Proliferation of oligodendrocyte progenitor (OP) cells is a crucial process controlling myelination in the CNS. Previous studies demonstrated a correlation between OP proliferation rate and cyclin E/cyclin-dependent kinase-2 (cdk2) activity. To establish a causal link between cyclin E/cdk2 activity and OP proliferation, we selectively modulated cdk2 activity in vitro by transfection of cultured OP cells. Dominant-negative (Dn)-cdk2 overexpression inhibited mitogen-induced OP cell proliferation, whereas wild-type (wt)-cdk2 prevented cell cycle arrest caused by anti-mitotic signals. Dn-cdk2- or wt-cdk2-mediated regulation of G1/S transition, per se, did not influence initiation of OP differentiation. To study the function of cyclin E/cdk2 in OP cells during development in vivo, we analyzed cdk2 and cyclin E expression in cells acutely isolated from transgenic mice expressing the green fluorescent protein (GFP) under the control of the 2′-3′-cyclic nucleotide 3′-phosphodiesterase gene promoter. Both cyclin E/cdk2 protein levels and activity were decreased in GFP+oligodendrocyte lineage cells between postnatal days 4 and 30. Immunostaining of NG2+/GFP+ OP cells in brain tissue sections showed a 90% decrease in overall cell proliferation and cdk2 expression between perinatal and adult cells. However, cdk2 expression within the proliferating (i.e., expressing the proliferating cell nuclear antigen) OP cell population was maintained throughout development. Our data indicate that: (1) cyclin E/cdk2 activity plays a pivotal function in OP cell cycle decisions occurring at G1/S checkpoint; (2) initiation of OP differentiation is independent of cyclinE/cdk2 checkpoint, and (3) intrinsic differences in cyclin E/cdk2 expression and activity may underlie the slowly proliferative state that characterizes so-called “quiescent” adult OP cells in vivo.
- cell cycle
- cell differentiation
- checkpoint genes
- fluorescence-activated cell sorting
In the oligodendroglial lineage, oligodendrocyte progenitor (OP) cells undergo a limited period of proliferation during development, before they exit the cell cycle and terminally differentiate into myelinating oligodendrocytes (Temple and Raff, 1986). Nevertheless, some oligodendrocyte precursor cells have also been shown to lack replicative senescence in vitro and to proliferate indefinitely in the presence of mitogens and in the absence of hydrophobic stimulation (i.e., thyroid hormone) (Tang et al., 2001). This would suggest that OP proliferation is regulated by extracellular environmental signals acting on molecular intrinsic targets, which are yet undefined, but ultimately linked to the cell cycle machinery.
A significant number of cells with an OP phenotype is still present in the adult CNS, both in the white and gray matter (Levine et al., 2001). Adult OPs proliferate much more slowly than perinatal OPs (Shi et al., 1998; Noble, 2000) and, after CNS damage, they display drastic changes in morphology and increases in NG2 immunoreactivity and mitotic rate (Dawson et al., 2000; Levine et al., 2001). In areas of demyelination, NG2+ adult OPs proliferate and increase in number, but do not appear to differentiate into myelinating oligodendrocytes. In vitro studies demonstrated lineage continuity between perinatal and adult OPs, however the molecular basis of their differences in cell proliferation and cell cycle length are still unknown (Noble, 2000).
Cell cycle progression, including G1-S phase transition, is governed by a complex network of biochemical interactions involving the activity of essential components, the cyclin-dependent kinases (cdks) (Morgan, 1997; Kohn, 1999; Roberts, 1999). Because of its intrinsic complexity, cell cycle progression can be regulated at many molecular levels (Sherr and Roberts, 1999) and distinct cdks play pivotal roles in G1-S transition in different neural cell types (Ross, 1996; Ohnuma et al., 2001). The precise roles and relative importance of specific cdks in different neural cell types is still a matter of debate. For example, cdk2 has been shown to play a prominent role in the regulation of Schwann cell proliferation (Tikoo et al., 2000), whereas the cdk4/6 signaling pathway is crucial for cortical neural progenitor cell proliferation (Ferguson et al., 2000). In astrocytes, cdk4 is an essential component of cell division (Holland et al., 1998). Clearly, the identification of critical cell type-specific cell cycle regulator(s) under defined physiological conditions will help elucidate mechanisms underlying the growth characteristics of different neural cells in development and disease.
Previous analysis in cultured OP cells demonstrated a decrease of cyclin E/cdk2 kinase activity after permanent cell cycle exit or after reversible cell cycle arrest triggered by extracellular signals (Ghiani et al., 1999a,b; Ghiani and Gallo, 2001). Conversely, cyclin D-associated cdk4/6 kinase activity was only marginally affected during permanent cell cycle withdrawal (Ghiani and Gallo, 2001). In contrast with proliferation, specific molecular targets involved in OP differentiation remain elusive, and little is known about the relationship between cell cycle-related pathways and OP differentiation (Tokumoto et al., 1999; Zezula et al., 2001; Huang et al., 2002).
In the present study, we first used a cell transfection procedure to modify cdk2 activity in cultured perinatal OP cells. We then extended our analysis of cyclin E/cdk2 expression and activity to a transgenic mouse selectively expressing the green fluorescent protein (GFP) in the oligodendrocyte lineage (Belachew et al., 2001; Yuan et al., 2002). Our results establish a direct causal link between cyclin E/cdk2 activity and OP cell proliferation, pointing to a pivotal function of cdk2 in changes of OP cell cycle kinetics.
MATERIALS AND METHODS
Cell culture. The CG-4 oligodendrocyte progenitor cell line was cultured as previously described (Louis et al., 1992). Twenty-four hours before transfection, CG-4 cells were plated onto poly-l-lysine-treated 12 mm glass coverslips at a density of 50,000 cells per coverslip.
Purified cortical OP cell cultures were prepared as previously described from embryonic day 20 Sprague Dawley rats, using a standard experimental protocol (McCarthy and de Vellis, 1980) with slight modifications (Ghiani et al., 1999a,b). OP cells were expanded for 2 weeks in poly-l-lysine-treated 100 mm plastic Petri dishes (passaged every 4 d) in DME-N1 supplemented with B104 conditioned medium (30%, v:v) (Louis et al., 1992). One day before transfection, OP cells were transferred onto poly-l-lysine-treated 12 mm glass coverslips at a density of 50,000 cells per coverslip in DME-N1medium in the presence of platelet-derived growth factor (PDGF, 10 ng/ml) (Upstate Biotechnology, Lake Placid, NY).
DNA constructs and transfection procedure. We used pCytomegalovirus (CMV)-neo-BAM vectors containing cDNAs of wild-type cdk2 and dominant-negative mutants of cdk2, cdk4, and cdk6 (van den Heuvel and Harlow, 1993). BamHI-BamHI fragments encompassing wild-type (wt)- and dominant negative (Dn)-cdk genes were subcloned into Bluescript SK (+/−) vectors (Stratagene, La Jolla, CA), to allow the following step of downstream cloning of an internal ribosomal entry site (IRES)–enhanced green fluorescent protein (EGFP) sequence (catalog #6064-1; Clontech, Palo Alto, CA). Wt-cdk2-, Dn-cdk2-, Dn-cdk4-, and Dn-cdk6-IRES–EGFP fragments were then inserted back into the native pCMV-neo-BAM backbone vector, to form plasmids named pCMV:WT2:IRES–GFP, pCMV:DN2:IRES–GFP, pCMV:DN4:IRES–GFP, and pCMV:DN6:IRES–GFP, respectively (Fig.1 A).
All plasmid constructs were introduced into cultured cells by liposomal transfection in 24-well dishes, using 0.3 μg of DNA and 0.8 μl of lipofectamine in OPTI-MEM (Life Technologies, Gaithersburg, MD) for each well (50,000 cells per coverslip). The duration of the transfection procedure was 6 hr. After transfection, the cells were placed in a mitogenic environment, i.e., DME-N1supplemented with PDGF (10 ng/ml) for primary OP cells, and DME-N1 supplemented with 30% (v:v) B104 conditioned medium for the CG-4 cells. The average efficiency of the transfection procedure (evaluated at 48 hr after transfection) in primary OP cells (GFP+ cells per total cells, mean ± SD) was 9.6 ± 4.2‰ for Dn-cdk2 construct, 13.4 ± 4.2‰ for Dn-cdk4, 16.1 ± 7.0‰ for Dn-cdk6, and 19.5 ± 7.8‰ for wt-cdk2. The rate of transfection in CG-4 cells was between 30 and 50% for all the constructs.
2′-3′-cyclic nucleotide 3′-phosphodiesterase–GFP transgenic mouse. The 2′-3′-cyclic nucleotide 3′-phosphodiesterase (CNP)–GFP transgenic mouse was generated by using the 3.7 kbXbaI-HindIII sequence, which contains the type I and II promoter core elements of the mouse CNP gene (Gravel et al., 1998; Belachew et al., 2001; Yuan et al., 2002). This fragment was ligated upstream of a 1 kb SmaI–XhoI fragment encompassing the coding sequence of EGFP gene (catalog #6064–1; Clontech) (Belachew et al., 2001; Yuan et al., 2002). A polyadenylation signal from bovine growth hormone was cloned downstream of the fusion gene. A 4.7 kb XbaI–XhoI fragment was excised, purified, and injected into fertilized oocytes from F1 females from FVB/NxCB6 genetic background. To identify positive transgenic animals, mice were analyzed by both PCR and Southern blot analysis. All animal procedures were performed according to National Institutes of Health guidelines. Immunohistochemical and electrophysiological characterization of GFP+ cells in tissue sections from transgenic mice revealed GFP expression in the oligodendroglial lineage (Yuan et al., 2002).
Isolation of oligodendroglial cells by fluorescence-activated cell sorting. Brains were dissected out from postnatal day 4 (P4)–P8, P15, and P30 CNP–GFP mice and from wild-type littermates. Meninges were removed. Brain tissues were cut into small pieces and incubated in enzyme solution (papain, 15 U/ml; type I deoxyribonuclease, 100 U/ml diluted in PBS solution) at 37°C for 20 min. Digested tissues were gently dissociated by passing through successive needles (gauge, 19 × 1 inch, 21 × 1 inch, and 23 × 1 inch). Cell suspensions were then filtered through a 70 μm cell strainer, centrifuged, and resuspended in DME-N1 plus 10% fetal bovine serum (FBS) at a density of 107 cells/ml for subsequent fluorescence-activated cell sorting (FACS). Cells were analyzed for light forward and side scatter using a FACS Vantage SE instrument (Becton Dickinson, San Jose, CA). To detect GFP fluorescence, cells were analyzed through a 530 nm bandpass filter, because the excitation wavelength was set at 488 nm of the argon–ion laser. Cells from negative littermates were used to set the background fluorescence, and a size threshold was used to gate out debris and small fragments. The sorting speed was 2000–4000 cells/sec. After FACS, cells were washed twice and harvested in ice-cold PBS solution.
Western blot and cdk2 activity assay. FACS-sorted cell pellets were resuspended in 100 μl of sample buffer [50 mm HEPES, pH 7.4, 150 mmNaCl, 1 mm EDTA, 2.5 mmEGTA, 1% Nonidet P-40, 1 mmNa3VO4, 4 μm NaF, 10 μg/ml leupeptin, 10 μg/ml aprotinin, 10 μg/ml pepstatin, and 1 mm4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF)] and lysed for 45 min on ice, followed by a brief sonication. The lysate was clarified by centrifugation at 12,000 rpm for 5 min, and the supernatant was collected. An aliquot was taken for protein determination using the Pierce (Rockford, IL) BCA protein assay kit. Protein extracts (18 μg) were resolved on a 4–20% mini-SDS polyacrylamide gel and transferred to Immobilon polyvinylidene difluoride membranes. Equal protein loading was verified by Ponceau S solution (Sigma, St. Louis, MO) reversible staining of the blots. Blots were processed as previously described (Ghiani et al., 1999a,b; Ghiani and Gallo, 2001), using anti-cdk2 (1:500, clone M2, sc-163; Santa Cruz Biotechnology, Santa Cruz, CA) and anti-cyclin E (1:500, clone M20, sc-481; Santa Cruz Biotechnology) antibodies.
Immunostaining in tissue sections and in cell cultures. For cdk2 immunohistochemical experiments, P6 and P30 sections from CNP–GFP transgenic mice were prepared as previously described (Yuan et al., 2002). Frozen sections were blocked at room temperature for at least 1 hr in blocking solution (1% BSA, 0.3% Triton X-100, and 20% normal goat serum in 1× PBS). Anti-cdk2 (1:200, clone M2, sc-163; Santa Cruz Biotechnology), anti- proliferating cell nuclear antigen (PCNA) (1:50, 05–347; Upstate Biotechnology, Lake Placid, NY), and NG2 (1:1000, AB5320; Chemicon, Temecula, CA) primary antibodies were diluted using carrier solution (1% BSA, 0.3% Triton X-100 in 1× PBS). Brain sections were incubated with primary antibody at 4°C overnight. For PCNA staining, tissue sections were treated with cold methanol at 20°C for 10 min before blocking. Rinse was performed in carrier solution at room temperature with three changes of solution every 15 min. Anti-rabbit and anti-mouse tetramethylrhodamine isothiocyanate- and Cy5-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA) were diluted 1:200 in carrier solution. Incubation was performed at room temperature for 1 hr and followed by three washes, as described above. Sections were then transferred into 1× PBS, mounted with Mowiol, and later imaged using a Nikon fluorescence microscope equipped with a Bio-Rad (Hercules, CA) MRC1024 laser confocal scanning system.
For immunocytochemistry of acutely purified OP cells from CNP–GFP mice or transfected CG-4 and OP cell cultures, NG2, O4, O1, and BrdU stainings were performed as previously described (Yuan et al., 1998,2002). For BrdU immunostaining, because the necessary alkaline denaturation of cell DNA (NaOH 0.07N) was partially quenching GFP fluorescence, GFP immunostaining (1:100; polyclonal anti-GFP, Santa Cruz Biotechnology) was performed before BrdU staining to allow accurate detection of transfected cells.
Cdk2 is a crucial mediator of OP cell cycle progression either in a pro- or anti-mitotic environment
To establish a causal link between cyclin E/cdk2 activity and OP proliferation, we used cell transfection procedure and performed gain/loss of function experiments. We designed plasmids containing dominant-negative mutants of cdk2, 4, and 6 (Dn-cdk2, 4 and 6) (Fig.1 A) placed under the control of the cytomegalovirus (CMV) promoter with GFP as a reporter gene. These dominant inhibitory mutants hold a single amino acid mutation located in the catalytic cleft of the enzyme that renders them inactive (van den Heuvel and Harlow, 1993). However, Dn-cdks retain their ability to bind their cyclin partners and inactivate the function of endogenous cdks by sequestering the required cyclins (van den Heuvel and Harlow, 1993).
Because previous studies on OP cell cycle kinetics were performed with rat tissue (Ghiani et al., 1999b; Ghiani and Gallo, 2001), transfection experiments were done with purified rat primary cortical OP cells prepared following an established standard culture protocol (McCarthy and de Vellis, 1980) and with progenitors from the rat oligodendroglial cell line CG-4 (Louis et al., 1992). Proliferation was assessed by using bromodeoxyuridine (BrdU) incorporation assays to quantitate GFP-expressing S-phase precursors. In the presence of mitogenic stimulation after transfection, Dn-cdk2 overexpression strongly decreased the number of dividing cells both in primary OP and CG-4 cultures by 58 and 49%, respectively (Fig. 1 B,C). Dn-cdk4 overexpression only inhibited proliferation of primary OP cells by 27%, whereas Dn-cdk6 was devoid of effect (Fig.1 B,C).
In parallel with loss of function experiments, which pointed to a crucial role of cdk2 in mitogen-induced OP proliferation, we also observed that overexpression of the wild-type version of cdk2 (wt-cdk2) in the presence of mitogens did not result in any significant gain of function (Fig. 1 B,C). Conversely, when deprived from platelet-derived growth factor (PDGF) stimulation, overexpression of wt-cdk2 was sufficient to maintain a higher percentage of primary OP cells within the cell cycle (Fig.2 A, E–G). These results are consistent with the notion that wt-cdk2 overexpression can compensate the downregulation of endogenous cdk2 that constitutively occurs in OP cells after PDGF withdrawal (Ghiani and Gallo, 2001).
We demonstrated that OP cell cycle progression in vitro andin situ (Gallo and Ghiani, 2000) could be influenced by extracellular signals, such as glutamate (Bergles et al., 2000; Gallo and Ghiani, 2000) and norepinephrine (Ghiani et al., 1999b) or by pharmacological K+ channel blockade (Ghiani et al., 1999a). In OP cells overexpressing wt-cdk2, reversible G1/S cell cycle arrest induced by kainic acid, isoproterenol, and tetraethylammonium in the presence of mitogen, was partially but significantly prevented (Fig.2 B–D).
OP cell differentiation is cdk2- and cell cycle number-independent
To determine whether modifications of OP cell cycle progression could also affect the differentiation process, we studied the emergence of O4 +pre-oligodendrocytes and O1 + mature oligodendrocytes in transfected cultures. Before transfection, primary OP cultures maintained in PDGF contained <5% of O4 + and no O1 + cells (>90% of the cells were GD3 + or A2B5+ under these conditions) (Ghiani et al., 1999b; Ghiani and Gallo, 2001) (data not shown). When OP cultures were maintained in the presence of PDGF for 4 d after transfection, the overexpression of Dn-cdk2, 4, 6, and wt-cdk2 did not influence oligodendrocyte differentiation, which remained very low (<2% of O1 + cells) (Fig. 3 A,B). In this mitogenic environment, thyroid hormone (T3) increased, as expected, the proportion of O4 + cells (Fig. 3, compare A, C) and induced the appearance of 15–20% of O1 + cells (Fig. 3, compare B, D). Furthermore, the presence of T3 did not stimulate growth-arrested OP cells overexpressing Dn-cdk2 or 4 to differentiate more rapidly than OP cells transfected with the control vector (Fig. 3 C,D). Finally, in mitogen-free medium, we also observed an increase of OP cell differentiation (Fig. 3, compare A, E and B, F), but overexpression of wt-cdk2, despite maintaining OP cells within the cell cycle (Fig. 2 A), did not induce a more immature phenotype (Fig. 3 E–G)
Developmental regulation of cyclin E/cdk2 expression and activity in OP cells in vivo
To quantitate cyclin E and cdk2 expression in the oligodendroglial lineage in vivo, we took advantage of a transgenic mouse expressing the GFP under the control of the CNP promoter (Belachew et al., 2001; Yuan et al., 2002). In this transgenic mouse, GFP expression was targeted to the oligodendroglial lineage (Belachew et al., 2001;Yuan et al., 2002). Acutely dissociated GFP+ oligodendroglial cells from CNP–GFP mice at different developmental stages (P4–P30) were purified by FACS. Unlike in tissue sections (Yuan et al., 2002), we observed that the proportion of NG2+ and O4+ cells among acutely purified CNP–GFP+ oligodendroglial cells was not significantly different between FACS-sorted cell suspensions derived from early (P4) or adult (P30) transgenic mice (Fig.4 A). On the other hand, the proportion of O1+ cells significantly increased between P4 and P30 (Fig. 4 A). The steady percentage of NG2+ and O4+ cells in the GFP+ cell suspensions between P4 and P30 most likely reflects the better survival of immature oligodendroglial cells during the dissociation procedure. At least in part, this could be caused by a loss of cell-to-cell contacts of mature oligodendrocytes in cell suspension. Importantly, as previously demonstrated (Yuan et al., 2002), all the NG2+ OP cells isolated from our CNP–GFP mice were GFP+. We never distinguished a subpopulation of NG2+/GFP-negative cells, as reported in the transgenic mouse strain expressing the GFP under the proteolipid protein (PLP) promoter (Mallon et al., 2002).
Western blot analysis of protein extracts from FACS-purified GFP+ cells revealed higher levels of cdk2 and cyclin E expression at the early postnatal period (P4–P8) (Fig.4 B,C). In FACS-purified CNP–GFP+ oligodendrocyte lineage cells, a 65% decrease of cdk2 and a 95% reduction of cyclin E protein expression were observed between P4–P8 and P30 (Fig.4 B,C). Similarly, we also observed a drastic (>50%) reduction in cyclinE/cdk2 activity between P4 and P30 (Fig.4 D,E).
To determine whether the developmental regulation of cyclinE/cdk2 occurred in OPs during in vivo development, we performed double immunostaining of GFP+ cells in tissue sections from CNP–GFP transgenic mice at P6 and P30. Immunodetection of NG2 chondroitin sulfate proteoglycan was used to identify both perinatal and adult OPs in the subventricular zone (SVZ), subcortical white matter, and cerebellar white matter (Levine et al., 2001). NG2+/GFP+ OP cells were immunostained with anti-cdk2 antibodies or with antibodies raised against the PCNA, a DNA replication-associated protein expressed in the nucleus of dividing cells, which was used as a marker of cell proliferation (Hyde-Dunn and Jones, 1997). In the SVZ at P6, the majority of NG2+/GFP+ OP cells displayed cytoplasmic staining with anti-cdk2 antibodies (Fig.5 A–D) and nuclear staining with anti-PCNA antibodies (Fig. 5 E–H). Consistent with the Western blot results shown in Figure 4, cell counting after immunohistochemical analysis of tissue sections demonstrated a 9- to 10-fold decrease in the percentage of NG2+/GFP+ OP cells expressing cdk2 or PCNA in the SVZ, subcortical white matter, and cerebellar white matter between P6 and P30 (Fig.6 A,B).
Double staining of GFP+ cells with anti-cdk2 and anti-PCNA in the same brain regions showed that, both at P6 and at P30, >90% of the GFP+/PCNA+cells were also cdk2+ (Figs.7,8 A). This finding indicates that, within the GFP+oligodendroglial cell population, proliferating cells always expressed cdk2, whereas most of the cdk2-negative cells were nonproliferating (Fig. 7). At P6, 80–90% of the cdk2-expressing cells were also PCNA+ (Fig. 8 B), but at P30, PCNA expression was detected only in 40–60% of the cdk2+ cells (Fig. 8 B).
Altogether, these data indicate that cdk2 is highly expressed in proliferating oligodendroglial cells in vivo throughout development, and that the significant downregulation of cdk2 expression in adult NG2+ OP cells is correlated with the event of cell cycle withdrawal.
In the present study, using in vitro transgenesis to overexpress dominant-negative mutants of different cdk genes, we provide evidence that cyclin E/cdk2 activity constitutes an essential component of the mechanism of mitogen-dependent OP cell proliferation. In parallel, we also demonstrate that a decrease of cdk2 activity causally underlies OP cell cycle exit triggered by extracellular signals acting at glutamate and β-adrenergic receptors, or K+ channels. Therefore, our results show that, in OPs, cell cycle pathways activated either in a pro- or in an anti-mitotic environment converge on the common molecular target cdk2.
The finding that cdk4/6 activity was less limiting than cdk2 in the regulation of OP entry into S-phase is consistent with the function of cyclinD/cdk4/6 complexes in different cell types (Jones and Kazlauskas, 2000). In fact, it is likely that reduced cdk4/6 activity is still sufficient to initiate pRb phosphorylation in early G1 (Jones and Kazlauskas, 2000). This would trigger release of E2F factors, transcriptional activation of cyclin E, and formation of the cyclin E/cdk2 complex, which in turn would maintain pRb phosphorylation until G1/S checkpoint (Jones and Kazlauskas, 2000).
Our experiments also demonstrate that cdk2-mediated regulation of OP cell cycle progression did not modify the time course of oligodendrocyte differentiation in the presence of PDGF. Hence, OP cell cycle arrest is a necessary but insufficient condition to trigger differentiation. The uncoupling between OP cell proliferation and differentiation has been demonstrated in cells treated with growth factors or overexpressing cdk inhibitors such as p18, p21, or p27 (Tikoo et al., 1998; Zezula et al., 2001; Tokumoto et al., 2002), however in all these experimental paradigms effects on multiple intracellular pathways could not be excluded. In our cdk2 loss-of-function or gain-of-function experiments, we are selectively targeting one component of the cell cycle machinery and demonstrating that by the specific modification of cdk2 activity we can arrest proliferation without altering differentiation.
Oligodendrocyte differentiation induced by PDGF withdrawal or T3 treatment was not modified by cdk2-mediated manipulations of OP cell cycle kinetics achieved by overexpressing Dn- or wt-cdk2. Thus, agents that stimulate OP differentiation may cause irreversible OP cell cycle withdrawal by complex and likely redundant mechanisms, which appear to be unrelated to cyclin E/cdk2 G1/S checkpoint. Furthermore, because cdk2-dependent cell cycle arrest did not result in a more immature OP phenotype, nor did cdk2-dependent cell cycle progression accelerate differentiation, our results also provide evidence that OP cells do not “count” the number of cell cycles before they differentiate. Thus, it appears that the yet undefined molecular cues which regulate the differentiation program of OP cells are cell cycle number-independent.
A recent study suggests that at least two molecularly distinct intracellular pathways may be involved in the timer mechanism that triggers OP cell irreversible cell cycle exit and differentiation (Tokumoto et al., 2001). T3-mediated OP cell differentiation appears to depend on a p53 family protein (for review, see Levrero et al., 2000), whereas spontaneous differentiation of OP cells in the presence of PDGF or OP cell differentiation induced by PDGF withdrawal are p53-independent (Tokumoto et al., 2001). T3 treatment and PDGF starvation can trigger changes in expression of Cip/Kip and Ink4 CKIs (Ghiani et al., 2001;Tokumoto et al., 2001) that in turn influence cdk2 activity. However, our data show that transgenic modulation of cdk2 activity, per se, did not influence T3- and PDGF withdrawal-induced differentiation. Therefore, the unknown endpoint mechanism of irreversible cell cycle exit associated with these pro-differentiating conditions are likely to involve distinct pathways that are independent from cyclin E/cdk2 activity. It can be hypothesized that nuclear transcription factors, such as helix-loop-helix proteins, known to stimulate oligodendrocyte differentiation (Kondo and Raff, 2000; Wang et al., 2001) could bypass cyclin E/cdk2 checkpoint and trigger irreversible OP cell cycle exit by directly affecting the phosphorylation state of pRb (Huang et al., 2002).
To correlate our findings on the essential role of cyclin E/cdk2 activity in OP proliferation in culture with OP proliferation in vivo, we took advantage of a transgenic mouse model in which GFP was selectively expressed in the oligodendroglial lineage throughout embryonic and postnatal development (Belachew et al., 2001; Yuan et al., 2002). After a FACS procedure that allowed us to purify GFP+ oligodendroglial cells at the progenitor stage (Yuan et al., 2002), we were able to demonstrate a downregulation of cyclin E/cdk2 activity and cyclin E and cdk2 protein expression in OP cells acutely isolated from adult brain.
We also performed a developmental immunohistochemical study of the proportion of cdk2-expressing cells within NG2+/GFP+ OP cells in the SVZ, cerebellar white matter, and subcortical white matter. We detected a drastic decrease of both proliferation and cdk2 expression in NG2+ OP cells between the perinatal period and adulthood. Furthermore, we observed that nearly all the few OP cells which remain proliferative in the adult brain sustained a high level of cdk2 expression. Conversely, about half of cdk2-expressing OPs of the adult brain were found to be nonproliferative, which suggests that the decrease of cyclin E/cdk2 complex activity and proliferation in postnatal OPs may be initially triggered by a reduction in cyclin E expression. This hypothesis is supported by our Western blot analysis of cyclin E and cdk2 expression in FACS-purified GFP+ cells (Fig. 4), indicating a more rapid developmental decline in cyclin E than in cdk2 levels. Thus, altogether our data strongly suggest that regulation of cyclin E and cdk2 expression and cyclin E/cdk2 activity within the oligodendroglial lineage in vivo causally underlies the progressive breakdown of OP cell proliferative potential that occurs during postnatal maturation (Wolswijk et al., 1990; Calver et al., 1998; Shi et al., 1998).
Despite an established lineage continuity, adult OP cells are known to differ from their neonatal counterparts in cell cycle time, rate of migration and time course of differentiation (Ffrench-Constant and Raff, 1986; Wolswijk et al., 1990; Shi et al., 1998). Our data show that, besides extrinsic environmental differences at the level of mitogen supply and response (Chan et al., 1990; Wolswijk and Noble, 1992; Engel and Wolswijk, 1996; van Heyningen et al., 2001), a major reason for the slower cell cycle kinetics of adult OPs relies on the intrinsic developmental regulation of the molecular machinery controlling cell cycle G1/S checkpoint. These findings may shed a new light on our understanding of the “quiescent” state of OP cells in human normal adult white matter and in multiple sclerosis lesions (Wolswijk, 1998; Chang et al., 2000;Maeda et al., 2001). Consistent with our hypothesis, a recent study bySim et al. (2002) demonstrated that the significant attenuation in remyelination efficiency observed in adult OPs as compared to their perinatal counterpart is at least in part due to an impairment in recruitment, that involves both proliferation and migration.
Our results not only provide insights into OP cell cycle decisions and the relationship between OP proliferation and differentiation, but also open new perspectives with respect of our understanding of: (1) the deregulation of proliferation underlying the genesis of OP-derived gliomas (Shoshan et al., 1999), and (2) the failure of remyelination in spite of the persistence of OP cells in CNS demyelinating lesions (Chang et al., 2000). We propose here that, besides acting on extrinsic cues, the molecular strategies aimed at modulating adult OP cell proliferation may need to focus on specifically targeting the cyclinE/cdk2 complex. With regard to genetic engineering techniques, as demonstrated in previous studies (Belachew et al., 2001; Yuan et al., 2002; Mallon et al., 2002), the use of CNP or PLP promoter-driven genes appears to be best suited for loss/gain-of-function experiments selectively targeting adult OP cells.
This work was supported by the National Institute of Child Health and Human Development Intramural Program. S.B. was supported by the Fonds National de la Recherche Scientifique (Belgium). A.A. was supported by a predoctoral fellowship from Consejo Nacional de Ciencia y Tecnologia (Mexico). We are grateful to Dr. Sander van den Heuvel for generously providing the pCMV-neo-BAM vectors containing the cDNAs for the dominant-negative mutants of cdk2, 4, and 6 and wild-type cdk2. We are particularly grateful to Cristina Ghiani for valuable discussion and for help during the initial phase of this project and to Ann Baron-Van Evercooren for valuable discussion. We thank Li-Jin Chew, Ramesh Chittajallu, Doug Fields, Cristina Ghiani, Chris McBain, Sergio Schinelli, and Beth Stevens for their critical comments on this manuscript.
Correspondence should be addressed to Dr. Vittorio Gallo, Center for Neuroscience Research, Children's Research Institute, Children's National Medical Center, 111 Michigan Avenue, Washington, DC 20010-2970. E-mail:.
S. Belachew's present address: Center for Cellular and Molecular Neurobiology, Department of Neurology, University of Liège, Belgium; and Center for Neuroscience Research, Children's Research Institute, Children's National Medical Center, Washington, DC 20010-2970.
A. Aguirre's and V. Gallo's present address: Center for Neuroscience Research, Children's Research Institute, Children's National Medical Center, Washington, DC 20010-2970.