Simultaneous suppression of cdc2 and cdk2 activities induces neuronal differentiation of PC12 cells.

The involvement of cdc2 and cdk2 during neuronal differentiation in rat pheochromocytoma PC12 cells was examined. When PC12 cells were cultured with nerve growth factor (NGF), expression of cdc2 decreased significantly after day 5, while expression of cdk2 decreased gradually after day 7. Cells overexpressing cdc2 or cdk2 were resistant to NGF-induced differentiation and growth suppression, and maintained high cdc2 or cdk2 kinase activity, respectively, during NGF treatment. In contrast, the NGF-treated parental cells showed a marked decline in these kinase activities after day 3. When PC12 cells were treated with specific inhibitors of cdc2/cdk2 (butyrolactone-I, olomoucin), they showed marked neurite extension and up-regulation of microtubule-associated protein 2 expression. In addition, treatment with mixtures of antisense oligonucleotides for cdc2 and cdk2 resulted in down-regulation of both cdc2 and cdk2 kinase activities as well as significant neurite outgrowth and up-regulation of microtubule-associated protein 2 expression. However, neurite outgrowth was not observed in cells treated with either single antisense oligonucleotide, or antisense cdc2 + cdk4 or cdk2 + cdk4 oligonucleotide mixtures. These results suggest that simultaneous down-regulation of cdc2 and cdk2 activity is sufficient and necessary for neuronal differentiation in PC12 cells.

The commitment to cellular differentiation is controlled by a series of strictly regulated steps, each requiring specific signals for survival and simultaneous loss of proliferative potential. The decision to proliferate or differentiate is made during the G 1 phase of the cell cycle by the interaction of G 1 cyclins, cyclin-dependent kinases (cdks) 1

and cdk inhibitors (1-4).
Cyclin-cdk complexes have been shown to play a crucial role in the G 1 /S transition, thus regulation of their functions is also critical for the commitment to cellular differentiation. Indeed, we have previously demonstrated that ectopic expression of cdk2, but not of cdk4 or any of the cyclins, inhibits NGFmediated neuronal differentiation of PC12 cells (5,6). Furthermore, ectopic expression of D-type cyclins or cdk4 was also reported to inhibit granulocyte-colony stimulating factor-induced differentiation of murine myeloid cells or erythroleukemia cells, respectively (7,8). In addition, cell differentiation has been reported to correlate with the suppression or activation of kinase activity in several cell lines, including those of the neuronal, hematopoietic, and myocytic lineages (9 -16). Taken together, it is clear that cdks play a key role in the switch between cell proliferation and differentiation, and that the responsible cdk varies with cell type. However, it still remains unclear whether a change in cdk activity directly induces neuronal differentiation or whether this change is secondary to differentiation.
We have been investigating the involvement of cell cycle regulator molecules in the neuronal differentiation of rat pheochromocytoma PC12 cells. PC12 cells are a good model of neuronal differentiation since they are particularly responsive to growth factors, such as nerve growth factor (NGF), basic fibroblast growth factor, and overexpression of oncogene products, v-src (17)(18)(19)(20)(21). In the previous study, we showed that suppression of cdk2, but not of cdk4 activity, is a crucial step in the neuronal differentiation of PC12 cells (5,6). Thus, in the present study, we investigated more specifically the roles of cdc2 and cdk2 during this process and, herein, demonstrate that suppression of both cdc2 and cdk2 activities induces neuronal differentiation of PC12 cells, independent of the differentiation inducing agents described above.

EXPERIMENTAL PROCEDURES
Cell Lines-PC12 cells were grown on plastic dishes with collagen coating in Dulbecco's modified Eagle's medium, containing 10% heatinactivated horse serum and 5% fetal bovine serum. For NGF treatment (20 or 50 ng/ml in final concentration) or induction of neuronal differentiation, cells were cultured in Dulbecco's modified Eagle's medium containing 5% horse serum and 2.5% fetal bovine serum. Cells having one or more neurites with the length more than twice the diameter of the cell body were defined as "differentiated" (5,22). For cell proliferation assay, cells were plated at 1 ϫ 10 5 cells/plate (9 cm in diameter) in triplicate on day 1. Twenty-four hours later (day 0), the cells were fed with fresh medium. The attached cells were harvested at day 0, 1, 3, and 7 and counted by using a FACScan flow cytometer (Becton & Dickinson). Each experiment was performed in triplicate.
Reagents-Reagents in this study were obtained as follows: puromycin was purchased from Sigma, NGF was from Wako Ltd. Butyrolactone-I (BL-I) was purified as described previously (23,24). Olomoucin (OLM) was provided by Dr. T. Akiyama, University of Tokyo.
Plasmids and Transfection-Human cdc2 and cdk2 cDNAs, subcloned into the pcDL-SR␣ 296 vector were provided by Dr. K. Sagawa, Keio University (25,26), and used for transfection into PC12 cells. PC12 cells (5 ϫ 10 6 cells) were transfected with 50 g of cDNA(s) along with 5 g of the puromycin resistance gene pcDNA-puro (Invitrogen). Transfection was achieved by electroporation with a Bio-Rad electroporator followed by selection for puromycin resistance (Life Technologies, Inc., 5 g/ml). Four weeks after transfection, 50 clones were isolated for each cDNA transfection and maintained in medium containing 3 g/ml puromycin. For each cDNA transfectant, the 3 clones showing the highest expression were selected by immunoblotting analysis and used for further experiments.
Antibodies-Monoclonal antibody against cdc2 and polyclonal antibody against cdk2 were purchased from Santa Cruz Biotechnology and monoclonal antibody against microtubule-associated protein 2 (MAP2, clones AP18 and AP20) was obtained from Lab Vision Corp.
Immunoblotting and Immunoprecipitation-For immunoblotting analysis, cells were lysed in high-salt lysis buffer as described previously (5). The protein concentration of each sample was determined using a protein assay kit (Bio-Rad), 50 g of protein was analyzed on a SDS-polyacrylamide electrophoresis gel (12% of acrylamide gel) for cdc2 and cdk2, 100 g of protein on a gradient gel of 3.5-15% (w/v) for MAP2 (27), and then transferred to a nitrocellulose membrane (Hybond-C, Amersham Pharmacia Biotech). Following blocking in skim milk for 3 h, the membrane was incubated with primary antibodies to human cdc2 (1:400 dilution), cdk2 (1:200), or MAP2 (mixture of AP 18 and AP20, each 1:200). Each protein was detected by sequential binding of a specific primary antibody followed by alkaline phosphatase-conjugated secondary antibody (1:6000 dilution, Promega).
In Vitro cdc2 Kinase Reactions-In vitro cdc2 kinase assay was performed with MESACUP cdc2 Kinase Assay Kit (MBL). This kit utilizes a synthetic peptide substance and a monoclonal antibody recognizing the phosphorylated form of the cdc2-specific peptide substance (antibody 4A4). In brief, cells were lysed in cdc2 sample buffer (50 mM Tris, pH 7.5, 5 mM EDTA, 0.5 M NaCl, 50 mM 2-mercaptoethanol, 0.01% Brij 35, 25 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 50 g/ml leupeptin) and the resultant lysates were sonicated on ice four times for 10 s each (31). Lysates were clarified by centrifugation at 10,000 ϫ g for 50 min and aliquots (20 g of protein) were incubated in a 50-l reaction solution (25 mM HEPES, pH 7.5, 10 mM MgCl 2 , 0.1 mM ATP) containing biotinylated peptide (SLYSSSPG-GAYC) for 30 min at 30°C. After terminating the reaction with 200 l of phosphorylation stop reagent (50 mM EDTA in PBS), 100 l of the reaction mixture was transferred to microwells coated with monoclonal antibody, 4A4, and incubated for 60 min at 25°C. Then, 100 l of peroxidase-conjugated streptavidin was added to the microwells and the reactions were incubated for 30 min followed by incubation with 100 l of peroxidase-substrate solution (5.3 mg/ml o-phenylenediamine in H 2 O 2 ) for 5 min. After terminating the reaction with 20% H 3 PO 4 , the absorbance of each well was read at 492 nm with a microplate reader (Bio-Rad). Independent experiments were performed in triplicate.
In Vitro cdk2 and cdk4 Kinase Reactions-Cells were lysed in solubilizing buffer for cdk2 and in immunoprecipitation buffer for cdk4 as described previously (5,32). Lysates (100 g of protein) were incubated with anti-cdk2 antibody (diluted 1:150) or anti-cdk4 antibody (1:100) for 1 h, followed by an additional 1-h incubation with protein A-Sepharose beads at 4°C. A bacterially expressed fragment of the retinoblastoma protein (pRB) fused to glutathione S-transferase was purified by affinity chromatography on glutathione-Sepharose and used as a substrate (0.5 g of protein) in 50 l of kinase reaction buffer containing 20 mM [␥-32 P]ATP (5 Ci; 1 Ci ϭ 37 kBq) (5). After incubation for 10 min at room temperature, the sample was analyzed by SDS-polyacrylamide gel electrophoresis followed by autoradiography. Independent experiments were performed in triplicate.
Densitometric Analysis-Densitometric quantification of the data obtained by immunoblotting analysis (nitrocellulose filters) and in vitro kinase assays (x-ray films) were done using a GT6500ARTS scanner (Epson) and all scans were incorporated into NIH images (version 1.56) for densitometric analysis (33). Since more than one band was detected in the cdc2 and cdk2 immunoblots, the most rapidly migrating bands were subjected to quantification.

RESULTS
Expression of Endogenous cdc2 and cdk2 during Differentiation of PC12 Cells-PC12 cells were cultured in the presence FIG. 1. cdc2 and cdk2 protein levels as well as MAP2 expression during NGF-mediated neuronal differentiation of PC12 cells. Lysates were prepared from cells that had been cultured in medium containing 20 ng/ml NGF, harvested at the indicated times, and subjected to immunoblotting analysis to detect cdc2, cdk2, and MAP2. The positions of the cdc2 and cdk2 doublets as well as that of MAP2 are indicated by arrows. The intensity of the fast migrating bands of cdc2 and cdk2, and a single band of MAP2 were quantitated and expressed as ratios relative to those obtained on day 0.

FIG. 2. Forced expression of ectopic cdc2 (a) and cdk2 (b) in PC12 cells.
Representative clones and control clones (parental and vector-introduced cell lines) were lysed as described under "Experimental Procedures." Lysates (50 g of protein) were subjected to immunoblotting analysis to detect cdc2 (a) and cdk2 (b). The positions of the cdc2/cdk2 doublet are indicated by arrows. The intensity of the fast migrating bands were quantitated and expressed as ratios relative to those obtained from parental cells. of 5% horse serum, 2.5% fetal bovine serum, and 20 ng/ml NGF for 10 days. At days 0, 1, 3, 5, 7, and 10, cells were lysed and the expression of cdc2 and cdk2 was evaluated by immunoblotting analysis using specific antibodies. Changes in MAP2 expression were also examined to confirm the neuronal differentiation of PC12 cells induced by NGF (27,34). As shown in Fig. 1, the expression of MAP2 increased gradually, reaching 1.9-fold higher levels through the 10 days of observation. In contrast, the expression of cdc2 decreased significantly after 5 days to 50% of the initial level and decreased gradually thereafter to 12% by day 10. The expression of cdk2 was more or less stable for the first 5 days, then decreased to 60% of the initial level after 7 days, and to 25% after 10 days.
Introduction of cDNAs Encoding Human cdc2 and cdk2-Expression constructs containing human cdc2 or cdk2 cDNAs were transfected into PC12 cells, and clonally derived cell lines expressing significantly higher levels of each protein were obtained. Immunoblotting analysis of those cell lines showed that the bands corresponding to ectopically overexpressed cdc2 and cdk2 proteins were almost identical in size to those of the endogenous proteins (Fig. 2). The highest levels of the overexpressed cdc2 and cdk2 proteins were, approximately, 6.8-and 6.1-fold higher, respectively, than those of the endogenous proteins.
Morphological Changes Induced by NGF-mediated Differen- for the cdk2-expressing transfectants, are shown since all other transfectants showed similar morphology. tiation in Transfected Cells-All of the established cell lines overexpressing cdc2, cdk2, or transfected with the empty vector showed morphologies very similar to the parental PC12 cells in the absence of NGF (Fig. 3a). When the cells were stimulated with NGF (50 ng/ml), approximately 70% of the parental cells and 60% of the vector-transfected cells displayed neurite outgrowths by day 3 (Fig. 3b). Neurite outgrowths were the most prominent at day 7 in those two cell lines (Fig. 3c). By contrast, each of the three independent cell lines overexpressing cdc2 or cdk2 did not show any significant neurite extension throughout the 7 days of observation (Fig. 3c).
Effects of NGF Stimulation on the Growth of PC12 Cells-NGF-induced differentiation of PC12 cells is accompanied by a decline in the growth rate (17). Thus, the effects of cdc2 or cdk2 overexpression on the growth of PC12 cells were examined by culturing these cell lines in the absence or presence of NGF (50 ng/ml). In the absence of NGF, all of the cell lines exhibited similar growth rates (Fig. 4a). In the presence of NGF, the parental and vector-transfected cells showed marked growth suppression after 3 days of culture, while cell lines overexpressing cdc2 (clones cdc2-4, -9, and -25) exhibited no NGF-mediated growth suppression and no significant reduction in the growth rate until day 7. The cell lines overexpressing cdk2 (clones cdk2-6, -7, and -14) also did not show any significant decline in the growth rate (Fig. 4b).
Changes in cdc2 Kinase Activity in Parental and Transfected Cells-The experimental results described above suggested that suppression of cdc2 and cdk2 kinase activities could be a crucial factor in neuronal differentiation. We, therefore, examined the changes in cdc2 kinase activity that occurred during NGF-induced differentiation (20 ng/ml NGF). cdc2 kinase activity was assayed by its ability to phosphorylate a cdc2-specific synthetic oligopeptide. As shown in Fig. 5a, a rapid decline in kinase activity to about 51% of the starting levels was detected in parental cells at day 3, 14% at day 7, and 2.5% at day 10. An almost identical time course was observed in vector-transfected cells (Fig. 5b). However, in the cell lines overexpressing cdc2 (clones cdc2-4, -9, and -25) which initially showed 1.4-fold higher level, cdc2 kinase activity slightly decreased but was still maintained at 1.2-fold higher levels of the parental cells after 10 days (Fig. 5c, only the data obtained from one cell line, cdc2-25, is shown). In contrast, in cell lines overexpressing cdk2, cdc2 activities decreased to 54, 16, and 6% of the initial levels at day 3, 7, and day 10, respectively (Fig. 5d, only the data obtained from one cell line, cdk2-6, is shown).
Changes in cdk2 Kinase Activity in Parental and Transfected Cells-We next examined the changes in cdk2 kinase activity in lysates of these cell lines during NGF-induced differentiation (20 ng/ml NGF). cdk2 immunoprecipitates were assayed for their ability to phosphorylate a fragment of recombinant pRB (5). As shown in Fig. 6, a rapid decline in kinase activity to about 7 and 18% of the starting levels was detected in parental cells and a vector-transfected cell line, respectively, at day 3. In cell lines overexpressing cdc2, cdk2 activity decreased to 14 and 0% of the initial levels at days 3 and 7, respectively ( Fig. 6; only the data obtained from one cdc2-overexpressing cell line, cdc2-25, is shown). However, cdk2 kinase activity in all of the cell lines overexpressing cdk2 remained high; 1.9-fold of that in the parental cells at the initial level and 1.3-fold at day 10 ( Fig. 6).
Changes in Morphology and Protein Expression Induced by Specific Inhibitors of cdc2/cdk2-Since our results indicate that overexpression of cdc2 or cdk2 can inhibit PC12 neuronal differentiation, we next examined whether the cdc2/cdk2 inhibitors BL-I or OLM could induce differentiation in these cells. Both of these agents have been previously demonstrated to suppress cdc2/cdk2 activity (23,24,35). In addition, changes in MAP2 expression were also examined to confirm the neuronal differentiation of PC12 cells by NGF (27,34). A control group, treated with 0.1% dimethyl sulfoxide alone did not show significant neurite outgrowth and expression of MAP2 was much lower than in cells treated with NGF (50 ng/ml) for 3 days (Fig.  7, a and d). When PC12 cells were cultured in 20 -80 g/ml BL-I, cells displayed prominent neurite extension by 3 days and MAP2 expression was significantly higher than the control group and comparable to cells treated with NGF (Fig. 7, b and  d). This effect of BL-I was optimal at a concentration around 50 g/ml. When cells were treated with more than 80 g/ml BL-I, neurite extension was unclear, probably due to the cytotoxic effect of the drug. Treatment with OLM at concentrations between 30 and 50 M also caused neurite extension in PC12 cells (Fig. 7c). In these cells, MAP2 was much higher than that of control cells, reaching levels almost comparable to the NGFtreated group (Fig. 7d). OLM at concentrations less than 20 M did not show any significant effect, while treatment at more than 60 M caused shrinkage of the cytoplasm which obscured clear formation of the neurites.
Effect of cdc2, cdk2, and cdk4 Antisense Oligonucleotides on Morphology and Neuronal Phenotype-To determine whether the induction of neuronal differentiation by BL-I and OLM is the result of the suppression of cdc2 and/or cdk2 activity, PC12 cells were treated with antisense oligonucleotides. As controls, FIG. 5. cdc2-associated kinase activities. cdc2-associated kinase activity in parental (a), vector-transfected (b), and derivative clones overexpressing cdc2 (c) or cdk2 (d) during NGF treatment (20 ng/ml) are shown. Lysates prepared at the indicated times were subjected to cdc2 kinase assay as described under "Experimental Procedures." The histograms were generated by measuring the absorbance at 492 nm at the indicated times. All values were obtained from independent triplicate experiments and are expressed as mean ratios Ϯ two standard deviations relative to that of parental cells on day 0. Only the data obtained from the representative clones cdc2-25 and cdk2-6 are shown because all 3 cdc2-or cdk2-overexpressing cells revealed almost similar results. cdk4 sense and antisense oligonucleotides were also tested. When cells were cultured in the presence of any of the sense oligonucleotides or cdk4 antisense oligonucleotides, no neurite extension was observed at day 3 (Fig. 8a). When cultured with cdc2 or cdk2 antisense oligonucleotides, cells generated short processes but no apparent neurites through 3 days of treatment (Fig. 8a, lower row). Next, cells treated with mixtures of any two sense oligonucleotides or antisense cdc2 and cdk4, or cdk2 and cdk4 oligonucleotides generated occasional short processes but no significant neurite extension (Fig. 8b). However, treatment with a mixture of cdc2 and cdk2 antisense oligonucleotides, resulted in the marked formation of significant neurite extensions at day 3 (Fig. 8b, lower row). Accordingly, only cells treated with a mixture of cdc2/cdk2 antisense oligonucleotides showed higher expression of MAP2 protein, similar to the control NGF-treated group (Fig. 8c).
Changes in cdc2, cdk2, and cdk4 Kinase Activity after Treatment with Oligonucleotides-The experimental results described above suggested that a reduction in cdc2 and cdk2, but not in cdk4 kinase activities, correlates well with the extent of morphological neuronal differentiation. We therefore examined changes in cdc2 and cdk2 as well as cdk4 kinase activity to confirm that their kinase activities were suppressed by the corresponding oligonucleotides and that the kinase activities of cdc2 and cdk2 correlated with neurite extension induced by antisense oligonucleotides. In extracts from cells treated with sense oligonucleotide alone, cdc2, cdk2, and cdk4 kinase activity declined slightly to about 80% of the control cells treated FIG. 6. cdk2-associated kinase activities. cdk2-associated kinase activity in parental (PC12), vector-transfected (vector), and derivative clones overexpressing cdc2 (cdc2-PC12), or cdk2 (cdk2-PC12) during NGF treatment (20 ng/ml) are shown. Lysates prepared at the indicated times were immunoprecipitated with anti-cdk2 antibody. The immune complex was assayed for kinase activity using glutathione S-transferase-RB fusion protein as a substrate. The intensity of radioactive signals obtained from independent triplicate experiments was quantitated on x-ray film and is expressed as mean ratios relative to that of parental cells on day 0. Only the data obtained from the representative clones cdc2-25 and cdk2-6 are shown. with PBS (Fig. 9, a-c). In contrast, in extracts from cells treated with cdc2-antisense oligonucleotide alone, cdc2 kinase activity declined to about 8.2% of the control cells treated with PBS, while cdk2 and cdk4 activities were maintained at 84 and 81%, respectively, at day 3 ( Fig. 9, a-c). In cells treated with cdk2 antisense oligonucleotide alone, cdk2 activity declined markedly to 6.3% while cdc2 and cdk4 activities were maintained at 79 and 82% of the control level, respectively (Fig. 9, a-c). In cells treated with cdk4 antisense oligonucleotides, cdk4 activity decreased to 11%, while cdc2 and cdk2 activities declined slightly to 85 and 82%, respectively, of the control level (Fig. 9, a-c). These results suggest that each kinase activity was suppressed by the corresponding antisense oligonucleotide without significant cross-inhibition. Cells treated with mixtures of any of the sense oligonucleotides showed a slight decrease in all of kinase activities, to about 80% of the control cells treated with PBS alone (Fig. 9, d-f). However, cells treated with mixtures of cdc2 and cdk2 antisense oligonucleotides showed a marked decrease in both cdc2 and cdk2 activities, to 8.7 and 9.3%, respectively, while cdk4 kinase activity was maintained at 79% of the control levels ( Fig. 9, d-f). Mixtures of cdc2/cdk4 or cdk2/cdk4 similarly suppressed the activities of corresponding kinases without showing any significant inhibition of the other kinases (Fig. 9, d-f).

DISCUSSION
In the present study, we demonstrated the involvement of cdc2 and cdk2 in the differentiation of PC12 cells by overexpression or inhibition of these two molecules. Our results suggest that, both cdc2 and cdk2 are key regulators of neuronal differentiation in this cell type, as defined by neurite outgrowth and by up-regulation of MAP2 protein expression. We found that the activity of cdc2 and cdk2 decreased significantly following the addition of NGF, while constitutive overexpression of either cdc2 or cdk2 completely inhibited differentiation of PC12 cells. Furthermore, simultaneous suppression of FIG. 8. Changes of morphology and protein expression caused by specific antisense cdc2, cdk2, and/or cdk4 oligonucleotides. a and b, morphological changes in PC12 cells caused by treatment with sense or antisense oligonucleotides. Cells were cultured in the presence of the following oligonucleotides for 3 days. Sense oligonucleotides (a, upper row), antisense oligonucleotides for cdc2, cdk2, or cdk4 (a, lower row), or the mixtures of their sense oligonucleotides (b, upper row) and their antisense oligonucleotides, cdc2/cdk2, cdc2/cdk4, or cdk2/ cdk4 (b, lower row). c, the expression level of MAP2 was evaluated in the lysates obtained from the cells treated with sense or antisense oligonucleotides for 3 days described in a and b. Lysates obtained from cells treated with PBS (PC12) or NGF (50 ng/ml) for 3 days (NGF) were also indicated as control groups. The intensity of the bands of MAP2 was quantitated and expressed as ratios relative to that of the cells treated with PBS alone. both cdc2/cdk2, but not either alone, induced neuronal differentiation.
Cdc/cdk family kinase activity seems to be a critical determinant in switching between cell proliferation and differentiation, although reports describing their roles have varied, probably due to tissue or cell type-specific differences. In hematopoietic cell lineages, cdk4/cdk6 has been shown to be a major determinant in this switching, since forced expression of cdk4 inhibits differentiation in erythroleukemia cells and p16, one of the cdk4 inhibitors, has been reported to promote differentiation in acute lymphoblastic leukemia cells (8,15).
Although as a transformed cell line, PC12 cells may have aberrant cell cycle control mechanisms, suppression of cdk2 activity and sustained activation of the MAP kinase pathway have been described as crucial steps in neuronal differentiation (5, 22, 36 -38). The novel findings in this study add further weight to the idea that cdc2 and cdk2 kinase activities are key to the differentiation of PC12 cells. cdc2 and cdk2 probably exert their inhibitory effects on neuronal differentiation by forming complexes with endogenous cyclins (1,39,40). Our results suggest that it is the levels of cdc2 and cdk2 that are limiting, and not those of the cyclins. Thus, these kinases function in a dose-dependent manner: endogenous levels of cdc2/cdk2 are sufficient to promote cell proliferation and inhibit spontaneous differentiation in the absence of NGF, but not sufficient to inhibit differentiation in the presence of NGF. Suppression of cdc2/cdk2 kinase activity allows cells to spontaneously differentiate even in the absence of NGF. Conversely, overexpression of cdc2/cdk2 renders cells resistant to differentiation even in the presence of NGF, and they continue to proliferate.
Although overexpression of either cdc2 or cdk2 is sufficient to FIG. 9. Changes in cdc2, cdk2, and cdk4 kinase activities after treatment with oligonucleotides. The changes in cdc2, cdk2 and cdk4 kinase activity of the cells treated with each oligonucleotide. a-c, PC12 cells were treated with PBS, sense or antisense oligonucleotide for cdc2, cdk2, or cdk4 for 3 days. cdc2 kinase activity was assayed using MESACUP kit (a) and cdk2 or cdk4 kinase activities were assayed using glutathione S-transferase-RB fusion protein substrate (b and c). d-f, PC12 cells were treated with PBS, a mixture of sense or antisense oligonucleotides for cdc2, cdk2, and cdk4 for 3 days, and cdc2, cdk2, and cdk4 kinase activities were assayed. For cdc2 kinase activity, the absorbance values at 492 nm obtained from independent duplicate experiments were measured and indicated as mean ratios relative to those of the sample treated with PBS, with ranges of measured values (a and d). For cdk2 and cdk4 kinase activities, the intensity of radioactive signals on the x-ray films obtained from independent duplicate experiments was quantitated and expressed as mean ratios relative to those of the samples treated with PBS (b, c, e, and f).
inhibit differentiation, suppression of both cdc2 and cdk2 are necessary to induce differentiation. These results imply that the signal cascade determining neuronal differentiation in PC12 cells is negatively regulated by these two redundant kinases. However, cdk4 does not seem to be critically involved in this cascade, as shown by this study and by our previous results which demonstrated that constitutive overexpression of cdk4 did not prevent differentiation of PC12 cells (6). As for other cyclins, we have previously demonstrated that overexpression of cyclins A, D1, or E had no effect on inhibiting NGF-mediated differentiation (5).
The fact that suppression of cdc2 and cdk2, but not cdc2 and cdk4 or cdk2 and cdk4, is critical in this process, is noteworthy, although the reasons for this selectivity is unclear. cdc2 and cdk2 phosphorylate several cellular proteins crucially involved in cell cycle regulation which would otherwise prevent cell cycle transition and possibly induce further differentiation (2, 4). One candidate substrate for these kinases is pRB, which is phosphorylated by cdk2 and cdk4 (4,41) and which in addition, plays a crucial role in early neuronal and hematopoietic development (42)(43)(44). Thus, NGF-induced down-regulation of cdc2 or cdk2 activity could result in the accumulation of hypophosphorylated pRB. Accordingly, constitutive overexpression of cdc2/cdk2 may block differentiation by driving the phosphorylation of pRB. However, it may be difficult to explain our results solely by inhibition of pRB hyperphosphorylation: treatment of cells with antisense oligonucleotides for cdk2 and/or cdk4, did not cause neurite outgrowth, despite the fact that both kinases phosphorylate pRb (4,41).
Alternatively, the fact that only the combination of antisense oligonucleotides for cdc2 and cdk2 induced differentiation may suggest that inhibition of kinases in G 1 , S and G 2 /M phases is important for cellular differentiation. Simultaneous inhibition of cdc2 and cdk2 may restrain the cell cycle and keep the cells in a static condition.
One important mechanism of cyclin/cdk inactivation in neuronal cells involves inhibition by p21 and p27 (2,3). Which cdk inhibitor is responsible depends on the cell type; p21 predominates in PC12 cells (45), but p27 predominates in oligodendrocyte (9, 13) and immortalized hippocampal cells (10). These differences may be partially explained by the fact that p21 and p27 sometimes show reciprocal expression and also function in a redundant manner. For example, in certain kinds of human malignant tumors and in p21-deficient mice or p15-deficient human melanoma cells, only p21 or p27 are expressed and one compensates for the other to inhibit cdk (46 -48). Although induction of differentiation by transfection of a cdk inhibitor alone has not been observed, induction of p21 was reported to enhance susceptibility of PC12 cells for NGF-induced differentiation (5,11,12). Probably, stable cell lines in which p21 was inducibly overexpressed, may not be able to cause sufficient suppression of cdc2/cdk2 activity to trigger spontaneous differentiation. Again, there may be a dose-dependent effect of cdc/ cdk2 suppression by inhibitor: sufficient suppression of cdc2/ cdk2 activity triggers spontaneous differentiation even in the absence of NGF, and insufficient inactivation of cdk(s) only commits neuronal cells to a sensitive state for differentiation. Another cdk inhibitor, the INK4 family, have rarely been reported to be involved, consistent with the fact that their target protein, cdk4, is not crucially involved in induction of neuronal differentiation as described previously (2,3,6,49).
In contrast, a recent report has shown that activation of cdk5 induced neurites outgrowth in hippocampal cell lines, indicating that cdk5 activation is a key factor in the differentiation of several kinds of neuronal cells (10). Since PC12 cells are transformed cells, and may have aberrant cell cycle properties, this model does not necessarily reflect the general nature of neuronal cells. We also cannot exclude the possibility that cdk5 is also involved in the differentiation of PC12 cells, since some of the inhibitors of cdc2/cdk2 kinase activity also affect this kinase (50,51). However, taken together, our results strongly suggest that cdc2/cdk2 play a key role in the differentiation of PC12 cells: differentiation is inhibited when these kinases are overexpressed, and differentiation is induced when these kinases are inhibited by antisense oligonucleotides or chemical reagents.
We think the next step in understanding the mechanism of neuronal differentiation in PC12 cells will require the identification of those cdc2/cdk2 substrates whose phosphorylation directly affects cell proliferation/differentiation.