their final destinations in the cortical plate (Fig.1) (Edmondson and Hatten, 1987). The earliest-generated neurons of the telencephalon form the preplate, which subsequently becomes subdivided into the marginal zone and subplate by the insertion of the later-generated neurons destined for the cortical plate.
We also analyzed the expression pattern of p19INK4d by the cells of the RMS, the pathway traversed by the mitotic progeny of cells located within a distinct region of the anterior part of the postnatal subventricular zone of the forebrain designated the SVZa. After the SVZa-derived cells reach the olfactory bulb, they permanently cease division and differentiate into interneurons of the granule cell and glomerular layers of the olfactory bulb. SVZa neuronal progenitor cells possess a unique mitotic capacity compared with the progenitor cells of the rest of the telencephalon and other regions of the developing CNS. One distinguishing feature is that the SVZa and SVZa-derived cells in the RMS continue to proliferate throughout life. Unlike the progenitor cells of the embryonic telencephalic VZ, SVZa progenitors initiate differentiation without becoming postmitotic, such that proliferation and differentiation occur concurrently. In addition, on the basis of their phenotypic characteristics, they are full-fledged neurons, but they retain the ability to undergo division as they migrate, although the extent of mitosis steadily decreases as the cells approach the olfactory bulb (Menezes et al., 1995; Smith and Luskin, 1998). To investigate what accounts for the ability of SVZa neuronal progenitor cells to divide as they migrate, p19INK4dwas analyzed because of its pivotal role in regulating the mechanisms of the cell cycle.
Spatiotemporal pattern of p19INK4dexpression in the developing telencephalon
The developing telencephalon was analyzed to characterize the expression pattern of p19INK4d during cortical neurogenesis (E14–E20). By correlating the positions of cells as they undergo interkinetic nuclear migration with their stage of the cell cycle and pattern of p19INK4dexpression, our data suggest that the VZ can be further subdivided into a sublamina in which the majority of the cells are p19INK4d(+) and a sublamina in which the majority of the cells are p19INK4d(−) (Fig. 2). The p19INK4d(+) sublamina is situated in the apical portion of the VZ, adjacent to the ventricle, referred to as VZ lower (or abbreviated VZl). Interestingly there is a sharp boundary between the VZl and the VZ upper (VZu), which would not be predicted by the process of interkinetic nuclear migration. The particular environmental influences that presumably set up the boundary have not been determined. However, from BrdU and tritiated-thymidine studies (Brand and Rakic, 1979), one can conclude that the VZl contains newly generated postmitotic neurons that have just withdrawn from the cell cycle as well as cells at different stages of the mitotic cycle, including progenitor cells undergoing cytokinesis along the ventricular surface and cells in the G1 phase. Moreover, our recognition that the VZ is a bilaminar structure, on the basis of its p19INK4d-staining pattern combined with our previous knowledge from the literature of the kinetics of the VZ cells, suggested to us that the cells are expressing p19INK4d in the VZl are most likely immature neurons that have withdrawn recently from the cell cycle. Alternatively, we cannot exclude the possibility that the daughter cells destined to initiate a new round of interkinetic nuclear migration temporarily express p19INK4d(see below).
The other subdivision of the VZ revealed by its p19INK4d expression is the p19INK4d(−) sublamina in the basal or upper half of the VZ or VZu (Fig. 2). The VZu, like the VZl, is also a mixture of cells in different stages of the cell cycle, including progenitor cells going through, entering, or leaving S phase. In addition, the VZu is also comprised of immature neurons that are beginning their ascent to the cortical plate. The relatively low numbers of p19INK4d-expressing cells in the VZu may partly be because the cells do not express p19INK4d while they are actively undergoing DNA synthesis. The relatively few p19INK4d(+) cells in the VZu probably represent the postmitotic migrating neurons that withdrew recently from the cell cycle in the VZl and are en route to the cortical plate. In addition, to account for the low fraction of p19INK4d(+) cells in the VZu, we hypothesized that the postmitotic neurons traversing the VZu may temporarily downregulate their expression of p19INK4d on the basis of the observation that the VZu is flanked by the p19INK4d(+) VZl and the p19INK4d-immunoreactive intermediate zone. Evidence of this conjecture is given below.
As the neurogenesis of the cerebral cortex proceeds, the VZ becomes thinner and is completely depleted by the time all of the cortical neurons have been generated (approximately E20). However, as long as the cells of the VZ are undergoing proliferation, it has a bilaminar pattern of p19INK4d expression (Fig. 2). Furthermore, although the overall thickness of the VZ becomes smaller as cortical neurogenesis proceeds, we observed that the ratio of the thickness of VZu/VZl decreases. In other words, proportionally the p19INK4d(−) VZu becomes thinner and the p19INK4d(+) VZl becomes thicker (Fig. 2). The subventricular zone overlying the VZ, thought to be comprised of progenitor cells that undergo cytokinesis in the same position where they synthesize their DNA, also contains p19INK4d(+) cells. These p19INK4d(+) cells may represent newly generated neurons or glia that arise from the SVZ or cells traversing the SVZ destined for the cortical plate. In addition to the expression of p19INK4d by the postmitotic cells of the VZl, we observed that the postmitotic cells of the intermediate zone, cortical plate, and marginal zone also express p19INK4d. Furthermore, after the VZ is depleted of its cells, the part of the overlying SVZ that generates glia (posterior to the SVZa) also contains p19INK4d(+) cells.
Expression of p19INK4d by postmitotic neurons of the developing cerebral cortex
To determine more precisely when after cytokinesis newly generated neurons begin to express the cell cycle inhibitor p19INK4d, we examined the relationship between the pattern of expression of p19INK4d and that of neuron-specific type III β-tubulin, one of the earliest markers of immature neurons. We demonstrated previously that some immature neurons of the embryonic telencephalic VZ start to express neuron-specific type III β-tubulin (recognized by the antibody TuJ1) shortly after they exit the cell cycle in the VZ (Menezes and Luskin, 1994). Similarly, in this study, we noted a low number of TuJ1(+) cells in the VZ, which however increases over time (Fig. 2). When we stained sections of the embryonic telencephalon with TuJ1, in conjunction with anti-p19INK4d, we observed that the vast majority of TuJ1(+) cells in the VZ as well as in SVZ were also p19INK4d(+), indicating that p19INK4d is indeed expressed by the postmitotic neurons of the developing cerebral cortex. We noted, however, that higher levels of p19INK4dwere expressed by the TuJ1(+) cells of the intermediate zone, cortical plate, and marginal zone (Fig. 2), which indicates that type III β-tubulin and p19INK4d both persist and are expressed more intensely by the more mature neurons of the developing cerebral cortex.
To address the question of whether the progeny of a given progenitor cell express p19INK4d at the moment of their withdrawal from the cell cycle, we administered the cell proliferation marker BrdU to embryos aged E14–E20 45 min or 3 hr, 4 hr, or 9 hr before their perfusion and then compared the single-labeled [BrdU(+)/p19INK4d(−)] with the double-labeled [BrdU(+)/p19INK4d(+)] cells. Shorter BrdU exposure times (i.e., 45 min) allowed us to identify the VZ progenitor cells that are in the process of migration from the VZu (site of DNA synthesis and BrdU incorporation) to the ventricular surface, before mitosis (Fig.3A). When we stained the developing telencephalon with antibodies to p19INK4d and BrdU and analyzed the superficial portion of the VZl 3 hr after BrdU, we observed BrdU(+) cells, but we did not detect double-labeled cells (Fig. 3D), indicating that the progenitor cells that synthesized their DNA and were moving to the VZ surface during their G2phase do not express p19INK4d. We also examined whether the cells were single- and double-labeled in the VZ 9 hr after BrdU. The administration of BrdU to embryos at such a relatively long time period before their perfusion (i.e., 9 hr) revealed both BrdU(+)/p19INK4d(−) and BrdU(+)/p19INK4d(+) populations of cells in the superficial portion of the VZl (Fig. 3E). Our interpretation is that the BrdU(+)/p19INK4d(−) cells are the progenitor cells that took up BrdU at later time points and are moving toward the ventricular surface. On the other hand, the BrdU(+)/p19INK4d(+) cells are presumably the immature neurons migrating to the cortical plate whose ancestors incorporated BrdU at earlier time points and had enough time to migrate down to the VZ, undergo mitosis, and subsequently begin to migrate toward the cortical plate. These conclusions are consistent with the studies that showed a 12–16 hr cell cycle length for the VZ progenitors during cortical neurogenesis (Takahashi et al., 1995). Our results argue that the p19INK4d is expressed by the progeny of the VZ progenitors at the actual time of withdrawal from the cell cycle at the ventricular surface.
Changes in subcellular distribution of p19INK4din the cells of the developing telencephalon
To determine whether the subcellular distribution of p19INK4d in postmitotic cortical neurons is related to their state of differentiation as they proceed from the VZ to the cortical plate, we analyzed the intracellular localization of p19INK4d in TuJ1(+) cells. As shown above and described previously (Menezes and Luskin, 1994), the postmitotic immature neurons begin to express type III β-tubulin in the VZ presumably immediately after undergoing cytokinesis. The neuron-specific tubulin is expressed throughout the soma and gradually becomes more intense as cells differentiate and migrate through the intermediate zone to the overlying layers of the developing cortex. In contrast to the expression of type III β-tubulin, we observed that not only does the intensity of p19INK4dincrease but the subcellular distribution of p19INK4d also changes as cells undergo progressive differentiation during their ascent from the VZ to the cortical plate. Furthermore, although TuJ1 stains the entire soma, p19INK4d immunoreactivity was always more restricted to the cytoplasmic–nuclear border.
We detected layer-specific changes in the expression of p19INK4d by cells of the developing cerebral cortex, suggesting that p19INK4dis dynamically regulated. Initially, when the progeny of the progenitor cells in the VZl withdraw from the cell cycle, most of the postmitotic cells express p19INK4d outlining the cytoplasmic–nuclear border in the apical half of the cell in a crescent shape (Fig. 2). In the VZu, we observed lower levels of p19INK4d immunoreactivity than in the VZl or overlying intermediate zone, suggesting that cells traversing the VZu may downregulate their expression of p19INK4d (Fig. 2). As the immature neurons migrating through the intermediate zone start to reexpress the p19INK4d, the p19INK4d immunoreactivity no longer appears to form a crescent around the nucleus but becomes more restricted in its distribution. It is still localized to the cytoplasmic–nuclear border, but in a punctate manner (Figs. 2, 4). The intracellular distribution of p19INK4d undergoes another change as the cells enter the cortical plate and become more differentiated. p19INK4d again caps the apical half of the nucleus but appears more intense than in the VZl (Figs. 2, 4). The mechanism(s) that regulates the highly orchestrated series of changes in the intracellular localization of the p19INK4d has not been elucidated.
Spatiotemporal pattern of p19INK4d expression along the rostral migratory stream
We examined the spatiotemporal pattern of p19INK4d expression in the neonatal (P0–P2) rat forebrain and, in particular, in and around the RMS to find out when and where the SVZa-derived cells, which traverse the pathway, express the cell cycle inhibitor p19INK4d. We observed a gradient of p19INK4d expression in the RMS (Fig.5), such that as the progenitor cells originating in the SVZa migrate toward the subependymal zone in the center of the olfactory bulb, they express increasingly higher levels of p19INK4d. Furthermore, the number of cells expressing p19INK4d becomes greater along the RMS from the SVZa to the subependymal zone. The increase in the density of p19INK4d(+) cells, however, is not linear; rather it is most pronounced at the transition from SVZa to the vertical limb of the RMS (Fig. 5). The p19INK4d expression persists in the olfactory bulb in the permanently postmitotic SVZa-derived cells of the granule cell and glomerular layers, as well as in the earlier-generated neurons of the mitral cell layer of the olfactory bulb. Furthermore, the relatively quiescent ependymal cells that line the lateral ventricles showed significantly higher levels of p19INK4d expression in comparison with that of the subjacent SVZa cells (Figs. 5, 6). This pattern of p19INK4d staining of the ependyma and RMS is the inverse of the gradient of actively dividing cells. Accordingly, in the RMS we observed a stepwise reduction in the extent of BrdU incorporation along the RMS, minimal incorporation of BrdU in the subependymal zone in the middle of the olfactory bulb, and negligible labeling of the ependymal cells lining the lateral ventricles (Fig. 5). Our p19INK4d findings in conjunction with those from our BrdU studies demonstrate that an increasing proportion of the SVZa-derived cells withdraw from the cell cycle as they approach the olfactory bulb.
Expression of p19INK4d by the neuronal progenitor cells of the rostral migratory stream and their progeny
To determine how the spatiotemporal expression pattern and intracellular distribution of p19INK4dcorrelate with the phenotype, proliferative activity, and stage of differentiation of the cells in the RMS, we compared the pattern of p19INK4d expression with that of neuron-specific type III β-tubulin using immunocytochemistry. In this study, we confirmed our previous results that virtually all the cells along the RMS are immunoreactive for the neuron-specific antibody TuJ1 (Menezes et al., 1995). Furthermore, as expected from the pattern of TuJ1 staining and role of p19INK4d, we did not detect p19INK4d(+)/TuJ1(+) cells in the SVZa (Fig. 6). However, there was a pronounced gradient in the RMS of double-labeled cells as the olfactory bulb was approached; regions closer to the olfactory bulb expressed a higher proportion of p19INK4d(+)/TuJ1(+) cells. In fact the majority of the cells in the horizontal limb of the RMS coexpress p19INK4d and type III β-tubulin. Moreover, although most cells in the subependymal zone of the olfactory bulb were also double-labeled, they showed a greater intensity of staining for both p19INK4d and type III β-tubulin. We also detected a low fraction of cells that were TuJ1(+) but p19INK4d(−) in addition to the double-labeled cells in the RMS, suggesting that some cells destined for the olfactory bulb may downregulate their p19INK4d expression to undergo subsequent rounds of cell division (see below).
Although nearly all of the cells of the RMS express p19INK4d, we examined whether the intracellular distribution of p19INK4dchanges as a function of the phase of the cycle and/or the state of differentiation of a cell, as was found to be the case in the developing cerebral cortex. We found that p19INK4d was expressed at the cytoplasmic–nuclear border of the cells all along the RMS (Figs. 5, 6), just as in the p19INK4d-expressing cells of the different layers of developing cerebral cortex (Figs. 2, 4). However, in contrast to the developing cerebral cortex, in the cells of the RMS the p19INK4d was always expressed in a punctate manner both in dividing as well as in the postmitotic cells. Furthermore, in most cells of the RMS, the p19INK4d immunoreactivity was concentrated in the half of the cell closest to the leading process, similar to its distribution in immature neurons migrating toward the cortical plate. Collectively, these findings indicate that the SVZa-derived cells may regulate their expression of p19INK4ddifferently from cells destined for the cerebral cortex.
To determine whether the p19INK4d-expressing cells within the RMS undergo cell division, we administered an intraperitoneal injection of BrdU to neonatal pups and perfused them at various subsequent time points. We analyzed the distribution of p19INK4d(+)/BrdU(+) cells as a function of the position of a cell along the anterior–posterior axis of the RMS. Because the length of the cell cycle of SVZa-derived cells is ∼12 hr (Smith and Luskin, 1998), we perfused pups 3 hr after BrdU (while the proliferating cells were in S/G2 phase) and 9 hr after BrdU (to give the cells sufficient time to incorporate BrdU and proceed through the M phase of the cell cycle). Because SVZa-derived cells migrate at 23 μm/hr (Luskin and Boone, 1994), we surmised that where we see BrdU(+) cells is probably relatively close to their site of proliferation; SVZa-derived cells cannot traverse more than a fraction of the pathway within 3 or 9 hr time intervals. An analysis of the distribution of labeled cells using confocal microscopy revealed the presence of BrdU(+)/p19INK4d(−) cells in the SVZa 3 and 9 hr after BrdU administration (Fig.5D,F). These data indicate that the neuronal progenitor cells in the SVZa undergo cell divisions without expressing p19INK4d, unlike the immature neurons of the developing cerebral cortex. Conversely, the presence of BrdU(+)/p19INK4d(+) cells in the horizontal limb of the RMS 9 hr after BrdU (Fig. 5G), but not 3 hr after BrdU (Fig. 5E), suggests that cells may downregulate their p19INK4d, incorporate BrdU within the 9 hr interval after injection, and then reexpress p19INK4d in the pathway.
DISCUSSION
By blocking the cell cycle at the G1 phase, CDKIs and, in particular, p19INK4d might couple proliferation arrest to the terminal differentiation of the cells in the developing CNS. To determine whether the progenitor cells of the telencephalic VZ differ in their cell cycle kinetics from that of the neonatal RMS, we characterized the spatiotemporal expression pattern of p19INK4d in both progenitor cell populations. We showed that the telencephalic VZ can be divided into a predominantly p19INK4d(+) sublamina, in the apical half of the VZ (VZl), and a predominantly p19INK4d(−) sublamina, in the basal half of the VZ (VZu). Despite previous studies that have analyzed the genesis of cells in the VZ (Brand and Rakic, 1979; Luskin and Shatz, 1985; Menezes and Luskin, 1994; Bittman et al., 1997; Kornack and Rakic, 1998), a strict sublamination was not evident previously. We further demonstrated that p19INK4d is localized to the cytoplasmic–nuclear border in both the apical domain of dividing VZ cells and postmitotic neurons. Moreover, p19INK4d undergoes a set of subcellular rearrangements and varies in its distribution around the nucleus as a function of the laminar position of a cell. In contrast to telencephalic VZ cells, SVZa-derived neuronal progenitor cells of RMS exhibit a different developmental sequence of p19INK4d expression. Our data suggest that the SVZa-derived cells undergo multiple rounds of cell division by repetitively downregulating their p19INK4dexpression. In addition, there was an anteriorhigh–posteriorlowgradient of p19INK4d expression along the RMS. We showed that the p19INK4dexpression persists in the postmitotic neurons arising in both the VZ and SVZa. Taken together, our findings indicate that after the newly generated neurons of the developing cerebral cortex leave the cell cycle, they remain forever postmitotic, whereas SVZa-derived cells may successively downregulate their p19INK4dexpression and reenter the cell cycle before becoming permanently postmitotic neurons of the olfactory bulb.
The pattern of p19INK4d immunoreactivity in the telencephalic ventricular zone reveals distinct p19INK4d(+) and p19INK4d(−) sublaminae
It has long been known that the neuroepithelium of the developing telencephalon has two zones of proliferating cells, the VZ and the later-arising SVZ. In this study, we provide evidence that the VZ has two distinct subdivisions. On the basis of the expression of p19INK4d, one sublamina, situated in the apical part of the VZ adjacent to the ventricular lining (the VZl), has predominantly p19INK4d(+) cells, and the other sublamina, in the basal or upper half of the VZ (the VZu), contains mostly p19INK4d(−) cells. We also showed that the VZl becomes wider relative to the VZu as neurogenesis proceeds, in agreement with the fact that the pool of postmitotic neurons increases as the dividing cell population is depleted.
Although previous studies, using in situ hybridization, have analyzed the expression pattern of the CDKIs in the developing mouse brain (Zindy et al., 1997b), to better understand the spatiotemporal distribution of p19INK4d in the developing forebrain, immunocytochemical analysis is also required. In situ hybridization studies showed that the p19INK4d mRNA was expressed throughout all laminae of the developing neocortex. From this finding it was concluded that it is present in both proliferating and differentiating cell populations of the cerebral cortex. In contrast, our immunocytochemical results showed a differential expression pattern of p19INK4d among the different layers of the developing cerebral cortex. Specifically, VZu progenitors that are in the S phase of the cell cycle do not express p19INK4d. However, many cells in the VZl do express p19INK4d, and most likely these are immature neurons that have withdrawn recently from the cell cycle. Furthermore, reports that regulation of CDKIs occurs at the post-translational level (Pagano et al., 1995) underscore the need for immunocytochemical analyses of their expression. Although our data demonstrate that p19INK4d appears to be differentially regulated by the cells of the VZl and VZu, these data do not exclude the possibility that VZl progenitor cells temporarily express p19INK4d before they initiate a new round of interkinetic nuclear migration.
A few studies have suggested that under specific circumstances, p19INK4d is expressed by cells undergoing division, disputing the idea that p19INK4dis only expressed by postmitotic cells. In particular, Hirai et al. (1995) and Thullberg et al. (2000) demonstrated the oscillation of p19INK4d expression by cultured macrophages and fibroblasts, after they were induced to reenter the cell cycle from a quiescent serum-deprived state. In these cells, the p19INK4d mRNA levels were low during G1, highest in S, and low again as they approached the subsequent G1. When retroviral-mediated gene transfer was used to express p19INK4d constitutively in cultures of cycling fibroblasts, however, the cells were arrested at the G1 phase. This indicates that in fibroblasts the induction of p19INK4d during the G1 phase is sufficient to block the G1–S progression. Our data argue that the temporal sequence of expression of p19INK4d, described for macrophages and fibroblasts, is not what occurs when the progenitor cells of the VZ undergo division in vivo. Instead, our analysis of the temporal pattern of p19INK4d expression during interkinetic nuclear migration demonstrates that there is a negligible level of p19INK4d expression when the VZ progenitor cells are in the S phase and indicates that p19INK4d is expressed at high levels in the VZl by newly generated postmitotic neurons. An alternative explanation to the scenario that only postmitotic cells of the VZl express p19INK4d is that some actively dividing VZ progenitor cells temporarily express p19INK4d before they undergo another round of interkinetic nuclear migration. Although our studies cannot exclude this as a formal possibility for VZ cells, it is likely to be the case for the cells of the RMS. However, the cells situated within the SVZa show the highest levels of BrdU incorporation and the lowest levels of p19INK4d expression. Therefore, we have proposed that cells traversing the RMS may undergo dedifferentiation before successive rounds of cell division. The different patterns of p19INK4d expression exhibited by the dividing cells of the VZ and SVZa strengthen the conclusion that the kinetics of p19INK4d expression during the cell cycle may be cell-type specific.
p19INK4d expression in the neonatal rostral migratory stream is inversely correlated with the distribution of mitotically active cells
Previous studies have demonstrated that the SVZa has a higher proliferative capacity than the subependymal zone in the middle of the olfactory bulb (Menezes et al., 1995; Smith and Luskin, 1998). In agreement with the studies based on BrdU incorporation, our study demonstrated a spatiotemporal gradient of p19INK4d expression by SVZa-derived cells all along the RMS; the cells of the subependymal zone expressed higher levels of p19INK4d than did those of the proximal RMS. This anteriorhigh–posteriorlowgradient indicates that the SVZa-derived cells withdraw from the cell cycle as they approach the olfactory bulb. In addition, the ependymal cells that line the lateral ventricles express significantly higher levels of p19INK4d also in agreement with their low rate of BrdU incorporation and proliferation (Doetsch et al., 1999). Our data suggest that in contrast to VZ progenitor cells, the cell cycle kinetics of the SVZa-derived cells is differentially regulated by endogenous and/or exogenous signals, which permit their ongoing mitosis even though they express a neuronal phenotype.
The subcellular redistribution of p19INK4d by postmitotic neurons as they undergo migration and differentiation
Our findings revealed systematic changes in the subcellular distribution of p19INK4d as postmitotic neurons migrate from the VZ to the cortical plate, in addition to the differential expression of p19INK4d by the cells in the sublaminae of the VZ (VZl vs VZu) (Figs. 2, 4). The changes that p19INK4d undergoes are somewhat analogous to the differential subcellular distribution of proteins, such as Notch (Sestan et al., 1999), in the different layers of the developing cerebral cortex. Although the underlying mechanism(s) regulating the dynamic redistribution of p19INK4d is not known, layer-specific regulatory signals could be involved.
Several studies have suggested that the phenotype of a cell is related to whether it is the product of a symmetric or asymmetric division. For example, during cortical development, the fate of a cell is believed to be established by the differential distribution of Notch to the postmitotic apical daughter cell and Numb to the proliferative basal daughter cell after an asymmetric cell division (Chenn and McConnell, 1995; Doe and Spana, 1995; Zhong et al., 1996). Although the specific functions subserved by p19INK4d in the progenitor cells of the telencephalon have not been determined, our data suggest that when a VZ progenitor cell undergoes an asymmetric cell division, p19INK4d becomes segregated to the cytoplasmic–nuclear border of the apical postmitotic daughter cell. The significance of the concentration of p19INK4d in the SVZa-derived cells at the cytoplasmic–nuclear border, in the portion closest to the leading process, remains to be determined.
SVZa progenitor cells potentially dedifferentiate as they successively divide and migrate in the rostral migratory stream
Previous studies have demonstrated that the SVZa neuronal progenitor cells exhibit atypical properties of proliferation in comparison with other CNS progenitors. Specifically, SVZa progenitor cells continue to divide as they migrate despite expressing a neuronal phenotype. Our previous and current findings of the presence of numerous TuJ1(+)/BrdU(+) cells along the RMS substantiate the premise that SVZa-derived neurons are dividing and that large numbers of neurons are added to the rat olfactory bulb during postnatal life. Despite the presence of BrdU(+) cells in the RMS, we also demonstrated the expression of p19INK4d all along the RMS. This raises a discrepancy because our analysis of the developing cerebral cortex has shown that p19INK4d is expressed exclusively by postmitotic cells that no longer incorporate BrdU. Our results showing TuJ1(+)/p19INK4d(−) neurons in the RMS in addition to BrdU(+)/p19INK4d(+) cells may reconcile this apparent contradiction (Fig.7). We propose that SVZa-derived cells in the RMS can downregulate their expression of p19INK4d and undergo subsequent rounds of cell division and that they repetitively undergo dedifferentiation and division to generate the vast number of cells destined for the olfactory bulb.
Footnotes
- Received August 29, 2000.
- Revision received February 2, 2001.
- Accepted February 8, 2001.
This work was supported by the National Institute on Deafness and Other Communication Disorders Grant RO1 DC03190 to M.B.L. We thank Christopher P. Noyes for his technical assistance and the members of the lab of M.B.L. for their helpful comments on this manuscript. We are also grateful to Drs. Douglas Falls and Kevin Moses for their suggestions and valuable comments on this manuscript.
Correspondence should be addressed to Dr. Marla B. Luskin, Department of Cell Biology, Emory University School of Medicine, 1648 Pierce Drive, Atlanta, GA 30322. E-mail: luskin{at}cellbio.emory.edu.
- Copyright © 2001 Society for Neuroscience