Developmental and regional distribution of aspartoacylase in rat brain tissue

The function of N-acetyl-aspartate (NAA), a predominant molecule in the brain, has not yet been determined. However, NAA is commonly used as a putative marker of viable neurones. To investigate the possible function of NAA, we determined the anatomical, developmental and cellular distribution of aspartoacylase, which catalyses the hydrolysis of NAA. Levels of aspartoacylase activity were measured during postnatal development in several brain regions. The differential distribution of aspartoacylase activity in purified populations of cells derived from the rat CNS was also investigated. The developmental and anatomical distribution of aspartoacylase correlated with the maturation of white matter tracts in the rat brain. Activity increased markedly after 7 days and coincided with the time course for the onset of myelination in the rat brain. Gray matter showed little activity or developmental trend. There was a 60-fold excess in optic nerve (a white matter tract) when compared with cortex at 21 days of development. In the adult brain there was a 18-fold difference in corpus callosum compared with cortex (stripped of corpus callosum). Cellular studies demonstrated that purified cortical neurons and cerebellar granular neurones have no activity. Primary O-2A progenitor cells had moderate activity, with three-fold higher activity in immature oligodendrocyte and 13-fold increase in mature oligodendrocytes (myelinating cells of the CNS). The highest activity was seen in type-2 astrocytes (20-fold difference compared with O-2A progenitors) derived from the same source. Aspartoacylase activity increased with time in freshly isolated astrocytes, with significantly higher activity after 15 days in culture. We conclude that aspartoacylase activity in the developing postnatal brain corresponds with maturation of myelination, and that the cellular distribution is limited to glial cells.     

Macroglial cell development in embryonic rat brain: studies using monoclonal antibodies, fluorescence activated cell sorting, and cell culture

Astrocytes, ependymal cells, and oligodendrocytes have been shown to develop on the same schedule in dissociated cell cultures of early embryonic rat brain as in vivo. Subsequent studies showed that there are two major types of astrocyte (type-1 and type-2), which, in cultures of perinatal optic nerve, develop as two distinct lineages. In such cultures, type-2 astrocytes and oligodendrocytes develop from the same, bipotential, (O-2A) progenitor cells, which differentiate into type-2 astrocytes in 10% fetal calf serum (FCS) and into oligodendrocytes in less than or equal to 0.5% FCS. In light of these findings, we now have extended our studies on macroglial cell development in rat brain and show the following: (i) The first astrocytes to develop have a type-1 phenotype, while astrocytes with a type-2 phenotype do not develop until almost 2 weeks later, just as in the optic nerve. (ii) Most importantly, type-2 astrocytes, like the other macroglial cells, develop on the same schedule in cultures of early embryonic (less than or equal to E15) brain as they do in vivo. (iii) By contrast, both oligodendrocytes and type-2 astrocytes develop prematurely in cultures of E17 brain, and FCS influences this development in the same way it does in perinatal optic nerve cultures. (iv) Type-2 astrocyte precursors are labeled by the A2B5 monoclonal antibody, as shown previously for oligodendrocyte precursors in brain and for O-2A progenitor cells in optic nerve. Taken together with our previous findings, these results suggest that oligodendrocytes and type-2 astrocytes in brain develop from bipotential O-2A progenitor cells, whose choice of developmental pathway and timing of differentiation depend on mechanisms that operate independently of brain morphogenesis.     

Clostridium perfringens Epsilon Toxin Targets Granule Cells in the Mouse Cerebellum and Stimulates Glutamate Release

Epsilon toxin (ET) produced by C. perfringens types B and D is a highly potent pore-forming toxin. ET-intoxicated animals express severe neurological disorders that are thought to result from the formation of vasogenic brain edemas and indirect neuronal excitotoxicity. The cerebellum is a predilection site for ET damage. ET has been proposed to bind to glial cells such as astrocytes and oligodendrocytes. However, the possibility that ET binds and attacks the neurons remains an open question. Using specific anti-ET mouse polyclonal antibodies and mouse brain slices preincubated with ET, we found that several brain structures were labeled, the cerebellum being a prominent one. In cerebellar slices, we analyzed the co-staining of ET with specific cell markers, and found that ET binds to the cell body of granule cells, oligodendrocytes, but not astrocytes or nerve endings. Identification of granule cells as neuronal ET targets was confirmed by the observation that ET induced intracellular Ca²⁺ rises and glutamate release in primary cultures of granule cells. In cultured cerebellar slices, whole cell patch-clamp recordings of synaptic currents in Purkinje cells revealed that ET greatly stimulates both spontaneous excitatory and inhibitory activities. However, pharmacological dissection of these effects indicated that they were only a result of an increased granule cell firing activity and did not involve a direct action of the toxin on glutamatergic nerve terminals or inhibitory interneurons. Patch-clamp recordings of granule cell somata showed that ET causes a decrease in neuronal membrane resistance associated with pore-opening and depolarization of the neuronal membrane, which subsequently lead to the firing of the neuronal network and stimulation of glutamate release. This work demonstrates that a subset of neurons can be directly targeted by ET, suggesting that part of ET-induced neuronal damage observed in neuronal tissue is due to a direct effect of ET on neurons.     

An inducer protein may control the timing of fate switching in a bipotential glial progenitor cell in rat optic nerve

In rat optic nerve, oligodendrocytes and type-2 astrocytes develop from a common (O-2A) progenitor cell. The first oligodendrocytes differentiate at birth, while the first type-2 astrocytes differentiate in the second postnatal week. We previously showed that the timing of oligodendrocyte differentiation depends on an intrinsic clock in the O-2A progenitor cell. Here we provide evidence that the timing of type-2 astrocyte differentiation, by contrast, may depend on an inducing protein that appears late in the developing nerve. We show that extracts of 3- to 4-week-old, but not 1-week-old, rat optic nerve contain a protein (apparent Mr approximately 25,000) that induces O-2A progenitor cells in culture to express glial fibrillary acidic protein (GFAP), an astrocyte-specific marker in the rat central nervous system.     

Type-2 astrocyte development in rat brain cultures is initiated by a CNTF-like protein produced by type-1 astrocytes

O-2A progenitor cells are bipotential glial precursors that give rise to both oligodendrocytes and type-2 astrocytes on a precise schedule in the rat CNS. Studies in culture suggest that oligodendrocyte differentiation occurs constitutively, while type-2 astrocyte differentiation requires an exogenous inducer such as fetal calf serum. Here we describe a rat brain cell culture system in which type-2 astrocytes develop on schedule in the absence of exogenous inducers. Coincident with type-2-astrocyte development, the cultures produce an approximately 20 kd type-2-astrocyte-inducing factor(s). Purified cultures of type-1 astrocytes can produce a similar factor(s). Under conditions where they produce type-2-astrocyte-inducing factor(s), both brain and type-1 astrocyte cultures produce a factor(s) with ciliary neurotrophic (CNTF)-like activity. Purified CNTF, like the inducers from brain and type-1 astrocyte cultures, prematurely induces type-2 astrocyte differentiation in brain cultures. These findings suggest that type-2 astrocyte development is initiated by a CNTF-like protein produced by type-1 astrocytes.     

Analysis of the cell-cell interactions that control type-2 astrocyte development in vitro

Oligodendrocytes and type-2 astrocytes develop sequentially from O-2A progenitor cells in the rat CNS. We have reproduced this sequential development in a simplified, serum-free in vitro system: in cultures of newborn optic nerve cells treated with platelet-derived growth factor to maintain O-2A progenitor cell proliferation, progenitor cells differentiate into oligodendrocytes during the first week in vitro and into type-2 astrocytes during the second week. Thus all of the signals needed for type-2 astrocyte development are made by serum-free optic nerve cultures, indicating that neurons are not required. By manipulating the cellular composition of the cultures, we provide evidence that type-2 astrocyte development does not depend on oligodendrocytes, but instead requires non-O-2A lineage cells, which are also responsible for timing this development.     

PDGF receptors on cells of the oligodendrocyte-type-2 astrocyte (O-2A) cell lineage

It has been shown previously that cultures of rat optic nerve contain three types of macroglial cells--oligodendrocytes and two types of astrocytes. Type-1 astrocytes develop from their own precursor cells beginning before birth, while oligodendrocytes and type-2 astrocytes develop postnatally from a common bipotential precursor called the O-2A progenitor cell. Proliferating O-2A progenitor cells give rise to postmitotic oligodendrocytes beginning around birth, and to type-2 astrocytes beginning in the second postnatal week. Studies in vitro have suggested that platelet-derived growth factor (PDGF), secreted by type-1 astrocytes, plays an important part in timing oligodendrocyte development: PDGF seems to keep O-2A progenitor cells proliferating until an intrinsic clock in the progenitor cells initiates the process leading to oligodendrocyte differentiation. The clock apparently determines when a progenitor cell becomes unresponsive to PDGF, at which point the cell stops dividing and, as a consequence, automatically differentiates into an oligodendrocyte. Here we have used radiolabelled PDGF to show that O-2A progenitor cells have PDGF receptors, suggesting that these cells respond directly to PDGF. The receptors resemble the type A PDGF receptor previously described on human fibroblasts and are initially retained when progenitor cells stop dividing and develop in vitro into oligodendrocytes. The latter finding indicates that receptor loss is not the reason that progenitor cells initially become mitotically unresponsive to PDGF.     

Two proliferative stages of the oligodendrocyte lineage (A2B5+O4- and O4+GalC-) under different mitogenic control

Cell proliferation during successive stages of oligodendrocyte development was delineated in the rat brain and optic nerve. Surface antigens, A2B5, O4, and galactocerebroside (GalC) identified three cell populations emerging in sequence; the incorporation of bromodeoxyuridine into newly synthesized DNA identified the proliferative cells. In vivo, progenitor cells with phenotypes A2B5+O4- and A2B5+O4+GalC- were both proliferative, whereas differentiated GalC+ oligodendrocytes were not. Under basal conditions of culture, the proliferation of both progenitor cell types of the optic nerve was nearly abolished. Activity was restored for A2B5+O4- precursor cells with medium conditioned by either type-1 astrocytes, meningeal cells, or cerebellar interneurons. In contrast, intermediate O4+GalC- cells (proligodendrocytes) were refractory to the astroglial and meningeal signals, but remained as responsive as their precursor cells to the neuronal stimulus. These data further characterize the O4+GalC- proligodendrocyte as a distinct developmental stage, one that specifies a changing response of the cell to environmental mitogens.     

Infection by coronavirus JHM of rat neurons and oligodendrocyte-type-2 astrocyte lineage cells during distinct developmental stages.

Primary telencephalic cultures derived from neonatal Wistar Furth rats were able to support the growth of coronavirus JHM if a viable neuronal population was maintained. This occurred under serum-free defined, but not serum-supplemented, growth conditions. The importance of neurons in establishing infections in mixed cultures was confirmed by immunocytochemical and electron microscopic studies. Glia, although more abundant than neurons in these cultures, were less frequently infected during the initial 48 h postinoculation. The two glial lineages present in mixed telencephalic cultures were separated into type-1 astrocytes and oligodendrocyte-type-2 astrocyte (O-2A) lineage cells and individually assessed for their ability to support virus growth. Infection could not be established in type-1 astrocytes regardless of the culture conditions employed, consistent with our previous study (S. Beushausen and S. Dales, Virology 141:89-101, 1985). In contrast, infections could be initiated in selected O-2A lineage cells grown in serum-free medium. Virus multiplication was however significantly reduced by preconditioning the medium with mixed telencephalic or enriched type-1 astrocyte cultures, suggesting that intercellular interactions mediated by soluble factor(s) can influence the infectious process in O-2A lineage cells. This presumption was supported by eliciting similar effects with basic fibroblast growth factor and platelet-derived growth factor, two central nervous system cytokines known to control O-2A differentiation. The presence of these cytokines, which synergistically block O-2A cells from differentiating into oligodendrocytes was correlated with specific and reversible resistance to JHM virus (JHMV) infection. These data, combined with our finding that accelerated terminal differentiation of the oligodendrocyte phenotype confers resistance to JHMV (Beushausen and Dales, Virology, 1985), suggest that the permissiveness of O-2A cells for JHMV is restricted to a discrete developmental stage.     

A Simplified Method for Isolating Highly Purified Neurons, Oligodendrocytes, Astrocytes, and Microglia from the Same Human Fetal Brain Tissue

Elucidation of the underlying pathogenic mechanisms leading to apoptosis of neurons and oligodendrocytes and activation of microglia and astrocytes in different neurodegenerative and neuroinflammatory disorders remains a challenge in neuroscience. In order to overcome the challenge and find out therapeutic remedies, it is important to study live and death processes in each and every cell type of the brain. Here we present a protocol of isolating highly purified microglia, astrocytes, oligodendrocytes, and neurons, all four major cell types of the CNS, from the same human fetal brain tissue. As found in vivo, these primary neurons and oligodendroglia underwent apoptosis and cell death in response to neurodegenerative challenges. On the other hand, astroglia, and microglia, cells that do not die in neurodegenerative brains, became activated after inflammatory challenge. The availability of highly purified human brain cells will increase the possibility of developing therapies for different neurodegenerative disorders. M. Jana and A. Jana have equal contribution to the work.     

A role for platelet-derived growth factor in normal gliogenesis in the central nervous system

The bipotential progenitor cells (O-2A progenitors) that produce oligodendrocytes and type-2 astrocytes in the developing rat optic nerve are induced to proliferate in culture by type-1 astrocytes. Here, we show that the astrocyte-derived mitogen is platelet-derived growth factor (PDGF). PDGF is a potent mitogen for O-2A progenitor cells in vitro. Mitogenic activity in astrocyte-conditioned medium comigrates with PDGF on a size-exclusion column, competes with PDGF for receptors, and is neutralized by antibodies to PDGF. PDGF dimers can be immunoprecipitated from astrocyte-conditioned medium, and mRNA encoding PDGF is present in rat brain throughout gliogenesis. We propose that astrocyte-derived PDGF is crucial for the control of myelination in the developing central nervous system.     

Ornithine-σ-Aminotransferase Exhibits Different Kinetic Properties in Astrocytes, Cerebral Cortex Interneurons, and Cerebellar Granule Cells in Primary Culture

To evaluate a possible role of ornithine-delta-aminotransferase (EC 2.6.1.13; Orn-T) as a rate-limiting enzyme for the synthesis of transmitter glutamate and gamma-aminobutyric acid (GABA), respectively, its activity and kinetic properties were analyzed in cultured astrocytes as well as in neuronal cultures consisting mainly of glutamatergic neurons (cerebellar granule cells) or GABAergic neurons (cerebral cortex interneurons). For comparison the activity and kinetics of Orn-T were also assayed in mouse brain homogenates. The highest activity of Orn-T was found in astrocytes and in cerebral cortical neurons (5.3 +/- 0.5 and 5.3 +/- 0.4 nmol X mg-1 X min-1, respectively) whereas the activities of Orn-T in cerebellar granule cell cultures and in mouse brain were found to be about half of these values (3.1 +/- 0.3 and 2.8 +/- 0.1 nmol X min-1 X mg-1, respectively). From a kinetic study of Orn-T in the different preparations only a relatively low affinity for the enzyme with respect to ornithine was found in cerebellar granule cells, astrocytes, and whole brain [apparent Km values (at 0.5 mM alpha-ketoglutarate): 4.7 +/- 0.9, 4.3 +/- 2.2, and 6.8 +/- 2.2 mM, respectively] whereas the corresponding Km value for Orn-T in cerebral cortex interneurons was found to be significantly lower (apparent Km: 0.8 +/- 0.3 mM). The enzyme was not found to be inhibited by GABA (range 0.1 - 10 mM) in any of the preparations.     

In vitro expression of N-acetyl aspartate by oligodendrocytes: implications for proton magnetic resonance spectroscopy signal in vivo

Magnetic resonance spectroscopy (MRS) provides a noninvasive means of assessing in vivo tissue biochemistry. N-Acetyl aspartate (NAA) is a major brain metabolite, and its presence is used increasingly in clinical and experimental MRS studies as a putative neuronal marker. A reduction in NAA levels as assessed by in vivo 1H MRS has been suggested to be indicative of neuronal viability. However, temporal observations of brain pathologies such as multiple sclerosis, mitochondrial encephalopathy with lactic acidosis and stroke-like episodes (MELAS), and hypothyroidism have shown reversibility in NAA levels, possibly reflecting recovery of neuronal function. A knowledge of the cellular localisation of NAA is critical in interpreting these findings. The assumption that NAA is specific to neurones is based on previous immunohistochemical studies on whole brain using NAA-specific antibodies. The neuronal localisation was further substantiated by cell culture experiments in which its presence in the oligodendrocyte-type 2 astrocyte progenitors and immature oligodendrocytes, but not in the mature oligodendrocytes, was observed. More recently, studies on oligodendrocyte biology have revealed the requirement for trophic factors to promote the generation, maturation, and survival of oligodendrocytes in vitro. Here, we have used this new information to implement a more pertinent cell cultivation procedure and demonstrate that mature oligodendrocytes can express NAA in vitro. This observation brings into question whether the NAA changes observed in clinical in vivo 1H MRS studies reflect neuronal function alone. The data presented here support the hypothesis that oligodendrocytes may express NAA in vivo and contribute to the NAA signal observed by 1H MRS.     

Expression of fibronectin and laminin by different types of mouse glial cells cultured in a serum-free medium

The expression of fibronectin and laminin by cultured glial cells was studied. The glial culture from neonatal mouse cerebra maintained in a chemically defined, serum-free medium consisted of type-1 astrocytes, oligodendrocyte-type-2 astrocyte (O-2A) progenitor cells, oligodendrocytes and type-2 astrocytes. Double-labelling immunofluorescent experiments performed using the mixed glial culture indicated that fibronectin and laminin are expressed in different patterns among the glial subtypes. The staining intensities with anti-fibronectin or anti-laminin antibodies decreased in the order: type-1 astrocytes, O-2A progenitor cells and type-2 astrocytes. Both molecules were deposited in a fibrillar matrix underneath type-1 astrocytes, whereas only intracytoplasmic localization of these molecules was observed with O-2A progenitor cells and type-2 astrocytes. Western blot analysis showed that glial fibronectin has a slightly higher molecular weight than mouse plasma fibronectin (230 kDa) and that glial laminin is a variant with a 220 kDa B chain present and the 400 kDa A chain missing. Using enzyme-linked immunosorbent assays (ELISA), these molecules were detected in the glial extracellular matrix at the concentration of 4 ng/10⁶ cells. A large amount of fibronectin (82 ng/10⁶ cells) was secreted into the culture medium, while secretion of laminin was not detected.     

Primary structure of a dynamin-related mouse mitochondrial GTPase and its distribution in brain, subcellular localization, and effect on mitochondrial morphology

A new member of the dynamin GTPase family (OPA1) was recently identified in humans and shown to be mutated in patients with dominant optic atrophy. To understand better the function of mammalian OPA1, we isolated a mouse ortholog (mOPA1) from brain and raised a specific antibody against its C terminus. The subcellular distribution of mOPA1 overexpressed in COS-7 cells largely overlapped that of endogenous cytochrome c, a well known mitochondrial marker, and dramatically affected mitochondrial morphology, altering it from tubular to vesicular. Mitochondrial targeting was mediated by the N-terminal region of mOPA1 as follows: deletion of the 124 N-terminal amino acids eliminated mitochondrial targeting, although fusion of the N-terminal 60 or 90 amino acids of mOPA1 with green fluorescent protein resulted in its mitochondrial targeting. mOPA1 was expressed widely in the mouse brain, especially in neurons of olfactory bulb, cerebral cortex, piriform cortex, hypothalamus, hippocampus, red nucleus, cochlear nucleus, motor trigeminal nucleus, facial nucleus, cerebellar nucleus, and Purkinje cells. Within dissociated cerebellar cells, mOPA1 protein was clearly observed in the dendrites and somas of neuronal cells, as well as in astrocytes and meningeal cells. In each case, it was distributed in the vesicular pattern seen in other cell types.     

Irradiation in vitro discriminates between different O-2A progenitor cell subpopulations in the perinatal central nervous system of rats

The effects of X irradiation on oligodendrocyte-type-2-astrocyte (O-2A) progenitor cells derived from different regions of the perinatal central nervous system (CNS) of rats were investigated in vitro. The O-2A progenitor cells can differentiate into either oligodendrocytes or type-2 astrocytes. The depletion of these cells could lead to demyelination, seen as a delayed reaction after irradiation of the CNS in vivo. To quantify cell survival, O-2A progenitor cells were grown on monolayers of type-1 astrocytes. Monolayers of type-1 astrocytes stimulate O-2A progenitor cells to divide. O-2A progenitor cells were irradiated in vitro and clonogenic cell survival was measured. The O-2A progenitor cells derived from perinatal optic nerve were quite radiosensitive in contrast to O-2A progenitor cells derived from perinatal spinal cord and perinatal corpus callosum. Furthermore, O-2A progenitor cells derived from the optic nerve formed smaller colonies, with most colonies showing early differentiation into oligodendrocytes. In contrast, more than half of the colonies derived from corpus callosum did not show any differentiation after 2 weeks in vitro and kept growing. These differences support the view that perinatal O-2A progenitor cells derived from the optic nerve are committed progenitor cells while the O-2A progenitor cells derived from the perinatal corpus callosum and the perinatal spinal cord have more stem cell properties.     

Functions of N-acetyl-L-aspartate and N-acetyl-L-aspartylglutamate in the vertebrate brain: role in glial cell-specific signaling

N-Acetyl-L-aspartate (NAA) and its derivative N-acetylaspartylglutamate (NAAG) are major osmolytes present in the vertebrate brain. Although they are synthesized primarily in neurons, their function in these cells is unclear. In the brain, these substances undergo intercompartmental cycles in which they are released by neurons in a regulated fashion and are then rapidly hydrolyzed by catabolic enzymes associated with glial cells. Recently, the catabolic enzyme for NAA hydrolysis has been found to be expressed only in oligodendrocytes, and the catabolic enzyme for NAAG expressed only in astrocytes. These results indicate an unusual tricellular metabolic sequence for the synthesis and hydrolysis of NAAG wherein it is synthesized in neurons from NAA and L-glutamate, hydrolyzed to NAA and L-glutamate by astrocytes, and further hydrolyzed to L-aspartate and acetate by oligodendrocytes. Since the discovery that the NAA and NAAG anabolic products of neurons are specifically targeted to oligodendrocytes and astrocytes, respectively, this unique metabolic compartmentalization also suggests that these substances may play an important role in cell-specific glial signaling. In this review, it is hypothesized that a key function of NAA and NAAG in the vertebrate brain is in cell signaling and that these substances are important in the regulation of interactions of brain cells and in the establishment and maintenance of the nervous system.     

Extracellular matrix of cultured glial cells: selective expression of chondroitin 4-sulfate by type-2 astrocytes and their progenitors

We have studied the extracellular matrix composition of cultured glial cells by immunocytochemistry with different monoclonal and polyclonal antibodies. Double immunofluorescence experiments and metabolic labeling with [3H]glucosamine performed in different types of cerebellar and cortical cultures showed that bipotential progenitors for type-2 astrocytes and for oligodendrocytes (recognized by the monoclonal antibody LB1 at early stages of their development) synthesize chondroitin sulfate (CS) and deposit this proteoglycan in their extracellular matrix. The distribution of the various [3H]glucosamine-labeled glycosaminoglycans between the intracellular and the extracellular space was different. CS was present both within the cells and in the culture medium, although in different amounts. Bi-potential progenitors became also O4-positive during their development in vitro. At the stage of O4-positivity they were still stained with antibodies against CS. However, when the progenitor cells were maintained in serum-free medium and differentiated into Gal-C-positive oligodendrocytes, they became CS-negative. In the presence of fetal calf serum in the culture medium, the bipotential progenitors differentiated into GFAP-positive type-2 astrocytes. These cells still expressed CS: their Golgi area and their surface were stained with anti-CS antibodies. Staining with monoclonal antibodies specific for different types of CS (4-sulfate, 6-sulfate, and unsulfated) revealed that both bipotential progenitors and type-2 astrocytes synthesized only chondroitin 4-sulfate. Type-1 astrocytes were negative for both the polyclonal and the monoclonal anti-CS antibodies. Finally, type-2 astrocytes and their progenitors were weakly stained with anti-laminin antibodies and unstained with anti-fibronectin. Type-1 astrocytes were positive for both anti-laminin and anti-fibronectin antibodies and appeared to secrete fibronectin in the extracellular space.     

Clonal analysis of oligodendrocyte development in culture: evidence for a developmental clock that counts cell divisions

The clonal development of oligodendrocytes was studied by culturing individual oligodendrocyte--type-2 astrocyte (O-2A) progenitor cells on monolayers of type-1 astrocytes, which stimulate O-2A progenitor cells to divide. Oligodendrocytes developed by a proliferative lineage in which clonal progeny differentiated together after a number of cell divisions. Most O-2A progenitor cells had similar cell cycle times (1-2 days), but their proliferative capacity varied greatly: some divided only once while others divided up to eight times before differentiating. sister cells behaved similarly when recultured separately on astrocyte monolayers. These findings are consistent with the cell-division-counting hypothesis previously proposed to explain the timing of oligodendrocyte differentiation. They also unambiguously establish the phenotype of O-2A progenitor cells in vitro and demonstrate that these cells respond directly to growth factors produced by type-1 astrocyte monolayers.