(Received for publication, June 13, 1996, and in revised form, September 18, 1996)
From the Molecular Biology Program, Memorial Sloan Kettering
Institute, New York, New York 10021 and the Department of
Neurology and Neuroscience and the ¶ Department of Cell Biology
and Anatomy, Division of Hematology/Oncology, Cornell University
Medical College, New York, New York 10021
The generation of different glial cell types in the central nervous system depends upon a wide variety of proliferative and differentiative signals. Here we report that changes in the levels of cyclin-dependent kinase 2 (CDK2) and the cell cycle inhibitor p27kip1 accompany the differentiation of central glia-4 (CG-4) progenitor cells to an astrocytic cell phenotype in the presence of fetal calf serum. Although a decrease in CDK2 levels was observed in both oligodendrocyte and astrocyte cells derived from CG-4 cells, a striking increase in the levels of p27 was observed during the differentiation of astrocyte cells. In astrocyte cell extracts, inhibition of CDK2 activity could be overcome with exogenously added cyclin E. Furthermore, depletion of p27 from astrocyte extracts lowered the amount of cyclin E required for CDK2 activation. Taken together, these results suggest that the inhibitory action of p27 upon cyclin E-CDK2 may prevent entry of cells into the S phase and regulate the progression of CG-4 cells toward an astrocytic lineage.
Glial cell development is marked by cell proliferation, growth arrest, cellular specialization, and programmed cell death (1, 2). The mechanisms that dictate glial cell differentiation and survival depend upon signal transduction events that are intrinsic to each cell type and are derived from extracellular growth factor signals (3). A major challenge has been to explain how similar intracellular molecules are activated to give diverse and opposing biological responses. A crucial regulatory mechanism to ensure appropriate cell differentiation involves the decision whether to proceed through the cell cycle or to undergo cell cycle arrest during the G1 phase (4).
A useful model for studying glial cell development has been the O-2A1 bipotential progenitor cell, which possesses the capability of differentiating into either type 2 astrocytes or oligodendrocytes (5). In the presence of growth factors such as fibroblast growth factor and platelet-derived growth factor, O-2A precursor cells continue to divide. In the absence of mitogens, O-2A cells differentiate into oligodendrocytes. These progenitor cells can also differentiate into type 2 astrocytes, a cell type that is not observed in vivo but that most resembles protoplasmic astrocytes.
To address cell cycle components that may influence glial cell development, a bipotential oligodendrocyte precursor cell line, CG-4, has been utilized as a model system. This cell line was established from primary cultures of postnatal rat oligodendrocytes and resembles O-2A progenitor cells (6, 7). Growth in conditioned media derived from the B104 neuroblastoma line promotes proliferation of CG-4 cells. Upon withdrawal of B104 conditioned media, CG4 cells stop dividing and differentiate into mature oligodendrocytes. Interestingly, treatment of CG-4 cells with high levels of serum results in differentiation to type 2 astrocytes in vitro.
We have focused upon changes in the late G1 phase of the cell cycle that accompany CG-4 cell differentiation. The results indicate that increases in the level of the cell cycle inhibitor p27 can block activity of cyclin E-CDK2 complexes when CG-4 cells are induced to differentiate into astrocytes. These findings suggest that glial cell growth arrest and differentiation may be regulated by specific cell cycle inhibitors, working on key cyclin-CDK complexes during the G1 phase of the cell cycle.
Reagents
Antibodies against CDK2, CDK4, CDK5, CDK6, cyclin E, cyclin A, and cyclin D3 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibody against cyclin D1 and D2 was obtained from Oncogene Science (Manhasset, NY). Antibody against p21waf1 was obtained from Pharmingen Biotechnology, Inc. (San Diego, CA). Antibody against p27kip1 was generated by immunizing a rabbit with a mouse p27 fusion protein (8). Recombinant proteins were prepared using baculovirus-infected SF9 insect cell lysates overexpressing CDK2, cyclin E, or cyclin E-CDK2 complex as described previously (9).
Cells and Cell Culture
B104 Cell Conditioned MediumB104 neuroblastoma cells were
maintained in the logarithmic phase of growth in DMEM (Life
Technologies, Inc.) supplemented with 10% heat-inactivated fetal calf
serum (Life Technologies, Inc.), 2 mM glutamine, and 100 units/ml penicillin/streptomycin. For conditioned medium production,
cultures (upon attaining a density of 100 cells/mm2) were
washed two times with Hanks' salt solution (Life Technologies, Inc.)
and then incubated in serum-free DMEM containing 2 mM
glutamine, 100 units/ml penicillin/streptomycin, and N1 supplement (50 µg/ml transferrin, 5 µg/ml insulin, 100 µM
putrescine, 20 nM progesterone, 20 nM selenium,
and 10 ng/ml biotin). Three days later the medium was removed, filtered
(Nalgene 0.45-µm filter), and stored at 20 °C.
CG-4 precursor cells were grown in DMEM supplemented with 30% B104 conditioned medium, N1 supplement, 2 mM glutamine, 100 units/ml penicillin/streptomycin, 5 µg/ml insulin, and 10 mM HEPES buffer, pH 7.5. Cells were passed every 5 days and plated on polyornithine-coated dishes at a density of 100 cells/mm2. To induce differentiation into astrocytes, CG-4 cultures were exposed 24 h after passage to DMEM supplemented with 2 mM glutamine, 100 µg/ml penicillin/streptomycin, and 20% fetal calf serum. To induce differentiation into oligodendrocytes, cultures were first washed with Hanks' salt solution and then incubated with DMEM containing N1 supplement, 2 mM glutamine, 100 units/ml penicillin/streptomycin, and 5 µg/ml insulin.
Immunocytochemistry
Cells were fixed for 10 min at room temperature in 4% paraformaldehyde in PBS (pH 7.4) and then blocked in normal serum for 1 h at room temperature. Incubation in primary antibody was carried out overnight at 4 °C at the following dilutions: 1:1 in PBS for anti-galactocerebroside and A2B5 staining; 1:2000 in PBS + 0.3% Triton X-100 + 3% normal goat serum for glial fibrillary acidic protein (Dako) staining; and 1:1000 in PBS + 0.3% Triton X-100 + 3% normal horse serum for myelin basic protein (Boehringer Mannheim) staining. After 1 h of incubation at room temperature in secondary antibody solution (1:500 in PBS), cells were stained using Vectastain Elite ABC according to the manufacturer's instructions and visualized using diaminobenzidine as substrate.
Bromodeoxyuridine Incorporation
Cells were given a 5-h pulse with 10 µM BrdUrd and then fixed in ice-cold acetone:methanol (1:1). After 30 min in 1 N HCl, cells were incubated in anti-BrdUrd; antibody (Dako; 1:100) in PBS + 0.3% Tween 20 + 3% horse serum at 4 °C overnight. After 1 h of incubation at room temperature in secondary antibody solution (1:500 in PBS), cells were stained using Vectastain Elite ABC according to the manufacturer's instructions and visualized using diaminobenzidine as substrate.
Western Blotting
Cells were detached from culture plates by scraping and then
collected in Hanks' salt solution. Cells were washed in PBS and then
resuspended in HKM buffer (30 mM HEPES-KOH, pH 7.4, 7.5 mM MgCl2), 0.5 mM dithiothreitol, 2 mM phenylmethylsulfonyl fluoride, 5 µg of leupeptin, and
3.5 µg/ml aprotinin. Cells were lysed by sonication and centrifuged
to separate the cell pellet. Protein extracts were adjusted to 0.1 N NaCl and then stored at 70 °C.
Protein extracts (20-100 µg/lane) were analyzed by 12.5% SDS-PAGE. Protein was transferred from the gel to a polyvinylidene fluoride membrane by a transfer apparatus at 34 V for 6 h. The membrane was then blocked with 5% nonfat milk (Carnation) and incubated with primary antibody against cyclin, CDK, or inhibitor. After incubating with an anti-rabbit or anti-mouse horseradish peroxidase-conjugated secondary antibody (Boehringer Mannheim), protein was visualized using an enhanced chemiluminescence system (Amersham Corp.).
Histone H1 Kinase Assay
Protein extracts were obtained from precursor CG-4 cells, oligodendrocytes, or astrocytes by sonication as above. These extracts (40 µg/sample) were incubated for 30 min at 37 °C in the presence of assembly buffer containing 30 mM HEPES-KOH, pH 7.5, 7 mM MgCl2 (10), and (depending on the experiment) with the addition of exogenous recombinant cyclin E, CDK2, or cyclin E-CDK2 complex.
Addition of physiologic levels of exogenous recombinant cyclin E during the assembly reaction allowed better detection of the kinase reaction without changing the intrinsic activity of the protein extracts (10). Addition of supraphysiologic levels of cyclin or CDK2 was performed during the assembly reaction in certain experiments in an attempt to overcome inhibition of CDK2 activity. Physiological levels of cyclin E and CDK2 were determined by immunoblotting CG-4 precursor extracts against varying amounts of recombinant cyclin E or CDK2 obtained from SF9 insect cell lysates overexpressing these proteins (11).
The assembled extracts were then subject to immunoprecipitation in
0.5% Nonidet P-40-radioimmune precipitation buffer (50 mM
Tris, pH 7.4, 250 mM NaCl, 0.5% Nonidet P-40, 5 mM EDTA) for 1 h at 4 °C in the presence of
antibody against CDK2 followed by a 1-h incubation with immobilized
protein A-Sepharose beads (Pharmacia Biotech, Inc). The beads were then
washed twice with Nonidet P-40-radioimmune precipitation buffer and
four times with kinase assay buffer (20 mM Tris-HCl, pH
7.4, 7.5 mM MgCl2, 1 mM dithiothreitol). Phosphorylation of histone H1 kinase was performed by
incubating these beads in a 50-µl reaction mixture containing 10 µCi of [-32P]ATP (DuPont NEN), 30 µM
lithium ATP (Boehringer Mannheim), and 1 µg of histone H1 (Boehringer
Mannheim) at 37 °C for 30 min. After incubation, the samples were
boiled in SDS-PAGE sample buffer and then resolved by 12% SDS-PAGE.
The gel was dried and subject to autoradiography. Radiolabeled histone
H1 bands were then quantitated using a PhosphorImager (Molecular
Dynamics, Sunnyvale, CA) and associated ImageQuant software.
Immunodepletions/Immunoprecipitations
Saturating levels of antibody to p27kip1 (20 µl) or a control antibody (against the Fas antigen) was adsorbed to protein A-Sepharose at 4 °C for 20 min. Protein extract (500 µg) from astrocytes was then added and the mixture allowed to incubate for 30 min. The supernatants were removed and then incubated a second time to freshly prepared antibody-protein A-Sepharose beads. These resulting supernatants, cleared of p27 protein, were then utilized for Western blotting and kinase assays.
Immunoprecipitation with CDK2 antibody, followed by histone H1 kinase assays, was performed on astrocyte extracts. Astrocyte extracts (300 µg) were immunoprecipitated with CDK2 antibody (20 µl) and then washed with Nonidet P-40-radioimmune precipitation buffer + 1 N NaCl solution. To the resulting immunoprecipitates was added exogenous cyclin E at 0, 1, and 3 times physiologic levels, or exogenous cyclin D2 (as a control). This mixture was incubated in the presence of an assembly buffer (see above) for 30 min at 37 °C and then washed with Nonidet P-40-radioimmune precipitation buffer followed by kinase assay buffer. A histone H1 kinase reaction was then performed as above.
In culture, O-2A progenitor cells differentiate into either oligodendrocytes or astrocytes, suggesting that these cell lineages arise from a common precursor (5). To characterize these lineage-specific events on a biochemical level, we chose a cell line that mimics primary cultures of O-2A cells. CG-4 cells, which spontaneously arose in O-2A cultures, are capable of expansion in culture (6). CG-4 cells can be maintained in an undifferentiated state in the presence of DMEM supplemented with either B104 conditioned medium or with the mitogens basic fibroblast growth factor and platelet-derived growth factor. Under these conditions more than 95% of the cells remain in the cell cycle and express the antigen A2B5 (6). The ability to propagate CG-4 cells allows for the generation of large numbers of cells for protein analysis.
Withdrawal of conditioned medium or addition of serum resulted in
terminal exit from the cell cycle and assumption of differentiated phenotypes. In the absence of serum, most CG-4 cells withdraw from the
cell cycle and express oligodendrocyte markers such as galactocerebroside (Fig. 1), and myelin basic protein
and lose A2B5 reactivity. In the presence of serum, greater than 95%
of the CG-4 cells withdraw from the cell cycle and express the
astrocyte-specific marker glial fibrillary acidic protein (Fig. 1) and
also the A2B5 antigen. Furthermore, the morphology of each cell type is
markedly different. CG-4 cells are typically bipolar in shape, whereas oligodendrocytes express a very branched morphology. Astrocytes are
flatter in shape and display fibrous processes. These distinct phenotypes were another indication of differentiation of these cell
types.
In later passages of CG-4 cells, we have observed that CG-4 cell differentiation along the oligodendrocyte lineage was compromised. The extent of oligodendrocyte differentiation varied between 40 and 80%. Astrocyte differentiation was unaffected. To determine if withdrawal from the cell cycle was affected in later passages, we determined the incorporation of bromodeoxyuridine in cultures of cells grown for 4 days in differentiation medium. Whereas BrdUrd incorporation in astrocyte cultures was reduced more than 20-fold compared with precursor cells, only a 2-fold change was observed in late passage oligodendrocyte cultures (Table I). Therefore, in this culture system astrocyte differentiation was correlated tightly with withdrawal of the cell cycle as measured by BrdUrd incorporation, whereas oligodendrocyte differentiation displayed more variability depending upon passage number. These observations suggest the CG-4 cells may represent a useful model system to explore the biochemical basis during glial cell differentiation into astrocytes.
|
The ability to proceed along different differentiation pathways in the presence or absence of serum suggested that the status of cell cycle regulatory molecules might play an important role in cell fate determination. To explore this possibility, we examined the expression of both the positive regulators of cell cycle transitions, the cyclin and CDK proteins, and the negative regulators, the cyclin-dependent kinase inhibitors, as CG-4 cells withdrew from the cell cycle along different differentiation pathways. Protein extracts were isolated from proliferating CG-4 cells, oligodendrocytes, or astrocytes and subjected to Western blot analysis using antibodies specific for cyclins and CDK proteins.
We focused upon the proteins involved in the G1 phase of
the cell cycle. CG-4 cells were found to express all three D-type cyclins, both major forms of cyclin E, and cyclin A. Furthermore, these
cells express CDK4 and CDK2, the catalytic subunits of the G1 cyclins
(Fig. 2). In addition, exponentially growing CG-4 cells contain small amounts of the cell cycle inhibitor proteins
p21waf1/cip1 and p27kip1
(Fig. 2, Table II). The expression of other CDK4 and
CDK6 inhibitor proteins, such as p16INK4A, was
not reliably detected in these cells.
|
After 4 days in differentiation conditions, the expression of these cell cycle proteins was examined in oligodendrocytes and astrocytes. An example of results from CG-4 precursor cells compared to extracts isolated from differentiated oligodendrocytes and astrocytes is shown in Fig. 2. Little change in the amounts of cyclin D1, cyclin D2, cyclin D3, cyclin E, and p21 was observed in oligodendrocyte or astrocyte cultures. In contrast, there was a slight induction of CDK5 protein and a reduction in the amount of CDK2 and CDK4 proteins. An increase in the amount of cyclin A was detected in astrocytes and there was no change in oligodendrocyte cultures.
Intriguingly, we observed cell-specific alteration in the steady state
levels of p27 protein (Fig. 2). In astrocytes derived from CG-4 cells
under conditions of high serum, the amount of p27 was greatly
increased. In contrast, there was little change in the amount of p27 in
oligodendrocytes derived from CG-4 cells under serum-free conditions. A
time course of astrocyte cell differentiation indicated that along with
the increase in p27, a decrease in the amount of CDK2 protein was
observed (Fig. 3). Similar changes in the amount of CDK2
were observed in oligodendrocytes. No significant changes in cyclin E
were detected. These data imply that increased levels of p27 might
antagonize the growth-promoting effect of serum by acting as an
inhibitor of CDK2 and CDK4 at the same time that CDK2 protein levels
were found to decrease. Furthermore, these results suggest that p27
might be involved in the determination process of glial cell
differentiation.
p27 Suppresses CDK2 Activation in Astrocytes
The activity of cyclin E-CDK2 is maximal at the G1/S transition and is a target of the machinery that leads to cell cycle arrest. Therefore, we determined if p27 directly mediated the inactivation of CDK2 to permit differentiation along the astrocyte lineage. To measure CDK2 activity in extracts prepared from either astrocyte or oligodendrocyte cultures, we used histone H1 as a substrate. We detected approximately equal amounts of cyclin E and CDK2 in both astrocytes and oligodendrocytes (see Fig. 2). Strikingly, CDK2 activity was absent in astrocyte extracts. The loss of CDK2 activity might be due to intrinsic modifications of CDK2 that make it refractory to activation or to the presence of p27, which would block the activation of cyclin E-CDK2 complexes.
To determine if the CDK2 protein was refractory to activation, we immunoprecipitated CDK2 from the extracts and added recombinant cyclin E (and CDK activating kinase) to the immunoprecipitates. After a suitable interval allowing complex formation, we washed out the proteins not interacting with CDK2 and assayed the immunoprecipitates for histone H1 kinase activity. We found that CDK2 activity from astrocytes could be induced, suggesting that inhibition was not due to intrinsic changes in the CDK2 protein (data not shown).
To determine if a titratable inhibitor blocked CDK2 activation, we
added increasing amounts of recombinant cyclin E into extracts and
measured immunoprecipitable CDK2 activity. Addition of cyclin E to
oligodendrocyte extracts resulted in the induction of endogenous CDK2
activity (Fig. 4). This suggests that oligodendrocytes
are fully competent to assemble and activate CDK2. Thus, the partial withdrawal of these cells from the mitotic cycle may not be due solely
to inhibition of CDK2 activity. In contrast, the addition of cyclin E
to extracts derived from astrocytes was not capable of activating the
endogenous CDK2 protein (Fig. 4). A 16-fold molar excess of recombinant
cyclin E (compared to the level of endogenous cyclin E) was inefficient
in activating CDK2. This suggests that during differentiation along the
astrocyte pathway CDK2 is actively inhibited, presumably through the
induction and action of p27.
Overriding the Inhibitory Effect of p27
The p27 protein acts as a potent inhibitor of cyclin-CDK complexes but does not interact efficiently with either cyclins or CDKs alone (11, 12). Consequently, the inability of cyclin E to activate CDK2 in astrocytes might be explained by the decrease in the total amount of CDK2. In these extracts, CDK2 might be limiting so that only some of the cyclin E is bound to CDK2 (cyclin E is in excess). These cyclin E-CDK2 complexes would be subject to p27 binding and inhibition. In this model the amount of p27 present in these extracts will be sufficient to inhibit all CDK2 enzyme that is capable of forming complexes with the cyclins.
In order to test the relative levels and effects of cyclin E-CDK2 and
p27 proteins, we titrated increasing amounts of recombinant CDK2 with
cyclin E into astrocyte extracts and measured histone H1 kinase
activity. Mixing recombinant cyclin E and CDK2 in the absence of
astrocyte extracts allowed the assembly and activation of CDK2 activity
due to CDK activating kinase activities in the SF9 lysates (Fig.
5A). Increasing the amount of CDK2 leads to a
marked increase in the amount of cyclin E-dependent kinase
activity (Fig. 5A). A plateau of activity was eventually
reached beyond which the addition of CDK2 had no more effect because
the amount of cyclin E became limiting. Addition of increasing amounts
of astrocyte extract to a fixed amount of cyclin E-CDK2 complex
decreased the amount of CDK2 activity (Fig. 5B), verifying
the presence of an inhibitory activity in astrocyte extracts. Taken
together, these results confirm that astrocytes derived from CG-4 cells contain an inhibitory activity that blocks cyclin E-CDK2 complexes.
Immunodepletion of p27
Since p27 is the most obvious cell
cycle component that is increased during astrocytic differentiation, it
is a candidate for the cyclin E-CDK2 inhibitory factor. To determine
whether p27 was responsible for the inhibition of CDK2 in astrocyte
extracts, we depleted p27 from these extracts by immunoprecipitation
and measured the ability of recombinant cyclin E to activate the
endogenous CDK2 in the depleted lysates. Depletion of p27 was confirmed
by immunoblot of lysates from astrocyte extracts treated with anti-p27 antibodies (Fig. 6C). Immunoprecipitation of
p27 resulted in a 3-fold increase in the ability of recombinant cyclin
E to activate CDK2 in a histone H1 phosphorylation assay (Fig. 6,
A and B). Control depletions with an antibody
against the Fas antigen did not affect the degree of inhibition. This
experiment indicates that p27 acts as an inhibitor of the CDK2 during
astrocyte differentiation from CG-4 cells.
Glial progenitor cells such as O-2A proliferate in vivo at the time of birth and give rise to postmitotic oligodendrocytes over a period of several weeks. In vitro, O-2A cells stop dividing and differentiate into oligodendrocytes in serum-free medium, whereas differentiation into type 2 astrocytes depends upon extracellular signals such as ciliary neurotrophin factor or serum (13). In both cases oligodendrocytes and type 2 astrocytes are postmitotic (14). It has been postulated that proliferating O-2A cells possess an intrinsic program in which a limited number of cell divisions takes place before differentiation is initiated. Such a clock mechanism is supported by single cell culture experiments of oligodendrocytes (15); however, the contribution of cell cycle components to this differentiation pathway has not been fully defined.
In this study we used CG-4 cells, a clonal glial cell line with characteristics of O2-A progenitor cells (6). CG-4 cells, like O-2A cells, are capable of giving rise to oligodendrocytes and type 2 astrocytes, depending upon culture conditions, and are capable of myelination in vivo after implantation in the central nervous system (7).
The results reported here indicate that G1 cyclins and their associated catalytic subunits, CDKs, are expressed by CG-4 cells. These cells display abundant cyclin E-dependent CDK2 activity. However, upon differentiation into astrocytes, this activity is blocked by increasing levels of p27kip1 inhibitor protein. Interestingly, CDK2 activity measured in oligodendrocytes derived from CG-4 cells was not affected upon differentiation, consistent with only a modest decline in BrdUrd incorporation (Table I). The differential behavior of CDK2 enzymatic activity between oligodendrocytes and astrocytes produced from CG-4 cells might be a critical difference during the G1 phase of the cell cycle. We cannot evaluate at this time the contribution of p27 and CDK2 upon oligodendrocyte growth arrest due to the variability in CG-4 cell differentiation. The mechanism that accounts for oligodendrocyte growth arrest is not known but could involve members of other inhibitory proteins, the activity of the CDK activating kinase, or target a completely different G1-CDK complex (16).
CDK inhibitors such as p27 negatively regulate G1 phase
progression by acting on cyclin D-CDK4 complexes and cyclin E-CDK2 complexes. The binding of p27 appears to be stoichiometric and acts to
prevent threonine phosphorylation of cyclin E-CDK complexes (10) by CDK
activating kinase. The effect of p27 upon the activity of cyclin E-CDK2
complexes in CG-4 astrocytes is consistent with the general mechanism
of cell cycle arrest induced by cell-cell contact and transforming
growth factor- treatment of epithelial cells (11). The amount of
CDK4 associated with p27 in astrocytes is reduced in CG-4
cells.2 The level of CDK2 also decreased as
differentiation proceeded. CDK2 binding to cyclin E and its activation
by positive and negative phosphorylation events do not appear to be
influenced by the transition of CG-4 cells to astrocytes. The decrease
in CDK2 was strikingly accompanied by the accumulation of p27 in
astrocytes. Removal of p27 from astrocyte lysates resulted in a
demonstrable increase in CDK2 activity, indicating that the presence of
p27 has a direct effect upon the late G1 cell cycle activities of
cyclin E-CDK2.
As has been found for cells grown in primary culture, CG-4 astrocytic differentiation depends upon extracellular signals, whereas oligodendrocyte differentiation depends upon an intrinsic program. However, although astrocyte growth arrest could be easily achieved using CG-4 cells, a large percentage of oligodendrocytes did not withdraw from the cell cycle arrest under continuous culture conditions. Hence, the exact mechanisms of oligodendrocyte growth arrest are unclear. More experimentation into the cell cycle controls in oligodendrocytes and astrocytes from primary culture is needed. In addition, the use of cells derived from transgenic animals lacking the cyclin-dependent kinase inhibitors and the introduction of p27 in precursor cells will clarify the signals that dictate glial cell growth arrest and differentiation.
We thank J. deVellis for providing the CG-4 cells and J. Orlinick for the anti-Fas antibody.